INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The principal mechanism by which effective viral vaccines confer protective immunity is by induction of antibodies capable of neutralizing prevalent strains of virus (21). Efforts to develop a vaccine to protect against human immunodeficiency virus type 1 (HIV-1) infection are complicated by the fact that virus strains in infected patients tend to be highly resistant to neutralization. Studies that clarify the mechanisms responsible for this neutralization resistance may provide important leads regarding possible methods for the induction of potent neutralizing antibody responses capable of neutralizing these primary isolates. Previously, we described the selection and characterization of a neutralization-resistant mutant of the MN strain of HIV-1, which generally resists neutralization by human sera. The resistance phenotype was found to be determined by polymorphisms located in the C-terminal region of gp120 and N-terminal region of gp41 (16, 17). There were two residues in gp120 and four in gp41 involved. Genetic analysis demonstrated that all six residues participated in a complex interaction that was responsible for the neutralization resistance phenotype. Five of the same mutations were also responsible for a high-infectivity phenotype. The two gp120 mutations were located at the base of the CD4 binding pocket and the center of the putative coreceptor binding domain (14, 32). These critically located gp120 residues interact functionally with residues in the leucine zipper domain of gp41 to modulate neutralization resistance and infectivity. Additional findings indicated that these phenotypes were apparently the result of increased efficiency of the overall fusogenicity of the envelope protein (16, 17).

In this study, we report additional characteristics of the HIV-1 MN envelope, which reveal properties associated with the neutralization resistance and high-infectivity phenotypes. Resistance to neutralization extended to monoclonal antibodies (MAbs) directed against multiple gp120 epitopes. Conformational changes affecting exposure of epitopes in gp120 and gp41 were particularly associated with one of the mutations responsible for neutralization resistance, suggesting that two or more distinct types of alterations of structure-function relationships are likely to account for the phenotypes. The results are consistent with the interpretation that neutralization resistance and high infectivity are the result of enhanced efficiency of one or more aspects of the envelope functions involved in the fusion process.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

MAbs. The human MAbs specific for the CD4 binding site (CD4bs), the CD4-induced (CD4i) epitope, and the V3 loop, (15e, 48d, and 19b, respectively) were gifts from James Robinson (1, 4, 15, 26, 33). The human anti-V3 MAbs 391-95D, 694-98D, and 447-52D and the anti-gp41 MAbs 1281D, 50-69, and 126-6 have been described (9, 7, 10, 13, 34). The following antibodies were obtained through National Institutes of Health AIDS Research and Reference Reagent Program (NIH-ARRRP): anti-V3 mouse MAb R/V3-50.1 (provided by Repligen Corporation [8]), the anti-V3 loop human MAbs 257D and 268D and the human anti-gp41 MAbs 56-69 and 126-6 (provided by Susan Zolla-Pazner [9, 11, 12]), the human anti-CD4bs MAbs b12 (provided by D. Burton and C. Barbas [3, 6, 23]), and F105 (provided by M. Posner [18, 19]), the human anti-CD4i MAb 17b (provided by J. Robinson [26, 33]), the human anti-gp120 antibody 2G12 (provided by H. Katinger [5, 27]), and the human anti-gp41 antibody 2F5 (provided by H. Katinger [5, 20]).

Mutant envelope gene plasmids. The MN-V5 and E6 envelope gene expression plasmids and most of the mutant V5 envelope genes used in this study have been previously described (16, 17). The additional mutant envelope gene, chimera J, was constructed by digesting two chimeric env clones that had been constructed previously, chimeras C and D, with restriction endonucleases BamHI and Bsu36I (17). A fragment in the N terminus of the env gene, located between BamHI and Bsu36I, of chimera D was replaced with the corresponding sequence of chimera C. A map of chimera J is shown in Fig. 1. It consists of sequences derived from the V5 clone, except for E6 sequences coding for the N terminus of gp120 through the V2 region, and the regions of gp120 and gp41 which carry the neutralization resistance mechanisms. This gene was constructed to allow us to determine whether differences between E6 and V5 envelopes, which were not attributable to the neutralization resistance mutations, were attributable to mutations in the V1-V2 region.

View larger version (14K): [in this window] [in a new window]   FIG. 1.   Chimera J constructed by exchanging fragments of previously constructed env genes (see Materials and Methods) (16, 17). (A) Schematic diagram of the HIV-1 env gene. Locations are indicated for the coding sequences for the cleavage site between gp120 and gp41, restriction enzyme recognition sequences, and landmarks in mature envelope proteins (V1 to V5, variable regions of gp120; F, fusion domain; LZ, leucine zipper-like alpha-helical domain; AH, membrane proximal alpha-helical domain; TM, transmembrane domain; C, cytoplasmic domain). (B) Mutant env genes. The V5 gene encodes the neutralization-sensitive envelope characteristic of the laboratory adapted MN virus. The E6 gene encodes the MN variant selected in vitro with the neutralization-resistant, high-infectivity phenotype. Chimera J comprises the N-terminal region from E6 which includes the coding sequences for the V1-V2 region, the junction region between gp120 and gp41 from the E6 clone (which encodes the neutralization resistance phenotype), and other regions from V5. (C) Specific mutations which are responsible for the E6 neutralization resistance phenotype.

Pseudovirus preparation and neutralization assays with MAbs and sCD4. Pseudoviruses expressing envelope glycoproteins derived from various env plasmids were constructed as described before (7a, 16, 17, 35). Briefly, 293T cells were cotransfected with env-expressing plasmids and with the complementing viral genome-reporter gene vector, pNL4-3.Luc.ER. Cell culture fluid harvests containing pseudoviruses were used to infect PM1 cells, and the luciferase activity of infected cells was measured after 48 h. For neutralization assays, pseudovirus suspensions were diluted appropriately, and 25 µl of each pseudovirus suspension was incubated with 25 µl of serially diluted MAbs for 1 h in triplicate in 96-well plates. The pseudovirus dilutions were chosen to yield input inocula which gave luminescence generally between 10 and 100 times the background, in the absence of neutralization. The virus-MAb mixtures were used to infect PM1 cells, and the neutralization endpoint was determined for 90% reduction of luminescence compared to the control wells in which the same amount of pseudovirus was incubated with HIV-1-negative human serum. Neutralization of pseudoviruses by sCD4 was tested similarly using serially diluted sCD4 instead of MAbs. Soluble CD4 was obtained through NIH-ARRRP (provided by Ray Sweet, SmithKline Beecham [30]).

Flow cytometric analysis of envelope glycoproteins expressed on transfected cells with MAbs against various epitopes. The binding of MAbs to cells expressing different envelope glycoproteins was analyzed by flow cytometry. The 293T cells were transfected with plasmids by the same method used for pseudovirus preparation. Cells were harvested 48 h after transfection and washed twice in prechilled phosphate-buffered saline (PBS), and 5 × 10⁵ cells were used for staining with MAb. The cells were resuspended and incubated in 100 µl of each MAb, appropriately diluted. Cells were washed twice with washing solution (PBS containing 10% fetal bovine serum [FBS] and 0.05% sodium azide) and incubated with biotinylated anti-human or anti-mouse antibodies (0.5 µg/ml). After being washed, cells were stained with streptavidine-phycoerythrin conjugate (Sigma). Each reaction was performed in 100 µl of solution at room temperature for 20 min. Cells were finally fixed in 2% formaldehyde prepared in PBS and analyzed on a Coulter EPICS XL flow cytometer analyzer. For negative controls, cells incubated with PBS or isotype control antibodies, mouse immunoglobulin G2a (IgG2a) or human IgG1 (Sigma), were used. To test the effect of sCD4 binding on cell staining with MAbs, the same number of cells were preincubated with 100 µl of sCD4 at 10 µg/ml, washed three times with cold PBS, and then incubated with the same amount of each MAb. To calculate the percentages and median fluorescence intensity values for the positively stained cells, values for the negative control were subtracted from those of each sample as shown in panels B, D, F, H, and J of Fig. 2. Subtraction analysis was performed by WinList for Win32 4.0 (Verity Software House, Inc.) (2).

View larger version (42K): [in this window] [in a new window]   FIG. 2.   Comparative binding of the anti-V3 MAb, 257D, and the anti-gp41 MAb, 50-69, to cells expressing envelope glycoproteins of MN strain variant clones V5 and E6. Transfected 293T cells were prepared as described in Materials and Methods and analyzed for immunofluorescence by using flow cytometry. The percent fluorescence-positive cells (% Cells +) and median cell fluorescence (MCF) were calculated after background subtraction by using WinList software (2). Histograms are shown for total cell populations (unfilled areas) and positive cell populations (filled areas).

gp120 dissociation assay and ELISA. Spontaneous and ligand-induced gp120-gp41 dissociation was assessed by enzyme-linked immunosorbent assay (ELISA) using methods for virus processing described previously (16, 24). Pseudoviruses were filtered, sedimented by centrifugation of transfected cell culture supernatants at 15,000 rpm for 2 h (Tomy Tech USA, Inc.), washed twice with prechilled PBS by repeated centrifugation, and resuspended in PBS with 10% FBS in 1/40th of the initial volume (16). Each aliquot of concentrated pseudovirus was incubated at 37°C for 1 h with MAb solution at the same concentration that was used for the fluorescence-activated cell sorter analysis. Replicate pseudovirus suspensions were incubated with 10 µg of sCD4 or PBS per ml. After exposure to ligand for 30 min at 4°C, pseudovirus particles were separated from dissociated gp120 by centrifugation. The level of gp120 dissociation was determined by comparing gp120 antigen in the samples of the supernatants and pellets measured by ELISA. The amounts of p24 antigen in both supernatant and pellet samples were also measured. The ELISA assay was conducted by antigen capture, as described previously (16). Briefly, each well of the microtiter plates was coated with a human anti-HIV-1 IgG; then antigen prepared in lysis buffer and diluted in blocking reagent was applied, and bound antigen was detected by using either sheep anti-gp120 or rabbit anti-p24 antibody. Bound detection antibodies were assayed by using biotinylated anti-sheep or anti-rabbit antibody, followed by avidin-conjugated horseradish peroxidase, and then orthophenylenediamine substrate development. Standard antigen controls used in the assays consisted of serial dilutions of p24 and MN strain gp120, each obtained from the NIH-ARRRP (catalog numbers 3927 and 382, contributed by ImmunoDiagnostics, Inc., and M. Quiroga, respectively).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Phenotypes of chimera J. The neutralization resistance and infectivity phenotypes of the envelope protein encoded by chimera J were tested using pseudotyped virus, as described previously (16, 17). The neutralization resistance, tested with neutralizing human sera, and the infectivity of chimera J were similar to those properties of the E6 clone (16). Chimera J was 5- to 10-fold more infectious and 128-fold more resistant to neutralization than clone V5 (data not shown).

Neutralization of pseudoviruses expressing V5 and E6 envelope glycoproteins by MAbs. The concentration of each MAb which inhibited 90% of the viral infectivity was determined. Consistent neutralization results were obtained in each independent experiment, and the average results of experiments testing each MAb are shown in Table 1. (The amounts of MAbs 17b and 2G12 available were only sufficient for single tests.) Of the MAbs tested, most of the antibodies targeted against the V3 loop neutralized the HIV-1 MN V5 strain pseudovirus. In contrast, pseudovirus bearing the E6 envelope was less sensitive to neutralization by most of the anti-V3 antibodies. The MAb 19b neutralized V5 and E6 at comparable concentrations. Although these concentrations were somewhat higher than the concentrations of other anti-V3 MAb required to neutralize V5 pseudovirus, the absence in difference in sensitivity of this V5 and E6 to this MAb clearly distinguished it from other MAb tested. Most of these antibodies recognize contiguous sequences of amino acids located near the apex of loop, while 19b recognizes a discontinuous epitope spanning the apex of the V3 loop.

Binding of MAbs to envelopes measured by flow cytometry. The binding of MAbs to envelope glycoproteins was measured by immunofluorescence using flow cytometry. The success of transfections was monitored for each infectious clone in each experiment by comparing the infectivity titers of pseudoviruses produced from the same transfections. In almost all cases the proportional infectivity of the various clones was similar to that previously described. In a few cases the infectivity of specific clones was significantly lower than expected, and the experiments were repeated. In each case the infectivity in the repeat experiment was similar to that expected. Comparative binding patterns of the V5 and E6 envelopes with selected MAbs are presented in Fig. 2. The results obtained with isotype control antibodies were not significantly different from negative controls, which were incubated with PBS instead of primary antibody solution (data not shown).

MAb epitope accessibility on envelopes containing mutation(s) of E6. The effect of each mutation which determined the neutralization phenotypes of the E6 and V5 clones on the binding of MAbs to cells expressing each envelope was evaluated. The median fluorescence intensity of positively stained cells was compared among different clones for each MAb. The relative median fluorescence intensity obtained with each clone is expressed in proportion to that of V5. Results are presented in Fig. 3. The mutant clones were tested with three anti-V3 MAbs and two MAbs each against CD4bs and gp41 epitopes. They were also tested on multiple occasions with the anti-CD4i mAb, 48d (shown in Fig. 3), and once with MAb 17d (not shown). Each envelope reacted similarly with each MAb in the multiple experiments performed, and the 48d and 17b MAbs reacted similarly in the comparative experiment. The only point mutation which resulted in significantly reduced binding of all MAbs against gp120 epitopes was L544P. The magnitude of the reduction in binding of anti-V3 and -CD4bs MAb to this mutant was not as great as the reduction in binding to E6, compared to V5.

View larger version (29K): [in this window] [in a new window]   FIG. 3.   Comparative binding of MAb to mutant MN strain envelopes. The E6 clone is neutralization resistant. The six mutations which encode this phenotype were introduced singly or in combinations into the neutralization sensitive V5 clone. Chimera J incorporates the neutralization resistance mutations and the V1-V2 regions from E6 into the V5 sequence (see Fig. 1). Immunofluorescence analyses were performed by using flow cytometry, as described in Materials and Methods. Results are expressed as ratios: median cell fluorescence (MCF) obtained in testing the indicated clones/MCF for V5 tested in parallel. Each point indicates the result of an individual experiment. The results shown for the antibody 48d were obtained after preincubation of the transfected cells with sCD4 at 10 µg/ml at 4°C for 30 min.

sCD4 neutralization and effects on binding of envelopes by MAbs. Neutralization of V5 and E6 envelopes by sCD4 was tested. Unlike the effects of MAb 15e or b12, which target epitopes overlapping with the CD4bs, sCD4 neutralized V5 and E6 pseudoviruses equivalently. In each case 90% inhibition of infectivity was achieved at 50 ng/ml. The effect of a subneutralizing concentration of sCD4 (16 ng/ml) on neutralization by MAbs 257D, b12, and 48d was examined. No change in the neutralizing activity of each MAb against V5 or E6 was observed to be caused by preincubation with sCD4 (data not shown).

gp120-gp41 dissociation. The dissociation of gp120 shedding from the envelope glycoprotein complex was tested. There was significant spontaneous gp120 dissociation observed in the control samples which were preincubated in PBS with no ligands (Table 4). Preincubation with the 257D MAb caused no significant additional dissociation compared to the control samples. However, gp120 was dissociated significantly from the envelope complex when pseudoviruses were preincubated with sCD4 (10 µg/ml). There was no difference in the dissociation inducing effect of sCD4 on the V5 and E6 pseudoviruses. These findings support the interpretation that the differences in immunofluorescence reactivity of MAb with the different envelopes reflected differences in MAb binding rather than differences in MAb-induced gp120-gp41 dissociation.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

We previously described a neutralization escape mutant variant of the HIV-1 MN strain envelope which is broadly resistant to neutralization, compared to the MN laboratory adapted strain, by all human sera tested. This variant, designated clone E6, contains a number of mutations, six of which are responsible for the neutralization resistance phenotype (17). Two of the responsible mutations are located in the carboxy terminus (C4 and V5 regions) of gp120, and four are located in the amino terminus (leucine zipper region) of gp41. Five of these mutations, excluding the L544P mutation in gp41, are also responsible for a high-infectivity phenotype of the neutralization resistant clone, E6, compared to the neutralization sensitive clone, V5. Both phenotypes appear to be the result of enhanced efficiency of one or more functions of the envelope involved in the infection process (16). In the present study we attempted to elucidate the nature of the functions responsible for the high-infectivity and neutralization resistance phenotypes. The results demonstrated that the variant was resistant to neutralization by MAbs directed at multiple V3, CD4bs, and CD4i gp120 epitopes, that the L544P mutation impaired capacity of the gp120 to bind neutralizing MAb and increased the capacity of gp41 to bind MAb, and that resistance did not involve altered sensitivity to sCD4 effects. At least two mechanisms for alteration of neutralization resistance and infectivity appeared to be involved, one which affects conformation and one which does not, both of which probably modulate events which follow envelope-CD4 binding.

The neutralization-resistant variant, E6, was selected by growth in the presence of a human serum, the neutralizing activity of which could be completely blocked by synthetic V3 peptide (16, 17). Nevertheless, the variant was resistant to neutralization by human sera with or without demonstrable anti-V3 neutralizing activity (16, 17). The global neutralization resistance phenotype, which was suggested by these earlier studies, was confirmed in the present study. The E6 envelope was significantly more resistant to neutralization than the V5 envelope by MAb against V3, CD4bs, and CD4i epitopes. The MN strain of HIV-1 is a T-cell-line-adapted, syncytium-inducing, CXCR4-dependent virus. The extent to which neutralization resistance mechanisms reported here pertain to primary, R5-dependent HIV-1 isolates is, therefore, uncertain. However, the global neutralization resistance phenotype we have found is similar to that found in primary isolates, and the selection of the neutralization resistant variant using polyclonal human serum is similar to that which occurs in vivo.

Others have reported previously the association of resistance to neutralization by MAbs against multiple gp120 epitopes with a point mutation in gp41 at the residue corresponding to residue 583 in our clones (22, 25, 28, 29, 31). Since this residue is adjacent to residue 582, one of the mutations contributing to the neutralization resistance in our clone, it is likely that the mechanisms of effect of these mutations is similar. This sort of global neutralization resistance, which we saw in the E6 clone, is well known to be characteristic of primary patient isolates of HIV-1.

MAbs against gp120 neutralization epitopes bound the E6 envelope expressed on infected cells less well than the V5 envelope, while the reverse was true of MAb against gp41 epitopes. Our previous data demonstrated that the amount of gp120 produced by V5 and E6 transfected cells and the amount incorporated into pseudoviruses produced by these cells were equivalent (16). A number of the transfected cell cultures used in the present study for immunofluorescence assays were also analyzed by ELISA for gp120 expression, as was done in the previous study (data not shown). The comparability of gp120 production was confirmed in these assays. These findings indicate that the differences in MAb binding to cells expressing V5 and E6 did not reflect differences in protein expression. The results shown in Table 4, demonstrating equivalent spontaneous dissociation of gp120 and gp41 from V5 and E6 and the lack of effect of anti-V3 MAb on dissociation in either case indicate that differential gp120-gp41 dissociation is not the explanation for the differential MAb reactivity with the two envelopes. It is likely, therefore, that the differential MAb binding results from differences in conformation of the two envelopes that affect access of these MAb to their cognate epitopes.

The possible effects of conformational changes on neutralization resistance or infectivity were studied by examining the effects of mutations responsible for the latter two phenotypes on MAb neutralization and binding. Among the six mutations in E6 that we have shown previously to be responsible for the neutralization resistance phenotype, the V420I and L544P mutations have the greatest resistance enhancing effects (16). With regard to conformational changes in this study, the V420I mutation had no detected effect, while the L544P mutation was the only one that affected the binding of all MAbs tested. Two of the other mutations had lesser effects on MAb binding, and the remaining two had no measured effects. Thus, neutralization resistance is mediated by mutations, some of which do and some of which do not induce conformational changes affecting binding of neutralizing MAb.

The dominant role of the L544P mutation in determining the conformational change associated with neutralization resistance was particularly evident upon comparison of the V5 clone containing five of the six neutralization resistance enhancing mutations and the clone containing all six mutations. The five-mutation clone bound MAb equivalently to the V5 clone, while all of the conformation change attributable to the six mutations resulted from introduction of the L544P mutation into the five-mutation clone (Fig. 3). The mechanism by which the L544P mutation mediates conformational change is probably related to the functional interactions between gp41 mutations revealed in our previous studies. This functional interaction affected the interaction between gp41 and gp120. The possibility is suggested that the L544P mutation causes a conformational change in gp41 that requires a transition, such that there is a change in the specific residues in gp120 with which individual residues in the gp41 leucine zipper bond noncovalently. If this interpretation is correct, the mutations at the 562, 565, and 582 residues may be needed to allow the transition to proceed. It would not be surprising if changes in bonding over such a large area would affect conformation of regions of gp120 which associate with the gp41 leucine zipper region.

Our previous and present data support the interpretation that the high infectivity resulting from neutralization resistance mutations in clone E6 is determined by increased efficiency of envelope functions mediating the infection process. Furthermore, the infectivity differences did not appear to be the result of differences in sensitivity to effects of CD4 binding, since no differences between V5 and E6 were seen in assays of sCD4-induced gp120-gp41 dissociation, sCD4 neutralization, or post-sCD4-binding exposure of gp120 epitopes. The absence of evidence of differential responsiveness to sCD4 increases the likelihood that the mechanisms of enhanced infectivity and neutralization resistance involve a post-receptor-binding component(s) of the infection process.

The comparative reactivity of MAb 19b with the neutralization resistant and sensitive envelopes is of interest. This antibody recognizes a conformation-dependent epitope at the tip of the V3 loop. This was the only anti-gp120 neutralization epitope MAb which we tested that neutralized pseudoviruses expressing the V5 and E6 envelopes equivalently. Binding of this antibody to E6 was reduced compared to V5. However, binding to the V5 clone with the six neutralization resistance mutations was not reduced compared to nonmutated V5. Thus, the full neutralization resistance phenotype was not associated with either reduced neutralization or binding by the 19b antibody. This MAb has been reported to be active in neutralization of primary isolates of HIV-1, although conflicting data have also been reported (8, 15). The variability of these reported data may reflect antigenic variation in the epitope rather than the presence or absence of exposure of the epitope. Our own experience indicates that some primary isolates are and some are not neutralized by this MAb (unpublished data). In a separate study, our laboratory has found that antibodies against the same or a very similar conformational V3 epitope can mediate primary virus cross-reactive neutralization (P. F. Zhang et al., unpublished data). The global neutralization resistance we evaluated in the present study involves resistance to neutralization by antibodies directed against CD4bs and CD4i epitopes and some epitopes in V3. The selective retention of sensitivity to neutralization by antibodies directed against a conformation-dependent V3 epitope may be a common feature of the global neutralization resistance phenotype and an important lead to understanding the mechanisms of neutralization resistance of primary HIV-1 strains.