|Home | About | Journals | Submit | Contact Us | Français|
The tetherin/BST2/CD317 protein blocks the release of HIV-1 and other enveloped viruses by inducing tethering of nascent particles to infected cell surfaces. The HIV-1 Vpu protein antagonizes the antiviral activity of human but not monkey tetherins and many simian immunodeficiency viruses (SIVs) do not encode Vpu. Here, we show that the apparently ‘missing’ anti-tetherin activity in SIVs has been acquired by several SIV Nef proteins. Specifically, SIVMAC/SIVSMM, SIVAGM and SIVBLU Nef proteins can suppress tetherin activity. Notably, tetherin antagonism by SIV Nef proteins is species-specific, is genetically separable from other Nef activities and is most evident with simian rather than human tetherin proteins. Accordingly, a critical determinant of sensitivity to SIVMAC Nef in the tetherin cytoplasmic tail is variable in nonhuman primate tetherins and deleted in human tetherin, likely due to selective pressures imposed by viral antagonists, perhaps including Nef proteins.
The tetherin protein (also known as BST2 or CD317) is a potent inhibitor of the release of enveloped viruses. It was recently identified as a factor in human cells that blocks release of HIV-1 particles from infected cells and is counteracted by the viral protein Vpu(Neil et al., 2008; Van Damme et al., 2008). Its precise mechanism of action is not well defined at present, but in cells constitutively expressing tetherin, protease-sensitive tethers retain fully-formed and mature HIV-1 particles on the cell surface and tetherin colocalizes with puncta of Gag that likely represent nascent virions (Jouvenet et al., 2009; Neil et al., 2006; Neil et al., 2007; Neil et al., 2008). Recently, we and others have shown that human tetherin (hu-tetherin) has broad antiviral specificity and inhibits the release of particles assembled using structural proteins from all retroviruses tested, as well as filoviruses and arenaviruses (Jouvenet et al., 2009; Kaletsky et al., 2009; Sakuma et al., 2009).
The mechanism by which HIV-1 Vpu antagonizes hu-tetherin is not fully understood, but overexpressed HIV-1 Vpu reduces the overall levels of tetherin in cells and inhibits its appearance at the cell surface(Bartee et al., 2006; Van Damme et al., 2008). Furthermore, HIV-1 Vpu and hu-tetherin co-localize, and Vpu prevents the co-localization of hu-tetherin with nascent HIV-1 particles(Jouvenet et al., 2009; Neil et al., 2008). However, while tetherin proteins from non-hominid primates are potent inhibitors of HIV-1 particle release, they cannot be counteracted by HIV-1 Vpu(McNatt et al., 2009). Portions of primate tetherin genes, including sequences encoding the transmembrane domain that governs sensitivity to antagonism by Vpu, are unusually divergent, and exhibit clear evidence of positive selection(McNatt et al., 2009). Thus, HIV-1 has apparently acquired a biological activity (i.e. Vpu), that has specifically evolved to antagonize the tetherin variant expressed in its host species.
Although hu-tetherin inhibits the release of particles assembled using a diverse array of retroviral structural proteins, only a subset of the primate lentiviruses encode Vpu. Thus, it seemed reasonable to suppose that SIVs have evolved alternative mechanisms to evade tetherin in their natural hosts. Indeed, earlier work indicated that the HIV-2 envelope protein could enhance particle release from cells that were subsequently shown to express hu-tetherin(Abada et al., 2005; Bour et al., 1996; Bour and Strebel, 1996; Varthakavi et al., 2003). Given this precedent, it was quite plausible that the envelope proteins of SIVs might have a similar function. Consistent with this idea, the Ebola virus envelope protein has recently been reported to be a tetherin antagonist(Kaletsky et al., 2009).
However, as reported herein, we found that the envelope protein of SIVMAC, a macaque lentivirus that is closely related to HIV-2, did not antagonize macaque tetherin proteins. Rather, Nef proteins from SIVMAC and several other SIVs antagonize primate tetherins. Notably, tetherin antagonism by SIV Nef proteins was species-specific, and each SIV Nef was poorly active against human tetherin. Furthermore, the cytoplasmic tail of tetherin, which, like the transmembrane domain, has been evolving under positive selection in primates(McNatt et al., 2009), contains a discrete motif that is deleted in humans and variable in other primates and governs sensitivity to antagonism by SIVMAC Nef. Thus, several primate lentiviruses that lack Vpu have acquired the ability to antagonize tetherin using their Nef proteins.
Hu-tetherin can inhibit the release of particles assembled using the structural proteins (Gag and/or GagPol) of a wide variety of retroviruses(Jouvenet et al., 2009), raising the question of how retroviruses that lack a Vpu gene are efficiently released from infected cells that might ordinarily express tetherin. Among the retroviruses previously tested for sensitivity to hu-tetherin were the primate lentiviruses, SIVMAC and SIVAGMSab, neither of which encode a Vpu protein(Jouvenet et al., 2009). However, it has previously been shown that at least some strains of HIV-2, a virus that shares a recent common ancestor with SIVMAC, encode an envelope protein that has Vpu-like activity (Abada et al., 2005; Bour et al., 1996; Bour and Strebel, 1996; Varthakavi et al., 2003). Therefore, we tested whether the release of particles generated by a full-length SIVMAC239 proviral construct could be inhibited by tetherin proteins found in macaque species, specifically two variants from rhesus macaques (rh-tetherin-1 and -2) and one from pig-tailed macaques (pgt-tetherin). Strikingly, each of the macaque tetherin proteins was a rather poor inhibitor of SIVMAC239 particle release (Figure 1A). This was the case whether or not the SIVMAC239 proviral DNA construct encoded a functional Env protein. Thus, these results suggested that SIVMAC239 might encode a tetherin antagonist other than the Env protein.
Next we co-expressed an SIVMAC239 provirus with varying amounts of human and macaque tetherin proteins. As previously shown for SIVMAC239 particles assembled using only the Gag and Pol proteins, the release of authentic SIVMAC239 virions appeared highly sensitive to hu-tetherin (Figure 1B, D). In contrast, tetherin proteins from macaques appeared to be significantly less effective inhibitors of SIV MAC239 virion release (Figure 1C, D, Figure S1).
These findings suggested the possibility that another SIVMAC protein might be a macaque-tetherin-specific antagonist. The ability of Nef to regulate the levels of cell surface proteins prompted us to examine whether it might fulfill this role. Strikingly, the release of virions generated by a proviral plasmid lacking Nef (SIVMAC239(delNef)) appeared more sensitive to inhibition by macaque tetherins than SIVMAC239(WT) (Figure 1C,D Figure S1). In contrast, hu-tetherin inhibited SIVMAC239(WT) release almost as well as it inhibited SIVMAC(delNef) release (Figure 1B, D). Thus, Nef deletion specifically sensitized SIVMAC239 to inhibition by macaque tetherins.
Scanning electron microscopic analysis of cells cotransfected with plasmids expressing SIVMAC239 Gag and either rh-tetherin-1 or hu-tetherin revealed that either tetherin induced the accumulation of SIV MAC239 VLPs on the surface of cells (Figure 1E). Moreover, we have previously reported that HIV-1 Gag-GFP VLPs tethered to the plasma membrane by hu-tetherin can be internalized resulting in the accumulation of intracellular HIV-1 Gag-GFP (Neil et al., 2006; Neil et al., 2008) and similar findings were obtained with SIVMAC239 Gag-GFP and rh-tetherin-1 (Figure 1F), Specifically, in 293T cells stably expressing rh-tetherin-1, fluorescence microscopic analysis indicated that nearly half the cells transiently expressing SIVMAC239 Gag-GFP contained prominent intracellular accumulations of Gag-GFP (Figure 1F,G). Crucially, these accumulations decreased in frequency when increasing amounts of an SIVMAC239 Nef expression plasmid were cotransfected, or when cells that did not express tetherin were used(Figure 1F,G). Thus, macaque and human tetherin proteins block particle release in the same manner, and SIVMAC239 Nef antagonizes the tethering activity of macaque tetherin.
Vpu is present in only a minority of primate lentiviruses, while Nef is present in all primate lentiviruses isolated thus far. Therefore, to test whether Nef proteins from lentiviruses other than SIVMAC might also be capable of counteracting tetherin, several of them were expressed in place of HIV-1 Nef in the context of a replication competent HIV-1 proviral construct. The HIV-1 Vpu protein was inactivated in this construct, to test the effects of Nef on particle release in the absence of confounding factors. Proviral plasmids bearing the various Nef proteins were co-transfected with increasing amounts of plasmids expressing hu-, rh-1-, pgt-, and African green monkey (agm-) tetherin proteins. Western blot analyses indicated proviruses expressing the various Nef proteins generated HIV-1 virions with similar efficiency in the absence of tetherins (Figure 2A–H).
The generation of extracellular virions by cells transfected with the parental proviral plasmid (which lacked both Nef and Vpu) was strongly inhibited by each of the four tetherins (Figure 2A, Figure S2A), in the absence of effects on cell associated Gag expression. A matched HIV-1 provirus, which lacked Nef but retained Vpu, was largely resistant to hu-tetherin. However, Vpu did not antagonize agm-tetherin, rh-tetherin or pgt-tetherin (McNatt et al., 2009), (Figure 2B, Figure S2B). A construct that lacked Vpu but retained HIV-1 Nef behaved essentially the same as the construct that lacked both Vpu and Nef (Figure 2C, Figure S2C). Thus, HIV-1 Vpu specifically antagonized hu-tetherin but not monkey tetherins, while HIV-1 Nef did not antagonize any of the tetherin proteins.
Notably, replacement of HIV-1 Nef with SIVMAC Nef resulted in substantial resistance to both pgt- and rh-tetherin (Figure 2D, Figure S2D). Importantly, this effect was quite specific to the macaque tetherins; and SIVMAC Nef only marginally enhanced particle release when inhibition was imposed by hu-tetherin or agm-tetherin. Conversely, SIVAGMSab Nef conferred a greater degree of resistance to agm-tetherin than did SIVMAC Nef (Figure 2E, Figure S2E), and SIVAGMSab Nef was also active against the macaque tetherins. However, SIVAGMSab Nef was similar to SIVMAC Nef in that it was a poor antagonist of hu-tetherin (Figure 2E, Figure S2E). Although this sample size is small, we note that in the two cases examined where the parental viruses lack a vpu gene (SIVMAC and SIVAGMSab), the Nef proteins were able to antagonize the antiviral activity of tetherins found in the corresponding simian host species.
We also tested Nef proteins from three other SIVs, with the caveat that the SIV Nef proteins were not derived from the same host species as the tetherins. Nonetheless, an HIV-1 proviral construct expressing the Nef protein from SIVBLU was largely resistant to pgt-tetherin and rh-tetherin (Figure 2F, Figure S2F). Conversely, Nef proteins from SIVRCM and SIVGSN appeared largely unable to antagonize the tetherin proteins (Figure 2G, H Figure S2G, H). However, in addition to the fact that Nef and tetherin proteins from non-orthologous species were tested, SIVGSN naturally encodes a Vpu protein(Courgnaud et al., 2002), which potentially would obviate the requirement for this activity in Nef. An additional caveat with the aforementioned experiment is that the SIV Nef proteins were expressed in the Nef position of an HIV-1 provirus, and may be present at higher or lower levels than in their parental proviruses. Additionally, antibodies to the SIVRCM and SIVGSN proteins were not available, therefore we were unable to demonstrate that these Nef proteins were expressed, although we have previously shown that infection by nearly identical proviruses caused efficient CD4 downregulation (Schindler et al., 2006). Nonetheless, three of the five SIV Nef proteins tested in these assays exhibited apparent tetherin antagonizing activity. The sensitivity or otherwise of a given tetherin protein to antagonism by Nef did not correlate with its expression level (Figure S3), which was assessed under identical conditions to those used for the assays in Figure 2 and Figure S2.
To exclude the possibility the aforementioned findings were the result of some unanticipated consequence of inserting foreign nef genes into the HIV-1 genome in cis, we determined whether Nef proteins, expressed in trans, would relieve tetherin imposed blocks. A Vpu and Nef-defective HIV-1 proviral plasmid was cotransfected with a fixed amount of each tetherin protein, and varying amounts of plasmids expressing the Nef proteins that exhibited apparent tetherin antagonizing activity (Figure 2, Figure S2). We also included HIV-1 Nef as a control, and an additional SIVAGM Nef protein, from SIVAGMTan. As expected, western blot and infectivity assays revealed that particle release was strongly inhibited by each of the tetherin proteins, and co-expression of HIV-1 Nef in trans had little, if any, ability to relieve these tetherin-imposed blocks (Figure 3A, Figure S4A). Conversely, SIVMAC Nef relieved the inhibition of particle release imposed by macaque tetherins but its effect on particle release in the presence of hu-tetherin or agm-tetherin was marginal (Figure 3B, Figure S4B). Both SIVAGMSab and SIVAGMTan Nef restored particle release from cells expressing rh-, pgt- or agm-tetherin but were less active in relieving the block imposed by hu-tetherin (Figure 3C,D, Figure S4C,D). Notably these two SIVAGM derived proteins were the only Nef proteins that effectively relieved inhibition by agm-tetherin. Finally, SIVBLU Nef efficiently antagonized the macaque tetherin proteins, but it also had some ability to enhance particle release in the presence of hu- and agm-tetherin (Figure 3E Figure S4E). However, effects on hu-tetherin and agm tetherin were clearly reduced in magnitude as compared to effects on macaque tetherins. Overall, the effects of the various Nef proteins, expressed in trans, on HIV-1 particle release corroborated those obtained when they were expressed in cis. However, we were able to detect additional, albeit slight, anti-tetherin effects when Nef was expressed in trans (for example, SIVBLU Nef slightly antagonized hu-tetherin (Figure 3E)). We suspect that the small differences in results obtained using the two assays reflect differences in the level of Nef expression, which is likely higher when Nef is expressed in trans. Notably, a modest degree of variation in the levels of expression of the various tetherin proteins (Figure S5), assessed under identical conditions to those used in the assays in Figure 3 and Figure S4, did not correlate with the ability of a given tetherin protein to be antagonized by Nef. Moreover, the Nef proteins that did not antagonize tetherins (HIV-1 Nef) or Nef proteins that did antagonize monkey tetherins (SIVMAC239 Nef or SIVAGMTan Nef) did not appear to inhibit expression of any of the tetherin proteins (Figure S5).
The effects of Nef proteins on particle yield measured using infectious virion yield assays correlated well with those assessed using western blotting (Figure S4). However, in addition to their tetherin antagonizing effects, each Nef protein tested enhanced the intrinsic infectiousness of HIV-1 particles in the absence of tetherins. Nonetheless, the tetherin-independent effect on virion infectiousness was quite similar, and modest, irrespective of which Nef protein was expressed (Figure S4). Together, however, the compound effect of Nef proteins on particle release and infectiousness, resulted in quite large effects on overall infectious virion yield, approaching two orders of magnitude in some cases.
If endogenously expressed monkey tetherins inhibit virion release and are antagonized by SIV Nef proteins then Nef should enhance particle release from simian cells, particularly when they are treated with IFNα, a known enhancer of tetherin expression. We first determined the effect of Nef on SIVMAC239 particle release from rhesus macaque 221 T-cells in single cycle replication assays, following infection with VSV-G-pseudotyped SIVMAC239(WT) or SIVMAC239(del Nef). Notably, 221 expressed tetherin in the absence of IFNα, and the levels of tetherin mRNA were modestly increased by IFNα treatement (Figure S6A). Western blotting analysis indicated that equivalently infected cells release fewer SIVMAC239(del Nef) virions than SIVMAC239(WT) virions, even though Gag was equally expressed (Figure 4A, Figure S6B). This effect was exacerbated when the infected 221 cells were treated with IFN-α. Notably, the discrepancy in particle release between SIVMAC239(WT) and SIVMAC239(del Nef) was not observed when human 293T cells were used in an otherwise identical single cycle replication assay (Figure 4A, Figure S6C). Thus, stimulation of particle release by Nef was clearly host cell-type dependent. Notably, transmission electron microscopic analysis of 221 cells revealed very obvious accumulation of virions associated with the plasma membrane of cells when they were infected with SIVMAC239 (delNef) but not when they were infected with SIVMAC239(WT) (Figure 4B). This phenotype was observed in otherwise untreated 221 cells but was especially striking when the infected 221 cells were treated with IFNα. Furthermore, these virions often appeared to be tethered together (Figure 4B, (iv) inset), Thus, the inefficient release of Nef-defective SIVMAC239 particles was accompanied by accumulation of virions at the plasma membrane, recapitulating previous observations of Vpu-defective HIV-1 virions in hu-tetherin expressing human cells.
We also examined whether SIVAGMSab Nef could enhance particle release from African green monkey COS-7 cells. In this case, VSV-G pseudotyped HIV-1 constructs encoding either no Nef, HIV-1 Nef or SIVAGMSab Nef were used to infect COS-7 cells. Subsequent western blotting analysis indicated that SIVAGMSab Nef significantly increased the yield of HIV-1 particles from HIV-1 infected COS-7 cells, as compared to HIV-1 Nef or no Nef (Figure 4C, Figure S6D) Again this effect was present in the absence of IFNα treatment, suggesting that COS-7 cells express some tetherin, but was exacerbated by IFN-α. Overall, both SIVMAC239 and SIVAGMSab Nef proteins promoted particle release and alleviated the accumulation of tethered virus particles at the plasma membrane, in cells derived from relevant host-species.
The amino-terminus of tetherin is the only portion of the protein that is ordinarily exposed to the cell cytoplasm and was, therefore, the most likely determinant of sensitivity to antagonism by SIV Nef proteins. We constructed two plasmids expressing chimeric tetherins, in which the N-termini of rh-tetherin and hu-tetherin (residues 1–21 and 1–26, respectively) were reciprocally exchanged, generating huCT-rh-tetherin and rhCT-hu-tetherin respectively (Figure 5A). Each of these chimeric tetherin proteins inhibited the release of HIV-1 particles in the absence of Vpu or Nef (Figure 5B). Notably, hu-tetherin and the huCT-rh-tetherin chimera also inhibited the release of HIV-1 particles generated by proviral plasmids that encoded SIVMAC239 Nef in cis, assessed using both western blot or infectious virion assays (Figure 5B, C). Conversely, the rhCT-hu-tetherin protein, like rh-tetherin did not efficiently inhibit the generation of HIV-1 particles generated in the presence of SIVMAC239 Nef. Similar experiments showed that the release of Nef-defective SIVMAC239 virions was inhibited by each of the intact and chimeric tetherins. Conversely, the intact, Nef-expressing SIVMAC239 construct was sensitive to inhibition by hu-tetherin and huCT-rh-tetherin, but was resistant to the rhCT-hu-tetherin and rh-tetherin proteins (Figure 5D,E). Thus, determinants in the cytoplasmic tails of rh- and hu-tetherins govern their sensitivity and resistance, respectively, to antagonism by SIVMAC Nef.
The cytoplasmic tails of rh-tetherins differ from hu-tetherin in three short clusters of amino-acids, including a five-residue sequence that is deleted specifically in the human lineage. Therefore, we constructed three plasmids expressing hu-tetherin in which each variant amino acid cluster was substituted for its equivalent in rh-tetherin-1; (hu(PIL), hu (KM) and hu(GDWIK), Figure 6A). Each of these tetherin proteins was expressed and inhibited the release of SIVMAC239(del Nef) virions with approximately equal efficiency, as assessed by western blotting assays (Figure 6B, D). Moreover, hu(PIL) and hu(KM) tetherin proteins efficiently inhibited the release of SIVMAC239(WT) virions and were, thus, apparently insensitive to antagonism by SIVMAC239 Nef (Figure 6C, D). Conversely, restoration of the ‘missing’ 5 amino acid GDIWK motif in hu-tetherin conferred sensitivity to antagonism by SIVMAC239 Nef. Indeed, hu(GDIWK) was as sensitive to antagonism by SIVMAC239 Nef as was rhCT-hu protein that encoded the complete rh-tetherin-1 cytoplasmic tail (Figure 6C, D).
The cytoplasmic tail of tetherin has been evolving under positive selection, and the 5-residue motif that is deleted in humans contains codons that exhibit a high probability of being positively selected in other primate species(McNatt et al., 2009). Notably, rh-tetherin-1 and rh-tetherin-2 differ at one amino acid within this motif, encoding GDIWK and DDIWK respectively (Figure S7A). Moreover, agm-tetherin, which is insensitive to antagonism by SIVMAC239 Nef (Figure 2, Figure 3) contains a single substitution relative to rh-tetherin-2 tetherin at this motif and encodes DDICK (Figure S7A). To further determine the role of this variable motif in determining susceptibility to antagonism by SIVMAC239 Nef, we generated plasmids expressing rh-tetherin lacking this motif (rh(ΔGDWIK)), as well as single amino acid substitution mutants (W17C) of rh-tetherin-1 and rh-tetherin-2 (rh-1(GDICK) and rh-2(DDICK)). Notably, either the 5 amino acid deletion or the single amino-acid substitution, in the context of either the rh-tetherin-1 or rh-tetherin-2, conferred resistance to antagonism by SIVMAC239 Nef (Figure S7B, C). Thus, naturally occurring deletions and substitutions within the same small motif, found in humans and African green monkeys, can confer resistance to antagonism by Nef proteins found in an SIV from another species.
We next determined whether mutations that specifically abolish known biological activities of SIVMAC239 Nef affected its ability to antagonize tetherin. As is the case with most other biological activities of Nef proteins, mutation of the N-terminal myristate acceptor (G2A) abolished the ability of SIVMAC239Nef to antagonize rh-tetherin (Figure 7A, B). Conversely, mutation of a single residue (Y223F) that abolishes the ability of SIVMAC239 Nef to downregulate MHC-I but leaves other activities of SIVMAC239 Nef, such as CD4 downregulation, intact (Schindler et al., 2004) had little effect on tetherin antagonizing activity. However, a series of three mutants that share the D204R substitution and have lost CD4 and CD28 downregulation activity, but maintain MHC-I downregulation activity(Schindler et al., 2004), were not effective rh-tetherin-1 antagonists (Figure 7A, B). Notably, none of these mutations had any effect on Nef expression levels (Schindler et al., 2004). Moreover, an analysis of 5 naturally occurring Nef proteins derived from the same lentivirus lineage as SIVMAC239 revealed clear differences in tetherin antagonizing activity (Figure 7C,D). Specifically, two SIVSMM derived Nef proteins were efficient rh-tetherin-1 antagonists, while 2 of 3 HIV-2 Nef proteins were devoid of tetherin antagonizing activity and a third exhibited weak rh-tetherin-1 antagonizing activity. Because these Nef proteins all share the ability to downregulate CD4, MHC-I and CD3(Munch et al., 2005), these data show that tetherin antagonism is genetically separable from several known biological activities of Nef, but may share some mechanistic properties with CD4 downregulation. Additionally, these data hint that natural variation in the ability of closely related Nef proteins to act as tetherin antagonists may be influenced by the host species in which the corresponding virus replicates.
Here we report that Nef proteins can enhance virion particle release via antagonism of tetherin proteins. Notably, this activity was not present in HIV-1 Nef, but 5 out of 7 SIV Nef proteins tested, from 5 different SIV lineages, exhibited tetherin-antagonizing activity. Importantly, this Nef function exhibited striking species specificity; none of the SIV Nef proteins were highly active against hu-tetherin, but in both cases where tetherins and Nef proteins were derived from the same host species (macaque and agm) efficient antagonism was observed.
Previously, we showed that the HIV-1 Vpu protein also antagonizes tetherins in a species-specific manner. Determinants of sensitivity to Vpu reside in the hu-tetherin transmembrane domain, which exhibits a strong signature of positive or diversifying selection, suggesting the possibility that Vpu, or a similar viral antagonist of tetherin, has imposed selective pressure on this portion of the tetherin gene. Notably, we also observed signatures of positive selection in sequences encoding the tetherin N-terminal cytoplasmic tail(McNatt et al., 2009). Obviously, given the finding that the cytoplasmic tail of tetherin can determine sensitivity or resistance to antagonism by SIVMAC Nef, then Nef proteins are reasonable candidates for antagonists that impose diversifying selection pressure on this tetherin domain.
The mechanism by which SIV Nef proteins antagonize tetherins remains to be determined. Preliminary experiments did not reveal a clear effect of SIVMAC or SIVBLU Nef proteins on rh-tetherin cell surface expression (data not shown). Nonetheless, HIV-1 and SIV Nef proteins downregulate several other cell surface proteins, most notably CD4 and MHC-I (Collins et al., 1998; Garcia and Miller, 1991; Schindler et al., 2006; Schwartz et al., 1996; Stumptner-Cuvelette et al., 2001), by recruiting these receptors to the endocytic machinery or by redirecting them to lysosomes (reviewed in (Kirchhoff et al., 2008; Roeth and Collins, 2006)). The HIV-1 Vpu protein excludes tetherin from sites of particle assembly at the plasma membrane and inhibits cell surface expression of tetherin(Jouvenet et al., 2009; Van Damme et al., 2008). However, further work will be required to determine precisely how antagonism of tetherin function by Nef and Vpu proteins is achieved.
The ability to antagonize tetherin has apparently arisen at least twice, perhaps three times during the evolution of primate lentiviruses and, strikingly, the acquisition of this function has occurred in completely different ways, employing Nef, Vpu and perhaps Env proteins. Of interest is the fact that both Nef and Vpu proteins, to varying degrees, remove or antagonize molecules (CD4 and tetherin) that are both capable of inhibiting the release of infectious viral particles from the cell surface(Lama et al., 1999; Ross et al., 1999). Moreover, both appear to be multifunctional proteins whose activities partially coincide. Overall these findings underscore the functional plasticity of primate lentiviral genomes in general, and vpu and nef genes in particular.
While the ability of Vpu and Nef proteins to antagonize tetherin is clearly demonstrable, it remains unproven that these activities are critical in vivo. SIVMAC239 chimeras that encode HIV-1 Nef, and therefore lack a predicted tetherin antagonist, are less pathogenic than intact SIVMAC239, but are capable of causing simian AIDS in some animals(Alexander et al., 1999; Kirchhoff et al., 1999). Thus, tetherin antagonism might not be essential for pathogenesis. However, given the plasticity of Nef function it will be interesting to determine whether HIV-1 Nef can acquire tetherin antagonizing activity during SIVMAC239/SHIV replication in vivo. Additionally, HIV-1 Vpu, despite lacking the ability to antagonize macaque tetherins, enhances replication in the context of SHIV infection of pig-tailed macaques(Stephens et al., 2002). Thus, other functions of Vpu and Nef, may be critical in vivo. Nonetheless, the independent acquisition of tetherin antagonizing activity that is specific to the host species of a given HIV/SIV by Nef and Vpu strongly suggests that these activities confer a selective advantage in vivo.
Finally, we note that three of the five known ‘non-essential’ accessory genes in primate lentiviruses, namely vif, vpu and nef, are now demonstrated to encode antagonists of innate or intrinsic immunity. Overall, these findings emphasize the extent to which intrinsic and innate host defenses are likely to have imposed evolutionary pressure on primate lentiviruses, influenced species tropism and provided the impetus for the acquisition of new genes and biological activities to enable immunodeficiency virus replication in an intrinsically hostile environment.
Plasmids expressing the various HA-epitope tagged tetherin proteins were constructed as previously described (Jouvenet et al., 2009; McNatt et al., 2009) using the cDNA sequences and oligonucleotides listed in Supplementary table 1. Two variants of rhesus macaque tetherin were used (rh-1 and rh-2) that differ at a single amino acid position in the cytoplasmic tail. No functional difference between the two variants was detected, and the variants were used interchangeably. Mutant human and rhesus tetherins were generated by PCR using long oligos directed to the 5’ end of the coding sequence (Supplementary table 1). To generate the SIVMAC proviral clones used in this study, the full-length sequence of SIVMAC239 was assembled from plasmids p239SpSp5’ and p239SpE3’ (Kestler et al., 1990)(NIH AIDS Reagent program) into a low copy number (pXF3) backbone. For the SIVMAC(delNef) derivative of this clone, overlapping primers and PCR were used to introduce premature stop codon at codons 58 and 59 of Nef, immediately 3’ to the Env stop codon, using the primers listed in Supplementary table 2. HIV-1-based proviral plasmids expressing Nef proteins from various lentiviruses have been previously described(Munch et al., 2007; Schindler et al., 2006). To generate vpu-defective derivatives thereof, restriction fragments encompassing the nef region were cloned into pBRHIV-1NL4-3Δvpu containing a deletion of nucleotides 3 to 120 of the vpu coding frame. For the experiments in which Nef proteins were expressed in trans, Nef coding sequences from HIV-1, SIVMAC239, SIVBLU, SIVAGMSab and SIVAGMTan were amplified by PCR using primers introducing EcoRI and SalI sites at the 5’ end and 3’ end of the coding sequence respectively (Supplementary table 2). The PCR products were inserted into the pIRES2-GFP expression vector (Clontech). SIVMAC239 Nef was also inserted into pCR3.1. pCG-IRESGFP vectors expressing the SIVMAC239 mutant, HIV-2 and SIVSMM Nef proteins used in Figure 7 have been described previously(Schindler et al., 2004).
293T and TZMbl cells (which express CD4 and CCR5 and contain a lacZ reporter gene under the control of an HIV-1 LTR) were maintained under standard conditions. All transfection experiments were performed in 293T cells, which do not constitutively express tetherin, that were seeded in 24-well plates at a concentration of 2.5×105 cells/well and transfected the following day using polyethylenimine (PolySciences). A 293T-;derived cell line stably expressing rh-tetherin was derived by inserting the rh-tetherin-1 cDNA into a retroviral vector (LHCX) which was then used to transduce 293T cells, which were then selected in hygromycin.
To measure tetherin inhibition of SIVMAC239 particle release, cells were transfected with 400ng of an SIVMAC239 proviral plasmid or the SIVMAC (delNef) derivative, along with increasing amounts (3.12ng, 6.25ng, 12.5ng, 25ng, or 50ng) of each tetherin-HA expression plasmid. In some experiments (Figure 5), the amount of SIVMAC proviral plasmid DNA was increased to 500ng and the amounts of tetherin plasmid were increased (to 0ng, 12.5ng, 25ng, 50ng, 100ng, 200ng/well). For assays using the HIV-1 proviral plasmids encoding various SIV Nef proteins, cells were transfected with 200ng of each proviral plasmid and varying amounts (0ng, 3.12ng, 6.25ng, or 12.5ng) of each tetherin-HA plasmid. Alternatively, 500ng of HIV-1 proviral plasmid and 0ng, 50ng, 100ng, and 200ng of tetherin expression plasmid were used.
For assays in which Nef proteins were expressed in trans, cells were transfected with 400ng of the HIV-1 proviral plasmid that lacked both Nef and Vpu, along with 25ng of each tetherin-HA expression plasmid and varying amounts (0ng, 25ng, 50ng or 100ng) of each HIV-1 or SIV Nef expression plasmid. In all transfection experiments, the total amount of DNA was held constant within the experiment by supplementing the transfection with empty expression vector.
To generate the VSV-G pseudotyped viruses in Figure 4, 10µg of proviral construct and 2µg of VSV-G expression plasmid were used to transfect a 10cm dish of 293T cells. Viral stocks were titered on TZMbl indicator cells.
At 48 hrs post transfection, virion containing culture supernatants were harvested, clarified by low speed centrifugation and filtered (0.2 µm). Infectious virus release was determined by inoculating TZMbl indicator cells, plated the previous day in 96 well plates at 8×103 cells/well, with 50µl of serially diluted supernatants. At 48 hrs after infection, β-galactosidase activity was determined using GalactoStar reagent (Perkin Elmer). The remainder of the virion containing supernatant (750µl) was layered onto 400µl of 20% sucrose in PBS and centrifuged at 20,000 g for 2 hours at 4°C. Virion pellets, and corresponding virion producing cells were dissolved in SDS PAGE loading buffer. Virion and cell lysates were separated on 4–12% acrylamide gels, blotted onto nitrocellulose membranes and probed with anti-HIV-1-p24CA (183-H12-5C, which also recognizes SIVMAC239 p27CA) or anti-HA (Covance) monoclonal antibodies. Subsequently, blots were probed with anti-mouse HRP-conjugated goat secondary antibodies (Jackson) and proteins revealed using chemiluminescence detection reagents (Pierce). Alternatively, for quantitative western blotting, blots were simultaneously or individually probed with anti-CA and rabbit anti-HA antibodies, followed by goat anti-rabbit and anti-mouse antibodies conjugated to IRDye680 and IRDye800CW, respectively. Fluorescent signals were detected and quantitated using a LICOR Odyssey scanner.
Cells were seeded on 3.5-cm, glass-bottomed dishes coated with poly-L-lysine (Mattek) The following day, they were transfected with 1.6µg of a mixture of plasmids expressing untagged SIVMAC239 Gag and SIVMAC239 Gag-GFP, along with 1.6 µg of empty pCR3.1 or pCR3.1/SIVMAC239 Nef, using Lipofectamine 2000. Cells were fixed 24 h later and observed by deconvolution microscopy using an Olympus IX70-based Deltavision microscopy suite (Tokyo, Japan) as described previously(Neil et al., 2006). For transmission electron microscopy studies, 221 cells were fixed 48 hrs post-infection, using 4% paraformaldehyde, for 10 min. Cells were then pelleted by centrifugation and processed for electron microscopy as described previously(Neil et al., 2006). Alternatively, HT1080 cells were cotransfected with 200ng each of a plasmid expressing a codon optimized SIV Gag-IRES GFP cassette and either rh-tetherin-1 or hu-tetherin expression plasmids inspected by fluorescent and scanning electron microscopy using a Hitachi S4700 field emission SEM (University of Missouri Electron Microscopy Core Facility) as described previously(Zhadina et al., 2007).
We thank members of the Bieniasz Lab and Stuart Neil for helpful discussions. This work was supported by grants from the NIH, R01AI078788 (to TH), R01AI050111 (to PDB), and R01AI067057 (to FK), the Deutsche Forschungsgemeinschaft and the Howard Hughes Medical Institute.