ABSTRACT
BST-2/CD317/tetherin is a host transmembrane protein that potently inhibits human immunodeficiency virus type 1 (HIV-1) virion release by tethering the nascent virions to the plasma membrane. Viral protein U (Vpu) is an accessory protein encoded by HIV-1 as well as by some simian immunodeficiency viruses (SIVs) infecting wild chimpanzees, gorillas, or monkeys (SIVcpz, SIVgor, or SIVgsn/SIVmon/SIVmus, respectively). HIV-1 Vpu directly binds to and downregulates human BST-2. The antagonism is highly species specific because the amino acid sequences of BST-2 are different among animal species. Here, we show that Vpu proteins from several SIVcpz, SIVgsn, SIVmon, or SIVmus isolates fail to antagonize human BST-2. Only Vpu from an SIVgsn isolate (SIVgsn-99CM71 [SIVgsn71]) was able to antagonize human BST-2 as well as BST-2 of its natural host, greater spot-nosed monkey (GSN). This SIVgsn Vpu interacted with human BST-2, downregulated cell surface human BST-2 expression, and facilitated HIV-1 virion release in the presence of human BST-2. While the unique 14AxxxxxxxW22 motif in the transmembrane domain of HIV-1NL4-3Vpu was reported to be important for antagonizing human BST-2, we show here that two AxxxxxxxW motifs (A22W30 and A25W33) exist in SIVgsn71 Vpu. Only the A22W30 motif was needed for SIVgsn71 Vpu to antagonize GSN BST-2, suggesting that the mechanism of this antagonism resembles that of HIV-1NL4-3 Vpu against human BST-2. Interestingly, SIVgsn71 Vpu requires two AxxxxxxxW (A22W30 and A25W33) motifs to antagonize human BST-2, suggesting an as-yet-undefined way that SIVgsn71 Vpu works against human BST-2. These results imply an evolutionary impact of primate BST-2 on lentiviral Vpu.
IMPORTANCE Genetic alterations conferring a selective advantage in protecting from life-threating pathogens are maintained during evolution. In fact, the amino acid sequences of BST-2 differ among primate animals and their susceptibility to viral proteins is species specific, suggesting that such genetic diversity has arisen through the evolutionarily controlled balance between the host and pathogens. The M (main) group of HIV-1 is thought to be derived from SIVcpz, which utilizes Nef, but not Vpu, to antagonize chimpanzee BST-2. SIVcpz Nef is, however, unable to antagonize human BST-2, and Vpu was consequently chosen again as an antagonist against human BST-2 in the context of HIV-1. Studies on how Vpu lost and acquired this ability, together with the distinct mechanisms by which SIVgsn71 Vpu binds to and downregulates human or GSN BST-2, may help to explain the evolution of this lentiviral protein as a result of host-pathogen interactions.
INTRODUCTION
Animals (including primates) have a complex immune system that protects the host from viral and microbial infections, including innate immunity (1), acquired immunity (2), and intrinsic immune (3). Host proteins referred to as “restriction factors” constitute the initial line of the defense system and suppress virus replication in a cell-autonomous manner (3–5). However, many viruses have strategies to overcome these host defense mechanisms. In fact, HIV-1 has four accessory genes (vpu, vif, vpr, and nef) that function to antagonize host restriction factors.
BST-2, also known as tetherin, CD317, and HM1.24, is an interferon-inducible type II transmembrane (TM) glycoprotein and has been identified as a restriction factor inhibiting release of vpu-deleted HIV-1 (6, 7). Viral protein U (Vpu), an HIV-1 accessory protein encoded by the vpu gene, has the ability to counteract BST-2. Vpu is a type I transmembrane protein consisting of 77 to 86 amino acid residues. In the absence of Vpu, BST-2 tethers nascent virions at the surface of infected cells (6, 7). HIV-1 Vpu physically interacts with human BST-2 via its transmembrane (TM) domains (8–17) and downregulates human BST-2 from the plasma membrane (16–22). A mutational analysis predicted that residues A14, W22, and A18 in the TM domain of HIV-1 NL4-3 Vpu form one face of the Vpu TM alpha helix and mediate its binding to human BST-2 (14). Mutation of A14 and/or W22 ablated the interaction of Vpu with human BST-2 (12, 14), while A18 was reported to be important for enhancement of virus release (14).
HIV-1 group M (main), which is responsible for the ongoing AIDS pandemic, is thought to have been derived from lentiviruses infecting wild chimpanzees in Africa (SIVcpz) (23, 24), and SIVcpz is thought to have arisen from recombination of two simian immunodeficiency viruses (SIVs) (25). Aside from HIV-1 and its precursor SIVcpz (26), the vpu gene is found in SIVgor from gorillas (27), SIVgsn from greater spot-nosed (GSN) monkey (28), SIVmon from mona monkey (29, 30), SIVmus from mustached monkey (29), and SIVden from Cercopithecus denti (31). Since SIVden was isolated from a pet monkey, the vpu gene in SIVcpz is more likely to have originated from SIVgsn, SIVmon, or SIVmus isolated from wild monkeys in Africa (32).
The BST-2 antagonism by Vpu of primate lentiviruses appears to be species specific (32–39). Most Vpu proteins from HIV-1 group M can antagonize human BST-2 but not monkey BST-2 (9, 32–40), except for Vpu from clinical HIV-1 isolates, which are capable of antagonizing macaque BST-2 (41). On the other hand, Vpu from SIVgsn, SIVmon, or SIVmus was reported to be able to antagonize monkey BST-2s but not human BST-2 (32). Interestingly, SIVcpz employs Nef to antagonize chimpanzee BST-2 albeit the vpu gene exists in its genome (32, 42, 43). Indeed, SIVcpz Vpu is unable to antagonize chimpanzee or human BST-2 but is able to induce CD4 degradation (32, 43). In addition, Vpu from SIVcpzMB897, which is thought to be phylogenetically close to HIV-1 group M, cannot bind to chimpanzee BST-2 (8) or human BST-2 (44). It was reported previously that W22 is invariant in HIV-1/SIVcpz Vpu proteins whereas A14 and A18 are highly conserved among Vpu alleles from HIV-1 groups M and N but not among those from group O or SIVcpz, which lack the ability to antagonize human BST-2 (14, 45). This prompted us to investigate how Vpu of HIV-1 group M acquired the ability to bind to and downregulate human BST-2 and whether any other SIV Vpu has such activities.
In this study, we found that of the Vpu proteins tested, only SIVgsn-99CM71 (SIVgsn71) Vpu was able to antagonize human BST-2 efficiently and that it enhanced virus release in the presence of human BST-2. Two AxxxxxxxW motifs (A22W30 and A25W33) of SIVgsn71 Vpu were required for this antagonism, whereas only the A22W30 motif was needed to antagonize BST-2 of its natural host, greater spot-nosed monkey.
RESULTS
SIVgsn71 Vpu is able to interact with human BST-2.Based on classical phylogenetic analyses, HIV-1 is assumed to be derived from SIVcpz (26), which was thought to represent a hybrid of SIVs (25). Among those simian immunodeficiency viruses (SIVs), SIVgsn, SIVmon, SIVmus, and SIVcpz carry the vpu gene (26, 28–30) and HIV-1Vpu is thought to have originated from them. To study the relationship between these SIV Vpu proteins and human BST-2, we evaluated their ability to physically interact by employing a bimolecular fluorescence complementation (BiFC) assay. The BiFC assay is based on the formation of the reporter protein monomeric Kusabira-Green (mKG), which is divided into two nonfluorescent fragments (mKGN and mKGC) (46). These mKGC and mKGN fragments were fused to the human or monkey BST-2 and Vpu, respectively (Fig. 1A), and the mKG fragments were able to form an active structure and emit fluorescence only when these proteins were bound to each other as shown previously (8, 9, 44). The fluorescence of reconstituted mKG proteins in live cells was measured by flow cytometry, and the geometric mean (Geo Mean) of fluorescence represents the relative binding efficiency. We analyzed a set of 8 vpu genes from viruses that included HIV-1 (group M), SIVcpz, and other SIVs (SIVgsn, SIVmon, and SIVmus). An amino acid alignment is shown in Fig. 1B. Interestingly, SIVgsn Vpu proteins have two sets of an AxxxxxxxW motif, which was previously shown to be essential for the interaction of HIV-1 NL4-3 Vpu with human BST-2 (Fig. 1B; highlighted). We evaluated if these SIV Vpu proteins can bind to the natural host BST-2 by BiFC assay. KGN-tagged Vpu proteins from SIVgsn71, SIVgsn-99CM166 (SIVgsn166), SIVmon-99CMCML1 (SIVmonCML1), or SIVmus-01CM1085 (SIVmus1085) produced higher BiFC signals with KGC-tagged greater spot-nosed monkey (GSN), mona monkey (MON), or mustached monkey (MUS) BST-2 proteins than the corresponding negative-control KGN tags (Fig. 1C to E). A mutant of HIV-1 NL4-3 Vpu (HIV-1 NL4-3 Vpu14/22), which has A14L and W22A mutations in the transmembrane (TM) domain, was reported to be unable to bind to human BST-2 (12, 14) and therefore was used as a negative control. Expression of Vpu proteins and human BST-2 was verified by Western blotting (Fig. 1F and G, bottom). The BiFC signals were much weaker for the HIV-1 NL4-3 Vpu14/22 mutant or the KGN tag than for HIV-1 NL4-3 Vpu (Fig. 1F, top). Low levels of BiFC signals were also observed for SIVcpz Vpu proteins from the MB897, MB66, and LB7 strains (Fig. 1F, top), even though the expression levels of both Vpu and human BST-2 in cells expressing these SIVcpz Vpu proteins were higher than in cells expressing HIV-1 NL4-3 Vpu (Fig. 1F, bottom). This is consistent with previous reports that SIVcpz Vpu failed to antagonize human BST-2 (32, 43). Vpu from SIVmus1085 induced somewhat higher BiFC signals with human BST-2 than HIV-1 NL4-3 Vpu14/22. In contrast, Vpu proteins from SIVgsn166 and SIVmonCML1 showed BiFC signals similar to those seen with HIV-1 NL4-3 Vpu14/22 (Fig. 1G, top). Notably, SIVgsn71 Vpu showed much higher levels of BiFC signal than the other SIV Vpu proteins (Fig. 1G, top). To confirm this result, we carried out coimmunoprecipitation analyses and found that HIV-1 NL4-3 Vpu and SIVgsn71 Vpu but not HIV-1 NL4-3 Vpu14/22 physically interacted with human BST-2 (Fig. 1H). These results indicate that Vpu from SIVgsn71 is able to bind to human BST-2.
SIVgsn71 Vpu is able to interact with human BST-2. (A) A schematic diagram of the KGN-tagged Vpu and KGC-tagged BST-2 proteins used in this study is shown. (B) An amino acid alignment of the HIV and SIV Vpu proteins used in this study is shown. The amino acids constituting the AxxxxxxxW motifs are highlighted. (C to G) The physical interaction between GSN BST-2 (C), MON BST-2 (D), MUS BST-2 (E), or human BST-2 (F and G) and SIV Vpu proteins was assessed by the BiFC assay. (C to E) HEK293T cells were transfected with pKGC-GSN BST-2 (C), pKGC-MON BST-2 (D), or pKGC-MUS BST-2 (E) together with a plasmid expressing KGN-tagged Vpu of SIVs isolated from the same species. (F and G) HEK293T cells expressing KGC-tagged human BST-2 together with KGN-tagged HIV-1 NL4-3 Vpu, HIV-1 NL4-3 Vpu14/22, or SIVcpz Vpu proteins (F) or SIVgsn71 Vpu, SIVgsn166 Vpu, SIVmonCML1 Vpu, or SIVmus1085 Vpu (G) were harvested 48 h after transfection. Cells were subjected to flow cytometry and Western blotting. The KGN tag represents the amino-terminal part of monomeric Kusabira-Green (mKG) expressed by the vector phmKGN-MN and serves as a negative control here. (Top) The geometric mean (Geo Mean) value of BiFC signal in cells expressing human BST-2 and HIV-1 NL4-3 Vpu (column 2) was arbitrarily set as 100 (F and G). Data represent means of results from four independent experiments ± SEM, and statistical significance compared to the BiFC signal of the KGN tag (C to E) or HIV-1 NL4-3 Vpu14/22 (F and G) was determined. (***, P < 0.001; **, P < 0.01; *, P < 0.05; n.s., not significant [P > 0.05]). (Bottom) The levels of expression of human BST-2 and HIV-1 or SIV Vpu proteins were assessed with anti-KGC or anti-KGN antibodies, respectively. α-Tubulin was detected as a loading control with the same membranes. The arrow indicates nonspecific bands. One representative Western blotting result is shown. (H) 293T cells expressing FLAG-tagged human BST-2 and the indicated HA-tagged Vpu proteins were harvested 48 h after transfection. Vpu proteins were immunoprecipitated with an anti-HA antibody from the cell lysate with protein A agarose beads. The immunoprecipitates (top) and cell lysates (bottom) were subjected to Western blotting and analyzed with anti-FLAG and HA antibodies.
SIVgsn71 Vpu is able to downregulate cell surface human BST-2.To further assess if these SIV Vpu proteins downregulate human BST-2, we generated an HIV-1 proviral DNA pNL4-3EGFP (enhanced green fluorescent protein) Δenv Δnef strain (47) whose vpu gene was replaced by SIV vpu genes. A proviral DNA pNL4-3EGFP Δenv Δnef Δvpu (Udel) strain was used as a negative control which cannot express the Vpu protein. To confirm whether SIV vpu genes inserted into the HIV proviral DNA were functionally expressed in infected cells, we verified downregulation of monkey BST-2s by SIV Vpu proteins. For this purpose, we first knocked out (KO) endogenous human BST-2 in HeLa cells using the CRISPR/Cas9 system followed by constitutive expression of various monkey BST-2s (Fig. 2A). HeLa cells expressing GSN, MON, or MUS BST-2 were then infected with the recombinant viruses carrying vpu from SIVgsn, SIVmon, or SIVmus, respectively. Cell surface or intracellular expression of monkey BST-2 was measured by flow cytometry (Fig. 2B to D), and EGFP-derived fluorescence was monitored as an infection marker (x axis). Infected cells (EGFP positive; right gate [R3] in Fig. 2B) expressing SIVgsn71 Vpu or SIVgsn166 Vpu, but not HIV-1 NL4-3 Vpu or Udel (a Δvpu mutant), showed lower cell surface GSN BST-2 expression than uninfected cells (EGFP negative; left gate [R2] in Fig. 2B), indicating specific downregulation of cell surface GSN BST-2 by SIVgsn Vpu proteins (Fig. 2B, top). In contrast, in cells infected with virus expressing HIV-1 NL4-3 Vpu or Udel, both intracellular and cell surface levels of GSN BST-2 were increased, whereas intracellular levels of GSN BST-2 in cells infected with virus expressing SIVgsn71 Vpu or SIVgsn166 Vpu were unaffected (Fig. 2B, bottom). Similarly, SIVmonCML1 Vpu downregulated cell surface expression of MON BST-2 and SIVmus1085 Vpu downregulated cell surface expression of MUS BST-2 (Fig. 2C and D). Thus, we validated the hypothesis that all SIV Vpu proteins were functionally expressed in infected cells because they successfully downregulated cell surface expression of natural host BST-2. We next infected HeLa cells with these viruses to assess whether SIV Vpu proteins would downregulate cell surface human BST-2 (Fig. 3, top). As expected, endogenous human BST-2 was downmodulated from the cell surface by NL4-3EGFP Δenv Δnef capable of expressing HIV-1 NL4-3 Vpu but not by virus lacking Vpu expression (Udel). In contrast, Vpu from SIVcpzMB897 failed to downregulate human BST-2 at the cell surface (Fig. 3, top), which is consistent with a previous report showing that SIVcpzMB897 Vpu cannot antagonize human BST-2 (32). Neither Vpu from SIVgsn166 nor Vpu from SIVmonCML1 or SIVmus1085 showed downregulation of cell surface human BST-2 (Fig. 3, top). Surprisingly, SIVgsn71 Vpu downregulated cell surface human BST-2 in HeLa cells as efficiently as HIV-1 NL4-3 Vpu (Fig. 3, top). This effect was evident also in human T-cell line MT4 (data not shown). Of note, cells expressing HIV-1 NL4-3 Vpu or SIVgsn71 Vpu had no effect on the intracellular expression level of human BST-2 (Fig. 3, bottom). Interestingly, however, intracellular BST-2 levels were significantly elevated in cells infected by viruses expressing Vpu proteins that were unable to downregulate cell surface human BST-2 (i.e., SIVcpzMB897, SIVgsn166, SIVmonCML1, and SIVmus1085). From these results, we conclude that SIVgsn71 Vpu is able to downregulate cell surface expression of endogenous human BST-2.
Vpu of SIVgsn, SIVmon, or SIVmus specifically downregulates BST-2 of each natural host monkey. (A) Cell surface expression of BST-2 in HeLa, HeLa BST-2 KO, and HeLa BST-2 KO cells which stably express KGC-tagged GSN, MON, or MUS BST-2 was monitored by flow cytometry with anti-BST-2 antibody. The x axis data represent cell surface BST-2 expression. ctrl, control. (B, C, and D) HeLa KO-GSN BST-2 cells (B), HeLa KO-MON BST-2 cells (C), and HeLa KO-MUS BST-2 cells (D) were infected with a VSV-G-pseudotyped HIV-1 isolate whose vpu gene was intact (HIV-1 NL4-3 Vpu), deleted (Udel), or replaced with that of SIVs isolated from each monkey species and were subjected to flow cytometry to detect cell surface (top) and intracellular (bottom) expression of monkey BST-2 with anti-BST-2 antibody. In the bar graphs, the Geo Mean fluorescence intensities of BST-2 (y axis) in EGFP-negative isolates (uninfected, left gate) and EGFP-positive isolates (infected, right gate) are shown as means of results from three independent experiments ± SEM, and statistical significance compared to the Geo Mean of EGFP-negative cells was determined. The Geo Mean of EGFP-negative cells in each sample was arbitrarily set as 100. One representative dot plot is shown. neg ctrl, negative control.
SIVgsn71 Vpu is able to downregulate cell surface expression of human BST-2. HeLa cells were infected with a VSV-G-pseudotyped HIV-1 isolate whose vpu gene was intact (HIV-1 NL4-3 Vpu), deleted (Udel), or replaced with that of SIVcpz, SIVgsn, SIVmon or SIVmus. Cells were stained with anti-BST-2 antibody and subjected to flow cytometry. The cell surface (top) and intracellular (bottom) human BST-2 were detected separately. Data in graphs are presented as described for Fig. 2B. One representative dot plot is shown.
SIVgsn71 Vpu antagonizes human BST-2 via its transmembrane domain.HIV-1 Vpu consists of an N-terminal short extracellular domain, a TM domain, and a C-terminal cytoplasmic domain (CTD) (48). Human BST-2 and HIV-1 Vpu are reported to interact via their TM domains (8–17). To understand whether the TM domain in SIVgsn71 Vpu is responsible for the antagonism, we constructed a mutant SIVgsn71 Vpu (SIVgsn71 Vpu-TMRD; Fig. 4A) whose TM domain was replaced by the TM domain of the NL4-3 Vpu-TMRD mutant, which cannot interact with human BST-2 (8, 9, 15, 17). After verifying the expression of all Vpu proteins and BST-2, the levels of BiFC signals were measured by flow cytometry (Fig. 4B). The BiFC signal for SIVgsn71 Vpu-TMRD with human BST-2 was significantly lower than that for SIVgsn71 Vpu, suggesting that SIVgsn71 Vpu requires its TM domain for binding to human BST-2. In addition, this mutation impaired BST-2 downregulation, as the level of cell surface expression of human BST-2 on SIVgsn71 Vpu-TMRD-expressing (EGFP-positive) cells was similar to that seen with uninfected (EGFP-negative) cells (Fig. 4C, top). Intracellular expression of human BST-2 was enhanced in cells infected with Udel or SIVgsn71 Vpu-TMRD-expressing virus (Fig. 4C, bottom). To further investigate whether SIVgsn71 Vpu enhances virus release in the presence of human BST-2, we transfected HeLa cells or those whose bst-2 genes were knocked out (HeLa BST-2 KO cells) with proviral DNA (pNL4-3EGFP Δenv Δnef carrying different vpu genes) together with a plasmid expressing vesicular stomatitis virus G (VSV-G) (pMISSION-VSV-G). We set the experimental condition such that transfections of HeLa BST-2 KO cells with each recombinant construct achieved similar luciferase activities in TZM-bl reporter cells (Fig. 4D) or CAp24 in the culture supernatant (Fig. 4E). This may explain the fluctuations in intracellular levels of Pr55gag and CAp24 in the producer cells (Fig. 4D and E, bottom). We then evaluated the release of infectious virus from these cells using TZM-bl reporter cells which express luciferase when they are infected with HIV-1 vector. While we found similar levels of infectious virus in all culture supernatants from HeLa BST-2 KO cells (Fig. 4D, right columns), HIV-1 NL4-3 Vpu and SIVgsn71 Vpu, but not SIVgsn71 Vpu-TMRD, enhanced release of infectious virus from HeLa cells which express endogenous human BST-2 (Fig. 4D, left columns). Moreover, we also quantified the amounts of virus released into the culture supernatants by HIV-1 p24 enzyme-linked immunosorbent assay (ELISA) and obtained results comparable to the TZM-bl assay results (Fig. 4D). We found that HIV-1 NL4-3 Vpu and SIVgsn71 Vpu enhanced virus release in the presence of human BST-2 (Fig. 4E, left columns) under conditions where these viruses produce similar amounts of p24 in the absence of human BST-2 (HeLa BST-2 KO cells; Fig. 4E, right columns). Our results indicate that SIVgsn71 Vpu, like HIV-1 Vpu, interacts with and downregulates human BST-2 to facilitate virus release in a TM domain-dependent manner, although less efficiently than HIV-1 Vpu.
SIVgsn71 Vpu antagonizes human BST-2 via its transmembrane domain. (A) Schematic diagram of SIVgsn71 Vpu (white box) and its TMRD mutant (gray box represents the TM domain of NL4-3 Vpu-TMRD; white box represents the cytoplasmic domain of the SIVgsn71 Vpu). (B) HEK293T cells expressing KGC-tagged human BST-2 together with KGN-tagged SIVgsn71 Vpu or its TMRD mutant were harvested and subjected to flow cytometry (top) and Western blotting (bottom) as described for Fig. 1B. The Geo Mean value of BiFC signal in cells expressing human BST-2 and SIVgsn71 Vpu (the 2nd column) was arbitrarily set as 100. Data represent means of results from three independent experiments ± SEM, and statistical significance compared to the BiFC signal of SIVgsn71 Vpu was determined. The arrow indicates nonspecific bands. One representative Western blotting result is shown. (C) HeLa cells were infected with a VSV-G-pseudotyped HIV-1 isolate whose vpu gene was intact (HIV-1 NL4-3 Vpu), deleted (Udel), or replaced with that of SIVgsn71 or SIVgsn71 Vpu-TMRD and was subjected to flow cytometry. Data in graphs are presented as described for Fig. 2B. (D) Infectious virus release in the presence or absence of human BST-2 was assessed by TZM-bl assay. Viruses were prepared by transfecting HeLa or HeLa BST-2 KO cells with HIV-1 env-deficient proviral DNA which carried the vpu gene from HIV-1 NL4-3 or SIVgsn71 together with pMISSION-VSV-G. TZM-bl cells were infected with released virus, and relative luciferase activity levels were measured (top). The level of luciferase activity caused by infection with viruses released from HeLa cells was arbitrarily set as 100. Data represent means of results from three independent experiments ± SEM, and statistical significance was determined. Producer cells were harvested for Western blotting (bottom) to monitor Gag protein expression with anti-p24 antibody. α-Tubulin was detected as a loading control with the same membrane. One representative Western blotting result is shown. (E) The amounts of virus released into the supernatants in the presence (HeLa cells) or absence (HeLa BST-2 KO cells) of human BST-2 were assessed by HIV-1 p24 ELISA. Viruses were prepared as described for panel D. The amounts of p24 released from producer cells are shown in the bar graph (top). Producer cells were harvested for Western blotting (bottom) with anti-p24 and anti-α-tubulin antibodies.
SIVgsn71 Vpu antagonizes human BST-2 through two AxxxxxxxW motifs.Previous reports have shown that an AxxxxxxxW motif in HIV-1 NL4-3 Vpu is essential for the binding to human BST-2 (12, 14). In SIVgsn71 Vpu, we found two AxxxxxxxW motifs (A22W30 and A25W33) and mutated them to LxxxxxxxA separately (SIVgsn71 Vpu-m1 and SIVgsn71Vpu-m2, respectively; Fig. 5A). The expression of these mutants were verified by Western blotting, and the BiFC signal was measured by flow cytometry to assess the effect of these mutations on the binding to human BST-2. SIVgsn71 Vpu-m1 and Vpu-m2 showed levels of BiFC signals significantly lower than those seen with SIVgsn71 Vpu but similar to those seen with SIVgsn71 Vpu-TMRD (Fig. 5B). We next evaluated the cell surface expression of human BST-2 in HeLa cells infected with virus expressing these SIVgsn71Vpu mutants. Neither SIVgsn71 Vpu-m1 nor Vpu-m2 downregulated human BST-2 whereas wild-type SIVgsn71 Vpu did (Fig. 5C). Similar results were obtained in MT4 cells infected with NL4-3EGFP expressing SIVgsn71 Vpu or its mutants (data not shown). To further assess the ability of these SIVgsn71 Vpu mutants to facilitate virus release in the presence of human BST-2, we transfected HeLa and HeLa BST-2 KO cells with proviral DNA (pNL4-3EGFP Δenv Δnef) carrying SIVgsn71 vpu or mutants together with pMISSION-VSV-G. Reporter cell (TZM-bl) assays revealed that all three SIVgsn71 Vpu mutants (TMRD, m1, and m2) failed to facilitate infectious virus release in the presence of human BST-2 (HeLa cells), whereas efficient release of infectious virus was observed in the absence of human BST-2 (HeLa BST-2 KO cells; Fig. 5D). These results indicate that the two AxxxxxxxW motifs play important roles in the human BST-2 antagonism by SIVgsn71 Vpu, albeit only one AxxxxxxxW motif in NL4-3 Vpu is important for the human BST-2 antagonism by NL4-3 Vpu (12, 14).
SIVgsn71 Vpu antagonizes human BST-2 through two AxxxxxxxW motifs. (A) Schematic diagram of HIV-1 NL4-3 Vpu (light gray box; the AxxxxxxxW motif is shown in bold), SIVgsn71 Vpu (white box; the two AxxxxxxxW motifs are shown with closed and open circles) and its two AxxxxxxxW motif mutants (closed and open circles indicate the location of mutations A22LW30A and A25LW33A, respectively). (B) HEK293T cells expressing KGC-tagged human BST-2 together with KGN-tagged SIVgsn71 Vpu or its AxxxxxxxW motif mutants were harvested and subjected to flow cytometry (top) and Western blotting (bottom) as described for Fig. 1B. The Geo Mean value of BiFC signal in cells expressing human BST-2 and SIVgsn71 Vpu (column 2) was arbitrarily set as 100. Data represent means of results from three independent experiments ± SEM, and statistical significance compared to the BiFC signal of SIVgsn71 Vpu was determined. The arrow indicates nonspecific bands. One representative Western blotting result is shown. (C) HeLa cells were infected with a VSV-G-pseudotyped HIV-1 isolate whose vpu was replaced with SIVgsn71, SIVgsn71 Vpu-TMRD, SIVgsn71 Vpu-m1, or SIVgsn71 Vpu-m2 and subjected to flow cytometry. Data in graphs are presented as described for Fig. 2B. One representative dot plot is shown. (D) Infectious virus release in the presence (HeLa cells) or absence (HeLa BST-2 KO cells) of human BST-2 was assessed with TZM-bl assay as described for Fig. 4D (top). The level of luciferase activity of TZM-bl cells infected with virus expressing SIVgsn71 Vpu released from HeLa cells was defined as 100. Data represent means of results from three independent experiments ± SEM, and statistical significance was determined. Producer cells were harvested for Western blotting with anti-p24 or anti-α-tubulin antibodies (bottom). One representative Western blotting result is shown.
SIVgsn71 Vpu antagonizes GSN BST-2 through one AxxxxxxxW motif.To assess whether SIVgsn71 Vpu antagonizes BST-2 of its natural host (GSN BST-2) and whether this depends on the two AxxxxxxxW motifs, we examined the binding of SIVgsn71 Vpu mutants with GSN BST-2 by BiFC assay. The levels of expression of these mutants were confirmed by Western blotting, and the levels of the BiFC signals were measured by flow cytometry (Fig. 6A). As expected, SIVgsn71 Vpu interacted with GSN BST-2 and Vpu-TMRD did not show efficient interaction (Fig. 6A). Interestingly, SIVgsn71 Vpu-m1 showed significantly lower binding activity than the wild-type strain whereas the level seen with SIVgsn71 Vpu-m2 was similar to that of the wild type (Fig. 6A). We also evaluated downregulation of GSN BST-2 by these two mutants using HeLa cells lacking human BST-2 and expressing GSN BST-2 (HeLa KO-GSN BST-2 cells; see Fig. 2A). SIVgsn71 Vpu but not SIVgsn71 Vpu-TMRD was able to downregulate cell surface GSN BST-2 (Fig. 6B). SIVgsn71 Vpu-m1 did not downregulate cell surface GSN BST-2, while SIVgsn71 Vpu-m2 was as efficient as SIVgsn71 Vpu in downregulating BST-2 (Fig. 6B). Next, we examined the effect of SIVgsn71 mutants on virus release in the presence or absence of GSN BST-2 using HeLa KO-GSN BST-2 or HeLa BST-2 KO cells, respectively. Results of a reporter cell assay showed that SIVgsn71 Vpu and Vpu-m2 facilitated release of infectious virus in the presence of GSN BST-2 in HeLa KO-GSN BST-2 cells, whereas neither SIVgsn71 Vpu-TMRD nor Vpu-m1 promoted this process (Fig. 6C). We also carried out CAp24 ELISA to assess the amount of virus released into the culture supernatants (Fig. 6D) and confirmed that both Vpu-m1 and Vpu-m2 failed to promote virus release in the presence of human BST-2 and that Vpu-m1 but not Vpu-m2 had lost the activity against GSN BST-2. Taken together, these results indicate that only the first AxxxxxxxW motif (A22W30) is required for GSN BST-2 antagonism by SIVgsn71 Vpu.
SIVgsn71 Vpu antagonizes GSN BST-2 via one AxxxxxxxW motif. (A) HEK293T cells expressing KGC-tagged GSN BST-2 together with KGN-tagged SIVgsn71 Vpu or its AxxxxxxxW motif mutants were harvested and subjected to flow cytometry (top) and Western blotting (bottom) as described for Fig. 1B. The Geo Mean value of BiFC signal in cells expressing GSN BST-2 and SIVgsn71 Vpu (column 2) was arbitrarily set as 100. Data represent means of results from seven independent experiments ± SEM, and statistical significance compared to the BiFC signal of SIVgsn71 Vpu was determined. The arrow indicates nonspecific bands. One representative Western blotting result is shown. (B) HeLa BST-2 KO cells expressing GSN BST-2 (HeLa KO-GSN BST-2 cells) were infected with a VSV-G-pseudotyped HIV-1 isolate whose Vpu was replaced with SIVgsn71, SIVgsn71 Vpu-TMRD, SIVgsn71 Vpu-m1, or SIVgsn71 Vpu-m2 and subjected to flow cytometry. Data in graphs are presented as described for Fig. 2B. One representative dot plot is shown. (C) Infectious virus release in the presence (HeLa KO-GSN BST-2 cells) or absence (HeLa BST-2 KO cells) of GSN BST-2 was assessed with TZM-bl assay as described for Fig. 4D (top). The level of luciferase activity of TZM-bl cells infected with virus expressing SIVgsn71 Vpu released from HeLa KO-GSN BST-2 cells was defined as 100. Data represent means of results from three independent experiments ± SEM. Producer cells were harvested for Western blotting with anti-p24 or anti-α-tubulin antibodies (bottom). One representative Western blotting result is shown. (D) The amounts of virus released into the supernatants in the absence (HeLa BST-2 KO cells) or presence of human (HeLa cells) or GSN BST-2 (HeLa KO-GSN BST-2 cells) were assessed by HIV-1 p24 ELISA as described for Fig. 4E. The amounts of p24 released from producer cells are shown (top). Producer cells were harvested for Western blotting (bottom) with anti-p24 and anti-α-tubulin antibodies.
SIVgsn166 Vpu is unable to antagonize human BST-2.SIVgsn71 and SIVgsn166 are two strains isolated from GSN monkeys (28). Interestingly, even though Vpu proteins from SIVgsn71 and SIVgsn166 have two sets of the AxxxxxxxW motif (Fig. 1B), they displayed contrasting phenotypes of human BST-2 downregulation (Fig. 3, top). The enhancement of virus release by SIVgsn166 Vpu was assessed in HeLa cells and HeLa BST-2 KO cells by TZM-bl reporter assay (Fig. 7A) and HIV-1 p24 ELISA (Fig. 7B). SIVgsn166 Vpu failed to enhance virus release in the presence of human BST-2, which corresponds to its lack of binding and its downregulation (see Fig. 1G and Fig. 3, respectively). Eight amino acids in the TM domain are different between SIVgsn71 Vpu and SIVgsn166 Vpu (Fig. 1B). To investigate whether this difference can explain their contrasting activities against human BST-2, SIVgsn71 Vpu and SIVgsn166 Vpu chimeras consisting of the TM domain and CTD of different origins (SIVgsn Vpu 71TM-166CTD and 166TM-71CTD) were constructed (Fig. 7C). Cell surface human BST-2 expression was evaluated in HeLa cells infected with viruses expressing SIVgsn71 Vpu, SIVgsn166 Vpu, or the chimeras. SIVgsn Vpu 71TM-166CTD was able to downregulate cell surface human BST-2 similarly to SIVgsn71 Vpu. In contrast, SIVgsn Vpu 166TM-71CTD and SIVgsn166 Vpu were unable to downmodulate human BST-2 (Fig. 7D). These results reinforce the notion that the TM domain of SIVgsn71 Vpu is important for antagonism of human BST-2. As two of the eight different amino acid residues are located at positions 23 and 24 in the first AxxxxxxxW motif (A22 to W30) but not in the second one (A25 to W33; Fig. 1B), mutants of SIVgsn71 Vpu (SIVgsn71 Vpu IF/LL) and SIVgsn166 Vpu (SIVgsn166 LL/IF) were constructed to address the importance of these two amino acid residues (Fig. 7C). Both SIVgsn71 Vpu IF/LL and SIVgsn166 LL/IF showed downregulating activity similar to that of the wild-type Vpu proteins (SIVgsn71 Vpu and SIVgsn166 Vpu, respectively) (Fig. 7D). These results collectively indicate that the structure of the TM domain of SIVgsn Vpu is the determinant for the antagonism of human BST-2 and that the failure of SIVgsn166 Vpu to antagonize human BST-2 cannot be attributed solely to the two amino acid difference in the 22AxxxxxxxW30 motif.
SIVgsn166 Vpu is unable to antagonize human BST-2. (A) Infectious virus release in the presence (HeLa cells) or absence (HeLa BST-2 KO cells) of human BST-2 was assessed with TZM-bl assay as described for Fig. 4D (top). The level of luciferase activity of TZM-bl cells infected with virus expressing HIV-1 NL4-3 Vpu released from HeLa cells was defined as 100 (top). Data represent means of results from three independent experiments ± SEM, and statistical significance was determined. Producer cells were harvested for Western blotting with anti-p24 or anti-α-tubulin antibodies (bottom). One representative Western blotting result is shown. (B) The amounts of virus released into the supernatants in the presence (HeLa cells) or absence (HeLa BST-2 KO cells) of human BST-2 were determined by HIV-1 p24 ELISA as described for Fig. 4E. The amounts of p24 released from producer cells are shown (top). Producer cells were harvested for Western blotting (bottom) with anti-p24 and anti-α-tubulin antibodies. (C) Schematic diagram of SIVgsn71 Vpu (white box; amino acid sequence of the AxxxxxxxW motifs is shown), SIVgsn166 Vpu (gray box; amino acid sequence of the AxxxxxxxW motifs is shown), SIVgsnVpu chimeras (SIVgsnVpu 71TM-166CTD and SIVgsnVpu 166TM-71CTD), SIVgsn71 Vpu IF/LL (I23 and F24 were mutated to L23 and L24; black square) and SIVgsn166 Vpu LL/IF (L23 L24 were mutated to I23 F24; white square). (D) HeLa cells were infected with a VSV-G-pseudotyped HIV-1 isolate whose vpu gene was intact (HIV-1 NL4-3 Vpu), deleted (Udel), or replaced with that of SIVgsn71, SIVgsn166, or the mutants listed in panel C. Cells were stained with anti-BST-2 antibody and subjected to flow cytometry to detect cell surface human BST-2 expression. Data in graphs are presented as described for Fig. 2B. One representative dot plot is shown.
DISCUSSION
In cross-species transmissions, viruses encounter orthologous host restrictive proteins with structure and properties slightly different from those of their natural host, and to overcome such new barriers, they change the antagonist (e.g., Vpu versus Nef versus Env) or undergo mutations in existing accessory proteins that enable counteracting restrictions in a new host (49). On the other hand, the host genes against pathogens would have evolved under selective pressure by life-threating pathogens, resulting in amino acid differences among orthologous host restrictive proteins. The species specificity of physical and functional interactions between primate BST-2s and lentiviral Vpu proteins observed in this study is based on the amino acid differences among orthologues of primate BST-2 and those among Vpu proteins. These observations inspired us to assume that such genetic diversity has arisen as a result of combat between host and pathogens.
First of all, we assessed the physical interaction between Vpu proteins and human BST-2 by BiFC assay based on fluorescent protein reconstitution, which can quantify their binding in live cells as reported previously (8, 9, 44). As shown in Fig. 1F and G, HIV-1 NL4-3 Vpu bound to human BST-2, and a A14LW22A mutation in HIV-1 NL4-3 Vpu caused loss of the binding activity to human BST-2, which is consistent with previous results obtained by coimmunoprecipitation assay (14, 16, 17, 50) and nuclear magnetic resonance (NMR) spectroscopy (12). In this assay, we showed that none of the Vpu proteins from SIVcpz can interact with human BST-2 (Fig. 1F), a phenotype similar to that of SIVcpzMB897 Vpu as shown previously (8, 44). Because binding of Vpu to BST-2 was reported to be necessary for the antagonism (8), it is reasonable to assume that the lack of interaction between SIVcpz Vpu proteins and human BST-2 accounts for their lack of competence in antagonizing human BST-2 (32, 42). Although Vpu proteins from SIVgsn166, SIVmonCML1, and SIVmus1085 could not reduce the levels of cell surface human BST-2 expression (Fig. 3), they successfully downregulated BST-2 of their natural host monkey (Fig. 2B to D). These results are consistent with those of a previous study showing that these SIV Vpu proteins enhanced virus release in the presence of monkey BST-2s but not in the presence of human BST-2 (32). Efficient binding of these Vpu proteins to BST-2 of their natural host but not to human BST-2 is likely to constitute the basis of BST-2 antagonism. Meanwhile, Vpu from SIVgsn166, which was isolated through the same serologic testing and is highly related to SIVgsn71 (28), failed to downregulate cell surface human BST-2 and facilitate virus release in HeLa cells (Fig. 3 and 7A and B). This is consistent with a previous report showing that SIVgsn166 Vpu did not enhance virion production in the presence of human BST-2 and that a chimeric Vpu consisting of SIVgsn166 Vpu TM domain and HIV-1 NL4-3 Vpu cytoplasmic domain did not downregulate human BST-2 in HEp-2 cells or enhance virion production in the presence of human BST-2 (51). SIVgsn166 Vpu, SIVmonCML1 Vpu, and SIVmus1085 Vpu were reported to antagonize not only BST-2 from natural host monkey but also that from some cognate monkeys (32). Taking the data together, human BST-2 is considered to be more resistant to SIV Vpu proteins than monkey BST-2.
Interestingly, SIVgsn71 Vpu binds not only to BST-2 of its natural host (GSN monkey; Fig. 1C and 6A) but also to human BST-2 and does so at the highest level of efficiency among SIV Vpu proteins (Fig. 1G), which explains why it antagonizes human BST-2 (Fig. 4). In fact, only SIVgsn71 Vpu induced downregulation of human BST-2 from the cell surface of infected cells, which is consistent with a previous report showing its downregulation in cells transfected with a plasmid expressing SIVgsn71 Vpu (42). This effect of SIVgsn71 Vpu was also shown in infected T-cell line MT4 cells (data not shown) as well as in infected HeLa cells (Fig. 3, 4C, and 5C). In addition, release of infectious virus in the presence of BST-2 was quantified with TZM-bl reporter cells in this study as reported previously (6, 8, 9, 33, 52). SIVgsn71 Vpu facilitated HIV-1 escape from human BST-2 (Fig. 4D and E), although the effect was milder than that seen with HIV-1 NL4-3 Vpu. SIVgsn71 Vpu mutants, including SIVgsn71 Vpu-TMRD (Fig. 4B) and Vpu-m1 and Vpu-m2 (Fig. 5B), that showed less binding activity to human BST-2 had lost the ability to reduce cell surface human BST-2 expression and to enhance release of infectious virus (Fig. 4C and D and 5C and D). Collectively, these results corroborate our conclusion that the physical interaction of Vpu and BST-2 is essential for the human BST-2 antagonism by SIVgsn71 Vpu. The AxxxxxxxW motif in the TM domain of HIV-1 NL4-3 Vpu is crucial for antagonizing human BST-2 and, more specifically, for binding to human BST-2 (12, 14). However, motifs in SIV Vpu essential for antagonizing BST-2 of their natural host have not been explored. Surprisingly, we identified two AxxxxxxxW motifs (A22W30 and A25W33) in SIVgsn71 Vpu (Fig. 5A) and mutagenesis analyses showed that only one motif (A22W30) is required for the GSN BST-2 antagonism (Fig. 6). This result is similar to those reported from previous studies investigating HIV-1 NL4-3 Vpu interaction with human BST-2 (12, 14). From the viewpoint of viral evolution, HIV-1 group M is believed to be derived from SIVcpz but not directly from SIVgsn. Interestingly, Vpu proteins from three SIVcpz isolates (MB897, MB66, and LB7), which are thought to be closely related to HIV-1 group M (53, 54), have one tryptophan residue (W) but not an AxxxxxxxW motif in the TM domain (Fig. 1B). Thus, it is tempting to speculate that the AxxxxxxxW motif for interacting with BST-2 of natural host animals was acquired by SIVgsn and HIV-1 independently by convergent evolution rather than by direct succession from SIVgsn to HIV-1 group M. However, since we do not know whether SIVcpz transmitted from infected chimpanzees to human beings expressed Vpu proteins that were structurally and functionally similar to the Vpu proteins of current SIVcpz, more genetic information will be needed to understand how HIV-1 Vpu got the ability to bind to and antagonize human BST-2.
Interestingly, two AxxxxxxxW motifs of SIVgsn71 Vpu are needed to antagonize human BST-2 (Fig. 5). This suggests that the mechanism by which SIVgsn71 Vpu works on human BST-2 is distinct from and perhaps more complicated than that by which it works on GSN BST-2, which could partly explain the more resistant phenotype of human BST-2 against SIV Vpu proteins. In view of the coevolution of host and pathogens, human BST-2 should also have evolved under life-threating pressure by viruses expressing Vpu or proteins functionally closely related to Vpu. The results reported in this paper may shed light on the tailored loss and/or gain of function of Vpu during cross-species transmissions and on the coevolution of lentiviral proteins and host restriction factors.
MATERIALS AND METHODS
Cells and transfection.HEK293T, HeLa, and TZM-bl cells were cultured in Dulbecco’s modified essential medium (DMEM) (Nacalai Tesque, Kyoto, Japan) supplemented with 10% heat-inactivated fetal bovine serum (FBS; Sigma-Aldrich Co., St. Louis, MO, USA) and penicillin-streptomycin (Nacalai Tesque). These cells were cultured at 37°C with 5% CO2. HEK293T cells and HeLa cells were transfected using polyethylenimine (PEI; PolyScience, Niles, IL, USA).
Plasmids.The plasmids expressing KGN-tagged NL4-3 Vpu, NL4-3 Vpu14/22, and SIVcpzMB897 Vpu and the plasmid expressing KGC-tagged human BST-2 were described previously (8, 9). The codon-optimized vpu genes from SIVcpzMB66 (GenBank accession no. ABD19480), SIVcpzLB7 (GenBank accession no. ABD19489), SIVgsn-99CM71 (GenBank accession no. AAM90227), SIVgsn-99CM166 (GenBank accession no. AAM90236), SIVmon-99CMCML1 (GenBank accession no. AAR02384), and SIVmus-01CM1085 (GenBank accession no. AAR02375) were amplified by single-step assembly of 6 oligonucleotides (55) and cloned into phmKGN-MN vector (MBL International, Woburn, MA, USA), as described previously (8, 9). The SIVgsn-99CM71 Vpu-TMRD mutant was constructed by overlap extension PCR with an N-terminal TM domain of pNL4-3 VpuRD-KGN (amino acid residues 1 to 27) (8) and a C-terminal cytoplasmic domain of pSIVgsn-99CM71 Vpu-KGN (amino acid residues 29 to 76). The SIVgsn-99CM71 Vpu AxxxxxxxW mutants (m1 and m2) were generated and inserted into phmKGN-MN vector by overlap extension PCR, using the following primers: N3-A22LW30A (CACTTGTCCGCGGCCAGGTACAGGGCGAAAATGAGCAG), C5-A22LW30A (CTGCTCATTTTCGCCCTGTACCTGGCCGCGGACAAGTG), N3-A25LW33A (CCCTTAATCGCCTTGTCCCAGGCCAGGTACAGGAGGAAAATG), and C5-A25LW33A (CATTTTCCTCCTGTACCTGGCCTGGGACAAGGCGATTAAGGG). The bst-2 genes of greater spot-nosed monkey (GSN), mona monkey (MON), and mustached monkey (MUS) were amplified from pCG_GSN tetherin internal ribosome entry site (IRES) DSRed, pCG_MON tetherin IRES DSRed, and pCG_MUS tetherin IRES DSRed (kindly given by Daniel Sauter and Frank Kirchhoff), respectively, and were tagged with KGC fragment in phmKGC-MC vector (MBL International).
The HIV-1 proviral vector pNL4-3EGFP Δenv Δnef has been described previously (47). The HIV-1 Vpu-deficient proviral vector pNL4-3EGFP Δenv Δnef Δvpu was generated as reported for pNL4-3 Udel (56). The proviral vectors carrying vpu or vpu mutants from SIVs were generated by replacing the HIV-1 NL4-3 vpu gene with full-length SIV vpu genes, which were amplified by PCR using the pVpu-KGN plasmids as templates, and inserted into the EcoRI (position 5743) and BamHI (position 8465) sites of the pNL4-3EGFP Δenv Δnef vector by overlap extension PCR together with the other two fragments before and after the HIV-1 NL4-3 vpu gene amplified from the pNL4-3EGFP Δenv Δnef vector. The proviral vectors carrying SIVgsn71 and SIVgsn166 vpu chimeras (SIVgsnVpu71TM-166CTD and SIVgsnVpu166TM-71CTD) or mutants (SIVgsn71VpuIF/LL and SIVgsn166VpuLL/IF) were generated by overlap extension PCR using the pNL4-3EGFP Δenv Δnef ΔvpuSIVgsn71vpu and pNL4-3EGFP Δenv Δnef ΔvpuSIVgsn166vpu vectors as templates and the following primers: N3-SIVgsn71IF2324LL (GGTACAGGGCCAGCAGGGCCAGACAGAAG), C5-SIVgsn71IF2324LL (CTGTCTGGCCCTGCTGGCCCTGTACC), N3-SIVgsn166LL2324IF (GGTACAGGGCGAAAATGGCCACACACAG), C5-SIVgsn166LL2324IF (CCTGTGTGTGGCCATTTTCGCCCTGTACCTGGCCTG), N3-chimera71&166 (CCACTTGTCCCAGGCCAGGTACAGGGC), and C5-chimera71&166 (GCCCTGTACCTGGCCTGGGACAAGTGG).
The pFLAG-huBST-2 was constructed by amplifying the human bst-2 gene from pKGC-huBST-2 and inserted into the EcoRI and EcoRV sites in p3×FLAG-CMV-10 (Sigma-Aldrich) using the following primers: N5-EcoRI huBST-2 (ACGTGGAATTCAATGGCATCTACTTCGTATGAC) and KG3′-hu&chiBST (GTTCTCGAGTTTCACTGCAGCAGAGCGCTGAG). pNL4-3Vpu-HA (hemagglutinin), pNL4-3Vpu14/22-HA, and pSIVgsn71Vpu-HA were generated by amplifying the corresponding vpu genes from pVpu-KGN plasmids, which were inserted into the KpnI and XbaI sites in phmKGN-MN (MBL International) with the following primers: N5-univ pBiFC-MN (CGCCCCATTGACGCAAA) and C3-pBiFC-MN-HAtag-XbaI (GCTCTAGATTATGCATAATCTGGCACATCATATGGATAGGCGGCCGCTGATCCAAGCTTTG).
Establishment of stable cell lines.HeLa cells were transfected with a plasmid expressing CRISPR/Cas9 [pSpCas9(BB)-2A-puro(PX459)V2.0 {Addgene catalog no. 62988; a kind gift from Feng Zhang}] and a single guide RNA (sgRNA) targeting part of human bst-2 gene (TAAGCGCTGTAAGCTTCTGC). After drug selection performed with 4 μg/ml of puromycin for 20 h, cells were subjected to limiting dilution to get single-cell clones. The resultant human bst-2 knockout cells (clone 61H5) were transfected with KGC-tagged GSN BST-2, MON BST-2, or MUS BST-2 expression plasmid and were subjected to drug selection with 500 μg/ml of G418 and limiting dilution to get single-cell clones. The cell surface expression of human or monkey BST-2 was verified using a FACSCalibur flow cytometer (BD Bioscience, San Diego, CA, USA) with anti-BST-2 rabbit polyclonal antibody (57).
BiFC assay.HEK293T cells were seeded at a density of 8 × 105 cells per well of a 6-well plate the day before transfection. Cells were cotransfected with 1 μg of mKGC-tagged human BST-2 or GSN BST-2 plasmids, mKGN-tagged HIV or SIV Vpu plasmids (the amounts of these plasmids were optimized to achieve similar expression levels on Western blotting), and 2 μg of pmCherry, which served as transfection marker (8). At 48 h after transfection, cells were harvested and resuspended with 2 ml of phosphate-buffered saline (PBS). The cell suspension was divided into two aliquots, and the aliquots were subjected to flow cytometry and Western blotting separately. One aliquot of the cells was fixed with 0.4% paraformaldehyde (PFA; Nacalai Tesque)–PBS and analyzed using a FACSCalibur flow cytometer (BD Bioscience) as previously reported (8, 9, 44). The Geo Mean values of bimolecular fluorescence complementation (BiFC) signal in mCherry-positive cells were calculated and analyzed by using BD Cell Quest Pro software (BD Bioscience). The other aliquot was pelleted and resuspended with 150 μl of PBS for Western blotting. After adding an equal volume of 2× sample buffer (0.125 M Tris-HCl [pH 6.8], 10% [vol/vol] 2-mercaptoethanol [2-ME], 4% [wt/vol] sodium dodecyl sulfate [SDS], 10% [wt/vol] sucrose, 0.01% [wt/vol] bromophenol blue [BPB]), cells were incubated at 95°C for 10 min and 10 μl of supernatant was loaded for Western blotting, where proteins were detected with anti-KGC antibody (MBL International) or anti-KGN antibody (MBL International). α-Tubulin was detected as a loading control with anti-α-tubulin antibody (Sigma-Aldrich). Horseradish peroxidase (HRP)-conjugated rabbit anti-mouse IgG (American Qualex International Inc., San Clemente, CA, USA) and Western Lightning Plus-ECL (PerkinElmer, Waltham, MA, USA) were used to visualize the proteins with a Li-Cor Odyssey imaging system (Li-Cor Biosciences, Lincoln, NE, USA).
Immunoprecipitation.HEK293T cells were seeded at a density of 8 × 105 cells per well of a 6-well plate the day before transfection. Cells were transfected with 2 μg of p3×FLAG-huBST-2 plasmid together with 2 μg of pNL4-3Vpu-HA, pNL4-3Vpu14/22-HA, or pSIVgsn71Vpu-HA. At 48 h after transfection, cells were lysed on ice for 30 min with 500 μl of lysis buffer (50 mM Tris [pH 7.4], 150 mM NaCl, 1% [vol/vol] digitonin). After centrifugation, the supernatants were precleared with 10 μl of Pierce protein A agarose beads (Thermo Scientific, Rockford, IL, USA) for 1 h at 4°C. The precleared cell lysate was then incubated with another 20 μl of Pierce protein A agarose together with 1 μg of HA probe Y-11 (Santa Cruz Biotechnology, Santa Cruz, CA, USA) for 4 h at 4°C. The beads were washed 5 times with lysis buffer, resuspended with 20 μl of 2× sample buffer, and boiled for 5 min at 95°C. Cell lysates and immunoprecipitates were subjected to Western blotting, where proteins were detected with anti-FLAG BioM2-Biotin antibody (Sigma-Aldrich), anti-HA antibody (MBL International), and anti-α-tubulin antibody (Sigma-Aldrich). HRP-conjugated rabbit anti-mouse IgG or streptavidin-HRP (Invitrogen, Frederick, MD, USA) and Western Lightning Plus-ECL were used to visualize the proteins with a Li-Cor Odyssey imaging system.
Staining of cell surface and intracellular BST-2.For generating VSV-G pseudotyped HIV-1 vector, HEK293T cells were cotransfected with 3 μg of proviral DNA pNL4-3EGFP Δenv Δnef vector expressing HIV or SIV Vpu and 1 μg of pMISSION-VSV-G (Sigma-Aldrich). The virus-containing supernatant was harvested 48 h after transfection, filtered using 0.45-μm-pore-size filters (Merck Millipore, Burlington, MA, USA), and stored at –80°C. Approximately 6 × 105 HeLa cells or HeLa BST-2 knockout cells expressing GSN, MON, or MUS BST-2 (HeLa KO-GSN BST-2, HeLa KO-MON BST-2, or HeLa KO-MUS BST-2 cells) were incubated with an optimized amount of viral supernatant in 6-well plates such that approximately 40% to 50% of the cells become EGFP positive. At 48 h after infection, cells were harvested. To detect cell surface BST-2 expression, cells were incubated with anti-BST-2 rabbit polyclonal antibody (57) for 30 min on ice followed by washing with PBS and staining with goat anti-rabbit IgG conjugated Alexa Fluor 647 (Jackson laboratory, Bar Harbor, ME, USA) for another 30 min on ice. After washing with PBS and fixing with 0.4% paraformaldehyde (PFA; Nacalai Tesque)–PBS were performed, cells were subjected to flow cytometry. To detect intracellular BST-2 expression, cells were fixed with 0.4% paraformaldehyde (PFA; Nacalai Tesque) for 10 min at room temperature before washing was performed with intracellular staining permeabilization wash buffer (BioLegend, San Diego, CA, USA). Permeabilized cells were incubated with anti-BST-2 rabbit polyclonal antibody and goat anti-rabbit IgG-conjugated Alexa Fluor 647 as described above. The Geo Mean values corresponding to the Alexa Fluor 647 signal in EGFP-positive cells and EGFP-negative cells were calculated by using BD Cell Quest Pro software (BD Bioscience).
Quantification of virus release.HeLa, HeLa human BST-2 knockout (HeLa BST-2 KO), or HeLa KO-GSN BST-2 cells were seeded at a density of 1.6 × 106 cells per well of a 6-well plate the day before transfection. Cells were cotransfected with proviral DNA pNL4-3EGFP Δenv Δnef vector carrying HIV or SIV vpu genes (the amount of each plasmid was optimized to give luciferase values in TZM-bl assays similar to those measured for viruses released from HeLa BST-2 KO cells) and 1 μg of pMISSION-VSV-G. The culture medium was changed 8 h after transfection. The virus-containing supernatant was harvested 24 h after transfection, centrifuged at 8,000 rpm for 1 min to remove cell debris, and subjected to TZM-bl assay and ELISA with an HIV-1 p24 ELISA kit (Rimco Corporation, Okinawa, Japan). The producer cells were lysed with 200 μl of lysis buffer (25 mM Tris [pH 7.8], 8 mM MgCl2, 1 mM dithiothreitol [DTT], 1% [vol/vol] Triton X-100, 15% [vol/vol] glycerol) and subjected to Western blotting. The TZM-bl assay was performed as described previously (8, 52). Briefly, TZM-bl cells were seeded at a density of 5 × 104 cells per well of a 12-well plate the day before infection. Virus-containing supernatant (100 μl for each well) was used for infection in triplicate. At 47 h later, TZM-bl cells were lysed with 150 μl of lysis buffer. The luciferase activity was measured using 40 μl of cell lysate and a GloMax multidetection system (Promega Corp., Madison, WI, USA). Protein concentrations were measured by Bradford assay for normalization of the luciferase activity. For Western blotting, whole-cell lysates prepared from producer cells were mixed with an equal volume of 2× sample buffer and boiled at 95°C for 10 min. The total protein concentration was determined using Bradford assay. A 30-μg volume of each sample was subjected to 10% SDS-PAGE and transferred to polyvinylidene difluoride (PVDF) membrane (Merck Millipore). Gag proteins were detected with anti-p24 antibody (Abcam, Inc., Cambridge, MA, USA), and α-tubulin used as a loading control was detected with anti-α-tubulin antibody (Sigma-Aldrich). HRP-conjugated rabbit anti-mouse IgG and Western Lightning Plus-ECL were used to visualize the proteins with a Li-Cor Odyssey imaging system (Li-Cor Biosciences).
Statistical analysis.All the statistical analyses were performed using GraphPad Prism 6. The data are presented as means ± standard errors of the means (SEM) of results from at least three independent experiments. Significance values were calculated with Student's t test or with one-way analysis of variance (ANOVA) and with Dunnett’s multiple-comparison test (***, P < 0.001; **, P < 0.01; *, P < 0.05; n.s., not significant [P > 0.05]).
ACKNOWLEDGMENTS
We thank Daniel Sauter and Frank Kirchhoff for pCG_GSN tetherin IRES DSRed, pCG_MON tetherin IRES DSRed, and pCG_MUS tetherin IRES DSRed; Feng Zhang for pSpCas9(BB)-2A-Puro (PX459) V2.0 plasmid; Yoshio Koyanagi for plasmids expressing KGN-tagged NL4-3 Vpu or KGC-tagged human BST-2; Amy Andrew for anti-BST-2 polyclonal antibodies; Kenzo Tokunaga for technical instructions; and Natsumi Tamura for technical assistance. T.Y. also thanks Yuko Yoshida for lively discussions.
This work was supported by grants 26860302 to T.Y., 18K07143 to S.Y. from the Ministry of Education, Culture, Sports, Science, and Technology, research grant 2902 to T.Y. from Nakatsuji Foresight Foundation, and collaborative research grant 2A175 to H.T. with Shionogi & Co. Ltd. in Japan. W.Y. was supported by the China Scholarship Council (CSC). We also recognize the support from Binlian Sun and Hirofumi Akari.
T.Y. designed the study. W.Y., T.Y., and S.H. performed the experiments. W.Y., T.Y., S.H., H.T., K.S., and S.Y. analyzed and interpreted the data. W.Y., T.Y., H.T., K.S., and S.Y. contributed reagents and materials. W.Y., T.Y., and S.Y. wrote the paper. All of us reviewed the manuscript.
We have declared that no competing interests exist.
FOOTNOTES
- Received 30 September 2019.
- Accepted 21 October 2019.
- Accepted manuscript posted online 30 October 2019.
- Copyright © 2020 American Society for Microbiology.
