Coupling of Human Immunodeficiency Virus Type 1 Fusion to Virion Maturation: a Novel Role of the gp41 Cytoplasmic Tail

Virus-cell fusion assay.HIV-1 particles carrying a β-lactamase reporter protein were produced by cotransfection of 293T cells with cloned HIV-1 proviral DNA and pMM310, a construct encoding β-lactamase (BlaM) fused to the amino terminus of the virion protein Vpr (BlaM-Vpr), thereby targeting BlaM to the virion. To assay fusion, we inoculated SupT1 cells (100,000) by centrifugation at 600 × g for 1.5 h at 35°C with dilutions of viruses in 100-μl total volumes and incubated them at 37°C for 2 h to allow virus-cell fusion. CCF2/AM (35) (20 μM; Aurora Biosciences Corp.) was added, and the cultures were incubated for 14 h at room temperature. Cells were pelleted and resuspended in phosphate-buffered saline (PBS), and the fluorescence was measured at 447 and 520 nm in a microplate fluorometer after excitation at 409 nm. Fluorescence ratios were calculated after subtraction of the average background fluorescence of control cultures containing no virus (blue values) and wells containing PBS (green values). Duplicate determinations were performed for each virus dilution, and the fusion values typically agreed to within 10%.

Viruses.The viral mutants used in this study were derived from the wild-type infectious HIV-1 proviral clone pNL4-3 (1). Viruses were produced by transient transfection of 293T cells in 100-mm-diameter dishes with 20 μg of proviral DNA and 7 μg of pMM310, encoding a BlaM-Vpr fusion protein, by a calcium phosphate method. For the generation of HIV-1(VSV) pseudotyped viruses, the proviral DNA was reduced to 18 μg and 2 μg of pHCMV-G was included. Virus stocks were harvested 36 to 48 h after transfection and clarified through 0.45-μm-pore-size syringe filters, and aliquots were buffered with HEPES (10 mM) prior to storage at −80°C. HIV-1 stocks were assayed for p24 by enzyme-linked immunosorbent assays (ELISAs) (31) after incubation for 10 min at 95°C in the presence of 1% sodium dodecyl sulfate (SDS) to disrupt immature particles and subsequent dilution of at least 100-fold to prevent SDS interference with antigen-antibody interactions. This procedure was capable of accurately quantifying the levels of both processed and unprocessed Gag proteins, as demonstrated by Western blot analysis of Gag cleavage mutants with a variety of HIV-specific antisera after the analysis of equal quantities of wild-type and mutant virions by ELISA.

The pNL-MA/CA mutant, containing a point mutation preventing cleavage at the MA-CA junction, was previously described (33). pNL-MA/p2 and pNL-MA/NC were created by transferring the SpeI-ApaI fragments of the previously described CA5 and CA6 mutants (32), respectively, into pNL-MA/CA. The remaining mutations were created by PCR fragment segment overlap extension. pNL-MA/p6 produces HIV-1 particles containing defects in all of the known Pr55^Gag cleavage sites and was generated by substituting Ile codons for the Asn and Phe codons at the P1 positions of the NC-p1 and p1-p6 junctions, respectively. The Asn-to-Ile mutation was generated by a PCR with the forward primer 5′-GCTATTTTTTTAGGGAAGATCTGGCCATCC. This mutant was then used as a template for the Phe-Ile substitution. The forward primer used for this mutation was 5′-CCAGGGAATATTCTTCAGAGTCGACCAGAGCCAACA. An ApaI-BclI fragment containing both mutations was transferred into pNL-MA/NC to generate pNL-MA/p6. PCR-amplified regions of the proviral clones were sequenced to confirm the presence of the desired mutations and that there were no additional mutations. env-defective versions of the Pr55^Gag cleavage site mutants were generated by replacing the EcoRI-BamHI fragments with those from pNL4-3Env⁻ and pNL4-3CTdel144-2 (21).

An immunoblot analysis of pelleted virions was performed as previously described (33). Blots were probed sequentially with a monoclonal antibody against HIV-1 gp120 (902; NIH AIDS Research and Reference Reagent Program), 2F5 monoclonal antibody against gp41, a monoclonal antibody against Escherichia coli β-lactamase (QED Biosciences, Inc.), and a rabbit antiserum against HIV-1 CA (from Didier Trono). Proteins were detected by chemiluminescence after binding of the appropriate horseradish peroxidase-conjugated secondary antibodies.

An electron microscopic analysis of HIV-1 particles was performed with pellets of transfected 293T cells. Cells were collected, fixed in paraformaldehyde, and washed four times in PBS. Cell pellets were postfixed in OsO₄ and embedded in Spurr's plastic resin, and ultra-thin sections were generated. Mounted sections were poststained with uranyl acetate and Reynold's lead citrate. Grids were examined in a JEOL 1200 EX transmission electron microscope operating at 60 kV.

Use of a quantitative assay of HIV-1 virus-cell fusion.To determine whether immature HIV-1 particles are defective in the ability to fuse with target cells, we adapted a recently developed reporter assay for the quantification of HIV-1 particle entry that measures BlaM activity delivered to target cells upon the fusion of virions containing BlaM fused to the HIV-1 Vpr protein (BlaM-Vpr) (5, 29a). After incubation with HIV-1, the SupT1 target cells were loaded with the fluorescent BlaM substrate CCF2/AM. This compound is taken up by cells and accumulates due to cleavage of the acetoxymethyl ester (AM) side chain by cellular esterases. Uncleaved CCF2 fluoresces green due to fluorescence resonance energy transfer between the coumarin and fluorescein groups; however, cleavage by BlaM results in the dissociation of these fluorophores and the emission spectrum shifts to blue. Thus, the ratio of blue to green cellular fluorescence is proportional to the overall extent of virus-cell fusion. In this assay, the fluorescence signal is dependent on the expression of CD4 and an appropriate coreceptor on target cells and on the viral envelope (Env) glycoproteins and is blocked by specific inhibitors of HIV-1 fusion, including T-20, neutralizing antibodies, and coreceptor antagonists (5; also data not shown). The signal was also proportional to the amount of input virus, thereby providing a quantitative measure of virus-cell fusion (Fig. 1a).

Immature HIV-1 particles are impaired for fusion with cells.In initial studies, we found that HIV-1 particles lacking a functional protease contain markedly reduced levels of active BlaM; consequently, these virions were inactive in the fusion assay. An immunoblot analysis revealed the presence of cleaved BlaM-Vpr in virions containing an active protease but not in PR-defective virions, suggesting that PR-mediated cleavage of the BlaM-Vpr fusion protein in the virion is required for BlaM activity. It remains unclear why BlaM-Vpr is apparently inactive as a fusion protein. Regardless, these observations precluded the use of a PR⁻ mutant as a source of immature HIV-1 reporter particles. To circumvent this limitation, we constructed an HIV-1 mutant, designated MA/p6, containing substitutions preventing cleavage of the major Pr55^Gag cleavage sites but encoding a wild-type PR sequence. Wild-type and MA/p6 virions were then tested in the CCF2/AM reporter assay. SupT1 cells were inoculated with similar quantities of wild-type and MA/p6 reporter virions, as quantified by p24 ELISA after SDS lysis of virions to disrupt the stable immature particles. For minimization of potential differences in attachment between wild-type and mutant virions and to increase the overall fusion signal, the cells were infected by centrifugal inoculation (22). MA/p6 virions exhibited reduced fusion activities, achieving levels of 10 to 20% those of wild-type HIV-1 particles (Fig. 1a). An immunoblot analysis of viral lysates demonstrated that wild-type and MA/p6 particles contained similar quantities of gp41 and gp120 (Fig. 1b). Probing of the blot with a CA-specific antiserum revealed the presence of a major band representing uncleaved Pr55^Gag in lysates of MA/p6 virions (Fig. 1b, lane 2). A secondary band of lower molecular weight was also detected for the MA/p6 lysates. This band was not observed when the mutant virions were produced in the presence of an HIV-1 protease inhibitor (data not shown), suggesting that it may correspond to a fragment of the Pr160^Gag-Pol precursor arising from PR cleavage of the p6^pol segment. To control for possible differential incorporation of the BlaM-Vpr reporter, we assayed the pelleted virions for BlaM protein by immunoblotting and for BlaM activity by using a chromogenic substrate after lysis of the virions in detergent. Wild-type and MA/p6 virions contained similar quantities of cleaved and uncleaved BlaM-Vpr and exhibited equivalent levels of BlaM activity in vitro (Fig. 1b and data not shown), thus confirming that mature and immature virions exhibited similar BlaM-Vpr incorporation, processing, and activation. Collectively, these results indicate that immature virions are impaired for fusion with target cells.

Truncation of the gp41 CT relieves the impaired fusion ability of immature HIV-1 particles.Previously, we showed that the gp41 CT mediates a strong association of Env proteins with immature HIV-1 virions, suggesting that the CT binds tightly to Gag in these particles (33). These observations suggested a functional role of the CT in regulating HIV-1 fusion. To test whether the gp41 CT links HIV-1 fusion competence to core maturation, we quantified the fusion competence of wild-type and MA/p6 virions lacking the C-terminal 144 amino acids of gp41 (ΔCT). Truncation of the gp41 CT reduced the fusion efficiency of mature HIV-1 50 to 60% (Fig. 1a). In contrast, the fusion of immature particles was enhanced by the CT truncation, resulting in levels of fusion similar to those of the corresponding mature virions (Fig. 1a, compare MA/p6-ΔCT and NL-ΔCT). Thus, removal of the gp41 CT resulted in the equivalent fusion of immature and mature HIV-1 particles with target cells, suggesting that the gp41 CT couples HIV-1 fusion to virion maturation. These results further demonstrate that the impaired fusion ability of immature virions bearing full-length Env is not due to defective BlaM release in target cells or to the greater structural stability of immature HIV-1 particles. An immunoblot analysis of viral lysates demonstrated that similar amounts of gp120 were present on mature and MA/p6 particles lacking the gp41 CT (Fig. 1b). However, we observed increased levels of truncated gp41 on mature and immature virions, likely resulting from higher levels of Env expressed on the cell surface due to the loss of the tyrosine-based endocytic motif within the gp41 CT (27). This result was confirmed by using several different gp41-specific monoclonal and polyclonal antibodies (data not shown). Thus, the enhanced fusion of MA/p6 upon truncation of the CT could be due to the elevated levels of gp41 on the surfaces of the virions. To test this possibility, we produced mature and immature HIV-1 particles containing decreasing quantities of gp41-truncated Env by cotransfecting the proviruses with matched Env-defective clones at various ratios. Mature and immature HIV-1 particles bearing truncated gp41 exhibited similar fusion efficiencies over a range of Env levels (Fig. 1c). We concluded that the removal of the CT uncouples HIV-1 fusion from virion maturation by a mechanism that is independent of the overall levels of Env on the virions.

Pseudotyping by VSV-G rescues the impaired fusion of immature HIV-1 particles.To further probe the role of the gp41 CT in inhibiting fusion, we determined whether a heterologous Env protein containing a short CT would permit the fusion of immature HIV-1 particles. For this purpose, Env-defective HIV-1 particles were pseudotyped with the glycoprotein of the vesicular stomatitis virus (VSV-G). In contrast to the case with normal HIV-1 particles, infection by HIV-1(VSV) particles requires endosomal acidification (3), consistent with the known pH dependence of VSV-G-induced fusion (17). Particles were produced by the cotransfection of env-defective proviral constructs with plasmids expressing VSV-G and BlaM-Vpr, and the virions were assayed for fusion. Pseudotyping by VSV-G restored fusion of the MA/p6 particles to a level identical to that of mature pseudotyped particles (Fig. 1d). The pseudotyped virions exhibited higher signals than wild-type HIV-1 in the fusion assay, consistent with the higher overall infectivity of HIV-1(VSV) (3). Assays employing approximately threefold dilutions of HIV-1(VSV) yielded reduced signals, but the fusion activity remained similar for particles containing mature and MA/p6 cores, confirming that the results were not due to assay saturation (Fig. 1d and data not shown). Based on these results, we conclude that the fusion of immature HIV-1 particles is specifically inhibited by the gp41 CT.

HIV-1 fusion is activated by cleavage of Gag between MA and NC.The impaired fusion of immature (MA/p6) virions suggested that HIV-1 fusion is activated during virion maturation by the specific cleavage of one or more sites in Pr55^Gag. To identify the cleavage event(s) required for fusion, we constructed a set of proviral clones, allowing selective processing of one or more Pr55^Gag cleavage sites (Fig. 2a), and quantified their fusion activity levels. The MA/CA and MA/p2 mutants were 70 to 80% as active as wild-type virions, as quantified by comparison of the linear regions of the dose-response curves (Fig. 2b). This result was somewhat surprising, since we had previously shown that the MA/CA mutant is highly resistant to detergent, similar to PR-defective virions (33). Thus, the impaired ability of immature HIV-1 particles to fuse with target cells did not appear to result from the overall structural stability of the virions. In contrast, the MA/NC particles exhibited only 30% of the wild-type HIV-1 fusion level (Fig. 2b), suggesting that the release of NC from Pr55^Gag is required for the activation of HIV-1 fusion with target cells. To determine whether cleavage between CA and NC is required for efficient HIV-1 fusion, we tested a previously described HIV-1 mutant (CA6) that specifically prevents cleavage between CA and NC while permitting cleavage of the MA-CA and NC-p6 junctions (32). CA6 virions exhibited fusion activity levels that were comparable to that of wild-type HIV-1 (Fig. 2b). Western blot analysis of viral lysates confirmed that the combined mutations resulted in cleavage at the intended sites and that the virions contained similar quantities of BlaM, gp120, and gp41 (Fig. 2c).

HIV-1 maturation mutants exhibit expected internal morphologies.For determination of whether the cleavage site substitutions resulted in the expected structural alterations, 293T cells transfected with the mutant proviruses were analyzed by electron microscopy. The results revealed the presence of similar quantities of budding and cell-associated budded virions. No consistent differences in particle size or numbers were detected for the mutant virions. However, none of the mutant viruses exhibited conical cores characteristic of mature HIV-1 particles (Fig. 3). Virions containing mutations preventing cleavage between CA and NC lacked electron-dense cores, consistent with impaired formation of the viral ribonucleoprotein complex, as previously reported (32). Virions that failed to cleave between MA and CA exhibited a thickened layer of protein at the membrane, consistent with a previous analysis of a similar MA/CA mutant (11). Based on these results, we conclude that the inhibition of HIV-1 fusion mediated by the gp41 CT requires an intact segment of Pr55^Gag extending from the MA domain through NC.

Functional implications of the regulation of HIV-1 fusion by the gp41 cytoplasmic domain.The ability to couple HIV-1 fusion competence to virion maturation represents a novel function of the gp41 CT. In contrast to those of simple retroviruses, the CTs of lentiviral fusion proteins are unusually long (reviewed in reference 13). The gp41 CT facilitates the incorporation of the Env protein complex into HIV-1 particles in a cell-type-dependent manner (21). Although the gp41 CT is not required for fusion, removal of the CT inhibits the ability of HIV-1 to replicate in primary T cells and in many T-cell lines due to a defect in Env protein incorporation. The CT is also required for simian immunodeficiency virus (SIV) replication in vivo. Truncations in the CT arise spontaneously during culturing of SIV in human cells, but these mutations revert during viral passaging in monkey cells and in infected macaques (12, 14, 16). Our results demonstrate that in addition to its role in Env incorporation, the gp41 CT inhibits the fusion of immature HIV-1 particles, likely through interactions with Gag. This regulation may enhance viral fitness in several ways. By preventing premature HIV-1 fusion, the gp41 CT may enhance the frequency of fusion events that result in productive infections. Although maturation is initiated during HIV-1 budding from infected cells, it is likely that during continuous viral replication, a significant fraction of infection events occurs when virions are transmitted between infected and uninfected cells that are in close contact (for example, in lymph nodes). In such an environment, particle maturation could be a rate-limiting step in replication; hence, even a slight delay in fusion could give the virus a replicative advantage. In this context, Gag binding to gp41 in nascent immature virions could also limit the exposure of sensitive epitopes on gp120 or gp41, thereby enhancing resistance to neutralizing antibodies. The binding of Gag to gp41 may also limit fusion of infected and uninfected cells, thereby prolonging cell survival and leading to larger yields of progeny virions. Our results further suggest that the therapeutic efficacy of HIV protease inhibitors may be due in part to the reduced ability of immature virions to fuse with target cells.

Our findings are reminiscent of observations reported for other retroviruses, including murine leukemia virus, Mason-Pfizer monkey virus, and equine infectious anemia virus (2, 4, 24-26). The CTs of these viral Env proteins are cleaved by their respective viral proteases, and for murine leukemia virus and Mason-Pfizer monkey virus Env proteins, cleavage appears to enhance Env-mediated fusion. In contrast, the 152-amino-acid gp41 CT of HIV-1 is not cleaved in virions (D. J. Wyma and C. Aiken, unpublished observations). Instead, HIV-1 fusion activity is regulated by the gp41 CT, likely through a novel mechanism involving an association with Gag. Our data, together with studies of other retroviral Env proteins, suggest that the coupling of fusion competence to virion maturation may be a general requirement for efficient retrovirus replication in vivo.

Proposed mechanism for regulation of virus cell fusion by the gp41 CT.Although the mechanism by which the gp41 CT inhibits the fusion of immature HIV-1 particles remains to be elucidated, we propose a model in which the binding of gp41 to Gag inside the virion suppresses the receptor-induced conformational changes in the gp41 domains on the virion outer surface that mediate fusion. This hypothetical mechanism is based in part on our previous work in which we showed that the HIV-1 Env proteins gp120 and gp41 are strongly associated with immature HIV-1 particles through the gp41 CT, suggesting that the gp120-gp41 association is stabilized on immature HIV-1 particles (33). Consistent with this hypothesis, we observed a decreased ratio of gp120 to gp41 on virions lacking the gp41 CT. In further support of the model, studies employing cell-to-cell fusion assays have reported an inhibitory effect on membrane fusion mediated by the CTs of HIV-1, HIV-2, and SIV through the modulation of Env conformation (9, 19, 28, 29). Our data also indicate that the ability of HIV-1 particles to fuse with target cells is activated by the cleavage of Gag between the MA and NC domains, i.e., at the MA-CA or CA-NC junction. The known RNA binding ability of NC suggests a possible role for viral RNA in inhibiting the fusion of immature HIV-1 particles. The viral genomic RNA may lock gp41 in an inactive state by maintaining Gag, while bound to the gp41 CT, in an extended conformation. The release of NC by proteolytic cleavage may allow for rearrangement of the remaining MA-CA fusion protein, thereby releasing the structural constraints on gp41. Although this model is conceptually appealing, another possibility is that by tethering Env to Gag, the gp41 CT prevents the clustering of Env trimers on the surfaces of immature virions that may be required for expansion of the fusion pore. These two mechanisms are not mutually exclusive, and further studies will be required to determine the relative contribution of each.