Hepatitis B Virus Polymerase Disrupts K63-Linked Ubiquitination of STING To Block Innate Cytosolic DNA-Sensing Pathways

ABSTRACT

The cellular innate immune system recognizing pathogen infection is essential for host defense against viruses. In parallel, viruses have developed a variety of strategies to evade the innate immunity. The hepatitis B virus (HBV), a DNA virus that causes chronic hepatitis, has been shown to inhibit RNA helicase RIG-I-mediated interferon (IFN) induction. However, it is still unknown whether HBV could affect the host DNA-sensing pathways. Here we report that in transiently HBV-transfected Huh7 cells, the stably HBV-producing cell line HepAD38, and HBV-infected HepaRG cells and primary human hepatocytes, HBV markedly interfered with IFN-β induction and antiviral immunity mediated by the stimulator of interferon genes (STING), which has been identified as a central factor in foreign DNA recognition and antiviral innate immunity. Screening analysis demonstrated that the viral polymerase (Pol), but not other HBV-encoded proteins, was able to inhibit STING-stimulated interferon regulatory factor 3 (IRF3) activation and IFN-β induction. Moreover, the reverse transcriptase (RT) and the RNase H (RH) domains of Pol were identified to be responsible for the inhibitory effects. Furthermore, Pol was shown to physically associate with STING and dramatically decrease the K63-linked polyubiquitination of STING via its RT domain without altering the expression level of STING. Taken together, these observations suggest that besides its inherent catalytic function, Pol has a role in suppression of IFN-β production by direct interaction with STING and subsequent disruption of its K63-linked ubiquitination, providing a new mechanism for HBV to counteract the innate DNA-sensing pathways.

IMPORTANCE Although whether and how HBV infection induces the innate immune responses are still controversial, it has become increasingly clear that HBV has developed strategies to counteract the pattern recognition receptor-mediated signaling pathways. Previous studies have shown that type I IFN induction activated by the host RNA sensors could be inhibited by HBV. However, it remains unknown whether HBV as a DNA virus utilizes evasion mechanisms against foreign DNA-elicited antiviral signaling. In recent years, the cytosolic DNA sensor and key adaptor STING has been demonstrated to be essential in multiple foreign DNA-elicited innate immune signalings. Here, for the first time, we report STING as a new target of HBV to antagonize IFN induction and identify the viral polymerase responsible for the inhibitory effect, thus providing an additional molecular mechanism by which HBV evades the innate immunity; this implies that in addition to its inherent catalytic function, HBV polymerase is a multifunctional immunomodulatory protein.

INTRODUCTION

Hepatitis B virus (HBV) is one of the most important pathogens causing liver diseases. Worldwide, approximately 350 to 400 million individuals are chronically infected, many of whom are at increased risk of developing cirrhosis and hepatocellular carcinoma (HCC) (1, 2). Although the underlying mechanisms leading to chronic HBV infection remain to be clearly defined, the outcome of HBV infection is thought to be the result of complex interactions between replicating HBV and the host immune system (3).

The innate immunity constitutes the first line of defense against invading pathogens, which recognizes the pathogen-associated molecular patterns (PAMPs) through germ line-encoded pattern recognition receptors (PRRs). Viral infection usually activates one or more PRRs, leading to type I interferon (IFN) (including IFN-α and IFN-β) and inflammatory activities (4, 5). However, viruses, including HBV, have developed a variety of strategies to counteract the host immune responses for their survival. It has been reported that HBV surface antigen (HBs), HBV e antigen (HBeAg), and HBV virions could inhibit Toll-like receptor (TLR)-mediated production of type I IFN and proinflammatory cytokines in murine liver cells (6). In addition, HBV x protein (HBx) was reported to negatively regulate retinoic acid-inducible gene I (RIG-I)-mediated antiviral responses (7–9), while the viral polymerase (Pol) was shown to suppress type I IFN induction through impairing RIG-I- and TLR3-stimulated signaling (10, 11), both of which are RNA-sensing pathways. Considering that HBV is a DNA-containing virus with a genome size of 3.2 kb and that there are at least two types of viral DNAs distinct from the host DNA, i.e., relaxed circular DNA (rcDNA) and covalently closed circular DNA (cccDNA), during its life cycle, we thus speculate that HBV may also have strategies to interfere with the host DNA-sensing pathways.

Significant progress has been made in recent years in understanding how the innate immune system detects nonself DNA molecules or DNA-containing pathogens. Several proteins, including DNA-dependent activator of IFN regulatory factors (DAI) (12, 13), absent in melanoma 2 (AIM2) (14–16), the member of the PYHIN protein family IFI16 (17), the member of the DEXDc family of helicases DDX41 (18), and cyclic GMP-AMP (cGAMP) synthase (cGAS) (19, 20), have been identified as DNA sensors. Interestingly, the downstream signaling activated by most of these DNA sensors converges on an essential signal transducer, the stimulator of interferon genes (STING) (also known as MITA, ERIS, TMEM173, and MPYS) (21–24). STING is reported to be a direct innate immune sensor of cyclic di-GMP (c-di-GMP), a bacterial second messenger (25). Collectively, STING, functioning at the signaling “traffic junction,” plays a critical role in the regulation of the immune response to microbial nucleic acids, particularly the cytosolic DNA and cyclic dinucleotides (CDNs). However, little is known about whether and how HBV disturbs STING signaling in human hepatocytes.

Therefore, the aim of this study was to investigate the possible impact of HBV replication on the STING-mediated type I IFN induction pathway. The results revealed a novel mechanism employed by HBV to escape the innate immunity and provided evidence for a new role of the viral polymerase in the inhibition of IFN-β production through disrupting the K63-linked ubiquitination of STING.

MATERIALS AND METHODS

Plasmids and viruses.

The following expression plasmids were generously provided: pIFN-β-Luc (Rongtuan Lin, McGill University, Canada); pEF-STING and p-55C1B-Luc (Takashi Fujita, Kyoto University, Japan); pFlag-MITA, pcDNA3.1-HA-MITA, pFlag-TRIM32, and pFlag-TRIM56 (Hongbing Shu, Wuhan University, China); and pcDNA3.1-CMV-HBV1.1 (Youhua Xie, Fudan University, China), which contains the wild-type HBV 1.1-mer overlength genomic sequence. pHBV1.3, pFlag-Pol (pcDNA3.1-Flag-Pol and pQCXIP-Flag-Pol), the series of truncated mutants of Pol, pcDNA3.1-Flag-Core, pcDNA3.1-Flag-HBx, pcDNA3.1-Flag-Precore, pcDNA3.1-Flag-HBsAg, pFlag-TBK1, pHA-RIG-I, pCMV-Myc-Pol, and pcDNA3-HA-Ub were described previously (10, 26–29). pHBV1.3-ΔPol and pCMV-HBV-ΔPol, defective in viral polymerase expression, were derived from pHBV1.3 and pcDNA3.1-CMV-HBV1.1, respectively, by introducing a frameshift mutation into the Pol gene after codon 108. The mutation strategy was described previously (30). pHBV1.3-YMHD and pFlag-Pol-YMHD, lacking the polymerase catalytic active site, were derived from pHBV1.3 and pFlag-Pol, respectively, by introducing a substitution mutation into the Pol gene. The amino acid was changed from YMDD to YMHD in the polymerase sequence. The primer was GGCTTTCAGTTATATGCATGATGTGGT (the mutated nucleotide is underlined). The mutation strategy was described previously (31). The carboxy-terminally Flag-tagged Pol pPol-Flag was constructed by inserting a DNA fragment encoding three repeats of the Flag epitope tag at the carboxy terminus of Pol on the pcDNA3.1 vector.

The mutants pHA-Ub-K48R and pHA-Ub-K63 were derived from pcDNA3-HA-Ub. In mutant pHA-Ub-K48R, the lysine at position 48 was replaced with arginine by in vitro mutagenesis. In pHA-Ub-K63, all lysines except that at position 63 (including the lysines at positions 6, 11, 27, 29, 33, and 48) were mutated to arginines. pRL-TK was purchased from Promega. Newcastle disease virus tagged with green fluorescent protein (NDV-GFP) (a gift from Yan Yuan, University of Pennsylvania, PA, USA) was propagated and purified from chicken eggs maintained under specific-pathogen-free conditions.

Antibodies and reagents.

Anti-β-actin, anti-Flag M2, anti-GFP, and anti-hemagglutinin (anti-HA) antibodies were purchased from Sigma-Aldrich. Anti-c-Myc and anti-IFN regulatory factor 3 (anti-IRF3) (FL425) antibodies were obtained from Santa Cruz Biotechnology. Anti-STING antibody was obtained from Proteintech. Anti-Core, anti-HBs, and antiubiquitin (anti-Ub) antibodies were purchased from Dako, Shanghai Long Island Biotech, and Covance, respectively. Anti-TRIM32 antibody was a gift from Hongbing Shu (Wuhan University, China). The proteasome inhibitor MG-132 was purchased from Calbiochem. cGAMP and poly(dA · dT) were purchased from Biolog and Amersham Biosciences, respectively.

Cell culture and transfection.

The cell lines HEK293, HEK293T, Vero (obtained from the Cell Bank of the Chinese Academy of Science [Shanghai, China]), and Huh7 were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (FBS) (Gibco), penicillin (100 IU ml⁻¹) (Gibco), and streptomycin (100 μg ml⁻¹) (Gibco) in a 5% CO₂ atmosphere at 37°C. The HepG2-derived HBV-producing stable cell line HepAD38, which was a kind gift from Yumei Wen, was maintained as described previously (32). Additionally, HepAD38 cells were grown in presence or absence of 1 μg ml⁻¹ doxycycline (Dox) to regulate HBV pregenomic RNA transcription. PH5CH8, a simian virus 40 large T antigen-immortalized nonneoplastic human hepatic cell line with intact capacity for type I IFN induction, was maintained as described previously (33). HepaRG cells (34) were purchased from Biopredic International, France. To obtain the differentiated HepaRG (dHepaRG) cells, the cells were cultured for 2 weeks in standard medium and then for two more weeks in medium supplemented with 1.8% dimethyl sulfoxide according to the manufacturer's instructions. Primary human hepatocytes (PHHs) were purchased from Shanghai RILD Inc. (Shanghai, China). The cell culture was performed as described previously with slight modification (35). The cell pellet containing PHHs was resuspended in the plating medium of Williams E medium supplemented with 10% FBS, 5 μg ml⁻¹ transferrin, 5 ng ml⁻¹ sodium selenite, 3 μg ml⁻¹ insulin, 2 mM l-glutamine, 100 U ml⁻¹ penicillin, and 100 μg ml⁻¹ streptomycin. The cells were plated in collagen I-precoated chambers (Lab-Tek) or cell culture plates. At 5 h after plating, the medium was changed to primary hepatocyte maintenance medium (PMM), which is Williams E medium supplemented with 5% FBS, 5 μg ml⁻¹ transferrin, 10 ng ml⁻¹ epidermal growth factor (EGF), 3 μg ml⁻¹ insulin, 2 mM l-glutamine, 18 μg ml⁻¹ hydrocortisone, 40 ng ml⁻¹ dexamethasone, 5 ng ml⁻¹ sodium selenite, 2% DMSO, 100 U ml⁻¹ penicillin, and 100 μg ml⁻¹ streptomycin.

Transient transfection was performed with the indicated plasmids using Fugene HP (Roche) or Lipofectamine 2000 (Invitrogen), according to the manufacturer's instructions. Digitonin permeabilization was utilized to deliver cGAMP into cultured cells as described previously (36).

HBV infection of HepaRG cells and PHHs.

HBV infection of the hepatocytes was performed as described previously with slight modification (37). The dHepaRG cells or PHHs were incubated overnight with HBV-positive sera at approximate 100 genome equivalent copies of HBV per cell (1 volume of infectious pooled sera from 10 chronic hepatitis B (CHB) patients diluted in 10 volumes of culture medium) containing 4% polyethylene glycol (PEG) 8000. After incubation, cells were rinsed three times. The HBeAg secretion and the viral DNA in the medium were determined every 2 or 3 days postinfection (data not shown). The cells successfully infected with HBV were processed for further treatment and experiments at least 9 days postinfection.

The HBV-positive sera were collected from CHB patients in Shanghai Public Health Clinical Center with informed consent and the approval of the institutional ethics committee.

Dual-luciferase reporter assay.

Cells (1 × 10⁵) seeded in a 24-well plate were cultured overnight and then transfected with the indicated stimulator plasmid (such as pSTING), a reporter plasmid (pRL-TK, pIFN-β-Luc, or p-55C1B-Luc), and the indicated plasmids using either Lipofectamine 2000 (Invitrogen) or Fugene HP (Roche) according to the manufacturer's instructions. At 36 h posttransfection, cells were lysed with passive lysis buffer and assayed using a dual-luciferase assay kit (Promega). Data were processed as described previously (10). Data are expressed as the mean fold induction ± standard deviation (SD) relative to control levels. The results are representative of three independent experiments, each performed in triplicate.

Real-time reverse transcription-PCR (RT-PCR).

Total cellular RNA was extracted with TRIzol reagent (Invitrogen) and then subjected to reverse transcription using a kit from TaKaRa according to the manufacturer's instructions. The cDNA samples were subjected to real-time PCR using primers specific for human IFN-β or human interferon-stimulated gene 56 (ISG56) and human glyceraldehyde 3-phosphate dehydrogenase (GAPDH): IFN-β forward, 5′-GATTCATCTAGCACTGGCTGG-3′; IFN-β reverse, 5′-CTTCAGGTAATGCAGAATCC-3′; ISG56 forward 5′-TAGCCAACATGTCCTCACAGAC-3′; ISG56 reverse, 5′-TCTTCTACCACTGGTTTCATGC-3′; GAPDH forward, 5′-GGTATCGTGGAAGGACTCATGA-3′; GAPDH reverse, 5′-ATGCCAGTGGCTTCCCGTTCAGC-3′. For comparisons, transcription of IFN-β and ISG56 was normalized to that of GAPDH. Data are expressed as the mean fold induction ± SD relative to control levels. The results are representative of three independent experiments, each performed in triplicate.

IFN-β ELISA.

The level of IFN-β in the culture medium was measured using an enzyme-linked immunosorbent assay (ELISA) kit for human IFN-β (PBL Interferon Source) according to the manufacturer's instructions.

Immunoprecipitation and Western blot analysis.

For immunoprecipitation, HEK293T (1 × 10⁷) or HEK293 (4 × 10⁷) cells were transfected with various combinations of plasmids using Lipofectamine 2000 (Invitrogen) for 48 h and then lysed in immunoprecipitation buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 0.5% [vol/vol] NP-40, 10% glycerol, 1 mM EDTA) supplemented with a protease inhibitor cocktail (Roche). Cell debris was removed by centrifugation at 12,000 × g for 10 min at 4°C. The supernatants were collected, precleared with protein A/G Plus agarose beads (Santa Cruz) for 30 min, and incubated with anti-Flag or anti-HA antibodies at 4°C. After 2 h, 25 μl of a 1:1 slurry of protein A/G Plus agarose beads was added and incubated for additional 2 h. The immunoprecipitates were washed five times with lysis buffer and boiled in 1% (wt/vol) SDS sample buffer, followed by SDS-PAGE and Western blot analysis as described previously (27).

Confocal microscopy.

Immunofluorescence was performed as described previously (10). The fluorescence was observed under a confocal fluorescence microscope (Leica TCS SP2).

Southern blot analysis.

Intracellular HBV core particle-associated DNA in Huh7 cells was extracted at 48 h posttransfection, and Southern blot analysis was performed as described previously (29).

Native PAGE.

For the detection of IRF3 dimerization, 3 × 10⁵ HEK293 cells were seeded into a 12-well plate, cultured overnight, and then transfected with the indicated amounts of pEF-STING and pQCXIP-Flag-Pol (empty vector was used to balance the total amount of DNA) using Lipofectamine 2000. After 24 h, cells were selected with puromycin (1.5 μg ml⁻¹) for 36 h. The cells were harvested with 40 μl ice-cold lysis buffer (50 mM Tris-HCl, [pH 7.5], 150 mM NaCl, and 0.5% NP-40 containing 1× Roche protease inhibitors). After centrifugation at 13,000 × g for 10 min, supernatants were quantified using a bicinchoninic acid (BCA) assay (Thermo Scientific) and diluted with 5× native PAGE sample buffer (312.5 mM Tris-HCl [pH 6.8], 75% glycerol, and 0.25% bromophenol blue), and then 60 μg of total protein was applied to a prerun 6% native gel for separation. After electrophoresis, the proteins were transferred onto a nitrocellulose membrane for Western blotting.

For the detection of HBV core capsids, cells were harvested with the above-described lysis buffer at 36 h posttransfection and then centrifuged at 13,000 × g for 10 min. The quantified supernatants were run on a native agarose gel and blotted onto a nitrocellulose membrane, and the formation of core capsids was detected using the anti-Core antibody.

In vivo ubiquitination assay.

HEK293T cells in a 60-mm-diameter dish were cotransfected with the indicated plasmids using Lipofectamine 2000 (Invitrogen). For the detection of STING ubiquitination, at 36 h posttransfection, cells were treated with the proteasome inhibitor MG-132 at a final concentration of 20 μM for 6 h. The cells were washed twice with PBS and lysed with ice-cold radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris-HCl [pH 7.5]), 150 mM NaCl, 1% NP-40, 0.5% deoxycholic acid sodium, and 0.1% SDS, supplemented with a protease inhibitor cocktail [Roche] and 20 μM MG132) for 5 to 10 min. Cell debris was removed by centrifugation at 12,000 × g for 10 min at 4°C. The supernatants were divided into two aliquots. One aliquot (5%) was prepared for Western blotting. The second aliquot (95%) was sonicated five times to shear DNA. The soluble lysates were then immunoprecipitated with anti-HA antibody, followed by three washes with RIPA buffer. HA-tagged proteins were resolved by SDS-PAGE and sequentially blotted with the indicated antibodies.

Infection protection assay.

Huh7 cells were seeded in a 24-well plate and transfected with the indicated plasmids. After 36 h, the cells were mock treated or treated with 100 nM cGAMP for 16 h. The supernatants were collected and filtered with 0.22-μm filters (Millipore). Vero cells in 24-well plate were incubated with the filtered Huh7 supernatants overnight and then challenged with 80 hemagglutinin units (HAU) ml⁻¹ of NDV-GFP for an additional 24 h, followed by observation under a fluorescence microscope and analysis by Western blotting.

Statistical analysis.

Data were analyzed using Student's t test and presented as the mean ± standard deviation. A P value of <0.05 was considered to be statistically significant.

RESULTS

HBV inhibits STING-mediated IFN-β induction and the antiviral immune response.

To explore whether HBV could regulate STING-mediated type I IFN induction, we first test the overexpressed STING-stimulated IFN-β promoter activation in Huh7 cells transfected with two kinds of HBV replicative constructs, pHBV1.3 or pCMV-HBV, in which HBV replication is driven by either the HBV native promoter or the more potent cytomegalovirus (CMV) promoter, respectively. Meanwhile, we monitored the HBV replication level by detecting the formation of core capsids using a native agarose gel. As shown in Fig. 1A (left), STING-mediated IFN-β promoter activation was abrogated in a dose-dependent manner in both types of HBV replicative plasmid-transfected cells, particularly in the cells transfected with pCMV-HBV, in which a higher level of core capsids was detected (Fig. 1A, right). Interferon regulatory factor 3 (IRF3), as a key transcriptional factor, is essential for STING-stimulated signaling transduction and IFN-β induction (22, 24). To further confirm the effect of HBV on STING-mediated IFN-β induction, the effect of HBV on STING-mediated IRF3 activation was tested using the reporter plasmid p-55C1B-Luc, in which the luciferase expression was driven by a promoter containing repeated IRF-binding sites. The results showed that HBV effectively inhibited the STING-mediated IRF3 activation (Fig. 1B).

FIG 1.

HBV inhibits STING-mediated IFN-β induction and antiviral response. (A and B) Huh7 cells were cotransfected with pIFN-β-Luc (A) or p-55C1B-Luc (B), pRL-TK, and 100 ng, 200 ng, or 400 ng of the plasmid pHBV1.3 or pCMV-HBV together with pSTING (80 ng) or empty vector. After 36 h, the cells were extracted and assayed for luciferase activity. The expression of STING in cells was examined by Western blotting, and the expression of β-actin was examined as a loading control. The level of core capsids was detected using anti-Core antibody as described in Materials and Methods (A, right). (C) Measurement of endogenous STING expression in the indicated cell lines by Western blotting with anti-STING antibodies. (D) Huh7 cells were cotransfected with pIFN-β-Luc, pRL-TK, and 125 ng, 250 ng, or 500 ng of the plasmid pHBV1.3 or pCMV-HBV together with pSTING (25 ng) or empty vector. After 36 h, the cells were mock treated or treated with the indicated concentration of cGAMP for an additional 16 h and then harvested for luciferase assay. Data are expressed as the mean fold induction ± SD relative to control levels. (E and F) Huh7 cells were cotransfected with 125 ng, 250 ng, or 500 ng of pHBV1.3 together with pSTING (25 ng) or empty vector. After 36 h, the cells were mock treated or treated with 100 nM cGAMP for 16 h. The amount of IFN-β in the supernatants was determined by ELISA (E). The remaining supernatants from Huh7 cells were then collected and transferred to Vero cells for overnight incubation with a positive control using IFN-α (500 IU ml⁻¹) treatment. Vero cells were then infected with NDV-GFP (80 HAU ml⁻¹) for an additional 24 h. The GFP expression levels were determined by microscopy observation and Western blotting (F). (G) HepAD38 cells cultured in the presence or absence of doxycycline were cotransfected with pIFN-β-Luc, pRL-TK, and pSTING (25 ng) or empty vector. After 36 h, the cells were mock treated or treated with the indicated concentration of cGAMP for an additional 16 h and then harvested for luciferase assay. (H) HepaRG cells were seeded onto a chambered cover glass, proliferated, and differentiated. At day 10 after HBV infection of the dHepaRG cells, the cells were mock treated or treated with 200 nM cGAMP for an IRF3 nuclear translocation assay by immunofluorescence (left). The number of cells with or without nuclear IRF3 from HBs-negative and HBs-positive hepatocyte-like cells in three different visual fields at low magnification was counted and presented as the mean percentage ± SD (right). (I) PHHs were seeded onto a chambered cover glass or 24-well plates. At day 10 after HBV infection, the cells were either collected for immunofluorescence (left) or mock treated or treated with 200 nM cGAMP for 6 h and then harvested for the measurement of IFN-β mRNA by quantitative RT-PCR (right). The results are representative of three independent experiments.*, P < 0.05; **, P < 0.01 (based on an unpaired, two-tailed Student t test).

Cyclic GMP-AMP (cGAMP) has been identified as a specific upstream activator of STING to induce type I IFN production (19, 20, 38, 39). Because STING expression was barely detectable in Huh7 cells and HepaAD38 cells (Fig. 1C), as reported previously (40), we cotransfected the cells with STING-encoding plasmids to reconstitute the responsiveness of the cells to cGAMP when reporter plasmids and plasmids containing HBV genome or control vectors were transfected into the cells. Upon cGAMP stimulation, significant activation of the IFN-β promoter was observed; however, it was markedly impaired in the cells transfected with pHBV1.3 or pCMV-HBV (Fig. 1D). Moreover, the amount of cGAMP-triggered IFN-β production in Huh7 cells was also decreased in the presence of HBV replication in a dose-dependent manner (Fig. 1E), indicating that the cGAMP-STING-mediated signaling pathway was suppressed by HBV.

To further evaluate the inhibitory effect of HBV on the STING-mediated antiviral response, we performed a virus infection protection assay, as described previously (10). For this, the antiviral activities of supernatants harvested from IFN-α- or cGAMP-treated Huh7 cells with or without HBV replication were assayed in Vero cells challenged with Newcastle disease viral particles tagged with green fluorescent protein (NDV-GFP) by determination of the extent of green fluorescence. The results showed that NDV-GFP expression was diminished in Vero cells pretreated with the supernatants from cGAMP- and IFN-α-treated Huh7 cells (Fig. 1F, lanes and panels 4 and 8). In contrast, the supernatants from Huh7 cells transfected with the pHBV1.3 construct lost antiviral activity depending on the transfected dose, as judged by dim green fluorescence (Fig. 1F, lanes and panels 5, 6, and 7), implying that the innate antiviral response induced by cGAMP in Huh7 cells was remarkably decreased in the presence of HBV.

To better understand the interference of HBV with STING-induced type I IFN production in the stably HBV-producing cell model, we tested the cGAMP-stimulated IFN-β promoter activation in the HepaAD38 cell line, which expresses HBV under the control of an inducible tetracycline-off (Tet-off) promoter (32). The cells were pretransfected with the STING-expressing plasmid before cGAMP stimulation because, as mentioned above, the endogenous expression level of STING is rather low in HepAD38 cells. As shown in Fig. 1G, both 100 nM and 400 nM cGAMP-stimulated IFN-β promoter activation were statistically lower, almost 35%, in the cells without doxycycline than in those with doxycycline.

To further validate whether the suppression by HBV of STING signaling occurred in a natural infection of hepatocytes, differentiated HepaRG (dHepaRG) cells and PHHs infected with HBV-positive sera were mock treated or treated with cGAMP. In dHepaRG cells, the immunofluorescence results showed that the IRF3 protein was localized predominantly in the cytoplasm in the mock-treated cells. Upon cGAMP stimulation, IRF3 translocated into the nucleus in about 70% of hepatocyte-like cells, but the nuclear translocation of IRF3 was significantly reduced in those HBV-infected cells (using HBs as a positive marker for HBV infection) (Fig. 1H). In PHHs, which support a more efficient viral infection (about 30% to 40%) (Fig. 1I, left), the cGAMP-elicited IFN-β gene production in HBV-infected cells was obviously lower than that in uninfected cells (Fig. 1I, right). The data strongly support the results obtained in the overexpression system that the presence of HBV replication dampened STING-activated signaling pathways.

Overall, these results indicated that HBV may have developed some strategies to evade the STING-mediated host antiviral immune responses.

HBV polymerase (Pol) suppresses STING-induced IFN-β production.

To determine which viral proteins are responsible for the inhibition of STING-mediated IFN-β induction by HBV, we performed a functional screening assay by cotransfecting constructs encoding HBV proteins, including polymerase (Pol), Precore, Core, HBx, or HBs, together with the STING-expressing plasmid and the reporter plasmids into Huh7 cells. The results showed that only expression of HBV Pol, and not that of other viral proteins, obviously suppressed the IFN-β promoter activation induced by ectopic expression of STING in Huh7 cells (Fig. 2A), and the inhibitory effect was shown to be dose dependent by using both the amino- and carboxy-terminally Flag-tagged Pol constructs (Fig. 2B). To further confirm the suppression by Pol of the STING-mediated endogenous IFN-β production, HEK 293 cells were cotransfected with plasmids encoding HBV Pol and STING, and the cellular mRNA levels of IFN-β and interferon-stimulated gene 56 (ISG56) were quantified by real-time qPCR. The results showed that the expression of Pol resulted in decreased production of both IFN-β and ISG56 mRNAs (Fig. 2C). We next examined the effect of HBV polymerase on STING-mediated IRF3 activation in HEK293 cells by detecting the dimerization of IRF3, which is considered a hallmark of IRF3 activation and IFN induction (41). The results demonstrated that STING-induced dimerization of IRF3 was inhibited by Pol in a dose-dependent manner (Fig. 2D).

FIG 2.

HBV polymerase (Pol) suppresses STING-mediated IFN-β induction. (A) Huh7 cells were cotransfected with pIFN-β-Luc, pRL-TK, and the plasmids (300 ng) encoding Flag-tagged HBV Pol, Precore, Core, HBx, and HBs together with pSTING (120 ng) or empty vector. After 36 h, the cells were collected for luciferase assay. Expression of the plasmids encoding the indicated proteins was detected by Western blotting using anti-Flag and anti-STING antibodies (right panel; asterisks indicate the expression of the relevant proteins). (B) Huh7 cells were cotransfected with pIFN-β-Luc, pRL-TK, and increasing amounts of the plasmids pFlag-Pol (the amino-terminally Flag-tagged Pol construct) and pPol-Flag (the carboxy-terminally Flag-tagged Pol construct) together with pSTING or empty vector. After 36 h, the cells were extracted and assayed for luciferase activity. The expression of these two Pol constructs is shown (right). The full-length Pol is indicated with an asterisk. (C) 293 cells were cotransfected with Flag-tagged Pol (300 ng or 600 ng) and pSTING (200 ng). The cells were harvested at 36 h posttransfection for the measurement of IFN-β and ISG-56 mRNAs by quantitative RT-PCR. (D) 293 cells were cotransfected with pQCXIP-Flag-Pol (150 ng, 300 ng, or 450 ng) and pSTING (150 ng). At 24 h posttransfection, the cells were treated with puromycin (1.5 μg ml⁻¹) for 36 h to remove any cells without the plasmids. The cells were then harvested for the analysis of IRF3 dimerization by native PAGE. The dimer and monomer of IRF3 are indicated as “D” and “M,” respectively. (E) Huh7 cells were cotransfected with pIFN-β-Luc, pRL-TK, and Flag-tagged Pol (62.5 ng, 125 ng, or 250 ng) together with pSTING (25 ng) or empty vector. After 36 h, the cells were mock treated or treated with cGAMP (100 nM) for an additional 16 h and then harvested for luciferase assay. (F) PH5CH8 cells were cotransfected with pQCXIP-Flag-Pol (250 ng, 500 ng, or 1 μg) or pQCXIP-GFP. At 24 h posttransfection, the cells were treated with puromycin (1.5 μg ml⁻¹) for 36 h to remove the cells without the plasmids. The cells were then transfected with poly(dA · dT) (0.3 μg ml⁻¹) for an additional 8 h and then harvested for the measurement of IFN-β mRNA by quantitative RT-PCR. (G) Huh7 cells were cotransfected with pQCXIP-Flag-Pol (100 ng, 200 ng, or 300 ng) or pQCXIP-GFP and pSTING (120 ng). The assay was performed as described for Fig. 1E except the concentration of NDV-GFP was 40 HAU ml⁻¹. (H) Huh7 cells were cotransfected with pIFN-β-Luc, pRL-TK, and pHBV1.3 (500 ng) or pHBV-ΔPol (500 ng) together with pSTING (25 ng), mock treated or treated with cGAMP (100 nM) for 16 h, and then harvested for luciferase assay. (I) Huh7 cells were cotransfected with pIFN-β-Luc, pRL-TK, and increasing amounts of the plasmids pHBV1.3 and pHBV1.3-YMHD (top) or pFlag-Pol and pFlag-Pol-YMHD (bottom) together with pSTING or empty vector, After 36 h, the cells were extracted for luciferase reporter assay. −, transfection of the same amounts of empty vector plasmid with the other corresponding groups. Data are expressed as the mean fold induction ± SD relative to control levels. The results are representative of three independent experiments. *, P < 0.05; **, P < 0.01 (based on an unpaired, two-tailed Student t test).

We further confirmed the inhibitory effect of Pol on the STING-mediated IFN-β induction by using cGAMP and poly(dA · dT) as STING activators. First, we used cGAMP as an IFN-β stimulus. Consistently, Pol also inhibited STING-mediated IFN-β promoter activation upon cGAMP stimulation in a dose-dependent manner (Fig. 2E). Previous reports have showed that STING plays an important role in the regulation of IFN-β induction in response to transfected poly(dA · dT), which is a synthetic double-stranded DNA (dsDNA) that mimics dsDNA virus (21, 42). We verified that poly(dA · dT)-induced IFN-β production is mostly dependent on STING by knocking down STING expression in PH5CH8 cells, which have the intact capacity for type I IFN induction (43) (data not shown). Based on this, PH5CH8 cells expressing Pol or not were further transfected with poly(dA · dT) to activate the STING signaling. The results showed that less IFN-β mRNA was detected in cells expressing Pol than in control cells upon poly(dA · dT) stimulation (Fig. 2F). In addition, the impaired antiviral response mediated by STING was also observed in Huh7 cells expressing Pol (Fig. 2G).

To examine whether Pol could act as an IFN-β production blocker in the context of HBV replication, we compared the responses of Huh7 cells harboring the wild-type pHBV1.3 or the polymerase-null mutant pHBV1.3-ΔPol upon cGAMP stimulation. As shown in Fig. 2G, in Huh7 cells transfected with the wild-type HBV, IFN-β promoter activation induced by cGAMP was obviously reduced, while it was not much affected in those cells lacking HBV polymerase expression (Fig. 2H). Similar results were obtained in Huh7 cells transfected with either the wild-type pCMV-HBV construct or the corresponding Pol-null mutant pCMV-HBV-ΔPol (data not shown), thereby verifying the inhibitory effect of HBV Pol on the STING-mediated IFN-β production in the context of viral replication.

Considering that Pol has an indispensable role in HBV genome replication as a reverse transcriptase, we intended to explore whether Pol's catalytic activity is involved in its inhibitory effect on STING signaling. To test this point, we constructed the polymerase-inactive mutant pHBV1.3-YMHD, with which none of HBV DNA replication intermediates was detected (data not shown), and tested its effect on STING-mediated IFN-β production. We found that the mutated Pol exhibited a suppression effect on the IFN-β promoter activation that was comparable to that of the wild-type Pol in Huh7 cells (Fig. 2I, top). A similar observation was made with the mutant pFlag-Pol-YMHD (Fig. 2I, bottom). Thus, the results indicate that Pol suppresses STING signaling in a manner that is independent of its catalytic activity.

Collectively, these results demonstrate that Pol is predominantly responsible for the suppression of STING-mediated IFN-β production by HBV in human hepatic cell lines.

HBV polymerase physically associates with STING.

To gain further insight into how HBV polymerase disrupts STING signaling, we determined whether Pol interacts with STING to inhibit its function. We first examined whether Pol and STING colocalize in Huh7 cells by immunofluorescence. As shown in Fig. 3A (upper panels), Flag-tagged Pol exhibited a cytoplasmic punctate pattern of immunofluorescence which moderately colocalized with STING, implying that Pol may interact with STING. In contrast, the viral core protein did not show any colocalization with STING (Fig. 3A, lower panels). In view of these results, we further confirmed the association of Pol with STING by coimmunoprecipitation analysis. The Pol-expressing stable cell line 293T-Flag-Pol and the control cell line 293T-GFP were transfected with HA-tagged STING. The results showed that the Flag-tagged Pol, but not the GFP control, could efficiently coprecipitate with the STING protein (Fig. 3B), suggesting that Pol binds to STING. However, we did not detect any interaction between Pol and the retinoic acid-inducible gene (RIG-I), which is known as an RNA sensor that can induce IFN-β production and also be inhibited by Pol (10, 11) (Fig. 3C). These results suggested that the association between Pol and STING is specific.

FIG 3.

Pol associates with STING. (A) Huh7 cells were cotransfected with Flag-tagged Pol or Core and pSTING. After 36 h, cells were collected for immunofluorescence assay as described in Materials and Methods. Representative confocal images are shown. (B and C) Equal amounts of HA-tagged STING (B) or HA-tagged RIG-I (C) were transfected into the stable cell line 293T-Flag-Pol or 293T-GFP. The cells were harvested at 36 h posttransfection for coimmunoprecipitation analysis as described in Materials and Methods. The results are representative of three independent experiments.

The RT/RH domain of Pol binds to STING and is responsible for the inhibitory activity of Pol.

HBV polymerase is composed of four distinct regions, i.e., the terminal protein (TP), the spacer, the reverse transcriptase (RT), and RNase H (RH). To delineate the region of Pol interacting with STING, a series of truncated mutants of Pol was constructed (Fig. 4A). We cotransfected cells with the plasmids expressing Flag-tagged truncated mutants of Pol together with HA-tagged STING and examined the association between them by coimmunoprecipitation analysis. The results showed that HA-tagged STING could bind to the full-length Pol, RT/RH, RT, and RH domains of Pol but barely to the TP/spacer region. Moreover, binding of the RT/RH and RT domains with STING exhibited high affinity (Fig. 4B, upper panel).

FIG 4.

Regions of Pol responsible for binding to STING and the inhibitory effect. (A) Schematic diagram of the truncated constructs of Pol. (B) 293T cells were cotransfected with HA-tagged STING and Flag-tagged Pol or its truncated constructs TP/SP, RT/RH, RT, and RH. After 36 h, the cells were collected for coimmunoprecipitation analysis. The asterisks indicate the expression of the relevant proteins. (C and D) Huh7 cells were cotransfected with pIFN-β-Luc (C) or p-55C1B-Luc (D), pRL-TK, and the plasmids (300 ng) encoding Flag-tagged truncated constructs of HBV Pol together with100 ng of pSTING or empty vector. After 36 h, the cells were collected for luciferase assay. (E) Huh7 cells were cotransfected with pIFN-β-Luc, pRL-TK, and Flag-tagged plasmids encoding the RT or RH domain of Pol (80 ng, 160 ng, or 320 ng), together with 80 ng of pSTING or empty vector. After 36 h, the cells were collected for luciferase assay. Data are expressed as the mean fold induction ± SD relative to control levels. The results are representative of three independent experiments. *, P < 0.05; **, P < 0.01 (based on an unpaired, two-tailed Student t test).

To test whether the association of Pol with STING resulted in the inhibition of STING-mediated IFN-β production, the truncated mutants of Pol were further cotransfected with the STING-expressing and IFN-β reporter plasmids into Huh7 cells. The results showed that the IFN-β promoter activation was greatly inhibited in the cells expressing full-length Pol, RT/RH, RT, and RH but was rarely affected in the cells expressing the TP and spacer regions of Pol. The greatest inhibition, of more than 80%, was exhibited by the RT/RH region of Pol (Fig. 4C), suggesting that the protein sequence within the RT/RH region of Pol is critical to the inhibitory effect. A similar effect of the RT/RH domains on the IRF-regulating p-55C1B promoter activation is shown in Fig. 4D. To further test the inhibitory effect of the RT and RH domains on IFN-β production, we conducted a dose-response assay using different amounts of the RT and RH mutants. The results demonstrated that just 100 ng of the Pol RT domain could decrease the IFN-β promoter activation by almost 70%, while the RH mutant also had an inhibitory effect on the IFN-β promoter activation, although at a lower rate (Fig. 4E). Additionally, overexpression of the ΔRH mutant reduced the IFN-β promoter activation by approximately 50% compared to the control; in contrast, the ΔRT mutant only modestly downregulated the IFN-β promoter activation (Fig. 4C). Thus, the RT domain of HBV polymerase makes relatively more of a contribution to the inhibition of STING-mediated IFN-β induction by Pol, while the RH domain is subordinately responsible for the inhibitory effect. These results can be combined with the above data to suggest that the binding of Pol to STING was associated with its inhibitory activity on interferon induction, and the binding affinity positively correlated with the degree of the inhibitory effect. We thus concluded that Pol may interact with STING through the RT and RH domains and consequently suppresses STING-mediated IFN-β induction in the transfection system.

The K63-linked polyubiquitination of STING is drastically decreased by HBV polymerase.

Modification of signaling molecules by ubiquitination plays a central role in transmitting signals important for activation of innate antiviral responses (44). Recent studies indicate that STING polyubiquitination is required for the activation of IFN induction pathway (45, 46). Ubiquitination events structurally contain several distinct ubiquitin chain linkage types (47). The best characterized linkages to date are the K48 ubiquitin chain and the K63 ubiquitin chain (48). K48-linked polyubiquitin targets mainly proteins for proteasomal degradation, whereas K63-linked polyubiquitin more often regulates protein function, subcellular localization, or protein-protein interaction. In view of our above-described observations and to define how the interaction between Pol and STING abrogates the STING-mediated IFN-β production, we first determined if Pol impairs STING polyubiquitination by in vivo ubiquitination analysis in HEK293T cells. A dramatic decrease in STING ubiquitination was observed in cells expressing Pol compared to that in control cells (Fig. 5A). Meanwhile, the expression levels of STING in the presence and absence of HBV Pol expression remained comparable, suggesting that the linkage of STING ubiquitination impaired by Pol is most likely K63 (Fig. 1A and 5A, middle panels). To confirm this point, we employed two plasmids, i.e., pHA-Ub-K48R, in which the lysine at position 48 of the ubiquitin was replaced with arginine, and pHA-Ub-K63, in which all lysines except K63 were mutated to arginines. The cells transfected with pHA-Ub-K48R still showed a reduction in the level of ubiquitinated STING in the presence of Pol (Fig. 5B). Moreover, the K63-linked ubiquitination of STING, which has been shown to be important for the cellular antiviral response (45, 46), was markedly decreased in the cells expressing Pol, indicating that HBV Pol inhibits IFN-β production by specifically targeting the K63- rather than the K48-linked ubiquitination of STING (Fig. 5C). Subsequently, we confirmed this in the context of HBV replication using pHBV1.3 or pHBV1.3-ΔPol in Huh7 cells and found that the K63-linked ubiquitination of STING was decreased in the cells transfected with the wild-type HBV genome; however, this attenuation was rescued in the cells containing the HBV Pol-null mutant (Fig. 5D).

FIG 5.

Effect of Pol on the ubiquitination of STING. (A) 293T cells transfected with HA-tagged STING and Flag-tagged Pol or empty vector were treated with 20 mM MG132 for 6 h prior to harvest. The cell lysates were immunoprecipitated with anti-HA antibody and then subjected to Western blotting with anti-Ub to evaluate ubiquitinated proteins (upper panels). Whole-cell lysates (WCL) were blotted to evaluate expression of the indicated proteins (lower panels). β-Actin expression was examined as a protein loading control. (B and C) 293T cells were cotransfected with HA-tagged Ub-K48R (B) or Ub-K63(C) and Flag-tagged STING in the presence or absence of Myc-tagged Pol. After 36 h, the cell lysate was prepared for coimmunoprecipitation analysis as described for panel A. (D) Huh7 cells were cotransfected with HA-tagged Ub-K63, Flag-tagged STING, and pHBV1.3 or pHBV1.3-ΔPol. After 48 h, the cell lysate was prepared for coimmunoprecipitation analysis as described above. The level of HBV DNA replication intermediates was determined by Southern blotting, using a ³²P-radiolabeled HBV DNA probe. The positions of relaxed circular (RC) and single-stranded (SS) DNAs are indicated. (E) 293T cells were cotransfected with plasmids encoding the full-length or truncated constructs of Pol and Flag-tagged STING together with HA-tagged Ub-K63. The assay was performed as described for panel A. The results are representative of three independent experiments.

Next, to define which domain(s) of Pol contributes to the disruption of STING ubiquitination, we cotransfected the series of Pol truncated mutants together with STING and HA-Ub-K63 plasmids, followed by examination of STING ubiquitination. Expression of truncated mutants of Pol downregulated K63-linked ubiquitination of STING to different extents, with the greatest decrease exhibited by the full-length Pol, RT/RH domain, and RT domain, followed by the ΔRT and ΔRH mutants, and with a weak decrease exhibited by the TP/spacer and RH domains (Fig. 5E, top panel). These results were largely in agreement with our previous observations on their inhibitory effects on STING signaling (Fig. 4B to E) and suggest that the impairment of K63-linked ubiquitination of STING by Pol may depend on the binding of Pol to STING, thereby leading to a weakened IFN-β production and antiviral response.

HBV polymerase does not affect the expression of the E3 ubiquitin ligases TRIM32 and TRIM56 and their interaction with STING.

The E3 ubiquitin ligases tripartite motif protein 32 (TRIM32) and TRIM56 have been identified to target STING for K63-linked ubiquitination (45, 46). Thus, we speculated on whether Pol disrupts the K63-linked ubiquitination of STING by competitively binding to STING that associates with these two E3 ligases. To this end, we examined the impact of Pol on the interaction of TRIM32 or TRIM56 with STING by coimmunoprecipitation analysis. The results showed that the expression levels of TRIM32 and TRIM56 were comparable, regardless of the presence or absence of Pol (Fig. 6A, input). Moreover, the interaction between STING and TRIM32 or TRIM56 was not apparently affected by Pol in the overexpression system (Fig. 6A, top panel). To further confirm this result, we pulled down the endogenous TRIM32 or TRIM56 with Flag-tagged STING with or without Pol. Analysis of the immunoprecipitated products showed that Pol had little effect on the binding of STING to endogenous TRIM32 (Fig. 6B). However, we failed to detect interaction between STING and endogenous TRIM56 in 293 cells, which may be because the expression of TRIM56 in the cells is rather low (data not shown). Taken together, these data suggested the suppression of STING K63-linked ubiquitination by Pol was not mediated by a disturbance of the expression of TRIM32 or TRIM56 and their interaction with STING.

FIG 6.

Effect of HBV polymerase on the expression levels of the E3 ligases TRIM32 and TRIM56 and their interaction with STING. (A) 293T cells were cotransfected with pHA-STING and pFlag-TRIM32 or pFlag-TRIM56 in the presence or absence of pMyc-Pol. (B) 293 cells were cotransfected with pFlag-STING and pMyc-Pol or empty vector. After 36 h, the cells were harvested, and the cell lysates were immunoprecipitated with anti-Flag antibody. The coimmunoprecipitation assay was performed as described in Materials and Methods. The results are representative of three independent experiments.

DISCUSSION

It has long been controversial whether HBV induces innate immune responses during infection. A previous study using the chimpanzee model showed that HBV did not activate the host innate antiviral responses in the liver, and thus it was described as a “stealth virus” (49, 50). Nevertheless, several recent observations in clinic specimens, chimeric mouse models and infectious cell models have shown that HBV may induce some type I IFN and type III IFN production, suggesting that the host can sense the HBV infection and then initiate the innate immune responses (51–55). A possible explanation for the conflicts is that HBV has evolved numerous strategies to block the host antiviral responses early after infection (6–11, 56–59), and thus it is difficult for investigators to detect the host innate antiviral responses during infection. In the present study, we found that HBV could escape the host innate immune response via antagonizing IFN-β induction mediated by STING, a key molecule in the host cytosolic DNA-sensing pathways (17, 18, 20, 38, 60). Our results revealed, for the first time, that the HBV polymerase interacts with STING and inhibits its K63-linked ubiquitination, which is associated with loss of STING function and consequent impairment of IRF3 activation and IFN-β induction and the antiviral response.

Given the key role of STING in the regulation of the host antiviral response, many viruses have been reported to target it for subversion of the host innate immunity by various elaborate mechanisms. It has been shown that HCV, which is also a hepatotropic virus with a high rate of persistence, inhibits STING-mediated signaling through the viral NS4B protein (40, 61). The protease NS2B3 of dengue virus can cleave STING to block type I IFN induction (62, 63), and the papain-like proteases of coronaviruses interfere with the STING-mediated signaling pathway by acting as viral deubiquitinases (64, 65). For HBV, though the PRRs and related signaling pathways involved in sensing HBV infection and triggering antiviral innate immune responses remain to be clarified clearly, many reports have shown that activation of the PRR pathways specifically inhibits HBV replication (6, 26, 54, 66, 67). Moreover, DAI, a cytosolic DNA sensor that also utilizes STING as an adaptor (68), was also shown to inhibit HBV replication in our previous work (69), implying that suppression of the activation of STING, the converging point of the PRR signaling, may be an efficient strategy for HBV, like many other viruses, to evade the innate antiviral activity at an early stage.

Although several cellular biological events, such as STING ubiquitination and dimerization and its recruitment of TANK-binding kinase 1 (TBK1) (41, 70), are reported to be critical for STING-mediated signal transduction, more investigations are needed to characterize the detailed mechanisms that control STING activation and signaling (71). In this study, we demonstrated that HBV Pol markedly impaired K63-linked ubiquitination of STING by physically associating with it through the RT/RH region, though it had a moderate effect on STING dimerization, which is essential for STING self-activation and subsequent downstream signaling culminating in the induction of type I IFN (21) (data not shown). The recruitment of TBK1 and IRF3 by STING is also required for the activation of STING signaling and appears to be a process that is regulated by STING ubiquitin modification. Although K63-linked ubiquitination of STING was significantly inhibited by Pol, the results obtained in 293T cells did not show that Pol could affect the interaction between overexpressed TBK1 and STING (data not shown). Since the K63 ubiquitination of STING not only is important for STING to recruit TBK1/IRF3 but also is critical for STING as a scaffold protein to promote TBK1 and IRF3 activation, our observations that Pol-mediated disruption of K63-linked ubiquitination of STING did not affect the recruitment of TBK1 to STING but resulted in significant inhibition of IRF3 phosphorylation (data not shown) and dimerization and IFN induction were reasonable. Additionally, while the essential role of STING in DNA-sensing pathways is well established, it has also been shown to play a role in RIG-I-mediated IFN-β production (22, 24). Furthermore, HBV Pol has been reported to inhibit RIG-I-mediated signaling by disrupting the interaction between TBK1/IKKε and DDX3 (10, 11). Therefore, there is a possibility that the suppression effect of Pol on RIG-I signaling is involved in its antagonism to STING-mediated IFN-β induction. To test this hypothesis, we examined whether overexpression of DDX3 could restore the inhibitory effect of HBV polymerase on STING signaling and found that the suppressed IFN-β promoter activity could be efficiently rescued by overexpressed DDX3 when RIG-I but not STING acted as the stimulator (data not shown). This result partly excludes the possibility that Pol inhibits STING signaling to a large extent through its action on RIG-I signaling. Therefore, we concluded that HBV Pol inhibits STING-mediated type I IFN production mainly by dampening its K63-linked ubiquitination. Nevertheless, the precise delineation of how Pol affects STING-mediated signaling requires further understanding of the molecular mechanisms involved in STING activation.

HBV Pol consists of four domains. The RT domain is responsible for reverse transcription and nucleocapsid assembly of HBV (72–74), while the TP domain serves as a protein primer to initiate reverse transcription (75). The C-terminal RNase H domain directly mediates the degradation of pregenomic RNA template after reverse transcription to maintain the stability of the HBV genome. It should be noted that Pol is commonly believed to be inefficiently translated and rapidly encapsidated. However, some investigators detected nonencapsidated polymerase that accumulated in the cytoplasm in a manner similar to that for nonencapsidated duck hepatitis B virus (DHBV) polymerase (76, 77), which implies that Pol may have additional functions besides those of a reverse transcriptase. Recent reports indicate that Pol displays immunomodulatory activities through binding to a series of host factors (78). Besides blocking RIG-I- and TLR3-mediated IFN-β induction via dampening the interaction between TBK1/IKKε and DDX3 as mentioned above, Pol was also shown to abrogate IFN-α signaling by suppressing the nuclear translocation of STAT1 and STAT2 in Huh7 cells (79). Interestingly, detailed studies of the TP domain of Pol revealed its association with protein kinase C-delta (PKC-δ), resulting in inhibition of PKC-δ phosphorylation, while the RH domain was able to interfere with nuclear transportation of STAT1/2 through competitive binding to importin-α5 (80). Here, we demonstrated that the RT and RH domains of Pol were responsible for the inhibition of STING-activated signaling transduction through associating with STING and inhibiting its activation. Overall, these investigations support the idea that besides its inherent catalytic role in viral replication, Pol has multifunctional activities, including regulating host immune signaling pathways by binding to diverse host targets, which may contribute to HBV evasion of the innate immunity. Nevertheless, further work should aim to determine the significance of Pol-mediated immune modulation during HBV infection at more physiological levels of Pol expression in vivo.

In summary, a new role of HBV polymerase in inhibition of IFN-β production through selectively targeting the cytosolic DNA sensor and the adaptor molecule STING has been proposed here. We believe that these findings may promote our understanding of the molecular mechanisms by which HBV escapes the innate immunity and establishes chronic infection and may provide new target for designing novel therapeutic interventions to combat HBV.