128, 101-I

CELLULARIMMUNOLOGY

17(1990)

CD4-Derived Synthetic Peptide Blocks the Binding of HIV-1 GPI 20 to CD4-Bearing Cells and Prevents HIV-1 Infection ORITSHAPIRA-NAHOR*T'HANA GOLDING,~ LUBA K.VUJCIC,~ SANDRARESTO-RUIZ,$RAY L. FIELDS,* AND FRANK A. ROBEY* *The Peptide and Immunochemistry Unit, Laboratory of Cellular Development and Oncology, *Laboratory ofImmunology, National Institute ofDental Research, National Institute ofHealth, and t The Division of Virology, CBER, FDA, Bethesda, Maryland 20892 Received November 17, 1989; accepted February 13, 1990 The T cell surface glycoprotein CD4 plays an important role in mediating cellular immunity and serves as the receptor for human immunodeficiency virus. In order to identify primary sequenceswithin the CD4 molecule that may be involved in the binding of the HIV-I envelope, we synthesized various peptides corresponding to the VI, V2, V3, and V4 domains of CD4. We tested the ability of these peptides to block the binding of purified HIV-I gp 120to CD4+ human lymphoblastic leukemia cells (CEM) using fluorescence-activated cell sorting. One of these peptides, corresponding to CD4 amino acids (74-95), when preincubated with gp120, blocked its subsequent binding to CEM cells by 80%. A truncated form of this peptide (8 l-95), was found to be as efficient as the longer peptide (74-95) in inhibiting the binding of gp120 to CEM cells. The same peptide did not block the binding of OKT4A or Leu3A anti-CD4 monoclonal antibodies, which were previously shown to block HIV-I binding to CD4. The peptides were also tested for their ability to block HIV-I infection of a T cell line in vitro. Only CD4 peptide (7495) and the shorter fragment (81-95) succeeded in protecting T cells against infection with different HIV-I strains. All the other peptides examined had no effect on gp120 binding to CEM cells and did not block syncytia formation. Goat polyclonal antibodies against the CD4 peptide (74-95) gave modest interference of gp120 binding to CEM cells. These data suggestthat the CD4 region (74-95) participates in the CD4-mediated binding and/or internalization of HIV-I VhiOn.

0 1990 Academic

Press, Inc.

INTRODUCTION CD4+ helper T cells are important regulatory and effector cells in the immune system. The CD4 molecules on these cells may play an important accessoryrole during T cell activation. By binding to conserved determinants of class II major histocompatibility complex (MHC) molecules on antigen-presenting cells (APC), and to the T cell receptor (TcR) (l-6) they stabilize the complex of antigen, MHC ’ To whom correspondence should be addressed at current address NIAID, Bl.4 Room 388, NIH, Bethesda, MD, 20892. * Abbreviations used: ELISA, enzyme-linked immunosorbant assay; FACS, fluorescence-activated cell sorter; FITC, fluorescein isothiocyanate; HIV, human immunodeficiency virus; mAb, monoclonal antibody. 101 0008-8749/90 $3.00 Copyright 0 1990 by Academic Press, Inc. All rights ofreproduction in any form reserved.

102

SHAPIRA-NAHOR

ET AL.

classII, and TcR. This results in an increasing T cell-APC adhesion and T cell activation (7- 11). Human immunodeficiency virus (HIV-I) shows specific tropism toward the CD4 receptor. The virus predominantly attacks cells that bear CD4 receptors such as helper T cells and monocytes (12- 16), and induces the formation of multinucleated giant cells (syncytia) ( 17- 18). This infection can be blocked by soluble CD4 and by certain anti-CD4 murine monoclonal antibodies (mAb) in an epitope-specific manner ( 13- 1519-22). Furthermore, genetransfer experiments formally established that expression of CD4 molecules is required and sufficient to render a human cell susceptible to HIV-I infection (23). Recently it was demonstrated that a soluble form of the CD4 molecule lacking all but the first 180 amino acids blocked HIV-I binding to CD4+ cells (22, 24). Sitedirected mutagenesis of the CD4 gene led to the identification of regions and single amino acids that appear to be directly involved in binding to the HIV-I gp 120 envelope protein (26-30). This study was designed to test the ability of short peptides derived from the primary sequenceof CD4 to block HIV envelope binding to CD4+ T cells and to inhibit infection by various HIV-I strains. Such an approach may have a practical application in designing simple viral intervention therapy for individuals already exposed to the virus. It would have an advantage over the use of the entire CD4 receptor, which may lead to disturbances of normal immune functions. The present study suggeststhat in addition to epitopes previously identified as important in HIV-I gp120 binding that are recognized by the murine mAb OKT4A, OKT4D, and Leu3A (residues 35-58), there is an additional region of CD4, mapping to the V 1-J 1junction (residues 74-95), that may be involved in viral envelope binding. Furthermore, a small synthetic peptide encompassing this region, at relatively low concentrations, can provide CD4+ T cells full protection against HIV-I infection in vitro. Our results are in agreement with the data of Lifson et al. (3 1) who demonstrated that a derivatized CDCsynthetic peptide (76-94) blocked HIV infection and syncytia formation. MATERIALS AND METHODS Synthesis ofpeptides. CDcderived peptides were synthesized as amides on an Applied Biosystems (Foster City, CA) 430A automated peptide synthesizer, according to a previously described procedure (32). The amino acid sequence of the human CD4 used as a basis for the peptide selection was published by Maddon et al. (33). The sequences of the individual peptides and their positions in CD4 are shown in Table 1. After synthesis, the peptide amides were deprotected and cleaved from the solid-phase resin by using hydrogen fluoride and 10% anisole. The cleaved peptides were dissolved in 10% acetic acid or 0.1 A4 NH4HC03, filtered to remove the resin, and lyophilyzed. Peptides were used without further purification. The correct amino acid composition was verified by amino acid analysis. Chromatographic fmctionation of unpurified synthetic CD4 (74-95) is presented in Fig. 1. Two milligrams of the synthetic peptide was fractionated by high-performance liquid chromatography (HPLC) on a Vydac C4 reverse-phasecolumn, in a linear gradient of 0- 100%B over 30 min, where A = 0.1% TFA and B = 70% CH3CN in 0.1% TFA. The profile suggests the presence of some heterogeneity in the peptide preparation which results from

BLOCKING OF HIV-I INFECTION BY CD4-DERIVED

103

PEPTIDE

TABLE 1 Description of CD4Derived Synthetic Peptides and Their Biological Functions

Peptide sequence 1. [CD4 (74-95)] K-N-L-K-I-E-D-S-D-T-YI-C-E-V-E-D-Q-K-E-E-V-NH2 2. [CD4 (81-95)] S-D-T-Y-I-C-E-V-ED-Q-K-E-E-V-NH2 3. [CD4 (74~SO)] K-N-L-K-I-E-D-NH2 4. [CD4(132-147)] C-R-S-P-R-G-K-N I-Q-G-G-K-T-L-S-NH1 5. [CD4 (230-249)] T-F-D-L-K-N-K-E-V-SV-K-R-V-T-Q-D-P-K-L-NH2 6. [CD4 (318-335)] S-L-K-L-E-N-K-E-A-KV-S-K-R-E-K-A-V-NH*

Sequence length

Blocking of gpl20 binding to CEM cells

Inhibition of HIV-I-infection”

22

+

++ (12.5 &ml)

15

+

+ (50 a/ml)

7

-

-

16

-

-

20

-

-

18

-

-

Note. The amino acid sequenceshown (single letter code) is that of CD4-derived synthetic peptides that were examined in this study. Blocking ofgpl20-binding to CEM cells was described in the legend to Fig. 2. ’ Inhibition of HIV-I infection was described in the legend to Fig. 3. Numbers in brackets represent the concentration of peptide that gives >50% inhibition ofgiant cells 5 days after infection.

residual side chain protecting groups due to incomplete deprotection of the peptide and from deletion peptides where some poor coupling yields may have occurred during synthesis. The major peak contains a 55-67% concentration of whole peptide material. An amount of 1 pg/ml of CDCderived peptide (74-95) corresponds to 0.372 &fand 1 p/ml ofthe shorter peptide (81-95) corresponds to 0.57 PM. Cell lines. Human acute lymphoblastic leukemia T cells of the CEM line (CD4-F) [American Type Culture Collection (ATCC), Rockville, MD] were used for measure210 nm 0.6 0.5 0.4 0.3 02 0.1 00 5

10

15 Mrnutes

20

25

30

FIG. 1. HPLC chromatographic spectrum of CM-synthetic peptide (74-95). CDCpeptide (74-95) (2 mg/ml) was fractionated on a Vydac C4 reverse-phasecolumn.

104

SHAPIRA-NAHOR

ET AL.

ments of direct gp 120 binding. The human erythroleukemia cell line K562 (CD4-) (ATCC) was used as negative control. Viral infectivity assays(see below) were conducted with the human T cell leukemia line MOLT3. The cells were grown in RPM1 1640 supplemented with 10% FBS, 50 units/ml of penicillin, and 50 pg/ml streptomycin, at 37°C and 5% carbon dioxide. Reagents. Soluble HIV-I gp 120 was a gift of Dr. L. 0. Arthur, Program Resources, Inc., NCI-FCRF (Frederick, MD). HIV gp120 was isolated from the culture fluids of HIV-I (IIIB)-infected H9 cells by immunoaffinity chromatography and was then purified to homogeneity by polyacrylamide gel electrophoresis (34). At the highest concentration tested, the gp 120 preparations had < 1 ng of endotoxin per milliliter as determined by the limulus amebocyte lysate assay(35). Murine mAb OKT4A (antiCD4) and OKT8 (antiCD8) were from Ortho Diagnostic Systems (Raritan, NJ), Leu3A (anti-CD4) was from Becton-Dickinson (Oxnard, CA), and BL4 (anti-CD4) was from Pell-Freeze (Brown Deer, WI). Human anti-gp120 sera was a gift from Dr. M. Robert-Guroff, Laboratory of Tumor Cell Biology, NIH. The antiserum (heated for 30 min at 60°C) was used at a 1:100 dilution to detect gp 120 binding to CEM cells. Affinity-purified FITC-goat anti-human (Fab) IgG (Cappel Laboratories, West Chester, PA) was used in a 1:400 dilution to stain cells preincubated with gpl20 and anti-gp 120 serum. Generation of goat anti-CD4-derived peptide antisera. For immunization, N-chloroacetyl-derivatized peptides (32) were conjugated to the protein carrier, bovine serum albumin (Sigma, St. Louis, MO), which was pretreated with 2-iminothiolane (Sigma). The molar ratio of peptide to protein was 54: 1. Antisera to peptide conjugateswere raised in goats by subcutaneous injection of 10 mg of conjugate suspended in complete Freund’s adjuvant, followed by five additional injections of 5 mg conjugated peptide in incomplete Freund’s adjuvant, at 2-week intervals. Sera were collected before immunization and every two weeks after immunization. The titer of goat anti-CD4 peptide antibodies was measured by ELISA. Plasma was collected 10 days after the last boost and the anti-CD4 antibodies were affinity purified on a CHSepharose-4B (Pharrnacia, Uppsala, Sweden) column to which CDCderived peptide had been conjugated. Immunojluorescence analysis: Blocking of gp120 surface binding by CDI-derived peptide. CEM cells were resuspended in Hanks solution containing 2% FBS and 0.1% sodium azide, and were incubated (5 X 1O5cells/tube) with 200 rig/ml of HIV-I gp 120 at 37°C for 1 hr. The cells were washed and incubated at 4°C for 30 min with 2.5 pg/ ml of the following mAb to cell surface antigens: FITC-conjugated OKT4A, OKT8, BL4, and Leu3A, or with human anti-gp120 serum followed by FITC-goat anti-human IgG. The ability of soluble CD4 peptides to block gp120 binding was assessed by prior incubation of soluble gp120 with an excess amount (200 pg/ml) of CD4derived peptides at 37°C for 1 hr. In experiments to block surface binding of antiCD4 mAb, the mAb were preincubated with the CD4 peptides at 4°C or at 37°C for 2 hr before the cells were stained. After staining, the cells were washed twice, resuspended in 250 ~1 of the above buffer, and analyzed by flow microfluorometry with a FACS Star cell sorter (Becton-Dickinson). Enzyme-linked immunosorbent assay (ELISA). ELISA was performed by coating 96-well polystyrene plates (Falcon, Becton-Dickinson) with peptides at 10 pg/ml in carbonate buffer (pH 9.6) overnight at 4°C followed by 1%gelatin to block unbound sites. Goat anti-peptide antibodies were incubated at various dilutions for 2 hr at

BLOCKING

OF HIV-I

INFECTION

BY CD4-DERIVED

PEPTIDE

105

room temperature. This was followed by incubation with rabbit anti-goat affinitypurified IgG conjugated to alkaline phosphatase (1: 1000), (Sigma) for 1 hr at room temperature. Between each step extensive washing was performed with Tris buffer saline containing 0.1% Tween 20. Color was developed by addition of the substrate, pnitrophenyl phosphate (Sigma). The optical densities were read at 405 nm using microplate reader (Molecular Devices Corporation, Palo Alto, CA). HIV-infectivity assay. (i) Blocking infectivity of HIV-I isolates was done as described (36), by incubation of (1.5 X 104)MOLT3 cells with HIV-I virus, in 96-well plates, in the absence or presence of CDCderived peptides (at different concentrations) at 37°C and 5% C02. After 5 days, the number of giant cells per culture was determined and percentage blocking was calculated according to [(giant cells in absence of peptide) - (giant cells in presence of peptide)/(giant cells in absence of peptide)] X 100. In some experiments peptides were added 24 hr after infection with the virus. HIV-I strains used were III-B, RF, and Zaire. (ii) Syncytia formation assay. CEM and Molt3 cells were infected with the gp 160 vaccinia vector vsc25 or with the control vector vsc8 at 2-3 PFU/cell. Both vectors were produced in the laboratory of Bernard Moss (NIAID, NIH). Assays were performed in 96-well plates using duplicate wells containing 5 X lo5 cells per well. After addition of virus, the plates were spun for 45 min at 1700 rpm at room temperature, and incubated overnight at 37°C in 5% C02. Syncytia were scored between 24 and 30 hr. Proliferative responsesin vitro. The DP2-allospecific T cell clone (G9 16-3-53, 2.5 X 1O4cells/well) was stimulated for 3 days with the EBV-transformed B cell line (Z2B, 2.5 X lo4 cells/well). Normal PBL (5 X 104/well) were stimulated for 7 days with tetanus toxoid (25 pg/ml) or with the mitogen PHA (2 pg/ml). [3H]Thymidine (1 &i/well) was added to wells for the last 16 hr of culture. Blocking antibodies or CD4derived peptides were added to the responder cells 30 min prior to the addition of stimulators or antigens, and were present throughout the culture period. Murine mAb used: anti-CD4,OKT4A and anti-CDS, OKT8 (from Ortho Diagnostics). AntiDP mAb, B721, was a gift from W. Biddison, Neurobiology Branch, NIMH, NIH. These mAb were added to the culture at 1:200, final dilution. The peptides were used at 100-200 pg/ml. RESULTS Binding of soluble gp120 to CD4+ T cells. In order to identify CD4 epitopes that are involved in HIV infection, it was important to develop a system that would mimic the first steps involved in the in vivo infection of human CD4+ T cells by HIV-I envelope glycoprotein (gp120). For this purpose, we measured, in two different assays, the ability of increasing concentrations of soluble gp 120 [isolated from HIV-I (IIIB)-infected H9 cells] to bind to the CD4+ leukemic T cell line, CEM. Following incubation with gp120, the cells were stained with anti-gp120 human antisera followed by FITC-anti-human IgG reagent, and were subjected to FACS analysis (Fig. 2, open circles). Alternatively, CEM cells were preincubated with gp120 and then stained with FITC-OKT4A mAb, which was previously shown to bind to human CD4 at or near the binding site for HIV envelope glycoprotein (gp 120) (Fig. 2, closed circles). It was found that soluble gp 120 binds to CEM cells in a concentration-dependent fashion in the range used ( 12.5-300 r&ml). Binding of gp 120 to CEM cells did not interfere with their staining with the anti-CD8 mAb OKT8 (Fig. 2, open dia-

106

SHAPIRA-NAHOR

go120

ET AL.

concentration fng/ml)

FIG, 2. Assessmentof gp 120binding and inhibition of OKT4A binding to CL%bearing cells. Two methods were used. First (5 X 10’) CEM cells were incubated with various concentrations of gp120 for 1 hr at 37°C. The cells were then washed and incubated at 4°C with anti-gpl20 human serum (1: 100) followed by FITC-goat anti-human IgG (0). In the second method, the CEM cells were incubated with gp120 as above and then incubated with the murine mAb OKT4A-FITC (0). The results are expressed as percentages of the fluorescence intensity of control cells incubated in the absence of gp120. (0). Percentage binding of OKT8 to CEM cells in the presence of 200 rig/ml gp 120.

mond). In contrast, preincubation ofthese cells with 2200 rig/ml of gp 120 completely inhibited their subsequent staining with OKT4A (Fig. 2, closed circles). Becausethe second method was a more sensitive way to measure functional binding of gp 120 to CD4, this system was next used to evaluate the ability of CDCderived peptides to block gp 120 binding to CD4+ cells. CDI-derived synthetic peptide can block gp120 binding to CD4+ cells. The CD4derived synthetic peptides used in this study are summarized in Table 1. We looked for amino acid sequencesthat are highly charged (hydrophilic) and are thus expected to be immunogenic and at least partly exposed on the surface of the native CD4 molecule. The peptides were derived from the V 1-J 1 junction, and the V2, V3, and V4 CD4 domains. To identify potential CD4-gp 120 interaction regions, gpl20 (at 200 rig/ml) was preincubated with various peptides (200 pg/ml), and the mixtures were then added to CEM cells, which were subsequently stained with the CD4-specific mAb OKT4A and Leu3A. In preliminary experiments it was established that none of the peptides used in this study interfered with direct binding of OKT4A, Leu3A, BL4, and OKT4 to CD4+ cells, nor did they bind to these antibodies as measured by ELISA (not shown). As depicted in Fig. 3, the staining of CEM cells with OKT4A and Leu3A (A, B) can be completely blocked if the cells are preincubated with soluble gp 120 (200 rig/ml) prior to staining (C, D). When gp 120 was preincubated with CD4 peptide derived from VI-J1 junction (corresponding to amino acids 74-95) at 74.4 &4, its ability to block anti-CD4 mAb binding to CEM cells was significantly decreased,as demonstrated by the staining with OKT4A-FITC (80% of

BLOCKING

OF HIV-I

INFECTION

BY CDCDERIVED

Staining

1

10

PEPTIDE

Reagent

100 Log,o

Fluorescence

Intensity

FIG. 3. CD4-derived peptide can block gp 120 binding to CD4+ CEM cells. HIV-I gp 120 (200 rig/ml) was preincubated with various peptides (200 rg/ml) for 1 hr at 37°C before addition to (5 X 105) CEM cells for a further I hr incubation under the same conditions. The cells were then washed and were incubated with the mAb OKT4A-FITC (A, C, E, G, I) or Leu3A-FITC (B, D, F, H, J) for 30 min at 4°C and prepared for FACS analysis. The broken line represents binding of OKT4A and Leu3A after preincubation of CEM cells with gp120 (C-H). The dotted lines represent the binding of OKT4A or Leu3A to CEM cells when gp120 was preincubated with the indicated peptides (E-J). This figure represents one of six experiments which produced similar results.

control, Fig. 3E) or Leu3A-FITC (85% of control, Fig. 3F). Similar reduction in gp 120 binding to CEM cells following preincubation with this peptide was also determined by direct measurement using anti-gp 120 human antisera in FACS analysis (79% inhibition, data not shown). The blocking of gp120 binding by CD4 peptide (74-95) was dose dependent. No blocking was seenwith doseslessthan 50 &ml (19

108

SHAPIRA-NAHOR

ET AL.

PM). Preincubation of gp120 with peptides derived from other regions of CD4 (V2, V3, V4) did not interfere with its ability to bind to CEM cells (Figs. 31 and 35 and Table 1). In order to map the sequence that is responsible for the observed inhibition of gp120 binding by CD4 peptide (74-95), we made shorter fragments of this peptide, encompassing amino acids (74-80) and (8 l-95), and tested the ability of the short fragments to cause inhibition of gp 120 binding to CEM cells. When the truncated peptide (8 l-95) was examined, it was found to be as efficient as the longer peptide (74-95) in inhibiting gp120 binding to CEM cells (Figs. 3G and 3H). In contrast, the N-terminal truncated peptide (residues 74-80) had low or no effect on gp 120 binding to CEM cells in both assays(Table 1). Interestingly, when the active peptide (8 l-95) was further shortened from its N-terminal, its ability to block gp120 binding was gradually reduced. It was found that the fragments: 83-95,85-95, and 87-95 showed 80, 55, and 0% inhibition of gp 120 binding, respectively, in comparison with the peptide 8 l-95 that gave above 90% blocking of gp120 binding. Therefore, we have identified a short peptide (8 l-95) derived from the V 1-J 1 region of CD4, as a potent inhibitor of HIV-I envelope glycoprotein binding to CD4-bearing cells. This peptide includes one of the cysteines which participate in disulfide bond formation in the native CD4 molecule. CDI-peptide (74-95) can protect CD4+ cells from HIV-I infection. In order to establish the biological relevance of the blocking of gp 120 binding to CD4+ cells which was observed with the CD4-derived peptide (74-95), it was essential to test this peptide (and its shorter fragments) in HIV-I infectivity assays.This was particularly important, since this peptide maps downstream of the sites recognized by OKT4A and Leu3A, which were previously shown to be important in gp120 binding (25, 27, 28, 30). For that purpose, cells of a CD4+ T cell line, MOLT3, were infected with HIV-I (IIIB) in the absence or presence of decreasing concentrations of the CD4-derived peptides or soluble recombinant CD4. As can be seen in Fig. 4, in agreement with the findings in the binding assay,CD4 peptide (74-95) protected MOLT3 cells from infection by HIV-I in a dose-dependent fashion, giving 50% protection at 12.5 &ml (4.5 j&) and 100% protection at 100 pg/ml(37.2 pA4). The shorter fragment (8195) was needed at a higher concentration (250 gg/ml, 142 PM) in order to block 100% of syncytia formation following infection with HIV-I (IIIB) (Fig. 4). These data indicate that the HIV-I infectivity assaywas twofold more sensitive to blocking with the CD4 peptide compared to the gpl20 binding assay.Full protection was obtained with 100 pg/ml of peptide in comparison to the gp 120-binding assaywhere 200 pg/ ml peptide was required for blocking. The other peptides examined, which provided no inhibition of gp 120 binding to CEM cells, also did not block infection of MOLT3 cells by HIV-I (Fig. 4 and Table 1). Kinetic experiments were conducted to determine whether protection could be achieved by adding the inhibitory peptide after induction of viral infection. It was observed that the addition ofpeptide (74-95) to MOLT3 cells 24 hr after induction of infection with HIV-I did block syncytia formation on Days 4-5 after infection. However, this effect was only seen when higher peptide concentrations were added 24 hr postinfection [250 pg/ml(93 pM) and 125 &ml (46.5 PM) reduced syncytia formation by 100%and 50% respectively]. These findings suggestthat the addition of peptide to the HIV virion prior to infection of cells, on Day 0, can actually block viral internalization and thus reduce the viral load or number of infected cells, while later addition of the peptide can still be beneficial in blocking giant cell formation and potentially blocking cell to cell transfer of HIV-I.

BLOCKING OF HIV-I INFECTION BY CD4-DERIVED

PEPTIDE

109

60 ,z g 5 ;

70 60

:g

50

f+ 8

40 30

-

Soluble CD4 Peptlde 74-95 Peptlde 81 95

0-O

Peptides:

20

132.147 318

Peptide

Concentration

335

I@glmll

FIG. 4. CD4-peptide (74-95) can block giant cell formation in cell cultures infected by HIV-I. (1.5 X 104) MOLT3 cells were incubated with HIV-I virus in the presence or absence of various concentrations of CD4-derived peptides (X, 0, q ), or soluble recombinant CD4 (0). Cultures were incubated at 37°C for 5 days, and then were scored for giant cell formation, Percentage inhibitions were calculated according to [(giant cells in absence of peptide) - (giant cells in presence of peptide)/(giant cells in absence of peptide)] X 100. Results are expressedas the means + SEM of values obtained from five experiments.

CD4 peptide (74-95) was also found to protect MOLT3 cells against infection by other HIV-I strains, including RF and Zaire (the latter being more divergent from IIIB strain than the RF) (37), with the same dose responsesas that observed with the IIIB strain (data not shown). In a second syncytia assay, CEM and MOLT3 cells were infected with the gp160 vaccinia vector in the presenceor absenceof CD4-derived peptides. In this assayvery high levels of gpl60 are expressedby the infected cells within 16-20 hr postinfection and syncytia can be scored at that time. Peptides (74-95) and (8 l-95) gave 50% blocking at 5 pg/ml(2 PM) and 40 fig/ml (23 p&f), respectively (Table 2). Complete deprotection of CD4 peptide (74-95) diminished its biological activity. As mentioned under Materials and Methods and shown in Fig. I, the CD4 peptide (74-95) which is responsible for the inhibitory activity in the various HIV-I-binding and infectivity assaysretained a low level of side chain-protecting groups (as benzyl esters) due to incomplete deprotection of the synthetic peptide. The correct amino acid composition was always verified by amino acid analysis. Lifson et al. (3 1) have recently described a similar finding with a peptide derived from the same region (7694) of the CD4 molecule. They assumed that the remaining benzyl groups have an important role for the biological activity of the peptide. The peptide (74-95) in the present study was subjected to HPLC purification. The major eluted peak (Fig. 1, 13.4- 14.5 min) had no inhibitory activity in the gp120-binding assay. However, in the syncytia formation assayusing vaccinia gpl60-infected CEM cells, the pure peptide gave 50% inhibition at I5 PM as compared to 50% inhibition seenwith 2 pA4 of the pre-HPLC peptide (Table 2), suggestinga seven- to eightfold reduction in activity. We also made an attempt to modify the HPLC-purified peptide by other chemical

110

SHAPIRA-NAHOR

ET AL.

TABLE 2 Inhibition of gpl60 Vaccinia-Mediated Syncytia of CEM Cells by ClWDerived Peptides CD4 peptide aa 74-95 (unpurified) aa 74-95 (HPLC purified)* aa 8 l-95 (unpurified) aa 312-147 (unpurified)

Concentration

20 &ml 10 a/ml 5 rtiml 40 fig/ml 20 &ml 10 &ml 40 &ml 20 &ml 10 &ml 20 &ml 10 fig/ml 5 a/ml

% Inhibition of gpl60 vaccinia-mediated syncytia“ 82% 76% 50% 50% 44% 30% 44% 27% 0% 11% 0% 0%

n CEM cells (5 X 104)were infected with vSC25 (gpl60 vaccinia, 2 PFU/cell) in duplicate, in the presence or absence of peptides at the indicated concentrations. Syncytia were scored after 24 hr. Control cultures (no peptide added) contained 132 syncytia per well. ’ Major peak in the HPLC profile depicted in Fig. 1.

modifications of the cysteine (at position 86). N-Ethylmaleimide-modified cysteine did not regain its biological activity. The correct amino acid composition in the purified peptide was always verified and compared with the unpurified peptide. It seems therefore, and according to Lifson et al., that the benzyl side groups modify the peptide in a unique manner which is essential for its biological activity. Goat antibodies to CD4 peptides bind to CDI-bearing cells.The CD4derived synthetic peptides (Table 1) were conjugated to BSA and were used to elicit antibodies against the CD4 molecule in goats. Affinity-purified antibodies and preimmune control sera were then evaluated for their anti-peptide reactivity in ELISA, and for their binding to CDCbearing cells in indirect immunofluorescence assays.The affinitypurified antibodies were found to be specific for their respective immunizing peptides. For example, the antibodies raised against peptide (74-95) reacted specifically with this peptide, but not with the other CDCderived peptides (residues 132-147, 230249, and 3 18-335). When the antibodies were evaluated for their ability to recognize “native” CD4 receptor molecules on CDCpositive CEM cells or CDCnegative K562 cells, they were found to stain CEM cells to various degrees,with the highest fluorescence intensity observed using goat anti-peptide (3 18-335) antibodies, followed by anti-peptide (74-95) antibodies, anti-peptide (132-147) antibodies and anti-peptide (230-249) antibodies (Fig. 5, I, C, E, and G, respectively). It is important to note that the brightest staining with the goat anti-peptide (318-335) antibodies (I) was 70-80% of the fluorescent intensity obtained by staining with OKT4A (A) while staining with the other goat anti-peptide antibodies did not exceed 25% of the fluorescence intensity obtained following staining of the same cells with the murine CDCspecific mAb (Fig. 5, A, C, D, and E). These different levels of staining may reflect the relative accessibility of these epitopes in the native form of the CD4 to antibody binding, or may reflect

BLOCKING OF HIV-I INFECTION BY CD4-DERIVED CD4( + 1 Cells ICEMI

PEPTIDE

111

CD41 - 1 Cells (K562)

500 7,

J

E j :, : i ; :

i :

1

10

1

100 Log Fluorescence

., ‘: ‘...,

10

100

Intensity

FIG. 5. Goat antibodies against CD4-peptides can recognize “native” CD4. CEM (CD4+) or K562 (CD4-) cells (1 X 106)were stained with: A, B: . . . . , FITC-goat anti-mouse IgG; -, OKT4A (2.5 &ml); C-J: . . . . , FITC-rabbit anti-goat IgG; -, affinity-purified goat anti-peptide antibodies (25 &ml) followed by FITC-rabbit anti-goat IgG. C: ---, goat anti-peptide (74-95) antibodies preincubated with soluble pep tide (74-95).

the heterogeneity of the goat polyclonal antibodies which may differ in their specificities and binding affinities. Nevertheless, these data show that three out of the four goat anti-CD4 peptide antibodies did recognize the CD4 molecule on CD4+ cells

112

SHAPIRA-NAHOR

100

-

90

-

80

-

g

70-

‘; m %

60

-

50

-

.g 2 z k

ET AL.

e

4030

-

20

-

10

-

ogpl20 Goat antiCD4 Peptide Preincubation with peptide

I

_ _

-

74-95

74-95

_

74-95

132.147

-

318.335

-

FIG. 6. Antibodies against peptide (74-95) interfere with gpl20 binding to CDCbearing cells. CEM cells (5 X 105)were preincubated with goat antibodies raised against different CD4 peptides (25 &ml) for 30 min at 4°C washed, and then incubated with (200 @ml) soluble gp120 for 1 hr at 37°C. Fluorescence intensity was measured followed staining with human anti-gp120 antisera (1: 100) followed by FITC-antihuman (Fab) IgG. Results are expressed as the percentages of fluorescence intensity seen in control cells incubated with free gp120. The specific inhibition was tested by preincubation of the goat anti-peptide (74-95) antibodies with the specific peptide before incubation with gp120.

(but did not stain CD4- cells). The specificity of the binding of these antibodies to CDbbearing cells was also established by preabsorption with the relevant or irrelevant peptides prior to CEM staining. Blocking of goat anti-CD4 peptide (74-95) antibodies binding to CEM cells by the free CD4 peptide (74-95) is shown in Fig. 5C. These affinity-purified goat anti-peptide antibodies were next tested for their ability to block gp120 binding to CEM cells. As shown in Fig. 6, the only goat antibodies that gave partial blocking (40%) of gp 120 binding were specific for the CD4 peptide (74-95). This blocking effect can be reversed by preincubation of the antibodies with the specific free peptide (Fig. 6), but not with irrelevant peptides (not shown). The other two antisera, which did stain CD4+ cells, had no effect on gp 120 binding. These results are in concurrence with our earlier findings that peptide (74-95) [but not pep tides (132- 147) and (3 18-335;; was effective in blocking gp 120 binding and protection of CDCpositive cells against HIV-I infection. None of the goat anti-peptide antibodies blocked the binding of monoclonal anti-CD4 antibodies (OKT4A, Leu3A and BL4) to CEM cells (not shown). The antibodies against peptide (74-95) were also tested in ELISA for their binding to truncated forms of this peptide, and were found to bind primarily to the (81-95)

BLOCKING OF HIV-I INFECTION BY CDCDERIVED

PEPTIDE

113

0.6 Peptides:

0.5

Normal

z 9

0.4

d ti F

0.3

(74.951. O-O (66.951. M 183.951. o--o 181.951. M 17%951. FT 137.58). w Goat IgG- x---.x

7 5 .-: $ cz

0.2

1.1500 antibodes

1:4-500 dhtlon

FIG. 7. Goat anti-peptide (74-95) antibodies bind primarily to the (81-95) fragment. Affinity-purified antibodies against peptide (74-95) were added at various dilutions to ELISA plates coated with different truncated forms ofpeptide (74-95), or to irrelevant peptide (37-58) (X - X). Binding of normal Goat IgG to peptide-coated plates (X . . X). Results are expressedas the means + SEM of triplicate OD 405 absorption values from three experiments.

fragment (Fig. 7). The affinity-purified anti-peptide (74-95) antibodies were tested in the HIV-I viral infectivity assay, but were found ineffective in protecting MOLT3 cells. This finding was not surprising in light of their only moderate ability in blocking gp 120 binding to CD4+ cells. CD4-derivedpeptides do not inter&ee with normal T cell responses.Becauseof the pivotal role of CD4+ T helper cells in many immune responsesand the contribution of the CD4 molecule in cellular interactions, it was important to determine whether the CD4-peptide identified in this study as an inhibitor of gp 120 binding to cells has any effect on the proliferative responsesof normal CD4+ T cells. To test this possibility, the CD4-derived peptides were added to cultures of CD4+ DPZspecific T cell line G9 16-3-53 stimulated with the relevant stimulator cells (Table 3). The proliferative responseof the T cell clone is very sensitive to blocking with the CDCspecific OKT4A mAb. The addition of CDCpeptides or peptide-specific goat antisera to these cultures did not inhibit these proliferative responses.Similarly, none of the CD4-derived peptides had any inhibitory effect on the tetanus toxoid (CDCdependent) or PHA (CD4independent) responsesof normal PBL (data not shown). Therefore, it would appear that the CD4 region identified in this study as a potential blocker of HIV-I gp120 binding to T cells is not involved in normal immune functions and would probably not interfere with immune responsesif administered in vivo. DISCUSSION Recombinant soluble CD4 has recently been shown to be capable of inhibiting HIV-I infection and syncytium formation in vitro (2 1, 22). These findings have led

114

SHAPIRA-NAHOR

ET AL.

TABLE 3 CD4-Derived Peptides Do Not Interfere with Activation of Normal CD4+ T Cell Clones

Responder

Stimulator

T cell clone (G9 16-3-53)

B cell line Z2B

Inhibitor

Uptake cpm

% Control

anti-CD4 anti-DP peptide 14-95 peptide 132-147 peptide 229-249 peptide 3 18-335 Goat anti-74-95

15,798 1,964 600 16,188 14,917 15,516 14,775 14,595

100 12 4 102 94 98 93 92

Note. Responder T cells and stimulator B cells (Dp’, 10,000 rads irradiated), were cultured at 1:1 ratio (2.5 X 104/well) for 3 days in the presence or absence of the murine mAb OKT4A (l:4OO/well) and B72 1 (aDP, 1:1,000) or with CD4derived peptides ( I25 &ml) or goat anti-peptide (74-95) ( 1:20). [3H]Thymidine was added for the last 16 h. All the cpm values represent net cpm = Total cpm - (cpm of responders + cpm of stimulators only).

to proposals for the use of recombinant soluble CD4 as an antiviral intervention agent in HIV-I-infected individuals. Concerns exist, however, that long term treatment with soluble CD4 could act as an immunosuppressant by competing with cell surface CD4 for binding to its putative ligand in the MHC class II molecules expressed by antigen-presenting cells. This would result in a further reduction in the activation potential of CDCpositive helper lymphocytes in patients undergoing treatment. An alternative approach which was taken in the present study is to identify short fragments within the CD4 molecule which are intimately involved in binding and/or internalization of the HIV-I virus. In the short term, large scale production of these fragments in the form of synthetic peptides is considerably easier than mass production of the recombinant CD4. In addition, CDCderived short peptides are less likely to be immunosuppressant or to lead to the generation of undesirable anti-CD4 autoantibodies in treated individuals. Along this line, previous studies utilized techniques such as site-directed mutagenesis and immunoselection of CD4 mutants, and identified epitopes in the CD4 molecule recognized by various murine anti-CD4 mAb (e.g., Leu3A, OKT4A, and OKT4D) which are e5cient blockers of HIV-I binding (2630). Unfortunately, the same CD4 epitopes were found in earlier immunological studies to be potent inhibitors of normal immune functions. Thus, it could be argued that synthetic peptides tailored according to sites recognized by these monoclonal antibodies could be as immunosuppressant as the intact soluble CD4 in vivo. The data presented in this study describe the biological study of a 22-amino acid peptide (aa 74-95) mapped to the VI-J1 junction of CD4. This peptide as well as a shorter derivative (aa 81-95) was found to be effective in blocking the binding of soluble gp120 to live CD4+ cells, while three other peptides taken from the V2, V3, and V4 domains of CD4 had no biological activity in this assay (Fig. 3, Table 1). More importantly, this V 1-J 1-derived peptide was found to be a potent inhibitor of HIV-I infection of CD4+ T cells in vitro using different, divergent strains of HIV-I (Fig. 4, Table 1). These findings suggestthat this region (74-95) of CD4 binds to a

BLOCKING OF HIV-I INFECTION BY CD4-DERIVED

PEPTIDE

115

conserved region on gpl20. Inhibition of syncytium formation was also seen (albeit at higher peptide concentrations) when the peptide was added 24 hr after the induction of viral infection. In a separateset of experiments, these CDCderived peptides were added to vaccinia gpl60-infected cells. In this assay, infected cells express very high levels of HIV-I gp 120 on their surface and form large syncytia with neighboring CD4+ cells within 24 hr. The CDCderived peptide (74-95) could effectively block syncytia formation in this assayas well at low concentrations (80% inhibition by 8 PM). In an independent study it was reported that a similar CD4 peptide (aa 76-94) could inhibit HIV-induced syncytia formation, and this biological activity was dependent on benzyl modification of the cysteine in position 86 (3 1). In the present study we confirm and extend these findings by showing that the same region of the CD4 molecule blocks the binding of soluble gp 120 to CD4+ cells (Fig. 3), and that antisera against this peptide provide partial (40%) but not complete protection against gp 120 binding to CD4+ cells (Fig. 6). More importantly, we showed that this peptide blocks HIV-I infection of CD4+ cells, by various divergent HIV strains, and also reduces syncytia formation after infection has been established. The HPLC-purified peptide (74-95) was inactive in blocking gp 120 binding to CEM cell, while it was still active in a vaccinia gp 160 syncytia assay.However, it was seven- to eightfold less potent in its ability to block the syncytia formation. Thus, we are in agreement with the finding of Lifson et al. regarding the possibility that the biological activity of the peptide is not restricted just to the linear amino acid sequence (74-95) but may be dependent on contributions by the benzyl side groups (3 1). However the remaining protecting groups on the cysteine in peptide (74-95) could not in themselves be responsible for the biological activity of this peptide, since even purified peptide (in our study) had low biological activity, and the other peptides which were used in our study also contained residual protecting groups, yet showed no blocking activity. In addition, when we reduced the size of the unpurified truncated peptide (81-95) from the Nterminal, the blocking activity of the peptide was reduced or lost. Thus, the two amino acids serine and aspartic acid (aa 8 1,82, which were not included in the short peptide tested by Lifson et al,) were found to be important for the biological activity of the peptide. These findings indicate that both the primary amino acid sequence and the secondary structure assumedby this peptide are critical for its biological activity. This is to be expected since it contains one of the cysteines participating in the disulfide bonds of the first Vl domain of CD4. The CD4 region identified in our study was not recognized by the murine OKT4A and Leu3A mAb. However, it is possible that in the three-dimensional structure this region is close to the CD4 region recognized by these mAb. It was also important to establish that the CDCderived peptide (74-95) does not interfere with normal immune functions. Therefore, this peptide (as well as control peptides) was added to cultures of human PBL stimulated with tetanus toxoid or to cultures of an alloreactive classII-specific T cell clone activated by allogeneic stimulators. In both casesproliferative responsescould be efficiently blocked by OKT4A and Leu3A mAb, but were not affected by the addition to culture of the CDCderived peptide (74-95) or the goat antibodies raised against it (Table 3). Therefore, it seems that this region of CD4 may not be important in normal CD4-MHC class II interactions and could provide the differential inhibition of HIV-I viral infection but not normal immune cellular interactions. This characteristic of the CD4 peptide (74-95)

116

SHAPIRA-NAHOR

ET AL.

provides it with a distinct advantage over the CD4 molecule asa potential therapeutic agent. The disadvantage is that a larger amount of the peptide is needed compared to soluble CD4. ACKNOWLEDGMENTS We thank Drs. Basil Golding, J. E. Folk, L. Salzman, L. M. Wahl, and E. Wolf for their critical review of this manuscript.

REFERENCES 1. Reinhetz, E. L., and Schlossman, S. F., Cell 19,821,1980. 2. Swain, S. L., Proc. Natl. Acad. Sci. USA 78,1101, 1981. 3. Marrack, P., Endres, R., Shimonkovitz, R., Zlotnik, A., Dialynas, D., Fitch, F., and Kappler, J., J. Exp. Med. 158,1077,1983. 4. Swain, S. L., Immunol. Rev. 74,129, 1983. 5. Gay, D., Maddon, P., Sekaly, R., Talle, M. A., Godfrey, M., Long, E., Goldstein, G., Chess,L., Axel, R., Kappler, J., and Marrack P., Nature (London) 328,626, 1987. 6. Gay, D., Bum, S., Pastemak, J., Kappler, J., and Marrack, P., Proc. Natl. Acad. Sci. USA 85, 5629, 1988. 7. Doyle, C., and Strominger, J. L., Nature (London) 330,256, 1987. 8. Bierer, B. E., and Burakoff, S. J., FASEB J. 2,2584, 1988. 9. Rogozinski, L., Bass,A., Glickman, E., Talle, M. A., Goldstein, G., Wang, J., Chess,L., and Thomas, Y., J. Immunol. 132,735, 1984. 10. Saizawa, K., Rojo, J., and Janeway, C. A., Nature (London) 328,260,1987. 11. Janeway, C. A., Carding, S., Jones, B., Murray, J., Portoles, P., Rasmussen, R., Rojo, J., Saizawa, K., West, J., and Bottomly, K., Immunol. Rev. 101,39, 1988. 12. Dalgleish, A. G., Beverley, P. L. C., Clapham, P. R., Crawford, D. H., Greaves, M. F., and Weiss, R., Nature (London) 312,763, 1984. 13. Klatzmann, D., Champagne, E., Chamaret, S., Gruest, J., Guetard, D., Hercend, T., Gluckman, J. C., and Montanier, L., Nature(London) 312,167, 1984. 14. McDougal, J. S., Mawle, A., Cort, S. P., Nicholson, J. K. A., Cross, G. D., Shempple-Campbell, J., Hicks, D., and Sligh, J. M., J. Immunol. 135,3151, 1985. 15. McDougal, J. S., Kennedy, M. S., Sligh, J. M., Cort, S. P., Mawle, A., and Nicholson, J. K. A., Science 231,382,1986. 16. Sattentau, Q. S., and Weiss, R. A., Cell 52,63 1, 1988. 17. Sodroski, J., Goh, W., Rosen, C., Campbell, K., and Haseltine, W., Nature (London) 322,470,1986. 18. Lifson, J. D., Feinberg, M. B., Reyes, G. R., Rabin, L., Banapour B., Chakrabarti, S., Moss, B., WongStaal, F., Steimer, K. S., and Engleman, E. G., Nature (London) 323,725, 1986. 19. Fauci,A. S., Science239,617, 1988. 20. Sattentau, Q. J., Dalgleish, A. G., Weiss, R. A., and Beverley, P. C. L., Science 234, 1120, 1986. 21. Smith, D. H., Bym, R. A., Marsters, S. A., Gregory, T., Groopman J. E., and Capon, D. J., Science 238,1704,1987. 22. Traunecker, A., Luke, W., and Karjalainen, K., Nature (London) 331,84,1988. 23. Maddon, P. J., Dalgleish, A. G., McDougal, J. S., Clapham, P. R., Weiss, R. A., and Axel, R., CeN47, 333,1986. 24. Berger, E. A., Fuerst, T. R., and Moss, B., Proc. Natl. Acad. Sci. USA 85,2351,1988. 25. Richardson, N. E., Brown, N. R., Hussey, R. E., Vaid, A., Matthews, T. J., Bolongnesi, D. P., and Reinherz, E. L., Proc. Natl. Acad. Sci. USA 85,6 102,1988. 26. Landau, N., Warton, M., and L&man, D. R., Nature (London) 334,159,1988. 27. Jameson, B. A., Rao, P. E., Kong, L. I., Hahn, B. H., Shaw, G. M., Hood, L. E., and Kent, S. B. H., Science240,1335,1988. 28. Mizukami, T., Fuerst, T. R., Berger, E. A., and Moss, B., Proc. Natl. Acad. Sci. USA 85,9273,1988. 29. McDougal, J. S., Nicholson, J. K. A., Cross, G. D., Cort, S. P., Kennedy, M. S., and Mawle, A., J. Immunol. 137,2931, 1986. 30. Peterson, A., and Seed,B., Cell 54,65,1988.

BLOCKING OF HIV-I INFECI-ION BY (XX-DERIVED

PEITIDE

117

31. Lifson, J. D., Hwang, K. M., Nava, P. L., Fraser, B., Padgett, M. Dunlop, N. M., and Eiden, L. E., Science241,712, 1988. 32. Linder, W., and Rohey, F. A., Int. J. Pept. Prot. Rex 30,794, 1987. 33. Maddon, P. J., L&man, D. R., Godfrey, M., Maddon, D. E., Chess, L., and Axel, R., Cell 42, 93, 1985. 34. Robey, W. G., Arthur, L. O., Matthew, T. J., Langlois, A., Copeland, T. D., Lerche, N. W., Orosxlan, S., Bolongnesi, D. P., Gilden, R. V., and Fischinger, P. J., Proc. Natl. Acad. Sci. USA 83, 7023, 1986. 35. Yin, E. T., Galanos, C., Kinsky, S., Bradshaw, R. A., Wessler, S., Luderitz, O., and Sarmiento, M. E., B&hem. Biophys. Acta. 261,284, 1972. 36. Vujcic, L. K., Shepp, D. H., IUutch, M., Wells, M. A., Hendry, R. M., Wittek, A. E., Krilov, L., and Quinnan, G. V., Jr., J. Infect. Dis. 157,1047, 1988. 37. Willey, R. W., Rutiedge, R. A., Dias, S., Folks, T., Theodore, T. Buckler, C. E., and Martin, M. A., Proc. Natl. Acad. Sci. USA 83,5038, 1986.