MECHANISM OF HIV ENTRY: SURFACE

STIFFNESS AND STERIC DEFENSE

by

Yu Shi

A thesis submitted to the faculty of The University of Utah

in partial fulfillment of the requirements for the degree of

Master of Science

Department of Biochemistry

The University of Utah

December 2007

Copyright © Yu Shi 2007

All Rights Reserved

THE UNIVERSITY OF UTAH GRADUATE SCHOOL

SUPERVISORY COMMITTEE APPROVAL

of a thesis submitted by

Yu Shi

This thesis has been read by each member of the following supervisory committee and by

majority vote has been found to be satisfactory.

Chair: Michael S. Kay

Christopher P. Hill

THE UNIVERSITY OF UTAH GRADUATE SCHOOL

FINAL READING APPROVAL

To the Graduate Council of the University of Utah:

I have read the thesis of Yu Shi in its final form and have found that (1) its format, citations, and bibliographic style are consistent and acceptable; (2) its illustrative materials including figures, tables, and charts are in place; and (3) the final manuscript is satisfactory to the supervisory committee and is ready for

submission to The Graduate School.

<l-(J-67

Date Michael S. Kay

Chair: Supervisory Committee

Approved for the Major Department

Dana Carroll

Chair/Dean

Approved for the Graduate Council

�\.n� _ _ .

David S. Chapman

Dean of The Graduate School

ABSTRACT

HIV undergoes a maturation process required for infectivity after leaving the cell.

Significant internal morphological rearrangements occur during the maturation process,

converting the immature HIV particle with a thick protein shell into a mature particle

with a thin protein shell and a prominent conical core. In this study, we explored the

linkage between virion morphology and mechanical properties during maturation. Our

results showed that the stiffness of HIV drops dramatically during maturation, and that

there is a correlation between virion stiffness and entry ability. We also showed that

virion stiffness is mainly regulated by the Env CT domain.

During HIV entry, HIV Env (gp120/gp41 complex) undergoes a senes of

conformational changes that induce viral and target cell membrane fusion. Membrane

fusion is driven by the formation of gp41 six-helix bundle composed of the N-trimer

region surrounded by three helices from the C-peptide region (C-peptides). Prevention of

six-helix bundle formation inhibits men1brane fusion and viral entry. This critical role in

entry and its highly conserved sequences make the gp41 N-trimer a key drug target.

gp41 C-peptides bind the N -trimer and potently inhibit HIV entry by preventing

binding of the endogenous C-peptide region. Extensive efforts have been made to find

potent broadly neutralizing Abs against the N-trimer, but without success. Our lab

previously demonstrated that there is a steric block on the viral side of N-trimer region,

which may explain why such Abs are difficult to discover.

In this study, we first explored the geometry of the N-trimer steric block using

bulky cargo proteins attached to a C-peptide in different orientations. We demonstrated

that a less severe block also exists on the cellular side ofN-trimer. Next, a modified HIV

entry assay was used to dissect sources of the block. Our results suggest viral components

to be the major source. The information and reagents obtained from these studies will be

used to design, test and optimize mimics of the N-trimer with an artificial sterk block to

select for neutralizing Abs that overcome the steric block.

v

To my parents

TABLE OF CONTENTS

ABSTRACT ......................................................................................... iv

LIST OF TABLES .................................................................................. ix

LIST OF FIGURES .... , ........................ " ................................................... x

CHAPTERS

INTRODUCTION ......................................................................... 1

AIDS Epidemic ............................................................................. 1

HIV Components .... , ......................................... ' ......................... 1

HIV Lifecycle .............................................................................. 3

HIV Mahlration ............................................................................. 4

Env Mediated HIV Entry ........... " ....................... , ............................ 5

gp41 N-trimer Is an Attractive Drug Target .......................................... 9

gp41 N-trimer Steric Block ............................................................... 9

References .................................................................................. 12

2 A STIFFNESS SWITCH IN HUMAN IMMUNODEFIENCY VIRUS ... 15

Introduction ............... ,................................................................ 16

Materials and Methods .................................................................... 17

Results and Discussion ................................................................... 18

Supplementary MateriaL ................................................................. 21

References .................................................................................... 21

3 GEOMETRY AND SOURCE OF THE GP41 N-TRIMER STERIC

BLOCK ............................................ , ......................................... 23

Materials and Methods ................................................................. 25

Reagents .. "., ................. ,' ................................................... 25

Protein Expression and Purification ........................................... 25

Expression and purification of C-peptide inhibitors ....... " .... 25

Control inhibitor ...................................................... 26

Expression and purification of soluble CD4012 (sCD4) .. 26

Viral Infectivity Assays .......................................................... 27

Standard assay ........................................................... 27

sCD4 assisted fusion assay ........................................... 27

Infectivity data analysis ............................................... 28

Biacore SPR Analysis ........................................................... 28

Inhibitor Degradation and Aggregation Assay .............................. 29

Results ....................................................................................... 29

N-trimer Steric Block Is Asymmetric ......................................... 29

Viral Components Are the Major Source of the Steric Block ........... 30

C-terminal Fusion to C37 Has Little Effect on Binding Affinity

for the N-trimer .................................................................. 35

Fusion Proteins Remain Stable During the Time of Assay ................. 35

Discussion ................................................................................ 36

Future Directions ......................................................................... 38

gp120 as the Major Source of the Steric Block .............................. 38

Biochemical confirmation of the sCD4-activated state

steric block .............................................................. 38

Disulfide-trapped fusion .............................................. 38

Env deglycosylation mutants ........................................ 39

Functional Mimicry of the N-trimer Steric Block Using Designed

Antigens ........................................................................... 40

References .................................................................................. 44

VIll

LIST OF TABLES

2.1 The mechanical properties of HIV particles .......................................... 20

3.1 ICso (in nM) of fusion proteins in standard and sCD4 assisted

infectivity assays .............. " ........................... 0 •••••••••••••••••••••••••••••• 32

LIST OF FIGURES

Figure Page

1.1 Schematic models for HIV ................................................................ 2

1.2 Crystal structure of HIV -1 gp41 core and schematic representation

of gp41 protein .............................................................................. 7

1.3 Model of HIV membrane fusion .......................................................... 8

1.4 Inhibitory activity for C37 and cargo-C37 proteins .................................. 11

2.1 Schematic representation of HIV Gag protein don1ains .............................. 1 7

2.2 HIV particle shape and dimensions .................................................... 18

2.3 Measuring the point stiffness of the virus by indentation type

experiments ................................................................................. 19

2.4 Averaged point stiffness of HIV virus ................................................. 19

2.5 Viral reporter particle entry assay ...................................................... 20

3.1 Exploring the geometry of the steric block ............................................ 24

3.2 Inhibitory activity of cargo-C37 and C37-cargo proteins in standard

assay .......................................................................................... 31

3.3 Inhibitory activity of cargo-C3 7 and C3 7 -cargo proteins in sCD4

assisted fusion assay ...................................................................... 34

3.4 Schen1atic representation of IZN36-DSL20 crossliked to gp 120 via

SOS-like disulfide bond ................................................................. 43

CHAPTER!

INTRODUCTION

AIDS Epidemic

An estimated 39.5 million people are currently living with acquired

inlmunodeficiency syndrome (AIDS) worldwide. There were 4.3 million new cases and

2.9 million deaths reported in 2006 alone [1]. These statistics indicate that despite more

than 20 years of intense scientific investigation, the disease is still not well controlled.

AIDS is caused by human immunodeficiency virus (HIV). Current reconlmended AIDS

treatment, known as highly active antiretroviral therapy (HAART), uses a "cocktail" of

antiHIV drugs to target multiple events in the HIV lifecyc1e [2]. However, given the high

mutation rates of the HIV genome, HIV strains with resistance to all available drugs are

an emerging threat [3, 4]. Even patients whose disease is under control still have to be on

medication for their entire life to suppress the residual virus. Additional research on HIV

and AIDS is urgently needed to develop new drugs for more effective therapy.

HIV Components

The schenlatic structure of HIV is shown in Fig. 1.1. The HIV genome consists of

diploid single-stranded positive sense RNA. CAJp24 protein forms a conical core that

2

A Gag

__ ......... __ ._ -- ~-... -- --- -1

MA CA NC

B Env

Immature Mature

Figure 1.1. Schematic models for HIV. (A) Schematic representation of HIV Gag protein domains. MA (blue); CA (red); NC (green). For simplicity, other Gag domains, such as p2 and p6, that are not structural proteins are not shown. (B) In the immature form, a thick protein shell (~25 nm) composed of Gag is observed beneath the membrane. Viral maturation is induced by the proteolytic processing of Gag into three major structural domains: MA, CA, and NC. The mature HIV particle has a thin protein shell (~5 nm) and conical core. Maturation has no effect on the virus dimensions; both mature and immature particles have a diameter of ~ 100 nm. Env trimers are shown on the surface of the virion and do not undergo proteolytic processing during maturation. The Env ectodomain is in brown, and the transmembrane and CT domains are in tan. Figure taken from [5] with permission.

3

contains the viral genon1e and enzymes such as reverse transcriptase, integrase and

protease. Surrounding the conical core is a protein shell formed by MA/p 17 protein. MA

binds to the inner side of the viral membrane and provides structural support. The only

viral protein on the surface of HIV is Envelope protein (Env). Env has two interacting

subunits gp120 and gp41. Env is a transmembrane protein and interacts with MA via its

cytoplasmic tail (CT) domain [6].

HIV Lifecycle

The lifecycle of HIV can be roughly divided into three stages: entry, replication, and

release. HIV entry is mediated by Env. Env interacts with host cell receptors and triggers

the fusion between the viral and host membranes. Once inside the cell, HIV releases its

genome and enzymes into the cytoplasm. The ssRNA genome of HIV is converted into

dsDNA by HIV reverse transcriptase, and viral DNA then enters the nucleus and is

integrated into the host genome by HIV integrase. After integration, viral genes are

transcribed into RNA and translated into proteins by host cell n1achinery. Viral RNA and

proteins self-assemble at the cell membrane into viral particles and new virus is released

by budding from the cell membrane. Many of the viral proteins are made in the form of

nonfunctional polyprotein precursors and are cleaved into individual proteins by HIV

protease after budding.

Being a critical step in HIV lifecycle and exposing highly conserved sequences, HIV

entry is an ideal drug target. Moreover, preventing entry can literally prevent all

4

downstream infections. However, current antiHIV medications mostly focus on reverse

transcriptase and protease. Of the 29 FDA approved drugs in the treatment of HIV

infection, only one is an entry inhibitor [7]. To delay drug resistance and increase the

potency of current therapy, more research is needed to reveal the mechanisms of HIV

entry and open up opportunities for anti entry drug discovery.

HIV Maturation

After releasing from the host cell and before entering a new cell, HIV and other

retroviruses undergo a maturation process crucial for their infectivity. During maturation,

the viral structural protein Gag/pr55 is cut by HIV protease into three major proteins:

MA/pI7, CA/p24 and NC/p7 (schematics shown in Fig. 1.IA). The proteolytic

maturation is accompanied by dramatic morphological rearrangements inside the virion.

In cryogenic electron microscopy (cryoEM) pictures, immature virions have a thick

spherical protein shell, while mature virions have a significantly thinner protein shell and

a prominent conical core [8]. A schen1atic model of HIV maturation has been predicted

based on biochemical and EM studies (Fig. I.IB). In immature virions, uncut Gag protein

constitutes the thick protein shell. In mature virions, MA protein remains associated with

the inner membrane and forms the thin protein shell, while CA protein condenses to form

the conical core. NC protein forms a complex with the RNA genome inside the conical

core.

Although HIV maturation has been extensively characterized and proven to be

S

required for infectivity, the mechanism behind this process remains poorly understood.

Why does HIV assemble as noninfective immature virus and mature only after leaving

the host cell? Given the different demands for virus particle at different stages of lifecycle,

it is likely that by adopting different morphologies, virions can have distinct physical

properties each advantageous in performing certain task. For example, immature virions

may be more durable, allowing more distant spread of infection, while mature virions

may be more fragile and therefore easily deformed during membrane fusion. To

investigate whether there is a linkage between morphology and physical properties in

HIV, we compared the stiffness of HIV particles before and after maturation using

nanoindentation experiments performed by atomic force microscopy (AFM) (Chapter 2).

Env Mediated HIV Entry

HIV entry is mediated by the interaction of HIV Env with host cell receptors (CD4

and a chemokine coreceptor, typically CXCR4 or CCRS). Env is composed of two

noncovalently associated subunits, gp 120 and gp41, which form trimers and are the only

viral proteins exposed on the surface of HIV. gp 120 and gp41 are derived from the

proteolytic cleavage of Env protein gp 160 by a cellular protease. gp 120 recognizes host

cell receptors, while gp41 is anchored in the viral membrane and directly mediates fusion

[9].

A schematic domain map of gp41 is shown in Fig. 1.2A. The gp41 ectodomain

contains two alpha-helical heptad repeat regions that have intrinsic affinity for each other,

6

namely the N- and C-peptide regions. The crystal structure of gp41 in the postfusion

conformation is a six-helix bundle, with trimeric N-peptides in the center and three

C-peptides packed in an anti parallel manner into the hydrophobic grooves on the surface

of N-peptides (Fig. 1.2B) [10]. This structure has been very infoffilative for producing a

model of HIV fusion. The postfusion structure of gp41 closely resembles the postfusion

conformation of influenza virus and moloney murine leukemia virus (Mo-MLV) Envs,

suggesting a common fusion mechanisnl [9].

Our current model of HIV entry was derived based on the gp41 core structure and

other biochemical and structural studies (shown in Fig. 1.3). Initially, gp160 folds into a

poorly understood native conformation. Following proteolytic cleavage, gp41 becomes

metastable since a more stable conformation is now available (6-helix bundle). However,

the association of gp41 with gp120 prevents adoption of this more stable conformation.

The interaction of gp 120 with cellular receptor CD4 induces large conformational

changes in gp 120 that propagate to gp41 via the gp 120/ gp41 interface. These changes

cause gp41 to insert its N-terminal hydrophobic fusion peptide into the target cell

membrane and adopt a conformation called the "prehairpin intermediate." This

intermediate is thought to be long-lived (minutes) and exposes the N- and C-peptide

regions of gp41 in an extended conformation. At this stage, the association of gp 120 with

gp41 prevents the interaction between the N- and C-peptide regions. gp120 then interacts

with a second cellular receptor (chemokine coreceptor CXCR4 or CCR5) that causes

further conformational changes that loosen its association with gp41. This loosening

7

A Heptad repeat 1 Heptad repeat 2

tm

B

C34

Figure 1.2. Crystal structure of HIV-l gp41 core and schematic representation of gp41 protein. (A) Schematic representation of gp41 protein. (B) side view (left) and end view (right) of gp41 core crystal structure. Gray, N36, residues 546-581 of gp160. Black, C34, residues 628-661 of gp 160. fp, fusion peptide. tm, transmembrane domain. cyto, cytoplasmic domain. Figure drawn from the results in [10].

Native cell membrane

gp120 .. gp41

...

viral membrane

.1. 1.1.

~ I ,

Post-fusion

CD4

co-receptor

Prehairpin Intermediate

err I

Cargoes

Hairpin/fusion

8

Figure 1.3. Model of HIV membrane fusion. Formation of the trimer-of-hairpins drives the viral and cellular membranes together, leading to fusion. The N-peptide region (gray), C-peptide region (blue), gp120 (green), gp41 (light blue), gp41 fusion peptide (orange), and transmembrane domain (dark blue) are shown. gp120 is removed from the prehairpin intermediate for clarity. Also shown are a series of C37 fusion proteins of different sizes and an anti-N-trimer Ab attempting to access the N-trimer but potentially blocked by a steric restriction. Sizes of the Ab and fusion proteins are only approximately to scale. Redrawn from [11] with permission .

9

allows the C-peptide region of gp41 to fold back antiparallel to the N-peptide region to

form the "trimer-of-hairpins" (6-helix bundle) structure. Formation of this structure

induces fusion by forcing the viral and host cell membranes together [9].

gp41 N-trimer Is an Attractive Drug Target

As HIV's only extracellular target, Env, especially the N-trimer region of gp41, is an

ideal vaccine and drug target. The gp41 N -trimer forms a critical part of the fusion

machinery and contains highly conserved epitopes including a deep hydrophobic pocket

[10, 12]. Mutations in the pocket region dranlatically lower the efficiency of fusion

[12-17]. Several peptide entry inhibitors derived from the C-peptide region of gp41

("C-peptides") bind to the N-trimer and prevent subsequent interaction of N-trimer with

its endogenous C-peptide binding partner. One of these entry inhibitors (Fuzeon, T-20) is

in current clinical use [18]. Fuzeon, when used with "cocktail" drugs, is effective for

patients who have previously failed other antiHIV medications. However, this drug has

substantial proteolysis problem in vivo and requires frequent injection of large doses at a

high cost. Improved ways of targeting the N-trimer are needed.

gp41 N-trimer Steric Block

Since the gp41 N-trimer is an attractive drug target, extensive efforts have been made

to discover neutralizing antibodies (Abs) against it. A vaccine capable of eliciting broadly

neutralizing Abs is the ultimate strategy for controlling the HIV epidemic. AIDS vaccine

research has been extremely difficult since the virus uses multiple strategies to evade host

10

immunity including glycosylation, steric protection of its surface proteins to avoid

recognition by Abs, and selective down-regulation of MHC-I receptors to disguise

infected cells from NK cells [19, 20]. A very small number of broadly neutralizing Abs

against gp41 and gp 120 have been isolated from HIV infected individuals, but none of

them target the highly conserved gp41 N-trimer region [21]. The N-trimer and N-trimer

pocket region have been authentically mimicked by several available designed peptides

(e.g., 5-helix, IZN36, IQN36, IZNI7, IQNI7) [22-24]. Using these artificial N-trimer

mimics, many groups were able to discover Abs that bind N -trimer mimics with high

affinity and specificity but have very weak or no inhibitory activity in standard viral entry

assays [25-29]. D5, the best Ab of these, mainly binds to the gp41 N-trimer hydrophobic

pocket and surrounding residues and has a mid pM KD for N-trimer mimics but is >

1000x less potent in vivo [29, 30]. A key question was raised by these results: why does

the N-trimer region have such poor accessibility for Abs?

Our lab recently discovered a steric defense around the N-trimer region of HIV that

prevents large proteins (like Abs) from accessing the N-trimer and likely explains the

ineffectiveness of current anti-N-trimer Abs. Briefly, a small C-peptide inhibitor C37 was

fused to various sizes of cargo proteins at its N-terminus via poly-Gly/Ser linker (Fig.

1.3). The potency (ICso) of these inhibitors was tested in a standard viral infectivity assay.

Increasing cargo size decreased the inhibitory potency of these C37 fusion proteins. The

inhibitory potency of a bulky inhibitor was partially restored by extension of the linker

connecting it to C37 (Fig. 1.4). The decrease in potency is not due to weakened affinity

250

200

0 .. 150 C'O ~ 0 It') 100 0

50

0

C37 BPTI-C37

Cargo 0 6.5

kDa

Ub- Mb- GFP- MBP- MBP1- MBP2· C37 C37 C37 C37 C37 C37

8.6 17 27 41 41 41

II Syncytia

I

·OJRFL • HXB2

11

Figure 1.4. Inhibitory activity for C37 and cargo-C37 proteins. Black bars are results from HXB2 syncytia assay (cell-cell fusion). White and gray bars are results from viral assay. All proteins have a 7-residue Gly/Ser linker between C37 and the cargo protein except MBPI-C37 (20-residue) and MBP2-C37 (33-residue). Plotted from the results of [11].

12

for the N-trimer, since all bulky inhibitors showed sin1ilar KDs for an N-trimer mimic to

the original C-peptide by surface plasmon resonance (SPR) binding assay [11]. The

discovery of the N-trimer steric defense can help to improve the search for potent broadly

neutralizing Abs. Sterically restricted N-trimer mimics could be designed and used as

antigens to screen for Abs that can overcome the steric block. Here, as an initial effort

towards this goal, we further investigated the geometry, magnitude and likely source of

the N-trimer steric block (Chapter 3). Infom1ation obtained here will be used to guide

future antigen design.

References

1. UNAIDS/WHO, AIDS epidemic update. http://www.unaids.org, 2007.

2. AIDSinfo, Recommended HIVTreatment Regimens. http://aidsinfo.nih.gov, 2007.

3. HIV Drug Resistance Database. http://hivdb.stanford.edu, 2007.

4. D'Aquila, R.T., et aI., Drug Resistance Mutations in HIV-l. Top HIV Med, 2002.

10(5): p. 21-25.

5. Kol, N., et aI., A stiffness switch in human immunodeficiency virus. Biophys 1,

2007.92(5): p. 1777-83.

6. Coffin, 1.M.H., Stephen H.; Vam1us, Harold E, Retroviruses. 1997: Cold Spring

Harbor Laboratory Press.

7. FDA, HIV and AIDS http://www.fda.gov/oashi/aids/hiv.html, 2007.

8. Briggs, 1.A., et aI., The stoichiometry of Gag protein in HIV-l. Nat Struct Mol

BioI, 2004. 11(7): p. 672-5.

9. Eckert, D.M. and P.S. Kim, Mechanisms of viral membrane fusion and its inhibition. Annu Rev Biochem, 2001. 70: p. 777-810.

13

10. Chan, D.C., et aI., Core structure of gp41 from the HIV envelope glycoprotein. Cell, 1997.89(2): p. 263-73.

11. Hamburger, A.E., et aI., Steric accessibility of the HIV-l gp41 N-trimer region. J

BioI Chern, 2005.280(13): p. 12567-72.

12. Chan, D.C., C.T. Chutkowski, and P.S. Kim, Evidence that a prominent cavity in the coiled coil of HIV type 1 gp41 is an attractive drug target. Proc Nat! Acad Sci

USA, 1998.95(26): p. 15613-7.

13. Cao, J., et ai., Effects of amino acid changes in the extracellular domain of the human immunodeficiency virus type 1 gp41 envelope glycoprotein. J Virol, 1993.

67(5): p. 2747-55.

14. Chen, S.S., et aI., Mutational analysis of the leucine zipper-like motif of the human immunodeficiency virus type 1 envelope transmembrane glycoprotein. J Virol, 1993.67(6): p. 3615-9.

15. Dubay, J.W., et aI., Mutations in the leucine zipper of the human immunodeficiency virus type 1 transmembrane glycoprotein affect fusion and infectivity. J Virol, 1992.66(8): p. 4748-56.

16. Weng, Y. and C.D. Weiss, Mutational analysis of residues in the coiled-coil domain of human immunodeficiency virus type 1 transmembrane protein gp41. J Virol, 1998.72(12): p. 9676-82.

17. Furuta, R.A., et aI., Capture of an early fusion-active conformation of HIV-l gp41. Nat Struct Bioi, 1998.5(4): p. 276-9.

18. Trimeris, http://wwwjuzeon.com. 2007.

19. Cohen, G.B., et aI., The selective down regulation of class I major histocompatibility complex proteins by HIV-l protects HIV-infected cells from NK cells. Immunity, 1999. 10(6): p. 661-71.

20. Wei, X., et aI., Antibody neutralization and escape by HIV-l. Nature, 2003.

422(6929): p. 307-12.

21. McGaughey, G.B., et ai., Progress towards the development of aHIV-l gp41-directed vaccine. CUIT HIV Res, 2004.2(2): p. 193-204.

14

22. Eckert, D.M. and P.S. Kim, Design of potent inhibitors of HIV-l entry from the gp41 N-peptide region. Proc Natl Acad Sci USA, 2001.98(20): p. 11187-92.

23. Root, MJ., M.S. Kay, and P.S. Kim, Protein design of an HIV-l entry inhibitor. Science, 2001. 291(5505): p. 884-8.

24. Eckert, D.M., et aI., Inhibiting HIV-l entry: discovery of D-peptide inhibitors that target the gp41 coiled-coil pocket. Cell, 1999.99(1): p. 103-15.

25. Weiss, C.D., HIV-l gp41. mediator offusion and target for inhibition. AIDS Rev, 2003.5(4): p. 214-21.

26. Golding, H., et aI., Dissection of human immunodeficiency virus type 1 entry with neutralizing antibodies to gp41 fusion intermediates. J Virol, 2002. 76(13): p. 6780-90.

27. Opalka, D., et aI., Analysis of the HIV-l gp41 specific immune response using a multiplexed antibody detection assay. J Immunol Methods, 2004. 287(1-2): p. 49-65.

28. Louis, J.M., et aI., Covalent trimers of the internal N-terminal trimeric coiled-coil of gp41 and antibodies directed against them are potent inhibitors of HIV envelope-mediated cell fusion. J BioI Chern, 2003. 278(22): p. 20278-85.

29. Miller, M.D., et aI., A human monoclonal antibody neutralizes diverse HIV-l isolates by binding a critical gp41 epitope. Proc Natl Acad Sci USA, 2005.

102(41): p. 14759-64.

30. Luftig, M.A., et aI., Structural basis for HIV-l neutralization by a gp41 fusion intermediate-directed antibody. Nat Struct Mol BioI, 2006. 13(8): p. 740-7.

CHAPTER 2

A STIFFNESS SWITCH IN HUMAN IMMUNODEFICIENCY

VIRUS

Reprinted from Biophysical Journal, 2007 March 1; 92(5): 1777-83, with permission by

the Biophysical Society_

Biophysical Journal Volume 92 March 2007 1777-1783 1777

A Stiffness Switch in Human Immunodeficiency Virus

Nitzan Kol: Yu Shi, t Marianna Tsvitov: David Barlam, t Roni Z. Shneck, * Michael S. Kay,§ and Itay Rousso' 'Department of Structural Biology, Weizmann Institute of Science, Rehovot 76100, Israel; tDepartment of Mechanical Engineering and 'Department of Materials Engineering, Ben-Gurion University of the Negev, Beer-Sheva 84105, Israel; and §Department of Biochemistry, University of Utah School of Medicine, Salt Lake City. Utah 84112-5650

ABSTRACT After budding from the cell, human immunodeficiency virus (HIV) and other retrovirus particles undergo a maturation process that is required for their infectivity. During maturation, HIV particles undergo a significant internal morphological reorganization, changing from a roughly spherically symmetric immature particle with a thick protein shell to a mature particle with a thin protein shell and conical core. However, the physical principles underlying viral particle production, maturation, and entry into cells remain poorly understood. Here, using nanoindentation experiments conducted by an atomic force microscope (AFM), we report the mechanical measurements of HIV particles. We find that immature particles are more than 14-fold stiffer than mature particles and that this large difference is primarily mediated by the HIV envelope cy10plasmic tail domain. Finite element simulation shows that for immature virions the average Young's modulus drops more than eightfold when the cytoplasmic tail domain is deleted (930 vs. 115 MPa). We also find a striking correlation between the softening of viruses during maturation and their ability to enter cells, providing the first evidence, to our knowledge, for a prominent role for virus mechanical properties in the infection process. These results show that HIV regulates its mechanical properties at different stages of its life cycle (i.e., stiff during viral budding versus soft during entry) and that this regulation may be important for efficient infectivity. Our report of this maturation-induced "stiffness switch" in HIV establishes the groundwork for mechanistic studies of how retroviral particles can regulate their mechanical properties to affect biological function.

INTRODUCTION

Retroviruses are complex self-assembled structures that are specifically designed to spread infection. The viral Gag protein alone is necessary and sufficient for production of virus-like particles (l). The other major structural protein of human immunodeficiency virus (HIV) particles is the enve­lope glycoprotein (Env, gp 160). Env is synthesized as a precursor that is proteolytically cleaved into two subunits: a receptor-binding subunit (gp 120) and a transmembrane subunit (gp41). The gp120/gp41 complex is required for receptor binding and viral entry (2-5). HN and other len­tiviruses differ from most retroviruses in that they have very long (~150 residue) Env cytoplasmic tails (CT). These IT domains have been shown to interact with the matrix (MA) region of Gag and are important for Env localization to sites of virus budding and efficient Env incorporation into virions (6-9). Mutations within the MA domain (4,8) and deletions in the gp 160 CT domain (6,8,9) block Env incorporation into virions.

After budding from the cell, HIV and other retrovirus particles undergo a maturation process that is required for their infectivity. Virus maturation is induced by the enzy­matic cleavage of the viral Gag protein by virus-encoded protease (PR) into three main structural protellls: MA, capsid

Submitted July 23.2006. and accepted jor publication Novemher 2/.2006.

Addrcss reprint requests to [tay Rousso. Dept. of Structural Bio[ogy, Wcizmann Institute of Science, Rehovot 76[ 00, Israel. Tel.. 972-8-9343479; Fax: 972-8-9344136; E-mai[ : ik1y.rousso@weizmann.ac.i[; or Michael S. Kay, Dept. of Biochemistry, University of Utah School of Medicine, [5 N. Medical Drive East Rm. 4[00, Salt Lake City, UT 84[ [2-5650. Tel.: 801-585-5021; Fax: 801-581-7959; E-mai[: kay@biochem.utah.edu.

© 2007 by the Biophysica[ Society

0006-3495/07/03/1777/07 $2.00

(CA), and nucleocapsid (NC) (10) (Fig. 1 A). Viral matura­tion has been extensively studied using biochemical methods and a vruiety of electron microscopy (EM) imaging tech­niques. During maturation, HN particles undergo a significant internal morphological reorganization, as observed by EM, changing from a roughly spherically symmetric immature particle with a thick protein shell to a mature particle with a thin protein shell and a prominent conical core (\ 1) (sche­matically shown in Fig. 1 B).

Despite substantial progress in morphological and bio­chemical characterization of the virus life cycle, the physical principles underlying virus production, maturation, and entry into cells remain poorly understood. A virion must satisfy several potentially conflicting demands during its lifetime­spontaneous assembly during budding, durability in the out­side environment, and then efficient membrane fusion during entry into the target cell. It is therefore reasonable to spec­ulate that the virus adopts a different set of physical proper­ties at different stages of its life cycle.

In this study we analyze the mechanical properties of HN particles using nanoindentation experiments conducted by an atomic force microscope (AFM). The AFM has been successfully used to measure the mechanical properties of another retrovirus, Moloney murine leukemia virus (ML V) (12), as well as CAs of bacteriophage (13), cowpea chlorotic mottle virus (14), and minute virus (15). We show that the HIV maturation process is accompanied by a dramatic soft­ening of the virion surface. This "stiffness switch" is an ex­ample of a complex macromolecular assembly drastically altering its mechanical properties by spontaneous internal

doi: 10.1 529/biophysj. 106.0939 [4

16

1778

A Gag

-.-.~- . - - - --.

MA CA NC

B Env

Immature Mature

FIGURE I (A) Schematic representation of HIV Gag protein domains. MA (blue); CA (red); NC (green). For simplicity, other Gag domains, such as p2 and p6, that are not structural proteins are not shown. (B) Schematic models for HfY mature and immature states. In the immature form, a thick protein shell (- 25 nm) composed of Gag is observed beneath the membrane. Viral maturation is induced by the proteolytic processing of Gag into three major structural domains: MA, CA, and NC. The mature HfY particle has a thin protein shell (- 5 nm) and conical core. Maturation has no effect on the virus dimensions; both mature and immature particles have a diameter of ----100 run. Env trimers are shown on the surface of the virion and do not undergo proteolytic processing during maturation. The Env ectodomain is in brown, and the transmembrane and CT domains are in tan.

rearrangement. Recently, HIV maturation was shown to affect the ability of virus particles to enter target cells (16,17) using a fluorescence-based assay (18). The entry activity of immature particles is almost lO-fold lower than that of ma­ture particles. Truncation of the viral envelope protein (En v) CT domain restores the entry ability of immature virus par­ticles (16,17). Strikingly, here we show that viral entry activity correlates with its mechanical changes, providing the first evidence, to our knowledge, of a link between mechan­ical and biological properties of a virus. These results show that HIV regulates its mechanical properties at different stages of its life cycle (i.e., stiff during viral budding versus soft during entry), and this regulation may be important for efficient infectivity.

MATERIALS AND METHODS

Virus preparation

Pseudovirion particlcs used in this study were produced by cotransfection of 293T cells with an toEnv HfV - I genome containing an inactivating integrase mutant (DHIV3-GFP-DI16G (19), provided by V. Planelles) and an Env expression vector (pEBB-HXB2 (20), provided by B. Chen). Immature par­ticles were produced by cloning Gag with all PR sites deleted (pNL-MNp6 (17), provided by C. Aiken) into the toEnv, lnt- HIV-I genome. toCT HXB2 Env (to147 (6)) was provided by E. Hunter and cloned into pEBB-HXB2.

Biophysical Journal 92(5) 1777-1783

Kol et al.

Virus particles were collected and purified by centrifugation through a sucrose cushion (20% sucrose in TNE buffer: 0.1 M NaC! , I mM EDTA, JO mM Tris, pH 7.6) at 20,000 X g for 90 min at 4"C. Virus pellets were then resuspended in TNE buffer. During all measurements, virus particles were kept in a physiological buffer (TNE). Purified viruses were attached to microscope glass slides that were pretreated with hexamethyldisilazane (HMDS) vapors. Western blots were developed using sheep polyclonal anti­gpl20 (contributed by M. Phelan. National Institutes of Health AIDS Research and Reference Reagent Program) and rabbit polyclonal anti-CA (provided by W. Sundquist). Blots were quantified using Li-Cor's Odyssey instrument (Lincoln, NE).

Sample preparation for AFM imaging and force measurements

Microscope glass slides were cleaned by boiling in HCI solution. dried. and then incubated overnight in HMDS vapors to enable virus particles attachment (13). Before depositions, purified virus solutions were filtered through a 0.45-JLm filter. A JO'JLI droplet of virus supernatant was then deposited onto a glass slide and left to absorb to the substrate for 15 min. The glass surface was then rinsed with TNE buffer to remove unbound material. All measurements were carried out under TNE buffer.

Virus entry assays

HfV entry assays were performed essentially as described (17). Briefly. HIV particles were incubated with HOS-CD4-CXCR4 cells (provided by Benjamin Chen) for 2 h at 37"C. After rcmoving unbound virus. entry events were detected using CCF2-AM dye (Invitrogen, Carlsbad, CAl as de­scribed in the manufacturer's instructions. Fluorescence was detected using a PolarStar Optima (BMG LABTECH, Offenburg, Germany) plate reader.

AFM imaging and indentation experiments

All AFM experiments were carried out using a Bioscope with a Nanoscope fY controller (Veeco, Santa Barbara, CAl equipped with a dimension XY closed loop scanner mounted on an inverted optical microscope (Axiovert 200M, Carl Zeiss AG, Jena, Germany). Images of virus particles were acquired in AFM tapping mode in a fluid environment and rendered using the WSxM software (Nanotec Electronica, Madrid, Spain, http://www. nanotec.es/progcorn.htm). Pyramidal silicon nitride probes (with a measured averaged stiffness of 0.22 N/m (DNP, Veeco) or 1.55 N/m (NSC36, Micro­ma.sch, Tallin, Astonia) were used, their spring constants being determined experimentally by measuring the thermal fluctuations of the cantilevers (21). Both probe types have a nominal tip radius of 20 nm. To measure the mechanical properties of an individual virus, an indentation experiment was performed with the microscope operated in the force-distance (FD) mode. Before beginning an indentation experiment, the probe was positioned at the center of the virus surface, and the AFM operation was switched from tap­

ping to contact mode by reducing the driving amplitude to 0 mY. For each virus measurement, -100 FD curves were performed at a scan rate of 0.5 Hz.

Data analysis for calculating the virus point stiffness

To obtain the measured point stiffness of a virus particle from a set of roughly 100 successive FD curves, each curve was shifted, first along the z axis to set the tip-sample contact point to a distance of zero, and then along the y axis to set the deflection in the noncontact region to zero. We further analyzed each experiment by plotting the individual measured point stiffness as a histogram and as a function of the measurement count (see Fig. 3 8). Virus measured stiffness (k",e,,) was derived mathematically from the slope

17

A Stiffness Switch in HIV

of the FD curve. A linear function was fitted to the upper 75% of the FO curve (scc Fig. 3 A). Virus particles whose point stiffness values decreased consistently during experimentation were discarded, since they underwent irreversible defonnation, probably due to fatigue or even breakage. Ncxt, a maximal deflection threshold value was set. Curvcs failing to reach this value were discarded, and the remaining aligned curves were averaged. The averaged FO curves were then converted from deflection units (V) to loading force (N) by multiplying by the deflection sensitivity (in nmN , derived from a FO curve perfonned on mica) and the spring constant (N/m) of the cantilever. Virus meaC)ured stiffness (kmeas' in N/m) was derived mathemat­ically from the slope of the averaged FO curve as described above. The stiffness of the virus (k,'ruJ was computed according to Hooke's law on the assumption that our experimental system can be modeled as two springs (the virus and the cantilever) arranged in series:

Oata analysis was carried out using MATLAB software (The Math­Works, Natick, MA). To calculate the Young's modulus from the measured virus stiffness, we utilized the finite element method as previously described (12).

RESULTS AND DISCUSSION

Native virus particles were imaged in a physiological buffer with the AFM operating in tapping mode to minimize possible damage. Under these imaging conditions, vims particles maintained their native dimensions, as determined by their cross sectional profiles. In addition, all of the vims particle types used in this study were found to have similar size (Fig. 2). We analyzed the point stiffness of viral particles by measuring FD curves, as shown in Fig. 3 A, for a mature HIV particle (FD curves for immature HIV particles are available online as Supplementary Material). As seen in Fig. 3 A, even at the maximal loading force, indentation depths are well below 10% ofthe sample thickness-the indentation limit determined by Bueckle's law (22), above which the rigid supporting substrate begins to contribute to the mea­sured stiffness. Low penetration depths are essential for minimizing damage to the virus during the experiment and also ensure that the CA core (in the mature state) will not contribute significantly to the measured stiffness. The measured stiffness values derived from - 100 FD curves are plotted as a histogram to which a Gaussian curve is fitted (Fig. 3 B). During each experiment, the measured stiffness values derived from the individual FD curves were found to distribute normally around a mean without systematic deviation upon repeated measurements (Fig. 3 B, inset), which suggests that the virus did not undergo irreversible deformation during measurement.

The point stiffness measured for mature HIV particles is 0.22 2: 0.0l N/m (n = 37) (Fig. 4). Strikingly, immature particles are more than 14-fold stiffer, with a point stiffness of 3.152: 0.09 N/m (n = 26). To test if Env, the other major stIUctural protein of HIV, might be contributing to the large difference in stiffness between the mature and immature virions, we measured HIV particles lacking Env (~Env). The

10

5

o

5

o

C Mature AEnv

Immature En.

100 150 Virus diameter Inm]

F

1779

10

5

0

5

- '-"""---+ 0 Immature 10

CT

5

100 150

Virus diameter Inm]

FIGURE 2 HIV particle shape and dimensions. (A) An AFM topographic image, acquired in tapping mode, of mature HlV virus in TNE butter (scan area 3 X 3 !Lm, 170 X 170 pixels). Height distributions of mature virus particles (8), mature lacking Env particles (C) , immature virus (D), irrunature lacking Env particles (E), and immature particles lacking the CT domain (P). The corresponding averaged heights in nanometers are 98 (SO = 12, n = 23), 94 (SD = 22, n = 26), WI (SO = 23, n = 23), 98 (SO = 21 , n = 22). Virus size was detennined by height rather than width. because the width of the virus is larger and less accurate due to convolution between the AFM tip and the virus.

presence of Env in mature HIV particles has little effect on stiffness (~Env is 0.21 2: 0.01 N/m (n = 24) vs. 0.22 N/m with Env). In contrast, ~Env immature particles have a dramatically decreased stiffness (0.52 2: 0.02 N/m (n = 23» compared to immature particles with Env (3.15 N/m) (Fig. 4). These ~Env immature particles are still more than twice as stiff as mature particles (with or without Env) but are more than sixfold softer than Env-containing immature viruses. Previous cryo-EM studies of HIV particles show that the protein shell ofthe immature form is nearly five times thicker than the mature form, and this thickness is not affected by the presence of Env (23). Based on EM analysis as well as our previous MLV study (12), we expected that a virus with a thicker protein shell will be stiffer than one with a thinner shell. Our current results indicate that, in fact, the impact of the protein shell thickness (mature ~Env versus immature ~Env) is much smaller than of Env on vims stiffness (immature with versus without Env). Env likely does not have an appreciable impact on mature virus stiffness since

Biophysical Journal 92(5) 1777-1783

18

1780

Z 1.0 .s ~ .2 0>

0.5 ,6 .."

~ .... 0.0

50

~ 25 ::J o U

o

A

0

I / t/'/«: tJ..z ;)} ,!::>'li I .' ?;:,ilJ / ~ ,..,.!~\ ....

0' / ",'

.~0.:s. I ,,;;,,1 ~I 11-' (;'Ii II ,,..-" I .,...p""(

,// 5 10

Z distarlOB [nmJ

01 0.2 Poinl stillness [Nlml

FIGURE 3 Measuring the point stiffness of the virus by indentation type

experiments. (A) Typical force distance curve for a mature virus attached to

HMDS-pretreated glass slide (solid line) and the deflection of the cantilever

(dolled black line). Curves were shifted along the z axis to set the tip-sample

contact point (Zo) to a distance of zero. For each experiment, ---100 curves

were acquired. The virus indentation depth is defined as the difference

between the z position of the virus and cantilever deflection at a given

loading force (labeled as ~Z). (8) A histogram. over which a normal

distribution curve is fitted, of the individual measured point stiffness values

derived from the consecutive force distance curves, The inset shows the

individual measured point stiffness values obtained for a virus during a

single measurement against the experiment number (count). These plots,

together with the observed distribution of the individual measured spring

constants, demonstrate that the virus did not undergo significant irreversible

defonnation during the indentation measurements.

the Gag-Env interaction is broken during maturation by the proteolytic processing of Gag.

To detennine if the HN CT domain mediates the stiff­ening of immature virions, we measured the stiffness of immature particles with Env lacking the CT domain (!lCT). !lCT virions incorporate a nonnal amount of Env because of high surface expression (caused by the loss of an endocytosis signal in the CT domain (24-26)). As seen in Fig. 4, the point stiffness of immature !lCT HIV particles is 0,39 ± 0.01 N/m (n = 22), similar to the !lEnv immature particles (0.52 N/m), Thus, the CT domain appears to be the main contributor to the greatly increased rigidity of the immature state.

The Young's modulus is an inherent material property that in contrast to point stiffness does not depend on the geometry of the sample. Thus, it provides an insight into the average interactions between the building blocks of the virus su-

Biophysical Journal 92(5) 1777-1783

Kol e\ al.

FIGURE 4 Averaged point stiffness of HlV virus. Each value was cal­

culated from the average of ---100 FD curves obtained from individual virus

particles. The bars represent the standard error ofthe mean, and the number

of virions analyzed is indicated by the number shown within each column.

The distribution of the measured point stiffness values as well as the stiffness

as a function of the virus particles' diameter are available online as Sup·

plemcntary Material.

pramolecular shell. To estimate the average Young's moduli (E) of virus particles from the measured virus stiffness, we described the mechanical behavior of the virus as a homo­genous, linear elastic material. Within this framework we have modeled our indentation experiments by using a finite element method as previously described by Kol et al. (12). All virus particle types were modeled as hollow spheres with an outer radius of 50 nm and inner radius of 45 and 25 nm for the mature and immature states, respectively. Virus dimen­sions were adopted from an HIV electron cryomicroscopy study (23). The calculated Young's modulus values are listed in Table I.

Finite element simulation (FES) provides further support for the role of Env CT in stabilizing the immature virus protein shell as indicated by the -8-fold increase in the average Young's modulus when the CT domain is present (lIS vs. 930 MPa).

The number of Env trimers (and therefore, Env CT domains) on the surface of an HIV particle is currently controversial but is thought to be low (7-72 trimers, (27-32)). The dramatic effect these relatively small numbers of Env trimers exert on the global stiffness of an -100 nm viral particle is remarkable. To exclude the possibility that this large and unexpected Env effect is due to overincorporation of En v, we confinned by Western blot analysis that the level of Env incorporation into the pseudotyped virions used in this study is similar or less than that of authentic virions (with Env contained in the viral genome, data not shown).

19

A Stiffness Switch in HIV

TABLE 1 The mechanical properties of HIV particles

Rex' Wall thickness K Estimated E State [nml [nml IN/ml [MPaJ

Mature 50 0.22 440 Immature 50 25 3.16 930 Immature l1ef 50 25 0.39 115

We have recently reported the effect of maturation on the mechanical properties of another retrovirus, ML V (12). Comparison between the properties of the two retroviruses provides two main observations: I), The Young's moduJus of mature HIV particles (440 MPa) is more than twofold lower than mature ML V (1.03 GPa) (12). This result sug­gests that the protein-protein interactions in the ML V mature shell are stronger that those in the HIV mature shell. 2), In sharp contrast to the dramatic stiffness switch observed here with HIV, MLV particles undergo a much more subtle (-2-fold) decrease in stiffness during maturation . The mature state of HIV and ML V virions have a very similar stiffness, whereas immature HIV is - 5-fold stiffer than immature MLV. Additionally, in our previous MLV mechanical anal­ysis (12), we find that the Young's modulus of the mature state is -4-fold higher than the immature state. Here we find that the Young's modulus of the HIV mature state is more than twofold lower than the immature state. Interestingly, the changes in MLV stiffness and Young' s modulus with mat­uration are quite similar to those we observe between mv mature and DoCT immature virions. The difference in stiff­ness between the HIV and ML V immature forms can be explained by the fact that ML V is not a lentivirus and thus does not possess a long CT domain. Altematively, the dif­ference may be rationalized by the presence of a poorly ordered layer in the ML V immature shell, the pp 12 domain (33), localized between the MA and CA domains, which is likely to destabilize the shell of the ML V immature state. Such a poorly ordered domain is not present in HIV Gag.

The dramatic effect of Env on the virus stiffness may be explained by the following two possibilities: a) , Interactions between Env (via CT) and Gag (via MA) at positions on the virus protein shell propagate throughout the Gag layer to stabilize the entire shell; and alternatively b), the assembly of Gag proteins during viral formation may depend on Env CT. In the absence of Env CT, Gag proteins may self-assemble into a different structural arrangement that is less stable than the organization of Gag formed when Env CT is present, which is manifested as decreased stiffness.

To correlate virus mechanical properties and biological activity, we measured the entry activity of virus particles used in this work using a fluorescence-based assay (18). In agreement with previously reported results (17,34), we find that immature virus particles enter target cells very ineffi­ciently' but truncation of the Env CT domain rescues their entry ability (Fig. 5). These results correlate well with the stiffness of the virus shell . Soft mature virus particles can

1781

FIGURE 5 Viral reporter particle entry assay. Viruses are packaged with a Vpr-/3-lactamase fu sion protein. HOS-CD4-CXCR4 cells are loaded with GeneBlazer dye (Invitrogen), which is cleaved by /3-lactamase upon viral entry (causing a green to blue shift in fluorescence). Blue cells indicate entry cvents. Shown arc fluorescence micrographs of (A) uninfected cell control, (B) immature virions with wild-type Env, and (C) immature virions with l1CT Env . The results are representative of at least two independent experiments.

enter cells efficiently, whereas the stiffer immature form cannot. Removal of the Env CT domain in the immature virus dramatically softens the virus shell and restores the entry activity of these particles.

These results show that the very high stiffness of immature HIV particles depends on the presence of the Env CT domain, whereas the thickness of the protein shell plays a

Biophysical Journal 92(5) 1777-1783

20

1782

much smaller role. Intriguingly, we find a strong correlation between virus stiffness and its ability to enter target cells. Recently, we provided evidence that mature ML V virus particles have an elastic and brittle shell and postulated that this shell undergoes deformation during fusion (12). Based on our results, we speculate that immature virions cannot enter cells efficiently because their shell is too stiff to easily undergo deformation. The entry ability of virus particles will likely also depend on additional factors, such as changes in the conformation of Env related to its CT domain ("inside­out" signaling (16,17)) and the lateral diffusion of Env trimers in the membrane.

In summary, our discovery of an Env-mediated stiffness switch that correlates with viral entry activity provides, to our knowledge, the first evidence for a possible role of virus mechanical properties in the infection process. This work establishes the groundwork for future mechanistic studies on virus self-assembly and, more generally, how biological sys­tems regulate their mechanical propelties, as well as how this regulation can be employed to control biological function.

SUPPLEMENTARY MATERIAL

An online supplement to this article can be found by visiting BJ Online at http://www.biophysj.org.

I.R. is the incumbent of the Robert Edwards and Roselyn Rich Manson Career Development Chair. We thank C. Aiken. D. Eckert, C. Hill, W. Sundquist, and S. Weiner for discussions and critical review of the

manuscript.

This work was supported in part by the Jean-Jacques Brunschwig Fund for the Molecular Genetics of Cancer (I.R.). the Kimmelman Center for Macromo­lecular Assemblies (J.R.), and the National Institutes of Health (M.K.).

REFERENCES

I. Wills, J. W., and R. C. Craven. 1991. Form, function, and use of retroviral Gag proteins. AIDS. 5:639-{}54.

2. Berger, E. A., P. M. Murphy, and J. M. Farber. 1999. Chemokine receptors as HIV~l coreceptors: roles in viral entry, tropism, and disease. Anna. Rev. Immunal. 17:657-700.

3. Chan, D. c., and P. S. Kim. 1998. HIV entry and its inhibition. Cell. 93 :681-{}84.

4. Freed, E. 0., and M. A. Martin. 1995. Virion incorporation of envelope glycoproteins with long but not short cytoplasmic tails is blocked by specific, single amino acid substitutions in the human immunodefi­ciency virus type I matrix. 1. Viral. 69:1984--1989.

5. Moore, J. P., and J. Binley. 1998. HIV. Envelope's letters boxed into shape. Nature. 393:630-{}31.

6. Dubay, J. W. , S. J. Roberts, B. H. Hahn, and E. Hunter. 1992 Truncation of the human immunodeficiency virus type 1 transmem­brane glycoprotein cytoplasmic domain blocks virus infcctivity. 1. Viral. 66:6616-{}62S.

7. Freed, E. 0., and M. A. Martin. 1995. The role of human Immuno­deficiency virus type 1 envelope glycoprotcins in virus infection. 1. Bioi ChCln. 270:23883-23886.

8. Freed, E. 0., and M. A. Martin. 1996. Domains of the human Immunodeficiency virus type I matrix and gp41 cytoplasmic tail re­quired for envelope incorporation into virions. 1. Viral. 70:341-351.

Biophysical Journal 92(5) 1777-1783

Kol et at.

9. Spies, C. P., G. D. Ritter Jr., M. J. Mulligan, and R. W. Compans. 1994. Truncation of the cytoplasmic domain of the simian immuno­deficiency virus envelope glycoprotein alters the conformation of the external domain. 1. Virol. 68:585-591.

10. Swanstrom, R., and J. W. Wills. 1997. Synthesis, assembly, and processing of viral proteins. In Retroviruses. J. M. Coffin, S. H. Hughes, and H. E. Varmus, editors. Cold Spring Harbor Laboratory Press, Plainview, NY. 263-334.

II. Coffin, J. M., S. H. Hughes, and H. E. Varmus, editors. 1997. Retroviruses . Cold Spring Harbor Laboratory Press, Plainview, NY.

12. Kol, N., M. Gladnikoff, D. Barlam, R. Z. Shncek, A. Rein, and I. Rousso. 2006. Mechanical properties of murine leukemia virus particles. Effect of maturation. Biophys. 1. 91 :767-774.

13. Ivanovska, I. L., P. J. de Pablo, B. Ibarra, G. Sgalari, F. C. MacKintosh, J. L. Carrasco sa, C. F. Schmidt, and G. J. Wuite. 2004. Bacteriophage capsids: tough nan as hells with complex clastic properties. Proc. Natl. Acad. Sci. USA. 101:7600--7605.

14. Michel, J. P., I. L. Ivanovska, M. M. Gibbons, W. S. Klug, C. M. Knobler, G. J. Wuite, and C. F. Schmidt. 2006. Nanoindentation studies of full and empty viral capsids and the effects of capsid protein mutations on elasticity and strength. Proc. Natl. Acad. Sci. USA. 103:6184-{}189.

15. Carrasco, c., A. Carreira, I. A. Schaap, P. A. Serena, J. Gomez­Herrero, M. G. Mateu, and P. J. de Pablo. 2006. DNA-mediated anisotropic mechanical reinforcement of a virus. Proc. Natl. Acad. Sci. USA. 103:13706--13711.

16. Murakami, T., S. Ablan, E. O. Freed, and Y. Tanaka. 2004. Rcgulation of human immunodeficiency virus type 1 Env-mediated membrane fusion by viral protease activity. 1. Viral. 78:1026--1031.

17. Wyma, D. J., J. Jiang, J. Shi, J. Zhou, J. E. Lineberger, M. D. Miller, and C. Aiken. 2004. Coupling of human immunodeficiency virus type I fusion to virion maturation: a novel role of the gp41 cytoplasmic tail. 1. Viral. 78:3429-3435.

18. Cavrois, M., C. De Noronha, and W. C. Greene. 2002. A sensitive and specific enzyme-based assay detecting HIV-I virion fusion in primary T lymphocytes. Nat. Biatechnal. 20:1151-1154.

19. Dehart, J. L., J. L. Andersen, E. S. Zimmerman, O. Ardon, D. S. An, J. Blackett, B. Kim, and V. Planelles. 2005. The ataxia telangiectasia­mutated and Rad3-relatcd protein is dispensable for retroviral integra~ tion.l. Viral. 79:1389-1396.

20. Chen, B. K., K. Saksela, R. Andino, and D. Baltimore. 1994. Distinct modes of human immunodeficiency virus type I proviral latency revealed by superinfection of nonproductively infected ccll lines with recombinant luciferase-encoding viruses. 1. Viral. 68:654-{}60.

21. Hutter, J. L., and J. Bechhoefer. 1993. Calibration of atomic-force microscope tips. Rev. Sci. Instrum. 64: 1868-1873.

22. Bueckle, H. 1973. Use of hardness to determine other material prop­erties. In The Science of Hardness Testing and Its Research Applica­tions. J. W. Westbrook and H. Conrad, editors. American Society for Metals, Materials Park, OH.

23. Wilk, T., I. Gross, B. E. Gowen, T. Rutten, F. de Haas, R. Welker, H. G. Krausslich, P. Boulanger, and S. D. Fuller. 2001. Organization of immature human immunodeficiency virus type 1.1. Viral. 75:759-771.

24. Berlioz-Torrent, c., B. L. Shacklett, L. Erdtmann, L. Delarnarre, I. Bouchaert, P. Sonigo, M. C. Dokhelar, and R. Benarous. 1999. Interactions of the cytoplasmic domains of human and simian retroviral transmembrane proteins with components of the clathrin adaptor complexes modulate intracellular and cell surface expression of envelope glycoproteins. 1. Virol. 73: 1350-1361.

25. Boge, M., S. Wyss, J. S. Bonifacino, and M. Thali. 1998. A membrane­proximal tyrosine-based signal mediates internalization of the HIV-I envelope glycoprotein via interaction with the AP-2 clathrin adaptor. 1. BioI. Chern. 273: 15773-15778.

26. Rowell, J. F., P. E. Stanhope, and R. F. Siliciano. 1995. Endocytosis of endogenously synthesized HIV-I envelope protein. Mechanism and role in processing for association with class II MHC. 1. Immunol. 155:473-488.

21

A Stiffness Switch in HIV

27. Chertova, E., 1. W. Bess Jr., B. J. Crise, I. R. Sowder, T. M. Schaden, J. M. Hilburn, J. A. Hoxie, R. E. Benveni ste, J. D. Lifson, L. E. Hcnderson, and L. O. Arthur. 2002. Envelope glycoprotein incorpo­ration, not shedd ing of surfacc envclope glycoprotein (gpI20/SU), is the primary determinant of SU content of purified human immunode­ficicncy vims type I and simian immunodeficiency vims. 1. Virol. 76: 5315-5325.

28. Gelderblom, H. R. 1991. Assembly and morphology of HlY: potcntial effect of structurc on viral function. AIDS. 5:617--{j37.

29. Hockley, D. J., R. D. Wood, J. P. Jacobs, and A. J. Garrett. 1988. Electron microscopy of human immunodeficiency virus. 1. Gen. Viral. 69:2455-2469.

30. Ozel, M., G. Pauli, and H. R. Geldcrblom. 1988. The organization of the envelope projections on the surface of HIV. Arch. Virol. 100: 255-266.

1783

31. Zhu, P., E. Chertova, J. Bess Jr., J. D. Lifson, L. O. Arthur, J. Liu, K. A. Taylor, and K. H. Roux. 2003. Electron tomography analysis of envelope glycoprotein trimers on HlY and simian immunodeficicncy virus virions. Proc. Natl. Acad. Sci. USA. 100:15812-15817.

32. Zhu, P., J. Liu, J. Bess Jr., E. Chertova, 1. D. Lifson, H. Grise, G. A. Ofek, K. A. Taylor, and K. H. Roux. 2006. Distribution and three­dimensional structure of AIDS virus envelope spikes. Nature. 441: 847-852.

33. Yeager, M., E. M. Wilson-Kubalek, S. G. Weiner, P. O. Brown, and A. Rein. 1998. SupramolecuJar organization of immature and matufC murine leukemia virus revealed by electron cryo-microscopy: implications for re­troviral a"cmbty mechanisms. Proc. Natl. Acad. Sci. USA. 95:7299-7304.

34. Jiang, 1., and C. Aikcn. 2006. Maturation of thc viral core enhances the fusion of HIY -I particles with primary human T cells and monocyte­derived macrophages. Virology. 346:460-468.

Biophysical Journal 92(5) 1777-1783

22

CHAPTER 3

GEOMETRY AND SOURCE OF THE GP41

N-TRIMER STERIC BLOCK

We have previously shown that gp41 N-trimer has a steric block that reduces the

potency of bulky inhibitors. In this study, we explored the geometry and magnitude of the

gp41 N-trimer steric block by comparing the potency of bulky inhibitors with cargos

pointing towards the viral vs. cellular membranes (Fig. 3.1). Our results demonstrate that

bulky inhibitors are more potent with cargo pointing to the cell side vs. to the virus side,

indicating a less severe steric block on the cell side of the N -trimer. To infer the source of

the steric block, we utilized a modified sCD4 assisted fusion pathway to remove the

possible steric hindrance from a target cell. In this system, if any steric block is still seen,

it would suggest that the block comes from viral sources. We measured the potency of

bulky inhibitors with cargoes pointing towards both virus and cell sides in sCD4 assisted

fusion system, and found that their potency relative to C37 was similar to that in the

standard system. These results indicate that a steric block still exists on sCD4 activated

virus (in the absence of host cell), which suggests viral component, especially gp120,

likely to be the major source of the steric block. Knowledge about the source,

cell

N-trimer N-trimer

C-term c c

virus Cargoes

cell

virus

Cargoes

Steric block?

.. C-term

N-term

24

Figure 3.1. Exploring the geometry of the steric block. Prehairpin intermediate is shown with an exposed N-trimer. See Figure 1.3 for the color legend. Also shown are a series of C37 fusion proteins of different sizes with cargo proteins pointing at different orientations. ("viral side" and "cellular side" cargos are shown on the left and right panels, respectively). Adapted from [1] with permission.

25

approximate shape and magnitude of the block will aid the rational design of sterically

restricted antigens that mimic the steric environment of the N-trimer on virions.

Materials and Methods

Reagents

pET20b vector (Novagen), pET9a-CD4012 (gift from R. Varadarajan) [2],

pEBB-HXB2 and pEBB-1RFL (gifts from B. Chen) [3], BLR(DE3)pLysS E. coli

(Novagen), BL21-Gold(DE3) E. coli (Stratagene), amylose resin (New England BioLabs),

T7-Tag antibody agarose (Novagen). HisTrap HP (Amersham Biosciences) was used for

all Ni affinity purifications. The National Institutes of Health AIDS Research and

Reference Program provided the following reagents: pNL4-3.Luc.R-E- (N. Landau) [4,

5], HOS-CD4-fusinlCCR5 cells (N. Landau) [6, 7], Cf2Th-CXCR4/synCCR5 cells (T.

Mirzabekov and 1. Sodroski) [8].

Protein Expression and Purification

Expression and purification of C-peptide inhibitors. C37H6, and cargo-C37

proteins were produced as described [1]. For C3 7 -cargo proteins, all proteins were cloned

into the NdeIlXhoI sites of the pET20b vector and expressed in either BLR(DE3)pLysS

or BL21-Gold(DE3) E. coli. The cargo proteins used in this study were human ubiquitin

(Ub) and E. coli maltose-binding protein (MBP). They were cloned onto the C-terminus

of C37 with the following linker in between: (Ser)s(GlY)2. C37-MBPl had a linker

sequence of (Ser)3(GlY)3Ser(GlY)3Ser(GlY)3(Ser)4(GlY)2. All proteins contained an

26

N-terminal T7 tag and a C-terminal His tag (His)6. All proteins were purified with two

steps. MBP-containing proteins were first purified with amylose resin followed by Ni

affinity chromatography. C37-Ub was first purified by Ni affinity chromatography

followed by reverse phase HPLC with a C 18 column. Purified proteins were lyophilized

for long-term storage. C37-MBP and C37-MBPI were shown to be >90% pure via

SDS-PAGE and coomassie blue staining (with the major contaminant identified via mass

spectrometry as a 20-amino acid N-terminal truncation of C37 which was present during

protein expression). C37-Ub was shown to be >980/0 pure via SDS-PAGE and coon1assie

blue staining. Protein masses were confirmed by e1ectrospray mass spectrometry

(University of Utah Core Facility). Working protein stocks were prepared by dissolving

lyophilized protein in PBS (Gibco) and centrifuging at 18,000 g for 10 minutes to remove

any aggregates. These stocks were aliquoted and flash frozen. For each experiment, a

fresh aliquot was thawed and centrifuged and the protein concentration was determined

via absorbance at 280 nm.

Control inhibitor. To verify that the T7 tag on the N-terminus of C3 7 had no effect

on the antiviral inhibitory activity, a control C37 with an N-terminal T7 tag was made and

compared to our standard C37H6 control peptide. No difference in inhibitory activity was

seen (data not shown), so we continued to use C37H6 as the control inhibitor.

Expression and purification of soluble CD4D12 (sCD4). sCD4 was expressed and

purified as previously described [2]. Briefly, pET9a-CD4DI2 was transformed into E. coli

strain BL21 (DE3). The culture was grown in 30 ug/ml Kanamycin and was induced with

27

80 mg/l IPTG during mid-log growth. After cell lysis, inclusion bodies were resuspended

in 6 M guanidine hydrochloride (GdnCI) and centrifuged at 9,000g for 30 minutes at 4°C.

The supernatant were subject to N i affinity purifications. Protein was > 90% pure by

SDS-polyacrylamide gel electrophoresis (PAGE).

Viral Infectivity Assays

Standard assay. Viral infectivity was measured using a luciferase-based assay

essentially as described [1], with the following modifications. During viral production,

viral supernatants were collected 30 and 36 hours after transfection and were filtered

through a 0.45um membrane. During infectivity assay, virus and inhibitor were removed

20-24 hours after infection.

sCD4 assisted fusion assay. sCD4 assisted fusion assay was performed in a similar

way to the standard assay with a few modifications. Pseudovirions were produced as

described above. For HXB2, virus and inhibitors were added to Cf2Th-CXCR4 cells in

96 well plates in the presence of 50nM sCD4 (co-incubation). For JRFL, pseudovirions

were incubated with inhibitors in the presence of 200nM sCD4 for 2 hours at 37°C before

adding to Cf2Th-synCCR5 cells (preincubation). All assays included 4ug/ml

DEAE-dextran to enhance fusion and 0.5% PBS (Gibco). Cells in 96 well plates were

centrifuged at 2,600 rpm at 37°C for 30 minutes using a Sorvall RT6000c centrifuge to

promote viral attachment. After spinoculation, cells were incubated at 37°C. Media

change and assay development were performed as previously described [I].

28

Infectivity data analysis. ICsos for infectivity assays were calculated by fitting

data to Langmuir equation essentially as previously described except for the following

changes [1]. Curve fit is weighed by the standard error of each point (minimum standard

err is set to 1 %). For standard assay, k is fixed to 1. For sCD4 assisted fusion assay, k is

set to the highest normalized luciferase value to deal with the overshoot effect, which is

explained in more detail below.

In some inhibition curves, infectivity above 1000/0 is observed for low concentration

inhibitor points. Using a fixed zero inhibitor point results in poor fitting of the inhibition

curves. The overshoot is especially severe for sCD4 preincubation assay. The overshoots

also vary among different inhibitors and seem to be related to media composition as well

(e.g., more PBS in media seemed to amplify the overshoot). A possible explanation for

the overshoot is that C-peptide inhibitors can preserve the durability of activated virus,

presumably by stabilizing the N-trimer and prevent aggregation. As a result, some virus

that would otherwise die ends up fusing with cells once inhibitor dissociates. We have not

been able to avoid overshoot problem in our system.

Biacore SPR Analysis

To compare the KDs of C37-cargo proteins and C37H6, SPR binding experiments

were performed as previously described [1].

29

Inhibitor Degradation and Aggregation Assay

To check for aggregation of the C37 fusion inhibitors (which could account for the

reduced antiviral potency), the inhibitors were incubated overnight in a 96-well plate with

HOS-CD4-fusin cells in standard tissue culture medium (DMEM + 100/0 FCS) to

simulate the conditions of the antiviral assay. The mediun1 was analyzed by Western blot

with an antiHis antibody both before and after centrifugation at 18,000 g for 10 minutes.

No pellet was seen after centrifugation and no difference was seen in the Western blot

signal between medium before and after centrifugation.

To check for degradation of the C37 fusion inhibitors and the release of free C3 7, the

inhibitors were incubated in a 96-well plate as described above. Afterwards, the inhibitors

were purified with T7-Tag antibody agarose and analyzed via SDS-PAGE and silver stain.

T7-tagged C37 was used as a positive control for purification and silver staining. No

degradation was visualized in the assay, which allowed us to define a maximum possible

degradation of <1 %.

Results

N-trimer Steric Block Is Asymmetric

The previously published studies used cargos attached to the N-terminus of C37

(viral-side cargo), named cargo-C3 7. In this study, bulky inhibitors were constructed by

fusing different sized proteins onto the C-terminus of C37 (cellular-side cargo) via a

7-residue Gly/Ser linker, named C37-cargo. We tested the potency of these inhibitors in

30

our standard fusion assay to detennine the magnitude of the steric block on the cellular

side of the N -trimer.

For both HXB2 and JRFL viruses, C37-cargo proteins showed decreasing potency

with increasing cargo sizes in a trend similar to that of cargo-C37 proteins (Fig. 3.2A) [1].

Interestingly, C37-cargo proteins seemed to access the N-trimer better than cargo-C37

proteins. C37-MBP was ~2.5x more potent than MBP-C37 against HXB2, and was ~4x

more potent against JRFL (Fig. 3.2B, Table 3.1). This result suggests that the cellular side

of the N-trimer also has a steric block, but may be less severe than the viral side block

(i.e., the steric block is asymmetric). Unexpectedly, unlike MBP-C37, an extended

flexible linker between C37 and a cell-side cargo did not restore the potency as

effectively. C37-MBPI has slightly improved IC50 over C37-MBP in HXB2 strain, and

has no improvement in JRFL strain.

Viral Components Are the Major Source of the Steric Block

The source of the N -trimer steric block was investigated using a modified fusion

assay. In our standard assay, CD4 and coreceptor are both supplied by target cells. In a

modified sCD4 assisted assay, only coreceptor is present on the target cell, and CD4 is

provided in trans by sCD4. Incubation of virus with sCD4 in the absence of target cell or

coreceptor was shown to expose the N-trimer of gp41 in both CXCR4 and CCR5 tropic

HIV-l strains [9-14]. sCD4 assisted fusion provides a useful tool to analyze early entry

events in the absence of target cell. In the sCD4 assisted assay, inhibitors were able to

A ~ 1 --"'-C37-Ub ...... . s; --'-C37H6 :;:::;

()

<C 0.8 .~-.-- C37 -MBP Q) en ~ 0.6 ~

'(3 :::J

...J 0.4 "'0 Q)

.~ ctl 0.2 E '-

o l 0 z

0 200 400 600 800 1000 concentration (nM)

B Ratio of IC50s of C37 -cargo and cargo-C37

140

~120 o -100 (.0

::I: ...... 80 M

o 60 o ..... o 40 .. C\1 c:: 20

o

in Standard Assay

C37H6 C37·Ub C37- C37- Ub·C37 MBP· MBP1· MBP MBP1 C37 C37

~HXB2

.JRFL

31

Figure 3.2. Inhibitory activity of cargo-C37 and C37-cargo proteins in standard assay. Data represent averages of at least quadruplicate measurements in at least two independent assays. (A) Inhibition curve for cargo-C37 proteins, HXB2. (B) Calculated ICso ratios of cargo-C37 and C37-cargo proteins relative to C37H6.

Tab

le 3

.1

ICso

(in

nM

) o

f fus

ion

prot

eins

in

stan

dard

and

sC

D4

assi

sted

infe

ctiv

ity

assa

ys

Fus

ion

Sta

ndar

d S

tand

ard

Sta

ndar

d sC

D4

sCD

4 sC

D4

part

ner

assa

y IC

so

Sta

ndar

d as

say

ICso

P

rote

in

mol

ecul

ar

assa

y as

say

assa

y IC

so

assa

y as

say

infe

ctiv

ity

rati

o in

fect

ivit

y in

fect

ivit

y ra

tio

infe

ctiv

ity

rati

o JR

FL

m

ass

HX

B2

HX

B2

JRF

L

HX

B2

HX

B2

JRF

L

C37

H6

0 1.

3 1.

0 20

.8

1.0

0.31

1.

0 2.

8 C

37-U

b 8.

6 3.

4 2.

5 77

.5

3.7

0.78

2.

6 14

.1

C37

-MB

P

41.0

46

.3

34.6

53

6.0

25.8

7.

2 23

.6

71.0

C

37-M

BP

1 41

.0

30.0

22

.4

515.

3 24

.8

5.1

16.6

84

.5

Ub-

C37

8.

6 4.

5 3.

3 38

.5

1.9

0.84

2.

8 7.

4

MB

P-C

37

41.0

11

6.5

86.9

22

49.5

10

8.4

19.2

62

.9

204.

8

MB

P1-

C37

41

.0

39.0

29

.1

1034

.0

49.8

9.

1 29

.8

95.6

Not

e: I

Cso

rat

ios

are

rela

tive

to

C37

H6.

sCD

4 as

say

ICso

ra

tio

JRF

L

1.0

5.1

25.7

30

.6

2.7

74.2

34

.6

w

tv

33

bind exposed N-trimer on sCD4-activated virus before viral attachment and fusion with

target cells.

The effect of sCD4 assisted fusion on the potency of bulky cargo-C37 inhibitors was

tested using two protocols, coincubation for HXB2 and preincubation for JRFL. Ideally,

we want inhibitors to bind virus that are synchronized to sCD4-activated state before

proceeding to fusion. But for both HXB2 and JRFL strains, sCD4 exposure causes a

time-dependent decrease in viral titer. The decay of sCD4-activated virus is especially

fast for HXB2. For JRFL, we were able to do a 2-hour sCD4 preincubation and get

reasonable titer, but we were not able to use preincubation for HXB2.

All C-peptide inhibitors tested, including C37H6, showed improved potency in sCD4

assisted assay compare to the standard assay (~4-6x for HXB2, llx for JRFL),

suggesting a longer kinetic window of N -trimer exposure (Table 1). In JRFL, the potency

increase correlates positively with the length of sCD4 preincubation, and reaches

maximum at 2-3 hours (data not shown). This result confinns the exposure of N-trimer

on virus by incubation with sCD4. Aside from the kinetic effect, the potency of

cargo-C37 inhibitors still correlates negatively with the size of cargoes (Fig. 3.3A). This

result demonstrates that a sterk block still exists on sCD4-activated virus.

The cellular side sterk block was also examined to characterize the geometry of the

block on sCD4-activated virus. Inhibitory potency of C3 7 -cargo proteins was n1easured

using sCD4 assisted assay and compared to cargo-C37 proteins. The negative correlation

between cargo size and potency is still present, demonstrating that a steric block also

A

B

90

o 80 r..n (.) 70

~ 60 {:;; 50 (.) 40 o .... tn 30 o ;; 20 ctS

0::: 10

o

cargo-C37 ----C37 -Ub-C37 -..• ,~,., MBP-C37

o 500 1000 1500 2000 2500 3000 3500

concentration (nM)

----C37 -C37-Ub

C37-cargo

· ...... -C37-MBP

• o 200 400 600 800 1000 1200 1400 1600

concentration (nM)

Ratio of IC50s of C37 .. cargo and cargo .. C37 in sCD4 Assisted Assay

C37H6 C37 -Ub C37· MBP

C37-MBP1

Ub-C37 MBP­C37

MBP1· C37

~HXB21 .... IRFL

34

Figure 3.3. Inhibitory activity of cargo-C37 and C37-cargo proteins in sCD4 assisted fusion assay. (A) Inhibition curve for cargo-C37 proteins (left) and C37-cargo proteins (right), JRFL. (B) Calculated IC50 ratios of cargo-C37 and C37-cargo proteins relative to

C37H6.

35

exists on the cell side of the N-timer (Fig. 3.3A). The magnitude of the block is also

smaller on the cell side, as C37-MBP is ,....,3x more potent than MBP-C37 against both

HXB2 and JRFL (Fig. 3.3B). Also, the IC50 ratios of cargo-C37 inhibitors relative to

C37H6 in sCD4 assisted assay remained in the same range to those in the standard assay

(Fig. 3.2B & 3.3B). Taken together, these data show that sCD4-activated virus has a

steric block of comparable geometry and magnitude to host cell CD4-activated virus,

implicating viral components are the major source of the steric block.

C-terminal Fusion to C37 Has Little Effect

on Binding Affinity for the N-trimer

To exclude the possibility that the linker or cargo on C37 fusion proteins affect its

ability to bind the N-trimer, an SPR binding assay with an N-trimer mimic IZN36 was

performed for C37-cargo proteins. Similar to previous findings with cargo-C37 proteins,

the affinity of C37-cargo proteins showed only small deviations from that of C37H6

(within 3-fold, data not shown), which may account for the loss of potency of C37-Ub

but not C37-MBP [1]. These results showed that the difference in potency between

C37H6 and bulky C-peptide inhibitors is mainly due to steric hindrance, not differences in

binding affinity.

Fusion Proteins Remain Stable During the Time of Assay

To control for the stability of C37H6 and C37-cargo proteins, each of these proteins

was incubated under simulated assay conditions before subject to degradation and

36

precipitation analysis. cargo-C37 proteins were previously shown to have very little

degradation «20/0) and no precipitation after incubation [1]. Analysis of C37-cargo

proteins also showed very little degradation «10/0) and no noticeable precipitation over

the incubation period (data not shown). These results suggest that proteolysis or

aggregation does not account for the loss of potency of bulky C-peptide inhibitors.

Discussion

The way virus attaches to the host cell (inserting N -terminal fusion peptide into the

cell membrane) places the N-trin1er closer to the cell membrane than viral membrane. So,

one might expect that the cell side of the N-trimer should have a more severe block. On

the contrary, bulky C-peptide inhibitors with cellular-side cargo are more potent than

those with viral-side cargo. These results suggest that the viral side has more obstruction

than just the viral membrane itself, probably also from the protein around the N-trimer.

To better elucidate the possibility of viral source of the block, we performed sCD4

assisted fusion.

In sCD4 preincubation, all the activated virus were allowed to bind with inhibitors

for 2 hours before encountering host cells, giving inhibitors maximum chance to

overcome the steric block. In sCD4 coincubation, activated virus could either bind

inhibitor first or fuse with host cell before meeting inhibitor. In this case, only a fraction

of virus has the chance to be inhibited. Given the big mismatch in ICso and Ko for bulky

C-peptide inhibitors, even partial release of the steric block could be detected in either

37

incubation setting. However, no improvement in potency was observed. So we conclude

that the steric block still exists in the absence of host cell, whkh means the sterk block

likely comes from virus itself. On sCD4-activated virus, the only components around the

gp41 N-trimer are gp120 and sCD4 that binds gp120. At this stage, gp120 remains

associated with gp41 at disulfide loop (DSL) and C-peptide regions that are adjacent to

the N-trimer (Sunghwan Kim et aI., submitted), making gp120 the best candidate for

creating the steric hindrance.

A caveat of this study is that the efficiency of sCD4 activation on virus is still

unknown. So we do not know if our conclusion is based on the behavior of the majority

of virus or only a small fraction.

Learning the source, symmetry and magnitude of the N-trimer sterk block can help

us to better depict the fusion pathway. More importantly, it also establishes the

groundwork for discovering ways to overcome the sterk block. Our results showed that

access to the N-trimer on a virus is sterically protected fron1 both the viral and cellular

membrane side, with a more severe block on the viral membrane side. Current N -trimer

mimics have a sterically open environment, making it difficult to enrich for Abs with

special penetrating shapes. One possible solution is to make N-trimer n1imics with

sterically blocked environment similar to that found in vivo. Such mimics can be used to

screen and enrich for Abs that overcome the sterk block.

38

Future Directions

In this study, we have characterized the sterk block on the gp41 N-trimer region of

HIV-l. Our results demonstrated that whether virus is attached to a target cell or is

activated by sCD4 only, there is a steric block on both sides of the N-trimer. A more

severe block was seen on the viral membrane side of the N-trimer. Together with

biochemical studies on the interaction of gp 120 and gp41, these results indicate that

gp 120 is likely to be the major source of the block. Future studies will aim at further

demonstrating the role of gp 120 in the steric block, and applying this information to the

antigen design ofN-trimer mimics.

gp 120 as the Major Source of the Steric Block

Biochemical confirmation of the sCD4-activated state steric block. F or both

HXB2 and JRFL, we will measure the sterk accessibility of the N-trimer on

sCD4-activated virus using a similar protocol as described[9, 15]. The steric accessibility

of the sCD4-activated state will be investigated using the bulky inhibitors described

above (cargo-C37, C37-cargo). If bulky inhibitors bind poorly to sCD4-activated virions

compared to their non bulky controls, the presence of the sterk block in the

sCD4-activated state will be confirmed.

Disulfide-trapped fusion. A modified fusion pathway will be used to probe the

correlation of gp120 conformational change with the presence of the sterk block.

Disulfide-trapped fusion uses an engineered gp 120 that is covalently linked to gp41

39

via a labile disulfide bond (SOS Env) [16]. Pseudovirions with this engineered disulfide

get trapped at a late stage of fusion after coreceptor binding but before membrane fusion

due to the inability of gp120 to dissociate from gp41. Adding low concentration of DTT

can reduce the disulfide bond and resume fusion [17, 18]. Using this systen1, a previous

study demonstrated that that the N-trimer is lOX more accessible to a C-peptide inhibitor

T-20 after bond reduction compare to before bond reduction [18].

We plan to use disulfide-trapped fusion to compare the potency of our bulky

inhibitors before and after bond reduction as described in the previous study with T-20.

The IC50s of the bulky inhibitors will be calculated as ratios relative to C37H6 in order to

normalize any kinetic factors. If we can show restored potency of bulky inhibitors

during/after disulfide reduction but not before, it will further suggest gp 120 to be the

most likely cause of the steric block before it loosens from gp41.

Env deglycosylation mutants. Since viral components are inferred as n1ajOr

source of the steric block, Env modifications will be performed to further dissect how

Env contributes to the steric block. Extensive glycosylation of Env has been shown to

shield antigenic interfaces from Abs [19-21].

To investigate the general role of glycosylation on the N-trimer accessibility,

enzymatic deglycosylation of Env on the viral surface will be performed using

glycosidases. If deglycosylated virus shows more sensitivity to neutralization by bulky

inhibitors but similar sensitivity by C37, it would indicate release of the sterk block.

Single glycosylation sites could then be rationally mutated to identify specific residues

40

contributing to the block (with a focus on glycosylation sites near the gp41 N-trimer

region, including the gp41 DSL and C-petide regions, as well as gp 120 residues near the

gp41/gp120 interface).

Functional Mimicry of the N-trimer Steric Block

Using Designed Antigens

By designing mimics of the N-trimer region with artificial steric blocks, we can test

our understanding about the overall size and geometry of the block and screen for

antibodies that may overcome the block. Even an approximate functional mimic will still

be very useful to screen and enrich for neutralizing Abs in future studies. Structural

characterization of even one such Ab would be invaluable for designing improved

antigens. Our ultimate goal of this work is to create an authentic mimic of the N-trimer

steric environment on the virus. This goal may not be achievable, but the importance of

the problem demands our attempt.

We have accumulated a test set of reagents to test the steric properties of our

artificially designed antigens, including cargo-C37 and C37-cargo fusions, as well as the

weakly neutralizing anti-N-trimer Ab D5 IgG and affinity matured ScFvs. We will use

this test set to challenge how well each iteration of artificial antigen mimics the

N-trin1er's natural steric block. For the ideal antigen, the measured ICsos of our set of

bulky inhibitors and Abs would now agree with their KDs measured against our designed

sterically blocked antigen. Information on the geometry and magnitude of the block will

41

help shape the optimization of our designed antigens.

Our basic strategy is to attach different sized proteins or other bulky materials to the

N- and/or C-terminus of existing N-trimer peptide mimics to create artificial steric blocks

of varying stringencies. These bulky agents can be proteins, inert chemical groups,

liposomes, or beads. We will take two general approaches to this problem: (1) designing

artificial mimics of the N-trimer sterk environment by adding artificial bulky materials to

existing N-trimer mimics and (2) producing mimics of the N-trimer steric environment

using gp 120.

For the first approach we have made a crude first generation mimic, IZN36-MBP,

containing the full N-trimer and steric bulk (MBP) on the C-terminal (viral) side. We

tested the artificial sterk block on this mimic using coprecipitation studies with our

cargo-C37 series of bulky proteins. Our preliminary results show that this crude mimic

already displays the qualitative features of the viral-side N-trimer steric block. We will

further evaluate this mimic using other members of our bulky test set and make further

optimizations.

For the second approach, we can apply the results of our gp411gp120 biochemical

studies (Sunghwan Kim et aI., submitted) to design a gpl20-linked N-trimer antigen. Our

preliminary data show that the primary gp411gp120 contact interface lies between gp41 's

DSL20 (DSL plus 20 residues of C-peptide region adjacent to DSL) sequence and the

gp120 C5 region. To mimic this interaction in an antigen, we will attach gp120 to a

IZN 17/36-DSL20 construct via a disulfide bond to the DSL20 region (in a similar

42

manner to the previously reported SOS Env [16]).

The final gp120-IZN17/36-DSL20 product of this crosslinking is shown in Fig. 3.4.

Once we have generated suitable gp120-IZN17/36-DSL20 (as judged by solubility,

stability, trimeric structure, and the ability to bind small N-trimer ligands such as C37),

we will validate this antigen using our bulky test set as described above. Any

shortcomings in this validation process will guide future refinement of our design.

Validated designed antigen will be used in large scale Ab phage display screens to enrich

for Abs that can overcome the block. Winners from the selection will be authenticated by

ELISA and tested for their inhibitory activity in our standard entry assay. The Abs that

can overcome the block will be characterized structurally to learn how they overcome the

steric block and to inform the design of second generation antigens. Success from this

study would mark the first step down a long road towards making an antigen usable in

humans that would direct the production of such special Abs in high enough

concentration to prevent HIV infection.

43

gp120

Figure 3.4. Schematic representation of IZN36-DSL20 crosslined to gp120 via SOS-like disulfide bond. This cartoon shows an idealized version of this antigen. This interaction between gp 120 and DSL20 may simulate their orientations in the prehairpin intermediate and mimic the natural steric block. gpl20 is not drawn to scale.

44

References

1. Hamburger, A.E., et aI., Steric accessibility of the HIV-l gp41 N-trimer region. J BioI Chern, 2005.280(13): p. 12567-72.

2. Varadarajan, R., et aI., Characterization of gp120 and its single-chain derivatives, gp120-CD4D12 and gp120-M9: implications for targeting the CD4i epitope in human immunodeficiency virus vaccine design. J Virol, 2005.79(3): p. 1713-23.

3. Chen, B.K., et aI., Distinct modes of human immunodeficiency virus type 1 proviral latency revealed by superinfection of nonproductively infected cell lines with recombinant luciferase-encoding viruses. J Virol, 1994. 68(2): p. 654-60.

4. He, J., et aI., Human immunodeficiency virus type 1 viral protein R (Vpr) arrests cells in the G2 phase of the cell cycle by inhibiting p34cdc2 activity. J Virol, 1995. 69(11): p. 6705-11.

5. Connor, R.I., et aI., Vpr is required for efficient replication of human immunodeficiency virus type-l in mononuclear phagocytes. Virology, 1995. 206(2): p. 935-44.

6. Deng, H., et aI., Identification of a major co-receptor for primary isolates of HIV-l. Nature, 1996.381(6584): p. 661-6.

7. Landau, N.R. and n.R. Littman, Packaging system for rapid production of murine leukemia virus vectors with variable tropism. J Virol, 1992. 66(8): p. 5110-3.

8. Mirzabekov, T., et aI., Enhanced expression, native purification, and characterization of CCR5, a principal HIV-l coreceptor. J BioI Chern, 1999. 274( 40): p. 28745-50.

9. Furuta, R.A., et aI., Capture of an early fusion-active conformation of HIV-l gp41. Nat Struct BioI, 1998.5(4): p. 276-9.

10. Melikyan, G.B., et aI., Evidence that the transition of HIV-l gp41 into a six-helix bundle, not the bundle configuration, induces membrane fusion. J Cell BioI, 2000. 151(2): p. 413-23.

11. Henderson, H.I. and TJ. Hope, The temperature arrested intermediate of virus-cell fusion is afunctional step in HIVinfection. Virol J, 2006.3: p. 36.

45

12. Si, Z., et aI., Small-molecule inhibitors of HIV-l entry block receptor-induced conformational changes in the viral envelope g~ycoproteins. Proc Natl Acad Sci U

SA, 2004. 101(14): p. 5036-41.

13. Reeves, J.D., et aI., Sensitivity of HIV-l to entry inhibitors correlates with envelopelcoreceptor affinity, receptor density, and fusion kinetics. Proc Natl Acad Sci USA, 2002. 99(25): p. 16249-54.

14. He, Y., et aI., Peptides trap the human immunodeficiency virus type 1 envelope glycoprotein fusion intermediate at two sites. J Virol, 2003. 77(3): p. 1666-71.

15. Root, MJ., M.S. Kay, and P.S. Kim, Protein design of an HIV-l entry inhibitor. Science, 2001. 291(5505): p. 884-8.

16. Binley, J.M., et aI., A recombinant human immunodeficiency virus type 1 envelope glycoprotein complex stabilized by an intermolecular disulfide bond between the gp 120 and gp41 subunits is an antigenic mimic of the trimeric virion-associated structure. J Virol, 2000. 74(2): p. 627-43.

17, Abrahamyan, L.G., et aI., Human immunodeficiency virus (ype 1 Env with an intersubunit disulfide bond engages coreceptors but requires bond reduction after engagement to induce fusion. J ViroI, 2003. 77(10): p. 5829-36.

18. BinIey, J.M., et aI., Redox-triggered infection by disulfide-shackled human immunodeficiency virus type 1 pseudovirions. J Virol, 2003. 77(10): p. 5678-84.

19. Wei, X., et aI., Antibody neutralization and escape by HIV-l. Nature, 2003. 422(6929): p. 307-12.

20. Koch, M., et aI., Structure-based, targeted deglycosylation of HIV-l gp120 and effects on neutralization sensitivity and antibody recognition. Virology, 2003. 313(2): p. 387-400.

21. McCaffrey, R.A., et aI., N-linked glycosylation of the V3 loop and the immunologically silent face of gp120 protects human immunodeficiency virus type 1 SF162 from neutralization by anti-gp120 and anti-gp41 antibodies. J ViroI, 2004. 78(7): p. 3279-95.