Atomic Force Microscopy and Force Spectroscopy of...

Atomic Force Microscopy and Force Spectroscopy ofBiomembranes and Applications to Nanotechnology

Kislon Voıtchovsky

Condensed Matter Physics

Clarendon Laboratory, University of Oxford

Parks Road, Oxford OX1 3PU, United Kingdom

DPhil Thesis

Trinity Term 2006

Atomic Force Microscopy and Force Spectroscopy ofBiomembranes and Applications to Nanotechnology

Kislon VoıtchovskyCondensed Matter Physics, Clarendon Laboratory, University of Oxford, Oxford OX1 3PUDPhil Thesis, Trinity Term 2006

Abstract

Biological membranes form the interface between the cell and its environment. Taskssuch as energy transduction and chemical communication between the inside and theoutside of the cell are achieved by membrane proteins. However, molecular level un-derstanding of membrane proteins function is largely hindered by the lack of structuralinformation available, especially in biologically relevant environments.Bacteriorhodopsin (bR) naturally found in halophilic Purple Membranes (PM) is a light-driven proton pump, and by far the preferred model system for ion transporters.In this thesis, I have used atomic force microscopy (AFM) to investigate the relationbetween bR local mechanical properties and function, demonstrating that PM exhibitsdifferential leaflet stiffness which can be related to interactions with ions at the molecularlevel and to PM halophilic adaptation.Single molecule force spectroscopy revealed that bR functions on two levels: in bR cyto-plasmic region, highly fluctuating ion-mediated interactions with the solvent favour largestructural rearrangements, while the extracellular region is rigidified by steric, specificinteractions which involve tryptophan resides, and allow bR to control ion binding atPM extracellular surface. The asymmetric force field induced by bR to trap, release, andlaterally transfer protons, was characterized with sub-nanometer resolution at both PMsurfaces. Based on this result, it was possible to develop a new imaging approach whichprovided unprecedented resolution of PM, also under conditions never probed before.This method was further applied to successfully image PM at sub-molecular resolutionwith high-speed AFM, and at scan rates allowing, for the first time, real-time observa-tion of bR photo-induced conformational changes in native-like conditions. The largestructural rearrangement observed brought some new insight in the controversial datagathered by X-ray and electron diffraction studies.Finally, several steps towards the realization of a bR-based single molecule photo-detectorwere undertaken, including the development of a low current measurement setup suc-cessfully used to sense protein attachment to nitrogen doped carbon nanotubes.

3

Contents

Abstract 3

1 Introduction 91.1 Biological Membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

1.1.1 General considerations . . . . . . . . . . . . . . . . . . . . . . . . 91.1.2 Membrane Lipids . . . . . . . . . . . . . . . . . . . . . . . . . . . 91.1.3 Membrane Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . 10

1.2 A special case: Purple Membrane from Halobacterium salinarium . . . . 131.2.1 Bacteriorhodopsin . . . . . . . . . . . . . . . . . . . . . . . . . . . 151.2.2 The Purple Membrane lipids . . . . . . . . . . . . . . . . . . . . . 181.2.3 Halophilism and electrostatic interactions within Purple Membrane 211.2.4 Purple Membrane as a model system . . . . . . . . . . . . . . . . 24

1.3 Atomic Force Microscopy: a powerful tool for biology . . . . . . . . . . . 261.3.1 Atomic Force Microscopy vs other techniques . . . . . . . . . . . 27

1.4 Goals and Motivations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321.4.1 Relating specific interactions with membrane structure and activity 331.4.2 AFM to study membrane proteins and vice versa . . . . . . . . . 341.4.3 Studying membrane proteins for medical and technological appli-

cations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351.4.4 Overview of the thesis . . . . . . . . . . . . . . . . . . . . . . . . 36

1.5 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

2 Atomic Force Microscopy of Membrane Proteins: A short Review 572.1 Bacterial and Mammalian Rhodopsins . . . . . . . . . . . . . . . . . . . 57

2.1.1 Bacteriorhodopsin . . . . . . . . . . . . . . . . . . . . . . . . . . . 572.1.2 Halorhodopsin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 592.1.3 Bovine Rhodopsin . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

2.2 Membrane channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 602.2.1 Ion transporters and channels . . . . . . . . . . . . . . . . . . . . 602.2.2 Porins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 612.2.3 Gap junctions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

2.3 Photosynthesis related systems . . . . . . . . . . . . . . . . . . . . . . . 632.3.1 The Light Harvesting complexes 1 and 2 . . . . . . . . . . . . . . 63

2.4 Molecular motors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 632.4.1 F1F0 bacterial and Spinach ATP-synthase . . . . . . . . . . . . . 632.4.2 φ-29 rotary motor . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

2.5 Toxins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 652.5.1 Membrane associating toxins . . . . . . . . . . . . . . . . . . . . . 652.5.2 Other toxins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

2.6 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

5

3 Differential Stiffness and lipid mobility in both leaflets of Purple Mem-brane 753.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

3.1.1 The natural asymmetry between Purple Membrane leaflets . . . . 753.1.2 Chapter overview . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

3.2 Distinction of Purple Membrane extracellular and cytoplasmic sides usingAC-AFM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 783.2.1 High resolution imaging and Phase information . . . . . . . . . . 783.2.2 Membrane protrusions on the cytoplasmic side and pH effect . . . 79

3.3 Observation of the asymmetric effect of salt on Purple Membrane . . . . 813.3.1 Imaging with no salt, in liquid and in air . . . . . . . . . . . . . . 813.3.2 Side-specific Force Spectroscopy . . . . . . . . . . . . . . . . . . . 83

3.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 853.4.1 Lipid mobility and density . . . . . . . . . . . . . . . . . . . . . . 853.4.2 Estimation of Purple Membrane differential stiffness by force spec-

troscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 883.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 943.6 Material and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 963.7 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

4 Probing purple membranes specific molecular interactions and ioniceffects with AFM 1034.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

4.1.1 The formation of an AFM image . . . . . . . . . . . . . . . . . . 1034.1.2 Chapter overview . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

4.2 General theory of molecular interactions: the AFM insight . . . . . . . . 1054.2.1 The DLVO approximation . . . . . . . . . . . . . . . . . . . . . . 1054.2.2 Ions specific effects and hydration forces . . . . . . . . . . . . . . 1084.2.3 Failure of continuum theories at physiological ionic concentrations 1094.2.4 AFM as an alternative approach to study the molecular interac-

tions governing biology . . . . . . . . . . . . . . . . . . . . . . . . 1104.3 Hydration force microscopy reveals the molecular forces that trap and

release protons in bacteriorhodopsin . . . . . . . . . . . . . . . . . . . . . 1104.3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1104.3.2 The Hydration Force Microscopy approach . . . . . . . . . . . . . 1114.3.3 Probing the local force field at the surface of Purple Membrane . 1144.3.4 The effect of salt imaged at sub-molecular resolution . . . . . . . 1184.3.5 Biological significance . . . . . . . . . . . . . . . . . . . . . . . . . 1204.3.6 Material and Methods . . . . . . . . . . . . . . . . . . . . . . . . 121

4.4 The intra-molecular forces underlying bacteriorhodopsin structure andmechanics: Tuning electrostatic and steric interactions . . . . . . . . . . 1234.4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123

6

4.4.2 Side specific unfolding of bacteriorhodopsin at different ionic con-centration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

4.4.3 Fitting the unfolding curves with the Worm-Like Chain model . . 1264.4.4 Interpretation of the unfolding patterns and discussion . . . . . . 1304.4.5 Towards a functional picture of bR interactions . . . . . . . . . . 1394.4.6 Error estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . 1414.4.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1414.4.8 Material and Methods . . . . . . . . . . . . . . . . . . . . . . . . 142

4.5 Effects of different types of cations on Purple Membrane . . . . . . . . . 1434.5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1434.5.2 High resolution imaging PM in buffers containing different ions . 1434.5.3 Comparison and Discussion . . . . . . . . . . . . . . . . . . . . . 1474.5.4 Material and Methods . . . . . . . . . . . . . . . . . . . . . . . . 150

4.6 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150

5 Real-time observation of membrane proteins function using high-speedAFM 1615.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161

5.1.1 Time resolution in biology . . . . . . . . . . . . . . . . . . . . . . 1615.1.2 High-Speed AFM (hsAFM) . . . . . . . . . . . . . . . . . . . . . 1635.1.3 Chapter overview . . . . . . . . . . . . . . . . . . . . . . . . . . . 164

5.2 High-resolution at high speed . . . . . . . . . . . . . . . . . . . . . . . . 1655.2.1 Practical considerations . . . . . . . . . . . . . . . . . . . . . . . 1655.2.2 Theoretical considerations . . . . . . . . . . . . . . . . . . . . . . 1665.2.3 Achieving molecular resolution of PM at high-speed . . . . . . . . 1705.2.4 Imaging with dynamical pH changes . . . . . . . . . . . . . . . . 173

5.3 Real-time observation of bacteriorhodopsin photocycle: large structuralrearrangements and functional implications . . . . . . . . . . . . . . . . . 1745.3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1745.3.2 Direct imaging of bacteriorhodopsin photo-activity with hsAFM . 1785.3.3 Material and Methods . . . . . . . . . . . . . . . . . . . . . . . . 182

5.4 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183

6 Towards bio-nanotechnological applications 1916.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191

6.1.1 A single molecule photo-detector . . . . . . . . . . . . . . . . . . 1916.1.2 Chapter Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . 193

6.2 Preliminary studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1946.2.1 Monomerization of bR . . . . . . . . . . . . . . . . . . . . . . . . 1946.2.2 Electrical conductivity properties of Purple Membrane . . . . . . 1946.2.3 Nano-gap electrodes for bio-molecules . . . . . . . . . . . . . . . . 1966.2.4 Trapping bacteriorhodopsin between nano-electrodes . . . . . . . 198

6.3 Measurement setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202

7

6.3.1 Implementation of a low current measurement system . . . . . . . 2026.3.2 Current vs Time and Voltage curves; the effect of light . . . . . . 2036.3.3 Bio-sensing metalloproteins with CNx carbon nanotubes . . . . . 207

6.4 Discussion and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . 2086.5 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210

7 General discussion and conclusions 215

Acknowledgments 217

8 Appendix 2198.1 Energy dissipation with small amplitude AC-AFM . . . . . . . . . . . . . 219

8.1.1 From UHV to liquid: adapting the models . . . . . . . . . . . . . 2198.1.2 Estimation of the damping induced by lateral movement (scan-

ning) with AC-AFM . . . . . . . . . . . . . . . . . . . . . . . . . 2218.1.3 Maximizing energy dissipation at small amplitudes . . . . . . . . 222

8.2 Estimation of the number of bacteriorhodopsin molecule excited per sec-ond given a characterized incident light beam . . . . . . . . . . . . . . . 225

8.3 Analysis of protein unfolding curves . . . . . . . . . . . . . . . . . . . . . 2268.4 Monomerization of Purple Membrane . . . . . . . . . . . . . . . . . . . . 2278.5 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230

9 Publications and Conference Meetings 2319.1 List of publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2319.2 Conferences and Meetings . . . . . . . . . . . . . . . . . . . . . . . . . . 232

8

Introduction

1 Introduction

1.1 Biological Membranes

1.1.1 General considerations

Biological membranes are present in all living organism. The primordial role of these bar-

riers is to physically separate the inside from the outside of cells. They act as protective

walls preventing unwanted molecules to diffuse inside the cell, but maintain a selective

permeability through membrane proteins in order to reach the cell requirements (Berg

et al. 2002; Houslay & Stanley 1982). Cell membranes are self-assembled sheet-like

structures composed of lipids and proteins and are typically 6− 10 nm thick.

1.1.2 Membrane Lipids

In aqueous solution, membrane lipids spontaneously assemble into a bilayer. Most of

the membrane lipids are composed of two hydrophobic fatty acid tails connected to a

charged or polar head-group. The length and the degree of unsaturation of the tail alkyl

chains influences the lipid-lipid and protein-lipid interactions within the membrane, and

therefore the membrane stability (Berg et al. 2002; Houslay & Stanley 1982). There

are three main categories of membrane lipids: phospholipids, glycolipids, and cholesterol.

Phospholipids represent the major class of lipids with two fatty acid tails linked to a

phosphate head-group via a glycerol. An alcohol group is linked to the phosphate and

determines the phospholipid identity. Glycolipids head groups contain a sugar residue

and orient always with the sugar residue on the extracellular side of the membrane.

Cholesterol is a steroid composed of hydrocarbon rings only present in eukaryotic cells.

The amphipathic nature of membrane lipids is responsible for the membrane bilayer self-

assembly which minimizes the system energy by isolating the lipids apolar tails from the

9

Introduction

aqueous surrounding. A polarity profile of a cell membrane shows important variation

between the highly hydrophobic core and the polar/charged lipids head-groups (Fig. 1.1).

Most of the membrane proteins structure and composition reflects this profile, exhibiting

an excess of apolar residues where in contact with the lipid tails. This strategy assures

their anchoring in the membrane, and specific interactions with surrounding lipids can

frequently be found (Palsdottir & Hunte 2004). Since only thermodynamical forces are

responsible for the membrane self-assembly and integrity (no covalent interactions are

present), its structure is highly dynamical and proteins and lipids can diffuse laterally

if no specific interactions are present; a lipid molecule can diffuse laterally up to 2 µm

per second (Jacobson et al. 1995; Berg et al. 2002; Houslay & Stanley 1982). The

hydrophobic core of cell membranes provides an efficient barrier for soluble molecules,

as illustrated in Fig. 1.1.C, since the necessary energy required to cross the apolar

region if far more important than the thermal bath kT (with k Boltzmann’s constant

and T the temperature). It also prevents membrane proteins and lipids to reverse their

orientation, which allows membranes to be asymmetric and to show different properties

on their extracellular and cytoplasmic sides. Such an asymmetry is ubiquitous in cell

membranes, and necessary for differentiating the inside from the outside of the cell where

the conditions are different. Most of the membranes are polarized and maintain different

ionic types and concentrations inside and outside the cell, as required for the cell activity

and survival (De Weer 2000; Berg et al. 2002; Houslay & Stanley 1982).

1.1.3 Membrane Proteins

Membrane proteins regulate the communication of the cell with the outside world. They

are responsible for the membrane selective permeability which is achieved by active and

passive transport of molecules and ions through the membrane, enzymatic activity and

cell tags (Berg et al. 2002; Houslay & Stanley 1982; Byrne & Iwata 2002). The role

10

Introduction

10-12 10-10 10-8 10-6 10-4 10-2

Na+

K+

Cl- Glucose

Tryptophan

UeraGlycerol

Indole

H2O

membrane permeability [cm/s]

3020100-10-20-30 3020100-10-20-30

H20H20carbonyl

glycerolphosphatecholine

interface interfacehydrocarbon core

pro

ba

bili

ty

cha

rge

den

sity

distance from HC center [A]distance from HC center [A]

A B C

Figure 1.1: (A) Profile of a DOPC bilayer structure obtained by X-ray and neutron diffraction(White & Wimley 1998; White et al. 2001). The spatial distribution of the different groupsis given respectively to the hydrocarbon core hc location. (B) Polarity profile of the DOPCbilayer shown in (A), calculated from absolute values of atomic partial charges (White &Wimley 1998). (C) Permeability of on a phospholipid bilayer to different molecules (adaptedfrom (Berg et al. 2002)).

played by membrane proteins is essential for the survival and development of the cell,

and account for 20% to 30% of the genome of any living organism (Wallin & Von-Heijne

1998; Krogh et al. 2001).

There are two types of membrane proteins: peripheral and integral. Peripheral proteins

are anchored to the lipid bilayer but remain at the surface and solutions of high ionic

strength (e.g. 1M salt) are generally sufficient to solubilize them. Integral membrane

proteins, in contrast, are embedded in the bilayer that they can span (e.g. channels

and pumps) and their solubilization can only be achieved through the use of a detergent

(Berg et al. 2002).

Structure determination and general properties

The inherent non-solubility of membrane proteins makes their three-dimensional crys-

tallization difficult, and therefore hinders their structure determination through crystal-

lographic techniques. Recent progress has been obtained in membrane protein crystal-

11

Introduction

lization using a technique based on the properties of certain lipids (lipidic cubic phase

crystallization (Landau & Rosenbusch 1996)), but the number of resolved membrane

proteins structures still represents less than 5% from all the proteins structures available

(www.pdb.org 2006; White 2004). Important efforts have been invested in computational

research, with the aim of predicting membrane protein’s structures from their amino acid

sequence1. (Bowie et al. 1991; Krogh et al. 2001; Becker et al. 2004; Zhang & Skolnick

2005; Nayeem et al. 2006; Eisenberg 2003). Most of the membrane proteins high reso-

lution structures available at the present time are integral transmembrane proteins (107

of the 114 unique structures available in September 2006 (White 2006)) and most of the

computational and experimental studies have consequently concentrated their attention

on integral membrane proteins. Some common trends about the way they interact with

surrounding lipids can be observed:

1. The residues accessible to lipid tails tend to be more hydrophobic than the residues

buried inside the protein (Eyre et al. 2004; Ulmschneider et al. 2004).

2. Charged or polar residues can be exposed to the lipids, but they will interact either

with the lipids head-groups or with other polar or charged residues of the protein.

3. Polar aromatic residues (Trp, Tyr) can also interact with lipids, but tend to be

located in the polar-apolar interfacial region, near the lipids carbonyls (Killian &

von Heijne 2000; Palsdottir & Hunte 2004), and act as anchor for the protein in

the bilayer (de Planque et al. 2003).

The available high-resolution structures of membrane proteins can be classified in es-

sentially five major families relatively to their type, role and structure (Byrne & Iwata

1 owing the fact that there are only two known secondary structures for membrane proteins: α-helicesand β-sheets. β-sheets are typically found in outer membrane of Grahm-negative bacteria, mitochondriaand chloroplasts, forming pores so-called β-barrels whereas trans-membrane α-helix based proteins arefar more common and involved in a broader range of functionalities and complexities.

12

Introduction

2002): The outer membrane proteins, the photosynthetic membrane proteins, the mem-

brane transporters and channels, the respiratory enzymes, and the receptors and related

proteins.

The family of receptors and related proteins includes notably the class of G-protein-

coupled receptors (GPCRs). The GPCRs are principally found in eucaryotic membranes.

They are seven transmembrane α-helices proteins of particular interest for pharmaceuti-

cal purposes since their malfunction is responsible for numerous diseases (Brunton et al.

2006; Drews 2000; Flower 1999; Becker et al. 2004).

1.2 A special case: Purple Membrane from Halobacterium sali-

narium

The so-called Purple Membranes (PM) are part of the membrane of the archaea Halobac-

terium salinarium, an extremophile naturally found in salt-saturated water (Oesterhelt

& Stoeckenius 1971; Stoeckenius et al. 1979) (Fig. 1.2, Table 1.1). This particular

biological membrane contains only one protein, bacteriorhodopsin (bR) and about seven

different unusual membrane lipids (Oesterhelt & Stoeckenius 1971; Kates et al. 1982;

Corcelli et al. 2002) assembled in a hexagonal lattice of bR trimers (p3 crystallographic

point group, 6.2 nm lattice constant (Henderson 1977; Cartailler & Luecke 2003)) (Fig.

1.2.C). The protein-lipid ratio has recently been determined to be 10 lipids per bR

(Corcelli et al. 2002). The proteins represent 75% of PM mass and cover 50% of the

membrane surface (Henderson 1977; Cartailler & Luecke 2003). bR acts as a light-driven

vectorial proton pump which converts photonic energy into a proton gradient across the

membrane (Lanyi 1995). The induced electrochemical gradient is subsequently used by

ATP-synthase as an alternative driving force to produce ATP in mediums particulary

poor in oxygen (Falk et al. 1980; Teissie et al. 1990). The absorption spectrum of bR

13

Introduction

A B C

Figure 1.2: (A) Salt lake in Moab, Utah. The purple color comes from PM of halobacteria (B)Scheme of the Halobacterium salinarium with bR embedded in its PM patches. (C) Schemeof the hexagonal lattice of bR trimers in PM, computed with MOLMOL, 1m0l bR quaternarystructure. bR molecules are arranges in trimers (purple) and the X-ray resolved lipids surroundthem (grey). The lipids filling the inter-trimer gaps are more mobile (not shown).

in its ground state (maximum absorbtion at 565 nm, Fig. 1.4.B) is responsible for the

purple colour of PM (Fig. 1.2.A) (Oesterhelt & Stoeckenius 1971; Birge 1990b).

Element Agar peptone growing medium H. salinarium intracell. medium

Na 4.281 1.27± 0.11

P − 0.19± 0.01

S 0.081 0.19± 0.02

Cl 4.307 4.59± 0.14

K 0.036 3.97± 0.13

Mg 0.081 0.036± 0.004

Ca − 0.025± 0.002

Table 1.1 Ionic concentrations (in mol/liter) of the main elements present in H. sali-narium intracellular medium. The first column displays the concentrations added tothe standard Agar peptone medium in which H. salinarium was grown. Adapted from(Engel & Catchpole 2005).

14

Introduction

1.2.1 Bacteriorhodopsin

bR is an integral 26KDa membrane protein composed 248 amino acids arranged in 7

transmembrane α-helices (Khorana et al. 1979; Henderson et al. 1990; Grigorieff et al.

1996) that surround a central cavity which encloses a chromophore, the retinal, linked

to the protein backbone (Lys216 residue, helix G) via a Schiff base (Fig. 1.3) (Pebay-

Peyroula et al. 1997; Mitsuoka et al. 1999; Luecke et al. 1999a). When the protein

is in its fundamental/ground state (called bR state), the retinal chromophore adopts

an all-trans configuration (Henderson et al. 1990; Birge 1990b; Lanyi 1995). Upon

light absorption by the protein, the retinal can isomerize around its C13 = C14 double

bond, adopting a 13-cis configuration (Albeck et al. 1986; Birge 1990b). This new chro-

mophore conformation triggers a series of structural rearrangements in bR (detailed in

the next paragraph), leading to the pumping of a proton from the inside to the outside of

the archea. In PM, bR molecules lateral and rotational movements are prevented by the

Asp96

Asp85

Trp182

Trp86

Trp189

Glu194-204

Arg82

Wat402

Retinal

BA

Figure 1.3: (A) bR photo-induced pumping mechanism step by step, from (Luecke et al.1999b). (B) The retinal binding pocket with some of the most relevant residues for the protonrelease and uptake (see text). The key water molecules (red) are also presented (made withMOLMOL on the pdb file 1C3W (Luecke et al. 1999a))

crystalline assembly (Ahl & Cone 1984; Korenstein & Hess 1978; Sherman et al. 1976;

15

Introduction

Jain & Wagner 1980), and unlike most of the metabolically active membranes which

show an important fluidity (Jacobson et al. 1995; Berg et al. 2002), PM is rather rigid

(Korenstein & Hess 1978) with a viscosity typically 103−104 higher than that of halobac-

terial membrane lipids (Sherman et al. 1976). Within one trimer, the proximity between

the bR molecules allows direct protein-protein interactions through the formation of salt

bridges, hydrogen bonds and van der Waals interactions between the α-helices of the

different proteins involved (Isenbarger & Krebs 1999). These interactions are however

not sufficient to explain the important trimer stability (Korenstein & Hess 1978; Haltia

& Freire 1995; Brouillette et al. 1989) in which specific protein-lipid interactions are

involved (Sato1999, Weik1998). The exceptional cohesion within a single trimer (Haltia

& Freire 1995; Korenstein & Hess 1978) is believed to be related to the protein activity

through cooperativity (Cassim 1992; Varo et al. 1996; Tokaji 1998).

bR photocycle and the proton pumping mechanism

Once activated by a photon, bR undergoes a serie of conformational rearrangements

triggered by the isomerization of the retinal chromophore around its C13 = C14 double

bond (Albeck et al. 1986; Birge 1990b). These changes modify temporarily the position

of charged residues within the protein (Luecke et al. 1999b; Der & Keszthelyi 2001),

allowing directional pumping of one proton over a ∼ 10−20 ms photo-cycle (Lanyi 1995;

Luecke et al. 1999b; Neutze et al. 2002). The isomerization induces modifications of

the retinal local electrical environment that can be observed spectroscopically (Ippen

et al. 1978; Groma et al. 1984; Mathies et al. 1988; Kobayashi et al. 2001) and at

least 6 different states named bR, K, L, M , N and O are distinguishable (Lozier et al.

1975; Kung et al. 1975; Xie et al. 1987; Birge 1990b; Braiman et al. 1991; Heyn et al.

2000), reflecting some of the different conformational states of bR during the photocycle

(Subramaniam et al. 1999) (Fig. 1.4.A). At the present time, it is generally accepted

16

Introduction

that the protein photocycle follows a reversible homogenous succession of steps (Lozier

et al. 1992). This view is however contested by several studies which have proposed

structural heterogeneity bR molecules during the photocycle and suggested cooperative

work of the proteins (Hendler 2005; Varo et al. 1996; Tokaji 1998). It takes the protein

bR

K

L

M2

N

M1

Oλmax = 570 nm

λmax = 590 nmλmax = 640 nm

λmax = 550 nmλmax = 560 nm

λmax = 410 nm

τ ~ 4 ps

τ ~ 1 µs

τ ~ 40 µs

τ ~ 350 µs

τ ~ 5 ms

τ ~ 5 ms

τ ~ 5 ms

H+ (extracellular)

H+ (cytoplasmic)

400 500 600 700

2

4

6

8

[nm]

OD x 10-2

WT bR

A B

Figure 1.4: (A) bR photocycle, as seen with the homogeneous reversible steps model. Themaximum absorption wavelength at each step (blue) and the inter-step time (red) are given.(B) Absorption spectrum of bR in its ground (bR) state. The spectrum was obtained from adiluted sample as used for atomic force microscopy studies (typ. ∼ 25 µg/ml).

about 4− 5 ps to build up the K-intermediate state, following the isomerization of the

retinal (Polland et al. 1986). This state has been attributed to the expulsion of a key

water molecule (Wat402) (Fig. 1.3.B) from its initial position near the retinal (Maeda

et al. 2004; Schobert et al. 2002; Neutze et al. 2002), inducing a reduction of the pKa

difference between the Schiff base and the neighboring Asp85 (Fig. 1.3.B). The mutual

attraction of the two groups triggers a slight bend of the C-helix (Gat & Sheves 1993;

Neutze et al. 2002). The proton transfer from the Schiff base to the Asp85 happens

a few microseconds later, during the L state, and cancels the mutual attraction of the

two groups, thus assuring the vectoriality of the proton pumping process (Royant et al.

2000). The protein then rearranges and major conformational changes are observed. A

17

Introduction

steric clash between the isomerized retinal C13 methyl group with Trp182 triggers an

important bending of the F -helix during the M1-intermediate state (or Ms) (Koch et al.

1991; Kamikubo et al. 1996; Sass et al. 1997; Brown et al. 2002). The M2-intermediate

state (or Mf ), spectrally identical to M12 (Lozier et al. 1992), corresponds to the pro-

ton release on the protein extracellular side via the Arg82, the Glu194 and the Glu204

respectively (Otto et al. 1990; Luecke et al. 1999b; Zscherp et al. 1999; Sanz et al.

2001)(Fig. 1.3.B). The Schiff base then reprotonates from Asp96 (N -intermediate state)

(Schobert et al. 2003; Vonck 2000; Otto et al. 1989), allowing the retinal to re-isomerize

to its all -trans conformation (state O-intermediate) and the protein to reach its ground

state (bR-state) (Zscherp & Heberle 1997; Dioumaev et al. 2001; Neutze et al. 2002).

1.2.2 The Purple Membrane lipids

PM contains many different lipids some of which are not found in any other membrane

(Cartailler & Luecke 2003; Kates et al. 1982; Corcelli et al. 2002) and are necessary for

bR activity (Hendler et al. 2003; Hendler & Dracheva 2001), also in reconstituted mem-

branes (Saito et al. 2003). To date, seven different lipids have been reported (Kates

et al. 1982; Corcelli et al. 2002): three phospholipds, two glycolipids, squalene and

traces of vitamin MK8 (Fig. 1.5). The requirement of a fixed membrane composition

indicates that selective interactions occur between bR and certain lipid molecules and

that these interactions are essential for lattice assembly and bR function (Cartailler &

Luecke 2003; Krebs & Isenbarger 2000; Lee 2003).

The phospholipids

2 The amount of light exciting PM affects the proportion M1/M2 (Varo et al. 1996) and it has beenproposed that these states represent an heterogeneity of excited/non-excited proteins, consistently withbR cooperativity models (Tokaji 1998)

18

Introduction

squalene

Glyclipid sulphate S-TGA-1

phosphatidylglycerophosphate methy ester PGP-Me

archaealcardiolipin BPG

phosphatidylglycerol PG

phosphatidylglycerosulphate PGS

vitamin MK8

PM neutral lipids

PM polar lipids

archaealglycocardiolipin GlyC

Figure 1.5: PM polar and neutral lipids as determined by (Corcelli et al. 2002).

19

Introduction

Phospholipids represent ∼ 40% of the lipids (Cartailler & Luecke 2003; Corcelli et al.

2002); their head-group comports two negative charges which makes them particularly

hydrophilic. The main phospholipid present in PM is phosphatidylglycerophosphate

methylated (PGP-Me) (Kates et al. 1982) with a molar ratio of 2.4:1 to the retinal

(Corcelli et al. 2002). The other archaeal phospholipids are phosphatidylglycerol (PG),

phosphatidylglycerosulphate (PGS), and archaeal cardiolipin (BPG). Fluorescence stud-

ies have demonstrated that phospholipids are located mainly in PM cytoplasmic leaflet

inter-trimer space (Renthal & Cha 1984; Sternberg et al. 1992; Belrhali et al. 1999)

where they seem to interact with bR and each other mainly through non-specific long

range electrostatic forces (Sternberg et al. 1992; Grigorieff et al. 1995). X-ray diffrac-

tion patterns of their head-groups provided large B-factors for most of them, suggesting

an important mobility. One phospholipid could, however, be precisely located between

each trimer bR molecule and is believed to strongly interact with the proteins, acting like

”trimer glue” (Sato et al. 1999). This is supported by the fact that bR-DMPC recon-

stitution in the presence of > 2 M NaCl exhibits the well known hexagonal crystalline

array formation of PM only when PGP or PGS are present (Sternberg et al. 1992; Watts

1995) and that reconstitution in model membranes without those natural lipids showed

conformational changes (Saito et al. 2003).

The glycolipids

The glycolipids are archaeal glycocardiolipin (GlyC) and triglycosyl lipid (S-TGA-1)

and represent 30% of PM lipids (Corcelli et al. 2002). Unlike phospholipids, glycolipids

show clear patterns in PM diffraction experiments (Grigorieff et al. 1996) (smaller B-

factors), suggesting that they are specifically and tightly bound to bR (Weik et al. 1998;

Henderson et al. 1978). Several studies have shown that the adjunction of non-polar

detergent is indeed not sufficient to solubilize glycolipids (Essen et al. 1998; Patzelt et al.

20

Introduction

1997; Weik et al. 1998). Neutron diffraction experiments confirmed the presence of two

S-TGA-1 molecules per bR molecule, both located in the extracellular leaflet with one

in the inter-trimer and one in the intra-trimer space (Weik et al. 1998) and contribute

to the trimer stability (Belrhali et al. 1999; Kates et al. 1982; Luecke et al. 1999a;

Weik et al. 1998). It has been proposed that glycolipids are located exclusively in the

extracellular leaflet of the membrane (Henderson et al. 1978).

The squalene

The squalene seems to be located in a crevice near bR cytoplasmic surface despite the

hydrophobic repulsion with several negatively charged residues. This apparently unfa-

vorable interaction has been explained invoking a better control of the protein shape

through mutual repulsion, enhancing bR activity (Hendler et al. 2003).

1.2.3 Halophilism and electrostatic interactions within Purple Membrane

The particular properties of PM such as the presence of unusual lipids, the important

and asymmetric surface charge of the membrane (Fig. 1.6.A) and the constrained ar-

rangement of bR molecules in a two dimensional crystal are related to the extreme salt

concentration of Halobacterium salinarium natural environment (Oren 1994). PM lipids

show the characteristics of archaea membrane lipids, allowing them to be more stable

and resistant to oxidation. First, the non-polar alkyl chains are linked to the glycerol

backbone via ether (rather than ester) which provide them more resistance to hydrolysis.

Second, the alkyl chains are branched with fully saturated carbons, providing a better

fluidity despite specific interactions, and finally, the stereochemistry of the glycerol is

inverted by comparison with eubacteria and plants membrane lipids (Kates et al. 1982;

Berg et al. 2002; De Rosa et al. 1994). Further to these particularities, PM lipids

have an unusually high negative charge density on their head-groups, especially in the

21

Introduction

case of phospholipids. bR is also globally more negatively charged on its cytoplasmic

half (see §3.1.1 for details). The excess of acidic residues over basic residues showed by

bR is characteristic of extremophile protein (Lanyi 1974; Oren et al. 2002; Dym et al.

1995). This reinforces the polarization of PM, but creates important repulsive electro-

static forces between bR and the surrounding phospholipids. The salt ions present in PM

natural environment are necessary to shield the electrostatic repulsions and are believed

to mediate interactions within PM, allowing the membrane to remain stable and cohesive

(Cartailler & Luecke 2003; Griffiths et al. 1996; Lancaster 2003). Some studies have

demonstrated that ion binding to bR is necessary for the proton pumping activity (Varo

et al. 1999) and have suggested several possible binding sites with different affinities

(Eliash et al. 2001; Sanz et al. 2001; Yonebayashi et al. 2004; Pardo et al. 1998). The

different nature of the ions in contact with both sides of PM (Table 1.1) is believed to

be related to the membrane stability and/or to enhance bR activity (Oren 1994; Engel

& Catchpole 2005; Cartailler & Luecke 2003), but little experimental data accounting

for specific ionic effects in PM is available. Generally, salt plays an important role for

tuning the interactions between lipids, bR residues and water molecules in PM. Ions are

not only involved in point charge shielding but can also polarize molecules and mediate

H-bonds (Fig. 1.6.B-C) as a way to screen a charge, (Collins 2004; Lancaster 2003) which

adds to the complexity of PM ion-mediated interactions.

The role of PM lattice and the role of bR trimeric arrangement is not yet fully under-

stood, despite some experiments suggesting a cooperative activity of the protein within

a trimer (Tokaji 1998; Varo et al. 1996; Cassim 1992; Komrakov & Kaulen 1995).

Inter-trimer interactions involve relatively long distances and do not allow proteins from

different trimers to interact directly, but are rather mediated through long-range electro-

static interactions and protein-lipid specific interactions (Sternberg et al. 1992; Grigorieff

et al. 1995; Cartailler & Luecke 2003; Weik et al. 1998). The regular and static trimers

22

Introduction

HHO

HHO

HHO

HHO H

HO

HHO

HHO H

HO

++

+ --

+

A B C

chaotrope

kosmotrope

water

cytoplasmic extracellular

Figure 1.6: (A) Cytoplasmic (left) and extracellular (right) surface electrostatic potential ofPM calculated from its 1m0l structure with MOLMOL. Red is negatively charged and bluepositively. Lipids appear in orange as they are mainly negatively charged, but their contribu-tion was not calculated. (B) The three main charge screening mechanisms. The charge canorient surrounding molecules if in a dielectric medium (up), balance a potential created bystatic dipoles (middle) or screen a punctual charge with opposite sign (down) (modified from(Lancaster 2003)). (C) Differences in interactions of ions with the first hydration layer (grey)depending on their nature. Kosmotropic ions can strongly interact with H2O molecules andpromote order whereas chaotropic ions tend to interact more weakly with H2O through highlyfluctuating interaction. (modified from (Collins 2004))

23

Introduction

assembly may contribute to the observed membrane rigidity (Zaccai 2000; Korenstein &

Hess 1978) despite the fluctuating electrostatic interactions. The lattice could also play a

role in the proton transfer to ATP-synthase sites (Teissie et al. 1990; Georgievskii et al.

2002) located in the surrounding red membrane (Stoeckenius et al. 1979; Cartailler &

Luecke 2003) by establishing a regular network of bound ions.

1.2.4 Purple Membrane as a model system

Since reported for the first time in 1971 by Oesterhelt and Stoeckenius (Oesterhelt &

Stoeckenius 1971) PM has provoked an enormous interest among scientists. The se-

quence of bR, the first discovered membrane protein, has been fully determined nearly

30 years ago (Khorana et al. 1979) and more than 10 different models of bR struc-

tures obtained by X-ray and electron diffraction are currently available (White 2006).

Diffraction studies on mutant protein have provided model structures of the protein in

most of the different states of the photocycle (apart for N and O) allowing an unambigu-

ous determination of the pumping process3. Many studies have also covered PM lipids

(Lee 2003, 2004; Hendler & Dracheva 2001; Cartailler & Luecke 2003) and their inter-

actions with bR and each other. This wealth of information made PM one of the best

known biological membrane and an important model system to confront new theories

and experimental methods. bR was indeed the first protein which structure was deter-

mined by X-ray crystallography of three-dimensional crystals grown in cubic lipid phase

(Luecke et al. 1998), refining the structure obtained earlier by electron diffraction of

two-dimensional crystals (Grigorieff et al. 1996). Local probe techniques have also been

used to study purple membrane (Tamayo et al. 1995) and bR was the first membrane

3 There has been some controversy about bR pumping chloride, but it was demonstrated that a mutationis necessary for bR to achieve chloride pumping (Sasaki et al. 1995) and a natural chloride pump hasbeen later identified in the same family as halorhodopsin (Kolbe et al. 2000). So far most of the evidencesupport bR as a proton pump.

24

Introduction

protein extensively studied by atomic force microscopy (Butt et al. 1990; Muller et al.

1995) and force spectroscopy (Oesterhelt et al. 2000).

Generally, due to its ready availability and to its important thermal and photochemical

stability (Brouillette et al. 1989; Jackson & Sturtevant 1978; Dancshazy et al. 1999;

Tan & Birge 1996), bR has become one of the most studied proteins, raising an enormous

interest not only in biochemistry, but also in physics and engineering where potential

applications of bR in photonic devices drive an important research area (Birge et al.

1999; Crittenden et al. 2003; Jussila et al. 2002; Bhattacharya et al. 2002).

The analogy between bR and GPCR proteins is also an important factor contributing

to the use of bR as a model system (Birge et al. 1999; Henderson et al. 1990). Unlike

most of the prokaryotic α-helical membrane proteins which are composed of six or twelve

transmembrane helices (Wallin & Von-Heijne 1998), eukaryotic α-helical membrane pro-

teins show a preference for seven transmembrane α-helices (Wallin & Von-Heijne 1998;

Koonin et al. 2002) mainly because of the GPCRs (Kroeze et al. 2003; Metpally &

Sowdhamini 2005). The bR structure with seven transmembrane α-helices enclosing

the active group is a characteristic of GPCRs (Kroeze et al. 2003; Flower 1999). The

first GPCR structure resolved, bovine rhodopsin, was achieved only recently (Palczewski

et al. 2000) and most of the previous pharmaceutical studies aiming at GPCR were

based on the structure of bR (Flower 1999; Klabunde & Hessler 2002; Henderson et al.

1990). To date, the only GPCR structure resolved remains rhodopsin, a representant

of the largest GPCR sub-family (A-family with 90% of GPCRs). Despite the fact that

bR is not a GPCR, it is related to rhodopsins (Henderson et al. 1990) and remain an

important model for pharmacological studies (Flower 1999).

25

Introduction

1.3 Atomic Force Microscopy: a powerful tool for biology

The invention of Atomic Force Microscopy (AFM) in 1986 by Binnig and coworkers

(Binnig et al. 1986) opened the field of local probes to the electrically non-conducting

biological samples. Over the last twenty years many studies have exploited the unique

ability offered by AFM to both image and probe biological systems in air and in liquid

under native-like conditions. (Alessandrini & Facci 2005; Jena & Horber 2002). At the

cell level, AFM studies have revealed some of the cellular mechanical properties (Rad-

macher et al. 1992), allowed the observation of cell membrane exocytosis (Schneider

et al. 1997) and adhesion to different substrate (Benoit et al. 2000; Kim et al. 2003).

At the single molecule level, globular proteins such as titin (Rief et al. 1997; Marszalek

et al. 1999; Williams et al. 2003), tenascin (Oberhauser et al. 1998), spectrin (Rief

et al. 1999), fibronectin (Rief et al. 1998) and ubiquitin (Cieplak & Marszalek 2005)

have been investigated through their unfolding and refolding patterns using force spec-

troscopy. The same technique has been applied to measure specific interactions such as

molecular recognition between proteins (Leckband 2001; Lee et al. 1994; Hinterdorfer

et al. 1996; Hinterdorfer & Dufrene 2006; Florin et al. 1994). The high sensitivity of

force spectroscopy allows precise quantification of the intermolecular forces involved, but

independent information on the system investigated (e.g. protein structure in the case of

un/re-folding studies or the nature of the interactions involved in the case of molecular

recognition) is necessary in order to provide a relevant interpretation of the results. High

resolution imaging also allowed the observation of antibodies structures (San Paulo &

Garcıa 2000), the activity chaperonins proteins (Mou et al. 1996), ordering of collagen

fibers (Paige et al. 1998) and direct visualization of enzymes at work (Radmacher et al.

1994; Guthold et al. 1999). AFM has equally been extensively use to investigate nucleic

acids properties such as base-pair interactions forces (Boland & Ratner 1995; Sattin et al.

2004; Bustamante et al. 1994; Hansma et al. 2004), super-coiling of DNA (Lyubchenko

26

Introduction

& Shlyakhtenko 1997; Jett et al. 2000) and mechanisms for the implication of RNA in

viral diseases (Kiselyova et al. 2001; Andersen et al. 2004).

In order to study membrane proteins and their interactions with lipids in real cell mem-

branes, it is first necessary to stably adsorb the membrane on a substrate (Jena & Horber

2002, chap. 11 by Czajkowsky & Shao). Considerable efforts have been made to exam-

ine the properties of natural and synthetic lipid bilayers involving mixtures of different

lipids (Rinia et al. 2001; Connell & Smith 2006; Grandbois et al. 1998; Milhiet et al.

2001; Dufrene et al. 1997; Tokumasu et al. 2003). Membrane proteins could then be

attached or reconstituted in a relevant lipid bilayer (Reviakine et al. 1998; Slade et al.

2002; Czajkowsky et al. 1998; Epand et al. 2001; Mou et al. 1995; Kim et al. 2005)

in order to provide useful biological information. High resolution imaging of the sur-

face of integral membrane proteins and force spectroscopy experiments in native and

reconstituted membranes could then be achieved, providing valuable information about

their interactions with the surrounding proteins and lipids (see chap. 2) (Muller et al.

2002; Scheuring et al. 2005). Furthermore, the determination of many integral mem-

brane protein structures over the past few years has opened new possibilities for relating

AFM observation with specific intra- and inter-molecular interactions involved in the

membrane proteins anchoring, stability and activity.

1.3.1 Atomic Force Microscopy vs other techniques

Several experimental techniques other than AFM can access and probe molecular details

of single proteins. Generally, each technique presents different advantages and difficul-

ties which makes it more suitable to address certain types of questions. At the present

time, the main investigation techniques relevant for single proteins study are diffractions

techniques (such as X-ray, neutron and electron diffraction), nuclear magnetic resonance

(NMR), scanning electron microscopy (SEM), traditional AFM and high-speed AFM

27

Introduction

(hsAFM). Some of the characteristics of these techniques are summarized in table 1.2,

and the main advantages and drawbacks of each technique are briefly discussed.

resolution

relative gap with native conditions

aquisition timeNMR

StandardAFM

SEM

0.1 nm1 nm

10 nm

dayhour

min

msecsec

High-speedAFM

week

TEM Neutron

Pulsed Xray

Xray

Idealtechnique

Increasing difficulty

Figure 1.7: Comparison of the main experimental techniques providing microscopic informa-tion relatively to the resolution they provide, the gap between the investigated protein nativeconditions and the measurement conditions, and the duration of measurement. The color scaleis an indication of the ease of these techniques, taking into account sample preparation (e. g.labelling or crystallization of proteins)

X-ray crystallography

Among the techniques providing atomic resolution, X-ray diffraction is the most widely

used, mainly for protein and DNA structures determination (∼ 83% of the Protein Data

Bank structures), but also for non-organic molecules. X-ray photons interact mainly

with valence electrons surrounding the sample nuclei, providing information about the

interacting electron spatial distributions and allowing localization of the nuclei posi-

tions. Its main advantage is the ability to resolve the structure of whole complexes (e.g.

28

Introduction

protein-lipid complex) with an outstanding precision (Lecomte et al. 2004). Since rel-

atively short exposition time are required (typically minutes for a complete structure),

time-resolved experiments with pulsed X-rays can provide picosecond time resolution

on specific systems (e.g a group of few atoms) (Schotte et al. 2003). The main dis-

advantage of this technique is the requirement of 3-dimensional crystal of the structure

to be resolved, which can be difficult to achieve, especially for membrane proteins4. In

order to reduce the crystal thermal motion and the radiation damages, it is necessary

to work at low temperatures (liquid nitrogen or helium), and the complex kinetics and

structural fluctuations associated with the protein activity can not be resolved. Besides

crystallization difficulties, the resolution will also depend on a stable X-ray source (e.g.

synchrotron or rotating-anode type source).

Neutron crystallography

Unlike X-ray photons, neutron interact mostly with nuclei through strong nuclear forces

and can only provide little information about electrons (spin coupling). Atomic resolu-

tion of the nuclei is however possible and neutron crystallography offers better contrast

over lipids and water molecules than X-ray diffraction (Gutberlet et al. 2001). Further-

more, information about protein dynamics and rigidity can be obtained through neutron

scattering experiments which are insensitive to static inhomogeneities (Zaccai 2000). The

disadvantages are similar than for X-ray crystallography and time resolved experiments

not possible.

Cryo-Transmission Electron Microscopy

This technique uses electron diffraction, and is therefore mainly sensitive to electrons

4 The development of lipid cubic phases (Landau & Rosenbusch 1996) has greatly improved possibili-ties for membrane protein structure determinations. There has however been some concern about themembrane protein stability in phospholipid based crystals (Lunde et al. 2006)

29

Introduction

charge densities (Coulomb forces) and to magnetic fields. This affects the resolution

and crystal structure determination is practically difficult. However, Cryo-Transmission

Electron Microscopy (Cryo-TEM) shows an important advantage over other diffraction

techniques; since electrons are transmitted through the sample, it requires only a 2-

dimensional crystal which in principle facilitates greatly membrane protein structure

determination (Torres et al. 2003). This advantage also reflects a limitation: the elec-

tron transmission imposes a maximum sample thickness of ∼ 100nm. The 3-dimensional

structure studied has to be reconstructed from projection maps (Frank 2002). The other

limitations are the low temperature requirements as for X-ray and neutron diffraction,

and a long integration time (table 1.2).

Scanning Electron Microscopy

Scanning electron microscopy (SEM) provides a substantially lower resolution than diffrac-

tion techniques and molecules . 100KDa can not be studied with this technique. SEM

is however useful to visualize larger structures and its rapidity (< min/scan) offers a

possibility for studying molecular dynamics. Its main advantage is the ease of use, but

vacuum conditions are required. Environmental SEM (ESEM) can operate in controlled

gases and vapors (e.g. water) and no constraints on temperature are necessary.

Nuclear Magnetic Resonance

Nuclear magnetic resonance (NMR) can provide not only atomic resolution structures

of proteins (∼ 10% of PDB structures) but also some information about the molecule

dynamics at the atomic level and does not require crystallization of the proteins (Luca

et al. 2003; Jonas 2002). Furthermore, since this technique can operate at room tem-

perature and in liquid, the information obtained is directly relevant for real biological

systems (Torres et al. 2003; Jonas 2002). The best resolution is however obtained in a

30

Introduction

controlled atmosphere (solid-state NMR, ssNMR) (Luca et al. 2003; Torres et al. 2003).

The main problem arise from the necessity of labelling proteins in order to probe them,

imposing an enormous amount of work in order to resolve complete structures. It also

requires important magnetic fields (typ. 10-45T ) and large amounts of sample.

Traditional Atomic Force Microscopy

Traditional AFM allows local probing of single molecules in native-like conditions (Muller

et al. 2002). AFM can provide quantitative information about intra- and inter-molecules

interactions (Oesterhelt et al. 2000). Its main advantages are the ability to study wild-

type proteins in a native-like environment, and the fact that it uses a local probe, thus

allowing detection and imposition of various constraint at the single molecule level di-

rectly. The interactions between the probe and the sample can be tuned through the

properties of the buffer when working in liquid (Muller et al. 1999). As a local probe,

its main limitation comes from the fact that AFM can only image the surface of the

sample studied and with a resolution lower than diffraction techniques or NMR. It also

imposes mechanical constraints on the sample while scanning which requires the sample

to be attached to a substrate.

High-speed Atomic Force Microscopy

High-speed AFM (hsAFM) presents the same advantages that traditional AFM with the

exception of time resolution. hsAFM can achieve sub-molecular resolution while main-

taining scanning rates of ∼ 100 frames/sec. At the present time, hsAFM is the only

technique that can directly image structural changes of single molecules in real time and

in native conditions (Ando et al. 2001). The traditional AFM limitations described

above also apply to hsAFM.

31

Introduction

Technique Type Resolution Conditions Duration

X-ray diffraction indirect atomic (1-2A) low temperature min (ps-ms with time res.(typ. < 100K) exp. on specific syst.)

Neutron diffraction indirect atomic-5A low temp. days(typ. < 100K)

Cryo-TEM indirect atomic-5A low temp. days(typ. < 100K)

SEM direct 5-50A light vacuum < 1min(Environ. SEM)

NMR indirect atomic dry (HR-ssNMR) hours-daysor liquid

traditional AFM direct 5-20A native-like min

High-speed AFM direct 5-20A native-like ms

Table 1.2 Comparison of the main microscopy techniques relevant to study of singleproteins.

1.4 Goals and Motivations

The main part of the work presented in this report uses AFM, hsAFM and force spec-

troscopy to investigate the effects of ions on the interactions governing structure, activity

and properties of a model membrane protein in a native-like environment: bR in PM. The

choice of bR as a model system is motivated first by the wealth of information available

which allows a better interpretation of experimental results, but also by the potential

biomedical and technological applications which constitute an important driving force.

The motivations can be classified in three main categories:

1. Biophysical understanding of the nature and the specific role of the different inter-

actions dynamically controlling bR structure and function

32

Introduction

2. Understanding the nature of the AFM tip-sample interactions in order to control

and optimize imaging while gathering biologically relevant information.

3. Pharmacological and technological applications

1.4.1 Relating specific interactions with membrane structure and activity

The structure of a protein depends on both its amino acid sequence and its surrounding

environment. In the case of membrane proteins, the environment involves lipids, water

molecule and ions and can be highly specific with well defined interactions maintaining

the protein structure while enhancing its activity (Lee 2003; Petrache et al. 2005; White

et al. 2001).

The AFM ability to operate locally and in a controlled environment (temperature, pH,

ionic type and density, chemical composition of the buffer, relative humidity) allows isola-

tion of the different types of interactions responsible for the protein structure and activity

using imaging and force spectroscopy obtained under different environmental conditions.

Further confrontation of the results with available data such as the protein sequence,

high-resolution structure and combination with other techniques can provide insight in

the way the protein interacts with water, ions and surrounding molecules (McAllister

et al. 2005; Grudzielanek et al. 2005) and ultimately relate specific interaction with the

protein activity.

Recently, the development of hsAFM has provided real-time observations of single molecule

activity in solution (Ando et al. 2001, 2005), opening a new way to study single-protein

dynamics in their native environment. This powerful technique combines the advantages

of traditional AFM but with unprecedented time resolution (typically 1000 to 5000 times

faster) making it one of the most promising tools to directly image and probe dynamical

processes in real systems and at molecular level.

33

Introduction

1.4.2 AFM to study membrane proteins and vice versa

The ability of AFM to image single proteins under physiological-like conditions and to

measure molecular interaction forces makes it an important tool for membrane protein

systems investigation (see review chap. 2). Furthermore, since AFM is a local probe

technique, no chemical or biological modification of the studied system such as mutation

or labelling is generally required, and the AFM tip can be used as a channel for imposing

various constraints locally on the system studied (Frederix et al. 2005; Afrin et al. 2004;

Kessler et al. 2006). AFM has therefore an enormous potential in molecular level studies

(Chap. 2). Pharmacological, medical research (Brody et al. 2006; Edwardson & Hen-

derson 2004; Lee et al. 2006; Stolz et al. 2000; Young et al. 2003; Blackley et al. 2000)

and biotechnologies also rely on AFM to image, sense and characterize single molecule

systems under physiological conditions (Lang et al. 2002) as well as for nano-fabrication

(Tang et al. 2004).

With the increasing amount of studies addressing biological systems by the means of

AFM, there is a strong need for a theoretical framework able to describe and predict

the multitude of interactions observed at the molecular level. Furthermore, the com-

plexity and the multiplicity of the interactions occurring between the AFM tip apex

and the sample investigated makes any interpretation of AFM data such as images and

force spectroscopy curves difficult. So far, several AFM studies present sub-nanometer

resolution images of proteins in air and in liquid (Muller et al. 1995; San Paulo &

Garcıa 2000) but the different interactions involved appear to be highly non-linear and

vary in types and range, which makes it difficult to establish criteria for achieving such

high-resolution (Giessibl 2003; Hu & Raman 2006; Tamayo & Garcia 1998). The work

achieved in liquid, the most relevant for biology, was explained using the Derjaguin-

Landau-Verwey-Overbeek (DLVO) theory which is based on continuum considerations

(more details in chap. 4). However at resolutions < 1 nm, the size of individual hydrated

34

Introduction

ions and the role of local hydration forces can not be neglected, since most of the biolog-

ical processes require the presence of specific ions with different relative concentrations

(typically 20 mM − 150 mM). The AFM study of a membrane protein of known

structure such as bacteriorhodopsin offers the possibility of experimentally investigating

the role of local hydration forces in the image formation by varying the ionic composition

and concentration of the working buffer. AFM can indeed locally probe relative changes

in the tip-sample interactions, and relating these changes to the operating parameters

can provide an experimental base for the development of a discrete theory and establish

biophysically relevant criteria for high-resolution.

1.4.3 Studying membrane proteins for medical and technological applica-

tions

The transmembrane nature of integral membrane proteins make them occupy a critical

position in every living organism for which survival and good development depends on

the efficiency and reliability of the different tasks carried out by membrane proteins (Berg

et al. 2002; Houslay & Stanley 1982; Byrne & Iwata 2002). Many diseases are directly

correlated with malfunction of certain membrane proteins, and the main effort of phar-

maceutical research is aimed at these particular proteins (Drews 2000). At the present

time, more than 70% of available drugs target membrane proteins, 45% of which are

aimed at GPCRs (Brunton et al. 2006; Drews 2000; Flower 1999; Klabunde & Hessler

2002; Becker et al. 2004).

Besides the important pharmaceutical interest in membrane proteins, part of the research

on these molecular machines aims at utilizing their unique properties for technological

applications. Bacteriorhodospin, was proposed to be used as a multiple-state memory

unit for bio-computing (Birge 1990a; Birge et al. 1999). Other applications such as

single-molecule detectors (Ivanov et al. 2006; Nicolini et al. 1999), bio-sensors (Mur-

35

Introduction

phy 2006) and nano-switches aim at using membrane proteins specific recognition ability

and the multiple states characterizing their activity (Giardi & Pace 2005; Sergi et al.

2004; Danelon et al. 2004). Technological applications are also motivated by the fact

that membrane protein are highly reproducible, generally far smaller than an equivalent

semi-conductor solution and of very low cost to produce, especially if self assembly can

be achieved.

1.4.4 Overview of the thesis

Chapter 2 presents a brief review of the main AFM contributions on membrane proteins

systems. Its aim is to provide an overview of the typical question addressed so far with

AFM in membrane biology, and to demonstrate the ability of AFM to investigate mem-

brane proteins in native-like environments at the molecular level.

In chapter 3, a new AFM approach is used to distinguish both sides of an asymmetric

membrane (PM) without requiring molecular resolution. This technique is further com-

bined with force spectroscopy and applied to side-specifically measure the evolution PM

both leaflets stiffness with ionic concentration. The results demonstrate differential stiff-

ening and lipid mobility within the leaflets when increasing the salt concentration but

no clear dependence on pH variations, apart for the membrane assembly. The results are

analyzed in terms of electrostatic and specific protein-lipid and lipid-lipid interactions

within the membrane.

Chapter 4 further investigates the asymmetric properties of PM, through PM side-specific

dynamical (oscillating cantilever) force spectroscopy which allows mapping of the tip-

sample local force field over a wide range of ionic concentrations. Based on theses results,

a new theoretical approach is developed and provides criteria for high resolution. When

applied, this new method demonstrated for the first time sub-molecular resolution of

36

Introduction

both PM sides over the range of ionic concentrations considered as unfavourable by pre-

vious studies. The membrane structure exhibit changes in roughness and trimer packing

evolving with the ionic concentration. The results suggest specific ion binding on the

extracellular side of the membrane while less specific interactions occur at the cytoplas-

mic surface. Ion mediated hydrations forces are seen to be essential for bR structure and

function. In order to relate bR specific intra- and inter-molecular interactions with its

activity and the overall membrane structure and cohesion, unfolding curves have been

taken from both protein termini distinctly while varying the ionic concentration. This

approach allows distinction of steric interactions from electrostatic contributions while

precisely locating, if relevant, the protein amino-acid responsible for the unfolding poten-

tial barrier observed. The results highlight the role of tryptophan residues in anchoring

and supporting the protein through steric interactions and hydrogen bonds. Based on

this robust scaffolding located exclusively in the PM extracellular leaflet, the protein uses

electrostatic interactions to dynamically adapt its cytoplasmic structure and enhance the

pumping activity. Specific ion binding to bR extracellular loops could be observed and

is related to proton mobility along the PM extracellular surface. Finally, the role of spe-

cific ions, naturally present in PM native environments is addressed by imaging PM in

different ionic environments. The high-resolution imaging conditions and the structure

of the membrane appear to be different when varying the hydrated ion size, confirming

previous results about the role of ions on electrostatic interactions within the membrane.

Aiming to image protein working in real time, chapter 5 accounts for studies achieved

with high-speed AFM. At lower resolution, it is possible to image dynamical tip-sample

interactions changes through continuous modification of the buffer. While such obser-

vation are useful to understand the interactions forming the image observed, the main

goal of high-speed AFM remains direct observation of single protein activity, combining

high-speed with sub-molecular resolution. Applying the high-resolution theory devel-

37

Introduction

oped in chapter 4, it was possible to image photo-induced conformational changes of

single bR α-helices in native-like environment at scanning rate higher than 50 frame/sec

(10 msec/frame). The results show a large conformational change induced by a twist of

helices E and F away from the protein center, with always a single protein per trimer

isomerizing at a time.

The last experimental chapter, chapter 6, concentrates on technological applications.

Several steps towards the building of a bR based single-molecule photodetector are un-

dertaken, providing encouraging preliminary results. One of the main steps is the devel-

opment of a low current measurement system with hardware and software optimization

for low noise recordings. The system is successfully used for biosensing protein attach-

ment to nitrogen doped carbon nanotubes.

1.5 References

Afrin, Rehana, Yamada, Takafumi, & Ikai, Atsushi. 2004. Analysis of forcecurves obtained on the live cell membrane using chemically modified AFM probes.Ultramicroscopy, 100(3-4), 187.

Ahl, P. L., & Cone, R. A. 1984. Light activates rotations of bacteriorhodopsin inthe purple membrane. Biophys. J., 45(6), 1039–1049.

Albeck, Amnon, Friedman, Noga, Sheves, Mordechai, & Ottolenghi,Michael. 1986. Role of Retinal Isomerizations and Rotations in the Photocycle ofBacteriorhodopsin. J. Am. Chem. Soc., 108(15), 4614–4618.

Alessandrini, Andrea, & Facci, Paolo. 2005. AFM: a versatile tool in biophysics.Meas. Sci. Tech., 16(6), R65.

Andersen, Ebbe S., Antoranz Contera, Sonia, Knudsen, Bjarne,Damgaard, Christian K., Besenbacher, Flemming, & Kjems, Jorgen. 2004.Role of the Trans-activation Response Element in Dimerization of HIV-1 RNA. J. Biol.Chem., 279(21), 22243–22249.

Ando, Toshio, Kodera, Noriyuki, Takai, Eisuke, Maruyama, Daisuke,Saito, Kiwamu, & Toda, Akitoshi. 2001. A high-speed atomic force micro-scope for studying biological macromolecules. Proc. Natl. Acad. Sci. USA, 98(22),12468–12472.

38

Introduction

Ando, Toshio, Kodera, Noriyuki, Uchihashi, Takayuki, Miyagi, Atsushi,Nakakita, Ryo, Yamashita, Hayato, & Matada, Keiko. 2005. High-speedAtomic Force Microscopy for Capturing Dynamic Behavior of Protein Molecules atWork. e-Journ. Surf. Sci. Nanotech., 3, 384–392.

Becker, Oren M., Marantz, Yael, Shacham, Sharon, Inbal, Boaz, Heifetz,Alexander, Kalid, Ori, Bar-Haim, Shay, Warshaviak, Dora, Fichman,Merav, & Noiman, Silvia. 2004. G protein-coupled receptors: In silico drug dis-covery in 3D. Proc. Natl. Acad. Sci. USA, 101(31), 11304–11309.

Belrhali, Hassan, Nollert, Peter, Royant, Antoine, Menzel, Christoph,Rosenbusch, Jurg P., Landau, Ehud M., & Pebay-Peyroula, Eva. 1999.Protein, lipid and water organization in bacteriorhodopsin crystals: a molecular viewof the purple membrane at 1.9 A resolution. Structure, 7(8), 909–917.

Benoit, Martin, Gabriel, Daniela, Gerisch, Gunther, & Gaub, Hermann E.2000. Discrete interactions in cell adhesion measured by single-molecule force spec-troscopy. Nat. Cell Biol., 2(6), 313.

Berg, Jeremy M., Tymoczko, John L., & Stryer, Lubert. 2002. Biochemistry.5th edn. New York: W. H. Freeman and Company.

Bhattacharya, Pallab, Xu, Jian, Vr, Gyorgy, Marcy, Duane L., & Birge,Robert R. 2002. Monolithically integrated bacteriorhodopsin-GaAs field-effect tran-sistor photoreceiver. Opt. Lett., 27(10), 839–841.

Binnig, G., Quate, C. F., & Gerber, Ch. 1986. Atomic Force Microscope. Phys.Rev. Lett., 56(9), 930.

Birge, R. R. 1990a. Photophysics and Molecular Electronic Applications of theRhodopsins. Annu. Rev. Phys. Chem., 41(1), 683–733.

Birge, R. R., Gillespie, N. B., Izaguirre, E. W., Kusnetzow, A., Lawrence,A. F., Singh, D., Song, Q. W., Schmidt, E., Stuart, J. A., Seethara-man, S., & Wise, K. J. 1999. Biomolecular Electronics: Protein-Based AssociativeProcessors and Volumetric Memories. J. Phys. Chem. B, 103(49), 10746.

Birge, Robert R. 1990b. Nature of the primary photochemical events in rhodopsinand bacteriorhodopsin. Biochim. Biophys. Acta, 1016(3), 293–327.

Blackley, H. K. L., Sanders, G. H. W., Davies, M. C., Roberts, C. J.,Tendler, S. J. B., & Wilkinson, M. J. 2000. In-situ atomic force microscopystudy of β-amyloid fibrillization. J. Mol. Biol., 298(5), 833.

Boland, T., & Ratner, B. D. 1995. Direct Measurement of Hydrogen Bonding inDNA Nucleotide Bases by Atomic Force Microscopy. Proc. Natl. Acad. Sci. USA,92(12), 5297–5301.

39

Introduction

Bowie, J. U., Luthy, R., & Eisenberg, D. 1991. A method to identify proteinsequences that fold into a known three-dimensional structure. Science, 253(5016),164–170.

Braiman, M. S., Bousche, O., & Rothschild, K. J. 1991. Protein Dynamics inthe Bacteriorhodopsin Photocycle: Submillisecond Fourier Transform Infrared Spectraof the L, M, and N Photointermediates. Proc. Natl. Acad. Sci. USA, 88(6), 2388–2392.

Brody, Sarah, Anilkumar, Thapasimuthu, Liliensiek, Sara, Last, Julie A.,Murphy, Christopher J., & Pandit, Abhay. 2006. Characterizing NanoscaleTopography of the Aortic Heart Valve Basement Membrane for Tissue EngineeringHeart Valve Scaffold Design. Tissue Eng., 12(2), 413–421.

Brouillette, Christie G., McMichens, Ruth B., J. Stern H., Lawrence,& Khorana, Gobind. 1989. Structure and thermal stability of monomeric bacteri-orhodopsin in mixed pospholipid/detergent micelles. Proteins, 5(1), 38–46.

Brown, Leonid S., Needleman, Richard, & Lanyi, Janos K. 2002. Confor-mational change of the E-F interhelical loop in the M photointermediate of bacteri-orhodopsin. J. Molec. Biol., 317(3), 471.

Brunton, Laurence L., Lazo, John S., & Parker, Keith L. 2006. The Phar-macological Basis of Therapeutics. 11th edn. Goodman & Gilman’s. McGraw-Hill.

Bustamante, Carlos, Marko, J. F., Siggia, E. D., & Smith, S. 1994. EntropicElasticity of λ-Phage DNA. Science, 265, 1599–1600.

Butt, H. J., Downing, K. H., & Hansma, P. K. 1990. Imaging the membraneprotein bacteriorhodopsin with the atomic force microscope. Biophys. J., 58(6), 1473–1480.

Byrne, Bernadette, & Iwata, So. 2002. Membrane protein complexes. Curr. Op.Struct. Biol., 12(2), 239.

Cartailler, J.P., & Luecke, H. 2003. X-ray and crystallographic analysis of lipid-protein interactions in the bacteriorhodopsin purple membrane. Annu. Rev. Biophys.Biomol. Struct., 32, 285–310.

Cassim, Joseph Y. 1992. Unique biphasic band shape of the visible circular dichroismof bacteriorhodopsin in purple membrane Excitons, multiple transitions or proteinheterogeneity? Biophys. J., 63(5), 1432–1442.

Cieplak, Marek, & Marszalek, Piotr E. 2005. Mechanical unfolding of ubiquitinmolecules. J. Chem. Phys., 123(19), 194903.

Collins, Kim D. 2004. Ions from the Hofmeister series and osmolytes: effects onproteins in solution and in the crystallization process. Methods, 34(3), 300.

40

Introduction

Connell, Simon D., & Smith, D. Alastair. 2006. The atomic force microscopeas a tool for studying phase separation in lipid membranes (Review). Molec. Memb.Biol., 23(1), 17–28.

Corcelli, Angela, Lattanzio, Veronica M. T., Mascolo, Giuseppe, Pa-padia, Paride, & Fanizzi, Francesco. 2002. Lipid-protein stoichiometries in acrystalline biological membrane: NMR quantitative analysis of the lipid extract of thepurple membrane. J. Lipid Res., 43(1), 132–140.

Crittenden, S., Howell, S., Reifenberger, R., Hillebrecht, J., &Birge, R. R. 2003. Humidity-dependent open-circuit photovoltage from a bacte-riorhodopsin/indium tin oxide bioelectronic heterostructure. Nanotechnology, 14(5),562.

Czajkowsky, Daniel M., Sheng, Sitong, & Shao, Zhifeng. 1998. Staphylococcala-hemolysin can form hexamers in phospholipid bilayers. J. Mol. Biol., 276(2), 325.

Dancshazy, Zsolt, Tokaji, Zsolt, & Der, Andras. 1999. Bleaching of bacteri-orhodopsin by continuous light. FEBS Lett., 450(2), 154–157.

Danelon, C., Lindemann, M., Borin, C., Fournier, D., & Winterhalter,M. 2004. Channel-forming membrane proteins as molecular sensors. IEEE Trans.NanoBioscience, 3(1), 46.

de Planque, M. R. R., Bonev, B. B., Demmers, J. A. A., Greathouse, D. V.,Koeppe, R. E., Separovic, F., Watts, A., & Killian, J. A. 2003. InterfacialAnchor Properties of Tryptophan Residues in Transmembrane Peptides Can Domi-nate over Hydrophobic Matching Effects in Peptide-Lipid Interactions. Biochemistry,42(18), 5341–5348.

De Rosa, M., Morana, A., Riccio, A., Gambacorta, A., Trincone, A., &Incani, O. 1994. Lipids of the Archaea: a new tool for bioelectronics. Biosens.Bioelec., 9(9-10), 669.

De Weer, Paul. 2000. A Century of Thinking About Cell Membranes. Ann. Rev.Physiol., 62(1), 919–926.

Der, A., & Keszthelyi, L. 2001. Charge Motion during the Photocycle of Bacteri-orhodopsin. Biochemistry (Moscow), 66(11), 1234–1248.

Dioumaev, A. K., Brown, L. S., Needleman, R., & Lanyi, J. K. 2001. Couplingof the Reisomerization of the Retinal, Proton Uptake, and Reprotonation of Asp-96 inthe N Photointermediate of Bacteriorhodopsin. Biochemistry, 40(38), 11308–11317.

Drews, Jurgen. 2000. Drug Discovery: A Historical Perspective. Science, 287(5460),1960–1964.

41

Introduction

Dufrene, Y. F., Barger, W. R., Green, J. B. D., & Lee, G. U. 1997.Nanometer-Scale Surface Properties of Mixed Phospholipid Monolayers and Bilayers.Langmuir, 13(18), 4779–4784.

Dym, O., Mevarech, M., & Sussman, J. L. 1995. Structural Features That Stabi-lize Halophilic Malate Dehydrogenase from an Archaebacterium. Science, 267(5202),1344–1346.

Edwardson, J. Michael, & Henderson, Robert M. 2004. Atomic force mi-croscopy and drug discovery. Drug Discovery Today, 9(2), 64.

Eisenberg, David. 2003. The discovery of the α-helix and β-sheet, the principalstructural features of proteins. Proc. Natl. Acad. Sci. USA, 100(20), 11207–11210.

Eliash, Tamar, Weiner, Lev, Ottolenghi, Michael, & Sheves, Mordechai.2001. Specific Binding Sites for Cations in Bacteriorhodopsin. Biophys. J., 81(2),1155–1162.

Engel, Milton B., & Catchpole, Hubert R. 2005. A microprobe analysis ofinorganic elements in Halobacterium salinarum. Cell Biol. Int., 29(8), 616.

Epand, Raquel F., Yip, Christopher M., Chernomordik, Leonid V., LeDuc,Danika L., Shin, Yeon-Kyun, & Epand, Richard M. 2001. Self-assembly ofinfluenza hemagglutinin: studies of ectodomain aggregation by in situ atomic forcemicroscopy. Biochimi. Biophys. Acta, 1513(2), 167.

Essen, L. O., Siegert, R., Lehmann, W. D., & Oesterhelt, D. 1998. Lipidpatches in membrane protein oligomers: Crystal structure of the bacteriorhodopsin-lipid complex. Proc. Natl. Acad. Sci. USA, 95(20), 11673–11678. Sulfoglycolipids inEC leaflet.

Eyre, Tina A., Partridge, Linda, & Thornton, Janet M. 2004. Computationalanalysis of a-helical membrane protein structure: implications for the prediction of 3Dstructural models. Prot. Eng. Design Select., 17(8), 613–624.

Falk, K. E., Karlsson, K. A., & Samuelsson, B. E. 1980. Structural analysisby mass spectrometry and NMR spectroscopy of the glycolipid sulfate from Halobac-terium salinarium and a note on its possible function. Chem. Phys. Lipids, 27(1),9–21.

Florin, E. L., Moy, V. T., & Gaub, H. E. 1994. Adhesion forces between individualligand-receptor pairs. Science, 264(5157), 415–417.

Flower, Darren R. 1999. Modelling G-protein-coupled receptors for drug design.Biochim. Biophys. Acta, 1422(3), 207.

42

Introduction

Frank, Joachim. 2002. Single-particle Imaging of Macromolecules by Cryo-ElectronMicroscopy. Annu. Rev. Biophys. Biomol. Struct., 31(1), 303–319.

Frederix, Patrick L. T. M., Gullo, Maurizio R., Akiyama, Terunobu,Tonin, Andreas, Rooij, Nicolaas F. de, Staufer, Urs, & Engel, Andreas.2005. Assessment of insulated conductive cantilevers for biology and electrochemistry.Nanotechnology, 16(8), 997. 0957-4484.

Gat, Y., & Sheves, Mordechai. 1993. A Mechanism for Controlling the pK, of theRetinal Protonated Schiff Base in Retinal Proteins. A Study with Model Compounds.J. Am. Chem. Soc., 115(9), 3772–3773.

Georgievskii, Yuri, Medvedev, Emile S., & Stuchebrukhov, Alexei A. 2002.Proton Transport via the Membrane Surface. Biophys. J., 82(6), 2833–2846.

Giardi, Maria Teresa, & Pace, Emanuela. 2005. Photosynthetic proteins fortechnological applications. Trends Biotech., 23(5), 257.

Giessibl, Franz J. 2003. Advances in atomic force microscopy. Rev. Mod. Phys.,75(3), 949.

Grandbois, Michel, Clausen-Schaumann, Hauke, & Gaub, Hermann. 1998.Atomic Force Microscope Imaging of Phospholipid Bilayer Degradation by Phospho-lipase A2. Biophys. J., 74(5), 2398–2404.

Griffiths, J. A., King, J., Yang, D., Browner, R., & El-Sayed, M. A. 1996.Calcium and Magnesium Binding in Native and Structurally Perturbed Purple Mem-brane. J. Phys. Chem., 100(3), 929–933.

Grigorieff, N., Beckmann, E., & Zemlin, F. 1995. Lipid Location inDeoxycholate-treated Purple Membrane at 2.6 A. J. Mol. Biol., 254(3), 404–415.

Grigorieff, N., Ceska, T.A., Downing, K.H., Baldwin, J.M., & Henderson,R. 1996. Electron-crystallographic refinement of the structure of bacteriorhodopsin.J. Mol. Biol., 259, 393–421.

Groma, G. I., Szabo, G., & Varo, Gy. 1984. Direct measurement of picosecondcharge separation in bacteriorhodopsin. Nature, 308(5959), 557.

Grudzielanek, Stefan, Jansen, Ralf, & Winter, Roland. 2005. SolvationalTuning of the Unfolding, Aggregation and Amyloidogenesis of Insulin. J. Mol. Biol.,351(4), 879.

Gutberlet, Thomas, Heinemann, Udo, & Steinera, Michael. 2001. Proteincrystallography with neutrons - status and perspectives. Acta Crystal. D, 57(2), 349–354.

43

Introduction

Guthold, Martin, Zhu, Xingshu, Rivetti, Claudio, Yang, Guoliang, Thom-son, Neil H., Kasas, Sandor, Hansma, Helen G., Smith, Bettye, Hansma,Paul K., & Bustamante, Carlos. 1999. Direct Observation of One-DimensionalDiffusion and Transcription by Escherichia coli RNA Polymerase. Biophys. J., 77(4),2284–2294.

Haltia, Tuomas, & Freire, Ernesto. 1995. Forces and factors that contribute tothe structural stability of membrane proteins. Bioenergetics, 1228(1), 1–27.

Hansma, Helen G., Kasuya, Kenichi, & Oroudjev, Emin. 2004. Atomic forcemicroscopy imaging and pulling of nucleic acids. Curr. Op. Struct. Biol., 14(3), 380.

Henderson, R. 1977. The Purple Membrane from Halobacterium Halobium. Annu.Rev. Biophys. Bioeng., 6(1), 87–109.

Henderson, R., Jubb, J.S., & Whytock, S. 1978. Specific labelling of the proteinand lipid on the extracellular surface of purple membrane. J. Mol. Biol., 123, 259–274.

Henderson, R., Baldwin, J. M., Ceska, T. A., Zemlin, F., Beckmann, E.,& Downing, K. H. 1990. Model for the structure of bacteriorhodopsin based onhigh-resolution electron cryo-microscopy. J. Mol. Biol., 213(4), 899.

Hendler, R. W. 2005. An Apparent General Solution for the Kinetic Models of theBacteriorhodopsin Photocycles. J. Phys. Chem. B, 109(34), 16515–16528.

Hendler, R. W., & Dracheva, S. 2001. Importance of Lipids for BacteriorhodopsinStructure, Photocycle, and Function. Biochemistry (Moscow), 66(11), 1311–1314.

Hendler, Richard W., Barnett, Steven M., Dracheva, Swetlana, Bose,Salil, & Levin, Ira W. 2003. Purple membrane lipid control of bacteriorhodopsinconformational flexibility and photocycle activity. An infrared spectroscopic study.Eur. J. Biochem., 270(9), 1920–1925.

Heyn, Maarten P., Borucki, Berthold, & Otto, Harald. 2000. Chromophorereorientation during the photocycle of bacteriorhodopsin: experimental methods andfunctional significance. Biochim. Biophys. Acta, 1460(1), 60.

Hinterdorfer, Peter, & Dufrene, Yves F. 2006. Detection and localization ofsingle molecular recognition events using atomic force microscopy. Nat. Meth., 3(5),347.

Hinterdorfer, Peter, Baumgartner, Werner, Gruber, Hermann J.,Schilcher, Kurt, & Schindler, Hansgeorg. 1996. Detection and localizationof individual antibody-antigen recognition events by atomic force microscopy. Proc.Natl. Acad. Sci. USA, 93(8), 3477–3481.

44

Introduction

Houslay, Miles D., & Stanley, Keith K. 1982. Dynamics of biological membranes:influence on synthesis, structure, and function. New York: Wiley Interscience.

Hu, Shuiqing, & Raman, Arvind. 2006. Chaos in Atomic Force Microscopy. Phys.Rev. Lett., 96(3), 036107.

Ippen, E. P., Shank, C. V., Lewis, A., & Marcus, M. A. 1978. Subpicosecondspectroscopy of bacteriorhodopsin. Science, 200(4347), 1279–1281.

Isenbarger, T. A., & Krebs, M. P. 1999. Role of Helix-Helix Interactions in As-sembly of the Bacteriorhodopsin Lattice. Biochemistry, 38(28), 9023–9030.

Ivanov, Yuri D., Govorun, Vadim M., Bykov, Victor A., & Archakov,Alexander I. 2006. Nanotechnologies in proteomics. Proteomics, 6(5), 1399–1414.

Jackson, Meyer B., & Sturtevant, Julian M. 1978. Phase transitions of thepurple membranes of Halobacterium halobium. Biochemistry, 17(5), 911–915.

Jacobson, K., Sheets, ED., & Simson, R. 1995. Revisiting the fluid mosaic modelof membranes. Science, 268(5216), 1441–1442.

Jain, M. K., & Wagner, R. C. 1980. Introduction to Biological Membranes. NewYork: John Wiley and Sons.

Jena, Bhanu P., & Horber, J. K. Heinrich. 2002. Atomic Force Microscopy inCell Biology. Vol. 67. San Diego: Academic Press.

Jett, Stephen D., Cherny, Dmitry I., Subramaniam, Vinod, & Jovin,Thomas M. 2000. Scanning force microscopy of the complexes of p53 core domainwith supercoiled DNA. J. Mol. Biol., 299(3), 585.

Jonas, Jiri. 2002. High-resolution nuclear magnetic resonance studies of proteins.Biochim. Biophys. Acta, 1595(1-2), 145.

Jussila, T., Li, M., Tkachenko, N.V., Parkkinen, S., Li, B., Jiang, L.,& Lemmetyinen, H. 2002. Transient absorption and photovoltage study of self-assembled bacteriorhodopsin/polycation multilayer films. Biosens. Bioelec., 17(6),509–515.

Kamikubo, Hironari, Kataoka, Mikio, Varo, Gyorgy, Oka, Toshihiko,Tokunaga, Fumio, Needleman, Richard, & Lanyi, Janos K. 1996. Struc-ture of the N intermediate of bacteriorhodopsin revealed by x-ray diffraction. Proc.Natl. Acad. Sci. USA, 93(4), 1386–1390.

Kates, M., Kushwaha, S. C., & Sprott, G. D. 1982. Lipids of purple membranefrom extreme halophiles and of methanogenic bacteria. Page 98 of: Methods Enzymol.,vol. Volume 88. Academic Press. Lester Packer.

45

Introduction

Kessler, Max, Gottschalk, Kay E., Janovjak, Harald, Muller, Daniel J.,& Gaub, Hermann E. 2006. Bacteriorhodopsin Folds into the Membrane againstan External Force. J. Mol. Biol., 357(2), 644.

Khorana, H. Gobind, Gerber, Gerhard E., Herlihy, Walter C., Gray,Christopher P., Anderegg, Robert J., Nihei, Kayoko, & Biemann, Klaus.1979. Amino Acid Sequence of Bacteriorhodopsin. Proc. Natl. Acad. Sci. USA, 76(10),5046–5050.

Killian, J. Antoinette, & von Heijne, Gunnar. 2000. How proteins adapt to amembrane-water interface. Trends Biochem. Sci., 25(9), 429.

Kim, David T., Blanch, Harvey W., & Radke, Clayton J. 2005. Imaging ofreconstituted purple membranes by atomic force microscopy. Colloids Surf. B, 41(4),263.

Kim, H., Arakawa, Hideo, Osada, Toshiya, & Ikai, Atsushi. 2003. Quantifica-tion of cell adhesion force with AFM: distribution of vitronectin receptors on a livingMC3T3-E1 cell. Ultramicroscopy, 97(1-4), 359.

Kiselyova, O. I., Yaminsky, I. V., Karger, E. M., Frolova, O. Yu,Dorokhov, Y. L., & Atabekov, J. G. 2001. Visualization by atomic force mi-croscopy of tobacco mosaic virus movement protein-RNA complexes formed in vitro.J. Gen. Virol., 82(6), 1503–1508.

Klabunde, Thomas, & Hessler, Gerhard. 2002. Drug Design Strategies for Tar-geting G-Protein-Coupled Receptors. ChemBioChem, 3(10), 928–944.

Kobayashi, Takayoshi, Saito, Takashi, & Ohtani, Hiroyuki. 2001. Real-timespectroscopy of transition states in bacteriorhodopsin during retinal isomerization.Nature, 414(6863), 531.

Koch, M. H. J., Dencher, N. A., Oesterhelt, D., Plhn, H.-J., Rapp, G., &Bldt, G. 1991. Time-resolved X-ray diffraction study of structural changes associatedwith the photocycle of bacteriorhodopsin. EMBO J., 10(3), 521–526.

Kolbe, Michael, Besir, H., uuml, seyin, Essen, Lars-Oliver, & Oesterhelt,Dieter. 2000. Structure of the Light-Driven Chloride Pump Halorhodopsin at 1.8 AResolution. Science, 288(5470), 1390–1396.

Komrakov, Andrey Yu, & Kaulen, Andrey D. 1995. M-decay in the bacte-riorhodopsin photocycle: effect of cooperativity and pH. Biophys. Chem., 56(1-2),113.

Koonin, Eugene V., Wolf, Yuri I., & Karev, Georgy P. 2002. The structureof the protein universe and genome evolution. Nature, 420(6912), 218.

46

Introduction

Korenstein, Rafi, & Hess, Benno. 1978. Immobilization of bacteriorhodopsin andorientation of its transition moment in purple membrane. FEBS Lett., 89(1), 15.

Krebs, M.P., & Isenbarger, T.A. 2000. Structural determinants of purple mem-brane assembly. Biochim. Biophys. Acta, 1460, 15–26.

Kroeze, Wesley K., Sheffler, Douglas J., & Roth, Bryan L. 2003. G-protein-coupled receptors at a glance. J. Cell Sci., 116(24), 4867–4869.

Krogh, Anders, Larsson, Bjorn, von Heijne, Gunnar, & Sonnhammer, ErikL. L. 2001. Predicting transmembrane protein topology with a hidden markov model:application to complete genomes. J. Mol. Biol., 305(3), 567.

Kung, M. Chu, Devault, Don, Hess, Benno, & Oesterhelt, Dieter. 1975.Photolysis of Bacterial Rhodopsin. Biophys. J., 15, 907–911.

Lancaster, C. Roy D. 2003. The role of electrostatics in proton-conducting membraneprotein complexes. FEBS Lett., 545(1), 52.

Landau, Ehud M., & Rosenbusch, Jurg P. 1996. Lipidic cubic phases: A novelconcept for the crystallization of membrane proteins. Proc. Natl. Acad. Sci. USA,93(25), 14532–14535.

Lang, H. P., Hegner, M., Meyer, E., & Ch, Gerber. 2002. Nanomechanicsfrom atomic resolution to molecular recognition based on atomic force microscopytechnology. Nanotechnology, 13(5), R29.

Lanyi, Janos K. 1974. Salt-Dependent Properties of Proteins from ExtremelyHalophilic Bacteria. Bacteriol. Rev., 38(3), 272–290.

Lanyi, J.K. 1995. Bacteriorhodopsin as a model for proton pumps. Nature, 375,461–463.

Leckband, Deborah. 2001. Force as a probe of membrane protein structure andfunction. Curr. Op. Struct. Biol., 11(4), 433.

Lecomte, C., Guillot, B., Muzet, N., Pichon-Pesme, V., & Jelsch, C. 2004.Ultra-high-resolution X-ray structure of proteins. Cell. Molec. Life Sci., 61(8), 774–782.

Lee, Anthony G. 2003. Lipid-protein interactions in biological membranes: a struc-tural perspective. Biochim. Biophys. Acta, 1612, 1–40.

Lee, Anthony G. 2004. How lipids affect the activities of integral membrane proteins.Biochim. Biophys. Acta, 1666(1-2), 62.

47

Introduction

Lee, Gil U., Kidwell, David A., & Colton, Richard J. 1994. Sensing DiscreteStreptavidin-Biotin Interactions with Atomic Force Microscopy. Langmuir, 10(2),354–357.

Lee, Jason Wei-Ming, Chu, Justin Jang-Hann, & Ng, Mah-Lee. 2006. Quanti-fying the Specific Binding between West Nile Virus Envelope Domain III Protein andthe Cellular Receptor alphaVbeta3 Integrin. J. Biol. Chem., 281(3), 1352–1360.

Lozier, R. H., Xie, A., Hofrichter, J., & Clore, G. M. 1992. Reversible Stepsin the Bacteriorhodopsin Photocycle. Proc. Natl. Acad. Sci. USA, 89(8), 3610–3614.

Lozier, Richard H., A., Bogomolni Roberto, & Stoeckenius, Walther. 1975.Bacteriorhodopsin: A light-driven proton pump in Halobacterium halobium. Biophys.J., 15, 955–962.

Luca, S., Heise, H., & Baldus, M. 2003. High-Resolution Solid-State NMR Appliedto Polypeptides and Membrane Proteins. Acc. Chem. Res., 36(11), 858–865.

Luecke, H., Schobert, B., Richter, H.T., Cartailler, J.P., & Lanyi, J.K.1999a. Structure of Bacteriorhodopsin at 1.55 Aresolution. J. Mol. Biol., 291, 899–911. location of STGA.

Luecke, Hartmut, Richter, Hans-Thomas, & Lanyi, Janos K. 1998. ProtonTransfer Pathways in Bacteriorhodopsin at 2.3 AResolution. Science, 280(5371), 1934–1937.

Luecke, Hartmut, Schobert, Brigitte, Richter, Hans-Thomas, Car-tailler, Jean-Philippe, & Lanyi, Janos K. 1999b. Structural Changes in Bac-teriorhodopsin During Ion Transport at 2 Angstrom Resolution. Science, 286(5438),255–260.

Lunde, Christopher S., Rouhania, Shahab, Facciottib, Marc T., &Glaeser, Robert M. 2006. Membrane-protein stability in a phospholipid-basedcrystallization medium. J. Struct. Biol., 154(3), 223–231.

Lyubchenko, Yuri L., & Shlyakhtenko, Luda S. 1997. Visualization of super-coiled DNA with atomic force microscopy in situ. Proc. Natl. Acad. Sci. USA, 94(2),496–501.

Maeda, A., Verhoeven, M. A., Lugtenburg, J., Gennis, R. B., Balashov,S. P., & Ebrey, T. G. 2004. Water Rearrangement around the Schiff Base in theLate K (KL) Intermediate of the Bacteriorhodopsin Photocycle. J. Phys. Chem. B,108(3), 1096–1101.

Marszalek, Piotr E., Lu, Hui, Li, Hongbin, Carrion-Vazquez, Mariano,Oberhauser, Andres F., Schulten, Klaus, & Fernandez, Julio M. 1999.Mechanical unfolding intermediates in titin modules. Nature, 402(6757), 100.

48

Introduction

Mathies, R. A., Brito Cruz, C. H., Pollard, W. T., & Shank, C. V. 1988.Direct observation of the femtosecond excited-state cis-trans isomerization in bacteri-orhodopsin. Science, 240(4853), 777–779.

McAllister, Chad, Karymov, Mikhail A., Kawano, Yoshiko, Lush-nikov, Alexander Y., Mikheikin, Andrew, Uversky, Vladimir N., &Lyubchenko, Yuri L. 2005. Protein Interactions and Misfolding Analyzed by AFMForce Spectroscopy. J. Mol. Biol., 354(5), 1028.

Metpally, Raghu, & Sowdhamini, Ramanathan. 2005. Genome wide survey ofG protein-coupled receptors in Tetraodon nigroviridis. BMC Evol. Biol., 5(1), 41.

Milhiet, Pierre Emmanuel, Vie, Veronique, Giocondi, Marie-Cecile, &Le Grimellec, Christian. 2001. AFM Characterization of Model Rafts in Sup-ported Bilayers. Single Mol., 2(2), 109–112.

Mitsuoka, Kaoru, Hirai, Teruhisa, Murata, Kazuyoshi, Miyazawa, Atsuo,Kidera, Akinori, Kimura, Yoshiaki, & Fujiyoshi, Yoshinori. 1999. Thestructure of bacteriorhodopsin at 3.0 A resolution based on electron crystallography:implication of the charge distribution. J. Mol. Biol., 286(3), 861.

Mou, J., Sheng, S., Ho, R., & Shao, Z. 1996. Chaperonins GroEL and GroES:views from atomic force microscopy. Biophys. J., 71(4), 2213–2221.

Mou, Jianxun, Yang, Jie, & Shao, Zhifeng. 1995. Atomic Force Microscopy ofCholera Toxin B-oligomers Bound to Bilayers of Biologically Relevant Lipids. J. Mol.Biol., 248(3), 507.

Muller, Daniel J., Buldt, Georg, & Engel, Andreas. 1995. Force-inducedconformational change of bacteriorhodopsin. Journal of Molecular Biology, 249(2),239.

Muller, Daniel J., Fotiadis, Dimitrios, Scheuring, Simon, Mller,Shirley A., & Engel, Andreas. 1999. Electrostatically Balanced Subnanome-ter Imaging of Biological Specimens by Atomic Force Microscope. Biophys. J., 76(2),1101–1111.

Muller, Daniel J., Janovjak, Harald, Lehto, Tiina, Kuerschner, Lars,& Anderson, Kurt. 2002. Observing structure, function and assembly of singleproteins by AFM. Prog. Biophys. Molec. Biol., 79(1), 1–43.

Murphy, Lindy. 2006. Biosensors and bioelectrochemistry. Curr. Op. Chem. Biol.,10(2), 177.

Nayeem, Akbar, Sitkoff, Doree, & Krystek, Stanley, Jr. 2006. A compara-tive study of available software for high-accuracy homology modeling: From sequencealignments to structural models. Protein Sci, 15(4), 808–824.

49

Introduction

Neutze, Richard, Pebay-Peyroula, Eva, Edman, Karl, Royant, Antoine,Navarro, Javier, & Landau, Ehud M. 2002. Bacteriorhodopsin: a high-resolutionstructural view of vectorial proton transport. Biochim. Biophys. Acta, 1565(2), 144–167.

Nicolini, C., Erokhin, V., Paddeu, S., Paternolli, C., & Ram, M. K. 1999.Toward bacteriorhodopsin based photocells. Biosens Bioelec., 14(4), 427.

Oberhauser, Andres F., Marszalek, Piotr E., Erickson, Harold P., &Fernandez, Julio M. 1998. The molecular elasticity of the extracellular matrixprotein tenascin. Nature, 393(6681), 181.

Oesterhelt, D., & Stoeckenius, W. 1971. Rhodopsin-like protein from the purplemembrane of Halobacterium halobium. Nat. New Biol., 233(39), 149.

Oesterhelt, F., Oesterhelt, D., Pfeiffer, M., Engel, A., Gaub, H. E.,uuml, & ller, D. J. 2000. Unfolding Pathways of Individual Bacteriorhodopsins.Science, 288(5463), 143–146.

Oren, Aharon. 1994. The ecology of the extremely halophilic archaea. FEMS Micro-biol. Rev., 13(4), 415–439.

Oren, Aharon, Heldal, Mikal, Norland, Svein, & Galinski, Erwin. 2002. In-tracellular ion and organic solute concentrations of the extremely halophilic bacteriumSalinibacter ruber. Extremophiles, 6(6), 491.

Otto, H., Marti, T., Holz, M., Mogi, T., Stern, L. J., Engel, F., Khorana,H. G., & Heyn, M. P. 1990. Substitution of Amino Acids Asp-85, Asp-212, andArg-82 in Bacteriorhodopsin Affects the Proton Release Phase of the Pump and thepK of the Schiff Base. Proc. Natl. Acad. Sci. USA, 87(3), 1018–1022.

Otto, Harald, Marti, Thomas, Holz, Martin, Mogi, Tatsushi, Lindau,Manfred, Khorana, H. Gobind, & Heyn, Maarten P. 1989. Aspartic Acid-96 is the Internal Proton Donor in the Reprotonation of the Schiff Base of Bacteri-orhodopsin. Proc. Natl. Acad. Sci. USA, 86(23), 9228–9232.

Paige, Matthew F., Rainey, Jan K., & Goh, M. Cynthia. 1998. Fibrous LongSpacing Collagen Ultrastructure Elucidated by Atomic Force Microscopy. Biophys. J.,74(6), 3211–3216.

Palczewski, Krzysztof, Kumasaka, Takashi, Hori, Tetsuya, Behnke,Craig A., Motoshima, Hiroyuki, Fox, Brian A., Trong, Isolde Le,Teller, David C., Okada, Tetsuji, Stenkamp, Ronald E., Yamamoto,Masaki, & Miyano, Masashi. 2000. Crystal Structure of Rhodopsin: A G Protein-Coupled Receptor. Science, 289(5480), 739–745.

50

Introduction

Palsdottir, Hildur, & Hunte, Carola. 2004. Lipids in membrane protein struc-tures. Biochim. Biophys. Acta, 1666(1-2), 2.

Pardo, Leonardo, Sepulcre, Francesc, Cladera, Josep, Dunach, Mireia,Labarta, Amilcar, Tejada, Javier, & Padros, Esteve. 1998. Experimen-tal and Theoretical Characterization of the High-Affinity Cation-Binding Site of thePurple Membrane. Biophys. J., 75(2), 777–784.

Patzelt, Heiko, Ulrich, Anne S., Egbringhoff, Hermann, Dx, Petra,Ashurst, Jennifer, Simon, Bernd, Oschkinat, Hartmut, & Oesterhelt,Dieter. 1997. Towards structural investigations on isotope labelled native bacte-riorhodopsin in detergent micelles by solution-state NMR spectroscopy. J. Biomol.NMR, 10(2), 95–106.

Pebay-Peyroula, E., Rummel, G., Rosenbusch, J. P., & Landau, E. M. 1997.X-ray structure of bacteriorhodopsin at 2.5 Afrom microcrystals grown in lipidic cubicphases. Science, 277(5332), 1676–1681.

Petrache, H. I., Kimchi, I., Harries, D., & Parsegian, V. A. 2005. MeasuredDepletion of Ions at the Biomembrane Interface. J. Am. Chem. Soc., 127(33), 11546–11547.

Polland, H.-J., Franz, M. A., Zinth, W., Kaiser, W., Klling, E., & Oester-helt, D. 1986. Early picosecond events in the photocycle of bacteriorhodopsin. Bio-phys. J., 49, 651–662.

Radmacher, M., Tillamnn, R. W., Fritz, M., & Gaub, H. E. 1992. Frommolecules to cells: imaging soft samples with the atomic force microscope. Science,257(5078), 1900–1905.

Radmacher, M., Fritz, M., Hansma, H. G., & Hansma, P. K. 1994. Directobservation of enzyme activity with the atomic force microscope. Science, 265(5178),1577–1579.

Renthal, Richard, & Cha, C. H. 1984. Charge asymmetry of the purple membranemeasured by uranyl quenching of dansyl fluorescence. Biophys. J., 45(5), 1001–1006.

Reviakine, Ilya, Bergsma-Schutter, Wilma, & Brisson, Alain. 1998. Growthof Protein 2-D Crystals on Supported Planar Lipid Bilayers Imaged in Situ by AFM.J. Struct. Biol., 121(3), 356.

Rief, Matthias, Gautel, Mathias, Oesterhelt, Filipp, Fernandez,Julio M., & Gaub, Hermann E. 1997. Reversible Unfolding of Individual TitinImmunoglobulin Domains by AFM. Science, 276(5315), 1109–1112.

51

Introduction

Rief, Matthias, Gautel, Mathias, Schemmel, Alexander, & Gaub, Her-mann E. 1998. The Mechanical Stability of Immunoglobulin and Fibronectin IIIDomains in the Muscle Protein Titin Measured by Atomic Force Microscopy. Biophys.J., 75(6), 3008–3014.

Rief, Matthias, Pascual, Jaime, Saraste, Matti, & Gaub, Hermann E. 1999.Single molecule force spectroscopy of spectrin repeats: low unfolding forces in helixbundles. J. Mol. Biol., 286(2), 553.

Rinia, Hilde A., Snel, Margot M. E., van der Eerden, Jan P. J. M., &de Kruijff, Ben. 2001. Visualizing detergent resistant domains in model membraneswith atomic force microscopy. FEBS Lett., 501(1), 92.

Royant, Antoine, Edman, Karl, Ursby, Thomas, Pabay-Peyroula, Eva,Landau, Ehud M., & Neutze, Richard. 2000. Helix deformation is coupledto vectorial proton transport in the photocycle of bacteriorhodopsin. Nature, 406,645–648.

Saito, H., Yamamoto, K., Tuzi, S., & Yamaguchi, S. 2003. Backbone dynamicsof membrane proteins in lipid bilayers: the effect of two-dimensional array formation asrevealed by site-directed solid-state 13C NMR studies on [3-13C]Ala- and [1-13C]Val-labeled bacteriorhodopsin. Biochim. Biophys. Acta, 1616(2), 127–136.

San Paulo, Alvaro, & Garcıa, Ricardo. 2000. High-Resolution Imaging of An-tibodies by Tapping-Mode Atomic Force Microscopy: Attractive and Repulsive Tip-Sample Interaction Regimes. Biophys. J., 78(3), 1599–1605.

Sanz, Carolina, Marquez, Mercedes, Peralvarez, Alex, Elouatik, Samir,Sepulcre, Francesc, Querol, Enric, Lazarova, Tzvetana, & Padros, Es-teve. 2001. Contribution of Extracellular Glu Residues to the Structure and Func-tion of Bacteriorhodopsin. PRESENCE OF SPECIFIC CATION-BINDING SITES.J. Biol. Chem., 276(44), 40788–40794.

Sasaki, J., Brown, L. S., Chon, Y. S., Kandori, H., Maeda, A., Needleman,R., & Lanyi, J. K. 1995. Conversion of bacteriorhodopsin into a chloride ion pump.Science, 269(5220), 73–75.

Sass, H. J., Schachowa, I. W., Rapp, G., Koch, M. H. J., Oesterhelt, D.,Dencher, N. A., & Buldt, G. 1997. The tertiary structural changes in bacteri-orhodopsin occur between M states: X-ray diffraction and Fourier transform infraredspectroscopy. EMBO J., 16(7), 1484–1491.

Sato, Hidenori, Takeda, Kazuki, Tani, Koji, Hino, Tomoya, Okada, Tet-suji, Nakasako, Masayoshi, Kamiya, Nobuo, & Kouyama, Tsutomu. 1999.Specific lipid-protein interactions in a novel honeycomb lattice structure of bacteri-orhodopsin. Acta Crystallogr. D, 55(7), 1251–1256.

52

Introduction

Sattin, Bernie D., Pelling, Andrew E., & Goh, M. Cynthia. 2004. DNAbase pair resolution by single molecule force spectroscopy. Nucl. Acids Res., 32(16),4876–4883.

Scheuring, Simon, Levy, Daniel, & Rigaud, Jean-Louis. 2005. Watching thecomponents of photosynthetic bacterial membranes and their in situ organisation byatomic force microscopy. Biochim. Biophys. Acta, 1712(2), 109–127.

Schneider, Stefan W., Sritharan, Kumudesh C., Geibel, John P., Oberlei-thner, Hans, & Jena, Bhanu P. 1997. Surface dynamics in living acinar cellsimaged by atomic force microscopy: Identification of plasma membrane structuresinvolved in exocytosis. Proc. Natl. Acad. Sci. USA, 94(1), 316–321.

Schobert, Brigitte, Cupp-Vickery, Jill, Hornak, Viktor, Smith,Steven O., & Lanyi, Janos K. 2002. Crystallographic Structure of the K Inter-mediate of Bacteriorhodopsin: Conservation of Free Energy after Photoisomerizationof the Retinal. J. Mol. Biol., 321(4), 715.

Schobert, Brigitte, Brown, Leonid S., & Lanyi, Janos K. 2003. Crystallo-graphic Structures of the M and N Intermediates of Bacteriorhodopsin: Assembly ofa Hydrogen-bonded Chain of Water Molecules Between Asp-96 and the Retinal SchiffBase. J. Mol. Biol., 330(3), 553.

Schotte, Friedrich, Lim, Manho, Jackson, Timothy A., Smirnov, Alek-sandr V., Soman, Jayashree, Olson, John S., Phillips, George N., Jr.,Wulff, Michael, & Anfinrud, Philip A. 2003. Watching a Protein as it Func-tions with 150-ps Time-Resolved X-ray Crystallography. Science, 300(5627), 1944–1947.

Sergi, Mauro, Zurawski, John, Cocklin, Simon, & Chaiken, Irwin. 2004. Pro-teins, recognition networks and developing interfaces for macromolecular biosensing.J. Mol. Recognit., 17(3), 198–208. 10.1002/jmr.671.

Sherman, Warren V., Slifkin, Michael A., & Caplan, S. Roy. 1976. Kineticstudies of phototransients in bacteriorhodopsin. Biochim. Biophys. Acta, 423(2), 238.

Slade, Andrea, Luh, Jeanne, Ho, Sylvia, & Yip, Christopher M. 2002. Sin-gle molecule imaging of supported planar lipid bilayer–reconstituted human insulinreceptors by in situ scanning probe microscopy. J. Struct. Biol., 137(3), 283.

Sternberg, Brigitte, L’Hostis, Chantal, Whiteway, Clare A., & Watts,Anthony. 1992. The essential role of specific Halobacterium halobium polar lipids in2D-array formation of bacteriorhodopsin. Biochim. Biophys. Acta, 1108(1), 21–30.

Stoeckenius, Walther, Lozier, Richard H., & Bogomolni, Roberto A. 1979.Bacteriorhodopsin and the purple membrane of halobacteria. Biochim. Biophys. Acta,505(3-4), 215.

53

Introduction

Stolz, Martin, Stoffler, Daniel, Aebi, Ueli, & Goldsbury, Claire. 2000.Monitoring Biomolecular Interactions by Time-Lapse Atomic Force Microscopy. J.Struct. Biol., 131(3), 171.

Subramaniam, Sriram, Lindahl, Martin, Bullough, Per, Faruqi, A. R., Tit-tor, Jorg, Oesterhelt, Dieter, Brown, Leonid, Lanyi, Janos, & Hen-derson, Richard. 1999. Protein conformational changes in the bacteriorhodopsinphotocycle. J. Mol. Biol., 287(1), 145.

Tamayo, J., Saenz, J. J., & Garcıa, R. 1995. Scanning tunneling microscopymodification of purple membranes. J. Vac. Sci Technol A, 13, 1737–1741.

Tamayo, Javier, & Garcia, Ricardo. 1998. Relationship between phase shift andenergy dissipation in tapping-mode scanning force microscopy. Appl. Phys. Lett.,73(20), 2926–2928.

Tan, E. H., & Birge, R. R. 1996. Correlation between surfactant/micelle structureand the stability of bacteriorhodopsin in solution. Biophys. J., 70(5), 2385–2395.

Tang, Qian, Shi, San-Qiang, & Zhou, Limin. 2004. Nanofabrication with AtomicForce Microscopy. J. Nanosci. Technol., 4(8), 948–963.

Teissie, J., Prats, M., LeMassu, A., Stewart, L. C., & Kates, M. 1990.Lateral proton conduction in monolayers of phospholipids from extreme halophiles.Biochemistry, 29(1), 59.

Tokaji, Zsolt. 1998. Quantitative model for the cooperative interaction of the bacte-riorhodopsin molecules in purple membranes. FEBS Lett., 423(3), 343.

Tokumasu, Fuyuki, Jin, Albert J., Feigenson, Gerald W., & Dvorak,James A. 2003. Nanoscopic Lipid Domain Dynamics Revealed by Atomic ForceMicroscopy. Biophys. J., 84(4), 2609–2618.

Torres, Jaume, Stevens, Tim J., & Samso, Montserrat. 2003. Membraneproteins: the ’Wild West’ of structural biology. Trends Biochem. Sci., 28(3), 137.

Ulmschneider, M. B., Tieleman, D. P., & Sansom, M. S. P. 2004. Interactionsof a Transmembrane Helix and a Membrane: Comparative Simulations of Bacteri-orhodopsin Helix A. J. Phys. Chem. B, 108(28), 10149–10159.

Varo, G., Needleman, R., & Lanyi, J. K. 1996. Protein structural change at thecytoplasmic surface as the cause of cooperativity in the bacteriorhodopsin photocycle.Biophys. J., 70(1), 461–467.

Varo, Gyorgy, Brown, Leonid S., Needleman, Richard, & Lanyi, Janos K.1999. Binding of Calcium Ions to Bacteriorhodopsin. Biophys. J., 76(6), 3219–3226.

54

Introduction

Vonck, Janet. 2000. Structure of the bacteriorhodopsin mutant F219L N intermediaterevealed by electron crystallography. EMBO J., 19(10), 2152–2160.

Wallin, E., & Von-Heijne, G. 1998. Genome-wide analysis of integral membraneproteins from eubacterial, archaean, and eukaryotic organisms. Protein Sci., 7(4),1029–1038.

Watts, Anthony. 1995. Bacteriorhodopsin: the mechanism of 2D-array formation andthe structure of retinal in the protein. Biophys. Chem., 55(1-2), 137.

Weik, Martin, Patzelt, Heiko, Zaccai, Giuseppe, & Oesterhelt, Dieter.1998. Localization of Glycolipids in Membranes by In Vivo Labeling and NeutronDiffraction. Mol. Cell, 1(3), 411–419.

White, Stephen H. 2004. The progress of membrane protein structure determination.Protein Sci., 13(7), 1948–1949.

White, Stephen H. 2006. Membrane proteins of Known 3D structure.http://blanco.biomol.uci.edu/Membrane Proteins xtal.html.

White, Stephen H., & Wimley, William C. 1998. Hydrophobic interactions ofpeptides with membrane interfaces. Biochim. Biophys. Acta, 1376(3), 339.

White, Stephen H., Ladokhin, Alexey S., Jayasinghe, Sajith, & Hristova,Kalina. 2001. How Membranes Shape Protein Structure. J. Biol. Chem., 276(35),32395–32398.

Williams, Philip M., Fowler, Susan B., Best, Robert B., Luis Toca-Herrera, Jose, Scott, Kathryn A., Steward, Annette, & Clarke, Jane.2003. Hidden complexity in the mechanical properties of titin. Nature, 422(6930),446.

www.pdb.org. 2006. RCSB Protein Data Bank.

Xie, A. H., Nagle, J. F., & Lozier, R. H. 1987. Flash spectroscopy of purplemembrane. Biophys. J., 51(4), 627–635.

Yonebayashi, Koka, Yamaguchi, Satoru, Tuzi, Satoru, & Sait, Hazime.2004. Cytoplasmic surface structures of bacteriorhodopsin modified by site-directedmutations and cation binding as revealed by 13C NMR. Eur. Biophys. J., 32(1),1432–1017.

Young, Paul M., Price, Robert, Tobyn, Michael J., Buttrum, Mark, &Dey, Fiona. 2003. Investigation into the effect of humidity on drug-Drug interactionsusing the atomic force microscope. J. Pharma. Sci., 92(4), 815–822. 10.1002/jps.10250.

55

Introduction

Zaccai, Giuseppe. 2000. How Soft Is a Protein? A Protein Dynamics Force ConstantMeasured by Neutron Scattering. Science, 288(5471), 1604–1607.

Zhang, Yang, & Skolnick, Jeffrey. 2005. The protein structure prediction problemcould be solved using the current PDB library. Proc. Natl. Acad. Sci. USA, 102(4),1029–1034.

Zscherp, C., & Heberle, J. 1997. Infrared Difference Spectra of the IntermediatesL, M, N, and O of the Bacteriorhodopsin Photoreaction Obtained by Time-ResolvedAttenuated Total Reflection Spectroscopy. J. Phys. Chem. B, 101(49), 10542–10547.

Zscherp, Christian, Schlesinger, Ramona, Tittor, Jorg, Oesterhelt, Di-eter, & Heberle, Joachim. 1999. In situ determination of transient pKa changesof internal amino acids of bacteriorhodopsin by using time-resolved attenuated totalreflection Fourier-transform infrared spectroscopy. Proc. Natl. Acad. Sci. USA, 96(10),5498–5503.

56

Atomic Force Microscopy of Membrane Proteins

2 Atomic Force Microscopy of Membrane Proteins:

A short Review

This chapter reviews some of the most successful results obtained by previous AFM

studies on integral membrane proteins. In the vast majority of cases, however, AFM is

used as a diagnostic tool and its contribution is limited to high-resolution imaging and

nano-dissection.

2.1 Bacterial and Mammalian Rhodopsins

2.1.1 Bacteriorhodopsin

By far the most studied system, bacteriorhodopsin (bR) from Purple Membranes (PM)

was the first membrane protein to be imaged with AFM (Worcester et al. 1988, 1990)

in air and in liquid (Butt et al. 1990, 1991), and the well known 6.2 nm constant of the

protein lattice (Henderson & Unwin 1975) could be measured. With the improvement of

AFMs, ultra-high resolution (< 0.5 nm) of the protein lattice surface could be obtained

(Muller et al. 1995a, 1999b; Moller et al. 1999), opening the way to thorough studies

of the system under native like conditions. Procedures averaging and symmetrizing the

trimers were however employed to enhance the resolution, thus removing single molecule

information. Highest resolution was obtained using contact mode AFM, and Muller et

al. were able to resolve both sides of PM, and the side assessment could be confirmed

scanning a mutant membrane with a functionalized tip (Muller et al. 1996). PM cyto-

plasmic side shows distinct trimers that protrude up to 0.64 nm above the membrane.

The main protrusion was attributed to the peptide loop connecting the transmembrane

helices E and F. Scanning the membrane with an excessive tip pressure induced reversible

conformational changes of the protein, mainly due to lateral flexion of the peptide loop

revealing other inter-helical loops (Muller et al. 1995b). The extracellular side is flatter

57


and only bR BC inter-helical loop, protruding up to 0.5 nm out of the membrane, could

be identified (Muller et al. 2000). The achievement of high resolution imaging was ex-

plained with the so-called DLVO theory (Derjaguin-Landau-Verwey-Overbeek, see chap.

4) and therefore a specific ionic concentration was required in order to balance van der

Waals attractive forces with Electrostatic Double Layer repulsive forces (Muller & Engel

1997; Muller et al. 1999a).

Single molecule force spectroscopy on the bR cytoplasmic terminal (Oesterhelt et al.

2000) revealed distinctive unfolding pathways, with certain α-helices unfolding preferen-

tially by pairs. Further work has studied the influence of pH (Muller et al. 2002b) and

temperature (Janovjak et al. 2003) on bR unfolding patterns. Extracellular loops have

been shown to play an important role in stabilizing the protein and it was suggested

that α-helices tend to unfold before their extraction from the membrane when stretched

over a potential barrier (Janovjak et al. 2004). In order to distinguish intra-molecular

interactions from inter-molecular forces anchoring the protein in the membrane, single

molecule force spectroscopy was carried out on orthorhombic purple membranes lattices,

obtained when re-crystallizing solubilized bR in presence of n-dodecyl trimethylammo-

nium chloride (DTAC) (Michel et al. 1980; Muller et al. 1999b), expecting modified

inter-molecular forces between bR molecules (Sapra et al. 2006). No clear contribution

could be identified suggesting that forces responsible for the unfolding patterns observed

do not depend on the trimeric arrangement of bR in native membrane. Several studies

have used dynamic force spectroscopy (Janovjak et al. 2005) and refolding of bR against

an external force (Kessler et al. 2006) to estimate the potential barriers observed. Fi-

nally, a side-specific force spectroscopy study showing different unfolding patterns of bR

when pulled from the protein cytoplasmic and extracellular termini respectively (Kessler

& Gaub 2006) confirmed the assessment of the barriers being mostly in the inter-helical

loops regions.

58


2.1.2 Halorhodopsin

Halorhodopsin (hR) belongs to the family of archeal rhodopsin, like bR and the sen-

sory rhodopsins (sRI and sRII) responsible for the archaea phototaxis. hR is a seven

transmembrane helices protein and functions as an inward light-driven chloride pump,

the only anion pump known to date (Kolbe et al. 2000). hR shows many similarities

to bR; both use a retinal linked to the G α-helix as chromophore, both undergo a well

charcterized pumping photocycle (Haupts et al. 1997) and they share an important part

of their respective amino acid sequence (Blanck & Oesterhelt 1987; Cisneros et al. 2005).

When over-expressed in Halobacterium salinarium, hR assembles in a 2D crystal with

P4212 symmetry (hR tetramers) (Havelka et al. 1993). AFM in native-like solution

provides sub-molecular resolution of the hR lattice (Persike et al. 2001; Cisneros et al.

2005) and can distinguish between the two sides of the crystal (whether the cytoplasmic

or extracellular side of hR is exposed to the solution). Single molecule spectroscopy

revealed unfolding patterns almost identical to that of bR except for some potential bar-

rier located in a short segment of helix E. (Cisneros et al. 2005). This result suggested

that non-identical aa sequences can generate identical potential barriers, ruling out the

necessity of a specific aa sequence. The molecular origin of the barriers could however

not be identified.

2.1.3 Bovine Rhodopsin

To date, the only GPCR atomic structure available is bovine rhodopsin (Palczewski

et al. 2000). This protein is naturally found in the rod outer-segment disc membrane

of vertebrate. When illuminated with light, a single photon can activate a rhodopsin

(Rho) which will in turn trigger an amplification cascade resulting in a nerve signal

(Stryer 1987). AFM of native rod outer-segment disc membrane demonstrated that Rho

are naturally arranged in dimers (Fotiadis et al. 2003), thus optimizing the surface

59


occupation of the disc membrane by photo-receptors. This controversial result (Chabre

et al. 2003) was further confirmed by AFM and Electron Microscopy (EM) on mouse disc

membranes (Liang et al. 2003; Fotiadis et al. 2004a) which provided a higher dimeric

organization level. Recently, single molecule force spectroscopy experiments on Rho

have provided some insight in the residues responsible for the protein stability (Sapra

et al. 2006; Fotiadis et al. 2006). Segments showing highly conserved residues among

GPCRs showed an increased mechanical stability, in particular the role of Cys110-Cys187

disulfide bond has been highlighted.

2.2 Membrane channels

2.2.1 Ion transporters and channels

Sodium-proton Antiporter

The sodium-proton antiporter of Escherichia coli(NhaH) is composed of 12 transmem-

brane α-helices (Williams 2000) and plays an important role in human cardiac disease.

When reconstituted into 2D crystals, high resolution (∼ 0.6 nm) AFM imaging revealed

dimeric NhaH molecules that assemble in regular rows (Kedrov et al. 2004). Single

molecule force spectroscopy experiments exhibit two types of curves that correspond to

unfolding NhaH from the N- and the C-terminal, as confirmed with trypsin cleaved NhaH

(Kedrov et al. 2004). Like for bR, secondary elements were reported to constitute the

main potential unfolding barriers with certain helices unfolding pairwise. The protein

could fully refold into the membrane after 10 of the 12 helices had been unfolded. Two

specific helices constantly showed a ”snap in” behaviour and proline residues were sug-

gested to play an important role in the folding kinetics (Kedrov et al. 2004). Using an

AFM with improved feedback, the experiment could be repeated controlling at the ms

level the time necessary for the protein to refold. Results provided some information

60


about the folding kinetics assessing different folding rates from 0.31s−1 to up to 47s−1

for the different structural segments (Kedrov et al. 2006).

Mucolipin-1 channel

Mucolipidis type IV (MLIV) is a neurogenic disorder causing neurological, ophthalmo-

logic and gastric dysfunction. Experiments have established that MLIV is caused by the

mutation of a gene encoding mucolipin-1 (ML1) a 65kDa protein of unknown function

(Sun et al. 2000). Channel recording experiments demonstrated that wild-type ML1 is

a pH gated non-selective cation channel forming multiple pores in the membrane (Ray-

chowdhury et al. 2004). Complementary AFM imaging of WT-ML1 reconstituted into

a lipid bilayer showed changes in size of the protein a low pH, suggesting channel closure,

and disassembly of the WT-ML1 complexes. The mutant ML-1 showed less sensitivity

to pH variations. Further experiments showed that unlike WT-ML1, the mutant protein

is not inhibited by Ca2+ transport (Cantiello et al. 2005). These observations provided

essential insight in the molecular basis of MLIV.

2.2.2 Porins

Outer Membrane Protein F Porin

Outer Membrane Protein F Porin (OmpF) is a trimeric channel protein from the outer

membrane of bacteria such as Escherichia coli and Rhodobacter capsulatus. A OmpF

monomer is composed of 340 aa forming 16 antiparallel β-strands folded into a hol-

low transmembrane cylindrical β-barrel which allows the passage of ions and soluble

molecules smaller than ∼ 600kDa. AFM imaging of OmpF reconstituted in a phos-

pholipid bilayer showed that the trimer lattice could assemble hexagonally (Schabert &

Engel 1994) and that the loops connecting adjacent β-strands protruded substantially

more out of the extracellular surface (1.3±0.2 nm above the lipid surface) than out of the

61


periplasmic surface (Schabert et al. 1995). An important flexibility could be measured

for the extracellular loops, suggesting that they play a role in the conformational changes

related to the channel activity. OmpF is a voltage gated channel which conformational

changes could be directly imaged by applying a voltage to a platinum electrode located

close to an AFM scanning tip (Muller et al. 1999b).

Aquaporins

The Aquaporins (AQPs) are a family of tetrameric transmembrane channel proteins

transporting water and, in some cases, other small solutes across biological membranes.

They are highly selective and can allow a large number of water molecules per second

through the membrane (up to 109) (Pohl et al. 2001).

Reconstituted in phospholids membranes, Aquaporin 0 (Major Intrinsic Protein), 1, 2

and Z could be imaged and nano-dissected by AFM with sub-molecular resolution (Fo-

tiadis et al. 2000, 2002; Schenk et al. 2005; Scheuring et al. 1999), providing structural

indications about their known respective properties such as lens fiber adhesion (AQP0),

Ca2+ binding (AQP1), and stiffness of the most prominent peptide loop (AQP Z). Single

molecule force spectroscopy of AQP1 (Moller et al. 2003) related the major potential

barriers observed in the unfolding curves to individual secondary structure elements and

conservative residues.

2.2.3 Gap junctions

Gap junctions are inter-cellular channels that allow direct transfer of small molecules

(< 1kDa) between two adjacent vertebrate cells. The channel is composed of two hex-

americ hemi-channels (connexon) coaxially aligned, each cell providing one half of the

channel. Nano-dissection and high resolution AFM imaging of connexons assemblies in

62


the cell membrane (Hoh et al. 1991; Muller et al. 2002a) allowed direct observation

of the channel conformational changes and suggested the flexibility of the connexons

cytoplasmic domains to be involved in the gating properties of the channel.

2.3 Photosynthesis related systems

2.3.1 The Light Harvesting complexes 1 and 2

In photosynthetic bacteria, algae and in plants, the initial steps of photosynthesis involve

light absorption by the light-harvesting photo-sensitive complexes (LH) with subsequent

transmission of the absorbed energy to the reaction center (RC) where charge separa-

tion across the membrane occurs. In photosynthetic bacteria, these two initial steps are

achieved by the so-called core complex featuring intimate association of LH1 with RC.

LH2 are an antennae complex which transmits their photo-absorbed energy further to

a LH1. AFM investigation and high-resolution imaging of several reconstituted photo-

synthetic complexes (Fotiadis et al. 1998, 2004b; Scheuring et al. 2003, 2001, 2004a,b)

provided insight in the spatial organization of the different complexes in the membrane.

The different subunits could be investigated by nano-dissection (Scheuring et al. 2003),

and different variable stoichiometries observed (Scheuring et al. 2004a; Barz et al. 1995;

Scheuring et al. 2005). High-resolution AFM of reconstituted LH2 complex provided

some basis to estimate the dipole strength for energy transfer between two neighboring

LH2 rings (Stamouli et al. 2003).

2.4 Molecular motors

2.4.1 F1F0 bacterial and Spinach ATP-synthase

The F1F0 ATP-synthase is an enzymatic rotary motor that uses the transmembrane

ionic gradient (H+ or Na+) to synthesize adenosine triphosphate (ATP) (Dimroth et al.

63


2006). High-resolution X-ray structure of F1 catalytic sites suggests that the motor ef-

ficiency (ATP/ion) directly depends on the number of rotor subunits. High-resolution

AFM images of the Ilyobacter tartaricus Na+-driven rotor reconstituted in lipid bilayer

revealed 11 c-subunits (Stahlberg et al. 2001). When reconstituted in a lipid bilayer,

F0 complex sometimes exhibits structural gaps with missing subunits, regardless of the

rotor’s packing density (Meier et al. 2005). The diameter of the rotor, however, remains

constant independently of the number of subunits missing (Muller et al. 2001). This

finding suggests that the rotor diameter and stoichiometry is determined by the shape of

the subunits and their neighboring interactions. The same result was also demonstrated

for spinach chloroplast ATP-synthase (Seelert et al. 2000) exhibiting 14 subunits. Re-

cently, AFM topographs of reconstituted Spirulina platensis ATP-synthase F0 complex

revealed a pentadecameric rotor (15 c-subunits), the largest c ring observed so far. In-

terestingly, with 15 c-subunits, this rotor does not show the classical F1/F0 mismatch to

the three-fold symmetry of the F1 domain (Pogoryelov et al. 2005).

2.4.2 φ-29 rotary motor

The bacteriophage φ-29 of the Bacillus subtilis is a virus composed of an icosahedral

prohead and a rotary head-tail connector that hydrolyzes ATP in order to pack newly

replicated genome in the φ-29 prohead (Smith et al. 2001). AFM of the bacteriophage

head-tail connector crystals resolved the 12 subunits of the wide connector domain and

showed a right-handed velocity (Muller et al. 1997). The other side of the connector

appears to be a flexible channel with a conical shape.

64


2.5 Toxins

2.5.1 Membrane associating toxins

The cholera toxin

The cholera toxin, secreted by Vibrio cholerae is composed of two subunits A and B,

with the A-subunit catalytically active and the B-subunit binding to a membrane re-

ceptor (a ganglioside, GM1) naturally hosted in the membrane of the attacked cell. The

insertion of the cholera toxin into a synthetic phospholipid bilayer containing ganglioside

(GM1) could be directly observed by AFM (Yang et al. 1993) and the B-subunit im-

aged with sub-molecular resolution (Yang et al. 1993; Mou et al. 1995). These studies

largely contributed in opening the field of membrane proteins to AFM by demonstrating

the feasibility of AFM studies on proteins reconstituted in model membranes. Recently,

single molecule force spectroscopy has been used to characterize the binding potential of

the toxin B-pentamer with the GM1 ganglioside membrane receptor (Cai & Yang 2003).

Cry endotoxins from Bacillus thuringiensis

The Cry endotoxins from Bacillus thuringiensis can cause death to a large variety of

insect larvae by inducing pores the host midgut epithelial cell membrane. Using Fourier

transform infrared spectroscopy (FTIR) combined with AFM, it was possible to demon-

strate that Cry1Aa spontaneously inserts into lipid mono- and bilayers of relevant com-

position (Vie et al. 2001). AFM imaging of the pores indicates that they are formed by

four subunits enclosing a 1.5 nm diameter central depression.

The Cry4Ba toxin causes larvae death for the Aedes and Anophles mosquitoes which

are involved in the transmission of dengue and malaria viruses. AC-AFM imaging of the

Cry4Ba under different conditions showed the toxin to have an overall negative charge

at a pH close to that of the larval midgut where it is normally activated (Puntheeranurak

65


et al. 2005). The images shown that the toxin was sensitive to hydration forces and

could easily undergo important conformational changes, in order to spontaneously inserts

in receptor-free lipid bilayers where it aggregate preferentially in tetrameric associations.

2.5.2 Other toxins

α-hemolysin

Staphylococcal α-hemolysin (α-HL) is a water soluble bacterial toxin that can spon-

taneously bind to the plasma membrane and assemble, with other monomers, into a

pore-forming homo-oligomer. The created pore is water filled and can induce the cell

death trough leakage of small soluble molecules and ions. X-ray crystallography of the

α-HL demonstrated that the oligomeric membrane-bound toxin assembly is a heptamer

(Song et al. 1996), ruling out previous electron microscopy data that suggested a hex-

amer (Ward & Leonard 1992). Using high resolution AFM, it was possible to image

in liquid, the oligomeric complex in supported phospholipid bilayer (Czajkowsky et al.

1998). The images revealed a distinct hexameric pore, contrarily to X-ray data, suggest-

ing that different oligomeric stable stoichiometry are possible. Incomplete forming pores

could also be resolved, providing insight into the pore formation mechanism.

Pefringolysin O

Perfringolysin O (PFO) is a bacterial toxin belonging to the large family of pore-forming

cholesterol-dependent cytolysins (CDCs). Each PFO monomer is composed of four do-

mains, with one containing two β-hairpins that can insert into the membrane to form

a transmembrane β-barrel pore (Rossjohn et al. 1997). A enigma was surrounding the

fact that the β-hairpins seemed to be located too far above the membrane to cross the

bilayer and form the pore. Using AFM, it was possible to demonstrate that the pre-

pore complex protrude substantially more from the membrane than the pore complexes

66


(Czajkowsky et al. 2004). Time-lapse AFM experiments combined with high resolution

imaging showed in real time a synchronized vertical collapse of the PFO, allowing their

β-hairpins to cross the membrane and form the β-barrel pore.

2.6 References

Barz, Wolfgang P., Andr, Vermglio., Francia, Francesco, Venturoli,Giovanni, Melandri, B. Andrea, & Oesterhelt, Dieter. 1995. Role of thePufX Protein in Photosynthetic Growth of Rhudubacter sphaeroides. 2. PufX Is Re-quired for Efficient Ubiquinone/Ubiquinol Exchange between the Reaction Center QBSite and the Cytochrome bc1 Complex. Biochemistry, 34(46), 15248–15258.

Blanck, A., & Oesterhelt, Dieter. 1987. The Halo-Opsin Gene. II. Sequence,Primary Structure of Halorhodopsin and Comparison with Bacteriorhodopsin. EMBOJ., 6, 265–273.

Butt, H. J., Downing, K. H., & Hansma, P. K. 1990. Imaging the membraneprotein bacteriorhodopsin with the atomic force microscope. Biophys. J., 58(6), 1473–1480.

Butt, Hans-Jurgen, Prater, C. B., & Hansma, Paul K. 1991. Imaging pur-ple membranes dry and in water with the atomic force microscope. Page 1193of: Proceedings of the Fifth International Conference on Scanning Tunneling Mi-croscopy/Spectroscopy, vol. 9. Baltimore, Massachusetts (USA): AVS.

Cai, X. E., & Yang, J. 2003. The Binding Potential between the Cholera ToxinB-Oligomer and Its Receptor. Biochemistry, 42(14), 4028.

Cantiello, Horacio, Montalbetti, Nicolas, Goldmann, Wolfgang, Ray-chowdhury, Malay, Gonzalez-Perrett, Silvia, Timpanaro, Gustavo, &Chasan, Bernard. 2005. Cation channel activity of mucolipin-1: the effect of cal-cium. Pflugers Arch. Eur. J. Physiol., 451(1), 304.

Chabre, Marc, Cone, Richard, & Saibil, Helen. 2003. Biophysics (commu-nication arising): Is rhodopsin dimeric in native retinal rods? Nature, 426(6962),30.

Cisneros, David A., Oesterhelt, Dieter, & Muller, Daniel J. 2005. ProbingOrigins of Molecular Interactions Stabilizing the Membrane Proteins Halorhodopsinand Bacteriorhodopsin. Structure, 13(2), 235.

Czajkowsky, Daniel M., Sheng, Sitong, & Shao, Zhifeng. 1998. Staphylococcala-hemolysin can form hexamers in phospholipid bilayers. J. Mol. Biol., 276(2), 325.

67


Czajkowsky, Daniel M., Hotze, Eileen M., Shao, Zhifeng, & Tweten, Rod-ney K. 2004. Vertical collapse of a cytolysin prepore moves its transmembrane b-hairpins to the membrane. EMBO J., 23, 3206–3215.

Dimroth, Peter, von Ballmoos, Christoph, & Meier, Thomas. 2006. Catalyticand mechanical cycles in F-ATP synthases: Fourth in the Cycles Review Series. EMBORep., 7(3), 276–282.

Fotiadis, Dimitrios, Muller, Daniel J., Tsiotis, Georgios, Hasler, Lorenz,Tittmann, Peter, Mini, Thierry, Jeno, Paul, Gross, Heinz, & Engel,Andreas. 1998. Surface analysis of the photosystem I complex by electron and atomicforce microscopy. J. Mol. Biol., 283(1), 83.

Fotiadis, Dimitrios, Hasler, Lorenz, Muller, Daniel J., Stahlberg, Hen-ning, Kistler, Jorg, & Engel, Andreas. 2000. Surface Tongue-and-grooveContours on Lens MIP Facilitate Cell-to-cell Adherence. J. Mol. Biol., 300(4), 779.

Fotiadis, Dimitrios, Suda, Kitaru, Tittmann, Peter, Jeno, Paul,Philippsen, Ansgar, Muller, Daniel J., Gross, Heinz, & Engel, Andreas.2002. Identification and Structure of a Putative Ca2+-binding Domain at the C Ter-minus of AQP1. J. Mol. Biol., 318(5), 1381.

Fotiadis, Dimitrios, Liang, Yan, Filipek, Slawomir, Saperstein, David A.,Engel, Andreas, & Palczewski, Krzysztof. 2003. Rhodopsin dimers in nativedisc membranes. Nature, 421(6919), 127.

Fotiadis, Dimitrios, Liang, Yan, Filipek, Slawomir, Saperstein, David A.,Engel, Andreas, & Palczewski, Krzysztof. 2004a. The G protein-coupledreceptor rhodopsin in the native membrane. FEBS Lett., 564(3), 281.

Fotiadis, Dimitrios, Qian, Pu, Philippsen, Ansgar, Bullough, Per A., En-gel, Andreas, & Hunter, C. Neil. 2004b. Structural Analysis of the ReactionCenter Light-harvesting Complex I Photosynthetic Core Complex of Rhodospirillumrubrum Using Atomic Force Microscopy. J. Biol. Chem., 279(3), 2063–2068.

Fotiadis, Dimitrios, Jastrzebska, Beata, Philippsen, Ansgar, Muller,Daniel J., Palczewski, Krzysztof, & Engel, Andreas. 2006. Structure ofthe rhodopsin dimer: a working model for G-protein-coupled receptors. Curr. Op.Struct. Biol., 16(2), 252.

Haupts, U., Tittor, J., Bamberg, E., & Oesterhelt, D. 1997. GeneralConcept for Ion Translocation by Halobacterial Retinal Proteins: The Isomeriza-tion/Switch/Transfer (IST) Model. Biochemistry, 36(1), 2–7.

68


Havelka, Wendy A., Henderson, Richard, Heymann, Jrgen A. W., &Oesterhelt, Dieter. 1993. Projection Structure of Halorhodopsin from Halobac-terium halobium at 6 Resolution Obtained by Electron Cryo-microscopy. J. Mol.Bio., 234, 837–846.

Henderson, R., & Unwin, P. N. T. 1975. Three-dimensional model of purple mem-brane obtained by electron microscopy. Nature, 257(5521), 28.

Hoh, J. H., Lal, R., John, S. A., Revel, J. P., & Arnsdorf, M. F. 1991. Atomicforce microscopy and dissection of gap junctions. Science, 253(5026), 1405–1408.

Janovjak, Harald, Kessler, Max, Oesterhelt, Dieter, Gaub, Hermann, &Muller, Daniel J. 2003. Unfolding pathways of native bacteriorhodopsin dependon temperature. EMBO J., 22(19), 5220–5229.

Janovjak, Harald, Struckmeier, Jens, Hubain, Maurice, Kedrov, Alexej,Kessler, Max, & Muller, Daniel J. 2004. Probing the Energy Landscape of theMembrane Protein Bacteriorhodopsin. Structure, 12(5), 871.

Janovjak, Harald, Muller, Daniel J., & Humphris, Andrew D. L. 2005.Molecular Force Modulation Spectroscopy Revealing the Dynamic Response of SingleBacteriorhodopsins. Biophys. J., 88(2), 1423–1431.

Kedrov, Alexej, Ziegler, Christine, Janovjak, Harald, Kuhlbrandt,Werner, & Muller, Daniel J. 2004. Controlled Unfolding and Refolding of aSingle Sodium-proton Antiporter using Atomic Force Microscopy. Journal of Molecu-lar Biology, 340(5), 1143.

Kedrov, Alexej, Janovjak, Harald, Ziegler, Christine, Kuhlbrandt,Werner, & Muller, Daniel J. 2006. Observing Folding Pathways and Kinet-ics of a Single Sodium-proton Antiporter from Escherichia coli. J. Mol. Biol., 355(1),2.

Kessler, Max, & Gaub, Hermann E. 2006. Unfolding Barriers in BacteriorhodopsinProbed from the Cytoplasmic and the Extracellular Side by AFM. Structure, 14(3),521.


Kolbe, Michael, Besir, H., uuml, seyin, Essen, Lars-Oliver, & Oesterhelt,Dieter. 2000. Structure of the Light-Driven Chloride Pump Halorhodopsin at 1.8 AResolution. Science, 288(5470), 1390–1396.

69


Liang, Yan, Fotiadis, Dimitrios, Filipek, Slawomir, Saperstein, David A.,Palczewski, Krzysztof, & Engel, Andreas. 2003. Organization of the GProtein-coupled Receptors Rhodopsin and Opsin in Native Membranes. J. Biol.Chem., 278(24), 21655–21662.

Meier, Thomas, Polzer, Patrick, Diederichs, Kay, Welte, Wolfram, &Dimroth, Peter. 2005. Structure of the Rotor Ring of F-Type Na+-ATPase fromIlyobacter tartaricus. Science, 308(5722), 659–662.

Michel, Hartmut, Oesterhelt, Dieter, & Henderson, Richard. 1980. Or-thorhombic Two-Dimensional Crystal form of Purple Membrane. Proc. Natl. Acad.Sci. USA, 77(1), 338–342.

Moller, Clemens, Allen, Mike, Elings, Virgil, Engel, Andreas, & Muller,Daniel J. 1999. Tapping-Mode Atomic Force Microscopy Produces Faithful High-Resolution Images of Protein Surfaces. Biophys. J., 77(2), 1150–1158.

Moller, Clemens, Fotiadis, Dimitrios, Suda, Kitaru, Engel, Andreas,Kessler, Max, & Muller, Daniel J. 2003. Determining molecular forces thatstabilize human aquaporin-1. J. Struct. Biol., 142(3), 369.

Mou, Jianxun, Yang, Jie, & Shao, Zhifeng. 1995. Atomic Force Microscopy ofCholera Toxin B-oligomers Bound to Bilayers of Biologically Relevant Lipids. J. Mol.Biol., 248(3), 507.

Muller, D. J., & Engel, A. 1997. The height of biomolecules measured with theatomic force microscope depends on electrostatic interactions. Biophys. J., 73(3),1633–1644.

Muller, D. J., Schabert, F. A., Buldt, G., & Engel, A. 1995a. Imagingpurple membranes in aqueous solutions at sub-nanometer resolution by atomic forcemicroscopy. Biophys. J., 68(5), 1681–1686.

Muller, D. J., Schoenenberger, C. A., Buldt, G., & Engel, A. 1996. Immuno-atomic force microscopy of purple membrane. Biophys. J., 70(4), 1796–1802.

Muller, Daniel J., Buldt, Georg, & Engel, Andreas. 1995b. Force-inducedconformational change of bacteriorhodopsin. Journal of Molecular Biology, 249(2),239.

Muller, Daniel J., Engel, Andreas, Carrascosa, J. L., & Velez, M. 1997.The bacteriophage f29 head-tail connector imaged at high resolution with the atomicforce microscope in buffer solution. EMBO J., 16(10), 2547–2553.

Muller, Daniel J., Fotiadis, Dimitrios, Scheuring, Simon, Mller,Shirley A., & Engel, Andreas. 1999a. Electrostatically Balanced Subnanometer

70


Imaging of Biological Specimens by Atomic Force Microscope. Biophys. J., 76(2),1101–1111.

Muller, Daniel J., Sass, Hans-Jurgen, Muller, Shirley A., Buldt, Georg,& Engel, Andreas. 1999b. Surface structures of native bacteriorhodopsin dependon the molecular packing arrangement in the membrane. J. Mol. Biol., 285(5), 1903.

Muller, Daniel J., Heymann, J. Bernard, Oesterhelt, Filipp, Mller,Clemens, Gaub, Hermann, Buldt, Georg, & Engel, Andreas. 2000. Atomicforce microscopy of native purple membrane. Biochim. Biophys. Acta, 1460(1), 27–38.

Muller, Daniel J., Dencher, Norbert A., Meier, Thomas, Dimroth, Peter,Suda, Kitaru, Stahlberg, Henning, Engel, Andreas, Seelert, Holger, &Matthey, Ulrich. 2001. ATP synthase: constrained stoichiometry of the trans-membrane rotor. FEBS Letters, 504(3), 219.

Muller, Daniel J., Hand, Galen M., Engel, Andreas, & Sosinsky, Gina E.2002a. Conformational changes in surface structures of isolated connexin 26 gap junc-tions. EMBO J., 21(14), 3598–3607.

Muller, Daniel J., Kessler, Max, Oesterhelt, Filipp, Mller, Clemens,Oesterhelt, Dieter, & Gaub, Hermann. 2002b. Stability of Bacteriorhodopsina-Helices and Loops Analyzed by Single-Molecule Force Spectroscopy. Biophys. J.,83(6), 3578–3588.

Oesterhelt, F., Oesterhelt, D., Pfeiffer, M., Engel, A., Gaub, H. E., &Muller, D. J. 2000. Unfolding Pathways of Individual Bacteriorhodopsins. Science,288(5463), 143–146.

Palczewski, Krzysztof, Kumasaka, Takashi, Hori, Tetsuya, Behnke,Craig A., Motoshima, Hiroyuki, Fox, Brian A., Trong, Isolde Le,Teller, David C., Okada, Tetsuji, Stenkamp, Ronald E., Yamamoto,Masaki, & Miyano, Masashi. 2000. Crystal Structure of Rhodopsin: A G Protein-Coupled Receptor. Science, 289(5480), 739–745.

Persike, Norbert, Pfeiffer, Matthias, Guckenberger, Reinhard, Rad-macher, Manfred, & Fritz, Monika. 2001. Direct observation of different surfacestructures on high-resolution images of native halorhodopsin. J. Mol. Biol., 310(4),773.

Pogoryelov, Denys, Yu, Jinshu, Meier, Thomas, Vonck, Janet, Dimroth,Peter, & Muller, Daniel J. 2005. The c15 ring of the Spirulina platensis F-ATP synthase: F1/F0 symmetry mismatch is not obligatory. EMBO Rep., 6(11),1040–1044.

71


Pohl, Peter, Saparov, Sapar M., Borgnia, Mario J., & Agre, Peter. 2001.Highly selective water channel activity measured by voltage clamp: Analysis of planarlipid bilayers reconstituted with purified AqpZ. Proc. Natl. Acad. Sci. USA, 98(17),9624–9629.

Puntheeranurak, Theeraporn, Stroh, Cordula, Zhu, Rong, Angsuthana-sombat, Chanan, & Hinterdorfer, Peter. 2005. Structure and distribution ofthe Bacillus thuringiensis Cry4Ba toxin in lipid membranes. Ultramicroscopy, 105(1-4), 115.

Raychowdhury, Malay K., Gonzalez-Perrett, Silvia, Montalbetti, Nico-las, Timpanaro, Gustavo A., Chasan, Bernard, Goldmann, Wolfgang H.,Stahl, Stefanie, Cooney, Adele, Goldin, Ehud, & Cantiello, Horacio F.2004. Molecular pathophysiology of mucolipidosis type IV: pH dysregulation of themucolipin-1 cation channel. Hum. Mol. Genet., 13(6), 617–627.

Rossjohn, Jamie, Feil, Susanne C., McKinstry, William J., Tweten, Rod-ney K., & Parker, Michael W. 1997. Structure of a Cholesterol-Binding, Thiol-Activated Cytolysin and a Model of Its Membrane Form. Cell, 89(5), 685–692.

Sapra, K. Tanuj, Besir, Huseyin, Oesterhelt, Dieter, & Muller, Daniel J.2006. Characterizing Molecular Interactions in Different Bacteriorhodopsin Assembliesby Single-molecule Force Spectroscopy. J. Mol. Biol., 355(4), 640.

Schabert, F. A., & Engel, A. 1994. Reproducible acquisition of Escherichia coliporin surface topographs by atomic force microscopy. Biophys. J., 67(6), 2394–2403.

Schabert, F. A., Henn, C., & Engel, A. 1995. Native Escherichia coli OmpF porinsurfaces probed by atomic force microscopy. Science, 268(5207), 92–94.

Schenk, Andreas D., Werten, Paul J. L., Scheuring, Simon, de Groot,Bert L., Muller, Shirley A., Stahlberg, Henning, Philippsen, Ansgar,& Engel, Andreas. 2005. The 4.5 A Structure of Human AQP2. J. Mol. Biol.,350(2), 278.

Scheuring, Simon, Ringler, Philippe, Borgnia, Mario, Stahlberg, Henning,Muller, Daniel J., Agre, Peter, & Engel, Andreas. 1999. High resolutionAFM topographs of the Escherichia coli water channel aquaporin Z. EMBO J., 18,4981–4987.

Scheuring, Simon, Reiss-Husson, Francoise, Engel, Andreas, Rigaud,Jean-Louis, & Ranck, Jean-Luc. 2001. High-resolution AFM topographs ofRubrivivax gelatinosus light-harvesting complex LH2. EMBO J., 20(12), 3029–3035.

Scheuring, Simon, Seguin, Jerome, Marco, Sergio, Levy, Daniel, Robert,Bruno, & Rigaud, Jean-Louis. 2003. Nanodissection and high-resolution imaging

72


of the Rhodopseudomonas viridis photosynthetic core complex in native membranesby AFM. Proc. Natl. Acad. Sci. USA, 100(4), 1690–1693.

Scheuring, Simon, Rigaud, Jean-Louis, & Sturgis, James N. 2004a. VariableLH2 stoichiometry and core clustering in native membranes of Rhodospirillum photo-metricum. EMBO J., 23, 4127–4133.

Scheuring, Simon, Sturgis, James N., Prima, Valerie, Bernadac, Alain,Levy, Daniel, & Rigaud, Jean-Louis. 2004b. Watching the photosynthetic ap-paratus in native membranes. Proc. Natl. Acad. Sci. USA, 101(31), 11293–11297.

Scheuring, Simon, Busselez, Johan, & Levy, Daniel. 2005. Structure of theDimeric PufX-containing Core Complex of Rhodobacter blasticus by in Situ AtomicForce Microscopy. J. Biol. Chem., 280(2), 1426–1431.

Seelert, Holger, Poetsch, Ansgar, Dencher, Norbert A., Engel, An-dreas, Stahlberg, Henning, & Muller, Daniel J. 2000. Structural biology:Proton-powered turbine of a plant motor. Nature, 405(6785), 418.

Smith, Douglas E., Tans, Sander J., Smith, Steven B., Grimes, Shelley,Anderson, Dwight L., & Bustamante, Carlos. 2001. The bacteriophage phi-29 portal motor can package DNA against a large internal force. Nature, 413(6857),748.

Song, Langzhou, Hobaugh, Michael R., Shustak, Christopher, Cheley,Stephen, Bayley, Hagan, & Gouaux, J. Eric. 1996. Structure of Staphylococcalalpha -Hemolysin, a Heptameric Transmembrane Pore. Science, 274(5294), 1859–1865.

Stahlberg, Henning, Muller, Daniel J., Suda, Kitaru, Fotiadis, Dimitrios,Engel, Andreas, Meier, Thomas, Matthey, Ulrich, & Dimroth, Peter.2001. Bacterial Na+-ATP synthase has an undecameric rotor. EMBO Rep., 2(3),229–233.

Stamouli, Amalia, Kafi, Sidig, Klein, Dionne C. G., Oosterkamp, Tjerk H.,Frenken, Joost W. M., Cogdell, Richard J., & Aartsma, Thijs J. 2003.The Ring Structure and Organization of Light Harvesting 2 Complexes in a Recon-stituted Lipid Bilayer, Resolved by Atomic Force Microscopy. Biophys. J., 84(4),2483–2491.

Stryer, L. 1987. The molecules of visual excitation. Sci. Am., 257(1), 50.

Sun, Mei, Goldin, Ehud, Stahl, Stefanie, Falardeau, John L., Kennedy,John C., Acierno, James S., Jr., Bove, Catherine, Kaneski, Christine R.,Nagle, James, Bromley, Matthew C., Colman, Matthew, Schiffmann,Raphael, & Slaugenhaupt, Susan A. 2000. Mucolipidosis type IV is caused bymutations in a gene encoding a novel transient receptor potential channel. Hum. Mol.Genet., 9(17), 2471–2478.

73


Vie, V., Van Mau, N., Pomarede, P., Dance, C., Schwartz, J. L., Laprade,R., Frutos, R., Rang, C., Masson, L., Heitz, F., & Le Grimellec, C. 2001.Lipid-induced Pore Formation of the Bacillus thuringiensis Cry1Aa Insecticidal Toxin.J. Memb. Biol., 180(3), 195.

Ward, R. J., & Leonard, K. 1992. The Staphylococcus aureus a-toxin channelcomplex and the effect of Ca2+ ions on its interaction with lipid layers. J. Struct.Biol., 109(2), 129–141.

Williams, Karen A. 2000. Three-dimensional structure of the ion-coupled transportprotein NhaA. Nature, 403(6765), 112.

Worcester, D. L., Miller, R. G., & Bryant, P.J. 1988. Atomic force microscopyof purple membranes. J. Microsc., 152(3), 817–821.

Worcester, D. L., Kim, H. S., Miller, R. G., & Bryant, P. J. 1990. Imagingbacteriorhodopsin lattices in purple membranes with atomic force microscopy. J. Vac.Sci. Technol. A, 8(1), 403–405.

Yang, Jie, Tamm, Lukas K., Tillack, Thomas W., & Shao, Zhifeng. 1993.New Approach for Atomic Force Microscopy of Membrane Proteins: The Imaging ofCholera Toxin. J. Mol. Biol., 229(2), 286.

74

The asymmetric properties of Purple Membrane

3 Differential Stiffness and lipid mobility in both

leaflets of Purple Membrane

3.1 Introduction

3.1.1 The natural asymmetry between Purple Membrane leaflets

Membrane proteins are essential for communication, side specific interactions and molec-

ular transport through the membrane of any living organism. In order to fulfil these

vectorial tasks, membrane proteins are extensively asymmetrically oriented (Houslay &

Stanley 1982; Op den Kamp 1979). Lipids generally show less defined asymmetrical

behaviour between the two membrane leaflets and adapt their location around proteins

(Houslay & Stanley 1982; Rothman & Lenard 1977). Phospholipids, however, tend to

be located within the cytoplasmic leaflet of plasma membranes (Bergelson & Barsukov

1977). PM also exhibits some asymmetry between its two leaflets’ lipid composition and

bR-lipid interactions, consistently with bR vectorial task and the strong ionic asymme-

try of Halobacterium salinarium natural environment (Engel & Catchpole 2005). PM

extracellular leaflet contains almost only glycolipids (Weik et al. 1998), some of which

specifically and tightly bound to bR, while phospholipids are located mainly within the

cytoplasmic leaflet (Renthal & Cha 1984), and seem less localized than the glycolipids

(see Introduction chapter, §1.2 for details). bR residues’ distribution across PM also

shows some asymmetry; the cytoplasmic part of the protein is richer in exposed acidic

residues1, some of which play an important role in the bR pumping process by acting as

an attractive antenna for the protons that will further be pumped throughout the protein

(Fig. 1.3) (Nachliel et al. 1996). The extracellular part of bR, in contrast, is richer in

aromatic residues (Fig. 4.9) and possesses an extensive network of water-mediated hy-

drogen bonds which rigidify the protein structure (Neutze et al. 2002). The net charge

1 namely Asp36, Asp38, Asp102, Asp104, Glu105, Glu161, Glu231, Glu233, Glu236 and Asp242

75


asymmetry exhibited by bR between the two PM leaflets is responsible for the important

permanent dipole moment of the protein (55D experimentally) (Porschke 1996). Most

of the charged residues present in bR extracellular half are part of the proton translo-

cation pathway (Asp85, Arg82, Glu204, Glu194 and possibly Glu74 and Glu75) with

only Glu74, Glu75, Glu194 and Glu204 located at the protein surface, forming the pro-

ton release cluster (Sanz et al. 2001). Neutral hydrophobic residues also constitute an

important part of the protein (21 Val residues, 35 Leu residues) but are evenly present,

mainly within the intra-membranar region where the protein is in contact with the sur-

rounding lipid tails.

bR utilizes its particular amino acid distribution and interactions with the surrounding

lipids to catalyze its photo-induced conformational changes and ensure vectorial proton

pumping (Neutze et al. 2002; Luecke et al. 2003; Lanyi 2004). In order to relate PM

asymmetric properties with bR activity, it is necessary to understand the different types

of interactions involved and their respective roles. Furthermore, the composition of PM

natural environment (Table 1.1) makes it is fundamental to consider the importance

of salt ions in PM local interactions, ideally investigating both membrane leaflets inde-

pendently. Experimentally, it is difficult to observe membrane asymmetry and labelling

techniques are generally required (Gale & Watts 1992; Renthal & Cha 1984; Muller et al.

1996). Using atomic force microscopy (AFM), its possible to study locally the overall

cohesion of PM in buffer solution, and particularly the influence of salt on PM mechan-

ical stiffness and lipid mobility. The natural asymmetry between the two sides of PM

(Gale & Watts 1992; Renthal & Cha 1984; Subramaniam et al. 1993) can be exploited

to distinguish them by alternate contact mode AFM (AC-AFM). Furthermore, probing

with a nanometre size tip allows quantitative investigation of inter-trimer protein-lipid

and lipid-lipid interactions in PM by inducing controlled mechanical stress on the mem-

brane (force curves). Since the AFM tip surface charge depends on the solution pH,

76


electrostatic changes at the membrane surface can also be probed.

3.1.2 Chapter overview

This chapter investigates PM asymmetry at a global level. A new way of differentiating

the two membrane sides using the phase signal available in alternate contact atomic

force microscopy (AC-AFM) is presented (Voıtchovsky et al. 2006). This method does

not require molecular resolution, and is applied to study stiffness and inter-trimer lipid

mobility in both leaflets of the PM independently, over a broad range of pH and salt

concentrations. PM stiffens with increasing salt concentration according to two differ-

ent regimes. At low salt concentration, the membrane Young’s normal modulus grows

quickly but differentially for both leaflets, exhibiting higher moduli values for the ex-

tracellular than for the cytoplasmic leaflet. At higher salt concentration, both leaflets

behave similarly and PM stiffness converges toward its value in a natural environment.

Changes in pH do not affect PM stiffness, however the protein-lipid crystal assembly is

less pronounced at pH> 10. Lipid mobility is high in the cytoplasmic leaflet, especially

at low salt concentration, but negligible in the extracellular leaflet regardless of pH or

salt concentration.

An independent lipid mobility study by solid-state nuclear magnetic resonance confirms

and quantifies the AFM qualitative observations. This ssNMR work was carried out by

Miya Kamihira and is consequently not presented here (see (Voıtchovsky et al. 2006;

Kamihira & Watts 2006)).

77


3.2 Distinction of Purple Membrane extracellular and cyto-

plasmic sides using AC-AFM

3.2.1 High resolution imaging and Phase information

When PM in solution is deposited on mica, adsorbed patches assemble into the char-

acteristic hexagonal lattice (Fig. 3.1). A small band (typically 5 nm to 50 nm wide)

of non ordered membrane can be observed around a developing patch. Patches are

5.3 ± 0.4 nm thick and 500 nm to 2 µm wide, and do not show any preferential ori-

entation for the adsorption (extracellular or cytoplasmic side facing mica). Due to the

structural asymmetry of PM, molecular level studies on PM mechanical or electrostatic

properties require leaflet specific experiments. All AFM experiments reported to date

showing side distinction used high resolution imaging (Muller et al. 1996; Moller et al.

1999). This method requires a careful balancing of the electrostatic double layer forces

between the membrane and the AFM tip. The thickness of the double layer is character-

ized by its Debye length, 1/κ (see §4.2 for more details). In the case of KCl in aqueous

solution (Israelachvili 1991):

1/κ ∼= 0.304/√

[KCl] (3.1)

where [KCl] is the concentration of KCl in mole per liter and 1/κ is in nanometers. Con-

sequently, high resolution imaging imposes a specific salt concentration of the imaging

buffer (Muller et al. 1999) which prevents PM side distinction away from the optimum

conditions. We have overcome this problem by using the phase information provided by

AC-AFM. Operated in very soft AC mode (A0/A close to 1, see §3.6 Material and Meth-

ods), the phase shift is very sensitive to changes in energy dissipation due to tip-sample

interactions (San Paulo & Garcıa 2001; Garcıa et al. 1999; Cleveland et al. 1998). The

natural asymmetry between the extracellular and cytoplasmic sides of PM (Renthal &

Cha 1984; Subramaniam et al. 1993) is sufficient to offer a clear phase contrast when

78


scanned in any buffer studied. The extracellular patches appear darker than the cy-

toplasmic patches in the phase image (Fig. 3.1.B and 3.1.C). High resolution imaging

under favorable buffer conditions confirmed the side assignment obtained through phase

shift imaging (Fig. 3.1.D and 3.1.E). Scanning in low salt buffer accentuates the asym-

metry resolved by the AFM tip between the two sides of PM, and thus the phase shift.

This new way of distinguishing the PM two sides has been extensively used in this study,

mainly for side-specific force spectroscopy.

3.2.2 Membrane protrusions on the cytoplasmic side and pH effect

Additional features which help to distinguish the two sides of PM when adsorbed onto

mica and present under any salt or pH conditions studied (20 mM KCl to 300 mM KCl,

pH 5 to pH 12), are the frequent and random protrusions exhibited by cytoplasmic

patches (patches with cytoplasmic side exposed for imaging and their extracellular side

facing the substrate), in contrast to the smoothness of the extracellular patches (Fig.

3.1.A, 3.1.B). These features, appearing only on cytoplasmic patches, have been ob-

served before (Stark et al. 2001) but never commented on or explained. High resolution

images of cytoplasmic patches showed that the PM lattice is still visible on top of the

protrusions (data not shown). They are consequently due to material located underneath

the membrane, between the extracellular leaflet and the substrate. Extracellular patches

do not exhibit such irregularities. Moreover they seem to prevent their formation and

cytoplasmic patch regions adjacent to extracellular patches exhibit a smooth band with-

out protrusions (Fig. 3.1.B), forming a ”smooth belt” around the extracellular patches.

Finally, changing the pH of the buffer solution does not seem to affect the global cohesion

of the membrane, but more the tip-sample interaction by changing the surface charge of

the AFM tip from negative (pH 10) to positive (pH 6). However, at pH> 10, despite

the good cohesion of the PM lattice, the assembly seems more difficult and large areas

79


A B D

E

C

F I

J

G

H

Figure 3.1: Distinction of PM extracellular and cytoplasmic sides using AC-AFM (AE).(A). Topographic image. The phase image (B) reveals a contrast between extracellular patches(darker) and cytoplasmic patches (lighter). A profile of (B) gives∼ 7 cytoplasmic-extracellularshift and ∼ 15 cytoplasmic-mica shift (C). High resolution topographic images of PM cyto-plasmic (D) and extracellular (E) side obtained by zooming respectively on the lighter anddarker patches in (B). The protrusions on the cytoplasmic patches and a smooth belt aroundthe extracellular patch are clearly visible in panels (A) and (B) (dashed line). Effect of saltremoval on the PM imaged by AC-AFM in air (FJ). The membrane has been rinsed withultrapure water and dried at room temperature. The extracellular (pitted) and cytoplasmic(cracked) sides of PM after drying (F). Higher magnifications of the extracellular patch (G)and (H). Hexagonal depressions related to the cytoplasmic leaflet and pits are visible. Highermagnification of the cytoplasmic cracked patch (I) and (J). Arrows point to larger holes cross-ing extracellular patches in panels (F) and (G) and to trimers in panel (J). The scale bar is 1µm for (A) and (B), 10 (C), 3 nm (D) and (E), 1 µm (F), 100 nm (G), 50 nm (H), 100nm (I), and 50 nm (J). Panels (F−I) are topographic images, and (J) is a phase image.

80


of a non-ordered thinner membrane are visible (Fig. 3.2) around the well assembled

patches. The lattice then assembles from these non-ordered regions (Neugebauer et al.

1983). This mis-assembled membrane could be due to the lysine residues of bR loosing

their positive charge at pH 10, thus weakening the bR interaction with highly negatively

charged lipids and making the lattice assembly less favorable. Since almost all the lysine

residues of bR are located in the cytoplasmic leaflet, once the membrane is assembled,

the cohesion is maintained by the extracellular leaflet specific interactions. Consistently

increasing the salt concentration of the buffer allowed a normal PM assembly. It should

be noted that around the cytoplasmic patches, the protrusions mentioned above are

already present in this non-ordered area (Fig. 3.2).

3.3 Observation of the asymmetric effect of salt on Purple

Membrane

3.3.1 Imaging with no salt, in liquid and in air

Imaging PM in liquid under very low salt concentration (< 20 mM KCl) is difficult due

to a significant increase of the double layer thickness (Eq. 3.1). In water (ultrapure wa-

ter), the AFM silicon nitride tip feels a strong electrostatic interaction scanning over the

negatively charged PM cytoplasmic side, thus preventing any non-destructive imaging.

However, it is possible to obtain images of the less charged extracellular side, revealing

a rough ”pitted” membrane, with many circular depressions the diameter of few bR

trimers and 1−2 nm deep (Fig. 3.3.C-D). In order to image the effect of salt removal on

the cytoplasmic side of the membrane, it is necessary to dry the sample for AC imaging

in air. Two distinct sides are revealed: a pitted side similar to the membrane imaged in

ultrapure water, and a cracked side, with the cracks following a hexagonal lattice (Fig.

3.1 F). These pitted and cracked patches had previously been observed by cryo-electron

81


0 0.2 0.4 0.6 0.8 1.0 1.2 1.4

1

0

-1

-2

-3-4

Position [µm]

Sect

ion

[nm

]

1

2

3

Figure 3.2: PM assembly at pH 10. The arrows 1, 2, and 3 show cytoplasmic patches,extracellular patches, and nonassembled membrane, respectively. The nonassembled membraneis ∼ 1 nm thinner than ordered lattice (profile line). Different lattice orientations are observablein assembled patches, and protrusions are visible in and around cytoplasmic patches. The scalebar is 400 nm.

82


microscopy (Fisher et al. 1978), and assigned to the cytoplasmic and extracellular mem-

brane sides respectively. However, contrary to this report, we attribute the cracks to

the cytoplasmic side, and the pits to the extracellular side. This was confirmed using a

mutant bR (M163C) that can covalently bind to gold via a cysteine residue located on its

cytoplasmic side, and prevents normal assembly of patches having their cytoplasmic side

in contact with gold (Fig. 3.3.A-B). Patches with their extracellular side on gold were

normally assembled, and also exhibited hexagonal cracks, similar to those shown in Fig.

3.1.I. Examination of the cracks in the cytoplasmic leaflet reveals that they are following

the bR hexagonal lattice (Fig. 3.1.I-J). The larger cracks traverse the membrane but the

smaller cracks are only about one leaflet deep (2 to 3 nm) with occasional deeper holes

corresponding to fully removed trimers. Phase imaging allows the identification of trimer

rows (Fig. 3.1.J). Cracks propagate between these rows and are probably due to missing

lipids, removed by capillary forces during the membrane drying, and condensation of the

lipid chains (Zaccai 1987). Some lipids are removed from the cytoplasmic leaflet inter-

trimer space, weakening the membrane, and occasionally allowing whole trimers to be

extracted along with the lipids. The extracellular leaflet does not show clear cracks, but

their presence is suggested in the opposite leaflet by depressions following the hexagonal

lattice (Fig. 3.1.G). Small round pits similar to the holes observed in PM in ultrapure

water (Fig. 3.1.H) and larger circular holes (50 − 150 nm in diameter) spanning the

whole membrane width, are visible.

3.3.2 Side-specific Force Spectroscopy

In order to quantify the differences reported above between the two PM leaflets, we

have taken series of force curves on both sides separately, varying salt concentration and

pH. All the curves are extension curves, showing the approach of the tip toward the

sample. The corresponding retraction curves (not shown) are identical except for some

83


A B C

D

Figure 3.3: PM containing M163C bR mutant that can covalently bind its cytoplasmic sideto gold (A-B). The patches showing a normal assembly are hexagonally cracked as in Fig.3.1 (F), (I). The attachment of the mutated residue to gold prevents normal assembly ofextracellular patches (cytoplasmic surface in contact with gold). (C) Scanning of PM in purewater; the sample has been extensively rinsed with ultrapure water after a normal adsorptionin buffer. The scanning tend to be destructive even at small amplitudes and with soft tapping.(D) Magnification of a patch in (C). The membrane shoes the same pitted appearance as inFig. 3.1 (G-H). Scale bars: 150 nm (A) and (B), 300 nm (C) and 30 nm (D). All imagesare topographic images.

possible adhesion of the tip to the sample. Representative force curves are presented in

Fig. 3.4; they are taken on both sides of PM and in a buffers containing 10 mM Tris

and 20 mM , 25 mM , 30 mM , 40 mM , 150 mM , and 300 mM KCl, respectively. The

pH was set to 8. At 20 mM and 25 mM KCl concentration, no obvious difference

can be seen between the force curves on mica and on the cytoplasmic side of PM. Both

exhibit a double layer electrostatic repulsion zone (Eq. 3.1) followed by a linear deflection

starting for forces smaller than 100 pN . This shows that the AFM tip feels almost no

membrane repulsion pressing on the cytoplasmic side. The tip penetrates very easily

through the whole membrane to reach the mica underneath. However, imaging the

sample after having taken force curves revealed no permanent damage of the membrane,

suggesting an immediate healing after a hole was made by the tip. The cytoplasmic

leaflet is weakened by a lack of shielding between highly charged phospholipids, and by

84


the mobility of the inter-trimer lipids. In contrast, the extracellular side of the membrane

is much more resistant to the tip pressure and continuously deflects the cantilever. At

30 mM KCl concentration in the buffer, force curves on the cytoplasmic side of PM

exhibit an intermediate behaviour between the curves on mica and on the extracellular

side due to partial shielding of electrostatic repulsion between the cytoplasmic leaflet

lipids. This could be explained by the tip perforating only the cytoplasmic leaflets. The

critical salt concentration past which the electrostatic repulsion within the cytoplasmic

leaflet is fully shielded is 40 mM KCl. Curves on extracellular and cytoplasmic side are

very similar. Both leaflets deform in the same way under the tip pressure. Increasing

the salt concentration hardens the membrane symmetrically up to the point where force

curves on the membrane and on mica show almost no difference (at 300 mM KCl).

Side-specific force curves have also been acquired at pH 6, 8 and 10 in a 150 mM KCl

buffer. No obvious difference could be observed between the force curves taken on the

cytoplasmic and the extracellular side of PM (Fig. 3.7.C).

3.4 Discussion

3.4.1 Lipid mobility and density

Despite the very high salt concentration of the Halobacterium salinarium native envi-

ronment (∼ 4 M NaCl), PM can still assemble on mica at low salt concentration, and

membrane patches extensively rinsed with ultrapure water exhibit the hexagonal lattice

(Fig. 3.1.F and 3.1.I) characteristic of native PM (Blaurock & Stoeckenius 1971; Hen-

derson & Unwin 1975). However, the lack of salt affects the membrane in an asymmetric

fashion. The cytoplasmic leaflet becomes very soft with weakly attached inter-trimer

lipids that can be removed by capillary forces, while the extracellular leaflet remains

almost unchanged. These observations indicate that highly branched and charged phos-

85


extracellularcytoplasmicmica

20 nm

500 pN

20 mM KCl

25 mM KCl

30 mM KCl

40 mM KCl

150 mM KCl

300 mM KCl

Figure 3.4: Force curves on the extracellular side (light shaded) and the cytoplasmic side (darkshaded) of PM and on mica (solid) at different salt concentrations (20 to 300 mM KCl). Thecurves have been offset for better visibility. The maximum force applied by the tip is ∼ 2nN(500 pN force scale) corresponding to the force beyond which all the curves look similar whensuperimposed, regardless of the sample.

86


pholipids located in the inter-trimer space of the cytoplasmic leaflet (PG, PGS, PGP-Me

and BPG), interact with bR and each other mainly through non-specific long range

electrostatic forces, yet still possessing high mobility within the leaflet. This is consis-

tent with other studies (Renthal & Cha 1984; Gale & Watts 1991), in which the charge

asymmetry of PM has been determined using fluorescence and NMR. The AFM images

emphasize the importance of phospholipids for the overall membrane cohesion, since

their removal induces the extraction of whole bR trimers by capillary forces (Fig. 3.1).

However, the main inter-trimer interactions stabilizing the membrane together appear

to be specific and to take place in the extracellular leaflet, as previously reported for

intra-trimer interactions (Grigorieff et al. 1996; Essen et al. 1998). These inter-trimer

interactions are sufficient to maintain the lattice structure of PM, even in very low salt

conditions when the cytoplasmic leaflet is dramatically affected or partially removed

(Fig. 3.1). Sulfoglycolipids are known to be predominantly located in the extracellu-

lar leaflet (Henderson et al. 1978) and in the inter-trimer space (Weik et al. 1998),

and to interact strongly and specifically with bR, mainly with its tryptophan residues.

Tryptophan residues have been shown to be determinant in the anchoring of proteins in

lipid membranes in general (Killian & von Heijne 2000; Antoranz Contera et al. 2005).

In PM, the asymmetric interaction of sulfoglycolipids with bR is underscored by the

position of the tryptophan residues in bR, since almost all of them are located in the

extracellular leaflet, some of them pointing toward the inter-trimer space (Luecke et al.

1999). More evidence of the specificity of the lipids and the interactions in each leaflet

can be found in Fig. 3.1.A, B where the cytoplasmic patches (extracellular side facing

mica) exhibit many protrusions due to material located under the patch. Since these

protrusions are already present in the membrane assembling edge (Fig 3.2), they are

probably lipid vesicles joining the membrane edge during the patch formation. These

vesicles are indeed able to fuse with the cytoplasmic leaflet of the membrane, allowing

87


PM to be smooth if they are sandwiched between mica and the cytoplasmic leaflet. This

suggests that the vesicles are composed of phospholipids, which would also explain why

they cannot cross the membrane to reach the cytoplasmic leaflet when trapped between

mica and the extracellular leaflet. The location of the vesicles exclusively under cyto-

plasmic patches demonstrates the specificity of the interactions within the extracellular

leaflet. Vesicles trapped close enough to an extracellular patch (cytoplasmic side down)

seem to reach it by lateral diffusion, and fuse to it, thus creating this ”smooth belt”

around it. This phenomenon, however, raises an interesting question about lipid density

within the cytoplasmic leaflet; if the above deduction is correct, the cytoplasmic leaflet

can find room for more lipids than the number initially present just after the assembly,

since vesicles are fusing to an already assembled patch. A possible explanation would be

the high mobility of the phospholipids, allowing them to diffuse within the whole cyto-

plasmic leaflet and leave it at its edges if the lipid density is too high and the head-group

repulsion excessive. The density of lipids has indeed been showed to be significantly

higher in the cytoplasmic leaflet (Blaurock & King 1977).

3.4.2 Estimation of Purple Membrane differential stiffness by force spec-

troscopy

We have estimated PM Young’s normal bulk modulus by fitting the force curves with

a theory recently developed by Chadwick (Chadwick 2002). This theory assumes a

sphere indenting a thin film located on a substrate composed of harder material. The

modulus has been calculated independently for both sides of the membrane at each salt

concentration. The AFM tip apex has a curvature radius of 10 nm to 20 nm, and thus

presses on at least a whole trimer when indenting the membrane. As shown from imaging

(Fig. 3.1), bR conserves its trimeric form both in the lattice and when extracted from

the membrane. bR is indeed more stable in trimeric than in monomeric form (Haltia &

88


Freire 1995; Lopez et al. 1999); this implies that force curves are mainly representative

of the inter-trimer protein-lipid and lipid-lipid interactions within PM. For a membrane

not bound to its substrate, the indentation δ of a sphere of radius R pressing with a

force F on the thin membrane is given by:

F =2πER

3hδ2 (3.2)

where h is the membrane thickness and E the Young’s modulus. The membrane is

considered incompressible (Poisson’s ratio of 1/2) which is a reasonable assumption for

biological membranes. The major improvement of Chadwick theory over the traditional

Sneddon’s modification of the Hertzian approach (Hertz 1882; Sneddon 1965) used in

previous AFM experiments (Domke & Radmacher 1998), is the assumption of a hard

substrate under the membrane, which is relevant in our case. The Sneddon’s approach

assumes an indenting cone on a semi-infinite medium, also giving a quadratic dependence

of the force in δ, but with different constant parameters:

F =3πE tan α

2δ2 (3.3)

where α is the half opening angle of the indenting cone and Poisson’s ration is also

1/2. We have fitted our data with both theories for comparison. Fig. 3.5 shows the

principle used to do it, taking the curve on mica as a reference. The indentation curve

is superimposed on the reference curve (on mica) for measuring δ. Fδmax corresponds

to the applied tip force beyond which the maximal possible indentation δmax on the

membrane is reached and the curve does not then represent the indentation properties of

the membrane, but more those of the substrate underneath (substrate region) (Domke

& Radmacher 1998). If F ≥ Fδmax, then δ = δmax and the force curve on the membrane

is similar to that for the substrate underneath (substrate region). When the AFM tip

89


δmax hδ

mica

PM

AFM tip

F

δmax − δ(F)

micaPM

Z-piezo extension

Forc

e

Electrostatic region

Indentationregion

Fδmax

Substrateregion

Figure 3.5: Principle of measurement of the effective indentation of the membrane by theAFM tip. The region taken for fitting the models, labeled Indentation region, corresponds tothe region where the equivalent force curve on mica is linear and the applied force is smallerthan Fδmax.

90


approaches the membrane, electrostatic effects arise from the tip crossing the electrostatic

double layer (electrostatic region). Membrane force curves were fitted in the region where

the tip is in contact with the sample and the corresponding curve on mica is linear,

between the electrostatic and the substrate regions. The tip radius R, the membrane

thickness h, the cantilever spring constant k and the half opening angle α of the tip

cone were assumed to be identical for each curve analyzed. Many independent fitting

trials suggested values of R = 20 nm and α = 30. The membrane thickness h is

5.3 nm (Fig. 3.1.C), and the cantilever spring constant assumed to be k = 0.08 N/m

(see Methods). Fig. 3.6.A shows the evolution of the Young’s modulus with increase of

KCl concentration in the buffer (pH 8), calculated using the Chadwick theory. Young’s

modulus calculated from Sneddon’s modification of the traditional Hertzian theory shows

a similar evolution with values systematically 35% higher than for Chadwick treatment

and are thus not shown for more clarity. Two different regimes are distinguishable and

the evolution of the Young’s modulus has been fitted with a bi-exponential curve of type:

E = Enat − Ele−c/Kl − Ehe

−c/Kh (3.4)

where Enat is the Young’s modulus of PM close to its natural medium salt concentration

(∼ 4 M NaCl), c is the KCl concentration and the two exponential terms represent

the changes in E at low and high salt concentration respectively. The salt concentration

parameters Kl = 12 mM and Kh = 650 mM have been fixed as global variables for both

sides of the membrane. This choice reflects the fact that when free, these parameters give

similar final results within 20% and that their globalization allows an easier interpretation

of the results. Interestingly, PM stiffness increases with the salt concentration despite a

better surface charge shielding. This is because at low salt concentration, the phospho-

lipids are mobile within the cytoplasmic leaflet and can accommodate the indentation

91


A

B

You

ng

's N

orm

al M

od

ulu

s [M

Pa]

extracellular

KCl concentration [mM]

cytoplasmic

15

20

5

10

0 50 100 150 200 250 300

KCl concentration [mM]

0 50 100 150 200 250 300

extracellularcytoplasmic

45

23

6

δ max

[n

m]

Figure 3.6: (A) Evolution of PM Youngs normal modulus with KCl concentration at pH8, measured from each side membrane independently by fitting force curves with Chadwicksmodel. The evolution of the moduli values obtained is fitted with a biexponential curve. (B)Evolution of the maximum tip indentation δmax in the membrane, also obtained from forcecurves fitting.

92


without significantly increasing the electrostatic potential. The effective membrane stiff-

ness is therefore decreased because the structure of the cytoplasmic leaflet is temporarily

modified. At low salt concentration (between 20 mM KCl and 40 mM KCl) the mea-

sured Young’s modulus differs depending on the side indented. The extracellular side

gives a value of 14.4 MPa for Eextracell.l , while Ecytoplasm.

l = 75 MPa for the cytoplasmic

side, giving a ratio for Ecytoplasm.l /Eextracell.

l ∼ 5. The stiffening is five times more im-

portant in the cytoplasmic leaflet. At higher salt concentration, the parameter Eh gives

comparable values for both sides of the membrane: 17 MPa for the extracellular side

and 16.4 MPa for the cytoplasmic side. Since the behaviour of the membrane Young’s

modulus does not seem to depend on the side indented at high salt concentration, Enat

was logically set as a global parameter, giving a value of Enat = 28 MPa. This value is

however an estimate, firstly because the highest salt concentration studied is still about

ten times less than that of the natural medium concentration for the bacterium, and also

due to the very large uncertainty in the moduli obtained from force curve fitting. It is

difficult to select the right region of the curve investigated to fit the model. In order to

be valid, the indentation curve should be fit between the double layer electrostatic region

and the substrate region where the indentation reaches its maximum δ = δmax (see Fig.

3.5). At high salt concentration, the membrane stiffens and it becomes more difficult to

determine the transition between the indentation region and the substrate region; this

problem is the main factor of uncertainty. The uncertainty arising from the moduli cal-

culated over the different curves obtained in one buffer is negligible (< 5%). The model

used assumes a linear deformation of the indented layer (Hooke deformation) which may

not be true at low salt concentration. However, since the force curves used for fitting were

identical in extension and retraction, the deformation of this membrane is elastic. The

model used also takes into account a hard substrate underneath (Emica = 15− 25 GPa),

and the use of small forces (< 2 nN) gives consistent results on comparison with surface

93


force apparatus (SFA) measurements on phospholipid bilayers (Benz et al. 2004; Cevc &

Marsh 1987). The systematic 35% over estimation of Young’s modulus using the Sned-

don’s modified Hertzian theory is consistent with the observations made by Dimitriadis

et al. (Dimitriadis et al. 2002) when comparing the two theories used here for thin

film indentation. Changing the pH of the buffer solution does not seem to modify the

stiffness of PM. The extracellular and cytoplasmic Young’s moduli obtained by fitting

the corresponding force curves provided no evidence for the membrane to be asymmet-

rically affected by pH changes within the range measured (Fig. 3.7.A-C). Finally, the

evolution of the maximal indentation δmax with changing KCl concentration (Fig. 3.6.B)

is consistent with a qualitative observation of the force curves. Both theories used for

fitting gave a similar value for δmax within 5%.

3.5 Conclusions

We have studied salt and pH effects on cohesion, stiffness and mobility in both leaflets

of PM independently, concentrating on inter-trimer lipids and protein-lipid interactions.

AFM imaging allowed a direct observation of how PM is asymmetrically affected by

salt removal. Side-specific indentation of the membrane provided a quantification of the

changes in stiffness, using a novel method for distinguishing the extracellular and the

cytoplasmic sides which exploits the phase shift information available when the AFM

is operated in AC mode. This method, based on PM natural asymmetry, does not

require high resolution imaging. Fitting the force curves separately obtained for the

extracellular and the cytoplasmic side of PM, we have calculated the evolution of PM

normal Young’s modulus with increase of the buffer pH and salt concentration, and

carried out an estimation of PM modulus under natural saline concentration. NMR

measurements offer an independent way of observing the phospholipid mobility and order,

showing an increased mobility of the head groups on salt removal. All the experiments

94


20 nm

500 pNextracellularcytoplasmicmica

pH 6

pH 8

pH 10

A

B

C

No

rma

l Mo

d. [

MP

a] ext.

cyt.

16

18

12

14

6pH

4

2

3

δ ma

x [n

m]

ext.cyt.

7 8 9 10

6 7 8 9 10pH

Figure 3.7: Influence of the pH on PM at 150 mM KCl. (A) Evolution of PM Youngsnormal modulus with pH, measured from extracellular and cytoplasmic side of PM. Curvesare fit using Chadwicks model. (B) Maximal tip indentation in PM, obtained from fit. (C)Side-specific force curves on PM at different pH.

95


tend to show that PM can be seen as a structure held together by specific protein-lipid

and lipid-lipid interactions within the extracellular leaflet. This scaffolding seems static

and is very resistant to pH or salt concentration changes of the surrounding buffer,

maintaining the well known stability of PM. The cytoplasmic leaflet, on the other hand,

contains very mobile inter-trimer phospholipids largely affected by salt and pH changes.

The mobility is due to non-specific long range electrostatic forces characterizing the

cytoplasmic leaflet. Further AFM experiments on PM and on asymmetric membranes

in general should take into account the different behaviour of the leaflets since results

obtained from both sides may not be similar.

3.6 Material and Methods

Purple membrane preparation

Wild type PM from Halobacterium salinarium strain S9 and the M163C-bR mutant

from strain L33 were grown by a standard method using a peptone media (L37, Oxoid,

Basingstoke, UK) and isolated and purified using established protocols (Oesterhelt &

Stoeckenius 1974). The membranes were either frozen at −25 C for storage (4 month

max.) or kept at 4 C for immediate use.

Sample preparation

PM was diluted to ∼ 15 µg/ml in the imaging buffer. A drop (30 µl) of PM solution

was adsorbed on freshly cleaved mica (9.9 mm mica discs, Agar Scientific, Essex, UK)

for 5-10 min and then gently rinsed with imaging buffer (2 ml). For work in liquid

with low salt concentration buffers (20 mM KCl to 100 mM KCl), the discs were

not rinsed because of the weak adsorption of PM on mica (33). For salt concentration

< 50 mM KCl PM adsorption was carried out by putting a drop (50 µl) of PM in

solution (50 mM KCl, 10 mM Tris − HCl) on freshly cleaved mica, and by diluting

96


with 10 mM Tris−HCl up to the required KCl concentration a few minutes later, in

order to prevent ill-formation or no adsorption of the membrane. Imaging was performed

after adding some more buffer (∼ 200 µl). For imaging in air, a drop (50 µl) of PM in

solution (50 mM KCl, 10 mM Tris−HCl) was adsorbed for 5− 10 min on mica, then

gently rinsed with ultra-pure water (2 ml of 18.2 MΩ, Maxima Ultrapure water system,

ELGA, High Wycombe, Bucks, UK) and dried at 20 C in a closed Petri dish. All the

buffers were made with ultrapure water and renewed every two weeks. Buffer chemi-

cals (KCl, HCl, Tris and NaOH) were purchased from Sigma-Aldrich (Sigma-Aldrich,

Dorset, UK). AFM imaging and force curves were acquired at 20± 1 C.

Imaging

High resolution images were recorded with a commercial Dimension 3100 AFM (Veeco,

Santa Barbara, CA) equipped with a 90 µm scanner operated in open loop. An AC

liquid-cell was used for imaging in buffer. Since AC-AFM is extremely sensitive to tip

or sample contamination (Moller et al. 1999) the liquid cell was sonicated in ultrapure

water, then cleaned with ethanol and finally rinsed with ultrapure water before imaging

in buffer. A new AFM tip was used for each sample. Imaging in liquid was carried out

with Olympus TR800 tips (SiN, nominal spring constant kn = 0.57 N/m, Olympus Ltd,

Japan). High resolution was achieved in 150 mM KCl 10mM Tris−HCl buffer at pH

8 (isoelectric point of the SiN (Biscan et al. 2000)) according to Moller et al (Moller

et al. 1999). Imaging in air was performed with Olympus AC240 tips (SiN, kn = 2 N/m,

Olympus Ltd, Japan). Before imaging, the system was left 2 or 3 hours scanning a blank

sample to reach thermal equilibrium. For each image, height, amplitude and phase in-

formation were acquired simultaneously. The scan speed was adjusted between 1 and 3

lines/sec for low magnification frames (> 400 nm) and between 4 and 9 lines/sec for high

resolution frames. Best results were obtained in ultrasoft AC, i.e. with a ratio of the

97


free oscillation amplitude A0 over the set-point amplitude A as close as possible to 1 and

at low oscillation amplitude (≤ 1 nm). The PM lattice was used for lateral calibration

of the scanner and quantification of the drift. Acquired images were corrected for drift,

but no averaging was done.

Force curves

Force curves were recorded in solution with a commercial MFP-3D AFM (Asylum Re-

search, Santa Barbara, CA) with a close loop in the x, y and z directions. Olympus

TR400 tips (SiN, nominal spring constant kn = 0.08 N/m, Olympus Ltd, Japan) were

used for increased sensitivity and to avoid sample damaging, especially in low salt buffer.

AC images of the sample were acquired before and after force curves were measured to

ensure reliability of the curves, and that the sample had not been damaged. In each

buffer condition, ∼ 300 curves were taken on both cytoplasmic and extracellular sides of

PM with the same cantilever and without re-engaging the AFM. Force curves where also

taken on mica before and after force curves acquisition on PM, in order to calculate the

inverse optical lever sensitivity (nm/V ) and to ensure the stability of the system. The

sampling rate was set to 300 nm/s. The z-piezo extension per curve was set to 100 or

150 nm in order to prevent excessive pressure on the membrane. The spring constant of

the cantilever was calculated using the thermal noise method (Butt & Jaschke 1995; Hut-

ter & Bechhoefer 1993), and theoretical predictions (Sader 1995). Thermal noise scans

always indicated a similar value for the cantilevers used (k = 0.11 N/m within < 10%),

suggesting a good reproducibility of the cantilever stiffness, but theoretical calculations

gave a spring constant of k = 0.085 N/m. Due to the discrepancy of the different force

calibration methods, the nominal value of 0.08 N/m was used to calculate forces, in-

cluding an error of 40%. In order to avoid systematic errors, each set of measurements

carried out in a specific buffer condition was made in a random order.

98


3.7 References

Antoranz Contera, Sonia, Lematre, Vincent, de Planque, Maurits R. R.,Watts, Anthony, & Ryan, John F. 2005. Unfolding and extraction of a trans-membrane a-helical peptide: Dynamic Force Spectroscopy and Molecular DynamicsSimulations. Biophys. J., 89, 3129–3140.

Benz, Marcel, Gutsmann, Thomas, Chen, Nianhuan, Tadmor, Rafael, & Is-raelachvili, Jacob. 2004. Correlation of AFM and SFA Measurements Concerningthe Stability of Supported Lipid Bilayers. Biophys. J., 86(2), 870–879.

Bergelson, L. D., & Barsukov, L. I. 1977. Topological asymmetry of phospholipidsin membranes. Science, 197, 224–230.

Biscan, J., Kallay, N., & Smolic, T. 2000. Determination of iso-electric point ofsilicon nitride by adhesion method. Colloids Surf. A, 165, 115–123.

Blaurock, Allen E, & King, G.I. 1977. Asymmetric structure of the purple mem-brane. Science, 196, 1101–1104.

Blaurock, Allen E, & Stoeckenius, Walter. 1971. Structure of the PurpleMembrane. Nat. New. Biol., 233(39), 152–155.

Butt, H. J., & Jaschke, M. 1995. Calculation of thermal noise in atomic forcemicroscopy. Nanotechnology, 6(1), 1–7.

Cevc, Gregor, & Marsh, Derek. 1987. Phospholipid bilayers: physical principleand models. Cell biology (Wiley), vol. 5. New York: Wiley.

Chadwick, R. S. 2002. Axisymmetric indentation of a thin incompressible elastic layer.SIAM J. Appl. Math., 62(5), 1520–1530.

Cleveland, J. P., Anczykowski, B., Schmid, A. E., & Elings, V. B. 1998.Energy dissipation in tapping-mode atomic force microscopy. Appl. Phys. Lett., 72(20),2613–2615.

Dimitriadis, E. K., Horkay, F., Maresca, J., Kachar, B., & Chadwick, R. S.2002. Determination of elastic moduli of thin layers of soft material using the atomicforce microscope. Biophys. J., 82(5), 2798–2810.

Domke, J., & Radmacher, M. 1998. Measuring the elastic properties of thin polymerfilms with the atomic force microscope. Langmuir, 14(12), 3320–3325.

Engel, Milton B., & Catchpole, Hubert R. 2005. A microprobe analysis ofinorganic elements in Halobacterium salinarum. Cell Biol. Intl., 29(8), 616.

99


Essen, L. O., Siegert, R., Lehmann, W. D., & Oesterhelt, D. 1998. Lipidpatches in membrane protein oligomers: Crystal structure of the bacteriorhodopsin-lipid complex. Proc. Natl. Acad. Sci. USA, 95(20), 11673–11678. Sulfoglycolipids inEC leaflet.

Fisher, K. A., Yanagimoto, K., & Stoeckenius, W. 1978. Oriented adsorptionof purple membrane to cationic surfaces. J. Cell Biol., 77(2), 611–621.

Gale, Paul, & Watts, Anthony. 1991. Characterization of phospholipid compo-sitions and physical properties of DMPC/Bacteriorhodopsin vesicles produced by adetergent-free method. Biochem. Biophys. Res. Comm., 180(2), 939–944.

Gale, Paul, & Watts, Anthony. 1992. Effect of bacteriorhodospin on orientationof headgroup od 1,2-dimyristoyl-sn-glycero-3-phosphocholine in bilayers: a 31P- and2H-NMR study. Biochim. Biophys. Acta, 1106, 317–324.

Garcıa, R., Tamayo, J., & San Paulo, A. 1999. Phase contrast and surface energyhysteresis in tapping mode scanning force microscopy. Surf. Interf. Anal., 27(5), 312–316.

Grigorieff, N., Ceska, T.A., Downing, K.H., Baldwin, J.M., & Henderson,R. 1996. Electron-crystallographic refinement of the structure of bacteriorhodopsin.J. Mol. Biol., 259, 393–421. electron diffraction study.

Haltia, Tuomas, & Freire, Ernesto. 1995. Forces and factors that contribute tothe structural stability of membrane proteins. Bioenergetics, 1228(1), 1–27.

Henderson, R., & Unwin, P.N. 1975. Three-dimensional model of purple membrane.Nature, 257, 28–32.


Hertz, H. 1882. Uber die Berhrung Fester Elasticher Korper. Journal fur die reine undangewandte Mathematik, 92, 156–171.

Houslay, Miles D., & Stanley, Keith K. 1982. Dynamics of biological membranes:influence on synthesis, structure, and function. New York: Wiley Interscience.

Hutter, Jeffrey L., & Bechhoefer, John. 1993. Calibration of atomic-forcemicroscope tips. Rev. Sci. Instr., 64(7), 1868–1873.

Israelachvili, Jacob N. 1991. Intermolecular and Surface Forces. Second editionedn. London: Academic Press.

Kamihira, M., & Watts, A. 2006. Functionally Relevant Coupled Dynamic Profile ofBacteriorhodopsin and Lipids in Purple Membranes. Biochemistry, 45(13), 4304–4313.

100


Killian, J. Antoinette, & von Heijne, Gunnar. 2000. How proteins adapt to amembrane-water interface. Trends Biochem. Sci., 25(9), 429.

Lanyi, Janos K. 2004. Bacteriorhodopsin. Annu. Rev. Physiol, 66(1), 665–688.

Lopez, Francesco, Lobasso, Simona, Colella, Matilde, Agostiano, An-gela, & Corcelli, Angela. 1999. Light-dependent and Biochemical Propertiesof Two Different Bands of Bacteriorhodopsin Isolated on Phenyl-Sepharose CL-4B.Photochem. Photobiol., 69(5), 599–604.

Luecke, H., Schobert, B., Richter, H.T., Cartailler, J.P., & Lanyi, J.K.1999. Structure of Bacteriorhodopsin at 1.55 Aresolution. J. Mol. Biol., 291, 899–911.location of STGA.

Luecke, Hartmut, Lanyi, Janos K., & Douglas, C. Rees. 2003. Structuralclues to the mechanism of ion pumping in bacteriorhodopsin. Adv. Prot. Chem., 63,111–130.


Muller, D. J., Schoenenberger, C. A., Buldt, G., & Engel, A. 1996. Immuno-atomic force microscopy of purple membrane. Biophys. J., 70(4), 1796–1802.

Muller, Daniel J., Fotiadis, Dimitrios, Scheuring, Simon, Mller,Shirley A., & Engel, Andreas. 1999. Electrostatically Balanced Subnanome-ter Imaging of Biological Specimens by Atomic Force Microscope. Biophys. J., 76(2),1101–1111.

Nachliel, E., Gutman, M., Kiryati, S., & Dencher, N. A. 1996. Protonationdynamics of the extracellular and cytoplasmic surface of bacteriorhodopsin in thepurple membrane. Proc. Natl. Acad. Sci. USA, 93(20), 10747–10752.

Neugebauer, D. C., Zingsheim, H. P., & Oesterhelt, D. 1983. Biogenesis ofpurple membrane in halobacteria. Methods Enzymol., 97, 218–26.


Oesterhelt, D., & Stoeckenius, W. 1974. Isolation of the cell membrane ofHalobacterium halobium and its fractionation into red and purple membrane. MethodsEnzymol., 31(Pt A), 667–678.

101


Op den Kamp, J. A. F. 1979. Lipid asymmetry in membranes. Annu. Rev. Biochem.,48, 47–71.

Porschke, D. 1996. Electrostatics and electrodynamics of bacteriorhodopsin. Biophys.J., 71(6), 3381–3391.

Renthal, Richard, & Cha, C. H. 1984. Charge asymmetry of the purple membranemeasured by uranyl quenching of dansyl fluorescence. Biophys. J., 45(5), 1001–1006.

Rothman, J. E., & Lenard, J. 1977. Membrane asymmetry. Science, 195, 743–753.

Sader, John Elie. 1995. Parallel beam approximation for V-shaped atomic forcemicroscope cantilevers. Rev. Sci. Instr., 66(9), 4583–4587.

San Paulo, Alvaro, & Garcıa, Ricardo. 2001. Amplitude, deformation and phaseshift in amplitude modulation atomic force microscopy: a numerical study for compli-ant materials. Surf. Sci., 471(1-3), 71–79.

Sanz, Carolina, Marquez, Mercedes, Peralvarez, Alex, Elouatik, Samir,Sepulcre, Francesc, Querol, Enric, Lazarova, Tzvetana, & Padros, Es-teve. 2001. Contribution of Extracellular Glu Residues to the Structure and Functionof Bacteriorhodopsin. J. Biol. Chem., 276(44), 40788–40794.

Sneddon, Ian N. 1965. The relation between load and penetration in the axisymmetricboussinesq problem for a punch of arbitrary profile. Internat. J. Engin. Sci., 3(1), 47–57.

Stark, Martin, Moller, Clemens, Muller, Daniel J., & Guckenberger,Reinhard. 2001. From Images to Interactions: High-Resolution Phase Imaging inTapping-Mode Atomic Force Microscopy. Biophys. J., 80(6), 3009–3018.

Subramaniam, S., Gerstein, M., Oesterhelt, D., & Henderson, R. 1993. Elec-tron diffraction analysis of structural changes in the photocycle of bacteriorhodopsin.EMBO J., 12(1), 1–8.

Voıtchovsky, Kislon, Antoranz Contera, Sonia, Kamihira, Miya, Watts,Anthony, & Ryan, J. F. 2006. Differential Stiffness and Lipid Mobility in theLeaflets of Purple Membranes. Biophys. J., 90(6), 2075–2085.


Zaccai, Giuseppe. 1987. Structure and hydration of purple membranes in differentconditions. J. Mol. Biol., 194(3), 569.

102

Probing PM specific interactions and ionic effects

4 Probing purple membranes specific molecular in-

teractions and ionic effects with AFM

4.1 Introduction

4.1.1 The formation of an AFM image

Despite being based on a conceptually simple principle, AFM relies on highly complex

and diverse interactions between the tip and the sample to deflect the cantilever. An

AFM image is formed through these interactions, representing a map of the tip-sample

forces over the scanned area. A consistent interpretation of AFM images can conse-

quently only be achieved if the nature of the main interactions responsible for deflecting

the cantilever is understood. Reciprocally, imaging a same sample under various condi-

tions can provide some insight in the nature, range and behaviour of the interactions gov-

erning the image formation. Understanding these forces subsequently allows an optimal

adjustment of the different parameters (scanning parameters, imaging medium composi-

tion and external parameters such as temperature, sample preparation, etc) in order to

utilize very steep force gradients and achieve the best possible resolution. Furthermore,

important information about the local force fields that govern biological function can

then be obtained.

The same investigation principle can be applied to understand the nature of potential

barriers responsible for the protein unfolding patterns obtained in force spectroscopy

experiments.


The previous chapter investigated the effect salt ions on Purple Membrane mechanical

properties and suggested that the observed leaflet-dependent lipid mobility relates to

the different types of interactions governing PM stability and bR dynamics. This chap-

103


ter aims at understanding these interactions at the molecular level and their specific

implications in controlling bR structure in order to enhance its activity. Using PM side-

specific AC force spectroscopy, bR unfolding experiments and high resolution imaging

with different ionic species and concentrations, functionally relevant interactions could

be observed and identified at the local and the global level. The results are interpreted

in terms of strategy adopted by PM to enhance bR proton uptake, release and transport

by the mean of a combination of specific steric interactions, electrostatic interactions and

ion binding.

The first section of the chapter gives an overview of the most commonly accepted theory

(so-called DLVO theory) for describing tip-sample interactions in a liquid medium. This

continuum theory, however, often fails to describe observations on biological samples,

especially at relevant salt concentrations and it is consequently important to develop

experimental approaches that can directly probe the local force field governing biological

function at the molecular level. In PM, this local force field is believed to govern bR pro-

ton uptake, release and transport. A method based on small amplitude alternate-contact

atomic force microscopy is proposed in the second part of the chapter, and applied to

characterize, through force spectroscopy and sub-nm resolution imaging, the complex

molecular interactions that govern these processes at local and global level. Using this

approach, it is possible to show that on the membrane cytoplasmic side, highly charged

mobile phospholipids and flexible and hydrated bR produce a dynamic force field that

effectively removes the order in the adjacent double-layer, producing very short-range

repulsive hydration interactions. On the extracellular side both lipids and bR are more

static, providing a flat and stiff surface that allows cation binding to specific locations and

produces a structured double-layer. We propose that the dynamic cytoplasmic double-

layer promotes ion-mediated proton uptake from the bulk solution while extracellular

protons are confined to a static double-layer that facilitates fast surface proton transfer

104


through a Grotthuss-like mechanism.

The third section of the chapter uses force spectroscopy to unfolded single bR molecule

from both protein termini and at different salt concentration. This approach allows iso-

lation of specific steric interactions from ionic related electrostatic contributions, and

relating these interactions to bR activity. The tryptophan residues, exclusively located

in bR extracellular half, use steric interactions to form a rigid scaffolding that stabilizes

and anchors bR into the membrane. Single ions can then specifically bind to extracellu-

lar loops, while the bR cytoplasmic part remains dominated by non-specific electrostatic

interactions that allow large structural rearrangements.

In the last section, the problem of PM ion specificity is addressed and the influence of

different salts on PM, namely LiCl, NaCl, KCl and CsCl, is investigated. Despite

being still at a preliminary stage, this last study already highlights important variations

in the condition required in the case of each salt to achieve high resolution. Molecu-

lar resolution of PM cytoplasmic side could be achieved in comparable conditions with

buffers containing LiCl, KCl and CsCl and structural differences can be observed. The

structural variations are discussed is terms of hydrated size of the different cations and

their affinity for ordering the surrounding water molecules. The differences observed

when imaging with the different ions reflect the importance of specific ions binding to

bR, shaping the protein surface in order to optimize the proton pumping activity.

4.2 General theory of molecular interactions: the AFM insight

4.2.1 The DLVO approximation

To date, the theory which has been the most widely used to interpret AFM data in

biology is the so-called Derjaguin-Landau-Verwey-Overbeek (DLVO) theory of colloidal

stability (Derjaguin & Landau 1941; Verwey & Overbeek 1948; Israelachvili 1991). This

105


theory, initially developed to describe interactions between colloids in a liquid, assumes

that the interaction forces can be approximated by a superimposition of van der Waal

attractive forces and electrical double layer repulsive contributions. The DLVO theory

treats the liquid as a continuum medium that is described by a single dielectric constant.

The double-layer forces are usually calculated reducing ions as point charges or uniformly

charged spheres. Both attractive and repulsive contributions are usually modelled with

power laws, typically using a Lennard-Jones potential.

van der Waals forces

van der Waals forces arise from the interaction of ions and dipole with polarized molecules.

They can include contributions from fixed or angle-averaged dipoles (Keesom forces),

dipole induced-dipole interactions (Debye forces) and interaction between spatially fluc-

tuating electron cloud distribution (London forces). London forces, also called dispersion

forces, are quantum mechanical in essence and require complex calculations for a rigorous

treatment (Landau et al. 1984; Mahanty & Ninham 1976). Their presence is ubiquitous

regardless of the conditions of the system considered. In the presence of charges, disper-

sion forces are, however, usually the weakest contribution to van der Waals interactions,

the main properties which can be described as follow (Israelachvili 1991):

1. They are long range forces effective on distances > 10 nm down to interatomic

distances ∼ 0.2 nm, depending on the conditions.

2. They can be repulsive or attractive and the dispersion forces between two molecules

can generally not be described by simple power laws.

3. Dispersion forces between two molecules are affected by the presence of surrounding

bodies and are consequently not additive.

4. Dispersion interactions not only bring molecules together but also tend to orient

106


and/or align them.

In the DLVO theory, van der Waals interactions are often assumed to be largely insen-

sitive to pH or to the electrolyte concentration of the liquid (Israelachvili 1991).

Electrical double layer forces

The electrical double layer forces between two identically charged surfaces originate from

the overlap of the counterions distributions near each surface. The electrolyte concen-

tration profile (co-ions and counterions) near a charged surface is usually described us-

ing the Gouy-Chapman theory which assumes point charges obeying Poisson-Bolzmann

statistics (Grahame 1953). The surface potential is related to the surface charge den-

sity, the temperature and the co/counter-ions valency (Grahame equation) (Israelachvili

1991). In the approximation of low potentials (. 25 mV ), the counterions density can

be approximated as a simple exponential decay when moving away from the surface

(Debye-Huckel approximation) and does not depend on the surface charge density. The

decaying distance is characterized by the so-called Debye length 1/κ where κ is given by:

κ =

√∑i

ρ∞ie2z2i /εε0kT (4.1)

with ρ∞i the density of the electrolyte i in the bulk solution, zi its valency, and 1/κ is in

meters. For monovalent ions in aqueous solution at 25 C, the Debye length is given by:

1/κ ≈ 0.304/√

C (4.2)

with C the concentration of ions and 1/κ in nanometers. Double layer repulsive forces

can be directly derived from the counterions density profile and generally follow an

exponential behaviour only within the Debye length (Ninham & Parsegian 1971) which

107


can be used as a first approximation for the effective non-linear range of double layer

forces.

4.2.2 Ions specific effects and hydration forces

In many model systems, the DLVO theory provides a good description of the interactions

measured (Pashley & Israelachvili 1984; Marra 1986; Smith et al. 1988), often down to

distances well below 1/κ. In several cases, however, deviations from DLVO predictions

have been observed, especially at high ionic strength where the Debye length is reduced

(Lips & Willis 1972; Behrens et al. 2000) or when the interacting surfaces offer particular

charge distributions (Bowen & Epstein 1979; Elimelech & O’Melia 1990). The presence

of a Stern and/or Helmoltz layer of counter ions (Israelachvili 1991; Bedzyk et al. 1990)

on the surfaces can seriously modify the DLVO predictions by shifting the potential near

the surface (Zeta potential) and require a refinement of the double-layer theory (Frens &

Overbeek 1972). At short distances, hydration, solvation and structural forces induced by

the interactions of ions with the surrounding molecules through hydrogen bonds can also

dramatically modify DLVO predictions which do not account for the discrete nature of

the molecule interacting, nor their specificity. The ordering of solvent molecules confined

between two approaching surfaces can induce well-known oscillatory forces (Horn &

Israelachvili 1981), and the hydration of salt ions can exhibit an ion-specific hydration-

repulsive behaviour, especially at high ionic strength (Pashley 1981a,b) with the range

and magnitude of the hydration forces following the hydration number (number of water

molecules interacting with the ion in aqueous solution) of the considered ion (Ohtaki &

Radnai 1993; Israelachvili 1991). The range of these forces is typically 0.1 − 1 nm and

the different ions hydration numbers rank according to the Hofmeister series (Hofmeister

1888) and relates to many effects essential and ubiquitous in biology such as salting in

and out effects (Setschenow 1889; Lanyi 1974; Collins & Washabaugh 1985; Collins 2004)

108


and chao-/kosmotropic effects of ions (Collins & Washabaugh 1985; Cacace et al. 1997).

These non-DLVO effects are important for proteins function and stability (Eppler et al.

2006; Dunach et al. 1988; Rosgen et al. 2005), and cannot be neglected when studying

the local forces governing proteins structure and activity.

4.2.3 Failure of continuum theories at physiological ionic concentrations

Besides hydration, oscillatory and structural forces, several other non-DLVO interac-

tions such as hydrophobic, steric and fluctuations forces (Leckband & Israelachvili 2001;

Woelki & Kohler 2004; Bostrom et al. 2003a) prevent the widely used continuum the-

ories1 to explain much of the experimental observations (Kunz et al. 2004), especially

at biological salt concentrations (above ∼ 100 mM) where the effects due to these forces

become important (Bostrom et al. 2001; Pashley 1981a,b). Furthermore, DLVO theory

separates electrostatic double-layer forces from attractive quantum mechanical electro-

dynamic fluctuation forces (van der Waals forces as described in the Lifshitz theory)

(Landau et al. 1984; Israelachvili 1991) whereas they are profoundly coupled (Kunz

et al. 2004). The most important failure, however, comes from the continuum liquid

hypothesis itself. At the molecular level, most of the biological process are governed by

specific interactions involving an ordered combination of chosen ions and oriented water

molecules (Bostrom et al. 2003b,a). In PM, the Hofmeister effects are essential for

the function and assembly of proteins (Cacace et al. 1997; Lanyi 1974; Perez-Jimenez

et al. 2004) as reflected by the asymmetric salt composition between the cytoplasmic

and extracellular side of the membrane (Engel & Catchpole 2005).

1 Born (self energy of an ion), Debye-Huckel (ions correlation free energy), Onsager Samaras (ionseffect on interfacial tensions) and DLVO (Israelachvili 1991; Kunz et al. 2004).

109


4.2.4 AFM as an alternative approach to study the molecular interactions

governing biology

The development of a theoretical model that can treat molecular interactions in a rele-

vant physiological environment and over multiple length scales remains one of the most

challenging task in biological physics. At the present time, the theories available often

fail to explain experimental data at physiological conditions, making predictions difficult

and unreliable. It appears therefore necessary to adopt an alternative approach in order

to investigate the force fields that control proteins structure and drive conformational

changes in response to external stimuli in fluid environment. The AFM’s ability to probe

local force fields and to quantify interactions at the sub-molecular level and in a liquid

environment allows a direct experimental approach to investigate the the local, discrete,

specific, short-range potentials that biology uses to tune interactions and function of

proteins in a molecular environment (Leckband & Israelachvili 2001). Furthermore, if

the imaging conditions are carefully selected, AFM can be used as a tool for studying

deviations from the DLVO theory (Uchihashi et al. 2005) by varying the composition of

the working medium.

4.3 Hydration force microscopy reveals the molecular forces

that trap and release protons in bacteriorhodopsin

4.3.1 Introduction

Despite an enormous amount of biochemical, genetic and spectroscopic investigations

dedicated to understand the vectorial proton-pumping mechanism of bR (§1.2), many

important questions remain unanswered, in particular neither the proton release nor the

proton uptake is understood (Lanyi 2000). These surface processes are controlled by very

complex solvent mediated local interactions in which the role of ions is expected to be

110


fundamental, especially since bR adaptation to its extreme and highly asymmetric native

environment has produced specificity of surface interactions between proteins, lipids and

ions (Oren 1999).

Usually AFM measurements of biological samples in liquid are performed in alternate

contact (AC) mode with feedback on amplitude, near the resonance frequency of the can-

tilever (on-resonance), at relatively high amplitudes (tapping mode). However tapping

mode images of bR have not reached sub-nm resolution (Moller et al. 1999). The high-

est resolution of proteins in fluid reported to date (0.5 nm) was inferred from DC-mode

(contact) measurements of bR, based on the DLVO approach in which the tip-membrane

interaction was treated as a balance of van der Waals and double-layer forces, therefore

sub-nm resolution was achieved only at a specific salt concentration (Moller et al. 1999;

Oesterhelt et al. 2000).

4.3.2 The Hydration Force Microscopy approach

In order to characterize and quantify the local (down to single α-helix level) and global

interactions that make possible proton uptake, release and transport at the surfaces

of bR and to study their non-DLVO character, a new approach based on AC-AFM was

developed. AC-AFM spatial resolution can be improved by confining the image formation

interaction to the frontmost atoms of the tip and by keeping the cantilever oscillations

stable. This can be achieved if the amplitude of the oscillation is very small, i.e. be

comparable to the range of rapidly decaying, short-range forces (Giessibl 2003) at the

surface of bR. If the amplitude is too large, the cantilever traverses highly non-linear

potentials arising from complicated combinations of attractive and/or repulsive forces,

the oscillations may become chaotic (Hu & Raman 2006), the resolution decreases and

quantitative information becomes difficult to obtain (Giessibl 2003; Hoffmann 2004).

111


Experimentally, the interactions responsible for high resolution image formation over

Distance tip-sample

0

0 1/κ

Tip-

sam

ple

forc

e

Outer Debyeregion

Inner Debyeregion

Fts

Repulsive hydrationdominated

van der Waalsdominated

Small amplitude AC-AFMSmall amplitude AC-AFMStandard tapping AFMStandard tapping AFM

Figure 4.1: Scheme of typical AFM tip-sample forces. The dominating repulsive or attractiveinteractions depend on the tip-sample system and the solvent. Ion specific hydration effectsbecome significant inside the inner Debye region. Small amplitude AC-AFM allows dynamicalprobing of the tip-sample interactions by oscillating along a locally monotonous force region.

PM were identified as the non-DLVO short-range repulsive hydration force (Israelachvili

1991; Leckband & Israelachvili 2001) (hence Hydration Force Microscopy, HFM). Using

very small amplitudes, images resolving details down to the interactions of single α-

helices at the cytoplasmic and extracellular sides of purple membrane were obtained in

a wide range of KCl concentrations (Fig. 4.2, 4.5). The conditions necessary for high

resolution reflect the characteristics of bR local force fields (tip-sample forces). These

conditions are:

1. Stable cantilever oscillations mapping short range interactions. At the liquid-

membrane interface, marginally chaotropic K+ and Cl− (Collins 2004) compete

with water to interact with purple membrane charged groups. Ionic size, charge

112


and concentration will determine the hydration and affect locally the surface po-

tential. To obtain a high resolution of the α-helices hydration and interactions

with ions it is necessary to adjust the tip oscillation amplitude to the range of the

local interactions; experimentally, we have determined that the Debye length 1/κ

provides an upper limit for the amplitude A, namely A < 1/κ, where 1/κ is given

by Eq. 4.2.

2. Optimized imaging energy dissipation. Experimentally, AC-AFM images of PM

are formed mainly by energy dissipated through tip-sample interactions (Pts) (Fig.

4.4.C). Pts reveals differences in the force field and local mechanical properties.

When A is small enough, the surface interaction potential probed by the tip can be

approximated as locally harmonic and tip-sample force Fts becomes (Hoffmann 2004):

Fts ≈ Fts0 − ktsz (4.3)

where z is the tip-sample distance and kts is the interaction stiffness. When the cantilever

oscillates near resonance and at small amplitudes, kts can be related to the free oscillation

amplitude A0, the phase shift relative to the driving signal ϕ and A (Hoffmann 2004):

kts = kA0 cos (ϕ)/A (4.4)

Then the second imaging condition translates as:

kts ≈ k (4.5)

This condition can be met by choosing k comparable to purple membrane stiffness and

adjusting A (Appendix 8.1.3), taking into account that the cytoplasmic leaflet is softer

113


than the extracellular leaflet and that both leaflets harden with increasing salt concen-

tration (Voıtchovsky et al. 2006).

4.3.3 Probing the local force field at the surface of Purple Membrane

In order to identify the global force interactions at the surface, amplitude and phase vs.

distance curves were simultaneously recorded at 30, 100 and 300 mM KCl respectively,

on both sides of the membrane and on mica. Applying Eq. 4.4, kts vs. distance z curves

were calculated, Fig. 4.3.A-C. A0 was the same as for high resolution images (Fig. 4.2),

at each salt concentration:

1. At 30 mM KCl, kts reflects the asymmetry of the membrane: the extracellular

leaflet is less charged than the cytoplasmic leaflet and exhibits longer range inter-

actions with the tip. Both leaflets show a strictly monotonic repulsive interaction,

which allows Eq. 4.3 to be applied, and therefore high resolution. On the other

hand, mica shows an attractive region (Fig. 4.3.A). This instability induces an

oscillation in the imaging AFM feedback as predicted in Fig. 4.1. This oscillation

becomes noise in mica images, Fig. 4.4.A. At this KCl concentration, DLVO is

still applicable except for the hydration layers immediate to the membrane.

2. At 100 mM KCl, Fig. 4.3.B shows kts obtained with a soft (∼ 0.39 N/m) and

the stiffer cantilever (∼ 0.57 N/m) that provided the high resolution images in

Fig. 1B. Both cantilevers show similar behaviour over the membrane, but the

softer cantilever feels non-linear interactions over mica. The extracellular leaflet

still shows a longer range interaction than the cytoplasmic leaflet.

3. At 300 mM KCl (Fig. 4.3.C), extracellular and cytoplasmic leaflets curves are

qualitatively similar to 100 mM curves but with a stiffer kts, in accordance with

(Voıtchovsky et al. 2006).

114


B C

D E F

30 mM 100 mM 300 mM

A

Figure 4.2: HFM 30 nm × 15 nm topography images of purple membrane in solution showingthe effect of KCl concentration on bR trimer structure and local solvent mediated interactions.A single trimer is highlighted in all the images. Conditions that maximize resolution: extra-cellular side at (A) 30 mM , A0 = 9 A, A = 5.3 A, k = 0.39 N/m; (B) 100 mM , A0 = 10 A,A = 6.7 A, k = 0.57 N/m; (C) 300 mM , A0 = 11 A, A = 3.7 A, k = 0.57 N/m. Cytoplasmicside at (D) 30 mM , A0 = 8.4 A, A = 5 A, k = 0.39 N/m; (E) 100 mM , A0 = 10 A, A = 8A, k = 0.57 N/m; (F) 300 mM , A0 = 11.6 A, A = 5.5 A, k = 0.57 N/m. D-F are displayedin 3D in the lower panels, the z-range is 5 Afor all the images. RMS roughness, reflecting bRprotrusion from the lipids is (A) 1.5 A, (B) 1.4 A, (C) 1.5 A, (D) 1.2 A, (E) 2.4 A, (F)2.4 A. Cytoplasmic trimers are malformed at low salt; as the salt concentration increases theyare more defined and protrude more from the lipid matrix. Extracellular trimers are distinctat any salt concentration and protrude less due to specific bR-lipid interactions. At high saltmonomers are barely distinguishable, probably due to K+ binding to extracellular loops. Theerror on the given cantilever stiffness k is rather large ∼ 30% (see §3.6) but the relative stiffnessdifferences between different cantilevers is < 10% from their thermal spectra (see Methods).

115


kts= - Ftsz

A

C

B

F

E

D

100 20 30 400

5

10

0

10

20

100 20

0

10

20

100

0

2

4

05

1510

0 10 20

100 20 30 40

100 20

100

0 10 20

0

2

4

6

0

4

6

12

02

64

kts

z

x 10-9 m

cytoplasmicextracellular

mica

x 1

0-2

N/m

x 1

0-1

6 J

/s

kts

z30 30

Pts

Pts

Pts

z

z

z

z

Pts

z

z

5

15

3030

x 10-9 m

30

kts

kts

Figure 4.3: Plots of tip-sample interaction kts (A-C) and corresponding Pts (D-F) vs. dis-tance z at different salt concentrations measured for extracellular and cytoplasmic sides com-pared with mica; the force fields correspond to Fig. 4.2 imaging conditions. Tip-sampleinteractions over purple membrane are monotonically repulsive, dominated by electrostaticsand repulsive hydration forces. The decay length of the double layer forces is longer for theextracellular leaflet reflecting purple membrane charge asymmetry. The shoulder detected inthe extracellular curves in (B-C) could be related to specific K+ interactions with PM.

116


The curves on both sides of purple membrane reveal a strictly and monotonically repul-

sive tip-membrane interaction; no attractive dispersion force was detected. We propose

that the exponentially decreasing short-range interaction detected is due to solvent me-

diated hydration repulsive force (Israelachvili 1991; Faraudo & Bresme 2005; Marcelja

1997). On a longer range (> 1 nm), the extracellular and cytoplasmic kts vs. distance

curves exhibit a different behaviour depending on the salt concentration and therefore

related to membrane electrostatics. The extracellular kts vs. distance curves show a

smoother transition from the electrostatic region into the hydration repulsion regime

than the cytoplasmic curves and present a distinctive shoulder at higher salt concentra-

tions. The fact that the best resolution at any salt concentration is obtained if A fulfils

condition 1 (generally A < 1 nm) shows that the short-range exponentially decaying

local interactions are responsible for the image formation. Small amplitudes near the

surface confine the image forming interaction to a very few atoms at the apex of the tip,

increasing the resolution. At small amplitudes, Pts can then be calculated by,

Pts =A2

0

2

kω0

Q

[A

A0

sin ϕ−(

A

A0

)2]

(4.6)

where Q is the quality factor. Although identical to the equation valid for high Q

(Anczykowski et al. 1999), Eq. 4.6 is derived assuming that most of the damping is

viscoelastic, regardless of Q (see Appendix 8.1). Pts vs. z profiles for both sides of

the membrane are shown in Fig. 4.3.D-F. The scanning set-point for the corresponding

images in Fig. 4.2 is marked in the Pts curves, indicating the necessity to dissipate more

energy on less charged extracellular leaflet than on the cytoplasmic leaflet, emphasizing

the role of local interactions in AFM image formation. Finally, a direct relation between

117


the phase and Pts can be obtained by introducing Eq. 4.4 and 4.5 in 4.6:

Pts =A2

0

2

kω0

Qcos ϕ(sin ϕ− cos ϕ) (4.7)

When approaching a sample, the tip can dissipate energy and ϕ decreases from π/2 (Pts =

0, away from the sample) down to a minimum of 3π/8 (maximum Pts) depending on the

tip-sample interactions. Cytoplasmic patches appear lighter (larger ϕ) than extracellular

patches in phase images (Fig. 4.4.C), reflecting the differences in energy dissipation on

both sides, which are due to their different Youngs elasticity modulus and surface charge

(Voıtchovsky et al. 2006).

B CA

Figure 4.4: (A) Topographic image of a purple membrane patch at 30 mM KCl. Image size1.14 µm × 1.14 µm. The non-linearity in the force field over mica (Fig. 4.3.A) produces a”noisy” image of the mica surface while the membrane cytoplasmic side appears free of noise,due to the monotonic repulsive force field. (B) 140 nm × 140 nm topographic image ofextracellular (left) and cytoplasmic (right) membrane sides; no obvious contrast visible. (C)Phase image corresponding to (B) showing a contrast between the extracellular side (darker)and the cytoplasmic (lighter) reflecting differences in energy dissipation during scanning.

4.3.4 The effect of salt imaged at sub-molecular resolution

Fig. 4.2 HFM images show the effect of KCl on purple membrane. On the extracellular

side, three distinct monomeric protrusions are distinguishable within the trimer at low

118


salt (30 mM). As the salt concentration increases, the trimers become more compact

and at 300 mM KCl only single trimeric protrusions are visible (Fig. 4.2.C). Salt con-

centration does not affect the monomer and trimer protrusions from the membrane (∼ 1

A) because the extracellular surface remains closely bound to the lipids through specific

interactions (Henderson et al. 1978) that are not electrostatic (Voıtchovsky et al. 2006).

Concomitantly, kts vs z curves on the membrane extracellular side (Fig. 4.3.A-C) present

a distinct shoulder visible only at higher salt concentrations (Fig. 4.3.B-C). We associate

this shoulder with an ordered ion-water double-layer structure at the membrane surface.

However, the length and shape of this shoulder (∼ 2 nm at 150 mM KCl and ∼ 2.4 nm

at 300 mM KCl) and its weak dependency on the salt concentration do not follow DLVO

predictions (1/κ ≈ 0.8 nm at 150 mM ; and 0.5 nm at 300 mM). The softer cantilever

employed can detect this double-layer at 150 mM KCl; the harder cantilever can only

detect the shoulder at higher salt concentrations (300 mM KCl). The salt concentra-

tion therefore affects more the stiffness than the structure of double-layer. The energy

dissipated Pts increases with presence of the shoulder; as the stiffness of the double-layer

increases, more energy is required to disrupt it (Fig. 4.3.E-F) as reflected also in the

phase image (Fig. 4.4). Consequently we propose that Fig. 4.2.B-C is a direct image of

the local ordered double layer topography at the membrane surface and specific extra-

cellular cation binding (Tuzi et al. 2001; Sanz et al. 2001). At the cytoplasmic side,

the trimeric arrangement of monomers is almost indiscernible at low salt, where proteins

and trimers are soft (Voıtchovsky et al. 2006) and malformed and bR α-helices are

visible (Fig. 4.2.D) and only protrude 1 A from the lipid matrix. As the salt concentra-

tion increases distinct trimers appear more detached from the surrounding lipids (Fig.

4.2.E-F). At 300 mM KCl individual α-helices become visible (Fig. 4.5.A) and pro-

trude approximately 2 A (approx. the size of a water molecule) from the membrane; this

suggests stabilization and hydration of individual helices at higher K+ concentration.

119


Increasing the salt concentration improves the shielding of the electrostatic repulsion

within cytoplasmic leaflet and increases its Youngs modulus (Voıtchovsky et al. 2006),

stabilizing the membrane. Like in other haloarchaea, in its native conditions the cyto-

plasmic surface is in contact with molar concentrations of KCl. The chaotropic nature

of weakly hydrated K+ (it reduces the surface tension of water) maybe important to

stimulate individual helix hydration (Collins 2004). Moreover there are many residues

with strong water binding affinity (glutamate and aspartate (Collins 2004)) and a clus-

ter of water binding moieties at the cytoplasmic surface has been proposed to act as a

proton-collecting antenna (Nachliel et al. 1996). The visibility of cytoplasmic monomers

at high salt may be possible due to non-specific character of cationic interactions, in-

teracting through water and inducing a global surface charge effect with water binding

tightly to the α-helices (Fig. 4.5.A).

4.3.5 Biological significance

Based on these observations, a general description of the forces at the surface of bR can be

proposed (Fig. 4.5.C). bR cytoplasmic surface is very negatively charged (Kimura et al.

1997) and undergoes large conformational changes during the photocycle (Kamikubo &

Kataoka 2005). The highly charged phospholipids located in the cytoplasmic leaflet are

mobile and interact non-specifically with cations and bR (Voıtchovsky et al. 2006). This

highly fluctuating electric field affects the double-layer; anisotropic long-range transla-

tional diffusion of water molecules and rotational motions of water parallel to PM have

been reported (Fitter et al. 1999). Here, the force spectroscopy and HFM images indi-

cate that the order in the double-layer is removed by the dynamic character of the electric

fields and that the strongest interaction is short range hydration repulsive, correspond-

ing to the hydration of α-helices and lipids (Fitter et al. 1999). This dynamic double-

layer would enhance proton mobility towards the proton-collecting antenna (Nachliel

120


et al. 1996) from the bulk solution. Theoretical calculations suggest that over 100 mM

ionic concentrations, proton uptake from the bulk can become the dominant mechanism

(Georgievskii et al. 2002a). On the other hand, the extracellular side is static, stiff,

interacts specifically with lipids (Voıtchovsky et al. 2006) and appears very flat (Fig.

4.2, 4.5). Its static character allows the presence of a stable and spatially defined force

field inducing the binding of cations to specific locations and the creation of an ordered

double-layer that can be directly observed in AFM images. The stability of the ordered

double layer increases with the salt concentration (Fig. 4.3). In purple membrane natural

hypersaline environment, this extracellular ordered double layer must present a robust

network of specifically bound cations and confined water molecules, favouring a fast mi-

gration of pumped protons within the double-layer (Fig. 4.5) through a Grotthuss-like

mechanism (Agmon 1995). In this way, the bacterium can keep pumped protons at its

surface (Fitter et al. 1999; Heberle et al. 1994) in order to transport them more ef-

ficiently (Georgievskii et al. 2002b) towards ATP-synthases that will, in turn, use the

protons to build much of the ATP that powers the cell. HFM provides a method for the

direct observation of the effect of ions on single molecules, quantifying and characterising

the local non-DLVO force fields necessary for biological function. The high spatial res-

olution of HFM combined with high temporal resolution of high-speed AFM will allow

the study of single protein function in real time (Antoranz Contera et al. 2006b).

4.3.6 Material and Methods

Purple membranes from Halobacterium salinarium (Sigma-Aldrich Co., St. Louis, M.,

U.S.A) were dissolved in buffers with different KCl concentrations (10 mM Tris and 30,

100 and 300 mM KCl, at pH 8) and deposited on mica surfaces following a procedure

detailed in §3.6. Experiments were carried out in solution with an AFM with close-loop

in x, y and z (MFP-3D, Asylum Research, Santa Barbara, CA, U.S.A.). SiN cantilevers

121


B

C

ion-waterdouble-layer

bulk solution

towardsATP synthase ?

Extracellular

Cytoplasmic

H+

A

H+

Figure 4.5: (A) 12 nm × 12 nm high-resolution image of the cytoplasmic surface of 3bR trimers at 300 mM KCl, individual hydrated α-helices can be clearly distinguished andidentified by comparison with (B). The resolution depends on the direction of the scanningwith respect to the monomers. (B) Scheme of cytoplasmic and extracellular (inset) of purplemembrane surfaces (pdb id.: 1m0l) showing the surface charges (red is negative, blue is positive,lipids are orange). (C) Proposed model for the role of molecular forces in proton uptake, releaseand transport. On the cytoplasmic side the dynamic and disordered double-layer promotesproton uptake from the bulk solution. On the extracellular side, the static and ordered doublelayer facilitates confined surface proton diffusion via Grotthuss-like mechanism.

122


were purchased from Olympus (Tokyo, Japan). SiN cantilevers of different stiffness

k were calibrated using their thermal spectrum. Despite a substantial uncertainty by

comparison with their nominal values (§3.6), the relative stiffness differences are reliable

within ∼ 10%. The AFM was operated in AC mode on-resonance, at a fixed frequency

and constant drive amplitude for both imaging and force spectroscopy. At pH 8 the

tip is weakly negatively charged. In Fig. 3, 50-100 measurements were averaged for

each curve. Extension and retraction velocity was 200 nm/s. X-axis alignment and zero

were determined using AC-deflection curves (Appendix 8.3). After the measurements,

high resolution AC-AFM images were obtained to verify that no permanent damage had

been made to the membrane. Set-points used for high resolution, indicated by arrows

(Fig. 4.3.D-F) are an over-estimation of the real scanning value because AC-curves are

acquired on a static sample. Schemes of bR structure in Fig. 4.5.B were prepared using

MOLMOL.

4.4 The intra-molecular forces underlying bacteriorhodopsin

structure and mechanics: Tuning electrostatic and steric

interactions

4.4.1 Introduction

To understand the molecular origins of the influence of salt ions on PM properties such

as differential leaflet stiffness (chap. 3), hydration of individual α-helices, the particu-

lar force field at PM extracellular surface (§4.3.3), and to identify the key interactions

controlling bR structure and activity, it is necessary to consider bR at a sub-molecular

level.

Recently, single molecule force spectroscopy has been extenisvely employed to study sta-

bility and anchoring of membrane proteins through their unfolding patterns (Oesterhelt

123


et al. 2000; Muller et al. 2002; Moller et al. 2003; Kedrov et al. 2004; Cisneros et al.

2005). Contrarily to soluble proteins for which the entropic loss induced by the protein

extension is often sufficient to explain the energy dissipated in the unfolding process

(Rief et al. 1997; Thompson et al. 2002; Bustamante et al. 1994), an important part

of the energy necessary to unfold a membrane protein accounts for its anchoring (An-

toranz Contera et al. 2005). Unfolding pathways of bR have been studied under different

pH and temperatures (Muller et al. 2002; Janovjak et al. 2003, 2004; Sapra et al. 2006;

Kessler & Gaub 2006). The magnitude and positions of the main potential barriers op-

posing the protein un/refolding have been proposed (Kessler et al. 2006), showing that

alpha-helices tend to unfold by pairs and emphasizing the role of extracellular loops in bR

stability. However, the complexity of the system still hinders a satisfying identification

and understanding of the interactions responsible for these barriers at a single amino acid

(aa) level and no information about the barriers nature nor about their relevance for bR

activity has been proposed. Importantly, the role of salt ions on the potential barriers

observed has not been addressed. The calculation of PM elasticity Young’s modulus at

different salt concentration showed that cytoplasmic leaflet is softer than the extracel-

lular leaflet and that both leaflets harden with increasing salt concentration (chap. 3).

This reflects the important impact of salt ions on protein-lipids and lipid-lipid electro-

static interactions within the cytoplasmic leaflet, but not on the extracellular leaflet.

In this study we directly probe the influence of salt ions on bR intra- and inter-molecular

interactions, by unfolding of a single bR molecule from both, its cytoplasmic and extracel-

lular termini (C-terminus and N-terminus respectively) at the salt concentrations known

to impose a differential stiffening of the membrane (20 mM to 40 mM KCl) (see chap.

3, (Voıtchovsky et al. 2006)). In this way, it is possible to unambiguously isolate specific

and steric interactions from salt-dependent electrostatic contributions shaping bR un-

folding patterns. The unfolding starts with the disruption of the retinal binding pocket.

124


The main unfolding barriers are steric and lie in the stiff and rigid extracellular half of bR

which gives a support for specific ion binding sites, identified in the extracellular loops

by the unfolding curves. The cytoplasmic side is dominated by non-specific ion-mediated

electrostatic interactions that increase its structural flexibility. These results effectively

connect the interactions determining bR structure, with the mechanical properties and

the force fields necessary to control bR activity.

4.4.2 Side specific unfolding of bacteriorhodopsin at different ionic concen-

tration

Unfolding bacteriorhodopsin from purple membrane cytoplasmic side

Fig. 4.6.a-c shows the unfolding patterns obtained by pulling from bR C-terminus (cy-

toplasmic curves) at different salt concentrations. Only curves corresponding to a full

unfolding of the protein (curves with their last force peak occurring beyond 60 nm) were

used (Oesterhelt et al. 2000; Muller et al. 2002). Since the tip can non-specifically bind

many of the amino-acids of the C-terminus, the curves were aligned using their second

unfolding step, around 25 nm away from the membrane (Oesterhelt et al. 2000; Muller

et al. 2002). At 20 mM KCl (Fig. 4.6.a), about 90% of the recorded cytoplasmic un-

folding events showed the characteristic four steps unfolding pattern reported previously

(Oesterhelt et al. 2000; Muller et al. 2002; Janovjak et al. 2003; Sapra et al. 2006)

(∼ 5 nm, ∼ 27 nm, ∼ 43 nm, ∼ 65 nm). Each step is single, apart from the first step

and the unfolding pattern is reproducible as shown by the statistical histogram. At 30

and 40 mM KCl (Fig. 4.6.b-c), the four main steps are still present on the curves his-

tograms, but new peaks corresponding to ionic-dependent interactions gradually appear

as KCl concentration increases, reflecting an improved bR stability due to better shield-

ing of bR intra and inter-molecular electrostatic repulsion (Voıtchovsky et al. 2006).

125


Unfolding from purple membrane extracellular side

Unfolding bR from its N-terminus located at PM extracellular surface (extracellular

curves, Fig. 4.6) presents a different pattern than from the cytoplasmic side (Kessler

& Gaub 2006). At low salt (Fig 4.6.d), no step is observed beyond ∼ 45 nm which

makes it difficult to establish a discrimination criterion equivalent to the criterion used

for cytoplasmic curves to retain only full unfolding events. At 20 mM KCl, statistical

analysis over all the unfolding events showed a trend comparable to the four unfolding

steps characteristic of the cytoplasmic curves: two main unfolding steps located around

25 nm and 45 nm respectively were visible in more than 60% of the unfolding curves

(Fig. 4.6.d, Table 4.1). Some curves (∼15%) also show a step around 3 nm. At 30

and 40 mM KCl, new steps appear, including a last step beyond 60 nm. The crite-

rion used for discriminating cytoplasmic could hence be applied to extracellular curves

(Fig. 4.6.e-f). These extracellular curves also exhibit the two steps present at low salt,

suggesting their presence to be a good discrimination criterion for extracellular curves

at 20 mM KCl. All the extracellular curves retained were aligned using the first of the

two steps (∼ 25 nm).

4.4.3 Fitting the unfolding curves with the Worm-Like Chain model

Fitting the unfolding curve steps with the worm like chain (WLC) model (Bustamante

et al. 1994) (persistence length of 0.4 nm, amino-acid length of 0.36 nm, (Rief et al.

1997; Muller et al. 2002)) provided an estimation of the number of amino acids (aa)

extended at each step. Using high resolution models of the protein (Luecke et al. 1999;

Mitsuoka et al. 1999), it was then possible to determine the particular location in the aa

sequence where the potential barrier responsible for the step occurs (Table 4.2, Fig. 4.8).

This was done assuming an aa length of 0.36 nm when extended (Rief et al. 1997) and

126


0 20 40 60 nm

200

0

0 20 40 60 nm

200

0

0 20 40 60 nm

0 20 40 60 nm

0

0

4

8

0 20 40 60 nm

20

mM

KC

l3

0 m

M K

Cl

40

mM

KC

l

0 20 40 60 nm

Cytoplasmic Extracellular

0.0

0.1

a

c

eb

d

f

0.0

0.1

pNpN

0

4

0

8

200200pNpN

freq. .10-2 freq. .10-2

10

0

0

4

0

4

0

200 200pNpN

88

freq. .10-2 freq. .10-2

freq.freq.

Figure 4.6: Statistical histograms of the unfolding steps position for both cytoplasmic andextracellular unfolding curves and at different salt concentrations. In each case, a representativecurve has been included as an example. At each salt concentration studied, the new stepsappearing have been fitted with Gaussian curves as an eye guide. The blue Gaussian curvesrepresent steps appearing at 20 mM KCl, the red curves steps appearing at 30 mM KCl andthe black steps appearing at 40 mM KCl

127


of 0.15 nm when ordered in a α-helix (Eisenberg 2003) (Fig. 4.7). In principle, due to

non-specific binding of the AFM tip to the protein (Oesterhelt et al. 2000), each curve

can represent an unfolding event carried out from a different aa of the bound terminal.

Aligning the curves on a common step (Muller et al. 2002) allows consistent statistics of

the inter-step distances, but every step position is shifted by the amount corresponding

to the number of aa between the terminus extremity and the aa from which the pro-

tein is actually unfolded. This shift is typically 10-15 aa depending on the tip-protein

interaction and on the nature of the terminal from which the protein is unfolded. It is

therefore necessary to identify the exact aa where the tip-binding happens in order to

determine accurately the aa positions corresponding to the steps observed (Table 4.2,

Fig. 4.7).

When unfolding bR from the C-terminus the tip binding shift was determined assuming

that the first step observed (first sub-step of the first main step Fig. 4.6.a) corresponds

to the position of the Lys216 residue which binds the retinal. A potential barrier cor-

responding to this residue has been reported (Muller et al. 2002) and simulations on

bacterioopsin suggest that even without the presence of retinal, a salt bridges network

involving Lys216 would oppose unfolding (Seeber et al. n.d.). The potential barrier

could also be justified by the retinal extraction from its rigid binding pocket pointing

again towards the Lys216 residue. This attribution provides a shift of 14 aa or 15 aa

depending on the protein model used (Table 4.2).

The binding shift of unfolding curves from the N-terminus were determined assuming

that the second step observed at low salt (∼ 26 aa, Table 4.2) is triggered by Trp80 (see

discussion) which provided a shift value of 7 aa to 8 aa. Such a shift is compatible with

the hypothesis that Trp10-12 are responsible for the first unfolding step (see discussion)

within the WLC model fitting uncertainty (Table 4.2).

128


CE

DB

G FA

LoutLout

Lout

LinLin

a b c

Figure 4.7: Unfolding principle. (a) The AFM tip binds to bR cytoplasmic terminus. The tipstarts to unfold the protein by stretching part of helix G up to a first potential barrier (Lys216).(b) When the pulling force is sufficient, the barrier is overcome and the helix can be stretch upto the next potential barrier (Trp189). (c) The effective protein length stretched is Lout + Lin

but the WLC only accounts for Lout. It is therefore necessary to know Lin (calculated from(Luecke et al. 1999) and (Mitsuoka et al. 1999)) in order to accurately determine the potentialbarrier location.

129


KCl conc. PM sidepulledfrom

Unfold.at-tempts

Averageappliedforce[nN ]

Sucessful un-fold. events(as % of at-tempts)

Full unfold.events (as% of unfold.events)

20 mMcytoplasm. 508 0.3 45 (8.8%) 40 (89%)extracell. 930 1.2 30 (3.2%) 19 (63%)

30 mMcytoplasm. 1174 3.2 109 (9.3%) 49 (45%)extracell. 1203 4 103 (8.6%) 53 (51%)

40 mMcytoplasm. 1507 6 111 (7.4%) 47 (42%)extracell. 1040 7.2 65 (6.3%) 25 (40%)

Table 4.1 Number of partial and full unfolding events at the different salt concentrationsstudied. The protein is unfolded from its both termini at each ionic condition. The in-crease of PM leaflets differential Young’s moduli with the salt concentration (Voıtchovskyet al. 2006) is reflected by the necessity of adjusting the force applied by the AFM tipto the membrane at each bR unfolding attempt (see text). ”Successful unfolding events”accounts for both partial and full unfolding of the protein, depending if the tip attachedto the terminus or to an inter-helical loop. The ”full unfolding events” represent onlycurves corresponding to the unfolding of the entire protein, as used for the histogramsin Fig. 4.6.

4.4.4 Interpretation of the unfolding patterns and discussion

Tip-protein binding and unfolding mechanism

In order to produce a non-specific attachment of single bR molecules to the AFM tip,

it is experimentally necessary to press with the tip on the membrane. The necessary

force producing non-specific binding depends on protein terminal targeted and on the

salt concentration (see Table 4.1). If the force is too low, almost no unfolding events

are observed (< 10/00 for cytoplasmic unfolding with 0.3 nN at 30 mM KCl, data

not shown). The minimum force required to induce bR-tip binding (Table 4.1) follows

the leaflet-dependant increase of PM stiffness with the salt concentration (Voıtchovsky

et al. 2006). This suggests that it is necessary to perturb intra-membrane interactions

to produce non-specific tip-protein binding. Nevertheless, the speed of the unfolding

130


process (50 nm/s) allows sufficient time for the unfolded protein to rearrange before the

first step is observed. When the pressure exerted by the tip on the membrane is sufficient,

the rate of successful binding/unfolding events hardly depends on the salt concentration,

but the percentage of full unfolding events recorded decreases as the ionic concentration

increases. Ions shield electrostatic repulsive interactions and consequently allow more

interactions to stabilize the protein (Voıtchovsky et al. 2006) which makes its unfolding

more difficult.

1. At 20 mM KCl, the membrane is soft and presents a reduced Young’s modulus

E (typically ∼25% of native value) (Voıtchovsky et al. 2006; Antoranz Contera

et al. 2006a). The interactions maintaining PM structure are non-electrostatic and

specific, and appear to be located mainly in the extracellular leaflet (Voıtchovsky

et al. 2006; Antoranz Contera et al. 2006a; Lehnert et al. 2002; Lee 2003) (Table

4.2, Fig. 4.6.a, d, Fig. 4.8.a-e, 4.8.n-p) which E is substantially larger than the

cytoplasmic leaflet E (Ecyto./Eextracell. ∼ 0.3). Although it is easier to pull proteins

out of the softer cytoplasmic leaflet, the positively charged Arg7 residue interacting

with the weakly negatively charged AFM tip (Biscan et al. 2000) may assist tip

binding to the extracellular terminal. Only the specific intra and inter-molecular

interactions maintaining the membrane cohesion (Voıtchovsky et al. 2006) are

responsible for the observed steps (Table 4.2, Fig. 4.6.a, 4.6.d).

2. At 30 mM KCl, Ecyto./Eextracell. is less important (∼ 0.6) (Voıtchovsky et al.

2006) allowing several electrostatic contributions to appear in the unfolding curves,

mainly corresponding to locations on the extracellular loops (Fig. 4.8.f, 4.8.i, 4.8.r,

4.8.t). The specific contributions observed at low salt are still present (Fig. 4.6.b,

4.6.e). Unfolding bR from the extracellular side of PM shows a percentage of

success almost as large as for cytoplasmic unfolding.

131


3. At 40 mM KCl, Ecytop./Eextracell. > 0.8 and most of the membrane electrostatic

repulsions are shielded (Voıtchovsky et al. 2006). This is reflected by the new

unfolding steps observed, corresponding exclusively to cytoplasmic locations where

most of bR charged residues are located (Fig. 4.8.k-m, 4.8.u-v). All the steps

observed at lower salt concentrations are still present.

cytoplasmic

AB C D

EFG

extracellular

30

mM

KC

l2

0 m

M K

Cl

40

mM

KC

l

Cytoplasmic unfolding Extracellular unfolding

a b c d e n o p

f g h i j q r s t

k l m u v

Figure 4.8: Cartoon of the potential barrier positions reported in Table 4.2 below, obtainedby fitting the step positions with the WLC model. The barriers are symbolized by sphere orellipsoid fitting the residue(s) suspected to be responsible for the step observed. At each saltconcentration, the steps are classified following the observation chronology when unfolding theprotein (Fig. 4.6 histograms) and the arrow showing the unfolding direction is the colour ofthe helix being unfolded. The colour code used for bR helices is showed at the bottom (idemas Fig. 4.7).

132


Cytoplasmic Extracellular

KCl conc. Step pos.(fromWLC)

Proposed cor-resp. aa/eventin the protein

Step pos.(fromWLC)

Proposed cor-resp. aa/eventin the protein

20 mM

4nm† (11aa) Lys216 (binds:234L, 232M)

2-4nm†,‡

(5-11aa)Trp10-12

12.5nm(35aa)

Trp189 (188L,186M)

25.9nm(72aa)

Trp80

31.2nm(87aa)

Trp137 (140L,137M)

48.1nm(134aa)

Trp136-137

51.2nm(142aa)

Trp80-86, Tyr83(84L, 83M)

76.3nm(214aa)

Trp10-12 (12L,10M)

30 mM

7.7nm (21aa) FG-loop (200L,200M) Glu

12nm†,‡

(33aa)Helix B kink

16.3nm†

(35aa)Pro186, helix Flooping

20.2nm(56aa)

BC-loop, Glu

35.1nm(87aa)

Extract. Trp137-138 cyto. surf.

58nm†,‡

(161aa)EF-loop

57.2nm(142aa)

BC-loop (65L,67M) Glu

71nm†,‡

(197aa)FG-loop, Glu204

78nm(216aa)

Glu231-233-236

40 mM21nm‡ (21aa) EF-loop 8nm‡ (22aa) Asp36-3844nm‡ (35aa) CD-loop (Asp103-

Glu104)34nm‡

(181aa)CD-loop (Asp)

65-70nm‡ Cyto. half helix B,Asp36-38

Table 4.2 Location of the unfolding steps obtained from fitting with the WLC modeland proposed corresponding potential barrier. At each salt concentration, only the newsteps appearing in the unfolding curves are reported in the table but the steps measuredat lower salt concentration are still present (see Fig. 4.6). For each step, the full length ofthe protein fraction unfolded (Lin+Lout, Fig. 4.7) is given in nanometres and in numberof amino acids extended (0.36nm/aa), obtained using two bR models (Luecke et al. 1999)(L) and (Mitsuoka et al. 1999) (M) when relevant. The proposed amino acid/potentialbarrier responsible for the step is also given. When pulling from PM cytoplasmic side,the last amino acid to which the tip attach is determined assuming that the first stepreported at low salt is triggered by Lys216. For extracellular unfolding, the position ofthe tip attachment is determined assuming that Trp80 is responsible for the first mainstep (see text).

133


† Measurements with reduced accuracy due to WLC fitting discrepancy (small fittingregion)‡ Steps with large discrepancy (> 3aa) in the position obtained from different curves.

Tryptophan residues determine the main unfolding steps through specific

steric interactions

Tryptophan (Trp) residues play an important role in bR stability (Cartailler & Luecke

2003; Mogi et al. 1989; Tuparev et al. 2000) and activity (Mogi et al. 1989; Wu et al.

1992; Schenkl et al. 2005; Petkova et al. 2002) and are the only large bR residues located

exclusively in PM extracellular leaflet (Luecke et al. 1999; Lee 2003) (Fig. 4.9). Three of

bR eight Trp residues (Trp86, Trp182 and Trp189) point mostly toward the inside of pro-

tein and compose part of the rigid retinal binding pocket (Luecke et al. 1999; Mogi et al.

1989; Pebay-Peyroula et al. 1997; Reat et al. 1998; Rothschild et al. 1989) essential to

control and catalyze the retinal photo-isomerisation process (Mogi et al. 1989; Petkova

et al. 2002; Tsuda et al. 1980). Steric interactions between the pocket residues and

the retinal are indeed responsible for the retinal stereospecific C13 = C14 isomerisation

and for the high quantum efficiency observed (Zgrablic et al. 2005; Nonella 2000; Becker

& Freedman 1985) and the Trp present are believed to play an important role in the

process (Schenkl et al. 2005; Rothschild et al. 1989; Weidlich et al. 1996). The other

Trp residues (Trp10, Trp12, Trp80, partially Trp137 and Trp138) point more toward the

outside of the protein (Luecke et al. 1999; Mitsuoka et al. 1999). They are not directly

involved in bR photo-activity (Rothschild et al. 1989) (apart for Trp137, (Mogi et al.

1989; Wu et al. 1992)) but seem to be involved in rigidifying the protein through a

hydrogen bonds network (Petkova et al. 2002; Reat et al. 1998; Neutze et al. 2002) and

are thought to anchor bR into PM (Cartailler & Luecke 2003; Voıtchovsky et al. 2006).

Studies on model peptides flanked with Trp have demonstrated the importance of these

134


bulky residues for anchoring the peptide into a lipid membrane (Antoranz Contera et al.

2005; de Planque & Killian 2003; de Planque et al. 2002) and shown that hydrophobic

interaction can be overcome by Trp-lipid interactions (de Planque & Killian 2003).

We propose that the extraction of Trp residues is responsible for the main unfolding

steps observed, especially at low salt where steric, specific interactions dominate the

membrane.

Unfolding from the C-terminal - When fitting cytoplasmic curves at 20 mM KCl

(Fig. 4.6.a) with the WLC model, the location obtained for each step can be interpreted

invoking resistance from a Trp (Table 4.2, Fig. 4.8.a-e), if the first step is assumed to

correspond to the position of Lys216 (See §4.4.3). The second sub-step of the first main

step then coincides with the position of Trp189. The second main step can be attributed

to the anchoring of helix E in the membrane by Trp137 and the third step with the

anchoring of helix C in the membrane by Trp80. Tyr83 forms a hydrogen bond with

Trp189 (Luecke et al. 1999) and could also play a role in the third step observed (see

next paragraph). The fourth step corresponds to the anchoring of helix A by Trp10 and

12 (Fig. 4.6.a, 4.8.a-e).

Unfolding from the N-terminus - Trp residues can also explain the steps observed at

low salt on extracellular curves. The second unfolding step (Fig 4.6.d) is assumed to be

caused by the extraction of Trp80 (see §4.4.3). This assumption suggests the first step to

be caused by Trp10 and 12. The low percentage of curves (15%) that presented this step

can be explained by the protein deformation induced by the tip when pressing on the

membrane and the shorter length of the N-terminus. The third step can be attributed

to Trp137-138. Helix F does not show any important contribution at low salt because

its Trp mostly points towards the inside of the protein fragment still remaining in the

membrane and the unfolding process, occurring from the cytoplasmic end of the helix,

will tend to bend the helix.

135


Thermodynamic considerations also support the role of steric interactions in the ob-

served steps and suggest the responsibility of bulky residues; independent studies of bR

unfolding have measured that the extraction force associated with the main steps de-

creases with temperature (Janovjak et al. 2003), and increases with the logarithm of the

pulling speed (Janovjak et al. 2004), suggesting a stick-slip protein pulling behaviour

(Antoranz Contera et al. 2005). If Trp steric interactions are responsible for the unfold-

ing steps, extracting these bulky residues from the membrane should indeed show the

stick-slip behaviour (Antoranz Contera et al. 2005) observed, characteristic of thermally

assisted bond rupture (Evans 2001; Dudko et al. 2003).

B C D E F GA

EW

W LA

LGT A

LM

GLGT L

V K GM

I

PRGTIE A G

G VSD P

DAK K

FT

A IT

T LV

PA

IAF T

MT

L SML

L GYG

L T M V P F GG EE

NPIY

W

12

74

Y F L

75

18980

194

86

A RYA

D WLF

TT

PL L

LL

DL AL

L VD

204

ADE

GTI LAI

V GA

DG

IM

I GTG

LVGA LT K V Y S

182

138

137

Y RFVW

WA

I S TA

AM

LYI

LYV

L PP GF

TS

K A E SM

10

216

RP E

VAS TFKV

LR

N VTVV

LWS A

Y

LI GS

E G AG

IV

PVVW

P L N IET L

LFM V

L D V SA K VG FG LI L

LRS R AI F

GEAEAP

EPSAGDGAAATSC-terminus

N-terminus

Figure 4.9: Scheme of bacteriorhodopsin amino acid sequence in the membrane. The trypto-phan residues are boxed; the lysine 216 (linking the retinal via a Schiff base) and the glutamatesinvolved in extracellular ion binding and are circled. The helices and the termini are labelled.

Possible steric contributions from other bulky residues

Other bulky residues like Tyrosine (Tyr), Phenylalanine (Phe) and Proline (Pro) could in

136


principle also play a role in the steps observed at low salt. They are however distributed

on both side of bR (Luecke et al. 1999; Mitsuoka et al. 1999) and can consequently

not fully account for the unfolding steps. Tyr residues could play a role similar to Trp

because of their size and ability to make H-bond, essential feature of Trp in bR activity

(Pebay-Peyroula et al. 1997; Heberle et al. 1994). Mutation studies showed that Tyr57

and 83 are important for bR function (Mogi et al. 1987; Faham et al. 2004) and in-

volved in the retinal binding pocket (Mogi et al. 1989). No unfolding steps were however

observed around these positions at low salt, apart from Tyr83 which forms a hydrogen

bond with Trp189 (Luecke et al. 1999) and could be responsible for the steps attributed

to Trp80-86. Mutations of Phe and Pro2 residues did not show any significant effect on

the protein (Seeber et al. n.d.; Faham et al. 2004).

Specific K+ binding to the Glu74-75 and Glu194-204 residues located on

bacteriorhodopsin extracellular inter-helical loops

When increasing the salt concentration from 20 to 30 mM KCl, new steps are observed,

some of which correspond to charged residue positions located on bR extracellular loops.

The increased membrane rigidity should favour ion binding to specific extracellular sites

(Reat et al. 1998; Neutze et al. 2002) by improving the stability of the electrostatic

interactions (Antoranz Contera et al. 2006a). Both extracellular and cytoplasmic un-

folding curves show two salt-dependent steps that can be attributed to bR extracellular

loops (BC and FG, Table 4.2, Fig. 4.8.f, 4.8.i, 4.8.r and 4.8.t). It has been shown that

ions specifically bind to the bR extracellular side (Antoranz Contera et al. 2006a; Tuzi

et al. 2001) and mutation studies have highlighted the interaction of Glu204 with Mn++

ions (Tuzi et al. 2001) and proposed Glu9 and Glu194-204 as potential cation binding

sites (Sanz et al. 2001). The present data suggest that Glu74-75 and Glu194-204 could

2 Pro50 and Pro86 form the centre for helix B and helix F kink respectively.

137


offer specific binding sites for the K+ ions. This is supported by the fact that the only

extracellular loops exhibiting salt-dependent interactions are the loops with Glu residues.

Glu194 and Glu204 are involved in the proton release mechanism (Balashov et al. 1997;

Garczarek et al. 2005), and specific ion binding could enhance this release. Furthermore

specifically bound ions would form an extracellular network (Antoranz Contera et al.

2006a) favouring lateral proton transfer towards ATP-synthases (Heberle et al. 1994).

Secondary electrostatic unfolding barriers are mostly located in bacteriorhodopsin

cytoplasmic region

The rest of secondary steps observed at 30 mM KCl depend on the side from which bR

is unfolded. These steps probably account for the different rearrangements of the protein

fragment remaining in the membrane, which depend on the previous steps and therefore

on the unfolding process.

Cytoplasmic curves show a step around 45 aa (helix F) which could be explained by

helix F looping around its known kinking point, Pro186 (Tuzi et al. 2001), (Fig. 4.8.g)

as predicted by simulation studies (Seeber et al. n.d.). A second cytoplasmic curve

step at 97 aa could correspond to the extraction of the Trp137-138 from bR cytoplasmic

side (Fig. 4.8.h). However, the step is not observed in extracellular curves, and it is

not possible to rule out the contribution of the DE-loop as reported in previous studies

(Muller et al. 2002; Kessler & Gaub 2006).

Extracellular curves show a step corresponding to the 33 aa location in helix B cy-

toplasmic half (Fig. 4.8.q), which could, with the step observed in the same region at

40 mM KCl (AB-loop, Table 4.2, Fig. 4.8.u) account for the resistance of the charged

residues located in this region (Asp36-38, Lys40-41). A second extracellular curve step

(∼ 161 aa) corresponds to the EF loop region (Fig. 4.8.s) and is confirmed by cytoplas-

mic curves at 40 mM KCl (see discussion below). Finally, steps corresponding to the

138


extraction of bR C-terminus have sometimes been observed at 30 mM KCl (Fig. 4.6.e,

Table 4.2) and reported in previous studies (Kessler & Gaub 2006) but the present lack

of supporting data (not observed at 40 mM KCl) does not allow a clear interpretation.

Further increasing of the ionic concentration up to 40 mM KCl reveals new steps all

corresponding to locations in bR cytoplasmic half (Fig. 4.8.k-m and 4.8.u-v). Some of

the new steps reported are already suggested in the steps histograms at lower salt con-

centration (Fig. 4.7.b-c and 4.7.e-f) highlighting the progressive shielding of electrostatic

repulsion by salt ions. However, due to the non-specific nature of the electrostatic inter-

actions within PM cytoplasmic leaflet (Voıtchovsky et al. 2006; Antoranz Contera et al.

2006a) these steps are sometimes poorly reproducible inducing broad step distributions.

Steps corresponding to locations around cytoplasmic loops (AB-loop and CD-loop) are

observable on both cytoplasmic and extracellular curves (Table 4.2, Fig. 4.8.m,u and

4.8.l,v). Several Asp and Arg residues involved in attracting H+ for bR pumping activity

(Luecke et al. 2003; Nachliel et al. 1996) are located precisely around/on these loops

and may form the salt bridges network (Seeber et al. n.d.) responsible for the steps

observed. Finally, a cytoplasmic curve step corresponding to a residue on bR EF-loop

can also be observed, confirming the same interaction observed at 30 mM on extracellu-

lar curves (Fig. 4.8.k, 4.8.s). The EF-loop presents an exposed glutamate (Glu161, Fig.

4.9) which probably interacts with salt ions.

4.4.5 Towards a functional picture of bR interactions

Extracellular Trp steric and specific interactions constitute bR fundamental scaffolding

and are therefore essential for the protein activity. bR extracellular rigidity achieved

through Trp interactions could play a dual role by supporting the retinal isomerisa-

tion setreoselectivity (Weidlich et al. 1996) and by controlling the transmission of the

photo-induced vibrational energy from the retinal to the protein (Schenkl et al. 2005;

139


Rothschild et al. 1989). The vibrational energy transmitted allows the Schiff base de-

protonation (Wu et al. 1992) by a partial disruption of the extracellular hydrogen bonds

network (Neutze et al. 2002). Furthermore, the vibrational energy could enhance the

large conformational changes occurring in the softer protein cytoplasmic region during

the second part of the photocycle (Reat et al. 1998; Vonck 1996) through amplifica-

tion of the small vibration amplitudes of the extracellular rigid network (Ferrand et al.

1993). The Trp scaffolding is indeed relatively static throughout bR photocycle (Reat

et al. 1998) which allows it to be anchored into the membrane through interactions

with extracellular lipids known to be tightly and specifically bound to the protein (Weik

et al. 1998; Henderson et al. 1978). Direct interactions between bR monomers are

however weak (Isenbarger & Krebs 1999) and to date no inter-protein interactions have

been observed in bR unfolding patterns (Sapra et al. 2006), regardless of photo-induced

structural changes on the protein (Muller et al. 2002). The bR rigid and stable extra-

cellular surface also offer ion specific locations to bind to the protein loops, creating an

ordered double layer (Antoranz Contera et al. 2006a) that could support a rapid lat-

eral migration of the released proton (Heberle et al. 1994) towards ATP-synthase sites.

PM cytoplasmic leaflet is, on the other hand, governed by highly dynamical, non-specific

electrostatic interactions. bR cytoplasmic half is rich in charged residues acting as anten-

nae for proton capture (Nachliel et al. 1996). Salt ions are necessary to screen mutual

protein charge repulsion, but no ordered layer can be established (Antoranz Contera

et al. 2006a) because of the dynamical nature of the electrostatic interactions involved.

This allows bR cytoplasmic half to adapt structurally and dynamically throughout the

photocycle in order to optimize proton uptake from the bulk and accommodate the large

structural rearrangements necessary to reprotonate the chromophore (Neutze et al. 2002;

Vonck 1996) and consequently increase the pumping efficiency .

140


4.4.6 Error estimation

There are two main sources of uncertainty in the determination of the potential barriers

corresponding to the different steps:

1. Curve fitting of small steps with the WLC model only allows a restrained fitting

region which limits the precision of the procedure, especially since certain steps

cannot be fitted well with WLC, introducing an uncertainty of up to ±1 nm (± ∼

3 aa) mainly at high salt and for the extracellular curves (Table 4.2). The poor

reproducibility of certain steps location, especially at high salt, also affects the

precision.

2. Non-specific binding of the protein terminal to the tip. This process, favoured

by mechanically pressing on the membrane, may induce some reversible structural

changes to the protein. This effect is more important on extracellular curves where

the shorter N-terminus makes the binding more difficult. Event-dependent effects

can however be discarded by the statistical analysis, and the coincident locations

of most of the interactions in extracellular and cytoplasmic curves allows a consis-

tent interpretation of the results (Kessler & Gaub 2006). Finally, suppressing the

electrostatic interactions at low salt simplifies the problem and the attribution of

the steps to the Trp residues can be done unambiguously.

4.4.7 Conclusions

To survive in their extreme environment, purple membranes have developed a sophis-

ticated strategy that exploits high ionic concentrations for enhancing bR activity. The

protein extracellular and cytoplasmic parts present very different, yet complementary

properties, which combination governs bR function. Side specific single molecule force

spectroscopy at different salt concentration revealed the nature of the interactions domi-

141


nating bR structure and function locally, and allowed, for the first time, relating unfold-

ing potential barriers with interactions involving specific aa and discrimination between

steric and electrostatic energy barriers.


PM of Halobacterium Salinarium (Sigma-Aldrich, Dorset, UK) were diluted in buffer

(30 µl of 10 mM Tris, 50 mM KCl, ph8) and adsorbed on freshly cleaved mica (Fisher

Scientific). In order to avoid miss-assembling of PM lattice, the final experimental salt

concentration was reached by further dilution with a buffer containing only Tris (pH8)

prior to imaging. Unfolding experiments were performed in static mode with an Asylum

MFP-3D AFM (Asylum Research Santa Barbara, CA) equipped with XYZ close-loop.

The approach speed between the tip and the sample was set to 200 nm/s. A dwelling

time of 1 sec with the AFM tip pressing with a constant given force on the membrane

(see Table 4.1) was allowed to favour non-specific attachment of a protein terminal to the

tip. The retraction speed was set to 50 nm/s (Oesterhelt et al. 2000) in order to allow

enough time for the attached protein to rearrange after an eventual deformation induced

by the tip pressure. The side of the membrane on which the experiment was performed

was determined prior to pulling by high resolution amplitude and phase imaging using

AC-mode AFM (Voıtchovsky et al. 2006; Antoranz Contera et al. 2006a). The sample

was imaged before and after pulling to ensure reliability of the unfolding curves ob-

tained. Both unfolding and imaging were done using silicon cantilever Olympus TR400

(Olympus, Tokyo, Japan) with nominal stiffness of 0.08 N/m, and at a temperature of

26 ± 1 C. Cantilever stiffness were calculated using their thermal spectrum (Butt &

Jaschke 1995) and geometrical considerations (Sader 1995). The determination of the

unfolding curves steps position, the Worm Like Chain model fitting of the steps and the

statistical analysis were carried out by routines programmed using Igor ProTM software

142


(Appendix 8.3).

4.5 Effects of different types of cations on Purple Membrane

4.5.1 Introduction

The the previous sections of this chapter and partly chapter 3 have addressed the essential

role played by salt ions in the interactions governing bR function and structure, and in the

protein interactions with the surrounding lipids in PM. All these studies have, however,

used only one salt: KCl. This choice was motivated by the fact that experimentally, ion-

mediated electrostatic interactions were found to be dominating mainly PM cytoplasmic

leaflet where this salt is naturally predominant. Furthermore, in order to be comparable,

quantitative results required the use of one type of ions, but with different concentrations.

In a natural environment, however, the different ionic species predominant in the solution

at PM cytoplasmic and extracellular surface (Engel & Catchpole 2005) strongly suggest

that ionic specificity is an important parameter for the pumping activity of bR and

potential lateral proton transfer along PM.

This section presents images of PM cytoplasmic side in solutions containing different

types of monovalent cations but with the same chloride anion. In this way, it is possible

to directly visualize the differences induced on PM by the different cations (Li+, K+

and Cs+) and to locate PM sites interacting preferentially with a specific type of ions.

This study is at its preliminary stage and the role of anions and multivalent ions will be

investigated in future work.

4.5.2 High resolution imaging PM in buffers containing different ions

Experimentally, the minimum ionic concentration required to image PM in the different

salt buffers studied with AC-AFM can vary importantly. The conditions are summarized

143


in Table 4.3.

Min. conc. Atomic rad. Hydrated rad. Hydration num. Res. achieved

Ion∗ (mM) (nm) (nm) (±1)

Li+ 150− 200 0.074 0.38 5-6 Molec.

Na+ > 200 0.102 0.36 4-5 —

K+ ∼ 20 0.138 0.33 3-4 Sub-molec.

Cs+ ∼ 20 0.170 0.33 1-2 Sub-molec.

∗Always combined with Cl− in the salt used for the different buffers employed.

Table 4.3 Minimum ionic concentration required when imaging PM (AC-AFM) andcharacteristic of the different cations used. The atomic radii, hydrated radii and thehydration numbers are adapted from (Shannon & Prewitt 1969; Israelachvili 1991). Foreach ions, the lifetime of interaction with water (exchange rate) is in the order of 10−9 swith a tendency for a cation lifetime to increase with its hydration number (Israelachvili1991). No high-resolution imaging was achieved in NaCl solution with the range of ionicconcentration tried (300 mM NaCl highest concentration tried).

The smaller cations: Li+ and Na+

The ionic concentrations necessary to adsorb and image PM shows a clear transition

between the the cations smaller (atomic radii) and larger than K+ (Table 4.3). When

using LiCl and NaCl buffers, the membrane adsorption on a mica substrate seems very

weak (Fig. 4.10) and high ionic concentrations are necessary for scanning by AC-AFM.

The scanning setpoint then requires to be as high as possible (A/A0 ∼ 1) in order not

to destroy/remove the membrane. The hydrated size of the Li+ and Na+ cations is

bigger than for K+ and Cs+ (Table 4.3) due to the larger number of water molecules

with which they interact (larger hydration number). In terms of protein stability and

144


BA

Figure 4.10: Pictures of PM in different buffers. In both cases, the ionic concentration of thebuffer is the minimum concentration allowing non-destructive scanning of the membrane. (A)60 nm × 60 nm image of PM 150 mM LiCl, the lattice is hardly visible. (B) 3.5 µm × 3.5µm of PM in 200 mM NaCl. Even at such a low magnification, the membrane tend to bydestroyed by the scanning.

ordering of the water layer at the immediate surface of the membrane, these cations are

considered kosmotropes (Collins 2004). This implies that they tend to favour the forma-

tion of hydrogen bonds with surrounding water molecules over direct ionic interactions

(e.g. salt bridges) with the protein/lipids (Cacace et al. 1997; Collins 2004). This effect

can indirectly reduce the membrane hydration (salting-out) which often translates as a

stability increase of the proteins through the entropy loss (Lanyi 1974; Collins 2004).

These considerations, however, do not account for the particular nature of PM and the

specific interactions of bR with cations, and the rule suffers many exceptions (Cacace

et al. 1997) with dependency on the ionic concentration (Ebel et al. 1999). The reduc-

tion of bR solubility and direct interactions with ions by salting-out actually destabilize

the protein, as reflected by Table 4.3 and Fig. 4.10. Furthermore, the fact that PM

seems more stable in a LiCl buffer than in a NaCl buffer despite a lower hydration

value for Na+ clearly demonstrates the presence of interactions involving specific ions.

145


The larger cations: K+ and Cs+

The K+ and Cs+ cations are generally considered chaotropic (Collins & Washabaugh

1985; Collins 2004), and unlike the smaller cations, the tend to disrupt ordered water

structure which allows them to interact more with the protein and the lipids directly.

Their better ability to stabilize PM is reflected in Table 4.3 and the effect of their direct

interaction with bR in the high resolution images that could be obtained (Fig. 4.11

below). According to Table 4.3, the lower hydration of Cs+ allows it to interact more

A B

Figure 4.11: 20 nm × 20 nm high resolution pictures of PM cytoplasmic side in 100 mM KCl(A) and 100 mM CsCl (B). Similar cantilevers and scanning parameters were used in bothcases. The highlighted trimers shows, in the case of CsCl many sub-molecular details and indi-vidual α-helices are visible, as in Fig. 4.5.A with a higher KCl concentration. This highlightsthe role of hydration forces in both the structure of PM and the tip-sample interactions

strongly with bR than K+ and to consequently further hydrate the protein by salting-in

effect. This is indeed observed when working at 100 mM ionic concentration with both

cations; the images of PM cytoplasmic surface in CsCl present a larger (more hydrated)

trimer and more sub-molecular details (Fig. 4.11.B). Further evidence of the stronger

K+ interaction with PM cytoplasmic surface can be found comparing Fig. 4.5.A and

Fig. 4.11.B; a higher concentration of K+ ions (300 mM) is necessary to provide an

146


image equivalent to that obtained with 100 mM Cs+. A stronger cation-mediated bR

hydration makes the protein protrude further out of the membrane and single inter-helical

loops become observable with AC-AFM. The interaction of cation with bR cytoplasmic

surface are however non-specific for the most (§4.4.4), inducing indirect hydration of

the membrane. The exchange rate of bound cations (with the protein) is consequently

relatively high (Table 4.3) and despite showing and actual hydrated size observable by

AFM, their direct observation seems unlikely.

4.5.3 Comparison and Discussion

A comparison of PM cytoplasmic side images obtained LiCl, KCl and CsCl is pre-

sented in Fig. 4.12. The trend described in the previous paragraphs is clearly visible,

with cations affecting and stabilizing more the membrane as their hydration number

decreases. Beside an increased hydration of bR (salting-in) by chaotropic cations, the

B CA

Figure 4.12: Comparison of 50 nm × 50 nm pictures of PM cytoplasmic surface in 150 mMLiCl (A), 100 mM KCl (B), and 100 mM CsCl (C). Reducing the size of the hydrated ionand their affinity for ordering water molecules (Cs+ is chaotropic whereas Li+ is kosmotropic)allows the visualization of more sub-molecular details.

membrane shows a general better ordering and the scanned patches appear flatter with

KCl and CsCl than with LiCl (Fig. 4.12) which suggests an overall better membrane

147


stability and cohesion with the chatropic cations. The molecular origin of the salt-

mediated PM cytoplasmic leaflet stability depends on the ability of salt ions to directly

interact with bR and PM lipids in order to allow shielding of electrostatic repulsions.

Furthermore, chaotropic ions will tend to disrupt ordered water-ions layers such as the

double layer observed on PM extracellular surface (§4.2.3) and favour bR proton up-

take from the bulk (§4.3.5). This dual role naturally achieved by K+ cations on PM

cytoplasmic leaflet is however not ideal for the membrane extracellular leaflet. First,

PM extracellular leaflet is less sensitive to salt changes (chap. 3) and dominated by

specific interactions (§4.4.5), suggesting that direct interaction with chatropic ions is

not necessary to maintain the leaflet stability. Second, some cations bind to bR specific

locations (§4.4.4) and to PM lipids (Szundi & Stoeckenius 1987), as part of the ordered

double layer naturally present on PM extracellular surface. The stability of this ordered

layer, believed to be involved in the transfer of pumped proton to ATP-synthase sites

(§4.3.5), would be enhanced in the presence of kosmotropic ions which promote water

structures and are more favourable candidates for specific binding. In H. salinarium, PM

extracellular surface is indeed in contact mainly with Na+ cations (Engel & Catchpole

2005) which allows PM to accentuates its asymmetrical strategy to uptake, pump, release

and transfer protons while improving the membrane stability. The fact that the cations

dominating the intra- and extracellular natural PM medium are weakly chao- respec-

tively kosmotropic3 allows PM to adapt easily to changes of its surrounding medium, as

demonstrated by the extracellular orderd double-layer already present in 100 mM KCl.

Unlike for most of the soluble and mammalian proteins stabilized by kosmotropic ions

(Cacace et al. 1997; Arakawa & Timasheff 1984; Collins 2004), the stability of bR and

more generally of PM depends on the salting in of its cytoplasmic leaflet by chaotropic

3 K+ and Na+ are considered to be weakly chaos- respectively kosmotropic since the limit is generallyadmitted to lie between the two ionic species (Collins 2004)

148


ions. This is consistent with the asymmetric composition of PM and the excess of acidic

bR residues at its cytoplasmic surface, requiring electrostatic shielding by ions (Szundi &

Stoeckenius 1989). The strong and non-specific interactions of chaotropic ions with bR

cytoplasmic surface (Szundi & Stoeckenius 1989) also supports the protein proton uptake

from the bulk intracellular medium by disfavouring the formation of ordered ion-water

structure. Such a structure is however possible on the membrane extracellular surface,

first because of the different nature of the interactions governing the leaflet, but also by

the kosmotropic nature of the ions naturally dominating PM extracellular medium.

Experimental limitations and future investigations

Previous work using contact mode AFM has reported that PM adsorption on mica is

favoured by kosmotropic ions (Muller et al. 1997) and adsorbed PM could be observed

in LiCl buffer with an ionic concentration as low as 20 mM . No high resolution imaging

was however reported, but these results suggest that it is the imaging technique (AC-

AFM) rather than the lack of adsorption which is responsible for the high ionic imaging

concentration required. Energy is dissipated by the tip into the membrane in order

to form an AC-AFM image (§4.1.1) and the minimum dissipation required seem very

sensitive to the buffer cations. The imaging ionic concentrations are however about ten

times less than the natural PM medium and the membrane should be expected to show

an increased sensitivity to ionic changes. Nevertheless, the observations tend to confirm

the hypothesis that salting-out ions destabilize PM cytoplasmic leaflet. It is however

necessary to carry out experiments on PM extracellular surface to confirm the ideas

developed above and to try to explain the sharp transition observed between Na+ and

K+ imaging conditions. This would also allow the study of PM extracellular ordered

double layer with kosmotropic ions.

149


Further imaging and force spectroscopy experiments varying the anionic species and with

divalent ions will provide more insight in the strategies adopted by halophilic proteins to

adapt to their environment and exploit the unusual conditions in order to enhance their

activity.


PM of Halobacterium salinarium (Sigma-Aldrich, Dorset, UK) were diluted in the dif-

ferent buffers following the same procedure as described in §4.3.6. AC-AFM imaging

was achieved with an Asylum MFP-3D AFM (Asylum Research Santa Barbara, CA)

equipped with XYZ close-loop. All the images were acquired using silicon cantilever

Olympus RC800 (Olympus, Tokyo, Japan) with nominal stiffness of 0.39 and 0.76 N/m,

at a temperature of 26 ± 1 C. And the imaging technique developed in §4.3.2 was

applied.

4.6 References

Agmon, Noam. 1995. The Grotthuss mechanism. Chem. Phys. Lett., 244(5-6), 456.

Anczykowski, B., Gotsmann, B., Fuchs, H., Cleveland, J. P., & Elings,V. B. 1999. How to measure energy dissipation in dynamic mode atomic force mi-croscopy. Appl. Surf. Sci., 140(3-4), 376.

Antoranz Contera, Sonia, Lematre, Vincent, de Planque, Maurits R. R.,Watts, Anthony, & Ryan, John F. 2005. Unfolding and extraction of a trans-membrane α-helical peptide: Dynamic Force Spectroscopy and Molecular DynamicsSimulations. Biophys. J., 89, 3129–3140.

Antoranz Contera, Sonia, Voıtchovsky, Kislon, & Ryan, J. F. 2006a. Hy-dration force microscopy reveals the molecular forces that trap and release protons ina bacterial proton-pump. Submitted.

Antoranz Contera, Sonia, Voıtchovsky, Kislon, Yamashita, Hayato, Uchi-hashi, Takayuki, Ando, Toshio, & Ryan, John F. 2006b. Small amplitudesub-molecular resolution AC-AFM and High-Speed AC-AFM of purple membranes

150


in fluid: non-continuum effects inside the double layer. Biophys. J. Suppl. MeetingAbstract.

Arakawa, Tsutomu, & Timasheff, Serge N. 1984. Mechanism of protein saltingin and salting out by divalent cation salts: balance between hydration and salt binding.Biochemistry, 23(25), 5912–5923.

Balashov, S. P., Imasheva, E. S., Ebrey, T. G., Chen, N., Menick, D. R.,& Crouch, R. K. 1997. Glutamate-194 to Cysteine Mutation Inhibits Fast Light-Induced Proton Release in Bacteriorhodopsin. Biochemistry, 36(29), 8671–8676.

Becker, Ralph S., & Freedman, Kenn. 1985. A comprehensive investigation ofthe mechanism and photophysics of isomerization of a protonated and unprotonatedSchiff base of 11-cis-retinal. J. Am. Chem. Soc., 107(6), 1477–1485.

Bedzyk, M. J., Bommarito, G. M., Caffrey, M., & Penner, T. L. 1990.Diffuse-double layer at a membrane-aqueous interface measured with x-ray standingwaves. Science, 248(4951), 52–56.

Behrens, S. H., Christl, D. I., Emmerzael, R., Schurtenberger, P., &Borkovec, M. 2000. Charging and Aggregation Properties of Carboxyl Latex Par-ticles: Experiments versus DLVO Theory. Langmuir, 16(6), 2566–2575.

Biscan, J., Kallay, N., & Smolic, T. 2000. Determination of iso-electric point ofsilicon nitride by adhesion method. Colloids Surfaces A, 165, 115–123.

Bostrom, M., Williams, D. R. M., & Ninham, B. W. 2001. Specific Ion Effects:Why DLVO Theory Fails for Biology and Colloid Systems. Phys. Rev. Lett., 87(16),168103.

Bostrom, M., Williams, D. R. M., Stewart, P. R., & Ninham, B. W. 2003a.Hofmeister effects in membrane biology: The role of ionic dispersion potentials. Phys.Rev. E, 68(4), 041902.

Bostrom, M., Williams, D. R. M., & Ninham, B. W. 2003b. Specific ion effects:The role of co-ions in biology. Europhys. Lett., 63(4), 610–615.

Bowen, Bruce D., & Epstein, Norman. 1979. Fine particle deposition in smoothparallel-plate channels. J. Coll. Interf. Sci., 72(1), 81.

Bustamante, Carlos, Marko, J. F., Siggia, E. D., & Smith, S. 1994. EntropicElasticity of λ-Phage DNA. Science, 265, 1599–1600.

Butt, H. J., & Jaschke, M. 1995. Calculation of thermal noise in atomic forcemicroscopy. Nanotechnology, 6(1), 1–7.

151


Cacace, M. G., Landau, E. M., & Ramsden, J. J. 1997. The Hofmeister series:salt and solvent effects on interfacial phenomena. Q. Rev. Biophys., 30, 241–277.


Cisneros, David A., Oesterhelt, Dieter, & Muller, Daniel J. 2005. ProbingOrigins of Molecular Interactions Stabilizing the Membrane Proteins Halorhodopsinand Bacteriorhodopsin. Structure, 13(2), 235.

Collins, K. D., & Washabaugh, M. W. 1985. The Hofmeister effect and thebehavior of water at interfaces. Q. Rev. Biophys., 18, 323–422.

Collins, Kim D. 2004. Ions from the Hofmeister series and osmolytes: effects onproteins in solution and in the crystallization process. Methods, 34(3), 300.

de Planque, M. R. R., Boots, J. W. P., Rijkers, D. T. S., Liskamp, R. M. J.,Greathouse, D. V., & Killian, J. A. 2002. The Effects of Hydrophobic Mismatchbetween Phosphatidylcholine Bilayers and Transmembrane α-Helical Peptides Dependon the Nature of Interfacially Exposed Aromatic and Charged Residues. Biochemistry,41(26), 8396–8404.

de Planque, Maurits R. R., & Killian, J. A. 2003. Protein-lipid interactionsstudied with designed transmembrane peptides: role of hydrophobic matching andinterfacial anchoring. Molec. Memb. Biol., 20(4), 271–284.

Derjaguin, B. V., & Landau, L. 1941. Theory of the stability of strongly chargedlyophobic sols and of the adhesion of strongly charged particles in solution of elec-trolytes. Acta Physicochim. URSS, 14, 633–662.

Dudko, O. K., Filippov, A. E., Klafter, J., & Urbakh, M. 2003. Beyond theconventional description of dynamic force spectroscopy of adhesion bonds. Proc. Natl.Acad. Sci. USA, 100(20), 11378–11381.

Dunach, M., Seigneuret, M., Rigaud, J. L., & Padros, E. 1988. Influence ofcations on the blue to purple transition of bacteriorhodopsin. Comparison of Ca2+

and Hg2+ binding and their effect on the surface potential. J. Biol. Chem., 263(33),17378–17384.

Ebel, C., Faou, P., Kernel, B., & Zaccai, G. 1999. Relative Role of Anionsand Cations in the Stabilization of Halophilic Malate Dehydrogenase. Biochemistry,38(28), 9039–9047.

Eisenberg, David. 2003. The discovery of the α-helix and β-sheet, the principalstructural features of proteins. Proc. Natl. Acad. Sci. USA, 100(20), 11207–11210.

152


Elimelech, Menachem, & O’Melia, Charles R. 1990. Kinetics of deposition ofcolloidal particles in porous media. Env. Sci. Technol., 24(10), 1528–1536.

Engel, Milton B., & Catchpole, Hubert R. 2005. A microprobe analysis ofinorganic elements in Halobacterium salinarum. Cell Biol. Intl., 29(8), 616.

Eppler, Ross K., Komor, Russell S., Huynh, Joyce, Dordick, Jonathan S.,Reimer, Jeffrey A., & Clark, Douglas S. 2006. Water dynamics and salt-activation of enzymes in organic media: Mechanistic implications revealed by NMRspectroscopy. Proc. Natl. Acad. Sci. USA, 103(15), 5706–5710.

Evans, Evan. 2001. Probing the relation between forcelifetimeand chemistry in singlemolecular bonds. Annu. Rev. Biophys. Biomol. Struct., 30(1), 105–128.

Faham, Salem, Yang, Duan, Bare, Emiko, Yohannan, Sarah, Whitelegge,Julian P., & Bowie, James U. 2004. Side-chain Contributions to MembraneProtein Structure and Stability. J. Mol. Biol., 335(1), 297.

Faraudo, Jordi, & Bresme, Fernando. 2005. Origin of the Short-Range, StrongRepulsive Force between Ionic Surfactant Layers. Phys. Rev. Lett., 94(7), 077802.

Ferrand, M., Dianoux, A. J., Petry, W., & Zaccai, G. 1993. Thermal Motionsand Function of Bacteriorhodopsin in Purple Membranes: Effects of Temperature andHydration Studied by Neutron Scattering. Proc. Natl. Acad. Sci. USA, 90(20), 9668–9672.

Fitter, J., Lechner, R. E., & Dencher, N. A. 1999. Interactions of HydrationWater and Biological Membranes Studied by Neutron Scattering. J. Phys. Chem. B,103(38), 8036–8050.

Frens, G., & Overbeek, J. Th G. 1972. Repeptization and the theory of electrocraticcolloids. J. Colloid Interf. Sci., 38(2), 376.

Garczarek, Florian, Brown, Leonid S., Lanyi, Janos K., & Gerwert,Klaus. 2005. Proton binding within a membrane protein by a protonated watercluster. Proc. Natl. Acad. Sci. USA, 102(10), 3633–3638.

Georgievskii, Yuri, Medvedev, Emile S., & Stuchebrukhov, Alexei A.2002b. Proton transport via coupled surface and bulk diffusion. J. Chem. Phys.,116(4), 1692.

Georgievskii, Yuri, Medvedev, Emile S., & Stuchebrukhov, Alexei A.2002a. Proton Transport via the Membrane Surface. Biophys. J., 82(6), 2833–2846.

Giessibl, Franz J. 2003. Advances in atomic force microscopy. Rev. Mod. Phys.,75(3), 949.

153


Grahame, David C. 1953. Diffuse Double Layer Theory for Electrolytes of Unsym-metrical Valence Types. J. Chem. Phys., 21(6), 1054.

Heberle, Joachim, Riesle, Jens, Thiedemann, Gerd, Oesterhelt, Dieter,& Dencher, Norbert A. 1994. Proton migration along the membrane surface andretarded surface to bulk transfer. Nature, 370(6488), 379.


Hoffmann, Peter M. 2004. Small-Amplitude Atomic Force Microscopy. Pages 3641– 3654 of: Dekker Encyclopedia of Nanoscience and Nanotechnology.

Hofmeister, F. 1888. Zur Lehre von der Wirkung der Salze II. Arch. Exp. Pathol.Pharmakol., 24, 247–260.

Horn, Roger G., & Israelachvili, Jacob N. 1981. Direct measurement of struc-tural forces between two surfaces in a nonpolar liquid. J. Chem. Phys., 75(3), 1400.

Hu, Shuiqing, & Raman, Arvind. 2006. Chaos in Atomic Force Microscopy. Phys.Rev. Lett., 96(3), 036107.

Isenbarger, T. A., & Krebs, M. P. 1999. Role of Helix-Helix Interactions in As-sembly of the Bacteriorhodopsin Lattice. Biochemistry, 38(28), 9023–9030.

Israelachvili, Jacob N. 1991. Intermolecular and Surface Forces. Second editionedn. London: Academic Press.

Janovjak, Harald, Kessler, Max, Oesterhelt, Dieter, Gaub, Hermann, &Muller, Daniel J. 2003. Unfolding pathways of native bacteriorhodopsin dependon temperature. EMBO J., 22(19), 5220–5229.

Janovjak, Harald, Struckmeier, Jens, Hubain, Maurice, Kedrov, Alexej,Kessler, Max, & Muller, Daniel J. 2004. Probing the Energy Landscape of theMembrane Protein Bacteriorhodopsin. Structure, 12(5), 871.

Kamikubo, Hironari, & Kataoka, Mikio. 2005. Can the Low-Resolution Struc-tures of Photointermediates of Bacteriorhodopsin Explain Their Crystal Structures?Biophys. J., 88(3), 1925–1931.

Kedrov, Alexej, Ziegler, Christine, Janovjak, Harald, Kuhlbrandt,Werner, & Muller, Daniel J. 2004. Controlled Unfolding and Refolding ofa Single Sodium-proton Antiporter using Atomic Force Microscopy. J. Mol. Biol.,340(5), 1143.

Kessler, Max, & Gaub, Hermann E. 2006. Unfolding Barriers in BacteriorhodopsinProbed from the Cytoplasmic and the Extracellular Side by AFM. Structure, 14(3),521.

154



Kimura, Yoshiaki, Vassylyev, Dmitry G., Miyazawa, Atsuo, Kidera, Aki-nori, Matsushima, Masaaki, Mitsuoka, Kaoru, Murata, Kazuyoshi, Hi-rai, Teruhisa, & Fujiyoshi, Yoshinori. 1997. Surface of bacteriorhodopsin re-vealed by high-resolution electron crystallography. Nature, 389(6647), 206.

Kunz, W., Lo Nostro, P., & Ninham, B. W. 2004. The present state of affairswith Hofmeister effects. Curr. Op. Coll. Interf. Sci., 9(1-2), 1.

Landau, L. D., Lifshitz, E. M., Pitaevskii, L. P., & Kearsley, M. J. 1984.Electrodynamics of continuous media. Landau and Lifshitz Course of Theoreticalphysics, vol. 8. Oxford: Pergamon.

Lanyi, Janos K. 1974. Salt-Dependent Properties of Proteins from ExtremelyHalophilic Bacteria. Bacteriol. Rev., 38(3), 272–290.

Lanyi, Janos K. 2000. Bacteriorhodopsin. Biochi. Biophys. Acta, 1460(1), 1.

Leckband, Deborah, & Israelachvili, Jacob. 2001. Intermolecular forces in bi-ology. Q. Rev. Biophys., 34(02), 105.

Lee, A.G. 2003. Lipid-protein interactions in biological membranes: a structural per-spective. Biochim. Biophys. Acta, 1612, 1–40.

Lehnert, U., Rat, V., Kessler, B., Oesterhelt, D., & Zaccai, G. 2002. Quan-tification of thermal motions in the purple membrane. Appl. Phys. A, 74(0), s1287.

Lips, A., & Willis, E. 1972. Low angle light scattering technique for the study ofcoagulation. J. Chem. Soc., Faraday Trans. 1, 69, 1226–1236.

Luecke, H., Schobert, B., Richter, H.T., Cartailler, J.P., & Lanyi, J.K.1999. Structure of Bacteriorhodopsin at 1.55 Aresolution. J. Mol. Biol., 291, 899–911.

Luecke, Hartmut, Lanyi, Janos K., & Douglas, C. Rees. 2003. Structuralclues to the mechanism of ion pumping in bacteriorhodopsin. Advances in ProteinChemistry, Volume 63, 111–130.

Mahanty, J., & Ninham, B. W. 1976. Dispersion Forces. London: Academic Press.

Marcelja, S. 1997. Hydration in electrical double layers. Nature, 385, 689–690.

Marra, J. 1986. Direct measurement of the interaction between phosphatidylglycerolbilayers in aqueous electrolyte solutions. Biophys. J., 50(5), 815–825.

155


Mitsuoka, Kaoru, Hirai, Teruhisa, Murata, Kazuyoshi, Miyazawa, Atsuo,Kidera, Akinori, Kimura, Yoshiaki, & Fujiyoshi, Yoshinori. 1999. Thestructure of bacteriorhodopsin at 3.0 A resolution based on electron crystallography:implication of the charge distribution. J. Mol. Biol., 286(3), 861.

Mogi, T., Marti, T., & Khorana, H. G. 1989. Structure-function studies on bacteri-orhodopsin. IX. Substitutions of tryptophan residues affect protein-retinal interactionsin bacteriorhodopsin. J. Biol. Chem., 264(24), 14197–14201.

Mogi, Tatsushi, Stern, Lawrence J., Hackett, Neil R., & Khorana,H. Gobind. 1987. Bacteriorhodopsin Mutants Containing Single Tyrosine to Pheny-lalanine Substitutions are All Active in Proton Translocation. Proc. Natl. Acad. Sci.USA, 84(16), 5595–5599.


Moller, Clemens, Fotiadis, Dimitrios, Suda, Kitaru, Engel, Andreas,Kessler, Max, & Muller, Daniel J. 2003. Determining molecular forces thatstabilize human aquaporin-1. J. Struc. Biol., 142(3), 369.

Muller, D. J., Engel, A., & Amrein, M. 1997. Adsorption of biological moleculesto a solid support for scanning probe microscopy. J. Struct. Biol., 119(2), 172–188.

Muller, Daniel J., Kessler, Max, Oesterhelt, Filipp, Mller, Clemens,Oesterhelt, Dieter, & Gaub, Hermann. 2002. Stability of Bacteriorhodopsina-Helices and Loops Analyzed by Single-Molecule Force Spectroscopy. Biophys. J.,83(6), 3578–3588.

Nachliel, E., Gutman, M., Kiryati, S., & Dencher, N. A. 1996. Protonationdynamics of the extracellular and cytoplasmic surface of bacteriorhodopsin in thepurple membrane. Proc. Natl. Acad. Sci. USA, 93(20), 10747–10752.

Neutze, Richard, Pebay-Peyroula, Eva, Edman, Karl, Royant, Antoine,Navarro, Javier, & Landau, Ehud M. 2002. Bacteriorhodopsin: a high-resolutionstructural view of vectorial proton transport. Biochim. Biophys. Acta, 1565(2), 144.

Ninham, Barry W., & Parsegian, V. Adrian. 1971. Electrostatic potential be-tween surfaces bearing ionizable groups in ionic equilibrium with physiologic salinesolution. J. Theoret. Biol., 31(3), 405.

Nonella, M. 2000. Electrostatic Protein-Chromophore Interactions Promote theall-trans -¿13-cis Isomerization of the Protonated Retinal Schiff Base in Bacteri-orhodopsin: An ab Initio CASSCF/MRCI Study. J. Phys. Chem. B, 104(47), 11379–11388.

156


Oesterhelt, F., Oesterhelt, D., Pfeiffer, M., Engel, A., Gaub, H. E., &Muller, D. J. 2000. Unfolding Pathways of Individual Bacteriorhodopsins. Science,288(5463), 143–146.

Ohtaki, Hitoshi, & Radnai, Tamas. 1993. Structure and dynamics of hydrated ions.Chem. Rev., 93(3), 1157–1204.

Oren, Aharon. 1999. Bioenergetic Aspects of Halophilism. Microbiol. Mol. Biol. Rev.,63(2), 334–348.

Pashley, R. M. 1981a. DLVO and hydration forces between mica surfaces in Li+, Na+,K+, and Cs+ electrolyte solutions: A correlation of double-layer and hydration forceswith surface cation exchange properties. J. Colloid Interf. Sci., 83(2), 531.

Pashley, R. M. 1981b. Hydration forces between mica surfaces in aqueous electrolytesolutions. J. Colloid Interf. Sci., 80(1), 153.

Pashley, R. M., & Israelachvili, J. N. 1984. DLVO and hydration forces betweenmica surfaces in Mg2+, Ca2+, Sr2+, and Ba2+ chloride solutions. Journal of Colloidand Interface Science, 97(2), 446.

Pebay-Peyroula, E., Rummel, G., Rosenbusch, J. P., & Landau, E. M. 1997.X-ray structure of bacteriorhodopsin at 2.5 angstroms from microcrystals grown inlipidic cubic phases. Science, 277(5332), 1676–1681.

Perez-Jimenez, Raul, Godoy-Ruiz, Raquel, Ibarra-Molero, Beatriz, &Sanchez-Ruiz, Jose M. 2004. The Efficiency of Different Salts to Screen ChargeInteractions in Proteins: A Hofmeister Effect? Biophys. J., 86(4), 2414–2429.

Petkova, A. T., Hatanaka, M., Jaroniec, C. P., Hu, J. G., Belenky, M.,Verhoeven, M., Lugtenburg, J., Griffin, R. G., & Herzfeld, J. 2002. Tryp-tophan Interactions in Bacteriorhodopsin: A Heteronuclear Solid-State NMR Study.Biochemistry, 41(7), 2429–2437.

Reat, Valerie, Patzelt, Heiko, Ferrand, Michel, Pfister, Claude, Oester-helt, Dieter, & Zaccai, Giuseppe. 1998. Dynamics of different functional partsof bacteriorhodopsin: H-2H labeling and neutron scattering. Proc. Natl. Acad. Sci.USA, 95(9), 4970–4975.

Rief, Matthias, Gautel, Mathias, Oesterhelt, Filipp, Fernandez,Julio M., & Gaub, Hermann E. 1997. Reversible Unfolding of Individual TitinImmunoglobulin Domains by AFM. Science, 276(5315), 1109–1112.

Rosgen, Jorg, Pettitt, B. Montgomery, & Bolen, David Wayne. 2005. Pro-tein Folding, Stability, and Solvation Structure in Osmolyte Solutions. Biophys. J.,89(5), 2988–2997.

157


Rothschild, Kenneth J., Gray, Daniel, Mogi, Tatsushi, Marti, Thomas,Braiman, Mark S, Stern, Lawrence J., & Khorana, H. Gobind. 1989. Vi-brational Spectroscopy of Bacteriorhodopsin Mutants: Chromophore IsomerizationPerturbs Tryptophan-86. Biochemistry, 28(17), 7052–7059.

Sader, John Elie. 1995. Parallel beam approximation for V-shaped atomic forcemicroscope cantilevers. Rev. Sci. Instr., 66(9), 4583–4587.

Sanz, Carolina, Marquez, Mercedes, Peralvarez, Alex, Elouatik, Samir,Sepulcre, Francesc, Querol, Enric, Lazarova, Tzvetana, & Padros, Es-teve. 2001. Contribution of Extracellular Glu Residues to the Structure and Func-tion of Bacteriorhodopsin. PRESENCE OF SPECIFIC CATION-BINDING SITES.J. Biol. Chem., 276(44), 40788–40794.

Sapra, K. Tanuj, Besir, Huseyin, Oesterhelt, Dieter, & Muller, Daniel J.2006. Characterizing Molecular Interactions in Different Bacteriorhodopsin Assembliesby Single-molecule Force Spectroscopy. J. Mol. Biol., 355(4), 640.

Schenkl, S., van Mourik, F., van der Zwan, G., Haacke, S., & Chergui,M. 2005. Probing the Ultrafast Charge Translocation of Photoexcited Retinal inBacteriorhodopsin. Science, 309(5736), 917–920.

Seeber, Michele, Fanelli, Francesca, Paci, Emanuele, & Caflisch,Amedeo. Sequential unfolding of individual helices of bacterioopsin observed inmolecular dynamics simulations of extraction from the purple membrane. Biophys.J., 91.

Setschenow, J. 1889. Uber die Konstitution der Salzlosungen auf Grund ihres Ver-haltens zu Kohlensaure. Z. phys. Chem., 4, 117–125.

Shannon, R. D., & Prewitt, C. T. 1969. Effective ionic radii in oxides and fluorides.Acta Crystal. B, 25(5), 925–946.

Smith, Christopher. P., Maeda, Mayumi, Atanasoska, Ljiljana, White,Henry S., & McClure, D. J. 1988. Ultrathin Platinum Films on Mica and theMeasurement of Forces at the Platinum/Water Interface. J. Phys. Chem., 92, 199–205.

Szundi, I., & Stoeckenius, W. 1989. Surface pH controls purple-to-blue transitionof bacteriorhodopsin. A theoretical model of purple membrane surface. Biophys. J.,56(2), 369–383.

Szundi, Istvan, & Stoeckenius, Walther. 1987. Effect of Lipid Surface Chargeson the Purple-To-Blue Transition of Bacteriorhodopsin. Proc. Natl. Acad. Sci. USA,84(11), 3681–3684.

158


Thompson, James B., Hansma, Helen G., Hansma, Paul K., & Plaxco,Kevin W. 2002. The Backbone Conformational Entropy of Protein Folding: Ex-perimental Measures from Atomic Force Microscopy. J. Mol. Biol., 322(3), 645.

Tsuda, Motoyuki, Glaccum, Moira, Nelson, Burr, & Ebrey, Thomas G.1980. Light isomerizes the chromophore of bacteriorhodopsin. Nature, 287(5780),351.

Tuparev, N, Dobrikova, A, Taneva, S, & Lazarova, T. 2000. Bacteriorhodopsinthermal stability: influence of bound cations and lipids on the intrinsic protein fluo-rescence. Zeitschrift fur Naturforschung. C, 55(5-6), 355–360.

Tuzi, Satoru, Hasegawa, Jun, Kawaminami, Ruriko, Naito, Akira, & Saito,Hazime. 2001. Regio-selective Detection of Dynamic Structure of Transmembranea-Helices as Revealed from 13C NMR. Biophys. J., 81(1), 425–434.

Uchihashi, T., Higgins, M., Nakayama, Y., Sader, J. E., & Jarvis, S. P. 2005.Quantitative measurement of solvation shells using frequency modulated atomic forcemicroscopy. Nanotechnology, 16(3), S49.

Verwey, E. J. W., & Overbeek, J. T. G. 1948. Theory of Stability of LypophobicColloids. Amsterdam: Elsevier.

Voıtchovsky, Kislon, Antoranz Contera, Sonia, Kamihira, Miya, Watts,Anthony, & Ryan, J. F. 2006. Differential Stiffness and Lipid Mobility in theLeaflets of Purple Membranes. Biophys. J., 90(6), 2075–2085.

Vonck, J. 1996. A Three-Dimensional Difference Map of the N Intermediate in theBacteriorhodopsin Photocycle: Part of the F Helix Tilts in the M to N Transition.Biochemistry, 35(18), 5870–5878.

Weidlich, O., Schalt, B., Friedman, N., Sheves, M., Lanyi, J. K., Brown,L. S., & Siebert, F. 1996. Steric Interaction between the 9-Methyl Group of theRetinal and Tryptophan 182 Controls 13-cis to all-trans Reisomerization and ProtonUptake in the Bacteriorhodopsin Photocycle. Biochemistry, 35(33), 10807–10814.


Woelki, Stefan, & Kohler, Hans-Helmut. 2004. Effect of dispersion forces onthe potential of charged interfaces. Chem. Phys., 306(1-3), 209.

Wu, S., Chang, Y., el Sayed, M. A., Marti, T., Mogi, T., & Khorana, H. G.1992. Effects of tryptophan mutation on the deprotonation and reprotonation kineticsof the Schiff base during the photocycle of bacteriorhodopsin. Biophys. J., 61(5),1281–1288.

159


Zgrablic, Goran, Voıtchovsky, Kislon, Kindermann, Maik, Haacke, Ste-fan, & Chergui, Majed. 2005. Ultrafast Excited State Dynamics of the ProtonatedSchiff Base of All-trans Retinal in Solvents. Biophys. J., 88(4), 2779–2788.

160

Molecular resolution with High-speed AFM

5 Real-time observation of membrane proteins func-

tion using high-speed AFM

5.1 Introduction

5.1.1 Time resolution in biology

Biological process can happen on a very broad range of timescales, depending on the

level at which changes are considered. Generally, a relevant description of any mecha-

nism requires consideration of several timescales simultaneously (Elber 2005; Berg et al.

2002). Atomic and electronic motions within molecules typically range from atto- to

femtoseconds (10−18 − 10−15 s), whereas structural changes happening at the level of

large biomolecules typically take milliseconds and process involving whole cells or micro-

organism can take hours. In essence, any biological process originates from changes

occurring at the atomic level, but the enormous disparity in timescales (up to 10 order

of magnitude) between such changes and their consequences at the molecular/organism

level prevents most of the bottom-up approaches to describe the entire process. Sev-

eral investigation techniques have been developed to probe the dynamics of biological

mechanisms on different timescales. The most common experimental approaches use

time-resolved spectroscopy (IR, visible, UV, and X-ray diffraction), but important infor-

mation can also be derived from nuclear magnetic resonance (NMR), electron paramag-

netic resonance (EPR) and fluorescence resonant energy transfer (FRET) for lower time

resolution (Coppens & Novozhilova 2002, Special issue: ”Time resolved chemistry”).

Theoretical approaches, mainly molecular dynamics simulations, can also provide signif-

icant and very detailed information, especially when this information is not or difficultly

accessible experimentally.

Time-resolved spectroscopy has been extensively applied to the study of protein dy-

namics. This technique interrogates the investigated system by sending a first pulse of

161


photons characterized spatially, temporally and energetically1 and probing its effect on

molecular electronics with second pulse reaching the sample with a given, adjustable

delay (Chen et al. 1997; Cubeddu et al. 2002). Different types of information about the

protein dynamics such as movements of its secondary structure (Kim et al. 2003), elec-

tronic rearrangements (Schenkl et al. 2005; Chen et al. 1999) and molecular vibrations

(Williams et al. 1996; Brixner et al. 2005) can be obtained by tuning the wavelength of

the exciting and probing photons, and consequently the nature of the electronic reaction

induced and observed. Recently, some studies have used X-ray photons as a probing pho-

tons and achieved time-resolved X-ray crystallography of proteins (Genick et al. 1997).

However, despite being operated at near-room temperature, time-resolved X-ray crystal-

lography still requires three dimensional protein crystals, but with the proteins still able

to function normally (Moffat 2002). The temporal resolution achieved by time-resolved

spectroscopy range typically from femto- to nanoseconds, depending on the experimental

setup, the type of experiment run and the system studied.

NMR and EPR can, similarly to time-resolved spectroscopy techniques, use pulsed

electromagnetic fields to extract dynamical information from a biological system and

probe timescales ranging typically from pico- to milliseconds (Levanon & Mobius 1997;

Palmer 1997). As for continuous wave measurements, the sample requires labelling and

very detailed structural information can be obtained (see §1.3.1).

FRET can provide structural information of single proteins in a native environment with

details about sub-molecular movements (Michalet et al. 2006), but it requires the stud-

ied system to be relevantly tagged with fluorophores. The temporal resolution achieved

depends, in essence, on the resonant transfer efficiency and on the fluorescence lifetime of

stimulated fluorophores which typically lasts for nanoseconds. Practically, however, the

1 This is typically achieved using pulsed lasers. Since the photons obey quantum mechanical rules,their energy and their spatio-temporal characterization is however limited by Heisenberg uncertaintyprinciple.

162


actual temporal resolution achieved is limited by the detection system and is typically

in the order of milliseconds (Michalet et al. 2006).

Molecular dynamics simulation of a proteins constitute the main theoretical ap-

proach to investigate structural changes induced by internal (e.g. hydrophobic mismatch,

thermal fluctuations) or external constrains (e.g. mechanical stress) to the simulated sys-

tem (Karplus & McCammon 2002; Sansom et al. 2002; Isralewitz et al. 2001). The

strength of this approach is to explain the physical basis of molecular processes by re-

lating specific interactions between atoms to mechanisms at the molecular level. This

technique is however very costly in terms of calculation power, especially for simulating

whole proteins in a relevant environment. Currently the timescale of simulation is typi-

cally reaches tens of nanoseconds (Daggett 2000).

These techniques can follow specific mechanism of the studied biomolecule dynamics in

great details and on different timescale, but they can rarely provide an extensive and

global real-time picture of the protein structural rearrangements on the timescale that

governs its activity. Furthermore, most of them are indirect techniques and often require

averaging over many proteins or events. The development of high-speed AFM (hsAFM)

opens a new way among experimental approaches, by allowing direct real-time imaging

of single proteins activity in native-like conditions (Ando et al. 2001). Like for tradi-

tional AFM, no particular sample preparation such as labelling is required, apart from

adsorption for scanning.

5.1.2 High-Speed AFM (hsAFM)

Traditional AFM offers a whole range of possibilities to study single molecules in various

conditions (see §1.3.1). The main limitation of this techniques comes from its lack of

temporal resolution; typically the acquisition of an image takes 30 - 150 s depending on

the resolution and the scan speed. However, to obtain a relevant temporal information

163


when studying biological systems, it is necessary to reach at least sub-second imaging.

Most of the biological process at the level probed by AFM (typically 1 nm) indeed occur

over milliseconds.

During the last decade, several studies have extended the temporal capabilities of tra-

ditional AFM, adopting different strategies such as direct actuation of the cantilever

(Manalis et al. 1996; Sulchek et al. 2000b,a), reduction of the cantilever size (Viani

et al. 1999), the use of an active Q-control (Sulchek et al. 2000b), combined feedback-

feedforward controllers (Schitter et al. 2004a,b) and the use of a micro-resonator as

scan stage (Humphris et al. 2005). hsAFM is however still an emerging field, subject

to constant fundamental developments (Fantner et al. 2006; Uchihashi et al. 2006) and

few results about the dynamics of single biomolecules are available (Viani et al. 2000).

A major improvement has however been obtained by the group of T. Ando which devel-

oped a hsAC-AFM capable of imaging in liquid single myosin V molecules moving onto

mica (Ando et al. 2001). This hsAC-AFM system was successfully used on other bio-

logical systems such as DNA cleavage by enzymes (Yokokawa et al. 2006) and imaging

of structural changes of single bR molecules during their photcycle (§5.3.2 following).

This enormous potential of hsAFM for medical, technological and fundamental research

opens a complete new world of possibilities for investigating single molecule processes

in real time and under physiological conditions, without requiring any particular sample

preparation such as labelling or crystallization.


The first part of this chapter describes practical and theoretical considerations inherent

to the achievement of simultaneous high spatial and temporal resolution with hsAFM.

Key hardware components of the hsAFM are briefly reviewed and a simple model pro-

viding a theoretical scan-speed limit for achieving a given resolution is proposed. The

164


applicability to hsAFM of the high-resolution method developed in §4.3.2 for traditional

AC-AFM imaging of PM is also discussed. Illustration of the hsAFM possibilities such

as direct mapping of PM extracellular ordered layer (see §4.3.3 for details) and imaging

of PM under dynamical pH changes are exposed. High spatio-temporal resolution of

both PM surfaces is finally demonstrated, showing sub-molecular bR details captured in

liquid at scanning rates > 50 frames/s.

In second part of the chapter, single bR photo-induced conformational changes are di-

rectly imaged, showing a large twist of helices E and F away from the rest of the protein

during the second part of the photocycle. The changes observed could also happen in

the dark (thermal isomerization) but more rarely. Only one protein per trimer is seen to

undergo conformational changes at a given time. The result is discussed in terms protein

cooperativity/heterogeneity and compared with data from other techniques such as with

X-ray, neutron and electron diffraction and with time resolved spectroscopy.

5.2 High-resolution at high speed

5.2.1 Practical considerations

The hsAFM used for this work was developed by Ando’s group in Kanazawa, Japan.

Since the first published result (Ando et al. 2001), the hsAFM has been further op-

timized (Ando et al. 2006), with each component designed to maximize the imaging

bandwidth.

A dynamic proportional, integral and derivative (PID) feedback control was developed

using analog based electronics (Ando et al. 2001, 2003) in order to react fast enough

to changes of tip-sample interactions. Recently a feedforward compensation system was

implemented (Uchihashi et al. 2006) to minimize the damage caused by the tip to bio-

logical samples.

165


The scanner constitutes the main bandwidth limiting factor (Ando et al. 2001). The

geometry of the scanner and the stiffness of its material will determine the resonance

frequencies which have to be avoided when scanning (Ando et al. 2001; Kodera et al.

2005). The piezo-actuators also possess intrinsic resonances that have to be considered,

and the scanner was specially designed to minimize unwanted resonances. The devel-

opment of an active damping system (Kodera et al. 2005) further contributed to the

scanner improvement.

The scanning cantilevers constitute an important part of the hsAFM and their fabrica-

tion presents a considerable challenge. On the one hand, high-speed AC imaging requires

the AFM tip to oscillate as fast as possible in order to obtain enough oscillation cycles

per scanned surface, suggesting the use of stiff material for the cantilever. On the other

hand, biological samples require soft cantilevers to achieve non-destructive imaging. The

solution lies in the reduction of the cantilever size, as suggested by previous studies (Viani

et al. 1999). The cantilever used for hsAFM are typically a 10 µm long, 2 µm wide

and 140 nm thick silicon nitride beam back coated with gold and further coated with

osmium (Ando et al. 2001, 2003). Such cantilevers can achieve resonance frequencies of

500− 800 kHz in liquid while keeping a stiffness of 0.15− 0.2 N/m. They are however

difficult to manufacture and the tips have to be grown subsequently using electron beam

deposition (Ando et al. 2001, 2003).

5.2.2 Theoretical considerations

Estimation of the maximum possible scan speed to achieve a given resolution

Besides scan speed limitations inherent to the hsAFM components, the maximum scan

speed allowing a given resolution is limited by purely geometrical considerations. First,

as for traditional AFM, the spacial gradient of the tip-sample interaction brings the

ultimate resolution limit. Assuming a scanning cantilever oscillating at a given frequency

166


ν, the scan speed v is limited by the number tap/area (cycle/area) necessary to resolve

the features studied. This relates to the energy dissipated by the tip into the sample

during each tap, necessary to the image formation.

In the following paragraph, it is assumed that the hsAFM scan speed limit imposed by

its different components is always beyond the geometrical limitations discussed.

Let us consider a sample the topography of which is described by a function T (x), scanned

at a speed v with a tip oscillating at a frequency ν (Fig. 5.1). Let us assume that ε

describes the smallest detectable lateral change of the tip-sample interaction (ε > 0),

and that r is the resolution aimed for, i.e. the size of the sample feature resolved in the

AFM image (Fig. 5.1). It is reasonable to assume (e.g. for PM) that T (x) is r-periodic.

v

εr

1/η=v/ν

X

T(x)

Figure 5.1: Scheme of the proposed model. The tip, oscillating at a frequency ν, scans at aspeed v over the sample. The lateral distance travelled between two successive taps is given by1/η = v/ν and the AFM resolves features of size r about the sample topography T (x).

We can define the spacial scanning frequency η by:

η = ν/v (5.1)

η corresponds to the number of tap per distance scanned. If we consider a feature of

size s, and an AFM scan speed vs satisfying s > 1/η(vs), then all the features of size

larger than > s can be resolved. When v → 0, the system tends towards its maximum

167


resolution. Since the maximum resolution is limited by ε > 0, the maximum resolution

is actually reached for any scan speed satisfying the following condition:

v < vε = νε (5.2)

If we define T ′(x) as the topography function resolved by the AFM, Eq. 5.2 allows the

decomposition of T ′(x) in a convergent Fourier series, which terms follow the spatial

frequencies ξn = 2πnr

, n ∈ N. Let us consider the amplitudes An and Bn of the series

sin and cos terms, respectively. Using ε, it is possible to define the smallest integer

nε satisfying 1/ξnε < ε. nε represents the Fourier series term beyond which the spatial

frequencies are smaller than ε. Hence we can write:

An = Bn = 0 ∀n > nε (5.3)

which guarantees the series convergence. Practically, it is reasonable to model the be-

haviour of the An and Bn coefficients as a decreasing exponentially with n:

|An|, |Bn| ∝ e−ρξn (5.4)

with ρ depending on the topography T ′(x) (Fig. 5.2.A). For more convenience, we can

extend Eq. 5.4 to a continuous function C(η) defined by:

C(η) = δe−ρη (5.5)

with δ a constant. It is now possible to define the resolution function R(η) ∈ [0; 1]

(maximum possible resolution corresponds to R = 1 and no resolution to R = 0) using

168


R(η) = 1− C(η) and imposing that:

R(η) = 0 ∀η 61

rand lim

η→∞R(η) = 1 (5.6)

These conditions lead to:

R(η) =

1− e−rη+1 if η > 1/r

0 if η < 1/r(5.7)

with r the size of the feature resolved by the AFM. This last equation can be rewritten

in term of scan speed:

R(v) =

1− e−r νv+1 if v 6 rν

0 if v > rν(5.8)

Fig. 5.2.B shows a example of a typical resolution curve for hsAFM.

0.0

1.0

0.8

0.6

0.4

0.2

0 100 µm/s

R(v)R(v)

1/ε1/r η

An, Bn

C(η)C(η)

200

A B

Figure 5.2: (A) Example of the convergence of the Fourier terms An and Bn, and of thefunction C(η) used to model them. (B) Resolution of a system scanning features of sizer = 0.5 nm with a cantilever oscillating at a frequency of 600kHz. Almost no limiting effectis visible below a scan speed of v =∼ 60µm/s. The resolution then decreases almost linearlyuntil 300 µm/s where the lateral distance η travelled between two successive taps is equal to r.

Extensibility of the high-resolution theory to high speed

In §4.3.2, a method (hydration force microscopy) was proposed in order to achieve high

169


resolution imaging of PM under a broad range of salt concentrations. This method was

based on small amplitude AC-AFM, adjusting the cantilever stiffness k to that of the

interaction kts.

The first condition which requires scanning amplitudes comparable or smaller than the

Debye length 1/κ can be met by the hsAFM.

It is possible to demonstrate (Appendix 8.1.3) that the second condition corresponds,

given a free oscillation amplitude, to maximizing the energy dissipation by the tip into

the sample. Since the image formation originates mainly from variations in the energy

dissipated by the tip into the sample, the high resolution method can be expected to apply

at high speed if the ratio energy dissipated into the sample/surface scanned is maintained

from traditional AFM high resolution parameters. This equivalence is however difficult

to quantify because forces not considered in the model such as friction forces (Dudko

et al. 2003) can provide important dissipation channels at higher speeds. A direct

comparison of the number of tap/surface2 may consequently be an oversimplification of

reality.

5.2.3 Achieving molecular resolution of PM at high-speed

Amplitude and phase feedback

The most common feedback parameter for AC-AFM is the cantilever oscillation ampli-

tude. Despite showing an often lower signal/noise ratio, the phase signal is generally

more sensitive to changes in tip-sample interaction. This was illustrated in §4.3.3, where

simultaneous acquisition of amplitude, phase and AC-deflection curves was run on both

surfaces of PM. Over 100 mM KCl concentration, the tip-sample interaction stiffness

exhibited a distinctive shoulder, accounting for an ordered double layer (see §4.3.3). This

2 typically 500 to 1000 tap/nm2 for high resolution with traditional AFM whereas hsAFM providedequivalent resolution (§5.2.4 following) with comparable tip oscillation amplitude but with only 200−500tap/nm2

170


shoulder was however mostly visible in the phase curves, despite a lower signal/noise ra-

tio (Fig. 5.3.C). This suggests phase to be a valuable parameter to feedback on, in order

to obtain complementary information from amplitude feedback. The hsAFM has been

developed with possibility of fast feedback on both amplitude and on phase (Yamashita

et al. 2006). In both cases, molecular resolution of PM could be achieved, but amplitude

images still provided a better resolution when scanning in identical conditions (Fig. 5.3).

This is because a higher scanning amplitude is required with phase feedback due to the

lower signal/noise ratio, and the first high-resolution condition mentioned above can not

be applied.

1.0

0 10 20 30 [nm]

0.80.60.4

90

858075

[nm]

[°]

A B Camplitude

phase70

Figure 5.3: Resolution of PM lattice obtained with phase (A) and amplitude feedback (B).Both pictures were captured with the same cantilever and at similar free oscillating amplitudes.The lattice is hardly visible in (A) but well resolved in (B). (C) Example of further sensitivityobtained through phase over amplitude information; a shoulder corresponding to PM extra-cellular ordered layer disruption (§4.3.3) is better resolved in the phase vs distance curve thanin the amplitude-distance curve, despite a lower signal/noise ratio. Scale and scan times: A100 nm × 100 nm captured over 594 ms, B 80 nm × 80 nm captured over 486 ms. A andB where obtained in 150 mM KCl, pH 8 and at the same scan speed. Curves in (C) wereacquired in 100 mM KCl.

Direct imaging of the extracellular ordered layer

Interestingly, the ordered double layer measured with traditional AC-AFM over PM ex-

tracellular surface (§4.3.3, (Antoranz Contera et al. 2006)) can be directly imaged by

171


hsAFM in both phase and amplitude feedback mode (Fig. 5.4), especially on larger

frames and with higher amplitudes. These conditions, not optimal for high resolution,

force the tip to travel across the whole double layer at each cycle, and at a frequency

which does not leave enough time for the disturbed layer to rearrange. Consequently,

scanning over PM extracellular surface often prevents the tip from reaching the actual

membrane surface, by bouncing on the ordered double layer. This phenomenon makes

PM extracellular surface recognizable when imaging with hsAFM and allows direct map-

ping of the ordered double layer (Fig. 5.4). At the patch edges, the ordered layer thick-

ness decreases continuously, and patch defects or cracks clearly affect the ordered layer

topography (Fig. 5.4.C) by disrupting it.

A B C

Figure 5.4: Imaging of the PM extracellular ordered layer (visible as a shoulder in phase andamplitude vs distance curves, Fig. 5.3) with high-speed AFM. The ordered layer prevents directcontact of tip with extracellular surface in larger frames (arrows in A. Progressive magnificationover extracellular patches (B-C) allows imaging of the membrane when the ordered layer ispartially disrupted. The extracellular surface is still not fully revealed and the remainingordered layer shows some hexagonal symmetry according with PM lattice (C). All the imageswere acquired in 300 mM KCl at pH 8. The scales and acquisition times are: (A) 1 µm × 1 µm,2.5 s, (B) 715 nm × 715 nm, 1.8 s and (C) 400 nm × 400 nm, 1.0 s.

Sub-molecular resolution of Purple Membrane at high-speed

Applying the high-resolution method, it was possible to image PM with sub-molecular

resolution (Fig. 5.5). Single α-helices are visible, protruding out of the membrane cy-

172


toplasmic surface. Sub-molecular details are also visible on PM extracellular surface

despite a general lower resolution (see previous paragraph). The frame size was reduced

to a rectangle in order to further minimize the scanning time per frame. Such a res-

olution could be achieved at scan rates of about 50 frames/s (19 or 21 ms/frame)

(Fig. 5.5.A-B) with some lower resolution images collected at more than 110 frames/s

(9 ms/frame) (Fig. 5.5.C). In terms of scan speed, this corresponds to ∼ 50 µm/s and

A B C

Figure 5.5: High resolution pictures of PM extracellular (A) and cytoplasmic (B-C) obtainedat high speed. Sub-molecular details are visible within a single protein (highlighted trimers(A-B) at scan rated of ∼ 20 ms/frame but disappear when doubling the scan speed (C).Scale and acquisition time: (A) 20 nm × 6 nm, 21 ms, (B) 15 nm × 4.5 nm, 19 ms, (C)15 nm × 4.5nm, 9 ms

∼ 130µm/s respectively. Comparing these curves with the geometrical limitation obtain

in §5.2.1, Fig. 5.2.B, shows that such limitations are already operating and can explain

the lower resolution obtained for the fastest frames. Higher speeds would require the

use of higher oscillation frequency in order to achieve similar resolution. Nevertheless,

sub-molecular resolution of PM cytoplasmic surface was obtained at speeds relevant to

directly study the protein photocycle.

5.2.4 Imaging with dynamical pH changes

hsAFM offers the possibility to image a sample while dynamical changes are imposed to

the buffer environment. If these changes affect the tip sample interaction, their effect

can be mapped in real-time. To illustrate this possibility, PM was imaged in a solution

containing 500 mM KCl and 10 mM Tris at pH 10. Some hydrochloric acid ([HCl] =

173


0.5 M) was then continuously injected into the buffer. A total of ∼ 83 µl of HCl were

injected into 150 µl of buffer. The injections process was however not homogeneous,

and the actual pH in the scanned region could not be reliably calculated. Despite being

a purely qualitative experiment, changes in the membrane appearance can be observed

(Fig. 5.6), demonstrating the ability of hsAFM to record not only dynamical changes of

the sample itself, but also of the imaging medium.

pH decrease

Figure 5.6: Dynamical pH changes imaged with hsAFM. Each frame (1.5 µm × 1.5 µm) wasacquired over 2 s. The pH change affects the tip-sample interactions as reflected in the images.(see text for pH details)

5.3 Real-time observation of bacteriorhodopsin photocycle: large

structural rearrangements and functional implications

5.3.1 Introduction

Early after the discovery of bR (Oesterhelt & Stoeckenius 1971) and before the first

atomic structure was proposed (Henderson et al. 1990), spectroscopy techniques have

been extensively used to characterize bR dynamics and resolve the ∼ 10 ms photocycle,

revealing several spectroscopically distinct steps (Kung et al. 1975; Lozier et al. 1975).

Following these reports, a considerable effort using time-resolve spectroscopy has aimed

174


at detailing the multiple timescale kinetics of the proton pumping mechanism (Ippen

et al. 1978; Groma et al. 1984; Xie et al. 1987; Mathies et al. 1988; Birge 1990).

At the present time, the most widely held view of this process is that of a succession

of homogeneous and reversible steps involving a linear sequence of spectrally distinct3

intermediate states (Lozier et al. 1992)

bR→ K L M N O → bR

triggered by the absorption of a photon by the retinal (Kobayashi et al. 2001). The

proton pumping mechanism is correlated with the photocycle (see §1.2.1); the proton

is released from the retinal Schiff base to the extracellular medium during the first

half of the photocycle (L → M). The retinal reprotonation occurs from the protein

cytoplasmic side during the photocycle second half and is accompanied by large confor-

mational changes in bR cytoplasmic region. Several spectroscopic studies have however

demonstrated that the proton pumping mechanism exhibits a complex dynamic and put

forward an unresolved picture of the photocycle that challenges the simplicity of a linear

and reversible succession of intermediate states (Hendler 2005). Different alternative

photocycle models assuming unidirectional, parallel heterogeneous cycles have been pro-

posed to explain the complex kinetics observed (Hendler 2005; Tokaji 1995). There are

two intriguing and controversial spectroscopic phenomena observed in purple membrane

(and not with bR monomers):

1. The intensity of the exciting light affects the number of proteins undergoing the

photocycle and the kinetics of the M intermediate (Dancshazy & Tokaji 1993;

Shrager et al. 1995; Luchian et al. 1996; Mostafa 2004). The existence of two

spectrally undistinguishable M states proposed earlier on as M1 and M2 (Chizhov

3 A supplementary step M1 → M2 was later proposed (Chizhov et al. 1996) to explain the differentdecay time constants obtained from time-resolved spectroscopic experiments.

175


et al. 1996), was reinterpreted as slow (Ms) and fast (Mf ) decay coexisting bR

species. The effect is seen as an increase in the proportion of slow-decaying Ms

species over fast-decaying Mf species when increasing the actinic light intensity.

2. The biphasic band shape of visible circular dichroism (CD) of bR in purple mem-

brane (Muccio & Cassim 1979; Cassim 1992; Wu & El-Sayed 1991).

It has been proposed that both phenomena are caused by heterogeneity of bR proteins

(El-Sayed et al. 1989; Mollaaghababa et al. 2000). Alternatively these results have

been explained by strong lipid-mediated bR-bR interactions (Bryl & Yoshihara 2001)

resulting in photocooperation of the proteins within the crystal (Shrager et al. 1995;

Tokaji 1993, 1998; Varo et al. 1996; Komrakov & Kaulen 1995). These cooperative

interactions can happen via structural changes (Vonck 2000) or in the case of the CD

spectra, via excitonic coupling of transition dipole moments of the retinal chromophores

(Cassim 1992).

Following spectroscopic studies, high resolution X-ray diffraction studies of bR structures

became possible with the development cubic phase bilayer method (Landau & Rosen-

busch 1996) to produce crystals of bR wild type and different mutant proteins and by

collecting data of frozen crystals containing reasonable concentrations of proteins (more

than 25%) in the desired state (Pebay-Peyroula et al. 1997; Neutze et al. 2002). These

structures have given insight in the local processes involved in the photocycle and re-

lated bR conformational changes to the different states of the accepted model (§1.2.1).

However, unlike most of the spectroscopy studies of bR dynamics, the low temperature

and the relatively long integration time necessary for X-ray crystallography prevents

the capture of the complex dynamic nature of protein function and the structural fluc-

tuations essential to protein activity (Ferrand et al. 1993; Manneville et al. 1999).

Moreover experimental difficulties inherent to proteins crystallization make the X-ray

176


data contentious (Kamikubo & Kataoka 2005); the cubic phase method can alter the

conformation and kinetics of bR (Schenkl et al. 2003; Efremov et al. 2006; Lunde et al.

2006). Nevertheless, time-resolved X-ray (Koch et al. 1991) and certain continuous

wave diffraction studies (Kamikubo et al. 1996; Kamikubo & Kataoka 2005; Sass et al.

1997) observed the large conformational changes in bR cytoplasmic half associated with

the Schiff base reprotonation during the M → N transition, in agreement with neutron

diffraction (Dencher et al. 1989; Reat et al. 1998), infrared spectroscopy (Ormos 1991;

Efremov et al. 2006) and electron diffraction (Subramaniam et al. 1993; Vonck 2000)

reports. This large structural rearrangement, involving important movements of helices

E and F away from the trimer center, has been proposed to be at the origin of the

biphasic decay observed by spectroscopic techniques (Vonck 2000). Neutron diffraction

experiments (Ferrand et al. 1993) and time-resolved spectroscopy experiments (Varo &

Lanyi 1991) have shown that the free energy initially stored in the chromophore is trans-

ferred to the protein in an irreversible transition between an early M state and a late M

with associated drops in enthalpy and entropy. It was proposed that the enthalpy drives

a large conformational change which results in a drop of entropy. Subsequent steps in the

photocycle are entropy driven as the protein relaxes to its ground state. Consistently,

the efficiency of the photocycle is strongly dependant on thermal motion within bR; the

M → bR transition is correlated with large-amplitude anharmonic atomic motion, and

a temperature decrease below 230 K slowed down the pumping activity by several order

of magnitude (Ferrand et al. 1993; Lehnert et al. 1998).

All these considerations suggest that the complex dynamics involving structural hetero-

geneity and/or protein cooperation may be related the large structural changes related

to the Schiff base reprotonation. We propose that the decay from M to the ground (bR)

state can control the kinetics of the photocycle and may be necessary for structural com-

munication between bR monomers within PM (Vonck 2000). This idea is supported by

177


mutation studies which demonstrated that if the protein is restrained from undergoing

large conformational changes within the membrane, the proton pumping mechanism is

still possible, but reduced by ∼ 2 orders of magnitude (Tittor et al. 2002).

5.3.2 Direct imaging of bacteriorhodopsin photo-activity with hsAFM

The hsAFM developed by Ando’s group has demonstrated an unprecedented ability to

combine high spatial and temporal resolution by providing sub-molecular resolution of

individual bR molecules at the membrane surface and on a timescale relevant to study

the large conformational changes accompanying the transition M → bR (§5.2.1). In this

study we have applied this exceptional ability to image conformational changes of single

bR molecules in PM, in a saline solution and at room temperature. This direct and new

insight into the problem of bR structural heterogeneity/cooperativity related to the M

decay confirms the associated large conformational change of bR, and suggests that only

one protein per trimer can isomerize at a time, consistent with electron microscopy ob-

servations (Vonck 2000). Further investigations with variation of the actinic intensities

and imaging temperature should allow direct verification of the protein cooperativity

theory within PM. Besides the biophysical question, this study demonstrates the capa-

bility of hsAFM to provide a direct ”movie” of protein functioning in real time, and in

biologically relevant conditions, and with no sample preparation requirements such as

labelling or three-dimensional crystallization.

Results

Fig. 5.7 shows photo-induced conformational changes of individual bR, obtained by

high-speed scanning of PM cytoplasmic surface. The photo-excited molecules open their

cytoplasmic half, exhibiting two distinct, well separated protrusions instead of the usual

single bundle visible for each bR monomer. We always observed only one bR molecule

178


per trimer exhibiting the conformational change.

The effective scanning time required to resolve a single monomer within a picture is

. 5 ms, which is fast enough to capture the dynamics of the M → bR transition within

a frame. The imaging rate is however comparable to that of bR photocycle, and does

not allow acquisition of several successive pictures of a protein dynamical conformational

change. Consistently, the images captured immediately before and after the observed

photo-event do not show conformational changes for the relevant bR monomer (Fig. 5.7).

When scanning dark-adapted PM, similar structural changes were rarely observed (see

discussion). In order to assess the tip effect on the observed photo-events, dynamical

tip-imposed structural changes of bR (Muller et al. 1995) were induced by continuously

decreasing the scanning setpoint while imaging at high-speed and in the dark. The cap-

tured images did not show any similarity with the events in Fig. 5.7, but rather an

global increase of the inter-trimer space. It is therefore unlikely for the tip to induce the

individual photo-events observed.

Discussion and conclusions

The large photo-induced conformational change showing a double protrusion observed

with hsAFM supports the idea that bR helices E and F tilt away from the rest of the

protein and clash with the next trimer, thus preventing more than one third of the bR

to isomerize at a time (Vonck 2000). Consistently, always one isomerizing protein per

trimer was observed at a time. The isomerizations visible in Fig. 5.7 are most likely

photo-induced considering the density of the photon flux on the scanned membrane (Ap-

pendix 8.2). The hypothesis of a single photo-isomerization per trimer however requires

further confirmation with experiments varying the intensity of the exciting light beam.

Thermal isomerizations are also possible and were sometimes observed in the dark also

179


A

B

C

D

Figure 5.7: Photo-isomerization of bR in PM. All the events are successive frames capturedwhile the scanned sample was illuminated with a green laser (λ = 532 nm, 7 mW focused onspot of ∼ 5 µm radius). Using PM extinction coefficient at 532 nm, it possible to show thatthe light intensity used is sufficient to photo-excite each bR molecule over a frame scan time(see Appendix 8.2). The scale and acquisition time are 15 nm × 4.5 nm and 19 ms for (A)and 15 nm × 15 nm 64 ms for (B-D). All the pictures where captured in 150 mM KCl, 10mM Tris, pH 8.

180


not as frequently (data not shown). Some studies reported that extraction of the retinal

from dark adapted PM showed ∼ 2/3 chromophores to be in the 13-cis state whereas the

retinal from light adapted membrane were almost fully in the all-trans state (Scherrer

et al. 1989). Thermal adaptation of the membrane in darkness takes typically ∼ 20 min

(Milder 1991) and can be observed with hsAFM. Full darkness adaptation of PM is how-

ever not possible when scanned with hsAFM due to the wavelength of the AFM laser

(675 nm). This wavelength is absorbed by both, water and carbon-carbon single/double

bonds. Absorption by the water is likely to induce local variations of the sample temper-

ature, which can in principle be overcome by implementing temperature control of the

sample (e.g. with a Peltier-effect stage). The IR laser absorption by the retinal C = C is

however more problematic because it can stimulate all-trans 13-cis retinal transition

by inducing vibration of the C13 = C14 double bond, known to precede photo-induced

isomerization of the chromophore and to unlock the π − π orbital coupling (Gonzalez-

Luque et al. 2000).

Concomitant to photo-isomerization events, spatial fluctuations of the membrane could

be observed. Such fluctuations are expected to be induced by both thermal motion/scanning

friction (Ferrand et al. 1993) and the proteins activity (Manneville et al. 1999; Ra-

maswamy et al. 2000). A reliable interpretation of these fluctuations in terms of bR

cooperativity and possible activity enhancement induced by membrane fluctuation (As-

tumian 2002; Frauenfelder et al. 1991) will however require the development of a software

capable of analyzing these fluctuations systematically, and of establishing their frequency

spectrum and temporal correlation with PM illumination. The energy dissipated by the

tip into the sample also probably influence the dynamics of membrane fluctuations but

is very difficult to quantify. Repeating the experiment at different temperatures may

provide a better understanding of the tip-induced membrane fluctuations.

Generally, like traditional AFM, hsAFM uses tip energy dissipation to interrogate the

181


system studied. The dissipation channel employed can however vary from traditional

AFM, and friction components can become important at high speed. Decoupling ther-

mal and tip induced effects presents a challenge for any hsAFM studies, and temperature

dependent work will require careful investigation. In the case of single bR conformational

changes, tip induced isomerization could be ruled out and the protein activity reliably

imaged.

hsAFM enabled us to image, for the first time, dynamical conformational changes of

single bR proteins with sub-molecular resolution and on a millisecond timescale. The

pictures could resolve the twist of two helices (E and F) away from the trimer center

during the reprotonation of the Schiff base (protein M -state). It also opened new way

to investigate larger scale motions within the membrane, known to strongly influence

proteins activity in nature. This newly developed tool will provide valuable insight in

the dynamics of biomolecular systems, and allow direct imaging of proteins at work in

their natural environment.


PM of Halobacterium salinarium (Sigma-Aldrich, Dorset, UK) were diluted in buffer (5

µl of 10 mM Tris, 150 mM KCl, pH8) and adsorbed on freshly cleaved mica (1 mm

diameter discs). Sample were imaged using the hsAFM developed and built in the

biophysics laboratory of T. Ando’s research group, Kanazawa (Ando et al. 2001). The

cantilever used were manufactured specially by Olympus (Olympus, Tokyo, Japan) such

as to achieve a typical resonance frequency of ∼ 600 kHz in liquid with an estimated

stiffness of 0.15 − 0.2 N/m (§5.1.2, (Ando et al. 2001)). The tip were grown on the

cantilever in the laboratory using focused electron beam. The tip lifetime was typically

5000 − 10000 images. A green laser (λ = 532 nm) was focused in a spot of ∼ 10µm in

diameter just next to the scanning tip, but attention was paid not to directly focus on

182


the cantilever end in order to avoid deflecting the cantilever when the laser is on. The

intensity was set to 7 mW and the illumination was modulated by means of a shutter

and directly controlled from the image acquisition software.

When running the experiment, the system was first set so as to provide stable high-

resolution over a chosen membrane patch. Excitation of bR molecules was then triggered

by successive opening and closing of the shutter. The illumination/darkness times were

set such as to allow the capture of ∼ 40 images in each case.

5.4 References

Ando, Toshio, Kodera, Noriyuki, Takai, Eisuke, Maruyama, Daisuke,Saito, Kiwamu, & Toda, Akitoshi. 2001. A high-speed atomic force micro-scope for studying biological macromolecules. Proc. Natl. Aacad. Sci. USA, 98(22),12468–12472.

Ando, Toshio, Kodera, Noriyuki, Naito, Yasuyuki, Kinoshita, Tatsuya,Furuta, Ken’ya, & Toyoshima, Yoko Y. 2003. A High-speed Atomic ForceMicroscope for Studying Biological Macromolecules in Action. Chem. Phys. Chem.,4, 1196–1202.

Ando, Toshio, Uchihashi, Takayuki, Kodera, Noriyuki, Miyagi, Atsushi,Nakakita, Ryo, Yamashita, Hayato, & Sakashita, Mitsuru. 2006. High-Speed Atomic Force Microscopy for Studying the Dynamic Behavior of ProteinMolecules at Work. Jap. J. Appl. Phys., 45(3B), 1897–1903.

Antoranz Contera, Sonia, Voıtchovsky, Kislon, & Ryan, J. F. 2006. Hydra-tion force microscopy reveals the molecular forces that trap and release protons in abacterial proton-pump. Submitted.

Astumian, R. D. 2002. Protein conformational fluctuations and free-energy transduc-tion. App. Phys. A, 75(2), 193.

Berg, Jeremy M., Tymoczko, John L., & Stryer, Lubert. 2002. Biochemistry.5th edn. New York: W. H. Freeman and Company.

Birge, Robert R. 1990. Nature of the primary photochemical events in rhodopsinand bacteriorhodopsin. Biochim. Biophys. Acta, 1016(3), 293–327.

183


Brixner, Tobias, Stenger, Jens, Vaswani, Harsha M., Cho, Minhaeng,Blankenship, Robert E., & Fleming, Graham R. 2005. Two-dimensional spec-troscopy of electronic couplings in photosynthesis. Nature, 434(7033), 625.

Bryl, K., & Yoshihara, K. 2001. The role of retinal in the long-range protein-lipid interactions in bacteriorhodopsin-phosphatidylcholine vesicles. Eur. Biophys. J.,29(8), 628.

Cassim, Joseph Y. 1992. Unique biphasic band shape of the visible circular dichroismof bacteriorhodopsin in purple membrane Excitons, multiple transitions or proteinheterogeneity? Biophys. J., 63(5), 1432–1442.

Chen, E., Wittung-Stafshede, P., & Kliger, D. S. 1999. Far-UV Time-ResolvedCircular Dichroism Detection of Electron-Transfer-Triggered Cytochrome c Folding. J.Am. Chem. Soc., 121(16), 3811–3817.

Chen, Eefei, Goldbeck, Robert A., & Kliger, David S. 1997. Nanosecond time-resolved spectroscopy of biomolecular processes. Ann. Rev. Biophys. Biomol. Struct.,26(1), 327–355.

Chizhov, I., Chernavskii, D. S., Engelhard, M., Mueller, K. H., Zubov,B. V., & Hess, B. 1996. Spectrally silent transitions in the bacteriorhodopsin pho-tocycle. Biophys. J., 71(5), 2329–2345.

Coppens, P., & Novozhilova, I. 2002. Time-resolved chemistry at atomic resolution.Faraday Discuss., 122, 1–11.

Cubeddu, R., Comelli, D., Andrea, C. D., Taroni, P., & Valentini, G. 2002.Time-resolved fluorescence imaging in biology and medicine. J. Phys. D, 35(9), R61.

Daggett, Valerie. 2000. Long timescale simulations. Curr. Op. Struct. Biol., 10(2),160.

Dancshazy, Z., & Tokaji, Z. 1993. Actinic light density dependence of the bacteri-orhodopsin protocycle. Biophys. J., 65(2), 823–831.

Dencher, N. A., Dresselhaus, D., Zaccai, G., & Buldt, G. 1989. Struc-tural Changes in Bacteriorhodopsin during Proton Translocation Revealed by NeutronDiffraction. Proc. Natl. Aacad. Sci. USA, 86(20), 7876–7879.

Dudko, O. K., Filippov, A. E., Klafter, J., & Urbakh, M. 2003. Beyond theconventional description of dynamic force spectroscopy of adhesion bonds. Proc. Natl.Aacad. Sci. USA, 100(20), 11378–11381.

Efremov, R., Gordeliy, V. I., Heberle, J., & Buldt, G. 2006. Time-ResolvedMicrospectroscopy on a Single Crystal of Bacteriorhodopsin Reveals Lattice-InducedDifferences in the Photocycle Kinetics. Biophys. J., 91(4), 1441–1451.

184


El-Sayed, M. A., Lin, C. T., & Mason, W. R. 1989. Is There an ExcitonicInteraction or Antenna System in Bacteriorhodopsin? Proc. Natl. Aacad. Sci. USA,86(14), 5376–5379.

Elber, Ron. 2005. Long-timescale simulation methods. Curr. Op. Struct. Biol., 15(2),151.

Fantner, Georg E., Schitter, Georg, Kindt, Johannes H., Ivanov, Tzve-tan, Ivanova, Katarina, Patel, Rohan, Holten-Andersen, Niels, Adams,Jonathan, Thurner, Philipp J., Rangelow, Ivo W., & Hansma, Paul K.2006. Components for high speed atomic force microscopy. Ultramicroscopy, 106(8-9),881.

Ferrand, M., Dianoux, A. J., Petry, W., & Zaccai, G. 1993. Thermal Motionsand Function of Bacteriorhodopsin in Purple Membranes: Effects of Temperature andHydration Studied by Neutron Scattering. Proc. Natl. Aacad. Sci. USA, 90(20), 9668–9672.

Frauenfelder, H., Sligar, S. G., & Wolynes, P. G. 1991. The energy landscapesand motions of proteins. Science, 254(5038), 1598–1603.

Genick, Ulrich K., Borgstahl, Gloria E. O., Ng, Kingman, Ren, Zhong,Pradervand, Claude, Burke, Patrick M., Srajer, Vukica, Teng, Tsu-Yi,Schildkamp, Wilfried, McRee, Duncan E., Moffat, Keith, & Getzoff,Elizabeth D. 1997. Structure of a Protein Photocycle Intermediate by MillisecondTime-Resolved Crystallography. Science, 275(5305), 1471–1475.

Gonzalez-Luque, Remedios, Garavelli, Marco, Bernardi, Fernando, Mer-chan, Manuela, Robb, Michael A., & Olivucci, Massimo. 2000. Computa-tional evidence in favor of a two-state, two-mode model of the retinal chromophorephotoisomerization. Proc. Natl. Acad. Sci. USA, 97(17), 9379–9384.

Groma, G. I., Szabo, G., & Varo, Gy. 1984. Direct measurement of picosecondcharge separation in bacteriorhodopsin. Nature, 308(5959), 557.

Henderson, R., Baldwin, J. M., Ceska, T. A., Zemlin, F., Beckmann, E.,& Downing, K. H. 1990. Model for the structure of bacteriorhodopsin based onhigh-resolution electron cryo-microscopy. J. Mol. Biol., 213(4), 899.

Hendler, R. W. 2005. An Apparent General Solution for the Kinetic Models of theBacteriorhodopsin Photocycles. J. Phys. Chem. B, 109(34), 16515–16528.

Humphris, A. D. L., Miles, M. J., & Hobbs, J. K. 2005. A mechanical microscope:High-speed atomic force microscopy. Appl. Phys. Lett., 86(3), 034106.

Ippen, E. P., Shank, C. V., Lewis, A., & Marcus, M. A. 1978. Subpicosecondspectroscopy of bacteriorhodopsin. Science, 200(4347), 1279–1281.

185


Isralewitz, Barry, Gao, Mu, & Schulten, Klaus. 2001. Steered moleculardynamics and mechanical functions of proteins. Curr. Op. Struct. Biol., 11(2), 224.


Kamikubo, Hironari, Kataoka, Mikio, Varo, Gyorgy, Oka, Toshihiko,Tokunaga, Fumio, Needleman, Richard, & Lanyi, Janos K. 1996. Struc-ture of the N intermediate of bacteriorhodopsin revealed by x-ray diffraction. Proc.Natl. Aacad. Sci. USA, 93(4), 1386–1390.

Karplus, Martin, & McCammon, J. Andrew. 2002. Molecular dynamics simula-tions of biomolecules. Nat. Struct. Mol. Biol., 9(9), 646.

Kim, J. E., Pan, D., & Mathies, R. A. 2003. Picosecond Dynamics of G-ProteinCoupled Receptor Activation in Rhodopsin from Time-Resolved UV Resonance RamanSpectroscopy. Biochemistry, 42(18), 5169–5175.

Kobayashi, Takayoshi, Saito, Takashi, & Ohtani, Hiroyuki. 2001. Real-timespectroscopy of transition states in bacteriorhodopsin during retinal isomerization.Nature, 414(6863), 531.

Koch, M. H. J., Dencher, N. A., Oesterhelt, D., Plhn, H.-J., Rapp, G., &Bldt, G. 1991. Time-resolved X-ray diffraction study of structural changes associatedwith the photocycle of bacteriorhodopsin. EMBO J., 10(3), 521–526.

Kodera, Noriyuki, Yamashita, Hayato, & Ando, Toshio. 2005. Active dampingof the scanner for high-speed atomic force microscopy. Rev. Scient. Inst., 76, 053708–1–053708–5.

Komrakov, Andrey Yu, & Kaulen, Andrey D. 1995. M-decay in the bacte-riorhodopsin photocycle: effect of cooperativity and pH. Biophys. Chem., 56(1-2),113.

Kung, M. Chu, Devault, Don, Hess, Benno, & Oesterhelt, Dieter. 1975.Photolysis of Bacterial Rhodopsin. Biophys. J., 15, 907–911.

Landau, Ehud M., & Rosenbusch, Jurg P. 1996. Lipidic cubic phases: A novelconcept for the crystallization of membrane proteins. Proc. Natl. Aacad. Sci. USA,93(25), 14532–14535.

Lehnert, Ursula, Reat, Valerie, Weik, Martin, Zaccai, Giuseppe, & Pfis-ter, Claude. 1998. Thermal Motions in Bacteriorhodopsin at Different HydrationLevels Studied by Neutron Scattering: Correlation with Kinetics and Light-InducedConformational Changes. Biophys. J., 75(4), 1945–1952.

186


Levanon, H., & Mobius, K. 1997. Advanced EPR spectrsocopy on electron trans-fer processes in photosynthesis and biomemetic model systems. Annual Review ofBiophysics and Biomolecular Structure, 26(1), 495–540.

Lozier, R. H., Xie, A., Hofrichter, J., & Clore, G. M. 1992. Reversible Stepsin the Bacteriorhodopsin Photocycle. Proc. Natl. Aacad. Sci. USA, 89(8), 3610–3614.

Lozier, Richard H., A., Bogomolni Roberto, & Stoeckenius, Walther. 1975.Bacteriorhodopsin: A light-driven proton pump in Halobacterium halobium. Biophys.J., 15, 955–962.

Luchian, Tudor, Tokaji, Zsolt, & Dancshazy, Zsolt. 1996. Actinic light densitydependence of the O intermediate of the photocycle of bacteriorhodopsin. FEBS Lett.,386(1), 55.

Lunde, Christopher S., Rouhania, Shahab, Facciottib, Marc T., &Glaeser, Robert M. 2006. Membrane-protein stability in a phospholipid-basedcrystallization medium. J. Struct. Biol., 154(3), 223–231.

Manalis, S. R., Minne, S. C., & Quate, C. F. 1996. Atomic force microscopy forhigh speed imaging using cantilevers with an integrated actuator and sensor. Appl.Phys. Lett., 68(6), 871.

Manneville, J. B., Bassereau, P., Levy, D., & Prost, J. 1999. Activity ofTransmembrane Proteins Induces Magnification of Shape Fluctuations of Lipid Mem-branes. Phys. Rev. Lett., 82(21), 4356.

Mathies, R. A., Brito Cruz, C. H., Pollard, W. T., & Shank, C. V. 1988.Direct observation of the femtosecond excited-state cis-trans isomerization in bacteri-orhodopsin. Science, 240(4853), 777–779.

Michalet, X., Weiss, S., & Jager, M. 2006. Single-Molecule Fluorescence Studiesof Protein Folding and Conformational Dynamics. Chem. Rev., 106(5), 1785–1813.

Milder, Steven J. 1991. Correlation between absorption maxima and thermal iso-merization rates in bacteriorhodopsin. Biophys. J., 60, 440–446.

Moffat, Keith. 2002. The frontiers of time-resolved macromolecular crystallography:movies and chirped X-ray pulses. Faraday Discuss., 122, 65–77.

Mollaaghababa, R., Steinhoff, H. J., Hubbell, W. L., & Khorana, H. G.2000. Time-Resolved Site-Directed Spin-Labeling Studies of Bacteriorhodopsin: Loop-Specific Conformational Changes in M. Biochemistry, 39(5), 1120–1127.

Mostafa, Hamdy I. A. 2004. Heterogeneity based on bending of purple membranecontaining bacteriorhodopsin. FEBS Lett., 571(1-3), 134.

187


Muccio, D. D., & Cassim, J. Y. 1979. Interpretation of the absorption and circulardichroic spectra of oriented purple membrane films. Biophys. J., 26(3), 427–440.

Muller, Daniel J., Buldt, Georg, & Engel, Andreas. 1995. Force-inducedconformational change of bacteriorhodopsin. J. Mol. Biol., 249(2), 239.


Oesterhelt, D., & Stoeckenius, W. 1971. Rhodopsin-like protein from the purplemembrane of Halobacterium halobium. Nat. New Biol., 233(39), 149.

Ormos, P. 1991. Infrared Spectroscopic Demonstration of a Conformational Change inBacteriorhodopsin Involved in Protein Pumping. Proc. Natl. Aacad. Sci. USA, 88(2),473–477.

Palmer, I. I. I. Arthur G. 1997. Probing molecular motion by NMR. Curr. Op.Struct. Biol., 7(5), 732.

Pebay-Peyroula, E., Rummel, G., Rosenbusch, J. P., & Landau, E. M. 1997.X-ray structure of bacteriorhodopsin at 2.5 Afrom microcrystals grown in lipidic cubicphases. Science, 277(5332), 1676–1681.

Ramaswamy, Sriram, Toner, John, & Prost, Jacques. 2000. NonequilibriumFluctuations, Traveling Waves, and Instabilities in Active Membranes. Phys. Rev.Lett., 84(15), 3494.

Reat, Valerie, Patzelt, Heiko, Ferrand, Michel, Pfister, Claude, Oester-helt, Dieter, & Zaccai, Giuseppe. 1998. Dynamics of different functional partsof bacteriorhodopsin: H-2H labeling and neutron scattering. Proc. Natl. Aacad. Sci.USA, 95(9), 4970–4975.

Sansom, Mark S. P., Shrivastava, Indira H., Bright, Joanne N., Tate, John,Capener, Charlotte E., & Biggin, Philip C. 2002. Potassium channels: struc-tures, models, simulations. Biochim. Biophys. Acta, 1565(2), 294.

Sass, H. J., Schachowa, I. W., Rapp, G., Koch, M. H. J., Oesterhelt, D.,Dencher, N. A., & Buldt, G. 1997. The tertiary structural changes in bacteri-orhodopsin occur between M states: X-ray diffraction and Fourier transform infraredspectroscopy. EMBO J., 16(7), 1484–1491.

Schenkl, S., van Mourik, F., van der Zwan, G., Haacke, S., & Chergui,M. 2005. Probing the Ultrafast Charge Translocation of Photoexcited Retinal inBacteriorhodopsin. Science, 309(5736), 917–920.

188


Schenkl, Selma, Portuondo, Erwin, Zgrablic, Goran, Chergui, Majed,Suske, Winfried, Dolder, Max, Landau, Ehud M., & Haacke, Stefan.2003. Compositional Heterogeneity Reflects Partial Dehydration in Three-dimensionalCrystals of Bacteriorhodopsin. J. Mol. Biol., 329(4), 711.

Scherrer, P., Mathew, M. K., Sperling, W., & Stoeckenius, W. 1989. RetinalIsomer Ratio in Dark-Adapted Purple Membrane and Bacteriorhodopsin Monomers.Biochemistry, 28, 829–834.

Schitter, G., Stark, R. W., & Stemmer, A. 2004a. Fast contact-mode atomicforce microscopy on biological specimen by model-based control. Ultramicroscopy,100(3-4), 253.

Schitter, G., Allg, F., wer, & Stemmer, A. 2004b. A new control strategy forhigh-speed atomic force microscopy. Nanotechnology, 15(1), 108.

Shrager, Richard I., Hendler, Richard W., & Bose, Salil. 1995. The Abilityof Actinic Light to Modify the Bacteriorhodopsin Photocycle. Heterogeneity and/orPhotocooperativity? Eur. J. Biochem., 229(3), 589–595.

Subramaniam, S., Gerstein, M., Oesterhelt, D., & Henderson, R. 1993. Elec-tron diffraction analysis of structural changes in the photocycle of bacteriorhodopsin.EMBO J., 12(1), 1–8.

Sulchek, T., Hsieh, R., Adams, J. D., Minne, S. C., Quate, C. F., & Adder-ton, D. M. 2000a. High-speed atomic force microscopy in liquid. Rev. Sci. Inst.,71(5), 2097.

Sulchek, T., Hsieh, R., Adams, J. D., Yaralioglu, G. G., Minne, S. C.,Quate, C. F., Cleveland, J. P., Atalar, A., & Adderton, D. M. 2000b.High-speed tapping mode imaging with active Q control for atomic force microscopy.Appl. Phys. Let., 76(11), 1473.

Tittor, J., Paula, S., Subramaniam, S., Heberle, J., Henderson, R., &Oesterhelt, D. 2002. Proton Translocation by Bacteriorhodopsin in the Absenceof Substantial Conformational Changes. J. Mol. Biol., 319(2), 555.

Tokaji, Z. 1993. Dimeric-like kinetic cooperativity of the bacteriorhodopsin moleculesin purple membranes. Biophys. J., 65(3), 1130–1134.

Tokaji, Zsolt. 1995. Cooperativity-regulated parallel pathways of the bacteri-orhodopsin photocycle. FEBS Lett., 357(2), 156.

Tokaji, Zsolt. 1998. Quantitative model for the cooperative interaction of the bacte-riorhodopsin molecules in purple membranes. FEBS Letters, 423(3), 343.

189


Uchihashi, Takayuki, Kodera, Noriyuki, Itoh, Hisanori, Yamashita, Hay-ato, & Ando, Toshio. 2006. Feed-Forward Compensation for High-Speed AtomicForce Microscopy Imaging of Biomolecules. Jap. J. Appl. Phys., 45(3B), 1904–1908.

Varo, G., Needleman, R., & Lanyi, J. K. 1996. Protein structural change at thecytoplasmic surface as the cause of cooperativity in the bacteriorhodopsin photocycle.Biophys. J., 70(1), 461–467.

Varo, Gyrgy, & Lanyi, Janos K. 1991. Thermodynamics and Energy Coupling inthe Bacteriorhodopsin Photocycle. Biochemistry, 30(20), 5016–5022.

Viani, M. B., Schaffer, T. E., Paloczi, G. T., Pietrasanta, L. I., Smith,B. L., Thompson, J. B., Richter, M., Rief, M., Gaub, H. E., Plaxco,K. W., Cleland, A. N., Hansma, H. G., & Hansma, P. K. 1999. Fast imagingand fast force spectroscopy of single biopolymers with a new atomic force microscopedesigned for small cantilevers. Rev. Sci. Inst., 70(11), 4300.

Viani, Mario B., Pietrasanta, Lia I., Thompson, James B., Chand, Ami,Gebeshuber, Ilse C., Kindt, Johannes H., Richter, Michael, Hansma,Helen G., & Hansma, Paul K. 2000. Probing protein-protein interactions in realtime. Nat. Struct. Mol. Biol., 7(8), 644.

Vonck, Janet. 2000. Structure of the bacteriorhodopsin mutant F219L N intermediaterevealed by electron crystallography. EMBO J., 19(10), 2152–2160.

Williams, S., Causgrove, T. P., Gilmanshin, R., Fang, K. S., Callender,R. H., Woodruff, W. H., & Dyer, R. B. 1996. Fast Events in Protein Folding:Helix Melting and Formation in a Small Peptide. Biochemistry, 35(3), 691–697.

Wu, Shuguang, & El-Sayed, Mostafa A. 1991. CD spectrum of bacteriorhodopsin:Best evidence against exciton model. Biophys. J., 60(1), 190–197.

Xie, A. H., Nagle, J. F., & Lozier, R. H. 1987. Flash spectroscopy of purplemembrane. Biophys. J., 51(4), 627–635.

Yamashita, Hayato, Uchihashi, Takayuki, Voıtchovsky, Kislon, An-toranz Contera, Sonia, Yamamoto, Daisuke, Ryan, John F., & Ando,Toshio. 2006. High-resolution dynamic imaging of membrane proteins by high-speedAFM. Jap. Biophys. Soc. Meeting Abstract.

Yokokawa, M., Yoshimura, S. H., Naito, Y., Ando, T., Yagi, A., Sakai, N.,& Takeyasu, K. 2006. Fast-scanning atomic force microscopy reveals the molecularmechanism of DNA cleavage by ApaI endonuclease. IEE Proc. Nanobiotech., 153(4),60.

190

Towards bio-nanotechnological applications

6 Towards bio-nanotechnological applications

6.1 Introduction

6.1.1 A single molecule photo-detector

Reaching the level of single molecules is one of the ultimate goals of bionanotechnology.

It is indeed the fundamental starting point of any hybrid technology that would combine

biological molecules and electronics. Controlling that point would allow the design of

new devices that would take advantage of the highly efficient solutions which already

exist for specific biological tasks.

The development of a single bio-molecule based photo-detector presents a good starting

point because it offers a possibility to evaluate the device by correlating the signal mea-

sured with a photon flux on the detector. Besides, such a device would present many

advantages over traditional photo-detectors. First the size of the pixels would be about

three order of magnitude smaller (e.g. with bR) than on a typical Coupled Charge Device

(CCD), providing an unprecedented resolution with color sensitivity and on a timescale

of milliseconds. The high quantum efficiency of the bR photo-detection process (> 60%

(Govindjee et al. 1990)) also exceeds that of most of the common semiconductor based

detectors. Finally, such a technology exploits the naturally optimized biological solution

available and could be achieved through self-assembly processes, which would guarantee

a low production cost. From a point of view of fundamental science, a single bio-molecule

device would open a whole new world of possibilities for single molecules studies by us-

ing the device as a nano-laboratory that can monitor the molecule activity in different

environments while applying various constraints.

Numerous studies have already exploited bR in bio-electronics, (Birge et al. 1999; Bhat-

tacharya et al. 2002; Crittenden et al. 2003; Jussila et al. 2002; Wise et al. 2002)

191


bio-optics (Birge 1990; Chen & Birge 1993; Gu et al. 1996; Hillebrecht et al. 2004),

bio-computing (Birge 1992; Gu et al. 1996) and nano-patterning (Douglas et al. 1992).

Besides its photo-induced proton pumping activity which involves optical and structural

changes, bR also presents a long-term stability against thermal, chemical and photochem-

ical degradation (Shen et al. 1993). So far, the bR-based devices achieved, required

a large number of bR molecules and multiple layers of PM in order to be functional

and/or reach detectable signals. With the recent development of local probe techniques

and nanolithography, single molecule devices become feasible, although very challenging.

There are still many difficulties to overcome and only very few of the results presented so

far involve a single molecule system (Uppenbrink & Clery 1999, Special edition: ”Fron-

tiers in Chemistry: Single Molecules”). The main problems are first the molecule trap-

ping, and second the achievement of contacts that relevantly link the trapped molecule

with the outside, in order to be able to detect its activity. Furthermore, at the single

molecule level, signal detection becomes a challenge in itself due to the weakness of the

signal. In the case of bR, the protein could act as a light driven switch (Birge 1990).

Sandwiched between two nano-electrodes, the photo-activation of bR could be monitored

through changes in tunnelling current between the two nano-electrodes, induced by bR

large conformational changes occurring during the photocycle (Kamikubo & Kataoka

2005). This is motivated by the fact that bR can naturally conduct protons by reso-

nant tunnelling through a path involving its chromophore. It has indeed been showed

that tunnelling through the protein is naturally three to four orders of magnitude more

efficient than through a lipid bilayer, and that the magnitude of tunnelling current de-

pends on the activation state of the protein (Jin et al. 2006). Beside the difficulties

mentioned above for single-molecule systems, the development of such a device requires

nano-electrodes with a gap comparable to the protein size, and the ability to measure

very small current with a good signal to noise ratio.

192


6.1.2 Chapter Overview

The development of a single-biomolecule device presents an enormous potential but also

a considerable challenge for both science and technology and the achievement of such

devices is long term project. The work presented in this chapter is still in progress and

many difficulties need to be overcome before the achievement of a reliable and repro-

ducible single biomolecule-based detector.

The first part of this chapter describes some preliminary studies towards the realization

of a bR-based single molecule photo-detector. PM monomerization and electrical prop-

erties are studies by AFM. Different strategies are tried for trapping single bR molecules

between nano-electrodes, and are discussed with issues related to the electrodes em-

ployed. Starting from a ”naive” approach, each step of the development is then guided

by the experience previously acquired. Whenever possible, self-assembly procedures have

been used, in order to find solution relevant for technology and large scale production.

In the second part of the chapter, the development of a special low current measure-

ment setup is presented. Its ultimate goal is to record changes in the current tunnelling

through single proteins, and the setup signal to noise ratio is optimized for low currents

measurements. The setup is successfully used for bio-sensing proteins attachment to

nitrogen-doped carbon nanotubes.

The last section summarizes and discuss the different issues related to single molecule

bio-sensing and suggests ways for future improvements.

193


6.2 Preliminary studies

6.2.1 Monomerization of bR

Since bR naturally occurs as PM, it is necessary to extract the protein from its natural

membrane in order to use a single bR molecule. The most logical way to proceed is

through bR monomerization with a detergent (Triton X100) that solubilizes the protein

into micelles and small vesicles (Schertler et al. 1993). The procedure followed is detailed

in Appendix 8.4. AFM imaging of the bR vesicles in a salt containing buffer (300 mM

KCl, 10 mM Tris, pH 7.8) showed a size distribution between 20 nm and 50 nm (Fig.

6.1). The absorption spectrum of the monomerized membrane allowed procedure control

(spectral shift (Dencher et al. 1982)) and verification of the protein functionality (spectra

not shown). The solution containing the vesicle could be stored at 4 C for up to

one month without aggregation. It however required adjunction of a small amount

of detergent (0.01% DDM, n-dodecyl-β-maltoside) and mild sonication for ∼ 1 min

prior to use. Some native PM lipids are specifically and tightly bound to bR and their

removal affects the functionality of the protein (Huang et al. 1980; Cartailler & Luecke

2003). The monomerization procedure followed does not remove these lipids, which may

account for the lower size limit of the vesicles observed. The weak attachment of the

smaller vesicles to the substrate could also prevent their visualization by AFM. Drying

the sample revealed smaller but non-identifiable features (Fig. 6.1.B) and highlighted

the necessity to limit the amount of detergent (< 0.05% DDM) in solution for using the

monomerized sample in air (Fig. 6.1.C).

6.2.2 Electrical conductivity properties of Purple Membrane

The feasibility of a device using bR molecules as light-driven current nano-switches de-

pends on a reliable amplification of the process. It is therefore important to know how

194


A B C

Figure 6.1: Monomerized bR adsorbed on mica AC-AFM in ultrapure water with 0.01%DDM (centrifugation and resuspension from the storage buffer, see Appendix 8.4) (A) and inair (B). An excess of detergent in the monomerized bR adsorption buffer (0.1% DDM) formslarge aggregates when dried (C). The scales are (A) 100 nm × 100 nm, (B) 80 nm × 80 nm,and (C) 3 µm × 3 µm.

much current can tunnel through the protein natural channel when a voltage is applied.

Tunnelling charges will preferentially flow through bR molecules rather than lipids (Jin

et al. 2006), but the actual current-voltage (IV ) dependency is unclear because it highly

depends on the metal-protein contacts. Several studies of PM in ambient conditions1

using STM (Wang et al. 1990; Niemi et al. 1993; Tamayo et al. 1995; Garcıa et al.

1997) and larger electrodes (Jin et al. 2006) have provided IV characteristics ranging

from 10−18 A/V per bR trimer when estimated from large electrodes (0.2 mm2) up to

10−10 A/V per bR trimer from STM2. This is hardly surprising since the tunnelling cur-

rent depends exponentially on the tunnelling distance and on the protein-metal contact.

Other parameters such as the shape of the electrode/STM tip, the voltage applied and

the relative humidity can modify the current measured in a highly non-linear fashion.

The best reproduction of a system featuring bR between two nano-electrodes can be

achieved using conducting AFM to mimic nano-electrodes. The size of the tip is com-

parable to that of the electrodes and its position in respect to the membrane is better

1 In air, at room temperature and with 40-85% relative humidity 2 This is assuming the current flowsfrom the STM tip to the substrate through a single trimer

195


controlled than with STM, because determined by a feedback on the cantilever deflec-

tion rather than on a tunnelling current. PM deposited on gold and graphite (Highly

Oriented Polygraphite, HOPG) was scanned in contact mode AFM with tip-sample bias

between 0.5 V and 3 V (Fig. 6.2). Recording the current tunnelling through the sample

while scanning allows the superimposition of a conductivity map of the sample to the

topographic image. The tunnelling current contrast between a single layer of PM and

the substrate is of about 200 pA (also in the case of PM on gold, not shown) and reverse

tip-sample voltages consistently provide reversed contrasts of the current. However, the

effective sensitivity of the current measurement system (∼ 5 pA/V max. sens.) did

not allow the resolution of local variations of the tunnelling current through the mem-

brane apart for possible holes and defects (Fig. 6.2.B-C), but the recent appearance of

commercial systems able to monitor current down to a few tens of fA should allow this

difficulty to be overcome. Importantly, it will also allow a reduction of the tip-sample

bias. Applying a bias is indeed problematic because it modifies the tip-sample interac-

tion and consequently the tip-substrate absolute distance. This impacts exponentially

on the tunnelling current measured. Furthermore, the important electric field at the tip

apex3 can induce local electro-chemical changes of the sample such as pH changes and

redox reactions. It is therefore preferable to work with a tip-sample voltage as small as

possible. Scanning PM with voltages > 3V destroyed the sample and the tip.

6.2.3 Nano-gap electrodes for bio-molecules

The electrodes were kindly provided by Dr. Philip Steinmann from the research group of

Prof. John M. R. Weaver, Nanoelectronics Research Centre, Department of Electronics

and Electrical Engineering, Glasgow University, Glasgow G12 8LT, United Kingdom.

3 If we model the AFM tip apex with a sphere of radius r, then E ∝ CV/r2 ∼ V/r with V the tip-samplebias and C the tip-sample capacitance. Typically r ∼ 10−8m.

196


A

B C

Figure 6.2: PM on HOPG imaged in air, in contact mode with conducting AFM. (A) Topo-graphic image. (B-C) Current images with a tip-sample bias of +3 V and −3 V respectively.Sections of (B) and (C) are shown below each picture. Scales are 500 nm × 500 nm for allthe images and 50 pA for the sections of (B) and (C) (vertical graduation).

One of the main difficulties in the development of a single molecule device involving nano-

electrode lies in a reproducible manufacturing of the electrodes. Conventional lithogra-

phy techniques can reliably achieve structures with a precision of a about 10 − 20 nm

(Steinmann & Weaver 2004; Kronholz et al. 2006). In order to build sub-10 nm gap

electrodes, further metal evaporation (Steinmann & Weaver 2004, 2005) or electrochemi-

cal deposition (Wang et al. 2005) is necessary, leading to less reproducible gap distance.

In order to trap bio-molecules such as bR in the nano-gap, it is necessary to immerse the

gap region in an solution containing the protein while applying a voltage (DC or dielec-

trophoretic trapping) and recording the inter-electrode tunnelling current. However, the

ionic current induced by the work in liquid makes the detection of sub-pA signals very

difficult. The accuracy of the measurement therefore necessitates electrical insulation

around the gap. Some studies have demonstrated the feasibility of insulated electrodes

(Kronholz et al. 2006), but not with gaps reaching the size of a single bR molecule. We

197


opted for nano-gap electrodes (Steinmann & Weaver 2004) coated with a ∼ 100 nm layer

of insulant (Polymethyl Methacrylate, PMMA) except for the gap region (Fig. 6.3.A-B).

PMMA is hydrophobic and disfavors bR adsorption, which allows the confinement of

the adsorption to the uncoated nano-gap region. The removal of PMMA from the gap-

region is carried out by electron-beam which imposes a lower limit of ∼ 20 µm in the

diameter size of the uncoated region. Further improvement of the measurements noise

level has been imagined using the AFM to scratch fully coated electrodes over the gap,

thus revealing only the gap itself when the PMMA insulation is partially plasma-etched

(Fig. 6.3.D). Although feasible, the method is difficult to reproduce because of the poor

control over the scratch process (data not shown).

6.2.4 Trapping bacteriorhodopsin between nano-electrodes

At the single molecule level, there is no obvious way to trap and orient bR between two

nano-electrodes and different options have been considered. The most common trapping

methods can be divided in two categories; the methods using an electric field (electro-

or dielectrophoretic trapping) (Hughes 2000; Bezryadin et al. 1997; Lifeng et al. 2003)

and the methods relying on capillary forces (Cui et al. 2004).

The first strategy tried for trapping bR combines the two above methods. The idea is

to use a DC electric field generated by the electrodes to trap bR vesicles diffusing in the

immediate gap region and orient them. bR molecules possess a natural permanent dipole

(∼ 55D experimentally (Porschke 1996)) because of its charge asymmetry between the

cytoplasmic and extracellular sides of the protein. The possible orienting of the protein

dipole with an external field has already been demonstrated (Kimura et al. 1984) and

the direction would be relevant for tunnelling through the protein natural proton path.

A ∼ 3 µl drop of solution containing monomerized bR vesicles was deposited on the gap

region of the electrodes. Diffusion of the protein vesicles across the trap can in principle

198


Profile view

Top view

AuSiO2

PMMA AFM tip Oxygen plasma

21 43

A B C

D

Figure 6.3: Optical microscope (A) and AFM image (B-C) of the nano-electrodes. Thehole in PMMA is visible in (A) (arrow). The nanogap (arrow in (B), (C)) is located on aguard electrode visible in (B). (D) describes the AFM-based manufacturing of the electrodeinsulation. The scale bars are 20µm (A), 3 µm (B) and 50 nm (C).

199


be induced by making the edge of the drop meniscus cross the gap while drying (Cui

et al. 2004). Fig. 6.4.A shows a typical trapping trial using this method. The distri-

bution of bR vesicles does not allow us to conclude that the trapping is effective despite

an apparent sufficient vesicle density. This can be due to several reasons preventing a

vesicle diffusing across the gap to be trapped; first the ions present in the solution form

a double layer on the electrodes and the vesicles surfaces, shielding the static potential.

The water molecules also orient according to the electric potential, attenuating it over

short distances (Ataka et al. 1996). Finally, vesicles containing several bR molecules

probably show a reduced effective dipole moment. The trapping DC electric field ac-

knowledges an upper limit due to the limited voltage that can be applied to the electrode

without inducing electrochemical damage to the electrodes. As for the AFM tip apex,

the curvature radius of the electrodes at the gap enhance the intensity of the field and

destructive effects are observed for voltages & 1 V (Fig. 6.4.E). As a general rule, the

absolute value of the electrode bias was kept below 1 V in further trials.

A second strategy was tested using PM extensively rinsed in ultra-pure water (18.2 MΩ

0 400200

0

-20

20

-40[s]

[fA]

ON ON

OFFOFF

A B C D E

Figure 6.4: bR trapping tried with monomerized PM (A), and with PM in ultrapure water(B-C). (C) is the phase image of (B), allowing visualization of PM patches adsorbed overthe electrode. (D) is an example of light-induced FET effect on the nano-electrodes. (E)Electrodes destroyed by a too large trapping DC voltage (+5 V ) inducing electro-chemicalreactions in buffer. The scale bars are (A) 200 nm, (B-C) 150 nm and (E) 1 µm.

MiliQ) by successive centrifugations and re-suspensions. The idea is to adsorb whole

200


PM patches on the electrodes, but operating in ultra-pure water. A DC electric field

would then allow a near single bR molecule or trimer to be extracted from the weak-

ened membrane (see chap. 3). Furthermore, working in ultra-pure water minimizes ionic

current and therefore the measurement noise level. Fig. 6.4.B-C shows PM patches

adsorbed on the nano-gap after a trapping trial. The images do not allow however a

conclusion of whether bR molecules are trapped in the electrodes gap. IV curves taken

before and after the trapping trial exhibit some differences in the tunnelling current (Fig.

6.6), suggesting that some molecules are trapped in the gap or close to it. When illumi-

nated with white light, the system showed a response correlated with the photon flux on

the gap (Fig. 6.4.D). However, the typical shape and time characteristics of the signal

(exponential decay over ∼ 2 − 3 s) seems too slow to represent the activity of a small

number of molecules. A control experiment carried out on unused nano-electrodes ex-

hibited a similar light response, suggesting that the signal is generated by photo-induced

electron-hole pairs in the electrodes Si substrate. The gold layer constituting the elec-

trodes (∼ 30 nm) and the insulating SiO2 layer are not sufficient to stop all the incident

photons, and the system electrode/insulator layer/semiconductor substrate functions as

a light-gated FET. It is then very difficult to discriminate the signal induced by electrode

photo-charges and possible bR induced effects. Several steps have been undertaken to

minimize the electrodes photo-effect, mainly by limiting the spatial distribution and the

wavelength of the excitation light employed (see §6.3.2 below).

The most promising method for single molecule trapping (monomerized bR) is dielec-

tric trapping. Using this method, a recent study has reported trapping of a single

R-phycoerythrin4 molecule between two nano-electrodes (Holzel et al. 2005). Despite

some controversy over the number of molecules trapped (Ying et al. 2006; Holzel et al.

2006), the result demonstrates the ability of dielectric trapping methods to operate effi-

4 R-phycoerythrin is a 240KDa protein from red algae and produces an intense auto-fluorescence

201


ciently in liquid on bio-molecules.

6.3 Measurement setup

6.3.1 Implementation of a low current measurement system

In order to be able to measure changes in the current tunnelling through a single trapped

protein, the measurement setup must show an outstanding sensitivity. This has been

achieved by optimizing a KeithleyTM 6340 low current measurement system for further

reduction of the noise level while improving the sampling speed. An important noise

reduction was obtained through extensive electromagnetic and acoustic shielding of the

wiring, connections and setup (Fig. 6.5). The probes (contacting the gold electrodes)

are enclosed in a grounded Faraday cage and the connection to the preamplifier passes

through the shielding to a second cage (Fig. 6.5.A-C). The main cage is acoustically

isolated with foam and mounted on an an active anti-vibration table. This insulation

reduced the noise by two order of magnitude; the noise level measured over the whole

system is typically ∼ 4 fA with a constant voltage applied and ∼ 10 fA on a voltage

scan from between −1 V and 1 V (Fig. 6.5.E, 1 s sampling time/point). An acquisition

software able to optimize the sensitivity level of the setup according to the sampling

time/point desired while applying voltages has been developed using LabviewTM . In

order to increase the accuracy and the measurement speed of the setup for recording at

the pA level, a four probes sensing was implemented (Fig. 6.5.B, 6.5.D), but attention

was payed to keep it independent from the lower noise two probe sensing system. Finally

two supplementary probes able to transmit high frequency signals have been implement

to the setup, in order to apply alternating voltages for dielectric trapping (Fig. 6.5.D).

202


-1 1-0.5 0 0.5

0-4-8

[fA]

[V]

A B C D

E

Figure 6.5: Low current measurement setup. (A) Main Faraday cage with probes inside. Thecage is mounted on an active anti-vibration table. (B) Nano-electrode contacting pads withfour probes (lower part), and guard probe contact (arrow). (C) Shielded connection box withindependent two-probes/four-probes capability. (D) All probes and lenses system focusing thelight excitation beam (arrow) coupled to an optical fiber. (E) Typical baseline noise over a0 V → +1 V → −1 V → 0 V sweep. Each point is sampled over 1 s.

6.3.2 Current vs Time and Voltage curves; the effect of light

Using the setup described in the previous section, the tunnelling current through nano-

gap electrodes could be measured reliably and reproducibly. When applying a voltage to

a nano-electrode contacting pads (Fig. 6.5.B) the tunnelling current measured followed

a multiple-exponential decrease with time, characterizing the capacitor formed by the

nano-gap and the measurement setup (Fig. 6.6.D). Typical current-voltage (IV) curves

and current vs time at a given applied voltage (IT) curves are presented in Fig. 6.6.

The discharging of the nano-gap capacitor typically occurs over several tens of seconds.

Fig 6.6 shows IV-curves before and after PM adsorption on the electrodes nano-gap (see

section ”Trapping of bacteriorhodopsin”). A blank trial with ultrapure water (Fig. 6.6.A)

shows that if enough time is allowed after drying, IV curves measured are similar to these

before the electrode immersion and the wetting/drying process is fully reversible. With

a solution containing PM, however, the curves after drying remain almost unchanged

with time (Fig. 6.6.C), exhibiting a shape characteristic of an immersed nano-gap (Fig.

203


before

1-1 -0.5 0 0.5-1

-0.5

0

0.5

1

1.5

[V]

[pA]

1-1 -0.5 0 0.5 [V]

-60

-20-40

02040

[pA]

1-1 -0.5 0 0.5 [V]

0

0.5

-0.5

1 h

15 min30 min

[pA] dryH2O

adsorption

1 h

15 min30 min

0

20

10

30

40

50

[pA]60

0 100 200 300 400 500 600 7000 100 200 300 400

0

0.2

0.4

0.6[pA]

experimentaldouble exp. fit

PM adsorption

A B C

D E

Figure 6.6: (A) IV curves of a new/clean electrode in air (before) and drying after having beenimmersed in ultrapure water. The times specified in the legend represent minutes waited afterthe disappearances of the water droplet on the nano-gap (observed by optical microscopy). (B)Same as (A) but measured with the nano-electrode gap actually immersed. (C) Same principleas (A) but measured with PM adsorbed on the nano-electrodes (Fig. 6.4.B-C) immediately(adsorption) and at different times after the disappearance of the solution droplet. (D) Typicaltemporal decay of the nano-gap tunnelling current after a change (0 V → +1 V ) of the appliedbias. Several timescales are necessary for a suitable fit of the decay (2.3 s and 38 s with animportance ratio of ∼ 3). (D) Recording of the tunnelling current during PM adsorption (C)

204


6.6.B). This suggests that PM is adsorbed on the nano-gap (as visible in Fig. 6.4.B-C),

and maintains, at least partially, its hydration in ambient conditions. Despite differences

in the gap capacitance accounting for the adsorbed membrane, light induced activity of

bR cannot be reliably measured because of the light-induced FET effect of the electrodes,

which superimposes to a potential bR signal (Fig. 6.7). In order to characterize and

ultimately discard the electrode related light induced effect, two strategies have been

adopted. Firstly, the size of illumination spot has been considerably reduced and focused

to the gap region by the mean of an optical fiber coupled to a system of lenses (Fig.

6.5.D). Secondly, a sample of different well characterized (emission spectrum, intensity)

light emitting devices (LEDs) has been implemented, allowing to couple directly the

optical fiber to the different sources without modifying the focus of the beam on the

nano-gap. The effect of the different wavelength could then be recorded on a blank

electrode with and without the lens system (Fig. 6.7) Interestingly, focusing the light

beam on the gap region was sufficient to reduce the light-induced FET effect to a level not

detectable within the measurement accuracy (Fig. 6.7.B). It also presents the advantage

of increasing the density of photons on the nano-gap (typically > 100 photons/s ·nm2)5.

The effective absorption section σ at of single bR molecule can be deduced from PM

extinction coefficient at a given wavelength (see Appendix 8.2) For λ = 530nm (green),

σ ∼ 1.9 ·10−2 nm2. Compared with the photon density on the gap, each bR molecule can

be excited up to ∼ 2 per second and isomerize (assuming a monolayer of bR molecules).

Applying an alternating voltage to the nano-electrodes slightly increases the noise level

but also reduces the impact of external stimuli such as the light-induced FET effects,

provided that the amplitude of the alternating voltage is larger than any imposed voltage

offset and that the alternating frequency is high enough. Alternative voltages could

5 Estimated for the green LED, with the approximation that all the light intensity is emitted from thepeak emitting wavelength

205


20 s

20 s

20 fA/cd

20 fA/cd

ON OFF630 nm

587 nm

525 nm

470 nm

white

A

B

Figure 6.7: (A) Light-induced FET effect observed with IT curves, recorded for nano-electrodes stimulated with different LEDs. The illumination was obtained without the lensessystem focusing the exciting beam on the nano-gap (only the fiber). (B) Same as (A), butwith the lens system focusing white light (LED) on the nano-gap. All the curves have beenrecorded with a constant bias of +1 V . The tunnelling current measured was divided by theluminous intensity (candela, cd) for each LED in order to allow comparison. The samplingtime was set to 1 s.

206


indeed be useful for dielectric trapping, in order to reduce ionic and polar shielding of

the applied field by the buffer. Fig 6.5 presents the nano-electrode tunnelling current

response to different frequencies, each applied with a 2 V peak to peak amplitude (2 Vpp)

and with a voltage offset of +0.5 V . At frequencies higher than ∼ 10 Hz already, the

current level is greatly reduced (Fig. 6.7.A) and for frequencies > 100 Hz the current

measured is within the noise level, apart for the voltage offset inducing the characteristic

exponential decay (Fig. 6.7.B). The effect of light, measured with a 1 MHz, 2 Vpp

AC voltage, 0.5 V offset and the white LED could not be detected (not shown). These

-4

4

2

-2

0

2 s 2 s

-0.3

-0.2

-0.15

-0.25

[pA][pA] 10 Hz100 Hz1 kHz

10 kHz100 kHz1 MHz

1 Hz2 Hz3 Hz

4 Hz5 Hz8 Hz

10 Hz

A B

Figure 6.8: Tunnelling current measured for the nano-electrodes with an imposed AC voltageof 2 Vpp and superimposed with a DC offset of +0.5 V . Measurements at frequencies lowerthan 10 Hz (A) are substantially affected whereas little effect is visible for higher frequencies(B)

developments show promising results, both for trapping and measuring bR activity using

nano-gap electrodes. New trials and measurements making use of theses improvements

will reveal how effective they are, and open the way for further developments.

6.3.3 Bio-sensing metalloproteins with CNx carbon nanotubes

Using the low current measurement setup described above, reliable and reproducible sens-

ing of metalloproteins could be achieved with a nitrogen doped carbon nanotube based

207


sensor developed by H. Burch (Burch et al. 2006c).

Carbon nanotubes (CNTs) present particular electronic properties that can be affected

by their chemical environment (Kong et al. 2000), and which makes them a material

of choice for label-free biosensing. In the recent years, many studies have developed

CNT-based devices capable of reliably and reproducibly detecting biomolecules in small

amounts of solution (typically microliters) (Chen et al. 2003, 2004; Wang et al. 2004).

CNTs can be doped with foreign atoms or molecules such as nitrogen (Czerw et al. 2001),

potassium and boron (Lee et al. 1997). The doping affects the CNTs properties and

multiwalled nitrogen doped CNTs (CNx MWNTs) show a consistent high conductance,

independently of their chirality (Burch et al. 2006a). Furthermore, the nitrogen located

in the CNx MWNT sidewalls present hydrophilic sites suitable for allowing the binding of

proteins without affecting their folding (Burch et al. 2006b), thus making CNx MWNTs

particularly good candidates for biosensing. Using a CNx MWNTs-based electro-sensor

(Burch et al. 2006c), it was possible to measure binding of metalloproteins (ferritin

and cytochrome c) to the CNx MWNTs in solution (Fig. 6.9). Binding of anti-ferritin

to an already ferritin-coated nanotube triggers a measurable conductance change (Fig.

6.9.B). Control experiments with the apoprotein (without the metal atom) and with the

anti-protein demonstrated that metal atom did not play a major role in the conductance

signal detected (Burch et al. 2006c).

6.4 Discussion and Conclusions

Biosensing at the single-molecule level gathers many advantages such as unprecedented

size and sensitivity, low cost and the possibility to built higher level complex architec-

tures through bottom-top assembly techniques. To date different types of biosensors are

available commercially, already developed on a large scale. None of these sensors has

208


ferritin

anti-ferritin IgG

ferritinanti-ferritin

buffer

0 80 160 240 320 400 4800

50

100

150

-100

-50

50

100

0

-1 10 [V] [s]

[µA][µA]

A B

Figure 6.9: Biosensing of ferritin and anti-ferritin binding to CNx MWNTs. (A) IV curvesfor the biosensor in a buffer containing 5 mM NaH2PO4 5 mM NaCl, pH 6.01 (squares) andwith subsequent adjunction of 1 µM of ferritin (circles) or 1 µM of anti-ferritin (diamonds)in solution in the buffer. (B) Biosensing IT trace recorded in buffer at +1 V . Consecutiveadjunction of 1 µM ferritin and 1 µM anti-ferritin to the stabilized system reveals clear stepsin current.

however reached the single molecule level and more developments are necessary before

being able to control reliably and routinely the integration of single biomolecule in a

functional device. One possible approach is presented in this chapter, through the use

of nano-gap electrodes as electro-mechanical contacts for a trapped biomolecule. The

strong field-effect present at the nano-gap can be utilized for trapping and orienting

the sensing biomolecule, allowing control of the assembly at the single molecule level.

Aiming towards a single bR molecule based system, a setup capable of applying arbi-

trary voltage sequences while recording the tunnelling current at fA level between the

nano-electrodes has been developed, and the response of the nano-electrodes to different

wavelength characterized. Further experiments are required to study the effectiveness of

dielectric trapping on bR, and characterize the protein-metal contacts. The main dif-

ficulty, however, comes from the lack of reproducibility when manufacturing electrodes

with so small gaps. Below ∼ 10 nm the control of the gap size becomes problematic due

209


the manufacturing process involving metal evaporation (Steinmann & Weaver 2005). A

reproducible gap size is necessary to achieve trapping and sensing experiments that are

comparable, and despite some progress in the manipulation of single biomolecules, the

main fundamental limitation of this approach appears to come from the nano-electrodes.

Future developments will have to overcome this limitations in order to be relevant for

large scale productions, either by finding ways to circumvent the need of metallic nano-

structures or by achieving reproducible templates that allow integration single-molecules.

Conducting carbon nanotubes could provide a reproducible and efficient alternative to

metallic nano-structures (Burch et al. 2006a). An nitrogen doped carbon nanotubes

based biosensor could be achieved (Burch et al. 2006c) and characterized with the

low current measurement setup, demonstrating a molecule-selective response of the de-

vice. Combining conducting carbon nanotubes with the developments achieved for nano-

electrodes could provide a new base for single-molecule biosensors.

6.5 References

Ataka, K., Yotsuyanagi, T., & Osawa, M. 1996. Potential-Dependent Reorien-tation of Water Molecules at an Electrode/Electrolyte Interface Studied by Surface-Enhanced Infrared Absorption Spectroscopy. J. Phys. Chem., 100(25), 10664–10672.

Bezryadin, A., Dekker, C., & Schmid, G. 1997. Electrostatic trapping of singleconducting nanoparticles between nanoelectrodes. App. Phys. Lett., 71(9), 1273.

Bhattacharya, Pallab, Xu, Jian, Vr, Gyorgy, Marcy, Duane L., & Birge,Robert R. 2002. Monolithically integrated bacteriorhodopsin-GaAs field-effect tran-sistor photoreceiver. Opt. Lett., 27(10), 839–841.

Birge, R. R. 1990. Photophysics and Molecular Electronic Applications of theRhodopsins. Annu. Rev. Phys. Chem., 41(1), 683–733.

Birge, R. R. 1992. Protein-based optical computing and memories. Computer, 25(11),56–67.

Birge, R. R., Gillespie, N. B., Izaguirre, E. W., Kusnetzow, A., Lawrence,A. F., Singh, D., Song, Q. W., Schmidt, E., Stuart, J. A., Seethara-

210


man, S., & Wise, K. J. 1999. Biomolecular Electronics: Protein-Based AssociativeProcessors and Volumetric Memories. J. Phys. Chem. B, 103(49), 10746.

Burch, Hilary J., Davies, J. A., Brown, E., Hao, L, Antoranz Contera,S., Grobert, N., & Ryan, J. F. 2006a. Electrical conductance and breakdown inindividual CNx multiwall nanotubes. Appl. Phys. Lett., in press.

Burch, Hilary J., Antoranz Contera, Sonia, de Planque, M. R. R.,Grobert, Nicole, & Ryan, J. F. 2006b. Functionalization of CNx multi-wallednanotubes with correctly folded metalloproteins. J. Am. Chem. Soc., submitted.

Burch, Hilary J., Antoranz Contera, Sonia, Voıtchovsky, Kislon,Grobert, Nicole, & Ryan, J. F. 2006c. Sensing of metalloproteins with a CNx

nanotude device. in preparation.


Chen, R. J., Choi, H. C., Bangsaruntip, S., Yenilmez, E., Tang, X., Wang,Q., Chang, Y. L., & Dai, H. 2004. An Investigation of the Mechanisms of ElectronicSensing of Protein Adsorption on Carbon Nanotube Devices. J. Am. Chem. Soc.,126(5), 1563–1568.

Chen, Robert J., Bangsaruntip, Sarunya, Drouvalakis, Katerina A.,Wong Shi Kam, Nadine, Shim, Moonsub, Li, Yiming, Kim, Woong, Utz,Paul J., & Dai, Hongjie. 2003. Noncovalent functionalization of carbon nanotubesfor highly specific electronic biosensors. Proc. Natl. Acad. Sci. USA, 100(9), 4984–4989.

Chen, Zhongping, & Birge, Robert R. 1993. Protein-based artificial retinas.Trends Biotechnol., 11(7), 292.

Crittenden, S., Howell, S., Reifenberger, R., Hillebrecht, J., &Birge, R. R. 2003. Humidity-dependent open-circuit photovoltage from a bacte-riorhodopsin/indium tin oxide bioelectronic heterostructure. Nanotechnology, 14(5),562.

Cui, Y., Bjork, M. T., Liddle, J. A., Sonnichsen, C., Boussert, B., &Alivisatos, A. P. 2004. Integration of Colloidal Nanocrystals into LithographicallyPatterned Devices. Nano Lett., 4(6), 1093–1098.

Czerw, R., Terrones, M., Charlier, J. C., Blase, X., Foley, B., Ka-malakaran, R., Grobert, N., Terrones, H., Tekleab, D., Ajayan, P. M.,Blau, W., Ruhle, M., & Carroll, D. L. 2001. Identification of Electron DonorStates in N-Doped Carbon Nanotubes. Nano Lett., 1(9), 457–460.

211


Dencher, Norbert A., Heyn, Maarten P., & Lester, Packer. 1982. Prepara-tion and properties of monomeric bacteriorhodopsin. Page 5 of: Methods Enzymol.,vol. Volume 88. Academic Press.

Douglas, Kenneth, Devaud, Genevieve, & Clark, Noel A. 1992. Transferof Biologically Derived Nanometer-Scale Patterns to Smooth Substrates. Science,257(5070), 642–644.

Garcıa, R., Tamayo, J., & Bustamante, C. 1997. Scanning tunneling microscopyimaging and selective modification of purple membranes. Intl J. Imag. Syst. Technol.,8(2), 168–174.

Govindjee, R., Balashov, S. P., & Ebrey, T. G. 1990. Quantum efficiency of thephotochemical cycle of bacteriorhodopsin. Biophys. J., 58(3), 597–608.

Gu, Li-Qun, Zhang, Chun-Ping, Niu, An-Fu, Li, Jia, Zhang, Guang-Yin,Wang, Yong-Mei, Tong, Ming-Rong, Pan, Ji-Lun, Wang Song, Q., Par-sons, Bruce, & Birge, Robert R. 1996. Bacteriorhodopsin based photonic logicgate and its applications to grey level image subtraction. Opt. Comm., 131(1-3), 25.

Hillebrecht, J. R., Wise, K. J., Koscielecki, J. F., & Birge, R. R. 2004.Directed evolution of bacteriorhodopsin for device applications. Methods Enzymol.,388, 333.

Holzel, Ralph, Calander, Nils, Chiragwandi, Zackary, Willander, Mag-nus, & Bier, Frank F. 2005. Trapping Single Molecules by Dielectrophoresis. Phys.Rev. Lett., 95(12), 128102.

Holzel, Ralph, Calander, Nils, Chiragwandi, Zackary, Willander, Mag-nus, & Bier, Frank F. 2006. Holzel et al. Reply. Phys. Rev. Lett., 96(19), 199802.

Huang, Kuo-Sen, Bayley, Hagan, & Khorana, H. Gobind. 1980. Delipidationof Bacteriorhodopsin and Reconstitution with Exogenous Phospholipid. Proc. Natl.Acad. Sci. USA, 77(1), 323–327.

Hughes, Michael Pycraft. 2000. AC electrokinetics: applications for nanotechnol-ogy. Nanotechnology, 11(2), 124.

Jin, Yongdong, Friedman, Noga, Sheves, Mordechai, He, Tao, & Cahen,David. 2006. Bacteriorhodopsin (bR) as an electronic conduction medium: Currenttransport through bR-containing monolayers. Proc. Natl. Acad. Sci. USA, 103(23),8601–8606.

Jussila, T., Li, M., Tkachenko, N.V., Parkkinen, S., Li, B., Jiang, L.,& Lemmetyinen, H. 2002. Transient absorption and photovoltage study of’ self-assembled bacteriorhodopsin/polycation multilayer films. Biosens. Bioelec., 17(6),509–515.

212



Kimura, Y., Fujiwara, M., & Ikegami, A. 1984. Anisotropic electric properties ofpurple membrane and their change during the photoreaction cycle. Biophys. J., 45(3),615–625.

Kong, Jing, Franklin, Nathan R., Zhou, Chongwu, Chapline, Michael G.,Peng, Shu, Cho, Kyeongjae, & Dai, Hongjie. 2000. Nanotube Molecular Wiresas Chemical Sensors. Science, 287(5453), 622–625.

Kronholz, S., Karthauser, S., Meszaros, G., Wandlowski, Th, van derHart, A., & Waser, R. 2006. Protected nanoelectrodes of two different metalswith 30 nm gapwidth and access window. Microelec. Engin., 83(4-9), 1702.

Lee, R. S., Kim, H. J., Fischer, J. E., Thess, A., & Smalley, R. E. 1997.Conductivity enhancement in single-walled carbon nanotube bundles doped with Kand Br. Nature, 388(6639), 255.

Lifeng, Zheng, Shengdong, Li, Burke, P.J., & Brody, J.P. 2003. Towards singlemolecule manipulation with dielectrophoresis using nanoelectrodes. Nanotechnology,IEEE-NANO 2003, 1, 437–440.

Niemi, H. E. M., Ikonen, M., Levlin, J. M., & Lemmetyinen, H. 1993. Bacteri-orhodopsin in Langmuir-Blodgett films imaged with a scanning tunneling microscope.Langmuir, 9(9), 2436–2447.

Porschke, D. 1996. Electrostatics and electrodynamics of bacteriorhodopsin. Biophys.J., 71(6), 3381–3391.

Schertler, Gebhard F. X., Bartunik, Hans D., Michel, Hartmut, &Oesterhelt, Dieter. 1993. Orthorhombic Crystal Form of Bacteriorhodopsin Nu-cleated on Benzamidine Diffracting to 3.6 AResolution. J. Mol. Biol., 234(1), 156.

Shen, Yi, Safinya, Cyrus R., Liang, Keng S., Ruppert, A. F., & Rothschild,Kenneth J. 1993. Stabilization of the membrane protein bacteriorhodopsin to 140 Cin two-dimensional films. Nature, 366(6450), 48.

Steinmann, P., & Weaver, J. M. R. 2005. Nanometer-scale gaps between metallicelectrodes fabricated using a statistical alignment technique. Appl. Phys. Lett., 86(6),063104.

Steinmann, Philipp, & Weaver, John. 2004. Fabrication of sub-5 nm gaps betweenmetallic electrodes using conventional lithographic techniques. J. Vac. Sci Technol B,22(6), 3178–3181.

213


Tamayo, J., Saenz, J. J., & Garcıa, R. 1995. Scanning tunneling microscopymodification of purple membranes. J. Vac. Sci Technol A, 13, 1737–1741.

Uppenbrink, Julia, & Clery, Daniel. 1999. Single Molecules. Science, 1593–1804.

Wang, Chengyin, Chen, Yujing, Wang, Fengxiang, & Hu, Xiaoya. 2005. Fab-rication of nanometer-sized carbon electrodes by the controllable electrochemical de-position. Electrochim. Acta, 50(28), 5588.

Wang, J., Liu, G., & Jan, M. R. 2004. Ultrasensitive Electrical Biosensing ofProteins and DNA: Carbon-Nanotube Derived Amplification of the Recognition andTransduction Events. J. Am. Chem. Soc., 126(10), 3010–3011.

Wang, Z., Hartmann, T., Baumeister, W., & Guckenberger, R. 1990. Thick-ness Determination of Biological Samples with a z-Calibrated Scanning TunnelingMicroscope. Proc. Natl. Acad. Sci. USA, 87(23), 9343–9347.

Wise, Kevin J., Gillespie, Nathan B., Stuart, Jeffrey A., Krebs, Mark P.,& Birge, Robert R. 2002. Optimization of bacteriorhodopsin for bioelectronicdevices. Trends Biotechnol., 20(9), 387.

Ying, Liming, Zhou, Dejian, & Bruckbauer, Andreas. 2006. Comment on”Trapping Single Molecules by Dielectrophoresis”. Physical Review Letters, 96(19),199801.

214

Conclusion

7 General discussion and conclusions

Purple membranes are complex systems featuring a broad range of inter- and intramolec-

ular interactions among bR and the lipids. A subtle combination of hydrogen bonds, van

der Waals, steric and electrostatic interactions leads to an efficient functioning of bR. Salt

ions naturally present in PM native medium play a essential role in the local force field

that governs bR structural changes during the photocycle, the proton uptake, release

and lateral transfer at the membrane surface, phospholipids mobility and PM mechani-

cal properties. Both the nature and the concentration of the different ions are important

parameters for tuning PM electrostatic interactions and consequently its metabolic ac-

tivity. AFM high resolution imaging, force spectroscopy and bR unfolding have allowed

a detailed description the interactions governing of bR structure at the molecular level

and highlighted the contributions from tryptophan residues steric interactions, ions spe-

cific binding and water mediated electrostatic interactions in the protein activity. PM

provides a good example of a membrane which leaflets asymmetry fully supports and

catalyze its membrane protein function. Both leaflets exhibit different properties with

specific stable interactions dominating the extracellular leaflet in which bR is anchored,

while ion and water mediated electrostatic interactions govern the cytoplasmic leaflet.

Throughout the work presented in this report, bR has been used as a model system for

integral membrane proteins. This was motivated by the wealth of information already

available on this protein and by bR potential in technological and biomedical applica-

tions. Despite being a particular protein because of its high affinity for salt, several

properties such as the role of tryptophan residues in the protein anchoring and rigidity,

the different ways developed by the protein to exploit ions in order to enhance its activity

and the membrane asymmetry can be used as a relevant base to investigate other sys-

tems. Furthermore, the different AFM and hsAFM approaches developed to investigate

215

Conclusion

PM and bR can be confronted to systems of wider biological and biomedical significance

such as GPCRs. AFM and hsAFM are indeed promising tools to investigate membrane

proteins, not only because they can probe locally and in native-like conditions, but also

because the imaging resolution directly depends on a good understanding of the nature

of the interactions governing the system studied, as illustrated in chap. 4. hsAFM has

also demonstrated its ability to directly image single proteins at work, in real-time, in

native-like conditions and with sub-molecular resolution. These outstanding capabilities

of hsAFM makes it one of the best tools to obtain real-time global and detailed under-

standing of molecular functions and interactions in biological systems.

The biophysical study of membrane proteins is also motivated by bio-nanotechnological

applications that can take advantage of the unequaled efficiency and small size exhibited

by membrane proteins to integrate them in an artificially controlled system. Integration

of biomolecules in artificial system has already been achieved, but at a level involving

typically several thousands or more molecules. The main difficulty lies in the reduction

of number of bio-molecules involved and in mimicking natural amplification processes.

Doped carbon nanotubes could play a decisive role in this amplification process (chap.

6) and open new ways for single bio-molecule devices design.

216

Acknowledgments

Acknowledgments

The work presented in this report could not have been achieved without the help of

several peoples that I would like to thank.

First, I would like to thank Dr. Sonia Antoranz Contera for her enormous effort of su-

pervision and scientific input all along my doctorate. Thank you for teaching me AFM,

for constantly stimulating my interest in new biophysical problems. Thank you also for

you constant support, the time you gave me and for your patience; I could not have

hoped for better direct supervision.

I would like to thank Prof. J. F. Ryan, my supervisor, for the support and freedom of

research he gave me. Despite being very busy, he always found time to discuss scientific

problems and providing useful advice.

I am grateful to the Marquis of Amodio whose funding through the Lord Florey schol-

arship allowed me to concentrate only on my research work.

Collaborating with Takayuki Uchiashi and Hayato Yamashita from Kanazawa university

was a very exciting experience. Thank you for all your efforts in constantly improving the

hsAFM and your interest in biophysics. I hope we will be able to keep on collaborating

together.

The low noise performance of the current measurement setup could not have been

achieved without the expert soldering of J. Voıtchovsky. Thank you for your time and

your help.

The setup could then successfully be used by Hilary Burch for bio-sensing; thank you

Hilary for making good use of the setup. Thank you also for pointing out some weak-

nesses of the setup and for interesting conversation.

Miya Kamihira from the research group of Anthony Watts supervised the monomeriza-

tion of bR and harvesting of the PM used in chap. 3. Philip Steinmann from John

217

Acknowledgments

Weaver research group provided the nano-electrodes used in chap. 6.

I would finally like to thank Dr. Maurits de Planque and all my laboratory collaborators

for stimulating conversations and useful advice.

C’est bien au dela des considerations scientifiques et des aleas de la recherche que vivent

les reelles motivations et la force qui m’a ete necessaire a l’aboutissement de ce projet.

Rien n’aurait ete possible sans un support constant de mes amis, mais surtout de ma

famille. Je ne pourrais jamais decrire a quel point vous m’etes cher; c’est vous qui donnez

un sens a mon travail.

218

Appendix

8 Appendix

8.1 Energy dissipation with small amplitude AC-AFM

8.1.1 From UHV to liquid: adapting the models

Previous work have modelled the movement of an AFM tip operated in AC-mode, by

a mass attached to a damped harmonic oscillator (Anczykowski et al. 1999). The

cantilever plays the role of a spring driven by a a piezo oscillating with a frequency

ν, so that the the driving force of the damped harmonic oscillator can be modelled by

Fdrive = Ad cos(ωt), with ω = 2πν. Two kinds of damping are considered: viscoelastic

damping of the cantilever in its fluid environment and intrinsic damping of the cantilever.

Solving the equation of motion for the tip provides solution of type:

z(t) = A cos(ωt− ϕ) A > 0 and 0 < ϕ < π (8.1)

where ϕ is the dephasing induced by the damping. In order to establish the energy

dissipated by the tip into the sample, Anczykowski et al. proposed to subtract to the

averaged energy fed to the system the average losses due to damping. The average energy

put in the system is given by:

P in =1

2kcωAAd sin ϕ (8.2)

which is maximum for a dephasing of ϕ = π/2. kc is the cantilever stiffness and A the

tip oscillation amplitude. The intrinsic average energy dissipation is given by:

P out−intr =1

πα1ω

2A

[(A− Ad cos(ϕ)) arcsin

(A− Ad cos(ϕ)√

A2 + A2d − 2AAd cos(ϕ)

)+ Ad sin(ϕ)

](8.3)

219

Appendix

and the viscoelastic average energy dissipation by:

P out−visc =1

2α2ω

2A2 (8.4)

where α1 and α2 are the damping factors for the intrinsic and viscoelastic damping

respectively. Considering the fact that the free oscillating amplitude A0 of the tip (away

from the sample) is given by A0 = QAd, if A ∼ A0 Ad i.e. Q 1 (typ. Q ∼ 1000 in

UHV), eq. (3) simplifies to: P out−intr ' 1/2α1ω2A2.

In liquid, due to the viscoelastic damping of the cantilever, the driving amplitude can be

almost as big as and tip oscillation amplitude (A0 & Ad) and the Q-factors are smaller

(typ. Q ∼ 10). The simplification made for UHV is not straightforward anymore.

However, since viscoelastic damping is much more important than intrinsic damping, we

can write:

α2 α1 (8.5)

Using eq. (5) back in eq. (3), we obtain:

P out−intr = 1πα1ω

2A

[(A− Ad cos(ϕ)) arcsin

(A− Ad cos(ϕ)√

A2 + A2d − 2AAd cos(ϕ)

)︸︷︷︸

≤π/2

+Ad sin(ϕ)

]

≤ 12α1ω

2 AA0︸︷︷︸∼A2

(1− 1

Qcos(ϕ) +

1

Qsin(ϕ)

)︸︷︷︸

<2, since Q&1

. α1ω2A2 1

2α2ω

2A2 = P out−visc

It therefore reasonable to neglect intrinsic damping terms over viscoelastic dissipation,

and posing α2 = α = kc/Qω, and working on-resonance (ω = ω0) we obtain the same

formula than derived by (Anczykowski et al. 1999) for the average energy P ts dissipated

220

Appendix

by the tip into the sample:

P ts = P in − ΣP out =1

2

kcω0

Q

[A0A sin(ϕ)− A2

](8.6)

8.1.2 Estimation of the damping induced by lateral movement (scanning)

with AC-AFM

When scanning over the sample, the cantilever/tip moves horizontally at a speed v. The

fluid speed in respect to the cantilever depends of the distance between the driving part

and the sample (see Fig. 8.1), as described by the laws of hydrodynamic. To calculate the

x

z A

Z(x,t) speedprofileof liquid

V

sample

Figure 8.1: Motion of the tip, scanning in a liquid. The speed profile of the fluid changesalong the tip path during a cycle.

corresponding cantilever energy dissipation at given time t, it is necessary to consider

the contribution of each cantilever point separately and to integrate along the beam.

Let’s consider the function Z(x, t) describing the position (x(t); z(x(t))) of each points

of the cantilever at a given time t (Fig. 8.1). The fluid speed is small enough to assume

a Stokes regime and the average friction force exerted by the fluid on the cantilever is

given by:

|F fric| ∼ω

2π

∫ 2πω

0

∫cant.

ηw

∣∣∣∣∣∂Z(x, t)

∂x

∣∣∣∣∣v(Z(x, t)) d`dt (8.7)

221

Appendix

were η is the fluid viscosity and w the width of a rectangular cantilever. If we consider

a small oscillation amplitude (i.e. typ. A ∼ 10−9 m L with L the cantilever length),

we can make the following approximations:

∣∣∣∣∣∂Z(x, t)

∂x

∣∣∣∣∣ w A

2L| cos(ωt− ϕ)| and v(Z(x, t)) ≡ v ∀x, t (8.8)

Using eq. (7) and (8), we can calculate the average cantilever power dissipation induced

by the scanning:

P scan ∼ωηw

2π

∫ 2πω

0

∫cant.

A

2L| cos(ωt− ϕ)|v2 d`dt =

ηwAv2

π(8.9)

Experimentally, we typically have v ∼ Aω, η ∼ 10−3 and w ∼ 10−5 m. We can then

write:

P scan ∼ 10−17A2ω2 P out−visc (8.10)

The scanning contribution can therefore be neglected at small oscillation amplitude and

Eq. 8.6 still provides a good approximation of the average energy dissipated by the tip

into the sample.

8.1.3 Maximizing energy dissipation at small amplitudes

The interest of working at small oscillating amplitude is the near-linearity of the tip-

sample force over the tip path during the oscillation. It is possible to model the whole

system (tip, sample, liquid) with a damped, driven harmonic oscillator (Fig 8.2). If the

oscillation amplitude is small enough, the tip-sample interaction force can be written as

(Hoffmann 2004):

Fts = Fts0 − ktsz (8.11)

222

Appendix

αc

αts

kc

kts

cantilever drive

sample

tip

sample

substrate

forcefield tip

Figure 8.2: Scheme of the model used to linearize tip-sample interactions. The first springkc represents the cantilever, harmonically driven. The second spring kts stands for the sample,and is therefore attached to a hard substrate underneath. Both springs are coupled with adamping system accounting for the previously mentioned viscoelastic damping, and for sample-related damping respectively. Sample damping can be seen as a sum of non-conservativeprocesses induced by the tip motion in the tip-sample force field (e.g. ions and water moleculedisplacement).

with z the tip-sample separation distance. kts can be understood as the tip-sample

interaction ”stiffness” which characterizes the force field the tip feels when approaching

the sample. From this approximation, the equation of motion reduces to (adapted from

(Hoffmann 2004)):

mz + Cz + (kts + kc)z = kcAd cos(ωt) (8.12)

with m the effective mass of the oscillator, z = dz/dt and C = αc+αts the damping factor

(Fig. 8.2). This equation can be solved assuming a solution identical to Eq. 8.1. When

driven on the cantilever resonance frequency far from the sample (ω = ω0f =√

kc/m),

the solution can be written as

A

A0

cos(ϕ) =Qx

Q2 + (1 + x)2, x =

kc

kts

(8.13)

and

tan(ϕ) =x + 1

Q, x =

kc

kts

(8.14)

223

Appendix

These last equations directly relate the tip-sample interaction stiffness with the macro-

scopic observable of the system. In first approximation, the motion is fully determined

by the ratio x = kc/kts and by the quality factor Q of the cantilever.

If changes in the energy dissipated by the tip into the sample are responsible for the

image formation, optimizing the resolution can be achieved through maximization of

the tip-sample energy dissipation P ts (Eq. 8.6). It is easy to verify that the amplitude

maximizing P ts is unique and given by:

∂P ts

∂A= 0 ⇔ A = A0 sin(ϕ)/2 ∀ϕ ∈ [0, π/2] (8.15)

However A and ϕ are related through the tip equation of motion. Considering a linear tip-

sample force field (assuming a scan at small enough oscillation amplitudes i.e. Hoffmann’s

model) we can use Eq. 8.13 and 8.14 as solution of the motion. This imposes a condition

on x = kc/kts through sin(ϕ) cos(ϕ) = 2Qx

Q2+(1+x)2

tan(ϕ) = x+1Q

(8.16)

This system of equations admits only one solution:

x = 1 , tan(ϕ) = 2/Q (8.17)

Eq. 8.17 gives an indication of the relationship that should be kept between the cantilever

spring constant and the tip-sample interaction (i.e. kc ∼ kts) in order to maximize the

energy dissipation in the sample. Experimentally, the sample stiffness can be taken as a

first approximation for kts.

224

Appendix

8.2 Estimation of the number of bacteriorhodopsin molecule

excited per second given a characterized incident light beam

The extinction coefficient of PM has been determined to be ε = 63000± 300 at 562 nm

(Rehorek & Heyn 1979). If we consider the absorption spectrum of PM (Fig 1.4.B), the

absorption coefficient at 530nm, relevant for the experiments carried out, is∼ 50000. The

effective or efficient absorption section σ of a molecule is defined by the section within

which an incident photon will interact with the molecule (σ ∝ oscillator strength), and

is related to the extinction coefficient through the following relation:

ε = σNa log(e) (8.18)

where Na is Avogadro’s number. Using this formula and the appropriate units, we obtain

for PM:

σ(530 nm) ≈ 1.9 · 10−2 nm2 (8.19)

(or 1.9 · 108 barn).

The experiment run in chapter 5, (§5.3.2) uses a 532 nm CW laser operated at an

average intensity I of ∼ 7 mW . The beam is focused on a spot of ∼ 10 µm in diameter,

illuminating a surface S of ∼ 80 µm2. It is then possible to calculate the number N of

incident photons per surface and per second at λ = 530 nm:

N530 nm =I

S

λ

hc≈ 2.4 · 108 [photon · s−1 · nm−2] (8.20)

where h is Planck’s constant and c the speed of light in vacuum.

Comparing Eq. 8.19 and 8.20, it appears that each bR molecule in a PM monolayer can

potentially interact with an incident photon about 106 times per second.

225

Appendix

8.3 Analysis of protein unfolding curves

In order to objectively analyze a batch of single-protein unfolding curves and locate the

position of the different steps, it is necessary to find criteria that can be systematically

applied by a software program to all the curves of the analyzed batch. Such a software

was developed in IgorTM , in parallel with a user friendly interface.

Given a batch of curves, the first problem is to define the origin (both vertically and

horizontally) of the unfolding curves. By definition, the origin means the point where the

AFM tip touches the membrane, but without exerting any pressure on it (i.e. without

any deflection of the cantilever). The origin determination for one curve was achieved

by minimizing a function A(x) defined as in Fig. 8.3. The origin was defined as the

tip-sample distance

cantileverdeflection

a

bA(x)

x

d1

d2

Figure 8.3: Scheme of the procedure followed to find the origin of an unfolding curve. Thepoints a and b are fixed while x runs along the curve. For each x, two first order regression d1

and d2 of the unfolding curve are calculated for the intervals [a, x] and [x, b]. The area A(x)covered by the difference between the regression curves and the actual unfolding curve is thenminimized using a dichotomic technique.

intersection of d1 with d2 at the point xm for which A(xm) is minimum. Among different

trials, this technique provided the best results, especially for curves presenting an im-

226

Appendix

portant cantilever deflection. Fine alignment of each batch was consequently achieved

by taking the curve showing the largest deflection and aligning all the other curves to

it through area difference minimization. Advanced options allowed fine alinement using

only specified regions. Such options were used to align unfolding curves on a particular

step.

Once all the curves are aligned and admit a relevant origin, the position of the steps is

determined for each curve separately by running two articulated rods along the curve

(Fig. 8.4.A) while recording the slope difference between the rods. This process can

be seen as a differentiation of the curve, but with a smoothing factor dictated by the

rod length as illustrated in Fig. 8.4.B. The effective step where then determined by the

regions where the rods slope difference is larger than σ, with σ an adjustable parameter.

Different combinations of rod length and σ values lead to differences in the steps position

determination. Intuitively, longer rod length combined with larger σ should allow better

localization of the actual steps and discrimination of the signal noise. This intuition was

confirmed by many trials, provided that the rods length is small enough to resolve the

steps, as visible in the histograms of Fig. 8.4.C. General steps statistics such as presented

in Fig. 4.6 were obtained by adding up each curve step-histogram for a batch and by

further division of the result by the total number of curves in the batch considered.

8.4 Monomerization of Purple Membrane

The monomerization of bR has been Supervised by Dr. Miya Kamihira from the research

group of Prof. Anthony Watts, Department of Biochemistry, University of Oxford.

The procedure followed is detailed in (Schertler et al. 1993).

bR in solubilized buffer ∼ 2ml (1% Triton, 10mM NaPi (NaH2PO4), pH = 5.6)

227

Appendix

rod length =10

rod length = 13

rod length = 16

rod = 10, σ = 0.9

rod = 13, σ = 0.9

rod = 16, σ = 1.2

0 20 40 60 80 100

0 20 40 60 80 100

0 20 40 60 80 100

Unfolding event

α0

-1

2

1

[nm]

[nm]

[nm]

[nm]

freq.

1

1

slopediff.

defl. A

B

C

Figure 8.4: Determination of the unfolding steps by a programmed routine. Two articulatedrods of a given length run along the investigated curve (A) while their slope difference αis recorded for each point of the curve analyzed (B). The locations of the steps are thendetermined (C) as the positions where (B) is larger than a adjustable parameter σ. (B) and(C) illustrate different possible combinations of values for the rods length and for σ. Theresults are offset for better visibility.

228

Appendix

↓ overnight, 34C on shaker with glass beads

↓ 35k rpm, 20min, 15C

DEAE-Sepharese column ← wash with 500mM NaPi (pH = 5.6, ∼ 5ml − 10ml)

+ equilibrium with solubilized buffer (∼ 5ml − 10ml)

↓

Load bR (1% DDM, 10mM NaPi pH = 5.6, ∼ 30ml of exchange buffer)

+

Elution with high salt buffer (1% DDM, 500mM NaPi, pH = 5.6)

↓

Dialysis overnight (room T C) against low salt buffer (∼ 400ml)

12kDa− 14kDa dialysis membrane

Event. concentration of monomeric bR solution by centrifugation (better storage)

Different buffer solutions

• Solubilized buffer

1% Triton, 10mM NaH2PO4, pH = 5.6

Make 5ml of

100mM NaH2PO4 (1ml NaH2PO4 0.5M)

10% Triton (weight/vol.) (0.5g)→ 5.0ml

⇒ diluted x10 with miliQ water for use

• Column washing buffer

500mM NaH2PO4 (pH = 5.6)

(Make 500ml because all other buffers made from this one)

229

Appendix

• Exchange buffer - to load bR

10% DDM, 100mM NaH2PO4 (pH = 5.6) Make 5ml

(also diluted x10 with miliQ water for use)

• Elution buffer

1% DDM, 500mM NaH2PO4 (pH = 5.6)

(0.5g DDM → 50ml)

• Low salt buffer - for dialysis

10mM NaH2PO4 (pH = 5.6)

8.5 References

Anczykowski, B., Gotsmann, B., Fuchs, H., Cleveland, J. P., & Elings,V. B. 1999. How to measure energy dissipation in dynamic mode atomic force mi-croscopy. Appl. Surf. Sci., 140(3-4), 376.

Hoffmann, Peter M. 2004. Small-Amplitude Atomic Force Microscopy. Pages 3641– 3654 of: Dekker Encyclopedia of Nanoscience and Nanotechnology.

Rehorek, Mike, & Heyn, Maarten P. 1979. Binding of all-trans-Retinal to thePurple Membrane. Evidence for Cooperativity and Determination of the ExtinctionCoefficient. Biochemistry, 18(22), 4977–4983.

Schertler, Gebhard F. X., Bartunik, Hans D., Michel, Hartmut, &Oesterhelt, Dieter. 1993. Orthorhombic Crystal Form of Bacteriorhodopsin Nu-cleated on Benzamidine Diffracting to 3.6 AResolution. J. Mol. Biol., 234(1), 156.

230

Publication and Conferences

9 Publications and Conference Meetings

9.1 List of publications

1. Real-time imaging of single bacteriorhodopsin large conformational changes

by high-speed atomic force microscopy

Kislon Voıtchovsky, Hayato Yamashita, Takayuki Uchihashi, Sonia Antoranz Con-

tera, John F. Ryan and Toshio Ando, 2006, in preparation.

2. Sensing of metalloproteins with a CNx nanotube device

Hilary J. Burch, Sonia Antoranz Contera, Kislon Voıtchovsky, Nicole Grobert and

John F. Ryan, 2006, in preparation.

3. Local electrostatic and steric forces underlying bacteriorhodopsin struc-

ture and mechanics

Kislon Voıtchovsky, Sonia Antoranz Contera and John F. Ryan, 2006 submitted.

4. Hydration force microscopy reveals the molecular forces that trap and

release protons in a bacterial proton-pump

Sonia Antoranz Contera*, Kislon Voıtchovsky* and John F. Ryan, 2006, submitted.

5. Differential Stiffness and Lipid Mobility in the Leaflets of Purple Mem-

branes

Kislon Voıtchovsky, Sonia Antoranz Contera, Miya Kamihira, Anthony Watts and

John F. Ryan, 2006, Biophys. J., 90 (6) 2075-2085.

6. Ultrafast Excited State Dynamics of the Protonated Schiff Base of All-

trans Retinal in Solvents

0 *equal contribution of these authors

231


Goran Zgrablic, Kislon Voıtchovsky, Maik Kindermann, Stefan Haacke and Majed

Chergui, 2005, Biophys. J., 88 (4) 2779-2788.

7. Ultrafast photophysics of the protonated Schiff base of retinal in alcohols

studied by femtosecond fluorescence up-conversion

Goran Zgrablic, Kislon Voıtchovsky, Maik Kindermann, Majed Chergui and Stefan

Haacke, 2004, In Femtochemistry and Femtobiology: Ultrafast Events in Molecular

Science. M. Martin and J. T. Hynes, editors. Elsevier, Paris.

9.2 Conferences and Meetings

1. High-resolution dynamic imaging of membrane proteins by high-speed

AFM

Hayato Yamashita, Takayuki Uchihashi, Kislon Voıtchovsky, Sonia Antoranz Con-

tera, Daisuke Yamamoto, John F. Ryan and Toshio Ando

Japanese Biophysical Society Meeting, November 2006, Okinawa, Japan (Poster)

2. Biosensing with CNx multi-wall carbon nanotubes

Hilary J. Burch, Sonia Antoranz Contera, Nashville C. Toledo, Maurits R.R. de

Planque, Kislon Voıtchovsky, Nicole Grobert and John F. Ryan

International Conference on Solid State Devices and Materials (SSDM06), Septem-

ber 2006, Yokohama, Japan (Poster)

3. High-Resolution Dynamic Imaging of Membrane Proteins with High-

Speed AFM

Takayuki Uchihashi, Kislon Voıtchovsky, Hayato Yamashita, Sonia Antoranz Con-

tera, John F. Ryan and Toshio Ando

International Conference on Nanoscience and Technology (ICN+T06), August 2006,

Basel, Switzerland (Talk)

232


4. Sub-molecular level images and single molecule force spectroscopy show

local effects of salt concentration on protein-lipid interactions in halophilic

biomembranes

Kislon Voıtchovsky, Sonia Antoranz Contera and John F. Ryan

Biophysical Society Meeting, February 2006, Salt Lake City, USA (Poster)

5. Small amplitude sub-molecular resolution AC-AFM and High-Speed

AFM of purple membranes in fluid: non-continuum effects inside the

double layer

Sonia Antoranz Contera, Kislon Voıtchovsky, Hayato Yamashita, Takayuki Uchi-

ashi, Noriyuki Kodera, Toshio Ando and John F. Ryan

Biophysical Society Meeting, February 2006, Salt Lake City, USA (Talk)

6. High-resolution dynamic imaging of membrane proteins by high-speed

AFM

Hayato Yamashita, Kislon Voıtchovsky, Takayuki Uchiashi, Sonia Antoranz Con-

tera, Noriyuki Kodera, John F. Ryan and Toshio Ando

Biophysical Society Meeting, February 2006, Salt Lake City, USA (Poster)

7. Nitrogen-doped nanotubes as biosensors

Hilary J. Hamnett, Sonia Antoranz Contera, Nicole Grobert, Kislon Voıtchovsky

and John F. Ryan

Sixth International Conference on the Science and Application of Nanotubes (NT05),

June 2005, Goteborg, Sweden (Poster)

8. Bacteriorhodopsin for bionanotechnology

Kislon Voıtchovsky

Heythrop Park BioNanotechnology Workshop, June 2005, Chipping Norton, UK

(Talk)

233


9. Evidence of asymmetry in cohesion and assembly of both leaflets of

purple membranes

Kislon Voıtchovsky, Sonia Antoranz Contera, Miya Kamihira, Anthony Watts and

John F. Ryan

Biophysical Society Meeting, February 2005, Long Beach, USA (Poster)

10. Towards Single molecule biophotonics

Kislon Voıtchovsky, Miya Kamihira, Sonia Antoranz Contera, Philipp Steinmann,

Anthony Watts, John Weaver and John F. Ryan

Rank Prize Fund Nanotechnology conference, April 2004, Grasmere, UK (Invited

talk)

234

Atomic Force Microscopy and Force Spectroscopy of...

Documents

Transcript of Atomic Force Microscopy and Force Spectroscopy of...