Cells and drugs.

Adherent cells (293T, TE671, HeLa, and Mus dunni tail fibroblasts) were maintained in Dulbecco's modified Eagle medium. A human T-cell line, CEM-SS, was maintained in RPMI medium. All media were supplemented with 10% fetal bovine serum and antibiotics. CsA (Bedford Laboratories) was prepared in dimethyl sulfoxide at 10 mg/ml and diluted in tissue culture medium to 2.5 μM for each experiment. Dextran sulfate (5 mg/ml; Sigma) was prepared in H₂O and was used at 100 μg/ml as described previously (8).

Viruses.

Vectors and viruses were produced by transfecting 10⁶ 293T cells/well in 6-well plates using Lipofectamine 2000 (Invitrogen). NL4-3_GFP vectors were produced by cotransfecting cells with 0.3 μg of pMDG and 3.7 μg of either pNL4-3_GFP or pNL4-3_GFPG89V. HIV-1_GFP wild-type and HIV-1_GFP A92E vectors were produced using 2 μg of CSGWΔN, 0.3 μg of pMDG, and 1.7 μg of p8.9NΔSB, either wild type or A92E mutant. Full-length infectious HIV-1 viruses were produced by transfecting 293T cells with 4 μg of pNL4-3 or pNL4-3/G89V. To produce MSCV-based shRNA vectors, 293T cells were cotransfected with 1.7 μg of pCL-Eco, 0.3 μg of pMDG, and 2 μg of either pSRP-CypA, pSRP-TRIM5, or pSRP-Luc. To produce HIV-1-based shRNA vectors, 293T cells were cotransfected with 1.7 μg of p8.9NΔSB, 0.3 μg of pMDG, and 2 μg of either pshRLP-TRIM5 or pshRLP-Luc. For production of N- or B-tropic MLV_GFP vectors, cells were cotransfected with 2 μg of pMIG, 0.3 μg of pMDG, and 1.7 μg of either pCIG3-N or pCIG3-B. For production of HIV-1 VLPs, 4.5 × 10⁶ 293T cells/10-cm-diameter plate were cotransfected with 18 μg of p8.9NΔSB and 2 μg of pMDG using Ca₂PO₄.

All virus- and VLP-containing supernatants were harvested 48 h posttransfection, spun at 300 × g for 5 min to remove cell debris, and filtered through a 0.45-μm-pore-size filter (Pall Acrodisc). For infectious NL4-3 viruses and VLPs, clarified supernatant was layered onto a 25% sucrose cushion and accelerated at 100,000 × g for 2 h in an SW41 rotor. The pelleted virions were resuspended in RPMI medium with 10% fetal bovine serum and stored at −80°C. Virions were normalized prior to infection by measuring reverse transcriptase (RT) activity as previously described (24).

Infection.

Adherent cells were plated in 48-well plates (2 × 10⁴ cells/well). Cells in suspension were seeded in 48-well plates at 8 × 10⁴ cells/well. Virion stocks were added to the cells in a final volume of 250 μl/well. GFP expression in infected cells was analyzed 48 h postinfection by flow cytometry (see below). For infections with full-length virus, dextran sulfate was added 16 h postinfection (100 μg/ml) to prevent spread of the virus (8). Forty-eight hours postinfection, cells were fixed in 4% formaldehyde-phosphate-buffered saline and assayed for intracellular p24 using a monoclonal anti-p24 antibody (183-H12-5C; NIH AIDS Research & Reference Reagent Program).

To generate TE671 and HeLa cells with stable expression from pSRP vectors, cells were seeded into 24-well plates at 3 × 10⁴ cells/well, covered with 500 μl of clarified 293T supernatant, and spinoculated for 70 min at 1,200 × g. Forty-eight hours posttransduction, cells were subjected to puromycin selection with quantum increases in drug concentration (1, 5, 10, 15, 20, 30, 40, and 50 ng/μl) over a period of 10 days.

To generate CEM-SS cells with stable expression from pshRLP vectors, cells were seeded into 24-well plates at 10⁵ cells/well and infected with 50 μl of clarified 293T supernatant in a total volume of 500 μl. All cells were subjected to puromycin selection 48 h posttransduction as described above.

Flow cytometry.

FACSCalibur and Cellquest Pro software (Becton Dickinson) were used to record GFP fluorescence in the FL1 channel. A quantity of 10⁴ live cells was analyzed per sample.

Inhibition of HIV-1 infectivity by factors that disrupt the CA-CypA interaction is TRIM5α_hu independent.

To investigate whether HIV-1 infectivity is more susceptible to TRIM5α_hu-mediated restriction when the CypA-CA interaction is disrupted, TE671-TR5-shRNA cells and TE671-Luc-shRNA cells were infected with HIV-1_GFP in the presence of CsA, a competitive inhibitor of the CA-CypA interaction (35). The magnitude of reduction of HIV-1 infectivity by CsA in TE671-TR5-shRNA cells was identical to the magnitude of reduction by the drug in the control TE671-Luc-shRNA cells (Fig. 2A and B). Similarly, the titer of HIV-1_GFP G89V, a CA mutant that disrupts interaction with CypA (67), was equally decreased in TE671-TR5-shRNA and TE671-Luc-shRNA cells (Fig. 2C and D). These results show that the reduction in HIV-1 infectivity that results from CsA or the G89V mutation does not result from increased susceptibility to TRIM5α_hu-mediated restriction.

Open in a separate window

FIG. 2.

Reduction of HIV-1 infectivity by CsA or CA mutant G89V is not mediated by TRIM5α_hu. VSV-G-pseudotyped, RT-normalized, HIV-1_GFP wild-type and G89V mutant virions were used to infect TE671-Luc-shRNA (A and C) and TE671-TR5-shRNA (B and D) cells in the presence or absence of CsA, as indicated. The percentage of GFP-positive (infected) cells was determined by flow cytometry. Shown are representative results of a single experiment. Identical results were obtained on three separate occasions using independently produced viral stocks.

Reduction of HIV-1 infectivity by CypA knockdown is independent of TRIM5α_hu.

To verify that the CsA-mediated decrease in HIV-1 titer resulted from the inhibitory effect of CsA on CypA, TE671 cells were transduced with a construct delivering an shRNA specific for CypA, as previously described (48, 51). After selection of a pool of transduced cells with puromycin, CypA was undetectable by Western blotting (Fig. ​(Fig.3A).3A). CypA downregulation resulted in a three- to fivefold decrease in HIV-1 infectivity compared to control cells (Fig. ​(Fig.3B).3B). The magnitude of reduction of HIV-1 infectivity in the context of the CypA knockdown was almost identical to the magnitude of reduction of HIV-1 by CsA in control TE671-Luc-shRNA cells (Fig. ​(Fig.3B),3B), indicating that CsA reduces HIV-1 infectivity via effects on CypA. As expected, since the G89V mutation disrupts the CypA binding site in CA, HIV-1 G89V infectivity was decreased no further by CypA knockdown (Fig. ​(Fig.3C3C).

Open in a separate window

FIG. 3.

Knockdown of CypA reproduces the effect of CsA and G89V on HIV-1 infectivity. TE671 cells were transduced with MLV-based vectors delivering shRNA expression constructs specific for CypA (CypA-shRNA) or luciferase (Luc-shRNA). (A) Cells were normalized by number, and CypA knockdown was demonstrated by Western blotting for CypA and β-actin. VSV-G-pseudotyped, RT-normalized, HIV-1_GFP wild-type (B) or G89V mutant (C) vectors were used to infect TE671-CypA-shRNA or TE671-Luc-shRNA cells in drug-free media or in the presence of 2.5 μM CsA, as indicated. The percentage of GFP-positive (infected) cells was determined by flow cytometry. Shown are representative results of a single experiment. Identical results were obtained on three separate occasions.

To directly test the interdependence of TRIM5α_hu and CypA in mediating HIV-1 restriction, we generated cell lines in which both TRIM5α_hu and CypA were downregulated using shRNA. To accomplish this, the singly transduced TE671-TR5-shRNA and TE671-Luc-shRNA cells (Fig. ​(Fig.1)1) were subjected to a second round of transduction either with the shRNA specific for CypA or with the control construct targeting luciferase. CypA knockdown was confirmed by Western blotting (Fig. ​(Fig.4A).4A). Both TE671-TR5/CypA-shRNA and TE671-TR5/Luc-shRNA cell lines still failed to restrict N-MLV, confirming that the TRIM5α_hu knockdown was stable (data not shown).

Open in a separate window

FIG. 4.

Reduction of HIV-1 infectivity by CypA knockdown is independent of TRIM5α_hu. TE671 cells were transduced with an MLV-based vector delivering an shRNA expression construct specific for human TRIM5 (TR5-shRNA) or luciferase (Luc-shRNA). Each pool of cells was then transduced a second time with shRNA vectors specific for CypA or luciferase. (A) CypA knockdown in the indicated pools of cells was demonstrated by Western blotting for CypA and β-actin. VSV-G-pseudotyped HIV-1 wild-type (WT) (B) or G89V mutant (C) virions were used to infect the indicated pools of cells. The percentage of GFP-positive (infected) cells was determined by flow cytometry. Shown are representative results of a single experiment. Identical results were obtained on three separate occasions.

As before (Fig. ​(Fig.11 and ​and3),3), cells in which TRIM5α_hu was downregulated were more permissive for HIV-1 infection than control cells, and cells in which CypA was knocked down were less permissive for HIV-1 infection (Fig. ​(Fig.4B).4B). The magnitude of the infectivity decrease due to CypA knockdown was the same in TE671-TR5/CypA-shRNA and TE671-Luc/CypA-shRNA cells, compared to the corresponding control cell lines TE671-TR5/Luc-shRNA and TE671-Luc/Luc-shRNA (Fig. ​(Fig.4B).4B). Similarly, the magnitude of the increase in infectivity due to disruption of TRIM5α_hu was the same whether CypA was expressed or not, or whether HIV-1 wild type or G89V was used for infection (Fig. 4B and C). As expected, infectivity of HIV-1 G89V remained unaltered by CypA knockdown (Fig. ​(Fig.4C).4C). These results demonstrate that TRIM5α_hu and CypA independently modulate HIV-1 infectivity. In other words, disruption of TRIM5α_hu does not rescue the reduction in HIV-1 infectivity caused by CypA knockdown.

CypA and TRIM5α_hu independently modulate HIV-1 infectivity in human CD4⁺ T cells.

To determine whether the results described above were peculiar to the TE671 rhabdomyosarcoma cell line, the CD4⁺ T-cell line CEM-SS was transduced with an HIV-1-based vector delivering a pSUPER-shRNA cassette (14) specific for human TRIM5 (54). A construct delivering an shRNA specific for luciferase (48) was used as a control. Pools of transduced cells were then selected in puromycin.

As described above for the TE671 cells, TRIM5α_hu knockdown in CEM-SS was confirmed by challenging transduced cells with N- and B-tropic MLV_GFP (Fig. 5A and B). HIV-1_GFP infectivity was reduced to the same extent by CsA in CEM-SS-TR5-shRNA as it was in CEM-SS-Luc-shRNA (Fig. 5C and D). Similarly, the titer of HIV-1_GFP G89V was decreased in both TRIM5α_hu knockdown and control cells (Fig. 5C and D). These results confirm that CypA acts independently of TRIM5α_hu to modulate HIV-1 infectivity in human CD4⁺ T cells.

Open in a separate window

FIG. 5.

Reduction of HIV-1 infectivity by CsA or the G89V mutant is independent of TRIM5α_hu in human T cells. CEM-SS cells were transduced with an HIV-based vector delivering an shRNA expression construct specific for human TRIM5 or luciferase as a control. VSV-G-pseudotyped, N- and B-tropic MLV_GFP virions were used to infect CEM-SS-Luc-shRNA cells (A) or CEM-SS-TR5-shRNA cells (B). VSV-G-pseudotyped, HIV-1_GFP wild-type, or G89V mutant vectors were used to infect CEM-SS-Luc-shRNA cells (C) and CEM-SS-TR5-shRNA cells (D) in drug-free medium or in the presence of 2.5 μM CsA, as indicated. The percentage of GFP-positive (infected) cells was determined by flow cytometry. Shown are representative results of a single experiment. Identical results were obtained on three separate occasions using independently produced viral stocks.

Effects of CypA and TRIM5α_hu on HIV-1 infectivity are independent of the route by which virions enter target cells.

Previous studies indicated that pseudotyping with VSV-G renders HIV-1 insensitive to CsA (3). VSV-G and the HIV-1 env-encoded glycoproteins (gp120/gp41) are believed to promote virion entry into target cells via two distinct mechanisms. VSV-G enters cells via an endocytic pathway, while the HIV-1 env glycoproteins mediate direct fusion into the target cell cytoplasm (5, 36, 53). To determine whether the functional interdependence of CypA and TRIM5α_hu is altered by the entry mechanism, CEM-SS-TR5-shRNA and CEM-SS-Luc-shRNA cells were challenged with full-length, infectious HIV-1_NL4-3. In the presence of CsA, or the G89V mutation, HIV-1 infectivity was decreased to the same extent in CEM-SS-TR5-shRNA cells as it was in control CEM-SS-Luc-shRNA cells (Fig. ​(Fig.6).6). These findings were confirmed in TE671 cells transfected with CD4 (data not shown). Thus, CypA and TRIM5α_hu act independently to regulate HIV-1 infectivity, irrespective of the route of entry utilized by the virus.

Open in a separate window

FIG. 6.

Promotion of HIV-1 infectivity by CypA is independent of TRIM5α_hu, regardless of the route of viral entry. HIV-1_NL4-3 virions were used to infect CEM-SS-Luc-shRNA (A) and CEM-SS-TR5-shRNA cells (B), in the presence or in absence of CsA, as indicated. HIV-1_NL4-3 virions, either wild type or G89V mutant, as indicated, were used to infect CEM-SS-Luc-shRNA (C) and CEM-SS-TR5-shRNA cells (D). The percentage of infected cells was determined by flow cytometry after immunostaining for CA. Shown are representative results of a single experiment. Identical results were obtained on three separate occasions using independently produced viral stocks.

Disruption of the CypA-CA interaction enhances the efficiency by which HIV-1 VLPs saturate restriction.

Prior to the discovery that TRIM5 restricts retroviruses in primate cells (48, 54), and prior to the widespread use of RNA interference in mammalian systems, HIV-1 VLPs were shown to enhance N-MLV infectivity, if the CA-CypA interaction was disrupted using CsA or the CA G89V mutation (61). Given that TRIM5α_hu exhibits a mild inhibitory effect on wild-type HIV-1 (Fig. ​(Fig.1),1), one would expect wild-type HIV-1 VLPs to also increase the N-MLV titer, though small phenotypes in such technically difficult saturation experiments might be difficult to detect. The effect of HIV-1 VLPs on N-MLV titer was reexamined here in the context of CypA and TRIM5 knockdown cells.

TE671-CypA-shRNA cells or control TE671-Luc-shRNA cells were infected with a constant amount of N-MLV_GFP in the presence of increasing amounts of HIV-1 VLPs, either wild type or G89V. Consistent with previous observations (61), HIV-1 G89V VLPs enhanced N-MLV titer in both TE671-CypA-shRNA and control TE671-Luc-shRNA cells (Fig. ​(Fig.7A).7A). In TE671-Luc-shRNA cells, wild-type HIV-1 VLPs increased N-MLV titer approximately threefold, though these effects were only evident at the highest quantities of VLPs administered (Fig. ​(Fig.7B).7B). The same VLPs were much more efficient at abrogating restriction to N-MLV in TE671-CypA-shRNA cells, indicating the importance for the saturation mechanism of a factor that selectively recognizes HIV-1 CA in the absence of CypA (Fig. ​(Fig.7B7B).

Open in a separate window

FIG. 7.

Abrogation of N-MLV restriction in TE671 cells by HIV-1 VLPs is enhanced by CypA knockdown. TE671-CypA-shRNA or TE671-Luc-shRNA cells were infected with constant amounts of N-MLV_GFP in the presence of increasing quantities of HIV-1 G89V VLPs (A) or HIV-1 WT VLPs (B). TE671-Luc-shRNA (C) or TE671-TR5-shRNA (D) cells were infected with constant amounts of HIV-1_GFP G89V in the presence of increasing quantities of N- or B-MLV VLPs. The percentage of GFP-positive (infected) cells was determined by flow cytometry. Identical results were obtained on three separate occasions using independent VLP stocks.

A VLP titration experiment was also performed in the context of TRIM5α_hu knockdown. Here cells were challenged with a fixed amount of HIV-1_GFP G89V in the presence of increasing quantities of either N-MLV or B-MLV VLPs. Compared with B-MLV VLPs, N-MLV VLPs specifically increased HIV-1_GFP G89V titer in TE671-Luc-shRNA cells (Fig. ​(Fig.7C).7C). No infectivity increase was observed in TE671-TR5-shRNA cells (Fig. ​(Fig.7D).7D). These results indicate that TRIM5α_hu is required for saturation of restriction machinery by N-MLV VLPs.

Effects of CypA on HIV-1 infectivity are independent of endogenous TRIM5α_hu.

Endogenous TRIM5α_hu and CypA both influence HIV-1 infectivity, but, in contrast to findings in Old World primates, where HIV-1 restriction by TRIM5α is CypA dependent (6), it was clearly demonstrated here that endogenous TRIM5α_hu and CypA influence HIV-1 infectivity independently of each other (Fig. ​(Fig.22 and ​and4).4). These findings held for HIV-1 infection of adherent cells and of CD4⁺ T cells (Fig. ​(Fig.5).5). In apparent contrast to a previous study with CsA (3), the effects of the two host factors were the same whether HIV-1 entered target cells via direct fusion with the cytoplasm or via an endocytic pathway (Fig. ​(Fig.6).6). The discrepancy with the earlier report can be explained by the fact that the phenomena studied here concern CypA's role in postentry events (26, 51, 61) and the previous study examined the effect of CsA when the drug was administered during virion production (3). The inhibitory mechanism of the drug given during virion production, in fact, does not involve disruption of CypA or the CA-CypA interaction (51).

Evidence that CypA protects HIV-1 from a restriction factor.

Since the discovery of the HIV-1 CA/CypA interaction (35), several models have been proposed to explain the role of CypA in HIV-1 replication. The finding that HIV-1 is particularly good at saturating the TRIM5α_hu-dependent restriction of N-MLV when CypA is removed (61) set the stage for the most recent model, namely, that CypA protects HIV-1 from restriction by TRIM5α_hu. This model was supported by the discovery of the owl monkey fusion protein TrimCyp (48), the existence of which hinted that the functions of CypA and TRIM5 were somehow associated. In cells from rhesus macaques and African green monkeys, the TRIM5α orthologues are indeed CypA dependent (6). Ever-more-convincing evidence was provided when a TE671 clone (17H1) that had been selected for loss of N-MLV restriction was found to be permissive for HIV-1 in a CypA-independent manner (47). Though all the results described above indicate that CypA protects HIV-1 from a host restriction factor, it was disappointing to find that neither TRIM5α_hu expression levels nor the TRIM5α_hu sequence was altered in clone 17H1 (47) (data not shown). To accommodate these reports, and the data presented here, the model requires modification: CypA does not protect HIV-1 from TRIM5α_hu but from an unknown restriction factor. Though much progress has been made over the past 13 years, the exact role of CypA in HIV-1 replication remains a mystery.

Possible mechanisms by which CypA regulates HIV-1 CA recognition by the restriction machinery.

CypA catalyzes the cis-trans isomerization of peptidyl-prolyl bonds (20, 33, 34, 56), and HIV-1 CA has been shown to be a catalytic substrate by nuclear magnetic resonance (11). In the absence of CypA, nearly 90% of the covalent bonds between glycine 89 and proline 90 are in the trans conformation (23). If restriction machinery only recognizes CA in its trans conformation, then CypA might protect HIV-1 by assisting in the formation of the restriction-resistant cis-isoform. Alternatively, CypA binding might mask CA residues or induce an allosteric change in CA that prevents recognition by the restriction machinery.

We thank David Sayah and Sarah Sebastian for reagents and ideas.

This work was supported by the National Institutes of Health grant RO1AI36199 to J.L.