We characterized HIV-1 Envs cloned from a series of well characterized primary PA-R5 and A-R5 viruses. These included a cross sectional panel of four PA-R5 viruses (NB23, NB24, NB25 and NB27) and four A-R5 HIV-1 viruses (NB2, NB6, NB7 and NB8) [34,36], as well as PA-R5 and A-R5 viruses isolated sequentially from one subject (IK1) [13,42] (Table 1). All viruses are of R5 phenotype, but compared to the PA-R5 viruses the A-R5 viruses have enhanced M-tropism, reduced CD4- and CCR5-dependence, reduced sensitivity to inhibition by HIV-1 entry inhibitors and RANTES, and increased sensitivity to neutralization by the Env mAb IgG1b12 [13,34,36,42]. Thus, the A-R5 HIV-1 isolates have unique biological properties distinguishing them from the PA-R5 isolates that serve to enhance Env-receptor interactions, and most likely map to the env gene. However, it is important to note that the PA-R5 virus from subject IK1 was isolated just prior to the onset of CD4+ T-cell loss and progression toward AIDS [13], whereas the PA-R5 viruses from the cross sectional panel were isolated at earlier stages of HIV-1 infection, including one virus isolated from an acute seroconverter [34]. Therefore, although all the A-R5 viruses were isolated from patients with AIDS, there is considerable heterogeneity among the PA-R5 viruses with respect to the stage of asymptomatic HIV-1 infection from which they were isolated.

To identify viral determinants which underlie the unique biological properties of A-R5 HIV-1 viruses and may contribute to the pathogenesis of R5 HIV-1 variants, the env gene was cloned into the pSVIII-HXB2 Env expression vector using KpnI and BamHI restriction sites. Three to four independent and functional Envs cloned from each virus were identified by single round entry assays in JC53 or Cf2-CD4/CCR5/CXCR4 cells using Env-pseudotyped GFP reporter virus, and by fusion assays (Table 1, and data not shown). Western blot analysis of Env expression in transfected 293T cells showed distinct gp160 and gp120 proteins in 36/37 primary Envs, similar to the control R5 ADA, YU2, JRFL and JRCSF Envs (Fig. 1). To determine the coreceptor specificity of the cloned Envs, Env-pseudotyped GFP reporter viruses were used in single round entry assays with Cf2th cell lines stably expressing CD4/CCR5 or CD4/CXCR4 (Table 1). The X4 HXB2, R5 ADA, and R5X4 89.6 Envs were used as positive controls. A non functional Env, ΔKS Env, was used as a negative control to determine background levels of GFP expression. As expected, HXB2 Env used CXCR4, ADA Env used CCR5, and 89.6 Env used both CXCR4 and CCR5 for HIV-1 entry. All 37 primary Envs used CCR5 for HIV-1 entry, similar to the coreceptor specificity of the primary isolates from which they were cloned. Thus, we established and validated a bank of functional PA-R5 and A-R5 Envs cloned from well characterized primary R5 HIV-1 isolates.

Alterations in Env that augment fusogenicity contribute to the pathogenesis of SHIV infection [20,23-26]. In addition, HIV-1 fusogenicity is evident as MNGCs in tissues such as brain, which frequently harbors highly fusogenic R5 HIV-1 strains that share a number of phenotypic characteristics with blood-derived A-R5 viruses [43-46]. We used a quantitative cell-cell fusion assay to determine whether A-R5 Envs are more fusogenic than PA-R5 Envs. In this assay, cells were sampled at 2-hourly intervals until maximal fusion levels were reached at 12 hours post-mixing of Env-expressing effector cells and CD4/CCR5-expressing target cells. A-R5 Envs from the cross-sectional panel caused greater levels of cell-cell fusion than the PA-R5 Envs, which was particularly evident at 10 and 12 h post-mixing (Fig. 2A). The differences in fusogenicity between PA- and A-R5 Envs were not due to differences in cell surface Env expression levels on effector cells (Fig. 2B). When maximal fusion levels attained by PA- and A-R5 Envs were stratified across very low (+/-), low (+), moderate (++) or high (+++) maximal levels, the majority of A-R5 Envs had either moderate or high maximal fusion levels, whereas the majority of PA-R5 Envs had either very low or low maximal fusion levels (Fig. 2C). Thus, the A-R5 Envs from the cross-sectional panel are more fusogenic than the PA-R5 Envs.

We next tested whether A-R5 Envs from IK1 are more fusogenic than PA-R5 Envs cloned from this subject. A-R5 Envs from IK1 caused greater levels of cell-cell fusion than matched PA-R5 Envs, which was evident at 8, 10 and 12 h post-mixing (Fig. 3A). The differences in fusogenicity between the PA-R5 and A-R5 Envs were not due to differences in cell surface Env expression levels on effector cells (Fig. 3B). In fact, in these experiments PA-R5 Envs were expressed to greater levels on effector cells than A-R5 Envs. The extent of cell-cell fusion is directly related to the level of Env expression on effector cells (J. Sterjovski and P.R. Gorry, unpublished data). Thus, in subject IK1, A-R5 Envs are more fusogenic than PA-R5 Envs, and the differences shown in Figure 3A are likely to be conservative.

Enhanced fusogenicity of A-R5 Envs suggests that these Envs may be less sensitive to the antiviral effects of the HIV-1 fusion inhibitor T-20 than PA-R5 Envs. Therefore, we next determined the sensitivity of A-R5 and PA-R5 Envs from the cross sectional viruses to inhibition by T-20. Quantitative cell-cell fusion assays were carried out in the presence of 10-fold increasing concentrations of T-20 ranging from 0.001 to 10 μg per ml, and IC₅₀ and IC₈₀ values calculated by regression analysis of inhibition curves. The results demonstrate a significant increase in the IC₅₀ (Fig. 4A) and a non-significant trend toward an increase in the IC₈₀ (Fig. 4B) for T-20 against A-R5 Envs compared to PA-R5 Envs. These differences were not due to differences in cell surface Env expression levels on effector cells (Fig. 4C). Thus, highly fusogenic A-R5 Envs appear to be less sensitive to the inhibitory effects of T-20 in cell-cell fusion assays compared to less fusogenic PA-R5 Envs.

To identify amino acid variants associated with A-R5 Envs that may contribute to enhanced fusogenicity, the gp120 region of the 37 Envs was sequenced and analyzed. Phylogenetic analysis of Envs demonstrated tight clustering of nucleotide sequences according to virus isolate (data not shown). Multiple sequence alignments verified that all sequences were unique (data not shown). Together, these data demonstrate that the Envs are independent clones and not the result of sequence resampling.

A-R5 and PA-R5 Envs could not be segregated based on the total number of potential N-linked glycosylation sites (PNGS) in gp120 (range, 16 to 24 PNGS; median 19), length of the V1V2 variable loops (range, 68 to 83 amino acids; median 72), net charge of the V1V2 (range, -3 to +4; median +1) or V3 (range, +3 to +8, median +5) amino acid sequence, or number of PNGS in the V4 (range, 3 to 6 PNGS, median 4) or V5 (range 0 to 2 PNGS, median 1) sequence (data not shown), which are parameters shown previously to affect the biological activity of HIV-1 Envs [47-55]. Net charge of the V3 variable loop region did not predict coreceptor usage, consistent with results of previous studies [44,46,56,57].

Signature pattern analysis of A-R5 and PA-R5 Envs from the cross-sectional viruses identified an amino acid variant, N362, that was present more frequently in A-R5 Envs (14/15 Envs; 93%) than PA-R5 Envs (6/13 Envs; 46%) (Fig. 5). No additional Env changes that could potentially distinguish A-R5 Envs from PA-R5 Envs were identified. Consistent with these results, database analysis of published Env sequences where sufficient clinical information was present to confidently assign Envs as A-R5 or PA-R5 [34,38,58-65] demonstrated N362 is significantly more frequent in A-R5 Envs (74%; n = 142) compared with PA-R5 Envs (49%; n = 77) (p = 0.0004, Fisher's exact test). N362 is located in the C3 Env region immediately N-terminal to residues in the CD4bs [50], suggesting that N362 could potentially influence Env-CD4 binding. N362 is also present in ADA and YU2 R5 Envs, which are highly fusogenic, similar to the majority of A-R5 Envs (data not shown). In contrast, threonine is present at this position in JR-CSF Env, which is poorly fusogenic, similar to the majority of PA-R5 Envs (data not shown). Thus, the presence of N362 is associated with A-R5 Envs from the cross-sectional panel and other published Env sequences, and may contribute to enhanced fusogenicity. However, since N362 is present in a relatively high proportion of PA-R5 Envs from the cross sectional viruses (6/13) and other published studies (49%), any effect N362 may have on the biological activity of A-R5 Envs is likely to be strain-specific and/or context dependent.

The preceding studies showed a predominance of A-R5 Envs containing N362, but Envs from two PA-R5 viruses, NB23 and NB27, also contained N362. Furthermore, although the fusogenicity of A-R5 Envs from the cross-sectional panel was significantly greater than that of the PA-R5 Envs, there was still considerable overlap. To better understand the relationship between the presence of N362, fusogenicity, and sensitivity to T-20, these data were stratified based on the presence or absence of N362 (Fig. 6). R5 Envs containing N362 had significantly greater maximal levels of cell-cell fusion than R5 Envs lacking N362 (Fig. 6A). The differences in cell-cell fusion between Envs containing or lacking N362 were not due to differences in cell surface Env expression levels on effector cells (Fig. 6B). Envs containing N362 had a significantly higher IC₈₀ and a trend toward a higher IC₅₀ for T-20 than Envs lacking N362 (Fig. 6C,D). Together, these data demonstrate an association between the presence of N362 in R5 Envs from the cross sectional panel and enhanced fusogenicity, and an additional association between the presence of N362 and sensitivity to the inhibitory effects of T-20 in cell-cell fusion assays.

Previous studies of brain-derived Envs identified the N283 variant in the C2 region of gp120 within one of the CD4 contact sites, which increases Env-CD4 affinity and enhances M-tropism [43]. Since blood-derived A-R5 viruses have enhanced M-tropism compared to PA-R5 viruses [34,36], and N362 is located immediately N-terminal to another CD4 contact site in the C3 gp120 region [Fig. 5 and [50]], we hypothesized that N362 may potentially affect Env structure and CD4 binding. N362, a potential site for N-linked glycosylation, was modelled on the unliganded crystal structure of SIV gp120 and CD4-liganded crystal structure of HIV-1 JRFL gp120. The CD4bs in the unliganded gp120 is located in the outer domain and consists of a disordered loop flanked by the β-14 and β-16 strands. N362 is positioned just distal to the β-14 strand in the disordered loop region of the CD4 binding motif (Fig. 7A). Upon binding to CD4, conformational changes lead to interactions between the β-14, β-18 and β-24 strands to form an antiparallel sheet (Fig. 7B), which is one of two major conformational changes that occur in gp120 upon CD4 binding [66]. Analysis of interatomic contacts within the antiparallel sheet region showed that N362 has the potential to form hydrogen bonds with one or more residues from within the β-14 strand and/or neighbouring strands of the β-sheet, including V360, F361, H363, F468, and R469 (Fig. 7C). Modeling glycosylated residues on the CD4-bound gp120 demonstrated N362 is in close proximity to N392, another potentially glycosylated residue in the β-18 strand (data not shown). Thus, N362 may be important in formation of the CD4-bound structure of gp120, and may contribute to the stability of the CD4-bound conformation of gp120 by forming intramolecular hydrogen bonds with residues from neighbouring strands and/or interaction with other glycosylated residues.

The structural models suggest N362 may influence CD4 binding and thus, the efficiency of HIV-1 entry. To better understand how N362 may enhance the entry of A-R5 Envs, we produced single-round luciferase reporter viruses pseudotyped with a subset of A-R5 Envs containing N362 (NB6-C2, NB6-C3, NB6-C4, NB7-C2, NB7-C4, NB8-C2 and NB8-C4) or with a subset of PA-R5 Envs lacking N362 (NB24-C1, NB24-C2, NB24-C3, NB24-C4, NB25-C2, and NB25-C4). Using time-of-addition studies with T-20, we compared the kinetics of HIV-1 entry between reporter viruses pseudotyped with A-R5 Envs containing N362 and PA-R5 Envs lacking N362 (Fig. 8A). In this assay, the maximal delay time after addition of virus to cells when addition of T-20 can still completely inhibit HIV-1 entry was measured; shorter delay times indicate faster entry kinetics, and longer delay times indicate slower entry kinetics. Viruses pseudotyped with Envs containing N362 had significantly shorter delay times than those pseudotyped with Envs lacking N362. Thus, the presence of N362 in the A-R5 Envs studied is associated with faster entry kinetics.

The prediction from structural models that N362 may contribute to stabilizing the CD4-liganded gp120 structure, and the observed faster entry kinetics by Envs containing N362 suggests that N362 may increase the exposure of the CD4bs in gp120. To determine the relationship between the presence of N362 in A-R5 Envs and CD4bs exposure in gp120, we compared the sensitivity of reporter viruses pseudotyped with A-R5 Envs containing N362 or PA-R5 Envs lacking N362 to neutralization by Env mAbs or HIV-Ig. Viruses pseudotyped with Envs containing N362 were more sensitive to neutralization by the Env mAb IgG1b12 than viruses pseudotyped with Envs lacking N362, as shown by significant differences in the IC₅₀ and IC₉₀ for IgG1b12 (Fig. 8B). In fact, 7/7 of the A-R5 Envs tested were highly sensitive to neutralization by IgG1b12 whereas in comparison, only 2/6 PA-R5 Envs were neutralized by IgG1b12 and 4/6 PA-R5 Envs were completely resistant to neutralization by IgG1b12 (IC₅₀ and IC₉₀ > 50 μg per ml). However, there were no differences in sensitivity to neutralization by the Env mAb 2G12 (Fig. 8C) or polyclonal HIV-Ig (Fig. 8D) between viruses pseudotyped with Envs either containing or lacking N362. Since the binding site for IgG1b12 overlaps the CD4bs in gp120 [67], these results suggest that the A-R5 Envs containing N362 have increased CD4bs exposure compared to the PA-R5 Envs lacking N362. This conclusion is supported by results of FACS-based antibody binding studies, which showed that A-R5 Envs containing N362 bound to non-saturating concentrations of IgG1b12 more efficiently than PA-R5 Envs lacking N362 when expressed to equivalent levels on the surface of 293T cells (data not shown).

To determine whether N362 contributes to enhanced fusogenicity of R5 Envs, a panel of Env mutants was generated that either introduced or removed Asn at position 362 in gp120. Western blot analysis of transfected 293T cells demonstrated equivalent levels of Env expression between wild type Envs and Env mutants (Fig. 9A). The effect of N362 on fusogenicity was first determined in R5 control Envs (Fig. 9B). Wild type ADA and YU2 Envs have N362 and are highly fusogenic, whereas wild type JR-CSF Env lacks N362 and is, by comparison poorly fusogenic (data not shown). Replacement of Asn with Lys at position 362 in ADA and YU2 Envs resulted in significant reductions in fusogenicity. Conversely, introducing Asn at position 362 in JR-CSF Env resulted in a significant increase in fusogenicity. Thus, N362 contributes to enhanced fusogenicity of ADA and YU2 Envs and increases fusogenicity of JR-CSF Env.

To better understand the contribution of N362 to enhanced fusogenicity of A-R5 Envs, Asn at position 362 in gp120 of the highly fusogenic NB2-C4, NB6-C3, NB7-C1 and NB8-C4 Envs was replaced with Lys. The removal of Asn at position 362 resulted in significant reductions in fusogenicity of NB2-C4, NB6-C3 and NB8-C4 Envs, but resulted in an apparent increase in fusogenicity by NB7-C1 Env (Fig. 9C). However, increased fusogenicity by the NB7-C1 Env mutant was marginal and significant only at one early timepoint (6 h). Thus, whether substituting Asn for Lys at position 362 modulated fusogenicity of NB7-C1 Env is presently unclear. Together, results of the mutagenesis studies indicate N362 contributes to fusogenicity of the majority of the R5 Envs studied. However, the effect of N362 on fusogenicity of A-R5 Envs appears to be strain-dependent, suggesting the presence of additional factors contributing to fusogenicity of R5 Envs.

In this study, we utilized two independent panels of well characterized primary R5 HIV-1 isolates. The first was a cross sectional panel of R5 HIV-1 viruses isolated from immunocompetent subjects with asymptomatic HIV-1 infection (n = 4 isolates) or from patients with AIDS (n = 4 isolates) [34,36]. The second comprised R5 HIV-1 viruses isolated sequentially from one subject (IK1) from chronic HIV-1 infection to AIDS (n = 2 isolates) [13,42]. For the purpose of this study, the viruses isolated from patients with AIDS are referred to as AIDS R5 (A-R5) viruses, and those isolated from the earlier times are referred to as pre-AIDS R5 (PA-R5) viruses.

A detailed characterization of the cross sectional panel of R5 HIV-1 viruses including analysis of quasispecies diversity, coreceptor usage, replication kinetics, and clinical characteristics of the subjects from whom they were isolated, has been described previously [34,36]. Briefly, the PA-R5 viruses NB23, NB24, NB25 were isolated from peripheral blood mononuclear cells (PBMC) of individuals with CDC category II disease (asymptomatic infection) with CD4 counts of >500 cells/μl. PA-R5 virus NB27 was isolated from PBMC of an individual with CDC category I disease (acute seroconversion) with CD4 count of >750 cells/μl. A-R5 viruses NB2, NB6, NB7 and NB8 were isolated from PBMC of individuals with CDC category IV disease (AIDS) and CD4 counts of <50 cells/μl.

The longitudinal R5 HIV-1 viruses isolated from sequential PBMC samples from subject IK1 are designated IK1-PA and IK1-A. These viruses were referred to previously as 435–531 and 435–3415, respectively [13]. A detailed characterization of these HIV-1 isolates including analysis of quasispecies diversity, coreceptor usage, replication kinetics, and clinical characteristics of the subjects from whom they were isolated, has been described previously [13,42].

PBMC were purified from blood of healthy HIV-1-negative donors, stimulated with 5 μg of phytohemagglutinin (PHA) (Sigma, St. Louis, MO) per ml for 3 days, and cultured in RPMI 1640 medium supplemented with 10% (vol/vol) fetal calf serum (FCS), 100 μg of penicillin and streptomycin per ml, and 20 U of interleukin-2 (IL-2) (Roche, Basel, Switzerland) per ml. Cf2-Luc cells [25] are derived from the Cf2th canine thymocyte cell line [77], and stably express the luciferase gene under the control of the HIV-1 long terminal repeat and were cultured in Dulbecco modified Eagle medium (DMEM) supplemented with 10% (vol/vol) FCS, 100 μg of penicillin and streptomycin per ml, and 0.7 mg of G418 per ml. Cf2-CD4 cells [78] were cultured in DMEM supplemented with 10% (vol/vol) FCS, 100 μg of penicillin and streptomycin per ml, and 0.5 mg of G418 per ml. Cf2-CD4/CCR5 cells [79] were cultured in DMEM supplemented with 10% (vol/vol) FCS, 100 μg of penicillin and streptomycin per ml, 0.5 mg of G418 per ml, and 0.1 mg of hygromycin per ml. Cf2-CD4/CXCR4 cells were constructed by transduction of the Cf2-CD4 cell line [78] with pBABE-puro vectors expressing CXCR4 [80,81] followed by selection and expansion in DMEM supplemented with 10% (vol/vol) FCS, 100 μg of penicillin and streptomycin per ml, 0.5 mg of G418 per ml, and 1 μg of puromycin per ml. Cf2-CD4/CCR5/CXCR4 cells [56] were cultured in DMEM supplemented with 10% (vol/vol) FCS, 100 μg of penicillin and streptomycin per ml, 0.5 mg of G418 per ml, 0.1 mg hygromycin per ml, and 1 μg puromycin per ml. JC53 cells are derived from the HeLa cell line and stably express high levels of CD4, CXCR4 and CCR5 on the cell surface [82], and were cultured in DMEM supplemented with 10% (vol/vol) FCS, and 100 μg of penicillin and streptomycin per ml. 293T cells were cultured in DMEM supplemented with 10% (vol/vol) FCS, and 100 μg of penicillin and streptomycin per ml.

Genomic DNA was extracted from PBMC infected with NB2, NB6, NB7, NB8, NB23, NB24, NB25 and NB27 primary isolates using a QIAmp genomic DNA purification kit (Qiagen) according to the manufacturers' protocol. Viral RNA was isolated from 1 ml of IK1-PA and IK1-A primary isolates using a QIAmp UltraSense viral RNA isolation kit (Qiagen), according to the manufacturers' instructions. cDNA was reversed transcribed from viral RNA using SuperscriptIII RT (Invitrogen) and random hexamers, according to the manufacturers' protocol. An approximately 2.1 kb fragment spanning the KpnI to BamHI restriction sites in HIV-1 env (corresponding to nucleotides 6348 to 8478 in HXB2) was amplified by PCR using nested primers and Expand high fidelity DNA polymerase (Roche diagnostics), as described previously [56,57]. The outer primers were env1A and env1M [83], and the inner primers were Env-KpnI and Env-BamHI [56,57]. PCR cycling consisted of an initial denaturation step at 94°C for 2 min followed by 9 cycles of 94°C for 15 s, 60°C for 30 s and 72°C for 2 min, then a further 20 cycles of 94°C for 15 s, 60°C for 30 s and 72°C for 2 min but with a 5 s increasing extension time for each cycle, followed by a final extension at 72°C for 7 min. The products of 3 independent PCR reactions were purified and pooled, then cloned into the pSVIII-HXB2 Env expression plasmid [83] by replacement of the 2.1 kb KpnI to BamHI HXB2 env fragment. Thus, the resulting Env clones contain the entire gp160 coding region of primary virus-derived env genes except for 36 amino acids at the N terminus and 105 amino acids at the C terminus, which are derived from HXB2. Three to 4 functional Env clones from each virus were identified by the ability to support entry when pseudotyped onto Env-deficient GFP reporter viruses and used in single round entry assays in JC53 or Cf2-CD4/CCR5/CXCR4 cells, and by Western blot analysis of gp120/gp160 in transfected 293T cells and fusion assays. The coreceptor usage of functional Env clones was verified by single round entry assays in Cf2-CD4/CCR5 and Cf2-CD4/CXCR4 cells infected with Env-pseudotyped GFP reporter viruses, as described previously [56,84]. Envs were sequenced by Big Dye terminator sequencing (Applied Biosystems) and analyzed using a model 3100 Genetic Analyzer (Applied Biosystems).

Mutagenesis to introduce or remove Asn at position 362 in gp120 was performed using the QuikChange II site-directed mutagenesis kit (Stratagene) according to the manufacturer's instructions. Mutagenesis primers were designed to span the CD4 binding loop region corresponding to nucleotides 7297 to 7325 of HXB2 for control R5 Env clones ADA, YU2, and JR-CSF and primary A-R5 Env clones NB2-C4, NB6-C3, NB7-C1 and NB8-C4. The primers used in the mutagenesis studies are shown in Additional file 1. Envs mutants were sequenced to confirm the presence of the mutated residue.

For analysis of Env expression, 293T cells were co-transfected with 8 μg of pSVIII-Env plasmid and 2 μg pSVL-Tat plasmid using the calcium phosphate method. At 72 h after transfection, cells were rinsed twice in PBS and resuspended in 400 μl of ice cold lysis buffer (0.5% [vol/vol] NP-40; 0.5% [wt/vol] sodium deoxycholate; 50 mM NaCl; 25 mM Tris-HCl [pH 8.0]; 10 mM EDTA, 5 mM benzamidine HCl; and a cocktail of protease inhibitors) for 10 min, followed by centrifugation at 15,300 × g for 10 min at 4°C to remove cellular debris. Cell lysates were separated in 8.5% (wt/vol) sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels and analyzed by Western blotting using rabbit anti-gp120 polyclonal antisera. Env proteins were visualized using horseradish peroxidase-conjugated anti-rabbit immunoglobulin G antibody and enhanced chemiluminescence (Promega).

Fusion assays were conducted as described previously [43,45,71,85] with minor modifications. Briefly, 293T effector cells co-transfected with 3.4 μg of Env-expressing plasmid and 0.6 μg pSVL-Tat plasmid using Lipofectamine 2000 (Invitrogen) were mixed with Cf2-Luc target cells that had been co-transfected with 2 μg of pcDNA3-CD4 and 6 μg of pcDNA3-CCR5, and incubated at 37°C in replicate wells of 96-well tissue culture plates containing 200 μl of culture medium. Cells from replicate wells were harvested at 2, 4, 6, 8, 10 and 12 h post-mixing and assayed for luciferase activity (Promega) according to the manufacturers' protocol. 293T cells transfected with pSVL-Tat alone were used as negative controls to determine the background level of luciferase activity. In experiments analysing the ability of T-20 to inhibit Env function in fusion assays, the 10 h time point was used for analysis. To control for cell surface Env expression levels in 293T effector cells, Env expression was measured by flow cytometry using pooled AIDS serum and a FITC-conjugated anti-human F(ab')2 Ig (Chemicon). To account for both the number of Env expressing cells and the fluorescence intensity, the relative fluorescence was calculated from fluorescence-activated cell sorter (FACS) profiles by multiplying the percentage Env-expressing cells by the mean channel fluorescence, as described previously [86].

Env-pseudotyped, luciferase reporter viruses were produced by transfection of 293T cells with pCMVΔP1ΔenvpA, pHIV-1Luc and pSVIII-Env plasmids using Lipofectamine 2000 (Invitrogen) at a ratio of 1:3:1, as described previously [56,79,87,88]. Supernatants were harvested 48 h later and filtered through 0.45 μm filters. Recombinant luciferase reporter viruses were ultracentrifuged through a 25% (vol/vol) sucrose cushion at 25,000 rpm for 2 h at 4°C using a Beckman Ultra high speed centrifuge and a SW28 rotor, resuspended in 2 ml culture medium, aliquotted and stored at -80°C. The TCID₅₀ of virus stocks was determined by titration in JC53 cells.

Time-of-addition experiments using the HIV-1 fusion inhibitor T-20 and Env-pseudotyped luciferase reporter viruses were conducted to measure the entry kinetics of Env. This method was based on that recently described by Olivieri et al., [68] with the following modifications. Two hundred TCID₅₀ of Env-pseudotyped luciferase reporter virus (equating to an MOI of 0.02) was added to replicate wells of JC53 cells. Fifty micrograms of T-20 per ml was added to the replicate wells at 0, 10, 20, 30, 40, 50 60 and 120 min post-infection. This concentration of T-20 was empirically determined to be completely inhibitory for each of the Env-pseudotyped luciferase reporter viruses tested when added to virus before infection of JC53 cells (data not shown). After 2 h, the virus inoculum was removed and the cells were washed twice prior to addition of fresh culture medium containing 50 μg of T-20 per ml. Cells were harvested 48 h later and assayed for luciferase activity (Promega) according to the manufacturers' protocol. The maximum delay time after addition of virus to JC53 cells when addition of 50 μg per ml of T-20 can still completely inhibit HIV-1 entry was calculated (Max. delay of entry). For these analyses, inhibition of entry was defined as luciferase activity measurements <3-fold above background, as determined using reporter virus pseudotyped with a non-functional Env (ΔKS Env) [25]. Luciferase activity was plotted against time of T-20 addition using Prism version 4.0c (GraphPad Software, San Diego, CA.). Data were fitted with a nonlinear function, and Max. delay of entry was calculated by least squares regression analysis of time-of-addition curves. Envs with lower Max. delay of entry values were interpreted to have faster entry kinetics, and vice-versa.

Human mAbs against HIV-1 gp120 (IgG1b12, 2G12) and the polyclonal antibody HIV-Ig have been described previously [89-93]. The ability of these antibodies to neutralize the infectivity of Env-pseudotyped luciferase reporter viruses was assayed using JC53 cells. Two hundred TCID₅₀ of each Env-pseudotyped luciferase reporter virus (equating to an MOI of 0.02) was incubated with 10-fold increasing concentrations of each mAb (0.0005 to 50 μg/ml) or HIV-Ig (1 to 10,000 μg/ml) for 1 h at 37°C. The virus-Ab mixtures were then used to inoculate JC53 cells overnight at 37°C. Cells were rinsed twice with culture medium to remove residual virus inoculum and incubated a further 48 h at 37°C. Virus infectivity was then measured by assaying luciferase activity in cell lysates (Promega), according to the manufacturers' protocol. Negative controls included mock-infected cells that were incubated with culture medium instead of virus, and cells treated with luciferase reporter virus pseudotyped with the non-functional ΔKS Env. After subtracting background luciferase activity, the amount of luciferase activity in the presence of antibody was expressed as a percentage of the amount produced in control cultures containing no antibody. The percent inhibition was calculated by subtracting this number from 100. Data were fitted with a nonlinear function, and fifty percent inhibitory concentration (IC₅₀) and IC₉₀ values were calculated by least squares regression analysis of inhibition curves, as described previously [36,45,85].

Proteins structures of unliganded SIV gp120 (2BF1) [66,94] and the V3 loop-containing protein structure of JR-FL (2B4C) [95] were obtained from the RCSB Protein Data Bank. Structural modelling was performed using Swiss PDB Viewer. Hydrogen bond analysis was performed using CSU software [96].

The Env nucleotide sequences reported here have been assigned GenBank accession numbers EU308533 to EU308568.