The US3 Kinase of Herpes Simplex Virus Phosphorylates the RNA Sensor RIG-I To Suppress Innate Immunity

ABSTRACT

Recent studies have demonstrated that the signaling activity of the cytosolic pathogen sensor retinoic acid-inducible gene-I (RIG-I) is modulated by a variety of posttranslational modifications (PTMs) to fine-tune the antiviral type I interferon (IFN) response. Whereas K63-linked ubiquitination of the RIG-I caspase activation and recruitment domains (CARDs) catalyzed by TRIM25 or other E3 ligases activates RIG-I, phosphorylation of RIG-I at S8 and T170 represses RIG-I signal transduction by preventing the TRIM25-RIG-I interaction and subsequent RIG-I ubiquitination. While strategies to suppress RIG-I signaling by interfering with its K63-polyubiquitin-dependent activation have been identified for several viruses, evasion mechanisms that directly promote RIG-I phosphorylation to escape antiviral immunity are unknown. Here, we show that the serine/threonine (Ser/Thr) kinase US3 of herpes simplex virus 1 (HSV-1) binds to RIG-I and phosphorylates RIG-I specifically at S8. US3-mediated phosphorylation suppressed TRIM25-mediated RIG-I ubiquitination, RIG-I-MAVS binding, and type I IFN induction. We constructed a mutant HSV-1 encoding a catalytically-inactive US3 protein (K220A) and found that, in contrast to the parental virus, the US3 mutant HSV-1 was unable to phosphorylate RIG-I at S8 and elicited higher levels of type I IFNs, IFN-stimulated genes (ISGs), and proinflammatory cytokines in a RIG-I-dependent manner. Finally, we show that this RIG-I evasion mechanism is conserved among the alphaherpesvirus US3 kinase family. Collectively, our study reveals a novel immune evasion mechanism of herpesviruses in which their US3 kinases phosphorylate the sensor RIG-I to keep it in the signaling-repressed state.

IMPORTANCE Herpes simplex virus 1 (HSV-1) establishes lifelong latency in the majority of the human population worldwide. HSV-1 occasionally reactivates to produce infectious virus and to facilitate dissemination. While often remaining subclinical, both primary infection and reactivation occasionally cause debilitating eye diseases, which can lead to blindness, as well as life-threatening encephalitis and newborn infections. To identify new therapeutic targets for HSV-1-induced diseases, it is important to understand the HSV-1-host interactions that may influence infection outcome and disease. Our work uncovered direct phosphorylation of the pathogen sensor RIG-I by alphaherpesvirus-encoded kinases as a novel viral immune escape strategy and also underscores the importance of RNA sensors in surveilling DNA virus infection.

INTRODUCTION

The innate immune system provides a rapid and often effective defense mechanism against invading pathogens. It is composed of several classes of pattern-recognition receptors (PRRs) that, upon recognition of invading microbes, elicit expression of antiviral molecules and proinflammatory cytokines that create an environment hostile to pathogen infection and prevent spread. PRRs can be distinguished based on the cellular compartments in which they reside and the pathogen-associated molecular patterns (PAMPs) that they recognize (1). Among the PRRs, RIG-I-like receptors (RLRs) are a family of cytoplasmic sensors essential for innate immune detection of viral infection in nearly all cell types (2).

The well-characterized PAMPs recognized by the two principal RLR members, RIG-I and MDA5, are short dsRNA molecules containing 5′-tri or di-phosphate modifications and long dsRNA molecules, respectively (reviewed in references 3–5). These nucleic acid species are not typically present in uninfected host cells but are found in the genomes of many RNA viruses or arise during their life cycle as replication intermediates. Additionally, mislocalized and/or unmasked host RNAs were recently found to serve as potent RLR agonists during DNA virus infections, in particular for HSV-1, Epstein-Barr virus (EBV), and Kaposi's sarcoma-associated herpesvirus (KSHV) (6–9). In line with these observations, RIG-I and its adaptor protein mitochondrial antiviral signaling protein (MAVS) have been shown to play a pivotal role in the type I interferon (IFN) response to several herpesviruses, including HSV-1 and KSHV, as well as hepatitis B virus (HBV) and adenoviruses (6, 8, 10–12). Furthermore, it has been shown that in rig-I knockout mouse embryonic fibroblasts (MEFs) the induction of IFN-β in response to HSV-1 is reduced compared to the induction in wild-type (WT) MEFs (13). Therefore, while RLRs have long been believed to specifically detect RNA virus infections (e.g., influenza viruses, paramyxoviruses, and flaviviruses), growing evidence suggests that RLRs and in particular RIG-I also play an important role in the innate immune response to DNA viruses including herpesviruses, and in turn, herpesviruses have evolved a variety of RLR-evasion mechanisms (14–17). Thus, the RLRs critically contribute to herpesvirus sensing in addition to the well-characterized DNA sensors cGAS and IFI16, which recognize viral DNA genomes and induce signaling via the adaptor protein STING (reviewed in references 18 and 19), as well as the membrane-bound Toll-like receptors (TLR3 and TLR9), which sense dsRNA and unmethylated CpG DNA, respectively. Furthermore, extensive cross talk between the different signaling pathways has recently been reported, and cGAS, TLRs, and RLRs likely play temporal and/or cell-type-specific roles in immune surveillance of herpesvirus infection (20, 21).

RIG-I and MDA5 are composed of three domains: two amino-terminal tandem caspase activation and recruitment domains (CARDs), a central DExD/H box RNA helicase domain, and a carboxy-terminal domain (22, 23). The CARDs mediate RIG-I and MDA5 interactions with their common downstream interaction partner, MAVS, and the helicase and C-terminal domains are important for the recognition of RNA ligands and RIG-I autorepression. RLR activation upon PAMP recognition triggers a signaling cascade that leads to the recruitment and activation of several effector proteins, resulting in the phosphorylation, dimerization, and activation of the transcription factors interferon regulatory factor (IRF) 3 and IRF7, which translocate from the cytoplasm to the nucleus where they mediate transcriptional activation of type-I IFNs and other antiviral or proinflammatory genes. Type-I IFNs can act directly on cellular processes and also upregulate a large set of ISGs that carry out a variety of antiviral functions, including the suppression of viral protein synthesis and the induction of apoptosis (24).

Because the RLR-mediated antiviral response is characterized by high levels of cytokine production and the activation of a diverse set of genes that control essential cellular processes, aberrant activation of RLR signaling can be severely detrimental to the host. Many points of regulation therefore exist along the RLR signaling pathway to maintain a balance between viral clearance and host preservation (reviewed in references 25 and 26). Post-translational modifications of the RIG-I and MDA5 CARDs are a key mechanism by which RLR activity is regulated (27). In uninfected cells, RIG-I and MDA5 are maintained in an inactive state by several phosphorylations including those at S8/T170 and S88, respectively (28–33). In the case of RIG-I, both S8 and T170 were identified as potential target phosphorylation sites for the Ser/Thr kinases PKA and PKC by bioinformatics approaches (30). However, detailed biochemical analysis revealed that PKC-α and PKC-β are the primary kinases that phosphorylate S8 and T170 of RIG-I (30). The kinase(s) that phosphorylate(s) MDA5 S88 has not yet been identified. Upon virus infection, RLR activation requires the removal of these inhibitory phosphorylation marks by the protein phosphatases PP1α/γ (31). In the case of RIG-I, dephosphorylation is followed by K63-linked (non-proteasome-targeting) polyubiquitination of K172 by tripartite motif-containing protein 25 (TRIM25) or other E3 ligases (2, 28, 34). In the case of MDA5, dephosphorylation by PP1 induces ISGylation of the CARDs (35). Polyubiquitination (RIG-I) or ISGylation (MDA5) facilitates RLR multimerization and translocation to mitochondria-associated membranes, where they interact with MAVS (34, 36, 37). These regulatory steps are essential for RLR activation and also serve as targets for viral antagonism. Whereas many studies have focused on elucidating the mechanisms by which RNA viruses antagonize RLRs (38, 39), much less is known about specific RLR-evasion mechanisms employed by DNA viruses such as HSV-1.

In this study, we identify the HSV-1 Ser/Thr kinase US3 as a RIG-I antagonist that inhibits the type-I IFN response by phosphorylating the critical regulatory residue S8 in RIG-I. A recombinant HSV-1 encoding a catalytically-inactive mutant US3 induced a more robust RIG-I-dependent cytokine response compared to the parental virus or a revertant virus. Our results further suggest that the US3-dependent RIG-I evasion mechanism is conserved among the alphaherpesvirus US3 kinase family.

RESULTS

The HSV-1 US3 serine/threonine kinase induces phosphorylation of RIG-I at regulatory residue S8.

Given the crucial role of dephosphorylation in the activation of RIG-I and MDA5 (31), we hypothesized that virus-encoded kinases may counteract the induction of an innate immune response by phosphorylating the RLRs to keep them in an inactive state. Examination of the RIG-I S8/T170 and MDA5 S88 phosphorylation sites revealed that the sequence surrounding RIG-I S8 conforms to the consensus phosphorylation sequence described for HSV-1 US3—which resembles the target sequences of cellular PKA and PKC (40)—while the RIG-I T170 and MDA5 S88 sequences do not fit this motif (Fig. 1A). Therefore, to determine the effect of HSV-1 US3 on the phosphorylation state of these regulatory residues in RIG-I and MDA5, we tested the effect of US3 coexpression on the phosphorylation of S8 and T170 (RIG-I) using phospho-specific antibodies that recognize the individual phosphorylated regulatory residues. We observed that, in the presence of US3, phosphorylation of RIG-I 2CARD at S8 was strongly enhanced, while phosphorylation at T170 remained unchanged (Fig. 1B and C). In line with the current model that RIG-I phosphorylation precludes the K63-linked ubiquitination of RIG-I by TRIM25, we also observed that the enhanced phosphorylation of RIG-I 2CARD in the presence of ectopically expressed US3 correlated with diminished ubiquitination (Fig. 1B and C). In contrast, the phosphorylation of MDA5 at S88 was unaffected by US3 expression, while it was increased upon expression of a known antagonist of MDA5 dephosphorylation, the measles virus V protein (41), which served as a positive control (Fig. 1D).

FIG 1.

HSV-1 kinase US3 promotes phosphorylation of RIG-I at S8. (A) Top: consensus target sequences for HSV-1 kinase US3 and cellular protein kinases (PK) C and PKA. For US3, n is ≥2, X can be R, A, V, P, S, or absent; Y can be any amino acid except an acidic residue. For PKC, X can be any residue. For PKA, Φ can be any hydrophobic residue. Bottom: regulatory phosphorylation sites (in red) in the RIG-I and MDA5 CARDs. HSV-1 US3 GenBank accession number AJE60290. (B, C) HEK293T cells were cotransfected with GST-RIG-I 2CARD and either empty vector or FLAG-US3 as indicated. Whole cell lysates (WCLs) were prepared 48 h posttransfection (hpt) and subjected to precipitation with glutathione-Sepharose beads (PD: GST), followed by immunoblot (IB) analysis with antibodies specific for RIG-I phospho(p)-S8 or p-T170, ubiquitin (Ub) or GST. WCLs were immunoblotted with anti-FLAG or anti-β-actin (loading control). Arrow indicates RIG-I pT170 band. Asterisk indicates unspecific band. (D) HEK293T cells were transfected with HA-MDA5 together with empty vector, increasing amounts of FLAG-US3, or FLAG-tagged measles virus V protein (MeV V; positive control). WCLs were subjected to immunoprecipitation (IP) using anti-HA-coupled beads (IP: anti-HA), followed by IB analysis using antibodies specific for MDA5 p-S88 or HA. WCLs were immunoblotted with anti-FLAG or anti-β-actin. (E) HEK293T cells were transfected with full-length FLAG-RIG-I together with empty vector or V5-US3. At 48 hpt, WCLs were prepared and subjected to IP with FLAG M2 antibody-conjugated-Sepharose beads (IP: anti-FLAG), followed by IB with the indicated antibodies. (F) HEK293T cells were transfected with GST-RIG-I 2CARD together with either empty vector or V5-tagged US3 WT or the catalytically-inactive mutant K220M. At 48 hpt, cells were lysed and subjected to GST-PD, followed by IB with the indicated antibodies. (G) HEK293T cells were transfected with FLAG-tagged full-length RIG-I and either empty vector or HA-tagged US3 WT or K220A. At 48 hpt, cells were lysed and subjected to IP with anti-FLAG, followed by IB with the indicated antibodies. Data shown are representative of at least two independent experiments.

In uninfected cells, RIG-I assumes a closed, auto-inhibited conformation that limits exposure of the CARDs. Because the cellular kinases that maintain constitutive phosphorylation of RIG-I act on this inactive form of RIG-I, we next tested whether US3 is able to induce phosphorylation of full-length RIG-I. Coexpression of US3 with full-length RIG-I resulted in a robust enhancement of RIG-I S8 phosphorylation, suggesting that US3 is similarly able to access this 2CARD phosphorylation site in the auto-inhibited form of RIG-I (Fig. 1E). We also determined whether the kinase activity of US3 is required for the enhancement of RIG-I phosphorylation by using previously described catalytically-inactive mutants of US3, K220M and K220A (42, 43). Whereas wild-type (WT) US3 robustly enhanced the S8 phosphorylation of GST-RIG-I 2CARD or FLAG-RIG-I, the US3 K220M and K220A mutants did not noticeably affect the S8 phosphorylation state, demonstrating that the viral kinase activity is required for RIG-I phosphorylation by US3 (Fig. 1F and G). Taken together, these results showed that HSV-1 US3 enhances the phosphorylation of RIG-I at S8—but not RIG-I at T170 or MDA5 at S88—in a manner dependent on its kinase activity.

US3 interacts with RIG-I and inhibits critical RIG-I activation steps.

To determine whether RIG-I is a direct target of US3, we examined whether US3 interacts with RIG-I by performing coimmunoprecipitation experiments. First, HEK293T cells were transfected with FLAG-tagged RIG-I together with HA-tagged US3, followed by anti-FLAG immunoprecipitation to pull down RIG-I. We observed efficient coprecipitation of HA-US3 with FLAG-RIG-I (Fig. 2A). Next, we evaluated whether RIG-I and US3 interact in the context of virus infection. Endogenous RIG-I was precipitated from HSV-1-infected HEK293T cells using a specific antibody, and subsequent IB analysis resulted in efficient coprecipitation of US3 with RIG-I (Fig. 2B). These results showed that HSV-1 US3 interacts, directly or indirectly, with RIG-I, suggesting that S8 in RIG-I is a direct substrate of the viral kinase.

FIG 2.

HSV-1 US3 binds to RIG-I and blocks its CARD activation steps. (A) HEK293T cells were transfected with HA-tagged US3 together with either empty vector or FLAG-tagged RIG-I. Following chemical protein cross-linking at 40 hpt, WCLs were prepared and subjected to anti-FLAG IP, followed by IB analysis with the indicated antibodies. β-actin served as loading control. (B) Interaction between endogenous RIG-I and HSV-1 US3 in HEK293T cells infected with HSV-1 at a multiplicity of infection (MOI) of 1. Twenty hours postinfection (hpi), cells were lysed and subjected to IP with anti-RIG-I, followed by IB with the indicated antibodies. (C) HEK293T cells were transfected with V5-tagged US3 for 24 h and then infected with 50 HAU/mL of SeV for 24 h, or left uninfected. Cells were lysed and subjected to IP with anti-TRIM25 (+), or without antibody (−), followed by IB with the indicated antibodies. (D) HEK293T cells were transfected with V5-tagged US3 and an HA-tagged ubiquitin mutant in which all lysines except K63 were mutated (HA-K63-Ub) as indicated. At 24 hpt, cells were infected with 50 HAU/mL of SeV for 24 h, or left uninfected, after which the cells were lysed and subjected to IP with (+) or without (−) anti-RIG-I, followed by IB analysis with the indicated antibodies. (E) HEK293T cells were transfected with GST-RIG-I 2CARD, FLAG-MAVS CARD-proline rich domain (PRD), and either empty vector or V5-US3 as indicated. At 48 hpt, WCLs were subjected to anti-FLAG IP, followed by IB with the indicated antibodies. Data shown are representative of at least two independent experiments.

Dephosphorylation of RIG-I at S8 is a prerequisite for its interaction with the E3 ubiquitin ligase TRIM25, which mediates K63-linked ubiquitination of RIG-I to induce activation of downstream signaling (31, 34). Therefore, we asked whether US3 inhibits the interaction between RIG-I and TRIM25. To test this, we assessed the effect of ectopic expression of US3 on the interaction between endogenous RIG-I and TRIM25 in cells infected with Sendai virus (SeV), a strong and specific activator of RIG-I signaling (34). Immunoprecipitation and IB analysis showed that SeV infection efficiently triggered RIG-I binding to TRIM25 in the absence of US3; however, the RIG-I-TRIM25 interaction was reduced in the presence of US3 (Fig. 2C). In line with the compromised interaction that we observed between TRIM25 and RIG-I, the K63-linked ubiquitination of endogenous RIG-I following SeV infection was potently suppressed in the presence of US3 (Fig. 2D).

After undergoing K63-linked ubiquitination, RIG-I is primed for translocation to the mitochondria or mitochondria-associated membranes (MAM), where the RIG-I 2CARD domains mediate an interaction with the CARD domain of MAVS. Thus, we next tested the effect of US3 expression on the interaction between FLAG-tagged MAVS CARD-proline-rich domain (CARD-PRD) and GST-RIG-I 2CARD. An interaction between GST-RIG-I 2CARD and MAVS-CARD-PRD was readily observed in the absence of US3, yet in the presence of US3, this binding was strongly reduced, indicating that US3 inhibits the RIG-I-MAVS interaction that is required for antiviral signaling (Fig. 2E).

Together, these results indicated that US3 interacts with RIG-I during virus infection and inhibits critical downstream activation steps of this RNA sensor.

HSV-1 US3 suppresses RIG-I-mediated IFN induction, and this antagonism is dependent on S8 in RIG-I.

We next asked whether US3 blocks RIG-I-mediated type I IFN induction. To this end, we coexpressed titrating amounts of US3 with an IFN-β-promoter-driven luciferase reporter plasmid. SeV infection resulted in robust IFN-β luciferase reporter induction; however, this response was largely suppressed by US3 in a dose-dependent manner (Fig. 3A). Furthermore, in contrast to WT US3, the catalytically-inactive US3 K220M mutant was ineffective at suppressing SeV-induced IFN-β promoter activity (Fig. 3B). To determine whether inhibition of the IFN-β response by US3 is specifically due to its ability to phosphorylate S8 in RIG-I, we compared the effect of US3 on IFN-β luciferase reporter activation induced by RIG-I 2CARD WT or the S8A mutant, which cannot be phosphorylated at S8. In addition, we also tested the effect of US3 on IFN induction by RIG-I 2CARD T170A in which T170, which is also important for regulating RIG-I signaling (28–30) but is not phosphorylated by US3, is mutated to alanine. We observed that US3 expression suppressed the IFN-β promoter activation by RIG-I 2CARD WT or T170A in a dose-dependent manner (Fig. 3C to F). In contrast, IFN-β luciferase reporter activation by RIG-I 2CARD S8A was hardly affected by US3 (Fig. 3G and H). The modest reduction in 2CARD S8A-mediated reporter activation caused by high US3 expression may be due to other immunomodulatory functions of US3 (44, 45). These results indicated that HSV-1 US3 antagonizes the RIG-I-mediated type I IFN response and that the antagonism by US3 is dependent on S8 in RIG-I.

FIG 3.

US3-mediated inhibition of the RIG-I-dependent IFN response relies on S8 in RIG-I. (A) HEK293T cells were transfected with an IFN-β luciferase reporter construct (IFN-β-luc) and a β-galactosidase plasmid (pGK-β-gal), along with empty vector or increasing amounts (5 to 100 ng) of FLAG-US3. At 48 hpt, cells were infected with 5 HAU/mL of SeV for 16 h (or left uninfected), after which IFN-β promoter activity was measured by luciferase assay. (B) HEK293T cells were transfected with IFN-β-luc and pGK-β-gal along with 200 ng empty vector or FLAG-tagged US3 WT or K220M mutant. At 48 hpt, cells were infected with 5 HAU/ml of SeV for 16 h and IFN-β promoter activity was measured by luciferase assay. (C–H) HEK293T cells were transfected with IFN-β-luc, pGK-β-gal, and increasing amounts (5, 50, or 200 ng) of FLAG-US3, along with GST-RIG-I 2CARD WT (C and D), T170A (E and F), or S8A (G and H). IFN-β promoter activity was measured by luciferase assay at 48 hpt (C, E, G), and representative protein expression of FLAG-US3 was determined in the WCLs by IB with anti-FLAG (D, F, H). Immunoblotting for β-actin served as a loading control. Luciferase data were normalized to β-gal values, and induction relative to empty vector-transfected or uninfected samples is presented as mean ± SD of three biological replicates. Data shown are representative of at least two independent experiments. *, p <0.05; **, p <0.01; ***, p <0.001; ns, not significant (one-way ANOVA with Dunnett’s correction for multiple comparisons).

US3 counteracts RIG-I S8 dephosphorylation and blocks innate immune activation during HSV-1 infection.

We next sought to characterize the role of US3 in RIG-I S8 phosphorylation and innate immune inhibition in the context of HSV-1 infection. To this end, we generated a recombinant HSV-1 (F-strain) encoding the US3 K220A mutation (HSV-1 K220A) that abolishes US3 kinase activity, using a bacterial artificial chromosome (BAC) system as described previously (46). First, to evaluate the effect of viral US3 on RIG-I S8 phosphorylation during HSV-1 infection, we assessed the phosphorylation state of FLAG-RIG-I in either uninfected cells or cells infected with HSV-1 WT or HSV-1 K220A. Of note, infection conditions that resulted in a similar infection rate for WT and mutant virus were chosen, allowing for unambiguous assessment of RIG-I phosphorylation. RIG-I immunoprecipitated from cells infected with HSV-1 WT had similar levels of S8 phosphorylation as RIG-I purified from uninfected cells (Fig. 4A). In contrast, RIG-I from cells infected with HSV-1 K220A displayed low S8 phosphorylation, which is indicative of the inability of mutant US3 to counteract S8 dephosphorylation.

FIG 4.

US3 counteracts RIG-I dephosphorylation and innate signaling during HSV-1 infection. (A) HEK293T cells were transfected with FLAG-RIG-I as indicated and, 48 h later, infected with HSV-1 WT or HSV-1 K220A (each MOI 0.1) for 18 h, or left uninfected (mock). WCLs were subjected to IP with anti-FLAG, followed by IB with anti-RIG-I pS8 and anti-FLAG. WCLs were probed with anti-ICP8 (HSV-1). (B) qRT-PCR analysis of IFNB1 mRNA in the indicated cell types infected with HSV-1 WT or HSV-1 K220A (each MOI 1) for 16 h. (C) qRT-PCR analysis of the indicated ISG and chemokine transcripts in HDF cells infected with HSV-1 WT, HSV-1 K220A, or HSV-1 revertant (each MOI 1) for 8, 12, 16, or 24 h. (D) Left: qRT-PCR analysis of IFNB1 mRNA in HDF cells transfected for 48 h with either nontargeting control siRNA (si.Con) or siRNA targeting RIG-I (si.RIG-I) and then infected with HSV-1 WT or HSV-1 K220A (each MOI 1) for 16 h. Right: Knockdown efficiency of endogenous RIG-I (DDX58) was confirmed by qRT-PCR. (E) NHLF cells were infected with HSV-1 WT, HSV-1 K220A, or HSV-1 revertant (each MOI 0.01). Viral titers in the cell supernatant were determined by standard plaque assay at the indicated times after infection. Dotted line indicates the detection limit. Data shown are representative of two (A) or three (B to E) independent experiments; mean ± SD of three biological replicates in B–E. *, p <0.05; **, p <0.01; ***, p <0.001; ns, not significant (two-sided Student's t test).

Next, we asked whether the failure of HSV-1 K220A to phosphorylate S8 and thereby counteract RIG-I dephosphorylation-dependent activation results in higher IFN-β induction during HSV-1 K220A infection compared to HSV-1 WT infection. To test this, we compared IFNB1 transcript induction in primary human foreskin fibroblast (HFF), normal human lung fibroblasts (NHLF), and hTERT-immortalized human dermal fibroblasts (HDF) upon infection with HSV-1 WT or HSV-1 K220A. In the three cell types tested, we detected significantly higher levels of IFNB1 transcripts in HSV-1 K220A-infected cells compared to HSV-1 WT-infected cells (Fig. 4B). HSV-1 K220A infection also elicited higher levels of the ISGs RSAD2, IFIT2, and OAS1 as well as the chemokine CCL5/RANTES in HDF cells compared to infection with HSV-1 WT or with a revertant HSV-1 in which the WT US3 sequence was restored (Fig. 4C). To test whether the enhanced innate immune activation elicited by HSV-1 K220A was triggered in a RIG-I dependent manner, we depleted RIG-I by RNAi. In HDF cells transfected with non-targeting control siRNA (si.Con), infection with HSV-1 K220A induced significantly higher levels of IFNB1 than infection with HSV-1 WT (Fig. 4D). In contrast, in cells transfected with RIG-I-targeting siRNA, HSV-1 K220A induced similar levels of IFNB1 transcripts as did HSV-1 WT. Of note, HSV-1 K220A exhibited reduced replication compared to HSV-1 WT or the US3 revertant virus, as reported previously (45, 47–49), ruling out that the elevated cytokine response induced by HSV-1 K220A was due to higher virus replication that may result in enhanced availability of RLR-PAMPs (Fig. 4E). Collectively, these results showed that the HSV-1 US3 kinase activity counteracts RIG-I S8 dephosphorylation, thereby suppressing the RIG-I-mediated type-I IFN response during HSV-1 infection.

The ability to phosphorylate RIG-I S8 and to inhibit innate immunity is conserved among the alphaherpesvirus US3 kinase family.

Next, we asked whether other herpesvirus kinases also phosphorylate RIG-I and inhibit RIG-I-mediated IFN gene expression. Besides the US3 serine/threonine kinases that are uniquely encoded by the alphaherpesvirus subfamily (40, 50), a second family of protein kinases, the conserved herpesvirus-encoded protein kinases (CHPKs), are encoded by all three herpesvirus subfamilies (51). We first tested the ability of these herpesviral kinases to phosphorylate the regulatory residue S8 in RIG-I. We observed that ectopic expression of each of the US3-related kinases, HSV-1 US3, HSV-2 US3, and VZV ORF66, enhanced S8 phosphorylation of GST-RIG-I 2CARD compared to empty vector cotransfection (Fig. 5A to C). In contrast, none of the tested CHPKs (HSV-2 UL13, HCMV UL97, EBV BGLF4, KSHV ORF36, or MHV68 ORF36) noticeably increased GST-RIG-I 2CARD S8 phosphorylation levels (Fig. 5D to H). Of note, HCMV UL97 and EBV BGLF4 expression even reduced RIG-I 2CARD S8 phosphorylation. These results suggested that the ability to phosphorylate RIG-I S8 is conserved specifically among the alphaherpesvirus US3 kinase family and does not extend to the more widely expressed CHPK kinase family.

FIG 5.

The ability to phosphorylate RIG-I S8 is conserved among the alphaherpesvirus US3 kinase family. HEK293T cells were transfected with GST-RIG-I 2CARD along with either empty vector or titrating amounts of epitope-tagged US3-related kinases HSV-1 US3, HSV-2 US3, or VZV ORF66 (A–C), or CHPK kinases HSV-2 UL13, HCMV UL97, EBV BGLF4, KSHV ORF36, or MHV68 ORF36 (D–H). Forty-eight h later, WCLs were subjected to GST-PD, followed by IB analysis with the indicated antibodies. Data shown are representative of at least two independent experiments.

Next, we examined whether the ability of the other alphaherpesvirus US3 kinases to phosphorylate RIG-I S8 also correlates with inhibition of RIG-I-mediated type I IFN responses, as we observed for HSV-1 US3. Similar to HSV-1 US3 and in line with their ability to enhance RIG-I S8 phosphorylation, both HSV-2 US3 and VZV ORF66 suppressed RIG-I 2CARD-induced IFN-β promoter activation in a dose-dependent manner (Fig. 6A to C). In contrast, CHPK HSV-2 UL13, which failed to enhance RIG-I 2CARD S8 phosphorylation, was also ineffective at inhibiting RIG-I 2CARD-mediated IFN-β induction (Fig. 6D). These results suggested that RIG-I antagonism through direct phosphorylation of the sensor at S8 is specifically conserved among alphaherpesvirus US3-related kinases. However, the relative contribution of RIG-I phosphorylation by HSV-2 US3 and VZV ORF66 to effective immune evasion remains to be determined.

FIG 6.

Inhibition of RIG-I CARD-mediated signaling is conserved among the alphaherpesvirus US3 kinase family. HEK293T cells were transfected with IFN-β-luc, pGK-β-gal, and GST-RIG-I 2CARD together with either empty vector or increasing amounts of FLAG-tagged HSV-1 US3 (A), or HA-tagged HSV-2 US3 (B), VZV ORF66 (C), or HSV-2 UL13 (D). Forty-eight h later, IFN-β promoter activity was measured by luciferase assay. Fold inductions (mean ± SD of three biological replicates) are shown relative to the sample without GST-RIG-I 2CARD. Data shown are representative of at least three independent experiments. *, p <0.05; **, p <0.01; ***, p <0.001; ns, not significant (one-way ANOVA with Dunnett’s correction for multiple comparisons).

DISCUSSION

RLRs are highly regulated by post-translational modifications, which are mediated by a specific repertoire of cellular proteins (2). Modifications in the RLR CARDs in particular follow a prescribed sequence of events. In the case of RIG-I, two critical early steps in its activation are dephosphorylation of the CARD residues S8 and T170 and subsequent K63-linked ubiquitination by TRIM25 and other E3 ligases. Recent studies have uncovered a variety of mechanisms utilized by viruses to subvert these regulatory steps to evade the innate immune response (38). The TRIM25-mediated K63-linked ubiquitination step of RIG-I is inhibited by the NS1 proteins of influenza A virus and respiratory syncytial virus, the E6 protein of human papillomaviruses, and by the DENV sfRNA (52–55). Furthermore, a number of viral deubiquitinating enzymes remove K63-linked ubiquitination from RIG-I (56–59). The dephosphorylation step of RLR activation is also manipulated by viruses for innate immune evasion. For example, the V protein of measles virus prevents removal of the inhibitory phosphorylation mark at S88 in MDA5 through direct antagonism of the phosphatase PP1 (41). These findings highlight the importance of phosphorylation-dependent inhibition of RLRs, and therefore we postulated that some viruses may employ mechanisms that, rather than prevent RLR dephosphorylation, mimic cellular kinases and promote phosphorylation of RLRs at regulatory sites. We speculated that one such mechanism could be direct phosphorylation of RLRs by virus-encoded kinases. In this study, we found that the HSV-1-encoded Ser/Thr kinase US3 potently enhanced phosphorylation of RIG-I at S8 while we observed no effect on the phosphorylation levels of RIG-I T170 or MDA5 S88. Examination of the amino acid sequence surrounding the RIG-I S8 phosphorylation site revealed that it fits the consensus phosphorylation motif described for US3.

We further showed that US3-mediated phosphorylation of RIG-I prevents subsequent activation steps of RIG-I, including TRIM25 binding, K63-linked CARD ubiquitination, and the CARD-CARD interaction with MAVS. Two inhibitory phosphorylation marks have been described for the RIG-I CARDs, and removal of both appears to be required for full RIG-I activation (28, 30, 31). Therefore, it is not surprising that US3-mediated phosphorylation of only a single CARD residue appears to be sufficient for effective RIG-I inhibition. Our finding that US3 suppresses signaling mediated by RIG-I 2CARD WT, but not by RIG-I 2CARD S8A, suggests that the ability of US3 to inhibit the RIG-I-induced type I IFN response is largely dependent on its ability to manipulate the phosphorylation of RIG-I at S8. We also show that US3 and RIG-I interact during HSV-1 infection and, by using the catalytically-inactive US3 mutants K220A and K220M, that RIG-I S8 is a direct substrate of US3 kinase activity. Collectively, these findings suggest that US3 binds to and directly phosphorylates RIG-I S8 in HSV-1-infected cells. US3 has been described to perform a myriad of functions, including inhibition of apoptosis and NF-κB activation as well as regulation of viral egress (60–64). It is tempting to speculate that a particular subcellular population, or isoform, of US3 may be responsible for RIG-I antagonism. US3 is expressed late during infection and is also found in the tegument of mature virions (65). Future studies will be required to determine the precise kinetics of RIG-I phosphorylation by US3 during infection and whether RIG-I phosphorylation is mediated by tegument US3 released during primary infection and/or by de novo-expressed US3. In HSV-1 and pseudorabies virus, US3 contains an alternative start site that in the case of HSV-1 gives rise to a truncated US3 protein (US3.5) that lacks the first 76 N-terminal residues (66). This truncation does not affect the kinase activity and, although the known target proteins of US3 and US3.5 are highly similar (66), it remains to be determined whether these isoforms have differential RIG-I-binding and -inhibitory abilities.

US3 is considered a promiscuous kinase, and while many putative substrates have been identified (especially in in vitro studies), not all may be of equal importance for viral replication and pathogenesis. Therefore, we investigated whether US3 phosphorylates RIG-I in the context of viral infection, and whether this antagonistic mechanism contributes to inhibition of type I IFN responses in HSV-1-infected cells. RIG-I purified from cells that were infected with WT HSV-1 was found to have similar levels of S8 phosphorylation as RIG-I purified from uninfected cells, whereas RIG-I immunoprecipitated from cells infected with HSV-1 K220A had undergone efficient S8 dephosphorylation. These findings suggest that HSV-1 WT effectively counteracts RIG-I dephosphorylation by encoding an active US3 kinase that facilitates the phosphorylation of RIG-I S8. On the other hand, HSV-1 K220A, which encodes a catalytically-inactive US3 kinase, loses the ability to maintain RIG-I S8 phosphorylation and is unable to prevent the dephosphorylation-dependent activation of RIG-I. We also observed that infection with HSV-1 encoding US3 K220A induced higher levels of IFN-β and ISGs than infection with the WT or revertant virus. Importantly, IFNB1 induction elicited by HSV-1 K220A was markedly reduced upon RIG-I silencing and similar to that elicited by WT HSV-1. These results indicated that the enhanced immune activation by HSV-1 K220A is RIG-I-dependent.

Our results point to a model in which US3 exerts its major inhibitory activity on the type-I IFN response by counteracting RIG-I CARD dephosphorylation. US3 mimics the activity of cellular kinases PKC-α/β to mediate phosphorylation-dependent repression of RIG-I signaling, providing another example of how viral proteins can take advantage of preexisting cellular regulatory mechanisms to antagonize the antiviral innate immune response. Furthermore, our findings suggest that this RIG-I targeting strategy is conserved among the US3 family of alphaherpesvirus kinases; however, additional studies in the context of virus infection are needed to confirm whether RIG-I antagonism by HSV-2 US3 and VZV ORF66 are relevant for suppressing innate immunity. Along these lines, it will be interesting to explore whether the same tactic is employed by viruses beyond the Herpesviridae family, given that all large DNA virus families, including Poxviridae and Baculoviridae, encode conserved Ser/Thr kinases (67).

MATERIALS AND METHODS

Cell culture and viruses.

HEK293T (ATCC), Vero (ATCC), Normal Human Lung Fibroblasts (NHLF, Clonetics), hTERT-immortalized human dermal fibroblasts (HDF, kindly provided by Patrick Hearing (68)), and human foreskin fibroblasts (HFF, ATCC) were cultured in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 10% (vol/vol) fetal bovine serum (FBS), 2 mM glutamine (GlutaMAX; Gibco), and 1% (vol/vol) penicillin-streptomycin (Gibco) under standard tissue culture conditions.

The US3 K220A mutant recombinant HSV-1 (HSV-1 K220A) was generated using a bacterial artificial chromosome (BAC) clone of HSV-1 strain F (pYEbac102; kindly provided by Yasushi Kawaguchi, University of Tokyo, Japan). To facilitate detection of infected cells, pYEbac102 was modified by insertion of the egfp (enhanced green fluorescent protein) gene into mini-F vector sequences, resulting in the pYEbac102-G construct that was used for further manipulation. Expression of EGFP was driven by the human cytomegalovirus (HCMV) immediate early (IE) promoter. En passant mutagenesis (46) was performed to introduce a mutation at amino acid residue 220 of the US3 protein kinase where the original lysine (K) was substituted with alanine (A), leading to the BAC pYEbac102-K220A-G construct. The following primers were used to amplify the PCR product that allowed for mutagenesis to occur: (forward: 5′-TGACAGCAGCCACCCAGATTACCCCCAACGGGTAATCGTGGCGGCGGGGTGGTACACGAGCACTAGGGATAACAGGGTAATCGAT-3′; reverse: 5′-GCAGTCGCGCCTCGTGGCTCGTGCTCGTGTACCACCCCGCCGCCACGATTACCCGTTGGGGGTCCAGTGTTACAACCAATTAACC-3′; (nucleotides underlined and bold represent the desired mutation). After confirmation of the constructs by restriction fragment length polymorphism (RFLP) and DNA sequencing (LGC Genomics, Germany), 1 μg each of purified BAC DNA was transfected into Vero cells at 90% confluence using polyethylenimine (PEI; Polysciences) according to procedures described elsewhere (69). At 72 h posttransfection, when visible fluorescent plaques appeared, the respective mutant virus was harvested following 3 cycles of freeze-thawing. Following the same strategy, a revertant HSV-1 (HSV-1 US3 revertant) was prepared in which the mutated alanine was restored to the original lysine at position 220.

HSV-1 infection of indicated cells was performed in DMEM supplemented with 1% FBS, 2 mM glutamine, and 1% (vol/vol) penicillin-streptomycin at the indicated multiplicity of infection (MOI) for 1 h at 37°C, after which the infection medium was replaced by fresh supplemented DMEM for the remainder of the experiment.

DNA constructs and transfections.

Plasmids encoding GST-RIG-I 2CARD, GST-RIG-I 2CARD S8A, GST-RIG-I 2CARD T170A, FLAG-RIG-I, HA-MDA5, and FLAG-MAVS-CARD-PRD have been described previously (28, 31, 34). Plasmids encoding HSV-1 US3 (FLAG-US3, V5-US3, HA-US3) are described in (44). Plasmids encoding HSV-2 US3 (pSG5-US3) and HSV-2 UL13 (pSG5-UL13) were gifts from Lynda Morrison (Saint Louis University) (70). The plasmid encoding VZV ORF66 (pGK2-HA66) was a gift from Paul Kinchington (University of Pittsburgh) (71). The plasmid encoding HCMV UL97 (pHA-UL97) was a gift from Donald Coen (Harvard University) (72). The plasmid encoding MHV68 ORF36 was a gift from Pinghui Feng (University of Southern California). The plasmid encoding EBV BGLF4 (pcDNA-BGLF4–FLAG) was a gift from Manfred Marschall (Friedrich-Alexander Universität Erlangen-Nürnberg) (73). The plasmid encoding KSHV ORF36-myc (pcDNA4-ORF36-myc/His) was a gift from Frank Neipel (Friedrich-Alexander Universität Erlangen-Nürnberg) (74). The plasmid encoding V5-US3 K220M was generated by site-directed mutagenesis using primers containing the desired mutation. The plasmid encoding K63-only-Ub (#17606) cloned into the pRK5-HA vector was a gift from T. Dawson (Addgene (75)).

Transfection of plasmids into cells was performed using calcium phosphate (Clontech), Lipofectamine and Plus reagent, or Lipofectamine 2000 (both Life Technologies), according to the manufacturer’s instructions.

Coimmunoprecipitation, pull-downs, and immunoblot analysis.

GST, FLAG, and hemagglutinin (HA) immunoprecipitation assays were performed essentially as described previously (28, 34). HEK293T cells were transfected with the indicated constructs for 48 h, after which the cells were collected by centrifugation and lysed in NP-40 lysis buffer (50 mM HEPES pH 7.4, 150 mM NaCl, 1% (vol/vol) NP-40, protease inhibitor cocktail [Sigma] and Ser-Thr phosphatase inhibitor cocktail [Sigma]). After centrifugation for 20 min at 13,000 rpm at 4°C, cleared lysates were mixed with glutathione-conjugated Sepharose beads (Amersham Biosciences), anti-FLAG M2 Sepharose beads (Sigma), or anti-HA-conjugated beads (Sigma) and incubated at 4°C for 4 h. Alternatively, for the RIG-I and TRIM25 coimmunoprecipitations, lysates were incubated with anti-RIG-I antibody (Alme-1, Adipogen) or anti-TRIM25 antibody (2/EFP, BD Biosciences) that was captured by subsequent incubation with protein A/G agarose beads (Pierce). Precipitates were then washed extensively with lysis buffer and eluted with Laemmli gel SDS loading buffer by boiling for 5 min at 95°C. Where indicated, proteins were chemically cross-linked by treating the cells with 1 mM DSP (Sigma) for 2 h on ice followed by addition of 20 mM Tris-HCl (pH 8.0) for inactivation, before proceeding with cell lysis and immunoprecipitation as described.

For immunoblot analysis, proteins were resolved by Bis-Tris-PAGE and transferred onto poly-vinylidene difluoride (PVDF) membranes (Bio-Rad) using a Trans-Blot SD semidry transfer cell (Bio-Rad). After blocking with 5% (wt/vol) nonfat dry milk in PBS-Tween 20 for 1 h, membranes were probed with primary antibodies for either 1 h at room temperature or at 4°C overnight. The following primary antibodies were used: anti-FLAG (1:2000, M2, Sigma), anti-HA (1:2000, HA-7, Sigma), anti-β-Actin (1:5000, A1978, Sigma), anti-ICP8 and anti-US3 (1:10,000 and 1:1000, respectively (76, 77)), anti-RIG-I (1:2000, Alme-1, Adipogen), anti-TRIM25 (1:1000, 2/EFP, BD Biosciences), anti-GST (1:2000, GST-2, Sigma), anti-ubiquitin (1:500, P4D1, Santa Cruz), anti-V5 (1:2000, R960-25, Invitrogen), anti-myc (1:2000, 9E10, Covance), and phospho-specific anti-RIG-I pS8, anti-RIG-I pT170, and anti-MDA5 pS88 antibodies that have been previously described (28, 30). Goat anti-mouse- or goat anti-rabbit-horseradish peroxidase (HRP) secondary antibodies (both 1:2000) were purchased from Cell Signaling Technologies (cat #7076S and #7074S, respectively). Protein bands were visualized with the enhanced SuperSignal West Pico chemiluminescence reagent (Thermo Fisher) and were detected by a luminescent imaging system (LAS-4000; Fujifilm).

Luciferase reporter assays.

HEK293T cells were seeded into 12-well plates and transfected with 200 ng IFN-β luciferase reporter construct and 300 ng β-galactosidase-expressing pGK-β-gal using Lipofectamine and Plus reagent (Life Technologies) according to the manufacturer’s instructions (52). To induce RLR signaling, cells were either cotransfected with 1 ng of plasmid encoding GST-RIG-I 2CARD (WT, S8A, or T170A) or infected with Sendai virus (SeV, 5 HAU/ml, Cantell strain, Charles River Laboratories) as indicated. At the indicated times after transfection or infection, the cells were harvested and firefly luciferase and β-galactosidase activities were determined using the Luciferase Assay System and β-Galactosidase Enzyme Assay System (Promega), respectively. Luminescence and absorbance were measured using a BioTek Synergy Microplate Reader in 96-well plates using 10 μL and 25 μL of cell lysates, respectively. Luciferase activity was normalized to β-galactosidase values, and luciferase induction was calculated relative to vector-transfected or mock-infected samples that were set to 1.

Quantitative reverse transcription-PCR (qRT-PCR).

Total RNA was purified from cells using the HP Total RNA Kit (OMEGA Bio-tek) following the manufacturer’s instructions. RNA quality and quantity were assessed using NanoDrop 2000. qRT–PCR was performed with equal amounts (25–500 ng) of the purified RNA using the SuperScript III Platinum One-Step qRT–PCR kit with ROX (Invitrogen) on a 7500 Fast Real-Time PCR Machine (Applied Biosystems) according to the manufacturer’s instructions. TaqMan primers and probes for each individual gene were acquired as premixed master mixes (IDT). Fold expression level of each target gene relative to mock-treated cells was calculated by normalizing against GAPDH or 18S RNA using the Comparative CT Method (ΔΔCT Method) and presented relative to the values for mock-treated cells, set to 1.

siRNA-mediated knockdown of RIG-I.

HDF cells were seeded into 12-well plates (∼0.5 million cells per well) and transfected with 80 nM siRNAs targeting RIG-I (siGENOME SMARTpool M-012511-01, Horizon) or nontargeting control siRNAs (D-001206-14) using Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer's instructions. Silencing efficiency of the target gene was determined by analyzing the transcript levels by qRT–PCR.

Data availability.

The data that support the findings of this study are available from the corresponding author upon request.