Strain-Specific Differences in the Impact of Human TRIM5α, Different TRIM5α Alleles, and the Inhibition of Capsid-Cyclophilin A Interactions on the Infectivity of HIV-1

Reagents.Debio-025 was kindly provided by Debiopharma (Lausanne, Suisse), dissolved in dimethyl sulfoxide (DMSO; final concentration, 10 mM), and stored in aliquots at −80°C. Alpha interferon (IFN-α) (Sigma-Aldrich) was dissolved in deionized water (final concentration, 2 × 10⁵ U/ml) and stored in aliquots at −80°C.

Cell culture.Three cell lines were used as target cells. U373-X4 cells were derived from the human glioblastoma cell line U373-MG by sequential (i) transduction with a vector permitting constitutive expression of CD4 under the control of a retroviral long terminal repeat (LTR) promoter, (ii) transfection with a plasmid permitting tat-inducible expression of β-galactosidase, and (iii) transfection with a plasmid permitting constitutive expression of CXCR4 under the control of the cytomegalovirus (CMV) immediate-early promoter (28). P4 cells were derived from the human cervical carcinoma cell line HeLa by sequential (i) transfection with a plasmid permitting tat-inducible expression of β-galactosidase, and (ii) transduction with a vector permitting expression of CD4 under the control of the mouse phosphoglycerate kinase promoter (12). MT4-R5 cells were obtained by transducing the lymphoblastoid T-cell line MT4 with a lentiviral vector permitting constitutive expression of CCR5 under the control of the CMV promoter (4). These cell lines and 293T cells were cultured as previously described (36). The procedures used to generate U373-X4 cell lines expressing TRIM5γ or a pre-miRNA targeting TRIM5α, and CRFK cells expressing different TRIM5α alleles are described in the supplemental material. These cell lines were maintained in the presence of 5 μg/ml (CRFK cells) or 8 μg/ml (U373-X4 cells) blasticidin.

Peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll gradient centrifugation. Cells were resuspended at 7 × 10⁵ cells/ml in RPMI 1640 supplemented with 10% fetal calf serum, 100 U/ml penicillin G, and 100 μg/ml streptomycin, stimulated with 0.5 μg/ml purified phytohemagglutinin (PHA) (Murex, Kent, United Kingdom) for 24 h, and maintained for up to 10 days in medium containing 100 U/ml interleukin-2 (ProSpec, Revohot, Israel).

Recombinant viruses.The construction of the recombinant HIV-1 viruses has been described previously (36). Briefly, pNL4-3XCS-derived proviral vectors were used in which (i) a sequence coding for Renilla luciferase has been inserted in the place of nef, (ii) a 580-bp deletion has been introduced in env, and (iii) Gag-protease sequences from five different clinical isolates of HIV-1 had been cloned into the BssHII and ClaI restriction sites present in the original vector. All patient-derived sequences were obtained from individuals infected with subtype B viruses and who were naïve to antiretroviral therapy. None of the viral sequences had protease resistance mutations. To produce vesicular stomatitis virus (VSV)-pseudotyped viruses with Env deleted, 293T cells were cotransfected with the constructs and a VSV-G expression plasmid (phCMV-G). Virus-containing culture supernatants were harvested 48 h after transfection, filtered (0.45-μm pore size), and stored in aliquots at −80°C.

Recombinant N-MLV and B-tropic MLV (B-MLV) vectors were prepared as previously described (8), using plasmids provided by Jonathan Stoye.

Infectivity assays.HIV-1 infectivity was measured by determining luciferase activity in target cells 40 h after infection, as previously described (36), except that U373-X4 cells were initially plated at 2 × 10⁴ cells/96-well plate and CRFK cells were plated at 1 × 10⁴ cells/96-well plate. PBMCs (maintained in the continuous presence of 100 U/ml human interleukin-2) were plated at 1 × 10⁵ to 2 × 10⁵ cells/96-well plate, cultured in the presence or absence of 100 U/ml IFN-α for 20 h, and infected with viral suspensions containing 50 to 100 ng/well p24 in the presence of 4 μg/ml DEAE-dextran.

To measure MLV infectivity, U373-X4 cells (5 × 10⁴/well) or CRFK cells (1 × 10⁵/well) were cultured overnight in 12-well plates and infected with serial dilutions of N-MLV or B-MLV in the presence of 4 μg/ml DEAE-dextran. Forty hours after infection, cells were trypsinized, washed, and resuspended in 0.5 ml 1% paraformaldehyde in phosphate-buffered saline (PBS) without calcium or magnesium. The proportion of yellow fluorescent protein (YFP)-expressing cells was determined by cytofluorometry using an LSR cell analyzer (Becton Dickinson) fitted with an enhanced green fluorescent protein (EGFP)/YFP filter set (XCY-500, Omega Optical); data were analyzed using CellQuest-3 software.

Real-time PCR.Cell suspensions were washed with PBS, and mRNA was extracted from 2 × 10⁶ cells using NucleoSpin RNA II kits (Macherey-Nagel). CypA, TRIM5α, and GAPDH (glyceraldehyde-3-phosphate dehydrogenase) cDNAs were quantified by real-time PCR, using previously described techniques (36). Viral DNA produced during reverse transcription was measured by real-time PCR, using procedures described in the supplemental material.

Western blotting.Cell suspensions were washed two times with PBS, and 3 × 10⁶ cells were pelleted and resuspended in 75 μl of lysis buffer (1% NP-40, 150 mM NaCl, 1 mM EDTA, 1% protease inhibitor cocktail [Sigma], 20 mM HEPES, pH 7.8). Lysates were cleared by centrifugation (11,000 × g; 10 min), and protein content was determined using the Bradford assay (Bio-Rad). Cellular proteins (10 μg) were separated by electrophoresis into 10% SDS-PAGE gels, transferred to Protran BA85 membranes (Whatman), and blocked in Li-Cor blocking buffer (7 h at 20°C). To detect hemagglutinin (HA)-tagged TRIM5α isoforms, the membranes were probed with a peroxidase-conjugated rat anti-HA epitope antibody (3F10; Roche) (1/10,000 for 18 h at 4°C) and revealed using CCL Advance (GE Healthcare). To detect actin, used as a loading control, the same membrane was subsequently incubated with a rabbit antiactin antibody (A2066; Sigma) (1/5,000 for 1 h at 20°C), washed, incubated with IRDye 800CW-conjugated donkey anti-rabbit IgG antibody (1/10,000 for 45 min at 20°C), and scanned using an Odyssey Infrared imaging system (Li-Cor).

Effect of IFN-α pretreatment on viral infectivity in different target cells.The treatment of human cells by IFN-α increases the expression of TRIM5α (5, 10, 45). Thus, in an attempt to evaluate possible strain-specific differences in the susceptibility of HIV-1 to human TRIM5α, we first examined the impact of IFN-α on the infectivity of recombinant HIV-1 viruses expressing different Gag-protease sequences in human cells. We have previously described a series of VSV-pseudotyped pNL4-3-based recombinant viruses with env deleted in which the Gag-protease sequences were derived from clinical isolates obtained from patients who had never received protease inhibitors and in which a sequence coding Renilla luciferase was inserted in place of nef (36). The ability of these viruses to infect three different target cells (MT4-R5, HeLa-derived P4, and U373-X4 cells) pretreated for 20 h with a range of IFN-α concentrations was assessed by measuring luciferase expression 40 h after infection. The effect of IFN-α pretreatment was cell type dependent (Fig. 1). IFN-α pretreatment of MT4-R5 cells led to a strong, dose-dependent inhibition of the infectivity of all viruses tested. IFN-α pretreatment of P4 cells also led to a dose-dependent inhibition of viral infectivity compared to that of unpretreated cells, but the extent of this effect (20 to 30%) was considerably less than that observed in MT4-R5 cells. IFN-α pretreatment of U373-X4 cells resulted in an effect intermediate between that seen in MT4-R5 cells and P4 cells.

• Open in new tab
• Download powerpoint

FIG. 1.

Effect of pretreatment with IFN-α on the infectivity of recombinant viruses carrying Gag-protease sequences from clinical isolates. MT4-R5 cells (top), HeLa-derived P4 cells (middle), or U373-X4 cells (bottom) were pretreated with the indicated doses of IFN-α for 20 h and then infected with the indicated recombinant viruses (listed in the bottom panel). IFN-α treatment was renewed at the time of infection. Forty hours after infection, medium was removed, the cells were lysed, and luciferase activity was determined by luminometry. The results are the mean ± standard error of the mean (SEM) for three (MT4-R5 and P4 cells) or six (U373-X4 cells) experiments. The recombinant virus NRC10 was only studied in U373-X4 cells.

The extent that IFN-α pretreatment inhibited viral infectivity was virus dependent in U373-X4 cells (Fig. 1). Indeed, while the infectivities of NL4-3 and NRC3 were inhibited by 38% in U373-X4 cells pretreated with 100 U/ml IFN-α, the infectivities of NRC2, NRC9, and NRC10 were inhibited by 64 to 69% (P < 0.01 comparing NL4-3 with NRC2, NRC9, and NRC10). Apparent virus-specific differences in the effect of IFN-α pretreatment were also observed in MT4-R5 cells, although the magnitude of this effect was clearly less than that seen in U373-X4 cells, and the differences from NL4-3 were significant for NRC9 (P < 0.05), but not for NRC2. Very similar results were obtained when the three cell lines were infected using the same panel of recombinant viruses expressing the HIV-1 envelope (data not shown). Thus, the entry pathway used during infection did not influence these results.

The pretreatment of phytohemagglutinin (PHA)-stimulated human peripheral blood lymphocytes from two different donors with 100 U/ml IFN-α also led to a 4- to 5-fold decrease in viral infectivity compared to that observed in unpretreated cells (Fig. 2). Similar to results obtained with U373-X4 cells, the infectivities of NRC9 and NRC10 were inhibited to a significantly greater extent than NL4-3 in lymphocytes from both donors. As was observed in studies comparing MT4-R5 cells and U373-X4 cells (Fig. 1), however, the magnitudes of virus-specific differences in the effect of IFN-α pretreatment on infectivity appeared to be different in the two donors.

• Open in new tab
• Download powerpoint

FIG. 2.

Effect of pretreatment of PBMCs with IFN-α on the infectivity of recombinant viruses carrying different Gag-protease sequences. PBMCs from two donors were pretreated or not with 100 U/ml IFN-α for 20 h and infected with the indicated recombinant viruses. Forty hours after infection, the cells were lysed and luciferase activity was determined by luminometry. The results are expressed as a percentage of infectivity observed for cells not pretreated with IFN-α and are the mean ± SEM for 5 (donor 1) or 3 (donor 2) experiments. **, P < 0.01, and *, P < 0.05, compared to results obtained for NL4-3, using repeated-measures analysis of variance (ANOVA) followed by the Newman-Keuls multiple comparisons test.

The antiviral activity of IFN-α in U373-X4 cells is essentially mediated by TRIM5α.We next evaluated the extent to which TRIM5α accounted for the IFN-α-induced inhibition of viral replication in U373-X4 cells. To inhibit TRIM5α activity, two different approaches were used. TRIM5α mRNA expression was inhibited by transducing U373-X4 cells with a vector permitting expression of a pre-microRNA (pre-miRNA) targeting a sequence in the 3′-untranslated region of TRIM5α. Alternatively, TRIM5α activity was blocked by transducing U373-X4 cells with a vector resulting in overexpression of TRIM5γ. TRIM5 isoforms lacking a C-terminal PRYSPRY domain, such as TRIM5γ, heterodimerize with and exert a dominant-negative effect on endogenous TRIM5α activity (40, 50). To serve as controls, untransduced cells, cell lines overexpressing LacZ, and those expressing a pre-miRNA targeting LacZ were used. Evaluation of the replication of N-MLV and B-MLV in these cell lines indicated that TRIM5α activity had been reduced to undetectable levels in the U373-X4-TRIM5γ and U373-X4-TRIM5α-miRNA cell lines but was retained in the corresponding control cell lines (see Fig. S1 in the supplemental material).

IFN-α pretreatment of the cell lines expressing TRIM5α activity led to a dose-dependent decrease in viral replication (Fig. 3). As before, the extent of viral inhibition was virus dependent. In contrast, IFN-α pretreatment of the U373-X4-TRIM5γ cells had no significant effect on viral replication at any of the doses of IFN-α evaluated (P > 0.05 for all viruses). Similarly, the antiviral effect of IFN-α pretreatment was abolished in the U373-X4-TRIM5α-miRNA cell line.

• Open in new tab
• Download powerpoint

FIG. 3.

The IFN-α-induced antiretroviral activity in U373-X4 cells is TRIM5α. U373-X4 cells transduced with lentiviral vectors expressing an miRNA targeting TRIM5α (blue triangles), TRIM5γ (inverted blue triangles), an miRNA targeting LacZ (inverted red triangles), or LacZ (red diamonds) or untransduced cells (red triangles) were pretreated for 20 h with the indicated doses of IFN-α and infected with equal amounts (1 ng p24) of the indicated VSV-pseudotyped recombinant viruses carrying Gag-protease sequences from pNL4-3 or five different clinical isolates. Forty hours after infection, medium was removed, the cells were lysed, and luciferase activity was determined by luminometry. The results are the mean ± SEM for four experiments expressed relative to the infectivity of the virus in unpretreated cells.

Recombinant viruses carrying Gag-protease sequences from clinical isolates show differential sensitivity to TRIM5α in U373-X4 cells.Because TRIM5α is responsible for the IFN-α-induced effects on viral infectivity, the results shown in Fig. 1 suggest that the different recombinant viruses studied show differential sensitivity to human TRIM5α. Evaluation of the infectivity of these viruses in U373-X4 cells that did and did not express TRIM5α confirmed this conclusion. In cells not pretreated with IFN-α, the infectivity of viruses carrying Gag-protease sequences from NL4-3 and the clinical isolates NRC1 and NRC3 was not significantly increased comparing cell lines in which TRIM5α activity had been inhibited and the control cell lines. In contrast, inhibition of TRIM5α activity led to a 50% increase in the infectivity of NRC2 and NRC9 and a 300% increase in the infectivity of NRC10 (Fig. 4, open bars).

• Open in new tab
• Download powerpoint

FIG. 4.

Recombinant HIV-1 viruses expressing Gag-protease proteins derived from clinical isolates express differential sensitivity to TRIM5α. Untransduced U373-X4 cells or U373-X4 cells transduced with lentiviral vectors expressing an miRNA targeting LacZ, LacZ, an miRNA targeting TRIM5α, or TRIM5γ that were unpretreated (open bars) or had been pretreated with 100 U/ml IFN-α (solid bars) were infected with equal amounts (1 ng p24) of the indicated VSV-pseudotyped recombinant viruses carrying Gag-protease sequences from pNL4-3 or five different clinical isolates. Forty hours after infection, medium was removed, the cells were lysed, and luciferase activity was determined by luminometry. The results, expressed relative to the infectivity of the viruses in untransduced U373-X4 cells, are the mean ± SEM for five experiments. Asterisks indicate the results of statistical analyses comparing infectivity relative to untransduced cells as determined using ANOVA, followed by Dunnett's multiple-comparison test: *, P < 0.05; **, P < 0.01.

Following pretreatment of the cells with IFN-α, the infectivity of viruses carrying Gag-protease sequences from NL4-3 and the clinical isolates NRC1 and NRC3 was somewhat higher in cell lines in which TRIM5α activity had been inhibited (10%, 40%, and 30%, respectively), but these small differences did not achieve statistical significance, whereas the infectivity of NRC2, NRC9, and NRC10 was substantially increased in cell lines in which TRIM5α activity had been inhibited (250%, 250%, and 600% increases, respectively) (Fig. 4, solid bars).

Induction of TRIM5α by IFN-α in different human cell lines.The results in Fig. 1 demonstrated substantial differences in the impact of IFN-α pretreatment on the induction of antiviral activity in the different cell lines. To evaluate the possibility that differences in the extent that IFN-α induced TRIM5α expression could explain these findings, we measured the effect of IFN-α on TRIM5α mRNA levels. Treatment of MT4-R5 cells and U373-X4 cells for 8 h with 100 U/ml IFN-α led to a 4- to 5-fold increase in TRIM5α mRNA (Fig. 5 A). This induction appeared to be relatively transient and had decreased by 24 h in both cell types. In contrast, little induction of TRIM5α mRNA was observed in P4 cells at either time point. This finding offers a likely explanation for the very weak induction of antiviral activity by IFN-α observed in P4 cells (Fig. 1, middle panel). Treatment with 100 U/ml IFN-α for 8 or 24 h had no effect on CypA expression in any of the cell types (Fig. 5B).

• Open in new tab
• Download powerpoint

FIG. 5.

Effect of IFN-α on TRIM5α and CypA mRNA expression in different cell lines. MT4-R5 cells, HeLa-derived P4 cells, and U373-X4 cells were cultured in the presence or absence of 100 U/ml IFN-α for 8 h or 24 h. Cells were recovered, and mRNA was extracted and reverse transcribed using Moloney murine leukemia virus reverse transcriptase and random hexamers. TRIM5α, CypA, and GAPDH cDNA was quantified in each sample, and expression (copies/1,000 GAPDH sequences) was determined. The results show the mean ± SEM (n = 3) for the fold change in TRIM5α expression (A) and CypA expression (B) observed in IFN-α-treated cells compared to that of untreated cells.

U373-X4 cells and MT4-R5 cells express different TRIM5α alleles.Differences in the allelic forms of TRIM5α expressed by the cell lines are a possible explanation for the more pronounced differential sensitivity of the viruses to TRIM5α activity in U373-X4 cells. To begin to test this possibility, we sequenced the TRIM5α alleles expressed in each cell line (see Table S1 in the supplemental material). A single allelic form was identified in U373-X4 cells. Two allelic variants were identified in MT4-R5 cells, both of which differed from the allele identified in U373-X4 cells. One differed only by the Q136R polymorphism and the second differed by Q136R and G249D. HeLa-derived P4 cells also expressed two different alleles: one identical to that in U373-X4 cells and the second expressing the Q136R polymorphism.

Recombinant viruses carrying Gag-protease sequences from clinical isolates show differential sensitivity to TRIM5α allelic variants.To directly compare the sensitivity of the recombinant viruses carrying Gag-protease sequences from different isolates to allelic variants of TRIM5α, CRFK cells expressing N-terminal HA-tagged versions of each of the allelic variants present in U373-X4 cells and MT4-R5 cells were prepared. The level of TRIM5α expression was comparable in each of these cell lines (Fig. 6 A). As shown in Fig. 6B, the infectivity of NL4-3 was inhibited to a greater extent by the 136Q allele than by the 136R allelic variant. This difference has been previously described (21), but it was not observed in two other studies (17, 46). The 136R-249D variant also had reduced activity against NL4-3 compared to the 136Q allele.

• Open in new tab
• Download powerpoint

FIG. 6.

Recombinant HIV-1 viruses expressing Gag-protease proteins derived from clinical isolates express differential sensitivity to TRIM5α allelic variants. (A) Untransduced CRFK cells and those expressing the indicated HA-tagged TRIM5α allelic variants or LacZ were lysed, and 10 μg of cellular proteins was separated on 10% acrylamide gels. Proteins were transferred to Protran BA85 membranes, and TRIM5α and actin were detected as described in Materials and Methods. All samples were on the same membrane, but not in adjacent lanes. (B) Untransduced CRFK cells and those expressing the indicated TRIM5α allelic variants or LacZ were infected with equal amounts (1 ng p24) of the indicated recombinant viruses. Forty hours after infection, medium was removed, the cells were lysed, and luciferase activity was determined by luminometry. Results are the mean ± SEM (n = 4) expressed relative to the infectivity observed in cells transduced with the lentiviral vector expressing LacZ. *, P < 0.05, and **, P < 0.01, compared to the 136Q allele. †, P < 0.05, comparing the 136R and 136R-249D alleles as determined by ANOVA followed by the Newman-Keuls multiple-comparison test. (C) As in panel B, except that results are expressed as fold change in the reduction of antiviral activity relative to the 136Q allele. †, P < 0.05, comparing the 136R and 136R-249D alleles by paired t tests.

Evaluation of the activity of the different TRIM5α alleles against the recombinant viruses carrying Gag-protease sequences from clinical isolates showed virus-specific differences in sensitivity. The 136Q TRIM5α allele was more active than the 136R or 136R-249D alleles against all viruses. The reduction in activity of the 136R allele relative to the 136Q allele was similar for all viruses (Fig. 6C). In contrast, the reduction of the activity of the 136R-249D allele was virus specific. The reduction of the activity of this allele relative to the 136Q allele was not significantly different from that of the 136R allele for NL4-3, NRC3, and NRC1, but the reduction in the activity of the 136R-249D allele for NRC2, NRC9, and NRC10 was significantly greater than that of the 136R allele. Because the activity of the 136R-249D allele was preferentially reduced for the viruses that were most sensitive to the 136Q allele of TRIM5α, these findings offer an explanation for the observation that IFN-α pretreatment led to greater virus-specific differences in infectivity in U373-X4 cells than in MT4-R5 cells (Fig. 1).

CypA-CA interactions modulate TRIM5α activity in U373-X4 cells.Using the U373-X4 cell lines that did or did not express TRIM5α activity, we also evaluated the impact of CypA-CA interactions on TRIM5 activity. In these studies, cells were pretreated with 100 U/ml IFN-α in order to accentuate the differences in TRIM5α activity comparing the control cell lines and those either expressing the miRNA targeting TRIM5α or overexpressing TRIM5γ. In the cell lines not expressing TRIM5α activity (Fig. 7, blue curves), the shape of the Debio-025 dose-response curves was virus dependent. Incubation of cells in the presence of increasing doses of Debio-025 led to a progressive inhibition of NL4-3, NRC3, and NRC9, although maximal inhibition was less for NRC9 than NL4-3 or NRC3. Debio-025 had no significant effect on the replication of NRC1. For NRC2 and NRC10, low doses of Debio-025 (40 nM) improved viral replication to varous extents. In cells incubated with 1 μM Debio-025, the replication of NRC2, but not NRC10, remained higher than that seen in untreated cells.

• Open in new tab
• Download powerpoint

FIG. 7.

CypA-CA interactions modulate the sensitivity of viruses to TRIM5α. U373-X4 cells transduced with lentiviral vectors expressing a miRNA targeting TRIM5α (blue triangles), TRIM5γ (inverted blue triangles), an miRNA targeting LacZ (inverted red triangles), or LacZ (red diamonds) or untransduced cells (red triangles) were pretreated for 20 h with 100 U/ml IFN-α and infected with equal amounts (1 ng p24) of the indicated VSV-pseudotyped recombinant viruses in the presence of the indicated doses of Debio-025. Forty hours after infection, medium was removed, the cells were lysed, and luciferase activity was determined by luminometry. The results, expressed relative to the infectivity of that of cultures not treated with Debio-025, are the mean ± SEM for 3 experiments. Asterisks indicate the results of statistical analyses comparing infectivity in the transduced cell lines relative to that in untransduced cells treated with 1 μM Debio-025, as determined using ANOVA, followed by Dunnett's multiple comparison test: *, P < 0.05; **, P < 0.01.

When cell lines expressing TRIM5α were infected (Fig. 7, red curves), the inhibition of infectivity by Debio-025 was significantly greater for NL4-3, NRC3, NRC2, and NRC9, indicating that inhibition of CypA-CA interactions rendered these viruses more susceptible to TRIM5α activity. In contrast, in the presence of Debio-025, the infectivity of NRC10 was significantly greater than that observed in cells not treated with Debio-025. Thus, for NRC10, under conditions where CypA-CA interactions were permitted, viral replication was quite sensitive to inhibition by TRIM5α (Fig. 3 and 4), but under conditions where CypA-CA interactions were inhibited, viral replication became less sensitive to this effect.

Antiviral activities induced by IFN-α and revealed by Debio-025 act at the same stage of the viral life cycle in MT4-R5 cells.Both pretreatment with IFN-α and treatment with Debio-025 lead to an inhibition of the replication of NL4-3 in MT4-R5 cells. To characterize the step in the viral life cycle at which these effects occur, we evaluated viral DNA synthesis and two-LTR circle formation at various times after infection using real-time PCR. In these studies, the recombinant virus with envelope deleted carrying the NL4-3 Gag sequence was used. Because MT4-R5 cells have been transduced with a lentiviral vector carrying the CCR5 receptor, early DNA synthesis could not be evaluated by measuring strong-stop DNA. Thus, we evaluated the synthesis of a sequence in the 3′ end of luciferase just upstream of the 3′ polypurine tract. Following infection, synthesis of the minus strand of this region begins within the first hour after infection, whereas plus-strand synthesis occurs between 2 and 4 h after infection (54). To evaluate later steps in DNA synthesis, a sequence in the 5′ portion of gag was evaluated. Synthesis of both the minus strand and plus strand of this region occurs between 2 and 5 h after infection. The kinetics of DNA synthesis and two-LTR circle formation are shown in Fig. S2 in the supplemental material. Synthesis of the luc gene began within the first hour and was complete by 5 h. Synthesis of gag was detectable at 5 h, but not at 1 h after infection. In untreated MT4-R5 cells, gag synthesis increased somewhat between 5 and 20 h, but this was not observed in cells treated with IFN-α and/or Debio-025. Two-LTR circles were barely detectable at 5 h and accumulated over the ensuing 15 h.

As illustrated in Fig. 8, 1 h after infection, only a modest deficit in viral DNA synthesis was present for cells treated with IFN-α and/or Debio-025 (Fig. 8A). This deficit progressively increased for each treatment condition over the ensuing 15 h, reaching a maximum for gag synthesis at 20 h (Fig. 8B). The deficit in the formation of two-LTR circles in cells treated with IFN and/or Debio-025 appeared to parallel that observed for the synthesis of gag (Fig. 8C), and the extent of inhibition of gag synthesis and two-LTR formation at 20 h were not significantly different for cells treated with IFN-α and/or Debio-025 (P > 0.27 for all comparisons). Thus, we found no evidence for an additional block occurring during nuclear import or integration. The deficit in gag synthesis also appeared to fully explain the reduction in the expression of luciferase activity measured in parallel cultures at 40 h (Fig. 8D), because the extents of inhibition of gag synthesis and luciferase activity were not significantly different in cells treated with IFN-α and/or Debio-025 (P > 0.19 for all comparisons).

• Open in new tab
• Download powerpoint

FIG. 8.

Comparison of the effects of IFN-α and Debio-025 on viral DNA synthesis and infectivity. The synthesis of viral DNA corresponding to Renilla luciferase (luc) at 1 h (A), gag at 20 h (B), and two-LTR circles at 20 h (C) was measured by real-time PCR as described in the supplemental material. (D) Forty hours after infection, medium was removed from parallel cultures, the cells were lysed, and luciferase activity was determined by luminometry.

TRIM5α and IFN-α.We found that IFN-α pretreatment increased the expression of TRIM5α mRNA in U373-X4 cells and MT4-R5 cells and was associated with increased antiviral activity against HIV-1 in both of these cell lines, as well as in mitogen-stimulated human peripheral blood lymphocytes. Further, the IFN-α-induced increase in antiviral activity in U373-X4 cells was entirely absent in cells in which TRIM5α activity had been inhibited. Several authors have previously shown that IFN-α treatment increases the expression of TRIM5α in various cell types (5, 10, 45). Although IFN-α pretreatment of TE671 cells and HeLa cells has been shown to increase the antiviral activity against N-MLV, no increases in antiviral activity against HIV-1 expressing the NL4-3 CA were observed (10, 45). We also found that IFN-α-pretreatment of HeLa-derived P4 cells had little impact on antiviral activity, but in our study, this appeared to reflect poor stimulation of TRIM5α expression in P4 cells, which is not the case for HeLa cells (10, 45). As previously discussed (10), several parameters can influence the extent that increasing TRIM5α expression in target cells by IFN-α modulates HIV-1 infectivity, including the basal levels of TRIM5α expression, the amount of virus used to infect the cells, and the intrinsic sensitivity of a given isolate to the TRIM5α allele expressed by the target cells. These factors are likely to explain the differences between our results and previously published studies. Differences in these factors may also account for our observation that IFN-α pretreatment had a greater impact on the replication of NL4-3 in MT4-R5 cells than in U373-X4 cells. In this regard, NL4-3 is intrinsically less sensitive to the TRIM5α alleles expressed by MT4-R5 cells than that expressed in U373-X4 cells, and it might be expected that increasing TRIM5α expression in MT4-R5 cells would have a proportionally greater effect. In U373-X4 cells, TRIM5α appears to be the principal factor inhibiting HIV-1 replication during the early steps of viral infection up to integration, consistent with the finding that TRIM5α is the main mediator of IFN-α-induced N-MLV restriction in HeLa and Vero cells (10). These observations need to be extended to other cell types.

Viral sensitivity to TRIM5α.An important finding in our study was the observation that viruses carrying Gag sequences from primary isolates could show differences in their sensitivity to human TRIM5α. For the most sensitive isolate studied (NRC10), infectivity in IFN-α-treated U373-X4 cells was reduced 6- to 8-fold compared to that observed in cells not expressing TRIM5α activity, whereas the infectivity of NL4-3 was reduced only 1.2-fold comparing cells that did or did not express TRIM5α activity. The mechanism responsible for the variability in the sensitivity of these isolates to TRIM5α remains to be defined. We have found, however, that three different clonal isolates from NRC10 showed similar sensitivities to TRIM5α, suggesting that this property is shared by the swarm of viruses in this patient and is not due to a rare deleterious point mutation present only in a single clone (data not shown). Further studies are required to identify the viral determinants that modulate sensitivity to human TRIM5α.

Activity of TRIM5α alleles.Our results indicate that the sensitivity of a given viral isolate to TRIM5α is also dependent on the TRIM5α allele that it encounters. The sensitivity of the six different viruses to the three allelic forms studied here followed the same order, but the 136Q allele had significantly greater activity than either the 136R allele or the 136R-249D allele. In addition, the decrease in activity of the 136R-249D allele relative to the 136Q allele was more pronounced for viruses that displayed increased sensitivity to TRIM5α. The mechanisms accounting for the allelic differences in TRIM5α activity have not been fully defined (21). The presence of polymorphisms in the coiled-coil region (Q136R) may affect dimerization (22, 37, 41) and therefore impact cellular levels of functional TRIM5α multimers. The L2 linker region has been found to play a role in CA recognition (30, 51), and therefore polymorphisms in this region (G249D) may influence TRIM5α binding. These considerations could explain the differences in the activities of the 136Q and 136R alleles and why the impaired activity of the 136R-249D allele was more pronounced against TRIM5α-sensitive viruses.

TRIM5α-CypA interactions.A novel finding in these studies was the observation that CypA-CA interactions modulate the activity of human TRIM5α. We found that the inhibition of CypA-CA interactions increased the antiviral activity of TRIM5α against 4 of the 6 viruses tested, but viruses for which inhibiting CypA-CA interactions either had no effect on TRIM5α activity (NRC1) or led to a paradoxical increase in viral infectivity (NRC10) were also identified. Several groups have reported that inhibition of CypA-CA interactions appeared to have no effect on the activity of human TRIM5α (23, 47, 52). The discrepancy with our results is likely due to two reasons. First, these groups evaluated viruses carrying the CA protein from either the NL4-3 strain or the closely related HXB2 laboratory-adapted strains. Although we did find that blocking CypA-CA interactions increased the sensitivity of NL4-3 to the activity of TRIM5α, NL4-3 is relatively resistant to human TRIM5α, and the differences were more pronounced in cells in which TRIM5α expression had been increased by IFN-α pretreatment. Second, prior studies used either TE671 cells, which express the 136R TRIM5α allele (19), or canine Cf2Th cells expressing the 136R TRIM5α allele. Because NL4-3 is less sensitive to this TRIM5α allele than the 136Q allele expressed by U373-X4 cells, the effect of blocking CypA-CA interactions would be further attenuated.

Our findings provide insights into the potential mechanisms through which CypA and TRIM5α interact to influence viral infectivity. For viruses in which inhibiting CypA-CA interactions reduced viral infectivity (NL4-3, NRC3, and NRC9), TRIM5α produced an additional 50% decrease in infectivity, because CypA binding inhibited, in part, the interaction of TRIM5α with the CA, and/or because the CypA-deficient CAs became more sensitive to the destabilizing effects of TRIM5α. For other viruses, however, inhibition of CypA-CA interactions in TRIM5α-expressing cells either had little impact or significantly increased infectivity. Because CypA and TRIM5α interact with adjacent regions of the viral CA (33, 39), it is possible that CypA binding to the viral CA directly or indirectly inhibits the recognition of the viral CA by TRIM5α and/or modulates the impact of TRIM5α on viral infectivity. This model does not easily explain our observations for the virus NRC10. For this virus, blocking CypA-CA interactions in cells expressing TRIM5α led to an increase in viral replication. CypA binding may, under some circumstances, facilitate the recognition of the CA by TRIM5α and/or render the capsid more sensitive to the destabilizing activity of TRIM5α. In this case, the inhibition of CypA-CA interactions would be expected to improve viral replication. Indeed, this mechanism has been proposed as an explanation for the observation that the inhibitory effects of rhesus TRIM5α on HIV-1 replication are partially blocked by inhibiting CypA-CA interactions (23, 39, 52, 56). NRC1, which carries mutations in the CypA-binding loop that render the virus insensitive to the effects of CypA on infectivity (11, 16, 20, 26, 36, 39, 51), was the only virus for which modulation of CypA-CA interactions had no impact on its sensitivity to TRIM5α. This finding would also be expected if the effects of CypA-CA interactions on CA stability were responsible for modulating sensitivity to TRIM5α.

Consistent with our finding that CypA-CA interactions influence TRIM5α activity, we found that the effects of increasing TRIM5α expression by IFN-α and the inhibition of CypA-CA interactions on the replication of NL4-3 occurred at the same point in the viral life cycle, and the two interventions reinforced each other. The inhibition of DNA synthesis fully explained the impairment in viral infectivity, and no additional impact of either intervention on two-LTR circle formation was observed, suggesting that these interventions did not influence nuclear transport or integration.

Taken together, our findings support a model in which the sequence of the viral CA determines numerous independent but interlinked properties that can influence viral CA stability, and therefore viral infectivity, including (i) the intrinsic stability of the CA, (ii) the extent that CypA binding modulates CA stability, (iii) the direction of the effect (increased or decreased stability) of CypA binding, (iv) the sensitivity of the virus to a particular TRIM5α allele, and (v) the extent that CypA binding modulates sensitivity to TRIM5α.