|Home | About | Journals | Submit | Contact Us | Français|
Vpu proteins of pandemic HIV-1 M strains degrade the viral receptor CD4 and antagonize human tetherin to promote viral release and replication. We find that Vpus from SIVgsn, SIVmus and SIVmon infecting Cercopithecus primate species also degrade CD4 and antagonize tetherin. In contrast, SIVcpz, the immediate precursor of HIV-1, whose Vpu shares a common ancestry with SIVgsn/mus/mon Vpu, uses Nef rather than Vpu to counteract chimpanzee tetherin. Human tetherin, however, is resistant to Nef and thus poses a significant barrier to zoonotic transmission of SIVcpz to humans. Remarkably, Vpu from non-pandemic HIV-1 O strains are poor tetherin antagonists while those from the rare group N viruses do not degrade CD4. Thus, only HIV-1 M evolved a fully functional Vpu following the three independent cross-species transmissions that resulted in HIV-1 groups M, N, and O. This may explain why group M viruses are almost entirely responsible for the gobal HIV/AIDS pandemic.
Primate lentiviruses such as HIV and SIV encode several accessory proteins. These include Vif, Vpr, Vpu, Vpx, and Nef, and are usually dispensable for viral growth in vitro. However, they are important for viral replication in vivo because they counteract host restriction factors, enhance viral replication and virion infectivity, or facilitate viral evasion of the adaptive immune response (reviewed in Malim and Emerman, 2008). Vif, vpr and nef are found in the genomes of all primate lentiviruses. In contrast, vpu genes were initially only found in HIV-1 and its precursor SIVcpz from chimpanzees (Pan troglodytes), but not in HIV-2 or other SIV strains (Cohen et al., 1988; Huet et al., 1990; Gao et al., 1999; Santiago et al., 2003). Subsequently, molecular characterization of SIVs from additional primate species revealed that the genomes of SIVgsn, SIVmon, SIVden and SIVmus from greater spot-nosed (Cercopithecus nictitans), mona (C. mona), Dent's mona (C. denti) and Mustached monkeys (C. cephus), as well as SIVgor from gorillas (Gorilla gorilla gorilla) also carry vpu genes (Barlow et al., 2003; Courgnaud et al., 2002; Courgnaud et al., 2003; Dazza et al., 2005; Takehisha et al., 2009). SIVcpz, which gave rise to pandemic (M, main) and non-pandemic (O, outlier and N, non-M, non-O) groups of HIV-1 and also to SIVgor (which is closely related to HIV-1 O; Van Heuverswyn et al., 2006), is the product of successive cross-species transmission and recombination events involving precursors of today’s SIVgsn/mus/mon/den and SIVrcm from red-capped mangabeys (Cercocebus torquatus) (Bailes et al., 2002). Thus, all vpu genes likely originated from a common ancestor of the SIVgsn/mus/mon/den lineage of primate lentiviruses (Kirchhoff, 2009).
Vpu is an ~80 amino acid integral class I membrane phosphoprotein (Cohen et al., 1988; Maldarelli et al., 1993). Studies performed with HIV-1 NL4-3 Vpu have established two main functions. First, Vpu induces the degradation of the primary viral receptor CD4 by a multi-step process that involves direct binding of Vpu to the cytoplasmic tail of CD4 in the endoplasmic reticulum (Bour et al., 1995, Margottin et al., 1998; Willey et al., 1992). Second, Vpu promotes the release of progeny virions from HIV-1-infected human cells (Strebel et al., 1988; Klimkait et al., 1990; Gottlinger et al., 1993) by antagonizing a recently identified restriction factor, termed tetherin (also known as CD317, BST2 or HM1.24), that is induced by interferon-alpha and results in the ‘tethering’ of nascent virions to the cellular plasma membrane (Neil et al., 2008; Van Damme et al., 2008). Tetherin has a broad antiviral activity and inhibits the release of various enveloped viruses (Jouvenet et al., 2009; Kaletsky et al., 2009; Sakuma et al., 2009). Furthermore, monkey and rodent tetherins block virion release but are not counteracted by HIV-1 Vpu (Goffinet et al., 2009; Gupta et al., 2009; McNatt et al., 2009; Wong et al., 2009).
To date, functional data have almost exclusively been derived from the Vpu protein of the T-cell line adapted HIV-1 NL4-3 molecular clone. It is thus unknown to what extent CD4 degradation and tetherin antagonism are conserved among the diverse Vpu proteins found in HIV-1 and SIVs and how Vpu function evolved following zoonotic transmissions of primate lentiviruses from monkeys to chimpanzees and, ultimately, to humans. To address these questions, we analyzed a panel of vpu constructs representing nearly the entire spectrum of primate lentiviruses known to encode this accessory gene, i.e. HIV-1 M, O and N, SIVcpz, SIVgor, SIVgsn, SIVmon and SIVmus. We show that all Vpu proteins, except those found in HIV-1 group N, degrade human CD4, whereas tetherin is antagonized in a more species-specific manner. Unexpectedly, we found that SIVcpz employs Nef rather than Vpu as a tetherin antagonist. Following transmission of SIVcpz to humans, HIV-1 group M Vpu and not Nef acquired an anti-tetherin activity in addition to its abilty to degrade CD4, most likely because the human tetherin has a deletion in its cytoplasmic domain that disrupts its susceptibility to Nef (Zhang et al., 2009). In contrast, Vpus of non-pandemic HIV-1 group O and N strains that resulted from independent zoonotic transmissions (Hahn et al., 2000; Van Heuverswyn and Peeters, 2007) either do not antagonize human tetherin (group O strains) or are unable to degrade CD4 (group N strains). Thus, the evolution of a fully functional Vpu protein may have been an important prerequisite for the efficient spread of HIV-1 group M in the human population.
To examine to what extent CD4 degradation and anti-tetherin activity are conserved among primate lentiviral Vpus, we analyzed a large panel of HIV-1 and SIV vpu alleles (Table S1). Our collection included Vpu constructs from 26 strains of HIV-1 group M, 11 strains of HIV-1 O, three strains of HIV-1 N, six strains of SIVcpzPtt from central chimpanzees (CPZ), four strains of SIVcpzPts from eastern chimpanzees, one strain of SIVgor from a Western lowland gorilla (GOR), two strains of SIVgsn from greater spot-nosed monkeys (GSN), one strain of SIVmon from a mona monkey (MON), and three strains of SIVmus from mustached monkeys (MUS). The HIV-1 group M Vpus included all nine subtypes, as well as ten vpu alleles (8 clade B and 2 clade C) from full-length infectious molecular clones of transmitted/founder viruses obtained by single genome amplification of plasma virion RNA (Salazar-Gonzales et al., 2009). Notably, all but two SIV vpu genes were amplified directly from the blood or fecal material of naturally infected monkeys or apes (Table S1).
An amino acid alignment of representative primate lentiviral Vpus shows that they are highly variable. Nonetheless, some features, including a hydrophobic N-terminal membrane anchor domain, a putative central α-helical region with hydrophobic, basic and acidic residues and an acidic C-proximal region, are usually conserved (Figure 1). Two serine residues (S52 and S56), known to be phosphorylated by casein kinase II and required for CD4 degradation by NL4-3 Vpu (Schubert and Strebel, 1994; Paul and Jabbar, 1997), were present in all HIV-1 and SIVcpzPtt Vpus. In the remaining Vpus, the number and location of serine residues varied, although at least one potential serine phosphorylation site was always preserved (Figure 1). Thus, despite their exceedingly high variability, primate lentiviral Vpus exhibit common sequence signatures.
To determine which primate lentiviral Vpus prevent cell surface expression of CD4, we cotransfected 293T cells with vectors coexpressing Vpu and GFP (or GFP alone for control) together with a human (HU) CD4 expression construct. In the absence of Vpu, cells expressed high levels of CD4 and GFP (Figure 2A). However, upon coexpression of most Vpus, CD4 surface expression was markedly decreased (examples are shown in Figure 2A). This effect was Vpu specific since it was abrogated by a single S52A mutation in the cytosolic domain of the HIV-1 NL4-3 Vpu, previously shown to be critical for CD4 degradation (Schubert and Strebel, 1994; Paul and Jabbar, 1997). Quantitative analysis of CD4 surface levels (Fig. 2B) showed that most SIVgsn, SIVmon, SIVmus, SIVcpz and SIVgor, as well as HIV-1 M and O Vpus reduced CD4 cell surface expression by more than 80% (Figure 2B). In striking contrast, each of three HIV-1 group N Vpu proteins lacked this function (Figure 2B). These functional differences were statistically highly significant (p<0.0001). Analysis of Vpus containing a C-terminal AU-1-tag confirmed these results and showed that all but one construct expressed proteins of the expected size (Figure S1 and data not shown). These results show that the ability of blocking CD4 cell surface expression is highly conserved among primate lentiviral Vpus - with the notable exception of HIV-1 N Vpu, which lacks this function (Figure 2B).
It has recently been shown that the HIV-1 NL4-3 Vpu antagonizes tetherin in a species-specific manner, i.e. it exhibited activity against human but not monkey (i.e. rhesus macaque, RM and African green monkey, AGM) or rodent tetherins (Goffinet et al., 2009; Gupta et al., 2009; McNatt et al., 2009). To determine the anti-tetherin activity of other primate lentiviral Vpus we generated vectors that expressed GSN, MON, MUS, CPZ, GOR, HU, AGM and RM tetherin molecules (Table S2). As shown in Figure 3A, the sequence of the transmembrane (TM) domain of tetherin, which has been evolving under positive selection and determines its susceptibility to HIV-1 NL4-3 Vpu (McNatt et al., 2009; Gupta et al., 2009), varies substantially between higher primates and monkeys, but is conserved between humans, gorillas and chimpanzees (except for a V35M variation between HU and CPZ tetherins; Figure 3A). In comparison, the TM domains of the six monkey tetherin sequences differed at a total of four amino acid positions. All tetherin proteins were expressed and suppressed HIV-1 release in the absence of Vpu (Figure S2).
To test whether the various HIV-1 and SIV Vpus are capable of counteracting tetherins from their respective hosts, as well as those from other species, we measured infectious virus yields from 293T cells following cotransfection of a vpu-deleted (ΔVpu) HIV-1 proviral construct (Rucker et al., 2004) with Vpu and tetherin expression plasmids. Our data showed that only HIV-1 Vpus were effective antagonists of HU, CPZ and GOR tetherins (Figure 3B). In comparison, SIVgsn, SIVmon and SIVmus Vpus counteracted the tetherins of their respective host, as well as other monkey species (Figure 3B). The exception was SIVmon CML1 Vpu, which was inactive against the MUS tetherin, which differs from the other monkey tetherins by a L36F substitution and a single amino acid deletion (ΔL42) in the TM domain (Figure 3A, B). Thus, HIV-1 group M Vpus potently antagonized tetherins of apes and humans but were less effective against the monkey restriction factors, whereas SIVgsn, SIVmon and SIVmus Vpus showed the opposite phenotype. Interestingly, the HIV-1 JR-CSF and YU-2 Vpus exhibited some activity against GSN and AGM tetherins (Figure 3B), indicating that certain HIV-1 Vpus have a broader anti-tetherin activity than the NL4-3 Vpu (Goffinet et al., 2009; Gupta et al., 2009; Jia et al., 2009; McNatt et al., 2009). In agreement with previous studies (McNatt et al., 2009), an exchange of the TM domain in HU tetherin with that of AGM or RM rendered it resistant to HIV-1 Vpus but conferred sensitivity to SIVmon and SIVmus Vpus (Figure S3).
While the functional profiles of HIV-1, SIVgsn, SIVmon and SIVmus Vpus were as expected, the observed anti tetherin activities of ape virus Vpus came as a surprise: all SIVcpz Vpus, including those derived from the closest relatives of the HIV-1 group M (i.e. LB7, MB897, MB66) (Keele et al., 2006), failed to antagonize all eight different tetherins including that derived from their cognate CPZ host (Figure 3B). Similarly, the Vpu protein of SIVgor had only poor anti-tetherin activity. These results were confirmed by expressing the various Vpu proteins in cis in the context of infectious HIV-1 molecular clones, using both tetherin transfected 239T cells as well as HeLa cells expressing endogenous tetherin (see Supplemental Results and Figure S4, Figure S5 and Figure S6). To control for intracellular viral gene expression, we also examined the anti-tetherin activity of primate lentiviral Vpus by Western blot analysis. The results confirmed that HIV-1 Vpus efficiently counteracted HU and CPZ tetherins, but were less effective against those of AGM or RM, whereas Vpu proteins from SIVgsn and SIVmus showed the opposite phenotype. They also confirmed that SIVcpz Vpus lack significant anti-tetherin activity (Figure S7). Importantly, the Western blots also demonstrated that the various Vpus acted specifically at the level of virion release into the culture supernatants, and did not influence cell-associated levels of p24 antigen expression.
Our finding that SIVcpz Vpus are very poor tetherin antagonists suggested that HIV-1 group M Vpus gained anti-tetherin activity during adaptation to the new human host. Thus, we next examined whether Vpus from non-pandemic HIV-1 groups O and N strains (which represent independent cross-species transmissions; Hahn et al., 2000; Van Heuverswyn and Peeters, 2007) also evolved the capability of antagonizing HU tetherin. Strikingly, we found that (similarly to SIVcpz and SIVgor Vpus) HIV-1 group O Vpus generally displayed little if any activity against the human tetherin (Figure 3C). In contrast, one HIV-1 group N Vpu (YBF30) efficiently counteracted HU tetherin, while two others had more modest effects (Figure 3C). These results were confirmed at various levels of HU tetherin expression (Figure S8). Thus, only HIV-1 group M and N, but not HIV-1 group O Vpus, have acquired significant anti-tetherin activity during viral adaptation to the human host.
Recently, it has been shown that the Nef proteins of some SIVs that lack a vpu gene (i.e. those of SIVsmm/mac, SIVagm and SIVblu) counteract tetherin (Jia et al., 2009; Zhang et al., 2009). We therefore examined whether SIVcpz, SIVgor and HIV-1 group O employ Nef rather than Vpu to antagonize this restriction factor. Since primate lentiviral Nef proteins enhance virion infectivity (Münch et al., 2007), it was expected that expression of Nef would enhance the infectious virus yield but not p24 antigen release in the absence of tetherin (examples shown in Figure S9). To account for this confounding variable, we normalized all infectious virus yields to that obtained in the presence of Nef but the absence of tetherin (100%). After this adjustment the relative levels of virus release measured by infectivity assays and by p24 antigen ELISA correlated extremely well (R2 =0.93; Figure S9). We also measured the levels of cell-associated p24 expression by ELISA and Western blot and found that neither Nef nor tetherin affected these values.
First, we examined whether HIV-1, SIVcpz and SIVgor Nefs (Table S3) antagonize tetherin from apes and humans by cotransfecting 293T cells with an HIV-1 proviral construct lacking vpu and nef, tetherin expression vectors and constructs expressing various Nef proteins. The results demonstrated that SIVcpz and SIVgor Nefs antagonize the tetherin molecules from their cognate hosts but not the tetherin variant found in humans (Figure 4A). Insertion of five amino acids (DDIWK) that are missing in the cytoplasmic domain of HU tetherin (Figure 3A) and are known to confer sensitivity to SIVmac and SIVagm Nef (Jia et al., 2009; Zhang et al., 2009) fully restored its susceptibility to SIVcpz and SIVgor Nef antagonism (HU-INS, Figure 4A). The disruptive effect of the N-terminal deletion in HU tetherin was highly significant (p<0.0001). Some HIV-1 Nefs displayed modest activities against the CPZ, GOR and HU-INS tetherins, but did not antagonize HU tetherin. The SIVgor Nef was more effective against the GOR tetherin than any of six SIVcpz Nefs. Notably, the GOR and CPZ tetherin proteins differ at only two amino acid positions in the cytoplasmic domain (D15A, W17L) at the same location that governs sensitivity to other Nef proteins (Figure 4A). This strongly suggests that SIVgor Nef has specifically adapted to antagonize tetherin in gorillas. To confirm these findings we generated vpu-defective mutants of replication-competent HIV-1 NL4-3 proviral constructs that coexpress GFP and different HIV-1 and SIVcpz nef alleles from bicistronic RNAs (Schindler et al., 2006) and analyzed viral protein expression and release using Western blot analyses. Our data confirmed that all SIVcpz Nefs antagonize CPZ (but not HU) tetherin (Figure 4B). Notably, Nef proteins from other SIVs that encode Vpu proteins, i.e. SIVgsn, SIVmon and SIVmus, did usually not counteract tetherins from their respective hosts (Figure 4C). In agreement with published data (Jia et al., 2009; Zhang et al., 2009), we found that the nef alleles from SIVmac and SIVagm antagonize RM and AGM tetherins, albeit less efficiently than SIVcpz Nefs (Figure 4C, and data not shown).
To detect possible weak anti-tetherin activities, we next compared the effects of various primate lentiviral vpu and nef alleles on virus release at different doses of plasmids expressing the tetherins from their respective hosts. These titration experiments confirmed that SIVgsn, SIVmus and SIVmon use mainly Vpu to antagonize tetherin (Figure 5A). The exception was SIVgsn116 Nef which also displayed significant anti-tetherin activity, suggesting that this primate lentivirus may use both accessory proteins to counteract this restriction factor. The titration analyses also confirmed that SIVcpz and SIVgor use Nef and HIV-1 groups M and N use Vpu to antagonize tetherin (Figure 5A, B). We also assessed whether HIV-1 O Nefs counteract HU tetherin at lower concentrations, since a weak enhancement of infectious virus release in the presence of HU tetherin was observed in some experiments (Figure 5B). These titration analyses showed that HIV-1 group O Nefs are just as inactive against HU tetherin as those of SIVcpz (Figure 5C). Moreover, HIV-1 group O Nefs were less potent antagonists of CPZ tetherin than SIVcpz Nefs (Figure 4A, Figure 5C). Thus, HIV-1 group O strains appear to have lost some of their Nef-associated anti-CPZ-tetherin activity following transmission to humans. Taken together, our results demonstrate that SIVcpz and SIVgor (but not HIV-1 groups M, N and O) use Nef rather than Vpu to antagonize tetherin.
In this study, we examined the evolution of Vpu function in the primate lentiviruses whose ancestors gave rise to HIV-1. We show that the two major functions of the HIV-1 NL4-3 Vpu protein, i.e. CD4 degradation and enhancement of virion release by counteracting tetherin, are conserved in other vpu-containing non-human primate lentiviruses (summarized in Figure 6A). Nonetheless, important differences were noted. First, Vpu appears to require some host-specific adaptation in order to be an effective tetherin antagonist since Vpus from SIVgsn, SIVmus and SIVmon were only active against monkey tetherins and those from HIV-1 group M were only effective against ape and human tetherins (Figure 3B). In comparison, host-specific adaptation seems less important for CD4 downmodulation since the great majority of primate lentiviral Vpu and Nef proteins is active against human CD4 (Figure 2; Schindler et al., 2006). Second, we found that SIVcpz employs Nef rather than Vpu to antagonize tetherin (Figure 3C and and4C).4C). This finding was unexpected given that the vpu genes of SIVcpz, SIVgsn, SIVmus and SIVmon share a common ancestor (Schindler et al., 2006) and because previous studies suggested that only SIVs that do not contain a vpu gene use Nef to antagonize tetherin (Jia et al., 2009; Zhang et al., 2009). Third, we found that SIV Nefs generally fail to counteract HU tetherin because a deletion in the TM domain of the human restriction factor disrupts its susceptibility to Nef (Figure 4A) (Jia et al., 2009; Zhang et al., 2009). Together, these findings suggest that tetherin poses a significant - albeit not insurmountable - barrier to transmission of non-human primate lentiviruses to humans. Moreover, it seems that only the pandemic HIV-1 M group has entirely cleared this hurdle because its Vpu protein evolved to become an effective tetherin antagonist, yet maintained its ability to degrade CD4. We therefore speculate that the acquisition of a fully functional Vpu protein by the main group of HIV-1 contributed to the global spread of the AIDS pandemic.
It may seem surprising that SIVcpz uses Nef to antagonize tetherin but switched to use Vpu following transmission to humans. Although further studies are necessary to fully elucidate the evolution of Vpu and Nef functions, a possible scenario is outlined in Figure 6. Recent data show that some primate lentiviruses that do not encode a vpu gene use Nef to antagonize tetherin (Jia et al., 2009; Zhang et al., 2009). Our observation that Vpus from SIVgsn, SIVmus and SIVmon are all capable of counteracting their cognate tetherin suggests that this function was already present in the vpu-containing common ancestor of these viruses (Figure 6B, upper left). Evolutionary analyses indicate that SIVcpz received vpu from the precursor of the SIVgsn/mus/mon lineage, while its nef gene is derived from the SIVrcm lineage (Bailes et al., 2003; Schindler et al., 2006). Thus, during the adaptation of SIVcpz to its chimpanzee host, either Nef or Vpu could have served as the tetherin antagonist. Because both Nef and Vpu antagonize tetherin in a host-specific manner, it is likely that initially they were both only poorly active against the CPZ tetherin (Figure 6B, middle). Our data show that Nef subsequently evolved to become an effective antagonist of CPZ tetherin, possibly because the cytoplasmic tail (CT) of the chimpanzee tetherin differs less from that of monkey tetherins than the TM domain which governs susceptibility to Vpu (Figure 3A). Furthermore, the recombinant virus that comprises SIVcpz (Bailes et al., 2003) may have had a selective advantage over its non-mosaic precursors because its Vpu was capable of degrading CD4 while its Nef was capable of exerting an anti-tetherin activity. Upon cross-species transmission of SIVcpz to humans, however, the Nef-mediated tetherin antagonism was lost, most likely because a deletion in the cytoplasmic domain of the human tetherin abrogated its sensitivity to Nef. This deletion, which may have evolved in humans following encounters with other pathogens, made it difficult, if not impossible, for Nef to restore its anti-tetherin function. Hence, none of the 14 HIV-1 group M, N and O Nefs that we tested had an appreciable activity against HU tetherin. In contrast, the SIVgor Nef was able to adapt to antagonize the GOR tetherin, most likely because that restriction factor differs from that of CPZ only by two amino acid substitutions in the CT domain. Consistent with this hypothesis are recent findings showing that SIVsmm and SIVmac Nefs are antagonists of monkey but not human tetherin (Jia et al., 2009; Zhang et al., 2009). Furthermore, HIV-2, which originated from SIVsmm, appears to use Env to facilitate the release of progeny virions from human cells (Bour and Strebel, 1996), although it remains to be elucidated whether this effect is due to Env-mediated tetherin antagonism. Our data show that the HIV-1 group M Vpu evolved to become an effective antagonist of HU tetherin (Figure 6B, upper right). Thus, the flexibility to switch between Nef and Vpu (which is lacking in HIV-2) to counteract tetherin may have facilitated two cross-species transmission events which may otherwise have been less successful: (i) the transmission of SIVgsn/mus/mon to chimpanzees and (ii) the transmission of SIVcpz to humans (Figure 6).
In contrast to their effect on tetherin, the ability of SIV Vpus to degrade CD4 has been well preserved, and this is also true for most SIV Nef proteins (Schindler et al., 2006). Thus, at least two primate lentiviral accessory genes, i.e. nef and vpu, that antagonize host restriction factors in a species-specific manner also have activities that seem largely independent of the host species. It thus seems beneficial for primate lentiviruses to target antiviral factors by proteins that exhibit functional plasticity. A viral gene encoding a product that becomes entirely non-functional, and cannot readily adapt to a new host environment after cross-species transmission is likely prone to elimination by the rapid accumulation of mutations. However, genes encoding proteins capable of performing multiple functions, including at least one that is not species-specific, will be maintained and may evolve to regain lost activities, sometimes perhaps even at the cost of losing another activity. The acquisition of anti-tetherin activity by HIV-1 group M Vpus, and the loss of anti-CD4 activity by HIV-1 group N Vpus may represent examples of this.
While the specificity of the Vpu/tetherin interaction resembles that of other anti-retroviral host restriction factors and their viral antagonist, such as APOBEC3G and Vif (Gaddis et al., 2004; Sawyer et al., 2004; Schröfelbauer et al., 2004; Zennou and Bieniasz, 2006), there is one important difference: other restriction factors such as APOBEC3G and TRIM5α are unlikely to pose a barrier to cross-species transmission of SIVcpz from chimpanzees to humans. Specifically, the SIVcpz Vif protein is fully capable of eliminating the antiviral activity of human APOBEC3G (Gaddis et al., 2004), while the SIVcpz capsid is not susceptible to restriction by human TRIM5α (Kratovac et al., 2008). Altogether, these previous studies suggested that SIVcpz did not have to overcome a major species barrier to transmission to humans because chimpanzees and humans are genetically closely related. Our finding that SIVcpz Vpus and Nefs were generally poorly active in antagonizing HU tetherin challenges this dogma and strongly suggests that tetherin constitutes a significant barrier to transmission. Our data also provide compelling evidence that host-specific adapation allowed pandemic HIV-1 M strains to overcome this restriction because its Vpu protein gained anti-tetherin activity while maintaining its ability to degrade CD4. In the absence of such adaptation, SIVcpz strains that have crossed the species barrier may not spread efficiently. This may explain why just one of three independent chimpanzee-to-human transmissions of SIVcpz has led to a pandemic. Our finding that Vpus from all group M clades are capable of antagonizing tetherin suggests that this function was acquired early in the evolution of HIV-1 group M. Moreover, the ten vpu alleles from transmitted/founder HIV-1 strains that are responsible for productive infection in humans all antagonized tetherin. In contrast, Vpu proteins of HIV-1 group O are poor tetherin antagonists and this is also true for this group’s Nef proteins. It will thus be interesting to investigate whether HIV-1 O has evolved other mechanisms to counteract this restriction factor (similar to HIV-2). Finally, Vpu alleles from some HIV-1 group N strains exhibit Vpu mediated anti-tetherin activity (Figure 3C) but appear to have lost the CD4 degradation function (Figure 2B). Thus, imperfect adaptation of HIV-1 O and N to the human host may explain why they have spread substantially less efficiently in the human population than HIV-1 M.
In summary, our data suggest that evolutionary switches between Nef- and Vpu-mediated tetherin antagonism facilitated the cross-species transmission that ultimately led to the emergence of pandemic HIV-1 strains. The differences in HIV-1 group M, O and N Vpu function identified here should provide important insights into the role of Vpu and tetherin antagonism in virus transmission and pathogenesis. It has been previously shown that the primate lentiviral lineage that gave rise to HIV-1 lost its ability to suppress T-cell activation (Schindler et al., 2006). Thus, HIV-1 may need a particularly effective tetherin antagonist for effective virus release because it induces higher levels of inflammation (and thus possibly tetherin) than non-pathogenic natural SIV infections (Kirchhoff, 2009). Finally, the present report shows that - with respect to Nef and Vpu function - SIVgsn, SIVmon and SIVmus more closely resemble HIV-1 M, than SIVcpz or other known SIVs. Thus, studies of these natural SIV infections may provide relevant insights into the pathogenesis of AIDS in humans.
Constructs expressing N-terminally HA-tagged tetherins have been described previously (Neil et al., 2008; McNatt et al., 2009). Vpu alleles from the different subtypes of HIV-1 M and some tetherin variants were chemically synthesized (GenScript) (Table S1, Table S2). Cloning of vpu, nef and tetherin alleles into the bi-cistronic CMV-based pCGCG expression vector coexpressing the green fluorescent protein (GFP) was essentially performed as described previously (Greenberg et al., 1998; Schindler et al., 2003). See the Supplemental Experimental Procedures for details.
Generation of HIV-1 (NL4-3-based) proviral constructs containing a deletion of nucleotides 3 to 120 of the vpu coding frame (ΔVpu) or functional nef genes followed by an IRES and the eGFP gene (Nef/eGFP) has been described previously (Rucker et al., 2004; Schindler et al., 2003; Schindler et al., 2006). To generate vpu-defective derivatives of the HIV-1 Nef/eGFP constructs (ΔVpu/Nef/eGFP), restriction fragments encompassing the nef-IRES-eGFP region were cloned into the ΔVpu proviral HIV-1 construct using the unique StuI and NotI restriction sites in the env gene and the flanking vector sequence, respectively. Proviral HIV-1 NL4-3 vectors expressing various vpu alleles independently of overlapping env gene sequences were generated using splice-overlap extension PCR and standard cloning techniques as outlined in the Supplemental Experimental Procedures. All PCR-derived inserts were sequenced to confirm their accuracy.
Blood samples were obtained from HIV-1 infected people living in Cameroon. Genomic DNA was extracted from buffy-coat and fragments spanning the nef and vpu genes were amplified by nested or semi-nested PCR. For details, see the Supplemental Experimental Procedures.
Cells were cultured and transfected as described in the Supplemental Experimental Procedures.
To determine the effect of Vpu on CD4 cell surface expression, 293T cells were transfected by the calcium phosphate method in triplicate with 1 µg of a CD4 expression vector and 5 µg of pCGCG GFP/Vpu or GFP/Vpu-AU1 constructs expressing GFP alone or together with untagged or tagged Vpus. Two days post-transfection CD4 expression was examined by FACS analysis as described previously (Schindler et al., 2003). See Supplemental Experimental Procedures for details.
Virus yield and infectivity assay were conducted in principle as described (Munch et al., 2007; Neil et al., 2008; McNatt et al., 2009; Zhang et al., 2009). For details, see the Supplemental Experimental Procedures.
The effects of Vpu and Nef on viral particle release were monitored as described previously (Neil et al., 2008; NcNatt et al., 2009; Zhang et al., 2009). Expression of Vpu and tetherin, as well as of viral genes by proviral NL4-3 UIE constructs was examined as described in the Supplemental Experimental Procedures.
Tetherin and Vpu amino acid sequences were aligned using Multalign and ClustalW2, respectively, followed by some manual editing. Newly derived vpu, nef and tetherin sequences will be submitted to the GenBank sequence database and accession numbers will be provided upon acceptance of the manuscript.
All statistical calculations were performed with a two-tailed unpaired Students-t-test using Graph Pad Prism Version 5.0. P values <0.05 were considered significant. Correlations were calculated with the linear regression module.
We thank Thomas Mertens for support, Daniela Krnavek, Kerstin Regensburger, Martha Mayer and Nicola Schrott for excellent technical assistance, Guido Silvestri, Ingrid Bennett and Nathalie Arhel for critical reading of the manuscript, and Welkin Johnson for the SM tetherin. The molecular p89.6, pYU-2 and pYK-JRCSF clones and TZM-bl cells were obtained through the NIH AIDS Research and Reference Reagent Program. This work was supported by the Deutsche Forschungsgemeinschaft (DFG) and grants from National Institutes of Health to FK (RO1 AI06757), BHH (R37 AI50529, R01 AI58715, P30 AI27767) and TH (R01 AI078788).