Download - Reviews in Fluorescence Annual Volumes 2006 Reviews in Fluorescence

Reviews in Fluorescence 2006

Chris D. Geddes Joseph R. Lakowicz (Eds.)

Reviews in Fluorescence 2006

Springer

Chris D. Geddes Joseph R. Lakowicz Institute of Fluorescence Center for Fluorescence Spectroscopy University of Maryland University of Maryland

Biotechnology Institute Baltimore, MD 21201 Baltimore, MD 21201 USA USA [email protected] [email protected]

ISBN-10: 0-387-29342-6 ISBN-13: 978-0387-29342-4

Printed on acid-free paper.

© 2006 Springer Science-hBusiness Media, Inc. All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Springer Science-f-Business Media, Inc., 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights.

Printed in the United States of America. (TB/MVY)

9 8 7 6 5 4 3 2 1

springer.com

PREFACE This is the third volume in the Reviews in Fluorescence series. To

date, two volumes have been both published and well received by the scientific community. Several book reviews have also favorably described the series as an "excellent compilation of material which is well balanced from authors in both the US and Europe". Of particular mention we note the recent book review in JACS by Gary Baker, Los Alamos.

In this 3rd volume we continue the tradition of publishing leading edge and timely articles from authors around the world. We hope you find this volume as useful as past volumes, which promises to be just as diverse with regard to content.

Finally, in closing, we would like to thank Dr Kadir Asian for the typesetting of the entire volume and our counterparts at Springer, New York, for its timely publication.

Professor Chris D. Geddes Professor Joseph R. Lakowicz

August 20* 2005. Baltimore, Maryland, USA.

CONTRIBUTORS AND BIOGRAPHIES

Ousama M. A 'Amur. Boston University, Department of Biomedical Engineering, Boston, MA

Ousama M. A'Amar is Senior Research Associate at the Biomedical Engineering Department of Boston University, MA since 2001. He received his BS in Electronics Engineering in 1989. He received his MS (1993) and PhD (1997) in automatic control and digital signal processing from the National Polytechnic Institute of Lorraine, France. His MS and PhD research work focused on optical biomedical signals; mainly Autofluorescence and Induced-Fluorescence for cancer diagnosis and treatment. In 1996, he received the European Diploma in Medical Lasers from the University Nancy I, France and won the Young Researcher Prize of the French Society of Medical Lasers (SFLM). He worked as: Assistant Professor at the department of Biomedical Engineering, Amman University, Jordan (1998/1999-2002/2003); Postdoctoral Research Associate at the Bioscience Division of Los Alamos National Laboratory, NM (1999-2001). He works in the field of biomedical optics and his research activities focus on optical biomedical signals and optical spectroscopy for cancer diagnosis and Photodynamic Therapy.

Amit Agrawal. Emory University and Georgia Institute of Technology, Atlanta, GA

Amit Agrawal is a graduate student in the third year in the Nie research group in biomedical engineering department at Georgia institute of technology and Emory University. He has a Master's degree (5 yr) in chemistry from Indian institute of technology Kanpur. His research includes ultrasensitive biological detection inside living cells and developing material and nanoparticles for use in novel cancer diagnostics schemes. His work involves nanoparticle functionalization, delivery and targeting of nanoparticles and design of novel spectroscopic and imaging instrument set ups. He is the author of several conference papers and peer reviewed journal articles.

Onur AlptUrk. Department of Chemistry, Louisiana State University, Baton Rouge, LA

Christopher D. Anderson. Department of Surgery, Vanderbilt University Medical Center, Nashville, TN

vui CONTRIBUTORS

Renato J. Aguilera. Department of Biological Sciences, University of Texas at El Paso., El Paso, TX

Dr. Renato Aguilera obtained his Ph.D. from UC Berkeley in 1987 and was a professor at the University of California at Los Angeles from 1989 to 2002. Dr. Aguilera subsequently joined the biology department at the University of Texas at El Paso where he serves as the Director of the Biology Graduate Program and the RISE Research Scholars Program. He is also a member of the Board of Scientific Counselors of the National Institutes of Environmental Health and Safety (NIEHS). His work on the transcriptional regulation of the lymphocyte-specific Recombination Activating Genes (RAG) has been highly recognized and he has made significant contributions to others fields as well. Dr. Aguilera has many publications in high impact journals and holds a patent on an enzyme (DNase II) that is essential for engulfment-mediated DNA degradation. Most recently, Dr. Aguilera group has developed fluorescence-based assays for the rapid identification of cytotoxic and antimicrobial compounds generated by combinatorial chemistry.

Egidijus Auksorius. Imperial College London, U.K.

Richard K, P. Benninger. Imperial College London, U.K.

Axel Bergmann. Becker&Hickl GmbH, Nahmitzer Damm, Berlin, Germany.

Pieter de Beule. Imperial College London, U.K.

Irving J, Bigio. Boston University, Department of Biomedical Engineering and Electrical and Computer Engineering, Boston, MA

Franz Stanzel is head of the Bronchology Unit at the Asklepios Fachkliniken Munich-Gauting, Center for Respiratory Diseases and Thoracic Surgery, one of the biggest lung hospitals in Germany. He is a clinician of pulmonary medicine with a special interest on bronchology and the secretary of the Endoscopy Section of the German Society of Pneumology. Dr. Stanzel works since several years on interventional diagnostic and therapeutic procedures with the focus of lung cancer. The development of an autofluorescence bronchoscopy system together with Karl HauBinger braught early lung cancer into the center of his interest. He is an internationally accepted expert on autofluorescence bronchoscopy. Dr. Stanzel published a lot of scientific articles, papers, review articles and book chapters on bronchoscopy, interventional bronchology and fluorescent bronchoscopy

Rebecca A. Bozytn. University of Maryland School of Medicine, Baltimore, MD

CONTRIBUTORS ix

John D, Brennan. Department of Chemistry, McMaster University, Hamilton, Canada

John D. Brennan is an Associate Professor in the Department of Chemistry at McMaster University and holds the Canada Research Chair in Bioanalytical Chemistry. He has B.Sc, M.Sc. and Ph.D. degrees in analytical chemistry (fluorescence-based biosensors) from the University of Toronto and postdoctoral experience at the National Research Council of Canada in protein biophysics (time-resolved fluorescence). His current research primarily involves the entrapment of proteins within silica materials for the development of bioanalytical assays and devices. As part of this research, fluorescence methods are widely employed to examine the behaviour of proteins entrapped in silica. Dr Brennan has published over 80 scientific articles various aspects of protein immobilization and applications of fluorescence spectroscopy.

Denis Boudreau. Department of chemistry and Centre d'optique, photonique et laser, Universite Laval, Quebec, Canada

Denis Boudreau is Full Professor in the Department of Chemistry, and member of the Centre d'optique, photonique et laser (COPL) research center at Universite Laval, Quebec City, Canada. He has a B.Sc. from Universite de Sherbrooke, Canada, and a Ph.D. in analytical chemistry (plasma mass spectrometry) from the Universite de Montreal, Canada. He is the Editor of Spectrochimica Acta Electronica. Dr Boudreau has published over 40 scientific articles, papers, review articles and book chapters on various aspects of chemical trace analysis.

Jan Willem Borst. MicroSpectroscopy Centre, Wageningen University, Dreijenlaan Wageningen, The Netherlands

Ru-xiu Cai. Department of Chemistry, Wuhan University, Wuhan, China.

Cai Ruxiu is a professor. Supervisor of PhD, Director of the Group of Molecular spectroscopy (includes fluorescence, stopped-flow fluorescence. Catalytic kinetic fluorescence) in Analytical Science center at Wuhan University, China. She has a M.S from Wuhan University. She was visiting professor at Lawrence Berkeley National Laboratory, Energy and Environment Division, U.S.A in 1997, and worked at University of Arizona. Tucson in 1990, 1992. She is the committee of Editor of the Journal of Analytical Science. Professor Cai get continually National science foundation founding for six times.and has published 150 Scientific articles, papers, review articles and book chapters on the principles and applications of fluorescence spectroscopy, UV-Visible spectroscopy and kinetic Analysis.

Nils Calander, Physics Department, Chalmers University of Technology, Goteborg, Sweden.

X CONTRIBUTORS

Ravi S, Chart Department of Surgery, Vanderbilt University Medical Center, Nashville, TN

Ravi S. Chari is Associate Professor of Surgery and Cancer Biology, and Chief, Division of Hepatobiliary Surgery and Liver Transplantation at Vanderbilt University in Nashville, TN. He received his MD from the University of Saskatchewan, and his surgical training at Duke University. He is secretary-elect for the Society of University Surgeons and a member of the Scientific Committee of the International Hepato-Pancreato-Biliary Association (IHPB A) and was Program Chair of the 2004 IHPBA World Congress. He is a member of the Editorial Boards of the Journal of Surgical Research, HPB, World Journal of Surgery and Surgery. Dr Chari has published 100 scientific articles, papers, review articles and book chapters on liver and biliary surgery.

Herbert C. Cheung. Department of Biochemistry and Molecular Genetics, University of Alabama at Birmingham, Birmingham, AL.

Herbert C. Cheung is Professor of Biochemistry, University of Alabama at Birmingham School of Medicine. He holds joint appointments as Adjunct Professor in the Department of Physics and Senior Scientist in the Comprehensive Cancer Center. He received a master degree in physical chemistry from Cornell University, and a bachelor's degree in chemistry and a Ph. D. in physical chemistry and physics from Rutgers University. Following a period of industrial research in polymer physics, he was a senior fellow at the Cardiovascular Research Institute, University of California San Francisco, where he began a career in fluorescence spectroscopy and in the biophysics of muscle contraction. His current work is focused on use of FRET in both equilibrium and kinetic studies to study conformational switching in molecular motors and cardiac myofilaments.

Robert M, Clegg. Physics Department, University of Illinois Champaign-Urbana, Illinois.

Wen-Ji Dong. Department of Biochemistry and Molecular Genetics, University of Alabama at Birmingham, Birmingham, AL.

/aci#i/fl Z)'sowzfl. Department of Biological Sciences, Tata Institute of Fundamental Research Road, Mumbai, India.

RoryR. Duncan. Centre for Integrative Physiology, University of Edinburgh Medical School, Edinburgh, UK.

Christopher Dunsby. Imperial College London, U.K.

Guy Duportail. Faculte' de Pharmacie, Universite, Louis Pasteur de Strasbourg, Illkirch, France.

CONTRIBUTORS xi

Daniels. Elson. Imperial College London, U.K.

Jorge O. Escobedo. Department of Chemistry, Louisiana State University, Baton Rouge, LA.

Carol A, Fierke. University of Michigan, Ann Arbor, MI

PaulM. W, French. Imperial College London, U.K.

Xiaohu Gao. Emory University and Georgia Institute of Technology, Atlanta, GA

Xiaohu Gao is currently a postdoctoral fellow in the group of Dr. Shuming Nie. He earned his BS degree in chemistry from Nankai University (China), and his PhD degree in bioanalytical chemistry and nanotechnology from Indiana University - Bloomington. In the last 5 years, he published more than 20 papers, filed 4 patent applications, and delivered 15 invited talks at major conferences and academic institutions.

Neil Galletly. Imperial College London, U.K.

Anne Gibaud. Institut Curie, Paris, France.

Jean-Frangois Gravel. Department of chemistry and Centre d'optique, photonique et laser, Universite Laval, Quebec, Canada

Jean-Fran9ois Gravel is a Research Associate in the Department of Chemistry at Universite Laval, Quebec City, Canada. He has a B.Sc. in chemistry and a Ph.D. in analytical chemistry (laser spectrochemical analysis) from the Universite Laval, Canada. Dr Gravel has authored or co-authored over 15 scientific articles, papers, review articles and book chapters on laser spectrochemical analysis.

Laszlo Hegyi. Imperial College London, U.K.

Mark A, Hink. MicroSpectroscopy Centre, Wageningen University, Dreijenlaan Wageningen, The Netherlands

Arie van Hoek. MicroSpectroscopy Centre, Wageningen University, Dreijenlaan Wageningen, The Netherlands

Richard G.H. Immink. Laboratory for Biophysics, Wageningen University, Dreijenlaan, Wageningen, The Netherlands

Carey K. Johnson. Department of Chemistry, University of Kansas, Lawrence, KS.

xii CONTRIBUTORS

Kyu Kwang Kim. Department of Chemistry, Louisiana State University, Baton Rouge, LA

Mamata Kombrabail. Department of Chemical Sciences, Tata Institute of Fundamental Research, Mumbai, India.

G. Krishnamoorthy. Department of Chemical Sciences, Tata Institute of Fundamental Research, Mumbai, India.

G. Krishnamoorthy did his Masters in Science from University of Madras, India in 1974 and Ph.D. in Physical Biochemistry from the Tata Insititute of Fundamental Research, Mumbai, India in 1980. Subsequently he had postdoctoral research training at the Biochemistry department, Cornell University during 1981-84. Following his return to India, he joined the Faculty at the Tata Institute of Fundamental Research, Mumbai as Research Associate. At present he holds the position of Professor in the department of chemical sciences. His research interest covers application of time domain fluorescence spectroscopy to a variety of problems in macromolecular systems of biological and artificial origins. His current focus lies on the elucidation of site-specific dynamics in proteins, nucleic acids, DNA-protein complexes, cell membranes and cell interior with emphasis on correlation of dynamics and function.

Peter M. P. Lanigan. Imperial College London, U.K.

John Lever. Imperial College London, U.K.

Wei-Chiang Lin. Department of Neuro-Engineering, Miami Children's Hospital, Miami, FL

Zhi-hong Liu. Department of Chemistry, Wuhan University, Wuhan, China.

Bernard Malfoy. Institut Curie, Paris, France

C Mazzuca. Department of Chemical Sciences and Technologies, University of Roma Tor Vergata, Rome, Italy

James McGinty. Imperial College London, U.K.

YvesMely. Faculte' de Pharmacie, Universite, Louis Pasteur de Strasbourg, Illkirch, France.

P.M. Krishna Mohan. Department of Chemical Sciences, Tata Institute of Fundamental Research, Mumbai, India.

Jessica Montoya. Department of Biological Sciences, University of Texas at El Paso., El Paso, TX

CONTRIBUTORS xiii

Ian Munro. Imperial College London, U.K.

Nabanita Nag. Department of Chemical Sciences, Tata Institute of Fundamental Research, Mumbai, India.

Mark A. A, Neil. Imperial College London, U.K.

Isabella Nougalli-Tonaco. MicroSpectroscopy Centre, Wageningen University, Dreijenlaan Wageningen, The Netherlands

Shunting Nie. Emory University and Georgia Institute of Technology, Atlanta, GA

Shuming Nie is a Professor of Biomedical Engineering, Chemistry, Hematology, and Oncology, and also directs the program in cancer nanotechnology and bioengineering in the Winship Cancer Institute. He is the author of more than 80 peer-reviewed papers, the inventor of 12 patents, and the speaker of more than 250 invited talks and keynote lectures. After serving on the chemistry faculty at Indiana University for 8 years, he and his group moved to the Wallace H. Coulter Department of Biomedical Engineering at Georgia Tech and Emory in 2002. His research interest is primarily in the areas of biomolecular engineering and nanotechnology, with a focus on bioconjugated nanoparticles for cancer molecular imaging, molecular profiling, pharmacogenomics, and targeted therapy. Professor Nie has received many awards and honors including the Rank Prize (London, UK), the Georgia Distinguished Cancer Scholar Award, the Beckman Young Investigator Award, the National Collegiate Inventors Award, and the Distinguished Overseas Scholar Award. Professor Nie received his BS degree from Nankai University (China) in 1983, earned his MS and PhD degrees from Northwestern University (1984-1990), and did postdoctoral research both at Georgia Tech and Stanford (1991-1993).

John P, Nolan. La Jolla Bioengineering Institute, La Jolla, CA

John P. Nolan is a Senior Scientist and Principal Investigator at the La Jolla Bioengineering Institute, La Jolla, California. He has B.S. degrees from the University of Illinois, Urbana-Champaign in biology and chemistry and a Ph.D. in biochemistry from the Pennsylvania State University. He did post-doctoral work at Penn State and Los Alamos National Laboratory, where he was also a Technical Staff Member and Director of the NIH National Flow Cytometry Resource. Dr. Nolan's research interests are in the area of development and application of technology for the quantitative molecular analysis of biological systems.

xiv CONTRIBUTORS

Jamie K. Pero. Department of Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, NC.

Jamie K. Pero received an Honors B.S. degree in Chemistry in 2002 from the University of Utah and is currently a Ph.D. Candidate in Analytical Chemistry in the Department of Chemistry at the University of North Carolina at Chapel Hill. She has received several honors, has participated in a wide variety of community service and humanitarian projects, and has thus far published two scientific articles.

B. Pispisa. Department of Chemical Sciences and Technologies, University of Roma Tor Vergata, Rome, Italy

Basilio Pispisa is Full Professor of Physical Chemistry at the University of Roma Tor Vergata (Rome, Italy). He has a doctorate degree from the University of Pisa, and spent a few years in USA, at the Polymer Research Institute of the Polytechnic Institute of Brooklyn (New York). He is fellow of the American Peptide Society, of the Biophysical Society, of the Protein Society and of the European Peptide Society.

E. Shane Price. Department of Chemistry, University of Kansas, Lawrence, KS.

Todd P. Primm. Department of Biological Sciences, Sam Houston State University, Hunsville, TX.

T. Ramreddy. Department of Chemical Sciences, Tata Institute of Fundamental Research, Mumbai, India.

B,J, Rao. Department of Biological Sciences, Tata Institute of Fundamental Research Road, Mumbai, India

Jose Requejo-Isidro. Imperial College London, U.K.

Gang Ruan. Emory University and Georgia Institute of Technology, Atlanta, GA

Gang Ruan is a postdoctoral research fellow in the joint Department of Biomedical Engineering of Georgia Institute of Technology (School of Engineering) and Emory University (School of Medicine). He received his PhD from the National University of Singapore. He has published 9 scientific journal articles. Dr Ruan's current research interest is biomolecular engineering and bionanotechnology.

Oleksandr Rusin. Department of Chemistry, Louisiana State University, Baton Rouge, LA

CONTRIBUTORS xv

Ann Sandison. Imperial College London, U.K.

Brian D. Slaughter. Department of Chemistry, University of Kansas, Lawrence, KS.

Andrew M. Smith. Emory University and Georgia Institute of Technology, Atlanta, GA

Andrew Smith is a third-year graduate student in the biomedical engineering department at Georgia Institute of Technology and Emory University. He obtained his BS degree from Georgia Institute of Technology. His research interest is in the areas of biomolecular engineering and nanotechnology, with a particular focus on the development of near-infrared-emitting quantum dots for molecular profiling and imaging applications. He is the author of seven publications in the last two years.

Steven A. Soper. Department of Chemistry, Louisiana State University, Baton Rouge, LA

Steven A. Soper, Ph.D. is currently a professor of Chemistry and Mechanical Engineering at Louisiana State University (LSU) in Baton Rouge, LA. Steve received his Ph.D. from the University of Kansas in 1989 and then, was a postdoctoral fellow at Los Alamos National Laboratory where he was involved in developing fluorescence single molecule detection for high throughput DNA sequencing. He joined the faculty at LSU in 1991 and has been working on new fluorescence detection strategies for the analysis of DNA.

Pat Soutter. Imperial College London, U.K.

Gordon W. Stamp. Imperial College London, U.K.

L. Stella. Department of Chemical Sciences and Technologies, University of Roma Tor Vergata, Rome, Italy

Robert M. Strongin. Department of Chemistry, Louisiana State University, Baton Rouge, LA

Clifford Talbot. Imperial College London, U.K.

Nancy L. Thompson. Department of Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, NC.

Nancy L. Thompson received a Ph.D. in Physics from the University of Michigan at Ann Arbor in 1982 and was then a Damon Runyon - Walter Winchell Postdoctoral Fellow in the Department of Chemistry at Stanford University. She has been a member of the Faculty of the Department of Chemistry at the University of North Carolina at Chapel Hill since 1985 where

xvi CONTRIBUTORS

she currently holds the position of Professor of Chemistry. She has received several honors including a National Science Foundation Presidential Young Investigator Award, the Margaret Oakley Dayhoff Award from the Biophysical Society, a Dreyfus Teacher-Scholar Award, and the Hettleman Prize from the University of North Carolina at Chapel Hill. She has served on a variety of Editorial Boards and published numerous scientific articles in the fields of membrane biophysics and fluorescence microscopy.

Richard B, Thompson. University of Maryland School of Medicine, Baltimore, MD

Dr. Thompson was bom in Ohio and raised north of Chicago, Illinois. He received a B.A. in Biology from Northwestern University; while there, he began biochemical studies with E. Margoliash. He received the Ph.D. in Biochemistry from the University of Illinois in Urbana-Champaign working under the direction of Thomas O. Baldwin. He trained as a postdoctoral fellow in the laboratory of Joseph Lakowicz at the University of Maryland at Baltimore before moving to the U.S. Naval Research Laboratory as a National Research Council Associate. At the Naval Research Laboratory he began work on fluorescence-based biosensors under Paul Schoen and subsequently became a Supervisory Research Chemist under the direction of Frances Ligler; he received a Navy Special Act Award for activity related to Operation Desert Storm. He joined the faculty of the University of Maryland School of Medicine in the Department of Biochemistry and Molecular Biology where he is now Associate Professor. He serves on the Editorial Boards of the Journal of Fluorescence and the Journal of Biomedical Optics, as well as panels for the National Research Council, National Institutes of Health, National Science Foundation, and other agencies.

Dina Tleugahulova. Department of Chemistry, McMaster University, Hamilton, Canada

Dina Tleugahulova is Postdoctoral Fellow in the Department of Chemistry at McMaster University, Hamilton, Canada. She has a B.Sc. and M.Sc in physical chemistry from Moscow State University, Russia and a Ph.D. in biology from the University of Havana, Cuba. Dr. Tleugahulova has published scientific articles on protein separation, pharmaceutical analysis and principles and applications of fluorescence anisotropy.

Khuong Truong. IMSTAR. Paris France

Jay R. Unruh. Department of Chemistry, University of Kansas, Lawrence, KS.

Andrew Wallace. Imperial College London, U.K.

Armando Varela-Ramirez. Department of Biological Sciences, University of Texas at El Paso., El Paso, TX

CONTRIBUTORS xvii

M. VenanzL Department of Chemical Sciences and Technologies, University of Roma Tor Vergata, Rome, Italy.

Li Zhu. Department of Chemistry, Louisiana State University, Baton Rouge, LA.

Li Zhu came to LSU in the fall of 2000 as a Ph.D. student from Nankai University in Tianjin, China. Li's dissertation work focused on developing near-IR time-resolved fluorescence detection for multiplexing applications in genomics. She received her Ph.D. in the fall of 2005 and is working at GE Global Research Center in Niskayuna, NY.

Antonie J. W.G. Visser. MicroSpectroscopy Centre, Wageningen University, Dreijenlaan Wageningen, The Netherlands

Nicolas Vogt. Institut Curie, Paris, France

Jun Wang. Department of Chemistry, Wuhan University, Wuhan, China.

Weihua Wang. Department of Chemistry, Louisiana State University, Baton Rouge, LA

Xiangyang Xu, Department of Chemistry, Louisiana State University, Baton Rouge, LA

CONTENTS

PREFACE V

CONTRIBUTORS vii

1. THE HISTORY OF FRET

Robert M. Clegg

1.1. INTRODUCTION 1 1.2. PRELUDE TO THE HISTORICAL BACKGROUND 3

1.2.1. The End of the Dark Ages: the Pre-Field Era 3 1.2.2. Middle Ages: Experiments That Eventually Changed Our World View 5 1.2.3. Renaissance: Enter the Theory of Electrodynamics and Fields 6 1.2.4. The Beginning of the Modem Age: The Field Surrounding an Oscillating Charge 7

1.3. FIELDS, SPECTROSCOPY AND QUANTUM MECHANICS 9 1.3.1. Fields 9 1.3.2. Quantum Mechanics and Spectroscopy 9

1.4. THE FIRST EXPERIMENTAL OBSERVATION OF ENERGY TRANSFER AT A DISTANCE - SENSITIZED LUMINESCENCE IN VAPORS 11 1.4.1. Sensitized Fluorescence 11 1.4.2. Spectroscopic and Collisional Cross-Sections in Vapors 12

1.5. THE FIRST QUANTUM MECHANICAL THEORY OF ENERGY TRANSFER 12 1.5.1. A Few of the Pre-Quantum Theories that calculated the Spectroscopic Cross-Sections of Atomic Vapors 13 1.5.2. Some Details of the Kallmann and London Paper 14

1.6. LONDON FORCES (VAN DER WAALS) AND DEB YE AND KEESOM INTERACTIONS 17 1.6.1. London Interactions: Induced-Dipole-Induced-Dipole 17 1.6.2. Keesom and Debye Interactions: Dipole-Dipole and Dipole Induced-Dipole 19

1.7. FRET BETWEEN ORGANIC CHROMOPHORES IN CONDENSED SYSTEMS 20

XIX

XX CONTENTS

1.7.1. Experimental Observations of Energy Transfer in Solution .. 20 1.7.2. The Theories of J. PerrinandF. Perrin 21 1.7.3. A Derivation of the Perrins' Estimated Distances for Two Electron Oscillators in Exact Resonance 23 1.7.4. The Contribution of W. Arnold and J.R. Oppenheimer to FRET in Photosynthesis 30

1.8. FORSTER'S SEMINAL CONTRIBUTION: THE MODERN, PRACTICAL DEPICTION OF FRET (FORSTER RESONANCE ENERGY TRANSFER) 37

1.9. MATURATION OF FRET 41 1.10. EPILOGUE 42 1.11. ACKNOWLEDGMENTS 42 1.12. REFERENCES 43

2. TRICHOGIN TOPOLOGY AND ACTIVITY IN MODEL MEMBRANES AS DETERMINED BY FLUORESCENCE SPECTROSCOPY 47

B. Pispisa, L. Stella, C. Mazzuca, and M. Venanzi

2.1. INTRODUCTION 47 2.2. THE PROPERTIES OF THE FLUORESCENT ANALOGS 48 2.3. AGGREGATION IN WATER 50 2.4. WATER-MEMBRANE PARTITION AND AGGREGATION 53 2.5. BIOACTIVITY: MECHANISM OF MEMBRANE

PERTURBATION 60 2.6. POSITION OF TRICHOGIN INTO THE MEMBRANE:

TRANSLOCATION, DEPTH-DEPENDENT QUENCHING, AND DISTRIBUTION ANALYSIS 61 2.6.1. Peptide Translocation 61 2.6.2. Depth-Dependent Quenching and Peptide Distribution Analysis 64

2.7. PEPTIDES ORIENTATION INSIDE THE MEMBRANE 65 2.8. CONCLUDING REMARKS 67 2.9. ACKNOWLEDGMENTS 68 2.10. REFERENCES 68

3. THEORY OF METAL-FLUOROPHORE INTERACTIONS 71

Nils Calander

3.1. INTRODUCTION 71 3.2. SURFACE PLASMON RESONANCE 72

3.2.1. Plasma Oscillations 72 3.2.2. Surface Plasmon Resonances 73

CONTENTS xxi

3.3. THEORY OF SURFACE PLASMON RESONANCE AT PLANAR STRUCTURES 74 3.3.1. Basic Theory 75 3.3.2. Simulations 78 3.3.3. Conclusions of surface plasmon resonance at planar structures 84

3.4. THEORY OF SURFACE-PLASMON RESONANCE OPTICAL FIELD ENHANCEMENT AT PROLATE SPHEROIDS 85 3.4.1. The Field Enhancement at Spheroids 88 3.4.2. Conclusion of Surface-Plasmon Resonance Optical-Field Enhancement at Prolate Spheroids 94

3.4.3. Solving the Maxwell's Equations in Prolate Spheroidal Coordinates 94

3.5. OPTICAL TRAPPING OF SINGLE FLUORESCENT MOLECULES BY SURFACE PLASMON RESONANCE 98

3.6. REFERENCES 104

4. CURRENT DEVELOPMENT IN THE DETERMINATION

OF INTRACELLULAR NADH LEVEL 107

Zhi-hong Liu, Ru-xiu Cai, and Jun Wang

4.1. INTRODUCTION TO NADH 107 4.2. SIGNIFICANCE OF DETERMINING INTRACELLULAR NADH LEVEL 108 4.3. DETERMINATION OF INTRACELLULAR NADH LEVEL 110

4.3.1. Enzymatic Assays I l l 4.3.2. Fluorometric Methods 112 4.3.3. Micro-Fluorescence Photometry 116 4.3.4. Laser Scanning Confocal Microphotographics 117 4.3.5. Two-Photon Excitation Micrographics 117

4.4. REGULATION OF INTRACELLULAR NADH LEVEL 118 4.4.1. Effect of Vitamins on Intracellular NADH Level 118 4.4.2. The Time Course of Intracellular NADH in Yeast Apoptosis 119

4.5. REFERENCES 123

5. PREDICTION OF THERMAL TISSUE DAMAGE USING

FLUORESCENCE SPECTROSCOPY 125

Christopher D. Anderson, Wei-Chiang Lin, and Ravi S. Chari

5.1. INTRODUCTION 125 5.2. FLUORESCENCE SPECTROSCOPY TO DETECT THERMAL

TISSUE DAMAGE 126 5.3. MEASUREMENTS OF FLUORESCENCE SPECTRA IN VIVO... 127

xxii CONTENTS

5.4. SPECTRAL CORRELATES TO THERMAL DAMAGE 130 5.5. FLUORESCENCE SPECTRA CORRELATE WITH HISTOLOGIC

TISSUE DAMAGE 132 5.7. DETECTION OF ABSOLUTE CELL DEATH 133 5.8. CONCLUSIONS AND FUTURE DIRECTIONS 136 5.9. REFERENCES 137

6. DETECTION OF BIOLOGICAL THIOLS 139

Jorge O. Escobedo, Oleksandr Rusin, Weihua Wang, Onur Alpturk, Kyu Kwang Kim, Xiangyang Xu, Robert M. Strongin

6.1. INTRODUCTION 139 6.2. HOMOCYSTEINE METABOLISM 141 6.3. NEW PERSPECTIVES ON HOMOCYSTEINE'S ROLE IN

DISEASE 143 6.4. OVERVIEW OF KNOWN METHODS FOR BIOLOGICAL THIOL

DETECTION 144 6.5. DETECTION OF CYSTEINE AND HOMOCYSTEINE 147 6.6. HIGHLY SELECTIVE DETECTION OF CYSTEINE AND SITE-

SPECIFIC PEPTIDE LABELING 149 6.7. HIGHLY SELECTIVE DETECTION OF HOMOCYSTEINE 151 6.8. AUTOMATED POST-COLUMN DETECTION OF CYSTEINE AND

HOMOCYSTEINE 156 6.9. BIOTHIOL DETECTION BASED ON SIMPLE ARRAYS 156 6.10. CONCLUSIONS 158 6.11. ACKNOWLEDGMENT 158 6.12. REFERENCES 158

7. FLUORESCENT BRONCHOSCOPY 163

Franz Stanzel

7.1. INTRODUCTION 163 7.2. PHENOMENON AND TECHNIQUES 165

7.2.1. Drug-Induced Fluorescence 165 7.2.2. Autofluorescence 165

7.3. INVESTIGATIONS AND DATA 167 7.3.1. Drug-Induced Fluorescence 167 7.3.2. Autofluorescence Bronchoscopy 168

7.4. DISCUSSION 169 7.5 CONCLUSIONS 174 7.6. REFERENCES 175

CONTENTS xxiii

8. QUANTUM DOTS AS FLUORESCENT LABELS FOR

MOLECULAR AND CELLULAR IMAGING 181

Gang Ruan, Amit Agrawal, Andrew M. Smith, Xiaohu Gao, and Shuming Nie

8.1. INTRODUCTION 181 8.2. PROBE DEVELOPMENT 182 8.3. NOVEL OPTICAL PROPERTIES 185 8.4. DELIVERY OF QD PROBES INTO CELLS 187 8.5. APPLICATIONS IN INTRACELLULAR IMAGING 189

8.5.1 Cellular staining 189 8.5.2. Intracellular studies 189

8.6. ACKNOWLEDGMENT 191 8.7. REFERENCES 191

9. MOLECULAR ANALYSIS USING MICROPARTICLE-BASED

FLOW CYTOMETRY 195

John P. Nolan

9.1. INTRODUCTION 195 9.2. OPTICAL MEASUREMENTS USING FLOW CYTOMETRY 196 9.3. SOLID PHASE ASSAYS USING MICROPARTICLES 197 9.4. DETECTION AND SENSOR APPLICATIONS 200 9.5. MOLECULAR INTERACTIONS AND FUNCTION 202

9.5.1. Enzyme-Substrate Interactions 202 9.5.2. Ligand-Receptor Interactions 205 9.5.3. Protein Immobilization 207

9.6. GENETIC ANALYSIS 208 9.7. SUMMARY AND FUTURE DIRECTIONS 210 9.8. ACKNOWLEDGEMENT 210 9.9. REFERENCES 211

10. TOTAL INTERNAL REFLECTION-FLUORESCENCE

CORRELATION SPECTROSCOPY 215

Nancy L. Thompson, and Jamie K. Pero

10.1. ABSTRACT 215 10.2. INTRODUCTION 216 10.3. CONCEPTUAL BASIS AND EXPERIMENTAL DESIGN 217 10.4. THEORETICAL MODELS FOR DATA ANALYSIS 221 10.5. APPLICATIONS 225 10.6. FUTURE DIRECTIONS 229 10.7. ACKNOWLEDGEMENTS 233 10.8. REFERENCES 233

xxiv CONTENTS

11. FLUORESCENCE PROBES OF PROTEIN DYNAMICS AND CONFORMATIONS IN FREELY DIFFUSING MOLECULES 239

Carey K. Johnson, Brian D. Slaughter, Jay R. Unruh, and E. Shane Price

11.1. INTRODUCTION 239 11.2. FLUORESCENCE CORRELATION SPECTROSCOPY TO

PROBE PROTEIN DYNAMICS 240 11.2.1. PCS Measurements of Intramolecular Dynamics 241 11.2.2. PCS Cross-Correlation Measurements 242 11.2.3. PCS of Calmodulin 243

11.3. BURST-INTEGRATED SINGLE-MOLECULE ANALYSIS 246 11.4. TIME-REVOLVED FLUORESCENCE MEASUREMENTS 248

11.4.1. CaM Conformational Substates by Ensemble Time-Resolved Fluorescence Measurements 248 11.4.2. Associated-Anisotropy Analysis to Assess the Influence of Dye-Protein Interactions 252

11.5. CONCLUSION 255 11.6. ACKNOWLEDGEMENT 255 11.7. REFERENCES 256

12. BIOLOGICAL APPLICATION OF FLIM BY TCSPC 261

Axel Bergmann, and Rory R. Duncan

12.1. INTRODUCTION 261 12.2. PHYSICAL BACKGROUND OF FLUORESCENCE LIFETIME

IMAGING 262 12.2.1 Fluorescence Lifetime as a Separation Parameter 262 12.2.2. The Fluorescence Lifetime as an Indicator of the Local Environment 263 12.2.3 Fluorescence Resonance Energy Transfer 264

12.3. THE LASER SCANNING MICROSCOPE 265 12.3.1 Suppressionofout-of-focus light 265 12.3.2 Scan Rates 265 12.3.3 Two-Photon Excitation with Direct Detection 266

12.4. REQUIREMENTS FOR FLUORESCENCE LIFETIME IMAGING IN SCANNING MICROSCOPES 267 12.4.1 Efficiency 267 12.4.2 Principle of Time-Correlated Single Photon Counting 268 12.4.3 Imaging by Multi-Dimensional TCSPC 269

CONTENTS XXV

12.5. BIOLOGICAL APPLICATION OF FLIM 270 12.5.1 Biological FLIM Data Acquisition 270 12.5.2 TCSPC-FLIM 270 12.5.3 FLIM Data Analysis and FRET Calculations 271 12.5.4 FLIM to Measure FRET In Cells 271 12.5.5 Dual channel FLIM 273

12.6. FUTURE PERSPECTIVES 273 12.7. REFERENCES 274

13. TIME-RESOLVED FLUORESCENCE ANISOTROPY APPLIED TO SILICA SOL-GEL GROWTH AND SURFACE MODIFICATION. 277

Dina Tleugabulova, and John D. Brennan

13.1. INTRODUCTION 277 13.1.1 Characterization of Silica Growth and Modification 278

13.2. SURVEY OF TRFA THEORY 280 13.2.1 What Is Measured in TRFA? 281 13.2.2 TRFA Data Analysis 282 13.2.3 Rotational Diffusion of Probes in Non-Interacting Environments 284 13.2.4 Restricted Dynamics in the Presence of Probe-Host Interactions 287

13.3. PARTICLE GROWTH STUDIES 290 13.3.1 Particle Growth in DGS Sols 290 13.3.2 Particle Growth in SS Sols 295

13.4. MONITORING SILICA SURFACE MODIFICATION 296 13.4.1 Background 296 13.4.2 Ludox 297 13.4.3 Monitoring Adsorption by TRFA 297

13.5. CONCLUSIONS AND OUTLOOK 304 13.6. ACKNOWLEDGEMENTS 305 13.7. REFERENCES 305

14. DYNAMICS OF DNA AND PROTEIN-DNA COMPLEXES VIEWED THROUGH TIME-DOMAIN FLUORESCENCE 311

Nabanita Nag, T. Ramreddy, Mamata Kombrabail, P.M. Krishna Mohan, Jacinta D'souza, B.J. Rao, Guy Duportail, Yves Mely, and G. Krishnamoorthy

14.1. INTRODUCTION 311 14.2. FLUORESCENCE PROBES FOR DNA DYNAMICS 313 14.3. PROBING DNA DYNAMICS WITH NON-SPECIFIC PROBES 316

xxvi CONTENTS

14.3.1. DNA condensation 316 14.3.2. YOYO-1 as an indicator of DNA condensation 317 14.3.3. Structure and dynamics of condensed DNA 318 14.3.4. DNA condensation by the nucleocapsid protein probed by YOYO-1 Fluorescence 323 14.3.5. DNA Dynamics in Chromosomes from Picogreen Fluorescence 323

14.4. DNA DYNAMICS FROM SITE-SPECIFIC FLUORESCENCE PROBES 325 14.4.1. DNA dynamics in RecA-DNA filaments 326 14.4.2. Position-dependent DNA dynamics 328 14.4.3. Mismatch recognition and DNA dynamics 329

14.5. CONCLUSIONS 332 14.6. ACKNOWLEDGMENTS 332 14.7. REFERENCES 332

15. PROTEIN-PROTEIN INTERACTIONS IN VIVO: USE OF BIOSENSORS BASED ON FRET 341

Jan Willem Borst, Isabella Nougalli-Tonaco , Mark A. Hink, Arie van Hoek, Richard G.H. Immink, and Anionic J.W.G. Visser

15.1. INTRODUCTION 341 15.2. FRET COMBINATIONS IN CELL BIOLOGY 343 15.3. FRET SENSORS 344

15.3.1 Cameleons (Ycam) 344 15.3.2 Caspase sensor 345 15.3.3 FLAME 345

15.4. INTENSITY BASED FRET IMAGING 346 15.4.1 Confocal and wide-field FRET imaging 346 15.4.2 Spectral imaging 346 15.4.3 Acceptor photo-bleaching 346

15.5. FLUORESCENCE LIFETIME IMAGING MICROSCOPY (FLIM) 348 15.5.1 FLIM setup 349 15.5.2 FLIM analysis 349

15.6. APPLICATIONS WITH PLANT TRANSCRIPTION FACTORS 350 15.6.1 Sub-cellular localization via confocal microscopy 350 15.6.2 Molecular interaction imaging via FRET-FLIM 351 15.6.3 Molecular interaction imaging via FRET-FLIM 354

15.7. ACKNOWLEDGMENTS 355 15.8. REFERENCES 355

CONTENTS xxvii

16. SPECTROSCOPY FOR THE ASSESSMENT OF MELANOMAS . 359

Ousama M. A'Amar, and Irving J. Bigio

16.1. INTRODUCTION 359 16.2. SKIN MELANOMA 362 16.3. FLUORESCENCE SPECTROSCOPY 363

16.3.1. Instrumentation 363 16.3.2. Melanoma Diagnosis by Autofluorescence 363 16.3.3. Melanoma Diagnosis with exogenous fluorophores 368

16.4. ELASTIC SCATTERING SPECTROSCOPY 369 16.4.1. Principles of Elastic Scattering Spectroscopy 369 16.4.2. Instrumentation 371 16.4.3. Preclinical Trials 373 16.4.4. Clinical Studies 374

16.5. CORRECTION OF FLUORESCENCE ESf TURBID MEDIA 376 16.6. CONCLUSIONS 378 16.7. REFERENCES 379

17. QUANTITATIVE FLUORESCENCE HYBRIDIZATION USING AUTOMATED IMAGE CYTOMETRY ON INTERPHASE NUCLEI 387

Khuong Truong, Anne Gibaud, Nicolas Vogt, and Bernard Malfoy

17.1. INTRODUCTION 387 17.2. CHROMOSOME IMBALANCES IN HUMAN DISEASES 388

17.2.1. Cancers 388 17.2.2. Constitutional diseases 388

17.3. EXPERIMENTAL APPROACHES FOR THE IN SITU DETERMINATION OF CHROMOSOME IMBALANCES 389 17.3.1. Metaphase chromosomes 389 17.3.2. Interphase chromosomes 390

17.4. QUANTITATIVE FISH BY AUTOMATED IMAGE CYTOMETRY ON INTERPHASE NUCLEI 390 17.4.1. Chromosome 3 arms imbalances inbronchic cancers 395 17.4.2. Prenatal Diagnosis of Trisomy 21 395

17.5. CONCLUSIONS AND PERSPECTIVES 395 17.6. REFERENCES 397

xxviii CONTENTS

18. IMPORTANCE OF MEASURING FREE ZINC IN CELLS 399

Rebecca A. Bozym, Richard B. Thompson, and Carol A. Fierke

18.1. INTRODUCTION 399 18.2. TSQ DERIVATIVES 402 18.3. FLUORESCENT INDICATORS BASED ON FLUORESCEIN .. 402

18.3.1. The Zinpyr family 402 18.3.2. The Zinspy Family 404 18.3.3. The ZnAFs 405

18.4. ZINC INDICATORS BY MOLECULAR PROBES 406 18.5. A ZINC INDICATOR BASED ON BENZOXAZOLE 408 18.6. LANTHANIDE CHEMOSENSORS FOR ZINC 408 18.7. EXCITED-STATE INTRAMOLECULAR PROTON

TRANSFER 409 18.8. PEPTIDES AS ZINC INDICATORS 410 18.9. CARBONIC ANHYDRASE AS A BIOSENSOR FOR ZINC 411 18.10. CONCLUSION 414 18.11. REFERENCES 415

19. LASER-BASED DETECTION OF ATMOSPHERIC HALOCARBONS 421

Jean-Francois Gravel, and Denis Boudreau

19.1. ABSTRACT 421 19.2. INTRODUCTION 422 19.3. LASER INDUCED BREAKDOWN SPECTROSCOPY 423 19.4. LASER PHOTOFRAGMENTATION

FRAGMENT DETECTION 425 19.5. LONG RANGE REMOTE SENSING OF HALOCARBONS BY NON-LINEAR LASER PROPAGATION 431 19.6. FUTURE DIRECTIONS 438 19.7. ACKNOWLEDGEMENTS 438 19.8. REFERENCES 438

20. FRET STUDIES OF CONFORMATIONAL TRANSITIONS IN

PROTEINS 445

Herbert C. Cheung, and Wen-Ji Dong

20.1. INTRODUCTION 445

20.2. CALCIUM ACTIVATION OF CARDIAC MUSCLE 446 20.2.1. Equilibrium Conformation of Cardiac Troponin 446

20.3. KINETICS OF CONFORMATIONAL TRANSITIONS IN cTN . 453

CONTENTS xxix

20.4. CONFORMATION OF N-DOMAIN OF cTnC IN MYOFILAMENT 455

20.5. FRET-BASED CONSTRUCTION OF MOLECULAR MODELS 458 20.6. NUCLEOTIDE-DEPENDENT KINESIN CONFORMATIONS . 460 20.7. SUMMARY 461 20.8. ACKNOWLEDGEMENTS 462 20.9. REFERENCES 462

21. GREEN FLUORESCENT PROTEIN AS A BIOSENSOR FOR TOXIC COMPOUNDS 463

Renato J. Aguilera, Jessica Montoya, Todd P. Primm, and Armando Varela-Ramirez

2L1 ABSTRACT 463 2L2. BRIEF OVERVIEW ON THE PROPERTIES OF GFP 464 21.3. GFP AS A BIOSENSOR 464 2L4. GFP-BASED TOXICITY ASSAYS IN MULTICELLULAR

ORGANISMS 466 21.5. RECENT GFP-ASSAYS FOR DRUG DISCOVERY 467 21.6. USING THE HELA-GFP ASSAY TO DETERMINE THE

CYTOTOXICITY OF ANTIBACTERIAL COMPOUNDS 468 21.7. LARGE-SCALE SCREENING OF COMPOUNDS ON

EUKARYOTIC AND PROKARYOTIC CELLS 470 21.8. SUMMARY 472 21.9. ACKNOWLEDGEMENTS 473 21.10. REFERENCES 474

22. MULTIDIMENSIONAL FLUORESCENCE IMAGING APPLIED TO BIOLOGICAL TISSUE 477

Daniel S. Elson, Neil Galletly, Clifford Talbot, Jose Requejo-Isidro, James McGinty, Christopher Dunsby, Peter M. P. Lanigan, Ian Munro, Richard K. P. Benninger, Pieter de Beule, Egidijus Auksorius, Laszlo Hegyi, Ann Sandison, Andrew Wallace, Pat Soulier, Mark A. A. Neil, John Lever, Gordon W. Stamp, and Paul M. W. French

22.1. INTRODUCTION 477 22.2. FLUORESCENCE LIFETIME 480

22.2.1. Fluorescence lifetime of endogenous fluorophores 482 22.3. FLUORESCENCE LIFETIME DETERMINATION 483

22.3.1. Single-point measurement of fluorescence lifetime 484 22.3.2. Fluorescence Lifetime Imaging (FLIM) 487 22.3.3. Complex decay profiles and the stretched exponential ftinction 492

XXX CONTENTS

22.3.4. Wide-field time-domain FLIM instrumentation 493 22.4. MULTIWELL PLATE IMAGING OF ENDOGENOUS

FLUOROPHORES 495 22.5. FLIM MICROSCOPY OF BIOLOGICAL TISSUE 495

22.5.1. Cartilage 497 22.5.2. Artery wall and atherosclerotic plaques 500 22.5.3. Neoplastic tissue 501

22.6. TOWARDS IN VIVO IMAGING 503 22.6.1. Real-Time FLIM 504 22.6.2. Endoscopic FLIM 506

22.7. EMERGING TECHNOLOGY FOR FLIM AND MDFI 508 22.7.1. Tunable continuum source for fluorescence excitation 509 22.7.2. Hyperspectral FLIM instrumentation 514

22.8. CONCLUSIONS 517 22.9. ACKNOWLEDGEMENTS 518 22.10. REFERENCES 518

23. MULTIPLEXED FLUORESCENCE DETECTION FOR DNA SEQUENCING 525

Li Zhu, and Steven A. Soper

23.1. BACKGROUND AND RELEVANCE 525 23.1.1. What Is DNA Sequencing? 526 23.1.2. Gel Electrophoresis for DNA Sequencing 530 23.1.3. Fluorescence Detection for DNA Sequencing 535

23.2. DYE-PRIMER/DYE-TERMINATOR CHEMISTRY IN DNA SEQUENCING 536

23.3. FLUORESCENT DYES FOR DNA LABELING AND SEQUENCING 537

23.3.1. Visible Fluorescent Dyes 538 23.3.2. Near-IR Fluorescent Dyes 541

23.4. FLUORESCENCE-BASED DNA SEQUENCING STRATEGIES 546

23.4.1. Color Discrimination Methods 547 23.4.2. Lifetime Discrimination Methods 552 23.4.3. Combination of Color-Discrimination and Time-Resolved Methods 559

23.5. INSTRUMENTAL FORMATS FOR FLUORESCENCE-BASED DNA SEQUENCING 563 23.5.1. Fluorescence-scanning Detectors 565 23.5.2. Fluorescence-imaging Detectors 567

CONTENTS xxxi

23.5.3. Time-resolved Fluorescence Scanning Detectors 567 23.5.4. Time-resolved Fluorescence Imaging Detectors 569

23.6. REFERENCES 569

INDEX 589

THE HISTORY OF FRET:

From conception through the labors of birth

Robert M. Clegg^

1.1. INTRODUCTION

This chapter is an excursion into the historical development of energy transfer. This chapter is not concerned with a detailed review of applications, or a review of modem theoretical developments; this is available elsewhere (Van Der Meer et al, 1994; Wu and Brand, 1994; Clegg, 1996). The topic is the emergence of Forster resonance energy transfer FRET. I also examine the ideas, experiments and theories that formed the scientific backdrop that preceded and led up to FRET.

FRET is a physical process whereby the excited state energy of one chromophore molecule, the "donor", can be transferred to a neighboring chromophore, the acceptor, in the ground state. This can take place whenever the two molecules are close enough, usually separated by less that 7 nm provided certain other conditions are met.

FRET is one of the major experimental methods for discovering whether two molecules are in close proximity, or for determining the distance between two specific locations on macromolecules and in molecular complexes. Energy transfer is used to follow conformational changes of macromolecules, either statically or in real time. It has recently become a major experimental technique in the field of single molecules. Since the "efficiency" of energy transfer (that is, the fraction of energy absorbed by the donor that is transferred to the acceptor) is usually measured with fluorescence tools, and fluorescence is sensitive, specific and widely available, FRET has become very popular. The chromophores (donors and acceptors) that are used for accomplishing this measurement are usually attached (often covalently) to other macromolecules, such as proteins, nucleic acids, and lipids. The energy transfer can be detected relatively easily and it is often used qualitatively to signify intimate interaction

^ Robert M. Clegg, Physics Department, University of Illinois Champaign-Urbana, Illinois

2 R. M. CLEGG

between two "labeled" biomolecules. Sometimes one or both of the participating chromophores occur naturally in biological macromolecules, such as tryptophan or chlorophyll. However, the number and variety of synthetic fluorescence probes available for labeling has expanded tremendously in the last several years. Several readable reviews of FRET for a general audience are readily available (Clegg, 1992; Van Der Meer et al, 1994; Clegg, 1996; Clegg, 2004a).

The FRET measurement is now applied routinely with a wide variety of samples: micro structures (such as DNA and protein chips and micro/nano assay arrays), living biological cells, and even whole organisms. It is a very powerful technique, fairly simple, and can be carried out in most laboratories with their existing spectrometers and microscopes. Although the technique has been readily available and applied since the early 1950s, the use of FRET has literally exploded in the last few years, in academic research as well as industrial applications, especially in biotechnology and bioengineering. This flurry of activity has many reasons. First, FRET measures interactions and dynamics on a spatial scale that is unique. Also, our ability to produce well defined and pure macromolecules in the laboratory has increased dramatically in the last few years, and it is relatively easy to label them specifically with fluorophores. In the last several years we have developed the ability to produce hybrids of specific proteins with fluorescent proteins (for instance, GFP, YFP, CFP and RFP, respectively green-, yellow-, cyan-, and red-fluorescence proteins) that can be produced in vivo under genetic control in the living cell (and in tissue); certain pairs of these proteins can undergo FRET. These fluorescence proteins have revolutionized the field of biological fluorescence, especially the measurement of FRET, in the fluorescence microscope. A great number of excellent synthetic fluorophores are available commercially, with the required chemical groups attached for specific labeling to biomolecules. In addition there have been many instrumentation improvements and innovations that make the FRET measurement much more sensitive and convenient. These chemical, biological, and instrumentation advances have expanded tremendously the range of applications, and the ease of carrying out the experiments.

In spite of the wide spread use of such a well known and useful technique, and the availability of several excellent treatise and reviews of the underlying theory, not to mention the hundreds of experimental applications published every year, little is published about the historical development of the major concepts. The historical events are not only interesting in themselves, but understanding and appreciating the major theoretical insights realized by the pioneers of energy transfer, and the scientific context in which they worked, provides insight into the mechanism, and leads to a better appreciation of the original contributions. A short history of the contributions of the Perrins and Foerster to FRET has been published recently (Clegg, 2004b). This chapter is a more extensive examination of the state of affairs and the general state of knowledge that was prevalent in physics at the time, leading up to the first observations and theoretical explanations of energy transfer.

THE HISTORY OF FRET 3

1.2. PRELUDE TO THE HISTORICAL BACKGROUND

Although the practical applications of FRET started after 1950, the awareness that energy could be transferred between two atoms or molecules over distances larger than their physical collisional radii took place much earlier. The first experimental observation of energy transfer happened after 1900 and this chapter will only deal with FRET literature between 1900 and 1970. However, I will follow the thread back into the 19* century, when the ideas of an electromagnetic (EM) field and spectroscopy were being formed, and the dilemmas leading to quantum mechanics (QM) at the turn of the century were starting to appear. These theoretical concepts were essential for the observation and correct interpretation of non-radiative energy transfer. The notion of EM fields entered compellingly into the mainstream of physics only a few decades before the first observations of energy transfer. The first theoretical attempts explaining FRET were applications of this classical EM theory. And the first quantum mechanical theories of FRET were developed concurrently with the new theories of Heisenberg, Schrodinger and Dirac (Heisenberg, 1925; Dirac, 1926; Schrodinger, 1926b). So we will take a look at the historical scientific context in which the first experiments and theoretical accounts of energy transfer took place.

The aim of the first part of this chapter is to indicate the scientific atmosphere in which the idea of energy transfer at a distance was bom. We start by taking a short journey through the development of the concept of electromagnetic fields, move quickly through the quantum theory of atoms and spectroscopy, and then into the first experimental and theoretical discoveries of energy transfer. This initial time frame is from 1820 to 1920. I have decidedly selected the topics, emphasizing only those concepts important for FRET. After reviewing the emergence of these physical concepts, this chapter will only deal with FRET literature between 1920 and 1960.

1.2.1. The End of the Dark Ages: the Pre-Field Era

The fundamental paradigms of physics were undergoing radical changes in the 19* century, and these ideas were critical for understanding that atomic (molecular) interactions could extend over distances long compared to their atomic (molecular) radii. A lively account of how the concept of a field in a void entered into physics is given in an easily readable account by Einstein and Infeld (Einstein and Infeld, 1966). The concept of electric and magnetic fields is of course now common place, but we will see it was not until Maxwell that this concept was set on firm footing.

The notion that magnetism and electricity were somehow related had been suspected for some time before 1800, because of the formal similarities between static electricity and magnetism. Hans Christian Oerstead^ in 1820 was the first

^ Oerstead, a professor of Natural Philosophy in Copenhagen, received his PhD in 1799 in the medical faculty of Copenhagen; his topic dealt with Kant's philosophy.

4 R. M. CLEGG

to demonstrate the interrelationship of magnetism and electricity^ In a short four page article written in Latin (Greiner, 1986) he reported that a magnet's needle held next to a current-carrying wire was deflected, and oriented itself perpendicular to the line of current^ This interaction happened at a distance, decreased in effectiveness with increasing distance, and surprisingly the force on the needle was perpendicular to the line between the wire and the magnet^ This discovery, easy to reproduce^, was the first direct demonstration of the connection between electricity (a current) and magnetism, and it was first done by accident at the end of a lecture demonstration (Whitaker, 1989a). He wanted to show that if the magnet was perpendicular to the current flow there was no effect (which was true). At the end of the lecture he inadvertently oriented the magnet parallel to the wire, and there was a pronounced deflection. It was a phenomenally significant and completely unanticipated discovery, especially since magnetism had been known from antiquity, and was conceived by many as somewhat magical with supernatural powers^ and the compass and its use for navigation had been known for a long time. Oerstead's short report instigated immense interest throughout Europe, not only in the physics community, but was also enthusiastically received by workers in all scientific (and medical) disciplines. At that time, science was not as topically separated and divided as now. Even the general public heard of, and enthusiastically discussed, his experiment. Interestingly, Oerstead wrote later that it was his interest in romanticism and the movement of romantic natural philosophy that inspired him to carry out these experiments. He was also a passionate and tireless lecturer, and this may explain some of his influence (although, apparently, some scientists of that day did not appreciate his romantic outlook, and thought of him as a lucky, amateurish, dreamy opportunist - he earned handsomely from this discovery). Whatever, one might say that this was the inauguration of a great paradigm change in physics.

Two months later it was announced in Paris by Dominique Fran9ois Jean Arago^ the famous French astronomer, who had just returned from Denmark. Andre-Marie Ampere (and others, e.g. Jean-Baptiste Biot and Felix Savart)

^ Of course, this discovery could not take place before one had the ability of making currents (Voltaic piles) and had wires. Both these requirements had only been available since approximately 1800.

^ Oerstead got the idea to carry out his experiment with a galvanic circuit because it was known that deflections in a compass needle take place during lightning bursts in thunderstorms.

^ Oerstead did not determine quantitative aspects of this discovery; this was done by others soon after his discovery.

^ Interestingly, Oerstead was apparently all "thumbs" in the lab, and all his experiments had to be carried out by his (enthusiastic) students and assistants.

^ Magnets had been purported since antiquity to have healing powers. In 1780-1800 Franz Anton Mesmer, a doctor considered a charlatan by many but a medical savior by others, became a sensation "mesmerizing" his patients by passing magnets - in the appropriate mystifying setting -over the location in their body of their suspected ailments (or heads, if they were mentally distressed). To be fair, he realized later that his method had to do with suggestion, and did not require magnets. His work was the forerunner of the later work of Puysesur, Braid, Charot and Freud on hypnosis.

^ Arago discovered in 1811 the rotation of the plane of polarization when polarized light passes through an optically active crystal.


immediately repeated Oerstead's experiments. Ampere was adept in the manipulation (and participated in the theoretical development) of partial differential equations (he was also well educated and took part in the world of literature and philosophy), and in 1822 he soon came up with a theory of electromagnetic interactions involving currents^. He found that current-carrying wires would attract or repel each other depending on whether the currents were in the same direction or opposite. As a result of his interest in Oerstead's experiments, he suggested the possibility of a telegraph together with Jacques Babinet. This is the first mention that communication between two places could take place via electromagnetic interactions^^. I mention this, because electromagnetic communication between two locations is the basic physical event in FRET.

1.2.2. Middle Ages: Experiments That Eventually Changed Our World View

In the realm of magnetism and electricity, the pictorial, intuitive representation of lines or tubes of force was introduced by Faraday in 1821 (Whitaker, 1989a) ^ Faraday was perhaps the most thorough, dedicated, likable and honest of all experimenters. His interest in the inter-conversion and transformation of forces between different forms was a lifelong goal (this was before the concept of the conservation of energy and the equivalence of heat and energy by Julius Robert Meyer in 1842). Faraday handled his lines of force as though they were real physical entities, and not just as abstract helpful mathematical concepts^^. He pictured these lines of force as the mechanism by which electrical and magnetic substances interact with themselves and with each other; these tubes of force were so to speak for him the carriers of forces through space. He discovered the "Faraday induction" (i.e. changing magnetic fields produce circulating electrical fields), which is the basis of modem electric motors. The idea of forces at a distance had of course been a topic for a long time (e.g. Newton and gravitation). However, Faraday's ideas of tubes of force

^ Ampere's first law gives the force on a current carrying wire placed in a magnetic field; his second

law is that the magnetic field B circulates around an enclosed electrical current y ( ^ ) ; that is, in

vector notation, v x 5 =

^ The electromagnetic telegraph was later implemented by Karl Friedrich Gauss and Wilhelm Edward Weber in 1833, who developed the practical telegraph; they sent the first telegraph message in Gottingen from Gauss's observatory outside the city limits to Weber's laboratory in the city. Interestingly, there was much debate among the population of Gottingen whether the wires strung over the house tops were possible health hazards because of the magnetic effects emanating from the current carrying wires. Gauss developed a system of units for measuring magnetism that was based on length, mass and time, and this is the basis for the unit system in EM called the Gaussian Units. The unit Gauss, measures the strength of a magnetic field.

^ Actually, this concept was reported by Niccolo Cabeno, as early as 1628, and referred to as early as 1629 by Aristotelian-scholastic philosophers.

^ Faraday had no knowledge of formal mathematics. It has been mentioned to me by a mathematical physicist that he was perhaps the supreme mathematical physicist, with an incredible spatial imagination that did not need the crutches of equations.

6 R. M. CLEGG

did not sit well with many of the mathematical physicists of the day, who were used to expressing all in terms of differential equations^^. Nevertheless, he was an exceptionally capable experimenter; his experiments and graphical explanations were remarkably innovative and provided astounding, intuitive insight. His ideas set the stage for the next step important for the background for FRET, which was then carried out by Maxwell.

1.2.3. Renaissance: Enter the Theory of Electrodynamics and Fields

The basic FRET phenomenon involves the electrodynamic interaction between two molecules over distances that are large compared to their diameters; and this description requires the idea of an EM field (for FRET this is a dipole interaction, which arises from a multi-pole approximation to the Coulomb interaction). Faraday did not deal with fields, but with tubes of force. It was James Clerk Maxwell who introduced the first field theory. His equations describe the EM field; the objects (electrical or magnetic) enter only through boundary conditions. The ideas and experiments of Faraday played a major role in Maxwell's theoretical development. He created a complete mathematical representation of Faraday's descriptions of electricity and magnetism (Maxwell, 1873; Simpson, 1997). He admired the work of Faraday, and read all what Faraday had written before undertaking the task of formulating his ideas in mathematics^" . He was of the opinion that Faraday had articulated his discoveries "in terms as unambiguous as those of pure mathematics". Maxwell's accomplishment is enshrined in his famous classical equations of electrodynamics, which are familiar to all physics students (and often the cause of much sweat and toil). In addition to the concepts he borrowed from Faraday, he introduced the notion of displacement current - the circulatory magnetic field caused by a time-varying electric field (also in empty space). These equations describe all classical electrodynamic phenomena, and they are the starting point for describing energy transfer. His equations fumished the theoretical setting to predict electromagnetic radiation (e.g. the classical theory of fluorescence emission).

It is a fascinating story how Maxwell wrestled with physical and mathematical analogies, experimental results and mathematical formulations in order to arrive at his equations, as well as how his equations predicted light as an electromagnetic field. He created the term electrodynamics, as this quote from him exemplifies: "The theory I propose may therefore be called a theory of the Electromagnetic Field, because it has to do with the space in the neighborhood of the electric or magnetic bodies, and it may be called a theory

' Faraday eventually repudiated a reality that consisted of separate entities of matter (atoms) and void: "The difference between a supposed little hard particle and the powers around it, I cannot imagine".

^^ The titles of Maxwell's first two of his three main papers on EM were: "On Faraday's Lines of Force" (1855) and "On Physical Lines of Force" (1861). The first shows his respect and enthusiasm for Faraday's ideas, and the second signifies his new paradigm of the physical field concept. His third paper on this subject "A Dynamical Theory of the Electrodynamic Field" was presented in (1864).


of the Dynamical Theory, because it assumes that in that space there is matter in motion, by which the observed electrodynamic phenomena are produced" (this quote is from his third paper in this series - "A Dynamical Theory of the Electrodynamic Field"). This is actually a partial description of what happens in FRET. Maxwell understood Faraday's lines of force as "a line passing through any point of space so it represents the direction of the force exerted ...". This depicts nicely the vector representation of the EM field, which is now given in every EM textbook, and is the way the orientational dependence of the interaction between a FRET pair is portrayed.

The impact on the physics community and the conceptual revolutions that were initiated by Faraday and formulated by Maxwell are perhaps difficult to appreciated^. The theory of Maxwell, based on the original ideas of Faraday, turned much of physics on its head. Whereas Newton's laws conserve energy and momentum in the motions and collisions of bodies. Maxwell's field theory is concerned with the energy and momentum of the field, and does not take account of the bodies, except as boundaries. Because this concept of a field is so critical for understanding FRET, and this is a historical account, I emphasize this with two quotes. The first is from Ludwig Boltzmann^^, who, to express his admiration for Maxwell's equations, quoted Goethe: "War es ein Gott, der diese Zeichen schrieb?" - Was it a god, who wrote these expressions?". The second is from Maxwell himself '': " In speaking of the Energy of the field, however, I wish to be understood literally. All energy is the same as mechanical [...]. The energy in electromagnetric phenomena is mechanical energy. The only question is. Where does it reside? On the old theories it resides in the electrified bodies, conducting circuits, and magnets, in the form of an unknown quality called potential energy, or the power of producing certain effects at a distance. On our theory it resides in the electromagnetic field.". This was an enormous paradigm change. The difficulty to imagine energy in a void led to the introduction of the "ether". Ether was supposed to be the inert medium through which all electric and magnetic phenomena were transmitted. This was not required by Maxwell's equations. But scientists trained in the Cartesian view of physical phenomena found the ether necessary (including Maxwell). It was difficult to imagine how the transverse undulations of light (required and predicted by Maxwell's equations and verified by Hertz) could take place without a medium. As we know, the ether was shown not to exist. (Whitaker, 1989a; Whitaker, 1989b).

1.2.4. The Beginning of the Modern Age: The Field Surrounding an Oscillating Charge

The next critical step for FRET was carried out by Heinrich Hertz, with his famous Hertzian oscillating dipole. The electrodynamic field emanating from a vibrating electric dipole (the Hertzian oscillating dipole) is derived from

' Although eventually this impact was prevailing, Maxv^^ell's EM theory was not immediately accepted, and was even highly criticized my many scientists.

^ " Vorlesungen iiber Maxwells Theorie", L. Boltzmann ^ "A Dynamical Theory of the Electrodynamic Field" JC Maxwell (1864)

8 R. M. CLEGG

Maxwell's equations. This is the classical theoretical basis for understanding the production of light (e.g. emission of fluorescence from atoms and molecules, as well as the absorption), and is also the starting point of the first classical descriptions of FRET. Maxwell's equations predicted the identity of electromagnetism and light (expressed in his memoirs of 1868), and even foretold the quantitative properties of light (interference, refrangibility, polarization, as well as the speed of light). This was brilliantly confirmed by the experiments of Hertz. These experiments, and his theoretical description based on Maxwell's theory, forced the skeptics in the scientific community to accept the concepts inherent in Maxwell's field equations. Hertz carried out the critical experiments in 1888, and published the theory (derived from Maxwell's equations) to explain the EM fields surrounding his electric oscillator in 1889. This derivation is given in any intermediate or advanced electrodynamics textbook. Of course. Hertz was not specifically referring to an atomic oscillator, but to a macroscopic electric oscillator.

His first experiments were carried out by producing high frequency repetitive sparks in an air gap of a primary oscillating circuit (which acted as the source of the EM radiation). The electrodynamic disturbance was detected at a distance by a secondary circuit, resonant with the first, with a similar air gap. Sparks were observed in the secondary receiving circuit when it was resonant with the primary circuit. At first Hertz was primarily interested in proving the existence of propagating electromagnetic radiation (light, at frequencies of what is today radio frequencies), which was predicted by Maxwell's equations. Therefore, in the first experiments the distance between the primary and secondary circuit was long compared to the wavelength of the propagating Maxwell electromagnetic wave at that frequency of oscillation. The result was fully consistent with Maxwell's field equations, and this is of course the basis of all radio communications.

Hertz's theoretical description describes the electromagnetic disturbance in the near field (much less than a wavelength of the emitted radiation), in an intermediate zone, as well as in the far zone (at distances greater than a wavelength) where the electromagnetic energy escapes and is carried away as radiation with transverse oscillations. He calculated a very graphic field-line representation of the EM field of an oscillating dipole, demonstrating how the field lines are pinched off at approximately a distance of a wavelength, at which point transverse waves in the far field are formed (E and B fields of propagating radiation are perpendicular to the direction of photon travel). Only in the far field can we think of a photon (or in the language of pre-photon concepts, a traveling light wave). In the near field (distances small compared to the wavelength, which is where FRET takes place) both transverse and longitudinal components of the EM fields are present. However, in spite of the high energy density in the near-field, no propagating EM waves are present here (no emission of energy). This corresponds to the terminology in FRET that the energy in the near field is transferred non-radiatively. Experiments carried out in the near field verified Hertz's theoretical description.

The graphical and mathematical description of the oscillating electric field emanating from a Hertzian dipole, in particular in the near field, played a


critical role in the understanding, and the eventual theoretical description, of FRET. The oscillating E-field in the near field of a Hertzian dipole has the same effect on a receiving oscillator as when the receiver is in ihQ far field. The only difference between the near- and far-field effects on a receiver (acceptor in FRET) is the direction of the field vector (and the intensity), which is a sum of tangential and longitudinal components in the near field, and only tangential in the far field. The oscillating electric field surrounding the Hertz oscillator is indispensable for all theoretical descriptions of FRET.

1.3. FIELDS, SPECTROSCOPY AND QUANTUM MECHANICS

1.3.1. Fields

As we have seen, in a relatively short period of time the paradigm in physics that all interactions take place by mechanical contact and collisions, changed considerably (Einstein and Infeld, 1966). By 1900 the concept of a field had been generally accepted (this is only 22 years before the first report of energy transfer at a large distance between atoms in a vapor (Cario, 1922)). Nevertheless, because of the very successful dynamical gas theory (the groundwork of which was also set by Maxwell), most original interpretations of energy transfer and fluorescence quenching between gaseous atoms naturally assumed coUisional contact (mechanical interactions).

1.3.2. Quantum Mechanics and Spectroscopy

At the turn of the century there was another paradigm change in physics about to take place, which is the second requirement for understanding FRET. This took place because the theory of radiation, in spite of the success of Maxwell's theory, had reached a very unsatisfactory state. This had to do with the failure to explain the dispersion (frequency dependence of the energy emission) of blackbody radiation. As is well known, in 1900 Planck solved this problem by introducing the quantum concept - energy changes in matter could only take place in well defined quantum jumps. His successful theory for explaining black body radiation was presented December 14, 1900 at the German Physical Society (Planck, 1900) l His reasoning centered on an ensemble oi Hertzian oscillators (the oscillators were the atomic constituents of the walls), which he proposed could only exchange energy with the radiation field in jumps of energy quanta. His famous paper was published in 1901 (Planck, 1901). This was followed by the work of Einstein (Einstein, 1905;

^ In October of 1900 he had presented to the same society a phenomenological theory that was in agreement with black body experiments, which was a modification of previous work of his (in 1899) based on thermodynamic reasoning; but he considered this approach unsatisfactory because it contained undefined empirical constants.

10 R. M. CLEGG

Einstein, 1906), where he proposed that light itself behaved as though it were particle-like quanta of energy^^

1.3.2.1. Conway and Ritz

Up until 1907, the picture that scientists had of absorbing and emitting atoms was that each atom consisted of an electrical system with many natural periods of oscillation, all present simultaneously. In 1907 Arthur William Conway, from Dublin (Conway, 1907), proposed that the spectrum of an atom does not result from free vibrations as a whole, but that each atom produces spectral lines one at a time (that is, the actual spectrum observed at any time depends on the presence of many atoms). This is prior to the Bohr-theory of the atom, or even of the Rutherford model of an atom (1911). The idea of Conway was that in order to produce a spectral line, one electron in an atom must be in some sort of perturbed state (he had no way to describe this in detail), and that this electron is then stimulated to produce vibrations of a frequency corresponding to the observed spectral line. This disturbed state did not last indefinitely but would decay with time, emitting a fairly long train of vibrations (as was required by Maxwell's equations). This was remarkable insight considering that he did not know the later interpretations of atomic spectra (e.g. the Balmer and Rydberg series) in 1908 by Ritz (Ritz combination principle) (Ritz, 1908) who showed how the spectral lines could be interpreted by differences, taken in pairs, of certain distinct numbers.

1.3.2.2. Bohr

This discussion would not be complete without mention of the critical insight of Niels Bohr (Bohr, 1913; Hettema, 1995), who integrated and selected many of the ideas that were being considered at the time (including those of Conway (Gillispie, I960)). He selected what he considered to be superior ideas from the inferior ones, and produced the paradigm of the atom (the Bohr atom) that was enormously influential and placed spectroscopy on firm ground. This is now so common-place and is even sometimes introduced in grade school; however, his synthesis of a model that could explain quantitatively the spectroscopy of many simple one-electron systems was a real eye-opener, and was critical for all that followed. Because it is so well known, I spend no time describing Bohr's ideas. Of course, extensive experimental studies in spectroscopy were carried out at that time (Pringsheim, 1928). The spectroscopic experiments stimulated the theoretical work, and provided the data for motivating and checking the theories. The extensions of these basic ideas to complex molecules could only take place after the introduction of quantum mechanics by Heisenberg in 1925 (Heisenberg, 1925) and Schrodinger

' Although Planck had considered that energy of the field could be quantized, he resisted this because this would go against all that was known of the "continuous" Maxwellian light field; for this reason he only considered the quantized emission of energy from the oscillators, and not the absorption.


in 1926 (Schrodinger, 1926b), and we will see in the next section that the new quantum mechanics was immediately applied by Kallmann and London (Kallmann and London, 1928) to explain energy transfer in vapors, and by F. Perrin for energy transfer in solution (Perrin, 1932).

1.4. THE FIRST EXPERIMENTAL OBSERVATION OF ENERGY TRANSFER AT A DISTANCE - SENSITIZED LUMINESCENCE IN VAPORS

1.4.1. Sensitized Fluorescence

With the concept of EM fields, the experiments and theory of Hertz, the development of the older quantum theory, the spectroscopy data of atoms, and Bohr's theory of the atom, we have reached the point of entry for FRET. The first recorded measurements of energy transfer (observing the emission of the accepting atom) over distances larger than collision radii were made in 1922 by Carlo and Franck (Carlo, 1922; Carlo and Franck, 1922; Franck, 1922). Carlo observed emission from thallium in a mixture of mercury vapor and thallium vapor, when the vapor mixture was excited with wavelength of 253.6 nm, which can only excite the mercury atoms. This fluorescence emission from thallium was named "sensitized fluorescence". It was obviously due to the transfer of energy from the excited mercury atoms to the thallium atoms. Further experiments by many people showed sensitized fluorescence with the vapors of the alkali metals: silver, cadmium, lead, zinc and indium in the presence of mercury vapor. The importance of resonance between the energy levels of the sensitizer and the sensitized atoms was explicitly shown by fiirther experiments, especially with the later experiments of Beutler and Josephi (Beutler and Josephi, 1927; Beutler and Josephi, 1929), who studied the sensitized fluorescence of sodium vapor in the presence of mercury vapor. The sensitized fluorescence increased in intensity the smaller the energy differences between the states of the two participating atoms. This was consistent with "Franck's principle" (Franck, 1922), which had been articulated, in reference to fluorescence quenching, previous to the experiments of Carlo. This principle can be stated as: "the electronic energy of an excited atom cannot be transferred directly into kinetic energy of the colliding particles. If the excitation energy has to be taken over almost completely as internal energy of the quenching molecules, these must have some sort of excited states, which are in energy resonance with the primarily excited states", page 116 of Pringsheim's book (Pringsheim, 1949); my italics. Here we see already the entrance of "resonance" into the FRET story, which will play a central role.

Sensitized fluorescence was discovered during the copious spectroscopic experiments that were carried out at different temperatures and densities with vapors of many different atoms and diatomic molecules in the first two decades of the twentieth century. A thorough discussion of literature up to 1949 can be found in Pringsheim's treatise (Pringsheim, 1949). Dynamic fluorescence

12 R. M. CLEGG

quenching (due to collisions), photochemical reactions, resonance fluorescence and molecular associations were intensively studied in the early decades of the 1900s (Wood, 1934). Many of these experiments were naturally interpreted in terms of collision theory. The number of collisions per time could be calculated simply from gas theory; and the fraction of collisions that were effective (leading to quenching, chemical reactions, or sensitized fluorescence) could then be determined.

1.4.2. Spectroscopic and Collisional Cross-Sections in Vapors

The cross-section of molecular encounters (or what is equivalent, the frequency of effective collisions) gives estimates for the "spectroscopic size" of the reacting atoms. From the fraction of successful collisions, calculated by comparing the rate of successful quenching encounters to the collision rate from gas theory, one can calculate a ''spectroscopic cross-section'\ If the radius of this cross-section is smaller than the radius of the atoms or molecules (or equivalently, if the rate of collisions is smaller than calculated from the gas theory) then the conclusion is that only a certain percentage of the collisions are effective in quenching. If the spectroscopic cross section is larger than expected from the encounter radius, then it is assumed that there are interactions between the two collision partners that extend beyond their encounter radius. A larger radius of interaction than predicted from the theory of gas dynamics was found for many energy transfer measurements in the vapor. These large "spectroscopic" cross-sections constituted the first hint that many inter-atomic interactions could take place over larger distances. This discovery that energy transfer could take place over distances large compared to the encounter radii showed that hard physical collisions were not required for atoms (or molecules) to interchange energy.

1.5. THE FIRST QUANTUM MECHANICAL THEORY OF ENERGY TRANSFER

A quantum mechanical theory to explain the transfer of energy between atoms at longer distances compared to collisional radii was proposed by Kallmann and London20 (Kallmann and London, 1928). This theory assumed "almost resonance" between the energy levels of the interacting atoms. Essentially this is a second order perturbation calculation to calculate the energy of interaction. I will use their notation in this paragraph. They found that provided the corresponding transitions between the energy states of the two atoms were (spectroscopically) dipole-allowed, the effective cross-section q of

- 2 / 3 the two interacting atoms increases as a , where a is the difference between

' This is the same F. London who proposed the quantum mechanical description of van der Waals interactions, which also involved dipole-dipole interactions, similar to FRET.


the excitation energies of the two interacting systems. As cr ^ 0 , the cross section approaches a limiting value much larger than the coUisional radii. This work was the germ of the later quantum mechanical FRET theories, and is very similar to the description of London-van der Waals forces. Before discussing this paper further, we look at classical theories that were published just previous to Kallmann and London.

1.5.1. A Few of the Pre-Quantum Theories that calculated the Spectroscopic Cross-Sections of Atomic Vapors

Some of the previous pre-quantum papers (Holtsmark, 1925; Mensing, 1925; Nordheim, 1926) (all referenced by Kallmann and London) dealing with "spectroscopic cross-sections" are remarkable in their insight.

Mensing (Mensing, 1925) considered how intermolecular dipole-dipole interactions broaden spectra (using the Bohr-Sommerfeld atomic theory with elliptical orbitals) of atomic vapors. She derived a broadening due to dipole-dipole interactions to be approximately a width of

Sv^3.6\^e^/hjya^/dj\l + 3£^/2], where a and £ are the long axis and

eccentricity of the atomic orbitals, eis the electron charge, and d is the distance between the two molecules. She considered dipole-dipole interactions of circulating electrons (from a semi-classical point of view, this is similar to what is done by London when the two interacting molecules are not in the same energy state).

An article by Nordheim (Nordheim, 1926) presents a collision theory between atoms, whereby the atoms interact via dipole-dipole terms. Although the new quantum theory of Heisenberg (Heisenberg, 1925) and Bom and Jordon (Bom and Jordan, 1925) had been published already, Nordheim calculates the interactions classically. He justifies this choice because the results should be approximately the same, and he also wanted to by-pass the difficulties in analyzing the collisions quantum mechanically. Later, Forster, Ketskemety and Kuhn (Forster, 1951; Ketskemety, 1962; Kuhn, 1970) showed that the classical and quantum calculations arrive at the same result. Nordheim derives also higher multipole interaction energies, and uses the same classical conjugate dynamic equations of motion as Mensing. By considering collisions between the electrical multipoles that are oscillating at spectroscopic frequencies, he derives expressions for the energy transfer rates proportional to the product of the appropriate powers of the spectroscopic transition moments of the different multi-pole interactions. However, this derivation is very complex, because the rate depends on very complex averaging of the collision paths, and the velocities. The details of the theory are only applicable to vapor samples.

Holtsmark's paper (Holtsmark, 1925) is a purely classical approach aimed at understanding the average deviation of the energy levels from that of the free atom values of Na-vapors, in order to understand the broadening of the spectroscopic lines. The spectral widths were known to be broader than calculated from simple hard-core collision theory according to gas theory. A

14 R. M. CLEGG

classical perturbation approach (assuming that the atoms are classical electron oscillators) was taken whereby again the interaction of the atoms was assumed

to be dipole-dipole. Holtsmark calculated an interaction proportional to ijR^

for the interaction between any two atoms. Since he was interested in a random collection of atoms at a certain concentration, he integrated all interactions from the shortest distance of atomic approach {d) to infinity (including the orientation factor). This results in broadening of the lines proportional

Xoye^lm\ Idt (where e is the electron charge and m^ is the mass of the

electron); although he has assumed perfect electric oscillators, each one

contributing a factor of e jm^, the interaction would also be proportional to the

multiplication of the oscillator strengths. He did not assume exact resonance -

the total interactions were summed via a type of overlap integral. His factor d~ is the same dependence on the distance of closest approach as by Arnold and Oppenheimer (Arnold and Oppenheimer, 1950) (see below) for the rate of energy transfer in a condensed system with random acceptors.

The point of discussing these early theoretical accounts of dipole-dipole interactions leading to energy transfer is to show the type of analyses invoked at this time to account for the very large effective molecular spectroscopic cross sections. These cross sections were much larger than expected from simple atomic coUisional gas theory. This could only be explained by molecular interactions at large distances. These theories were published just a few years before Kallmann and London's quantum derivation, and at the same time as J. Perrin's work on energy transfer in solution. Already at this time, it was apparent to everyone that dipole-dipole (or higher multi-pole) interactions could extend the radius of inter-atomic interactions considerably. Some of the theories

showed explicitly the 1/7? dependence. In addition, it was shown that the

spectroscopic oscillator strengths (that is the spectra) had to overlap and the orientation of the dipoles were taken into account. It is clear that the pieces are starting to fall into place.

1.5.2. Some Details of the Kallmann and London Paper

The theory of Kallmann and London (Kallmann and London, 1928) dealt with energy transfer in vapors of atoms; their theoretical ideas were the basis of the later quantum mechanical theory proposed by F. Perrin (Perrin, 1932; Perrin, 1933) for energy transfer in condensed systems, which was subsequently improved and extended by Forster (Forster, 1948).

The article by Kallmann and London (K&L) has many interesting aspects that are important for understanding the later theoretical treatments of FRET. Because F. Perrin essentially used a very similar theoretical approach, and his theory is outlined in the section "F. Perrin's model" below, I refer the reader to that section for a discussion of the basic ideas of the derivation. K&L assumed that the two interacting atoms have two states, but the energy levels do not have


to be exactly the same. They write the differential equations derived from the Schrodinger wave mechanics pertaining to the coupled system for the case of two states (Schrodinger, 1927). The basic equations were given by originally by Schrodinger (Schrodinger, 1927), in an article titled "Energieaustausch nach der Wellenmechanik", or "Energy exchange according to wave mechanics". K&L refer to this paper (this will be important when we discuss Rabi oscillations below), and it is remarkable that Schrodinger essentially indicated the path for a solution to energy exchange between two atoms (molecules) in one of his first quantum mechanic papers. The solution for this problem, if the energy states are narrow, is oscillatory, and K&L give the probability that the system of two

atoms will have transferred energy to the other, c |, as (using their notation)

\c \=-p'

\^p' -sm

h

P -IW^^jG* . ^2 is th^ interchange integral of the perturbation

W between the two states of the system, a' = a + W^^ ~ W^^ . a is the

difference of the energy levels of the two participating molecules. They assumed for simplicity that the zero energy levels of both molecules were equal.

^ j and 22 ^ ^ ^^^ configuration integrals of the perturbation W for the two

different states (the first order energy change from the perturbation). W is the

dipole perturbation operator, and W = {ju^ju^)/R^ where//^ andju^ are the

transition dipole moments of the two molecules; this is the usual Coulomb interaction of dipoles (the QM version of the Hertzian dipole interaction energy). They are interested in collisions between atom gas molecules, and if the time for an atom to pass by another is long, compared to the oscillation of

sin^[--] term, they can take an average over this time, giving them the

probability G7 that, on the average, the energy will be transferred from the first combined molecular state to the other state. This is (skipping several steps - and carrying through several substitutions in order to show the relationship between their derivation and Forster's later theory)

1 J3'

m = •

j/2_

111 1 + ,-",

3 \ -

1/2

1/2

1 + R"

4(//,//.)V' o- ) 1 +

Xl2_

R'l 0 /

16 R. M. CLEGG

I have defined the variable R\={A{^ju^ju^y j ( J ^ \ in anticipation of the

normal R^ of energy transfer. Except for the factor of xjl, this is the standard

expression for the efficiency of FRET. The 1/2 arises because the solution is

oscillatory, and K&L have calculated the time average. In addition K&L have not integrated over a frequency spectrum of the two spectroscopic transitions, which in reality cannot be sharp lines (this would give us the overlap integral -see our discussion of Arnold and Oppenheimer's derivation (Arnold and Oppenheimer, 1950)). This equation by K&L is quite remarkable - it is the first indication, except for the classical calculations given in the last section, that for two atoms separated by R the probability of energy transfer obeys a

l / n + i?/i?'^ jrule. The validity of the equation depends on the ratio of the

energy differences and the size of the perturbation, and I cannot delve into the reasoning concerning this ratio when calculating molecular interactions in this paper (but see the short discussion by Knox (Knox, 1996)). The important point is that already in 1928 K&L derived essentially the correct dependence on the separation of the two molecules. By integrating this equation over R they arrived at their expression for the "spectroscopic" cross section q .

2 \ + Pl \ (J J

This is the effective cross-section of atomic collisions, which is significantly larger than the hard core cross sections due to dipole-dipole interactions between one excited molecule and one ground state molecule). The last equality holds for not too sharp resonance. This is the relationship

q oc a'^'^given two sections previous. And it is based on an efficiency of

energy transfer that varies as R'^. K&L calculate approximations to this integral, and find that the dipole-

dipole interaction at a distance results in an anomalously large cross section, and their equation compares well to the experimental results. The above derivation is valid if the two molecules are not exactly in resonance (that is, if the spectroscopic transitions of the two molecules are not identical). K&L

derive the case for exact resonance, and decide that this would give values R \

values far too large. This is partly because one would have to account for the unavoidable frequency spread of the molecular transitions. I discuss this below. This is essentially the same as the F. Perrin derivation below, and so I refer the reader to this section on F. Perrin for a discussion of this. We will see that even though F. Perrin used the same basic quantum calculation as K&L, which is that discussed by Schrodinger in 1927 (Schrodinger, 1927), he used exact resonance and calculated the rate of energy transfer in another way. F. Perrin calculated

the rate from the period of the oscillation in the sin^ [•••] term of the above


solution of K&L; and this gives the wrong distance dependence of energy transfer. In summary, KcfeL had the right idea, as well as the right dependence onR.

1.6. LONDON FORCES (VAN DER WAALS) AND DEB YE AND KEESOM INTERACTIONS

Before delving further into theories of FRET, we turn our attention to a closely related topic - the quantum description of van der Waals (1873) interactions (or London dispersion forces). The idea of dipole-dipole interactions at a distance between atoms and molecules was being applied by London to explain intermolecular van der Waals interactions concurrently with his work on energy transfer (previous section). The calculation of classical descriptions of dipole-dipole and dipole-induced-dipole interactions had already been introduced to explain intermolecular interactions (Keesom, 1912; Debye, 1920;Debye, 1921).

There is a close connection and concurrent historical development of the theories describing London's intermolecular interactions and FRET. The major difference is that one is interested in the energy of interaction for the van der Waals forces, and in the rate of energy exchange for FRET. For normal van der Waals interactions, both interacting atoms are in the ground state^^ In FRET, one of the interacting atoms (molecules) is in an electronically excited state. The theory of London's forces is also important with regard to the first quantum theories of FRET by F. Perrin.

1.6.1. London Interactions: Induced-Dipole-Induced-Dipole

Fritz London published his quantum mechanical description of these forces in 1930 (London, 1930; London, 1937). This was two years after he published the paper analyzing the transfer of energy between mercury and thallium with Kallmann (Kallmann and London, 1928). London's interaction energy is carried out by quantum mechanical second order perturbation theory. Normal London dispersion interactions involve fluctuating dipole-dipole interactions between atoms in their ground states.

An excellent review of classical and quantum mechanical theories of van der Waals interactions can be found in a 1939 review of Margenau (Margenau, 1939), and good accounts are also given in books by Davydov (Davydov, 1965) and Walter Kauzmann (Kauzmann, 1957). We limit our discussion to interactions between two atoms in their ground states; although, London forces

^ London forces usually refer to interactions between two atoms in their ground states; however, in the review by Margenau, he discusses the interaction between one atom in the excited state, and one atom in the ground state. This is essentially the same physical circumstance as FRET, except that London forces result from averaging rapidly fluctuating forces over time (and space), and one is looking at the energy of interaction, not a rate of transfer, which is what one is observing in FRET.

18 R. M. CLEGG

are summed to explain interactions between large, closely spaced macroscopic objects.

The expression for the London dispersion forces between two atoms involves the product of the corresponding visible and UV oscillator strengths of the optical transition for each interacting atom, just as FRET (Forster, 1951) (see section on Forster below). Usually the London dispersion energies are expressed in terms of the polarizabilities of the two atoms; however, the polarizabilities are related to the spectroscopic oscillator strengths (Margenau, 1939; Kauzmann, 1957). The separate components of the oscillator strengths are proportional to the squares of the (spectroscopic) transition moments for each optical transition of the atoms. The full expression for the London interaction is a sum over all significantly contributing optical transitions of the two interacting atoms (see the next paragraph, and see equation 7 of Margenau (Margenau, 1939)). If both atoms are in their ground states, the London forces are always attractive^l In addition, if the interacting atoms are not spherically symmetrical, and have anisotropic polarizabilities, their relative orientation will affect the interaction (just as the orientation dependence of FRET). The simple van der Waals interaction (not taking into account retarded potentials - which are involved in the Casmir effect (Casimir, 1948)) according to London's theory

decreases as R^, where R is the distance between the atoms; this is the same distance dependence as FRET between two chromophores.

The interaction term in the total Hamiltonian is that of two interacting

electrical dipoles, and is proportional to \/R^ . The London interaction energy is

calculated using second order perturbation theory (the first order perturbation term for atoms or molecules in their ground states is zero). According to second order perturbation theory, the total energy of interaction is proportional to a sum over all higher energy states of the combined two atom system, where each term of the sum is proportional to the square of the interaction matrix element.

Therefore, the total energy of interaction is proportional io\/R^, where R is

the distance between the two atoms. Each component of the sum is also proportional to the product of the squares of the transition dipole moments (these are identical to the spectroscopic transition moments) for each atom between the ground state and the excited state for that term of the sum. In addition, each term of the sum is divided by the difference between the combined energies of the two ground states and combined excited states corresponding to that term of the sum. Although the sum is extended over all higher energy states, the higher energy states are usually assumed to be small,

^ However, interestingly (especially within the historical context of FRET) if one interacting partner is in the excited state, the interaction can become repulsive. If one atom is in the excited state, the atomic interaction shows resonance; that is, the interaction becomes very strong when both the electronic transitions (absorption and emission oscillator strengths) have large values at very nearly the same frequency (energy). In addition, if one of the molecules is in the excited

state, the interaction energy can vary as \/R , because the energy contribution from the first

order perturbation calculation is no longer guaranteed to be small. This will become important when we deal with F. Perrin's FRET theory.


and are not included in calculations. This sum can be written in terms of the product of polarizabilities of both atoms. The denominator of each term is an energy difference, where the higher energy is subtracted from the lower energy. Therefore, each term contributes a negative energy component, meaning that the normal van der Waals interaction is universally attractive.

Although in many respects the theories of London interactions and FRET are similar, the two theories are not identical. One major difference is that the London interactions are calculated from second order time independent

perturbation theory; this is the origin of the X/R' dependence and the squares

of the transition matrix terms. The y R^ dependence and square of the

transition matrix terms in the Forster expression for energy transfer emerge from different reasoning (application of Fermi's Golden Rule to calculate a rate), as I will discuss later. Fermi's Golden Rule can only be applied when the interacting oscillators are dynamically incoherent; this point will be important when we discuss the Perrins' treatments of energy transfer, because they did not

find a XJR^ dependence, but a l/i?^ dependence, essentially assuming a

coherent interaction. However, as already mentioned, London found the correct

y R^ dependence.

Since the London interactions and energy transfer interactions originate from dipole-dipole interactions, it is not surprising that Kallmann and London was working concurrently on both interactions. He derived both quantum mechanical theories for the vapor state essentially simultaneously. It is remarkable that he achieved this straight away so soon after the quantum theories by Heisenberg and Schrodinger.

1.6.2. Keesom and Debye Interactions: Dipole-Dipole and Dipole-Induced-Dipole

Previous to London's theory, inter-molecular interactions between molecular dipoles had already been proposed as the basis of molecular interactions. The Keesom orientation effect considered the interaction between two permanent molecular dipoles (Keesom, 1912); if the dipoles are strong, this can orient the interacting dipoles. Debye described induction forces between a permanent molecular dipole, and an induced molecular dipole (Debye, 1920; Debye, 1921); this is known as the Debye interaction or the induction interaction. For both the Debye and Keesom interations the energy varies as the inverse 6 ^ power of the interatomic interactions, as do the London forces. The latter theory of London describes the van der Waals dispersion interactions as due to fluctuating oscillating induced molecular electric dipoles (London, 1930; London, 1937); that is, no permanent dipole need exist. All three interactions, dipole-dipole, dipole-induced dipole, and dipole induced - dipole induced, are usually included as components of van der Waals forces (Israelachvili, 1992). In contrast to the Keesom and Debye effects, the van der Waals interaction as described by London's dispersion forces is a pure quantum mechanical effect.

20 R. M. CLEGG

and is always present (although classical non-rigorous derivations are often given). The designation - dispersion - refers to the dispersion of light in the visible and UV spectral region (remember that the energy of van der Waals interactions can be expressed in terms of the oscillator strengths of the interacting molecules). As we will see, Forster's expression for the rate of FRET also involves the optical oscillator strengths (usually expressed in terms of the absorption and emission spectra).

1.7. FRET BETWEEN ORGANIC CHROMOPHORES IN CONDENSED SYSTEMS

We've now arrived at the time when the first attempts were made to explain observations indicating energy transfer over long distances in condensed systems (in solution). The first classical description of this by J. Perrin (Perrin, 1925; Perrin, 1927) predated London's quantum mechanical derivation of energy transfer in the vapor phase, but followed Cario and Franck's studies of energy transfer in vapors. As indicated above, dipole-dipole interactions were well understood at this time, and the description of the Hertz oscillator contains all the essentials needed to explain FRET on the basis of classical models of atomic electric oscillators. The classical description of FRET in condensed matter systems, with imposed restrictions from the old quantum theory, involves the comparison of the energy escaping to the far field from a Hertzian oscillator when it is alone, to the energy escaping when another molecule or atom (the acceptor) is in the near field zone. All the concepts that have been introduced above come into play: the oscillating electric field of a Hertzian dipole (near and far field), the quantum states of the oscillators (atoms), the requirement that the two communicating oscillators be in resonance and that the orientation of the two oscillators be favorable, and the idea of competition between emission of radiation and energy transfer. All these concepts were standing by, ready to be put into place when J. Perrin approached this problem for fluorophores in condensed solvents. However, because the new quantum mechanics had just begun, some rather subtle concepts in time-dependent quantum systems were not obvious.

1.7.1. Experimental Observations of Energy Transfer in Solution

The experiments that led J. Perrin to attempt a theoretical interpretation of energy transfer between molecules, involved fluorescence polarization in a solution of a single chemical species of fluorophore. It had been discovered by Weigert (Weigert, 1920) and by Gaviola and Pringsheim (Gaviola and Pringsheim, 1924) that the polarization of fluorescence emission from solutions of dye molecules began to decrease rapidly when the concentration was raised to a critical value (approximately 3 mM). This happened even when the fluorescence intensity (corrected for trivial absorption of fluorescence) was still linearly increasing with concentration. A fluorescence polarization


measurement determines the degree of rotational freedom of the fluorophores. If the fluorescent molecules are in a rigid environment or highly viscous solution so that they cannot rotate within the time of fluorescence decay (usually between 1-10 nanoseconds), the polarization contribution from individual molecules will have a high value (a maximum of 0.5). If the fluorescent molecules act independently, the polarization should be concentration independent. It could be shown for several well known dyes that the polarization was appreciably reduced when the molecules were on the average separated by about 50-80 Angstroms, much larger than the combined radii of the molecules. This distance was also much greater than the distance over which the excited fluorophores could diffuse within their excited state lifetimes (especially in high viscosity solvents, or solid solutions). This was the conundrum that led eventually to the discovery of FRET in condensed systems.

1.7.2. The Theories of J. Perrin and F. Perrin

1.7.2.1. J. Perrin's Model

A simple classical model to explain this polarization decrease was developed by J. Perrin (for details see below) (Perrin, 1925; Perrin, 1927). He hypothesized that the transfer of the excitation energy could hop from one molecule to the other through interactions between oscillating dipoles of closely spaced molecules. According to the classical theory of electrons in a molecule and the early quantum ideas, an excited molecule will oscillate at the frequency V corresponding to the magnitude of the excitation energy (the correct view

according to the Bohr model is that the energy of the emitted light, AE' . = hv ,

is the difference between two energy levels, but it was known that the classical electron oscillator explained much of spectroscopy). Thus, he modeled the participating molecules classically as Hertzian electric dipoles (Hertz, 1888; von Hippel, 1954). As we know, close to a Hertzian dipole the oscillating electric field resembles a static dipole. Perrin assumed that if the molecules were separated by a sufficiently small distance, the energy could be transferred to the acceptor molecule non-radiatively. He called this ''transfert d'activation''. This is of course the same type of transfer considered by Mensing, Holtsmark and Nordheim (Holtsmark, 1925; Mensing, 1925; Nordheim, 1926) in their theories on atomic vapors. However, his derivation took quite a different route.

According to his model of two interacting identical Hertzian dipoles, Perrin

calculated that this distance is approximately XJln;, where X is the wavelength of a free electric field oscillating at the frequency of the atomic electric field, v , A = c/v where c is the speed of light. A quantitative account is given below. Because he assumed that the two molecules were identical, A is also the wavelength of the light used to excite the original donor. In J. Perrin's model the molecules had the exact same frequency of their electron oscillations; that is, the two Hertzian oscillators are in exact resonance. The two dipoles will

22 R. M. CLEGG

exchange energy similar to the resonance exchange of energy between two identical weakly interacting classical mechanical harmonic oscillators (just as two identical balls hanging on identical springs attached through a rod). Using the Hertzian dipoles as a model, and assuming exact resonance, he reasoned that if the molecules were separated by less than a critical distance (which he

calculated to be XJIK ), one could detect this energy transfer between identical

molecules by measuring a decrease in the polarization of the fluorescence emission, as had been determined experimentally. Because the acceptor would on the average not have parallel transition dipole as the donor, the fluorescence emission of the acceptor would be depolarized compared to the originally excited donors, and this would lead to a decrease in the measured extent of polarization. Already, we can see one problem; molecules cannot have exact resonance with each other at all times due to the uncertainty principle. In addition, their energies will also be broadened by collisions and thermal motion, and strong interactions with the solvent broaden the spectra considerably.

1.7.2.2. Where did the J. Perrin's Idea of Dipole-Dipole Interaction Come From?

As discussed above, the concept of interacting dipoles had been considered in other contexts for some time; so it was natural for J. Perrin to consider this once it was clear that the interaction between chromophores took place at a distance large compared to the molecular diameters. The classical model of absorption and emission of radiation, which involves oscillating dipoles (Kauzmann, 1957; Stepanov and Gribkovskii, 1968; Heitler, 1984), was also well developed by this time. Perrin's model is an application of these ideas of dipole-dipole interaction to the case where one of the molecules is in an excited electronic state, and the other is in the ground electronic state. He reasoned correctly that this dipole-dipole near-field Coulombic interaction could lead to the transfer of the excitation energy from the donor to the acceptor, without direct mechanical interaction of the two molecules and without the emission of a photon from the donor. Unfortunately J. Perrin's model says that the transfer can take place over distances of about 1000 Angstroms, which he realized was greater by a factor of 20 than the experimental results (see below). A subtle point here is that J. Perrin proposed that the intervening solvent (e.g. water at 55 molar concentration) did not participate in the ''transfert d'activation''. This is not a trivial point when we remember that the surrounding solvent did not play a role in the energy transfer experiments or theories in the vapor phase. In other words, he assumed that the solvent acted only as a dielectric bath.

1.7.2.3. F. Perrin's Model

F. Perrin (J. Perrin's son) was one of the pioneers of fluorescence (Perrin, 1929), He contributed extensively to the basic concepts. He extended J. Perrin's theory of energy transfer by developing a quantum mechanical model (Perrin,


1932; Perrin, 1933) (for details see below), similar to what had been suggested for transfer of energy between different atoms in gases (Kallmann and London, 1928). However, he concluded, as had J. Perrin, that the rate of transfer takes

place proportional to y R^ ; this results in energy transfer at much longer

distances than found experimentally. F. Perrin also later considered collisions between the chromophores and the solvent molecules, as well as Doppler effects (Perrin, 1932; Perrin, 1933). These collisions broaden the spectrum of the absorption and emission of the molecules, which had been originally assumed by J. Perrin to be infinitely sharp in order to guarantee effective interaction. Such collisional and Doppler effects had been the subject of much spectroscopic research on vapors in the first two decades of the 19 ^ century, and had also been considered by Kallmann and London in the vapor phase. The broadening of the spectra is important for the following reason (this will also play a central role in Forster's theory). The energy lost by the donor must exactly equal the energy gained by the acceptor. The probability that the energy levels of the donor and acceptor molecules will simultaneously have exactly the precise values necessary to conserve energy during the transfer is much less than one. Each molecule has only a certain probability (weighted by the spectral dispersion) of being anywhere within the small spectral distributions caused by the collisions. This decreases the probability of resonance, because the two interacting dipoles must be closer than found by J. Perrin for a successful transfer of energy to take place. F. Perrin used the known theory of spectral collisional broadening to show that in the case of collisions the new distance for

energy transfer is reduced to approximately ( / l /2;r)(F/r) /^

the time between collisions of the solvent with the molecule (at

most« 10^^ sec), and T is the fluorescence lifetime (« 10~ sec ). This would reduce the distance to about 200-250 Angstroms, which was still much too long, and would lead to transfer at about 100 micromolar concentrations, instead of the experimentally determined 3-5 millimolar. In addition, the theories of the Perrins (classical and quantum mechanical) did not provide a simple means to interpret the solution experiments. This discrepancy remained a puzzle for about 20-25 years, perhaps so long due to the Second World War.

1.7.3. A Derivation of the Perrins' Estimated Distances for Two Electron Oscillators in Exact Resonance

I have given in the previous sections the results of the Perrins' calculations; the following two sections are for those more quantitative aficionados who are curious how one can arrive at their answers. I only outline the basic line of attack. The reader who is not interested in these quantitative calculations can skip the next two sub-sections; but an understanding of the dynamic rates of energy transfer when the two molecules are limited to two states, and when the molecules are in exact resonance, is central for understanding why the Perrins

where t is

24 R. M. CLEGG

calculated such long distances. This will also be important when I present the major contributions of Forster.

1.7.3.1. The Classical Derivation of J. Perrinfor two Hertzian Oscillators in Exact Resonance - with a Pinch of Old Quantum Theory

The electric field Ej^ in the near field zone surrounding a "donor"

oscillating dipole has the same form as the field of a static dipole. This is

n R

Where n is the index of refraction, R is the vector from the dipole (assume to

be a point dipole) to the point of observation, and ju^ is the dipole moment. The

arrows indicate a vector quantity, and the carrot signifies a unit vector. If

another dipole (the acceptor / /^) is placed in the E^ -field, the energy of

interaction E is

n R

where the last equality is because in J. Perrin's theory the donor and acceptor dipoles are assumed to be identical, K is the orientation factor,

' = [/>./iz,-3(/i.-^)(A>-^)]-

Using a bit of Planck's old quantum theory, we can set the energy of interaction equal to a frequency (corresponding to the energy of interaction) and therefore calculate the time period of oscillation.

E = KLi^ n^R^ = ho) ^ fi T , or , r = filE = fin^Rj KI/ la r~ I int / int ' ' int I la 11^

In other words, the rate of transfer k^ is k^ = l/r. ^ = Kf/ jfinR!' .

From the theory of a Hertzian oscillating dipole, we know that an isolated quantized Hertzian dipole radiates its energy with a time constant of

r = 3hc/a/jul , where co is the frequency of oscillation of the donor Hertzian

oscillator, c is the speed of light.


Now we simply find the distance R^ where the natural decay time of the

oscillator is equal to the time that the energy is transferred, r = r. ^ ^ ; this gives

, 3KC^ 3K: . . R'= = A' « . 0 U ' ;

n (o [In) n

or R^ « 0.2A . So we see that according to this model, the energy transfer would

take place over a distance approximately that of 1/5 the wavelength of light

radiating from the oscillating dipole oscillator. This would be about 100 nm, which is much too large.

The reason for the \j R^ dependence, which leads to such a large value

oiR^, is the assumption of exact resonance. We should be clear what this

means. Exact resonance between any two individual molecular oscillators is required in order to conserve energy. However, in an ensemble of molecules in solution, there is a distribution of energies, and the width of this distribution must be taken into account correctly (this was done in gases by Kallmann and London, and by Mensing, Nordheim and Holtsmark, by considering the broadening effects of collisions and Doppler shifts). The critical role that this broadening plays will become clear in our discussion below of Forster's first theory. In the following section I show that this is also the major reason that F. Perrin's quantum mechanical derivation arrived at a distance that was too large. However, we do not want to lose sight of the historical context. The basic model of J. Perrin was correct, and he used all the ideas prevalent at the time: the Maxwellian electromagnetic field of an oscillating Hertzian dipole, the decay constant of a quantized Hertzian dipole (semiclassical description), the Bohr condition of quantum energy jumps and the condition of resonance between the two Hertzian dipoles exchanging energy. When J. Perrin first developed his theory (Perrin, 1925), the two new quantum mechanic theories of Heisenberg (Heisenberg, 1925) and Schrodinger (Schrodinger, 1926b; Schrodinger, 1926a) were just being developed.

1.7.3.2. The Quantum Mechanical Derivation ofF. Perrin with Exact Resonance

This section is a resume of the quantum mechanical theory of F. Perrin (Perrin, 1932; Perrin, 1933). It is beyond the topic of this chapter to go into details, and a full understanding of this section requires some acquaintance with

^ The designation R was first given by Kallmann and London in their 1928 publication, and even

earlier by Holtsmark and Mensing; the same expression was then used by J. and F. Perrin, as well as Forster.

26 R. M. CLEGG

the quantum theory of two states. But it is important to show his approach, because it is the first detailed quantum mechanical description of energy transfer in solution. He draws on the ideas of Kallmann and London (Kallmann and London, 1928), but he develops a theory applicable for solution studies. He arrived at the same estimate as the classical derivation by J. Perrin; that is,

energy transfer over distances of R^ « 0.2A, where the wavelength is that of

the fluorescence of the donor. This distance is far too long, and the reasons for this are revealing from a historical point of view.

F. Perrin (Perrin, 1933) derived the rate of energy transfer between two identical molecules, each with very narrow energy states. The two molecules are only considered to have two states - the ground state and the excited state -and they are in resonance. A similar derivation was given later by Forster (Forster, 1965a), where he used this theory to illustrate exciton theory with a molecular dimer. Forster mentions that this oscillatory transfer rate for resonance between the two monomers of an exciton dimer would be difficult to measure if it took place exactly by this mechanism. Indeed, it is, and to observe optical resonance oscillations requires time resolution far beyond that available in Forster's time. In a normal FRET mechanism - i.e. Forster transfer - the transfer takes place between a single level of the donor to either a continuum, or many closely spaced states, in the acceptor; and the theory for this is different and usually uses the Fermi Golden Rule. The reader must consult F. Perrin's paper for details. Similar accounts of two interacting molecules with two states are available in QM textbooks (Pauling and Wilson, 1935; Davydov, 1965; Landau and Lifshitz, 1965; Schiff, 1968; Cohen-Tannoudji et al., 1977). A very good discussion of many aspects of two-state systems, with and without coupling to the environment, can be found in the textbook by Scully and Zubairy (Scully and Zubairy, 1997). Our short derivation follows Davydov's book (Davydov, 1965), but the derivations of other texts are identical. As mentioned above, the derivation of the basic equations is based on the work of Kallmann and London (Kallmann and London, 1928).

We consider one of the two molecules to be in an excited state, and one to be in the ground state; therefore, the first order perturbation to the energies does not go to zero, as in the London interactions between two ground state molecules (see the section on London's theory). The two wavefunctions (including their time dependence) of the whole system 4 , and 4 ^ (including both molecules) are sums of products of the stationary slates of each of the molecules ^ ( l ) and |^(2) (1 and 2 designate the different molecules)

* i ' , = - r K ( 0 n ( 2 ) + n ( 0 v ' „ ( 2 ) } e "

V2


Subscripts n and 0 designate which molecule is in the excited state n or the ground state 0 . The molecules are identical, and only one can be in an excited

state. £", and E^ are the energies of the two states (formed from the linear

combination of the products of the atomic orbitals) for which we have to solve. First order perturbation theory gives the correction to the energies of the combined system (the zero order energy of the two-molecule system, with one

molecule in the excited state, is E^ + E^) to be

2 2

K R

le^/RjD is the Coulomb perturbation between the two molecules.

Substituting the expressions for the wavefunctions and the perturbation, we get

2

AE^{R) = -AE^{R) = U{n\P\0)\'K(l,2) , K

where \n) and |o) designate that the corresponding molecule is in the n*

excited state or in the ground state. /^(l,2) is the geometric orientation factor

between the dipole moments and \{n\r\ 0)f is the square of the dipole transition

matrix element. In terms of the oscillator strength of the 0 -> n transition,

\(^n\r\o)\^ =hf^j2ju^o) (Forster, 1951; Kauzmann, 1957; Stepanov and

Gribkovskii, 1968). This derivation is very similar to the quantum mechanical theory of London's van der Waals forces; however, the London interaction

between two ground state molecules varies asl/i?^ (from the second order

perturbation). In the case where one molecule is in the excited state and we are

at exact resonance, the energy of interaction varies as l/i?^ (because in this case

we must use the first order perturbation). The excitation energy of the two states

Tj andT^ is distributed at any moment over both molecules; that is, the

separate molecules do not have well-defined energies at any time. Now we calculate the rate of energy exchange between the two molecules

according to F. Perrin. Using the calculated energies and wavefunctions, we can

write (leaving out all the details) a superposition of the states ^ ^ and ^^ •

r = - ^ [ * , + ^ J = K ( l ) l ^ „ ( 2 ) c o s ( v O + n ( l ) l ^ „ ( 2 ) s i n ( w ) } e ' " "^^ V2

v = e'l,K{\,2)l(2n,0}R')

28 R. M. CLEGG

This is a valid solution to the Schrodinger equation of the whole system, r can be substituted into the Schrodinger equation for the combined system, and the coefficients (the time dependent cosine and sine terms) in the above equation can easily be derived. So the time it takes to completely transfer the energy from 1 —> 2 (assuming that the energy is solely in molecule " 1 " at time zero) is

ITT it TicoR^ T: ^'

The rate of energy transfer in this case is therefore

K=- = ^^^^ = \^\{nm\ -(1,2) = V . {R)\ T lAjicoR h R h

We see that the rate of transfer is proportional to the energy splitting of the

two exciton energy levels —[A^^ (i^)|, which is proportional to the square of h

the transition moment |(«|^|o)| . F. Perrin's quantum rate of transfer of energy

shows the same XJR^ dependence as J. Perrin's classical derivation. Again, the

reason is that we have assumed exact resonance of the two oscillators at all times, and infinitely sharp energy levels have been assumed. In addition he chose to look at the dynamics of the oscillating term, which Kallmann and London did not. If we use the well known expression for the Einstein rate coefficient in terms of the oscillator strength, which is the natural rate of

fluorescence emission k^ competing with the energy transfer, and set the rate of

emission equal to the rate of energy transfer derived by F. Perrin, k^ - k^^, we

can calculate the distance where half the energy will be transferred within the

excited state lifetime. This gives 7?^ = 0.19A , which is again the same as was

obtained from the calculation of two classical oscillators in exact resonance; this is much too large to explain the distances of interaction measured in solution. F. Perrin calculates R^ - ^IIX . F. Perrin was aware that this distance was too

large, and he suspected that broadening of the spectra could lead to shorter R^

distances. As we learned in the section "F. Perrin's Model", he invoked

collisional broadening (by the solvent), which did decrease R^ to

approximately 25 nm. This was still much too long. We now know that collisional and Doppler broadening is miniscule compared to broadening caused by specific interactions with the solvent (such as polar effects).

The oscillation of this system between two interacting two-state molecules with sharp energy levels, where the two systems are in exact resonance, behaves


essentially the same as what is known as Rabi oscillators. Such resonant oscillators are named after Rabi (Rabi, 1936; Rabi, 1937) and were originally obtained by Giittinger (Giittinger, 1931) to describe a spin subjected to a time-dependent magnetic field. In the optical realm, this is called an optical Rabi oscillation (Cohen-Tannoudji et ah, 1977; Allen and Eberly, 1987; Loudon, 2000). The oscillatory exchange between the two states of an atom in resonance with an optical field (photons) near the resonance irequency is called optical nutation (taking over the terminology of spin nutation in magnetic resonance). The frequency is essentially the interaction energy divided by h,

( 4 / / ? ) | A £ {R)\. These are the solutions for the Bloch equations (Bloch, 1946)

describing the time dependent interaction of a two state quantum system with a weak perturbation of an oscillating electromagnetic field in exact resonance with the molecular system (Allen and Eberly, 1987). In the case of an externally applied optical field (light), the oscillatory energy exchange is between the two-state molecule and the electromagnetic field.

In F. Perrin's case the electromagnetic interaction is from the near field of the other identical molecule, and the energy exchange is between the two molecules. In general, if multiple discreet eigenstates are well isolated from all the other levels of the unperturbed Hamiltonian, then the transitions between the two levels are superpositions of Rabi oscillators. When the number of coupled states becomes very large (which is the normal case for fluorophores in solution) the different Rabi oscillations with different frequencies and amplitudes interfere, and then the system evolves with the normal irreversible character.

The important point is that the interacting molecules of F. Perrin were limited to only two states, and the two molecular oscillators were exactly the same and in exact resonance. The result is valid for the case of identical

molecules where the width in energy levels of the two states, ^^ ( l )^^ (2) and

y/^{\)\f/^{2) are less than 2 |A£ 'J , which is the energy splitting caused by the

perturbation. This results in a coherent interaction, and that is responsible for the prediction of the FRET interaction at much longer distances. However, if coherence is not lost during the interaction between two atoms or molecules with discreet eigenstates, the distance dependence of the rate of energy transfer

would b e l / i ? \ as calculated by J. and F. Perrin. This would happen, for

instance, in a vacuum between two isolated atoms. This is not an irreversible transfer of energy from a donor to an acceptor, because in this idealized case (for solution) there is an oscillatory "back and forth" transfer. See two papers by Robinson and Frosh(Robinson and Frosch, 1962; Robinson and Frosch, 1963) and references there-in for more detailed discussion of the introduction of irreversibility with multiple states, which leads to the normal exponential decay,. If the perturbation of the two state system is not resonant, a more involved calculation shows that the probability for complete transfer is less than

one. But the system still oscillates, and still shows the If R^ dependence of the

oscillation frequency (Cohen-Tannoudji et al, 1977).

30 R. M. CLEGG

1.7.4. The Contribution of W. Arnold and J.R. Oppenheimer to FRET in Photosynthesis

It is not commonly known that Oppenheimer reported the theory of FRET (with energy transfer at the correct distances) in 1941 (Oppenheimer, 1941). Perhaps even less well known (except in the field of photosynthesis) is that Oppenheimer's contribution to FRET, together with Arnold, led to a major advance in our understanding of photosynthesis. Although the full description of Arnold and Oppenheimer's contribution was not published until 1950 (Arnold and Oppenheimer, 1950) (probably because of his work in Los Alamos; in 1950 he was at the Advanced Institute at Princeton). Even though this latter publication is after the contributions of Forster (Oppenheimer's earlier abstract was not known by Forster at the time he developed his theory) it is clear that Oppenheimer had the correct solution in 1941; so I will consider this first.

1,7.4.1. Oppenheimer's Short Abstract of 1941

At the American Physical Society in 1941, a paper was presented by J. R. Oppenheimer, entitled "Internal Conversion in Photosynthesis" (Oppenheimer, 1941). As we know, Arnold had gone to Oppenheimer in 1940 consult about this problem (Arnold, 1991), and this short abstract was the result of their work together. Dutton wrote a historical account of sensitized photosynthesis, with a discussion of Arnold and Oppenheimer's contributions (Dutton, 1997)). In order to account for the rate of photosynthesis when light was absorbed by certain dyes (which absorbed where chlorophyll absorbed little), the energy must be transferred to chlorophyll, where it can then be transported to the photosynthetic reaction sites. In the abstract (Oppenheimer, 1941) Oppenheimer points out that the high efficiency of this transport of energy cannot be due to light emission and re-absorption (the probability for this is too small). However, the energy transfer could be enhanced if the chlorophyll molecules are much closer than the wavelength of the chlorophyll fluorescence (near field of a Hertzian dipole). In this abstract, the ratio of the number of quanta transferred to

the number of quanta emitted as fluorescence is given as naXjdt, where d is the closest distance of approach between the chlorophyll molecules, n is the chlorophyll concentration, a is the absorption coefficient, and 2n% is the wavelength of light in water. This is for the case of chlorophyll molecules located randomly in space, and he has already integrated the rate of energy transfer over all molecules from d to infinity (see the next section). Although in this abstract Oppenheimer does not show the solution for the rate of transfer

between only two molecules, the XJd^ result is obviously the result of

integrating \j/ from d to infinity. So it is clear that he had the correct

equation. He also assumed a quantum yield of one for isolated chlorophyll molecules.


There is a very interesting sentence in this abstract: "This transfer gives a large scale model of the internal conversion of nuclear gamma-rays.". Oppenheimer was well versed in the theory of internal conversion in nuclear physics - the non-radiative transfer of energy between a radioactive nucleus and tightly bound electrons, which is a process that competes with the emission of gamma rays. Very probably Oppenheimer already had the solution applicable to FRET at his fingertips. Arnold and Oppenheimer discuss this in more detail in their subsequent 1950 paper (Arnold and Oppenheimer, 1950). I will explain the remarkable analogy between FRET and nuclear internal conversion in the next section.

It is not surprising that this abstract was not noticed by many researchers, and certainly not by those interested in chemical and biological systems. The reasons are many: the war, the shortness of the abstract, the audience where it was presented, and the fact that he did not present a general expression for energy transfer - he only presented the integration over the distribution of acceptors that was relevant specifically to his particular problem. He became interested in this topic through his contact with William Arnold, an expert and pioneer in photosynthesis (Knox, 1996). There is no record of Oppenheimer's talk, other than this abstract, and I have not found anyone who heard Oppenheimer's presentation. At this time he was in the California Institute of Technology and Berkeley, and had not yet started to work on the war effort in Los Alamos. In an article by Arnold (Arnold, 1991), he recounts how Oppenheimer came to know about the photosynthesis problem^" . Emerson had told Arnold of the experiments indicating the transfer of energy from phycocyanin to chlorophyll. Arnold did some experiments to verify this, and then went to Berkeley in 1940 where Oppenheimer was at the time, and consulted with him about the apparent energy transfer. As soon as Oppenheimer heard about the problem, he realized the analogy with internal conversion in radioactive nuclei, and the connection to gamma rays - only the length scale was different by 10" (see the end of the next section). Arnold had known Oppenheimer since 1935 when he went to Berkley to audit Oppenheimer's course on quantum mechanics. We now turn our attention to the later paper by Arnold and Oppenheimer (A&O) from 1950 (Arnold and Oppenheimer, 1950).

1.7.4.2. Arnold and Oppenheimer's Derivation of the Rate of Energy Transfer of 1950.

In 1950, A&O published the work (Arnold and Oppenheimer, 1950) that was alluded to in the abstract of 194 PI In this paper, they provided a mechanism of energy transfer from phycocyanin (which is one of the accessory dyes in plants, in addition to e.g. carotene, xanthophylls and phycoerythrin) to chlorophyll in the blue green algae. The major question was "whether or not any of the light energy absorbed by these accessory pigments is used by the

" I thank Robert S. Knox for reminding me of this history. ^ This was Oppenheimer's last official scientific publication.

32 R. M. CLEGG

plant to reduce carbon dioxide" (all quotes in this section are from the A&O paper). It had been shown by Emerson and Lewis (Emerson and Lewis, 1942) that almost all of the energy absorbed by phycocyanin in Chroococus is used with an efficiency of approximately one in photosynthesis; the efficiency of photosynthesis was essentially as high as if the photons had been absorbed by chlorophyll. For our purposes we concentrate on A&O's contributions to the theory for energy transfer.

A&O consider three ways the energy could be transferred: 1) by direct collision, 2) by trivial emission and reabsorption of fluorescence by another molecule, and 3) by "internal conversion, or the resonance transfer of energy from one oscillator to another in resonance with it, and lying within the quasistatic rather than the wave zone field of the former". Direct collision only happens for distances on the order of atomic dimensions, and does not require resonance; in addition, the concentration of acceptors is too small, their mobility is highly restricted and the spatial separation is much too great for collisions to occur in the excited state lifetime. Trivial absorption-reabsorption happens only for distances larger than the wavelength of the emitted light, and in the far-field zone. In this case, they calculate the percentage of light

transferred is only roughly FanR = 10" which is much too small (F is the fluorescence quantum yield, <j is the absorption cross section of the dye, n is the concentration of chlorophyll - the acceptor - and R is the dimension of the cell). Therefore, they consider transfer in the near field of Hertzian dipole radiation, for which "the electric field of an oscillator (emitting primarily electric dipole radiation), which in the wave zone falls off linearly with the inverse distance from the emitter, increases, as the emitter approaches within distances small compared to a wave length, as the inverse cube of the distance". He also makes the important point, which is often not appreciated, that the energy transfer efficiency can be much greater that the fluorescence efficiency if the two molecules are close enough. This was very important for photosynthesis because it was known that the in vitro fluorescence quantum efficiency of the accessory pigments was often much smaller than the quantum efficiency of the energy that was transferred to the photosynthetic unit.

A&O then embark on a succinct two page derivation of the rate of energy transfer. We cannot give this derivation justice, but only skim the important points for this FRET history. The reader is referred to their paper (which is terse; it is not particularly easy to read unless you are versed in EM calculations). First they calculate the fluorescence yield of the Hertzian dipole by considering the Poynting vector, using the EM vector and scalar potentials (Greiner, 1986) at large distances (far field zone) from the donor oscillator; then they calculate the rate of absorption of the acceptor dipole that is in the near field zone of the donor oscillator, using again the vector and scalar potential of a Hertzian dipole, and using the expression for the absorption coefficient in terms

of the electric transition dipole moment M .


The electric field in this near zone is (keeping his notation)

2C, sin Invt E =

Inv/

3 ( 5 - r ) r

a is the amplitude of the vector potential A oscillation^^ C^ is the speed of light in the medium (water), v is the (spectroscopic) oscillator frequency, r is the distance from the oscillator to an acceptor, and h is Planck's constant. For some readers, the vector potential may be unfamiliar - it is a potential used in EM calculations with Maxwell equations (actually first introduced by Maxwell), which is an analogue of the scalar electric potential ^ ( r ) (which

Oppenheimer also uses)^\ The expression in square brackets represents the orientation factor (the direction of the E-field in vector notation).

The rate of absorption of an acceptor molecule in the near zone of the donor E-field is calculated as

[clM'[a' -^^{a-ry jr^'^l^hW

,2 M is the square of the transition dipole moment of the acceptor. Here

Arnold and Oppenheimer have assumed that the acceptor molecule, bathed in the EM field of the near zone of the donor Hertzian dipole, will absorb energy according to the same rate equations (Fermi's Golden Rule) as if the acceptor molecules were bathed in the EM field of radiation (light). The expression in

square brackets is a^ times the usual K^ orientation factor of FRET. This rate of energy transfer is the same as derived by Forster (see below), but in terms of the vector potential squared and it has been assumed that the fluorescence quantum yield of the donor is one (you just have to multiply the equation by the quantum yield). The equation for the rate of energy transfer (which requires knowing an expression for 5) is to be compared to the expression of Forster (see later).

Next, they average over all angles, giving the following equation for the rate of energy transfer to a randomly oriented acceptor at a distance of r (this

averages the orientational factor, to give K^ = 2/3).

^ The oscillating vector potential of a Hertzian dipole at position r is

A = (2a/r)cos2;rv(< - r/ C )

^ As a reminder for those who are familiar: the magnetic field in terms of the vector potential A is

B = S/ X A , and the electric field in terms of the scalar potential ^ and the vector potential is

E = -v<p- dAJdt .

34 R. M. CLEGG

T{r) = aa C,

An^hv r

They have substituted the expression M' = 3/2Cj where a is the

absorption cross section of the acceptor.

Finally they integrate over a uniform distribution of acceptor oscillators, with a density of n outside a sphere, surrounding the donor with a radius of d, which is the radius of closest approach.

aa C, n T = -

3;r hv d

This is the same rate of energy transfer given in Oppenheimer's abstract of 1941 (Oppenheimer, 1941), just in terms of other variables. Then A&O go on to calculate expectations from this theory and compare to experiments. Of course, when comparing to experimental data the overlap of the energy levels, the spectral distributions of the donor and acceptor must be taken into account (the overlap integral). 1 remind the reader of our discussion of Holtsmark's publication, where he arrived at a very similar equation with the same definition of (and dependence on) d (Holtsmark, 1925).

I now calculate some quantities that were not included in their publication.

Substituting their expression for M into the expression for ^ ( r ) above

(which is the rate of energy transfer to one acceptor at a distance of r), and setting this equal to the rate of emission from the Hertzian dipole

16;rV5^

and assuming that the orientation factor is 2/3 , we can calculate

3 C f a

64 TT' \vj ; this assumes that the quantum yield q = 1.

The absorption cross section and quantum yield are in fact dispersed over a

spectrum of frequencies; that is, cr(i/) and ^(v^). And the quantum yield

\q{v)dv is not usually 1. To account for the spectral spread of the donor and

acceptor spectra, we have to integrate over the total frequency interval. We also


remember that C, == c/« , where c is the speed of light in a vacuum and n is

the index of refraction pertaining to energy transfer. This gives

64

3 . — K

2

[^{y)q{y)j^.^ 9000 10 Kc' f f ^ ^ K M ^ ^ J , / ^ IOC ''\T J , / TT^n'^ *' v'* 128 n^nN^

To arrive at the last equality I have simply substituted the molar decadic molar

absorption coefficient of the acceptor £^{y) and the fractional quantum yield

factor ^^ ( v ) . A'^ is Avogadro's constant. This last expression is exactly the

expression for R^ that was derived by Forster (see below). I have put the

orientation factor, K back in both expressions. A&O did not carry this simple

calculation for R^ through, and they did not consider the overlap integral,

because they were interested in calculating the expression for the rate of transfer to all the acceptors located randomly, which is the expression for

T ^\jdt given above. But this shows that their derivation gives the same result

as Forster's. Although we have skipped over details of the derivation (which A&O also

did not furnish) it may have occurred to the reader that this derivation did not explicitly use a spectral distribution of frequencies of the donor and acceptor (until we added it at the end in an ad hoc manner). Why then, did the problem

that was encountered by Perrin (a XJ R^ behavior) not happen here too? It might

seem as if we are again dealing with a coherent interaction of two oscillators with identical single frequencies. The reason is subtle, but very important. In effect, the oscillators have already been considered to be incoherent, and an integration over the spectral distribution of acceptor energy levels has already been carried out, because he used the solution-phase absorption coefficient. That is, it is not a two-state system. When calculating quantum rates (e.g. rates of absorption and emission) resulting from time-dependent perturbation theory, where the perturbation and the perturbed quantum system are considered to have coherence times short compared to the interaction, one usually uses the Fermi Golden Rule. This rule is derived by taking into account the distribution of available quantum states (in the molecule, and/or in the field) by integrating over the frequency distribution. This is the same for Forster's theory. The spectral distribution is very important, but can be introduced in different ways. Oppenheimer was well aware of this. By 1941-1950, quantum mechanics had been refined a great deal after F. Perrin developed his quantum mechanical theory of energy transfer (Perrin, 1932; Perrin, 1933).

In this paper, Oppenheimer refers several times to the identical theoretical treatment of energy transfer and "the process of internal conversion that we have in the study of radioactivity". Because this is a remarkable historical

36 R. M. CLEGG

connection between nuclear physics and fluorescence spectroscopy, which has apparently not been recognized very often, I discuss this mechanism shortly so the parallel is clear^^.

An excited nucleus (e.g. a radioactive substance) can undergo spontaneous emission of a photon (a gamma ray). Depending on the nuclear levels, as the nucleus undergoes a transition from one level to another one speaks of the emission of a photon, a gamma ray, of a particular multipole transition. However, this nuclear transition can also take place by transferring energy directly to one of the orbital atomic electrons (e.g. K electrons) by a non-radiative mechanism. The atomic electron absorbs the energy non-radiatively, and is subsequently ejected through the strong interaction of the electron and the nuclear currents and charges via the large electromagnetic field in the near zone. This is named internal conversion of the electric or magnetic multipole.

The ratio of the rate of electron emission W^ to the rate of gamma-ray emission

W is defined as the internal-conversion coefficient, a = W W, and there are

extensive tables of these coefficients. An identical ratio is what Oppenheimer calculated for FRET in this publication with Arnold. The dipole field of the nuclear transition is the same as that which is considered for FRET, only on a much smaller scale. Especially for the heavier atoms, the electrons have a relatively large probability of being very close to the nucleus, where the electron interacts with the near field of the nucleus. This interaction varies

asl/i?^ , and the rate of internal conversion in the near field zone varies

asl//?^ , just as for FRET. This near field effect increases the probability of

transfer (internal conversion) dramatically - just as in FRET. Actually, the theory of this nuclear internal conversion is identical to FRET, except for the scale. The rate expression, in terms of the square of the dipole transition moments, is identical. The internal conversion electron spectra (observing the properties of the ejected electrons) are similar to observing the fluorescence of the acceptor in FRET. The observation of the gamma rays is similar to observing the decreased emission of the donor in the presence of the acceptor. I will not go into this interesting comparison further, or all the information concerning the electron and nuclear states that can be derived (Siegbahn, 1965). The important historical point is that Oppenheimer realized these similarities; he had worked extensively with the theory of internal conversion in the nucleus, and he also realized that the theory of internal conversion could be applied to energy transfer in photosynthesis.

^ I thank Hans Frauenfelder, who gave me the initial literature reference for researching nuclear internal conversion.


1.8. FORSTER'S SEMINAL CONTRIBUTION: THE MODERN, PRACTICAL DEPICTION OF FRET (FORSTER RESONANCE ENERGY TRANSFER)

This brings us to the numerous contributions of Theodor Forster, which is the culmination and end point of our history. Forster's theory and the accompanying experimental work on energy transfer is the most widely known, and most influential, of all FRET publications. The major papers are listed here (Forster, 1946; Forster, 1947; Forster, 1948; Forster, 1949a; Forster, 1949b; Forster, 1951; Forster, 1959; Forster, 1960; Forster, 1965b; Forster, 1965a; Forster, 1993). The 1993 reference (Forster, 1993) is an English translation by Robert S. Knox of Forster 's 1948 paper. Forster provided an accessible theory in a form that was practical for experimenters. It is difficult to exaggerate the influence of his work. His papers are still referenced in every paper dealing with FRET. His famous book (Forster, 1951) has also been cited thousands of times, but I suspect, since the book still exists only in German, it is seldom read, which is unfortunate because it is excellent. Regrettably, several of his papers have never been officially translated; however, there are some excellent translations of critical papers; for instance (Forster, 1993). One interesting later English paper is (Forster, 1960). His influence is remarkable, considering that he may just hold a record of citations that have not been read, or even seen, by many authors. Because the literature discussing Forster's contributions is extensive, I will not dwell on details. But this should not diminish the fact that the extensive, widespread use of FRET in physics, engineering, chemistry, biology and medicine, are due to the description of energy transfer given to us by Forster. The following account is meant to highlight why this is so.

Forster apparently became interested in the energy transfer process because of the known effectiveness of photosynthesis (just as Oppenheimer); although, he was also aware of the work of the Perrins. Experiments (Emerson and Arnold, 1932a; Emerson and Arnold, 1932b) had shown that the capture and utilization of the light energy by plant leaves was much more effective than would be expected if it were required that photons exactly hit the reaction centers (that is, there are too few reaction centers in the leaves to explain the very effective extent of energy capture). Forster knew of these results, and reasoned that an extremely efficient transfer of energy between the chlorophyll molecules must be responsible for the eventual diffusion of the energy, which was absorbed over the whole surface of the leaf, into the relatively sparse reaction centers. He assumed that this energy diffusion is due to energy rapidly hopping between molecules. He was also aware of the earlier work of the Perrins, and of other data indicating energy transfer over distances longer than the molecular diameters. He gives a thorough account of this early work in his initial papers.

In his first paper on FRET (Forster, 1946), he correctly developed the basic theoretical background of FRET (an English account of this derivation has been given (Clegg, 1996)). First he reviewed the mechanisms proposed by the Perrins. He then proceeded to take three critically important steps that allowed

38 R. M. CLEGG

him to derive a quantitative theory of non-radiative energy transfer (Forster, 1948). The reader should keep in mind our discussion of A&O's work and that of J. and F. Perrin, while reading the following. Forster did not know of Arnold and Oppenheimer's derivation, I have included information there that will not be repeated here.

1) Forster was well versed in the new quantum theory describing the electronic structure of molecules. He knew that the atomic vibrations in complex molecules and interactions with the solvent in condensed media considerably broaden absorption and emission spectra. These theories of broad condensed system spectra had been developed mainly subsequently to the original work of the Perrins. The theoretical quantum mechanical treatments of spectroscopic transitions had clearly shown the necessity of taking into account the effect of broadened energy distributions when calculating the rate of a kinetic process between two quantum states (Dirac, 1927). This leads to the famous Fermi Golden Rule, which quantitatively relates the rate of transition between states of a quantum molecular system that is perturbed by an oscillating electromagnetic field, such as a light wave^^ In a similar manner, Forster took into account the broad spectral dispersion of the donor fluorescence and acceptor absorption. The actual experimentally determined spectral breadths correspond to much broader energy dispersions than that calculated by F. Perrin from collisions, or from the spectra in vapors. This significantly affects the estimates of the probability that the frequencies (energy differences between the excited and ground states) of the donor and acceptor molecules will be simultaneously nearly identical. See our discussion of this in the section "F. Perrin's Model". He correctly took into account the overlapping oscillation frequencies of the donors in the excited state and the acceptor molecules in the ground state. In his first paper (Forster, 1946), Forster treated this frequency overlap semi-classically and semi-quantitatively. Shortly thereafter (Forster, 1947; Forster, 1948) he gave a full quantum treatment. He showed how to express this overlap quantitatively in terms of the frequency dependent "oscillator strengths" of the classical or quantum spectroscopic transition dipoles (Forster, 1951). The frequency dependence of oscillator strengths of a spectroscopic transition is a theoretical way to represent the shape of the measured spectroscopic spectra. The stringent requirement that the two molecules be in resonance to effect energy transfer, is the reason that Forster emphasized the name "resonance". Of course, the requirement for resonance had been emphasized by all the previous theories, classical and quantum mechanical.

2) Forster realized that the classical theory of interacting oscillating dipoles, which he had shown could lead to an exchange of energy between molecules, is very similar to the interaction of a single molecular transition

^ The Fermi Golden Rule was actually derived by Dirac in 1927.The common attribution of the golden rule to Enrico Fermi is misleading; it was Dirac who developed time-dependent perturbation theory, including this formula. It was coined Fermi's Golden Rule because in Fermi's famous lectures in Chicago he used the rate expression, and called it the golden rule (of course, he did not call it Fermi's Golden Rule). Although Fermi never claimed, or suggested, any priority, his name has remained associated with this rate expression.


dipole with the oscillating electric field of light, which theoretically describes spectroscopic transitions of absorption and fluorescence. Thus he was able to develop a quantitative theory (see the equations below) of the rate of transfer from an excited donor molecule to a ground state acceptor molecule in terms of what is now known as the overlap integral. We have seen that A&O applied the same reasoning (which was also applied in the internal conversion of nuclear transitions). However, A&O's rate expression does not present the integrated overlap of the absorption and emission spectra, and they do not give an explicit expression for the orientation factor (that is, they leave it up to the reader to derive this from the vector expression). A&O were not interested in general applications of FRET; they were solving a particular problem; although their equations are quite general and are identical to Forster's. Forster explicitly deals with the vibrationally broadened spectra and he derives the overlap integral in terms of the measured absorption and fluorescence spectra. Forster also gives

helpful expressions for the orientation factor, and discusses K^ (and introduces this notation). The overlap integral is the integral of the product of the donor fluorescence spectrum and the acceptor absorption spectrum over the entire

frequency range, divided by v\ \ will not go into the factor v^, but it also arises in the expression for the rate of transfer I have given in the discussion of Arnold and Oppenheimer's work^^ The overlap integral represents the probability that the two molecular transition dipoles will have the same frequency. This was a major conceptual step, because these spectroscopic transitions can be measured experimentally, independent of the FRET measurement. It is also important to realize that no "spectroscopic transition" takes place; that is, there is neither the emission of a photon nor the absorption of a photon in the FRET process. Fluorescence comes into the picture because the method normally used to measure FRET is fluorescence; it is not part of the actual physical process. It just happens that the electromagnetic interaction between the donor and acceptor can be described in terms of the same theoretical expression as the normal absorption and emission of a photon^^ We should keep in mind that the earlier work of Mensing, Holtsmark, Nordheim, Kallmann & London, the Perrins, and Arnold and Oppenheimer also realized that the dipole-dipole interactions took place over these spectroscopic transition moments. However, Forster showed unambiguously how to connect this overlap to measured absorption and emission spectra, and gave an explicit expression for the overlap integral, opening the way to quantitative interpretation of experimental data. He included the effect of the index of refraction, which affects all electric interactions in condensed media at these

very high optical frequencies,« 10 sec' . The relative orientation of the two dipoles will control the strength of their interaction at a certain distance. This is the infamous kappa square, which has led to great discussion in the literature

° This requires a bit of calculation: a has a factor of v , and a has a factor oiv . ^^ It is the perturbation of an oscillating E-field (of the transition moment) interacting with the

electronic ground state of a molecule.

40 R. M. CLEGG

and is one of the major complications of FRET interpretations. The quantum yield of the donor fluorescence, or the fluorescence lifetime, in the absence of the acceptor, appear because these parameters are related to the strength of the oscillating dipole (classical model) or the transition dipole (quantum model).

3) Forster's model included quantitatively the ijR^ distance dependence of

the dipole-dipole interaction. This corrected the XJR^ dependence of the

Perrins' solutions for energy transfer in solution. Similar to previous theoretical

accounts, Forster calculated the distance R^ where the rate of the energy

transfer was equal to the rate of fluorescence emission. This distance is now

universally known as R^^^. He showed how R^ can be calculated from the

overlap integral, the quantum yield of the acceptor, the lifetime of the donor in the absence of an acceptor and the effective index of refraction, resulting in a very concise expression for the distance dependence of FRET efficiency (see the equations below).

His resulting famous equations for the rate of energy transfer and the efficiency are (Forster, 1948; Forster, 1951; Ketskemety, 1962) :

9000£«10 rate , , = ^ r = —

energy transfer ET ^ ,^c^

128;r n N T i? * V T \R

Efficiency of transfer = E = ^ET

+1

/ {y)s yv^dv/v is the overlap integral of the normalized fluorescence

spectrum of the donor and the extinction coefficient of the acceptor, K^ is the kappa square factor that takes into account the relative orientation of the two

transition dipoles, A^ is Avogadro's constant, n is the refractive index of the

environment of the donor and acceptor, c is the speed of light, r^ is the

^ The terminology R , or r , was introduced already in the very early papers of Mensing and

Holtsmark. " Forster's equation for the rate of energy transfer was first published with a printing mistake; but

one that seemed to plague several of his following publications, and even in his 1951 book this mistake persisted (there was apparently one printing of the book where this factor was correct).

This was the TT factor in this equation, which was printed in several of his publications as TT . The mistake was first noticed by Ketskemety, who in 1962 had offered a different fully classical derivation. Ketskemety's correction was acknowledged by Forster. It was undoubtedly a printing error, and is well known; but it is interesting that this error persisted even into much later works of Forster, (for instance, Forster, 1959, 1960). It was given correctly in Forster, 1965.


average dwell time of the donor in the excited state in the absence of the acceptor (this is the same time as the fluorescence lifetime), and R is the distance between the centers of the two molecular transition dipoles. R^ is, as

before, the distance between the two fluorophores where the rate of energy transfer is equal to the rate of fluorescence of the donor without the presence of the acceptor. These same equations were given in a short historical account of FRET, which had a few misprints.^^ Regarding the equation for the efficiency £", it is interesting to note, as mentioned above in the section on Kallmann and London's work, that Kallmann and London wrote this equation in the process of their derivation equation 17 of their paper (Kallmann and London, 1928)).

Forster's theory is the basis of our present understanding of FRET and most of its applications. There have been several extensions of the theory to other experimental conditions, also by Forster himself; however, within the validity of his model, which encompasses most of the applications, Forster's original theory still applies. Forster's original theoretical description of energy transfer set the stage for all subsequent applications of FRET in many fields of research (from pure physics to biology), and it is his theory that still is used to interpret experimental results. He extended the original ideas of the Perrins involving the well known interaction of molecular dipoles; however, his insight and great contribution was to provide the quantitatively correct and very practical description of the FRET process in terms of experimentally accessible parameters. By relating the rate of energy transfer to purely experimentally available parameters (except for the kappa square term, which can usually be estimated; there is an extensive and still hotly debated literature dealing with this factor), he provided the general theoretical framework for all FRET applications. FRET has been shown to be broadly applicable and extremely informative for determining molecular interactions, and to measure molecular distances that are impossible to determine otherwise.

1.9. MATURATION OF FRET

Since the seminal papers by Forster, there has literally been a flood of papers, theoretical and experimental, dealing with FRET. Although Kallmann and London, and J. and F. Perrin, as well as Mensing, Nordheim and Holtsmark, set the stage for the correct interpretation for FRET, Forster furnished the clear and explicit connection to experiment. As in all science endeavors, once a theory is developed that can easily be compared to experiment, this opens the door to wide-ranging experimentation in diverse fields.

Extensive Russian literature, much of it unavailable to non-Russian readers, contributed to the theory of energy transfer and to fluorescence in general. I

^^ In this earlier paper the exponents that are supposed to be to the 6*** power were mistakenly given

as the 1/6* power, and C was given as e ; and the 9 should be 9000.

42 R. M. CLEGG

have not mentioned these contributions, although Forster mentioned some of the early work in his first papers. Most references to this work can be found in an English book by Agranovich and Galanin (Agranovich and Galanin, 1982); the authors have included a short history of energy transfer. Also, many references to work following Forster can be found in a recent very readable book on FRET (Van Der Meer et aL, 1994). Another review (Clegg, 1996) also has many references to much of this literature. The chapters of a recent book edited by Andrews and Demidov (Andrews and Demidov, 1999) have accounts of more recent advances, with many references. I apologize to the many authors who have made vital contributors to FRET who have not been explicitly mentioned; this is simply because of lack of space, as well as the time-frame of this history.

1.10. EPILOGUE

Thus, our history of FRET closes with Forster. We have covered the major contributions to the ideas of energy transfer leading up to Forster. There have been many critical contributions since Forster's first papers. The history following Forster's early work (and also contributed by Forster) is extensive (actually enormous) and rich in innovative experimentation and theory. But this will have to await another chapter of FRET history.

A final point: This chapter has covered mostly work previous to Forster, leading up to his final, practical expression for FRET. Forster always gave explicit reference to the pioneers who preceded him. Yet I would recommend restricting the acronym FRET to Forster Resonance Energy Transfer. Forster put all the pieces together. His theory has been tested thoroughly, and if the conditions for which his derivation is applicable are met, his theory has always been found to be valid. There are other modes of energy transfer, and circumstances where Forster transfer is not valid; these require different theoretical foundations. However, reserving "Forster" for the "F" in FRET, whenever we mean Forster transfer, gives credit to the person who made it possible for us to gain valuable, quantitative insight into so many processes at the molecular scale, through relatively easy experiments.

1.11. ACKNOWLEDGEMENTS

I thank Hans Frauenfelder for suggesting a good source for internal conversion, Govindjee and Zigurts Majumdar for critical reading of the manuscript and for many conversations and sharing my enthusiasm for history, and Robert Knox for sharing a few historical tips with me. All responsibility for the reconstruction of the history is of course mine.


1.12. REFERENCES

Agranovich VM and Galanin MD (1982) Electronic Excitation Energy Transfer in Condensed Matter Agranovich VM, Maradudin A A, editors Amsterdam: North-Holland Publishing Company 371 pp.

Allen L and Eberly JH (1987) Optical resonance and two-level atoms New York: Dover Publications, Inc. 233 pp.

Andrews DL and Demidov A A (1999) Resonance Energy Transfer Chichester: John Wiley & Sons 468 pp.

Arnold W and Oppenheimer JR (1950) Internal Conversion in the Photosynthetric Mechanism of Blur-Green Algae J. Gen. Physiol. 33:423-435.

Arnold WA (1991) Experiments Photosynthesis Research 27:73-82. Beutler H and Josephi B (1927) Resonanz by Stossen zweiter Art in der Fluoreszenz und

Chemilumineszenz Naturwiss 15:540. Beutler H and Josephi B (1929) Resonanz by Stossen zweiter Art in der Fluoreszenz und

Chemilumineszenz. Z. f Phys. 53:747. Bloch F (1946) Nuclear induction Phys. Rev. 70:460-474. Bohr N (1913) On the Constitution of Atoms and Molecules Philosophical Magazine 26:1-25, 476-

502 and 857-875. Bom M and Jordan P (1925) On quantum mechanics Z. Phys. 34:858-888. Carlo G (1922) Uber Entstehung wahrer Lichtabsorption un scheinbare Koppelung von

Quantensprtingen. Z Physik 10:185-199. Carlo G and Franck J (1922) tJber Zerlegugen von Wasserstoffmolekiilen durch angeregte

Quecksilberatome. Z Physik 11:161 -166. Casimir HBG (1948) Proc. K. Ned. Akad. Wet. 51:79. Clegg R (1992) Fluorescence resonance energy transfer and nucleic acids. In: Lilley D, JE D,

editors Methods Enzymol San Diego: Academic Press, pp 353-388. Clegg R (2004a) Nuts and bolts of excitation energy migration and energy transfer. In: Govindjee

GCPa, editor Chlorophyll a Fluorescence: A Signature of Photosynthesis New York: Springer.

Clegg R (2004b) The vital contribution of Perrin and Forster Biophotonics International September:42-45.

Clegg RM (1996) Fluorescence resonance energy transfer In: Wang XF, Herman B, editors Fluorescence Imaging. Spectroscopy and Microscopy. New York: John Wiley & Sons, Inc.pp 179-252.

Cohen-Tannoudji C, Dui B and Laloe F (1977) Quantum Mechanics New York: John Wiley & Sons 903-1523 pp.

Conway AW (1907) Sci Proc. R. Dublin.Soc. xi(March):181. Davydov AS (1965) Quantum Mechanics additions tbDtHw, editor Oxford: Pergamon Press 680 pp. Debye P (1920) Die van der Waalsschen Kohaesionskraefte Phys. Z. 21:178-187. Debye P (1921) Molekularkraefte und ihre Elektrische Deutung Phys. Z. 22:302-308. Dirac P (1927) Proc. Roy. Soc. Al 14:243. Dirac PAM (1926) Theory of quantum mechanics Proc. Roy. Soc. (London) Al 12:661-677. Dutton HJ (1997) Carotenoid-sensitized photosynthesis: Quantum efficiency, fluorescence and

energy transfer Photosynthesis Research 52:175-185. Einstein A (1905) Ueber einen die Erzeugung und Verwandlung des Lichtes betreffenden

heuristischen Gesichtspunkt Ann. D. Phys. Xvii: 132-148. Einstein A (1906) Zur Theorie der Lichterzeugung und Lichtabsorption Ann. D. Physik Xx:199-

206. Einstein A and Infeld L (1966) The evolution of physics: from early concepts to relativity and

quanta New York: Touchstone Books, Simon and Schuster 302 pp. Emerson R and Arnold W (1932a) J. Gen. Physiol. 15:391-420. Emerson R and Arnold W (1932b) The Photochemical Reaction in Photosynthesis J. Gen. Physiol.

16:191-205. Emerson R and Lewis CM (1942) The Photosynthetic Efficiency of Phycocyanin in Chroococus

and the problem of Carotenoid Participation in Photosynthesis J. Gen. Physiol. 25:579-595.

Forster T (1946) Energiewanderung und Fluoreszenz. Naturwissenschaften 6:166-175.

44 R. M. CLEGG

Forster T (1947) Fluoreszenzversuche an Farbstoffmischungen. Angew Chem A 59:181-187. Forster T (1948) Zwischenmolekulare Energiewanderung und Fluoreszenz. Ann Phys 2:55-75. Forster T (1949a) Expermentelle und theoretische Untersuchung des zwischengmolekularen

Ubergangs von Elektronenanregungsenergie. A Naturforsch 4A:321-327. Forster T (1949b) Versuche zum zwischenmolekularen Ubergang von Elektronenanregungsenergie.

ZElektrochem 53:93-100. Forster T (1951) Fluoreszenz Organischer Verbindungen. Gottingen: Vandenhoeck & Ruprecht 315

pp. Forster T (1959) Transfer mechanisms of electronic excitation. Discuss Faraday Soc(27):7-17. Forster T (1960) Transfer mechanisms of electronic excitation energy. Radiat. Res. Suppl.(2):326-

339. Forster T, editor (1965a) 1. Delocalized excitation and excitation transfer New York: Academic

Press Part III Action of light and organic crystals 93-137 pp. Forster T (1965b) Delocalized excitation and excitation transfer. In: Sunanoglu O, editor Modem

Quantum Chemistry New York: Academic, pp 93-137. Forster T (1993) Intermolecular energy migration and fluorescence In: Mielczarek EV, Greenbaum

E, Knox RS, editors Biological Physics New York: Americal Institute of Physics, pp 148-160.

Franck J (1922) Einige aus der Theorie von Klein und Rosseland zu ziehende Folgerungen tiber Fluorescence, photochemische Prozesse und die Electronenemission gliihender Korper. Z. Physik. 9:259-266.

Gaviola E and Pringsheim P (1924) Uber den Einfluss der Konzentration auf die Polarisation der Fluoreszenz von Farbstofflosungen Z. Physik 24:24-36.

Gillispie CC (1960) The Edge of Objectivity: An Essay in the History of Scientic Ideas Princeton: Princeton University Press 562 pp.

Greiner W (1986) Theoretische Physik: Klassische Elektrodynamik. Giittinger P (1931) Das Verhalten von Atomen im Magnetishen Drehfeld Zeitschrift fur Physik

73:169-184. Heisenberg W (1925) Uber quantentheoretische Umdeutung kinematischer un mechanischer

Beziehungen Z. Phys. 33:879-893. Heitler W (1984) The Quantum Theory of Radiation New York: Dover 430 pp. Hertz H (1888) Ann. Physik 36:1. Hettema H (1995) Bohr's Theory of the Atom 1913-1923: A Case Study in the Progress of

Scientific Research Programmes Studies in History and Philosophy of Modem Physics 26(3):307-323.

Holtsmark J (1925) Uber die Absorption in Na-Dampf Z. Physik 34:722-729. Israelachvili J (1992) Intermolecular & Surfaqce Forces New York: Academic Press 450 pp. Kallmann H and London F (1928) Uber quantenmechanische Energieiibertragungen zwischen

atomaren Systemen. Z. Physik. Chem. B2:207-243. Kauzmann W (1957) Quantum Chemistry. Keesom W (1912) On the deduction of the equation of state from Boltzmann's entropy principle In:

Kammerlingh H, Ijdo E, editors Communications Physical Laboratory Leyden: University of Leiden, pp 3-20.

Ketskemety I (1962) Zwischenmolekulare Energieubertragung in fluoreszierenden Losungen. Z. Naturforsch. 17A(17A):666-670.

Knox RS (1996) Electronic Excitation Transfer in the Photosynthetic Unit: Reflections on Work of William Amold Photosynthesis Research 48:35-39.

Kuhn H (1970) Classical aspects of energy transfer in molecular systems. J. Chem. Phys. 53:101-108.

Landau LD and Lifshitz EM (1965) Quantum Mechanics. Non-Relativistic Theory Oxford: Pergamon Press 616 pp.

London F (1930) Zur Theorie und Systemmatik der Molekularkraefte Z. Phys. 63:245-279. London F (1937) The general theory of molecular forces Trans. Faraday Soc. 33:8-26. Loudon R (2000) The Quantum Theory of Light Oxford: Oxford University Press 438 pp. Margenau H (1939) Van der Waals Forces Reviews of Modem Physics 11(1): 1-35. Maxwell JC (1873) A Teeatise on Electricity and Magnetism London: Oxford. Mensing L (1925) Beitrag zur Theorie der Verbreitung von Spektrallinien Z. Physik 34:611-621. Nordheim L (1926) Zur Theorie der Anregung von Atomen durch Stosse Z. Physik 36:497-539.


Oppenheimer JR (1941) Internal conversion in photosynthesis Phys. Rev. 60:158. Pauling L and Wilson EB (1935) Introduction to Quantum Mechanics New York: McGraw-Hill

Book Company 322-325 pp. Perrin F (1929) La Fluorescence des Solutions Ann. de Phys. XII: 169-275. Perrin F (1932) Theorie quantique des transferts d'activation entre molecules de meme espece. Cas

des solutions fluorescentes. Ann. Chim. Phys. (Paris) 17:283-314. Perrin F (1933) Interaction entre atomes normal et activite. Transferts d'activitation. Formation

d'une molecule activitee. Ann. Institut Poincare 3:279-318. Perrin J (1925) Fluorescence et radiochimie Conseil de Chemie, Solvay, 2iem, Paris, 1924 Paris:

Gauthier & Villar. pp 322-398. Perrin J (1927) Fluorescence et induction moleculaire par resonance. C.R. Hebd. Seances Acad. Sci.

184:1097-1100. Planck M (1900) Uber cine Verbesserung der Wienschen Spektralgleichung (1) Verhandlung der

physikal. Gesellschaft 2:202-204. Planck M (1901) On the Law of Distribution of Energy in the Normal Spectrum Annalen der Physik

4:553. Pringsheim P (1928) Luminescence und Phosphorescence im Lichte der neueren Atomtheorie

Berlin: Interscience 234 pp. Pringsheim P (1949) Fluorescence and Phosphorescence New York: Interscience Publishers 794 pp. Rabi II (1936) On the Process of Space Quantization Phys. Rev. 49:324-328. Rabi II (1937) Spin Quantization in a Gyrating Magnetic Field Phys. Rev. 51:662. Ritz W (1908) Phys. Zs. ix:521. Robinson G and Frosch R (1962) Theory of electronic energy relaxation in the solid phase. J. Chem.

Phys. 37(1962-1973). Robinson G and Frosch R (1963) Electronic excitation transfer and relaxation. J. Chem.

Phys.(38):l 187-1203. Schiff L (1968) Quantum Mechanics New York: McGraw Hill Book Co. 544 pp. Schrodinger E (1926a) Quantisierung als Eigenwertproblem Ann. Physik 81:109-139. Schrodinger E (1926b) Quantisierung als Eigenwertungsproblem Ann. Phys. 79:361-376, 489-527. Schrodinger E (1927) Energieaustausch nach der Wellenmechanik Ann. Physik 83:956-968. Scully MO and Zubairy MS (1997) Quantum Optics Cambridge: Canbridge University Press 630

pp. Siegbahn K, editor (1965) Alpha-, Beta- and Gamma-Ray Spectroscopy: North-Holland Publishing

Company 1 & 2 1742 pp. Simpson TK (1997) Maxwell on the Electromagnetic Field: A Guided Study Flaumenhaft hM,

editor New Brunswick: Rutgers University Press 440 pp. Stepanov BI and Gribkovskii VP (1968) Theory of Luminescence Bristol: ILIFFE Books Ltd. 497

pp. Van Der Meer WB, Coker III G and Chen S-Y (1994) Resonance Energy Transfer: Theory and

Data New York: John Wiley & Sons 177 pp. von Hippel AR (1954) Dielectrics and Waves Cambridge: The MIT Press 284 pp. Weigert F (1920) Verh. d. D. Phys. Ges. 23:100. Whitaker SE (1989a) A history of the theories of aether and electricity New York: Dover

Publications 434 pp. Whitaker SE (1989b) A history of the theories of aether and electricity New York: Dover

Publications, Inc. 319 pp. Wood RW (1934) Physical Optics Washington, DC: Optical Society of America 846 pp. Wu P and Brand L (1994) Resonance energy transfer: Methods and applications. Anal. Biochem.

218:1-13.

TRICHOGIN TOPOLOGY AND ACTIVITY IN MODEL MEMBRANES AS DETERMINED

BY FLUORESCENCE SPECTROSCOPY

B. Pispisa, L. Stella, C. Mazzuca and M. Venanzi ®

2.1. INTRODUCTION

Antibiotic peptides, exhibiting activity against different types of microorganisms, are part of the innate immune system of most organisms.^ Their mechanism of action is basically that of altering the permeability of cell membranes, bringing about the cell death by collapse of transmembrane electrochemical gradients and osmolysis,-^ though, despite the large number of studies, the molecular details of the mechanism are still uncertain.^'^

In view of the fact that bioactivity relies on both peptide affinity for membranes and ability to self-associate, it is easily understandable why cationic and hydrophobic peptides behave differently. Cationic peptides can bind to the charged surface of membranes, but their insertion in the hydrophobic core of the phospholipid bilayer or their aggregation is hindered by electrostatic effects. Therefore, their activity is best described by the Shai-Matsuzaki-Huang model, also known as the "carpet" or "toroidal pore" model, in which peptides bind to the membrane surface in a carpet-like fashion, by insertion into the polar headgroups region only. As a result, an unfavorable elastic tension arises, leading to the formation of transient defects or pores.^ On the other hand, because hydrophobic peptides include in their sequence only a few or no charged amino acids, they tend to bring together, so that their mechanism of action is best described by the so-called "barrel stave" model, in which several peptide chains assemble in a transmembrane orientation, forming well defined channels. These two classes of peptides also differ in activity, in the sense that anionic lipids, such as those in bacteria, but not the zwitterionic bilayers, present in mammal or fungal cells, favor the binding of cationic peptides only.

Department of Chemical Sciences and Technologies, University of Roma Tor Vergata, Rome, Italy 1-00133

47

48 B. PISPISA ETAL.

Therefore, these latter peptides are more selective than the hydrophobic ones. A similar difference holds in aqueous solution, in that the entropy-driven hydrophobic interactions definitely favor gathering, while the opposite occurs for cationic peptides, owing to electrostatic effects.

Lipopeptaibols^ are members of a unique family of membrane active peptides, comprising a linear sequence of 6-10 aminoacids, a large amount of the helix-promoting Aib (a-aminoisobutyric acid),^'^ a N-terminal fatty acyl group and a 1,2-amino alcohol at the C-terminus. One of the main components of this family is Trichogin GA IV (TR), a natural peptide having the following primary structure, where Oct is «-octanoyl, and Lol leucinol.

Oct-Aib^-Gly^-Leu^-Aib^-Gly^-Gly^-Leu^-Aib^-Gly^-Ile^^-Lol (TR)

This peptide can modify membrane permeability, displaying both antibacterial and hemolytic effects. However, the full details of bioactivity are still unknown, and different models, including bilayer destabilization,^ channel formation ^ or diffusion through the membrane as ion carrier,^ ^ have been put forward.

We report here the results of a thorough investigation on the behavior of this peptide in both aqueous solution and model membranes. Because peptide-membrane interactions are best studied by fluorescence experiments, two fluorescent trichogin analogs, both having a leucine methyl ester at the C-terminus, replacing Lol, were employed. One, denoted A3, is labeled with azulene [Aal: P-(l-azulenyl)-L-alanine], replacing Leu 3, and the other, labeled with a fluorene moiety (Fmoc: fluorenyl-9-methylcarbonyl group) linked to the side-chain of 2n-diamino-L-butyric acid (Dab), replacing He 10, is denoted FIO. The primary structure of the analogs is as follows, where TR-OMe is the reference peptide, and Boc a protective group (/-butyloxycarbonyl).

Oct-Aib-Gly-Leu-Aib-Gly-Gly-Leu-Aib-Gly-Ile-Leu-OMe (TR-OMe) Oct-Aib-Gly-Aal-Aib-Gly-Gly-Leu-Aib-Gly-Dab(Boc)-Leu-OMe (A3) Oct-Aib-Gly-Leu-Aib-Gly-Gly-Leu-Aib-Gly-Dab(Fmoc)-Leu-OMe (FIG)

The fluorophores, chosen because they can act as a donor-acceptor pair in Forster energy transfer, ^ were incorporated in such positions as to substitute bulky side chains. This would minimize perturbations of peptide structure and activity.

2.2. PROPERTIES OF THE FLUORESCENT ANALOGS

To answer the question as to whether A3 and F10 peptides exhibit the same structural features and bioactivity of TR, both CD and peptide activity measurements, these latter based on peptide-induced leakage of small

TRICHOGIN TOPOLOGY AND ACTIVITY 49

unilamellar vesicles of ePC/cholesterol,* containing carboxyfluorescein, were carried out.

Figure 2.1 illustrates the far-UV CD spectra of TR-OMe, F10 and A3 in a structure-supporting solvent, such as methanol. The similarity of the spectra of Tr-OMe and FIO suggests that the Fmoc fluorescent label does not induce any significant structural perturbation, and that FIO analog exhibits the same structural features of TR. The CD spectrum of A3 is somewhat different, in the sense that the molar ellipticity is smaller than that of the other peptides. In all cases, the lowest accessible wavelength negative band falls at around 207 nm, and another negative band is observed at around 230 nm, the ratio of the molar ellipticity at the two wavelengths, [0]23o/[®]2O7' being around 0.4 for TR-OMe and FIO, and around 1 for A3. A value near 0.4 is suggestive of a right-handed 3io-helical conformation. ' "

Interestingly enough, despite the fact that the fluorene moiety in FIO absorbs in this spectral region, an induced dichroic contribution was not detected, indicating that the fluorophore is not significantly perturbed by the asymmetric peptide chain. This, in turn, suggests that fluorene has a relatively high conformational mobility, in agreement with the fact that it is linked to the butyric side-chain. By contrast, azulene in A3 is less mobile than Fmoc in FIO, because it is linked to the peptide backbone through a short side-chain, being thus capable of experiencing an induced dichroic contribution from the chiral

6 o

'o

0

-2.0

-4.0

^ _ /

/ /

L /

c

b "\

a

\

'--

1

— — "

,''

^ ' ' - ; > ^ ^

200 225 A.(nm)

Figure 2.1. Circular dichroism spectra of TR-OMe ( a), FIO ( b) and A3 ( c) in methanol.

250

' Liposomes were normally formed by egg phosphatidylcholine (ePC) and cholesterol (1:1 molar ratio). The experimental details are reported elsewhere.

50 B. PISPISA ETAL.

environment. This leads to a perturbation in the CD spectral patterns, as that observed, even if the structure is not altered.

As far as the CD spectra in water are concerned, they are reminiscent of unordered species, despite the presence of the helix-promoting Aib in the main chain. By contrast, when constrained in the bilayer, the peptides attain a a-helix conformation, as shown in Figure 2.2, where typical CD spectra in water and in membrane are reported. The conformational transition from a disordered state to a-helix on going from water to liposomes has been already observed for other peptide-membrane systems. It is often described as partition-folding coupling, which is thought to take place through a transient, prefolded state, into which the disordered peptide transforms, mediated by the membrane interface.^^ ' ^ However, excluded volume effects are very likely sufficient to bring about the transition to an ordered structure.

The membrane permeability was estimated by measuring the peptide-induced release of a fluorescent marker (carboxyfluorescein), entrapped inside the liposomes. The results are shown in Figure 2.3, indicating that the behavior of A3, FIO and TR-OMe is quite similar.

To summarize, both FIO and A3 analogs exhibit structural characteristics and membrane permeability very similar to those of TR-OMe. Therefore, they can be considered as a good model of the natural trichogin peptide.

2.3. AGGREGATION IN WATER

Trichogin experiences aggregation in water because of its high hydrophobicity, which is also responsible for its insolubility, even at concentrations as low as 50 fiM. Around the same concentration, aqueous solutions of TR-OMe, A3 or FIO are opalescent, suggesting that they aggregate,

o

I 6 O

0

200 220

Figure 2.2. CD spectra of TR-OMe in water (a) and in membrane (b).

240 Mnm)260


though ESR data show that this process is also accompHshed in apolar solvents, such as toluene or chloroform, ^ ^ thereby indicating that hydrophobicity is not the only driving force for aggregation to occur.

We then postulate the occurrence of the following two-state equilibrium: nM <^ M„ (1)

where M is the monomer, Mj the aggregate, n the number of monomers in the

aggregate and K the equilibrium constant, defined as:^^

(2) K'^-l= —

X' and X being the molar fraction of aggregate and monomer, respectively. The total molar fraction of peptide, X , is then:

Xt = X + nX' (3)

and the fraction of peptide chains partecipating to aggregates, f, is:

f = ^ (4) X,

Fluorescence decay experiments prove that equilibrium (1) applies to the analogs examined in aqueous solution. For instance. Figure 2.4 illustrates the

decay curve of FIO peptide at two concentrations, together with that of the reference, i.e. fluoren-9-acetic acid (Fmoc-OH). The decay curve of Fmoc-OH

is well described by a single exponential (x = 5.6 ± 0.1 ns; ^ = 1.0), while that

PH

1

0.8

0.6

0.4

0.2

0

-

•H .PAKW

A /u

11 I J - l l — 1 L_L.,i,IJ,.lJj

10-' 10" lO'"" 10-[peptide] (M)

Figure 2.3. Perturbation of the bilayer by TR-OMe (circles), FIO (triangles) and A3 (squares): peptide-induced release of carboxyfluorescein from the bilayer. [Lipid] = 60 j^M. The fractional release was determined 20 min after peptide addition.

52 B. PISPISA ETAL.

of FIO by a biexponential function, having lifetimes Xj = 5.6 ± 0.1 ns and X2 =

0.87 ± 0.08 ns (x^ = 1.0), irrespective of concentration. In principle, each decay time could arise either from a single species or from many species, all having a similar quenching rate. This latter hypothesis implies that all these species have very similar structural and geometric features, so that they are indistinguishable within the resolution time.

Where the pre-exponential factors of the shortest lifetime component is reported as a function of peptide concentration, the plot of Figure 2.5 is obtained. The data are well fitted by the following expression, derived from Eqs. (2)-(4):l'7

f = n(l-f)' '(KXt)""^ (5) This finding validates the occurrence of a two-state equilibrium, such as that of Eq. (1). Therefore, the longest lifetime component may be reasonably assigned to the monomeric species M, while the shortest one is assigned to the oligomeric species Mj . From the fitting, one obtaines n = 2.3 =t 0.3 and K

= (3.4 ± 0.2) 10^, n being a Hill-like parameter, because, according to Eq. (1), the aggregation equilibrium is cooperative, i.e., to a first approximation, intermediate aggregates other than the species M^ are minor. A narrow

distribution of aggregates is likely to reflect assemblies with similar size and geometry, fully compatible with a single lifetime. The parameter n thus represents the lowest limit of the number of peptide chains partecipating to the gathering process.

Time (ns)

Figure 2.4. Fluorescence decay curves of the reference (Fmoc-OH; a) and of FIO, at two concentrations: 1.9 \xM (b) and 11 |iM (c) in aqueous solution. The best fit of the reference is accomplished by a monoexponential function, while that of the peptide by a biexponential function.

= 265 ,K = 315 nm.


o

T3

bO

10 [F10](^M)

Figure 2.5. Pre-exponential factors of the short hfetime component of F10 decay in water, corresponding to the molar fraction of aggregated peptide (squares), as a function of [FIO]. The solid line represents the best fit to the experimental data, according to Eq. (5).

2.4. WATER-MEMBRANE PARTITION AND AGGREGATION

Preliminarly, we measured the change in fluorescence intensity caused by the addition of different amounts of liposome into peptide aqueous solutions. The results are reported in Figure 2.6, where the partiton curves show that, as peptide concentration increases, association with the phospholipid bilayer is progressively less favored, a finding suggestive of a complex partition equilibrium, very likely involving aggregated species, too.^^'^^

To answer the question as to whether the peptides are actually inserted into the membrane, and, in such a case, which species predominate, both collisional quenching and time-resolved fluorescence experiments were carried out. As far as the collisional measurements are concerned, the efficiency of water soluble quenchers was determined to assess the accessibility of membrane-bound fluorophores from the water phase.^^ Figure 2.7 illustrates the Stem-Volmer plots corresponding to the iodide quenching of 1.0 juM FIO and A3. The results clearly show that for both analogs the quenching caused by iodide ions is significantly reduced by the presence of liposomes, indicating that the peptides are inserted into the membrane. However, the probes still exhibit a partial accessibility to the quencher, azulene appearing somewhat more accessible than fluorene, probably because of its polarity, bringing it close to the polar headgroups for electrostatic effects.

54 B. PISPISA ETAL.

[lipid] (mM)

Figure 2.6. Water-membrane partition of FIO, as measured by the fluorescence intensity of the peptide (F), normalized by that in the absence of membrane (FQ). [FIO] = 1.1 ^M (a), 11 iM (b)

and 30 |iM (c).

Time-resolved fluorescence measurements, within the 5-10" - 65-10"^ range of the total peptide/lipid molar ratio explored (at two phospholipid concentrations: 0.2 and 2.0 mM), were performed under the condition that the membrane-bound peptide is in equilibrium with the free peptide in water. As expected, the fluorescence decay was complex, because of the contribution of the species in the bilayer, besides those present in water. A typical decay curve is shown in Figure 2.8, referring to FIO. After different trials, a good fit to the experimental data was obtained by a four exponential function, assuming that two species are present in the membrane phase, too. Besides the two lifetime components observed in pure water (Figure 2.5), that were taken fixed at 5.6 and 0.87 ns, the other two components were left free for the global fit of the fluorescence decay. As a resuU, one obtains x^ = 0.95 and a satisfactory residuals distribution.

The two decay components associated to the membrane-bound species were T'J = 7.0 ± 0.2 ns and T'2 = 2.2 ± 0.2 ns, both being significantly longer than in

water, as expected for the increase in fluorescence intensity caused by membrane binding. The relative population of each component is reported in Figure 2.9. From the results, it appears that the weight of the longest lifetime in the membrane decreases as peptide concentration increases, as one would predict for a membrane-bound monomer; consequently, the other component (T'2) is assigned to the oligomer. This hypothesis is supported by the finding that the aggregate is almost completely quenched as compared to the monomer, as already observed in water. A multi-exponential decay could be also due to other effects, such as a slow relaxation of membrane system, a different environment around the fluorophore, different states of the peptide,^^ and,


<

o

1.4

1.3

1.2

1.1

l l

-

-

ip7

/// y

^AS^

A/

^..-^^

n

y /

J

/

-

A

1 _.__P_

o

- 2

1 0 0.01 0.02 0.03 0.04 0.05

[KI] (M)

Figure 2.7. Stem-Volmer plots of collisional quenching by iodide ions for A3 (full symbols) and FIO (empty symbols). [A3] and [FIO] = 1 jj,M in water (triangles) and in the presence of liposomes (circles).

finally, a concentration-induced change in orientation of the peptide inside the bilayer.21 The observation that the fluorescence decay of membrane-bound peptide depends on peptide concentration rules out both conformational heterogeneity of the peptide and solvent/membrane relaxation phenomena. Therefore, besides the monomer-aggregate equilibrium, a concentration-induced change in orientation may contribute to the observed multi-exponential decay.

The following scheme summarizes the overall results.

Aqueous solution Membrane phase

M, M M' ^ = ^ M',

K K, K'

In this scheme, the equilibrium Mj -> M'j^. is not taken into account, because

the aggregate in the two phases exhibit a different size and, possibly, structure (see below), ascribable to the different environment. Therefore, Mj and M\,

species can not be related by a simple partition equilibrium. Furthermore, the prime refers to the species in the membrane, K and K' to the equilibrium

56 B. PISPISA ETAL.

o o

G ^ 5

• * -

"

5-

r \ s .

r

^ * ^

1 —

^ " ^ ^

10 Time (ns)

Time (ns)

Figure 2.8. Fluorescence decay curves of FIO (50 \M) in the presence of 2 mM ePC/cholesterol ( l : l ) ; ; i _ = 265, ?ie = 315 nm. The dashed line represents the fit to the experimental data by a

triple exponential function, with two lifetimes fixed at 5.6 and 0.87 ns, as those observed in pure

water, and the third one left free (x = 4.6). By contrast, the solid line represents the fit to the data by a four exponential function, in which two lifetimes were still maintained fixed, and the other two

were left free for the global fit of the fluorescence decay (x = 0.95). The residuals clearly show how the four exponendal fiinction (a) is definitely better than the triple one (b).

constant of aggregation in water and membrane, respectively, and IC to the equilibrium constant of the water-membrane partition, linked to the change in the overall standard free energy of transfer by the following expression:^^'^^

AGO,, = .RXinKp = A G 0 , „ „ + AGO,,^^^, (6)

In Eq (6), AG g iy is the change in the standard free energy of solvation of the peptide and membrane upon partitioning, and AG j n that arising from both perturbation of the peptide conformation caused by the bilayer, and perturbation of the lipid structure upon partitioning and binding processes.^^'^^" '^^ No coulombic term is present because the peptides examined have no


o

O

0.5-

.25-

0- 1

"v --

b

a

d

1 — € >

50 100 [F10](^iM)

Figure 2.9. Pre-exponential factors from fluorescence decay curves of FIO in 2 mM ePC/cholesterol (1:1), corresponding to the following lifetimes: 0.87 ns (a), 2.2 ns (b), 5.6 ns (c) and 7.0 ns (d); see text.

charged residues. Owing to the composite nature of both free energy changes, it is quite difficult to estimate AG^g j and AG jj yg . The sole quantity that can be confidently evaluated is the total standard free energy change of transfer, i.e. AG^ j. = -1.8 kcal -mol'^ as obtained from the value of IC reported below.

Interestingly, the value of AG^ j. is much smaller than that of, say, the transfer of hydrocarbons from water to an organic solvent. For instance, for n-C|QHJ2 at 25^C, AG jj ~ - 11 kcal-mol"^.^^ In our case, however, there are contributors of opposite sign that partially balance, giving rise to a relatively small value of AG^ j.. For example, contributors with positive sign are the

variation in conformational free energy and, possibly, in desolvation free energy of the peptide upon transferring from water to the membrane.

We next examined each equilibrium reported in the foregoing scheme, separately.^^ The results are shown in Figures 2.10 and 2.11. The main inference to be drawn from these Figures is 3-fold. Firstly, the water-membrane partition equilibrium is characterized by a constant value of the fraction of the

monomer associated to the membrane, within the 5-10"^ - 65*10" range of the total peptide/lipid molar ratio explored (data not shown). This implies the occurrence of a true equilibrium, whose constant is Kp =(1.3 ± 0.3)10^.

Secondly, the results referring to aggregation in water are similar to those obtained in pure water and already reported in Figure 2.5, all data being satisfactorily fitted by a single curve, as illustrated in Figure 2.10. Thirdly, the curve describing the aggregation equilibrium in membrane (Figure 2.11) shows

58 B. PISPISA ETAL.

that this process is much less favored and much more cooperative than that in water. The best fit of these latter data by eq (5) gives n' = 8.0 ± 3 and K' = 154 ± 8, implying that the average size of the aggregate in membrane is definitely larger than in water. This very likely reflects the peptide state of being ordered and constrained.

.o '•0 o

<

5 10"

Molar fraction of F10 in water

1 10"

Figure 2.10. Aggregation data of FIO in water (empty circles), as obtained by the time-resolved curves in the presence of liposomes (Figure 2.9). The results referring to the aggregation in pure water are also reported as full circles, for comparison. The solid line is the best fit to the overall data, according to Eq. (5).

o

bij

bO

<

0.02 0.04

Molar fraction of F10 in the membrane phase

Figure 2.11. Aggregation data of FIO in membrane, as obtained by the results reported in Figure 2.9. The solid line is the best fit, according to Eq. (5).


To get further evidence of peptide aggregation inside the membrane, fluorescence resonance energy transfer (FRET) experiments were carried out by mixing together FIO and A3 in 2 mM lipid solution, the probes in the analogs being able to act as a FRET donor-acceptor pair. The Forster radius for the fluorene-azulene pair, calculated according to the following expression,^'^^'^^ is RQ = 22 A.

Rn -ocn O 3

D-'

1/6

(7)

In Eq. (7), n is the refractive index of the medium, Oj) the quantum yield of the

donor, a a constant (8.785-10"^^ M-cm"^), and J the overlap integral of the normalized fluorescence spectrum of the donor and the absorption spectrum of the acceptor.

Owing to the strong distance dependence of FRET, the transfer efficiency is expected to vary significantly on going from a random distribution of fluorophores to the clustering caused by aggregation.^^'^^ This is indeed the case, as shown in Figure 2.12, where the energy transfer efficiency as a fimction of acceptor (A3) concentration is reported together with that theoretically expected for a random distribution of peptide in the bilayer.^^ The quenching efficiency was assessed by the decrease in the average fluorescence lifetime of

'o

o

&

9 12 [A3] (jiM)

Figure 2.12. Efficiency of intermolecular fluorescence resonance energy transfer between FIO and A3 analogs in membrane, as a function of the acceptor concentration (full symbols). [Lipid] = 2 mM, [FIO] = 2 jiM, X^^ = 265, X^^ = 315 nm. The quenching efficiency theoretically calculated

for a random distribution of peptides in membrane is also reported, for comparison (broken line). The 2 mM membrane concentration corresponds approximately to a complete binding of the peptide.

60 B. PISPISA ETAL.

the donor, to avoid inner-filter effects, caused by the acceptor absorption.^^ The large difference between the experimental and theoretical curves definitely indicate that gathering occurs in membrane, whose driving force can rely on different effects. It has been reported^^ for instance, that hydrophobic mismatch, i.e. the difference between the thickness of the hydrophobic core of the bilayer and the size of the transmembrane inclusion, may improve self-association. Moreover, specific intermolecular interactions are likely to arise from the GxxxG sequence motif, present in the primary structure of trichogin, where G stands for glycine and x for any aminoacid,^^ because the two glycine residues lie on the same face of the helical peptide, thus allowing a close, energetically favored intermolecular packing.

Finally, it is worth anticipating that a change in orientation of the peptide inside the membrane does take place, too, as the the membrane-bound peptide/lipid molar ratio (r) increases. The two phenomena are very likely concerted, in the sense that a transition to a transmembrane topology makes enough room for the peptide to aggregate.

2.5. BIOACTIVITY: MECHANISM OF MEMBRANE PERTURBATION

We next addressed the problem of the role played by the two membrane-bound species, M' and M'j , in the bioactivity. The peptide-induced membrane permeability was then investigated, searching for a relationship between the variation in the concentration of M' or M' . and peptide activity. Such relationship was indeed observed only when the membrane-bound M'j , species

is taken into account, a finding suggestive that only the aggregate is responsible for membrane perturbation, and hence bioactivity. This is shown in Figure 2.13, where the membrane-bound aggregate, as determined by the relative population of the 2.2 ns lifetime component in the fluorescence decay, and bioactivity, expressed in terms of the initial rate of carboxyfluorescein release from the membrane, are reported as a function of F10 concentration. In the same Figure, the membrane-bound monomer concentration is also reported, whose trend emphasizes the close relationship existing solely between aggregate formation and membrane permeability.

To gather further informations on the mechanism of bioactivity, both light scattering and release experiments were carried out. Where the liposome content is released by disruption of the bilayer through a detergent-like mechanism, a drastic decrease in particle size, and hence in light scattering, is expected to occur. ^ This is indeed the case when a detergent, such as Triton, is added to the membrane, because a 15 times decrease in the intensity of light scattering is observed for liposome micellization. By contrast, upon addition of FIO, A3 or TR-OMe at a concentration that causes an almost complete release of liposome content, the light scattering modestly varies (less than 10 %), implying that, in this case, the size of vesicles is not significantly perturbed. In addition, where the release of carboxyfluorescein and Texas-red labeled


dextran® is compared, the latter molecule is released in a much smaller quantity than the former one, at all peptide concentrations examined (data not shown) J Both these results rule out any mechanism of peptide-induced perturbation other than that involving peptide aggregates, which can be safely considered as the major contributor to the formation of transmembrane pores.

JT-' ' V5

1 t5 &

>

0.08

0.04

Q

-A

_ A

A

__A . •

U l L 0

A

D

•

A

n

• A •

D

1

10

•

D

A

1

n

A _

[FIO] (JAM)

2

1

^

Q CD o B' CO 3

22 ^

OQ ^

g- ^ ^ CD

*

2.13. Relationship between the concentration of FIO aggregate in membrane (empty squares; right scale) and peptide activity (full squares; left scale), as a function of [FIO]. Peptide activity is expressed in terms of the initial rate of carboxyfluorescein release from 0.2 mM ePC/cholesterol (1:1), after 20 min of peptide addition. The trend of the membrane-bound monomer concentration is also reported (triangles; right scale), showing the lack of any correlation with activity.

2.6. POSITION OF TRICHOGIN INTO THE MEMBRANE: TRANSLOCATION, DEPTH-DEPENDENT QUENCHING, AND DISTRIBUTION ANALYSIS

To tackle the problem of peptides translocation across the bilayer, FRET measurements, relying on the comparison of three different energy-transfer efficiencies rather than only two, as usually done,^ ' '* were carried out. The position of trichogin analogs into the bilayer was also investigated by means of depth-dependent quenching experiments,^^ and distribution analysis of membrane quenching.^^'^^

2.6.1. Peptide Translocation

Liposomes containing the NBD energy-transfer acceptor, and denoted as C6-NBD-PC(l-palmitoyl-2-[6-((7-nitrobenz-2-oxa-l,3-diazol-4- yl) amino)

-^ M.w. of carboxyfluorescein: 376; average m.w. of Texas-red labeled dextran: 10000. ' NBD = 7-nitrobenz-2-oxa-l,3-diazol-4-yl.

62 B. PISPISA ETAL.

caproyl]-L-a-phospha-tidylcholine), were used for FRET experiments. Three samples were prepared, differing for the position of NBD in the bilayer, i.e. i) a symmetrically labeled liposome (s-L), where the fluorescent lipid is inserted in both layers by mixing it to the other components before membrane formation; ii) an outer layer labeled liposome (o-L), where the fluorescent probe is incorporated only in the external region of the membrane, as obtained by adding it to a bilayer suspension after liposome formation; iii) an inner layer labeled liposome (i-L), as obtained by chemically quenching the label in the outer layer of s-L liposomes.^^

FRET depends on the inverse of the sixth power of the distance, the 50% quenching efficiency occurring at the interprobe distance RQ (Forster

radius).^'^^'^^ In the case of FIO, the fluorene-NBD pair has RQ = 24 A, while

the thickness of the bilayer is 42 A.^^ This ensures that a transbilayer FRET does not occur because of the small value of the Forster radius as compared to the thickness of the membrane. Since the fluorophore in C6-NBD-PC is located in the region of the polar headgroups,"* ' ^ quenching of the peptide lying in the outer layer or distributed in both leaflets of the bilayer is expected to be quite different. Upon translocation, one can predict that the quenching efficiency of i-L and o-L liposomes is approximately half that of the symmetrically labeled liposomes (s-L), where all peptides are surrounded by FRET acceptors. This because, on the average, half peptide molecules bound to i-L or o-L samples are in a layer void of acceptors. On the other hand, the relative increase in lipid fluorescence, as due to peptide binding, should be the same in all cases, because all labeled lipids are surrounded by peptide donors.

The aforementioned predictions were fully matched, as shown in Table 2.1, where the efficiency of relative fluorescence quenching of the donor [E = 1 -(F/FQ)], and the intensity increase of the acceptor (FVFQ'), are reported. The relative fluorescence quenching of the donor was obtained normalizing the emission intensity F in the presence of acceptor labeled liposomes (s-L, o-L or i-L) by the fluorescence of the unlabeled sample (FQ). By contrast, the intensity

increase in the acceptor emission caused by FRET was determined by measuring the lipid fluorescence in the presence (F') and absence (FQ') of the

fluorescent peptide. By inspection of the Table, it appears that i) the intensity increase of the

acceptor is practically the same for all lipidic samples used. This because, once translocation occurs, in each layer the acceptor encounters, on the average, the same number of donors; ii) the quenching efficiency of the donor nearly doubles on going from o-L and i-L to s-L, as expected. These data clearly indicate the occurrence of a complete translocation of the peptide across the membrane. It is worth noting that, where translocation had not occurred, the i-L liposomes would have caused negligible quenching of peptide fluorescence, because all peptides would be lain in the outer layer, while the effect of o-L and s-L liposomes would have been the same, because both these membranes have the same amount of acceptors in the outer layer. As a result, the increase in lipid fluorescence would have been quite different from that observed.


1

Q>

ys 2 0.8 tat)

1 0,6

8 0''

8 CO

1 0.2

E n

# / ^ \

a / ^ "" ' A L. / / i^

/ ' " ' \ - > i ' It-"-A'-- * / •' \ \

71 * \* * / JB V*A

• / . . ' \ 1 •; ,1 1 , 1 J

10 15 20

R'(A)

Figure 2.14. Fluorescence quenching profiles as a function of the distance from the bilayer center, R' (A), generated by the distribution analysis of the membrane-inserted FIO peptide, at 0.5 |iM (a), 3.5 |iM (b) and 10.8 ^M (c), in 2 mM labeled liposomes aqueous solution Q^^y^ = 290, X^^ = 315

nm). The distribution appears doubly peaked, as explicitly shown by the fitting of curve c, because a single gaussian is unable to fit all experimental data (see text). In all cases, the second peak is much closer to the center of the bilayer. A schematic representation of a lipid molecule is reported at the bottom of the Figure, in which the possible quencher positions along the stearoyl chain are indicated by a full circle.

Table 2.1. Results of translocation experiments of FIO inside the three labeled liposomes, expressed as quenching efficiency for the fluorene donor (E) and intensity increase (FVF ') for the NBD acceptor.

Liposomes

o-L i-L s-L

E = 1 - (F/Fo) (donor)

0.19 ±0.02 0.23 ± 0.03 0.34 ±0.03

FVFJ (acceptor)

2.0 ±0.2 1.9 ±0.2 1.9 ±0.2

a. All data were obtained using a peptide concentration of 0.5 |LIM

64 B.PISPISA^r^i:.

2.6.2. Depth-Dependent Quenching and Peptide Distribution Analysis

A detailed information on the position of a fluorophore within a membrane can be obtained by the method of depth-dependent quenching,^^ based on exploiting quenchers covalently bound to the phospholipid acyl chains. By varying the quencher position along the lipid acyl chain, the depth of a fluorophore within a membrane can be explored. The experiments were carried out with FIO at three concentrations, i.e. 0.5, 3.5 and 10.8 \xM, covering the whole activity range, corresponding to the release of membrane-entrapped carboxyfluorescein of 0%, 50% and 100%, respectively. The results are illustrated in Figure 2.14, where the profiles generated by the distribution analysis of the membrane-inserted peptide are reported as a function of the distance from the bilayer center, R' (A). The inferences to be drawn from the data of this Figure are the following, i) In all cases, the distribution appears doubly peaked, because a single gaussian is inadequate to fit all experimental data. The second peak, explicitly shown only in the case of curve c, for the sake of clarity, is much closer to the center of the bilayer; ii) the ratio of the area under the peaks was estimated to be approximately 11 (curve a), 7 (b) and 4 (c), indicating that the amount of the peptide deeply buried into the membrane increases as FIO concentration increases; iii) at 0.5 \xM, the highest decrease in fluorescence intensity is caused by the liposome having the quencher in position 7, which is near the polar headgroups; iv) at both 3.5 and 10.8 |iM (curves b and c in Figure 2.14), i.e. at concentrations high enough to determine liposome leakage, the relative quenching efficiency that increases significantly is that corresponding to the liposome with the deepest quencher, according to the ratio of the area under the peaks, which is almost halved or quartered, respectively. These results once againg indicate that the peptide becomes depeely buried in the bilayer upon increasing the membrane-bound peptide/lipid molar ratio, r.

An independent confirmation of the foregoing results was obtained by performing similar experiments, but with different labeled lipids. As r rises, the fluorophore is becoming, on the average, more accessible to the quencher positioned close to the center of the bilayer. This is shown in Figure 2.15, where the ratio of fluorescence intensities, measured in the presence of liposomes containing stearic acids labeled in position 5 or 16 of the acyl chain, is reported. As may be seen, the F(16)/F(5) ratio decreases with a non-linear trend as the membrane-bound peptide/lipid molar ratio increases, because of the concomitant variation of the two intensities, the first increasing and the other one decreasing. The same effect is observed with A3, despite the fact that in this peptide the fluorescent label is close to the N-terminus rather than to the C-terminus, as in FIO. This implies that the increase in the relative efficiency of the quencher located deeply in the bilayer is the same, irrespective of the direction taken by the peptide in going inside the membrane.


F(16)/F(5)

1

30 ^ 40 10 r

Figure 2.15. Relative quenching of F10 by liposomes containing stearic acid labeled on the 5* or 16* position by doxyl moiety, expressed as the ratio of fluorescence intensities of the peptide interacting with the two samples, as a function of the membrane-bound peptide/lipid molar ratio, r. [Lipid] = 2 mM; X^^ = 265, , XQYYI ~ 315 nm.

2.7. PEPTIDES ORIENTATION INSIDE THE MEMBRANE

The results reported in Section 2.6.2 prove that trichogin changes its position in the lipid bilayer as the membrane-bound peptide/lipid molar ratio rises. At low r values, the peptides lie close to the polar headgroups region, but as r increases, reaching a value corresponding to a peptide concentration able to cause membrane leakage, the quenching efficiency of the doxyl group positioned approximately in the middle of the bilayer increases significantly. Therefore, under these conditions, the peptide in the membrane experiences both a gathering (Figure 2.12) and a diving-like process (Figure 2.15).

The change in orientation is thus strictly related to aggregation, as further illustrated in Figure 2.16, where both the fluorescence quenching data of F10 in lipids labeled with a quencher on the 5** or 16* position and the fraction of membrane-bound aggregate are plotted as a function of the membrane-bound peptide/lipid molar ratio. The {I(5)-I(16)}/[F10] quantity is, in fact, linearly related to the fraction of deeply buried peptide, representing the difference between the fluorescence intensities of F10 in the lipids labeled with a quencher on the 5* or 16* position, as obtained from the data of Figure 2.15, normalized by the peptide concentration in the membrane. A similar trend was observed for A3 analog, too. From the results, it appears that there is a strict relationship between the two sets of data, suggesting that FIO, and hence trichogin, populates two states only, i.e. a monomeric, surface bound and inactive form, and a buried, aggregated state, responsible for membrane leakage. There is,

66 B. PISPISA ETAL.

0

o

Figure 2.16. Dependence of the fraction of membrane-bound FIO aggregate (full symbols) and of the same peptide deeply buried into the bilayer (empty symbols) on the membrane-bound peptide/lipid molar ratio, r. The {I(5)-I(16)}/[F10] quantity is linearly related to the fraction of deeply buried peptide, [FIO] representing the peptide concentration in the membrane (see text).

therefore, a threshold for the transition to a transmembrane orientation, and the interconversion between the two states is controlled by the membrane-bound peptide/lipid molar ratio.

A concentration-induced orientational transition, like that above described, was already reported for a few other antibiotic peptides,^^ as well as for the antifungal polyene nystatin.^^ This phenomenon can be explained by considering that a perturbation of the surface tension of the membrane arises from the binding of the peptide to the surface,^ and that at high peptide concentration excluded volume effects come also into play. " As a result, at fixed lipid concentration, a transmembrane orientation becomes thermodinamically favored as the amount of peptide increases.

The features of the transmembrane arrangement depend on the charge state of the peptide. For neutral or weakly charged peptides, such as alamethicin, a full insertion into the lipid bilayer is feasible, while for highly charged peptides, such as magainin, the insertion in the apolar region of the membrane is unfavored. In the latter case, an increase in peptide concentration leads to the formation of bilayer defects, the so-called "toroidal pores", which is another way in which a relaxation process for the accumulated surface tension is thought to occur.2 Accordingly, the foregoing results indicate that a high value of r forces the neutral lipopeptide trichogin to go deeply inside the bilayer.


An earlier study^^ - in which both a trichogin analog containing the quencher aminoacid TOAC and liposomes labeled with phosphatidylcholine analogs, bearing the fluorophore BODIPY at different positions along the acyl chain, were used - apparently contradicts this conclusion. Because the fluorescence quenching was found to be independent of the position of TOAC in the peptide and of the peptide/lipid ratio, the idea was that trichogin was lying along the membrane surface, at all concentrations examined. However, it has been conclusively shown" ^ that the BODIPY group attached to phospholipids exhibits a clear tendency to reside in the polar headgroups region of the bilayer, irrespective of its position along the acyl chain. Therefore, we are highly inclined to think that, using this fluorophore, in no case a transition in peptide orientation could be detected. This idea is supported by the lack of any clear dependence of fluorescence quenching on BODIPY position along the lipid acyl chain (Figure 2.2 of ref 45), indicating that the fluorescent lipid analog employed makes it impossible to determine the actual peptide position inside the membrane. By contrast, the data presented here have not such limitation,^^ because the use of doxyl labeled lipids and fatty acids is a very well established method for determining the membrane position of fluorescent probes.^^'^^ It must be also mentioned that some EPR results'*^ support the hypothesis put forward in ref. 45. However, the 0.1 mM peptide concentration used in this study is definitely higher than that normally needed for antimicrobial and membrane perturbing activity, and too high to make the EPR data and those reported here worthy of comparison, also in view of the complex interplay between the strongly concentration-dependent aggregation and membrane-peptide partition phenomena.'*^

2.8. CONCLUDING REMARKS

From the data set considered here, four major conclusions can be drawn. Firstly, the analogs of trichogin GA IV investigated exhibit structural features and bioactivity very similar to that of the natural peptide, and undergo a monomer-aggregate equilibrium both in water and model membranes. The aggregates in the two phases differ, however, in size and, possibly, structure. Secondly, fluorescence quenching measurements, carried out using water soluble quenchers and quenchers positioned in the membrane at different depths, indicate that at low membrane-bound peptide/lipid molar ratio (r) trichogin lies close to the region of polar headgroups, while, as r increases until membrane leakage occurs, a cooperative transition takes place, leading to an arrangement that sees the peptide deeply buried into the bilayer. Thirdly, the transitions from a surface to a transmembrane topology and from monomers to oligomers are very likely concerted, in the sense that a transmemrane arrangement makes enough room for the peptide to aggregate. Fourthly, Forster

TOAC = 4-amino-4 carboxy-2,2,6,6-tetramethylpiperidino-l-oxyl; BODIPY = 4,4-difluoro-4-bora-3a,4a- diaza-S-indacene).

68 B. PISPIS A ETAL.

energy transfer measurements indicate that, on the average, trichogin is equally distributed between the outer and inner leaflet of the membrane.

The mechanism of trichogin action can be then envisaged as a two-state transition controlled by peptide concentration. One state is the monomeric, surface bound and inactive peptide, and the other state is a buried, aggregated form, which is responsible for membrane leakage and bioactivity. Since trichogin suffers of hydrophobic mismatch, a complex supramolecular structure is likely to form when the peptide is buried into the bilayer, "^ an hypothesis which is under investigation.

2.9. ACKNOWLEDGMENTS

We thank Prof. C. Toniolo for kindly providing the trichogin analogs, and the Ministry of University and Research for financial support.

2.10. REFERENCES

1. N. Sewald, and H.-D. Jakubke, Peptides .Chemistry and Biology (Wiley-Vch, Weinheim, 2002). 2. Y. Shai, Mode of action of membrane active antimicrobial peptides, Biopolymers 66, 236-248

(2002). 3. R. M. Epand, and H. J. Vogei, Diversity of antimicrobial peptides and their mechanisms of

action, Biochim. Biophys. Acta 1462, 11-28 (1999). 4. K. Matsuzaki, Molecular mechanisms of membrane perturbation by antimicrobial peptides, in

Development of Novel Antimicrobial Agents: Emerging Strategies, edited by K. Lohner (Horizon Scientific Press, Wymondham, 2002), pp. 167-181.

5. L. Yang, T. A. Harroun, T. M. Weiss, L. Ding, and H. W. Huang, Barrel stave or toroidal model? A case study on melittin pores, Biophys. J. 81, 1475-1485 (2001).

6. F. Y. Chen., M. T. Lee, and H. W. Huang, Sigmoidal concentration dependence of antimicrobial peptide activities: a case study on alamethicin, Biophys. J. 82, 908-914 (2002).

7. C. Toniolo, M. Crisma, F. Formaggio, C. Peggion, R, F. Epand, and R. M. Epand, Lipopeptaibols, a novel family of membrane active, antimicrobial peptides, Cell. Mol Life Sci. 58, 1179-1188(2001).

8. B. Pispisa, L. Stella, M. Venanzi, A. Palleschi, F. Marchiori, A. Polese, A., and C. Toniolo, A spectroscopic and molecular mechanics investigation on a series of Aib-based linear peptides and a peptide template, both containing tryptophan and a nitroxide derivatives as probes, Biopolymers 53, 169-181 (2000).

9. B. Pispisa, L. Stella, M. Venanzi, A. Palleschi, C. Viappiani, A. Polese, F. Formaggio, and C. Toniolo, Quenching mechanisms in bichromophoric, 3io-helical Aib-based peptides, modulated by chain-length dependent topologies, Macromolecules 33, 906-915 (2000).

10. P. Scrimin, P. Tecilla, U. Tonellato, A. Veronese, M. Crisma, F. Formaggio, and C. Toniolo, Zinc(II) as an allosteric regulator of liposomal membrane permeability induced by synthetic template-assembled tripodal polypeptides., Chem. Eur. J. 8, 2753-2763 (2002).

11. A. D. Milov, Y. D. Tsvetkov, F. Formaggio, M. Crisma, C. Toniolo, and J. Raap, Self-assembling and membrane modifying properties of a lipopeptaibol studied by CW-ESR and PELDOR spectroscopies, J. Pept. Sci. 9, 690-700 (2003).

12. B. Pispisa, A. Palleschi, C. Mazzuca, L. Stella, A. Valeri, M. Venanzi, F. Formaggio, C. Toniolo, and Q. B. Broxterman, The versatility of combining FRET measurements and molecular mechanics results for determining the structural features of ordered peptides in solution, J. Fluoresc. 12, 213-217 (2002).

13. C. Toniolo, A. Polese, F. Formaggio, M. Crisma, and J. Kamphuis, Circular dichroism spectrum of a peptide 3io-helix, J. Am. Chem. Soc. 118, 2744-2745 (1996).


14. B. Pispisa, C. Mazzuca, A. Palleschi, L. Stella, M. Venanzi, F. Formaggio, C. Toniolo, and Q. B. Broxterman, Structural features and conformational equilibria of 3io-helical peptides in solution by spectroscopic and molecular mechanics studies, Biopolymers (Biospectroscopy) 67, 247-250 (2002).

15. W. C. Wimley, and S. H. White, Experimentally determined hydrophobicity scale for proteins at membrane interfaces, Nat. Struct. Biol 3, 842-848 (1996).

16. S. H. White, and W. C. Wimley, Hydrophobic interactions of peptides with membrane interfaces, Biochim. Biophys. Acta 1376, 339-352 (1998).

17. L. Stella, C. Mazzuca, M. Venanzi, A. Palleschi, M. Didone, F. Formaggio, C. Toniolo, and B. Pispisa, Aggregation and water-membrane partition as major determinants of the activity of the antibiotic peptide trichogin GA IV, Biophys. J. 86, 936-945 (2004).

18. V. Rizzo, S. Stankowski, and G. Schwarz, Alamethicin incorporation in lipid bilayers: a thermodynamic study. Biochemistry 26, 2751-2759 (1987).

19. M. Castanho, and M. Prieto, Filipin fluorescence quenching by spin-labeled probes: studies in acqueous solution and in a membrane model system, Biophys. J. 69, 155-168 (1995).

20. A. S. Ladokhin, and S. H. White, Alphas and taus of tryptophan fluorescence in membranes, Biophys. J. 81, 1825-1827 (2001).

21. H. W. Huang, Action of antimicrobial peptides: two-state model. Biochemistry 39, 8347-8352 (2000).

22. S. H. White, and W. C. Wimley, Membrane protein folding and stability: physical principles, Amu Rev. Biophys. Biomol. Struct. 28, 319-365 (1999).

23. T. J. Mcintosh, A. Vidal, and S. A. Simon, The energetics of peptide-lipid interactions: modulation by interfacial dipoles and Cholesterol, in Peptide-Lipid Interactions, edited by T. J. Mcintosh, and S. A. Simon (Academic Press, San Diego 2002), pp. 309-338.

24. C. J. Russell, T. E. Thogeirsson, and Y. -K. Shin, The membrane affinities of the aliphatic amino acid side chains in an alpha-helical context are independent of membrane immersion depth. Biochemistry 38, 337-346 (1999).

25. C. Tanford, The Hydrophobic Effect: Formation of Micelles and Biological Membranes, 2nd ed. (Wiley, New York, 1980).

26. B. Pispisa, M. Venanzi, L. Stella, A. Palleschi, A., and G. Zanotti, Photophysical and structural features of covalently bound peptide-protoporphyrin-peptide compounds carrying naphtalene chromophores, J. Phys. Chem. B 38, 8172-8179 (1999).

27. L. Stella, M. Venanzi, M. Carafa, E. Maccaroni, M. E. Straccamore, G. Zanotti, A. Palleschi, and B. Pispisa, Structural features of model glycopeptides in solution and in membrane phase: a spectroscopic and molecular mechanics investigation, Biopolymers 64,44-56 (2002)

28. M. Schumann, M., Dathe, M., Wieprecht, T., Beyermann, M., and M. Bienert, The tendency of magainin to associate upon binding to phospholipid bilayers. Biochemistry 36, 4345-4351 (1997).

29. J. Strahilevitz, A. Mor, P. Nicolas, and Y. Shai, Spectrum of antimicrobial activity and assembly of dermaseptin-b and its precursor form in phospholipid membranes. Biochemistry 33, 10951-10960(1994).

30. B. Pispisa, C. Mazzuca, A. Palleschi, L. Stella, M. Venanzi, M. Wakselman, J. -P. Mazaleyrat, M. Rainaldi, F. Formaggio, and C. Toniolo, A combined spectroscopic and theoretical study of a series of conformational^ restricted hexapeptides carrying a rigid binaphthyl-nitroxide donor-acceptor pair, Chem. Eur. J. 9, 2-11 (2003).

31. G. R. Jones, and A. R. Cossins, Physical methods of study, in Liposomes: a Practical Approach, edited by R. R. C. New (Oxford University Press, Oxford., 1990) pp. 183-220.

32. A. R. Curran, and D. M. Engelman, Sequence motifs, polar interactions and conformational changes in helical membrane proteins, Curr Op. Struct. Biol. 13, 412-417 (2003).

33. K. Matsuzaki, O. Murase, N. Fujii, and K. Miyajima, Translocation of a channel forming antimicrobial peptide, magainin 2, across lipid bilayers by forming a pore, Biochemistry 34, 6521-6526(1995).

34. W. C. Wimley, and S. H. White, Determining the membrane topology of peptides by fluorescence quenching. Biochemistry 39, 161-170 (2000).

35. E. London, and A. S. Ladokhin, Measuring the depth of amino acid residues in membrane-inserted peptides by fluorescence quenching, in Peptide-Lipid Interactions, edited by T.J. Mcintosh, and S.A Simon (Academic Press, San Diego 2002) pp. 89-115.

70 B.PISPISA^r^L.

36. A. S. Ladokhin, Distribution analysis of depth-dependent fluorescence quenching in membranes: a practical guide, Meth. Enzymol. 278, 462-473 (1997).

37. A. S. Ladokhin, Analysis of protein and peptide penetration into membranes by depth-dependent fluorescence quenching: theoretical considerations, Biophys. J. 76, 946-955 (1999).

38. J.C. Mclntyre, and R.G. Sleight, Fluorescence assay for phospholipid membrane asimmetry, Biochemistry 30, 11819-11827 (1991).

39. L. J. Lis, M. McAlister, N. L. Fuller, R. P. Rand, and V. A. Parsegian, Measurement of the lateral compressibility of several phospholipid bilayers, Biophys. J. 37, 667-672 (1982).

40. S. Mazeres, V. Schram, J. F. Tocanne, and A. Lopez, 7-Nitrobenz-2-oxa-l,3-diazole-4-yl-labeled phospholipids in lipid membranes: differences in fluorescence behavior, Biophys. J. 71, 327-335 (1996).

41. D. E. Wolf, A. P. Winiski, A. E. Ting, K. M. Bocian, and R. E. Pagano, Determination of the transbilayer distribution of fluorescent lipid analogues by nonradiative fluorescence resonance energy transfer. Biochemistry 31, 2865-2873 (1992).

42. M. Lee, F. Chen, and H. W. Huang, Energetics of pore formation induced by membrane active peptides. Biochemistry 43, 3590-3599 (2004).

43. A. Coutinho, and M. Prieto, Cooperative partition model of nystatin interaction with phospholipid vesicles, Biophys. J. 84, 3061-3078 (2003).

44. M. Zuckermann, and T. Heimburg, Insertion and pore formation driven by adsorption of proteins onto lipid bilayer membrane-water interfaces, Biophys. J. 81, 2458-2472 (2001).

45. R. F. Epand, R. M. Epand, V. Monaco, S. Stoia, F. Formaggio, M. Crisma, and C. Toniolo, The antimicrobial peptide trichogin and its interaction with phospholipid membranes, Eur. J. Biochem. 266, 1021-1028 (1999).

46. R. D. Kaiser, and E. London, Determination of the depth of BODIPY probes in model membranes by parallax analysis of fluorescence quenching, Biochim. Biophys. Acta 1375, 13-22(1998).

47. C. Mazzuca, L. Stella, M. Venanzi, F. Formaggio, C. Toniolo, and B. Pispisa, Mechanism of membrane activity of the antibiotic trichogin GA IV: a two-state transition controlled by peptide concentration, Biophys. J. 88, 3411-3421 (2005).

48. V. Monaco, F. Formaggio, M. Crisma, C. Toniolo, P. Hanson, and G. L. Millhauser, Orientation and immersion depth of a helical lipopeptaibol in membranes using TO AC as an ESR probe, Biopolymers 50,239-253(1999).

49. M. R. R. de Planque, and J. A. Killian, Protein-lipid interactions studied with designed transmembrane peptides: role of hydrophobic matching and interfacial anchoring, Mol Membr Biol. 20, 271-284 (2003).

THEORY OF METAL-FLUOROPHORE INTERACTIONS

Nils Calander

This chapter is mainly a review of research done by the author, concerning theory of surface plasmon resonance interaction with fluorophores. Surface plasmon coupled emission (SPCE) is studied theoretically and compared to experiment. Surface plasmon resonance optical field enhancement is investigated at elongated particles by solving the Maxwell's equations with the use of spheroidal vector wave functions. Finally, the theoretical possibility of trapping fluorophores by optical gradient forces at surface plasmon enhanced hot spots is examined.

3.1. INTRODUCTION

Surface plasmon resonance (SPR) is a phenomenon that may occur at metal interfaces. It is also the name of a method that is used to investigate biomolecular binding to surfaces "'*. It relies on the absorption of optical energy by surface plasmons and is very sensitive to the optical properties near the surface.

The opposite to optical absorption by excitation of surface plasmons is that surface plasmon energy couple out into an optical field. This is called surface plasmon coupled emission (SPCE) and has recently " been shown to have a potential for enhanced detectability of fluorophores near a planar surface. The radiation is concentrated at specific angles and polarizations making it possible to collect a larger proportion of the emitted photons.

The Raman cross-section can be greatly enhanced by surface plasmons. This is called surface enhanced Raman scattering (SERS). The method has been refined and found uses in the life sciences. It has been found that the Raman signal can be increased several orders of magnitude ^"" by surface plasmon

Physics Department, Chalmers University of Technology, SE-412 96 Goteborg, Sweden e-mail: [email protected]

71

72 N. CALANDER

resonance at colloidal metal particles, preferentially silver or gold. This has made Raman spectroscopy of single molecules possible. It has been suggested that the strong optical field gradient may attract molecules to the hot spots by gradient forces ^ , in a way enhancing the Raman spectroscopy.

The opposite of the optical field enhancement at hot spots is that the radiation out from the hot spot is also enhanced, i.e. radiative enhancement. An excited fluorophore nearby couple out its energy more efficiently by the surface plasmons, an antenna or transformer action, resulting in radiative decay enhancement. This technique is sometimes called radiative decay engineering (RDE) ' ' -2 ^ which has a number of advantages in fluorescence spectroscopy. The quantum yield of a fluorophore is greatly enhanced. The fluorescence signal is also enhanced, as is the time before bleaching of the fluorophore. These advantages may make weak intrinsic fluorescence of biomolecules come into play and tagging by fluorescent labels may be unnecessary. Also, the radiative decay enhancement may make the optical trapping of fluorophores possible.

3.2. SURFACE PLASMON RESONANCE

3.2.1. Plasma Oscillations

Interacting charged particles may show plasma oscillations where the restoring force is electrical, as illustrated in Figure 3.1. If the free electrons in the metal are displaced from the lattice of positively charged ions an electrical restoring force tries to restore neutrality. Since the electrons are rather free an oscillating behavior results at the so called plasma frequency.

b)

* l » »*l» ' • I * 4*4* i*fe« »4» ,-*^ «|» tA* »|« *i|« '%" -*|» *il«' •^

+ 4 *4^ -4 '4 * * f4 ' ' 4 *4 ' ' f 4 *4 ' * f ' - t *4 *

uLt JL n^. Ju, iiL, JL :ifL, JL «JL J ^ ^ JL a|« ol'' «J«

^ '^J'J'J'J^J'J^J^J'J^J^J'J^JC^

-

Mm*

Figure 3.1. a): A weight hanging in a spring. Oscillations are due to the inertia of the weight and restoring forces in the spring, b): Assuming the electrons moving as a whole in the metal compared to the fixed positive background of ions, displacements of the electrons cause electrical restoring forces turning the metal back to the homogenously neutral state. These restoring forces make the electrons oscillate, i.e. plasma oscillations.

THEORY OF METAL-FLUOROPHORE INTERACTIONS 73

3.2.2. Surface Plasmon Resonances

The plasma oscillations near a surface may interact with the electromagnetic field from the charged particles in such a way that the plasma oscillations are localized to the surface. These oscillations are called surface plasmons and are illustrated in Figure 3.2.

Phenomenologically, the metal can at optical frequencies be seen as having a permittivity of negative real part. The imaginary part is due to losses. In the quantum theory language the plasma oscillations are quantized in the energy quanta hcop where cOp is the oscillation frequency, and are called plasmons. The surface plasma oscillations are quantized in units of surface plasmons. The quantized theory is not needed for explaining the surface plasmon resonance effects in this chapter, and is therefore not discussed further. The theoretical background of the surface plasma oscillations in this chapter only relies on the Maxwell's equations and the macroscopic behavior of the metal in terms of permittivity, believed to be valid down to feature sizes of below a nanometer. The Maxwell's equations are solved separately in the metallic and non-metallic domains and joined together by appropriate boundary conditions. The boundary conditions are that the normal component of the electric field times the permittivity and the in-plane component of the electric field are continuous at the interface. The magnetic field is also continuous. The permittivity versus wavelength is shown in Figure 3.3 for silver and gold.

Figure 3.2. Surface plasma oscillations at a plane and spherical interface between two media, one of positive and one of negative permittivity. The plasma oscillation is a solution to the Maxwell's equations with the appropriate boundary conditions. Two evanescent plane waves in the two media are matched at the interface. The waves are traveling or standing in the plane and exponentially decaying away from the interface. The electromagnetic field is accompanied by oscillating surface charges at the interface. A condition for surface plasma oscillations at the plane interface is that £1+82 < 0 .

74 N. CALANDER

Silver Gold

0

-20

-40

-60

on

r^ ii^^jiiiiiiii'

i

• Ima^inbry iHtmi, . } . . . • . . . .7

i RealV^

i i '- 1 0.2 0.4 0.6 0.8 1

X{\irr\) 1.2 0.2 0.4 0.6 0.8 1 1.2

A,{^m)

Figure 3.3. The permittivity (dielectric function) of silver and gold. The real parts are negative due to plasma oscillations. The imaginary part is due to electric losses in the metal. Data from Reference

3.3. THEORY OF SURFACE PLASMON RESONANCE AT PLANAR STRUCTURES

Surface plasmon resonance at planar structures is studied by solving the Maxwell's equations using Fresnel theory. The electromagnetic fields from oscillating dipoles are expressed as integrals over plane waves in order to fit into the Fresnel scheme. In this way the interaction of fluorophores with planar structures possessing surface plasmons is studied. In particular surface plasmon coupled emission is theoretically investigated. This section is mainly a review of Reference .

Fluorescence spectroscopy is an important method for high sensitivity detection and analysis in many fields of biotechnology. One way of improving the sensitivity and selectivity is by resonance enhancement and modification of the fluorescence signal. In recent publications ' ' ^ experimental studies of surface plasmon-coupled directional emission (SPCE) are reported.

Surface plasmons in a thin metal layer, excited by fluorophores, can radiate into a glass prism at sharply defined angles determined by the emission wavelength and the optical properties of the respective layers and the glass. SPCE is related to surface plasmon resonance (SPR) \ a well-known and used phenomenon in which absorption of light takes place at specific angles of incidence, which are sensitive to the optical properties near the metal surface. SPCE is in a sense the opposite; emission is detected rather than absorption. The directional property of SPCE can be used to observe the emission selectively. Coupling to the surface plasmons and the collection of the emission can be made much more efficient than with standard methods. Only fluorophores close to the surface contribute to the SPCE, which means that


background emission is highly rejected. These properties of SPCE will surely find numerous applications in biotechnology and chemistry.

One aim of this chapter is to show that a theoretical / simulation approach to SPCE is a useful tool in order to design experimental setups and structures, fine-tune measurements, predict results, explain experimental findings, and to show promise for further refinement of SPCE. Quantities simulated are SPCE angles, power levels, decay enhancements and decay times. In this chapter a theoretical simulation approach for SPCE is compared to experimental results. Theoretical explanations of some experimental findings thus become evident from these simulations.

3.3.1. Basic Theory

The radiation from a dipole can be decomposed into an integral over plane waves, see Figure 3.4. This is a 2-dimensional Fourier transform. The decomposition is also called the Weyl identity ^^. The electromagnetic field is also divided into a p- and an s-polarized part. Fresnel theory ^ ' , i.e. the theory of refraction of plane waves in a dielectric planar structure, can then be applied, see Figure 3.5. Proper matching of the plane waves has to be done at the boundaries and also at the dipole. A pre-integration (pre-summation) in azimuth (cp) is done by using Bessel integrals, see Reference ^ , which means that only one-dimensional numerical integration (in 0) has to be done in order to calculate the electromagnetic fields at any point. The method can also be explained in terms of the Sommerfeld identity ^ . No integration is needed for calculation of the far-fields. Parseval's equation is used for the power flow.

iillliilililsli;-;""' ''"••"Jlilillliillliii

Figure 3.4. The electromagnetic field from an oscillating dipole is fit into the Fresnel's scheme by decomposing it into an integral (or sum) of plane waves in the two main directions of the planar structure. The evanescent waves, i.e. non-propagating plane waves or plane waves with imaginary normal components of the wave-vector have to be included. When incorporated into the layered structure (Figure 3.5), appropriate matching of the electromagnetic fields at the surface indicated by the straight vertical line above, has to be done, with the assumption of including an electromagnetic

76 N. CALANDER

k,

K

2

k2

K

; 3

1 k3

{ 14 \

W§m N

\ k

Figure 3.5. The Maxwell's equations for the electromagnetic fields are solved by using Fresnel's theory of plane wave propagation in a planar dielectric structure. Two plane waves with mirrored wave-vectors are assumed to propagate in each homogenous layer and boundary-matched to the plane waves in the neighboring layers. The in-plane components of all couples of plane wave's wave-vectors in all layers are the same (Snell's refraction law). Appropriate matching of the electromagnetic fields has to be done at all boundaries. When including an oscillating dipole in one of the layers as a radiating source, appropriate matching of the electromagnetic fields also has to be done at the dipole.

Here some equations used in the simulations are given. The quantities in the equations are defined in

Table 3.1. The integrals for the electromagnetic fields at an arbitrary position in the layered structure, from a dipole normal to the plane are:

1 3 ^ Y\^

H = - ^ c p , j d n ^ _ p ^ j , (k„n^p)(ae' «"' +be • »"' ) (1) 47C 0 -y/nd-np

V"<|-"p (2) 4nnn d 0

-zn^J„(kon^p)(ae^'<«"'^+be-»'«"'^)]

The power per solid angle, normalized to the total output power from the fluorophore in bulk medium, for the far-fields from a dipole normal to the plane is:

P = 3n^sin(0) cos(0) . .2

87cn, n^-n'sin(e)' (3)


For a dipole in parallel to the plane the p-polarized part of the normalized power per solid angle is:

3n^cos(e)^, ,2 . .2

87inj (4)

The s-polarized part is:

3n'cos(e)'

87in, nf -n ' s in (e ) [a| sin((p) (5)

The power flow in the layered structure is for a dipole normal to the plane given by:

P 3 iK#^(-^)K-^*) K-"p (6)

For a dipole in parallel to the plane, it is given by:

1 P 3

P, 8 jdnpUp

, n?n n, n? -n^, , V n ' - n f ( a - b ) ( a * + b*) (7)

The power flow is normalized to the total radiated power from a fluorophore in a continuous bulk medium similar to the medium of the layer in which the fluorophore radiates. The free vacuum space is not chosen for the normalization because there seems to be an uncertainty about the behavior of the radiation decay enhancement in a dielectric medium, see Reference ^ . By classical electromagnetic theory the radiated power of a dipole in a homogenous dielectric medium should be proportional to the refractive index, but it is very sensitive to the local environment. In the empty-cavity model the dipole is considered to be inside an empty spherical cavity in the dielectric of refractive index n, giving an enhancement, compared to vacuum, by:

3n^

2 n ' + l (8)

This model is compared to other models in Reference ^ . Reference ^ refers to a model that predicts a square dependence on the refractive index, which for refractive indices between one and two compares well with the empty cavity model.

78 N. CALANDER

Table 3.1. Descriptions or definitions of the quantities occurring in the equations.

1 Notation ko

1 ^ 1 Hd

k kd

1 np 1 n 1 Co 1 to 1 ^

b

1 ^ 1 * 1 ^

Jm

--

1 x,y,z

P

1 ^ 1 ^ Lid

Description or definition | Wave-vector in vacuum Refractive index 1 Refractive index of layer containing dipole 1 nko 1 Hdko 1 ko=kono, where ko is in-plane component of wave-vector

Speed of light in vacuum 1 Vacuum permeability 1 Electric field coefficient for the forward electromagnetic component in the layered 1 structure Electric field coefficient for the backward electromagnetic component 1 Radiating dipole moment 1 Unit vector in azimuth 1

Unit vector in direction of incidence 1

Bessel function of order m I Magnetic field I Electric field 1 Cartesian coordinates 1

(xW)'" Normalized power per steradian Power flow in z-direction 1 Total power from dipole in bulk medium |

3.3.2. Simulations

The experiments in References " ' ^ are simulated using the theory above. Figure 3.6 illustrates the basic setup. A layered structure is built on a hemispherical prism. At the planar surface of the prism a glass substrate is glued by index-matching glue. A silver layer of 50 nm thickness is deposited on the glass. The silver is protected by a 5 nm thick layer of Si02, which also serves as a spacer. A layer of polyvinyl-alcohol (PVA) is spin-coated on top of the Si02. The PVA layer contains fluorophores, i.e. sulforhodamine 101, rhodamine 123 or Py2. The thickness used in the first simulations is either 15 or 30 nm. Excitation light at 514 nm wavelength, from an argon ion laser, is either coming from the left (front side. Reverse Kretschmann configuration, RK) at normal incidence, exciting the fluorophores in the PVA layer directly, or from the right (back side, Kretschmann configuration, KR) at the surface plasmon angle, exciting the fluorophores in the PVA layer via the SPR evanescent wave. Fluorescence light near 600 nm wavelength is detected at specific angles at the back side, and more diffuse at the front side. The fluorescence light at specific angles at the back side is the so called SPCE, and is the main topic of these simulations. It is assumed to originate from surface plasmon resonances ^ in the silver layer, excited by the nearfield light from the fluorophore.


The calculated far field is illustrated in Figure 3.7, and its azimuthal dependence in Figure 3.8. The light field from the fluorophore in the layered structure is calculated and illustrated in Figure 3.9. Figures 3.10 - 3.12 show different aspects of radiated and dissipated power, and decay times, to illustrate the theory compared to the experiments. The refractive indices are assumed to be 1.5 for the glass prism, the Si02 and the PVA layers, and are taken from Reference . The complex refractive index for the silver layer is assumed to be 0.1243+3.73161, and is taken from Reference ^l

Front side

PVA layer coiitaiiiingNv I fluorophores^

Back side

SPCE radiation

Fhiorophon

i ^

Tlie layers

5 nm SiOa

: / \ \ ' i ,<=? * -< - -T .V-' -* * J

50 nm

O

I m

Figure 3.6. The basic configuration. Excitation light is either coming from the left (front side, A) at normal incidence, exciting the fluorophores in the PVA layer directly, or from the right (back side, B) at the surface plasmon angle, exciting the fluorophores in the PVA layer via the SPR evanescent wave.

Figure 3.7. Left: Normalized radiated power per steradian from a dipole oriented normal to the plane. A dipole in this orientation only radiates p-polarized light. The normalization is to the output from a radiating dipole in bulk PVA. Right: Maximum normalized radiated power per steradian from a dipole oriented parallel to the plane. The solid line shows the p-polarized part which is coupled out by SPCE. The dotted line shows the s-polarized part. The angular distribution in azimuth is illustrated in Figure 3.8.

80 N. CALANDER

Figure 3.8. The SPCE light intensity as it should be seen projected on a screen at the backside. The direct laser beam should hit a spot at the center of the circle. The SPCE light varies in azimuth, i.e. the angle cp in Figure 3.6, depending on the dipole orientation. For a dipole oriented normal to the layered structure, the SPCE light is equally distributed in the light cone, illustrated by the left circle. The output from a dipole oriented in parallel to the layered structure is illustrated in the middle. Only the p-polarized part of the light contributes to the SPCE in that case, and the light is not equally distributed in the circle. In a real situation the dipoles are more or less randomly oriented, due to the distribution of fluorophore orientations and the exciting light. In this case the non-equal distribution of light is less pronounced. This is illustrated by the right circle.

-The layers The layers

-20 -10

Figure 3.9. Left: Electromagnetic energy density of the radiation from a dipole oriented normal to the plane. Right: A magnification that shows an instant of the absolute value of the electric field. The dipole has position (0,0). The position of the layers, i.e. the silver, Si02, and PVA layers, is indicated in the figure. The thickness of the PVA layer is 15 nm and of the SiOi layer 5 nm, as in Reference .


The farfield power density is simplest calculated by Equations (3), (4), and (5). It is illustrated in Figure 3.7 for a fluorophore in the middle of the 15 nm thick PVA layer, with dipole oriented either normal or parallel to the layered structure. It is seen to be highly directed at the back side, at angles of 46.8°, an almost perfect match to the 47° reported in References ^''^. For the 30 nm thick PVA layer, the SPCE direction is 50.4°, also an almost perfect match to the 50° reported in Reference . The SPCE angle is nearly independent of the fluorophore position and orientation within the PVA layer. The ftiU widths at half maximum (FWHM) are 1.1° for the 15 nm thickness and 1.8° for the 30 nm thickness of the PVA layer, for all orientations and positions of the fluorophores.

The azimuthal distribution around the SPCE cone is illustrated in Figure 3.8, and can be compared to Figure 3.6 in Reference . It depends on the dipole orientation. Only p-polarized light contributes here to the SPCE. A dipole oriented normal to the layered structure only radiates p-polarized light, and does it equally distributed in azimuth. This is illustrated in the left part of the figure. A dipole oriented in parallel to the layered structure radiates a combination of p-and s-polarized light. Since SPCE only has p-polarization this means that there is almost no SPCE radiation perpendicular to the dipole, illustrated in the middle part. In reality there is a more or less random distribution of dipoles, depending on the distribution of fluorophore orientation and the exciting light. For 90° excitation light from the front side, only the corresponding parallel component of the fluorophore dipole contributes to excitation, whereas for backside SPR light, both the normal and a parallel component are contributing. It is not reported concerning the azimuthal intensity variations in References " .

The light field is calculated by Equations (1) and (2), which hold inside the layered structure, outside for the nearfield as well as for the farfield. The light from a fluorophore in the middle of the 15 nm thick PVA layer is illustrated in Figure 3.9. Its radiating dipole moment is normal to the layered structure. The SPCE light at the back side is seen to really be highly directed in the glass. At the front side it is seen to be more diffuse.

The power flow inside the layered structure, and outside, is calculated by Equations (6) and (7). Figure 3.10 shows the power flow in the case of the 15 nm thick PVA layer, from a dipole oriented normal or parallel to the layered structure. Most of the power drawn from the dipole by the surface plasmon resonance ^ is dissipated in the silver. The closer to the silver the more power is drawn from the dipole and dissipated. This is due to that more of the evanescent nearfield from the dipole reaches the silver layer, where it is eventually dissipated, not coupled out, by Snell's law. Since the refractive indices of the glass and the PVA layers are the same, the evanescent nearfield from the dipole is not coupled out at all; only non-evanescent field is coupled out. Instead, the PVA and Si02 layers take part in the surface plasmon resonance in the silver layer, by interfering reflections at the PVA-air interface, influencing the behavior. SPCE-angles as well as power levels are affected.

82 N. CALANDER

Air PVA Si02 Silver Glass

Figure 3.10. The normalized power flow in the layers. The fluorophore is situated at 0. The solid line is for normal orientation to the planes and dotted line for parallel orientation. The power is normalized to the power from a fluorophore in bulk PVA. The power is strongly dissipated in the silver layer, suggesting thinner silver layers for less dissipation. The decay enhancement is equal to the sum of the normalized powers immediately to the left and right of the fluorophore.

One may think that a thinner silver layer would dissipate less of the dipolar output. That matter is analyzed in Figure 3.11 and it indeed seems to be the case to a certain extent, but for the very thin layers another surface plasmon mode is excited (not the one described in ^^), dissipating the power further. The highest output SPCE power seems to be at a thickness of 22.5 nm. Unfortunately the SPCE peak widens for thinner silver layers, so there is a compromise between output power and peak sharpness. The highest peak SPCE power per solid angle is reached at a thickness of 48.3 nm.

The decay times of fluorophores at different positions and orientations in the PVA layer are shown in Figure 3.12. Due to the increased power drawn from the fluorophores near the silver surface by the plasmon resonance, the decay times are much shortened there. An accurate comparison with the experimental decay times reported in Reference ^ gives an indication of that the fluorophores are concentrated to the surfaces rather than randomly positioned in


10

10 h

10 k

10 >

10

10

1 !,. .<^J*^m»»^»

V X

L ;

1 i 1 1 1

1 1 1 » 1 1 1 % 1 ' 1 1 > 1 1 1 1

I I s ; ;; 1 1 :

: : : ^ ^ '

t 1 i 1 1 t i

1

: . : : ^

i 1 I i 1 1 1 1

10 iim

20 50 100

Figure 3.11. The thickness of the silver layer is varied. Radiating dipole normal to plane. Solid curve: normalized output power to the back side. Maximum output at 22.5 nm. Dashed curve: Decay enhancement. Dotted curve: Probability of radiation into the back side. Dot-dashed curve: normalized output power into the front side. A highly evanescent surface plasmon resonance mode dissipates lots of energy at thin layers. Principally the same behavior occurs for a radiating dipole parallel to the plane.

the PVA layer, because the experimental decay times show a higher representation of the shortest and longest decay times.

For thick PVA layers a new phenomenon occurs. Both highly peaked p-polarized and s-polarized light appear at the backside. This is found experimentally and reported in Reference ^ . It is illustrated for a PVA thickness of 482 nm in Figure 3.13, and in Figure 3.14 where the electromagnetic field and energy density are shown. This phenomenon is due to that the PVA layer acts as a Fabry-Perot resonator, coupled together with the surface plasmon resonance in the silver layer. The alternating p- and s-polarized highly directed peaks are due to the different Fabry-Perot modes in the PVA layer.

84

10°

s OB <J

N

(

» . • . « . • . .

1

....|...

r

solid line: ±, dashed line: || T f f

xL.x I l...^

r \ - \ - r ? : \ \ : : \ * : L .\ . .Li. !

::::::::::::::::::ix::li£:::::::::::::!::::::::

r \* r

N

1

v-r -N

\ ^ s. . > ( . — V

r \

i i i i i

) 5 10 15 20 mil in PVA la^er

.CALANDER

25 30

Figure 3.12. The normalized decay times at different orientations and positions in the PVA layer. Values given for both the 15 and the 30 nm thick PVA layers.

3.3.3. Conclusions of surface plasmon resonance at planar structures

The theoretical approach to SPCE developed in this section reproduces the experimental findings in References ' ' ^ well. Because the fluorophores are confined in a layer (PVA) with the same refractive index as the glass prism, the waves that couples out into the prism are not at all evanescent in the PVA layer. The dependence of SPCE emission intensity, and backward to forward ratio, on position and thickness, are due to cooperative resonance of surface plasmons with the PVA-Si02 layers, not to evanescent field weakening with distance from the fluorophore. Thick PVA layers act like standing wave resonators (Fabry-Perot resonators) coupled to the surface plasmon resonances. For thin PVA layers only single p-polarized SPCE occurs. For ticker PVA layers alternating multi-peaked p- and s-polarized emission occur. The decay times of fluorophores near the silver layer are shown to shorten due to quenching, which is due to dissipation of surface plasmon enhanced highly evanescent waves into the silver layer. It seems like the fluorophores in the experimental situation are concentrated to the surfaces, at the near and far ends of the PVA layer, if the data is interpreted as large proportions of short and long decay times.

Concerning the merits of the two different arrangements for the excitation, they depend very much of what should be analyzed. Excitation from the backside (B in Figure 3.6), by a p-polarized beam, via the surface plasmon resonance, give stronger excitation field near the surface, with a strong component perpendicular to the plane. This is particularly good if strong SPCE emission is wanted. On the other hand, if polarization aspects of the


fluorophores are important, using the azimuthal variation of the SPCE emission (Figure 3.8), excitation from the front side (A in Figure 3.6) may be preferred, exciting the fluorophores with a polarization in parallel to the plane.

The comparison of the theoretical and experimental results show that the theoretical approach in this section is a useful tool for design of experimental structures and setups, for fine-tuning experiments, and for explanation of experimental results. It is also a useful tool for further development of SPCE techniques.

3.4. THEORY OF SURFACE-PLASMON RESONANCE OPTICAL-FIELD ENHANCEMENT AT PROLATE SPHEROIDS

The optical field enhancement from plasmon resonance at spheroids is studied by solving the Maxwell's equations using spheroidal vector wave functions. This treatment is an extension of the Mie theory for spheres. The phase retardation or dephasing effects is clearly shown by this method, confirming suggestions of previous research ^^ Nevertheless, the optical field enhancement is shown to be substantial under certain resonance conditions. It is suggested that the positions of the resonances in parameter space are determined by global antenna properties and the magnitude of the field enhancement by local plasmon resonance. This section is mainly a review of Reference ' .

The optical field enhancement at the end of a sharp tip has been used for a number of purposes. For example, in surface enhanced Raman spectroscopy (SERS) ' ' ' ' ^ the optical field enhancement, possibly together with some other phenomena, will increase the Raman signals such that spectra of single molecules are possible^ ' ' ' . Single molecule SERS experiments on Ag and Au colloids have indicated signal to noise enhancement factors of 10 - 15 orders of magnitude^ ' ^. Since the spectroscopic signal to noise enhancement goes as the fourth power of the field enhancement (light power two ways), the field enhancement would then be around 3 orders of magnitude at the detection spots.

86 N. CALANDER

k %

Figure 3.13. Same as in Figures 3.7 and 3.8 but PVA thickness 482 nm and no Si02 layer as in Reference^l Alternating semi-rings of p- and s-polarized light appear at the back side.


-The layers The layers

-20 -10 0 10 20 30 40 50

Figure 3.14. Same as in Figure 3.9 but PVA thickness 482 nm and no SiOi layer as in Reference . The two spread out rays in the energy density picture corresponds to the two narrow peaks in Figure 3.13, lower left diagram. Narrow peaks in far-field by necessity must correspond to spread out electromagnetic near-field due to the Huygens principle and Fourier transforms. Note the different directions of the dipoles.

Other examples: The strongly localized optical field enhancement at sharp tips may give nanometer resolution in aperture-less near-field scanning optical microscopy (NSOM) '*. The field enhancement is used for frequency mixing in scanning tunneling microscopes^^. The strongly localized field and strong field gradient at sharp tips has been proposed for optical nanotweezers^^. In the next section attraction of fluorescent molecules to enhancement spots is investigated^^.

A number of theoretical studies of the optical field enhancement have been done, using different methods^ ' '" ^ Exact methods exist for some simple shapes. The treatment will be further simplified if the quasistatic (electrostatic) limit is considered, i.e. the structures are much smaller than the wavelength. The main topic of this chapter is field enhancement at spheroids in the non-static and exact case. Multiple multipole methods^^ (MMP) and finite element methods (FEM) such as finite element time domain methods^ (FETD) have been used for more complex structures. Since the accuracy of these methods sometimes may be questioned, they have been tested^ on the simple structures where exact methods exist. Here is contributed an exact treatment in the case mentioned above.

Theoretical studies of the optical field enhancement at for example prolate spheroids show strong enhancements in the quasi-static limit^^. In previous research ^ the field enhancement factor at the end of a spheroid is calculated by the FETD method. The enhancement factor strongly depends on the ratio of the

88 N. CALANDER

wavelength to the structure size. The enhancement falls off rapidly as the structure size is enlarged; this is confirmed by calculations using spheroidal vector wave functions ' . This is attributed to phase retardation or dephasing effects.

The main objective of this section is to show, by using the exactly solvable model of a spheroid, that there are plasmon resonance regions with strong field enhancements in the non-static regime, and how their positions in parameter space and their magnitudes are related. In fact, there is a strong dependence on the parameters wavelength, spheroidal length, axial ratio, polarization, angle of incidence, and material properties. These sometimes very sensitive spots in the space spanned by these parameters may be used for specific sensing of the near environment of the spheroids. One may think of optimized nanostructures, not necessarily spheroidal, for this purpose.

3.4.1. The Field Enhancement at Spheroids

The Maxwell's equations have been solved in prolate spheroidal coordinates and the fields are expressed as sums of prolate spheroidal vector wave functions. The appropriate derivations are shown and references given in Section 3.0 below.

The incident field is a plane wave linearly polarized. The optical electric field near the spheroid is highly evanescent and usually elliptically polarized. If the incident wave is directed perpendicular to and polarized parallel to the main axis, the electric field at the top and bottom of the spheroid is, in the appropriate cases studied in this section, almost linearly polarized and directed along the main axis. The optical field enhancement y is defined as the ratio of the root mean squares of the scattered and the incident electric fields; see Section 3.0 for further details.

The optical field enhancement y at the spheroids is determined by the length a, the axial ratio a/b, the material, the wavelength X, the direction of the incident optical radiation 0, and the polarization a, illustrated in Figure 3.15. The only material property used is the complex relative permittivity 8r (dielectric function), which is calculated from the complex refractive index n = n+ik by 8r = n . The refractive indices are taken fi-om Reference^^ The relative permeability |Xr is assumed equal to one.

Figure 3.16 shows the dependence of y on a/X, for two values of a/b, for gold spheroids, for incidence perpendicular to and polarization parallel to the main axis, i.e. 9 = 90° and a = 0°. The calculation is based on prolate spheroidal vector wave functions. The strong dephasing effect mentioned in Reference^ is confirmed.


Figure 3.15. The geometry of a prolate spheroid and the incoming plane wave. Two of the principal axes are assumed equal to b and the remaining axis equal to a, the long axis. The axial ratio is defined as a/b, 9 is the angle of the wavevector to the main axis of the spheroid, and a is the polarization angle of the incoming E-field.

Figure 3.16. Starting with a small spheroid (the quasi-static limit, a/A, = 0) the field enhancement increases and culminates at a resonance peak, and then declines due to the dephasing effect. The dephasing effect means that different parts of the structure are out of phase. Solid line: a/b = 3.0. Dashed line: a/b = 3.21. Gold spheroid with X = 633 nm and e = -10.84 + 0.762i.

The dephasing is due to that field scattered fi-om different parts of the structure (in this case spheroid) are out of phase and cause destructive interference. If the scattered fields are in phase at the scatterers (electrons) they reinforce each others fields cooperatively, i.e. the structures sustain a surface plasmon resonance. The plasmon field enhancement is therefore due to the phasing of the scatterers depending on the structure and the wavelength, which is seen in Figure 3.17.

90 N. CALANDER

a)

13

Z 9

20

1S

10

5

1

l.^

^

• '"" '1 t

^

...--20-

' f %-

}. - 1 200

" / r ^ 100

"iM.ii t i

" ' » »""""" •

20 .

^0

'•: ' \

Mi l t

" • " 1

%

>

^ 30

^

to

1

'1 1

^0

>a 1 . ' ^

1 i 1

l i ., il

0.2 0.4 0.6 0.8 1.2 1.4 1.6

length (fuu)

b)

5K

W!\ 'M'i /•''••'•' > ''V'

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6

length (^m)

Figure 3.17. a) Contour plot of y versus length a and axial ratio a^, at the top of a silver spheroid when X = 0.633 jiim, incidence perpendicular and polarization parallel to the main axis, b) The same but for a perfectly conducting spheroid.

In Figure 3.17 the field enhancement at the top or bottom of a silver spheroid (Figure 3.17a) is compared to the enhancement at a perfectly conducting spheroid (Figure 3.17b). The angle of incidence is 6 = 90° and the polarization is a = O*'. The enhancements show rather different patterns. In the


perfectly conducting case strong enhancement show up when the length a is near an odd integer times a half wavelength a = (n+/4)X-. The enhancement gets stronger for higher axial ratios and gets closer to the odd integer times the half wavelength from below. Compare this with Figure 3.1 in Reference^\ In the silver case enhancement peaks show up at localized islands according to different resonance modes. The axial ratios do not have to be as high in the silver case as in the perfectly conducting case for high enhancement y to occur. This is due to the surface plasmon resonance effect in the silver spheroid. The high enhancement in the perfectly conducting case at the very high axial ratios is due to crowding of the field lines at the small radius of curvature, also called the "lightning rod effect".

It is seen that the very high enhancement factors are strongly diminished for the larger spheroids. This is attributed to the dephasing effect. It is also seen that new resonance regions occur due to higher modes illustrated in Figure 3.19. Some of the resonances are rather localized, i.e. a small change in wavelength or axial ratio causes a substantial change in field enhancement.

250

Figure 3.18. Field enhancement y at silver spheroid of length a = 0.2 }im, for 0 = 90° and a = 0°. Axial ratio a/b = 10 and wavelength A- = 1.1 p,m. A (11 l)-plane for silver is drawn to scale in the right figure, giving a feeling of the tip size. The maximum field enhancement is a theoretical value for a smooth spheroid.

92 N. CALANDER

Figure 3.18 shows the field enhancement at a silver spheroid of length a = 0.2 |Lim, axial ratio a/b = 10 which gives the width b = 0.02 jiim, and at the wavelength X = 1.1 |Lim, for 6 = 90* and a = 0°. Although the field enhancement is rather large near the tip due to the calculations, the tip is so sharp in this case such that the atomic structure may come into play.

Figure 3.19 shows the field around the spheroid to illustrate different modes and Figure 3.20 shows y when the incidence angle 6 and the wavelength X are varied. The field enhancement dependence on the polarization angle a seems to be mainly on the projection of the field onto the main spheroidal axis.

determines the position of the maxima and depends on the length a, the axial ratio a/b, the wavelength X, and the relative permittivity 8r.

Table 3.3 shows the local maxima of the enhancement in Figure 3.17a and their positions. It is found that the field enhancement at the local maxima, except the first which is quasi-static, is related to the curvature of the tip. From Figures 3.18 and 3.19 it is seen that the field enhancement is strongly localized to the ends of the spheroid, at least for the more elongated structures. It is conjectured that the field enhancement at the tip is a combination of a local and a global effect. The local effect determines the size of y at the maxima and depends on the curvature at the tip, the wavelength X, and the relative permittivity 8 . As is discussed in Reference^^ in the quasi-static case, the local effect has two contributions, the "lightning rod effecf due to crowding of field lines to the small radius of curvature, and a "resonant effect" due to excitation of local surface plasmons. The global effect, due to global antenna properties,

Figure 3.19. Field enhancement y at silver spheroid of length a = 0.6 jum and axis ratio a/b = 12 (width b = 0.05 }im), for 8 = 90° and a = 0°. The whiter the more field enhancement. Left: >. = 1.2 |im. Middle: X = 0.9 |im. Right: X = 0.73 |Lim. This figure shows enhancement for two kinds of resonance modes and the enhancement between the modes.


250

200

150

100

0.7 0.8 A(ftm)

Figure 3.20. Plot showing y dependence on y and X. The polarization a= 0°. The spheroidal length a = 0.6 jLim and width b = 0.6 jiim. Some local maxima are shown as x, and some local minima as •.

determines the position of the maxima and depends on the length a, the axial ratio a/b, the wavelength X, and the relative permittivity 8r.

Table 3.2. The local maxima in Figure 3.3a are here tabled. The radius of curvature R at the tip of a spheroid is given by b^/2a. It seems like the enhancement maxima are related to the radii of curvature, except for the first which is quasi-static.

r# 1 2 3

[_4

Y 250.% 85.467 64.584 56.189

a(um) 0.00000 0.50177 1.03160 1.62317

a^ 4.8037 9.4121 12.921 15.822

R(um) 0.00000 2.8321 3.0895 3.2420

YR'Clim') 1 0.00000 0.68549 0.61647 0.59059 1

94 N. CALANDER

3.4.2. Conclusion of Surface-Plasmon Resonance Optical-Field Enhancement at Prolate Spheroids

In this work " the enhancement factor y at spheroids have been calculated by using spheroidal vector wave functions, and compared to the results of other methods used in previous works. The spheroid with its elongated shape has a much better model geometry than other geometries accessible to exact calculations such as the sphere.

An advantage of this accurate calculational method is that it uncovers narrow very high enhancement regions in the non-static case; where the finite element methods tend to give worse results, see for example Figure 3.16 and Reference ^^ An application of the narrow enhancement regions may be for sensing of molecules nearby. The accurate calculations on the spheroid as a model make it possible to draw conclusions about the positions and magnitudes of the plasmon resonances, see Table 3.2.

3.4.3. Solving the Maxwell's Equations in Prolate Spheroidal Coordinates

Most definitions used here are found in Reference"^ . The prolate spheroidal coordinates (r|,^,(p) are defined by the following expressions for the Cartesian coordinates (x,y,z):

x = f^(l-Ti^)(^^-l cos(cp) (8)

y = f^(l-T1^)(^^-l)sin((p) (9)

z = fTi (10)

The distance between the foci is denoted f The ranges of the prolate spheroidal coordinates are -l<r|<l, 1<^, and 0<(p<27c, and Figure 3.21 gives an illustration.


^ = # 0

Figure 3.21. Illustration of the spheroidal coordinates.

3.4.3.1. The Quasi-Static Limit

The quasi-static limit may be calculated in the same way as an electrostatic problem. The procedure for that may be found in many textbooks of electromagnetics"^^. The problem to solve in our case is V^V(r|,5,(p) = 0 for the potential V with the boundary conditions that V(r|,^,cp) and 8(r|,§,(p)6V(r|,§,(p)/6§ should be continuous across the spheroidal boundary ^ = ^o, that V(r|,^,(p) ^•-Ez = -Er|^ when ^ —> oo, and that V is finite everywhere. The solution outside the spheroid is

V = E[^+AQ,(^)]II

where

^ „ ^ ( ^ o ) - ^ Q . ( ^ o )

(11)

(12)

Q>(^) = | l n ^ - 1

-1 is a Legendre function of the second kind

The field enhancement y at the top or bottom of the spheroid (r| = ±1, § = ^o)is

Ttop ''-t<^)^ (13)

Equation (13) is illustrated in Figure 3.23. The equation is given in a different form in References^ ' .

96 N. CALANDER

Figure 3.22. The optical field enhancement y at the spheroidal ends, in the quasi-static limit.

3.4.3.2. The Prolate Spheroidal Wave Functions

The definitions and calculations of the scalar and vector prolate spheroidal wave functions are described in Reference"^ . The vector wave functions should satisfy V^F+k^F=0 and V-F=0 (underline means vector). There are solutions defined by M mn==V\j/mnXa and N^nin=(l/k)VxM^mn where M/nm=Smn(r|)Rmn( )exp(im(p) and a-x±iy, a=z or a-r. ^^nM) and R,nn( ) are scalar spheroidal wave fiinctions defined in Reference"^ . Here 2^ y and z are the three Cartesian unit vectors and r the radius vector.

3.4.3.3, Solving the Maxwell's Equations at a Dielectric Spheroid

The solution method is to a certain extent similar to the one used in References^^-^^ We use the functions with a=x= iy and a=z . It has also been tried to use the flmctions with a=r ^ " '*, which give simpler boundary conditions, but found that the convergence in this case is much worse for the more elongated structures.

The electric and magnetic fields are expanded in terms of prolate spheroidal vector wave functions. Functions of only the first kind are used inside the spheroid and functions of both the first and third kinds outside. The joining at the surface of the spheroid is that the tangential components of the electric and magnetic fields should be continuous. The incident plane wave is


expanded in terms of functions of the first kind with known expansion coefficients'*^. The expansion coefficients of the outgoing functions of the third kind and the functions inside the spheroid are determined by the boundary conditions.

The scalar wave equation is separable in spheroidal coordinates, as is for example the vector wave equation in spherical coordinates. The vector wave equation is unfortunately not separable"^^ in spheroidal coordinates, which show up when joining functions for satisfying boundary conditions. The functions are separable in the cp-direction, but not in the r|-direction. Functions with different n's must match at the surface, which is solved by joining them in a least squares sense. The functions can in principle be matched pointwise, but a better method, a method we use, is to find the smallest deviations of integrals of the expansions with Legendre functions. We apply a slightly different method for calculation of the matching integrals, than the methods used in the References. Integrals like

iz.v = f p : ( ^ ) ^ ^ ^ (14)

where pe{0,l}, qe{-1,0,1}, re{-5,-4,-3,-2,-1,0,1} and se{0,1,2}, are solved by recurrence of m and n, starting from zeros. These recurrence relations are derived from known recurrence relations for the Legendre fimctions^^. All start integrals are evaluated exact, for example

1 1 1 ^ r i \ (15)

v^y

where K denotes a complete elliptic integral^^.

3.4.3.5. The Definition of Optical Field Enhancement

The incident electric field is a linearly polarized plane wave whereas the scattered electric field is highly evanescent and usually elliptically polarized. In this chapter we define the optical enhancement y as the ratio of the root mean squares of the scattered and the incident electric fields. The mean is taken of the square of the absolute magnitude of the electric field over a frequency cycle.

The physical electric field is given by 9i[Eexp(-ia)t)], where E denotes a complex vector quantity, an outcome of the calculations. 9? denotes taking the real part and o the angular frequency. The extremes (maximum and minimum) of the absolute magnitude of the physical electric field during a cycle are then given by (V2(EE ilEEl))"^, and the root mean square over a cycle is given by ( /sEE*)" . The * means complex conjugation and the • complex scalar product. Since in our calculations the incident field is normalized such that Ei-Ei =1 this

98 N. CALANDER

means that the optical filed enhancement y is given by (E-E*)' ' , which is the sameas(|E,p+|E/+|E^|Y\

3.5. OPTICAL TRAPPING OF SINGLE FLUORESCENT MOLECULES BY SURFACE PLASMON RESONANCE

Here a scheme of optical trapping of fluorescent molecules based on the strongly enhanced optical field due to surface plasmon resonances at laser illuminated metal tips or particles is proposed. Since there is a large optical field enhancement at small localized spots the optical field gradient must be huge and since the optical polarizability at resonance of a fluorophore is large one may ask how the gradient forces will affect the fluorophore. One may also ask whether the also strongly enhanced radiation out from the fluorophore (radiative decay enhancement) is enough for cooling the fluorophore in the intense optical field. A semi-classical approach is compared to a quantum-mechanical. Attractive as well as repulsive gradient forces are possible depending on the wavelength of the optical field. The trapping potential is shown to be strong enough to overcome the Brownian motion in water solution for common optical tweezers light intensities. The radiative decay enhancement is shown to keep the diffusion of the fluorophore in check. Single molecule SER(R)S probes are particularly well suited for the trapping scheme. This section is mainly a review of Reference '^.

Optical manipulation of particles has been achieved in water solution at room temperature, by the use of optical tweezers and optical scissors ^ . This technology has developed into an important research tool in cell biology. The objects manipulated are cells, organelles, and larger molecules. The spot of manipulation, or attraction, is conventionally created by confocal concentration of the optical energy supplied by a laser. Another way of concentrating optical energy is by the surface plasmon resonance at nanoprobes, which has been suggested for nanometric optical tweezers ^ , able of manipulating smaller particles than the conventional optical tweezers, particles typically down to a few nanometers. Here the possibility that the nanoprobes should also be able to manipulate even smaller particles, if they are fluorescent molecules, is investigated.

Optical detection and spectroscopy of single molecules have been achieved in water solution at room temperature. Particularly interesting is single molecule Raman spectroscopy (SERS) or single molecule resonance Raman spectroscopy (SERRS) ^ " ' ^ made possible by the surface plasmon resonance at colloidal particles acting as antennas or transformers amplifying the signal. The Raman signal is rather rich in structure compared to other optical spectroscopies, and is well suited for precise molecule discrimination. Here the possibility of attracting fluorescent molecules to the detection spots by optical means is investigated^^


The theory developed in Reference ^ is used and compared to some formulas derived here from a simple classical model "^ of a fluorophore. The theory in Reference ^ was primarily intended for laser cooling of atoms.

The optical force is divided into a scattering force and a gradient force. The scattering force is pointing in the direction of the incident light. The gradient force is pointing in the direction of the intensity gradient. From the previous section and Reference ^^, it seems like the huge field enhancement due to surface plasmon resonance, may be mainly considered as a standing inhomogeneous wave around the particle. A standing wave has no direction of propagation since it is composed of a number of (at least two) waves whose directions of propagation cancel. Therefore only the gradient force applies.

Now the simple classical mechanical model of the fluorophore is described^^ It consists of an electron in a three-dimensional harmonic potential, which by definition has unity oscillator strength, a characteristic of strong fluorescence transitions. The radiation damping is introduced via a damping force on the electron. The gradient force is described by the potential:

U = - l 9 i ( a ) | E f (16)

Here a is the linear polarizability (assumed the same in all directions) of the molecule, !R is taking the real part, and E is the electric field.

The equation of motion for the electron is:

me(l + r r + cOorW-eE (17)

Here mg is the electron mass, r the electron position, F the radiation damping, CDQ the resonance angular frequency, and e the charge of the electron. A dot means taking the time derivative.

Assuming harmonic oscillations (time derivative -^ -ico) the polarizability of such a model fluorophore is then given by:

«H=V%F- (18) (OQ - CO - iFco

The radiation damping is given by Fermi's golden rule combined with the conditions for unity oscillator strength:

2 2

27i8C^m^ r = i ^ (19)

Putting equations (16) and (18) together give:

100 N. CALANDER

2 / 2 2\

U = r-^ ^ M (20) 2mMoi^-(olf -^oy^r^

A realistic two-level molecule has an upper limit of the polarization (and the energy) which our model does not have. Therefore the energy of the oscillations is limited to hcoo (h is Planck's constant divided by 2n). The saturated polarization is then given by:

Ich Psat -J (21)

I m^coj e^O

By integrating the energy from zero electric field one then gets the saturated potential (sign(9i(a)) means the sign of 91(a)):

Usa ,=^J^-Psa .Es ign(9?(a ) ) (22)

The quantum mechanical treatment is based on the optical Bloch equations and the dressed atom approach, and is taken from reference ^ . Only gradient forces are considered. A molecule with two electronic states in a photon field with an amplitude gradient is considered. By reference ^ the gradient force is described by a potential given by

^ ^ M C O - C O Q ) , / 1+- (23)

Here Q. is the Rabi frequency, which is given by:

p E Q = ^^^-^ (24)

Here ptr is the transition dipole moment and is for a unit oscillator strength fluorophore given by:

3e^h Ptr=J- (25)

f 2m coo

Saturation is also seen in this case, the potential has a logarithmic dependence on the electric field when "saturated".


Now the field enhancement due to surface plasmon resonance is considered. The optical field enhancement may be rather large. From references ^ ' ^ the SERS signal enhancement may be 10 to 15 orders of magnitude. Since the SERS signal enhancement goes as the fourth power of the amplitude enhancement (SERS signal is power enhancement to-ways), the amplitude enhancement is 300 to 5000. This is also in line with the results for the prolate spheroid in the previous section. The enhanced field is strongly evanescent and short-range. Since a broad range of photon modes are coupled to the strong field enhancement, the rate of spontaneous emission from a molecule in the enhancement region is also affected. Describe the enhancement of the electric field E by the coefficient y:

E -^ yE (26)

Then for the rate of spontaneous emission:

r->xY^r (27)

The square dependence for the spontaneous emission rate is due to the square dependence of the transition matrix element in the Fermi's golden rule. % is about one third, since <cos^0>=V3, for a field enhancement particle small compared to the wavelength and less than one third otherwise.

The increased spontaneous emission rate affects the trapping potential capabilities by making it lower, see Equations (16), (18), and (23). There is also a good side of the increased spontaneous emission, because it lowers the diffusion. In reference ^ a dipole diffusion coefficient for the molecule momentum due to the optical field is:

^ d i p - - (28)

It is in reference ^ warned that the diffusion coefficient increases more rapidly with the laser intensity than the trapping potential. Fortunately, the spontaneous emission rate also increases rapidly with field enhancement and balances it out in the diffusion constant. The diffusion constant above, Ddip, is for the particle momentum. For a particle moving in a viscous medium the relation between momentum, force and particle position is:

p = F = Kr (29)

Here p is the momentum, r the position, F the force, and K the frictional coefficient in the viscous medium. A dot means time derivative. This means that the diffusion coefficient for the particle position is given by:

102 N. CALANDER

D = — ^ GO)

For a spherical particle of radius R the frictional coefficient is given by:

K = 67iRr| (31)

Here r| is the viscosity. The ordinary positional diffusion constant in a viscous liquid is given by

D v i s c = ^ (32)

Here ke is Bolzmann's constant and T the absolute temperature. For a calculation of a typical measure of the diffusion coefficients, let us

use a molecule of diameter 1 nm and the values in Figure 3.25. One gets that Dpos = 3.8-10" ^ m^/s which should be compared to Dvisc = 4.110'^^ m /s which is about an order of magnitude larger. If the increase of the spontaneous emission rate is not considered the diffusion constant due to the optical field may be orders of magnitudes larger than the ordinary diffusion coefficient which would be disastrous for the trapping.

As is seen in Figure 3.23 the three descriptions give rather different results. They all almost coincide far from the resonance. Near the resonance the saturation come into play. It is unfortunate that the simple unsaturated classical model does not work; it should have given an easy way to manipulate the fluorophore at a very specific frequency. That effect should have been even more pronounced than in Figure 3.23 if the decay rate did not increase (but the decay rate is necessary in order to keep the diffusion in check, see above). The modified simple model, with saturation, also show too good results compared to the quantum mechanical model from reference ^ . The quantum mechanical (most realistic) model still show trapping but it is not very precise in frequency.

Figure 3.24 shows the trapping potential at a spheroid and between two spheres. The calculation of the field enhancement assumes the Rayleigh limit (quasi-static) which in this case gives a good approximation. (Exact calculation using spheroidal vector wave functions, see previous section and Reference ^'^, for the spheroid and extended Mie theory for the spheres could be done instead). Other shapes with sharper ends would show stronger field enhancement. The spots showing largest attraction also show the highest sensitivity for Raman spectroscopy (SERS).


L4 0.5 0.6 0 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6

Figure 3.23. The potential acting on a unit oscillator strength fluorophore. The laser intensity at the probe is 50 mW/fim^, the field enhancement at the probe is 1000, the surrounding medium is water and the resonance wavelength of the fluorophore is 0.7 |im^. Unfortunately the most realistic model has the poorest trapping potential, but still it has a trapping potential of 10 ksT at A, = 0.8 - 1 (xm, which is enough to overcome the Brownian motion. A real probe has its own resonance frequency, taking out only a small portion of the diagram. The wavelength region A, = 0.8 - 1 fxm is particularly convenient to minimize destruction of cell components.

In order for this trapping technique to work well, one may have to put some effort on probe optimization. Theoretical calculations on simple shapes have been done as has already been mentioned. High field enhancement is achieved at the ends of the more elongated spheroids (se previous section or Reference ' ). It has also been achieved between almost touching spheres ^ . This is shown in Figure 3.24. High field enhancement seems to have been achieved experimentally on irregular shapes where the hot spots are believed to be at sharp irregularities or between particles ^ . Design criteria for a good fluorophore trap would be high enhancement at as large volume as possible, together with coupling to a broad range of photon modes to increase the spontaneous emission rate.

There are some questions concerning heating. It has been shown in this section that the enhanced decay brings the diffusion due to the discrete nature of the photons down. It is likely that heating due to other effects also is brought down, due to the enhanced decay rate, but it is difficult to judge on theoretical grounds only.

Poor fluorophores are lifted to high quantum yield fluorophores when the strong radiation channel appears. The trapping of molecules possessing several fluorescent groups would benefit from their cooperative action.

A metal particle possessing a strong surface plasmon resonance spot should be easy to select by optical tweezers, since the resonance also enhances the forces holding the particle.

104

Figure 3.24. This figure shows the optical trapping potential in units of ksT at a silver spheroid and between two silver spheres in liquid water. The calculation is done in the Rayleigh limit, i.e. feature sizes of the structures much smaller than the wavelength. Spheroidal axial ratio 4.78, distance between spheres one tenth of a radius. Wavelength at spheroid is 0.8 im and at spheres 0.4 [im. Field intensity before enhancement is 500 mW/ im . Maximal trapping potential at spheroid in this case is 12 keT and between spheres 14 keT. Note that the gray scale is logarithmic.

3.6. REFERENCES

1. J. Homola, S.S. Yee and G. Gauglitz, Surface plasmon resonance sensors: review. Sensors and Actuators B-Chemical, 1999. 54 (1-2): p. 3-15.

2. B. Liedberg, I. Lundstrom and E. Stenberg, Principles of Biosensing with an Extended Coupling Matrix and Surface-Plasmon Resonance. Sensors and Actuators B-Chemical, 1993. 11 (1-3): p. 63-72.

3. B. Liedberg, C. Nylander and I. Lundstrom, Biosensing with Surface-Plasmon Resonance - How It All Started. Biosensors & Bioelectronics, 1995. 10 (8): p. R1-R9.

4. S. Lofas, M. Malmqvist, I. Ronnberg, E. Stenberg, B. Liedberg and L Lundstrom, Bioanalysis with Surface-Plasmon Resonance. Sensors and Actuators B-Chemical, 1991. 5 (1-4): p. 79-84.

5. J.R. Lakowicz, J. Malicka, L Gryczynski and Z. Gryczynski, Directional surface plasmon-coupled emission: a new method for high sensitivity detection. Biochemical and Biophysical Research Communications, 2003. 307 (3): p. 435-439.

6. J.R. Lakowicz, Radiative decay engineering 3. Surface plasmon-coupled directional emission. Analytical Biochemistry, 2004. 324 (2): p. 153-169.

7. L Gryczynski, J. Malicka, Z. Gryczynski and J.R. Lakowicz, Radiative decay engineering 4. Experimental studies of surface plasmon-coupled directional emission. Analytical Biochemistry, 2004. 324 (2): p. 170-182.

8. J. Malicka, I. Gryczynski, Z. Gryczynski and J.R. Lakowicz, DNA hybridization using surface plasmon-coupled emission. Analytical Chemistry, 2003. 75 (23): p. 6629-6633.

9. N. Calander, Theory and simulation of surface plasmon-coupled directional emission from fluorophores at planar structures. Analytical Chemistry, 2004. 76 (8): p. 2168-2173.


10. K. Kneipp, Y. Wang, H. Kneipp, L.T, Perelman, I. Itzkan, R. Dasari and M.S. Feld, Single molecule detection using surface-enhanced Raman scattering (SERS). Physical Review Letters, 1997. 78 (9): p. 1667-1670.

U.K. Kneipp, H. Kneipp, V.B. Kartha, R. Manoharan, G. Deinum, I. Itzkan, R.R. Dasari and M.S. Feld, Detection and identification of a single DNA base molecule using surface-enhanced Raman scattering (SERS). Physical Review E, 1998. 57 (6): p. R6281-R6284.

12. K. Kneipp, H, Kneipp, I. Itzkan, R.R. Dasari and M.S. Feld, Surface-enhanced non-linear Raman scattering at the single-molecule level. Chemical Physics, 1999. 247 (1): p. 155-162.

13. S.M. Nie and S.R. Emery, Probing single molecules and single nanoparticles by surface-enhanced Raman scattering. Science, 1997. 275 (5303): p. 1102-1106.

14. N. Calander and M. Willander, Theory of surface-plasmon resonance optical-field enhancement at prolate spheroids. Journal of Applied Physics, 2002. 92 (9): p. 4878-4884.

15. N. Calander and M. Willander, Optical trapping of single fluorescent molecules at the detection spots of nanoprobes. Physical Review Letters, 2002. 89 (14).

16. J.R. Lakowicz, Radiative decay engineering: Biophysical and biomedical applications. Analytical Biochemistry, 2001. 298 (1): p. 1-24.

17. J.R. Lakowicz, Y.B. Shen, S. D'Auria, J. Malicka, Z. Gryczynski and I. Gryczynski, Radiative decay engineering: Biophysical applications. Biophysical Journal, 2002. 82 (1): p. 426A-426A.

18. J.R. Lakowicz, Y.B. Shen, S. D'Auria, J. Malicka, J.Y. Fang, Z. Gryczynski and I. Gryczynski, Radiative decay engineering 2. Effects of silver island films on fluorescence intensity, lifetimes, and resonance energy transfer. Analytical Biochemistry, 2002. 301 (2): p. 261-277.

19. J.R. Lakowicz, J. Malicka, I. Gryczynski, Z. Gryczynski and CD. Geddes, Radiative decay engineering: the role of photonic mode density in biotechnology. Journal of Physics D-Applied Physics, 2003. 36 (14): p. R240-R249.

20. J.R. Lakowicz, Radiative decay engineering 5: metal-enhanced fluorescence and plasmon emission. Analytical Biochemistry, 2005. 337 (2): p. 171-194.

21. K. Asian, I. Gryczynski, J. Malicka, E. Matveeva, J.R. Lakowicz and CD. Geddes, Metal-enhanced fluorescence: an emerging tool in biotechnology. Current Opinion in Biotechnology, 2005. 16(1): p. 55-62.

22. I. Gryczynski, J. Malicka, K. Nowaczyk, Z. Gryczynski and J.R. Lakowicz, Effects of sample thickness on the optical properties of surface plasmon-coupled emission. Journal of Physical Chemistry B, 2004. 108 (32): p. 12073-12083.

23. E.D. Palic, Handbook of Optical Constants of Solids. 1985, New York: Academic. 24. M. Bom and E. Wolf, Principles of Optics. 1980, Oxford: Pergamon. 25. CW. Chew, Waves and Fields in Inhomogeneous Media. 1995, New York: Van Nostrand

Reinhold. 26. R.E. Benner, R. Domhaus and R.K. Chang, Angular Emission Profiles of Dye Molecules

Excited by Surface-Plasmon Waves at a Metal-Surface. Optics Communications, 1979. 30 (2): p. 145-149.

27. M. Abramowitz and LA. Stegun, eds. Handbook of Mathematical Functions. 1 ed. 1965, Dover Publications, Inc: New York.

28. A. Sommerfeld, Partial Differential Equations in Physics. 1949, New York: Academic Press. 29. F.J.P. Schuurmans and A. Lagendijk, Luminescence of Eu(fod)(3) in a homologic series of

simple alcohols. Journal of Chemical Physics, 2000. 113 (8): p. 3310-3314. 30. W.H. Weber and CF. Eagen, Energy-Transfer from an Excited Dye Molecule to the Surface-

Plasmons of an Adjacent Metal. Optics Letters, 1979. 4 (8): p. 236-238. 31. Y.C Martin, H.F. Hamann and H.K. Wickramasinghe, Strength of the electric field in

apertureless near-field optical microscopy. Journal of Applied Physics, 2001. 89 (10): p. 5774-5778.

32. D.J. Maxwell, S.R. Emory and S.M. Nie, Nanostructured thin-film materials with surface-enhanced optical properties. Chemistry of Materials, 2001. 13 (3): p. 1082-1088.

33. H.X. Xu, E.J. Bjemeld, M. Kail and L. Borjesson, Spectroscopy of single hemoglobin molecules by surface enhanced Raman scattering. Physical Review Letters, 1999. 83 (21): p. 4357-4360.

34. F. Zenhausem, M.P. Oboyle and H.K. Wickramasinghe, Apertureless near-Field Optical Microscope. Applied Physics Letters, 1994. 65 (13): p. 1623-1625.

106 N. CALANDER

35. T. Gutjahr-Loser, A. Homsteiner, W. Krieger and H. Walther, Laser-frequency mixing in a scanning tunneling microscope at 1.3 mu m. Journal of Applied Physics, 1999. 85 (9): p. 6331-6336.

36. L. Novotny, R.X. Bian and X.S. Xie, Theory of nanometric optical tweezers. Physical Review Letters, 1997. 79 (4): p. 645-648.

37. J. Jersch, F. Demming, L.J. Hildenhagen and K. Dickmann, Field enhancement of optical radiation in the nearfield of scanning probe microscope tips. Applied Physics a-Materials Science & Processing, 1998. 66 (1): p. 29-34.

38. C. Girard and A. Dereux, Near-field optics theories. Reports on Progress in Physics, 1996. 59 (5): p. 657-699.

39. H.X. Xu, J. Aizpurua, M. Kail and P. Apell, Electromagnetic contributions to single-molecule sensitivity in surface-enhanced Raman scattering. Physical Review E, 2000. 62 (3): p. 4318-4324.

40. W. Denk and D.W. Pohl, Near-Field Optics - Microscopy with Nanometer-Size Fields. Journal of Vacuum Science & Technology B, 1991. 9 (2): p. 510-513.

41. J.P. Kottmann, O.J.F. Martin, D.R. Smith and S. Schultz, Field polarization and polarization charge distributions in plasmon resonant nanoparticles. New Journal of Physics, 2000. 2: p. 271-279.

42. C. Flammer, Spheroidal Wave Functions. 1957, Stanford, California: Stanford University Press. 43. J.D. Jackson, Classical Electrodynamics. 2 ed. 1975, New York: John Wiley & Sons. 44. B.P. Sinha and R.H. Macphie, Electromagnetic Scattering by Prolate Spheroids for Plane-Waves

with Arbitrary Polarization and Angle of Incidence. Radio Science, 1977. 12 (2): p. 171-184. 45. B.P. Sinha and R.H. Macphie, Electromagnetic Plane-Wave Scattering by a System of 2 Parallel

Conducting Prolate Spheroids. leee Transactions on Antennas and Propagation, 1983. 31 (2): p. 294-304.

46. M.F.R. Cooray, I.R. Ciric and B.P. Sinha, Electromagnetic Scattering by a System of 2 Parallel Dielectric Prolate Spheroids. Canadian Journal of Physics, 1990. 68 (4-5): p. 376-384.

47. M.F.R. Cooray and I.R. Ciric, Scattering by Systems of Spheroids in Arbitrary Configurations. Computer Physics Communications, 1991. 68 (1-3): p. 279-305.

48. M.F.R. Cooray and I.R. Ciric, Scattering of Electromagnetic-Waves by a System of 2 Dielectric Spheroids of Arbitrary Orientation. leee Transactions on Antennas and Propagation, 1991. 39 (5): p. 680-684.

49. M.F.R. Cooray and I.R. Ciric, Scattering of Electromagnetic-Waves by a Coated Dielectric Spheroid. Journal of Electromagnetic Waves and Applications, 1992. 6 (11): p. 1491-1507.

50. S. Nag and B.P. Sinha, Electromagnetic Plane-Wave Scattering by a System of 2 Uniformly Lossy Dielectric Prolate Spheroids in Arbitrary Orientation. leee Transactions on Antennas and Propagafion, 1995. 43 (3): p. 322-327.

51. S. Asano and G. Yamamoto, Light-Scattering by a Spheroidal Particle. Applied Optics, 1975. 14 (1): p. 29-49.

52. S. Asano, Light-Scattering Properties of Spheroidal Particles. Applied Optics, 1979. 18 (5): p. 712-723.

53. S. Asano and M. Sato, Light-Scattering by Randomly Oriented Spheroidal Particles. Applied Optics, 1980. 19 (6): p. 962-974.

54. Y.P. Han and Z.S. Wu, Scattering of a spheroidal particle illuminated by a Gaussian beam. Applied Optics, 2001. 40 (15): p. 2501-2509.

55. M.P. Sheetz, ed. Laser Tweezers in Cell Biology. Methods in Cell Biology. Vol. 55. 1998, Academic Press: New York.

56. C.C. Tannoudji, J. Dupont-Roc and G. Grynberg, Atomic-Photon Interactions—Basic Processes and Applications. 1992, New York: Wiley.

CURRENT DEVELOPMENT IN THE DETERMINATION OF INTRACELLULAR NADH

LEVEL

Zhi-hong Liu, Ru-xiu Cai*^ and Jun Wang

4.1. INTRODUCTION TO NADH

Early in 1930s, Nicotinamide adenine dinucleotide (NAD) was separated and studied on its construction features. Now it has been well established that NAD plays very important role in many life-related processes such as genetic transcription, life-saving and disease-treating. For example, Kelly et al have found that the diabetics' ability of regulating NAD level decreased ^

As the reduced form of NAD (figure 4.1), NADH is a ubiquitous biological molecule that participates in many metabolic reactions, especially in the process of energy metabolism. NADH plays a significant role in the process of electron transfer. It acts as electron transporter and presides over the transportation of H2. The oxidation of NADH in the electron-transport chain releases energy, much of which is captured by mitochondrial ATP-synthesizing system. It is known that one NADH molecule produces three ATP molecules. NADH stimulates the biosynthesis of tyrosine hydroxylase and dopamine, thus it is effective in the treatment of Parkinson's syndrome. NADH was recommended for therapy by Birkmayer in 1993 instead of bendopa^. NADH was also accredited by FDA in 1998 for curing chronic fatigue because it is one of the unwonted nutritious extender which can stimulate the function of brain and supply energy.

Address of authors: Department of Chemistry, Wuhan University, Wuhan 430072, China

107

108 Z.H. LIU ETAL.

B H OH HO klf; H H

O O H 11

QH QH

P

NAD^

o o tt H

OH OH

NH2

IH.0J?'

Hi |H OH OH

H | [ ^ H OH OH

NADH

N

Figure 4.1. Structure of NAD^ and NADH.

4.2. SIGNIFICANCE OF DETERMINING INTRACELLULAR NADH

LEVEL

NADH is a necessary coenzyme in cellular metabolism, which is indispensable to all activities in life. Generally, intracellular NADH level is considered as an important biochemical criterion on many physiological and pathological events. For example, Jonas et al found that the conversion rate of NADH in hyperplasia cancer cells was much higher than the normal cells . It has also been found that the decrease of intracellular NADH content is linked to cell senescence and death. Harden and Young firstly found that NADH was related with glycolysis ^. There is an interdependent relation as follow between intracellular NADH and NAD^.

DETERMINATION OF INTRACELLULAR NADH 109

NADH 4- //^ + -O2 ^ NAD^ + H^O

NAD^ + ADP + R^ NADH + H^ + ATP

From the equations we can see that the NADH metabolism is very closely related to the transformation between ADP and ATP. Under anaerobic condition, NADH level (ratio of NAD/NADH) maintained sustained oscillations in glycolysis. Now we have known that, in a general way, ATP/ADP oscillation is dependent on the concentration of NAD. A tiny change in NAD concentration will alter the metabolism mode. It has been approved that NADH oscillates with a phase shift of 180° with respect to ATP oscillation. That is to say, when intracellular NADH content reaches to the minimum, ATP gets to the maximum. A further research conducted with pyruvic-acid sensor has confirmed that NAD/NADH increases with the perturbation of acetaldehyde, accordingly, the intracellular ADP/ATP ratio decreases . Therefore, it is deduced that the NAD level acts as an indicator of the intracellular ADP/ATP ratio, which represents the energy metabolism status. Reactive oxygen species (ROS) and its scavengers both affect the intracellular oscillating behaviors of NAD(H) . Correspondingly, the NAD level is able to indicate the oxidative stress of cells.

It becomes clearer and clearer from more and more biological experimental evidences that the redox state in vivo, especially the NADH level (NAD/NADH), is closely related to many important life processes, such as the regulation of longevity in yeast. Up to now, it is accepted that there are some relations between longevity and metabolism. Latest researches proved that the NADH level could decide the length of life-span in PKA-activity-decreasing mutants . It is found that the over-expression of deacetylase Sir2, which is dependent on the NADH level, will result in the 'dreariness' of Global Gene, through which the life-span is prolonged. Such phenomenon exists in all living organisms, from bacterium to mammal. 'Intake-calorie restrictively' (CR) is the commonly accepted pathway to prolong life-span. If the intake calorie decreases by 30-50%, the survival time of rat will be prolonged by about 1/3 . The fact is also ascribed to the intracellular NAD level and accordingly, the decrease of Sir2 activity and the 'dreariness' of Global Gene (Figure 4.2).

In summary, the intracellular NAD^/NADH ratio can not only regulate the redox state in vivo, but also be regarded as an indicator of metabolistic status that relates with numerous biological processes. Studying on the NADH level has been the focus of researches in fields of biology, medicine and chemistry. Especially, to dynamicly describe the real-time intracellular NADH level is a challenging subject in analytical chemistry, biochemistry and biology as well.

110 Z.H. LIU ETAL.

Age-associated diseases

CR

;

NAD

NAD: NADH ratio

I Sir2 activi^

Longevity

Figure 4.2. NADH level acts as a metabolic regulator of longevity and disease

4.3. DETERMINATION OF INTRACELLULAR NADH LEVEL

Cell, as a fundamental unit of life movement, is the minimal unit to put up the entire characteristics of living organism. All the key problems of life sciences can be traced back to cell problems. However, cells are so tiny, the compounds in cells are so infinitesimal and complicated, and chemical reactions in cells are so fleet that the cell research, especially the single living cell research is very difficult. As we know, the diameter of a cell is about lO' m, and the volume is about 10'' -10"^^L. Biochemical reactions in vivo usually occur within microseconds, and the amount of those important biologically active species in a cell is always below lO'^mol. What is more, the composition of cells is very complex. According to these facts, researchers always destroy cell membrane by hypotonic homogenization, ultrasonication, rubbing and so on to get a mixture of organelles such as nucleolus, mitochondrial, chlorophyll, endoplasmic reticulum, dictyosome, lysosome and cytoplasm. Then cellular components are obtained through differential centrifugation of the mixture for analysis. For example, Yamazaki extracted components from yeast by homogenization and centrifugation, thus the function and mechanism of NAD+/NADH ratio in glycolysis oscillation was discussed with using pyruvic-acid sensor .

In recent years, analytical apparatuses with high sensitivity, high precision and multiple functions have been utilized in cell analysis widely, and some approaches such as enzymatic analysis, fluorometric analysis, fluorescence


micro-imaging techniques, fluorescence microscopy combined with AFM or SEM, laser scanning confocal microphotographics and two-photon excitation micrographics have been appHed in the determination of intracellular NADH level.

4.3.1. Enzymatic Assays

As is known, NADH participates in more than 300 kinds of enzymatic oxidation-reduction reactions in vivo, and enzymatic analysis has many virtues such as good specificity, easy detection, soft reaction conditions, no pollution and high sensitivity. Therefore, methods based on enzymes that associated with NADH have been extensively utilized in the determination of intracellular NADH level.

Krebe and his colleagues firstly made use of the concentration of both oxidized form of dehydrogenase and reductive zymolyte to calculate the specific value of NAD/NADH . To say, lactate dehydrogenase can catalyze the transformation of pyruvic acid to lactic acid, namely the transformation from NADH to NAD. The NAD/NADH ratio can be related with the ratio of pyruvic acid/lactic acid as follow:

[NAD^]/[NADH]=[ pyruvic acid]/[ lactic acid] x K

In the above equation, the concentration of pyruvic acid and lactic acid can be gained by detecting the change of absorbency at 340nm through an enzymatic analysis based on lactate dehydrogenase, and K is the equilibrium constant from lactate dehydrogenase. Therefore, one can obtain the specific value of NAD+/NADH by calculation. Krebe et al also studied the metabolism of white rats under three different conditions, i.e. carefully fed, in hunger and ill, respectively. They found that the specific value of NAD/NADH were different in the three situations, decreasing in turn markedly.

Zhang and his colleagues ^ made use of enzymatic analysis to monitor the change of NAD/NADH so as to investigate the co-repressor's inhibiting effect on transcription factor. It was found that the inhibition was controlled by the ratio of NAD/NADH.

Enzymatic cycling assay is a very important and interesting method, which has contributed a lot to the NADH level determination. As we know, both NAD and NADH are important electron carriers in a variety of metabolic reaction *\ therefore separate measurement of the reduced and oxidized cofactors will provide useful information about metabolic, bioenergetic, and redox status of organisms. Hence methods with proper quantification are desired for the purpose of separate determination of NAD and NADH.

Experiments based on enzymatic cycling assay for quantification of NAD and NADH have been performed * ' '*. For accurate measurement of NAD and NADH in organisms, usually a proper extraction combined with quantitative

112 Z.H. LIU ETAL.

determination is required. For extraction, two types of method have been described: (1) Separate extraction method, (2) Single extraction method.

In separate extraction method, the NAD and NADH are separately extracted under different pH values. NAD is extracted in an acidic solution, in which NADH will decompose, and then NADH is extracted in an alkaline solution to decompose NAD. In contrast, in the single extraction method, both NAD and NADH are extracted in a single procedure with a neutral or mild basic solution that keeps both forms stable. NAD and NADH in the extract are distinguished by substrate specificity of an enzyme or by their difference in heat stability.

Ken proposed a novel method to measure NAD and NADH in monolayers of neuroblastoma cell line, which is based on modifying a single extraction procedure originally developed for erythrocytes and enzymatic cycling assay using a dye that absorbs in visible range. The method was simple and inexpensive and effective in determination of NAD and NADH in cell monolayers ^ .

4.3.2. Fluorometric Methods

4.3.2.1. Direct Methods

Because NADH has strong fluorescence (>-ex'=340nm, >-em 460nm) while NAD^ does not, the change of the fluorescence signal can reflect the change of the NADH level. Fluorometry is of high sensitivity and good specificity, which has been used in cell investigation for many years. It always imports electron-transfer inhibitor or proton- transfer-uncoupling reagent to adjust the specific value of NAD+/NADH. As we know, the process of electron transfer is closely connected with phosphorylation. In its state of rest, oxidizing phosphorylation is of its lowest level. Under such conditions, the electrochemical gradient of the membrane in chondriosome can restrain proton pump from moving, therefore electron transfer is also restrained. Electron-transfer inhibitor refers to substances that can interdict the electron transfer in a certain part of respiratory chain. CN", N3", carbon monoxide, rotenone, amytal and antimycin are normally used electron-transfer inhibitor. For instance, when KCN acts on chondriosome respiratory chain site, it can interdict the transfer of electrons in the respiratory chain, prevent NADH from being oxidized, strongly stop synthesis of ATP, and accelerate the process of hydrolysis. Consequently, it will reduce the specific value of NAD+/NADH. The effect of proton-transfer-uncoupling agent is to separate the process of electron transfer and the formation of ATP. It only restrains the process of ATP formation, but not the process of electron transfer. Because this kind of reagent let the electron-transfer procedure out of control, ion gradient would not be formed. It will reduce the mitochondria membrane potential, restrain the formation of ATP, promote cellular respiration and maintain the specific value of NAD+/NADH. The direct fluorometric assay can be easily expressed as the equation below:


NAD+/NADH=(NADHn,ax" NADHn,in)/( NADHctri- NADH^n)

Where NADH^ax refers to the concentration of NADH when getting the largest reductive quantity with the existence of electron-transfer inhibitor, NADHmin refers to the concentration of NADH when getting the largest oxidizing quantity and NADHctri refers to the controlled concentration of NADH. In the case, only the autofluorescence of cells is needed. In this way, one can get the specific value of NAD+/NADH, and then NADH level.

Rex and his colleagues detected the change of fluorescence from NADH in pallium by laser-induced fluorescence and found that the fluorescence intensity increased as blood vessel of brain was enlarged by vasodilation drugs ^ . When the rat put up cortex diffusible depressed representation which was the response to poison, fluorescence intensity of NADH decreased momently (within about 1 minute), afterward, it increased again (lasting about 5 to 10 minutes). This phenomenon indicates that cortex NADH level is related to excitement and metabolism of central nervous cell.

Fluorometric assay for NADH is direct and simple, however, it is always difficult to detect the fluorescence of NADH in suspended living cells because the autofluorescence of NADH is comparatively weak and the background is rather strong. Then NADH is often extracted from cells and detected out of real time for accuracy. Usually, NADH is extracted using the following methods of cell disruption: mechanical methods (high-speed bead mill, high-pressure homogenization, ultrasonication, etc.) and chemical methods (enzymatic lysis and chemical permeation).

Yeast cell is one of the perfect models for cell research. It is quite similar in many aspects with mammalian cells and is relatively ease to be operated. So many scientific findings in human cells are first investigated in yeast cells. In our recent work concerning the intracellular NADH level in an apoptotic process of yeast cells ^ , NADH is extracted with snailase lysing and SDS permeating since mechanical methods is not effective enough for yeast cells due to the layer of quite firm cell wall. The fluorescence excitation and emission spectrum of NADH in cell extracts are recorded, and as a comparison, the spectra of NADH standard solution are also recorded. As is illustrated in figure 4.3, the maximum excitation and emission wavelength of cell extracts are located at 350nm and 430nm, respectively, which are consistent with that of the NADH standard solution. No interfering fluorescence is detected within the scanned wavelength range.

114 Z.H. LIU ETAL.

250 300 350 400 450 500 550 600

wavelength(nm)

Figure 4.3. The fluorescence spectra of NADH. a, b: the excitation and emission spectrum of NADH in yeast cell extracts; a', b': the excitation and emission spectrum of NADH standard solution.

Fluorescence from naturally occurring pyridine nucleotides can be used as an indicator of cellular respiration and therefore as an intrinstic probe to study cellular metabolism. Because, as is mentioned above, NADH is fluorescent whereas NAD^ is not, the changes in the cellular redox state lead to changes in the autofluorescence signal. That is to say, the increase of NADH/NAD radio will lead to an increase in autofluorescence intensity. However, the major problem with this approach is that NADH is not a very ideal fluorophore. It has a small absorption cross-section and rather low quantum yield. Thereby it is in urgent need of developing new highly sensitive methods. We have already studied a highly sensitive and simple spectrofluorimetric method for the determination of NADH ' , which is based on its inhibitory effect on a hemoglobin catalyzed reaction. Under the optimum conditions, the degree of inhibitory effect was linear to the NADH concentration in the range of 5.0 x 10' ~2.0 X lO'^mol/L. The relative standard deviation was 3.6% for eleven determinations of 2.0 xlO'^ mol/L NADH and the detection limit was 2.0 x 10' mol/L. Further experimental results revealed that the inhibition of NADH on this system was of the competitive type.

We have also revealed the relation between NADH fluorescence intensity and the number of yeast cells in our recent work (Figure 4.4). An exact linearity ranging in 1.8 ~ 3.6 X lOVml of yeast cells was found. The linear response can be fitted to an equation as follows:


FI = (111.84 ± 15.87) + (199.24 ± 6.27) x 10'

(r = 0.9985, n = 5)

Where " C " is the cell number, " r " and " « " are the linear correlation

coefficient and the number of experiments, respectively.

900

2.0 2.5 3.0 3.5

Cells number (loVml)

4.0

Figure 4.4. Relationship between NADH fluorescence and cell number

It was further found that intracellular NADH fluorescence intensity is related to the survival rate of cells, as is shown in Figure 4.5. We can see that the NADH fluorescence intensity decreased with the increase of survival rate.

0 20 40 60 80 100

Survival rate

Figure 4.5. Relation between Fluorescence intensity and survival rate

116 Z.H. LIU ETAL.

4.3.2.2. Indirect Methods

There is another method to determine intracellular NADH level besides making use of its natural fluorescence, which is the so called indirect method.

As we know NADH is able to reduce resazurin, which is not fluorescent, to resorufm with strong fluorescence in the presence of diaphorase, and then one can measure the fluorescence of resorufm. This method was able to detect NADH with a concentration as low as lO'^mol/L, which had a sensitivity two times higher than that of direct methods. Zhou et al ^ improved this method and detected 100 pmol/L NADH by using Amplex Red which is not fluorescent. Nicola invented a sensitive NADH fluorescent probe (reagent A in figure 4.6), in which the fluorescence of the fluorophore is quenched. Through the reaction between NADH and reagent A, the fluorophore is released and its fluorescence is detected^^.

fluorophore

fluorophof^"

Figure 4.6. The reaction scheme of the NADH probe

4.3.3. Micro-Fluorescence Photometry

With micro-fluorescence photometry one can analyze cells and organs quantitatively in virtue of the high resolving power of microscope, the principle of which is the same as usual fluorescence analysis. Its strengths include high sensitivity and high particularity, and it does not subject to the influence of the distributing of fluorescent substances. The usually adopted operation is to detect total fluorescence of substances over a long enough span of spectrum. The main shortcoming of this technique is that the fluorescence is easy to decline and even to be quenched. So the experimental conditions should be fit exactly when detecting fluorescence. This method is mainly used to detect fluorescent substances in cells and organs. It is also applied in dynamic studies in vivo. Combined with fluorescent probes, substances in cells such as DNA, RNA, protein, [Ca^], and products of intracellular reactions can be quantitatively determined. In a study of glucose metabolism, Piston found a good linear relationship between NADH level in cytoplasm and mitochondria and the fluorescence intensity. Thus the NADH level in cells was determined with micro-fluorescence photometry ^'.


Micro-fluorescence photometry is especially suitable for kinetic studies of living organs and cells and unfixed slice. It will be better if it is combined with time-resolved fluorescence technology. Now this method has been used broadly to detect NADH level in the process of glucose metabolism in organs such as heart, brain, sperm and pancreas.

Images of cells can be observed with the microscope combined with computers. Thus the information of the cell structure and the inclusion content can be obtained. The application of fluorescence images in life sciences has been reviewed " ^

Combining a PMT at the port of the fluorescence microscope is the simplest and most commonly used fluorescent imaging method to get the information of the intracellular NADH. The cell metabolism can be researched with the signal given by the cell. With replacing the PMT with CCD, spatial resolution can be improved greatly. Although NADH is of ultraviolet absorption, and its fluorescence is very weak, the planar fluorescence images of NADH in various cells have been obtained with conventional fluorescence microscope ' , with which the quantitative analysis of NADH in the cell has been performed.

4.3.4. Laser Scanning Confocal Microphotographics

In laser scanning confocal microscope, there is a pinhole which is confocal to the laser focus to restrain the scattering light. When the scanner moves on the sample, the emission through the pinhole is collected to form three-dimensional images so that the longitudinal spatial resolution is improved (<l[im). This technique is widely used in the quantitative analysis of intracellular substances, the distribution of calcium ion, the statistic of optical density, and cell configuration. The resolvable size of laser scanning confocal microscope is much smaller than the size of cells. Therefore the fluorescence intensity of every pixel denotes the ion concentration. But the excite wavelength in the laser scanning confocal microphotographics is limited by light resource, which always is Ar or Ar-Kr laser. And the time resolution is not high enough. If the ultraviolet light resource is used, the temporal resolution will be improved largely ^ . Nieminen et al observed the oxidation of NADH and the generation of active oxygen in vivo with laser scanning confocal microscope ^ .

4.3.5. Two-Photon Excitation Micrographics

In two-photon excitation micrographics, the substance is excited by two photons. In this method the interference from scattering light on fluorescence detection is markedly eliminated. Since the excitation is only occurred near the focus, the light damage to cells is minimized. Meanwhile, the resolving power is notably improved. When excited by two 730nm photons, NADH can give out fluorescence with the peak located at 450nm, while NAD^ can not. With this method, one can easily probe the cell metabolism since the relative concentration of intracellular NADH and NAD^ indicates the cell metabolism

118 Z.H. LIU ETAL.

status. For example, Master ^ has studied the metabolism status in human skin through determining the fluorescence of NADH with two-photons microscope whose light resource wavelength is 730 nm.

4.4. REGULATION OF INTRACELLULAR NADH LEVEL

Because of the extensive participation of NADH in cell metabolism, the intracellular NADH level is likely affected with many factors, which are intracellular or extracellular species acting on cells. And contrarily, the intracellular NADH level may sometimes be considered as a very effective indicator of the metabolic status of cells.

4.4.1. Effect of Vitamins on Intracellular NADH Level

As we have known, plasma membrane redox systems are import electron chains, which promote electron transport in plasma membrane by catalyzing the oxidation of electron donors and the reduction of electron acceptors. Usually, NADH and NADPH are the electron donors, and ferricyanide, cytochrome c, vanadate and molecular oxygen are the extracellular electron acceptors ^ ' . It has been known that plasma membrane redox systems of eukaryotic cells are related to several vital functions which are involved in cell growth and nutrition \xi^\.dkQ^^'^^. It is found that ferricyanide reduction by the plasma membrane redox system of HL-60 cells was strongly enhanced by ascorbate or dehydroascorbate. Vitamin K is a kind of potential electron carriers. Plasma membrane redox activity in soybean hypocotyls was promoted by vitamin K3 ^ .

We have investigated the effect of Vitamin K on Intracellular NAD level in yeast cell by fluorescence spectrum in our recent work^ . And the plasma membrane redox system was found to be greatly promoted by addition of vitamin K3 and vitamin Ki. Ferricyanide reduction catalyzed by vitamin K was found to be accompanied by the decrease in intracellular NADH concentration and the increase in intracellular NAD level of yeast cells.

In our work, with the results being shown in Figure 4.7, it is found that cellular fluorescence intensity decreased gradually with the increasing of concentrations of vitamin K3 (a) and vitamin Ki (b). In other words, plasma membrane redox system stimulated by vitamin K3 was accompanied by the decrease in intracellular NADH concentration. Molecular oxygen is a natural electron acceptor in plasma membrane redox system. Electron is transferred from NADH to O2 in absent of exterior electron acceptors. Redox reaction between NADH and O2 occurred in the presence of NADH oxidase. NADH oxidase is part of electron chain. In action of NADH oxidase and other oxidase, electron is transferred to O2. However, medium between NADH and O2 is not clear presently. It is suggested that electron transport chain of plasma membrane NADH-O2 be NADH —flavoprotein-*COQ —Cyt—O2. In addition, it may include SH, ascorbic acid and free radical and so on.

DETERMINATION OF INTRACELLULAR NADH

4.4.2. The Time Course of Intracellular NADH in Yeast Apoptosis

119

Apoptosis is a physiologically programmed cell death, which plays a major role in the development and homeostasis of multicellular organisms ' . Failure to invoke appropriate cell death can result in cancer or autoimmunity, whereas increased programmed cell death can lead to degenerative processes such as immunodeficiency and neurodegenerative diseases. In recent years, many great breakthroughs in molecular mechanism of cell apoptosis have been achieved, which helps people realize the life phenomena in molecular, cellular and organic levels more and more deeply. These discoveries are or will be benefit to keep people healthy and to cure many diseases. However, the regulative mechanism of cell apoptosis in molecular level remains to be fully elucidated.

500 ^

\ (a)

\

^ , ^ ^

^^-\^^^

1 1 1 1 1 1 1 1 1 1 1

700

600

500

400

300

200

•

\

" \ - \

- X (b)

~-~-B

0.5 1.0 1.5 2.0 2.5

(10"" mol/L)

3.0 0.0 0.8 1.2

C ^ (10"'mol/L)

Figure 4.7. Effect of vitamins on activity of yeast cells (a): vitamins K3; (b): vitamins Ki, Cell amount: 4.8x10V mL

Mitochondria have been classically recognized as the most important cellular source of free radicals, as the main target for free radical regulatory and toxic actions, and as the source of signaling molecules that command cell cycle, proliferation and apoptosis. Mitochondrially regulated cell death pathway in mammalian cells is critical for morphogenesis and programmed cell death after withdrawal of survival factors. The prominent effect of mitochondria in apoptosis has been established, but the status of mitochondria in apoptosis is still attractive for lots of researchers, such as what exceptional changes occur in the enzymatic system of mitochondria respiratory chain in apoptotic process. Especially, how does the important electron transportation conductor -NADH act during the process?

120 Z.H. LIU ETAL.

It is significant to examine the time course of NADH content in the apoptotic process due to its crucial roles in reductive reactions and energetics. As an indispensable coenzyme in cell energy metabolism, NADH acts as electron transporter and presides over the transportation of H2. NADH is closely related to the oxidation phosphorylating process, from which ATP is synthesized. Recent studies show that NAD and its derivatives NADH, NADP and NADPH are involved in many assimilatory pathways as well as maintaining the intracellular redox state ^ . NAD depletion is considered as a critical factor in precipitating cell death during oxidative stress due to compromised energetics. However, as is mentioned above, the detailed time course of intracellular NADH content in apoptotic process is still unclear by far. Answers to the question will contribute to revealing the complicated molecular mechanism of apoptosis.

In our recent work ^ , the time course of intracellular NADH content is detected in the apoptotic process induced by H2O2 and ONOO". For comparison, the time course of intracellular NADH is also detected in processes of necrosis and reversible injury.

The yeast apoptosis and necrosis was confirmed with morphological analysis (figure 4.8) DNA staining and fluorescence microphotographics. As we know, AO is capable of crossing the intact cell membrane, combining with nuclear DNA and emitting yellow fluorescence, whereas EB is capable of crossing only the injured cell membrane, combining with nuclear DNA and emitting red fluorescence ^ . So their different ability of penetrating cell membrane and their distinguishable fluorescence emissions are helpful in discriminating between different pathways of cell injury. The morphological changes of the cell nuclear during the process of apoptosis and necrosis are compared in figure 4.8. It is seen that the nuclear of normal yeast cell emits bright yellow fluorescence and presents natural structural traits. In this situation, only AO has crossed the cell membrane and combined with nuclear DNA. But in the case of apoptosis, the nuclear is found to gradually rupture to fragments and to form membrane-enclosed apoptotic bodies finally. It is also found that the fluorescence emission has clearly indicated different stages of the apoptotic process. As is seen, the nuclear of the inchoate apoptotic cell emits yellow fluorescence; while that of the terminal apoptotic cell emits salmon pink fluorescence. In our opinion, the salmon pink fluorescence must have resulted from the mixing of the yellow and the red fluorescence, which implies that both AO and EB have crossed the cell membrane. As to the case of necrosis, the nuclear is found to be thoroughly disaggregated, and carmine fluorescence is observed during the whole process. Different amounts of H2O2 were taken to treat yeast cells so as to initiate the apoptotic/necrotic and reversible injury process. Concentrations of NADH are determined timely during the processes of the three different pathways of cell injury. Results are shown in figure 4.9(A). The control experiment (curve d) indicates that the NADH content in normal yeast cells keeps nearly unchanged within the period of research. From the curve b in the figure we can found the

DETERMINATION OF INTRACELLULAR NADH

I6h

121

m\

^ ^ ^ ^ ^ # ^

Figure 4.8. Morphological changes of cell nucleus during the process of apoptosis and necrosis with AO and EB dying. (See color insert section.)

interesting and significant kinetic change of intracellular NADH concentration during the whole apoptotic process. As a response to oxidative stress, the concentration of intracellular NADH gradually ascends to some degree within the early stage of the apoptotic process, which lasts for a certain time, i.e., 2h in our experiments. And then it continuously descends to a level lower than the original one. As we know that the electron transport in respiratory chain of mitochondria begins with the oxidation of NADH (transferred to NAD), hence any inhibition of respiratory chain will lead to the accumulation of intracellular NADH. On the other hand, the continuous decrease of intracellular NADH content is considered to have indicated the cell death ^ . Therefore, in our experiments, the accumulation of NADH emerged in the beginning of the apoptotic process might have indicated the temporary inhibition of mitochondrial respiratory chain by H2O2. And the followed continuous decrease of NADH content suggests that the respiratory chain has been thoroughly destroyed and meanwhile, cells go to die. The time course of intracellular NADH content in the process of necrosis (curve a) and reversible injury (curve c) have suggested other physiological responses. The most remarkable difference between the apoptotic and the necrotic process is that the NADH content in the latter case changes much more sharply. It is easy to understand that in the case of apoptosis, which is a relatively gentle way of cell death as compared with necrosis, cells have higher resistibility to oxidative stress (the death signal) so that the intracellular NADH content changes more relaxedly. Further in the case of reversible injury (curve c), the time course of intracellular NADH content is

122 Z.H. LIU ETAL.

completely different from the above two pathways of cell death. It just decreases to some degree firstly and then resumes to the original level, which means that the oxidative injury is reparable when cells are treated with a less amount of H2O2. To our thinking, the cellular antioxidizing mechanism is activated promptly in this situation and the process of electron transportation becomes quicker, which makes the concentration of intracellular NADH decrease. And then, along with time, the oxidative injury is remedied by the cell itself and the NADH concentration resumes.

t(h)

t(h)

Figure 4.9. The time courses of intracellular NADH autofluorescence in ROS-induced apoptosis, necrosis and reversible injury. (A): With H2O2 as the stimuli. Line a: necrosis; Line b: apoptosis; Line c: reversible injury; Line d: the control. (B): With ONOO as the stimuli. Line a: necrosis; Line b: apoptosis; Line c: reversible injury; Line d: the control.


In order to draw a general conclusion, parallel experiments are performed using ONOO', a more reactive and more harmful oxidative reagent, as the stimuli. As is shown in figure 4.9(B), the kinetic characteristics of intracellular NADH content are just the same to those obtained with H2O2. Hereby it is concluded that the time course of intracellular NADH content obtained in our experiments is a common event in yeast cells, which reveals the physiological response and accommodation of cells to ROS-induced oxidative injury.

4.5. REFERENCES

[I] D.E.Kelley, J.He, E.V.Menshikova, et al. Dysfunction of mitochondria in human skeletal muscle in type 2 diabetes, Diabetes 51, 2944-2950 (2002)

[2] J.G.D.Birkmayer, C.Vrecko, D.Volc, et al. Nicotinamide adenine dinucleotide (NADH)- a new therapeutic approach to Parkinson's disease. Comparison of oral and parenteral application, Acta NeurolScand. 146, 32-35 (1993)

[3] S.K.Jonas, C.Benedetto, A.Flatman, et al. Increased activity of 6-phosphogluconate dehydrogenase and glucose-6-phosphate dehydrogenase in purified cell suspensions and single cells from the uterine cervix in cervical intraepithelial neoplasia,5r. J. Cancer 66, 185-191 (1992)

[4] A.Harden, W.J.Young, Proc. Roy. Soc. London B 78, 369 (1906) [5] S.Yamazaki, K.Miki, K.Kano, et al. Mechanic Study on the Role of the NAD+/NADH Ratio in

the Glycolytic Oscillation with a Pyruvate Sensor, /. Electroanal. Chem. 516, 59-65 (2001) [6] B.M.Douglas, E.Frank, L.David, K.Hiroshi, Involvement of glutathione in the regulation of

respiratory oscillation during a continuous culture of Saccharomyces cerevisiae, Microbiology 145, 2739-2745 (1999)

[7] T.Finkel, N.J.Holbrook, Oxidants, Oxidative Stress and the Biology of Aging (Insight Review), Nature 408, 239-247 (2000)

[8] C.Mclay, M.Crowell, L.Maynard, J. Nutr. 10, 63-79 (1935) [9] D.H.Williamson, P.Lund, H.A.Krebs, The Redox State of Free Nicotinamide Adenine

Dinucleotide in The Cytoplasm and Mitochondria of Rat Liver, Biochem. J. 103, 514-527 (1967)

[10] Q.Zhang, D.W.Piston, R.H.Goodman, Regulation of corepressor function by nuclear NADH, Science 295,1895-1897 (2002)

[II] A.K. Al-Ali, Pyridine Nucleotide redox potential in erythrocytes of sandi subjects with sickle cell disease, Acta Haematol. 108,19-22 (2002)

[12] T.C.Wagner, M.D.Scott, Single extraction method for the spectrophotometric quantification of oxidized and reduced Pyridine Nucleotides in erythrocytes Anal.Biochem. Ill, 417-426 (1994)

[13] C.R.Zerez,S.J.Lee,K.Tanaka, spectrophotometric determination of oxidized and reduced Pyridine Nucleotides in erythrocytes,y4«a/.5zoc/zem.l64, 367-373 (1987)

[14] D.A.Lane, D.Nadeau, Determination of Pyridine Nucleotide contents of Cell monolayers by bioluminescence, J.Biochem.Biophys. Method 17,107-118 (1988)

[15] U.Ke, K.Hideo, Determination of oxidized and reduced nicotinamide adenine dinucleotide in cell monolayers using a single extraction procedure and a spectrophotometric assay, Anal.Biochem. 338, 131-135 (2005)

[16] A.Rex, L.Pfeifer, F.Fink, et al. Cortical NADH during pharmacological manipulations of the respiratory chain and spreading depression in vivo, /. Neurosci. Res. 57, 359-370 (1999)

[17] J. Liang, Z.H.Liu, R.X.Cai, Intracelluar NADH metabolism in H202-induced yeast cell apoptosis. Oral presentation on Xith ISLS. 2004, Tsinghua Univ.

[18] Y.H.Chen, R.X.Cai, Determination of NADH with inhibited fluorometry, Chin. J. Anal. Chem. 32,719(2004)

124 Z.H. LIU ETAL.

[19] M.Zhou, Z.Diwu, V.N.Panchuk, et al. A stable nonfluorescent derivative of resorufin for the fluorometric determination of trace hydrogen peroxide: applications in detecting the activity of phagocyte NADPH oxidase and other oxidases, Anal. Biochem. 253, 162-168 (1997)

[20] L.Nicola, M.Maidwell, R.Reza, et al. On the development of NAD(P)H-sensitive fluorescent probes, J. Chem. Soc, Perkin Trans. 1, 1541-1546 (2000)

[21] D.W.Piston, S.M.Knobel. Real-time analysis of glucose, metabolism by microscopy. Trends Endocrinol Metab. 10,413-417(1999)

[22] X.F.Wang, C.K.Florine, J.J.Lemasters, et al. J. Fluoresc. 5, 71 (1995) [23] S.Lahooti, H.K.Yueh, A.W.Neumann, Collids Surf. B. 3, 333 (1995) [24] M.Dellinger, M.Geze, R.Santus, et al . Imaging of cells by autofluorescence. A new tool in the

probing of biopharmaceutical effects at the intracellular level, Biotechnol. Appl. Biochem. 28, 25-32 (1998)

[25] T.Hama, A.Takahashi, A.Ichihara, et al. Real time in situ confocal imaging of calcium wave in the perfused whole heart of the rat. Cell Signal. 10, 331-337 (1998)

[26] A.L.Nieminen, A.M.Byme, B.Herman, et al. Mitochondrial permeability transition in hepatocytes induced by t-BuOOH: NAD(P)H and reactive oxygen species. Am. J. Physiol. Cell Physiol. 272, C1286-C1294 (1997)

[27] B.R.Masters, P.T.So, E.Gratton, et al., Multiphoton excitation microscopy of in vivo human skin. Functional and morphological optical biopsy based on three-dimensional imaging, lifetime measurements and fluorescence spectroscopy, Ann. NY Acad. Sci. 838, 58-67 (1998)

[28] M. V. Wright, T. B. Kuhn. CNS neurons express two distinct plasma membrane electron transport systems implicated in neuronal viability. J. Neurochem. 83, 655-664 (2002)

[29] H. J. Gong, K. M. Chen, G C. Chen, et al. Redox system in the plasma membranes of two ecotypes of reed (Phragmites communis Trin.) leaves from different habitats. Colloids & Surfaces B: Biointerfaces 32, 163-168 (2003)

[30] C. Kim, F. L. Crane, W. P. Faulk, et al. Purification and Characterization of a Doxorubicin-inhibited NADH-quinone (NADH-ferricyanide) Reductase from Rat Liver Plasma Membranes. J. Biol. Chem. Ill, 16441-16447 (2002)

[31] M. Belting. Heparan sulfate proteoglycan as a plasma membrane carrier. Trends in Biochemical Sciences 2%, 145-151 (2003)

[32] A. Bridge, R. Barr, and D. J. Morre. The plasma membrane NADH oxidase of soybean has vitamin K(l) hydroquinone oxidase activity, Biochim Biophys Acta. 1463, 448-458 (2000)

[33] Y.H.Chen, Z.H.Liu, R.X.Cai, Study on Effect of Vitamin K on Intracellular NAD Level in Yeast Cell by Fluorescence spectrum, paper submitted.

[34] M.J.Arends, A.H.Wyllie, Apoptosis: mechanisms and roles in pathology. Internation Review of Experimental Pathology. 32, 223-254 (1991)

[35] N.Averet, H.Aguilaniu, O.Bunoust, L.Gustafsson, Rigoulet. NADH is specifically channeled through the mitochondrial porin channel in Saccharomyces cerevisiae. M J Bioenerg Biomembr. 34, 499-506 (2002)

[36] J.Liang, Z.H.Liu, R.X.Cai, P.Shen, The time course of intracellular NADH in ROS-induced yeast apoptosis, paper submitted.

[37] D.A.Carson, S.Seto, D.B.Wasson, C.J.Carrera, DNA strand breaks, NAD metabolism, and programmed cell death. Exp. Cell Res. 164, 273-281 (1986)

PREDICTION OF THERMAL TISSUE DAMAGE USING FLUORESCENCE SPECTROSCOPY

Christopher D. Anderson, Wei-Chiang Lin, Ravi S. Chari,

5.1. INTRODUCTION

Tissue themial ablation techniques, especially radiofrequency ablation (RFA), have become important implements in the arsenal of therapies for treatment of primary and metastatic tumors of the liver. RFA can be performed percutaneously, laparoscopically, or during celiotomy as a primary therapy of an unresectable tumor. It may also be used as an adjunct to partial hepatectomy, which remains the gold standard for the treatment of hepatic tumors. The technique of RFA requires image guided placement of the needle electrode within the targeted tumor. The imaging techniques used are most often ultrasound or computed tomography (CT), but it can also be done using magnetic resonance imaging (MRI). Once activated, the RFA needle electrode causes destruction of the tumor cells via thermal conversion of the radiofrequency energy. The intrinsic electrical impedance of the tissue transforms the electrical current into heat, elevating the local tissue temperature and leading to coagulation necrosis.

Clinically, RFA is a relatively safe procedure with a complication rate reported between 5 and 13% in most series (Bilchik et al. 2001; Curley et al. 2000; Wong et al. 2001). The efficacy of RFA, however, is not clear. There are no long term data regarding disease free survival, and the largest of clinical series report local recurrence rates as high as 39% (Bilchik et al. 1999; Curley et al. 1999; de Baere et al. 2000). Most clinicians believe that the large recurrence rates are due to either initial misplacement of the electrode or inadequate ablation of the tumor. The basic principles of oncologic surgery hold that a margin of non-cancerous tissue must be removed surrounding a tumor to insure

Christopher D. Anderson and Ravi S. Chari, Department of Surgery, Vanderbilt University Medical Center, Nashville, TN 37232-4753. Wei-Chiang Lin, Department of Neuro-Engineering, Miami Children's Hospital, Miami, FL 33174.

125

126 CD. ANDERSON ET. AL.

that both the macroscopic and microscopic margins of the resection are free of tumor. With this in mind, the RFA procedure must ablate not only the tumor, but a surrounding rim of non-cancerous hepatic parenchyma in order to assure complete tumor destruction.

To estimate the completeness of ablation, there are three separate protocols recommended by the three manufacturers of available RFA systems. One system relies simply on time, another on impedance and the third on temperature as measurable variables to gauge the volume of an ablation. Impedance measurement determines completeness of ablation based on depletion of cellular energetics and high resistance to current. Achievement of impedance roll-off, however, does not imply uniform ablation throughout the distribution of the RFA electrode array. The thermo-couple based feedback uses achievement of temperature during the procedure to determine completeness. Currently, there is no system which provides an accurate measure of ablation margins during the RFA procedure. In general, assumptions about heating time and ablation power are made to estimate the size of the ablation. These deficiencies in monitoring the completeness of an ablation in real-time are well recognized (Gazelle et al. 2000). Several groups have investigated methods to improve monitoring of an ongoing ablation. Preliminary reports have been published using magnetic resonance thermal imaging, improved ultrasound algorithms to allow better thermal monitoring, and ultrasound based tissue elastography (Quesson et al. 2000; Varghese et al. 2002b; Varghese et al. 2002a). Of these, tissue elastography is the only method, other than the method reviewed here, which does not rely solely on temperature for estimation of the size of ablation.

Fluorescence spectra of tissue are dependent on physical cell characteristics such as protein conformation. Changes induced by thermal damage will alter these characteristics and produce detectable changes in these spectra. The work reviewed in this chapter is the result of our interest in using optical spectroscopic techniques to develop a real-time feedback system to detect an RFA margin.

5.2. FLUORESCENCE SPECTROSCOPY TO DETECT THERMAL TISSUE DAMAGE

The extent of thermal damage can be predicted using the theory of rate processes. According to the Arrhenius model, thermal damage can be quantified as an integral that depends on the local temperature-time history of the tissue, and the tissue dependent empirical constants: frequency factor {A) and the activation energy barrier {Ed) (Pearce and Thompsen 1995). This model can be modified based on specific experimental conditions to predict tissue thermal damage, and has been used to predict the thermal coagulation of tissue induced by various energy sources. As we reported (Buttemere et al. 2003; Buttemere et al. 2004), these theoretical thermal damage time courses can be compared to a correlate of thermal damage such as tissue fluorescence.

FLOURESCENCE SPECTROSCOPY AND THERMAL TISSUE DAMAGE 127

In these experiments, freshly harvested canine liver was cut into small slabs roughly 3mm X 3cm. These samples were then subjected to thermal damage by submersion in 50 and 70°C water baths for varying times (Buttemere et al. 2003). Samples were allowed to cool to room temperature before fluorescence spectra were obtained. Fluorescence spectra were measured from these samples using a fiberoptic based spectroscopy system with fluorescence excitation at 337nm via a nitrogen gas laser (repetition rate of 20 Hz and a pulse energy of 50 |LIJ at the tissue surface) (Buttemere et al. 2003; Buttemere et al. 2004). Spectral data were obtained using an integration time of 1 second over a spectral range of 350 to 850 nm.

The results of this in vitro study demonstrated several important findings regarding the fluorescence spectra as the hepatic tissue progressed from its non-damaged state to full coagulation. The two especially noticeable changes are (1) the peak of the fluorescence emission spectrum shifted from 490 nm in the native samples to 510 nm in the fully-coagulated ones (70 °C, 60 minutes) and (2) the primary emission peak from the fully coagulated samples is broader than that from native hepatic tissue. Other line shape changes in the spectra from coagulated tissue samples were also noted, such as the appearance of a shoulder at 610 nm in the fluorescence spectra (Buttemere et al. 2003; Buttemere et al. 2004). These changes are demonstrated and summarized in Figure 5.1.

These experimental results were then compared to the Arrhenius thermal damage prediction model for comparison. In order to track the peak shift in fluorescence spectra a ratio of intensities (F510/F480) were determined for each time point and then normalized to the F510/F480 of native liver. Generally, there was good agreement between the time courses of the spectral features selected as thermal damage correlates and the Arrhenius thermal damage prediction model predictions (Buttemere et al. 2003; Buttemere et al. 2004). The F510/F480 time courses at 50 and 70 °C are plotted in Figure 5.2. These experiments suggested that fluorescence spectroscopy should accurately predict full hepatic tissue coagulation based on spectral changes alone.

5.3. MEASUREMENTS OF FLUORESCENCE SPECTRA/TV VIVO

To measure fluorescence spectral changes in vivo, a fiberoptic based spectroscopy system was designed and built in the Biomedical Optics laboratory at Vanderbilt University. This system was designed to continuously acquire fluorescence spectra from tissue, specifically liver, undergoing RFA. Measurements were taken from very localized regions including both surface and intraparenchymal points of interest. This was accomplished by the use of a micro-interrogation probe consisting of two 200 |Lim excitation fibers, two 400 )Lim collection fibers, and one 500 |Lim collection fiber; all of which are ensheathed in a stainless steel catheter. The excitation light source was a nitrogen laser operated at a repetition rate of 20 Hz and pulse energy of 50 yJ at

128 CD. ANDERSON £7: ^L.

the tissue interrogation site. The spectrometer was equipped with an A/D (analog/digital) converter, which communicated with a computer via a universal serial bus connection. A proprietary computer software developed by the vendor of the spectrometer (OOIBase32, Ocean Optics) was used to control spectral acquisition.

- • - N o o n a l Liver -•<-Coagulated Lhf

360 400 4S0 SOO S60 eOO 6S0 TOO r&O 800

Wavelength (mn)

Figure 5.1. Representative fluorescence spectra from native (normal) liver tissue and liver tissue heated at 70 °C for 60 minutes (coagulated). The fluorescence spectra were normalized to their respective peak intensities.

Time (win.)

Figure 5.2. The experimental time courses of F510/F480 at heating temperatures of 50 and 70 °C plotted together with the model-predicted time courses. The F510/F480 experimental curve reaches a plateau (indicating complete thermal damage) within the first 5 minutes of heating at 70 °C, which closely agrees with the theoretical curve. However, the theoretical 50 °C curve predicts 75% thermal damage after 60 minutes, while the experimental 50 °C curves imply a lesser degree of tissue thermal damage.


In this in vivo study, the micro-interrogation probe, was deployed within the expected RFA zone of ablation (as conceptualized in figure 5.3) in order to acquire fluorescence spectra from intra parenchymal sites, and for our experiments, we directed the probe to within the expected RFA zone of ablation (Anderson CD et al. 2004a). Using this method, fluorescence spectra were acquired from liver tissue prior to, during and following RFA. The ablations were performed until the desired spectral changes occurred as temperature at the site of interrogation was not monitored. The spectral changes were correlated to the amount of thermal damage at the site of interrogation both grossly and histologically.

RFA Electrode Array

MIP

Liver

Figure 5.3. The placement of the micro-interrogation probe within the zone of ablation. MIP=micro interrogation probe

As the ablation progressed, noticeable changes in tissue fluorescence spectra occurred between 400 and 650 nm. Theses are demonstrated in Figure 5.4. Undamaged liver shows a strong fluorescence emission between 450 and 550 nm (Anderson CD et al. 2004a). Consistent with the in-vitro studies described above, tissue thermal damage induced several significant spectral alterations. The fluorescence intensity decreased as thermal ablation occurred and the primary emission peak (482 nm) decreased as much as 45% (Buttemere et al. 2004). A shift in the primary fluorescence emission peak from 482 to 508 nm as well as a spectral broadening was also observed in the latter part of the ablation course (Anderson CD et al. 2004a).

130 C D . ANDERSONfr.^z.

5.4. SPECTRAL CORRELATES TO THERMAL DAMAGE

This data allowed several spectral correlates for tissue thermal damage to be defined. As shown (figure 5.4), very little alteration in the fluorescence emission occurred at 610 nm during the course of RFA. This allowed tracking of the fluorescence intensity decrease by using the ratio: F610/F480, and tracking of the shift in peak fluorescence emission by using the ratio: F510/F480. F510/F480, and F610/F480 were analyzed as a function of ablation time(Anderson CD et al. 2004a; Buttemere et al. 2004). Representative plots of the time courses of these spectral correlates are shown in Figure 5.5.

Time (sec.) Waveleng^ (nm)

B I 1 y • • • " 1 • •• " v

1 ; i i 1

Waveleogth (nm)

Figure 5.4. A) The complete spectral time course for representative radio frequency ablation monitored using fluorescence spectroscopy. The asterisks denote the shift in peak fluorescence from 480 nm to 508 nm which is similar to the shift noted in vitro. B) The initial and final fluorescence spectra for the same ablation sites as in A. (cu = calibrated units)


Comparison among samples were facilitated by normalizing all data points in a given time course to the initial (i.e., native hepatic tissue) point for that data set. When the tissue surrounding the micro-interrogation probe was completely coagulated the time course of the fluorescence spectral correlates exhibited three distinct phases of change (figure 5.5)(Anderson CD et al. 2004a). Phase one spectral changes showed minimal deviation from the spectral fingerprint of native liver parenchyma. During phase two, a rapid increase in intensity ratio occurred until a plateau was reached at which point no further spectral change occurs. Phase three represents this plateau region. Independent of all other factors, phase three consistently corresponded with gross and histological evidence of complete tissue coagulation. These spectral changes in the tissue did not change when the tissue was allowed to cool (Anderson CD et al. 2004a; Buttemere et al. 2004).

These findings from these in vivo studies suggested that spectral changes induced by thermal damage are irreversible. Thus, fluorescence spectroscopy represents a method of detecting thermal damage to hepatic parenchyma. However, the correlation between the degree of thermal damage and the corresponding fluorescence spectral correlate were needed in order to evaluate the sensitivity of this method of thermal damage detection.

i 1 I

100 150

Time (sec.)

300

Figure 5.5. The F510/F480 and F610/F480 time courses for a representative ablation sites. All data points are normalized to the native liver. These spectral changes can be divided into 3 phases. Phase one (1) is a period of negligible intensity change in comparison with that of native liver tissue. Phase two (2) shows rapid increases the intensity ratios. Phase three (3) shows a plateau of intensity ratios.

132 C D. ANDERSON £r.4L.

5.5. FLUORESCENCE SPECTRA CORRELATE WITH HISTOLOGIC TISSUE DAMAGE

A partial ablation is defined as an ablation in which phase 3 spectral changes were not achieved. A series of this type of ablation was performed to compare spectral correlates with histologic tissue destruction (Anderson CD et al. 2004a). During these ablations, RFA was stopped at varying times, such as following a 15, 30, or 45% decrease in fluorescence intensity. The position of the position of the micro interrogation probe was registered with India ink and the ablation zone was excised for gross and histological study. The partial ablation experiments can be summarized by two results: 1) phase three spectral changes do not occur unless the micro-interrogation probe lies within coagulated tissue, 2) phase two spectral changes are seen when the micro-interrogation probe lies within the advancing hemorrhagic edge of the zone of ablation (Anderson CD et al. 2004a).

Because no histological grading system for hepatic thermal damage existed, we developed one based on the available literature (Anderson CD et al. 2004a). This grading system assigned a score between 1 and 6 where 6 was complete coagulation of the tissue. Using this histologic grading system, spectral data from the partial ablation experiments were correlated to a histologic thermal damage score. When plotted, the tissue thermal damage scores, regardless of phase, seem to correlate linearly with normalized F610/480 as demonstrated in figure 5.6 (Anderson CD et al. 2004a).

1.00 2.00

F 610'480

Figure 5.6. Scatter plot showing the correlation of normalized FeioAso with the total tissue thermal damage scores from all ablation sites. The correlation coefficient, R2, is 0.74.

These results suggest that fluorescence spectra correlate with histologic tissue thermal damage well enough to predict the margin of an RF ablation. However, defining the actual "margin" of a RF ablation zone is


difficult. Cell death from RFA occurs when thermal damage is severe enough to overwhelm the usual cellular repair mechanisms. As tissue is exposed to higher temperatures or prolonged heating times, tissue structural proteins undergo denaturation and conformational changes; this is thermal coagulation. Macroscopically, tissue coagulation appears as whitening of tissue (Pearce and Thompsen 1995). Microscopically, H&E staining can demonstrate coagulation changes. Thermally coagulated cells and intracellular organelles shrink and undergo characteristic conformational changes (Pearce and Thompsen 1995). In addition to coagulation, superheated water vapor is generated during heating, which leads to formation of vacuoles. As these vacuoles expand and coalesce, the walls rupture and tissue fragmentation occurs (Pearce and Thompsen 1995). These histological changes were included in the thermal damage scoring system (Anderson CD et al. 2004a). Thus, our findings demonstrate that the fluorescence spectroscopic techniques presented can accurately detect a RFA "margin" if it is defined as the achievement of tissue coagulation.

However, this does not imply that the fluorescence spectral correlates correspond to absolute cell death. There may be cells outside of the fully coagulated zone, in the advancing hemorrhagic edge of the ablation (cells in the zone 2 of spectral changes) which are dead. This region may be a better representation than full coagulation of the true RF ablation margin.

5.7. DETECTION OF ABSOLUTE CELL DEATH

To determine if fluorescence spectroscopy could predict absolute cell death, RFA was performed in swine and monitored using the same methods discussed above (Anderson et al. 2004b). Random areas of healthy liver parenchyma were selected for RFA. The micro interrogation probe was placed into the expected zone of ablation at varying orientations and distances to the RF electrode. These ablations were used for transmission electron microscopy studies. Following ablations, the animals were euthanized, and each ablation zone was excised. The specimens were then bivalved along the axis of the micro interrogation probe to expose the interrogation tract; thus spatial correlations between interrogation sites and degree of tissue thermal damage could be assessed.

Fluorescence spectra obtained from the porcine liver both in vivo and in vitro were analyzed for all wavelengths monitored during the procedure. The fluorescence emission intensity for all wavelengths of non-ablated tissue was identical in both the in vivo and in vitro samples. Peak fluorescence emission for porcine liver was determined to be 470 nm (Anderson et al. 2004b). In addition, studies were performed to verify that spectral changes similar to the canine liver occur in the porcine liver. Similar to the results discussed above, the peak fluorescence emission wavelength shifted upward with ongoing ablation. In addition the overall intensity , especially the intensity at 470 nm decreased during ablation. A plateau phase was noted consistently with full ablation (Anderson et al. 2004b). These results suggests that the findings


presented in the previous sections are not unique to canine liver and can be generalized to other animals, including human liver.

Calculation of spectral ratios as correlates hindered the ability to make immediate intra-operative assessments. Therefore, to facilitate real-time analysis, changes in absolute fluorescence intensity were utilized as the spectral correlate in this study. Because of the reproducibility demonstrated, the percentage decrease of fluorescence intensity at 470 nm (F470) was compared among the samples (Anderson et al. 2004b). This percentage was calculated by using non-ablated liver tissue from either the beginning of an experiment (for in vivo studies) or a region of non-ablated liver near the ablation zone (for in vitro studies).

Following excision of the ablation zones, biopsies were taken from four areas of each specimen: 1: normal liver, 2: hemorrhagic zone/normal liver interface, 3: hemorrhagic zone/coagulation zone interface, and 4: coagulation zone (figure 5.7). Biopsies were made using a 2mm punch biopsy and specimens were immediately fixed in 2.5% glutaraldehyde. Transmission electron microscopy (TEM) was used to evaluate each sample for evidence of cellular and subcellular injury such as chromatin clumping, nuclear membrane integrity, and cellular organelle integrity (Anderson et al. 2004b). These tissue characteristics are well accepted markers of tissue thermal damage (Pearce and Thompsen 1995). Fluorescence spectra were obtained in vitro from the exact location of the biopsy (figure 5.7). This was performed to ensure that the spectral data were obtained from the exact point to be studied using TEM.

When the biopsies were subjected to TEM, the appearance of non-ablated liver demonstrated the normal appearance of all the cellular and subcellular structures including the nucleus, nuclear membrane, endoplasmic reticulum, and mitochondria. In contrast, TEMs from the coagulation zone showed complete destruction of subcellular organelles, chromatin clumping at the nuclear margins, and extensive disruption of nuclear membranes (Anderson et al. 2004b). These findings are indicative of complete cell destruction (Pearce and Thompsen 1995). Samples taken from the coagulation zone/hemorrhagic zone interface show similar features indicative of irreversible cellular damage (figure 5.8). However, tissue at the interface of the hemorrhagic zone and normal liver showed a lesser degree of cellular damage. Cells in this zone had a lack of recognizable subcellular organelles, but there were various degrees of nuclear membrane disruption and chromatin clumping within each sample. These TEM findings are consistent with incomplete cell destruction. Hence, there may be hepatocytes at the interface between the hemorrhagic zone and normal liver which remain viable. Therefore within an ablation, absolute cellular death occurs at the interface of the coagulation zone with the hemorrhagic zone, not at the advancing edge of the hemorrhagic zone.

We correlated the fluorescence spectral measures from the coagulation/hemorrhagic zone interface biopsy sites with TEM findings (Anderson et al. 2004b). This biopsy site demonstrated 87.5% +/- 9% decrease in the F470 emission intensity as compared to non-ablated tissue. Thus, an 87.5% +/- 9% reduction in F470 emission intensity corresponds to TEM evidence of absolute cell death.


Wavelength (nm)

Figure 5.7. Representative ablation zone from liver. 1) normal liver, 2) the advancing edge of the ablation zone (hemorrhagic zone/normal liver interface), 3) hemorrhagic zone/coagulation zone interface, 4) coagulation zone. The number mark areas where biopsies and TEM sampling was performed. The corresponding fluorescence intensities for each biopsy are shown.


Figure 5.8. Transmission electron photomicrographs demonstrating A) normal hepatic parenchyma, and B) hepatic tissue biopsied from the hemorrhagic zone/coagulation zone interface following RFA. Figure A demonstrates normal subcellular organelle architecture whereas following a 87.5% decrease in fluorescence emission intensity, there is complete destruction of these subcellular organelles (figure B). In addition chromatin clumping and deterioration of the nuclear membrane occurs (not shown).

5.8. CONCLUSIONS AND FUTURE DIRECTIONS

These studies characterized the fluorescence spectra of hepatic tissue and verified the spectral changes associated with tissue thermal damage from radiofrequency ablation. The findings reviewed herein support the hypothesis that fluorescence spectroscopy can be used to assess liver tissue thermal damage. The major findings from this work are that fluorescence spectra can be measured in vivo in perfused liver tissue undergoing radio-frequency ablation and that the observed spectral alterations correlate well with histological markers of thermal damage. Moreover, fluorescence spectral data can predict absolute cell death in real-time.

Because clinical RFA was the inspiration for this study, this is a very important finding. Currently, a number of algorithms are used to predict the size (or volume) of ablation achieved. In most clinical instances, ablations are carried out based upon the maximum calculated volume of tissue capable of being ablated based on these algorithms. However, the achievement of inadequate ablation margins continues to be a major contributor to the large local recurrence rate following RFA. The real-time fluorescence spectroscopic monitoring can detect varying degrees of thermal damage and absolute cell death is a current shortcoming of all other RFA monitoring methods. These studies indicate the feasibility of developing an optical spectroscopy feedback system which can provide margin detection in real-time during thermal ablation procedures. Improved ablation margin detection could, in turn, improve upon the local tumor recurrence rate following radiofrequency ablation.

The field of optical spectroscopy, specifically fluorescence, has grown rapidly. In addition to our work reviewed in this chapter, there are many emerging studies using fluorescence spectroscopy which have clinical relevance. For example, fluorescence spectroscopy has been demonstrated by


many groups, including ours, to accurately distinguish between tissue types. Distinction between normal and malignant tissue has been demonstrated, and is beginning to have some clinical usefulness (Bogaards et al. 2002; Lin et al. 2000; Lin et al. 2001; Nijssen et al. 2002; Nordstrom et al. 2001). Also of great interest is a recent publication by Fitzgerald et al. in which fluorescence spectral changes of renal parenchyma were found to correlate with warm ischemic time of donor kidneys(Fitzgerald et al. 2004). Future work in this area holds great promise for the field of transplantation if these spectral correlates can predict post transplantation graft function.

Future work in the area of hepatic surgical oncology will concentrate on refining thermal damage assessment while also incorporating tumor margin detection and localization. In addition, expansion to other fields where RFA is an important surgical oncology tool will be untaken. Preliminary data from our group suggests that fluorescence spectroscopy can accurately detect thermal damage and irreversible cell death in renal parenchyma during RFA(Lin et al. 2004).

5.9. REFERENCES

Anderson CD, Lin WC, Buttemere CR, Washington MK, Mahadevan-Jansen A, Pierce J, Nicoud IB, Pinson CW, and Chari RS. 2004. "Real-Time Spectroscopic Assessment of Thermal Damage: Implications for Radiofrequency Ablation." J.Gastrointest.Surg. 8:660-669.(a)

Anderson,C.D., W.C.Lin, J.Beckham, A.Mahadevan-Jansen, C.R.Buttemere, J.Pierce, I.B.Nicoud, P.C.Wright, and R.S.Chari. 2004. "Fluorescence spectroscopy accurately detects irreversible cell damage during hepatic radiofrequency ablation." Surgery. 136:524-531.(b)

Bilchik,A.J., D.M.Rose, D.P.Allegra, P.J.Bostick, E.Hsueh, and D.L.Morton. 1999. "Radiofrequency ablation: a minimally invasive technique with multiple applications." Cancer J.Sci.Am. 5:356-361.

Bilchik,A.J., T.F.Wood, and D.P.Allegra. 2001. "Radiofrequency ablation of unresectable hepatic malignancies: lessons learned." Oncolgist.24-33.

Bogaards,A., M.C.Aalders, C.C.Zeyl, S.de Blok, C.Dannecker, P.Hillemanns, H.Stepp, and H.J.Sterenborg. 2002. "Localization and staging of cervical intraepithelial neoplasia using double ratio fluorescence imaging." J.Biomed.Opt. 7:215-220.

Buttemere,C., R.Chari, C.Anderson, A.Mahadevan-Jansen, and W.Lin. 2003. "Feedback control of liver thermotherapy using optical spectroscopy." 2003 Photonics West Conference Proceedings.

Buttemere,C., R.Chari, C.Anderson, M.K.Washington, A.Mahadevan-Jansen, and W.Lin. 2004. "/« Vivo Assessment of Thermal Damage in the Liver Using Optical Spectroscopy." J.Biomedical Optics. 9:1018-1027.

Curley,S.A., F.Izzo, P.Delrio, L.M.ElUs, J.Granchi, P.Vallone, F.Fiore, S.Pignata, B.Daniele, and F.Cremona. 1999. "Radiofrequency ablation of unresectable primary and metastatic hepatic malignancies: results in 123 patients." ««.5'wrg. 230:1-8.

Curley,S.A., F.Izzo, L.M.Ellis, V.J.Nicolas, and P.Vallone. 2000. "Radiofrequency ablation of hepatocellular cancer in 110 patients with cirrhosis." Ann.Surg. 232:381-391.

de Baere,T., D.Elias, C.Dromain, M.G.Din, V.Kuoch, M.Ducreux, V.Boige, N.Lassau, V.Marteau, P.Lasser, and A.Roche. 2000. "Radiofrequency ablation of 100 hepatic metastases with a mean follow-up of more than 1 yQar." AJR Am.J.Roentgenol. 175:1619-1625.

Fitzgerald,J.T., S.Demos, A.Michalopoulou, J.L.Pierce, and C.Troppmann. 2004. "Assessment of renal ischemia by optical spectroscopy." J.Surg.Res. 122:21-28.

Gazelle,G.S., S.N.Goldberg, L.Solbiati, and T.Livraghi. 2000. "Tumor ablation with radio-frequency energy." Radiology. 217:633-646.

138 C D . ANDERSON ET. AL.

Lin,W.C., S.A.Toms, M.Johnson, E.D.Jansen, and A.Mahadevan-Jansen. 2001. "In vivo brain tumor demarcation using optical spectroscopy." Photochem.Photobiol. 73:396-402.

Lin,W.C., S.A.Toms, M.Motamedi, E.D.Jansen, and A.Mahadevan-Jansen. 2000. "Brain tumor demarcation using optical spectroscopy; an in vitro study." J.Biomed.Opt. 5:214-220.

Lin,W.C., J.Beckham, D.B.Parekh, S.D.Harrel, C.D.Anderson, Chari R.S., and A.Mahadevan-Jansen. Optical spectroscopy for guiding thermotherapies of tumors. 2004. The 2004 Biomedical Optics Conference. 2004.

Nijssen,A., T.C.Bakker Schut, F.Heule, P.J.Caspers, D.P.Hayes, M.H.Neumann, and G.J.Puppels. 2002. "Discriminating basal cell carcinoma from its surrounding tissue by Raman spectroscopy." J.Invest Dermatol. 119:64-69.

Nordstrom,R.J., L.Burke, J.M.Niloff, and J.F.Myrtle. 2001. "Identification of cervical intraepithelial neoplasia (CIN) using UV-excited fluorescence and diffuse-reflectance tissue spectroscopy." Lasers Surg.Med. 29:118-127.

Pearce,J. and S.Thompsen. 1995. "Rate process analysis of thermal damage." In A.Welch and M.van Gemert, editors. Optical-thermal response of laser-irradiated tissue. Plenum Press. New York. 561-606.

Quesson,B., J.A.de Zwart, and C.T.Moonen. 2000. "Magnetic resonance temperature imaging for guidance of thermotherapy." JMagw Reson.Imaging. 12:525-533.

Varghese,T., J.A.Zagzebski, Q.Chen, U.Techavipoo, G.Frank, C.Johnson, A.Wright, and F.T.Lee, Jr. 2002a. "Ultrasound monitoring of temperature change during radiofrequency ablation: preliminary in-vivo results." Ultrasound Med.Biol. 28:321-329.

Varghese,T., J.A.Zagzebski, and F.T.Lee, Jr. 2002b. "Elastographic imaging of thermal lesions in the liver in vivo following radiofrequency ablation: preliminary results." Ultrasound Med Biol. 2S:\461-\413.

Wong,LH., W.Yeo, T.Leung, W.Y.Lau, and P.J.Johnson. 2001. "Circulating tumor cell mRNAs in peripheral blood from hepatocellular carcinoma patients under radiotherapy, surgical resection or chemotherapy: a quantitative evaluation." Cancer Lett. 167:183-191.

DETECTION OF BIOLOGICAL THIOLS

Jorge O. Escobedo, Oleksandr Rusin, Weihua Wang, Onur Alpttirk, Kyu Kwang Kim, Xiangyang Xu and Robert M. Strongin*

6.1. INTRODUCTION

Naturally occurring thiols exhibit a variety of structures and important physiological properties. The amino acid homocysteine (Hey, 1) has attracted significant recent attention. At elevated levels 1 has been proposed as an important potential risk factor for Alzheimer's and cardiovascular diseases.^ However, it is not clear why 1 (Figure 6.1) is considered a risk factor in many serious diseases, while 2, 3 and related thiols are typically viewed as non-threatening and often associated with antioxidant activity. Our work is aimed at answering this and related questions while generating useful materials and methods for detecting 1-5.

A need exists for improved methods for biological thiol detection.^ Thiols are readily prone to oxidation. Many have similar structures and are typically colorless and non-fluorescent in the visible region. Thiol derivatizations are typically based on chromophores or fluorophores that can be non-selective and/or unstable.^ Current widely used detection methods include chromatographic separations or immunoassays.

N H V ^ Y ^ S H

o HO'

°r^°> o JX .nA^NH. U y - - O' NH O Q

,A^^SH OH OyJ HO-S^SH ^° X KILI <~»i I Til 1 ^ 1

NH HO^

HS-NH2 OH NH2 SH

L-homocysteine glutathione L-cysteine N-acetylcysteine penicillamine 1 2 3 4 5

Figure 6.1. Selected bioactive thiols.

Department of Chemistry, Louisiana State University, Baton Rouge, LA 70803 USA.

139

140 J. O. ESCOBEDO ETAL.

We have achieved the selective, non-chromatographic colorimetric and fluorimetric detection of cysteine and homocysteine in the presence of each other and also in biological media, despite the similarity of their structures. We use simple organic dyes. We do not need immunogenic reagents.^ In addition, new HPLC postcolum detection methods are under development which potentially overcome some of the drawbacks in current technology. A longer term goal of these efforts is to make multianalyte disease-specific assays. Until the roles of homocysteine and other specific thiols in disease are more clearly understood, the creation of disease-specific methods remains future work.

H2N^H-S

Hcys*

A

0

0 0A/NH2

1 Protein

B

NH2 1

Protein>^^ ^ / v . ^"^^^^^^c^ ^ 1

s-s Y^^oo NH2

Proteinv„ „ " ^ ^ ^ ^ ^ ^ 0

s-s ^ ^ coo^ • 1

c Figure 6.2. We are interested in the following: how might homocysteine chemistry fundamentally

differ from that of other biothiols? Can its unique chemistry serve as the basis for selective

spectroscopic detection methods? Can its fundamental chemistry help explain its role in disease?

Chemical hypotheses based on free (A) and protein homocysteinylated materials (B,C):

(A) The geometrically-favored intramolecular alpha-H abstraction coupled electron transfer by Hey

thiyl to afford an alpha amino-carbon-centered radical. The same process in Cys and GSH proceeds

via less readily formed 4- and 9-membered ring transition states, respectively. This analysis was

proposed initially by Zhao et al and recently shown by us to promote highly selective Hcy-induced

colorimetric and fluorescence signals. This allows us to selectively detect Hey over other thiols'*''

and additionally to propose new studies of Hcy's biochemical roles {vide infra).

(B) Proposed stabilization of the S-centered radical cation of a disulfide-conjugated Hey. This

radical cation has been postulated by Armstrong in the case of homocystine. Its favored formation

compared to the analogous radical cation of cystine was reported in 1981. Armstrong noted that the

amino lone pairs would stabilize the cation; more recently, Butterworth presented strong evidence

that the analogous process occurs in C-terminal methionine residues, stabilized (and promoted) by

the carboxylate anion. This latter Met-derived species is strongly implicated in Alzheimer's and

lipid peroxidation. Hcy-conjugated proteins have not yet been studied for S-centered radical

formation, to the best of our knowledge.

(C) Armstrong also proposed that perthiyl (RS-S*) radicals may form more readily from

homocystine than cystine and related disulfides. Hcy-conjugated proteins have not yet been studied

perthiyl radical formation, to the best of our knowledge.

We thus propose that structures/processes depicted in B and C may cause free radical damage to

peptide backbones. Importantly, Hey is found physiologically as mainly the protein bound disulfide

cysteine residue conjugate.

The processes depicted in B and C are assisted by the amino and/or the carboxylate moieties in Hey.

Protein-bound Hey is commonly disulfide-linked. It would thus not be necessary (as in the case for

methionine) to be specifically bound at a terminus. This is because the amino and carboxylate are

free to assist in these processes in a Cys residue-Hcy disulfide.

DETECTION OF BIOLOGICAL THIOLS 141

6.2. HOMOCYSTEINE METABOLISM

The universal methylating agent S-adenosylmethionine (SAM) is synthesized from methionine and ATP (Figure 6.3). SAM is essential for one carbon metabolism. Demethylation of SAM produces S-adenosyl homocysteine (SAH). This reaction is followed by the enzymatic hydrolysis of SAH by S-adenosyl homocysteine hydrolyase (SAHH) to afford adenosine and Hey. A transsulfuration pathway leading from Hey to Cys begins at this point.

Hey reacts with serine via Cystathionine Beta-Synthase (CBS), the vitamin B6-dependent enzyme, to afford cystathionine. Cystathionine reacts to form cysteine, a source of glutathione, sulfate and sulfite.^

Homocysteine may also be methylated, released into the extracellular medium or deaminated. Folate-dependent Hey methylation to methionine can be promoted by methionine synthase or via betaine homocysteine methylase.^

When Hey metabolism is disrupted, export of Hey from the cellular to the extracelluar medium becomes imbalanced. Lowered Hey cellular levels produce elevated export levels, resulting in more Hey in plasma and urine. Higher Hey levels in plasma and urine are thus directly related to lower methionine synthase activity and folate or vitamin Bu deficiency. Hyperhomocysteinemia is the condition where plasma Hey concentration exceeds 12-15 fiM. It has thus been proposed that vitamin or folate therapy may be useful for hyperhomocysteinemia-related disorders.

Hey is found in several forms after being released into plasma. The sum of all these forms is the plasma total homocysteine level. Oxidation to disulfides in plasma is coupled to O2 reduction, leading to oxidative stress. Reactive oxygen species (ROS) levels can be diminished by peroxidases. Hyperhomocysteinemia appears to inhibit the expression of peroxidases.^

Nitric oxide (NO) released by endothelial cells can react with Hey to furnish S-nitrosohomocysteine (SNOHO), a strong antiplatelet and vasodilator agent. The consequence of nitrosylation is the repression of peroxide production and therefore inhibition of ROS formation.^ When present at hyperhomocysteinemic levels. Hey is not significantly affected by this mechanism.

Low-density lipoprotein oxidized by ROS suppresses endothelial nitric oxide synthase expression.^ Hey has been cited as an agent involved in lowering NO availability upon its nitrosylation.^ NO is a neurotransmitter and involved in muscle relaxation and microphage cytotoxicity.^ Importantly, Hey impairs endothelial cell function in the absence of NO. Although the mechanism is not perfectly understood, it is believed that the direct action of homocysteine on endothelial cells could either involve enhanced oxidative stress or result from a direct effect of the oxidation products of homocysteine.^

The impairment of endothelial cells in hyperhomocysteinemics is believed to be an origin of cardiovascular diseases. It is believed that Hey switches


endothelial cell phenotype from anticoagulant to procoagulant. Some studies show that homocysteine-mediated cardiovascular risk can be as strong as that of hyperlipidemia. ' ^"^^ It has been reported that high homocysteine levels were detected in up to 20 % of people suffering from heart disease.

Based on the effect of homocysteine on endothelial cells, studies have potentially linked hyperhomocysteinemia to Raynaud's syndrome. Since blood vessels carry oxygen to the brain and heart, oxidative stress generated by hyperhomocysteinemia may cause brain damage and Alzheimer's disease.

The overexpression of glutathione peroxidases can occur in Alzheimer patients, linking the disease to oxidative stress in the brain. In addition, elevated levels of plasma homocysteine have been detected under the same conditions. Further evidence for the role of oxidative stress is that antioxidant supplement delays the Alzheimer's-related complications. Increased risks of birth defects, ^"^^ and renal failure^^ are other diseases also related to hyperhomocysteinemia.

Hyperhomocysteinemia and associated disorders may be treatable by vitamin therapy. In fact, diagnosis may help prevent neural tube defect pregnancies, ischemic heart disease and strokes, and possibly colon cancer. The risk for heart disease can be reduced up to 40 %. ^ Again, nutritional folate deficiency is often attributed to hyperhomocysteinemia. Folic acid supplementation is recommended against many of these and related conditions.

Triphosphate

CH-CH2-S-O-C-CHNH2 — ^ H2N-CHC-OH — • ^ Glutathione

NH2 CH2 CH2

OH SH Cystathionine Cysteine

Figure 6.3. Some of the pathways involved in homocysteine metabolism.


6.3. NEW PERSPECTIVES ON HOMOCYSTEINE'S ROLE IN DISEASE

After many years of study, it is still not yet known if homocysteine causes cardiovascular and other disease, is a consequence of them, or is just a biomarker.^^ Our work addresses elucidating unique chemistry of homocysteine (Figure 6.2) . We plan to build upon our recent finding that proves homocysteine forms an alpha amino-carbon-centered radical at room temperature and neutral pH more readily compared to all other amino acids tested to date." ^

The chemistry of biological thiols is complex. This has prompted investigators to assume that homocysteine at elevated levels causes oxidative stress.^^ Jacobsen has pointed out that cysteine, however, undergoes similar chemistry and is present at higher (20-30 fold) concentrations than homocysteine but is not often associated with oxidative damage." ^ Jacobsen has proposed the "molecular target hypothesis," suggesting that the study of Hey should focus on elucidating its interactions with specific biomolecules. Hey promotes detachment-mediated programmed cell death, and contributes to the development of atherosclerosis in hyperhomocysteinemia. Hey decreases the expression of a wide range of antioxidant enzymes.^^ Hey inhibits glutathione peroxidase activity. "^ Homocysteine disrupts folding in the endoplasmic reticulum.^^ Protein homocysteinylation is now under intensive study. Conjugation to bovine serum albumin (BSA), for instance, was shown to occur via a thiolate anion and disulfide exchange reactions.^^

Most (ca. 70 %) plasma homocysteine is protein-bound. Armstrong (Figure 6.2, B and C) has stated that homocystine (Hey disulfide) may, for example, be more prone to perthiyl radical formation (due to favorable, specific homocystine pj a effects) or radical cation formation at sulfur (due to homocystine-specific favorable geometry allowing for stabilization by the amino lone pairs).^^ This allows us to propose homocystine, mixed Hcy-disulfide or Hcy-protein conjugates as potential specific sites of free radical damage induction (Figure 6.2).

Earlier research involving methionine residues as sites of free radical damage initiation may be relevant to Hey and disease. Posttranslational modifications of proteins are known to be initiated by disulfide radical cations.^^ Additionally, evidence exists showing that the oxidation of methionine to the radical cation may result in amyloid and prion-induced disease with backbone scission, aggregate formation and/or lipid peroxidation.^^ Recently, new studies involving the "apoE" knockout mouse model of atherosclerosis were reported. The aopE KG mouse develops atherosclerosis spontaneously. When the animals are on a high methionine diet (low folate diet, or both), the progression of atherosclerosis accelerates dramatically. Thus, the apoE KG mouse with hyperhomocysteinemia is highly prone to developing atherosclerosis. These findings have greatly bolstered the field of Hey research. ^


There is growing interest in studying metabolites of hyperhomocystei nemia. ^ This prompts us to extend our prior work in characterizing the products of Hey carbon-centered-radical formation."^^ It is well-known that such radicals form NH3 and several other radical termination and disproportionation products.^^ Due to the ready ability of Hey to form carbon-centered-radicals,"^^ we propose that the formation of NH3 and other toxins may play a role in Hcy-related disease. For instance, hyperammonemia is associated with Alzheimer's: however, no link between Hey and ammonia formation has yet been studied. This may be of particular importance in homocystinuric patients. They often have tHcy (total plasma) levels approaching 500 jaM. Many homocystinurics, if left untreated, have mental retardation, often severe.^^ The study of the extent of ammonia formation and other byproducts is planned. The thorough identification of products and their yields will be studied via the treatment of Hey with a variety of common oxidants at room temperature in neutral media.

6.4. OVERVIEW OF KNOWN METHODS FOR BIOLOGICAL THIOL DETECTION

There are numerous procedures for detecting thiols, which mainly include chromatographic separations, immuno- and enzymatic assays, electrochemical, mass spectrometric and flow injection technology. Electrochemical detection is complicated by interference from oxidizable impurities. '* Electrochemical detection of thiols by CE is hampered by the need for precision electrode alignment and isolation of the detector from the separation voltage. Amperometric post-column detection of cysteine and homocysteine also can suffer from low selectivity and high background current as cysteine exhibits irreversible oxidation requiring positive overpotential.^^ Small volumes in the separation capillaries in CE require that the detector be placed in-line to minimize line-broadening. Good sensitivity often requires dual electrode configurations. The stability of the detection cell components is another concern, depending on the type of analysis. Mercury and mercury amalgam electrodes have been widely used for thiols but have severe limitations including toxicity and poor stability.^^ Chemically modified electrodes require complex preparation, can exhibit poor stability and need controlled working conditions.

o o

,1'-thiocarl NPM •••"••-' d i imidazole

BTHZC CH3

mBrB 1,1'-thiocarbonyl

5-iodoacetamidofluorescein

Figure 6.4. Representative thiol derivatizing agents.


Fluorescence polarization immunoassays (FPIA) and enzyme immunoassays (EIA) are very useful but have shown some (inter-laboratory) imprecision (vide infra).^^ Enzymes are unstable (low shelf life compared to common organic reagents) and relatively expensive, making enzyme-based assays less attractive in spite of high specificity.

Radioimmunoassays involve toxic substances. STE (Substrate-Trapping-Enzyme) technology necessitates a batch chromatography step and exhibits low precision.^^ Mass spectrometry coupled to HPLC requires complex and expensive equipment. GC-MS also uses cumbersome equipment and requires tedious procedures not suitable for routine diagnostic applications. Gas chromatography-electron capture detection and flame photometry detection require tedious sample preparations and/or high operating temperatures.^^ Trap & Release Membrane Introduction Mass Spectrometry (T&R MIMS) requires time-consuming derivatizations and sophisticated instrumentation rendering it unsuitable for routine analysis."^^

The determination of specific thiols is often carried out in conjunction with HPLC separations.^ Thiol derivatizing agents often contain electrophilic alkylating groups for reaction with sufhydril moieties. They include iodoacetamides/^ maleimides,"^^ and monobromobimanes (mBrB)." ^ They are non-selective among the thiols. Other interferences are of concern as well. For instance, iodoacetamides can react with histidine, tyrosine and methionine."^^ Other reagents such as l,r-thiocarbonyl diimidazole can derivatize cysteine and penicillamine; however, derivatives can only be detected at short wavelength.' ' Thiol derivatization conditions can lead to several other problems. Excess derivatization agents must often be removed from the reaction mixture. In addition, conditions can be time consuming and complex. For instance, the reaction with OPA (orthophthalaldehyde) is pH-dependent. No adduct formation takes place at pH less than pH 9." ^ In some cases, the derivatives are prone to unwanted further reactions. For instance, the products of isothiocyanates and succinimidyl esters with biological thiols have limited stabilities and undergo further reactions with neighboring amines to afford thioureas. Maleimide-based derivatization agents have similar problems. Although maleimides are inert to the interferences of iodoacetamides, amines have been reported to crosslink the derivatized products.

A related observation has been reported concerning the detection of penicillamine via NBD-F." ^ The possible instability of NBD-thiols leading to S-N migrations has been proposed as an explanation (Figure 6.5). NBD-F has been used successfully via precolumn derivatization; however, the sample labeling/pretreatment process takes 2 h." ^

Terminal cysteine residues converted to S-carbamoylmethyl-cysteine during the standard alkylation step with iodoacetamide afforded new insight into the difficulties of some current labeling procedures. Cyclization produced a ring structure that corresponded to a loss of 17 Da in the mass spectrum, by loss of ammonia. This modification was reported."^^ Other findings showed that every peptide known to have a cysteine at the iV-terminus was modified.


affording a lowering in m/z value of 17 Da. About 50% of the peptides concerned are lost, causing a large decrease in mass spectrometer sensitivity."^^

Some thiol-chromophore/fluorophore derivatives also are sensitive to light and hydrolysis. The OPA-Hcy adduct is stable only in dark. ^ Monobromobimane is well-known to produce fluorescent hydrolysis products.^^ When thiols are derivatized with certain maleimides, hydrolysis peaks are encountered at the beginning and at the end of chromatographic elution.^^ Hexaiodoplatinate, on the other hand, produces no hydrolysis products.^^ However, hexaiodoplatine exhibits a broad reactivity: thioethers, thiazolidines and ascorbic acids are among the reported interferences. Some derivatization agents themselves are prone to instability. lodiacetamides"^^ are unstable to light. In addition, monobromobimane is known to be photosensitive and unstable in water. " The instability of certain maleimides in aqueous conditions necessitates the use of water-miscible organic cosolvents.^^

The most widely used universal thiol derivatizing agent in clinical studies is monobromobimane.^*' In addition to problems with fluorescent hydrolysis byproducts formed during pre-column derivatization (vide supra), there have been older reports describing the need to remove excess monobromobimane (e.g., by Sepharose-SH treatment prior to HPLC analyses).^^ Researchers have also noted that levels of fluorescent impurities in monobromobimane vary significantly from batch to batch.^^ Monobromobimane is stable in water-acetonitrile but at temperatures < -20 °C . Solutions must be replaced after a few days.^^

Thiol and sulfide quantitation kits are available. The procedure necessitates an enzymatic reaction to release the thiols followed by their determination by Ellman's reagent. However, enzymes are expensive and fragile.

P + HS NH2

N02

S-N migration

, N ^ ^ NH2

NO2

Figure 6.5. NBD-F adduct S-N migrations.

Refsum, Ueland and co-workers, as well as several others, have made pioneering advances in developing HPLC assays towards homocysteine and biological thiol analysis. These methods are used in large-scale population studies which have been ongoing for two decades. Refsum and Ueland addressed the issues of removing excess monobromobimane and other


(unknown) fluorescent impurities prior to HPLC analyses in 1989. ^ They used a chromatographic "heart-cut," with a column switching device and two-solvent delivery system. They later^^ ingeniously devised an automated system (which is still used today^^) requiring less tedium, instrumental manipulation and wear as compared to their 1989 method. The improved method requires very precise pH control. It is based on the fact that the retention times of the impurity peaks due to monobromobimane are pH-sensitive. At pH 3.50, homocysteine and cysteine peaks are overlapped by an unknown fluorescent impurity peak (apparently a common monobromobimane reaction hydrolysis product). At pH 3.80, homocysteine and glutathione peaks overlap. At pH 3.65, the homocysteine peak is resolved; however, nitric acid is added to the mobile phase to attain separation of homocysteine from an unknown latereluting peak. Elevated (800 JLIM) amounts of monobromobimane and added EDTA must be used in order to obtain maximal fluorescence yields of cysteine and cysteinylglycine. Dithioethryrytol is added to increase the fluorescence of all thiols.^^ Blood is collected and centrifuged at 2000 x g at 0-2 °C. An autosampler is filled with 60 samples (20 h analysis tune) and the derivatization reagents. The vials and most reagent reservoirs are stored in chilled racks at 0 °C. The sample processor collects the sample and reagents into a vial where derivatization is carried out. The monobromobimane derivatization is quenched by addition of glacial acetic acid after 3 min. ^ As stated by Refsum and Ueland. et al. in their comprehensive 2004 review (which is co-authored by a group of the world's leading experts in this field)^^ many analytical methods afford results that do not fulfill baseline bias and imprecision criteria. Interlaboratorv. inter and intrapatient errors in homocysteine and biological thiol assays persist. Specifically: "the different t(total) Hey measurements give comparable results, but the variation among methods and laboratories are considerable. Ideally, from the known biological inter- and intraindividual variation in tHcy, the bias should be < 10 %. and the imprecision no higher than 5 %. but many methods do not fulfill these criteria."^^ They state that there is a significant need for standardization, despite recent advances.^^ While HPLC monitoring with a universal thiol derivatizing reagent affords simultaneous monitoring of thiols, this technique is "not suited for many labs," according to Refsum et al. It requires "skilled staff and (is) labor intensive, and throughput may be low."^^ The commercial immunoassays (e.g., the fluorescence polarization immunoassay run on Abbott's Imx and AxSYM platforms^^) are easier to use than HPLC but can only monitor one analvte and utilize relatively fragile biological materials.

6.5. DETECTION OF CYSTEINE AND HOMOCYSTEINE

We have reported prior progress towards the colorimetric and fluorimetric detection of mono- and oligosaccharides. Our studies featured new functionalized xanthenes which we found arose in situ from ring-opened


450 500

Wavelength (nm)

Figure 6.6. UV-Vis spectra of 6 (2.5 x 10" M) and cysteine (4 x 10' M - 8 x l o ' M) in 0.1 M carbonate butYer pH 9.5 at room temperature. Each spectrum was acquired after 5 min. As the concentration of cysteine increases, a red shitY from 480 nm to 505 nm is observed. Reprinted with permission from J. Am. Chem. Soc, 126(2), 438-439 (2004). Copyright 2004 American Chemical Society.

resorcinarenes and related materials. Our interest in biological thiols arose initially from the reports that cysteine interferes with known sialic acid (an important saccharide cell-surface residue) determinations.^"* Our studies of the possible interference of cysteine with xanthene dye-based detection of sugars led us to find that our sugar detection methods did not suffer interference from cysteine.

During the course of our saccharide sensing work, we discovered novel methods for Hey and Cys detection. Upon addition of cysteine or homocysteine (1.0 X 10" M) to a solution of 6 previously used as a synthetic intermediate for a zinc sensor^^ (1.0 x 10' M, H2O, pH 9.5), a solution color change from bright yellow to brownish-orange is observed. Similar color changes are observed on Ci8-bonded silica. UV-Vis absorbance changes of cysteine-6 solutions, readily monitored in the 10" -10" M cysteine concentration range, exhibit a 25 nm red shift (Figure 6.6). Addition of cysteine to solutions of 6 results in monitorable fluorescence quenching (Figure 6.7). The reaction leads to thiazolidine (Cys 7a, Hey 7b heterocycle, not shown) adducts. Their structures have been unambiguously confirmed by H NMR and MS. UV-Vis spectra of solutions containing 6 and other common thiols (L-methionine, mercaptoethanol, glutathione), other amino acids (L-glutamine, L-serine, glycine, L-glutamic acid), and amines (D-glucosamine hydrochloride and CH3NH2 (8 x 10'" M, pH 9.5) confirm selectivity of 6 for cysteine and homocysteine. At most a 15 % change in absorbance at 480 nm is observed in response to the aforementioned analytes, even when they are applied in 10-fold molar excess relative to cysteine."*^


c

0) o c o w o o 3

450 500 550 600

Wavelength (nm)

650

Figure 6.7. Fluorescence emission spectra of solutions of 6 (1.0 x 10^ M) and cysteine excited at

460 nm. Inset: FO represents the maximum fluorescence intensity in the absence of analyte and F

represents the corresponding intensity in the presence of analyte. Reprinted with permission from J.

Am. Chem. Soc, 126(2), 438-439 (2004). Copyright 2004 American Chemical Society.

Solutions containing 6 and bovine serum albumin or urease also exhibit only small absorbance (less than 15 %) decreases and no wavelength shifts. Compound 6 can be used to readily detect cysteine and homocysteine in the range of their physiological levels (healthy plasma total homocysteine concentrations are ca. less than 12-15 |aM; cysteine concentrations are typically 20-30 times that of homocysteine). Again, interference from amines, amino acids and certain thiols and proteins is minimal. Importantly, we have demonstrated that both Cys and Hey can be detected in human blood plasma within range of their physiological levels." ^

6.6. HIGHLY SELECTIVE DETECTION OF CYSTEINE AND SITE-SPECIFIC PEPTIDE LABELING

Respective solutions of 6 containing identical concentrations of Cys and Hey exhibit similar spectrophotometric changes."^^ However, (Figure 6.8), we have found that 6 can be used in combination with other aldehydes such as 4-(dimethylamino)-cinnamaldehyde (DMA) to afford selective detection of cysteine. To the best of our knowledge this is the first example of such a high degree of direct Cys selectivity over other biological thiols, without using any separations or biological materials. Since the DMA signal is at 400 nm and responsive only to Cys, one may also determine Hey indirectly by differences.


1.2 n

1.1 -

0.9

0.8

0.7

0.6 -

n ^ -

i

0.0000

•

I

0.0005

•

•

E

•

0.0010

i

•

• L-Cys

• DL-Hcy

•

#

0.0015

Figure 6.8. UV-Vis titration data from mixtures of 4-(dimethylamino)-cinnamaldehyde (DMA, 3 x 10^ M) and L-cysteine and homocysteine in 0.1 M carbonate buffer (pH 9.5) monitored at 400 nm at room temperature. Concentrations of the amino acids were increased from 0 to 1.7 x 10' M. The error bars are based on three runs. This shows that DMA affords reproducible selectivity for Cys.

using non-selective absorbance changes at longer wavelengths promoted by both Cys and Hey in the presence, for example, of 6.

Increasingly more electrophilic a,P-unsaturated aldehydes such as cinnamaldehyde and 4-nitrocinnamaldehyde afford successively poorer selectivity for Cys over Hey. The electron donating dimethylamino group of DMA renders the dye less electrophilic, resulting in enhanced Cys selectivity. Presumably, the enhaced reactivity towards Cys is a result of more facile 5- vs. 6-membered ring (as expected for Hey) thiazolidine formation.

There is current interest in the site-specific labeling of peptides and proteins. One approach involves the fluorescent tagging of existing or pre-engineered A/ -terminal cysteine residues.^^ The fluorophores used include fluorescein or NIR dyes with appended thioester groups. Pretreatment with O-methylhydroxylamine is needed to free the cysteine residues that exist as carbonyl (thiazolidine) adducts.^^ The reactions may require the use of GdmCl to proceed effectively; however, this impacts protein folding. These ligations also typically require thiol co-factors such as thiophenol or benzylmercaptan. They are irreversible. They are subject to thioester hydrolytic instability and side-reactions.^^ Thus, our aldehyde-appended-dye techniques should find use in forming reversible and strong, covalent site-specific fluorescent labels. Figure 6.9 exhibits this approach towards the selective and facile labeling/detection with simple dipeptide models.


320 370 420 470

Wavelength (nm)

520

Figure 6.9. UV-Vis spectra of 4-(dimethylamino)-cinnamaldehyde (left, 3 x 10 M, 400 nm) and 6 (right, 3 X 10' M, 480-500 nm) with varying concentrations of the dipeptide cys-gly (4 x 10" M - 8 X 10^ M) in 0.1 M carbonate buffer pH 9.5 at room temperature. Each spectrum was acquired after 5 min. Spectrophotometric changes were insignificant for gly-gly (at most < 3 % absorbance changes), as expected, under the same conditions, for each of the two dye solutions.

6.7. HIGHLY SELECTIVE DETECTION OF HOMOCYSTEINE

Fluorone black (8) is a commercially available xanthene dye. Upon addition of 1 equiv Hey (1.0 x 10' M final cone.) to a solution of 70 % MeOH:H20 (MeOH is used for solubility) at pH 7.3 and containing 8 (1.0 x 10" ^ M), a change (increase) in absorbance at 510 nm is observed. A similar, but smaller absorbance change is observed for Cys, GSH and «-propanethiol. Other amino acids (glycine, alanine, serine, methionine, glutamine, lysine, arginine, threonine) or the disulfide homocystine do not promote a detectable absorbance change.' '

We investigated the effect of PPha on this process, initially in order to mimic the conditions used to prepare biological samples for analysis via disulfide reduction (bound thiol liberation). When PPhs (45 equiv to dye) is present in the dye solution, an absorbance change only occurs in the presence of homocysteine. No change is observed for Cys or any other biological thiols. If a 30-fold molar excess of Cys (to Hey) is added to solution of Hey (at levels approximating both the homocysteine risk level in plasma and the proportion of Cys to Hey), 8 (1.0 x 10" M) and PPhs (45 equiv), no change in the


IflOOOOO -T - " -

^ 1600000

c 1400000

Ic 1200000

g 1000000

S 800000

g 600000

0 400000

E 200000

0 1

480

< % ^ ^ ® / \ - » - 8 + Hcy

/ / \ \ —*— 8 + Cys

530 580 630

Wavelength (nm)

CO

•e < , 0.65 -

c

1 O 0.6 -

<

C

*,^^^

) 2 4 6 8

Concentration ( lO'^M)

• Hey

• Cys

10 12

0.00.

-0.40

O.OE+00 5.0E-07 1.0E-06. 1.5E-06 2.0E-06 2.5E-06 3.0E-06

[Hey] (Molar)

Figure 6.10. Upper: Fluorescence emission spectrum of 8 (1.0 x 10^ M) in the presence of PPhs

(4.5 X 10" M) and Hey or Cys (1.0 x 10" M). Middle: Concentration vs. absorbance dependency

for Hey and Cys (10^ M) in the presence of 8 (1.0 x 10^ M) and PPhs (45 equiv) showing Hey

selectivity. Bottom: Fluorescence emission intensity at 550 nm of solutions of 8 (1.0 x 10" M) in

the presence of PPha (4.5 x 10' M) and varying concentrations of 1 (0 3.0 f M). The excitation

wavelength was set to 510 nm. Plot is of I/Io (normal Stem-Volmer plot). The slope is negative

instead of positive due to fluorescence enhancement instead of quenching. These three experiments

have not yet been optimized to date. Reprinted with permission from J. Am. Chem. Soc, 126(11),

3400-3401 (2004). Copyright 2004 American Chemical Society.


spectrum compared to the original Hcy-8 solution is observed. Concentration-dependent dye absorbance increases are thus observed only for Hey under these conditions. Importantly, we have successfully used a standard addition technique to determine Hey in a commercial sample of human blood plasma. The method was validated by standard statistical protocols and afforded an excellent standard recovery of 102.9 %.'^^

Figure 6.10 shows the selectivity evidenced by the UV-Vis data is confirmed via fluorescence spectroscopy. An increase in fluorescence emission is observed only for PPhs and Hcy-containing solutions of 8. Cys does not affect the fluorescence emission observed in solutions of PPha and 8 but does promote an intensity increase in the absence of PPha.

Solutions of fluorescein, unlike 8, do not exhibit spectral changes in response to Cys or Hey under these conditions. We propose that relatively facile semiquinone radical formation in fluorone black (8) is responsible for the signal changes observed. Electron transfer raises the pK^ of the dye, as the observed spectral changes correspond to the absorption increase of the dye which occurs upon added base.

The role of PPha in heightening Hey selectivity is due to the fact that it serves as a competing reducing agent to the thiol-derived reducing alpha amino-carbon-radicals (Figure 6.2A, and vide infra). Beyond a threshold level of at least 100 equiv PPhs the Hcy-promoted absorbance changes are diminished. Only 45 equiv PPhs is needed to suppress the interaction of other thiols. We find that 8 enhances 0=PPh3 formation in solutions of 8 and PPhs as clearly monitored via ^ P NMR spectroscopy. After 1 h, 0= PPhs formation is not observed in buffered PPhs solutions. In identical solutions containing PPhs but with added 8, 0=PPh3 formation is clearly observed. These are conclusive and reproducible results. Since 0=PPh3 formation must proceed via the PPha radical cation (i.e., oxidized PPhs) we conclude that 8 acts as an oxidant to both PPhs and any added thiols. Since Hey can serve as the best reducing agent of the series 1-5 (due to its facile transformation to the carbon-centered radical. Figure 6.2A), we utilize a threshold level of competing PPhs which allows for only Hcy-generated spectrophotometric changes in solutions of 8- PPha. Importantly, the UV-Vis spectrum of 8 affords the same signaling patterns to PPhs as it does to added Hey (increased absorbance), again showing that both the phosphine and thiol act as reducing agents towards 8.

The selective signal responses to Hey may be explained on the basis of its function as a more potent reducing agent than other thiols. It is known, for example, that Hey is a significantly more potent reducing agent towards dehydroascorbic acid than either GSH or Cys. ^ Furthermore, homocysteine has been found to strongly inhibit the oxidation of dihydrorhodamine and luminal in the presence of strong oxidants.^^ It also rapidly reduces ferrylmyoglobin to metmyoglobin.^^ Strong support for the free radical-induced fragmentation of Hey in the presence of 8 derives from mass spectra which exhibit prominent peaks for glycine (loss of ethylene and H2S) and glycine


dimer. Glycine and glycine dimer (and formation of ammonia) are known products of alpha-amino carbon-centered radicals."^^

Oxidizing thiyl radicals can rapidly equilibrate to reducing, captodative a-amino carboncentered radicals under physiological, aerobic conditions.^^ Additionally, reducing disulfide radical anions rapidly decay to the reducing a-aminoalkyl radicals.^^ The equilibria involved in the free radical chemistry of biological thiols are pH-dependent and include several radical and recombinant species. Hey can thus function as a relatively potent reducing agent, though it is also believed to be a causative agent of oxidative stress. The understanding of hyperhomocysteinemia and its associated pathogenicity continues to be of significant current interest. The dication methyl viologen (9) has been previously used as an oxidant during a detailed investigation of the equilibrium kinetics of both the reducing disulfide and the a-amino carbon-centered radicals derived from Hey, Cys and GSH. ^ Reducing radical formation was monitored via changes in the UV-Vis spectra of solutions containing the methyl viologen radical cation which formed in the presence of the biological thiols.^^ Importantly, it was surmised that the formation of the reducing alpha-aminoalkyl radical derived from Hey should be particularly favorable.^^ This was attributed to an intramolecular hydrogen abstraction mechanism which involves a five-membered ring transition state (Figure 6.11). ^^ In contrast, in the cases of Cys and GSH, H-atom abstraction to afford a reducing carbon-centered radical leading to formation of the a-amino carbon radical proceeds via less-favored four- and nine-membered ring transition state geometries.^^ It should thus be possible to develop conditions for the selective detection (right)- of Hey. We find that upon heating colorless solutions of 9 (4.0 mM) at a very gentle reflux (5 min, pH 7.5, 0.1 M tris buffer, 100 % H2O, 17 mM aminothiol), visual signaling selective for Hey is observed. The color formation can be monitored via the appearance of absorptions at 398 nm and 605 nm. Importantly, solutions of Cys, GSH and 9 remain colorless."^^

The solution containing Hey turns blue selecUvely, from colorless, as readily seen by visual inspection. This result embodies a beginning towards the development of Hey and related thiol-selective HPLC post-column reagents

® ooc ® 00c

H ^ N - ^ H,NJn

Figure 6.11. Proton abstraction leading to formation of the a-aminoalkyl radical from the thiyl radical of Hey (left) and Cys (right). This process is favored for Hey. Reprinted with permission from J. Am. Chem. Soc, 126(11), 3400-3401 (2004). Copyright 2004 American Chemical Society.


which are colorless and nonfluorescent (low background) until they interact with target analytes. Absorbance changes due to the formation of reduced 9 were previously^^ found to be promoted by not only Hey, but also GSH and Cys. In addition, the rate constants experimentally derived for the proton abstraction step were found to be surprisingly similar for GSH and Hey. Simulated kinetics of the decay of the disulfide radical anion to the reducing alkyl radical agreed with experimental results only in the case of GSH. These experiments were each performed at pH 10.5. No selective color changes due to Hey were reported.

Based on the significant pH dependence of the equilibria involving biothiol chemistry, as well as the relatively favored proton abstraction mechanism previously proposed for Hey, we found appropriate experimental conditions for the highly selective detection of Hey. Selective colorimetric detection of Hey is attained in our lab at neutral pH.

Our result thus embodies the strongest evidence to date for the geometrically-favored intramolecular proton-coupled electron transfer mechanism (Figure 6.11) for Hey compared to other thiols. Our work towards sensor development thus far has contributed to the understanding as to why Hey can fimction biochemically as a more potent reducing agent compared to Cys. GSH and other structurally-related thiols.

Two additional experimental findings include the facile deproteinization of plasma via simple solvent precipitation (MeOH or MeCN), a commonly used sample pretreatment before thiol analysis, allows us to perform thiol detection in a fashion similar to our initially-used more cumbersome centrifugation and filtering experiments (Figure 6.12).

370 390 410 430 450 470 490

Wavelength (nm)

510 530 550

Figure 6.12. Interaction of 6 (4 x 10" M) with L-cysteine in human blood plasma (pH 9.5) at room temperature. Plasma was deproteinized with MeCN. The concentration of Cys is increased from 4.9 X 10" to 7.4 X 10"* M. Reprinted with permission from J. Am. Chem. Soc, 126(2), 438-439 (2004). Copyright 2004 American Chemical Society.


Optimization of many known solvent deproteinization methods is ongoing. A related second finding is that pretreatment of solutions of 8 with light leads to more sensitive absorbance responses to the presence of Hey (improvement to ca. 25 % absorbance change so far, as compared to earlier work, ca. 15 %). We are continuing to study other related oxidants and several means of increasing their oxidizing power, while retaining Hey selectivity.

6.8. AUTOMATED POST-COLUMN DETECTION OF CYSTEINE AND HOMOCYSTEINE

We have developed a chromatographic method based on the post-column reaction of Cys and Hey with methyl viologen. The selective detection of thiols can be accomplished using a common UV-Vis detector set at 610 nm. Figure 6.13 exhibits the baseline separation and detection of Cys and Hey.

6.9. BIOTHIOL DETECTION BASED ON SIMPLE ARRAYS

The interaction of biothiols with metals is of well-known biomedical importance. We envision that studies of organometallic dyes as biothiol binding agents have a good likelihood in aiding us towards further elucidating biothiol chemistry, as we have done in our other dye-analyte studies. A screen (UV-Vis) of our reagents are summarized in both Figure 6.14 and Table 6.1. The data shows that even a relatively very small set of our reagents displays variability in selectivity and spectral features upon interacting with thiols. This study should lead to ready access of array-type sensors for biothiol discrimination via visible range spectroscopic techniques.

0.002 n

% 0.001

S 0000

1 o -0.001 n ^ .n 009

C

Cys

1 T in J l K ^~^uu_. ) 2 4

Retention time (min)

6

Figure 6.13. HPLC post-column detection of Cys and Hey (5 )ig respectively). Conditions:

column: Cig 5}a (250 x 4.6 mm) mobile phase: trifluoroacetic acid 0.01 M; flow rate: 1.5 ml/min;

reagent: methyl viologen 0.0025 M in carbonate buffer 0.5 M pH 9.5 reagent flow rate: 0.5 ml/min

reactor temperature: 80 °C.


D Homocysteine

n Cysteine

• Penicillamine

• Glutathione

Figure 6.14. Relative changes in absorbance (AA) upon the interaction of 8 (as a chelator of each

of the various metals shown) at the wavelengths shown upon addition of 1-3 or 5. Table 6.1 below

summarizes additional data generated from this preliminary selectivity and spectral response screen.

Table 6.1 below summarizes additional data generated from this spectral response screen.

Table 6.1. Each analyte and 8/metal complex exhibits wavelength shifts upon addition of thiols (plus denotes red shift, minus denotes blue shift). In addition, absorption intensity varies (Figure 6.14): w = weak, m = moderate and s = small absorptions. The symbol A denotes absorbance increases at the wavelength corresponding to each thiol. Concomitant (ratiometric) absorbance decreases of the 8/metal complexes occur at their respective wavelength maxima. Likewise • denotes absorbance decrease upon added thiol, which results in absorbance increases of the corresponding 8/metal complexes. The results of this small screen strongly suggest that one should be able to generate sufficient descriptors and specific spectroscopic (UV-Vis or fluorescence) signatures for each of 1-5 and related biomolecules.

dye

added thiol (0.1 mM)

metal

1 3 5 2

^max (nm) AAmax (niTi) A>„max (niTi) AA,max (nm) AAmax (nm)

8

8

8

8

Fe(lll)

Co{ll)

Cu(ll)

Zn(ll)

510

530

535

515

+90 A

(w)

-25 A (s)

-25 T (w)

-5 T (m)

+90 A (m)

-25 A (s)

-25 • (m)

-5 T (m)

+90 A

(m)

-25 A (s)

-25 T (w)

-5 T (m)

+90 A

(w)

-25 A (m)

-25 T (w)

-5 T (m)


6.10. CONCLUSION

Enhanced understanding of the chemistry of aminothiols promotes both the development of novel chemosensors and new insights into their biochemistry.


We are very grateful to the National Institutes of Health for supporting this research via grant ROl EB002044.

6.12. REFERENCES

1. (a) Review: Refsum, H.; Ueland, P. M.; Nygard, O.; VoUset, S. E. "Homocysteine and Cardiovascular Disease," Annu. Rev. Med. 1989, 49, 31. (b) Seshadri, S.; Beiser, A.; Selhub, J.; Jacques, P. F.; Rosenberg, I. H.; D'Agostino, R. B.; Wilson, P. W. F. "Plasma Homocysteine £is a Risk Factor for Dementia and Alzheimer's Disease," N. Engl. J. Med. 2002, 346, 476.

2. Recent reviews: (a) Nekrassova, O.; Lawrence, N. S.; Compton, R. G. "Analytical Determination of Homocysteine: A Review," Talanta, 2003, 60, 1085. (b) Refsum, H.; Smith, A. D.; Ueland, P. M.; Nexo, E.; Clarke, R.; McPartlin, J.; Johnston, C ; Engbaek, F.; Schneede, J.; McPartlin, C ; Scott, J. M. "Facts and Recommendations about Total Homocysteine Determinations: An Expert Opinion," Clin. Chem. 2004, 50, 3.

3. Review: Shimada K.; Mitamura K. "Derivatization of Thiol-Containing Compounds," J. Chromatogr. B 1994, 659, 227.

4. (a) Rusin, O.; St.Luce, N. N. Agbaria, R. A.; Escobedo, J. O.; Jiang, S.; Warner, I. M., Strongin, R. M. "Visual Detection of Cysteine and Homocysteine," J. Am. Chem. Soc. 2004, 126, 438. (b) Wang, W.; Escobedo, J. O.; Lawrence, C. M.; Strongin, R. M. "Direct Detection of Homocysteine," J. Am. Chem. Soc. 2004, 126, 3400.

5. Selhub J. "Homocysteine Metabolism," Annu. Rev. Nutr. 1999, 19, 217. 6. Medina, M. A.; Urdiales, J. L.; Amores-Sanchez, M. I. "Roles of Homocysteine in Cell

Metabolism," Eur. J. Biochem. 2001, 268, 3871. 7. Liao, J. K.; Shin, W. S.; Lee, W. Y.; Clark, S. L. "Oxidized Low-Density Lipoprotein

Decreases the Expression of Endothelial Nitric Oxide Synthase," J. Biol. Chem. 1995, 1,319. 8. Jacobsen, D. W. "Homocysteine-to Test and to Treat," DPC Technical Report. 2001. 9. Girard, P.; Potier, P. "NO, Thiols, and Disulphides," FEBS Lett. 2000, 320, 7. 10. Zhang, X.; Li H.; Jin, H.; Ebin, Z.; Brodsky, S.; Goligosrksy, M. S. "Effects of Homocysteine

on Endothelial Nitric Oxide Production," Am. J. Physiol. Renal Physiol. 2000, 279, F671. 11. Boushey, C ; Beresford, S.; Omenn, G.; Motulsky, A. "A Quantitative Assessment of Plasma

Homocysteine as a Risk Factor for Vascular Disease: Probable Benefits of Increasing Folic Acid Intake," JAMA 1995, 274, 1049.

12. Graham, I. M.; Daly, L.; Refsum, H.; Robinson, K.; Brattstrom, L.; Ueland, P. M. "Plasma Homocysteine as a Risk Factor for Vascular Disease:the European Concerted Action Project," JAMA 1997, 277, 1775.5482.

13. Yi, P.; Melnyk, S.; Pogribna, M.; Popribny, I. P.; Hine, J.; James, S. J. "Increase in Plasma Homocysteine Associated with Parallel Increases in Plasma S-Adenosylhomocysteine and Lymphocyte DNA Hypometylation," J. Biol. Chem. 2000, 275, 29318.

14. Eskes, T. K. "Open or closed? A World of Difference: a History of Homocysteine Research," Nutr. Rev. 1998,56,236.

15. Mills, J. L.; Scott, J. M.; Kirke, P. N.; McPartlin, J. M.; Conley, M. R.; Weir, D. G.; Molloy, A. M.; Lee, Y. J. "Homocysteine and Neural Tube Defects," J. Nutrition. 1996, 126, S756.

16. James, S. J.; Pogribna, M.; Pogribny, I. P.; Melnyk, S.; Hine, R. J.; Gibson, J. B.; Yi, P.; Tafoya, D. L.; Swenson, D. H.; Wilson, V. L.; Gaylor, D. W. "Abnormal Folate MetaboUsm


and Mutation in The Methylenetetrahydrofolate Reductase Gene may be Maternal Risk Factors for Down Syndrome," Am. J. Clin. Nutr. 1999, 70, 495.

17. Kapusta, L.; Haagmans, M. L.; Steegers, E. A.; Cuypers, M. H.; Blom, H. J.; Eskes, T. K.; "Congenital Heart Defects and Maternal Derangement of Homocysteine Metabolism," J. Pediatr. 1999, 135,773.

18. Guldener, C. V.; Robinson, K. "Homocysteine and Renal Disease," Semin. Thromb. Hemostasis. 2000, 26, 313.

19. Law, M. "Fortifying Food with Folic Acid," Semin. Thromb. Hemostasis. 2000, 26, 349. 20. Lentz, S.R.; Haynes, W.G. "Homocysteine: a Clinically Important Cardiovascular Risk

Factor?" Cleveland Clinic J. Med. 2004, 71, 730. 21. (a) Welch, G. N.; Loscalzo, J. "Homocysteine and Atherothrombosis," N. Engl. J. Med. 1998,

338, 1042-1050. (b) Starkebaum, G.; Harlan, J. M. "Endothelial-Cell Injury Due to Copper-Catalyzed Hydrogen-Peroxide Generation From Homocysteine," J. Clin. Invest. 1986, 77, 1370.

22. (a) Jacobsen, D. W. Arterioscler. Thromb. Vase. Biol. 2000, 20, 1182. (b) Jacobsen, D. W. in Homocysteine in Health and Disease 2001 (Carmel, R.; Jacobsen, D. W., eds) pp 425-440, Cambridge University Press, Cambridge, UK.

23. Dutinen, P. A.; Sood, S. K.; Pfeifer, S. I.; Podor, T. J.; Li, J. Weitz, J. I.; Austin, R. C. Blood 1999,94,959.

24. Upchurch, G. R.; Welch, G. N.; Fabian, A. J.; Freedman, J. E.; Johnson, J. L.; Keaney, J. F.; Loscalo, J. J. Biol. Chem. 1997, 272, 17012.

25. Boushey, C. J.; Beresford, S. A. A.; Omenn, G. S.; Motulski, A. G. JAMA 1995, 274, 1049. 26. For example: Lim, A. et al. "In Vitro and in Vivo Interactions of Homocysteine with Human

Plasma Transthyretin," J. Biol. Chem. 2003, 278, 49707. 27. Elliot, A. J.; McEachem, R. J.; Armstrong, D. A. "Oxidation of Amino-containing Disulfides

by Br2-.cntdot. and OH. A Pulse-radiolysis Study," J. Phys. Chem. 1981, 85, 68. 28. Nauser, T.; Felling, J.; Schoneich, C. "Thiyl Radical Reaction with Amino Acid Side Chains:

Rate Constants for Hydrogen Transfer and Relevance for Posttranslational Protein Modification," Chem. Res. Toxicol. 2004, 17, 1323.

29. Varadarajan, S.; Kanski, J.; Aksenova, M.; Lauderback, C ; Butterfield, D. A. "Different Mechanisms of Oxidative Stress and Neurotoxicity for Alzheimer's AP(l-42) and A|3(25-35)," J. Am. Chem. Soc. 2001, 123, 5625.

30. (a) Zhou, J.; MoUer, J.; Danielsen, C. C ; Bentzon, J.; Ravn, H. B.; Austin, R. C ; Falk, E.. "Dietary supplement with methionine and homocysteine promotes early atherogenesis but not plaque rupture in apoE-deficient mice," Arterioscler. Thromb. Vase. Biol. 2001, 21, 1470. (b) Hofmann, M. A.; Lolla, E.; Lu, Y.; Ryu Gleason, M.; Wolf, B. M.; Tanji, N., Ferran, L. J. Jr.; Kohl, B.; Rao, V.; Kisiel, W.; et al.. "Hyperhomocysteinemia Enhances Vascular Inflammation and Accelerates Atherosclerosis in a Murine Model," J. Clin. Invest. 2001, 107, 675. (c) Wang, H.; Jiang, X.; Yang, F.; Gaubatz, J. W.; Ma, L.; Magera, M. J.; Yang, X.; Berger, P. B.; Durante, W.; Pownall, H. J.; et al.. "Hyperhomocysteinemia Accelerates Atherosclerosis in Cystathionine Beta-Synthase and Apolipoprotein E Double Knock-Out Mice with and without Dietary Perturbation," Blood 2003, 101, 3901. (d) Lentz, S. R.; Sobey, C. G.; Piegors, D. J.; Bhopatkar, M. Y.; Faraci, F. M.; Malinow, M. R.; and Heistad, D. D.. "Vascular Dysfunction in Monkeys with Diet-Induced Hyperhomocyst(E)Inemia," J. Clin. Invest. 1996, 98, 24. (e) Devlin, A. M.; Aming, E.; Bottiglieri, T.; Faraci, F. M.; Rozen, R.; Lentz, S. R.. "Effect of Mthfr Genotype on Diet-Induced Hyperhomocysteinemia and Vascular Function in Mice," Blood 2004, 103, 2624. (f) Ungvari, Z.; Csiszar, A.; Edwards, J. G.; Kaminski, P. M.; Wolin, M. S.; Kaley, G.; KoUer, A. "Increased Superoxide Production in Coronary Arteries in Hyperhomocysteinemia: Role of Tumor Necrosis Factor-Alpha, NAD(P)H Oxidase, and Inducible Nitric Oxide Synthase," Arterioscler. Thromb. Vase. Biol. 2003,23,418.

31. Boger, R. H.; Bode-Boger, S. M.; Sydow, K.; Heistad, D. D.; Lentz, S. R.. "Plasma Concentration of Asymmetric Dimethylarginine, an Endogenous Inhibitor of Nitric Oxide Synthase, is Elevated in Monkeys with Hyperhomocyst(E)Inemia or Hypercholesterolemia," Arterioscler. Thromb Vase. Biol. 2000, 20, 1557.

32. Stadtman, E. R. "Oxidation of Free Amino Acid Residues in Proteins by Radiolysis and by Metal-catalyzed Reactions," Annu. Rev. Biochem. 1993, 62, 797-821.


33. (a) Li, S. C; Stewart, P. M. "Homocystinuria and Psychiatric Disorder: A Case Report," Pathology 1999, 31, 221-4. (b) Bracken, P.; Coll, P. "Homocystinuria and Schizophrenia. Literature Review and Case Report," J. Nerv. Ment. Dis. 1985, 173, 51.

34. Fahey, R. C. In Glutathione: Chemical, Biochemical and Medical Aspects Dolphin, D.; Poulson, R.; Avramovic, O., Ed.; Wiley, New York, 1989; Chapter 9, p. 303.

35. (a) Mefford, L; Adams, R. N. "Determination of Reduced Glutathione in Guinea Pig and Rat Tissue by HPLC with Electrochemical Detection," Life Sci. 1978, 23, 1167. (b) Kreuzig, F.; Frank, J. "Rapid Automated Determination of D-Penicillamine in Plasma and Urine by Ion-Exchange High-Performance Liquid Chromatography with Electrochemical Detection using a Gold Electrode," J. Chromatogr. 1981, 218, 615.

36. (a) Yamashita, G. T.; Rabenstein, D. L. "Determination of Penicillamine, Penicillamine Disulfide and Penicillamine-Glutathione Mixed Disulfide by High-Performance Liquid Chromatography with Electrochemical Detection," J. Chromatogr. 1989, 491, 341. (b) Dupuy, D.; Szabo, S. "Measurement of Tissue Sulfhydryls and Disulfides in Tissue Protein and Nonprotein Fractions by High Performance Liquid Chromatography using Electrochemical Detection," J. Liq. Chromatogr. 1987, 10, 107.

37. Ubbink, J. B.; Delport, R.; Riezler, R.; Hayward-Vermaak, W. J. "Comparison of Three Different Plasma Homocysteine Assays with Gas Chromatography-Mass Spectrometry," Clin. Chem. 1999,45,670.

38. Chen, G.; Tran, H.; Datta, A.; Yuan, C. "Determination of Total Plasma Homocysteine using STE Technology," Clin. Chem. 2002, 48, AlOO.

39. (a) Kataoka, H.; Takagi, K.; Makita, M. "Determination of Total Plasma Homocysteine and Related Aminothiols by Gas Chromatography with Flame Photometric Detection," J. Chromatogr. B 1995, 664, 421. (b) Myung, S. W.; Chang, Y. J.; Yoo, E. A.; Park, J. H.; Min, H. K.; Kim, M. S. "Determination of Homocysteine, Cysteine and Methionine in Human Plasma by Gas Chromatography with Electron Capture Detector," Analytical Science & Technology 1999, 12, 408. (c) Kataoka, H.; Tanaka, H.; Fujimoto, A.; Noguchi, I.; Makita, M. "Determination of Sulphur Amino Acids by Gas Chromatography with Flame Photometric Detection," Biomed. Chromatogr. 1994, 8, 119.

40. Haddad, R.; Hoehr, N. F.; Mendes, M. A.; Eberlin, M. N. "Amino Acid Quantitation in Aqueous Matrices Via Trap and Release Membrane Introduction Mass Spectrometry: Homocysteine in Human Plasma," Analyst 2001, 126, 1212.

41. Causee, E,; Siri, N.; Bellet, H.; Champagne, S.; Bayle, C ; Valdigiue, P.; Salvayre R.; Couderc F. "Plasma Homocysteine Determined by Capillary Electrophoresis with Laser-Induced Fluorescence Detection," Clin. Chem. 1999, 45, 412.

42. (a) Nakashima, K.; Uwekawa, C ; Yoshida, H.; Nakatsuji, S.; Akiyama, S. "High-Performance Liquid Chromatography-Fluorometry for the Determination of Thiols in Biological Samples using N-[4-(6-dimethylamino-2-benzofuranyl)phenyl]-maleimide," J. Chromatogr. 1987, 414, 11. (b) Nakashima, K.; Uwekawa, C ; Nakatsuji, S.; Akiyama, S.; Givens, R. S. "High-Performance Liquid Chromatography/Chemiluminescence Determination of Biological Thiols with N-[4(6-dimethylamino-2-benzofuranyl)-phenyl]maleimide," Biomed. Chromatogr. 1989, 3, 39. (c) Takahashi, H.; Nara, Y.; Meguro, H.; Tuzimura, K.; "A Sensitive Fluorometric Method for the Determination of Glutathione and some Thiols in Blood and Mammalian Tissues by High Performance Liquid Chromatography," Agric. Biol. Chem. 1979, 43, 1439. (d) Anzai, N.; Kimura, T.; Chida, S.; Tanaka, T.; Takahashi, H.; Meguro, H. "A Sensitive Fluorometric Determantion of N-acetyl-L-cysteine in Biological Fluids by High-Performance Liquid Chromatography," Yakugaku Zasshi 1981, 101, 1002. (e) Takahaski, H.; Yshida, T.; Meguro, H. "Fluorometric Analysis of Thiols by High Performance Liquid Chromatography with Post Column Derivatization," Bunseki Kagaku 1981, 30, 339. (f) Kageal, B.; Kallberg, M. "Reversed-phase Ion-pair High-Performance Liquid Chromatography of Mercaptoacetate and N-Acetylcysteine after Derivatization with N-(l-pyrene)maleimide and N-(7-dimethylamino-4-methyl-3-coumarinyl)maleimide," Chromatogr. 1982, 229, 409. (g) Kanaoka, Y.; Machida, M.; Ando, K.; Sekine, T. "Fluorescence and Structure of Proteins as Measured by Incorporation of Fluorophore. IV. Synthesis and Fluorescence characteristics of N-(p-(2-benzimidazolyl)phenyl)maleimide," Biochim. Biophys. Acta. 1970, 207, 269. (h) Miners, J. O.; Fearbley, I.; Smith, K. J.; Birkett, D. J.; Brooks, P. M.; Whitehouse, M. W. "Analysis of D-Penicillamine in Plasma by Fluorescence Derivatization with N-[p-(2-benzoxazolyl)-phenyl]maleimide and High-Performance Liquid Chromatography," J.


Chromatogr. 1983, 275, 89. (i) Tsuruta, Y.; Moritani, K.; Date, Y.; Kohashi, K. "N-[4-(5,6-dimethoxy-2-phthalimidyl)phenyl]maIeimide as Precolumn Fluorescence Derivatization Reagent for Thiols," Anal. Sci. 1992, 8, 393.

43. (a) Fahey, R. C; Newton, G. L.; Dorian, R.; Kosower, E. M. "Analysis of Biological Thiols: Quantitative Determination of Thiols at the Picomole Level Based upon Derivatization with Monomoromobimanes and Separation by Cation-Exchange Chromatography," Anal. Biochem. 1981, 111, 357. (b) Newton, G. L.; Dorian, R.; Fahey, R. C. "Analysis of Biological Thiols: Derivatization with Monobromobimanes and Separation by Reverse-Phase High-Performance Liquid Chromatorgrapy," Anal. Biochem. 1981, 114, 383. (c) O'Keefe, D. O.; Lee, A. L.; Yamazaki, S. "Use of Monobromobimane to Resolve two Recombinant Proteins by Reversed-Phase High-Performance Liquid Chromatography based on their Cysteine Contents," Chromatogr. 1992, 627, 137. (d) Demoz, A.; Netteland, B.; Svardal, A.; Mansoor, M. A.; Berge, R. K. "Separation and Detection of Tissue CoASH and Long Chainacyl-CoA by Reversed-Phase High-Performance Liquid Chromatography after Precolumn Derivatization with Monobromobimane," J. Chromatogr. 1993, 635, 251.

44. Amamath, V.; Amamath, K. "Specific Determination of Cysteine and Penicillamine through Cyclization to 2-thioxothiazolidine-4-carboxylic acids," Talanta 2002, 56, 745.

45. (a) Fermo, I.; Arcelloni, C; De Vecchi, E.; Vigano, S.; Paroni, R. "High-Performance Liquid Chromatographic Method with Fluorescence Detection for the Determination of Total Homocyst(e)ine in Plasma," J. Chromatogr. 1992, 593, 171. (b) Svedas, V. J. K.; Galaev, I. J.; Borisov, I. L.; Berezin, I. V. "The Interaction of Amino Acids with o-Phthaldialdehyde: a Kinetic Study and Spectrophotometric Assay of the Reaction Product," Anal. Biochem. 1979, 101, 188.

46. Al-Majed, A. A. "Specific Spectrofluorometric Quantification of D-Penicillamine in Bulk and Dosage Forms after Derivatization with 4-fluoro-7-nitro-benzo-2-oxa-l,3-diazole," Anal. Chim. Acta 2000, 408, 169.

47. Review: Uchiyama, S.; Santa, T.; Okiyama, N.; Fukushima, T.; Imai, K. "Fluorogenic And Fluorescent Labeling Reagents with a Benzofurazan Skeleton," Biomedical Chromatography 2001, 15,295.

48. Geoghegan, K. F.; Hoth, L. R.; Tan, D. H.; Borzillerl, K. A; Withka, J. M.; Boyd, J. G. "Cyclization of N-terminal S-carbamoylmethylcysteine Causing Loss of 17 Da from Peptides and Extra Peaks in Peptide Maps," J. Proteome Res. 2002, 1, 181.

49. Krokhin, O. V.; Ens, W.; Standing, K. G. "Characterizing degradation products of peptides containing N-terminal Cys residues by (off-line high-performance Uquid chromatography)/matrix-assisted laser desorption/ionization quadrupole time-of-flight measurements," Rapid Commun. Mass Spectrom. 2003, 17, 2528.

50. Fermo, I.; Arcellonic, C; Mazzola, G.; D'angelo, A.; Paroni, R. "High-Performance Liquid Chromatographic Method for Measuring Total Plasma Homocysteine Levels," J. Chromatogr. B. 1998,719,31.

51. (a) Kaniowska, E.: Chwatko, G.; Glowacki, R.; Kubalczyk, P.; Bald, E. "Urinary Excretion Measurement of Cysteine and Homocysteine in the Form of their S-pyridinium Derivatives by High-Performance Liquid Chromatography with Ultraviolet Detection," J. Chromatogr. A 1998, 798, 27. (b) Baeyens, W.; Van der Weken, G.; Ling, B.; Lin Moerloose, P. D. "HPLC Determination of N-acetylcysteine in Pharmaceutical Preparations after Precolumn Derivatization with Thiolyte MB using Fluorescence Detection," Anal. Lett. 1988, 21, 741.

52. Winters, R. A.; Zukowski, J.; Ercal, N.; Matthews, R. H.; Spitz, D. R. "Analysis of Glutathione, Glutathione Disulfide, Cysteine, Homocysteine, and other Biological Thiols by High-Performance Liquid Chromatography following Derivatization by n-(l-pyrenyl)maleimide," Anal. Biochem. 1995, 227, 14.

53. Harada, D.; Naito, S.; Kawauchi, Y.; Ishikawa, K.; Koshitani, O.; Hiraoka, I.; Otagiri, M. "Determination of Reduced, Protein-Unbound, and Total Concentrations of N-Acetyl L-Cysteine and L-Cysteine in Rat Plasma by Postcolumn Ligand Substitution High-Performance Liquid Chromatography," Anal. Biochem. 2001, 290, 251.

54. Ivanov, A. R.; Nazimov, I. V.; Baratova, L. A. "Qualitative and Quantitative Determination of Biologically Active Low-Molecular-Mass Thiols in Human Blood by Reversed-Phase High-Performance Liquid Chromatography with Photometry and Fluorescence Detection," J. Chromatogr. A 2000, 870, 433.


55. Parmentier, C ; Leroy, P.; Wellman, M.; Nicolas, A. "Determination of Cellular Thiols and Glutathione-Related Enzyme Activities: Versatility of High-Performance Liquid Chromatography-Spectrofluorimetric Detection," J. Chromatogr. B 1998, 719, 37.

56. Jacobsen, D. W.; Gatautis, V. J.; Green, R. "Determination of Plasma Homocysteine by High-Performance Liquid Chromatography with Fluorescence Detection," Anal. Biochem. 1989, 178,208.

57. Baeyens, W.; Van Der Weken, G.; Ling Ling, B.; De Moerloose, P. "HPLC Determination of N-Acetylcysteine in Pharmaceutical Preparations after Pre-Column Derivatization with Thiolyte MB Using Fluorescence Detection," Anal. Lett. 1988, 21, 741.

58. Ivanov, A. R.; Nazimov, L V.; Baratova, L. A. Qualitative and Quantitative Determination of Biologically Active Low-Molecular-Mass Thiols in Human Blood by Reversed-Phase High-Performance Liquid Chromatography with Photometry and Fluorescence Detection," J. Chromatogr. A 2000, 870, 433.

59. Refsum, H.; Ueland, P.; Svardal, A. M. "Fully Automated Fluorescence Assay for Determining Total Homocysteine in Plasma," Clin. Chem. 1989, 35, 1921.

60. Fiskerstrand, T.; Refsum, H.; Kvalheim, G.; Ueland, P. "Homocysteine and Other Thiols in Plasma and Urine: Automated Determination and Sample Stability," Clin. Chem. 1993, 39, 263.

61. El-Khairy, L.; Vollset, S. E.; Refsum, H.; Ueland, P. M. "Plasma Total Cysteine, MortaUty, and Cardiovascular Disease Hospitalizations: The Hordaland Homocysteine Study," Clin. Chem. 2003, 49, 895.

62. (a) Shipchandler, M. T.; Moore, E. G. "Rapid, Fully Automated Measurement of Plasma Homocyst(e)ine with the Abbott Imx Analyzer," Clin. Chem. 1995, 41, 991. (b) Pemet, P.; Lasnier, E.; Vaubourdolle, M. "Evaluation of the AxSYM Homocysteine Assay and Comparison with the Imx Homocysteine Assay," Clin. Chem. 2000, 46, 1440.

63. (a) Davis, C. J.; Lewis, P. T.; McCarroll, M. E.; Read, M. W.; Cueto, R.; Strongin, R. M. "Simple and Rapid Visual Sensing of Saccharides," Org. Lett. 1999, 1, 331. (b) Lewis, P. T.; Davis, C. J.; Cabell, L. A.; He, M.; Read, M. W.; McCarroll, M. E.; Strongin, R. M. "Visual Sensing of Saccharides Promoted by Resorcinol Condensation Products," Org. Lett. 2000, 2 589. (c) He, M.; Johnson, R. J.; Escobedo, J. O.; Beck, P. A.; Kim, K. K.; St. Luce, N. N. Davis, C. J.; Lewis, P. T.; Fronczek, F. R.; Melancon, B. J.; Mrse, A. A.; Treleaven, W. D.: Strongin, R. M. "The Mechanism of Chromophore Formation in Resorcinarene Solutions and Visual Detection of Mono- and Oligosaccharides," J. Am. Chem. Soc. 2000, 124, 5000. (d) Kim, K. K.; Escobedo, J. O.; Rusin, O.; Strongin, R. M. "Post-column Detection of Mono-and Oligosaccharides with a Chemosensor," Org. Lett. 2003, 5, 5007. (f) Yang, Y.; Lewis, P. T.; Escobedo, J. O.; St. Luce, N. N.; Treleaven, W. D.; Cook, R. L.; Strongin, R. M. "Mild Colorimetric Detection of Sialic Acid," Collect. Czech, Chem. Commun. 2004, 69, 1282. (g) Rusin, O.; Alpturk, O.; He, M.; Escobedo, J. O.; Jiang, S.; Dawan, F.; Lian, K.; McCarroll, M. E.; Warner, I. M.; Strongin, R. M. "Macrocycle-Derived Functional Xanthenes and Progress Towards Concurrent Detection of Glucose and Fructose," J. Fluoresc. 2004, 14, 611.

64. Warren, L. "The Thiobarbituric Acid Assay of Sialic Acids," J. Biol. Chem. 1959, 234, 1971. 65. Burdette, S. C ; Walkup, G. K.; Spingler, B.; Tsien, R. Y.; Lippard, S. J. "Fluorescent Sensors

for Zn2+ Based on a Fluorescein Platform: Synthesis, Properties and Intracellular Distribution," L Am. Chem. Soc. 2001, 123, 7831.

66. (a) Gentle, I. E. De Souza, D. P; Baca, M. "Direct Production of Proteins with N-Terminal Cysteine for Site-Specific Conjugation," Bioconjug. Chem. 2004, 15, 658. (b) Schuler, B.; Pannell, L. K. "Specific Labeling of Polypeptides at Amino-Terminal Cysteine Residues Using Cy5-benzyl Thioester," Bioconjug. Chem. 2002, 13, 1039.

66. Park, J. B. "Reduction of Dehydroascorbic Acid by Homocysteine," BBA-Gen Subjects 2001, 1525,173.

67. Zappacosta, B.; Mordente, A.; Persichilli, S.; Minucci, A.; Carlino, P.; Martorana, G. E.; Giardina, B.; De Sole, P. "Is Homocysteine a Pro-Oxidant?," Free Radical Res. 2001, 35, 499.

68. (a) Zhao, R.; Lind, J.; Merenyi, G.; Eriksen, T. E. "Kinetics of One-Electron Oxidation of Thiols and Hydrogen Abstraction by Thiyl Radicals from Alpha-Amino C-H Bonds," J. Am. Chem. Soc. 1994, 116, 12010. (b) Zhao, R.; Lind, J.; Merenyi, G.; Eriksen, T. E. "Significance of the Intramolecular Transformation of Glutathione Thiyl Radicals to Alpha-Aminoalkyl Radicals. Thermochemical and Biological Implications," J. Chem. Soc, Perkins Trans. 2 1997, 569, and references cited therein.

FLUORESCENT BRONCHOSCOPY

Application to lung cancer screening

Franz Stanzel

7.1. INTRODUCTION

Lung cancer is the most common cancer in the world. The incidence is 1.2 million new cases annually representing 12.3% of all cancers.^ In 2001 lung cancer was responsible for over 1 million deaths worldwide, in Germany for about 40000 deaths. Despite progress of detection and treatment of lung cancer, prognosis is still poor. Less than 15% of all patients diagnosed with lung cancer survive their disease.^ Unfortunately most lung cancer patients present with "advanced stage" disease, where cure occurs in only a small percentage of cases.^ Advanced-stage lung cancer is almost always fatal. More than two thirds of the patients have mediastinal lymph node involvement or distant metastatic disease at the time of diagnosis."* Prognosis strongly depends on the stage of the disease at diagnosis. The five year survival rate in patients with stage-I disease is about 70% and exceeds 90% in stage-la disease. The only patients who achieve long-term survival are those with resectable early-stage disease.^ Early stage disease is defined as primarily lung carcinoma in situ (CIS) (TigNoMo) or stage 0 in the International System for Staging Lung Cancer.^ CIS includes malignant cellular changes throughout the full thickness of the mucosa but an intact basement membrane. Microinvasive carcinoma is described as a few millimetres of bronchial invasion but not involving the muscle or cartilage. Perhaps this is truly stage lA.^ These facts indicate the need for diagnosis in an early preclinical stage and detection of preinvasive lesions. This pins our hope on improving the prognosis and decrementing mortality. Treating CIS is based on the assumption that it will progress to invasive carcinoma and that treatment at this stage will yield a higher cure rate.

Franz Stanzel, Asklepios Fachkliniken Munich-Gauting, Center for Respiratory Medicine and Thoracic Surgery, Gauting, Germany D-82131

163

164 F. STANZEL

On the other hand, carcinogenesis is interpreted as a multi-step process taking a lot of time. It is believed to result from the multistep accumulation of genetic and molecular alterations involving cell proliferation and apoptosis. The process from early changes of metaplasia and different stages of dysplasia to severe dysplasia is estimated to take 3-4 years 7' ^ Severe dysplasia proceeds to carcinoma in situ (CIS) in a time span of about 6 months and to invasive cancer in 2 to 10 years7~^ Obviously, it is important to detect at least these more advanced lesions because approximately 10% of moderate dysplasia and 40 to 83% of severe dysplasia will progress to invasive cancer. ^"^^ But whether all early lesions even CIS will progress is not known. In addition, 50-60% of the squamous cell carcinomas develop in the central airways^^ which can be reached with the bronchoscope.

But a major obstacle is detection.^ Patients with early lesions are usually asymptomatic and radiographically quiescent. And if they get bronchoscopy the majority of these early stages, even when centrally located, are missed by conventional white-light (WL) bronchoscopy. Bronchoscopic detection of dysplasia and CIS has been limited to roughly 30% of the total number. '* The lesions are usually small and only a few cell layers thick, ^ and with an intraepithelial growth pattern.^^ Only subtle changes are common. Lam^^ measured such lesions in 19 patients using a grid. 28 mild dysplasias, 9 moderate or severe dysplasias, and 1 CIS were included. 55% of these were smaller than 1.5mm in size and 45% ranged from 1.5 to 3mm. The largest was a CIS of 3mm size. Subtle changes are hard to differentiate from non-specific inflammatory changes. Some discrete spur thickening, subtle mucosal irregularity, minor vessel dilatation, localised edema, erythema, granular mucosa or distorted light reflex^^ can easily be missed even by experienced bronchoscopists. Flat or superficially spreading lesions of 20mm or more in surface diameter and nodular or polypoid lesions of more than 2mm in size are usually visible. However, flat and superficial lesions smaller than 5mm in diameter cannot be seen by WL bronchoscopy.^ ' ^ Approximately 75% of CIS are of the superficial or flat type. So it is no wonder that the majority will be missed using conventional WL bronchoscopy with optical fiberbronchoscopes.

Another reason for the importance of detection of early malignant lesions is the high rate of second primaries in heavy smokers. Serial sectioning of the bronchial tree in patients who died from lung cancer found synchronous CIS in 13.2 to 22.5% of the patients.^' ^ In resected material from patients who were operated and in sectional material of patients who died from lung cancer, metaplasias, dysplasias, or CIS were found in the bronchial tissue in up to 30% of the patients.^^ For resected non-small-cell lung cancer, the rate of metachronous second primary lung cancers was 3.6-4.2% per year. ' ^ A study in Italy found that among patients with completely resected stage-I lung cancer, 17% developed a second primary tumor during a median follow-up period of 3 years. " This reflects the challenge of detecting synchronous or metachronous second primaries at an early stage^^ in preoperative staging and in follow-up.

FLUORESCENT BRONCHOSCOPY 165

7.2. PHENOMENON AND TECHNIQUES

7.2.1. Drug-Induced Fluorescence

Originally, drug-induced fluorescence had been in use for tumor detection.^^ However, preprocedural application of a photosensitizer is required. Hematoporphyrin derivative (HpD) or its partially purified form, dihematopor-phyrin ether/ester (Photofrin®), which has been licensed by the FDA for photo-dynamic therapy of lung cancer, ^"^^ and 5-aminolevulinic acid (ALA) are preferentially retained in tumors.^^ When excited by violet light HpD, Photofrin®, or ALA emits red fluorescence.^^' ^ This leads to increased red fluorescence within abnormal tissues while green fluorescence remains low, thereby enhancing the contrast between normal and abnormal areas.^ Imaging devices can detect this phenomenon. Thus tumors can be distinguished by their more intense fluorescence.^^ The main disadvantages are high costs and potentially dangerous side effects. In particular there is a long-lasting photosensibilisation of the skin.^''' ^ ' ^ Increasing tumor selectivity and decreasing skin sensibilisation has been the field of investigation of the last years^\ ALA was applied orally or topically by inhalation.^ ' ^ ' " First results indicate a high sensitivity for detection of dysplasia and CIS. ^ A main disadvantage is the low specificity caused by a highly variable deposition among individuals, and non-specific uptake of ALA by inflamed mucosa. Even trials to achieve a reproducible and homogeneous deposition in the central airways by specially designed inhalers did not sufficiently solve the problem.^^ The most important disadvantage is that drug-induced fluorescence bronchoscopy cannot be performed without prepara-

The D-Light AF-System contains a special blue-light mode for drug-induced fluorescence, the so-called ALA mode. Its wavelengths are a narrower band of blue light (380-445nm), slightly different from that of the AF mode. The filter in the eyepiece of the bronchoscope for drug-induced fluorescence differs from those used in the AF detection mode, too.

7.2.2. Autofluorescence

The basic discovery was that detection of early malignant lesions could be achieved without using any drugs.^^^^ It has been known since the early part of last century that tissues fluoresce when exposed to light of a suitable wavelength and that infiltrating tumors, by disturbing these properties of fluorescence, could be detected more easily.^ Fluorescence of tissue is a property enabling differentiation between normal and premalignant or malignant areas. Illuminated by violet or blue light (400-440nm) normal bronchial mucosa emits fluorescent light with a major peak of about 520nm in the green range and a minor peak of about 630nm in the red range.^^ Premalignant or early malignant changes such as severe dysplasia or CIS lead to an almost 10-fold reduction of total fluorescence and a change of the green/red ratio from 5:3 to 2:3.^^' ^ There

166 F. STANZEL

are different explanations for the decrease in autofluorescence^^ and the contrast between normal and tumor tissues. The changes on the way to carcinoma include thickening of the epithelial cell layers, tumor hyperemia, redox changes in the tumor matrix, and a reduced fluorophore concentration.^^' "^ The epithelial layer contributes less than 5% to the overall detected fluorescence."^^' ^ The bulk of the fluorescence signal comes from the submucosa."*^ However, this autofluo-rescence is very weak.

For capture, different systems have been developed. The first device, the Lung Imaging Fluorescence Endoscopy (or Laser Induced Fluorescence Endoscopy = LIFE) System, was developed by the British Columbia Cancer Agency staff together with Xillix Technologies, Vancouver, Canada."*" In 1996 the US Food and Drug Administration approved the use of the LIFE system. A helium-cadmium laser generates a monochromatic blue light of 442nm as excitation light. The system consists of a dual-channel CCD camera system and an intensi-fier board with a colour monitor. The LIFE System works with different types of fiberbronchoscopes. The optical bundle of a conventional bronchoscope is used for illumination and detection of the autofluorescent light. In addition to WL bronchoscopy LIFE bronchoscopy is carried out as a routine fiberbroncho-scopy in local anesthesia. '" '"*^ Normal mucosa has a bright green appearance, while areas of reduced fluorescence display brown-red discoloration.'*^ The procedure adds an extra time of 5 to 15 minutes ' "^ to conventional bronchoscopy.

Sequential systems, which had been developed, use a xenon lamp to produce non monochromatic blue light of high intensity. Incoherent blue light has an adequate emission from the fluorophores, but the light may penetrate to deeper layers of the tissue because of the lower range of blood absorption.^^ The blue remission and the red fluorescent light are not reduced by the same amount as the green fluorescent light. ^ The colour of malignant and premalignant areas appears bluish-red and darker. In contrast, the colour of normal mucosa is dominated by pale green fluorescence light. '" ^

The Karl Storz Company, Tuttlingen, Germany, the Laser Research Laboratories of the Urological Clinic of Munich-GroBhadem and the Clinic for Pneumology of the Asklepios Fachkliniken Munich-Gauting, Germany, developed and tested the D-Light AF-System as the first non-laser system.^ ' ^ The system can be used with both fiberscopes and rigid telescopes with a filter integrated into the eyepiece. Detection is possible with the naked eye or by a CCD camera and a monitor connected to the eyepiece. The Pentax Company brought the SAFE 1000 System (Pentax Asahi Optical, Tokyo, Japan) to the market. ^ Other developments are the Wolf system called DAFE and an update of the LIFE system using a Xenon light source (Xillix Onco-LIFE) compatible to nearly every fiberbronchoscope type. These systems are based on a xenon light source, which is cheaper and more effective.

Pentax is responsible for the latest step in this development. SAFE-3000 is the first chipvideobronchoscopy system with the possibility of autofluorescence examination. It is an integrated processor containing separate light sources for white light and autofluorescence examination, based on the videobronchoscope EB-1970AK. The white light source is a xenon lamp, the blue light source a


diode laser of 408nm. The image is displayed on a full screen monitor as WL image, or AF image, or both simultaneously, but smaller. The advantage is that the bronchoscopist can compare the WL image with the AF image every time.

7.3. INVESTIGATIONS AND DATA

The measure for evaluation of white light and fluorescence techniques is the histological classification of biopsy specimens taken in optically detectable lesions or control biopsy specimens. To quantify the value of fluorescence techniques Lam introduced "relative sensitivity" as the ratio between the number of lesions found by combined examination (WL + fluorescence) and the number of lesions found by WL alone. But the true sensitivity of fluorescence mode remains unknown, because it is unclear how many lesions are missed by both, WL and fluorescence mode.^^

7.3.1. Drug-Induced Fluorescence

ALA-induced fluorescence for detection of early cancers was evaluated in a pilot study.^ ' ^ 88 patients with radiological or clinical suspicion of lung cancer or with a suspicious cytology (sputum, bronchoscopy) were included. Among 360 biopsies the relative sensitivity for all degrees of dysplasia and CIS was 1.8. The main disadvantages were a fast bleaching effect within a few minutes besides the need of a drug administered before bronchoscopy. This limits its use in clinical practice.^^ The non-reliable representation of larger visible cancers and the non-detection of the borderlines of invasive cancers were disappointing. The study by Awadh " using LIFE for detection of ALA-induced fluorescence after oral administration of therapeutical dosages of ALA demonstrated widespread false-positive fluorescence in 5 out of 6 patients. Several biopsies of these areas revealed inflammation, metaplasia, or mild dysplasia. ALA-induced protoporphyrine IX-fluorescence was not specific for tumor tissue.

Diagnostic fluorescence bronchoscopy can be performed pre-eminently when photodynamic therapy by HpD or Photofrin® was planned, and the drug was given intravenously. This is a favourable situation for studying drug induced fluorescence. However, side effects and costs limit clinical use. ALA-induced fluorescence appears promising because there is no long-lasting skin hypersensibilisation. However, AF bronchoscopy is a more promising tool providing more reliable results without the need of a preprocedural drug. Huber^^ mentioned the optional use of a combination of autofluorescence technique and ALA-induced fluorescence. To our knowledge, such a combined system is not clinically used and does not clearly improve the sensitivity for early detection.^^

168 F. STANZEL

7.3.2. Autofluorescence Bronchoscopy

Several investigations using the LIFE-System for detection of early lung cancer have been published. In summary, WL bronchoscopy and LIFE increased the relative sensitivity of detecting dysplasia and CIS by a factor of 1.1 to 6.3 compared with conventional WL bronchoscopy alone. ^"^^ Lam^^ summarised the worldwide experience with the LIFE-Lung device based on the examination of over 1000 patients.^^' ' ' ^^^ ' ~ ^ The numbers include lesions found by WL mode, by AF mode, and by random biopsies, the latter meaning findings by chance. WL bronchoscopy alone found 40% of the detected lesions. But all these studies show a high variability. The values range from 27% in USA/Canada, to 47% in Europe, and 51% in Japan/Singapore. The addition of LIFE increased the detection rate 2-fold, ranging from 2.6-fold (71%) in USA/Canada, 1.8-fold (83%) in Europe, to 1.7-fold (88%) in Japan/Singapore.

Venmans^^ noted that in the studies LIFE always had been applied after conventional bronchoscopy, possibly causing a bias in favour of LIFE-System. Their study had two protocols. In one protocol, LIFE followed WL, in the other, WL followed LIFE. They detected no difference between the two protocols. However, Kato and Ikeda^^ recommend WL bronchoscopy first for general information, followed by AF bronchoscopy for more detailed aspects. In another study by Hirsch^^ the sequences of bronchoscopists and examination mode are randomised. Neither the examiner nor the sequence had a significant influence. The so called detection ratio (DR) incorporates sensitivity and specificity. More biopsy specimens were taken on the basis of abnormalities detected by LIFE than by WLB. This over-sampling could lead to increased sensitivity at the expense of specificity. But the analyses of DRs demonstrated that LIFE had a better diagnostic ability to detect high-grade dysplasia, irrespective of the number of biopsies per subject. Clearly, LIFE was better than WLB in the detection of high-grade premalignant changes with respect to overall sensitivity and DRs.

Fewer results with non-laser light systems have been reported. We published the results of a pilot study in 1999,"* this dealt with 60 patients older than 40 years, including patients with clinically or radiologically suspected lung cancer, patients with resected non-small-cell lung cancer, and patients with suspicious cytological findings (sputum, bronchoscopy). Of the 264 biopsies that were taken, there were 5 mild dysplasias, 6 moderate/severe dysplasias and 1 CIS, and in addition 36 invasive carcinomas. The relative sensitivity for detecting dysplasia or CIS was 2.8. The results were limited by the small number of premalignant lesions and the relatively high number of invasive cancers included.'^'^'

The results of a European prospective randomised multicenter study for evaluating the Storz D-Light AF-System in early lung cancer detection are accepted for publication.^^ 1173 patients were included, stratified by 4 risk groups, and randomised in two study groups. The patients of arm A were examined by WL and AF bronchoscopy, the patients of arm B underwent WL bronchoscopy only. The results are based on the comparison of prevalences and not


of relative sensitivities. The prevalence of patients with preinvasive lesions in arm B was 2.7% compared with 5.1% in arm A. For patients with dysplasia II-III, WLB and AFB increased the detection rate by a factor of 2.1, while for CIS the factor was only 1.24. The biopsy based sensitivity of WLB alone and WLB+AFB for detecting dysplasia II-III and CIS was 57.9% compared with 82.3% (that means a 1.42-fold increase). The corresponding specificities were 62.1% vs. 58.4%.^^

Published data using the SAFE 1000 System are sparse. In one study^^ the prevalence of precancerous lesions was very low with only 5 moderate dysplasias found in 157 biopsy samples. They reported an increased detection rate of 3.75-fold. In a Japanese study of 79 biopsies^^ the relative sensitivity of the SAFE 1000 System was found to be 1.9. However, dysplasias were not differentiated as to their severity, and 24 were invasive carcinomas. Another Japanese study^^ found an increase of 1.2 for overall detection rate of dysplasia and worse, for dysplasia alone 1.33. 108 patients were included in the study. The authors report the most significant benefit by AFB over WLB, if severe dysplasia is detected. But 27% of severe dysplasias were detected neither by WLB nor by AFB. For SAFE-3000 there are no data available up to now.

7.4. DISCUSSION

WLB with optical fiberbronchoscopes is ineffective in detecting premalig-nant changes in the bronchial epithelium and in detecting early lung cancers.' ^ LIFE and other fluorescence techniques have been reported to increase the sensitivity for detecting moderate dysplasia or worse. But there are differences in the results (table 7.1). The study designs and calculations are based on different abnormalities ranging from dysplasia up to CIS or microinvasive cancer, partially all mixed together. Therefore, direct comparisons of the data are difficult. The study populations are not always comparable too, because the inclusion criteria vary as well. When a high rate of invasive carcinomas is included, there is a higher sensitivity for WL detection which will decrease the WL/fluorescence ratio. In the published studies, rates reaching from 7 invasive tumors out of a total number of 83 lesions^^ to 15 out of 19 " were found. The North-American study^^ included 40 invasive lesions out of a total number of 142. However, there is a surprisingly high rate of WL-negative invasive tumors in the study. Thus the detection of early malignant lesions in WL-mode had a sensitivity of 9%. It is not surprising that the relative sensitivity of the combined examination mode compared to WL mode is 6.3. However, the AF detection rate was only 56%.^^ There is a leaming curve for the use of fluorescence techniques sharpening the eye of even experienced endoscopists.^^' ^ Higher numbers of biopsies were performed during the leaming phase (4.6 versus 4.0 biopsies per patient), and the numbers of preinvasive lesions found were lower (20 vs. 102 lesions)."^^ Even in centers with a relatively high rate of early cancer diagnoses, indicating a high degree of experience of investigators,^^' ^ ' ^ addition of AF bronchoscopy led to an improvement of the diagnostic rate.^^

170 F. STANZEL

Experience with early cancer and fluorescence methods increases the sensitivity of WL bronchoscopy, too. So the relative sensitivity for recent studies was lower, possibly due to the higher WL sensitivity and the improved WL equipment.

Table 1 Details of discussed publications (adapted from HauBinger^^)

First author

LW^

Yokomise^''

Lam^^

Kurie^^

Khanavkar^^

Kakihana^^

Venmans^^ Kusunoki^^

Shibuya''

Sato'" Hirsch'^

Miyazu^^

Moro-Sibilot^^ HauBinger^^

No. of subjects/biopsies

223/451

30/51

173/700

38/245

165/162

72/147

114/790 65/216

64/212

50/123 55/391

54/67

244/354

1173/2907

Inclusion criteria

LC45 LCres21 +ENTC 4 S30 LC, LC res, Cyt all nr LC, LC res, Cyt all nr LC res, ENTC, S all nr LC, LC res, Cyt, S, Care all nr LC 59 nr S50nr Cyt 22 nr LCnr LC32 LC res 23 Cyt 40 Cyt, hemoptysis allnr Cytnr LC, S, Cyt allnr LC, Cyt, PDT all nr LC res, ENTC, S all nr LC 55.7/55.1 LC res 30.2/31 Cyt4.6/4.3 COPD or occ exp 9.5/9.6 all S and older than 40 y

Prevalences Dys No. (

nr

13.3

nr

0

nr

30.6

12.0 nr

40.6

80 52.7

1.0

4.0

3.9

II/III, CIS Df patients

Sensitivity Dys II/III, CIS

WLB/AFB+WLB

(%) Dys II/III 38.5/73.1 CIS 40/91.4

65/90 (Dys and cancer) 8.8/55.9

-

31.8/86.4

51/88

53/84 64/88

68/91

85/94 21.9/78.3

nr

35.7/85.7

57.9/82.3

LC=lung cancer, LC res=follow up of resected lung cancer, ENTC=ENT cancer, S=(former) smokers, Cyt=positive cytology, Carc=carcinogenic exposure, PDT=photodynamic therapy, occ exp=occupational exposure, nr=percentage not reported, Dys Il/lII=moderate/severe dysplasia, CIS=carcinoma in situ


In the studies the prevalence of early malignant lesions varies w idely. Lam" ^ had a 1.6% prevalence of CIS and 19% of moderate or severe dysplasia. This is a very high rate, higher than found by sputum cytology studies in patients with chronic obstructive lung disease.^^ Sato " reported an extremely high prevalence of 80% of premalignant finding including mild dysplasia. He investigated AFB exclusively in patients with abnormal sputum cytology. In the European Multicenter Trial we found a lower overall prevalence of premalignant lesion of only 3.9%. However in another study^^ only 3% of the biopsy specimens showed metaplasia and/or mild dysplasia, no biopsy was worse. However, an inclusion criterion was a history of 20 pack years. Reasons for such differences are different criteria for recruiting patients, especially prepro-cedural cytological examinations or special cytological examinations such as computer-assisted image analysis"^^ and different smoking habits of the patients included. In the European study the group with abnormal sputum cytology comprises the lowest number of patients, but the highest prevalence of preinvasive lesions. This is confirmed by previous studies.^^' ^^'^^' ^ Moreover, a reason may be the inclusion of patients with previous bronchial or upper aerodigestive cancers, and genetic variations.

However, it is difficult to assess the role of the pathologist."^ ' ^ ' '' Also the standardisation of early malignant changes is problematic. There is a great variety in classification of biopsies of early malignancies between two pathologists. In a study by Venmans^^ the site pathologist reported 45 preinvasive lesions among 343 biopsies. A reference pathologist confirmed only 20 of them, upgraded one preinvasive lesion, and downgraded or found unsatisfactory 35 lesions. But evaluation of fluorescence bronchoscopy depends critically on biopsy interpretation as its gold standard. A lack of agreed standards as well as borderline lesions leads to an increase in interobserver variation.^^ The classification into mild, moderate, and severe dysplasia and CIS should follow the criteria of the World Health Organization published not before 1999." ^ Thereby, diagnostic accuracy must be improved, and interobserver variation must be minimised. The development of classification methods such as quantitative image cytometry may implement objective criteria.

The natural history of premalignant lesions remains uncertain.^^' ^ To date there are more open questions about the concept of a gradual progress of premalignant changes to CIS. No gradual pattern of morphological changes could be determined. The changes were very dynamic in a non-gradual manner. The dynamism of each lesion during follow-up makes it rather difficult to predict any outcome by using the initial morphological classification as a starting point.^^ Also, the current sampling method of bronchial tissues may contain many non-specific changes and so disturb the natural history.^^ Recent studies^^' " pointed out data in contrast to former studies contradictory. Out of 9 lesions classified as CIS, 56% progressed to invasive cancer.^^ Because some of the lesions were treated endoscopically, this number is reduced from a possible 78%. Another study " was performed in a high-risk group of 104 patients over a 2-year period. Lowest-grade lesions up to metaplasias progressed only rarely to invasive cancer (2.5%). The same is observed with moderate dysplasia (3.5%). High-grade lesions (27 severe dysplasias, 32 CIS) were treated if they remained

172 F. STANZEL

high-grade or progressed. After 2 years only 47% of the severe dysplasia had progressed or remained high-grade. But in this study, the natural history of CIS was not assessed, because the majority of the patients was treated endoscopi-cally. Especially according to recent data the natural history remains still unclear.

Another important factor is the definition of the lesions. Frequently, in borderlines of invasive cancers, different degrees of early malignant changes can be found. ^ In our opinion such lesions should be valued separately. We only include clearly different lesions that are at least 2cm distant from visible invasive tumors. The result of the biopsies depends on the quality and size of the specimens too. Larger and deeper biopsies are of high quality; smaller biopsies are not as conclusive.

LIFE cannot replace WL bronchoscopy. In AF mode much detail and spatial resolution are lost because of the monocular image." ^ However, most of the recent AF bronchoscopy systems enable autofluorescence examination without first using white-light mode. By starting with autofluorescence examination, artificial changes which disturb the fluorescence image can be avoided. However, this procedure requires optimal conditions for illumination, which may be reduced in patients with hyperemic mucosa, and an experienced bronchoscopy technique.

Control or random biopsies established that there are false-negative lesions. The real number of false negatives is not known. To date, only one study is a matter of WL and AF bronchoscopy followed by a complete pathological examination of the resected bronchial system.^^ 30 operable lung cancer patients were first examined by WL and AF bronchoscopy. After resection of the tumor, 163 thin sections were carried out. The sensitivity of WLB was 90% for cancer, 31% for dysplasia and of AFB 97% for cancer, 50% for dysplasia. This was a relative sensitivity of 1.61 for dysplasia. The specificities were 88% and 84%. One explanation for a negative autofluorescence image in a true malignant lesion may be a different type of tumor growth. Adenocarcinomas or small-cell lung cancers grow almost exclusively in the subepithelial level. Malignant changes at the subepithelial level will not be visible with autofluorescence, since the light sources do not sufficiently penetrate the bronchial wall. The phenomenon of hyperfluorescence can typically be seen in necroses. This may overt tumor growth in autofluorescence mode. ^ Inflammation with its hyper-vascularisation and granulation tissue rich in capillaries and previous biopsy sites cause reduced fluorescence leading to false positive findings. It is assumed that the causes of false-positive lesions occur at the molecular level, indicating very early changes beyond visibility. Thus false-positive measures correlate better with morphometric measurement and molecular genetic analysis than with the conventional histopathologic grade.^^ It is not known how long a false-positive finding after performing a biopsy can be observed.

There are some technical differences among the systems (table 7.2), although LIFE and SAFE 1000 detected all preinvasive lesions in a comparison study.^^ The difference between these systems was that LIFE led to a higher number of false positives and to a longer examination time. Another study led to comparable results between LIFE and D-Light AF, but examination time was


shorter and handling was easier with D-Light AF. ^ On the one hand the light sources are different. The LIFE had a HeCd laser as monochromatic blue light source. The recent developments have a xenon light source and use a broad band of blue light for excitation. SAFE-3000 has a xenon light source for WLB and a diode laser of 408nm for AFB. On the other hand most of the systems are based on optical fiberbronchoscopy except SAFE-3000, which works with a videobronchoscope

Tabic 2 Technical details of AFB systems

Manufacturer

Blue light source

Colors

Comfort

Compatible to (fiberscopes)

LIFE" Onco-LIFE^

Xillix, Canada HeCd laser" Xenon^ 2 (R/G)" 3 (R/G/B)^ Low" Veryhigh^ Olympus" more''

D-Light

Storz, Germany Xenon

3 (R/G/B)

Very high

Storz

System AF

(others adaptable)

SAFE-1000' SAFE-3000^

Pentax, Japan Xenon' Diode laser'* 1(G)' 2 (R/G)'* Low' Very high'* Pentax

DAFE

Wolf, Germany Xenon

3 {R/G/B]

Very high

Wolf (and others)

based on chip technology. Whether there is a difference in detecting early lesions is unknown. To date there are no studies comparing different of the recent autofluorescence systems prospectively, so we cannot exclude that technical factors may have a significant impact on the fmdings.^^

The role of other procedures must be taken into account. Fluorescence bronchoscopy plays a special role in localising early lung cancer. Technical progress has led to the development of a new generation of the videobroncho-scopes with excellent structural information of the bronchial surface.^ ' ^ Fluorescence bronchoscopy will still be advantageous as the colour information is additional to the structural information. However, the difference will become smaller. High magnification videobronchoscopy enables the observation of complex vascular networks in the bronchial mucosa.^^ Narrow band imaging (NBI) is a new technique based on a filter with two blue light bands (Bl: 400-430nm, B2: 420-470nm) and a green band of 560-590nm. High magnification bronchoscopy and NBI combined detect capillary blood vessels in angiogenic squamous dysplasia (ASD). ASD may play a role as a risk factor for progression to cancer.

Fluorescence bronchoscopy will be more effective when embedded in a program of preprocedural evaluation by sputum cytometry of high-risk patients.' ' ^ ' ' ^ Other methods of early detection, such as immunostaining of transformed epithelial cells, and polymerase chain reaction based assays to detect oncogene mutations, will help detect more early cancers and smaller lesions. To determine the penetration depth of the bronchial wall, and to detect

174 F. STANZEL

lymph-node invasion, endoscopic ultrasound is a new tool. Currently, the whole question of screening population groups at high risk for the development of lung cancer is being re-examined. Carefully designed epidemiological studies making use of new tools are needed to determine whether or not lung cancer screening can save lives at affordable costs. A program including screening cytology methods, such as cytometry carried out in risk groups examined by low-dose spiral CT, conventional white-light bronchoscopy, and fluorescence bronchoscopy, staged by endobronchial ultrasound, may improve the poor prognosis of lung cancer, if resection or endoscopical treatment in earlier stages follow immediately. Improvements of the fluorescence systems are ongoing. Spectrometry or spectrofluorometry^^ gives a more objective basis than evaluation by eye. But this is not part of the daily routine of fluorescence bronchoscopy. Optical coherence tomography (OCT) enables high-resolution imaging at or near the cellular level. But to date systems that enable OCT bronchoscopy are still in experimental development. Endomicroscopy provides something comparable but has been introduced the last year only into gastroscopy and coloscopy. It will take a while until the first bronchomicroscope will be introduced.

7.5. CONCLUSIONS

Fluorescence bronchoscopy as a highly sensitive test detects lung cancer at an early stage. ' ' ' The results of the first randomised autofluorescence bronchoscopy study confirmed the superiority of AFB over WLB in detecting premalignant lesions.^^ AFB can not be recommended as a general screening tool for lung cancer^^ but seems to be especially useful in patients with suspicious sputum cytology and normal chest X-ray, as well as in patients with a clinical suspicion of lung cancer. It is not broadly accepted for screening or case finding. Squamous cell cancer accounts for only 30% of all lung cancer. Fluorescence bronchoscopy can be used not only for squamous cell cancer, but this is the only disease clearly studied so far. ^ Though the incidence of this cancer type is declining, especially in industrialised nations, it still plays an important role. The effect depends on the impairment of the superficial layers, so AFB will be helpful despite of histology type if the tumor is affecting the surface of central airways.

Advantages of fluorescence bronchoscopy have been reported for a number of clinical situations. Fluorescence bronchoscopy is superior to conventional bronchoscopy and segmental brushing in the localisation of premalignant or malignant lesions in patients having positive sputum cytology in mass screening for lung cancer.^^ It is superior to conventional bronchoscopy, too, if carried out in a preoperative staging program to detect synchronous lung cancer, " and in a follow-up program to detect metachronous lung cancer.^^ Nevertheless, there is a need for larger studies with longer follow-up and a consistent study design to find more answers to open questions. Improved detection and localisation of early malignancies may give us more information about the genetics and natural


history of these lesions, which is useful for studying carcinogenesis and chemo-prevention.^^ Thus, strategies can be developed for handling these early lesions clinically.

Autofluorescence bronchoscopy is not listed in the Como conference recommendations, which was held on the topic of lung cancer screening for early diagnosis.^ Among diagnostic methods, sputum cytology for the high-risk group and bronchoscopy for patients with positive sputum cytology were evaluated as recommendable. For fluorescence bronchoscopy, there was insufficient evidence to conclude that it is efficious, summarized a Japanese group.^^ The guidelines from the American College of Chest Physicians for interventional pulmonary procedures list as indications to autofluorescence bronchoscopy known or suspected lung cancer by abnormal cytology findings, inspection for synchronous tumors, surveillance following cancer resection, and primary screening among high-risk patients.^^ A pilot study carried out in British Columbia found a detection rate of 3.1% for lung cancer among volunteers -smokers or former smokers, older than 50 years. All subjects had sputum analysis by automated quantitative image cytometry and CT scan at baseline. The majority had autofluorescence bronchoscopy, too. The overall detection rate was improved and the histological cell-type distribution became similar to the general population.^^ This is a promising result, and may pave the way. But the goal is still to increase the survival rate of patients suffering from lung cancer.

7.6. REFERENCES

1. G. M. Strauss, L. Dominioni, J. R. Jett, M. Freedman, and F. W. Grannis. Como international conference position statement. Lung cancer screening for early diagnosis 5 years after the 1998 Varese conference. Chest 111, 1146-1151 (2005).

2. W. A. Fry, H. R. Mende, and D. P. Winchester. The national cancer data base report on lung cancer. Cancer 11, 1947-1955 (1996).

3. F. D. Sheski and P. N. Mathur. Diagnosis and treatment of early lung cancer: As it stands today. Sem. Resp. Crit. Care Med. 25, 387-397 (2004).

4. D. C. Ihde. Chemotherapy of lung cancer. A . Engl J. Med 321, 1434-1441 (1992). 5. T. Naruke, R. Tsuchiya, H. Kondo, H Asamura, and H. Nakayama. Implications of staging in

lung cancer. Chest 112, 242S-8S (1997). 6. C. F. Mountain. Revisions in the international system for staging lung cancer. Chest 111, 1710-

1717(1997). 7. O. Auerbach, A. P. Stout, E. C. Hammond, and L. Garfinkel. Changes in bronchial epithelium in

relation to cigarette smoking and in relation to lung cancer. A . Engl. J. Med. 265, 253-267 (1961).

8. G. Saccomanno, R. P. Saunders, V. E. Archer, O. Auerbach, M. Kuschner, and P. A. Beckler. Cancer of the lung: The cytology of sputum prior to the development of carcinoma. Acta Cytol. 9(6), 413-423 (1965).

9. G. Saccomanno, V. E. Archer, and O. Auerbach. Development of carcinoma of the lung as reflected in exfoliated cells. Cancer 33, 256-270 (1974).

10. E. K. Risse, G. P. Vooijs, and M. A. van't Hof. Diagnostic significance of "severe dysplasia" in sputum cytology. Acta Cytol. 32, 629-634 (1988).

11. P. Band, M. Feldstein, and G. Saccomanno. Reversibility of bronchial marked atypia: implication for chemoprevention. Cancer Detect. Treat. 9, 157-160 (1986).

12. J. K. Frost, W. C. Ball Jr., M. L. Levin, M. S. Tockman, Y. S. Erozan, P. K. Gupta, J. C. Eggleston, N. J. Pressman, M. P. Donithan, and A. W. Kimball. Sputum cytology: use and poten-

176 F. STANZEL

tial in monitoring the workplace environment by screening for biological effects of exposure. J. Occup. Med. 28, 692-703 (1986).

13. G. Saccomanno, O. Auerbach, M. Kuschner, N. H. Harley, R. Y. Michels, M. W. Anderson, and J. J. Bechtel. A comparison between the localization of lung tumors in uranium miners and in nonminers from 1947 to 1991. Cancer 11, 1278-1283 (1996).

14. L. B. Woolner. Pathology of cancers detected cytologically. In: Atlas of Early Lung Cancer. Edited by National Cancer Institute, National Institutes of Health, US Department of Health and Human Services (Igaku-Shoin, Tokyo, 1983), pp. 107-213.

15. Y. Hayata, H. Kato, and C. Onaka. Photodynamic therapy in early stage lung cancer. Lung Cawcer 9, 287-294 (1993).

16. B. Khanavkar. Autofluorescence bronchoscopy. J. Bronchol. 1, 60-66 (2000). 17. S. Lam, C. MacAulay, J. C. LeRiche, and B. Palcic. Detection and localization of early lung

cancer by fluorescence bronchoscopy. Cancer 89(Suppl.), 2469-2473 (2000). 18. N. Nagamoto, Y. Saito, T. Imai, H. Suda, K. Hashimoto, T. Nakada, and H. Sato. Roentgeno-

graphically occult bronchogenic squamous cell carcinoma: location in the bronchi, depth of invasion and length of axial involvement of the bronchus. Tohoku J. Exp. Med. 148, 241-256 (1986).

19. H. Kato, T. Okunaka, and H. Shimatani. Photodynamic therapy for early stage bronchogenic carcinoma. J. Clin. Laser Med Surg. 14, 235-238 (1996).

20. O. Auerbach, E. C. Hammond, and L. Garfmkel. Changes in bronchial epithelium in relation to cigarette smoking. A . Engl. J. Med. 300, 381-385 (1979).

21. K. Kayser, T. Runtsch, S. Bach, and I. Vogt-Moykopf. Dysplastische und hyperplastische Bronchusschleimhautveranderungen beim operierten Bronchuskarzinom. Verh. Dtsch. Ges. Pathol. 12, 449 (\9SS).

22. L. B. Woolner, R. S. Fontana, D. A. Cortese, D. R. Sanderson, P. E. Bematz, W. S. Payne, P. C. Pairolero, J. M. Piehler, and W. F. Taylor. Roentgenographically occult lung cancer: pathologic findings and frequency of multicentricity during a 10-year period. Mayo Clin. Proc. 59, 453^66 (1984).

23. P. Thomas, L. Rubinstein, and the Lung Cancer Study Group. Cancer recurrence after resection of Ti No non-small cell lung cancer. Ann Thorac Surg 48, 242-247 (1990).

24. U. Pastorino, M. Infante, M. Maioli, G. Chiesa, M. Buyse, P. Firket, N. Rosmentz, M. Clerici, E. Soresi, and M. Valente. Adjuvant treatment of stage 1 lung cancer with high-dose vitamin A. J. Clin. Oncol. 11, 1216-1222(1993).

25. S. Lam and H. D. Becker. Future diagnostic procedures. Chest Surg. Clin. North Am. 6, 363-379(1996).

26. A. E. Profio and O. J. Balchum. Fluorescence diagnosis of cancer. Adv. Exp. Med. Biol. 193, 43-50(1985).

27. F. Stanzel, K. HauBinger, D. Geiger, J. Piehler, W. Sauer, and H. Stepp. Photodynamische Therapie in der Pneumologie. Pneumologie 54, 269-277 (2000).

28. K. HauBinger, F. Stanzel, C. Nimmermann, R. M. Huber, and R. Baumgartner. Photodynamic therapy: actual and future aspects. Atemw. Lungenerkrkh. 5, 295-300 (1996).

29. F. D. Sheski and P. N. Mathur. Endoscopic treatment of early-stage lung cancer. Cancer Control 7, 35-44 (2000).

30. S. Lam, C. MacAulay, and B. Palcic. Detection and localization of early lung cancer by imaging techniques. Chest 103(Suppl.), 12S-14S (1993).

31. R. M. Huber, F. Gamarra, H. Hautmann, K. HauBinger, S. Wagner, M. Castro, and R. Baumgartner. 5-Aminolaevulinic acid (ALA) for the fluorescence detection of bronchial tumors. Diag Ther. Endosc. 5, 113-118 (1999).

32. T. J. Dougherty, M. T. Cooper, and T. S. Mang. Cutaneous phototoxic occurrences in patients receiving Photofrin. Lasers Surg. Med 10, 485-488 (1990).

33. F. Stanzel and K. HauBinger. Bronchoscopic fluorescence diagnosis by 5-aminolevulin acid (ALA). Pneumologie 5, 250s (1997).

34. N. Awadh, C. MacAulay, and S. Lam. Detection and treatment of superficial lung cancer by using delta-aminolevulinic acid: a preliminary report. J. Bronchol. 4, 13 17 (1997).

35. K. HauBinger, J. Piehler, F. Stanzel, A. Markus, H. Stepp, A. Morresi-Hauff, and R. Baumgartner. Autofluorescence bronchoscopy: the D-Light System, in: Interventional Bronchoscopy, edited by C. T. Bolliger and P. N. Mathur (Karger, Basel, 2000), pp. 243-252.

36. J. Hung, S. Lam, J. C. LeRiche, and B. Palcic. Autofluorescence of normal and malignant bronchial tissue. Lasers Surg. Med. 11, 99-105 (1991).


37. B. Palcic, S. Lam, J. Hung, and C. MacAiilay. Detection and localization of early lung cancer by imaging techniques. Chest 99, 742-743 (1991).

38. S. Lam, J. Hung, S. M. Kennedy, J. C. LeRiche, S. Vedal, B. Nelems, C. MacAulay, and B. Palcic. Detection of dysplasia and carcinoma in situ by ratio fluorometry. Am. Rev. Respir. Dis. 146, 1458-1461 (1992).

39. L. Freitag, A. Korupp, L Itzigehl, F. Dankwart, E. Tekolf, G. Reichle, H. J. Kullmann, and H. N. Macha. Experience with laser induced fluorescence diagnostic and photodynamic therapy in a multimodal therapeutical concept for operated, recurrent bronchial cancer. Pneumologie 50, 693-699(1996).

40. S. Lam and H. Shibuya. Early diagnosis of lung cancer. Clin. Chest Med. 20, 53-61 (1999). 41. D. M. Benson and J. A. Knopp. Effect of tissue absorption and microscope optical parameters

on the depth of penetration for fluorescence and reflectance measurements of tissue samples. Photochem. Photobiol. 39, 495-502 (1984).

42. J. Qu, C. MacAulay, S. Lam, and B. Palcic. Laser-induced fluorescence spectroscopy at endoscopy: tissue optics, Monte Carlo modelling, and in vivo measurements. Opt. Eng. 34, 3334-3343(1995).

43. J. Qu, C. MacAulay, S. Lam, and B. Palcic. Optical properties of normal and carcinoma bronchial tissue. Appl. Opt. 33, 7397 (1994).

44. S. Lam and H. D. Becker. Future diagnostic procedures. Chest Surg. Clin. North Am. 6, 363-379(1996).

45. B. Khanavkar. How golden is the gold standard? The elusive pathology of early lung cancer (editorial). J. Bronchol. 7, 205-206 (2000).

46. F. Stanzel and K. Haufiinger. Fluoreszenzdiagnostik in der Pneumologie: Autofluoreszenzbron-choskopie, in: Klinische Fluoreszenzdiagnostik undphotodynamische Therapie, edited by R. M. Szeimies, D. Jocham, and M. Landthaler (Blackwell, Berlin, Wien, 2003), pp. 131-146.

47. J. A. Nakhosteen and B. Khanavkar. Autofluorescence bronchoscopy: The laser imaging fluorescence endoscope, in: Interventional Bronchoscopy, edited by C. T. Bolliger and P. N. Mathur (Karger, Basel, 2000), pp. 236-242.

48. S. Lam, J. Hung, and B. Palcic. Mechanism of detection of early lung cancer by ratio fluorometry. Lasers Life Sci. 11, 99-105 (1991).

49. K. HauBinger, F. Stanzel, R. M. Huber, J. Pichler, and H. Stepp. Autofluorescence detection of bronchial tumors with the D-Light/AF. Diag. Ther. Endosc. 5, 105-112 (1999).

50. M. Leonhard. New incoherent autofluorescence/fluorescence system for early detection of lung cancer. Diag. Ther. Endosc.5, 71-75 (1999).

51. R. Adachi, T. Utsui, and K. Furusawa. Development of the autofluorescence endoscope imaging system. Diag Ther. Endosc. 5, 65-70 (1999).

52. F. Stanzel and K. HauBinger. Fluorescence bronchoscopy, in: Interventional pulmonary medicine, edited by J. F. Beamis, P. N. Mathur, and A. C. Mehta (Marcel Dekker, New York, Basel 2004), pp. 355-384.

53. S. Lam, C. MacAulay, J. LeRiche, N. Ikeda, and B. Palcic. Early localization of bronchogenic carcinoma. Diag. Ther. Endosc. 1, 75-78 (1994).

54. H. Yokomise, K. Yanagihara, T. Fukuse, T. Hirata, O. Ike, H. Mizuno, H. Wada, and S. Hitomi. Clinical experience with lung-imaging fluorescence endoscope (LIFE) in patients with lung cancer. J. Bronchol. 4, 205-208 (1997).

55. H. Kato and N. Ikeda. The role of fluorescence diagnosis in the early detection of high-risk bronchial lesions (editorial). J. Bronchol. 5, 273-274 (1998).

56. B. Khanavkar, F. Gnudi, A. Muti, W. Marek, K. M. Miiller, Z. Atay, T. Topalidis, and J. A. Nakhosteen. Principles and results of autofluorescence-(LIFE® System) bronchoscopy compared to conventional bronchoscopy in the detection of early lung cancer. Pneumologie 52, 71-76 (1998).

57. J. M. Kurie, J. S. Lee, R. C. Morice, G. L. Walsh, F. R. Khuri, A. Broxson, J. Y. Ro, W. A. Franklin, R. Yu, and W. K. Hong. Autofluorescence bronchoscopy in the detection of squamous metaplasia and dysplasia in current and former smokers. / . Natl. Cancer Inst. 90, 991-995 (1998).

58. S. Lam, T. Kennedy, M. Unger, Y. E. Miller, D. Gelmont, V. Rusch, B. Gipe, D. Howard, J. LeRiche, A. Goldman, and A. F. Gazdar. Localization of bronchial intraepithelial neoplastic lesions by fluorescence bronchoscopy. Chest 113, 696-702 (1998).

178 F. STANZEL

59. B. J. W. Venmans, A. J, M. van Boxem, E. F. Smit, P. E. Postmus, and T. G. Sutedja. Results of two years experience with fluorescence bronchoscopy in detection of bronchial neoplasia. Diag. Then Enclose. 5, 77-84 (1999).

60. L. Thiberville, T. Sutedja, P. Vermylen, V. Ninane, S. Lam, P. Spinelli, P. Diaz-Jimenez, F. Schrammel, B. Khanavkar, and J. A. Nakhosteen. A multicenter European study using the light induced fluorescence endoscopy system to detect precancerous lesions in high risk individuals. Eur. Respir. J. 14(Suppl.), 2475s. (1999).

61. Y. Kusunoki, F. Imamura, H. Uda, M. Mano, and T. Horai. Early detection of lung cancer with laser-induced fluorescence endoscopy and spectrofluorometry. Chest 118, 1776-1782 (2000).

62. Y. Miyazu, T. Miyazawa, Y. Iwamoto, K. Kano, and N. Kurimoto. The role of endoscopic techniques, laser-induced fluorescence endoscopy, and endobronchial ultrasonography in choice of appropriate therapy for bronchial cancer. J. Bronchol. 8, 10-16 (2001).

63. K. Shibuya, T. Fujisawa, H. Hoshino, M. Baba, Y. Saitoh, T. lizasa, M. Suzuki, M. Otsuji, K. Hiroshima, and H. Ohwada. Fluorescence bronchoscopy in the detection of preinvasive bronchial lesions in patients with sputum cytology suspicious or positive for malignancy. Lung Cancer 32, 19-25 (2001).

64. M. Sato, A. Sakurada, M. Sagawa, M. Minowa, H. Takahashi, T. Oyaizu, Y. Okada, Y. Matsu-mura, T. Tanita, and T. Kondo. Diagnostic results before and after introduction of autofluores-cence bronchoscopy in patients suspected of having lung cancer detected by sputum cytology in lung cancer mass screening. Lung Cancer 32, 247-253 (2001).

65. F. R. Hirsch, S. A. Prindiville, Y. E. Miller, W. A. Franklin, E. C. Dempsey, J. R. Murphy, P. A. Bunn, and T. C. Kennedy. Fluorescence versus white-light bronchoscopy for detection of preneoplastic lesions: a randomized Study. 7. Natl. Cancer Inst. 93, 1385-1391 (2001).

66. D. Moro-Sibilot, M. Jeanmart, S. Lantuejoul, F. Arbib, H. Laverriere, E. Brambilla, and C. Brambilla. Cigarette smoking, preinvasive bronchial lesions and autofluorescence bronchoscopy. Chest 111, 1902-1908 (2002).

67. J. A. Nakhosteen, B. Khanavkar, A. Muti, and W. Marek. Early diagnosis of lung cancer with autofluorescent (LIFE) bronchoscopy and automated sputum cytometry: the next generation. Atemw. Lungenkrkh. 23, 211-217 (1997).

68. P. Vermylen, C. Roufosse, P. Pierard, A. Verhest, J. P. Sculier, and V. Ninane. Detection of preneoplastic lesions with fluorescence bronchoscopy (FB). Eur. Respir. J. 10(Suppl.), 425s. (1997).

69. N. Ikeda, K. Kim, T. Okunaka, K. Furukawa, T. Furuya, M. Saito, C. Konaka, H. Kato, and Y. Ebihara. Early localization of bronchogenic cancerous/precancerous lesions with lung imaging fluorescence endoscope./)/ag. Ther. Enclose. 3, 197-201 (1997).

70. N. Ikeda, H. Honda, T. Katsumi, T. Okunaka, K. Furukawa, T. Tsuchida, K. Tanaka, T. Onoda, T. Hirano, M. Saito, N. Kawate, C. Konaka, H. Kato, and Y. Ebihara. Early detection of bronchial lesions using lung imaging fluorescence endoscope. Diag. Ther. Enclose. 5, 85-90 (1999).

71. B. J. Venmans, H. van der Linden, T. J. van Boxem, P. E. Postmus, E. F. Smit, and T. G. Sutedja. Early detection of preinvasive lesions in high-risk patients. A comparison of conventional flexible and fluorescence bronchoscopy. J. Bronchol. 5, 280-283 (1998).

72. K. HauBinger, F. Stanzel, A. Markus, C. T. BoUiger, J. Pichler. Early diagnosis of bronchial carcinoma. Technical endoscopic progress -a step toward new screening concepts? Pneumologie 53,77-82(1999).

73. K. HauBinger, H. Becker, F. Stanzel, A. Kreuzer, B. Schmidt, J. Strausz, S. Cavaliere, F. Herth, M. Kohlhaufl, K. M. Muller, R. M. Huber, U. Pichlmeier, and C. T. Bolliger, Thorax, accepted for publication (2005).

74. T. Horvath, M. Horvathova, F. Salajka, B. Habanec, L. Foretova, J. Kana, H. Koukalova, P. Paflco, F. Wurst, E. Novotna, J. Pecina, V. Vagunda, R. Vrbacky, R. Talac, H. Coukova, and Z. Pacovsky. Detection of bronchial neoplasia in uranium miners by autofluorescence endoscopy (SAFE-1000). Diag. Ther. Endosc. 5, 91-98 (1999).

75. M. Kakihana, K. K. II, T. Okunaka, K. Furukawa, T. Hirano, C. Konaka, H. Kato, and Y. Ebihara. Early detection of bronchial lesions using system of autofluorescence endoscopy (SAFE) 1000. Diag. Ther. Endosc. 5, 99-104 (1999).

76. K. Furukawa, N. Ikeda, T. Miura, M. Kakihana, M. Saito, and H. Saito. Controversy. Is autofluorescence bronchoscopy needed to diagnose early bronchogenic carcinoma? Pro: Autofluorescence bronchoscopy. J. Bronchol. 10, 64 69 (2003).


77. T. C. Kennedy, S. P. Proudfoot, W. A. Franklin, T. A. Merrick, G. Saccomanno, M. E. Corkill, D. L. Mumma, K. E. Sirgi, Y. E. Miller, P. G. Archer, and A. Prochazka. Cytopathological analysis of sputum in patients with airflow obstruction and significant smoking histories. Cancer /?e5. 56,4673-^678(1996).

78. B. J. Venmans, H. C. van der Linden, H. R. Elbers, T. J. van Boxem, E. F. Smit, P. E. Postmus, and T. G. Sutedja. Observer variability in histopathological reporting of bronchial biopsy specimens: influence on the results of auto fluorescence bronchoscopy in detection of preinvasive bronchial neoplasia. J. Broncholl, 210-214 (2000).

79. S. S. Raab, K. R. Geisinger, J. F. Silverman, P. A. Thomas, and M. W. Stanley. Interobserver variability of a Papanicolaou smear diagnosis of atypical glandular cells of undetermined significance. Am. J. Clin. Pathol 110, 653-659 (1998).

80. A. C. Mehta and T. R. Gildea. Controversy. Is autofluorescence bronchoscopy needed to diagnose early bronchogenic carcinoma? Con: autofluorescence bronchoscopy. J. Bronchol 10, 70-74 (2003).

81. G. Sutedja. New techniques for early detecdon of lung cancer. Eur. Resp. J. 21(Suppl. 39) 57s-66s (2003).

82. W. N. Hittelmann. Clones and subclones in the lung cancer field. J. Natl. Cancer Inst. 91, 1796-1799(1999).

83. B. J. W. Venmans, T. J. M. van Boxem, E. F. Smit, P. E. Postmus, and T. G. Sutedja. Outcome of bronchial carcinoma in situ. Chest ill, 1572-1576 (2000).

84. S. Bota, J. B. Auliac, C. Paris, J. Metayer, R. Sesbolie, G. Nouvet, and L. Thiberville. Follow-up of bronchial precancerous lesions and carcinoma in situ using fluorescence endoscopy. Am. J. Respir. Crit. Care Med. 168, 1688-1693 (2001).

85. N. Ikeda, T. Hiyoshi, M. Kakihana, H. Honda, Y. Kato, T. Okunaka, K. Furukawa, T. Tsuchida, H. Kato, and Y. Ebihara. Histopathological evaluation of fluorescence bronchoscopy using resected lungs in cases of lung cancer. Lung Cancer 41, 303-309 (2003).

86. P. Pierard, B. Martin, J. M. Verdebout, J. Faber, M. Richez, J. P. Sculier, and V. Ninane. Fluorescence bronchoscopy in high-risk patients. J. Bronchol. 8, 254-259 (2001).

87. F. J. F. Herth, A. Ernst, and H. D. Becker. Autofluorescence bronchoscopy - a comparison of two systems (LIFE and D-Light). Respiration 70, 395-398 (2003).

88. K. Shibuya, H. Hoshino, M. Chiyo, K. Yasufiiku, T. lizasa, Y. Saitoh, M. Baba, K. Hiroshima, H. Ohwada, and T. Fujisawa. Subepithelial vascular patterns in bronchial dysplasias using a high magnification bronchovideoscope. Thorax 57, 902-907 (2002).

89. K. Shibuya, H. Hoshino, M. Chiyo, A. lyoda, S. Yoshida, Y. Sekine, T. lizasa, Y. Saitoh, M. Baba, K. Hiroshima, H. Ohwada, and T. Fujisawa. High magnification bronchovideoscopy combined with narrow band imaging could detect capillary loops of angiogenic squamous dysplasia in heavy smokers at high risk for lung cancer. Thorax 58, 989-995 (2003).

90. P. W. Payne, T. J. Sebo, A. Doudkine, D. Gamer, C. MacAulay, S. Lam, J. C. LeRiche, and B. Palcic. Sputum screening by quantitative microscopy: a re-examinadon of a pordon of the Na-donal Cancer Institute cooperative early lung cancer study. Mayo Clinic Proc. 72, 697-704 (1997).

91. W. Marek, S. Krampe, N. J. Dickgreber, L. Nielsen, A. Mud, B. Khanavkar, K. M. Miiller, Z. Atay, T. Topalidis, and J. A. Nakhosteen. Automated quantitative image cytometry of bronchial washings in suspected lung cancer compared to cytology, histology, and final diagnosis. Pneumologie 53, 583-595 (1999).

92. F. Stanzel. Fluorescent bronchoscopy: contribution for lung cancer screening? Lung Cancer 45(Suppl. 2), S29-S37 (2004).

93. M. Sato, A. Sakurada, M. Sagawa, M. Minowa, H. Takahashi, T. Oyaizu, Y. Okada, Y. Matsu-mura, T. Tanita, and T. Kondo. Diagnostic results before and after introduction of autofluorescence bronchoscopy in padents suspected of having lung cancer detected by sputum cytology in lung cancer mass screening. Lung Cancer 32, 247-253 (2001).

94. M. T. van Reus, F. M. Schramel, J. R. Elbers, and J. W. Lammers. The clinical value of lung imaging fluorescence endoscopy for detecting synchronous lung cancer. Lung Cancer 32, 13-18 (2001).

95. T. L. Weigel, P. J. Kosco, S. Dacic, V. W. Rusch, R. J. Ginsberg, and J. D. Luketich. Postoperative fluorescence bronchoscopic surveillance in non-small cell lung cancer patients. Ann. Thorac. Surg. 11, 967-970 (2001).

180 F. STANZEL

96. L. Thiberville, P. Payne, J. Metayer, J. Vielkinds, J. LeRiche, B. Palcic, and S. Lam. Molecular follow-up of a preinvasive bronchial lesion treated by IS-cw-retinoic acid. Hum. Pathol 28, 108-110(1997).

97. A. Sakurada, C. Endo, M. Sato, T. Kondo. Management of central-type early lung cancer: An evidence-based guideline. Nippon Geka Gakkai zasshi 105, 388-391 (2004).

98. A. Ernst, G. A. Silvestri, and D. Johnstone. Interventional pulmonary procedures. Guidelines from the American College of Chest Physicians. Chest 123, 1693-1717 (2003).

99. A. McWilliams, J. Mayo, S. MacDonald, J. C. LeRiche, B. Palcic, E. Szabo, and S. Lam. Lung cancer screening - a different paradigm. Am. J. Respir. Crit. Care Med. 168, 1167-1173 (2003).

QUANTUM DOTS AS FLUORESCENT LABELS FOR MOLECULAR AND CELLULAR IMAGING

Gang Ruan, Amit Agrawal, Andrew M. Smith, Xiaohu Gao, and Shuming Nie*

8.1. INTRODUCTION

Quantum dots (QDs) are light-emitting particles on the nanometer scale, with novel optical and electronic properties that are not available from either isolated molecules or bulk solids. They are under intense development for a broad range of biological applications, including single molecule biophysics, biomolecular profiling, optical barcoding, molecular and cellular imaging. In comparison with organic dyes and fluorescent proteins, semiconductor QDs offers several unique advantages such as size- and composition-tunable emission from visible to infrared wavelengths, large absorption coefficients across a wide spectral range, and very high levels of brightness and photostabilityV Due to their broad excitation profiles and narrow/symmetric emission spectra, high-quality QDs are also well suited for combinatorial optical encoding, in which multiple colors and intensities are combined to encode thousands of genes, proteins, or small-molecule compounds^'^.

Despite their relatively large sizes (2-8 nm diameter), recent research has shown that bioconjugated QD probes behave like fluorescent proteins (4-6 nm), and do not suffer from serious binding kinetic or steric-hindrance problems^"^ . In this "mesoscopic" size range, QDs also have more surface areas and functionalities that can be used for linking to multiple diagnostic (e.g., radioisotopic or magnetic) and therapeutic (e.g., anticancer) agents. Furthermore, polymer-encapsulated QDs have been found to be essentially nontoxic to cells and animal models. This review will focus on the development and applications of QDs for molecular and cellular imaging. We

Departments of Biomedical Engineering, Emory University and Georgia Institute of Technology, 1639 Pierce Drive, Suite 2001, Atlanta, GA 30322, USA. Email: [email protected]. telephone: 404-712-8595, FAX: 404-727-9873.

181

182 G. RUAN ETAL.

will discuss recent advances in delivering QD probes across the cell membrane and in real-time observation of single QDs inside living cells. For further information on QD ftmdamentals and applications, excellent review articles are available in the literature^' ^ ' ^^.

8.2. PROBE DEVELOPMENT

Probe development has focused on the synthesis, solubilization, and bioconjugation of highly luminescent and stable QDs. The particles are generally made from hundred to thousands of atoms of groups II and VI (e.g., CdSe and CdTe) or groups III and V (e.g., InP and InAs) elements. Recent advances have allowed a precise control of the particle size, shape (dots, rods, or tetrapods)^^"^ , and internal structures (core-shell, gradient alloy, or homogeneous alloy)^ ' ^. In particular, quantum dots have been synthesized using both two-element systems (binary dots) as well as three element systems (ternary alloy dots). Their fluorescence emission wavelength can be continuously tuned from 400 to 2000 nm by changing both the particle size and chemical composition, with fluorescence quantum yields as high as 85% at room temperature^^.

High-quality QDs are typically prepared at elevated temperatures in organic solvents, such as TOPO and HDA (both of which are high boiling-point solvents containing long alkyl chains). These hydrophobic organic molecules not only serve as the reaction medium, but also coordinate with unsaturated metal atoms on the QD surface to prevent formation of bulk semiconductors. As a result, the nanoparticles are capped with a monolayer of the organic ligands and are soluble only in non-polar hydrophobic solvents such as chloroform. For biological imaging applications, these hydrophobic dots can be solubilized by using amphiphilic polymers that contain both a hydrophobic segment or side-chain (mostly hydrocarbons) and a hydrophilic segment or group (such as polyethylene glycol or multiple carboxylate groups). A number of polymers have been reported including octylamine-modified low molecular weight polyacrylic acid, PEG derivatized phospholipids, block copolymers, and polyanhydrides^' ' • . As schematically illustrated in Figure 8.1, the hydrophobic domains strongly interact with tri-n-octylphosphine oxide (TOPO) on the QD surface, whereas the hydrophilic groups face outward and render QDs water soluble. Note that the coordinating organic ligands (TOP or TOPO) are retained on the inner surface of QDs, a feature that is important for maintaining the optical properties of QDs and for shielding the core from the outside environment. To achieve binding specificity or targeting abilities, polymer-coated QDs are linked to bioaffinity ligands such as monoclonal antibodies, peptides, oligonucleotides, or small-molecule inhibitors, and also to polyethylene glycols or similar ligands (leading to improved biocompatibility and reduced nonspecific binding).

QUANTUM DOTS FOR MOLECULAR AND CELLULAR IMAGING 183

QD capping iigand TOPO

Amphlphlilc polymer coating

iminlti^ Ifgindls: unillli&dy^, |3€»pd4@, small mciieciuie dttig, inhibilixir, t@tc

Figure 8.1. Structure of a multifunctional QD probe, showing the capping ligand TOPO, an encapsulating copolymer layer, tumor-targeting ligands (such as peptides, antibodies, or small-molecule inhibitors), and polyethylene glycol (PEG)

QD bioconjugation can be achieved by several approaches including passive adsorption, multivalent chelation, or covalent bond formation (Figure 8.2). Two popular cross-linking reactions are carbodiimide-mediated amide formation, and active ester maleimide-mediated amine and sulfhydryl coupling. An advantage for the carboxylate-amine condensation method is that most proteins contain primary amine and carboxylic acid groups, and do not need any chemical modifications prior to QD conjugation. In contrast, free sulfhydryl groups are rare in native biomolecules and are often unstable in the presence of oxygen. Depending on the available chemical groups, other conjugation reactions can also be used. For example, Pellegrino et al. reported the use of a pre-activated amphiphilic polymer for nanoparticle solublization^^. This polymer contains multiple anhydride units, and is highly reactive toward primary amines without addition of coupling reagents. This procedure deserves further attention because polyanhydrides represent a class of biodegradable polymers that are under intense development for use in sustained drug delivery and tissue engineering^^'^^.

Several strategies can be used to manipulate the molecular orientations of the attached ligands as well as their molar ratios with respect to QDs (Figure 8.2). But "perfect" QD probes with precisely controlled ligand orientations and molar ratios are still not available. Mattoussi et al. first explored the use of a fusion protein as an adaptor for IgG antibody coupling^^. The adaptor protein has a positively charged leucien zipper domain for electrostatic interaction with

184 G. RUAN ETAL.

X A \ antibody - ^ ^

antibody

»# j^ ^@f

Sragmants

\X^ Binding site

^ ^ p - a , N H . . . . . . . 4 * - , 4 . w — ^ ^ ^ m

^ A ^ ' Hjs tagged A, I NTA-Ni peptide ^

Figure 8.2. Methods for conjugating QDs to biomolecules. (a) Traditional covalent cross-linking chemistry using EDAC as a catalyst, (b) Conjugation of antibody fragments to QDs via reduced sulfhydryl-amine coupling, (c) Conjugation of antibodies to QDs via an adaptor protein, (d) Conjugation of histidine-tagged peptides and proteins to NTA-Ni modified QDs, with potential control of the attachment site and QD:ligand molar ratios.

QDs, and a protein G domain that binds to the antibody Fc region. As a result, the Fc end of the antibody is connected to the QD surface, with the target-specific F(ab)2 domain facing outward (Fig. 8.2D). In a dramatically different approach, we have linked QDs to a chelating compound (nickel-nitrilotriacetic acid or Ni-NTA) that quantitatively binds to hexahistidine-tagged biomolecules with controlled molar ratio and molecular orientation (Figure 8.2C). Early studies using genetically engineered peptides showed excellent tumor targeting abilities (Gao, Yang, Nie, unpublished data). This indirect "his-tag" coupling method have several advantages such as a controlled or known orientation of the binding ligand (a histidine tag can be conveniently fused to proteins and peptides at a particular site), compact overall probe sizes (which should improve binding efficiencies), and low production costs (direct coupling and rapid purification).


8.3. NOVEL OPTICAL PROPERTIES

As briefly noted above, QDs are made from inorganic semiconductors, and have novel optical properties that can be used to optimize signal-to-background ratios. First, QDs have very large molar extinction coefficients on the order of 0.5 - 5 X 10^ M- cm' , which makes them brighter probes under photon-limited in vivo conditions (where light intensities are severely attenuated by scattering and absorption). In theory, the lifetime-limited emission rates for single quantum dots are 5-10 times lower than those of single organic dyes because of their longer excited state lifetimes (20-50 ns). In practice, however, fluorescence imaging usually operates under absorption-limited conditions, in which the rate of absorption is the main limiting factor of fluorescence emission. Since the molar extinction coefficients of QDs are about 10-50 times larger than that (5-10 x 10" M' cm" ) of organic dyes, the QD absorption rates will be 10-50 times faster than that of organic dyes at the same excitation photon flux (number of incident photons per unit area). Due to this increased rate of light emission, individual QDs have been found to be 10-20 times brighter than organic dyes (Fig. 8.3A)^ ' ^ . In addition, QDs are several thousand times more stable against photobleaching than dye molecules, and are thus well suited for continuous tracking studies over a long period of time.

Second, the longer excited state lifetimes of QDs provide a means for separating the QD fluorescence from background fluorescence, in a technique known as time-domain imaging^ ' ^^. Figure 8.3C shows a comparison of the excited state decay curves of QDs and organic dyes. Assuming that the initial fluorescence intensities of QDs and dyes after a pulse excitation are the same and that the fluorescence lifetime of QDs is one order of magnitude longer, one can estimate that the QD and dye intensity ratio (IqD/Idye) will increase rapidly from 1 at time t = 0 to -100 in only 10 nanoseconds (t = 10 ns). Thus, the image contrast (measured by signal-to-noise or signal-to-background ratios) can be dramatically improved by time-delayed data acquisition.

Third, the large Stokes shifts of QDs (measured by the distance between the excitation and emission peaks) can be used to further improve detection sensitivity. This factor becomes especially important for in-vivo molecular imaging due to the high autofluorescence background often seen in complex biomedical specimens. As shown in Figure 8.3D, the Stokes shifts of semiconductor QDs can be as large as 300-400 nm, depending the wavelength of the excitation light. Organic dye signals with a small Stokes shift are often buried by strong tissue autofluorescence, whereas QD signals with a large Stokes shift are clearly recognizable about the background. This "color contrast" is only available to QD probes because the signals and background can be separated by wavelength-resolved or spectral imaging.

186 G. RUAN ETAL.

t\fm {m\n\, tsmmhum C^ts

0 ^ 40 $s m 100 Am im im tm

^

1 3 w 400

y. ODn^SJ^SiSOurd f V ^ 1 M«KistSklfii

Mm t^m im m^

WaviMfiifigth (nrnj

Figure 8.3. Novel optical properties of QDs for improving the sensitivity of in-vivo bioimaging. (a) Comparison of fluorescence light emission from organic dyes (TRITC - left vial), green QDs (middle vial), and red QDs (right vial) under normal room light illumination and at the same molar concentration (1.0 |iM for dyes and QDs). Bright fluorescence emission is observed from QDs but not from the dye, due to the large absorption cross sections of QDs. (B) Photobleaching curves showing that QDs are several thousand times more photostable than organic dyes under the same excitation conditions, (c) A comparison of the excited state decay curves (monoexponential model) between QDs and common organic dyes. The longer excited state lifetimes of QD probes allow the use of time-domain imaging to discriminate against the background fluorescence (short lifetimes), (d) Comparison of mouse skin and QD emission spectra obtained under the same excitation conditions, demonstrating that the QD signals can be shifted to a spectral region where the autofluorescence is reduced. (See color insert section.)

A further advantage is that multicolor QD probes can be used to image and track multiple molecular targets simultaneously. This is a very important feature because most complex human diseases such as cancer and atherosclerosis involve a large number of genes and proteins. Tracking a panel of molecular markers at the same time will allow scientists to understand, classify, and differentiate complex human diseases^^. Multiple parameter imaging, however, represents a significant challenge for magnetic resonance imaging (MRI), positron emission tomography (PET), computed x-ray tomography (CT), and related imaging modalities. On the other hand, fluorescence optical imaging provides both signal intensity and wavelength information, and multiple wavelengths or colors can be resolved and imaged simultaneously (color imaging). Therefore, different molecular or cellular targets can be tagged with different colors. In this regard, QD probes are particularly attractive because


their broad absorption profiles allow simultaneous excitation of multiple colors, and their emission wavelengths can be continuously tuned by varying particle size and chemical composition. For organ and vascular imaging in which micrometer-sized particles could be used, optically encoded beads (polymer beads embedded with multicolor QDs at controlled ratios) could allow multiplexed molecular profiling in vivo at high sensitivities^^""^ .

8.4. DELIVERY OF QD PROBES INTO CELLS

A number of methods have been reported for delivery of QD probes into living cells, and each of them has its own advantages and limitations, (i) Endocytosis. QDs have been conjugated to endocytosis ligands such as transferrin for transport across the cell membrane. But this method is slow (several hours or longer), and QDs are often trapped in intracellular vesicles and are not available for binding to molecular targets^' ' . (i) Carrier-mediated transfer. Carriers such as cell-penetrating peptides, cationic lipids and dendrimers are able to deliver QDs into cells with high speed and efficiency" "' . However, these carriers are often cationic and cause QD aggregation due to charge-charge interactions. In particular, QD-Tat peptide conjugates were found to enter the nucleus shortly after they were incubated with cells^^. This method deserves fiirther exploration for intracellular delivery of QDs because it has been used successful for delivery and targeting of a variety of macromolecules'*'*' . QD aggregation could be minimized by optimizing the carrier/cargo ratio" . The QD cargo could also be separated from the delivery peptides immediately after cell entry so that the probes will be available for binding to cytoplasmic targets" " ' ^^. (ii) Electroporation. This method can deliver QDs quickly and efficiently into living cells, but its disadvantages are low cell viability and the need for special equipment. In one report, QD probes were found to aggregate after electroporation delivery for reasons that are still not clear" . (iii) Microinjection. This is the most reliable way to deliver QDs into cells without causing QD aggregation. But it is also most labor intensive since one cell is injected at a time, (iv) Toxin deliverv. Work in our own group has shown that an endotoxin Streptolysin-0 (SLO) is able to deliver single QDs into cells (Figure 8.4). SLO is a prototype of the cholesterol-binding family of bacterial exotoxins. Multiple copies of the toxin can insert themselves into the membrane bilayers and produce nanometer-sized pores, allowing macromolecules and nanoparticles to move across the membrane. The maximum size of the particles/molecules that can enter cells by SLO is about 40 nm" . A potential problem with this delivery method is that the number of QDs that can enter cells may be limited, since the delivery mechanism is limited by diffusion.

188 G. RUAN ETAL.

Figure 8.4. Delivery of single QDs into cells by SLO toxin. The image is one frame of a movie taken by a Perkin-Elmer Spinning Disk Confocol microscope. The illumination sources of both bright field (to show the cell) and fluorescence (to show the QDs) were open. The following lines of evidence support that the QDs in cells are single: (a) these QDs have similar brightness and spot size; (b) the brightness of these QDs is not higher than that of single QDs on a coverslip; and (c) the intracellular QDs show intermittent on/off light emission (called blinking), a characteristic of single dot behavior. (G. Ruan, A. Agrawal, and S. Nie, unpublished data, 2005).

Figure 8.5. Immunostaining of F-actin with phalloidin-biotin followed by QD-streptavidin in fixed NIH/3T3 cells. (G. Ruan, A. Agrawal, and S. Nie, unpublished data, 2005).


8.5. APPLICATIONS IN INTRACELLULAR IMAGING

8.5.1 Cellular staining.

In fixed cells, Wu et al. ^ used QDs to label various molecular targets in cell membrane, cytoplasm and nucleus effectively and specifically. The following advantages of QDs compared with organic dyes were demonstrated: brightness, photostability, and multiplexing capability. Figure 8.5 showed an epi-fluorescence image of F-actin in fixed NIH/3T3 cells stained with phalloidin-biotin followed by QD-streptavidin.

In live cells, however, QD delivery and targeting is more challenging because of the following difficulties: (i) delivery of single and monodispersed QDs into cells; (ii) the high viscosity of intracellular environment hinders the moving and binding of QDs; (iii) the QDs unbound to the target cannot be washed away. In one report, Voura et al. introduced QDs into cancer cells by cationic lipids, and tracked the extravasation of these cells in vivo"^^. In another report, Jaiswal et al. labeled D. discoideum cells with QDs by endocytosis, and monitored the cell growth and development on culture plates^. In both studies, the use of cationic lipids or endocytosis led to QD aggregation in cells.

8.5.2 Intracellular studies.

Lidke et al.^ showed that QD-epidermal growth factor (EGF) conjugates binded with EGF receptor (erbBl) on cell membrane and were internalized into cells by endocytosis (Fig. 8.6). The erbB/HER family of transmembrane receptor tyrosin kinases (RTKs) mediate cellular responses to EGF and related ligands. Using QD-EGF conjugates along with erbB 1/2/3 fused with visible fluorescent proteins, Lidke et al. ^ were able to follow the early stages of RTK-dependent signaling in live cells. A new mechanism of retrograde transport to the cell body was revealed. It was also demonstrated that erbB2 but not erbB3 heterodimerized with erbBl after EGF stimulation, thereby modulating EGF-induced signaling.

QDs have also been conjugated with targeting peptides such as nuclear localization sequence (NLS) peptide and mitochondrial localization sequence (MLS) peptide to image organelles in cells (Fig. 8.7)" . However, microinjection had to be used to deliver the QD probes into cells as both carrier-mediated delivery and electroporation caused aggregation.

190 G. RUAN ETAL.

.10§ s »* *\ 0 s •% 13.S s »• 17 s

181 s Figure 8.6. Dynamics of endosomal fusion. Selected time points in the cytoplasm of a CHO cell expressing erbBl-eGFP, 30 min after addition of 200 pM 6:1 QD-EGF. The vesicles show Brownian movement interrupted by directed movement and fusions. At time 0, the lower-right vesicle (red arrow) moved towards the upper-left vesicle (blue arrow), with which it fused irreversibly at 25 s (purple arrow). Scale bar, 5 |xm. This figure demonstrates that QD-EGF can illuminate the RTK-dependent signaling pathway. (Reprinted from ref.7 with permission of the Nature Publishing Group). (See color insert section.)

Figure 8.7. Subcellular localization of QDs. QDs were conjugated to localization sequence peptides, which permit active transport to the nucleus (NLS, A) or mitochondria (MLS, B), and were delivered to 3T3 fibroblast cells by microinjection. A) Fluorescence and phase micrographs of a cell 24h after co-injection of QD-NLS with 70 kDa rhodamine dextran control. The four spots in the nucleus that are not stained with QDs are the nucleoli. B) Fluorescence and phase micrographs 24h after injection of QD-MLS. Colocalization with Mictotracker Red confirms mitochondrial labeling. C) QDs remain fluorescent after 8min of continuous mercury lamp exposure, while conventional MitoTracker dye (D) bleaches beyond detection after 30s of continuous excitation. Different cells were imaged for (C, D). (Reprinted from ref 43 with permission of Wiley-VCH Verlag GmbH & Co KG). (See color insert section.)



This work was supported by grants from the National Institutes of Health, the US Department of Energy Genomes to Life Program, and the Georgia Cancer Coalition (GCC). One of the authors (A. M. S.) acknowledges the Whitaker Foundation for generous fellowship support.

8.7. REFERENCES

1. W. C. W. Chan, D. J. Maxwell, X. H. Gao, R. E. Bailey, M.Y. Han, and S. M. Nie, Luminescent quantum dots for multiplexed biological detection and imaging, Cur. Opin. Biotechnol 13 (1), 40-46 (2002).

2. M. Y. Han, X. H. Gao, J. Z. Su, and S. Nie, Quantum-dot-tagged microbeads for multiplexed optical coding of biomolecules, Nat. Biotechnol. 19 (7), 631-635 (2001).

3. X. H. Gao and S. M. Nie, Doping mesoporous materials with multicolor quantum dots, J. Phys. Chem. B, 107 (42), 11575-11578 (2003).

4. X. H. Gao and S. M. Nie, Quantum dot-encoded mesoporous beads with high brightness and uniformity:Rapid readout using flow cytometry, Anal. Chem. 76 (8), 2406-2410 (2004).

5. X. Wu, H. Liu, J. Liu, K. N. Haley, J. A. Treadway, J. P. Larson, N. Ge, F. Peale, and M. P. Bruchez, Immunofluorescent labeling of cancer marker Her2 and other cellular targets with semiconductor quantum dots, Nat. Biotechnol. 21 (1), 41-46 (2003).

6. M. Dahan, S. Levi, C. Luccardini, P. Rostaing, B. Riveau, and A. Triller, Diffusion dynamics of glycine receptors revealed by single-quantum dot tracking. Science 302 (5644), 442-445 (2003).

7. D. S. Lidke, P. Nagy, R. Heintzmann, D. Amdt-Jovin, J. N. Post, H. E. Grecco, E. A. Jares-Erijman, and T. M. Jovin, Quantum dot ligands provide new insights into erbB/HER receptor-mediated signal transduction, Nat. Biotechnol. 22 (2), 198-203 (2004).

8. Y. Xiao and P.E. Barker, Semiconductor nanocrystal probes for human metaphase chromosomes. Nucleic Acids Res. 32 (3), e28 (2004).

9. J. K. Jaiswal, J. K. Mattoussi, J. M. Mauro, and S. M. Simon, Long-term multiple color imaging of live cells using quantum dot bioconjugates, Nat. Biotechnol. 21 (1), 47-51 (2003).

10. L. Medintz, A. R. Clapp, H. Mattoussi, E.R. Goldman, B. Fisher, and J. M. Mauro, Self-assembled nanoscale biosensors based on quantum dot FRET donors Igor, Nat. Mater. 2 (9), 630-638(2003).

11. L L. Medintz, J. H. Konnert, A. R. Clapp, L Stanish, M. E. Twigg, H. Mattoussi, J. M. Mauro, and J. R. Deschamps, A fluorescence resonance energy transfer-derived structure of a quantum dot-protein bioconjugate nanoassembly, Proc. Natl. Acad. Sci. USA. 101 (26), 9612-9617(2004).

12. S. J. Rosenthal, L Tomlinson, E. M. Adkins, S. Schroeter, S. Adams, L. Swafford, J. McBride, Y. Wang, L. J. DeFelice, and R. D. Blakely, Targeting cell surface receptors with ligand-conjugated nanocrystals, J. Am. Chem. Soc. 124 (17), 4586-4594 (2002).

13. A. P. Alivisatos, Semiconductor clusters, nanocrystals, and quantum dots. Science. 271 (5251), 933-937 (1996).

14. A. P. Alivisatos, The use of nanocrystals in biological detection, Nat. Biotechnol. 22 (1), 47 - 52 (2004).

15. L. Manna, D. J. Milliron, A. Meisel, E. C. Scher, and A. P. Alivisatos, Controlled growth of tetrapod-branched inorganic nanocrystals, Nat. Mater. 2 (6), 382-385 (2003).

16. D. J. Milliron, S. M. Hughes, Y. Cui, L. Manna, J. Li, L. W. Wang, and A. P. Alivisatos, Colloidal nanocrystal heterostructures with linear and branched topology, Nature. 430 (6996), 190-195(2004).

17. K. A. Dick, K. Deppert, M. W. Larsson, T. Martensson, W. Seifert, L. R. Wallenberg, and L. Samuelson, Synthesis of branched 'nanotrees' by controlled seeding of multiple branching events, Nat. Mater. 3 (6), 380-384 (2004).

192 G. RUAN ETAL.

18. W. W. Yu, Y. A. Wang, and X. G. Peng, Formation and stability of size-, shape-, and structure-controlled CdTe nanocrystals: Ligand effects on monomers and nanocrystals, Chem. Mater. 15 (22), 4300-4308 (2003).

19. M. A. Hines and P. Guyot-Sionnest, Synthesis and characterization of strongly luminescing ZnS-capped CdSe nanocrystals, J. Phys. Chem. 100 (2), 468-471 (1996).

20. X. G. Peng, M. C. Schlamp, A. V. Kadavanich, and A. P. Alivisatos, Epitaxial growth of highly luminescent CdSe/CdS core/shell nanocrystals with photostability and electronic accessibility, J. Am. Chem. Soc. 119 (30), 7019-7029 (1997).

21. B. O. Dabbousi, J. RodriguezViejo, F. V. Mikulec, J. R. Heine, H. Mattoussi, R. Ober, K. F. Jensen, and M. G. Bawendi, (CdSe)ZnS core-shell quantum dots: Synthesis and characterization of a size series of highly luminescent nanocrystallites, J. Phys. Chem. 101 (46), 9463-9475 (1997).

22. R. E. Bailey and S. M. Nie, Alloyed semiconductor quantum dots: Tuning the optical properties without changing the particle size, J. Am. Chem. Soc. 125 (23), 7100-7106 (2003).

23. L. H. Qu and X. G. Peng, Control of photoluminescence properties of CdSe nanocrystals in growth, J. Am. Chem. Soc. 124 (9), 2049-2055 (2002).

24. B. Dubertret, P. Skourides, D. J. Norris, V. Noireaux, A. H. Brivanlou, and A. Libchaber, In vivo imaging of quantum dots encapsulated in phospholipid micelles, Science 298 (5599), 1759-1762(2002).

25. X. H. Gao, Y. Cui, R. M. Levenson, L. W. K. Chung, and S. Nie, In vivo cancer targeting and imaging with semiconductor quantum dots, Nat. Biotechnol. 11 (8), 969-976 (2004).

26. T. Pellegrino, L. Manna, S. Kudera, T. Liedl, D. Koktysh, A. L. Rogach, S. Keller, J. Radler, G. Natile, and W. J. Parak, Hydrophobic nanocrystals coated with an amphiphilic polymer shell:A general route to water soluble nanocrystals, Nano Lett. 4 (4), 703-707 (2004).

27. E. Ron, T. Turek, and E. Mathiowitz, Controlled release of polypeptides from polyanhydrides, Proc. Natl. Acad. Sci. USA. 90 (9), 4176-4180 (1993).

28. K. S. Anseth, V. R. Shastri, and R. Langer, Photopolymerizable degradable polyanhydrides with osteocompatibility, Nat. Biotechnol. 17 (2), 156-159 (1999).

29. H. Mattoussi, J. M. Mauro, E. R. Goldman, G. P. Anderson, V. C. Sundar, F. V. Mikulec, and M. G. Bawendi, Self-assembly of CdSe-ZnS quantum dot bioconjugates using an engineered recombinant protein, y. Am. Chem. Soc. 122 (49), 12142-12150 (2000).

30. C. A. Leatherdale, W. K. Woo, F. V. Mikulec, and M. G. Bawendi, On the absorption cross section of CdSe nanocrystal quantum dots, J. Phys. Chem. B. 106 (31), 7619-7622 (2002).

31. M. Bruchez, M. Moronne, P. Gin, S. Weiss, and A. P. Alivisatos, Semiconductor nanocrystals as fluorescent biological labels, Science. 281 (5385), 2016-2018 (1998).

32. W. C. W. Chan and S. Nie, Quantum dot bioconjugates for ultrasensitive nonisotopic detection. Science. 281 (5385), 2016-2018 (1998).

33. S. Jakobs, V. Subramaniam, A. Schonle, T. M. Jovin, and S. W. Hell, EGFP and DsRed expressing cultures of Escherichia coli imaged by confocal, two-photon and fluorescence lifetime microscopy, FEBSLett. 479 (3), 131-135 (2000).

34. R. Pepperkok, A. Squire, S. Geley, and P. I. H. Bastiaens, Simultaneous detection of multiple green fluorescent proteins in live cells by fluorescence lifetime imaging microscopy. Cur. Bio. 9 (5), 269-272 (1999).

35. X. H. Gao and S. M. Nie, Molecular profiling of single cells and tissue specimens with quantum dots, Trends Biotechnol. 21 (9), 371-373 (2003).

36. K. Ogawara, M. Yoshida, K. Higaki, T. Kimura, K. Shiraishi, M. Nishikawa, Y. Takakura, and M. Hashida, Hepatic uptake of polystyrene microspheres in rats: Effect of particle size on intrahepatic distribution, J. Control. Release. 59 (1), 15-22 (1999).

37. S. Flacke, S. Fischer, M. J. Scott, R. J. Fuhrhop, J. S. Allen, M. McLean, P. Winter, G. A. Sicard, P. J. Gaffney, S. A. Wickline, and G. M. Lanza, Novel MRI contrast agent for molecular imaging of fibrin implications for detecting vulnerable plaques. Circulation. 104 (11), 1280-1285(2001).

38. L. C. Katz, A. Burkhalter, and W. J. Dreyer, Fluorescent Latex Microspheres as a Retrograde Neuronal Marker For In Vivo and In Vitro Studies of Visual-Cortex, Nature. 310 (5977), 498-500(1984).

39. G. L. Chien, C. G. Anselone, R. F. Davis, and D. M. Van Winkle, Fluorescent vs. radioactive microsphere measurement of regional myocardial blood flow, Cardiovasc. Res. 30 (3), 405-412(1995).


40. R. Pasqualini and E. Ruoslahti, Organ targeting in vivo using phage display peptide libraries, Nature. 380 (6572), 364-366 (1996).

41. F. Chen and D. Gerion, Fluorescent CdSe/ZnS nanocrystal-peptide conjugates for long-term, nontoxic imaging and nuclear targeting in living cells, Nano Lett. 4(10), 1827-1832 (2004).

42. A. Hoshino, K. Fujioka, T. Oku, S. Nakamura, M. Suga, Y. Yamaguchi, K. Suzuki, M. Yasuhara, and K. Yamamoto, Quantum dots targeted to the assigned organelle in living cells, Microbiol. Immunol 48 (12), 985-994 (2004).

43. A. M. Derfus, W. C. W. Chan, and S. N. Bhatia, Intracellular delivery of quantum dots for live cell labeling and organelle tracking. Adv. Mater. 16 (12), 961-966 (2004).

44. A. Joliot and A. Prochiantz, Transduction peptides: from technology to physiology, Nat. Cell Biol. 6 (3), 189-196(2004).

45. M. C. Morris, J. Depollier, J. Mery, F. Heitz, and G. Divita, A peptide carrier for the delivery of biologically active proteins into mammalian cells, Nat. Biotechnol. 19 (12), 1173-1176 (2001).

46. N. Nitin, P. J. Santangelo, G. Kim, S. M. Nie, and G. Bao, Peptide-linked molecular beacons for efficient delivery and rapid mRNA detection in living cells. Nucleic Acids Res. 32 (6), e58 (2004).

47. I. Walev, S. C. Bhakdi, F. Hofmann, N. Djonder, A. Valeva, K. Aktories, and S. Bhakdi, Delivery of proteins into living cells by reversible membrane permeabilization with Streptolysin-0, Proc. Natl. Acad Sci. USA 98 (6), 3185-3190 (2001).

48. E. B. Voura, J. K. Jaiswal, H. Mattoussi, and S. M. Simon, Tracking metastatic tumor cell extravasation with quantum dot nanocrystals and fluorescence emission-scanning microscopy, Nat. Med 10 (9), 993-998 (2004).

MOLECULAR ANALYSIS USING MICRO?ARTICLE-BASED FLOW CYTOMETRY

John P. Nolan^

9.1. INTRODUCTION

A major goal for biomedical research in the 21st century will be to integrate molecular information at the level of genes, proteins, and their substrates and products into working models of cell and organism function from which predictions can be made. The realization of this goal requires a significant change in focus for experimental biology from one that focuses on individual molecules to one that includes the interactions of many different molecules as they function in networks of biochemical pathways in living systems. The rationale for pursuing such an ambitious goal stems from the very significant advances in molecular analysis that enable the sequencing of whole genomes, the highly parallel analysis of gene expression levels, and large scale identification of proteins in hcomplex samples. These advances resulted from new molecular reagents and assay chemistries, new instrumentation with improved sensitivity and throughput, new computational tools, and new conceptual approaches to approaching biological problems.

However, just as these new technologies have enabled the rapid acceleration of data collection and interpretation, continued progress toward transforming this information into biological understanding is dependent on continued improvement in analytical technologies. In particular, it is critical to augment qualitative analysis methods that allow the identification of important molecules with quantitative measurements of their abundance and function. The ability to make quantitative measurements of the concentrations of many individual proteins, their interactions and the formation of macromolecular assemblies, and the measurement of these assemblies in live cells and organisms represent major challenges to understanding the systems of molecular networks and pathways that underlie physiology and disease. These

1 John P. Nolan, La JoUa Bioengineering Institute, La JoUa, California, 92037

195

196 J. P. NOLAN

challenges are being addressed through developments in microscopic imaging, optical spectroscopy, and mass spectrometry. In this article, I review and highlight the applications of an emerging platform for high throughput and high resolution molecular analysis: microparticle-based flow cytometry.

9.2. OPTICAL MEASUREMENTS USING FLOW CYTOMETRY

The term flow cytometry refers the measurement of the optical properties of single particles in a flowing sample stream. The name originates from its initial application to the study of individuals cells, and the emergence of an industry that develops and markets instruments and reagents for this purpose has established the term in general use, although several important applications, including those described in this paper, are concerned with the measurement of microspheres, molecules, and other particles rather than cells. Flow microfluorimetry, a term which was used in the early days of the development of this technology might be a more appropriate general name, although recent developments demonstrate that the power of flow-based analysis is not limed to fluorescence signals, or even to optical signals. However, because the widely available instruments that can be used to perform the molecular analyses I describe here are called flow cytometers, these biological applications are considered sub-categories of flow cytometry.

In flow cytometry (Figure 9.1), liquid samples are introduced into the instrument such that a sheath fluid hydrodynamically focuses the sample into a narrow stream. This sheath fluid carries the sample particles single file through a focused laser beam so that the sample is interrogated in a small probe volume defined by the dimensions of the sample stream and laser spot, generally less than twenty picoliters. Light collected from this probe is spectrally filtered, typically with optical filters and mirrors, and detected with a photodetector such as a photomultiplier or photodiode. Even the most basic commercial instruments will measure fluorescence from multiple spectral bands, as well as laser light scatter, and high-end instruments can be configured with multiple lasers and detectors to detect more than a dozen distinct fluorophores from an individual particle. For most commercial instruments, particles spend only a few microseconds in the laser beam, allowing the analysis of thousands of particles per second.

The ability of flow cytometers to rapidly measure multiple optical signals from samples in very small probe volumes provides for some unique analytical features that enable not only the cellular analyses that are now routine, but also the emerging microparticle-based applications that are the topic of this article. First, the small probe volume ensures that background signals are low, allowing for very sensitive fluorescence measurements, down to a single fluorescent molecule in specially configured instruments^'^. This small probe volume also enables the instrument to resolve particle-bound fluorescent probe from free fluorophore over a wide fluorophore concentration range. This allows the configuration of homogeneous assays in which there is no need to wash

CROPARTICLE-BASED FLOW CYTOMETRY 197

away unbound probe before analysis, a great advantage for many kinetic, high throughput, and high resolution molecular measurements. Finally, the ability to simultaneously measure multiple optical signals from individual particles and high rates provides the opportunity to encode particles for multiplexed analysis applications

Sample tube

Photomultiplier tubes (PMTs)

Figure 9.1. Schematic Diagram of a Flow Cytometer

9.3. SOLID PHASE ASSAYS USING MICROPARTICLES

The potential of particle-based analysis using flow cytometry have been recognized for nearly three decades. In the 1970's, it was proposed that microspheres coated with antigens could be used to capture antibodies from circulation for detection using flow cytometry, and that by using different sized microspheres that could be distinguished by their light scatter, it was possible to detect the presence of antibodies to multiple antigens simultaneously in a sample" . This potential was partially realized in the 1980's and early 1990's in the development of a number of immunoassays targeting one to a few analytes^'^. In parallel, it was recognized that microspheres could be functionalized with a range of biomolecules, including proteins, nucleic acids, and lipids which led to the demonstration of a range of molecular analyses^. Recent advances in biomolecule conjugation and in particle encoding are spurring renewed development of microparticle-based flow cytometry.

198 J. P. NOLAN

©

Immobiilzatfon Fluorescent labels DNA: avidln-blotin, covalent Protein: affinity tags Lipid: adsorption, self assembly

synthetic, enzymatic chemical, expressed synthetic, enzymatic

Figure 9.2. Microsphere-based Molecular Analysis.

A generalized scheme for microparticle-based molecular.analysis is presented in Figure 9.2. A biomolecule of interest (A) is immobilized on the particle surface, and one or more interacting molecules (B, C, and D) may be labeled in solution. Interaction between A and one of its interacting partners can be detected as a change in the amount of label associated with the particle.

As will be described, there are many variations of this simplest scheme, but it serves to point out how developments in molecular biology and biochemistry have expanded the options for immobilizing and labeling biomolecules for this type of analysis. For immobilization, the simplest and most widely used approach is to immobilize biomolecules via covalent attachment chemistries such as carbodiimide based- crosslinking of primary amino groups to carboxyl groups. While this approach is often adequate for immobilizing antibodies, it can result in the loss of activity for less stable proteins. Expression of proteins as epitope tag fusions, which is often used to aid in purification, provides a convenient handle for the site-directed, noncovalent immobilization of a protein. Proteins expressed as his-tagged, biotinylated, or glutathione-S-transferase fusions may be immobilized on metal-, avidin-, or glutathione-coated particles, and these are becoming available as commercial reagents. For the labeling of soluble biomolecules, covalent labeling with small organic fluorophores is generally successful, but expression of proteins as fluorescent protein-^ or FlAsH tagged^ genetic fusions expands the options available to label less stable proteins.


o

a> N

CO,

1 O

CO sz

§ £

^

R1

R6

E]

R2 R3 R4 R5

(Diii R7 R8 R9 R10

e®0®

^

o o C "" (D 0)

LL T3

°o 10 10^ 10^

Red Fluorescence 10 10'' 10-"

Orange Fluorescence

Figure 9.3. Optical encoding schemes for microsphere-based assays. Discrete intensities of one color of fluorescence can be combined with different sized beads, which produce discrete light scatter intensities, or with a second color of fluorescence to enable the identification of many particle populations within a sample.

The other major factor stimulating microparticle-based molecular analysis is the development of optically encoded microparticles for multiplexed analysis. While duplex and triplex assays had been demonstrated using different sized beads, the introduction of microsphere sets bearing different amounts of dye that gave discrete levels of fluorescence^ ' ^ increased the multiplexing capacity by more than an order of magnitude, and potentially much more. Different intensities of a single dye can allow approximately ten different beads to be distinguished. Size (light scatter) can be used to increase this number by a factor of two to four, while the addition of another dye with as many as ten discrete intensities of a second color expands the multiplex set to nearly 100. Sets of optically encoded microparticles, sometime referred to as suspension arrays, thus constitute a variant of the popular microarray technology, in which the surface on which an assay is performed is defined not by a physical spot on a two dimensional surface, but by the features of the encoding elements it exhibits. Pioneered by Luminex Corp (Austin, TX), encoded microspheres, assay kits, and dedicated microsphere reading instruments are now available from a number of commercial sources. An attractive feature of this platform is its open nature, in that beads and assay kits from any manufacturer can in principle be used with any flow cytometer with the appropriate detection capabilities. While the most popular application for microsphere-based flow cytometry is the ELISA-like sandwich immunoassay, the number of other potential applications is quite large.

200 J. P. NOLAN

9.4. DETECTION AND SENSOR APPLICATIONS

A key step in realizing the potential of multiplexed analysis using encoded microparticle arrays is the development of quantitative single bead assays for the analysis of molecular interactions and functions. As mentioned above, the first and still the most popular multiplexed microparticle application is the sandwich immunoassay, in which a soluble analyte is captured by an immobilized antibody and detected with a second, labeled, antibody resulting in the formation of an antibody-analyte-antibody complex on the particle surface and an increase in signal from the microsphere. This approach, which is analogous to the widely used ELISA assay, has advantages over that method in that it can be performed with fewer or no wash steps and, more importantly, can be used to measure multiple analytes in a single sample using encoded particles.

A recent example of this approach applied to infectious disease detection is presented in Figure 9.4. To detect and differentiate between circulating influenza virus types, monoclonal antibodies to influenza types A and B immobilized on different sized beads captured virus particles from a sample. The antibody-bound virus was then detected with a cocktail of fluorescence-labeled polyclonal antibodies against the A and B type viruses. Using this approach, the two virus types could be uniquely identified with a limit of detection of less than 1 ng of viral protein (Figure 9.5), more than 100-times less than required by two commercial point of care kits'^. This approach has also been adapted to fluorescently encoded beads^^ and is being adapted to detect influenza virus sub-types as well. This immunoassay sandwich format has also been applied to the detection of other pathogens''^'^^.

^ • 1 ^ MAbtoFluA

^> — - h )

94m P^tonuA^ tMfid Flu Vlru» A

O

^ ^ ^ MAbtoFhtll ^

^ ^ ^ '^^'~ r'^ FluowplK»r»-eor^u9»tw» 4.5 |im PAI> to Flu A*8

5.4 Mm

t 4.5 j m

FiuVlru«8 F8C-H

Figure 9.4. Schematic of a duplexed influenza virus typing assay. Capture antibodies specific for different types of influenza virus are immobilized on different sized microspheres. After incubation with the sample and capture of the virus particles, fluorescence-labeled anti influenza virus reporter antibodies are used to detect the captured virus. The particles are then analyzed by flow cytometry to identify the particles on the basis of their light scatter intensity and to measure the amount of reporter antibody associated with each bead. Reprinted with permission from reference 12.


800

1 a 700

600-j

^ 500

i , 400*1 ! J u.

300

200-

Oui exsd aissay, msdiing on mHiumma A specific b«a(te (54 pm)

«—A/Klw/301/94(H3N2)

*—t^^s r t :

0.57 ng

800

6Q0H

200H

0.01 0.1 1 10

VIraJ protein ( ^ n % (ng)

Duplexed a»$ay. maoKrig on

*—AA<wv/30im (H3M2)

0.5? na

I ,^

0.01 0.1 1 10 viral prols»i qudn^ty (ng)

Figure 9.5. Influenza virus typing assay perfonnance. Reprinted with permission from reference 12.

While the antibody sandwich format is well established, there is a great deal of interest in developing assay chemistries that are simpler, faster, and more sensitive. Towards this end, the specific binding of receptors by proteins has been exploited for the detection of bacterial toxins. Cholera toxin, a multi-subunit protein secreted by the pathogenic bacteria Vibrio cholerae, binds to specific glycolipid receptors on mammalian cells in order to gain entry to the cell. This receptor, the ganglioside Gml, can move laterally within the plane of the cell membrane and is aggregated by the pentavalent cholera toxin. This aggregation forms the basis of a very sensitive FRET-based assay for cholera toxin^ " . Fluorescently labeled analogues of Gml were synthesized with a fluorescence donor (BODIPY-TMR) or acceptor (BODIPY-TR) and were incorporated at low mole fractions into artificial membrane vesicles (liposomes), which were then assembled onto the surface of glass microspheres. Upon excitation at 514 nm in a flow cytometer, the supported bilayer membranes containing the fluorescent Gml analogues exhibited significant BODIPY-TMR fluorescence, but low levels of BODIPY-TR fluorescence, as that dye is inefficiently excited at 514 nm. Upon addition of cholera toxin (or its B subunit, which contains the receptor binding activity) the GMl analogues were brought into proximity such that energy transfer between BODIPY-TMR and BODIPY-TR occurs, resulting in a decrease in the fluorescence of the donor and an increase in the fluorescence of the acceptor (Figure 9.6). The ratio of the acceptor to donor fluorescence was proportional to the amount of cholera toxin present, and it was possible to detect less than 50 pM with only a five minute incubation (Figure 9.7). This sensitivity is comparable to the traditional ELISA immunoassay, but is much simpler as it involves only a single reagent

202 J. P. NOLAN

200000

Wavefength (nm)

Figure 9.6. Receptor aggregation-induced FRET. Cholera toxin at the indicated concentrations was added to microsphere-supported bilayers bearing BODIPY-TMR and BODIPY-TR GMl analogues, and the emission spectra measured by spectrofluorimetry (excitation wavelength: 530 nm).Reprinted with permission from reference 17.

(the receptor-bearing microspheres) and no washing or other sample processing steps. Adapted for other receptors using different-sized beads, this assay approach could in principle be used for single step detection of multiple analytes.

9.5. MOLECULAR INTERACTIONS AND FUNCTION

The features of sensitive, homogeneous detection have been applied to the analysis of molecular interactions and function for a range of different systems. Here, I present some examples that illustrate some of the key approaches and their advantages relative to other methods.

9.5.1 Enzyme-Substrate Interactions

The interactions between enzymes and substrates play key roles in health and disease, and high resolution methods for their analysis are key to understanding of basic mechanisms as well as for the development of potential therapeutic agents. The structure specific nuclease human flap endonuclease-1 (FEN-1) recognizes and cleaves a particular DNA structure characterized by a double stranded character with a single stranded 5' flap. This flap structure is thought to be an intermediate in DNA replication and repair, and mutations in FEN-1 have been associated with defects in DNA replication associated with


i

i DC,

O

vc £

1.0-

0.8-

0.6-

0.4.

0,2-

0.0-

'0.2-

1 ' —

J y

^ X

..ny 1

, r r 1 ^ 1 i » f ^ 1 r T 1

• ' / A / ,•" -1

' J

/' A . J

/ , •* 1

v

X

,x

T • 1 ' 1 ' —

4

Bead [BTMROM1MBTRGM1J H

m 0Jmg/3nM/3nM [J

A 0.3mg / 6ftM / 6nM n

X 0.3mg/!2nM/l2fiM H

n ' 1 »-—r—-^ r—-1

CT (nM)

Figure 9.7. Sensitive detection of cholera toxin using receptor induced aggregation on microspheres. The indicated concentrations of cholera toxin was incubated with microsphere-supported bilayers containing different amounts of fluorescent GMl analogues for 5 minutes, followed by analysis using flow cytometry. Reprinted with permission from reference 17.

certain types of cancer^^. Thus, the mechanisms of action of this enzyme are of importance in understanding human disease. The traditional approach to measuring FEN-1 activity involves the generation of a synthetic oligonucleotide substrate with a radioactive label at the 5' end of the flap strand, and using gel electrophoresis to resolve the intact substrate from the shorter, cleaved product. This approach is slow and labor intensive, and not well suited to kinetic analysis and mechanistic studies of enzyme structure and function.

To develop an improved approach for these types of studies we replaced the radioactive label with a fluorophore, added a biotin to one of the ends, and immobilized the flap substrate on the surface of a streptavidin microsphere (Figure 9.8). When the enzyme FEN-1 was added in the presence of its essential cofactor Mg^^, the flap strand bearing the fluorophore was cleaved, releasing the fluorophore and resulting in a decrease in microsphere fluorescence. Because the flow cytometer can resolve free from microsphere-bound fluorophore, it was possible to measure this decrease continuously, without the need for a wash step.

Apart from the homogeneous assay format, which enables continuous kinetic resolution, the microsphere based approach has important advantages in sensitivity of other methods. In a typical assay, approximately 50,000 substrate molecules were immobilized per bead, which represents a fairly bright signal

204 J. P. NOLAN

Figure 9. 8. Schematic of the flap DNA substrate. The synthetic substrate composed of a biotinylated template strand, fluorescence-labeled flap strand, and invading adjacent strand, is attached to a streptavidin microsphere.

for the flow cytometer. However, at the microsphere concentration used, typically 500,000 per ml, this translates to a solution substrate concentration of ~ 10 pM, 2-3 orders of magnitude less than is practical with radioisotope labeling. This allowed us to provide enzyme in excess to perform pre-steady state, single turnover kinetic analysis which are particularly powerful kinetic approaches.

This combination of sensitivity and continuous kinetic resolution allowed us to easily perform kinetic analysis that would have been difficult using radioisotope methods. For example, we could readily determine the enzyme concentration dependence of the overall reaction time course (Figure 9.9A). We could evaluate the cleavage step independently from the binding step by allowing the enzyme and substrate to bind in the absence of Mg^ , and the start the reaction by the addition of Mg^^. These studies were performed using a specially-constructed stopped flow cytometer that was able to mix reagents and initiate measurements in less than 0.3 seconds, allowing accurate estimation of the cleavage rate constant (Figure 9.9B).

To examine the binding step in detail, we took advantage of a poly-his epitope tag at the amino terminus of the FEN-1 enzyme to immobilize it on the surface of Ni^^-chelate beads (Figure 9.10A). In the absence of Mg^ , it was possible to measure the kinetics of substrate binding to the immobilized enzyme (Figure 9.1 OB). Addition of Mg^^ resulted in cleavage of the substrate and loss of fluorescence from the microsphere. We could then use these independent measurements of binding and cleavage in combination with the overall reaction time course and a kinetic model of the reaction mechanism to estimate the rate constants for enzyme-substrate association, dissociation, and cleavage^^\ Apart from providing detailed information about the FEN-1 reaction mechanisms, the microsphere-based assay approach also provided a convenient tool to investigate the role of individual amino acid residues in FEN-1 function " "


70 ..

1 ^^"^h ''*'* '

»00 200 300 400 50O 600 TIME <:SECONDS)

'f.*«**AW*«f*

V 20 30 40 50 «0 70 «0 90

Tifcte (SCCONDS)

Figure 9.9. Measurement of enzyme-substrate binding and catalysis. Microspheres bearing a fluorescent flap DNA substrate were incubated with (A) the indicated concentrations of FEN 1 in the presence of Mg++ to measure the kinetics of binding and cleavage or (B) a saturating concentraton of FEN 1 in the absence of Mg++ followed by the addition of MgC12 to measure the kinetics of cleavage. Reprinted with permission from reference 20.

The example of the study of FEN-l's interaction with its substrate serves to illustrate several configurations of the bead-based platform. Similar approaches have been used to study the interactions of ligands and receptors 25-

receptors and G proteins ' , and proteases and their substrates.

9.5.2 Ligand-Receptor Interactions

The study of ligand-receptor interactions poses a particular challenge because cell surface receptors most often are in a lipid membrane environment that is critical for their function. This is true of both protein and glycoprotein receptors and glycolipid receptors. In order to better understand the mechanistic details of bacterial toxin binding to cells, we employed the microsphere-based approach in conjunction with supported lipid bilayers bearing specific toxin receptors. Cholera toxin recognizes the glycolipid ganglioside GMl to bind with high avidity to cell membranes. This interaction has been studied in vitro with surface plasmon resonance (SPR) based measurements, but these studies were unable to extensively investigate the receptor concentration because of the sensitivity of SPR. These studies also used simple single site reversible binding models, even though it is well known that the toxin is multivalent (see Section 3 above). To overcome some of these limitations we prepared microsphere-supported bilayers bearing mole fractions of Gml ranging from 1 mole % to 0.01 mole %, and measured the binding of fluorescent cholera toxin using flow cytometry^^.

206 J. P. NOLAN

20000

60 120 180 240

TIME (SECONDS)

Figure 9.10. Enzyme-substrate binding. His-tagged FENl was immobilized on the surface of Ni-chelate michrospheres (A), and incubated with increasing concentrations of fluorescent flap DNA substrate in the absence of Mg++ (B).

0.26

*" 0.20

0.16

S7 0.10

0.05

0.00 10 20 30

[CTBioiai(Mx10-*)

40 10 20 30

(RTC.CTBltota,(Mx10-»)

Figure 9.11. Binding of fluorescent cholera toxin B subunit to GMl-bearing microsphere supported bilayers. (A) Equilibrium binding. (B) Competitive binding of unlabeled cholera toxin B subunit. Reprinted with permission from reference 35

Binding of the fluorescent cholera toxin B subunit (F-CTB) to GMl-containing supported bilayers is saturable and specific, with little non-specific binding to supported bilayers lacking GMl (Figure 9.11 A), and is inhibited when beads are preincubated with unlabeled B subunit (CTB; Figure 9.1 IB). The kinetics of binding at several GMl and F-CTB concentrations (Figure 12) were determined and analyzed a using a kinetic model that considered the pentavalent stoichiometry of the CTB-GMl interaction. The dissociation kinetics were also analyzed and these data were used to estimate the individual rate constants in the multistep binding mechanism and to predict and demonstrate incubation conditions that would result in full and partial occupancy of CTB by the receptor GMl. These results are the highest resolution analysis of the cholera toxin-GMl interaction to date, and was made possible by the unique combination of quantitative sensitivity and continuous kinetic resolution provided by flow cytometry.


100 20O 300 400 600

Time (s)

100 200 300 400 500

T\fm (s)

100 200 300 400 500

Time (s)

100 200 300 400 500

Time {$}

Figure 9.12. Kinetics of F-CTB binding to GMl supported bilayers. F-CTB (2, 5, 10, 20 nM) was added to microsphere-supported bilayers containing 0.01 (A), 0.05 (B), 0.1 (C) or 0.5 (D) mole % GMl and the kinetics of binding measured using flow cytometry. The solid lines represent the results of a simultaneous fit of a multi-step binding model to the complete data set.. Reprinted with permission from reference 35.

9.5.3. Protein Immobilization

A general challenge for molecular assembly applications of microparticle-based flow cytometry involves the immobilization of biomolecules on the surface of microparticles in an active state for binding or activity studies. To address this we sought to further develop the approach taken for analysis of FEN-1, namely the immobilization of his tagged proteins on metal chelate microspheres. We investigated two approaches to preparing metal chelate microspheres^^. One (Figure 9.13A) uses Ni2+-nitrilotriacetic acid (Ni-NTA) covalently attached to the surface of a polystyrene microsphere. The other (Figure 9.13B) uses a NI-NTA-conjugated lipid incorporated into a bilayer membrane supported on glass microspheres. The protein binding and flow cytometry.characteristics of these two surfaces was characterized using a his-tagged version of the enhanced green fluorescent protein (EGFP).

208 J. P. NOLAN

Arrifrfified products

SBE Labeled Primers <MNTPs

-l\

TangetDNA

Extended primers

A

'X '3 A primers

Bead with captured

Figure 9.13. Single base extension reaction using microspheres and flow cytometry.

We found that his-EGFP showed specific binding to both covalently attached NI-NTA polystyrene and silica microspheres, as well as to silica microspheres bearing supported blilayer membranes containing Ni-NTA lipid (Figure 9.14). The silica microspheres showed relatively less nonspecific binding than the hydrophobic polymer beads. Washing and competitive binding studies showed that the his-EGFP was stably attached to the microspheres after washing ft)r a period of hours, but that the addition of competing his-proteins or imidizole could rapidly displaced the bound protein. Thus, these microsphere reagents are suitable for short term studies, but that improved his-binding ligands are required for long term binding stability.

9.6. GENETIC ANALYSIS

An area that is in particular need of high throughput and high resolution analysis tools is that of genetics and genomics. For both basic and clinical research, methods to readily analyze many genetic features in many samples are in great demand. The flat surface microarrays (DNA chips) are effective for the analysis of large number of genetic features in small numbers of samples, but are less well suited to large numbers of samples. Conversely, microplate based methods are efficient for analyzing a single genetic feature in many samples, but less so when it is necessary to analyze many sites in a genome. Multiplexed analysis using encoded microparticle arrays effectively span this middle ground where many sites must be analyzed in many samples.

CROP ARTICLE-BASED FLOW CYTOMETRY 209

8.0e+4 r ? |A

1 Jc 6.0e*4 1

SL 8 i *

1 10 2.0e^^ i

S "*" 0.0 ^ « •

0 10

J S" 1.0^+6 g b 8.0e*5

H «i 6.0«+5

X J2 4.0e+S i / S. 2.0e+5 ,

«»»«*• * 0.0 f 0 10

• * • • »

^ V

•»

. • • . . « . . . • . • .

20 30 40 50 60

• , . * ^ ' * ^

1 « « « « « »

20 30 40 50 60

V

• 70

•

• 70

f

> 80

•

•

80

3.5e+6 1

a-oe+e 1 ® ^ * * 2.5e+6 i ^ * *

2.0e^6 * : •

1.$e-f-$ j * 1.0e*6 y

0 , 0 4 » » • • • - • • • • • • • • • • • • - ^

0 10 20 30 40 50 60

ps^GFP|^ , (nM)

2 .00-1^6 p.............v............— 0

\ * # * * 1,5e-MI 1 •

* 1.0e*6 i m

\ * 6.0e+5 1

\* QJQ ikm-p*- • • • # • • • •'••••

0 10 20 30 40 50 60

*

...^--70

*

• • • • • - •

70

4 1

--4 80

•

1

4 80

Figure 9.14. Binding of his-EGFP to Ni-NTA microspheres. Saturable, high affinity binding of his-EGFP to microspheres with (A, B) covalently attached Ni-NTA polystyrene microspheres (prepared (A) as in (ref x) and (B) commercial Beadlyte microspheres), (C) ) covalently attached Ni-NTA silica microspheres, or (D) Ni-NTA lipid in silica microsphere-supported bilayer membranes. Reprinted with permission from reference 36.

A case in point is the analysis of single nucleotide polymorphisms (or SNPs) which are locations in the genome where a single nucleotide base varies among individuals^^. It is estimated that the human genome contains more than one million SNPs, and that the identity of the nucleotide base at these sites can be used to predict disease susceptibility or drug response^^. For this reason there is great interest in methods that can efficiently genotype SNPs in patient samples^^, and we sought to apply the advantaged of multiplexed analysis using encoded microparticle arrays. There are several assay chemistries that have been developed to genotype SNPs. AUele-specific hybridization is widely used, and when optimized is highly effective. However, optimal conditions depend on local sequence context, and it can be difficult to identify conditions for the analysis of many different SNPs simultaneously. An alternate approach is the use of enzymes, such as DNA polymerase or DNA ligase, which have the ability to discriminate single bases with high specificity and little dependence on local sequence context. One approach, single base extension (SBE) of an oligonucleotide primer, has been shown to be effective in a variety of different assay formats, including solution-based assays, microwell plates, and flat microarrays.

210 J. P. NOLAN

In order to configure the SBE approach for use with encoded microparticles, while taking advantage of solution-phase kinetics, we devised a system of bead-based address tags, and SNP-specific capture tagged primers (Figure 9.13). The primers are incubated with the sample DNA and anneal directly adjacent to the site of interest. In the presence of labeled dideoxynucleotide analogues, the annealed primer will be extended by exactly on base, and the base that is incorporated will reveal the identity of the complementary base on the sample DNA strand. This reaction is similar to that used in DNA sequencing, and is very high fidelity. More over, it can be performed in a thermal cycler with a thermal-stable enzyme allowing each strand of sample DNA to be interrogated repeatedly by multiple primers, thereby amplifying the signal. After the primer extension phase, the SNP-specific primers for each SNP are captured by hybridization onto a unique population of encoded microspheres via the 5' capture tag, and analyzed by flow cytometry. The use of microsphere-bound address tags and primer-bound capture tags allows reactions to be performed in solution, where hybridization to the large sample DNA molecules is most efficient, and allows one set of address tagged microspheres to be used for many different assays for which new, captured tagged primers are synthesized.

This approach has proven very efficient for large scale SNP genotyping studies. The development of large sets of address and capture tags" ^ and software tools to facilitate multiplexed primer design" ^ and data analysis"^ , have made it possible to rapidly configure new multiplexed SNP genotyping assays. The results of these assays compare well with those obtained from conventional DNA sequencing"^ '' , and enable the efficient genotyping in studies involving hundreds of samples"^ .

9.7. SUMMARY AND FUTURE DIRECTIONS

In this article, I have given an overview of the application of microparticle-based flow cytometry to high resolution molecular analysis. The applications described range from kinetic analysis of interaction mechanisms to highly sensitive and parallel screening assays. Some of these types of applications are now widely used, supported by commercial kits and reagents. Others are being integrated into high throughput screening environments for disease marker and drug discovery. Current research in new optical encoding schemes for microparticles, new detection schemes for fluorescence and other optical signals, and microparticle fabrication and surface chemistries promise to lead to even more flexible and robust assay systems for molecular analysis


I would like to thank the many colleagues and coauthors who worked with me on these projects, especially Xiaomei Yan, Xuedong Song, Sabine Lauer, Rhiannon Nolan, Byron Goldstein, Binghui Shen, Min Park, Larry Sklar, Hong Cai, Scott White, and Steve Graves. I am also grateful to the National


Institutes of Health grants RR01401, RR01315, and EB03824, which funded much of this and continuing work.

9.9. REFERENCES

1. Wilkerson,C.W., P.M.Goodwin, W.P.Ambrose, J.C.Martin, and R.A.Keller. 1993. Detection and Lifetime Measurement of Single Molecules in Flowing Sample Streams by Laser-Induced Fluorescence. Applied Physics Letters 62:2030-2032.

2. Van Orden,A. and R.A.Keller. 1998. Fluorescence correlation spectroscopy for rapid multicomponent analysis in a capillary electrophoresis system. Analytical Chemistry 70:4463-4471.

3. Goodwin,P.M., W.P.Ambrose, and R.A.Keller. 1996. Single-molecule detection in liquids by laser-induced fluorescence. Accounts of Chemical Research 29:607-613.

4. Horan,P.K. and L.L.Wheeless, Jr. 1977. Quantitative single cell analysis and sorting. Science \9%\U9-\51.

5. McHugh,T.M., D.P.Stites, C.H.Casavant, and M.J.Fulwyler. 1986. Flow cytometric detection and quantitation of immune complexes using human Clq-coated microspheres. J. Immunol Methods 95:57-61.

6. Fulwyler,M.J. and T.M.McHugh. 1990. Flow microsphere immunoassay for the quantitative and simultaneous detection of multiple soluble analytes. Methods Cell Biol. 33:613-629.

7. Nolan,J.P. and L.A.Sklar. 1998. The emergence of flow cytometry for sensitive, real-time measurements of molecular interactions. Nature Biotechnology 16:633-638.

8. Shaner,N.C., R.E.Campbell, P.A.Steinbach, B.N.G.Giepmans, A.E.Palmer, and R.Y.Tsien. 2004. Improved monomeric red, orange and yellow fluorescent proteins derived from Discosoma sp red fluorescent protein. Nature Biotechnology 22:1567-1572.

9. Adams,S.R., R.E.Campbell, L.A.Gross, B.R.Martin, G.K.Walkup, Y.Yao, J.Llopis, and R.Y.Tsien. 2002. New biarsenical Ligands and tetracysteine motifs for protein labeling in vitro and in vivo: Synthesis and biological applications. Journal of the American Chemical Society 124:6063-6076.

10. Fulton,R.J., R.L.McDade, P.L.Smith, L.J.Kienker, and J.R.Kettman, Jr. 1997. Advanced multiplexed analysis with the FlowMetrix system. Clin.Chem. 43:1749-1756.

11. Kettman,J.R., T.Davies, D.Chandler, K.G.Oliver, and R.J.Fulton. 1998. Classification and properties of 64 multiplexed microsphere sets. Cytometry 33:234-243.

12. Yan,X., E.G.Schielke, K.M.Grace, C.Hassell, B.L.Marrone, and J.P.Nolan. 2004. Microsphere-based duplexed immunoassay for influenza virus typing by flow cytometry. J.ImmunolMethods 284:27-38.

13. Yan,X., A.Tang, E.G.Schielke, W.Hang, and J.P.Nolan. 2004. Development of a microsphere-based multiplexed immunoassay for influenza virus typing and sub-typing by flow cytometry. Proceedings Influenza V Conference: International Congress Series 1263:342-345.

14. Dunbar,S.A., C.A.Vander Zee, K.G.Oliver, K.L.Karem, and J.W.Jacobson. 2003. Quantitative, multiplexed detection of bacterial pathogens: DNA and protein applications of the Luminex LabMAP system. J.Microbiol.Methods 53:245-252.

15. McBride,M.T., D.Masquelier, B.J.Hindson, A.J.Makarewicz, S.Brown, K.Burris, T.Metz, R.G.Langlois, K.W.Tsang, R.Bryan, D.A.Anderson, K.S.Venkateswaran, F.P.Milanovich, and B.W.Colston. 2003. Autonomous detection of aerosolized Bacillus anthracis and Yersinia pestis. Analytical Chemistry 75:5293-5299.

16. Song,X., J.Shi, J.Nolan, and B.Swanson. 2001. Detection of multivalent interactions through two-tiered energy transfer. Anal.Biochem. 291:133-141.

17. Song,X., J.P.Nolan, and B.I.Swanson. 1998. An optical biosensor based on fluorescence self-quenching and energy transfer: Ultrasensitive and specific detection of protein toxins. Journal of the American Chemical Society 120:11514-11515.

18. Song,X.D., J.Nolan, and B.I.Swanson. 1998. Optical signal transduction triggered by protein-ligand binding: Detection of toxins using multivalent binding. Journal of the American Chemical Society 120:4873-4874.

212 J. P. NOLAN

19. Liu,Y., H.I.Kao, and R.A.Bambara. 2004. Flap endonuclease 1: A central component of DNA metabolism. Annual Review of Biochemistry 73:589-615.

20. NolanJ.P., B.Shen, M.S.Park, and L.A.Sklar. 1996. Kinetic analysis of human flap endonuclease-1 by flow cytometry. Biochemistry 35:11668-11676.

21. Shen,B., J.P.Nolan, L.A.Sklar, and M.S.Park. 1996. Essential amino acids for substrate binding and catalysis of human flap endonuclease 1. J.Biol. Chem. 271:9173-9176.

22. Shen,B., J.P.Nolan, L.A.Sklar, and M.S.Park. 1997. Functional analysis of point mutations in human flap endonuclease-1 active site. Nucleic Acids Res. 25:3332-3338.

23. Frank,G., J.Qiu, M.Somsouk, Y.Weng, L.Somsouk, J.P.Nolan, and B.Shen. 1998. Partial functional deficiency of E160D flap endonuclease-1 mutant in vitro and in vivo is due to defective cleavage of DNA substrates. J.Biol.Chem. 273:33064-33072.

24. Gary,R., M.S.Park, J.P.Nolan, H.L.Comelius, O.G.Kozyreva, H.T.Tran, K.S.Lobachev, M.A.Resnick, and D.A.Gordenin. 1999. A novel role in DNA metabolism for the binding of Fenl/Rad27 to PCNA and implications for genetic risk. Mol.Cell Biol. 19:5373-5382.

25. Buranda,T., G.P.Lopez, P.Simons, A.Pastuszyn, and L.A.Sklar. 2001. Detection of epitope-tagged proteins in flow cytometry: fluorescence resonance energy transfer-based assays on beads with femtomole resolution. Anal.Biochem. 298:151-162.

26. Sklar,L.A., B.S.Edwards, S.W.Graves, J.P.Nolan, and E.R.Prossnitz. 2002. Flow cytometric analysis of ligand-receptor interactions and molecular assemblies. Annu.Rev.Biophys.Biomol.Struct. 31:97-119.

27. Waller,A., P.C.Simons, S.M.Biggs, B.S.Edwards, E.R.Prossnitz, and L.A.Sklar. 2004. Techniques: GPCR assembly, pharmacology and screening by flow cytometry. Trends Pharmacol.Sci. 25:663-669.

28. Waller,A., P.Simons, E.R.Prossnitz, B.S.Edwards, and L.A.Sklar. 2003. High throughput screening of G-protein coupled receptors via flow cytometry. Comb.Chem.High Throughput.Screen. 6:389-397.

29. Simons,P.C., M.Shi, T.Foutz, D.F.Cimino, J.Lewis, T.Buranda, W.K.Lim, R.R.Neubig, W.E.McIntire, J.Garrison, E.Prossnitz, and L.A.Sklar. 2003. Ligand-receptor-G-protein molecular assemblies on beads for mechanistic studies and screening by flow cytometry. Mol.Pharmacol. 64:1227-1238.

30. Bennett,T.A., T.A.Key, V.V.Gurevich, R.Neubig, E.R.Prossnitz, and L.A.Sklar. 2001. Real-time analysis of G protein-coupled receptor reconstitution in a solubilized system. J.Biol.Chem. 276:22453-22460.

31. Simons,P.C., S.M.Biggs, A.Waller, T.Foutz, D.F.Cimino, Q.Guo, R.R.Neubig, W.J.Tang, E.R.Prossnitz, and L.A.Sklar. 2004. Real-time analysis of ternary complex on particles: direct evidence for partial agonism at the agonist-receptor-G protein complex assembly step of signal transduction. J.5/W.C/zem. 279:13514-13521.

32. Sarvazyan,N.A., A.E.Remmers, and R.R.Neubig. 1998. Determinants of gilalpha and beta gamma binding. Measuring high affinity interactions in a lipid environment using flow cytometry. J.Biol.Chem. 273:7934-7940.

33. Sarvazyan,N.A., W.K.Lim, and R.R.Neubig. 2002. Fluorescence analysis of receptor-G protein interactions in cell membranes. Biochemistry ^\:\2^5^-\2%61.

34. Lan,K.L., N.A.Sarvazyan, R.Taussig, R.G.Mackenzie, P.R.DiBello, H.G.Dohlman, and R.R.Neubig. 1998. A point mutation in Galphao and Galphail blocks interaction with regulator of G protein signaling proteins. J.Biol.Chem. 273:12794-12797.

35. Lauer,S., B.Goldstein, R.L.Nolan, and J.P.Nolan. 2002. Analysis of cholera toxin-ganglioside interactions by flow cytometry. Biochemistry 41:1742-1751.

36. Lauer,S.A. and J.P.Nolan. 2002. Development and characterization of Ni-NTA-bearing microspheres. Cytometry 48:136-145.

37. Brookes,A.J. 1999. The essence of SNPs. Gene 234:177-186. 38. Milos,P.M. and A.B.Seymour. 2004. Emerging strategies and applications of

pharmacogenomics. Hum.Genomics 1:444-455. 39. Chen,X. and P.F.Sullivan. 2003. Single nucleotide polymorphism genotyping:

biochemistry, protocol, cost and throughput. Pharmacogenomics.J. 3:77-96. 40. D'Yachkov,A.G., P.L.Erdos, A.J.Macula, V.V.Rykov, D.C.Tomey, C.S.Tung,

P.A.Vilenkin, and P.S.White. 2003. Exordium for DNA codes. Journal of Combinatorial Optimization 7:369-379.


41. Kaderali,L., A.Deshpande, J.P.Nolan, and P.S.White. 2003. Primer-design for multiplexed genotyping. Nucleic Acids Res. 31:1796-1802.

42. Cleland,C.A., P.S.White, A.Deshpande, M.Wolinsky, J.Song, and J.P.Nolan. 2004. Development of rationally designed nucleic acid signatures for microbial pathogens. Expert Rev.Mol.Diagn. 4:303-315.

43. Deshpande,A., J.P.Nolan, P.S.White, Y.E.Valdez, W.C.Hunt, C.L.Peyton, and C.M.Wheeler. 2005. TNF- alpha Promoter Polymorphisms and Susceptibility to Human Papillomavirus 16-Associated Cervical Cancer. J.Infect.Dis. 191:969-976.

TOTAL INTERNAL REFLECTION-FLUORESCENCE CORRELATION

SPECTROSCOPY

Nancy L. Thompson^ and Jamie K. Pero

10.1. ABSTRACT

The combination of total internal reflection illumination with fluorescence correlation spectroscopy (TIR-FCS) is an emerging technique. This method allows measurement of at least three key properties of fluorophores very close to surface/solution interfaces, including the local fluorophore concentration, the local fluorophore translational mobility, and the kinetic rate constants which describe the reversible association of fluorophores with the interface. This review describes the conceptual basis of TIR-FCS, aspects important to the experimental realization of this technique, and methods for theoretically modeling and analyzing data. Previous experimental applications of TIR-FCS are also summarized, including studies of protein dynamics very close to substrate-supported planar membranes, measurement of the kinetic dissociation rate for fluorescent ligands specifically and reversibly associating with receptors in substrate-supported planar membranes, the interaction of small dye molecules with chromatographic surfaces, investigations of fluorophore behavior in sol-gel films, and the monitoring of intracellular vesicular motion near the inner leaflet of plasma membranes in intact, live cells. A number of potential future directions for TIR-FCS are indicated, for example, the use of very high refractive index substrates or small, metallic substrate-deposited structures, the use of photon counting histograms, the use of a single fluorescent

Department of Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599-3290 (919) 962 0328 (telephone) (919) 966 3675 (fax) [email protected] (electronic mail)

215

216 N. L. THOMPSON ETAL

reporter molecule to provide kinetic rate constants for nonfluorescent molecules which compete for the same surface sites, and more sophisticated methods of data analysis including two-color cross-correlation and high-order autocorrelation.

10.2. INTRODUCTION

Total internal reflection fluorescence microscopy (TIRFM) is a type of fluorescence microscopy in which the excitation light is the interfacial evanescent field created by internally reflecting a light source at a planar dielectric interface. The use of evanescent illumination in fluorescence microscopy encompasses a set of related methods in biophysics as well as other fields such as cell biology, optical biosensors and polymer chemistry. Many TIRFM methods are now well-established; others are in development. Some of the more well-developed TIRFM methods include the use of evanescent illumination to measure surface densities of bound, fluorescent molecules (e.g., for the purpose of measuring surface binding isotherms), to monitor fluorescence recovery after photobleaching (e.g., for measuring surface binding kinetic constants), to examine processes near cell-substrate contact regions (e.g., for characterizing exocytotic events), and to carry out polarization-sensitive fluorescence intensity measurements (e.g., for measuring order parameters in thin films deposited on transparent planar surfaces). TIRFM has been the subject of a number of recent reviews in a variety of contexts (Thompson and Pero, 2005; Schneckenburger, 2005; Axelrod, 2003; Sako and Uyemura, 2002; Toomre and Manstein, 2001; Steyer and Aimers, 2001; Krishnan et al., 2001).

Fluorescence correlation spectroscopy (FCS) is also a set of methods which encompass both mature and developing variants applicable to a wide range of questions in biophysics and related fields. In FCS, temporal fluctuations in the fluorescence measured from a small volume are analyzed to obtain information about the processes giving rise to the observed fluctuations. Some of the FCS methods for which firm theoretical and experimental foundations have been developed include fluorescence fluctuation autocorrelation to measure translational mobility (e.g., for defining apparent diffusion coefficients in different environments), two-color fluorescence fluctuation cross-correlation (e.g., for detecting bimolecular binding), photon counting histograms (e.g., for characterizing molecular aggregation) and spatial fluorescence fluctuation analysis (e.g., for monitoring cell-surface receptor clustering). FCS has also been described in a number of recent reviews as well as a book (Vukojevic et al., 2005; Gosch and Rigler, 2005; Haustein and Schwille, 2004; Enderlein et al., 2004; Pramanik, 2004; Weiss and Nilsson, 2004; Kahya et al., 2004; Levin and Carson, 2004; MuUer et al., 2003; Hink et al., 2003; Thompson et al , 2002; Frieden et al., 2002; Rigler and Elson, 2001).

In this review, the combination of total internal reflection with fluorescence correlation spectroscopy (TIR-FCS), an emerging technique both in the set of methods included in TIRFM and in the set of methods included in FCS, is

TIR-FCS 217

O O o o o °

o o o o o o o o

Figure 10.1. Total Internal Reflection Fluorescence. A laser beam with vacuum wavelength Xo traveling from a high refractive index (ni) medium into a lower refractive index (n2) medium is totally internally reflected at a planar interface when the incidence angle a is greater than the critical angle ttc (Eq. 3) The incidence plane is defined as the x-z plane and the interfacial plane as the x-y plane. Internal reflection creates a thin layer of light in the lower refractive index medium, called the "evanescent field", which decays exponentially in intensity along the z-axis with a characteristic depth d (Eq. 1). The depth d depends on 8o, a, ni and nj (Eq. 2). The polarization of the incident light is defined by the electric field direction, as denoted by angle v|/ measured from the incidence plane. When \\f = 0°, the incident light is p-polarized; when \\f = 90°, the incident light is s-polarized. Only molecules very close or bound to the surface are illuminated and subsequently fluoresce.

described. Recently developed applications of TIR-FCS include characterizing the local concentration and mobility of fluorescent molecules very close to biologically and chemically important interfaces as well as measuring the kinetic rate constants which describe the reversible association of fluorescent molecules with interfaces. These applications are described below. Potential future developments in TIR-FCS, in a variety of contexts, are also suggested.

10.3. CONCEPTUAL BASIS AND EXPERIMENTAL DESIGN

As described by elementary and idealized optical theory, total internal reflection occurs when a plane wave traveling in a higher refractive index medium impinges on a planar interface of a lower refractive index medium, at an angle (defined from the normal to the interface) greater than the "critical angle" (defined below) (Figure 10.1). When internal reflection occurs, the plane wave completely reflects into the higher refractive index medium. In addition, a surface-associated "evanescent" field is generated in the lower refractive index medium. The evanescent light propagates parallel to the interface and penetrates into the lower refractive index medium with a distance on the order of the incident light wavelength. In this work, the interface is defined by coordinates (x,y) and the incidence plane by coordinates (x,z). The properties of evanescent fields have been discussed previously in extensive detail (Girard


et al., 2000; Knoll, 1998; Thompson et al., 1993; Axelrod et al., 1984; Agudin andPlatzeck, 1978).

For a plane wave, the evanescent intensity as a function of the distance from the interface, z, is

/(z) = /oexp(-4) (1) a

where IQ is the intensity at the interface and d is the characteristic distance of evanescent penetration into the lower refractive index medium. The distance d is given by

d= , / " (2) 47^^jn^ singer-«2

where Xo is the vacuum wavelength of the incident light, a is the incidence angle, and ni and n2 are the higher and lower refractive indices, respectively. The incidence angle must be greater than the critical angle Vc where

a^ = s i n " ^ ( - ^ ) (3)

Typical values for these parameters are XQ ~ 500 nm (visible light), ni ~ 1.5 (glass or fused silica), n2 ~ 1.33 (water), a ~ 75°, ac ~ 65", and d ~ 70 nm.

The evanescent intensity at the interface, lo, depends on the two refractive indices and the incidence angle. For typical conditions, IQ is on the order of the intensity of the incident light (Axelrod et al., 1984). The polarization of the electric field is somewhat unusual for evanescent waves created by total internal reflection as compared to freely propagating plane waves. When the incident light is polarized perpendicular to the incidence plane (s-polarized), the evanescent field is also polarized perpendicular to the incidence plane (along the y-axis). However, an atypical evanescent polarization is predicted when the incident light is polarized in the incidence plane (p-polarized). Here, the evanescent electric field, although it lies completely in the incidence plane, contains components both perpendicular to the direction of evanescent propagation (along the z-axis) and parallel to the direction of propagation (along the X-axis) (Axelrod et al., 1984).

In the simplest version of TIR-FCS, the incident light beam is internally reflected through a prism mounted on the stage of an inverted fluorescence microscope (Figure 10.2a). If the incident light is a laser beam loosely focused at the point of internal reflection,the intensity of the evanescent field in the x-y plane varies approximately as an elliptical Gaussian and the evanescent depth is approximately equal to the expression in Eq. (2) throughout the illuminated area

TIR-FCS 219

a)

Glycerin

Oil y - r -Evanescently ^^^^V Objective \ c o v e r s l i p

Solution # \ \ \ \ ^ \ Planar Supported Bilayer

Fluorescence

Figure 10.2. Prism-Based Total Internal Reflection Fluorescence Microscopy, (a) The most common method for generating evanescent fields in TIR-FCS, thus far, has been to internally reflect a laser beam through a prism which is optically coupled to a substrate and mounted on the stage of an inverted optical microscope, (b) If the laser beam is roughly focused at the point of internal reflection, the evanescent intensity in the x-y plane has an elliptical Gaussian shape with 1/e^-radii on the order of 10-100 }xm. The radius along the x-axis is larger than the one along the y-axis. Fluorescence detection is restricted to a small area within the illuminated area by using a pinhole placed at a confocal back image plane.

(Burghardt and Thompson, 1984a) (Figure 10.2b). A circular pinhole is placed at a confocal back image plane, thus defining an area on the surface much smaller than the illuminated area (Figure 10.3a). The observed volume is Tch d where h is the confocal pinhole radius in the sample plane (.1 jim). The fluorescence arising from the observed volume is usually measured either with a sensitive single-photon counting photomultiplier or with a silicon avalanche photodiode. In some cases, imaging detectors are employed (e.g., Johns et al., 2001).

If a small number of molecules is within the observed volume, the measured fluorescence fluctuates significantly with time as individual fluorescent ligands diffuse into the volume, bind to sites on the surface, dissociate, and diffuse out of the volume (Figure 10.3b). The fluorescence fluctuations are autocorrelated as

G(r) < SF{t)SF{t + r) >

<F>^ (4)

where F(t) is the fluorescence at time t, <F> is the time-averaged fluorescence, and *F(t) is the time-dependent fluctuation of the fluorescence from its average value. G(r) typically decays to zero with the lag-time T. The magnitude of this function contains information about the density of surface-bound molecules and the concentration of molecules in solution. The rate and shape of the time decay of G(r) contain, in addition, information about the processes giving rise to the fluorescence fluctuations; for example, association rate constants for the


0 0

o 0

aperture i : . - . . . : - . . . . : . > : ^

>

0 0 0 0

0 0 0

observation area {TTH^)

to detector

f \

0 0

_ focusing ~ optics

i .,,^:-..\ i

f

b) dF{t)

<F>

Figure 10.3. Total Internal Reflection with Fluorescence Correlation Spectroscopy, (a) A small sample volume is defined by the depth of the evanescent intensity, d, in combination with a circular aperture placed at an intermediate image plane of the microscope that defines an area of radius h in the sample plane, (b) The fluorescence measured from the small sample volume adjacent to the surface fluctuates with time.

surface sites, dissociation rate constants for the surface sites, and diffusion coefficients in solution.

Several additional aspects of TIR-FCS experimental design are worth noting. (1) A recent paper has demonstrated that TIR-FCS can be carried out with a through-objective as opposed to a through-prism apparatus for generating evanescent light (Hassler et al., 2005). Through-objective optics have a number of advantages including better image quality and possibly higher detection sensitivity. However, this type of apparatus is not readily adapted to the use of very high refractive index substrates (see below), the evanescent polarization may be easier to control with through-prism optics and the background signal associated with the through-objective set-up can be higher due to the fact that the excitation light passes through the microscope interior. (2) The angular dependence of the emission of fluorophores next to surfaces is dramatically different from that in homogeneous space (Enderlein, 1999; Hellen and Axelrod, 1987; Thompson and Burghardt, 1986; Burghardt and Thompson,

TIR-FCS 221

1984b). This phenomenon means that the fluorescence collection efficiency and the effective evanescent wave depth can depend significantly on the precise parameters of the apparatus used for fluorescence collection (Steyer and Aimers, 2001). (3) For measurements using avalanche photodiodes as detectors, the signal can be directed to two different avalanche photodiodes and cross-correlated to eliminate contributions arising from diode after-pulsing (e.g., Licht et al., 2003). (4) When the background light is significant compared to the fluorescence arising from the molecules of interest, the measured value of G(T) can be corrected by a simple multiplicative factor (Thompson, 1991; Starr and Thompson, 2002).

10.4. THEORETICAL MODELS FOR DATA ANALYSIS

A significant amount of theoretical work has been developed in which the nature of the fluorescence fluctuation autocorrelation function for TIR-FCS has been defined (Lieto and Thompson, 2004; Starr and Thompson, 2001; Thompson, 1982; Thompson et al, 1981). For the purpose of this article, we have restricted discussion of these theoretical models to the case in which the sample volume radius h is enough higher in magnitude compared to the evanescent wave depth d so that the problem is reduced to a single dimensionality (the z-axis), although generalizations in which this assumption is not made have been presented for some cases (Hassler et al., 2005; Holt et al., 2004; Thompson et al., 1981). We also restrict the discussion to situations in which the time scale of the fluctuations of interest is slow enough and the incident light intensity is dim enough so that photochemical effects such as transient triplet state population do not significantly affect G(r), although generalizations in which this assumption is not made have also been described for some situations (Hassler et al., 2005; Widengren et al., 1995).

When the fluorescence fluctuations arise solely from diffusion through the depth of the evanescent intensity in a direction perpendicular to the sample plane (Starr and Thompson, 2001), G(T) is given by

G, (r) = G, (0)|(1 - 2Rj)M{i{Rjy" ] + 2 ( ^ ) " 4 (5)

where

C . ( 0 ) = ^ (6,

NA = Tch dA is the average number of fluorescent molecules in the observation volume for solution concentration A,


w(4)^cxp(-^')erM-i^) (7)

and Rf: is the rate for diffusion with coefficient D through the evanescent depth; i.e.,

R. D

(8)

The factor of two in the denominator of Eq. (6) arises from the manner in which the observation volume is defined. Ge(r) as predicted by Eqs. (5)-(8) is illustrated in Figure 10.4a. In general, the magnitude Ge(0) increases with decreasing NA (corresponding to lower solution concentrations, thinner evanescent wave depths or smaller observation area sizes) and the decay rate increases for higher diffusion coefficients or thinner evanescent depths. The functional form shown in Eqs. (5)-(8) does not account for surface effects which may alter concentration or diffusion very close to the interface; i.e., in the derivation of this equation, it was assumed that D and the average value of A did not change with z.

A second special case is one in which the fluorescence fluctuations arise solely from a reversible bimolecular reaction at the interface. This case is applicable when contributions from solution are negligible because the fluorescence arising from molecules in solution but close to the interface is much smaller than the fluorescence arising from molecules transiently bound to the surface. Here (Starr and Thompson, 2001),

j

1

B-5 1e-4 1e-3

T(sec)

0 08

a

0.04

V ^ 0.02 •

b)

- ^ 0.00 1 1e-5 1e-4 1e-3 1e-2

T(sec)

1e-1 1e+0

Figure 10.4. Theoretical Models for G(T). In these examples, D = 5 x 10^ cm^/sec and d = 75 nm. (a) Fluorescent molecules freely diffuse through the evanescent light. GeCx) was calculated from Eqs. (5)-(8) with NA = 5. The half-time for decay is 3.3Re \ (b) Fluorescent molecules reversibly bind to the surface with ka = 10 M'^sec', kd = 1 sec"\ K = 10 M'\ h = 0.5 im, S = 50 molecules/fxm^, and KS = 0.83 ^m. The total number of binding sites (occupied and unoccupied) in the observation volume is Ns = nh^S = 39. GS(T) was calculated from Eqs. (9)-(14) with (line) A = 0.03 iM, p - 0.77, Gs(0) = 0.083, R = 1.3 sec ' and Rt = 2.1 x 10 sec ^ (dash) A - 0.1 \xM, P = 0.500, Gs(0) = 0.025, R, = 2 sec \ Rt = 1.2 x 10^ and (dot) A = 0.3 iM, P = 0.25, Gs(0) = 0.008, R = 4 sec^ and Rt = 1.9 x 10"*. In this example, Rr is always much smaller than Rt and Gs(t) is nearly identical to Eq. (15). Measurement of the decay rate as a function of A would generate values for both ka and k . The values of Nc = (l-P)Ns and N A = rfdA, respectively, are (line) 9.1 and 1.1; (dash) 20 and 3.5; (dot) 29 and 11. Thus, Nc > NA.

TIR-FCS 223

a_^,â_^,Ê£l±M4nzKj±l(M^ ,,, R,-R.

where

GAO) = ^ (10)

InEq. (10)

J3 = —^ (11)

is the fraction of surface sites that are free, on the average, for a given solution concentration A and equilibrium association constant K, and Nc is the average number of fluorescent molecules bound to the surface w ithin the observed area. Gs(0) increases for lower values of Nc, as expected, but also depends on p. The P-dependence arises from the fact that the observed surface area, unlike the observed solution volume, is not completely open because it contains a finite number of binding sites. Thus, the magnitude is determined by binomial rather than Poisson statistics (Lieto and Thompson, 2004). The rates are given by

Rate Rr is the relaxation rate for a pseudo first-order reaction

R^=kÂ + k, (13)

where ka and kd are the kinetic association and dissociation rate constants for surface binding, respectively. The rate Rr increases with the solution concentration A; by measuring Rr as a function of A, the constants ka and k j may in principle be determined. Rate Rt is a rate describing transport in solution through distance KSP^

R, = , „ i ? 2 . 2 (14) (KSJ3')


where S is the total surface site density (occupied plus unoccupied). The relative magnitude of R and Rt reflects the degree to which dissociated molecules rebind to the surface within the observed area (Lagerholm and Thompson, 1998). When the relaxation rate is much smaller than the transport rate (Rr « Rt), rebinding is negligible, and the autocorrelation function assumes the simple form

G,(r) = G,(0)exp[-(^„^ +A:,)r] (15)

When the transport rate is much smaller than the relaxation rate (Rt « Rr), rebinding is prominent, and

KSP' GM) = GM^\'2 ] (16)

Eqs. (9)-(15) are illustrated in Figure 10.4b. In most experimental systems of interest, the surface binding kinetics are

much slower than Rg. In this case (Starr and Thompson, 2001; Lieto and Thompson, 2004),

G(r) = G,(r) + G,(r) (17)

where Ge(i ) is given by Eq. (5), Gs(x) is given by Eq. (9), and

G, (0 ) = ^ J (18)

G^(0) = ^^^ ^ (19)

Because Gs(0) is approximately inversely related to Nc, if the surface site density is high enough, then GS(T) is too small to be measurable in the midpoint of the binding isotherm where KA = 1. One method for circumventing this problem is to mix a small concentration of fluorescently labeled ligands with a larger concentration of unlabeled ligands (Lieto and Thompson, 2004). In this case, in the absence of rebinding and for large R^ G( x) is given by Eqs. (5), (17) and (18) with

G,(0){,g/7exp[-(/:^4 + ^ j ^ ] + (l-;y)exp(-/:,r)} (20)

TIR-FCS 225

In Eqs. (20) and (21), r| denotes the fraction of ligands on the surface that are fluorescently labeled and At denotes the total (labeled plus unlabeled) ligand concentration. For very low 0, the first term in Eq. (20) is of negligible magnitude and only the dissociation rate constant, kd, is measureable.

Of particular interest is the generalization of this situation to the case in which two different ligands, one fluorescent and one not fluorescent, compete for the same surface binding sites (Lieto and Thompson, 2004). In this case, G(T) is given by Eqs. (5), (17), (18), (21) and

G,(r) = G,(0)[jexp(-;L,r) + ( l-j)exp(-; i ,r)] (22)

where

\ i - •

(«^ +a„+ k^ +k,„)±[{a^ -a„+ k,^ -k,„ f + 4«^a„] '

7 = — ^ 2

(23)

(24)

af = kafAf and an = kanAn. In Eqs. (23) and (24), kaf and kan denote association rate constants, kdf and k n denote dissociation rate constants, and Af and An denote solution concentrations, for the fluorescent and nonfluorescent species, respectively. The remarkable property of Eqs. (22)-(24) is that GS(T) contains information about the kinetic rate constants (and not just the equilibrium binding constant) of the nonfluorescent competitor. Thus, it may be possible to use a single fluorescent ligand to carry out a kinetic screen for a series of nonfluorescent ligands.

10.5. APPLICATIONS

Solution-based PCS was first introduced during a time when other autocorrelation-based spectroscopies (i.e., dynamic light scattering) were also making their debut (Elson and Magde, 1974). Shortly thereafter, PCS was combined with total internal reflection in an apparatus called a virometer


X (msec)

Figure 10.5. Representative TIR-FCS Data in the Absence of Surface Binding. The autocorrelation curve is for 30 nM fluorescently labeled IgG in 25 mM Tris with 225 mM NaCl at pH 6.4 diffusing through the evanescent wave near a planar membrane. Reproduced with permission from J. Phys. Chem. B, 2002,106:2365.

(Hirschfeld and Block, 1911 a; Hirschfeld et al., 1911 b). This virometer was a rudimentary, proof-of-principle instrument proposed as a method for identifying viral particles stained with ethidium bromide based on their time for diffusion through the evanescent depth. This work demonstrated the experimental feasibility of TIR-FCS. Several years later, a firm theoretical backbone for TIR-FCS was presented (Thompson et al., 1981). This first theoretical paper outlined the predicted forms for TIR-FCS autocorrelation functions when fluorescent molecules in solution reversibly bind to specific surface sites. This paper also presented a theoretical basis for analyzing recovery curves obtained with a very similar technique in which total internal reflection is combined with fluorescence recovery after photobleaching. Subsequently a second theoretical paper briefly proposed an extension of TIR-FCS for measuring the surface binding kinetics of nonfluorescent species (Thompson, 1982). In this method, fluorescent molecules are combined with nonfluorescent molecules which compete for the same surface binding sites. The second experimental demonstration of TIR-FCS was one in which the kinetics associated with the nonspecific, reversible adsorption of rhodamine-labeled immunoglobulin and insulin to serum albumin-coated fused silica slides were examined (Thompson andAxelrod, 1983).

TIR-FCS has now seen a range of applications with additional ones most likely forthcoming. Two recent studies have used TIR-FCS to measure the evanescent wave depth, an important parameter in many TIRFM measurements. The evanescent wave depth has traditionally been inferred from Eq. (2), given known values for ni, n2, Xo and a, rather than directly measured. In one study, TIR-FCS was used to watch the diffusion of fluorescein through the evanescent wave and therefore measure the wave depth by using Eq. (5), Eq. (8) and an assumed value for the diffusion coefficient of fluorescein in solution (Harlepp et al., 2004). The measured values of d agreed well with the values predicted by Eq. (2). The second study also monitored the difftision of fluorescein in solution through the evanescent wave and found good agreement between Eq. (2) and Eq. (8) (Hassler et al., 2005). However, the purpose of this work was not wave depth determination, but to prove that through-objective TIR-FCS

TIR-FCS 227

(see above) can achieve count rates high enough to be used in single molecule applications.

TIR-FCS has seen much of its activity in cellular biophysics. In particular, cell membranes provide a bountiful group of participants whose behavior can be investigated by using TIR-FCS. Due to the surface confinement provided by evanescent illumination,TIR-FCS is particularly applicable to processes occurring at or near substrate-supported planar model membranes or the basal plasma membranes of surface-adherent, intact cells. In one recent example, TIR-FCS was used to measure the diffusion coefficients and concentrations of fluorescently labeled IgG in close proximity to planar model membranes (Figure 10.5) (Starr and Thompson, 2001; Starr and Thompson, 2002). These measurements were carried out for membranes with different compositions and surface charges and for solutions of varying pH and ionic strength. The purpose was to determine if previously observed nonidealities associated with the kinetics of specific ligand-receptor interactions at membranes are related to local deviations in the ligand concentration or mobility very near membrane surfaces. The results, for an evanescent wave depth of approximately 70 nm, showed no statistically significant change in the local IgG concentration for the numerous sample conditions explored, and no change in the local IgG mobility for membranes with different charges or for solutions with different pH values spanning the IgG isoelectric point. However, a significant decrease in the local mobility was detected with increasing ionic strength. This result suggests that strong hydrodynamic interactions occur between proteins and membrane surfaces and that these interactions are amplified by increasing ionic strength. Similar results have been seen for colloidal spheres diffusing close to planar walls (e.g., Bevan and Prieve, 2000).

A next step in the investigation of the nonidealities seen in ligand-receptor kinetics near membranes would be to use TIR-FCS to look directly at ligand-receptor interactions and specifically the kinetics of ligand-receptor interactions. One way to mimic cellular conditions is to embed receptors in substrate-supported planar membranes and then allow fluorescently labeled ligands to freely diffuse, bind to, and dissociate from the surface-associated receptors.

0.25

0.20

0.15

0.10

0.05

0 00

• C/,%

(a)

FcyRII in planar membrane

10 100 1000 10000

X (msec)

Figure 10.6. Representative TIR-FCS Data for Specific Ligand-Receptor Kinetics a) with Surface Binding and b) without Surface Binding. The autocorrelation functions are for 10 nM fluorescently labeled IgG with 1|LIM total IgG in solution. The sample in a) contained the mouse Fc receptor FcyRII embedded in the planar membrane while the sample in b) contained no receptor. Reproduced with permission from Biophys J., 2003, 85:3294.


while the system is in equilibrium. This approach has recently been used with TIR-FCS to examine the kinetics of fluorescently labeled IgG specifically and reversibly associating with the mouse Fc receptor FcyRII which was purified and reconstituted into supported membranes (Figure 10.6) (Lieto et al., 2003; Schwille, 2003; Butkus, 2003). The experimental parameters required for successful implementation of this type of measurement are not completely straightforward. First, to ensure that a high enough fraction of the evanescently excited fluorescence arises from surface-bound fluorescent ligands as opposed to those merely close to the surface, a high enough receptor density must be used. Second, large enough ligand concentrations must be used both to avoid working far below the midpoint of the binding isotherm where rare, tight, nonspecific binding sites might dominate the surface-bound species and to ensure that the fluorescence fluctuation autocorrelation function reflects surface kinetic rates rather than bulk diffusion (mass transport). However, these constraints can be contradictory to the mandates of FCS where a small number of observed, fluorescent molecules is required so that the magnitude of the fluorescence fluctuations relative to the mean fluorescence value is large enough to be accurately measured. In other words, meeting these conditions can result in a value of Gs(0) that is too small (Eq. 19). One method for circumventing this difficulty is to mix nonfluorescent ligands with a much smaller amount of fluorescently labeled ligands. Consequently, a comprehensive theory describing the nature of the TIR-FCS autocorrelation function when fluorescent and nonfluorescent ligands compete for surface binding sites was developed (Lieto and Thompson, 2004). This theory was used to interpret the data obtained for IgG - FcyRII interactions (Lieto et al., 2003).

TIR-FCS has also seen application to problems in fields other than membrane biophysics. The technique was used in a molecular counting study with rhodamine 6G as a probe in water/methanol solutions at bare silica surfaces and at silica surfaces derivatized with C-18 alkyl chains (Hansen and Harris, 1998a). This system is of particular interest in analytical chemistry because it mimics a reversed-phase chromatography column. Since the magnitude of the fluorescence fluctuation autocorrelation function is approximately inversely proportional to the number of fluorescent molecules in the observed volume (Eqs. (18) and (19)), G(0) can be used to obtain fluorophore surface densities. This property of TIR-FCS was used to determine the equilibrium association constants for the reversible adsorption of rhodamine 6G to the chromatographic support mimics for a variety of solution conditions. A second study extended the work to measure the dye adsorption and desorption kinetic rates (Hansen and Harris, 1998b). This second work was the first application of the theory predicting the manner in which the rates of surface binding, surface dissociation, diffusion through the evanescent wave, and mass transport combine to determine the rate and shape of decay of the TIR-FCS autocorrelation function (Starr and Thompson, 2001; Thompson et al., 1981).

TIR-FCS 229

TIR-FCS has also been used to examine molecular transport in sol-gel films (McCain and Harris, 2003). In this work, the translational mobilities of rhodamine 6G dye molecules inside sol-gel films made by dip-coating porous silica particles onto microscope slides were measured fi^om the time-decay of TIR-FCS autocorrelation ftmctions. These measurements defined the tortuosities for films of differing thicknesses, and both before and after ethanol annealing. A subsequent paper began to address the question of interparticle vs. intraparticle molecular mobilities in these sol-gel films (McCain et al., 2004a). Towards this goal, fluorescent probes of varying size were created as polyamidoamine dendrimers labeled with carboxyrhodamine 6G. Fluorescently labeled dendrimers were chosen as probes because they are available in monodisperse sizes ranging from 3-20 nm, thus making them amenable to the exploration of pore sizes within the films. By using TIR-FCS, it was found that the dendrimers nonspecifically adsorbed to bare silica surfaces, a property that would significantly complicate their use as probes of molecular transport in the sol-gel films. To circumvent this difficulty, the positively charged primary amine groups on the dendrimers were covalently modified to negatively charged carboxylic acid groups. This modification significantly reduced dendrimer adsorption, presumably because the silica surfaces are negatively charged. The carboxylated dendrimers were then placed inside sol-gel films and the mobilities were very carefully examined using TIR-FCS with the goal of understanding intraparticle vs. interparticle mobility as well as tortuosities (McCain et al., 2004b). This set of three papers is perhaps the most comprehensive use of TIR-FCS to date and outlines in a very thorough manner the way in which this method can be used to quantitatively define interfacial behavior.

Although the use of TIR-FCS in live cell experiments has thus far been limited, at least two studies have invoked its partial application. In one study, the motions of secretory granules near the plasma membranes of adherent bovine chromaffin cells were examined in detail by using various TIRFM approaches (Johns et al., 2001). In these measurements, secretory granules were labeled with a GFP conjugate of pro-atrial natriuretic peptide, and the behavior of the highly fluorescent granules was monitored before and during exocytosis. Evanescent light was used to illuminate only those vesicles close to the basal membrane. As part of this comprehensive work, the granule velocities, as measured with evanescent excitation and sequential imaging, were autocorrelated to determine if granule motion was Brownian in nature or if the autocorrelation data showed evidence of possible granule caging or tethering. The data, quite interestingly, were indicative of the latter situation in some cases. In a second study, TIR-FCS was used to measure effective diffusion coefficients for fluorescently labeled synaptic vesicles near ribbon synapses in retinal bipolar cells (Holt et al , 2004).

10.6. FUTURE DIRECTIONS

Thus far, most if not all TIR-FCS measurements have relied solely on the first-order fluorescence fluctuation autocorrelation function defined in Eq. (4).


However, there are a number of alternative and more sophisticated methods for analyzing fluorescence fluctuations, many of which have been used in solution-based FCS, which may prove to be advantageous for TIR-FCS studies. One alternative method for characterizing fluorescence fluctuations is to count the photons detected from the observed volume in sequential time elements and to use this information to generate the probability distribution for detecting a given number of photons within the observation time. This method is called photon-counting histogram (PCH) analysis or fluorescence intensity distribution analysis (FIDA) (Chen et al , 2002; Kask et al., 2000; Kask et al., 1999; Chen et al., 1999). The intensity histogram contains much of the same information as G(0) and is particularly sensitive to molecular clustering. For monodisperse samples, PCH/FIDA might allow determination of the concentration of fluorescent molecules as a function of the distance from the interface in a manner similar to that previously carried out for spheres next to planar surfaces (e.g., Bevan and Prieve, 2000). Another method for analyzing fluorescence fluctuations which is particularly sensitive to molecular clustering is to calculate high-order fluorescence fluctuation autocorrelation functions, defined as

G(T) = — ^ ^— (25) <F >'"^"

where m and n are integers greater than one. This method has been used previously to characterize polydisperse solutions (Palmer and Thompson, 1989a,b; Palmer and Thompson, 1987) but has not yet been explored with respect to evanescent excitation. A third possible approach, which has been rather fully developed for solution FCS but not explored in TIR-FCS is to cross-correlate fluctuations arising from two different molecular species which emit in spectrally separable regions (Widengren et al., 2001; Wallace et al., 2000; Schwille et al., 1997). The cross-correlation function is defined as

<^,(O^F,(/^r)> G{T) = (26)

< F, >< F^ >

where Fi and F2 are the fluorescence intensities detected in the two different channels and 8Fi(t) and 5F2(t) are the fluctuations from the mean fluorescence values. This type of measurement is particularly sensitive to cases in which the two different molecular species combine to form a complex. Two-color cross-correlation can also be used to monitor the structural dynamics of dually labeled single molecules via fluorescence resonance energy transfer. A final possibility might be to examine the behavior of multidimensional correlation functions; i.e..

G(T,,T,,.,. rJ = T J ^

(27)

TIR-FCS 231

One experimental question is the method by which evanescent excitation is generated. As described above, both through-objective and through-prism optical geometries have been used in TIR-FCS, and these two different methods have complementary advantages and disadvantages. Also worth mention is the possibility of using very high refractive index substrates to generate very thin evanescent fields; for example, single crystal Ti02 or SrTiOs (Starr and Thompson, 2000). These materials have refractive indices of Ui ~ 2.5 in the visible. Thus, the critical angle for total internal reflection at an aqueous interface with n2 = 1.334 is ac ~ 32° (Eq. 3) and, for XQ = 488 nm and an incidence angle of a ~ 85°, the evanescent depth d is predicted to equal the extremely low value of 18 nm. Another interesting approach might be to use thin metal films to quench fluorescence very close to the interface (Axelrod et al., 1992). Combined with evanescent excitation in which the incidence angle and therefore the evanescent wave depth are varied, one might be able to tune, to a certain degree, the z-axis sensitivity. A third possibility is to use wavelengths suitable for two-photon excitation (Gryczynski et al., 1997). When fluorophores are excited by two-photon absorption, the excitation probability is proportional to the square of the excitation intensity. Thus, the characteristic depth of penetration, d, for two-photon excitation is halved (Eq. 1). This halved depth of penetration theoretically gives a two-fold better discrimination for surface-adjacent fluorophores as compared to those in solution. However, for the same fluorophore, excited by one-photon or two-photon absorption, the wavelength is approximately doubled in the latter case, giving rise to a depth d which is doubled (Eq. 2). Therefore, two-photon absorption is not predicted to significantly enhance surface selectivity. Nonetheless, it is possible that other advantages accompanying two photon excitation, such as reduced cellular autofluorescence, might make this type of excitation useful in TIR-FCS. In addition, for two-color cross-correlation measurements, some fluorophore pairs can be excited via two-photon absorption with the same wavelength but nonetheless emit in spectrally separable regions (Heinze et al., 2000). Thus, two-photon excitation in two-color cross-correlation TIR-FCS could have the advantage of not requiring two excitation sources with different wavelengths.

New types of sample chamber designs might also prove to be useful in TIR-FCS. One possibility is to use flow to generate the fluorescence fluctuations rather than relying on diffusion, although the flow profile next to the surface should be carefully considered (e.g., Jennissen and Zumbrink, 2004; Jin et al., 2004; Hansen and Harris, 1998a,b). Of particular interest is the possible use of nanofabricated microfluidic channels, which could significantly reduce sample preparation time and the amount of sample required. These features can limit experimental feasibility in cases where the problem of interest involves multiple interacting biological species or other sample parameters. Opportunities also exist in this area for the design of microarray chip-based screening devices. Microfluidic chambers have previously been used both with FCS (e.g., Foquet et al., 2004; Foquet et al., 2002; Dittrich and Schwille, 2002) and with TIRFM (Jin et al., 2004; Yang et al., 2003; Yang et al., 2001; Jakeway and de Mello, 2001) but have not yet been implemented in TIR-FCS.


One of the most advantageous aspects of TIR-FCS is the dual reduction in sample volume achieved by limiting the illuminated volume by a surface confined evanescent wave and then further restricting what is detected by precisely placing a pinhole at the confocal image plane. However, recent years have presented a number of alternative methods by which sample volume reduction and single molecule sensitivity can be attained. A recent brief review of these methods provides a starting point from which to discuss them (Laurence and Weiss, 2003). The use of metals in fluorescence spectroscopy has received much attention over the past five years and might be used to achieve volume reduction due to localized enhanced excitation (Asian et al., 2005; Maliwal et al., 2003; Lakowicz et al, 2004; Lakowicz, 2001). A twist on the use of metals in fluorescence has produced zero-mode waveguides composed of tiny holes in metal films that avert light propagation but permit evanescent waves (Levene et al., 2003). These tiny holes consequently limit the sampling region. Stimulated emission depletion (STED) uses a visible beam to excite fluorophores and a pinhole to limit detection. However, what is unique to STED is that a second near infrared beam is used to quench the fluorescence immediately surrounding the focal spot and to further localize the sample volume (Klar et al., 2000). Near field scanning optical microscopy (NSOM) uses an optical probe with a subwavelength pinhole to scan in close proximity to an optically active sample. This set-up produces a confined observation volume (Krishnan et al., 2001). A particularly interesting possibility is the use of supercritical angle fluorescence which reports surface confined fluorescence in a manner similar to evanescent excitation except that internal reflection of the incident light is not required (Axelrod, 2001; Ruckstuhl and Verdes, 2004; Ruckstuhl and Seeger, 2004). Small microfluidic channels can also be used to limit detection volume sizes (Foquet et al., 2004).

Thus far, TIR-FCS has been used primarily to characterize mobility close to interfaces and to examine the kinetics of reversible surface binding. Both of these types of measurements have not yet been fully explored. For mobility measurements, it may be possible, by using different evanescent wave depths and molecular sizes, to obtain more detailed information about the manner in which the mobility and concentration depend on the distance from the surface of interest. TIR-FCS has thus far been used to examine the kinetics of specific surface binding for only one system (see above). Further development is needed in this area to precisely define the conditions for which this type of measurement can be made (e.g., in terms of the receptor density on the surface and the ligand concentration) and to demonstrate further applicability to other biochemical systems of interest.

A number of additional, possible future applications for TIR-FCS are mentioned here. Particularly when combined with new data analysis methods (see above), TIR-FCS may be able to address the kinetics of surface binding when the mechanism is more complex than a simple bimolecular reaction between fluorescent molecules in solution and surface binding sites. One particular system of interest is the case in which one fluorescent species in solution competes with a different nonfluorescent species in solution for the same surface binding sites. In this case, the autocorrelation function contains

TIR-FCS 233

information about the kinetic rates of the nonfluorescent species. Thus, this approach might be developed into a kinetic screen for nonfluorescent ligands. Because the evanescent field is polarized, it may be possible to monitor fluorescence fluctuations arising from rotational motions or macromolecular flexing. Finally, the nature of G(T) when surface binding sites are laterally mobile has not yet been fully explored.

When adherent cells are illuminated by evanescent light, only fluorescent molecules bound or close to the basal membrane are excited (Axelrod, 1981). Thus, if intracellular fluorescent molecules are reversibly associating with the cytoplasmic face of the plasma membrane, the kinetics of this process might be characterized by TIR-FCS. Similar measurements have been carried out either by combining evanescent illumination with fluorescence recovery after photobleaching (Sund and Axelrod, 2000) or by imaging single fluorescent molecules as they bind to and dissociate from the cytoplasmic membrane leaflet (Mashanov et al., 2004; Ueda et al., 2001). When combined with an imaging format, this type of measurement produces a kinetic map of the basal cell membrane. Finally, for very flat, adherent cells, the evanescent wave can penetrate through the entire cell and illuminate the exterior face of the apical plasma membrane. This arrangement would allow use of TIR-FCS to examine the behavior of fluorescent molecules in solution which reversibly interact with sites on the cell exterior.


This work was supported by NSF grant MCB-0130589.

10.8. REFERENCES

Agudin, J.L., and Platzeck, A.M., 1978, Fermat's principle and evanescent waves, J. Opts. 9:101. Asian, K., Gryczynski, I., Malicka, J., Matveeva, E., Lakowicz, J.R., and Geddes, CD., 2005,

Metal-enhanced fluorescence: an emerging tool in biotechnology, Curr. Opin. Biotech. 16:55. Axelrod, D., 1981, Cell-substrate contacts illuminated by total internal reflection fluorescence, J.

Cell Biol. H9:l4l. Axelrod D., Burghardt, T.P., and Thompson, N.L., 1984, Total internal reflection fluorescence,

Annu. Rev. ofBiophys. and Bioeng. 13:247. Axelrod, D., Hellen, E.H., and Fulbright, R.M., 1992, Topics in Fluorescence Spectroscopy Volume

3: Biochemical Applications, Plenum Press, New York, pp. 289-343. Axelrod, D., 2001, Selective imaging of surface fluorescence with very high aperture microscope

objectives, J. Biomed. Opt. 6:6. Axelrod, D., 2003, Total internal reflection fluorescence microscopy in cell biology. Meth.

Enzymol. 361:1. Bevan, M.A., and Prieve D.C., 2000, Hindered diffusion of colloidal particles very near to a wall:

revisited, J. Chem. Phys. 113:1228. Burghardt, T.P., and Thompson, N.L., 1984a, Evanescent intensity of a focused Gaussian light

beam undergoing total internal reflection in a prism. Opt. Eng. 23:62.

Burghardt, T.P., and Thompson N.L., 1984b, Effect of planar dielectric interfaces on fluorescence emission and detection: evanescent excitation with high-aperture collection, Biophys. J. 46:729.


Butkus, B.D., 2003, Fluorescence correlation spectroscopy quantifies ligand-receptor binding, Biophoton. Int. 10:55.

Chen, Y., MuUer, J.D., So, P.T.C., and Gratton, E., 1999, The photon counting histogram in fluorescence fluctuation spectroscopy, Biophys J. 77:553.

Chen, Y., Muller, J.D., Ruan, Q.Q., and Gratton, E., 2002, Molecular brightness characterization of EGFP in vivo by fluorescence fluctuation spectroscopy, Biophys. J. 82:133.

Dittrich, P.S., and Schwille, P., 2002, Spatial two-photon fluorescence cross-correlation spectroscopy for controlling molecular transport in microfluidic structures, Anal. Chem. 74:4472.

Elson, E.L., and Magde, D., 1974, Fluorescence correlation spectroscopy. I. conceptual basis and ihQory, Biopolymers. 13:1.

Enderlein, J., 1999, Fluorescence detection of single molecules near a solution / glass interface - an electrodynamic analysis, Chem. Phys. Lett. 308:263.

Enderlein, J., Gregor, I., Patra, D., and Fitter, J., 2004, Art and artefacts of fluorescence correlation spectroscopy, Curr. Pharm. Biotech. 5:155.

Foquet, M., Korlach, J., Zipfel, W.R., Webb, W.W., and Craighead, H.G., 2004, Focal volume confinement by submicrometer-sized fluidic channels, Anal. Chem. 76:1618.

Foquet, M., Korlach, J., Zipfel, W.R., Webb, W.W., and Craighead, H.G., 2002, DNA fragment sizing by single molecule detection in submicrometer-sized closed fluidic channels, Anal. Chem. 74:1415.

Frieden, C , Chattopadhyay, K., and Elson, E.L., 2002, What fluorescence correlation spectroscopy can tell us about unfolded proteins, Adv. Protein Chem. 62:91.

Girard, C , Joachim, C , and Gauthier, S., 2000, The physics of the near-field. Rep. Progress Phys. 63:893.

Gosch, M., and Rigler, R., 2005, Fluorescence correlation spectroscopy of molecular motions and kinetics, <^. Drug Deliv. Rev. 57:169.

Gryczynski, I., Gryczynski, Z., and Lakowicz, J.R., 1997, Two-photon excitation by the evanescent wave from total internal reflection. Anal. Biochem. 247:69.

Hansen, R.L., and Harris, J.M., 1998a, Total internal reflection fluorescence correlation spectroscopy for counting molecules at solid/liquid interfaces, Anal. Chem. 70:2565.

Hansen, R.L., and Harris, J.M., 1998b, Measuring reversible adsorption kinetics of small molecules at solid/liquid interfaces by total internal reflection fluorescence correlation spectroscopy, Anal. Chem. 10:4241.

Harlepp, S., Robert, J., Damton, N.C., and Chatenay, D., 2004, Subnanometric measurements of evanescent wave penetration depth using total internal reflection microscopy combined with fluorescent correlation spectroscopy, Appl. Phys. Lett. 85:3917.

Hassler, K., Anhut, T., Rigler, R., Gosch, M., and Lasser, T., 2005, High count rates with total internal reflection fluorescence correlation spectroscopy, Biophys. J. 88:L1.

Haustein, E., and Schwille, P., 2004, Single-molecule spectroscopic methods, Curr. Opin. Struct. Biol.,U:53\.

Heinze, K.G., Koltermann, A., and Schwille, P., 2000, Simultaneous two-photon excitation of distinct labels for dual-color fluorescence cross-correlation analysis, Proc. Natl. Acad. Sci. U.S.A. 97:10377.

Hellen, E.H., and Axelrod, D., 1987, Fluorescence emission at dielectric and metal-film interfaces. J. Opt. Soc. B 4: 337.

Hink, M.A., Borst, J.W., and Visser, A.J.W.G., 2003, Fluorescence correlafion spectroscopy of GFP fusion proteins in living plant cells, Meth. Enzymol. 361:93.

Hirschfeld, T., and Block, M.J., 1977, Virometer: real time virus detection and identification in biological fluids. Opt. Eng.\6'.406.

Hirschfeld, T., Block, M.J., and Mueller, W., 1977, Virometer: an optical instrument for visual observation, measurement, and classification of free viruses, J. Histochem. Cytochem. 25:719.

Holt, M., Cooke, A., Neef, A., and Lagnado, L., 2004, High mobility of vesicles supports continuous exocytosis at a ribbon synapse, Curr. Biol. 14:173.

Jakeway, S.C, and de Mello, A.J., 2001, Chip-based refracfive index detection using a single point evanescent wave probe guide. Analyst 126:1505.

Jennissen, H.P., and Zumbrink, T., 2004, A novel nanolayer biosensor principle, Biosens. Biolect. 19:987.

Jin, S., Huang, P., Park, J., Yoo, J.Y., and Breuer, K.S., 2004, Near-surface velocimetry using evanescent wave illumination, Expt. Fluids 37:825.

TIR-FCS 235

Johns, L.M., Levitan, E.S., Shelden, E.A., Holz, R.W., and Axelrod, D., 2001, Restriction of secretory granule motion near the plasma membrane of chromaffin cells, J. Cell Biol. 153:177.

Kahya, N., Scherfeld, D., Bacia, K., and Schwille, P., 2004, Lipid domain formation and dynamics in giant unilamellar vesicles explored by fluorescence correlation spectroscopy, J. Struct. Bio.Ul'.ll.

Kask, P., Palo, K., Fay, N., Brand, L., Mets, U., Ullmann, D., Jungmann, J., Pschorr, J., and Gall, K., 2000, Two-dimensional fluorescence intensity distribution analysis: theory and applications, Biophys.J.n-.XlQ?^.

Kask, P., Palo, K., Ullmann, D., and Gall, K., 1999, Fluorescence intensity distribution analysis and its application in biomolecular detection technology, Proc. Natl. Acad. Sci. U.S.A. 96:13756.

Klar, T.A., Jakobs, S., Dyba, M., Egner, A., and Hell, S.W., 2000, Fluorescence microscopy with diffraction resolution barrier broken by stimulated emission, Proc. Natl. Acad. Sci. U.S.A. 97:8206.

Knoll, W., 1998, Interfaces and thin films as seen by bound electromagnetic waves, Annu. Rev. Phys. Chem. 49:569.

Krishnan, R.V., Varma, R., and Mayor, S., 2001, Fluorescence methods to probe nanometer-scale organization of molecules in living cell membranes, J. Fluoresc. 11:211.

Lagerholm, B.C., and Thompson, N.L., 1998, Theory for ligand rebinding at cell membrane surfaces, Biophys. J. 74:1215.

Lakowicz, J.R., 2001, Radiative decay engineering: biophysical and biomedical applications. Anal. Biochem.29S:\.

Lakowicz, J.R., Geddes, CD., Gryczynski, I., Malicka, J., Gryczynski, Z., Asian, K., Lukomska, J., Matveeva, E., Zhang, J.A., Badugu, R., and Huang, J., 2004, Advances in surface-enhanced fluorescence,/. Fluoresc. 14:425.

Laurence, T.A., and Weiss, S., 2003, How to detect weak pairs, Science 299:667. Levene, M.J., Korlach, J., Turner, S.W., Foquet, M., Craighead, H.G., and Webb, W.W., 2003,

Zero-mode waveguides for single-molecule analysis at high concentrations. Science. 299:682. Levin, M.K., and Carson, J.H., 2004, Fluorescence correlation spectroscopy and quantitative cell

biology, Z)(^re«/. 72:1. Licht, S.S., Sonnleitner, A., Weiss, S., and Schultz, P.G., 2003, A rugged energy landscape

mechanism for trapping of transmembrane receptors during endocytosis, Biochem. 42:2916. Lieto, A.M., Cush, R.C., and Thompson, N.L., 2003, Ligand-receptor kinetics measured by total

internal reflection with fluorescence correlation spectroscopy, Biophys. J. 85:3294. Lieto, A.M., and Thompson, N.L., 2004, Total internal reflection with fluorescence correlation

spectroscopy: nonfluorescent competitors, Biophys. J. 87:1268. Maliwal, B.P., Malicka, J., Gryczynski, L, Gryczynski, Z., and Lakowicz, J.R., 2003, Fluorescence

properties of labeled proteins near silver colloid surfaces, Biopoly. 70:585. Mashanov, G.I., Tacon, D., Peckham, M., and Molloy, J.E., 2004, The spatial and temporal

dynamics of pleckstrin homology domain binding at the plasma membrane measured by imaging single molecules in live mouse myoblasts, J. Biol. Chem. 279:15274.

McCain, K.S., and Harris, J.M., 2003, Total internal reflection fluorescence-correlation spectroscopy study of molecular transport in thin sol-gel films, Anal. Chem. 75:3616.

McCain, K.S., Schluesche, P., and Harris, J.M., 2004a, Modifying the adsorption behavior of polyamidoamine dendrimers at silica surfaces investigated by total internal reflection fluorescence correlation spectroscopy, Anal. Chem. 76:930.

McCain, K.S., Schluesche, P., and Harris, J.M., 2004b, Poly(amidoamine) dendrimers as nanoscale diffusion probes in sol-gel films investigated by total internal reflection fluorescence spectroscopy. Anal. Chem. 76:939.

Muller, J.D., Chen, Y., and Gratton, E., 2003, Fluorescence correlation spectroscopy, Meth. Enzymol. 361: 69. Palmer, A.G., and Thompson, N.L., 1987, Molecular aggregation characterized by high order

autocorrelation in fluorescence correlation spectroscopy, Biophys. J. 52:257. Palmer, A.G., and Thompson, N.L., 1989a, Fluorescence correlation spectroscopy for detecting

submicroscopic clusters of fluorescent molecules in membranes, Chem. Phys. Lipids 50:253. Palmer, A.G., and Thompson, N.L., 1989b, High-order fluorescence fluctuation analysis of model

protein clusters, Proc. Natl. Acad. Sci. U.S.A. 86:6148. Pramanik, A., 2004, Ligand-receptor interactions in live cells by fluorescence correlation

spectroscopy, Curr. Pharm. Biotech. 5:205.


Rigler, R., and Elson, E.L., 2001, Fluorescence Correlation Spectroscopy: Theory and Applications. Springer, Berlin.

Ruckstuhl, T. and Verdes, D., 2004, Supercritical angle fluorescence (SAP) microscopy. Opt. Exp. 12:4246.

Ruckstuhl, T., and Seeger, S., 2004, Attoliter detection volumes by confocal total-intemal-reflection fluorescence microscopy, Opt. Lett. 29:569.

Sako, Y., and Uyemura, T., 2002, Total internal reflection fluorescence microscopy for single-molecule imaging in living cells. Cell Struct. Func. 27:357.

Schneckenburger H (2005) Total internal reflection fluorescence microscopy: technical innovations and novel applications, Curr. Opin. Biotech. 16:13.

Schwille, P., Meyer-Almes, F.J., and Rigler, R., 1997, Dual-color fluorescence cross-correlation spectroscopy for multicomponent diffusional analysis in solution, Biophys. J. 72:1878.

Schwille, P., 2003, TIR-FCS: staying on the surface can sometimes be better, Biophys. J. 85:2783. Starr, T.E., and Thompson, N.L., 2000, Formation and characterization of planar phospholipid

bilayers supported on TiOi and SrTiOs single crystals, Langmuir 16:10301. Starr, T.E., and Thompson, N.L., 2001, Total internal reflection with fluorescence correlation

spectroscopy: combined surface reaction and solution diffusion, Biophys. J. 80:1575. Starr, T.E., and Thompson, N.L., 2002, Local diffusion and concentration of IgG near planar

membranes: measurement by total internal reflection with fluorescence correlation spectroscopy, J. Phys. Chem. B. 106:2365.

Steyer, J.A., and Aimers, W., 2001, A real-time view of life within 100 nm of the plasma membrane. Nature Rev. Molec. Cell Biology 2:268.

Sund, S.E., and Axelrod, D., 2000, Actin dynamics at the living cell submembrane imaged by total internal reflection fluorescence photobleaching, Biophys. J. 79:1655.

Thompson, N.L., Burghardt, T.P., and Axelrod, D., 1981, Measuring surface dynamics of biomolecules by total internal reflection fluorescence with photobleaching recovery or correlation spectroscopy, Biophys. J. 33:435.

Thompson, N.L., 1982, Surface binding rates of nonfluorescent molecules may be obtained by total internal reflection with fluorescence correlation spectroscopy, Biophys. J. 38:327.

Thompson, N.L., and Axelrod, D., 1983, Immunoglobulin surface-binding kinetics studied by total internal reflection with fluorescence correlation spectroscopy, Biophys. J. 43:103.

Thompson, N.L., and Burghardt, T.P., 1986, Total internal reflection fluorescence: measurement of spatial and orientational distributions of fluorophores near planar dielectric interfaces, Biophys. Chem. 25:9\.

Thompson, N.L., 1991, Topics in Fluorescence Spectroscopy Volume 1, Plenum Press, New York, pp. 337-378.

Thompson, N..L, Pearce, K.H., and Hsieh, H.V., 1993, Total internal reflection fluorescence microscopy: application to substrate-supported planar membranes, Europ. Biophys. J. 22:367.

Thompson, N.L., Lieto, A.M., and Allen, N.W., 2002, Recent advances in fluorescence correlation spectroscopy, Curr. Opin. Struct. Biol. 12:634.

Thompson, N.L., and Pero, J.K., 2005, Fluorescence Spectroscopy in Biology: Advanced Methods and Their Applications to Membranes, Proteins, DNA and Cells, Springer-Verlag, Berlin, pp79-103.

Toomre, D., and Manstein, D.J., 2001, Lighting up the cell surface with evanescent wave microscopy. Trends Cell Biol. 11:298.

Ueda, M., Sako, Y., Tanaka, T., Devreotes, P., and Yanagida, T., 2001, Single molecule analysis of chemotactic signaling in Dictyostelium cells, Science 294:864.

Vukojevic, V., Pramanik, A., Yakovleva, T., Rigler, R., Terenius, L., and Bakalkin, G., 2005, Study of molecular events in cells by fluorescence correlation spectroscopy. Cell. Molec. Life

Sciences. 62:535. Wallace, M.I., Ying, L.M., Balasubramanian, S., and Klenerman, D., 2000, FRET fluctuation

spectroscopy: exploring the conformational dynamics of a DNA hairpin loop, J. Phys. Chem. B 104:11551.

Weiss, M., and Nilsson, T., 2004, In a mirror dimly: tracing the movements of molecules in living cells. Trends Cell Biol. 14:267.

Widengren, J., Mets, U., and Rigler, R., 1995, Fluorescence correlation spectroscopy of triplet states in solution: a theoretical and experimental study, J. Phys. Chem. 99: 13368.

TIR-FCS 237

Widengren, J., Schweinberger, E., Berger, S., and Seidel, C.A.M., 2001, Two new concepts to measure fluorescence resonance energy transfer via fluorescence correlation spectroscopy: theory and experimental realizations, J. Phys. Chem. A 105:6851.

Yang, T.L., Baryshnikova, O.K., Mao, H.B., Holden, M.A., and Cremer, P.S., 2003, Investigations of bivalent antibody binding on fluid-supported phospholipid membranes: the effect of hapten density, J. Amer. Chem. Soc. 125:4779.

Yang, T.L., Jung, S.Y., Mao, H.B., and Cremer, P.S., 2001, Fabrication of phospholipid bilayer coated microchannels for on-chip immunoassays, Anal Chem. 73:165.

FLUORESCENCE PROBES OF PROTEIN DYNAMICS AND CONFORMATIONS IN FREELY

DIFFUSING MOLECULES

Single-molecule resonance energy transfer and time-resolved fluorescence methods

Carey K. Johnson, Brian D. Slaughter, Jay R. Unruh, and E. Shane Price^

11.1. INTRODUCTION

Over the past decade, powerful new fluorescence methods have emerged giving researchers the ability to map out the conformations and dynamics of single molecules.^"^ These methods are particularly suited to systems such as proteins that are heterogeneous and characterized by complex potential energy surfaces. Particularly promising is the class of single-molecule fluorescence methods that probe freely diffusing molecules in solution. A significant advantage of this approach is the simplicity of sample handling. Because the molecule of interest is freely diffusing, potential pitfalls associated with immobilization of molecules are avoided.

Methods to probe single molecules freely diffusing in solution have their origins in fluorescence correlation spectroscopy (FCS)" " and related techniques that detect fluorescence fluctuations or bursts as fluorophores diffuse through the focal region of a fluorescence microscope. The first theories for fluorescence correlation spectroscopy (PCS) appeared in the early 1970's.' ' The core idea of PCS, illustrated in Pig. 11.1, is the measurement of correlations in fluctuations in the fluorescence signal recorded in a time series. Magde, Elson, and Webb showed that those fluctuations could be correlated with physical behavior of the fluorophore such as diffusion in and out of the focal volume. A breakthrough in PCS sensitivity came when Rigler and coworkers reached the single-molecule detection limit by pairing PCS with

Department of Chemistry, University of Kansas, Lawrence, KS 66045. 239

240 C. K. JOHNSON ETAL,

Figure 11.1. Illustration of FCS. As molecules diffuse through the focused laser beam, they are excited and emit bursts of fluorescence.

confocal microscopy. The result was greatly improved signal-to-noise while minimizing the excitation volume to femtoliter levels.^' ^

Related methods of analysis have emerged to characterize multiple fluorescing species diffusing in solution. These methods include the photon counting histogram (PCH)/^' ' fluorescence intensity distribution analysis (FIDA)/^ and fluorescence intensity multiple distribution analysis (FIMDA).^^

11.2. FLUORESCENCE CORRELATION SPECTROSCOPY TO PROBE PROTEIN DYNAMICS

The focus of this review will be on the application of FCS and related techniques to detect intramolecular dynamics of diffusing species such as proteins and to determine the timescales of intramolecular dynamics. In single channel FCS measurements the autocorrelation function of the fluorescence signal F(t) is computed to characterize the diffusing sample:

C..W-(F(t))^

(1)

The autocorrelation function is typically fit to a function that describes diffusion through a three-dimensional Gaussian observation volume with adjustable parameters including the diffusion time (r^), the local concentration (Q, the focal volume dimensions (r/z), and the effective focal volume iYej])'^' ^^

PROBES OF PROTEIN DYNAMICS AND CONFORMATIONS 241

1 1 1

v..(c) 1 +

Vv^V ^oy

Equation (2), however, describes fluorescence fluctuations resulting only from the diffusion of the fluorophore in and out of the focal volume. Thus, fluorescence fluctuations due to intramolecular dynamics are not accounted for.

11.2.1. FCS Measurements of Intramolecular Dynamics

Difftision is not the only process that can cause fluctuations in the fluorescence signal. Any process associated with a change in the fluorescence intensity can be tracked by FCS. One such process is triplet state dynamics. Triplet blinking was detected by FCS by Widengren and coworkers in the mid 1990's.^^ The FCS autocorrelation ftinction with triplet dynamics was described by incorporation of a time-dependent factor to describe fluctuations due to triplet dynamics:

G,,(r) = Gl° -[(l - T ) + Te' / ^ ] (3)

where Tis the fraction in the triplet state and rf is the sum of the rates of entry and exit from the triplet state. By fitting the FCS autocorrelation data with Eq. (3) it was possible to measure the triplet dynamics. ^

In Eq. (3) the autocorrelation function is factored into two parts, one describing translational diffusion (as in Eq. 2) and the other describing intramolecular dynamics. Separation of the autocorrelation function in this manner introduces the assumption that the intramolecular dynamics and translational diffusion are not correlated. The conditions under which this assumption is valid have been analyzed by Thompson and Palmer. ^ These workers showed that the autocorrelation function is separable if (i) the diffusion is slow compared to the dynamics, or (ii) the diffusion coefficients for all of the fluorescing species are equal. These conditions are met for photophysical phenomena such as triplet blinking or for intramolecular dynamics that either do not significantly change the shape of the molecule or are slow relative to the transit time of a molecule through the beam (TJ in Eq. 2). In contrast, fluorescence fluctuations resulting from large scale protein conformational changes on the time scale of the transit time may be correlated with translational diffusion and in this case Eq. (3) is not valid.

Other intramolecular dynamics can be studied by this approach. For example, microsecond dynamic motions have been detected in intestinal fatty acid binding protein by FCS based on the sensitivity of fluorescein fluorescence to protein conformational fluctuations.^^' ^ Photoinduced isomerization of the cyanine dye, Cy5, has also been detected by FCS. ^

242 C. K. JOHNSON ETAL.

11.2.2. FCS Cross-Correlation Measurements

In the absence of an environmentally sensitive probe whose fluorescence properties change with dynamic fluctuations of the biomolecule, the autocorrelation decay of the fluorescence from the probe reports only the diffusion of the biomolecule. The presence of two different fluorophores with distinct emission spectra opens a new dimension for FCS measurements. The cross-correlation of the signals from two probes by fluorescence cross-correlation spectroscopy (FCCS) allows one to obtain more detailed information about the system. Dual-color FCCS analyzes the correlation between the fluorescence fluctuations of the two fluorophores.^^'^^

The cross-correlation is given by:

where D(t) and Aft) are the fluorescence signals from two fluorophores. Fluorescence fluctuations from the two fluorophores will be correlated if, for example, they are attached to the same diffusing entity and therefore diffuse jointly, while the fluorescence from two fluorophores diffusing independently will not contribute to the cross-correlation function. Thus, in the absence of Forster resonance energy transfer (FRET) between the two probes, the initial amplitude of the cross-correlation curve can be used to learn about the extent of interaction or co-diffusion of two labeled species. FRET between the two probes in this case complicates the analysis of co-diffusion. FCCS measurements can be implemented by simultaneous excitation of two fluorophores, either by superposition of two excitation beams or by excitation with one beam at a wavelength chosen to overlap the excitation bands of both fluorophores. Such methods have been applied to detect phenomena such as protein-protein interactions or protein-ligand binding that cause the two fluorophores to diffuse in a correlated manner.^^

One of the most valuable applications of FCCS, however, may lie in its combination with single-pair FRET (spFRET) to detect intramolecular dynamics. Conformational changes and intramolecular fluctuations that cause changes in the FRET efficiency between the two dyes will lead to an intramolecular contribution to the FCCS function. A fluctuation leading to an increase in FRET efficiency will result in a decrease in D(t) and an increase in A(t), while a fluctuation leading to a decrease in FRET efficiency will result in the opposite. The result is an anti-correlated contribution to the FCCS signal. This is a negative contribution to the cross-correlation function that decays on the timescale of the conformational change. " ' ^

If the intramolecular dynamics and translational diffusion are separable as described above, then the cross-correlations can be fit to determine the time


scale of intramolecular fluctuations. FCCS detection of FRET fluctuations has been demonstrated in several laboratories. In each case, anticorrelated contributions to the cross-correlation were detected. Eid observed microsecond dynamics in the 70S prokaryotic ribosome and in a calmodulin-CFP-YFP "cameleon."^^ Seidel and co-workers detected dynamics on the time scale of ca. 1 ms in the protein syntaxin 1 ?^ These results demonstrate the potential of FCCS for measurement of intramolecular dynamics in proteins. Our application of this method to calmodulin dynamics is described in the next section.

11.2.3. FCS of Calmodulin

In our laboratory we have carried out FCCS measurements on calmodulin (CaM), a calcium signaling protein involved in many cellular fiinctions.^"^ We generated a doubly-labeled CaM construct with AF488 and Texas Red (TR) conjugated to cysteine residues 34 and 110, on opposing globular domains of CaM (denoted CaM-DA).^^ There is negligible direct excitation of TR, thus emission from TR is due solely to energy transfer from AF488.

The attachment of two-probes to the same biomolecule assures co-diffusion, and in the presence of FRET, such a construct can be used to obtain dynamic information on time scales that are difficult to access by alternative methods. In this case, the cross-correlation of signals from AF488 and TR possesses a rising component due to anti-correlation between the two probes. For comparison with the FCCS signal from CaM-DA, we also measured the

c .o " • 4 - *

03 0

O

O

0.10-

0.08-

0.06-

0.04-

0.02-

0.00-

- . ^ Autocorrelation of CaM-AF488

^ " " ^ / i

^

\ j l Ca^'-CaM

Cross Correlation of \fu D and A of CaM-DA iM,

• / ' '-^ V'^^v^rH*^4^1>L" 1 1 ^nHlLii

"" '^^^Hlfcii 1 1 1 1 1 | 1 1 1 r 1 1 1 1 ) 1 1—1 1 1 1 1 1 | 1 1 1 1 1 1 1 / | ' ^

0.01 0.1 1

Time (ms) 10

Figure 11.2. Autocorrelation and cross-correlation curves for Ca ^ CaM. The autocorrelation function for CaM-AF488 shows the time dependence due to translational diffusion, fit to Eq. (2). The cross-correlation for CaM-DA manifests the additional time dependence resulting from anti-correlation of donor and acceptor. The smooth line for the cross-correlation is a fit to Eq. (5).


autocorrelation fixnction for CaM labeled with a single fluorescent dye, Alexafluor 488 (AF488). Direct comparison of the cross-correlation of CaM-DA and the autocorrelation of CaM with a single probe reveals this rising component as a leveling off of the cross-correlation curve relative to the autocorrelation curve (Fig. 11.2).

Because the same experimental parameters were maintained, including microscope objective, excitation power, and detection source, the different time dependence of the two curves is due only to intramolecular dynamics. We fit the cross-correlation function for CaM-DA to a function analogous to that in Eq. (3) above:

G „ ( r ) = Gi : ' ( r ) -E( r ) (5)

where Gaci^) is the time dependence of the autocorrelation function due to diffusion of CaM determined from the autocorrelation of singly-labeled CaM-AF488. Two exponential time constants, TFI and TF2, with amplitudes / and g, respectively, were found necessary to fit the difference between the two curves:

E(r) = (^l-f-e"'^''-g-e"^"j (6)

revealing dynamic motion between the lobes of CaM on the time scale of hundreds of microseconds and a few milliseconds.^^ The autocorrelation curve of CaM-AF488 is shown in Fig. 11.2, along with the cross-correlation curves of CaM-DA. The least-squares fits of the curves according to Eq. (6) are also shown. The resulting time constants, listed in Table 11.1, show that the dynamics change upon Ca^^ binding by CaM. This result suggests that the intramolecular dynamics play a functional role in CaM and that the relationship between dynamics and function for CaM may be as important as the widely recognized structure-function relationship.

An alternative approach to fit complex data is the Maximum Entropy method (MEM). " ^ MEM fits data to a series of time constants with amplitudes selected to minimize y^ while maximizing the smoothness of fit. In the MEM, peaks in the distribution of time constants appear only if their inclusion results in a reduction in y^ that is significant enough to offset the corresponding reduction in the entropy function that results from the extra peak. The MEM program MemExp^^ was used to fit the the ratio of the cross-correlation data by the autocorrelation data, yielding the time dependence of the anticorrelation component:

G (T)

- ^ ^ ^ = E(r) (7)

This method of fitting was judged to be superior to least squares because it removes uncertainty deriving from fitting the autocorrelation decay Gac(T).


Table 11.1. Time constants for CaM dynamics. Ca^^ concentration

150 nM 100 iM

150 nM 100 jiM

^From nonlinear least-squares ^From MEM fit to Eq. (7)

fit to Eq. (6)

T F I ( / / S ) TF2(ms)

least-squares fits^ 515 >10 188 1.4

MEM fits^

398 3.8 150 2.3

Furthermore, this approach requires no a priori assumption regarding the number of time constants expected. An example of data treatment according to Eq. (7) is shown in Fig. 11.3 for Ca^^-CaM. The corresponding MEM fit is also shown in Fig. 11.3 and the peak time constants are listed in Table 11.1. As the decay of the auto-correlation approachs zero, the division of the cross correlation by the auto-correlation begins to fluctuate between values approaching positive and negative infinity. Therefore, the quotient was only calculated and fit out to 10 ms.

The MEM fit is consistent with the least squares fit shown in Fig. 11.2. With both fitting procedures, two time constants of motion between probes attached to CaM were observed, and the faster time constant was slightly faster in the presence of Ca^^ than in the absence of Ca^^ (Table 11.1). The dynamics

1.0.

0.8.

0 • D

.-i 0.6. "5. E <

0.4-1

. • lOO^iMCa'"

- 1 0 1 2 Log Time (ms)

Figure 11.3. MEM for FCCS of Ca^^-CaM-DA according to Eq. (7).


of CaM in the presence of Ca^^ could have important impUcations for its biological function. Such motions may endow CaM with the flexibility needed for rapid conformational change upon Ca^^ binding or upon encountering a target. A second, longer cross-correlation time constant on the order of milliseconds may reflect dynamic motion due to conformation exchange^^ and may be essential for conformational changes in CaM that occur upon target recognition.

11.3. BURST-INTEGRATED SINGLE-MOLECULE ANALYSIS

The high sensitivity of laser-induced fluorescence methods led to attempts to push the detection limit for molecules in solution down to the single-molecule level. Beginning in the late 1980's, reports appeared from several laboratories demonstrating detection of fluorescence bursts from the passage of single molecules through a focused laser beam. ' ^ Once single-molecule burst detection was demonstrated, efforts were also directed at generating additional information from other fluorescence properties of the molecule, such as fluorescence lifetime,'* "' ^ optical anisotropy,'* ' ^^ or spFRET efficiency^ ' "^ as molecules diffuse freely in solution.

The measurement of single-molecule properties affords the opportunity to map out distributions of a physical property, as opposed to an ensemble average. For example, spFRET probes the distribution of distances between donor and acceptor probes rather than merely the average distance. spFRET, first demonstrated in the laboratory of Shimon Weiss^^ has become a powerful technique for examining conformational heterogeneity in DNA, proteins, and

25 29-31 49 50

protem-protem mteractions. ' ' ' Using the CaM-DA construct, we have generated distance distributions for

CaM as a function of Ca^ , pH, ionic strength, and oxidative modification of methionine residues. Fluorescence from freely diffusing CaM-DA molecules was collected in 300-|is bins. Relative donor and acceptor counts were related to the distance between probes and thus between N and C-terminal domains of CaM by:

R = I.

c ( I , - b I J (8)

where I^ and 4 are the fluorescence intensities of donor and acceptor, respectively, b accounts for donor fluorescence detected in the acceptor channel ("bleedthrough"), c corrects for differences in donor and acceptor quantum efficiencies and detection efficiencies, and RQ is the Forster radius, the distance at which the FRET efficiency is 50%. ^ For this technique to be useful and accurate, a number of parameters for the system must be well understood.


Direct excitation of the acceptor would distort the results. Similarly, the extent of donor fluorescence detected in the acceptor channel (the factor b in Eq. 8) must be determined and accounted for. For quantitative, and not merely relative, distance results, it is also important to diagnose possible differences in the brightness of the two probes and in the detection efficiency of the two channels (the parameter c in Eq. 8). In addition, acceptor photobleaching or blinking would cause changes in FRET efficiency unrelated to the distance between probes. Only once all these parameters have been considered can spFRET distance distributions be considered reliable. Fortunately, careful choice of the FRET pair and filter sets can alleviate a number of these concems. For example, there is very small spectral overlap of the emission profiles of the FRET pair in CaM-DA, AF488 and TR, as well as negligible direct excitation of TR by a 488 nm laser line. TR is highly photostable, which reduces the likelihood of it bleaching or blinking during the observation time. However, due to possible uncertainty due to some of these variables, it is prudent to draw conclusions from comparisons of distance distributions obtained under identical experimental conditions.

Distance distributions of CaM-DA at saturating and below saturating Ca^^ are shown in Fig. 11.4. The distance distributions of CaM reveal three components, showing that CaM exists in solution in at least three distinct conformational substates. These substates persist with time bins of up to 800 |is, showing that the interchange among them occurs on the millisecond time scale or longer. The 20-30 A component is consistent with the published crystal structure of compact Ca^^-CaM^ , and the 50-60 A component is consistent with the distance reported between residues 34 and 110 for the crystal structure of Ca^^-CaM with a helical central linker domain.^ ' ^^ Interestingly, the dominant conformation for Ca^^-CaM is not consistent with either of the crystal structures, but is consistent with the idea that CaM has a bend in its central linker, allowing flexibility between domains. The dominant 30-40 A apoCaM conformation is consistent with the distance predicted by the solution structure of apoCaM.^^ It seems likely that the presence of multiple conformation substates serves an important biological function. In binding to various enzymes that it activates, CaM is known to undergo large conformational changes and to adopt a range of different conformations.^ ' ^ The presence of multiple conformational substates may therefore allow CaM to recognize and bind to a diverse range of target enzymes in different binding conformations.

We have also applied spFRET to investigate the response of CaM to solution conditions.^^ Figure 11.4D shows the change in distance distributions with change in Ca^^ concentration. This result shows that the change in average FRET-efficiency upon Ca^^-binding (Fig. 11.4C) is not due simply to a shift in the conformation of CaM, but is due to a decrease in the amplitude of an extended conformation and an increase in amplitude of a compact conformation, as evidenced by the percentage of bursts above 0.95 and below 0.40 in Fig. 11.4D.

The changes in populations upon Ca^^ binding illustrates an important point: the response of CaM to different solution conditions may involve shifts in the equilibrium among the conformational substates, rather than changes in


u (/}

"GO i _ D

CO

o

1

16-

12-

8-

4-

0

•

0 0

•

• • 0

•

• • •

0 0

• o

%E %E

•

0

>0.9! <0.4(

R(A) Free Ca * (M)

Figure 11.4. Single-molecule distance distribution for Ca^^-CaM (A) and ApoCaM (B) generated from fluorescence bursts recorded in 300-|as bins. The change in FRET efficiency with Ca ^ concentration (C) shows a decrease in average between donor and acceptor upon Ca ^ binding. An analysis of bursts reveals (D) changes in populations of extended and compact conformational substates. Reprinted with permission from ref. 59. Copyright (2005) American Chemical Society

Structure of a single conformation. In another example, ensemble studies have shown that at reduced pH, CaM becomes more compact.^^ The single-molecule distribution at reduced pH in Fig. 11.5 shows that the change in the average is not due simply to a shift in conformation, but rather to a nearly complete absence of the extended population at reduced pH. For apoCaM at reduced ionic strength (Fig. 11.5) the distribution becomes very broad, indicative of increased conformational heterogeneity on the sub-millisecond time scale. Oxidative modification of methionine residues changes the average distance between probes at sites 34 and 110 significantly, but again this change is due to an increase in amplitude of the extended conformation of CaM relative to the compact one.

11.4. TIME-REVOLVED FLUORESCENCE MEASUREMENTS

11.4.1. CaM Conformational Substates by Ensemble Time-Resolved Fluorescence Measurements

Although single-molecule burst measurements provide the most straight forward means of detecting conformational distributions, information concerning the distribution of FRET efficiencies is also embedded in the fluorescence decay profile of the donor. Time-resolved detection of donor fluorescence in a FRET construct can therefore, in principle, give the


20 30 40 50 60 70

R(A) Figure 11.5. Dependence of single-molecule distance distributions obtained by spFRET for CaM-DA on pH, ionic strength, and oxidative modification of CaM, showing shifts in populations of conformational substates of CaM. Reprinted with permission from ref 59. Copyright (2005) American Chemical Society

distribution of distances between donor and acceptor. In the past, however, applications of this method have been limited by several factors. First, it is rare to find a FRET donor whose intrinsic decay is single exponential. This is crucial for unambiguous determination of the donor-acceptor distance distribution because any complexity in the fluorescence decay of the donor fluorophore would be complicated by the presence of multiple FRET states. In addition, a heterogeneous lifetime distribution of the donor itself indicates multiple donor quantum yields, further complicating the analysis. A second and related problem is the inability of traditional fluorescence lifetime fitting methods to unambiguously assess the shape of the fluorescence lifetime distribution. This problem can be circumvented by fitting the fluorescence decay by MEM, ^ which makes no assumptions about the shape of the distribution, but minimizes the extent of heterogeneity in lifetime distribution, ensuring that only statistically valid lifetime peaks are observed.^^

The CaM-DA construct described above represents an excellent FRET construct for time-resolved studies, as the AF488 probe is characterized by a single-exponential fluorescence decay with a lifetime of 4.0 ns. In addition, time-resolved FRET can add a significant amount of information about this


12% R = 56A

0.1 1

Lifetime (ns)

Figure 11.6. Lifetime distributions of CaM-DA in calcium loaded (A) and apo (B) forms. Fluorescence decays of AF488 in CaM-DA were recorded by time-correlated single photon counting and analyzed by MEM.

construct due to its high sensitivity to short fluorescence lifetime components, whereas single molecule FRET distributions are limited at short distances due to the poor signal statistics of the donor at these distances. Figure 11.6 shows the MEM lifetime distribution of CaM-DA in the presence of saturating Ca^ . It is clear from Fig. 11.6 that there are three fluorescence decay components corresponding to three conformational states of the protein. The short lifetime distribution is centered around 80 ps, with an amplitude of 61%. This component corresponds to a distance for this conformational state of 24 A. The intermediate lifetime distribution has a 28% amplitude and a center lifetime of 1.1 ns, corresponding to a distance of 40 A. The long lifetime distribution has a 12% amplitude with a center lifetime of 3.0 ns, corresponding to a distance of 56 A. For apoCaM (in the presence of 400 jiM EGTA), the lifetime distribution changes dramatically as seen in Figure 11.6. The amplitude of the short


lifetime component decreases to 28%, while the amplitudes of the intermediate and long lifetime components increased to 47% and 25%, respectively. The distances corresponding to these states do not change by more than 4 A from the distances observed with Ca^^-CaM, but the amplitudes are altered, as observed in the spFRET distributions described above.

To verify that we have detected all of the fluorescence states of CaM-DA (i.e. that there are no completely quenched species), we calculated the average FRET efficiency of the CaM-DA species from the lifetime distribution and compared it to the steady-state FRET efficiency. The average FRET efficiency calculated for the lifetime distribution of CaM-DA in the presence of saturating calcium was 84%, compared to an efficiency of 87%o calculated from steady state fluorescence measurements. In the absence of calcium the average FRET efficiency from the lifetime distribution was 67% compared to a steady-state FRET efficiency of 70%. In both cases, the time-resolved measurements underestimate the energy transfer efficiency by 3%, indicating that any unresolved short-lifetime components consist of at most a small fraction of CaM-DA molecules.

In most respects the MEM distribution agrees with the single-molecule distributions. The MEM analysis differs significantly from the single molecule burst analysis only in the values of the relative amplitudes of the short distance components. The MEM analysis predicts a dominant short component with a distance of 24 A while the single molecule measurements show this component to have a significantly lower amplitude than the intermediate component. One possible explanation for this is the high energy-transfer efficiency associated with this short distance, which leads to poor donor photon counting statistics in the single molecule measurement. Single molecule measurements rely on high signal-to-background levels. States with extreme energy transfer efficiencies (close to 1 or 0), such as the short distance state of CaM-DA, are most affected by background levels. This possibility is further suggested by the fact that the single molecule measurements predict a distance for this state that is 4 A longer than the distance predicted by MEM, possibly indicating that shorter distance species are being eliminated in the single molecule analysis. It is also possible that the amplitudes predicted by the MEM analysis are not fully reliable. While MEM analysis selects the smoothest distribution of lifetimes (i.e. with the highest entropy) from those with similar ^ values, these amplitudes may not correspond to the actual distance distribution. Thus, although MEM analysis unambiguously assesses whether a distinct peak exists in the distribution of lifetimes, it may not be able to accurately quantitate the amplitude of that peak. It therefore seems advisable to undertake both methods of analysis.

As we have shown here, the MEM time-resolved FRET analysis agrees quite well on several important characteristics of the CaM-DA conjugate, reducing the ambiguity in such results. First, the MEM analysis confirms the presence of three conformational substates, and, in particular, the presence of a high energy-transfer efficiency state is verified by the MEM analysis. This indicates that the state with high FRET efficiency state observed in the single molecule distribution is not simply a result of photon counting artifacts or


background cutoff levels. Second, the increase in the short lifetime component upon increase in calcium concentration is verified by the MEM analysis.

11.4.2. Associated-Anisotropy Analysis to Assess the Influence of Dye-Protein Interactions

Single-molecule fluorescence methods necessarily require the use of high quantum yield fluorophores. These dyes are often bulky and have fairly long linkers connecting them to the protein of interest. It has been observed in a number laboratories, including ours, that these dyes typically experience a degree of rotational coupling to the proteins or nucleic acids to which they are attached. " ' ' " This raises an important issue because of the dependence of the energy transfer efficiency on the orientational averaging of the dyes within the fluorescence lifetime.^^ The observation of segmental motion of the dye with respect to the protein or DNA molecule can be accounted for by two alternative models of the reorientational dynamics of the dye. The first is a homogeneous model, which assumes that all of the fluorophores undergo restricted segmental motion of the dye. This motion can be modeled by the motion-in-a-cone model,^ ' ' ^ in which reorientational motion of the dye is restricted to within a cone characterized by its half-angle, Og, within which the dye wobbles freely. Because this motion is restricted, the orientational motion of the dye is also coupled to the protein to which it is attached. In this model, the fluorophore is characterized by a multi-exponential fluorescence anisotropy decay in which the short rotational correlation time corresponds to the correlation time for segmental motion in the cone, and the long rotational correlation times represent tumbling and internal motions of the protein.

Motion within the cone results in partial orientational averaging of the notorious FRET orientational factor K\^^ In the case of homogeneous reorientational dynamics of the fluorophore, the uncertainty in the Forster radius RQ resulting from incomplete rotational averaging can be estimated from the steady-state anisotropics of the donor and acceptor fluorophores.^^' ^ ' ^ If the steady state anisotropy is sufficiently small, the analysis of FRET data can be carried out unhindered by concerns about incomplete orientational averaging of the donor or acceptor dyes. For example, if the steady state anisotropy of the donor is 0.12, and the steady state anisotropy of the acceptor is 0.20, error in the measured distance can be estimated to be ±10 to ±15% or less.^ ' ^ ' ^ It is important to note that these numbers represent the limit where the segmental motion of the dye is fast compared to the fluorescence lifetime, and the overall rotation of the donor-acceptor labeled system does not occur rapidly enough to significantly lower the steady state anisotropy.

There is another possible scenario for reorientational dynamics of the donor or acceptor dyes. In this picture, described by a heterogeneous model, the dye exists in an equilibrium between a state in which it sticks to the protein and is therefore rotationally coupled to the protein, and a state in which it rotates freely. The anisotropy decay of fluorophore in this case is characterized by two


components, one reflecting fast segmental motion of the freely rotating population and the other representing the reorientational dynamics (tumbling and internal motion) of the population stuck to the protein. As in the homogeneous model, the fluorescence anisotropy decay in the heterogeneous model consists of a short component related to segmental motion of the dye and a long component related to overall tumbling. Thus, it is not possible to distinguish between the homogeneous and heterogeneous models based on the anisotropy decay alone.

However, the effect on the distribution of observed FRET efficiencies may be quite different for the two models. In the heterogeneous model, the relative orientations of donor and acceptor transition dipoles would not be orientationally averaged for the population of fluorophores stuck to the protein. Therefore it may not be possible to analyze the measured FRET efficiencies to yield the distances between the donor and acceptor dyes, because knowledge of their relative orientations is typically not available. It is therefore important for the analysis of single-molecule FRET distributions to determine the nature of the interaction between the protein and the dyes.

In order to rule out the possibility that heterogeneity in a FRET distribution is due to heterogeneity in the orientational mobility of the dye, one must verify that different measured FRET efficiencies do not result from states having distinct reorientational mobilities of the dye. One method of carrying out this analysis, demonstrated by Seidel and co-workers, employs a four-channel detection system that measures simultaneously the intensity of parallel and perpendicular polarizations for both the donor and acceptor detection channels."^^ In this way, both donor and acceptor mobilities can be assessed for all FRET states observed. The successful implementation of this method enabled Seidel and co-workers to demonstrate that the distinct FRET states that they observed did not result from heterogeneous orientational mobilities of the fluorophores."^^ Although this technique pushes the limits of single-molecule detection and requires a more complex optical setup, the information gained can be crucial for correct interpretation of the FRET data.

Time resolved associated anisotropy analysis provides another means to determine whether different FRET states are correlated with different orientational mobilities of the fluorophores. This analysis involves measurement of the rotational mobility associated with each of the lifetimes of the donor fluorophore. We have employed this technique in our laboratory to rule out the possibility of dye-protein interactions contributing to FRET heterogeneity for CaM-DA. Associated anisotropy analysis has been described in detail previously.^^ Figure 11.7 compares the predicted associated anisotropy on a system characterized by two fluorescence lifetimes and two rotational correlation times for homogeneous and heterogeneous models of the orientational mobility of the donor fluorophore. If the short fluorescence lifetime state is rotationally mobile, while the long lifetime state is rotationally immobile, then the anisotropy will decay much more slowly than the donor-only anisotropy. In addition, the anisotropy displays a distinct minimum at short times as seen in Fig. 11.7. If the opposite is true, the anisotropy will initially decay more slowly than that of the donor-only species, but at longer


times it will decay much more quickly. As a result, as fluorescence from the short lifetime decays to zero, the anisotropy also decays to zero. In this way, if there are significant differences in the anisotropy decay of different FRET species, the anisotropy decay will show significant differences from the anisotropy decay of the donor in the absence of acceptor.

Figure 11.7 shows the anisotropy decays of CaM-34-AF488 and CaM-DA. The anisotropy decays overlay one another. This result demonstrates that AF488 reorientation follows the homogeneous model, verifying that the FRET distributions of CaM-DA are not a result of AF488-CaM interactions. This does not rule out the possibility that Texas Red-CaM interactions contribute to the FRET distribution. It is not possible to carry out the associated anisotropy

0.4-

>^0.3. a.

2 S 0.2 c <

o.H

0.0

0.3 4

5 10 Time (ns)

Figure 11.7. (A) Predicted anisotropy decays for an orientationally immobile state associated with a long lifetime and an orientationally mobile state associated with a short lifetime (dotted line), an orientationally mobile state associated with a long lifetime and an orientationally immobile state associated with a short lifetime (dashed line), and the case where both lifetimes display the same rotational mobility (solid line). (B) Anisotropy decay curves for CaM-DA (dots), and CaM-34-AF488 (line) consistent with homogeneous orientational mobility states.


analysis on the acceptor dye for several reasons. First, the rapid negative amplitude FRET components in the acceptor fluorescence are statistically much more difficult to resolve than the decaying amplitudes of the donor decay. Secondly, the initial anisotropy of the acceptor depends strongly on the relative orientation of donor and acceptor. Since it is unlikely that the dipoles are always collinear, the initial fluorescence anisotropy of the acceptor will be quite low after energy transfer, further reducing the ability to resolve anisotropy components. Nevertheless, we have shown that the homogeneous model is valid for the donor dye. With the steady state anisotropy of the donor being 0.11, even if the acceptor were completely stuck to the protein, the error in the measured distance would be on the order of ±13%.^^ All of the distances measured here by burst and by MEM analysis are separated by more than 26% of the median values between them. This proves that the dynamics and heterogeneity we observe with CaM-DA are due to heterogeneity in the donor-acceptor distances rather than changes in orientational mobility of the fluorescent dyes.

11.5. CONCLUSIONS

The results described here demonstrate that single-molecule fluorescence spectroscopy of molecules freely diffusing in solution can yield unique information about both the dynamics and conformations of proteins. We used single-molecule FRET to detect the dynamics in CaM on the timescales of lOO's of microseconds and a few milliseconds by cross-correlation analysis of the donor and acceptor signals. In addition, analysis of the distribution of donor-acceptor distances by single-molecule FRET demonstrates the presence of three distinct conformational substates of CaM in solution. Together these results paint a picture of a protein that is dynamic and flexible, accessing a range of distinct conformation substates in solution.

It has further been the purpose of this review to show how time-resolved fluorescence decay measurements can be used in concert with single-molecule measurements. Analysis of the donor fluorescence decay by MEM confirmed the presence of three distinct conformational substates of CaM. The anisotropy decay of the donor in the presence of the acceptor showed that donor orientational motion can be described by a homogeneous orientational model, and thus the conformational substates do not result from distinct dynamic states of the donor fluorophore. The predicted uncertainty in the FRET orientational factor further shows that the observed single-molecule distance distributions cannot result from heterogeneous conformations of the acceptor dye. Thus, conformational substates are an intrinsic feature CaM in solution.


We acknowledge support for this research from NIH ROI GM58715. B.D.S. and J.R.U acknowledge support from the Dynamic Aspects of Chemical


Biology NIH Training Grant (NIH 5 T32 GM08545-09). E.S.P. acknowledges support from the Pharmaceutical Aspects of Biotechnology NIH Training Grant (NIGMS 08359).

11.7. REFERENCES

1. X. S. Xie. Single-molecule spectroscopy and dynamics at room temperature, Ace. Chem. Res. 29(12), 598-606 (1996).

2. W. E. Moemer, and M. Orrit. Illuminating single molecules in condensed matter, Seience 283, 1670-1676(1999).

3. S. Weiss. Measuring conformational dynamics of biomolecules by single molecule fluorescence spectroscopy, Nat. Struet. Biol. 7(9), 724-729 (2000).

4. D. Magde, E. Elson, and W. W. Webb. Thermodynamic fluctuations in a reacting system-measurement by fluorescence correlation spectroscopy, Phys. Rev. Lett. 29, 705-708 (1972).

5. E. L. Elson, and D. Magde. Fluorescence correlation spectroscopy. I. Conceptual basis and theory, Biopolymers 13(1), 1-27 (1974).

6. D. Magde, E. L. Elson, and W. W. Webb. Fluorescence correlation spectroscopy. II. Experimental realization, Biopolymers 13(1), 29-61 (1974).

7. R. Rigler, U. Mets, J. Widengren, and P. Kask. Fluorescence correlation spectroscopy with high count rate and low background: Analysis of translational diffusion, Eur. Biophys. J. 22(3), 169-175(1993).

8. S. T. Hess, S. Huang, A. A. Heikal, and W. W. Webb. Biological and chemical applications of fluorescence correlation spectroscopy: A review. Biochemistry 41(3), 697-705 (2002).

9. P. Schwille, U. Haupts, S. Maiti, and W. W. Webb. Molecular dynamics in living cells observed by fluorescence correlation spectroscopy with one- and two-photon excitation, Biophys. J. 77(4), 2251-2265 (1999).

10. R. Rigler, and J. Widengren. Biosceince 3, 180-183 (1990). 11. Y. Chen, J. D. Muller, P. T. C. So, and E. Gratton. The photon counting histogram in

fluorescence fluctuation spectroscopy, Biophys. J. 77(1), 553-567 (1999). 12. Y. Chen, J. D. Muller, S. Y. Tetin, J. D. Tyner, and E. Gratton. Probing ligand protein binding

equilibria with fluorescence fluctuation spectroscopy, Biophys. J. 79(2), 1074-1084 (2000). 13. P. Kask, K. Palo, D. Ullmann, and K. Gall. Fluorescence-intensity distribution analysis and its

application in biomolecular detection technology, Proc. Natl. Acad. Sci. U.S.A. 96(24), 13756-13761 (1999).

14. K. Palo, U. Mets, S. Jager, P. Kask, and K. Gall. Fluorescence intensity multiple distributions analysis: Concurrent determination of diffusion times and molecular brightness, Biophys. J. 79(6), 2858-2866 (2000).

15. S. R. Aragon, and R. Pecora. Fluorescence correlation spectroscopy as a probe of molecular dynamics, J. Chem. Phys. 64(4), 1791-1803 (1976).

16. J. Widengren, R. Rigler, and U. Mets. Triplet-state monitoring by fluorescence correlation spectroscopy, J. Fluoresc. 4(3), 255-258 (1994).

17. J. Widengren, U. Mets, and R. Rigler. Fluorescence correlation spectroscopy of triplet states in solution: A theoretical and experimental study, J. Phys. Chem. 99(36), 13368-13379 (1995).

18. A. G. Palmer, 3rd, and N. L. Thompson. Theory of sample translation in fluorescence correlation spectroscopy, Biophys. J. 51(2), 339-343 (1987).

19. K. Chattopadhyay, E. L. Elson, and C. Frieden. The kinetics of conformational fluctuations in an unfolded protein measured by fluorescence methods, Proc. Natl. Acad. Sci. U.S.A. 102(7), 2385-2389 (2005).

20. K. Chattopadhyay, S. Saffarian, E. L. Elson, and C. Frieden. Measurement of microsecond dynamic motion in the intestinal fatty acid binding protein by using fluorescence correlation spectroscopy, Proc. Natl. Acad Sci. U.S.A. 99(22), 14171-14176 (2002).

21. J. Widengren, and P. Schwille. Characterization of photoinduced isomerization and back-isomerization of the cyanine dye Cy5 by fluorescence correlation spectroscopy, J. Phys. Chem. A 104(27), 6416-6428 (2000).


22. P. Schwille, P.-J. Meyer-Almes, and R. Rigler. Dual-color fluorescence cross-correlation spectroscopy for multicomponent diffusional analysis in solution, Biophys. J. 72(4), 1878-1886(1997).

23. U. Kettling, A. Koltermann, P. Schwille, and M. Eigen. Real-time enzyme kinetics monitored by dual-color fluorescence cross-correlation spectroscopy, Proc. Natl Acad. Set U.S.A. 95(4), 1416-1420(1998).

24. B. D. Slaughter, M. W. Allen, J. R. Unruh, R. J. B. Urbauer, and C. K. Johnson. Single-molecule resonance energy transfer and fluorescence correlation spectroscopy of calmodulin in solution., J. Phys. Chem. B 108(29), 10388-10397 (2004).

25. M. Margittai, J. Widengren, E. Schweinberger, G. F. Schroder, S. Felekyan, E. Haustein, M. Konig, D. Fasshauer, H. GrubmuUer, R. Jahn, and C. A. M. Seidel. Single-molecule fluorescence resonance energy transfer reveals a dynamic equilibrium between closed and open conformations of syntaxin \,Proc. Natl. Acad. Sci. U.S.A. 100(26), 15516-15521 (2003).

26. J. S. Eid Two-photon dual channel fluctuation correlation spectroscopy: Theory and application Ph.D. Thesis, University of Illinois, Urbana-Champaign (2002).

27. M. W. Allen, R. J. B. Urbauer, A. Zaidi, T. D. Williams, J. L. Urbauer, and C. K. Johnson. Fluorescence labeling, purification and immobilization of a double cysteine mutant calmodulin fusion protein for single-molecule experiments. Anal. Biochem. 325, 273-284 (2004).

28. T. Ha, T. Enderle, D. F. Ogletree, D. S. Chemla, P. R. Selvin, and S. Weiss. Probing the interaction between two single molecules: Fluorescence resonance energy transfer between a single donor and a single acceptor, Proc. Natl. Acad. Sci. U.S.A. 93(13), 6264-6268 (1996).

29. M. Dahan, A. A. Deniz, T. Ha, D. S. Chemla, P. G. Schultz, and S. Weiss. Ratiometric measurement and identification of single diffusing molecules, Chem. Phys. 247(1), 85-106 (1999).

30. J. R. Grunwell, J. L. Glass, T. D. Lacoste, A. A. Deniz, D. S. Chemla, and P. G. Schultz. Monitoring the conformational fluctuations of DNA hairpins using single-pair fluorescence resonance energy transfer, J. Am. Chem. Soc. 123(18), 4295-4303 (2001).

31. A. A. Deniz, T. A. Laurence, G. S. Beligere, M. Dahan, A. B. Martin, D. S. Chemla, P. E. Dawson, P. G. Schultz, and S. Weiss. Single-molecule protein folding: Diffusion fluorescence resonance energy transfer studies of the denaturation of chymotrypsin inhibitor 2, Proc. Natl. Acad Sci. U.S.A. 97(10), 5179-5184 (2000).

32. D. C. Nguyen, R. A. Keller, J. H. Jett, and J. C. Martin. Detection of single molecules of phycoerythrin in hydrodynamically focused flows by laser-induced fluorescence, Anal. Chem. 59(17), 2158-2161 (1987).

33. K. Peck, L. Stryer, A. N. Glazer, and R. A. Mathies. Single-molecule fluorescence detection -auto-correlation criterion and experimental realization with phycoerythrin, Proc. Natl. Acad. Sci. U.S.A. 86(11), 4087-4091 (1989).

34. S. A. Soper, E. B. Shera, J. C. Martin, J. H. Jett, J. H. Hahn, H. L. Nutter, and R. A. Keller. Single-molecule detection of rhodamine-6G in ethanolic solutions using continuous wave laser excitation. Anal. Chem. 63(5), 432-437 (1991).

35. S. A. Soper, Q. L. Mattingly, and P. Vegunta. Photon burst detection of single near-infrared fluorescent molecules. Anal. Chem. 65(6), 740-747 (1993).

36. W. B. Whitten, J. M. Ramsey, S. Arnold, and B. V. Bronk. Single-molecule detection limits in levitated microdroplets, ^wa/. Chem. 63(10), 1027-1031 (1991).

37. M. Sauer, K. H. Drexhage, C. Zander, and J. Wolfrum. Diode laser based detection of single molecules in solutions, Chem. Phys. Lett. 254(3-4), 223-228 (1996).

38. S. Nie, D. T. Chiu, and R. N. Zare. Probing individual molecules in solution with confocal fluorescence microscopy, 5'c/ewce 266, 1018-1021 (1994).

39. U. Mets, and R. Rigler. Submillisecond detection of single rhodamine molecules in water, J. Fluoresc. 4(3), 259-264 (1994).

40. S. A. Soper, L. M. Davis, and E. B. Shera. Detection and identification of single molecules in solution, J. Opt. Soc. Am. B 9, 1761-1769 (1992).

41. C. W. Wilkerson, P. M. Goodwin, W. P. Ambrose, J. C. Martin, and R. A. Keller. Detection and lifetime measurement of single molecules in flowing sample streams by laser-induced fluorescence, Appl. Phys. Lett. 62(17), 2030-2032 (1993).

42. L. Edman, U. Mets, and R. Rigler. Conformational transitions monitored for single molecules in solution, Proc. Natl. Acad Sci. U.S.A. 93(13), 6710-6715 (1996).

258 C. K. JOHNSON ETAL,

43. C. Zander, M. Sauer, K. H. Drexhage, D. S. Ko, A. Schulz, J. Wolfrum, L. Brand, C. Eggeling, and C. A. M. Seidel. Detection and characterization of single molecules in aqueous solution, Appl. Phys. B 63(5), 517-523 (1996).

44. R. Miiller, C. Zander, M. Sauer, M. Deimel, D.-S. Ko, S. Siebert, J. Arden-Jacob, G. Deltau, N. J. Marx, K. H. Drexhage, and J. Wolfrum. Time-resolved identification of single molecules in solution with a pulsed semiconductor diode laser, Chem. Phys. Lett. 262, 716-722 (1996).

45. L. Brand, C. Eggeling, C. Zander, K. H. Drexhage, and C. A. M. Seidel. Single-molecule identification of coumarin-120 by time-resolved fluorescence detection: Comparison of one-and two-photon excitation in solution, J. Phys. Chem. A 101, 4316-4321 (1997).

46. C. Eggeling, J. R. Fries, L. Brand, R. Gunther, and C. A. M. Seidel. Monitoring conformational dynamics of a single molecule by selective fluorescence spectroscopy, Proc. Natl. Acad. Sci. U.S.A. 95(4), 1556-1561 (1998).

47. J. Schaffer, A. Volkmer, C. Eggeling, V. Subramaniam, G. Striker, and C. A. M. Seidel. Identification of single molecules in aqueous solution by time-resolved fluorescence anisotropy, J. Phys. Chem. A 103(3), 331-336 (1999).

48. P. J. Rothwell, S. Berger, O. Kensch, S. Felekyan, M. Antonik, B. M. Wohrl, T. Restle, R. S. Goody, and C. A. M. Seidel. Multiparameter single-molecule fluorescence spectroscopy reveals heterogeneity of hiv-1 reverse transcriptase:Primer/template complexes, Proc. Natl. Acad. Sci. U.S.A. 100(4), 1655-1660 (2003).

49. B. Schuler, E. A. Lipman, and W. A. Eaton. Probing the free-energy surface for protein folding with single-molecule fluorescence spectroscopy, Nature 419, 743-747 (2002).

50. Y. Chen, D. Hu, E. R. Vorpagel, and H. P. Lu. Probing single-molecule t4 lysozyme conformational dynamics by intramolecular fluorescence energy transfer, J. Phys. Chem. B 107(31), 7947-7956 (2003).

51. L. Stryer, and R. P. Haugland. Energy transfer: A spectroscopic ruler, Proc. Natl. Acad. Sci. U.S.A. 58(2), 719-726 (1967).

52. J. L. Fallon, and F. A. Quiocho. A closed compact structure of native calcium-calmodulin. Structure 11(10), 1303-1307 (2003).

53. R. Chattopadhyaya, W. E. Meador, A. R. Means, and F. A. Quiocho. Calmodulin structure refined at 1.7 A resolution, J. Mol. Biol. 228(4), 1177-1192 (1992).

54. M. A. Wilson, and A. T. Brunger. The 1.0 A crystal structure of calcium-bound calmodulin: An analysis of disorder and implications for functionally relevant plasticity, J. Mol. Biol. 301(5), 1237-1256. (2000).

55. H. Kuboniwa, N. Tjandra, S. Grzesiek, H. Ren, C. B. Klee, and A. Bax. Solution structure of calcium-free calmodulin, Nat. Struct. Biol. 2(9), 768-776 (1995).

56. K. P. Hoeflich, and M. Ikura. Calmodulin in action: Diversity in target recognition and activation mechanisms. Cell 108(6), 739-742 (2002).

57. S. W. Vetter, and E. Leclerc. Novel aspects of calmodulin target recognition and activation, Eur. J. Biochem. 270(3), 404-414 (2003).

58. A. Yamniuk, P., and H. Vogel, J. Calmodulin's flexibility allows for promiscuity in its interactions with target proteins and peptides, Mol. Biotechnol. 27(1), 33-58. (2004).

59. B. D. Slaughter, J. R. Unruh, M. W. Allen, R. J. B. Urbauer, and C. K. Johnson. Conformational substates of calmodulin revealed by single-pair fluorescence resonance energy transfer: Influence of solution conditions and oxidative modification. Biochemistry 44(10), 3694-3707 (2005).

60. H. Sun, D. Yin, and T. C. Squier. Calcium-dependent structural coupling between opposing globular domains of calmodulin involves the central hehx. Biochemistry 38(38), 12266-12279 (1999).

61. J. C. Brochon. Maximum-entropy method of data-analysis in time-resolved spectroscopy, Meth. Enzymol 240, 262-311 (1994).

62. R. Swaminathan, and N. Periasamy. Analysis of fluorescence decay by the maximum entropy method: Influence of noise and analysis parameters on the width of the distribution of lifetimes, Proc. Indian Acad. Sci. Chem. Sci. 108(1), 39-49 (1996).

63. J. R. Unruh, G. Gokulrangan, G. H. Lushington, C. K. Johnson, and G. S. Wilson. Orientational dynamics and dye-DNA interactions in a dye-labeled DNA aptamer, Biophys. J. 88, 3455-3465 (2005).


64. A. Dietrich, V. Buschmann, C. MuUer, and M. Sauer. Fluorescence resonance energy transfer (FRET) and competing processes in donor-acceptor substituted DNA strands: A comparative study of ensemble and single-molecule data, Rev. Mol. Biotechnol. 82(3), 211-231 (2002).

65. R. E. Dale, J. Eisinger, and W. E. Blumberg. The orientational freedom of molecular probes. The orientation factor in intramolecular energy transfer, Biophys. J. 26, 161-194 (1979).

66. K. K. Kinosita Jr, S.; Ikegami, A. A theory of fluorescence polarization decay in membranes, Biophys. J. 20(3), 289-305 (1977).

67. E. Haas, E. Katchalski-Katzir, and I. Z. Steinberg. Effect of the orientation of donor and acceptor on the probability of energy transfer involving electronic transitions of mixed polarization. Biochemistry 17, 5064-5070 (1978).

68. B. W. van der Meer. Kappa-squared: From nuisance to new sense, Rev. Mol. Biotechnol. 82(3), 181-196(2002).

69. L. Brand, J. R. Knutson, L. Davenprot, J. M. Beechem, R. E. Dale, D. G. Walbridge, and A. A. Kowalczyk. in Spectroscopy and the dynamics of molecular biological systems, edited by P. M. Bayley, and R. E. Dale (Academic Press, London, 1985), pp. 259-305.

BIOLOGICAL APPLICATION OF FLIM BY TCSPC

Axel Bergmann and Rory R. Duncan*^

12.1. INTRODUCTION

Since their recent introduction, confocal (see Minsky 1988) and multi-photon laser scanning microscopes (Denk 1990) have initiated a breakthrough in biomedical fluorescence imaging. The high image quality obtained in these instruments mainly results from the fact that of out-of-focus light is strongly suppressed by a pinhole, or - in case of two-photon excitation - not excited. As a result, high contrast images are obtained, permitting true 3D imaging. Moreover, the scanning technique makes detection in several wavelength channels and multi-spectral detection relatively easy. More features, such as excitation wavelength scanning, polarization imaging, and second-harmonic imaging have been added in the recent years. These multi-dimensional features make laser scanning microscopes an almost ideal choice for steady-state fluorescence imaging of biological samples (Pawley 1995; Biskup 2004; Dumas et al 2004; Day and Schaufele 2005). Importantly, however, the fluorescence of organic molecules is not only characterized by the emission spectrum, but also possesses a distinctive lifetime. Including the fluorescence lifetime in the imaging process provides a direct approach to all effects involving energy transfer between different fluorophores and / or their local environment. Typical examples are the probing of the local environment parameters of a fluorophore via lifetime changes, probing distances on the nanometer scale by FRET, and separation of fractions of the same fluorophore in different binding states to proteins, lipids, or DNA. Fluorescence lifetime imaging is particularly attractive in combination with multi-photon excitation, as these microscopes not only

,* Axel Bergmann: Becker&Hickl GmbH, Nahmitzer Damm 30, 12277 Berlin, Germany. Rory R. Duncan: Centre for Integrative Physiology, University of Edinburgh Medical School, George Square, Edinburgh, EH8 9XD, UK.

261

262 A. BERGMAN and R. R. DUNCAN

provide the required pulsed excitation source (Denk 1990), they also avoid crosstalk of the lifetimes at different depths of thick tissue.

12.2. PHYSICAL BACKGROUND OF FLUORESCENCE LIFETIME IMAGING

The considerations above show that laser scanning fluorescence microscopy is an excellent technique to obtain spatial and spectral information from biological samples. Fluorescence lifetime imaging (FLIM) not only adds a parameter to separate the signals of different fluorophores but also provides a direct approach to all processes involving energy transfer between different fluorophores and / or their local environment. The following paragraph gives a brief summary of the practically relevant effects governing the decay of fluorescence, and their potential application.

12.2.1 Fluorescence Lifetime as a Separation Parameter The most relevant molecular states and relaxation processes of

fluorescent molecules are described here. The ground state is SO, the first excited state SI. By absorption of a photon of the energy SI-SO, the molecule transits into the SI state. A molecule can also be excited by absorbing two photons simultaneously. The sum of the energy of the photons must be at least the energy difference bewteen the SI and the SO state. Simultaneous two-photon excitation requires a high photon flux. Because two photons are required to excite one molecule the excitation efficiency increases with the square of the photon flux. Efficient two-photon excitation requires a pulsed laser and focusing into a diffraction-limited spot. Due to the nonlinearity of two-photon absorption, the excitation is the almost entirely confined to the central part of the diffraction pattern and is thus contained within a very small volume. Higher excited states, S2, S3, do exist, but decay at an extremely rapid rate into the SI state. Moreover, the electronic states of the molecules in condensed matter are strongly broadened by vibration. Therefore, a molecule can be excited by almost any energy higher than the gap between SO and S1. Without interaction with its environment the molecule can return from the SI state by emitting a photon or by internal conversion of the absorbed energy internally into heat. The probability that one of these effects occurs is independent of the time after the excitation. The fluorescence decay function measured at a large number of similar molecules is therefore single-exponential. The lifetime the molecule had in absence of any radiationless decay processes is the 'natural fluorescence lifetime', !„. For molecules in solution the natural lifetime is a constant for a given molecule and a given refraction index of the solvent. Because the absorbed energy can also be dissipated by internal conversion the effective fluorescence lifetime, Xo, is shorter than the natural lifetime, Xn. The

TCSPC IMAGING FOR BIOLOGY 263

'fluorescence quantum efficiency', i.e. the ratio of the number of emitted and absorbed photons, reflects the ratio of the radiative and total decay rate. Using the fluorescence lifetime to separate the signals of different fluorophores may appear somewhat artificial at first glance. A wide variety of fluorophores is available, and normally a sample can be stained with fluorophores with distinct fluorescence spectra. Separating the signals by the emission wavelength is certainly easier than by using the lifetimes (Chen 2004). Nevertheless, FLIM has proved to be useful for imaging histological samples (Bugiel 1989). The fluorescence lifetime is particularly important when it comes to autofluorescence imaging of biological tissue. Usually a large number of endogenous fluorophores are present in tissue, most of which have poorly defined fluorescence spectra (Richards-Kortum 2003). Fluorescence lifetime imaging then becomes a powerful imaging tool (Draaijer 1995). The lifetime as a separation parameter has also been used to track the progress of photoconversion of dyes for photodynamic therapy (Riick 2003), the internalization and aggregation of dyes in cells (Kelbauskas 2002), and to verify laser-based transfection of cells (Tirlapur 2002).

12.2.2. The Fluorescence Lifetime as an Indicator of the Local Environment

An excited molecule can also dissipate the absorbed energy by interaction with another molecule, thus opening an additional return path to the ground state. The fluorescence lifetime, x, becomes shorter than the normally observed fluorescence lifetime, TQ. The fluorescence intensity decreases by the same ratio as the lifetime. The effect is called fluorescence quenching. The quenching intensity depends linearly on the concentration of the quencher. Typical quenchers are oxygen (Gerritsen 1997), heavy metal ions (Lakowicz 1996), and a large number organic molecules (Ameer-Beg 2003). Many fluorescent molecules have a protonated and a deprotonated form (Heikal 2001), isomers (Gautier 2001; Zacharias et al 2002), or can form complexes with other molecules (Duncan 2004). The fluorescence spectra of these species can be virtually identical, but the fluorescence lifetimes may be different, and it is not always clear whether or not these effects are related to fluorescence quenching. In practice, it is only important that for almost all dyes the fluorescence lifetime depends more or less on the concentration of ions, on the oxygen concentration, on the pH value or, in biological samples, on the binding to proteins, DNA or lipids. The lifetime can therefore be used to probe the local environment of dye molecules on the molecular scale, independently of the variable, and usually unknown concentration of the fluorescing molecules. Fluorophores often exists in different conformational or binding states and therefore deliver multi-exponential decay profiles (Gautier 2001; Tramier 2002; Emiliani 2003; Duncan 2004). Typical examples are the mapping of cell parameters (Gerritsen 1997; Centonze 1998; Emiliani 2003; Chen 2004; Duncan 2004; Treanor et al 2005), and probing protein or DNA structures by the environment-dependent lifetime of dyes (Bastiaens and Squire 1999; Chan 1999; Cotlet 2001; Hink et al 2002; Zacharias, Violin et al 2002; Bereszovska


2003; Dumas, Gaborit et ah 2004). The decay components and their intensity factors of endogenous proteins also depend directly on their local environment. An example is NADH whose lifetime increases from 400 - 600 ps to 1.4 - 2.4 ns upon binding to proteins (Lakowicz 1992).

12.2.3 Fluorescence Resonance Energy Transfer A particularly efficient energy transfer process is fluorescence

resonance energy transfer, or FRET. FRET describes an interaction of two fluorophore molecules, where the emission spectra of one dye (the Donor) overlaps the absorption band of the other (the Acceptor). In this case the energy from the donor is transferred immediately to the acceptor. The energy transfer itself does not involve any light emission and absorption. Forster resonance energy transfer, or resonance energy transfer (RET), are synonyms of the same effect. FRET results in an extremely efficient quenching of the donor fluorescence and, consequently, a decrease in the donor fluorescence lifetime. The energy transfer rate from the donor to the acceptor decreases with the sixth power of the distance. Therefore, it is apparent only at distances shorter than 10 nm (Stryer and Haugland 1967; Stryer 1978). At the critical distance where 50% of the donor energy is transferred to an acceptor - the Forster radius (Forster 1948) - the donor emission and fluorescent lifetime are each reduced by 50%, and sensitized emission {acceptor emission specifically under donor excitation) is increased. FRET has become an important tool in cell biology (Bastiaens and Squire 1999; Periasamy 1999; Hink, Bisselin et al 2002; Jares-Erijman and Jovin 2003; Niggli and Egger 2004; Day and Schaufele 2005; Voss et al. 2005). FRET in cell biology is used commonly to verify whether labeled proteins are physically linked: by measuring the FRET "efficiency", distances on the nm scale can thus be determined. As protein complexes have diameters on the nanometer range, the detection of FRET can be assumed to report a molecular interaction. An obvious difficulty in steady-state FRET measurements (i.e. intensity-based measurements) in cells is that the concentrations of the donor and acceptor are variable and unknown, the emission band on the donor extends into the emission band of the acceptor, and the absorption band of the acceptor extends into the absorption band of the donor. A further complication is that usually only a fraction of the donor molecules are linked with an acceptor molecule. These effects are hard to distinguish in steady-state FRET measurements. Nevertheless, a number of FRET techniques based on steady-state imaging have been developed (Gordon et al 1998; Periasamy 1999; Xia and Liu 2001; Gu 2004). The techniques need several measurements, including images of cells containing only the donor and the acceptor, or are destructive and therefore not applicable to living cells. FLIM-based FRET techniques have the benefit that the results are obtained from a single lifetime image of the donor (Bastiaens and Squire 1999; Wouters and Bastiaens 1999; Tramier 2002; Duncan 2004; Peter et al. 2005; Wallrabe and Periasamy 2005). These approached do not need calibration by different cells, and are non-destructive. Moreover, FLIM may able to resolve the interacting and non-interacting donor fractions.


12.3. THE LASER SCANNING MICROSCOPE

The term iaser scanning microscope' is used for a number of very different instruments. Scanning can be accomplished by galvano-driven mirrors in the beam path, by piezo-driven mirrors, by a Nipkow disc, or by a piezo-driven sample stage. This chapter refers to microscopes with fast beam scanning by galvano-driven driven mirrors. The optical principle of these microscopes is shown in Figure 1. Laser scanning microscopes can be classified by the way they excite the fluorescence in the sample, and by the way they detect the fluorescence. Single-photon microscopes use a near-UV or visible continuous wave laser to excite the sample. Two-photon - or 'multiphoton' - microscopes use a femtosecond laser of high repetition rate. The fluorescence light can be detected by feeding it back through the scanner and through a pinhole (albeit one of maximum diameter) . This is termed 'descanned' detection. A second type of detection is achieved by diverting the fluorescence directly behind the microscope objective. The principle is termed 'direct' or 'non-descanned' detection.

12.3.1 Suppression of out-of-focus light

One of the most relevant features of the scanning technique is its suppression of out-of-focus light and sectioning capability. Scanning in combination with confocal detection and, more efficiently, two-photon excitation, also reduces lateral crosstalk. These features become particularly important in combination with FLIM. Mixing the decay functions of different pixels or focal planes must be avoided to obtain clean lifetime results. Lateral and vertical crosstalk is avoided by point-detection FLIM techniques, which are therefore the first choice for the laser scanning microscope.

12.3.2 Scan Rates

Commercial laser scanning microscopes scan the sample with pixel dwell times down to a few 100 ns. There are two reasons for the high scanning rate. The first one is that a high frame rate is required to record fast image sequences. Of course, single frames recorded at pixel dwell times this short deliver a poor signal-to-noise ratio. However, image correlation techniques are able to recover transient effects even from the sequence of extremely noisy images. Although generally possible, image correlation techniques have not been used in conjunction with FLIM yet. Therefore the second benefit of high scan rates is more important: At the high excitation power density used in scanning a considerable fraction of the fluorophore is accumulated in the triplet state. Molecules in the triplet state do not fluoresce and are lost for the build-up of the image. Typical triplet lifetimes are in the range from 10 to 100 ^s. Fast scanning therefore reduces the fraction of the fluorophore molecules in the triplet state at the current scan position.


12.3.3 Two-Photon Excitation with Direct Detection

With a titanium-sapphire (Ti:Sa) laser or another high-repetition rate femtosecond laser the sample can be excited by simultaneous multi-photon absorption (Buurman 1992; Biskup 2004). Two-photon excitation is used almost exclusively for biological application; nevertheless, such microscopes are normally called 'Multi-photon' microscopes. The excitation wavelength is twice the absorption wavelength of the molecules to be excited. Multiphoton excitation (MPE) (Denk 1990; Xu et al 1996; Straub et al 2000; White et ah 2001; Cahalan et al 2002) uses a pulsed laser to provide ultra-short (femtosecond duration), rapid (megahertz repetition rates) pulses of excitation energy (Cahalan, Parker et al. 2002), meaning that the average absorbed excitation energy is lower than in conventional laser scanning microscopy. This follows the principle of two-photon excitation (TPE), allowing the use of near-infra red excitation energy, which can be less phototoxic to cells than conventional laser energy. Because two photons of the excitation light must be absorbed simultaneously the excitation efficiency increases with the square of the excitation power density. Due to the high power density in the focus of a high numerical aperture microscope objective and the short pulse width of a titanium-sapphire laser, two-photon excitation works with remarkable efficiency. Excitation is obtained essentially in the volume of the diffraction pattern around the geometric focus of the objective lens. Consequently, depth resolution is an inherent feature of two-photon excitation, even if no pinhole is used. Since the scattering and the absorption at the wavelength of the two-photon excitation are small the laser beam penetrates through relatively thick tissue. The loss on the way through the tissue can easily be compensated by increasing the laser power. The increased power does not cause much photo-damage because the power density outside the focus is small. The fluorescence emission has a shorter wavelength than the excitation photons and the scattering coefficient at the fluorescence wavelength is higher. Fluorescence photons from deep tissue layers therefore emerge from a relatively large area of the sample. To make matters worse, the surface is out of the focus of the objective lens, and thus the fluorescence from deep tissue layers cannot be efficiently focused into a pinhole. The preferred detection technique for two-photon imaging is therefore direct (or non-descanned) detection. Direct detection splits off the fluorescence light immediately behind the microscope lens and directs it to a large-area detector. Consequently, acceptable light collection efficiency is obtained even for deep layers of highly scattering samples. Two-photon imaging with non-descanned detection can therefore be used to image tissue layers several 100 jiim (in extreme cases 1 mm) deep (Berland 1995; Gratton 2003). The absence of a pinhole in a two-photon microscope with non-descanned detection makes the optical path relatively easy to align. Two-photon microscopes can be built by upgrading a one-photon system or by attaching an optical scanner to a conventional microscope (Biskup 2004; Biskup 2004). The downside of the large light-collection area of non-descanned detection is that the systems are very sensitive to daylight. For thin samples, such as single cells, two-photon excitation is therefore used also with descanned detection. The


pinhole is usually opened wide (up to 1 mm), and mainly used to suppress daylight leaking into the objective lens.

12. 4. REQUIREMENTS FOR FLUORESCENCE LIFETIME IMAGING IN SCANNING MICROSCOPES

12.4.1 Efficiency

An ideal lifetime detection technique would record the fluorescence decay function without loss of photons, over a time interval much longer than the fluorescence decay time, in a large number of time channels, and with an infinitely short temporal instrument response function. The standard deviation of the fluorescence lifetime for a number of recorded photons, N, would be

and the signal-to-noise-ratio, SNR

That means, a single-exponential fluorescence lifetime can ideally be derived from a given number of photons per pixel with the same accuracy as the intensity. A lifetime accuracy of 10% can be obtained from only 100 photons. However, the required N increases dramatically for multi-exponential lifetime analysis. The resolution of double-exponential decay profiles requires at least 1000 photons per pixel (Kollner 1992), depending on the ratio of the lifetimes and intensity factors (Cotlet 2001). Therefore, the required number of photons in FLIM is normally larger than in steady-state imaging. Obtaining a large number of photons from the sample means either long exposure or high excitation power. Therefore photobleaching (Eliceiri 2003) and photodamage (Chemomordik 2002) become a problem in precision FLIM experiments.

It is therefore important that a lifetime detection technique comes as closely as possible to the ideal signal-to-noise ration (SNR) for a given number of detected photons. The efficiency of a lifetime technique is often characterized by the 'Figure of Merit', F. The figure of merit compares the SNR (signal-to-noise ratio) of an ideal recording device to the SNR of the technique under consideration:

SNK,^,

The loss of SNR in a real technique can also be expressed by the counting efficiency. The counting efficiency, E, is the ratio of the number of photons ideally needed and the number needed by the considered technique:

E = l /F^


It should be noted that the practically achieved values of F and E also depend on the numerical stability of the lifetime analysis algorithm. Moreover, F was originally defined for a single-detector device and single exponential decay. The definition of F is therefore not directly applicable to multi-wavelength TCSPC and multi-exponential decay analysis.

12.4.2 Principle of Time-Correlated Single Photon Counting The TCSPC technique makes use of the fact that for low level, high

repetition rate signals the light intensity is so low that the probability to detect one photon in one signal period is far less than one. Therefore, the detection of several photons in one signal period can be neglected. It is then sufficient to record the photons, measure their time in the signal period, and build up a histogram of the photon times (Becker 2001; Cole 2001; Cotlet 2001; Deprez 2001; Becker 2002; Bird 2004). The detector signal is a train of randomly distributed pulses corresponding to the detection of the individual photons. There are many signal periods without photons, other signal periods contain a single photon. Periods with more than one photon are very rare. When a photon is detected, the time of the corresponding detector pulse in the signal period is measured. The events are collected in a memory by adding a ' 1' in a memory location with an address proportional to the detection time. After many photons, in the memory the distribution of the detection times, i.e. the waveform of the optical pulse builds up. Although this principle looks complicated at first glance, TCSPC records light signals with an amazingly high time resolution and near-ideal efficiency. As mentioned above, the time-resolution is limited only by the transit-time-spread of the detector. With multichannel plate (MCP) PMTs a width of the instrument response function shorter than 30 ps is achieved. The drawback of classic TCSPC devices was the limited speed of the nuclear instrumentation modules (NIMs) used for signal processing. The slow acquisition can be considered a feature of the early instruments. A more severe drawback is, however, that the principle is intrinsically one-dimensional. It only delivers the intensity versus time. Its application to laser scanning systems therefore requires the recording of a full fluorescence decay curve in one pixel, the read out of the data, then the next pixel, and so on. Such systems have indeed been used for FLIM (Becker 2004), but were restricted to slow scanning and low count rates. A new generation of TCSPC devices abandoned the NIM technique entirely and integrated all the building blocks on a single printed circuit board. The electronic system was optimized as a whole, resulting in time-shared operation of TAC, ADC and memory access. Together with new time-to-digital conversion principles the count rate of TCSPC was increased by two orders of magnitude. Moreover, advanced TCSPC devices use a multidimensional histogram process. They record the photon density not only as a function of the time in the signal period, but also of other parameters, such as the wavelength, spatial coordinates, location within a scanning area, the time from the start of the experiment, or other externally measured variables.


12.4.3 Imaging by Multi-Dimensional TCSPC

Multi-dimensional TCSPC is used for fluorescence lifetime imaging in laser scanning microscopes. At the input of the detection system are a number of photomultipliers (PMTs), typically detecting the fluorescence signal in different wavelength intervals. As mentioned above, TCSPC is based on the presumption that the detection of more than one photon per laser pulse period in a single detector is unlikely. Under this condition, the detection of several photons in different detectors is unlikely as well. The single-photon pulses of all detectors can therefore be combined into a common timing pulse line and sent through a single time-measurement channel. To identify the origin of the photon pulses a 'router' delivers a digital expression of the number of the PMT in which the photon was detected. The number of the detector is stored in the 'channel' register of the TCSPC device. The times of the photons in the laser pulse period are measured in the time-measurement channel of the TCSPC device. A constant-fraction discriminator removes the amplitude jitter from the pulses. The time is converted into data word via a time-to-amplitude converter, TAC, and a fast analog-to-digital converter, ADC. The principle of the time measurement channel is identical with classic TCSPC. However, new conversion principles have increased the maximum count rate by two orders of magnitude (Bacskai 2003; Duncan 2004). The third building block of the TCSPC device is the scanning interface. The scanning interface receives the scan clock pulses of the scanner in the microscope, and for each photon delivers the location of the laser beam in the scanning area, x and y. The channel number, n, the time in the laser period, t, and the coordinates of the laser spot, x and y, are used to address a histogram memory in which the detection events are accumulated. Thus, in the memory, the distribution of the photon density over x, y, t, and n builds up. As described above for classic TCSPC, the width of the temporal instrument response function is determined mainly by the transit time spread of the detectors. With MCP PMTs an IRF width of less than 30 ps (full width at half-maximum) is obtained. Thus, short lifetimes down to 50 ps can be measured with almost ideal efficiency. Lifetimes down to 10 ps can certainly be measured but are observed only for fluorophores of low quantum efficiency. The data acquisition can be run at any scanning speed of the microscope. Under typical conditions, the pixel rate is higher than the photon count rate. This makes the recording process more or less random. The acquisition process is controlled by the scanner. Therefore, no changes in the microscope hardware or software are required. The regular zoom and image rotation functions of the microscope can be used in the normal way. As many frame scans as necessary to obtain an appropriate signal-to-noise ratio can be accumulated. It should be pointed out that the recording process does not use any time gating, wavelength scanning, or detector multiplexing. Under reasonable operating conditions all detected photons contribute to the result, and a maximum signal-to-noise ratio for a given fluorescence intensity and acquisition time is obtained. For all detector channels the counting efficiency, E, and the figure of merit, F, are close to one. In practice the signal processing of a recorded photon causes a 'dead time' during which the TCSPC electronics is unable to process another photon.


As long as the photon detection rate is small compared to the reciprocal dead time of the TCSPC module the counting efficiency remains close to one. Currently the fastest TCSPC devices have a dead time of 100 ns, and can be reasonably used up to 10 MHz detector count rate.

FLIM on samples with several fluorophores often requires the recording of the fluorescence in different wavelength channels. Wavelength resolution is particularly useful in autofluorescence experiments. Also FRET experiments benefit from simultaneously recording the donor and acceptor fluorescence, and fluorescence anisotropy measurements require the fluorescence to be recorded under an angle of 0° and 90° from the excitation. Photo-bleaching usually precludes the recording these signals consecutively. Multi-detector capability is therefore another important feature of a FLIM technique.

12.5. BIOLOGICAL APPLICATION OF FLIM

12.5.1 Biological FLIM data acquisition

(This chapter describes a typical TCSPC-FLIM system which was used by the authors to perform FRET experiments..) All imaging experiments were performed using a Zeiss LSM 510 Axiovert confocal laser scanning microscope, equipped with a pulsed excitation source (MIRA 900 Ti:Sapphire femtosecond pulsed laser, with a coupled VERDI lOW pump laser (Coherent, Ely, UK)). The laser was tuned to provide a TPE wavelength of 800 nm, which efficiently excited ECFP, without any detectable excitation / emission from EYFP in the absence of FRET from a donor. Live cells on glass coverslips (37 mm) were imaged using an incubation chamber (H. Saur, Reutlingen, Germany) ; fixed cells were mounted using FLUORSAVE (Calbiochem, San Diego, CA). TPE data acquisition was performed using 512 x 512 or 1024 x 1024 pixel image sizes, with 4 x frame averaging, using a Zeiss Plan NeoFLUAR 1.3 NA 40 x oil immersion, or a Zeiss C-Apochromat 1.2 NA 63 x water corrected immersion objective lens. Band pass (BP) and long pass (LP) emission filters were used, as detailed in the text, in conjunction with a Schott (New York, NY) BG39 IR filter to attenuate the TPE light.

12.5.2 TCSPC-FLIM

TCSPC imaging requires that the scan control pulses of the microscope, i.e. the frame clock, line clock, and, if possible the pixel clock pulses be available. All newer microscopes have access to these signals. Although the standard PMTs of the microscope can generally be used for TCSPC they do not yield an instrument response function (IRF) shorter than 500 ps full width half-maximum (fwhm). It is therefore better to attach a fast detector at a suitable optical output of the microscope. TCSPC measurements


were made under 800 nm TPE, using a non-descanned detector (Hamamatsu R3809U-50; Hamamatsu Photonics UK Ltd, Herts, UK) multichannel plate-photomultiplier tube (MCP-PMT), coupled directly to the rear port of the Axiovert microscope and protected from room light and other sources of overload using a Uniblitz shutter (Rochester, NY). This MCP-PMT is a key to measuring very fast fluorescent lifetimes as it achieves a transit time spread (TTS; the limiting factor for TCSPC measurements) of 30 ps, and is free of afterpulses. The count rates due to the dark noise of the detector and ambient room light were 10 - 10 photons per second. The MCP-PMT was operated at 3 kV, and signal pulses were pre-amplified using a Becker & Hickl HFAC-26 26 dB, 1.6 GHz preamplifier. TCSPC recording used the "reversed start stop" approach, with accurate laser synchronisation using a Becker & Hickl SPC-730 card together with a PHD-400 reference photodiode, routinely at 79.4 MHz. In contrast to conventional TCSPC devices, the SPC boards use a novel analog-to-digital (AD) conversion (ADC) technique that cancels the unavoidable errors of an ultra-fast ADC chip. Together with a speed-optimised time-amplitude-converter (TAC), this achieves an overall dead time of only 125 ns per photon. BP and LP filters were used, as detailed in the text, to dissect components of ECFP emission and also to enable spectral separation of donor and acceptor FRET- and sensitized-emissions. 3-6 mm Schott BG39 filters were positioned directly in front of the MCP-PMT. TCSPC recordings were acquired routinely for between 5 s and 25 s, mean photon counts were between 10 - 10^ counts per second. Images were recorded routinely with 128 x 128 pixels, from a 512 x 512 scan, with 256 time bins per pixel, or 256x256 pixels from a 1024 x 1024 image scan with 64 time bins.

12.5.3 FLIM data analysis and FRET Calculations

Off-line FLIM data analysis used pixel-based fitting software (SPCImage, Becker & Hickl), able to import the binary data generated with the FLIM module.

The fluorescence was assumed to follow a multi-exponential decay. In addition an adaptive offset-correction was performed. A constant offset takes into consideration the time-independent baseline due to dark noise of the detector and the background caused by room light, calculated from the average number of photons per channel in front of the rising part of the fluorescence trace. To fit the parameters of the multi-exponential decay to the fluorescence decay trace measured by the system, a convolution with the instrumental response function was carried out. The optimisation of the fit parameters was performed by using the Levenberg-Marquardt algorithm, minimizing the weighted chi-square quantity.

12.5.4 FLIM to measure FRET in cells

Transfected PC 12 or HEK293 cells, expressing ECFP or CY24, were imaged as described using 800 nm TPE, enabling efficient excitation of ECFP,


with no detectable excitation or emission from EYFP in the absence of FRET. The steady-state fluorescence image data revealed ECFP or the CY fiision to be distributed throughout the cell cytoplasm (Fig 2, a). To quantify donor fluorescence lifetime and energy transfer in the fixed-distance construct, we applied TCSPC FLIM to cells expressing the ECFP alone or CY24 constructs, acquiring data from a 512 x 512 pixel image (146 nm x 146 nm pixel dimensions) using 128 x 128 binned TCSPC pixels (i.e. 4 x TCSPC binning) and 256 time bins per pixel. Image acquisition times of 5 s provided a mean photon per pixel count across the entire image, of ~ 100, with a peak count of ~ 1000 photons/pixel for the bright regions of the image. TCSPC data acquisition using different BP or LP filters to separate spectral components of the ECFP emission revealed that ECFP (alone) fluorescent decay data were best fit using the Levenberg-Marquardt algorithm to a bi-exponential decay (average reduced weighted chi-squared residual (x^) value < 1.1), as previously described (Pepperkok et al. 1999; Tramier 2002). These data yielded a long lifetime component of 2.19 ± 0.24 ns (Duncan 2004). A short lifetime component (tl) was present, with lifetimes of 0.42 ±0.12 ns, (Duncan 2004). These combined data yielded a mean time constant valueof 1.57 ± 0.06 ns (mean ± s.d., n = 12; (Duncan 2004)). TCSPC analyses of intra-molecular FRET between tandem ECFP and EYFP moieties revealed a specific, significant decrease in the donor lifetime participating in FRET. However, no decrease in either the long or the short FRET lifetimes could be resolved unless spectral filtering was used to separate the quenched donor emission from the sensitized emission, supporting the conclusion that FRET occurred between the ECFP and EYFP moieties (as EYFP is not directly excited under these conditions). If donor emission was selected using a Zeiss 435-485 IR nm BP filter, the emission-specific decrease in ECFP fluorescence lifetime under FRET conditions was resolved for both lifetime components, thus strengthening the conclusion that the lifetime quenching was due to energy transfer. These intra-molecular FRET data were best fit to a bi-exponential decay, with a statistically significant donor-specific decrease in the mean lifetime from 1.57 ± 0.06 ns (for ECFP alone) to 1.28 ± 0.18 ns (Mann-Whitney rank sums test, p < 0.0001, n = 8) for CY24. Previous work demonstrated that FLIM analyses using ECFP as a donor in FRET reactions are complicated by the donor non-FRET bi-exponential decay (Tramier et al. 2002). The treatment of these data depends upon the physical reason(s) for the existence of the complex decay behavior of ECFP in non-(hetero) -FRET conditions; our calculations assume the existence of two spectroscopically distinct forms of ECFP. As both the long and the short lifetime components are affected by energy transfer in our experiments, we were able to resolve a statistically significant effect of FRET upon the donor mean lifetime. To provide fiirther confirmatory evidence that the donor-specific decrease in the mean fluorescence lifetime was due to energy transfer, we photo-bleached specifically the acceptor, EYFP, fluorophore (Fig 2, b). Photo-bleaching required 500 iterations from a 514 nm laser line, at 100%(~50 laser power, in a defined intracellular region of interest. FLIM imaging after acceptor photobleaching revealed that the mean fluorescence lifetime of the donor.


ECFP, fluorophore had increased within the photo-bleached region (Fig 2, c). These data were plotted as lifetime vs pixel frequency distributions (Fig 2, d), emphasising the appearance of a longer mean donor lifetime (- 1500 ps, comparable with that measured for ECFP alone in a non-FRET system, Table 1) in the image after photo-bleaching. Interestingly, this longer lifetime species is also apparent as a minor peak in the pre-photo-bleached image (in a perinuclear location) and frequency distribution plot, perhaps indicating a folding intermediate of the C Y fusion protein in the endoplasmic reticulum or the Golgi apparatus, revealed as a change in FRET efficiency.

12.5.5 Dual channel FLIM

Almost all measurements of FRET in cells using time domain FLIM have to date used single channel measurements of donor lifetimes (Bastiaens and Squire 1999; Wouters and Bastiaens 1999; Tramier 2002; Duncan 2004; Peter, Ameer-Beg et al. 2005; Wallrabe and Periasamy 2005). Using this approach, a decrease in donor lifetime in the presence of an acceptor is ascribed to FRET. However, a number of physical and chemical processes other than FRET can result in lifetime changes, as described above. These processes are difficult to control experimentally, especially in the complex intracellular milieu. Multi-detector FLIM therefore can provide an additional level of information by quantifying the lifetime behavior of the supposed donor and acceptor simultaneously. We recently applied this approach to a pair of interacting proteins known to interact in the regulation of exocytosis, the process where cells secrete hormones and neurotransmitters (Sudhof 1995). Using the multi-detector approach, we measured both donor (cerulean, an optimized ECFP with a mono-exponential fluorescence decay (Rizzo et al 2004)) and acceptor (EYFP) lifetimes. This approach revealed that the donor lifetime was indeed quenched in the presence of acceptor, and that a bi-exponential lifetime could be measured for the acceptor in pixels containing FRET. Importantly, a lifetime component for the acceptor could be measured with a negative amplitude, indicative of a delayed fluorescence process, and absolutely diagnostic of FRET from the donor. This assay is typically very content rich; the intensity, co-localization, lifetime, FRET, and rate of transfer can be determined for both the donor and acceptor simultaneously and within a (short, sub-minute?) few seconds acquisition time.

12.6. FUTURE PERSPECTIVES

TCSPC-FLIM has reached a stage of maturity where it is accessible to many cell biologists. The advent of multi-detector FLIM with its content-rich assays will increase the interest in the approach both in the basic science fields and in the pharmaceutical industry. However, TCSPC FLIM as described here is very expensive and technically difficult to establish and perform. The


development of new laser sources and simplified platforms will help to increase the uptake of this technology, and hopefully provide answers to important questions in cell biology.

12.7. REFERENCES

Ameer-Beg, S. M., N. Edme, M. Peter, P.R. Barber, T. Ng, B. Vojnovic (2003). "Imaging Protein-Protein Interactions by Multiphoton FLIM." Proc. SPIE 5139: 180-189.

Bacskai, B. J., J. Skoch, G.A. Hickey, R. Allen, B.T. Hyman (2003). "Fluorescence resonance energy transfer determinations using multiphoton fluorescence lifetime imaging microscopy to characterize amyloid-beta plaques." J. Biomed. Opt 8(3): 368-375.

Bastiaens, P. I. and A. Squire (1999). "Fluorescence lifetime imaging microscopy: spatial resolution of biochemical processes in the cell." Trends Cell Biol 9(2): 48-52.

Becker, W., A. Bergmann, C. Biskup, T. Zimmer, N. Klocker, K. Benndorf (2002). "Multi-wavelength TCSPC lifetime imaging." Proc. SPIE 4620: 79-84.

Becker, W., A. Bergmann, K. Konig, U. Tirlapur (2001). "Picosecond fluorescence lifetime microscopy by TCSPC imaging." Proc. SPIE 4262: 414-419.

Becker, W., A. Bergmann., G. Biscotti., A. Ruck (2004). "Advanced time-correlated single photon counting technique for spectroscopy and imaging in biomedical systems." Proc. SPIE 5340: 104-112.

Bereszovska, O., B.J. Bacskai, B.T. Hyman (2003). "Monitoring Proteins in Intact Cells." Science of Aging Knowledge Environment., SAGE KE 14.

Berland, K. M., P.T.C. So, E. Gratton (1995). "Two-photon fluorescence correlation spectroscopy: Method and application to the intracellular environment." Biophys. ^ 68: 694-701.

Bird, D. K., K.W. Eliceiri, C-H. Fan, J.G. White (2004). "Simultaneous two-photon spectral and lifetime fluorescence microscopy." Applied Optics 43: 5173-5182.

Biskup, C , A. Bohmer, R. Pusch, L. Kelbauskas, A. Gorshkov, I. Majoul, J. Lindenau, K. Benndorf, F-D. Bohmer (2004). "Visualization of SHP-1-target interaction." J. Cell. Sci 117:5155-5178.

Biskup, C , L. Kelbauskas, T. Zimmer, K. Benndorf, A. Bergmann, W. Becker, J.P. Ruppersberg, C. Stockklausner, N. Klocker (2004). "Interaction of PSD-95 with potassium channels visualized by fluorescence lifetime-based resonance energy transfer imaging." J. Biomed Opt 9(4): 135-759.

Bugiel, I., K. Konig, H. Wabnitz (1989). "Investigations of cells by fluorescence laser scanning microscopy with subnanosecond time resolution." Lasers in the Life Sciences 3(1): 47-53.

Buurman, E. P., R. Sanders, A. Draaijer, H.C. Gerritsen, J.J.F. van Veen, P.M. Houpt, Y.K. Levine (1992). "Fluorescence lifetime imaging using a confocal laser scanning microscope." Scanning \4: 155-159.

Cahalan, M. D., I. Parker, S. H. Wei and M. J. Miller (2002). "Two-photon tissue imaging: seeing the immune system in a fresh light." Nat Rev Immunol 2(11): 872-80.

Centonze, V. E., J.G. White (1998). "Multiphoton excitation provides optical sections from deeper within scattering specimens than confocal imaging." Biophys J15: 2015-2024.

Chan, B., K. Weidemaier, W-T. Yip, P.F. Barbara, K. Musier-Forsyth (1999). "Intra-tRNA distance measurements for nucleocapsid protein-dependent tRNA unwinding during priming of HIV reverse transcripsion." PNAS 96: 459-464.

Chen, v., A. Periasamy (2004). "Characterization of Two-photon Excitation Fluorescence Lifetime Imaging Microscopy for Protein Localization." Microsc Res. Tech 63: 72-80.

Chemomordik, V., D.W. Hattery, D. Grosenick, H. Wabnitz, H. Rinneberg, K.T. Moesta, P.M. Schlag, A. Gandjbakhche (2002). "Quantification of optical properties of breast tumor using random walk theory." J. Biomed. OptT. 80-87.

Cole, M. J., J. Siegel, S.E.D. Webb, R. Jones, K. Dowling, M.J. Dayel, D. Parsons-Karavassilis, P.M. French, M.J. Lever, L.O. Sucharov, M.A. Neil, R. Juskaitas, T. Wilson (2001).


"Time-domain whole-field lifetime imaging with optical sectioning." J. Microsc 203(3): 246-257.

Cotlet, J. H., S. Habuchi, G. Dirix, M. Van Guyse, J. Michiels, J. Vanderleyden, F.C. De Schryver (2001). "Identification of different emitting species in the red fluorescent protein DsRed by means of ensemble and single molecule spectroscopy." PNAS9S: 14398-5006.

Day, R. N. and F. Schaufele (2005). "Imaging molecular interactions in living cells." Mol Endocrinol.

Denk, W., J.H. Strickler, W.W.W. Webb (1990). "Two-photon laser scanning fluorescence microscopy." Science 248: 73-76.

Deprez, E., P. Tauc, H. Leh, J-F. Mouscadet, C. Auclair, M.E. Hawkins, J-C. Brochon (2001). "DNA binding induces dissociation of the multimetric form of HIV-1 integrase: A time-resolved fluorescence anisotropy study." PNAS9S: 10090-10095.

Draaijer, A., R.Sanders, H.C. Gerritsen (1995). Fluorescence lifetime imaging, a new tool in confocal microscopy. NYC, Plenum Press.

Dumas, D., N. Gaborit, L. Grossin, B. Riquelme, C. Gigant-Huselstein, N. De Isla, P. Gillet, P. Netter and J. F. Stoltz (2004). "Spectral and lifetime fluorescence imaging microscopies: new modalities of multiphoton microscopy applied to tissue or cell engineering." Biorheology AXQ-A): 459-67.

Duncan, R. R., A. Bergmann, M.A. Cousin, D.K. Apps, M.J. Shipston (2004). "Multi-dimensional time-correlated single-photon counting (TCSPC) fluorescence lifetime imaging microscopy (FLIM) to detect FRET in cells." J. Microsc 215(1): 1-12.

Eliceiri, K. W., C.H. Fan, G.E. Lyons, J.G. White (2003). "Analysis of histology specimens using lifetime multiphoton microscopy." J. Biomed. Opt 376-3S0: 376-380.

Emiliani, V., D. Sanvito, M. Tramier, T. Piolot, Z. Petrasek, K. Kemnitz, C. Durieux, M. Coppey-Moisan (2003). "Low-intensity two-dimensional imaging of fluorescence lifetimes in living cells." AppL Phys. left. 83(12): 2471-2473.

Forster, V. T. (1948). "(German)." ««. Phys. 6: 54-75. Gautier, I., M. Tramier, C. Durieux, J. Coppey, R.B. Pansu, J-C. Nicolas, K. Kemnitz, M. Copey-

Moisan (2001). "Homo-FRET microscopy in living cells to measure monomer-dimer transition of GFP-tagged proteins." Biophys. 7 80: 3000-3008.

Gerritsen, H. C, R. Sanders, A. Draaijer, Y.K. Levine (1997). "Fluorescence lifetime imaging of oxygen in cells." J. Fluorescence 7: 11-16.

Gordon, G. W., G. Berry, X. H. Liang, B. Levine and B. Herman (1998). "Quantitative fluorescence resonance energy transfer measurements using fluorescence microscopy." Biophys J 74(5): 2702-13.

Gratton, E., S. Breusegem, J. Sutin, Q. Ruan, N. Barry (2003). "Fluorescence lifetime imaging for the two-photon microscope: Time-domain and frequency domain methods." J. Biomed. 0/7/8(3): 381-390.

Gu, Y., W.L. Di, D.P. Kelsell, D. Zicha (2004). "Quantitative fluorescence resonance energy transfer (FRET) measurement with acceptor photobleaching and spectral unmixing." J. Microsc 162: 162-173.

Heikal, A. A., S.T. Hess, W.W.W. Webb (2001). "Multiphoton molecular spectroscopy and excited-state dynamics of enhanced green fluorescent protein (EGFP): acid-base specificity." Elsevier Chem. Ph^^s. 274: 37 55.

Hink, M. A., T. Bisselin and A. J. Visser (2002). "Imaging protein-protein interactions in living cells." Plant Mol Biol 50i^): 871-83.

Jares-Erijman, E. A. and T. M. Jovin (2003). "FRET imaging." Nat Biotechnol 2\_{\ 1): 1387-95. Kelbauskas, L., W. Dietel. (2002). "Internalization of aggregated photosensitizers by tumor cells:

Subcellular time-resolved fluorescence spectroscopy on derivates of pyropheophorbide-a ethers and chlorin e6 under femtosecond one- and two-photon excitation." Photochem. Photobiol. 76(6): 686-694.

KoUner, M., J. Wolfrum (1992). "How many photons are necessary for fluorescence-lifetime measurements?" P/zy5. Chem. Lett. V. 199-204.

Lakowicz, J. R. (1996). "Emerging applications of fluorescence spectroscopy to cellular imaging: lifetime imaging, metal-ligand probes, multi-photon excitation and light quenching." Scanning Microsc Suppl iO: 213-24.

Lakowicz, J. R., H. Szmacinski, K. Nowaczyk, M.L. Johnson (1992). "Fluorescence lifetime imaging of free and protein-bound NADH." PNAS 89:1271-1275.


Minsky, M. (1988). "Memoir on inventing the confocal microscope." Scanning 10;. 128-138. Niggli, E. and M. Egger (2004). "Applications of multi-photon microscopy in cell physiology."

Front Biosci9\ 1598-610. Pawley, J. (1995). Handbook of biological confocal microscopy. NYC, Plenum. Pepperkok, R., A. Squire, S. Geley and P. I. Bastiaens (1999). "Simultaneous detection of multiple

green fluorescent proteins in live cells by fluorescence lifetime imaging microscopy." Cwrr^/o/9(5): 269-72.

Periasamy, A., R.N. Day (1999). Visualizing protein interactions in living cells using digitizded GFP imaging and FRET microscopy.. Academic Press.

Peter, M., S. M. Ameer-Beg, M. K. Hughes, M. D. Keppler, S. Prag, M. Marsh, B. Vojnovic and T. Ng (2005). "Multiphoton-FLIM quantification of the EGFP-mRFPl FRET pair for localization of membrane receptor-kinase interactions." Biophys J SS(2): 1224-37.

Richards-Kortum, R., R. Drezek, K. Sokolov, I. Pavlova, M. Follen. (2003). Survey of endogenous biological fluorophores. NYC, Basel, Marcel Dekker Inc.

Rizzo, M. A., G. H. Springer, B. Granada and D. W. Piston (2004). "An improved cyan fluorescent protein variant useful for FRET." Nat Biotechnol 22(4): 445-9.

Riick, A., F. Dolp, C. Happ, R. Steiner, M. Beil. (2003). "Time-resolved microspectrofluorometry and fluorescence lifetime imaging using ps pulsed laser diodes in laser scanning microscopes." Proc SPIE 5139: 166-172.

Straub, M., P. Lodemann, P. Holroyd, R. Jahn and S. W. Hell (2000). "Live cell imaging by multifocal multiphoton microscopy." Eur J Cell Biol 79(10): 726-34.

Stryer, L. (1978). "Fluorescence energy transfer as a spectroscopic ruler." Annu Rev Biochem 47: 819-46.

Stryer, L. and R. P. Haugland (1967). "Energy transfer: a spectroscopic ruler." Proc Natl Acad Sci US A Smy 719-26.

Sudhof, T. C. (1995). "The synaptic vesicle cycle: a cascade of protein-protein interactions." Nature 375(6533): 645-53.

Tirlapur, U. K., K. Konig (2002). "Targeted transfection by femtosecond laser." Nature 418: 290-291.

Tramier, M., I. Gautier, T. Piolot, S. Ravalet, K. Kemnitz, J. Coppey, C. Durieux, V. Mignotte, M. Coppey-Moisan (2002). "Picosecond-hetero-FRET microscopy to probe protein-protein interactions in live cells." Biophys. 7 83- 3570-3577.

Treanor, B., P. M. Lanigan, K. Suhling, T. Schreiber, I. Munro, M. A. Neil, D. Phillips, D. M. Davis and P. M. French (2005). "Imaging fluorescence lifetime heterogeneity applied to GFP-tagged MHC protein at an immunological synapse." J Microsc 217(Pt 1): 36-43.

Voss, T. C , I. A. Demarco and R. N. Day (2005). "Quantitative imaging of protein interactions in the cell nucleus." Biotechniques 38(3): 413-24.

Wallrabe, H. and A. Periasamy (2005). "Imaging protein molecules using FRET and FLIM microscopy." Curr Opin Biotechnol 16( 1): 19-27.

White, J. G., J. M. Squirrell and K. W. Eliceiri (2001). "Applying multiphoton imaging to the study of membrane dynamics in living cells." Traffic 2(11): 775-80.

Wouters, F. S. and P. I. Bastiaens (1999). "Fluorescence lifetime imaging of receptor tyrosine kinase activity in cells." Curr Biol 9(19): 1127-30.

Xia, Z. and Y. Liu (2001). "Reliable and global measurement of fluorescence resonance energy transfer using fluorescence microscopes." Biophys JS\(4): 2395-402.

Xu, C , W. Zipfel, J. B. Shear, R. M. Williams and W. W. Webb (1996). "Multiphoton fluorescence excitation: new spectral windows for biological nonlinear microscopy." Proc Natl Acad Sci US A 93(20): 10763-8.

Zacharias, D. A., J. D. Violin, A. C. Newton and R. Y. Tsien (2002). "Partitioning of lipid-modified monomeric GFPs into membrane microdomains of live cells." Science 296(5569): 913-6.

TIME-RESOLVED FLUORESCENCE ANISOTROPY APPLIED TO SILICA SOL-GEL

GROWTH AND SURFACE MODIFICATION

Dina Tleugabulova and John D. Brennan'*

13.1. INTRODUCTION

The sol-gel process involves the hydrolysis and condensation of suitable metal alkoxides to form highly porous metal oxide glasses. Suitable precursors include silicon, titanium, hafnium or zirconium alkoxides, which result in silica, titania, hafnia or zirconia materials. By far, the bulk of studies involving the sol-gel process have focused on silica-based systems, and thus this review will focus exclusively on such systems.

While the sol-gel process appears to be conceptually simple, the process is in fact quite complex, and is dependent on a myriad of factors, including solvent type, water:silicon ratio, pH, ionic strength and temperature. By controlling such factors, one can modulate surface area, pore volume and diameter, density and other properties of the final material. Thus, detailed understanding of the growth processes involved in silica formation is crucial for optimizing the final materials.

In recent years, an area where sol-gel based systems have found increased use is in the formation of silica-based biocomposite materials, particularly in the field of protein-doped silica materials.^'^ Such materials have found significant use in areas such as biosensing, immunoextraction, solid-phase biocatalysis and affinity chromatography. The need for biocompatible silica materials has led to a range of new silica precursors, and to the need for modification of silica surfaces to promote a more compatible environment for entrapped proteins. In

Dina Tleugabulova, John D. Brennan, Department of Chemistry, McMaster University, 1280 Main Street West, Hamilton, Canada, L8S4M1. Phone: 905-525-9140 x. 27033; E-mail: [email protected].

277

278 TLEUGABULOVA ETAL.

developing a biocomposite material, it is important to understand the nature of the interactions between biomolecules and the silica surface, and methods that can be used to control such interactions. Further issues arise when developing sol-gel based bioaffinity phases, including control over small molecule:silica interactions and potential methods for modifying silica surfaces to either be more protein compatible or more resistant to non-selective interactions.

More recently, biologically modified silica nanoparticles (NPs) have been utilized for ultrasensitive bioanalysis.^'"^ Silica NPs have attracted attention owing to the versatility of silica in synthesis of various particle sizes, and the ease of surface modification. Depending on experimental conditions, silica NPs form stable sols,^ polymerized gels^ or mesoporous materials and organized porous solids in the presence of templating agents.^ Methods to control the morphology of such materials are still being explored, but it is generally accepted that the final structure depends on the whole kinetic evolution of the silica, starting from the silica oligomers and particle nuclei formed initially in the silica sol.^

In this review, we summarize our recent work on the use of time-resolved fluorescence anisotropy (TRFA) as a tool to characterize both the growth mechanisms of silica from biocompatible precursors in aqueous solution and the extent of modification of silica surfaces with polymers and organosilanes. We describe some fundamental insights gained into the interpretation of TRFA decays in silica systems, and highlight recent contributions to particle growth and surface modification studies with special emphasis on biocompatible silica precursors, such as diglycerylsilane^DGS), A^-[(3-triethoxysilyl)propyl]glucon-amide (GLTES) ^ and 7V-[(3-triethoxysilyl)propyl]malton-amide (MLTES/ (Figure 13.1). These precursors undergo hydrolysis and condensation at neutral pH under aqueous conditions and provide stabilization of entrapped proteins,^^' ^ improved cure characteristics, less susceptibility to pH effects and lower degrees of shrinkage.' ''*^ We also highlight recent studies on applying TRFA to adsorption of small biomolecules to silica, and suggest future work needed to extend such studies.

13.1.1 Characterization of Silica Growth and Modification

Since silica gels typically form in aqueous or aqueous-alcohol solvent systems, there is a need to characterize silica growth under aqueous conditions in a rapid and nondestructive way. As a response to this need, a growing number of methodologies have been proposed in the last few years for real-time monitoring of the particle growth and structural transitions of silica. These include time-resolved fluorescence anisotropy, ^ small-angle X-ray scattering (SAXS), ' " ' ' ^ time-resolved laser-induced incandes-cence,'^ NMR spectroscopy,^^'^^ electron paramagnetic resonance spectroscopy,^^'^' X-ray diffraction,^^' ^ second harmonic generation, "^ refractive index measurements^^ and time-resolved fluorescence quenching. ^ These methods complement existing techniques (dynamic light scattering, transmission and scanning electron microscopy)^^ regarding the information they provide, although each method

TRFA FOR SOL-GEL STRUCTURE ELUCIDATION 279

has its own advantages and disadvantages in terms of sampling, ease with which measurements can be carried out and in terms of which systems can be studied.

Recently, the potential of TRFA for the analysis of silica colloids has been recognized. Geddes and Birch used TRFA to study the rotational mobility of a long-lifetime cationic fluorescent probe in an aqueous suspension of Ludox

was proposed that the longer correlation time (~ 400 ns) in the TRFA decays of the probe might reflect Brownian rotation of the colloidal silica NPs with surface-attached probe molecules. This hypothesis was later corroborated for other, short-lifetime cationic dyes dispersed in aqueous sodium silicate (SS) and tetramethylorthosilicate (TMOS) over a broad pH range, which has led to the establishment of TRFA as a feasible approach for particle size

10 Oft 10

measurements. ' ' A detailed survey of sol-gel chemistry, advanced TRFA studies involving two-photon excitation and gated sampling, and the use of TRFA for nanoparticle metrology can be found in Geddes' review from 2002.^^ In the present review, we will describe our recent work on the growth mechanisms of silica using the novel silica precursor diglycerylsilane, which allows investigation of even the earliest processes in silica condensation, and hence provides new insights into the mechanism of silica growth.

Besides extending TRFA analysis to the particle-growth kinetics studies,^^' ^ the nanoparticle metrology approach also provided a new framework for interpretation of TRFA decays of fluorescent probes.^^' ' ^' ^ It was realized that the ability to examine changes in anisotropy decays as a function of silica composition could also be used to study the modification of silica surfaces and interactions of both small molecules and biomacromolecules with silica. ^ ' ^ ' ^ As noted above, such interactions are critical to the performance of both protein-doped bioglasses and bioaffinity stationary phases. While TRFA has been widely used to probe the dynamics of proteins within silica, "" ^ and has provided useful insights into the interactions of entrapped proteins with the silica matrix, the size and complexity of proteins results in a wide array of different motions that occur on different timescales. Furthermore, proteins may be entrapped in a range of different environments within a sol-gel derived material, making the overall dynamics picture highly complex and very difficult to interpret from the anisotropy decay. The use of silica particle dispersions removes the problem of multiple environments found in silica gels and leads to easily interpretable dynamic motions for molecules adsorbed onto NPs, while providing sufficient complexity to allow insights into the parameters that affect protein-silica interactions. The TRFA-based analysis of surface modification makes use of the ability of the anisotropy decays of cationic probes dispersed in silica sols to reflect the real-time equilibrium between the free probe in solution, which rotates rapidly, and that ionically bound to the silica NPs, which thus undergoes slow rotational diffusion. As the silica NPs are modified by a non-fluorescent target or a biomolecule, the surface available for the binding of probe decreases, which allows calculation of the extent of silica modification. The use of NPs to examine silica modification and biomolecule:silica interactions are a key topic for this review.

280 TLEUGABULOVA ETAL

CH3CH2]

H.C

'Cr^o. P—Si-O

P

HO

DGS

H O H O W A H O -

o

Si(0Et)3

^Si(0Et)3

MLTES

Figure 13.1. Structures of R6G and biocompatible silica precursors diglycerylsilane (DGS), 7V-[(3-triethoxysilyl)-propyl]gluconamide (GLTES) and 7V-[(3-triethoxysilyl)propyl]maltonamide (MLTES).

13.2. SURVEY OF TRFA THEORY

Motions of molecules are quite complex and have been the subject of thorough theoretical analysis over the years." ^ Here we will only give a brief survey of TRFA methodology applied to the analysis of aqueous silica sol-gels,


focusing on the specific acquisition and analysis methods used in our laboratories.

13.2.1 What Is Measured in TRFA?

With the TCSPC fluorimeter (IBH 5000U, Edinburgh, UK) and a pulsed LED source, the emission intensity of a fluorescent probe is recorded as the actual number of photon counts in / channels (/ = 1, ..., 4095). Rhodamine 6G (R6G, A.ex = 495 nm, Aem = 551 nm) is the most common probe for the ionic labeling of silica NPs, mainly because of its strong binding, well-understood single-lifetime photophysics and the high experimental limiting anisotropy. In the measurement, a sample containing ^ 1 |LIM R6G is irradiated with vertically ("F') polarized light (reference direction). From the photon counts collected in the vertical lyy (/) or horizontal ly^^ (/) planes perpendicular to the polarization

plane of the incident radiation, the intensity histograms are constructed. The measurement is normally carried out to a total intensity difference D(i) of 10" counts

K-l K-1 (l)

D(i)=j;^lyy(i)-Gj^lyfjil) = lx\0'

with both lyy (/) or ly^ (/)collected over the same time period, where K is the number of channels (K = 4096) and G accounts for the different transmission efficiencies of the lyy (/) and lyn (0 components. The difference function is normalized over the total fluorescence intensity decay, also regarded as the sum function S(i), which is constructed from the convolution of the instrument response function P(m) with the excitation light pulse Y(i-mf^

S{i) = Iyyii) + 2GIyHii) = Y.Pim)Y{i-M) + , + b = Ar^5, / = 1,..., -1 (2)

where N the sampling interval, m the zero-time shift, ^ the light-scattering correction parameter and b is the constant background. Equation (2) reflects the dependence of S(i) on the probe emission intensity, the optical quality of the sample and the fitting procedure used to generate S(i).

From the F(i) histograms (F(i) = IVH{I), Jw(i), S(i) or D(i)) the continuous functions Ivnit), I wit), S{t) and D(t) are obtained

F(i)= ^F(t)dt, t,=ih (3)

where h is the time calibration constant (ns/channel). Since the time scale of the emission process is limited to 1-10 ns for most fluorescent probes, the intensities are generally collected over a time window of 50 ns.


From convolution with the instrument response function P(m), the TRFA decay is obtained

jPit)Dit-T)S{t-T)dT

r(t) = (4)

jP(T)S{t-T)ciT 0

Formally, Eq. (4) can be re-written as

(O + lGIy^it) S{t)

The measurement is totally automated and carried out in a simple and reliable way.

Figure 13.2 shows typical experimental r(t) decays collected for R6G in silica sols and gels using the methods described above, and demonstrates the kinds of alterations in anisotropy decay behaviour exhibited upon binding of probes to silica surfaces.

13.2.2 TRFA Data Analysis

Essentially, the measured r(t) decay contains information about the probe motion through its relationship to the Fourier transform of the orientational time correlation function P2(t) of a rotating emission dipole" ^

r{t) = P2{t) = ^{cos'a{t))-^ (6)

where a is the displacement angle between the absorption and emission dipole moments of the probe. Depending on the probe structure, shape and its environment, several mathematical models have been developed to describe P2(0.^^'^^ThQ fit of a particular model to the experimental r(t) decay is usually achieved via an algorithm"^ ' ^ that minimizes the difference between the experimental and simulated data and finds the global minimum in the fitting function. The acceptance criteria for goodness of the fit include a satisfactory

reduced chi-squared value ( ZR ^^) ^^^ ^ random distribution of weighted residuals. Since more than one mathematical solution can potentially exist that will fit the experimental r(t) data equally well,'*^ a statistically good fit may generate meaningless decay parameters. ^ Hence, the fulfillment of the statistical requirements by itself cannot be regarded as a definitive proof of the diffusion model. The validity of the model is judged by its physical significance in describing the system under study.


R6G in water r(0 = roexp(-//^)

R6G

\ ^

10 15 20

R6G-DGS KO = ^o/i exp(-r/^) + ro/2 exp(-//<^2)

• Silica NP

5 10 15 20

R6G-Ludox r(t) = r^f^ exp(-r / <z ) + 0/2 exp(-^ / 2 ) + ^n)

0.4 -

0.3-

0.2 -

0.1 -

n U i

C

r(th

0.4 -

0.3 -

0.2

0.1

n u

(

,

^^^^^^Wg i 1 ' \ 1

) 5 10 15 20

R6G-SS r{t)«To •j^

L|^^^ll^^|d» ^^^^'WBiR • • ^ 5 T ^ » * i ^

• »

1 f ^ 1

3 5 10 15 20 time, ns

Figure 13.2. TRFA decays of R6G in water and in silica sols. SiOa: 0.3 wt%; pH 9.


The rotational diffiision of small and large solutes in the aqueous solution within silica monoliths is determined by the solute structure and properties as well as by the composition, organization and geometry of the silica environment.^^' ^ ' ^ Usually, the silica network restricts the mobility of large solutes, whereas small molecules are able to rotate freely in the water-filled pores, unless their mobility is impaired by electrostatic binding to silica surface.^^ Both the diffusion and electrostatic binding are greatly influenced by the presence of alcohol, which is a byproduct of the tetramethylorthosilicate (TEOS) or TMOS hydrolysis, as discussed by Geddes.^^ This leads to significant differences between the TRFA decays for R6G in TMOS sols and those measured in aqueous, solvent-free silica systems at near-neutral pH, such as sodium silicate.^^ This indicates that observation of the growth or modification of silica materials requires solvent systems that are free from alcohol co-solvents; in such a case the dye should be able to bind strongly to the silica surface.

13.2.3 Rotational Diffusion of Probes in Non-Interacting Environments

13.2.3.1 Model of Rigid Sphere

In diffusion studies, the rotation of a fluorescent probe in water is usually taken as a reference point. The probe molecule is approached as a small rigid body (sphere, ellipsoid or rod) and its motion in water is modeled as isotropic Brownian rotation in an unhindered environment. The simplest model is that of the rigid sphere, "* which describes P2(t) as a single-exponential function

r(0 = Abexp(-//<zJ) (7)

where TQ is the limiting anisotropy at time zero and (/> is the rotational correlation time. The VQ value indicates the initial anisotropy after internal conversion and vibrational relaxation, prior to probe rotation, and should nearly correspond to the steady-state anisotropy of the fluorescent probe in a glassy frozen solvent. The theoretical value of TQ depends upon the nature of the excitation process: maximum values are ro = 0.4 using one-photon excitation"^^ or ro = 0.57 using two-photon excitation.^^ Two-photon excitation provides a higher degree of molecular orientation during photoselection, leading to the higher TQ value. A close agreement between the fitted and the steady-state or theoretical ro values is one of the most powerful criteria of goodness of the fit.

The rotational correlation time (^) indicates how fast the initially polarized emission is randomized due to Brownian diffusion. For the rotating rigid sphere, (/> is related to the viscosity of the medium 77, temperature T and the volume of the sphere F through the classic Debye-Stokes-Einstein relation^^


where k is the Boltzmann constant. As neutral and anionic fluorescent probes do not bind to the silica surface and thus are free to diffuse in the water-filled pores, the viscosity of silica sols and gels can be measured from the correlation time (j) of these probes in the entrapped state.^^

The TRFA decay of R6G in water at 295K (Figure 13.2) gives ^ = 0.16 ± 0.01 ns, ro= 0.38 and XR^ = l.^^'^^The TQ value is in excellent agreement with the steady-state anisotropy of R6G in glassy frozen solvents ^ whereas ^ corresponds to the radius of the R6G sphere of 0.56 nm, ^ in agreement with previous reports.^^

13.2.3.2 Viscosity Measurements

In the absence of interactions with silica surfaces, the diffusion of anionic and neutral fluorescent probes can be affected not only by the sol microviscosity, but also by collisions with other molecules dispersed in sol-gel pores (molecular crowding). The described model of rigid sphere provides theoretical support to study these aspects by TRFA. The TRFA decay of the anionic probe 8-hydroxyl-l,3,6-trisulfonated pyrene (pyranine) entrapped in a sodium silicate sol prior gelation (Table 13.1) shows unexpectedly high molecular mobility in the densely polymerized and aged silica network.^^ The decay is monoexponential, indicating that pyranine samples a single, average fluid environment with no effects from distribution between different environments. The 60% increase in the average rotational correlation time of pyranine upon aging of the SS hydrogel is likely related to shrinkage of the silica network leading to a higher solvent viscosity. These findings have been recently supported by NMR/EPR studies of the rotational mobility of neutral and negatively charged small solutes in TMOS sol-gels. ^ This puts in serious doubt the existence of presumed "viscosity domains",^^' ^ questioned earlier by Geddes,^^ and supports the view that the sol-gel interior is composed of a rigid silica network with relatively uncrowded aqueous pores, in which small molecules can rotate freely.

Since the binding of a probe to the silica surface severely restricts its rotation, the use of the rigid sphere model in the presence of probe-host interactions can be misleading.^^ The binding of cationic probes to silica can be altered in the presence of alcohols or at extreme pH values to generate a fraction of free, silica-unbound probe. The free probe molecules diffuse to solvent-rich regions with a characteristic short correlation time^ ' ^ that can be used for the viscosity measurement. In this approach, the second correlation time reflects the diffusion of the silica-bound probe. As only one time component is available for the viscosity calculation, the presence of multiple viscosity domains, if they exist, might be overlooked.

While the presence of multiple viscosity domains in silica remains an open question, such domains have been detected in aqueous solutions of water-soluble polymers.^^' ' ^ Polymer solutions are good host systems for testing the rotational diffusion of small and large molecules in the absence of rigid


geometrical confinement as found in the silica sol-gel network. The TRFA decay of R6G in diluted aqueous solutions of poly(ethylene oxide) (PEO) is monoexponential, but shows the statistical significant contributions from two correlation times upon increasing the polymer concentration to 30 wt % (see Table 13.1)

r{t) = P^ Gxpi-t/(^) + P2 expH/<zJ2) = ^o/i Qxp(-t/^)-hrj2 Q^Vi'tI(l>2) (9)

where Pi and P2 are the pre-exponential terms (0 < p < ro, Pi+ P2 ^ i"o) representing the extent to which the probe emission is depolarized by each correlation time; ro = {pi + P2) is the experimental initial anisotropy at / = 0 and /i = P\MPi(f>\-^P2(f>2) and/2 = P2MP\(f>\^P2(h% 0 < / < 1, are the fractional contributions of each time component to r(t).

As both (/>i and ^ correspond to faster motions than the typical segmental mobility of a water-soluble polymer,^^ R6G binding to PEO can be ruled out. On the other hand, both (f>i and (/>2 (^1 < ^2) are larger in value than the correlation time of R6G in water, indicating that there are barriers for R6G diffusion in the presence of non-interacting, dynamic polymer coils. Several factors can be responsible for the retarded R6G dynamics in PEO solutions as compared to that in bulk water: the orientational dynamics of polymer chains,^^ the restricted movement of water molecules in the vicinity of polymer chains^^ and/or collisions due to molecular crowding.^^ Since an increase in the polymer concentration usually correlates with the larger fi-action of polymer clusters and aggregates in the solution,^^ 1 and ^ in Eq. (9) have been assigned to the diffusion of R6G in two different environments, outside and inside the polymer coils and clusters. Formally, ^\ and ^ give viscosities of 1-1.7 Pa s x 10 and 4-5 Pa s X 10 (Eq. 8), respectively. These values may report on a combination of above-mentioned effects. The existence of many intermediate "viscosity regions" cannot be ruled out and that would require fitting r(t) to a distribution of correlation times

Table 13.1. TRFA decay parameters for pyranine (1 |iM) in different environments. Decays were acquired with a pulsed violet laser diode (repetition rate, 1 MHz; pulse duration, 100 ps; X-ex = 404 nm) and fitted to Eq. (7). ^ X-em = 515 nm; T= 295K; r, fluorescence Hfetime.

Sample Pyranine in water Pyranine in 50 wt% glycerol Pyranine in SS sol, 3 wt% Si02 Pyranine in SS gel, 3 wt% SiOz Pyranine in aged SS gel, 3 wt% Si02

T, ns 5.25 ± 0.05 4.81 ±0.05 5.25 ± 0.05 5.08 ± 0.03 4.95 ± 0.03

ro 0.380 0.375 0.347 0.349 0.335

^, ns 0.11 ±0.01 0.55 ±0.01 0.12 ±0.01 0.10 ±0.02 0.20 ±0.01

XR

1.01 1.04 1.03 1.05 1.02


Table 13.2. TRFA decay parameters for R6G (1 |LIM) in aqueous PEO (MW 10 000) solutions. Decays were acquired with a pulsed LED (repetition rate, 1 MHz; pulse duration, 1.3 ns; Aex = 495 nm). y, ^2, rotational correlation times; To, limiting anisotropy; XR^ reduced chi-squared from fits to Eq. (9); Aem = 551 nm; 7 = 295K; fluorescence lifetime, 3.87 ± 0.01 ns; g < 0.05 (from fits to Eq. 15)." ^ All errors represent one standard deviation.

Sample (j)i^ns (1)2, ns // RQ XR R6G in water 0.16 ±0.01 - 1 0.380 1.02 R6Gin0.1 wt%PEO 0.15 ±0.01 - 1 0.385 0.90 R6Ginl.0wt%PEO 0.20 ±0.01 1 1 0.350 1.07 R6G in 30 wt% PEO 0.28 ± 0.03 0.95 ± 0.07 0.21 0.359 1.00

K0 = S>^7^^PH/<^7) (10)

where y]y^/=^o ^^^ J ^'^ number of time components. The statistical

resolution of multiple correlation times may be hampered by small Pj values for one or more components. "^ Based on the discussion above, we can conclude that while our data for pyranine rotation in SS materials suggests a single environment based on the good fit to a monoexponential decay model, the possibility of a distribution of viscosity domains in such materials is possible.

13.2.4 Restricted Dynamics in the Presence of Probe-Host Interactions

1.2.4.1 Wobble-in-Cone Diffusion Model

A key consideration for using R6G to probe the growth of silica particles is strong coupling between the rotation of the NP and that of the probe. This requires that the probe not be able to undergo local motion when bound to the silica surface. In general, the interaction between a fluorescent probe and a macromolecule, even one involving a covalent bond, restricts but does not completely remove the independent mobility of the probe molecule relative to the macromolecule. As a result, the probe dipole is subjected to a fast local depolarization with a short correlation time / c ("loc", local) and a slower tumbling of the probe-macromolecule complex as a whole (^g, "g", global) " ' ^

^(0 = fiocro exp(-r / 4 , ) + /^ro exp(-/ / ^) (11)

As the local motion is somewhat independent from global motion, it decreases the fraction of anisotropy available to extract (pg Bxidfg. Hence, if the fraction of local motion is significant, the parameters for the global rotation cannot be accurately measured.^^ Furthermore, if the correlation times for local and global motion are sufficiently similar, as might be expected during early stages of silica growth, then it is possible that neither correlation time will be accurately recovered. Thus, it is imperative that in studies of silica NPs the


extent of local motion be minimal so as to accurately recover the global rotation of the silica particle.

Recent experimental evidence from our labs has shown that R6G does indeed bind rigidly to the surface of silica (see section 2.4.3),^^ which overcomes the problem of local motion found in previous experimental work with covalently attached fluorophores.^^ This is due in part to multi-point contact between the probe and the silica surface, an in part to the rigidity of silica NPs, which removes the problem of internal flexibility that is common for most water-soluble macromolecules/^ As a result, the R6G-silica complex rotates as one rigid sphere of the radius R, which corresponds to the radius of the silica NP.

It is important to ensure that no more than one R6G molecule binds to each NP. If several R6G molecules bind to the same silica particle, it is possible for short range energy transfer to occur between two R6G molecules, which would inevitably lead to significant depolarization and a concomitant alteration in the TRFA parameters. Hence, it is important to use R6G/silica ratios at which no more than one R6G molecule is bound per silica particle.

13.2.4.2 Mixture of Rigid Spheres

In evaluating the growth or modification of silica NPs, it must be noted that the sample will inherently contain a distribution of particle sizes. For a mixture of j isotropic spheres of different sizes, there must be a distribution of correlation times (Eq. 10) due to the particle size distribution. Since the binding of R6G to the silica NP is not covalent, a small amount of unbound dye may also exist in the silica sol and the fraction of free dye may become significant if the adsorption equilibrium is altered by the presence of organic solvent or at extreme pH. Since the unbound dye molecules diffuse much faster than the silica NPs, Eq. (10) can be re-written as

r(0 = /?,exp(-//^) + 5]y^^exp(-//^^) (12)

where (/>] (^ « ^ ) is the correlation time of free R6G and ^ the correlation time of the j silica NP. Since the photophysical properties of R6G (absorption coefficient, emission wavelength, fluorescence lifetime and quantum yield) do not change in the silica-adsorbed state, the fractional fluorescence of each correlation time in Eq. (12) corresponds to the relative concentration of R6G in each environment (free versus bound to different NPs). In aqueous silica sols, R6G is mostly adsorbed to the silica NPs and thus J3j < 0.01.^^' ^

If the silica NPs are normally distributed around a mean radius R, the anisotropy due to depolarization of the silica-bound R6G decays with one average correlation time ^ :

r(t) = /],Qxp{-t/^) + fi2^M-t/^2) (13)


.12,27, 30 where ^ is related to R by the equation

^ = ^ ^ (14) ^ 3kT

An issue with the use of short-lifetime probes, such as R6G, for silica labeling is that it restricts the time scale of measurable motions to ^ < 40 ns, which corresponds to R < 2.5 nm. The binding of R6G to larger NPs (R > 2.5 nm) leads to significant residual anisotropy r^> 0 (Figure 13.2) and thus adds the third term to Eq. (13), addressed as a hidden or non-decaying

4.28„30

component "

r(t) = y i exp(-/l(k) + P2 exp(-//^2) + oo = n)/i exp(-^/^) + ro/2 exp(-^/^2) + ^'b (15)

where g = rjr^ is the fractional contribution from the non-decaying component, fx'^fi^g^ 1 • If the NPs are normally distributed around R > 2.5 nm, the upper part of the normal particle distribution is hidden in the non-decaying component, whereas the lower part of the distribution, which will no longer be normally distributed, but rather will be a truncated distribution, is averaged as ^ . Thus, caution must be used when interpreting the average size of silica NPs based on the value of ^ in cases where g ^ 0, as under these conditions one cannot assume an underlying normal distribution of particle sizes, and indeed a highly asymmetric particle distribution could exist.

13.2.4.3 R6G Mobility in Porous Monolithic Silica

In the absence of surface modification, it is possible to have essentially all R6G bound to the silica surface and thus the (j>\ component in Eq. (15) disappears (/l = 0). This situation has been seen in the TRFA decays of R6G in aqueous silica sols, but is more prevalent in gelled sodium silicate samples. On the other hand, in TMOS sols and gels, f\ ~ 0.7^^ due to the presence of methanol, highlighting the importance of removing alcohol co-solvents.

Gelation of aqueous sols leads to the formation of a continuous network of macroscopic dimensions immersed in a liquid medium exhibiting no steady-state flow.^^ From this definition, it is obvious that most R6G molecules entrapped in the sol-gel compartments of an aqueous silica gel should be bound to the walls of the rigid silica network, rather than to discrete NPs (Figure 13.2). ^ Assuming that this rigid network is immobile, Eq. (11) can be rewritten

KO = fioJQ exp(-^/^^,,) + r^ (16)

As R6G binds rigidly to the silica surface,^ the local motion is absent (fioc = 0) and thus r(t) = Too. Hence, the residual anisotropy reflects not only the binding


of R6G to large silica NPs (Eq. 15), but also senses the binding to the silica network. Importantly, the ability to achieve a situation where r(t) = r^ rules out any possibility for local probe motion owing to "wobbling" or exchange equilibria.^^ Hence, this probe is ideally suited for studies of silica particle growth.

13.3. PARTICLE GROWTH STUDIES

13.3.1 Particle Growth in DGS Sols

Most theoretical models of the growth of silica particle have been developed using TEOS as a silica precursor and propose the nucleation of primary silica NPs (-' 20 Si atoms)^^ and their fast evolution into secondary NPs and higher order aggregates. However, the notion that the primary NPs may actually exist in solution as discrete, stable structures has been viewed with skepticism for many years, mainly because of the extremely fast kinetics of the silica growth in water and the analytical limitations to follow such fast processes occurring on the nanoscale.^^ Prior work by Geddes and Birch demonstrated that the real-time growth of silica NPs in an aqueous SS sol is too fast (- 7 s) to follow the initial stages of nucleation and silica particle growth.^^ Primary silica NPs could be detected in TMOS sols,^^ where their growth occurs over a very limited range, from ~ 0.8 to 1.1 nm in radius, however the later stages in silica particle formation could not be directly observed by TRFA since particles did not grow beyond 1.1 nm.

As part of our research into entrapment of proteins within silica, it became apparent that conventional alkoxysilane precursors, such as TEOS and TMOS, were not ideal for formation of bioglasses. As a result, our group developed the biocompatible silane precursor diglycerylsilane (DGS, Figure 13.1), which releases the protein-compatible solute glycerol as a byproduct of hydrolysis. A somewhat unexpected benefit of DGS was the ability of this precursor to undergo rapid hydrolysis (< 1 min) in aqueous solvent over a broad pH range (pH 5-11),^ removing the need to use multiple pH regimes to initiate hydrolysis and condensation. More importantly, the glycerol evolved in the initial hydrolysis reaction acts to suppress silica condensation reactions that nonnally occur very rapidly at neutral pH in an aqueous environment, thus slowing the rate of formation of the silica NPs. This precursor therefore appeared to be ideal to prove the existence of primary particles, and the overall validity of the nanoparticle metrology approach.

Figure 13.3 shows TRFA decays for R6G-DGS collected after one hour following DGS dissolution and the corresponding residual plots from the fits to Eq. (15). Figure 13.4 shows the fractions of R6G fluorescence in DGS sols associated with ^7, 2 and g as a component ^ , indicating that the DGS sol is mainly composed of primary silica NPs. With time, ^ is increased to 3.7 ns (t = 50 min) and subsequently to 7.45 ns (/ = 150 min). AFM imaging of the diluted DGS sol shows individual particles that are close to spherical in shape and


compact with smooth interfaces (Figure 13.5A). The phase image (Figure 13.5B) shows that the particles contain a dense nucleus (white zone) and a less dense outer shell (dark layer). The radius of the dense nuclei is in the range of 1-2 nm, according to the particle height on the substrate surface, which minimizes tip-convolution effects. The dark outer shell is likely composed of oligosiloxanes and glycerol molecules, which are dispersed in water in the initial sol, but become adsorbed to the particle surface after drying the sample for AFM imaging. These layers would be expected to provide steric barriers to NP aggregation during imaging, but, on the other hand, make the imaging process difficult. The DGS sample sticks to the AFM probe and is dragged across the surface producing streaks. When the silicon substrate containing the DGS sample is heated for 20 min at 70°C, the dark layers around the NPs disappear (Figure 13.6B), indicative of condensation cure. Figures 13.5 and 13.6 clearly show predominance of individual spherical NPs of different sizes in the diluted DGS sol, but provide no evidence for large aggregates. Thus, the AFM data confirm the TRFA data, and provide direct evidence for existence of stable sols composed essentially of primary silica NPs. As the silica NPs are of spherical shape (Figures 13.5 and 13.6), one can estimate the average particle radius from ^ using Eq. (14). Since ;; = 0.894 x 10" Pa s and 7 = 298 K, the average particle radius isR = 0.9 nm in the DGS sol (0.3 wt % SiOi, pH 9.2) 11 min after DGS dissolution. Higher silica concentrations in the initial sol lead to a larger size of the particle nuclei (Figure 13.7). For the next 20 min after DGS dissolution, the initial size of the nuclei continuously increases, whereas the fraction of ^ does not change (Figure 13.4) indicating no further nucleation. The fraction of primary NPs remains essentially constant for at least 3 weeks,^^ while the average radius of these NPs attains a maximum value of ~ 2.8 nm after 7 days, after which the NPs contract slightly in size, owing to continued syneresis. The particle size measurements are consistent with the formation of stable primary NPs in the DGS sol that do not undergo significant aggregation or further growth into secondary NPs. The TRFA decays collected at pH 8.2 show significantly different parameters than those collected at pH 9.2 (Figure 13.3c). In 40 min after DGS dissolution, the fraction of ^ decreases rapidly to zero. Both the initial and maximum ^ values are considerably higher (12 and 20 ns, respectively) than those measured for R6G-DGS at pH 9.2 (Table 13.3), pointing to a larger particle size at pH 8.2. The fraction of ^ initially increases and then decreases (Figure 13.4B). Unlike the situation at pH 9.2, the decreased fractions of ^i and ^ correlate to time-dependent increases in the fraction of the non-decaying component, g. During the first 15 min, g and /^ increase simultaneously at the expense of a decreased/i value. After 15 min, g continues to increase, bu t^ decreases. Overall, this behavior is consistent with more rapid formation and subsequent aggregation of primary NPs as the pH value approaches neutrality.

292 TLEUGABULOVA ETAL

6.5 9.5 12.5 15.5 18.5 21.5 24.5

£, ~u ti k JI • iL U i iU l i i fa I. J

0

-2

6.5 9.5 12.5 15.5 18.5 21.5 24.5

t ime, ns

Hi^yiJIu.kiiliiJiiikiU^t.

6.5 9.5 12.5 15.5 18.5 21.5 24.5

2

0

-2

6.5 9.5 12.5 15.5 18.5 21.5 24.5

t ime, ns

9.5 12.5 15.5 18.5 21.5 24.5 6.5 9.5 12.5 15.5 18.5 21.5 24.5

6.5 9.5 12.5 15.5 18.5 21.5 24.5 time, ns

6.5 9.5 12.5 15.5 18.5 21.5 24.5

time, ns

Figure 13.3. TRFA decays and residual plots for R6G-DGS in 20 mM borate buffer. (A) 0.3 wt% Si02, pH 9.2; (B) 0.3 wt% SiOa, pH 9.2; (C) 3 wt% Si02, pH 8.2 and (D) 3 wt % SiOs, pH 7. Measurements were performed 1 h after DGS dissolution. R6G, l^M; X^x ~ 495 nm; A-em = 551 nm; T = 298 K. The starting point (t = 6.5 ns) corresponds to the prompt response. Reprinted with permission from Ref 31, Copyright 2004 American Chemical Society.


50 75 100

time, min

150

B

0.9

0.7

1 °- o 2 0.3

0.1

.n 1 i ^^^^^ o 1 X l Vi

A —

—

\j

A

9

\J

25 50 75 100

time, min

125 150

Figure 13.4. Fractions of R6G fluorescence in DGS sols associated with ^j (o), ^ (•) andg (A) as a function of time following DGS dissolution. Si02: 0.5 wt%. (A) pH 9.2; (B) pH 8.2. Relative errors on fractions are < 3%. Reprinted with permission from Ref 31, Copyright 2004 American Chemical Society.

Currently, there are two extreme models describing silica particle growth: one model is a "growth-only" model ' ^ based on LaMer's synthesis of monodisperse colloidal particles by homogeneous precipitation.^^ This model neglects aggregation and holds that a narrow distribution of colloids can be achieved by a brief "burst" of nucleation, followed by diffusion-limited growth. Growth is defined as the addition of soluble species directly to the particle surface. Van Blaaderen et al?^ proposed a related model whereby nucleation is

294 TLEUGABULO VA ETAL.

Table 13.3. TRFA decay parameters for R6G-DGS in 20 mM borate buffer, pH 9.2. Decays were recorded in - 10 min-time intervals following DGS dissolution and fitted to Eg. (15). T= 298 K;r = 3.89 ± 0.02 ns; < 0.003.^"

time after DGS dissolution ^7, ns (^2, ns ro _IR_ 0 11 min 20min 29 min 38 min 47 min 56 min 150 min

0.16 + 0.01 0.13 ±0.06 0.21 ±0.08 0.16 ±0.05 0.15 ±0.06 0.17 ±0.07 0.11 ±0.08 0.17 ±0.03

-0.69 ±0.17 1.83 ±0.20 2.81 ±0.12 4.34 ±0.16 3.03 ± 0.20 3.66 ± 0.05 7.45 ± 0.07

1 0.41 0.32 0.10 0.06 0.07 0.04 0.01

0.380 0.336 0.330 0.359 0.346 0.355 0.365 0.356

1.02 0.96 1.00 1.00 0.96 0.98 1.05 1.01

controlled by the aggregation of soluble species, but subsequent growth is controlled by the surface reaction. The other extreme model describing how the particle size distribution matures is an "aggregation-only" model. ' " Much recent evidence provided by ^^Si-NMR, conductometry and photon correlation spectroscopy of the Stober synthesis process supports the aggregation model.^^ Our study of DGS growth at different pH values showed that in the case of DGS, there is no contradiction between these two models; by setting up different initial conditions of the sol, such as the silica concentration and the pH, the particle growth can be driven by one mechanism or the other. Under conditions of low silica concentration (< 0.5 wt % Si02) and at pH 9.2, the only route available for consumption of the monomeric DGS is nucleation; the

Figure 13.5. Height (A) and phase contrast (B) images of the DGS sol dispersed on a silicon surface. SiOz, 0.2 wt %; pH 9.2. Image field, 500 nm. Reprinted with permission from Ref 31, Copyright 2004 American Chemical Society.


nucleation stage is brief and ends up with the formation of « 1 nm radius nuclei, which grow slowly with time until reaching a stable size on the 2-3 nm scale. Upon increasing the silica concentration to 1 wt % Si02, there is clearly an increase in the initial radius of nuclei formed, which subsequently grow to a larger size (> 4.5 nm radius) or aggregate with time.

At pH 8.2, particle formation is much faster and thus the nucleation and aggregation occur simultaneously in the early stage of silica formation. Hence, primary silica particles form rapidly in the early stages of silica formation and further growth occurs by aggregation of these particles. This results in the formation of larger secondary silica structures in the sol. These results highlight the potential for TRFA analysis, since the method is amenable to in situ studies under a variety of solution conditions.

13.3.2 Particle Growth in SS Sols

In protein entrapment studies, the SS sol is prepared by acidification of SS via reaction with Dowex resin.^^ The R6G dye added to the freshly made SS sol (< 3 wt% Si02) exhibits a high steady-state anisotropy indicating the presence of already grown NPs and aggregates.^^ This corroborates previous findings of Geddes and Birch that the real-time growth of silica NPs in aqueous SS sols is too fast to follow the initial stages of nucleation and silica particle growth.^^ The anisotropy value is even higher (r = 0.35) if the SS sol is submitted to conditions that promote fast silica particle growth and aggregation.^^' ^ The R6G dye dispersed in such sols exhibit no fluorescence depolarization and r(t) ^ ro (Figure 13.2), indicating complete restriction of R6G diffusion. As noted

Figure 13.6. Height (A) and phase contrast (B) images of DGS-based NPs dispersed on the silicon substrate and heated at 70° C for 20 min. SiOi, 0.2 wt %; pH 9.2. Image field, 500 nm. Reprinted with permission from Ref 31, Copyright 2004 American Chemical Society.


2.5

E = 2

•§1 .5

re Q.

0 . 5 ^

X X X X X X X >K

X

2 o o o o ^ ^ ^ • • • •

o •

25 50 75 time, min

100 125 150

Figure 13.7. Silica particle radius (± 0.1 nm) in DGS sols (pH 9.2) as a function of time following DGS dissolution. SiOz: 0.3 wt% ( • ) , 0.5 wt% Si02 (0) and 0.8 wt%(*). Reprinted with permission from Ref 31, Copyright 2004 American Chemical Society.

above, this constitutes the strongest experimental evidence for the rigid bindingof R6G to the silica surface supported by theoretical considerations^^ and corroborated by several independent studies of rhodamine adsorption onto silica surfaces by second harmonic generation in combination with another spectroscopic methods. ' ' ' These studies show that the R6G dye, spin-coated on the surface of fused silica or embedded in layered silicate films, penetrates into the interlayer spaces of silicates and orients with the molecular plane of the xanthene ring almost parallel to the silica surface.

13.4. MONITORING SILICA SURFACE MODIFICATION

13.4.1 Background

High residual anisotropics in the TRFA decays of R6G in aqueous silica sols reduce the fraction of anisotropy available to reconvolute ^i and ^ from the decay, and thus obscure information on the ps-ns dynamics. However, the high residual anisotropics are an advantage when TRFA is used for monitoring the modification of silica surfaces.

One of the commonly used practices to modify particle surface properties is to have molecules with specific properties physically adsorbed or covalently attached to the NP surface. The choice of target molecules for the adsorption studies is determined by their relevance in suppressing imr ! J rij to the silica surface (APTES, tertiary amines), ^ promoting native-like protein function (GLTES, MLTES),^^' ^ stabilizing colloidal dispersions (PEO)^^ or modeling biological systems (peptides).^^ It is therefore important that we have the experimental ability to quantitatively detect molecular adsorption and


discuss it in the context of colloidal properties based on particle surface conditions.

Some of the intrinsic properties of colloidal silica make it difficult to employ as a sorbent in adsorption studies. In water, the silica surface continually evolves due to the dissolution of silicate species and reduction of the available surface area. The silica dissolution affects the surface charge density calculated from potentiometric titrations, especially when the volume fraction of the silica NPs is low. The silica NPs may flocculate as a result of a decreased surface potential (by increasing pH), increased concentration of electrolyte (counterions) or due to a reduced degree of hydration. In any silica suspension formed via nucleation and growth processes one never has a truly monodisperse system. The polydispersity of silica sols has been demonstrated by SAXS^^ and dynamic light scattering,^^ which measure the hydrodynamic radius of solvated particles. A difference between the average radius of solvated particles and the radius of vacuum-dried particles measured by TEM points toward the existence of superficial gel layers or "surface steric barriers" on the NP surfaces composed of poly silicates and bound cations. All these features of colloidal silica cause it to deviate from the ideal model of a solid surface, and complicate the interpretation of adsorption phenomena in colloidal silica systems.

13.4.2 Ludox

Ludox AM-30 dispersion (30 wt%, pH = 8.9) consists of stable silica NPs (mean radius of 6 nm) whose negative charge is mostly independent of pH due to substitution of a small portion of tetravalent silicium ions by trivalent aluminum ions (Figure 13.8). ' " ' ^ Ludox NPs are normally distributed around an average radius R = 6 nm (TEM).^ These properties makes Ludox a good model of colloidal silica as a class, given that the silica surface charge does not depend on the synthesis procedure used to obtain the NPs.^ However, some challenges still remain. Despite the fact that vacuum-dried Ludox NPs appear as perfect spheres under the electron microscope, they are highly hydrated in water and, as discussed above, exhibit gel-like layers on their surfaces. ^ Although such layers contribute to the stabilization of Ludox against flocculation, on the other hand, they lead to an abnormally high concentration of surface-exposed silanol groups.^ As a result, the actual number of the adsorption sites on Ludox cannot be accurately estimated from the particle dimensions. Even so, Ludox is an excellent model system to allow investigation of silica surface modification and biomolecule:silica interactions.

13.4.3 Monitoring Adsorption by TRFA

13.4.3.1 Conditions for Monitoring Adsorption

One important requirement for the use of TRFA in adsorption studies is the absence of interactions between the target molecules and the R6G dye. In order


Nrtf OH Na*0"

V ^ ! k ^^K ^^^ / • •

Figure 13.8. Surface of Ludox particle. Thick line denotes the particle surface. Reprinted with permission from Ref 51, Copyright 2005 American Chemical Society.

to prove this, the emission parameters of R6G (emission spectra, quantum yield, fluorescence lifetime, TRPA decay) must be measured in aqueous solutions containing different amounts of the dissolved modifier (in the absence of Ludox). A characteristic lifetime of r~ 4 ns, a correlation time of ^< 0.2 ns and the absence of residual anisotropy in the decay are good indicators of the absence of interactions between the modifier and R6G. The adsorption of non-fluorescent targets onto Ludox NPs is measured by competitive adsorption of R6G, which is added to the modified Ludox sol immediately after the modification has been done. The TRFA decay of R6G in the modified Ludox is modeled as the mixture of rigid spheres (Eq. 15). The coating of the Ludox surface by target molecules leads to a decreased fraction of R6G bound to large Ludox NPs (g) and an increased fraction of free, unbound dye (/i), as compared to the TRFA parameters of R6G in the plain Ludox suspension. This situation is shown schematically in Figure 13.9, where addition of a surface modifying agent leads to blockage of the silica surface, and hence a higher proportion of free R6G in solution.

It is worth noting that the use of R6G as the adsorption marker is conceptually different from what is commonly done in another analytical methods. ' ' ' ^ Usually, the adsorption marker is added in sufficient amounts to create surface saturation and then a target is added. The adsorbed target displaces the adsorption marker from the particle surface and thus the adsorption is monitored as a decrease of the signal due to the displacement of the dye to solution. In contrast, TRFA uses extremely low concentrations of the fluorescent probe, below one R6G molecule per silica particle. Even so, statistically, the adsorption of surface modifiers will lead to a shift in the R6G binding equilibrium in a manner that depends on both the affinity of the modifier for the surface and the concentration of modifier in solution.^^' ' ^

One issue with using Ludox for adsorption studies is that the adsorption of target molecules onto Ludox NPs usually decreases the repulsive energy barrier


large Ludox NP small Ludox NP R6G PEO

Figure 13.9. Schematic representation of competitive adsorption of R6G and PEO onto Ludox NPs. As more PEO chains bind to Ludox, less silica surface is available for R6G adsorption, leading to an increase in free dye. This alters the relative contribution of each decay parameter ifufi and g) depending on the extent of surface modification.

existing between plain NPs. This decrease may be due to the reduced particle surface charge, ^ a lower degree of hydration^ or both. If the repulsive potential between modified Ludox NPs is low, but their surface is not completely covered by target molecules, the NPs collide, form Si-O-Si bonds and flocculate. Since the flocculation affects the optical transparency of the sample, the concentration of a modifier in the Ludox sample for the TRFA measurement must remain below the critical flocculation concentration.

Table 13.4. TRFA decay parameters for R6G in PEO-modified Ludox suspensions.^^

Target (l>hns (f>2,ris fi 3 wt% Si02 0.55 kD PEO, 1 wt% 0.55 kD PEO, 30 wt% lOkDPEO, 1 wt% lOkDPEO, 30wt% 100kDPEO,0.1 wt% lOOkDPEO, 1 wt%

0.16 ±0.01 0.16 ±0.03 0.55 ± 0.02 0.20 ± 0.06 0.37 ± 0.04 0.12 ±0.03 0.24 ±0.16

5.6 ±0.6 2.7 ±0.3 2.8 ±0.4 4.1 ±0.3 2.5 ± 0.3 4.4 ± 0.6 3.5 ±0.6

0.01 0.09 0.47 0.02 0.19 0.04 0.05

0.79 0.52 0.18 0.68 0.38 0.78 0.54


13.4.3.2 Adsorption of Polymers

The first system we examined by TRFA was the interaction between PEO and silica. PEO can undergo hydrogen bonding and hydrophobic interactions with siHca, which are similar to those found in proteins and other biological materials. Additionally, PEO is known to stabilize colloidal dispersions,^^ act as an antifouling agent on membranes,^^ resist the adsorption of proteins and increase the intracellular uptake of modified NPs. ^ Furthermore, sol-gel derived silica doped with PEO and other polymers has been shown to provide enhanced activity for entrapped enzymes '*"^^ and to alter the dynamics of entrapped biomolecules."^ ' ^ These aspects make the incorporation of PEO chains into silica-based nanocomposites the basis of a general methodology for the entrapment of bioactive molecules into ceramics and glasses,^^ and make such materials useful for biomedical and environmental immunochromato-graphy and immunosensing.^

Table 13.4 shows the decays of R6G in unmodified and PEO modified Ludox samples. The first point to note is that under appropriate conditions, the fraction of free dye present in unmodified silica systems, given by/i, is on the order of 1%. Thus, 99% of the R6G is bound to the silica surface. A second key point is that the decays of R6G in the PEO-modified Ludox exhibit a concentration-dependent increase in /i and a corresponding decrease in the g value. However, a high PEO concentration (30 wt%) is required cause substantial changes in the/i and g values.

An important aspect of TRFA measurements is the ability to assess fractional surface coverage using the TRFA parameters. Considering the 10 kDa PEO samples, the g value gradually decreased from 0.8 (plain Ludox) to 0.38 (30 wt % PEO), with the fractional fluorescence of the nanosecond component increasing from 20% to 43% and that of the picosecond component increasing from ~ 1 % to 19%. Since the emission properties of R6G do not change upon adsorption, the total fluorescence from the nanosecond and residual anisotropy components (1-fi) corresponds to the fraction of bound dye. Thus, the fraction of rigidly bound dye is reduced from -99% to 81% upon adsorption of the highest concentration of 10 kDa PEO to the Ludox suspension, indicative of a surface coverage of 18%.

Using this approach, the extent of blockage of dye adsorption can be calculated as a function of the molecular weight and concentration of added polymer. As shown in Figure 13.10, more significant changes in the fraction of free dye were obtained after introduction of low molecular weight PEO to the Ludox samples. More importantly, the higher fraction of free, silica-unbound R6G as a result of the silica coverage by the modifiers, indicates a higher degree of Ludox modification by these agents. The majority of the changes occurred at low polymer concentrations in the case of 0.55 kDa PEO. These data clearly show that the lower molecular weight polymer is able to form close packed layers that block the adsorption of the dye to the silica surface, and thus suggest that the smaller polymer chains more thoroughly coat the silica NP surface. This is not unexpected, as the larger polymer would lose


10 15 20

Wt% of Polymer

35

Figure 13.10. Fractional contributions from free probe to the total anisotropy decay of R6G in polymer-doped Ludox sols (3 wt % SiOz) determined from TRFA measurements. ( • ) 10 kDa PEG modified Ludox; ( • ) 0.55 kD PEO modified Ludox. Reprinted with permission from Ref 36, Copyright 2004 American Chemical Society.

more degrees of freedom relative to a smaller polymer if all segments adsorbed to the silica surface.

13.4.3.3 Adsorption of Covalent Modifiers

One area where anisotropy measurements are particularly useful is for monitoring the modification of silica by organosilanes species. Adsorption of neutral sugarsilanes, cationic aminopropylsilane species and cationic buffer diallyldimethylammonium chloride (DADMAC) to the surface of Ludox NPs was examined by steady-state anisotropy and TRFA. ^ However, relatively low levels of modifier were used to avoid flocculation of the Ludox. Maximum concentrations of the various modifiers were 2.5 wt% for GLTES, 2.0 wt% for MLTES, 0.3 wt% for DADMAC and 0.1 wt% for APTES. Although flocculation was an issue, it was possible to examine samples containing up to 5

Table 13.5. TRFA decay parameters for R6G in organosilane-modified Ludox suspensions.^^

Target ^7, ns (l>2,ns fi Neutral:

Cationic:

GLTES, 2.5 wt% MLTES, 2 wt% APTES, 0.005 wt% APTES, 0.01 wt% DADMAC, 0.3 wt%

0.15 + 0.02 0.16 ±0.01 0.20 ±0.01 0.17 ±0.01 0.19 ±0.01

2.3 ±0.5 2.1 ±0.2 2.2 ± 0.9 1.7 ±0.9 2.0 ± 0.4

0.22 0.16 0.73 0.80 0.55

0.49 0.54 0.18 0.09 0.14


wt% GLTES or MLTES, since flocculation of the aqueous Ludox sol (3 wt% Si02) required 15 min, providing enough time to carry out the measurement of R6G steady-state anisotropy.^^ Steady-state anisotropy values dropped from 0.336 (unmodified Ludox) to 0.117 (GLTES-Ludox) and 0.09 (MLTES-Ludox), and further increases in the concentration of silane agents has little effect on R6G anisotropy, with minimum values approaching 0.087 and 0.078, respectively. Even lower final steady-state anisotropy values were obtained from cationic modifiers, with values of 0.061 being obtained for 0.3 wt% DADMAC and 0.020 being obtained for 0.1 wt% APTES. If we attribute this decrease to a higher fraction of free, silica-unbound R6G as a result of the silica coverage by the modifiers, the lower anisotropy values should correspond to a higher degree of Ludox modification by these agents.

Table 13.5 shows the TRFA parameters obtained for the R6G-Ludox system at the highest level of each modifier. As was the case with PEO, the presence of the modifier led to marked decreases in the g value and correspondingly higher/i values. Such results are consistent with the covalent binding of silane coupling agents to the silica surface, although hydrogen bonding may contribute to the binding. Consistent with the steady-state anisotropy results, the observed increases in/i and decreases in g followed the trend of APTES > DADMAC > GLTES > MLTES, indicative of a higher degree of adsorption for the cationic species relative to the neutral compounds. Hence, the ionic binding of small cationic molecules seems to provide a more efficient silica surface coverage.

In the case of the covalent modifiers, the fractional surface coverage was assessed using both steady-state and time-resolved anisotropy data. For steady-state anisotropy, the fractional coverage of the surface can be estimated by comparing the minimum and maximum (limiting) steady-state anisotropy values as 1 - {r^JrQ), where r^Jr^ is the fraction of anisotropy remaining due to bound probe after maximum coverage of the surface with the modifier, and thus 1 -( minM)) is the fraction of sites that are inaccessible to dye, and thus assumed to be bound to the silane coupling agent. Based on this simple equation, the coverage values are thus estimated to be on the order of 45% for the sugarsilane coupling agents at the flocculation limits of 2.5 wt% (GLTES) or 2 wt% (MLTES), 82% for DADMAC (0.3 wt%) and 94% for APTES (0.1 wt%).

Using the TRFA data and assessing surface coverage based no changes in (l-/i), significantly different surface coverage values are obtained: 16% for MLTES, 22% for GLTES, 55% for DADMAC and 80% for APTES. In all cases, the estimated surface coverage from steady-state anisotropy measurements of silica modified surfaces is higher than those obtained by time-resolved anisotropy. The difference in these values lies in the presence of the nanosecond component, which corresponds to bound dye that has relatively rapid rotational motion. In general, the fractional components for the nanosecond component and residual anisotropy are consistent with a redistribution of dye between large and small particles. This contributes to a decrease in R6G steady-state anisotropy, which can be mistakenly interpreted as a higher fraction of free dye, and thus a higher surface coverage of silica, when


only average anisotropy values are measured. This clearly demonstrates that steady-state anisotropy values overestimate the amount of "free" probe, and hence the extent of surface modification. Using TRFA, it is possible to distinguish between free dye and dye that is bound to rapidly rotating particles, providing a more accurate measurement of the true amount of adsorption. This highlights the importance of using time-resolved anisotropy decays to accurately assess the distribution of dye in Ludox systems.

13.4.3.4 Adsorption of Tripeptides

More recent studies have extended our previous work with the R6G:Ludox system to the area of peptide:silica interactions. Four peptides were chosen for this study: the cationic tripeptides Lys-Trp-Lys (KWK) and N-acetyl-Lys-Trp-Lys (Ac-KWK) and the anionic peptides Glu-Trp-Glu (EWE) and N-acetyl-Glu-Trp-Glu (Ac-EWE). The number of positively charged ammonium groups in these tripeptides vary from 0 (Ac-EWE) to 3 (KWK) and in the case of EWE and Ac-EWE there are also carboxylate groups of the Glu side chains, which should be repelled from the anionic surface of Ludox. The relatively small size of the tripeptides provides a higher flocculation limit than larger cationic peptides such as polylysine,^^ and thus is likely to lead to greater coverage of the silica surface, while the tryptophan residue provides a spectroscopic handle to allow direct observation of peptide adsorption via Trp fluorescence anisotropy.

Tripeptide adsorption was monitored both directly, via Trp steady-state anisotropy, and indirectly, using both steady-state and time-resolved fluorescence anisotropy of R6G, by monitoring the competition of the peptide with R6G for binding to the silica surface. In the case of Trp anisotropy, increases in the silica:peptide molar ratio generally led to increases in anisotropy, with the extend of the anisotropy increase correlating with the number of cationic groups present in the peptide. However, the free rotation of the indole ring relative to the peptide backbone led to relatively low limiting anisotropy values, and the Trp anisotropy appeared to depend on both the degree of adsorption and the binding geometry of the peptide, making it difficult to use Trp anisotropy for quantitative assessment of the extent of peptide binding.

On the other hand, the competitive binding assay using TRFA of the R6G/Ludox system provided a usefiil method to probe peptide:silica interactions, with adsorption depending on the number of cationic sites present in the peptide. Referring to Table 13.6, it is apparent that the differences in adsorption of the four peptides fall into two categories. The anionic peptides show essentially no adsorption, based on the minimal changes in both/i and g. The cationic peptides, on the other hand, show both increased/i and decreased g values relative to unmodified Ludox, with the tricationic KWK showing larger variations in these parameters than was observed for the dicationic Ac-KWK at similar peptide concentrations. More importantly, it was possible to use the TRFA decay parameters to assess the extent of surface modification for each peptide. At the flocculation limit, both KWK (5 mM) and Ac-KWK (6 mM)


Table 13.6. TRFA decay parameters for R6G in peptide-modified Ludox suspensions.^^

Anionic:

Cationic:

Target plain Ludox 0.7 wt% Si02 Ac-EWE, 10 mM EWE, lOmM KWK, 0.6 mM KWK, 1 mM KWK, 2 mM KWK, 5 mM Ac-KWK, 5 mM

^h^S

0.1810.01 0.1710.01 0.23 ± 0.07 0.1710.04 0.1810.03 0.1610.02 0.1910.02 0.1910.03

(l>2.ns

3.910.5 3.910.3 1417 3.910.3 4.110.4 3.010.4 3.410.5 3.310.9

/ /

0.05 0.05 0.05 0.20 0.24 0.33 0.44 0.35

g

0.67 0.66 0.72 0.42 0.37 0.26 0.02 0.40

showed surface coverages of-40%, while EWE and Ac-EWE (both at 10 mM) showed essentially no binding to the surface.

Overall, the surface modification studies above show that monitoring the R6G anisotropy during the incremental addition of various modifiers or peptides to aqueous Ludox sols resulted in the ability to measure the degree of surface modification of silica particles. Potential advantages of the indirect TRFA method for probing interactions of compounds with surfaces include: (1) no need for labeling the competing compound, which may be difficult, or may introduce unwanted changes; (2) the ability to use the method for many different species (polymers, peptides, proteins) and (3) the ability to change the nature of the probe to assess different types of interactions (electrostatic, hydrogen bonding, hydrophobic), providing a more versatile method for studying interactions of compounds with surfaces. On the other hand, the indirect method has difficulties detecting the adsorption of weakly binding species, such as EWE, the adsorption of which might be observed by the direct measurement of Trp anisotropy. This suggests that weakly binding molecules cannot effectively compete with R6G for binding to the surface and thus a weakly binding fluorescent probe may be more suitable. It should also be noted that at this point TRFA cannot provide absolute quantitation of the surface coverage, but rather provides a relative assessment of the adsorption as a fianction of the charge and concentration of the modifier. Even so, the TRA measurement method still has significant utility, particularly for evaluating potential surface modification methods that may be used to reduce adsorption. Thus, TRFA measurements should be useful for characterization of biomaterials, antifouling surfaces and new chromatographic stationary phases.

13.5. CONCLUSIONS AND OUTLOOK

With the implementation of nanoparticle metrology by Geddes and Birch, the TRFA method has moved beyond the doors of the analytical laboratory and is becoming a promising technique for the analysis of problems related to colloid and interface science. The feasibility of this methodology has been


proven using aqueous DGS sols. The nanoparticle metrology approach has impacted the interpretation of TRPA decays allowing a more realistic understanding of the sol-gel interior in terms of microviscosity. The nanoparticle metrology also revealed the potential of TRFA decays to monitor the adsorption equilibrium between the soluble and silica-bound fractions of a fluorescent probe dispersed in a silica sol. This finding has been exploited in monitoring the surface modification of colloidal silica suspensions by water-soluble polymers, amines, silica precursors and small peptides.

While this review has highlighted the potential of TRPA for studies of silica growth and modification, much remains to be explored. First, methods must be developed to allow assessment of absolute surface coverage, in molecules per nm^ surface area, to allow comparison of TRFA modification data to that obtained by other methods. Second, TRFA methods must be developed to assess the potential of modified surfaces to resist or promote protein adsorption. Using our current method based on R6G equilibrium between bound and free states, there are only minor amounts of probe remaining on highly modified surfaces such as APTES coated Ludox. Thus, adsorption of secondary species such as proteins, particularly to APTES-coated areas, will not lead to redistribution of the dye. Covalent attachment of dyes to Ludox NPs, or examination of changes in the correlation times themselves rather than the fractions of the decay components, may allow for observation of the adsorption of large species, such as proteins. Another area where TRFA needs to be further explored in terms of the surface modification and protein adsorption is the sol-gel monolithic matrices. This can avoid issues with flocculation and thus provide a higher range of modifier concentrations to be studied. Studies are now underway to address these issues, and will be reported in due course.


The authors thank the Natural Sciences and Engineering Research Council of Canada, MDS-Sciex, the Canadian Foundation for Innovation, and the Ontario Innovation Trust for financial support of this work. John D. Brennan holds the Canada Research Chair in Bioanalytical Chemistry.

13.7. REFERENCES

1. W. Jin and J.D. Brennan, Properties and applications of proteins encapsulated within sol-gel derived materials, Anal. Chim. Acta 461, 1-36 (2002).

2. I. Gill, Bio-doped nanocomposite polymers: sol-gel bioencapsulates, Chem. Mater. 13, 3404-3421 (2001).

3. W. Tan, K. Wang, X. He, X. J. Zhao, T. Drake, L. Wang and R. P. Bagwe, Bionanotechnology based on silica nanoparticles, Med Res Rev. 24, 621-638 (2004).

4. A. A. Vertegel, R. W. Siegel and J. S. Dordick, Silica nanoparticle size influences the structure and enzymatic activity of adsorbed lysozyme, Langmuir 20, 6800-6807 (2004).

5. Du Pont. Ludox Colloidal Silica, Properties, Uses, Storage and Handling (Data sheet, 1987). 6. R. K. Her, The Colloidal Chemistry of Silica (ACS, Washington, 1994). 7. H. Eckert and M. Ward, Controlling the length scale through "soft" chemistry: from organic-

inorganic nanocomposites to functional materials, Chem. Mater. 13, 3059-3060 (2001).


8. M. A. Brook, Y. Chen, K.Guo, Z. Zhang, W. Jin, A. Deisingh, J. Cruz-Aguado and J. D. Brennan, Proteins entrapped in silica monoliths prepared from glyceroxysilanes, J. Sol-Gel ScL Technol. 31, 343-348 (2004).

9. M. A. Brook, Y. Chen, K. Guo, Z. Zhang and J.D. Brennan, Sugar-modified silanes: precursors for silica monoliths, J. Materials Chem. 14, 1469-1479 (2004).

10. J. Cruz-Aguado, Y. Chen, Z. Zhang, N. Elowe, M.A. Brook and J.D. Brennan, Ultrasensitive ATP detection using firefly luciferase entrapped in sugar-modified sol-gel derived silica, J. Amer. Chem. Soc.

126, 6878-6879 (2004). 11. J. Cruz-Aguado, Y. Chen, M.A. Brook and J.D. Brennan, Activity and inhibition of Src protein

tyrosine kinase entrapped within sugar-modified sol-gel derived silica, Anal. Chem. 76, 4182-4188(2004).

12. C. D. Geddes, J. Karolin and D. J. S. Birch, Sol-gel nanometrology: gated sampling can reveal initial sol formation kinetics, J. Fluoresc. 12, 113-117 (2002).

13. D. Pontoni, T. Narayanan and A. R. Rennie, Time-resolved SAXS study of nucleation and growth of silica colloids, Langmuir 18, 56 - 59 (2002).

14. G. Beaucage, H. K. Kammler, R. Mueller, R. Strobel, N. Agashe, S. E. Pratsinis and T. Narayanan, Probing the dynamics of nanoparticle growth in a flame using synchrotron radiation. Nature Materials^, 370 - 373 (2004).

15. P. Agreen, M. Linden, J. B. Rosenheim, R. Schwarzenbacher, M. Kriechbaum, H. Amenitsch, P.Laggner, J. Blanchard and F. Schuth, Kinetics of cosurfactant-surfactant-silicate phase behavior 1. Short-chain alcohols, JPhys Chem B 103, 5943-5948 (1999).

16. F. Ne, F. Testard, T. Zemb and J.-M Petit, Fast kinetics study of mesoporous material growth by small-angle X-ray scattering, ESRF News Lett. 33, 23-25 (1999).

17. S. Dankers and A. Leipertz, Determination of primary particle size distributions from time-resolved laser-induced incandescence measurements. Applied Optics 43, 3726-3731 (2004).

18. J. Frasch, B. Lebeau, M. Soulard, J. Patarin and R. Zana, In situ investigations on cetyltrim-ethylammonium surfactant/silicate systems, precursors of organized mesoporous MCM-41-type siliceous materials, Langmuir 16, 9049 - 9057 (2000).

19. J. Trebosc, J. W. Wiench, S. Huh, V. S.-Y. Lin and M. Pruski, Solid-state NMR study of MCM-41-type mesoporous silica nanoparticles, J. Am. Chem. Soc. 127, 3057-3068 (2005).

20. J. Zhang, Z. Luz and D. Goldfarb, EPR studies of the formation mechanism of the mesoporous materials MCM-41 and UCU-5Q,J Phys Chem, B 101, 7087-7094 (1997).

21. A. Galameau, F. Di Renzo, F. Fajula, L. Mollo, B. Fubini and M. F. Ottaviani, Kinetics of formation of micelle-templated silica mesophases monitored by electron paramagnetic resonance, J Colloid Interface Sci 201, 105-117 (1998).

22. M. F. Ottaviani, A. Galamneau, D. Desplantier-Giscard, F. Di Renzo and F. Fajula, EPR investigations on the formation of micelle-templated silica, Microporous Mesoporous Mater, Volume: 1(8), 44-45 (2001).

23. C. L. Landry, S. H. Tolbert, K. W. Gallis, A. Monnier, G. D. Stucky, P. Norby and J. C. Hanson, Phase transformation in mesostructured silica/surfactant composites. Mechanisms for change and applications to materials synthesis, Chem Mater 13, 1600-1608 (2001).

24. L. Liu, L. Xu, Zh. Hou, Zh. Xu, J. Chen, W. Wang and F. Li, In situ SHG investigation on the gelation process of organic doped silica film, Phys. Lett. A 262, 206-211 (1999).

25. L. Wenga, X. Zhoua, X. Zhangc, J. Xua, and L. Zhang, In situ monitoring gelation process of N,N-dimethylacrylamide by refractive index technique, Polymer 43, 6761-6765 (2002).

26. L. D'Souza, A. Suchopar and R. M. Richards, In situ approaches to establish colloidal growth kinetics, J. Colloid Interface Sci. 219, 458-463 (2004).

27. C. D. Geddes, K. Apperson and D. J. S. Birch, New fluorescent quinolinium dyes —^Applications in nanometre particle sizing, Dyes Pigments 44, 69-74 (2000).

28. C. D. Geddes and D. J. S. Birch, Nanometre resolution of silica hydrogel formation using time-resolved fluorescence anisotropy, /. Non-Cryst. Solids 270,191-204 (2000).

29. C. D. Geddes, J. Karolin, D. J. S. Birch, 1- and 2-photon fluorescence anisotropy decay in silicon alkoxide sol-gels: interpretation in terms of self-assembled nanoparticles, J. Phys. Chem. B. 106, 3835-3841 (2002).

30. C D . Geddes, 1 and 2-photon fluorescence anisotropy decay to probe the kinetic and structural evolution of sol-gel glasses: a summary, J. Fluoresc. 12, 343-367 (2002).


31. D. Tleugabulova, A. M. Duft, Z. Zhang, Y. Chen. M. A. Brook and J. D. Brennan, Evaluating formation and growth mechanisms of silica particles using fluorescence anisotropy decay analysis, Langmuir 20, 5924-5932 (2004).

32. D. Tleugabulova, Z. Zhang and J. D. Brennan, Evolution of sodium silicate sols through the sol-to-gel transition assessed by the fluorescence-based nanoparticle metrology approach, J. Phys. Chem. 5 107, 10127-10133 (2003).

33. S. Mukhopadhyay, Ira, G. Krishnamoorthy and U. Maitra, Dynamics of bound dyes in a nonpolymeric aqueous gel derived from a tripodal bile salt, J. Phys. Chem. B 107, 2189-2192 (2003).

34. J. W. Gilliland, K. Yokoyama and W. T. Yip, Comparative study of guest charge-charge interactions within silica sol-gel, J. Phys. Chem. B 109, 4816 -4823 (2005).

35. X. Sui, J. A. Cruz-Aguado, Y. Chen, Z. Zhang, M. A. Brook and J. D. Brennan, Properties of human serum albumin entrapped in sol-gel-derived silica bearing covalently tethered sugars, Chem. Mater. 17, 1174-1182 (2005).

36. D. Tleugabulova, A. M. Duft, M. A. Brook and J. D. Brennan, Monitoring solute interactions with poly(ethylene oxide)-modified colloidal silica nanoparticles via fluorescence anisotropy decay, Langmuir 20, 101-108 (2004).

37. D. Tleugabulova, Z. Zhang, Y. Chen, M. A. Brook and J. D. Brennan, Fluorescence anisotropy in studies of solute interactions with covalently modified colloidal silica nanoparticles, Langmuir 20, 848-854 (2004).

38. J. Sui, D. Tleugabulova and J. D. Brennan, Direct and indirect monitoring of peptide-silica interactions using time-resolved fluorescence anisotropy, Langmuir, 21, 4996-5001 (2005).

39. R. Wang, U. Narang, P. N. Prasad and F. V. Bright, Affinity of antifluorescein antibodies encapsulated within a transparent sol-gel glass. Anal. Chem .65, 2671-2675 (1993).

40. K. Flora and J. D. Brennan, Effect of matrix aging on the behavior of human serum albumin entrapped in a tetraethyl orthosilicate-derived glass, Chem. Mater. 13, 4170-4179 (2001).

41. D. S. Gottfried, A. Kagan, B. M. Hoffman and J. M. Friedman, Impeded rotation of a protein in a sol-gel matrix J. Phys. Chem. B 103, 2803-2807 (1999).

42. J. D. Jordan, R. A. Dunbar and F. V. Bright, Dynamics of acrylodan-labeled bovine and human serum albumin entrapped in a sol-gel-derived biogel, AnalChem. 67, 2436-2443 (1995).

43. G. A. Baker, J. D. Jordan and F. V. Bright, Effects of poly(ethylene glycol) doping on the behavior of pyrene, rhodamine 6G and acrylodan-labeled bovine serum albumin sequestered within tetramethylorthosilane-derived sol-gel-processed composites, J. Sol-Gel Sci. Technol. 11, 43-54(1998).

44. M. A. Doody, G. A. Baker, S. Pandey and F. V. Bright, Affinity and mobility of polyclonal anti-dansyl antibodies sequestered within sol-gel-derived biogels, Chem. Mater. 12, 1142-1147 (2000).

45. I. Khan, C. F. Shannon, D. Dantsker, A. J. Friedman, J. P-G.-de Apodaca and J. M. Friedman, Sol-gel trapping of functional intermediates of hemoglobin: geminate and bimolecular recombination studies, Biochem. 39, 16099-16109 (2000).

46. J. R. Lakowicz, Principles of Fluorescence spectroscopy (Kliiwer Academic Plenum Publishers, New York, 1999).

47. D. V. O'Connor and D. Phillips, Time-Correlated Single Photon Counting (Academic Press, London, 1984).

48. A. Perico and M. Guenza, Viscoelastic relaxation of segment orientation in dilute polymer solutions. II. Stiffness dependence of fluorescence depolarization, J. Chem. Phys. 84, 510-516 (1986).

49. J. M. Beechem, , J. R. Knutson and L. Brand, Global analysis of multiple dye fluorescence anisotropy experiments on proteins, Biochem. Soc. Trans. 14, 832-835 (1986).

50. B. Imhof, Decay Analysis Software Support Manual, (IBH Consultants Ltd, Glasgow, 1989). 51. D. Tleugabulova, J. Sui, P. W. Ayers and J. D. Brennan, Evidence for rigid binding of

rhodamine 6G to silica surfaces in aqueous solution based on fluorescence anisotropy decay analysis, J. Phys. Chem. B. 109,7850-7858 (2005).

52. B. Dunn and J. I. Zink, Probes of pore environment and molecule-matrix interactions in sol-gel materials,/ Chem. Mater. 9, 2280-2291 (1997).

53. T. Keeling-Tucker and J. D. Brennan, Fluorescent probes as reporters on the local structure and dynamics in sol-gel-derived nanocomposite materials, Chem. Mater 13, 3331 -3350 (2001).

54. F. Perrin, La fluorescence des solutions, Ann. Phys. Paris 12, 169-175 (1929). 55. P. A. Egelstaff, An Introduction to the Liquid State (Clarendon, Oxford, 1994).


56. U. Narang, R. Wang, P. N. Prasad and F. V. Bright, Effects of aging on the dynamics of rhodamine 6G in tetramethyl orthosiHcate-derived sol-gels, J. Phys. Chem. 98, 17-22 (1994).

57. K. E. Wheeler, N.S. Lees, R. J. Gurbiel, S. L. Hatch, J. M. Nocek and B. M. Hoffman, Electrostatic influence on rotational mobilities of sol-gel-encapsulated solutes by NMR and EPR spectroscopies, J. Am. Chem. Soc. 126, 13459 - 13463 (2004).

58. C. D. Geddes, J. Karolin and D. J. S. Birch, Fluorescence anisotropy in sol-gels: microviscosities or growing silica nanoparticles offering a new approach to sol-gel structure elucidation,/. Fluoresc. 12, 135-137 (2002).

59. D. Tleugabulova, W. Czardybon and J. D. Brennan, Time-resolved fluorescence anisotropy in assessing side-chain and segmental motions in polyamines entrapped in sol-gel derived silica, J. Phys. Chem. B. 108, 10692-10699 (2004).

60. H. Shirota and H. Segawa, Time-resolved fluorescence study on liquid oligo(ethylene oxide)s: coumarin 153 in poly(ethylene glycol)s and crown ethers, J. Phys. Chem. A 107, 3719 - 3727 (2003).

61. S. Sen, D. Sukul, P. Dutta and K. Bhattacharya, Solvation dynamics in aqueous polymer solution and in polymer-surfactant aggregate, J. Phys. Chem. B 106, 3763 -3769 (2002).

62. S. Zorrilla, G. Rivas and M. P. Lillo, Fluorescence anisotropy as a probe to study tracer proteins in crowded solutions, J. MolRecognit. 17, 408-416 (2004).

63. M. Polverari and T. G. M. van de Ven, Dilute aqueous poly(ethylene oxide) solutions: clusters and single molecules in thermodynamic equilibrium, J. Phys. Chem. 100, 13687 - 13695 (1996).

64. I. Isenberg, On the theory of fluorescence decay experiments. II. Statistics, J. Chem. Phys. 59, 5708-5713 (1973).

65. K. Kinosita Jr., A. Ikegami and S. Kawato, On the wobbling-in-cone analysis of fluorescence anisotropy decay, Biophys. J. 37, 461-464 (1982).

66. P. Damberg, J. Jarvet, P. Allard, U. Mets, R. Rigler and A. Graslund, 13C-1H NMR relaxation and fluorescence anisotropy decay study of tyrosine dynamics in motilin, Biophys. J. 83, 2812-2825 (2002).

67. G. R. Ziegler and E. A. Foegeding, The gelation of proteins. Adv. Food Nutr. Res. 34, 203-298 (1990).

68. H. Boukari, J. S. Lin and M. T. Harris, Small-angle X-ray scattering study of the formation of colloidal silica particles from alkoxides: primary particles or not? J. Colloid Interface Sci. 194, 311-318(1997).

69. C. H. Byers, M. T. Harris and D. F. Williams, Controlled microcrystalline growth studies by dynamic laser-light-scattering methods, Ind. Eng. Chem. Res. 26, 1916-1923 (1987).

70. T. Matsoukas, E. Gulari, Monomer-addition growth with a slow initiation step: A growth model for silica particles from alkoxides, J. Colloid Interface Sci. 132, 13-21 (1989).

71. V. K. LaMer, Kinetics in phase transitions, Ind. Eng. Chem. 44, 1270-1277 (1952). 72. A. van Blaaderen, J. van Geest and A. Vrij, Monodisperse colloidal silica spheres from

tetraalkoxy si lanes: particle formation and growth mechanism, J. Colloid Interface Sci 154, 481-501 (1992).

73. G. H.Bogush, C. F. Zukoski, Studies of the kinetics of the precipitation of uniform silica particles through the hydrolysis and condensation of silicon alkoxides, J. Colloid Interface Sci. 142,1-18(1991).

74. G. H.Bogush, C. F. Zukoski, Uniform silica particle precipitation: An aggregative growth model, J. Colloid Interface Sci. 142, 19-34(1991).

75. K. Lee, J. L. Look, M. T. Harris and A. V. McCormick, Assessing extreme models of the Stober synthesis using transients under a range of initial composition, J. Colloid Interface Sci. 194, 78-88(1997).

76. R. B. Bhatia, C. J. Brinker, A. K. Gupta and A. K. Singh, Aqueous sol-gel process for protein encapsulation, Chem. Mater. 12, 2434 - 2441 (2000).

77. K. Ishibashi, O. Sato, R. Baba, D. A. Tryk, K. Hashimoto and A. Fujishima, Characterization of the chromophore orientation of rhodamine B amphiphiles in Langmuir Blodgett monolayers. J. Colloid Interface Sci 233, 361-363 (2001).

78. T. Kikteva, D. Star, Z. Zhao, T. L. Baisley and G. W. Leach, Molecular orientation, aggregation, and order in rhodamine films at the fused silica/air interface, J. Phys. Chem. B. 103, 1124-1133 (1999).

79. R. Sasai, T. Fujita, N. lyi, H. Itoh and K. Takagi, Aggregated structures of rhodamine 6G intercalated in a fluor-taeniolite thin film, Langmuir 18, 6578-6583 (2002).


80. J. Bujdak, N. lyi, Y. Kaneko, A. Czimerova and R. Sasai, Molecular arrangement of rhodamine 6G cations in the films of layered silicates: the effect of the layer charge, Phys. Chem. Chem. Phys. 5, 4680-4685 (2003).

81. R. J. Hodgson, Y. Chen, Z. Zhang, D. Tleugabulova, H. Long, X. Zhao, M. Organ, M. A. Brook and J. D. Brennan, Protein-doped monolithic silica columns for capillary liquid chromatography prepared by the sol-gel method: applications to frontal affinity chromatography. Anal Chem.. 76, 2780-2790 (2004).

82. S. R. Raghavan, H. J. Walls and S. A. Khan, Rheology of silica dispersions in organic liquids: new evidence for solvation forces dictated by hydrogen bonding, Langmuir 16, 7920 - 7930 (2000).

83. J. N. Cha, G. D. Stucky, D. E. Morse, T. J. Deming, Biomimetic synthesis of ordered silica structures mediated by block copolypeptides, Nature 403, 289-92 (2000).

84. P. van der Meeren, H. Saveyn, S. Bogale Kassa, W. Doyen, R. Leysen, Colloid-membrane interactions effects on flux decline during cross-flow ultrafiltration of colloidal silica on semi-ceramic membranes, Phys. Chem. Chem. Phys. 6, 1408-1412 (2004).

85. M. Kosmulski, Electrokinetic study of specific adsorption of cations on synthetic goethite. Colloid Surf. A 111, 113-118 (2003).

86. J. Laven and H. N. Stein, The electroviscous behavior of aqueous dispersions of amorphous silica (Ludox), J. Colloid Interface Sci. 238, 8-15 (2001).

87. H. Wang, E. C. Y. Yan, Y.Liu, K. B. Eisenthal, Energetics and population of molecules at microscopic liquid and solid surfaces, J. Phys. Chem. B. 102, 4446-4450 (1998).

88. H. Wang, E. Borguet, E. C. Y. Yan, D. Zhang, J. Gutow and K. B. Eisenthal, Molecules at liquid and solid surfaces, Langmuir 14, 1472-1477 (1998).

89. H. M. Eckenrode and H.-L. Dai, Nonlinear optical probe of biopolymer adsorption on colloidal particle surface: poly-L-lysine on polystyrene sulfate microspheres, Langmuir 20, 9202-9209 (2004).

90. M. A. Polizzi, R. M. Plocinik and G. J. Simpson, EUipsometric approach for the real-time detection of label-free protein adsorption by second harmonic generation, J. Am. Chem. Soc. 126,5001-5007(2004).

91. M. Y. Lin, H. M. Lindsay, D. A. Weitz, R. Klein, R. C. Ball and P. Meakin, Universal diffusion-limited colloid aggregation, J. Phys. Condens. Matter 2, 3093-3113 (1990).

92. J. Hester, P. Banerjee and A. Mayes, Preparation of protein-resistant surfaces on poly(vinylidene fluoride) membranes via surface segregation, Macromolecules 32, 1643-1650 (1999).

93. A. K. Gupta, A. S. Curtis, Surface modified superparamagnetic nanoparticles for drug delivery: interaction studies with human fibroblasts in culture, J. Mater. Sci. Mater. Med. 15, 493-496 (2004).

94. Q. Chen, G. L. Kenausis and A. Heller, Stability of oxidases immobilized in silica gels, J. Am. Chem. Soc.llQ, 4582-4585 (1998).

95. J. Heller, A. Heller, Loss of activity or gain in stability of oxidases upon their immobilization in hydrated silica: significance of the electrostatic interactions of surface arginine residues at the entrances of the reaction channels, J. Am. Chem. Soc. 120, 4586-4590 (1998).

96. T. Keeling-Tucker, M. Rakic, C. Spong, J. D. Brennan, Controlling the material properties and biological activity of lipase within sol-gel derived bioglasses via organosilane and polymer doping, Chem. Mater. 12, 3695-3704, (2000).

97. G. A. Baker, S. Pandey, E. P. Maziarz III and F. V. Bright, J. Sol-Gel Sci. Technol. 1999, 75, 37.

98. A. Bronshtein, N. Aharonson, D. Avnir, A. Tumiansky and M. Alstein, Sol-gel matrixes doped with atrazine antibodies: atrazine binding properties, Chem. Mater. 9, 2632-2639 (1997)..

99. T. Coradin, O. Durupty and J. Livage, Interactions of amino-containing peptides with sodium silicate and colloidal silica: a biomimetic approach of silicification, Langmuir, 18, 2331-2336 (2002).

DYNAMICS OF DNA AND PROTEIN-DNA COMPLEXES VIEWED THROUGH TIME-

DOMAIN FLUORESCENCE

Nabanita Nag\ T. Ramreddy^ Mamata Kombrabail\ P.M. Krishna Mohan^, Jacinta D'souza^, B.J. Rao^, Guy Duportail^, Yves Mely , and G. Krishnamoorthy

14.1. INTRODUCTION

In recent times, there has been increasing level of realization that activity of biomolecular systems arises from a combination of structure and dynamics '' . While this concept has gained substantial experimental support in proteins "" , the role of dynamics is comparatively less appreciated in DNA where the activity is assumed to be largely controlled by its sequence. This expectation might have arisen due to the fact that DNA is more rigid and structured than proteins and other bio-macromolecules. However, a variety of intramolecular as well as intermolecular interactions often influence the structure and dynamics of DNA. Double-stranded (ds) DNA is generally modeled as a string of rods (or disk) which follows Hookean twisting and bending much like a spring where the counter interacting torsional and flexural deformation play an important role in maintaining its rigidity^. The rigidity of DNA is often characterized by its persistence length which is -- 50 nm for ds-Watson-Crick B-form^ and ~ 1.0 nm when present as single stranded^. Furthermore, the structure of DNA double helix is not monotonous but instead exhibits marked sequence dependent variations^. Ds-DNA bends easily at one

1 2

Department of Chemical Sciences, Department of Biological Sciences, Tata Institute of Fundamental Research, Homi Bhabha Road, Mumbai 400 005, India. ^Laboratoire de Pharmacologic et Physicochimie des interactions cellulaires et moleculaires, UMR 7034 du CNRS, Faculte' de Pharmacie, Universite' Louis Pasteur de Strasbourg, 74 Route du Rhin, 67401 lUkirch, France.

311

312 N.NAG^r.^L.

plane when compared to the other thus giving rise to 'anisotropic flexibiHty^"^\ This in turn helps ds-DNA to form transient as well as stable complexes with proteins and other co-factors and drugs. Sequence-specific bends in DNA has been recognized as one of the mechanisms for specific binding of proteins^^. Additionally, sequence-specific dynamics along the backbone could be thought of as another interesting mechanism for specific recognition by proteins (see section 4.2 and 4.3).

The hierarchy of dynamics could be classified in timescale and in molecular structure. Fluorescence studies have shown that the interior of DNA has the diffusive and viscous dynamics characteristics of fluid, rather than the purely vibrational dynamics of crystals^^. These local motions in DNA could occur in sub-nanosecond timescales. This would include propeller twist and wobbling motion of base pairs involving relatively small segments of DNA. In contrast, movement of DNA backbone giving rise to twisting of double helix (torsion) and in tum long range bending and global tumbling of the DNA involve larger segments and would occur in the timescales of nanoseconds and beyond. Spontaneous and transient opening/ disruptions of double helix have been shown to be segmental rather than continuous^'^. Hydrogen exchange studies by NMR have revealed the occurrence of local fluctuations as well as melting and unstacking of the duplex* ' ^ . Such local and small perturbations could finally determine long range torsional and flexural rigidity and Brownian dynamics of DNA. Capturing high resolution snapshots of the various molecular motions under equilibrium conditions poses technical challenges.

A variety of physical techniques such as NMR, X-ray crystallography, electron microscopy, atomic force microscopy (AFM), electrophoretic mobility on gels and fluorescence spectroscopy are being used to study the structure and dynamics of DNA systems. By virtue of its selectivity, sensitivity and large temporal range, fluorescence spectroscopy has become an important and widely used technique for studying structure and dynamics of small as well as large biological macromolecules in solution^^"^ . The exquisiteness of this technique lies on the selectivity, where the small change in the environment around the fluorophore is reflected in the measurable change in the emission properties without interference from the rest of the system. The observation can be made with small amount of sample which gives it sensitivity. The most important feature of fluorescence spectroscopy is the characteristic time scale of the emission process over a time period ranging from picoseconds to nanoseconds which is comparable to the temporal range of biological macromolecular dynamics.

The dynamic nature of DNA structure, which is essential for its ftinction, gives its structural heterogeneity. The apparent level of heterogeneity depends upon the time window used for observation. Large time windows result in averaging of the structural parameters, whereas shorter windows produce an instantaneous snapshot of the structural variants populating a distribution. The time window set by fluorescence based methods is linked to the excited state life time of the fluorophore, which lies generally in the range of 10 ps to 10 ns. The rapid timescale of the emission process allows us to take snapshots of

DYNAMICS OF DNA AND PROTEIN-DNA COMPLEX. 313

various structural forms present due to various types of molecular movements including internal motion.

14.2. FLUORESCENCE PROBES FOR DNA DYNAMICS

The fluorescent quantum yield of native DNA which is due to its bases is too small (Of- 4x10'Y^'^^ and its fluorescent lifetimes too short (10-60 ps)^^. However, there have been several attempts to use the intrinsic fluorescence of DNA to gain information on its dynamics^^. Furthermore, due to its non-specific nature, its use is limited to study only overall dynamics of DNA such as internal motion of DNA bases "*.

Several extrinsic probes (Figure 14.1 A) have been developed over the years to make DNA fluorescently visible. They fall into two categories, namely specific and non-specific. Non-specific probes such as ethidium bromide^^, DAPI^^ propydium iodide^^, YOYO^ ' ^^ and PicoGreen^^ (PG) and their large range of analogs were originally developed as DNA stains to visualize DNA in fluorescence microscopic images. However, probes such as ethidium^^' ^ and YOYO-1^^' ^^ (see section 3) have found use in studying complex dynamic modes of DNA and as condensation indicators (see section 3.3 and 3.4). Most of these probes are nearly non-fluorescent in their free form and become highly fluorescent when bound to DNA (see section 3.3 and 3.4). Although many of these probes are either mono or bis intercalators, but DAPI^^ work as groove binders. Probe such as PicoGreen^^ has been used to assess the relative population of single-stranded regions and single-strand nicks in ds-DNA^ ' ^ . This capability arises presumably due to the modulation of fluorescence quantum yield by the various dynamic modes of DNA (see section 3.5).

Owing to its non-specific probing properties, the probes mentioned above are not useful in studying environmental change in a specific location of DNA. Hence site-specific fluorescent probes (Figure 14.IB) which are mainly nucleotide base analogues are most appropriate in studying the dynamics in a specific region of DNA. Selection of these nucleotide analogs is generally based on the criteria that the fluorescent analog mimics the parent nucleotide in its base-pair hydrogen bonding and other interaction potentials. 2-aminopurine (2-AP) which is being widely used as an analog of adenine forms hydrogen bonds with thymine very similar to those of adenine (Figure 14.2A). However, 2-AP:T base-pair itself is treated as a mismatched base-pair by the mismatch recognition protein, Mut-S^^' ^ indicating subtle differences in the base-pair structure and dynamics. Although the quantum yield of 2-AP gets substantially reduced (Of ~ 0.1) on incorporation into DNA, it is still sufficient for monitoring subtle changes in the structure and dynamics" " .

Very recently, 8-vinyl-deoxyadanine (8-VA) has been proposed as an improved substitute of 2-AP' . The quantum yield of this new adenine analog is significantly higher than that of 2-AP, when inserted in DNA. In addition, 8-VA

314 N, N AG ET.AL.

(i)

^ f ^

H^N—C

(ii)

Z = CI , I , CIO3 X = S, O RZ-R = aliphatic chains

(iii) R6 R5

CH3

^ , — ^ (CH2)3 N—CH3

V V CH2CH3

(iv)

r-; CH2)3—^N -(012)3—f;J -iCR,\

CH3 41

CH,

(V)

Figure 14.1A: Chemical structures of some of the non-specific fluorescence probes used for monitoring DNA dynamics, (i) Ethidium bromide, (ii) DAPI, (iii) PicoGreen (PG), (iv) Propidium iodide, (v) YOYO-1.

is able to adopt an anti conformation that preserves the Watson-Crick hydrogen bonding.

Another fluorescent analog pyrrolo-dC (PdC) was introduced few years back which can pair with Guanine'*' (Figure 14.2B). The quantum yield sensitivity of pyrrolo-dC is quite similar to that of 2-AP' ' ' ^^. All the three probes 2-AP, 8-VA and PdC show significant reduction in their quantum yield on base-pair formation and thus serve as indicators of local hybridization. However, PdC is likely to have greater potential since its excitation wavelength (347 nm) overlaps less with that of proteins. Thus PdC is better suited for studies on DNA-protein complexes. Furthermore, PdC has a very unique characteristic when it is mismatched in a ds-DNA. Its quantum yield in this condition is higher than that in the single-stranded species when the mismatched base is adenosine. This unusual behavior has been used to probe dynamics at the mismatched site" .


Several other fluorescent base analogs such as Etheno-A and Etheno-C (Figure 14.IB) developed earlier have limited use due their poor hybridizing nature'* . Other analogs of adenosine such as 6-MAP (4-amino-6-methyl-8-(2-deoxy-beta-d-riboftiranosyl)-7(8H)-pteridone)(Figure 14.IB) and derivatives of guanosine and other nucleotides have also been investigated^^' ^

One of the commonly used fluorescent labeling of DNA, especially at the ends of DNA, rely upon nucleoside bases derivatized with probes such as fluorescein, rhodamine, Cy3, Cy5 and Alexa derivatives with an aliphatic linker. The linkers are attached to C-5 or C-8 position of pyrimidine or purine bases respectively so that the linker arms are projected into the major groove of the DNA and thus interfere less with base-pairing interactions. These probes, when placed far from the ends, give useful information about the microenvironment inside the groove" ' ^ apart from being useful in foot printing studies^^' ^ involving DNA binding proteins. These fluorescent dyes along with the linker are also tethered to the 5' or 3' terminus of the DNA and extensively used in applications such as estimation of DNA bends^^, interaction with proteins from the change in its tumbling motion^^, formation of hairpin bends^^ and DNA hybridization^^' ' .

n NH

(i) (ii) fiii)

N'

HjN N

(iv)

N

Q (y) (vi)

Figure 14.IB: Chemical structures of some of the site-specific fluorescence probes used for monitoring DNA dynamics, (i) Etheno-A, (ii) Etheno-C, (iii) 6-MAP, (iv) 2-Aminopurine (2-AP), (v) 8-Vinyladenine (8-VA), (vi) Pyrrolo-dC (PdC).

316 N.NAGCT.^i.

H \ r ^ ' V/™'

A:T base-pair (X=H) ' 2AP:T base-pair 8-VA:T base-pair (X= -CH=CH2)

(A)

CH,

G : C base-pair

N — H O

G : PdC base-pair

N

(B)

Figure 14.2: Base-pairing schemes for the fluorescent base analogues. (A) adenine : thymine (A:T), 8-vinyl adenine : thymine (8-VA:T) and 2-aminopurine : thymine (2-AP:T) base-pairs; (B) guanine : cytocine (G:C) and guanine : pyrrolo-dC (G:PdC) base-pairs.

14. 3. PROBING DNA DYNAMICS WITH NON-SPECIFIC PROBES

14.3.1. DNA condensation

Biological interest in DNA is generally confined to its sequence. However, higher order structures and structural changes in DNA, which are sequence-independent, have profound control on several processes that involve DNA. Condensation of DNA, whereby DNA transforms from an open and


fluctuating architecture into a compact and relatively static form is encountered in a variety of situations both natural and artificial. Chromatinized DNA in the nucleus of mammalian cells^^ and DNA packaged into phage heads^^ exemplify the phenomenon of natural condensation. A major part of artificial condensation of DNA is related to the development of gene delivery (therapy) systems in which DNA is condensed and compacted by a variety of agents such as cationic lipids^^' ^ , polycations^^' ^ , cationic polymers, and cationic detergents '* before being taken up into cells by endocytic pathway^^' .

The process of DNA condensation has to achieve a fine balance between efficient compaction to economize the size while preserving the function such as gene expression in chromosomes. This makes studies on the structure and dynamics of condensed forms of DNA quite rewarding. Condensation of DNA has been studied by a variety of physical techniques mainly based on light and X-ray scattering^^' ^ , spectroscopy^" , imaging^^"^^ and viscometry . Atomic force microscopy ' and electron microscopy have revealed that condensed form of DNA has either toroidal or rod-like structures having dimensions in the range of 50-100 nm. Despite the use of a large variety of experimental techniques to DNA condensates, atomic level high-resolution structures are lacking. For example, even information such as relative localization of the condensing agent with respect to the DNA backbone is not known with certainty although it has been shown that condensation results in protection of DNA from nucleases^^. Similarly, questions such as how the therapeutic DNA which is condensed by various agents is integrated into the genome and does transcription of this DNA require decondensation, are largely unanswered. In this scenario it becomes quite relevant and interesting to look for detailed information on the comparative dynamics of DNA backbone and base in condensed and extended forms of DNA.

14.3.2. YOYO-1 as an indicator of DNA condensation

We had shown recently that YOYO-1, a DNA bis-intercalator, is a convenient and robust marker of DNA condensation^'*. The fluorescence quantum yield of YOYO-1 gets enhanced by several orders of magnitude on intercalation presumably due to suppression of internal rotation-induced non-radiative decay when bound to DNA " . This model for the remarkable level of enhancement in the fluorescence gets support from the observation that the quantum yield of YOYO-1 gets significantly enhanced in a viscous solvent such as glycerol. Bis-intercalation of analogs of YOYO-1 has been confirmed from NMR structures'^ When used at relatively high levels (YOYO-1: DNA phosphate, D:P > 1/50), the quantum yield of DNA-bound YOYO-1 shows a remarkable decrease following the condensation of DNA. Detailed spectroscopic investigations^^ showed that the formation of H-dimers of YOYO-1 (Figure 14.3A) as the cause of condensation-induced decrease in quantum yield. The major evidence for the formation of H-dimers comes from the observation of excitonic blue shift ' ' in the absorption spectrum of DNA-bound YOYO-1^^ Furthermore, the closeness of DNA-bound YOYO-1 in the presence of condensing agents was demonstrated by time-resolved

318 N. N AG ET.AL.

Intercalating dye ^ ; : H-aggtegates

^ ^,4^%v w - ^

\.<-CF'*>*- ^

Condensing agent

DNA Condensed DNA

Figure 14.3 A; Schematic diagram of DNA condensation induced by cationic polymers. The process of condensation is visualized from a decrease in the fluorescence intensity of YOYO-1 (marked as black diamonds) by the formation of H-dimers (marked in DNA condensates).

fluorescence anisotropy studies which reveal the presence of energy migration by homo-FRET^" ' ^ . The probability of formation of H-dimers is expected to increase with the increase in the probe to phosphate ratio (D: P) and hence condensation-induced fluorescence quenching is expected to show up only at high levels of probe/phosphate ratio as observed (Table 14.1). Thus it is clear that YOYO-1 fluorescence is a simple and robust marker of DNA condensation.

14.3.3. Structure and dynamics of condensed DNA

Despite the fact that condensation of DNA has been achieved by using a variety of agents and procedures, it is known that only a very few forms of condensed DNA are effective as transfection agents in cells^ ' ^ ' ^ . Some of the condensed forms of DNA interact with the extra cellular side of plasma membrane of cells and release the DNA which subsequently gets trapped on the cell surface making transfection ineffective^^. On the other hand, dimerizable detergents pack the DNA tightly and retard their release when the condensates are taken into the cells. This also reduces the transfection efficiency^^. Thus it is clear that information on the structure and dynamics of DNA condensates would be useful while designing newer strategies for condensing DNA. Figure 14.3B shows Atomic Force Microscopy (AFM) images of DNA and DNA condensed by interaction with the cationic polymer polyethyleneimine (PEI). It can be seen that DNA chains are organized in a regular side-by-side configuration in the presence of PEI.

Driven by such a motivation, some of the commonly used condensing agents (Table 14.1) were tested by using YOYO-1 as a probe " YOYO-1


»i0nim^<:i^t'i^:^!i^i$

'lAia/LiwiFh

(I) (11) Figure 14.3B: Typical AFM images (imaged in air) of plasmid DNA (I) and DNA condensed by PEI (II). (3X3 \xm^ scan was performed in the acoustic non-contact mode in air (relative humidity -30%) at room temperature (22 °C) by using a PicoPlus AFM setup from Molecular Imaging Corp. (Arizona, USA). The cantilevers (NCL) had a resonance frequency -190 kHz and force constant of -48 N/m). Parallel arrangement of DNA in PEI condensates (II) could be seen.

fluorescence changes, when used as condensation indicator, show that the efficient transfection agent, PEI packs the DNA tightly when compared to other agents. This is indicated by the higher level of fluorescence quenching in the presence of PEI (Table 14.1). Having demonstrated the use of YOYO-1 in monitoring DNA condensation, the motional dynamics of condensed DNA was studied again by the use of YOYO-1. For this purpose, the concentration of YOYO-1 was reduced significantly (D: P, ~ 1: 2000) such that the probability of formation of H-dimers is negligible. In this situation, the fluorescence signal would be controlled mainly by the dynamics of DNA.

Table 14.1: Fluorescence intensity of DNA-bound YOYO-1 in the presence of various condensing agents.

Condensing agent YOYO-1: DNA (phosphate) Relative fluorescence intensity (D:P) (X,=ei 70nm; Xem=510nm)

1.00 0.08 1.20 0.40 0.20 0.30 0.20 1.05

1. 2. 3. 4. 5. 6. 7. 8.

None PEI PEI CTAB Co ^ NCp7 PEG (mol. PEG (mol.

wt wt

IkD) IkD)

1:50 1:50 1:2,000 1:50 1:50 1:50 1:50 1:2,000

320 N. NAG ET.AL.

Motional dynamics of DNA-bound YOYO-1 inferred through picosecond time-resolved fluorescence anisotropy was used in revealing various dynamic modes of extended and condensed forms of DNA. It is interesting to note that the major part of the body of knowledge on torsional dynamics of DNA backbone has come from fluorescence anisotropy decay kinetics of ethidium bromide ' ' . However, ethidium is not an appropriate probe in the present situation since it gets expelled when DNA condenses^^. In contrast, YOYO-1 (and its analogs) remains bound even in the condensed state making it suitable to study condensed DNA. Furthermore, the anisotropy decay kinetics of DNA-bound YOYO-1 could be satisfactorily fitted to a sum of 2 or 3 exponentials which is in accordance with a restricted rotation model^^. This is unlike the anisotropy decay kinetics of ethidium which requires significantly more complex models^ ' ^ ' ^\ Figure 14.4 shows comparative decay kinetics of ethidium dimer and YOYO-1 in uncondensed DNA. This difference could be due to differences in the binding modes of the two probes and further work will shed more insights into the nature of these interactions.

It is also interesting to note that the motional dynamics of DNA-bound YOYO-1 is able to differentiate the level of supercoiling in DNA. Figure 14.5 shows the fluorescence anisotropy decay curves of YOYO-1 bound to DNA of varying supercoil contents. Although the origin of the difference in the motional dynamics of YOYO-1 is not clear it could still be used as an effective assay for the status of supercoiling.

Figure 14.4: Fluorescence anisotropy decay kinetics of plasmid (3kb) DNA bound ethidium dimer (A: gray line) and YOYO-1 (B: black line). The decay kinetics of YOYO-1 (B) fits to a sum of two exponentials statisfactionally (Table 14.2), unlike that of ethidium dimer (A) which does not fit to a sum exponentials but fits to the model of Schurr and coworkers^\


Figure 14.5: Decay of fluorescence anisotropy of YOYO-1 bound to either supercoiled plasmid DNA (A: gray line) or relaxed plasmid DNA (B: black line) caused by a single-strand cut of the supercolied DNA by the enzyme EcoRl. Parameters obtained from analyses of decay curves are given in Table 14.2.

As mentioned before, fluorescence anisotropy decay kinetics of YOYO-1 in uncondensed circular plasmid DNA could be fitted satisfactorily to a sum of 2 or 3 exponentials. In general, at least three types of molecular motions could be inferred from the decay kinetics of fluorescence anisotropy of fluorophores in macromolecules^^' ^ ' ^ . They are as follows: (a) the internal motion of the fluorophore with respect to the macromolecular matrix; a combination of motions such as helical twisting, propeller twisting, base tilting and rolling could contribute to internal motion " , (b) segmental motion of the region of the macromolecule to which the fluorophore is attached; this could represent the motion of a local region (say, a few nucleotides in a DNA segment) with respect to the entire macromolecule. Torsional and bending motions of adjacent base pairs also could contribute collectively to segmental dynamics; and (c) global tumbling dynamics of the entire macromolecular mass. This motion can be theoretically estimated from Stokes-Einstein relationship (for spherical shape, rotational correlation time, cp = r|V / kT, where V is the molecular volume and r| is the viscosity) and hence experimental value could be compared and identified. Small lengths of ds-DNA could be approximated by rigid cylinders and hydrodynamic equations for cylindrical shapes have been solved " .

The anisotropy value of DNA-bound YOYO-1 drops by ~ 40% within -30 ps (the instrument response time) which corresponds to the internal local motion of bound YOYO-l " ' ^ (Table 14.2). The dominant depolarization

322 N, NAG ET.AL,

Table 14.2: Parameters associated with the decay of fluorescence anisotropy of YOYO-1 in various DNA condensates^.

Rotational correlation time (amplitude), ns

1. 2. 3. 4. 5. 6.

Sample

Uncondensed DNA PEI-DNA CTAB-DNA Co^^-DNA NCp7-DNA PEG(lkD)-DNA

91 (Pi)

<0.1 (0.31) <0.1(0.35) <0.1(0.41) <0.1(0.47) <0.1 (0.41) <0.1 (0.27)

92 (P2)

7.30 (0.69) 2.10(0.16) 3.70 (0.09) 2.90 (0.53) 0.98(0.11) 0.90 (0.22)

93 (Pa)

>100(0.49) >100(0.50)

>100(0.48) 23.1 (0.51)

^Errors in Cp2, Pi, P2, P3 are about 10%.

process having time constant in the range of 4-10 ns could represent the segmental dynamics mentioned above since global tumbling times estimated for the plasmid DNA (2.7 kb) is expected to be several orders of magnitude longer. DNA dynamics shows remarkable changes during the transformation to the condensed form. While the internal dynamics is largely preserved during condensation, the segmental dynamics gets highly dampened '*. The most striking observation on various DNA condensates is the substantial variation in the correlation time associated with segmental dynamics '*. The long (> 100 ns) value of the correlation time (93) observed in the cases of PEl and CTAB indicates freezing of the DNA backbone when condensed by these agents. However, condensation by Co^^ led to an enhancement in the level of segmental flexibility in agreement with the observed decrease in the persistence length in the presence of Co^^ ^ . Furthermore, the observation of another intermediate level of rotational dynamics (92 ~ 1-3 ns with small amplitude, Table 14.2) indicates the presence of higher level of restricted segmental flexibility in the PEI condensate when compared to the CTAB condendate. It is likely that the higher transfection efficiency of PEI '' could be related to such modulations in the dynamics of condensed DNA.

Condensation of DNA has also been achieved by osmotic exclusion or macromolecular crowding offered by a variety of non-interacting polymers such as polyethylene glycol (PEG)^ ' . How do the DNA molecules condensed by either charge neutralizing agents such as cationic polymers/detergents mentioned above or molecular crowding agents such as PEG compare with each other? Are their structure and dynamics similar to each other? If not similar, what are the essential differences? To answer these questions, we compared the dynamics of base and the backbone of DNA condensed by either class of condensing agents. Our results show that DNA condensed by PEG is less compact and less rigid when compared to DNA condensed by binding agents (Table 14.2).


14.3.4. DNA condensation by the nucleocapsid protein probed by YOYO-1 fluorescence

The nucleocapsid protein, NCp7, of HIV-1 virus has a key role in the replication of the virus^^ It is a small basic protein having two copies of a highly conserved zinc finger motif CCHC which binds Zn ^ with very high affinity^ ^ One of the key roles of NCp7 is to coat the genomic RNA in the infectious virus^ ' ^ and protect the RNA against degradation by RNAases^" . NCp7 is also critical for efficient and complete proviral DNA synthesis by promoting both the initiation and the two strand transfer steps during reverse transcription^^"^ . From the observations that NCp7 also binds to ds-DNA^^"^^\ it is hypothesized that NCp7 binds to the newly synthesized viral DNA and protects it from cellular nucleases during reverse transcription and its transport to the nucleus for the integration step^ ' ' ° ' ^ . YOYO-1 fluorescence was used in elucidating the plausible mechanism of DNA protection by NCp7^^.

Using plasmid DNA as a model system it was shown that NCp7 binds to ds-DNA and causes it to condense^^. The YOYO-1 fluorescence-based method shown earlier (see section 3.2) was used to demonstrate DNA condensation. Here again, DNA condensation was inferred from the several fold decrease in the quantum yield of DNA-bound YOYO-1 (Table 14.1). That this decrease is not due to expulsion of YOYO-1 by NCp7 was demonstrated by the fact that the decrease in fluorescence intensity is seen only at sufficiently high levels of YOYO-1 such that H-dimers mentioned above could form on condensation.

The molecular mechanism of DNA condensation by NCp7 was delineated by following the condensation signal with either the truncated NCp7 or with the Zn-fingerless peptide. It was shown that efficient condensation of DNA requires the full length of the peptide along with the fingers. Thus it is clear that elements other than basic amino acids are required for efficient condensation of DNA.

The dynamics of DNA condensed by NCp7 was monitored by the fluorescence anisotropy decay kinetics of YOYO-1 bound sparsely^^ (Table 14.2). While all the NCp7 peptides caused freezing of DNA backbone, the truncated peptide NC(12-55) which retains the Zn fingers caused only a 3 fold decrease in the rate of segmental dynamics^^. Once again we could gain information on the mechanism of DNA protection by correlating the level of protection with dynamics. Thus from the fluorescence dynamics of YOYO-1, we could propose that the histone-like property of NCp7 leading to DNA condensation contributes to viral DNA stability.

14.3.5. DNA Dynamics in Chromosomes from Picogreen Fluorescence

Physical and chemical integrity of DNA in chromosomes is a matter of great concern. The presence and the level of single-stranded breaks and the extent of rigidity of DNA while bound to histones are some of the important issues. Several physiological processes control and rely upon the level of DNA integrity in chromosomes. For example, disintegration of chromosomes is

324 N. NAG ET.AL.

encountered during apoptosis and here the information on the early events occurring at the DNA level before the occurrence of visual disintegration of chromosomes is lacking^ " ' ^ ^ Similarly, the level of rigidity and compactness is a concern for efficient transcription to occur. Hence there is a tremendous need for new methodologies for monitoring the state of DNA in chromosomes.

The fluorescence probe PicoGreen (PG, Figure 14.1 A), which is an unsymmetric monomethine cyanine dye, has been finding interesting applications in assessing the level of ss-DNA in a mixture containing both ss and ds DNA^ . The difference in the fluorescence lifetime of PG when bound to either ss-DNA or ds-DNA seems to originate from the difference in the mode of binding^^' ^ . Furthermore, single strand nicks generated by y irradiation of DNA has also been quantified from the analysis of time-resolved fluorescence of PG^ . The fluorescence quantum yield of cyanine dyes such as PG and YOYO are controlled predominantly by molecular rotation-aided internal conversion^^ ' ^ . Hence, y irradiation sensitivity of fluorescence lifetime of DNA-bound PG could have arisen due to changes in the dynamics of DNA following irradiation-caused single strand nicks.

Damage to DNA during programmed cell death (apoptosis) is a well documented phenomenon. As mentioned above, there is a lack of information on the sequence of events occurring during apoptosis. For example, it would be useful to have a temporal map of the changes in the integrity of DNA during apoptosis. We used time-domain fluorescence of PG to study DNA dynamics during UV-irradiation induced apoptosis in the unicellular green algae, Chlamydomonas reinhardtii. Characteristic hallmarks of apoptosis have been observed when chlamydomonas cells were exposed to UV ^^l The mam motivation was to see whether DNA damage occurs well before other apoptotic markers show up and the actual event of cell death. Figure 14.6B shows typical curves of PG fluorescence intensity decays in Chlamydomonas cells and the effect of UV-irradiation. The data were collected either from single cell or from a group of cells under a fluorescence microscope coupled to a pico-second laser system^^^ (Figure 14.6A). Analysis of decay curves showed that the mean fluorescence lifetime of PG exhibited significant reduction following UV treatment (Table 14.3). That this decrease is not due to a direct effect of the UV dose on DNA was checked in experiments wherein either plasmid DNA or calf thymus DNA was irradiated with UV and the fluorescence lifetime of PG was measured. In this case, the decay curves were nearly identical to each other (Table 14.3) showing clearly that the decrease in the mean lifetime observed in chlamydomonas cells must have arisen due to


[IMAGE]

FLUORESCENCE MICROSCOPE

(ii)

LASER (i)

COMPUTER

D[ TAG

MCA

CFD TCSPC (Hi)

Figure 14.6A: Schematic diagram of the time-resolved fluorescence microscope setup used for aquiring time-resolved fluorescence data in single cells. The main components of the set-up are the following: (i) a pico-second laser system, (ii) an inverted fluorescence microscope, and (iii) a time-correlated single photon counting (TCSPC) spectrometer. Laser - Mode-locked Nd-YAG laser pumped dye laser/ Nd-Vanadate pumped Ti-Saphire, frequency doubled output; M- Mirror; P-polariser; F- Filter; PH- Pinhole; DM - Dichroic mirror; OBJ - Objective lens; PMT 2020 - Signal photomultiplier tube; PMT - Start signal photomultiplier; CFD - Constant Fraction Discriminator; TAC - Time-to-amplitude converter; MCA - Multi Channel Pulse Height Analyzer.

DNA damage induced by an apoptotic agent rather than by the direct action of UV on chromosomal DNA.

What could be the origin of the observed changes in fluorescence decay kinetics of PG in Chlamydomonas cells following UV irradiation? Information from the work of y-irradiation of DNAss-svand our data on

various model DNA systems (Table 14.3) suggests that single strand nicks as a possible cause. Furthermore, preliminary work has suggested that the changes observed in the cells occur several tens of hours prior to the sighting of well-recognized apoptotic markers. Future work using fluorescence probes is expected to offer deep insights into spatial and temporal maps of chromosomal DNA dynamics, in relation to a cellular process such as apoptosis.

14.4. DNA DYNAMICS FROM SITE-SPECIFIC FLUORESCENCE PROBES

14.4.1. DNA dynamics in RecA -DNA filaments

One of the unique applications of fluorescence is the ability to address the question why apparently similar structures do not elicit similar activities.

326 N. N AG ET.AL,

g 0.1 -{

0.01

" Time(ns) ^ 10

Figure 14.6B: Typical traces of decay of fluorescence intensity in PicoGreen (PG) labeled single cells of Chlamidomonas reinhardti (inset). Decay kinetics in control cells (A) and in cells irradiated with increasing doses of UV (B-D) are shown. Acceleration of intensity decay caused by UV-irradiation of cells can be seen (see Table 14.3 for parameters obtained from analyses of decay kinetics).

Table 14.3: Parameters associated with fluorescence intensity decay of PicoGreen (PG) in cells and in plasmid DNA.

Sample UV I

1. Chlamydomonas cells'

2. Plasmid DNA (pGEXKG)

)osage (a. u.)

0 200 600 800

0 800

Fuorescence lifetimes (amplitude) (ns)

X 1 ( a i )

2.85(0.61) 2.77 (0.58) 2.49 (0.43) 2.36 (0.39) 3.73 (0.71) 3.68 (0.68)

Meanlifetime

12 (a2)

0.77 (0.39) 0.76 (0.42) 0.75 (0.57) 0.72(0.61) 1.27(0.29) 1.22(0.32)

Tm (ns)

2.03 1.93 1.49 1.36 3.02 2.89

Chlamydomonas cells were irradiated with different doses of UV followed by labeling with PicoGreen.

Tm = Z a j Tj , where Tm is mean fluorescence lifetime, Xj and a\ are ith component of multi-exponential decay time and its corresponding amplitudes respectively.

This paradigm was tested recently in a DNA-protein system by site-specific labeling of DNA with 2-AP (Figure 14.IB). RecA, a protein from E.coli catalyzes DNA recombination by reciprocal exchange of strands between two


DNA molecules which have homologous regions in them^^ . One of the elementary steps in the early part of the recombination process is the formation of a three-stranded intermediate complex involving RecA-coated single strand and a homologous ds-DNA. Strand exchange between the homologous tracts requires hydrolysis of ATP by RecA, which has coated one of the strands. The non-hydrolysable analogue of ATP (ATPyS) does not allow the strand exchange although the three stranded structures gets formed^ ' ^ . The dynamics of DNA, monitored through NMR and fluorescence depolarization kinetics of 2-AP shows that RecA-catalysed DNA recombination is initiated by a high energy form of DNA in the nucleoprotein filaments, where the bases are highly unstacked and the backbone is highly unwound^^ ' ^ . Interestingly, only the energetics consequent to ATP binding, rather than its hydrolysis, seems sufficient to mediate such a high energy structural hallmark of a recombination filament.

What is the role of ATP hydrolysis in RecA-mediated DNA recombination, especially when we note that the three stranded structure, which is an obligatory intermediate as mentioned above, is seen even in the presence of ATPyS ? Is there an obligatory involvement of DNA dynamics? To provide answers to these questions, we used fluorescence dynamics of 2-AP placed at specific locations in DNA^ . 2AP, in single stranded DNA shows three fluorescence lifetimes (~ 0.5, - 2.0 and ~ 7.0 ns) possibly originating from three conformations of 2AP as compared to the single lifetime (12.0 ns) in the free form^ " ' ^ ' ^. A new short lifetime (50 -100 ps) component gets generated following the formation of ds-DNA by annealing with the complimentary strand. This short component has been assigned as a signature of base-pairing^ ' ^ . The amplitude of the shortest lifetime component, which is a quantitative measure of base-pairing status, would therefore provide important information about duplex DNA in RecA-nucleoprotein complexes. The reduction in the extent of base-pairing is revealed by a reduction in the amplitude. This reduction was more in complexes formed with ATPyS than that with ATP, an indication of the accentuated effect of RecA binding in the former as compared to the later. The decrease in the amplitude of the short life lifetime is consistent with the highly unstacked double helix in RecA-nucleoprotein complexes where base pairs are poised towards homologous pairing with the incoming sequence"^-"«'"I

The motional dynamics of 2-AP, monitored through time-resolved fluorescence anisotropy decay showed two rotational correlation times (-- 0.5 and > 3.0 ns) when incorporated into DNA. These could be interpreted as due to the internal (local) motion of 2-AP with respect to the strand backbone and either segmental or tumbling dynamics of the strand respectively^^. The longer correlation time which, in some cases, is associated with overall tumbling of the whole strand scales linearly with the molecular mass of strand hence gives the signature of nucleoprotein filament formation. 2-AP in RecA-DNA complexes exhibits a novel motional dynamics with a rotational correlation time of 7-10 ns, under ATP hydrolysis conditions. In presence of ATPyS (a non-hydrolysable

328 f^. NAG ET.AL.

analog), the fast dynamic motion ceased and the global tumbling motion with a rotational correlation time of >20 ns got revealed instead.

The correlation time of 7-10 ns observed in RecA-DNA complex during ATP hydrolysis represents a novel segmental dynamics in DNA hitherto unreported. The best estimates of turnover rate of ATP hydrolysis by RecA suggest that about 20-30 ATP molecules are hydrolysed per RecA monomer per sec with a high co-operativity in the filament^ ^ . In contrast, our observation on segmental dynamics is in ns time scales. This would raise the question as to how such a slow rate of ATP hydrolysis manifest its action in the ns rotational dynamics of the nucleofilament. One could envisage a picture wherein concerted hydrolysis of ATP by RecA monomers^ ^ in the filament could result in a standing wave of segmental dynamics of the filament. Subsequently, the presence of such a standing wave of perturbation would then become observable at all time scales.

Has the ATP-induced flexibility any role to play in the overall process of DNA recombination? Although ATPyS is sufficient to form RecA nucleofilaments, it is insufficient to generate RecA treadmilling and efficient strand exchange across long stretches of DNA, pushing the branch migration through sequence impediments such as insertions/deletions ^ ' ^ ' o j ^ - jjj giy that the main difference in the modes of action of ATP and its nonhydrolysable ATPyS lies in the dynamics offered by ATP hydrolysis process. This dynamics could be visualized as a concerted rocking motion of the RecA coat on the polydeoxynucleotide chain and thereby conferring the strand a significant level of flexibility. Thus our observations strongly suggest that the increased level of flexibility of the nucleotide backbone is the answer to the puzzle of ATP requirement for RecA-catalyzed DNA recombination in general and strand-realignment activity across repeats in particular. It is tempting to speculate that the enhanced segmental motion of a DNA-nucleotide triggered by a DNA-binder in the presence of ATP-hydrolysis could well be a general enough dynamic effect that could explain the results of many DN A-transactions specific to ATP-hydrolysis.

14.4.2. Position-dependent DNA dynamics

The information content of DNA is often assumed to be solely deposited in its sequence. While this assertion holds in a vast majority of situations, there are striking examples wherein the information retrieval from a sequence is modulated by the position of the sequence with respect to an open end. Such extra-sequence information could originate from position-dependence of structure and dynamics of DNA stretches. Position-dependent dynamics are implicated in a variety of situations such as specific and sequence-independent binding by regulatory proteins to ds-DNA ends (blunt ended or single stranded overhangs^^ ' ^^), ss-DNA and junctions of ss-DNA/ds-DNA^^^' ^^\ Fraying of end segments of ds-DNA has been inferred in several situations^^^ and implicated in some aspects of DNA repair processes^^^. Specific recognition of ds DNA ends by proteins such as translin^^^ is likely to be mediated by


end-specific DNA dynamics. It is clear that a specific sequence of DNA exhibits varying biological properties largely under the influence of neighboring sequences. The physical basis of such position-dependent properties of DNA could stem from interesting changes in the dynamics of local sequences in DNA.

We have used the fluorescent analog of adenine, 2-aminopurine (2-AP) to monitor the position-dependence of dynamics of base and the backbone of ss-and ds-DNA. As mentioned above, 2-AP, which forms base-pair with thymine, has been very effectively used to study both specific and non-specific aspects of structure and dynamics of DNA in a wide variety of situations ' ' . Dynamic fluorescence parameters such as excited state lifetime and rotational depolarization time of 2-AP located at specific positions were used to infer DNA dynamics. We find that the motional dynamics of base and DNA backbone depends on the position in unexpected ways (in some cases) in ss- , ds- and mixed DNAs.

Fraying at the ends of ds-DNA leading to dynamic rupture and reformation of base-pairs is an important aspect of DNA dynamics in repair processes related to broken ends of DNA. Similarly, the question as to how the base-pair dynamics is distributed along the length of DNA is quite relevant for understanding DNA-protein interactions. These could be quantitatively studied by site-specific labeling with 2-AP. Many of the results from these studies are highly counter-intuitive. For example, one could expect the nucleotide at the ends of ds-DNA to fray. However, we observed fraying even at the third base from the end (Table 14.4). The level of fi'aying was inferred either from the mean lifetime or from the shortest lifetime whose value and amplitude reflects the level of base-pairing^^' ^^. Another counter-intuitive observation was the increasing level of motional dynamics of 2-AP when it was moved from the end to the middle of ds-DNA. This was inferred from the amplitude associated with the shorter rotational correlation time extracted from analysis of fluorescence anisotropy decay kinetics (Table 14.4). It is tempting to speculate that this behavior, which is not observed in ss-DNA (Table 14.4), could have an origin in a putative correlated dynamics of base-pairs. Such a model of enhanced correlated dynamics of base-pairs far from the ends of ds-DNA is quite attractive and hence demands more work in future.

14.4.3. Mismatch recognition and DNA dynamics

DNA mismatch repair (MMR) system, an evolutionarily conserved biochemical pathway, plays an important role in regulating the genome by correcting base mismatches arising from replication errors (error rate lO" ) ^ " ^ . In E. coli, repair is initiated by MutS protein, a member of ABC ATPase superfamily that recognizes a mismatch/Insertion-Deletion-Loop (IDL) with an affinity that is only several folds higher than that of its binding to normal matched duplex^^ ' ^ . After mismatch recognition, MutS with the assistance of MutL initiates the mismatch repair by activating MutH that nicks the newly

330 N.NAG^r.^i .

Table 14.4: Parameters associated with the motional dynamics of 2-AP at various locations in DNA.

Rotational correlation times (amplitude), ns Mean fluorescence

1 2 3 4 5 6

A30(l)-T30 A30(3)-T30 A30(15)-T30 A30(l) A30 (3) A30(15)

1.00 0.70 0.40 1.70 2.30 2.40

DNA Sample^ lifetime, T„, (ns)*" cpi ( pi) 92 (P2)

0.18(0.37) 3.9(0.63) 0.33(0.53) 11.0(0.47) 0.33 (0.60) 12.0 (0.45) 0.15(0.60) 1.1 (0.40) 0.18(0.44) 1.9(0.56) 0.14(0.40) 3.0(0.60)

The location of 2-AP from the 5' end of poly A strand (A30) is indicated in parenthesis. A30 forms duplex DNA with T30.

Tm = Z a , Tj , where Tm is mean fluorescence lifetime, Tj and OLi are 1 component of multi-exponential decay time and its corresponding amplitudes respectively.

synthesized, unmethylated 'GATC sequence strand^^ ' ^ , following which a concerted action of helicase, exonuclease, polymerase, and ligase ftmctions ensue, thereby restoring the correct complementary sequence in the DNA

MutS, which exists in dimer-tetramer equilibrium at physiological conditions^^^ possess a conserved mismatch recognizing and ATPase domain^^ ' ^^^' ^ . Several studies have shown that MutS has variable affinity for different mismatches but it forms strongest complexes with G:T mismatches and single unpaired bases " ' ' . Crystal structure has shown that only the one monomer out of the dimer has specific mismatch-binding contacts, involving exclusively a Phe-X-Glu motif at the N-terminal mismatch-recognition domain " ' ^^^, The equivalent domain in the second monomer has only loose DNA backbone contacts elsewhere^^' ^^^' ^^^. Lamer et. al. ' ^ have shown, by crystallographic studies, that when E. coli MutS is bound to a G:T mismatch, the DNA which was mostly in B DNA conformation showed a strong kink of ~ 60^ at the mismatch towards the major groove. Through the combined action of the mismatch-recognition domain and the clamp domains, MutS involves other residues also to orient mismatched base-pair and continued to kink the DNA imparting a favorable interaction between the DNA and the protein playing a crucial role in mismatch recognition process^^' ^^^. AFM studies on MutS-DNA complex has revealed a single population of molecules in which DNA is bent (by ~ 50^) at MutS bound non-mismatch sites, but two populations of, bent and unbent (3:1), at MutS bound mismatch sites " . It could be suggested that MutS binds to DNA nonspecifically and bends it in search of mismatches and upon recognizing the mismatch it undergoes a conformation change resulting in a


Time (ns)

Figure 14.7: Typical traces of decay of fluorescence anisotropy of 2-AP located at a site either next to a mismatch ( Seq 1) (A: black line) or far away from the site of mismatch ( Seq 2) (B: gray line). A* denotes 2-AP. Seq 1: 5' — T ATG T G C — 3' Seq 2: 5' — T ATG TGC— 3'

3 ' — A TAG A* TG — 5 ' 3' — A* TAG ATG—5' Rotational correlation times obtained from analyses are as follows, (A) 0.15 (0.50); 1.5 (0.2); 7.0 (0.30), (B) 0.30 (0.35); 1.5(0.18); 9.3 (0.45). The values in parenthesis are associated amplitudes.

kink at the binding site in the DNA. Finally a further conformational change at the ultimate recognition complex makes an unbent DNA which could be functional for MMR ' ^

Thus, structural alterations of MutS at the mismatch site, which is different from that exerted at a non-mismatch site, could be thought of as a mechanism in mismatch recognition by MutS. Additionally, the suggestion that alterations in the dynamics of DNA at mismatch sites having contributions in mismatch recognition is an attractive proposal and is worthy of experimental verification. This suggestion stems from the current theme elaborated in this chapter; viz. DNA dynamics plays an important role in several functions of DNA.

To compare the dynamics at a mismatch and non-mismatch sites, we incorporated 2-AP next to a G:T mismatch and at several other non-mismatch positions. Rotational diffusion of 2-AP at these locations, monitored by time-resolved fluorescence anisotropy (Figure 14.7), clearly shows differences in the dynamics. Analysis of decay kinetics shows that the local motion of 2-AP placed next to the mismatch site is less restricted when compared to that at any location within matched regions (see legend to Figure 14.7). The loss of base-pairing at the mismatched G:T site seems to enhance the motional dynamics of even bases next to it. Thus it is tempting to speculate that this change in the dynamic feature in the mismatch region could be used by MutS in locating a mismatch sprinkled among matched regions.

332 N. N AG ET.AL.

14.5. CONCLUSIONS

One of the recurring themes in modem biology, namely, dynamics along with structure forms the basis of function of biomolecular systems is getting increasing level of support from various experimental and molecular dynamics simulation techniques. In this chapter, the power of time-resolved fluorescence spectroscopy in bringing to light various intricate aspects of DNA dynamics and, in some cases, their correlation with function has been shown. We believe that this is a very fruitful approach in seeking explanations for complex activity of DNA based systems especially in the light of the enormous level of structural information being gathered in recent times.


We thank Profs. J-L. Darlix (ENS, Lyon, France) and B. Roques (Univ. Rene Descartes, France) for their collaboration in the research work described here. We are grateful to Prof Periasamy (TIFR, Mumbai) for providing us with the software used in the analysis of time-resolved fluorescence data and advice in using them efficiently.

14.7. REFERENCES

1. P. W. Fenimore, H. Frauenfelder, B. H. McMahon, and F. G. Parak, Slaving: Solvent fluctuations dominate protein dynamics and functions, Proc. Natl Acad. ScL USA 99(25), 16047-16051 (2002).

2. 2. S. Hammes-Schiffer, Impact of enzyme motion on activity. Biochemistry 41 (45), 13335-13343(2002).

3. L. Brand, and M.L. John (eds) Meth. Enzymology 21H (1997) 4. P. M. Hwang, and L. E. Kay, Solution structure and dynamics of integral membrane

proteins by NMR: a case study involving the enzyme PagP. Meth. Enzymol. 394, 335-50 (2005).

5. J. M. Schurr, Rotational diffusion of deformable macromolecules with mean local cylindrical symmetry, Chem. Phys. 84, 71-96 (1984).

6. S. B. Smith, Y. Cui, and C. Bustamante, Overstretching B-DNA: the elastic response of individual double-stranded and single-stranded DNA molecules, Science 271(5250), 795-799 (1996).

7. Y. Seol, G. M. Skinner, and K. Visscher, Elastic Properties of a Single-Stranded Charged Homopolymeric Ribonucleotide, Phys. Rev Lett. 93( 11), 118102-1-4 (2004)

8. R. E. Dickerson, and H. R. Drew. Structure of a 5-DNA dodecamer : II. Influence of base sequence on helix structure, J. Molec. Bio. 49 (4), 761-786 (1981)

9. J. M. Schurr, Effect of anisotropic bending rigidity and fmite twisting rigidity on statistical properties of DNA model filaments, Biopolymers 24 (10), 1233-1246 (1985).

10. E. N. Trifonov, and J. L. Sussman, The Pitch of Chromatin DNA is Reflected in its Nucleotide Sequence, Proc. Natl Acad. Sci. USA 11 (7), 3816-3820 (1980)


11. H. R. Drew, and C. R. Calladine, Sequence- specific positioning of core histones on an 860 base-pair DNA, J. Molec. Bio. 195 (1), 143-173 (1987)

12. G. Obmolova, C. Ban, P. Hsieh, and W. Yang, Crystal structures of mismatch repair protein MutS and its complex with a substrate DNA, Nature. 407(6805), 703-710 (2000)

13. E. B. Brauns, Catherine J. Murphy, and Mark A. Berg. Local Dynamics in DNA by Temperature-Dependent Stokes Shifts of an Intercalated Dye, J. Am. Chem. Soc. 120 (10), 2449-2556 (1998).

14. G. S. Manning. Breathing and bending fluctuations in DNA modeled by an open-base-pair kink coupled to axial compression, Biopolymers. 22(2), 689-729 (1983).

15. M. Gueron, M. Kochoyan, and J-L. Leroy, A single mode of DNA base-pair opening drives imino proton exchange, Nature. 328 (6125), 89-92 (1987).

16. A. S. Benight, J. M. Schurr, P. F. Flynn, B. R. Reid, and D. E. Wemmer, Melting of a self-complementary DNA minicircle : Comparison of optical melting theory with exchange broadening of the nuclear magnetic resonance spectrum, J. Molec. Bio. 200(2), 377-399 (1988).

17. J. R. Lakowicz, Principles of Fluorescence Spectroscopy, 2nd edn. (Kluwer Academic/ Plenum Publishers, New York 2000).

18. T. Ramreddy, S. Sen, B. J. Rao, and G. Krishnamoorthy. DNA dynamics in RecA-DNA filaments; ATP hydrolysis-related flexibility in DNA, Biochemistry 42(41), 12085-12094(2003).

19. P. Sharma, R. Varma, R. C. Sarasij, Ira, K. Gousset, G. Krishnamoorthy, M. Rao, and S. Mayor, Nanoscale organization of multiple GPI-anchored proteins in living cell membranes. Cell 116(4), 577-89 (2004)

20. A. Anders, DNA fluorescence at room temperature excited by means of dye laser, Chem. Phy. Lett. 81(2), 270-272 (1981).

21. T.I. Aoki, The fluorescence of native DNA at room temperature, Chem. Phy. Lett. 92(3), 270-272 (1982).

22. S. Georghiou, T. M. Norlund, and A. M. Saim, Picosecond fluorescence decay time measurements of nucleic acids at room temperature in aqueous solution, Photochem. Photobiol. 41, 209-212 (1985).

23. S. Georghiou, T. D. Bradrick, A. Philippetis, and J.M. Beechem, Large-amplitude picosecond anisotropy decay of the intrinsic fluorescence of double-stranded DNA, Biophys. J. 70(4), 1909-1922 (1996)

24. C. R. Guest, R. A. Hochstrasser, C. G. Dupuy, D. J. Allen, S. J. Benkovic, D. P. Millar, Interaction of DNA with the Klenow fragment of DNA polymerase I studied by time -resolved fluorescence spectroscopy. Biochemistry 30(36), 8759-70 (1991)

25. W. A. Franklin, J. D. Locker, Ethidium bromide: a nucleic acid stain for tissue section, J. Histochem. Cytochem. 29(4), 572-6 (1981).

26. M. L. Barcellona, and E. Gratton, Torsional Dynamics and Orientation of DNA-DAPI Complexes. Biochemistry 35(1), 321-333 (1996)

27. B. G. Rami, L. S. Chin, B. E. Lazio, and S. K. Singh, Okadaic acid-induced apoptosis in malignant glioma cells, Neurosurg Focus 14 (2), 1-7 (2003)

28. B. Becker, J. Clapper, K. R. Harkins, and J. A. Olson, In situ screening assay for cell viability using a dimeric cyanine nucleic acid stain. Anal. Biochem., 221(1), 78-84(1994)

29. S. Gurrieri, K. S. Wells, I. D. Johnson, and C. Bustamante, Direct visualization of individual DNA molecules by fluorescence microscopy: characterization of the factors affecting signal/background and optimization of imaging conditions using YOYO. Anal. Biochem. 249(1), 44-53 (1997)

30. N. Ashley, D. Harris, and J. Poulton, Detection of mitochondrial DNA depletion in living human cells using PicoGreen staining, Exp. Cell Res. 303(2), 432-46 (2005)

334 N.NAG^r.^x.

31. J. M. Schurr, B. S. Fujimoto, P. Wu, and L. Song, in Topics in Fluorescence Spectroscopy VoL3, edited by J. R. Lakowicz, (Plenum Press, New York (1992) pp 137-229

32. Ph. Wahl, J. Paoletti, and J.-B. Le Pecq, Decay of fluorescence emission anisotropy of the ethidium bromide-DNA complex evidence for an internal motion in DNA, Proc. Natl Acad. Sci. USA 65 (2), 417-421 (1970)

33. J.-P. Clamme, S. Bemacchi, C. Vuilleumier, G. Duportail, and Y. Mely, Gene transfer by cationic surfactants is essentially limited by the trapping of the surfactant/DNA complexes onto the cell membrane: a fluorescence investigation. Biochim. Biophys. Acta 1467(2), 347-361 (2000).

34. G. Krishnamoorthy, G. Duportail, and Y. Me'ly, Structure and Dynamics of Condensed DNA Probed by l,r-(4,4,8,8-Tetramethyl-4,8-diazaundecamethylene)bis[4-[[3- methylbenz-1,3-oxazol-2-yl]methylidine]-1,4-dihydroquinolinium] Tetraiodide Fluorescence, Biochemistry 41(51), 15277-15287 (2002)

35. G. Cosa, K. S. Focsaneanu, J. R. McLean, J. P. McNamee, and J. C. Scaiano Photophysical properties of fluorescent DNA-dyes bound to single- and double-stranded DNA in aqueous buffered solution. Photochem Photobiol 73(6), 585-99 (2001)

36. C. Schweitzer, and J. C. Scaiano, Selective binding and local photophysics of fluorescent cyanine dye Picogreen in double stranded and single-stranded DNA, Phys. Chem. Chem. Phys. 5, 4911-4917 (2003)

37. G. Cosa, A. L. Vinnette, J. R. N. McLean, and J. C. Scaiano, DNA damage detection technique applying time-resolved fluorescence measurements, Anal. Chem. 74(24), 6163-6169 (2002)

38. J. Brown, T. Brown, and K. R. Fox, Affinity of mismatch-binding protein MutS for heteroduplexes containing different mismatches, Biochem. J. 354, 627-633 (2001).

39. E. P. Efimova, E. P. Delver, and A. A. Belogurov. 2-Aminopurine and 5-bromouracil induce alleviation of type I restriction in Escherichia coli: mismatches function as inducing signals? Mol Gen Genet. 214(2), 317-20 (1988)

40. D. C. Ward, E. Reich, and L. Stryer, Fluorescence Studies of Nucleotides and Polynucleotides. I formycin, 2-aminopurine riboside, 2,6-diaminopurine riboside, and their derivatives, J. Biol. Chem. 244(5), 1228 - 1237 (1969)

41. L. C. Sowers, G. V. Fazakerleyt, R. Eritjat, B. E. Kaplant, and M. F. Goodman, Base pairing and mutagenesis: Observation of a protonated base-pair between 2-aminopurine and cytosine in an oligonucleotide by proton NMR, Proc. Natl. Acad. Sci. USA 83(15), 5434-38 (1986)

42. J. M. Jean, and K. B. Hall, 2-Aminopurine fluorescence quenching and lifetimes: Role of base stacking, Proc. Natl. Acad Sci. USA 98(1), 37-41 (2001)

43. N. Ben Gaied, N. Glasser, N. Ramalanjaona, H. Beltz, P. Wolff, R. Marquet, A. Burger, and Y Mely, 8-vinyl-deoxyadenosine, an alternative fluorescent nucleoside Analog to 2'-deoxyribosyl-2-aminopurine with improved properties Nucleic Acids /?^5. 33(3), 1031-9.(2005)

44. C. Liu, and C. T. Martin, Fluorescence characterization of the transcription bubble in elongation complexes of T7 RNA polymerase, J. Molec. Bio. 308(3), 465-475 (2001).

45. C. Liu and C. T. Martin, Promoter Clearance by T7 RNA Polymerase initial bubble collapse and transcript dissociation monitored by base analog fluorescence. J. Biol. Chem. 277(4), 2725-2731 (2002).

46. S. C. Srivastava, S. K. Raza, and R. Misra, 1, N6-etheno deoxy and ribo adenosine and 3,N4-etheno deoxy and ribo cytidine phosphoramidites. Strongly fluorescent


structures for selective introduction in defined sequence DNA and RNA molecules, Nucleic. Acids Res. 22(1), 1296-1304 (1994)

47. M. E. Hawkins, W. Pfleiderer, O. Jungmann, and F. M. Balis, Synthesis and fluorescence characterization of pteridine adenosine nucleoside analogs for DNA incorporation, Anal. Biochem. 298(2), 231 -40 (2001)

48. M. E. Hawkins, W. Pfleiderer, F. M. Balis, D. Porter, and J. R. Knutson, Fluorescence properties of pteridine nucleoside analogs as monomers and incorporated into oligonucleotides, ^«a/. Biochem. 244(1), 86-95 (1997).

49. D. A. Barawkar, and K. N. Ganesh, Fluorescent d(CGCGAATTCGCG): characterization of major groove polarity and study of minor groove interactions through a major groove semantophore conjugate. Nucleic Acids Res. 23(1), 159-164(1995)

50. P. Hagmar, M. Bailey, G. Tong, J. Haralambidis, W. H. Sawyer, and B. E. Davidson, Synthesis and characterisation of fluorescent oligonucleotides. Effect of internal labelling on protein recognition, Biochim Biophys Acta 1244(2-3), 259-68 (1995).

51. M. Bailey, P. Hagmar, D. P. Millar, B. E. Davidson, G. Tong, J. Haralambidis, and W. H. Sawyer, Interaction between the Escherichia coli Regulatory protein TyrR and DNA: a fluorescence footprinting study. Biochemistry 34(48), 15802-12 (1995).

52. T. E. Carver Jr, R. A. Hochstrasser, D. P. Millar. Proofreading DNA: recognition of aberrant DNA termini by the Klenow fragment of DNA polymerase I. Proc Natl AcadSci USA 91(22), 10670-4 (1994)

53. K. M. Parkhurst, M. Brenowitz, and L. J. Parkhurst, Simultaneous binding and bending of promoter DNA by the TATA binding protein: real time kinetic measurements. Biochemistry 35(23), 7459-7465 (1996)

54. F. W. Sevenich, J. Langowski, V. Weiss, and K. Rippe, DNA binding and oligomerization of NtrC studied by fluorescence anisotropy and fluorescence correlation spectroscopy. Nucleic Acids Res. 26(6), 1373-81 (1998)

55. J. R Grunwell, J. L. Glass, T. D. Lacoste, A. A. Deniz, D. S. Chemla, and P. G. Schultz, Monitoring the conformational fluctuations of DNA hairpins using single-pair fluorescence resonance energy transfer. / . Am. Chem. Soc. 123(18), 4295-303 (2001).

56. R. Groben, and L. Medlin, In Situ Hybridization of Phytoplankton Using Fluorescently Labeled rRNA Probes, Methods Enzymol. 395,299-310 (2005).

57. S. Bemacchi, E. Piemont, N. Potier, A. Dorsselaer, and Y. Mely, Excitonic Heterodimer formation in an HIV-1 oligonucleotide labeled with a donor-acceptor pair used for fluorescence resonance energy transfer, Biophys J. 84(1), 643-54 (2003).

58. G. Arents, and E. N. Moudrianakis, Topography of the histone octamer surface: repeating structural motifs utilized in the docking of nucleosomal DNA, Proc. Natl. Acad Sci. USA 90(22), 10489-93 (1993).

59. J. A. Schellman, and N. Parthasarathy, X-ray diffraction studies on cation-collapsed DNA, J. Mo.lBiol 175(3), 313-29 (1984

60. A. D. Miller, Cationic Liposomes for Gene Therapy, Angew. Chem. Int. Ed. 37, 1768-1785(1998).

61. Y. S. Tarahovsky, V. A. Rakhmanova, R. M. Epand, and R. C. MacDonald, High temperature stabilization of DNA in complexes with cationic lipids, Biophys. J. 82(1 pt 1), 264-273(2002)

62. J.-P. Behr, Synthetic gene transfer vectors, Ace. Chem. Res. 26, 274-278 (1993)

336 N.NAG^r.^L.

63. M. X. Tang, and F. C. Szoka, The influence of polymer structure on the interactions of cationic polymers with DNA and morphology of the resulting complexes, Gene r/?er. 4(8), 823-832 (1997)

64. S. M. Mel'nikov, V. G. Sergeyev, Y. S. Melnikova, and K. Yoshikawa, Folding of long DNA chains in the presence of distearyl dimethyl ammonium bromide and unfolding induced by neural liposome, J. Chem. Soc, Faraday Trans. 93(2), 283-288 (1997)

65. A. Remy-Kristensen, J.-P. Clamme, C. Vuilleumier, J.- G. Kuhry, and Y. Mely, Role of endocytosis in the transfection of L929 fibroblasts by polyethylenimine/DNA complexes, Biochim. Biophys. Acta. 1514(1), 21-32 (2001)

66. W. T. Godbey, K. K. Wu, and A. G. Mikos, Tracking the intracellular path of poly(ethylenimine)/DNA complexes for gene delivery. Proc. Natl. Acad. Sci. USA 96(9), 5177-81 (1999)

67. B. Pitard, N. Oudrhiri, J. P. Vigneron, , M. Hauchecome, O Aguerre, R. Toury, M. Airiau, R. Ramaswamy, D. Scherman, J. Crouzet, J. M. Lehn, and P. Lehn, Structural characteristics of supramolecular assemblies formed by guanidinium-cholesterol reagents for gene transfection, Proc. Natl. Acad. Sci. USA 96(6), 2621-2626(1999)

68. E. Maccioni, L. Vergani, A. Dembo, G. Mascetti, and C. Nicolini, X-ray small angle scattering study of chromatin as a fiinction of fiber length, Mol. Biol. Rep. 25(2), 73-86 (1998)

69. D. D. Dunlap, A. Maggi, M. R. Soria, and L. Monaco, Nanoscopic structure of DNA condensed for gene delivery. Nucleic Acids Res. 25(15), 3095-3101 (1997)

70. R. Golan, L. I. Pietrasanta, W. Hsieh, and, H. G. Hansma, DNA toroids: stages in condensation, B/oc/?^ww/rv 38(42), 14069-14076 (1999).

71. N. V. Hud, and K. H. Downing, Cryoelectron microscopy of lambda phage DNA condensates in vitreous ice: the fine structure of DNA toroids Proc. Natl. Acad. Sci. USA 98 (26), 14925-14930 (2001)

72. S. He, P. G. Arscott, and V. A. Bloomfield, Condensation of DNA by multivalent cations: experimental studies of condensation kinetics. Biopolymers 53(4), 329-41 (2000)

73. I. Moret, J. E. Peris, V. M. Guillem, M. Benet, F. Revert, F. Dasi, A. Crespo, and J. Alino, Stability of PEI-DNA and DOTAP-DNA complexes: effect of alkaline pH, heparin and serum, J. Controlled Release 16{\-2), 169-181 (2001)

74. A. Larsson, C. Carisson, and M. Jonsson, Characterization of the binding of YO to [poly(dA-dT)]2 and [poly(dG-dC)]2, and of the fluorescent properties of YO and YOYO complexed with the polynucleotides and double-stranded DNA. Biopolymers 36(2), 153-167 (1995)

75. F. Johansen, and J. P. Jacobsen, IH NMR studies of the bis-intercalation of a homodimeric oxazole yellow dye in DNA oligonucleotides, J. Biomol. Struct. Dyn. 16(2), 205-22 (1998)

76. S. Bemacchi, and Y. Mely, Exciton interaction in molecular beacons: a sensitive sensor for short range modifications of the nucleic acid structure Nucleic Acids Res. 29(13), e62 (2001)

77. B. Z. Packard, D. D. Toptygin, A. Komoriya, and L. Brand, Profluorescent protease substrates: Intramolecular dimers described by the exciton model, Proc. Natl. Acad. Sci. USA 93(21), 11640-11645 (1996)

78. D. Goula, C. Benoist, , S. Mantero, G. Merlo, G. Levi, and B. A. Demeneix, Polyethylenimine-based intravenous delivery of transgenes to mouse lung. Gene Ther. 5 (9), 1291-1295 (1998).

79. T. Blessing, J.-S. Remy, and J.-P. Behr, Monomolecular collapse of plasmid DNA into stable virus-like particles A ^ c . Natl. Acad Sci. USA 95(4), 1427-1431 (1998)


80. D. P. Millar, R. J. Robbins, and A. H. Zewail, Direct observation of the torsional dynamics of DNA and RNA by picosecond spectroscopy. Proc. Natl Acad. Sci. t/5^ 77(10), 5593-7(1980).

81. G. Chirico, M. Collini, K. Toth, N. Brun, and J. Langowski, Rotational dynamics of curved DNA fragments studied by fluorescence polarization anisotropy. Eur. Biophys. J. 29(8), 597-606 (2001).

82. K. Kinosita Jr, S. Kawato, and A. Ikegami, A theory of fluorescence polarization decay in membranes. Biophys J. 20(3), 289-305 (1977)

83. R. Swaminathan, U. Nath, J. B. Udgaonkar, N. Periasamy, and G. Krishnamoorthy, Motional dynamics of a buried tryptophan reveals the presence of partially structured forms during denaturation of barstar. Biochemistry 35(28), 9150-9157 (1996)

84. J. Duhamel, J. Kanyo, G. Dinter-Gottlieb, and P. Lu Fluorescence emission of ethidium bromide intercalated in defined DNA duplexes: evaluation of hydrodynamics components, Biochemistry 35(51), 16687-16697 (1996)

85. G. Krishnamoorthy, B. Roques, J-L. Darlix, and Y. Mely, DNA condensation by the nucleocapsid protein of HIV-1: a mechanism ensuring DNA protection. Nucleic Acids Res. 31(18), 5425-5432 (2003)

86. C. G. Baumann, S. B. Smith, V. A. Bloomfield, and C. Bustamante, Ionic effects on the elasticity of single DNA molecules, Proc. Natl. Acad. Sci. USA 94(12), 6185-6190(1997)

87. O. Boussif, F. Lezoualc'h, M. A. Zanta, M. D. Mergny, D. Scherman, B. Demeneix, and J. p. Behr, A versatile vector for gene and oligonucleotide transfer into cells in culture and in vivo: polyethylenimine, Proc. Natl. Acad. Sci. USA 92(16), 7297-7310(1995).

88. L. S. Lerman, A transition to a compact form of DNA in polymer solutions Proc. Natl. Acad Sci. USA 68(8), 1886-90 (1971)

89. K. Yoshikawa, Controlling the higher-order structure of giant DNA molecules. Adv. DrugDeliv. Rev. 52(3), 235-44 (2001)

90. V. V. Vasilevskaya, A. R. Khokhlov, Y. Matsuzawa, and K. Yoshikawa, Collapse of single DNA molecule in poly (ethylene glycol) solution, J. Chem. Phys. 102 (16), 6595-6602(1995).

91. Y. Mely, H. d. Rocquigny, N. Morellet, B. P. Roques, and D. Gerad. Zinc binding to the HIV-1 nucleocapsid protein: a thermodynamic investigation by fluorescence spectroscopy, Biochemistry, 35(16), 5175-82 (1996)

92. J. M. Coffm, Structure of the retroviral genome. In RNA Tumor Viruses. Edited by R. Weiss, N. Teich, H. Varmus, and J. Coffm, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, (1984) pp. 261-368.

93. J-. L. Darlix, M. Lapadat-Tapolsky, H. d. Rocquigny, and B. P. Roques, First glimpses at structure-function relationships of thenucleocapsid protein of retroviruses, J. Mol. Biol. 254(4), 523-537 (1995).

94. V. Tanchou, C. Gabus, V. Rogemond, and J-. L. Darlix, Formation of stable and functional HIV-1 nucleoprotein complexes in vitro, J. Mol.Biol. 252(5), 563-571 (1995).

95. C. Barat, V. Lullien, O. Schatz, G. Keith, M.T. Nugeyre, and F. Gruninger-Leitch, HIV-1 reverse transcriptase specifically interacts with the anticodon domain of its cognate primer tRNA. EMBOJ. 8(11), 3279-3285 (1989)

96. M. Lapadat-Tapolsky, C. Pemelle, C. Borie, and J-. L. Darlix, Analysis of the nucleic acid annealing activities of nucleocapsid protein from HIV-1, Nucleic Acids Res. 23(13), 2434-2441 (1995)

338 f^, NAG ET.AL.

97. B. Allain, M. Lapadat-Tapolsky, C. Berlioz, and J. L. Darlix, Transactivation of the minus-strand DNA transfer by nucleocapsid protein during reverse transcription of the retroviral genome, EMBOJ. 13(4), 973-981(1994)

98. J.C. You, and C. S. McHenry, Human immunodeficiency virus nucleocapsid protein accelerates strand transfer of the terminally redundant sequences involved in reverse transcription, J. Biol. Chem. 269(50), 31491-31495 (1994)

99. M. Lapadat-Tapolsky, H. De Rocquigny, D. Van Gent, B Roques, R. Plasterk, and J. L. Darlix, Interactions between HIV-1 nucleocapsid protein and viral DNA may have important functions in the viral life cycle. Nucleic Acids Res. 21(4), 831-839 (1993).

100. M. A. Urbaneja, M. Wu, J. R. Casas-Finet, and R. L. Karpel, HIV-1 nucleocapsid protein as a nucleic acid chaperone: spectroscopic study of its helix-destabilizing properties, structural binding specificity and annealing activity, J. Mol. Biol 318(3), 749-764 (2002)

101. A. H. Maki, A. Ozarowski, A. Misra, M. A. Urbaneja, and J. R. Casas-Finet, Phosphorescence and optically detected magnetic resonance of HIV-1 nucleocapsid protein complexes with stem-loop sequences of the genomic Psi-recognition element. Biochemistry 40(5), 1403-1412 (2001)

102. J. S. Buckman, W. J. Bosche, and R. J. Gorelick, Human immunodeficiency virus type 1 nucleocapsid zinc fingers are required for efficient reverse transcription, initial integration process and protection of newly synthesized viral DNA, J. Virol, 77(2), 1469-1480(2003).

103. K. Gao, R. J. Gorelick, D. G. Johnson, and F. Bushman, Cofactors for human immunodeficiency virus type 1 cDNA integration in vitro, J. Virol. 77(2), 1598-1603 (2003)

104. C. J. Norbury, and B. Zhivotovsky, DNA damage-induced apoptosis. Oncogene 23(16), 2797-808 (2004)

105. J. Z. Parrish, and D. Xue, Functional genomic analysis of apoptotic DNA degradation in C. elegans, Mol. Cell 11(4), 987-96 (2003)

106. J. Saltiel, Perdeuteriostilbene. The Role of Phantom States in the cis-trans Photoisomerizationof Stilbenes,y. Am. Chem. Soc. 89(4), 1036-1037 (1967)

107. J. Saltiel, N. Tarkalanov, and D. F. Sears, Conformer-specific adiabatic cis .fwdarw. trans photoisomerization of cis-1-(2-Naphthyl)-2-phenylethene. A striking application of the nEER principle, J. Am. Chem. Soc 117(20), 5586 - 5587 (1995).

108. S. Moharikar, J. S. D'souza, A. B. Kulkami, and B J Rao, UV-C induced apoptosis cell death in the unicellular chlorophyte Chamydomonas reinhaedtii (manuscript under preparation)

109. A. Srivastava, and G. Krishnamoorthy Cell type and spatial location dependence of cytoplasmic viscosity measured by time-resolved fluorescence microscopy. Arch Biochem. Biophys. 340(2), 159-67 (1997)

110. A. I. Roca, and M. M. Cox, RecA protein: structure, function, and role in recombinational DNA repair. Prog. Nucleic Acid Res. Mol. Biol. 56, 129-223 (1997).

111. J. I. Kim, M. M. Cox, and R. B. Inman, On the role of ATP hydrolysis in RecA protein-mediated DNA strand exchange. I. Bypassing a short heterologous insert in one DNA substrate. J. Biol. Chem. 267(23), 16438-16443 (1992).

112. W. Rosselli, and A. Stasiak, The ATPase activity of RecA is needed to push the DNA strand exchange through heterologous regions, Embo J. 10(13), 4391-4396 (1991).

113.T. Nishinaka, A. Shinohara, Y. Ito, S. Yokoyama, and T. Shibata, Base-pair switching by interconversion of sugar puckers in DNA extended by proteins of


RecA-family: a model for homology search in homologous genetic recombination. Proc. Natl Acad. ScL USA 95(19), 11071-6 (1998)

114. W. C. Lam, E. J. Van der Schans, L. C. Sowers, and D. P. Millar, Interaction of DNA polymerase I (Klenow fragment) with DNA substrates containing extrahelical bases: implications for proofreading of frameshift errors during DNA synthesis. Biochemistry 38(9), 2661-2668 (1999).

115. E. L. Rachofsky, R. Osman, and J. B. Ross, Probing structure and dynamics of DNA with 2-aminopurine: effects of local environment on fluorescence.. Biochemistry 40(4), 946-56 (2001)

116. R. A. Hochstrasser, T. E. Graver, L. C. Sowers, and D. P. Millar, Melting of a DNA helix terminus within the active site of a DNA polymerase, Biochemistry 33(39), 11971-9 (1999).

117. E. H. Egelman, and A. Stasiak, Structure of helical RecA-DNA complexes, Complexes formed in the presence of ATP-gamma-S or ATP, J. Mol Biol 191(4), 677-97 (1986)

118. T. Nishinaka, Y. Ito, S. Yokoyama, and T. Shibata, An extended DNA structure through deoxyribose-base stacking induced by RecA protein, Proc. Natl Acad. Sc.i USA 94(13), 6623-8 (1997)

119. J. M. Cox, O. V. Tsodikov, and M. M. Cox, Organized unidirectional waves of ATP hydrolysis within a RecA filament, PLoS. Biol 3(2), e52 (2005).

120. K. J. MacFarland, Q. Shan, R. B. Inman, and M. M. Cox, RecA as a motor protein. Testing models for the role of ATP hydrolysis in DNA strand exchange, J. Biol Chem. 272(28), 17675-85(1997).

121. E. R. Goedken, S. L. Kazmirski, G. D. Bowman, M. O'Donnell, and J. Kuriyan, Mapping the interaction of DNA with the Escherichia coli DNA polymerase clamp loader complex. Nat. Struct. Mol Biol 12(2), 183-90 (2005)

122. J. C. Morales, and E. T. Kool, Importance of terminal base pair hydrogen-bonding in 3'-end proofreading by the Klenow fragment of DNA polymerase I, Biochemistry 39(10), 2626-32 (2000)

123. K. Sengupta, and B. J. Rao, Translin binding to DNA: recruitment through DNA ends and consequent conformational transitions, Biochemistry 41(51), 15315-26 (2002)

124. N. Haruta, X. Yu, S. Yang, E. H. Egelman, and M. M. Cox, A DNA pairing-enhanced conformation of bacterial RecA proteins, J. Bio.I Chem. 278(52), 52710-23 (2003).

125. C. Cotta-Ramusino, D. Fachinetti, C. Lucca, Y. Doksani, M. Lopes, J. Sogo, and M. Foiani, Exol processes stalled replication forks and counteracts fork reversal in checkpoint-defective cells, Mol Cell 17(1), 153-9 (2005)

126. S. L. Lam, and L. N. Ip, Low temperature solution structures and base pair stacking of double helical d(CGTACG)(2), J. Biomo.l Struct Dyn. 19(5), 907-17 (2002)

127. L. B. Bloom, M. R. Otto, J. M. Beechem, and M. F. Goodman: Influence of 5'-nearest neighbors on the insertion kinetics of the fluorescent nucleotide analog 2-aminopurine by Klenow fragment, Biochemistry 32(41), 11247-11258 (1993)

128. P. Modrich, and R. Lahue, Mismatch repair in replication fidelity, genetic recombination and cancer biology, ««w. Rev. Biochem. 65, 101-133 (1996)

129. B. D. Harfe, and S. Jinks-Robertson, DNA mismatch repair and genetic instability, Annu. Rev. Genet. 34, 359-399 (2000)

130. D. T. Minnick, and T. A. Kunkel, DNA synthesis error, mutator and cancer, Cancer 5wrvey 28, 3-20 (1996)

131. S. S. Su, R. S. Lahuae, K. G. Au, and P. Modrich, Mispair specificity of methyl directed DNA mismatch correction in vitro, J. Biol Chem. 263(14), 6829-6835 (1988)

340 N.NAG^j.^x.

132. S. Gradia, S. Acharya, and R. Fishel, The human mismatch recognition complex hMSH2'-hMSH6 functions as a novel molecular switch, Cell 91(7), 995-1005 (1997).

133. D. L. Cooper, R. S. Lahue, and P. Modrich. Methyl directed mismatch repair is bidirectional, J. Biol Chem. 268(16), 11823-11829 (1993)

134. M. J. Schofield, and P. Hsieh. DNA Mismatch Repair: Molecular Mechanisms and Biological Function, ^««w. Rev. Microbiol 57, 579-608 (2003)

135. A. B. Buermeyer, S. M. Deschenes, S. M. Baker, and R. M. Liskay. MammaUan DNA mismatch repair, Annu. Rev. Genet. 33, 533-564 (1999).

136. W. Yang, Structure and function of mismatch repair proteins, Mutat. Res. 460(3-4), 245-256 (2000)

137. K. P. Hopfner, and J. A. Tainer, DNA mismatch repair: the hands of a genome guardian. Structure with Folding and Design. 8(12), R237-41 (2000)

138. K. P. Bjomson, L. J. Blackwell, S. Harvey, C. Batinger, D. Allen, and P. Modrich, Assembly and molecular activities of MutS tetramer, J. Biol. Chem. 278(36), 34667-34673 (2003)

139. S. S. Su, and Modrich, P. Escherichia coli mutS-encoded protein binds to Mismatched DNA base pairs, Proc. Natl Acad ScL USA 83 (14), 5057-5061 (1986)

140. L. T. Haber, and G. C. Walker, Altering the conserved nucleotide binding motif in the Salmonellatyphimurium MutS mismatch repair protein affects both its ATPase and mismatch binding activities, EMBOJ. 10(9), 2707-2715 (1991)

141. S. S. Su, R. S. Lahue, K. G. Au, and P. Modrich, Mispair specificity of methyl-directed DNA mismatch correction in vitro, J. Biol. Chem. 263(14), 6829- 6835 (1988).

142. B. O. Parker, and M. G. Marinus, Repair of DNA heteroduplexes containing small heterologous sequences in Escherichia coli. Proc. Natl. Acad. Sci. USA 89 (5), 1730-1734(1992)

143. M. H. Lamers, A. Perrakis , J. H. Enzlin, H. H. K. Winterwerp, N. Wind, and T. K. Sixma, The crystal structure of DNA mismatch repair protein MutS binding to G.T Mismatch, Nature 407(6805), 711-717 (2000)

144. M. J. Schofield, F. E. Brownewell, S. Nayak, C. Du, E. T. Kool, and P. Hsieh, The Phe-X-Glu DNA binding motif of MutS. The role of hydrogen bonding in mismatch recognition, J. Biol Chem. 276(49), 45505-8 (2001)

145. R Lavery, and H. Sklenar, Defining the structure of irregular nucleic acids: conventions and principles, J. Biomol Struct. Dyn. 6 (4), 655-667 (1989).

146. H. Wang, Y. Yang, M. J. Schofield, C. Du, Y. Fridman, S. D. Lee, E. D. Larson, J. T. Drummond, E. Alani, P. Hsieh, and D. Erie, DNA bending and unbending by MutS govern mismatch recognition and specificity, Proc. Natl. Acad. ScL USA 100(25), 14822-14827(2003).

PROTEIN-PROTEIN INTERACTIONS IN VIVO: USE OF BIOSENSORS BASED ON FRET

Jan Willem Borst ' , Isabella Nougalli-Tonaco^ Mark A. Hink ' , Arie van Hoek^" , Richard G.H. Immink'* and Antonie J.W.G.

1,2,5 Visser

15.1. INTRODUCTION

Sensing of molecules in living cells is a rapidly evolving discipline in modem biological and biomedical research. Biosensors make use of biological components to sense a molecule of interest. In the food, health care and pharmaceutical industry the biosensor technology is the prime developing target for the future. Optical biosensors are ideal candidates for industrial use because of the high specificity, selectivity and adaptability (Terry et al., 1995). A typical example is the application of disposable blood glucose biosensors, in which blood sugar levels of diabetic patients are monitored in real time (Wilson and Gifford, 2005). Biosensors used in cell biology generally imply real-time monitoring of molecular behavior activities or interactions of molecules in living cells (Bunt and Wouters, 2004). Many of the biosensors used nowadays are based on the Forster Resonance Energy Transfer FRET methodology. FRET is a fluorescence technique based on dipolar interactions between different donor and acceptor molecules making this technique very sensitive to intermolecular distances (Clegg, 1996; Wouters and Bastiaens, 1999; Bemey and Danuser, 2003; Jares-Erijman and Jovin, 2003; Gadella et al , 1999; Lakowicz, 1999). The main principles of FRET have been summarized in Table 15.1.

^ MicroSpectroscopy Centre, Wageningen University, Dreijenlaan 3, 6703 HA Wageningen, The Netherlands

^ Laboratory of Biochemistry, Wageningen University, Dreijenlaan 3, 6703 HA Wageningen, The Netherlands

^ Plant Research International, Bioscience, Bomsesteeg 65, 6708 PD Wageningen ^ Laboratory for Biophysics, Wageningen University, Dreijenlaan 3, 6703 HA Wageningen, The

Netherlands ^ Department of Structural Biology, Faculty of Earth and Life Sciences, Vrije Universiteit, 1081 HV

Amsterdam, The Netherlands

341

342 BORST ETAL

Table 15.1: FRET Theory

The fluorescence lifetime provides a time window for the detection of dynamic processes Dynamic processes that compete with fluorescence: quenching by external molecules resonance energy transfer (rate constant kr) The fluorescence lifetime of the donor becomes shorter in case of RET.

Donor

R ;; ;::; ^1

s

1

^ k 1

•• _ 1 .

Acceptor

* • ; • • * -

I

k^-\- k^ Becomes r 1

-k,+k„, + kr

Summary of FRET parameters

Ro = 0 . 2 1 1 ( K ' * n ' * Q D * J ) " ' (in A)

2 2 K = dipole orientation factor (0 < K < 4)

2

(for random and rapid orientations: K =2/3) Qp = quantum yield of the donor (QCFP = 0.39) J = spectral overlap between donor emission and acceptor absorbance spectra (J= 1.55479 X 10'' nm^M-lcm-l (CFP-YFP) n = refractive index of the sample (n = 1.4).

Em

H F M f?o = critical transfer distance, for CFP-YFP R^ = 5 0 nm

Wavelength, nm

The actual distance R is obtained from the transfer efficiency E - Ro / (Ro +R ) Example in case of donor CFP and acceptor YFP TD = 2.6 ns (no FR£T)rz)A "" 1.3 ns (FRET) E = 0.5 => /? = /?o = 5.0 nm => actual and critical distances are the same

New developments of optical biosensors in combination with fluorescence microscopy give insight in the responses of molecular components in the cell such as receptors, proteins or ligands, which are activated in a dynamic network. Classical biochemical approaches mostly referred to as in vitro experiments, have been proven to be valuable for identification and physical isolation of proteins. The cellular localization of proteins can be visualized using immunofluorescence techniques. Co-immuno-precipation and affinity chromatography are classical biochemical techniques for purification of protein complexes. Live cell imaging provides the time as an additional dimension and has evolved as an important approach in cell biology to monitor cell dynamics (Lippincott-Schwartz et al., 2001). Nowadays localization and

INTERACTIONS IN VIVO: USE OF BIOSENSORS BASED ON FRET 343

dynamics can be connected with system properties like active state, mobility, electric field, oxygen, calcium or proton fluxes to address different biological questions (Wouters et al., 2001). To visualize and quantify proteins of interest, advanced spectroscopic techniques are combined with microscopy and biosensors, so that specific molecular information of cells can be retrieved (Bunt and Wouters, 2004). The application of the green fluorescent protein GFP technology has been of crucial importance in the imaging of cellular proteins (Tsien, 1998; Matz et al., 1999). GFP can be coupled to the protein of interest via genetic approaches. After transformation or transfection of eukaryotic cells its fluorescence can be detected within several hours. Since the discovery of GFP, a variety of colored fluorescent proteins was developed by mutagenesis. Fluorescent proteins emitting colors from violet to red are currently available (Verkhusha and Lukyanov, 2004; Shaner et al., 2004) and allow to monitor simultaneously multiple proteins.

The question, whether the physical state of a receptor or protein is monomeric, dimeric, or multimeric can be addressed and answered by several different approaches. The phage display method and/or yeast two-hybrid system are the most significant examples of these techniques (Burch et al., 2004; Causier and Davies, 2002). However, these methods have the disadvantage that they lack spatial information. The introduction of high-resolution confocal microscopy gave the opportunity to investigate the co-expression of different proteins in their natural environment. The optical resolution of a microscope allows detection at sub-cellular level, but physical molecular interactions between proteins or receptors on nanometer scale cannot be visualized. The resolution of a typical confocal or wide-field image is diffraction limited, which means that, for example, the use of excitation light of 488 nm will result in an optical resolution around 220 nm (Herman, 1998). One possibility to go beyond the optical limitations is to apply FRET microscopy. FRET microscopy can be monitored using different optical methods, which will be discussed in this chapter. First, the FRET combinations in cell biology and different FRET biosensors will briefly reviewed. Then, the fluorescence intensity based approaches will be discussed. Subsequently, the methodology to observe FRET with fluorescence lifetime imaging microscopy FLIM will be addressed in more detail. Plant transcription factors labeled with enhanced Cyan Fluorescent protein ECFP and enhanced Yellow Fluorescent protein EYFP in plant cells will serve as an example.

15.2. FRET COMBINATIONS IN CELL BIOLOGY

Which FRET pair is the best choice? The most attractive method to add a fluorescent label is by genetic approaches (Hink et al., 2002). VisuaHzation of proteins of interest in plant or animal cells requires in vivo labeling using different fluorophore combinations. The enhanced forms of CFP and YFP are most commonly used in cell biology. The main reasons for that are the attractive spectroscopic characteristics of this FRET pair: i the spectral overlap between the donor emission and acceptor absorption is large ii the donor

344 BORST ETAL

fluorescence lifetime ECFP is relatively short T= 2.5 ns and iii the molar extinction coefficient of the acceptor EYFP is high. The resulting Forster radius is large Ro= 5 nm see Box 1, Recently, a monomeric and improved version of ECFP mCerulean has been developed, which is 2.5 times brighter than ECFP (Rizzo et al., 2004). In combination with the improved YFP version called 'Venus' mCerulean-Venus is a promising FRET pair that can be used for imaging of protein interactions in cell biology. Another frequently used combination is GFP and the Discosoma red fluorescent protein, DsRED, or related red fluorescent protein RFP. The disadvantage of DsRED as an acceptor is the maturation time, which is four times longer than for GFP. Furthermore, DsRED can tetramerize and it has a 'green' absorption band, which overlaps with the absorption band of GFP. Since 2002 several different red fluorescent proteins, dimer and monomer versions of DsRED have been cloned (Campbell et al., 2002; Zhang et al., 2002; Shaner et al., 2004). Recently, FLIM data of GFP tagged protein in combination with monomeric RFP mRFP have been presented (Peter et al., 2005). A clear interaction between a chemokine receptor and protein kinase C in carcinoma cells was found after adding a stimulus to the cells resulting in a strong reduction of the fluorescence lifetime of the GFP tagged protein (Peter et al., 2005).

15.3. FRET SENSORS

There are many examples how FRET can be used in current biological research. For example, a new displacement hybridization method is reported using a FRET biosensor to detect double stranded nucleic acid targets with Hoechst 33258 and Oregon Green488 as donor-acceptor pair (Ho and Hall, 2004). The visible fluorescent proteins appear to be useful for the detection of different processes that employ the intrinsic properties of the fluorescent protein itself It is known that GFP is sensitive to pH changes (Haupts et al., 1998; Kneen et al., 1998). The protonated form of GFP absorbs light at 400 nm and is hardly fluorescent upon 480-500 nm excitation, so that it can act as a pH sensor. Different other sensors based on GFP technology were developed, but will not be discussed in this review. Three FRET based sensors will be briefly described here below.

15.3.1 Cameleons (Yearn)

Fluorescent indicators for in vivo calcium imaging were developed (Miyawaki et al., 1997). Cameleons are genetically encoded fluorescent calcium indicators without cofactors and are targetable to specific intracellular locations (Miyawaki et al., 1999; Miyawaki et al., 1997). Calcium is a very important ion in cellular signaling, since it can act as an essential second messenger in signal transduction cascades. Ca^^ signals are found in the cytosol and different organelles, but these signals are often hard to measure. To overcome this problem a specific calcium sensor was developed consisting of tandem fusions of a CFP, calmodulin having four calcium binding sites, a


calmodulin binding peptide M13 and YFP (Miyawaki et al., 1997). The level of intramolecular FRET is dependent on the amount of Ca^^ binding to the calmodulin. Binding of Ca^^ makes calmodulin wrap around the Ml3 domain upon which an increase of the FRET signal will be observed. The dynamic range for Ca^^ concentrations is from 10' to 10' M. Several variants of cameleons are currently available and are continuously improved. The latest versions are pH independent and consist of the brighter EYFP variants Venus. Recently, also a biosynthetic indicator of endoplasmic reticulum in endocrine cells of mouse pancreas has been described based on the cameleon3.3er YC3.3er allowing real-time visualization of cell signaling in living tissues (Kara et al., 2004).

15.3.2 Caspase sensor

One of the cellular processes, which is involved in a variety of diseases, is apoptosis or programmed cell death (Mahajan et al., 1999). A class of cysteine proteases known as caspases is involved in the apoptosis process. After activation by apoptotic signals the caspases digest cellular proteins and induce cell death. These proteases exist in many isoforms and can recognize specific amino acid motifs in their target proteins (Morgan and Thorbum, 2001). A caspase sensor based on FRET was developed for in vivo visualization and quantification of caspase activity (Xu et al., 1998). Like the cameleon the caspase sensor consists of a tandem of cyan and yellow fluorescent protein coupled via a linker between 12 and 20 amino acids. There are several isoforms of the caspases, which recognize specific amino acid sequences. Xu and colleagues have demonstrated caspase 3 activity in vivo using this FRET sensor (Xu et al., 1998).

15.3.3 FLAME

Fluorescent biosensors like the cameleons are also applied for detecting intracellular phosphorylation states in which a conformational change will alter the energy transfer efficiency. Recently, a new FRET sensor the so-called FLAME fluorescent linked autophosphorylation monitor for EGFR, was developed by Bastiaens and colleagues. This FRET sensor monitors the in vivo phosphorylation state of the epidermal growth factor receptor EGFR signaling system (Offterdinger et a l , 2004). EGFR initiation occurs by ligand EGF binding followed by homodimerization and rapid receptor autophosphorylation. To monitor EGFR phosphorylation the translocation and binding of phosphotyrosine-binding domain PTB labeled with enhanced yellow fluorescent protein EYFP to enhanced cyan fluorescent protein ECFP-tagged EGFR was measured. Expression of this FLAME sensor in C0S7 cells demonstrated rapid and reversible changes in the EYFP/ECFP fluorescence intensity ratios, due to binding of the PTB domain to its consensus-binding site upon phosphorylation at the cell periphery. The results show that this sensor closely approaches the true dynamics of tyrosine kinase autophosphorylation and dephosphorylation (Offterdinger et al., 2004).

346 BORST ETAL

15.4. INTENSITY BASED FRET IMAGING

15.4.1 Confocal and wide-field FRET imaging

The combination of optical microscopy with FRET spectroscopy provides quantative, temporal and spatial information about binding and interaction of proteins in vivo (Herman, 1998; Wallrabe et al., 2003). FRET can be quantified using steady state or time-resolved fluorescence techniques. In the steady-state approach the fluorescence intensities of donor and acceptor are monitored in a fluorescence microscope by using either confocal or wide-field illumination and detection. This steady-state approach relies on the observation that the fluorescence intensity of the donor is reduced and the acceptor fluorescence is enhanced when energy transfer takes place. The disadvantage of this approach is that the signals are highly dependent on the concentrations of donor and acceptor molecules. Photobleaching needs to be avoided, because it alters the effective donor and acceptor concentrations. Artefacts like bleed-through of the donor fluorescence in the acceptor detection channel and direct excitation of the acceptor also need to be taken into account. In several publications corrected FRET imaging methods have been demonstrated and optimized (Gordon et al., 1998; Xia and Liu, 2001).

15.4.2 Spectral imaging

Another approach is to couple a spectrograph to the microscope for determining emission spectra of microscopic objects. Spectral resolution of fluorescence can provide the necessary imaging contrast when the spectral profile changes due to modulation of energy transfer efficiency. Based on sensitized acceptor emission and donor fluorescence quenching, spectral images can be analyzed to quantify the donor-acceptor energy transfer process within a cell (Gadella et al., 1999; Raicu et al., 2005).

15.4.3 Acceptor photo-bleaching

Another method in which FRET can be used to examine intracellular molecular interactions between proteins is acceptor photobleaching APB. This method has been applied to several studies (Kenworthy, 2001) and are not only restricted to intensity-based microscopic applications, but also to fluorescence lifetime imaging microscopy FLIM (Wouters and Bastiaens, 1999). APB is often used as a method to prove the occurrence of FRET. APB measurements in a confocal microscope have been critically assessed by Karpova and colleagues (Karpova et al., 2003). These APB experiments make use of a confocal laser scanning microscope, because specific laser lines are not only able to selectively excite the fluorescent dyes, but can also specifically bleach the dye of interest. The fluorescence intensities of donor and acceptor are determined before and after applying a strong laser pulse that photobleaches the acceptor fluorophore irreversibly. When donor and acceptor molecules interact, the destruction of the acceptor fluorophore will result in increased


430 nm 485 nm

Images before bleach S ^ Images after bleach Bleach pulse

514 nm A B Time(s)

Figure 15.1. Schematic illustration of the acceptor photobleaching principle (A). Fluorescence intensity images are taken before and after a strong bleach pulse, which irreversibly destroys the acceptor molecule (YFP). The integrated fluorescence intensities of donor (CFP) and acceptor (YFP) fluorescence are measured and plotted in a graph (B).

donor fluorescence, since the energy cannot be transferred to the acceptor molecule Fig. 15.1. The energy transfer efficiency E before the bleaching can be calculated according to:

E — IDA-IDR/I DA-ADB/ADA (1)

where IDB is the fluorescence intensity before bleaching and IDA is the fluorescence intensity after bleaching.

The measurement of FRET efficiency by the acceptor-photobleaching approach requires several checkpoints. First, selective bleaching of the acceptor is required, because bleaching of the donor will result in underestimation of donor dequenching (Kenworthy, 2001; Voss et al., 2005). Second, APB is still an intensity-based approach and therefore it is sensitive for donor and acceptor concentrations. By using a FRET indicator where only one CFP and one YFP molecule are connected by a linker this problem can be avoided. A general remark about the visualization and FRET determination of fluorescent proteins is that the use of strong constitutive promoters is often required. Then high expression levels of the fluorescent proteins can be reached and as a result the localization can be altered and large protein aggregates might appear. Furthermore, fluorescent proteins have the tendency to form dimers at high expression levels, although the proteins of interest are not necessarily related. It has been demonstrated that the main amino acid responsible for this "sticky" behavior of the visible fluorescent proteins is alanine 206. Replacement of this amino acid for a lysine resulted in truely monomeric fluorescent proteins (Zacharias et al., 2002). APB experiments are often performed in fixed samples. In hving cells a rapid redistribution of donor and acceptor- molecules may give rise to fast recovery and FRET is then difficult to determine quantitatively. APB experiments can also be combined with FLIM measurements determining the fluorescence lifetime of the donor. Interaction between a donor and an acceptor molecule will result in a faster

348 BORST^r^i:

fluorescence decay of the donor molecule, but destruction of the acceptor molecule will reverse the quenching effect and the initial fluorescence lifetime of the donor without acceptor is obtained (Bastiaens and Squire, 1999).

15.5. FLUORESCENCE LIFETIME IMAGING MICROSCOPY (FLIM)

The fluorescence lifetime of a molecule is dependent on local environmental factors such as the presence of Ca^ , effects of pH or refractive index (Suhling et al., 2002; Borst et al., 2005) or FRET. In general, the fluorescence lifetime is determined by interplay of radiative and non-radiative decay rates. The radiative decay rate is considered to be constant for a given fluorophore, while the non-radiative decay rate can vary with the environment. Most FLIM measurements for determining FRET are based on the effect that an acceptor molecule specifically interacts with a donor molecule resulting in quenching of donor fluorescence. This donor fluorescence quenching will result in the reduction of the fluorescence lifetime since energy transfer will introduce an additional relaxation path from the excited state to the ground state see Box 1. The amount of reduction is directly correlated with the FRET efficiency E via:

E = 1 - TDA/XD (2)

where TDA is the fluorescence lifetime of the donor in the presence of acceptor and TD is the fluorescence lifetime of the donor alone.

FLIM combines fluorescence lifetime measurements with microscopic resolution resulting in spatially resolved fluorescence lifetimes. A major advantage of this approach is that the results are relatively insensitive to the local fluorophore concentration, spectral cross talk or photobleaching. Fluorescence lifetimes can be determined via time-domain or frequency-domain methods (Gratton et al., 2003; Squire and Bastiaens, 1999; Gadella et al., 1999). In short, with the frequency-domain approach the sample is illuminated with sinusoidally modulated light at frequencies typically between 20 MHz and 1 GHz. As a result the fluorescence is also sinusoidally modulated and the fluorescence lifetime is calculated from the phase shift and demodulation relative to these parameters of the excitation light (Lakowicz, 1999; Gratton et al., 2003; Squire and Bastiaens, 1999). Frequency domain FLIM measurements consist of successive recording of several images to determine spatially resolved apparent lifetimes pixel by pixel from the phase shift between excitation and emission demodulation. The time-domain method uses ultrashort excitation pulses down to hundreds of femtoseconds upon which the fluorescence decay can be measured (Hink et al., 2002; Gerritsen and De Grauw, 1999). Nowadays commercially available photon counting electronic boards are widely used and integrated with a confocal microscope system Becker et al., 2004. The board uses the time-correlated single photon counting TCSPC principle to record the arrival time of the emitted photon in


the detector with respect to the laser pulse. Another time-domain configuration is the so-called lifetime module LIMO system in which the photons are collected in four variable time windows. From the relative content of the four time gates the fluorescence lifetime can be retrieved (Gerritsen et al., 2002)

15.5.1 FLIM setup

A schematic description of the instrumental set up and how TCSPC is incorporated in the microscope set up is given in Fig. 15.2.

The here presented experiments are performed on a multi-photon dedicated Biorad Radiance 2100 MP system. The fluorescence is detected in a non-descanned manner and FLIM measurements are performed by directing the fluorescence photons via a dichroic filter 670UVDCLP onto a Hamamatsu R3809U photomultiplier tube PMT. ECFP fluorescence is selected using a 480DF30 bandpass filter. The microcharmel-plate PMT allows single-photon detection at high time-resolution 50 ps. The output of the PMT is coupled to a Becker & Hickl single photon counting module SPC 830, as the start signal the single-photon timing process. The pulses from the Ti-Sapphire laser repetition rate 76 MHz serve as the SYNC or stop signal for the single-photon timing process. From the statistics of the time of arrival of fluorescence photons relative to the laser pulses the decay profile is built up. The experimental time window was divided into 64 channels and fluorescence was typically recorded for 90 seconds at a photon count rate of approximately 20 kHz. The pixel clock and line pre-divider signals from the Biorad scanhead were used to direct the single-photon timing data in appropriate memory blocks to create 2D lifetime images.

15.5.2 FLIM analysis

The FLIM data were obtained using the Becker & Hickl SPC830 acquisition card. The SPCImage 2.8 software package was used for data analysis. The software enables to fit the fluorescence decay data of every pixel in the image to a mono- or bi-exponential decay model. When in the analysis of the fluorescence data of a FRET system two decay time components are used, one fraction can directly denote the contribution of the non-interacting species donor fluorescence lifetime alone and the other fraction with a shorter lifetime component may reflect the presence of interacting molecules.

350 BORST ETAL

Confocal and FLIM Image

TCSPC Module

X-Y scan (64x64 pixels)!

Detector

Confocal Image i

Scan Headl

Ti:Sa laser 76 MHz. 150 fs

X-Y scan (64x64 pixels)

Fluoresence Lifetime Image (64x64 pixels)

u 10

c

FluiitMCcnci Decay

f^ ^"'*^'"*^«*1^

3 4 6 8 10 13

Thntdis)

Figure 15.2. Schematic illustration of a two-photon microscope in combination with a setup for measuring fluorescence lifetimes. Besides the conventional two-photon excited confocal imaging, the microscope can also collect the photons in a counting mode, while the scanhead of the microscope is synchronized with the TCSPC acquisition card.

The analysis software has the possibility to choose a pixel binning factor. The binning occurs according to the relationship p = (2b+l)^ where p is the number of pixels and b is the binning factor. All presented data were analyzed with binning b=l and comparison with true 1-pixel analysis b=0 resulted in the same fit parameters. The image resolution was set at 64 x 64 pixels. The increase of the number of pixels would not enhance the resolution, because with a 60x water immersive objective lens the size of the x,y pixels reaches already the diffraction limit 250 - 400 nm.

15.6. APPLICATIONS WITH PLANT TRANSCRIPTION FACTORS

15.6.1 Sub-cellular localization via confocal microscopy

In this paragraph an application of FRET microscopy in combination with FLIM will be described. The aim is the study of MADS box transcription factor interactions in living plant cells. MADS box genes encode for transcription factors involved in many developmental process in flowering plants, most notably in the determination of floral meristem and floral organ identity (Ferrario et al., 2004; Riechmann and Meyerowitz, 1997). This family of transcription factors acts in a complex network of physical protein-protein and protein-DNA interactions either as homo- or heterodimers, and are, most likely, able to form higher order complexes (Egea-Cortines et al., 1999; Honma and


Goto, 2001). To study this network of interactions, the use of the yeast two-hybrid system has become a powerful tool to get a first impression about the ability of the MADS box proteins to form specific dimers and/or higher-order complexes. Recently, comprehensive matrix-based screens for petunia and Arabidopsis MADS box transcription factor interactions revealed the ability of these factors to form specific homo- and heterodimers (Immink et al., 2003; de Folter et al., 2005). Moreover, these interactions are conserved between different plant species (Immink et al., 2002). Despite the fact that many studies have been done in yeast only little information is known about how these interactions occur in planta. In this work, different petunia MADS box proteins, the FLORAL BINDING PROTEINS FBP involved in ovule development were used for the study of protein-protein interactions in living cells. First, the genetic fluorescent labels ECFP and EYFP were cloned at the C-terminal side of the different FBPs. C-terminal fusions to various Arabidopsis MADS box proteins appeared to give no loss in biological activity of these proteins Angenent and Urbanus, unpublished. Subsequently, different combinations of the MADS box proteins were transfected in cowpea leaf protoplasts, which transfection has been described previously (Russinova et al., 2004).

The localization of the FBP proteins FBP 11, FBP2 and FBP24 has been described by Nougalli and colleagues (NougaUi-Tonaco et al., 2006). Differences in sub-cellular localization were observed among the proteins. Remarkably, the single transfected constructs FBP2 and FBP24 were localized in the nucleus, whereas FBPl 1 stays in the c3^oplasm, probably because of its inability to homodimerize, which seems to be a prerequisite for movement into the nucleus (Immink et al., 2002). Co-transfections of FBP2 and FBP24, as well as FBP 11 and FBP24, were performed and nuclear co-localization was observed in both cases. Considering that FBP 11 is expressed in the cytoplasm by its own, and taking into account the hypothesis that dimerization is essential for transport into the nucleus, the co-localization of both proteins is already a preliminary indication of the formation of dimers. Co-localization is a good indicator for interacting proteins, but, on the other hand, FRET-FLIM measurements can unambiguously prove the existence of physical interactions between the different FBPs.

15.6.2 Molecular interaction imaging via FRET-FLIM

Various techniques to measure FRET have been applied in animal (Bastiaens and Pepperkok, 2000) and in plant cells (Hink et al., 2002; Russinova et al., 2004). We have used FLIM by combining a two-photon excitation laser-scanning microscopy with time-correlated single photon counting as described above. First a fluorescence intensity image is obtained. Subsequently, the fluorescence decay of the donor molecule is determined for every pixel in the image. Depending on the fluorescent signal of the donor, an experiment typically takes 90 seconds (Borst et al., 2003). FLIM measurements of the single and double co-transfected cells, as described for the confocal imaging part, were repeated and these experiments started in a specific order. In protoplasts co-expressing the ECFP and EYFP tagged FBPs,

352 BORSTETAL

a protoplast was selected showing EYFP fluorescence in the epi-fluorescence mode while ECFP fluorescence is hardly visible in this configuration. If ECFP expression was observed in the 2-photon-imaging mode, this cell was chosen for a FLIM measurement. The excitation wavelength in plant cells was set to 860 nm, which mainly excites the ECFP molecules, whereas excitation of EYFP is minimal. The excitation of ECFP was also reported at 820 nm (Chen et al., 2003) in a ECFP-EYFP pair, but at this wavelength the auto-fluorescence detected with the ECFP band pass filter 480DF30 nm in plant protoplasts is higher then at 860 nm.

In Fig. If.3A, a monochromatic image represents the fluorescence intensity of the ECFP donor molecules in the absence of EYFP acceptor molecules. This is the control experiment to obtain the fluorescence lifetime of the FBP24-ECFP donor alone. The pixel where the blue crosshair is pointed is the actual fluorescence decay shown in Fig. 15.3B. The fluorescence lifetime image is calculated per pixel and displayed as a pseudo color image. In Fig. 15.3C the fluorescence Hfetime of FBP24-ECFP of 2.5 ns clearly shows up throughout the whole nucleus indicated as a blue color. The distribution of fluorescence lifetimes can be observed in Fig. 15.3D. The fluorescence lifetime was analyzed according to a mono-exponential decay fit model. For transcription factors in plant cells a fluorescence lifetime of typically 2.5 ns is found. After determination of the fluorescence lifetime of the donor FBP-ECFP alone, the analysis for all other combinations was performed according to a double exponential decay model, in which the value of the ECFP fluorescence lifetime was fixed to 2.5 ns. Upon FRET the fluorescence lifetime of the ECFP donor molecules will decrease and the amount of reduction is directly correlated with the FRET efficiency. The fluorescence lifetime images displayed in the figures represent the average lifetime values.

In Fig. 15.4 A-B the fluorescence intensity and lifetime images are presented for the co-transfection of FBP2-ECFP and FBP24-EYFP. The fluorescence intensity image clearly shows a nuclear localization. The fluorescence lifetime image indicates a significant reduction of the fluorescence lifetime (x) dark orange to green corresponds to t = 1.9 - 2.0 ns. The distribution of fluorescence lifetimes is given in the histogram in Fig. 15.4C. This distribution of fluorescence lifetimes in the case of the combination FBP2-ECFP and FBP24-EYFP was homogeneously spread around 2.0 ns resulting in a FRET efficiency of 20%. The localization of the co-transfection of FBP24-ECFP and FBPll-EYFP is also nuclear, but the distribution of fluorescence lifetimes in this combination is broader than that of the previous pair of proteins. The combination FBP24-ECFP and FBPll-EYFP suggests that there are sub-nuclear regions with T=1.9 ns and without interaction t =2.4 ns between the two FBP proteins. This effect is illustrated in Fig. 15.4E where green spots interaction are present in a 'blue' nucleus no interaction. In the histogram Fig. 15.4F a peak at 2.5 ns is seen that is absent in the histogram of Fig. 15.4C. The reciprocal combinations have been tested as well and gave the same results data not shown.


c»0** ndT^ yik:*0r^ TM^fi 8«-:*|r' tw*;-ji ftM^:*p~*~* t-jUT"*^ m . n s s ^

P^sSTf CW*»j' ''i430«l"" P

^ ^ M i

0.0 1.0 20 3.0 4.0 StO 6« 74 aO SO IftO I tO U t t

Figure 15.3. An image of a typical FLIM experiment using the Becker and Hickl analysis software. In panel A the fluorescence intensity of the nuclear localized MADS box protein, FBP24-ECFP, is shown. At the blue crosshair the fluorescence decay of the selected pixel is displayed (panel B). The fluorescence lifetimes are calculated per pixel and visualized as a pseudo color image (panels C and D). (See color insert section.)

Figure 15.4. FRET-FLIM analyses in double exponential decay model of transfected cowpea leaf protoplasts, expressing the following combinations FBP2-ECFP and FBP24-EYFP (panels A-C) and FBP24-ECFP and FBPll-EYFP (panels D-F). In panels A and D the fluorescence intensity image of the nucleus of a representative cell is shown, in panels B and E the fluorescence lifetime image of the same nucleus is shown as a pseudo color image, and in panels C and F the distribution of fluorescence lifetimes over the nucleus is presented. (See color insert section.)

354 BORST ETAL

Several studies revealed that homo- and/or heterodimerization of transcription factors have indicated their binding to specific DNA sequences (Pellegrini et al., 1995; Shore and Sharrocks, 1995). Also higher-order complex formation of MADS box transcription factors is stabilized by specific DNA binding (Egea-Cortines et al., 1999). It might be possible that the sub-nuclear regions represent places where the chromatin is available for transcription and to which the transcription factors are recruited, resulting in stabilization of the less stable or "transient" interactions (Nougalli-Tonaco et al., 2006). The FRET-FLIM data also have shown to be concentration independent. The fluorescence intensity signal over the nucleus is homogeneously distributed, but the fluorescence lifetimes showed sub-nuclear spots with reduced values indicating specific molecular interactions.

It was shown previously by yeast two-hybrid and FRET-FLIM that the petunia flowering gene PEG and FBP2 do not interact (Immink et al., 2002). This combination was used as negative control and the invariant fluorescence lifetimes for different locations in the nucleus indicated no interaction results not shown. As a positive control the combination FBP2-ECFP and FBPll-EYFP was transfected in cells and throughout the whole nucleus average fluorescence lifetimes of about 1.9 ns were measured data not shown (Immink et al., 2002).

15.6.3 Molecular interaction imaging via FRET-FLIM

FRET applications in combination whh cell biology have become a very powerful tool to obtain spatial and dynamic information of cellular processes in vivo. FRET can be used as a spectroscopic ruler to detect molecular interactions between proteins or molecules in living cells. In this review we have discussed different methodologies to measure FRET and FLIM having several advantages compared to intensity based methods. A good example of the additional valuable information is shown in the highly spatial resolved fluorescence lifetime images Fig. 15.4. FLIM measurements of the MADS box transcription factors showed sub-nuclear spots in the fluorescence lifetime images indicating molecular interactions between the different FBP's. The reduced fluorescence lifetimes in these regions may indicate possible transcriptional activation in the nucleus and these data could not be retrieved with other FRET techniques. Acceptor photobleaching (APB) is a good alternative method to measure FRET, but preferably in combination with FRET biosensors where the molar ratio of ECFP and EYFP is the same and concentration effects are minimal.

How can we improve FLIM measurements in the future? One option is to combine APB with FLIM. The fluorescence lifetime will be determined before and after a bleach pulse and the donor fluorescence lifetime should return to its original value after the destruction of the acceptor (Bastiaens and Squire, 1999). For this method the FLIM set up should be equipped with specific lasers to photobleach the acceptor and the cells need to be fixed to avoid redistribution of the acceptor. An alternative method is the simultaneous


determination of donor and acceptor fluorescence lifetimes. Upon physical interaction between a donor and acceptor molecules the donor fluorescence lifetime will become shorter and the acceptor fluorescence will grow in at the same rate as the donor fluorescence decay rate. In this configuration donor quenching and acceptor fluorescence in-growth is a direct internal prove for the existence of molecular inter-actions and donor quenching is not due to autofluorescent components from the cells.


We thank Gerco Angenent and Sacco de Vries for continuing interest and fruitful collaboration.

15.8. REFERENCES

Bastiaens, P. I. and Pepperkok, R., 2000, Observing proteins in their natural habitat: the living cell. Trends Biochem Sci 25, 631-637.

Bastiaens, P. I. and Squire, A., 1999, Fluorescence lifetime imaging microscopy: spatial resolution of biochemical processes in the cell. Trends Cell Biol 9, 48-52.

Becker, W., Bergmann, A., Hink, M. A., Konig, K., Benndorf, K. and Biskup, C , 2004. Fluorescence lifetime imaging by time-correlated single-photon counting. Microscopy Research and Technique 63, 58-66.

Bemey, C. and Danuser, G., 2003, FRET or no FRET: A quantitative comparison. Biophys J 84, 3992-4010.

Borst, J. W., Hink, M. A., van Hoek, A. and Visser, A. J., 2005, Effects of refractive index and viscosity on fluorescence and anisotropy decays of enhanced cyan and yellow fluorescent proiQm^. J Fluor esc \S, 153-160.

Borst, J. W., Hink, M. A., van Hoek, A. and Visser, A. J., 2003, Multiphoton microspectroscopy in living plant cells. In Proceedings ofSPIE, vol. 4963 ed. A. Periasamy and P. T. C. So, pp. 231-238.

Bunt, G. and Wouters, F. S., 2004, Visualization of molecular activities inside living cells v ith fluorescent labels. Int Rev Cytol 231, 205-277.

Burch, L. R., Scott, M., Pohler, E., Meek, D. and Hupp, T., 2004, Phage-peptide display identifies the interferon-responsive, death-activated protein kinase family as a novel modifier of MDM2 andp21WAFl. JM?/5/o/337, 115-128.

Campbell, R. E., Tour, O., Palmer, A. E., Steinbach, P. A., Baird, G. S., Zacharias, D. A. and Tsien, R. Y., 2002, A monomeric red fluorescent protein. Proc Natl Acad Sci USA 99, 7877-7882.

Causier, B. and Davies, B., 2002, Analysing protein-protein interactions with the yeast two-hybrid system. Plant Mol Biol 50, 855-870.

Chen, Y., Mills, J. D. and Periasamy, A., 2003, Protein localization in living cells and tissues using FRET and FLIM. Differentiation 71, 528-541.

Clegg, R. M., 1996, in Fluorescence Imaging Spectroscopy and Microscopy X. Wang and B. Herman Eds. New York: Wiley.

De Folter, S., Immink, R. G., Kieffer, M., Parenicova, L., Henz, S. R., Weigel, D., Busscher, M., Kooiker, M., Colombo, L., Kater, M. M. , Davies, B. and Angenent, G.C., 2005, Comprehensive interaction map of the arabidopsis MADS Box transcription factors. Plant Cell 17, 1424-1433.

Egea-Cortines, M., Saedler, H. and Sommer, H., 1999, Ternary complex formation between the MADS-box proteins SQUAMOSA, DEFICIENS and GLOBOSA is involved in the control of floral architecture in Antirrhinum majus. EMBOJIH, 5370-9.

Ferrario, S., Busscher, J., Franken, J., Gerats, T., Vandenbussche, M., Angenent, G. C. and Immink, R. G., 2004, Ectopic expression of the petunia MADS box gene UNSHAVEN accelerates

356 BORST ETAL

flowering and confers leaf-like characteristics to floral organs in a dominant-negative manner. Plant Cell 16,1490-1505.

Gadella, T. W., Jr., van der Krogt, G. N. and Bisseling, T., 1999, GFP-based FRET microscopy in living plant cells. Trends Plant Sci 4, 287-291.

Gerritsen, H. C , Asselbergs, M. A., Agronskaia, A. V. and Van Sark, W. G., 2002, Fluorescence lifetime imaging in scanning microscopes: acquisition speed, photon economy and lifetime resolution. J Mcro5c 206, 218-224.

Gerritsen, H. C. and De Grauw, C. J., 1999, Imaging of optically thick specimen using two-photon excitation microscopy. Microscopy Research and Technique 47, 206-209.

Gordon, G. W., Berry, G., Liang, X. H., Levine, B. and Herman, B., 1998, Quantitative fluorescence resonance energy transfer measurements using fluorescence microscopy. Biophys J 14,2102-21X3.

Gratton, E., Breusegem, S., Sutin, J. and Ruan, Q. Q., 2003, Fluorescence lifetime imaging for the two-photon microscope: time-domain and frequency-domain methods. J Biomed Opt 8, 381-390.

Hara, M., Bindokas, V., Lopez, J. P., Kaihara, K., Landa, L. R., Jr., Harbeck, M. and Roe, M. W.,2004, Imaging endoplasmic reticulum calcium with a fluorescent biosensor in transgenic mice. Am J Physiol Cell Physiol 287, C932-938.

Haupts, U., Maiti, S., Schwille, P. and Webb, W. W., 1998, Dynamics of fluorescence fluctuations in green fluorescent protein observed by fluorescence correlation spectroscopy. Proc Natl Acad Sci USA 95, 13573-13578.

Herman, B., 1998, Fluorescence Microscopy. BIOS Scientific Publishers, New York. Hink, M. A., Bisseling, T. and Visser, A. J. 2002. Imaging protein-protein interactions in Uving

cells. Plant Mol Biol 50, 871 -883. Ho, F. M. and Hall, E. A., 2004, A strand exchange FRET assay for DNA. Biosens Bioelectron 20,

1001-1010. Honma, T. and Goto, K., 2001, Complexes of MADS-box proteins are sufficient to convert leaves

into floral organs. Nature 409, 525-529. Immink, R. G., Ferrario, S., Busscher-Lange, J., Kooiker, M., Busscher, M. and Angenent, G. C ,

2003, Analysis of the petunia MADS-box transcription factor family. Mol Genet Genomics 268, 598-606.

Immink, R. G. H., Gadella, T. W. J., Ferrario, S., Busscher, M. and Angenent, G. C , 2002, Analysis of MADS box protein-protein interactions in living plant cells. Proc Natl Acad Sci USA 99, 2416-2421.

Jares-Erijman, E. A. and Jovin, T. M. 2003. FRET imaging. Nat Biotechnol 21, 1387-1395. Karpova, T. S., Baumann, C. T., He, L., Wu, X., Grammer, A., Lipsky, P., Hager, G. L. and

McNally, J. G., 2003, Fluorescence resonance energy transfer from cyan to yellow fluorescent protein detected by acceptor photobleaching using confocal microscopy and a single laser. J Microsc 209, 56-70.

Kenworthy, A. K., 2001, Imaging protein-protein interactions using fluorescence resonance energy transfer microscopy. Methods 24, 289-296.

Kneen, M., Farinas, J., Li, Y. and Verkman, A. S. 1998. Green fluorescent protein as a noninvasive intracellular pH indicator. Biophys J14, 1591-1599.

Lakowicz, J. R., 1999, Principles of Fluorescence Spectroscopy, 2nd edition. New York: Kluwer Academic/Plenum Publishers.

Lippincott-Schwartz, J., Snapp, E. and Kenworthy, A., 2001, Studying protein dynamics in living cells. Nat Rev Mol Cell Biol 2, 444-456.

Mahajan, N. P., Harrison-Shostak, D. C , Michaux, J. and Herman, B., 1999, Novel mutant green fluorescent protein protease substrates reveal the activation of specific caspases during apoptosis. Chem Biol 6, 401-409.

Matz, M. v., Fradkov, A. F., Labas, Y. A., Savitsky, A. P., Zaraisky, A. G., Markelov, M. L. and Lukyanov, S. A., 1999, Fluorescent proteins from nonbioluminescent Anthozoa species. Nat Biotechnol 17, 969-973.

Miyawaki, A., Griesbeck, O., Heim, R. and Tsien, R. Y., 1999, Dynamic and quantitadve Ca ^ measurements using improved cameleons. Proc Natl Acad Sci USA 96, 2135-2140.

Miyawaki, A., Llopis, J., Heim, R., McCaffery, J. M., Adams, J. A., Ikura, M. and Tsien, R. Y., 1997, Fluorescent indicators for Ca ^ based on green fluorescent proteins and calmodulin. Nature 388, 882-887.


Morgan, M. J. and Thorbum, A., 2001, Measurement of caspase activity in individual cells reveals differences in the kinetics of caspase activation between cells. Cell Death Differ 8, 38-43.

Nougalli-Tonaco, I. A., Borst, J. W., Vries, S. C , Angenent, G. C. and Immink, R. G., 2006, In vivo imaging of MADS box transcription factor interactions. Submitted.

Offterdinger, M., Georget, V., Girod, A. and Bastiaens, P. I., 2004, Imaging phosphorylation dynamics of the epidermal growth factor receptor. J Biol Chem 219, 36972-36981.

Pellegrini, L., Tan, S. and Richmond, T. J., 1995, Structure of serum response factor core bound to DNA. Nature 376, 490-498.

Peter, M., Ameer-Beg, S. M., Hughes, M. K., Keppler, M. D., Prag, S., Marsh, M., Vojnovic, B. and Ng, T., 2005, Multiphoton-FLIM quantification of the EGFP-mRFPl FRET pair for localization of membrane receptor-kinase interactions. Biophys J .HS, 1224-1237.

Raicu, v., Jansma, D. B., Miller, R. J. and Friesen, J. D., 2005, Protein interaction quantified in vivo by spectrally resolved fluorescence resonance energy transfer. Biochem J 3H5, 265-277.

Riechmann, J. L. and Meyerowitz, E. M., 1997, MADS domain proteins in plant development. Biol Chem31H, 1079-1101.

Rizzo, M. A., Springer, G. H., Granada, B. and Piston, D. W., 2004, An improved cyan fluorescent protein variant useful for FRET. Nat Biotechnol 22, 445-449.

Russinova, E., Borst, J. W., Kwaaitaal, M., Cano-Delgado, A., Yin, Y., Chory, J. and de Vries, S. C , 2004, Heterodimerization and Endocytosis of Arabidopsis Brassinosteroid Receptors BRIl and AtSERK3 BhK\. Plant Cell 16, 3216-3229.

Shaner, N. C, Campbell, R. E., Steinbach, P. A., Giepmans, B. N., Palmer, A. E. and Tsien, R. Y., 2004, Improved monomeric red, orange and yellow fluorescent proteins derived from Discosoma sp. red fluorescent protein. Nat Biotechnol 22, 1567-72.

Shore, P. and Sharrocks, A. D., 1995, The MADS-box family of transcription factors. Eur J Biochem 229, 1-13.

Squire, A. and Bastiaens, P. I., 1999, Three dimensional image restoration in fluorescence lifetime imaging microscopy. JMicrosc 193, 36-49.

Suhling, K., Siegel, J., Phillips, D., French, P. M., Leveque-Fort, S., Webb, S. E. and Davis, D. M., 2002, Imaging the environment of green fluorescent protein. BiophysJS3, 3589-3595.

Terry, B. R., Matthews, E. K. and Haseloff, J., 1995, Molecular characterisation of recombinant green fluorescent protein by fluorescence correlation microscopy. Biochem Biophys Res Com 217,21-27.

Tsien, R. Y., 1998, The green fluorescent protein. Annu Rev Biochem 67, 67509-67544. Verkhusha, V. V. and Lukyanov, K. A. 2004. The molecular properties and applications of

Anthozoa fluorescent proteins and chromoproteins. Nat Biotechnol 11, 289-296. Voss, T. C , Demarco, I. A. and Day, R. N., 2005, Quantitative imaging of protein interactions in

the cell nucleus. Biotechniques 38,413-424. Wallrabe, H., Elangovan, M., Burchard, A., Periasamy, A. and Barroso, M., 2003, Confocal FRET

microscopy to measure clustering of ligand-receptor complexes in endocytic membranes. Biophys J%S,559-51\.

Wilson, G. S. and Gifford, R., 2005, Biosensors for real-time in vivo measurements. Biosens Bioelectron 20, 2388-2403.

Wouters, F. S. and Bastiaens, P. I., 1999, Fluorescence lifetime imaging of receptor tyrosine kinase activity in cells. Curr Biol 9, 1127-1130.

Wouters, F. S., Verveer, P. J. and Bastiaens, P. I. H., 2001, Imaging biochemistry inside cells. Trends Cell Biol 11, 203-211.

Xia, Z. and Liu, Y., 2001, Reliable and global measurement of fluorescence resonance energy transfer using fluorescence microscopes. Biophys J SI, 2395-2402.

Xu, X., Gerard, A. L., Huang, B. C , Anderson, D. C , Payan, D. G. and Luo, Y., 1998, Detection of programmed cell death using fluorescence energy transfer. Nucleic Acids Res 26, 2034-2035.

Zacharias, D. A., Violin, J. D., Newton, A. C. and Tsien, R. Y., 2002, Partitioning of lipid-modified monomeric GFPs into membrane microdomains of live cells. Science 296, 913-916.

Zhang, J., Campbell, R. E., Ting, A. Y. and Tsien, R. Y., 2002, Creating new fluorescent probes for cell biology. Nat Rev Mol Cell Biol 3, 906-918.

SPECTROSCOPY FOR THE ASSESSMENT OF MELANOMAS

Ousama M. A'Amar and Irving J. Bigio'

16.1. INTRODUCTION

The mortality rate due to cutaneous malignant melanoma has been growing worldwide ' ' '' . Successful management of melanoma depends on early diagnosis and excision. Prior to excision the diagnostic accuracy of identifying a pigmented skin lesion as malignant or benign relies essentially on the practitioner expertise. However, for non-specialists, many pigmented skin lesions are difficult to classify, especially when they present unusual features^. Chen et al. have reported that the probability of primary-care physicians to correctly determine that a pigmented lesion may be malignant and to make the appropriate management decision (either to order a biopsy or to refer the patient to a melanoma specialist) can be as low as 0.42^. Melanoma can sometimes be confused, even for the skin-cancer specialist, with other types of pigmented lesions (melanocytic nevi)^.

Early melanoma diagnosis is mostly based on the visual features of the suspected lesions such as the asymmetry, border, color and dimension (ABCD)^. Several computerized imaging techniques that could help in recognizing these criteria have been developed . Applying the ABCD criteria, a spectrophotometric imaging technique that uses selected spectral bands from

Ousama M. A'Amar, Boston University, Department of Biomedical Engineering, Boston, MA 02215. Irving J. Bigio, Boston University, Department of Biomedical Engineering and Electrical and Computer Engineering, Boston, MA 02215.

359

360 O. M. A'AMAR and I. J. BIGIO

420 to 1040 nm in order to enhance the ABCD criteria (and mainly the color of the suspected lesions) has been shown to improve the accuracy of detecting melanoma^ ' ' ' . Dermoscopy, also known as dermatoscopy, epiluminescence microscopy and skin surface microscopy, is an imaging technique that permits the visualization of morphological features that are not visible to the naked eye " . It is extensively used in clinical practice to enhance the detection of microscopic features (e.g. brown globules, pigment network, branched streak and homogenous blue pigmentation) of cutaneous melanocytic lesions^ ' . Dermoscopy has shown higher sensitivity values (90% - 95%) compared with the average sensitivity (70% - 80%) for clinical visual examination^^. However, the specificity (correctly identifying benign lesions) is still poor. A fiber-optic confocal imaging system, with associated fluorescent markers, has been tested on an animal model and was sensitive to melanoma tumors up to 0.2 mm under the skin surface^^. However, the usefulness of this system for clinical application in distinguishing benign melanocytic lesions from melanoma has not been demonstrated, and histopathological diagnosis remains the gold standard^^.

In general, research investment in new optical, non-invasive methods to improve clinical diagnosis has rapidly grown over the past two decades. Recent advances in several areas of optical technology have led to the development of a variety of "optical biopsy" techniques. These techniques offer a range of modalities for diagnosis, and include light-induced fluorescence spectroscopy (LIFS) elastic-scattering spectroscopy (ESS), optical coherence tomography (OCT)^ ' ' ^ diffuse optical tomography (DOT) ' ' ' ^ Raman and vibrational spectroscopy ' ' , confocal microscopy ' , among others. Optical characterization of tissues may offer several advantages over the conventional methods because of its noninvasive/minimally-invasive nature, and its ability to provide clinical information in real time.

Fluorescence-based techniques have attracted significant attention as potential methods for in vivo characterization of biological tissues. The use of LIFS in medicine is often associated with photodynamic therapy (PDT) for the quantification of exogenous fluorophores (e.g. photosensitizers) and monitoring their bio-distribution ^ ' ,33 jvjany other clinical applications using autofluorescence signals (due to native fluorophores in tissue) were further developed, including cancer screening and diagnosis tumor staging and surgical

For pigmented skin lesions such as melanoma, the absorption coefficient, predominantly resulting from melanin, is relatively high at both excitation and emission wavelengths of fluorescence by native fluorophores. Melanin is known to absorb light very strongly in the UV-visible spectral region, with the extinction coefficient rising for shorter wavelengths. The extinction spectrum of eumelanin ^^ (the class of melanin found in human black hair) is shown in Figure 16.1. Skin pigmentation of some pathological conditions may also be affected by the scattering of light by melanosomes and/or melanin aggregates' ^. Melanin cannot be regarded as a pure absorber because the size (0.4 - 0.8 jiim) and shape (ovoid) of melanosomes may result in them behaving as substantial

SPECTROSCOPY FOR THE ASSESSMENT OF MELANOMAS 361

500 600 Wavelength [nm]

900

Figure 16.1. Extinction spectrum of eumelanin. The data for this spectrum was obtained from the Laser Photomedicine and Biomedical Optics at the Oregon Medical Center (http://omlc.ogi.edu). They refer to T. Sama, HM Swartz as their source of data'*^

scattering centers of light" . Melanin granules may also strongly scatter light at shorter wavelengths as a result of Rayleigh scattering' '* . This suggests that the excitation light in the UV-visible spectral region w ould be strongly absorbed and/or scattered by melanin, consequently reducing the efficiency of exciting tissue fluorophores. In addition, the emitted fluorescence w ould be strongly reabsorbed before reaching the probe through the surface of the lesion. These conditions limit the suitability for assessing melanoma by LIFS.

ESS, sometimes inappropriately called diffuse-reflectance spectroscopy"^" , is a non-invasive technique that analyzes the variations of elastically backscattered light from different tissues illuminated with a broadband light source with a harmless wavelengths range (UV - NIR). (With ESS the optical geometry is such that collected photons have only scattered a few times, and analysis methods based on diffusion theory are not relevant.) Most cancers manifest alterations in the structure of cells" . Since the cancerous cell density, size and shape differ from that of the normally growing cells'* , they can be optically differentiated. ESS translates these morphological changes into spectral features" " . Hence, ESS may be an alternative and/or complementary technique to LIFS, since it provides spectra that contain information about the cellular and sub-cellular morphology (scattering centers) of the tissue as well as the chromophore content (e.g. hemoglobin and melanin)"^ . The information


derived from the ESS is comparable to that are commonly used on the analysis of microscopic structure in standard histopathological assessment ^^ On the other hand LIF provides information as to the biochemistry of tissue, especially compounds related to the energy cycle of cells.

This chapter focuses on the problem of pigmented skin lesions for which the standard histopathologic examination is the best method available for determining the presence or absence of melanoma. We review the application of LIFS and ESS for the in vivo assessment of pigmented skin lesions and melanoma in order to determine if different spectral pattems could differentiate malignant melanoma from dysplastic and benign nevi.

16.2. SKIN MELANOMA

The skin consists of two major layers: epidermis and dermis. The upper epidermis layer consists of five sub-layers. The stratum comeum is the outermost layer, comprising keritanized epithelial cells. The other layers are the stratum germinativum, the stratum spinosum, the stratum granulosum and the stratum lucidum" .

Malignant melanoma is a neoplasm that occurs for the most part on the skin. Neoplasms of melanocytes constitute a model of tumor progression that evolves in a stepwise fashion^^. The presence of intercellular junctions and tonofilaments are determinants for the diagnosis of melanosomes for melanoma"^ . The progression of the melanocyte to a malignant melanoma passes through various steps: development of benign nevocellular nevus, preneoplastic dysplastic nevus, primary melanoma, and metastatic melanoma^ The thickness, known as the Breslow thickness, of the lesion is an important prognostic factor for melanocytic lesions ' ' " . Survival closely correlates with the thickness of the melanoma. It is generally believed that recognizing the early signs of melanoma may lead to earlier diagnosis and removal of 'thin' lesions (< Imm)^^ before they become invasive.

Melanosomes are the small vesicles that contain melanin. The melanin content and composition, as well as the sizes of melanosomes, in the human epidermis vary considerably with both ethnicity and chronic sun exposure^^. The morphology and contents of melanosomes are important features for differentiating melanocyte-derived melanotic lesions such as melanosis and malignant melanoma^^. The cross-sectional size range of melanosomes in humans varies between 0.94 x 10" |im^ and 1.44 x 10' |im^ depending on the complexion (from light to dark)^^. In malignant melanoma in situ, abnormal melanosomes are pleomorphic and ellipsoid, with an increase in sulphur content and alkali elution rate. In invasive malignant melanoma, the irregular ellipsoid and spheroid melanosomes are found either as discrete bodies or compound melanosomes with further increased sulphur content and alkali elution. Abnormal melanosome morphology and high sulphur content are predictive markers for assessment of early or precancerous melanotic lesions and malignant melanoma^^.


16.3. FLUORESCENCE SPECTROSCOPY

The term Light Induced-fluorescence spectroscopy (LIFS) refers to a point measurement on tissue of a full emission spectrum for a specific excitation wavelength (UV-Visible), typically utilizing a small fiber-optic probe. LIFS can be classified under two categories: (1) Autofluorescence, which refers to the in situ analysis of native tissue fluorescence resulting from endogenous fluorophores that are naturally present in the tissue. (2) exogenous drug-fluorescence, that is commonly used to describe the process of analyzing the fluorescence spectrum in situ of an exogenous fluorophore (photosensitizer or marker) after a time interval of systemic or topical administration of the fluorophore to the patient. Some exogenous fluorophores preferentially concentrate in malignant or pre-malignant tissues, whereas others concentrate in areas of increased vascularity, which also, generally, accompanies tumors. Although the fluorescence from such fluorophores provides a large signal, which can be helpful in the detection process of malignancy in some organs^ ' , the selectivity (tumor/normal tissue concentration ratio) of most available fluorophores is still poor ' ' ' .

16.3.1. Instrumentation

Both autofluorescence and exogenous-drug fluorescence measurements can be carried out utilizing the basic fluorescence spectroscopy instrumentation shown in Figure 16.2. The instrumentation includes a monochromatic light source, a spectrograph, a multichannel detector (PDA or CCD) sensitive over the UV-visible region and an optical guide for excitation light delivery and fluorescence collection. The optical guide is usually a fiber-optic probe, composed of single bifurcated fiber, two fibers or a bundle of fibers for guiding excitation light and fluorescence signal.

16.3.2. Melanoma Diagnosis by Autofluorescence

Tissues naturally contain several fluorophores that are involved in cellular metabolic process such as nicotinamide adenine dinucleotide (NADH) and flavins (FAD) or are associated with their structural matrix (elastin and collagen). Other fluorophores include the aromatic amino acids (tryptophan) and endogenous porphyrins'^. Fluorescence spectroscopy is a sensitive technique for measuring the biochemical composition of the tissue and potentially providing semi-quantitative information on fluorophore bio-distribution in tissue. However, spectral distortion due to the scattering properties of tissue'^ and absorption by tissue chromophores (e.g. hemoglobin and melanin) cause fluorescence measurements in tissue to be significantly more complicated than measurement of a simple solution of several fluorophores ' ' ' . Despite the distortion effects due to scattering and absorption, the general spectral trends seen in ex vivo and in vivo studies have demonstrated the potential for fluorescence spectroscopy as a real-time noninvasive diagnostic technique. Using various optical configurations LIFS has


Mo nochromadc light soufve

D: Soui

C: CoUecdi^Wer Conirolier

Figure 16.2. Schematic view of the basic concept of spectroscopy instrumentation used for clinical studies of tissue autofluorescence and induced fluorescence.

been widely investigated as a method for the localization of precancerous lesions of the aerodigestive tract ^ ' 5' ' , bronchus'''"'''"''^' colon 2,72,73,74,75 cervix^ ' ' ' ' , skin^^ (non-melanoma) and other organs .

Pathological tissues, when excited with ultraviolet or visible light, exhibit lower autofluorescence intensity and altered spectral shape when compared to normal surrounding tissue. The fluorescence ratio between the red region and the green region of the fluorescence spectrum measured in cancers attains higher values than that measured in normal surrounding tissues. The drop of autofluorescence intensity in pathological tissue of one or several fluorophores may be due to the morphological modifications of tissue, the deficit of fluorophores and/or the degradation of their efficiency caused by an altered balance between the reduced and oxidized forms of certain fluorophores '*'' ' . In addition, the concentration of chromophores (absorbers) in some pathological tissue, may increase compared to normal surrounding tissue, resulting in a significant change in the relative light absorption by fluorescent and non-fluorescent molecules, which consequently has the effect of reducing the fluorescence efficiency ^ . (Since the total number of excited fluorescent molecules is reduced, the overall intensity will be reduced/) In an attempt to determine how changes in tissue fluorophores and chromophores affect the fluorescence signal. Ramanujam et al fitted in vivo fluorescence spectra to a model of fluorescence in turbid tissue, which included contributions from NAD(P)H, FAD, collagen and elastin and took into account absorption by hemoglobin^^. They observed an average increase in NAD(P)H content and an


average decrease in the contribution of collagen fluorescence in the cervix, as the tissue progresses from normal to carcinoma in situ . In contrast, NAD(P)H fluorescence of colonic tissue measured in vitro appears to decrease as tissue progresses from normal to abnormal '*. The increase in the contribution of collagen fluorescence seen by Ramanujam et al for cervical tissue is related to the results of Schomaker et al and Bottiroli et al, both of whom assert that changes in fluorescence are at least partly due to differences in the structural organization of the tissue ' ' . In particular, in the case of polyps, there is a thickening of the mucosa, which shields some of the underlying collagen fluorescence.

LIFS for the in vivo detection of skin pathologies has been investigated by several research groups. The skin contains the fluorophores elastin, collagen, keratin, and NADH, which are expected to contribute to the fluorescence spectrum, as well as the absorbing chromophores melanin and hemoglobin.

Lohmann et al. used fluorescence spectroscopy as a diagnostic tool in situ for nc oi: on

differentiating melanoma from dysplastic nevi and normal tissue ' ' They excited skin tissues in vivo using UV light (365 nm) and collected autofluorescence from normal tissues, nevi and melanoma. On 147 lesions, they were able to distinguish non-dysplastic nevi from melanoma and dysplastic nevi. Although, the spectra of all types of the examined tissues had a simple structure peaking around 475 nm, they noticed that melanoma mainly exhibited very low fluorescence intensity compared to normal surrounding tissues. They also observed an increase of fluorescence intensity in the zone adjacent to a melanoma compared to both the melanoma and normal surrounding tissues. The authors found that the increase in fluorescence intensity in the transition zone was helpful in differentiating melanoma from dysplastic nevi. The same technique was applied for determining ex vivo dysplasia and invasive carcinomas in the cervix uteri. Again, 365-nm excitation light was used and fluorescence was mainly observed at 475 nm. They observed an increase of fluorescence intensity with the degree of dysplasia compared to healthy tissue with higher intensity at the borders of the malignant lesion. In the tumor region, the intensity was very small. The authors attributed the naturally occurring fluorescence to NADH^^

Leffell et al. studied autofluorescence of skin with 325 nm excitation in vivo in 28 human volunteers in an attempt to correlate the results with chronological aging or photoaging of the skin^ . They found an increase in fluorescence intensity of 10%, in correlation with age. As the correlation was not significant, they concluded that the autofluorescence spectra were unrelated to age, pigmentation, or skin thickness. However, they found a correlation with photoaging that was related to photoexposure. This correlation is not believed to be caused by simple differences in melanin content.

Sterenborg et al. have used a more sophisticated autofluorescence technique for examining skin in vivo^^. The technique, called excitation-emission matrix spectroscopy, utilizes a sequence of excitation wavelengths and acquires the fluorescence spectrum at each excitation. Different excitation wavelengths might be expected to excite different fluorophores, resulting in more complex emission patterns with more information relevant to biochemical


changes than for single wavelength excitation, and with presumed greater likelihood of distinguishing malignancy from normal conditions. The results of Sterenborg et al. ' ^ disagree with those reported by Lohman et al. although they both used the same measurements protocol (365 nm excitation). Sterenborg et al concluded that there were no significant differences between the fluorescence of control sites and non-melanoma skin tumors. For the eight melanomas and eight benign pigmented lesions they stated that 'neither the shape of the fluorescence intensity distribution, nor the spatial distribution of the fluorescence intensity showed any signature specific to the histopathological nature of the lesions investigated'. Also, by removing the stratum comeum and measuring the fluorescence, they established that an important component of the fluorescence is from keratin in the stratum granulosum. Minor contributions to the fluorescence from other endogenous fluorophores were also noted in the tumors, but were not reliable enough to be used as a cancer diagnostic.

Kollias, et al. have used a bifurcated bundle of 0.1-mm fibers randomly arranged for analyzing the autofluorescence spectra of skin in an animal model and in humans in several studies. In one study they measured the excitation spectra in vivo of heat-separated epidermis and dermis, and on extracts of mouse skin to characterize the absorption spectra of the emitting fluorophores as a function of age and chronic UVB exposure^^ They attributed the emitted autofluorescence at maximum excitation peaks 295, 340, and 360 nm to tryptophan, pepsin digestable collagen crosslinks and collagenase digestable collagen crosslinks, respectively. The authors believe that the 295-nm excitation peak originates in the epidermis, while the peaks at 340 and 360 nm originate in the dermis. They found that autofluorescence excitation peaks remained unchanged in chronologically aged mice; however, their intensity distribution was related to age. For mice chronically exposed to UVB, a large increase in the 295-nm peak was observed in comparison with unexposed mice of same age. Meanwhile, the peaks at 340 and 360 nm were no longer distinct and two new peaks appeared in the chronically exposed mice at 270 nm and at 305 nm. The authors concluded that changes in skin autofluorescence were related to aging that caused predictable alterations in both epidermal and dermal fluorescence, whereas chronic UV exposure induced the activation of new fluorophores. In another study. Gillies et al. analyzed skin autofluorescence excitation and emission spectra in 24 normal and diseased human subjects. They found the same epidermal and dermal autofluorescence excitation peaks as in the animal model, with a slight shift for the collagen cross-links emission peaks (335 and 370 nm) ^ . An additional excitation peak at 440 nm was observed in human skin that was not found in mice. Fluorescence spectra obtained from lesions and healthy sites of psoriatic patients were characterized by a significantly larger signal at 295 nm excitation compared to those obtained from healthy volunteers that they contributed to cellular proliferation. Brancaleon et al. have investigated in vivo and ex vivo autofluorescence of non-melanoma skin lesions in 18 patients^^. Again, they observed that tryptophan was more intense in diseased tissue (basal cell carcinomas and squamous cell carcinomas) than in normal. The finings were also related to the epidermal thickening and/or hyperproliferation. Conversely, the fluorescence intensity associated with


dermal collagen crosslinks was generally lower in tumors than in the surrounding normal tissue, probably because of degradation or erosion of the connective tissue due to enzymes released by the tumor. Such results may be useful for differentiating normal from diseased skin tissue by LIFS. However, in these studies the skin diseases/malignancies were progressive, and the authors did not evidently prove the possibility of classifying skin pathologies by LIFS.

Chwirot et al. have conducted several studies using a digital imaging system for assessing pigmented skin lesions. In one study, they excited skin tissues with the 366 nm Hg line and collected autofluorescence images of spectrally resolved autofluorescence through a narrow band-pass filter centered at 475 nm. The examined population included 90 melanoma, 205 common melanocytic and dysplastic nevi, and 113 lesions of different types '*. Autofluorescence images showed characteristic patterns of intensity distributions reflecting the outlines of the lesions. The average maximum to minimum intensity ratio in normal skin areas of 3-5 cm in diameter was approximately 2. The ratio of highest intensity outside the lesion to the lowest intensity within the lesion was used for diagnosis. A ratio equals or higher than 7 indicated the presence of melanoma. With the diagnosis criterion, melanomas were detected with a sensitivity of 82.5% and a specificity of 78.6%. The positive predictive value for melanoma versus other pigmented lesions was 58.9%. In another study that included 7228 pigmented lesions (56 melanoma) in 4079 subjects in which 568 cases were correlated with histopathology, in situ digital imaging was capable of detecting melanoma with a sensitivity of 82.7% and a specificity of 59.9%^^. Meanwhile, the positive predictive value for melanoma versus all pigmented lesions was as low as 17.5%. In another ex vivo study using microscopic digital imaging technique and microspectrofluorimetry, Chwirot et al. investigated specimens fixed with formalin and embedded in paraffin from 50 malignant melanomas, 4 basal cell carcinomas and 58 benign lesions. Their results showed a sensitivity of 74% and a specificity of 59% in differentiating melanoma from other pigmented lesions^^.

The common conclusion for all 3 papers demonstrate that in vivo and ex vivo autofluorescence can be used as auxiliary techniques for the directing the assessment of melanoma. The fluorescence intensity variability over a small area and among different individuals can be problematic. . Autofluorescence imaging and spectroscopy cannot be used solely as a diagnostic technique. In vivo autofluorescence may be useful in screening large populations of patients to assist in recommending selected patients for examination by specialists, although a more robust measurement would be desirable.

The use of optical spectroscopy for non-invasive diagnosis of malignant melanoma can be seriously compromised by spatial variations and other factors in the optical properties of the tissue that are not related to malignancy^^. The variation of fluorescence intensity with photoaging or with pathology conditions using the ratio method may not be sufficient for discriminating between normal and malignant nevi. More elaborated approaches may be needed to take into account the natural variability among individuals and among different locations of tissue in the same individual. The correction for spatial


variations in the optical properties of the tissue should be considered. Other correction techniques based on ESS could offer more reliable correlation between histopathology and fluorescence spectral trends.

16.3.3. Melanoma Diagnosis with exogenous fluorophores

LIFS can be used for the measuring the compound accumulation in the tumor for pharmacokinetic evaluation of drugs used in PDT. The determination of the maximum fluorescence ratio (tumor/healthy tissue) should enable optimization of the time interval between drug administration and tumor irradiation. Although fluorescence measurements give a semi-quantitative estimate of tissue photosensitizer content^^, a correlation has been shown between fluorescence signal of certain exogenous fluorophores and internal tumor content measured by extraction and reverse-phase high-performance liquid chromatography (HPLC) .

However, most exogenously administered fluorophores (photosensitizers), that are currently in use for clinical applications (diagnosis and PDT), have poor selectivity and produce undesirable side effects such as toxicity and photosensitivity. The photosensitizer is retained in the tumor as well as in the normal surrounding tissue. Photosensitizers such as hemato-porphyrin derivative (HpD), meta-tetra (hydroxyphenyl) chlorin (mTHPC) and benzopophyrin derivative (BPD) are designed for PDT and are not necessarily suitable for cancer diagnosis^^. Such fluorophores are retained in normal skin for a longer time than other normal tissues. Hence, the use of exogenous fluorophores that are systemically administered to the patients is not the ideal solution for cancer detection since the selectivity of these agents is often poor, and they can result in skin photosensitivity. Furthermore, the procedure is invasive and requires a waiting time ranging between 6 and 96 hours to reach the maximum selectivity for diagnostic purposes.

A compound found naturally in the heme cycle, delta-aminolevulinic acid (ALA) is converted in the body to protoporphyrin IX (PplX) and has been clinically used for diagnosis and treatment of non-melanoma skin cancers^^^ and other skin diseases ^ ' ^ ' ^ . ALA is usually applied topically to the patients and only requires a few hours (2-6 hours) to reach its maximum selectivity in malignant lesions.

Sterenborg et al. have suggested the use of LIFS for diagnosis of malignant melanoma using an approach that allows eliminating the spatial variations in the optical properties of the tissue investigated. ALA was topically applied on moles and unpigmented control skin in human volunteers. Fluorescence and color measurements were performed before and after topical application of ALA. Single and double ratio techniques were used for the evaluation. The first was based on the ratio of the red to the yellow fluorescence (660-750 nm to 550-600 nm) at 405 nm excitation and the second was based on the red-to-yellow ratio at 405 nm excitation divided by the red-to-yellow ratio at 435 nm excitation. The single ratio technique showed a significant correlation between fluorescence and color. The double ratio was independent of the color of the lesion. The findings indicated that the double-ratio technique may be suitable


for in-vivo detection of local differences in concentration of fluorescent tumor-localizing drugs in pigmented lesions and consequently may significantly facilitate the development of a fluorescence diagnostic tool for malignant melanoma^^.

16.4. ELASTIC SCATTERING SPECTROSCOPY

The parameters most commonly used to characterize the optical properties of scattering media such as tissue are: (1) The scattering coefficient, |is(^)? defined as the inverse of the mean-free-path between scattering events; (2) the absorption coefficient, |ia(^)? defined as the inverse of the absorption mean-free-path; (3) P(0) representing the probability of scattering through an angle 0 and is also called the phase function; and the anisotropy factor, g = <cos0>, representing the mean cosine of the scattering angle. For most tissues the scattering is mostly forward as the value of g is typically greater than 0.85 . It is also convenient to use the reduced scattering coefficient, which can be regarded as the inverse of an effective distance between direction-randomizing scattering events, and is defined as ^s' = |Us(l~g)-

The scattering and absorption properties of a tissue depend on its cellular structure, architectural features and the wavelength of light used ^ . Absorption is due to specific chromophores such as water, hemoglobin, keratin and melanin^^ ' ^ , whilst scattering is caused by the different refractive indices of tissue components such as cell organelles, structural proteins and membranes'*^. ESS spectra from normal tissue often differ from those of malignant tissues, providing the potential for ESS to be utilized as a non-invasive diagnostic aid in melanocytic lesions.

Given the strong scattering and absorption by melanin granules, UV-visible and NIR elastic scattering spectroscopy may offer an useful tool for diagnosing pigmented skin lesions. When ESS is employed for tissue diagnosis, the tissue pathologies are detected and diagnosed using spectral measurements of the elastically-scattered light that characterizes the wavelength dependences of both scattering and absorption in the tissue. The use of a technique that is sensitive to the wavelength dependence of scattering efficiency and angles, as well as to absorption bands, results in composite signatures that appear to correlate well with differences in tissue types and pathologies.

16.4.1. Principles of Elastic Scattering Spectroscopy

Since the cellular components that are responsible for elastic scattering have dimensions typically of the order of visible to NIR wavelengths, the elastic scattering properties will exhibit a wavelength dependence that is more complex than for simple Rayleigh scattering (Ifk^), which is relevant for particles that are much smaller than the optical wavelength. In our approach to ESS for tissue diagnostics, separate optical fibers are used for illumination and collection, and the fiber tips are placed in gentle contact with the tissue surface. (See Figure


16.3). When the fibers are spaced close to each other (< 0.05 cm), as with an endoscope-compatible probe, the wavelength dependence is readily measured. Thus, for such geometries, changes in the size, densith and morphology of subcellular structures (e.g., nuclei and other organelles) can be expected to cause significant changes in the optical signature that is derived from the wavelength dependence of elastic scattering. The principles of ESS and its operating features have been described by Bigio and coworkers in earlier pubHcations ' '

With the probe in optical contact with the tissue under examination, and with separate illuminating and collecting fibers, there is no interference due to surface reflection (specular reflection). Thus, the light that is collected and transmitted to the analyzing spectrometer must first undergo a few scattering events through a small volume of the tissue before entering the collection fiber(s). The experimental conditions can be simulated by Monte Carlo (MC) computational methods. The scattering properties of the scattering centers can be calculated using Mie theory for similar sphere sizes, to yield the wavelength dependence of the scattering efficiency, as well as the phase function, although sometimes the Henyey-Greenstein (HG) approximation for the phase function is utilized.

Differences in lesion thickness and microscopic architecture changes are both likely to affect the ESS signal. UV-visible ESS has been demonstrated to be sensitive to architectural changes in tissue structure at the cellular and subcellular levels" '"* ' ^ ' The potential of this technique in vivo has been reported for diagnosing breast tissues and breast sentinel lymph nodes'*\ upper and lower GI tract mucosa''^'''\ bladder''"* and pigmented skin lesions'* . In these studies, ESS data showed repeatable spectral trends with the histopathological assignments. ESS may be an appropriate technique for melanoma diagnosis, with the advantage of noninvasively providing real time diagnostic signatures in situ that are related to histopathology.

As stated in the introduction, melanin is known to absorb and scatter light very strongly in the UV-visible spectral region, with the extinction coefficient rising for shorter wavelengths as result of Rayleigh scattering caused by the melanin granules"^. Optical absorption by melanin is much weaker in the near-infrared (NIR), where scattering dominates over absorption in most tissue^^'"^'"^ Thus, for one of our studies we were motivated to determine whether NIR-ESS may have the advantage (over UV-vis ESS) of being more sensitive to changes in scattering that accompany dysplasia, rather than being mainly sensitive to the presence of major absorbers (e.g. hemoglobin, keratin and melanin).


Light Detector

Light Source

Figure 16.3. A schematic diagram showing the optical geometry of a fiber-optic probe used in optical contact with the tissue for elastic-scattering spectroscopy.

16.4.2. Instrumentation

The basic schematic of ESS instrumentation is shown in Figure 16.4. The setup is similar to that used for LIFS (shown in figure 16.2) but with the monochromatic light source being replaced by a broadband light source such as an arc lamp, and eliminating the long-pass filter at the detection level. The probe must be used in optical contact with the examined tissue or designed in a way by which only light that has scattered through the tissue can enter the collection fiber. The intent with this approach is to generate spectral signatures of relevance to the tissue parameters that pathologists address.


^^Spectrometer

^ Detector Array

Optical-fiber tissue probe

in Con^uter and

Interface Electronics

Figure 16.4. Schematic of the system components for elastic scattering spectroscopy

The spectral response of the spectroscopic system is usually calibrated by recording a reference spectrum from a spectrally-flat diffuse reflector such as Spectralon®. Thus, the displayed spectrum is calculated according to:

I{A)- ^UL.-/U), tissue background

I{^h-Ii^\ (1)

ref background

where the subscript "background" refers to measurements made without triggering the light source, and includes ambient light and detector dark current.

The ESS probe geometry can be designed to be sensitive to both scattering and absorption properties. However, for certain diagnostic applications it is probably more appropriate to emphasize the scattering to the detriment of absorption properties in order to obtain more distinctive features between normal and malignant tissues.

Some groups have also studied ESS coupled to a single fiber probe for illuminating tissue and collecting the backscattered light. This geometry was shown to be sensitive to the particle size and consequently to the changes of scattering properties in the superficial layer, which is associated to premalignant conditions ^ ' ^ . This approach presents a challenge for clinical application in that the fiber self-reflectance must be accounted for in the system calibration, [not relevant to this paper]


For some types of diagnostic application of ESS it becomes desirable to enhance the sensitivity to photons that have undergone a single backscattering event or a small number of large angle scattering events. Such a circumstance optimizes sensitivity to variations in the large-angle components of the scattering phase function. It was shown that polarized measurements, which enhance sensitivity to singly-scatterd photons, facilitate extraction of information about the nuclear size ^^ '' \ and the sub-nuclear structure ^ .

For the application to measurements on pigmented lesions, we decided on a small center-to-center separation of about 350 jim between light-delivery and light-collecting fibers for measuring the wavelength-dependent spectra of elastically scattered light from cutaneous melanocytic lesions, making the device sensitive to both scattering and absorption properties ^ , since the interplay of those two properties can be indicative of the condition of melanosomes.

16.4.3. Preclinical Trials

In a recent study, A'Amar et al. reported on ESS measurements on an animal model. South American opossums were treated with dimethylbenz(a)anthracene on multiple dorsal sites, which induces both malignant melanomas and benign pigmented lesions"^ . In this animal model the melanosomes in melanocytic lesions are generally oval (-0.5 |im x 0.3 jLim), solitary and not clustered in compound melanosomes^ ' . Skin lesions were examined in vivo with ESS using UV-visible and NIR instrumentations, with wavelength ranges of 330-900 nm and 900-1700 nm, respectively. ESS measurements were made on the lesions, and spectral differences were grouped by diagnosis from standard histopathological procedure. Typical UV-visible and NIR spectra for pigmented benign lesions, melanomas and normal skin, are shown in Figure 16.5. Surprisingly, ESS spectra for melanoma in the short-wavelength range showed increased scattering, which appears to overtake the effect of increasing absorption. In contrast, the ESS signal from benign pigmented lesions either decreased, as expected, due to higher absorption at shorter wavelengths, or exhibited only a slight rise in signal. The authors suggested that melanosomes and melanin granules in malignant lesions act more strongly as Mie and Rayleigh scatterers than is the case in benign pigmented lesions, in agreement with the findings of Wolbarsht et al., in which they showed that, in addition to absorption properties, the melanin granules act as strong Rayleigh scattering centers ^ ^ (due to their large refractive index).

The common feature in the NIR spectra among the three tissue types is the broad absorption band due to water at around 1450 nm. However, the full width at half maximum (FWHM) of the melanoma scatter spectrum in the range of 920-1400 nm is significantly larger than that of benign lesions. Moreover, a small feature around 1150 nm also distinguishes the melanoma spectra from those of benign nevi. The spectrum for normal skin shows a small absorption feature due to water, around 1200 nm, which does not appear either in benign


xlO"-

200 400 600 800 1000 1200 Wavelength [nm]

1400 1600 1800

Figure 16.5. Normalized elastic-scattering spectra representative of typical data for pigmented benign lesions, melanomas and normal skin measured on opossums using UV-visible ESS system. B: benign pigmented lesion, M: melanoma and N: normal skin.

lesions or melanomas. This is consistent with deeper sampled depth in normal skin due to lower scattering without melanin.

It was concluded that both UV-visible and NIR elastic-scattering spectroscopy provide spectral criteria that correlate with the histopathological findings. The UV-visible ESS correctly classified 85% of the examined lesions, whereas the NIR-ESS correctly classified 92%, although the statistical significance is limited due to the small sample sizes in that study. The differences in lesion thickness (depth) had only a modest effect on the ESS spectral data in the NIR region for melanocytic lesions, because so much of the signal for this optical geometry used comes from scatterers at shallow depths (< 400 |im). Both UV-Visible and NIR ranges are sensitive to the melanosomes, which are the strongest sub-cellular scatterers in pigmented lesions.

16.4.4. Clinical Studies

Optical properties of melanoma and skin cancers have been explored using different configurations of ESS or diffuse reflectance spectroscopy by several research groups. Marchesini et al. have used diffuse reflectance spectrophotometry in the UV-visible region (400 to 780 nm), which incorporates an external integrating sphere, on different cutaneous pigmented lesions, including primary and metastatic malignant melanoma, pigmented nevi, lentigo and seborrhoeic keratosis^^^. They found that differences in reflectance spectra were more pronounced in the near IR region. Significant spectral difference (P < 10' ) between begin nevi and malignant melanomas were recorded. In another study that included 31 primary melanomas and 31 benign


nevi they were able to distinguish the two groups with a sensitivity of 90.3% and a specificity of 77.4%^^ . Their method provided discriminating power between melanoma and nevi equivalent to that of a specialized medical judgment, although the data set was small. They suggested the technique for screening skin pigmented lesions and developed a CCD-based imaging technique, which allows an objective evaluation of the color, size, shape and boundary of lesions to distinguish melanoma from other pigmented lesions. They used a set of interference filters at selected wavelengths ranging from 420 to 1040 nm to characterize these features. This technique has been proposed as a useful tool to discriminate cutaneous melanoma from other pigmented cutaneous lesions^^^. In another study by this group, including 67 melanoma and 170 non-melanomas pigmented lesions, melanomas were identified by lesion dimensions, mean reflectance, lesion roundness and border irregularity^^.

Another spectrophotometric technique, reported by Moncrieff et al., producing spectrally filtered images, was used on 348 pigmented lesions (52 melanomas). Certain diagnostic features of malignant melanoma were identified allowing to differentiate melanoma from non-melanoma skin lesions with sensitivity of 80% and specificity of 83%^^^.

Mcintosh et al. performed in vivo reflectance spectroscopy (400-2500 nm) on skin neoplasms (both pigmented and non-pigmented lesions). Significant spectral differences (p < 0.05) were found between normal skin and skin neoplasms in several regions of the NIR spectrum^^^. Using a bundle of 30-fiber-optic (18 sources and 12 detectors) as a probe, Wallace et al. recorded the reflectance spectra (320 - UOOnm) from 121 melanocytic benign and malignant skin lesions ^ . Encouraging retrospective correlations with pathology were reported.^^^. However, the probe geometry used in their experimentation incorporated a wide range of source-detector separations, which would reduce sensitivity to scattering changes and primarily respond to absorption.

As for any imaging technique, the various enhanced imaging methods for identifying melanoma require sophisticated computational pattern-recognition technologies or the interpretation by clinical experts. Non-imaging spectral techniques that provide spectral features specific to melanoma may provide a better approach for disease detection in a non-specialist setting (e.g., the principle-care provider). A research team at University College London Hospital carried out studies on human subjects with the aim of demonstrating proof-of concept for the ESS technique in the diagnosis of pigmented skin lesions^^^ They used a probe geometry identical to that used in the preclinical trials conducted by A'Amar et al. (refer to subsection 16.4.3.). Figure 16.6 shows examples of elastic scattering spectra from normal skin, benign nevi and malignant melanoma. In humans there can be range of melanin content in skin^ . When the concentration of melanin is significant in benign conditions, typical ESS spectra, with the same geometry used in the animal model, will typically exhibit strong absorption due to melanin at the shorter wavelengths, with less evidence of enhanced scattering effects. However, preliminary results using UV-visible ESS produced spectral trends in human melanoma similar to


140

120

100 ]

80

c 60 H w UJ

40

20

- Malignant Melanoma Normal Skin

~ Malignant Melanoma • Benign Mole

^\vd

300 400 500 600 Wavelength [nm]

700 800

Figure 16.6: Elastic scattering spectra from normal skin, a dysplastic naevus, a benign nevus and malignant melanomas. Data is obtained from our collaborators at University College London Hospital.

those found in opossum melanomas, with enhanced scattering at the shorter wavelengths.

Although most results are still preliminary, and the data set is too small for statistical analysis, the interpretation of the spectra can be related to the optical processes of scattering and absorption, and to the optical geometry used to collect the ESS spectra. The spectral trends showed clear differences between melanoma and non-melanoma pigmented lesions. The analysis of all the spectral measurements discussed in this section suggests that ESS shows promise as an aid in clinical diagnosis of cutaneous pigmented lesions.

16.5. CORRECTION OF FLUORESCENCE IN TURBID MEDIA

Light-induced fluorescence spectroscopy and fluorescence imaging techniques have been suggested for non-invasive diagnosis of malignant melanoma. However, the reliability of such optical measurements can be seriously affected by spatial variations in the optical properties of the tissue that are not related to malignancies^^. As various fluorophores (endogenous and exogenous) may contribute simultaneously to the global fluorescence signal, and the fluorescence spectra are distorted by the scattering and absorption properties of the tissue, the real characteristics of each fluorophore cannot be accurately determined^^. Consequently, the identification of these fluorophores in tissue ca be quite ambiguous. This is mostly due to the superposition of


fluorescence emissions and the effect of optical properties of tissue^ ' . Both effects strongly modify the spectral shape of the intrinsic fluorescence. Time-resolved spectroscopy may provide a solution for the superposition effect by decomposing the spectra based on the fluorescence lifetime^^. However, this technique does not correct for the effects of tissue attenuation, which are due to scattering (morphology) and absorption (eg. hemoglobin, keratin and melanin).

Sterenborg et al. have suggested an approach to eliminate the effects of spatial variations in the optical properties of the tissue for use with fluorescence spectroscopy of pigmented lesions^^ (refer to paragraph 3.3.). Their concept is based on performing fluorescence and color measurements on moles and unpigmented control skin before and after topical application of ALA.

In other attempts to enhance the characterization of pathological tissue by fluorescence several analytical and empirical correction models have been proposed. They are mostly based on ESS or spectral diffuse reflectance. Using the spectral data provided by the ESS, the tissue spectral-convolution effect can be extracted and used to correct the fluorescence spectra ^ ' ^ ' " . As an example of these correction models. The authors used an approximation of Beer's law combined with ESS for computing the total relative attenuation coefficient of light in a turbid medium^^ . They derived the following formula from Beer's law for correcting the fluorescence:

F,(l) = . ''-W Qxp[a^'\og{a'R{Ji))] ^^^

where,

Ffn is the normalized fluorescence spectrum measured at the surface of the tissue (distorted);

a is a proportionality constant to account for the geometry; al is a proportionality constant related to the fluorescence intensity

^/

Ls is the average effective pathlength of the elastically scattered light between source and detector;

Lf is the pathlength of the fluorescence light (through the medium to the detector surface);

is the calibrated elastic scattering spectrum that represents the ratio of the

elastically scattered light intensity II to the incident light intensity Z .

The parameters a, ai and a2 are empirically determined.


5 0.8

S 0.6 z .§•

0.4 h

0.2

Single

r i i

1 1

1 r ' J

1 / 1 1 J 1 If

1 // 1 ' ff

1 1 11 If

\lH wj

1

-fiber probe

• J

/ / r /

r /

\ S

1

\ N \

V

1 1

Pure fluorescence

Distorted fluorescence 1

^sw. Corrected fluorescence xis ^ H

" \ V

^^ V ^ ^ -> \

V \ X N^X \ V 1

N NX^ A ^ V v /A

VyxK J

500 550 600 650 Wavelength [nm]

700 750

Figure 16.7. Fluorescence spectra of fluorescein (2.5%) in aqueous solution (intrinsic) and in a tissue phantom (distorted) composed of fluorescein (2.5%), Direct Blue absorber (|aa=ll cm' at 500 nm) and polystyrene particles (|as'=20 cm"', g=0,82 at 500 nm). The corrected spectrum is shown in the same figure. The measurements were made using a single-fiber probe. The squared estimation error between the corrected fluorescence spectrum and the intrinsic fluorescence spectrum was 2.8x10 .

Figure 16.7 shows a normalized intrinsic fluorescence spectrum Ft of fluorescein in an aqueous solution, a normalized distorted fluorescence spectrum for the same fluorophore in tissue phantom Fm and the normalized corrected fluorescence spectrum F^. The model permitted to correct the distortion of fluorescence spectra due to the turbidity of the medium. Such an approach can be implemented in fluorescence spectral and imaging techniques to account for the spatial variation due to tissue optical properties in order to enhance the detection features that are related to the tissue malignancies.

16.6. CONCLUSIONS

We have reviewed several optical approaches being studied for assessing melanoma by imaging, light-induced fluorescence spectroscopy and elastic scattering spectroscopy. Current fluorescence spectroscopy and imaging techniques are convenient for noninvasive diagnosis of certain cancers and non-cancer diseases. Although the skin is easily accessible by optical techniques, there are challenges and limitations to the application of fluorescence spectroscopy for distinguishing malignant from benign nevi. The development of new optical reporters for melanoma, with higher selectivity and reduced skin photosensitivity and toxicity may lead to reestablishing larger investigations for assessing pigmented skin lesions. It is important to acquire fluorescence signatures that are independent of optical properties of tissue for the elucidation of fluorescence diagnostic signatures.


Elastic scattering spectroscopy is a sensitive technique for detecting changes in sub-cellar morphology that accompany the transformation of normal cells to cancerous cells . Preliminary measurements have shown promise for the differentiation of benign, dysplastic nevi and malignant melanocytic lesions. The ultimate utility of ESS will require larger studies to benchmark the diagnostic accuracy histology, and then compare the accuracy with the clinical assessment of dermatologists and general practitioners, in order to establish the clinical benefits of ESS in the management of melanoma in humans. Larger scale studies are in progress.

Finally, the combination of ESS and LIFS might provide a better solution for diagnosing melanoma With a slight modification, one instrumentation can be used to acquire both ESS and LIFS measurements. Such technique will provides wider information about tissue morphology and biochemistry for enhancing the differentiation between benign nevi and malignant melanoma.

16.7. REFERENCES

1. R. M. MacKie, D. Hole, J. A. A. Hunter, R. Rankin, A. Evans, K. McLaren, M. Fallowfield, A. Hutcheon and A. Morris, Cutaneous malignant melanoma in Scotland, incidence, survival, and mortality, 1979-94. British MedicalJournal 315, 1117-1121 (1997).

2. G. G. Giles, B. K. Armstrong, R. C. Burton, M. P. Staples and V. J. Thursfield, Has mortality from melanoma stopped rising in Australia? Analysis of trends between 1931 and 1994, British MedicalJournal 312, 1121-1125 (1996).

3. J. F. Aitken, J. M. Elwood, J. B. Lowe, D. W. Firman, K. P. Balanda and L T. Ring, A randomised trial of population screening for melanoma. Journal of Medical Screening 9, 33-37 (2002).

4. R. Marks, Epidemiology of melanoma. Clinical and Experimental Dermatology 25, 459-463 (2000).

5. S. L. Edwards and K. Blessing, Problematic pigmented lesions, approach to diagnosis. Journal of Clinical Pathology 53(6), 409-418 (2000).

6. S. C. Chen, D. M. Bravata, E. Weil and L Olkin, A comparison of dermatologists' and primary care physicians' accuracy in diagnosing melanoma. Archives of Dermatology 137, 1627-1634 (2001).

7. M. Elbaum, A. W. Kopf, H. S. Rabinovitz, R.G. B. Langley, H. Kamino, M. C. Mihm, A. J. Sober, G. L. Peck, A. Bogdan, D. GutkowiczKrusin, M. Greenebaum, S. Keem, M. Oliviero and S. Wang, Automatic differentiation of melanoma from melanocytic nevi with multispectral digital dermoscopy, A feasibility study. Journal of the American Academy of Dermatology 44(2), 207-218 (2001).

8. A. Bono, S. Tomatis, C. Bartoli, G. Tragni, G. Radaelli, A. Maurichi and R. Marchesini, The ABCD system of melanoma detection, a spectrophotometric analysis of the asymmetry, border, colour, and dimension. Cancer 85, 72-7 (1999).

9. R. Marchesini , A. Bono, C. Bartoli, M. Lualdi, S. Tomatis and N. Cascinelli, Optical imaging and automated melanoma detection: questions and answers, Melanoma Research 12(3), 279-86 (2002).

10. A. Bono, S. Tomatis, C. Bartoli, N. Cascinelli, C. Clemente, C. Cupeta and R. Marchesini, The invisible colours of melanoma. A Telespectrophotometric diagnostic approach on pigmented skin lesions, European Journal of Cancer 32A(4), 727-729 (1996).

11. S. Tomatis, C. Bartoli, A. Bono, N. Cascinelli, C. Clemente and R. Marchesini, Spectrophotometric imaging of cutaneous pigmented lesions, discriminant analysis, optical


properties and histological characteristics, Journal of Photochemistry and Photobiology B, Biology 42(\), 32-39 il99S).

12. B. Farina, C. Bartoh, A. Bono, A. Colombo, M. Lualdi, G. Tragni, R. Marchesini, A. Colombo, M. Lualdi and G. Tragni, Multispectral imaging approach in the diagnosis of cutaneous melanoma, potentiality and limits, Physics in Medicine and Biology 45(5), 1243-54 (2000).

13. S. Tomatis, A. Bono, C. Bartoli, M. Carrara, M. Lualdi, G. Tragni and R. Marchesini, Automated melanoma detection: multispectral imaging and neural network approach for classification, Medical Physics 30(2):212-21 (2003).

14. E. Ruocco, G. Argenziano, G. Pellacani and S. Seidenari, Noninvasive imaging of skin tumors, Dermatologic Surgery 30,301-310 (2004).

15. S. W. Menzies, C. Ingvar, W. H. McCarthy, A sensitivity and specificity analysis of the surface microscopy features of invasive melanoma. Melanoma Research 6, 55-62 (1996).

16. S. W. Menzies, A. Gustenev, M. Avramidis, A. Batrac, W. H. McCarthy, Short-term digital surface microscopic monitoring of atypical or changing melanocytic lesions. Archives of Dermatology 137, 1583-1589 (2001).

17. P. Carli, V. De Giorgi, B. Giannotti, Dermoscopy and early diagnosis of melanoma, Archives of Dermatology 137, 1641-1644 (2001).

18. P. Anikijenko, L. T. Vo, E. R. Murr, J. Carrasco, W. J. McLaren, Q. Chen, S. G. Thomas, P. M. Delaney and R. G. King, In vivo detection of small subsurface melanomas in athymic mice using noninvasive fiber optic confocal imaging. Journal of Investigative Dermatology 117(6), 1442-1448(2001).

19. J. D. Whited and J. M. Grichnik, Clinical diagnosis of moles vs melanoma, Journal of American Medical Association 280(10), 881-882 (1998).

20. G. J. Teamey, M E. Brezinski, B. E. Bouma, S. A. Boppart, C. Pitris, J. F. Southern and J. G. Fujimoto, In vivo endoscopic optical biopsy with optical coherence tomography. Science 276(3321), 2037-2039 (1997).

21. S. B. Colak, M. B. van der Mark, G. W.'t Hooft, J. H. Hoogenraad, E. S. van der Linden and F. A. Kuijpers, Clinical optical tomography and NIR spectroscopy for breast cancer detection, IEEE Journal of Selected Topics in Quantum Electronics 5(4), 1143-1158 (1999).

22. S. Jackie, N. Gladkova, F. Fledchtein, A. Terentieva, B. Brand G. Gelikonov, V. Gelikonov, A. Sergeev, A. Fritscher-Ravens, J. Freund, U. Seitz S. Schroder and N. Soehendra, In vivo endoscopic optical coherence tomography of esophagitis, Barrett's esophagus, and adenocarcinoma of the esophagus. Endoscopy 32(10), 750-755 (2000).

23. H. Liu, D. Boas, Y. Zhang and A. G. Yodh, B. Chance, Determination of optical properties and blood oxygenation in tissue using continuous NIR light, Physics in Medicine and Biology 40(11), 1983-1993(1995).

24. M. J. Holboke B. Tromberg, X. Li, N. Shah, J. Fishkin, D. Kidney, J. Butler, B. Chance and A. G. Yodh, Three-dimensional diffuse optical mammography with ultrasound localization in a human subject. Journal of Biomedical Optics 5(2), 237-247 (2000).

25. A. E. Cerussi, D. Jakubowski, N. Shah, F. Bevilacqua, R. Lanning, A. J. Berger, D. Hsiang, J. Butler, R. F. Holocombe and B. J. Tromberg, Spectroscopy enhances the information content of optical mammography Journal of Biomedical Optics 7(1), 60-71 (2002).

26. G. C. Tang, A. Pradhan and R. R. Alfano, Spectroscopic ditTerences between human cancer and normal lung and breast tissues. Lasers in Surgery and Medicine 9, 290-295 (1989).

27. N. N. Boustany, R. Manoharan, R. R. Dasari and M. S. Feld , Analysis of normal and diseased colon mucosa using ultraviolet resonance Raman spectroscopy. Advances in Laser and Light Spectroscopy to Diagnose Cancer and Other Diseases III, Optical Biopsy, Proc. SPIE 2619, 66-70(1996).

28. M. Gniadecka, P. Philipsen, L. Hansen, J. Hercogova, K. Rossen, H. Thomsen and H. Wulf, Malignant melanoma diagnosis by Raman spectroscopy and artificial neural network, Journal of Investigative Dermatology 113(3), 464 (1999).

29. M. Rajadhyaksha, M. Grossman, D. Esterowitz, R. H. Webb and R. R. Anderson, In vivo confocal scanning laser microscopy of human skin, melanin provides strong contrast Journal of Investigative Dermatology 104 (6), 946-952 (1999).


30. M. Rajadhyaksha, S. Gonzalez, J. M. Zavislan, R. R. Anderson and R. H. Webb, In Vivo Confocal Scanning Laser Microscopy of Human Skin II, Advances in Instrumentation and Comparison With Histology Journal of Investigative Dermatology 113 (3), 293 -303 (1999).

31. E. Bossu, O. A'Amar, R. M. Parache, D. Notter, P. Labrude, C. Vigneron and F. Guillemin, Determination of the maximal tumor/normal skin ratio after HpD or m-THPC administration in hairless mouse (Skh-1) by fluorescence spectroscopy, a non invasive method, Anti-Cancer Drugs H, 67-12 (1991).

32. H. Rezzoug, L. Bezdetnaya, O. A'Amar, M. Barberi-Heyob, J. L. Merlin and F. Guillemin, Parameters affecting photodynamic activity of Foscan® or meta-tetra (hydroxyphenyl) chlorin (mTHPC) in vitro and in vivo, Lasers in Medical Science 13(2), 119-125 (1998).

33. E. Bossu, J. J. Padilla, O. A'Amar, R. M. Parache, D. Notter, C. Vigneron and F. Guillemin, Determination of endogenous porphyrins and the maximal HpD tumor/normal skin ratio in SKH-1 hairless mice by light induced fluorescence spectroscopy, Artificial Cells Blood Substitutes and Immobilization Biotechnology 27(2), 109-117 (1999).

34. S. Andersson-Engels and B. C. Wilson In vivo fluorescence in clinical oncology. Fundamental and practical issues. Journal of Cell Pharmacology 3, 66-79 (1992).

35. I. J. Bigio and J. R. Mourant. Ultraviolet and visible spectroscopies for tissue diagnostics, fluorescence spectroscopy and elastic-scattering spectroscopy. Physics in Medicine and 5/Wo^ 42, 803-813, (1997).

36. G. A. Wagnieres, W. M. Star and B. C. Wilson, In vivo fluorescence spectroscopy and imaging for oncological applications, Photochemistry andPhotobiology 68(5), 603-632 (1998).

37. N. Ramanujam, Fluorescence spectroscopy in vivo. Encyclopedia of Analytical Chemistry, R A Meyers(Ed.), pp. 20-56, John Wiley & Sons Ltd, Chichester (2000).

38. S. Andersson-Engels, C. Klinteberg, K. Svanberg and S. Svanberg, In vivo fluorescence imaging for tissue diagnostics, Phys. Med. Biol. 42, 815-824 (1997).

39. T. Sama and H. M. Swartz, The physical properties of melanins, in The Pigmentary System, ed. JJ Nordlund et al., Oxford University Press, (1988).

40.1. A. Vitkin, J. Woolsey, B. C. Wilson and R. R. Anderson, Optical and thermal characterization of natural (Spia officinalis) melanin. Photochemistry and photobiology 59(4), 455-462 (1994).

41. T. Sama, H. M. Swartz, The physical properties of melanins, in The Pigmentary System, ed. JJ Nordlund et a l , Oxford University Press, (1988).

42. O. A'Amar, R. D. Ley and I. J. Bigio, Comparison between UV-visible and NIR elastic-scattering spectroscopy of chemically induced melanomas in an animal model. Journal of Biomedical optics 9(6), 1320-1326 (2004).

43. M. L. Wolbarsht, A. W. Walsh and G George, Melanin, a unique biological absorber, Applied Op^/c^ 20, 2184-6 (1981).

44. G. Zonios, L. T. Perelman, V. Backman, R. Manoharan, A. Nusrat, S. Shields, M. Fitzmaurice, J. Can Dam and M. S. Feld, Diffuse reflectance spectroscopy of human adenomatous colon polyps in vivo. Applied Optics 38, 6628-6637 (1999).

45. J. D. Pfeifer, D. A. Hill, M. J. O'Sullivan and L. P. Dehner, Diagnostic gold standard for soft tissue tumours: morphology or molecular genetics?, Histopathology 37, 485-500 (2000).

46. J. C. E. Underwood, General and systemic pathology, Chap. 24 Skin, Underwood JCE, Eds., pp. 667-697, Churchill Livingstone (2000).

47. R. Drezek, A. Dunn and R. Richards-Kortum, Light scattering from cells, finite-difference time-domain simulations and goniometric measurements. Applied Optics 38(16), 3651-3661 (1999).

48. L. T. Perelman, V. Backman, M. Wallace, G. Zonios, R. Manoharan, A. Nusrat, S. Shields, M. Seller, C. Lima, T. Hamano, I. Itzkan, J. Can Dam, J. M. Crawford and M. S. Feld, Observation of periodic fine structure in reflectance from biological tissue, a new technique for measuring nuclear size distribution. Physical Review Letters 80, 627-630 (1998).

49. M. Osawa and S. Niwa, A portable diffuse reflectance spectrophotometer for rapid and automatic measurementof tissue, Measurement Science and Technology 4, 668-76 (1993).

50. D. Elder, Tumor Progression, Early Diagnosis and Prognosis of Melanoma, Acta Oncologica 38(5), 535-548(1999).


51. R. R. Mehta, L. Bratescu, J. M. Graves, A. Shilkaitis, A. Green, R. G. Mehta, T. K. Das Gupta, In vitro transformation of human congenital naevus to malignant melanoma, Melanoma Research 12(1), 27-33 (2002).

52. A. Jackson, C. Wilkinson, M. Ranger, R. Pill and P. august. Can primary prevention or selective screening for melanoma be more precisely targeted through general practice? A prospective study to validate a self administered risk score, British Medical Journal 316, 7124-7134 (1998).

53. J. F. Aitken, J. M. Elwood, J. B. Lowe, D. W. Firman, K. P. Balanda and I. T. Ring, A randomised trial of population screening for melanoma, Journal of Medical Screening 9, 33-37 (2002).

54. C. M. Balch, S.-J. Soong, J. E. Gershenwald, J. F. Thompson, D. S. Reintgen, N. Cascinelli, M. Urist, K. M. McMasters, M. I. Ross, J. M. Kirkwood, M. B. Atkins, J. A. Thompson, D, G. Coit, D. Byrd, R. Desmond, Y. Zhang, P.-Y. Liu and G. H. Lyman, A Morabito, Prognostic factors analysis of 17,600 Melanoma patients, validation of the American Joint Committee on Cancer Melanoma Staging System, Journal of Clinical Oncology 19(16), 3622-3634 (2001).

55. J. J. Scarisbrick, C. D. O. Pickard, A. C. Lee, G. M. Briggs, K. Johnson, S. G. Bown, M. Novelli, M. R. S. Keshtgar, I. J. Bigio and R. Yu, Elastic scattering spectroscopy in the diagnosis of pigmented lesions. Comparison with clinical and histopathological diagnosis. Diagnostic Optical Spectroscopy 11, Proc. SPIE 5141 (2003).

56. H. Y. Thong, S. H. Jee, C. C. Sun and R. E. Biossy, The patems of melanosomes distribution in keratinocytes of human skin as one determining factor of skin color, British Journal of Dermatology, 149, 498-505 (2003).

57. N. Nagai, Y. J. Lee, N. Nagaoka, M. Gunduz, K. Nojima, H. Tsujigiwa, E. Gunduz, C. H. Siar and H. Nagatsuka, Elemental sulphur and alkali elutable melanin detected in oral melanosis and malignant melanoma by energy-filtering transmission electron microscopy. Journal of Oral Pathology & Medicine 31, 481-487 (2002).

58. A. E. Profio and O. J. Balchum, Fluorescence diagnosis of cancer Adv. Exp. Med. Biol. 193,43-50(1985).

59. D Kessel, Tumor locaHzation and photosensitization by derivatives of hematoporphyrin, a review, IEEE Journal of Quantum Electron 23, 1718 20 (1987)

60. G. A. Wagnieres, W. M. Star and B. C. Wilson, In vivo fluorescence spectroscopy and imaging for oncological applications, Photochemistry andPhotobiology, 68 (5) 603-632 (1998).

61. Cheong W, Prahl S A, Welch A J, A review of the optical properties of biological tissues, IEEE Journal of Quantum Electron. 26,2166-85 (1990)

62. K. T. Schomacker, J. K. Frisoli J., C. Compton, T. J. Flotte, J. M. Richter, N. S. Nishioka and T. F. Deutsch, Ultraviolet laser-induced fluorescence of colonic tissue: Basic biology and diagnostic potential. Lasers in Surgery and Medicine 12, 63-78 (1992)

63. O. A'Amar, D. Lignon, H. Begorre, F. Guillemin and E. Yvroud, Gaussian Analyzing Model of Tissue Light-Induced Fluorescence for the Detection of Early Cancers, Optical Tomography and Spectroscopy of Tissue: Theory, Instrumentation, Model, and Human Studies II, Proc. SPIE 2979, 574-584 (1997).

64. O. A'Amar, D. Lignon, O. Menard, H. Begorre, F. Guillemin and E. Yvroud, Autofluorescence spectroscopy of normal and malignant tissues, both in vivo and ex-vivo measurements in the upper aero-digestive tract and lung tissues, Proc. SPIE: Advances in Laser and Light Spectroscopy to Diagnose Cancers and Other Diseases III, Optical Biopsy 2679, 42-50, 1996.

65. M. Panjehpour, B. F. Overholt, J. L. Schmidhammer, C. Farris, P. F. Buckley and T. Vo-Dinh, Spectroscopic diagnosis of esophageal cancer, new classification model, improved measurement system. Gastrointestinal Endoscopy 41(6), 577-581.

66. K. K. Wang, K. Gutta, M. Laukka and J. Densmore, Laser induced fluorescence in the detection of esophageal carcinoma, Proc. SPIE: Optical Biopsy and Fluorescence Spectroscopy and Imaging 2324, 14-18, 1994.

67. T. Vo-Dinh, M. Panjehpour, B. F. Overholt, C. Farris, F. B. Buckley III and R. Sneed, In vivo cancer diagnosis of the esophagus using differential normalized fluorescence (DNF) indices, Lasers in Surgery and Medicine 16, 41-47, 1995.


68. A. E. Profio, D. R. Doiron, O. J. Balchum and G. C. Huth, Fluorescence bronchoscopy for localization of carcinoma in situ, Medical Physics 10,35-9 (1983).

69. J. Hung, S. Lam, J.-C. Leriche and B. Palcic, Auto fluorescence of normal and malignant bronchial tissue, Laser in Surgery and Medicine 11, 99-105, 1991.

70. S. Lam, C. MacAulay, J. LeRiche, N. Ikeda and B. Palcic, Fluorescence imaging of early lung cancer, Proc. SPIE, Optical Biopsy and Fluorescence Spectroscopy and Imaging 1^1^, 2-8, 1994.

71. G. C. Tang, A. Pradhan and R. R. Alfano, Spectroscopic differences between human cancer and normal lung and breast tissues, Lasers in Surgery and Medicine 9, 290-295, 1989.

72. R. Marchesini, M. Brambilla , E. Pignoli, G. Bottiroli, A. Cleta Croce, M. Dal Fante, P. Spinelli and S. Di Palma, Light-induced fluorescence spectroscopy of adenomas, adenocarcinoma and non-neoplastic mucosa in human colon. Journal of Photochemistry and Photobiology B: jg/o/ogv, 14,219-230(1992).

73. R. Richard-Kortum, R. P. Rava, R. E. Petras, M. Fitzmaaurice, M, Sivak and M. S. Feld, Spectroscopy diagnosis of colonic dysplasia, Photochemistry and Photobiology, 53(6), 777-786(1991).

74. G. Zonios, R. M. Cothem, J. Arendy, J. Wu, J. M. Crawford, J. Van Dam, R. Manoharan and M. S. Feld, Fluorescence spectroscopy for colon cancer diagnosis, Proc. SPIE: Optical Biopsy and Fluorescence Spectroscopy and Imaging 2324, 9-13 (1994).

75. T. Vo-Dinh, M. Panjehpour, B. F. Overholt, F. B. Buckley III and D. M. Edwards, Detection of colon malignancy using differential normalized fluorescence, in Advances in laser and light spectroscopy to diagnose cancer and other diseases III .Optical Biopsy, Proc. SPIE 2679, 26-33(1996).

76. W. Lohmann, J. Mussmann, C. Lohmann and W. Kunzel, Native fluorescence of the cervix uteri as a marker for dysplasia and invasive carcinoma, European Journal of Obstetrics & Gynecology an Reproductive Biology 31, 249-53 (1989).

77. W. S. Glassman, C. H. Liu, G. C. Tang, S. Lubicz and R. R. Alfano, Ultraviolet excited fluorescence spectra from non-malignant and malignant tissues of the gynecological tract. Laser in the Life Sciences, 1(2) 49-58, (1992)

78. A. Mahadevan, M. F. Mitchell, E. Silva, S. Thomsen and R. Richard-Kortum, Study of the fluorescence properties of normal and neoplastic human cervical tissue. Laser in Surgery and Medicine, 13, 647-665 (1993).

79. M. F. Mitchell, S. B. Cantor, N. Ramanujam, G. Tortolero and R. Richards-Kortum, Fluorescence spectroscopy for diagnosis of squamous intraepithelial lesions of the cervix. Obstetrics & Gynecology 93(3), 462-470 (1999).

80. Richards-Kortum R, Mitchell M F, Ramanujam N, Mahadevan A and Thomsem S In vivo fluorescence spectroscopy: potential for non-invasive automated diagnosis of cervical intraepithelial diagnosis of cervical intraepithelial neoplasia and use as a surrogate endpoint biomarker. Journal of Cell Science Supplement 19, 111-119 (1994).

81. M. Panjehpour, C. E. Julius, M. N. Phan, T. Vo-Dinh and S. Overholt, Laser-induced fluorescence spectroscopy for in vivo diagnosis of non-melanoma skin cancers, Lasers in Surgery and Medicine, 31(5), 367-73 (2002).

82. Bottiroli G., Balzarini P. and A. C. Croce, Autofluorescence properties of colonic mucosa: dependence on excitation wavelength, Proc. SPIE 2927, 173-179 (1996).

83. N. Ramanujan, M. F. Mitchell, A. Mahadeevan, S. Warren, S. Thomsen, E. Silva and R. Richards-Kortum ,In vivo diagnosis of cervical intraepithelial neoplasia using 337-nm-excited laser-induced fluorescence, Proc. National Academy of Sciences USA 91, 10 193-197 (1994).

84. R. Richards-Kortum, R. P. Rava, R. E. Petras, M. Fitzmaurice, M. Sivak and M. S. Feld, Spectroscopic diagnosis of colonic dysplasia. Photochemistry Photobiology 53, 777-86 (1991).

85. W. Lohmann and E. Paul, In situ detection of melanomas by fluorescence measurements, Naturwissenschaften 75, 201-2 (1988).

86. W. Lohmann and E. Paul Native fluorescence of unstained cryo-section of the skin and melanomas and nevi Naturwissenschaften 76, 424-426 (1989).


87. W. Lohmann, M. Nilles and R. H. Bodeker, In situ differentiation between nevi and malignant melanomas by fluorescence measurements, Naturwissenschaften 78, 456-457 (1991).

88. D. J. Leffell and M. L. Stetz, In vivo fluorescence of human skin, a potential marker of photoaging, Archives of Dermatology 124(10), 1514-18 (1988)

89. H. J. C. M. Sterenborg, M. Motamedi, R. F. Wagner, M. Duvic, S. Thomsen and S. L. Jacques, In vivo fluorescence spectroscopy and imaging of human skin tumors, Lasers Medical Science 9 191-201 (1994).

90. H. J. C. M Sterenborg, M. Motamedi, R. F. Wagner, S. Thomsen and S. L. Jacques In vivo fluorescence spectroscopy for the diagnosis of skin diseases, Proc. SPIE 2324 32-37 (1994).

91. N. Kollias, R. Gillies, M. Moran, I. E. Kochevar and R. R. Anderson, Endogenous Skin Fluorescence Includes Bands that may Serve as Quantitative Markers of Aging and Photoaging, Journal of Investigative Dermatology 111(5), 776 (1998)

92. R. Gillies, G. Zonios, R. R. Anderson and N. Kollias, Fluorescence excitation spectroscopy provides information about human skin in vivo, Journal of Investigative Dermatology 115 (4), 704 -(2000).

93. L. Brancaleon,, A. J. Durkin., J. H. Tu, G. Menaker, J. D. Fallon and N. Kollias, In vivo Fluorescence Spectroscopy of Nonmelanoma Skin Cancer, Photochemistry and Photobiology 73(2), 178-183(2001).

94. B. W. Chwirot, S. Chwirot, J. Redzinski and Z. Michniewicz, Detection of melanomas by digital imaging of spectrally resolved ultraviolet light-induced autofluorescence of human skin, European Journal of Cancer 34(11), 1730-1734 (1998).

95. B. W. Chwirot, S. Chwirot, N. Sypniewska, Z. Michniewicz, J. Redzinski, G. Kurzawski and W. Ruka, Fluorescence In Situ Detection of Human Cutaneous Melanoma, Study of Diagnostic Parameters of the Method, Journal of Investigative Dermatology 117(6), 1449-1451(2001)

96. B. W. Chwirot, N. Sypniewska and J. Swiatlak, Diagnostic potential of fluorescence of formalin-fixed paraffin-embedded malignant melanoma and pigmented skin lesions, quantitative study of fluorescence intensity using fluorescence microscope and digital imaging. Melanoma Research 11(6), 569-76 (2001).

97. H. J. Sterenborg, A. E. Saamak, R. Frank, M.Motamedi, Evaluation of spectral correction techniques for fluorescence measurements on pigmented lesions in vivo, Photochemistry Photobiology B 35(3): 159-65 (1996).

98. H. Rezzoug, O. A'Amar, L. Bezdetnaya and F.Guillemin, In vivo pharmacokinetics and photodynamic effect of Foscan® (mTHPC) in tumor tissue of mice bearing human colon adenocarcinoma. First Internet Conference, Photochemistry and Photobiology, C21 http://www.photobiologv.com/vl/rezzoug/pub.htm, (1997).

99. M. F. Grahn, M. L. De Jode, M. G. Dilkes, J. K. Ansell, D. Onwu, J. Maudsley and N. S. Williams, Tissue photosensitizer detection by low-power remittance fluorimetry. Lasers in Medical Sciences 12, 245-252 (1997).

100. S. Andersson-Engels, R. Berg, K. Svanberg, S. Svanberg,Multi-colour fluorescence imaging in connection with photodynamic therapy of -amino levulinic acid (ALA) sensitised skin malignancies, Bioimiging3, 134-143 (1995).

101. E. V. Ross, R. Romero, N. Kollias, C. Crum and R. R. Anderson, Selectivity of protoporphyrin IX fluorescence for condylomata after topical application of 5-aminolaevulinic acid: implications for photodynamic treatment, British Journal of Dermatology 137(5):736-42 (1997).

102. S. H. Ibbotson, Topical 5-aminolaevulinic acid photodynamic therapy for the treatment of skin conditions other than non-melanoma skin cancer, British Journal of Dermatology 146(2), 178-88 (2002).

103. C. Clark, A. Bryden, R. Dawe, H. Moseley, J. Ferguson and S. H. Ibbotson, Topical 5-aminolaevulinic acid photodynamic therapy for cutaneous lesions: outcome and comparison of light sources, Photodermatol Photoimmunol Photomed. 19(3): 134-141 (2003).

104. B. C. Wilson, Measurement of tissue optical properties: methods and theories, in Optical-Thermal Response of Laser-Irradiated Tissue eds. A. J. Welch, and M. J. C. van Gemert (Plenum, New York) 233-74 (1995).


105. J. Mourant, J Freyer, A Hielscher, A Eick, D Shen, T Johnson, Mechanisms of light scattering from biological cells relevant to noninvasive optical-tissue diagnostics, Applied Optics 37, 3586-3593(1998).

106. R. Marchesini, M. Brambilla, C. Clemente, M. Maniezzo, A. E. Sichirollo, A. Testori, D. R. Venturoli and N. Cascinelli, In vivo spectrophotometric evaluation of neoplastic and nonneoplastic skin pigmented lesions~I. Reflectance measurements. Photochemistry and Photobiology 53(1), 77-84 (1991).

107. D. Rallan, C. C. Harland, Skin imaging: is it clinically useful?, Clinical and Experimental Dermatology, 29(5):453-9 (2004).

108. J. Boyer, J. R. Mourant and I. J. Bigio, Monte Carlo investigations of elastic scattering spectroscopy applied to latex spheres used as tissue phantoms, Proc. SPIE 2389, 103-12 (1995).

109. I. J. Bigio, J. R. Mourant, J. Boyer, T. Johnson and T. Shimada, Noninvasive identification of bladder-cancer with sub-surface backscattered light, Proc. SPIE 2135 26-35 (1994).

110. L. G. Henyey and Greenstein JL, Astrophysics Journal 93, 70 (1941). 111.1. J. Bigio, S. G. Bown, G. Briggs, C. Kelley, S. Lakhani, D. Pickard, P. M. Ripley, I. Rose and

C. Saunders, Diagnosis of Breast Cancer using Elastic-Scattering Spectroscopy, Preliminary Clinical Results, Journal of Biomedical Optics 5, 221-228 (2000).

112. J. R. Mourant, I. J. Bigio, J. Boyer, T. M. Johnson, J. A. Lacey, A. G. Bohorfoush, and M. Mellow, Elastic-scattering spectroscopy as a diagnostic tool for differentiating pathologies in the gastrointestinal tract, preliminary testing. Journal of Biomedical Optics 1(2), 192-199 (1996).

113. D. C. Pickard, L. B. Lovat, M. Novelli, P. M. Ripley, C. Kelly, I. J. Bigio and S. G. Bown, Diagnosis of dysplasia in Barrett's oesophagus with in-situ elastic-scattering spectroscopy. Optical Biopsy and Tissue Optics, Proc. SPIE 4161, 122-130 (2000).

114. J. R. Mourant, I. J. Bigio, J. Boyer, R. L. Conn, T. M. Johnson and T. Shimada, Spectroscopic diagnosis of bladder cancer with elastic light scattering. Lasers in Surgery and Medicine 17, 350-357(1995).

115. M. L. Wolbarsht, A. W. Walsh and G. George, Melanin, a unique biological absorber, Applied Optics 20, 2184-2186 (1981).

116. R. Marchesini, C. Clemente, E. Pignoli and M. Brambilla, Optical properties of in vitro epidermis and their possible relationship with optical properties of in vivo skin. Journal of Photochemistry and Photobiology B, Biology 16(2), 127-140 (1992).

117. G. Zonios, J. Bykowski and N. Kollias. Skin melanin, hemoglobin and Hght scattering properties can be quantitatively assessed in vivo using diffuse reflectance spectroscopy, Journal of Investigative Dermatology 117(6), 1452-1457 (2001).

118 M. Canpolat and J. R. Mourant, Particle size analysis of turbid media with a single optical fiber in contact with the medium to deliver and collect white light, Applied Optics 40, 3792-3799(2001).

119 A. Amelink, M. P. L. Bard, S. A. Burgers, H. J. C. M. Sterenborg, Single-scattering spectroscopy for the endoscopic analysis of particle size in superficial layers of turbid media, Applied Optics 42, 4095-4101 (2003).

120 V. Backman, R. Gurjar, K. Badizadegan, I. Itzkan, R. RO Dasari and L. T. Perelman, MS Feld, Polarized light scattering spectroscopy for quantitative measurement of epithelial cellular structure in situ, IEEE Selected Topics in Quantum Electrons 5, 1019-1026 (1999).

121 K. Sokolov, R. Drezek, K. Gossage and R. Richards-Kortum, Reflectance spectroscopy with polarized light: is it sensitive to cellular and nuclear morphology. Optics Express 5, 302-317 (1999).

122 J. R. Mourant, T. M. Johnson, S. Carpenter, A. Guerra, T. Alda and J. P. Freyer, Polarized angular dependent spectroscopy of epithelial cells and epithelial cell nuclei to determine the size scale of scattering structures. Journal of Biomedical Optics 7, 378-387 (2002).

123. J. R. Mourant, J. Boyer, A. H. Hielscher and I. J. Bigio Influence of the scattering phase function on light transport measurements in turbid media performed with small source-detector separations. Optics Letters 21(7), 546-548 (1996).


124. D. F. Kusewitt, L. A. Applegate, C. D. Bucana and R. D. Ley, Naturally occurring malignant melanoma in the South American opossum (Monodelphis domestica), Veterinary Pathology 27, 66-68 (1990).

125. R. Marchesini, N. Cascinelli, M. Brambilla, C. Clemente, L. Mascheroni, E. Pignoli, A. Testori and D. R. Ventroli, In vivo spectrophotometric evalulation of neoplastic and non-neoplastic skin pigmented lesions. II: discriminant analysis between nevus and melanoma, Photochemistry and Photobiology 55(4), 515-522 (1992).

126. R. Marchesini, S. Tomatis, C. BartoH, A. Bono, C. Clemente, C. Cupeta, I. Del Prato, E. Pignoli, SichiroUo AE and Cascinelli N., In vivo spectrophotometric evaluation of neoplastic and non-neoplastic skin pigmented lesions. III. CCD camera-based reflectance imaging, Photochemistry and Photobiology 62(1),151-154 (1995).

127. M. Moncrieff, S. Cotton, E. Claridge, P. Hall, Spectrophotometric intracutaneous analysis: a new technique for imaging pigmented skin lesions, British Journal of Dermatology 146(3):448-457 (2002).

128. L. M. Mcintosh, R. Summers, M. Jackson, H. H. Mantsch, J. R. Mansfield, M. Howlett, A. N. Crowson and J. W. P. Toole. Towards non-invasive screening of skin lesions by near-infrared spQctroscopy, Journal of Investigative Dermatology 116, 175-181 (2001).

129. V. P. Wallace, D. C. Crawford, P. S. Mortimer, R. J. Ott and J. C. Bamber, Spectrophotometric assessment of pigmented skin lesions, methods and feature selection for evaluation of diagnostic performance, Physics in Medicine and Biology 45(3), 735-51 (2000).

130. V.P Wallace, JC Bamber, D. C. Crawford, R. J. Ott and P. S. Mortimer, Classification of reflectance spectra from pigmented skin lesions, a comparison of multivariate discriminant analysis and artificial neural networks. Physics in Medicine and Biology 45(10), 2859-2871 (2000).

131. J. J. Scarisbrick, C. D. O. Pickard, A. C. Lee, G. M. Briggs, K. Johnson, S. G. Bown, M. Novelli, M. R. S. Keshtgar and I. J. Bigio, R. Yu, Elastic scattering spectroscopy in the diagnosis of pigmented lesions: Comparison with clinical and histopathological diagnosis, Diagnostic Optical Spectroscopy II, Proc. SPIE 5U\ (2003).

132. N. N. Zhadin and R. R. Alfano, Correction of the internal absorption effect in fluorescence emission and excitation spectra from absorbing and highly scattering media: theory and experiment. Journal of biomedical optics, 3 (2), 171 -186 (1998).

133. Q. Zhang, M. G. MuUere, J. Wu and M. S. Feld, Turbidity-free fluorescence spectroscopy of biological tissue. Optics letters 25 (19), 1451-1453 (2000).

134. J. Y. Qu and J. Hua, Calibrated fluorescence imaging of tissue in vivo, Applied Optics 78 (25), 4040-4042(2001).

135. O. M. A'Amar and I. J. Bigio, Correction of fluorescence spectra using data from elastic scattering spectroscopy and a modified Beer's law. Optical Society of America Topical Meetings, Miami (2002).

QUANTITATIVE FLUORESCENCE HYBRIDIZATION USING AUTOMATED IMAGE

CYTOMETRY ON INTERPHASE NUCLEI

Applications on chromosome imbalances

Khuong Truong, Anne Gibaud, Nicolas Vogt, and Bernard Malfoy^

17.1. INTRODUCTION

Chromosome imbalances are the hallmark of malignancies in several human pathologies. In contrast to normal cells, alterations of the amount of chromosomes or chromosome arms (chromosome aneuploidy) are consistently observed in virtually all cancers [1]. There is much controversy about the cause and effect with regard to malignancy [ 2 - 4 ] . However, evidence such as specific gains or losses of chromosomes segments within specific tumour types, presence of aneuploidy in various pre-neoplastic conditions and increased frequency of genetic instability in aneuploid cell lines compared to diploid cells suggest that this phenomenon possibly plays an active role in carcinogenesis. Whatever their function may be, the presence of chromosome imbalances in somatic tissue, is a strong diagnostic indication for malignant transformation. Dramatic consequences of chromosome imbalances are also well established for several constitutional diseases. For example, trisomy of chromosome 21 (Down's syndrome) is the most frequently observed aberration in newborns [5]. Thus, the detection of numerical chromosome aberrations remains the main purpose of prenatal diagnosis. Several methods have been developed to detect chromosome number abnormalities in situ using fluorescence-based approaches. Data acquisition is currently performed by human observers using

Khuong Truong, IMSTAR. Paris France. Anne Gibaud, Nicolas Vogt, Bernard Malfoy. Institut Curie-CNRS-UPMC UMR7147. Institut Curie, 26 Rue d'Ulm 75248 Paris Cedex 5 France.

387

388 K. TRUONG ETAL.

digital cameras for image recording and visual counting of the signals. In order to improve the reliability of these molecular cytogenetic approaches, we have developed a quantitative two-colour fluorescence in situ hybridization (FISH) method based on differential chromosome or chromosome arm paintings for detection of imbalances on interphase nuclei using automated image cytometry.

17.2. CHROMOSOME IMBALANCES IN HUMAN DISEASES.

17.2.1. Cancers.

Considerable efforts have been devoted to characterize biological indicators of human cancers with relevance to clinical outcome. Cancer cells can be characterized by their genetic alterations having initiated the evolution from normal to malignant cells. Such alterations can be reflected either at the molecular or chromosomal level. In the latter case, the analysis of karyotypes reveales chromosome aberrations with gains and losses of chromosome segments leading to chromosome aneuploidy. Variations in the number of gene copies following an aneuploidy may induce overexpression (oncogenes, receptors,) or underexpression (suppressor genes) of genes at the origin of the tumour development. Thus, specific chromosome gains and losses may indicate how to establish tumour classifications for specific therapies. In human solid tumours, specific chromosome aberrations were particularly well characterized in breast and lung cancers.

The most recurrent clonal alterations within breast cancers affect chromosome 1, with principally gain of the long arm (Iq) and loss of the short arm (Ip) leading to Iq/lp imbalance [6,7]. These aberrations are regarded as early events in tumorigenesis resulting from various rearrangements between chromosomes 1 and/or 16 [8,9]. Bronchic cancers can be divided into two categories: small-cell lung cancer (SCLC) and non-small-cell lung cancer (NSCLC) [10]. The short arm of chromosome 3 is partially or totally lost in more than 90% of SCLC and 50-80% of NSCLC [11,12]. This loss is considered as an early event in the tumour progression since it is also observed in pre-neoplastic lesions [13,14]. In addition, recurrent gains of the long arm of chromosome 3 are also observed. In particular, in about 50% of NSCLC cases, gains in the whole 3q arm are found [15,16]. Thus, loss of 3p associated with gain of 3q leads to an imbalance between the long and short arms within many lung tumours.

17.2.2. Constitutional diseases

Children affected by chromosome trisomy usually have a range of birth defects, including delayed development and intellectual disabilities. Apart from Down's syndrome (trisomy 21), trisomy 13 and trisomy 18 allow live bom children [17 - 20]. However, their incidence remain lower, about one in every 5000 or 3000 births for trisomy 13 and 18, respectively, and one in 1000 births for trisomy 21. The presence of an extra copy of genes localized on the third chromosome leads to overexpression of some genes involved in the child's

QUANTITATIVE FLUORESCENCE HYBRIDIZATION 389

development. It is estimated that only a small percentage of these presently unidentified genes, may be involved in producing the symptoms of Down syndrome. The addition of an extra chromosome usually occurs spontaneously during conception. The cause of this is unknown and therapy remains difficult [5]. Prenatal diagnosis is the only way to provide informed consent to pregnant women.

17.3. EXPERIMENTAL APPROACHES FOR THE IN SITU DETERMINATION OF CHROMOSOME IMBALANCES.

17.3.L Metaphase chromosomes.

The 23 human chromosome pairs within metaphase are formed during the cell division by condensation of interphase chromatin. For experimental purposes, they are generally obtained by growing in vitro cells coming from in vivo material (tumour fragments or embryonic tissues). Preparations are spread out on slides and treated in classical cytogenetics to develop chromosome specific banding patterns (non fluorescent method) or processed using FISH (Fluorescence in situ hybridization, fluorescent method). Classical cytogenetic analysis of G or R-banded metaphase chromosomes has extensively been used for analysis of tumour aneuploidy [21,22] and remains the standard test for identification of trisomies in prenatal diagnosis [23,24]. However, its level of resolution is low and the use requires intervention of highly trained experts. Molecular cytogenetics has opened a new way for analysis of chromosome abnormalities, improving sensitivity of conventional banding analysis. In FISH, a DNA probe is labelled with chemically modified nucleotides. The modified probe hybridizes specifically to the homologous sequence on the spread chromosomes. Probes labelled with fluorescent adducts may be directly detected. Altematively, probes are localized using fluorescent labelled antibodies specific of adducts. Using adducts with different emission or excitation wavelengths allow simultaneous detection and localization of several probes on the same preparation. Several probe families of various sizes can be used. They range from a DNA fragment of a few thousand of base pairs specific of a single gene, to complex mixtures (painting) covering a whole chromosome or chromosome arm (review in [25 - 27]).

FISH on metaphase chromosomes is a very efficient method to determine chromosome aneuploidy. However, the method has several limitations. The main difficulty is to obtain metaphase chromosomes in sufficient quality and quantity for analysis. In fact, it is not always possible to obtain such preparations from a patient's sample. For example in breast cancers, metaphase chromosome preparations are obtained in no more than 50% of the cases, whereas in some bronchic cancers, this value does not exceed 10% [15,28]. Another bias is introduced using the in vitro culture. Only cells able to divide in the given conditions will be observed and a false evaluation of the actual clonal composition of the sample is possible [29]. In addition, metaphase

390 K. TRUONG ETAL.

chromosome preparation requires a minimum of 3 - 5 days inducing a long waiting time before the diagnostic result is known.

17.3.2. Interphase chromosomes.

In interphase nuclei, chromosomes adopt a more extended structure compared to metaphase chromosomes. FISH with DNA probes may be performed directly on interphase nuclei from various tissue samples (fine-needle samplings, pleural effusions, biopsies, tissue slides...). This approach solves numerous problems associated with metaphase chromosome preparations with loss in resolution. In contrast to metaphase chromosome where probes may be precisely localized on an identifiable chromosome, hybridization on interphase nuclei show fluorescent spots which may be enumerated without possibilities of precise location (review in [30 - 33]).

Whole chromosome and chromosome arm specific paintings generate a diffuse signal when hybridized on nuclei, not allowing a precise localization of the spots particularly in cases of complex chromosome rearrangements. Specific probes of centromeric repeated DNA may also be used. However, these probes do not detect chromosome rearrangements without involvement of a variation in the number of centromeres, such as formation of isochromosomes. Probes of several hundred of kilobase pairs such as Bacterial Artificial Chromosomes (BAC or PAC) specific of single sequences may be used. A commercial probe of this type is available for trisomy detection in prenatal diagnosis [34].

Whatever the used probe may be, a bias is remaining in tumour studies due to the ignorance of the cell ploidy. In fact, the copy number of a chromosome segment has not the same significance in a diploid or an aneuploid cell. The use of a reference probe located on another chromosome does not solve the problem as the number of copies of each chromosome may vary in tumour cells. In addition, enumeration of FISH spots either by human observers or by automated image cytometry reveals major difficulties concerning objective criteria for signal discrimination or segmentation. Hybridization efficiency and variable signal intensity within nuclei may hamper correct and objective spot enumeration. In order to go beyond these limitations, we developed an automated image cytometry approach [35 - 39].

17.4. QUANTITATIVE FISH BY AUTOMATED IMAGE CYTOMETRY ON INTERPHASE NUCLEI.

Image cytometers are automated imaging systems for cellular analysis on glass slides in fluorescence and bright field microscopy using a wide range of dedicated analysis modules. The platform that has been chosen for in our studies was Pathfinder^^ from IMSTAR (France). The basic unit is a microscope-based unit combined with a motorized stage, a high-resolution CCD camera, and fast auto-focus mechanism (Figure 17.1). The Pathfinder^^ instrument has been designed for easy set up and use, which is achieved through an intuitive graphical user interface fiilly adjustable for each user type, defined parameters ensuring optimal, reproducible analysies. A powerful combination


of multi-wavelength fluorescence imaging, including flexible analysis algorithms and full system automation enable quality information for the analysis. Results can be interactively verified due to permanent link between extracted data and the images. Thus, the system provides tools for rapid and objective quantitation of multiple cell parameters allowing human control. For chromosome or chromosome arms imbalance measurements with image cytometers, slides were processed using standard methods for FISH. DNA in the nuclei was stained by 4,6-diamino-2-phenylindole (DAPI). The two chromosomes or chromosome arms specific paintings were labelled using digoxigenin or biotin modified nucleotides and detected using fluorescein isothiocyanate (FITC) or Texas red labelled adduct specific antibodies. Reference slides with normal lymphocytes or fibroblasts were simultaneously hybridized to the samples. Quantitation of hybridization signals was performed as followed: images were acquired at different excitation wavelengths corresponding to DAPI (330 nm), FITC (green fluorescence, 485 nm) and Texas Red (red fluorescence, 577 nm). No bleeding of any red or green signal was observed when changing the filters. Detection of the nuclei was performed on the basis of DAPI counter-staining. After acquisition of 300 fields with a 40x objective, which contained altogether 1000-1500 nuclei, object segmentation and fluorescence measurements for each nucleus were performed automatically by the PATHFINDERS^ system. For each experiment, optimal integration time was determined with the reference slide and kept constant for all slides. Artefacts, falsely segmented nuclei and nuclei containing no green or red fluorescence were excluded interactively. The amount of excluded cells could reach up to 20% depending on the hybridization efficiency. A minimum of 800 cells with analyzable data was obtained. The histogram representing the chromosome or chromosome arm ratio of the whole cell population and its mean value were automatically generated by the software after background subtraction. Two examples with different applications are presented: chromosome 3 arms imbalance in bronchic cancer and prenatal diagnosis of trisomy 21.

17.4.1. Chromosome 3 arms imbalances in bronchic cancers.

A first series of analyses was performed on short-time-cultured lung tumour samples [36]. It was shown that high quality double-colour FISH could be obtained with on these samples (Figure 17.2A) and that the imbalance could be determined with reliability. To investigate the importance of our method in clinical work, we further analyzed a series of tissue sections from snap-frozen NSCLC and SCLC specimens [39] (Figure 17.2B) Integrated fluorescent signals from 3q and 3p arm-specific painting were separately quantified in each nucleus and their ratio measured. A histogram representing the 3q/3p ratio of the whole cell population was automatically generated by the software after background subtraction. Detection of peaks representing each cell population was performed automatically based on curve fitting of the Gaussian mode. The standard deviation of the mean fluorescence was taken as an indicator for dispersion of the values. The automated calculation based on statistical analysis of the histogram allowed the exclusion of nuclei containing no green or red

392 K. TRUONG ETAL.

Figure 17.1. Image cytometer Pathfinder (IMSTAR).

fluorescence as well as artefacts with abnormally high fluorescence values. The proportion of excluded cells could reach up to 10% depending on the hybridization efficiency.

An example of the quantified signals within nuclei is shown in a biopsy and in a control (Figure 17.3). Green/red (long arm painting/short arm painting) ratios were determined for each nucleus of a slide. Analysis of the resulting histograms, using specifically developed software allowed a determination of mean values for the whole preparation. Imbalances were obtained by dividing the mean value of the 3q/3p ratio of tumour samples by the mean value of the 3q/3p ratio from normal lymphocytes. When no imbalance was present, the expected ratio was 1, in case of excess of long arm, the ratio would be higher than 1; it would be below 1 if short arms were in excess. In the studied biopsies, calculated imbalances varied from 1.0 to 2.6 with a mean of 1.4 (median: 1.5) indicating excess of 3q [39]. No excess of 3p was observed. Variability of calculated imbalances were verified between successive experiments and adjacent sections. Data showed variations of about 10 %. 48 % of the samples showed an imbalance. In one sample, two populations were observed on the same biopsy, one revealing an imbalance, one without an imbalance in accordance with histological heterogeneity of this tumour (Figure 17.4). The distribution of individual nuclei carrying imbalances on the whole tissue section allowed the establishment of two populations with specific distributions (Figure 17.2C). However, the overlap between the experimentally established value of the imbalance in cell nuclei for normal and abnormal 1 populations limits the possibilities to detect the presence of cellular clone represented by only a few cells. Experiments performed on artificially composed samples of normal and aneuploid cells have shown that the current detection limit corresponds to the ratios of 1 to 20.


iaoo

200

|100

0 7"., .......'... 1 ,f 5 g "

- ]

„.X'iJ.»^fl.nss.._,

Figure 17.2. A. FISH on interphase nuclei and a partial metaphase of non-small-cell lung cancer cells. The chromosome 3 long arm and short arm specific paintings were revealed in green and red, respectively. Arrow: 2 normal metaphase chromosome 3; Arrow heads: interphase nuclei. Nuclei and chromosome are colored in orange by propidium iodide B. Interphase nuclei in a biopsie of a lung cancer after FISH of chromosome 3 painting (green: long arm, red: short arm., blue: Dapi. C. Chromosome 3 arm imbalance for lung cancer biopsy with two clones. The histogram displays the number of nuclei as a function of the imbalance measured on the slide. Cells with imbalance values around 1 and 1.6 were selected on the histogram in red and green, respectively, and located on the tissue slide using the same colors. Cells without imbalance were mainly present in the upper portion of the slide whereas cells with imbalances around 1.6 were located in the lower portion (after [1]). D. Field of an ammiotic fluid preparation after FISH of chromosome 21 (green), and 22 (red). Interphase nuclei are labelled in blue by Dapi. (See color insert section.)

394 K. TRUONG ETAL.

600 -

500 -

400 -

300 -

200 -

100 •

0 -

N Cancer Control 11

1 • • • • ' J J J

III i l l l l l

III I^BLJBLHUHLH^HLJHL^

1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , 1 , •, •, •, 1, , , • [

1 Ratio Figure 17.3. Determination of chromosome imbalance in lung cancer biopsy. Histograms representing the distribution of the fluorescence ratios for chromosome 3 long and short arms revealed by specific paintings. Measured nodal values were 1.05 for control (lymphocytes) and 2.75 for the cancer, thus imbalance was 2.6.

IN

laso

|200

150

[lOO

so

111 n 0.5 1.0 1.5

r-L_^ n Type Ox^eX Cwve2

I I 1 I T K '{ ' 2" r 0 »a(Whg Xo j U ^ P T H T

2.0 2 5 3.0 3 5 4i0 Ratio

Figure 17.4. Automated curve fitting. The curve enveloping of the histogram, can be decomposed in 2 Gaussian components. In the box, the characteristics of the gaussian components are presented: amplitude, position (Xo) and standard deviation (sigma) determined automatically. In comparaison to the control (not shown), the first gaussian (Xo=1.165) corresponds to an imbalance of 1. The position of the second gaussian gives the value of the imbalance of the second population: 1.6.


17.4.2. Prenatal Diagnosis of Trisomy 21.

For pre-natal diagnosis of trisomy 21, the ratio between the number of chromosome 21 and 22 was measured [38] (Figure 17.2D). The chromosome 22 was chosen as a reference because of its similar size with chromosome 21. Cultured blood of patients with Down's syndrome and of healthy donors were first used to establish the sensitivity of the technique. Then, samples from amniotic fluids from pregnant women at risk chosen after classical diagnostic tests were analyzed in parallel with standard metaphase analysis. Integrated fluorescent signals from chromosome 21 and 22 specific paintings were separately quantified and their ratio measured. The ratio of the signals of chromosomes 21 and 22 normalized with data from normal cells (normal lymphocytes or fibroblasts), which were processed simultaneously allowing determination of the copy amount of chromosome 21. When two copies of both chromosomes 21 and 22 were present in the cells, the ratio was expected to be 1, in case of trisomy, the ratio would be 1.5. An example of the quantified signals within interphase nuclei is shown in figure (Figure 17.5). Green/Red ratios were determined on each nucleus in every field and the resulting histogram allowed determination of the mean value for the whole sample.

The ratio calculated by comparison of blood from various donors, ranged from 0.95-1.05 for normal blood (not shown). For patients with trisomy, ratios ranged from 1.5 to 1.6, showing good correlation with the expected value of 1.5 (not shown). The low dispersion of the data allows clear discrimination between normal and trisomic blood specimens. For amniotic fluids, normal fibroblasts were used as a reference because results were less dispersed (smaller standard deviations) than using normal lymphocytes. The reason for this difference between the two cell types may lie in the size difference of the nuclei and the cytoplasm. The cytoplasm of the fibroblasts seems to have less permeability for probes and antibodies. Cells from amniotic fluids reacted in a similar way. Therefore, fibroblasts are recommended as a reference. On 20 analyzed cases, 7 were positive (imbalance range 1.45-1.6) and 13 were normal (imbalance range 0.95-1.1). Results showed 100 % concordance with conventional cytogenetics.

17.5. CONCLUSIONS AND PERSPECTIVES.

The data obtained indicated that chromosome or chromosome arm imbalances could be detected efficiently on interphase nuclei using a quantitative comparison between the integrated fluorescent signals from specific paintings. The possible bias during cell culture necessary for metaphase analysis may be eliminated. Our method presents several advantages compared to the currently used interphase FISH approaches. Enumeration of FISH spots by human observers showed difficulties related to objective and correct spot discrimination requiring presence of experienced observers. Our automated technique allows objective analysis of FISH signals on a great number of cells (500-1000 or more). It also avoids segmentation problems due to variable signals within one nucleus. Moreover, inter-experimental variations of the

396 K. TRUONG ETAL.

350 1

300 H

250 A

200 H

Control

150 H

100

I I'n'i I rrn i i

Ratio Figure 17.5. Histograms representing the distribution of the fluorescence ratios of chromosome 21 and 22 paintings in an amniotic fluid and in normal control fibroblasts. The mean ratio were 1.3 for the ammiotic fluid and 0.8 for the control. Thus, the measured imbalance was 1.6, indicating a trisomy 21.

signal do not interfere with the measurement because the imbalance is calculated with regard to a reference slide that is added at each experiment. Finally, the variability of the signal within the same nucleus does not affect the measurement either, since the entire integrated fluorescence is analyzed.

The efficiency of the approach was proved for detection of chromosome 21 trisomy. Aneuploidy of other chromosomes of interest as chromosomes 13, 18, X or Y could also be studied in a similar way using appropriate chromosome painting. For cancers, the data obtained indicate that chromosome imbalance can be detected efficiently on the nuclei of clinical samples. The method allows the detection of cancer cells directly on the histological heterogeneous tumour sample. The presence of sub-populations with different chromosome imbalances can be precisely established. The possibility to localize each nucleus including the chromosome imbalance value allows reconstitution of the multi-clonal panel of the apparently homogeneous section. This reconstitution could be a powerful approach for the understanding of tumour progression.


The described quantitative FISH method gathers the advantages of common interphase analysis with a faster and expert-independent way of analyzing chromosome aneuploidies for cancer or prenatal diagnosis. This approach could be a response to the challenge of performing reliable, fast and large-scale analysis of patients. In particular, this approach may be used for techniques of detection of rare cells such as micrometastasis or ciriculating fetal cells within maternal blood for non-invasive prenatal diagnosis.

17.6. REFERENCES

1. H. Rajagopalan and C. Lengauer (2004). Aneuploidy and cancer Nature 432(7015), 338-341. 2. G. Pihan and S. J. Doxsey (2003). Mutations and aneuploidy: co-conspirators in cancer?

Cancer Cell 4(2), 89-94. 3. Z. Storchova and D. Pellman (2004). From polyploidy to aneuploidy, genome instability and

cancer Nat Rev Mol Cell Biol 5(1), 45-54. 4. V. M. Draviam, S. Xie, and P. K. Sorger (2004). Chromosome segregation and genomic

stability Curr Opin Genet Dev 14(2), 120-125. 5. S. E. Antonarakis, R. Lyle, E. T. Dermitzakis, A. Reymond, and S. Deutsch (2004).

Chromosome 21 and down syndrome: from genomics to pathophysiology Nat Rev Genet 5(10), 725-738.

6. B. Dutrillaux (1995). Pathways of chromosome alteration in human epithelial cancers Adv Cancer Res ^159-^2.

7. F. Farabegoli, M. A. Hermsen, C. Ceccarelli, D. Santini, M. M. Weiss, G. A. Meijer, and P. J. van Diest (2004). Simultaneous chromosome Iq gain and 16q loss is associated with steroid receptor presence and low proliferation in breast carcinoma Mod Pathol 17(4), 449-455.

8. I. Bieche, M. H. Champeme, and R. Lidereau (1995). Loss and gain of distinct regions of chromosome Iq in primary breast cancer Clin Cancer Res 1(1), 123-127.

9. N. Kokalj-Vokac, A. Alemeida, M. Gerbault-Seureau, B. Malfoy, and B. Dutrillaux (1993). Two-color FISH characterization of i(lq) and der(l;16) in human breast cancer cells Genes Chromosomes Cancer 7(1), 8-14.

10. D. P. Carbone (1997). The biology of lung cancer Semin Oncol 24(4), 388-401. 11. Wistuba, II, C. Behrens, S. Milchgrub, D. Bryant, J. Hung, J. D. Minna, and A. F. Gazdar

(1999). Sequential molecular abnormalities are involved in the multistage development of squamous cell lung carcinoma Oncogene 18(3), 643-650.

12. K. L. Braithwaite and P. H. Rabbitts (1999). Multi-step evolution of lung cancer Semin Cancer Biol 9(4), 255-265.

13. C. A. Powell, S. Klares, G. O'Connor, and J. S. Brody (1999). Loss of heterozygosity in epithelial cells obtained by bronchial brushing: clinical utility in lung cancer Clin Cancer Res 5(8), 2025-2034.

14. J. Sanz-Ortega, M. C. Saez, E. Sierra, A. Torres, J. L. Balibrea, F. Hernando, J. Sanz-Esponera, and M. J. Merino (2001). 3p21, 5q21, and 9p21 allelic deletions are frequently found in normal bronchial cells adjacent to non-small-cell lung cancer, while they are unusual in patients with no evidence of malignancy J Pathol 195(4), 429-434.

15. J. R. Testa, Z. Liu, M. Feder, D. W. Bell, B. Balsara, J. Q. Cheng, and T. Taguchi (1997). Advances in the analysis of chromosome alterations in human lung carcinomas Cancer Genet Cytogenet 95(1), 20-32.

16. J. Pei, B. R. Balsara, W. Li, S. Litwin, E. Gabrielson, M. Feder, J. Jen, and J. R. Testa (2001). Genomic imbalances in human lung adenocarcinomas and squamous cell carcinomas Genes Chromosomes Cancer 31(3), 282-287.

17. M. J. Parker, J. L. Budd, E. S. Draper, and I. D. Young (2003). Trisomy 13 and trisomy 18 in a defined population: epidemiological, genetic and prenatal observations Prenat Diagn 23(10), 856-860.

18. I. Witters, K. Devriendt, E. Legius, G. Matthijs, D. Van Schoubroeck, F. A. Van Assche, and J. P. Fryns (2002). Rapid prenatal diagnosis of trisomy 21 in 5049 consecutive uncultured amniotic fluid samples by fluorescence in situ hybridisation (FISH) Prenat Diagn 22(1), 29-33.

398 K. TRUONG ETAL.

19. S. A. Rasmussen, L. Y. Wong, Q. Yang, K. M. May, and J. M. Friedman (2003). Population-based analyses of mortality in trisomy 13 and trisomy 18 Pediatrics 111(4 Pt 1), 111-l^A.

20. R. Wapner, E. Thom, J. L. Simpson, E. Pergament, R. Silver, K. Filkins, L. Piatt, M. Mahoney, A. Johnson, W. A. Hogge, R. D. Wilson, P. Mohide, D. Hershey, D. Krantz, J. Zachary, R. Snijders, N. Greene, R. Sabbagha, S. MacGregor, L. Hill, A. Gagnon, T. Hallahan, and L. Jackson (2003). First-trimester screening for trisomies 21 and 18 N Engl J Me^/349(15), 1405-1413.

21. B. Dutrillaux (2000). Chromosome and gene alterations in human cancers in relation to aging Chromosomes Today 13207-223.

22. F. J. Mitelman, B. Mertens, F.E., Mitelman database of chromosomes aberrations in cancers, 2001.

23. L. Verma, F. Macdonald, P. Leedham, M. McConachie, S. Dhanjal, and M. Hulten (1998). Rapid and simple prenatal DNA diagnosis of Down's syndrome Lancet 352(9121), 9-12.

24. B. Eiben, W. Trawicki, W. Hammans, R. Goebel, and J. T. Epplen (1998). A prospective comparative study on fluorescence in situ hybridization (FISH) of uncultured amniocytes and standard karyotype analysis Prenat Diagn 18(9), 901-906.

25. M. Muhlmann (2002). Molecular cytogenetics in metaphase and interphase cells for cancer and genetic research, diagnosis and prognosis. Application in tissue sections and cell suspensions Genet Mol Res 1(2), 117-127.

26. S. D. Mundle and I. Sokolova (2004). Clinical implications of advanced molecular cytogenetics in cancer Expert Rev Mol Diagn 4( 1), 71 -81.

27. T. Schwarzacher (2003). DNA, chromosomes, and in situ hybridization Genome 46(6), 953-962.

28. A. Flury-Herard, E. Viegas-Pequignot, H. De Cremoux, C. Chlecq, J. Bignon, and B. Dutrillaux (1992). Cytogenetic study of five cases of lung adenosquamous carcinomas Cancer Genet Cytogenet 59(1), 1-8.

29. K. Truong, M. N. Guilly, M. Gerbault-Seureau, B. Malfoy, P. Vielh, and B. Dutrillaux (1999). Evidence for in vitro selection during cell culturing of breast cancer: detection by flow and image cytometry Cancer Genet Cytogenet 114(2), 154-155.

30. J. M. Bartlett (2004). Fluorescence in situ hybridization: technical overview Methods Mol Med 9111-Sl.

31. T. Liehr, H. Starke, A. Weise, H. Lehrer, and U. Claussen (2004). Multicolor FISH probe sets and their applications Histol Histopathol 19(1), 229-237.

32. S. Langer, J. Kraus, I. Jentsch, and M. R. Speicher (2004). Multicolor chromosome painting in diagnostic and research applications Chromosome Res 12(1), 15-23.

33. K. K. Jain (2004). Current status of fluorescent in-situ hybridisation Med Device Technol 15(4), 14-17.

34. M. A. Hulten, S. Dhanjal, and B. Pertl (2003). Rapid and simple prenatal diagnosis of common chromosome disorders: advantages and disadvantages of the molecular methods FISH and Q¥-PCR Reproduction 126(3), 279-297.

35. K. Truong, M. N. Guilly, M. Gerbault-Seureau, B. Malfoy, P. Vielh, C. A. Bourgeois, and B. Dutrillaux (1998). Quantitative FISH by image cytometry for the detection of chromosome 1 imbalances in breast cancer: a novel approach analyzing chromosome rearrangements within interphase nuclei Lab Invest 78( 12), 1607-1613.

36. K. Truong, M. Gerbault-Seureau, M. N. Guilly, P. Vielh, G. Zalcman, A. Livartowski, A. Chapelier, M. F. Poupon, B. Dutrillaux, and B. Malfoy (1999). Quantitative fluorescence in situ hybridization in lung cancer as a diagnostic marker J Mol Diagn 1(1), 33-37.

37. K. Truong, P. Vielh, M. N. Guilly, J. Klijanienko, X. Sastre-Garau, F. Soussaline, B. Dutrillaux, and B. Malfoy (2002). Quantitative FISH analysis on interphase nuclei may improve diagnosis of DNA diploid breast cancers Diagn Cytopathol 26(4), 213-216.

38. K. Truong, A. Gibaud, J. M. Dupont, M. N. Guilly, F. Soussaline, B. Dutrillaux, and B. Malfoy (2003). Rapid prenatal diagnosis of Down syndrome using quantitative fluorescence in situ hybridization on interphase nuclei Prenat Diagn 23(2), 146-151.

39. K. Truong, A. Gibaud, G. Zalcman, M. N. Guilly, M. Antoine, F. Commo, C. Fouquet, J. Cadranel, T. Soussi, B. Dutrillaux, and B. Malfoy (2004). Quantitative fish determination of chromosome 3 arm imbalances in lung tumors by automated image cytometry Med Sci Monit 10(ll),BR426-432.

IMPORTANCE OF MEASURING FREE ZINC IN CELLS

Rebecca A. Bozym, Richard B. Thompson, and Carol A. Fierke

18.1. INTRODUCTION

Zinc is the second most abundant trace element in the body and essential for eukaryotic and prokaryotic organisms. Zinc is required as a cofactor or structural component for more than 300 metalloenzymes in all six classes of enzymes " . Zinc is a vital cofactor for zinc fingers of transcription factors ' j the immune system , and the reproductive system . More than 1% of human gene products are zinc finger proteins . A deficiency in zinc leads to impairment of growth, immune activity, and brain functions ' ^ . Interest in zinc is growing among researchers in medicine and neuroscience. In the brain, synaptically released zinc has both physiological and pathological relevance " ^ where the level of free zinc after release from vesicles may reach a range of 10-100 fiM in the synaptic cleft ^ . High amounts of vesicular zinc are seen in the hippocampus,fu cerebral cortex, and amygdala ' . Zinc released from synaptic vesicles has been suggested to inhibit N-methyl-D-aspartate (NMDA) and y-aminobutyric acid (GABAA) receptors, as well as potentiate a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors and act as an antagonist of voltage-gated calcium channels ^ ' ^ . Although zinc is essential for proper brain function, excessive zinc may also act as a neurotoxin ^ ; Frederickson and later Koh, et al., showed that neurons were filled with fi*ee zinc following prolonged seizure ^ or an ischemic insult ^ ' . Zinc has also been increasingly linked to apoptosis, although how it influences this process is still unclear. Several groups have reported that zinc at various levels (up to

* Rebecca A. Bozym and Richard B. Thompson, University of Maryland School of Medicine, Baltimore, MD, 21201. Carol A. Fierke, University of Michigan, Ann Arbor, MI, 48109.

399

400 BOZYM ETAL

hundreds of micromolar) induces apoptosis ^ . Canzoneiro, et aL, ^ have shown that chelation of extracellular zinc inhibits zinc influx and eventual neuronal death. However, they also found that addition of the membrane-permeant chelator TPEN caused neuronal death to occur, presumably due to excessive chelation of zinc, as also as shown by Truong-Tran et al ^^. Other investigators have shown that oxidative insults, including administration of nitric oxide, lead to the release of free zinc from intracellular sources ^ " . Maret, along with Bossy-Wetzel et al , have also shown that metallothionein can release bound zinc under oxidizing conditions ^ ' ^ .

Though zinc plays important biological roles, little is known about the process(es) of distribution of this metal in the body, or amongst metalloproteins which require it for function. Eukaryotic cells generally are rich in zinc with a total concentration in the range of 100 ^M \ The abundance of zinc ligands in cells, including, glutathione, histidine, and cysteine, as well as metallothionein (MT) and a myriad of proteins ^ ' , assures that the vast majority of cellular zinc is bound and not free. Thus it has been suggested that there is effectively very little (fM) or no free zinc in bacterial cells ^ . However, femtomolar free zinc equates to less than one atom of free zinc per bacterial cell and only a handful in eukaryotic cells, which seems improbable in view of the high total concentration and abundant ligands with medium affinity (nM to ^M).

An alternative hypothesis proposes that there is a "pool" of free and rapidly exchangeable zinc ions complexed with kinetically labile ligands which are available for inclusion into metalloproteins and for other purposes. Rapidly exchangeable zinc is likely the biologically active form of zinc, whose level is closely regulated. Metallothionein is thought to act as a buffer of cytosolic zinc ^ . Zinc is enriched in certain compartments within the cell such as endosomes/lysosomes which aid in storage or detoxification ^ . Similar types of organelles are also found in synaptic vesicles and secretory granules ^ ' . Significant amounts of chelatable zinc are seen in certain cell types: pancreatic P cells ^ , hippocampal neurons, and sperm cells ^''^\

Two families of zinc transporters also aid in regulating zinc homeostasis. The ZIP (ZRT, IRT-like protein) family facilitates influx of zinc into the cytosol ^^'^^. The CDF (cation diffusion facilitator) family acts in the efflux of zinc from the cytosol ^\ Different subtypes of these families are seen in different cells and different areas of the cell such as the CDF family member ZnT-3 which is highly expressed in the hippocampus, cerebral cortex, and amygdala "

It seems likely there should be a process for rapidly delivering zinc to upregulated transcription factors and the 300+ metalloenzymes that utilize zinc. In the case of copper, specific chaperones exist for each copper enzyme even though it fixnctions in only -12 enzymes. Presumably due to its reactivity and potential toxicity, copper is tightly regulated and there is very little, if any, free copper in the cell ^'^. Copper chaperones deliver copper from transporters to particular cytosolic proteins and to organelles. To date, no similar chaperone for zinc has been identified. Yet if free/exchangeable levels are truly picomolar or lower, the kinetics of binding to newly synthesized zinc proteins by diffusion alone would be very slow: for instance, the measured association rate

MEASURING FREE ZINCIN CELLS 401

constant for zinc binding in vitro to apocarbonic anhydrase II (kon ) is 1.1 x 10^ |iM" s" ^^ , such that even at 0.1 nM zinc concentrations would take more than an hour to saturate the apoenzyme, which seems unacceptably slow in comparison to the rest of the gene expression apparatus. If the intracellular free/rapidly exchangeable level is typically picomolar ^^ even proteins with diffusion-controlled zinc association rate constants would take hours to equilibrate. However, if intracellular free zinc levels are rapidly maintained at constant concentration (i.e., pM) by transporters and chelators then the kinetics of zinc incorporation will depend on the concentration of the apoprotein in question. In this case, the half time for zinc incorporation by diffusion-controlled binding is more rapid, varying from 70 seconds for nanomolar apoprotein to 70 msec for an apoprotein at micromolar levels. O'Halloran and his colleagues have suggested that zinc addition to apoproteins is under kinetic rather than thermodynamic control ' . One way to significantly increase the rate of zinc incorporation is for chelated zinc, which is at much higher concentration than free zinc, to serve as donor. It has long been known that 2,6-dipicolinate ^^ is able to catalyze the binding to and the dissociation from carbonic anhydrase of zinc. It remains to be shown whether other small molecules or proteins are also capable of catalyzing the association/dissociation of zinc. Also, apoproteins are thermodynamically less stable in their metal-free state, adding to the need for rapid zinc insertion either by small molecules or chaperones. However, the hundreds of proteins requiring zinc would seem to require a large number of specific protein chaperones, which makes small molecules perhaps more appealing as chaperones.

The determination of free zinc concentration in cells is important for a number of reasons: 1) the level of free (or rapidly exchangeable) zinc essentially reflects the occupancy of all other binding sites, and is most closely related to the zinc activity; 2) free levels (as opposed to total levels) establish thresholds of zinc toxicity, both high and low, with intracellular and extracellular concentrations maintained at different levels; 3) to establish if free zinc is involved in signaling in the brain or elsewhere then it must be quantitated, preferably with an imaging method; 4) free zinc concentrations may provide an overall measure of zinc nutriture, ultimately useful in assessing deficiency; 5) understanding the roles of zinc in pathological states and processes, such as amyloid plaque formation, seizures, apoptosis, ischemia and oxidative stress.

The free zinc level in cells has been observed under various conditions ^ "" j but many of these studies were performed in cells known to be rich in zinc (e.g., mossy fiber neurons, pancreatic islet cells). By comparison, the majority of resting cultured cells studied exhibit little cytoplasmic staining ^ , except for perinuclear punctuate staining ^ ^ , suggesting that cytoplasmic resting free zinc levels are lower than the detection limits of these indicators, generally not much under one nanomolar. Some difficulties in measuring free zinc at nanomolar levels and below are widely appreciated, such as potential interference by Ca^^ and Mg^^ which are present at much higher concentrations. Less well accepted are the influences of substantial (and variable) concentrations of medium strength ligands such as serum albumin as well as zinc itself in culture media.

402 R.A. BOZYM ETAL

18.2. TSQ DERIVATIVES

The first biologically useful fluorescent zinc indicator, TSQ (p-toluenesulfonamide quinoline) was used to observe free zinc in brain tissue sections ^ . TSQ shows an increase in fluorescence upon formation of the Zn:TSQ complex and is able to cross the cell membrane. This was the first indicator that was selective enough to measure zinc in the presence of high concentrations of calcium and magnesium ' . However TSQ may also form a water insoluble complex and is found to partition in the cell membrane as well. The fluorescence of TSQ is also pH dependent, and the stoichiometry of zinc binding is not unequivocal ^ , complicating the ability to quantitatively measure free zinc in cells. Due to this nature derivatives of TSQ were made for use in cell systems, such as zinquin ^ . Zinquin coordinates zinc by two nitrogen atoms with a large increase in fluorescence upon addition of zinc. Ca ^ and Mg^^ have little effect on the fluorescence of zinquin whereas Fe^^ and Cu^^ quench any fluorescence. The ethyl ester of zinquin is excited at 368 and emits at 490 nm ^ . Studies have found that zinquin A likely forms a ternary complex with zinc and the protein or biological molecule it is attached to ^ . This suggests that it is a semiquantitative indicator for measuring intracellular free zinc.

Another TSQ derivative used in imaging free zinc in mossy fiber boutons is TFLZn, N-(6-methoxy-8-quinolyl)-p-carboxybenzoylsulfonamide ^'^. This indicator is water soluble and exhibits an increase in fluorescence emission by 100-fold upon excitation at 360 nm in the presence of zinc ' . TFLZn has a IQ of approximately 20 ^M with no apparent interference by Ca and Mg at ImM ' . Due to the extremely low levels of available zinc, the affinity of this indicator renders it generally less useful for application in cells.

18.3. FLUORESCENT INDICATORS BASED ON FLUORESCEIN

18.3.1. The Zinpyr family

Fluorescein and its derivatives are widely used in biological applications such as flow cytometry, confocal scanning microscopy, DNA sequencing, and immunoreagents ^ . The high quantum yield, long wavelength excitation maxima, high extinction coefficient, and water solubility of fluorescein in comparison with the quinolines make it a good candidate for the development of zinc indicators. The many members of the Zinpyr (ZP) family (Figure 18.1) utilize the spectral properties of fluorescein while adding the selectivity of the zinc chelator di(2-picolyl)amine (DPA) forming a photoinduced electron transfer (PET) zinc indicator ^ ' ' ' ^ . The first-generation indicators ZPl and ZP2 contain two DPA moieties at the 4' and 5' positions of the xanthenone ring, forming a symmetrical indicator that exhibits enhanced fluorescence upon addition of zinc or cadmium ^ . In the absence of zinc the fluorophore is switched "off, likely due to electrons on the nitrogen donor atom that transfer to the fluorophore, subsequently quenching emission. Zinc coordinates to the nitrogen donor, reducing the quenching so that fluorescence


increases 3- to 6-fold ^ . However, ZPl and 2 demonstrate high background fluorescence and can bind two zinc atoms. Nolan et al. ^^ describe three new ZP indicators, ZP5, 6, and 7, based on the structure of ZP4, which contains an unsymmetrical fluorescein platform. ZP5-7 are electronic variations of ZP4 at the aniline nitrogen ^ . These indicators exhibit fluorescence enhancement upon the addition of zinc along with a slight blue shift of 10 nm. They also exhibit a lower quantum yield in the unbound state due to the reduced pKa of the nitrogen atom. For ZP5 and ZP6 the pKa's are 9.6 and 6.3 respectively, making ZP6 a good candidate because PET is not inhibited at physiological pH. Zinc binding affinities were measured using a dual-metal buffering system in which both Ca and EDTA are kept at a constant concentration while varying the zinc concentration ^\ The K^'s for both ZP5 and 6 are 0.50 ± 0.10 nM; however, the fluorescence intensity only increases approximately 2-fold for ZP5 and 6, and changes negligibly for ZP7. For both ZP5 and 6, Cd^^ and Zn^^ exhibit similar fluorescence increases, while there is no fluorescence enhancement in the presence of Ca^ , Mg^ , and Mn^^ (no Zn^^). In the presence of Fe^^, Co^ , Ni^^, and Cu^ , Zn^^ does not successfully compete to bind with the fluorophore, so there is no fluorescence enhancement. These indicators (ZP4-6) are not cell permeable.

By comparison ZP3 seems to be the most promising indicator as it is cell permeable and exhibits a six-fold increase in fluorescence in the presence of zinc ^\ However ZP3 is still an intensity-based indicator which means that it is dependent on the amount of indicator present, the excitation intensity, and sample conditions. ZP3 has been used to image endogenous pools of zinc in hippocampal neurons and slices using confocal microscopy ^ . More recently ZP3 has been shown to undergo two-photon excitation for microscopy with excitation at 800 nm ^ . Overall, ZPl-3 (Figure 18.1) exhibit a high quantum yield in the presence of zinc. ZP4-7 have a lower quantum yield in the presence of zinc. ZP8 has also been developed on the monosubstituted difluorescein platform and is asymmetrical, as are ZP4-7 *l Upon addition of zinc ZP8 exhibits an eleven-fold increase in fluorescence intensity with a IQ of 0.6 ± 0.1 nM ^ . ZP8 has the same response to other metals as ZP4-7 and is also not cell permeable.

Improvements to these indicators have been made in order to minimize background fluorescence due to protonation of the tertiary nitrogen atom of the PET switch ^\ In order to achieve this, ZP indicators were substituted with electron withdrawing groups on fluorescein ^\ ZPFl, ZPCU, ZPBrl, and ZPF3 all display pKa values below physiological pH with a low quantum yield ^

Other fluorescent indicators for zinc which have recently been described are also based on a fluorescein platform ^ ' ^ . The indicator Zin-naphthopyr 1 (ZNPl), described by Chang et al. ^ , is based on seminaphthofluorescein where it switches between tautomeric forms of fluorescein and naphthofluorescein. This indicator utilizes a single excitation source with dual emission at 624 nm and 528 nm. Upon addition of zinc the emission intensity increases at 624 nm with very minimal changes at 528 nm. The reported apparent K for zinc was 0.55 nM, with

404 R.A. BOZYM ETAL

ZPI:X = CI,Y = H ZP2:X»H,Y»H ZP3: X « F, Y = H ZPF1:X = CI»Y»F ZPCI1:X = CI,Y«CI ZPBrl:X = CI,Y = Br ZPF3: X = F, Y = F

ZP4:X = H,Y = H,Z = CI ZP5:X = F,Y = H,Z«CI ZP6:X«CI,Y = H,Z = a ZP7:X = 0Me,Y = H,Z = CI ZP8:X«H,Y = F,Z»F

Figure 18.1. Structures of the Zinpyr family of zinc indicators. ZPl-3 along with the electronically substituted ZP's are based on a symmetrical fluorescein platform with substitutions on both the xanthene ring and benzoic acid moiety. ZP4-8 are based on an asymmetrical fluorescein platform with substitutions on the xanthene ring. Redrawn from

an 18-fold increase in emission intensity ratio after zinc binding. When ZNPl is added to COS-7 cells as shown by Chang et al. ^ , minimal fluorescence of the cells is observed. Upon addition of 50 |iM Zn(pyrithione)2 an increase in cellular fluorescence.is observed

18.3.2. The Zinspy Family

The Lippard group is also developing zinc indicators based on a fluorescein reporting group and a pyridyl-amine-thioether ligand ^^ This new set of zinc indicators are water soluble and increase their fluorescence upon addition of zinc. The Zinspy (ZS) family of indicators undergo PET as well. ZSl contains dichlorofluorescein, ZS2 contains difluorofluorescein, ZS3 contains benzylbromide, and ZS4 contains carboxaldehyde (Figure 18.2). The change from Cl-1 to F-1 in ZSl to ZS2 results in a decrease of the quantum yield in the apo form, but does not lower the pKa of the tertiary amine nitrogen, unlike the ZP indicators 1 and 3. The increase in fluorescence intensity upon zinc binding is only 2 fold for ZS2 and 1.4 fold for ZSl. ZS3 has a pKa of 9.3 suggesting that protonation will interfere with PET quenching in the apo form. Upon addition of zinc there is no fluorescence increase rendering this indicator unsuitable for zinc sensing at neutral pH. ZS4 exhibits a quantum yield of 0.12 in the apo form and 0.5 upon zinc addition with a 4.5-fold increase in fluorescence intensity immediately after adding zinc. Unfortunately, several minutes after mixing there is approximately a 25% decrease in fluorescence


Figure 18.2. Structures of the Zinspy family; redrawn from^

and precipitation occurs after 15 minutes. In summary, ZSl and ZS2 exhibit relatively high quantum yields in the absence of zinc, with modest fluorescence emission enhancement upon zinc binding. ZS3 has a fairly high pKa for the tertiary amine responsible for PET and thus displays a high quantum yield at neutrality in the absence of zinc with no enhancement when zinc is added. ZP4 demonstrates a 4.5-fold increase in fluorescence emission upon addition of zinc; however, this signal also decreases over time with the formation of a precipitate.

18.3.3. The ZnAFs

The fluorescent indicators in the ZnAF family also detect zinc through the PET mechanism while employing fluorescein as the fluorophore with an attached N,N-Bis(2-pyridylmethyl)ethylenediamine (a TPEN derivative) (Figure 18.3) '*. ZnAF-1 and ZnAF-2 show little fluorescence in the absence of zinc with fluorescence enhancement of 17-fold and 51-fold, respectively, in the presence of zinc ^^. Below pH 7.0 the fluorescence of both ZnAF-1 and 2 decreases due to protonation of a hydroxyl group on fluorescein which has a pKa of 6.43 ^\ Therefore ZnAF-IF and ZnAF-2F were designed with 2 fluorine

406 R.A. BOZYM ETAL

atoms on fluorescein in order to lower the pKg value for use at physiological pH. The quantum yield in the absence of zinc is greatly reduced with ZnAF-lF and 2F. Calibrations of binding affinities for each ZnAF were performed by buffering zinc in HEPES with 10 mM NT A. Studies were performed with other cations such as Ca^ , Mg^ , Mn^ , Co^^, Ni^^, and Cu^^ to see if they interfere with the fluorescence intensity. At up to 5 mM, Ca^ , Mg^ , and Mn^ showed no enhancement of fluorescence; Fe^^, Co^ , and Ni ^ indicated a slight enhancement of fluorescence; and Cu^^ quenched fluorescence ^ . In order to achieve intracellular measurements of free zinc the diacetyl derivative of ZnAF-2F was made so that it can permeate the cell membrane ^^. Cell staining was demonstrated in mouse macrophage cells and rat hippocampal slices ' . This series of indicators is useful in determining the free zinc level in the range of 0.1-10 nM zinc however there are a few drawbacks in using these indicators to detect free zinc inside of cells. The ZnAFs are intensity-based indicators which mean that their emission is dependent on excitation intensity, indicator concentration, the environment around the indicator, and the thickness of the cells. These indicators also utilize UV excitation which might limit their use in fluorescence microscopy.

ZnAF-Rl and ZnAF-R2 have also been developed using benzofuran derivatives as fluorophores and TPEN as the chelator ^ . These indicators are excitation ratiometric and operate through the internal charge transfer (ITC) mechanism. They are designed with an electron donating group (amino) conjugated to an electron withdrawing group (carboxylate) that undergoes ITC from donor to the acceptor ^ . There is a cation-induced blue shift in absorbance which is used to ratio the excitation intensities at an emission wavelength of 495 nm. ZnAF-R2 has a better quantum yield in the presence of zinc than ZnAF-Rl; however, both have relatively low quantum yields. ZnAF-R2 was shown to be selective for zinc over other divalent cations such as 5 mM Mg^ , Ca^ , Na^ , and 5^M Fe^^, Ni^^, Mn^ , and Cu^^ with a large response to Cd^^^^ Metal ion selectivity was determined by adding the indicated amount of each cation with 5|LIM ZnAF-R2 in HEPES buffer and the fluorescence ratio (335/365 nm) was calculated for each. ZnAF-R2 was made cell permeable by formation of the ethyl ester derivative and was used to stain mouse macrophage cells producing ratiometric images (340/380 nm). Although ZnAF-R2 appears to be a useful tool for measuring intracellular zinc it is ratiometric in the UV range which may increase the number of artifacts due to cellular autorluorescence.

18.4. ZINC INDICATORS BY MOLECULAR PROBES (NOW NVITROGEN)

A number of the indicators which are used for Ca+2 detection actually bind zinc with higher affinity ^ . These BAPTA-based (l,2-bis(o-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid) Ca^^ indicators, such as Fura-2, have been used to measure zinc in the presence of low [Ca^^]. Fura-2 has an apparent IQ for zinc of 3 nM in the absence of Ca^ , and remains sensitive to zinc in the presence of 25-100 nM free Ca+2 68. Other related


l:ZiAF-Rl ZnAF-lF 2: ZiAF-iU

Figure 18.3. Selected structures of the ZnAF family; redrawn from ^'^^.

indicatorsfor Mg^^ , such as MagFura-2 and MagFura-5 (Kd for zinc of 20 nM), have also been used for measuring zinc in neurons ^ . Another series of indicators has been developed to measure free zinc in the 0.1-100 \xM range for use in monitoring synaptically released zinc. FuraZin-1 and lndoZin-1 exhibit excitation and emission shifts, respectively, upon zinc addition. These indicators have apparent Kd's for zinc of 8 iiM and 3 ^M. FluoZin-2 and RhodZin-1 show an increase in fluorescence emission upon zinc binding with Kd's of 2 jiM and 23 ^M, respectively. FluoZin-3 exhibits a Kd for zinc of 15 nM and has been shown to respond to calcium at concentrations above 1 M ' ^ ^ , making it of questionable utility for quantitating extracellular zinc ^\ Moreover, this indicator is intensity-based with an increase in fluorescence with increasing concentrations of zinc. The AM ester of RhodZin-3 localizes in the mitochondria, exhibiting a 75-fold increase in fluorescence with saturating zinc, having a Kd of 65 nM ^ . While this may be a useful tool in determining the role of zinc in apoptosis, it is a straight intensity indicator which is subject to the usual artifacts; it may be, however, that it exhibits a useful lifetime change. The Newport Green indicators exhibit ^M affinity for zinc (1 iM and 30 ^M) ^ with no interference by mM calcium or magnesium ^ . These indicators use DPA as the chelator and undergo fluorescence enhancement upon zinc binding; Newport Green DCF also exhibits a lifetime shift which was useftjl in quantifying intracellular zinc ^ . Overall the zinc indicators available through Molecular Probes are useful in measuring higher zinc concentrations, particularly where calcium is at a minimum. Substantial data indicate that much of the biology of zinc occurs at < 10' M [Zn jfree where these probes are unable to accurately measure fi*ee zinc ion. The majority of these indicators are intensity-based, making it difficult to use them for actual quantitation of free zinc in a biological system.

408 R.A. BOZYM ETAL

18.5. A ZINC INDICATOR BASED ON BENZOXAZOLE

As mentioned before ratiometric indicators for zinc provide a more accurate measurement of free zinc in biological systems. Using this approach the measurement is largely independent of indicator concentration, excitation intensity, or differences in the sample. Taki et al. '^^ have recently described an emission ratiometric indicator using benzoxazole as the fluorescent component and an aminomethyl pyridine moiety as the zinc ligand named Zinbo-5 ^^. Upon addition of zinc to Zinbo-5 there was a decrease in the absorbance band at 337 nm which is seen in the apo form, and an increase of a new band at 376 nm forming an isosbestic point ' . When excited at the isosbestic point at 356 nm there is a band at 407 nm that shifts to 443 nm with increasing concentrations of zinc. In determining the apparent K^ of this indicator a zinc/EGTA buffer system was used to achieve free zinc values ranging from 0 to 11 nM. The ratio of 443/395 nm was plotted versus zinc concentration giving rise to a IQ of 2.2 ± 0.1 nM '*. Two photon excitation microscopy was used to determine if this indicator is usefril in biological applications. When Zinbo-5 is added to mouse fibroblast cells, there is very minimal fluorescence seen (excitation 710 nm) even after the brightness of the image had been increased '^^. Zinbo-5 is responsive to changes in zinc concentration as addition of zinc with the ionophore pyrithione leads to an increase in the ratio image; whereas addition of TPEN caused the fluorescence to decrease and thus match the original ratio image. However all of this was performed in fixed cells (4% formaldehyde on ice for 20 min) and the staining process was also done on ice which thus may not reflect the actual resting level of free zinc inside these cells.

18.6. LANTHANIDE CHEMOSENSORS FOR ZINC

Lanthanides exhibit good spectral properties for sensing such as long luminescence lifetimes (up to msec), Stokes' shift larger than 200 nm, narrowband emission, and high water solubility ^ . Lanthanides are also suitable for time-resolved fluorescence (TRF) measurements, which have the ability to suppress background resulting in a better signal-to-noise ratio. The europium (Eu^^) complex in particular is useful in this area. The lanthanide requires a chromophore for high luminescence which will absorb the light and undergo ISC (intersystem crossing) to populate the triplet state. Then an intramolecular energy transfer occurs to the lanthanide which then emits at a particular wavelength. However it should also be mentioned that this type of fluorophore is also susceptible to water quenching, which would be difficult to overcome in a biological system. In the study by Hanaoka, et al., the chromophore is a quinoline-containing TPEN ligand which also acts as a sensitizer. This is combined with the Eu^^ - diethylenetriaminepentaacetic acid (DTPA) complex forming [Eu-7] ^ . This complex was characterized using time-delayed luminescence with a delay of 0.05 ms in aqueous solution. Upon addition of zinc, the luminescence emission increased by 8.5-fold. For IQ determinations the zinc concentration was controlled by using a NTA system


J L^ coo

Eu* (Eu-71

Figure 18.4. Structure of the lanthanide chemosensor [Eu-7],using a quinolyl ligand as the chromophore; redrawn from ^ .

below 398 nM but was unbuffered above 200 jiM zinc ^ . The apparent IQ for [Eu-7] was calculated to be 59 nM, which is less tight than expected ^ . The luminescence of [Eu-7] is stable at physiological pH. Various metals were added to [Eu-7] (50 |iM) in HEPES buffer in order to determine the selectivity. The presence of 5 mM Ca ^ and Mg^^ did not result in any luminescence emission enhancement. The same result was seen with the addition of 50 jiM Cu^^ Ni' ^ Co' ^ Fe^^ and Mn^l In the presence of zinc the luminescence of Eu-7 was quenched upon the addition of Ni , Cu , and Co . A large enhancement of luminescence was seen upon addition of Cd^^ to [Eu-7] ^ . While it is advantageous to use this indicator due to its long emission wavelength and lifetime, unless a lifetime change occurs it remains an intensity-based indicator with the attendant shortcomings.

18.7. EXCITED-STATE INTRAMOLECULAR PROTON TRANSFER

Ratiometric indicators offer a better solution for analyte detection than simple intensity indicators due to their dependence on the cation concentration, but not on indicator concentration, spectral properties of the instrument, etc. Henary et ah ^ describe a new emission ratiometric indicator that is based on inhibition of excited-state intramolecular proton transfer (ESIPT). ESIPT occurs when a proton is transferred either to a base or another atom on the molecule in the excited state which results in a large shift in emission. In the process described by Henary, et ah, they use a benzimidazole derivative which undergoes ESIPT ^ . When the fluorophore is in the excited state, the proton becomes highly acidic, thus the coordination of a metal cation will displace the proton and thus inhibit ESIPT ^ . Ligands 6b, 7b, 8b, and 9b are water soluble and the pKa of the sulfonamide nitrogen atom is above 8.0 for each one. The absorption spectra for 6b-9b are very similar in that they all have a peak around 300 nm in the protonated form and a red shifted band around 330 nm in the deprotonated state in aqueous solution. Upon excitation at the peak wavelength, 6b-9b exhibit a single band at 460 nm ' . In the presence of zinc the absorbance spectra of 6b, 7b, and 9b resemble the spectra seen in the

410 R.A. BOZYM ETAL

c / \

Figure 18.5. Structures of ligands 6b and 9b; redrawn from ''''.

deprotonated state, indicating loss of the nitrogen proton and zinc binding. Interestingly 8b does not contain a red shifted peak in the absorbance spectrum indicating that the nitrogen atom is protonated even in the presence of zinc and does not coordinate to zinc. The emission spectra in the presence of zinc resulted in a large blue shift of the peak emission around 460 nm to a peak around 400 nm with an isosbestic point at 428 nm ^ . In calculating the zinc binding affinities of 6b and 7b, metal ion buffers were evidently not used which calls into question the accuracy of the measured affinities. For 8b Zn/EGTA buffers were used to calculate the affinity for zinc and for 9b Zn/HEDTA (N-(2-hydroxyethyl)ethylenediaminetriacetic acid) buffers were used. The binding affinities reported for 6b-9b are fxM, mM, nM, and pM, respectively ^ . Calibration curves for 6b-9b were performed by taking the ratio of emission at 400 and 500 nm, indicating 9b and 8b can be used to measure zinc between 0.2 - 20 pM and 1-10 nM respectively. 7b has very low affinity for zinc thus it is not useful for biological applications. 6b is suitable for zinc concentrations between 30 |xM - 3 mM .

18.8. PEPTIDES AS ZINC INDICATORS

Other zinc indicators have been developed which utilize the properties of zinc finger domains ^ " . Although these indicators exhibited high selectivity for zinc the Imperiali group expanded their development to smaller peptides for easier use in biological systems. The second group of indicators developed by this group were only seven amino acids long with the reporting group being 8-methoxy-quinoline ^\ These indicators underwent chelation enhanced fluorescence (CHEF) upon zinc binding. More recently the Imperiali group has developed peptides with tunable affinity for zinc in the nanomolar to micromolar range ^ . The indicators make use of three components: a Zn ligand amino acid, a p-tum sequence, and the fluorophore containing zinc ligand. The fluorophore is incorporated into the peptide via a novel amino acid Sox, which contains 8-hydroxy-5-(N,N-dimethylsulfonamido)-2-methylquinoline ^ . Each


of the peptides synthesized are only 6-8 amino acids in length, all of which include 2 serine residues for solubility and cysteine or histidine residues for binding zinc ^ . Upon addition of zinc the Sox peptides emit at 500 nm with excitation at 360 nm exhibiting a 30-fold increase in fluorescence through CHEF. Metal competition experiments were performed to determine if these peptides also respond to other divalent cations either in a positive (cation binding results in fluorescence) or negative manner (cation binding compromises the ability of the peptide to bind zinc). Only a few Sox peptides (P7, PI2, and PI3) exhibited fluorescence in the presence of ImM Ca^^ and Mg^^ whereas all the peptides gave a positive response to Cd^ . However, Cu^^ prohibits Zn^^ binding for all peptides, Ni ^ quenches peptides that use imidazole and carboxylate ligands, and Mn^^ and Co^^ compete with the binding of zinc for peptides with weaker affinity for zinc ^ . It is possible to increase the zinc affinities of the Sox peptides by making a more flexible turn or increasing the number of ligands incorporated into the peptide. When combined with one another these indicators may be used to determine a range of zinc concentration when one peptide gives a response and others do not. However it would be difficult to utilize these Sox peptides over a broad range using just one color.

18.9. CARBONIC ANHYDRASE AS A BIOSENSOR FOR ZINC

We have been developing a fluorescent indicator system using apocarbonic anhydrase as a sensor transducer ' ' . Human carbonic anhydrase II (CA) exhibits a 4 pM affinity for zinc at pH 7.5 and its response is unaffected 10 mM Ca^^ or 50 mM Mg^^ , levels significantly higher than those encountered in vivo ^ . In using a biological molecule we are uniquely able to alter or adapt the molecule to our needs through site-directed mutagenesis ' . Single amino acid substitutions in the sequence can improve the selectivity, sensitivity, and response-time of carbonic anhydrase ^ , making it a uniquely flexible and responsive indicator. For instance, a variant CA with a single amino acid mutation (E117A) exhibits a several hundred-fold faster association rate constant than the wild type, with only a 10-fold drop in affinity ^ . When compared with other indicators, CA has excellent selectivity (100-fold) over cadmium ^ .

Apocarbonic anhydrase has been adapted to an excitation ratiometric fluorescent biosensor based on fluorescence resonance energy transfer (FRET) from a zinc-bound aryl sulfonamide to a fluorescent label on the protein ^ . Zinc levels are measured by taking the ratio of intensities at two different excitation wavelengths ^ . This approach, as depicted in Figure 18.6, is based on FRET from the zinc-dependent binding of Dapoxyl sulfonamide to the fluorescent label, Alexa Fluor 594, covalently attached to the protein. In the absence of zinc, Dapoxyl sulfonamide doesn't bind and there is only very weak emission from free Dapoxyl sulfonamide at 617 nm upon excitation at 365 nm. When zinc is bound to the active site of the enzyme, Dapoxyl sulfonamide binds to zinc replacing water as the fourth ligand, and there is strong emission at 617 nm with UV excitation, due to FRET from bound Dapoxyl exciting

412 R,A,BOZYM ETAL

H36C-CA H36C-CA

TAT • Zn'.

WEAK y N/ V 617nm*^ \ ^ n /

TAT

STRONG .:/n«mV CA KJU

e - 5 0 0 0

EM.eOOnm QY«0.01

QY-1

OAPOXYL N ^ P (BOUNO>

O J SULFONAMIDE (FREE)

EXC.365nm

Figure 18.6. Schematic of ratiometric zinc determination with apoTAT-H36C-Alexa Fluor 594 carbonic anhydrase and Dapoxyl sulfonamide. In the absence of zinc, Dapoxyl sulfonamide does not bind to CA, therefore no FRET occurs to the fluorescent label on the protein and very weak emission at 617 nm is observed. In the presence of zinc, Dapoxyl sulfonamide binds, is excited with UV light, and FRET occurs from Dapoxyl to the Alexa Fluor, thus exciting the fluorescent label causing emission to occur at 617 nm. The emission when excited with 365 nm is normalized to the amount of labeled protein by directly exciting Alexa Fluor 594 with 543 nm and dividing the two intensities.

Alexa Fluor. The change in emission when excited in the UV is normalized to the amount of labeled protein present by exciting the Alexa Fluor 594 directly (543 nm, where Dapoxyl sulfonamide does not absorb) and dividing the two intensities. Additional advantages of this approach (beyond being ratiometric) are that there is no interference from the emission of Dapoxyl sulfonamide bound to adventitious carbonic anhydrase, lipids, or other proteins because it emits in the blue, and the sensitive range and kinetics can be adjusted by small changes in the protein structure. Also the recognition and transduction elements are capable of being in separate moieties of the molecule; the fluorophore need not be the metal ligand ^ .

This approach was tailored to intracellular use by attaching a TAT (transactivator of transcription) peptide to carbonic anhydrase which then induces the cells to take up the protein ^ . The TAT peptide is encoded by HIV-1 and contains an 11 amino acid protein transduction domain (PTD) allowing the cell to carry the protein across the membrane and into the cytoplasm. Dapoxyl sulfonamide penetrates the cell readily. Thus intracellular free zinc levels are able to be imaged and quantitated using this ratiometric approach. Calibrations for this system were performed using both the steady state fluorometer and the epifluorescence microscope. For calibrations performed on the microscope, wells of a 1536 well-plate were filled with apoH36C-AF594-CA, Dapoxyl, and buffers of increasing free zinc concentration. The wells were imaged using the Nikon 4x objective. The mean ratio for each well was calculated and used to plot the calibration curve from the microscope against the one acquired on the fluorometer (Figure 18.7). The curve from the microscope reveals a IQ of 70 ±15 pM whereas the fitted Kj using the fluorometer is 137 ±18 pM. The optical trains of the microscope (using filters) and the fluorometer used for the cuvette measurements (using monochromators) are quite different, such that agreement to within a factor of two appears


- 0.8

i c

5 ^ 0-7

H 3 6 C - A F 5 9 6

DAPOXYL

EM. 617 nm

? 9

10-1* j(j-iJ ^^-10 jo^«

FREE [ Z n * * ] . N A N O M O L A R

Figure 18.7. Zinc dependent ratio of 366 nm over 548 nm with emission at 618 nm of apo-H36C-AF594 and Dapoxyl sulfonamide performed on the microscope in a 1536-well plate (o) and on a steady state fluorometer in a cuvette (n). Kd = 70 ±15 pM (o) and 137±18 pM (D).

Figure 18.8. Microscope calibration with a 1536 well plate. Ratio images of each well (365/543 nm) containing zinc ion buffers, apoH36C-AF594 CA and Dapoxyl sulfonamide; pseudo colored for visual comparison. (See color insert section.)

satisfactory. The ratio increases with increasing zinc concentration by almost 50%, which is slightly less than that achieved with the fluorometer. A representative set of wells was used to create Figure 18.8 in order to visually compare the intensity ratio of different zinc concentrations. The ratio images of the wells were pseudo colored using a rainbow color scheme in IPLab. The calibration color bar was formed using the values in the calibration curve and matching them with the well plate.

Measuring and imaging the free zinc level in cells is accomplished by adding apoTAT-H36C-AF594 CA and Dapoxyl sulfonamide. The TAT peptide quickly transports the label into the cells (incubation time of only 15 min) with bright fluorescence seen immediately. Figure 18.9 shows a representative set of cells after staining with apoTAT-H36C-AF594 CA and Dapoxyl sulfonamide. In the excitation 543 and 365 nm images (emission 617 nm), fluorescence is uniformly distributed throughout the cells. Overall the free zinc is approximately 5pM, which is low, but not zero and well above the femtomolar levels proposed for prokaryotic cells ^ . Similar results were obtained using the E117A variant which has a ten-fold lower affinity for zinc and faster zinc equilibration ^ , reinforcing the conclusion that our measurement is near equilibrium conditions and therefore the free zinc in cells is well above femtomolar.

414 R.A. BOZYM ETAL

Figure 18.9. Measuring free zinc inside of PC-12 cells with apoTAT-H36C-AF594 and Dapoxyl sulfonamide. Top left: bright field; top right: ratio image pseudo color rainbow; bottom left: excitation 543 nm, emission 617 nm, exposure time 20 ms, pseudo color red; bottom right: excitation 365 nm, emission 617 nm, exposure time 500 ms, pseudo color red. The calibration bar (at right) indicates the level of cytoplasmic free zinc. (See color insert section.)

18.10. CONCLUSION

The number of papers published on developing zinc indicators and the importance of zinc in biochemistry has been steadily increasing. There is substantial incentive to develop selective, sensitive, and quantitative zinc indicators in order to elucidate the many unknown roles of zinc in the body. How zinc is sequestered and bound to the vast number of enzymes is still unknown, although the likelihood of a cfuatalytic delivery mechanism is very high. The roles of zinc in neurodegenerative disorders (Alzheimer's), intracellular signaling, and apoptosis are the subject of substantial investigation which will be elucidated by the use of zinc indicators just as calcium indicators opened the door to calcium signaling. It is important to use the appropriate indicator for experiments. The kinetic parameters are also of importance in choosing an appropriate indicator. Faster association/dissociation rates are needed for measurements where the zinc concentration is going to change rapidly, as in neuronal activity. For a quantitative measurement of the free zinc level a ratiometric probe is most useful as it is much less affected by environmental factors (indicator concentration, excitation intensity, etc.). Intensity-based indicators are susceptible to these issues and it is difficult to obtain reliable quantitation of free zinc using these indicators. The comparable advantages of lifetime- and polarization-based indicators for these applications should not be overlooked. How the indicator is transported into cells is a main focus in many zinc indicator studies. Many indicators are not cell-permeable with some being injected into cells ^ , others use the AM ester form to get through the cell membrane, meanwhile others are developed to be transported into the cell. We have reported the novel use of the TAT peptide in transporting our CA-based indicator into cells. The TAT peptide allows the cell to readily take in the protein in an efficient manner without any apparent adverse affects to the cell, or need for disruptive procedures such as microinjection. Our


system is also excitation ratiometric which allows for a more accurate quantitation of the free zinc level in the cell.

18.11. REFERENCES

1. Vallee, B. L.; Falchuk, K. H., The biochemical basis of zinc physiology. Physiological Reviews 1993,13, (\),19-l\S.

2. Berg, J. M.; Shi, Y., The galvanization of biology: a growing appreciation for the roles of zinc. Science 1996, 271, 1081-1085.

3. Vallee, B. L.; Auld, D. S., Zinc coordination, function, and structure of zinc enzymes and other proteins. Biochemistry 1990, 29, (24), 5647-59.

4. O'Halloran, T. V., Transition metals in control of gene expression. Science 1993, 261, (715-725).

5. Fraker, P. J.; Telford, W. G., A reappraisal of the role of zinc in life and death decisions of cells. Proceedings of the Society for Experimental Biology and Medicine 1997, 215, 229-236.

6. Costello, L. C; Franklin, R. B., Novel role of zinc in the regulation of prostate citrate metabolism and its implications in prostate cancer. The Prostate 1998, 35, 285 - 296.

7. Venter, J. C ; Adams, M. D.; Myers, E. W.; Li, P. W.; Mural, R. J.; Sutton, G. G.; Smith, H. O.; Yandell, M.; Evans, C. A.; Holt, R. A.; Gocayne, J. D.; Amanatides, P.; Ballew, R. M. Huson, D. H.; Wortman, J. R.; Zhang, Q.; Kodira, C. D.; Zheng, X. H.; Chen, L.; Skupski, M. Subramanian, G.; Thomas, P. D.; Zhang, J.; Gabor Miklos, G. L.; Nelson, C ; Broder, S. Clark, A. G.; Nadeau, J.; McKusick, V. A.; Zinder, N.; Levine, A. J.; Roberts, R. J.; Simon, M.; Slayman, C ; Hunkapiller, M.; Bolanos, R.; Delcher, A.; Dew, I.; Fasulo, D.; Flanigan M.; Florea, L.; Halpem, A.; Hannenhalli, S.; Kravitz, S.; Levy, S.; Mobarry, C ; Reinert, K. Remington, K.; Abu-Threideh, J.; Beasley, E.; Biddick, K.; Bonazzi, V.; Brandon, R.; Cargill M.; Chandramouliswaran, L; Charlab, R.; Chaturvedi, K.; Deng, Z.; Di Francesco, V.; Dunn, P.; Eilbeck, K.; Evangelista, C ; Gabrielian, A. E.; Gan, W.; Ge, W.; Gong, F.; Gu, Z.; Guan, P.; Heiman, T. J.; Higgins, M. E.; Ji, R. R.; Ke, Z.; Ketchum, K. A.; Lai, Z.; Lei, Y.; Li, Z.; Li J.; Liang, Y.; Lin, X.; Lu, F.; Merkulov, G. V.; Milshina, N.; Moore, H. M.; Naik, A. K. Narayan, V. A.; Neelam, B.; Nusskem, D.; Rusch, D. B.; Salzberg, S.; Shao, W.; Shue, B. Sun, J.; Wang, Z.; Wang, A.; Wang, X.; Wang, J.; Wei, M.; Wides, R.; Xiao, C; Yan, C. Yao, A.; Ye, J.; Zhan, M.; Zhang, W.; Zhang, H.; Zhao, Q.; Zheng, L.; Zhong, F.; Zhong, W. Zhu, S.; Zhao, S.; Gilbert, D.; Baumhueter, S.; Spier, G.; Carter, C ; Cravchik, A.; Woodage T.; Ali, F.; An, H.; Awe, A.; Baldwin, D.; Baden, H.; Bamstead, M.; Barrow, L; Beeson, K. Busam, D.; Carver, A.; Center, A.; Cheng, M. L.; Curry, L.; Danaher, S.; Davenport, L. Desilets, R.; Dietz, S.; Dodson, K.; Doup, L.; Ferriera, S.; Garg, N.; Gluecksmann, A.; Hart, B.; Haynes, J.; Haynes, C; Heiner, C; Hladun, S.; Hostin, D.; Houck, J.; Howland, T. Ibegwam, C; Johnson, J.; Kalush, F.; Kline, L.; Koduru, S.; Love, A.; Mann, F.; May, D.: McCawley, S.; Mcintosh, T.; McMuUen, L; Moy, M.; Moy, L.; Murphy, B.; Nelson, K. Pfannkoch, C; Pratts, E.; Puri, V.; Qureshi, H.; Reardon, M.; Rodriguez, R.; Rogers, Y. H. Romblad, D.; Ruhfel, B.; Scott, R.; Sitter, C; Smallwood, M.; Stewart, E.; Strong, R.; Suh, E.; Thomas, R.; Tint, N. N.; Tse, S.; Vech, C; Wang, G.; Wetter, J.; Williams, S.; Williams, M. Windsor, S.; Winn-Deen, E.; Wolfe, K.; Zaveri, J.; Zaveri, K.; Abril, J. F.; Guigo, R. Campbell, M. J.; Sjolander, K. V.; Karlak, B.; Kejariwal, A.; Mi, H.; Lazareva, B.; Hatton, T. Narechania, A.; Diemer, K.; Muruganujan, A.; Guo, N.; Sato, S.; Bafna, V.; Istrail, S. Lippert, R.; Schwartz, R.; Walenz, B.; Yooseph, S.; Allen, D.; Basu, A.; Baxendale, J.; Blick, L.; Caminha, M.; Cames-Stine, J.; Caulk, P.; Chiang, Y. H.; Coyne, M.; Dahlke, C ; Mays, A. Dombroski, M.; Donnelly, M.; Ely, D.; Esparham, S.; Fosler, C ; Gire, H.; Glanowski, S. Glasser, K.; Glodek, A.; Gorokhov, M.; Graham, K.; Gropman, B.; Harris, M.; Heil, J. Henderson, S.; Hoover, J.; Jennings, D.; Jordan, C ; Jordan, J.; Kasha, J.; Kagan, L.; Kraft, C. Levitsky, A.; Lewis, M.; Liu, X.; Lopez, J.; Ma, D.; Majoros, W.; McDaniel, J.; Murphy, S. Newman, M.; Nguyen, T.; Nguyen, N.; Nodell, M.; Pan, S.; Peck, J.; Peterson, M.; Rowe, W. Sanders, R.; Scott, J.; Simpson, M.; Smith, T.; Sprague, A.; Stockwell, T.; Turner, R.; Venter, E.; Wang, M.; Wen, M.; Wu, D.; Wu, M.; Xia, A.; Zandieh, A.; Zhu, X., The sequence of the human genome. Science 2001, 291, (5507), 1304-51.

8. Frederickson, C. J., The neurobiology of zinc and of zinc-containing neurons. International Review of Neurobiology 1989, 31, 145-238.

416 R.A. BOZYM ETAL

9. Frederickson, C. J.; Hernandez, M. D.; McGinty, J. F., Translocation of zinc may contribute to seizure-induced death of neurons. Brain Research 1989,480, 317-321.

10. Frederickson, C. J.; Bush, A. I., Synaptically released zinc: Physiological functions and pathological effects. BioMetals 2001, 14, (3 - 4), 353 - 366.

11. Choi, D. W.; Koh, J. Y., Zinc and brain injury. Annual Review of Neuroscience 1998, 21, 347-375.

12. Vogt, K.; Mellor, J.; TOng, G.; Nicoll, R., The actions of synaptically -released zinc at hippocampal mossy fiber synapses. Neuron 2000, 26, 187 - 196.

13. Hambidge, M., Human zinc deficiency. JNutr 2000, 130, (5S Suppl), 1344S-9S. 14. Palmiter, R. D.; Cole, T. B.; Quaife, C. J.; Findley, S. D., ZnT-3, a putative transporter of zinc

into synaptic vesicles. Proc Natl Acad Sci USA 1996, 93, (25), 14934-9. 15. Harrison, N. L.; Gibbons, S. J., Zn2+: an endogenous modulator of ligand- and voltage-gated

ion channels. Neuropharmacology 1994, 33, (8), 935-52. 16. Smart, T. G.; Xie, X.; Krishek, B. J., Modulation of inhibitory and excitatory amino acid

receptor ion channels by zinc. Prog Neurobiol 1994,42, (3), 393-41. 17. Yokoyama, M.; Koh, J.; Choi, D. W., Brief exposure to zinc is toxic to cortical neurons.

Neuroscience Letters 1986, 71, 351-355. 18. Tonder, N.; Johansen, F. F.; Frederickson, C. J.; Zimmer, J.; Diemer, N. H., Possible role of

zinc in the selective degeneration of dentate hilar neurons after cerebral ischemia in the adult rat. Neuroscience Letters 1990, 109, 247-252.

19. Koh, J. Y.; Suh, S. W.; Gwag, B. J.; He, Y. Y.; Hsu, C. Y.; Choi, D. W., The role of zinc in selective neuronal death after transient global cerebral ischemia. Science 1996, 272, 1013-1016.

20. Truong-Tran, A. Q.; Carter, J.; Ruffin, R. E.; Zalewski, P. D., The role of zinc in caspase activation and apoptotic cell death. Biometals 2001, 14, (3-4), 315-30.

21. Canzoneiro, L. M. T.; Manzerra, P.; Sheline, C. T.; Choi, D. W., Membrane -permeant chelators can attenuate Zn^^-inducced cortical neuronal death. Neuropharmacology 2003, 45, 420-428.

22. Truong-Tran, A. Q.; Ruffin, R. E.; Zalewski, P. D., Visualization of labile zinc and its role in apoptosis of primary airway epithelial cells and cell lines. Am J Physiol Lung Cell Mol Physiol 2000, 279, (6), LI 172-83.

23. Cuajungco, M. P.; lees, G. J., Diverse effects of metal chelating agents on the neuronal cytotoxicity of zinc in the hippocampus. Brain Research 1998, 799, 97-107.

24. Aizenman, E.; Stout, A. K.; Hartnett, K. A.; Dinely, K. E.; McLaughlin, B.; Reynolds, L J., Induction of neuronal apoptosis by thiol oxidation:putative role of intracellular zinc release. Journal ojNeurochemistry 2000, 75, (5), 1878 - 1889.

25. Jiang, D.; Sullivan, P. G.; Sensi, S. L.; Steward, O.; Weiss, J. H., Zn(2+) induces permeability transition pore opening and release of pro-apoptotic peptides from neuronal mitochondria. J Biol Chem 2001, 276, (50), 47524-9.

26. Bossy-Wetzel, E.; Talantova, M. V.; Lee, W. D.; Scholzke, M. N.; Harrop, A.; Mathews, E.; Gotz, T.; Han, J.; EUisman, M.; Perkins, G. A.; Lipton, S. A., Crosstalk between Nitric Oxide and Zinc Pathways to Neuronal Cell Death Involving Mitochondrial Dysfuntion and p38-Activated K^ Channels. Neuron 2004,41, (February 5), 351-365.

27. Maret, W., The function of zinc metallothionein: a link between cellular zinc and redox state. J Nutr 2000, 130, (5S Suppl), 1455S-8S.

28. Perkins, D. J., A study of the effect of amino acid structure on the stabilities of the complexes formed with metals of Group II of the periodic classification. Biochemical Journal 1953, 55, 649 - 652.

29. Prasad, A. S.; Oberleas, D., Zinc in human serum: evidence for an amino acid-bound fraction. Journal of Laboratory and CLinical Medicine 1968, 1006.

30. Finney, L. A.; O'Halloran, T. V., Transition metal speciation in the cell: insights from the chemistry of metal ion receptors. Science 2003, 300, 931 - 936.

31. Gaither, L. A.; Eide, D. J., Eukaryotic zinc transporters and their regulation. Biometals 2001, 14, (3-4), 251-70.

32. Zhao, H.; Eide, D., The ZRT2 gene encodes the low affinity zinc transporter in Saccharomyces cerevisiae. J Biol Chem 1996, 271, (38), 23203-10.


33. Zhao, H.; Eide, D., The yeast ZRTl gene encodes the zinc transporter protein of a high-affinity uptake system induced by zinc limitation. Proc Natl Acad Sci USA 1996, 93, (6), 2454-8.

34. Grotz, N.; Fox, T.; Connolly, E.; Park, W.; Guerinot, M. L.; Eide, D., Identification of a family of zinc transporter genes from Arabidopsis that respond to zinc deficiency. Proc Natl Acad Sci USA 1998, 95, (12), 7220-4.

35. Coyle, P.; Philcox, J. C; Carey, L. C; Rofe, A. M., Metallothionein: the multipurpose protein. Cell Mol Life Sci 2002, 59, (4), 627-47.

36. Colvin, R. A., pH dependence and compartmentalization of zinc transported across plasma membrane of rat cortical neurons. American Journal of Physiology Cell Physiology 2002, 282, C317-C329.

37. Perez-Clausell, J.; Danscher, G., Intravesicular localization of zinc in rat telencephalic boutons. A histochemical study. Brain Res 1985, 337, (1), 91-8.

38. Beaulieu, C ; Dyck, R.; Cynader, M., Enrichment of glutamate in zinc-containing terminals of the cat visual cortex. Neuroreport 1992, 3, (10), 861-4.

39. Zalewski, P. D.; Millard, S. H.; Forbes, I. J.; Kapaniris, O.; Slavotinek, A.; Betts, W. H.; Ward, A. D.; Lincoln, S. F.; Mahadevan, I., Video image analysis of labile zinc in viable pancreatic islet cells using a specific fluorescent probe for zinc. J Histochem Cytochem 1994, 42, (7), 877-84.

40. Andrews, J. C ; Nolan, J. P.; Hammerstedt, R. H.; Bavister, B. D., Characterization of N-(6-methoxy-8-quinolyl)-p-toluenesulfonamide for the detection of zinc in living sperm cells. Cytometry 1995, 21, (2), 153-9.

41. Danscher, G.; Zimmer, J., An improved Timm sulphide silver method for light and electron microscopic localization of heavy metals in biological tissues. Histochemistry 1978, 55, (1), 27-40.

42. Rae, T. D.; Schmidt, P. J.; Pufahl, R. A.; Culotta, V. C; O'Halloran, T. V., Undetectable intracellular free copper: the requirement of a copper chaperone for superoxide dismutase. Science 1999, 284, 805 - 808.

43. Hunt, J. A.; Fierke, C. A., Selection of carbonic anhydrase variants displayed on phage: aromatic residues in zinc binding site enhance metal affinity and equilibration kinetics. Journal of Biological Chemistry 1997, 272, (33), 20364-20372.

44. Bozym, R. A.; Zeng, H. H.; Cramer, M. L.; Stoddard, A.; Fierke, C. A.; Thompson, R. B. In In vivo and intracellular sensing and imagine of free zinc ion. Advanced Biomedical and Clinical Diagnostic Systems II, San Jose, CA, 2004; Cohn, G. E., Grundfest, W.S., Benaron, D.A., Vo-Dinh, T., Ed. Proc. SPIE Int. Soc. Opt. Eng.: San Jose, CA, 2004; pp 34-38.

45. Hitomi, Y.; Outten, C. E.; O'Halloran, T. V., Extreme zinc-binding thermodynamics of the metal sensor/regulator protein, ZntR. J Am Chem Soc 2001, 123, (35), 8614-5.

46. Hunt, J. B.; Rhee, M. J.; Storm, C. B., A rapid and convenient preparation of apocarbonic anhydrase. Analytical Biochemistry 1977, 79, 614-617.

47. Budde, T.; Minta, A.; White, J. A.; Kay, A. R., Imaging free zinc in synaptic terminals in live hippocampal slices. Neuroscience 1997, 79, (2), 347-358.

48. Canzoniero, L. M. T.; Sensi, S. L.; Choi, D. W., Measurement of intracellular free zinc in living neurons. Neurobiology of Disease 1997, 4, 275 - 279.

49. Gee, K. R.; Zhou, Z. L.; Ton-That, D.; Sensi, S. L.; Weiss, J. H., Measuring zinc in living cells. A new generation of sensitive and selective fluorescent probes. Cell Calcium 2002, 31, (5), 245-51.

50. Chang, C. J.; Jaworski, J.; Nolan, E. M.; Sheng, M.; Lippard, S. J., A tautomeric zinc sensor for ratiometric fluorescence imaging: application to nitric oxide-induced release of intracellular zinc. Proc Natl Acad Sci USA 2004, 101, (5), 1129-34.

51. Walkup, G. K.; Burdette, S. C; Lippard, S. J.; Tsien, R. Y., A new cell-permeable fluorescent probe for Zn2+. Journal of the American Chemical Society 2000, 122, 5644 - 5645.

52. Burdette, S. C; Walkup, G. K.; Spingler, B.; Tsien, R. Y.; Lippard, S. J., Fluorescent Sensors for Zn2+ based on a fluorescein platform: synthesis, properties, and intracellular distribution. Journal of the American Chemical Society 2001, 123, 7831 - 7841.

53. Frederickson, C. J.; Kasarskis, E. J.; D. Ringo; Frederickson, R. E., A quinoHne fluorescence method for visualizing and assaying histochemically reactive zinc (bouton zinc) in the brain. Journal of Neuroscience Methods 1987, 20, 91-103.

418 R.A. BOZYM ETAL

54. Kimura, E.; Aoki, S., Chemistry of zinc(II) fluorophore sensors. Biometals 2001, 14, (3-4), 191-204.

55. Fahmi, C. J.; O'Halloran, T. V., Aqueous coordination chemistry of quinoline-based fluorescence probes for the biological chemistry of zinc. Journal of the American Chemical Society 1999, 121, 11448 - 11458.

56. Zalewski, P. D.; Forbes, I. J.; Betts, W. H., Correlation of apoptosis with change in intracellular labile Zn(II) using Zinquin [(2-methyl-8-p-toluenesulphonamido-6-quinolyloxy)acetic acid], a new specific fluorescent probe for Zn(II). Biochemical Journal 1993, 296, 403-408.

57. Hendrickson, K. M.; Geue, J. P.; Wyness, O.; Lincoln, S. F.; Ward, A. D., Coordination and fluorescence of the intracellular Zn2+ probe [2-methyl-8-(4-toluenesulfonamido)-6-quinolyloxyjacetic acid (Zinquin A) in ternary Zn2+ complexes. J Am Chem Soc 2003, 125, (13), 3889-95.

58. Haugland, R. P., Handbook of Fluorescent Probes and Research Chemicals. Sixth ed.; Molecular Probes, Inc.: Eugene, Oregon, 1996; p 679.

59. Burdette, S. C ; Frederickson, C. J.; Bu, W.; Lippard, S. J., ZP4, an improved neuronal Zn2+ sensor of the ZinPyr family. Journal of the American Chemical Society 2003, 125, 1778 -1787.

60. Nolan, E. M.; Burdette, S. C ; Harvey, J. H.; Hilderbrand, S. A.; Lippard, S. J., Synthesis and characterization of zinc sensors based on a monosubstituted fluorescein platform. Inorg Chem 2004, 43, (8), 2624-35.

61. Chang, C. J.; Nolan, E. M.; Jaworski, J.; Burdette, S. C ; Sheng, M.; Lippard, S. J., Bright fluorescent chemosensor platforms for imaging endogenous pools of neuronal zinc. Chem Biol 2004, 11, (2), 203-10.

62. Chang, C. J.; Nolan, E. M.; Jaworski, J.; Okamoto, K.; Hayashi, Y.; Sheng, M.; Lippard, S. J., ZP8, a neuronal zinc sensor with improved dynamic range; imaging zinc in hippocampal slices with two-photon microscopy. Inorg Chem 2004, 43, (21), 6774-9.

63. Nolan, E. M.; Lippard, S. J., The zinspy family of fluorescent zinc sensors: syntheses and spectroscopic investigations. Inorg Chem 2004, 43, (26), 8310-7.

64. Hirano, T.; Kikuchi, K.; Urano, Y.; Nagano, T., Improvement and biological applications of fluorescent probes for zinc, ZnAFs. J Am Chem Soc 2002, 124, (23), 6555-62.

65. Sun, W.-C; Gee, K. R.; Klaubert, D. H.; Haugland, R. P., Synthesis of Fluorinated Fluoresceins, y Org Chem 1997, 62, (19), 6469-6475.

66. Maruyama, S.; Kikuchi, K.; Hirano, T.; Urano, Y.; Nagano, T., A novel, cell-permeable, fluorescent probe for ratiometric imaging of zinc ion. J Am Chem Soc 2002, 124, (36), 10650-1.

67. Hirano, T.; Kikuchi, K.; Urano, Y.; Higuchi, T.; Nagano, T., Novel Zinc Fluorescent Probes Excitable with Visible Light for Biological Applications We thank Prof E. Kimura and Prof T. Koike for advice on the chemistry of macrocyclic polyamines. Angew Chem Int Ed Engl 2000,39,(6), 1052-1054.

68. Atar, D.; Backx, P. H.; Appel, M. M.; Gao, W. D.; Marban, E., Excitation-transcription coupling mediated by zinc influx through voltage-dependent calcium channels. J Biol Chem 1995, 270, (6), 2473-7.

69. Simons, T. J., Measurement of free Zn2+ ion concentration with the fluorescent probe mag-fura-2 (furaptra). JBiochem Biophys Methods 1993, 27, (1), 25-37.

70. Gee, K. R.; Zhou, Z. L.; Qian, W. J.; Kennedy, R., Detection and imaging of zinc secretion from pancreatic beta-cells using a new fluorescent zinc indicator. J Am Chem Soc 2002, 124, (5), 776-8.

71. Kay, A. R., Evidence for chelatable zinc in the extracellular space of the hippocampus, but little evidence for synaptic release of Zn. Journal of Neuroscience 2003, 23, (17), 6847 - 6855.

72. Sensi, S. L.; Ton-That, D.; Weiss, J. H.; Rothe, A.; Gee, K. R., A new mitochondrial fluorescent zinc sensor. Cell Calcium 2003, 34, (3), 281-4.

73. Thompson, R. B.; Peterson, D.; Mahoney, W.; Cramer, M.; Maliwal, B. P.; Suh, S. W.; Frederickson, C. J., Fluorescent zinc indicators for neurobiology. Journal of Neuroscience Methods 2002, 118,63-75.

74. Taki, M.; Wolford, J. L.; O'Halloran, T. V., Emission ratiometric imaging of intracellular zinc: design of a benzoxazole fluorescent sensor and its application in two-photon microscopy. J Am Chem Soc 2004, 126, (3), 712-3.


75. Hanaoka, K.; Kikuchi, K.; Kojima, H.; Urano, Y.; Nagano, T., Development of a zinc ion-selective luminescent lanthanide chemosensor for biological applications. J Am Chem Soc 2004, 126,(39), 12470-6.

76. Martell, A. E.; Smith, R. M. NIST Critically selected stability constants of metal complexes; Nation Insitute of Standards and Technology, US Department of Commerce: Gaithersburg, MD,2001.

77. Henary, M. M.; Wu, Y.; Fahmi, C. J., Zinc(II)-selective ratiometric fluorescent sensors based on inhibition of excited-state intramolecular proton transfer. Chemistry 2004, 10, (12), 3015-25.

78. Walkup, G. a. I. B., Fluorescent chemosensors for divalent zinc based on zinc finger domains. Enhanced oxidative stability, metal binding affinity, and structural and functional characterization. J Am Chem Soc 1997, 119, 3443-3450.

79. Walkup, G. K.; Imperiali, B., Design and evaluation of a peptidyl fluorescent chemosensor for divalent zinc. Journal of the American Chemical Society 1996, 118, 3053 - 3054.

80. Godwin, H. A.; Berg, J. M., A fluorescent zinc probe based on metal induced peptide folding. Journal of the American Chemical Society 1996, 118, 6514-6515.

81. Walkup, G. K.; Imperiali, B., Stereoselective Synthesis of Fluorescent -Amino Acids Containing Oxine (8-Hydroxyquinoline) and Their Peptide Incorporation in Chemosensors for Divalent Zinc. Journal of Organic Chemistry 1998, 63, (19), 6727-6731.

82. Shults, M. D.; Pearce, D. A,; Imperiali, B., Modular and tunable chemosensor scaffold for divalent zinc. J Am Chem Soc 2003, 125, (35), 10591-7.

83. Jotterand, N.; Pearce, D. A.; Imperiali, B., Asymmetric synthesis of a new 8-hydroxyquinoline-derived alpha-amino acid and its incorporation in a peptidylsensor for divalent zinc. J Org Chem 2001, 66, (9), 3224-8.

84. Thompson, R. B. In Fiber optic ion sensors based on phase fluorescence lifetime measurements, SPIE Conference on Advances in Fluorescence Sensing Technology, Los Angeles, CA, 1993; Lakowicz, J. R.; Thompson, R. B., Eds. SPIE: Los Angeles, CA, 1993; pp 290-299.

85. Fierke, C. A.; Thompson, R. B., Fluorescence-based biosensing of zinc using carbonic anhydrase. BioMetals 2001, 14, 205 - 222.

86. Thompson, R. B.; Cramer, M. L.; Bozym, R.; Fierke, C. A., Excitation ratiometric fluorescent biosensor for zinc ion at picomolar levels. Journal of Biomedical Optics 2002, 7, (4), 555 -560.

87. Thompson, R. B.; Maliwal, B. P.; Fierke, C. A., Selectivity and sensitivity of fluorescence lifetime-based metal ion biosensing using a carbonic anhydrase transducer. Analytical Biochemistry 199% 261, 185-195.

88. Thompson, R. B.; Maliwal, B. P.; Feliccia, V. L.; Fierke, C. A.; McCall, K., Determination of picomolar concentrations of metal ions using fluorescence anisotropy: biosensing with a "reagenfless" enzyme transducer. Analytical Chemistry 1998, 70, (22), 4717-4723.

89. Schwarze, S. R.; Ho, A.; Vocero-Akbani, A.; Dowdy, S. F., In vivo protein transduction: Delivery of a biologically active protein into the mouse. Science 1999, 285, 1569 - 1572.

90. Kiefer, L. L.; Patemo, S. A.; Fierke, C. A., Hydrogen bond network in the metal binding site of carbonic anhydrase enhances zinc affinity and catalytic efficiency. Journal of the American Chemical Society 1995, 117, 6831-6837.

LASER-BASED DETECTION OF ATMOSPHERIC HALOCARBONS

Jean-Franois Gravel and Denis Boudreau*

19.1. ABSTRACT

Among the various atmospheric trace species that play a role in environmental issues, volatile halogenated alcanes have attracted considerable interest because of their massive industrial use and of their documented effects on global warming and ozone depletion. Whereas the monitoring of halogenated alcanes at or near ground level can be accomplished routinely using a variety of established techniques, their determination at higher altitude poses considerable technological and/or logistic challenges. One class of detection techniques that has been experiencing accelerated development in recent years is the use of laser sources to remotely excite and/or fragment the target molecules, followed by their identification and quantification using their optical emission. The purpose of this review is to describe the current status and future trends in remote detection of atmospheric halocarbons by laser-based spectroscopic techniques, with a special emphasis on the use of luminescence techniques to detect the analytes of interest.

* Department of chemistry and Centre d'optique, photonique et laser (COPL), Universite Laval, Quebec (QC), Canada GIK 7P4; phone (418) 656-3287; fax (418) 656-7916; email: denis.boudreau @ chm.ulaval .ca

421

422 J.-F. GRAVEL AND D. BOUDREAU

19.2. INTRODUCTION

The key role of atmospheric trace species on many environmental issues such as global warming, ozone depletion and photochemical smog formation is now well recognized, and the knowledge of the fate and behavior of these species in the troposphere, as well as the ability to measure their concentration, lifetime and reactivity in different atmospheric layers are crucial in order to gain a better understanding of the complex mechanisms that lead to atmospheric or climatic changes and potential adverse health effects.'

Among these species, volatile halogenated alcanes are of particular interest. Owing to their enviable physicochemical properties, they were massively used over the last decades as refrigerants, propellants/blowing agents and flame retardants. Since the central role of chlorinated and brominated alcanes in the stratospheric ozone layer depletion phenomenon was recognized,^ many replacement products with similar properties have been proposed, first HCFCs (hydrogen-containing chlorofluorocarbons) and then - when these were also shown to have severe impacts on the stratospheric ozone balance^ and on human health'* ^ - hydrofluorocarbons (HFCs). The latter are very stable with respect to atmospheric degradation and have limited effects on the ozone layer depletion. However, because they tend to accumulate in the atmosphere^ and are relatively strong IR absorbers, they are now recognized as significant anthropogenic greenhouse gases and are included in a list of target compounds to control (Kyoto Protocol, 1997). Another class of halocarbons which are also cause for concern are the perfluorocarbons (PFCs), which are among the most potent greenhouse gases.^ ^ As a result, the development of monitoring methods for halogenated species has received increased attention in recent years. Their monitoring at or near ground level can be done routinely using a variety of sensitive and selective analysis techniques, often using on-site air sampling with cryogenic preconcentration, followed by gas chromatographic analysis. ^^ Despite the high detection sensitivities that these off-line techniques offer, they often involve hour-long sampling times and measurement cycles.

As highlighted by Sigrist,' spectroscopic techniques offer many significant advantages for on-line and in situ monitoring of trace atmospheric species; those based on laser excitation, in particular, offer the unique combination of real-time analysis with the high sensitivity and selectivity provided by the directionality, high irradiance and spectral purity of lasers. Moreover, the use of pulsed laser sources with synchronous signal detection allows the temporally and spatially-resolved monitoring of target atmospheric species from a remote location. A number of reviews on the application of laser-based techniques for remote atmospheric sensing have been pubhshed.'^

Halogenated alcanes have been monitored using many different spectroscopic approaches. For example, conventional open-path spectroscopic techniques such as Fourier transform infrared spectroscopy (FTIR), tunable diode laser absorption spectroscopy (TDLAS) as well as differential optical absorption spectroscopy (DOAS), allow for the in situ halocarbon determination

LASER-BASED DETECTION OF ATMOSPHERIC HALOCARBONS 423

by propagating a light beam through the space to be monitored. These techniques, based on the rovibrational absorption spectral signature of halogenated alcanes in the IR region from 7 to 12.5 |im,^ ^ ^ have provided detection limits reaching the ppb range and good analytical selectivity for a number of halocarbons (CF4, C2F6, C2H3CI, C2HCI3, C2CI4, C2H5CI, C2H4CI2, C4H5CI, CF2CI2, CFCla). ' ^ However, they usually require that the probe beam be propagated over long (sometimes km-long) absorption paths to reach the desired sensitivity, either using multiple reflection sampling cells or an open-path configuration which, in the latter case, limits the spatial resolution of the measurements.^^

Among the other spectroscopic approaches that have been applied to atmospheric monitoring and that could in principle be used for halocarbon detection, laser-induced fluorescence (LIP) generally offers in situ, on-line monitoring capability, high sensitivity, and spatially resolved measurements.^^ However, as pointed out by several authors,^ ^ " it is often the case for polyatomic molecules, such as halogenated alcanes, to have poor electronic transition probabilities together with broad, poorly defined spectral bands, which makes LIP unsuitable for ultrasensitive measurements.^^ This can be attributed to the existence of numerous closely spaced electronic energy levels, which aggravate signal losses through non-radiative decay processes, as well as fast intersystem crossing^^ and collisional quenching by major atmospheric constituents.^

Ongoing efforts towards the development of efficient laser-based in situ detection of halogenated alcanes therefore often involve breaking the molecule of interest into smaller, yet characteristic fragments, before probing the latter using various optical detection schemes. Among these, photofragmentation (PF) coupled to laser-induced fluorescence (LIP) has seen the widest application and perhaps the highest degree of success outside the research laboratory to perform real-time, spatially resolved, highly sensitive and selective monitoring of halogenated species in various environments.^^ Another technique, Laser induced breakdown spectroscopy (LIBS), while relying on atomic emission rather than fluorescence to detect the analyte of interest, has also shown a great potential for in situ measurements of halogenated alcanes and is included in the present review. Finally, the unique propagation characteristics of intense femtosecond laser pulses in air make the latter a very promising fragmentation and excitation source for the spatially-resolved, sensitive and selective monitoring of halogenated species and other pollutants ranging from ground-level, in situ measurements to remote sensing of the upper atmosphere.

19.3. LASER INDUCED BREAKDOWN SPECTROSCOPY

Laser induced breakdown spectroscopy (LIBS, also known as laser-induced plasma spectroscopy, LIPS) is a versatile analytical method that is applicable to almost any type of sample matrix including solids, liquids and


gases. " ^ It relies on the generation of a hot plasma obtained by tightly focusing laser pulses (typically Nd:YAG laser light in the NIR) to reach high intensities (typically in the GW/cm^ range). The time-resolved emission spectrum of the plasma plume is then measured with a gated detection system to provide fast, quantitative multielemental analysis. While most of the applications of LIBS are related to the analysis of solid samples, ^ it is also considered as an analytically useful gas sensing technique. This is particularly true for the analysis of gaseous halogenated alcanes bearing chlorine and fluorine atoms, owing to the high plasma temperatures (in the 20,000-25,000 K range^^) which are high enough for efficient collision-induced population of high-lying excited atomic states of chlorine and fluorine (e.g. the first radiatively-coupled excited state of fluorine lies -13 eV above the ground state).

Over 20 years ago, Cremers et al. ^ reported the first use of LIBS for the time-resolved detection of chlorine- and fluorine-containing halogenated alcanes in air at atmospheric pressure. They studied different parent molecules (i.e. (CH3)2CHC1, CCI2F2, C2CI3F3, CCI4) and showed that the LIBS technique was insensitive to the molecular structure of the analytes, as long as the plasma was not physically or chemically perturbed by the samples. They also reported that the emission spectrum was mainly composed of the atomic lines of major air constituents, namely oxygen and nitrogen, and that the most intense emission lines for CI (I) and F (I) ( X. = 837.6 nm and 685.6 nm, respectively) were free from interferences with these major constituents. As is generally the case for LIBS experiments, satisfactory analytical performance was reached following the optimization of experimental parameters such as laser pulse energy, detection gate delay and integration time relative to the onset of plasma ignition. This is due to the fact that the emission spectrum is first dominated by an intense, short-lived background continuum, followed by the more persistent atomic lines. Therefore, by using time-resolved detection, the signal-to-noise ratio (SNR) can be maximized. Interestingly, Cremers et al. noted that optimal gate delays and integration times strongly depended on the species to be detected.^ ' ^

Most of the work done with different halogenated alcane showed that the CI and F emission intensities were directly proportional to the total number of each atom contained in the parent molecule, an indication of complete dissociation of different species in the plasma.^' ^ Measurement of line intensity ratio for different elements allowed for the discrimination and prediction of the molecular structure of different halogenated alcanes. As an example, Dudragne et al. ^ measured the relative CI, F and C intensity ratios to distinguish between a series of six different HCFCs. Haisch et al." used the same approach with CI and H Hnes to discriminate between CCI4, CHCI3 and CH2CI2, while Williamson et al. '* used F and H lines to distinguish between a series of fluorocarbons and hydrofluorocarbons. However, as pointed out by Cremers et al.,^' this procedure is not applicable to evaluate the molecular concentration or stoechiometry for mixtures of different parent molecules. Therefore, LIBS can only be considered


as a selective technique to distinguish classes of compounds. Typical limits of detection (LOD) for the analysis of halogenated alcanes by LIBS are in the low ppmw range. For example, reported LODs for CI or Cl-containing compounds ranged from 2 to 90 ppmw for CCl CF2CI2 and CF4 parent molecules. ^ "

It is interesting to note that, for all references that provide information on the time required for halocarbon determination in air using LIBS, the analysis time per sample was on the order of a minute or less. Williamson et al. " even showed the possibility to obtain acceptable data from a single laser pulse. Haisch et al. ^ also showed the possibility to develop a miniaturized system based on a compact Nd:YAG laser (dimensions of 13 x 5 x 3 cm). The rapid data acquisition time combined with the possibility to develop compact, field deployable instrumentation for the in situ determination of halogenated alcanes led to the use of LIBS as a diagnostic tool in various applications such as the on-line monitoring of chlorine emissions from waste management facilities (such as incinerators, landfill sites or semiconductor etching processes^^), the sensitive detection of chlorine and fluorine for the monitoring of chemical weapons such as nerve and blister agents^^ or even the real-time and in situ detection of fluorine concentration for the evaluation of different replacement products for halons in severely hostile fire suppression environments such as military aircraft and ground combat vehicle engine compartments. "^

19.4. LASER PHOTOFRAGMENTATION - FRAGMENT DETECTION

The laser photofragmentation - fragment detection approach (hereafter called PF/FD) shares many characteristics with LIBS in terms of in situ detection capability combined with high temporal and spatial resolution. The most important features of PF-luminescence approaches include their applicability to non fluorescent species as well as the ability to target a class of compound rather than a single molecular species by monitoring commonly shared fragmented functional groups. Compared to LIBS, FD is generally more sensitive (ppm to low ppb detection limits) and can be made more selective (or species-specific), especially when LIF or photoionization detection is considered. It is outside the scope of the present paper to review and compare all the various detection approaches that can be implemented to monitor the photofragmentation products. Detailed and comprehensive reviews have been pubUshed by Simeonsson et al. ' " on this topic. Therefore, only the basics of PF/FD will be highlighted herein.

PF/FD detection schemes can be divided in two distinct processes, as illustrated in Figure 1. The analyte species R-F is first excited using one laser source, hv^ through linear or multiphoton absorption, which leads to the fragmentation of the parent molecule into the characteristic fragment F (and its counterpart R), which can be atomic or molecular species of typically up to 2-3 atoms. As depicted on the energy level diagram of Figure 1, the fragments of interest will be characterized by their internal energy distribution E(j,v,e),


denoting the rotational (]), vibrational (v), and electronic excitation (e) contributions. Of course, the photofragments themselves contribute to the selectivity of the method since they are characteristic of the chemical composition of the parent molecule. However, unlike the LIBS approach, the first fragmentation step can contribute to enhance the selectivity of the detection technique since molecular absorption occurs through a transition characteristic of the targeted R-F species. Since high laser intensities are generally required for the photofragmentation step, it should be kept in mind that potential interferences from other species must be considered and the selection of an appropriate detection scheme is of crucial importance to obtain both selective and sensitive measurements.

After fragmentation, a second laser source, hv2, can be used to probe the fragments. As shown on the energy level diagram of Figure 1, fragment detection can be accomplished by various means including the luminescence methods of laser-induced fluorescence (LIP), prompt emission (PE, sometimes termed fragment fluorescence), stimulated emission (SE) or laser ionization recombination emission (LIRE). Fragment detection via photoionization approaches (photoionization, PI; prompt photoionization, PPI) coupled to mass spectrometry is also possible and has been extensively studied but will not be covered in the current review. As stated by Simeonsson et al., among the most common optical detection schemes, PF-LIF has seen the widest application and perhaps the highest degree of success outside the research laboratory to perform real-time, spatially resolved, highly sensitive and selective monitoring of halogenated species in various environments.^^ "

Analytical applications as well as more fundamental studies of the use of PF-LIF, PF-PE and PF-SE can be found in the literature for the measurement of F-, C1-, Br- and I- containing halogenated alcanes and involve the detection of halogen-containing molecular fragments as well as halogen atoms. For the purpose of the present review article, we have selected the papers that best highlight the different detection strategies that can be implemented and the fields that have been covered the most in terms of applications.

Chlorinated hydrocarbons (CHCs) have certainly been the most targeted species due to their massive use in many industrial and domestic products combined with their negative impact to the environment. An attractive field of application for PF-based applications was the monitoring of incinerator performance during CHC-containing waste disposal, which must be evaluated in situ and in real time to ensure complete CHC destruction and avoid the release of toxic by-products such as tetrachloroethylene (TCE). "^^^^


I. Vhotofragmentation

R-F - ^ R + F

2. Fragment Detection rnrv-f r »it <? '

PPI

(vj.ej)

It t (F^+e-;

Figure 19.1. Schematic diagram illustrating several photofragmentation / fragment detection (PF/FD) processes. Reprinted with permission from Simeonsson et al,^"^ Trends in Analytical Chemistry (1998) 17, 542-550. Copyright (1998) Elsevier.

Pioneering research on the photochemistry of chlorine-containing halogenated alcanes has been conducted by several groups. Tiee at al. ^ irradiated CCI4 and CCI3F with an ArF excimer laser emitting at 193 nm and probed the fragment distribution with a dye laser to obtain the fluorescence spectrum of the CCl fragment {A^A-> X^II) around 280 nm. They also identified several other fragments via their spontaneous emission spectra in the 190 - 600 nm region and found that CCl and CI2 were commonly generated from both parent molecules and that the emission of those fragments dominated the recorded spectra. They also showed that the laser photolysis of CCI4 and CCI3F occurs via multiphoton absorption of up to 3 UV photons. A few years later, Kenner et al. ^ published a more detailed study of the different mechanisms involved in the photodissociation and subsequent fragment


emission for COCI2 and several chlorinated methane species in the UV and VUV regions. They also reported the CI2 fragment as one of the major photofragmentation products and showed that HCl can be one of the main photofragmentation products for hydrogen-containing chlorinated methanes such as CHCI3 and CH2CI2.

The analytical potential of the PF-luminescence approach for the sensitive and selective detection of non or poorly fluorescing molecules has been highlighted and studied for different species. ^ ^ Among common luminescent fragments, Whery et al. identified CCl as a potential probe for class-specific detection/^ stressing the need to understand and control the fragmentation process in order to relate the intensity of fragment luminescence to the concentration of the parent compound in the original sample. They also recalled theoretical calculations by Rodgers et al. ^ and Halpern et al." ^ that suggested the potential use of such an approach (more precisely PF-LIF) for ground-based or aircraft-based atmospheric remote sensing. Interestingly, Rodgers et al." ^ identified CH3Br and CH3I as candidates for detection.

Jeffries et al. ^ evaluated the figures of merit of the PF-LIF approach to address the applicability of the technique to monitor a series of CHCs (CH3CI, CH2CI2, CHCI3, CC1„ C2HCI3, C2CI4) in incinerator effluents. Their strategy involved photofragmentation with one laser at 193 nm followed by 2-photon excitation of ground state atomic chlorine through a spin-forbidden transition using a second, time-delayed laser emitting near 233 nm. Subsequent fluorescence detection in the 700-800 nm region was accomplished. As noted by the authors, all common chlorinated alcanes are photodissociated when irradiated by 193 nm laser light and produce ground state CI atoms,"* enabling the possibility to monitor the total content of Cl-containing compounds in the probed environment. Their study showed that the 700-800 nm spectral region was free from interference from fragment emission (including excited-state CI and background gases such as Ar, N^ and O2), especially when the 2-photon-excited fluorescence was delayed by a few microseconds after the 193 nm laser pulse. The sensitivity of the technique was evaluated for CH3CI in air at 475 torr and the authors obtained a detection limit of 20 ppb. However, they noted that CI2 and HCl were also detected by the LIF probe following 193 nm photofragmentation with almost equal sensitivity. Since these two species are desired end-products of CHC incineration and will be found in concentrations orders of magnitude higher than the target concentrations for CHC release in the effluents, the applicability of their approach was limited.

With respect to this last issue, Lucas et al. ^ reported on the use of PF-LIF to monitor CHCs in incinerator effluents, this time using the CCl fragment which is not naturally occurring in incinerator exhaust. As mentioned previously, it had already been observed that different CHC molecules yielded the CCl fragment which thus offered the possibility to monitor the total content of Cl-containing compounds in the sampling environment. In a first publication, the authors answered several practical and fundamental questions that were relevant to ensure the applicability of such an approach to hostile environments.


Notably, they investigated the effect of elevated temperatures (1000 K) at atmospheric pressure together with the presence of quenching species such as CO2 and H2O on the fragmentation yield, and found that C2H5CI was efficiently fragmented. They also optimized the laser-induced fluorescence detection scheme, involving excitation at 271.4 nm and fluorescence detection at 278 nm (A A ^^ X^ri). This was critical since molecular oxygen, which absorbs 193 nm laser light through the Schumann-Runge band system, was shown to emit around 300 nm. A time-resolved study of oxygen emission showed that the excited state of O2 was short-lived (picosecond regime) compared to the optimal delay for fluorescence excitation of the CCl fragment at 271.4 nm (microsecond regime), thus enabling the effective measurement of trace amounts of C2H5CI in the presence of atmospheric level concentrations of O2. The authors finally showed that CCl fluorescence could not be detected following the irradiation of non CHC-containing mixtures of C- and CI- containing gases such as HCl and C2H4, thus demonstrating the usefulness of their approach in terms of selectivity towards the targeted species. In a second publication,^^ the authors extended their work to several other chloroethanes and chloroethylenes, namely CH3CI, C2H4CI2, C2H3CI3, C/5-C2H2CI2, C2HCI3 and C2CI4. They noted that all these species were detectable via the CCl-based PF-LIF approach, the CCl yield being equivalent or superior to that of C2H5CI, thus offering what they called broadband sensitivity to CHCs. They also ruled out CI2 as a potential interfering species, thus enabling the discrimination in the effluents of CI originating from unburned CHCs vs. CI generated from the desired incineration end products. The detection limit for CCl generated from C2H5CI in air at atmospheric pressure was 5 ppb and the time required for the measurement was 5 seconds per sample. As stated by the authors, this analytical performance matched the required sensitivity for emission monitoring (<10 ppb), especially when considering that the species under study would yield more CCl and thus exhibit better sensitivities.

Because of the comparable effect on stratospheric ozone depletion of brominated vs. chlorinated alcanes, PF-based detection strategies for brominated alcanes were also studied, although to a lesser extent. Arepalli et al.'* studied the photofragmentation of Br2, HBr, CHsBr and C2H5Br, identifying the possibility to use a single laser around 260 nm both for the photofragmentation of the bromine compounds and to induce resonant 2-photon excitation of atomic bromine through a spin-forbidden transition. Simeonsson et ^j 12,44,45 dernonstrated the analytical potential of this approach for the detection of trace-level brominated compounds, using LIF as well as a novel stimulated emission (SE) approach to detect the Br fragments in the gas phase under ambient conditions. PF-LIF measurements resulted in better signal-to-noise (SNR) ratios than PF-SE measurements, with detection limits for CHBra and CHClBri in the ppb and ppm range, respectively. Despite the lower sensitivity obtained with PF-SE, the authors concluded to its great potential for rapid and convenient measurements of ppm level (and possibly lower) detection of brominated compounds in gaseous environments, emphasizing the fact that the


coherent and directional nature of SE could provide better access to remote or hostile environments, while the sensitivity of the SE technique could be enhanced by using longer focal length lenses to increase the gain length.

As noted by Simeonson et al., fewer examples can be found of the analytical applications of PF-luminescence techniques for the monitoring of iodine- and fluorine-containing alcanes.^^ Brewer et al." ^ studied the energy distribution of the photofragmented iodine atoms originating from a series of alkyl iodides of up to 4 carbons using 2-photon-excited fluorescence in the UV. The authors reported a detection capability of around 10 ^ atoms per cm^, corresponding roughly to 10 ppb according to their experimental parameters.^^ To the best of our knowledge, the detection of fluorine-containing alcanes using a combination of photofragmentation and LIP detection of atomic fluorine was not reported. This can be mainly attributed, as stated previously, to the considerable photon energy needed to reach the high-lying excited states of the fluorine atom, equivalent to the use of 2 VUV photons for the excitation step. ^ " Analytical applications would therefore be restricted to the use of evacuated or purged beam propagation paths.

Many studies of the photofragmentation of fluorine-containing alcanes have shown that fluorine-containing molecular fragments can be efficiently generated and detected via luminescence detection techniques. Several authors " " underlined the considerable interest towards the difluorocarbene radical (CF2) owing to its remarkable stability^^ (resulting from the relatively high strength of the C-F bond, typically 485 kJ/mol as compared to -350 kJ/mol for the C-C bond) and to its occurrence in the atmosphere due to the stratospheric photolysis of PFCs, CFCs and HCFCs by sunlight." ^ Aside from being an environmentally relevant species, CF2 has also attracted much interest due to its use in many fluorocarbon reactions as well as in ion plasma etching of silicon. ^^* Therefore, the interest for more knowledge about the molecular properties of CF2 species has motivated theoretical work involving the simulation of laser-induced fluorescence experiments carried out with the CF2 fragment^ ' ^ together with spectroscopic and kinetic studies to better understand the photochemistry and photophysics of CF2-containing halocarbons.^^'^^'^^

For example, King et al. " ' ^ reported on LIF measurements involving the A^Bj -> X^Aj transition around 300 nm to probe the energy distribution of the CF2 fragments produced at low pressure by IR laser photofragmentation of parent species such as CF2CI2. Interestingly, the authors suggested that LIF-based analytical applications involving the determination of the CF2 fragment would appear shortly. Rossberg et al. ^ used LIF to study the chemical processes responsible for the formation of CF2 and CFCl fragments following the IR photofragmentation of CF2CFCI at low pressure. The authors also reported the detection of highly excited CF which was generated with the LIF excitation laser. The potential precursor of this photolytically generated secondary product was shown to be CFCl, a primary product of the photodissociation of the parent molecule. Winkelmann et al. ^ and Cameron et al." used LIF to refine the


identification of the rotational and vibrational structure of the A^B, -> X^Aj electronic transition of the CF2 fragment. While Winkelmann et al. used 193 nm laser light to induce the photofragmentation of CF2CI2, Cameron et al." formed the desired fragment by pyrolysis of C2F4 at 1000°C followed by LIF detection in a supersonic jet expansion. Their work resulted in the unique assignment of several new series of band transitions as well as the definitive assignment of the two stretching frequencies in the 'B^ excited state.

19.5. LONG RANGE REMOTE SENSING OF HALOCARBONS BY NON- LINEAR LASER PROPAGATION

The work that has been reviewed so far involved the spectrometric detection of halogenated alcanes near or at ground level. When remote sensing is considered, whether it implies monitoring at higher altitude or at ground level over km-long distances, the task becomes somewhat more challenging. Measurement strategies usually involve either bringing the sample to the laboratory^ or bringing the instrument to the sample using balloon-, plane- or satellite-borne instruments based on PF ^^ or IR absorption^^' ^^'^^ techniques. While the latter strategy provides an unmatched richness of information, i.e., composition of the atmosphere with high sensitivity and dynamic range at a precise location in space and time, it cannot be easily implemented for routine or surveillance analysis. On the other hand, ground-based open-path techniques mentioned previously, namely FTIR, DOAS and TDLAS, because they only provide path-averaged concentration values over line-of-sight measurements, cannot be used to obtain altitude- or range-resolved halocarbon concentrations. Recently, the availability of compact and reliable high power, nanosecond pulsed Nd-YAG lasers has led to the demonstration of LIBS for medium range (up to 45 meters) remote sensing of solid samples (solid or molten steel, solid environmental samples). ^^ ' ^ However, this approach has not yet been reported for atmospheric species, presumably because typical breakdown energy thresholds are significantly higher in the gas phase than for solid samples. ^ Obviously, the use of LIBS and PF/FD for long range remote sensing is hindered by the difficulty to deliver the relatively high laser intensities required over long distances.

Certainly one of the most effective and useful tools for atmospheric remote sensing is the light detection and ranging technique, or LIDAR. Based on the time-resolved detection of atmospheric backscattered light generated by laser pulses propagating in the atmosphere, this approach enables the accurate determination of 3-D distributions of atmospheric trace gases and aerosols over large areas.*^ ^ Differential optical absorption spectroscopy (DOAS) has been used successfully in a LIDAR configuration (usually called DIAL for differential absorption LIDAR^ ' ^ ) to measure atmospheric trace species in the gas phase over moderate to long ranges with high sensitivity (typical range for SO2, NOx, O3 and Hg is 0.5-5 km with ppb detection limits^^). In that


configuration, one wavelength is set to an absorption peak of the target molecule while a second wavelength is set off-peak, and the difference measured in the backscattered signals is used to monitor the concentration of the target species along the propagation path. With LIDAR techniques, the ability to detect more than one pollutant - or even a single one, in the case of DIAL - generally relies on the use of several probing lasers, or on fast wavelength switching of a wide spectral range tunable laser such as liquid dye lasers and optical parametric oscillators. However, to the best of our knowledge, the use of LIDAR in an absorption spectroscopy mode for the monitoring of halogenated alcanes has not been reported, presumably due to the poor electronic transition probabilities of the latter.

Recent developments in the field of nonlinear propagation of high power, ultra-short laser pulses (femtosecond, multiterawatt) offer new possibilities for remote sensing of atmospheric species. The dynamic interplay between two nonlinear effects, i.e. Kerr self-focusing of the propagating laser pulses in the neutral gas and subsequent defocusing by the plasma that is generated at the self-foci through multiphoton/tunnel ionization of the molecules and atoms,^ ' ^ leads to the production of a series of low electron density plasma columns referred to as filaments.^^' ^ Furthermore, the nonlinear propagation of the intense femtosecond laser pulses in air induces a spectral broadening of the initial pulse through the physical process of supercontinuum generation (SCG), often termed white-light laser generation.^^' '^^ For a pump wavelength centered at 800 nm, this supercontinuum has been shown to extend from 300 nm down to 4.5 |im.^^ The backward scattering of the white-light laser pulse combined with linear absorption spectroscopy has been used for the remote sensing of major atmospheric constituents. For example, Kasparian et al. ^ described a novel approach based on a mobile fs-multiterawatt laser platform (called Teramobile) for the remote sensing of atmospheric species.

The Teramobile laser system is depicted in Figure 19.2. The Ti:sapphire laser provides 350 mJ pulses with a 70-fs duration, resulting in a peak power of 5 TW at a wavelength of 800 nm and a 10 Hz pulse repetition rate. An important feature of their instrument lies in the use of pulse chirping to precompensate for group velocity dispersion (GVD) in air. Briefly, GVD affects the propagatio of spectrally broad ultrashort pulses (ca. 20 nm in the present case) in a dispersive media such as the atmosphere, with the result that the spectral components of longer wavelength (so-called "red" components), which propagate faster than those of shorter wavelength ("blue" components), will temporally stretch the laser pulse during its propagation - which, ultimately, limits the ability to deliver high laser intensities at remote distances. However, by predispersing the spectral content of the laser pulse (or pulse chirping), so that blue components precede red components at the output of the fs-laser chain (i.e. negatively chirped pulses), one can produce laser pulses that shorten temporally while propagating in the atmosphere until its intensity reaches the conditions for filamentation and white-light generation, this feature has allowed Kasparian et al. to control the onset of filamentation and its length over


distances of a few kilometers. The authors have demonstrated the analytical potential of he backward scattering of white-light laser pulses combined with linear absorption spectroscopy by measuring major atmospheric constituents such as oxygen, water vapour and tropospheric ozone/ ' ^ In effect, they have demonstrated the possibility to perform atmospheric multicomponent analysis at remote distances with a single laser source.^ ' ^

While the technique described above looks promising, it implies, for species having low absorption cross-section (as is the case for halocarbons), the integration of absorption along an extended path, which affects the ability to make range-resolved measurements with high spatial resolution. Furthermore, LIDAR applications using the NIR and IR side of the supercontinuum - where halocarbons have their strongest spectral absorption features - remain a challenging task, because the intensity of the white light supercontinuum drops exponentially in intensity as a function of wavelength, i.e. by more than 4 orders of magnitude from the pump wavelength (800 nm) up to 2.5 jam, and by another order of magnitude from 2.5 ^m to 4.5 \xmJ^ This is of crucial importance since both sensitivity and selectivity are related to the spectral

"^j •vp^';pyi"fM:y

Chirp control Telescope

Figure 19.2. Schematic of experimental setup used in white-light LIDAR. Before launch into the atmosphere, the pulse is given a chirp, which counteracts group velocity dispersion (GVD) during its propagation in air. Hence, the pulse recombines temporally at a predetermined altitude, where white-light continuum is produced, and then is backscattered and detected by LIDAR. Reprinted with permission from Kasparian et al./^ Science (2003) 301, 61-64. Copyright (2003) AAAS.


brightness of the sourceJ^ Recently, the detection of LIDAR signals in the NIR region up to 1.7 j^m using the supercontinuum emitted from light filaments propagating in the atmosphere was demonstrated,^^ although the application of this technique to the actual detection of trace molecules in this spectral region has not yet been reported.

The use of the nonlinear filamentation process has also been applied to the design of long range LIBS detection systems, again highlighting the possibility to deliver high laser intensities over long distances. Rohwetter^^ and Stelmaszczyk^^ performed remote LIBS experiments, using the Teramobile system to launch the unfocussed laser beam towards metallic copper, steel and aluminum targets located at distances up to 90 meters. By comparing the spectra obtained using picosecond and femtosecond laser pulses, they found that femtosecond pulses generated cleaner spectra that were free of spectral interferences from the ambient gas (i.e. atomic oxygen and nitrogen line emission). Moreover, at a distance of 90 meters, they observed that only femtosecond pulses were able to generate a plasma over the target surface. However, perhaps their most interesting observation was the dependence of analyte emission intensity on pulse chirping. The authors observed that the use of minimal duration, Fourier transform-limited pulses at the target surface did not yield maximum analyte emission intensity. Rather, they found that a positively chirped pulse of around 600 fs in duration resulted in an increased LIBS signal, while a negatively chirped pulse of equal duration gave a lower signal enhancement. Several groups have reported on the ability of pulse chirping, or more generally pulse shaping, to yield optimal multiphoton excitation or ionization in atomic or molecular systems^^' ^^ or improved laser ablation performance.^^ According to Rohwetter^^ a better understanding of pulse shaping could improve the range and sensitivity of femtosecond-induced LIBS, while enhancing the detection specificity in cases where species to be identified are spectrally overlapped.

The use of filamentation to detect gaseous atmospheric species via multiphoton-ionization-induced fragmentation followed by spectral detection of the characteristic fragments has also been investigated. It has been shown that the peak intensity inside a filament in air is limited to ca. 5x10^^ W/cm^, due to the competition between self-focusing and defocusing, ^"^^ while Teramobile experiments'^^' ^ ' ^ described in the previous paragraphs have shown that such an intensity can be reached at a remote distance from the source. At these high intensities, most molecules will undergo multiphoton/tunnel ionization followed by fragmentation. Many of these fragments, produced in the excited state, will promptly relax through radiative pathways, thus emitting a characteristic luminescence spectrum. An example of this is the so-called "clean fluorescence" of nitrogen molecules and ions has been observed from filaments generated in atmospheric air. ' ^ Due to the low plasma electron density inside the filament (n « 10*"*-10* cm^ vs. lO' -lO^^ cm^ for nanosecond laser-induced plasmas^^), continuum emission from the plasma is quite low and does not mask characteristic spectral features, as is usually the case for the optical breakdown


induced by longer pulses. ^ ^ Recent experiments by Luo et al. have shown that clean nitrogen luminescence, free from white-light supercontinuum, can be detected in the backward direction, ^ and that the backscattered signal can be amplified through stimulated emission.^^ These results make femtosecond-induced luminescence a promising approach for remote sensing of atmospheric constituents.

Gravel et al. ^ recently reported on the use of ultrafast and intense laser pulses in air ^ for the in situ determination of halogenated alcanes in the atmosphere. The authors demonstrated that clean and characteristic fragment luminescence can be measured inside a filament for chlorine- and fluorine-containing halocarbons. As shown in Figure 19.3A, when filamentation was induced in a gas cell filled with air (no halocarbons) at atmospheric pressure, the characteristic spectrum of N / (first negative band, B^Z^ —> X^Z^) and N2 (second positive band, Cfn^ -> B'H) ' ' ' ' could be observed in the 300-400 nm region. It had been shown previously that nitrogen molecules are ionized and excited inside the filament by the ejection of an inner valence electron and that radiative relaxation from these excited states gives rise to nitrogen ion luminescence, "* ^ while electron-ion recombination then leads to luminescence in the second positive band of the neutral N2 molecule.

When filamentation was induced in pure CF3CHCIF (i.e. HCFC-124) at atmospheric pressure, several spectral bands were observed in the UV-VIS (Figure 19.3B). Notably, the authors observed the A^Bj -^ XÂj band from the excited difluorocarbene radical (CF2) as well as the AÂ —> XÎT Sind B^Z -^XÛ bands from the CH radical (around 430 nm and 389 nm, respectively), together with the band of the CFCl fragment in the 350-450 nm region. The authors noted, in particular, the absence of atomic emission lines (i.e. N, O for air and C, CI, F, H for HCFC-124) or continuum emission in the spectral region studied. These observations, consistent with previous work by Talebpour,^^' ^ Liu ^ and Rohwetter^^ illustrated the marked, fundamental difference between filamentation and optical breakdown processes. These results also contrasted with results typically reported in the monitoring of halogenated alcanes using LIBS, where continuum emission followed by atomic emission dominates the spectrum.^' ^


Wavelength (nm)

Figure 19.3. Spectra obtained for femtosecond-laser induced fluorescence: A) air; B) HCFC-124. Laser pulse energy = 5 mJ, detector gate delay set to 0 ns with respect to the pump laser, 25 ns gate width, 60 laser shots accumulated. Note the different intensity scales used for each spectrum. Reprinted with permission from Gravel et al.^^ Analytical Chemistry (2004) 76, 4799-4805. Copyright (2004) American Chemical Society.

The authors reported CF2 emission from two other halogenated alcanes, namely CF^ and C2F , and suggested that the detection of this characteristic spectral signature by femtosecond-induced filamentation could be a promising avenue for the remote sensing of such species in the atmosphere. However, in air and at low concentrations, they noted that the strong nitrogen emission was interfering with CF2 emission around 300 nm, preventing the detection of the targeted species in air at atmospheric pressure. As shown in Figure 19.4, time-resolved spectral analysis of the filament emission allowed the observation of significantly different radiative lifetimes for the CF2 and N2/N2^ excited species and the possibility to increase detection contrast using adequate time gating. The authors were thus able to measure clean CF2 spectra, free from N^fN^^ emission bands, from halocarbons HCFC-124, CF^ and C2F diluted in air. Analytical figures of merit were evaluated for C2F in air at atmospheric pressure, resulting in an instrumental detection limit of 8 ppmv, and the authors suggested that other molecular fragments having fluorescence intensities comparable to that of CF2 together with similar or longer radiative lifetimes - or emission spectral ranges differing from those N2/N2^ - could yield similar detection limits.


13

-5 ns T = 0 ns

u

•• 5 n s

g n t = 10ns —» C

(D

0 O (/) 0.00-

0) o Z3 0025-

Li-

300 320 340 360 280 300 320 340

T = 15 ns

T = 50 ns

280 300 320 340 360

T = 20 ns

340 360 280

T = 75 ns

340 360

0 025

300 320 340 360

.Amjw^z

T = 110 ns

280 300 320 340 280 300 320 340 360

Wavelength (nm)

Figure 19.4. Time-resolved fluorescence spectra obtained for a mixture of 25% C2F6 in air. Laser pulse energy = 5 mJ. For each delay with respect to the pump pulse (x), 200 shots were averaged. ICCD gate width = 5 ns for all measurements. Note the different intensity scales used for each row of spectra. Reprinted with permission from Gravel et al.^^ Analytical Chemistry (2004) 76, 4799-4805. Copyright (2004) American Chemical Society.

Finally, the authors observed that the CF^ spectrum obtained from CF^ showed a weaker fluorescence yield than that obtained from C^¥^ or HCFC-124, in the experimental conditions used for their study. This observation emphasized the role of molecular structure on the extent of the ionization and fragmentation mechanisms leading to the excited CF^ fragment, while suggesting that the application of this technique might be limited to class-specific detection, the molecular concentration or stoichiometry of mixtures of distinct parent molecules being impossible to retrieve from CF2 intensity measurements. However, it was pointed out that the measurement of CH and CFCl spectral bands could be used to obtain a qualitative estimate of the sample composition.


19.6. FUTURE DIRECTIONS

Considering the latest achievements and demonstrations of the capabiHties of nonlinear LIDAR approaches, it is reasonably fair to suggest that signal enhancement strategies, in particular temporal background signal discrimination, could be implemented for the long range detection of halocarbons by femtosecond laser-induced luminescence in a LIDAR configuration. For example, the length and distance of the filament could be determined from the short-lived N2 fluorescence signaf^ and, using this information to set the time-gating delay, one could temporally separate analyte fluorescence signals from those of background species. Notably, it has been recently reported that the strongest part of the N2 emission inside a filament generated in mid-range propagation experiments in air (over 100 m) comes from the earliest self-focusing region. ^ With adequately chosen time-gating parameters and laser pulse parameters carefully controlled to induce the formation of a strong, single filament, ^^^ background species could be efficiently distinguishable from analyte fluorescence. Another promising approach involves the use of adaptively-shaped ultrafast laser pulses, as recently suggested by several authors (Mejean et al. ^ ^^\ Rohwhetter^^), by dynamically changing the relative phase and intensity envelope of the femtosecond pulses to enable the selective control of physical processes such as ionization, fragmentation and excitation.^^^ ^^ '" Such an approach could result in improved analytical performances in terms of sensitivity and selectivity, thus enhancing the analyte to background fluorescence ratio while allowing to distinguish between molecules such as CF4, C^F^, and CF3CHCIF. Therefore, and more generally speaking, fragment fluorescence in the UV-VIS could become a complementary technique to the white light approach based on absorption, especially for molecules that mostly absorb in the mid-IR, where major atmospheric constituents such as water vapor or carbon dioxide limit atmospheric transmission and degrade detection sensitivity as well as selectivity.


The authors acknowledge the Natural Sciences and Engineering Research Council of Canada (NSERC), and the Fonds Quebecois de Recherche sur la Nature et les Technologies (FQRNT) for their financial support.

19.8. REFERENCES

1. M. W. Sigrist, in: Air monitoring by spectroscopic techniques, edited by J. D. Winefordner (Wiley, New York, 1994), pp. 1-26.

2. M. J. Molina, F. S. Rowland, Stratospheric sink for chlorofluoromethanes - Chlorine atom-catalyzed destruction of ozone, Nature 249(5460), 810-812 (1974).


3. K. Stemmler, S. O'Doherty, B. Buchmann, S. Reimann, Emissions of the refrigerants HFC-134a, HCFC-22, and CFC-12 from road traffic: Results from a tunnel study (Gubrist Tunnel, Switzerland), Environmental Science and Technology 38(7), 1998-2004 (2004).

4. W. Dekant, Toxicology of chlorofluorocarbon replacements, Environmental Health Perspectives Supplements 104(Suppl. 1), 75-83 (1996).

5. P. Hoet, M. L. M. Graf, M. Bourdi, L. R. Pohl, P. H. Duray, W. Chen, R. M. Peter, S. D. Nelson, N. Verlinden, D. Lison, Epidemic of liver disease caused by hydrochlorofluorocarbons used as ozone-sparing substitutes of chlorofluorocarbons, Lancet 350(9077), 556-559 (1997).

6. E. J. Dolin, PFC emissions reductions: The domestic and international perspective, Light Metal ^ge 57(1,2), 56-67(1999).

7. G. Rhoderick, P. Chu, E. Dolin, J. Marks, T. Howard, M. Lytle, L. McKenzie, D. Altman, Development of perfluorocarbon (PFC) primary standards for monitoring of emissions from aluminum production, Fresenius' Journal of Analytical Chemistry 370(7), 828-833 (2001).

8. G. Mangani, A. Berloni, M. Maione, Cold solid-phase microextraction method for the determination of volatile halocarbons present in the atmosphere at ultra-trace levels. Journal of Chromatography, A 988(2), 167-175 (2003).

9. R. A. Rasmussen, Atmospheric halocarbon monitoring techniques. World Meteorological Organization, 460(Air PoUut. Meas. Tech., Part-II), 1977, pp. 199-207.

10. J. Koch, M. Okruss, J. Franzke, S. V. Florek, K. Niemax, H. Becker-Ross, Element-selective detection of gas chromatographic eluates by near infrared echelle optical emission spectrometry on microwave-induced plasmas, Spectrochimica Acta, Part B: Atomic Spectroscopy 59(2), 199-207 (2004).

11. M. M. Abdillahi, Gas chromatography-microwave-induced plasma for the analysis of halogenated hydrocarbons. Journal of Chromatographic Science 28(12), 613-616 (1990),

12. J. B. Simeonsson, R. C. Sausa, Trace detection of ambient brominated compounds by laser-induced photofragmentation/fragment detection spectrometry, Spectrochimica Acta, Part B: Atomic Spectroscopy 49B(\2-14), 1545-1555 (1994).

13. S. Toyoda, T. Tominaga, Y. Makide, Cryogen-free automated gas chromatograph system for monitoring of halocarbons in the atmosphere at background concentration levels. Analytical Sciences 14(5), 917-923 (1998).

14. U. Panne, Laser remote sensing, TrAC, Trends in Analytical Chemistry 17(8-9), 491-500 (1998).

15. P. M. Chu, Advances in optical methods for trace gas analysis. Analytical and Bioanalytical Chemistry 376(3), 305-307 (2003).

16. P. L. Hanst, S. T. Hanst, in: Air monitoring by spectroscopic techniques, edited by J. D. Winefordner (Wiley, New York, 1994), pp. 335-470.

17. H. A. Gamble, G. I. Mackay, D. R. Karecki, J. T. Pisano, H. I. Schiff, Development of a TDLAS based methodology for monitoring perfluorocarbon production during the aluminium smelting process. Light Metals, 275-281 (2001).

18. A. Mayer, J. Comera, H. Charpentier, C. Jaussaud, Absorption coefficients of various pollutant gases at CO2 laser wavelengths - Application to remote sensing of those pollutants, Applied 0/? /c5 17(3), 391-393 (1978).

19. M. W. Sigrist, Air monitoring by spectroscopic techniques (Wiley, New York, 1994). 20. D. Lucas, C. P. Koshland, C. S. McEnally, R. F. Sawyer, The detection of ethyl chloride using

photofragmentation, Combustion Science and Technology 85(1-6), 271-281 (1992). 21. T. A. Cool, B. A. Williams, Optical diagnostics in CHC combustion. Hazardous Waste &

Hazardous Materials 7(1), 21-39 (1990). 22. C. S. McEnally, R. F. Sawyer, C. P. Koshland, D. Lucas, Sensitive in situ detection of

chlorinated hydrocarbons in gas mixtures, Applied Optics 33(18), 3977-3984 (1994). 23. J. B. Simeonsson, R. C. Sausa, A critical review of laser photofragmentation fragment detection

techniques for gas-phase chemical analysis. Applied Spectroscopy Reviews 31(1-2), 1-72 (1996).

24. J. B. Simeonsson, R. C. Sausa, Laser photofragmentation/fragment detection techniques for chemical analysis of the gas phase, TrAC, Trends in Analytical Chemistry 17(8-9), 542-550 (1998).

25. J. D. Winefordner, I. B. Gomushkin, D. Pappas, O. L Matveev, B. W. Smith, Novel uses of lasers in atomic spectroscopy. Journal of Analytical Atomic Spectrometry 15(9), 1161 - 1189 (2000).


26. D. A. Rusak, B. C. Castle, B. W. Smith, J. D. Winefordner, Recent trends and the future of laser-induced plasma spectroscopy, TrAC, Trends in Analytical Chemistry 17(8-9), 453-461 (1998).

27. D. A. Rusak, B. C. Castle, B. W. Smith, J. D. Winefordner, Fundamentals and applications of laser-induced breakdown spectroscopy, Critical Reviews in Analytical Chemistry 27(4), 257-290(1997).

28. L. J. Radziemski, D. A. Cremers, Laser-induced plasmas and applications (M. Dekker, New York, 1989).

29. S. Palanco, J. M. Baena, J. J. Lasema, Open-path laser-induced plasma spectrometry for remote analytical measurements on solid surfaces, Spectrochimica Acta, PartB: Atomic Spectroscopy 57B(3), 591-599 (2002).

30. J. B. Simeonsson, A. W. Miziolek, Time-resolved emission studies of argon monofluoride-laser-produced microplasmas. Applied Optics 32(6), 939-947 (1993).

31. D. A. Cremers, L. J. Radziemski, Detection of chlorine and fluorine in air by laser-induced breakdown spectrometry, Analytical Chemistry 55(8), 1252-1256 (1983).

32. L. Dudragne, P. H. Adam, J. Amouroux, Time-resolved laser-induced breakdown spectroscopy: application for qualitative and quantitative detection of fluorine, chlorine, sulfur, and carbon in air. Applied Spectroscopy 52( 10), 1321 -1327 (1998).

33. C. Haisch, R. Niessner, O. I. Matveev, U. Panne, N. Omenetto, Element-specific determination of chlorine in gases by laser-induced-breakdown-spectroscopy (LIBS), Fresenius' Journal of Analytical Chemistry 356(1), 21-26 (1996).

34. C. K. Williamson, R. G. Daniel, K. L. McNesby, A. W. Miziolek, Laser-induced breakdown spectroscopy for real-time detection of halon alternative agents. Analytical Chemistry 70(6), 1186-1191 (1998).

35. J. B. Jeffries, G. A. Raiche, L. E. Jusinski, Detection of chlorinated hydrocarbons via laser-atomization/laser-induced fluorescence. Applied Physics B: Photophysics and Laser Chemistry B55(l), 76-83 (1992).

36. J. J. Tiee, F. B. Wampler, W. W. Rice, UV laser photochemistry of carbon tetrachloride and trichlorofluoromethane. Journal of Chemical Physics 72(5), 2925-2927 (1980).

37. R. D. Kenner, H. K. Haak, F. Stuhl, CI2 and HCl emissions in the ArF-laser photolysis of chlorinated compounds: Identification and mechanism of generation. Journal of Chemical Physics 85(4), 1915-1923 (1986).

38. J. G. Jinkins, J. Schendel, E. L. Wehry, Laser photolytic fragmentation-fluorescence spectrometry of nonfluorescent organic and organometallic molecules, ASTM Special Technical Publication 1066(Laser Tech. Lumin. Spectrosc), 123-132 (1990).

39. E. L. Wehry, R. Hohmann, J. K. Gates, L. F. Guilbault, P. M. Johnson, J. S. Schendel, D. A. Radspinner, Fragmentation-fluorescence spectrometric determination of nonfluorescent compounds, Applied Optics 26(17), 3559-3565 (1987).

40. M. O. Rodgers, K. Asai, D. D. Davis, Photofragmentation laser-induced fluorescence - A new method for detecting atmospheric trace gases. Applied Optics 19(21), 3597-3605 (1980).

41. J. B. Halpem, W. M. Jackson, V. Mccrary, Multi-photon sequential photo-dissociative excitation - A new method of remote atmospheric sensing, Applied Optics 18(5), 590-592 (1979).

42. M. J. Berry, Chloroethylene photochemical lasers - Vibrational energy content of HCl molecular elimination products. Journal of Chemical Physics 61 (8), 3114-3143 (1974).

43. S. Arepalli, N. Presser, D. Robie, R. J. Gordon, The detection of bromine atoms by resonant multiphoton ionization. Chemical Physics Letters 117(1), 64-66 (1985).

44. R. C. Sausa, J. B. Simeonsson, Detector of halogenated compounds based on laser photofragmentation/fragment stimulated emission, U.S. Pat, 5,866,073, 1999.

45. R, C. Sausa, J. B. Simeonsson, Laser-induced photofragmentation/fragment detection spectrometry of brominated compounds. Army Res. Lab., Aberdeen Proving Ground, MD, USA, 1996, pp. 1-38.

46. P. Brewer, P. Das, G. Ondrey, R. Bersohn, Measurement of the relative populations of atomic iodine C^^i/i) and atomic iodine ( P°3/2) by laser induced vacuum ultraviolet fluorescence. Journal of Chemical Physics 79(2), 720-723 (1983).

47. W. K. Bischel, D. Bamford, M. J. Dyer, G. C. Herring, L. E. Jusinski, Two-photon spectroscopy of atomic fluorine and oxygen. Springer Series in Optical Sciences 55(Laser Spectrosc. 8), 178-180 (1987).


48. W. K. Bischel, Two photon detection techniques for atomic fluorine, SRI Int., Menlo Park, CA, USA., 1988, pp. 1-49.

49. G. C. Herring, M. J. Dyer, L. E. Jusinski, W. K. Bischel, Two-photon-excited fluorescence spectroscopy of atomic fluorine at 170 nm. Optics Letters 13(5), 360-362 (1988).

50. G. Winkelmann, R. Kuhls, R. Osmanow, E. Linke, Laser induced fluorescence from carbon difluoride radical excited at 248 nm. Laser Chemistry 13(1), 43-55 (1993).

51. R. Vetter, W. Renter, S. D. Peyerimhoff, Theoretical spectroscopy of difluoro-methylene in the visible and ultraviolet region. Chemical Physics 161(3), 379-392 (1992).

52. F.-T. Chau, J. M. Dyke, E. P.-F. Lee, D.-C. Wang, Franck-Condon analysis of photoelectron and electronic spectra of small molecules. Journal of Electron Spectroscopy and Related Phenomena 97(1-2), 33-47 (1998).

53. M. R. Cameron, S. H. Kable, G. B. Bacskay, The electronic spectroscopy of jet-cooled difluorocarbene (CF2): the missing A-state stretching frequencies, Journal of Chemical Physics 103(11), 4476-4483 (1995).

54. D. S. King, P. K. Schenck, J. C. Stephenson, Spectroscopy and photophysics of the difluorocarbene radical A^Bi-X^Ai system. Journal of Molecular Spectroscopy 78(1), 1-15 (1979).

55. H. Okabe, Photochemistry of small molecules (Wiley, New York, 1978). 56. J. P. Booth, H. Abada, P. Chabert, D. B. Graves, Radical kinetics in an inductively-coupled

plasma in CF4, AIP Conference Proceedings 740(1), 252-267 (2004). 57. G. Cunge, J. P. Booth, CF2 production and loss mechanisms in fluorocarbon discharges:

Fluorine-poor conditions and polymerization. Journal of Applied Physics 85(8, Pt. 1), 3952-3959(1999).

58. G. Cunge, P. Chabert, J. P. Booth, Absolute fluorine atom concentrations in fluorocarbon plasmas determined from CF2 loss kinetics. Journal of Applied Physics 89(12), 7750-7755 (2001).

59. P. Fendel, A. Francis, U. Czametzki, Sources and sinks of CF and CF2 in a cc-RF CF4-plasma under various conditions. Plasma Sources Science & Technology 14(1), 1-11 (2005).

60. G. Hancock, J. P. Sucksmith, Optical diagnostics of radio-frequency plasmas containing CHF3 and CHF3/O2: Laser-induced fluorescence of CF2, CF, and O atoms, and optical emission from H, F, and O, Journal of Vacuum Science & Technology, A: Vacuum, Surfaces, and Films 20(1), 270-277 (2002).

61. M. Nakamura, M. Hori, T. Goto, M. Ito, N. Ishii, Spatial distributions of the absolute CF and CF2 radical densities in high-density plasma employing low global warming potential fluorocarbon gases and precursors for film formation. Journal of Vacuum Science & Technology, A: Vacuum, Surfaces, and Films 19(5), 2134-2141 (2001).

62. D. S. King, P. K. Schenck, J. C. Stephenson, Spectroscopy and photophysics of laser formed difluoromethylene. Lasers Chem., Proc. Conf 340-344 (1977).

63. M. Rossberg, J. Wolbrandt, W. Strube, E. Linke, Simultaneous LIF and REMPI analysis of fragments after photolysis of highly vibrationally excited CF2CFCI, Journal of Molecular Structure 348(421-424 (1995).

64. R. M. Measures, Laser remote chemical analysis (Wiley, New York, 1987). 65. J. G. Anderson, A simultaneous measurement of chlorine and chlorine oxide in the earth's

stratosphere, Univ. Michigan, Ann Arbor, MI, USA., 1977, pp. 111-113. 66. S. R. Drayson, J. M. Russell, III, J. H. Park, Satellite sensing of stratospheric halogen

compounds by solar occultation. Part I. Low resolution spectroscopy, Radial Atmos., Pap. Int. Symp 24S-249 (1911).

67. J. M. Vadillo, P. L. Garcia, S. Palanco, D. Romero, J. M. Baena, J. J. Lasema, Remote, realtime, on-line monitoring of high-temperature samples by noninvasive open-path laser plasma spectrometry, Analytical and Bioanalytical Chemistry 375(8), 1144-1147 (2003).

68. S. Palanco, S. Conesa, J. J. Lasema, Analytical control of liquid steel in an induction melting furnace using a remote laser induced plasma spectrometer. Journal of Analytical Atomic Spectrometry 19(4), 462-467 (2004).

69. J. R. Bettis, Correlation among the laser-induced breakdown thresholds in solids, liquids, and gasQS, Applied Optics 31(18), 3448-3452 (1992).

70. U. Panne, Laser remote sensing, Trac-Trends in Analytical Chemistry 17(8-9), 491-500 (1998). 71. S. Svanberg, in: Air monitoring by spectroscopic techniques, edited by J. D. Winefordner

(Wiley, New York, 1994), pp. 85-161.


72. V. P. Kandidov, O. G. Kosareva, I. S. Golubtsov, W. Liu, A. Becker, N. Akozbek, C. M. Bowden, S. L. Chin, Self-transformation of a powerful femtosecond laser pulse into a white-light laser pulse in bulk optical media (or supercontinuum generation), Applied Physics B: Lasers and Optics 11(2-3% 149-165 (2003).

73. A. Couairon, Filamentation length of powerful laser pulses. Applied Physics B-Lasers and Optics 16(1), 789-792 (2003).

74. A. Brodeur, C. Y. Chien, F. A. Ilkov, S. L. Chin, O. G. Kosareva, V. P. Kandidov, Moving focus in the propagation of ultrashort laser pulses in air. Optics Letters 22(5), 304-306 (1997).

75. H. R. Lange, G. Grillon, J.-F. Ripoche, M. A. Franco, B. Lamouroux, B. S. Prade, A. Mysyrowicz, E. T. J. Nibbering, A. Chiron, Anomalous long-range propagation of femtosecond laser pulses through air: moving focus or pulse self-guiding?. Optics Letters 23(2), 120-122(1998).

76. S. L. Chin, A. Brodeur, S. Petit, O. G. Kosareva, V. P. Kandidov, Filamentation and supercontinuum generation during the propagation of powerful ultrashort laser pulses in optical media (white light laser). Journal of Nonlinear Optical Physics & Materials 8(1), 121-146(1999).

77. J. Kasparian, R. Sauerbrey, D. Mondelain, S. Niedermeier, J. Yu, J. P. Wolf, Y. B. Andre, M. Franco, B. Prade, S. Tzortzakis, A. Mysyrowicz, M. Rodriguez, H. Wille, L. Woste, Infrared extension of the supercontinuum generated by femtosecond terawatt laser pulses propagating in the atmosphere, Optics Letters 25(18), 1397-1399 (2000).

78. J. Kasparian, M. Rodriguez, G. Mejean, J. Yu, E. Salmon, H. Wille, R. Bourayou, S. Frey, Y. B. Andre, A. Mysyrowicz, R. Sauerbrey, J. P. Wolf, L. Woste, White-light filaments for atmospheric analysis, Science 301(5629), 61-64 (2003).

79. P. Rairoux, H. Schillinger, S. Niedermeier, M. Rodriguez, F. Ronneberger, R. Sauerbrey, B. Stein, D. Waite, C. Wedekind, H. Wille, L. Woste, C. Ziener, Remote sensing of the atmosphere using ultrashort laser pulses, Applied Physics B: Lasers and Optics 71(4), 573-580 (2000).

80. G. Mejean, J. Kasparian, E. Salmon, J. Yu, J. P. Wolf, R. Bourayou, R. Sauerbrey, M. Rodriguez, L. Woeste, H. Lehmann, B. Stecklum, U. Laux, J. Eisloeffel, A. Scholz, A. P. Hatzes, Towards a supercontinuum-based infrared lidar. Applied Physics B: Lasers and Optics 77(2-3), 357-359 (2003).

81. P. Rohwetter, J. Yu, G. Mejean, K. Stelmaszczyk, E. Salmon, J. Kasparian, J. P. Wolf, L. Woeste, Remote LIBS with ultrashort pulses: characteristics in picosecond and femtosecond regimes. Journal of Analytical Atomic Spectrometry 19(4), A31AAA (2004).

82. K. Stelmaszczyk, P. Rohwetter, G. Mejean, J. Yu, E. Salmon, J. Kasparian, R. Ackermann, J.-P. Wolf, L. Woste, Long-distance remote laser-induced breakdown spectroscopy using filamentation in air. Applied Physics Letters 85(18), 3977-3979 (2004).

83. B. Chatel, J. Degert, S. Stock, B. Girard, Competition between sequential and direct paths in a two-photon transition, Physical Review A: Atomic, Molecular, and Optical Physics 68(4A), 041402/041401 -041402/041404 (2003).

84. N. Dudovich, B. Dayan, S. M. G. Faeder, Y. Silberberg, Transform-limited pulses are not optimal for resonant multiphoton transitions, Physical Review Letters 86(1), 47-50 (2001).

85. R. Stoian, M. Boyle, A. Thoss, A. Rosenfeld, G. Kom, I. V. Hertel, E. E. B. Campbell, Laser ablation of dielectrics with temporally shaped femtosecond pulses. Applied Physics Letters 80(3), 353-355 (2002).

86. J. Kasparian, R. Sauerbrey, S. L. Chin, The critical laser intensity of self-guided light filaments in air. Applied Physics B: Lasers and Optics 71(6), 877-879 (2000).

87. A. Becker, N. Akozbek, K. Vijayalakshmi, E. Oral, C. M. Bowden, S. L. Chin, Intensity clamping and re-focusing of intense femtosecond laser pulses in nitrogen molecular gas, Applied Physics B: Lasers and Optics 73(3), 287-290 (2001).

88. H. R. Lange, A. Chiron, J. F. Ripoche, A. Mysyrowicz, P. Breger, P. Agostini, High-order harmonic generation and quasiphase matching in xenon using self-guided femtosecond pulses. Physical Review Letters 81(8), 1611-1613 (1998).

89. A. Talebpour, M. Abdel-Fattah, S. L. Chin, Focusing limits of intense ultrafast laser pulses in a high pressure gas: Road to new spectroscopic source, Optics Communications 183(5,6), 479-484 (2000).

90. A. Talebpour, M. Abdel-Fattah, A. D. Bandrauk, S. L. Chin, Spectroscopy of the gases interacting with intense femtosecond laser pulses, Laser Physics 11(1), 68-76 (2001).


91. Q. Luo, S. A. Hosseini, B. Ferland, S. L. Chin, Backward time-resolved spectroscopy from filament induced by ultrafast intense laser pulses, Optics Communications 233(4-6), 411-416 (2004).

92. Q. Luo, W. Liu, S. L. Chin, Lasing action in air induced by ultra-fast laser filamentation. Applied Physics B: Lasers and Optics 76(3), 337-340 (2003).

93. J. F. Gravel, Q. Luo, D. Boudreau, X. P. Tang, S. L. Chin, Sensing of halocarbons using femtosecond laser-induced fluorescence. Analytical Chemistry 76(16), 4799-4805 (2004).

94. A. Talebpour, A. D. Bandrauk, J. Yang, S. L. Chin, Multiphoton ionization of inner-valence electrons and fragmentation of ethylene in an intense Ti:sapphire laser pulse. Chemical Physics Letters 313(5,6), 789-794 (1999).

95. A. Becker, A. D. Bandrauk, S. L. Chin, S-matrix analysis of non-resonant multiphoton ionization of inner-valence electrons of the nitrogen molecule. Chemical Physics Letters 343(3,4), 345-350 (2001).

96. W. Liu, Q. Luo, S. L. Chin, Competition between multiphoton/tunnel ionization and filamentation induced by powerful femtosecond laser pulses in air., Chinese Optics Letter 1(1), 56-59 (2003).

97. S. A. Hosseini, Q. Luo, B. Ferland, W. Liu, N. Akoezbek, G. Roy, S. L. Chin, Effective length of filaments measurement using backscattered fluorescence from nitrogen molecules, Applied Physics B: Lasers and Optics 77(6-7), 697-702 (2003).

98. G. Mejean, J. Kasparian, J. Yu, S. Frey, E. Salmon, J. P. Wolf, Remote detection and identification of biological aerosols using a femtosecond terawatt lidar system. Applied Physics B: Lasers and Optics 78(5), 535-537 (2004).

99. H. Wille, M. Rodriguez, J. Kasparian, D. Mondelain, J. Yu, A. Mysyrowicz, R. Sauerbrey, J. P. Wolf, L. Woste, Teramobile: A mobile femtosecond-terawatt laser and detection system, European PhysicalJournal-Applied Physics 20(3), 183-190 (2002).

100. Q. Luo, S. A. Hosseini, W. Liu, J. F. Gravel, O. G. Kosareva, N. A. Panov, N. Akoezbek, V. P. Kandidov, G. Roy, S. L. Chin, Effect of beam diameter on the propagation of intense femtosecond laser pulses. Applied Physics B: Lasers and Optics 80(1), 35-38 (2004).

101. G. Mejean, F. Courvoisier, J. Kasparian, V. Boutou, E. Salmon, J. Yu, J.-P. Wolf, Nonlinear aerosol lidar for remote detection and identification of bioaerosols in clouds, European Space Agency (Special Publication) SP-561(1), 25-28 (2004).

102. R. J. Levis, G. M. Menkir, H. Rabitz, Selective bond dissociation and rearrangement with optimally tailored, strong-field laser pulses. Science 292(5517), 709-713 (2001).

103. T. Brixner, N. H. Damrauer, P. Niklaus, G. Gerber, Photoselective adaptive femtosecond quantum control in the liquid phase. Nature 414(6859), 57-60 (2001).

FRET STUDIES OF CONFORMATIONAL TRANSITIONS IN PROTEINS

Herbert C. Cheung and Wen-Ji Dong^

20.1. INTRODUCTION

In recent years, fluorescence spectroscopy has been increasingly recognized as a useful tool for probing changes in localized and global conformations of macromolecules. Both intrinsic and extrinsic fluorophores can be used for these studies and for determination of intermolecular interactions. In proteins, the three aromatic amino acid residues have emission properties that can be exploited to reveal localized conformational changes. Such changes are usually described by changes in some basic spectroscopic properties of the fluorophores, rather than by a quantitative parameter to indicate the magnitude of the changes. When two fluorophores attached to the same protein or two proteins in a complex are used, the separation between their sites (R) can be quantified from measurements of the Forster type of resonance energy transfer (FRET) from an initially excited donor fluorophore (D) to an acceptor fluorophore (A). This transfer occurs without the appearance of a photon and results from a long-range dipole-dipole interaction between the donor emission dipole and the acceptor absorption dipole. Because the acceptor emission dipole is not involved in the transfer, the acceptor probe need not be fluorescent if its absorption spectrum overlaps with the donor emission spectrum. The rate of transfer (and the efficiency of transfer) is dependent upon the extent of the spectral overlap between the long-wavelength region of donor emission spectrum and the short-wavelength region of the acceptor absorption spectrum.

Department of Biochemistry and Molecular Genetics, University of Alabama at Birmingham, 1530 3*^ Avenue South, Birmingham, AL 35294-0005, U.S.A.

445

446 H.C. CHEUNG and W-J. DONG

Because the transfer rate is inversely proportional to R , FRET is a sensitive method to detect small changes in inter-site distances. The Forster distance (Ro) is the distance at which FRET is 50% efficient. The range of Ro for many donor-acceptor pairs that are useful for studies of proteins is in the range 20-75 A, although the Ro for a few pairs with lanthanide ions as donors extends the range to 90 A. Measurements of R are reliable only within a factor of two of RQ. FRET offers a convenient experimental approach to the determination of molecular distances that are comparable to the molecular dimensions of many proteins. FRET studies can be carried out in both equilibrium and kinetic modes. In addition, a sufficiently large number of FRET distances can be used as distance restraints for construction of molecular models of single-polypeptide proteins or protein complexes. In the present review, we summarize our recent studies of striated muscle proteins, using different FRET approaches to investigate conformational transitions that play critical roles in activation and regulation of cardiac muscle contractility. Also reviewed are studies of kinesin motor domain complexed with ADP and AIF4.

20.2. CALCIUM ACTIVATION OF CARDIAC MUSCLE

In relaxed striated muscle, the intracellular Ca^^ concentration of myocytes is in the sub-micromolar range, and actomyosin ATPase and tension development are inhibited. These inhibitions, which are due to a strong interaction between a component (Tnl) of the heterotrimeric troponin complex (Tn) and actin, are relieved when the Ca^^ concentration increases to the submillimolar range upon activation. In the activated state, the elevated level of Ca^^ binds to a Ca ^-specific regulatory site in a second component (TnC) of the troponin complex. This site is located in the N-terminal domain of cTnC, and Ca ^ binding to this site is the key molecular event that triggers activation of striated muscle.

The cardiac isoforms of TnC (cTnC) and Tnl (cTnl) differ from the corresponding isoforms from skeletal muscle in that cTnC has only one Ca^^ regulatory site and cTnl has a unique N-terminal extension (32-34 residues) that is absent in the skeletal isoform. Within this N-terminal extension, two adjacent serine residues (Ser23 and Ser24) are targets of protein kinase A (PKA). Phosphorylation of these cTnl residues within the troponin complex results in a loss of Ca^^ sensitivity in cardiac muscle function. cTnl also contains specific sites (Ser44, Ser45, and Thrl44) that can be phosphorylated by protein kinase C (PKC). Certain aspects of cardiac contractility are modified also by PKC phosphorylation.

Ca^^ binding to the regulatory site in cTnC induces several conformational changes in the troponin complex. These include (1) reorientation of the helices in the N-terminal domain of TnC, (2) alteration of the secondary structure of a region of Tnl (inhibitory region), and (3) movement of another region (regulatory region) of Tnl into a hydrophobic pocket of the Ca^^-bound open N-terminal domain of TnC. The inhibitory region is strongly bound to actin in relaxed muscle, and the Ca^^-induced alteration of its secondary structure

FRET STUDIES OF CONFORMATIONAL TRANSITIONS IN PROTEINS 447

weakens or breaks its contact with actin thus making it possible for strong interaction of actin with myosin motor to occur. The latter interaction leads to tension development in muscle. Figure 20.1 illustrates the global conformations of cTnC and cTnl within the troponin complex and their relationships to actin, tropomyosin, and myosin motor before and after regulatory Ca^^ binds to cTnC. In the relaxed stated (Mg^^-saturated), the C-terminal domain of cTnC is saturated with two Mg^^ ions and the N-domain is devoid of bound cation. In the Ca^^-activated state (Ca^^-saturated), the N-domain is saturated with Ca^^ at the single regulatory site, and bound Mg^^ in the Mg^^-saturated C-domain is exchanged with bound Ca^ . In the following sections, we describe the experimental systems that we used for FRET measurements to determine Ca^^-induced changes in inter-site distances and the kinetics of these changes involving the troponin subunits.

20.2.1. Equilibrium Conformation of Cardiac Troponin

Conformations of two of the three subunits (cTnC and cTnl) of cardiac troponin within the heterogeneous trimeric complex were studied by FRET with

N-domairiofTnC N domain of TnC

Tjp20

Deactivated State (Mg^^-saturated)

Activated State (Ca^^-saturated)

Figure 20.1. A model for the proximity relationship of troponin, tropomyosin (Tm), and actin to illustrate Ca^^-induced changes in the tertiary structure of cTnC (blue) and the secondary structure of cTnl (red). For simplicity, cTnT is omitted. In the deactivated state, the C-domain of cTnC is saturated with two Mg^^ and the N-domain is unoccupied by Ca^ . In the activated state, the N-domain of cTnC is saturated by Ca^^ at its single site, and bound Mg^^ in the C-domain is replaced by Ca^ . The disposition of myosin head is arbitrary. Residues in cTnC that are involved in FRET measurements are shown. The positions of several key residues in cTnl are indicated to show changes in secondary structure (inhibitory region) and movement of the regulatory region into the N-domain of cTnC. (See color insert section.)


donor and acceptor probes placed in strategic positions in the proteins. Initial studies were focused on intramolecular and intermolecular inter-probe (inter-site) distances at the two ends of the contractile cycle: deactivated and activated states. Samples for these states were prepared by controlling [Ca^] in the solution with EGTA. In the deactivated state, the solution contained 2 mM Mg^^ and 10* ^ M Ca^^ (pCa 7.5). In the activated state, the solution contained 2 mM Mg^^ and 0.1 mM Ca^^ (pCa 4) to achieve full saturation by Ca^ . Inter-probe distances in these two biochemical states were calculated from the distribution of donor-acceptor distances determined from measurements of donor intensity decay in the time domain. In other studies, Ca ^ titration experiments were performed from pCa 7.5 to 4 to establish the effect of [Ca^^] on FRET (and inter-probe distance) over the entire range of Ca^^concentration. The FRET signal in these experiments was determined by the steady-state method and was corrected for environmental sensitivity of donor probe at each [Ca^^]. The titration experiments provided physiologically relevant information for a functional parameter of the system. The free [Ca^^] at which the change in distance is 50% of the maximum (pCaso) is taken as the apparent affinity of the system for Ca^ . A shift of the observed pCaso value to lower values indicates a loss of Ca^^ sensitivity and a shift to higher values indicates a gain in Ca^^sensitivity. The Hill coefficient indicating Ca ^ binding cooperativity can be calculated from the titration curve.

20.2.1.1. Conformational Markers for the cTnC N-domain

Cardiac TnC is a single polypeptide protein (161 amino acid residues) and is dumbbell-shaped with the N-terminal and C-terminal regions folded into two globular domains. The N-domain has a single Ca^^-specific regulatory site, five a-helices, and a hydrophobic pocket that is shielded from solvent in the absence of bound Ca^ . The C-domain has two metal sites that bind both Mg^^ and Ca ^ competitively. In relaxed muscle, the regulatory site is unoccupied and the two metal sites in the C-domain are occupied by Mg^^ because the intracellular concentration of Mg^^ in myocytes is in the millimolar range. Cardiac muscle is activated upon binding of Ca^^ to the regulatory site in the N-domain. This binding induces reorientation of some of the helices and exposes the hydrophobic pocket. Reorientation of the helices gives rise to a more open tertiary structure of the N-domain (Fig. 20.1). The open conformation of the Ca^^-saturated N-domain is energetically unstable, but stabilized by an interaction between the exposed hydrophobic pocket and a predominantly hydrophobic regulatory region of cTnl (residues 150-165). The sliding of the regulatory region of cTnl into the open cTnC N-domain breaks the contact between the inhibitory region of cTnl (residues 130-149) and actin. This physical model of Ca^^ activation is based on biochemical, FRET, and heteronuclear NMR studies.

In our initial work which demonstrated a Ca^^-induced transition from the closed tertiary conformation to an more open conformation (Dong et al., 1999), we used full-length cTnC mutants containing a single tryptophan in position 12 (Phel2Trp) or 20 (Phe20Trp) as energy donor, and 1,5-IAEDANS [5-


(iodoacetamidoethyl) aminonaphtha-lene-sulfonic acid] covalently attached to a single cysteine in position 51 (Asn51Cys) as energy acceptor. Ro for this donor-acceptor pair is ca. 20-22 A. Since native cTnC has two endogenous cysteines in positions 35 and 84, these residues were substituted by serine. A tryptophaness cTnl mutant (Trpl92Cys) was used to prepare the binary cTnC»cTnI complex. In the Mg^^-saturated state in which the N-domain of cTnC contained no bound Ca^^ the Trpl2-Cys51 and Trp20-Cys51 distances were 19.3 and 15.4 A, respectively. In the Ca^^-saturated state (N-domain of cTnC saturated with bound Ca^^), both of these distances increased by 6.5 A. These increases reflect movement of the region of Cys51 away from residues 12 and 20, resulting in a more open conformation. Similar increases were also observed with the ternary complex cTnC*cTnI*cTnT in subsequent studies. The distance changes were reversible upon removal of bound Ca^^ by EGTA. In isolated cTnC, the Ca^^ effect was negligible (0.8 A). The open cTnC conformation in troponin elicited by bound Ca^^ is obligatory for other conformational transitions within the troponin complex that occur during the activation process.

We recently developed double-cysteine markers to monitor domain opening and closing using Phel2Cys/Asn51Cys (Dong et al., 2004) or Leul3 Cys /Asn51Cys (Dong et al., 2005). The two-cysteine mutants were first labeled with 1,5-IAEDANS as energy donor at Cysl2 or Cysl3, and with DDPM (N-[4-dimethylamino)-3,5-dinitrophenyl]-maleimide) as energy acceptor at Cys51. The double-cysteine cTnC was first labeled with the donor probe under conditions resulting in partial labeling of both cysteines. The reaction products contained a heterogeneous population of cTnC: unlabeled, doubly labeled, and singly labeled at either Cys51 or Cysl2/Cysl3. These species were separated in a cationic DEAE column with a salt gradient. Ionic strength and pH were the critical factors for good separation. The two singly labeled species were pooled, followed by incubation with the acceptor probe to label the other cysteine. It was not critical to ascertain which site was labeled with the donor and which was labeled with the acceptor. The Ro for the AEDANS-DDPM donor-acceptor pair is in the range 28-30 A, suitable for determination of the Cysl2-Cys51 and Cysl3-Cys51 distances. In reconstituted Mg^^-saturated trimeric cTn (cTnC»cTnI»cTnT), both Cysl2-Cys51 and Cysl3-Cys51 distances were comparable (23.2 and 25.9 A, respectively). In the Ca^^-saturated complex, the Cysl2-Cys51 and Cysl3-Cys51 distances increased by 7.3 and 7.8 A, respectively. These changes are comparable to those determined with Trpl2 as the energy donor. The advantage of the double-cysteine markers is that the N-domain conformation of cTnC can be studied in fully reconstituted myofilaments, which consist of the trimeric troponin, actin, tropomyosin, and myosin motor domain, without having to use tryptophanless mutants of troponin I, troponin T, actin and myosin.

20.2.1.2. FRET Markers for c Tnl Secondary Structure

As depicted in Fig. 20.1, the sequence between residues 129 and 165 is of physiological interest regarding the role of cTnl in activation and regulation of


cardiac myofilaments. There are two functionally important regions in cTnl: (1) the inhibitory region (residues 130-149) and (2) the regulatory region (residues 150-165). We were interested in the secondary structure of these regions, especially potential changes in the structure triggered by Ca^^ binding. We determined four inter-probe FRET distances within these regions using tryptophan as the energy donor and AEDANS attached to a single cysteine as energy acceptor (Dong et al., 2001; Dong et al., 2003a). Distance A is between Trpl29 and Cysl52 and spans the inhibitory region, and distance B is between residues 150 and 167, spanning the regulatory region. In the relaxed state (Mg^^-saturated cTnC), the secondary structure of Distance A is depicted by a helix-tum-helix motif. This motif is suggested by the amino acid sequence of the segment. Distance A changed from 19.8 A to 29.4 A, an increase of 9.6 A, in the Ca^^-activated state (Ca^^-saturated) (Dong et al., 2001). This large increase suggests that the helix-tum-helix motif of the inhibitory region became extended in which the turn was straightened out. The Ca^^ titration curve of the FRET signal from Distance A was biphasic, and the pCaso values for the two phases were 0.64 pCa^^ units apart (Dong et al., 2003b). These results suggest Ca^^ binding to two sets of sites with different affinities over the pCa range 7.5-4, both the high-affinity Ca^VMg^^ sites in the C-domain and the single low-affinity Ca^^-specific site in the N-domain. Distance B was unaltered in the Ca^^-activated state, changing from 22.7 A to 22.6 A. Distance C (between residues 129 and 160) spanned the inhibitory region and the N-terminal portion of the regulatory region, extending approximately to its midpoint. Upon addition of Ca^ , Distance C increased by 14.7 A from 24.5 A to 39.2 A. Distance D (residues 129-167), spanning both the inhibitory and regulatory regions, increased by 18.9 A from 26.8 A to 45.7 A in the presence of Ca^ . These FRET results suggest that Ca^^ activation of cTn is accompanied by changes in the secondary structure of the cTnl inhibitory region in which the helix-tum-helix motif switches to an extended quasi-a-helical secondary structure, as the contacts between the inhibitory region and actin are broken. During Ca^^activation, the regulatory region remains a-helical, moving as a rigid segment into the open N-domain of cTnC. This movement is facilitated by the change in the secondary stmcture of the inhibitory region. The inter-probe distances were calculated using experimental values of Ro based on a value of % for the orientation factor (K^). If the assumption of rapid and random tumbling of donor and acceptor probes (dynamic averaging) were not appropriate, the calculated distances would be in error. Since our main interest was in the difference in R determined in two biochemical states, it was more important to ascertain that K was similar in both states. We determined axial depolarization factors of both donor and acceptor for all four distances and found negligible Ca^^ effects on these factors, suggesting that Ca^^ binding had little effect on the orientational freedom of the probes.

The experimental mean FRET inter-probe distances (22.7 and 22.6 A) for Distance B (residues 150-167) are within 4 A of the Ca-Ca distance of an 18-residue native a-helix (26.4 A). The -^4-A distance difference can be attributed to donor and acceptor probe-Ca offset distances. In general, this offset distance is determined by probe stmcture/size and environmental factors. Because the


same donor-acceptor pair was used for determination of all four distances within the inhibitory and regulatory regions, it is reasonable to assume that a good approximation of the Ca-Ca distance for the other three inter-probe distances is obtained by adding 4 A to the experimental mean FRET distances to correct for probe-Ca offset. Using these crudely corrected distances to set long-range constraints, we constructed a reduced representation molecular model of the inhibitory/regulatory regions of cTnl in the Mg^^-saturated and Ca^^-saturated states consistent with Distances A-D (Dong et al., 2003a). The experimental FRET distances are compared with the model distances (Ca-Ca) in Table 1.

Table 20.1: Comparison of Observed FRET and Model Distances

Distance

A B C D

Observed distance, A

Residues

129-152 150-167 129-160 129-167

Mg^^

23.9 26.7 28.5 30.8

Ca^^

32.5 26.6 43.2 49.7

Chang!

8.6 -0.1 14.7 18.9

Model distance, A

e Mg^^

23.9 26.4 27.2 32.2

Ca^"

32.5 26.4 41.6 51.9

Change

8.6 0.0 14.4 19.7

A 4-A probe-Ca offset has been added to the observed FRET distances, and this adjustment has yielded good agreement between the adjusted experimental distances and model results. Importantly, the difference in the distances between the two biochemical states is not affected by the addition of probe-Ca offset. These results demonstrate that under favorable conditions, as in our system, FRET measurements can yield reliable results on differences between inter-probe separations even without adjustment of probe-Ca offset. This is important because for many experimental systems it is not always possible to estimate probe-Ca offsets in order to correct FRET distances. This correction does not appear necessary if the important structural information being sought is the change in inter-probe separation.

Conformational transitions that occur within the inhibitory region of cTnl in the activation-deactivation cycle are the structural basis in the Ca^^ switch in muscle contraction. High-resolution structural data are not available to show these transitions. The most recent crystal structure of cardiac troponin saturated with Ca^^ was determined with truncated cTnl and cTnT in which certain functionally important regions in these two proteins were absent (Takeda et al., 2003). Although the cTnl inhibitory and regulatory regions are intact in the crystal structure, no crystallographic information is available for the inhibitory


region presumably because of its flexibility. In addition, the absence of the corresponding structure without bound regulatory Ca^^ in the N-domain of cTnC precludes an analysis of structural differences in the inhibitory region (or any other region) between the activated and deactivated states. Experimental results of these differences need to be considered in formulation of molecular mechanistic models of cardiac muscle contraction. As summarized in this section, FRET studies have provided such information needed to understand the structural basis of the Ca^^ switching mechanism.

20.2,1.3. FRET Markesrfor the cTnC*cTnIInterface

Figure 20.1 shows a third conformational feature that was studied by FRET. The interface between cTnC and cTnl is far apart in the region between the central helix of cTnC and the C-terminal helix of the inhibitory region of cTnl in the deactivated state. These regions of the two proteins are expected to come closer toward each other in the Ca^^-activated state. The FRET distance between cTnI(Trpl50) and cTnC(Cys89-AEDANS) was 29.4 A in the deactivated state and 22.7 A in the activated state, a decrease of 6.7 A (Dong et al., 2003b). This change resulted from Ca^^ binding to two sets of sites as reflected by a biphasic Ca^^ titration curve with two pCaso values (0.55 pCa^^ units apart). This distance decrease is consistent with the Ca^^-induced transition of the inhibitory region of cTnl from a helix-tum-helix motif to an extended configuration. The relatively long separations between residues of cTnC and cTnl in these regions is consistent with NMR results (Abbott et al., 2001) showing no interactions between the central helix region of cTnC and cTnl residues 129-166, and with crystal structures (Takeda et al., 2003) showing no specific interactions of this cTnC region with other parts of cTnl and cTnT.

o c

e m H m

22

20

18 o c S m Q

0.5

0.4 I

16

14 0,00 0.02 0.04 0.06

Tlme/s

0.08 0.10

Figure 20.2. Kinetics of the closing of cTnC N-domain triggered by dissociation of bound Ca from that domain. The FRET transient monitors increase in energy transfer, and the distance transient follows decrease of inter-probe distance.


20.3. KINETICS OF CONFORMATIONAL TRANSITIONS IN cTN

FRET studies have shown different equilibrium conformations of cTnC, cTnl, and the interface of these two proteins at the two extreme levels of [Ca^^] that correspond to the intracellular environment of cardiac muscle cells in the deactivated and activated states. Ca^^ titration experiments show that the magnitude of each conformational change is proportional to changes in [Ca^^] between pCa 7.5 and 4. The Hill coefficients for Ca^^ binding as revealed in the titration curves are in the range 1.8-2.1, suggesting some degree of cooperativity. It is of interest to establish whether these conformational transitions between the two pCa values occur with rates that are physiologically relevant to activation and deactivation. The first set of kinetic experiments was performed to determine the transition rates triggered by Ca^^ dissociation with cTn preparations that contained the Trp-AEDANS donor-acceptor markers for the Trp20-Cys51 distance in cTnC N-domain (Dong et al., 2003b). Experiments were carried out by mixing a Ca^^-saturated cTn sample with a solution containing EGTA to trigger removal of bound Ca^^ in stopped-flow experiments. As expected, the FRET signal of the Trp20-Cys51 distance in cTnC increased rapidly upon mixing, indicating an increase in energy transfer and a decrease in inter-probe distance (domain closing). Since the donor probe was sensitive to Ca^ , a control stopped-flow run was carried out with a donor-only sample and this transient was used to correct the apparent FRET transient from which the distance transient was constructed. The corrected FRET transient is single-exponential (Fig. 20.2). On the other hand, The FRET transients for changes in the secondary structure of the cTnl inhibitory region and the separation between cTnl and cTnC in their interface were biphasic, with a fast phase and a very slow phase. The rate constants fi^om these three sets of experiments are shown in Table 2, together with the corresponding FRET distances derived from equilibrium experiments.

The half-time (t./J of the transient for cTnC domain closing was 5.4 ms, and the distance decrease (AR) derived fi*om the kinetic tracing was in good agreement with that determined from equilibrium measurements. The fractional amplitudes of the two phases of the other two distance transients allowed estimates of the AR associated with each kinetic phase. The sum of the two AR values in each case was in excellent agreement with the observed equilibrium value. The time constants ty, for the two fast phases were 5.2 ms, and those for the slow phases were 70-100 folds slower.

To interpret the origin of the biphasic feature of FRET transients, we determined directly the dissociation of bound Ca^^ from Ca^^-saturated cTnC in cTn (three bound Ca^ , one bound to the single low-affinity regulatory site in the N-domain and two bound to the high-affinity Mg^VCa^^ sites in the C-domain), using the fluorescent chelator Quin-2. Its fluorescence was enhanced upon chelating Ca^ . The kinetic tracing was biphasic with rate constants 199 s" for the fast phase (1/3 of total amplitude) and 1.45 s" for the slow phase (2/3 of total amplitude). The relative amplitudes indicated that one of three bound Ca^^ dissociated with the fast rate (ti/, - 3 . 5 ms) and the other two bound Ca^^


Table 20.2. Kinetic Parameters for Conformational Transitions Induced by Ca^^ Dissociation

Kinetics Equilibrium

Transition rate, ti (ms)

Transition

3^ structure change in cTnC (N-domain closing)

2^ structure change in cTnl

Fast Slow

5.4

5.2 526

AR,A AR,A

Fast Slow

-5.4 -- -5.9

-5.7 -3.1 -8.6 (shortening of inhibitory region)

cTnl-cTnC interface (increaseofcTnl-cTnC distance) 5.2 361 +5.7 +1.1 +6.7

dissociated with the slow rate (iy, ~ 478 ms). Previous studies in several groups including our own showed that Ca^^ dissociation from the regulatory site was two orders of magnitude faster than the dissociation from the two sites in the C-domain. This information suggested that the fast phase detected by Quin-2 with a t./, of 3.5 ms tracked Ca ^ dissociation from the single regulatory site in the N-domain, and the slow phase monitored dissociation of Ca^^ from the two sites in the C-domain. To demonstrate further that the slow phase was due to Ca^^ dissociation from the sites in the C-domain, we used a cTnC mutant in which two acidic residues were substituted by valine and alanine, cTnC(Asp64Val/Asp66Ala), in the kinetic experiments. Asp64 and Asp66 are two of the six ligands within the 12-residue Ca^^- binding loop for regulatory Ca^^ in the N-domain. The mutations are known to abolish Ca ^ binding completely at this site. The FRET transient obtained with cTn reconstituted with this cTnC mutant was monophasic, with a t./, of ~ 480 ms. The previously observed fast phase was completely eliminated. These results unequivocally support the conclusion that the fast phase of the FRET transient is triggered by Ca^^ dissociation from the regulatory site. Thus, the closing of cTnC N-domain was triggered by Ca^^ dissociation from the regulatory site and occurred in one step with a rate that was a factor of 1.5 slower than Ca^^ dissociation. The other two transitions occurred in two steps, a fast step associated with Ca^^ dissociation from the regulatory site and a slow step associated with dissociation from the two sites in the C-domain. During deactivation in which bound Ca^^ rapidly dissociated from the regulatory site. Distance A in cTnl (the inhibitory region) was shortened by 5.7 A and the interface between cTnl and cTnC was expanded by 5.7 A. These fast distance changes were followed by additional, but smaller, changes associated with the slow dissociation of Ca^^


from the C-domain over a longer time scale. Because of the very slow kinetics, the second step of the transition is not relevant to the physiology of cardiac contractility. The rates (t./,) of the three physiologically important confor mational transitions were essentially the same and suggested that the transitions were coupled to one another in a cooperative manner in restoring their conformations in the deactivated state. The FRET kinetic results define the magnitudes of structural changes relevant to Ca^^ switching between activation and deactivation in cardiac muscle. Equilibrium studies cannot delineate the fraction of the total distance change between the activated and deactivated states resulting from deactivation, i.e., induced by dissociation of bound regulatory Ca^ . FRET-based kinetic results provide the answer. A question arises as to why the markers of the conformations of the Distance A in cTnl and the cTnl-cTnC interface sensed Ca^^ dissociation from both the N-domain and C-domain. The two donors cTnI(Trpl29) and cTnI(Trpl50) of these FRET markers were located within cTn closer to the C-domain and than the N-domain of cTnC in the deactivated state. This disposition may be favorable for the donor-acceptor pairs to be sensitive to Ca^^ dissociation from both cTnC domains.

20.4. CONFORMATION OF N-DOMAIN OF cTnC IN MYOFILAMENT

With the FRET markers in place, the reversible opening and closing of the cTnC N-domain was studied under conditions that simulate Ca^^ activation and deactivation that occur in myocytes. The preparation was the regulated thin filament, which was assembled by reconstituting troponin (Tn), tropomyosin (Tm), and actin (A) into thin filaments with the stoichiometry TnTmAy. The Tn was first reconstituted with cTnC containing FRET markers attached to the double cysteine mutant Cysl2/Cys51 for monitoring domain opening and closing. The FRET signal in the regulated thin filament was measured from pCa 7.5 to 4 to obtain a complete Ca^^ titration curve. This titration was also performed with Tn alone, the complexes TnTm, and TnTmA7(S 1 + MgADP) in which SI was myosin subfragment 1 (motor domain). All four titration curves were monophasic, each with a well defined mid-point (pCaso). The measured mean distances, pCaso, and the Hill coefficients for the four preparations are given in Table 3 (Dong et al., 2003b).

The size of domain opening was in the range 5.3-7.2 A, and the pCaso values were identical in the first three preparations. The Hill coefficients indicated considerable cooperativity in Ca^^ binding to cTnC in these three preparations (Hill coefficient > 2). The N-domain of cTnC in isolated troponin appeared to be quite similar to that in the fully reconstituted system. However, pronounced differences were observed with the fourth preparation. The pCaso was shifted to a higher value by 0.31 units. This is a significant shift and indicated a small gain in Ca^^ affinity. The gain in affinity resulted from a strong interaction of myosin with the regulated thin filament, but diminished Ca^^ binding cooperativity in comparison with the regulated thin filament itself (the Hill coefficient decreasing from 2.89 to 1.69).


Table 20.3. Ca^^ Titration of Myofilament Monitored by cTnC N-domain Markers

Preparation

Tn TnTm TnTmAy TnTmA7(Sl+ADP)

Distance in cTnC N-domain (Residues 12-51)

Mg^^

22.5 23.1 23.0 22.5

Ca^"

29.2 29.8 28.3 29.7

,A AR

6.7 6.7 5.3 7.2

pCaso

5.33 5.33 5.33 5.64

Hill coefficient

2.37 2.50 2.80 1.69

We also performed titration experiments with preparations similar to those shown in Table 3, but using conformational markers for the cTnC-cTnl interface (Robinson et al., 2004). The results have led to the conclusion that, in the presence of actin, both Ca^^ and strong cross-bridges (bound myosin SI in the presence of ADP) are required for full activation. Actin desensitized the Ca^^ regulatory switch to Ca^^ and produced cooperativity in the Ca^^ activation. Strongly bound cross-bridges eliminated cooperativity, but re-sensitized the system to Ca^^ (pCaso shifting to higher values by 0.36 units).

We investigated the kinetics of opening and closing of the cTnC domain with the same four preparations in stopped-flow measurements, using the change in FRET signal to monitor changes in domain conformation. The kinetics of domain opening triggered by Ca ^ binding was monitored by mixing a protein preparation with a large excess of Ca^^ under pseudo-first order conditions. As expected, the FRET signal decreased rapidly upon mixing, indicating an increase in inter-probe distance (domain opening). Domain closing was monitored by mixing a preparation fully saturated with Ca^^ with a solution of EGTA to remove bound Ca^ . The FRET signal increased upon mixing due to dissociation of bound Ca^ . All kinetic tracings showing domain opening and domain closing in the four preparations were single exponential. The rate constants determined from the different preparations are compared in Fig. 20.3. Both opening and closing rates were affected when cTn was reconstituted into different species. From the physiological point of view, the interesting results are those obtained from the regulated thin filament (cTnTmAy) and its strong complex with myosin SI. If the kinetics of domain opening is correlated with the kinetics of the movement of the regulatory region of cTnl into the open domain, the domain opening rate is expected to be dependent upon the kinetics of the transition of the downstream cTnl inhibitory region from the helix-tum-helix motif to an extended domain opening rate is expected to be slower than in the absence of bound actin. Consistent with this possibility is that the domain opening rate was 50% slower in cTnTnAy than in


400

C 300 iS m

8 ^^ i5 fit: 200

150

366 376

259

202

318

Closing Rate Opening Rate

345

cTn cTnTm cTnTnfiAy cTnTmA^SI

251

210

350 O

I c

h300 iS

400

F250

200

h150

O U

Figure 20.3. Rate constants of opening and closing of the N-domain of cTnC triggered by Ca^^ binding to and dissociation from the N-domain. The cTnC was reconstituted into different complexes.

cTnTm. In the complex cTnTmA7(Sl + ADP), weakening of the contacts between actin and cTnl inhibitory region was facilitated by the presence of myosin S1, thus allowing the conformational transition of the inhibitory region to occur with a higher rate that is comparable to that in cTnTm. The closing rate in cTnTmAy is faster than in the complex because in the reverse conformational transition the inhibitory region finds its actin site without interference from bound myosin on the actin filament. cTnl binds to cTnC in multiple sites some of which are Ca^^ independent and others are Ca^^-dependent. The Ca^^independent interaction always exists in relaxed muscle, and the Ca^^-dependent interaction involving the regulatory region and the N-domain of cTnC occurs cyclically during cycles of activation and deactivation. In a complete cycle, Ca^^ binds to the N-domain of cTnC and the regulatory region of cTnl binds to the open N-domain, followed by Ca^^ dissociation and dissociation of the regulatory region from the N-domain. An important question regarding the sequence of these reversible interactions is which interaction is rate-limiting in activation and which interaction is rate-limiting in deactivation. Additional FRET-based stopped data have suggested that in activation Ca' binding is rate-limiting, whereas in deactivation dissociation of the regulatory region of cTnl from the N-domain of cTnC is rate limiting (Dong et al., 2005).

2+


The present FRET-based equilibrium and kinetic results provide a structural basis to describe several Ca^^ triggered conformational transitions that are components of the mechanisms of activation of cardiac myofilaments. Such information is difficult to obtain by other physical techniques. Unlike X-ray crystallography or heteronuclear NMR, FRET studies are not dependent upon protein crystallization in different biochemical states and, in principle, are not limited to small proteins. FRET is particularly useful for investigations of changes in global conformations as shown in our recent works.

20.5. FRET-BASED CONSTRUCTION OF MOLECULAR MODELS

The currently available high-resolution method to construct molecular models based on solution data is the heteronuclear NMR spectroscopy. This method uses short nuclear Overhause effect (NOE) distances (less than ca. 5 A) between two nuclei as distance restraints. With the same formal approach, FRET distances also can be used as distance restraints to construct molecular models. Our initial interest was in the structure of the cTnl'cTnC complex in the critical region of the cTnl inhibitory and regulatory regions (residues 129-160) since most of our FRET studies were directed to this region. Toward this goal, we used simulated annealing methods to study the Ca^^-saturated cTnC'cTnl complex in the regions of interest. This was done by using thirty inter-probe FRET distances determined with full-length binary complexes or ternary complexes, which included the cTnT subunit (Sheldahl et al., 2003). Twenty-eight of these were intermolecular distances between cTnC and cTnl. They were from seven cTnC residues distributed from the N-domain to the C-domain (residues 35-159) to four residues in the cTnl inhibitory and regulatory regions. The four cTnl residues were among those that were used in studies of conformational transitions in those regions. The intermolecular distances were determined with single-cysteine cTnC mutants covalently attached to MIANS [2-(4'-maleimidylanilino)naphthalene-6-sulfonate] as energy donor, and single-cysteine cTnl mutants linked to DDPM as energy acceptor. These distance ranged from 21.4 to 37.0 A. In addition, we also used one intramolecular distance within cTnl (between residues 129 and 152) and one long distance within cTnC. The intramolecular distance in cTnC was between Ca^^-binding site III in the C-domain and Cys35 in the N-domain. This inter-domain distance was determined with a single-tryptophan cTnC mutant (TyrlllTrp), using bound Tb^^ at site III as the energy donor and lAATMR (5- and 6-idoactamidotetramethylrhodamine) linked to the cysteine residue as energy acceptor. The luminescence of bound Tb^^ at site III was generated through sensitization by the tryptophan located within the 12-residue metal-binding loop of site III upon irradiation at 295 nm. The sensitized Tb^^ luminescence at 545 nm was the donor signal transferred to the acceptor. This distance was determined to be 46.7 A when the cTnC was complexed with cTnl, and 47.0 A when cTnC was reconstituted into cTn (Dong et al , 2000).

A large number of random structures of the cTnl residues 129-160 was generated and used as the starting conformations for simulated annealing


molecular dynamics runs, followed by energy minimization with the FRET distance restraints (Sheldahl et al., 2003). During these steps, the FRET distances were incorporated as harmonically restrained bonds in a manner analogous to the use of NOE distance restraints derived in NMR experiments. Force constants for these bonds were determined from the half-widths of the distributions of the FRET distances, using the relationship k = RT/a^, where k is the bond force constant, RT is the thermal energy at 300 K (0.596 kcal/mol), and o is the standard deviation of the experimental FRET distance distribution and is given by o^ = half-width/2.35 (Lakowicz et al., 1988). Several of the constructed structures satisfy the experimental FRET restraints fairly well. This initial study shows possible local structures of the cTnl inhibitor and regulatory regions and their proximity relationship to a portion cTnC. Current work is being extended to cover a much larger portion of the troponin complex with many more intermolecular distances in the three pairs of binary complexes (cTnC'cTnl, cTnI»cTnT, and cTnT«TnC) as well as intramolecular distances. Studies with fifteen inter-subunit distances across cTnT and cTnl showed large Ca^^-induced reduction in fluctuation dynamics in the region between the C-terminal segment of cTnT and the inhibitory region of cTnl (Xing et al., 2005). This reduction is consistent with changes in the secondary structure of the inhibitory region and anchoring of the regulatory region in the open N-domain ofcTnC.

An advantage of this approach to construct molecular models of the structure of cardiac troponin is that FRET distance restraints can be determined with full-length proteins not only in the deactivated and activated states, but also with cTnl and cTnT in different phosphorylation states to study the structural basis of regulation of cardiac contractility by phosphorylation. The difficulties here are the need to have a large number of single-cysteine mutants for labeling with donors or acceptors, and evaluation of the labeled mutants for functional integrity. Suitable donor-acceptor pairs are needed for long inter-site distances. In preliminary studies, we used the lAEDANS-DABMI (4-dimethylaminophenylazophenyl-4'-maleimide) pair (RQ ~ 40-42 A) to determine eight inter-probe distances between cTnl residues in the inhibitory and regulatory regions and two cTnT residues in the C-terminus. These inter-probe distances range from 43.6 to 78.4 A. The Ca-Ca distances between these cTnl and cTnT residues are from 40.8 to 76.6 A derived from the crystal structure of the truncated cardiac troponin. The differences between these Ca-Ca distances and the FRET distances are from < 1 A to about 4 A. These small differences suggest that use of inter-probe FRET distances to construct molecular models is a reasonable approach. The FRET structures are of low-resolution since the FRET distances are all > 10 A. In spite of the resolution, FRET models can be compared with structures derived from other experimental studies.


20.6. NUCLEOTIDE-DEPENDENT KINESIN CONFORMATIONS

Although crystal structures were available for kinesin in one biochemical state, in which ADP occupied the active site, the structure for the kinesin*ATP state was not available. The absence of models for the ATP state made it difficult to characterize structural changes that occurred in this molecular motor through the course of its ATP cycle. As an approach to gain insight into these structural changes, we determined five inter-probe distances between single tryptophans in monomeric kinesin mutants and the bound fluorescent nucleotide analogue 2'-deoxy-mant-ADP, in the absence and presence of AIF4. The five selected residues for substitution with tryptophan were in regions of kinesin implicated in nucleotide or microtubule binding or in force transduction. The mutants had functional properties similar to those of wild type kinesin. ADP and AIF4 together bind to the active catalytic site of kinesin tightly and the conformation of the kinesin motor domain in this nucleotide-bound state is usually assumed similar to the conformation of the ATP-bound state. The five kinesin mutants were Tyr228Trp, Phe318Trp, Val329Trp, Val238Trp, and Ala260Trp. The inter-probe distances determined with the mutants in the presence of ADP only and in (ADP + AIF4) are listed in Table 4 (Xing et al., 2000).

The magnitudes of AR between the two states were 1.2-3.4 A for the first three distances, whereas the AR for the last two distances was negligible. To investigate whether the small AR values could be considered significant, we examined and compared the dependence of X R on the distances of the donor-acceptor distance distribution for each pair of distances. The X R surfaces for the six distances determined with the first three kinesin mutants were sharp. Each pair of surfaces for the ADP and ADP + AIF4 complexes intersected above the 68% cutoff level of X R (one standard deviation). We concluded that small, but statistically significant, differences in the mean inter-probe distances between the two biochemical states were detected in three locations. These results suggested movements of three residues during the strong-to-weak transition in kinesin. These observed conformational transitions predicted corresponding changes in the mant emission when monitored by FRET stopped-flow measurements. These kinetic results and the measured rate of phosphate release unequivocally corroborated the equilibrium FRET results.


Table 20.4 Inter-probe Distances Between Single Tryptophans in Kinesin and Bound Nucleotide (A)

Position of Tryptophan Residue

Bound Ligand Tyr228Trp Phe318Trp Val329Trp Val238Trp Ala260Trp

ADP 28.2 18.4 30.0 18.2 28.6

ADP+AIF4 26.0 17.2 26.6 18.8 29.2

20.7. SUMMARY

We have summarized in this review some of our recent studies of several protein systems that were based on analyses of Forster resonance energy transfer measurements, in both equilibrium and kinetic modes. The studies of cardiac muscle proteins and their assembly into myofilaments demonstrate that FRET is a power and useful technique to detect global conformational changes. Equilibrium measurements of such changes yield quantitative data on inter-probe distance changes induced by biochemical perturbations. If the total change results from more than one specific perturbation, FRET-based structural kinetics can resolve quantitatively contributions from different perturbations to the overall global conformational change. These studies have yielded considerable mechanistic insights on Ca^^-activation in cardiac myofilament, and movements of specific residues in kinesin in a transition between two biochemical states even in the absence of microtubules. We also show that FRET distances can be used as distance restraints to construct molecular models by using protocols such as simulated annealing molecular dynamics and energy minimization. Such constructed models are low-resolution since FRET distances are longer than 10 A. When inter-probe distances are determined from the distributions of distances, the mean distances are incorporated as harmonically restrained bonds during simulated annealing and energy minimization. Force constants for these bonds can be obtained for the half-widths of the distance distributions. Unlike currently available high-resolution structural methods, FRET measurements can be carried out in solution and are not restricted to small proteins. They can be carried out for complexes of proteins since intermolecular distances can be readily determined. Against these advantages, extrinsic probes are needed and substitutions of specific amino acid residues in the proteins are required for labeling with the probes.


20.8. ACKNOWLEDGMENT

We acknowledge support from the National Institutes of Health (HL52508 to HCC and HL80186 to WJD) and American Heart Association ( 03301 TON to WJD).

20.9. REFERENCES

Abbott, M. B., Dong, W.-J., Dvoretshy, A., DaGue, D., Caprioli, R. M., Cheung, H. C , and Rosevear, P. R., 2001, Modulation of cardiac troponin C-troponin I regulatory interactions by the amino-terminus of cardiac troponin I, Biochemistry 40: 5992.

Dong, W-J., Xing, J, Villain, M, Hellinger, M, Robinson, J. R., Murali, M., Solaro, R. J., Umeda, P. K., and H. C. Cheung, 1999, Conformation of the regulatory domain of cardiac muscle troponin C in its complex with cardiac troponin I, J. Biol Chem. 274:31382.

Dong, W.-J., Robinson, J. R., Xing, J., and Cheung, H. C , 2000, An interdomain distance in cardiac troponin C determined by fluorescence spectroscopy. Protein Sci. 9:280.

Dong, W.-J., Xing, J., Robinson, J. R., and Cheung, H. C , 2001, Ca ^ Induces an extended conformation of the inhibitory region of troponin I in cardiac muscle troponin,. J. MoL BioL3U:5\.

Dong, W.-J., Robinson, J. R., Stagg, S., Xing, J., and Cheung, H. C , 2003a, Ca^^-induced conformational transition in the inhibitory and regulatory regions of cardiac troponin I, J. Biol. Chem. 278:8686.

Dong, W.-J., Robinson, J. R., Xing, J., and Cheung, H. C , 2003b, Kinetics of conformational transitions in cardiac troponin induced by Ca2+ dissociation determined by Forster resonance energy transfer, J. Biol. Chem. 278:42394.

Dong, W.-J., Robinson, J. R., Lin, C.-Y., Ruzaics, B, and Cheung, H. C , 2004, Ca^^-induced opening of the N-domain of cTnC in regulated cardiac thin filament, Biophys. J. 86 (#1, pt. 2):396a.

Dong, W.-J., Robinson, J. R., Xing, J., and Cheung, H. C , 2005, FRET kinetic study of Ca ^ and cTnl induced cTnC N-domain opening, Biophys. J. 88 (#1, pt. 2): 130a.

Lakowicz, J. R., Gryczynski, I., Cheung, H. C , Wang, C.-K., Johnson, M. L., and Joshi, N., 1988, Distance distributions in proteins recovered by using frequency-domain fluorometry.

Applications to troponin I and its complex with troponin C, Biochemistry 27:9149. Robinson, J. R., Dong, W.-J., Xing, J., and Cheung, H. C , 2004, Switching of troponin I: Ca ^ and

myosin- induced activation of heart muscle, J. Mol. Biol. 240:295. Sheldahl, C, Xing, J., Dong, W.-J., Harvey, S. C , and Cheung, H. C , 2003, The calcium-saturated

cTnl/cTnC complex: structure of the inhibitory region of cTnl, Biophys. J. 84:1057. Takeda, S A.,Yamashita, Y., Maeda, K, and Maeda, Y, 2003, Structure of the core domain of

human cardiac troponin in the Ca^^-saturated form. Nature 424:35. Xing, J., Wriggers, W., Jefferson, G. M., Stein, R., Cheung, H. C, and Rosenfeld, S. S., 2000,

Kinesin has three nucleotide-dependent conformations, J. Biol. Chem. 275:35413. Xing, J., Dong, W., Robinson, J. R., and Cheung, H. C , 2005, Inter-subunit distance between

cardiac troponin T and troponin I in reconstituted cardiac troponin, Biophys. J. 88 (#1, pt. 2): 131a.

GREEN FLUORESCENT PROTEIN AS A BIOSENSOR FOR TOXIC COMPOUNDS

Renato J. Aguilera*^ Jessica Montoya*, Todd P. Primm**, Armando Varela-Ramirez*

21.1 ABSTRACT

In this brief review, we present recent results in the development of fluorescence-based assays for the detection of compounds with cytotoxic, anticancer and anti-microbial properties. As other reviews have explored various aspects related to these topics, this review will focus on the use of the Green Fluorescent Protein (GFP) for the detection of potentially toxic and/or therapeutic compounds. Since high-throughput screening of chemical compounds can be both expensive and laborious, development of low cost and efficient cell-based assays to determine biological activity should greatly enhance the early screening process. In our recent studies, we have developed a couple of GFP-based assays for the rapid screening of compounds with cytotoxic and bacteriocidal properties. As will be described in more detail in subsequent sections, a new 96-well assay has recently been developed that allows for the simultaneous screening of test compounds on gram positive and negative bacteria as well as mammalian (human cancer) cells. Our results demonstrate that both mammalian cells and bacteria can be analyzed in tandem to rapidly determine which compounds are specifically toxic to one of these cell types. The parallel screening of both eukaryotic and prokaryotic cells was found to be feasible, reproducible, and cost effective.

*Departmeiit of Biological Sciences, University of Texas at El Paso, 500 W. University Dr., El Paso, TX 79968-0519. E-mail: [email protected] ** Department of Biological Sciences, Sam Houston State University, Hunsville, TX, 77341

463

464 R. J. AGUILERA ETAL.

21.2. BRIEF OVERVIEW ON THE PROPERTIES OF GFP

Few methods exist that allow the visualization of living cells as they undergo apoptotic or necrotic modes of death. Since labeling of cells with dyes is not always efficient and may lead to interference of cellular functions, the preferred method for non-invasive detection is via the use of fluorescent markers. A fluorescent marker that has achieved prominence in visualization assays is GFP and its derivatives. These proteins have greatly facilitated the in vivo and in situ visualization of proteins involved in complex cellular functions.

GFP was first detected in the jellyfish, Aequorea victoria, in 1961 [1], but the cloning of the GFP gene did not take place until 1992 [2]. Subsequent expression of the GFP protein revealed that it retained its fluorescence properties in non-jellyfish species (for extensive reviews see [3; 4]). GFP is an exceptional protein since it does not require co-factors or substrates, is stably expressed as a fusion protein, is relatively non-toxic, and can be readily detected by fluorescence microscopy and other fluorometric techniques (see [3; 4]). The wild type GFP protein has two excitation peaks at 395nm and 470nm, and its emission peak is at 509 nm [4]. However, the most unique feature of GFP is the highly fluorescent chromophore composed of Ser-Tyr-Gly residues which are cyclized and oxidized to form the chromophore [3; 4]. Due to these unusual characteristics, this protein has been used to explore complex biochemical and cellular processes in living cells and in whole prokaryotic and eukaryotic organisms (reviewed by [3; 5]). GFP variants have also been created with shifted absorbance and emission spectra and improved folding and expression properties. The creation of blue, cyan, yellow and red GFP variants coupled with new fluorescence imaging techniques has greatly enhanced our ability to perform localization and kinetics studies of GFP-tagged proteins (see [3] for a review of these techniques). It is important to note that the majority of recent studies make use of an enhanced version of GFP (EGFP) which is not only codon optimized for mammalian expression but also contains point mutations that make it brighter and more stable [3].

21.3. GFP AS A BIOSENSOR

The use of GFP as a biosensor for genotoxic compounds has also been gaining momentum over the past few years [5-8]. A good example of these assays, is the GreenScreen genotoxicity assay that can simultaneously measure both toxicity and genotoxicity in yeast [9-11]. This assay relies on genetically modified yeast carrying a DNA damage-inducible RAD54 promoter upstream of a yeast-enhanced GFP gene that is expressed when DNA repair is induced by genotoxic agents [6; 11]. A reduction in cell proliferation was used to determine toxicity and to test environmentally relevant substances such as metal ions, solvents, and pesticides. Most recently, this assay was modified for environmental monitoring using portable instrumentation [9]. Using a similar strategy, Normal et al, generated GFP-reporters based on the inducible DNA damage SOS response of Escherichia coli [12]. Although several constructs

GFP AS A BIOSENSOR FOR TOXIC COMPOUNDS 465

were tested, the SOS-GFP biosensor was found to be highly sensitive to the detection of carcinogens and genotoxins [12]. Apart from an increase in sensitivity, as compared to other E. coli GFP-based models, this system should be readily applicable to high-throughput screening assays.

The use of GFP-expressing mammalian cells for cytotoxic compound screening was first described by Sandman, et al., [13]. In this study, the inducible Tet-On system was utilized to drive the expression of EOF? in HeLa cells to evaluate the cytotoxicity of platinum complexes [13]. Treatment of the inducible HeLa cells with cisplatin and other platinum complexes resulted in a strong correlation between GFP inhibition and cytotoxicity [13]. The observed decrease in GFP expression was mediated by transcriptional downregulation possibly due to platinum-mediated DNA crosslinking [13]. In another study using well-known cytotoxic agents, Steff, et al., [14] demonstrated that GFP-based assays yielded comparable kinetics and sensitivity to more traditional apoptosis assays. After toxicant exposure, a decrease in GFP fluorescence was detected by flow cytometry and fluorescence-based microplate assays [14]. Although the mechanism for the loss of fluorescence was not determined, cytoplasmic GFP-fluorescence was found to decrease during cell death but this decrease in signal was apparently not due to protein degradation [14]. In yet another study, the fluorescence signal of a fusion protein consisting of GFP and the nuclear pore membrane protein POM121 that targets to the nuclear membrane was also found to dissipate after induction of apoptosis in a neuroblastoma cell line [15]. In this study, the loss of GFP-signal was found to correlate with the degree of chromatin condensation and nuclear DNA fragmentation associated with apoptosis [15].

A novel murine bone marrow stromal cell line has also been established for the assessment of p53 (tumor suppressor) protein responses to genotoxic stress [16]. In this study, EGFP was used to assess the transactivation response of p53 to chemical and physical stimuli. As was expected, GFP expression was significantly enhanced upon exposure to p53 inducers and this induction correlated well with p53 protein accumulation making this model system very useful for toxicological studies. Using a similar approach, Quinones and Rainov [17], established a human cell line (HEK293-TP53::EGFP) that expresses a functionally stabilized p53 protein fused to EGFP to monitor p53 protein expression as well as its subcellular localization. In this system, DNA damaging agents caused a significant increase of intracellular p53-EGFP levels, which was dependent on the endogenous p53 status. These GFP-based reporter systems should be very useful for the identification of mutagens and carcinogens that induce p53 expression.

An interesting twist to the aforementioned assays is the use of GFP-tagged cells in cell-mediated cytotoxicity assays. Analysis of cytotoxic T-lymphocyte (CTL) or natural killer (NK) cell activity generally involves the use of radiolabeled target cells and therefore the creation of sensitive non-radioactive assays is highly desirable. A GFP-based "fluorolysis" method was developed that is significantly more sensitive than the standard ^^Cr-release assay for CTL activity [18]. In this assay, the well characterized T-cell target cell line P815 was transfected with the GFP gene and used as a target in CTL assays. Using


flow cytometry, the percentage of GFP labeled target cells was easily detected allowing for the identification of a smaller number of activated CTLs [18]. A variant of this assay has recently been employed for the detection of NK-mediated cytolysis. In this case, a GFP-transfected Wilms' tumor cell line was established that is highly susceptible to human NK cells [19]. Measurement of GFP release by NK-lyzed tumor cells was found to correlate well with radioisotopic assays and proved to be a convenient alternative for monitoring human activated NK-cell mediated killing activity [19].

21.4. GFP-BASED TOXICITY ASSAYS IN MULTICELLULAR ORGANISMS

Apart from the large variety of prokaryotic and eukaryotic cell based assays, recent breakthroughs allow GFP-detection in live multicellular organisms. For example, a chick-embryo metastatic cancer model has recently been described that makes use of Lewis lung carcinoma cells that express GFP [20]. Injection of the GFP-tagged carcinoma cells into 12th day chick embryos resulted in metastases in the brain, heart, and stemum, which was readily visualized by fluorescence after several days. Co-injection of the marked cells with therapeutic agents resulted in complete inhibition of metastases indicating that this model can be used for screening of anti-cancer/metastatic drugs [20]. Several GFP transgenic animals have also been generated that could potentially be used to screen the effects of toxic compounds in vivo. Most recently, transgenic zebrafish [21; 22], medaka [23], and nematode [24] models have been described for the detection of pollutants and other environmental genotoxic compounds. Transgenic mice expressing GFP have been used for a variety of purposes including analysis of tumor angiogenesis [25] and metastasis [26-28]. A good example of targeted expression of GFP is the noninvasive prostate imaging model in live mice [29]. In this model, transgenic mice carrying the human kallikrein 2 promoter coupled to luciferase and GFP genes, via a bi-cistronic vector, yielded animals that specifically expressed GFP in the prostate lobes [29]. In addition, several GFP-tagged cancer cells introduced into immunocompromised mice have been developed that have allowed visualization of tumor growth, metastasis and its associated vascularization [27]. These systems offer great advantages as the GFP-tagged tumor cells are readily detected in normal tissues by fluorescence microscopy without extensive sample preparation. Since the process of metastasis is not fully understood and since prevention of tumor-spread is of great importance for cancer therapy, any insights provided by these cancer-models can lead to the discovery of novel therapeutic approaches. In recent years, intravital microscopic methods and whole body imaging techniques have been developed to detect GFP-tagged tumor progression in animals [25; 26; 30-33]. In fact, real time detection of such labeled cells has been used to detect single cells (reviewed by [27]) and used to monitor the metastatic process [34; 35]. These ex vivo detection techniques are highly relevant for the screening of novel therapeutic agents and they may allow the detection of tumor cell clearance during drug treatment.


21.5. RECENT GFP-ASSAYS FOR DRUG DISCOVERY

Since high throughput screening of chemical compounds can be both expensive and laborious, development of low cost and efficient cell-based assays to determine biological activity should greatly enhance the early screening stages [36-38]. Several computational methods are currently in use to predict the potential toxicity of novel compounds but it is highly unlikely that these methods will fully replace animal or cell-based assays [39-42].

In our recent studies, we developed two GFP-based assays for the rapid screening of compounds possessing cytotoxic and bacteriocidal properties [43; 44]. In our first attempt to develop a simple cytotoxicity assay, we made use of a human cancer cell line, HeLa-GFP [45] that constitutively expresses a recombinant EGFP gene fused to histone H2B [43]. Expression of the H2B-GFP fusion protein results in nuclei that can be readily detected by fluorescence microscopy [45]. An obvious advantage of using GFP-marked cells to examine cytotoxicity is that the same cells can be analyzed over a prolonged period of time. As can be seen in Fig. 21.1, the cytotoxic effects of plumbagin (Ql), a known cytotoxic naphthoquinone, was monitored by fluorescent microscopy over a 10 hour period. Prototypical signs of apoptosis were detected by 6 hours of exposure as evidenced by nuclear DNA condensation and membrane blebbing. After 10 hours of exposure, GFP fluorescence decreased significantly, which was most evident when the cells were treated with another cytotoxic naphthoquinone, 3-Phenyl-l,4-naphthoquinone oxide; Q23, see Table 21.1 and ref. [43]). This decrease in GFP-signal is likely due to DNA degradation and the loss of nucleosomal structure. Interestingly, clouds of GFP fluorescence were occasionally detected adjacent to the GFP-depleted nuclei presumably due to release of histones/nucleosomes after inter-nucleosomal DNA cleavage

Figure 21.1. Microscopic analysis of the toxic effects of plumbagin (Ql) and 3-phenyl-l,4-naphthoquinone oxide (Q23). Using fluorescence microscopy, a group of HeLa-GFP cells was monitored for signs of cell death at two hour intervals over a 10 hour period. Obvious signs of cell-death induction were clearly visible by 6 hours (hrs) of chemical exposure by membrane "blebbing" (see filled arrows) and by an increase in nuclear DNA condensation (open arrows). Note that Q23 (see Fig. 21.3) is the same compound referred to as NQl 1 in our previous analysis [43]. (See color insert section.)


To eliminate the possibility that the H2B-GFP fiision protein contributed to the cytotoxic effects of the quinone compounds, several cell lines were also subjected to the same microscopic analysis [43]. As can be seen in Fig. 21.2A, when Ql and Q23 were tested on several adherent cell lines (NIH3T3, L-cells, L-tk-, HeLa and HeLa-GFP), very similar toxicity effects were detected. Although the morphological changes of the dying cells resembled the changes attributed to apoptosis, it is clear that naphthoquinones can cause necrosis as well (see ref. [43] and references within). In order to determine if the two compounds, Ql and Q23, induce similar modes of death, flow cytometry was utilized to help determine the percentages of cells undergoing necrosis and/or apoptosis. Annexin V-FITC staining, which detects phosphatidylserine expo sure on the surface of apoptotic cells, and Propidium iodide (PI), which stains the nuclei of dying cells, were used in conjunction to determine the modes of death induced by Ql and Q23. As shown in Fig. 21.2B, treatment of a nonadherent lymphocyte cell line with Ql and Q23 resulted in a significant induction of apoptosis (> 30%; see [43]) consistent with prior results on the action of plumbagin and other naphthoquinones [46-49].

Although the cytotoxic effects of both quinones were quite similar, Q23 treatment resulted in a slightly higher level of necrosis (-30%) than apoptosis (-25%) at the highest concentrations tested (Fig. 21.2B). As will be discussed in the next section, it is likely that some of the naphthoquinone compounds that were recently screened will exhibit different modes of action.

21.6. USING THE HELA-GFP ASSAY TO DETERMINE THE CYTOTOXICITY OF ANTIBACTERIAL COMPOUNDS

Mycobacterial infections are on the increase worldwide, particularly in immunocompromised population, as is the case of M. avium in AIDS patients. Although these infections have devastating effects, there have been no new drugs specifically developed against these organisms since the 1960's ([50] and references within). A recent screen of acetophenone derivatives revealed that some of these compounds have significant anti-mycobacterial properties [50]. Acetophenone-derivaties with the strongest antibiotic properties were subsequently tested in the HeLa-GFP cytotoxicity assay to determine if any of these compounds exhibited cytotoxic effects against the human cell line (see Fig. 21.3; [50]). This analysis uncovered compounds with selective cytotoxicity toward eukaryotic cells (see asterisks. Fig. 21.3) and compounds with selective antibacterial activities (see arrows. Fig. 21.3; [50]). Thus, compounds likely to have distinct mechanisms of action in prokaryotic and eukaryotic organisms were uncovered. As will be described in more detail in the following section, a new 96-well assay has recently been developed that allows for the simultaneous screening of test compounds on gram positive and negative bacteria as well as on mammalian cells.


DMSO

B

o

o Q.

c c <

c CO

K

8 Q.

50 -

45 -

4 0 -

35 -

3 0 -

25-

20-

15

10

5 -

0_

D Necrosis

• Late Apoptosis/Necrosis

^ Early Apoptosis

[

( DMSO 1 Control 1 (

i

H202 Control

1.0

Q1

4.0 1.0

Q23

[

4.0^g/ml

Figure 21.2. Analysis of the toxic effects of plumbagin (Ql) and Q23. (A) Cytotoxicity screening of Ql and Q23 on various adherent cell lines. Plumbagin was tested at 5.3 (1.0) and 0.21 (.04) |j,M while Q23 was tested at 4.0 and 0.16 \xM. DMSO (1 \xl), which was the compound solvent, was added as a control at the same concentration as the test samples. (B) Graphical representation of results obtained by two-color flow cytometry (see [43]) at low and high compound concentration. Note that Annexin V-FITC and PI double staining should detect late apoptosis while single PI staining should be indicative of necrosis (see [43] for more details). Early apoptosis is defined as cells that are only Annexin V positive. Compounds were tested at the same concentration as in (A) for the lower concentration and four times that amount for the higher concentration.


21.7. LARGE-SCALE SCREENING OF COMPOUNDS ON EUKARYOTIC AND PROKARYOTIC CELLS

In order to rapidly screen novel compounds for their toxic properties, a relatively simple GFP-based assay was recently developed to simultaneously screen several compounds without having to perform other elaborate assays [43]. Although easy to implement, this assay could not be applied to high-throughput analysis as it relied on microscopic visualization [43]. Given this obvious limitation, the assay was modified to facilitate the simultaneous screening of multiple compounds in a 96-well format using an automated fluorescence plate reader. Since small quantities of compounds are analyzed, this assay is particularly well suited for the screening of small combinatorial chemical libraries. Another advantage of these microplate assays is the ability to perform all assays in duplicate or triplicate to derive more reliable results. Using this assay, we recently screened thirty nine naphthoquinone derivatives with significant success (see Fig. 21.4 [44]).

100

o x o

c

(D

(1>

Q_ Q_ Q. 0 . Q. 0 . CL Q. Q_ Q. Q-Q_ Q. Q. Q_ Q_ < Q. Q_ Q. CL CL Q_ Q. Q_ Q_ CL T= < < < < < < < < < < < < < < < < m < < < C Q C Q < < < 2 : < o

^^a^^^ ^coQ-^co^if ^t CO o ^ ^ ^ c o -^CO ^tCOQ I—

Compounds

Figure 21.3. Screening of acetophenone compounds for toxicity with the HeLa-GFP assay. HeLa-GFP ceils were exposed to various acetophenone compounds for 24 hours at two concentrations, 10 and 50 \iglm\. Compounds selectively toxic to bacteria (4 Nitroacetophenone, NAP; 4 Bromoacetophenone, 4BAP; and 4-Piperidinoacetophenone, PAP) with cytotoxicities of <25% and minimal inhibitory concentrations (MIC) of < 605 are indicated with open arrows. Compounds selectively toxic to the HeLa cells (Hydroxybenzoic acid, HBA; 4-Chloroacetophenone, 4CAP; and 4-Aminoacetophenone, 4AAP) with cytotoxity >25% and MIC >9500 are marked with asterisks (Ctrl, control without compound; see [50] for additional details).


plumbagin juglone naphthazarin lapachol menadione Type I Q1 Q2 Q3 Q4 Q5 Q6

Type II Q23

rV W OH O OH O

Q24 Q25 Q26 Q27 Q28

Figure 21.4. Chemical structure of some of the naphthoquinone compounds that were recently tested in the 96-well toxicity assays (see [44] for additional details). Type I compounds are prototypical naphthoquinones like plumbagin while Type II compounds represent naphthoquinone epoxides.

The HeLa-GFP assay was subsequently modified to test the same compounds on GFP-expressing Escherichia coli and Mycobacterium avium strains in an attempt to uncover potential anti-microbial agents [44]. As shown in Table 21.1, plumbagin (Ql) was clearly the most toxic of the test compounds on both the mammalian cells and the two bacterial strains. In addition, two similar compounds, Q3 (naphthazarin) and Q5 (menadione; Fig. 21.4) were also found to be highly toxic to all cell types. Apart from plumbagin, lapachol (Q4) was found to be the most toxic to M avium as previously determined with other assays [51]. Interestingly, lapachol and similar quinones have recently been shown to exhibit strong inhibitory properties against two species of Leishmania associated to tegumentar leishmaniasis [52]. Although lapachol was found unsuitable as an anticancer drug due to its toxic side effects [53], it has recently been shown to exhibit anti-metastatic effects by inhibiting the invasiveness of cancer cells [54].

Most of the compounds tested exhibited the high levels of toxicity at higher concentrations, although some were still very toxic at a ten-fold dilution (see Ql-3 and Q23-25 at 2 |ig/ml, Table 21.1) on the HeLa-GFP cell line. Apart from differences in overall toxicity, two general types of toxic compounds were detected in these assays, those that exhibited toxicity to two or all three of the cell types (Ql-7 and Q23) and those that were primarily toxic to the HeLa cells (Q24-28; see Fig. 21.4). As these two sets of compounds apparently target different cell types, it is likely that these compounds will have different modes of action. Experiments are currently in progress to elucidate if these compounds have different molecular targets or if these differences are just a matter of membrane permeability. It is important to point out that the observed differences would not have been detected if the compounds had been tested on a single cell type.

A structural/physical analysis of the test compounds indicates that they exhibit significant differences in size, solubility (log P values) and toxicity (see Table 21.2 and [44] for additional details). For example, Type-1 compounds (see Fig. 21.4 and Table 21.1) generally exhibit high toxicity against all tested organisms, although E. coli was the most intrinsically resistant. In contrast, Type-2 compounds (see Table 21.1) were selectively toxic against the HeLa


Table 21.1. Percent GFP-fluorescence after chemical exposure

Compound

01 02 03 04 05 06 07 Q23 024 025 026 027 028

HeLa-GFP* 20ng/ml**

9.0 29.0 4.0 3.0 22.0 20.0 25.0 3.0

23.0 18.0 27.0 21.0 4.0

HeLa-GFPlE. colh-GFVMM.avium-G¥P\ 2^g/ml 20^g/inl | 20 ig/ml

13.0 14.0

0 70.0 13.0 78.0 13.0 14.0 5.0 0

26.0 21.0 17.0

1.0 111.0 41.0 56.0 58.0 105.0 89.0 56.0 87.0 118.0 110.0 117.0 115.0

7.0 72.0 38.0 26.0 69.0 55.0 35.0 66.0

202.0 133.0 97.0 133.0 97.0

^ HeLa and M avium-G¥? cells were incubated with compound for 18 hrs.

*Compounds were tested at concentrations of 20 |ag/ml or 2 |ig/ml as indicated (see ref [44]) . ^E. co/Z-GFP was incubated with compound for 10 hrs

cells, perhaps due to some unique biotransformation in these cells. This analysis revealed that molecular weight and solubility (quantified as log?, see Table 21.2) of the compounds significantly correlated with toxicity to HeLa cells, with a weaker correlation of activity against M. avium, and did not correlate with killing of E. coli (for actual statistical values see [44]). Interestingly, toxicity against the two prokaryotic organisms strongly and significantly correlated with each other, but not with the eukaryotic cell line. This suggests that the mechanism of action of quinones is different against prokaryotic and eukaryotic cells, which has been observed in other studies as well [51]. In general, M avium and E. coli were more tolerant to the quinones than the HeLa cells. This is likely due to the great permeability barrier of mycobacteria [55] and the fact that E. coli often uses efflux pumps against quinones [56]. Compounds containing a 2,3-epoxide were highly toxic to eukaryotic cells (see Q23-Q28. Table 21.1), but were relatively non-toxic to prokaryotic cells. It thus appears that the differences between the mammalian and prokaryotic cells could be attributed to membrane permeability and/or cellular quinone metabolism.

21.8. SUMMARY

The use of GFP as a biosensor has gained significant popularity over the last decade because this molecule can be utilized in a great variety of assays. Through the use of cell-based assays, this fluorescent marker has facilitated the


Table 21.2. Names and physical characteristics of quinone compounds

[ Type-1* Compounds

Ql Plumbagin

Q2 Juglone

Q3 Naphthazarin

Q4 Lapachol

Q5 Menadione

Q6

Type-2 Compounds

Q23

Q24

Q25

Q26

Q27

Q28

Compound name 2-Methyl-5-hydroxy-1,4-naphtho quinone 5-Hydroxy-1,4-naphthoquinone

5,8-Dihydroxy-1,4-naphthoquinone

2-Hydroxy-3-(3-methylbut-2-enyl)-1,4-naphthoquinone

2-Methyl-1,4-naphthoquinone

2,3-Diphenyl-1,4-naphthoquinone

MW 188.18

174.15

190.15

242.27

172.18

310.35

LogP 0.87

1.92^

1.79^

2.02

2.20^

3.99

R-group Ri=CH3j R2~H R3=OH;R4=H Ri=H; R2=H R3-OH; R4=H Ri=H; R2=H R3=OH; R4=OH R,=OH R2= CH2CH=C(CH3)2 R3—Hj R4=H 1 Ri=CH3j R2~H R3=rij R4=ri | Ri=Ph; R2=Ph R3—rlj R4—11 1

3-Phenyl-1,4-naphthoquinone oxide

3-Butyl-l,4-naphthoquinone oxide

3-Hexyl-l,4-naphthoquinone oxide

2,3-Propyl-1,4-naphthoquinone oxide 2-Ethyl-3-propyl-l,4naphthoquinone oxide 2-Ethyl-3-(ethylperoxymethyl)-1,4-naphthoquinone oxide

250.25

230.26

258.31

258.31

244.29

274.27

1.74

1.67

2.50

2.55

2.13

1.42

R,=Ph 1 R2=H R,-(CH2)3CH3 R2=H

R,=(CH2)5CH3 R2=H Ri~CH2CH2CH3 1 R2= CH2CH2CH3 R[= CH2CH3 1 R 2 - CH2CH2CH3 Ri=CH2CH3 R2= CO2CH2CH3 1

*Typesl-2 naphthoquinones are based on structures shown in Fig. 21.4. + Experimental values for n-octanol-water, 25 °C.

in vivo and in situ visualization of proteins that are involved in complex cellular functions. Although several GFP-based models have already been applied to the screening of toxic, anticancer, and antibacterial compounds, it is clear that improvements in these assays will greatly advance the field of drug discovery. As described in this review, we have developed simple and effective assays to determine the toxicity of various compounds on various cell types. Furthermore, these assays can be performed in parallel to screen for toxic compounds on both eukaryotic and prokaryotic cells in a cost effective and reproducible manner. The availability of sophisticated instrumentation and robotic systems should soon allow for the implementation of the aforementioned assays to high-throughput screening of thousands of compounds from established and/or new chemical libraries


We thank Dr. E.D. Rael and members of our research groups for critically reading this manuscript. The authors also thank Drs. G.M. Wahl and T. Kanda for the generous gift of the HeLa-GFP cell line. This work was supported by

474 R. J. AGUILERA ETAL,

MBRS-SCORE (S06 GM8012-34) subproject grant to R.J.A and an institutional NIH RCMI grant 2G12RR08124. T.P.P. was funded by NIH K22 grant AI01812-02 and a grant from the Paso del Norte Health Foundation. J.M. was supported by the RISE (R25GM069621-04) and MARC*USTAR (5T34GM008048) Training Programs at UTEP.

21.10. REFERENCES

1. O.Shimomura, The discovery of aequorin and green fluorescent protein, J.Microsc. 217, (2005) 1-15.

2. D.C.Prasher, V.K.Eckenrode, W.W.Ward, F.G.Prendergast, M.J.Cormier, Primary structure of the Aequorea victoria green-fluorescent protein, Gene 111, (1992) 229-233.

3. J.Lippincott-Schwartz, G.H.Patterson, Development and use of fluorescent protein markers in living cells, Science 300, (2003) 87-91.

4. D.C.Prasher, Using GFP to see the light, Trends Genet. 11, (1995) 320-323. 5. J.C.March, G.Rao, W.E.Bentley, Biotechnological applications of green fluorescent protein,

Appl.Microbiol.Biotechnol. 62, (2003) 303-315. 6. A.W.Knight, Using yeast to shed light on DNA damaging toxins and irradiation. Analyst 129,

(2004) 866-869. 7. O.Griesbeck, Fluorescent proteins as sensors for cellular functions, Curr.Opin. Neurobiol. 14,

(2004)636-641. 8. S.J.Rosochacki, M.Matejczyk, Green fluorescent protein as a molecular marker in

microbiology, Acta Microbiol.Pol. 51, (2002) 205-216. 9. A.W.Knight, P.O.Keenan, N.J.Goddard, P.R.Fielden, R.M.Walmsley, A yeast-based

cytotoxicity and genotoxicity assay for environmental monitoring using novel portable instrumentation, J.Environ.Monit. 6, (2004) 71-79.

10. A.W.Knight, N.J.Goddard, N.Billinton, P.A.Cahill, R.M.Walmsley, Fluorescence polarization discriminates green fluorescent protein from interfering autolluorescence in a microplate assay for genotoxicity, J.Biochem.Biophys. Methods 51, (2002) 165-177.

11. P.A.Cahill, A.W.Knight, N.Billinton, M.G.Barker, L.Walsh, P.O.Keenan, C.V. Williams, D.J.Tweats, R.M.Walmsley, The GreenScreen genotoxicity assay: a screening validation programme, Mutagenesis 19, (2004) 105-119.

12. A.Norman, H.L.Hestbjerg, S.J.Sorensen, Construction of a ColD cda promoter-based SOS-green fluorescent protein whole-cell biosensor with higher sensitivity toward genotoxic compounds than constructs based on recA, umuDC, or sulA promoters, Appl.Environ.Microbiol. 71, (2005) 2338-2346.

13. K.E.Sandman, S.S.Maria, G.Zlokamik, S.J.Lippard, Rapid fluorescence-based reporter-gene assays to evaluate the cytotoxicity and antitumor drug potential of platinum complexes, Chem.Biol. 6,(1999)541-551.

14. A.M.Steff, M.Fortin, C.Arguin, P.Hugo, Detection of a decrease in green fluorescent protein fluorescence for the monitoring of cell death: an assay amenable to high-throughput screening technologies, Cytometry 45, (2001) 237-243.

15. M.Beckman, M.Kihlmark, K.Iverfeldt, E.Hallberg, Degradation of GFP-labelled POM121, a non-invasive sensor of nuclear apoptosis, precedes clustering of nuclear pores and extemalisationof phosphatidylserine, Apoptosis. 9, (2004) 363-368.

16. N.V.Gorbunov, J.E.Morris, J.S.Greenberger, B.D.Thrall, Establishment of a novel clonal murine bone marrow stromal cell line for assessment of p53 responses to genotoxic stress, Toxicology 179, (2002) 257-266.

17. A.Quinones, N.G.Rainov, Identification of genotoxic stress in human cells by fluorescent monitoring of p53 expression, Mutat.Res. 494, (2001) 73-85.

18. N.Kienzle, S.Olver, K.Buttigieg, A.Kelso, The fluorolysis assay, a highly sensitive method for measuring the cytolytic activity of T cells at very low numbers, J.Immunol.Methods 267, (2002)99-108.

19. H.Harada, K.Saijo, I.Ishiwata, T.Ohno, A GFP-transfected HFWT cell line, GHINK-1, as a novel target for non-RI activated natural killer cytotoxicity assay. Hum.Cell 17, (2004) 43-48.


20. V.Bobek, J.Plachy, D.Pinterova, K.Kolostova, M.Boubelik, PJiang, M.Yang, R.M.Hoffman, Development of a green fluorescent protein metastatic-cancer chick-embryo drug-screen model, Clin.Exp.Metastasis 21, (2004) 347-352.

21. P.H.Krone, S.R.Blechinger, T.G.Evans, J.A.Ryan, E.J.Noonan, L.E.Hightower, Use offish liver PLHC-1 cells and zebrafish embryos in cytotoxicity assays. Methods 35, (2005) 176-187.

22. S.R.Blechinger, J.T.Warren, Jr., J.Y.Kuwada, P.H.Krone, Developmental toxicology of cadmium in living embryos of a stable transgenic zebrafish line, Environ.Health Perspect. 110, (2002) 1041-1046.

23. K.Kurauchi, Y.Nakaguchi, M.Tsutsumi, H.Hori, R.Kurihara, S.Hashimoto, R.Ohnuma, Y.Yamamoto, S.Matsuoka, S.Kawai, T.Hirata, M.Kinoshita, In vivo visual reporter system for detection of estrogen-like substances by transgenic medaka, Environ.Sci.Technol. 39, (2005) 2762-2768.

24. A.L.Graves, W.A.Boyd, P.L.Williams, Using transgenic Caenorhabditis elegans in soil toxicity testing, Arch.Environ.Contam Toxicol. 48, (2005) 490-494.

25. M.Yang, L.Li, P.Jiang, A.R.Moossa, S.Penman, R.M.Hoffman, Dual-color fluorescence imaging distinguishes tumor cells from induced host angiogenic vessels and stromal cells, Proc.Natl.Acad.Sci.U.S.A 100, (2003) 14259-14262.

26. M.Yang, E.Baranov, J.W.Wang, P.Jiang, X.Wang, F.X.Sun, M.Bouvet, A.R.Moossa, S.Penman, R.M.Hoffman, Direct external imaging of nascent cancer, tumor progression, angiogenesis, and metastasis on internal organs in the fluorescent orthotopic model, Proc.Natl.Acad.Sci.U.S.A 99, (2002) 3824-3829.

27. S.Paris, R.Sesboue, Metastasis models: the green fluorescent revolution? Carcinogenesis 25, (2004)2285-2292.

28. R.M.Hoffman, Imaging tumor angiogenesis with fluorescent proteins, APMIS 112, (2004) 441-449.

29. X.Xie, Z.Luo, K.M.Slawin, D.M.Spencer, The EZC-prostate model: noninvasive prostate imaging in living mice, Mol.Endocrinol. 18, (2004) 722-732.

30. K.Yamauchi, M.Yang, P.Jiang, N.Yamamoto, M.Xu, Y.Amoh, K.Tsuji, M.Bouvet, H.Tsuchiya, K.Tomita, A.R.Moossa, R.M.Hoffman, Real-time in vivo dual-color imaging of intracapillary cancer cell and nucleus deformation and migration, Cancer Res. 65, (2005) 4246-4252.

31. R.M.Hoffman, In vivo imaging with fluorescent proteins: the new cell biology. Acta Histochem. 106, (2004) 77-87.

32. G.Choy, P.Choyke, S.K.Libutti, Current advances in molecular imaging: noninvasive in vivo bioluminescent and fluorescent optical imaging in cancer research, Mol.Imaging 2, (2003) 303-312.

33. T.Umeoka, T.Kawashima, S.Kagawa, F.Teraishi, M.Taki, M.Nishizaki, S.Kyo, K.Nagai, Y.Urata, N.Tanaka, T.Fujiwara, Visualization of intrathoracically disseminated solid tumors in mice with optical imaging by telomerase-specific amplification of a transferred green fluorescent protein gene, Cancer Res. 64, (2004) 6259-6265.

34. J.Condeelis, R.Singer, J.E.Segall, The Great Escape: When cancer cells hijack the genes for chemotaxis and motility, Annu.Rev.Cell Dev.Biol. (2005).

35. E.Sahai, J.Wyckoff, U.Philippar, J.E.Segall, F.Gertler, J.Condeelis, Simultaneous imaging of GFP, CFP and collagen in tumors in vivo using multiphoton microscopy, BMC.Biotechnol. 5, (2005) 14.

36. D.B.Kassel, Applications of high-throughput ADME in drug discovery, Curr.Opin.Chem.Biol. 8, (2004) 339-345.

37. J.R.Kenseth, S.J.Coldiron, High-throughput characterization and quality control of small-molecule combinatorial libraries,Curr.Opin.Chem.Biol. 8, (2004) 418-423.

38. L.Zemanova, A.Schenk, M.J.Valler, G.U.Nienhaus, R.Heilker, Confocal optics microscopy for biochemical and cellular high-throughput screening, Drug Discov.Today 8, (2003) 1085-1093.

39. W.H.van de, E.Gifford, ADMET in silico modelling: towards prediction paradise? Nat.Rev.Drug Discov. 2, (2003) 192-204.

40. M.T.Cronin, Prediction of drug toxicity Farmaco 56, (2001) 149-151. 41. K.S.Lam, Application of combinatorial library methods in cancer research and drug discovery.

Anticancer Drug Des 12,(1997) 145-167.


42. M.D.Burke, S.L.Schreiber, A planning strategy for diversity-oriented synthesis, Angew.Chem.Int.Ed Engl. 43, (2004) 46-58.

43. J.Montoya, A.Varela-Ramirez, A.Estrada, L.E.Martinez, K.Garza, R.J.Aguilera, A fluorescence-based rapid screening assay for cytotoxic compounds, Biochem.Biophys.Res.Commun. 325, (2004) 1517-1523.

44. J.Montoya, A.Varela-Ramirez, M.Shanmugasundram, L.E.Martinez, T.P.Primm, R.J.Aguilera, Tandem screening of toxic compounds on GFP-labeled bacteria and cancer cells in microtiter plates, Biochem.Biophys.Res.Commun. (2005; in press).

45. T.Kanda, K.F.Sullivan, G.M.Wahl, Histone-GFP fusion protein enables sensitive analysis of chromosome dynamics in living mammalian cells, Curr.Biol. 8, (1998) 377-385.

46. J.J.Inbaraj, C.F.Chignell, Cytotoxic action of juglone and plumbagin: a mechanistic study using HaCaT keratinocytes, Chem.Res.Toxicol. 17, (2004) 55-62.

47. D.W.Lamson, S.M.Plaza, The anticancer effects of vitamin K, Altem.Med.Rev. 8, (2003) 303-318.

48. M.Krishnaswamy, K.K.Purushothaman, Plumbagin: A study of its anticancer, antibacterial & antifungal properties, Indian J.Exp.Biol. 18, (1980) 876-877.

49. M.Kaminski, M.Karbowski, Y.Miyazaki, J.Kedzior, J.H.Spodnik, A.Gil, M.Wozniak, T.Wakabayashi, Co-existence of apoptotic and necrotic features within one single cell as a result of menadione treatment. Folia Morphol.(Warsz.) 61, (2002) 217-220.

50. L.Rajabi, C.Courreges, J.Montoya, R.J.Aguilera, T.P.Primm, Acetophenones with selective antimycobacterial activity, Lett.Appl.Microbiol. 40, (2005) 212-217.

51. T.Tran, E.Saheba, A.V.Arcerio, V.Chavez, Q.Y.Li, L.E.Martinez, T.P.Primm, Quinones as antimycobacterial agents, Bioorg.Med.Chem. 12, (2004) 4809-4813.

52. N.M.Lima, C.S.Correia, L.L.Leon, G.M.Machado, M.F.Madeira, A.E.Santana, M.O.Goulart, Antileishmanial activity of lapachol analogues, Mem.Inst.Oswaldo Cruz 99, (2004) 757-761.

53. J.B.Block, A.A.Serpick, W.Miller, P.H.Wiemik, Early clinical studies with lapachol (NSC-11905), Cancer Chemother.Rep.2 4, (1974) 27-28.

54. I.T.Balassiano, S.A.De Paulo, S.N.Henriques, M.C.Cabral, M. da Gloria da Costa Carvalho, Demonstration of the lapachol as a potential drug for reducing cancer metastasis, Oncol.Rep. 13,(2005)329-333.

55. H.Nikaido, Preventing drug access to targets: cell surface permeability barriers and active efflux in bacteria, Semin.Cell Dev.Biol. 12, (2001) 215-223.

56. G.Tegos, F.R.Stermitz, O.Lomovskaya, K.Lewis, Multidrug pump inhibitors uncover remarkable activity of plant antimicrobials, Antimicrob. Agents Chemother. 46, (2002) 3133-3141.

MULTIDIMENSIONAL FLUORESCENCE IMAGING APPLIED TO BIOLOGICAL TISSUE

Daniel S. Elson, Neil Galletly, Clifford Talbot, Jose Requejo-Isidro, James McGinty, Christopher Dunsby, Peter M. P. Lanigan, Ian Munro, Richard K. P. Benninger, Pieter de Beule, Egidijus Auksorius, Laszlo Hegyi, Ann Sandison, Andrew Wallace, Pat Soutter, Mark A. A. Neil, John Lever, Gordon W. Stamp, and Paul M. W. French*

22.1. INTRODUCTION

This chapter is intended to present our work at Imperial College London investigating the potential of intrinsic tissue autofluorescence using fluorescence lifetime imaging (FLIM) and related techniques to obtain useful label-free contrast for clinical and biomedical research applications. We review the development of FLIM instrumentation and our related imaging technology that resolves fluorescence lifetime and other spectroscopic parameters, including excitation and emission wavelength, in a single data acquisition. This technology facilitates a multi-dimensional fluorescence imaging (MDFI) approach that aims to obtain as much useful information as possible from a sample in the shortest possible time. Ultimately, the ambition of this work is to discover whether autofluorescence signals can be used to provide useful contrast in vivo and in vitro and to develop appropriate instrumentation for clinical applications.

Fluorescence provides a powerful means of achieving optical molecular contrast in a wide range of instruments including cuvette-based systems, microscopes, endoscopes and multiwell plate readers. Fluorescent molecules (fluorophores) can be used as "labels" to tag specific molecules of interest or the fluorescence properties of the target molecules themselves may be exploited

Imperial College London, London SW7 2BZ, U.K. 477

478 B.S.ELSON ETAL.

to provide label-free contrast. As well as providing information about the properties of the fluorophores the fluorescence process can also be extremely sensitive to the local environment surrounding the fluorophore, providing a sensing function. In principle different species of fluorophores may be characterised according to their excitation and emission spectra, their quantum efficiency, their polarisation response and their fluorescence lifetime. For example, the quantum yield is the efficiency of the fluorescence process (defined as the ratio of the number of fluorescent photons emitted compared to the number of excitation photons absorbed) and can change as a function of changes in the local viscosity, temperature, refractive index, pH, calcium and oxygen concentration, electric field, etc. Spectrally resolved and time-resolved measurements can also be sensitive to variations in the local fluorophore environment that change the molecular electronic configuration and time-resolved measurements provide a sensitive means to detect changes in de-excitation pathways via their impact on the fluorescence decay rate. Further, time-resolved polarisation measurements can determine the rotational correlation time - or molecular tumbling time - which can report ligand binding or local solvent properties and steady state polarisation resolved measurements (of polarisation anisotropy) can report on the fluorophore orientation.

For many applications including tissue imaging, however, it is not possible to determine the quantum efficiency since this requires knowledge of the photon excitation and detection efficiencies, as well as the concentration of the fluorophores, and any quantitative measurements of intensity are hindered by optical scattering, internal re-absorption of fluorescence (inner filter effect) and background fluorescence from other fluorophores present in a sample. More robust measurements can be made using ratiometric techniques. In the spectral domain, wavelength ratiometric imaging works on the assumption that most of these unknown quantities will be approximately the same in two or more spectral windows and may be effectively "cancelled out". Such spectrally resolved images have been demonstrated to provide useful contrast between e.g. malignant and normal tissue. ^ In the time domain, fluorescence lifetime imaging is effectively also a ratiometric technique in that the various unknown quantities do not change significantly during the fluorescence decay time (typically ns).

In practice, while single channel excitation, emission, lifetime and polarisation-based spectroscopy is widely undertaken in cuvette based instruments such as spectrophotometers and spectrofluorometers, only a fraction of the potential information available from fluorescence is typically exploited for imaging applications. This is partly due to instrumentation limitations - particularly with respect to the availability of suitable light sources for excitation and lifetime spectroscopy - and partly due to signal ambiguity and noise considerations associated with the (relatively weak) fluorescence signals available from biological samples.

The last decade, however, has seen tremendous changes in imaging tools and technology that are dramatically changing the microscopy landscape and user aspirations. A key catalyst has been the wide-spread uptake of multi-photon microscopy ^ implemented with the widely tunable femtosecond

MULTIDIMENSIONAL FLUORESCENCE IMAGING APPLIED TO BIOLOGICAL TISSUE 479

Ti:Sapphire laser. Such instruments confer the ability to excite most of the commonly used fluorophores with a single excitation laser - and such systems are now available under computer control to permit automated excitation spectroscopy. ^ The proliferation of multiphoton microscopes with their ultrafast pump lasers has in turn stimulated the uptake of fluorescence lifetime imaging (FLIM)" as a relatively straightforward and inexpensive "add-on" that provides significant new spectroscopic functionality - requiring only an appropriate detector and external electronic components to record the fluorescence lifetime data. In parallel, the explosive impact of green fluorescent protein (GFP)^'^ has created a myriad of imaging opportunities to probe molecular biology that has only begun to be realised with current technology. These opportunities are in turn driving the adoption and development of more sophisticated fluorescence imaging technology. One example is the widespread deployment of Forster resonant energy transfer (FRET)^ techniques. These can facilitate a so-called ''spectroscopic ruler"^ that can determine when fluorophores are located within ~ 10 nm of each other, to provide the ultimate in colocalisation and the possibility to "image" biochemical processes such as ligand binding in cells. This is an immensely powerful approach when combined with genetically-expressed fluorophores but is proving not so straightforward to implement with conventional intensity-based imaging techniques, where a number of correction calculations must be performed.^ Increasingly FLIM and other spectroscopic techniques are being applied to improve the reliability of FRET experiments, e.g.^ ' \

For medical applications, there is an increasing interest in exploiting tissue autofluorescence for non-invasive clinical diagnosis and research. Spectrally-resolved imaging of autofluorescence is relatively well established, e.g. , and FLIM is now being actively investigated as a means of obtaining or enhancing intrinsic autofluorescence contrast in tissues. ' ' To date, however, the limited range of excitation sources for tissue autofluorescence and the complexity of imaging technology have restricted this work to relatively few laboratories and the literature concerning useful lifetime contrast of endogenous fluorophores is sparse. There is therefore a requirement to develop systems with flexible fluorescence excitation, emission and lifetime resolution in order to rapidly and efficiently survey the very large number of clinically interesting tissue types that are often uncharacterised. The line-scanning hyperspectral FLIM microscope system discussed in section 22.0 is one approach that addresses this requirement.

In general, when maximising the contrast between different fluorophores or states of fluorophores for diagnostic or other purposes, it is naiVe to restrict a measurement to one spectral dimension, e.g. emission spectrum or lifetime, since it may often be possible to achieve a higher contrast (higher information content) in a measurement spanning two or more dimensions, e.g. exploring the excitation-emission matrix or applying spectrally-resolved lifetime measurements. Applied to imaging, we believe MDFI (multidimensional fluorescence imaging) adds significant value to microscopy, endoscopy and assay technology. A vital consideration, however, is that a fluorescence sample will only emit a limited number of fluorescence photons before the onset of

480 D. S.ELSONfr^/:.

photobleaching or damage. There is therefore a limited photon ''budget'* that must be "spent" carefully to maximise the information that can be obtained from a sample. This photon budget may be reduced for many biological experiments, particular those undertaken in vivo, that require imaging on a certain time scale or that can only tolerate limited excitation irradiance. Signal-to-noise ratios will be reduced if the photon budget is allocated over more measurement dimensions and so it is imperative to maximise the photon efficiency and to judiciously sample the measurement dimensions such that the useful information/photon is also maximised. In practice this can require consideration of how precisely to resolve the fluorescence signal with respect to lifetime, wavelength and spatial dimensions. Such consideration must be informed by prior knowledge and by what the investigator wants to learn from the sample.

In this chapter we initially discuss the principles of FLIM and then describe how it may be used to provide intrinsic image contrast in biological tissue, providing a review of the properties of pertinent tissue fluorophores. We then briefly review the time- and frequency-domain approaches to FLIM and discuss various methods for analysing fluorescence lifetime data. Appropriate fitting methods are vital to obtain useful information and should inform the development and optimisation of the FLIM instrumentation. We address fitting different models to complex fluorescence decay profiles including the stretched exponential decay model.

We then summarise experimental results, starting with single-point or averaged fluorescence lifetime data from endogenous fluorophores. Extending this to imaging tissue, we present FLIM microscopy data of various, mostly human, tissue samples including normal and diseased cartilage cervical and pancreatic tissue and discuss the effect of tissue fixation. We then describe the instrumentation being developed for in vivo FLIM, addressing the issues of high-speed (real-time) imaging and endoscopic imaging.

Finally there is a discussion of novel MDFI instrumentation that is in development. This includes an ultrafast broadband supercontinuum laser source that can be continuously (electronically) tuned throughout the visible spectrum for excitation spectroscopy. We also describe systems for one and two-photon excited hyperspectral FLIM to fully image the spectro-temporal properties of the sample.

22.2. FLUORESCENCE LIFETIME

The fluorescence lifetime is the average time a fluorophore takes to radiatively decay after having been excited from its ground energy level. In biological tissue fluorescence lifetime provides a contrast parameter for endogenous fluorophores and therefore has the potential to discriminate between healthy and diseased tissue.


After a molecule has been optically excited into an upper energy level, it may decay back to the ground state either radiatively by emitting a photon, or non-radiatively without emitting a photon (see figure 22.1). Depending on the type of molecule and on its environment, these processes have a different probability of occurring and it is useful to characterise their likelihoods using the rate constants kr and knr- The fluorescence lifetime is the average time the molecule spends in an upper energy level before returning to the ground state and can be expressed in terms of k and knr as follows:

T = -1

k^ +k (1)

The parameter knr contains contributions from a number of non-radiative processes including quenching, energy transfer, and solvent interactions. To generalise, the radiative decay constant depends mainly on the molecular energy structure and the refractive index of the surrounding medium, ' ^ whereas the non-radiative decay constant also has a contribution from the immediate environment. Therefore fluorescence lifetime can report on different molecular species (fluorophores) present in the sample and on variations in the local fluorophore environment.

In addition there are also a number of other "photobleaching" processes that may deplete the upper state population, which can result from physical or chemical changes to the molecule, often leading to irreversible photodamage and sometimes associated with biochemical damage to the system. ^ One significant mechanism is the formation of singlet oxygen, a radical which is phototoxic and can lead to photobleaching and biological photodamage.

Intensity

l=lue<t^%const

Time

(b) Figure 22.1. (a) simplified molecular energy level diagram highlighting the spectro-temporal properties of fluorescence and (b) single exponential fluorescence decay plot.

482 D.S.EhSON ETAL.

After instantaneous optical excitation of a large ensemble of identical molecules, or a large number of excitations of the same molecule, the fluorescence intensity will exhibit a monoexponential decay that may be described as:

I(t) = 1^6 ^ + const. (2)

where IQ is intensity of the fluorescence immediately upon excitation (at time zero) and the constant term represents any background signal. In practice, interactions with the environment can easily cause the fluorescence decay to contain multiple exponential decay components, and many molecules exhibit decays containing multiple exponential components caused by different possible de-excitation pathways. Such complex fluorescence decays may then be fitted using an N-component multi-exponential decay equation:

7(0 = ^C.e "' + const. (3)

where each pre-exponential amplitude is represented by the value Ci. This treatment is also applicable to the case where several fluorophore species are present, as is often the case for autofluorescence of biological tissue. As has been previously demonstrated, however, it is important to acquire data with a high signal to noise ratio in order to accurately analyse such decays ^ ' ^ and this can be difficult to achieve experimentally in an imaging application due to the large number of photons that must be collected. Analysis of complex fluorescence decay profiles is therefore much better suited to single-point cuvette measurements and relatively little FLIM incorporating complex decay analysis has been reported - indeed, for many applications of FLIM to biological samples, the signal to noise ratio is barely sufficient to accurately fit single exponential decay profiles and techniques such as spatial averaging or global analysis ^ are often used to address this issue. For complex decay profiles the problem is much simplified if some a priori knowledge of the pre-exponential factors or lifetimes can be obtained.

22.2.1. Fluorescence lifetime of endogenous fluorophores

To date investigations of tissue autofluorescence lifetime spectroscopy have mainly been single point measurements that have showed the potential of lifetime as a useful contrast parameter. ' ' ^ There are numerous fluorophores found in tissue, mostly exhibiting fluorescence lifetimes in the range of 100 ps to 10 ns, ^ although most are either present in miniscule quantities or have low quantum yields.^^ The main tissue fluorophores of interest can generally be excited in the UV, including the aromatic amino acids, tryptophan, tyrosine and


phenylalanine with fluorescence lifetimes of 3.1, 3.6 and 6.8 ns respectively.^ ' ^ The structural proteins collagen and elastin can be excited with UV and also visible radiation, with the autofluorescence being attributed to the cross-links, ' " with reported fluorescence lifetimes of 5.3 and 2.3 ns respectively.^^ There are, however, many different types of these structural proteins with broad or multi-peaked, absorption spectra^^ that make quantitative comparisons with figures quoted in the literature rather unreliable: see for example the work by Ashjian et al. where different relative expressions of collagen were contrasted through the fluorescence lifetime after excitation at 337 nm, ^ or the work by Tinker on fluorescence from different types of elastin. "^ Cross-linkages in collagen can either be formed during development to produce the autofluorescent enzymatic cross-links hydroxylysyl pyridinoline and lysyl pyridinoline,^^ or alternatively cross-linking can be age related via the process of glycation.^^ The origin of elastin fluorescence is less well understood, but Thomhill showed that the fluorescence is isolated to cross-linked portions of the protein " and Deyl et al. attribute the fluorescence to a cross-link, pyridinoline, that is also a collagen cross-link. ^ Other, as yet uncharacterised, elastin fluorescence has also been discovered, however, that emits yellow^^ or blue fluorescence,^^ which can not be attributed to pyridinoline.

The metabolic cofactors NADH (nicotinamide adenine dinucleotide) and flavins FAD (flavin adenine dinucleotide) and FMN (flavin mononucleotide) are major sources of cellular fluorescence. NADH is an electron transporter in respiration and locates mainly in the mitochondria and cytoplasm. Its lifetime increases from 0.4 ns to 5 ns when it binds to a protein, and there is also an increase in quantum yield. The fluorescence properties of NADH are almost indistinguishable from those of NADPH (nicotinamide adenine dinucleotide phosphate), which is involved in cellular reduction processes.^^ FAD and FMN have lifetimes of 4.7 and 2.3 ns, and may be strongly quenched after binding with a protein^^ and both have absorption/emission maxima at 450/525 nm.' '* Since NADH is only fluorescent in its reduced form and flavins are fluorescent in oxidised form, it is possible to probe the metabolic state of a tissue using a wavelength ratiometric autofluorescence technique.^^

The final important group of endogenous fluorophores are porphyrins, which emit a strong red fluorescence^^ specifically protoporphyrin, which has a lifetime of 5.2 ns. ^ It has long been realised that necrotic tumours exhibit red fluorescence and that protoporphyrins fluorescence could be a useful diagnostic fluorescence signal.

22.3. FLUORESCENCE LIFETIME DETERMINATION

In general, fluorescence lifetime measurement and fluorescence lifetime imaging techniques are categorized as time-domain or frequency-domain techniques, according to whether the instrumentation measures the fluorescence signal as a function of time delay following pulsed excitation or whether the

484 D.S.ELSON^r^i.

lifetime information is derived from measurements of phase difference between a sinusoidally modulated excitation signal and the resulting sinusoidally modulated fluorescence signal. In principle the two approaches provide equivalent information but specific implementations result in different tradeoffs in performance according to the target applications. Historically frequency-domain methods were initially preferred due to the simpler electronic instrumentation and excitation source requirements, but time-gated detection and time-correlated single-photon counting techniques were developed more or less in parallel, and today time and frequency domains are both widely applied. References '" ^ provide excellent reviews of the various techniques and in this section we will briefly outline how the main time- and frequency-domain methods are applied to single-point fluorescence lifetime measurements and to fluorescence lifetime imaging. Later we will describe recent advances in the wide-field time-gated imaging technique that we are developing at Imperial College London for tissue imaging.

22.3.1. Single-point measurement of fluorescence lifetime

22.3.1.1. Frequency-domain lifetime measurements

The principle of frequency-domain fluorescence detection is that when the sample is excited with modulated light, the fluorescence emission is also modulated but with a different modulation depth and with a phase delay relative to the excitation signal. This is illustrated in figure 22.2 which shows how the two pertinent parameters m (the relative modulation) and (j) (the phase delay) can be used to calculate the fluorescence lifetime.^^ These parameters can be determined using appropriate electronic circuits (e.g. using a lock-in amplifier) or optically using a modulator. If the excitation light takes the form of a pure sinusoid, then m and (|) can be calculated by modulating the fluorescence at the detector at the same frequency (homodyne detection) and recording the intensity at different detector phase angles (the minimum number of angles being three to fully reconstruct the fluorescence sinusoid). Altematively the fluorescence can be modulated at a slightly different frequency to the excitation frequency (heterodyne detection), in which case the fluorescence modulation can be obtained from the time-varying signal output from the detector.

For a perfect single exponential decay, the lifetime can be calculated from m or (j) using the following equations:

1 r =—tan^zJ (4)


Phase shift tan (|) = tOTp

m=[l+co2T^2-|-i/2

Demodulation factor

te m = b/B a/A

Figure 22.2. Schematic of principle of fluorescence lifetime determination in the frequency-domain.

co\m (5)

If the fluorescence decay is not a single exponential, however, the lifetimes calculated from these equations will not be the same and the experiment must be repeated at different excitation modulation frequencies to build up a full (multi-exponential) picture. This is usually calculated by fitting the results to a set of dispersion relationships, " ' ^ although the increased acquisition and data processing time can be disadvantageous. For simplicity it is possible to use only one modulation frequency and calculate an average fluorescence lifetime by taking the average of the two lifetimes found from Eqs. (4)-(5).

The ability to use sinusoidally modulated diode lasers or LED's as excitation sources makes this technique attractive for simple low-cost applications, for which frequencies of less than 100 MHz are sufficient to provide ns lifetime resolution. GHz frequencies are required to provide ps lifetime resolution, however, and such high frequency modulation of lasers can introduce significant complexity and expense. Instead, one can exploit the high harmonic content of ultrafast mode-locked laser output radiation for excitation in frequency domain FLIM imaging systems, as has been demonstrated with multiphoton microscopy.^^

22.3.1.2. Time-domain lifetime measurements

Single-point time-domain measurement of fluorescence decay profiles can be undertaken using a range of different instrumentation. Fast (GHz bandwidth) sampling oscilloscopes and streak cameras have been used in conjunction with

486 D.S.ELSON^r^x.

ultrashort pulse lasers for decades but recently there has been increasing uptake of photon counting techniques that build up histograms of the decay profiles and are well suited to record weak fluorescence signals. Probably the current most widely used technique is time-correlated single-photon counting (TCSPC), which is a well established technique for single-point cuvette type experiments,^^ and has enjoyed significant popularity with confocal^^ and two-photon fluorescence scanning microscopes, see for example commercial systems developed by Becker and Hickl GmbH^^ or PicoQuant GmbH," ^ whose websites provide much useful information. The principle of TCSPC is that at low fluorescence fluxes a histogram of photon arrival times can be built up by recording a series of voltage signals that depend on the arrival time of individual detected photons relative to the excitation pulse. By also recording spatial information from the scanning electronics of a confocal/multiphoton microscope, fluorescence lifetime images may be acquired. This is a straightforward way to implement FLIM on a scanning microscope since it only requires adding electronic components after the detector and may be "bolted on" to almost any system. Its main perceived drawback is a relatively low acquisition rate, owing to the requirement to operate at sufficiently low incident fluorescence intensity levels to ensure single photon detection at a rate limited by the "dead-time" between measurement events, which is associated with the time-to-amplitude (TAC) and constant fraction discriminator (CFD) circuitry that determine the photon arrival times. However, for modem TCSPC instrumentation this limit is usually less significant than problems caused by "classical" photon pile up, which limit the maximum count rate to approximately 5 % of the repetition rate of the laser.

An alternative single-point photon-counting time-domain technique is based on temporal photon-binning, for which the arrival times of photons are swept into one of a number of different time-bins and a histogram is built up accordingly."*^ This method does not have the same dead-time and pulse-pile-up limitations as TCSPC and so may be used with higher photon fluxes to provide higher imaging rates when implemented on a scanning fluorescence microscope. To date it has not been implemented with the same precision as TCSPC and the current commercial implementation from Nikon ^'^ provides a faster measurement capability that can provide FLIM at rates approaching real time but with lower lifetime precision.'*^ Of course the temporal resolution of fluorescence lifetime instrumentation is also limited by the temporal impulse response of the detector, as well as the electronic circuitry. Photon counting photomultipliers typically exhibit a response time of ~ 200 ps although faster multichannel plate (MCP) devices can have response times of a few lO's of ps. One significant exception to the above observation is the pump-probe approach where a second (probe) beam, which is delayed with respect to the excitation (pump) beam, interrogates the upper state population. A particularly elegant implementation of this technique for scanning fluorescence microscopy also provides optical sectioning in a manner analogous to two photon microscopy." "


Frequency-domain (heterodyne detection)

m. Time-binning

Time-correlated single photon counting (TCSPC) iiiiiiiiiiii

ki. Time-domain direct measurement

I e.g. streak camera with line-scanning

Figure 22.3. Techniques for FLIM implemented in scanning microscopes

22.3.2. Fluorescence Lifetime Imaging (FLIM)

22.3.2.1. Scanning FLIM microscopy

Figure 22.3 shows a schematic overview of different approaches to realise FLIM in scanning microscopes. As well as TCSPC and photon-binning, scanning FLIM microscopy is also implemented using frequency domain techniques. It is straightforward to use a sinusoidally modulated excitation laser, e.g. ^^, and apply synchronous detection, e.g. using a "lock-in" amplifier to determine the phase difference and change in modulation depth between the excitation signal and the resulting sinusoidally modulated fluorescence signal. One can also take advantage of the pulsed excitation source in two photon microscopy and exploit the harmonic content of the resulting fluorescence.^^ This technique is essentially analogue and is not limited by the dead-time associated with single photon counting and can provide high speed FLIM.

TCSPC is widely regarded as the most accurate method of lifetime determination due to the high photon efficiency, low temporal jitter and high temporal precision (large number of bins in the photon arrival time histogram) and high dynamic range (typically millions of photons can be recorded without saturation). Reductions in detector dead time and increased performance of

488 D.S.ELSONEr^i:.

electronic circuitry and laser repetition rates have transformed TCSPC into a much faster acquisition method.^^ The photon time-binning approach (and the frequency-domain approach) can be faster still, but there remains the requirement to scan the sample pixel by pixel in order to produce a FLIM image (except if using custom detectors),"^^ usually resulting in image acquisition times of at least a few seconds even when reducing the number of photons collected to only a few hundred per pixel (the minimum required to achieve a -10 % accuracy in the lifetime). '" ^ In practice however, reaching real-time imaging rates in a scanning microscope would require such high excitation powers and fluorophore concentrations that it is impractical for many biological samples. A recent direct comparison of TSCPC and (analogue) frequency domain FLIM concluded that while TCSPC provided a better signal to noise ratio (SNR) for weak fluorescence signals, the frequency domain approach was faster and exhibited less distortion for bright samples." ^

An obvious way to increase the imaging rate is to introduce parallel pixel acquisition and adopt a line-scanning approach. This has recently been implemented using multiphoton excitation with a multiple beam array rapidly scanned to a line to produce a line of fluorescence emission that is relayed to the input slit of a streak camera." ^ FLIM images have been acquired in less than one second using this approach, which is currently limited by the read-out rate of the streak camera system.

22.3.2.2. Wide-field FLIM techniques

The parallel nature of wide-field imaging techniques can support FLIM imaging rates of lO's to lOO's Hz, e.g. ^ ' \ although the maximum acquisition speed is still of course limited by the minimum number of photons/pixel required to determine an accurate lifetime. Wide-field FLIM may be implemented with frequency or time domain approaches, as represented in figure 22.4.

MJ

Figure 22.4. Wide-field approaches to FLIM.


The frequency modulated (FM) approach is the most established with several groups demonstrating systems around 1990, e.g. ^ ' , that employed frequency-modulated laser excitation and utilized a microchannel plate (MCP) image intensifier with frequency-modulated gain to analyse the resulting fluorescence by acquiring a series of "intensified"images with the phase of the MCP signal shifted with respect to the excitation signal. Originally the optical output image from the intensifier was read out using a linear photodiode array but this was rapidly superseded by CCD camera technology. ' ' ^

Wide-field time-domain FLIM has also utilized modulated MCP image intensifiers coupled to CCD cameras in an approach described as time-gated imaging. The MCP image intensifiers are "switched on" for short periods of time after optical excitation so that the fluorescence can be sampled as it decays. Initially this approach was limited to measuring relatively long decay components since the shortest gate widths that could be applied to the MCP were over 5 ns but the technology quickly developed to provide subnanosecond gating times^^ and then the use of a wire mesh proximity-coupled to the MCP photocathode led to devices with sub 100 ps resolution.^^'^^'^^ These ultrashort gated optical intensifiers (GOI) may be applied to resolve fluorescence lifetimes from <100 ps to )us and form the basis of our wide-field FLIM instrumentation at Imperial College London, as discussed in section 22.0.

While both frequency and time-domain FLIM offer massively parallel pixel acquisition compared to scanning microscopy techniques, they both suffer from inherently reduced photon efficiency owing to the time-varying gain applied to the MCP image intensifier. This is particularly significant in the time domain when very short time gates may be applied in order to precisely sample the fluorescence decay profile. In many situations, however, such short time gates are not necessary since two gates are sufficent to describe a monoexponential decay (providing the background level is known) and the gate widths can be of the same order as the fluorescence lifetime. Sub-nanosecond gate widths are useful for sampling complex decays, for which high temporal resolution is important. For some situations, however, it may be advantageous to use longer gate widths - albeit with steep (sub-100 ps) rising and falling edges - to collect more of the fluorescence signal. Using gate-widths rather longer than the fluorescence lifetime, it is possible to realise photon efficiencies of lO's %. In the frequency domain, similar photon efficiencies can be achieved using three phase resolved image acquisitions to determine the fluorescence lifetimes. Such simple sampling strategies must be "tuned" to the expected fluorescence lifetime of the sample under investigation. If the sample exhibits a significant range of Ifuorescence lifetimes, it is necessary to use more than two time gates, or more than three phases delays, to obtain accurate lifetime values.

More recent developments of wide-field FLIM have addressed the characterisation of complex exponential fluorescence decay profiles, higher-speed imaging and progress towards lower cost FLIM instrumentation.

To characterise complex fluorescence decay profiles in the frequency domain, one can determine the sample response at a range of modulation frequencies and then extract multiple exponential decay components. A

490 D.S.ELSON^r^L.

particularly elegant approach exploits a non-sinusoidal detector gain (a fast time-gated MCP detector response) to simultaneously measure at multiple harmonic frequencies in the excitation signal.^^ This can reduce the sample exposure time, limiting photobleaching and photodamage. In the time domain, it is straightforward to sample a complex fluorescence decay profile using time-gating and fit the data to different complex decay models, e.g. using a nonlinear least squares Levenberg Marquardt algorithm. This requires more than two time gates and sufficient time resolution to sample the fastest decay components.

The most common approach to analyse complex fluorescence decay profiles is to fit to a number of discrete lifetime components, which could correspond to a number of different fluorophore species or states. Fitting to a double exponential decay is straightforward providing that the data exhibits sufficient signal to noise - some -lOOOO's photons/pixel are required for an accurate double exponential fit compared to -lOO's for a monoexponential decay profile." ^ In general, the "goodness of fit" will tend to improve as a data set is fitted to more parameters (exponential decay components) but the impact of noise will also increase and the resulting FLIM images are often degraded. Fitting to multiple discrete decay components can therefore be misleading. It is also often inappropriate because complex biological tissue samples may contain many different fluorophore species or states and the assumption of two or three discrete decay components is essentially arbitrary. Other models for complex decay profiles may be employed with at least equal validity, as discussed in section 22.0, including the stretched exponential decay model that we have applied at Imperial College London to tissue autofluorescence.

With respect to high-speed frequency domain imaging, it is possible to implement wide-field FLIM with a single modulation frequency using three phase measurements.^^ Using such an approach to calculate the apparent (i.e. assuming a monoexponential decay) fluorescence lifetime, an endoscope-based system acquiring phase-resolved images at 25 Hz achieved real-time FLIM for a field of view of 32 x 32 pixels has been demonstrated.^^ A microscope-based system achieved a FLIM rate of 0.7 Hz for a field of view of 300 x 220 pixels and also provided "lifetime-resolved" images (based on the difference between only two phase resolved images) at up to 55 Hz for 164 x 123 pixels.^^ Wide-field time domain FLIM has achieved even higher frame rates with a single-shot technique reaching 100 Hz for a field of view of 126 X 128 pixels. " Highspeed time domain FLIM has been a particular emphasis of our research at Imperial and recent results using both single shot and rapid sequential image acquisition techniques are presented in section 22.0.

To reduce the cost of FLIM instrumentation, it is necessary to address both the excitation source and the detector technology. While diode lasers have been applied to frequency domain FLIM in scanning microscopes^^'^^ and to time domain FLIM in scanning^^ and wide-field^^ FLIM systems, their relatively low output power (typically < ImW) and sparse spectral coverage limit their range of applications. LEDs, however, are increasingly interesting for fluorescence excitation, particularly as their power increases and their spectral coverage extends to the deep ultraviolet, e.g. 280 nm. ^ Their low spatial coherence and ability to be directly (electrically) modulated make them attractive for


wide^ield frequency domain FLIM-in comparison to more expensive and complex gas lasers that are typically temporally modulated using acousto-optic technology and whose output beams must be "conditioned" to reduce the spatial coherence, e.g. using a diffuser or a vibrating multi-mode fibre, in order to avoid speckle. Wide-field FLIM excited by a frequency modulated LED has been demonstrated^^ and commercialised by Lambert Instruments.^^ We note that an LED has also recently been used to realise frequency domain FLIM in a scanning microscope. ^ With respect to lower cost detection technology, it would be desirable to remove the requirement for the modulated MCP image intensifier and to this end the direct modulation of a CCD sensor for FLIM has been demonstrated.^^ To date, however, the repetition rate is limited to lOO's kHz and only nanosecond lifetime resolution has been achieved. "

Comparisons are often made between the frequency and time domain approaches to FLIM with a view to deciding the best approach for a particular application. In principle they can provide equivalent information but there are a number of issues that lead to different trade-offs and advantages for each approach. As discussed above, both approaches can analyse single and double exponential decay profiles. Fitting data to more complex decay models, such as a stretched exponential decay profile, appears to be more straightforward using time-resolved data. When using the most efficient sampling strategies, the photon efficiency and signal to noise ratio are similar in the frequency and time-domains and both can provide fast (real-time) FLIM applied to samples exhibiting simple monoexponential decays. The processing time is faster for the time domain, however, since a simple analytic method exists to calculate lifetimes,^^ whereas in the frequency domain a slower pixel-by-pixel Fourier analysis or fitting is required.^^ Ultimately it is possible to carry out such calculations in hardware, thus increasing the calculation speed of both time- and frequency-domain methods to the point that this does not become a practical issue.^^

The frequency domain approach is often considered more attractive when developing low-cost FLIM instrumentation since modulated lasers, particularly semiconductor lasers, have traditionally been rather cheaper than ultrafast lasers, although this is changing with the commercial availability of picosecond pulsed diode lasers and lower cost ultrafast diode-pumped solid-state and fibre lasers. Undertaking FLIM at a single modulation frequency can be problematic for samples exhibiting diverse fluorophores with very different lifetimes since high frequencies are required for good time resolution but the modulation period should also be long relative to the fluorescence lifetime. In the time domain one can achieve high time resolution using ultrashort pulses but accommodate longer fluorescence lifetimes by using a lower pulse repetition rate. Having said this, many practitioners use mode-locked Ti:Sapphire lasers, particularly in multiphoton microscopes, which tend to be fixed at ~ 80 MHz pulse repetition rate. This frequency is higher than optimal for many biological fluorophores.

492 D.S.ELSON^r^i.

22.3.3. Complex decay proxies and the stretched exponential function

As discussed above, fluorescence decay profiles are often complex, i.e. they do not fit well to monoexponential decay models. It is expected that complex biological tissues will comprise many different fluorophores and ranges of fluorophore environments such that the autofluorescence will necessarily be complex. In other samples, a multiplicity of decay pathways can introduce more complicated temporal dynamics, for instance collisional quenching mechanisms produce non-exponential decay behaviour as can the process of Forster resonance energy transfer (FRET).^^ In some situations, a priori knowledge of the sample will determine the appropriate model to fit fluorescence decay data - such as multiple discrete decay components when different species are known to be present. With sufficient a priori knowledge, one can apply global analysis techniques'^ and significantly improve the signal to noise ratio and resulting FLIM images. With biological tissue, however, the origin of the autofluorescence is often not fully understood and there is often no a priori fluorescence decay model. In these situations it is common to fit a single exponential decay since the resulting FLIM image will often provide lifetime contrast, even though the fitting programme may report a poor fit to the model.

To better represent complex fluorescence decay data, more sophisticated fitting models have been proposed e.g. power law^^ and Laguerre functions. ^ At Imperial, we have investigated the stretched exponential decay function (StrEF),^^ which corresponds to a continuous spectrum of lifetime components and has been applied to explain a number of physical phenomena.^^ The form of the StrEF equation is

I(t) = I^e ^''^''^ + const. (6)

where T ^W is the decay parameter and h is known as the heterogeneity. It can be shown that this equation is equivalent to a continuous distribution of fluorescence lifetimes:

I(t) = \e~'p{T)dT (7) 0

and that it is possible to calculate an average lifetime^^ using

(T)=hT^nh) (8)

where T is the gamma function.


10 15 20 time (ns)

Figure 22.5. Stretched exponential model.

This model has only one additional fitting parameter, h, compared with the single exponential decay and can provide a reasonable representation of real biological decay data. ^ The physical significance of h is that it corresponds to the width of the lifetime distribution, and may itself be a useful contrast parameter in situations where the local fluorophore environment is complex or changing. See reference^^ for a comparison of the single, double and stretched exponential decay models applied to tissue autofluorescence data.

Further examples of the use of this function will be explored in section 22.0 of this chapter, showing pure endogenous fluorescence results imaged in multiwell plates.

22.3.4. Wide-field time-domain FLIM instrumentation

The instrumentation developed at Imperial College London for wide-field time-gated FLIM^^ will be described here since it has been used to obtain many of the tissue FLIM images presented in this chapter but is not a commercially available system. The experimental configuration is illustrated in figure 22.6. The output beam from an ultrafast laser system, which is typically a pulsed diode laser, a frequency doubled femtosecond TiiSapphire laser or the output from a femtosecond Ti:Sapphire laser-pumped optical parametric oscillator, is passed through a rotating diffuser wheel to reduce its spatial coherence in order to eliminate any laser speckle noise. This excitation beam is directed onto the sample using Kohler illumination and the resulting wide-field fluorescence image is detected at the photocathode of a gated optical image intensifier (GOI). The photocathode voltage is applied for short (gate) periods following each excitation pulse after a delay that can be electronically controlled by a computer. The resulting time-gated photoelectron "image" is subsequently amplified by a microchannel plate in the optical intensifier, converted back into photons by the phosphor screen at the output of the intensifier and the resulting optical signal is imaged onto a CCD camera so that the fluorescence image can be digitised and stored on a computer. As well as applying this wide-field FLIM technology to

494 D.S.ELSON^r^L.

Figure 22.6. Typical experimental set-up for time-gated FLIM using a gated optical intensifier

fluorescence microscopy in a conventional (inverted) wide-field microscope, we also apply it to a home-built multiwell plate reader and to endoscopic FLIM.

The GOI technology, which has been developed by Kentech Instruments Ltd, " falls into two main classes. For operation at pulse repetition rates of- 40 MHz to 1 GHz the gated image intensifier technology (model HRI) provides time gates from ~ 300 ps - 5 ns width and exhibits a spatial resolution of-10 lines/mm. For lower pulse repetition rates up to -10 kHz it is possible to achieve shorter gatewidths down to -80 ps duration with an improved spatial resolution of >10 lines/mm (model GOI). This is achieved by applying the gating voltage to a mesh in front of the input window of the MCP, rather than to the photocathode. The delay between the pulsed excitation and the triggering of the GOI gating voltage is controlled with a precision of 25 ps using a programmable delay generator. " The values quoted here for gatewidths include all temporal jitter in the system, such as the interpulse jitter from the ultrafast lasers (which is typically < 20 ps). For high-speed wide-field FLIM, we have also worked with a special multi-channel GOI prototype that has a segmented photocathode that enables it to acquire four time-gated images simultaneously with a different delay of the time gate applied to each. This facilitates single-shot FLIM and has been used to acquire FLIM images of tissue autofluorescence at 10 frames/second, as described in section 22.0. ^

The software used to control the data acquisition and analysis has been written in house using Lab VIEW. The typical "high precision" mode of operation uses between 8 and 16 time gates to sample the fluorescence signal and a nonlinear weighted least squares algorithm to fit the acquired data to single, double or


stretched exponential models of fluorescence decay profiles. Total data acquisition times for unstained tissue autofluorescence are lO's of seconds and the computation time required to calculate the FLIM images is on the order of seconds. When examining FLIM data, we typically initially view a false colour fluorescence lifetime map, an integrated fluorescence intensity image (obtained by summing the acquired time-gated fluorescence intensity images) and a "merged" FLIM "image" that is obtained by modulating the false colour fluorescence lifetime map with the integrated intensity image.

22.4. MULTIWELL PLATE IMAGING OF ENDOGENOUS FLUOROPHORES

Multiwell plate FLIM systems^ ' ^ provide a convenient way of analysing multiple samples in parallel and we have investigated endogenous fluorophores in isolation to better understand the lifetime contrasts that are observed in bulk tissues. Usually this is achieved in a macro-imaging instrument where the wide-field imaging capability of our FLIM instrument is exploited to simultaneously record multiple measurements of the lifetime of each well.^^ This means that it is possible to record a lifetime histogram for a single sample using only one acquisition run, where the standard deviation of the histogram gives an indication of the error of the average lifetime.

A FLIM map of five important endogenous fluorophores is shown in figure 22.7, from top to bottom porphyrins, NADH, collagen, elastin and flavins, excited using a blue (405 nm) picosecond laser diode.^^ The single exponential lifetimes found agree well with those reviewed in section 22.0 but it should be noted that these decays are in some cases fitted much better with a stretched exponential decay (for instance the reduced chi squared parameter, x r? that gives an indication of the goodness of fit^^ is reduced from 8.4 with the single exponential to 1.2 using a StrEF for collagen.)^^

22.5. FLIM MICROSCOPY OF BIOLOGICAL TISSUE

While there have been several reports of single point measurements of fluorescence lifetimes in tissue, ' ' there has as yet been little work on FLIM, although there are some useful reviews to introduce the field."^'^^ In this section we will introduce our preliminary work on FLIM of cartilage, artery, pancreas and cervix. These results are very much work in progress and it is not our intention to do more than illustrate the potential of FLIM to provide label-free contrast in unstained tissues.

One of the major aims of our FLIM programme is to investigate the promising lifetime contrast observed in many tissue sections in order to be able to apply FLIM as a useful diagnostic tool in intact tissue or in vivo. Tissue

496 D.S.ELSON^r^L.

sections provide an accessible opportunity to investigate this contrast since problems associated with tissue scattering, background fluorescence and other practical issues are minimized. Furthermore, fixed or frozen sections are the most convenient way to acquire and transport clinical tissue samples. It is well known however, that the fluorescence lifetime can be affected by the local fluorophore environment, and that the fluorescence generated from prepared ex vivo tissue can provide only an approximate picture of that which may be observed in vivo. We have recently used a Krumdieck tissue slicer which has shown greater contrast consistent with living tissue. However, comparison with artificially fixed tissue is a necessary precursor to in vivo imaging and the opportunity to compare high magnification FLIM images with serial sections that have been stained with conventional histopathological preparations (e.g. Haematoxylin and Eosin - H&E) is invaluable to discover the structural and chemical composition.

Porphyrins NADH

Collagen Elastin Flavins

(c)

5000 1 10 Delsy Ops)

:in Data Stretched exponential Single exponential

5000^ , 110^ Delay (ps) 1.5 10

(d)

Figure 22.7. Endogenous fluorophores and the StrEF. (a) shows the fluorescence lifetime map of the five fluorophores and the lifetime histogram across all pixels is shown in (b). The stretched exponential and single exponential fits and residuals for the collagen decay are compared in (c) and (d). (See color insert section.)


T = 800-2600 ps,formalin-rixed human pancreas, 10 |im section.

T = 400-2900 ps, frozen human colon, 10 )Lim section.

T = 1000-2500 ps, fresh mouse kidney, 200 jj.m section.

Figure 22.8. Examples of FLIM images of fixed, frozen and fresh tissue, (See color insert section.)

To address the important issue for FLIM concerning the transition between imaging tissue sections using a microscope and imaging tissue in vivo, perhaps using an endoscope, we have begun to investigate FLIM contrast for different tissue preparations, including fixed, frozen and fresh tissue sections. Figure 22.8 shows an example of each.

There is currently considerable uncertainty concerning the extent to which different methods of tissue preparation affect the fluorescence lifetime images. To date we have not been able to undertake a systematic study of the variation in FLIM contrast between fresh and frozen tissue sections and tissue sections prepared with different fixing agents. We have observed, however, that different fixing agents such as methanol, formalin and gluteraldehyde can produce different FLIM images^ ' ^ but that the lifetime contrast tends to be preserved while the lifetime values vary. Of the various fixing agents we have investigated, formalin seems to produce FLIM images most similar to fresh tissue sections although the average fluorescence lifetime values are slightly different and much less cellular autofluorescence is observed. These preliminary results are encouraging as they appear to validate the extrapolation from formalin fixed to fresh tissue but of course much more work needs to be done to make the transition to in vivo FLIM contrast.

22.5.1. Cartilage

This section presents results from what we believe to be the first study of FLIM applied to healthy and diseased specimens of human articular cartilage.^^ The major tissue constituents of articular cartilage are collagen type II and aggrecan. Degradation of these components is known to occur in degenerative joint diseases such as osteoarthritis and rheumatoid arthritis.^ ' ^ Three samples will be discussed here comprising one healthy tissue sample, one diseased sample with heavy erosion of cartilage, and one diseased sample that is only partly eroded.

498 D.S.ELSON^r^.

Femoral heads that had been removed during hip replacement procedures were used in this study. The femoral heads were fixed in 10 % buffered formalin and a 4 mm slab taken through the specimen was decalcified with formic acid and cut into blocks that were processed to paraffin wax. Sections (4 ^m thick) were cut and mounted on glass slides. Alternate sections were stained with H&E and examined with conventional light microscopy. The adjacent unstained sections were used for fluorescence studies. The standard wide-field FLIM set-up shown in figure 22.6 was m^y] to record the fluorescence lifetime maps shown in figure 22.9. Ihe cACiiutiviii iacsci 5uuice

was a diode-pumped ultrafast Ti:Sapphire oscillator-amplifier system pumping a frequency quadrupled optical parametric amplifier system (Spectra Physics Ltd, Hurricane and OP A system). This was tuned to provide 200 fs excitation pulses at 355 nm at a repetition rate of 5 kHz with an average output power of 5 mW (attenuated to 1 mW).

Within hyaline cartilage, the currently identified fluorophores (excitation/emission wavelengths) are the collagen pyridinium crosslinks (325/400 nm)P'^^ pentosidine (335/385 nm) " and 2,6-dimethyldifuro-8-pyrone (DDP) (305/395 nm). ^ Since the excitation wavelength for our study was 355 nm and emission detection window was >375 nm, we expect to detect fluorescence from more than one of these fluorophores. The lifetime of DDP has not been measured but collagen powder of unspecified type has been measured to have a bi-exponential decay, with lifetimes of 2.7 ns and 8.9 ns, ^ and pentosidine in water has a measured lifetime of 4.4 ns. ^ We note that fluorophore lifetimes of components measured within cartilage may well be different from those measured after chemical purification. It must also be noted that our study used fixed tissue and that fixation might have affected the lifetimes of the autofluorescent components.

The fluorescence lifetimes observed for the healthy cartilage sample 1 were different between areas of the joint with a relatively thick layer of overlying cartilage and areas with thinner layers of articular cartilage (1.6 versus 1.1 ns). The thin region of cartilage also had an area of lower lifetime at its surface, suggestive of a difference in composition or structure. In the thick region, the pixel to pixel variation (standard deviation) was found to be much greater (280 ps) than variations in average lifetime when moving from the sub-chondral region to the surface. This suggests microscopic heterogeneity that is similar across the whole thickness of the cartilage. The thinner region of the cartilage had a lower standard deviation of about 150 ps which suggests that the thinner region has a more uniform structure - except at its surface. These differences between thin and thick regions may be related to different load-bearing histories since the thick region generally supports more weight.

In sample II, some regions of cartilage showed erosion but the lifetime of the extra-cellular matrix in less degraded areas were similar to those observed in the healthy tissue (sample I). However, an increase of >200 ps in the lifetime of chondrocytes and their lacunae may indicate cellular changes related to (and/or causing) articular degeneration. A few regions exhibiting lower fluorescence lifetimes were observed around fissures in the damaged cartilage (region D) but


Region A

Sample I: healthy

Region D RegiqnC I

H&E Stain

Sample II: diseased Region C

2000ps

400p8

1500ps

Region D Ji700ps

1700ps

BOOps

Region E

Figure 22.9. H&E images and FLIM maps of different regions of healthy and diseased human cartilage. The FLIM maps were recorded for the areas highlighted on the H&E images. Regions A and B are from thin and thick regions of healthy cartilage respectively. C contains the frayed cartilage surface and subchondral bone, D contains a fissure with some regions of short lifetime, (See color insert section.)

these were very localized and occurred in different positions relative to different fissures (figure 22.9).

It was noted that the surface of the damaged cartilage, away from fissures was frayed compared to the smooth surface of the healthy tissue. In the

500 D.S.ELSON£r.4£.

FLIM image of region E, a region of lower lifetime at the surface of the cartilage can be seen. Although a similar phenomenon was noted in the healthy tissue, the region extends about five times deeper into the diseased tissue and the change is more gradual. This is perhaps indicative of structural changes that occur before or accompanying fraying. In the severely damaged sample (not presented in figure 22.9) the fluorescence lifetimes exhibited a larger standard deviation (350 ps), which is more than twice that observed in healthy tissue, indicating that the cartilage is more heterogeneous. The lifetime also varied markedly within the sample, with large differences between adjacent regions. This suggests that FLIM may be used to differentiate between healthy and diseased cartilage.

To summarise, our results show that there are fluorescence lifetime variations within healthy tissue that can change depending on the thickness of cartilage. This indicates a difference in chemical composition or microscopic structure which perhaps results from the differences in mechanical loading of the two regions. For tissue in the early stages of disease, we found that the fluorescence lifetime characteristics of the chondrocytes and their lacunae preceded changes in the extra-cellular matrix. In the more advanced stages of disease, the lifetime characteristics of the whole sample were different from healthy tissue.

22.5.2. Artery wall and atherosclerotic plaques

Within arterial tissue, tryptophan, with an emission peak at 325 nm, and collagen and elastin, both with peaks at 380 nm, have been identified as the dominant fluorophores, accounting for 95 % of the fluorescence when exciting at 310 nm. ^ It has also been found that the best excitation wavelengths for spectroscopic studies of atherosclerotic samples are between 314 nm and 344 nm^ ' , except for calcified plaques, which require excitation at greater than 380 nm. ^ It is important to note that the optimum wavelengths to obtain the best fluorescence lifetime contrast between different tissue components may be different from these excitation maxima.

Several studies have shown that fluorescence of arteries is affected by the presence of plaques. Artificially increasing the levels of collagen types I and III and cholesterol in normal canine aorta induced significant changes in fluorescence intensity of the tissue. ' ^ Several more studies have utilized spectroscopic data to distinguish normal aortas from those with plaques in varying stages of development, see for example.' ^ ' ^ ' ' ^^ ' ^ ' "^ A few fluorescence lifetime studies of arteries have also been carried out, showing an enhanced capability for identification and classification of atherosclerotic plaques but, to date, these have all been carried out in non-imaging modalities. « ' ««' «^

We have undertaken preliminary experiments in formalin fixed rabbit aorta using excitation wavelengths between 355 nm and 520 nm in order to investigate the potential of FLIM for studying and diagnosing arterial plaques. ^ The atherosclerotic rabbit aortas used in this study were obtained


from 2 month old male New Zealand White (NZW) rabbits. The rabbits had a supplemented diet with 1 % (w/w) cholesterol (Sigma) for 8 weeks. Following sacrifice, the abdominal aorta was formalin fixed in situ, excised, then postfixed also in 10 % buffered formalin and stored at 4°C. Prior to the imaging experiments, the aorta was soaked in saline buffer solution at a neutral pH in order to remove any formaldehyde. 10 jiim thick cross-sections of aorta were cut at 30°C using an embedding compound (Cryo-M-Bed) and mounted onto glass microscope slides.

The samples were typically imaged at xlO magnification and atherosclerotic lesions were identified by the thickening and irregular surface of the intima. We observed that it was possible to differentiate between atherosclerotic lesions and artery wall using FLIM when exciting from 355-440 nm. The strongest lifetime contrast between the artery wall and atherosclerotic lesions was obtained for 355 nm excitation, illustrated in figure 22.10, although contrast was still observed for excitation wavelengths up to 440 nm, and at 400 nm the intima was also distinguishable. For comparison, we also examined these samples with conventional spectrally-resolved fluorescence emission imaging. Using multi-spectral analysis we found that the level of contrast obtained using the fluorescence lifetime could not be recreated for these excitation wavelengths.

22.5.3. Neoplastic tissue A particularly exciting application of FLIM would be to demonstrate

contrast between normal and neoplastic tissue. We have recently started work with fresh tissue biopsies with the Krumdieck Tissue Slicer to contrast FLIM appearance of normal and neoplastic tissue and have obtained some promising initial results from cervical biopsies.

Figure 22.11 shows FLIM images of 200 ^m thick fresh tissue sections obtained from a normal cervical biopsy and a cervical biopsy with extensive precancerous neoplastic change (CIN - cervical intraepithelial neoplasia -grades II and III). We observe that, in the abnormal biopsy, areas of epithelium containing immature dysplastic cells show greater autofluorescence intensity

n 0.01 ] E

i mk

f 1 P " Plaque 1

1 —.Media

1 /AA A. %^

2050 2250

Lifetime (ps)

Figure 22.10. Fluorescence lifetime map and corresponding lifetime histogram of a section of artery showing lifetime contrast between the media and the plaque. (See color insert section.)

502 jy.S.ELSON ETAL.

(b)

Epithelium

iilipisem,^, ,;iS^umembr|

Figure 22.11. TCSPC FLIM - intensity images (a,c,f) and intensity-weighted FLIM maps (b,d,g) of fresh 200 ^im-thick tissue sections of cervix and the corresponding H^ E-stained 6 |im-thiclc tissue sections cut subsequently after fixing the 200 ^m sections in formalin. (a,b) Normal (post-menopausal) cervix; (c-e) CIN II; (f-h) CIN III. The CIN II images shows a 'tide mark' within the epithelium (dashed line) representing the change from immature dysplastic cells in the basal half of the epithelium to maturing dysplastic cells showing cytoplasmic keratinisation in the upper half; in the CIN III images the full epithelial thickness is occupied by immature dysplastic cells. Note the autofluorescence from the areas containing immature dysplastic cells shows shorter lifetimes and greater fluorescence intensity than from areas occupied by maturing keratinised dysplastic cells or normal epithelium. This difference may be due to greater metabolic activity and higher NADH levels in the immature cells. (Two photon excitation wavelength = 740 nm; detection range 385-600 nm; x20 microscope objective). (See color insert section.)

and significantly shorter mean fluorescence lifetimes than areas of more mature keratinised dysplastic cells or normal cervical epithelium, as shown in figure 22.11. We believe that this difference is due to greater metabolic activity and higher NADH levels in the immature cells.

In addition, we have also investigated a number of fixed tissue sections that have contained areas of neoplasia. Figure 22.12 illustrates one example that is relevant to the ability of FLIM to detect abnormal tissue. It should be


Figure 22.12. Wide-field time-gated FLIM of a formalin-fixed 10 |am-thick tissue section of pancreas showing partial invasion of an Islet of Langerhans (dotted line) by malignant cells (solid line and arrows), (a) Light microscope image of an H+E stained section; (b) False colour FLIM map; (c) Intensity weighted FLIM image. FLIM reveals the uninvaded portion of the islet to be contrasted to both the invaded area and the surrounding malignant parenchyma due to its shorter fluorescence lifetime. There is however no apparent contrast between the malignant parenchyma and the normal pancreatic parenchyma. (Excitation wavelength = 400nm, detection range > 435 nm, microscope objective = x40). (See color insert section.)

emphasised that this is an isolated observation rather than the result of a protocol designed to identify fluorescence lifetime changes associated with neoplasia and that such lifetime changes associated with neoplasia were not been seen in all FLIM images of neoplastic tissues - for example FLIM images of 10 |Lim fixed sections of MIN mouse colon did not reveal any appreciable changes in lifetime between normal colonic mucosa and areas of neoplasia and dysplasia. However, our recent work with fresh tissues indicates that fixation may reduce or distort lifetimes and/or contrast so that subtle changes may not be apparent in sections but would be manifest in living, intact tissues.

22.6. TOWARDS IN VIVO IMAGING

Exploiting the intrinsic contrast from autofluorescence for the diagnosis and monitoring of disease provides a strong motivation to develop in vivo FLIM systems. These will be initially applied to easily accessible tissues, (e.g. skin, eyes, oral mucosa), or to internal surfaces using an endoscope (e.g. GI tract, cervix, bronchial tree, bladder). An obvious application would be to provide a "red flag" technique identifying areas of early or pre-cancer not readily apparent under white-light visual inspection to guide biopsy, e.g. for GI endoscopy or colposcopy. There will also be applications involving exogenous fluorescence markers. One example is the diagnosis of skin tumours^^^ using the topically applied marker 5-aminolevulinic acid (5-ALA). 5-ALA is metabolised intracellularly by the haem biosynthetic pathway to produce the fluorophore protoporphyrin IX (PpIX).^^^ PpIX accumulates with some degree of selectivity in tumour cells due to their low levels of activity of the PpIX metabolizing enzyme ferrochelatase. Thus the presence of cancerous tissue is

504 D.S.ELSON^r^L.

signalled by the longer fluorescence lifetime of the Protoporhyrin IX compared to the surrounding tissue. Another possible exogenous FLIM probe is BCECF, which has been used to study the pH gradient in skin using two-photon excited

Unfortunately the FLIM technology discussed in the previous sections is not practical for most in vivo applications, being too slow, too complex and too cumbersome. A clinical FLIM instrument should be able to acquire and display the lifetime images in real-time. At Imperial we have been working towards developing compact FLIM instrumentation capable of real-time imaging and endoscopic application. We note that real-time FLIM would also find a range of other applications including live cell imaging, particularly for FLIM/FRET experiments, and imaging microfluidic processes. ^"^

22.6.1. Real-Time FLIM

As discussed in section 3, a number of groups have explored routes towards high-speed FLIM. In the frequency domain an average fluorescence lifetime can be calculated by modulating the excitation and detection at a single frequency and acquiring images at three detector phases, thereby creating lifetime maps at up to video-rate.^ ' ^ In the time-domain high-speed FLIM can be implemented using an analytical approach to calculate the single exponential decay lifetime based on acquiring two-time gated images: ' ^^

-M (9)

ln(A/A)

where Di and D2 are the intensities of two time-gated images separated by At. This assumes that the background level is zero or can be subtracted. If both gates are of equal width, it is found from error analysis that the most accurate lifetimes can be calculated if At is 2.5 times the lifetime being investigated, and the gate widths are equal to At. ^ This optimised measurement is only accurate for a narrow range of sample fluorescence lifetimes. To achieve high accuracy over a wider range of sample lifetimes, one can increase the width of the second time gate^^^ or use three or more time gates, albeit at the expense of imaging speed. There are also analytic expressions for calculating lifetimes of a single exponential decay with an unknown background^^^ or a biexponential decay using four time-gated images. ^^ At Imperial we have demonstrated video-rate wide-field FLIM using equation 9 to calculate FLIM maps from two sequentially acquired time-gated fluorescence images (and a previously acquired background image) using a GOI based system. ' " An important step


towards real-time FLIM was the use of wide time gates (several ns), albeit with steep sides (of ~ 100 ps). Increasing the gate width means that more of the emitted photons are detected during each acquisition, thereby increasing the sensitivity and imaging speed. For monoexponential decays, the lifetime resolution is independent of the gate width and, even for the complex decays associated with tissue autofluorescence, it is our experience that longer gates yield superior FLIM images (fitted to monoexponential decay profiles). A second important factor was the development of a delay control unit that could switch in less than 2 ms. " The main instrumentation factor now limiting the maximum possible wide-field FLIM frame rate is the persistence of the phosphor screen at the GOI output, which can be lO's of ms.

The maximum frame rate of wide-field FLIM based on sequential image acquisition can also be limited by severe lifetime artefacts that arise when imaging samples that move or change between acquisitions.^^ Such artefacts occur for both time and frequency domain FLIM and even a sample movement of one pixel between sequential image acquisitions can change the apparent lifetime by over 50 %. Using fast acquisition times with minimal delay between the sequential acquisitions will reduce this problem but it can only be eliminated though "single-shot" FLIM, for which all time-gated (or frequency modulated) images are acquired simultaneously in parallel. Two approaches to single-shot FLIM have been proposed to date.

The first, by Agronskaia et al., ^ involves dividing the fluorescence image from the sample into two sub-images using an intensity beam splitter and relaying these two sub-images onto separate regions of a GOI/CCD detector with one of the sub-images having been delayed by 5 ns with respect to the other using an optical imaging delay line. By registering the pixels of the two time-gated sub-images, one can calculate the fluorescence lifetimes using equation 9 with no motion artefacts. This elegant approach has been demonstrated at up to 100 Hz imaging a rat neonatal myocyte stained with the calcium indicator, Oregon green BAPTA-1. Its main drawbacks are that the field of view is reduced with respect to the conventional sequential acquisition approach and the complexity of the optical imaging delay line makes it difficult to adjust the delay between sub-images, in order to accommodate a range of sample fluorescence lifetimes. These issues may be addressed using multiple independently gated GOI detectors to maintain the field of view field, ^ ^ although this significantly increases the system size, complexity and cost.

We have addressed the second issue more simply by using a single GOI detector with a segmented photocathode that is divided into four quadrants that can be independently gated through the introduction of resistive elements.^^ This permits four time-gated sub-images to be acquired simultaneously with the relative gate delays being independently electronically adjustable. The four sub-images are obtained using an optical image-splitter with beam-splitter reflectivity's being chosen to ensure that the later sub-images receive more of the incident light. After pixel registration these four time-gated sub-images can be used to calculate the fluorescence lifetime analytically using equations similar to equation 9 or by linear or non-linear fitting of the decay model. The acquisition of four time-gated sub-images with different, electronically

506 D.S.ELSON^r^L.

adjustable, delay times confers considerable flexibility. The FLIM settings can be changed on-the-fly to suit different samples and high-speed FLIM of multiple fluorophores with different (or unknown) lifetimes is possible since the optimum combination of time-gated sub-images can be determined in postprocessing.^^ Figure 22.13 shows results obtained with this system applied to an unstained human pancreatic tissue section imaged with a xlO microscope objective. We have also applied this instrument to imaging arrays of chemicals in a multiwell plate reader at 20 frames per second. The ultimate frame rate of single-shot FLIM is dependent on the number of photons emitted by the sample but video-rate FLIM appears to be practical for unstained tissue autofluorescence and should provide useful clinical instruments.

22.6.2. Endoscopic FLIM

Our initial work on endoscopic FLIM involved applying the wide-field time-gated FLIM system described in section 22.0 through a flexible fibre bundle endoscope.^^ While this approach demonstrated that endoscopic FLIM was possible, the low frame-rate, sensitivity and sub-optimal design made it impractical for clinical experiments. Our recent work has focussed on applying the real-time (sequential) FLIM approach described in the previous section to rigid and flexible optical endoscopes.^"'^^^ Rigid endoscopes, such as the arthroscopes used for orthopaedic applications, relay the image through a series of rod lenses and provide relatively high throughout and image quality. Flexible optical endoscopes utilize fibre bundles to relay the image, resulting in loss of image quality and intensity compared to rigid endoscopes. Flexible endoscopes are usually longer (up to several metres) than arthroscopes and the resulting increase in group velocity dispersion (GVD) can temporally stretch fluorescence signals. For both rigid and flexible endoscopes, we generally use a separate low GVD optical fibre to bring the excitation light to the sample at the distal end of the endoscope. The excitation light for our endoscopic FLIM experiments is provided by a frequency tripled Nd:YV04 laser (Spectra Physics Inc, Vanguard 350-HMD355), which produces 10 ps pulses at 355 nm at a repetition rate of 80 MHz , with up to -350 mW average output power. This laser is relatively compact (60 x 15 x 10 cm) with a portable power supply that

Figure 22.13. FLIM of autofluorescence of a 10 |im human pancreatic tissue acquired at 10 frames per second. (See color insert section.)


does not require water cooling. An optical attenuator, consisting of a half wave plate and polarizing beam splitter, is used to control the excitation power.

For our first demonstration of endoscopic FLIM using a rigid endoscope, the sample was directly illuminated, using a holographic diffiiser to produce a cone of illumination with a full angle of-60°. The fluorescence emission was imaged through the conventional (Smith and Nephew), arthroscope pictured in figure 22.14 (a), passed through a 435 nm long pass filter to block any reflected or scattered excitation light, and imaged onto the photocathode of the GOI. A "real-time" FLIM image of the autofluorescence from a sample of lamb's kidney acquired at an update rate of 7.2 Hz is presented in figure 22.14 (c). For comparison, a conventional wide-field FLIM image of the same tissue (eight 400 ps time gates acquired over 15 seconds and fitted using a weighted nonlinear least squares iterative algorithm) is also shown. It will be seen that the "live" real-time image is noisier than that obtained with the conventional acquisition but the fluorescence lifetime contrast, particularly between the calyces and the medulla, is still apparent. We note that the "live" FLIM images appear less nosy that this "snapshot" because the human eye performs a degree of real-time image processing.

Figure 22.14. (a) Arthroscope used to obtain endoscopic FLIM images of bisected lambs kidney and FLIM images acquired using (b) conventional WNLLS at 0.1 Hz and (c) real time (7.2 Hz) FLIM acquisition and processing. (See color insert section.)

508 D.S.ELSON^r^L.

Figure 22.15. (a) Flexible 10 mm diameter endoscope and (b) FLIM image acquired at 5.5 Hz of the mucosal surface of piece of a bisected lamb's kidney viewed en-face with 355nm excitation. (See color insert section.)

Real-time FLIM has also been realised with the flexible endoscope shown in figure 22.15 (a). Even though the coupling and transmission losses of the fibre imaging bundle are significantly higher than for the rigid endoscope, this system is still capable of producing FLIM images of the endogenous fluorescence of fresh tissue at frame rates of several Hz. For this system the excitation light was coupled into the lower order modes of a 7 m long multimode fibre chosen for low intrinsic fluorescence when transmitting 355 nm light. This excitation fibre was then introduced through the biopsy port of a flexible endoscope and a light shaping diffuser was used at the distal end to produce a cone of illumination with a full angle of -60°. This was larger than the imaging angle of the endoscope (50°) and completely filled the field of view, providing a reasonably uniform illumination. The imaging channel of the flexible endoscope comprised a graded index objective cemented to a flexible coherent imaging bundle with -30,000 fibres. The output from the endoscope was imaged directly onto the photocathode of the gated optical imager (GOI) through a 435 nm long pass filter. Figure 22.15 (b) shows a fluorescence lifetime image of the bisected lamb's kidney acquired at an update rate of 5.5 Hz.

22.7. EMERGING TECHNOLOGY FOR FLIM AND MDFI

FLIM is a rapidly maturing field with an increasing range of applications in biology and medicine. As more potential applications are investigated, the need to combine fluorescence lifetime resolution with spectrally res( ;> u sion and versatile excitation spectroscopy becomes more apparent. The achievable FLIM contrast of tissue autofluorescence can often be improved by optimising the excitation wavelength or the detection spectral pass band. In order to design effective clinical instrumentation, e.g. for diagnosis or guided biopsy, it is important to develop tools to fully characterise autofluorescence in order to be able to design the most effective (simple) medical instrumentation and also to better understand the physiological origins of the observed fluorescence contrast.


In this last section we will outline some of the most recent advances in instrumentation that allows more functional information to be acquired from tissue autofluorescence. This includes a novel (electronically) tunable ultrafast light source spanning ~ 435-1200 nm and spectrally-resolved FLIM instrumentation capable of resolving fluorescence with respect to x, y, z, i & A, in a single acquisition.

22.7.1. Tunable continuum source for fluorescence excitation

The excitation source is perhaps the most critical component in a fluorescence imaging system. In wide-field fluorescence microscopes a filtered thermal or arc-generated white light source may provide tunable excitation across the u.v., visible and NIR but for more sophisticated optically sectioning microscopes (e.g. confocal microscopes) a spatially coherent laser source is usually required. The limited range of visible wavelengths available from convenient lasers (e.g. gas, solid-state and semiconductor laser sources) continues to constrain the design and utility of fluorescent probes for molecular biology and limits the information that can be gleaned from tissue autofluorescence. If one considers FLIM, then the availability of suitable ultrafast light sources is further restricted. This situation is mitigated by the development of convenient mode-locked Ti.Sapphire lasers and the opportunities for multiphoton excitation. Multiphoton microscopy has become the technique of choice for many fluorescence imaging laboratories, not always because the longer excitation length permits imaging to deeper depths in scattering media such as biological tissue but often because a single, user-friendly Ti: Sapphire can excite a wide range of fluorophores. For some applications, however, the higher excitation efficiency, superior spatial resolution and reduced nonlinear photodamage make single photon excitation of fluorescence desirable. A spatially coherent source that is conveniently tunable across the visible spectrum would be invaluable for fluorescence imaging and would provide new opportunities, e.g. automated excitation fingerprinting, enhanced separation of multiple fluorescence probes and in situ measurement of excitation cross-sections. FLIM requires a rapidly modulated or pulsed light source and currently, (tunable) ultrafast light sources covering the visible spectrum are complex, expensive and usually require a significant degree of expert adjustment to achieve tunability. We are now working with tunable continuum sources (TCS) that provide an electronically tunable, ultrafast and spatially coherent excitation source covering the visible and near infrared spectrum.

Recent developments in microstructured photonic crystalline fibre (PCF) and tapered fibres^^^ have enabled high average power continua of radiation to be generated using ultrashort pulse trains (e.g. directly from Ti:Sapphire femtosecond laser oscillators) propagating near the fibre zero group velocity dispersion (GVD) wavelength. ^ ^ These continua are spatially coherent, emerging from single mode fibres, and retain their ultrashort pulse

510 D. S.ELSON^J^.

characteristics although the pulses can be broadened to lO's ps by GVD. Spectral selection from such a supercontinuum provides a simple route to tunable excitation radiation for a wide range of fluorescence instrumentation. This has recently been proposed^ ^ and demonstrated for confocal fluorescence intensity imaging, ^^^ multiphoton microscopy ^^ , electronically tunable excitation spectroscopy ^ ^ and FLIM in confocal and wide-field imaging systems. ^

Our TCS provides a continuously electronically tunable (435-1150 nm) ultrafast source for fluorescence imaging applications and is derived from a visible continuum generated by injecting infrared ultrashort pulses in to a microstructured fibre. We routinely used this source for confocal and wide-field fluorescence microscopy, as well as multiwell plate imaging. Figure 22.16 shows the experimental configuration. A modelocked Ti: Sapphire laser is typically used as the pump source and provides up to 920 mW average power of 120 fs pulses at 80 MHz repetition rate centred at 790 nm. This pump beam is directed through an isolator arrangement to prevent back reflections from the fibre input interrupting the mode-locked operation of the Ti:Sapphire laser oscillator. A polarizing attenuator is used to adjust the power for alignment purposes and a x2 telescope is used to match the beam to the numerical aperture of the PCF, such that the pump light is typically coupled into the PCF with > 50 % efficiency. Using ~ 30 cm length of PCF (Crystal Fibre AS, #NL-740-

2.0), with a core diameter of 2 [xm and a zero-GVD wavelength at 740 nm, an incident input power of 770 mW produces a continuum spanning from 435-1150 nm at the -10 dB level with an average power of 220 mW, an example is shown in figure 22.16. Of this, approximately 70 mW was contained in the continuous visible spectral window from 435 nm to 700 nm. To realise convenient and electronically adjustable spectral selection of this continuum, the output light from the PCF is collimated using an Olympus x 40 achromatic objective lens (L4) and directed into the Fourier spectral selection set-up consisting of a lens and prism arrangement. In the back focal plane of the lens a triangular-shaped aperture is located directly in front of the mirror M3 and mounted on a (motorized) translation stage. Horizontal translation of this aperture adjusts the central wavelength and vertical translation adjusts the spectral width of the retroreflected radiation that is picked off at mirror M4. A more extensive description of the TCS and its applications is given in the reference.'^^

Figure 22.17 illustrates the applicationof the TCS to confocal microscopy, showing fluorescence intensity and lifetime images of a human B cell that has been labelled with green fluorescent protein (GFP). The spectrally selected light from the TCS (Aex = 480-500 nm, Pex = 1.8 mW) was coupled into an inverted scanning confocal microscope (Leica DMIRE2 with confocal scanning unit TCS SP2) via the u.v. port and using a 30/70 (reflection/transmission) intensity beam splitter. The ability to continuously adjust the excitation


Isolator Variable

attenuator Beam

expander

PBS1 PBS2 XI2 PBS3

400 600 800 1000 1200

Wavelength (nm)

Figure 22.16. Experimental setup. PBSl-3 polarizing beam splitters, PR Faraday rotator, Xll half wave plates, LI-5 lenses, Ml-4 mirrors, P prism, BD beam dump. Mirror M3 has a triangular-shaped slit attached directly onto the front of the mirror. Inset: output spectrum from fibre plotted on both linear and logarithmic scales. Reproduced from Journal of Physics D: Applied Physics © 2004 TOP Publishing Ltd. •''

wavelength is extremely convenient for intensity imaging as the emission detection bandwidth can be readily adjusted using the Leica TCS SP2 prism-based tunable emission filter. This facility was exploited to investigate the fluorescence intensity of the intracellular GFP in situ, as a function of excitation wavelength. Figure 22.17(b) shows the curve obtained using alOnm excitation bandwidth with the adjustable long pass filter (Leica TCS SP2) set at 20 nm above the mean excitation wavelength. To correct for variations in the excitation intensity, a (blank) reference area in the sample plane was imaged using the transmitted light detector to permit the fluorescence data acquired at each excitation wavelength to be normalized.

As the TCS is inherently ultrafast, FLIM could be directly implemented on the confocal microscope by simply using a photon counting detector and TCSPC control card (Becker & Hickl). The FLIM image corresponding to figure 22.17(a) is shown in figure 22.17(c). The TCS is also applicable to wide-field imaging and FLIM.

Figure 22.18 shows a widc-ficld intensity and time domain FLIM image of autofluorescence from an unstained section of human pancreas obtained in a conventional wide-field inverted microscope using the TCS. The potential of a rapidly electronically controlled tuning excitation source is also attractive for

512 D,S,ELSON ETAL.

(a)

475 500 525

Wavelength (nm)

(b) (c)

Figure 22.17. (a) Fluorescence intensity (inset: transmitted light) image (26x26 |am) and (b) in situ measurement of the total (relative) fluorescence intensity of GFP in a human B cell (with GFP-labelled GPI) for a small region of the cell membrane as a function of excitation wavelength (lines joining data points are to guide the eye only); (c) intensity merged FLIM image of a GFP labelled cell excited at 480-500 nm using the TCS. Reproduced from Journal of Physics D: Applied Physics ' Publishing Ltd

2004 lOP

automated fluorescence imaging applications such as multiwell plate assays. As a preliminary demonstration of this potential, we used the TCS to excite a range of fluorophore solutions in a home-built FLIM multiwell plate reader.^^ The TCS illuminated a multiwell plate and the resulting fluorescence was imaged using our wide-field time-domain FLIM system (with 12 gates of 400 ps width). The results of four consecutive FLIM acquisitions with four different excitation bands are shown in figure 22.18(c), together with the corresponding intensity images.

It is important to recognise the ease, simplicity and speed with which tunable excitation is achieved using the TCS compared to current methods for obtaining spatially coherent, tunable, ultrafast, visible radiation. By combining the TCS with acousto-optic tunable filter, liquid crystal or digital micromirror array technology, we anticipate that it will be possible to arbitrarily shape excitation spectra on millisecond timescales. Such flexibility may be useful for experiments in biology and elsewhere, permitting the multiplexing of more fluorescence labels, enhancing contrast between fluorophores and increasing fluorescence signal through more optimised excitation. The TCS may be applied to almost any optical instrument and is rapidly finding new applications. We have recently demonstrated its efficacy with a tandem scanning confocal microscope utilising a Nipkow disc and wide-field FLIM detection.'^^

This could provide a useful tool for rapid fluorescence imaging and FLIM/FRET of live cells. A particularly exciting aspect of the TCS is the potential for low cost devices based on compact, high average power fibre amplifier technology. ^ Figure 22.19 shows intensity and FLIM images of mouse embryo tissue stained with Troma that were obtained with a wide-field microscope using a low component cost (< £20,000) TCS pumped by a mode-locked fibre laser (IPC, model YLP-8-1060-PS). We believe that such a source could replace gas lasers for most imaging applications where tunable (visible)


radiation is required. This spectral versatility also relaxes constraints on fluorescence probes, for which excitation by argon ion laser lines will no longer be a key requirement.

Figure 22.18. (a) Fluorescence intensity and (b) fluorescence lifetime images of a fixed unstained 10 fxm section of human pancreas obtained in a wide-field time-domain FLIM system with X^x = 440-450 nm (A -small artery, I - islet of Langerhans and C - connective tissue), (c) composite wide-field fluorescence intensity and FLIM images of a multiwell plate array of fluorescent dyes excited with four different wavelength bands of 20 nm width centred on 500, 515, 570 and 610 nm respectively. The upper part shows the first time-gated fluorescence intensity image and the lower part shows the corresponding FLIM images. Reproduced fi-om Journal of Physics D: Applied Physics'^^ © 2004 lOP Publishing Ltd. (See color insert section.)

(a) (b)

Figure 22.19. Wide-field fluorescence (a) intensity and (b) lifetime images of a Troma stained section of mouse lung tissue. (See color insert section.)

514 D. S.ELSON^r^.

22.7.2. Hyperspectral FLIM instrumentation

As we have worked towards optimising the FLIM contrast obtained from tissue autofluorescence, the importance of optimising the excitation wavelength and emission spectral detection window has become apparent. As discussed in the previous section, the TCS, or multiphoton excitation, provide convenient means to address the first issue. It is straightforward to address the second by using different interference filters or dichroic beamsplitters in a fluorescence microscope but this procedure is extremely slow. In order to optimise the achievable FLIM contrast from an unknown sample, it is desirable to be able to collect the full fluorescence lifetime at each wavelength over the emission spectrum profile for each image pixel in as short a time as possible. We describe such a data acquisition as ''hyperspectral Fl!\^" Hypers pec tral imaging differs from multispectral imaging in that the latter involves resolving into only a small number of discrete spectral windows, rather than fully sampling the spectral profile. We previously reported a multispectral wide-field time-gated FLIM system that employed an image splitter (Optical Insights, Inc., Dual-View) to produce two spectrally separated replicas of the input image at the photocathode of the GOI. While this instrument was convenient to acquire lifetime images of distinct fluorophores with known spectral profiles, it is not useful to characterise tissue autofluorescence or optimise FLIM contrast.

We note that differences in the fluorescence spectra can themselves distinguish different states or types of tissue and spectrally resolved imaging of autofluorescence has been used to contrast healthy and diseased tissue. ^ ' ^ ' ^ Analysis of emission spectra can yield information concerning the distributions of fluorophores and local fluorophore environment in an analogous manner to fluorescence lifetime, although the measured lifetime may be less affected than the measured spectrum by variations in excitation or emission intensity due to intervening absorbers. In general, although changes to fluorescence spectra or decay profiles can both report changes in fluorophore electronic energy level configurations, they do not always manifest such changes to the same extent. The simultaneous measurement of the energy spectrum of the fluorescence emission and the temporal evolution of the emission process provides more information concerning the biochemical and structural properties of molecular tissue components than spectral or temporal measurements alone - and adding a further dimension to the fluorescence analysis provides increased opportunity to achieve intrinsic contrast through autofluorescence.

To date there have been few reports of fluorescence imaging instruments that acquire both spectrally and temporally resolved fluorescence data. Single point fluorescence measurement systems have been developed, however 'tnd applied to biological tissue.^^ Spectrally-resolved (hyperspectral) FLIM has recently been demonstrated in a multiphoton scanning microscope using TCSPC with an array of photon counting detectors applied to the fluorescence spectrum dispersed by a diffraction grating. ^ ^ While this system provides high quality data, it is relatively slow - with the total acquisition time being increased compared to single channel TCSPC by the number of spectrally


Figure 22.20. Generalised experimental set-up for hyperspectral FLIM microscopyLlandL2 are anamorphic optics to produce the slit illumination for single photon excitation; L3 represents the imaging optics; IS the imaging spectrograph and GOI the gated optical image intensifier.

resolved detectors, although multiple parallel TCSPC detectors could overcome this problem. Hyperspectral FLIM may be achieved at higher speeds by combining wide-field FLIM with an appropriate spectrally resolved imaging technique. Such wide-field hyperspectral imaging techniques might include approaches based on a liquid crystal tunable filter (LCTF) filter/^^ acousto-optic tunable filter (AOTF), '* a Fourier transform interferometer system^^^ or a spatial light modulator (SLM) performing Hadamard transforms. ^ ^ Of these wide-field approaches, only that using the SLM with Hadamard transform has so far been combined with FLIM, for which a frequency domain system was used. '"

An alternative approach is to implement hyperspectral "push-broom" imaging in a slit-scanning microscope, in which a line excitation (x axis) of the sample is imaged to the entrance slit of a spectrograph and a wide-field detector is placed at the output to capture the emerging (x-A.) fluorescence light image. By scanning the slit (along the y axis), the full x-y-A, data cube may be acquired. This line-scanning technique is faster than confocal point-scanning, and is photon efficient. Line-scanning microscopy also provides optical sectioning capabilities^^^ and is readily adapted for hyperspectral FLIM by using a wide-field FLIM system at the output of the spectrograph. Such an approach has previously been demonstrated for studying lanthanide chelates with lifetimes on microsecond timescales.^^^ We describe here what we believe to be the first picosecond resolved hyperspectral FLIM system based on a slit scanning microscope. Given that we can acquire a wide-field hyperspectral FLIM data stack in as little as -60 seconds, ^^^ we also believe it to be the fastest hyperspectral FLIM system reported to date. Figure 22.20 is a generalised

516 D.S.ELSON^r^L.

schematic of our experimental configuration. The pulsed source may be an ultrafast infrared laser for two photon excitation or it may be an ultrafast visible laser (TCS or frequency-doubled Ti:Sapphire laser) for single photon excitation.

For single photon excitation, our home-built system incorporates cylindrical optics to produce a line excitation on the sample in an inverted microscope and uses an imaging spectrograph (ImSpector, Specim, Finland) and wide-field FLIM system as described above. When we use the TCS for excitation, we have the potential to investigate the excitation, emission and lifetime properties of a fluorescent sample. Figure 22.21 shows an example of this TCS-excited hyperspectral FLIM system applied to a safranin-fast green preparation of a transverse section of the rhizome of a Lily-of-the-valley (Johannes Lieder, Ludwigsburg), which illustrates how the fluorescence lifetime can vary with both excitation and emission wavelength.

For multiphoton microscopy, we use a commercial multi-photon multi-foci microscope (TriMScope, La Vision Biotec GmbH, Bielefeld, Germany), ^^^ which can scan up to 64 excitation spots from a mode-locked Ti:sapphire laser to produce a line excitation (as well as being able to rapidly scan the whole field of view for wide-field detection). In line-scanning mode, we use a motorised translation stage to scan the sample through the line excitation and image the resulting fluorescence onto the entrance slit of an imaging spectrograph (SP-300i, Acton Research Corp., Acton MA). At the output of the spectrograph we use our standard wide-field time-gated FLIM system based on a GOI with CCD readout. Using a Ti:Sapphire laser that provides 100 fs pulses at 80 MHz with ~ 1 W of average output power at 780 nm, we can typically acquire a hyperspectral FLIM data set in a few minutes.

Figure 22.22 shows an example of hyperspectral-FLIM of tissue autofluorescence, illustrating how such a multidimensional data set can be analysed in different ways.

Figure 22.22(a) & (b) display FLIM images of a human carotid artery section excited at 800 nm for emission bands (a) 415-455 nm and (b) 570-635 nm respectively while

figure 22.22(c) & (d) show how the mean wavelength and spectral width, averaged over the lipid rich and fibrin rich areas indicated in

figure 22.22(a) & (b), vary during the fluorescence decay.

fe-* " f i

(a) (b) (c) Figure 22.21. FLIM images calculated from excitation-emission-lifetime data set for (a) 480 nm excitation, 490-620 nm emission band; (b) 550 nm excitation, 490-620 nm emission band and (c) 480 nm excitation, 620-835 nm emission band. (See color insert section.)


Lipid rich

iMOOpt

dfdift rich

^ 520

f 5 1 5 •35

I 510

§ ^^ ^ 500

(c)

70

1 . 60

^ 50

TO ' • O

5 30

D D

1 °

pm, ^ . . ^ o o o o <xw«a)oooc^ffP<'^ ^ '

J" U^

D L

OF arin rich

0 2000 4000 6000 8000 10000 12000 Delay (ps)

14000

, o o . o o n o o n -

^ i P i P °

1 ^^SScO000Ci()O0C!C> O <> O <> <> ^ ^

D Lipid o Fibrin

rich

rich

(d) 2000 4000 6000 8000 10000 12000 14000

Delay (ps)

Figure 22.22. Fluorescence lifetime images of carotid artery in emission bands (a) 415-455 nm and (b) 570-635 nm. (c) shows the mean wavelength and (d) shows the spectral width as a function of the delay time after excitation for the lipid rich and fibrin rich regions highlighted in (a). (See color insert section.)

22.8. CONCLUSIONS

Following the considerable impact of the application of convenient ultrafast lasers to multiphoton microscopy on biomedical imaging, it seems to us that FLIM and MDFI continue the trend in which advances in instrumentation will facilitate new discoveries - and modes of discovery - in biology and medicine. We hope we have shown the reader that fluorescence lifetime can provide intrinsic molecular contrast in unstained tissue and that the prospects for in vivo application are exciting. We believe that the capability to excite fluorophores at almost any excitation wavelength and the opportunities to extract more information from fluorescence signals by resolving with respect to lifetime, excitation and emission spectrum and also polarisation, will have a major impact on the ability to identify and exploit intrinsic contrast and on investigations of molecular biology. There the combination of new fluorescence probe technology, including genetically-expressed labels and nano-engineered devices, with new modes of interrogation and analysis, will continue to ftiel the astounding advances in this field. There is a real prospect that our ability to ask and test biological questions will cease to be limited by the availability of suitable instrumentation. Rather it is likely to be limited by our ability to analyse and comprehend the (rapidly increasing volume of) data that we collect.

518 D. S . E L S O N E J ^ X .


Funding for this research is gratefully acknowledged from the Biotechnology and Biological Sciences Research Council (BBSRC), the Engineering and Physical Sciences Research Council (EPSRC) Glaxo-SmithKline R&D Ltd, the European Community (Framework VI Integrated Project 'Integrated technologies to in vivo molecular imaging' contract number LSHG-CT-2003-503259), the Higher Education Funding Council for England (Joint Infrastructure Fund (V) Award), the Department of Trade and Industry (Beacon Award) and the Wellcome Trust (Showcase Award). RKP Benninger and PMP Lanigan acknowledge CASE QUOTA studentships from Kentech Instruments Ltd, and the EPSRC and BBSRC respectively. We also thank Damian Schimpf, Jan Siegel, Klaus Suhling, Stephen Webb and Sandrine Leveque-Fort for their helpful contributions.

22.10. REFERENCES

1. S. Andersson-Engels, J. Johansson, U. Stenram, K. Svanberg, and S. Svanberg, Malignant tumour and atherosclerotic plaque diagnosis using laser induced fluorescence, IEEE J. Quantum Electron. 26(12), 2207-2217 (1990).

2. W. Denk, J. H. Strickler, and W. W. Webb, Two-photon laser scanning fluorescence microscopy. Science 248 73-76 (1990).

3. M. E. Dickinson, E. Simbuerger, B. Zimmermann, C. W. Waters, and S. E. Eraser, Multiphoton excitation spectra in biological samples, Journal of Biomedical Optics 8(3), 329-338 (2003).

4. R. Cubeddu, D. Comelli, C. D'Andrea, P. Taroni, and G. Valentini, Time-resolved fluorescence imaging in biology and medicine, J. Phys. D-Appl. Phys. 35(9), R61-R76 (2002).

5. R. Y. Tsien, The green fluorescent protein, Annu. Rev. Biochem. 67 509-544 (1998). 6. M. Zimmer, Green fluorescent protein (GFP): Applications, structure, and related

photophysical behavior, Chem. Rev. 102(3), 759-781 (2002). 7. R. M. Clegg, O. Holub, and C. Gohlke, in: Method Enzymol, edited by G. M. and I. Parker

(Academic Press, San Diego, 2003), pp. 509-442. 8. L. Stryer, Fluorescence energy transfer as a spectroscopic ruler, Annu. Rev. Biochem. 47, 819-

846(1978). 9. G. W. Gordon, G. Berry, X. H. Liang, B. Levine, and B. Herman, Quantitative fluorescence

resonance energy transfer measurements using fluorescence microscopy, Biophys. J. 74(5), 2702-2713(1998).

10. P. I. H. Bastiaens, and A. Squire, Fluorescence lifetime imaging microscopy: spatial resolution of biochemical processes in the cell. Trends Cell Biol. 9(2), 48-52 (1999).

11. E. A. Jares-Erijman, and T. M. Jovin, FRET imaging, Nat. Biotechnol. 21(11), 1387-1395 (2003).

12. R. Richards-Kortum, and E. M. Sevick-Muraca, Quantitative optical spectroscopy for tissue diagnosis, Annual Review of Physical Chemistry 47 555-606 (1996).

13. B. B. Das, F. Liu, and R. R. Alfano, Time-resolved fluorescence and photon migration studies in biomedical and model random media. Rep. Prog. Phys. 60 227-291 (1997).

14. K. Dowling, M. J. Dayel, M. J. Lever, P. M. W. French, J. D. Hares, and A. K. L. Dymoke-Bradshaw, Fluorescence lifetime imaging with picosecond resolution for biomedical applications, Opt. Lett. 23(10), 810-812 (1998).


15. D. Toptygin, R. S. Savtchenko, N. D. Meadow, S. Roseman, and L. Brand, Effect of the solvent refractive index on the excited-state lifetime of a single tryptophan residue in a protein, J. Chem. Phys. 106(3724-3734 (2002).

16. K. Suhling, D. M. Davis, and D. Phillips, The influence of solvent viscosity on the fluorescence decay and time-resolved anisotropy of green fluorescent protein, J. Fluorescence 12(1), 91-95(2002).

17. J. R. Lakowicz, Principles of Fluorescence Spectroscopy 2nd edition (Kluwer Academic/Plenum Publishers: New York, 1999).

18. A Grinwald, On the analysis of fluorescence decay kinetics by the method of least-squares, AnaL Biochem. 59, 583-593, (1974).

19. D. R. James, and W. R. Ware, A fallacy in the interpretation of fluorescence decay parameters. Chemical Physics Letters 120(4-5), 455-459 (1985).

20. P. J. Verveer, A. Squire, and P. I. H. Bastiaens, Global analysis of fluorescence lifetime imaging microscopy data, Biophys. J. 78(4), 2127-2137 (2000).

21. O. S. Wolfbeis, in: Molecular luminescence spectroscopy, methods and applications, edited by S. G. Schulman (John Wiley, New York, 1985), pp. 167-370.

22. P. Urayama, and M. A. Mycek, in: Handbook of Biomedical Fluorescence, edited by M. A. Mycek and B. W. Pogue (Marcel Dekker, New York, 2003), pp. 211-236.

23. D. Fujimoto, K. Akiba, and N. Nakamura, Isolation and characterization of a fluorescent material in bovine achilles tendon collagen, Biochem. Biophys. Res. Commun. 76(4), 1124-1129(1977).

24. D. P. Thomhill, Separation of a series of chromophores and fluorophores present in elastin, BiochemicalJounal 14 (7) 215-219 (1975).

25. J. M. I. Maarek, L. Marcu, W. J. Snyder, and W. S. Grundfest, Time-resolved fluorescence spectra of arterial fluorescent compounds: Reconstruction with the Laguerre expansion iQc\m\q\xQ,Photochem. Photobiol. 71(2), 178-187 (2000).

26. P. Ashjian, A. Elbarbary, P. Zuk, D. A. DeUgarte, P. Benhaim, L. Marcu, and M. H. Hedrick, Noninvasive in situ evaluation of osteogenic differentiation by time-resolved laser-induced fluorescence spectroscopy. Tissue Eng. 10(3-4), 411-420 (2004).

27. D. H. Tinker, and A. L. Tappel, A Partial characterization of the major fluorophore of bovine ligamentum elastin. Connect. Tissue Res. 11(4), 309-319 (1983).

28. D. R. Eyre, M. A. Paz, and P. M. Gallop, Cross-linking in collagen and elastin, Annu. Rev. Biochem. 53, 717-748 (1984).

29. A. J. Bailey, R. G. Paul, and L. Knott, Mechanisms of maturation and ageing of collagen, Mech. Ageing. Dev. 106(1-2), 1-56 (1998).

30. Z. Deyl, K. Macek, M. Adam, and Vancikova, Studies on the Chemical Nature of Elastin Fluorescence, Biochimica et Biophysica Acta 625(2), 248-254 (1980).

31. B. Chance, B. Schoener, R. Oshino, F. Itshak, and Y. Nakase, Oxidation-reduction ratio studies of mitochondria in freeze-trapped samples. NADH and flavoprotein fluorescence signals, J. Biol. Chem. 254(11), 4764-4771 (1979).

32. F. N. Ghadially, and W. J. P. Neish, Porphyrin fluorescence of experimentally produced squamous cell carcinoma, Nature 188, 1124 (1960).

33. H. Schneckenburger, K. Konig, K. Kunzi-Rapp, C. Westphal-Frosch, and A. Ruck, Time-resolved in-vivo fluorescence of photosensitizing porphyrins, Journal of photochemistry and photobiology, B Biology 11, 143-147(1993).

34. E. Gratton, and M. Limkeman, A continuously variable frequency cross-correlation phase fluorometer with picosecond resolution, Biophys. J. 44(3), 315-324 (1983).

35. J. R. Lakowicz, and B. P. Maliwal, Construction and performance of a variable-frequency phase- modulation fluorometer. Biophysical Chemistry 21(1), 61-78 (1985).

36. P. T. C. So, T. French, W. M. Yu, K. M. Berland, C. Y. Dong, and E. Gratton, Time-resolved fluorescence microscopy using two-photon excitation, Bioimaging 3, 49-63 (1995).

37. D, V. O'Connor, and D. Phillips, Time-Correlated Single-Photon Counting (Academic Press, New York, 1984).

520 D.S.ELSON^r^x.

38. I. Bugiel, K. Konig, and H. Wabnitz, Investigation of cells by fluorescence laser scanning microscopy with subnanosecond time resolution, Lasers in the Life Sciences 3(1), 47-53 (1989).

39. W. Becker, A. Bergmann, M. A. Hink, K. Konig, K. Benndorf, and C. Biskup, Fluorescence lifetime imaging by time-correlated single-photon counting, Microsc. Res. Tech. 63(1), 58-66 (2004).

40. Y. L. Zhang, S. A. Soper, L. R. Middendorf, J. A. Wurm, R. Erdmann, and M. Wahl, Simple near-infrared time-correlated single photon counting instrument with a pulsed diode laser and avalanche photodiode for time-resolved measurements in scanning applications, Appl. Spectrosc. 53(5), 497-504 (1999).

41 . E. P. Buurman, R. Sanders, A. Draaijer, H. C. Gerritsen, J. J. F. Vanveen, P. M. Houpt, and Y. K. Levine, Fluorescence lifetime imaging using a confocal laser scanning microscope, ScanmngU{?>), 155-159(1992).

42. Nikon LIMO system (www.nikon-instruments.com/) 43. H. C. Gerritsen, M. A. H. Asselbergs, A. V. Agronskaia, and W. G. J. H. M. Van Sark,

Fluorescence lifetime imaging in scanning microscopes: acquisition speed, photon economy and hfetime resolution, J. Microsc. 206(3), 218-224 (2002).

44. C. Y. Dong, P. T. C. So, T. French, and E. Gratton, Fluorescence lifetime imaging by asynchronous pump-probe microscopy, Biophys. J. 69(6), 2234-2242 (1995).

45. K. Carlsson, and A. Liljeborg, Simultaneous confocal lifetime imaging of multiple fluorophores using the intensity-modulated multiple-wavelength scanning (IMS) technique, J. Microsc. 191(2), 119-127(1998).

46. K. Kemnitz, L. Pfeifer, and M. R. Ainbund, Detector for multichannel spectroscopy and fluorescence lifetime imaging on the picosecond timescale, Nucl. Instrum. Methods Phys. Res. Sect. A-Accei Spectrom. Dect. Assoc. Equip. 387, 86-87 (1997).

47. M. KoUner, and J. Wolfrum, How many photons are necessary for fluorescence-lifetime measurements? Chemical Physics Letters 200(1-2), 199-204 (1992).

48. E. Gratton, S. Breusegem, J. Sutin, and Q. Q. Ruan, Fluorescence lifetime imaging for the two-photon microscope: Time-domain and frequency-domain methods, Journal of Biomedical 0/7//C5 8(3), 381-390 (2003).

49. R. V. Krishnan, H. Saitoh, H. Terada, V. E. Centonze, and B. Herman, Development of a multiphoton fluorescence lifetime imaging microscopy system using a streak camera. Rev. Sci. Instrum. 74(5), 2714-2721 (2003).

50. J. Requejo-Isidro, J. McGinty, 1. Munro, D. S. Elson, N. Galletly, M. J. Lever, M. A. A. Neil, G. W. H. Stamp, P. M. W. French, P. A. Kellet, J. D. Hares, and A. K. L. Dymoke-Bradshaw, High-speed wide-field time-gated endoscopic fluorescence lifetime imaging, Opt. Lett. 29(19), 2249-2251(2004).

51. A. V. Agronskaia, L. Tertoolen, and H. C. Gerritsen, High frame rate fluorescence lifetime imaging, J. Phys. D-Appl. Phys. 36 1655-1662 (2003).

52. E. Gratton, B. Feddersen, and M. van de Ven, Parallel acquisition of fluorescence decay using array detectors, in Time-resolved laser spectroscopy in biochemistry 7/21-25 (SPIE vol. 1204, 1990).

53. C. G. Morgan, A. C. Mitchell, and J. G. Murray, Fluorescence decay time imaging using an imaging photon detector with a radiofrequency photon correlation system, in Time-Resolved Laser Spectroscopy in Biochemistry II19S-S01 (SPIE \o\. 1204, 1990).

54. J. R. Lakowicz, and K. W. Bemdt, Lifetime-selective fluorescence imaging using an rf phase-sensitive camera. Rev. Sci. Instrum. 62(7), 1727-1734 (1991).

55. T. W. J. Gadella, T. M. Jovin, and R. M. Clegg, Fluorescence lifetime imaging microscopy (FLIM) - spatial resolution of structures on the nanosecond timescale, Biophysical Chemistry 48 221-239(1993).

56. X. F. Wang, T. Uchida, D. M. Coleman, and S. Minami, A two-dimensional fluorescence lifetime imaging system using a gated image intensifier, Appl. Spectrosc. 45(3), 360-366 (1991).

57. J. D. Hares, Advances in sub-nanosecond shutter tube technology and applications in plasma physics, in X Rays from laser plasmas 165-170 (SPIE vol. 831, 1987).


58. A. D. Scully, A. J. Mac Robert, S. Botchway, P. O'Neill, A. W. Parker, R- ^ Ostler, and D. Phillips, Development of a laser-based fluorescence microscope with subn^'^^^^^^"^ ^^^^ resolution, J. Fluorescence 6(2), 119-125 (1996).

59. K. Dowling, M. J. Dayel, M. J. Lever, P. M. W. French, J. D. Hares, and A. K. L. Dymoke-Bradshaw, Fluorescence lifetime imaging with picosecond resolution for biomedical applications. Opt. Lett. 23(10), 810-812 (1998).

60. A, Squire, P. J. Verveer, and P. L H. Bastiaens, Multiple frequency fluorescence lifetime imaging microscopy, 7. Microsc. 197(2), 136-149(2000).

61. J. R. Lakowicz, H. Szmacinski, K. Nowaczyk, K. W. Bemdt, and M. Johnson, Fluorescence Lifetime Imaging, Anal. Biochem. 202(2), 316-330 (1992).

62. J. Mizeret, T. Stepinac, M. Hansroul, A. Studzinski, H. van den Bergh, and G. Wagnieres, Instrumentation for real-time fluorescence lifetime imaging in endoscopy, Rev. Set Instrum. 70(12), 4689-4701 (1999).

63. O. Holub, M. J. Seufferheld, C. Gohlke, Govindjee, and R. M. Clegg, Fluorescence lifetime imaging (fli) in real-time - a new technique in photosynthesis research, Photosynthetica 38(4), 581-599(2000).

64. A. V. Agronskaia, L. Tertoolen, and H. C. Gerritsen, High frame rate fluorescence lifetime imaging, y. Phys. D-Appl. Phys. 36 1655-1662 (2003).

65. C. Y. Dong, C. Buehler, P. T. C. So, T. French, and E. Gratton, Implementation of intensity-modulated laser diodes in time-resolved, pump-probe fluorescence microscopy, Appl. Optics 40(7), 1109-1115(2001).

66. M. J. Booth, and T. Wilson, Low-cost, frequency-domain, fluorescence lifetime confocal microscopy, J. Microsc.-Oxf. 214 36-42 (2004).

67. M. Bohmer, F. Pampaloni, M. Wahl, H. J. Rahn, R. Erdmann, and J. Enderlein, Time-resolved confocal scanning device for ultrasensitive fluorescence detection. Rev. Sci. Instrum. 72(11), 4145-4152(2001).

68. D. S. Elson, J. Siegel, S. E. D. Webb, S. Leveque-Fort, M. J. Lever, P. M. W. French, K. Lauritsen, M. Wahl, and R. Erdmann, Fluorescence lifetime system for microscopy and multiwell plate imaging with a blue picosecond diode laser. Opt. Lett. 27(16), 1409-1411 (2002).

69. C. D. McGuinness, K. Sagoo, D. McLoskey, and D. J. S. Birch, A new sub-nanosecond LED at 280 nm: Application to protein fluorescence, Meas. Sci. Technol. 15(11), L19-L22 (2004).

70. L. K. van Geest, and K. W. J. Stoop, FLIM on a wide field fluorescence microscope, Lett. Pept. Sci. 10(5-6), 501-510 (2003).

71. LIFA system, www.lambert-instruments.com 72. P. Herman, B. P. Maliwal, H.-J. Lin, and J. R. Lakowicz, Frequency-domain fluorescence

microscopy with the LED as a light source, J. Microsc. 203(2), 176-181 (2001). 73. A. C. Mitchell, J. E. Wall, J. G. Murray, and C. G. Morgan, Direct modulation of the effective

sensitivity of a CCD detector: a new approach to time-resolved fluorescence imaging, J. Microsc. 206(3), 225-232 (2002).

74. A. C. Mitchell, J. E. Wall, J. G. Murray, and C. G. Morgan, Measurement of nanosecond time-resolved fluorescence with a directly gated interline CCD camera, /. Microsc. 206(3), 233-238 (2002).

75. R. M. Ballew, and J. N. Demas, An error analysis of the rapid lifetime determination method for the evaluation of single exponential decays. Anal. Chem. 61(1), 30-33 (1989).

76. P. C. Schneider, and R. M. Clegg, Rapid acquisition analysis and display of fluorescence lifetime-resolved images for real-time applications, Rev. Sci. Instrum. 68(11), 4107-4119 (1997).

77. J. Wlodarczyk, and B. Kierdaszuk, Interpretation of fluorescence decays using a power-like model, Biophys. J. 85(1), 589-598 (2003).

78. J. A. Jo, Q. Y. Fang, T. Papaioannou, and L. Marcu, Fast model-free deconvolution of fluorescence decay for analysis of biological systems, Journal of Biomedical Optics 9(4), 743-752 (2004).

522 D.S.ELSON^r^JL.

79. K. C. B. Lee, J. Siegel, S. E. D. Webb, S. Leveque-Fort, M. J. Cole, R. Jones, K. Dowling, M. J. Lever, and P. M. W. French, Application of the stretched exponential function to fluorescence lifetime imaging, Biophys. J. 81(3), 1265-1274 (2001).

80. J. Laherrere, and D. Somette, Stretched exponential distributions in nature and economy: "fat tails" with characteristic scales, Eur. Phys. J. B 2(4), 525-539 (1998).

8L F. Alvarez, A. Alegria, and J. Colmenero, Relationship between the time-domain Kohlrausch-Williams-Watts and frequency-domain Havriliak-Negami relaxation functions, Phys. Rev. B 44(14), 7306-7312 (1991).

82. J. Siegel, D. S. Elson, S. E. D. Webb, K. C. B. Lee, A. Vlanclas, G. L. Gambaruto, S. Leveque-Fort, M. J. Lever, P. J. Tadrous, G. W. H. Stamp, A. L. Wallace, A. Sandison, T. F. Watson, F. Alvarez, and P. M. W. French, Studying biological tissue with fluorescence lifetime imaging: microscopy, endoscopy, and complex decay profiles, Appl. Optics 42(16), 2995-3004 (2003).

83. S. E. D. Webb, Y. Gu, S. Leveque-Fort, J. Siegel, M. J. Cole, K. Dowling, R. Jones, P. M. W. French, M. A. A. Neil, R. Juskaitis, L. O. D. Sucharov, T. Wilson, and M. J. Lever, A wide-field time-domain fluorescence lifetime imaging microscope with optical sectioning. Rev. Sci. Instrum. 73(4), 1898-1907 (2002).

84. www.kentech.co.uk 85. D. S. Elson, L Munro, J. Requejo-Isidro, J. McGinty, C. Dunsby, N. Galletly, G. W. Stamp,

M. A. A. Neil, M. J. Lever, P. A. Kellett, A. Dymoke-Bradshaw, J. Hares, and P. M. W. French, Real-time time-domain fluorescence lifetime imaging including single-shot acquisition with a segmented optical image intensifier. New J. Phys. 6 art. no.-180 (2004).

86. E. Waddell, Y. Wang, W. Stryjewski, S. McWhorter, A. C. Henry, D. Evans, R. L. McCarley, and S. A. Soper, High-resolution near-infrared imaging of DNA microarrays with time-resolved acquisition of fluorescence lifetimes. Anal. Chem. 72(24), 5907-5917 (2000).

87. D. S. Elson, Development of ultrafast laser technology and its application to fluorescence lifetime imaging, thesis, (University of London, 2003).

88. Q. Y. Fang, T. Papaioannou, J. A. Jo, R. Vaitha, K. Shastry, and L. Marcu, Time-domain laser-induced fluorescence spectroscopy apparatus for clinical diagnostics. Rev. Sci. Instrum. 75(1), 151-162(2004).

89. M. J. Cole, Fluorescence lifetime imaging for biomedical applications, thesis, (University of London, 2000).

90. S. E. D. Webb, Development and application ofwidefieldjluorescence lifetime imaging, thesis, (University of London, 2003).

91. C. B. Talbot, J. Requejo-Isidro, I. Munro, D. S. Elson, A. Sandison, A. Wallace, M. A. A. Neil, P. M. W. French, and M. J. Lever, Fluorescence lifetime imaging applied to articular cartilage to monitor the development of disease, submitted to Osteoarthr. Cartilage. (2005).

92. H. Nagase, and J. F. Woessner, in: Joint Cartilage Degredation, edited by J. F. Woessner (Marcel Dekker, Basel, 1993), pp. 159-186.

93. A. R. Poole, J. S. Mort, and P. J. Roughley, in: Joint Cartilage Degradation, edited by J. F. Woessner (Marcel Dekker, Basel, 1993), pp. 225-260.

94. D. R. Sell, and V. M. Monnier, Structure elucidation of a senescence cross-link from human extracellular-matrix - implication of pentoses in the aging process, J. Biol. Chem. 264(36), 21597-21602(1989).

95. H. K. Gahunia, R. Vieth, and K. P. H. Pritzker, Novel fluorescent compound (DDP) in calf, rabbit, and human articular cartilage and synovial fluid, J. Rheumatol. 29(1), 154-160 (2002).

96. L. Kessel, S. Kalinin, R. H. Nagaraj, M. Larsen, and L. B. A. Johansson, Time-resolved and steady-state fluorescence spectroscopic studies of the human lens with comparison to argpyrimidine, pentosidine and 3-oh-kynurenine, Photochem. Photobiol. 76(5), 549-554 (2002).

97. J. J. Baraga, R. P. Rava, M. Fitzmaurice, L. L. Tong, P. Taroni, C. Kittrell, and M. S. Feld, Characterization of the fluorescent morphological structures in human arterial-wall using ultraviolet-excited microspectrofluorimetry, Atherosclerosis 88(1), 1-14 (1991).

98. A. L. Alexander, C. M. C. Davenport, and A. F. Gmitro, Comparison of illumination wavelengths for detection of atherosclerosis by optical fluorescence spectroscopy. Optical Engineering 33( 1), 167-174 (1994).


99. F. Bosshart, U. Utzinger, O. M. Hess, J. Wyser, A. Mueller, J. Schneider, P. Niederer, M. Anliker, and H. P. Krayenbuehl, Fluorescence spectroscopy for identification of atherosclerotic tissue, Cardiovasc. Res. 26(6), 620-625 (1992).

100. W. D. Yan, M. Perk, A. Chagpar, Y. Wen, S. Stratoff, W. J. Schneider, B. L Jugdutt, J. Tulip, and A. Lucas, Laser-induced fluorescence .3. Quantitative-analysis of atherosclerotic plaque content, Lasers Surg. Med. 16(2), 164-178 (1995).

101. '^' G. Filippidis, G. Zacharakis, A. Katsamouris, A. Giannoukas, and T. G. Papazoglou, Single and double wavelength excitation of laser-induced fluorescence of normal and atherosclerotic peripheral vascular tissue,/. Photochem. Photobiol B-Biol 56(2-3), 163-171 (2000).

102. ^^ L. I. Deckelbaum, S. P. Desai, C. Kim, and J. J. Scott, Evaluation of a fluorescence feedback-system for guidance of laser angioplasty. Lasers Surg. Med. 16(3), 226-234 (1995).

103. ^^ A. J. Morguet, R. E. Gabriel, A. B. Buchwald, G. S. Werner, R. Nyga, and H. Kreuzer, Single-laser approach for fluorescence guidance of excimer laser angioplasty at 308 nm: Evaluation in vitro and during coronary angioplasty, Lasers Surg. Med. 20(4), 382-393 (1997).

104. A. J. Morguet, B. Korber, B. Abel, H. Hippler, V. Wiegand, and H. Kreuzer, Autofluorescence spectroscopy using a XeCl excimer-laser system for simultaneous plaque ablation and fluorescence excitation, Lasers Surg. Med. 14(3), 238-248 (1994).

105. R. T. Strebel, U. Utzinger, M. Peltola, J. Schneider, P. F. Niederer, and O. M. Hess, Excimer laser spectroscopy: Influence of tissue ablation on vessel wall fluorescence, J. Laser Appl. 10(1), 34-40 (1998).

106. S. Warren, K. Pope, Y. Yazdi, A. J. Welch, S. Thomsen, A. L. Johnston, M. J. Davis, and R. Richards-Kortum, Combined ultrasound and fluorescence spectroscopy for physicochemical imaging of atherosclerosis, IEEE Trans. Biomed. Eng. 42(2), 121-132 (1995).

107. M. Stavridi, V. Z. Marmarelis, and W. S. Grundfest, Spectro-temporal studies of Xe-Cl excimer laser-induced arterial-wall fluorescence, Med. Eng. Phys. 17(8), 595-601 (1995).

108. L. Marcu, M. C. Fishbein, J. M. Maarek, and W. S. Grundfest, Discrimination of human coronary artery atherosclerotic lipid-rich lesions by time-resolved laser-induced fluorescence s^QCixoscoxfy, Arteriosclerosis, Thrombosis & Vascular Biology 21(1), 1244-50(2001).

109. J. M. I. Maarek, L. Marcu, M. C. Fishbein, and W. S. Grundfest, Time-resolved fluorescence of human aortic wall: Use for improved identification of atherosclerotic lesions. Lasers Surg. MeJ. 27(3), 241-254 (2000).

110. C. B. Talbot, P. A. A. De Beule, C. Dunsby, J. Requejo-Isidro, D. S. Elson, M. A. A. Neil, P. M. W. French, and M. J. Lever, Fluorescence lifetime imaging of atherosclerotic plaques, manuscript in preparation.

111. R. Cubeddu, A. Pifferi, P. Taroni, A. Torricelli, G. Valentini, F. Rinaldi, and E. Sorbellini, Fluorescence lifetime imaging: An application to the detection of skin tumors, IEEE J. Sel. Top. Quantum Electron. 5(4), 923-929 (1999).

112. Q. Peng, T. Warloe, K. Berg, J. Moan, M. Kongshaug, K. E. Giercksky, and J. M. Nesland, 5-aminolevulinic acid-based photodynamic therapy - clinical research and future challenges, Cancer 79(12), 2282-2308 (1997).

113. K. M. Hanson, N. P. Barry, M. J. Behne, T. M. Mauro, E. Gratton, and R. M. Clegg, Two-photon fluorescence lifetime imaging of the skin's stratum comeum pH gradient, Biophys. J. 82(1), 2415 (2002).

114. R. K. P. Benninger, O. Hofmann, J. McGinty, J. Requejo-Isidro, I. Munro, M. A. A. Neil, A. J. deMello, and P. M. W. French, Time-resolved fluorescence imaging of solvent interactions in microfluidic devices, accepted for publication in Opt. Express (2005).

115. P. D. Devries, and A. A. Khan, An efficient technique for analyzing deep level transient spectroscopy data, J. Electron. Mater. 18(4), 543-547 (1989).

116. S. P. Chan, Z. J. Fuller, J. N. Demas, and B. A. DeGraff, Optimized gating scheme for rapid lifetime determinations of single-exponential luminescence lifetimes, Anal. Chem. 73(18), 4486-4490(2001).

117. R. M. Ballew, and J. N. Demas, Error analysis of the rapid lifetime determination method for single exponential decays with a non-zero baseHne, Analytica Chimica Acta 245 121-127 (1991).

118. K. K. Sharman, A. Periasamy, H. Ashworth, J. N. Demas, and N. H. Snow, Error analysis of the rapid lifetime determination method for double-exponential decays and new windowing schemes, ^«a/. Chem. 71 947-952 (1999).

524 D.S.ELSON£r.iL.

119. P. E. Young, J. D. Hares, J. D. Kilkenny, D. W. Phillion, and E. M. Campbell, 4-frame gated optical imager with 120-ps resolution, Rev. Set Instrum. 59(8), 1457-1460 (1988).

120.1. Munro'N. Galletly, J. McGinty, D. S. Elson, J. Requejo-Isidro, C. Dunsby, M. A. A. Neil, M. J. Lever, G. W. Stamp and P. M. W. French, Towards the clinical application of time-domain fluorescence lifetime imaging, accepted for publication in J. Biomed Opt. (2005).

121. T. A. Birks, W. J. Wadsworth, and P. S. Russell, Supercontinuum generation in tapered fibers. Opt. Lett. 25(19), 1415-1417 (2000).

122. J. K. Ranka, R. S. Windeler, and A. J. Stentz, Visible continuum generation in air-silica microstructure optical fibers with anomalous dispersion at 800 nm. Opt. Lett. 25(1), 25-27 (2000).

123. H. Birk and R. Storz, United States patent no. 6,611,643 (2001). 124. G. McConnell, Confocal laser scanning fluorescence microscopy with a visible continuum

source, Opt. Express 12(13), 2844-2850 (2004). 125. J. E. Jureller, N. F. Scherer, and T. A. Birks, in: Ultrafast Phenomena XIII, (Springer-Verlag,

Berlin, 2003), pp. 2844-2850. 126. C. Dunsby, P. M. P. Lanigan, J. McGinty, D. S. Elson, J. Requejo-Isidro, I. Munro, N.

Galletly, F. McCann, B. Treanor, B. Onfelt, D. M. Davis, M. A. A. Neil, and P. M. W. French, An electronically tunable ultrafast laser source applied to fluorescence imaging and fluorescence lifetime imaging microscopy, J. Phys. D-Appl. Phys. 37 3296-3303 (2004).

127. D. M. Grant, D. S. Elson, D. Schimpf, C. Dunsby, J. Requejo-Isidro, I. Munro, M. A. A. Neil, P. M. W. French, E. Nye, G. W. Stamp, and P. Courtney, Wide-field optically-sectioned fluorescence lifetime imaging using a Nipkow disk microscope and a tunable ultrafast continuum excitation source, accepted for publication in Opt. Lett. (2005).

128. P. A. Champert, S. V. Popov, and J. R. Taylor, Generation of multiwatt, broadband continua in holey fibers, Opt. Lett. 27(2), 122-124 (2002).

129. R. Drezek, C. Brookner, I. Pavlova, I. Boiko, A. Malpica, R. Lotan, M. Follen, and R. Richards-Kortum, Autofluorescence microscopy of fresh cervical-tissue sections reveals alterations in tissue biochemistry with dysplasia, Photochem. Photobiol. 73(6), 636-641 (2001).

130. G. A. Wagnieres, W. M. Star, and B. C. Wilson, In vivo fluorescence spectroscopy and imaging for oncological applications, Photochem. Photobiol. 68(5), 603-632 (1998).

131. R. R. Alfano, and Y. L. Yang, Stokes shift emission spectroscopy of human tissue and key biomolecules, IEEE J. Sel. Top. Quantum Electron. 9(2), 148-153 (2003).

132. D. K. Bird, K. W. Eliceiri, C. H. Fan, and J. G. White, Simultaneous two-photon spectral and lifetime fluorescence microscopy, Appl. Optics 43(27), 5173-5182 (2004).

133. D. L. Farkas, C. W. Du, G. W. Fisher, C. Lau, W. H. Niu, E. S. Wachman, and R. M. Levenson, Non-invasive image acquisition and advanced processing in optical bioimaging, Comput. Med. Imaging Graph. 22(2), 89-102 (1998).

134. E. S. Wachman, W. H. Niu, and D. L. Farkas, Imaging acousto-optic tunable filter with 0.35-micrometer spafial resolution, Appl. Optics 35(25), 5220-5226 (1996).

135. Z. Malik, D. Cabib, R. A. Buckwald, A. Talmi, Y. Garini, and S. G. Lipson, Fourier transform mukipixel spectroscopy for quantitative cytology, 7. Microsc.-Oxf. 182 133-140 (1996).

136. Q. S. Hanley, P. J. Verveer, and T. M. Jovin, Spectral imaging in a programmable array microscope by Hadamard transform fluorescence spectroscopy, Appl. Spectrosc. 53(1), 1-10 (1999).

137. Q. S. Hanley, D. J. Amdt-Jovin, and T. M. Jovin, Spectrally resolved fluorescence lifetime imaging microscopy,^/?/?/. Spectrosc. 56(2), 155-166 (2002).

138. T. Wilson, Optical aspects of confocal microscopy, in: Confocal Microscopy, edited by T. Wilson (Academic Press Ltd., London, 1990).

139. G. Vereb, E. Jares-Erijman, P. R. Selvin, and T. M. Jovin, Temporally and spectrally resolved imaging microscopy of lanthanide chelates, Biophys. J. 74(5), 2210-2222 (1998).

140. P. A. A. De Beule, C. B. Talbot, R. K. P. Benninger, J. Requejo-Isidro, D. S. Elson, C. Dunsby, I. Munro, A. Sandison, N. Sofat, H. Nagase, M. J. Lever, M. A. A. Neil, and P. M. W. French, One and two photon hyperspectral fluorescence lifetime imaging, in Focus on Microscopy (Jena, Germany, 2005).

141. www.lavisionbiotec.com

MULTIPLEXED FLUORESCENCE DETECTION FOR DNA SEQUENCING

Li Zhu and Steven A. Soper*

23.1. BACKGROUND AND RELEVANCE

The completion of the Human Genome Project (HGP) in 2003 is believed to be a significant milestone in the scientific world and represents one of the greatest achievements of humanity. It has provided a wealth of information including the number and average size of human genes, the fraction of the genome that codes for proteins, and the degree of sequence similarity, both among humans and compared with other organisms.^' ^ The availability of this information greatly facilitates the identification and isolation of genes that contribute to many human diseases, provides probes that can be used in genetic testing, diagnosis of diseases, and drug development and offers important information about many basic cellular processes as well.

The Human Genome Project, with goals of identifying all of the approximate 36,000 genes and determining the primary structure of the entire human genome comprised of its 3 billion base pairs, was greatly spawned by new technologies evolving over the past 15 years. Tremendous improvements have occurred in every field related to DNA sequencing, including developments in electrophoretic separation methods, instrumentation for fluorescence detection, DNA purification and cloning methods, and computational methods for the analysis of sequences. The successful completion of the Human Genome Project, however, does not signify an end to further pursuing novel techniques for providing sequence information for improving the throughput and reducing the cost. To the contrary, the number of known sequences increases the need for even more sequence data for verification (comparative genomics) and diagnostic purposes.^' " This chapter focuses primarily on advances in using fluorescence detection for DNA sequencing, including the development of fluorescent labeling dyes, various

* Department of Chemistry, Louisiana State University, Baton Rouge, LA, USA 70803

525

526 L. ZHU and S.A. SOPER

fluorescence detection strategies, as well as instrumental formats. Other basics associated with DNA sequencing will be described briefly.

23.1.1. What Is DNA Sequencing?

Located inside the nucleus of each cell of any living organism, Deoxyribonucleic Acid (DNA) carries genetic instructions, which consist of a master code that directs the building of all cellular structures and functions. This code is written in a language using only four genetic letters, which represent four different nucleotide bases. Adenine (A), Cytosine (C), Guanine (G), and Thymine (T). DNA is a well-known macromolecule that consists of many nucleotides as the basic building blocks. These building blocks form a long continuous chain, which is tightly coiled together as a double helix. These double strands are organized into chromosomes and compressed by associated proteins. For humans, a total of 3 billion nucleotides are ordered and arranged within 23 pairs of chromosomes. Each chromosome contains many genes, pieces of DNA that contain information for directing protein synthesis in a cell. They are the basic functional and physical units of heredity and comprise about 2% of the human genome; the remainder consists of non-coding regions, whose ftinctions may include providing chromosomal structural integrity and regulating where, when, and in what quantity proteins are made. Simply put, it is the order of the three billion A, C, G and T bases arranged along the chromosomal DNA, which spell out who we are and how our cells operate.

23.7. /. 1. Chemical Nature of DNA

The structure of DNA can be considered at three levels of increasing complexity: the primary, secondary, and tertiary structures. The primary structure of DNA consists of a large number of repeating units joined together by phosphodiester linkages (see Figure la). Each unit, a deoxyribonucleotide (simply called nucleotide), contains a deoxyribose sugar unit, a phosphate, and a nitrogen-containing base. There are four different types of nucleotides in DNA, differing in their nitrogenous base. Each is given a one-letter abbreviation; A for adenine, G for guanine, C for cytosine and T for thymine. The nitrogenous bases always form a covalent bond with the 1' -carbon atom of the deoxyribose sugar. All nucleotides of DNA are joined together by phosphodiester bonds that connect the 5'-phosphate group of one nucleotide to the 3'-carbon atom of the next. Therefore, in the primary structure, DNA exists as a biopolymer with a sugar-phosphate linkage serving as its backbone.

The secondary structure refers to DNA's stable three-dimensional double helical configuration. In 1953, James Watson and Francis Crick discovered that DNA has a double-helix structure, a spiral consisting of two DNA strands wound around each other. The alternating sugar-phosphate backbones orient on the outside of the helix and the bases project into the interior of the molecule (Figure lb). The bases on opposing strands are paired via hydrogen bonds such that adenine always pairs with thymine and guanine always pairs with cytosine. The specificity of the base pairing indicates that the

MULTIPLEXED FLUORESCECNE DETECTION FOR DNA SEQUENCING 527

t'ytttsiiic (C)

(;ii4ti(ne (€>

romptfur-geiierattif image of

if T i

0=-.:=-.:I< 1

Figure 23.1. (a) Primary structure of DNA. Four different nucleotides (A, C, G, and T) are linked together by phosphodiester bonds, (b) The Watson-Crick model of the DNA double helix structure. The computer-generated image shows the double-helical nature of DNA and also an expanded region of the DNA showing hydrogen bonding between nucleotide bases based on the Watson-Crick base-pairing rules (A to T and G to C).

two polynucleotide strands of a DNA molecule are complementary. Spiraling of the nucleotide strands generates major and minor grooves in the helix, which is important for the binding of proteins that regulate the expression of the genetic information.

The tertiary structure of DNA refers to the complex packing arrangements of double-stranded DNA in chromosomes. As already described, DNA exists in the form of very long molecules. Therefore, they must be tightly packed to fit into the nucleus in which it resides.

2 3 J. 1.2. Shotgun Sequencing Strategies for Whole Genome Processing

The main obstacle to sequencing an entire genome is their immense size. For example, the 3-billion base pairs of the human genome are distributed among 23 pairs of chromosomes, with the smallest chromosome containing 50 million bases and the largest 250 million bases. In spite of the refinements of the sequencing methods, it is still impossible to sequence from one end of a chromosome and continue through to the other end, because the length of the longest contiguous DNA stretch that can be sequenced is limited by the length that can be read by gel electrophoresis. The shotgun sequencing method has been widely adopted in production-scale sequencing to address this problem. In this method, DNA is first cut into smaller more manageable pieces. Each of


these fragments is then sequenced individually and finally puzzled together to create the original contiguous sequence. There are two major shotgun strategies for large genome sequencing. One is named Ordered Shotgun Sequencing (OSS) or map-based sequencing.^ It involves an initial mapping step that basically involves placing landmarks (Sequence Tagged Sites, STS) throughout the genome. The primary steps involved in the sequencing process using OSS are depicted in the flow chart shown in Figure 2. The genome is first broken into pieces of approximately 100 to 500 kilobase pairs (kbp), which are then cloned into Yeast Artificial Chromosomes (YACs) or Bacterial Artificial Chromosomes (BACs) to yield a library containing many copies of each fragment of the genome. These large-insert clones are ordered by STS mapping or by fingerprinting to result in a minimally overlapping tiling path, which covers the whole genome. Each clone is then individually sequenced using a shotgun strategy. The BAC or YAC clones are sheared into fragments ranging in size from 1000 to 2000 bp followed by sub-cloning into a single stranded M13 bacteriophage, to produce high quality single stranded DNA. These small-insert clones, with an appropriate size for actual sequencing (-1000 bp), are then sequenced typically by using Sanger dideoxy-chain termination methods that will be described in the next section. Following sequencing, these small-insert clones are finally reassembled into a contiguous strand with the final gap closure accomplished by directed reads.

High Ri^oiulion I'byskxl Maip

IkoMlion ofC'liromoitomal DNA

K«niiom ilydrodynuntie Shi'iirifts (100-500 kbp)

Kitndoro C'lcminR into U\('%

C%ngx(h$kfa|»)

Sobcloninft into MlS vectors (Pruiiurc <i<i icqucncing tcniptoteii)

Stf<|ttenciiiK Unci* of.VI13 ( loot'*

Fiiitng (i»px to Cumpktt' 8gt|titfO€c of BAC Clotic

Figure 23.2. Processing flow chart of ordered shotgun sequencing of DNA.


Another strategy is called Whole Genome Shotgun Sequencing (WGSS). The whole genome of an organism is fragmented randomly without a prior mapping step. Small-insert clones are prepared directly from these DNA segments and then sequenced, often from both ends 7 Using this method, computer assembly of sequence reads is more demanding than in the OSS strategy due to the lack of positional information. This method was adapted for whole human genome sequencing by Celera Genomics .

23.1.1.3. Sanger Chain Termination for Determining the Primary Structure Of DNA

The first methods for sequencing DNA to determine its primary structure, the order of the four bases along the backbone, were developed in 1970s. Allan Maxam and Walter Gilbert developed a method based on the chemical cleavage of DNA^ while Frederick Sanger and his colleagues created the dideoxy sequencing method based on the elongation (polymerization) of DNA.^ The Sanger method has served as the cornerstone for most genome sequencing efforts since 1977 and dominates today's production sequencing due to its ability to be easily automated. Therefore, this method will be described in this section.

O ~S»-~-,-5

,!i-y!,

ftipho^phati^ {tt«IN 11*1

TT ^^ y M IP

a! H'

y >H \\'

>\\ft' . M i l l '

MAMi.M%.Ui

Figure 23.3. Schematic diagrams of Sanger chain-termination DNA-sequencing. (a) Chemical structures of a dNTP and ddNTP. The latter lacks a hydroxyl group on the 3' position of the deoxyribose sugar, (b) Steps involved in the Sanger sequencing method of DNA.


The Sanger method of DNA sequencing, which is also called the dideoxy method, is based on the process of replication. The sequencing starts with an Ml 3 clone that contains an insert of foreign DNA to be sequenced. This single strand DNA fragment is used as a template to produce a series of complementary DNA molecules by using a polymerase enzyme. The Sanger sequencing reaction protocol is depicted in Figure 3. Four specially designed DNA nucleotide analogs, dideoxyribonucleoside triphosphates (ddNTP) are used. They are ddATP, ddCTP, ddGTP, ddTTP, which are identical to deoxyribonucleosides (dNTPs) except that they have an -H instead of an -OH group on the 3'-carbon position of the deoxyribose sugar. The reaction is carried out by adding to the template DNA, a primer, which is a short polynucleotide with a complementary sequence to a certain 3' position of the template DNA, a polymerase enzyme and a mixture of all four dNTPs with one of the four potential ddNTPs. Either one of the ddNTPs or the primer is tagged with a label, typically with a fluorescent label, so that it can be detected. The primer anneals to a complementary site on the template DNA providing a 3' -OH group to initiate the DNA synthesis. The polymerase enzyme assembles dNTP into a new strand of DNA from the primer, forming a complementary strand to its template. Each incoming nucleotide is selected by virtue of Watson-Crick base pairing rules. It uses at random either a dNTP or ddNTP to introduce a base in the newly synthesized strand. In the reaction mixture, dNTP is present in a much larger quantity than that of ddNTP (~1%). Therefore, dNTP is incorporated most often, allowing DNA synthesis to continue. However, when ddNTP is occasionally incorporated into the strand, synthesis of the growing DNA chain is terminated since there is no 3'-OH group on ddNTP to form a phosphodiester bond with another incoming nucleotide. This termination occurs randomly at different positions in different copies, producing a nested set of DNA fragments of different length, each ending with the same base. Equivalent reactions take place in four tubes, each ending in a particular nucleotide base. Thus, new DNA chains of all possible lengths that are complementary to the unknown sequence are generated. The reaction is carried out for many cycles (usually 20 - 50) under an appropriate temperature sequence that allows denaturation, primer annealing and extension occurring for a short period of time in each cycle to achieve complementary fragment amplification. By correlating the length of the terminated fragments with the identity of the terminating base that was present in the reaction, one can determine the order of the nucleotide bases through the size of the nested fragments and, hence, the corresponding nucleotide sequence. It should be noted that the sequence obtained is not that of the target DNA but that of its complement. After the completion of sequencing reactions, the products are loaded onto an electrophoresis gel, size fractionated and finally sequences are read out from the gel.

23.1.2. Gel Electrophoresis for DNA Sequencing

Electrophoresis involves the application of a high voltage across a conductive medium, within which charged molecules move with


distinguishable migration rates to achieve separations. Differences in shape, size, and overall charge of the solute molecules result in characteristic electrophoretic mobilities, providing the basis of the electrophoretic separation. As described previously, DNA is a biopolymer that consists of many repeating units called nucleotides. The electrophoretic mobility in free solution is typically dependent on the ratio of charge to molecular size and configuration. Each nucleotide carries a constant net negative charge residing on the phosphate group (pH > 5) and due to its free draining behavior, results in a constant ratio of charge density to its size. Because of this property of DNAs, free solution electrophoretic separations are impossible. It was first realized in 1966 that if DNA electrophoresis is performed within a gel, size-based separation will occur. ^ Electrophoresis of DNAs in a gel matrix can be thought of as a type of "gel filtration" where a mixture of nucleic acid molecules of different sizes are forced to move through the pores of the matrix under the influence of an electric field. Any DNA fragments to be fi'actionated encounter the gel network of polymer threads or pores, which increase the effective friction and consequently, lower the velocity of movement of the molecules. Small molecules move more rapidly through the gel, while large molecules move relatively slow. The existence of the gel medium contributes significantly to the observed electrophoretic mobility of the DNA molecules. Each DNA fragment of a unique size possesses a unique electrophoretic mobility in a particular gel and it migrates to a unique position within the electric field in a given length of time. The pore size in the gel matrix plays a critical role in determining the relative electrophoretic mobility and the separation efficiency of DNA fragments.

For DNA sequencing, electrophoresis is mainly used in one of three formats: slab gel, capillary, or microchip. The slab gel format employs a thin slab of chemically cross-linked gel and is traditionally used for the separation of biological macromolecules. The capillary format typically utilizes a small-bore capillary tube containing various matrices that are used as molecular sieves. The microchip format employs a separation channel embedded in a planar substrate. A schematic diagram of a typical set-up for each mode is presented in Figure 4.

23.1.2.1. Slab Gel Electrophoresis (SGE)

Slab gel electrophoresis is a traditional electrophoretic format used to separate DNA fragments in which the gel medium for electrophoresis is layered between two flat glass sheets. Figure 4a shows a typical slab gel electrophoresis unit. The thickness of the gel is determined by the dimensions of the spacers used along the sides of the gel during polymerization, typically ranging from 0.2 to 1 mm. When electrophoresis is performed, multiple samples, as many as 96, can be run in parallel in the same gel that is submersed in buffer. Samples are loaded into wells formed at one end of the gel during polymerization. The field strength that can be applied across the gel medium ranges from 50 to 80 V/cm, with the upper limit determined by heating (Joule) resulting from current flow through the gel. In this electrophoresis format, heat cannot be efficiently dissipated through the thick gel resulting in zone broadening, which limits the


C^apillary

a

High v<iU«g0 j

s if 1 St

i f

Catlimli-

1 i " i

c

'V'-' ''~~--.. U«?ti'<?.tor

/ tlfrV^ 1 /Vttwtle

1 mk I Hiirri-r ^m -; 1 ^ • j

1 Jligli %-ol»aigtt i 1 ! ptm t'r Hii|»pl) 1

mmmsmsm

—-,, <• ha i in t - l

Detector *

Figure 23.4. Schematic representations of three separation platforms for DNA sequencing: (a) Slab gel electrophoresis system; (b) capillary gel electrophoresis; (c) microchip electrophoresis system.

upper level of the electric field strength that can be effectively applied. Since all samples are present in the same gel, the electrophoresis is uniform from lane to lane. The sample loading volume is typically 1-10 |iL and analysis times range from a few hours to overnight with a rate of 100 bases per hour per lane. Despite its well-established reliability, slab gel electrophoresis suffers from several disadvantages. The most significant limitation of this platform is its poor dissipation of Joule heating, which limits its use to low electric fields. Moreover, the entire process, from the casting of the gel, preparation and loading of samples, is a series of cumbersome and time-consuming tasks.

23.1.2.2. Capillary Gel Eelectrophoresis (CGE)

Capillary electrophoresis (CE) has steadily gained popularity since its introduction in the late 1980's because of its inherent simplicity, the ability to provide online detection, full automation and superior throughput. The capillary filled with gel has been recognized for its potential to replace slab gel for DNA sequencing and is known as capillary gel electrophoresis (CGE). Capillaries employed in CGE are typically made of fused silica with an internal diameter of 20 to 100 jim and lengths of 20 to 100 cm. They are externally coated with a polymeric substance, polyimide, as a protection layer to give the capillary mechanical strength. The small dimensions of capillaries require only milliliter quantities of buffer and sample volumes in the nanoliter range. The efficient dissipation of Joule heat afforded by the high surface-area-to-volume ratio of a capillary tube virtually eliminates thermal and gravitational convection, allowing high electric strengths to be used resulting in short analysis times. ^ In


DNA separations, field strengths of about 300 V/cm are typically used, providing 25-times faster separations compared to conventional slab gel electrophoresis.^^

The basic components of CGE include a high voltage power supply, a capillary, two buffer reservoirs that can accommodate both the capillary and the electrodes connected to the power supply, and a detector (see Figure 4b). The high electric field strength that can be applied during electrophoresis results in a shorter electrophoresis time and enhanced separation efficiency compared to slab gel. In CGE, the capillary is filled with sieving polymer (gel) containing electrolyte or running buffer and its ends are submerged into reservoirs containing the same buffer. Samples are injected by dipping one end of the capillary into the sample and applying an electric field (electrokinetic injection) or by applying pressure (hydrodynamic injection). After a small sample plug is introduced into the capillary, the capillary is returned to the buffer reservoirs and electrophoresis is started. The species eventually pass a detector placed at the far end of the capillary where information is collected and stored by a data acquisition system. In free solution CE, the migration of the ionic species in the sample plug is driven by an electroosmotic flow (EOF), a bulk flow of solvent moving from the positive electrode toward the negative electrode. EOF is caused by development of an electrical double layer at the silica/buffer interface due to ionization of surface silanol groups at a certain pH value. When an electric field is applied, the mobile part of the diffuse double layer migrates toward the cathode in response to the electric force dragging the bulk solution behind it. However, EOF is undesirable when gel-filled capillaries are employed for DNA separations. DNA molecules are negatively charged and have an electrophoretic mobility toward the anode. In bare silica capillaries, a large magnitude of the EOF drives DNA molecules towards the cathode and therefore distorts the separation of DNAs. In some cases, electroosmotic forces within the capillary may cause the gel to migrate out of the capillary. In addition, analyte-wall interactions due to ionic interactions and hydrogen bonding significantly interfere with the separation as well. Because of these issues, it is desirable to suppress EOF within the capillary for gel separations of DNA. EOF can be controlled by applying a proper coating, either a covalent or a dynamic coating at the inner surface of the capillary . The coatings can increase, decrease, reverse, or eliminate the EOF depending on the presence or absence of certain functional groups contained within the coating materials. Coated capillaries also prevent sample interactions with the capillary and feature good separation efficiency and excellent run-to-run migration-time reproducibility. Most DNA separations by capillary gel electrophoresis are conducted in coated capillary tubes, where EOF is completely eliminated or significantly reduced to achieve high-resolution separations.

For all separation-based technologies, resolution is critical, especially in DNA sequencing since the read length (i.e., number of bases that can be accurately called in a single gel run) is dependent upon the electrophoretic band separation. Resolution is typically determined by two important electrophoretic parameters; (1) efficiency, which is measured in plates and; (2) selectivity, which is determined by the properties of the sieving gel and the molecular


properties of the analytes. The plate numbers are typically dependent upon the degree of band broadening or zonal dispersion, which describes processes that broaden zones during their migration through the capillary or other separation conduit. In gel electrophoresis, the major contributor to band broadening for well-designed systems is longitudinal diffusion of the solute molecules as they move through the separation conduit. Assuming that zonal dispersion is dominated by longitudinal difftision, the theoretical plate number (N), characteristic of column efficiency, can be directly related to the molecular diffusion by:

N^ju (1) 2D

where \i is the electrophoretic mobility, D is the diffusion coefficient of the solute in the separation gel-buffer system, and / is the effective column length. This equation suggests that higher applied electric fields and lower solute diffusion coefficients will result in higher separation efficiency. Another mathematical expression for separation efficiency is given as:

N = 5 . 5 4 ( - ^ ) ' (2) ^1/2

where tR is the migration time of the peak and W1/2 is the width at half-height of a Gaussian peak, which is related to dispersion processes with tend to broaden electrophoretic bands as they move down the separation tube.

The resolution (Rs) can be calculated from the following equation for electrophoretic separations:

Rs = }-^f^!EL.M^" (3) r*app,avg

where Ajiiapp is the difference in mobility between two neighboring bands (selectivity), |iapp,avg is the mean mobility of two neighboring components and N is the plate number (efficiency). As can be seen from this equation, the resolution in electrophoresis is determined by the relative mobilities between two components and the separation efficiency. When Rs = 0.75, two bands are considered as baseline resolved.


Table 23.1. Comparison of three platforms used for DNA fragment separations.

Slab Gel Electrophoresis Capillary Gel

Electrophoresis Micro-Gel

Electrophoresis

96-Lane Format

Yes

Yes

Yes

Load Volume 1-5 ^L

1-10 nL

0.1-1 nL

Field Strength

20-50 V/cm

100-300 V/cm

100-300 V/cm

Development Time

6-8 hrs

2-4 hrs

0.5-1 hrs

Gel Pouring

Yes

No

No

23.1.2.3. Microchip Gel Electrophoresis (ju-CGE)

Transitioning capillary gel electrophoresis to planar chips is considered as a revolutionary leap in analytical instrumentation and is having a large impact in the area of DNA sequencing. Microchip electrophoresis, being considered as a scaled-down version of conventional CE, can speed up analysis and allow massively parallel measurement systems to be envisioned. It also allows sophisticated sample processing steps to be carried out in an automated fashion while greatly reducing reagent and waste stream volumes. Chip-based DNA separation systems have been shown to have several advantages over their conventional analogues (SGE, or CGE, see Table 1).

Typical microchips consist of microstructures embedded into a glass or polymer substrate, ranging in design from a single electrophoretic separation channel and its injector to complex lab-on-a-chip or micro-total analysis systems (jiTAS) that include processing steps such as sample input, pre- and post-column reaction chambers, separation columns and detectors. Electrophoretic separations in microchips are performed directly in the fabricated micro-channels, with wells at the end of channels serving as reservoirs. The simplest layout of a microchip containing a cross structure is shown in Figure 4c. The use of cross-channels for sample injection allows for the injection of narrow, well-defined sample zones, which can significantly reduce the separation distance required for adequate resolution of DNA samples.^^ Samples are loaded electrokinetically into a cross-injector region and the separations are performed by applying an electric field across the long separation channel. The separated solute bands are then most frequently detected with laser-induced fluorescence (LIF) at the end of the separation channel.

23.1.3. Fluorescence Detection for DNA Sequencing

Following the electrophoretic separation of DNA sequencing fragments, they need to be detected and identified. Since the first demonstrations of the use of fluorescence tags to label DNA oligonucleotides,^^' ^ fluorescence detection has been accepted as the predominant detection protocol in DNA sequencing. It has allowed DNA sequencing to be performed in an "automated" fashion with base-calling performed during the gel


separation. In this method, Sanger sequencing fragments terminated with different bases are fluorescently labeled. Following gel electrophoresis, the individual components are identified based on unique properties of the fluorescent dyes used to tag each of them. Fluorescence detection possesses high sensitivity, intrinsic simplicity and the ability to perform on-line detection. More importantly, multiplexing capabilities, which allow multiple tracks of information to be processed in a single lane, allow high throughput applications. In addition, the recent focus on miniaturizing the separation platforms onto planar chips for sequencing has placed severe demands on detection due to the smaller sample injection volumes. Fluorescence-based sequencing instrumentation and strategies as well as a variety of fluorescent dyes used for gel-based DNA sequencing strategies will be described in the following sections.

23.2. DYE-PRIMER/DYE-TERMINATOR CHEMISTRY IN DNA SEQUENCING

Fluorescence-based DNA sequencing can generally be divided into two categories: dye-primer or dye-terminator sequencing. This categorization is based on the position of where the fluorescent dyes are attached to the substrates used for sequencing, either the oligonucleotide primer or the dideoxynucleotides (i.e., terminators).

In dye primer sequencing, four fluorescent dyes are attached to each sequencing primer and four separate sequencing reactions are carried out. Each reaction contains one dye-labeled primer, four dNTPs and a particular ddNTP. Dye-primer chemistry is widely used in various sequencing applications due partly to the fact that dye-labeled primers are typically less expensive compared to dye-labeled terminators. Also, in most sequencing, small pieces of DNA (1-2 kbp in length) are cloned into Ml3 vectors for propagation. The M13 vectors have a known sequence and can serve as ideal priming sites. However, dye-labeled primers do present some disadvantages. For example, the sequencing reactions must be run in four separate tubes during polymerization and then pooled prior to the gel electrophoresis. In addition, unextended primer can result in a large electrophoretic peak with high fluorescence intensity at the beginning of the trace, which often masks the bases close to the primer-annealing site and makes them difficult to be called.

In dye-terminator sequencing, the fluorescent dyes are attached to the dideoxynucleotides (ddNTPs). Sanger sequencing reactions are run using unlabeled primers, four dNTPs and four fluorescently-labeled ddNTPs, with each ddNTP labeled with a unique reporter (dye). The primary advantage of using dye-labeled terminators is that the sequencing reactions can be performed in a single reaction tube, which reduces reagent consumption and minimizes sample transfer steps. Dye-labeled terminators also can be appealing in certain applications, for example, when using primer-walking strategies . In primer-walking, the sequence of the DNA template is initiated from a common priming site, using a primer that is complementary to that site. After reading the


sequence at that site, the template is subjected to another round of sequencing, with the priming site occurring at the end of the first read. In this way, a long DNA can be sequenced by walking down the template. Dye terminators are particularly attractive in primer-walking strategies since primer sequence changes frequently and the use of unlabeled primers simplifies the synthetic preparation of these primers. Additionally, in many cases dye terminators improve the quality of sequencing data since the excess terminators can be easily removed prior to electrophoresis and as such, give clean gel reads free from intense primer peaks. However, many polymerase enzymes are very sensitive to the type of dye attached to the ddNTP, which can produce uneven peak heights (broad distribution of fluorescence intensities) during electrophoresis. Uneven peak heights are due to differences in incorporation efficiency of the dye-modified ddNTPs by the particular polymerase enzyme. In addition, the labeling of ddNTPs can be quite expensive. In dye-primer chemistry, this disparity is absent due to the large displacement of the dye from the polymerization site. ^

The instrumentation that can be used for dye-terminator reads can also be used for dye-primer sequencing as well. The difference rests in terms of the sample preparation protocols and the software corrections required for the sequencing data, such as different mobility-correction factors. Using dye-terminator chemistry, a purification step is necessary following DNA polymerization to remove excess dye-labeled terminators because they are negatively charged and can mask the sequencing data embedded within the trace. This can be efficiently accomplished using size exclusion columns and/or a cold ethanol precipitation step.

23.3. FLUORESCENT DYES FOR DNA LABELING AND SEQUENCING

A dye set for DNA sequencing is typically comprised of four dyes, with each one specific to a particular nucleotide base. An ideal dye set should possess the following properties: (1) Each dye in the set is spectroscopically distinct so as to be distinguishable. In most cases, the discrimination is based on differences in the emissive properties of the dye. The dyes should also be carefully chosen to minimize spectral leakage (cross-talk) among different detection channels. Other physical properties associated with the dyes, such as the fluorescence lifetime, can also be used to identify them. (2) The dye set should possess good photophysical properties such as high extinction coefficients and large quantum yields. These properties are necessary to ensure high sensitivity and detectability. (3) The dye set should show favorable chemical stability at high temperatures. This is because typical Sanger sequencing reactions utilize multiple temperature cycles that include a 95°C denaturing step. Therefore, the dyes are required to have good thermal stability


Xj^ " 5! 7 BB»

"^''•'o:p'''

i ^ " 5 6 0 am

k„," 5».3 nro

)ROX>

Ai, ' 5«t7 nro

iV IA i/\ \ ^

S20 S40 StfO S80 IMO 020 £140

WAVaEN&TH (mm)

• ntwi [3»«tw2 ^mir i i [!}»t«f«

Figure 23.5. Chemical structures of the FAM, JOE, TAMRA and ROX dye series used for labeling primers or chain terminators for DNA sequencing. The functional group on each dye is a carboxylic acid, which readily be conjugated to a primary amine group appended to a sequencing primer or the nucleobase used for chain termination (ddNTP). Also shown are the emission spectra of the dye set and the filters (shadowed area) used to select the appropriate dye color and process it onto each detection channel

for extended periods of time. (4) Dyes within a set should produce minimal mobility shifts during the gel separation. The post-run mobility corrections can be very complex and involved. The mobility shift is not only dependent upon the dye and linker structure, but also upon the separation platform used. For example, dyes which show uniform mobility shifts in slab gel electrophoresis, \may not show the same effect in capillary gel electrophoresis. Uniform mobilities of all dyes are desired to alleviate the need for extensive post electrophoresis data corrections. (5) An ideal dye set should require minimum numbers of excitation sources (lasers) and detection channels and possess highly efficient excitation and detection. These properties dramatically simplify the instrumentation and the cost of the sequencer. (6) The dyes must not disrupt the activity of the polymerase enzyme. This is particularly an issue in dye-terminator chemistry, since the proximity of the dyes to the polymerase enzyme can dramatically influence its ability to be incorporated into the DNA chain.

23.3.1. Visible Fluorescent Dyes

Traditional dye sets used in many automated fluorescence-based DNA sequencers consist of a set of four dyes that have absorption and emission spectra in the visible region of the electromagnetic spectrum (400-650 nm). This is due primarily to the readily available excitation sources, such as Ar, He-Ne and Kr ion lasers, and the high photon detection efficiencies of detectors like


photomultiplier tubes (PMTs) in this region. Also, fluorophores that can be detected in the visible are readily available from various commercial sources. The initial set of four fluorescent dyes used for sequencing were introduced by Smith et al. ^ These were fluorescein, 4-chloro-7-nitrobenzo-2-l-diazole (NBD), tetramethyl-rhodamine, and Texas Red. The dyes were covalently attached to a sequencing primer. A set of interference filters, centered at 520 nm, 550 nm, 580 nm and 610 nm were was used to spectrally resolve the four dyes based on their emission maxima. The use of these four dyes allowed fluorescence 4-color sequencing in a single gel lane that also provided real-time base calling and thus, increasing data throughput. Following this work, alternative dye sets were synthesized and linked to primers for sequencing. ' ' ^ ' ^ Among them, the most commonly used dye set included carboxyfluorescein (FAM), carboxy-4', 5'-dichloro-2', 7'-dimetoxyfluorescien (JOE), carboxyte-tramethyl-rhodamine (TAMRA), and carboxy-X-rhodamine (ROX). FAM and JOE are Applied Biosystems trade-names for fluorescein-based dyes that can be excited by the 488 nm line from an argon ion laser. TAMRA and ROX are rhodamine-based dyes that can be excited by the 514.5 nm line from the argon ion laser. These dyes contain a carboxy-group that can be converted to a succinimidyl ester for facile conjugation to amine terminated sequencing primers or terminators with appended primary amine groups. These dyes also possess fairly well-resolved emission profiles. Figure 5 shows the chemical structures of the FAM/JOE/TAMRA/ROX dye set and their emission spectra as well as the filter set that is used to isolate the emission from the dyes onto the appropriate detection channel.

Another dye set, dipyrrometheneborondifluoride fluorophores (BODIPY) were shown to have better spectral characteristics than conventional rhodamine and fluorescein dyes.^^ The bandwidths of the fluorescence emission profiles for this dye set were less than those associated with the FAM/JOE/TAMRA/ROX dye set and therefore, introduced less spectral leakage between detection channels. These dyes also showed uniform electrophoretic mobilities, high fluorescence intensities and thus, consumed 30% less reagent per reaction compared to rhodamine or fluorescein-type dyes. However, a drawback associated with this dye set was their chemical instabilities when subjected to extended heating at high temperatures during thermal cycling.

One common problem associated with the aforementioned dye sets is that their absorption spectra are dispersed over a relatively large spectral range, which provides poor excitation efficiency even with dual-color laser systems (488 and 514 or 543). As a consequence, the red-dyes are usually used at higher concentrations during DNA polymerization. To overcome this problem without sacrificing spectral dispersion in the emission profiles to allow efficient identification, dye sets based on fluorescence resonance energy-transfer (FRET) have been introduced that can be used for four-color sequencing.^ ' ^ FRET uses non-radiative energy transfer from a donor dye, whose emission spectrum overlaps with the absorption spectrum of an acceptor dye, to the acceptor dye. By employing FRET, a single donor dye with a broad emission spectrum can be


used to excite a suite of acceptor dyes, thus requiring only a single laser to excite all of the acceptor dyes without sacrificing excitation efficiency. This system requires a donor/acceptor dye pair to label each of the four bases. The paradigm of FRET-based sequencing was introduced by Ju and co-workers using the FAM/JOE/TAMRA/ROX dye-set.^^ These FRET-constructs (shown in Figure 6) featured FAM as a donor attached to the 5'end of the sequencing primer, and JOE/TAMRA/ROX acting as acceptors attached to an internal modified base (T*) that possessed a linker structure with a primary amine. The spacer size and the length between the donor and acceptor were optimized for energy-transfer efficiency and to provide uniform electrophoretic mobilities. Figure 6 shows the absorbance and emission profiles of the energy-transfer primer series. While the absorption spectra showed bands from both the acceptor and donor dyes, the emission profiles were dominated by fluorescence from the acceptor dyes. The emission intensity of the ET primers was found to be significantly higher than their single-dye primer counterparts when excited by 488-nm laser light. This translated into improved fluorescence sensitivity of these ET primers during electrophoretic sorting. Therefore, the major advantages of ET dye sets are that they can be efficiently excited by a single wavelength, producing stronger and more uniform fluorescence emissions. Other advantages of these ET primer sets are that their dye-dependent electrophoretic mobility shifts can be minimized by adjusting the linker structures and they require smaller amounts of template in the sequence analysis compared to the single dye sets. Since publication of the work by Ju et al , the ET cassette has been optimized with different donor/acceptor pairs, varied spacing between the donor and the acceptor, various spacer chemistries, and improved primer sequences.^^' ^

Efforts made in dye labeling technology have produced dye-conjugates that can be used with dye-terminator chemistry. The first dye-labeled dideoxy terminators were developed by Prober et al. in 1987. ^ Further achievements have been made by translating the concept of ET dyes into dye-terminator protocols. ' ' Impressive advances have been made in this field, the most notable of which are the DYEnamic^^ dye sets from Amersham Biosciences and BigDye^^ dye sets from Applied Biosystems. With these improvements, performing cycle sequencing with energy-transfer terminators has become routine and yields sequencing data of high quality.

While many successful sequencing experiments have been performed with the above mentioned visible dyes, they are susceptible to some technical limitations. The primary limitation is the broad emission profiles associated with many of these dyes, which makes it difficult to discriminate among the four colors. The inability to correctly match each detected photon to its source creates an inherent background (cross-talk) in each detection channel, which limits the accuracy of the base-call during sequencing. A limited number of excitation sources in the visible region of the spectrum, the wide, overlapping emission spectra of the visible dyes, a high intrinsic background, and different electrophoretic mobilities of fluorescent dye dyes are all complicating factors that must be accounted for during DNA sequence reconstruction.


F10F FAH-S'-GITTTCCCAGT'CACGACG'S'

(CH)2(C0)-NfHCHa)rNH^AM

0 FI0J FAM'5-GTTrrCCCAGfCAC^AC6-3*

(CH)2(COH^H^CK2)54JH-C- IO£

0 m FAM-S' TFT'TWAGTCACGACG-y

(CHyCOHH-fCH^Ji-NH^TAITRA

0 F ^ FAI^-5m*TCCCAGTCACQACG-3'

(iHycoj-NiHCjyg-NH-c-ftox I 0

40O 450 £00 550 600 650 700

Figure 23.6. Structures of the FRET (ET) primers and their emission spectra, (a) Structures of four ET primers, FIOF, FIOJ, F3T and F3R. The naming of these ET primers follows the convention: donor-spacer (number of bases)-accepter, where F = FAM; J = JOE; T = TAMRA and R = ROX (see Figure 5 for chemical structures of these dyes), (b) The absorption (dark line) and emission spectra for both the ET primers and the single-dye-labeled primers, which are shown for comparison purposes. The number in parenthesis is the excitation wavelength used for collecting the emission profile.^^

23.3.2. Near-IR Fluorescent Dyes

Performing fluorescence experiments in the near-IR region (650-1000 nm) has some unique advantages:^^ (1) Background interference from impurity molecules in the sample matrix can be substantially reduced in the near-IR since very few biological molecules possess intrinsic fluorescence (autofluorescence) in this region. On the other hand, many biological compounds have autofluorescence in the visible region, which is difficult to reduce using optical filters without compromising the sensitivity of the measurement. As such, detection of limits in the near-IR can be significantly improved. (2) Enhanced sensitivity can also be achieved through significant reduction of scattering effects. The amplitude of Raman or Rayleigh scattering is inversely proportional to the 4* power of the excitation wavelength QC^).^^ As the wavelength of the excitation light increases, the efficiency of the Raman scattering decreases dramatically, thus decreasing the background signal. (3) With the availability of laser diodes and photodiode transducers, the instrumentation required for near-IR detection can be rather simple and


inexpensive. These solid-state lasers and detectors can be run for extended periods of time while supplying ample power and requiring minimal maintenance or operator expertise. Semiconductor laser diodes are now available from 630 nm to longer wavelengths and can be operated either in a continuous or pulsed mode, allowing time-resolved fluorescence measurements as well. Moreover, most photodetectors, such as single photon avalanche diodes (SPADs), show high single photon detection efficiencies in the near-IR. Additionally, with the recent major push towards miniaturization of separation platforms, some aiming at portable systems, the compact size and low cost of these semiconductors operated in near-IR provide the ability to build complete miniaturized systems. Table 2 summarizes the comparison of a typical laser diode to an air-cooled argon-ion laser.^^

All the above merits associated with near-IR fluorescence detection make it a very attractive alternative to visible fluorescence for DNA sequencing. In sequencing, the separation and detection occur within a highly scattering medium, the gel matrix. The reduced amplitude of scattering and the intrinsically lower backgrounds associated with near-IR fluorescence compared to the visible region allow ultra-sensitive detection to be achieved. Several groups of researchers have explored the use of near-IR dyes for sequencing.

The near-IR fluorophores that have been typically used belong to the cyanine-class of dyes. ^ The most frequently employed are the dicarbocyanines or the tricarbocyanines. The carbocyanines consist of heteroaromatic structures linked by a polymethine chain containing conjugated carbon/carbon double bonds. The absorption and emission maxima can be altered by changing the length of this polymethine chain or changing the heteroatom within the heteroaromatic fragments. For example, dicarbocyanines show absorption maxima near 630 nm, while the tricarbocyanines show absorption maxima near 780 nm. These dyes have large extinction coefficients and can be augmented with various functional groups to increase their solubility in water. '*' ^ However, these dyes have poor photochemical stabilities in aqueous media and relatively low fluorescence quantum yields, which result primarily from high rates of internal conversion.^^

Near-IR fluorescence has been demonstrated in sequencing applications using slab gel electrophoresis, where the detection sensitivity has been reported to be -2000 molecules.^^ Williams and Soper demonstrated

Table 23.2. An operational comparison of a far-red laser diode and argon laser as excitation sources for fluorescence.^^

Wavelength (nm) Life span

Optical power output Power consumption Replacement cost

Laser diode 785 nm

> 100,000 hr 20 mW

0.150 W <$150

Argon-ion laser 488 nm

-3000 hr 20 mW 1800 W >$5000


superior detection limits attainable using a tricarbocyanine dye for sequencing by CGE. ^ A direct comparison of detection limits between visible and near-IR excitation (488 nm and 780 nm) was made, by electrophoresing sequencing primers, one labeled with FAM and the other labeled with a near-IR dye. It was determined that the detection limits for these systems were 3.4x10'^^ moles for 780 nm excitation and 1.5x10" ^ moles for 488 nm excitation. The improvement in the limit of detection for the near-IR system resulted primarily from the significantly lower background observed in the near-IR, in spite of the fact that the fluorescence quantum yield of the near-IR dye was far lower than that of the FAM dye.

Table 23.3. Chemical structures and photophysical properties of heavy-atom modified tricarbocyanine dyes, which contain an isothiocyanate labeling group.^^

Sulfonated heavy-atom modified near-IR dyes X=I,

1 Br, CI, or F

ifcifeg

k-p? ^ D. Sm

Dye

Br CI F

Absorption

766 768 768 768

Emission

796 798 797 796

e (M-'cm-*)

216000 254000 239000 221000

(ps)

947 912 880 843

^f

0.15 0.14 0.14 0.14

Strekowski et al. reported the synthesis of several heptamethine cyanine derivatives that contained an isothiocyanate labeling group appropriate for primary amine containing targets. ^"^^ In their subsequent work, they reported the use of these heptamethine derivatives as labels for DNA sequencing primers."^^ Flanagan et al. developed a series of heavy-atom-modified tricarbocyanine dyes, which possessed succinimidyl esters or isothiocyanate groups for labeling DNA primers that contained primary amine groups (Table 3). The interesting feature of these dyes was that they possessed an intramolecular heavy atom to perturb the fluorescence lifetime of the base chromophore. The incorporation of the heavy atom, however, did not alter the absorption and emission profiles of the dyes. Therefore, the dyes possess similar absorption and emission wavelengths but distinct lifetimes, the values of which depend upon the identity of the heavy atom modification. This group of dyes has been targeted for use in DNA sequencing based on lifetime discrimination, in which a single excitation source can excite all the dyes, and a single detection channel will process the fluorescence signals. An additional advantage of these dyes is that the primers labeled with these dyes show uniform electrophoretic mobilities and therefore, post-electrophoresis corrections would not be necessary.


Figure 23.7. Base structure of (a) Phthalocyanine (Pc) family of near-IR fluorescence compounds and (b) Naphthalocyanine (NPc) dyes. The M represents the metal center and can be, for example, Ni, Ga, Pt, Pd, Cu, Al, Si, and Zn. The identity of the metal center affects both the fluorescence and spectroscopic properties of the metal Pc and NPc dyes.

An attractive alternative to the cyanine-based near-IR fluorophores are the phthalocyanines (Pc) or naphthalocyanine (NPcs) family of compounds. The Pc's and NPc's possess a conjugated ring structure, which is linked together by four aromatic-dicarbonitrile fragments. The absorption and emission maxima of these dyes can easily be altered by appending different functional groups to the dye macrocycle."* ' "" For example, the annulation of benzene rings onto the Pc core will produce the NPc dyes, which have absorbance maxima red-shifted by 50-100 nm compared to the Pc dyes." ' '* The basic structures of Pc and NPc dyes are shown in Figure 7. Although the Pc and NPc dyes are very hydrophobic and essentially insoluble in water, functionalization of the

periphery of the dye with charged groups (SO^ , CO2 , PO^^ ~, etc.) or

attachment of a polar functionality to the metal center (PEG-OSi) will improve water solubility.'^'' Like the tricarbocyanines, Pc and NPc dyes can be functionalized for conjugation to biomolecules. The attractive feature associated with the Pc dyes are their superior quantum yields and favorable photochemical stabilities compared to the tricarbocyanines."^^"^^ In addition, Pc and NPc dyes typically have longer fluorescence lifetimes compared to the carbocyanines, which can be adjusted by coordinating different metals to the core of the macromolecule.^^ Table 4 shows a comparison of a few photophysical parameters of Pc and NPc dyes to several other classes of near-IR dyes. Unfortunately, the synthetic route for preparation of these dyes typically produces symmetrically substituted structures. Preparing asymmetrical analogues with fimctional groups for facile conjugation to target molecules has been somewhat difficult since purifying the Pc isomers can be problematic. Hammer and co-workers have prepared and isolated asymmetric metallo-phthalocyanine dyes containing a reactive isothiocyanate functional group for covalent labeling of oligonucleotide primers,^^ which has shown potential for use in near-IR-based DNA analysis. The same group is proposing the route of making a series Pc and NPc dyes with different metal centers for lifetime-based DNA sequencing.


Table 23.4. Comparison of the photophysical properties associated with several classes of near-IR dyes.

Absorption maximum Emission maximum Quantum yield (^{)

Lifetime (if) Photochemical Stability ((|)d)

Photon yield /Molecule {(<^{/^d)

Phthalocyanine 676 n m ' 684 n m '

0.37' 5.0 n s '

1.7 X 10- ' 2.2 X 10

Naphthalocyanine 769 nm ^ 777 nm ^

0.25'' 2.5 ns '

3.0 X 10"^' 83

Tricarbocyanines 764 n m ' 73 nm'* 0.05 '^ 1

0.076 ns '^ 7 .0x10 '^

7.1 a. Tetra-sulfonated Al-Pc; aqueous; from Ref. 53. b. Tetra-sulfonated Al-Naphthalocyanine,; aqueous; from Ref. 49. c. Measured in aqueous buffer pH=9.3. Ref. 50. d. Measured in aqueous buffer pH=8.3. Ref. 50. e. Unpublished data, Soper, et al.

The ability to tailor the spectroscopic properties with subtle changes in dye structure makes the carbocyanine, Pc and NPc dyes excellent candidates for bioanalytical applications. However, near-IR dyes for sequencing have relatively limited commercial sources. IRD700 and IRD800 are two near-IR dyes developed by LiCOR Inc. and they have been successfully employed in multi-lane sequencing instruments using a slab gel format (see www.LICOR.com). Given the interest in sequencing, accompanied by the vigorous growth in solid-state technology, one can expect continued efforts in development of new near-IR dyes for bioanalytical applications such as DNA sequencing.

Efforts have also been directed toward developing the deep-red cyanine-based fluorescent dyes for DNA sequencing applications. These dyes have high extinction coefficients and are available as phosphoramidites and NHS esters for labeling oligonucleotides (primers) or dideoxy terminators. For example, Cy5 and Cy5.5 dye-labeled terminators have been used in single-color sequencing experiments . It was demonstrated that deep red cyanine dye-labeled terminators, when combined with a suitable DNA polymerase, could produce uniform band patterns and yield high quality sequencing data. Structures of a Cy5.5 labeled ddATP is shown in Figure 8. The Cy-dye series (Cy 3, 5, 5.5) can also be used to produce ET cassettes for sequencing . With a donor being Cy3 that is excited by an Nd:YAG laser (532 nm), Cy3, 3.5, 5 and 5.5 can be used as acceptor dyes for preparing ET primers. Other near-IR dyes, such as IRDye 700 and IRDye 800 are available from LiCOR Inc. and they have also been successfully employed in multi-lane sequencing in slab gel formats (http://www.licor.com/bio/IRDves/index.isp). Molecular Probes has also developed the AlexaFluor series of dyes that span across the visible and near-IR. AlexaFluor 680 and 750 have spectral properties almost identical to Cy5.5 and Cy7 dyes, respectively, but with greater brightness and better photostability (http://probes.invitrogen.com/handbook/sections/0103.html). Another class of cyanine-based near-IR dyes is called the WellRED dye set, which has been introduced by Beckman Coulter Inc. for use with their CEQ^"^ Series Genetic Analysis system (http://www.beckman.com/products/specifications/


Figure 23.8. Near-IR labeled dye terminator. In this case, the terminator labeled is a ddATP. The linker is a propagyl amine, which contains a triple bond to provide rigidity to the linker to minimize dye interaction with the DNA polymerase during incorporation events. The dye is attached to the nucleobase on a non-hydrogen bonding site of the ddNTP. The dye used here is a tricarbocyanine, which contains 4 water solubilizing groups (sulfonates).

geneticanalvsis/ceq/wellreddves con stat.asp). Excitation of these dyes is achieved with 650 nm and 750 nm diode lasers. The dye set is available in the form of dideoxy terminator conjugates and facilitates four-color sequencing on Beckman's CEQ capillary array machine.

23.4. FLUORESCENCE-BASED DNA SEQUENCING STRATEGIES

One of the primary advantages of using fluorescence detection for DNA sequencing is its multiplexing capability, in other words, its ability to identify multiple analytes in a single assay. The fluorescent probes attached to each nucleotide base can be identified simultaneously based on their characteristic fluorescence properties. Two properties associated with fluorescent probes have been used for identification purposes include, spectral characteristics (color discrimination) and fluorescence lifetimes (lifetime-discrimination). In each method, several different strategies can be implemented to reconstruct the sequence. These strategies may vary in terms of the number of dyes used, the number of detection channels required, or the need for running 1-4 parallel electrophoresis lanes. For example, if only one dye is used to label all four bases, the electrophoresis must be run in four different lanes, one for each base comprising the DNA molecule. However, if four different dyes are used, the electrophoresis can be reduced to one lane, and as a consequence, the


throughput of the instrument goes up by a factor of 4. In the following sections, a discussion of the various fluorescence-based nucleotide base calling strategies will be discussed. In addition, a critical assessment of each protocol will be supplied with the following questions to be answered: Which configuration is better from both read-length and accuracy points of view? Which detection format produces the best signal-to-noise ratio (SNR) in the measurement? What is the necessary instrumentation required for detection?

23.4.1. Color Discrimination Methods

Color or spectral discrimination is the most common method used in DNA sequencing in which spectrally distinct reporters are attached either to a sequencing primer or to a dideoxynucleotide and are identified by their unique emission maxima. The sequencing strategies that will be discussed here are categorized by the number of color channels and electrophoresis lanes used in the base-calling process.

23.4.1.1. Single-color/Four-lane

In this processing format, only a single laser and single detection channel are used to excite and detect the sequencing fragments produced following Sanger chain termination. All bases are labeled with the same dye. Since no spectral discrimination capability is implemented, the products of the four sequencing reactions must be run in four separate lanes of a slab gel, one for each base, similar to the format used in traditional radiographic detection. One of the first demonstrations of using fluorescence detection for DNA sequencing used this single-color/four-lane strategy^^ used a primer that was labeled with rhodamine at its 5' end. A single laser excited the fluorescent dye from all lanes at the end of the gel and fluorescence was detected by an array of photodiodes. Using a similar scheme, Middendorf et al. sequenced a DNA template using a near-IR dye on a slab gel sequencer.^^ In this case, the system used a microscope head containing the collection optics, a diode laser, filters, and an avalanche photodiode to read the fluorescence from the gel. Four traces, each from an individual lane of the slab gel, were overlaid to allow reconstruction of the sequence of the DNA template. The time required to secure the data was 6 hours and possessed a read-length of-600 bps. While this single-color, four-lane approach is reasonable for sequencing in slab gel formats, it is difficult to implement in CGE applications due to the poor run-to-run reproducibility in the migration rates of the fragments traveling through different capillaries. This is due to differences in the gel from capillary to capillary as well as differences in the integrity of the wall coatings used to suppress the electroosmotic flow. In the slab gel format, reproducibility in the migration times between lanes becomes less of a problem since all of the lanes are run in the same gel matrix.


23.4.1.2. Single-color/Single-lane

In the single-color/single-lane method, a sequencing reaction is carried out using a single dye-labeled primer and four different molar ratios of dideoxynucleotides to deoxynucleotides used during polymerase chain extension. The different concentration ratios of the terminators used during DNA enzymatic polymerization results in variations in the relative fluorescence intensity associated with each electrophoretic band. For example, if the concentration of the terminators used during DNA polymerization is in a ratio of 8:4:2:1 (ddA:ddC:ddG:ddT), a series of fluorescence peaks should be generated following electrophoretic sizing with an intensity ratio of 8:4:2:1. The identification of the terminal base is then carried out by categorizing the peaks according to their peak amplitude with the largest peaks representing A, the next largest C, the next largest G and the smallest T. " ' ^ Because only a single fluorophore is used, this approach requires particularly simple instrumentation — a single laser wavelength to excite the fluorophore and a single detection channel to process the fluorescence. The throughput of this sequencing protocol is higher compared to the single-dye/four-lane strategy because the sequence is reconstructed from a single electrophoresis lane not four. However, this protocol requires uniform incorporation of the terminators during DNA polymerization. In other words, the ability of the DNA polymerase enzyme to incorporate the terminators must be nearly uniform to ensure a high degree of base calling accuracy. Thus, this approach is restricted to the use of dye-primer chemistry in most cases, since dye-labeled terminators serve as poor substrates for most polymerases. Because this sequencing method requires sequencing products generated at uniform rates, significant amount of work has been invested into discovering and developing polymerase enzymes that incorporate ddNTPs efficiently and uniformly. For example, Tabor and Richardson developed pyrophosphatase, a genetically modified T7 DNA polymerase, that incorporates dideoxynucleotides efficiently and generates reaction peaks of nearly equal height in a sequencing gel.^ ' ^

A similar coding scheme has been used but only three ddNTPs are used in a concentration ratio of 4:2:1.^^' ^ ' ^ The missing ddNTP was assigned to gaps between electrophoretic peaks from the gel traces. Figure 9 presents an example of sequencing data obtained using this base-calling strategy. In this example, the accuracy in calling bases was estimated to be 84% up to 250 bases from the primer-annealing site, with readable bases up to 400 but with accuracy deterioration to -60%.

In most cases, this simple single color peak-height encoded sequencing protocol produces poor accuracy due to the difficulty of controlling peak height precisely. Particularly, when the scheme contains null coding, it lessens the base calling accuracy due to non-uniform electrophoretic peak widths, which creates ambiguities when contiguous null signals must be identified, which can lead to insertion or deletion errors.


300 bp

10000 3200

Time (s)

Figure 23.9. Single-color/single-lane capillary-gel sequencing data. The template was an M13mpl8 phage. The primer was labeled with a single near-IR fluorescent dye. The terminators were adjusted to a concentration ratio of 4:2:1:0 during extension to allow identification based on fluorescence intensities. The separation was performed at a field strength of 250 V/cm, and the gel column contained a 3%T/3%C crosslinked polyacrylamide gel.^°

23.4.1.3. TwO'Color/Single-lane

To improve the base-calling accuracy associated with the single-color/single-lane strategy without significantly complicating the sequencing readout instrumentation, a two-color format can be used to identify the four terminal bases. In this approach, one or two lasers are used to excite one of two spectrally distinct dyes used for labeling sequencing primers, and the fluorescence is processed onto one of two detection channels, consisting of bandpass filters and photon transducers. Dovichi and co-workers developed a two-color peak-height encoded sequencing protocol for applications in CGE.^ ' ^ In their work, fluorescein-labeled primers with ddATP and ddCTP at a concentration ratio of 3:1 were used for one set of sequencing reactions while rhodamine labeled primers were used with ddGTP and ddTTP also in a 3:1 ratio for the second set of reactions. The DNA sequence was determined based on both spectral and intensity information processed in two color channels. The read length achieved was -400 bases with a base-calling accuracy >97%. Researchers from this same group also demonstrated DNA sequencing using an "internal labeling" strategy using the same two-color, intensity encoded protocol. Each fluorescent dye was situated on the deoxynucleotides (dATPs) instead of primers or terminators.^^ These dye-labeled dNTPs were inexpensive to make and could be incorporated more uniformly during polymerization


compared to dye-labeled terminators. The base-calling accuracy in this example was 97% at 500 bp.

Huang et al. developed another two-dye protocol based on binary coding.^^ They used a single laser (488 nm) to excite two labeling dyes, FAM and JOE, which possessed similar absorption maxima. This dye pair was selected because it produced sequencing fragments that showed identical mobility shifts. Red and green fluorescence was processed onto two separate detection channels. During DNA polymerization, four separate reactions were run with each set of DNA sequencing fragments labeled differently; A terminated fragments contained an equimolar mixture of FAM and JOE dye, G fragments were labeled only by JOE, T were labeled by FAM, and C fragments contained no dye. Following separation in a single capillary, the bases were distinguished by ratioing the signal in the red (JOE) to green (FAM) channel, which was called binary coding. The read length using this strategy was 350 bases with an accuracy of 95.7%. A majority of the errors were attributed to C determinations, since a null signal was used to indicate the presence of this base.

23.4.1.4. Four-color/Single-lane

The commonly used approach in most commercial automated DNA sequencers incorporating fluorescence detection is by a four-color/single-lane strategy. Four spectrally resolvable fluorescent probes are used to label either the sequencing primers or the terminators in Sanger reactions. Fluorescence emissions are sorted onto four color-channels followed by base identification. Accuracy of base calling by this method is in large part determined by the spectral resolution of the four labels and the degree of spectral leakage into different color channels. The primary advantage in using this four-color/single-lane approach is that the throughput is high, due to the fact that all bases can be called in a single gel tract. In addition, it provides high accuracy in the base calling, especially for long reads. Unfortunately, a four-color detector requires extensive optical components to sort the four-color fluorescence due to, in some cases, the need for multiple excitation sources to efficiently excite the fluorophores used to identify the individual ddNTPs. In addition, post-electrophoresis software corrections may be required to account for severe spectral leakage into detection channels because of the broad emission profiles of typical dye-sets used for fluorescence labeling.

Smith and co-workers reported the first four-color DNA sequencing.^^ In this method, four spectrally distinguishable sequencing primers were used for the Sanger reactions. The sequencing reaction products were pooled and separated in a single lane of the electrophoresis gel. The separated bands were detected near the bottom of the gel by an LIF spectrometer that consisted of an Ar ion laser operated in a multi-line mode and a conventional photomultiplier tube (PMT). Filter wheels were placed in front of the laser and the PMT to select the appropriate excitation wavelength (488 or 514.5 nm) and emission color. The filter pairs used during fluorescence readout were 488/520 nm.


488/550 nm, 514/580 nm, 514/610 nm. Carefully choosing the dyes was required to minimize spectral cross-talk.

Recently, Lewis et al. developed a different four-color/one-lane approach based on pulsed multi-line excitation (PME).^^ Four lasers are operated in a continuous-wave mode (399, 488, 594 and 685 nm) and are modulated by mechanical shutters. The beams are converged into a single coaxial beam by a prism sequentially illuminating a single capillary (see Figure 10a). The ordered pulses of each laser cycle were generated by electronically controlling the mechanical shutters at a frequency of 5 Hz. A unique fluorescence dye set that had absorption maxima spanning across the entire visible region of the electromagnetic spectrum was used. The four dyes were carefully chosen to match each laser's wavelength, resulting in strong emission signals while showing negligible cross-talk when exposed to the remaining laser wavelengths. Their absorption and emission spectra are shown in Figure 10b. The resulting pulsed fluorescence signals were imaged onto a single PMT. By correlating the sequence of excitation pulses from four lasers with PMT responses of the emission intensities from fluorescently labeled DNA fragments, the authors claimed a "color-blind" base calling — identifying the dye by its emission intensity pattern on the four color channels. The primary advantages of using the PME method are; (1) the high and uniform fluorescence signals resulting from the optimized excitation and emission of the dye and; (2) the elimination of cross-talk between dyes, since temporal separation of the laser pulses leads to the separation of the corresponding fluorescence.

a

10 460 4m

Figure 23.10. (a) Schematic diagram of the pulsed multi-line excitation (PME) fluorescence system. Each laser operates in a continuous-wave mode with mechanical shutters modulating the different excitation beams in sequential order, (b) Excitation and emission spectra of the four dye-labeled primers used in the PME experiments (solid lines are excitation and dotted lines are emission scans) and the appropriate filter systems. Fl, 420-nm long-pass filter; F2, F3, and F4, are 488, 594.1, and 685 nm Notch-plus filters. The four lasers are lasing at 399, 488, 594 and 685 nm, respectively.^^


23.4.2. Lifetime Discrimination Methods

While most sequencing applications using fluorescence utilize spectral discrimination to identify the terminal bases during electrophoretic sizing, an alternative or complementary approach is to use the fluorescence lifetimes of the labeling dyes to identify each of the bases. The fluorescence lifetime (xf) is an intrinsic photophysical property of fluorophores that measures the average time difference between electronic excitation and fluorescence emission. The monitoring and identification of multiple dyes by lifetime discrimination during a gel separation can allow for an additional identification protocol when combined with color discrimination to provide high multiplexing capabilities. Several principal advantages associated with fluorescence lifetime identification protocols include:

1. The calculated lifetime is immune to concentration differences. As such, dye-labeled terminators can potentially be used as well as dye-labeled primers with a wide choice in polymerase enzymes to suit the particular sequencing application; the base identification can be accomplished with high accuracy irrespective of the intensity of an electrophoretic band.

2. Lifetimes can be determined with higher precision than fluorescence intensities under appropriate conditions, improving base-calling accuracy.

3. Lifetime determinations do not suffer from spectral leakage due to broad fluorophore emission profiles.

4. Multi-dye fluorescence can potentially be processed on a single detection channel without the need for spectral sorting to multiple detection channels.^^

Several problems do arise when considering lifetime discrimination for DNA sequencing. One potential difficulty is the poor photon statistics (low number of photocounts in a decay profile from which the lifetime is extracted), especially when utilizing micro-separation techniques such as capillary and microchip gel electrophoresis, which have low sample loading masses and short residence time of fluorophores within the excitation beam (1-5 s). Basically, the low number of photocounts acquired to construct the decay profile during electrophoresis can produce low precision in lifetime measurements since the fluorescence lifetime is extracted and calculated directly from the decay. In addition, the high scattering medium in which the fluorescence is measured (gel matrix) can produce large levels of background photons that would be included into the decay profile, lowering precision in the measurement. The poor precision would consequently affect the accuracy in the base call. An additional concern with lifetime measurements for calling bases in DNA sequencing is the complex instrumentation required for lifetime determinations as well as the complex algorithms that are required for extracting the lifetimes from the decay profiles. Nevertheless, many of these concerns have been addressed using near-IR fluorescence. The increased availability of pulsed diode lasers and single photon avalanche photodiodes has


had a tremendous impact on the ability to assemble simple time-resolved instruments appropriate for sequencing applications, with performance characteristics comparable to those using visible wavelengths. The use of near-IR fluorescence also reduces the background and scattering photons during detection, potentially increasing the sensitivity of the instrument and improving the photon statistics.

There are two different methods for measuring fluorescence lifetimes; frequency-resolved^" ' ^ and time-resolved.^ ' ^ In frequency-resolved methods, also called phase-modulation techniques, the excitation source is intensity modulated at a high frequency, typically employing sine wave modulation. When the fluorescent sample is excited by the modulated light, the emission responds at the same modulation frequency, but with a time delay compared to the excitation with the phase delay related to the lifetime of the sample. This time delay is characterized by a phase shift, which is then used to calculate the fluorescence lifetime. In addition, the fluorescence lifetime can be measured from the intensity of the demodulated signal. The ability to measure short lifetimes depends on the frequency of the modulation and the efficiency of the phase-sensitive measuring electronics.

In the case of time-resolved fluorescence, the fluorophore is excited by a pulsed light source at a relatively high repetition rate. The time duration of the light pulse needs to be as short as possible preferably much shorter than the fluorescence lifetime being measured. The emitted photons from the sample are time-correlated to the excitation pulse from which the lifetime can be determined. At present, most time-resolved measurements are performed using the time-correlated single photon counting (TCSPC) technique. A typical TCSPC device consists of a pulsed excitation source, a fast detector, and timing electronics that include a constant fraction discriminator (CFD), an analog-to-digital converter (ADC), a time-to-amplitude converter (TAC), and a multichannel analyzer (MCA). The time difference between the excitation and the arrival of the resulting fluorescence photon to the detector is recorded electronically. The recorded time differences over many excitation-emission cycles are placed in the appropriate time channels of the MCA and a statistical histogram is constructed representing the decay profile of the fluorophore. A calculation algorithm is then applied to the decay to extract the fluorescence lifetime. The shortest lifetime that can be measured reliably by this method is determined by the response time of the instrument (Instrument Response Function, IRF), which depends on the width of the excitation pulse, the spread of the travel times of photoelectrons in the photon detector, and the jitter in the measuring electronics. The IRF can be deconvolved from the collected decay profile using a deconvolution algorithm to provide a more accurate representation of the fluorescent lifetime. The time-resolved mode is a digital (photon counting) method, therefore it typically shows better signal-to-noise than a frequency-resolved measurement, making it more attractive for separation platforms that deal with minute amounts of sample. In addition, the use of time-resolved methods allows for the use of time-gated detection in which background photons, which are coincident with the laser pulse (scattered photons), can be gated out electronically improving the signal-to-noise ratio in


the measurement. However, TCSPC typically shows a limited dynamic range due to pulse pile up effects at high counting rates.

A device that has been used for making on-line lifetime measurements during CGE is shown in Figure 11 . ^ The light source consisted of an actively pulsed solid-state GaAlAs diode laser lasing at 780 nm with a repetition rate of 80 MHz and an average power of 5.0 mW. The pulse width of the laser was -50 ps (FWHM). The detector selected for this instrument was a single photon avalanche diode (SPAD) possessing an active area of 150 |im offering a high single photon detection efficiency (>60% above 700 nm). The counting electronics were situated on a single TCSPC board, which was plugged directly into a PC-bus exhibiting a dead time of <260 ns, allowing efficient processing of single photon events at counting rates exceeding 2x10^ counts/s. This set of electronics allowed for the collection of 128 sequential decay profiles with a timing resolution of 9.77 ps per channel. The instrument possessed a response function of approximately 275 ps (FWHM), adequate for measuring fluorescence lifetimes in the sub-nanosecond regime.

Diode Laser (G«AU»)

FIcQferoiidCiirreAt Source

^jjf Flltcra I

[;:;;;::J-|[^ Stort Fulsc

MCS

Stiirt Pube

ADC U i r A c CFD

DV SFAD

Stop fitlie JICIAMP

Figure 23.11. Block diagram of a near-IR time-correlated single-photon counting (TCSPC) detector for CGE. The laser is focused onto a capillary column with the emission collected using a 40x microscope objective (NA = 0.85). The fluorescence is imaged onto a slit and then spectrally filtered and focused onto the SPAD. L, laser singlet focusing lens; C, capillary; BD, beam dump; MOl, collecting microscope objective; M02, focusing microscope objective; SPAD, single photon avalanche diode; AMP, amplifier; CFD, constant fraction discriminator; TAC, time-to-amplitude converter; ADC, analog-to-digital converter; and MCS, multi-channel scaler. ^


One of the most important aspects associated with lifetime measurements in sequencing applications is the processing or calculation algorithm used to extract the lifetime value from the resulting decay. The accuracy of the base call depends directly on the lifetime differences between fluorophores in the dye set and the relative precision in the measurement. Algorithms used in this application require not only the calculation of the lifetimes precisely, even under the situation of poor photon statistics, but also the ability to perform on-line measurements during the electrophoresis. The typical calculation algorithm for lifetime determinations, nonlinear least squares (NLLS), can deconvolve the IRF from the overall decay and provide a more accurate lifetime value. Unfortunately, this algorithm is calculation intensive and it produces large errors in cases where photon statistics are poor. Moreover, it is more suitable for static measurements rather than dynamic (on-line). Two other simple algorithms for on-the-fly fluorescence lifetime determinations that have been evaluated are the maximum likelihood estimator (MLE) and the rapid lifetime determination method (RLD).^^

MLE calculates the lifetime via the following relation:^^

TIT mTlr ^ l + (e ' ^ -\)-m{e ' ^ -l)' = N;' Z iN. (4)

/ - I ^ where m is the number of time bins within the decay profile, At is the number of photocounts in the decay spectrum, Ni is the number of photocounts in time bin i, and T is the width of each time bin. A table of values using the left-hand side (LHS) of the equation is calculated by setting m and T to the experimental values and using lifetime values (if) ranging over the anticipated values. The right-hand side (RHS) of the equation is constructed from actual decay data over the appropriate time range. The fluorescent lifetime is determined by matching the value of the RHS obtained from the data with the table entry from the LHS. The relative standard deviations in the MLE can be determined using

Fluorescence lifetimes are calculated using the RLD method by integrating the number of counts within the decay profile over a specified time interval and using the following relationship:^^

T^ =-At/ln(D^/DQ) (5)

where At is the time range over which the counts are integrated. Do is the integrated counts in the early time interval of the decay spectrum, and Dj represents the integrated number of counts in the later time interval. Both the MLE and RLD methods can extract only a single lifetime value from the decay, which in the case of multi-exponential profiles would represent a weighted average of the various components comprising the decay.

Wolfrum and co-workers have developed a special pattern recognition technique for calling bases using lifetime discrimination methods.^^ Basically,


the method involves comparing a simulated decay pattern to the measured decay and searches for the pattern that best fits the measurement. This algorithm is equivalent to the minimization of a log-likelihood ratio, where fluorescence decay profiles serve as the pattern. Since the pattern recognition algorithm uses the full amount of information present in the data, it potentially has the lowest error or misclassification probability.

Soper and co-workers demonstrated the feasibility of performing online lifetime determinations during capillary gel electrophoresis (CGE) separation of DNA sequencing ladders.^^ C-terminated fragments produced from Sanger chain-termination protocols and labeled with a near-IR fluorophore at the 5' end of the sequencing primer were electrophoresed and the lifetimes of various components within the electropherogram were determined. The average lifetimes determined using the MLE method was found to be 581 ps with a standard deviation of ±9 ps (RSD=1.9%). This result indicated that MLE could produce high precision, even for ultra-dilute conditions. The favorable accuracy and precision was aided by the use of near-IR fluorescence detection, which minimized scatter contributions into the decay as well as background fluorescence.

23.4.2.1. Four-lifetime/One-lane

In the four-lifetime/one-lane method, four DNA ladders of differently labeled fragments generated by Sanger chain termination reactions are separated in one lane, typically in a single capillary gel column, and the base calling is done with lifetime discrimination as opposed to spectral discrimination. Obviously, the dye sets suitable for color discrimination are not necessarily appropriate for the use in lifetime discrimination. New dye-sets must be developed that suit this identification method. In lifetime discrimination, it is not necessary to use dyes with discrete emission maxima and therefore, structural variations in the dye set can be relaxed. For instance, Flanagan et al. developed a dye set that consisted of a series of structurally similar near-IR tricarbocyanines that possessed identical absorption (765 nm) and emission (794 nm) maxima.^^ The lifetimes of the dyes were varied by incorporating a single halide (I, Br, CI, or F) into the molecular structure. This dye set, with lifetimes ranging from 947 ps to 843 ps when measured in a polyacrylamide gel, have been suggested for use in a four-lifetime/one-lane sequencing experiment. These dye-labeled sequencing primers demonstrated uniform mobilities in gel electrophoresis applications, irrespective of the linker structures (see Table 3). However, since these were tricarbocyanine dyes, the lifetimes were found to be < 1.0 ns and the lifetime differences among the dye set was found to be somewhat small (Axf = 70 ps, - 8 % relative difference). Lassiter et al. optimized the experimental conditions for using these dyes for DNA sequencing by lifetime discrimination.^^

Wolfrum and co-workers demonstrated the use of a four-lifetime/one-lane approach for DNA sequencing using CGE. ^ In their work, four dye labels selected from the rhodamine, cyanine, and oxazine families were covalently tethered to the 5'end of oligonucleotide primers. The dyes exhibited similar


$0000 -| I * f 138

/

» ! ^ - ^ •s ^''^3cr^A

11

V

%mot t n o

3.8

3,2

h 3.0

r ^^ h 2.0

Figure 23.12. Electropherogram (solid line), fluorescence decay times (symbols), and errors (standard deviations in the decay time determination) of a small part of a sequencing run to illustrate the lifetime distribution inside and between peaks7^

absorption and emission maxima and were excited efficiently with a 630 nm pulsed diode laser operated at a repetition rate of 22 MHz. The labels exhibited fluorescence lifetime values of 1.6, 2.2, 2.9, and 3.7 ns, with the difference between the dyes adequate for efficient identification in sequencing applications (see Figure 12). This dye set allowed for the use of a simple detection system that was equipped with a single laser and a single avalanche photodiode. The instrumental response function of the entire system was measured to be -600 ps (FWHM). The time-resolved data were managed using the TCSPC technique. Using appropriate linker structures, dye-dependent mobility shifts were minimized, eliminating the need for post-electrophoresis corrections. The dye-labeled sequencing fragments were identified by both MLE and pattern recognition algorithms, with the latter method providing higher overall base-calling accuracy. Using an Ml3 template, Wolfrum et al. were able to demonstrate a read length of 660 bases with a probability of correct classification >90%.

23.4.2.2. Two-lifetime/TwO'lane

To show that fluorescence lifetimes could also be obtained from multiple electrophoretic lanes, a scanning system for measuring fluorescence lifetimes from multi-lane slab gels has been reported by Lassiter et al. '* In that report, a modified microscope head was inserted into an automated slab gel


a

Figure 23.13. Front-panel displays of the data analysis in analyzing time-resolved data from gel images, (a) The 3upper portion of the screen shows the gel intensity image obtained from a time-resolved fluorescence scanner system. The lower half of the screen shows the instrument response function profile constructed from a region on the gel that was defined by the vertical and horizontal cursors (location free from any dye), (b) The top portion of the screen shows the intensity electropherogram of a single base DNA sequencing tract. The bottom portion displays a decay profile representing a single image pixel selected by the cursor (single vertical line) from the electropherogram. '

sequencer, which consisted of a near-IR time-resolved scanning imager and implemented a two-lifetime/two-lane sequencing approach. Two dyes in each lane were identified by their lifetimes on-line during gel electrophoresis. Two commercially available cyanine-based near-IR dyes, IRD700 and Cy5.5 were chosen as fluorescence reporters and were labeled at the 5' end of a sequencing primer. A-terminated bases were labeled with IRD700 and T-terminated bases labeled with Cy5.5 were electrophoresed in one lane while C (IRD700) and G (Cy5.5) tracts occupied an adjacent lane. The similar absorption and emission properties of these two dyes allowed efficient processing of the emission on a single detection channel and excitation with a single source. The lifetimes for


these two dyes were calculated by the MLE algorithm and determined to be 718 ± 5 ps and 983 ± 13 ps for IRD700 and Cy5.5, respectively. Figure 13 shows an intensity electropherogram and the data analysis screen for this time-resolved fluorescence experiment. The base calling accuracy for an Ml3 template using this approach was 99.7% for 670 bases, better than the 95.7% calling accuracy obtained using a single-color/four-lane strategy carried out on the same instrument. The improved base calling accuracy using lifetime identification was a consequence of increased information content in the electrophoretic bands, particularly in those which were experiencing poor electrophoretic resolution.

23.4.3. Combination of Color-Discrimination and Time-Resolved Methods

Using fluorescence for readout has proven to be a viable multiplexed method for DNA sequencing. In this method, multiple analytes can be run in one track, increasing the information content (multiplexing) that can be processed simultaneously. Up to four reporters, one for each nucleotide base, have been identified in a single electrophoresis lane using either color (spectral) or lifetime multiplexing. Indeed, many commercially available automated DNA sequencers are equipped with four-color capabilities. However, the use of only color multiplexing limits the number of probes that can be identified simultaneously due to the broad emission profiles associated with most molecular labeling dyes. The coupling of lifetime discrimination with traditional color discrimination allows increased fluorescence multiplexing capability, enabling DNA sequencing to be performed with greater throughput. The basic reason is that during time-resolved measurements, all intensity-based data are preserved, while lifetime data adds an additional layer of information. For example, in a color and lifetime hybrid instrument, different dyes can be identified by color and in each color channel, different labels can be distinguished through their characteristic fluorescent lifetimes. Based on this concept, Soper and co-workers developed a two-color, time-resolved fluorescence microscope using near-IR fluorescence and first demonstrated the combination of color and lifetime discrimination schemes.^^

The hybrid microscope built for color/lifetime discrimination could acquire both steady state and time-resolved fluorescence data on-line during gel electrophoresis in either a capillary or a microchip format for DNA sequencing applications.^^' ^ A diagram of the optical fiber-based, dual-color, time-resolved microscope is shown in Figure 14. It consisted of tv^ pulsed diode lasers (680 and 780 nm), both operated at a repetition frequency of 40 MHz, which were coupled to the microscope head using single-mode fibers. The laser light was directed onto a focusing objective using a dichroic mirror, and the resulting emission was collected by the same objective and focused onto a multimode fiber, which transports the luminescence to two SPADs. The emission was sorted spectrally using a second dichroic beam splitter and isolated by appropriate interference filters before reaching one of the two SPADs (710-nm channel or 810-nm


mmm Reference puh^$

Z[:zi:™z3:,; S^4|MI

"SSHldi^'ij'

40 X 680-7S0OBDR 40 X

Figure 23.14. Schematic diagram of a dual-color, time-resolved, near-IR hybrid microscope. The microscope used two diode lasers providing excitation at 680 and 780 nm and a pair of SPADs for photon transduction.^^

channel). A PC-based card in a computer contained all the photon counting electronics to process the time-resolved data. A router was used to properly register photons generated from different detectors, allowing simultaneously measurements in two channels. The dual-color microscope demonstrated a time-response of 450 ps (FWHM) and 510 ps for the 710 and 810-nm channels, respectively. The use of near-lR fluorescence detection greatly simplified the hardware and allowed superior detection limits. The mass limits of detection were determined to be 7.1x10"^^ and 3.2x10'^^ mol (SNR = 3) for the two detection channels by electrophoresing two near-IR dye-labeled sequencing primers through a capillary gel column.

One of the benefits of applying time-resolved measurements is the ability to eliminate the cross-talk between different color channels. For example, using the two-color, time-resolved microscope for DNA sequencing, the problem of cross-talk accompanying almost all color discrimination methods was successfully solved by implementing optical delay filtering. Briefly, an extra length of optical fiber was inserted into the 680-nm laser pulse


train to introduce a phase shift of this laser relative to the 780 nm laser. The time delay of the 680 nm pulse train with respect to the 780 nm pulse train resulted in a phase shift of the corresponding fluorescence emission. As such, spectral leakage of a dye in the wrong channel was separated from the fluorescence emission resulting from the other dyes in the right detection channel. Then, by setting the counting electronics for each color channel to process photons only over the selected time frame while excluding the delayed spectral leakage signals, the elimination of cross-talk was achieved. The fluorescence intensity data, appearing as normal electropherograms, were constructed from the time-resolved data by integrating all photoevents for a preselected time interval (e.g. 1 s) and plotting versus electrophoresis time.

23.4.3.1. Two-color/Two-lifetime Sequencing

Coupling lifetime discrimination capabilities with color discrimination allows one to further increase the multiplexing power of DNA sequencing to provide reduced cost in obtaining sequencing information and also, increasing throughput. The common method for increasing system throughput is to add additional electrophoresis lanes, which requires additional sample processing pipelines in the system with the limit of lanes that can be added determined by the amount of real-estate that can be interrogated by the fluorescence detector without sacrificing signal-to-noise. For example, work has been demonstrated on performing 384 capillary gel separations simultaneously on a single device with the fluorescence readout accomplished using a rotary scanner. The throughput of such a device could be doubled without adding more separation lanes if 8 fluorescent dyes could be efficiently distinguished allowing the ability to run 2 different samples in the same gel lane. The use of color discrimination exclusively is difficult when processing multiple dye sets due to the broad emission profiles associated with most molecular systems. Therefore, the ability to build hybrid systems that provide fluorescence information via both color and lifetime is attractive for high multiplexing opportunities in DNA sequencing.

A two-color/two-lifetime sequencing run was carried out as an initial test of the color and time-resolve hybrid system approach.^^ Two dye-pairs, IRD700 and AlexaFluor680 were processed in a 710-nm color channel and IRD800 and near-IR-Br dye^^ were used in an 810-nm color channel. Each pair of dyes showed minimal differences in their excitation and emission profiles, but possessed distinguishable lifetimes in the sequencing matrix used for electrophoretic sorting of the DNA. The decay profiles for all four dyes as well as the instrument response function in each color channel are shown in Figure 15. Four single-base tracts were run prior to the sequencing experiment to determine the lifetime value of each dye label under typical electrophoresis conditions. The average lifetime values, calculated using the MLE algorithm, for each dye label were 913±9 ps, 1493111 ps, 454+7 ps and 744±16 ps, for IRD700, Alexa680, IRD800 and NIR-Br, respectively. Based on the


710 nm Channel

1' ^ N monrn Prompt

^L:-

_ , , , J

810 nm Channel

Ffom»5

^%;;

Cnannet Numbef {/12 2 p&) Channe.' Number (71? 2 ps)

Figure 23.15. Instrument response function and fluorescence decay profiles for a two-color/two lifetime hybrid instrument, (a) Fluorescence decay profiles for IRD700-labeled G- and AlexaFluor680-labeled T-fragments as well as the instrument response function for the 710 nm color channel, (b) Fluorescence decay profiles for IRDSOO-labeled C- and NlR-Br-labeled A-sequencing fragments and the instrument response function generated in the 810 nm color channel. Individual dye-labeled DNA ladders were analyzed using CGE with POP6 as the sieving matrix. Decays were constructed by integrating photocounts over 5 pixels (integration time of 5 s) centered on individual electrophoretic peaks from each sequencing trace. The instrument response functions were accumulated over 5 pixels from the gel track prior to migration of the DNA fragments into the detection volume.^^

predetermined lifetime values, an automatic peak recognition algorithm was applied to assist the identification of the terminal bases. The raw data were subjected to mobility shift corrections but without going through any other data manipulations that are normally involved in many automated sequencingmachines, such as removal of cross-talk, baseline adjustment or signal normalization. The calling accuracy was 95.1% over a read length of 650 base pairs, with the majority of errors occurring at late times within the electrophoresis trace due to poor separation efficiency at the end of the run.

The same strategy (two-color/two-lifetime) was also tested on a microchip electrophoresis format. ^ The sequencing samples were required to be purified and pre-concentrated prior to electrophoretic sorting to improve the precision of the lifetime determinations due to the low loading levels associated with the microchip format. The fluorescence lifetimes of all dyes were determined with favorable precisions on the microchips.

23.4.3.2. Potential Applications

The primary motivation for coupling time-resolved measurements with color discrimination is to increase the fluorescence multiplexing capability to increase the information content obtainable in a single gel lane, thereby increasing the data throughput. With appropriate dye sets, a two-color, four-


lifetime sequencing strategy could be envisioned to allow the identification of 8 unique reporters. This scheme could, for example, allow simultaneous forward and reverse reads from both ends of a double-stranded DNA. The final steps of most shotgun sequencing strategies involve sequencing from both ends of selected Ml3 clones to build a scaffold (map) and fill in map gaps using directed reads . The two-color/four-lifetime strategy could be effectively used here for front and back-end reads in a single lane, where 8 probes need to be analyzed simultaneously. Sequencing from both ends also yields important positional information since the distance between the read pair is known.^^

As multiplexed dye-systems using color discrimination are fiirther developed, lifetime identification methods could be incorporated into any color method to further increase the multiplexing. The potential extension of the hybrid strategy may also involve the use of FRET systems. For example, a measurement of four different fluorescence lifetimes in four spectrally distinct channels that use FRET would allow sixteen different parameters to be measured in a single lane and still, using a single excitation source to efficiently excite the donor dye. Such a dye set, however, remains to be developed.

23.5. INSTRUMENTAL FORMATS FOR FLUORESCENCE-BASED DNA SEQUENCING

Proper instrumentation design of a detector system is the cornerstone of any successful implementation of DNA sequencing, since the performance of the detector will dramatically affect read length and calling accuracy. Reading fluorescence during the electrophoretic separation of DNA sequencing ladders and accurately and efficiently identifying each terminal base offers a challenging task. A good sequencing instrument should possess the following characteristics: high sensitivity, high base-calling accuracy, high-throughput ability and system robustness. High sensitivity is of particular importance because separation platforms used to fractionate DNA are being developed with miniaturization as a major goal. Loading amounts as low as the attomole range

are often demanded. The detector must be able to read fluorescence signatures with reasonably high signal-to-noise in order to accurately call the base. As systems are further scaled-down, increases in sensitivity of the fluorescence reader will have to occur.

A successful DNA sequencer also requires excellent base identification capabilities. This has been carried out by accurately processing the fluorescence based either on their emission wavelength (color) or their fluorescence lifetimes. For color identification, the hardware must provide efficient excitation of the dye set and effectively sort the color into appropriate channels without producing significant amounts of cross-talk. For lifetime discrimination, the hardware must possess a favorable instrument response function so as to eliminate the need for deconvolution to simplify data


processing. Common to both color and lifetime discrimination are the ability to provide excellent discrimination of different fluorescence reporters. The dye sets used should possess broad spacing between their emission maxima for color sorting and broad spacing between their fluorescence lifetimes for lifetime discrimination.

In addition, system throughput is another important criterion by which to evaluate a sequencer. High-throughput requires many separation channels or lanes incorporated into a separation platform to process samples in parallel. As noted previously, system throughput can also be expanded by incorporating a higher level of multiplexing capability into the system as well. For multi-lane systems, all lanes need to be illuminated simultaneously with acceptable laser power and detected with high sensitivity at high duty cycles, since signal aliasing can be introduced affecting electrophoretic performance. Numerous designs and improvements have been made towards developing comprehensive instruments that are capable of fast, automated and sensitive operation even when operated in a high throughput mode.

DNA sequencing was initially performed using slab gels, in which up to 96 lanes could be run simultaneously with loading volumes in the tens of microliters. With the success of capillary-based separation technologies and the desire of increased automation, capillary-based instruments have become the second generation sequencing instruments. To achieve high-throughput production of DNA sequencing information for CGE, parallel sample processing in multiple capillaries, termed as capillary array electrophoresis (CAE), has been introduced. With the recent advances in miniaturizing capillary dimensions on planar microfabricated devices, chip-based DNA sequencing has offered increases in speed and throughput compared to capillary array electrophoresis. The increase in throughput using microchips has been realized not only by speeding up individual runs but also by implementing array formats on a single chip. As such, fluorescence detection systems for chip-based DNA sequencing need to be developed. For example, a 4-color rotary confocal scanner has been demonstrated for a 96-lane microchip array for DNA sequencing. "^

The instrument development for fluorescence-based DNA sequencing has led to the realization of high-throughput, high-performance automated DNA sequencers. For instance, the successful development of a 96-capillary CAE instrument, such as the MegaBACE and ABI PRISM® 3700 DNA analyzer, greatly contributed to the early completion of the Human Genome Project (HGP). The throughput of a slab gel machine was estimated to be -600 bases/hour. ^ Today, automated CAE sequencing instruments that use fluorescent dyes and laser scanners or imagers for production-scale sequencing allow processing 50,000 to 60,000 bases of sequence information in only a few hours.

Irrespective of the separation platform used, the fluorescence detection systems required for high throughput arrays can be categorized into two types — scanning or imaging. The following sections provide a few typical examples of each type.


23.5.1. Fluorescence-scanning Detectors

In scanning systems, the excitation beam is tightly focused and irradiates only a single lane at a time, with the relay optic rastered over the lanes of the gel or array of capillaries and the fluorescence from each lane processed sequentially on a set of 4 detection channels. The 96-capillary MegaBACE 1000 DNA sequencer from Molecular Dynamics is based on the confocal scanning design developed by Mathies and co-workers.^^' ^

A typical example of a confocal scanning system is depicted in Figure 2^ 61,95 jj^-g system used a confocal geometry with epi-illumination, in which the objective used to collect the emitted fluorescence from the center of each capillary also served to focus the laser beam into individual capillaries. Following fluorescence collection by the objective, the emission was focused onto a spatial filter at the secondary image plan of the objective. The laser light was directed into the objective using a dichroic beam splitter, which also allowed transmission of the emitted radiation onto the detection channels. The capillaries were positioned into a linear array. As can be seen, the laser irradiated one capillary at a time. However, since the beam was tightly focused (diameter =10 |Lim), the electronic transition can be saturated at relatively low laser powers, improving the signal-to-noise in the fluorescence measurement.^^ In addition, the

mmm-—

^ 1 0 ^ ^ HiAMmirTE*'

y^msmrt**^

Figure 23.16. Schematic diagram of a two-color, confocal-fluorescence capillary array scanner. The excitation source was an Ar ion laser operating at 488 nm. One PMT had a red filter (bandpass with center wavelength = 590 nm, half band width = 35 nm) while the other had a green filter (center wavelength = 525 nm, half band width = 10 nm). ^


noise can be significantly suppressed in this system, since a pinhole was used at the secondary image plane of the microscope objective, preventing scattered out-of-focus light generated at the walls of the capillary from passing through the optical system. The capillary array was scanned at a rate of 20 mm/s with the fluorescence sampled at 1500 Hz/channel (color channel) resulting in a pixel image size of 13.3 jam. The fluorescence was collected by a 32X microscope objective (numerical aperture = 0.4) with a geometrical collection efficiency of approximately 12%. Once the fluorescence had been collected by the objective, it was passed through the primary dichroic and processed onto one of two different PMTs, with each PMT sampling a different color (spectral discrimination). While this system was configured with only two color channels, the system could easily be configured to do four-color processing by inserting appropriate dichroics and bandpass filters into the optical train. The concentration limit of detection of this system was determined to be 2 x 10' ^ M (SNR = 3) by flowing a solution of fluorescein through an open capillary. The scanning could also be performed in a rotating fashion with capillaries set in a circular configuration or with a microchip situated on top of the scanner. ' ' ^

One important parameter that requires evaluation for a scanning system is its duty cycle, which takes into account the loss in signal due to multiple-lane sampling. For any type of scanning system, the sampling of the electrophoresis lanes is done in a sequential fashion. For example, a scanning system sampling 96 capillaries produces a duty cycle of approximately 1%. However, in the imaging systems, all capillaries are sampled continuously and the duty cycle is nearly 100%. Therefore, comparisons of detection limits for any system must include a term for the duty cycle, since lower duty cycles degrade the limit of detection.

bu^r solution

:n ! j I

<hea9i'l3«

0«S$ mm Q^cipiiary

Figure 23.17. Diagram of an imaging laser-induced fluorescence detector for reading four-color fluorescence from a capillary-gel array, (a) View of the multiple sheath-flow cell. DNA fragments eluted from the gel-filled capillaries and were focused by a sheath-flow stream and drawn into the lower open capillaries situated 1 mm away, (b) Schematic of a multiple sheath-flow gel capillary array instrument. Combined beams from an argon ion laser (488 nm) and Nd:YAG laser (532 nm) traverse below the output of the gel columns and fluorescence is collected with a lens system and imaged onto a CCD camera.^^


23.5.2. Fluorescence-imaging Detectors

In a fluorescence imaging system, all of the electrophoresis lanes are irradiated by a laser(s) simultaneously, with the fluorescence readout using a multi-channel detector, such as a CCD. The ABI PRISM® 3700 DNA sequencer by Applied Biosystems employs a sheath flow design based on Kambara and co-workers' results, which incorporated an imaging system for reading fluorescence from multiple capillaries simultaneously.^^' ^ In this example, an array of capillaries was aligned inside a rectangular sheath-flow quartz cell as illustrated in Figure 17. A buffer solution was pumped through the interstitial space between the capillaries and the walls of the cuvette. The liquid sheath flow entrained the DNA sequencing fragments eluting from the columns and the flow was tapered into a smaller diameter with one stream per capillary. The laser beams (488 nm from Ar ion; 532 nm from frequency doubled Nd:YAG) were made collinear using a dichroic mirror and were focused into the cuvette, exciting the fluorescent molecules across all flow streams at a position slightly below the exit end of the capillaries. This geometry eliminated the need for a laser beam traveling through each capillary as required by a scanning system, which would cause significant scattering and reduce the intensity of the beam as it traveled through the array. In order to achieve multicolor processing, the collected fluorescence emission was sent through an image-splitting prism to produce four separated (spectrally) line images on a CCD camera. A set of narrow bandpass filters was placed in front of the prism to assist in isolating the appropriate colors for data processing. The collection optic and focusing optic produced a total magnification of 1 and resulted in a geometrical collection efficiency of 1%. The main advantage of this off-column detection strategy is the improved detection limits, since the major contributor to the background, scattering from the gel matrix, is absent here. ^^ The detection limit reported for this system was found to be 2 x 10' ^ M when operated in the four-color mode. The main challenges in this scheme when considering scaling up to multiple capillaries includes the alignment of the individual sheath flows, matching the laser beam waist to the core diameters of the fluid streams containing the eluted fragments and attenuation of the laser beam over a long distance due to cumulative absorption.

23.5.3. Time-resolved Fluorescence Scanning Detectors

The above imaging and scanning high-throughput sequencing systems are all based on the color discrimination. High-throughput applications can also be carried out through time-resolved measurements using scanning or imaging systems as well.

Neumann and co-workers have described a time-resolved scanner for reading 16 capillary arrays.^^ This scanner consisted of a pulsed diode laser, operating at 640 nm with a repetition rate of 50 MHz and an avalanche diode detector. A confocal imager configured in a time-correlated single-photon-counting arrangement was kept stationary. The capillary array was linearly


Avi^aclMs HK)tfid^^

Figure 23.18. Block diagram of a time-resolved near-IR laser-induced scanner. The laser was mounted at 56° with respect to the scanning surface to minimize light reflection from the glass surfaced"*

translated through the detection zone. Up to 16 capillaries were interrogated at scan rates approaching 0.52 Hz.

Lassiter and co-workers reported on the integration of a near-IR time-solved fluorescence scanner with an automated slab gel sequencer for lifetime-based DNA sequencing. Due to the size of the slab gel and the gel plates, the gel medium could not be translated underneath the relay optics of the detector. Therefore, a scanner containing the entire detection optics that could move in a linear fashion over the gel plate was constructed (see Figure 18). A pulsed diode laser operating at 680 nm was mounted on a microscope head at a 56° angle with respect to the boro-float gel plates to minimize reflected radiation from being coupled into the optical system. The laser diode was driven by an electrical short-pulse generator, which supplied a repetition rate of up to 80 MHz. The laser radiation was focused onto the surface of the gel plates using an ^71.4 lens that produced a spot on the gel of approximately 20 |im x 30 |Lim (elliptical beam shape of diode laser). The emission generated from the gel was subsequently collected byy71.2 optics mounted in the microscope head, filtered with a single band-pass filter and then focused onto the face of an actively quenched avalanche photodiode with a large photoactive area. All photon counting electronics were integrated into a PC board and situated in a computer to process the time-resolved data. The simple instrumental reconfiguration


implemented in this work demonstrated that many existing machines, which use steady-state fluorescence, could be easily modified to do time-resolved measurements as well acquiring both steady-state and the time-resolved data.

23.5.4. Time-resolved Fluorescence Imaging Detectors

The ability to construct imaging time-resolved systems has been made feasible for DNA sequencing applications with the advent of multi-channel detectors, such as CCDs, that demonstrate ultra-fast time resolution . These multi-channel detectors are similar to conventional CCD cameras used for steady-state measurements, except that they also consist of gated-image intensifiers that provide the ultra-fast time resolution. The intensifier is a phosphor plate generating nearly 200 photons per incident photo-event, thus providing gain (~200-fold) for high sensitivity measurements. The phosphor plate acts as an optical shutter as well providing the prerequisite time resolution. For a time-resolved measurement to construct the decay profile from which the lifetime could be extracted, the image intensifier is triggered with the excitation pulse from the laser to initiate image acquisition. Therefore, the system also requires a laser operating in a pulsed-mode. This acquisition can be delayed and shifted in time with respect to the excitation laser pulse by 25 ps steps (for 25 ps steps over a range of 12 ns, each decay profile would consist of 480 data points) to construct the entire decay curve. The timing resolution of such a system would depend on the speed of the gate, the pulse width of the excitation laser and the bin width of the MCA used to accumulate the individual photon events of the decay profile. The optical arrangement would be similar to that used for steady state measurements required for spectral discrimination, in that the pulsed laser would need to simultaneously irradiate either an array of capillaries or a series of microchannels configured on a chip. The resulting fluorescence would then be collected by a relay optic and imaged onto the multi-channel, time-resolved detector.

23.6. REFERENCES

1. E. S. Lander; et al Initial sequencing and analysis of the human genome, Nature (London, United Kingdom) 409, 860-921 (2001).

2. F. S. Collins; E. D. Green; A. E. Guttmacher; M. S. Guyer A vision for the future of genomics research, Nature (London, United Kingdom) 422, 835-847 (2003).

3. P. Onyango The role of emerging genomics and proteomics technologies in cancer drug target discovery. Current Cancer Drug Targets 4, 111-124 (2004).

4. R. C. Hardison Comparative genomics, PLoSBiology 1, 156-160 (2003). 5. J. D. Watson; F. H. Crick Molecular structure of nucleic acids; a structure for deoxyribose

nucleic acid. Nature 171, 737-738 (1953). 6. E. Y. Chen; D. Schlessinger; J. Kere Ordered shotgun sequencing, a strategy for integrated

mapping and sequencing of YAC clones, Genomics 17, 651-656 (1993). 7. J. L. Weber; E. W. Myers Human whole-genome shotgun sequencing, Genome Research 7,

401-409(1997).

570 L- ZHU and S.A. SOPER

8. A. M. Maxam; W. Gilbert A new method for sequencing DNA, Proceedings of the National Academy of Sciences of the United States of America 74, 560-564 (1977).

9. F. Sanger; S. Nicklen; A. R. Coulson DNA sequencing with chain-terminating inhibitors, Proceedings of the National Academy of Sciences of the United States of America 74, 5463-5467(1977).

10. H. V. Thome Electrophoretic characterization and fractionation of polyoma virus DNA, Journal of Molecular Biology 24, 203-211 (1967).

11. D. A. McGregor; E. S. Yeung Optimization of capillary electrophoretic separation of DNA fragments based on polymer filled capillaries, Journal of Chromatography 652, 61-1?^ (1993).

12. H. Drossman; J. A. Luckey; A. J. Kostichka; J. D'Cunha; L. M. Smith High-speed separations of DNA sequencing reactions by capillary electrophoresis, Analytical Chemistry 62, 900-903 (1990).

13. A. T. WooUey; R. A. Mathies Ultra-High-Speed DNA Sequencing Using Capillary Electrophoresis Chips, Analytical Chemistry 67, 3676-3680 (1995).

14. W. Ansorge; B. Sproat; J. Stegemann; C. Schwager; M. Zenke Automated DNA sequencing: ultrasensitive detection of fluorescent bands during electrophoresis. Nucleic Acids Research 75,4593-4602(1987).

15. W. Ansorge; B. S. Sproat; J. Stegemann; C. Schwager A non-radioactive automated method for DNA sequence determination, Journal of Biochemical and Biophysical Methods 13, 315-323 (1986).

16. J. M. Prober; G. L. Trainor; R. J. Dam; F. W. Hobbs; C. W. Robertson; R. J. Zagursky; A. J. Cocuzza; M. A. Jensen; K. Baumeister A system for rapid DNA sequencing with fluorescent chain-terminating dideoxynucleotides. Science (Washington, DC, United States) 238, 336-341 (1987).

17. L. M. Smith; J. Z. Sanders; R. J. Kaiser; P. Hughes; C. Dodd; C. R. Connell; C. Heiner; S. B. H. Kent; L. E. Hood Fluorescence detection in automated DNA sequence analysis. Nature (London, United Kingdom) 321, 674-679 (1986).

18. L. G. Lee; C. R. Connell; S. L. Woo; R. D. Cheng; B. F. McArdle; C. W. Fuller; N. D. Halloran; R. K. Wilson DNA sequencing with dye-labeled terminators and T7 DNA polymerase: effect of dyes and dNTPs on incorporation of dye-terminators and probability of termination fragments, Nucleic Acids Research 20, 2471-2483 (1992).

19. H. Swerdlow; R. Gesteland Capillary gel electrophoresis for rapid, high resolution DNA sequencing. Nucleic Acids Research 18, 1415-1419 (1990).

20. S. Carson; A. S. Cohen; A. Belenkii; M. C. Ruiz-Martinez; J. Berka; B. L. Karger DNA sequencing by capillary electrophoresis: use of a two-laser-two-window intensified diode array detection system, Analytical Chemistry 65, 3219-3226 (1993).

21. A. E. Karger; J. M. Harris; R. F. Gesteland Multiwavelength fluorescence detection for DNA sequencing using capillary electrophoresis. Nucleic Acids Research 19, 4955-4962 (1991).

22. M. L. Metzker; J. Lu; R. A. Gibbs Electrophoretically uniform fluorescent dyes for automated DNA sequencing, Science (Washington, D. C.) 271, 1420-1422 (1996).

23. J. Y. Ju; C. C. Ruan; C. W. Fuller; A. N. Glazer; R. A. Mathies Fluorescence Energy-Transfer Dye-Labeled Primers for DNA- Sequencing and Analysis, Proceedings of the National Academy of Sciences of the United States of America 92, 4347-4351 (1995).

24. J. Y. Ju; 1. Kheterpal; J. R. Scherer; C. C. Ruan; C. W. Fuller; A. N. Glazer; R. A. Mathies Design and Synthesis of Fluorescence Energy-Transfer Dye- Labeled Primers and Their Application for DNA-Sequencing and Analysis, Analytical Biochemistry 231, 131-140 (1995).

25. J. Y. Ju; A. N. Glazer; R. A. Mathies Cassette labeling for facile construction of energy transfer fluorescent primers. Nucleic Acids Research 24, 1144-1148 (1996).

26. S. C. Hung; R. A. Mathies; A. N. Glazer Optimization of spectroscopic and electrophoretic properties of energy transfer primers. Analytical Biochemistry 252, 78-88 (1997).

27. S. C. Hung; R. A. Mathies; A. N. Glazer Comparison of fluorescence energy transfer primers with different donor-acceptor dye combinations. Analytical Biochemistry 255, 32-38 (1998).

28. B. B. Rosenblum; L. G. Lee; S. L. Spurgeon; S. H. Khan; S. M. Menchen; C. R. Heiner; S. M. Chen New dye-labeled terminators for improved DNA sequencing patterns. Nucleic Acids Research 25, 4500-4504 (1997).

29. S. Kumar; S. Nampalli; B. F. McArdle; C. W. Fuller; (Amersham Pharmacia Biotech, Inc., USA). Application: WO, 1999, 40 pp.


30. D. C. Williams; S. A. Soper Ultrasensitive Near-IR Fluorescence Detection for Capillary Gel Electrophoresis and DNA Sequencing Applications, Analytical Chemistry 67, 3427-3432 (1995).

31. T. R. Gilson; P. J. Hendra Laser Raman Spectroscopy, 1970. 32. L. R. Middendorf; J. C. Bruce; R. C. Bruce; R. D. Eckles; S. C. Roemer; G. D. Sloniker A

versatile infrared laser scanner/electrophoresis apparatus, Proceedings of SPIE-The International Society for Optical Engineering 1885, 423-434 (1993).

33. G. Patonay; M. D. Antoine Near-infrared fluorogenic labels: new approach to an old problem. Analytical Chemistry 63, 321A-322A, 324A-327A (1991).

34. R. B. Mujumdar; L. A. Ernst; S. R. Mujumdar; A. S. Waggoner Cyanine dye labeling reagents containing isothiocyanate groups, Cytometry 10, 11-19 (1989).

35. L. Strekowski; M. Lipowska; T. Gorecki; J. C. Mason; G. Patonay Functionalization of near-infrared cyanine dyes. Journal of Heterocyclic Chemistry 33, 1685-1688 (1996).

36. S. A. Soper; Q. L. Mattingly Steady-State and Picosecond Laser Fluorescence Studies of Nonradiative Pathways in Tricarbocyanine Dyes: Implications to the Design of Near-IR Fluorochromes with High Fluorescence Efficiencies, Journal of the American Chemical Society 116, 3744-3752 (1994).

37. L. R. Middendorf; J. C. Bruce; R. C. Bruce; R. D. Eckles; D. L. Grone; S. C. Roemer; G. D. Sloniker; D. L. Steffens; S. L. Sutter; et al. Continuous, on-line DNA sequencing using a versatile infrared laser scanner/electrophoresis apparatus, Electrophoresis 13,487-494 (1992).

38. J. H. Flanagan, Jr.; C. V. Owens; S. E. Romero; E. Waddell; S. H. Kahn; R. P. Hammer; S. A. Soper Near-infrared heavy-atom-modified fluorescent dyes for base-calling in DNA-sequencing applications using temporal discrimination, Analytical Chemistry 70, 2676-2684 (1998).

39. L. Strekowski; M. Lipowska; G. Patonay Facile derivatizations of heptamethine cyanine dyes. Synthetic Communications 22, 2593-2598 (1992).

40. L. Strekowski; M. Lipowska; G. Patonay Substitution reactions of a nucleofugal group in heptamethine cyanine dyes. Synthesis of an isothiocyanato derivative for labeling of proteins with a near-infrared chromophore, Journal of Organic Chemistry 57,4578-4580 (1992).

41. M. Lipowska; G. Patonay; L. Strekowski New near-infrared cyanine dyes for labeling of proteins. Synthetic Communications 23, 3087-3094 (1993).

42. D. B. Shealy; M. Lipowska; J. Lipowski; N. Narayanan; S. Sutter; L. Strekowski; G. Patonay Synthesis, Chromatographic Separation, and Characterization of Near-Infrared Labeled DNA Oligomers for Use in DNA Sequencing, Analytical Chemistry 67, 247-251 (1995).

43. D. Wrobel; A. Boguta Study of the influence of substituents on spectroscopic and photoelectric properties of zinc phthalocyanines. Journal of Photochemistry and Photobiology, A: Chemistry 150, 67-76 (2002).

44. N. Kobayashi; H. Ogata; N. Nonaka; E. A. Luk'yanets Effect of peripheral substitution on the electronic absorption and fluorescence spectra of metal-free and zinc phthalocyanines, Chemistry—A European Journal 9, 5123-5134 (2003).

45. E. F. Bradbrook; R. P. Linstead Preparation of the ten dicyanonaphthalenes and the related naphthalenedicarboxylic acids, Journal of the Chemical Society, Abstracts 1739-1744 (1936).

46. E. I. Kovshev; E. A. Luk'yanets Phthalocyanines and related compounds. XL Substituted 2,3-naphthalocyanines, Zhurnal Obshchei Khimii 42, 1593-1597 (1972).

47. W. M. Sharman; S. V. Kudrevich; J. E. van Lier Novel water-soluble phthalocyanines substituted with phosphonate moieties on the benzo rings. Tetrahedron Letters 37, 5831-5834 (1996).

48. B. L. Wheeler; G. Nagasubramanian; A. J. Bard; L. A. Schechtman; M. E. Kenney A silicon phthalocyanine and a silicon naphthalocyanine: synthesis, electrochemistry, and electrogenerated chemiluminescence, Journal of the American Chemical Society 106, 7404-7410(1984).

49. G. A. Casay; J. Lipowski; T. Czuppon; N. Narayanan; G. Patonay Spectroscopic investigations of a tetrasubstituted aluminum naphthalocyanine near-infrared compounds. Spectroscopy Letters 27, A\l-A31 (1994).

50. I. McCubbin; D. Phillips The photophysics and photostability of zinc(II) and aluminum(III) sulfonated naphthalocyanines. Journal of Photochemistry 34, 187-195 (1986).

51. T. Shen; Z. Yuan; H. Xu Fluorescent properties of phthalocyanines, Lfyes and Pigments 11, 77-80(1989).

572 L- ZHU and S.A. SOPER

52. R. P. Hammer; C. V. Owens; S.-H. Hwang; C. M. Sayes; S. A. Soper Asymmetrical, water-soluble phthalocyanine dyes for covalent labeling of oligonucleotides, Bioconjugate Chemistry 13, 1244-1252 (2002).

53. M. Ambroz; A. Beeby; A. J. MacRobert; M. S. C. Simpson; R. K. Svensen; D. Phillips Preparative, analytical and fluorescence spectroscopic studies of sulfonated aluminium phthalocyanine photosensitizers, Journal of Photochemistry and Photobiology, B: Biology 9, 87-95(1991).

54. W. Ansorge; J. Zimmermann; C. Schwager; J. Stegemann; H. Erfle; H. Voss One label, one tube, Sanger DNA sequencing in one and two lanes on a gel. Nucleic Acids Research 18, 3419-3420(1990).

55. S. Tabor; C. C. Richardson DNA sequence analysis with a modified bacteriophage T7 DNA polymerase. Effect of pyrophosphorolysis and metal ions, Journal of Biological Chemistry 265,8322-8328(1990).

56. S. Tabor; C. C. Richardson DNA sequence analysis with a modified bacteriophage T7 DNA polymerase, Proceedings of the National Academy of Sciences of the United States of America 84,4161-4111 (1987).

57. S. L. Pentoney, Jr.; K. D. Konrad; W. Kaye A single-fluor approach to DNA sequence determination using high performance capillary electrophoresis. Electrophoresis 13, 467-474 (1992).

58. H. Swerdlow; J. Z. Zhang; D. Y. Chen; H. R. Harke; R. Grey; S. Wu; M. J. Dovichi; C. Fuller Three DNA sequencing methods using capillary gel electrophoresis and laser-induced fluorescence, Analytical Chemistry 63, 2835-2841 (1991).

59. D. Chen; H. R. Harke; N. J. Dovichi Two-label peak-height encoded DNA sequencing by capillary gel electrophoresis: three examples. Nucleic Acids Research 20, 4873-4880 (1992).

60. H. R. Starke; J. Y. Yan; J. Z. Zhang; K. Muehlegger; K. Effgen; N. J. Dovichi Internal fluorescence labeling with fluorescent deoxynucleotides in two-label peak-height encoded DNA sequencing by capillary electrophoresis. Nucleic Acids Research 22, 3997-4001 (1994).

61. X. C. Huang; M. A. Quesada; R. A. Mathies DNA sequencing using capillary array electrophoresis. Analytical Chemistry 64, 2149-2154 (1992).

62. E. K. Lewis; W. C. Haaland; F. Nguyen; D. A. Heller; M. J. Allen; R. R. MacGregor; C. S. Berger; B. Willingham; L. A. Bums; G. B. I. Scott; C. Kittrell; B. R. Johnson; R. F. Curl; M. L. Metzker Color-blind fluorescence detection for four-color DNA sequencing. Proceedings of the National Academy of Sciences of the United States of America 102, 5346-5351 (2005).

63. S. A. Soper; B. L. Legendre, Jr.; D. C. Williams Online Fluorescence Lifetime Determinations in Capillary Electrophoresis, Analytical Chemistry 67, 4358-4365 (1995).

64. S. L. Mcintosh; B. K. Nunnally; A. R. Nesbit; T. G. Deligeorgiev; N. 1. Gadjev; L. B. McGown Fluorescence Lifetime for On-the-Fly Multiplex Detection of DNA Restriction Fragments in Capillary Electrophoresis, Analytical Chemistry 72, 5444-5449 (2000).

65. B. K. Nunnally; H. He; L.-C. Li; S. A. Tucker; L. B. McGown Characterization of Visible Dyes for Four-Decay Fluorescence Detection in DNA Sequencing, Analytical Chemistry 69, 2392-2397(1997).

66. L. Li; L. B. McGown Effects of gel material on fluorescence lifetime detection of dyes and dye-labeled DNA primers in capillary electrophoresis. Journal of Chromatography, A 841, 95-103(1999).

67. L. Li; L. B. McGown Comparison of sieving matrices for on-the-fly fluorescence lifetime detection of dye-labeled DNA fragments, Fresenius' Journal of Analytical Chemistry 369, 267-272(2001).

68. L.-C. Li; H. He; B. K. Nunnally; L. B. McGown On-the-fly fluorescence lifetime detection of labeled DNA primers, Journal of Chromatography, B: Biomedical Sciences and Applications 695,85-92(1997).

69. L.-C. Li; L. B. McGown On-the-Fly Frequency-Domain Fluorescence Lifetime Detection in Capillary Electrophoresis, Analytical Chemistry 68, 2737-2743 (1996).

70. L.-C. Li; L. B. McGown Improving signal to background ratio for on-the-fly fluorescence lifetime detection in capillary electrophoresis. Electrophoresis 21, 1300-1304 (2000).

71. H. He; L. B. McGown DNA Sequencing by Capillary Electrophoresis with Four-Decay Fluorescence Detection, Analytical Chemistry 72, 5865-5873 (2000).

72. H. He; B. K. Nunnally; L.-C. Li; L. B. McGown On-the-Fly Fluorescence Lifetime Detection of Dye-Labeled DNA Primers for Multiplex Analysis, Analytical Chemistry 70, 3413-3418 (1998).


73. K. Chang; R. K. Force Time-resolved laser-induced fluorescence study on dyes used in DNA sequencing, Applied Spectroscopy 47, 24-29 (1993).

74. S. J. Lassiter; W. Stryjewski; B. L. Legendre, Jr.; R. Erdmann; M. Wahl; J. Wurm; R. Peterson; L, Middendorf; S. A. Soper Time-resolved fluorescence imaging of slab gels for lifetime base-calling in DNA sequencing applications, Analytical Chemistry 72, 5373-5382 (2000).

75. S. J. Lassiter; W. Stryjewski; C. V. Owens; J. H. Flanagan, Jr.; R. P. Hammer; S. Khan; S. A. Soper Optimization of sequencing conditions using near-infrared lifetime identification methods in capillary gel electrophoresis. Electrophoresis 23, 1480-1489 (2002).

76. U. Lieberwirth; J. Arden-Jacob; K. H. Drexhage; D. P. Herten; R. Mueller; M. Neumann; A. Schulz; S. Siebert; G. Sagner; S. Klingel; M. Sauer; J. Wolfrum Multiplex Dye DNA Sequencing in Capillary Gel Electrophoresis by Diode Laser-Based Time-Resolved Fluorescence Detection, Analytical Chemistry 70, 4771-4779 (1998).

77. M. Neumann; D. P. Herten; A. Dietrich; J. Wolfrum; M. Sauer Capillary array scanner for time-resolved detection and identification of fluorescently labelled DNA fragments, Journal of Chromatography, A 871, 299-310 (2000).

78. M. Sauer; J. Arden-Jacob; K. H. Drexhage; F. Gobel; U. Lieberwirth; K. Muhlegger; R. Muller; J. Wolfrum; C. Zander Time-resolved identification of individual mononucleotide molecules in aqueous solution with pulsed semiconductor lasers, Bioimaging 6, 14-24 (1998).

79. M. Sauer; B. Angerer; K. T. Han; C. Zander Detection and identification of single dye labeled mononucleotide molecules released from an optical fiber in a microcapillary: First steps towards a new single molecule DNA sequencing technique, Physical Chemistry Chemical Physics 1, 2471-2477 (1999).

80. M. Sauer; B. Angerer; W. Ankenbauer; Z. Foldes-Papp; F. Gobel; K. T. Han; R. Rigler; A. Schulz; J. Wolfrum; C. Zander Single molecule DNA sequencing in submicrometer channels: state of the art and future prospects, Journal of Biotechnology 86, 181-201 (2001).

81. S. Seeger; G. Bachteler; K. H. Drexhage; J. Arden-Jacob; G. Deltau; K. Galla; K. T. Han; R. Mueller; M. Koellner; et al. Biodiagnostics and polymer identification with multiplex dyes, Berichte der Bunsen-Gesellschaft 97, 1542-1548 (1993).

82. E. Waddell; S. Lassiter; C. V. Owens, Jr.; S. A. Soper Time-resolved near-IR fluorescence detection in capillary electrophoresis. Journal of Liquid Chromatography & Related Technologies 23, 1139-1158 (2000).

83. E. Waddell; Y. Wang; W. Stryjewski; S. McWhorter; A. C. Henry; D. Evans; R. L. McCarley; S. A. Soper High-Resolution Near-Infrared Imaging of DNA Microarrays with Time-Resolved Acquisition of Fluorescence LifQtimQs, Analytical Chemistry 72, 5907-5917 (2000).

84. Y. Zhang; S. A. Soper; L. R. Middendorf; J. A. Wurm; R. Erdmann; M. Wahl Simple near-infrared time-correlated single photon counting instrument with a pulsed diode laser and avalanche photodiode for time-resolved measurements in scanning applications, Applied Spectroscopy 53,497-504 (1999).

85. S. A. Soper; J. H. Flanagan, Jr.; B. L. Legendre, Jr.; D. C. Williams; R. P. Hammer Near-infrared, laser-induced fluorescence detection for DNA sequencing applications, IEEE Journal of Selected Topics in Quantum Electronics 2, 1129-1139 (1996).

86. L. Zhu; W. Stryjewski; S. Lassiter; S. A. Soper Fluorescence Multiplexing with Time-Resolved and Spectral Discrimination Using a Near-IR Detector, Analytical Chemistry 75, 2280-2291 (2003).

87. L. Zhu; W. J. Stryjewski; S. A. Soper Multiplexed fluorescence detection in microfabricated devices with both time-resolved and spectral-discrimination capabilities using near-infrared fluorescence. Analytical Biochemistry 330, 206-218 (2004).

88. B. L. Legendre, Jr.; C. C. Williams; S. A. Soper; R. Erdmann; U. Ortmann; J. Enderlein An all solid-state near-infrared time-correlated single photon counting instrument for dynamic lifetime measurements in DNA sequencing applicafions. Review of Scientific Instruments 67, 3984-3989(1996).

89. S. A. Soper; B. L. Legendre, Jr. Error analysis of simple algorithms for determining fluorescence lifetimes in ultradilute dye solutions, Applied Spectroscopy 48,400-405 (1994).

90. P. Hall; B. Selinger Better estimates of exponential decay parameters. Journal of Physical Chemistry 85, 2941-2946 (1981).

91. R. M. Ballew; J. N. Demas An error analysis of the rapid lifetime determination method for the evaluation of single exponential decays. Analytical Chemistry 61, 30-33 (1989).


92. M. Koellner; A. Fischer; J. Arden-Jacob; K. H. Drexhage; R. Mueller; S. Seeger; J. Wolfrum Fluorescence pattern recognition for ultrasensitive molecule identification: comparison of experimental data and theoretical approximations, Chemical Physics Letters 250, 355-360 (1996).

93. A. Edwards; C. T. Caskey Closure strategies for random DNA sequencing, Methods (San Diego, CA, United States) i , 41-47 (1991).

94. B. M. E. Paegel, C.A.; Wedemayer, G.J.; Scherer, J.R.; Mathies, R.A. High Throughput DNA Sequencing with a Microfabricated 96-Lane Capillary Array Electrophoresis Bioprocessor, Proceedings of the National Academy of Sciences of the United States of America 99, 574-579 (2002).

95. X. C. Huang; M. A. Quesada; R. A. Mathies Capillary array electrophoresis using laser-excited confocal fluorescence detection, Analytical Chemistry 64, 967-972 (1992).

96. S. A. Soper; H. L. Nutter; R. A. Keller; L. M. Davis; E. B. Shera The photophysical constants of several fluorescent dyes pertaining to ultrasensitive fluorescence spectroscopy. Photochemistry andPhotobiology 57, 972-977 (1993).

97. J. R. Scherer; I. Kheterpal; A. Radhakrishnan; W. W. Ja; R. A. Mathies Ultra-high throughput rotary capillary array electrophoresis scanner for fluorescent DNA sequencing and analysis. Electrophoresis 20, 1508-1517 (1999).

98. H. Kambara; S. Takahashi Multiple-sheathflow capillary array DNA analyser, Nature 361, 565-566(1993).

99. S. Takahashi; K. Murakami; T. Anazawa; H. Kambara Multiple Sheath-Flow Gel Capillary-Array Electrophoresis for Multicolor Fluorescent DNA Detection, Analytical Chemistry 66, 1021-1026(1994).

100. H. Swerdlow; S. Wu; H. Harke; N. J. Dovichi Capillary gel electrophoresis for DNA sequencing: laser-induced fluorescence detection with the sheath flow cuvette. Journal of Chromatography 516, 61-67 (1990).

Figure 4.8. (Page 121, Liu et al.) Morphological changes of cell nucleus during the process of apoptosis and necrosis with AO and EB dying.

Wav^Iengtti {nm)

Figure 8.3. (Page 186, Ruan et al.) Novel optical properties of QDs for improving the sensitivity of in-vivo bioimaging. (a) Comparison of fluorescence light emission from organic dyes (TRITC - left vial), green QDs (middle vial), and red QDs (right vial) under normal room light illumination and at the same molar concentration (1.0 DM for dyes and QDs). Bright fluorescence emission is observed from QDs but not from the dye, due to the large absorption cross sections of QDs. (B) Photobleaching curves showing that QDs are several thousand times more photostable than organic dyes under the same excitation conditions, (c) A comparison of the excited state decay curves (monoexponential model) between QDs and common organic dyes. The longer excited state lifetimes of QD probes allow the use of time-domain imaging to discriminate against the background fluorescence (short lifetimes), (d) Comparison of mouse skin and QD emission spectra obtained under the same excitation conditions, demonstrating that the QD signals can be shifted to a spectral region where the autofluorescence is reduced.

575

576 COLOR INSERTS

Figure 8.6. (Page 190, Ruan et al) Dynamics of endosomal fusion. Selected time points in the cytoplasm of a CHO cell expressing erbBl-eGFP, 30 min after addition of 200 pM 6:1 QD-EGF. The vesicles show Brownian movement interrupted by directed movement and fusions. At time 0, the lower-right vesicle (red arrow) moved towards the upper-left vesicle (blue arrow), with which it fused irreversibly at 25 s (purple arrow). Scale bar, 5 ^m. This figure demonstrates that QD-EGF can illuminate the RTK-dependent signaling pathway.

Figure 8.7. (Page 190, Ruan et al.) Subcellular localization of QDs. QDs were conjugated to localization sequence peptides, which permit active transport to the nucleus (NLS, A) or mitochondria (MLS, B), and were delivered to 3T3 fibroblast cells by microinjection. A) Fluorescence and phase micrographs of a cell 24h after co-injection of QD-NLS with 70 kDa rhodamine dextran control. The four spots in the nucleus that are not stained with QDs are the nucleoli. B) Fluorescence and phase micrographs 24h after injection of QD-MLS. Colocalization with Mictotracker Red confirms mitochondrial labeling. C) QDs remain fluorescent after 8min of continuous mercury lamp exposure, while conventional MitoTracker dye (D) bleaches beyond detection after 30s of continuous excitation. Different cells were imaged for (C, D).

COLOR INSERTS 577

|J^^||||||ji§|:jte^|||^ m

Figure 15.3. (Page 353, Borst et al.) An image of a typical FLIM experiment using the Becker and Hickl analysis software. In panel A the fluorescence intensity of the nuclear localized MADS box protein, FBP24-ECFP, is shown. At the blue crosshair the fluorescence decay of the selected pixel is displayed (panel B). The fluorescence lifetimes are calculated per pixel and visualized as a pseudo color image (panels C and D).

T(ns)

Figure 15.4. (Page 353, Borst et al.) FRET-FLIM analyses in double exponential decay model of transfected cowpea leaf protoplasts, expressing the following combinations FBP2-ECFP and FBP24-EYFP (panels A-C) and FBP24-ECFP and FBPll-EYFP (panels D-F). In panels A and D the fluorescence intensity image of the nucleus of a representative cell is shown, in panels B and E the fluorescence lifetime image of the same nucleus is shown as a pseudo color image, and in panels C and F the distribution of fluorescence lifetimes over the nucleus is presented.

578 COLOR INSERTS

Figure 17.2. (Page 393, Truong et al.) A. FISH on interphase nuclei and a partial metaphase of non-small-cell lung cancer cells. The chromosome 3 long arm and short arm specific paintings were revealed in green and red, respectively. Arrow: 2 normal metaphase chromosome 3; Arrow heads: interphase nuclei. Nuclei and chromosome are colored in orange by propidium iodide B. Interphase nuclei in a biopsie of a lung cancer after FISH of chromosome 3 painting (green: long arm, red: short arm., blue: Dapi. C. Chromosome 3 arm imbalance for lung cancer biopsy with two clones. The histogram displays the number of nuclei as a ftinction of the imbalance measured on the slide. Cells with imbalance values around 1 and 1.6 were selected on the histogram in red and green, respectively, and located on the tissue slide using the same colors. Cells without imbalance were mainly present in the upper portion of the slide whereas cells with imbalances around 1.6 were located in the lower portion (after [1]). D. Field of an ammiotic fluid preparation after FISH of chromosome 21 (green), and 22 (red). Interphase nuclei are labelled in blue by Dapi.

COLOR INSERTS 579

Figure 18.8. (Page 413, Bozym et al) Microscope calibration with a 1536 well plate. Ratio images of each well (365/543 nm) containing zinc ion buffers, apoH36C-AF594 CA and Dapoxyl sulfonamide; pseudo colored for visual comparison.

Figure 18.9. (Page 414, Bozym et al) Measuring free zinc inside of PC-12 cells with apoTAT-H36C-AF594 and Dapoxyl sulfonamide. Top left: bright field; top right: ratio image pseudo color rainbow; bottom left: excitation 543 nm, emission 617 nm, exposure time 20 ms, pseudo color red; bottom right: excitation 365 nm, emission 617 nm, exposure time 500 ms, pseudo color red. The calibration bar (at right) indicates the level of cytoplasmic free zinc.

580 COLOR INSERTS

N- domain of TnC

C

rtp20

of c Till

Deactivated State (Mg^^-saturated)

Activated State (Ca^^-saturated)

Figure 20.1. (Page 447, Cheung). A model for the proximity relationship of troponin, tropomyosin (Tm), and actin to illustrate Ca^^-induced changes in the tertiary structure of cTnC (blue) and the secondary structure of cTnl (red). For simplicity, cTnT is omitted. In the deactivated state, the C-domain of cTnC is saturated with two Mg^^ and the N-domain is unoccupied by Ca^ . In the activated state, the N-domain of cTnC is saturated by Ca" ^ at its single site, and bound Mg" ^ in the C-domain is replaced by Ca^ . The disposition of myosin head is arbitrary. Residues in cTnC that are involved in FRET measurements are shown. The positions of several key residues in cTnl are indicated to show changes in secondary structure (inhibitory region) and movement of the regulatory region into the N-domain of cTnC.

Figure 21.1. (Page 467, Aguilera et al) Microscopic analysis of the toxic effects of plumbagin (Ql) and 3-phenyl-1,4-naphthoquinone oxide (Q23). Using fluorescence microscopy, a group of HeLa-GFP cells was monitored for signs of cell death at two hour intervals over a 10 hour period. Obvious signs of cell-death induction were clearly visible by 6 hours (hrs) of chemical exposure by membrane "blebbing" (see filled arrows) and by an increase in nuclear DNA condensation (open arrows).

COLOR INSERTS 581

Porphyrins

NADH

Collagen

Elastin

Flavins

Figure 22.7. (Page 496, Elson et al) Endogenous fluorophores and the StrEF. (a) shows the fluorescence lifetime map of the five fluorophores and the lifetime histogram across all pixels is shown in (b).

T =800-2600 ps, formalin-fixed x = 400-2900 ps, frozen x = 1000-2500 ps,fresh human pancreas, 10 jam section.human colon, lOjum section, mouse kidney, 200 |Lim section.

Figure 22.8. (Page 497, Elson et al.) Examples of FLIM images of fixed, frozen and fresh tissue.

w °°2

>< ^ 0.015

© ^ 0.01

E

0.005

0 A / ^ J H ^ . ,-

l i i 1, \M\ '- Wl f V

— Plaque

— Media

— Whole

-..^-AV\./Wy' ISOiji ^'^•'^3ii:^ti!x~^

Lifetime (ps)

Figure 22.10. (Page 501, Elson et al.) Fluorescence lifetime map and corresponding lifetime histogram of a sectionof artery showing lifetime contrast between the media and the plaque. (Figure 22.9 follows on the next page.)

582 COLOR INSERTS

Region A

Region

Sample I: healthy

2Q00ps

H&E stain

Sample II: diseased 4C0ps

Region C

15(X)ps

ITOOps

Region D

1700ps

eoops Region E

Figure 22.9. (Page 499, Elson et ah) H&E images and FLIM maps of different regions of healthy and diseased human cartilage. The FLIM maps were recorded for the areas highlighted on the H&E images. Regions A and B are from thin and thick regions of healthy cartilage respectively. C contains the frayed cartilage surface and subchondral bone, D contains a fissure with some regions of short lifetime, whilst E shows a region of cartilage which appears healthy according to the H&E stain.

COLOR INSERTS 583

Basement membrane

Figure 22.11. (Page 502, Elson et al) TCSPC FLIM - intensity images (a,c,f) and intensity-weighted FLIM maps (b,d,g) of fresh 200 |Lim-thick tissue sections of cervix and the corresponding H+E-stained 6 |im-thick tissue sections cut subsequently after fixing the 200 |im sections in formahn. (a,b) Normal (post-menopausal) cervix; (c-e) CIN II; (f-h) CIN III. The CIN II images shows a 'tide mark' within the epithelium (dashed line) representing the change from immature dysplastic cells in the basal half of the epithelium to maturing dysplastic cells showing cytoplasmic keratinisation in the upper half; in the CIN III images the full epithelial thickness is occupied by immature dysplastic cells. Note the autofluorescence from the areas containing immature dysplastic cells shows shorter lifetimes and greater fluorescence intensity than from areas occupied by maturing keratinised dysplastic cells or normal epithelium. This difference may be due to greater metabolic activity and higher NADH levels in the immature cells. (Two photon excitation wavelength = 740 nm; detection range 385-600 nm; x20 microscope objective).

584 COLOR INSERTS

Figure 22.12. (Page 503, Elson et al) Wide-field time-gated FLIM of a formalin-fixed 10 jLim-thick tissue section of pancreas showing partial invasion of an Islet of Langerhans (dotted line) by malignant cells (solid line and arrows), (a) Light microscope image of an H+E stained section; (b) False colour FLIM map; (c) Intensity weighted FLIM image. FLIM reveals the uninvaded portion of the islet to be contrasted to both the invaded area and the surrounding malignant parenchjona due to its shorter fluorescence lifetime. There is however no apparent contrast between the malignant parenchyma and the normal pancreatic parenchyma. (Excitation wavelength = 400nm, detection range > 435 nm, microscope objective = x40).

Figure 22.13. (Page 506, Elson et al) FLIM of autofluorescence of a 10 ^m human pancreatic tissue acquired at 10 frames per second.

Figure 22.14. (Page 507, Elson et al) (a) Arthroscope used to obtain endoscopic FLIM images of bisected lambs kidney and FLIM images acquired using (b) conventional WNLLS at 0.1 Hz and (c) real time (7.2 Hz) FLIM acquisition and processing.

COLOR INSERTS 585

Figure 22.15. (Page 508, Elson et al) (a) Flexible 10 mm diameter endoscope and (b) FLIM image acquired at 5.5 Hz of the mucosal surface of piece of a bisected lamb's kidney viewed en-face with 355nm excitation.

o c

1 W T-

c

i 3 o O

8 c E

1 QC

CO O m

2 0) ; - .

O ) CO

CO CO i 5

Figure 22.18. (Page 513, Elson et al.) (a) Fluorescence intensity and (b) fluorescence lifetime images of a fixed unstained 10 jim section of human pancreas obtained in a wide-field time-domain FLIM system with A,ex = 440-450 nm (A - small artery, I - islet of Langerhans and C - connective tissue), (c) composite wide-field fluorescence intensity and FLIM images of a multiwell plate array of fluorescent dyes excited with four different wavelength bands of 20 nm width centred on 500, 515, 570 and 610 nm respectively. The upper part shows the first time-gated fluorescence intensity image and the lower part shows the corresponding FLIM images. Reproduced from Journal of Physics D: Applied Physics © 2004 lOP Publishing Ltd.

586 COLOR INSERTS

(a) (b)

Figure 22.19. (Page 513, Elson et al.) Wide-field fluorescence (a) intensity and (b) lifetime images of a Troma stained section of mouse lung tissue.

Figure 22.21. (Page 516, Elson et al.) FLIM images calculated from excitation-emission-lifetime data set for (a) 480 nm excitation, 490-620 nm emission band; (b) 550 nm excitation, 490-620 nm emission band and (c) 480 nm excitation, 620-835 nm emission band.

COLOR INSERTS 587

1 D D

1 ° D LJ

c R - - - - ' ^ ^ ' ^ ' ^ O%0CO3C00CC<^F><'<> " "^

n"^ rO

kifp o Fibrin rich

2000 4000 6000 8000 10000 12000 14000

Delay (ps)

D ° ° n ^

o o o o o ^ o o

a Lipid rich

2000 4000 6000 8000 10000 12000 14000

Delay (ps)

Figure 22.22. (Page 517, Elson et al.) Fluorescence lifetime images of carotid artery in emission bands (a) 415-455 nm and (b) 570-635 nm. (c) shows the mean wavelength and (d) shows the spectral width as a ftmction of the delay time after excitation for the lipid rich and fibrin rich regions highlighted in (a).

INDEX

Page numbers with / and/indicate table and figure respectively.

1,5-IAEDANS [5- (iodoacetamidoethyl) aminonaphtha-lene-sulfonic acid], 448-449 2,6-dipicolinate, 401 2-aminopurine (2-AP)

motional dynamics of, 330^ role of, 329

3-phenyl-l ,4-naphthoquinone oxide (Q23), microscopic analysis of toxic effects of, 467/ 4-(dimethylamino)-cinnamaldehyde (DMA), 149 4,6-diamino-2-phenylindole (DAPI), 391 4-chloro-7-nitrobenzo-2-1 -diazole (NBD), 539 6-MA P (4-amino-6-methyl-8-(2-deoxy-beta-d-riboftiranosyl)-7(8H)-pteridone), 315 70S prokaryotic ribosome, 243 8-methoxy-quinoline, 410 8-vinyl-deoxyadanine (8-VA), 313

a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor, 399 A/D (analog/digital) converter, 128 Aal: p-(l-azulenyl)-L-alanine, 48 ABCD criteria, 359 ABI PRISM®3700 DNA sequencer, 564, 567 ablation

completeness of, 126 margins, during RFA, 126

absolute cell death, detection of, 133 actin (A), 455 ADP, 460 Aequorea victoria, 464 AF488 reorientation, 254 AF488-CaM interactions, 254

aggregation-only model, 294 ALA-induced fluorescence, 167 ALA-induced protoporphyrine IX-fluorescence, 167 alamethicin, 66 aldehyde-appended-dye techniques, 150 Alexa Fluor 588, 411-412 Alexafluor 482 (AF488), 244 AlexaFluor680, 561 allele-specific hybridization, 209 alpha-amino carbon-centered radicals, 154 Alzheimer's disease, 142 amino acid homocysteine (Hey, 1), 139 aminoacid TO AC, 67 AmplexRed, 116 analog-to-digital conversion (ADC) technique, 269, 271 analog-todigital converter (ADC), 553 aneuploid cell lines, 387 angiogenic squamous dysplasia (ASD), 173 anionic lipids, 47 annexin V-FITC staining, 468 antibiotic peptides, 47, 66 antifungal polyene nystatin, 66 apocarbonic anhydrase, 411 apoE knockout mouse model of atherosclerosis, 143 apoptosis. See programmed cell death apoptotic and necrotic process, difference between, 120 Arabidopsis MADS box transcription factor interactions, 351 Arrhenius model, 126 ATP cycle, 460 ATP hydrolysis in RecA-mediated DNA recombination, role of, 327-328 auto fluorescence, 165-166, 172 autocorrelation decay, 244

589

590 INDEX

autofluorescence bronchoscopy (AFB), 167-170, 172 technical details of, 173/

automated image cytometry, 388 avalanche photodiodes, 221

B bacterial artificial chromosomes (BACs) See Yeast Artificial Chromosomes (YACs), BAPTA-based (1,2-bis(oaminophenoxy) ethane-N,N,N' ,N' -tetraacetic acid) Ca+2 indicators, 406 barrel stave model, 47 Beer's law, 377 benzopophyrin derivative (BPD), 368 Bessel integrals, 75 BigDye™ dye sets, 540 bioactivity mechanism experiments, 60 biocompatible silica materials, 277

characterization of, 278 biological FLIM data acquisition, 270 biological fluorescence, 2 biological macromolecules, 2 biomedical fluorescence imaging, 261 biothiol detection based on simple arrays, 156-157 Bloch equations, 29 BODIPY-TMR fluorescence, 201 BODIPY-TR fluorescence, 201 Bohr condition of quantum energy jumps, 25 Bohr-theory of atom, 10 bound fluorescent nucleotide analogue 2 deoxy-mant-ADP, 460 bovine serum albumin (BSA), 143, 149 Breslow thickness, 362

calcium activation of cardiac muscle, 446 Calcium

binding of, 345 importance of, 344

calmodulin (CaM), 243 calmodulin-CFP-YFP cameleon, 243 CaM (CaM-DA), globular domains of, 243-244, 249 CaM construct, generation of, 243, 245 CaM dynamics, 245/ CaM-DA construct, 249-250 CaM-DA, MEM lifetime distribution of, 250 capillary array electrophoresis (CAE), 564

capillary gel electrophoresis (CGE), 556 carboxy-4', 5'-dichloro-2', 7'-dimetoxyfluorescien (JOE), 539-540, 550 carboxyfluorescein (FAM), 539-540, 550 carboxyfluorescein release, rate of, 60 carboxyfluorescein, 49 carboxylate-amine condensation method, advantages of, 183 carboxytetramethyl-rhodamine (TAMRA), 539-540 carboxy-X-rhodamine (ROX), 539-540 carcinogens, detection of, 465 carcinoma in situ (CIS) (TisNOMO), 163-164 cardiac troponin, 447 cardiovascular diseases, origin of, 141 carmine fluorescence, 120 carpet model, See Shai-Matsuzaki-Huang model Casmir effect, 18 cationic peptides, 47-48 CCD camera technology, 166,489, 567, 569 CD spectra in water, 50 CDF family member ZnT-3, 400 cell apoptosis, 119 cell membrane, ways to destroy

hypotonic homogenization, 110 rubbing, 110 ultrasonication, 110

cell-mediated cytotoxicity assays, 465 cellular fluorescence, sources of

FMN (flavin mononucleotide), 483 flavins FAD (flavin adenine dinucleotide), 483 NADH, 483

cellular injuries, types of cellular organelle integrity, 134 chromatin clumping, 134 nuclear membrane integrity, 134

chelation enhanced fluorescence (CHEF), 410 chip-based DNA separation systems, 535 chipvideobronchoscopy system, 166 Chlamydomonas reinhardtii, 324 chlorinated hydrocarbons (CHCs), 426, 428 chlorophyll, 2 chromophore molecule, 1-2 chromosome aneuploidy, 388 chromosome arm paintings, 387

INDEX 591

chromosome imbalances, 387-390 in situ determination of, 389

chromosome trisomy, 388 cisplatin complexes, 465 codon optimized for mammalian expression, 464 collision theory, 12-13 color discrimination methods in DNA

sequencing four-color/single-lane, 550-551 single-color/four-lane, 547 single-color/single-lane, 548 two-color/single-lane, 549-550

combination of color-discrimination and time-resolved methods potential applications, 562-563 two-color/two-lifetime sequencing, 561-562

comparison for DNA fragment separations, 535/ computed tomography (CT), 125 computed x-ray tomography, 186 conformational transitions, kinetic parameters for, 453^54 constant fraction discriminator (CFD) circuitry, 486, 553 Conway theory of atom, 10 Coulomb perturbation between the two molecules, 27 Cr-release assay, 465 CTL assays, 465 cTnC domains, 455 cTnC mutant, 448, 454 cTnC N-domain markers, 456 cTnC N-domain, opening and closing of, 455-456 cTnC-cTnl complex, 449, 459 cTnC-cTnl interface, FRET markers for, 452 cTnl-cTnlcTnC interface, 455 cTnl mutant (Trpl92Cys), 449, 458 cTnl residues, 458 CTNI secondary structure, FRET markers for, 449 CTnl-cTnC complex, 458 CTnl-cTnT copmplex, 459 CTnT-TnC copmplex, 459 CY fusion, 272 CY24,271 Cy5.5, 558 cyanine dye, photoinduced isomerization of, 242 Cy-dye series, 545

cysteine in position 51 (Asn51 Cys), 449 cysteine. See also zinc ligands

automated post-column detection of, 156 detection of, 147 selective detection of, 149-151

cytoplasmic GFP-fluorescence, 465 cytotoxic naphthoquinone, 467 cytotoxic T-lymphocyte (CTL), analysis of, 465

D 5-aminolevulinic acid (ALA), 165 DAFE,5'ee Wolf system DAPI, See DNA dynamics, fluorescence probes for DDPM (N-[4-dimethylamino)-3,5-dinitrophenylj-maleimide), 449, 458 deacetylase Sir2, 109 DEAL column, 449 Debye interaction, 19 Debye-Stokes-Einstein relation, 284 delta-aminolevulinic acid (ALA), 368 deoxynucleotides (dATPs), 549 Deoxyribonucleic Acid (DNA), 526

chemical nature of, 526 polymerization of, 529 structure of, 526-529

deoxyribonucleosides (dNTPs), 530 dephasing effect, 88-89, 91 depth-dependent quenching, method of, 64 dermoscopy, 360 dextran@,61 DOS sols, particle growth in, 290-295 DIAL, 431 dideoxy sequencing method, 529 dideoxyribonucleoside triphosphates

(ddNTP), different DNA nucleotide analogs of, ddATP, 530, 549 ddCTP, 530, 549 ddGTP, 530, 549 ddTTP, 530, 549

differential optical absorption spectroscopy (DOAS), 422, 430 diffuse-reflectance spectroscopy, 361 diglycerylsilane8(DGS), 278 dihematoporphyrin ether/ester, 165 diode-pumped ultrafast Ti: Sapphire oscillator-amplifier system, 498 dipole-dipole interactions, 22

592 INDEX

dipyrrometheneborondifluoride fluorophores (BODIPY), 539 D-Light AF-System, 165, 168

development of non-laser system, 166, 172

DNA condensation by nucleocapsid protein, 323 physical techniques applied to study, 317 structure and dynamic of condensed DNA, 318-322 YOYO-1 marker for DNA condensation, 317-318

DNA damage-inducible RAD54 promoter, 464 DNA dynamics in chromosomes from picogreen fluorescence, 323-325 DNA dynamics, fluorescence probes for

non-specific, 313, 316 site-specific fluorescent probes, 313 specific, 313

DNA dynamics, site specific probes for DNA dynamics in RecA -DNA filaments, 325-328 mismatch recognition and DNA dynamics, 329-331 position-dependent DNA dynamics, 328-329

DNA labeling and sequencing, fluorescent dyes for, 537 near-IR fuorescent dyes, 541-546 visible fluorescent dyes, 538-541

DNA mismatch repair (MMR) system, 329 DNA oligonucleotides, 535 DNA sequencing, 525-526

color discrimination method, 547 fluorescent detection for, 535-536 gel electrophoresis, 530-535 strategies, 546-551

DNA sequencing, categories of, dye-primer sequencing, 536 dye-terminator sequencing, 536

DNA chemical nature of, 526-527 double helix structure of, 311 dynamic nature of, 312 fluorescent labeling of, 315 rigidity of, 311 separations, 533 synthesis, 530

DNA-bound YOYO-1, fluorescence intensity of, 319^ double ratio techniques, 368 Down's syndrome, See trisomy 21 dreariness of Global Gene, 109 drug-induced gluorescence, 165 dual channel FLIM, 273 dye-modified ddNTPs, 537 dynamic fluorescence quenching, 11-12 dynamical theory, 7 dysplasia, bronchoscopic detection of, 164

ECFP, 271 273 fluorescence lifetime, 272 fusion, 272 moieties, 272

EGTA, 447, 456 elastic scattering spectroscopy

absorption properties, 369 clinical studies, 374-376 instruments, 371-373 parameters in, 369 preclinical trials, 373-374 principles of, 369-370

elastin fluorescence, origin of, 483 electrical and magnetic substances interaction, 5 electrodynamics, theory of, 6 electrokinetic injection, 533 electromagnetic (EM) field, 3, 6, 8

concept of, 9, 12 electromagnetic communication, 5 electromagnetic radiation, 6 electron paramagnetic resonance spectroscopy, 278 electron-transfer inhibitor, 112 electroosmotic flow (EOF), 533 ELISA-like sandwich immunoassay, 199-201 Ellman's reagent, 146 electromagnetic theory, 3 empirical correction models of fluorescence in turbid media, 376-377 endogenous fluorophores, important

types, collagen, 495 elastin, 495 flavins, 495 NADH, 50 porphyrins, 495

INDEX 593

energy transfer between FIO and A3 analogs, 59/ changes due to water-membrane partition, 57 derivation of rate of, 31 due to peptide aggregation inside the membrane, 59 F.Perrin's calaculation of dynamic rates of, 23 from phycocyanin, 31 J.Perrin's calaculation of dynamic rates of, 24 theoretical observations of, in solution, 20-21 total standard free energy change of transfer, calculation of, 57 ways of, 32

energy transfer, illustration of application of quantum mechanical theory, 12-17 sensitized fluorescence, 11-12 spectroscopic and collisional cross-sections in vapour, 12

enhanced green fluorescent protein (EGFP), 464^65 his-tagged version of, 207-208

entropy-driven hydrophobic interactions, 48 enzymatic cycling assay, 111 enzymes, use of, 209 epitope tag fusions, 198 Escherichia coli, DNA damage SOS response of, 464 Escherichia coli, GFP-based models of, 465 ESS data, application of, 370 ethidium bromide. See DNA dynamics, fluorescence probes for ethyl ester of zinquin, 402 Eu-7 complex, 408 eukaryotic cells, 400

based assays, 466 screening of, 470-472

eukaryotic organisms, 399 european multicenter trial for finding dysplasia, 170 europium (Eu ) complex, 408 evanescent illumination, 217

intensity of, at interface, 218 measurement of, wave depth, 226 polarization, 218 use of, 216

excitation energy, pulses of femtosecond duration, 266 rapid, 266

excitation sources, types of ion lasers Ar, 538 He-Ne, 538 Kr,538

excitation-emission matrix spectroscopy, 365 excited-state intramolecular proton transfer (ESIPT), 409-410 exciton energy levels, 28 EYFP moieties, 272-273

Fabry-Perot resonator, 83 FAM dye, 543 Faraday induction, 5 fast (GHz bandwidth) sampling oscilloscopes, 485 femtosecond laser, 266 femtosecond Ti: Sapphire laser, 478-479 Fermi's golden rule, 19, 33, 35, 38, 101 fiberbronchoscopy in local anesthesia, 166 filaments, 432, 434, 449, 455 finite element methods (FEM), 87 finite element time domain methods (FETD), See finite element methods FlAsH tagged genetic fusions, 198 flat surface microarrays (DNA chips), 208 FLIM instrumentation, 480

limiting factor of, 505 FLIM microscopy of biological tissue

artery wall and atherosclerotic plaques, 500-501 cartilage, 497-500 neoplastic tissue, 501-503

FLIM-based FRET techniques, 264 advantages of, 348 data analysis, 349 instrumental setup, 349 molecular interaction imaging via FRET-FLIM, 351-355 sub-cellular localization via confocal microscopy, 350-351

flow cytometry, 203-205, 210 assays, 465 measurement of optical properties using, 196-197 particle-based analysis using, 197-198

594 INDEX

fluorescein isothiocyanate (FITC), 391 fluorescence bronchoscopy, 167, 173

advantages of, 174 fluorescence correlation spectroscopy

(FCS),216 use of metals, 232

fluorescence correlation spectroscopy (FCS), 239 intramolecular dynamics, measurements of, 241 intramolecular dynamics of diffiising species, techniques to detect, 240

fluorescence crosscorrelation spectroscopy (FCCS), 242 fluorescence decay experiments, 51-52, 54-55, 56/ fluorescence decay profiles, 492 fluorescence excitation, fiberoptic based spectroscopy system with, 127 fluorescence fluctuation autocorrelation functions, high-order, 230 fluorescence in situ hybridization (FISH), 387, 389-390,395 fluorescence intensity distribution analysis (FIDA), 230, 240 fluorescence intensity multiple distribution analysis (FIMDA), 240 fluorescence lifetime imaging (FLIM),

261-266, 346, 348, 477, 483, 485 biological applications of, 270-273 data analysis, 271,349 identification protocols, advantages of, 552 images, 486, 497, 505 methods for analyzing, 480 principles of, 480 requirements for, 267-270 setting of, 506 study of, 497 time-and frequency-domain approaches to, 480

fluorescence lifetime, 480-483 complex decay profiles and the stretched exponential function, 492-493 imaging, 487^91 of endogenous fluorophores, 482-483 single point measurement of, 484-487

wide-field time-domain FLIM instrumentation, 493-495

fluorescence lifetimes, observations for healthy cartilage sample I femoral heads, 498, 500 healthy cartilage sample II femoral heads, 498-500

fluorescence microscope, 2 fluorescence modulation, 484 fluorescence polarization, 20-21

F. Perrin model, 22-23 J. Perrin model, 21-22

fluorescence quantum efficiency, 263 fluorescence quenching, 263

relative, 62 fluorescence spectroscopy, 74, 131, 445

instrumentation, 363 melanoma diagnosis by autofluorescence, 363-368 melanoma diagnosis with exogenous fluorophores, 368-369 to detect thermal tissue damage, 126-127

fluorescence techniques for biopsy specimen classification autofluorescence bronchoscopy, 168-169 drug-induced fluorescence, 167

fluorescence tools, 1 fluorescence-based DNA sequencing

strategies color discrimination methods, 547-551 combination of color-discrimination and time-resolved methods, 559-563 lifetime discrimination methods, 552-559

fluorescence-based microplate assays, 465 fluorescent bronchoscopy, 163

phenomenon and techniques, 165 fluorescent cholera toxin B subunit (F-CTB), binding of, 206 fluorescent indicators based on

fluorescein zinpyr family, 402^04 zinspy family, 404-405 ZnAFs, 405^06

fluorescent proteins, hybrids of, CFP,2 GFP,2

INDEX 595

YFP,2 RFP,2

fluorometric methods, 112 fluorone black, 150, 153 fluorophore BODIPY, 67 fluorophore, anisotropy decay of, 252, 254 fluorophores types in hyaline cartilage,

2,6- dime thyld ifuro-8-pyrone (DDF), 498 collagen pyridinium crosslinks, 498 pentosidine, 498

fluorophores, 2, 64 decay times of, 82, 84/ interactions, 71-72, 74

FluoZin-3, 407 Fmoc: fluorenyl-9-methylcarbonyl group, 48 Forster resonance energy transfer

(FRET), 1,445^46,479 background for, 6 between organic chromophores in condensed systems, 20 definition of, 1 experiments, 62 principles of, 342^ quantum mechanical theories of, 3 standard expression for the efficiency of, 16

Forster's theory, 23, 25, 35, 37, 41 Fourier transform infrared spectroscopy (FTIR), 422, 430 fragment detection approach (PF/FD),

425 laser ionization recombination emission (LIRE), 426 luminescence methods of laser-induced fluorescence (LIF), 426, 428^30 prompt emission (PE), 426 stimulated emission (SE), 426, 429

Franck's principle, 12 free radical damage, 143 free sulfhydryl groups, 183 free zinc concentration determination, importance of, 401 frequency modulated (FM) approach, 489 frequency modulated LED, 491 frequency quadrupled optical parametric amplifier system, 498

frequency-domain approach, 487 frequency-domain fluorescence detection, 484 frequency-domain lifetime measurements, 484 Fresnel theory, 74-75 FRET combinations used in cell biology

CFP enhanced forms, 343 DsRED, 344 red fluorescent protein (RFP), 344 Venus mCerulean-Venus, 344 YFP enhanced forms, 343

FRET distances, distributions of, 459 FRET imaging, intensity based

acceptor photo-bleaching, 346-348 confocal and wide-field FRET imaging, 346 spectral imaging, 346

FRET kinetic results, 455 FRET measurement, samples of,

micro structures, 2 living biological cells, 2 whole organisms, 2

FRET sensors Cameleons (Ycam), 344-345 Caspase sensor, 345 FLAME, 345

FRET signal, 456 full width half-maximum (fwhm), 270, 373 Fura-2, 406 FuraZin-1,407 fusion protein, 183

y-aminobutyric acid (GAB A A) receptor, 399 galvano-driven mirrors, 265 gated image intensifier technology, 494 gated optical image intensifier (GOI) 489,493-494, 508 Gaussian mode, cell population detection using, 391 Gaussian peak, 534 gel electrophoresis

capillary electrophoresis (CE), 532-535 capillary Gel electrophoresis (CGE), 532-535 microchip gel electrophoresis ( i-CGE), 535

596 INDEX

slab gel electrophoresis (SGE), 531-532

genotoxins, detection of, 465 GFP based assays, 465

advantage of, 470 for drug discovery, 467-468

GFP fluorescence, 465 GFP inhibition and cytotoxicity, correlation between, 465 GFP, targeted expression of, 466 GFP-based fluorolysis method, 465 GFP-based toxicity assays in muhicellular organisms, 466 GFP-expressing Escherichia coli strains, testing of, 471^72 GFP-expressing mammalian cells, use of, 465 GFP-expressing Mycobacterium avium strains, testing of, 471-472 GFP-fluorescence level after chemical exposure, 472^ glutathione. See zinc ligands glycation, process of, 483 gradient force, 99 green fluorescent protein (GFP), 479

application of, 343 biosensor, as, 464 cloning of, 464 properties of, 464 use of, 463

GreenScreen genotoxicity assay, 464 group velocity dispersion (GVD), 432, 506 growth-only model, 293

H H&E staining, 133 HAD solvent, 182 HCFCs (hydrogen-containing chlorofluorocarbons), 422 Hey cellular levels, effect of, 141 heavy-atom modified tricarbocyanine dyes, chemical structures and photophysical properties of, 543/ HeLa cells, 465 HeLa-GFP assay, 467

determination of cytotoxicity of antibacterial compounds, 468-469

helium-cadmium laser, generation of excitation light, 166 hemato-porphyrin derivative (HpD), 165, 368

Henyey-Greenstein (HG) approximation, 370 hepatic parenchyma, 131 hepatic thermal damage, system for, 132 Hertzian oscillating dipole, 7-9 Hertzian oscillator, 20 heterodyne detection, 484 heteronuclear NMR spectroscopy, 458 high magnification bronchoscopy, 173 high-performance liquid chromatography (HPLC), 368 his-tag coupling method, 184 histidine. See zinc ligands histologic grading system, 132 histologic tissue damage, 132 histological classification, 167 Hoechst 33258, 344 homeostasis of multicellular organisms, 119 homocysteine

automated post-column detection of, 156 detection of, 147 149 metabolism, 141-142 potent reducing agent, function of, 153-154 role of, in disease, 143-144 selective detection of, 151-156

homodyne detection, 484 HPLC postcolum detection methods, 140 human articular cartilage, specimens of, 497 human carbonic anhydrase II (CA), 411 human cell line (HEK293-TP53::EGFP), 465 human flap endonuclease-1 (FEN-1), 202-204 human genome project (HGP), 525, 564 hydrodynamic injection, 533 hydrophobic mismatch, 60 hydrophobic peptides, 47 hyperhomocysteinemia, 141

development of atherosclerosis in, 143 effect of oxidative stress by, 142

hyperspectral FLIM instrumentation, 514-516

IAATMR(5-and6-idoactamidotetramethylrhodamine), 458

INDEX 597

lAEDANS-DABMI (4-dimethylaminophenylazophenyl-4'-maleimide) pair, 459 IgG antibody coupling, 183 in vivo imaging

endoscopic FLIM, 506-508 real-time FLIM, 504-506

inchoate apoptotic cell, 120 IndoZin-1,407 induction interaction. See Debye interaction instrument response function (IRF), 270, 553 instrumental formats for fluorescence-

based DNA sequencing fluorescence-imaging detectors, 567 fluorescence-scanning detectors, 565-566 time-resolved fluorescence imaging detectors, 569 time-resolved fluorescence scanning detectors, 567-569

intake-calorie restrictively (CR), 109 intermolecular dipole-dipole interactions,

16-17 in FRET, 17-19 in van der Waal forces, 17-19

internal charge transfer (ITC) mechanism, 406 internal labeling strategy, 549 international system for staging lung cancer, 163 Intracellular imaging, application of

cellular staining, 189 intracellular studies, 189-190

intracellular NADH fluorescence intensity, 115-116 iodoacetamides, 145 IR absorption techniques, 430 IRD700, 558, 561

K kappa square, 39 Keesom orientation effect, 19 kinesin conformations, nucleotide-dependent, 460 kinesin ATP state, 460 Krumdieck tissue slicer, 496, 501

labeled biomolecules, 2 Lab VIEW software, 494

lactate dehydrogenase, 111 lanthanide chemosensors, 408-409 laser induced breakdown spectroscopy (LIBS), 423-425 laser induced fluorescence endoscopy. See lung imaging fluorescence endoscopy laser photofragmentation, 425

laser scanning confocal microscope, 117 laser scanning microscope, 265

scan rates, 265 suppression of out-of-focus light, 265 two-photon excitation with direct detection, 266

laser-based spectroscopic techniques, 421 laser-induced fluorescence (LIF), 423,

535 use of, 360-361

Legendre functions, 97 Leica TCS SP2 prismbased tunable emission filter, 511 Levenberg Marquardt algorithm, 271, 490 Lewis lung carcinoma cells, use of, 466 LIBS, 425-426,430

detection systems, 434 signal, 434

LIDAR applications, 433, 438 configuration, 430 signals, 434 techniques, 432

LIFE bronchoscopy, 166 lifetime determination, algorithms used,

nonlinear least squares (NLLS), 555 maximum likelihood estimator (MLE), 555-556 rapid lifetime determination method (RLD), 555

lifetime discrimination methods, 552 four-lifetime/one-lane, 556-557 two-lifetime/two-lane, 557-559

lifetime module (LIMO) system, 349 lightning rod effect, 91-92 limited photon budget, 480 limits of detection (LOD), 425 linear photodiode array, 489 line-scanning approach, 488 line-scanning microscopy, 515 lipopeptaibols, 48 liposome micellization, 60

598 INDEX

liver parenchyma, 131 London dispersion forces, 17-19

differences with FRET, 19 longitudinal spatial resolution, 117 Ludox AM-30 dispersion, 297 lung cancer screening, 163 lung imaging fluorescence endoscopy (LIFE), 166, 168, 172 lysyl pyridinoline, 483

M macroscopic electric oscillator, 8 MADS box genes, 350 MADS box proteins, 351 magainin, 66 magnetic resonance imaging (MRI), 125, 186 magnetism and electricity, relation between, 3-4 maleimides, 145 map-based sequencing. See Ordered Shotgun Sequencing (OSS), 528 maximum entropy method (MEM), 244-245,249 maximum fluorescence ratio, determination of, 368 Maxwell's field theory, 7 Maxwellian electromagnetic field of an oscillating Hertzian dipole, 25 MCP image intensifiers, 489 MCP-PMT,271 MDFI (multidimensional fluorescence

imaging), 479 instrumentation, 480

MegaBACE 1000 DNA sequencer, 564^ 565 melanin, 360-361, 370, 374 melanoma diagnosis, 359

techniques used for, 360 MEM analysis, 251,255 MEM distribution, 251 MEM program MemExp, 244 membranepermeant chelator TPEN, 400 metal chelate microspheres, approaches to preparing, 207 metal ion buffers, use of, 410 metallothionein (MT), See zinc ligands meta-tetra (hydroxyphenyl) chlorin (mTHPC), 368 MIANS [2-(4'-maleimidylanilino)naph-thalene-6-sulfonate], 458 micro interrogation probe, 132

microchannel plate (MCP) image intensifier, 489 micro-fluorescence photometry, 116-117 micro-interrogation probe, 129 microinvasive carcinoma, 163 microplate based methods, 208 microsecond dynamic motions, 241 microsphere-bound address tags, use of, 210 micro-total analysis systems (fiTAS), 535 Mie theory for spheres, 85, 102, 370 mitochondria, 119 mitochondrial localization sequence (MLS) peptide, 189 mitochondrial respiratory chain, 121 model of rigid sphere, 284-285 models to represent complex

fluorescence decay data Laguerre functions, 492 power law, 492

molecular interactions and functions analysis enzyme substrate interactions, 202-205 ligand-receptor interactions, 205-207 protein immobilization, 207-208

molecular models, FRET-based construction of, 458 molecular target hypothesis, 143 monobromobimanes (mBrB), 145-146 Monte Carlo (MC) computational methods, 370 morphogenesis cell death, 119 multichannel analyzer (MCA), 553 multichannel plate (MCP) devices, 268, 486 multichannel platephotomultiplier tube (MCP-PMT),271 multi-lane sequencing instruments, 545 multi-photon absorption, 266 multiphoton excitation (MPE), 266, 488 multiphoton laser scanning microscopes, 261 multiphoton microscopy, 485 multi-photon multi-foci microscope, 516 multiple multipole methods (MMP), 87 multiple parameter imaging, 186 multiwell plate FLIM system, 495 murine bone marrow stromal cell line, 465 myofilaments, 449-450, 458,461

INDEX 599

N N,N-Bis(2-pyridylmethyl)ethylenediamine, 405 A^-[(3-triethoxysilyl)propyl]gluconamide (GLTES), 278 A^-[(3-triethoxysilyl)propyl]malton-amide (MLTES), 278 N2 fluorescence signal, 438 NAD(H), intracellular oscillating behaviors of, 109 NAD(P)H fluorescence of colonic tissue, 365 NAD/NADH, calculation of, 111 NADH, autofluorescence of, 113 NADHmolecule, 107, 363

determination of intracellular NADH level, 110-118 regulation of intracellular NADH level, 118-123 role of, in energy metabolism, 107 significance of determining intracellular, 108-109 structure of, 108/ uses of, 108

NADH level in cytoplasm and mitochondria, relationship between, 116 NADH level, determination of

intracellular approaches used in, 110, 111 enzymatic assays, 111-112 fluorometric methods, 112-116 laser scanning confocal microphotographics, 117 micro-fluorescence photometry, 116-117 two-photon excitation micrographics, 117-1118

NADH level, regulation of intracellular effect of vitamins on, 118 time course in yeast apoptosis, 119-123

NADPH (nicotinamide adenine dinucleotide phosphate), 483 naphthalocyanine (NPcs), 544 narrow band imaging (NBI), 173 natural killer (NK), analysis of, 465 NBD energy-transfer acceptor, 61 NCp7 of HIV-1 virus, role of, 323 NdiYAG laser, 425, 430, 545 near field scanning optical microscopy (NSOM), 87, 232 near-IR fluorescence detection, 542

near-IR fluorophores, 542 near-IR-based DNA analysis, 544 necrosis and reversible injury, processes of, 120, 122/ necrotic tumours, 483 neurotoxin, 399 Newport Green DCF, 407 Newton's theories, 7 nickel-nitrilotriacetic acid or Ni-NTA, 184 nicotinamide adenine dinucleotide

(NAD), 107 role of, 107 structure of, 108/

nitrosylation, 141 NK-cell mediated killing activity, 466 NK-lyzed tumor cells, 466 N-methyl-D-aspartate (NMDA) receptor, 399 NMR spectroscopy, 278 nonlinear filamentation process, 434 non-radiative energy transfer, 3 non-small-cell lung cancer (NSCLC), 388 normal mucosa, 166 novel therapeutic agents, screening of, 466 iV-terminal cysteine residues, fluorescent tagging of, 150 nuclear instrumentation modules (NIMs), 268 nuclear localization sequence (NLS) peptide, 189 nuclear pore membrane protein POM121, 465 nucleotide bases of DNA, four different

types, adenine (A), 526 cytosine (C), 526 guanine (G), 526 thymine (T), 526

O Oct-Aib-Gly-Aal-Aib-Gly-Gly-Leu-Aib-

Gly-Dab(Boc)-Leu-OMe (A3), 48-50 azulene in, 49 CD spectrum of, 49 FRET experiments of, 59 Stem-Volmer plots, 53, 55/

Oct-Aib-Gly-Leu-Aib-Gly-Gly-Leu-Aib-Gly-Dab(Fmoc)-Leu-OMe (FIO), 48-50, 54

600 INDEX

aggregation data in membrane, 58/ aggregation data in water of, 58/ concentration of, 6 1 / 64 Fmoc, 49 fluorescence decay curves of, 56/ FRET experiments of, 59 Stem-Volmer plots, 53, 55/ translocation experiments of, 63/

Oct-Aib-Gly-Leu-Aib-Gly-Gly-Leu-Aib-Gly-Ile-Leu-OMe (TR-OMe), 48-50 oligonucleotide primer, 536 optical biosensors, application of, 341 optical Bloch equations, 100 optical coherence tomography (OCT), 174 optical fiberbronchoscopes, 169 optical field enhancement

application of, 85, 87 definition of, 88, 97-98

optical manipulation of particles, application of, 98 optical Rabi oscillation, 29 optically encoded microparticles, application of, 199-201 ordered shotgun sequencing (OSS), 527-530 Oregon Green488, 344 organelles, mixture of

chlorophyll, 110 cytoplasm, 110 dictyosome, 110 endoplasmic reticulum, 110 lysosome, 110 mitochondrial, 110 nucleolus, 110

overlap integral, 39 oxidative stress, 109, 120, 121

p53 (tumor suppressor) protein, 465 Parseval's equation, 75 partial ablation, definition of, 132 Particle growth

DGS sols, 290-295 SS sols, 295-296

Pathfinder™ instrument, 390-391 PEO, 300 peptide activity. See carboxyfluorescein release, rate of peptide concentration, effect of, 53-55, 61 peptide distribution analysis, 64

peptide orientation inside membrane, 65-67 peptide translocation, 61-63 peptide-induced membrane permeability, 60 peptides as zinc indicators, 410-411 perfluorocarbons (PFCs), 422 perthiyl radical formation, 143 PET, 404^05 petunia MADS box transcription factor interactions, 351 PF techniques, 430 PF-luminescence approach, 428 phosphodiester bonds, 526 phospholipid acyl chains, 64 photodynamic therapy, 167 photofragmentation (PF), 423, 428-^30 Photofrin®), 165 photoinduced electron transfer (PET) zinc indicator, 402 photomultiplier tube (PMT), 269, 539, 550-551 photon counting histogram (PCH), 240

analysis, 230 photon counting method, 553 photon counting photomultipliers, 486 photon time-binning approach, 487 photonic crystalline fibre (PCF), 509 phthalocyanines (Pc), 544 PicoGreen (PG) in cells and in plasmid DNA, fluorescence intensity decay of, 326/ PicoGreen (PG), 324, See also DNA dynamics, fluorescence probes for pixel-based fitting software, 271 PKA-activity-decreasing mutants, 109 plasma frequency, 72 plasma membrane redox systems, 118 plasmon field enhancement, 89 plasmons, 73 platinum complexes, 465 platinum-mediated DNA crosslinking, 465 plumbagin (Q1), microscopic analysis of toxic effects of, 467/ PME method, advantages of using, 551 polynucleotide strands of a DNA molecule, 527 polyvinyl-alcohol layer, 78, 81-83 porphyrins, 483 positron emission tomography (PET), 186 PPh3,roleof, 151,153

INDEX 601

primer-bound capture tags, use of, 210 process of (FRET), 492-493 programmed cell death (apoptosis), 119, 143,164,324,345,399 prokaryotic cell based assays, use of, 466 prokaryotic cells, 400

screening of, 470-472 prolate spheroidal wave functions, 96 propydium iodide. See DNA dynamics, fluorescence probes for protein dynamics, probes of, 247-253 protein homocysteinylation, 143 protein syntaxin 1, 243 proton-transfer-uncoupling agent, effect of, 112 protoporphyrins fluorescence, 483 pulse chirping, 432

radiofrequency energy, thermal conversion of, 125 radioisotopic assays, 466 Raman spectroscopy, 72, 85 rate processes, theory of, 126 ratiometric techniques, 478 Rayleigh scattering, 369 Raynaud's syndrome, 142 reactive oxygen species (ROS) levels, 109, 141 refractive index for silver layer, 79 refractive index measurements, 278 resonance, importance of, 12 RFA zone of ablation, 129 rhodamine 123, 78 rhodamine 6G, 281 rigid endoscopes, 506 Ritz combination principle, 10

QD probe delivery into cell, methods of carrier-mediated transfer, 187 electroporation, 187, 189 endocytosis, 187 microinjection, 187, 189 toxin delivery, 187

QD probe delivery of, into cells, 187 multicolor, advantages of, 186 structure of a multifunctional, 183/

quantitative fluorescence hybridization, 391-393 quantum dots (QDs), 181

advantages of, 181 bioconjugation, 183 development of high quality, 182 excited state lifetimes of, 185 molar extinction coefficients of, 185 optical properties of, 185-187 probe development of, 182 Stokes shifts, application of, 185

quantum mechanics (QM), 3, 9-10

R R6G mobility in porous monolithic silica, 289-290 radiative decay engineering (RDE), 72, 77 radio communications, basis of, 8 radiofrequency ablation (RFA), 125, 127, 132-133

S-adenosyl homocysteine (SAH), 141 S-adenosylmethionine (SAM), 141 SAFE 1000 System, 166, 169, 172 SAFE-3000, 166, 173 salmon pink fluorescence, 120 Sanger method for DNA sequencing, 529-530 Sanger sequencing reaction protocol, 530 scanning FLIM microscopy, 487-488 scattering force, 99 Schrodinger equation, 28 second harmonic generation, 278 second order perturbation theory, 18 sensitized fluorescence, See energy transfer, illustration of sequence tagged sites (STS), 528 Shai-Matsuzaki-Huang model, 47 shotgun sequencing strategies for whole genome processing, 527-529 signal to noise ratio (SNR), 424,488, 547 silica nanoparticles, 279, 281, 287 single base extension (SBE) of oligonucleotide primer approach, 209 single extraction method, 112 single molecule Raman spectroscopy (SERS), 98 single molecule resonance Raman spectroscopy (SERRS), 98 single nucleotide polymorphisms (SNPs), analysis of, 209 single photon avalanche diode (SPAD), 542, 554, 559 single ratio techniques, 368

602 INDEX

single stranded Ml3 bacteriophage, 528 single-molecule analysis, fluorescence

properties of fluorescence lifetime, 246 optical anisotropy, 246 spFRET efficiency, 246

single-molecule detection limit, breakthrough in FCS sensitivity, 239 single-molecule resonance energy transfer, 239 single-pair FRET (spFRET), 242 sinusoidally modulated diode lasers (LED), 485 skin melanoma, 362 small-angle X-ray scattering (SAXS), 278, 297 small-cell lung cancer (SCLC), 388 Snell'slaw, 81 SNR (signal-to-noise ratio), 267 sol-gel process, 277

application of, 277 Sommerfeld identity, 75 SOS-GFP biosensor, 465 Sox peptides, 411 SP-300i,516 spatial light modulator (SLM) performing Hadamard transforms, 515 spectral changes, phases of, 131 spectral collisional broadening, theory of, 23 spectrally-resolved (hyperspectral) FLIM, 514 Spectralon®, 372 spectroscopic cross-sections of atomic

vapors, pre-quantum theories of, 12-14, 16 broadening of spectroscopic lines, 13 effective cross-section of atomic coUisions, 16 intermolecular dipole-dipole interactions, 13 Na-vapors, deviation of energy levels of, 13-14 spectroscopic oscillator strengths, 14

spectroscopic ruler, 479 spFRET efficiency, 246-247 SS sols, particle growth in, 295-296 static electricity and magnetism, relation between, 3 stimulated emission depletion (STED), 232

Stober synthesis process, 294 stoichiometry TnTmA7, 455 streak cameras, 485 stretched exponential decay function (StrEF), 492, 495 STS mapping, 528 submucosa, 166 sulforhodamine 101, 78 supercontinuum generation (SCG), 432 surface enhanced Raman scattering (SERS), 71 surface plasmon coupled emission (SPCE),71,74,78,81 surface plasmon resonance (SPR), 71, 74,

81,205 field enhancement, 99, 101-103 in silver spheroid, effect of, 91 optical trapping of fluorescent molecules by, 98 plasma oscillations, 72 surface plasmons, 73

surface plasmon resonance optical field enhancement at prolate spheroids, 85-93 application of spheroidal vector wave functions, 88-93 advantage of using spheroidal vector wave functions in, 94 solving of Maxwell's equations, 94-98

surface plasmon resonance theory at planar structures basic theory, 75-78 simulations, 78-84

synthetic fluorescence, 2

TAT (transactivator of transcription) peptide, 412-413 /-butyloxycarbonyl, 48 TCSPC, 268-270,486-488, 553

imaging, 270 instrumentation, 486 principle of, 486

teramobile laser system, 432, 434 terminal apoptotic cell, 120 ternary complex cTnC«cTnI*cTnT, 449 Tet-On system, 465 tetrachloroethylene (TCE), 426 tetramethyl-rhodamine, 539 Texas Red, 539 TFLZn, N-(6-methoxy-8-quinolyl)-p-carboxybenzoylsulfonamide, 402