NOVEL TARGETS WITHIN THE HEPATITIS C VIRUS NONSTRUCTURAL PROTEIN NS4B AND THEIR INHIBITION

USING DISTINCT CLASSES OF SMALL MOLECULES

A DISSERTATION SUBMITTED TO THE DEPARTMENT OF MICROBIOLOGY AND IMMUNOLOGY

AND THE COMMITTEE ON GRADUATE STUDIES OF STANFORD UNIVERSITY

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

Paul David Bryson December 2009

http://creativecommons.org/licenses/by-nc/3.0/us/

This dissertation is online at: http://purl.stanford.edu/bs187wm8852

© 2010 by Paul David Bryson. All Rights Reserved.

Re-distributed by Stanford University under license with the author.

This work is licensed under a Creative Commons Attribution-Noncommercial 3.0 United States License.

ii

I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.

Jeffrey Glenn, Primary Adviser

I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.

Harry Greenberg

I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.

Shoshana Levy

I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.

Peter Sarnow

Approved for the Stanford University Committee on Graduate Studies.

Patricia J. Gumport, Vice Provost Graduate Education

This signature page was generated electronically upon submission of this dissertation in electronic format. An original signed hard copy of the signature page is on file inUniversity Archives.

iii

iv

Abstract

Hepatitis C Virus (HCV) is the causative agent of significant liver disease,

including cirrhosis and hepatocellular carcinoma. This virus infects greater than 2% of

the world’s population, and treatment options for these patients are limited; thus, this

virus represents a major public health problem. In this thesis, we aim to enhance our

ability to solve this problem by seeking to unravel the role that NS4B, one of the HCV

nonstructural proteins, plays in the viral life cycle. In focusing on this viral protein, we

not only achieve a better understanding of HCV’s biology, but we also identify

multiple small molecules that may directly lead to better treatment options for those

afflicted with this virus.

Drawing on similarities to other plus-sense RNA viruses, we identify an RNA

binding activity in NS4B, demonstrate that this activity is specific for the 3’ terminus

of the HCV genome template, and characterize the domains of NS4B that are

responsible for this activity. We utilize the microfluidic technology developed for this

assay to perform a high-throughput screen for small molecule inhibitors of this RNA

binding activity, and identify clemizole hydrochloride, among others, as an effective

inhibitor. Cell-based studies show that inhibition of RNA binding through either

pharmacologic or genetic methods inhibits HCV replication. Genetic analysis further

identifies two mutations in the HCV NS3 protein that each can overcome the genetic

disruption of NS4B’s RNA binding domain.

In addition to clemizole, our work identifies a small molecule inhibitor of HCV

that affects a different activity of NS4B, its ability to form membrane-associated foci.

v

We provide genetic, biochemical, and cell biological evidence that NS4B is the target

of this drug, and an in vitro light scattering assay further suggests that it is the second

amphipathic helix of NS4B that is the target. In total, our results demonstrate that two

different activities of NS4B are each a valid pharmacological target for HCV antivirals

and uncover two candidate compounds that have potential for further pharmaceutical

development.

vi

Acknowledgments

I owe deep gratitude to many people who have helped me through the winding

road to this dissertation. First and foremost, I want to thank my PI, Jeffrey Glenn, who

has helped steer me through this path. I have to confess that occasionally—especially

when I got stuck on a project—I would feel somewhat less than “happy.” But

fortunately, you always had a 1-day experiment to turn things around and keep me

moving forward.

I am also grateful for the excellent colleagues I met in Jeff’s group. Thank you

to Menashe and Shirit for getting me started, to Namjoon, Choongho, and Ella for

keeping me going, to Wei for helping me out at all hours of the day, to Rick for

keeping me well hydrated, and to all for making the experience not just long, but

enjoyable. I will miss you guys.

I am very fortunate to have had the opportunity to obtain the level of education

I received at Stanford. My interactions with faculty really solidified my ambitions to

become a faculty member one day. I would like to thank my committee, Harry, Peter

and Shoshana for adding a dose of reality to my zany ideas. Joe Lipsick and James

Nelson were excellent teaching mentors, and I hope to follow in their lead some day.

Upi Singh and John Boothroyd both left me with the desire to explore new fields.

Thanks to all the faculty who put together the teaching curriculum for us M&I

students. And finally, I really appreciate all the staff, especially Julie, Mary Jeanne,

and Wanapa, who allowed me to focus on the science and not the business of being a

graduate student.

vii

On a somewhat more personal note, I am grateful to have entered with a class

of students who ended up being good friends and colleagues. You all have made this

period so much more enjoyable than it could have been. Special thanks to Jeff, Drew

and Drew for providing me with a roof over my head and to others for offering to help

me finish my work here.

Most importantly, I couldn’t have made it this far without the love and support

of my family. Xiaochin, sometimes I think you like science more than I do, and it

thrills me every time I get to explain my latest experiment to you. Mom, thanks for

always believing in me and pushing me to succeed at the next level. Hallie, I really

appreciate the grounding effect you and your family has had for me. And Copland,

thanks for listening to my practice talks—here’s a treat, or “bark bark, bark bark bark

treat!”

viii

Table of Contents ABSTRACT .......................................................................................................................................... IV

ACKNOWLEDGMENTS.................................................................................................................... VI

TABLE OF CONTENTS.................................................................................................................. VIII

LIST OF FIGURES................................................................................................................................X

LIST OF TABLES..................................................................................................................................X

CHAPTER 1. INTRODUCTION...........................................................................................................1

HCV EPIDEMIOLOGY.............................................................................................................................2 HCV THERAPY ......................................................................................................................................2 HCV MOLECULAR BIOLOGY ..................................................................................................................3 HCV NS4B ...........................................................................................................................................5 THEMES UNCOVERED IN THIS RESEARCH ...............................................................................................7

CHAPTER 2. DISCOVERY OF A HEPATITIS C TARGET AND ITS PHARMACOLOGICAL INHIBITORS BY MICROFLUIDIC AFFINITY ANALYSIS ...........................................................9

2.1 INTRODUCTION ..............................................................................................................................10 2.2 RESULTS ........................................................................................................................................12

Microfluidic assay validation: HuD binding to RNA.....................................................................13 NS4B binds HCV RNA and Kd is determined by microfluidics ......................................................16 NS4B specifically binds the 3’ terminus of the (–) viral strand .....................................................17 An ARM in NS4B is essential for RNA binding and HCV replication............................................19 High-throughput screening for inhibitory compounds...................................................................21 Inhibitors of HCV RNA replication................................................................................................25 Clemizole-resistant mutants...........................................................................................................25

2.3 DISCUSSION ...................................................................................................................................29 2.4 METHODS ......................................................................................................................................32

Plasmids.........................................................................................................................................32 In vitro RNA transcription and fluorescent and radioactive labeling............................................33 Device design.................................................................................................................................34 RNA binding assay.........................................................................................................................35 Screening of inhibitory compound library .....................................................................................39 Determination of IC50 for in vitro RNA binding.............................................................................40 Expression and purification of recombinant NS4B........................................................................40 GST pull down assay......................................................................................................................41 RNA Filter Binding Assay..............................................................................................................42 Cell cultures and electroporation ..................................................................................................42 Viability assay................................................................................................................................43 Luciferase assay.............................................................................................................................44 Real-time PCR ...............................................................................................................................45 Selection of resistant mutants ........................................................................................................45 Whole cell RNA electroporation ....................................................................................................46 Statistical analysis .........................................................................................................................46

CHAPTER 3. MUTATIONS IN NS3 COMPENSATE FOR THE DISRUPTION OF AN RNA-BINDING DOMAIN IN NS4B .............................................................................................................47

3.1 INTRODUCTION ..............................................................................................................................48 NS4B contains an RNA-binding activity ........................................................................................48 NS4B interacts with NS3................................................................................................................48

3.2 RESULTS ........................................................................................................................................49

ix

RBD mutations prevent HCV replication.......................................................................................49 RBD mutations do not affect HCV localization .............................................................................53 Primary-site revertants are not observed ......................................................................................55 Secondary-site mutations in NS3 restore HCV replication in KR247AA replicons .......................57

3.3 DISCUSSION ...................................................................................................................................59 METHODS ............................................................................................................................................62

Plasmids.........................................................................................................................................62 Colony formation assay .................................................................................................................63 Infection-transfection.....................................................................................................................63 Picking colonies, RT-PCR, & sequencing......................................................................................64 Luciferase assay.............................................................................................................................64

CHAPTER 4. A SMALL MOLECULE INHIBITS HCV REPLICATION AND DISRUPTS NS4B’S SUBCELLULAR DISTRIBUTION ......................................................................................65

4.1 INTRODUCTION ..............................................................................................................................66 4.2 RESULTS ........................................................................................................................................67

The small molecule, anguizole, inhibits HCV RNA replication. ....................................................67 Mutations conferring resistance to anguizole map to NS4B..........................................................70 Anguizole treatment leads to an altered subcellular distribution pattern of the NS4B protein. ....74 A resistance mutation also alters the subcellular distribution pattern of the NS4B protein. .........77 Anguizole interacts with the second amphipathic helix of NS4B. ..................................................78

4.3 DISCUSSION ...................................................................................................................................82 4.4 MATERIALS AND METHODS...........................................................................................................85

DNA constructs and peptides.........................................................................................................85 Drugs and antibodies.....................................................................................................................86 Cell culture, NS4B-GFP transfections, and fluorescence microscopy...........................................86 Stable luciferase replication assays ...............................................................................................87 Transient luciferase replication assays..........................................................................................88 Analysis of resistance mutants .......................................................................................................89 Dynamic Light Scattering (DLS)....................................................................................................90

CHAPTER 5. CONCLUSIONS ...........................................................................................................92

SUMMARY ...........................................................................................................................................93 SIGNIFICANCE......................................................................................................................................94

REFERENCES ......................................................................................................................................96

APPENDIX A. SUPPLEMENTARY MATERIAL TO CHAPTER 2 ..................................................................97 Calibration curve ...........................................................................................................................97 RNA binding experiment with high ionic strength buffer...............................................................97 ATA inhibits RNA binding by NS4B in a dose-dependent manner.................................................98 Specificity of hits identified in the small molecule screen..............................................................98

APPENDIX B. SUPPLEMENTARY MATERIAL TO CHAPTER 4 ................................................................109 BIBLIOGRAPHY..................................................................................................................................111

x

List of Figures FIGURE 1.1. THE HCV GENOME STRUCTURE. ............................................................................................4 FIGURE 1.2 REPLICATION COMPLEX OF HCV .............................................................................................6 FIGURE 2.1 PROTEIN-RNA INTERACTIONS MEASURED ON MICROFLUIDIC PLATFORM..............................15 FIGURE 2.2 NS4B BINDS SPECIFICALLY TO THE 3’ TERMINUS OF THE HCV NEGATIVE STRAND RNA......18 FIGURE 2.3 IDENTIFICATION OF RNA BINDING DOMAINS WITHIN NS4B. .................................................20 FIGURE 2.4 SMALL-MOLECULE SCREEN REVEALS THAT CLEMIZOLE HYDROCHLORIDE INHIBITS RNA

BINDING BY NS4B AND HCV RNA REPLICATION IN CELL CULTURE. .............................................24 FIGURE 2.5 CLEMIZOLE-RESISTANT MUTANT. ..........................................................................................28 FIGURE 3.1. PUTATIVE RNA BINDING DOMAINS ARE NECESSARY FOR HCV REPLICATION. .....................52 FIGURE 3.2. DISRUPTION OF RNA BINDING DOMAINS DOES NOT AFFECT THE FORMATION OF PUTATIVE

HCV REPLICATION COMPLEXES......................................................................................................54 FIGURE 3.3. REPLICATING COLONIES CONTAINING KR247AA MUTATION CONTAIN SECONDARY-SITE

MUTATIONS.....................................................................................................................................56 FIGURE 3.4. MUTATIONS IN NS3 ENHANCE REPLICATION OF RBD DEFECTIVE MUTANTS. .......................58 FIGURE 3.5. LOCATION OF COMPENSATORY MUTATIONS IN NS3-4A DIMER. ...........................................61 FIGURE 4.1. A SMALL MOLECULE INHIBITS HCV REPLICATION................................................................69 FIGURE 4.2. RESISTANT MUTATIONS MAP TO NS4B. ................................................................................71 FIGURE 4.3. CHARACTERIZATION OF THE H94R RESISTANCE MUTATION. ................................................73 FIGURE 4.4. ANGUIZOLE ALTERS THE SUBCELLULAR DISTRIBUTION OF NS4B-GFP IN TRANSIENTLY

TRANSFECTED CELLS. .....................................................................................................................76 FIGURE 4.5. THE H94R MUTATION ALTERS NS4B-GFP’S SUBCELLULAR DISTRIBUTION. ........................79 FIGURE 4.6. ANGUIZOLE INTERACTS WITH THE SECOND AMPHIPATHIC HELIX (AH2) OF NS4B. ..............81 FIGURE A.1. MICROFLUIDIC BASED RNA BINDING ASSAY. ....................................................................101 FIGURE A.2. SIGNAL TO NOISE RATIO IN OUR RNA BINDING ASSAY.......................................................103 FIGURE A.3. MICROFLUIDICS-BASED ANALYSIS OF RNA BINDING BY ANOTHER HUMAN PROTEIN FROM

THE ELAV-LIKE FAMILY, HUR (ELAV L1). ................................................................................105 FIGURE A.4. BINDING OF NS4B TO HCV RNA BY CONVENTIONAL METHODS.......................................106 FIGURE A.5. ATA INHIBITS RNA BINDING BY NS4B IN A DOSE DEPENDENT MANNER...........................107 FIGURE A.6. CLEMIZOLE INHIBITS HCV REPLICATION BY REAL-TIME PCR ASSAYS. .............................108 FIGURE B.1. SUBCELLULAR DISTRIBUTION PATTERN OF NS4B-GFP VARIES IN ANGUIZOLE-TREATED

CELLS............................................................................................................................................109 FIGURE B.2. DISTRIBUTION PATTERNS OF CELLULAR MARKERS DO NOT VARY WITH ANGUIZOLE

TREATMENT. .................................................................................................................................110

List of Tables TABLE 3.1. PRIMER SEQUENCES…................................................................................................. 63

1

Chapter 1. Introduction

2

HCV epidemiology

Hepatitis C Virus is a serious cause of liver disease. Patients infected with this

virus can develop chronic hepatitis, liver cirrhosis, and hepatocellular carcinoma.

Throughout the world, more than 170 million people are thought to be infected1 and

HCV is responsible for roughly 27% of liver cirrhoses and 25% of hepatocellular

carcinomas worldwide. Based on global mortality estimates, HCV is thus directly

responsible for 366,000 deaths per year2.

In the United States, prevalence is somewhat lower, though HCV is still the

most common bloodborne infection in the US3 and end-stage liver disease caused by

chronic hepatitis C is the most common cause of liver transplantation4. Approximately

1.6% of the American population has been exposed to HCV, and 1.3% is still positive

for HCV RNA5. This infection is clearly a significant public health problem, and

means to treat its pathogenicity are badly needed.

HCV therapy

Current standard of care regimens for HCV include treatment with Pegylated

interferon- and ribavirin. However, this treatment has significant downsides. First, it

is successful in only 54-56% of patients treated6. Second, these treatments are often

associated with significant side effects, and treatment is not recommended for many

patients with various contraindications; as a result, a sustained viral response to

Pegylated interferon- + ribavirin is achieved in as little as 15% of patients who

present with HCV7. Thus, the current treatment for patients is not effective at curing

3

this disease. Compounding the ineffectiveness of the current antiviral therapy is the

fact that no vaccine is yet available for this pathogen. Therefore, there is a great need

for improved antiviral therapies, and a large amount of effort is being put into the

search for such therapies. In order to develop specific antiviral drugs, an

understanding of the viral life cycle has been crucial (discussed in detail below).

HCV molecular biology

HCV is a member of the Hepacivirus genus, in the Flaviviridae family. Its 9.6

kb genome is composed of a single strand of positive-sense RNA. Encoded within this

genome are three structural proteins, Core, E1 and E2, which together form the viral

particle. Downstream of these proteins are the nonstructural proteins, which include

p7, NS2, NS3, NS4A, NS4B, NS5A, and NS5B (Figure 1.1). A number of model

systems have been utilized to examine the role of each of these proteins, as well as that

of RNA elements of the genome (reviewed in 8). Of relevance to this dissertation are

the 5’ untranslated region (UTR) of the RNA genome and the proteins NS3, NS4B,

and NS5B.

The 5’ UTR forms a highly ordered structure composed of four domains, I-IV.

These form an internal ribosome entry site (IRES) that allows for the cap-independent

translation of the viral genome. Notably, elements of the 5’UTR that are not necessary

for translation are essential for RNA replication9. As will be discussed in Chapter 2,

this region appears to encode the RNA sequence necessary for efficient binding

between NS4B and HCV RNA.

4

Figure 1.1. The HCV Genome structure.

HCV contain an RNA genome of 9.6kb. The 5’UTR contains four ordered domains,

which form an internal ribosome entry site (IRES). The HCV genome encodes 10

proteins, divided into structural (Core, E1, E2) and nonstructural (p7, NS2, NS3,

NS4A, NS4B, NS5B). Their known functions are shown here (reprinted by permission

from Macmillan Publishers Ltd: Nature Reviews Microbiology (5, 453-463),

copyright 2007).

5

NS3 is a multifunctional protein that contains both a protease domain and a

helicase domain. Its protease activity is a common target of antiviral drugs, some of

which have progressed into Phase IIb clinical trials10, 11. Furthermore, this protein has

been shown to interact genetically with NS4B, which will be discussed in Chapter 3.

NS5B is the RNA-dependent RNA polymerase that is responsible for copying

the HCV genome. As described in Chapter 4, this protein is another major target of

antiviral therapies, and various nucleoside and non-nucleoside inhibitors are in

development12.

HCV NS4B

The main subject of this dissertation is the NS4B viral protein. This protein has

been the subject of less investigation than the other nonstructural proteins.

Nevertheless, several groups have revealed structural and functional insights into this

protein’s role, outlined below.

NS4B was first characterized to be sufficient for the formation of the

membranous web, a vesicular structure observed in HCV-infected cells13 (Figure

1.2A). All RNA viruses appear to replicate in association with membranes, and this

membranous web is hypothesized to be the site of replication for HCV. As discussed

in Chapter 4, many of the HCV NS proteins, as well as nascent HCV RNA, have been

localized to this web14, 15. In addition, the equivalent of the membranous web is

thought to be visible with light microscopy as membrane-associated foci that adopt an

endoplasmic reticular pattern16 (Figure 1.2B).

6

Figure 1.2 Replication complex of HCV

(A) Electron micrograph of the “membranous web,” a network of vesicles observed in

the perinuclear region (reprinted by permission from the American Society for

Microbiology: J. Virology (9, 5487-5492), copyright 2007). (B) Fluorescence

microscopy image of NS4B’s “membrane-associated foci,” thought to represent the

sites of viral replication (reprinted by permission from the Society for General

Microbiology: J. General Virology (86, 1415-1421), copyright 2005).

7

Structually, NS4B is associated with membranes and is thought to contain four

or five transmembrane domains17. Also, it contains several amphipathic helices, which

play a role in its membrane localization18, 19. These functions will be examined in

Chapter 4.

In addition, NS4B has been described to contain a nucleotide binding motif

that is essential for its GTPase activity20. Also, NS4B has been described to

oligomerize, for which lipid modifications on the C-terminus are important.21

Themes uncovered in this research

This thesis enumerates the cumulative results of several different projects

performed during the author’s dissertation. All three of the chapters are centered on

NS4B’s role in the replication of viral RNA, but they are different in the approach

taken to that role. One project focuses on the cellular biology of how NS4B may be

involved in the formation of the complexes on which replication occurs (Chapter 4).

The other project takes both a biochemical and genetic approach in order to examine

one functional aspect of the NS4B protein, its RNA-binding activity (Chapters 2 and

3). By examining disparate aspects of NS4B’s role in the HCV life cycle, this thesis

aims to elucidate a broader understanding of this protein’s evolution as an essential

part of the HCV genome.

Much of this work was performed in collaboration with others in Jeffrey

Glenn’s lab and beyond. Specifically, Chapter 2 was a collaboration with Shirit Einav

of the Glenn lab and Doron Gerber of the Stephen Quake lab, and this chapter has

been published in Nature Biotechnology22. Paul Bryson, the author of this thesis, was

8

responsible for the developing the RNA-binding hypothesis, identifying potential

RNA-binding domains, designing the RNA-binding experiments, executing some of

the binding experiments, and contributing a portion of the final manuscript. Chapter 3

has not yet been published as a journal article, and all the material presented therein is

primarily the responsibility of this thesis’s author. Chapter 4 was a collaboration with

members of the Genelabs Technologies company and has been submitted for

publication. Paul Bryson was responsible for the design of the project, the execution of

some of the replication assays and resistant mutant identification, and of all the

fluorescence microscopy. Furthermore, he was responsible for the data analysis and

the composition of the manuscript.

9

Chapter 2. Discovery of a hepatitis C target and its pharmacological inhibitors by microfluidic affinity

analysis

10

2.1 Introduction

Over 150 million people are infected with Hepatitis C Virus (HCV)

worldwide. Unfortunately, many of these individuals are unable to clear their infection

with the current standard of care, which consists of a combination of interferon and

ribavirin23. Moreover, this treatment is associated with significant side effects,

precluding its use by many individuals. Thus, current therapies are inadequate for the

majority of the patients23, and there is a pressing need for new HCV drugs23.

The 9.6-kb, positive, single-stranded RNA HCV genome encodes a 3,000-

amino-acid polyprotein which is proteolytically processed into structural proteins,

which are components of the mature virus, and nonstructural proteins (NS), which are

involved in replicating the viral genome24. Like other positive strand RNA viruses25,

HCV appears to replicate in association with intracellular membrane structures. In the

case of HCV, the structures are termed the membranous web26 and are believed to be

induced by the NS4B protein. NS4B is also required to assemble the other viral NS

proteins within the apparent sites of RNA replication18. It is not known how viral

RNA, especially the negative strand template required for production of progeny

genomes, might be incorporated or maintained at these replication sites.

NS4B and HCV RNA have been shown to colocalize to the membranous

web27, 28, suggesting that NS4B is in intimate contact with viral RNA in the context of

authentic viral RNA replication. The hepatitis A and polio picornaviruses have

proteins termed 2C which are required for replication, bind RNA29, 30, and have an N-

terminal amphipathic helix and a nucleotide binding motif29-31. NS4B contains the

11

same structural features, and both of them are required for HCV replication18, 20. We

thus hypothesized that NS4B may similarly bind RNA, that this interaction might be

critical for the HCV life cycle, and that RNA binding by NS4B could be amenable to

pharmacologic disruption. To test these hypotheses, we sought to establish an in vitro

RNA binding assay in a format enabling simultaneous analysis of multiple conditions,

mutants, and replicates, quantitative dissociation constant (Kd) measurements, and

high-throughput screening for potential pharmacologic inhibitors.

Like the majority of drug targets32, NS4B is a membrane protein. Membrane

proteins are notoriously difficult to express and characterize biochemically, especially

in the quantities required for pharmaceutical screening. Moreover, solubilization by

detergents can alter their natural membrane associated topology. To ameliorate these

problems, we used microfluidic tools to perform binding assays with nanoliter protein

consumption. This enabled use of an in vitro cell lysate expression system, which was

supplemented with microsomal membranes to create more natural folding conditions,

under which the best NS4B topology data available to date has been obtained33. The

reduced yield relative to conventional expression methods is offset by low sample

consumption.

Previous microfluidic tools to measure drug interactions have been limited to

enzymatic targets which can catalyze formation of a fluorescent substrate34. In this

case we directly measured binding constants by using mechanical trapping of

molecular interactions (MITOMI), a microfluidic affinity assay that has previously

been used to measure interactions between transcription factors and DNA35. We have

extended the previous work by showing that MITOMI can be used both to measure

12

binding constants of membrane protein-RNA interactions and to measure inhibition of

such interactions by small molecules in a high throughput screen. The latter point was

particularly surprising in that the elastomer used to fabricate the device is known to

have limitations in chemical compatibility36, 37; here we show that this does not

prevent its use in a drug screen nor does it prevent discovery of a small molecule with

the desired pharmacological properties. Taken together, the results of this paper reveal

a novel HCV target and are the first demonstration that microfluidic technology can be

used to discover a new pharmaceutical, thereby successfully consummating more than

a decade of effort to apply microfluidic tools to drug discovery38, 39.

2.2 Results

We validated the use of this platform for RNA binding by studying two human

proteins from the embryonic lethal abnormal visual system (ELAV) family, the RNA

binding activity of which has been previously well-characterized40-42. We then applied

this methodology to study RNA interactions with the transmembrane HCV NS4B

protein. We used this platform to (i) test the hypothesis that HCV NS4B binds RNA,

(ii) determine the Kd for this interaction, (iii) study the substrate specificity of this

binding, (iv) determine the amino acids within NS4B required for RNA binding, and

(v) screen a compound library for pharmacologic inhibitors of NS4B RNA binding

and HCV replication.

13

Microfluidic assay validation: HuD binding to RNA

HuD is a host cell protein from the ELAV-like family with well characterized

RNA binding activity. It is a cytoplasmic protein that contains 3 conserved RNA

recognition motifs through which it binds AU-rich elements (ARE) in the 3’ UTR (un-

translated region) of genes such as those encoding cytokines and proto-oncogenes.

Binding was tested against two Cy3-labeled RNA probes: AU3 and a known AU3

mutant to which HuD binding is impaired41, 42 (Figure 2.1). RNA binding experiments

with MITOMI were performed essentially as described by Maerkl and Quake35, except

that in this case RNA was used instead of DNA (Figure A.1). Briefly, we spotted a

microarray of target RNA sequences labeled with Cy3 onto an epoxy-coated slide.

These arrays were used to program the microfluidic devices by aligning each spot in

the array to a unit cell in the device. After bonding the microfluidic device to a

microarray it was subjected to surface patterning that resulted in a circular area coated

with biotinylated anti-histidine antibodies within each unit cell. The device was then

loaded with in vitro transcription/translation mixture containing DNA templates

coding for HuD fused in frame with a C-terminal V5-6 histidine tag (HuD-V5-his) or

Gus protein fused in frame with a C–terminal 6 histidine tag (Gus-his). Bodipy-labeled

tRNALys was added for protein labeling. Each unit cell was then isolated by the control

of three micromechanical valves followed by an incubation to allow protein synthesis,

binding of the synthesized protein to the surface biotinylated anti-his antibodies,

solvation of target RNA, and equilibration of proteins and target RNA. MITOMI was

then performed by actuation of a “button” membrane to trap surface-bound complexes

while expelling any solution phase molecules. After a brief wash to remove untrapped

14

unbound material, the trapped molecules and expressed protein were subsequently

detected with an array scanner. The ratio of bound RNA to expressed protein was

calculated for each data point by measuring the median signal of Cy3 to median signal

of bodipy.

15

Figure 2.1 Protein-RNA interactions measured on microfluidic platform.

(a) Target RNA sequences used to study binding of HuD to RNA and comparison of

Kd values measured using microfluidic affinity analysis to values previously measured

by gel shift assay43, 44. ND, not determined. (b) Binding curve of HuD to increasing

concentration of AU3 () and AU3 mutant () RNA, as determined in the microfluidic

affinity assay. Normalized mean values for 10–20 replicates measured in two

independent experiments are shown for each graph. Error bars represent s.d. (c)

Fluorescent images from microfluidic chip. Left: NS4B-GFP and NS5A(AH)-GFP

were anchored to the microfluidic device surface via its interaction with anti-GFP.

Middle: an RNA probe corresponding to the 3’ terminal region of the negative viral

strand was labeled with Cy5 and incubated with the proteins on the device. Cy5 signal

representing bound RNA is shown following a brief wash. Right: an overlay of the

Cy5 signal and GFP signal representing bound RNA to protein ratio. (d) In vitro

binding curve of NS4B to serial dilutions of the RNA probe. Each data point

represents the mean of 10–20 replicates, and the bars represent the standard error.

16

The assay detected strong binding of HuD to the AU3 RNA probe; background

binding by Gus-his was 7-16 fold lower than the HuD signal (Figure 2.1A). This

background level did not increase with RNA probe concentration and was subtracted

from all chambers (Figure A.2). The binding affinity of HuD to the AU3 probe was

much greater than to the AU3 mutant probe: the Kd for the AU3 binding was

determined to be 23+5 nM and for the AU3 mutant 268+95 nM (Figure 2.1A). These

values agree with previous measurements in a gel shift assay41, 42 and serve to validate

the MITOMI microfluidic affinity assay for RNA-protein interactions. RNA binding

analysis of another protein from the ELAV-like family, HuR, is shown in Figure A.3.

NS4B binds HCV RNA and Kd is determined by microfluidics

We then tested whether HCV NS4B binds RNA. Since NS4B is important in

viral RNA replication, and initiation of positive-strand RNA synthesis is likely to start

at the 3’ terminus of the negative-strand RNA, we first tested binding of NS4B to this

region, using a probe designated 3’ negative terminus. A fusion of the amphipathic

helix (AH) of NS5A to the N-terminus of GFP18 was used as a negative control. This

protein also binds to membranes and can thus anchor the microsomal membranes to

the device surface via the interaction of GFP with anti-GFP (Figure 2.1C). The Kd of

NS4B binding to the 3’ terminus of the HCV negative strand RNA was measured at

3.43+1 nM (Figure 2.1D). To the best of our knowledge this is the first report that

HCV NS4B binds RNA. Conventional GST pull down assays and RNA filter binding

assays using recombinant forms of purified NS4B expressed in E. coli confirmed this

17

finding (Figure A.4), although these were less convenient and amenable to the types of

analyses and high throughput format that we sought to perform.

NS4B specifically binds the 3’ terminus of the (–) viral strand

We measured the substrate specificity of the observed NS4B-HCV RNA

interaction with three additional HCV probes (Figure 2.2a). The probes designated 5’

UTR pos and 3’ UTR pos correspond to the 5’ UTR and 3’ UTR sequences of the

positive viral strand, respectively, and 5’ negative terminus corresponds to the 5’

terminus of the negative strand. The RNA binding experiment was repeated and

binding of NS4B to equimolar concentrations (3 nM) of the various probes was

compared. Whereas NS4B binds all four HCV probes, its apparent affinity to the 3’

negative terminus probe is 5- to 12-fold greater than to the other three HCV RNA

regions tested or to an unrelated delta virus RNA genome sequence45, 46 (see Figure

2.2B). These findings suggest that it is the RNA sequence, and likely the secondary

RNA structure, that determines the specificity of NS4B binding to the 3’ terminus of

the negative strand.

18

Figure 2.2 NS4B binds specifically to the 3’ terminus of the HCV negative strand

RNA.

(a) Four HCV probes were designed. 5’ UTR pos and 3’ UTR pos corresponded to the

5’ UTR and 3’ UTR sequences of the positive viral strand, respectively, and 5’

negative terminus and 3’ negative terminus corresponded to the 5’ and 3’ terminal

regions of the negative strand, respectively. The position of these sequences with

respect to the HCV open reading frame (ORF) is shown. (b) Fractional binding of

NS4B to equimolar concentrations of the four HCV probes and to a non-HCV RNA

(delta virus RNA) probe. The 3’ terminus of the negative genome strand is favored by

> 5X. Each data point represents the mean of 10–20 replicates, and the bars represent

the standard error.

19

An ARM in NS4B is essential for RNA binding and HCV replication

Various structural motifs responsible for the interaction between proteins and

RNA have been reported47. One of these is the arginine rich motif, (ARM). ARMs

were originally defined as short (10 to 20 amino acids) arginine-rich sequences found

in viral, bacteriophage, and ribosomal proteins (Figure 2.3a)30, 47, 48. There is little

identity between ARM sequences, other than the preponderance of arginine residues.

Subsequently, ARM-like motifs, consisting of longer sequences containing fewer

arginines were identified30, 48. Inspection of the primary sequence of NS4B reveals the

presence of multiple positively-charged amino acids (see Figure 2.3B). The majority

of these arginine residues are within the last 71 amino acids, in the C-terminal region

of NS4B (Figure 2.3B, C). This region is predicted to form a cytoplasmic segment

based on empirical topology studies using glycosylation markers33. Elements of this

region conform with previously described ARM-like motifs.

20

Figure 2.3 Identification of RNA binding domains within NS4B.

(a) Arginine-rich sequences known to confer RNA binding found in viral and

bacteriophage proteins49-51. Elements in the terminal loop of NS4B that conform with

arginine rich–like motifs are shown at the bottom. Proteins listed are HIV Rev and

Tat25; l, F21 and P22 bacteriophage antiterminator N proteins49; triple gene block

protein 1 (TGBp1) of the bamboo mosaic virus (BaMV)51 and poliovirus 2C50. (b)

Positively charged amino acids (highlighted) within the primary sequence of NS4B

(genotype 1a). (c) Schematic diagram indicating predicted transmembrane and

intracellular domains of NS4B. Conserved positively charged amino acids in the C-

terminal segment of NS4B are shown in red. (d) Arginine residues mediate RNA

binding by NS4B. Binding of wild-type NS4B-GFP, RRa and RRb NS4B-GFP

mutants and NS5A(AH)-GFP to the 3’ terminus of the negative viral RNA strand (1.5

nM) was determined by microfluidic affinity analysis. Bars represent s.d.

21

Using a series of point mutations, we tested if RNA binding by NS4B is

mediated by some of its positively-charged residues. A substitution of Arg-Arg in

positions 192-193 with Ala-Ala was termed “RRa mutant” and a similar substitution

in positions 247-248 was designated “RRb mutant”. While the RRb mutant decreased

binding by NS4B to the 3’ negative terminus probe by 5 fold, RNA binding activity to

the RRa mutant was completely eliminated at the same probe concentration (1.5nM;

Figure 2.3D). These results suggest that the tested arginine residues mediate RNA

binding by NS4B in vitro. To test the hypothesis that these positively charged residues

are also important for HCV replication, we introduced these mutations into high

efficiency subgenomic HCV replicons and assayed them in standard replicon colony

formation assays (data not shown). The various mutations significantly impaired HCV

RNA replication and the degree of replication inhibition correlated with the degree of

NS4B RNA binding impairment (Bryson P et al, manuscript in preparation). These

results suggest that efficient binding of HCV RNA is required for viral replication in

vitro. Alignment of the sequences of natural HCV isolates currently available in

databases reveals that these positively-charged residues are highly conserved across all

HCV genotypes. This conservation suggests that there is a requirement for the RNA

binding residues for productive viral infection in vivo.

High-throughput screening for inhibitory compounds

We screened a compound library for small molecules that could inhibit the

RNA-NS4B interaction. As shown in Figure 2.4A, 1280 compounds from a small

molecular library were spotted on epoxy-coated slides as a microarray. The array was

22

allowed to dry, and was then aligned and bonded to a microfluidic device. The rest of

the assay was performed as before, except that the device was loaded with NS4B-GFP

followed by Cy5-labeled 3’ terminus negative RNA probe. In the primary screen, the

compounds were spotted at a concentration of ~1mM. The entire library was screened

in duplicate using only two microfluidic devices. Out of 1280 compounds, 104 were

found to have an inhibitory effect (>90% inhibition) on RNA binding by NS4B. In

addition, there were 110 compounds for which there was a significant discrepancy

between the two tested replicates or to which one or two of the measurements were

disrupted due to technical reasons.

The 214 compounds (104+110) identified in the primary screen were subjected

to a secondary screen (Figure 2.4A). This was done in a similar manner except that a

smaller device was used, the spotted compound concentration was 10 fold lower than

in the primary screen and 5 replicates were spotted for each compound. 18 compounds

were confirmed to significantly inhibit RNA binding by NS4B out of the 214

compounds tested (Figure 2.4B). Most of the identified compounds did not inhibit

binding of HuR protein to its own 4A target RNA sequence (Figure A.3) nor did they

inhibit HuR binding to its previously described HCV RNA target: the 3’ terminus of

the negative viral strand52, suggesting that these hits are specific. This data and

additional data supporting the validity of the screen are presented in Appendix A and

Figure A.5.

23

24

Figure 2.4 Small-molecule screen reveals that clemizole hydrochloride inhibits

RNA binding by NS4B and HCV RNA replication in cell culture.

(a) The first screen represented a low-stringency measurement of inhibition of 1,280

compounds where the latter were categorized as having high (green), ambiguous

(blue) or no (red) inhibition. Based on the initial screen, 214 compounds were then

measured again with higher stringency and with a greater number of replicates and the

best 18 inhibitors were tested for their ability to inhibit HCV replication via a cellular

assay. (b) In vitro inhibition of NS4B-RNA binding by the top 18 small molecules. (c)

HCV luciferase reporter–linked cellular assay showing that clemizole inhibits HCV

replication (left axis, blue ) with no measurable toxicity to the cell as measured by

alamarBlue (right axis, red ). (d) In vitro NS4B-RNA binding: inhibition curve of

clemizole. Each data point represents the mean of 10–20 replicates and the bars

represent the standard error.

25

Inhibitors of HCV RNA replication

We measured the in vivo antiviral effect on HCV RNA replication of the

inhibitory compounds identified in the above screen. Following electroporation with a

full-length HCV RNA genome harboring a Luciferase reporter gene53, Huh7.5 cells

were grown in the presence of increasing concentrations of these compounds.

Luciferase assays were performed at 72hr. In parallel, the viability of cells in the

presence of the compounds was assessed by an Alamar Blue-based assay. 6 of the

compounds showed some antiviral effect above that solely attributable to cellular

toxicity. One of these, clemizole hydrochloride (1-p-Chlorobenzyl-2-(1-pyrrolidinyl)

methylbenzimidazole hydrochloride), an H1 histamine receptor antagonist, was found

to significantly inhibit HCV replication. A tenfold decrease in viral replication was

measured at 20µM concentration of the drug, with an EC50 of ~8 µM (Figure 2.4C). At

these concentrations there was no measurable cellular toxicity (Figure 2.4C). Similar

results were obtained by real-time PCR assays performed in clemizole-treated Huh7.5

cells infected with the infectious HCV clone (J6/JFH) 54 (Figure A.6). The in vitro

half-maximal inhibitory concentration (IC50) of clemizole for RNA binding by NS4B

is ~ 24 1 nM (Figure 2.4D).

Clemizole-resistant mutants

The mechanism of action of clemizole’s antiviral activity was further

substantiated by selecting for clemizole resistant HCV mutants. Established HCV

replicon-harboring cells and Huh7 cells electroporated de novo with a genotype 1b

subgenomic HCV replicon (Bart 79I)55 were passaged in the presence of the drug,

26

yielding ~60 colonies that were able to grow in the presence of the compound. 11

individual colonies were successfully expanded, passaged 5-10 times and the HCV

RNA replicating in the cells was subjected to sequence analysis. In addition, RNA

from a pool of clemizole-resistant colonies was isolated and subjected to a similar

analysis. 3 colonies were found to harbor replicons with mutations that mapped to the

NS4B region and 6 colonies were found to harbor replicons with mutations that

mapped to the negative strand 3’ terminus RNA region. In addition there was one

colony with mutations that mapped to both NS4B and the negative strand 3’ terminus,

and one where we have yet to find the location of the mutation conferring resistance to

clemizole. No such mutations were identified in 10 replicon colonies that were

passaged in parallel in the absence of the drug.

Two of the clemizole-resistant NS4B mutants were characterized in detail. The

first, W55R is depicted in Figure 2.5A. It involves the substitution of an arginine for

the tryptophan at amino acid 55 within a predicted cytoplasmically-oriented segment

of NS4B. This mutation is sufficient to confer a clemizole-resistant phenotype in cells:

Huh7.5 cells transfected with either whole cell RNA extracted from the W55R mutant

cells (Figure 2.5B) or with in vitro-transcribed J6/JFH RNA encoding this point

mutation and a linked luciferase reporter gene (Figure 2.5C) were unaffected by 10µM

clemizole. EC50 of clemizole on the W55R mutant J6/JFH RNA was ~18 µM (2.25

times the EC50 on the wild type RNA). Similar to other HCV mutants resistant to an

NS3 protease inhibitor56, the absolute level of replication of the W55R mutant was

lower than that of the wild type genome, suggesting that the drug-resistant mutation

comes at the cost of impaired replication fitness.

27

28

Figure 2.5 Clemizole-resistant mutant.

Replicon cells were passaged in the presence of clemizole and individual colonies

were isolated, propagated and the HCV genomes harbored within were subjected to

sequence analysis. (a) Schematic diagram indicating predicted transmembrane and

intracellular segments of NS4B. Conserved positively charged amino acids are shown

in red. The clemizole-resistant mutation, W55R, is shown in green. (b) HCV

replication in Huh7.5 cells electroporated with 50 g of whole-cell RNA extracted

from cells harboring either wild type or the W55R mutant clone, followed by growth

in the absence (white bars) or presence (gray bars) of 10 M clemizole. Results

represent relative numbers of colonies obtained compared to each corresponding

untreated control. (c) HCV replication assays initiated by electroporation of in vitro

transcribed luciferase reporter– linked wild-type or W55R mutant HCV RNA

genomes performed in the absence (white bars) or presence (gray bars) of 10 M

clemizole. Results represent replication level of each genome relative to its untreated

level. (d) HCV RNA binding of wild-type NS4B and the W55R NS4B mutant as

measured in vitro by microfluidics in the presence of 10 nM clemizole (gray bars) and

摧 Ѝ� � �摧 �m m ce (white bars). (e) In vitro binding curves of W55R NS4B mutant

(solid line,) and wild-type NS4B (broken line, ) to serial dilutions of the RNA

probe. Each data point represents the mean of 10–20 replicates, and the bars represent

the standard error.

29

We also introduced this mutation into the NS4B-GFP vector and tested it for

its RNA binding activity using the in vitro microfluidic assay. Although both mutant

and wild type NS4B proteins experienced ~ a 2 fold reduction of RNA binding in the

presence of 10nM clemizole, because the baseline RNA binding of the mutant is

higher, the residual amount of RNA bound by the mutant in the presence of clemizole

was comparable to that bound by the wild type in the absence of clemizole (Figure

2.5D). Furthermore, this mutant demonstrates greater apparent affinity to the viral

RNA with a Kd of 0.75nM (vs 3.4 nM for wild type NS4B) (Figure 2.5E).

The second clemizole-resistant mutation, termed R214Q, was identified in a

resistant colony as well as in pooled resistant cells. It involves the substitution of a

glutamine for the arginine at amino acid 214 within the cytoplasmic C-terminal

segment of NS4B. Similar results to the first mutation were obtained in cellular and in

vitro analyses done on this mutation, with an EC50 of 40.3µM (~5 times higher than

the EC50 on the wild type RNA) in the luciferase reporter linked replication assay and

a Kd of 0.6nM in the in vitro binding assay (data not shown). Presumably both of these

mutations alter the conformation of NS4B so as to increase its affinity for the viral

RNA. Indeed the Kds measured by the in vitro RNA binding assay reflect this. Taken

together, the above data provides compelling genetic and biochemical evidence for

clemizole’s mechanism of action.

2.3 Discussion

The above results demonstrate that HCV NS4B binds RNA and that this

binding is specific for the 3’ terminus of the negative strand of the viral genome.

30

Because the in vivo specificity of this interaction has not yet been determined

quantitatively, one should use the usual caution in extrapolating the numerical value of

our in vitro results. We also showed that an arginine rich–like motif in NS4B is

essential for RNA binding and for HCV replication. Clemizole hydrochloride was

found to have a substantial inhibitory effect on HCV RNA replication mediated by its

suppression of NS4B’s RNA binding.

Because NS4B is associated with membranes18 and is known to induce

membrane replication sites26, its RNA binding activity offers a mechanism to

incorporate the viral genome into the HCV replication compartment. This may

facilitate the initiation of synthesis of nascent positive strand from the membrane

anchored negative strand. NS4B may also act by recruiting the polymerase complex to

the HCV RNA, via its interaction with the NS5B polymerase or other components of

the replication complex57.

The in vitro IC50 of clemizole for RNA binding by NS4B is ~ 24 ± 1nM

(Figure 2.4d). It is possible that poor cellular permeability accounts for the ~400 fold

difference between the IC50 measured for in vitro RNA binding by NS4B and the EC50

measured for the antiviral effect in cells. We hypothesize that improved drug delivery

and optimization of the compound following structure-activity relationship (SAR)

analysis are expected to improve its potency as an antiviral agent, and that the

microfluidic platform can facilitate this process.

This study also demonstrates the utility of a microfluidic approach. The

microfluidic affinity assay has several important advantages over currently used

31

methods which make it a promising and general tool for drug discovery, especially

against membrane bound targets. First, protein synthesis by mammalian reticulocyte

lysates in the presence of microsomal membranes provides the natural conditions

required for protein folding. Second, the microfluidic assay eliminates the need for a

high level of protein expression and purification, a problem faced by currently used

methods. Third, this assay is capable of detecting transient and low affinity

interactions35. Fourth, the microfluidic device is capable of making many parallel

measurements, and therefore can be used for high throughput screening. Finally, we

have shown in spite of known chemical compatibility issues with the elastomer used to

fabricate the device, one can successfully screen and discover small molecules with

desired pharmacological properties.

As with HIV and tuberculosis, effective pharmacologic control of HCV will

likely best be achieved by a cocktail of multiple drugs against independent virus-

specific targets. Combination of even a moderate NS4B inhibitor with other emerging

anti-HCV agents represents an attractive paradigm for increasing current virologic

response rates. Although we hypothesize that more potent inhibitors than clemizole

can be obtained, because clemizole has already been extensively used in humans

(albeit for a different indication) it may find immediate use as a critical component of

next generation anti-HCV strategies.

32

2.4 Methods

Plasmids

Standard recombinant DNA technology was used to construct and purify all

plasmids. All regions that were amplified by PCR were analyzed by automated DNA

sequencing. Plasmid DNAs were prepared from large-scale bacterial cultures and

purified by a Maxiprep kit (Marligen Biosciences). Restriction enzymes were

purchased from New England Biolabs.

The open reading frames (ORF) of HuR and HuD were obtained from the

ORFeome library of cDNA clones (Open biosystems). These ORFs were inserted into

the expression vector pcDNA-Dest 40 vector (Invitrogen) by the use of gateway

technology (Invitrogen) allowing addition of a C-terminal V5-his tag.

The plasmid pcDNA-NS4B-GFP encoding NS4B of genotype 1a fused in

frame with a C-terminal GFP was previously described 18. The mutations RRa, RRb

and W55R were introduced into this plasmid by site-directed mutagenesis (using the

QuikChange kit (Stratagene)). The plasmid NS5A(AH)GFP was constructed as

previously reported . Gus-his vector was obtained from Roche.

The plasmids pcDNA3.1-5’ UTR pos which encodes the 5’ UTR of the

positive viral strand, was generated by amplification of the 5’ UTR positive sequence

from the Bart79I plasmid58 with primers containing EcoRV restriction sites, digestion

with EcoRV and ligation into the corresponding site in pcDNA3.1 (Invitrogen). The

plasmid pcDNA3.1- 3’ negative terminus which encodes the 3’ terminal region of the

33

negative RNA strand was generated the same way except that the EcoRV-flanked

insert was ligated in an inverse orientation.

The plasmids pcDNA3.1 – 3’ UTR pos and 5’ negative terminus were

similarly generated except that the inserted gene was flanked by HindIII and Xho1

restriction sites.

The vector encoding the delta virus genomic RNA sequence was cloned by

inserting NheI flanked HDV sequence into a pSPT19 vector (Roche Diagnostics) cut

with XbaI45, 59.

The plasmid FL-J6/JFH-5’C19Rluc2AUbi that consists of the full-length HCV

genome and expresses Renilla luciferase was a gift from Dr. Charles M. Rice53. The

W55R mutation was introduced into this plasmid by site-directed mutagenesis (using

the QuikChange kit (Stratagene)).

In vitro RNA transcription and fluorescent and radioactive labeling

Plasmid DNA of the 5’ and 3’ terminal regions of the negative and positive

viral strands were linearized with XbaI. The plasmid DNA of the delta genomic

sequence was linearized with XbaI. Linearized plasmids were then treated with

proteinase K, followed by phenol-chloroform extraction and precipitation with

ethanol. The DNA was resuspended in RNasefree water to a final concentration of 1

µg/µl. 4 µgs of DNA was used as a template for transcription with the T7 MEGAscript

(Ambion) according to the manufacturer’s protocol. The template DNA was digested

by the addition of 5 U of RQ1 DNase (Ambion) and a 15-min incubation at 37°C. The

unincorporated ribonucleotides were removed by size exclusion with a Micro Bio-

34

Spin P-30 column (Bio-Rad), and the transcribed RNA was extracted with phenol-

chloroform, followed by precipitation in ethanol. The RNA pellet was washed with

70% ethanol and resuspended in H2O. Determination of the RNA concentration was

performed by measurement of the optical density at 260nm. The integrity of the RNA

and its concentration were confirmed by 1% agarose gel electrophoresis and ethidium

bromide staining.

The RNA sequences were labeled with Cy5 by using Label IT kit (Mirus)

according to the manufacturer protocol followed by purification on a microspin

column (Mirus) and ethanol precipitation. The number of fluorescent labels per RNA

molecule was determined by measuring the spectrophotometric absorbance of the

nucleic-dye conjugate at 260nm and the λMAX for Cy5 (650nm). This was proportional

to the probes length and was used to adjust our binding experiments results.

Cy3-labeled RNA probes used to study RNA binding by HuR and HuD were

purchased from IDT.

Radioactive labeling of RNA probes with 32P was done as previously

described.60

Device design

Device fabrication and design was done essentially as described35. In brief, a

“control” layer, which harbors all channels required to actuate the valves, is situated

on top of a “flow” layer, which contains the network of channels being controlled. All

biological assays and fluid manipulations are performed on the flow layer. A valve is

created where a control channel crosses a flow channel. The resulting thin membrane

35

in the junction between the two channels can be deflected by hydraulic

actuation.Using multiplexed valve systems allows a large number of elastomeric

microvalves to perform complex fluidic manipulations within these devices.

Introduction of fluid into these devices is accomplished through steel pins inserted into

holes punched through the silicone. Computer-controlled external solenoid valves

allow actuation of multiplexors, which in turn allow complex addressing of a large

number of microvalves. Each unit cell is controlled by three micromechanical valves

as well as a“button” membrane (Supplementary Figure 2.1). The “button membrane”

of each unit cell masks a circular area within the flow channel between two

“sandwich” valves. Aligning and bonding a microarray to the device, allows

positioning of the spotted material (RNA probes or inhibitory compounds) within the

“RNA chamber” of each unit cell. This chamber can then be opened to the flow

channel by the control of a “neck valve”. The used devices had either 640 or 2400 unit

cells.

RNA binding assay

The approach used was a modification of the previously described method for

protein-DNA interactions35 (Supplementary Figure 2.1 online).

1. For soluble proteins:

A. RNA arraying and device alignment.First, dilution series of Cy3-labeled

target RNA sequences (IDT) were spotted as a microarray onto epoxy coated glass

substrates slide (CEL Associates) by OmniGrid Micro (GeneMachines) microarrayer

using a CMP3B pin (TeleChem International, Inc.) with column and row pitches of

36

680 mm and 320 mm, respectively. Each sample solution contained 1% BSA in dH2O

to prevent covalent linkage of the target RNA to the epoxy functional groups as well

as for visualization during alignment. After spotting the arrays were quality controlled

on a GenePix4000b (Molecular Devices). The arrays were then aligned to a

microfluidic device by hand on a SMZ1500 (Nikon) stereoscope and bonded overnight

in the dark on a heated plate at 40ºC.

B. Surface chemistry (Supplementary Figure 2.1c online). All devices were

driven between 10 and 15 psi in the control line and between 4 and 6 psi for the flow

line. For the initial surface derivatization steps the chamber valves remained closed to

prevent liquid from entering the chambers containing the spotted RNA targets. First,

all accessible surface area was derivatized by flowing a solution of biotinylated BSA

(Pierce) resuspended to 2 mg/mL in dH2O for 30 min through all channels, followed

by a 10 min PBS wash. Next a 500 μg/mL Neutravidin (Pierce) solution in PBS was

flown for 20 min, followed by a 10 min PBS wash. Next, the”button” membrane was

closed and the PBS wash continued for an additional 5 min. Then all remaining

accessible surface area except for the circular area of 60 μm masked by the button was

passivated with the same biotinylated solution as above for 30 min, followed by a 10

min PBS wash. Finally a 1:1 solution of biotinylated-6-histidine antibody (Qiagen) in

2% BSA in PBS was loaded for 2-5 min, after which the ”button” membrane was

opened and flow continued for 20 min allowing specific functionalization of the

previously masked circular area. This was again followed by a 10 min PBS wash

completing the surface derivatization procedure.

37

C. Protein synthesis and MITOMI. Next a standard 25 μL TNT T7 coupled

wheat germ extract mixture (Promega) was prepared and spiked with 1 μL

tRNALys−bodipy−fl (Promega) and 2 μL of plasmid DNA template coding for the

appropriate protein (HuR, HuD or Gus). The mixture was immediately loaded into the

device and flushed for 5 min, after which the chamber valves were opened allowing

for dead end filling of the chambers with wheat germ extract. The chamber valves

were again closed and flushing continued for an additional 5 min. Next the segregation

valves separating each unit cell were closed followed by opening of the chamber

valves allowing for equilibration of the unit cell by diffusion. The entire device was

heated to 30ºC on a temperature controlled microscope stage and incubated for up to

90 min to complete protein synthesis, binding of protein to the surface bound

biotinylated-6-histidine antibody, solvation of target RNA, and equilibration of

proteins and target RNA. MITOMI was then performed by closing the ”button”

membrane as well as the chamber valves allowing trapping surface-bound complexes

while expelling any solution phase molecules. Radial closure prevents solvent pockets

from forming between the two interfaces and effectively creates zero dead volume

while preserving the equilibrium concentrations of the molecular interactions to be

detected. This was followed by a 5 min PBS wash to remove untrapped unbound

material. The device was then imaged on a modified arrayWoRxe (AppliedPrecision)

microarray scanner to detect protein synthesis and the trapped molecules.

D. Image and Data Analysis. All images were analyzed with GenePix3.0

(Molecular Devices). For each experiment an image was taken after MITOMI and the

final PBS wash. This was used to determine the concentration of surface bound

38

protein (FITC channel) as well as surface bound target RNA (Cy3 channel). Each unit

cell generated a single data point that consisted of the ratio of median signal of Cy3 to

median signal of bodipy; thus, representing the ratio of surface bound RNA to surface

bound expressed protein. Data from multiple data points were averaged and their

standard deviations calculated. Results were then normalized to 1. Between 10-20

replicates were included for each probe tested. A control protein, Gus-his (Roche) was

included in each experiment and subjected to the same analysis. When present,

detected signal with Gus-his was used for background subtraction.

2. For membrane-associated proteins.

Detection of RNA binding by NS4B was done similarly to the described above

except for several modifications. NS4B was expressed off the microfluidic device by

using coupled transcription/translation rabbit reticulocyte system (Promega). 2 μL

canine microsomal membranes (Promega) were added to the reaction mixture to allow

appropriate protein folding. An NS4B-GFP construct was used thus eliminating the

need for tRNALys−bodipy−fl in the reaction mixture. A biotinylated anti-GFP antibody

(Abcam) was used to specifically functionalize the protected circular area defined by

the button. The protein was anchored to the chip via its interaction with the surface

bound anti-GFP antibodies. The surface bound protein and the Cy5-labeled HCV RNA

probes were loaded into the chip and allowed to incubate for 5-20 min in a 50mM

Hepes KOH (pH 6.8) buffer. Next, MITOMI was performed followed by a brief wash

in Hepes KOH (pH 6.8) buffer. The device was then scanned and results quantified as

described above. NS4B-GFP mutants and NS5A(AH)-GFP were assayed the same

way. Signal detected with NS5A(AH)-GFP was used for background subtraction.

39

Screening of inhibitory compound library

The 1280 compounds of the Lopac library (Sigma) solubilized in Dimethyl

sulfoxide (DMSO) were spotted onto epoxy coated glass substrates slide (CEL

Associates) by OmniGrid Micro (GeneMachines) microarrayer using a CMP3B pin

(TeleChem International, Inc.) as a microarray. For the primary screen compounds

were spotted at a high concentration (~1mM) in duplicates. The array was allowed to

dry, and was then aligned and bonded to a microfluidic device. Two large devices

(2400 unit cells per device) were used for the primary screen. The rest of the

procedure was done similarly to the described above (RNA binding assay). In brief,

the device was subjected to surface patterning that resulted in a circular area coated

with biotinylated anti-GFP antibodies within each unit cell. Next, NS4B-GFP

expressed off chip using coupled transcription/translation rabbit reticulocyte system

(Promega) in the presence of microsomal membranes (Promega) was loaded into the

chip and bound to the surface biotinylated anti-GFP antibodies. Cy5-labeled 3’ neg

terminus RNA probe was then loaded at a concentration of 1.5nM. Each unit cell was

then isolated followed by a 30 min incubation to allow binding of the protein to

surface biotinylated anti-GFP antibodies, solvation of library compounds, and

equilibration of proteins and target RNA. Next, MITOMI was performed trapping

surface-bound complexes while expelling any solution phase molecules. After a brief

wash to remove untrapped unbound material the device was scanned and results

analyzed. The ratio of bound RNA to expressed protein was calculated for each data

point by measuring the median signal of Cy3 to median signal of bodipy. Results were

normalized to signal measured in unit cells containing no inhibitory compound.

40

Greater than 90% inhibition was defined as the cutoff for inhibition in the primary

screen. 104 compounds which were above this cutoff and additional 110 yielding

ambiguous results were subjected to a secondary screen. This was performed

similarly, except that two smaller devices (640 unit cells per each) were used, the

spotted compound concentration was 10 fold lower than in the primary screen and 5

replicates were spotted for each compound. Inhibition greater than 2.5 fold was

considered significant. 18 compounds identified in this screen were further analyzed

for their antiviral effect on HCV RNA replication.

Determination of IC50 for in vitro RNA binding

For an accurate measurement of IC50s, serial dilutions of the inhibitory

compound were loaded onto the microfluidic device by continuous flow while

maintaining a steady concentration of the compound in the flow channel. This helped

to avoid expected losses of the spotted compounds from incomplete solubilization

and/or binding of the compound to PDMS. The experiment was performed essentially

as described in RNA binding assay for transmembrane proteins except that the

expressed protein and the Cy5-labeled HCV RNA probes were incubated in the device

in the presence of the inhibitory compounds or their absence. IC50s were measured as

described in the Statistical Analysis section below.

Expression and purification of recombinant NS4B

GST-NS4B and GST were expressed in E.coli BL21 and purified as described

elsewhere 20. NS4B was fused in frame with an N-terminal 6his-MISTIC protein

(membrane-integrating sequence for translation of IM protein constructs)61. Overnight

41

cultures of E.Coli transformed with mistic-NS4B plasmid were diluted 1:100 in 400ml

of fresh medium and grown at 37ºC to an OD of 0.8. Isopropyl--D-

thiogalactopyranoside (IPTG) (Invitrogen) was then added to a final concentration of

0.1mM. After 3 hours growth at 37ºC cells were pelleted and resuspended in 30ml

lysis buffer (50mM NaCl, 50mM Tris HCL (pH8), 100mM Imidazole, 10mM

Decylmaltoside, Complete EDTA free protease inhibitors (Roche Applied Science)).

Cells were lysed by one cycle in a French Press at a pressure of 10,000 psi for 1

minute, followed by centrifugation at 12,000xg for 5 minutes at 4ºC. The supernatant

was loaded on nickel column (Amersham). Following washes, protein was eluted in a

buffer containing 400mM Imidazole. Glycerol was added at a final concentration of

20% and samples were stored at -20ºC. Purification was monitored by SDS-PAGE.

Total protein concentration was measured using the RC-DC assay (Bio-Rad). NS4B-

mistic was identified by Western blot analyses using monoclonal antibodies against 6-

his (Santa Cruz Biotechnology) and NS4B (Virostat). 6his-mistic was expressed and

purified in a similar manner.

GST pull down assay

Similar to a previously described strategy 62, 1g of purified GST-NS4B or

GST was incubated for an hour at 37ºC in a 50l reaction mixture containing 32P-

labeled in vitro transcribed HCV RNA (corresponding to the 3’ terminus of the

negative viral strand) and binding buffer (10mM DTT, 10mM Na HEPES (pH 7.4),

33mM NaCl, 0.1mM EDTA, 10mM MgCl2). 50l of tRNA pre-coated glutathione-

agarose beads (Sigma) were then added, followed by 1 hour incubation at 4ºC to allow

42

binding of GST to the beads. The beads were then washed three times in binding

buffer and bound RNA was measured by liquid scintillation counting of sample

aliquots. Control incubations with an RNA probe prepared in the absence of T7 RNA

polymerase were used for background subtraction.

RNA Filter Binding Assay

Assays were performed essentially as described 60. Briefly, various

concentrations of mistic-NS4B protein or mistic control were incubated for 1 hour at

30°C with 3.3 nM 32P-labeled in vitro transcribed HCV RNA probe in binding buffer

(50 mM HEPES pH 7.0, 50 mM NaCl, 1 mM MgCl2, 10 ng/µl tRNA, and 0.2mM

Decylmaltoside) in a final volume of 40 µl. Membranes were pre-soaked in the

binding buffer and assembled in a dot blot apparatus (Schleicher & Schull) from top to

bottom as follows: nitrocellulose (Biorad), Hybond N+ (Amersham Biosciences),

Whatman 3mm filter paper. The binding reactions were loaded onto the dot-plot

apparatus and filtered through the membranes. After washing, the membranes were

air-dried and visualized by Phospho-imaging. Results represent percentage of bound

RNA calculated by dividing the signal detected in the nitrocellulose membrane by the

sum of the signals detected in the nitrocellulose and the Hybond membranes.

Cell cultures and electroporation

Huh-7.5 cells were maintained in Dulbecco’s modified minimal essential

medium (Gibco) supplemented with 1% L-glutamine (Gibco), 1% penicillin, 1%

streptomycin (Gibco), 1X nonessential amino acids (Gibco) and 10% fetal bovine

serum (Omega Scientific). Cell lines were passaged twice weekly after treatment with

43

0.05% trypsin–0.02% EDTA and seeding at a dilution of 1:5. Subconfluent Huh-7.5

cells were trypsinized and collected by centrifugation at 700 X g for 5 min. The cells

were then washed three times in ice-cold RNase-free PBS (BioWhittaker) and

resuspended at 1.5*107 cells/ml in PBS. Wild type or mutant FL-J6/JFH-

5’C19Rluc2AUbi RNA for electroporation was generated by in vitro transcription of

XbaI-linearized DNA templates using the T7 MEGAscript kit (Ambion), followed by

purification, essentially as described above (In vitro RNA transcription and fluorescent

labeling). 5 µg of RNA was mixed with 400µl of washed Huh-7.5 cells in a 2mm-gap

cuvette (BTX) and immediately pulsed (0.82 kV, five 99 µs pulses) with a BTX-830

electroporator. After a 10 min recovery at room temperature, pulsed cells were diluted

into 10ml of prewarmed growth medium. Cells from several electroporations were

pooled to a common a stock and seeded in 6 well plates (5*105 cells per well). After

24 hr, medium was replaced and cells were grown in the presence of serial dilutions of

the various inhibitory compounds (Sigma) identified in the screen. 17 commercially

available compounds, out of the 18 identified, were analyzed. Untreated cells were

used as a negative control for water soluble compounds. For compounds solubilized in

DMSO, untreated cells were grown in the presence of corresponding concentrations of

the solvent as a negative control. Medium was changed daily. After 72hr of treatment

cells were subjected to an alamar blue based viability assay and luciferase assay.

Viability assay

Following 72 hrs of treatment cells were incubated for 3 hrs at 37ºC in the

presence of 10% Alamar Blue reagent (TREK Diagnostic Systems). Plates were then

44

scanned and fluorescence was detected by using FLEXstation II 384 (Molecular

Devices, Inc.). Depending on the inhibitory compound’s solvent, water or DMSO,

signal was normalized relatively to untreated samples or samples grown in the

presence of DMSO, respectively.

Luciferase assay

Viral RNA replication was determined using Renilla luciferase assay

(Promega). The same samples subjected to the viability assay described above were

analyzed in this assay. According to the manufacturer protocol, cells were washed

with ice cold PBS and scraped off the plate into 250 µl of ice-cold Renilla lysis buffer.

20 µl of the cell lysates were then loaded onto 96 well plates. 100 µl of the Renilla

luciferase assay buffer containing the assay substrate were injected and luciferase

activity was measured using a Berthold LB 96 V luminometer. As above, signal was

normalized relative to untreated samples or samples grown in the presence of the

corresponding concentration of DMSO.

Luciferase activity detected in samples treated with 100 u/ml Interferon alpha

B2 (PBL biomedical labs) was used as a positive control, demonstrating three log

reduction at 72hr treatment. The experiment was repeated four times, each time with

triplicates. IC50s were measured by fitting data to a three parameter logistic curve

using the formula Y=a+(b-a)/(1+10^(X-c)) (BioDataFit, Chang Bioscience, Inc).

45

Real-time PCR

5x104 Huh7.5 ells were infected with cell culture-grown HCV titered at

1.4x104 TCID50/ml, as described 54. 2 hours after infection, cells were washed three

times in culture medium. Cells were then treated daily with various concentrations of

clemizole. After 72 hours samples were subjected to the viability assay described

above, following which TRIzol Reagent (Invitrogen) was added and total cell RNA

was extracted in triplicates according to the manufacturer’s instructions. Reverse

transcription was then performed using random hexamers and Superscript II reverse

transcriptase (Invitrogen, Carlsbad, CA). Real-time PCR was performed on the

resulting cDNA to quantify the amounts of HCV and actin RNA (in separate

reactions) in each sample. Standards were made using an in vitro-transcribed HCV

RNA and human actin standard (Applied Biosystems, Foster City, CA). HCV was

quantified using primers AGAGCCATAGTGGTCT and

CCAAATCTCCAGGCATTGAGC and probe 6-carboxyfluorescein-

CACCGGAATTGCCAGGACGACCGG-6-carboxytetramethylrhodamine. Actin was

quantified using beta-actin control reagents (Applied Biosystems) according to the

manufacturer’s instructions. HCV RNA level was adjusted to actin level and

normalized relative to untreated samples.

Selection of resistant mutants

Established HCV replicons-harboring cells and Huh7 cells electroporated de

novo with a genotype 1b subgenomic HCV replicon (Bart 79I) 55 were passaged in the

46

presence of neomycin and increasing concentration of clemizole (1-16 µM). Colonies

that were able to grow in the presence of the compound were isolated and propagated

for 5-10 passages. Eleven colonies (out of ~60) survived the passages and were

subjected to sequence analysis, as previously described 20.

Whole cell RNA electroporation

Whole cell RNA was extracted from clemizole-resistant replicon clones and

from untreated replicon cells using TRIzol reagent (Invitrogen). Equal amounts of

whole cell RNA (50µg) were electroporated into Huh7.5 cells as described above.

Cells were grown under G418 selection in the presence or absence of clemizole for 3

weeks. The number of colonies was determined using Image J (NIH) following

fixation and staining with Crystal violet.

Statistical analysis

Dissociation equilibrium constants were determined by fitting data to the

equation describing equilibrium binding; Y = a*X/(b+X) (a and b represent maximum

binding and Kd, respectively) by nonlinear least squares regression fitting (BioDataFit,

Chang Bioscience). IC50s were measured by fitting data to a three parameter logistic

curve using the formula Y=a+(b-a)/(1+10^(X-c)) (a, b and c represent minimum

binding, maximum binding and logEC50, respectively)(BioDataFit, Chang Bioscience,

Inc).

47

Chapter 3. Mutations in NS3 compensate for the disruption of an RNA-binding domain in NS4B

48

3.1 Introduction

NS4B contains an RNA-binding activity

As described in Chapter 2, NS4B harbors the ability to bind HCV RNA.

Molecular determinants of this interaction were identified in both the protein and RNA

partners. The 3’ terminus of the negative-sense HCV RNA genome was shown to bind

to NS4B with higher affinity than the other regions of the genome (Figure 2.2).

Meanwhile, two putative RNA-binding domains (RBDs) were identified in the NS4B

sequence, RR192 and RR247 (Figure 2.3). When both arginines of either of these two

domains were mutated to alanines, binding was inhibited. In the case of RR192,

binding was almost completely abrogated, while for RR247, binding was inhibited

roughly 5-fold (Figure 2.3D). Therefore, these two domains are important for RNA

binding in vitro.

In addition, clemizole hydrochloride was identified as an inhibitor of this RNA

binding activity (Figure 2.4). We determined that this compound not only inhibited

RNA binding in vitro, but also impaired replication in cells. These data support our

hypothesis that RNA binding is important for HCV replication in cells, a hypothesis

we will further explore in this chapter.

NS4B interacts with NS3

As described in Chapter 1, HCV replication takes place in a replication

complex. All the viral nonstructural proteins have been localized to this complex, and

it is likely that these proteins physically interact with each other in order to achieve

49

efficient viral replication. In support of this idea, numerous protein-protein

interactions have been observed among the HCV NS proteins63, 64. Because these

proteins are so intimately involved with one another, it is likely that defects in one

nonstructural protein can be suppressed by mutations in another. In fact, NS3 has been

shown to perform such a role in the context of a defective NS4B; a number of amino

acid substitutions in the former have been identified that compensate for defects in the

latter65.

In this chapter, we hypothesize that the putative RBDs are important for HCV

replication in cell culture. We further hypothesize that defects in these RBDs can be

suppressed by secondary-site mutations in the virus. To test these hypotheses, we

examined the replication capacity of HCV replicons containing arginine to alanine

mutations at the RBDs. We also isolated revertant colonies, identified secondary-site

mutations and tested them for suppression of the KR247AA defect.

3.2 Results

RBD mutations prevent HCV replication

To test whether NS4B’s putative RNA-binding domains are important for

HCV replication, we constructed HCV replicons that contained one of the following

mutations: two consecutive arginines at amino acid 192-193 of the NS4B sequence to

two alanines (RR192AA) or a lysine-arginine sequence at amino acid 247-248 to two

alanines (KR247AA)a. Along with a wild-type and polymerase-defective mutant

a Note that the replicon studied is genotype 1b, not 1a as was tested in the in vitro experiments of Chapter 2. Thus, the 247th amino acid is a lysine, not an arginine.

50

construct, these constructs were electroporated into Huh7 cells, and propagated for

three weeks under G418 selection. The resulting colonies were stained and counted

(Figure 3.1).

Wild-type RNA gave many more colonies than both RBD mutants, yielding a

mean of 3330 colonies per g of RNA electroporated (N = 3, SEM = 450). For the

RR192AA mutant, only 4 colonies were observed over 5 experiments, for a mean of

0.8 colonies / g RNA. For the KR247AA mutant, the observed mean was 72 colonies

/ g RNA (N = 3, SEM = 8), roughly 2% of wild-type. No colonies were observed for

the polymerase-defective mutant. These data are consistent with the clemizole

inhibition experiments of Chapter 2; there, the pharmacological disruption of RNA

binding also led to an inhibition of viral replication.

51

52

Figure 3.1. Putative RNA binding domains are necessary for HCV replication.

Colony formation assays were performed with HCV replicons containing either (A)

wild-type, (B) RR192AA, (C) KR247AA, or (D) polymerase defective sequence. The

number of colonies was counted for each sample, and the results are displayed in (E).

Error bars represent SEM.

53

RBD mutations do not affect HCV localization

One possibility for the reduction in viral replication is that the genetic

disruption of the RBD impairs the expression or proper folding of NS4B and that this

inhibits replication. To test this possibility, we monitored the subcellular distribution

of the HCV nonstructural protein replicase complex. Because antibodies for NS4B are

not readily available, we chose to follow the distribution of another nonstructural

protein, NS5A. NS4B’s proper localization has been shown to be essential for the

membrane-associated foci distribution pattern of NS5A18; therefore, we hypothesized

that if NS4B was improperly folded, NS5A’s localization pattern would be disrupted.

To test this hypothesis, we infected Huh7 cells with a vaccinia virus expressing

the T7 polymerase, then transfected these cells with replicon constructs driven by a T7

promoter that contained a wild-type, RR192AA, or KR247AA HCV replicon. After 4

hours, these cells were fixed and subjected to immunofluorescence analysis. As seen

in Figure 3.2, NS5A’s pattern of subcellular distribution in each of the mutant samples

was reticular, with a predominance of foci in the perinuclear region. This localization

pattern matches that which we observed and has been reported elsewhere for wild-

type; therefore, it is likely that these mutations do not affect NS4B’s subcellular

localization.

54

Figure 3.2. Disruption of RNA binding domains does not affect the formation of

putative HCV replication complexes.

Cells were infected with a vaccinia virus expressing the T7 polymerase and

transfected with HCV replicons containing either (A) wild-type, (B) RR192AA, or (C)

KR247AA sequence. Immunofluorescence was performed with an anti-NS5A

antibody. All three samples show a similar pattern of foci spread through the

endoplasmic reticulum.

55

Primary-site revertants are not observed

In addition to staining and counting colonies from the electroporations, some

colonies were isolated and clonally expanded, in order to determine the mechanism by

which they overcame the RBD defect. RNA was harvested from these clones and was

reverse-transcribed, amplified and sequenced. Notably, no HCV sequence was

detected in the 3 isolated RR192AA colonies. Of the 20 colonies isolated from the

KR247AA samples, all were HCV positive. However, none of the clones had reverted

to the original KR sequence at position 247 of the NS4B region. Instead, some clones

had accumulated one or more mutations in other parts of the HCV genome. In others,

we did not detect any mutations in the genome.

All the mutations observed in the revertant clones are identified in Figure 3.3.

Figure 3.3A highlights the location of all observed mutations in the HCV genome,

while Figure 3.3B lists the type and location of the substitution in the full-length HCV

polyprotein. In addition, Figure 3.3B lists whether or not these mutations suppressed

the KR247AA defect in a transient transfection assay (described below). Those that

did suppress the defect are highlighted in red in Figure 3.3A.

56

Figure 3.3. Replicating colonies containing KR247AA mutation contain

secondary-site mutations.

RNA harvested from cells replicating under G418 selection was reverse-transcribed,

amplified, and sequenced. No reversions to KR at the RBD were observed. (A)

Twelve other substitutions were observed throughout the genome (*). Two of these

mutations, Q1112R and S1200Y, were confirmed to restore replication to wild-type

levels or above in a luciferase assay (indicated in red). (B) All the observed mutations

are listed, including in which clone they were found, whether they improved

replication in KR247AA replicons, and whether they also improved replication in

wild-type replicons. n.d. = not determined.

57

Secondary-site mutations in NS3 restore HCV replication in

KR247AA replicons

Having identified a dozen mutations that arose under G418 selection, we

assayed each of these mutations to determine whether it could suppress the original

KR247AA defect in NS4B. We generated constructs containing the KR247AA

mutation and one of each of the additional observed secondary-site mutations. These

constructs were generated in a luciferase reporter vector, which was used to test the

replication capacity of the constructs in transient transfection assays.

The majority of the constructs did not improve replication levels above that of

the KR247AA alone mutant (Figure 3.3B). However, two of the mutations in NS3 did

enhance replication levels, Q1112R and S1200Y. To our surprise, the KR247AA

Q1112R mutant replicated 3-fold more efficiently than the wild-type replicon (Figure

3.4). Meanwhile, the KR247AA S1200Y mutant replicated at virtually the same level

as the wild-type. Therefore, these mutations in the N-terminal region of NS3 suppress

the defect of the KR247AA mutation in NS4B.

58

Figure 3.4. Mutations in NS3 enhance replication of RBD defective mutants.

Replication assays were performed with various constructs of a luciferase-linked HCV

replicon. Pictured are the mean levels of replication for each of 8 different samples

relative to the wild-type sample. (A) Relative luciferase activity (RLU) of each sample

4 hours after electroporation. All samples express roughly equal amounts of firefly

luciferase. (B) 72 hours after electroporation, luciferase activity is much higher in

some samples (wt, KR + Q1112R, KR + S1200Y, wt + Q1112R, wt + S1200Y) than

in others (KR, KR + L1153I, KR + A1835T). Notably, the Gln 1112 to Arg mutation

(blue) and the Ser 1200 to Tyr mutation (purple/pink) enhanced replication in both

wild-type and KR247AA replicons containing the. Error bars represent SEM. N = 3.

59

Finally, we sought to determine whether these two mutations in NS3 enhance

HCV replication only in the context of the KR247AA mutation or whether they also

enhance replication in an otherwise wild-type replicon. We generated two constructs

in a wild-type backbone, one containing the Q1112R mutation and the other

containing the S1200Y mutation, and tested them in the transient transfection assay.

As shown in Figure 3.4, these mutations also enhanced the replication capacity of

wild-type replicons. The introduction of the Q1112R mutation to a wild-type replicon

enhanced replication by 10-fold, while the introduction of this mutation to a

KR247AA replicon increased replication by 30-fold. The S1200Y mutation enhanced

wild-type replication by 3.6-fold, while it enhanced KR247AA replication by 9.6-fold.

Thus, these mutations enhance HCV replication in a general manner, and they increase

replication more dramatically in the context of the KR247AA mutants.

3.3 Discussion

In this chapter, we have shown that the domains important for RNA binding in

vitro are also important for HCV replication in cells. Furthermore, one of these

domains, RR192, was found to be essential for replication, while the other could be

complemented. These results correlate with our previous in vitro binding data, in

which RR192 was essential for RNA binding, while KR247 appeared to play a

secondary role. We also determined that the formation of the replication complex does

not appear to be affected in either of these mutants, and that revertant colonies did not

contain a primary-site reversion. Finally, two mutations in the NS3 region of the HCV

genome were found to suppress the KR247AA defect.

60

These two mutations, Q1112R and S1200Y, are both located in the protease

domain of NS3. In their study of NS4B:NS3 interactions, Paredes et al.65 found that

the exact same mutation Q1112R could compensate for a more general genetic defect

in NS4Bb. Moreover, they found several additional mutations within the NS3 protease

domain, P1115R, E1202G, and M1205K. As described in that paper, and as shown in

Figure 3.5, all these mutations cluster together in the same region of the quaternary

protein structure. This region is distal to the protease active site and may be important

in interacting with the NS4B protein. In agreement with their data, the S1200Y

mutation is also located in this region, supporting the hypothesis that these mutations

might suppress the NS4B defect through a common mechanism.

As demonstrated by Figure 3.4, each of these two mutations causes a

significant increase in viral replication in wild-type viruses. This suggests that these

mutations may allow for more efficient replication in general. Such an improvement in

efficiency may compensate for the moderate impairment conferred by the KR247AA

mutation. Interestingly, these data do not agree with the report in which Q1112R was

originally identified65. While our luciferase assay showed an increase in replication of

greater than 10-fold (Figure 3.4), Paredes et al. found an increase of less than 1.5-fold.

This disparity may be partly due to minor differences in the HCV replicons used, to

differences in the cell lines used, or perhaps differences in the sensitivities of the assay

used in each case.

b To impair NS4B, the authors created a hybrid replicon containing a genotype 1a NS4B within the context of a genotype 1b backbone.

61

Figure 3.5. Location of compensatory mutations in NS3-4A dimer.

A ribbon diagram of the 3-D structure, as obtained from the protein data bank

(accession number 1CU1), of an NS3-4A dimer is shown. On the upper monomer

(green), the amino acids comprising the protease active site and the compensatory

mutations identified in this manuscript are shown in red. On the lower monomer

(blue), the same amino acids, as well as those previously identified to be

compensatory in the NS3 protease domain are shown in red. These mutations cluster

together in one region of the protein that is not part of the protease active site.

Underlined amino acids indicate those found in this study. The image was generated

with 3D Molecule Viewer of the Vector NTI suite (Invitrogen).

62

In addition, data from patient isolates show that these two substitutions,

Q1112R and S1200Y, are not commonly found in vivo66. Position 1112 is not highly

conserved in patients, and proline, glutamine, leucine, serine, alanine, histidine, and in

one case, arginine, have been observed at that location. In contrast, position 1200 is

highly conserved, with 71% of isolates containing a serine at that location, and 99%

containing either a serine, threonine, or alanine. Tyrosine has not been observed at that

position. Based on these data, it is likely that any enhancement of replication observed

in vitro by these mutations is counterbalanced by a selective disadvantage in vivo. In

the context of developing antiviral drugs that can achieve the same effect as the

KR247AA mutation, such an observation is encouraging in that these Q1112R and

S1200Y resistant mutants may not evolve in infected patients.

While the KR domain has been shown to be important for RNA binding in

vitro, it is possible that this domain may serve another function in the context of viral

replication. Therefore, future experiments will be targeted towards testing NS4B’s

RNA-binding activity in cells and determining the importance of these RBDs to such

an activity. To accomplish this, infection-transfection experiments may be performed,

as in Figure 3.2, in combination with biochemical RNA:protein interaction assays.

Methods

Plasmids

For the colony formation assays, the Bart79I plasmid20 was used. For the

luciferase assays, the Bart79ILuc* (Chapter 4) was used. Cloning of all mutations in

63

these plasmids was performed using the Quikchange method (Stratagene). Primers for

each of these mutations are listed in Table 3.1.

Primer name Primer Sequence (5’ to 3’) RR192AA AGCGATACTGGCTGCGCACGTGGGCCCA KR247AA TCAGCTGCTGGCGGCGCTTCACCAGTGGA Q1112R CGTCGGCTGGCGAGCGCCCCCCG P1142S GGCATGCCGATGTCATTTCGGTGCGCC L1153I CAGGGGGAGCCTAATCTCCCCCAGG R1157K GAGCCTACTCTCCCCCAAGCCCGTCTC S1200Y CTTTGTACCCGTCGAGTATATGGAAACCACTATGC A1835T GCGCCGGCATCACTGGAGCGGCT S2466N CATCTCGCAGCGCAAACCTGCGGCAGAAG Q2750H GCGGGGACCCATGAGGACGAGGC

Table 3.1. Primer sequences

Colony formation assay

Colony formation assays were performed as described20, using a minimum of

three separate electroporations for each of the wild-type, RR192AA, KR247AA, or

pol- RNA. Plates were harvested 20 days post-electroporation and stained with 0.1%

crystal violet.

Infection-transfection

Infection-transfection assays were performed as described18. 2 g of Bart79I

plasmid was transfected after infection with vaccinia virus at a MOI of 5. Coverslips

were fixed at 4 hours post-transfection and immunostaining was performed as

described in Chapter 4, using a monoclonal NS5A antibody67 (1:1000).

64

Picking colonies, RT-PCR, & sequencing

Colonies growing under G418 selection were picked using sterile cloning disks

and expanded under G418 selection until growth had exceeded a 6-well dish. Cellular

RNA was harvested with Trizol according to the manufacturer’s instructions

(Invitrogen). HCV replicon RNA was reverse-transcribed and amplified with

Superscript II (Invitrogen), using gene-specific primers. The PCR products were

directly sequenced and compared to wild type sequence. Of the 20 colonies isolated,

all were sequenced in the NS4B and 5’UTR regions. For the remainder of the HCV

genome, approximately 4 clones were sequenced for each region.

Luciferase assay

Luciferase assays were performed as described in Chapter 4. 5 g of plasmid

DNA was electroporated into Huh7.5 cells. Plasmids used contained wild-type, the

KR247AA mutation, or a combination of the KR247AA mutation and one of Q1112R,

P1142S, L1153I, R1157K, S1200Y, A1835T, S2466N, Q2750H, or wild-type and one

of Q1112R or S1200Y. Cells were harvested at 4 hours, 72 hours, and 96 hours post-

electroporation. At the time of harvest, an alamarBlue© viability assay was performed

according to the manufacturer’s recommendations (Invitrogen). Reported luciferase

values are normalized to each samples viability value.

65

Chapter 4. A small molecule inhibits HCV replication and disrupts NS4B’s subcellular distribution

66

4.1 Introduction

Worldwide, 170 million people are infected with Hepatitis C Virus (HCV), a

(+) sense, single-stranded RNA virus1. These individuals experience significant

morbidity related to their infection, and mortality rates are high. Moreover, treatment

options for this disease are inadequate for the majority of patients. The standard of

care regimen, which relies on pegylated interferon- plus ribavirin, is successful in

only about 50% of patients; moreover, it is associated with significant side effects8.

Therefore, better antiviral therapies are needed to address this public health problem.

As a better picture of the molecular biology of HCV has emerged, a number of

specific antivirals have been identified, and several pharmaceuticals that target the

viral nonstructural proteins NS3 and NS5B have progressed into phase II clinical

trials12. Notably, resistant mutants have already been observed for both these types of

compounds68; therefore, it is likely that a combination of antivirals will provide the

best sustained response in patients.

One potential target for a complementary antiviral is the HCV nonstructural 4B

(NS4B) protein. Several functions have been ascribed to this protein, including the

ability to rearrange cellular membranes into a so-called membranous web, visualizable

through electron microscopy and believed to represent the viral replication platform13,

and to form membrane-associated foci (MAF), believed to reflect, in part, the light

microscopic equivalent of the membranous web16. Furthermore, NS4B can bind and

hydrolyze GTP20, bind to HCV RNA22, and interact physically with other

67

nonstructural components of the HCV replication complex63. Any one of these

functions represents a potential target for pharmacological disruption.

Recently, Chunduru et al. have identified 23 small molecules that bind to

NS4B with a dissociation constant (KD) to NS4B of 5 M or less69. Of those leads, 4

molecules were shown to inhibit HCV replication with an EC50 of 1 M or less, as

measured by an ELISA for NS5A protein levels. Here, we focus on one of these

compounds, which we designate “anguizole.” We show that this compound indeed

inhibits HCV RNA replication with little toxicity to host cells. Furthermore, we

provide genetic and biochemical evidence that NS4B is the target of this drug, and we

demonstrate that its mechanism of action proceeds through the disruption of MAF

formation. Finally, our in vitro dynamic light scattering assay suggests that the second

N-terminal amphipathic helix (AH2) of NS4B interacts with this compound, and we

identify a mutation in the N-terminal region that confers resistance to the drug.

4.2 Results

The small molecule, anguizole, inhibits HCV RNA replication.

Chunduru et al.69 have recently performed a large-scale screen for small

molecules that could bind to recombinant NS4B. This assay was based on monitoring

changes in intrinsic protein fluorescence of recombinant NS4B as an indirect measure

of candidate ligand binding. One of the compounds identified in the screen, 7-

[chloro(difluoro)methyl]-5-furan-2-yl-N-(thiophen-2-ylmethyl)pyrazolo[1,5-

a]pyrimidine-2-carboxamide, is shown in Figure 4.1A, and is hereon referred to as

68

“anguizole”. In the screen, this compound was shown to inhibit HCV protein

expression, as detemined by an ELISA-based HCV replication assay. Initially, we

sought to confirm that anguizole does indeed possess specific anti-HCV activity in the

HCV replicon system.

Luciferase-linked HCV replicons were treated with various concentrations of

anguizole, and luciferase activity was assayed as a measure of HCV replication. A

representative experiment is shown in Figure 4.1b. Treatment with anguizole

significantly inhibited viral RNA replication in both genotype 1b and 1a replicons.

The mean 50% effective concentration (EC50) of the compound in genotype 1b was

310 nM and in genotype 1a was 560 nM. Anguizole appeared to have no significant

cytotoxic effect at the concentrations studied, with a mean 50% cytotoxicity

concentration (CC50) of > 50 M, as measured by the cell proliferation reagent, WST-

1(Roche). Notably, we were unable to determine the EC50 for genotype 2a replicons,

suggesting that this compound’s effect on viral replication may be genotype-specific.

69

Figure 4.1. A small molecule inhibits HCV replication.

(A) Structure of the compound, referred to as “Anguizole,” which was originally

reported as a ligand of NS4B69. (B) Mean 50% effective concentrations (EC50) of the

compound were determined by treating luciferase-linked HCV replicons with various

concentrations of anguizole and assaying luciferase activity as a measure of HCV

replication. Pictured above is a representative experiment for genotype 1b, in which

anguizole concentrations ranging from 0.0001 to 7.2 M were tested. Replication and

cell viability levels are reported as a percentage of the nontreated control. The mean

EC50 was 310 nM, while the CC50 was greater than 50 M. Each data point is the

mean of three replicates, and error bars indicate SEM.

70

Mutations conferring resistance to anguizole map to NS4B.

In order to determine anguizole’s mechanism of action, we first sought to

obtain genetic evidence that anguizole interacts with NS4B. Therefore, we selected for

anguizole-resistant mutants. Replicon cells were grown in the presence of the drug at

either 5- or 10-fold the EC50, and resistant colonies were isolated and propagated.

RNA was extracted from 18 colonies, and the NS4B coding region was reverse-

transcribed, amplified, and sequenced. Several mutations within NS4B were identified

in the resistant colonies (Figure 4.2A), while no NS4B mutations were observed in

untreated replicon cells grown in parallel.

Present in 14 of the 18 isolated clones, the most commonly selected mutation

conferring anguizole resistance was a substitution of histidine to either arginine

(H94R) or asparagine in position 94 of the NS4B amino acid sequence, corresponding

to position 1805 of the HCV polyprotein. To confirm that it was responsible for the

phenotypic resistance to anguizole, the H94R mutation was cloned into a firefly

luciferase-linked HCV reporter construct and electroporated into Huh7.5 cells. The

cells were then dosed with various concentrations of anguizole solubilized in DMSO,

and cells were treated with DMSO alone as a negative control. Luciferase expression

levels were measured after 5 days of treatment. As shown in Figure 4.3A, the replicon

harboring the H94R mutation was indeed much less sensitive to anguizole treatment

(EC50 = 7.5 M, 95% CI = 3.2-17.4 M), relative to the wild-type replicon (EC50 =

0.20 M, 95% CI = 0.08-0.52 M).

71

Figure 4.2. Resistant mutations map to NS4B.

Colony formation assays were performed at either 5- or 10 times the EC50 for this

compound. Replicating colonies were isolated and cDNA from these colonies was

sequenced. (A) Several point mutations in the genotype 1b NS4B coding region were

identified in resistant colonies. Colonies containing such mutations and the stringency

of selection are indicated. At the top are listed the wild-type amino acid found at the

indicated location in the NS4B polypeptide sequence. Observed mutations are listed

below. Highlighted is the most common mutation observed, a mutation of NS4B’s 94th

amino acid—corresponding to the HCV polyprotein’s 1805th amino acid—from a

histidine to either an arginine or an asparagine. (B) A predicted topology of the NS4B

protein, as modified from17. The location of the H94 mutation is indicated. Two

predicted N-terminal amphipathic helices are shown (AH1 and AH2), as are the four

predicted transmembrane domains (TM1-4).

72

To evaluate the replication capacity of the H94R NS4B mutant replicon,

nontreated WT and H94R replicons were monitored as a function of time over the

course of 6 days post-electroporation (Figure 4.3B). Replication kinetics of the

luciferase-linked H94R replicon were similar to those of the wild-type replicon;

however, the efficiency of replication was roughly two-fold lower for the mutant.

Therefore, the H94R mutation appears to carry a significant fitness cost in the

replication of the virus.

Taken together, these findings suggest that anguizole does not merely target

NS4B in vitro, but also likely in the context of viral replication in cells; furthermore,

its antiviral activity appears to depend on the sequence of NS4B.

73

Figure 4.3. Characterization of the H94R resistance mutation.

(A and B) Transient luciferase replication assays were performed with genotype 1b

HCV replicon constructs containing either a histidine (WT, black) or an arginine

(H94R, dark gray) at amino acid 94 in the NS4B sequence. (A) Luciferase assays were

performed following 5 days of treatment with various concentrations of anguizole.

Replication levels (RLU) are shown relative to the maximal luminescence observed

for each electroporation, and they are normalized to cell viability measurements for

each sample. EC50 values were calculated to be 0.20 M for WT and 7.5 M for the

H94R mutant. (B) Replication kinetics were tested for these constructs, along with a

polymerase defective control (Pol-, light gray), over a 6-day period in the absence of

anguizole. Replication levels are shown relative to the maximal luminescence

observed for all electroporations. Compared to wild-type, the H94R mutant is

impaired in its replication ability. Each data point is the mean of three replicates and

error bars represent SEM.

74

Anguizole treatment leads to an altered subcellular distribution

pattern of the NS4B protein.

We next sought to determine the mechanism by which anguizole affects NS4B.

Since NS4B alone has been shown to be sufficient for assembling the membrane-

associated foci (MAF) that are believed to correlate to the sites of viral replication16,

we hypothesized that anguizole might alter the integrity of the MAF.

To test this hypothesis, a plasmid expressing NS4B fused in frame with a C-

terminal green fluorescent protein (GFP) was transfected into Huh7.5 cells. As shown

in Figure 4.4, cells treated with 5 M anguizole for 48 hours displayed a NS4B

distribution pattern distinct from the MAF visible in non-treated cells. Rather than

appearing as small foci (Figure 4.4A), a significant amount of the NS4B proteins in

these treated cells appeared to form elongated snake-shaped structures (Figure 4.4B &

4C). As shown in the insets, these “snakes” were both longer and shaped differently

than the MAF observed in nontreated cells, suggesting that anguizole treatment was

altering the formation of NS4B MAF.

Notably, there was significant variation in the extent of the snake phenotype. A

pattern resembling a network of extremely elongated snakes (Figure 4.4C) was visible

in rare cells, while moderately elongated snakes (Figure 4.4B) were visible in roughly

one fifth of the cells expressing NS4B. In the remaining cells in which NS4B-GFP

was expressed, the distribution pattern of NS4B-GFP either adopted a diffuse

cytoplasmic pattern (Figure B.1A) or displayed a pattern of small, bright foci (Figure

B.1A) unlike the MAF observed in untreated cells.

75

76

Figure 4.4. Anguizole alters the subcellular distribution of NS4B-GFP in

transiently transfected cells.

Huh7.5 cells were cultured in the absence (A, D, F) or presence (B, C, E, G) of 5 M

anguizole, either alone (F, G) or following transfection with NS4B-GFP (A, B, C) or

GFP-NS5A (D, E). (A) In nontreated cells, NS4B-GFP displays an ER-associated

pattern of localization, along with several membrane-associated foci (MAF) of varying

size throughout the cytoplasm. (B and C) In the presence of anguizole, NS4B-GFP

appears to form elongated, curved structures (“snakes”), most commonly observed to

be small and dispersed (B), but occasionally observed to be dramatically long (C).

(Insets of A and B) Zoomed-in images highlight the differences in shape and length

between MAF and snakes. (D and E) The distribution pattern of a GFP-NS5A fusion

protein is equivalent in untreated (D) and treated (E) samples. (F and G) The pattern of

calnexin staining is equivalent in untreated (F) and treated cells (G). Scale bars

represent 50 m.

77

To test the specificity of anguizole’s interaction with NS4B, we examined the

effect of anguizole treatment on another viral protein and on subcellular structures in

Huh7.5 cells. First, when we treated cells expressing an NS5A-GFP fusion protein, no

snake-like structures were observed (Figure 4.4E). Also, when markers of the ER,

actin filaments, microtubules, or mitochondria were visualized, no differences

between the cellular structure of treated and nontreated cells were observed, both in

cells transfected with NS4B-GFP and in those not transfected (Figure 4.4G, B.2).

Therefore, we hypothesized that the observed effect is likely the result of a specific

interaction between anguizole and the NS4B protein.

A resistance mutation also alters the subcellular distribution

pattern of the NS4B protein.

Having identified a mutation in NS4B that confers resistance to anguizole

treatment, we hypothesized that this mutation may also affect NS4B-mediated MAF

formation. To test this hypothesis, we characterized the subcellular distribution of

H94R NS4B-GFP, both in the presence and absence of anguizole.

Transient transfections were performed with a construct of NS4B-GFP

carrying the H94R mutation, as described for Figure 4.4. In the absence of anguizole,

cells expressing the H94R NS4B-GFP mutant displayed a distribution pattern

significantly different from the wild-type MAF pattern (Figure 4.5A vs. Figure 4.4A).

Notably, this pattern was strikingly similar to the snake-like pattern seen in wild-type

cells treated with anguizole, with the exception that the H94R snakes tended to be

longer and more dramatic than those observed in the wild-type samples. These results

78

suggest that the H94R mutation itself is sufficient to disrupt the process of NS4B

MAF formation.

In contrast, cells treated with 5 M anguizole displayed a different distribution

of H94R NS4B-GFP (Figure 4.5B). In these cells, no snake patterns were detected,

and many large foci were present throughout the cytoplasm. These foci are

reminiscent of MAF, but their expression pattern is distinct from that of wild-type

NS4B-GFP in that H94R foci tend to be somewhat larger and present in higher

numbers than wild-type NS4B foci.

Anguizole interacts with the second amphipathic helix of NS4B.

Previously, our laboratory identified an N-terminal amphipathic helix (AH1) in

NS4B that is essential for wild-type distribution of NS4B throughout the cell18. We

have also identified a second, smaller amphipathic helix (AH2), consisting of amino

acids 43 – 65 of the NS4B protein. Recently, we have found that upon addition to a

monodisperse population of lipid vesicles, a synthetic peptide comprising this AH2

dramatically alters the size distribution of the vesicles by inducing massive vesicle

aggregation (N.J. Cho, H. Dvory-Sobol, C.H. Lee, S.J. Cho, P.D. Bryson, C. Frank,

and J.S. Glenn, manuscript in preparation). We hypothesized that the AH2 represents a

target for anguizole and that addition of anguizole might disrupt AH2’s lipid vesicle

aggregating activity.

79

Figure 4.5. The H94R mutation alters NS4B-GFP’s subcellular distribution.

The H94R mutation was cloned into the NS4B-GFP construct, and transient

transfections were performed as in Figure 4.4. (A) In the absence of anguizole, the

subcellular distribution of H94R NS4B-GFP adopts a snake-like pattern. (B) In cells

treated with 5 M anguizole, the snake-like pattern is not observed; instead, NS4B

appears to assemble into a novel pattern of many bright cytoplasmic foci. Scale bars

are 50 m.

80

To test this hypothesis, dynamic light scattering (DLS) measurements were

performed with POPC vesicles and AH2 peptide in the presence or absence of

anguizole. Clemizole hydrochloride—which has been recently shown to inhibit HCV

RNA:NS4B binding, for which a C-terminal domain mediating RNA binding is

essential22—was used as a negative control. As shown in Figure 4.6, the addition of

AH2 peptide to the POPC vesicles radically affected the effective diameter of vesicles

in solution, changing the 88 nm diameter untreated vesicles to aggregates with a mean

effective diameter of 5377 nm (Fig 6B vs. 6A). While clemizole exerted no significant

effect on AH2’s vesicle aggregating activity (Figure 4.6D), anguizole dramatically

inhibited AH2’s ability to aggregate vesicles (Figure 4.6C). Specifically, vesicles

incubated with AH2 and anguizole adopted a mean effective diameter of 126 nm,

comparable to what is observed for vesicles without AH2 (88 nm, Fig 6E). These

results suggest that anguizole specifically interacts with the AH2 segment of NS4B.

81

Figure 4.6. Anguizole interacts with the second amphipathic helix (AH2) of

NS4B.

Extruded synthetic vesicles composed of 1-Palmitoyl-2-Oleoyl-sn-Glycero-3-

Phosphocholine (POPC) lipids were incubated with synthetic peptide comprising the

AH2 region of NS4B and with either anguizole or clemizole hydrochloride. Dynamic

light scattering was employed to measure the effective diameter of vesicles in a

sample. (A – D) The size distribution of vesicle diameters is shown (gray bars), as

well as a gaussian fit of the data (black dashed line). Samples contained (A) 300 M

POPC lipid vesicles only, (B) POPC vesicles + 13 M AH2 peptide, (C) POPC

vesicles + AH2 peptide + 26 M anguizole, and (D) POPC vesicles + AH2 peptide +

26 M clemizole hydrochloride, as a negative control compound. The x-axis is

presented on a logarithmic scale. (E) The mean values of the effective vesicle

diameters in at least 3 independent experiments are shown for all four samples. Error

bars indicate SEM.

82

4.3 Discussion

In this study, we present evidence concerning the mechanism of a novel means

of inhibiting HCV replication. Anguizole, the small molecule studied here, is a

member of a previously uncharacterized class of inhibitors for this pathogen, and we

demonstrate that mutations in the viral NS4B coding region are sufficient to confer

resistance to this compound. Furthermore, we show that anguizole can disrupt the

formation of NS4B membrane-associated foci and that it also disrupts the interaction

of a specific peptide segment of NS4B with lipid vesicles, suggesting that anguizole

can directly bind this region of the protein. Thus, this compound employs a novel

mechanism of action for inhibiting HCV replication.

In addition to the data from the original screen that identified anguizole 69,

three pieces of evidence presented in this study suggest that anguizole’s inhibition of

viral replication is a result of directly targeting the viral NS4B protein. First, the

mutation of a single amino acid in the NS4B coding region is sufficient to

dramatically reduce the virus’ susceptibility to drug treatment (Figure 4.3). Second,

treatment with anguizole significantly alters the subcellular distribution pattern of

NS4B (Figure 4.4), but not that of another viral protein or cellular proteins. Third,

anguizole specifically blocks the function of a portion of the NS4B polypeptide, the

amphipathic helix comprising amino acids 43 – 65. Taken together, these data point to

NS4B as the target with which anguizole interacts. Because NS4B induces the

formation of membranous viral replication structures (the “membranous web”)13 and

is required for the proper localization of other nonstructural viral components18, 63, it is

83

logical that the disruption of functional NS4B would lead to a reduction in viral

replication. We propose that this is what happens upon either the addition of anguizole

or the mutation of NS4B histidine 94 (Figure 4.4C and Figure 4.5A).

Our in vitro dynamic light scattering experiments (Figure 4.6) suggest that

anguizole interacts directly with the second amphipathic helix of NS4B (amino acids

43-65). Interestingly, to date our resistant mutant analysis (Figure 4.2) has not

identified any mutations that map within this region of NS4B. The most common

mutation, which was found to be phenotypically resistant to anguizole, was located at

the 94th amino acid in the NS4B sequence. We interpret these potentially conflicting

results as follows: (1) mutations in the AH2 region are very poorly tolerated by the

virus; (2) mutations at position 94, in contrast, are well tolerated; and (3) the H94R

mutation, in particular, might cause a conformational change in the NS4B protein, one

which counteracts anguizole’s ability to inhibit HCV replication. In support of point

(2), we note that position 94 in NS4B is not strictly conserved in patient isolates. In

addition to histidine, glutamine (1a), asparagine (1b), serine (1b), and threonine (2a)

have each been observed in reference sequences of the indicated genotype66.

Furthermore, we found that the H94R mutation alone is sufficient to drastically alter

NS4B’s subcellular distribution (Figure 4.5A), supporting point (3). The nature of

such a conformational change is unclear, but one possibility is that it partially blocks

anguizole’s binding site on the AH2. Structural studies may help to further clarify the

molecular interactions occurring in this situation.

In addition to identifying a novel mechanism of HCV inhibition, our results

have enhanced our understanding of the biological processes that occur during HCV

84

replication. For example, we have shown that anguizole can disrupt both NS4B

AH2:lipid interactions in vitro (Figure 4.6) and the formation of MAF in cell culture

(Figure 4.4). These results are consistent with the hypothesis that NS4B’s AH2 is a

key determinant for the formation of MAF, a correlate of HCV replication complexes.

In addition, we have shown that the MAF formation process can be altered both

pharmacologically and genetically, resulting in the cytoplasmic distribution of

elongated snake-like structures. Previous work has shown that NS4B can physically

interact with other nonstructural components of the virus, including itself21, 63, and we

hypothesize that these snakes represent assemblies of NS4B proteins that have lost

their ability to form the proper protein:protein or protein:lipid interactions. Future

studies may reveal more insight into these interactions and, more generally, the

processes involved in replication complex formation.

To the best of our knowledge, anguizole is the first pharmacological compound

shown to affect NS4B subcellular distribution. Because targeting multiple stages of

the viral life cycle is likely to be essential for effective HCV therapy70, anguizole is

thus representative of an important new potential class of HCV inhibitors. Not only is

such a class likely to be complementary to inhibitors that target other viral proteins71,

72, but it may also complement inhibitors that target other functions of the same NS4B

protein. In support of such an outcome, our results show that clemizole hydrochloride

does not inhibit AH2’s vesicle-altering properties while anguizole does (Figure 4.6).

Similarly, anguizole does not inhibit NS4B RNA binding, while clemizole does (S.

Einav, unpublished data). Development of such compounds may lead to an efficacious

antiviral cocktail and help manage this disease for the millions of people it affects.

85

4.4 Materials and Methods

DNA constructs and peptides

Standard recombinant DNA technology was used to clone all constructs. The

plasmid pEF-NS4B-GFP was cloned previously20. To generate the H94R mutation in

this construct, site-directed PCR mutagenesis was performed as described20. The first

round of amplification used the primer sets [pEF-6879F, 5’-

GGCCAAGATCTGCACACTGGTATT-3’, to H94R_R, 5’-

TAAACAGGAGGGTACGTTGGGTGGTGAGCGG-3’] and [H94R_F, 5’-

CCGCTCACCACCCAACGTACCCTCCTGTTTA-3’, to pEF-1783R, 5’-

AGGCTGATCAGCGGGTTTA-3’]. The second round of PCR amplification was

achieved using the primer set pEF-6879F to pEF-1783R.

The Bart-Luciferase plasmid, Bart79ILuc*, was cloned from the Bart79I

parent20, an HCV genotype 1b Con1 sequence containing the adaptive mutation

S2204I, and the pGL3-Basic parent (Promega) as follows. A NotI site was introduced

after the 15th amino acid of Core in Bart79I using PCR mutagenesis. The luciferase

gene was PCR amplified from the pGA3Basic plasmid using the primers NotLucS,

GAATGCGGCCGCAATGGAAGACGCCAAAAACATAAAG, and LucDraAS,

CGATTTAAATTACACGGCGATCTTTCCGCCC. The PCR product was ligated

into the Bart79I plasmid following restriction digestions of both components with

NotI and DraI. In addition, the ScaI restriction site in the Luciferase coding region was

removed by introducing a silent mutation into the site by PCR mutagenesis.

86

A peptide comprising the second amphipathic helix region of the genotype 1b

NS4B protein was synthesized (Anaspec). This peptide, Pep4BAH2, contained the

sequence, H-WRTLEAFWAKHMWNFISGIQYLA-NH2.

Drugs and antibodies

7-[chloro(difluoro)methyl]-5-furan-2-yl-N-(thiophen-2-

ylmethyl)pyrazolo[1,5-a]pyrimidine-2-carboxamide (Chemical Block

A2828/0119446), which we refer to here as anguizole, was solubilized in DMSO, and

aliquots of 10mM stocks were maintained at -20° C. Immediately prior to use, these

stocks were diluted to 200-fold the desired final concentration, and this solution was

added directly to the dishes.

Reagents used for fluorescence microscopy included the following antibodies:

rabbit anti-calnexin (Stressgen SPA-860) rabbit anti-PDI (Stressgen SPA-890), mouse

anti--tubulin (Sigma T9026), and Alexa Fluor® 594 conjugated secondary antibodies

(Invitrogen A11012, A21044). Actin filaments and mitochondria were visualized with

phalloidin-TRITC (Sigma P1951) and the Mitotracker dye (Invitrogen M7512),

respectively.

Cell culture, NS4B-GFP transfections, and fluorescence

microscopy

Cells of the human hepatoma cell line, Huh7.5, were cultured in monolayers as

described73, with media consisting of DMEM (Mediatech) supplemented with 1% L-

glutamine (Mediatech), 1% penicillin, 1% streptomycin (Mediatech), and 10% fetal

87

bovine serum (Omega Scientific). Transfections of pEF-NS4B-GFP constructs were

performed with Lipofectamine-2000 (Invitrogen), according to the manufacturer’s

instructions. Following transfection, the cells were cultured in standard media

containing 5 M anguizole and 0.5% (70 mM) DMSO. Controls were cultured in

media containing 0.5% DMSO. Unless noted otherwise, the cells were treated for 48

hours, fixed in 4% formaldehyde, and mounted with Prolong® Gold antifade reagent

with DAPI (Invitrogen). Slides were visualized under a Nikon E600 fluorescence

microscope and images were captured with a SPOT digital camera and Openlab image

acquisition software (Improvision).

Immunofluorescence was performed as above and as described74, with the

following variations: following fixation, cells were permeabilized with 0.2% saponin,

and subcellular structures were visualized using the primary antibodies, rabbit anti-

calnexin (1:1000), rabbit anti-PDI (1:2000), and mouse anti--tubulin (1:1000),

followed by the appropriate secondary antibody (1:600). Actin was visualized with

TRITC-conjugated phalloidin (50 g/ml), and mitochondria were visualized with the

Mitotracker mitochondrial dye (100 nM), according to the manufacturer’s instructions.

Stable luciferase replication assays

The ET cell line was established by stably transfecting Huh7 cells with RNA

transcripts harboring a firefly luciferase-ubiquitin-neomycin phosphotransferase

fusion protein and EMCV IRES-driven genotype 1b Con1 NS3-5B polyprotein

containing the cell culture adaptive mutations E1202G, T1280I, and K1846T75. Cells

were plated at 0.5-1.0 x104 cells/well in 96-well plates and incubated for 24 hrs. Then,

88

anguizole was added to the cells to achieve 10 final concentrations ranging from 0.097

to 50 μM. Luciferase activity was measured 48 hours later by adding a lysis buffer and

the substrate (Promega E2661 and E2620). Under the same conditions, cytotoxicity of

the compounds was determined using the cell proliferation reagent, WST-1 (Roche),

according to the manufacturer’s instructions. Percent inhibition of replication was

determined relative to a no compound control. EC50 values were calculated using the

equation: % inhibition = 100%/[1 + 10(log EC50 – log (I)) * b)], where b is Hill’s coefficient.

Transient luciferase replication assays

Bart-Luciferase RNA was generated by linearizing the Bart79ILuc* vector

with the ScaI restriction enzyme (NEB). Subsequently, in vitro transcription was

performed with the Megascript T7 kit according to the manufacturer’s instructions

(Ambion).

In vitro-transcribed Bart-Luciferase RNA was transfected into Huh7.5 cells by

electroporation, as previously described76. Briefly, 5 g of RNA was transfected into 6

x 106 cells with five 99 µs pulses at 0.82kV over 1.1 seconds, using a BTX-830

electroporator. Cells were seeded into 6-well plates at 3 x 105 cells/well. For EC50

measurements, serial dilutions of anguizole were added to each well to achieve six

final concentration of anguizole ranging from 0.001 to 3.125 M (WT) or 0.01 to

10.24 M (H94R), while the percentage of DMSO in each well remained constant, at

0.5%. Luciferase activity was measured after 5 days of treatment.

For comparing replication kinetics between the wild-type and H94R replicons,

Bart-Luciferase RNA from the wild-type or mutant constructs was electroporated into

89

Huh7.5 cells as above. Bart-Luciferase RNA of a construct containing an NS5B

polymerase lethal mutation20 served as a negative control. Cells were seeded into 6-

well plates at either 5-, 3-, or 1.5 x 105 cells/well. Cells were lysed at 48 hour

timepoints throughout a 6-day experiment.

Prior to lysis of each sample, cell viability was measured with the

alamarBlue® reagent (TREK Diagnostic Systems), according to the manufacturer’s

recommendations.

Relative levels of replication were measured using the Luciferase Assay

System (Promega). Briefly, cells were lysed in 150 l Cell Culture Lysis Buffer,

scraped from the well, transferred into microcentrifuge tubes, and centrifuged at

16,000 x g for 1 minute. 20 l of the resulting supernatant was mixed with 100 µl of

the Firefly luciferase assay buffer containing the assay substrate and luminescence was

measured using a Berthold LB 96 V luminometer.

In all cases, replication levels were normalized to viability measurements. EC50

values were determined by fitting normalized replication data points to the equation

Y=Bottom + (Top-Bottom) / [1+10^(X-Log[EC50])], using the Graphpad Prism

software (GraphPad Software).

Analysis of resistance mutants

Mutations conferring resistance to anguizole treatment were identified as

previously described77. Briefly, ET cells stably harboring a subgenomic genotype 1b

HCV replicon were cultured in 10 cm dishes with standard Huh7.5 media containing

either 1.5 M or 3 M of anguizole and 1 mg/mL G418. An initial round of cell death

90

was followed by proliferation of replicon cells in the presence of the compound. RNA

was extracted from the cells with Trizol (Invitrogen), according to the manufacturer’s

protocol. RNA was then electroporated into 4 x 106 cells of the human hepatoma cell

line, Huh-Lunet, with a BioRad Gene Pulser at 270V, 960 uF and infinite resistance.

Transfected cells were plated in 25 cm2 dishes and selected with 1 mg/mL G418 and

1.5 or 3 M of anguizole until colonies proliferated. Colonies were picked and

propogated until growth had exceeded a T25 flask. Cells were then trypsinized and

lysed in Trizol. RNA was isolated according to the manufacturer’s instructions. HCV

replicon RNA was reverse-transcribed and amplified with Superscript II (Invitrogen),

using a primer specific to the HCV 3’UTR, and NS4B was amplified using gene-

specific primers. The PCR product was cloned, sequenced, and compared to wild type

NS4B sequence.

Dynamic Light Scattering (DLS)

Unilamellar light vesicles were prepared using the extrusion method, as

previously described78. Vesicle size distribution was determined by quasi-elastic

dynamic light scattering (DLS) as previously described 79. Briefly, DLS

measurements were performed by a 90Plus particle size analyzer from Brookhaven

Instrument Corp. 79 and results were analyzed by digital autocorrelator software

(Brookhaven Instruments Corp.). All measurements were taken at a scattering angle of

90, where the reflection effect is minimized. The hydrodynamic diameter of the

vesicles was determined via non-negatively constrained least squares fitting79. The

lipid and peptide concentrations for DLS measurements were 300 M and 13 M,

91

respectively. For inhibitor experiments, either anguizole or clemizole was first

incubated with the peptide at a 2:1 molar ratio. Then, the mixture was added to the

lipid vesicles. All of the experiments were thermostatically controlled at 25 C.

92

Chapter 5. Conclusions

93

Summary

The goal of this work was to better characterize the HCV viral life cycle and

identify small molecule inhibitors that could modulate this life cycle. Towards this

goal, we focused on understanding one viral protein, NS4B, and the multiple roles it

plays in HCV replication. We undertook a multi-pronged approach towards these

goals and focused on two activities, NS4B’s binding of viral RNA and its formation of

membrane-association foci.

In Chapter 2, we identified NS4B as an RNA binding protein. We showed that

this RNA binding activity is specific for the 3’ terminus of the negative-sense HCV

genome, and we determined that two domains in the C-terminus of NS4B contribute to

this activity. In addition, our high-throughput screen revealed several compounds that

could inhibit this activity, and one of these, clemizole hydrochloride, was shown to

also inhibit HCV replication. Finally, we identified a mutation in NS4B that conferred

resistance to clemizole and enhanced NS4B’s RNA-binding activity.

In Chapter 3, we further characterized NS4B’s RNA-binding domains. We

showed that each of these domains is important for HCV replication, and our

replication data correlated with our in vitro binding data. We demonstrated that the

subcellular distribution of NS5A in cells harboring RBD-mutated HCV replicons is

unchanged from that of cells containing wild-type replicons. Interestingly, we identify

two mutations in the protease domain of NS3 that suppress the defect in one of these

RNA binding domains.

94

In Chapter 4, we characterize the mechanism of inhibition of another HCV

antiviral, termed “anguizole.” We show that this compound targets NS4B and that it

can disrupt NS4B’s ability to form membrane-associated foci. Furthermore, we

identify resistant mutations in NS4B and utilize an in vitro light scattering assay to

show that the second amphipathic helix of NS4B is the likely target of the drug.

Significance

This dissertation has uncovered several new features of HCV biology, which

should be helpful in better understanding how this virus propagates. First, we have

shown that the NS4B protein can bind HCV RNA, and that this activity of NS4B is

important for viral replication. One can imagine many different ways in which an

RNA-binding activity may be important to the viral life cycle (discussed in Chapter

2.3), and future studies may uncover the exact stage at which this RNA-binding

activity becomes important. Though other proteins in the virus have been found to

bind RNA, our work was the first to show that the small molecule clemizole

hydrochloride can interrupt this protein:RNA interaction. This compound represents a

good lead towards a potential new class of HCV antivirals.

Our discovery of suppressive mutations in NS3 is consistent with the idea that

all the nonstructural proteins work in intimate association in order to replicate the viral

genome. How these two mutations may enhance HCV replication is unclear, but based

on the previous discovery of multiple mutations present in that location, it is likely that

that region of NS3 is important in interacting with NS4B. Perhaps small molecules can

be designed in the future to block NS4B from interacting with this region of NS3.

95

Other future studies might investigate the effect of these mutations on NS4B RNA

binding either in vitro, or in a cell-based assay.

The characterization of anguizole is another example of how many new

pathways exist for developing antiviral drugs. This compound is the first to be

reported to affect NS4B’s membrane associated foci formation, and likely will lead to

many more compounds that can target this activity. Furthermore, the experiments

presented in this thesis exemplify how useful small molecules can be in the study of

HCV biology. Using this compound, we identified regions of the protein that are a

likely target of the drug and show that these regions may also be involved in

membrane-associate foci formation. Future studies will likely examine the mechanism

by which revertants can overcome the anguizole inhibition. Structural studies may also

clarify how NS4B, anguizole, and lipid vesicles all interact with one another.

By elucidating the details of HCV’s life cycle and establishing new types of

small molecules as potential HCV antivirals, we hope that the new avenues of research

developed here will help lead to new classes of drugs to combat HCV infections.

Thus, we hope this thesis will contribute to making a significant difference in the lives

of millions of individuals around the world who currently must confront the ravages of

hepatitis C.

96

References

97

Appendix A. Supplementary material to Chapter 2

Calibration curve To confirm adequate solubilization of the dried spotted RNA probes and to

exclude the possibility that a large fraction of the solvents stick to the

polydimethylsiloxane (PDMS) composing the microfluidic device, we generated a

calibration curve. Known concentrations of soluble labeled RNA probes were loaded

into the device by continuous flow and mean values of Cy3 signal from multiple unit

cells were averaged. Based on this curve, we calculated the actual concentration of

target RNA sequences used in our assay. On average the actual assayed concentration

was 0.55±0.02 of the spotted one. Of note, a similar factor (0.57) was calculated based

on the ratio between a typical print volume and unit cell volume. All the presented

data were adjusted according to this factor.

RNA binding experiment with high ionic strength buffer The ionic strength of the buffer used in an RNA binding experiment may have

an effect on the protein-RNA interaction. To confirm that the high affinity interactions

detected in our RNA binding assay are not a result of enhancement of electrostatic

interactions due to a low ionic strength buffer we performed RNA binding

experiments of NS4B by microfluidics using high (150mM) ionic strength conditions.

RNA binding by NS4B in the presence of PBS (137 mM NaCl, 2.7 mM KCl, 10 mM

Na2HPO4, 2 mM KH2PO4) or 150mM Hepes was similar to that detected with 50mM

Hepes buffer, commonly used for RNA binding assays by others (1, 2). The Kd of

NS4B binding to the 3’ terminus of the HCV negative strand RNA in the high ionic

98

strength conditions was measured at 6.6+3.3 nM (vs. a Kd of 3.4+1.0 nM in the

presence of low (50mM) ionic strength conditions). These results suggest that there is

only a small electrostatic component to the observed RNA-binding.

ATA inhibits RNA binding by NS4B in a dose-dependent manner Prior to the library screen we wished to determine whether HCV RNA binding

by NS4B can be inhibited pharmacologically. Aurintricarboxilic acid (ATA) is a

compound known to inhibit interactions of proteins with nucleic acids 3. We thus

hypothesized that ATA may similarly inhibit binding of NS4B to HCV RNA. To test

this hypothesis, the NS4B RNA binding assay was repeated in the presence of

increasing concentrations of ATA. Indeed, ATA was found to inhibit binding of HCV

RNA by NS4B in a dose-dependent manner, with an IC 50 of 0.49±0.01μM (p value –

0.0003) (Figure A.5). To our surprise, having not previously realized that ATA was

one of the compounds included in the LOPAC library, ATA was one of the 18

compounds identified in the screen. Since we had previously shown that this

compound inhibits the RNA binding activity of NS4B, this turned to be a useful

internal control for the validity of our screen.

Specificity of hits identified in the small molecule screen To determine the specificity of the hits identified in our small molecule screen

we chose the HuR protein (one of the human RNA binding protein used to validate our

RNA binding assay). Interestingly, other than binding to its target RNA sequence, 4A

(Figure A.3), this protein has been previously shown by others 4 to bind the 3’

terminus of the negative HCV strand. Binding of HuR to the consensus 4A RNA

sequence was tested in the presence of the inhibitory molecules shown to inhibit RNA

99

binding by NS4B. No inhibitory effect on RNA binding was detected with the

majority of the hits including clemizole at a concentration of 0.1mM. Similarly,

0.1mM of the identified compounds didn’t have an inhibitory effect on binding of

HuR to the 3’ terminus of the negative HCV RNA strand. In contrast to the other hits,

ATA, known as a non-specific inhibitor of protein-nucleic acids interactions,

significantly inhibited HuR binding to both 4A and 3’ terminus of the negative HCV

strand. These results suggest that the identified hits including clemizole are indeed

specific to RNA binding by NS4B.

100

101

Figure A.1. Microfluidic based RNA binding assay.

3 individual unit cells (out of hundreds/thousands per microfluidic device) are shown

in this scheme.

a- Compartments and micromechanical valves. A valve is created where a control

channel crosses a flow channel. The resulting thin membrane in the junction between

the two channels can be deflected by hydraulic actuation.Using multiplexed valve

systems allows a large number of elastomeric microvalves to perform complex fluidic

manipulations within these devices.

b- Experimental protocol. ( ) represents Cy3 labeled RNA probe. ( ) represents

surface bound bodipy-labeled protein. 1) target RNA sequences labeled with Cy3 were

spotted onto an epoxy-coated slide as a microarray 2) The microfluidic device was

aligned and bonded to the slide allowing contact of each spot in the array with a unit

cell in the device. The device was subjected to surface patterning that resulted in a

circular area coated with biotinylated anti-histidine antibodies within each unit cell

(see c). The “neck valve” separating the flow channels from the RNA chamber

remained closed during this process. 3) The device was then loaded with in vitro

transcription/ translation mixture containing DNA template encoding a his-labeled

protein. Bodipy-labeled tRNALys was added for protein labeling. 4) Each unit cell

was then isolated by the control of two “sandwich” micromechanical valves. 5) the

“neck valve” was opened forming a single compartment by combining the flow

channel compartment with the RNA chamber. This was followed by an incubation to

allow protein synthesis, binding of the synthesized protein to the surface biotinylated

anti-his antibodies, solvation of target RNA, and equilibration of proteins and target

RNA. 6) MITOMI was then performed by actuation of a “button” membrane trapping

surface-bound complexes while expelling any solution phase molecules. 7) The

“sandwich” valves were opened followed by a brief wash to remove untrapped

unbound material. 8) the “button” membrane was opened and the device was scanned

by an array scanner. Trapped RNA molecules and expressed protein were detected.

The ratio of bound RNA to expressed protein was calculated for each data point by

measuring the median signal of Cy3 to median signal of bodipy (represented by ).

102

c- Surface patterning. 1)Accessible surface area was derivatized by flowing a

solution of biotinylated BSA ( ) through all flow channels 2) a Neutravidin solution

( ) was loaded 3) The “button” membrane was activated 4) all remaining accessible

surface area except for a circular area of 60 μm masked by the button was passivated

with biotinylated solution ( ) 5) the “button” membrane was opened 6)a solution of

biotinylated-antibody ( ) was loaded allowing specific functionalization of the

previously masked circular area 7) In vitro expressed protein ( ) was loaded into the

device and bound to the biotinylated antibody coating the discrete circular area. Each

of the described steps was followed by a PBS wash.

103

Figure A.2. Signal to noise ratio in our RNA binding assay.

a- Binding of HuD-his (●) and Gus-his (□) to increasing concentration of the AU3

RNA probe.

b- Binding of NS4B-GFP-microsomal membranes (mm) (●, continuous line) and

NS5A(AH)-GFP-mm (■, dashed line) to increasing concentrations of the 3’ terminus

of the negative strand.

The gray striped column represents the range of probe concentration for which the

signal to noise ratio was the greatest. All of our reported experiments were performed

within this range. Representative experiments are shown. Bars represent standard

deviation.

104

105

Figure A.3. Microfluidics-based analysis of RNA binding by another human

protein from the ELAV-like family, HuR (ELAV L1).

a. Target RNA sequences used to study binding of HuR to RNA and the phenotype

demonstrated by conventional RNA binding methods 5.

b. HuR binds RNA by microfluidics. A microarray of Cy3-labeled target RNA

sequences was used to program a microfluidic device, and binding of bodiby-labeled

proteins expressed on the device to the RNA sequences was assayed. Results represent

the ratio of bound RNA (median Cy3 signal) to expressed protein (median bodipy

signal). Normalized mean values for 10-20 replicates measured in two independent

experiments are shown. Error bars represent standard deviation. The gray bars

represent binding of HuR-his and the clear bars that of Gus-his, used as a negative

control.

c. HuR RNA binding is not affected by the 5 most active compounds, but is affected

by ATA. We tested binding of HUR to its 4A RNA target in the presence and absence

of NS4B RNA binding inhibitors. Data represent mean value of 10-20 replicates and

bars represent standard deviation.

106

Figure A.4. Binding of NS4B to HCV RNA by conventional methods.

NS4B was expressed in E.coli fused to GST or mistic-6his and purified.

a. Binding activity of GST-NS4B and GST to 32P-labeled RNA probe corresponding

to the 3’ terminus of the negative viral strand, as measured by a GST pull down assay.

b. Nitrocellulose membranes from a representative RNA filter binding assay. 32P-

labeled HCV RNA probe bound to 10mM (left panel) or 2mM (right panel) mistic-

NS4B or mistic control is shown.

c. Percentage of bound RNA to mistic-NS4B and mistic proteins at protein

concentrations of 10mM (gray bars) and 2mM (white bars), as measured by a filter

binding assay.

107

Figure A.5. ATA inhibits RNA binding by NS4B in a dose dependent manner.

A dose response curve of RNA binding by NS4B in the presence of increasing

concentrations of ATA, as measured by microfluidics. The Y category is bound RNA

to protein ratio relative to binding in the absence of ATA. The X category is ATA

concentration (μM). Data were fit to a 3 parameter logistic curve using the formula

Y=a+(b-a)/(1+10^(X-c)) (BioDataFit, Chang Bioscience) and IC50 was calculated at

0.49±0.01 μM (p value - 0.0003).

108

Figure A.6. Clemizole inhibits HCV replication by real-time PCR assays.

Real-time PCR assay in HCV infected cells showing that clemizole inhibits HCV

replication (left axis, blue ♦) with no measurable toxicity to the cell (right axis, red ■).

109

Appendix B. Supplementary material to Chapter 4

Figure B.1. Subcellular distribution pattern of NS4B-GFP varies in anguizole-

treated cells.

Experiments were performed as in Figure 4. Images show five different categories of

NS4B distribution: (A) diffuse cytoplasmic staining, (B) small, bright foci, (C) small

snakes, (D) elongated snakes, and (E) overexpression. Beneath each representative

image the frequency at which these patterns were observed is shown. The “small,

bright foci pattern”, which was the most common pattern observed, may represent an

early stage of snake formation.

110

Figure B.2. Distribution patterns of cellular markers do not vary with anguizole

treatment.

Representative fluorescent microscopy images for four different host cell markers are

shown. Markers were visualized with (Actin) phalloidin, (Tubulin) an anti--tubulin

antibody, and (Mitochon.) Mitotracker®. No obvious differences were observed in

cells treated with 5 M anguizole (+ Agl) compared to control cells treated with an

equivalent amount of DMSO (- Agl).

111

Bibliography

1. Pawlotsky, J.M. Pathophysiology of hepatitis C virus infection and related liver disease. Trends Microbiol 12, 96-102 (2004).

2. Perz, J.F., Armstrong, G.L., Farrington, L.A., Hutin, Y.J. & Bell, B.P. The contributions of hepatitis B virus and hepatitis C virus infections to cirrhosis and primary liver cancer worldwide. J Hepatol 45, 529-38 (2006).

3. Rustgi, V.K. The epidemiology of hepatitis C infection in the United States. J Gastroenterol 42, 513-21 (2007).

4. Afdhal, N.H. The natural history of hepatitis C. Semin Liver Dis 24 Suppl 2, 3-8 (2004).

5. Armstrong, G.L. et al. The prevalence of hepatitis C virus infection in the United States, 1999 through 2002. Ann Intern Med 144, 705-14 (2006).

6. Feld, J.J. & Hoofnagle, J.H. Mechanism of action of interferon and ribavirin in treatment of hepatitis C. Nature 436, 967-72 (2005).

7. Strader, D.B. Understudied populations with hepatitis C. Hepatology 36, S226-36 (2002).

8. Moradpour, D., Penin, F. & Rice, C.M. Replication of hepatitis C virus. Nat Rev Microbiol 5, 453-63 (2007).

9. Friebe, P., Lohmann, V., Krieger, N. & Bartenschlager, R. Sequences in the 5' nontranslated region of hepatitis C virus required for RNA replication. J Virol 75, 12047-57 (2001).

10. Sarrazin, C. et al. SCH 503034, a novel hepatitis C virus protease inhibitor, plus pegylated interferon alpha-2b for genotype 1 nonresponders. Gastroenterology 132, 1270-8 (2007).

11. Reesink, H.W. et al. Rapid decline of viral RNA in hepatitis C patients treated with VX-950: a phase Ib, placebo-controlled, randomized study. Gastroenterology 131, 997-1002 (2006).

12. Manns, M.P. et al. The way forward in HCV treatment--finding the right path. Nat Rev Drug Discov 6, 991-1000 (2007).

13. Egger, D. et al. Expression of hepatitis C virus proteins induces distinct membrane alterations including a candidate viral replication complex. J Virol 76, 5974-84 (2002).

14. El-Hage, N. & Luo, G. Replication of hepatitis C virus RNA occurs in a membrane-bound replication complex containing nonstructural viral proteins and RNA. J Gen Virol 84, 2761-2769 (2003).

15. Gosert, R. et al. Identification of the hepatitis C virus RNA replication complex in Huh-7 cells harboring subgenomic replicons. J Virol 77, 5487-92 (2003).

16. Gretton, S.N., Taylor, A.I. & McLauchlan, J. Mobility of the hepatitis C virus NS4B protein on the endoplasmic reticulum membrane and membrane-associated foci. J Gen Virol 86, 1415-21 (2005).

112

17. Lundin, M., Lindstrom, H., Gronwall, C. & Persson, M.A. Dual topology of the processed hepatitis C virus protein NS4B is influenced by the NS5A protein. J Gen Virol 87, 3263-72 (2006).

18. Elazar, M., Liu, P., Rice, C.M. & Glenn, J.S. An N-terminal amphipathic helix in hepatitis C virus (HCV) NS4B mediates membrane association, correct localization of replication complex proteins, and HCV RNA replication. J Virol 78, 11393-400 (2004).

19. Gouttenoire, J. et al. Identification of a novel determinant for membrane association in hepatitis C virus nonstructural protein 4B. J Virol 83, 6257-68 (2009).

20. Einav, S., Elazar, M., Danieli, T. & Glenn, J.S. A nucleotide binding motif in hepatitis C virus (HCV) NS4B mediates HCV RNA replication. J Virol 78, 11288-95 (2004).

21. Yu, G.Y., Lee, K.J., Gao, L. & Lai, M.M. Palmitoylation and polymerization of hepatitis C virus NS4B protein. J Virol 80, 6013-23 (2006).

22. Einav, S. et al. Discovery of a hepatitis C target and its pharmacological inhibitors by microfluidic affinity analysis. Nat Biotech 26, 1019-1027 (2008).

23. Liang, T.J., Rehermann, B., Seeff, L.B. & Hoofnagle, J.H. Pathogenesis, natural history, treatment, and prevention of hepatitis C. Ann Intern Med 132, 296-305 (2000).

24. Reed, K.E. & Rice, C.M. Overview of hepatitis C virus genome structure, polyprotein processing, and protein properties. Current topics in microbiology and immunology 242, 55-84 (2000).

25. Rice, C.M. Flaviviridae: The viruses and their replication. In B. N. Fields, D. M. Knipe, and P. M. Howley (ed.), Fields Virology. (Lippincott-Raven Publications, Philadelphia, Pa., 1996).

26. Egger, D. et al. Expression of hepatitis C virus proteins induces distinct membrane alterations including a candidate viral replication complex. J Virol 76, 5974-5984 (2002).

27. Gosert, R. et al. Identification of the hepatitis C virus RNA replication complex in Huh-7 cells harboring subgenomic replicons. J Virol 77, 5487-5492 (2003).

28. El-Hage, N. & Luo, G. Replication of hepatitis C virus RNA occurs in a membrane-bound replication complex containing nonstructural viral proteins and RNA. J Gen Virol 84, 2761-2769 (2003).

29. Kusov, Y.Y., Probst, C., Jecht, M., Jost, P.D. & Gauss-Müller, V. Membrane association and RNA binding of recombinant hepatitis A virus protein 2C. Arch Virol 143, 931-944 (1998).

30. Rodríguez, P.L. & Carrasco, L. Poliovirus protein 2C contains two regions involved in RNA binding activity. J Biol Chem 270, 10105-10112 (1995).

31. Echeverri, A.C. & Dasgupta, A. Amino terminal regions of poliovirus 2C protein mediate membrane binding. Virology 208, 540-553 (1995).

32. Overington JP, A.-L.B., Hopkins AL. How many drug targets are there? Nat Rev Drug Discov 5, 993-6 (2006).

113

33. Lundin, M., Monné, M., Widell, A., Von Heijne, G. & Persson, M.A.A. Topology of the membrane-associated hepatitis C virus protein NS4B. J Virol 77, 5428-5438 (2003).

34. Hadd AG, R.D., Halliwell JW, Jacobson SC, Ramsey JM. Microchip device for performing enzyme assays. Anal Chem 69, 3407-12 (1997).

35. Maerkl, S.J. & Quake, S.R. A systems approach to measuring the binding energy landscapes of transcription factors. Science 315, 233-237 (2007).

36. Toepke, M.W. & Beebe, D.J. PDMS absorption of small molecules and consequences in microfluidic applications. Lab on a chip 6, 1484-1486 (2006).

37. Lee, J.N., Park, C. & Whitesides, G.M. Solvent compatibility of poly(dimethylsiloxane)-based microfluidic devices. Analytical chemistry 75, 6544-6554 (2003).

38. Whitesides, G.M. The origins and the future of microfluidics. Nature 442, 368-373 (2006).

39. Kang, L., Chung, B.G., Langer, R. & Khademhosseini, A. Microfluidics for drug discovery and development: from target selection to product lifecycle management. Drug discovery today 13, 1-13 (2008).

40. Myer, V.E., Fan, X.C. & Steitz, J.A. Identification of HuR as a protein implicated in AUUUA-mediated mRNA decay. Embo J 16, 2130-2139 (1997).

41. Park, S., Myszka, D.G., Yu, M., Littler, S.J. & Laird-Offringa, I.A. HuD RNA recognition motifs play distinct roles in the formation of a stable complex with AU-rich RNA. Molecular and cellular biology 20, 4765-4772 (2000).

42. Park-Lee, S., Kim, S. & Laird-Offringa, I.A. Characterization of the interaction between neuronal RNA-binding protein HuD and AU-rich RNA. J Biol Chem 278, 39801-39808 (2003).

43. Park, S., Myszka, D.G., Yu, M., Littler, S.J. & Laird-Offringa, I.A. HuD RNA recognition motifs play distinct roles in the formation of a stable complex with AU-rich RNA. Mol Cell Biol 20, 4765-72 (2000).

44. Park-Lee, S., Kim, S. & Laird-Offringa, I.A. Characterization of the interaction between neuronal RNA-binding protein HuD and AU-rich RNA. J Biol Chem 278, 39801-8 (2003).

45. Glenn, J.S., Taylor, J.M. & White, J.M. In vitro-synthesized hepatitis delta virus RNA initiates genome replication in cultured cells. J Virol 64, 3104-7 (1990).

46. Glenn, J.S., Watson, J.A., Havel, C.M. & White, J.M. Identification of a prenylation site in delta virus large antigen. Science 256, 1331-3 (1992).

47. Burd, C.G. & Dreyfuss, G. Conserved structures and diversity of functions of RNA-binding proteins. Science 265, 615-621 (1994).

48. Wung, C.H. et al. Identification of the RNA-binding sites of the triple gene block protein 1 of bamboo mosaic potexvirus. J Gen Virol 80 ( Pt 5), 1119-1126 (1999).

49. Burd, C.G. & Dreyfuss, G. Conserved structures and diversity of functions of RNA-binding proteins. Science 265, 615-21 (1994).

50. Rodriguez, P.L. & Carrasco, L. Poliovirus protein 2C contains two regions involved in RNA binding activity. J Biol Chem 270, 10105-12 (1995).

114

51. Wung, C.H. et al. Identification of the RNA-binding sites of the triple gene block protein 1 of bamboo mosaic potexvirus. J Gen Virol 80 ( Pt 5), 1119-26 (1999).

52. Spångberg, K., Wiklund, L. & Schwartz, S. HuR, a protein implicated in oncogene and growth factor mRNA decay, binds to the 3' ends of hepatitis C virus RNA of both polarities. Virology 274, 378-390 (2000).

53. Tscherne, D.M. et al. Time- and temperature-dependent activation of hepatitis C virus for low-pH-triggered entry. J Virol 80, 1734-1741 (2006).

54. Lindenbach, B.D. et al. Complete replication of hepatitis C virus in cell culture. Science 309, 623-626 (2005).

55. Elazar, M. et al. Amphipathic helix-dependent localization of NS5A mediates hepatitis C virus RNA replication. J Virol 77, 6055-6061 (2003).

56. Tong, X. et al. Identification and analysis of fitness of resistance mutations against the HCV protease inhibitor SCH 503034. Antiviral Res 70, 28-38 (2006).

57. Dimitrova, M., Imbert, I., Kieny, M.P. & Schuster, C. Protein-protein interactions between hepatitis C virus nonstructural proteins. J Virol 77, 5401-5414 (2003).

58. Blight, K.J., Kolykhalov, A.A. & Rice, C.M. Efficient initiation of HCV RNA replication in cell culture. Science 290, 1972-4 (2000).

59. Glenn, J.S., Watson, J.A., Havel, C.M. & White, J.M. Identification of a prenylation site in delta virus large antigen. Science 256, 1331-1333 (1992).

60. Huang, L. et al. Hepatitis C virus nonstructural protein 5A (NS5A) is an RNA-binding protein. J Biol Chem 280, 36417-36428 (2005).

61. Roosild, T.P. et al. NMR structure of Mistic, a membrane-integrating protein for membrane protein expression. Science 307, 1317-1321 (2005).

62. Schilders, G., Raijmakers, R., Raats, J.M.H. & Pruijn, G.J.M. MPP6 is an exosome-associated RNA-binding protein involved in 5.8S rRNA maturation. Nucleic Acids Res 33, 6795-6804 (2005).

63. Gao, L., Aizaki, H., He, J.W. & Lai, M.M. Interactions between viral nonstructural proteins and host protein hVAP-33 mediate the formation of hepatitis C virus RNA replication complex on lipid raft. J Virol 78, 3480-8 (2004).

64. Dimitrova, M., Imbert, I., Kieny, M.P. & Schuster, C. Protein-protein interactions between hepatitis C virus nonstructural proteins. J Virol 77, 5401-14 (2003).

65. Paredes, A.M. & Blight, K.J. A genetic interaction between hepatitis C virus NS4B and NS3 is important for RNA replication. J Virol 82, 10671-83 (2008).

66. Kuiken, C., Yusim, K., Boykin, L. & Richardson, R. The Los Alamos hepatitis C sequence database. Bioinformatics 21, 379-84 (2005).

67. Elazar, M. et al. Amphipathic helix-dependent localization of NS5A mediates hepatitis C virus RNA replication. J Virol 77, 6055-61 (2003).

68. Mo, H. et al. Mutations conferring resistance to a hepatitis C virus (HCV) RNA-dependent RNA polymerase inhibitor alone or in combination with an

115

HCV serine protease inhibitor in vitro. Antimicrob Agents Chemother 49, 4305-14 (2005).

69. Chunduru, S.K. et al. 1686949-A2: (VIROPHARMA INC (VIRO-Non-standard) CHUNDURU S K (CHUN-Individual) BENATATOS C A (BENA-Individual) NITZ T J (NITZ-Individual) BAILEY T R (BAIL-Individual), 2005).

70. De Francesco, R. & Migliaccio, G. Challenges and successes in developing new therapies for hepatitis C. Nature 436, 953-60 (2005).

71. Lin, C., Kwong, A.D. & Perni, R.B. Discovery and development of VX-950, a novel, covalent, and reversible inhibitor of hepatitis C virus NS3.4A serine protease. Infect Disord Drug Targets 6, 3-16 (2006).

72. Malcolm, B.A. et al. SCH 503034, a mechanism-based inhibitor of hepatitis C virus NS3 protease, suppresses polyprotein maturation and enhances the antiviral activity of alpha interferon in replicon cells. Antimicrob Agents Chemother 50, 1013-20 (2006).

73. Sklan, E.H. et al. TBC1D20 is a Rab1 GTPase-activating protein that mediates hepatitis C virus replication. J Biol Chem 282, 36354-61 (2007).

74. Matto, M., Rice, C.M., Aroeti, B. & Glenn, J.S. Hepatitis C virus core protein associates with detergent-resistant membranes distinct from classical plasma membrane rafts. J Virol 78, 12047-53 (2004).

75. Lohmann, V., Hoffmann, S., Herian, U., Penin, F. & Bartenschlager, R. Viral and cellular determinants of hepatitis C virus RNA replication in cell culture. J Virol 77, 3007-19 (2003).

76. Lindenbach, B.D. et al. Complete replication of hepatitis C virus in cell culture. Science 309, 623-6 (2005).

77. Trozzi, C. et al. In vitro selection and characterization of hepatitis C virus serine protease variants resistant to an active-site peptide inhibitor. J Virol 77, 3669-79 (2003).

78. Cho, N.J., Kanazawa, K.K., Glenn, J.S. & Frank, C.W. Employing two different quartz crystal microbalance models to study changes in viscoelastic behavior upon transformation of lipid vesicles to a bilayer on a gold surface. Anal Chem 79, 7027-35 (2007).

79. Kolchens, S., Ramaswami, V., Birgenheier, J., Nett, L. & O'Brien, D.F. Quasi-elastic light scattering determination of the size distribution of extruded vesicles. Chem Phys Lipids 65, 1-10 (1993).