VIROLOGY

190,

134-142

(1992)

Amino Acid Residues of the Human lmmunodeficiency Virus Type I gpl20 Critical for the Binding of Rat and Human Neutralizing Antibodies That Block the gpl20-sCD4 Interaction J. A. McKEATlNG,*r’ J. CORDELL,*

M. THALl,t C. FURMAN,t S. KARWOWSKA,+ M. K. GORNY,+ S. ZOLLA-PAZNER,+$ J. SODROSKl,t AND R. A. WEISS*

*Institute of Cancer Research, Chester Beatty Laboratories, Fuiham Road, London, U.K. SW3 6/B; tDivision of Human Retrovirology, DanaFarber Cancer Institute, and Department of Pathology, Harvard Medical School, Boston, Massachusetts 02115; *Department of Pathology, New York University and Medical School, 550 First Avenue, New York, New York 100 16; and §Research Service, Veterans Affairs Medical Center, 423 East 23rd Street, New York, New York 100 10 Received

February

20,

1992;

accepted

May

2 1. 1992

We have characterized the discontinuous epitopes recognized by two rat and three human neutralizing monoclonal antibodies (mAb) by examining the effect of single amino acid changes in conserved residues of gp120 on mAb recognition. A human mAb derived from an infected individual, 448D, and two rat mAbs, 39.139 and 39.3b, respectively, derived by immunization with native recombinant gpl20, recognize similar epitopes. Recognition of the envelope glycoproteins by these mAbs was affected by changes in gpl20 amino acid residues 88,113,117,257,368, or 370. The gp120 amino acids 257, 368, and 370 have previously been reported to be important for CD4 binding, which is consistent with the ability of these mAbs to block the gp120-CD4 interaction. Residues 88, 113, and 117 are not thought to be important for CD4 binding, suggesting that the antibody epitopes overlap, but are distinct from, the CD4 binding region. We also found that some alterations in gpl20 residues 88, 117, 368, or 421 reduced the ability of polyclonal sera from HIV-l-infected individuals to inhibit the interaction of the mutant gpl20 glycoproteins with soluble CD4. Thus, changes in the HIV-1 gpl20 glycoprotein that minimally affect the receptor binding may allow escape from neutralizing antibodies directed against the CD4 binding region. o 1992 Academic PWSS, I~C.

INTRODUCTION

mation and which block the rgpl20-sCD4 interaction (Ho et a/., 1991; Posner et al., 199 1; Tilley et a/., 199 1; Karwowska et a/., 1992). The conformational nature of the epitope(s) recognized by these neutralizing mAbs is in agreement with data suggesting that the CD4 binding site is not a linear domain but involves several discontinuous conserved regions (Olshevsky et a/., 1990; Thali et al,, 1991 a). Recently, the discontinuous epitope of one human mAb that blocks the gpl20-CD4 interaction has been shown to consist of gpl20 amino acids located in at least four regions, two of which overlap regions important for receptor binding (Thali et al., 1991 b). We have previously reported that immunization of rats with rgpl20 in the presence of Freund’s adjuvant produced four mAbs capable of blocking gpl20 from binding to sCD4, two of which recognized conformation-dependent epitopes (Cordell eta/., 1991). Here we have used a panel of gpl20 mutants, previously characterized for CD4 binding (Olshevsky et al., 1990), to determine if the epitopes recognized by the two rat mAbs overlap with those of three human neutralizing mAbs that are able to block the gp120-CD4 interaction (Karwowska et a/., 1992). Identification of the residues in which changes affected the binding of both human and rat mAbs allowed us to study effects of these changes on the ability of polyclonal human sera

The induction of cross-reactive neutralizing antibodies for human immunodeficiency virus (HIV-l) is a primary goal for vaccine development. Sera from HIV-Iinfected individuals neutralize a broad spectrum of divergent HIV-1 isolates, suggesting the presence of conserved neutralization epitopes (Weiss et a/., 1985, 1986; Robert-Guroff et al., 1985). The conserved regions of gpl20 that are involved in the interaction with the CD4 receptor are an obvious target for such neutralizing antibodies. The CD4 binding site may be immunogenic in man since most HIV-1 -positive sera can block the binding of virus or recombinant gp120 (rgp120) to soluble CD4 (sCD4) (McDougal et al., 1986; Skinner et al., 1988; Schnittman eta/., 1988; Moore eta/., 1990; Back eta/., 1990). Recently it has been shown that much of the group-specific neutralization activity in human sera may be attributable to antibodies that are sensitive to gpl20 conformation and which block rgpl20-sCD4 binding (Steimer et a/., 199 1). Consistent with this, several neutralizing human monoclonal antibodies (mAb) have been cloned which are sensitive to gpl20 confor-

’ To whom 0042-6822/92

reprint

requests

should

$5.00

CopyrIght 0 1992 by Academic Press. Inc. All rlghts of reproduction in any form reserved.

be addressed 134

AMINO

to inhibit the interaction proteins with sCD4. \

MATERIALS

ACID

RESIDUES

of the mutant envelope glyco-

AND

METHODS

mAbs CBH/Cbi rats were immunized via Peyer’s patches with recombinant gpl20 (rgp120) (HIV-l IIIB; BHlO clone expressed in CHO cells) emulsified in complete Freund’s adjuvant. Hybridoma screening assays are described elsewhere (Cordell et al., 1991). MAbs 39.139 and 39.313 block rgpl20 from binding to sCD4 and bind to conformation-sensitive epitopes. In contrast mAb 38.1 a also blocks gpl20-sCD4 binding but binds to a linear peptide encompassing amino acids 430-447 (McKeating et al., 1992). MAb 41 .l i recognizes the V3 loop and is used as a control for gpl20 concentration in the cell lysates. Cell lines producing human mAbs to HIV were derived by fusing EBV-transformed peripheral blood cells from HIV-infected individuals with the SHM-D33 heteromyeloma, as described previously (Gorny et al., 1991). Cells producing antibodies specific for HlVwere identified byreactivityon ELlSAwith rgp120 (HIV-1 IIIB, expressed in baculovirus vectors) (Repligen, Cambridge, MA). Three human mAbs were used: 448-D, 559/64-D, and 558-D. All of the mAbs, human and rat, blocked the gpl20CD4 interaction and neutralized various HIV-1 isolates. Doses of antibody required to result in 50% neutralization ranged from 0.5-l 2.8 /*g/ml (Cordell et al., 1991, Karwowska et al., 1992). Binding of mAbs to gpl20

OF

HIV-I

teins at a predetermined saturating concentration of 10 pglml. Bound mAbs were detected with either alkaline phosphatase-conjugated anti-human Ig or anti-rat lg. The results are expressed as a ratio of the optical density (OD) of the mutant protein to wild type protein, and are termed the relative binding index.

Inhibition

of gpl20-sCD4

binding

Lysates containing the mutant and wild type proteins were bound to the solid phase as above at a concentration of 45 rig/ml in 2% milk powder in lysis buffer. Human sera (QC 1, 2, 4, 5, and 6; McKeating et al., 1989) were diluted l/1000 in 2% milk powder, 20% sheep serum in Tris-buffered saline (TMSS), and were allowed to bind to the D7324-immobilized envelope glycoproteins (triplicate wells) for 1 hr at room temperature. This dilution of human sera resulted in approximately 50% inhibition of the gpl20-sCD4 interaction. After washing away any unbound antibody, sCD4 (Smith Kline French, King of Prussia, PA) at 0.1 pg/ml in TMSS was added and incubated for 1 hr at room temperature. Bound sCD4 was detected with a rabbit antisCD4 serum (CBL-34) and a sheep anti-rabbit Ig alkaline phosphatase conjugate. The ability of the sera or mAbs to inhibit gpl20 from binding to sCD4 was determined. Relative sCD4 binding indices of the mutants 88N/P, 113 D/A, 117 K/W, 257 TIA, 257 T/G, 313 P/S, 368 D/T, 42 1 K/L, 432 K/A, 470 P/G, 475 M/S for sCD4 were 0.17, 1.01, 0.30, 0.84, 0.79, 1.03, 0.36, 0.59, 0.87, 0.76, and 1.08, respectively.

mutants

Mutations were introduced into the HXBc2 env gene and mutant glycoproteins transiently expressed in COS-1 cells as described previously (Olshevsky et al., 1990). Cells were lysed in the presence of 50 mMTris, 150 mM NaCI, 1% Triton, 0.1% SDS and the concentration of gpl20 in the lysates was determined by twin site ELISA as previously described, using recombinant CHO-derived gpl20 as a calibrant (Moore et al., 1990). Cell lysates were diluted in lysis buffer (see above) and were captured onto a solid phase by a sheep antibody D7324 to the C-terminus of gpl20, in the presence of 2% nonfat milk powder (Marvel, Cadbury, U.K.). Bound gpl2O/gpl60 was detected with a pool of HIV+ human sera and with alkaline phosphatase conjugated antihuman Ig (SeraLab, Crawley, U.K.). Mutant and wild type envelope glycoproteins were allowed to bind to the solid phase via D7324 at an input concentration of 15 rig/ml. All mAbs were tested for their binding to the various mutant envelope glycopro-

135

gp120

RESULTS Epitope mapping To identify amino acids important for recognition of gpl20 by the rat and human mAbs, their reactivity with a set of HIV-1 gpl20 mutants altered in conserved residues was examined. These mutant glycoproteins have been previously characterized with respect to gpl60 precursor processing, gpl20-gp41 association, and CD4 binding ability (Olshevsky et al., 1990). The wild type (E7) and mutant gpl2Os were tested by ELISA for their ability to bind to a saturating concentration of the various gpl20-sCD4-blocking mAbs (10 pug/ml), with a V3 loop mAb (41.1 i) included as a control to assess any variation in gp120 input. The results were expressed as a ratio of mAb bound to the mutant gpl20 in comparison to the wild type protein (Relative Binding Index) (Table 1). The V3 mAb bound equivalently to all the mutants with binding indices ranging from 0.86-l .lO (data not shown). Mutant glycoproteins with changes

136

McKEATlNG

ET AL.

TABLE BINDING OF SELECTED gpl20

MUTANTS

1

TO mAbs

39.313,

Relative Mutant

39.313

39.139

wt 69W/L 80 N/R 88 N/P 106 E/A 113 D/A 117 K/w 125 L’G 252 RAN 256 S/Y 257 TIG 262 NR 269 E/L 314 GIW 368 D/P 370 E/Q 384 Y/E 420 I/R 421 K/L 432 WA 457 D/G 470 PIG 475 MIS 4821314 ELYIGRA

1 .oo 1.05 0.72 0.38 0.92 0.16 0.25 0.88 0.94 0.16 0.25 0.14 0.72 0.89 0.35 0.14 0.44 0.86 0.51 0.92 0.97 0.71 0.59 0.31

39.139,

binding

index

38.1 a, 588,

559 AND 448

of mAb

38.la

588

559

448

1.00 1.05 0.89 0.40 0.99 0.14 0.27 1.11 1.02 0.16 0.25 0.14 0.70 0.81 0.48 0.16 0.59 0.98 0.55 0.95 0.92 0.70 0.61

1 .oo 1.07 0.71 1.15 0.98 0.65 0.88 0.77 0.90 1 .oo 0.90 0.75 0.67 1.02 1.06 1.07 0.88 1.04 1.05 0.74 0.95 0.89 0.93

1 .oo 1 .oo 0.85 0.34 1.05 0.76 0.42 0.61 1 .oo 0.13 0.42 0.05 1.06 0.84 0.60 0.93 0.98 1.04 0.91 1.03 0.88 0.96 1 .oo

1 .oo 1.00 0.87 0.51 1.03 0.85 0.08 0.75 0.98 0.06 0.31 0.06 1.07 0.94 0.73 1.08 1.02 1.09 0.88 1.05 0.93 1.01 0.86

1 .oo 0.98 0.83 0.25 0.94 0.03 0.18 0.57 0.95 0.07 0.40 0.12 1.03 1.02 0.05 1 .oo 0.84 1.14 0.73 1.26 0.80 0.89 1.03

0.39

0.10

0.84

0.87

0.50

Note. None of the mutants 40 Y/D, 102 E/L, 103 Q/F, 120/21 VWLE, 207 K/W, 266 A/E, 267 E/L, 281 NV, 298 R/G, 380/81 GE/YV, 382 F/L, 395 W/S, 427 WN and W/S, 429 K/L, 433 A/L, 435 Y/H, 438 P/R, 456 N/D, 463 N/D, 477 D/V, 485 WV, 491 I/F, 493 P/K, 495 G/K, or a mutant with a deletion of residues 1 19-205 had binding indices for any of the mAbs < 0.60, with the exception of mAb 38.1 a. In addition mAb 41.1 i bound to all the mutants tested with binding indices in the 0.92-l. 12 range. The binding indices of the following mutants for mAb 38.1 a were: 427 W/S, 0.26; 427 W/V, 1.12; 430 V/S, 0.10; 433 A/L, 0.05; 435 Y/H, 0.08; and 438 P/R, 0.08. The results represent the mean of triplicate wells where the standard deviation was within 15% of the mean for all samples. These results are from one experiment, but similar results were obtained in two separate experiments

in amino acids 88, 1 13, 1 17, 256, 257, 262, 368, 370, 421, 482-484 exhibited reduced binding to both rat mAbs. Only mutations at residues 1 13 D/A or D/R, 256 S/Y, 262 N/T, and 370 E/Q or E/R resulted in a complete loss of 39.3b and 39.13g binding (Binding index

Amino acid residues of the human immunodeficiency virus type I gp120 critical for the binding of rat and human neutralizing antibodies that block the gp120-sCD4 interaction.

We have characterized the discontinuous epitopes recognized by two rat and three human neutralizing monoclonal antibodies (mAb) by examining the effec...
1MB Sizes 0 Downloads 0 Views