The VP40 Protein of Marburg Virus Exhibits Impaired Budding and Increased Sensitivity to Human Tetherin following Mouse Adaptation

ABSTRACT

The Marburg virus VP40 protein is a viral matrix protein that spontaneously buds from cells. It also functions as an interferon (IFN) signaling antagonist by targeting Janus kinase 1 (JAK1). A previous study demonstrated that the VP40 protein of the Ravn strain of Marburg virus (Ravn virus [RAVV]) failed to block IFN signaling in mouse cells, whereas the mouse-adapted RAVV (maRAVV) VP40 acquired the ability to inhibit IFN responses in mouse cells. The increased IFN antagonist function of maRAVV VP40 mapped to residues 57 and 165, which were mutated during the mouse adaptation process. In the present study, we demonstrate that maRAVV VP40 lost the capacity to efficiently bud from human cell lines, despite the fact that both parental and maRAVV VP40s bud efficiently from mouse cell lines. The impaired budding in human cells corresponds with the appearance of protrusions on the surface of maRAVV VP40-expressing Huh7 cells and with an increased sensitivity of maRAVV VP40 to restriction by human tetherin but not mouse tetherin. However, transfer of the human tetherin cytoplasmic tail to mouse tetherin restored restriction of maRAVV VP40. Residues 57 and 165 were demonstrated to contribute to the failure of maRAVV VP40 to bud from human cells, and residue 57 was demonstrated to alter VP40 oligomerization, as assessed by coprecipitation assay, and to determine sensitivity to human tetherin. This suggests that RAVV VP40 acquired, during adaptation to mice, changes in its oligomerization potential that enhanced IFN antagonist function. However, this new capacity impaired RAVV VP40 budding from human cells.

IMPORTANCE Filoviruses, which include Marburg viruses and Ebola viruses, are zoonotic pathogens that cause severe disease in humans and nonhuman primates but do not cause similar disease in wild-type laboratory strains of mice unless first adapted to these animals. Although mouse adaptation has been used as a method to develop small animal models of pathogenesis, the molecular determinants associated with filovirus mouse adaptation are poorly understood. Our study demonstrates how genetic changes that accrued during mouse adaptation of the Ravn strain of Marburg virus have impacted the budding function of the viral VP40 matrix protein. Strikingly, we find impairment of mouse-adapted VP40 budding function in human but not mouse cell lines, and we correlate the impairment with an increased sensitivity of VP40 to restriction by human but not mouse tetherin and with changes in VP40 oligomerization. These data suggest that there are functional costs associated with filovirus adaptation to new hosts and implicate tetherin as a filovirus host restriction factor.

INTRODUCTION

Marburg viruses (MARV), which are negative-sense, enveloped RNA viruses classified along with Ebola viruses (EBOV) in the Filoviridae family, are zoonotic pathogens that likely use bats as reservoir hosts (1–3). While filoviruses appear to be relatively nonpathogenic in bats (4, 5), these viruses cause severe, often lethal, infections in humans and nonhuman primates (6). This is apparent in outbreaks of MARV in human populations, which occur sporadically, with reported case fatality rates ranging from 25 to 90% (6).

It is unclear why filoviruses are apathogenic in some species but extremely deadly in others. Rodents may be useful models to begin addressing such questions given that neither EBOVs nor MARVs kill mice or guinea pigs. However, mice lacking a functional alpha/beta interferon (IFN-α/β) receptor die following intraperitoneal (i.p.) inoculation with EBOVs or MARVs, and adaptation by serial passage in mice or guinea pigs yields viruses that are lethal in the respective species (7–13). These observations implicate the IFN-α/β response as a host determinant of virulence, and genetic changes acquired by adapted viruses may suggest molecular mechanisms that determine virulence in specific hosts.

Lethal mouse variants of the Ci67 and Ravn virus (RAVV) strains of Marburg virus have been generated by serial passage in mice, and genetic changes have accrued throughout the genome during adaptation (10, 11). Among the proteins acquiring changes was the VP40 protein, which functions as the viral matrix protein and as an inhibitor of Janus kinase 1 (JAK1) signaling (14).

A common assay for filoviral VP40 matrix protein function is a budding assay, where expression of VP40 alone is sufficient to induce the formation of virus-like particles (VLPs) (15). Determinants of VP40 budding efficiency include factors intrinsic to the viral protein as well as host factors. Late domains are among the best-studied sequence motifs present in VP40s that facilitate budding through interaction with host factors, and deletion or mutation of critical late domain amino acid residues impairs VP40 budding (16–19). For MARV VP40, the late domain PPPY associates with Tsg101 and Nedd4 (19, 20). In addition, amino acid sequences in VP40 located downstream of the traditional late domain motifs, including the motif LPLGIM, also influence budding (21). In the absence of these motifs, VLP release is reduced as VP40 oligomerization and plasma membrane localization are altered (21). A host factor that can impair VP40 budding is tetherin/bone marrow stromal cell antigen 2 (BST-2)/CD317. Tetherin, an IFN-induced type II transmembrane protein, inhibits the release of a variety of enveloped viruses from the cell surface (22–27). Expression of tetherin can block the release of filoviral VP40 VLPs (23, 24, 27). However, infectious filoviruses are resistant to the antiviral effects of tetherin, because filoviral glycoproteins (GP) counter tetherin function (24, 27, 28).

As an inhibitor of JAK1, MARV VP40s prevent IFN-α/β and IFN-γ signaling (14). However, RAVV VP40 was not an effective inhibitor of IFN signaling in mouse cell lines. In contrast, mouse-adapted RAVV (maRAVV) VP40 had gained the ability to inhibit interferon signaling in mouse cells. This gain of function was mapped to changes at VP40 amino acid residues 57 and 165 (29). These and 5 other amino acid mutations acquired during mouse adaptation did not affect the ability of maRAVV VP40 to inhibit IFN responses in human cells, and both the parental and maRAVV VP40s could efficiently bud from mouse cells (29).

The present study addresses the budding function of parental and maRAVV VP40 in human cell lines. We demonstrate a dramatic impairment of budding by maRAVV VP40 in either human-origin Huh7 or 293T cells and correlate this inhibition to restriction of maRAVV VP40 budding by human tetherin. In contrast, maRAVV VP40 is efficient in the mouse-derived Hepa1.6 cell line and is largely unaffected by overexpression of mouse tetherin. The restriction in human cells maps in part to residue 57, and change at this residue also alters RAVV VP40 oligomerization. This change at residue 57 was also important for VP40 acquisition of anti-IFN function in mouse cells (29). This suggests a model in which RAVV VP40 undergoes changes in its oligomerization potential as it is adapted to mice, enhancing its IFN antagonist function. However, this new capacity resulted in a cost for RAVV VP40 budding function in human cells. Therefore, these observations provide novel insight into the determinants and restrictions that may regulate host switching by filoviruses.

MATERIALS AND METHODS

Cells.

Huh7, Hepa1.6, and 293T cells were maintained in Dulbecco's modified Eagle's medium (DMEM; Invitrogen) supplemented with 10% fetal bovine serum (FBS; Invitrogen) and grown at 37°C in the presence of 5% CO₂. One Shot TOP10 chemically competent Escherichia coli cells (Invitrogen) were used for all cloning procedures following the manufacturer's protocol.

Plasmids.

FLAG-tagged RAVV and viral protein expression plasmids were described previously (14, 29). The hemagglutinin (HA)-tagged RAVV VP40 plasmids were constructed by restriction enzyme double-digestion reactions of FLAG-tagged VP40 plasmid with XhoI (NEB) and NotI (NEB). Purified product was subsequently ligated into HA-tagged pCAGGS vector with T4 DNA ligase (NEB). The maRAVV VP40 mutants were constructed using the QuikChange II XL site-directed mutagenesis kit (Stratagene). HA-tagged Nipah virus (NiV) M ΔYMYL mutant protein expression plasmid was described previously (30). Plasmids encoding HIV-1 gag-pol ΔVpu and Vpu (pcDNA) were kindly provided by Lubbertus Mulder and Viviana Simon (Icahn School of Medicine at Mount Sinai). Plasmids encoding tetherin were kindly provided by Peter Palese (Icahn School of Medicine at Mount Sinai). PCR primers were designed to amplify tetherin from the expression vector, and the purified PCR product was subcloned into pCAGGS such that it possessed an N-terminal HA tag.

VLP budding assay.

The budding assay was performed as previously described (30). Briefly, Huh7, Hepa1.6, or 293T cells were transfected with the indicated expression plasmids with or without small interfering RNA (siRNA) using Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. Forty-eight hours posttransfection, supernatants were harvested and clarified by centrifugation at 1,000 rpm at 4°C for 10 min. Clarified supernatant layered over a 20% sucrose cushion in NTE buffer (10 mM NaCl, 10 mM Tris [pH 7.5], 1 mM EDTA [pH 8.0]) was ultracentrifuged in a Beckman SW-41 rotor at 36,000 rpm for 2 h at 4°C. After ultracentrifugation, the supernatant was aspirated, and VLPs were resuspended in NTE buffer. Purified VLPs were treated with trypsin at 37°C for 1 h. Transfected cells were washed with phosphate-buffered saline (PBS) and lysed in radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris [pH 7.4], 150 mM NaCl, 0.1% SDS, 0.5% deoxycholate, 1% NP-40) supplemented with cOmplete protease inhibitor cocktail (Roche). Lysates and VLPs were analyzed by SDS-PAGE and visualized by Western blotting with anti-FLAG (Sigma-Aldrich) antibodies. Lysates were also analyzed by Western blotting with anti-β-tubulin, anti-HA (tetherin), anti-HIV-1 Vpu, anti-EBOV GP, and anti-MARV GP antibodies as indicated.

Real-time qPCR.

Total RNA was extracted from 293T cells either mock treated or transfected with validated Mission siRNAs targeting human tetherin (Sigma-Aldrich) or control siRNAs (Sigma-Aldrich). Each RNA sample was reverse transcribed using oligo(dT)s supplied in the SuperScript III first-strand synthesis system for reverse transcription (RT)-PCR (Invitrogen). Each cDNA was subsequently used in real-time quantitative PCRs (qPCRs) with human tetherin-specific primers (F, CCACCTGCAACCACACTG; and R, CCTGAAGCTTATGGTTTAATGTAGTG) designed using the Roche Universal Probe Library Assay Design Center. Tetherin signal was normalized to β-actin mRNA using primers previously described (31).

Immunofluorescence and confocal microscopy.

Huh7 and Hepa1.6 cells were seeded on glass coverslips coated with poly-l-lysine (Sigma-Aldrich) in a 24-well plate and transfected with the indicated plasmids using Lipofectamine 2000. Twenty-four hours posttransfection, cells were washed twice with PBS-CM (phosphate-buffered saline, 1 mM CaCl₂, 1 mM MgCl₂) and fixed with 4% paraformaldehyde in PBS-CM for 10 min. Cells were washed twice with PBS-CM, blocked with 4% normal goat serum, and stained with primary antibodies to FLAG-tagged RAVV VP40 and/or tetherin for 1 h at room temperature. Cells were washed three times with PBG (PBS, 0.5% bovine serum albumin [BSA], 0.15% glycine). Cells were then incubated with Hoechst nuclear dye 33342 (Invitrogen) and secondary antibodies conjugated to either Alexa Fluor 488 or Alexa Fluor 555 (Invitrogen) for 30 min at room temperature. Following incubation, cells were washed three times with PBG and mounted onto slides with VectaShield mounting medium (Vector Laboratories). Confocal images were acquired using a Leica TCS SP5 DM confocal microscope.

Co-IP assay.

Huh7 or Hepa1.6 cells were transfected with the indicated plasmids using Lipofectamine 2000. Twenty-four hours posttransfection, cells were washed with PBS and lysed in immunoprecipitation (IP) lysis buffer (50 mM Tris [pH 7.5], 150 mM NaCl, 0.5% NP-40, 1 mM EDTA [pH 8.0]) supplemented with cOmplete protease inhibitor cocktail (Roche). Lysates were incubated with anti-FLAG M2 beads (Sigma-Aldrich) overnight at 4°C. Precipitated proteins were analyzed by SDS-PAGE and Western blotted using anti-HA antibody. Whole-cell lysates (WCLs) were analyzed in a similar manner and blotted with anti-FLAG, anti-HA, and β-tubulin antibodies.

RESULTS

maRAVV VP40 fails to bud efficiently from human cell lines.

Mouse adaptation of RAVV resulted in 7 amino acid changes in the VP40 protein (11) (Fig. 1A). To determine whether mouse adaptation altered the ability of VP40 to bud as VLPs, the release of the parental RAVV VP40 and the mouse-adapted RAVV (maRAVV) VP40 was assayed in two human cell lines (Huh7 and 293T) and one mouse cell line (Hepa1.6). Cells were concurrently transfected with plasmids encoding wild-type (wt) NiV M and a previously described, budding-defective mutant, NiV M ΔYMYL, which served as positive and negative budding controls, respectively (30). Forty-eight hours posttransfection, cell culture supernatants were harvested, VLPs were pelleted through a 20% sucrose cushion, and the clarified supernatant was treated with trypsin to eliminate protein not protected by VLP membranes and then analyzed by Western blotting. In all three cell lines, the NiV M controls behaved as expected, with wt NiV M exhibiting budding and release and the mutant NiV M failing to be released (Fig. 1B) (30). In Hepa1.6 cells, both RAVV and maRAVV VLPs were detected at comparable levels, consistent with a previous study (Fig. 1B) (29). Interestingly, while RAVV VLPs were detected in supernatant harvested from transfected Huh7 and 293T cells, maRAVV VLP release from Huh7 and 293T cells was significantly impaired, despite similar expression levels of RAVV and maRAVV VP40 in all cell lines examined. Therefore, although VP40 budding from the mouse cell line is not affected by mouse adaptation, egress from the human cells is severely inhibited.

FIG 1.

maRAVV VP40 budding is impaired in human cells. (A) Schematic diagram depicting the seven amino acid differences between nonadapted and mouse-adapted RAVV VP40s. (B) VLP assay to assess budding of FLAG-RAVV and FLAG-maRAVV VP40 in mouse and human cell lines. Wild-type and mutant NiV M protein (NiV M and NiV ΔYMYL, respectively) expression plasmids were included as positive and negative controls, respectively. Hepa1.6, Huh7, or 293T cells were transfected with the indicated expression plasmids. Forty-eight hours posttransfection, VLPs were purified, and whole-cell lysates (WCLs) were prepared and analyzed for the indicated proteins by Western blotting. VLP and WCL blots were probed with anti-FLAG, anti-HA, and anti-β-tubulin antibodies.

VP40 budding is inhibited by tetherin in a species-specific manner.

A critical step in the release of filovirus particles from the cell surface is the localization of VP40 to the plasma membrane (21, 32). To assess whether amino acid changes acquired during mouse adaptation altered VP40 intracellular distribution, Huh7 cells were transfected with RAVV VP40 or maRAVV VP40. Twenty-four hours posttransfection, cells were fixed and analyzed by immunofluorescence microscopy. Both RAVV and maRAVV VP40 were found throughout the cytoplasm (Fig. 2A). Interestingly, the VP40 of cells transfected with maRAVV VP40 was localized to cell surface protrusions (Fig. 2A) reminiscent of the structures formed in cells expressing tetherin and VP40, as reported by Kaletsky et al. (24). The abundance of these tethered protrusions was markedly lower on the surface of cells transfected with RAVV VP40 (Fig. 2A). To determine whether the tethered structures identified on the surface of Huh7 cells transfected with VP40 correlated with functional inhibition of VP40 budding by tetherin, a VLP assay was performed with RAVV and maRAVV VP40 in Huh7 cells transfected with an expression plasmid for human tetherin. Consistent with the microscopy data, maRAVV VP40 did not bud, regardless of whether or not tetherin was overexpressed. RAVV VP40 budding, however, was decreased in a tetherin dose-dependent manner, but it was only completely inhibited at higher concentrations (500 to 1,000 ng/μl) of tetherin plasmid (Fig. 2B). This suggests that RAVV VP40 may possess some intrinsic resistance to tetherin. To further investigate restriction of maRAVV VP40 budding by human tetherin, VLP assays were performed with Hepa1.6 cells, a cell line conducive to both RAVV and maRAVV egress (Fig. 1B). Here, RAVV VP40 VLP release was only inhibited by higher concentrations of human tetherin (Fig. 2C). In contrast, maRAVV VLP release was significantly inhibited by concentrations of human tetherin plasmid as low as 10 ng/μl. Next, we sought to determine whether VP40 resistance to tetherin was species specific by performing a VLP assay with RAVV and maRAVV VP40 in the presence of increasing amounts of mouse tetherin. Interestingly, both RAVV and maRAVV VLPs were resistant to inhibition by mouse tetherin as budding was only inhibited at higher concentrations—500 to 1,000 ng/μl (Fig. 2D). These data indicate that human tetherin potently restricts budding by maRAVV VP40 but not the parental RAVV VP40. However, mouse tetherin poorly restricts either version of VP40.

FIG 2.

RAVV VP40 is resistant to both human and mouse tetherins, while maRAVV VP40 is resistant to only mouse tetherin. (A) Confocal microscopy images of tetherin-like protrusions on the surface of Huh7 cells transfected with RAVV VP40 and maRAVV VP40. Immunofluorescence analysis of Huh7 cells expressing RAVV (left) and maRAVV (right) VP40. DAPI, 4′,6-diamidino-2-phenylindole. (B) VLP assay of FLAG-RAVV and FLAG-maRAVV VP40 in the presence of increasing amounts of HA-tagged human tetherin in human cells. Huh7 cells were transfected with the indicated expression plasmids. Forty-eight hours posttransfection, VLPs were purified and the cells lysed. VLPs and WCLs were analyzed for protein expression by Western blotting. VLP and WCL blots were probed with anti-FLAG, anti-HA, and anti-β-tubulin antibodies. (C) VLP assay of RAVV and maRAVV VP40 with increasing amounts of HA-tagged human tetherin in mouse cells. The assay was performed as in panel B. (D) VLP assay of RAVV and maRAVV VP40 with increasing amounts of HA-tagged mouse tetherin in mouse cells. The VLP assay was performed as described above.

maRAVV VP40 budding is enhanced in human cells with reduced tetherin expression.

Because the budding capacity of RAVV and maRAVV VP40s correlated with restriction by human tetherin, it was of interest to determine whether tetherin could be implicated in the impaired budding of maRAVV VP40 from human cell lines. Experiments were performed with 293T cells, where maRAVV VP40 budding is severely impaired. Untreated 293T cells express low levels of tetherin, but tetherin is substantially upregulated in IFN-α-treated 293T cells (33). Two approaches were used. First, the HIV-1 Vpu protein was coexpressed with VP40 by transient transfection, as Vpu antagonizes tetherin by targeting it for proteasomal degradation and by other means (33–39). An expression plasmid encoding HIV-1 gag-pol but lacking Vpu (HIV-1 gag-pol ΔVpu) served as a control. Consistent with prior studies, HIV-1 gag was released from 293T cells even in the absence of Vpu. However, Vpu-expressing cells exhibited a modest increase in gag release, suggesting that some level of functional tetherin may be present in the 293T cells. Since the Vpu-expressing cells produced higher levels of gag in cell lysates, we cannot definitively ascribe increased release to suppression of tetherin. A similar pattern was seen in the RAVV VP40-transfected cells, which budded in the absence of Vpu but exhibited a modest increase in both expression and release in the presence of Vpu. More striking was the effect of Vpu on maRAVV VP40. Once again, maRAVV VP40 buds very poorly in the absence of Vpu, but release was substantially boosted in the presence of Vpu (Fig. 3A). For the latter samples, the Vpu-associated increase in budding was more dramatic than the corresponding increase in expression in the cell lysate. Additionally, 293T cells were cotransfected with either tetherin antagonist EBOV GP or MARV GP (24) and VP40 (Fig. 3B). Both EBOV and MARV GP enhanced RAVV VP40 release to levels observed with HIV-1 Vpu. For maRAVV VP40, VLP release was rescued in the presence of both EBOV and MARV GP. However, EBOV GP appears to function more efficiently than HIV-1 Vpu and MARV GP. We also transfected 293T cells with either control or tetherin-specific siRNAs and performed a VLP assay. Relative to cells not receiving tetherin-specific or control siRNAs, the tetherin-specific siRNA cells exhibited enhanced budding of both RAVV VP40 and maRAVV VP40 (Fig. 3C). The presence of tetherin mRNA in mock-transfected and control siRNA-transfected cells and the effective knockdown of tetherin mRNA in siRNA-transfected cells were confirmed by qRT-PCR (Fig. 3D).

FIG 3.

maRAVV VP40 release from human cells is restored by tetherin depletion. (A) VLP assay to monitor budding of FLAG-RAVV and FLAG-maRAVV VP40 in the presence of overexpressed HA-tagged human tetherin and HIV-1 Vpu in 293T cells as indicated. HIV-1 gag-pol ΔVpu plasmids transfected in the presence or absence of HIV-1 Vpu served as controls. Forty-eight hours posttransfection, VLPs were purified and the cells lysed. VLPs and WCLs were analyzed for protein expression by Western blotting. VLP and WCL blots were probed with anti-FLAG, anti-HA, anti-p24, anti-HIV-1 Vpu, and anti-β-tubulin antibodies. (B) VLP assay to monitor budding of FLAG-RAVV and FLAG-maRAVV VP40 in the presence of overexpressed HIV-1 Vpu, EBOV GP, or MARV GP in 293T cells as indicated. The VLP assay was performed as described above. VLP and WCL blots were probed with anti-FLAG, anti-HA, anti-p24, anti-HIV-1 Vpu, anti-EBOV GP, anti-MARV GP, and anti-β-tubulin antibodies. (C) VLP assay of RAVV and maRAVV VP40 in the presence of Mission siRNAs targeting human tetherin or control siRNAs. The VLP assay was performed as described above. VLP and WCL blots were probed with anti-FLAG and anti-β-tubulin antibodies. (D) Relative copy number of human tetherin mRNA in 293T cells. Cells were mock treated or transfected with control or human tetherin-specific siRNAs. Twenty-four and 48 h posttransfection, total RNA was extracted from cells, reverse transcribed, and used in real-time qPCRs. Tetherin mRNA was normalized to β-actin mRNA.

Replacing the mouse tetherin cytoplasmic tail with that of human tetherin allows restriction of maRAVV VP40 budding.

Tetherin inhibits viral egress by “tethering” virions to the surface of cells, a function attributed to the tetherin protein structure rather than its primary amino acid sequence (40). Human and mouse tetherins are only 38.7% homologous at the amino acid level, but their predicted structures are conserved. Tetherin has an N-terminal cytoplasmic tail (CT), a transmembrane (TM) domain, an extracellular (EC) domain, and a C-terminal glycosylphosphatidylinositol (GPI) domain (41). Human-mouse tetherin chimeras were constructed by replacing single domains of human tetherin (CT, TM, EC, or GPI) with the corresponding region of mouse tetherin. A reciprocal set of chimeras was also made in which individual mouse tetherin domains were swapped with the corresponding region of human tetherin. In a VLP assay performed with maRAVV VP40, the ability of human tetherin to restrict budding was impaired by swapping in single mouse tetherin domains to variable degrees, with the insertion of mouse TM or EC domains causing the greatest loss of activity (Fig. 4A). Interestingly, while mouse tetherin cannot significantly restrict budding of maRAVV VP40, swapping in the human CT domain was sufficient to confer restriction of budding (Fig. 4B).

FIG 4.

Chimeric human-mouse tetherin constructs reveal regions critical for restriction of maRAVV VP40 budding. (A) VLP assay to assess budding of maRAVV VP40 in the presence of the indicated overexpressed wild-type or chimeric human-mouse tetherins. Assay was performed in mouse Hepa1.6 cells as in Fig. 1. VLP and WCL blots were probed with anti-FLAG, anti-HA, and anti-β-tubulin antibodies. (B) VLP assay of maRAVV VP40 with various overexpressed wild-type or chimeric mouse-human tetherins in mouse cells. The assay was performed as described in the legend to Fig. 1. VLP and WCL blots were probed with anti-FLAG, anti-HA, and anti-β-tubulin antibodies.

Interchanging human and mouse domains in chimeric tetherin constructs does not significantly alter RAVV VP40 budding.

RAVV VP40 is resistant to inhibition by both human and mouse tetherins (Fig. 2). Using the two sets of human-mouse tetherin constructs described above, RAVV VP40 VLP assays were performed in Hepa1.6 cells. Each of the chimeras behaved like the parental tetherins and exhibited little suppression of RAVV VP40 budding (Fig. 5).

FIG 5.

Interchanging human and mouse domains in chimeric tetherin constructs does not significantly alter RAVV VP40 budding. (A) VLP assay of RAVV VP40 with various overexpressed wild-type or chimeric human-mouse tetherins in mouse Hepa1.6 cells. The assay was performed as described in the legend to Fig. 1. VLP and WCL blots were probed with anti-FLAG, anti-HA, and anti-β-tubulin antibodies. (B) VLP assay of maRAVV VP40 with various overexpressed wild-type or chimeric mouse-human tetherins in mouse Hepa 1.6 cells. The assay was performed as described in the legend to Fig. 1. VLP and WCL blots were probed with anti-FLAG, anti-HA, and anti-β-tubulin antibodies.

Residues involved in restricted budding of maRAVV VP40 map to three amino acid positions.

To determine which amino acid residue changes contribute to the loss of budding by maRAVV VP40 in 293T cells, individual residues were mutated back to the RAVV VP40 sequence. Changes at positions 57, 165, and 190 increased release of VP40 (Fig. 6).

FIG 6.

Identification of three amino acid positions as determinants of maRAVV VP40 restricted budding. A VLP assay was performed to assess budding of RAVV VP40, maRAVV VP40, and the indicated maRAVV VP40 mutants. Each of the seven mutations acquired during mouse adaptation was changed back to the RAVV VP40 sequence. 293T cells were transfected with the indicated expression plasmids. Forty-eight hours posttransfection, VLPs were purified and the cells lysed. VLPs and WCLs were analyzed for protein expression by Western blotting. VLP and WCL blots were probed with anti-FLAG, anti-HA, and anti-β-tubulin antibodies.

The identity of residue 57 modulates VP40-VP40 interaction and resistance to human tetherin.

EBOV VP40 can adopt multiple structural and oligomeric states that carry out different functions (42). To determine whether altered maRAVV VP40 budding reflects changes in VP40 oligomerization, a coimmunoprecipitation (co-IP) screen was performed to examine self-association of RAVV VP40, maRAVV VP40, and mutant maRAVV VP40s in which one or more individual residues were changed back to the parental VP40 sequence. By this analysis, RAVV VP40 coprecipitated with itself, whereas maRAVV VP40 did not (Fig. 7A). However, when residue 57 was mutated from the alanine present in maRAVV VP40 to valine, the residue present in RAVV VP40, coprecipitation was restored. Other single-residue mutants did not affect self-association. When the A57V change was combined with changes at the other positions that affected budding (Fig. 6), no further change in coprecipitation efficiency was noted (Fig. 7A). To determine whether the self-association properties of RAVV VP40, maRAVV VP40, or maRAVV VP40 A57V were cell type specific, co-IP assays were performed with Huh7 and Hepa1.6 cells. maRAVV VP40 self-interaction was abolished or significantly reduced in Huh7 and Hepa1.6 cells, respectively, and the A57V change was sufficient to restore self-association (Fig. 7B). In a reciprocal experiment, replacement of the valine at position 57 of RAVV VP40 with an alanine resulted in a loss of self-interaction in Huh7 and Hepa1.6 cells (Fig. 7B). Therefore, residue 57 modulates VP40 oligomerization.

FIG 7.

The valine at position 57 is critical for VP40-VP40 coprecipitation and for resistance to human tetherin. (A) Coimmunoprecipitation screen of VP40 protein single-point mutants. Huh7 cells were transfected with the indicated expression plasmids. Twenty-four hours posttransfection, cells were lysed and lysates immunoprecipitated with anti-FLAG antibody. Eluted proteins were analyzed by Western blotting and probed with anti-FLAG, anti-HA, and anti-β-tubulin antibodies. (B) Coimmunoprecipitation of VP40 proteins in various cell types. The assay was performed as described above with both Huh7 and Hepa1.6 cells. (C) VLP assay of RAVV V57A and maRAVV A57V VP40s in the presence of increasing amounts of human tetherin (hTetherin) in mouse cells. The assay was performed as described in the legend to Fig. 1. (D) VLP assay of RAVV V57A and maRAVV A57V VP40s in the presence of increasing amounts of mouse tetherin (mTetherin) in mouse cells. The assay was performed as described in the legend to Fig. 1.

We next determined whether VP40 homo-oligomerization, as assessed by coimmunoprecipitation assays, correlates with tetherin resistance. Hepa1.6 cells were transfected with RAVV, RAVV V57A, maRAVV, or maRAVV A57V VP40 and increasing amounts of human tetherin. Forty-eight hours posttransfection, VLPs were purified and analyzed by Western blotting. Oligomerization-competent RAVV VP40 was resistant to increasing concentrations of human tetherin, whereas maRAVV VP40 budded inefficiently, except in the absence of overexpressed tetherin. The oligomerization-defective RAVV V57A VP40 exhibited significantly reduced resistance and behaved similarly to maRAVV VP40, as VLP release was significantly inhibited by human tetherin concentrations as low as 10 ng/μl (Fig. 7C). However, oligomerization-competent maRAVV A57V gained resistance to lower concentrations of human tetherin and behaved more like RAVV VP40. Interestingly, repeating the VLP assay in the presence of increasing amounts of mouse tetherin revealed no change in VP40 resistance (Fig. 7D). Taken together, these data suggest that mouse adaptation resulted in alterations in VP40 oligomerization that influence its sensitivity to restriction by human tetherin and its ability to bud from human cell lines.

DISCUSSION

Filovirus VP40 proteins are multifunctional (32, 42, 43). Therefore, genetic changes influencing one function may impact other functions. Valmas et al. (29) demonstrated that the VP40 of mouse-adapted RAVV gained the ability to inhibit IFN signaling in mouse cells, without affecting its IFN antagonist function in human cells. Here we addressed the impact of mouse adaptation on RAVV VP40-driven VLP release. We found that after the RAVV strain was adapted to mice, budding was markedly reduced in Huh7 and 293T cells but unaltered in mouse Hepa1.6 cells. The efficient release from mouse cells indicates that maRAVV VP40 retains budding capability and that the defect is specific to the human cell lines tested.

In an effort to understand the basis for the restriction of maRAVV VP40 budding in two human cell lines that support the efficient release of both the parental RAVV VP40 and the NiV M protein, we examined the distribution of VP40 in transfected Huh7 cells. The cells expressing the restricted maRAVV VP40 exhibited many projections from their surface that resembled the “tethered” budding particles seen when EBOV VP40 was coexpressed with tetherin (24). This prompted us to examine the impact of expression of Flag-tagged tetherin on RAVV VP40 and maRAVV VP40 budding. Tetherin, an IFN-inducible type II transmembrane protein, has been shown to “tether” or prevent the release of a broad spectrum of enveloped viruses and virus-like particles from cells, including filovirus VLPs (22–27). We found maRAVV VP40 proteins egress from human 293T cells to be inhibited by tetherin. RAVV VP40, in contrast, was able to bud in the absence of overexpressed tetherin and retained some budding activity even when tetherin was overexpressed (Fig. 1 and 2). The Hepa1.6 cells, in which both RAVV and maRAVV VP40s bud efficiently, allowed us to assess the impact of overexpressed human or mouse tetherin on the budding of each VP40. In this assay, RAVV VP40 was only inhibited by large amounts of human tetherin, whereas maRAVV was exquisitely sensitive and was inhibited under all conditions where human tetherin was expressed. This conclusively demonstrates that mouse adaptation altered VP40 sensitivity to tetherin. A previous study demonstrated that mouse tetherin can inhibit budding of EBOV VP40 from 293T cells (24). In our Hepa1.6 cells, RAVV VP40 was inhibited at only the highest concentrations of mouse tetherin, and similar data were obtained with the maRAVV VP40. Altogether, these data suggest that mouse tetherin is less effective than human tetherin at inhibiting RAVV VP40 budding. The species-specific resistance to tetherin was unexpected given that tetherin functions broadly in a structure-dependent manner (40). However, there is precedent for our observation, as human tetherin, but not canine tetherin, could restrict budding of influenza A virus VLPs from MDCK cells (44). The molecular basis for the functional difference between human and mouse tetherin remains to be fully defined, but our mapping studies suggest a role for the tetherin cytoplasmic tail (Fig. 4).

We identified residue 57 of VP40 as a critical determinant of efficient budding in the human cell lines and as a determinant of resistance to human tetherin. A57V was one of three individual amino acid changes that partially restored budding of maRAVV VP40 in human cells. The identity of the residue was also demonstrated to influence the sensitivity of VP40 to restriction by overexpressed human tetherin. It is notable that previous work established that residue 57, along with residue 165, plays a central role in determining whether RAVV VP40 can effectively block IFN signaling in Hepa1.6 cells (29). Also of note, although mouse adaptation changed the tyrosine at position 19 of the RAVV VP40 late domain motif to a histidine, reversing the mutation did not restore VP40 budding (data not shown).

EBOV VP40 has been demonstrated to assume multiple oligomeric states. Because oligomerization influences EBOV VP40 function, we asked whether RAVV VP40, maRAVV VP40, or variants of maRAVV VP40 exhibit different capacities to interact with themselves in co-IP assays. maRAVV VP40 exhibited impaired self-association in this assay, relative to the parental RAVV VP40. This difference was seen in both the human and mouse cell lines tested. Further analysis demonstrated a critical role for residue 57 in VP40 self-association. When residue 57 was a valine, as in RAVV VP40, the VP40 co-IP showed interaction; when residue 57 was an alanine, as in maRAVV VP40, the VP40 co-IP failed to show efficient interaction. Previous studies on EBOV VP40 indicate that oligomerization is critical for VP40 budding (42, 43). Given that maRAVV VP40 can still bud from Hepa1.6 cells, it presumably retains oligomerization function, although we have not formally addressed this point. Nonetheless, these data suggest that maRAVV VP40, due to a mutation at position 57, has altered oligomerization properties. Given the critical role for residue 57 in IFN evasion, budding efficiency, and VP40 self-association, we propose the following model. The parental RAVV VP40 inhibits IFN responses efficiently in human cells but poorly in mouse cells (29). During mouse adaptation of RAVV, changes were selected for that enhanced VP40 IFN antagonist function. These changes altered VP40 self-association properties, perhaps facilitating interaction of VP40 with host factors relevant to suppression of IFN signaling. The changes incurred costs to VP40 in terms of budding capacity in human cells and increased sensitivity to human tetherin. These genetic changes were tolerated in the mouse system. This might reflect the inefficient inhibitory activity of mouse tetherin toward RAVV VP40 and/or the IFN-suppressing activity of VP40, which would suppress IFN-induced tetherin expression.

In the experiments described in this study, we correlate budding efficiency in the human cell lines with restriction by, or resistance to, tetherin in mouse Hepa1.6 cells. Two lines of evidence support a role for tetherin in the impaired maRAVV VP40 budding in human cells. First, expression of HIV-1 Vpu, EBOV GP, and MARV GP, all known IFN antagonists, allowed maRAVV VP40 to bud in 293T cells. Second, siRNAs targeting tetherin mRNA enhanced maRAVV VP40 budding in 293T cells. Despite the fact that 293T cells typically express little endogenous tetherin, even though expression is inducible by IFN-α (33), we detected tetherin mRNA in transfected 293T cells by qRT-PCR. Therefore, although we do not readily detect tetherin protein by Western blotting, our cells likely do express the protein. We hypothesize that tetherin may be induced by transfection in these cells, and maRAVV VP40 is likely extremely sensitive to tetherin, such that tetherin contributes to suppression of maRAVV VP40 budding.

The extent to which the amino acid changes in maRAVV VP40 influence replication of maRAVV in different cell types remains to be determined. Although budding of filoviral VP40 proteins is inhibited by tetherin, the EBOV GP has been demonstrated to antagonize the antibudding effects of tetherin (24, 27, 28). This antagonism likely explains why tetherin expression failed to block filovirus replication in cell culture (27). Therefore, it is possible that RAVV GP expression could partially or fully mitigate the effects of tetherin on maRAVV replication. Direct comparisons of parental and maRAVV replication and release in different cell types and in the presence and absence of tetherin should be performed to address this question. Even if the budding restriction seen in Huh7 and 293T cells is not entirely tetherin dependent, viral factors, such as GP and NP, which promote virus assembly and release, may facilitate replication despite alterations to VP40 function (45, 46, 47). Further characterization of the maRAVV and comparison of maRAVV and the parental RAVV should shed light on the capacity of filoviruses to switch host species and how adaptation to one host may impact viral fitness in other hosts.