Identification of an env-defective HIV-1 mutant capable of spontaneous reversion to a wild-type phenotype in certain T-cell lines

MT2, MT4, Jurkat, CEM, PM1 and SupT1 cells were all obtained through the NIH AIDS Research and Reference Reagent Program. All these cell types can efficiently support CXCR4-dependent HIV replication, and both MT2 and MT4 cells are standards for HIV replication studies. All cells were maintained in RPMI 1640 medium (Invitrogen) supplemented with 10% fetal bovine serum (Invitrogen), 1% L-glutamine, and 100 units/ml penicillin/100 μg/ml streptomycin under 5% CO₂ at 37°C. 293 T and TZM-bl cells were maintained in Dulbecco modified Eagle medium (Invitrogen). Culture media were changed every 3 days.

The HIV-1 HxB 2D clone was used as wild type (WT) virus. A defective env G367R mutant virus was produced by transfecting a proviral plasmid, that contained a mutation at position G367R in the CD4 binding site of gp 120, and that has been generated by PCR based site directed mutagenesis [17]. pVPack-VSV-G (Stratagene), which encodes the vesicular stomatitis virus (VSV) envelope glycoprotein, was used at 1:1 to produce VSV-G pseudotyped viruses. G367R mutant viruses and VSV-G pseudotyped viruses were generated in 293 T cells by transfection from the proviral plasmids with lipofectamine 2000 as recommended by the manufacturer (Invitrogen). At 48 hours after transfection, viruses were harvested, analyzed for p24 antigen and stored at −80°C.

R5 tropic subtype C Indie viruses were generated in 293 T cells by transfection from the Indie C proviral plasmid [28], also using lipofectamine 2000 as recommended by the manufacturer.

Virus stocks of WT were amplified in MT2 cells or in cord blood mononuclear cells (CBMCs), obtained through our Hospital Department of Obstetrics. Virus stocks were kept at −70°C until use.

2-5×10⁵ cells were infected with 0.5 ml of culture fluid containing ~500 ng of p24 of HIV for 3 hr at 37°C. Cells were washed three times with RPMI 1640 medium immediately following incubation to remove unbound virus and were resuspended in RPMI 1640 medium and split into wells at appropriate concentrations. Cytopathic effects (CPE) were monitored by microscopy and supernatants were evaluated for p24 antigen at 3 day intervals (Biomerieux, Netherlands).

For infection of 293 T or TZM-bl cells, 10⁵ cells plated the previous day were infected overnight by 0.5 ml of VSV-G pseudotyped viruses. Supernatants were removed and washed once and fresh medium was added. Supernatants were collected and cells were split at 3 day intervals and monitored for p24 antigen.

10⁴ 293 T or TZM-bl cells infected with pseudotyped G367R virus in a donor coculture were mixed with 5×10⁴ target cells. Alternatively, 10⁵ MT4 cells infected with pseudotyped G367R virus in a donor coculture were mixed with 2-4×10⁵ MT2 cells. CPE and levels of p24 antigen in supernatants were monitored to verify efficiency of infection and reversion.

The HIV entry inhibitors AMD3100, maraviroc, and DS003 were all obtained through the NIH AIDS Research and Reference Reagent Program. The endocytosis inhibitors chlorpromazine hydrochloride, Genistein and methyl-β-cyclodextrin were purchased from Sigma-Aldrich, Inc. Phorbol 12-myristate 13-acetate (PMA) was purchased from Cayman Chemical, Ann Arbor, MI.

We previously reported that viruses containing the env mutant G367R are defective in cell-free infection [17], similar to what has been observed for a mutation at position 368, D368R [25‐27]. Theoretically, expected reversion rates should be about 10⁻⁵ at a given site in single round infection studies, if the error rate of HIV reverse transcriptase is estimated at around 5×10⁻⁴. Thus, reversions might only be expected at a rate of 5×10⁻⁴, i.e., it would only be observed in a proportion of experimental studies when 10⁵ cells or fewer are seeded after infection at appropriate viral concentrations. However, we found higher reversion rates than expected when we employed the defective G367R virus that was pseudotyped with VSV-G and then used to infect MT2 cells (Figure 1 and Table 1). Figure 1 shows that the increase in p24 in the G367R-infected samples was slower than occurred when WT virus was used to mediate infection. This is not unexpected, considering that only a portion of cells are infected in a single round infection because of defective progeny viruses. However, CPE appeared in all the samples within 1–3 weeks, contrary to expectations. Reversion was confirmed by the fact that the progeny viruses gained the ability to participate in cell-free infections during second round infection experiments. This was demonstrated as well by documentation of a wild-type genotype, as shown by sequencing of the viral env gene (see below). In contrast, G367R alone (i.e., not pseudotyped) was noninfectious (Figure 1), consistent with our previous report [17]. Similar results were obtained for all samples tested over repeat experiments that involved seeding either 5×10⁴ or 10⁴ infected cells (Table 1). Thus, the reversion rate of G367R seems higher than the RT error rate over a single round of infection. As a consequence, we next investigated this phenomenon using a number of different cell types.

Four T cell lines, MT2, MT4, SupT1 and PM-1, were used to investigate the mechanism(s) whereby G367R can achieve reversion. All of these cells express both the CD4 receptor and the CXCR4 coreceptor, while PM-1 cells also express the CCR5 coreceptor. Variable amounts of virus and cell numbers were used in order to estimate the reversion rate. The growth curves and extent of CPE of G367R mutated virus in each of MT2, MT4, SupT1 and PM-1 cells are displayed in Figure 2 and in Table 1. The results show that reversion of G367R was cell-type dependent, with reversion occurring very frequently in both MT2 and SupT1 cells but rarely in MT4 and PM-1 cells (Tables 1 and 2). The initial infection of MT4 by pseudotyped G367R was successful on the basis of p24 value but viral replication then decreased gradually while PM-1 cells were not as sensitive to infection as MT4 or MT2 cells (Figure 2). Sequencing of the env gene in all the reverted MT2 and SupT1 cells, regardless the number of infected cells that were seeded demonstrating that G367R had reverted from GGA AGG GAC CCA to GGA GGG GAC CCA, corresponding to a change in amino acid sequence from GRDP to GGDP within the CD4 binding site (Table 3).

The time of reversion was closely related to the numbers of cells seeded after 3 hours of infection and the seeding of higher numbers of cells in each well led to faster reversion. Reversion of HIV from the same batch of infected cells occurred within 2 weeks when 10⁴ infected cells were seeded but only after 3 weeks if 100 infected cells were seeded (Table 1). Therefore, the reversion rate in one round, based on an inoculum of 100 cells was at least 500-fold higher than that expected only on the basis of the RT mutation rate and the estimation that each cell is successfully infected by a single virus. Otherwise, a single cell would need to be infected by 500 viruses simultaneously, representing an impossible situation.

Virus concentration may also affect the rate and time of reversion as shown in Table 2. Reversion appeared faster in samples infected by higher rather than lower amounts of virus. There was no reversion in MT4 cells when diluted pseudotyped G367R was used to initiate infection. This is reasonable because the fewer the viruses, the fewer infected cells in the first round. These findings indicate that reversion of the mutant G367R virus is unlikely to be generated over a single round of infection and that multiple rounds of viral replication may be needed in order to increase the chances for a revertant to emerge.

Similar results to those obtained with PM-1 cells were obtained with both Jurkat cells, CEM cells and CBMCs when infected by the G367R mutant pseudotyped with VSV-G (data not shown).

To further investigate the mechanisms involved, we carried out several experiments to eliminate interference, if any, by the VSV-G protein and by HTLV-1 that is constitutively produced by MT2 cells. First, MT2 cells were directly transfected with the G367R proviral plasmid instead of by infection. Reversion in the G367R transfected MT2 cells was comparable in terms of time required to that of infection by the VSV-G pseudotyped G367R mutant, indicating that reversion does not seem to be influenced by the VSV-G protein. The fact that SupT1 cells infected by G367R pseudotyped with VSV-G reverted in the same manner as seen with infection of MT2 cells negates the likelihood that the HTLV-1 in the MT2 cells contributed to the reversion. In support of this, MT2 cells that were infected by R5 tropic subtype C Indie viruses that had been pseudotyped with VSV-G did not produce progeny that could infect MT2, MT4 or SupT1 cells, because such progeny viruses are unable to infect these cells in second round infections (data not shown). Therefore, these results do not support the idea that the HTLV-1 envelope protein can pseudotype with HIV, since, HTLV-1 pseudotyped R5 viruses should also be infectious were this to occur, consistent with our recent report [29].

Finally, supporting evidence comes from results obtained with 293 T and TZM-bl cells that were infected by VSV-G pseudotyped G367R virus and then cocultured with MT2 or MT4 cells. The results all indicate that reversion requires viral transmission through a cell-to-cell route, as shown in Table 4. MT2 cells cocultured with the G367R positive 293 T or TZM-bl cells developed CPE within about 1–2 weeks while the cells infected by G367R positive supernatants did not. Reversion in the cocultured MT2 cells was confirmed over a second round of cell-free infection. In contrast, MT4 cells that were cocultured with G367R positive 293 T or TZM-bl cells did not develop CPE over a 4 week period. However, phenotypic reversion did occur in the G367R positive TZM-bl cells that were maintained for up to 10 weeks in culture, since both supernatants and cocultures were infectious at the time points that were tested (Table 4). In contrast, the infection of 293 T cells did not lead to reversion on the basis of the infectivity of supernatants and p24 values that decreased at week 10 (data not shown). This indicates that G367R mutant virus can infect or reinfect TZM-bl cells but cannot infect 293 T cells.

Next, we investigated whether donor or target cells determine the efficiency of reversion of the G367R mutant. As shown above, coculturing G367R positive 293 T and TZM-bl cells, but not their supernatants, with MT2 or MT4 cells confirmed that the cell-associated the G367R virus can infect MT2 cells and revert to WT but cannot infect MT4 cells (Table 4). This result indicates that the type of target cell is important in determining the reversion of G367R mutant. The fact that MT4 cells infected by diluted G367R viruses that were pseudotyped with VSV-G did not revert is also in agreement with this result (Table 2). This finding also indicates that the mutant G367R HIV-1 cannot transmit if MT4 cells serve as both donor and target cells, as p24 values decreased gradually (Figure 3). In contrast, MT4 donor cells can transmit HIV-1 G367R mutant to MT2 target cells. In our experiments, CPE appeared after about 3 weeks in cocultures of MT4 cells containing G367R provirus, i.e., MT4 donor cells with uninfected MT2 cells. Reversion of phenotype was confirmed by a steep increase in p24 values in culture fluids (Figure 3) and the ability of the supernatants to participate in a second round of infection via cell-free transmission (data not shown). Thus, the G367R mutant has no deficit in ability to bud from MT4 cells. Moreover, it can be concluded that the efficiency of the G367R reversion is completely determined by the type of target cell used and not by donor cells.

To determine whether the observed reversion was related to virus entry, three entry inhibitors were employed to determine their effect on transmission and reversion. A gp120 binding inhibitor DS003 and a CXCR4 inhibitor AMD3100 were used in MT2 cultures infected by G367R viruses pseudotyped with VSV-G. A CCR5 inhibitor Maraviroc was used as a control. The results showed that DS003 was not able to inhibit G367R reversion when used at a fairly low concentration, i.e. 2 nM, a concentration around the IC₅₀ (3.1 nM as determined in an assay of cell-free infection with WT virus) (Figure 4a). However, DS003 was able to inhibit G367R reversion at concentrations of 20 nM and 200 nM, since the p24 values in culture supernatants increased by day 14 and then decreased. In contrast, p24 values of all the samples tested after withdrawal of DS003 at day 7, continued to increase (data not shown) and CPE developed within 3 weeks. However, these viruses developed resistance quickly since CPE appeared in one of three experiments at each of 20 nM and 200 nM DS003 at 5 weeks after infection, accompanied by an increase in p24 value (Table 5). The resultant progeny was resistant to DS003 in a second round of cell-free infection (data not shown).

The CXCR4 inhibitor AMD3100 also delayed reversion but less efficiently (Figure 4b). A concentration of 180 μM that was 10,000 times above the IC₅₀ of 7.8 nM (as determined in an assay of cell-free infection with WT virus) was needed to delay G367R reversion (Figure 4b). Again, the virus developed resistance quickly (5 weeks after infection) in one of three experiments (Table 5). As expected, Maraviroc did not have any inhibitory effect. Taking together, it appears that G367R reversion requires the entry of the G367R mutant into cells and that replication must precede reversion. This process can be inhibited by entry inhibitors.

HIV binding to the viral receptor and coreceptor can trigger viral fusion with the plasma membrane; however, the G367R mutant might have lost this ability. One mechanism of HIV entry into cells is dependent on endocytosis [30]. In order to explore whether the spread of the G367R mutant virus might be affected by inhibitors of endocytosis, we employed three such molecules in our experiments, i.e., chlorpromazine which is thought to inhibit clathrin-dependent (CDE) processes as well as two clathrin-independent endocytosis(CIE) inhibitors, i.e., Genistein and methyl-β-cyclodextrin (MβCD). Genistein is a tyrosine-kinase inhibitor and causes local disruption of the actin network at the site of endocytosis while the mechanism of MβCD is to inhibit the endocytosis process that is dependent on the integrity of lipid rafts.All three inhibitors affected cell viability. Only a two-fold higher concentration of chlorpromazine and Genistein resulted in cell death in MT2 than in MT4 cells (data not shown). Only one or two subtoxic concentrations of chlorpromazine and MβCD as tested in 2-fold serial dilutions permitted cell growth yet inhibited G367R virus replication (Figure 5). 5 μg/ml of chlorpromazine slowed increases in p24 values and reversion occurred in only in 2 of 4 experiments after 25 days of infection. Lower concentrations of drug did not inhibit viral reversion. A concentration of 2.5 μM MβCD slowed cell viral growth by about 2-fold and delayed viral replication only slightly. Reversion appeared in all 4 experiments after 25 days of infection (Figure 5b) and lower concentrations of drug had no effect on viral reversion. In contrast, subtoxic concentrations of Genistein inhibited cell growth between 6-40 mM and cells became larger and round, indicating that the drug disrupted the structure of the cell skeleton. However, viral replication was not depressed as much as cell growth because p24 values increased even though the cells grew slowly. Reversion appeared in almost all experiments after 25 days (Figure 5c). These results indicate that the spread of the G367R mutant requires the process of endocytosis and is related to CDE but not to CIE, similar to WT HIV, because this process was not inhibited by MβCD and was not affected as much as Genistein-mediated inhibition of cellular replication.

Based on the above results, G367R reversion may be determined by undefined cellular mechanisms. Contrary to expectations, we found that IL2 was able to inhibit the reversion of G367R virus in MT2 cells at the same concentration used to maintain T lymphocyte growth (Figure 6a) without affecting cell viability, while it only slightly inhibited cell-free wt HIV by about 2 fold (data not shown). In contrast, IL2 had no influence on the reversion of G367R in SupT1 cells, based on both p24 measurements and CPE (Figure 6a). Again, there was no growth or reversion of G367R in MT4 cells, suggesting that a different mechanism of reversion of G367R may be involved in different cells.However, PMA and ionomycin inhibited the reversion of G367R virus in both MT2 and SupT1 cells (Figure 6b). A low concentration of PMA of 0.1 ng/ml or ionomycin at 25 ng/ml both inhibited the G367R reversion. These results suggest that the reversion of G367R that is dependent on transmission between cells may be protein kinase C (PKC) related.