- Email: [email protected]
Epitope exposure on functional, oligomeric HIV-1 gp41 molecules
VIROLOGY206, 713-717 (1995)
Epitope Exposure on Functional, Oligomeric HIV-1 gp41 Molecules QUENTIN J. SATI-ENTAU,*'1 SUSANZOLLA-PAZNER,t AND PASCAL POIGNARD* *Centre d'lmmunotogie de Marseille-Luminy, Case 906, 13288 Marseille Cedex 9, France; tLaboratory Service, Veterans Affairs Medical Center, 423 East 23rd Street, New York, New York 10010; and Department of Pathology, NYU Medical School, 550 First Avenue, New York, New York 10016 /?eeeived July 15, 1994; accepted September 26, 1994 We have used cells infected with the HIV-1 molecular clone HX10 to study the binding of monoclonal antibodies (mAbs) to different epitopes within the extracellular domain of the HIV-1 transmembrane glycoprotein gp41. Gp41 mAb binding to the infected cells at 4 ° was variable but weaker than the binding of an anti-gp120/V3 loop mAb and increased substantially for three of the gp41 antibodies at 37 °, Treatment of the cells with soluble CD4 (sCD4) at 37 ° increased gp41 mAb binding to epitopes spanning residues 5 2 1 - 6 6 3 , implying that these regions had probably been masked by gp120, which following interaction with sCD4 had subsequently dissociated from gp41. By contrast, the binding of a mAb to residues 6 6 2 - 6 6 7 which form a neutralization epitope was reduced by sCD4 binding. Another region which has been described as containing a neutralization epitope spans residues 7 2 5 - 7 5 0 . MAbs to this region bound equally well to noninfected and HIV-infected cells, and binding was not increased in the presence of sCD4. These data strongly imply that this epitope is not exposed on the external surface of the membrane, a finding in accord with the proposed cytoplasmic localization of this region. © 1995Academic Press, Inc.
The env gene of the human immunodeficiency virus type 1 (HIV-1) codes for a single chain glycoprotein precursor, gp160, which is cleaved to yield the mature glycoproteins gp120 and gp41 (reviewed in 1). These two subunits assemble into a heterodimer held together by noncovalent interactions and are expressed at the virion surface in a functional, multimeric form (2, 3). Virus binding takes place via a high affinity interaction between gp120, the surface envelope glycoprotein, and the HIV cellular receptor, the 0D4 molecule (4, 5). Subsequent entry of the virion capsid into the cell takes place by fusion of the virus and cell membranes, an event mediated by the transmembrane glycoprotein, gp41 (6). The N-terminus of gp41 has been shown to contain a hydrophobic fusion domain on the basis of structural homology with the fusion domains of other enveloped viruses (7, 8), and mutagenesis of this region leads to fusion-incompetent virions (9, 10). The precise mechanism by which gp41 induces virus-cell membrane fusion is not yet clear, but certain steps in the process are being elucidated. The binding of CD4 to gp120 triggers conformational changes in the envelope glycoproteins which are thought to trigger fusion in a process termed receptor-mediated activation of fusion (11, 12). In cellline-adapted viruses the end point of these molecular rearrangements is dissociation of gp120 and exposure of gp41 (13- 16). Infection of 0D4 + cell lines with HIV results in a chronic infection in which the cell expresses no detect-
able CD4 but high levels of the HIV envelope glycoproteins (15, 16, 23). This gp120/41 is functional as demonstrated by 0D4 binding and the rapid formation of syncytia in cocultures of HIV + and CD4 + cells (1). A number of monoelonal antibody (mAb) epitopes have been characterized on gp41, including two which are recognized by neutralizing mAbs; one in the extracellular membrane proximal portion (17) and one in the cytoplasmic domain (18). Gp41 expressed at the surface of HIV-infected cells in its oligomerio, functional form was probed with a panel of epitope-mapped gp41 mAbs to determine which regions are obscured by gp120 and which are constitutively exposed on the surface of HIVinfected cells. The gp41 mAbs were obtained from the following sources: IAM 41-3E)6, IAM 41-4D4, IAM 41-2502, IAM41-2F5, and IAM 41-3H 12 are human mAbs obtained from Viral Testing Systems (Houston, TX); 50-69 and 98-6 are human mAbs which preferentially recognize the oligomeric form of gp41 which were prepared and epitope mapped as previously described (19); and 1583 and 1908 were prepared by immunization of mice with recombinant chimaeric poliovirus containing the HIV gp41 sequence 735-752 (20). The anti-gp120/V3 loop mAb 110.5 was obtained from Genetic Systems, Inc. (Seattle, WA), and the CD4 domain 4-reactive mAb L120 prepared by Becton-Dickinson, Inc. (San Jose, CA) was obtained from the MRC AIDS Directed Programme, UK. H9 cells were infected with the HX10 clone of HIV-1 and cultured for 8 days before use, as previously described (16). At this time the cells expressed no detectable cell surface CD4 and high levels of the HIV envelope
~To whom correspondence and reprint requests should be addressed. Fax, (33) 91269430. 713
0042-6822/95 $6.00 Copyright© 1995by Academic Press, Inc. All rights of reproductionin any form reserved.'
714
SHORT COMMUNICATIONS
glycoproteins as determined by immunofluorescent staining with the anti-gp120 mAb 110.5. One million cells in a volume of 100 #1 were incubated for 2 hr with or without soluble CD4 (sCD4 from R. Sweet, Smith, Kline and Beecham, King of Prussia, PA; 21) at 10 ffg/ml at 4 or 37 ° with agitation, washed, and then stained with the anti-gp41 mAbs for 1 hr at 4 or 37° with agitation before washing and fixation overnight in 2% formaldehyde. After further washing an appropriate dilution of anti-human immunoglobulin coupled to phycoerythrin (Immunotech SA, Luminy Case 915, Marseille, France)was incubated with the fixed cells for 1 hr at 4° with agitation followed by washing as before. Flow cytometric analysis was carried out using a Becton-Dickinson FAOScan with Consort 30 software. The interaction of sOD4 with gp120 of cell-line-adapted HIV-1 isolates at 37 ° induces the dissociation of gp120 from gp41 (13-16). In order to determine which regions of gp41 are masked by gp120 and which are constitutively exposed, a panel of gp41 mAbs with known epitopes was used to stain uninfected or HIV-infected, envexpressing H9 cells in the presence or absence of prebound sCD4. The background staining with uninfected cells was subtracted from the specific signal obtained with the infected cells. In the absence of sOD4, staining of HIV-infected H9 cells with the gp41 mAbs at 4° was variable but relatively weak by comparison with the signal obtained with the gp120 mAb 110.5 (Fig. 1). This low-level gp41 staining is in accord withprevious studies (15, 16, 22, 23) and probably reflects the constitutive partial exposure of gp41 molecules in the presence of gp120, but may also imply the exposure of a small proportion of gp41 molecules resulting from a basal level of spontaneous gp120 dissociation. Following the addition of sOD4 at 4° , staining increased subtly for all mAbs except the neutralizing mAb 2F5 (Fig. 1). This small increase in mAb reactivity is probably the result of low-level gp120 dissociation which occurs with some isolates of HIV-1 in the presence of sCD4 at 4° (16). The binding of sCD4 to the cells was verified by staining with the anti-CD4 mAb L120; gp120 binding does not interfere with L120 since this mAb reacts with domain 4 of 0D4 (15). Incubation of the cells with sOD4 at 37 ° resulted in a diminution of cell surface gp120-0134 complexes as demonstrated bythe decreased binding of the 110.5 and L120 mAbs. The reduction in cell-surface gp120 induced by sOD4 was estimated from the decrease in sOD4 staining using mAb L120 (calculated by dividing the value obtained at 4° with that obtained at 37 °) and the decrease in gp120 staining using mAb 110.5 (calculated by dividing the value obtained with untreated ceils by that obtained with ceils treated with sCD4 at 37 °) and was 2.3- and 2.8-fold, respectively (mean of two separate experiments). By this analysis the decrease in 110.5 staining was consistently greater than the decrease in sCD4 staining, probably resulting from a small
reduction in signal obtained with sCD4 treatment at 4° as a result of low-level shedding. The binding of all gp41 mAbs with the exception of 2F5 was increased substantially after sOD4 pretreatment at 37°; 2502, 50-69, 3D6, and 98-6 showed maximum increases (mean of two separate experiments) over sOD4-untreated cells of 6.4-, 2.8-, 2.7-, and 5.8-fold, respectively. The greater reactivity of these mAbs following sCD4 treatment at 37 ° is very likely to be a direct result of gp120 dissociation revealing previously occluded portions of gp41, since it has been shown previously that the increase in gp41 staining correlates with the level of gp120 dissociation (16). The variation between mAbs in the level of binding enhancement probably reflects the ratio of preexposed to sOD4-induced exposure of gp41 epitopes; 50-69 and 3D6 bind gp41 relatively well in the absence of sOD4-induced gp120 dissociation, whereas 2502 and 98-6 bind poorly. Since the gp41 mAbs interact with HIV virions in vivo at physiological temperature, we decided to test the binding of the gp41 mAbs to the infected cells at 37° (Fig. 2). The signal obtained at 37 ° with the mAbs 2502, 50-69, 3D6, 98-6, and 2F5 increased by 1.6-, 0.9-, 1.0-, 4.3-, and 10.6-fold, respectively (mean of two experiments) over that observed at 4 °. Thus, raising the temperature induced a dramatic increase in the exposure of the two membrane proximal epitopes (644-663 and 662-667) and a modest increase in the NH2-terminal epitope (521 538). In the presence of sCD4 prebound at 37°, mAb reactivity at 37 ° was 1.8-, 0.9-, 1.1-, 3.4-, and 6.9-fold greater than that at 4° for mAbs 2502, 50-69, 3D6, 98-6, and 2F5, respectively, implying that increasing the temperature for the mAb binding step did not substantially alter the level of sOD4-induced gp120 dissociation. Surprisingly, sCD4 binding induced a decrease in 2F5 binding which was detectable at 4° (1.3-fold, mean of two experiments) and became more obvious at 37 ° (2.1-fold, mean of two experiments). Thus unlike the other gp41 mAbs tested, the 2F5 epitope is constitutively strongly exposed on the surface of infected cells at 37 °, and mAb reactivity is reduced in the presence of sCD4. The temperature dependence of this epitope may relate to its availability within the gp41 oligomer; at 4° the epitope may be obscured by gp120, by carbohydrate groups, by another gp41 molecule, or by other membrane components. Increasing the temperature may allow the oligomeric structure to "relax" somewhat within the membrane, thereby freeing the epitope. The sCD4-induced reduction in 2F5 signal is more difficult to explain, but may result from a conformational modification of the epitope induced indirectly by sCD4 binding to gp120, directly via an interaction between sOD4 and gp41, or from epitope masking caused by a receptor binding-induced interaction of gp41 with other molecules in the infected cell or virion membrane. Binding of the 2F5 mAb to gp41 may inhibit HIV fusion by interfering with one the above- .
SHORT COMMUNICATIONS 80 -
0o8° 14o25C2 (521-538)
715
50-69 (579-613)
6o5° 1 3D6 (604-617)
60
40
40"
20
20
r"
e-
O) 0e.
0 ,1
1
10
.... , 1000
100
0
1
10
100
.... , 1000
0
1
,1
10
100
1000
0 (n 0
2011598-6 (644-663)
20-
2F5 (662-667)
>,
15-
C
~
3o0.
"i
250'
gp120 staining
200' 10'
150 ' 100'.
so
5
~
-
o 110.5
0
.... =
.... l
1
10
100
1000
1
10
100
L120 mAbs
1000
MAb concentration(nM) FIG. 1. Effect of sCD4 on gp41 mAb binding. The gp41 mAbs were incubated at the concentrations shown with HIV-l-infected H9 cells which hadbeenpretreated or not with sCD4 at 4 or 37°, followed by an anti-human immunoglobulin-phycoerythrin-coupled second antibody. MAb binding wasanalyzedby flow cytometry using a FACScan with Consort 30 software. Each datum point represents the mean fluorescence intensity of 10,000 accumulatedeveners,from which the signal obtained with uninfected cells has been subtracted= mAb binding in the absence of sCD4 is represented by(O), in the presence of sCD4 at 37° (Iq), and at 4° (A). Note that the vertical scale (mean fluorescence intensity) for 2502, 50-69, and 3D6 differs fromthat of 98~6and 2F5. The epitope position for each mAb is presented in parentheses after the mAb names and is numbered according to the BH10amino-acid sequence; minimum epitopes for 50-69 and 98-6 were taken from Xu et al. (19), 2F6 from Muster et al. (123, and 2502 and 3D6 from the commercial literature accompanying these mAbs. The boxed histogram shows staining by mAbs 110.5 (gp120 V3 loop) and L120 (0D4 domain4) at 10 #g/ml of cells from the same experiment described above carried out under the same conditions. Open (El) bars represent mAb staining of untreated, HIV-infected H9 cells; hatched ([]) bars represent staining of infected cells pretreated with sOD4 at 4°; and filled (B) bars representstaining of infected cells pretreated with sCD4 at 37°.
mentioned pathways or may function by yet another unknown mechanism. Figure 3 illustrates graphically the location of the epitopes which, from the results presented in this study, are exposed in the presence of gp120, and those which are likely to be c o m p l e t e l y or partially occluded. It is clear that the majority of the extracellular portion of gp41, at least that comprising residues 5 2 1 - 6 6 3 , is likely to be partially or c o m p l e t e l y masked in the presence of g p l 20. This is in good a g r e e m e n t with tlhe w i d e l y spaced diversity of mutations indicated in Fig. 3 w h i c h disrupt the gp120/CD4 interaction. The only mAb epitope w h i c h appears well exposed in the presence of gp120 is that of 2F5 which is the most m e m b r a n e proximal (662-667), although mAb binding is highly temperature dependent. The cytoplasmic d o m a i n of gp41 has been reported to contain a neutralizing epitope spanning residues 7 3 5 752 (18, 20). MAbs have been prepared in mice against recombinant polioviruses containing this region of HIV gp41, which neutralize in a relatively type-specific man-
ner (24). We have tested two of these reagents and a human mAb with similar epitope specificity for binding to cell surface-expressed gp41 in the presence or absence of sOD4. Figure 4 c o m p a r e s the binding of four mAbs to uninfected H9 cells and sOD4 pretreated or untreated HX10-infected H9. The gp120 mAb 110.5 gave a strong signal, and pretreatment of the cells with sOD4 reduced the signal by about threefold. MAb 98-6 to the extracellular domain of gp41 reacted w e a k l y with infected H9 cells in the absence of sOD4 and substantially more strongly in the presence of sCD4 preadsorbed to the cells at 37 ° . Two neutralizing anti-HIV ( 7 3 5 - 7 5 2 ) mAbs, 1583 and 1908 (20, 24), gave intermediate binding to H9 and gave an identical signal on infected H9 cells. Treatment with sOD4 resulted in a small diminution in staining in this experiment, but this w a s not a reproducible finding. Similar results w e r e obtained for 3H12, a human mAb reactive with an epitope within the gp41 s e q u e n c e 6 9 3 - 8 5 6 (data not shown). These three mAbs w e r e demonstrated to be active in a peptide ELISA; all
716
SHORT COMMUNICATIONS
100 "
A
80"
60"
4-, ¢h eI!1
40-
20
C 0
2502
tJ' t" 0 m
100 •
50-69
3D6
98-6
2F5
B
0
in HIV infection. (1) The immunodominant disulfide-linked loop and flanking regions which have been suggested to form a conserved structural motif which functions in the association of the surface and transmembrane glycoproteins of lentiviruses (26). Moreover, peptides synthesized from the sequence of this region bind to cellular proteins of 45 and 80 kDa, an association which has been suggested to represent a potential HIV-coreceptor interaction (27). (2) A predicted o~-helical region spanning residues 558-595, structural conservation of which is importantfor infectivity, and peptides from which inhibit HIV-mediated fusion (28). Antiserum raised against this peptide binds better to HIV-infected cells in the presence of sCD4, a finding which concurs with our suggestion that this region is at least partially masked by gp120. (29).
80.
t~
60'
40"
)-69 20-
3DE 0 25C2
50-69
Monoclonal
3D6
98-6
2F5
antibodies
FIG. 2. Temperature dependence of gp41 mAb binding. HW-infected H9 cells either untreated ( i ) or preincubated with sCD4 (10 #g/ml) for 1 hr at 37° ([~) were washed and incubated for 2 hr with the mAbs shown at 37° (A) or 4 ° (B). After overnight fixation the cells were stained with anti-human conjugated phycoerythrin and analyzed by flow cytometry as described in the legend to Fig. 1. The histogram bars represent the mean fluorescence intensity obtained from the analysis of 10,000 accumulated events from which the signal obtained with uninfected H9 cells has been subtracted.
~mino ~rminus 2
gave strong signals with the peptide 735-752, but no significant activity with an irrelevant peptide (data not shown). It thus appears that the gp41-specific epitope(s) for these mAbs is not expressed at the surface of HIVinfected cells. This is in accord with the predicted intracellular location of the cytoplasmic tail of gp41, there being no evidence for a second transmembrane spanning region. These findings differ from results presented in another study (25), in which HIV-infected cells reacted more strongly than uninfected cells with an anti-735-752 peptide mAb. The reason for this discrepancy is unclear, and the mechanism of neutralization of HIV by these mAbs remains obscure. The HIV-cell fusion process is thought to be mediated by a direct interaction between the hydrophobic N-terminal fusion domain of gp41 and the cell membrane (1). Our current thinking suggests that this region is probably masked by gp120 until the virus and cell membranes are in close apposition, when a process termed receptormediated activation of fusion triggers the exposure of the fusion domain (11, 12). In addition to the fusion domain, there are other regions which are potentially functional
Extracellular
carboxy terminus ,~.
Intracellular JD
%®®-
FIG. 3. Regions of gp41 masked by gp120. The amino-acid backbone and proposed secondary structure are modified from Gallaher et al. (33). The thick black line represents the regions of gp41 which are likely to be partially or completely masked by gp120 in the functional giycoprotein oligomer. It is unclear from this study whether the Nterminal portion of the fusion peptide is masked by gp120, since the epitope of mAb 2502 is reported to begin at residue 521, and no reagents specific for the region N-terminal to this have been tested. The arrowheads, some of which are numbered according to the published sequence of HIV-1 BH10, indicate residues which are thought to be critical for gp120 association (taken from Cao et aL, 30). The epitope for each of the mAbs tested is boxed and labeled, and the epitope for mAb 2F5 (662-667) is shaded.
SHORT COMMUNICATIONS
717
3. HALLENBERGER,S., TUCKER, S. P., OWENS, R. J., BERNSTEIN, H. B,, and COMPANS, R. W. Virology 193, 510-514 (1993). 4. McDoUGAL, J. S., KENNEDY, M. S., SLIGH, J. M., OORT, S. P., MAWL, A., and NICHOLSON, J. K. A., Science 231, 382-386 (1986). 5. LASKY, L. A., NAKAMURA, G. M., SMITH, D. H., FENNIE, O., SHIMASAKi, C., PATZER, E., BERMAN, P. W., GREGORY,T., and CAPON, D. J., Ceil I!. //r ,~S....~... \ 50, 975-985 (1987). 6. KOWALSKI, M., POTZ, M., BASIRIPOUR, T., DORFMAN, T, GOH, W. C., 3:3 21 ."'...... "'.. ................ $ :,,:. . . . . . . . \ ~,.....; TERWILLIGER, E., DAYTON, A., ROSEN, C., HASELTINE, W. A., and 100 101 102 103 SODROSKI, J., Science 237, 1351-1355 (1987}. r7. GALLAHER, W. R., Cell 50, 327-328 (1987). 8. GONZALEZ-SOARANO, F., VVAXHAM, M. N., ROSS, A., and HOXlE, J. A., 0 AIDS Res. Hum. Retroviruses 3, 245-282 (1987). I !i. 9. FREED, E. O., MYERS, D. J., and RISSER, R., Proc. Natl. Acad. Sci. USA ~ -'.1\ 87, 4650-4654 (1990). .f" I ' 4 q 10. BERGERON, L,, SULLIVAN, N., and SODROSKI, J., J. Viro/. 66, 2389-2397 ~! ~(1992). i'[ ~. ,,{. ,i.. 11. ALLAN, J. S., Science 252, 1322 (1991). ....... ,~,4.., ' ' '"'~1 ........ . . . . . . . . 12. MOORE, J. P., MCKEATING, ]. A., WEISS, R. A., OLAPHAM, P. R., and 101 102 103 10° 101 102 103 SATIENTAU, O. ]., Science 252, 1322-1323 (1991). 13. HART, T. K., KIRSH, R., ELLENS, H., SWEET, R. W., LAMBERT, D. M., Mean fluorescence intensity PEqq-EWAY, S. R., LEARY, J., and BUGELSKI, P. J., Proc. Nat/. Acad. SeL USA 88, 2189-2193 (1991). FIG. 4. Staining of HIV-infected and uninfected cells by mAbs to gp41 14. MOORE, J. P., MOKEATING, ]. A., WEISS, R. A., and SAIq-ENTAU, O. J., cytoplasmic domain. HIV-l-infected H9 cells either pretreated (---) or Science 250, 1139-1142 (1990). not (. • .) with sCD4 and uninfected H9 cells (. • .) were incubated with 15. SATTENTAU, Q. J., and MOORE, J. P., J. Exp. Med. 174, 407-415 (1991 ). fluoreehrome conjugated second antibody alone ( - - ) or with the . . . . 1 16, SAI]-ENTAU, Q. J., MOORE, J. P., VIGNAUX, F., TRAINCARD, F., and PO~Gantibodies shown at a d dut~on of either ~-~ for ascites or a concentration NARD, P., ./. VlroL 67, 7383-7393 (1993). of 30 fig/el for purified immunoglobulin for 2 hr at 4 ° with agitation, 17, MUSTER, T., STEINDL, F., PURTSCHER, M., TRKOLA, A., KLIMA, A., HIMNi~ washed, and then fixed. After a further wash the cells were labeled LER, G., RUOKER, F., and KATINGER, H., J. VlroJ. 67, 6642-6647 with phycoerythrin-coupled second antibodies and then analyzed by (1993). flow cytometry on a FAOScan using Consort 30 software. Each peak 18. OHANH, T. C., DREESMAN, G. R., KANDA, P., LINE~FE, G. P., SPARROW, represents the accumulation of 10,000 gated events. Note that the J. T., HO, D. D., and KENNEDY, R. C., EMBOZ 5, 3065-3071 (1986). staining of uninfected and HIV-infected cells with mAbs 1583 and 1908 19. XU, J. Y., GORNY, M. K., PALKER,T., KARWOWSKA, S., and ZOLLA-PAZNER is superimposed in this experiment. S., J. Viro/. 65, 4832-4838 (1991). 20. EVANS, D. J., McKEATING, J., MEREDITH, J. M., BURKE, K. h., KATRAK, K., JOHN, A., FERGUSON, M., MINOR, P. D., WEISS, R. A., and ALMOND, (3) The coflserved neutralization epitope located close to J. W., Nature 339, 385-388 (1998). the membrane on the extracellular segment of gp41 (17). 21. DEEN, K. C., McDOUGAL, J., INACKER, R., FOLENA-WASSERMAN, G., ARTHOS, J., ROSENBERG, J., MADDON, P. J., AXEL, R., and SWEET, The data presented here imply that gp41 and gp120 interR. W., Nature 331, 82--84 (1988). act over a large surface, a conclusion which is not sur22. TYLER, D. S., STANLEY, S. D., ZOLLA-PAZNER, S,, GORNY, M. K., SHADprising considering the variety of mutations in both subDUCK, P. P., LANGLOIS, A. J., MAT]HEWS, T. ]., BOLOGNESI, D. P., units which interfere with this interaction (6, 30-32, Fig. PALKER, m. J., and WEINHOLD, K. J., J. Immune/. 145, 3276-3282 (1990). 3). Further studies using viral glycoproteins expressed 23. SPEAR, G. T., TAKEEMAN, D. M., SULLIVAN, B. L., LANOAY, A. L, and in a functional form at the cell surface may be useful for ZOLLA~PAZNER, S., J. Virol. 67, 53-59 (1993). the dissection of other structure-function relationships. 24. VELLA, O., FERGUSON, M., DUNN, G., MELOEN, R., LANGEDUK, H., EVANS, D., and MINOR, P. D., J. Gen. Virol. 74, 2603-2607 (1993). 25. DALGLEISH, A. G., CHANH, T. C., KENNEDY, R. C., KANDA, P., CLAPHAM, ACKNOWLEDG M ENTS P. R., and WEiss, R. A., Virology t65, 209-215 (1988). 26. SCHUL]2, T. F., JAMESON, B. A., LOPALCO, L., SICCARDI, A. G., WEISS, We thank H..Holmes and A. Newberry and the MRC AIDS Directed R. A., and MOORE, J. P., AiDS Res. Hum. Retroviruses 8, 1571Programme for reagents and time and K. Simon and A. Daser for their 1580 (1992). help with the manuscript. This work was supported by the CNRS, IN27. HENDERSON, L. A., and QUERSH~, M. N., .L BioL Chem. 268, 15291SERM, and a grant from the ANRS of France and in part by NIH Grants 15297 (1993). A1 32424 and A1 36085 and research funds from the Department of 28. WILD, C., OAS, T., MCDANAL, C., BOLOGNESI, D., and MAWHEWS, T., Veterans Affairs, NY. Proc Natl. AcacL ScL USA 89, 10537-10541 (1992). 29. MA]-FHEWS,T. J.,/n "Proceedings of the Seventh Colloque des Cents Gardes" (M. Girard and L. Valette, Eds.), pp. 117-121 (1992). REFERENCES 30. OAO, J., BERGERON, L., HELSETH, E., THALI, M., REPKE, H., and SODROSKI, J., Z VIroL 67, 2747-2755 (1993). 1. MOORE, J. P., JAMESON, B. A., WEISS, R. A., and SATTENTAU, Q. J., /n 31. HELSETH, E., OLSHEVSKY, U., FURAN, O., and SODROSKI, J., Z ViroL 65, "Viral Fusion Mechanisms" (J. Bentz, Ed.), pp. 233-289. CRC 2119-2123 (1991). Press, Boca Raton, FL, 1993. 32. IVEY-HOYLE, M., CLARK, R. K., and ROSENBERG, M., J. VlroL 65, 26822. PINTER, A., HONNEN, W. J., TILLEY, S. A., BONA, C., ZAGHOUANI, H., 2695 (1991). GORNY, M. K., and ZOLLA-PAZNER, S., J. ViroL 63, 2674-2679 33. GALLAHER, W. R., BALL, J. M., GARRY, R. F., GRIFFIN, M. C., and MONTE.LARO, R. C., AIDS Res. Hum. Retroviruses 5, 431-434 (1989). (1989).
110.5
E
~i
98-6
/)ii
100 101 102 103 1908 1583
I
100
I
,
, ......
I
.......
,I
.......
Epitope Exposure on Functional, Oligomeric HIV-1 gp41 Molecules QUENTIN J. SATI-ENTAU,*'1 SUSANZOLLA-PAZNER,t AND PASCAL POIGNARD* *Centre d'lmmunotogie de Marseille-Luminy, Case 906, 13288 Marseille Cedex 9, France; tLaboratory Service, Veterans Affairs Medical Center, 423 East 23rd Street, New York, New York 10010; and Department of Pathology, NYU Medical School, 550 First Avenue, New York, New York 10016 /?eeeived July 15, 1994; accepted September 26, 1994 We have used cells infected with the HIV-1 molecular clone HX10 to study the binding of monoclonal antibodies (mAbs) to different epitopes within the extracellular domain of the HIV-1 transmembrane glycoprotein gp41. Gp41 mAb binding to the infected cells at 4 ° was variable but weaker than the binding of an anti-gp120/V3 loop mAb and increased substantially for three of the gp41 antibodies at 37 °, Treatment of the cells with soluble CD4 (sCD4) at 37 ° increased gp41 mAb binding to epitopes spanning residues 5 2 1 - 6 6 3 , implying that these regions had probably been masked by gp120, which following interaction with sCD4 had subsequently dissociated from gp41. By contrast, the binding of a mAb to residues 6 6 2 - 6 6 7 which form a neutralization epitope was reduced by sCD4 binding. Another region which has been described as containing a neutralization epitope spans residues 7 2 5 - 7 5 0 . MAbs to this region bound equally well to noninfected and HIV-infected cells, and binding was not increased in the presence of sCD4. These data strongly imply that this epitope is not exposed on the external surface of the membrane, a finding in accord with the proposed cytoplasmic localization of this region. © 1995Academic Press, Inc.
The env gene of the human immunodeficiency virus type 1 (HIV-1) codes for a single chain glycoprotein precursor, gp160, which is cleaved to yield the mature glycoproteins gp120 and gp41 (reviewed in 1). These two subunits assemble into a heterodimer held together by noncovalent interactions and are expressed at the virion surface in a functional, multimeric form (2, 3). Virus binding takes place via a high affinity interaction between gp120, the surface envelope glycoprotein, and the HIV cellular receptor, the 0D4 molecule (4, 5). Subsequent entry of the virion capsid into the cell takes place by fusion of the virus and cell membranes, an event mediated by the transmembrane glycoprotein, gp41 (6). The N-terminus of gp41 has been shown to contain a hydrophobic fusion domain on the basis of structural homology with the fusion domains of other enveloped viruses (7, 8), and mutagenesis of this region leads to fusion-incompetent virions (9, 10). The precise mechanism by which gp41 induces virus-cell membrane fusion is not yet clear, but certain steps in the process are being elucidated. The binding of CD4 to gp120 triggers conformational changes in the envelope glycoproteins which are thought to trigger fusion in a process termed receptor-mediated activation of fusion (11, 12). In cellline-adapted viruses the end point of these molecular rearrangements is dissociation of gp120 and exposure of gp41 (13- 16). Infection of 0D4 + cell lines with HIV results in a chronic infection in which the cell expresses no detect-
able CD4 but high levels of the HIV envelope glycoproteins (15, 16, 23). This gp120/41 is functional as demonstrated by 0D4 binding and the rapid formation of syncytia in cocultures of HIV + and CD4 + cells (1). A number of monoelonal antibody (mAb) epitopes have been characterized on gp41, including two which are recognized by neutralizing mAbs; one in the extracellular membrane proximal portion (17) and one in the cytoplasmic domain (18). Gp41 expressed at the surface of HIV-infected cells in its oligomerio, functional form was probed with a panel of epitope-mapped gp41 mAbs to determine which regions are obscured by gp120 and which are constitutively exposed on the surface of HIVinfected cells. The gp41 mAbs were obtained from the following sources: IAM 41-3E)6, IAM 41-4D4, IAM 41-2502, IAM41-2F5, and IAM 41-3H 12 are human mAbs obtained from Viral Testing Systems (Houston, TX); 50-69 and 98-6 are human mAbs which preferentially recognize the oligomeric form of gp41 which were prepared and epitope mapped as previously described (19); and 1583 and 1908 were prepared by immunization of mice with recombinant chimaeric poliovirus containing the HIV gp41 sequence 735-752 (20). The anti-gp120/V3 loop mAb 110.5 was obtained from Genetic Systems, Inc. (Seattle, WA), and the CD4 domain 4-reactive mAb L120 prepared by Becton-Dickinson, Inc. (San Jose, CA) was obtained from the MRC AIDS Directed Programme, UK. H9 cells were infected with the HX10 clone of HIV-1 and cultured for 8 days before use, as previously described (16). At this time the cells expressed no detectable cell surface CD4 and high levels of the HIV envelope
~To whom correspondence and reprint requests should be addressed. Fax, (33) 91269430. 713
0042-6822/95 $6.00 Copyright© 1995by Academic Press, Inc. All rights of reproductionin any form reserved.'
714
SHORT COMMUNICATIONS
glycoproteins as determined by immunofluorescent staining with the anti-gp120 mAb 110.5. One million cells in a volume of 100 #1 were incubated for 2 hr with or without soluble CD4 (sCD4 from R. Sweet, Smith, Kline and Beecham, King of Prussia, PA; 21) at 10 ffg/ml at 4 or 37 ° with agitation, washed, and then stained with the anti-gp41 mAbs for 1 hr at 4 or 37° with agitation before washing and fixation overnight in 2% formaldehyde. After further washing an appropriate dilution of anti-human immunoglobulin coupled to phycoerythrin (Immunotech SA, Luminy Case 915, Marseille, France)was incubated with the fixed cells for 1 hr at 4° with agitation followed by washing as before. Flow cytometric analysis was carried out using a Becton-Dickinson FAOScan with Consort 30 software. The interaction of sOD4 with gp120 of cell-line-adapted HIV-1 isolates at 37 ° induces the dissociation of gp120 from gp41 (13-16). In order to determine which regions of gp41 are masked by gp120 and which are constitutively exposed, a panel of gp41 mAbs with known epitopes was used to stain uninfected or HIV-infected, envexpressing H9 cells in the presence or absence of prebound sCD4. The background staining with uninfected cells was subtracted from the specific signal obtained with the infected cells. In the absence of sOD4, staining of HIV-infected H9 cells with the gp41 mAbs at 4° was variable but relatively weak by comparison with the signal obtained with the gp120 mAb 110.5 (Fig. 1). This low-level gp41 staining is in accord withprevious studies (15, 16, 22, 23) and probably reflects the constitutive partial exposure of gp41 molecules in the presence of gp120, but may also imply the exposure of a small proportion of gp41 molecules resulting from a basal level of spontaneous gp120 dissociation. Following the addition of sOD4 at 4° , staining increased subtly for all mAbs except the neutralizing mAb 2F5 (Fig. 1). This small increase in mAb reactivity is probably the result of low-level gp120 dissociation which occurs with some isolates of HIV-1 in the presence of sCD4 at 4° (16). The binding of sCD4 to the cells was verified by staining with the anti-CD4 mAb L120; gp120 binding does not interfere with L120 since this mAb reacts with domain 4 of 0D4 (15). Incubation of the cells with sOD4 at 37 ° resulted in a diminution of cell surface gp120-0134 complexes as demonstrated bythe decreased binding of the 110.5 and L120 mAbs. The reduction in cell-surface gp120 induced by sOD4 was estimated from the decrease in sOD4 staining using mAb L120 (calculated by dividing the value obtained at 4° with that obtained at 37 °) and the decrease in gp120 staining using mAb 110.5 (calculated by dividing the value obtained with untreated ceils by that obtained with ceils treated with sCD4 at 37 °) and was 2.3- and 2.8-fold, respectively (mean of two separate experiments). By this analysis the decrease in 110.5 staining was consistently greater than the decrease in sCD4 staining, probably resulting from a small
reduction in signal obtained with sCD4 treatment at 4° as a result of low-level shedding. The binding of all gp41 mAbs with the exception of 2F5 was increased substantially after sOD4 pretreatment at 37°; 2502, 50-69, 3D6, and 98-6 showed maximum increases (mean of two separate experiments) over sOD4-untreated cells of 6.4-, 2.8-, 2.7-, and 5.8-fold, respectively. The greater reactivity of these mAbs following sCD4 treatment at 37 ° is very likely to be a direct result of gp120 dissociation revealing previously occluded portions of gp41, since it has been shown previously that the increase in gp41 staining correlates with the level of gp120 dissociation (16). The variation between mAbs in the level of binding enhancement probably reflects the ratio of preexposed to sOD4-induced exposure of gp41 epitopes; 50-69 and 3D6 bind gp41 relatively well in the absence of sOD4-induced gp120 dissociation, whereas 2502 and 98-6 bind poorly. Since the gp41 mAbs interact with HIV virions in vivo at physiological temperature, we decided to test the binding of the gp41 mAbs to the infected cells at 37° (Fig. 2). The signal obtained at 37 ° with the mAbs 2502, 50-69, 3D6, 98-6, and 2F5 increased by 1.6-, 0.9-, 1.0-, 4.3-, and 10.6-fold, respectively (mean of two experiments) over that observed at 4 °. Thus, raising the temperature induced a dramatic increase in the exposure of the two membrane proximal epitopes (644-663 and 662-667) and a modest increase in the NH2-terminal epitope (521 538). In the presence of sCD4 prebound at 37°, mAb reactivity at 37 ° was 1.8-, 0.9-, 1.1-, 3.4-, and 6.9-fold greater than that at 4° for mAbs 2502, 50-69, 3D6, 98-6, and 2F5, respectively, implying that increasing the temperature for the mAb binding step did not substantially alter the level of sOD4-induced gp120 dissociation. Surprisingly, sCD4 binding induced a decrease in 2F5 binding which was detectable at 4° (1.3-fold, mean of two experiments) and became more obvious at 37 ° (2.1-fold, mean of two experiments). Thus unlike the other gp41 mAbs tested, the 2F5 epitope is constitutively strongly exposed on the surface of infected cells at 37 °, and mAb reactivity is reduced in the presence of sCD4. The temperature dependence of this epitope may relate to its availability within the gp41 oligomer; at 4° the epitope may be obscured by gp120, by carbohydrate groups, by another gp41 molecule, or by other membrane components. Increasing the temperature may allow the oligomeric structure to "relax" somewhat within the membrane, thereby freeing the epitope. The sCD4-induced reduction in 2F5 signal is more difficult to explain, but may result from a conformational modification of the epitope induced indirectly by sCD4 binding to gp120, directly via an interaction between sOD4 and gp41, or from epitope masking caused by a receptor binding-induced interaction of gp41 with other molecules in the infected cell or virion membrane. Binding of the 2F5 mAb to gp41 may inhibit HIV fusion by interfering with one the above- .
SHORT COMMUNICATIONS 80 -
0o8° 14o25C2 (521-538)
715
50-69 (579-613)
6o5° 1 3D6 (604-617)
60
40
40"
20
20
r"
e-
O) 0e.
0 ,1
1
10
.... , 1000
100
0
1
10
100
.... , 1000
0
1
,1
10
100
1000
0 (n 0
2011598-6 (644-663)
20-
2F5 (662-667)
>,
15-
C
~
3o0.
"i
250'
gp120 staining
200' 10'
150 ' 100'.
so
5
~
-
o 110.5
0
.... =
.... l
1
10
100
1000
1
10
100
L120 mAbs
1000
MAb concentration(nM) FIG. 1. Effect of sCD4 on gp41 mAb binding. The gp41 mAbs were incubated at the concentrations shown with HIV-l-infected H9 cells which hadbeenpretreated or not with sCD4 at 4 or 37°, followed by an anti-human immunoglobulin-phycoerythrin-coupled second antibody. MAb binding wasanalyzedby flow cytometry using a FACScan with Consort 30 software. Each datum point represents the mean fluorescence intensity of 10,000 accumulatedeveners,from which the signal obtained with uninfected cells has been subtracted= mAb binding in the absence of sCD4 is represented by(O), in the presence of sCD4 at 37° (Iq), and at 4° (A). Note that the vertical scale (mean fluorescence intensity) for 2502, 50-69, and 3D6 differs fromthat of 98~6and 2F5. The epitope position for each mAb is presented in parentheses after the mAb names and is numbered according to the BH10amino-acid sequence; minimum epitopes for 50-69 and 98-6 were taken from Xu et al. (19), 2F6 from Muster et al. (123, and 2502 and 3D6 from the commercial literature accompanying these mAbs. The boxed histogram shows staining by mAbs 110.5 (gp120 V3 loop) and L120 (0D4 domain4) at 10 #g/ml of cells from the same experiment described above carried out under the same conditions. Open (El) bars represent mAb staining of untreated, HIV-infected H9 cells; hatched ([]) bars represent staining of infected cells pretreated with sOD4 at 4°; and filled (B) bars representstaining of infected cells pretreated with sCD4 at 37°.
mentioned pathways or may function by yet another unknown mechanism. Figure 3 illustrates graphically the location of the epitopes which, from the results presented in this study, are exposed in the presence of gp120, and those which are likely to be c o m p l e t e l y or partially occluded. It is clear that the majority of the extracellular portion of gp41, at least that comprising residues 5 2 1 - 6 6 3 , is likely to be partially or c o m p l e t e l y masked in the presence of g p l 20. This is in good a g r e e m e n t with tlhe w i d e l y spaced diversity of mutations indicated in Fig. 3 w h i c h disrupt the gp120/CD4 interaction. The only mAb epitope w h i c h appears well exposed in the presence of gp120 is that of 2F5 which is the most m e m b r a n e proximal (662-667), although mAb binding is highly temperature dependent. The cytoplasmic d o m a i n of gp41 has been reported to contain a neutralizing epitope spanning residues 7 3 5 752 (18, 20). MAbs have been prepared in mice against recombinant polioviruses containing this region of HIV gp41, which neutralize in a relatively type-specific man-
ner (24). We have tested two of these reagents and a human mAb with similar epitope specificity for binding to cell surface-expressed gp41 in the presence or absence of sOD4. Figure 4 c o m p a r e s the binding of four mAbs to uninfected H9 cells and sOD4 pretreated or untreated HX10-infected H9. The gp120 mAb 110.5 gave a strong signal, and pretreatment of the cells with sOD4 reduced the signal by about threefold. MAb 98-6 to the extracellular domain of gp41 reacted w e a k l y with infected H9 cells in the absence of sOD4 and substantially more strongly in the presence of sCD4 preadsorbed to the cells at 37 ° . Two neutralizing anti-HIV ( 7 3 5 - 7 5 2 ) mAbs, 1583 and 1908 (20, 24), gave intermediate binding to H9 and gave an identical signal on infected H9 cells. Treatment with sOD4 resulted in a small diminution in staining in this experiment, but this w a s not a reproducible finding. Similar results w e r e obtained for 3H12, a human mAb reactive with an epitope within the gp41 s e q u e n c e 6 9 3 - 8 5 6 (data not shown). These three mAbs w e r e demonstrated to be active in a peptide ELISA; all
716
SHORT COMMUNICATIONS
100 "
A
80"
60"
4-, ¢h eI!1
40-
20
C 0
2502
tJ' t" 0 m
100 •
50-69
3D6
98-6
2F5
B
0
in HIV infection. (1) The immunodominant disulfide-linked loop and flanking regions which have been suggested to form a conserved structural motif which functions in the association of the surface and transmembrane glycoproteins of lentiviruses (26). Moreover, peptides synthesized from the sequence of this region bind to cellular proteins of 45 and 80 kDa, an association which has been suggested to represent a potential HIV-coreceptor interaction (27). (2) A predicted o~-helical region spanning residues 558-595, structural conservation of which is importantfor infectivity, and peptides from which inhibit HIV-mediated fusion (28). Antiserum raised against this peptide binds better to HIV-infected cells in the presence of sCD4, a finding which concurs with our suggestion that this region is at least partially masked by gp120. (29).
80.
t~
60'
40"
)-69 20-
3DE 0 25C2
50-69
Monoclonal
3D6
98-6
2F5
antibodies
FIG. 2. Temperature dependence of gp41 mAb binding. HW-infected H9 cells either untreated ( i ) or preincubated with sCD4 (10 #g/ml) for 1 hr at 37° ([~) were washed and incubated for 2 hr with the mAbs shown at 37° (A) or 4 ° (B). After overnight fixation the cells were stained with anti-human conjugated phycoerythrin and analyzed by flow cytometry as described in the legend to Fig. 1. The histogram bars represent the mean fluorescence intensity obtained from the analysis of 10,000 accumulated events from which the signal obtained with uninfected H9 cells has been subtracted.
~mino ~rminus 2
gave strong signals with the peptide 735-752, but no significant activity with an irrelevant peptide (data not shown). It thus appears that the gp41-specific epitope(s) for these mAbs is not expressed at the surface of HIVinfected cells. This is in accord with the predicted intracellular location of the cytoplasmic tail of gp41, there being no evidence for a second transmembrane spanning region. These findings differ from results presented in another study (25), in which HIV-infected cells reacted more strongly than uninfected cells with an anti-735-752 peptide mAb. The reason for this discrepancy is unclear, and the mechanism of neutralization of HIV by these mAbs remains obscure. The HIV-cell fusion process is thought to be mediated by a direct interaction between the hydrophobic N-terminal fusion domain of gp41 and the cell membrane (1). Our current thinking suggests that this region is probably masked by gp120 until the virus and cell membranes are in close apposition, when a process termed receptormediated activation of fusion triggers the exposure of the fusion domain (11, 12). In addition to the fusion domain, there are other regions which are potentially functional
Extracellular
carboxy terminus ,~.
Intracellular JD
%®®-
FIG. 3. Regions of gp41 masked by gp120. The amino-acid backbone and proposed secondary structure are modified from Gallaher et al. (33). The thick black line represents the regions of gp41 which are likely to be partially or completely masked by gp120 in the functional giycoprotein oligomer. It is unclear from this study whether the Nterminal portion of the fusion peptide is masked by gp120, since the epitope of mAb 2502 is reported to begin at residue 521, and no reagents specific for the region N-terminal to this have been tested. The arrowheads, some of which are numbered according to the published sequence of HIV-1 BH10, indicate residues which are thought to be critical for gp120 association (taken from Cao et aL, 30). The epitope for each of the mAbs tested is boxed and labeled, and the epitope for mAb 2F5 (662-667) is shaded.
SHORT COMMUNICATIONS
717
3. HALLENBERGER,S., TUCKER, S. P., OWENS, R. J., BERNSTEIN, H. B,, and COMPANS, R. W. Virology 193, 510-514 (1993). 4. McDoUGAL, J. S., KENNEDY, M. S., SLIGH, J. M., OORT, S. P., MAWL, A., and NICHOLSON, J. K. A., Science 231, 382-386 (1986). 5. LASKY, L. A., NAKAMURA, G. M., SMITH, D. H., FENNIE, O., SHIMASAKi, C., PATZER, E., BERMAN, P. W., GREGORY,T., and CAPON, D. J., Ceil I!. //r ,~S....~... \ 50, 975-985 (1987). 6. KOWALSKI, M., POTZ, M., BASIRIPOUR, T., DORFMAN, T, GOH, W. C., 3:3 21 ."'...... "'.. ................ $ :,,:. . . . . . . . \ ~,.....; TERWILLIGER, E., DAYTON, A., ROSEN, C., HASELTINE, W. A., and 100 101 102 103 SODROSKI, J., Science 237, 1351-1355 (1987}. r7. GALLAHER, W. R., Cell 50, 327-328 (1987). 8. GONZALEZ-SOARANO, F., VVAXHAM, M. N., ROSS, A., and HOXlE, J. A., 0 AIDS Res. Hum. Retroviruses 3, 245-282 (1987). I !i. 9. FREED, E. O., MYERS, D. J., and RISSER, R., Proc. Natl. Acad. Sci. USA ~ -'.1\ 87, 4650-4654 (1990). .f" I ' 4 q 10. BERGERON, L,, SULLIVAN, N., and SODROSKI, J., J. Viro/. 66, 2389-2397 ~! ~(1992). i'[ ~. ,,{. ,i.. 11. ALLAN, J. S., Science 252, 1322 (1991). ....... ,~,4.., ' ' '"'~1 ........ . . . . . . . . 12. MOORE, J. P., MCKEATING, ]. A., WEISS, R. A., OLAPHAM, P. R., and 101 102 103 10° 101 102 103 SATIENTAU, O. ]., Science 252, 1322-1323 (1991). 13. HART, T. K., KIRSH, R., ELLENS, H., SWEET, R. W., LAMBERT, D. M., Mean fluorescence intensity PEqq-EWAY, S. R., LEARY, J., and BUGELSKI, P. J., Proc. Nat/. Acad. SeL USA 88, 2189-2193 (1991). FIG. 4. Staining of HIV-infected and uninfected cells by mAbs to gp41 14. MOORE, J. P., MOKEATING, ]. A., WEISS, R. A., and SAIq-ENTAU, O. J., cytoplasmic domain. HIV-l-infected H9 cells either pretreated (---) or Science 250, 1139-1142 (1990). not (. • .) with sCD4 and uninfected H9 cells (. • .) were incubated with 15. SATTENTAU, Q. J., and MOORE, J. P., J. Exp. Med. 174, 407-415 (1991 ). fluoreehrome conjugated second antibody alone ( - - ) or with the . . . . 1 16, SAI]-ENTAU, Q. J., MOORE, J. P., VIGNAUX, F., TRAINCARD, F., and PO~Gantibodies shown at a d dut~on of either ~-~ for ascites or a concentration NARD, P., ./. VlroL 67, 7383-7393 (1993). of 30 fig/el for purified immunoglobulin for 2 hr at 4 ° with agitation, 17, MUSTER, T., STEINDL, F., PURTSCHER, M., TRKOLA, A., KLIMA, A., HIMNi~ washed, and then fixed. After a further wash the cells were labeled LER, G., RUOKER, F., and KATINGER, H., J. VlroJ. 67, 6642-6647 with phycoerythrin-coupled second antibodies and then analyzed by (1993). flow cytometry on a FAOScan using Consort 30 software. Each peak 18. OHANH, T. C., DREESMAN, G. R., KANDA, P., LINE~FE, G. P., SPARROW, represents the accumulation of 10,000 gated events. Note that the J. T., HO, D. D., and KENNEDY, R. C., EMBOZ 5, 3065-3071 (1986). staining of uninfected and HIV-infected cells with mAbs 1583 and 1908 19. XU, J. Y., GORNY, M. K., PALKER,T., KARWOWSKA, S., and ZOLLA-PAZNER is superimposed in this experiment. S., J. Viro/. 65, 4832-4838 (1991). 20. EVANS, D. J., McKEATING, J., MEREDITH, J. M., BURKE, K. h., KATRAK, K., JOHN, A., FERGUSON, M., MINOR, P. D., WEISS, R. A., and ALMOND, (3) The coflserved neutralization epitope located close to J. W., Nature 339, 385-388 (1998). the membrane on the extracellular segment of gp41 (17). 21. DEEN, K. C., McDOUGAL, J., INACKER, R., FOLENA-WASSERMAN, G., ARTHOS, J., ROSENBERG, J., MADDON, P. J., AXEL, R., and SWEET, The data presented here imply that gp41 and gp120 interR. W., Nature 331, 82--84 (1988). act over a large surface, a conclusion which is not sur22. TYLER, D. S., STANLEY, S. D., ZOLLA-PAZNER, S,, GORNY, M. K., SHADprising considering the variety of mutations in both subDUCK, P. P., LANGLOIS, A. J., MAT]HEWS, T. ]., BOLOGNESI, D. P., units which interfere with this interaction (6, 30-32, Fig. PALKER, m. J., and WEINHOLD, K. J., J. Immune/. 145, 3276-3282 (1990). 3). Further studies using viral glycoproteins expressed 23. SPEAR, G. T., TAKEEMAN, D. M., SULLIVAN, B. L., LANOAY, A. L, and in a functional form at the cell surface may be useful for ZOLLA~PAZNER, S., J. Virol. 67, 53-59 (1993). the dissection of other structure-function relationships. 24. VELLA, O., FERGUSON, M., DUNN, G., MELOEN, R., LANGEDUK, H., EVANS, D., and MINOR, P. D., J. Gen. Virol. 74, 2603-2607 (1993). 25. DALGLEISH, A. G., CHANH, T. C., KENNEDY, R. C., KANDA, P., CLAPHAM, ACKNOWLEDG M ENTS P. R., and WEiss, R. A., Virology t65, 209-215 (1988). 26. SCHUL]2, T. F., JAMESON, B. A., LOPALCO, L., SICCARDI, A. G., WEISS, We thank H..Holmes and A. Newberry and the MRC AIDS Directed R. A., and MOORE, J. P., AiDS Res. Hum. Retroviruses 8, 1571Programme for reagents and time and K. Simon and A. Daser for their 1580 (1992). help with the manuscript. This work was supported by the CNRS, IN27. HENDERSON, L. A., and QUERSH~, M. N., .L BioL Chem. 268, 15291SERM, and a grant from the ANRS of France and in part by NIH Grants 15297 (1993). A1 32424 and A1 36085 and research funds from the Department of 28. WILD, C., OAS, T., MCDANAL, C., BOLOGNESI, D., and MAWHEWS, T., Veterans Affairs, NY. Proc Natl. AcacL ScL USA 89, 10537-10541 (1992). 29. MA]-FHEWS,T. J.,/n "Proceedings of the Seventh Colloque des Cents Gardes" (M. Girard and L. Valette, Eds.), pp. 117-121 (1992). REFERENCES 30. OAO, J., BERGERON, L., HELSETH, E., THALI, M., REPKE, H., and SODROSKI, J., Z VIroL 67, 2747-2755 (1993). 1. MOORE, J. P., JAMESON, B. A., WEISS, R. A., and SATTENTAU, Q. J., /n 31. HELSETH, E., OLSHEVSKY, U., FURAN, O., and SODROSKI, J., Z ViroL 65, "Viral Fusion Mechanisms" (J. Bentz, Ed.), pp. 233-289. CRC 2119-2123 (1991). Press, Boca Raton, FL, 1993. 32. IVEY-HOYLE, M., CLARK, R. K., and ROSENBERG, M., J. VlroL 65, 26822. PINTER, A., HONNEN, W. J., TILLEY, S. A., BONA, C., ZAGHOUANI, H., 2695 (1991). GORNY, M. K., and ZOLLA-PAZNER, S., J. ViroL 63, 2674-2679 33. GALLAHER, W. R., BALL, J. M., GARRY, R. F., GRIFFIN, M. C., and MONTE.LARO, R. C., AIDS Res. Hum. Retroviruses 5, 431-434 (1989). (1989).
110.5
E
~i
98-6
/)ii
100 101 102 103 1908 1583
I
100
I
,
, ......
I
.......
,I
.......