Equine Viperin Restricts Equine Infectious Anemia Virus Replication by Inhibiting the Production and/or Release of Viral Gag, Env, and Receptor via Distortion of the Endoplasmic Reticulum

ABSTRACT

Viperin is an endoplasmic reticulum (ER)-associated multifunctional protein that regulates virus replication and possesses broad antiviral activity. In many cases, viperin interferes with the trafficking and budding of viral structural proteins by distorting the membrane transportation system. The lentivirus equine infectious anemia virus (EIAV) has been studied extensively. In this study, we examined the restrictive effect of equine viperin (eViperin) on EIAV replication and investigated the possible molecular basis of this restriction to obtain insights into the effect of this cellular factor on retroviruses. We demonstrated that EIAV infection of primary equine monocyte-derived macrophages (eMDMs) upregulated the expression of eViperin. The overexpression of eViperin significantly inhibited the replication of EIAV in eMDMs, and knockdown of eViperin transcription enhanced the replication of EIAV in eMDMs by approximately 45.8%. Further experiments indicated that eViperin restricts EIAV at multiple steps of viral replication. The overexpression of eViperin inhibited EIAV Gag release. Both the α-helix domain and radical S-adenosylmethionine (SAM) domain were required for this activity. However, the essential motifs in SAM were different from those reported for the inhibition of HIV-1 Gag by human viperin. Furthermore, eViperin disrupted the synthesis of both EIAV Env and receptor, which consequently inhibited viral production and entry, respectively, and this disruption was dependent on the eViperin α-helix domain. Using immunofluorescence assays and electron microscopy, we demonstrated that the α-helix domain is responsible for the distortion of the endoplasmic reticulum (ER). Finally, EIAV did not exhibit counteracting eViperin at the protein level.

IMPORTANCE In previous studies, viperin was indicated as restricting virus replications primarily by the inhibition of virus budding. Here, we show that viperin may have multiple antiviral mechanisms, including the reduction of EIAV Gag budding and Env expression, and these activities are dependent on different viperin domains. We especially demonstrate that the overexpression of viperin inhibits EIAV entry by decreasing the level of virus receptor. Therefore, viperin restriction of viruses is determined largely by the dependence of virus on the cellular membrane transportation system.

INTRODUCTION

The first line of defense against viruses is the innate immune response, and macrophages play an important role in bridging the innate and adaptive immune responses to virus infection. Innate immune responses to viral infection generally include the production of type I interferon (IFN) and the subsequent induction of hundreds to thousands of interferon-stimulated genes (ISGs). However, the function of many ISG products remains unclear and requires further investigation. Some ISGs, including viperin, have direct antiviral activity.

Viperin (virus inhibitory protein, endoplasmic reticulum-associated, interferon inducible; also known as RSAD2) is a multifunctional protein that regulates virus replication and is characterized by broad antiviral activity (1–3). Viperin inhibits the replication of a wide range of viruses, including influenza A virus (4), hepatitis C virus (HCV) (5–8), Japanese encephalitis virus (JEV) (9), West Nile virus (WNV) (10), and Dengue virus (DENV) (11, 12). Intriguingly, viperin also enhances human cytomegalovirus (HCMV) infectivity by remodeling the cellular actin cytoskeleton (13). Studies on the antiviral activity of viperin against lentiviruses currently are mostly limited to human immunodeficiency virus type 1 (HIV-1), with inconsistent results. For instance, one study reported that viperin inhibits the replication of HIV-1 by inhibiting the viral egress in cultivated cells (14), while another study reported an inhibitory effect of viperin against only specific strains of HIV-1 (15). Additional investigations are needed to better elucidate the interaction between viperin and lentiviruses.

Equine infectious anemia virus (EIAV) is a lentivirus that is similar to HIV-1. It causes a persistent infection, characterized by recurring febrile episodes associated with viremia, fever, thrombocytopenia, and wasting symptoms in equids (16). EIAV is primarily macrophage tropic. Infected macrophages act as a reservoir of the viral quasispecies and are the source of the spreading of the virus throughout the course of EIAV infection (17, 18). EIAV has been studied extensively, and in this study, the antiviral activity of equine viperin (eViperin) against EIAV infection and replication was examined. In addition, the molecular basis of eViperin activity was investigated. This study provides additional data for the characterization of the biological behavior of viperin.

MATERIALS AND METHODS

Plasmids.

Equine viperin was cloned from cDNA derived from equine monocyte-derived macrophages (eMDMs) and expressed using the adenoviral expression vector pDC315 as a fusion with a Flag tag at the N terminus (peVIP-Flag). A set of expression vectors with deletion or substitution mutations in the eViperin cDNA was constructed. The 180 C-terminal amino acid residues were deleted by PCR using the eVIP-Flag forward primer and the eVIP-182 reverse primer. The deletion of 42 amino acid residues from the N-terminal α-helix domain was performed by PCR using the eVIP-42 forward primer and the eVIP reverse primer. A 111-amino-acid fragment of the radical S-adenosylmethionine (SAM) domain between residues 71 and 182 was removed by overlapping PCR using the eVIP-71 reverse primer and the eVIP-182 forward primer. Mutagenesis PCR was used to create mutations in the SAM motifs. The S1C83A mutant was generated using the C83A forward primer and the C83A reverse primer. The S1C87A and S1C90A mutants were constructed using the S1C87A forward and reverse primers and S1C90A forward and reverse primers, respectively. The S1, S2, and S3 mutants were constructed as described previously (14). All constructed mutants were confirmed by sequencing. Codon-optimized pEIAV-Gag and pEIAV-Env plasmids derived from the EIAV molecular clone FDDV3-8 were kindly provided by Meng et al. (19). The pGagRFP plasmid was constructed using the codon-optimized EIAV-Gag plasmid as the template with the EIAV-Gag forward and EIAV-Gag reverse primers, followed by subcloning into the pDsRed2-N1 vector (Clontech, USA). EIAV-CMV-3-8 is an infectious EIAV clone derived from cell-adapted EIAV_FDDV3-8 (FDDV3-8) by replacing the U3 region of its 5′ long terminal repeat (LTR) with the cytomegalovirus (CMV) promoter (20). Tat was amplified from EIAV-CMV-3-8 and cloned into the pEGFP-N1 mammalian expression vector (Clontech, USA) with a hemagglutinin (HA) tag. The Rev expression plasmid and the Env expression plasmid (non-codon-optimized for packaging the EIAV reporter) were constructed as previously described (21). The sequences of the PCR primers are summarized in Table S1 in the supplemental material.

Cells and virus stocks.

Equine monocyte-derived macrophages (eMDMs) were prepared from equine peripheral blood mononuclear cells (PBMCs) as previously described (22) and maintained in 1640 medium containing 60% fetal bovine serum (FBS). HEK293T and HEK293 cells were maintained in Dulbecco's modified Eagle's medium containing 10% FBS. The EIAV strain, FDDV_DLV34 (DLV34), was titrated using a reverse transcriptase (RT) assay kit (Roche, Switzerland) according to the manufacturer's instructions. The AdMax system (Microbix Biosystems, Canada) was used to construct E1/E3-deleted adenovirus vectors. pDC315 vectors containing the cDNA for enhanced green fluorescence protein (EGFP) or eViperin were cotransfected with the pBHGlox△E1 and E3Cre helper plasmids in HEK293 cells according to the manufacturer's instructions. Recombinant viruses were plaque purified three times, and the 50% tissue culture infection dose (TCID₅₀) was determined by titration in HEK293 cells.

Construction of the ELR1-expressing HEK293T cell line.

The cDNA of the sole EIAV receptor, equine lentivirus receptor 1 (ELR1), was cloned into a pFUGW lentiviral vector at the BamHI and EcoRI sites using the ELR1 forward and ELR1 reverse primers (see Table S1 in the supplemental material). The resulting plasmid was sequenced and then cotransfected with the helper plasmids psPAX2 and pMD2.G (Addgene, USA) to generate pseudotyped viruses. HEK293T cells were infected with the recombinant virus. A cell line that consistently expressed ELR1 was purified by limited dilution, proliferated, and named HEK293T/ELR1.

Construction of and infection with the luciferase-expressing reporter virus.

A luciferase-expressing EIAV reporter virus was constructed by packaging the viral genome with either vesicular stomatitis virus glycoprotein (VSV-G) or EIAV Env. HEK293T cells were cotransfected with 9 μg of pONY8.1-LUC, pEIAV-GagPol, and 1 μg of pMD2.0G (for VSV-G) or the EIAV envelope (Env) expression plasmid (21). The VSV-G or EIAV Env packaged virus was collected 48 h posttransfection (hpt) and used to infect the HEK293T/ELR1 cells seeded in 48-well plates. These cells were washed and subjected to luciferase analysis 24 h postinfection (hpi).

Measurement of eViperin expression by qPCR.

eMDMs were seeded into 96-well plates and incubated at 37°C for 12 h to obtain confluent monolayers. The DLV34 EIAV strain was added to these cells at 2.5 ng RT/well and incubated for 2 h. The cells then were washed three times with serum-free RPMI 1640 medium and incubated again with 100 μl of fresh RPMI 1640 containing 60% FBS. Total cellular RNA was extracted using the RNeasy plus minikit (Qiagen, Germany) and reverse transcribed into cDNA using the reverse transcription kit (TaKaRa, Japan) according to the manufacturer's instructions. The cDNA preparations were subjected to real-time quantitative PCR (qPCR) analysis using the SYBR green PCR mixture in a Stratagene 3000 system (Stratagene, USA). The VIP forward and VIP reverse primers were used to target eViperin. The cDNA of β-actin was prepared as a housekeeping control and quantitated using qPCR with the forward and reverse ACT primers (see Table S1 in the supplemental material). The expression levels of eViperin were measured at 0, 24, 48, 72, 96, and 120 hpi with EIAV. Relative fold changes in gene expression were determined using the 2^−ΔΔCT method.

Immunostaining and immunofluorescence microscopy.

HEK293 cells were cotransfected with the EIAV Gag-RFP plasmid (pGag-RFP) and serial eViperin expression plasmids. At 48 hpt, the cells were washed with cold phosphate-buffered saline (PBS) and fixed in 4% (vol/vol) cold formaldehyde for 15 min. The formaldehyde-fixed cells were permeabilized with 0.1% (vol/vol) Triton X-100 for 5 min before blocking in 5% (wt/vol) skim milk for 2 h. The cells were immunolabeled using a mouse anti-Flag monoclonal antibody (Sigma, USA). Immunoreactivity was detected with a goat fluorescein isothiocyanate (FITC)-conjugated anti-mouse IgG (Sigma). Nuclei were labeled with 4′,6-diamidino-2-phenylindole (DAPI) (Beyotime, China), and images were acquired with a confocal microscope (TE200; Nikon, Japan).

Knockdown of eViperin expression in eMDMs by siRNA.

Three viperin-specific short interfering RNAs (siRNAs) (siViperin 1 to 3; see Table S1 in the supplemental material for the sequences) and a scramble siRNA negative control (siScr) were synthesized by RiboBio, China. Equine MDMs were seeded in 96-well plates and cultivated for 3 days. These cells were washed twice with PBS before being transfected with either viperin-specific or negative-control siRNA, which was diluted to 50 nM in serum-free media. Knockdown efficacy of eViperin mRNA was verified by real-time PCR.

Transfection and Western blotting.

Cells were transiently transfected with the indicated plasmids using Lipofectamine 2000 (Invitrogen, USA) by following the manufacturer's instructions. At 48 hpt, the cells and the culture supernatants were collected separately. The cells were washed once with PBS and then lysed in radioimmunoprecipitation assay (RIPA) lysis buffer containing a protease inhibitor cocktail (Roche). The culture supernatant was centrifuged at 12,000 × g for 10 min at 4°C to remove cell debris and centrifuged again at 20,000 × g for 2 h at 4°C to precipitate the virions. The proteins in the cell lysates and supernatant precipitates were separated by SDS-PAGE, transferred to polyvinylidene difluoride (PVDF) membranes (Millipore, Germany), and probed with the indicated antibodies. MG132 and chloroquine, inhibitors of intracellular protein degradation pathways, were purchased from Sigma.

RESULTS

EIAV infection increases eViperin expression in eMDMs, and overexpression of eViperin reduces EIAV replication.

EIAV is a lentivirus that infects equids. To characterize the restrictive activity of eViperin against the EIAV equine lentivirus, the cDNA encoding this cellular factor was cloned, and its amino acid sequence was compared to sequences of other mammalian viperins, including human (GenBank accession number NM_080657.4), rhesus macaque (HQ596987), mouse (NM_021384.4), and swine (NM_213817.1). As shown in Fig. 1A, the first 70 residues are highly variable and include the N-terminal α-helix domain. Several leucine sites are conserved, suggesting that leucine is an important factor for the function of the α-helix. In addition, three cysteine residues (CxxxCxxC) in the radical SAM domain that are important for the antiviral activity of the investigated viperins are conserved in the eViperin sequence (Fig. 1A). To determine if EIAV infection plays a role in regulating eViperin expression in eMDMs, a relative real-time reverse transcriptase-PCR (RT-PCR) analysis was performed. Total RNA extracts were collected at 0, 24, 48, 72, 96, and 120 hpi with EIAV. Relative fold changes in gene expression were determined using the 2^−ΔΔCT method. Two upregulated peaks were observed at 24 and 96 hpi, and the size of the latter peak was increased 2.8-fold (Fig. 1B).

FIG 1.

Equine viperin (eViperin) inhibits EIAV replication. (A) Amino acid sequence alignment of mammalian viperins. Different amino acids are shaded in blue. The α-helix domain, radical SAM domain, and conserved C-terminal domain are marked by lines above the sequences. (B) eViperin mRNA expression was increased post-EIAV infection. Equine monocyte-derived macrophages (eMDMs) were infected with EIAV at 2.5 ng reverse transcriptase (RT) activity. The transcription levels of viperin and β-actin were quantified by real-time PCR. The numbers of viperin mRNA copies were normalized to those of β-actin. The data represent the means ± standard errors (SE) from three independent experiments. (C) Overexpression of eViperin inhibited EIAV replication. eMDMs were infected with 2.5 ng RT of EIAV, followed by infection with viperin-expressing adenovirus 1 h later. Viruses were collected at 72 h postinfection (hpi). The EIAV replication levels were titrated by measuring RT activity. The data represent the means ± SE from three independent experiments (*, P < 0.05). P values of <0.05 were considered statistically significant. (D) Knockdown of eViperin expression eliminated the restriction of EIAV replication. eMDMs were transfected with either 50 nM viperin-specific siRNA (siViperin) or 50 nM scrambled siRNA control (siScr) or were mock transfected (Mock). At 24 hpt, viperin mRNA levels were quantified by real-time PCR. This experiment was performed three times, and means ± SE are shown. **, P < 0.01. (E) Knockdown of eViperin expression increased EIAV replication in equine macrophages. eMDMs were transfected with 50 nM siViperin 1 or siScr for 12 h and then were infected with 2.5 ng RT of EIAV. The viral titers in the supernatant were assessed at 48 hpi by measuring the RT activity. The data represent the means ± SE from three independent experiments (*, P < 0.05). P values of <0.05 were considered statistically significant.

The increase in eViperin expression in eMDMs induced by EIAV infection prompted the investigation of the interaction between this cellular factor and EIAV in the target cells. eViperin was overexpressed in eMDMs to examine its effect on EIAV replication in these cells. Because the efficacy of transfecting the expression plasmids in eMDMs was low (unpublished data), an adenoviral vector was utilized to deliver the eViperin cDNA-expressing unit into the eMDMs. EIAV then was inoculated into the eMDMs and grown in these cells with or without the expression of recombinant eViperin. The EIAV produced in these cells was measured at 72 hpi. As indicated by viral reverse transcriptase (RT) activity, the amounts of EIAV produced in eMDMs was reduced by approximately half (average 43.8%) in cells expressing recombinant eViperin at 72 hpi compared with the virus grown in cells infected with the control adenovirus (Fig. 1C).

To determine if the endogenous eViperin also has antiviral activity against EIAV, specific siRNA was synthesized to target viperin mRNA in eMDMs, resulting in an approximately 72% knockdown of gene transcription (Fig. 1D). EIAV was inoculated into cells transfected with either specific (siViperin) or nonspecific scrambled siRNA (siScr) at 12 hpt, and the EIAV replication level was measured 48 h later. As shown in Fig. 1E, the knockdown of eViperin mRNA enhanced the production of EIAV by an average of 45.8%, as indicated by viral RT activity, compared to the siScr and mock transfection controls.

Equine viperin inhibits EIAV production by suppressing virus release.

HIV-1 and EIAV Gag proteins are the major viral proteins and often are considered indicators of viral replication. These proteins are translated on cytosolic polysomes and transported to the plasma membrane by mechanisms involving various components of intracellular vesicle-trafficking pathways. These proteins are either packaged by the viral envelope protein (Env) or self-assembled as virus-like particles (VLP) in the absence of Env and then egressed (23, 24). To investigate the mechanism of eViperin inhibition of EIAV production, the viperin expression plasmid was cotransfected into HEK293T cells with EIAV-CMV-3-8, an EIAV infectious clone. The amounts of EIAV-CMV-3-8 in the cell lysate and culture supernatant were measured as the levels of Gag protein (p55) and the viral capsid (CA; p26) cleaved from p55 by Western blotting at 48 hpi. As shown in Fig. 2AA, the levels of viral structural proteins in the cell lysate were not affected by eViperin. However, the level of p26 in viral particles released into the culture supernatant was decreased by up to 56.7% in a dose-dependent manner by the expression of recombinant eViperin (Fig. 2A and E). This result indicates that EIAV release is inhibited by eViperin.

FIG 2.

Inhibition of EIAV release by eViperin requires the α-helix and SAM domains. (A) Overexpression of eViperin reduced EIAV p26 antigenic proteins in the culture supernatant but not in the cell lysate. An EIAV infectious clone, EIAV-CMV3-8 (10 ng of DNA), was transfected into HEK293T cells along with increasing amounts of eViperin expression vector (0, 0.5, 1, and 2 μg). Virus production was quantified by Western blotting of the EIAV uncleaved (p55) and cleaved (p26) Gag using an anti-EIAV p26 monoclonal antibody. Exogenous eViperin was expressed as a fusion protein with a Flag tag at the N terminus by the expression vector peVIP-Flag and was detected by an anti-Flag monoclonal antibody. This experiment was performed three times, and a representative result is shown. (B) Overexpression of eViperin inhibited the release of coexpressed EIAV Gag protein. A codon-optimized gag gene expression vector, pEIAV-Gag (10 ng), was transfected into HEK293T cells alone or with the eViperin expression vector (0, 0.5, 1, and 2 μg). The expression levels of Gag and eViperin were detected by Western blotting. This experiment was performed three times, and a representative result is shown. (C) EIAV Gag and eViperin were colocalized in the cytoplasm and on the plasma membrane. HEK293 cells were transfected with peVIP-Flag and a red fluorescent protein (RFP)-tagged expression vector, pGagRFP (the fusion protein appears red when excited by UV light). At 24 hpt, the cells were stained with an anti-Flag antibody (green), nuclei were stained with DAPI stain (blue), and expression was analyzed by confocal microscopy. (D) The three functional domains of eViperin were deleted individually. Schematic structures of eViperin with the deleted domains are shown. (E) The effect of eViperin domain deletion on the production and release of Gag was determined. HEK293T cells were cotransfected with pEIAV-Gag and a panel of plasmids expressing wild-type eViperin (viperin) or one of the three deletion constructs. At 48 hpt, the cells were lysed, and the viral particles in the culture supernatant were precipitated by ultracentrifugation. Gag proteins in the transfected cells and released in the culture supernatant were analyzed by Western blotting. This experiment was performed three times, and a representative result is shown. Gag release was quantified by densitometry scanning of Western blot images from three independent experiments. *, P < 0.05. (F) The colocalization of EIAV Gag and the eViperin mutants was examined by confocal microscopy. HEK293 cells were cotransfected with pGagRFP and vectors expressing the eViperin mutants, which were tagged with Flag at the N terminus. At 24 hpt, the cells were stained with an anti-Flag antibody (green), nuclei were stained with DAPI stain (blue), and expression was analyzed by confocal microscopy. This experiment was performed three times, and a representative result is shown.

Human viperin (hViperin) inhibits HIV-1 production by decreasing viral budding by redistributing lipid rafts; this effect was characterized as reduced viral CA (p24) in the culture supernatant but not in the cell plasma (14). Therefore, we evaluated whether eViperin inhibits EIAV replication by interrupting the release of viral CA from cells. A codon-optimized Gag expression plasmid was cotransfected into HEK293T cells with increasing doses of the eViperin expression vector. The protein levels of Gag in the cell lysate and the culture supernatant were examined by Western blotting, which revealed that EIAV Gag was progressively decreased in the supernatant but not in the cell lysate with increasing expression levels of eViperin (Fig. 2B). This result indicates impaired budding of Gag.

To determine if eViperin interrupts EIAV Gag budding by direct interaction with the viral protein, the intracellular colocalization of these two proteins was examined further. The eViperin and GagRFP expression plasmids were cotransfected into HEK293 cells, and eViperin and GagRFP expression were detected by observing the fluorescence of RFP linked to Gag and an antibody against the Flag tag fused to eViperin by confocal microscopy. The images shown in Fig. 2C clearly indicate the colocalization of these two proteins. In addition, these proteins were clustered on the plasma membrane and in the plasma, in contrast to the colocalization of hViperin and HIV-1 p24, which was associated with CD81 compartments accumulating only on the plasma membrane. These results implicate similar but distinct mechanisms of viperin inhibition of viral production in EIAV and HIV-1.

To determine which domain(s) of eViperin is crucial for the suppression of EIAV release, the α-helix domain, radical SAM domain, and conserved C-terminal domain were individually deleted (Fig. 2D). The residual inhibitory activity of the eViperin deletion constructs revealed that the α-helix domain and the SAM domain were required for the reduction of EIAV Gag levels in the culture supernatant by eViperin. The average inhibitory effects of viperin and C-terminal domain-deleted mutants were 56.7% and 59.9%, respectively, as calculated from blots from three independent experiments. Deletion of the α-helix domain and the SAM domain almost abolished the suppressive effect of eViperin (Fig. 2E). Intriguingly, deletion of the C-terminal domain did not affect the reduction of Gag levels in the supernatant (in the suppression of EIAV release) but significantly decreased the intracellular Gag content, as indicated by the weakened p55 band in the cell lysate (average of 53.8% reduction from three independent blots) (Fig. 2E).

The colocalization of these three deleted mutants with EIAV Gag also was examined. Surprisingly, although the deletion of the α-helix domain and the radical SAM domain eliminated the inhibition of the release of the Gag antigen by eViperin, none of the deletions of these three functional domains impaired the colocalization of eViperin with Gag (Fig. 2F), indicating that viperin colocalization with EIAV Gag is necessary but not sufficient to inhibit EIAV and Gag release.

The radical SAM domain is required for eViperin suppression of EIAV Gag release.

There are four SAM motifs in hViperin, each of which is responsible for the restriction of the viperin activity for at least some examined viruses (1). In previous studies, point mutations of cysteine residues in the first motif (S1) of the radical SAM domain of hViperin increased the infectivity of HCV, WNV, and DENV in HEK293 cells (7, 12). The importance of the three cysteine residues in this eViperin motif in the antiviral activity against EIAV was examined in this study. First, the three cysteine residues in S1 were replaced with alanine separately (C83A, C87A, and C90A) or together (Viperin-S1) as shown in Fig. 3A. The inhibitory activity of these mutants against EIAV was examined by detecting the EIAV p55 Gag protein in the cells that were cotransfected with plasmids expressing either wild-type or mutant eViperin and the Gag expression plasmid pEIAV-Gag. The results of the Western blotting revealed that neither single mutations nor mutation of all three cysteine residues reduced the inhibition of EIAV Gag release by eViperin (Fig. 3B), indicating that anti-EIAV activity is based on a molecular mechanism that differs from that of hViperin against HCV.

FIG 3.

S2 and S3, but not S1, motifs of the SAM domain are required for the eViperin inhibition of Gag release and the colocalization of eViperin and Gag. (A) Site mutations of the S1-S3 motifs of the radical SAM domain were constructed. The schematic structure of the eViperin protein shows the sites and residues that were substituted. Residues in red indicate those being substituted. (B) The mutation of the S1 motif did not affect the inhibition of Gag release by eViperin. HEK293T cells were cotransfected with pEIAV-Gag (10 ng) and a panel of plasmids expressing either the wild-type or S1 mutated eViperin (2 μg). At 48 hpt, the cells were lysed and the Gag in the cell lysate and the culture supernatant were analyzed by Western blotting using an anti-p26 monoclonal antibody. This experiment was performed three times, and a representative result is shown. (C) Combined mutations in the S1-S3 motifs were performed as indicated. The Gag in the cell lysate and the culture supernatant were analyzed by Western blotting. This experiment was performed three times, and a representative result is shown. Gag release was quantified by densitometry scanning of Western blot images from three independent experiments. *, P < 0.05 compared to the vector control. (D) Colocalization of EIAV Gag and eViperin containing mutations at all three motifs (S1+S2+S3) was examined. HEK293 cells were cotransfected with a Flag-tagged vector expressing the S1+S2+S3 mutated eViperin and pGagRFP (red). At 24 hpt, the cells were stained with an anti-Flag antibody (green), nuclei were stained with DAPI (blue), and the expression was analyzed by confocal microscopy. This experiment was performed three times, and a representative result is shown.

Each of the first three motifs of the SAM domain was determined to be essential for anti-HIV-1 activity based on a series of mutants of these motifs (14). This approach was employed to examine the roles of the three SAM motifs in the inhibitory effect of eViperin on EIAV Gag release. The point-mutated residues are shown in Fig. 3A. As shown in Fig. 3B, substitution of one or all three cysteine residues in the first SAM motif (S1) did not reverse the inhibition of EIAV Gag release. In contrast, mutation in the second or the third SAM motif (S2 or S3) significantly reversed the inhibition of EIAV Gag release. In addition, all combined mutations of S1, S2, and S3 abolished the eViperin-mediated suppression (Fig. 3C). Interestingly, although the S1+S2+S3 mutant was unable to inhibit Gag release, the colocalization of this mutant with the EIAV p26 antigen was identical to that of wild-type eViperin (Fig. 3D); in contrast, the S1+S2+S3 mutant of hViperin no longer colocalizes with HIV-1 Gag (14).

Overexpression of eViperin decreases the production and release of the EIAV envelope protein.

hViperin has been shown to effectively block the release of HIV-1 Gag, but no data for other HIV-1 proteins have been documented. Gag and Env are the major structural proteins of HIV-1 and EIAV, and the maturation and packaging of these two proteins are associated with the cellular vascular membranes (23, 25). Given that eViperin inhibits EIAV budding by impairing the release of Gag, we next investigated whether eViperin also affects the production and release of EIAV Env. EIAV gp140 is the precursor of the gp90 surface protein and the gp45 transmembrane protein. The effect of eViperin on the production and release of gp140 was investigated. EIAV gp140 (Env) was coexpressed with increasing doses of eViperin in HEK293T cells, and the expression of gp140 in the cell lysate and culture supernatant was examined by Western blotting. Figure 4A) shows that the intracellular level of Env was 44.7% reduced as the expression level of eViperin increased. More importantly, Env nearly disappeared from the culture supernatant of cells coexpressing eViperin. Deletion of the functional domains of eViperin revealed that the N-terminal α-helix domain, but not the SAM domain and conserved C-terminal domain, was essential for the inhibition of plasma Env expression by eViperin (Fig. 4B). eViperin impaired the expression of EIAV Env (i.e., reduced the intracellular protein level) but not Gag, and the SAM domain did not appear to be necessary for the inhibition of Env expression but was necessary for the restriction of Gag release, indicating a difference in the inhibitory mechanism. A further analysis of mRNA expression by RT-PCR revealed that the coexpression of eViperin did not impact the transcription of the EIAV env gene (Fig. 4C), implying that eViperin restricts EIAV Env production via a posttranscriptional mechanism, such as protein synthesis or degradation.

FIG 4.

eViperin retards the production and release of EIAV Env, which requires the α-helix domain. (A) The overexpression of eViperin reduced Env production and release. HEK293T cells were cotransfected with 0.5 μg of the codon-optimized Env expression vector pEIAV-Env and increasing doses of peVIP-Flag (0, 3, and 6 μg). At 48 hpt the cells were lysed, and the culture supernatants were collected to analyze Env expression by Western blotting using an equine infectious anemia (EIA)-positive serum. This experiment was performed three times, and a representative result is shown. (B) The essential functional domain for eViperin inhibition of Env production was determined using deletion mutations. HEK293T cells were cotransfected with 0.5 μg of pEIAV-Env and 6 μg of plasmids expressing either wild-type eViperin or the eViperin mutants shown in Fig. 2D. The expression levels of Env at 48 hpt were examined by Western blotting. This experiment was performed three times, and a representative result is shown. (C) Overexpression of eViperin did not interfere with the transcription of the env gene. The expression vectors pEIAV-Env (0.5 μg) and peVIP-Flag (6 μg) were cotransfected into HEK293T cells. Total RNA was extracted, digested with DNA enzyme, and used as the template for cDNA synthesis. Env- and viperin-specific cDNA fragments were amplified by semiquantitative PCR and analyzed by agarose gel electrophoresis. This experiment was performed three times, and a representative result is shown. (D) The ubiquitin-proteasome-dependent pathway and the lysosome-dependent pathway were not involved in the eViperin inhibition of Env production. Env and eViperin were transiently expressed in HEK293T cells by cotransfection of pEIAV-Env (0.5 μg) and peVIP-Flag (6 μg). At 12 hpt, the cells were treated with either 20 μM MG132 or 5 mM chloroquine (Chlo) (inhibitors of the ubiquitin-proteasome and lysosome pathways, respectively). The same volume of DMSO was used as the vehicle control for the inhibitors. The protein levels of Env and eViperin in cells at 48 hpt were analyzed by Western blotting. This experiment was performed three times, and a representative result is shown.

The ubiquitin-proteasome-dependent pathway and the lysosome-dependent pathway are two major intracellular protein degradation pathways. To determine if these two pathways are involved in the eViperin-mediated inhibition of the production and release of EIAV structural proteins, these pathways were inhibited by either MG132 or chloroquine (inhibitors of the ubiquitin-proteasome and lysosome pathways, respectively) in HEK293T cells coexpressing EIAV Env and eViperin. The Western blot results shown in Fig. 4D indicate that neither of these two inhibitors reversed the restriction of EIAV Env expression by eViperin, suggesting that translation or posttranslational modification is the step(s) that is susceptible to the eViperin-mediated inhibition of Env production.

Overexpression of eViperin does not affect the production of the nonstructural EIAV proteins Tat and Rev.

The nonstructural EIAV proteins transactivator of transcription (Tat) and regulator of expression of viral proteins (Rev) are regulators of viral replication (16, 26, 27). To determine if eViperin also inhibited the production of these EIAV accessory proteins, HEK293T cells coexpressing either EIAV Tat (see Fig. S1A in the supplemental material) or Rev (see Fig. S1B) were given increasing doses of eViperin. The protein levels of Tat and Rev in the cell lysates were detected by Western blotting. No noticeable changes in the density of the specific bands were observed, indicating that the expression of these regulatory proteins was not affected by eViperin.

Overexpression of eViperin inhibits EIAV entry by decreasing the production of the viral receptor.

Equine lentivirus receptor-1 (ELR1) is the sole receptor of EIAV. ELR1 is located on the plasma membrane and mediates EIAV entry (28). Because the transportation and release of EIAV Gag and Env are associated with the plasma membrane and the results described above demonstrated that eViperin inhibits the release of EIAV Gag and Env, we investigated whether eViperin also inhibits EIAV entry by impairing ELR1 production.

To evaluate the effect of eViperin on EIAV entry, a HEK293T cell line (HEK293T/ELR1) consistently expressing ELR1 was developed. An EIAV pseudotyped virus that contained a luciferase reporter gene and was packaged by EIAV Gag-Pol and Env, which were expressed using separate expression vectors, was generated to test the effect of eViperin on EIAV entry. rhTRIM5α, which has been reported to block EIAV uncoating in target cells (29, 30), was employed as a positive control of inhibition in the early life cycle of EIAV. HEK293T/ELR1 cells were transfected with either the expression vector for rhTRIM5α or eViperin or an empty vector control. The EIAV pseudotyped virus was inoculated 24 hpt. The luciferase activities that were detected 24 hpi demonstrated that both rhTRIM5α and eViperin significantly inhibited EIAV entry or the early life cycle compared to the vector control (Fig. 5A). The functional domains required for eViperin inhibition of EIAV entry were examined next. Consistent with the eViperin restriction of EIAV Env production and Gag and Env release described previously, the α-helix domain also was critical for EIAV entry (Fig. 5B).

FIG 5.

eViperin inhibits EIAV entry by decreasing the expression of the EIAV receptor ELR1, which requires the α-helix domain. (A) Overexpression of eViperin decreased the activity of a luciferase reporter that was inserted in a one-life-cycle EIAV pseudotyped virus (EIAV-Luc). HEK293T/ELR1 cells, which constitutively express the recombinant EIAV receptor ELR1, were transfected with plasmids expressing either eViperin or rhesus macaque TRIM5α (RhTRIM5α; as a positive control for restricting the early life stages of EIAV). An empty plasmid vector was used as the negative control. Two doses of EIAV-Luc were inoculated at 24 hpt. Cells were lysed, and the luciferase activity in the cell lysates was measured at 48 hpt. The data represent the means ± SE from three independent experiments. (B) The α-helix domain was crucial for reduction of EIAV entry by eViperin. HEK293T/ELR1 cells were first transfected with plasmids expressing either wild-type eViperin or the eViperin mutants. Two doses of EIAV-Luc were inoculated at 24 hpt. The luciferase activity was measured at 48 hpt. The data represent the means ± SE from three independent experiments. (C) Exogenous eViperin decreased ELR1 expression. HEK293T cells were cotransfected with 1 μg of ELR1 expression plasmid and increasing doses of the eViperin expression plasmid (0, 3, and 6 μg). At 48 hpt, cells were lysed and protein levels of ELR1 were analyzed by Western blotting. Protein bands then were quantified by densitometry analysis to the levels of ELR1 and β-actin. This experiment was performed three times. A representative result and the summarized data are shown. (D) eViperin inhibition of EIAV entry is EIAV Env specific. HEK293T/ELR1 cells were transfected with eViperin, RhTRIM5α (positive control), and empty vector (negative control). VSV-G-pseudotyped EIAV-Luc was inoculated at 24 hpt. The luciferase activity was measured at 48 hpt. The data represent the means ± SE from three independent experiments (*, P < 0.05; **, P < 0.01; #, P < 0.05 compared to the 40-μl control in panels A and B; NS, no significance).

Because both EIAV Env and ELR1 are associated with the plasma membrane and eViperin appears to impair Env expression via a membrane-involved mechanism, we sought to determine whether eViperin inhibits EIAV entry by decreasing ELR1 production in the same way it reduces Env expression. ELR1 was coexpressed with eViperin, and the protein levels of these two recombinant proteins were examined by Western blotting. As shown in Fig. 5C, the expression of ELR1 was largely impaired (approximately 85.4% decrease) by the expression of eViperin in a dose-dependent manner. To further confirm that the decreased production of ELR1 is involved in the eViperin inhibition of EIAV infection, a VSV-G-pseudotyped EIAV (VSV-G-EIAV) was used to infect the HEK293T/ELR1 cell line, which also was transfected with either the eViperin or the rhTRIM5α positive-control expression vector. The titers of luciferase activity shown in Fig. 5D revealed that infection with VSV-G-EIAV was significantly inhibited by rhTRIM5α but not by eViperin; this finding indicates the importance of ELR1 in infection with the EIAV reporter virus and in the inhibitory effect of eViperin against EIAV infection shown in Fig. 5A and B and demonstrates that the mechanism involves the impairment of ELR1 expression and, consequently, the inhibition of EIAV entry.

Overexpressed of eViperin disrupts the ER.

The deletion of the α-helix domain led to a change in the distribution of eViperin in the plasma from a granular distribution to an evenly diffused distribution (Fig. 6A). Previous studies have demonstrated that overexpressed viperin accumulates on the ER membrane and distorts these membranes (31). The disrupted ER forms a type of crystalloid structure, which was observed by immunofluorescence microscopy as granular structures and by electron microscopy (EM) as a lattice-like pattern (31, 32). To further determine if eViperin causes similar damage to the ER, this equine restriction factor, as well as its mutants with deletions of either the α-helix domain, the radical SAM domain, or the conserved C-terminal domain, was transiently expressed in HEK293T cells. The eViperin-expressing cells were examined by EM. Crystalloid structures were observed in the cells overexpressing either wild-type (Fig. 6B), SAM domain-deleted, or the C-terminal domain-deleted but not the α-helix domain-deleted eViperin (see Fig. S2 in the supplemental material). Higher-magnitude images showed that these crystalloid structures were ER membranes that were distorted into a lattice-like pattern, a previously described damaged structure of the ER (32). Together with the data presented in Fig. 3A, these results indicate that the overexpression of eViperin disrupts the ER and that the α-helix domain is required for this effect.

FIG 6.

Overexpression of eViperin disrupts the ER. (A) The intracellular accumulation of eViperin was altered when its α-helix domain was deleted. HEK293 cells were transfected with peVIP-Flag. At 24 hpt, the cells were stained with an anti-Flag antibody (green), nuclei were stained with DAPI stain (blue), and the cells were analyzed by confocal microscopy. All of the experiments were performed three times, and a representative result is shown. (B) Crystalloid structures in the ER in cells overexpressing eViperin were observed by electron microscopy. HEK293T cells were transfected with peVIP-Flag or empty vector and were fixed and examined using a Hitachi H-7650 electron microscope at 48 hpt. Crystalloid structures were observed in the ER of cells overexpressing wild-type (WT) eViperin. A typical lattice-like pattern of distorted ER was visible when the crystalloid structures were observed in higher-magnitude fields (shown by an enlarged box).

EIAV infection does not decrease eViperin protein levels.

Viruses have evolved to counteract host antiviral factors through various mechanisms. For example, the human and macaque APOBEC3G, a major restriction factor to retroviruses, can be degraded largely by a proteolytic pathway mediated by the Vif protein of HIV-1 and simian immunodeficiency virus (SIV), respectively. It would be of interest to determine if EIAV, a lentivirus structurally similar to HIV-1 and SIV, counteracts the restriction of eViperin by a similar approach, i.e., viral protein-mediated degradation. The results obtained from the aforementioned experiments revealed that the protein level of eViperin is positively associated with its restriction activity against the production and release of EIAV Gag and Env. Given that eViperin transcription is modulated by EIAV infection, it is difficult to evaluate the possible degradation of the cellular factor mediated by an EIAV protein during the same experiment, such as that shown in Fig. 1B. Therefore, eViperin protein levels at the indicated time point with or without EIAV proteins were determined to evaluate the counteraction of EIAV to this restriction factor. In this study, HEK293T cells were cotransfected with the eViperin expression vector and either the proviral DNA clone EIAV-CMV-3-8 or empty vector. The Western blot results demonstrated that the coexpression of the EIAV proteins and/or EIAV replication did not reduce the protein level of eViperin (Fig. 7), indicating no direct degradation effect of EIAV replication and/or proteins on eViperin.

FIG 7.

Infection of EIAV does not reduce the eViperin protein level. HEK293T cells were cotransfected with 1 μg of peVIP-Flag and 6 μg of EIAV-CMV-3-8. Cells were lysed, and the protein levels of eViperin and EIAV were analyzed by Western blotting at 48 hpt. The experiments were performed three times, and a representative result is shown.

DISCUSSION

In this study, eViperin was analyzed as an innate antiviral protein against EIAV. Our data revealed that the mRNA expression of eViperin was upregulated in eMDMs by EIAV infection. However, this upregulation was much lower than the expression of hViperin in human MDMs induced by HIV-1 (14). Studies of hViperin have demonstrated that the expression of this antiviral factor can be induced in both an IFN-dependent and an IFN-independent manner (1). Viperin is induced via an IFN-dependent pathway in infections with Solenopsis invicta virus (SINV), pseudorabies virus (PRV), and HIV-1 (9, 15, 33). However, viperin is induced by HCMV, JEV, VSV, and Chikungunya virus (CHIKV) infection via an IFN-independent pathway (9, 33–35). We previously demonstrated that type I IFNs, particularly IFN-β, were upregulated both in vivo and in vitro (eMDMs) when cells were infected with EIAV (22, 36), and the increased levels were significantly different among different EIAV strains (37). In contrast, HIV-1 inhibited the induction of type I IFN in MDMs by impairing the nuclear translocation of interferon-regulated factor 3 (14). Whether these upregulated IFNs are responsible for or involved in inducing eViperin expression in EIAV infection remains to be determined.

Lim et al. demonstrated that although viperin was positively selected during the evolution of primates, its restriction to primate lentiviruses was very limited (15). Overexpression of viperin suppressed HIV-1 Lai replication by approximately 70% but failed to suppress several other strains of HIV-1, HIV-2, and detected strains of SIV. In addition, the knockdown of endogenous viperin did not affect HIV-1 Lai infectivity. hViperin did not affect the replication of feline immunodeficiency virus (FIV) and murine leukemia virus (MLV) (15). The inhibitory effect of viperin on HCV also was modest. Specifically, HCV JFH-1 replication was approximately 45% suppressed. The siRNA treatment only reversed IFN-induced inhibition in HCV infection from 69% and 66% to 45% and 37% (5). However, viperin appeared to be much more effective at inhibiting HCMV, with a reduction in virus production of 90% (34). Viperin also reduced the PFU count of influenza type A virus from 10⁷ PFU/ml to 10⁴ PFU/ml (4). Different mechanisms of viperin antiviral activity were proposed for the aforementioned viruses, which likely explain the various restriction effects of these viruses. EIAV is thought to be the first nonprimate lentivirus susceptible to viperin restriction. Despite the moderate strength of inhibition, eViperin exhibited an average of 43.8% suppression on EIAV replication. More importantly, the knockdown of endogenous viperin increased EIAV production by approximately 45.8%. In contrast, the siRNA approach did not affect the proliferation of HIV-1 Lai, a strain sensitive to overexpressed viperin, as reported by Lim et al (15). Considering the antiviral effect of endogenous eViperin and its multiple inhibitory pathways, this cellular protein is presumed to be actively involved in host restriction to retroviruses together with other innate mechanisms.

Viperin is a radical S-adenosylmethionine (SAM) enzyme that contains three functional domains: an amphipathic α-helix domain, a radical SAM domain, and a conserved C-terminal domain. The SAM domain is most important for the original function. However, the reported roles of these domains in viperin restrictive activity differ according to the virus. In this study, we demonstrated that both the α-helix and radical SAM domains of eViperin are crucial for inhibiting EIAV Gag budding and that the α-helix but not the other two domains was essential for restricting the release of Env and the expression of Env and the EIAV receptor ELR1, both of which are membrane-associated proteins.

The α-helix domain is vital for the subcellular location of viperin and is highly variable among viperins of different species (Fig. 1A), with the exception of several leucine residues that are conserved among species, indicating their importance (via the formation of a zipper structure). No significant interference in antiviral activity against HCV, DENV, and WNV was observed when this domain was deleted (7, 12), in contrast to the results we obtained from the test on antiviral activity against EIAV as well as the results of studies of HIV-1 (14) and CHIKV (38). The relative importance of the amphipathic α-helix for eViperin inhibition of the release of EIAV structural proteins emphasizes the difference in the mechanisms of viperin restriction against different viruses. Considering that the α-helix domain is required for both the budding of EIAV Gag and the expression/release of Env and the receptor, as well as the structural change of ER, this domain is presumed to play the most important role in suppressing EIAV infection.

There are four conserved motifs in the radical SAM domain, and the first three are reported to be important in viperin antiviral activity. Motif 1 contains a conserved element comprised of three cysteine residues, CxxxCxxC, which is vital for inhibiting the infectivity of HIV-1 and HCV (7, 14). In our study, replacement of any one or all three of the cysteine residues with alanine did not abrogate the inhibition of EIAV Gag release. In addition, although point mutations in the first SAM motif of hViperin alone significantly reduced the inhibition of HIV-1 replication and Gag budding (14), this mutation did not affect EIAV Gag release in our study, indicating that the eViperin inhibition of EIAV Gag assembly and budding occurs through a mechanism that is somewhat different from that for HIV-1 Gag. Previously, studies have indicated that EIAV and HIV-1 use related but distinct pathways that mediate Gag trafficking and assembly (39–41). For instance, EIAV is unique among the retroviruses in that its assembly and budding is insensitive to proteasome inhibitor treatments that deplete intracellular free ubiquitin. However, the budding of HIV-1 and some other retroviruses is inhibited by the same proteasome inhibitors (41, 42). There is no current consensus on the sites of retrovirus assembly and budding. It has been accepted generally that lentiviral Gag assembly and budding occurs at specific sites to form viral particles. The mutation of the first three motifs of the SAM domain (S1+S2+S3) in hViperin led to the loss of its colocalization with HIV-1 Gag (14). However, the same mutation in eViperin in this study did not affect the colocalization of this restriction factor with EIAV Gag (Fig. 3D), supporting the idea that EIAV and HIV-1 Gag proteins use distinct trafficking routes during viral assembly and budding (40).

Previous studies have demonstrated that the mechanisms of the broad-spectrum antiviral function of viperin are diverse. However, it is increasingly recognized that the antiviral activity of viperin involves lipid rafts, which play an important role in the entry and assembly stages of viral replication (43, 44). Proteins are synthesized in the cytoplasm or ER depending on their target destinations. We determined that the overexpression of eViperin in eMDMs distorted the ER by observing the distinct crystalloid structures in ER by EM. Therefore, it is reasonable that the expression of EIAV Gag, Tat, and Rev, which are synthesized in the cytoplasm, was not affected by the overexpression of eViperin (Fig. 2B; also see Fig. S1A and B in the supplemental material). However, because EIAV Env and ELR1 are membrane-targeting proteins, which generally are synthesized in the ER, disruption of the ER affected the production of these proteins. EM images showed that the existence or the lack of crystalloid structures was correlated with the appearance or disappearance of clustered viperin-specific granule images when cells expressing the wild-type or mutated eViperin were examined by confocal microscopy, suggesting a correlation between clustered fluorescent granules and ER damage. Although the deletion of the SAM domain reversed the eViperin inhibition of Gag release, it did not affect the formation of clustered fluorescent granules or the inhibition of Env production and EIAV entry (indicating the level of membrane-associated ELR1) (Fig. 2E and F, 4B, and 5B). Therefore, these results indicate that a distorted ER is required for eViperin inhibition of the production of membrane-associated EIAV Env and receptor but is not sufficient for the inhibition of Gag release. These results also suggest a damaged intracellular protein transportation system and/or a reduced exporting efficacy of the cell membrane and mechanisms of inhibition of the budding of EIAV Gag, and the expression/release of Env and receptor are different.

Previous studies of viperin antiviral activity have focused on the inhibition of the transport of soluble proteins in the cell and the budding of viruses, such as HIV-1 and influenza viruses (4, 14). In this study, we demonstrated that eViperin inhibits the expression of membrane-targeting proteins, both of viral origin (EIAV Env) and host origin (ELR1). Together with reports of the interference of an HCMV envelope glycoprotein, gB, by hViperin (34), experimental data suggest that viruses with envelope and virus receptors, many of which are membrane-targeting proteins, generally are targeted by viperin.

Innate immunities, including restriction factors, are important host responses to viral infections. However, viruses have evolved to counteract host innate immunities through various approaches. For example, the Vif protein of HIV-1 antagonizes APOBEC3G (45) by mediating APOBEC3G ubiquitination; the Vpu protein of this virus antagonizes the restriction factor Tetherin by altering its normal cellular localization (46), and the Vpx protein encoded by HIV-2 and related primate lentiviruses (such as SIVsm) degrades SAMHD1 (47, 48). A previous study showed that JEV antagonizes viperin through the ubiquitination pathway (9). However, to the best of our knowledge, there have been no reports to date on the counteraction of viperin by other viruses targeted by viperin, including EIAV, which was investigated in the present study. It would be logical to investigate how EIAV evades the effect of viperin, which is a broad inhibitor of viral replication and is upregulated by infection with the virus itself. One possible explanation is that the physiological level of eViperin is relatively low, even when induced by EIAV infection (approximately 2.8-fold increase in the mRNA level), and may not be sufficient to remarkably restrict EIAV replication. In addition, our results and those of other studies indicate that viperin restriction of viral replication occurs via distortion of the ER. This also results in disruption of normal cell function and viability; therefore, it may reduce the interaction between viruses and host cells. Another possible explanation is that EIAV counteracts viperin restriction through mechanisms, such as microRNA, other than degrading viperin at the protein level. Further exploration is needed to confirm these possible alternatives.

In conclusion, our study revealed that eViperin is an innate antiviral factor restricting EIAV replication by inhibiting the production and/or release of viral structural Gag and Env proteins, as well as the expression of the EIAV receptor, apparently via a mechanism that involves distortion of the cellular ER system, as shown in Fig. 8. These results expand the understanding of the inhibition of viral infection by viperin, particularly in lentiviruses.

FIG 8.

Schematic illustration of the proposed mechanism of eViperin restriction of EIAV replication. eViperin restricts EIAV replication at multiple levels: 1, inhibition of EIAV Gag and Env budding; 2, reduction in the synthesis of EIAV Env and receptor; 3, impairment of protein trafficking.