|Home | About | Journals | Submit | Contact Us | Français|
BST-2/tetherin is an interferon-inducible protein that restricts the release of enveloped viruses from the surface of infected cells by physically linking viral and cellular membranes. It is present at both the cell surface and in a perinuclear region, and viral anti-tetherin factors including HIV-1 Vpu and HIV-2 Env have been shown to decrease the cell surface population. To map the domains of human tetherin necessary for both virus restriction and sensitivity to viral anti-tetherin factors, we constructed a series of tetherin derivatives and assayed their activity. We found that the cytoplasmic tail (CT) and transmembrane (TM) domains of tetherin alone produced its characteristic cellular distribution, while the ectodomain of the protein, which includes a glycosylphosphatidylinositol (GPI) anchor, was sufficient to restrict virus release when presented by the CT/TM regions of a different type II membrane protein. To counteract tetherin restriction and remove it from the cell surface, HIV-1 Vpu required the specific sequence present in the TM domain of human tetherin. In contrast, the HIV-2 Env required only the ectodomain of the protein and was sensitive to a point mutation in this region. Strikingly, the anti-tetherin factor, Ebola virus GP, was able to overcome restriction conferred by both tetherin and a series of functional tetherin derivatives, including a wholly artificial tetherin molecule. Moreover, GP overcame restriction without significantly removing tetherin from the cell surface. These findings suggest that Ebola virus GP uses a novel mechanism to circumvent tetherin restriction.
Pathogenic viruses often have evolved mechanisms to neutralize host defenses that act at the cellular level to interfere with the virus life cycle. Such cellular restriction factors have been most extensively characterized for HIV-1 (38) and include the interferon-inducible membrane protein BST-2/HM1.24/CD317/tetherin (28, 40). If unchecked, tetherin blocks the release of newly formed HIV-1 particles from cells by physically tethering them at the cell surface (7, 28, 32, 40). In addition, tetherin has been shown to act against a broad range of enveloped viral particles, including retroviruses, filoviruses, arenaviruses, and herpesviruses (17, 18, 23, 35). In turn, certain viruses that are targeted by tetherin appear to have evolved counteracting activities, and anti-tetherin factors so far identified include HIV-1 Vpu; HIV-2 Env; simian immunodeficiency virus (SIV) Nef, Vpu, and Env proteins; Ebola virus GP; and Kaposi's sarcoma-associated herpesvirus (KSHV) K5 (11, 16, 18, 20, 23, 28, 36, 40, 44, 45).
Tetherin is a homodimeric type II integral membrane protein containing an N-terminal cytoplasmic tail (CT), a single-pass transmembrane domain (TM), an ectodomain-containing predicted coiled-coil regions, two glycoslyation sites, three conserved cysteines, and a C-terminal glycosylphosphatidylinositol (GPI) anchor (2, 19, 31). This unusual topology, with two independent membrane anchors, has led to the suggestion that the retention of virions at the cell surface arises from tetherin's ability to be inserted simultaneously in both host and viral membranes (28, 32, 41) or, alternatively, that dimers or higher-order complexes of tetherin conferred by the ectodomain mediate this effect (39). Interestingly, an artificial tetherin containing the same structural features as the native protein but constructed from unrelated sequences was able to restrict both HIV-1 and Ebola virus particles (32). This suggests that the viral lipid envelope is the target of tetherin and provides an explanation for tetherin's broad activity against diverse enveloped viruses.
A fraction of tetherin is present at the plasma membrane of cells (9, 14), and it has been proposed that viral anti-tetherin factors function by removing this cell surface fraction (40). This now has been shown to occur in the presence of HIV-1 Vpu (5, 7, 15, 26, 34, 40, 44), HIV-2 Env (5, 20), SIV Env (11), SIV Nef (15), and KSHV K5 (3, 23). In addition, certain anti-tetherin factors also may promote the degradation of tetherin, as has been observed for both HIV-1 Vpu (3, 5, 7, 10, 22, 26, 27) and KSHV K5 (3, 23), although Vpu also appears able to block tetherin restriction in the absence of degradation (8), and no effects on tetherin steady-state levels have been observed in the presence of either the HIV-2 or SIVtan Env (11, 20). Simply keeping tetherin away from the cell surface, or targeting it for degradation, may not be the only mechanism used by anti-tetherin factors, since it also has been reported that Vpu does not affect the levels of surface tetherin or its total cellular levels in certain T-cell lines (27).
The interactions between tetherin and viral anti-tetherin factors show evidence of species specificity, suggesting ongoing evolution between viruses and their hosts. HIV-1 Vpu is active against human and chimpanzee tetherin but not other primate tetherins (10, 25, 34, 36, 44, 45), while SIV Nef proteins are active against primate but not human tetherins (16, 36, 44, 45). This suggests that, unlike tetherin restriction, the action of the anti-tetherin factors may involve specific sequence interactions. Indeed, the TM domain has been recognized as a target for HIV-1 Vpu (10, 15, 16, 25, 34), while a single point mutation introduced into the extracellular domain of human tetherin can block its antagonism by the SIVtan Env (11).
In the present study, we investigated the roles of the different domains of tetherin in both promoting virus restriction and conferring susceptibility to the anti-tetherin factors encoded by HIV-1, HIV-2, and Ebola virus. We confirmed that tetherin restriction can be conferred by proteins that retain the two distinct membrane anchors, while signals for the cellular localization of the protein reside in the CT/TM domains of the protein. We found that the Vpu protein targets the TM domain of tetherin, while the HIV-2 Env targets the ectodomain of the protein. In contrast, the Ebola virus GP appears to use a non-sequence-specific mechanism to counteract tetherin restriction, since even an artificial tetherin could be successfully overcome by GP expression. Interestingly, Ebola virus GP counteracted tetherin restriction without removing the protein from the cell surface, suggesting that it is possible to overcome this restriction by mechanisms other than blocking tetherin's cell surface expression.
HeLa and 293T cells were obtained from the American Type Culture Collection; 293A cells were obtained from Qbiogene/MP Biomedicals (Irvine, CA). All cells were maintained in D10 medium: Dulbecco's modified Eagle's medium (DMEM) (Mediatech, Herndon, VA) supplemented with 10% fetal bovine serum (FBS) (Mediatech) and 2 mM glutamine (Gemini Bio-Products, West Sacramento, CA).
Plasmid pHIV-1-pack expresses HIV-1 Gag-Pol and Rev and generates HIV-1 VLPs (1). Plasmid pcDNA-Vphu encodes a human codon-optimized form of Vpu from HIV-1NL4-3 (Vphu) (29) and was provided by Klaus Strebel (NIH); plasmid Vphu-HcRed expresses Vphu with a C-terminal fusion of HcRed (42) and was provided by Paul Spearman (Emory University). An expression plasmid for the HIV-2 Env from virus strain ROD10 has been described previously (1, 4). An expression plasmid for human tetherin/BST-2 (pCMV6-XL5-Bst2) was obtained from Origene (Rockville, MD). In addition, tetherinA100D contained a point mutation described previously (11), and tetherinHA contained a hemagglutinin (HA) tag between amino acids 154 and 155, similarly to a construct previously described (25). TfR-tetherin derivatives were made in the pCMV6-XL5 plasmid backbone using splice-overlap PCR as follows: TfR(CM)-tetherin comprised the CT and TM of the human transferrin receptor 1 (TfR) (residues 1 to 100) fused to the ectodomain of tetherin (residues 44 to 180); TfR(M)-tetherin comprised the CT of tetherin (residues 1 to 21), the TM of TfR (residues 62 to 100), and the ectodomain of tetherin (residues 44 to 180); and TfR(C)-tetherin comprised the CT of TfR (residues 1 to 61) and the TM and ectodomain of tetherin (residues 22 to 180). N-terminal enhanced green fluorescent protein (EGFP)-tagged versions of tetherin and the TfR-tetherin derivatives were made by being cloned into vector pEGFP-C1 (Clontech, Mountain View, CA), with the addition of a 12-amino-acid linker, GHGTGSTGSGSS, between the two proteins. The construct mini-tetherin comprised residues 1 to 50 of tetherin, followed by a two-glycine residue linker, and was fused to EGFP (minus the Met start codon). Artificial tetherin-like protein (art-tetherin) (32) was kindly provided by Paul Bieniasz (Aaron Diamond AIDS Research Center) and cloned into the pCMV6-XL5 plasmid backbone. Ebola virus Zaire GP-8A, which expresses the full-length Ebola virus GP (43), was kindly provided by Heinz Feldmann (Public Health Agency of Canada) and cloned into pCI-neo (Promega, Madison, WI) to generate plasmid pEboGP. GP-8A also was cloned into plasmid pIRES2-DsRed2 (Clontech) to generate plasmid pEboGPiresDsRed. An in-house DsRed expression plasmid, pCCL-DsRedExpress, was used as a transfection marker in flow cytometry experiments.
HIV-1 VLPs were generated in 293A or 293T cells by transient transfection of 80 to 90% confluent cultures with plasmid pHIV-1-pack, plus any additional specified plasmids, using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) as previously described (1). Cell lysates were harvested, and viral particles were collected after 24 h from the supernatant and analyzed by Western blotting, as previously described (30). HIV-1 p24-reacting proteins were detected using rabbit HIV-1SF2 p24 antiserum (AIDS Research and Reference Reagent Program; ARRRP) at a 1:3,000 dilution, followed by horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG (1:12,000) (Santa Cruz Biotechnology, Santa Cruz, CA). Specific proteins were visualized using the enhanced chemiluminescence (ECL) detection system (Amersham International, Arlington Heights, IL). Exposed and developed films were scanned and quantified using the public domain NIH ImageJ software. The intensity of the p24-reacting bands on the Western blots was measured, and the ratio of the signal in VLPs to lysates was obtained. The percent VLP release was calculated by normalizing all values to the pHIV-1-pack-only control (100%). VLP release into the supernatant also was measured using an Alliance p24 antigen enzyme-linked immunosorbent assay (ELISA) kit (PerkinElmer, Waltham, MA), and the percent p24 release was calculated by normalizing data to the pHIV-1-pack-only control (100%). Data obtained by either the scanning of Western blots or p24 ELISA were highly correlative (Pearson's correlation coefficient; r > 0.9). The statistical significance of data was determined using one-way analysis of variance (ANOVA) followed by Dunnett's multiple comparison test from GraphPad Prism (GraphPad Software, La Jolla, CA).
Tetherin and derivatives were detected by the Western blotting of cell lysates using a 1:20,000 dilution of polyclonal rabbit anti-BST-2 serum (ARRRP; from Klaus Strebel), followed by HRP-conjugated goat anti-rabbit IgG (1:12,000) (Santa Cruz Biotechnology). EGFP-tagged constructs were detected using mouse anti-GFP 3E6 antibody (1:1,000) (Invitrogen), followed by HRP-conjugated goat anti-mouse IgG (1:10,000) (Pierce, Rockford, IL). TetherinHA and art-tetherin were detected using monoclonal mouse anti-HA IgG at 1:1,000 dilution (Roche Applied Science, Indianapolis, IN), followed by HRP-conjugated goat anti-mouse IgG (1:10,000). Viral proteins were detected using 1:1,000 dilutions of rabbit antisera raised against HIV-1NL4-3 Vpu (ARRRP; from Frank Maldarelli and Klaus Strebel), HIV-2ST Env (SU) (ARRRP; from Raymond Sweet), and Ebola virus sGP, (kindly provided by Paul Bates, University of Pennsylvania), followed by a 1:12,000 dilution of HRP-conjugated goat anti-rabbit IgG. Specific bands were visualized by ECL.
HeLa or 293A cells were transfected with specific expression plasmids in 10-cm dishes, and 18 to 24 h later they were seeded on coverslips coated with poly-l-lysine (Sigma-Aldrich). The cells were incubated for an additional 24 h at 37°C and then processed for antibody staining. Cells were fixed with 4% paraformaldehyde for 20 min at room temperature, washed three times in PBS, and permeabilized for 10 min in 0.1% Triton X-100 at room temperature, followed by three additional phosphate-buffered saline (PBS) washes. For the analysis of cell surface proteins, cells were placed at 4°C for 20 min, incubated with fresh D10 plus antibodies at 4°C for 30 min, washed with PBS, and then fixed in 4% paraformaldehyde. Tetherin was detected using a polyclonal mouse anti-BST-2 antibody, MaxPab H00000684-B02P (Abnova, Taipei City, Taiwan), at a 1:250 dilution. Mouse anti-GFP monoclonal antibody 3E6 (Invitrogen) was used at a 1:500 dilution. Vpu was detected using rabbit HIV-1NL4-3 Vpu antiserum (ARRRP) at a 1:1,000 dilution, and HIV-2 Env was detected using rabbit polyclonal serum against HIV-2ST SU (ARRRP) at a 1:1,000 dilution. The conjugated secondary antibodies used were donkey anti-mouse Alexa Fluor 488 and donkey anti-rabbit Alexa Fluor 594 (Invitrogen). Processed cells were mounted in Prolong Gold antifade reagent with 4′,6′-diamidino-2-phenylindole (DAPI) (Invitrogen). Images were acquired with an Ultraview ERS laser spinning disk confocal imaging system (PerkinElmer) at 100× magnification and processed using Volocity software (Improvision; PerkinElmer) and Adobe Photoshop Creative Suite 2.
293T cells were transfected with 200 ng of tetherin or tetherin derivative plasmids, 250 ng of pCCL-DsRedExpress, and 3 μg of viral anti-tetherin factors, as indicated. Plasmid amounts were kept constant by the addition of control CMV expression plasmids. Cell surface tetherin was stained using a 1:5,000 dilution of polyclonal rabbit anti-BST-2 (ARRRP), followed by a 1:100 dilution of donkey anti-rabbit IgG conjugated to Alexa Fluor 647 (Invitrogen). Cells were analyzed on a FACSCaliber (BD, Franklin Lakes, NJ). Viable cells were gated on forward and side scatter (FCS and SSC), and the DsRed-positive population was analyzed for tetherin expression. The surface expression of mini-tetherin was detected using a 1:200 dilution of mouse anti-GFP monoclonal antibody 3E6 (Invitrogen), followed by a 1:100 dilution of donkey anti-mouse Alexa Fluor 647 (Invitrogen) and fluorescence-activated cell sorting (FACS) analysis. Mean and peak fluorescence intensities were calculated. Percent maximum histograms were created using FlowJo analysis software (Ashland, OR).
Tetherin is a homodimeric type II integral membrane protein that is anchored in the membrane by both a transmembrane domain and a GPI anchor (19). Tetherin restriction has been shown to require both membrane anchors but specific sequences are not required, since an artificial tetherin that mimics the structure of the protein also can restrict virus release (32). We investigated the domains of tetherin necessary for both its characteristic cellular location and virus restriction properties by constructing chimeric proteins between tetherin and another type II integral membrane protein, the human transferrin receptor type 1 (TfR). Three different TfR/tetherin proteins were made, each preserving the entire ectodomain and glycosylation sites of tetherin and the GPI anchor but replacing either the CT and/or TM domain(s) (Fig. (Fig.1A).1A). We chose to extend the TM domain of TfR to include the first 11 amino acids of the ectodomain of the protein, since hydrophilic signals C terminal to the TM of type II membrane proteins can influence both membrane translocation and orientation (13, 21). An additional tetherin derivative, mini-tetherin, was constructed by fusing the first 50 amino acids of tetherin, comprising the CT and extended TM domain, to EGFP (Fig. (Fig.1A1A).
The expression of each tetherin chimera was confirmed by Western blotting, where both native tetherin and the derivatives ran as multiple bands between 25 and 40 kDa (Fig. (Fig.1B).1B). TfR(M)-tetherin ran at approximately the same size as tetherin, while the longer CT of the TfR protein meant that both TfR(CM)-tetherin and TfR(C)-tetherin ran higher. We also determined the level of cell surface expression of each protein by FACS analysis using an anti-tetherin antibody (Fig. (Fig.1C).1C). Here, we noted that although the surface levels of TfR(C)-tetherin were very similar to that of the wild-type tetherin, both chimeras containing the TM domain of TfR were present at significantly higher levels, as determined by relative fluorescence intensities.
We evaluated the ability of each tetherin derivative to inhibit the release of HIV-1 Gag-Pol virus-like particles (VLPs) produced from transiently transfected 293A cells, which are naturally nonrestrictive and express only a low level of endogenous tetherin (Fig. (Fig.1B).1B). In agreement with Van Damme et al. (40), we observed that exogenously transfected tetherin was extremely potent, with as little as 100 ng of plasmid DNA able to confer the profound restriction of VLP release (Fig. (Fig.1D).1D). We saw no reduction in VLP release when mini-tetherin was expressed, but a significant reduction occurred for all three of the TfR-tetherin chimeras (Fig. (Fig.1C).1C). Similar results have been observed with TfR-tetherin chimeras generated by Iwabu et al. (15). Taken together, these data demonstrate that the ectodomain of tetherin is both necessary and sufficient to confer restriction when presented by the membrane-targeting domain of another type II membrane protein. This also agrees with studies that have found that an artificial tetherin that retains the key structural features of the protein also is able to restrict virus release (32).
Tetherin cycles between the plasma membrane and an intracellular compartment that overlaps with cellular markers for the trans-Golgi network (TGN) (6, 12, 19, 24, 33). To facilitate the visualization of the cellular distribution of tetherin and its derivatives, we added N-terminal EGFP tags to all proteins and confirmed their expression by Western blotting with an anti-GFP antibody (Fig. (Fig.2A).2A). The chimeric proteins ran as multiple bands between 48 and 60 kDa, with additional slower-running bands that likely corresponded to dimers or higher-order multimers. In contrast, mini-tetherin, which does not contain either the cysteines or glycosylation sequences present in tetherin's ectodomain, ran as a single band at the expected size of 32 kDa. We noted that the EGFP-TfR(C)-tetherin construct was expressed at lower levels than those of the other chimeric proteins. The analysis of the effects of each chimera on VLP release confirmed that the EGFP-tagged proteins retained potent restriction activity (Fig. (Fig.2B2B).
In HeLa cells, endogenous tetherin can be detected at both the cell surface (using nonpermeabilized cells) and within an intracellular perinuclear compartment (Fig. (Fig.2C).2C). This intracellular compartment overlapped with markers for the TGN (data not shown), as has been previously reported (6, 19, 24, 33). The cellular distributions of EGFP-tetherin and derivatives were analyzed in 293A and HeLa cells, using both anti-tetherin and anti-EGFP antibodies (Fig. (Fig.2D).2D). The use of an anti-EGFP antibody allowed clearer visualization of cellular distribution than that obtained with either the anti-tetherin antibody or from EGFP fluorescence alone, in particular enabling good visualization of the cell surface rim, and it also revealed more intracellular puncta than the anti-tetherin antibody. The use of an anti-EGFP antibody also allowed the analysis of the proteins in HeLa cells against a background of endogenous tetherin. In both 293A cells and HeLa cells, we observed no major differences in the subcellular localization of EGFP-tagged tetherin or the tetherin derivatives compared to that of the native untagged protein present in HeLa cells (Fig. (Fig.2A2A).
We examined the ability of three anti-tetherin factors, HIV-1 Vpu, HIV-2 Env, and Ebola virus GP, to counteract the restriction caused by each of the functional tetherin derivatives. Tetherin and its derivatives were transiently expressed in 293A cells in the absence or presence of each viral factor, and the effects on VLP release were examined by either Western blotting or p24 ELISA (Fig. (Fig.3).3). Data obtained by either densitometry analysis of p24-reacting bands on the Western blots or the p24 ELISA were highly correlative (r = 0.96). We noted that the recovery of VLP release in the presence of both wild-type tetherin and the viral factors was incomplete, resulting in between 40 (Ebola virus GP) and 75% (Vpu) of the levels obtained in the absence of the tetherin. Such incomplete recovery previously has been observed when exogenous tetherin is expressed in human cells that are naturally nonrestrictive (16, 25, 40, 44).
When the activities of the viral anti-tetherin factors were examined against the chimeric TfR-tetherin proteins, we noted some significant differences. Vpu was unable to counteract the two tetherin derivatives lacking the native TM domain but could overcome TfR(C)-tetherin efficiently. These observations are in agreement with previous reports that have mapped the TM domains of both Vpu and tetherin as likely interacting partners (10, 15, 16, 25, 34). In contrast, both the HIV-2 Env and Ebola virus GP were able to counteract restriction conferred by all of the TfR-tetherin chimeras. Coimmunoprecipitation studies have suggested that both HIV-2 Env and Ebola virus GP directly interact with tetherin (18, 20), and our data further suggest that any such interactions would not require either the TM or CT domain of tetherin. Finally, we noted that the Ebola virus GP was significantly more active against the TfR-tetherin chimeras than the native protein, and that despite the fact that Ebola virus GP only restored VLP release to 40% of the control (no tetherin) levels when coexpressed with the wild-type protein, it was able to fully counteract the restriction imposed by each of the TfR-tetherin chimeras.
We also examined the effects that each viral antagonist had on the steady-state levels of the tetherin derivatives. We found that Vpu expression reduced levels of both wild-type tetherin and TfR(C)-tetherin, but for TfR(CM)-tetherin and TfR(M)-tetherin, where Vpu was unable to counter restriction, we observed no effects on tetherin levels. The reduction in levels of TfR(C)-tetherin indicates that Vpu's ability to degrade tetherin does not require the specific sequences present in the cytoplasmic domain of tetherin despite the fact that this region in the native protein has been shown to contain residues that are involved in Vpu-mediated degradation (8). In contrast, HIV-2 Env and Ebola virus GP had no effect on the levels of any tetherin protein, which is in agreement with previous observations made with wild-type tetherin (11, 18, 20).
It previously has been shown that both HIV-1 Vpu and HIV-2 Env have the ability to remove tetherin from the cell surface (5, 7, 15, 20, 26, 34, 40, 44). We next examined the effect of Vpu expression on the surface levels and cellular distribution of tetherin and its derivatives to determine whether there was a correlation between the ability of Vpu to counteract a tetherin derivative and remove it from the cell surface. Using FACS analysis we found that Vpu only reduced the surface levels of constructs containing the native tetherin TM domain, including the mini-tetherin protein (Fig. (Fig.4A).4A). Confocal analysis of EGFP-tagged tetherin derivatives confirmed that proteins containing the tetherin TM sequence also were removed from the cell surface by Vpu and, instead, concentrated in an intracellular compartment in both 293A cells (Fig. (Fig.4B)4B) and HeLa cells (data not shown). Although the cell surface expression of neither EGFP-TfR(CM)-tetherin nor EGFP-TfR(M)-tetherin was altered by Vpu expression, we did note that Vpu expression still caused the appearance of a brighter and larger concentration of the proteins in the perinuclear compartment, which may reflect the disruption of the TGN structure that has been reported previously following Vpu expression (41, 42). Overall, these data confirm that Vpu-tetherin interactions require the TM domain of tetherin and further reveal that the cytoplasmic tail and TM domains of tetherin, as present in mini-tetherin, are sufficient to confer this effect.
We next examined the ability of the HIV-2 Env to alter the cellular distribution of tetherin and its derivatives. FACS analysis revealed that all of the proteins except mini-tetherin were downregulated from the surface by HIV-2 Env coexpression (Fig. (Fig.5A).5A). This suggests that in contrast to the situation with Vpu, the tetherin ectodomain, but not the CT or TM domain, is the target of the HIV-2 Env. Confocal analysis of EGFP-tagged proteins in both 293A cells (Fig. (Fig.5B)5B) and HeLa cells (data not shown) confirmed the loss of cell surface staining for both EGFP-tetherin and the EGFP-TfR(C)-tetherin chimera, but significant levels of the CM and M chimeras remained at the surface. This likely reflects the higher overall levels of these two chimeras compared to that of EGFP-tetherin, since these constructs retained relatively high mean and peak fluorescence intensities even when downregulated by the HIV-2 Env (Fig. (Fig.5A).5A). For both tetherin and all three TfR-tetherin chimeras, HIV-2 Env expression increased the appearance of the proteins in the perinuclear compartment.
We next analyzed the effect of the Ebola virus GP on the cellular distribution of tetherin and its derivatives. To facilitate confocal analyses, Ebola virus GP was expressed from a plasmid that also expressed a DsRed reporter protein translated from an internal ribosome entry site. We confirmed that this Ebola virus GP expression plasmid also efficiently countered tetherin restriction and enhanced HIV-1 VLP release in 293A cells cotransfected with tetherin (Fig. (Fig.6A).6A). Interestingly, when we analyzed the effects of GP expression on the cellular distribution of EGFP-tetherin and derivatives in transfected 293A cells, we did not observe any differences (Fig. (Fig.6B).6B). FACS analysis confirmed that GP expression did not significantly remove these proteins from the cell surface (Fig. (Fig.6C),6C), although we sometimes observed a small shift to the left of the curves when GP was present, corresponding to the loss of the brightest population of cells. Taken together, these data suggest that Ebola virus GP does not counteract tetherin restriction by removing it from the cell surface, as we observed with Vpu and HIV-2 Env, but instead uses another mechanism to counteract its effects.
A recent study reported that an alanine at position 100 in the ectodomain of human tetherin was necessary for counteraction by the SIVtan Env, and that a point mutant, tetherinA100D, could be overcome by HIV-1 Vpu but not SIVtan Env (11). Since our observations suggested that the ectodomain of tetherin is targeted by HIV-2 Env and Ebola virus GP, we next analyzed their activities against this mutant tetherin. We confirmed that tetherinA100D restricted VLP release as efficiently as the wild-type protein and remained sensitive to Vpu (Fig. (Fig.7A).7A). However, similarly to the case with SIVtan Env, we found that HIV-2 Env had no activity against this mutant (Fig. (Fig.7A),7A), and FACS analysis further demonstrated that HIV-2 Env did not remove tetherinA100D from the cell surface (Fig. (Fig.7B).7B). In contrast, Ebola virus GP retained the ability to overcome restriction by the A100D mutant (Fig. (Fig.7A),7A), and, similarly to its lack of effect against the wild-type protein, had no effect on the cell surface level of this tetherin (Fig. (Fig.7B7B).
A wholly artificial tetherin that mimics the key structural features of tetherin has been described previously (32). Art-tetherin restricts virus release but is resistant to the HIV-1 Vpu. We examined the effects of art-tetherin on HIV-1 VLP release, where we determined that the protein was indeed able to block release, although it appeared to be significantly less potent than the native tetherin at equivalent levels of expression (Fig. (Fig.7C).7C). Using expression levels of both an HA-tagged native tetherin and art-tetherin that gave similar levels of restriction, we next examined the abilities of Vpu, HIV-2 Env, and Ebola virus GP to counteract their effects. As expected, HIV-1 Vpu had no effect against art-tetherin, and the HIV-2 Env also was not able to overcome this restriction. In contrast, the Ebola virus GP was fully active against art-tetherin. Indeed, its ability to completely overcome restriction resulting from art-tetherin contrasted with its less potent activity against the native protein, especially compared to HIV-1 Vpu or HIV-2 Env. In sum, these data indicate that the mechanism used by Ebola virus GP to overcome tetherin restriction does not require either cell surface removal or specific sequences in tetherin, suggesting that a more indirect mechanism is involved.
The block to virus release conferred by cell surface BST-2/tetherin presents a challenge to the replication of enveloped viruses, a growing number of which now are known to encode specific anti-tetherin proteins (reviewed in reference 39). The diversity of anti-tetherin factors, in particular, those found in the primate lentiviruses, highlights the importance of this activity for virus replication and also suggests that anti-tetherin factors represent a novel target for drug development.
Tetherin has an unusual structure, containing two membrane anchors, and its ability to restrict virus release is dependent on both of these features (15, 32). Currently, two general models have been proposed to explain how tetherin could restrict enveloped virus release (39). In the membrane-spanning model, a single tetherin monomer could be simultaneously anchored in both the cell and virus membranes through its TM domain and GPI anchor, effectively bridging the cell-virus gap. Alternatively, the ectodomain interaction model evokes dimeric or higher-order interactions between tetherin monomers based on interactions between the protein's ectodomains. Both models require tetherin to be either partially or fully incorporated into virions and explain the requirement for the basic structural elements of tetherin (TM domain, coiled-coil ectodomain, and GPI anchor). Furthermore, the demonstration that restriction can be conferred by an artificial tetherin with no significant sequence homology to BST-2 suggests that tetherin acts directly and without cofactors, and that its target is the viral lipid envelope (32).
We investigated aspects of both tetherin restriction and its antagonism using a series of TfR-tetherin chimeric proteins and an ectodomain-deleted mini-tetherin protein. Each of the proteins we constructed were designed to maintain the membrane-targeting sequence of a type II integral membrane protein while enabling us to examine the requirements for the CT, TM, or ectodomain regions in both tetherin restriction and its antagonism by different viral proteins. The ectodomain of tetherin includes three cysteine residues involved in disulfide bond formation, two glycosylation sites, predicted coiled-coil sequences, and a C-terminal GPI membrane anchor. We observed that all three TfR-tetherin derivatives localized at the cell surface and were able to restrict the release of HIV-1 VLPs, confirming that specific tetherin residues in the CT or TM domain are not required for restriction, as previously reported (15, 32). In contrast, a mini-tetherin derivative containing only the first 50 amino acids of the protein recapitulated the cellular distribution of tetherin, being localized to both the cell surface and a perinuclear compartment, but was unable to confer restriction. Therefore, tetherin restriction can be conferred by the protein's ectodomain alone, if presented by the trafficking machinery of a type II integral membrane protein, while the CT and TM domains are sufficient to confer the authentic cellular distribution.
Given the variety of viral tetherin antagonists so far identified, it seems reasonable to suppose that different proteins could have evolved different ways to counteract tetherin. We tested this hypothesis by first mapping the domains of tetherin targeted by the anti-tetherin factors HIV-1 Vpu, HIV-2 Env, and Ebola virus GP. Since we anticipated that the loss of cell surface tetherin would be a correlate of anti-tetherin activity, we analyzed both their ability to counteract the restriction of VLP release conferred by the panel of active tetherin derivatives and their effects on the cellular distribution of these proteins. For Vpu, we found that the antagonism of tetherin restriction required the specific TM domain of tetherin, which is in agreement with previous studies that have mapped the Vpu-tetherin interaction to this region (10, 15, 16, 25, 34). The ability of Vpu also to redistribute the mini-tetherin protein, which does not itself restrict virus release, further demonstrated that the ectodomain and GPI anchor of the protein are not required for the Vpu-tetherin interaction. Finally, we observed a complete correlation between the ability of Vpu to counteract restriction conferred by a tetherin derivative and to remove it from the cell surface. Vpu's ability to overcome tetherin restriction therefore fits with a simple model of TM interactions leading to the removal of tetherin from the cell surface, possibly by promoting its internalization, intracellular retention, and/or degradation.
Similarly to Vpu, we also observed that full-length tetherin was removed from the cell surface and sequestered intracellularly by the expression of the HIV-2 Env. However, unlike Vpu, this interaction did not lead to tetherin degradation. A point mutation (A598T) in the ectodomain of HIV-2 Env blocks the protein's ability to stimulate virus release (1, 4), and we now show that a point mutation, A100D, in the ectodomain of human tetherin also can block this interaction. Together with the finding that HIV-2 Env and tetherin can be coimmunoprecipitated (20), these observations suggest a direct physical interaction between the two proteins. This could lead to the increased internalization of tetherin through the AP-2-dependent pathway that the HIV-2 Env accesses, since we have previously shown that both AP-2 and its partner tyrosine motif in the cytoplasmic tail of the HIV-2 Env are necessary to increase virus release (1, 30).
The final viral protein that we analyzed was the Ebola virus GP, where we confirmed a previous report that GP counteracts the tetherin restriction of HIV-1 VLP release (18). When GP activity against the TfR-tetherin chimeras also was examined, we found that it antagonized all three of the proteins effectively, and that the resulting recovery of VLP release was more efficient than that achieved against the native tetherin. Interestingly, neither tetherin nor the TfR chimeras were significantly downregulated from the cell surface by GP, suggesting that Ebola virus GP uses a more indirect mechanism to counteract tetherin restriction. This was further supported by the surprising finding that GP also was active against an artificial tetherin, which was resistant to the action of both HIV-1 Vpu and HIV-2 Env.
How could Ebola virus GP block tetherin restriction without affecting the cell surface level of the protein and without targeting specific tetherin sequences? Previously, it has been reported that GP's anti-tetherin activity requires both its membrane anchor and proper trafficking to the cell surface (18), but that it does not require the mucin domain of the protein that has been implicated in the downregulation of a number of other cell surface proteins (37). GP also was reported to coimmunoprecipitate with tetherin, although the association only occurred with a less mature form of GP (18), so that a direct interaction between the mature forms of the proteins at the cell surface may not occur.
In the absence of specific sequence requirements, it is possible that some other feature of tetherin is targeted by GP, for example, an association with lipid rafts. Ebola virus GP is itself located in lipid rafts, and its presence in these membrane regions could in some way alter virus assembly sites, perhaps creating an environment that is incompatible with tetherin trafficking to, or retention at, such sites. Additionally, GP could act as a nonspecific steric block to physically prevent tetherin from accessing the viral lipid envelope or forming the higher-order multimers that have been proposed as alternate mechanisms to account for tetherin restriction (39). If the action of GP impacted only a relatively small fraction of tetherin at a virus-budding site or caused tetherin surface redistribution rather than large-scale removal, then this could account for the lack of the effect on overall tetherin cell surface levels that we observed. Further experiments will be needed to distinguish between these different possibilities.
This work was funded by NIH grants R01 AI068546 to P.M.C. and 5 R01 AI068546-S1 (Research Supplements to Promote Diversity in Health-Related Research Program) to L.A.L. and CHRP grant 1008-CHLA-038 to P.M.C.
Published ahead of print on 5 May 2010.