Hepatitis E Virus Inhibits Type I Interferon Induction by ORF1 Products

ABSTRACT

Hepatitis E virus (HEV) causes both endemic and epidemic human hepatitis by fecal-oral transmission in many parts of the world. Zoonotic transmission of HEV from animals to humans has been reported. Due to the lack of an efficient cell culture system, the molecular mechanisms of HEV infection remain largely unknown. In this study, we found that HEV replication in hepatoma cells inhibited poly(I·C)-induced beta interferon (IFN-β) expression and that the HEV open reading frame 1 (ORF1) product was responsible for this inhibition. Two domains, X and the papain-like cysteine protease domain (PCP), of HEV ORF1 were identified as the putative IFN antagonists. When overexpressed in HEK293T cells, the X domain (or macro domain) inhibited poly(I·C)-induced phosphorylation of interferon regulatory factor 3 (IRF-3), which is the key transcription factor for IFN induction. The PCP domain was shown to have deubiquitinase activity for both RIG-I and TBK-1, whose ubiquitination is a key step in their activation in poly(I·C)-induced IFN induction. Furthermore, replication of a HEV replicon containing green fluorescent protein (GFP) (E2-GFP) in hepatoma cells led to impaired phosphorylation of IRF-3 and reduced ubiquitination of RIG-I and TBK-1, which confirmed our observations of X and PCP inhibitory effects in HEK293T cells. Altogether, our study identified the IFN antagonists within the HEV ORF1 polyprotein and expanded our understanding of the functions of several of the HEV ORF1 products, as well as the mechanisms of HEV pathogenesis.

IMPORTANCE Type I interferons (IFNs) are important components of innate immunity and play a crucial role against viral infection. They also serve as key regulators to evoke an adaptive immune response. Virus infection can induce the synthesis of interferons; however, viruses have evolved many strategies to antagonize the induction of interferons. There is little knowledge about how hepatitis E virus (HEV) inhibits induction of host IFNs, though the viral genome was sequenced more than 2 decades ago. This is the first report of identification of the potential IFN antagonists encoded by HEV. By screening all the domains in the open reading frame 1 (ORF1) polyprotein, we identified two IFN antagonists and performed further research to determine how and at which step in the IFN induction pathway they antagonize host IFN induction. Our work provides valuable information about HEV-cell interaction and pathogenesis.

INTRODUCTION

Hepatitis E virus (HEV) is a viral pathogen transmitted by the fecal-oral route that causes acute hepatitis with a mortality rate at or below 3% in young adults and up to 30% in pregnant women in the third trimester (1, 2, 54). While previously thought to be a public health problem only for developing countries, hepatitis E has now been recognized frequently in industrialized countries (1). Isolation of HEV from pig, chicken, mongoose, rabbit, rat, ferret, bat, fish, and deer has been reported (3–5). Zoonotic transmission of HEV from animals to humans has been documented (1) and is considered a major transmission route for sporadic cases in the industrialized countries.

HEV contains a 7.2-kb single-stranded positive-sense RNA genome, which is capped and polyadenylated (6, 54). It has been classified as the sole member of the genus Hepevirus in the family Hepeviridae (2, 6). There are four major genotypes and a single known serotype for HEV (3, 7). There are three open reading frames (ORFs) in the HEV genome (8). ORF1 encodes a polyprotein that has all the nonstructural proteins for HEV replication. ORF2 encodes the capsid protein of the HEV virion. ORF3 encodes a small multifunctional protein with a molecular mass of 13 kDa (vp13).

As an invader, HEV faces host innate immune responses, which are mainly induced by activation of host pattern recognition receptors. For recognition of RNA viruses, those receptors include RIG (retinoic-acid-inducible gene)-I-like receptors (RLRs) and Toll-like receptors (TLRs). Stimulation of the RLR and TLR signaling pathways leads to activation of transcription factors, such as interferon-regulatory factor 3 (IRF-3), IRF-7, and NF-κB. These transcription factors mediate expression of type I interferons (IFNs) and inflammatory cytokines, which not only lead to an antiviral state of the neighboring uninfected cells, but also serve as regulators to evoke an adaptive immune response. Thus, viruses have evolved many strategies to evade host innate immune responses. Little is known about how HEV evades host IFN induction. Microarray analysis of hepatitis C virus (HCV)- and HEV-infected chimpanzees showed that HEV evoked a lesser magnitude of IFN response than HCV, indicating that HEV must employ an effective strategy to dampen host innate immune responses (9).

The objective of this study was to elucidate the mechanism of HEV interference with type I IFN induction. We found that HEV replication in S10-3 hepatoma cells inhibited IFN-β induction stimulated by poly(I·C), a double-stranded RNA (dsRNA) homologue. Further studies identified two putative domains (X and PCP) of ORF1 polyprotein as the IFN antagonists. The X domain (also known as the macro domain) inhibited poly(I·C)-induced IRF-3 phosphorylation, while the PCP domain led to deubiquitination of both RIG-I and TBK-1. These findings were also confirmed in hepatoma cells with HEV replication. Our findings provide valuable information about the function of the HEV ORF1 product and improve our understanding of HEV pathogenesis.

MATERIALS AND METHODS

Cells, transfection, viruses, and chemicals.

HEK293T and HEK293 cells were maintained in Dulbecco's modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS). S10-3 cells, a subclone of Huh-7 hepatoma cells (10), were maintained in DMEM-reduced serum (DMEM-RS) supplemented with 3% FBS. Transfection of HEK293T and HEK293 cells with plasmid DNA was performed by using FuGeneHD (Promega, Madison, WI) according to the instructions of the manufacturer.

Sendai virus (ATCC VR-907) was obtained from the ATCC. Propagation of Sendai virus was done in 10-day-old embryonated specific-pathogen-free (SPF) chicken eggs. The allantoic fluid was harvested 3 days postinfection, and the Sendai virus titer was determined by hemagglutination assay, as described previously (11).

Full-length RNAs of HEV and HEV-green fluorescent protein (GFP) were obtained by in vitro transcription from plasmids pSK-E2 and pSK-E2-GFP (12), respectively, using an AmpliCap-Max T7 High Yield Message Maker Kit (Cellscript, Madison, WI). Transfection of S10-3 cells with RNA was performed using an optimized protocol with DMRIE-C reagent (Invitrogen, Grand Island, NY). Briefly, after S10-3 cells reached 70% confluence in a 12-well plate, the medium was discarded and the monolayer cells were washed twice with phosphate-buffered saline (PBS), pH 7.4, followed by addition of 0.5 ml serum-free Opti-MEM to each well. For each well, 1 μg RNA was added to 50 μl Opti-MEM and mixed evenly with 4 μl DMRIE-C. After incubation at room temperature for 20 min, the mixture was added to the S10-3 cells. One milliliter DMEM-RS with 3% FBS was added to each well 5 h posttransfection. The cells were cultured at 34.5°C.

HEK293 cells stably expressing VenusC1-IRF-3 (see below) (HEK293-IRF-3) were established by transfection of the cells with VenusC1-IRF-3 and selection under the antibiotic G418. The surviving cells were cloned by limited dilution and cell sorting by flow cytometry.

Poly(I·C) (low molecular weight), a synthetic analog of dsRNA (Invivogen, San Diego, CA), was used to induce interferon production. The 293T and S10-3 cells were transfected with poly(I·C) at a concentration of 1 μg/ml and incubated for 12 h before being harvested for further analysis.

Plasmids.

Putative ORF1 domains were PCR amplified and cloned into the vector pCAGEN (Addgene; plasmid number 11160) with a hemagglutinin (HA) tag at the N terminus as previously reported (13, 14). VenusC1-IRF-3 was constructed in house by subcloning IRF-3 from pFLAG-CMV-IRF-3 (a gift from Michael Gale, Jr.). Full-length RIG-I cDNA from S10-3 cell RNA was cloned into the KpnI site in the pCMV-Flag-MAT-Tag-1 vector (Sigma). All primers used for plasmid construction are listed in Table 1. All in-house-constructed plasmids were subjected to DNA sequencing to confirm the inserts.

TABLE 1.

Primers used in this study

Primera	Sequence (5′ to 3′)b	Target gene
H1F1	GCGAATTCGAGGCCCATCAGTTTATCAAGG	Met
H1R7	CCCTCGAGTTAAATCCAGGAGCGCAGGTTGG
H1F3	GCGAATTCAGAACCACTAAGGTTACCGG	Y
H1R6	CCCTCGAGTTACTGAGCGTAGAACTCCAAC
H1F4	GCGAATTCTGTAGGCGCTGGCTCTCGGC	PCP
H1R5	CCCTCGAGTTAGAGATTGTGGCGCTCTGG
H1F5	GCGAATTCTCTTTTGATGCCAGTCAGAG	HPX
H1R4	CCCTCGAGTTAGCCGGCACAGGCCCGGCCG
H1F6	GCGAATTCTGTCGAGTCACCCCCGGCG	Hel
H1R3	GCGATATCTTAACCAGCAAGGAAAAAGTTATTA
H1F7	GCGAATTCGGCGAAATTGGCCACCAGCG	RdRp
H1R2	GCGATATCTCATTCCACCCGACACAGAAT
H1F10	CGCTCGAGCTTTTGATGCCAGTCAGAG	HV
H1R16	GCGAATTCTAAACAGCATCAACCTCCGAC
H1F9	GCGAATTCCCTAGTCCAGCCCAGCCCG	Pro
H1R18	GCGAATTCTAGCGATGCCGGGCCGTCTGG
H1F8	GCGAATTCCCGGATGGCTCTAAGGTG	X
H1R13	CCCTCGAGTGCTGTCCGCGCAACATCC
IRF3-F2	CGGAATTCGGGAACCCCAAAGCCACGGATC	IRF-3
IRF3-R2	CCGGTACCTCAGCTCTCCCCAGGGCCCTG
RIG-I-F1	TTAGGTACCATGACCACCGAGCAGCGAC	RIG-I
RIG-I-F2	GGAGGTACCTCATTTGGACATTTCTGCTG

^aF, forward primer; R, reverse primer. “H1” before a primer name indicates the primer is based on sequences of HEV ORF1.

^bThe italicized letters indicate restriction enzyme cleavage sites for cloning.

HEV replicons pSK-E2 and pSK-E2-GFP were described previously (12, 15, 16). The construction of Myc-RIG-I(N) (N-terminal domain) (17), MDA-5(N) (18), FLAG-TBK-1 (19), and FLAG-IKKε (20) plasmids was described previously.

IFA.

An immunofluorescence assay (IFA) was carried out as previously reported (21) by using chimpanzee antibody against the HEV capsid protein. Specific antibody-capsid reactions were detected with DyLight 549 goat anti-human immunoglobulin G (IgG) conjugate (Rockland Immunologicals, Gilbertsville, PA). The cover glass was mounted on a slide using SlowFade Gold antifade reagent containing 4′6′-diamidino-2-phenylinodole (DAPI) (Invitrogen) and observed using fluorescence microscopy.

Western blot analysis.

Cells were lysed in Laemmli sample buffer. The whole proteins in the lysate were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting as previously described (21, 22). Antibodies against GFP (Santa Cruz Biotechnology, Santa Cruz, CA), phosphorylated IRF-3-S396 (Cell Signaling Technology, Danvers, MA), HA (Thermo Fisher Scientific, Waltham, MA), IRF-3 (Santa Cruz), and tubulin (Sigma-Aldrich, St. Louis, MO) were used in the blotting. The chemiluminescence signal was recorded digitally using a ChemiDoc XRS imaging system (Bio-Rad Laboratories, Hercules, CA). Digital signal acquisition and densitometry analyses were conducted using the Quantity One Program, version 4.6 (Bio-Rad).

Reverse transcription and real-time quantitative PCR (RT-qPCR).

Total RNA was isolated from cells with TRIzol reagent (Invitrogen). RNase-free DNase was used to remove carryover DNA from the RNA isolation procedure. Reverse transcription using avian myeloblastosis virus (AMV) reverse transcriptase, along with oligo(dT) and a random 15-mer, was conducted. Real-time PCR with SYBR green (Invitrogen) detection of IFN-β was done as described previously (23). Transcripts of RPL32 (ribosomal protein L32) were also detected from the same samples to serve as an internal control for normalization. Gene expression was quantified by the 2^−ΔΔCT method (24). Primers for IFN-β and RPL32 were described previously (25).

Ubiquitination assay.

Immunoprecipitation (IP) was conducted as previously described (13, 26) with modifications. Briefly, S10-3 and HEK293T cells were lysed with lysis buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 0.2 mM EDTA, 2 mM EGTA, 0.5% IGEPAL CA-630, 10% glycerol, 1 mM sodium vanadate) as described previously (13), with supplementation with a protease inhibitor cocktail (Sigma) as recommended by the manufacturer and N-ethylmaleimide (NEM) (Sigma) at a final concentration of 50 μM. The lysate was clarified by centrifugation at 14,000 × g for 5 min at 4°C. Antibodies against RIG-I or TBK-1 were added to the supernatant. IP with protein G agarose (KPL Inc., Gaithersburg, MD) was done following the manufacturer's instructions. The IP samples were subjected to immunoblot analysis with antibody against ubiquitin.

Reporter assay and luminescent cell viability assay.

HEK293T cells were transfected with the reporter pGL3.0-IFN-β promoter and testing plasmids. The Renilla vector pRL-TK (Promega) was also transfected for normalization. At 48 h after transfection, the cells were lysed for luciferase activity assay of firefly and Renilla luciferases by following the manufacturer's instructions (Promega). Lysate of the cells transfected with empty vector was used as a control for calculation of the IFN-β promoter activation level. The relative firefly luciferase activity is shown after normalization with the Renilla level as described previously (27). A cell viability assay was conducted with a CellTiter-Glo Luminescent Cell Viability Assay kit by following the manufacturer's instructions (Promega).

Statistical analysis.

Differences in indicators between treatment samples, such as the IFN-β mRNA level between the group in the presence of HEV replication and the control sample, were assessed by the Student t test. A two-tailed P value of less than 0.05 was considered significant.

RESULTS

HEV replication inhibits IFN-β expression induced by poly(I·C) transfection.

We first tested whether HEV replication would interfere with interferon induction. S10-3 cells, a subclone of Huh-7 cells that is more susceptible to transfection with whole HEV genomic RNA (28), were used. HEV genomic RNA was transcribed from the pSK-E2 plasmid containing whole cDNA of the Sar55 (genotype I) HEV genome, as previously described (12). Full-length HEV RNA generated from the pSKE2-GFP construct, in which insertion of GFP disrupts the expression of ORF2 and ORF3 (12), was included for comparison. S10-3 cells were transfected with the two different HEV RNAs. HEV replication was detected by immunofluorescence assay with HEV antibody (Fig. 1A). The cells were transfected with poly(I·C) 10 days post-HEV RNA transfection to stimulate IFN production. RT-qPCR was then conducted to detect IFN-β transcript.

FIG 1.

HEV replication in S10-3 cells (a subclone of Huh-7 cells) reduces IFN production induced by poly(I · C). (A) HEV replication in S10-3 cells detected by immunofluorescence assay. The cells were transfected with full-length HEV genomic RNA from pSK-E2. The red fluorescence (left) indicates detection of the HEV ORF2 protein. Nuclear DNA (right) was counterstained using DAPI. (B) Reduction of IFN-β expression in HEV-infected S10-3 cells detected by RT-qPCR. The cells were transfected with poly(I · C) (PIC). Relative levels (fold) of IFN-β mRNA in comparison to control cells without HEV RNA transfection and without PIC stimulation are shown. The graph represents averages of three independent experiments. Significant differences from the HEV-negative group are denoted by asterisks (**, P < 0.01). (C) Cell viability assay comparing S10-3 cells with or without HEV replication and PIC transfection. Relative percentages are shown in comparison with normal S10-3 cells without PIC treatment. The error bars indicate standard error.

Compared with uninfected cells, S10-3 cells with HEV replication (from pSK-E2 RNA) had a significantly lower level of IFN-β transcript (3.2-fold) (Fig. 1B), which indicates that HEV replication reduced IFN expression evoked by poly(I·C). Similarly, the cells transfected with RNA from pSKE2-GFP had a significantly lower level of IFN-β transcript (4.8-fold). This result suggests that the ORF1 product was at least partially responsible for the inhibition, as GFP insertion interrupts the expression of both ORF2 and ORF3 in pSKE2-GFP.

To ensure that the impaired IFN-β mRNA level was not due to a cytotoxic effect induced by poly(I·C) or HEV replication, we performed a cell viability assay of mock-infected or HEV-infected S10-3 cells with or without poly(I·C) stimulation. The results showed that HEV replication or poly(I·C) stimulation had a minimal effect on the viability of S10-3 cells (Fig. 1C), which suggests the IFN-β mRNA reduction was due to inhibition by HEV replication.

The HEV ORF1 product inhibits IFN production.

After finding that the inhibition of IFN induction was potentially due to the ORF1 product, we cloned 6 fragments of ORF1 into the pCAGEN plasmid with an HA tag at the N terminus according to a previous analysis of the putative domains (29) (Fig. 2A). They were the methyltransferase domain (Met); the Y domain (Y); papain-like cysteine protease (PCP); a fragment covering the hypervariable region, the proline-rich domain, and the X domain (HPX); helicase (Hel); and RNA-dependent RNA polymerase (RdRp). Protein expression of these HA-tagged fragments in HEK293T cells transfected with the plasmids was confirmed with Western blotting using HA antibody (Fig. 2B). All the proteins except HPX appeared to be one band. The HPX lane had two bands, indicating excision of the product.

FIG 2.

Screening of ORF1 products for potential IFN antagonists. (A) Schematic illustration of ORF1 products. Met, methyltransferase domain; Y, Y domain; PCP, papain-like cysteine protease; HPX, hypervariable region, proline-rich domain, and X domain; Hel, helicase; RdRp, RNA-dependent RNA polymerase. The numbers above the boxes indicate the numbers of amino acids of the ORF1 polyprotein. (B) Cloning and expression of ORF1 fragments in HEK293T cells detected by Western blotting. The fragments were expressed as HA-tagged proteins. (C) Inhibition of poly(I · C)-induced IFN-β expression by ORF1 fragments in HEK293T cells detected by RT-qPCR. The cells were transfected with plasmids of the ORF1 fragments and then transfected with poly(I · C). At 12 h after poly(I · C) treatment, the cells were harvested for RNA isolation and real-time PCR to detect IFN-β transcript. PRRSV nsp1β was included as a control. Relative levels (fold) in comparison with untreated control cells are shown. EV, empty vector. Significant differences from the EV control are denoted by asterisks (**, P < 0.01). The graph represents averages of three independent experiments. The error bars indicate standard errors. (D) Cell viability assay comparing HEK293T cells transfected with HEV ORF1 domains. Relative percentages are shown in comparison with normal cells transfected with EV. (E) Inhibition of Sendai virus-evoked IFN-β expression by HEV PCP and HPX domains. HEK293T cells were transfected with PCP and HPX plasmids, along with IFN-β luciferase reporter and a Renilla control plasmid. The cells were infected with Sendai virus at an MOI of 1 48 h after transfection. Luciferase activity was assayed 48 h after infection and is shown as a relative level (fold) in comparison with the EV control. The graph represents the averages of three independent experiments.

Based on the size of the excised HPX band, it appeared that the excision occurred in the X domain. Recently, a highly conserved “glycine triad” (G815-G816-G817) was identified in the macro domain of HEV, which is homologous to the rubella virus protease substrate (G1299-G1300-G1301) (30). It is possible that the two bands of HPX product in HEK293T cells were due to the cellular protease-mediated cleavage at this glycine triad.

Next, we examined the effects of the proteins on IFN expression in HEK293T cells. An empty vector and porcine reproductive and respiratory syndrome virus (PRRSV) nonstructural protein 1β (nsp1β), which is known to inhibit IFN induction (31), were included as controls. At 48 h after transfection, the cells were transfected with poly(I·C) to stimulate IFN induction. RT-qPCR was conducted to detect the IFN-β mRNA level. Compared with the cells transfected with empty vector, PCP and HPX led to significant reductions in IFN-β expression (2.6- and 7.6-fold, respectively) (Fig. 2C). Likewise, PRRSV nsp1β led to a 4.6-fold reduction. However, the other fragments of ORF1 did not have an inhibitory effect on IFN-β induction, indicating that ORF1 PCP and HPX were IFN antagonists.

To exclude the possibility of a cytotoxic effect of HPX and PCP, we conducted cell viability assays for HEK293T cells transfected with empty vector and HEV ORF1 domains. The results showed that transfection with PCP and HPX plasmids had an undetectable cytotoxic effect (Fig. 2D). This also confirmed that the reduced IFN-β mRNA level in the cells with PCP and HPX was due to the antagonist functions of the two proteins.

To further confirm that HEV PCP and HPX are real IFN antagonists, we used Sendai virus, a well-known natural inducer of type I IFNs, to stimulate the HEK293T cells transfected with PCP and HPX, along with IFN-β reporter plasmids. The reporter assay showed that PCP and HPX significantly inhibited Sendai virus-induced IFN-β promoter activation (3.8- and 5.5-fold) (Fig. 2E). Taken together, our data indicate than HEV PCP and HPX function as antagonists to type I IFN induction.

The HEV X domain inhibits IFN induction by blocking phosphorylation of IRF-3.

After finding that the two ORF1 products inhibit IFN induction, we first selected HPX to determine its interference mechanism. Since overexpression of signal molecules of RLR pathways, such as RIG-I (N-terminal domain), MDA-5 (N-terminal domain), TBK-1, and IKKε, leads to the activation of the IFN-β promoter (31–34), HEK293T cells were transfected with plasmids encoding the IFN-β reporter, HPX, and the specific components of the RIG-I signaling pathway. Overexpression of the individual components was needed to determine which step in the RLR signaling HEV HPX blocks. Compared with the cells transfected with empty vector, HPX expression resulted in a significant reduction of RIG-I- and MDA-5-induced IFN-β reporter expression (1.8- and 2.2-fold, respectively) (Fig. 3A). Similarly, when the cells were transfected with a TBK-1 or IKKε plasmid, HPX reduced the luciferase yield significantly (2.26- and 2.35-fold, respectively) (Fig. 3B). PRRSV nsp1β reduced the luciferase yield of the IFN-β reporter, as expected. These results indicated that HPX interfered with IFN induction via the RLR pathway.

FIG 3.

Reporter assay showing that the ORF1 HPX product inhibits IFN-β induction. (A) The HPX product inhibits RIG-I- or MDA-5-activated IFN-β induction. HEK293T cells were transfected with IFN-β reporter plasmid, along with RIG-I(N) (N-terminal domain) or MDA-5(N) and nsp1β or HPX DNA. Firefly and Renilla luciferase activities were measured at 48 h posttransfection. PRRSV nsp1β was included as a positive control. EV, empty vector. Significant differences from the EV control are denoted by asterisks (**, P < 0.01). The graph represents the averages of three independent experiments. The error bars indicate standard errors. (B) HPX fragments inhibit TBK-1- or IKKε-activated IFN-β induction. Transfections were done as for panel A, with the replacement of RIG-I or MDA-5 with TBK-1 or IKKε. Calculations and presentation are done similarly to those in panel A.

As HPX inhibits the RLR pathway, we tested the activation status of IRF-3 in cells expressing HPX. Since the phosphorylation level of endogenous IRF-3 in HEK293T cells after poly(I·C) stimulation is hard to detect, we cotransfected the cells with FLAG-IRF-3 and HPX plasmids. At 40 h after transfection, the cells were transfected with poly(I·C) and harvested 8 h later for Western blotting. Compared with cells transfected with empty vector, the cells with HPX had much lower levels of IRF-3 phosphorylation (Fig. 4A), while the levels of total IRF-3 and tubulin were similar between the samples. This result indicated that HPX interfered with IRF-3 activation.

FIG 4.

Inhibition of IRF-3 phosphorylation by the X domain. (A) Expression of an HPX fragment inhibits IRF-3 phosphorylation. HEK293T cells were transfected with IRF-3 and empty vector (EV) or HPX plasmids and then transfected with poly(I · C) (PIC). At 8 h post-PIC treatment, the cells were harvested for detection of IRF-3 phosphorylation by Western blotting with antibody against phosphorylated IRF-3 (pIRF-3). Total IRF-3 and HPX expression was detected with IRF-3 and HA antibodies, respectively. (B) Identification of the X domain in inhibition of IRF-3 phosphorylation. HEK293-VenusC1-IRF-3 stable cells were transfected with HV, Pro, or X domain plasmids, and Western blotting was done as for panel A. The molecular mass marker for the HA blot is shown on the right.

The HPX region contains 3 putative domains, the hypervariable (HV), proline-rich (Pro), and X domains. To determine which domain was capable of antagonizing IRF-3 activation, we cloned the three domains individually into the pCAGEN expression vector, which encodes an HA tag at the N terminus of the target protein. When the plasmids were coexpressed with FLAG-IRF-3 in 293T cells, expression of IRF-3 was very low. Thus, FLAG-IRF-3 expression was interfered with in the cells transfected with the plasmids with HV and Pro domains, while endogenous IRF-3 expression was not affected (data not shown). To resolve the issue of plasmid interference in coexpression, we established HEK293 cells stably expressing VenusC1-IRF-3, which makes it easy to observe IRF-3 expression and to select positive clones. Phosphorylation of VenusC1-IRF-3 after poly(I·C) stimulation in the stable cells in the presence of the HV, Pro, or X domain was examined by Western blotting. The X domain was identified as the inhibitor of IRF-3 phosphorylation, while the HV and Pro domains had minimum effect (Fig. 4B).

The PCP domain deubiquitinates both RIG-I and TBK-1.

During our screening for the IFN antagonists from the ORF1 domains, the papain-like cysteine protease domain also inhibited poly(I·C)-induced IFN expression. It was reported that the HEV Met-PCP polyprotein had deubiquitination activity for ubiquitin-, SUMO-, and ISG15-conjugated cellular proteins (35). We reasoned that PCP could potentially lead to deubiquitination of both RIG-I and TBK-1, as ubiquitination of these two molecules is essential for their activation (36). To test this, we cotransfected 293T cells with plasmids expressing PCP and RIG-I or TBK-1, since the levels of endogenous RIG-I and TBK-1 are too low to be detected in Western blotting. Samples were immunoprecipitated with antibodies to RIG-I or TBK-1, and the ubiquitination status was assessed by Western blotting with ubiquitin antibody. Decreased ubiquitination levels of both RIG-I and TBK-1 were observed in cells cotransfected with PCP (3.8 and 3.2-fold, respectively) in comparison with the empty-vector control (Fig. 5A and B), while the total RIG-I and TBK-1 levels were not affected in whole-cell lysate. These results suggested that PCP may inhibit RIG-I and TBK-1 activation via its deubiquitinase activity.

FIG 5.

Deubiquitination of RIG-I and TBK-1 by the PCP domain. (A) RIG-I deubiquitination by PCP. HEK293T cells were transfected with PCP and RIG-I plasmids. At 48 h posttransfection, the cells were harvested for immunoprecipitation with FLAG antibody to pull down RIG-I. Western blotting (WB) with antibodies against ubiquitin (Ub) and RIG-I was conducted. The relative levels of ubiquitination signal intensity are shown below the blots after normalization with RIG-I. Whole-cell lysate (WCL) was used to detect ubiquitin, RIG-I, PCP, and tubulin. The molecular mass marker for the WCL blot is shown on the right. (B) TBK-1 deubiquitination by PCP. The experiment was done as for panel A with the exception that the TBK-1 plasmid was used.

HEV replication in hepatoma cells inhibits IRF-3 phosphorylation and ubiquitination of RIG-I and TBK-1.

The results described above showed that overexpression of the X and PCP domains of HEV ORF1 in HEK293T cells led to inhibition of IRF-3 phosphorylation and RIG-I and TBK-1 ubiquitination, respectively. We reasoned that the HEV ORF1 product could perform the same functions during whole-virus infection in hepatoma cells. To test this, we employed the HEV replicon pSK-E2-GFP. S10-3 cells that were transfected with RNA from E2-GFP and cultured for 10 days were stimulated with poly(I·C) for 10 h. Compared to the cells without HEV RNA, the cells with E2-GFP replication had much lower poly(I·C)-induced IRF-3 phosphorylation, while the total IRF-3 level had minimum change (Fig. 6A). This result was consistent with the findings for X domain expression in HEK293 cells.

FIG 6.

HEV replication leads to downregulation of IRF-3 phosphorylation and deubiquitination of RIG-I and TBK-1 in S10-3 cells. (A) Reduction of PIC-induced IRF-3 phosphorylation in S10-3 cells transfected with HEV RNA from the pSK-E2-GFP replicon. The cells were treated with PIC to induce IRF-3 activation. At 8 h post-PIC treatment, the cells were harvested for detection of IRF-3 phosphorylation by Western blotting with antibody against pIRF-3. Total IRF-3, GFP, and tubulin were also detected. (B) Deubiquitination of RIG-I in S10-3 cells with HEV replication. The cells were harvested for immunoprecipitation with RIG-I antibody. Western blotting with antibodies against ubiquitin and RIG-I was conducted. The relative levels of ubiquitination signal intensity are shown below the blots after normalization with RIG-I. Whole-cell lysate (WCL) was used to detect ubiquitin, RIG-I, PCP, and tubulin. The molecular mass marker for the WCL blot is illustrated on the right. (C) Deubiquitination of TBK-1 in S10-3 cells with HEV replication. The relative levels of ubiquitination signal intensity are shown below the blots after normalization with TBK-1. The ubiquitin level in the WCL is the same as in panel B.

In the S10-3 cells transfected with E2-GFP RNA, the ubiquitination levels of RIG-I and TBK-1 were also examined. IP results showed that both RIG-I and TBK-1 in S10-3 cells transfected with E2-GFP RNA had lower ubiquitination levels than the control without HEV replication (Fig. 6B and C). Again, the results were consistent with the observation of PCP overexpression in HEK293T cells. Endogenous RIG-I in S10-3 cells was also detectable, and no change in its expression was observed between control cells and E2-GFP RNA-transfected cells (Fig. 6B). There was no change between the samples for the total cellular ubiquitination level in the whole-cell lysate. However, endogenous TBK-1 in the whole-cell lysate of S10-3 cells was below detection level in Western blotting, though it was detectable after being enriched by IP (Fig. 6C). These results demonstrated that HEV replication inhibited poly(I·C)-induced IFN production by blocking phosphorylation of IRF-3 and ubiquitination of RIG-I and TBK-1.

DISCUSSION

Type I IFNs, such as IFN-α and IFN-β, are critical for innate immunity against viral infection and contribute to the modulation of adaptive immunity (37). Viruses have developed a variety of strategies to subvert or evade the innate immune response. IRF-3, a critical transcription factor, is frequently targeted by viruses to interfere with IFN induction. VP35 of Ebola virus inhibits IRF-3 activation, and this inhibition is related to the viral virulence (38). PRRSV nsp1β inhibits IRF-3 activation (31). Npro of classical swine fever virus induces proteasome-mediated degradation of IRF-3 (39, 40). In this study, we demonstrated that HEV replication inhibited IFN production. Furthermore, the ORF1 product was responsible for this inhibition. Out of 8 putative domains in the HEV ORF1 product, the PCP and X domains were demonstrated to inhibit RIG-I-mediated signaling in different steps. Sendai virus was used to evoke IFN induction, and the results confirm the inhibition of the HEV domains. PCP mediates the deubiquitination of RIG-I and TBK-1, while the X domain inhibits IRF-3 phosphorylation. A model is proposed to illustrate the HEV inhibition in the RLR pathway of IFN induction (Fig. 7).

FIG 7.

Interference in type I IFN production by HEV proteins. Activation of the RLR pathway and signaling by viral dsRNA are shown. Viral dsRNA is generated during HEV replication. “P” indicates phosphorylation. The red boxes indicate HEV proteins identified in this study as inhibiting the signaling steps indicated. The HEV X domain inhibits IRF-3 phosphorylation. HEV PCP inhibits K63-linked polyubiquitination of RIG-I and TBK-1. MAVS, mitochondrial antiviral signaling.

The X domain of HEV is known as a macro domain due to its homology with the C-terminal nonhistone domain of histone macroH2A, and such domains have been identified in a variety of bacterial, archaeal, and eukaryotic organisms (reviewed in reference 41). Studies of human macro domains indicate they possess a DNA binding motif and are involved in DNA repair, chromatin remodeling, and transcriptional regulation. Little information about viral macro domains is available. The coronavirus macro domains were shown to possess relatively poor ADP-ribose 1″-phosphohydrolase activities, implying they are functionally different from their human homologues (42).

In our study, the HEV X domain was found to inhibit IRF-3 phosphorylation, which appears to be a new function for a viral macro domain. However, we were unable to detect interaction between the X domain and IRF-3 or the upstream kinase TBK-1 in IP experiments (data not shown). Thus, the mechanism through which the X domain inhibits IRF-3 phosphorylation needs to be further elucidated. In addition, we found that the X domain of HEV could bind tightly with chromosomal DNA when overexpressed in HEK293T cells (data not shown). The human macro domains have a DNA binding motif and are involved in downregulation of gene activation, suggesting that the HEV X domain could potentially regulate host gene expression, such as IFN-stimulated genes, which may further dampen IFN signaling. However, due to the lack of an X antibody and an efficient cell culture system for HEV, we do not know whether the X domain performs the proposed function during HEV replication.

Another IFN antagonist we identified from HEV ORF1 is the PCP domain, which has moderate similarity to the protease domain of rubella virus (29). While the PCP domain encoded by rubella virus has been shown to be responsible for the proteolytic processing of its nonstructural protein (43), the role of the HEV PCP in proteolytically processing the ORF1 product into small subunits has yet to be confirmed. Recently, the connection of ubiquitination and activation of the IFN induction pathway was defined (reviewed in reference 44). Since cysteine proteases represent a large family of deubiquitinases (45), this implies that virus-encoded cysteine proteases could function as deubiquitinases to inhibit ubiquitination-dependent activation of host IFN induction pathways. Evidence from other viruses shows that virus-encoded cysteine proteases, such as arterivirus papain-like protease 2 (46) and the leader proteinase of foot-and-mouth disease virus (47), indeed possess a deubiquitinase function and inhibit host innate immunity.

In a previous report, the HEV Met-PCP polyprotein, which contains the Met, Y, and PCP domains, was found to act as a deubiquitinase for ubiquitin-, SUMO-, and ISG15-linked proteins in a cell-free system (35). However, Met-PCP was unable to cleave an LXGG motif, which should be recognized by cellular or viral PCP. In this study, we demonstrate that the PCP domain could act as the deubiquitinase for RIG-I and TBK-1 in HEK293T cells. In addition, our data indicate that HEV replication in S10-3 cells could lead to deubiquitination of RIG-I and TBK-1. A study of the foot-and-mouth disease virus indicates the deubiquitinase activity of the viral leader protease does not rely on its proteolytic activity (47). It is not known whether HEV PCP deubiquitinase activity relies on its protease activity. Since a Zn-binding finger is required for protease activity of PCP (48), a recent bioinformatics study identified the putative Zn-binding finger in the PCP domain as Cys457-His458-Cys459 and Cys481-Cys483 (49). Moreover, a recent report determined that the HEV PCP domain acts as a chymotrypsin-like protease when purified and renatured from a bacterial expression system (50).

What is more, the severe acute respiratory syndrome (SARS) virus processes a papain-like cysteine protease that is capable of cleaving ubiquitin- and ISG15-conjugated proteins and helping the virus evade host innate immunity (51, 52). One study demonstrated that targeting the viral cysteine protease with a virus-specific inhibitor blocks SARS virus replication while not affecting cellular deubiquitinases (53). This implies that the deubiquitinase function of viral PCP could be a potential therapeutic target.

In conclusion, our study identified the PCP and X domains of HEV ORF1 polyprotein as the IFN antagonists, and they inhibit RIG-I-mediated signaling in different steps, ubiquitination of RIG-I and TBK-1 and IRF-3 phosphorylation, respectively. This provides valuable information for understanding HEV pathogenesis and facilitating future development of antiviral drugs.