Characterization of the sensitivity of the HIV-1 V86M capsid mutant to TRIM5α_rh

Valine 86 is located in the cyclophilin A-binding loop of the HIV-1 capsid protein (Figure 1, A and B). Changes in the cyclophilin-binding loop of the capsid have been previously shown to reduce HIV-1 sensitivity to early-acting restriction factors in Old World monkey and owl monkey cells (Chatterji et al., 2005; Gatanaga et al., 2006; Hatziioannou et al., 2004; Ikeda et al., 2004; Kootstra et al., 2003, 2007; Nagao et al., 2009; Owens et al., 2004). Of note, some of these changes involve the histidine 87 residue, which is adjacent to valine 86. Other changes involving glycine 89 and proline 90 in the capsid cyclophilin-binding loop disrupt the binding of cyclophilin A, which in some cases appears to potentiate TRIM5α_rh restriction of HIV-1 (Berthoux et al., 2005; Stremlau et al., 2006b; Towers et al., 2003). We compared the effect of some of these capsid changes with that of the V86M change on HIV-1 infection of HeLa-CD4 cells expressing TRIM5α_rh. The effect of the capsid changes on the replication of the HIV-1_NL4-KB9 SEMQ virus was examined; the SEMQ alteration in Vif is reported to overcome partially the replication block conferred by rhesus monkey APOBEC3G (Schrofelbauer et al., 2006). The replication of the HIV-1_NL4-KB9 SEMQ virus with the wild-type and mutant capsids in HeLa-CD4-TRIM5α_rh cells is shown in Figure 2A. HIV-1 viruses containing the V86M or H87Q changes replicated efficiently in these cells, reaching a peak of RT activity around days 15 and 19, respectively. Beginning at approximately 19 days after infection, RT activity was detected in the culture infected by the P90 mutant; the RT activity in this culture reached a peak around day 33 post-infection. Reverse transcriptase remained at background levels in cultures inoculated with HIV-1_NL4-KB9 SEMQ bearing the wild-type capsids, even after 40 days in culture. In parallel, as a control, we incubated this panel of viruses with Cf2Th cells stably expressing human CD4 and CXCR4, but not TRIM5α_rh. All the viruses replicated efficiently in this cell line (Figure 2B); the delay in replication of the P90A mutant is consistent with the consequences of poor cyclophilin A binding. These results demonstrate that the V86M capsid change associated with virus passage in TRIM5α_rh-expressing cells increases the resistance of HIV-1 to the inhibitory effects of rhesus monkey TRIM5α.

Open in a separate window

Figure 1

The cyclophilin A-binding loop of the HIV-1 capsid protein. (A) The ribbon structure of a cyclophilin A-HIV-1 p24 capsid protein complex is shown (Gamble et al., 1996). The cyclophilin A-binding loop is colored yellow. The complex is oriented so that, in an assembled capsid, the exposed surface of the capsid protein faces the viewer. (B) The HIV-1 capsid protein is oriented as in (A), with the side chains of the amino acid residues relevant to this study colored according to the following key: valine 86 (red), other residues in the cyclophilin A-binding loop that have been shown to influence HIV-1 sensitivity to rhesus monkey restriction (green) (Owens et al., 2004), and residues that directly contact cyclophilin A (magenta). (C) Three HIV-1 capsid hexamers are shown as they are related in a crystal lattice, approximating their relationship in an assembled capsid (Pornillos et al., 2009). The surface of the assembled capsid faces the viewer, and the cyclophilin A-binding loops (yellow) and valine 86 residues (red) are highlighted. The right panel provides a close-up of the region at the intersection of the three hexamers in the left panel. The local 3-fold (triangle) and 2-fold (ellipses) axes of symmetry are shown. The cyclophilin A-binding loops at the distal end of each hexamer spoke are located near the depressions on the surface of the assembled capsid that track along the 2- and 3-fold axes of pseudosymmetry.

Open in a separate window

Figure 2

Replication of HIV-1 capsid mutants in HeLa-CD4 cells stably expressing TRIM5α_rh. (A,B) HeLa-CD4 cells expressing TRIM5α_rh (A) or control Cf2Th-CD4/CXCR4 cells (B) were incubated with 30,000 RT units of the indicated HIV-1_NL4-KB9 SEMQ variants for 14 hours and then washed once with PBS. Every 3 or 4 days, cell supernatants were removed and used for RT assays. Cells were then trypsinized, diluted 1/10 in fresh medium and replated. (C,D) Recombinant, single-round HIV-1 expressing GFP were incubated with HeLa-CD4 cells expressing TRIM5α_rh (C) or control HeLa-CD4 cells transduced with the empty LPCX vector (D). GFP expression was measured 48 hours later.

To examine whether the combination of the V86M or the H87Q capsid changes with the Vif_SEMQ changes were sufficient to allow HIV-1 to replicate in rhesus macaque cells, we infected the 221-89 lymphoid cell line (Alexander et al., 1997) and rhesus monkey PBMCs with HIV-1_NL4-KB9 viruses containing these changes. Virus replication was not detected in this rhesus T cell line nor in rhesus macaque PBMCs (data not shown). This result suggests that additional changes in HIV-1 are needed to allow efficient replication in rhesus monkey cells. This result was not unexpected, given the existence of rhesus monkey restriction factors such as BST2 (tetherin) that operate during the late phase of HIV-1 replication (Neil et al., 2008; Van Damme et al., 2008). Rhesus monkey BST2 is not effectively countered by the HIV-1 Vpu protein (Jia et al., 2009; McNatt et al., 2009). Moreover, the SEMQ changes in the Vif protein may be insufficient to allow HIV-1 to overcome APOBEC3G restriction and to replicate in the cells of Old World monkeys (Hatcho et al., 2008; Kamada et al., 2009).

To examine the ability of V86M and other capsid variants to escape specifically from the early restriction imposed by TRIM5α_rh, we infected cells with recombinant HIV-1 vectors containing the capsid changes. These viruses were pseudotyped with the vesicular stomatitis virus (VSV) G glycoprotein, which allows entry into a wide range of vertebrate cells. The recombinant viruses express either green fluorescent protein (GFP) or luciferase. The V86M change in the capsid protein allowed more efficient single-round infection of HeLa-CD4-TRIM5α_rh cells than was observed for the wild-type recombinant HIV-1 vector (Figure 2C). Because the R20W change in the matrix protein appeared in two clones during an early passage of the virus in HeLa-CD4-TRIM5α_rh cells, we also tested the contribution of this change. An HIV-1 vector containing the R20W change in the matrix protein in addition to the V86M change in capsid infected the HeLa-CD4-TRIM5α_rh cells similarly to the virus with only the V86M change. All three recombinant viruses infected the control HeLa-CD4 cells efficiently (Figure 2D). We conclude that the V86M change in the capsid protein confers an advantage during the infection of HeLa cells expressing TRIM5α_rh.

HeLa cells express human TRIM5α, which only minimally inhibits HIV-1 infection (Stremlau et al., 2004); the presence of human TRIM5α might influence the potency with which TRIM5α_rh restricts HIV-1 infection. Therefore, we used single-round, luciferase-expressing HIV-1 to examine the contribution of the V86M capsid change to the early phase of infection of another cell type expressing TRIM5α_rh. We compared the infectivity of recombinant HIV-1 with the V86M change, or with different changes in the capsid previously reported to affect sensitivity to TRIM5α restriction or cyclophilin A binding (Chatterji et al., 2005; Gatanaga et al., 2006; Hatziioannou et al., 2004; Ikeda et al., 2004; Kootstra et al., 2004, 2007; Owens et al., 2004; Towers et al., 2003). Canine Cf2Th cells, which do not express an endogenous TRIM5α protein (Sawyer et al., 2007), were used as target cells; in addition to the control cells transduced with the empty LPCX vector, we used Cf2Th cells expressing TRIM5α_rh and Cf2Th cells expressing the owl monkey TRIMCyp protein. The latter protein is expressed in the New World owl monkey lineage as a result of a retrotransposition that fuses the amino-terminal part of TRIM5 with the cyclophilin A protein (Nisole et al., 2004; Sayah et al., 2004). In control canine Cf2Th cells transduced with an empty LPCX vector, all of the HIV-1 capsid mutants, with the exception of H87Q, infected less efficiently than the wild-type HIV-1 (Figure 3A). An HIV-1 vector (SCA) containing the SIV_mac capsid (Owens et al., 2003) also infected these control cells less efficiently than wild-type HIV-1. In Cf2Th cells expressing TRIM5α_rh, the SCA virus exhibited a higher level of infectivity than wild-type HIV-1. The V86M and H87Q HIV-1 variants also infected these cells slightly more efficiently than wild-type HIV-1 (Figure 3B). The G89A and P90A HIV-1 capsid mutants, which are defective for cyclophilin A binding (Braaten et al., 1996; Franke et al., 1994), infected the TRIM5α_rh-expressing cells poorly. By contrast, these HIV-1 variants, as well as SCA, infected Cf2Th cells expressing owl monkey TRIMCyp much more efficiently than the other viruses, including wild-type HIV-1 (Figure 3C). This was expected, as neither the SIV_mac nor the G89A or P90A HIV-1 capsids bind TRIMCyp (Hatziioannou et al., 2004; Sayah et al., 2004; Towers et al., 2003). These results indicate that the V86M and H87Q changes in the HIV-1 capsid are specifically advantageous during the infection of Cf2Th-TRIM5α_rh cells.

Open in a separate window

Figure 3

Infectivity of HIV-1 capsid mutants in different cell lines. Target cells were infected with single-round recombinant reporter viruses expressing firefly luciferase. Forty-eight hours after infection, cells were lysed and the luciferase activity was measured using an EG&G Berthold LB 96V microplate luminometer. The target cells were: Cf2Th cells stably transduced with the empty LPCX vector (A), Cf2Th cells stably expressing TRIM5α_rh (B); Cf2Th cells stably expressing owl monkey TRIMCyp (C); primary rhesus monkey lung (PRL) fibroblasts (D), rhesus monkey kidney (MK2D) cells (E); and owl monkey kidney (OMK) cells (F). RLU, relative luciferase units.

We also examined the ability of these single-round recombinant viruses to infect primary rhesus monkey lung (PRL) fibroblasts, MK2D rhesus macaque cells and the OMK owl monkey kidney cell line. The PRL and MK2D cells express TRIM5α_rh, whereas the OMK cells express TRIMCyp from owl monkeys, a New World monkey species (Sayah et al., 2004; Stremlau et al., 2004; Wilson et al., 2008). Wild-type HIV-1 infection of PRL cells was nearly undetectable, whereas SCA infected these cells very efficiently (Figure 3D). The V86M and H87Q HIV-1 mutants and, to a slightly lesser extent, the G89A and P90A mutants, infected the PRL cells more efficiently than wild-type HIV-1. In the MK2D rhesus monkey cells, infection by wild-type HIV-1 and the G89A and P90A mutants was much less efficient than infection mediated by the V86M and H87Q mutants. In the OMK cells from the owl monkey, only SCA and the G89A and P90A HIV-1 mutants infected efficiently. Thus, as was observed in the Cf2Th cells above, the V86M and H87Q changes in the HIV-1 capsid confer a specific advantage during the early phase of infection of cells expressing rhesus monkey TRIM5α.

Competition of wild-type and mutant virus-like particles for restriction factors

Previous studies have shown that the restriction of HIV-1 in Old World monkey cells can be saturated by incubation of the cells with enveloped virus-like particles (VLPs) (Cowan et al., 2002; Hatziioannou et al., 2003; Munk et al., 2002). To test whether the HIV-1 capsid mutants were able to compete for restriction factors in primary rhesus monkey cells, we incubated PRL cells with the wild-type HIV-1 reporter virus encoding firefly luciferase along with wild-type or mutant VLPs. The presence of wild-type HIV-1 VLPs increased the infectivity of the HIV-1 reporter virus (Figure 4), suggesting that these VLPs are able to compete for the binding to restriction factors. The V86M, H87Q, G89A and P90A mutant VLPs exhibited reduced ability to compete for restriction factors in the PRL cells, compared with the wild-type HIV-1 VLPs. The YQ VLPs, which contain two changes in the HIV-1 capsid that reduce susceptibility to Old World monkey TRIM5α without affecting restriction factor binding (Owens et al. 2004), competed efficiently for restricting activity. As expected, HIV-1 VLPs without envelope glycoproteins or the SCA VLPs, which have an SIV_mac capsid in an HIV-1 background (Owens et al., 2003), did not compete for the PRL restriction factor(s). The results suggest that all of the HIV-1 VLPs with changes in the cyclophilin-binding loop interact less efficiently with restriction factors following entry into PRL cells.

Open in a separate window

Figure 4

Competition assay with HIV-1 VLPs. Target PRL cells were infected with 2,500 RT units of the wild-type recombinant HIV-1 expressing luciferase in the presence of 100,000 RT units of virus-like particles (VLPs) lacking a reporter gene. The capsid proteins of the VLPs were either wild-type HIV-1 (wt) or contained the indicated changes. All VLPs were pseudotyped with the VSV G glycoprotein, except control VLPs with wild-type HIV-1 capsids but with no added envelope glycoproteins (no Env). Forty-eight hours after infection, cells were lysed and the luciferase activity was measured.

Virus replication

Replication-competent HIV-1 viruses were generated by transfecting 20 μg of the pNL4-3 plasmid, which contains an infectious HIV-1_NL4-3 provirus (Adachi et al., 1986), into 2×10⁶ 293T cells using the calcium phosphate transfection method (Invitrogen). Forty-eight hours after transfection, supernatants containing viruses were harvested and cleared by low-speed centrifugation. The level of virus in the supernatant was determined by measuring reverse transcriptase (RT), as described previously (Rho et al., 1981).

Cells were infected with 30,000 RT units of HIV-1_NL4-KB9 variants for 14 hours and then washed once with PBS. Every 3 or 4 days, cell supernatants were removed and used for RT assays. Cells were trypsinized, diluted 1/10 in fresh medium and replated.

Selected gag mutations were introduced into the HIV-1_NL4-KB9 provirus encoding the SEMQ Vif variant (Schrofelbauer et al., 2006). Infectious viruses were generated and tested for replicative ability as described above.

Analysis of Gag and Pol sequences

The gag and pol genes of the adapted viruses were amplified by polymerase chain reaction (PCR) of genomic DNA isolated from infected cells with the QIAamp® DNA Blood Mini Kit (QIAGEN). A 4.5-kb fragment containing the full gag-pol sequence was generated by PCR with PfuUltra™ High-Fidelity DNA polymerase (Stratagene) and the primers gag-pol-forward: 5’-CTCTCGACGCAGGACTCG-3’ and gag-pol-reverse: 5’-AAACCAGTCCTTAGCTTTCCTTG -3’. The 4.5 kb fragment was cloned into the pCR®4blunt-TOPO plasmid (Invitrogen). The inserts from five individual clones were sequenced to obtain the consensus sequence in the gag and pol region of the adapted virus.

Site-directed mutagenesis

DNA sequence changes specifying the V86M alteration in capsid, which was associated with viral adaptation to TRIM5α_rh, were introduced by site-directed mutagenesis using the QuikChange II XL Site-Directed Mutagenesis protocol (Stratagene) into the pCMVΔP1ΔenvpA plasmid containing the gag, pol and vif regions of NL4-3 HIV-1. DNA sequences specifying the previously described capsid variants H87Q, Q50Y/T54Y, G89A and P90A (Bukovsky et al., 1997; Franke et al., 1994; Owens et al., 2004) were also introduced by site-directed mutagenesis into this plasmid for comparison purposes. The presence of the desired mutations was verified by automated DNA sequencing.

Single-round infection assay

The efficiency of a single round of HIV-1 infection was measured by using recombinant reporter viruses expressing firefly luciferase in place of Nef and pseudotyped with the vesicular stomatitis virus G glycoprotein (VSV-G). Recombinant luciferase-expressing HIV-1 viruses (Helseth et al., 1990) were generated by transfecting 293T cells using the calcium phosphate transfection method (Invitrogen) with 2 μg of the pHCMV-G plasmid expressing the VSV envelope glycoprotein G, 2 μg of a Rev-expressing plasmid, 4 μg of the pCMVΔP1ΔenvpA packaging plasmid expressing the corresponding CA mutant, and 12 μg of an HIV-1 vector plasmid. The HIV-1 vector plasmid expresses an RNA that can be packaged into virions, reverse transcribed, and integrated into a target cell, where it encodes firefly luciferase. Forty-eight hours after transfection, supernatants containing reporter viruses were harvested and cleared by low-speed centrifugation. The amounts of virus in the supernatants were quantified by measuring RT activity.

Target cells were seeded at a density of 6,000 cells/well in 96-well luminometer-compatible tissue culture plates. Twenty-four hours later, medium was changed and different amount of viruses were added to the cells. After 14 h of incubation, the supernatant was removed and fresh medium was added. Forty-eight hours later, the medium was removed and cells were lysed with 30 μl of passive lysis buffer (Promega). Luciferase activity was measured using an EG&G Berthold LB 96V microplate luminometer in accordance with the luciferase assay system technical bulletin (Promega).

Competition for restriction factors with virus-like particles (VLPs)

For the competition assay with VLPs, the PRL cells were infected with 2,500 RT units of wild-type luciferase reporter virus in the presence of 100,000 RT units of virus-like particles lacking a reporter. These VLPs were generated by co-transfecting 293T cells with 4 μg of the pHCMV-G plasmid expressing the VSV envelope glycoprotein G, 4 μg of a Rev-expressing plasmid (pRev), and 12 μg of the pCMVΔP1ΔenvpA packaging plasmid expressing the corresponding CA mutant. VLPs were concentrated by pelleting through a 20% sucrose cushion.

Cyclophilin A incorporation into virions

Eight million 293T cells were transfected with 3 μg of pcDNA-CypA-FLAG, 1 μg of pRev, and 8 μg of pCMVΔP1ΔenvpA harboring different capsid changes. Forty-eight hours after transfection, supernatants containing the viral particles were collected and spun at low speed to remove cell debris. Cleared supernatants were pelleted through a 20% sucrose cushion prepared in PBS for 2 h at 30,000 rpm in an SW41 rotor at 4°C. Pelleted VLPs were resuspended in RT suspension buffer. Similar amount of VLPs, normalized for RT activity, were analyzed by SDS-PAGE and Western blotting with anti-p24 and anti-FLAG antibodies. Western blots were exposed to film for time periods that allowed detection of signals but avoided saturation of the film.

We thank Yvette McLaughlin and Elizabeth Carpelan for manuscript preparation. We acknowledge Angela Carville and the New England Primate Research Center for providing the rhesus monkey blood. We acknowledge Dr. Ronald Desrosiers from the New England Primate Research Center for kindly providing the 221-89 rhesus cell line. We acknowledge ATCC for the 293T, Cf2Th, MK2D and OMK cell lines. This work was supported by grants from the National Institutes of Health (AI063987 and a Center for AIDS Research Award AI060354) and a gift from the late William F. McCarty-Cooper.