Voltar
Artigo Acesso aberto Revisado por pares
BIBTEX RIS

Naturally Occurring Capsid Substitutions Render HIV-1 Cyclophilin A Independent in Human Cells and TRIM-cyclophilin-resistant in Owl Monkey Cells

2005; Elsevier BV; Volume: 280; Issue: 48 Linguagem: Inglês

10.1074/jbc.m506314200

ISSN

1083-351X

Autores

Udayan Chatterji, Michael Bobardt, Robyn L. Stanfield, Roger G. Ptak, Luke A. Pallansch, Priscilla Ward, Maureen Jones, Cheryl A. Stoddart, Piétro Scalfaro, Jean‐Maurice Dumont, Kamel Besseghir, Brigitte Rosenwirth, Philippe Gallay,

Tópico(s)

Viral Infections and Immunology Research

Resumo

In this study, we asked if a naturally occurring HIV-1 variant exists that circumvents CypA dependence in human cells. To address this issue, we sought viruses for CypA independence using Debio-025, a cyclosporine A (CsA) analog that disrupts CypA-capsid interaction. Surprisingly, viral variants from the Main group replicate even in the presence of the drug. Sequencing analyses revealed that these viruses encode capsid substitutions within the CypA-binding site (V86P/H87Q/I91V/M96I). When we introduced these substitutions into viruses that normally rely on CypA for replication, these mutants no longer depended on CypA, suggesting that naturally occurring capsid substitutions obviate the need for CypA. This is the first demonstration that isolates from the Main group naturally develop CypA-independent strategies to replicate in human cells. Surprisingly, we found that these capsid substitutions render HIV-1 capable of infecting Owl monkey (OMK) cells that highly restrict HIV-1. OMK cell resistance to HIV-1 is mediated via TRIM-Cyp, which arose from a retrotransposition of CypA into the TRIM5 α gene. Interestingly, saturation experiments suggest that the Pro86/Gln87/Val91/Ile96 capsid core is “invisible” to TRIM-Cyp. This study demonstrates that specific capsid substitutions can release HIV-1 from both CypA dependence in human cells and TRIM-Cyp restriction in monkey cells. In this study, we asked if a naturally occurring HIV-1 variant exists that circumvents CypA dependence in human cells. To address this issue, we sought viruses for CypA independence using Debio-025, a cyclosporine A (CsA) analog that disrupts CypA-capsid interaction. Surprisingly, viral variants from the Main group replicate even in the presence of the drug. Sequencing analyses revealed that these viruses encode capsid substitutions within the CypA-binding site (V86P/H87Q/I91V/M96I). When we introduced these substitutions into viruses that normally rely on CypA for replication, these mutants no longer depended on CypA, suggesting that naturally occurring capsid substitutions obviate the need for CypA. This is the first demonstration that isolates from the Main group naturally develop CypA-independent strategies to replicate in human cells. Surprisingly, we found that these capsid substitutions render HIV-1 capable of infecting Owl monkey (OMK) cells that highly restrict HIV-1. OMK cell resistance to HIV-1 is mediated via TRIM-Cyp, which arose from a retrotransposition of CypA into the TRIM5 α gene. Interestingly, saturation experiments suggest that the Pro86/Gln87/Val91/Ile96 capsid core is “invisible” to TRIM-Cyp. This study demonstrates that specific capsid substitutions can release HIV-1 from both CypA dependence in human cells and TRIM-Cyp restriction in monkey cells. HIV-1 2The abbreviations used are: HIV-1human immunodeficiency virus, type 1X-gal5-bromo-4-chloro-3-indolyl-β-d-galactopyranosideGFPgreen fluorescent proteinCypAcyclosporine ACsAcyclophilin AOMKOwl monkey cellsVSVGvesicular stomatitis virus G protein. 2The abbreviations used are: HIV-1human immunodeficiency virus, type 1X-gal5-bromo-4-chloro-3-indolyl-β-d-galactopyranosideGFPgreen fluorescent proteinCypAcyclosporine ACsAcyclophilin AOMKOwl monkey cellsVSVGvesicular stomatitis virus G protein. specifically recruits the abundant host protein, cyclophilin A (CypA), to optimally replicate in human cells (1Franke E.K. Yuan H.E. Luban J. Nature. 1994; 372: 359-362Crossref PubMed Scopus (645) Google Scholar, 2Thali M. Bukovsky A. Kondo E. Rosenwirth B. Walsh C.T. Sodroski J. Gottlinger H.G. Nature. 1994; 372: 363-365Crossref PubMed Scopus (559) Google Scholar). It is thought that CypA binds HIV-1 capsid, a major structural component that forms the core that surrounds the viral genome (3Braaten D. Ansari H. Luban J. J. Virol. 1997; 71: 2107-2113Crossref PubMed Google Scholar, 4Dorfman T. Weimann A. Borsetti A. Walsh C.T. Gottlinger H.G. J. Virol. 1997; 71: 7110-7113Crossref PubMed Google Scholar, 5Gamble T.R. Vajdos F.F. Yoo S. Worthylake D.K. Houseweart M. Sundquist W.I. Hill C.P. Cell. 1996; 87: 1285-1294Abstract Full Text Full Text PDF PubMed Scopus (602) Google Scholar). Genetic and structural studies showed that HIV-1 binds CypA via a Gly89-Pro90 peptide within the unique exposed loop of capsid (1Franke E.K. Yuan H.E. Luban J. Nature. 1994; 372: 359-362Crossref PubMed Scopus (645) Google Scholar, 5Gamble T.R. Vajdos F.F. Yoo S. Worthylake D.K. Houseweart M. Sundquist W.I. Hill C.P. Cell. 1996; 87: 1285-1294Abstract Full Text Full Text PDF PubMed Scopus (602) Google Scholar). CypA-capsid interaction can be competitively disrupted by the immunosuppressive drug cyclosporine A (CsA), which binds to the hydrophobic pocket of CypA (2Thali M. Bukovsky A. Kondo E. Rosenwirth B. Walsh C.T. Sodroski J. Gottlinger H.G. Nature. 1994; 372: 363-365Crossref PubMed Scopus (559) Google Scholar, 6Franke E.K. Luban J. Virology. 1996; 222: 279-282Crossref PubMed Scopus (132) Google Scholar, 7Dorfman T. Gottlinger H.G. J. Virol. 1996; 70: 5751-5757Crossref PubMed Google Scholar). Preventing CypA-capsid interaction, either by the introduction of mutations in the CypA binding region of capsid or by the addition of CsA, decreases and delays HIV-1 infectivity in human cells (1Franke E.K. Yuan H.E. Luban J. Nature. 1994; 372: 359-362Crossref PubMed Scopus (645) Google Scholar, 2Thali M. Bukovsky A. Kondo E. Rosenwirth B. Walsh C.T. Sodroski J. Gottlinger H.G. Nature. 1994; 372: 363-365Crossref PubMed Scopus (559) Google Scholar, 3Braaten D. Ansari H. Luban J. J. Virol. 1997; 71: 2107-2113Crossref PubMed Google Scholar, 4Dorfman T. Weimann A. Borsetti A. Walsh C.T. Gottlinger H.G. J. Virol. 1997; 71: 7110-7113Crossref PubMed Google Scholar, 5Gamble T.R. Vajdos F.F. Yoo S. Worthylake D.K. Houseweart M. Sundquist W.I. Hill C.P. Cell. 1996; 87: 1285-1294Abstract Full Text Full Text PDF PubMed Scopus (602) Google Scholar, 6Franke E.K. Luban J. Virology. 1996; 222: 279-282Crossref PubMed Scopus (132) Google Scholar, 7Dorfman T. Gottlinger H.G. J. Virol. 1996; 70: 5751-5757Crossref PubMed Google Scholar). Furthermore, HIV-1 infection is attenuated in human T cells homozygous for a deletion of the gene encoding CypA (8Braaten D. Luban J. EMBO J. 2001; 20: 1300-1309Crossref PubMed Scopus (234) Google Scholar) or in cells treated with small interference RNA that target the CypA gene (9Liu S. Asparuhova M. Brondani V. Ziekau I. Klimkait T. Schumperli D. Nucleic Acids Res. 2004; 32: 3752-3759Crossref PubMed Scopus (52) Google Scholar, 10Sokolskaja E. Sayah D.M. Luban J. J. Virol. 2004; 78: 12800-12808Crossref PubMed Scopus (174) Google Scholar). Recent studies suggest that the CypA in target cells rather than the CypA in virus producer cells is required for optimal HIV-1 infectivity in human cells (10Sokolskaja E. Sayah D.M. Luban J. J. Virol. 2004; 78: 12800-12808Crossref PubMed Scopus (174) Google Scholar, 11Kootstra N.A. Munk C. Tonnu N. Landau N.R. Verma I.M. Proc. Natl. Acad. Sci. U. S. A. 2003; 100: 1298-1303Crossref PubMed Scopus (146) Google Scholar, 12Hatziioannou T. Perez-Caballero D. Cowan S. Bieniasz P.D. J. Virol. 2005; 79: 176-183Crossref PubMed Scopus (162) Google Scholar, 13Ikeda Y. Ylinen L.M. Kahar-Bador M. Towers G.J. J. Virol. 2004; 78: 11816-11822Crossref PubMed Scopus (78) Google Scholar). To date, the block of infectivity has been attributed to post-entry events prior to integration (14Steinkasserer A. Harrison R. Billich A. Hammerschmid F. Werner G. Wolff B. Peichl P. Palfi G. Schnitzel W. Mlynar E. J. Virol. 1995; 69: 814-824Crossref PubMed Google Scholar, 15Braaten D. Franke E.K. Luban J. J. Virol. 1996; 70: 3551-3560Crossref PubMed Google Scholar). Although the CypA-mediated enhancement of HIV-1 infectivity in human cells was discovered more than a decade ago (1Franke E.K. Yuan H.E. Luban J. Nature. 1994; 372: 359-362Crossref PubMed Scopus (645) Google Scholar, 2Thali M. Bukovsky A. Kondo E. Rosenwirth B. Walsh C.T. Sodroski J. Gottlinger H.G. Nature. 1994; 372: 363-365Crossref PubMed Scopus (559) Google Scholar), its role in the HIV-1 life cycle as well as the mechanisms that govern it remain obscure. Because CypA catalyzes the cis-trans isomerization of the Gly89-Pro90 peptide bond of capsid (16Bosco D.A. Eisenmesser E.Z. Pochapsky S. Sundquist W.I. Kern D. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 5247-5252Crossref PubMed Scopus (190) Google Scholar) and the capsid core likely undergoes an ordered uncoating in order to deliver the viral genome into the cytosol of target cells, it has been proposed that CypA orchestrates HIV-1 uncoating in human target cells (17Luban J. Cell. 1996; 87: 1157-1159Abstract Full Text Full Text PDF PubMed Scopus (116) Google Scholar), although no experimental evidence has been reported to support this model. However, observations in monkey cells may help us to understand the role of CypA in HIV-1 infection. Paradoxically, recent findings point to a detrimental role for CypA in HIV-1 infection in monkey cells. Towers et al. (18Towers G.J. Hatziioannou T. Cowan S. Goff S.P. Luban J. Bieniasz P.D. Nat. Med. 2003; 9: 1138-1143Crossref PubMed Scopus (324) Google Scholar) found that CypA-capsid interactions are responsible for the inability of HIV-1 to infect owl monkey (OMK) cells. OMK cells are normally highly refractory to HIV-1 infection; however, CsA renders them permissive to HIV-1 (18Towers G.J. Hatziioannou T. Cowan S. Goff S.P. Luban J. Bieniasz P.D. Nat. Med. 2003; 9: 1138-1143Crossref PubMed Scopus (324) Google Scholar). Moreover, the introduction of a mutation in the CypA binding region of capsid (G89V), which abolishes CypA-capsid contact, renders the mutant virus capable of infecting OMK cells (1Franke E.K. Yuan H.E. Luban J. Nature. 1994; 372: 359-362Crossref PubMed Scopus (645) Google Scholar). Supporting the hypothesis that CypA or CypA-like proteins are responsible for the refractivity of OMK cell to HIV-1, two studies elegantly showed that OMK cell resistance to HIV-1 is mediated via a TRIM-Cyp chimeric protein, which arose from a retrotransposition of CypA into the TRIM5 α gene (19Nisole S. Lynch C. Stoye J.P. Yap M.W. Proc. Natl. Acad. Sci. U. S. A. 2004; 101: 13324-13328Crossref PubMed Scopus (254) Google Scholar, 20Sayah D.M. Sokolskaja E. Berthoux L. Luban J. Nature. 2004; 430: 569-573Crossref PubMed Scopus (547) Google Scholar). Thus, CypA and CypA-like proteins possess the capacities to influence either positively (human cells) or negatively (OMK cells) HIV-1 infectivity of primates. The mechanisms that dictate whether CypA has a beneficial or detrimental effect on HIV-1 infection in primate cells are totally unknown. In the present study, we asked if a naturally occurring HIV-1 variant exists that circumvents both CypA dependence in human cells and TRIM-Cyp sensitivity in OMK cells. The identification of which molecular feature that renders HIV-1 susceptible or resistant to CypA dependence is imperative to shed light on the functions of CypA in HIV-1 infection in primate cells. human immunodeficiency virus, type 1 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside green fluorescent protein cyclosporine A cyclophilin A Owl monkey cells vesicular stomatitis virus G protein. human immunodeficiency virus, type 1 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside green fluorescent protein cyclosporine A cyclophilin A Owl monkey cells vesicular stomatitis virus G protein. Viruses and Cells—All HIV-1 isolates were obtained from the AIDS Research and Reference Reagent Program. The pNL4.3-GFP plasmid was provided by C. Aiken and D. Gabuzda. Capsid mutants were generated as described previously (21von Schwedler U. Song J. Aiken C. Trono D. J. Virol. 1993; 67: 4945-4955Crossref PubMed Google Scholar). To obviate specific cell entry requirements, all pNL4.3-GFP viruses were pseudotyped with the VSVG envelope generously provided by D. Trono. TZM-bl cells (contributed by J. C. Kappes, X. Wu, and Tranzyme, Inc.) were obtained through the AIDS Research and Reference Reagent Program. CCR5+ Jurkat T cells were provided by M. Emerman, whereas OMK cells were provided by G. Towers and C. Aiken. CsA and Debio-025 were provided by Sigma and Debiopharm, respectively. Capsid NL4.3 (R9) and pNL4.3-GFP mutant viruses were generated by PCR mutagenesis and by liposome-mediated transfection of 293T cells using Genejuice (Novagen). Viral supernatants were harvested 48 h post-transfection and filtered through a 0.2-μm pore size filter to remove cellular debris. Viral inoculum was standardized by exoRT assay or p24 enzyme-linked immunosorbent assay (PerkinElmer Life Sciences). Infections—TZM-bl and OMK cells (80,000 or 20,000 cells/well/ml) were seeded for 24 h pre-infection in 24-well plates. Cells were exposed to HIV-1 in the presence of 5 μg/ml polybrene. CsA or Debio-025, at the concentrations indicated, were added 15 min prior to virus addition. Infected cells were analyzed by fluorescence-activated cell sorting (GFP content), by X-gal staining (counting blue foci) or by β-galactosidase activity 48 h post-infection. For β-galactosidase activity, infected TZM-bl cells were washed twice with 1 ml of phosphate-buffered saline and lysed in 100 mm potassium phosphate, pH 7.8, containing 0.2% Triton X-100. Plates were stored at -80 °C for 16 h and thawed on ice, and 20 μl of lysate was transferred to a 96-well plate for detecting β-galactosidase activity. Galacton-Star substrate (Applied Biosystems, Bedford, MA) was diluted 1:50 in the reaction buffer diluent (100 mm sodium phosphate pH 7.5, 1 mm MgCl2, 5% Sapphire-II™ enhancer) to make the reaction buffer. 100 μl of reaction buffer was added to 20 μl of lysate, and the light emission was measured over 1 s in a microplate luminometer after 30 min. For saturation experiments, OMK cells were initially exposed to increasing amounts of wild-type or mutant NL4.3 (R9) viruses that do not encode GFP and subsequently (5 min later) exposed to wild-type or mutant NL4.3 viruses that do encode GFP. Western Blot Analysis—Purification and immunoblot analysis of viruses produced from 293T-transfected cells were conducted as previously described (21von Schwedler U. Song J. Aiken C. Trono D. J. Virol. 1993; 67: 4945-4955Crossref PubMed Google Scholar). CsA and Debio-025 were added 24 h post-infection, and viruses were collected 48 h post-transfection. Rabbit anti-capsid was obtained through the AIDS Research and Reference Reagent Program, whereas rabbit anti-CypA serum was obtained by immunization with recombinant human CypA protein. Pharmacological Prevention of CypA-Capsid Interactions—To identify HIV-1 variants refractory to both human and OMK CypA-mediated activities, we first searched for viruses that do not rely on CypA to optimally infect human cells. To address this issue, we sought conditions that abrogate CypA-capsid interactions in human cells. CsA binds CypA and prevents interaction between CypA and HIV-1 capsid, but it also exhibits an immunosuppressive activity, which complicates the interpretation of its effect on HIV-1 replication. To examine the effect of CypA-capsid interactions on HIV-1 replication exclusively, we sought CsA analogs, which prevent CypA-capsid interactions, but are not immunosuppressive. We used a CsA analog, Debio-025, which lacks the immunosuppressive effect usually associated with CsA but nevertheless retains an inhibitory activity against HIV-1. The non-immunosuppressive CsA analog Debio-025 is a derivative of the non-immunosuppressive CsA analog, SDZ-811, which prevents CypA-capsid interactions and HIV-1 infectivity (22Rosenwirth B. Billich A. Datema R. Donatsch P. Hammerschmid F. Harrison R. Hiestand P. Jaksche H. Mayer P. Peichl P. Antimicrob. Agents Che-mother. 1994; 38: 1763-1772Crossref PubMed Scopus (161) Google Scholar, 23Mlynar E. Bevec D. Billich A. Rosenwirth B. Steinkasserer A. J. Gen. Virol. 1997; 78: 825-835Crossref PubMed Scopus (35) Google Scholar, 24Billich A. Hammerschmid F. Peichl P. Wenger R. Zenke G. Quesniaux V. Rosenwirth B. J. Virol. 1995; 69: 2451-2461Crossref PubMed Google Scholar). We first compared the efficiencies of CsA and Debio-025 to prevent CypA-capsid interactions by examining their potential to inhibit CypA incorporation into virions. 293T cells were transfected with proviral HIV-1 NL4.3 plasmid in the presence of CsA or Debio-025. Viruses were analyzed for their CypA content as described previously (21von Schwedler U. Song J. Aiken C. Trono D. J. Virol. 1993; 67: 4945-4955Crossref PubMed Google Scholar) (Fig. 1). We found that 0.5 and 1 μm of CsA partially blocked CypA incorporation, whereas 8 μm significantly abrogated CypA incorporation. In contrast, Debio-025 at a concentration as low as 0.5 μm prevents CypA packaging. Thus, Debio-025 is particularly well-suited to identify HIV-1 variants that do not depend on CypA-capsid interactions for optimal infection in human cells. Naturally Occurring HIV-1 Isolates Derived from the Main Group Replicate in Human Cells Independently of CypA—To identify HIV-1 isolates that do not rely on CypA to infect human cells, we used TZM-bl HeLa cells that express β-galactosidase after infection with HIV-1. To determine if HIV-1 requires CypA-capsid interactions to infect TZM-bl cells, Debio-025 was added to target cells 15 min prior to addition of virus. By adding Debio-025 to target TZM-bl cells, we anticipate that the drug, by saturating cytosolic CypA, prevents CypA recruitment onto the incoming viral capsid and therefore decreases HIV-1 infectivity. Supporting this hypothesis, Debio-025, even at low concentrations, inhibits HIV-1 infectivity (Fig. 2). It is critical to emphasize that the 2- to 3-fold decrease of infectivity, although modest, is always consistent between experiments using TZM cells. Note that we obtained similar results using CCR5+ Jurkat cells (data not shown); however, given that the assay using TZM-bl cells is quicker (48 h instead of 2 weeks) than the assay using CCR5+ Jurkat cells that do not contain a reporter gene, we preferred using TZM-bl cells as targets to screen a large panel of isolates. Forty-two viruses from different clades and with different co-receptor usage were used to infect TZM-bl cells in the presence of Debio-025. We found that a majority of viruses are highly sensitive to Debio-025 and possess a mean 50% inhibitory concentration (IC50) of 0.097 μm± 0.029 (n = 27) (TABLE ONE). Furthermore, we found that a few viruses are partially resistant to Debio-025 with IC50 values of 1.98 μm± 0.48 (n = 10) (highlighted in green in TABLE ONE). Importantly, we found that two viruses derived from the Main group (M group) are highly resistant to Debio-025 with IC50 values of 8.5 μm± 0.1 (n = 2) (highlighted in pink in TABLE ONE). The definitions of “partially resistant” and “highly resistant” viruses are arbitrary. The “highly resistant viruses” are the two viruses that exhibit an IC50 of >2.5 μm Debio-025; similarly “partially resistant viruses” are the ten viruses that possess IC50 values between 1.5 and 2.5 μm of Debio-025. All the other viruses that exhibit an IC50 of <0.17 μm are termed “sensitive viruses”. Amounts of Debio-025 necessary to decrease the infectivity of the 42 viruses are lower (TABLE ONE) than those used to decrease NL4.3 infectivity in HeLa cells (Fig. 2), because we used less virus (100 pg instead of 1 ng). We obtained similar results for the resistant viruses, using Debio-025-treated CCR5+ Jurkat cells, by monitoring viral replication by p24 enzyme-linked immunosorbent assay over a period of 3 weeks (data not shown). Thus, we identified for the first time HIV-1 isolates from the Main Group that do not rely on CypA to optimally infect human cells. Interestingly, among the 42 isolates tested, only R5 and R5X4 viruses were found to partially or highly resist to Debio-025 (TABLE ONE).TABLE ONEResistance of HIV-1 to Debio-025 Open table in a new tab HIV-1 CypA-independent Infection of Human Cells Correlates with the Presence of Specific Residues within the CypA-binding Site of Capsid—We then asked if this CypA independence correlates with specific residues in the CypA binding region of HIV-1 capsid. To address this issue, we sequenced the CypA binding region of capsid for a majority of the 42 viruses amplified in CCR5+ Jurkat T cells (TABLE ONE). Interestingly, we found that four capsid residues in positions 86, 87, 91, and 96, which are normally well conserved among all HIV-1 subtypes (Fig. 3), were frequently substituted in viruses that are partially or highly resistant to Debio-025 (TABLE ONE). The most commonly occurring capsid sequence contains residues Val86, His87, Ile91, and Met96 (Fig. 3). Interestingly, the two viruses that are highly resistant to Debio-025, RU570 and RU132, have all four residues mutated: Pro86, Gln87, Val91, and Ile96 for RU570 and Gln86, Gln87, Phe91, and Ile96 for RU132 (TABLE ONE). It is also interesting to note that two consensuses contain frequent substitutions at these positions: consensus G that contains Gln86, Gln87, Ile91, and Ile96 and consensus O that contains Pro86, Ala87, Leu91, and Ile96 (Fig. 3). Supporting the possibility that substitutions at these specific positions of capsid mediate CypA independence of HIV-1, we found here that the two highly Debio-025-resistant RU570 and RU132 viruses are derived from consensus G (TABLE ONE). Moreover, previous work showed that two isolates from the Outlier group (group O) infect human cells even in the presence of cyclosporine A (25Braaten D. Franke E.K. Luban J. J. Virol. 1996; 70: 4220-4227Crossref PubMed Google Scholar, 26Wiegers K. Krausslich H.G. Virology. 2002; 294: 289-295Crossref PubMed Scopus (32) Google Scholar). Altogether these data indicate that naturally occurring mutations in capsid can substitute for the need for CypA for optimal HIV-1 infection of human cells. To determine if these residues govern CypA independence, we introduced them in the capsid region of a virus, which is normally highly sensitive to Debio-025 and NL4.3 (TABLE ONE) and which contains the most commonly occurring capsid residues (Val86, His87, Ile91, and Met96) (Fig. 3), and asked whether these mutations confer CypA independence. Using the highly resistant RU570 capsid sequence as a template, we created single, double, triple, and quadruple mutations in the context of the proviral molecular clone NL4.3. Viruses were produced from 293T-transfected cells and used to infect TZM-bl cells as above. Debio-025 significantly reduced the infectivity of wild-type NL4.3 virus (Fig. 4A). As a control, we used a capsid mutant, which cannot bind CypA (NL4.3 G89V) (1Franke E.K. Yuan H.E. Luban J. Nature. 1994; 372: 359-362Crossref PubMed Scopus (645) Google Scholar). This mutant, like the CsA-treated wild-type virus, exhibited a significant defect in infectivity, further reinforcing the notion that CypA is necessary for optimal HIV-1 infectivity in human cells. We found that single and double mutants are still sensitive to Debio-025 (Fig. 4A). In sharp contrast, the infectivity of the triple or quadruple mutant viruses is weakly influenced by Debio-025 (Fig. 4A). This strongly suggests that a combination of three to four capsid substitutions (TABLE ONE) renders HIV-1 CypA independent in human cells. We obtained similar Debio-025 resistance results using physiological peripheral blood mononuclear as target cells (data not shown). One possibility why naturally occurring HIV-1 variants derived from the Main group did not rely on CypA to infect human cells is that, like HIV-2 and simian immunodeficiency virus, they do not bind CypA, explaining their resistance to CsA and CsA-analogs (2Thali M. Bukovsky A. Kondo E. Rosenwirth B. Walsh C.T. Sodroski J. Gottlinger H.G. Nature. 1994; 372: 363-365Crossref PubMed Scopus (559) Google Scholar, 25Braaten D. Franke E.K. Luban J. J. Virol. 1996; 70: 4220-4227Crossref PubMed Google Scholar). However, we found that all Debio-025-resistant HIV-1 viruses, including the variant that contains the four capsid substitutions (Pro86/Gln87/Val91/Ile96), incorporate wild-type levels of CypA (data not shown and Fig. 4B). Moreover, Debio-025 prevented CypA incorporation into these viruses at a degree similar to that in the wild-type virus (Fig. 4B). This indicates that these naturally occurring HIV-1 variants derived from the Main group bind CypA well but do not rely on it to optimally infect human cells. Altogether these data suggest that HIV-1 variants, which naturally pre-exist in the Main group, contain specific residues in the unique exposed loop of capsid that permit infectivity of human cells independently of the CypA recruitment onto the capsid core in the cytosol of target cells. Capsid Substitutions That Permit CypA Independence in Human Cells Render HIV-1 Capable of Infecting Refractory Owl Monkey Cells—Although CypA-capsid interactions positively influence HIV-1 infectivity of human cells, they negatively influence HIV-1 infectivity of OMK cells. Specifically, CsA restores HIV-1 infectivity in OMK cells that normally highly restrict HIV-1 (18Towers G.J. Hatziioannou T. Cowan S. Goff S.P. Luban J. Bieniasz P.D. Nat. Med. 2003; 9: 1138-1143Crossref PubMed Scopus (324) Google Scholar). Moreover, a capsid mutant virus (G89V), which cannot bind CypA, is capable of infecting OMK cells (18Towers G.J. Hatziioannou T. Cowan S. Goff S.P. Luban J. Bieniasz P.D. Nat. Med. 2003; 9: 1138-1143Crossref PubMed Scopus (324) Google Scholar). Recent studies demonstrated that a TRIM-Cyp chimeric protein mediates HIV-1 restriction in OMK cells (19Nisole S. Lynch C. Stoye J.P. Yap M.W. Proc. Natl. Acad. Sci. U. S. A. 2004; 101: 13324-13328Crossref PubMed Scopus (254) Google Scholar, 20Sayah D.M. Sokolskaja E. Berthoux L. Luban J. Nature. 2004; 430: 569-573Crossref PubMed Scopus (547) Google Scholar). It has been proposed that the CypA moiety of TRIM-Cyp is responsible for the recognition of the incoming HIV-1 capsid core (19Nisole S. Lynch C. Stoye J.P. Yap M.W. Proc. Natl. Acad. Sci. U. S. A. 2004; 101: 13324-13328Crossref PubMed Scopus (254) Google Scholar). In this model, preventing TRIM-Cyp-capsid core interaction, either by CsA or by point mutation in the CypA-binding site of capsid, restores HIV-1 infectivity in OMK cells. Thus, after revealing the existence of naturally occurring HIV-1 variants that circumvent CypA dependence in human cells, we asked whether these variants could escape the CypA-mediated infectivity block in OMK cells. To address this issue, OMK cells were infected with wild-type virus (WT NL4.3) that binds CypA, a virus that does not bind CypA (G89V NL4.3), and a virus that binds CypA but does not require CypA to infect human cells (Pro86/Gln87/Val91/Ile96 NL4.3). We found that wild-type virus infects OMK cells poorly, even at a high inoculum (100 ng of p24) (∼1% of infected cells) (Fig. 5), suggesting that OMK cells contain restriction factors that are extremely potent at blocking HIV-1 infectivity. The addition of Debio-025 to OMK cells partially removes the block of infectivity of the wild-type virus (∼20% of infected cells at 100 ng of p24) (Fig. 5). Moreover, the virus, which does not bind CypA (G89V), infected OMK cells well (∼38 and 58% of infected cells at 50 ng and 100 ng of p24, respectively) (Fig. 5). The fact that the ablation of CypA-capsid interactions (G89V or Debio-025) promotes HIV-1 infectivity in OMK cells supports the notion that TRIM-Cyp mediates HIV-1 blockade in OMK cells (19Nisole S. Lynch C. Stoye J.P. Yap M.W. Proc. Natl. Acad. Sci. U. S. A. 2004; 101: 13324-13328Crossref PubMed Scopus (254) Google Scholar, 20Sayah D.M. Sokolskaja E. Berthoux L. Luban J. Nature. 2004; 430: 569-573Crossref PubMed Scopus (547) Google Scholar). Debio-025 had no effect on the infectivity of the virus, which did not bind CypA (G89V) (Fig. 5). Note that the infectivity of the virus that does not bind CypA (G89V) remains superior to that of wild-type virus in the presence of Debio-025 (Fig. 5). This suggests that either the Debio-025 saturation of TRIM-Cyp in OMK cells is not optimal, or that OMK cells contain additional inhibitory factors, which bind capsid via the residue Gly89 but are not sensitive to Debio-025. Importantly, the infectivity of the virus that binds CypA but does not require CypA to infect human cells (Pro86/Gln87/Val91/Ile96 NL4.3), was superior to that of the wild-type virus (i.e. ∼20-fold increase at 50 and 100 ng of p24) (Fig. 5). This indicates that the Pro86/Gln87/Val91/Ile96 substitutions render HIV-1 resistant (at least partly) to the CypA-mediated OMK restriction. Interestingly, the infectivity of Pro86/Gln87/Val91/Ile96 NL4.3 perfectly mirrored the infectivity of the wild-type virus in the presence of Debio-025 (Fig. 5), suggesting that the Pro86/Gln87/Val91/Ile96 NL4.3 escapes the block mediated by Debio-025-sensitive inhibitory factors in OMK cells (i.e. TRIM-Cyp). However, Debio-025 further enhances the capacity of Pro86/Gln87/Val91/Ile96 NL4.3 to infect OMK cells (∼3-fold) (Fig. 5). This suggests that the Pro86/Gln87/Val91/Ile96 capsid core is not totally impervious to Debio-025-sensitive OMK inhibitors (i.e. TRIM-Cyp). Supporting this notion, the infectivity of G89V NL4.3 was superior to that of Pro86/Gln87/Val91/Ile96 NL4.3 in the absence of Debio-025 (Fig. 5). Nonetheless, the rescue in infectivity of wild-type NL4.3 by Debio-025 was superior to that of Pro86/Gln87/Val91/Ile96 NL4.3 (10-fold versus 3-fold). Together these data strongly suggest that the wild-type capsid core is highly vulnerable to inhibitors OMK (i.e. whereas TRIM-Cyp), the Pro86/Gln87/Val91/Ile96 capsid core possesses intrinsic capacities to partly circumvent these factors. Capsid That Permits HIV-1 CypA Independence in Human Cells Does Not Saturate OMK Restriction Factors—When OMK cells are exposed to a first viral challenge (using virus-like particles), they become more susceptible to a second viral challenge (18Towers G.J. Hatziioannou T. Cowan S. Goff S.P. Luban J. Bieniasz P.D. Nat. Med. 2003; 9: 1138-1143Crossref PubMed Scopus (324) Google Scholar). This observation led researchers to postulate that OMK cells contain inhibitory factors that are saturable (18Towers G.J. Hatziioannou T. Cowan S. Goff S.P. Luban J. Bieniasz P.D. Nat. Med. 2003; 9: 1138-1143Crossref PubMed Scopus (324) Google Scholar). One possibility to explain why the Pro86/Gln87/Val91/Ile96 NL4.3 virus is more competent to infect OMK cells than wild-type virus is that its capsid core is less well recognized by the saturable inhibitory factors. To test this hypothesis, OMK cells were pre-challenged with either wild-type or Pro86/Gln87/Val91/Ile96 viruses and subsequently exposed to wild-type HIV-1. As above, in the absence of pre-challenge, wild-type HIV-1 barely infects OMK cells. In contrast, pre-challenging OMK cells with increasing amounts of wild-type virus (100, 500, and 1000 ng of p24) allows wild-type HIV-1 capable of infecting these cells (Fig. 6, right panel). This is in accordance with the hypothesis that an initial viral challenge can saturate inhibitory factors present in OMK cells (18Towers G.J. Hatziioannou T. Cowan S. Goff S.P. Luban J. Bieniasz P.D. Nat. Med. 2003; 9: 1138-1143Crossref PubMed Scopus (324) Google Scholar, 19Nisole S. Lynch C. Stoye J.P. Yap M.W. Proc. Natl. Acad. Sci. U. S. A. 2004; 101: 13324-13328Crossref PubMed Scopus (254) Google Scholar, 20Sayah D.M. Sokolskaja E. Berthoux L. Luban J. Nature. 2004; 430: 569-573Crossref PubMed Scopus (547) Google Scholar). Surprisingly, pre-challenging OMK cells with increasing amounts of the Pro86/Gln87/Val91/Ile96 virus renders target cells minimally more permissive to a subsequent wild-type virus exposure (Fig. 6, left panel). Similarly, we found that a pre-challenge with G89V NL4.3 virus does not affect the refractivity of OMK cells as previously reported (18Towers G.J. Hatziioannou T. Cowan S. Goff S.P. Luban J. Bieniasz P.D. Nat. Med. 2003; 9: 1138-1143Crossref PubMed Scopus (324) Google Scholar) (data not shown). Our observation that the Pro86/Gln87/Val91/Ile96 viral pre-challenge does not alleviate the block in OMK cells suggests that the Pro86/Gln87/Val91/Ile96 capsid core cannot saturate the OMK inhibitory factors. This strongly suggests that the Pro86/Gln87/Val91/Ile96 capsid core is less well recognized by the OMK inhibitory factors than wild-type capsid core. This capsid core “invisibility” may explain the enhanced capacity of the Pro86/Gln87/Val91/Ile96 virus to infect OMK cells. Together our data reveal that naturally occurring HIV-1 variants exist that circumvent both CypA dependence in human cells and TRIM-Cyp sensitivity in OMK cells. In this study, we identified naturally occurring HIV-1 isolates from the Main group that do not depend on CypA for optimal infection of human cells. Sequencing analyses revealed that these viruses contain specific residues at specific positions in the CypA-binding site of capsid. To determine whether these residues mediate CypA independence, we introduced them into viruses that normally highly depend on CypA. Importantly, these viruses no longer depend on CypA for infection after the introduction of these capsid substitutions. These findings demonstrate that naturally pre-existing residues located at specific positions within capsid can render HIV-1 capable of optimally infecting human cells even in the absence of CypA. An analysis of 2599 HIV-1 capsid sequences available from the Los Alamos Data base indicates that 3.35% of the capsid sequences contain the V86P substitution, 19.74% contain H87Q, 27.36% contain I91V, and 15.31% contain M96I (Fig. 7). We also found that 46.44% (1207/2599) have no change from the His87/Ile91/Met96 consensus, whereas 7% (182/2599) of the capsid sequences have a change in all three locations. Moreover, 45.48% (1182/2599) have no change from the Val86/His87/Ile91/Met96 consensus, whereas 3.89% (101/2599) of the capsid sequences have a change in all four locations. The two consensuses G and O contain frequent substitutions at these positions (Fig. 3). Supporting the possibility that substitutions at these specific positions of capsid mediate CypA independence, the two highly Debio-025-resistant RU570 and RU132 viruses are derived from consensus G (TABLE ONE) and two isolates from the consensus O, viruses 5180 and 9435, that infect human cells even in the presence of CsA (25Braaten D. Franke E.K. Luban J. J. Virol. 1996; 70: 4220-4227Crossref PubMed Google Scholar, 26Wiegers K. Krausslich H.G. Virology. 2002; 294: 289-295Crossref PubMed Scopus (32) Google Scholar) contain Pro86/Ala87/Leu91/Ile96 and Gln86/Ala87/Leu91/Ile96 capsid substitutions, respectively (26Wiegers K. Krausslich H.G. Virology. 2002; 294: 289-295Crossref PubMed Scopus (32) Google Scholar). During the course of this study, Ikeda et al. (13Ikeda Y. Ylinen L.M. Kahar-Bador M. Towers G.J. J. Virol. 2004; 78: 11816-11822Crossref PubMed Scopus (78) Google Scholar) also found that substitutions within capsid correlate with CsA resistance in human cells. Specifically, Ikeda et al. (13Ikeda Y. Ylinen L.M. Kahar-Bador M. Towers G.J. J. Virol. 2004; 78: 11816-11822Crossref PubMed Scopus (78) Google Scholar) transferred the full-length gag gene of eleven HIV-1 isolates into a CsA-sensitive NL4.3 backbone and tested the resulting chimeric viruses for their resistance to CsA in human cells. They found that the Gag region of four viruses (89.6, Gun WT, Gun V, and 93BR) also confers CsA resistance in human cells (13Ikeda Y. Ylinen L.M. Kahar-Bador M. Towers G.J. J. Virol. 2004; 78: 11816-11822Crossref PubMed Scopus (78) Google Scholar). They also identified a triple Pro86/Gln87/Val91 capsid substitution that renders HIV-1 partially resistant to CsA. We confirmed these data by showing that the Pro86/Gln87/Val91 resistance to Debio-025 is very similar to those of the two variants that we identified in this study, the triple Gln87/Val91/Ile96 and quadruple Pro86/Gln87/Val91/Ile96 HIV-1 variants (Fig. 4A). Interestingly, Kootstra et al. identified another CsA-resistant HIV-1 variant (Pro86/Gln87/Pro88/Val91) (11Kootstra N.A. Munk C. Tonnu N. Landau N.R. Verma I.M. Proc. Natl. Acad. Sci. U. S. A. 2003; 100: 1298-1303Crossref PubMed Scopus (146) Google Scholar), which distinct is from that of Ikeda et al. (Pro86/Gln87/Ala88/Val91) (13Ikeda Y. Ylinen L.M. Kahar-Bador M. Towers G.J. J. Virol. 2004; 78: 11816-11822Crossref PubMed Scopus (78) Google Scholar) and of ours (Val86/Gln87/Ala88/Val91/Ile96 and Pro86/Gln87/Ala88/Val91/Ile96). Given that we showed that the Gln87/Val91, Gln87/Ile96 and Val91/Ile96 double mutations do not suffice to render HIV-1 resistant to CsA or Debio-025, altogether these data suggest that a combination of three to four capsid substitutions allows HIV-1 capable of escaping the drug pressure and of replicating in a CypA-independent manner. Ikeda et al. (13Ikeda Y. Ylinen L.M. Kahar-Bador M. Towers G.J. J. Virol. 2004; 78: 11816-11822Crossref PubMed Scopus (78) Google Scholar) also presented evidence that the H87Q substitution alone renders NL4.3 more resistant to CsA. However, we found in this study that H87Q alone or in combination with I91V or M96I (double substitutions) does not suffice to render HIV-1 impervious to Debio-025 (Fig. 4A). Indeed, the H87Q substitution renders HIV-1 Debio-025-resistant and CypA-independent in human cells only in the context of the triple (H87Q/I91V/M96I) and the quadruple (V86P/H87Q/I91V/M96I) substitutions. Supporting the notion that the His87 substitution alone does not dictate CsA resistance, the His87 residue was also found substituted in two CsA-sensitive group O viruses, 13470 and 8913 (H87P and H87A for 13470 and 8913, respectively) (26Wiegers K. Krausslich H.G. Virology. 2002; 294: 289-295Crossref PubMed Scopus (32) Google Scholar). The apparent discrepancy between Ikeda's study and ours may simply result from the use of different target cells. Specifically, Ikeda et al. (13Ikeda Y. Ylinen L.M. Kahar-Bador M. Towers G.J. J. Virol. 2004; 78: 11816-11822Crossref PubMed Scopus (78) Google Scholar) used TE671 cells, whereas we used HeLa cells. Previous studies showed that the intrinsic cellular concentration of CypA modulates the resistance of HIV-1 to CsA (27Yin L. Braaten D. Luban J. J. Virol. 1998; 72: 6430-6436Crossref PubMed Google Scholar, 28Ackerson B. Rey O. Canon J. Krogstad P. J. Virol. 1998; 72: 303-308Crossref PubMed Google Scholar). Thus, we cannot exclude the possibility that HeLa cells contain more or less CypA than TE671 cells, explaining why the H87Q virus is sensitive to the drug in HeLa cells, but not in TE671 cells. Moreover, we used Debio-025, whereas Ikeda et al. used CsA. Our observation that Debio-025 is more efficient at disrupting CypA-capsid interaction (Fig. 1) may also explain why the H87Q virus is sensitive to Debio-25 but not to CsA. It is important to emphasize that Ikeda et al. and Kootstra et al. used HIV-1-based vectors, which contain a capsid region amplified by PCR, lack Nef (due to the inserted GFP gene), lack gp160 and are VSVG-pseudotyped. In contrast, in this study, we used “real” viruses: primary isolates derived from patients and amplified in physiological peripheral blood mononuclear cells. One of our goals of performing an extensive phenotypic analysis of HIV-1 variants for CypA dependence using “real” primary isolates rather than HIV-1-based vectors was to provide background information for the in vivo use of Debio-025 as a therapeutic agent. Previous work showed that single point mutations (Glu92 or Asp94) emerge in the capsid of NL4.3 under CsA selection, but in contrast to our naturally occurring CypA-independent viruses, the artificial A92E or G94D viruses require CsA to replicate in HeLa cells (29Aberham C. Weber S. Phares W. J. Virol. 1996; 70: 3536-3544Crossref PubMed Google Scholar, 30Braaten D. Aberham C. Franke E.K. Yin L. Phares W. Luban J. J. Virol. 1996; 70: 5170-5176Crossref PubMed Google Scholar). The A92E or G94D capsid substitutions were not found among our panel of HIV-1 isolates, suggesting that these artificially induced capsid mutations do not pre-exist in the HIV-1 population. Because CypA-independent viruses are poorly represented among the HIV-1 population and that a majority of simian immunodeficiency virus isolates do not require CypA (at least to infect human cells), this suggests that CypA recruitment represents a major advantage for HIV-1 infection in human cells. Nevertheless, our present findings suggest that HIV-1 may exploit CypA-dependent and -independent strategies to guarantee optimal infection of human cells. After revealing the existence of naturally occurring HIV-1 variants that circumvent CypA dependence in human cells, we asked whether these variants could also escape the TRIM-Cyp-mediated infectivity block in OMK cells. In contrast to wild-type HIV-1, which fails to infect OMK cells even at a high inoculum, we found that the CypA-independent Pro86/Gln87/Val91/Ile96 virus infects OMK cells significantly. During the course of this study, two studies showed that viruses containing the H87Q substitution also exhibit a higher capacity to infect OMK cells than that of wild-type virus (13Ikeda Y. Ylinen L.M. Kahar-Bador M. Towers G.J. J. Virol. 2004; 78: 11816-11822Crossref PubMed Scopus (78) Google Scholar, 31Owens C.M. Song B. Perron M.J. Yang P.C. Stremlau M. Sodroski J. J. Virol. 2004; 78: 5423-5437Crossref PubMed Scopus (112) Google Scholar). This further suggests that the CypA-binding site of capsid represents a major locus for TRIM-Cyp action. In contrast to wild-type virus, we found that the CypA-independent virus cannot saturate the OMK restriction. Given that it is thought that primate restrictions target the incoming virus at a post-entry step and that the locus for restriction resides in the viral capsid, our observations suggest that TRIM-Cyp fails to recognize the Pro86/Gln87/Val91/Ile96 incoming capsid core. This is intriguing given that native CypA likely binds Pro86/Gln87/Val91/Ile96 capsid. Indeed, we showed that CypA is packaged at wild-type level in the Pro86/Gln87/Val91/Ile96 virus (Fig. 4B). However, we cannot exclude the possibility that TRIM-Cyp binds differently to HIV-1 Gag than to the incoming capsid core. We tried to determine if TRIM-Cyp binds HIV-1 Gag by co-transfecting 293T cells with NL4.3 and OMK TRIM-Cyp and by looking for TRIM-Cyp incorporation into nascent viruses. Although TRIM-Cyp was abundantly expressed in transfected cells, we failed to detect TRIM-Cyp in pelleted particles by Western blot analysis (data not shown). This suggests that either TRIM-Cyp cannot bind NL4.3 Gag or that its subcellular localization prevents its packaging. Nevertheless, if CypA binds Pro86/Gln87/Val91/Ile96 capsid, we would expect TRIM-Cyp capable of recognizing the incoming Pro86/Gln87/Val91/Ile96 capsid core in the cytosol of OMK cells. Our observations, that Pro86/Gln87/Val91/Ile96 infectivity in OMK cells is lower than that of the G89V virus that cannot bind CypA and that Debio-025 boosts Pro86/Gln87/Val91/Ile96 infectivity, suggest that TRIM-Cyp recognizes, at least partially, incoming the Pro86/Gln87/Val91/Ile96 capsid core. Although further work is required to determine how these capsid substitutions render HIV-1 Debio-025-resistant and CypA-independent in human cells as well as competent for OMK infection, it is likely that these residues subtly modify the capsid loop structure in a way that circumvents CypA-mediated activities on the capsid core. Highlighting the importance of our findings, Debio-025 is currently being tested in a clinical trial for its anti-HIV-1 efficacy. By identifying naturally occurring primary isolates (and not VSVG-pseudotyped HIV-1-based vector-encoding mutated capsid) that resist to Debio-025, our study suggests that the use of Debio-025 or CsA analogs as therapeutic agents may fail in patients infected with viral variants that contain the capsid substitutions identified in this study (i.e. subtype G and O variants). Thus, our present data provide a molecular context for the interpretation of the results of this trial. We thank J. Kuhns for secretarial assistance, S. Turk for support of virus genotyping and scientific input, and B. Beer and G. Jandu for initial input on capsid sequences and alignments.

Referência(s)