USP4 Positively Regulates RIG-I-Mediated Antiviral Response through Deubiquitination and Stabilization of RIG-I

Abstract

Protein ubiquitination plays an essential role in the regulation of retinoic acid-inducible gene I (RIG-I) activation and the antiviral immune response. However, the function of the opposite process of deubiquitination in RIG-I activation remains elusive. In this study, we have identified the deubiquitinating enzyme ubiquitin-specific protease 4 (USP4) as a new regulator for RIG-I activation through deubiquitination and stabilization of RIG-I. USP4 expression was attenuated after virus-induced RIG-I activation. Overexpression of USP4 significantly enhanced RIG-I protein expression and RIG-I-triggered beta interferon (IFN-β) signaling and, at the same time, inhibited vesicular stomatitis virus (VSV) replication. Small interfering RNA (siRNA) knockdown of USP4 expression had an opposite effect. Furthermore, USP4 was found to interact with RIG-I and remove K48-linked polyubiquitination chains from RIG-I. Therefore, we identified USP4 as a new positive regulator for RIG-I that acts through deubiquitinating K48-linked ubiquitin chains and stabilizing RIG-I.

INTRODUCTION

Type I interferons (IFN-α and IFN-β) possess strong antiviral activity and play pivotal roles in defense against viral infection (1, 2). During viral infection, pattern-recognition receptors (PRRs), including Toll-like receptors (TLRs) and retinoic acid-inducible gene I (RIG-I)-like helicases (RLRs), activate immune cells to produce type I IFNs and proinflammatory cytokines, which are involved in the elimination of viruses (3). RLRs are cytoplasmic DExD–H-box RNA helicases and recognize viral RNA (4). There are three members in the RLR family: RIG-I, melanoma differentiation-associated gene 5 (MDA5), and laboratory of genetics and physiology 2 (LGP2) (3, 4). RIG-I is a key PRR for the detection of positive- and negative-stranded RNA viruses, including Sendai virus (SeV) and vesicular stomatitis virus (VSV) (5, 6). RIG-I is composed of two N-terminal caspase recruitment domains (CARDs), a central DExD/H box helicase/ATPase domain and a C-terminal regulatory domain (CTD) (6). N-terminal CARDs are responsible for the binding to the adaptor molecule mitochondrial antiviral signaling protein (MAVS) (also called interferon promoter stimulator 1 [IPS-1], Cardif, or virus-induced signaling adaptor [VISA]) (7–10). In the absence of viral infection, RIG-I is monomeric and is inactivated by masking of the central DExD/H domain with its N-terminal CARDs (6, 11). Upon infection and recognition of viral RNA, the CTD undergoes a conformational change and dimerizes, displacing the CARDs. RIG-I oligomers then interact with MAVS through their respective CARDs to induce oligomerization of MAVS. This MAVS complex then triggers the recruitment of various signaling components, such as stimulator of interferon genes (STING) (also called mitochondrial mediator of interferon regulatory factor 3 activation [MITA] or endoplasmic reticulum intermembrane small protein [ERIS]) (12–14) and TANK-binding kinase 1 (TBK1) (also called Nef-associated kinase [NAK])/IκB kinase ε (IKK-ε), leading to the phosphorylation, dimerization, and nuclear translocation of IFN regulatory factor 3 (IRF 3) and the production of IFN-β (3, 6, 15).

Given the important role of RIG-I in the antiviral innate response, it is not surprising that RIG-I activation is tightly controlled during viral infection. Up to now, many kinds of factors have been found to positively or negatively regulate RIG-I signaling (16). For example, a short isoform of zinc-finger CCCH-type antiviral protein 1 (ZAPS) is associated with RIG-I in the promotion of the oligomerization and ATPase activity of RIG-I, which leads to robust activation of IRF3 and NF-κB transcription factors (17). Atg5 and Atg12 associate directly with the CARD domain of RIG-I and this leads to an inactive status of RIG-I (18). ADP ribosylation factor-like protein 16 (ARL16) changes to GTP-binding status upon viral infection and binds with the RIG-I CTD to negatively control its signaling activity (19). Protein posttranslational modifications, such as phosphorylation (20, 21) and ISG15 conjugation (22, 23), have also been reported to be associated with the regulation of RIG-I activity.

Another well-studied protein modification that is essential in the regulation of RIG-I activity is ubiquitination. The ubiquitin ligases TRIM25 and Riplet were reported to deliver the K63-linked polyubiquitin moiety to RIG-I CARDs and the C-terminal domain, leading to the activation of RIG-I signaling (24, 25). In contrast, E3 ligase RNF125 was found to mediate K48-linked polyubiquitination preferentially in the N-terminal CARDs domain of RIG-I, leading to the proteasome degradation of RIG-I and negative regulation of RIG-I signaling (26). Although the function of ubiquitination in RIG-1 activation is well defined, the function of deubiquitination, an opposite process to ubiquitination, remains elusive. Deubiquitination is catalyzed by deubiquitinating enzymes (DUBs), which are proteases that cleave ubiquitin or ubiquitin-like proteins from target proteins (27). The human genome encodes about 100 putative DUBs that are divided into five subclasses based on their ubiquitin (Ub) protease domains. The ubiquitin-specific proteases (USPs) represent the largest subclass of DUBs (27).

In this study, we initially screened the members of the USP family that were differentially expressed upon RIG-I activation. We found that USP4 expression was inhibited by both poly(I·C) transfection and SeV infection in various cell types. Furthermore, we demonstrated that overexpression of USP4 significantly enhanced RIG-I-triggered IFN-β signaling and inhibited VSV replication, while small interfering RNA (siRNA) knockdown of USP4 expression had an opposite effect. As a deubiquitinating enzyme, USP4 was found to interact with RIG-I and remove K48-linked polyubiquitination chains from RIG-I, leading to the stabilization of this receptor. Therefore, our study identified USP4 as a new positive regulator for RIG-I through its deubiquitination of K48-linked ubiquitin chains and stabilization of RIG-I.

MATERIALS AND METHODS

Cells and reagents.

HEK293, HeLa, and Huh7.5 cells were obtained from the American Type Culture Collection (Manassas, VA). The cells were cultured at 37°C under 5% CO₂ in Dulbecco modified Eagle medium (DMEM) supplemented with 10% fetal calf serum (FCS) (Invitrogen-Gibco), 100 U/ml penicillin, and 100 μg/ml streptomycin, as described previously (28, 29). Poly(I·C) and poly(dA·dT) were purchased from InvivoGen (San Diego, CA) and used at a final concentration of 1 μg/ml for transfection. MG132 and the antibodies for USP4 and Flag were from Sigma (St. Louis, MO). The antibodies specific for HA, Ub, RIG-I, MAVS, β-actin, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and protein G agarose used for immunoprecipitation were from Santa Cruz Biotechnology (Santa Cruz, CA). The antibodies specific to Myc and TBK1 were from Cell Signaling Technology (Beverly, MA). Their respective horseradish peroxidase-conjugated secondary antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). SeV was purchased from the China Center for Type Culture Collection (Wuhan University, China).

Sequences, plasmid constructs, and transfection.

The pCMV6-Flag-USP4 expression plasmid was purchased from OriGene (Rockville, MD). USP4 cDNA was subcloned into pCMV-Myc plasmids (Promega). The IRF-3 5D mutant, in which residues at positions 396, 398, 402, 404, and 405 were replaced by the phosphomimetic aspartate amino acid, was generated using the KOD-Plus mutagenesis kit (Toyobo, Osaka, Japan). STING cDNAs were amplified from THP1 cells by PCR and cloned in a pCMV plasmid (Promega). All constructs were confirmed by DNA sequencing. Expression plasmids for RIG-I, MAVS, Toll-interleukin-1 receptor domain-containing adaptor-inducing beta interferon (TRIF), TBK1, IKK-ε, IRF3, hemagglutinin-tagged ubiquitin (HA-Ub), IFN-β, and IRF3 reporter plasmids were obtained as described previously (28). RIG-I truncation mutants were gifts from Danying Chen (Beijing University, China). USP4 mutants were provided by Ping Wang (East China Normal University, China). For transient transfection of plasmids, jetPEI reagents were used (Polyplus Transfection). For transient silencing, duplexes of small interfering RNA were transfected into cells with the Lipofectamine 2000 transfection reagent (Invitrogen) according to the standard protocol. Target sequences for transient silencing were 5′-GCUGGGACAUGUACAAUGU-3′(siRNA 1), 5′-GGCUCUGGAACAAAUACAU-3′(siRNA 2), and 5′-GGUCGCAGAUGUGUAUAAU-3′ (siRNA 3) for USP4; scrambled control sequences were 5′-UUCUCCGAACGUGUCACGU-3′.

RNA quantitation.

Total RNA was extracted with TRIzol reagent according to the manufacturer's instructions (Invitrogen). Specific primers used for reverse transcriptase PCR (RT-PCR) assays were 5′-CCTGGGCTCTGTGGACTTG-3′ and 5′-TGTTGATTTCGGCTTCATACTC-3′ for USP4, 5′-CAACAAGTGTCTCCTCCAAAT-3′ and 5′-TCTCCTCAGGGATGTCAAAG-3′ for IFN-β, and 5′-CAAGGTCATCCATGACAACTTTG-3′, and 5′-GTCCACCACCCTGTTGCTGTAG-3′ for GAPDH.

Immunoprecipitation and Western blot analysis.

For immunoprecipitation (IP), whole-cell extracts were collected 36 h after transfection and were lysed in IP buffer containing 1.0% (vol/vol) Nonidet P-40, 50 mM Tris-HCl (pH 7.4), 50 mM EDTA, 150 mM NaCl, and a protease inhibitor cocktail (Merck). After centrifugation for 10 min at 14,000 × g, supernatants were collected and incubated with protein G plus-agarose immunoprecipitation reagent (Santa Cruz) together with 1 μg monoclonal anti-Flag or 1 μg anti-Myc. After 6 h of incubation, beads were washed five times with IP buffer. Immunoprecipitates were eluted by boiling with 1% (wt/vol) SDS sample buffer. For Western blot analysis, immunoprecipitates or whole-cell lysates were loaded, subjected to SDS-PAGE, transferred onto nitrocellulose membranes, and then blotted as described previously (28, 29).

Assay of luciferase activity.

Luciferase activity was measured with the dual-luciferase reporter assay system according to the manufacturer's instructions (Promega), as described previously (28, 30). Data were normalized for transfection efficiency by division of the firefly luciferase activity with that of Renilla luciferase.

Ubiquitination assays.

Ubiquitination assays were performed as described previously (28, 29). For analysis of the ubiquitination of RIG-I, HEK293 cells were transfected with Flag-RIG-I, HA-Ub (wild type [WT]), or HA-Ub mutants and Myc-USP4, and then whole-cell extracts were immunoprecipitated with anti-Flag and analyzed by immunoblot with anti-HA antibody.

VSV plaque assay and detection of virus replication.

The VSV plaque assay was performed as described previously (31). The HEK293 cells or macrophages (2 × 10⁵) were transfected with the indicated plasmids or siRNA for 36 h prior to VSV infection (multiplicity of infection [MOI] of 0.1). At 1 h after infection, cells were washed with phosphate-buffered saline (PBS) three times and then medium was added. The supernatants were harvested at 24 h after washing. The supernatants were diluted 1:10⁶ and then used to infect confluent HEK293 cells cultured on 24-well plates. At 1 h postinfection, the supernatant was removed, and 3% methylcellulose was overlaid. At 3 days postinfection, the overlay was removed; cells were fixed with 4% formaldehyde for 20 min and stained with 0.2% crystal violet. Plaques were counted, averaged, and multiplied by the dilution factor to determine viral titer as PFU/ml. Total cellular RNA was extracted from HEK293 or macrophages transfected with VSV, and VSV RNA replicates were examined by quantitative RT-PCR. Primers used for VSV replicates were as follows: 5′-ACGGCGTACTTCCAGATGG-3′ (sense) and 5′-CTCGGTTCAAGATCCAGGT-3′ (antisense).

Statistical analysis.

All data are presented as means ± standard deviations (SD) of results from three or four experiments. Statistical significance was determined with the two-tailed Student t test; a P value of <0.05 was considered statistically significant.

RESULTS

USP4 expression is inhibited by RIG-I activation.

To investigate the function of USP in RIG-I signaling, we initially set out to identify USPs that were differentially expressed upon RIG-I activation. Among them, USP4 was found to have modulated expression upon RIG-I activation (Fig. 1). There was moderate expression of USP4 in HeLa cells (Fig. 1A and B). Transfection of poly(I·C), which has been shown to activate RIG-I signaling (6), substantially attenuated USP4 mRNA and protein expression in HeLa cells (Fig. 1A and B). Similarly, USP4 expression was also decreased upon SeV infection, which has been reported to activate RIG-I signaling (Fig. 1A and B). SeV infection also attenuated USP4 expression in murine peritoneal macrophages (Fig. 1C and D). Transfection of poly(I·C) and SeV infection also attenuated USP4 expression in HEK293 cells (data not shown). These data indicate that RIG-I signaling decreases USP4 expression in various types of cells.

Fig 1.

USP4 expression is inhibited by RIG-I activation. (A, B) RT-PCR and Western blot analysis of USP4 expression in HeLa cells transfected with poly(I·C) or infected with SeV for the indicated time periods. (C, D) RT-PCR and Western blot analysis of USP4 expression in mouse peritoneal macrophages infected with SeV for the indicated time periods. Similar results were obtained in three independent experiments.

USP4 enhances RIG-I-induced IFN-β production.

USP4 has been identified as a negative regulator of TLR/interleukin 1 receptor (IL-1R) signaling through deubiquitination of K-63-linked ubiquitin chains on tumor necrosis factor receptor-associated factor 6 (TRAF6) (32). The RIG-I-mediated decrease of USP4 expression suggests that USP4 might play a very important role in RIG-I signaling. To investigate the function of USP4 in RIG-I signaling, poly(I·C) was transfected into HEK293 cells with or without the USP4 expression plasmid, and the expression of IFN-β was measured by RT-PCR. As shown in Fig. 2A, transfection of poly(I·C) induced the expression of IFN-β in HEK293 cells (lane 3). Notably, overexpression of USP4 further increased poly(I·C)-induced expression of IFN-β (lane 4). To further confirm the regulatory role of USP4 on RIG-I-mediated IFN-β production, three small interfering RNAs (siRNAs) were synthesized and transfected into HEK293 cells to suppress endogenous USP4 expression. siRNA 2 was found to have a higher efficiency to knock down USP4 expression at both mRNA and protein level (Fig. 2B). Thus, siRNA 2 was used in the following experiments. As shown in Fig. 1C, USP4 knockdown through siRNA transfection significantly inhibited poly(I·C)-induced IFN-β expression in HEK293 cells. Similarly, poly(I·C)-induced IFN-β production was also greatly attenuated by USP4 knockdown in HeLa cells (Fig. 2C).

Fig 2.

USP4 enhances RIG-I-induced IFN-β production. (A) HEK293 cells transfected with the indicated plasmids were transfected with poly(I·C) or poly(dA·dT) for 12 h. Total RNA was prepared and analyzed by RT-PCR for the expressions of IFN-β,USP4, and GAPDH. (B) RT-PCR and Western blot analysis of USP4 expression in HEK293 cells transfected with control siRNA or USP4 siRNA 1, siRNA 2, and siRNA 3 for 36 h. (C) HEK293 or HeLa cells were transfected with control siRNA or USP4 siRNA 2 for 36 h, and then transfected with poly(I·C) or poly(dA·dT) for the indicated time periods. Total RNA was prepared and analyzed by RT-PCR for the expression of IFN-β and GAPDH. (D) HeLa cells were transfected with control siRNA or USP4 siRNA 2 for 36 h and then infected with SeV for 12 h. Total RNA was prepared and analyzed by RT-PCR for the expression of IFN-β and GAPDH. (E) HeLa cells were transiently transfected with IFN-β or the IRF3 reporter plasmid together with the USP4 expression plasmid or the control plasmid and analyzed for luciferase activity after SeV infection for 12 h. (F) HeLa cells were transfected with control siRNA or USP4 siRNA 2 and then transfected with the IFN-β or IRF3 reporter plasmid and analyzed for luciferase activity after SeV infection for 12 h. **, P < 0.01. Data are representative of three experiments (mean and SD of six samples in panels E and F).

Poly(dA·dT) has been reported to activate RIG-I signaling through RNA polymerase III-mediated transcription of poly(dA·dT) into poly(I·C), which can be recognized by RIG-I (33, 34). Consistent with the positive regulatory role of USP4 in RIG-I signaling, poly(dA·dT)-induced IFN-β production was either enhanced or attenuated with USP4 overexpression or knockdown, respectively (Fig. 2A and C).

SeV-induced IFN-β production was also substantially attenuated after knockdown of USP4 expression in HeLa cells (Fig. 2D). To further confirm the positive regulatory role of USP4 on RIG-I signaling, IFN-β promoter plasmid and IRF3 cis-reporting plasmid were transfected into HeLa cells, followed with SeV infection. As shown in Fig. 2E, USP4 overexpression significantly enhanced SeV-induced IFN-β promoter and IRF3 activation in HeLa cells. In contrast, USP4 knockdown significantly inhibited SeV-induced IFN-β promoter and IRF3 activation (Fig. 2F). Taken together, these data indicate that USP4 positively regulates RIG-I-mediated IFN-β signaling.

USP4 targets RIG-I.

To determine the molecular targets of USP4, we used RT-PCR to examine the effect of USP4 on the IFN-β expression mediated by various adaptors in RIG-I signaling. As shown in Fig. 3A, USP4 enhanced RIG-I-induced IFN-β mRNA expression, while USP4 overexpression did not impair IFN-β mRNA expression induced by MDA5, MAVS, TBK1, IKK-ε-, TRIF, and IRF3 5D (Fig. 3A, top). Similarly, RIG-I-induced IFN-β mRNA expression was also attenuated by USP4 overexpression in Huh7.5 cells (Fig. 3B). To further confirm USP4 function on RIG-I-mediated IFN-β mRNA expression under physiological conditions, we transfected USP4 siRNA into HeLa cells. As shown in Fig. 3A (bottom), USP4 siRNA transfection substantially decreased RIG-induced IFN-β mRNA expression, but not MDA5-, MAVS-, TBK1-, IKK-ε-, TRIF-, and IRF3 5D-induced IFN-β mRNA expression. Consistently, RIG-I-induced activation of the IFN-β promoter and IRF3 reporter was also significantly enhanced by USP4 overexpression (Fig. 3C). However, MAVS-, TBK1-, TRIF-, and IRF3-induced activation of the IFN-β promoter and IRF3 reporter was not affected by USP4 overexpression (Fig. 3C), indicating that USP4 targets RIG-I to modulate IFN-β signaling.

Fig 3.

USP4 targets RIG-I. (A) HEK293 cells were transfected with RIG-I, MAVS, TBK1, IRF3 5D, TRIF, IKK-ε, and MDA5 and analyzed by RT-PCR along with the USP4 plasmid and IFN-β mRNA (upper panel). HEK293 cells were transfected with RIG-I, MAVS, TBK1, IRF3 5D, and MDA5 and analyzed by RT-PCR along with USP4-specific siRNA or control siRNA and IFN-β mRNA (lower panel). (B) Huh7.5 cells were transfected with RIG-I, MAVS, TBK1, and IRF3 5D and analyzed by RT-PCR along with the USP4 expression plasmid and IFN-β mRNA. (C) HEK293 cells were transfected with TRIF, RIG-I, MAVS, TBK1, and IRF3 5D and analyzed along with IFN-β or the IRF3 reporter plasmid and USP4 plasmid and luciferase activity. **, P < 0.01. Data are representative of three experiments (mean and SD of six samples). (D) Western blot analysis of the lysates from HEK293 cells transfected with various tagged molecules (RIG-I, STING, TBK1, MDA5, IKK-ε, and IRF3) and Flag-tagged USP4. (E) Western blot analysis of the lysates from HEK293 cells transfected with Myc-RIG-I together with increasing concentrations of the Flag-USP4 expression plasmid. (F) Western blot analysis of the expression of RIG-I, MAVS, and USP4 in HeLa cells transfected with control siRNA or USP4 siRNA 1, 2, or 3 for 36 h. (G) Western blot analysis of the expression of RIG-I and TBK1 in HeLa cells transfected with control siRNA or USP4 siRNA 2 and then infected with SeV for the indicated time periods. Similar results were obtained in three independent experiments.

In order to clarify if USP4 targets RIG-I, we investigated the function of USP4 in the degradation of the molecules involved in RIG-I signaling. Various expression plasmids for RIG-I, MDA5, MAVS, STING, TBK1, IKK-ε, and IRF3 were cotransfected into HEK293 cells with USP4 expression plasmids, and the protein levels of the signaling molecules were analyzed by Western blotting 24 h after transfection. USP4 expression significantly enhanced RIG-I expression in a dose-dependent manner (Fig. 3D and E). In contrast, USP4 overexpression had no effects on MDA5, MAVS, STING, TBK1, IKK-ε, and IRF3 expression (Fig. 3D). To further confirm USP4-inhibited RIG-I degradation, USP4 expression was silenced by USP4 siRNA transfection in HeLa cells. Transfection of USP4 siRNA substantially decreased the RIG-I protein level but did not impair the MAVS protein level (Fig. 3F). Remarkably, siRNA2, which has the highest efficiency to knock down USP4 expression, has the highest potential to decrease RIG-I expression (Fig. 3F). SeV infection greatly enhanced RIG-I protein expression at 8 h and 12 h after infection. However, knockdown of USP4 expression significantly decreased RIG-I protein expression, especially at 8 h after SeV infection (Fig. 3G). In a control experiment, MAVS and TBK1 protein levels were not impaired (Fig. 3G). All together, these data indicate that USP4 targets RIG-I and positively regulates RIG-I-mediated IFN-β signaling through inhibition of RIG-I degradation.

USP4 interacts with RIG-I.

In order to study the molecular mechanisms of USP4 on the stabilization of RIG-I, we investigated the interaction between these two proteins. Myc-USP4 and Flag-RIG-I were cotransfected into HEK293 cells; 24 h after transfection, immunoprecipitation experiments were performed with Myc or Flag antibodies. As shown in Fig. 4A, Flag-RIG-I was coprecipitated with Myc-USP4 using anti-Myc antibody and vice versa. Endogenous interactions between RIG-I and USP4 in HeLa cells transfected with poly(I·C) were examined for various times by immunoprecipitation with USP4 antibody and Western blotting with RIG-I antibody. USP4 was found to interact constitutively with RIG-I (Fig. 4B). In a control experiment, the interaction could not be detected with normal IgG (Fig. 4B).

Fig 4.

USP4 interacts with RIG-I. (A) Lysates from HEK293 cells transiently cotransfected with the Flag-RIG-I and Myc-USP4 expression plasmids were subjected to immunoprecipitation with anti-Myc or anti-Flag antibody followed by Western blot analysis with anti-Flag or anti-Myc antibody, respectively. (B) HeLa cells transfected with poly(I·C) for the indicated time periods were subjected to immunoprecipitation with anti-USP4 or control IgG followed by Western blot analysis with anti-RIG-I antibody. Proteins in whole-cell lysate were used as positive controls (Input). (C) Schematic diagram of RIG-I WT and mutant constructs. RIG-I-N contains the two CARD domains. RIG-I-C contains the middle DExD/H domain and the C-terminal domain. (D) Lysates from HEK293 cells transiently cotransfected with Flag-tagged RIG-I-N and Myc-tagged USP4 expression plasmids were subjected to immunoprecipitation with anti-Myc or anti-Flag antibody followed by Western blot analysis with anti-Flag or anti-Myc antibody, respectively. (E) Western blot analysis of the lysates from HEK293 cells transfected with Flag-tagged RIG-I N-terminal mutant (RIG-I-N) and Myc-tagged USP4 for 24 h. (F) Schematic diagram of USP4 WT and mutant constructs. USP4 wild type (WT) contains a DUSP domain and a C-terminal USP domain. N260 lacks a C-terminal USP domain. C261 lacks the DUSP domain. (G) Lysates from HEK293 cells transiently cotransfected with Flag-tagged RIG-I and HA-tagged USP4 mutants were subjected to immunoprecipitation with anti-Flag antibody followed by Western blot analysis with anti-HA antibody. Similar results were obtained in three independent experiments.

RIG-I has an N-terminal domain composed of two CARD motifs and a C-terminal domain (Fig. 4C). To map the domains that are responsible for the interaction with USP4, two RIG-I truncations were used in the immunoprecipitation assays. As shown in Fig. 4D, USP4 was found to interact with the N-terminal domain of RIG-I. Consistently, overexpression of USP4 also increased the protein expression of the RIG-I mutant RIG-I-N harboring the N-terminal domain (Fig. 4E). USP4 is composed of two domains, the N-terminal dually specific phosphatase (DUSP) domain and the C-terminal USP domain (Fig. 4F). Coimmunoprecipitation experiments showed that RIG-I interacted with WT USP4 and the N-terminal DUSP domain but not with the C-terminal USP domain (Fig. 4G). Taken together, these data indicate that RIG-I and USP4 form a complex through the DUSP domain of USP4 and the N-terminal domain of RIG-I.

USP4 removes K48-linked polyubiquitination conjugate from RIG-I.

To test the role of USP4 in RIG-I ubiquitination, RIG-I was coexpressed with HA-ubiquitin and USP4 or a vector control. RIG-I ubiquitination was easily detected in the setting of HA-ubiquitin and RIG-I transfection. However, cotransfection of USP4 expression plasmid remarkably decreased RIG-I ubiquitination (Fig. 5A), indicating that USP4 has deubiquitination ability toward RIG-I. In order to investigate the forms of USP4-mediated deubiquitination of RIG-I, we performed the transfection assays using the ubiquitin mutant vectors K48 and K63, which contain arginine substitutions of all of ubiquitin's lysine residues except the ones at positions 48 and 63, respectively. As shown in Fig. 5B, transfection of both ubiquitin WT and the K48 and K63 mutants greatly increased the polyubiquitination of RIG-I (lanes 3, 5, and 7), consistent with the reported data that RIG-I can be ubiquitinated through both K48- and K63-linked polyubiquitin chains. However, cotransfection of USP4 greatly decreased the polyubiquitination of RIG-I in the setting of ubiquitin WT and K48-transfected cells (Fig. 5B, lanes 4 and 6) but not in the K63-transfected cells (Fig. 5B, lane 8), indicating that USP4 specifically deubiquitinates K48-linked polyubiquitin chains from RIG-I. In contrast, neither K48-linked nor K63-linked ubiquitination of MAVS was impaired by USP4 overexpression (Fig. 5C). To further confirm that USP4 deubiquitinates RIG-I, USP4 siRNA was transfected into HEK293 cells to knock down endogenous USP4 expression. The level of RIG-I polyubiquitination was greatly increased after transfection of USP4 siRNA in the setting of WT ubiquitin-transfected cells (Fig. 5D, lane 4) compared to the level in control siRNA-transfected cells (lane 3). Importantly, transfection of USP4 siRNA also substantially increased RIG-I polyubiquitination in ubiquitin mutant K48-transfected cells (Fig. 5D, lane 6) but not in mutant K63-transfected cells (Fig. 5D, lane 8). Finally, ubiquitination of endogenous RIG-I was investigated in USP4 siRNA-transfected cells. SeV infection greatly induced K48-linked polyubiquitination of RIG-I, and transfection of USP4 siRNA further increased K48-linked polyubiquitination of RIG-I in SeV-transfected cells (Fig. 5E). All together, these data indicate that USP4 removes K48-linked polyubiquitin chains from RIG-I, leading to the stabilization of RIG-I and the enhancement of RIG-I signaling.

Fig 5.

USP4 removes K48-linked polyubiquitination conjugate from RIG-I. (A) Lysates from HEK293 cells transiently cotransfected with Flag-RIG-I, Myc-USP4, and HA-Ub plasmids were subjected to immunoprecipitation with anti-Flag antibody followed by Western blot analysis with anti-HA antibody. (B) Lysates from HEK293 cells transiently cotransfected with Flag-RIG-I, Myc-USP4, or vector control and HA-Ub (WT), HA-Ub (K48), or HA-Ub (K63) plasmids were subjected to immunoprecipitation with anti-Flag antibody followed by Western blot analysis with anti-HA antibody. (C) Lysates from HEK293 cells transiently cotransfected with Myc-MAVS, Flag-USP4, or vector control and HA-Ub (WT), HA-Ub (K48), or HA-Ub (K63) plasmids were subjected to immunoprecipitation with anti-Myc antibody followed by Western blot analysis with anti-HA antibody. (D) Lysates from HEK293 cells transfected with control siRNA (Ctrl) or USP4 siRNA 2 (siRNA) and Flag-RIG-I and HA-Ub (WT), HA-Ub (K48), or HA-Ub (K63) plasmids were subjected to immunoprecipitation with anti-Flag antibody followed by Western blot analysis with anti-HA antibody. (E) Lysates from HeLa cells transfected with control siRNA (Ctrl) or USP4 siRNA 2 (siRNA) and infected with SeV for 4 h were subjected to immunoprecipitation with anti-RIG-I antibody followed by Western blot analysis with anti-Ub or anti-Ub(K48) antibody. Similar results were obtained in three independent experiments.

USP4 is involved in the cellular antiviral response.

We have demonstrated that USP4 expression is downregulated upon SeV infection and USP4 positively regulates RIG-I-mediated IFN-β signaling. Type I IFNs play critical roles in the innate immune responses against viral infection. Therefore, our results suggest that USP4 might play a very important role in antiviral immunity. To investigate the effect of USP4 on antiviral responses, vesicular stomatitis virus (VSV), a kind of single-stranded RNA (ssRNA) virus recognized by RIG-I, was used to infect HEK293 cells, followed by plaque assays. Plaque assays of HEK293 cells infected with VSV showed that overexpression of USP4 substantially decreased viral replication in these cells compared to control vector-transfected cells (Fig. 6A). Prestimulation of poly(I·C) inhibited VSV replication in control vector-transfected cells (Fig. 6A). Notably, poly(I·C)-induced inhibition of VSV replication was further decreased in USP4-overexpressing cells (Fig. 6A). In order to further confirm the function of USP4 in VSV replication under physiological conditions, we silenced USP4 expression by USP4 siRNA transfection, followed with VSV infection. Plaque assays showed that cells subjected to transfection of USP4 siRNA had greatly increased VSV viral replication in both the absence and presence of poly(I·C) stimulation compared to control siRNA transfected cells (Fig. 6B). Taken together, these data indicate that USP4 positively regulates RIG-I-mediated IFN-β signaling and antiviral immune responses.

Fig 6.

USP4 is involved in cellular antiviral response. (A) HEK293 cells (2 × 10⁵) were transfected with the indicated plasmids (1 μg each). Twenty-four hours later, cells were further transfected with poly(I·C) (0.1 μg) or left untreated. Eighteen hours after poly(I·C) transfection, cells were infected with VSV (MOI, 0.1), and the supernatants were harvested at 12 h postinfection. Supernatants were analyzed for VSV titers with standard plaque assays. Intracellular VSV RNA replicates were measured by quantitative RT-PCR (RT-PCR). **, P < 0.01. Data are representative of three experiments (mean and SD of six samples). (B) HEK293 cells (2 × 10⁵) were transfected with control siRNA (Ctrl) or USP4 siRNA (siRNA) and then treated as described for panel A.

DISCUSSION

Ubiquitination of RIG-I is essential for RIG-I signaling and antiviral immune responses. RIG-I can be modified by both K63-linked polyubiquitination and K48-linked polyubiquitination, leading to the activation and inhibition of RIG-I signaling, respectively. Therefore, it is not surprising that RIG-I deubiquitination is also very important for RIG-I activation. In this regard, the deubiquitinating enzyme cylindromatosis (CYLD) has been reported to remove K63-linked polyubiquitination chains from RIG-I, thereby inhibiting RIG-I-mediated signaling (35). Another deubiquitinating enzyme, USP17, has also been reported to regulate virus-induced type I IFN signaling through deubiquitination of RIG-I and MDA5 (36). Knockdown of USP17 potentiated both K63- and K48-linked ubiquitination of RIG-I. However, the deubiquitinating enzymes that specifically remove K48-linked polyubiquitination chains from RIG-I have not been identified. In this study, we demonstrated that USP4 specifically removed K48-linked polyubiquitination chains from RIG-I, leading to the stabilization of RIG-I, thereby enhancing RIG-I signaling. Furthermore, our data showed that MDA5-mediated IFN-β production was not affected by USP4. Thus, USP4 and USP17 are not functionally redundant in the regulation of RIG-I signaling. To the best of our knowledge, USP4 is the first reported deubiquitinating enzyme to specifically cleave K48-linked ubiquitination chains from RIG-I.

A member of the USP family, USP4 has been observed to repress TLR-, IL-1- and tumor necrosis factor alpha (TNF-α)-induced NF-κB activation by deubiquitinating K63-linked ubiquitin conjugates from TRAF2, TRAF6, and TAK1 (32, 37, 38). In addition to hydrolyzing K63-linked ubiquitination, USP4 also stabilizes molecules by deubiquitinating K48-linked ubiquitination. For example, USP4 interacts directly with and deubiquitinates ADP-ribosylation factor-binding protein 1 (ARF-BP1), leading to the stabilization of ARF-BP1 and subsequent reduction of p53 levels (39). USP4 binds to and deubiquitinates the adenosine A2A receptor and enhances the cell surface level of the receptor (40). As a deubiquitinating enzyme, USP4 was recently found to directly interact with type I transforming growth factor beta receptor (TβRI), leading to increases in the TβRI level at the plasma membrane and in transforming growth factor β (TGF-β) signaling (41). In this study, we demonstrated that USP4 directly interacts with RIG-I through the N-terminal domain and deubiquitinates K48-linked ubiquitination from RIG-I, leading to the stabilization of RIG-I. E3 ligase RNF125 was reported to mediate K48-linked polyubiquitination preferentially in the N-terminal CARDs (26). Consistent with this finding, we demonstrated that USP4 interacted with and stabilized the RIG-I truncation mutant with N-terminal CARDs (RIG-IN).

RIG-I signaling induced the activation of both NF-κB through TRAF6/TAK1/IKK and IRF3 through TBK1, leading to the production of proinflammatory cytokines and IFN-β, respectively (3). However, IRF3 activation and IFN-β production play more important roles than proinflammatory cytokines in the RIG-I-mediated antiviral responses (15). Therefore, USP4 might inhibit RIG-I-induced NF-κB activation through removal of K63-linked ubiquitin chains from TRAF6 and TAK1 downstream of RIG-I, which leads to a decrease in the production of proinflammatory cytokines. In fact, consistent with these reported data, we found that USP4 overexpression inhibited RIG-I-mediated activation of the NF-κB luciferase reporter (data not shown). We speculate that this RIG-I-mediated inhibition of NF-κB activation by USP4 might indicate that USP4 has a greater potential to remove K63-linked ubiquitin from TRAF6/TAK1 than to remove K48-linked ubiquitin from RIG-I. Although NF-κB has been demonstrated to cooperate with IRF3 to induce IFN-β expression, IRF3 activation and subsequent IFN-β production were found to be enhanced by USP4 through deubiquitination and stabilization of RIG-I, which is enough for the inhibition of viral replication. Therefore, in RIG-I signaling, USP4 performed two functions to remove K48-linked polyubiquitin from RIG-I and K63-linked polyubiquitin from TRAF6/TAK1, leading to the enhancement of IRF3 activation and inhibition of NF-κB activation, respectively.

Viruses use a variety of strategies to inhibit the immune response and evade the host immune system. One efficient way is to destroy the receptors that can recognize the virus and provoke the human immune response to limit viral replication. RIG-I is a key PRR for the detection of positive- and negative-stranded RNA viruses. Therefore, promotion of RIG-I degradation represents a novel pathway to limit the immune response and facilitate viral replication. RNF125 was found to mediate K48-linked polyubiquitination and proteasome degradation of RIG-I. Importantly, RNF125 was induced after viral infection (26). Thus, RIG-I will be destroyed through K48-linked ubiquitination by overexpressed RNF125 after viral infection. In this study, we found that USP4 expression was downregulated after viral infection. Because USP4 is a deubiquitinating enzyme for K48-linked polyubiquitination of RIG-I, downregulation of USP4 expression will lead to more K48-linked RIG-I ubiquitination and more proteasomal degradation. As a result, this will lead to the inhibition of IFN-β signaling and increase the viral replication. Therefore, USP4 and RNF125 might play important roles in facilitating viral replication through the regulation of RIG-I protein expression.

Activation of RIG-I signaling also leads to the production of proinflammatory cytokines such as TNF-α and IL-6, which are detrimental to the human body because they cause tissue damage (42, 43). Downregulation of USP4 expression and enhanced degradation of RIG-I also represent protective responses for the host after viral infection. Therefore, identification of USP4 as a positive regulator for RIG-I signaling might provide some therapeutic clues for drug design to prevent excessive inflammation and limit viral infection.