An Alternative Splicing Isoform of MITA Antagonizes MITA-Mediated Induction of Type I IFNs

The innate immune response plays a vital role in the host defense against pathogen invasion. It detects microbial invasion through the recognition of pathogen-associated molecular patterns by pattern recognition receptors (PRRs) and further activates the expression of type I IFNs and proinflammatory cytokines (1–4). Viral nucleic acid either as the viral genome or as replication intermediates serves as a common pathogen-associated molecular pattern to activate the innate immune response. A number of PRRs have been found to sense viral nucleic acid, including TLR3/7/8 (3, 5, 6) and retinoic acid–inducible gene (RIG)-I–like receptors (3, 6, 7) as viral RNA sensors, and TLR9, DNA-dependent activation of IFN regulatory factors (8), absent in melanoma 2 (9), RNA polylerase III (3), leucine-rich repeat (in FLII)–interacting protein 1 (10), DEAD box polypeptide 41 (11), IFN-γ–inducible protein 16 (12), and cGAMP synthase (13, 14) as viral DNA sensors (15). Upon binding to viral nucleic acid, these PRRs initiate signal transduction pathways leading to the activation of IFN regulatory transcription factor (IRF)3 and NF-κB, as well as the induction of type I IFNs.

Mediator of IRF3 activation (MITA) (also known as STING (16), TMEM173, MPYS (17), or ERIS (18)), which was independently identified by several groups, has been found to be an adaptor protein that links upstream pathogen sensing to downstream IFN induction. It is expressed in multiple tissues and cell lines and plays an important role in cytosolic nucleic acid–mediated induction of type I IFNs by activation of IRF3. Recently, it has been shown that the phosphorylation and ubiquitination of MITA by TANK-binding kinase 1 (TBK1) and tripartite motif containing 56, respectively, were important for MITA to exert its function (19, 20). The C-terminal tail of MITA has been demonstrated to function as a scaffold where IRF3 is phosphorylated by TBK1 (21). The MITA/TBK/IRF3 signaling axis is important in several antiviral innate immune pathways, such as retinoic acid–inducible gene I (RIG-I)–like receptor–mediated and some DNA sensor–mediated signaling pathways (11, 12, 19, 22).

The antiviral innate immune response is tightly controlled by a complex regulatory system to prevent harmful effects resulting from excessive activation. Several molecules, such as dihydroxyacetone kinase (23), suppressor of IKBKE (24), NLR family member X1) (25), and proteasome subunit α type 7 (26), have been identified as negative regulators in innate immune signaling pathways. In the case of MITA, it has been reported that as the E3 ligase, RNF5 targets MITA for ubiquitination and degradation. Additionally, autophagy related 9A, an autophagy related protein, inhibits the MITA–TBK1 interaction after dsDNA stimulation (27, 28). MITA is also a target molecule for microbial pathogens to escape the innate immune response. It has been shown that yellow fever virus, dengue virus, and hepatitis C virus nonstructure proteins target MITA to suppress MITA-mediated IFN induction (29–32).

In this study, we identified an alternatively spliced isoform of MITA, designated as MITA-related protein (MRP), and examined the function of MRP in the context of MITA-mediated signaling. Our results suggested that MRP acts largely as a dominant negative mutant of MITA and blocks MITA-mediated IFN induction via TBK1/IRF3 by disrupting the MITA–TBK1 interaction. However, MRP retained the ability to activate NF-κB. Our findings provide a new mechanism for the negative regulation of MITA-mediated signaling.

The human MITA, MRP and C-terminal truncation mutant of MITA (MRM, 1–253 aa) sequences were amplified by RT-PCR from 293T cells with specific primers (Table I) and confirmed by sequencing. These genes, together with RIG-I, virus-induced signaling adaptor (VISA), TBK1, inducible inhibitor of NF-KB kinase (IKKi), and IRF3 (provided by Prof. Hongbing Shu) were cloned into pcDNA3.1 and pXJ40-hemagglutinin (HA)/FLAG and named as pMITA, pMRP, pMRM, pHA-MITA, pHA-MRP, pHA-MRM, pFlag-MITA, pFlag-MRP, and pFlag-MRM. pMITA–enhanced GFP (EGFP), pMRP-EGFP, pEGFP-MITA and pEGFP-MRP were generated by cloning MITA and MRP into pEGFP-N1 or pEGFP-C1, respectively. The plasmids pRK-FLAG-TRIF and pRK-FLAG-MyD88 were provided by Prof. Hongbing Shu. The reporter plasmids pISRE-luc, pNF-κB-luc, pIRF3-luc, and pIFN-β-luc were purchased from Clontech. pRL-TK was purchased from Promega.

HeLa, 293T, Huh7, and HCT116 cells were cultured in DMEM supplemented with 10% FBS (Life Technologies) and 1× antibiotic-antimycotic (Life Technologies). All cells were cultured under humidified conditions with 5% CO₂ at 37°C.

Sendai virus (SeV) (provided by Prof. Hanzhong Wang), HSV-1 BAC with GFP and HSV-1 (provided by Prof. Chunfu Zheng) (33), and vesicular stomatitis virus (VSV)-GFP (provided by Prof. Hualan Chen) (34) were amplified, titrated, and used according to the standard protocols.

Rabbit polyclonal Abs against human MRP were produced by a commercial service (Abmart, Shanghai, China), and the specificity of this Ab was determined by Western blot (data not shown). Other Abs used in the present study include mouse mAbs against β-actin (Santa Cruz Biotechnology); anti-HA and anti-FLAG (Sigma-Aldrich); rabbit monoclonal Abs against HA (Sigma-Aldrich); anti-TBK1, anti–phospho-NF-κB p65 (Ser⁵³⁶), anti–NF-κB p65, and anti–phospho-IRF3 (Ser³⁹⁶) (Cell Signaling Technology); anti-IRF3 and anti-GAPDH (Proteintech); anti-VISA and anti-MITA (Abcam); and Alexa Fluor 488– and Alexa Fluor 568–conjugated secondary Abs (Invitrogen). Dynabeads protein A (Invitrogen), protein G-Sepharose 4 Fast Flow (GE Healthcare), cyclic diguanylate (c-diGMP; Biolog), and Hoechst 33258 (Beyotime) were provided by their respective suppliers. The small interfering (si)RNAs targeting human MRP and MITA were synthesized by GenePharma. The target sequence are as follows: siMRP-1, 5′-GCG GAA CCT GCA GAT GAC-3′; siMRP-2, 5′-GCA GCG GAA CCT GCA GAT-3′; siMRP-3, 5′-CGG GCA GCG GAA CCT GCA-3′; siMITA-1, 5′-ACT CTT CTG CCG GAC ACT T-3′; siMITA-2, 5′-CTC TTC TGC CGG ACA CTT G-3′; and siMITA-3, 5′-TAC AGT CAA GCT GGC TTT A-3′.

Transient transfection of 293T cells or HCT116 cells was performed with Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. Cells were seeded at a density of 1 × 10⁵/well in 24-well plates and transfected with the indicated amount of expression plasmids combined with 100 ng of the indicated reporter plasmids and 10 ng pRL-TK. The total DNA concentration for each transfection was kept constant by supplementing with the empty vector pXJ40-HA. For some experiments, cells were infected with SeV 18–24 h posttransfection (hpt). The reporter activity was measured with the Dual-Luciferase reporter assay system (Promega). Firefly luciferase activity was normalized to Renilla luciferase activity. All experiments were performed in triplicate and were repeated at least three times.

Total RNA was extracted from cultured cells with TRIzol reagent (Invitrogen). The first-strand cDNA was prepared by using the oligo(dT)15 primer (Promega) and Moloney murine leukemia virus reverse transcriptase (Promega) according to the manufacturer’s instructions. The primers used for PCR are listed in Table I.

Quantitative real-time RT-PCR was performed using the QuantiTect SYBR Green RT-PCR kit (Qiagen) on the ABI StepOne real-time PCR system (Applied Biosystems). Primers used are listed in Table I. The copy numbers of specific mRNAs were normalized according to the mRNA level of GAPDH in each sample. All experiments were performed in triplicate and repeated at least three times.

Immunoprecipitation (IP) and coimmunoprecipitation were performed as follows: total cellular proteins dissolved in IP buffer (50 mM Tris [pH 7.5], 1 mM EGTA, 1 mM EDTA, 1% Triton X-100, 150 mM NaCl, 2 mM DTT, 100 μM PMSF, 1 μg/ml proteinase inhibitor) were quantified using the Bradford kit (Bio-Rad). Cell lysates containing 30 μg protein were preserved in 2× loading buffer as whole cell lysate. From each sample, 1000 μg cellular protein was added to the mixture of protein A or protein G and the corresponding Ab and then incubated in a mixer at 4°C overnight. The mixtures were centrifuged, washed with IP buffer four times, denatured by boiling in SDS-PAGE sample buffer for 5–10 min, and subjected to SDS-PAGE. Immunoblot (IB) analysis was performed as previously described (35). An Ultra ECL kit, Western Lightning Ultra, was purchased from PerkinElmer.

To determine the subcellular location of MRP and MITA, transfected 293T cells were incubated with the MitoTracker Red or ER-Tracker Red dye (Invitrogen) for 15 min. The cells were then covered with normal growth medium and observed with a PerkinElmer UltraVIEW VoX confocal microscope. Indirect immunofluorescence was performed as described previously (36, 37).

The statistical significance of the obtained data was analyzed using a two-tail unpaired t test in GraphPad Prism (GraphPad Software, San Diego, CA). A p value <0.05 was considered statistically significant. Data are presented as the means ± SD.

During cloning of the full-length cDNA of human MITA by RT-PCR using RNAs from 293T cells, a cDNA clone with a short sequence was identified. Sequence analysis of the clone revealed that it contained a splice variant of MITA lacking exon 7. The absence of exon 7 resulted in a frame shift at nt 759, and the coding sequence ends at a stop codon at nt 850 (Fig. 1A). The putative protein encoded by this splice variant contains 283 aa and was designated as MRP (GenBank accession no. KF430638). The N-terminal portion of MRP is identical to aa 1–253 of MITA but possesses a unique 30-aa sequence at the carboxyl terminal (Fig. 1B).

To examine the expression pattern of MITA and MRP in different cell and tissue types, a pair of primers, MITA/MRP-F and MITA/MRP-R, was designed to distinguish MITA and MRP (Fig. 1A, Table I). The PCR products derived from the cDNAs of MITA and MRP have lengths of 405 and 218 bp, respectively. PCR analysis with the cDNA from different cell lines, including 293T, Huh7, and HeLa cells, and different human tissues showed that MITA and MRP transcripts were detectable in the tested cell types as well as in the liver, small intestine, and other organs and tissues (Fig. 1C, 1E). To further confirm the expression of MRP protein, a specific rabbit polyclonal Ab to MRP was generated using two peptides corresponding to the 253–262 aa and 272–283 aa region in the unique C-terminal portion of MRP. The specificity of the anti-MRP Ab was first verified via Western blot with samples of 293T cells overexpressing MITA or MRP (data not shown). The endogenous expression of MRP protein in different cell types at the baseline was low and could only be directly detected by Western blot using an Ultra ECL kit (Fig. 1D). Taken together, these results demonstrate that MRP is expressed in several different cell lines and human tissues, although at a low baseline level.

The absence of the C-terminal portion of MITA in MRP indicates that MRP may be functionally defective. Therefore, we compared the functionality of MITA and MRP. Reporter assays showed that MITA induced the IFN-β promoter and ISRE-, IRF3-, and NF-κB–dependent luciferase expression in a dose-dependent manner, whereas MRP was only able to activate NF-κB–dependent luciferase expression (Fig. 2A). To exclude the possibility that the newly identified C-terminal sequence affects the functionality of MRP, we constructed a truncated mutant of MITA, termed MRM, containing only aa 1–253 (Fig. 1A). Reporter assays with MRM indicated that MRM activated NF-κB–dependent gene expression but did not induce the IFN-β promoter, ISRE, or IRF3, similar to MRP (data not shown). Similarly, MRP overexpression induced the phosphorylation of NF-κB but not of IRF3, whereas MITA overexpression resulted in the phosphorylation of both factors (Fig. 2D, lanes 1–3). These results demonstrated that MRP was defective in triggering IFN-β but retained the ability to activate NF-κB–dependent signaling.

Defective components in the cellular signaling pathways play an important role in the regulation of signaling events. Therefore, we investigated whether MRP interferes with MITA-mediated signaling. Increasing amounts of pMRP were cotransfected with pMITA in the reporter assays. Fig. 2B shows that the coexpression of MRP, even at a low concentration, inhibited the induction of the IFN-β promoter and the expression of the IRF3- and ISRE-dependent reporters by MITA. In contrast, NF-κB–dependent reporter expression was enhanced by coexpression of MRP with MITA. The same results were obtained using pMRM instead of pMRP (data not shown). Similarly, expression of the IFN-stimulated genes (ISGs) IP10, IFIT1, and ISG15, which are mediated by IRF3 activation, was induced by MITA but inhibited by MRP coexpression (Fig. 2C). MITA-induced TNF-α expression, a marker for NF-κB activation, was enhanced by MRP coexpression (Fig. 2C). Western blot analysis also verified that MRP coexpression reduced MITA-induced IRF3 phosphorylation but not MITA-induced NF-κB phosphorylation (Fig. 2D). Alternatively, silencing of MRP by specific siRNA target MRP but not MITA (data not shown) increased MITA-mediated IFN-β activation in a dose-dependent manner, indicating a role of endogenous MRP in the MITA signaling (Fig. 2E and data not shown). MRP expression had no inhibitory effect on RIG-I–, VISA-, TBK1-, TRIF-, and MyD88-induced IFN-β signaling, as indicated by the reporter assays (data not shown). Thus, MRP specifically inhibits MITA-induced activation of IRF3-dependent but not NF-κB–dependent signaling.

It is known that the dimerization of MITA occurs for IFN induction, and MITA was shown to be a scaffold protein that interacts with some components of the cellular signaling pathways, including RIG-I, VISA, TBK1, IKKi, and IRF3 (18–20). As MRP contains a dimerization domain (Fig. 1B), we hypothesized that MRP may form heterodimers with MITA and thereby change the subsequent interaction of MITA with other signaling proteins, leading to NF-κB–dependent signaling but not IRF3-dependent signaling. We therefore attempted to demonstrate a direct interaction between MITA and MRP. In cotransfection experiments, MITA, MRP, and MRM could be coimmunoprecipitated with each other, indicating that they formed homodimers themselves or heterodimers with each other (Fig. 3A). This observation is consistent with the fact that the dimerization of MITA is mediated by the domains comprising the aa segments 153–173 or 155–180 (38–40). Additionally, MRP and MITA were coimmunoprecipitated with RIG-I, VISA, IKKi, and IRF3 (Fig. 3B). However, MRP and MRM could not be coimmunoprecipitated with TBK1, in contrast to MITA (Fig. 3B and data not shown). In an additional experiment, the expression plasmids pMITA and pMRP were cotransfected into 293T cells, and the binding of MITA and MRP to TBK1 was examined by coimmunoprecipitation. Increasing the expression of MRP reduced the coprecipitation of MITA by TBK1 (Fig. 3C), indicating that MRP disrupted the interaction between MITA and TBK1. Because MITA provides a platform for the interaction of TBK1 and IRF3, we further determined the role of MRP in TBK1 and IRF3 interaction using coimmunoprecipitation experiments. Transfection of pMITA was shown to enhance the pulldown of IRF3 through TBK1, whereas MRP and MRM had the opposite effect, consistent with the hypothesis that MRP and MRM prevent MITA-mediated signaling to TBK1 and downstream pathways (Fig. 3D). It has been reported that SeV infection induces the formation of the VISA complex (19). The participation of MRP in the VISA complex was shown by a coimmunoprecipitation experiment. Endogenous VISA and other components, including MRP, could be precipitated by Ab to VISA after SeV infection in 293T cells (Fig. 3E). Taken together, these results indicated that MRP could interact with VISA complex but interfered with the activation of the TBK1/IRF3 pathway by MITA.

MITA contains four transmembrane regions in the N terminus and resides in the endoplasmic reticulum (ER), according to previous reports (16, 18, 19). Stimulation with dsDNA induces the relocalization of MITA to the perinuclear punctuate structures; this relocalization is related to the activation of TBK1 (29). The subcellular localization of MRP was analyzed. ER-Tracker and MitoTracker were used for live cell imaging. For this purpose, different plasmids expressing MITA and MRP fused with EGFP at the amino- or C-terminal ends were constructed. The fusion of EGFP to the amino- or C-terminal ends of MITA or MRP did not change their functions related to the activation of the IFN-β promoter or ISRE- and IRF3-dependent reporter gene expression (data not shown). All MITA and MRP fusions with EGFP were located in the ER but not in the mitochondria (Fig. 4A and data not shown). MITA and MRP fused with small tags, such as HA and Flag, had the same subcellular localization as when they were fused with EGFP (data not shown). Furthermore, MITA and MRP were found to be colocalized in the cytoplasm after cotransfection of pHA-MITA and pFlag-MRP (Fig. 4B). However, a significant clustering of MITA but not MRP was found around the nucleus. Two different distribution patterns were observed in MITA and MRP coexpressing cells: colocalization of MITA and MRP with a high Pearson coefficient of 0.98 (Fig. 4B, upper panel) or a slightly lower Pearson coefficient of 0.82 (Fig. 4B, lower panel). The lower rate of colocalization of MITA and MRP in the latter case resulted from perinuclear clustering of MITA but not MRP and was found in ∼70% cells with MITA and MRP coexpression. SeV infection increased the fraction of cells with perinuclear clustering of MITA without MRP to 90% (data not shown). The subcellular localization of MITA and MRP relative to TBK1 was studied further. MITA and TBK1 were found to be colocalized and formed dot-like structures around the perinuclear region when TBK1 and MITA were coexpressed (Fig. 4C, upper panel). In contrast, MRP coexpression did not lead to this type of TBK1 distribution in cells (Fig. 4C, lower panel). There is evidence from the published literature that MITA localization is altered upon DNA signaling. In this study, we analyzed the influence of MRP on MITA localization upon HSV-1 infection. Results shown in Fig. 4D indicated that the silencing of endogenous MRP expression by specific siRNA did not change the translocation of MITA upon HSV-1 stimulation.

Because MRP inhibits the MITA-mediated activation of IRF3 but activates the NF-κB pathway, we investigated whether MRP expression plays a role in pathogen infection. SeV and HSV-1 were used as representative RNA and DNA viruses. Additionally, c-diGMP was used to mimic bacterial infection.

The expression patterns of MITA and MRP during viral infection were examined. The expression level of MITA did not significantly change during SeV infection (Fig. 5A), consistent with a previous report (19), but it increased after HSV-1 stimulation (Fig. 5B and data not shown). The expression of MRP increased transiently from 6 to 18 h and then decreased at 24 h after HSV-1 infection (Fig. 5B). In contrast, a gradual reduction of MRP expression was observed during SeV infection (Fig. 5A). This indicated that MRP may play different roles during SeV and HSV-1 infection.

In the reporter assay, SeV infection induced the activity of the IFN-β promoter. Whereas MITA overexpression enhanced SeV-induced IFN-β promoter activity in a dose-dependent manner, MRP overexpression had no effect on the induction of the IFN-β promoter by SeV infection (Fig. 5C). Additionally, cotransfection of pMRP blocked the effect of MITA (Fig. 5D). Silencing endogenous MRP by specific siRNA resulted in a significant elevation of SeV-induced IFN-β promoter activity (Fig. 5E and data not shown). Moreover, silencing of MRP inhibited VSV replication in 293T cells (Fig. 5F). Thus, MRP expression negatively regulated the induction of IFN signaling pathways during SeV infection, explaining the parallel decrease of MRP in SeV infection.

HSV-1 infection upregulated the expression of the tested reporter genes (Fig. 5G). Surprisingly, both MITA and MRP overexpression enhanced HSV-1–induced expression of the reporter genes, although the stimulation by MRP was slightly weaker than that by MITA (Fig. 5G). Silencing of MRP reduced the HSV-1–induced IFN-β promoter activity (Fig. 5H and data not shown). Thus, the transient increase of MRP expression after HSV-1 infection may contribute to the induction of IFN responses, unlike after SeV infection. It is not clear whether the different actions of MRP were due to the different mechanisms of activation of IFN signaling in SeV and HSV-1 infection.

Lastly, c-diGMP was described as an agonist of MITA (41) and has been shown to bind to its carboxyl domain. As shown in Fig. 5I, c-diGMP transfection alone only slightly stimulated IFN-β promoter activity in 239T cells. MITA overexpression mediated the stimulation of IFN-β promoter activity by c-diGMP. Owing to the lack of the carboxyl domain required for c-diGMP binding in MRP, MRP overexpression did not enhance the stimulation of the IFN-β promoter by c-diGMP, and it blocked the MITA-mediated c-diGMP stimulation of the IFN-β promoter (Fig. 5J). Similarly, the silencing of endogenous MRP by siMRP significantly improved the c-diGMP–induced IFN-β promoter activity (Fig. 5K and data not shown). Again, MRP acted as a negative regulator for MITA-induced activation of the IFN signaling pathway; this activation mimics bacterial infection.

A number of splice variants of proteins involved in the innate immune signaling pathways, such as the splice variant of MyD88 (42), the splice variants of IRAK2 (43), TAG (splice variant of TRAM) (44), the splice variant of TBK1 (45), and IRF3-CL (an isoform of IRF3) (46), have been reported. In this study, we identified MRP, a splice variant of MITA, and analyzed its functions. MRP was expressed in various tissues and cell lines. The coimmunoprecipitation experiments demonstrated that MRP is not able to interact with TBK1 and thus does not activate IRF3. However, it does activate NF-κB–dependent gene expression. MRP overexpression interfered specifically with MITA-mediated activation of IRF3 by disrupting the MITA–TBK1 interaction.

Previous reports have suggested that MITA is constitutively expressed and unchanged during virus infection, which is appropriate for its role as an adaptor. In the present study, we found that MRP expression was increased in HSV-1 infection and reduced in SeV infection. Experimentally, we measured higher MRP expression in HepG2.2.15 with hepatitis B virus replication than in parental HepG2 cells (data not shown). Transfection with poly(deoxyadenylic-deoxythymidylic) acid also increased MRP expression in HepG2 cells (data not shown). These results suggest that MRP expression is regulated through different mechanisms upon infection with DNA and RNA viruses, and they may also imply that MRP functions differently in response to different viral infections.

Full-length MITA is composed of 379 aa, and its N-terminal portion, aa 1–138, has been shown to be the transmembrane domain. The crystal structure of the C-terminal domain of MITA has been extensively examined by several groups (38, 40). The segments comprising aa 153–173 and aa 155–180 were shown to be cytosolic and to participate in the dimerization of MITA via hydrophobic interactions. The structure/function analysis suggests that an extra TBK1 binding site other than Ser³⁵⁸ exists within the region aa 139–344 (38). The first 253-aa portion of MRP is identical to MITA, harboring the dimerization domain but lacking the TBK1-binding domain. Our experimental results demonstrated that MRP could form homo- and heterodimers with itself and with MITA, respectively, but did not interact with TBK1; these results are fully consistent with the structural data. Additionally, a previous study has shown that there is no interaction between TBK1 and a truncated MITA mutant with only aa 1–240 (19). The phosphorylation of MITA by TBK1 at Ser³⁵⁸ is important for its function in the activation of IRF3. The phosphorylation site at Ser³⁵⁸ is missing in MRP. As a result, MRP could bind to IRF3; however, it does not contribute to IRF3 phosphorylation by recruiting TBK1. Additionally, MRP retained the ability to activate NF-κB–dependent gene expression, concordant with the result that MRP was able to bind to IKKi. This observation implies that the domain interacting with IKKi is different from the one that interacts with TBK1, which is located within the first 253 aa of MITA. Further investigation is needed to explore the mechanism of MITA- and MRP-mediated activation of NF-κB–mediated gene expression.

The translocation and binding of MITA to TBK1 have been shown to be required for the dsDNA-induced innate immune response. The addition of an ER retention signal to MITA decreases its ability to induce the IFN response (16). In our research, we found different subcellular distribution patterns of MITA and MRP. Although both MITA and MRP were located in the ER, perinuclear clustering was only observed for MITA when both were overexpressed via transfection with plasmids (Fig. 4). The ER retention sequence RIR at aa 178–180 has been experimentally proven to be an ER retention/retrieval signal (18). There is one additional putative ER retention sequence: RKR at aa 276–278 in MRP. The additional RKR sequence in MRP may result in the different subcellular location of MRP.

It has been demonstrated that MITA is a common signaling molecule in the innate immune signaling pathway induced by different factors but it functions in different mechanisms (16, 29, 41). MITA functions as a direct sensor of cyclic dinucleotides, such as c-diGMP, and the dimerization of MITA is essential for dinucleotide binding to MITA and activation of the pathway (38, 40, 41). MRP plays a role as a negative regulator in c-diGMP–induced IFN production owing to the lack of the c-diGMP binding domain and the formation of heterodimers of MITA and MRP. More importantly, we provide novel evidence to support the hypothesis that cyclic dinucleotide sensing and DNA sensing pathways are discrete, even though MITA is involved in both pathways (41), as MRP enhanced the HSV-1–induced IFN response. Translocation of MITA from the ER to punctate structures is specifically driven by dsDNA but not by dsRNA, implying that the singling pathways induced by DNA and RNA viruses are different (28). Knocking down the expression of endogenous MRP did not change the translocation of MITA during HSV-1 infection, consistent with the fact that MRP plays a role in HSV-1–stimulated IFN production. SeV infection induces the formation of the VISA complex including MITA (43). MRP could also incorporate into the VISA complex and interfere with the interaction between MITA and TBK1 after SeV infection, and then inhibit IFN production. The detailed mechanisms of MRP-mediated activation of IFN signaling in HSV-1 infection remain to be investigated.

We are grateful to Xue Hu and Yuan Zhou for assistance with cell culture and to Anna Du for confocal microscopy technical support.

The authors have no financial conflicts of interest.