Tripartite Motif-Containing Protein 38 Negatively Regulates TLR3/4- and RIG-I–Mediated IFN-β Production and Antiviral Response by Targeting NAP1

Pattern recognition receptors, including TLRs and RIG-I–like helicases (RLRs), play pivotal roles in defense against viral infection (1, 2). After recognizing microbial conserved pathogen-associated molecule patterns such as LPS, polyinosinic-polycytidylic acid [poly(I:C)], and viral RNA, TLRs and RLRs activate immune cells to produce type I IFN (IFN-α/β) and proinflammatory cytokines, which are involved in the elimination of viral infection (1, 2).

TLR3 and TLR4 initiate IFN-β signaling through the recruitment of TLR/IL-1R domain-containing adaptor protein inducing IFN-β (TRIF) (3–5). RLRs comprise three cytoplasmic DExD–H-box RNA helicases, RIG-I, melanoma differentiation-associated gene 5 (MDA5), and laboratory of genetics and physiology 2 (6, 7). The helicases RIG-I and MDA5 have been found to recognize viral RNAs and poly(I:C) in the cytoplasm and subsequently recruit another antiviral signaling adaptor, mitochondrial antiviral signaling protein (MAVS, also called IPS-1, Cardif, or VISA) to initiate IFN-β signaling (8–11). Recruitment of TRIF and MAVS mainly promotes the activation of TNFR-associated factor 3 (TRAF3)/NF-κB–activating kinase-associated protein (NAP1) and subsequent activation of TRAF family member-associated NF-κB activator (TANK)-binding kinase 1 (TBK1; also called NAK)/ IκB kinase (IKK) ε, leading to the phosphorylation, dimerization, and nuclear translocation of IFN regulatory factor (IRF) 3 and the production of IFN-β (12, 13). Although full activation of TLR and RIG-I signaling and secretion of type I IFNs are important for the elimination of invading microorganisms, uncontrolled expression of type I IFNs has manifested in diverse pathogenic autoimmune diseases, including systemic lupus erythematosus (14). Thus, TLR- and RIG-I–initiated IFN signaling must be tightly regulated.

Tripartite motif (TRIM) family proteins are composed of >70 members in humans (15). TRIM proteins are involved in a broad range of biological processes, including cell differentiation, apoptosis, transcriptional regulation, signal transduction, and immunity (16–21). The characteristic structure of TRIM proteins is the presence of a RING (R) domain, one or two B-boxes (B), and a coiled coil (CC) domain at the N-terminal. The C-terminal region is variable among different TRIM proteins. Although several TRIM proteins have been demonstrated to play very important roles in antiviral immune response, the function of the majority of TRIM proteins remains a mystery.

In this study, we identified TRIM38 (also known as RoRet) as a negative regulator in TLR3/4- and RIG-I–induced IFN-β signaling by targeting NAP1. Specifically, TRIM38 bound to NAP1 and promoted K48-linked polyubiquitination and proteasomal degradation of NAP1. Therefore, our results outline a new manner for the control of TLR and RLR response, and suggest TRIM38 as a potential target for the intervention of autoimmune diseases with uncontrolled IFN-β production.

C57BL/6J mice were obtained from Joint Ventures Sipper BK Experimental Animal (Shanghai, China). All animal experiments were undertaken in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals, with the approval of the Scientific Investigation Board of Medical School of Shandong University (Jinan, China). Mouse macrophage cell line RAW264.7 and human HEK293 cells were obtained from American Type Culture Collection (Manassas, VA). HEK293-TLR3/TLR4 cell lines were obtained from Invivogen (San Diego, CA). Mouse primary peritoneal macrophages were prepared, as described (22). The cells were cultured at 37°C under 5% CO₂ in DMEM supplemented with 10% FCS (Invitrogen-Life Technologies), 100 U/ml penicillin, and 100 μg/ml streptomycin. MG132, chloroquine, and LPS (Escherichia coli, 055:B5) were purchased from Sigma-Aldrich (St. Louis, MO), and LPS was repurified, as described (22). Poly(I:C) was purchased from Invivogen. LPS and poly(I:C) were used at a final concentration of 100 ng/ml and 10 μg/ml, respectively. The Abs specific for hemagglutinin (HA), Ub, β-actin, GAPDH, and protein G agarose used for immunoprecipitation (IP) were from Santa Cruz Biotechnology (Santa Cruz, CA). The Abs specific to Myc, IRF3, TBK1, STAT1, phospho-IRF3 at Ser³⁹⁶, and phospho-STAT1 at Ser⁷²⁷ were from Cell Signaling Technology (Beverly, MA). The Ab for Flag was from Sigma-Aldrich. The Abs for TRIM38 and NAP1 were from Abcam (Cambridge, MA). Their respective HRP-conjugated secondary Abs were purchased from Santa Cruz Biotechnology. Sendai virus was purchased from China Center for Type Culture Collection (Wuhan University).

pCMV6-Flag-TRIM38 (NM_006355) expression plasmid was purchased from OriGene (Rockville, MD). The TRIM38 C16A mutation was generated using the KOD-Plus-Mutagenesis kit (Toyobo, Osaka, Japan). TRIM38 wild-type (WT) and C16A mutant cDNA were cloned in pCMV-HA and pCMV-Myc plasmids (Promega). TANK, IRF3, and NAP1 cDNA were amplified from THP1 cells by PCR and cloned in pCMV plasmid (Promega). All constructs were confirmed by DNA sequencing. IFN-β and IRF3 reporter plasmids and TBK1 and TRIF plasmids were gifts of X. Cao (Second Military Medical University, Shanghai, China). Expression vectors for IKK-ε, HA-Ub WT, and mutant K48 and K63 were from H. Xiao (Institute Pasteur of Shanghai, Chinese Academy of Sciences, Shanghai, China). Flag-SINTBAD expression plasmid was a gift of F. Randow (Medical Research Council Laboratory of Molecular Biology). For transient transfection of plasmids into RAW264.7 cells, jetPEI reagents were used (Polyplus-transfection). For stable selection of cell lines overexpressing TRIM38, transfected RAW264.7 macrophages were selected with G418 (600 μg/ml) and were pooled for further experiments. For transient silencing, duplexes of small interfering RNA (siRNA) were transfected into cells with the Geneporter 2 Transfection Reagent (GTS, San Diego, CA), according to the standard protocol. Target sequences for transient silencing were 5′-GAGGAUCGUCGGCAAACAAUU-3′ (siRNA 1) and 5′-GACUAAAGGUGGAAGAUUAUU-3′ (siRNA 2) for TRIM38; scrambled control sequences were 5′-UUCUCCGAACGUGUCACGU-3′. In vivo siRNA transfection to knockdown TRIM38 expression in peritoneal macrophages was performed, as described (23). Briefly, female C57BL/6J mice (4 wk old) were i.p. injected with thioglycolate to elicit peritoneal macrophages. After 3 d, 6 nmol siRNA was incubated with Geneporter 2 Transfection Reagent, according to manufacturer’s instructions, and then i.p. injected. After 48 h, the mice were treated with i.p. administration of LPS for 1 h, and the secretion of IFN-β in the peritoneal lavage was measured by ELISA.

After cell stimulation, the concentrations of IFN-β in culture supernatants or peritoneal lavage were measured by ELISA kits (R&D Systems, Minneapolis, MN).

Total RNA was extracted with TRIzol reagent, according to the manufacturer’s instructions (Invitrogen). Specific primers used for RT-PCR assays were 5′-CAACAAGTGTCTCCTCCAAAT-3′ (sense) and 5′-TCTCCTCAGGGATGTCAAAG-3′ (antisense) for IFN-β; 5′-GCCCTCGCTGTCATCCTCA-3′ (sense) and 5′-CCCGAACCCATTTCTTCTCTG-3′ (antisense) for RANTES; 5′-ATGGCCTCAACCACCAGC-3′ (sense) and 5′-TCACCGACACTGGGGACAG-3′ (antisense) for TRIM38; and 5′-CAAGGTCATCCATGACAACTTTG-3′ (sense) and 5′-GTCCACCACCCTGTTGCTGTAG-3′ (antisense) for GAPDH. For IP, whole-cell extracts were collected 36 h after transfection and were lysed in IP buffer containing 1.0% (v/v) Nonidet P-40, 50 mM Tris-HCl (pH 7.4), 50 mM EDTA, 150 mM NaCl, and a protease inhibitor mixture (Merck). After centrifugation for 10 min at 14,000 × g, supernatants were collected and incubated with protein G Plus Agarose Immunoprecipitation reagent (Santa Cruz Biotechnology) together with 1 μg monoclonal anti-Flag or 1 μg anti-HA. After 6 h of incubation, beads were washed five times with IP buffer. Immunoprecipitates were eluted by boiling with 1% (w/v) SDS sample buffer. For Western blot analysis, immunoprecipitates or whole-cell lysates were loaded and subjected to SDS-PAGE, transferred onto nitrocellulose membranes, and then blotted, as described previously (22).

Luciferase activity was measured with the Dual-Luciferase Reporter Assay system, according to the manufacturer’s instructions (Promega), as described (24). Data were normalized for transfection efficiency by division of firefly luciferase activity with that of Renilla luciferase.

For analysis of the ubiquitination of NAP1, HEK293 cells were transfected with Flag-NAP1, HA-Ub (WT), or HA-Ub mutants and Myc-TRIM38 WT or Myc-TRIM38 C16A, and then whole-cell extracts were immunoprecipitated with anti-Flag and analyzed by immunoblot with anti-HA Ab.

Vesicular stomatitis virus (VSV) plaque assay was performed, as described (25). 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 of 0.1). At 1 h postinfection, cells were washed with PBS for 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 overlayed. At 3 d postinfection, 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 log₁₀ (PFU/ml). Total cellular RNA was extracted from HEK293, or macrophages transfected with VSV and VSV RNA replicates were examined by quantitative RT-PCR, as described (26). Primers used for VSV replicates were as follows: 5′-ACGGCGTACTTCCAGATGG-3′ (sense) and 5′-CTCGGTTCAAGATCCAGGT-3′ (antisense).

All data are presented as mean ± SD of three or four experiments. Statistical significance was determined with the two-tailed Student t test, with a p value <0.05 considered statistically significant.

TRIM38 is encoded within the MHC class I region, a region containing a large number of genes relevant to the immune response (27). We previously demonstrated TRIM38 as a negative feedback regulator for TLR-induced production of proinflammatory cytokines by targeting TRAF6 for ubiquitination and degradation (28). However, the function of TRIM38 especially in the antiviral immune response remains largely unknown. To investigate whether TRIM38 plays a role in antiviral immunity, we initially examined the effects of TRIM38 on TLR3- and TLR4-mediated IFN-β production in macrophages. Two siRNAs were transfected into primary peritoneal macrophages to suppress endogenous TRIM38 expression. The expression of TRIM38 protein was greatly decreased as measured by Western blotting with transfection of TRIM38-specific siRNA 1 and 2 (Fig. 1A). After transfection of TRIM38 siRNA, macrophages were with stimulated with LPS (TLR4 ligand) and poly(I:C) (TLR3 ligand), respectively. As shown in Fig. 1B, TRIM38 knockdown significantly increased LPS- and poly(I:C)-induced IFN-β production in mouse peritoneal macrophages. TRIM38 siRNA 2, which has a higher efficiency to knock down TRIM38 protein expression (Fig. 1A), has a greater potential to increase the TLR3- and TLR4-induced IFN-β production (Fig. 1B). Therefore, TRIM38 siRNA 2 was used in the following experiments.

To further confirm the negative regulatory role of TRIM38 on TLR3/4-mediated IFN-β production, RAW264.7 stable cell line with TRIM38 overexpression was constructed by transfecting Flag-TRIM38 expression plasmid. Overexpression of TRIM38 was confirmed by Western blotting with both TRIM38 and Flag Abs (Fig. 1C). As shown in Fig. 1D, TRIM38 overexpression significantly decreased LPS- and poly(I:C)-induced IFN-β production in RAW264.7 cells.

To investigate TRIM38 inhibiting LPS- and poly(I:C)-induced IFN-β expression at the transcriptional level, TRIM38 expression plasmid and IFN-β promoter luciferase reporter were cotransfected into RAW264.7 cells. As shown in Fig. 1E, TRIM38 overexpression significantly decreased LPS- and poly(I:C)-induced IFN-β promoter activation in RAW264.7 cells. Similarly, LPS- and poly(I:C)-induced IFN-β luciferase activation was also greatly attenuated by TRIM38 overexpression in HEK293/TLR4 and HEK293/TLR3 cells, which stably expressed TLR4 and TLR3, respectively (Fig. 1F).

Recognition of RNA virus through RIG-I may lead to the expression of IFN-β. To investigate the function of TRIM38 on RIG-I–mediated IFN-β signaling, poly(I:C) was transfected into HEK293 cells, which has been shown to activate IFN-β production through RIG-I and MDA-5 (29). As shown in Fig. 1G, transfection of poly(I:C) could induce the expression of IFN-β, RANTES, and TNF-α in HEK293 cells. Overexpression of TRIM38 greatly attenuated poly(I:C)-induced expression of IFN-β and RANTES. Collectively, these data indicate that TRIM38 negatively regulates TLR3/4- and RIG-I–induced IFN-β signaling.

Because TRIM38-deficient mice are not available, in vivo TRIM38 siRNA i.p. transfection was performed to knock down TRIM38 expression in thioglycolate-elicited peritoneal cells with Geneporter 2 Transfection Reagent. TRIM38 siRNA i.p. transfection successfully suppressed TRIM38 expression in i.p. cells (Fig. 2A). After i.p. transfection of TRIM38 siRNA, LPS was administrated i.p., and the secretion of IFN-β in peritoneal lavage was measured by ELISA. As shown in Fig. 2B, knockdown of TRIM38 expression significantly increased LPS-induced IFN-β production in peritoneal lavage. These findings demonstrate that TRIM38 could inhibit LPS-induced IFN-β production in vivo.

IRF3 is the key transcription factor that is responsible for the expression of IFN-β in TLR3/4 and RIG-I signaling. Activation of TLR3/4 and RIG-I led to the phosphorylation of IRF3 through a TBK1-dependent pathway, resulting in IRF3 dimerization and subsequent nuclear translocation and binding to the IFN-β promoter. To investigate the effect of TRIM38 on IRF3 activation, IRF3 cis-reporting plasmids were transfected into HEK293/TLR3 and HEK293/TLR4 cells. As shown in Fig. 3A, LPS and poly(I:C) greatly increased IRF3 activation in HEK293/TLR4 and HEK293/TLR3 cells, respectively. However, transfection of TRIM38 expression plasmid substantially attenuated LPS- and poly(I:C)-induced IRF3 activation.

To investigate TRIM38 inhibition of IRF3 activation under physiological conditions, endogenous TRIM38 expression was silenced with siRNA transfection in mouse peritoneal macrophages and IRF3 phosphorylation was measured upon LPS stimulation. Knockdown of endogenous TRIM38 expression substantially enhanced LPS-induced phosphorylation of IRF3 (Fig. 3B). Consistently, TRIM38 overexpression significantly decreased LPS-induced phosphorylation of IRF3 in RAW264.7 cells (Fig. 3C).

Virus-induced IFN-β production can activate the transcription factor STAT1. Phosphorylated STAT1 conjugates with STAT2 and IRFs to generate the IFN-stimulated gene factor 3 transcription complex to regulate the expression of IFN-stimulated genes. Consistent with the inhibitory function of TRIM38 on IFN-β production, knockdown of endogenous TRIM38 expression greatly increased LPS-induced STAT1 phosphorylation in mouse peritoneal macrophages (Fig. 3B). In contrast, TRIM38 overexpression inhibited LPS-induced STAT1 phosphorylation (Fig. 3C). All together, these data suggest that TRIM38 inhibits TLR-induced IRF3 activation and subsequent IFN-β signaling.

TLR3/4 and RIG-I recognize the microbial patterns and recruit adaptors TRIF and MAVS, respectively. TRIF and MAVS mediate IRF3 activation and IFN-β production through a complex cascade composed of various molecules, including TRIF, MAVS, TBK1, IKKε, and IRF3. To determine the molecular order and molecular targets of TRIM38 in TLR- and RIG-I–induced IFN-β signaling, the effects of TRIM38 overexpression on IFN-β promoter activation mediated by various molecules were examined in reporter assays. As shown in Fig. 4A and 4B, TRIF-, RIG-I–, and MAVS-induced IFN-β promoter activation and IRF3 activation were significantly inhibited by TRIM38 overexpression in a dose-dependent manner. However, TBK1 was still able to activate IFN-β promoter activation and IRF3 activation in the presence of TRIM38 expression plasmid (Fig. 4A, 4B). Similarly, TRIM38 expression could not inhibit IKKε-induced IRF3 activation (Fig. 4B). Therefore, we conclude that TRIM38 targets molecules upstream of TBK1 to inhibit the signal transduction.

Recently, Sasai et al. (30, 31) reported that NAP1 is a subunit of the kinase complex IKKε/TBK1 and can synergize with MAVS, TRIF, and TBK1 to increase IRF3 activation. Consistent with their data, we found cotransfection of NAP1 with TRIF or TBK1 greatly increased TRIF- and TBK1-induced IFN-β promoter activation (Fig. 4C). However, this activation was significantly inhibited with TRIM38 overexpression in a dose-dependent manner (Fig. 4C). Transfection of NAP1 alone could slightly induce IFN-β promoter activation, and this activation was also inhibited with TRIM38 overexpression (Fig. 4D). TANK and SINTBAD (similar to NAP1 TBK1 adaptor) are another two adaptors that can bind TBK1 and IKKε. Consistent with reported data, we found cotransfection of TANK or SINTBAD with TBK1 substantially increased TBK1-induced IFN-β promoter activation (Fig. 4E). In contrast to NAP1, this activation was not inhibited with TRIM38 overexpression (Fig. 4E). Taking all these reporter data together, we speculate NAP1 may be a TRIM38 target.

To confirm that TRIM38 targets NAP1, the function of TRIM38 on the degradation of the molecules in TLR3/4 and RIG-I signaling was investigated. As shown in Fig. 5A and 5B, TRIM38 promoted degradation of NAP1 in a dose-dependent manner. In contrast, TRIM38 overexpression had no effects on TBK1, TANK, IKKε, SINTBAD, STING, and IRF3 expression (Fig. 5B, 5C). To further confirm TRIM38-mediated NAP1 degradation under physiological condition, TRIM38 expression was silenced by TRIM38 siRNA transfection. TRIM38 knockdown greatly increased NAP1 protein level in peritoneal macrophages (Fig. 5D). As a control, TBK1, IRF3, and STAT1 protein levels were not impaired (Fig. 5D). All together, these data indicate that TRIM38 targets NAP1 for degradation to inhibit IFN-β production.

To understand the mechanisms by which TRIM38 degrades NAP1 expression, we first investigated the interaction between TRIM38 and NAP1. HA-TRIM38 and Flag-NAP1 were cotransfected into HEK293 cells; IP experiments were performed with HA or Flag Abs. Flag-NAP1 was coprecipitated with HA-TRIM38 and vice versa (Fig. 6A, 6B).

TRIM38 contains a cluster of domains composed of a RING finger, two B-boxes, and a C-terminal PRY/SPRY (16). The presence of RING-finger domain indicates TRIM38 may function as an E3 ligase. To test the role of TRIM38 in NAP1 ubiquitination, NAP1 was cotransfected with HA-ubiquitin and WT TRIM38 into HEK293 cells. Without MG-132, NAP1 ubiquitination was hardly detectable (data not shown). After MG-132 treatment, the level of NAP1 ubiquitination was markedly increased in the presence of TRIM38 expression plasmid (Fig. 6C, lane 3). Importantly, the TRIM38 point mutation (C16A) with substitution of the cysteine residue at position 16 within the RING domain with alanine lost the ability to promote polyubiquitination of NAP1 (Fig. 6C, lane 4), indicating TRIM38 could promote the ubiquitination of NAP1 though the RING-finger domain. To study the forms of TRIM38-mediated NAP1 polyubiquitination, ubiquitin mutant vectors K48 and K63, which contain arginine substitutions of all of its lysine residues except the one at positions 48 and 63, respectively, were used in the transfection assays. As shown in Fig. 6D, TRIM38-mediated NAP1 polyubiquitination could be detected in the presence of WT and K48 plasmid (lanes 2 and 4), but not with K63 plasmid (lane 6), indicating TRIM38 medicates K48-linked polyubiquitination of NAP1. K48-linked protein ubiquitination leads to the degradation of the corresponding protein by 26S proteasome. Consistently, TRIM38-induced degradation of NAP1 protein could be reversed by proteasome inhibitor MG-132, but not by lysosome inhibitor chloroquine (Fig. 6E). Accordingly, the TRIM38 mutant C16A lost the ability to inhibit TRIF-induced IFN-β activation, compared with WT TRIM38 (Fig. 6F). All together, these data demonstrate that TRIM38 interacts with NAP1 and promotes K48-linked polyubiquitination and proteasomal degradation of NAP1.

Type I IFNs play critical roles in the innate immune responses against viral infection. The fact that TRIM38 negatively regulates IFN-β production prompted us to investigate the function of TRIM38 in antiviral immunity. TRIM38 protein expression in primary macrophages was measured postinfection with Sendai virus (SeV). SeV infection substantially induced the expression of TRIM38 in macrophages at different time points after SeV infection (Fig. 7A), suggesting TRIM38 expression was modulated by virus infection. To investigate the function of SeV-induced TRIM38 expression, production of IFN-β and TNF-α was measured by ELISA, followed by SeV infection in TRIM38 siRNA-transfected macrophages. As shown in Fig. 7B, SeV infection greatly induced the production of IFN-β, whereas production of IFN-β was further increased after TRIM38 siRNA tranfection. Similarly, TNF-α production was also greatly increased in TRIM38 siRNA-transfected macrophages upon SeV infection (Fig. 7B). To directly investigate the effect of TRIM38 on antiviral responses, VSV, a kind of ssRNA virus recognized by RIG-I, was used to infect HEK293 cells and macrophages. Plaque assay of HEK293 cells infected with VSV showed that overexpression of TRIM38 substantially increased viral replication in the presence or absence of poly(I:C) (Fig. 7C). In sharp contrast, TRIM38 mutant C16A, which lost the ability to promote ubiquitination and degradation of NAP1, could not increase viral replication (Fig. 7C). Similarly, VSV RNA replicates in HEK293 cells were greatly increased in TRIM38-transfected cells, compared with control vector- or TRIM38 C16A-transfected cells, as measured by quantitative RT-PCR (Fig. 7C). To further confirm the function of TRIM38 on VSV replication under physiological conditions, TRIM38 expression was silenced by TRIM38 siRNA transfection in mouse peritoneal macrophages, and then the macrophages were infected with VSV. Plaque assay showed that transfection of TRIM38 siRNA greatly decreased VSV viral replication in macrophages in the presence or absence of poly(I:C) (Fig. 7D). Accordingly, knockdown of TRIM38 significantly decreased intracellular VSV RNA replicates, compared with control siRNA-transfected macrophages (Fig. 7D). Taken together, these data indicate that virus-induced TRIM38 negatively regulates production of IFN-β and antiviral immune responses, thus facilitating viral invasion of the immune system.

Innate immunity plays an essential role in the control and elimination of viral infection through the production of type I IFNs (7). Production of type I IFNs requires the recognition of viral pathogen-associated molecule patterns by TLRs and RIG-I–like receptors. TLR3/4 and RIG-I use TRIF and MAVS (also known as Cardif, IPS-1, or VISA) to mediate IFN-β signaling, respectively. Downstream of TRIF and MAVS, both TLR3/4 and RIG-I could activate a similar pathway that involves NAP1, TBK1, and IRF3 to induce the production of IFN-β (6, 7). Although production of type I IFNs is essential for the antiviral immune responses, uncontrolled production is harmful and has been found in diverse phathogenic autoimmune diseases, including systemic lupus erythematosus (14). Therefore, TLR3/4- and RIG-I–mediated IFN-β signaling must be tightly controlled.

TRIM family proteins are composed of >70 members in humans (15). Several TRIM proteins have been reported to play key roles either positively or negatively in TLR- or other pattern recognition receptor-mediated signaling and antiviral immunity through ubiquitination (17–21). TRIM25 interacts with the caspase-recruitment domain of RIG-I through its PRY/SPRY domain, which results in the K63-linked ubiquitylation of RIG-I and, consequently, the induction of type I IFN production and NF-κB activity (32). TRIM56 was required for dsDNA virus-triggered signaling through K63-linked polyubiquitination of STING (33). TRIM23 has been shown to catalyze K27-linked ubiquitination of NEMO, which is crucial for IRF3 and NF-κB activation downstream of TLR3 and RIG-I (34). TRIM21 has been shown to mediate K48-linked ubiquitination of various IRFs, and thereby directly inhibit production of type I IFNs (35–37).

In the current study, we have identified TRIM38 as a new regulator in TLR3/4- and RIG-I–mediated IFN-β signaling. Overexpression of TRIM38 inhibited TLR- and RIG-I–induced activation of IRF3 and production of IFN-β and downstream STAT1 phosphorylation. Moreover, siRNA knockdown of TRIM38 expression potentiated TLR- and RIG-I–induced activation of IRF3 and production of IFN-β. Relevant to the function of IFN-β in antiviral immunity, siRNA knockdown of TRIM38 or overexpression significantly inhibited or enhanced the replication of VSV, respectively. Therefore, to our knowledge, our study provided the first evidence to demonstrate that TRIM38 negatively regulates both TLR3/4- and RIG-I–induced IFN-β production, thus facilitating the viral immune evasion.

TRIM38, a member of TRIM family, is encoded within the MHC class I region (27). The characteristic structure of TRIM proteins is the presence of a RING (R) domain, one or two B-boxes (B), and a coiled coil (CC) domain (16). The presence of the RING domain indicates TRIM proteins may work as E3 ligases. Indeed, we have demonstrated that TRIM38 targets TRAF6 for ubiquitination and degradation to negatively regulate TLR-induced production of proinflammatory cytokines (28). In this study, we provided evidence to demonstrate that TRIM38 could also act as an E3 ligase to promote the ubiquitination and proteasomal degradation of NAP1. NAP1 is an adaptor protein for the kinase IKKε/TBK1 and plays a very important role in TLR3/4- and RLR-induced IRF3 activation and IFN-β production (30, 31, 38, 39). Activation of TLR7/8/9 could also lead to the production of type I IFNs. We found that siRNA knockdown of TRIM38 expression in macrophages greatly increased R848-induced IFN-β production (data not shown), indicating TRIM38 also negatively regulates IFN signaling in the TLR7/8/9 pathway. Given the fact that TLR7/8/9-mediated production of type I IFNs requires TRAF6 (40), we concluded that TRIM38-mediated ubiquitination and degradation of TRAF6 may account for the inhibitory effect of TRIM38 on TLR7/8/9-mediated IFN signaling. Thus, TRIM38-mediated ubiquitination and degradation of NAP1 and TRAF6 represent a new mechanism to terminate the excessive production of type I IFNs.

TANK and SINTBAD are another two adaptor proteins that were described to specifically bind TBK1/IKKε and participate in TLR/RLR-induced IFN production (41, 42). In contrast to NAP1, we found that TANK- and SINTBAD-induced IFN-β activation was not inhibited by TRIM38. Importantly, TRIM38 could not promote the degradation of TANK and SINTBAD. Therefore, TRIM38 specifically targets NAP1 for degradation to regulate the TLR3/4- and RLR-mediated IFN signaling. NAP1 as well as SINTBAD and TANK have been reported to bind TBK1 and IKKε via a TBK1/IKKε binding domain (42). Another conserved domain among these three adaptors is coiled coils-forming domain (CC). The different substrate specificity for TRIM38 toward NAP1, TANK, and SINTBAD may arise from the different binding activity between TRIM38 and the three adaptors. We have confirmed the binding between TRIM38 and NAP1. However, whether the interaction between TRIM38 and TANK and SINTBAD is present needs to be confirmed. Another possibility for the different substrate specificity may be the ubiquitin-receiving lysine residues that are different among these three adaptors. Therefore, the exact mechanism deserves further investigation.

In conclusion, we identified TRIM38 as a critical negative regulator of TLR3/4- and RIG-I–induced IFN-β production by targeting NAP1 for ubiquitination and degradation. Given the important roles of IFN-β in host antiviral immunity, virus may modulate TRIM38 expression to escape immune surveillance and facilitate the viral replication. Indeed, we found poly(I:C) treatment or SeV infection greatly increases TRIM38 expression. In contrast, excessive IFN-β production has manifested in diverse pathogenic autoimmune diseases (14). Our results provide a strategy to downregulate IFN-β production and suggest that TRIM38 may have therapeutic potential for the intervention of autoimmune diseases with uncontrolled IFN-β production.

We thank Drs. Xuetao Cao, Felix Randow, and Hui Xiao for providing plasmids.

The authors have no financial conflicts of interest.