The Capsid Protein of Nervous Necrosis Virus Antagonizes Host Type I IFN Production by a Dual Strategy to Negatively Regulate Retinoic Acid–Inducible Gene-I–like Receptor Pathways

As an important part of the innate immune system, the type I IFN system plays a pivotal role in the defense of virus infection (1). Host cells sense virus invasion through pattern recognition receptors, such as retinoic acid–inducible gene-I–like receptors (RLRs) and TLRs, and then activate IFN response through a series of signal transduction (2). RLRs, a class of cytoplasmic pattern recognition receptors, are vital for sensing RNA virus infection to induce antiviral innate immunity response. RLRs are activated by recognizing virus nucleic acid and then bind to a downstream adaptor molecule, the mitochondrial antiviral signaling protein (MAVS), which recruits and activates TNFR-associated factor 3 (TRAF3) and TANK-binding kinase 1 (TBK1). Subsequently, TBK1 phosphorylates IFN regulatory factor 3 (IRF3) to form homodimer and translocate into the nucleus for IFN transcription, thereby promoting the expression of IFN-stimulated genes (ISGs) to exert antiviral function (3).

In response to the host antiviral response, viruses have evolved various strategies to evade or suppress these signaling cascades for effective infection. It has been well documented that many viral proteins can directly or indirectly target distinct steps in the RLR signaling pathway to antagonize type I IFN response by different mechanisms, such as ubiquitylation and phosphorylation. For example, West Nile virus nonstructural protein antagonizes the expression of IFN-β by inhibiting the K63-linked polyubiquitination of retinoic acid–inducible gene-I (4). Newcastle disease virus V protein inhibits MAVS-mediated IFN response via E3 ubiquitin ligase RNF5 (5). Severe acute respiratory syndrome (SARS) coronavirus 2 nsp12 attenuates type I IFN production by inhibiting IRF3 nuclear translocation (6). Thus, understanding the molecular mechanisms of virus escaping or inhibiting the RLR-induced IFN response will provide insight for the control of virus infection.

Nervous necrosis virus (NNV), an important pathogen in fish, is a positive single-stranded RNA virus belonging to genus Betanodavirus within the Nodaviridae family. The NNV genome encodes a nonstructural protein RNA-dependent RNA polymerase (RDRP) and a structural protein capsid protein (CP) (7). RDRP is mainly responsible for replication of the viral genome and can promote viral replication on mitochondria by interacting with endoplasmic reticulum chaperone protein GRP78 (8). The latest report indicated that RDRP from orange-spotted grouper NNV could trigger type I IFN response in fish cells (9). CP is the basic unit that makes up the NNV capsid. It is also reported that CP of Greasy grouper (Epinephelus tauvina) NNV functions as an apoptosis inducer that induces the caspase-dependent apoptosis (10). B1 and B2 are two other nonstructural proteins encoded by a subgenome produced during NNV replication. They play opposite roles in the necrosis of the host cells to keep a balance, which allows the virus to make the most of the host cells for its replication and spread (11, 12). In addition, B2 protein enhanced viral replication by binding with and protecting the replicating intermediate dsRNA (13, 14). Until now, NNV has posed a serious threat to >170 marine and freshwater fish species, and its mortality rate can reach up to 100% in larvae and juveniles (15). Increasing evidence has shown that latent infection of NNV is very common in fish (16, 17), which is closely related to its powerful function of escaping from the host's antiviral immunity. However, the molecular mechanisms of NNV to evade or suppress the type I IFN response, specifically whether and how NNV encoding proteins participate in the modulation of the type I IFN response, are still largely unknown.

Sea perch (Lateolabrax japonicus) is an aquaculturally important fish suffering from NNV infection. In this study, we investigated the role of NNV CP in regulating the L. japonicus innate immune response in cell culture. We found that CP functioned as an IFN antagonist that interfered with the RLR-IFN signaling by targeting L. japonicus TRAF3 (LjTRAF3) and L. japonicus IRF3 (LjIRF3) for ubiquitination and degradation. Our findings suggest a novel strategy by which NNV subverts cellular innate immunity and evades host antiviral responses.

Fathead minnow (FHM) cells were grown at 28°C in M199 medium supplemented with 10% FBS (Life Technologies, Carlsbad, CA) (18). L. japonicus brain (LJB) cells were cultured in DMEM containing 10% FBS at 28°C (19). HEK293T cells were maintained in DMEM with 10% FBS at 37°C in 5% CO₂. Red-spotted NNV (RGNNV) was isolated from L. japonicus and propagated in LJB cells (19).

Anti-Flag (catalog number [Cat. no.] M20008), anti-GFP (Cat. no. P30010), anti-Myc (Cat. no. M20002), anti-HA (Cat. no. M20003), anti-His (Cat. no. M20001L), and anti-Actin (Cat. no. P30002) Abs were purchased from Abmart (Guagnzhou, China). Anti-ubiquitin (Ub) Ab (Cat. no. P4D1) was purchased from Cell Signaling Technology (Danvers, MA). Anti-IRF3 (Cat. no. bs-2993R), anti–p-IRF3 (Ser386) (Cat. no. bs-52170R), and anti-TRAF3 (Cat. no. bs-1185R) Abs were purchased from Bioss (Beijing, China). Anti-Lamin B1 Ab (Cat. no. CPA1693) was purchased from Cohesion Biosciences. Anti–L. japonicus ring finger protein 114 (LjRNF114) Ab was purchased from Kebiochem (Shanghai, China). Anti-CP Ab was produced by Kebiochem (Shanghai, China) (20). HRP-conjugated goat anti-rabbit IgG (Cat. no. A0208) and HRP-conjugated goat anti-mouse IgG (Cat. no. A0216) were purchased from Beyotime (Shanghai, China). Alexa Fluor 488–labeled donkey anti-rabbit IgG (Cat. no. R37118) and Alexa Fluor 555–labeled goat anti-mouse IgG (Cat. no. A-21422) secondary Abs were purchased from Invitrogen Corporation (Carlsbad, CA). Polyinosinic:polycytidylic acid [Poly(I:C)], MG132 (Cat. no. M7449), and NH₄Cl were purchased from Sigma (St. Louis, MO).

LjTRAF3 and LjIRF3 truncations with Flag-tag, including pCMV-Flag-LjTRAF3-RING, pCMV-Flag-LjTRAF3-ZF1, pCMV-Flag-LjTRAF3-ZF2, pCMV-Flag-LjTRAF3-CC, pCMV-Flag-LjTRAF3-MATH, pCMV-Flag-LjIRF3-DBD, pCMV-Flag-LjIRF3-TAD, and pCMV-Flag-LjIRF3-RD were constructed using standard molecular biology techniques. All plasmids were confirmed by DNA sequencing. Primers used for plasmid construction are listed in Supplemental Table I.

pGL3-LjIFNh-pro-Luc, pET32a-His-CP, pCMV-Myc-CP, pCMV-Flag-CP, pCMV-Flag-LjMAVS, pCMV-Flag-LjTRAF3, pCMV-Flag-LjTBK1, pCMV-Flag-LjIRF3, pCMV-Flag-LjRNF114, pEGFP-LjRNF114, and truncated mutants of CP with Myc-tag were prepared in our previous studies (1, 3, 21).

HA-Ub, HA-K27, HA-K48, and HA-K63 plasmids were purchased from YouBia (Guangzhou, China).

LJB cells were seeded into six-well plates (1 × 10⁶ cells/well) and cultured at 28°C for 24 h, followed by transfection with pCMV-Flag-CP or empty vector plasmids using Lipofectamine 8000 (Beyotime, Shanghai, China) according to the manufacturer’s instructions. At 24 h posttransfection, the cells were infected with RGNNV at a multiplicity of infection (MOI) of 2 for 12 and 24 h, respectively. Then cells were subjected to quantitative real-time PCR for measuring viral RNA copies and IFNh and ISGs transcripts. Primers are listed in Supplemental Table I. Cell supernatants were collected for virus titer assay as described previously with some modifications (22). The supernatants were 10-fold serially diluted and added to FHM cells seeded in 96-well plates (1 × 10⁴ cells/well). After 5 d of incubation at 28°C, viral titer was determined by the 50% tissue culture infective dose assay. For plaque assay, the supernatants at 24 h postinfection (hpi) were diluted 10 times and added (100 μl) onto a monolayer of FHM cells cultured in 24-well plates for 48 h. The supernatants were removed, and the monolayers were washed with PBS, followed by fixation with 4% paraformaldehyde (PFA) for 10 min, and stained with 1% crystal violet for 10 min.

Total RNAs were extracted using the TRIzol reagent (Invitrogen, Carlsbad, CA) and reverse transcribed into cDNAs using the PrimeScript First Strand cDNA Synthesis Kit (Takara, Dalian, China) according to the manufacturer’s instructions. qRT-PCR was performed in a LightCycler 480 II with the procedure of 95°C for 30 s, 40 cycles of 95°C for 10 s, 60°C for 10 s, and 72°C for 10 s. The sea perch β-actin was used as the reference gene. Primers are listed in Supplemental Table I. The 2^−ΔΔCt method was used to analyze the relative expression of target genes. Each sample was tested in triplicate.

Dual-luciferase reporter assay was performed as described previously with some modifications (23). FHM cells plated in 24-well plates were cotransfected with pGL3-LjIFNh-pro-Luc, pRL-TK, and indicated plasmids or empty vector, followed by poly(I:C) (5 μg/ml) stimulation or not. The luciferase activity was determined using the dual-luciferase reporter assay system with a GloMax 20/20 Luminometer (Promega, Madison, WI) according to the manufacturer’s instructions. The data were determined by normalization of the firefly luciferase activities to the renilla luciferase activities and were expressed as mean ± SD from three independent experiments performed in triplicates.

For cellular colocalization of CP with LjTRAF3, LjIRF3, or LjRNF114, FHM cells were cotransfected with pCMV-Myc-CP and pCMV-Flag-LjTRAF3, pCMV-Flag-LjIRF3, or pCMV-Flag-LjRNF114 for 24 h and were subjected to immunofluorescence assay as described previously (3). Cells were washed with PBS, fixed with 4% PFA for 10 min, permeabilized with 0.1% Triton X-100 for 10 min, and blocked with 5% BSA at room temperature (RT) for 1 h. Then cells were incubated with anti-Flag (1:200) and anti-Myc Abs (1:200) at 4°C overnight, followed by washing with PBS, incubating with secondary Abs for 1 h at RT, and staining with Hoechst 33342. Fluorescence images were obtained and analyzed using a laser scanning confocal microscope (LSM510; Zeiss, Jena, Germany).

The small interfering RNA (siRNA) targeting LjRNF114 was synthesized as described in our previous study (3). LJB cells were cotransfected with pCMV-Flag-CP and LjRNF114 siRNA mixture (100 nM) or negative control (NC) siRNA (100 nM), followed by RGNNV (MOI = 2) infection at 24 h posttransfection. Subsequently, cells were harvested at 24 hpi and subjected to qRT-PCR and Western blotting assays, respectively.

Coimmunoprecipitation (co-IP) assays were carried out as described previously (21). HEK293T cells were seeded into 25-cm² flasks and transfected with a different combination of indicated plasmids. At 24 h posttransfection, cells were collected and lysed in 500 μl of immunoprecipitation lysis buffer (Beyotime, Shanghai, China) via sonication on ice. After centrifugation for 10 min at 12,000 × g at 4°C, the supernatants were incubated with specific Abs and protein A/G magnetic beads overnight at 4°C. Then the beads were washed with immunoprecipitation lysis buffer for five times and resuspended in SDS loading buffer for Western blot assays.

His fusion protein expression and pull-down assays were performed as described previously with some modifications (3). Escherichia coli BL21 (DE3) transfected with pET32a-His-CP was cultured in 15 ml of Luria‒Bertani medium containing 0.5 mM isopropyl-1-thio-β-d-galactopyranoside at 18°C overnight with agitation. Then bacteria were harvested by centrifugation, resuspended in lysis buffer (100 mM sodium-phosphate [pH 8.0], 600 mM NaCl, 0.02% Tween 20), and crushed by ultrasonic treatment. After centrifugation for 10 min at 12,000 × g at 4°C, the supernatants were incubated with Dynabead His-Tag magnetic beads (Invitrogen) for 4 h at RT. The Ag-Ab-magnetic bead immune complex was captured by the magnetic force of a magnet and then was used to bind Flag-tagged proteins from the lysates of HEK293T cells transfected with indicated Flag-tagged plasmids, respectively. After incubation overnight at 4°C, the beads were washed and used for Western blotting assays.

For ubiquitination assay, HEK293T cells were cotransfected with indicated plasmids for 24 h, followed by MG132 (10 μM) treatment. Cells were harvested at 24 h posttransfection, lysed, and immunoprecipitated with the indicated Abs as described earlier. The whole-cell lysates or immunoprecipitates were then analyzed by Western blotting assays.

Western blotting assays were carried out as described previously (21). Cell lysates or immunoprecipitates were subjected to SDS-PAGE analysis using 10% SDS-PAGE gels. Resolved proteins were transferred to a polyvinylidene difluoride membrane (Millipore, Billerica, MA). The membranes were blocked with 10% skim milk for 1 h at RT and then probed with indicated Abs overnight at 4°C. After washing with 0.5% PBST three times, the membranes were incubated with secondary Abs for 1 h at RT. After washing three times, protein bands were detected by a chemiluminescence instrument.

Data were presented as the mean ± SD. Data were analyzed using either the Student t test for two-group comparisons or a one-way ANOVA for multiple-group comparisons. The p values <0.01 and <0.05 were considered statistically significant.

To determine the effect of CP on RGNNV infection, we infected LJB cells transfected with plasmids expressing CP protein with RGNNV. As shown in (Fig. 1A, LJB cells transfected with CP were more vulnerable to RGNNV infection. The viral gene RDRP expression and virus titration were significantly increased in LJB cells overexpressing CP compared with those in control cells (Fig. 1B, 1C). Furthermore, our results showed that the mRNA transcripts of IFNh, Mx, ISG15, and Viperin were significantly downregulated in CP overexpressed LJB cells (Fig. 1D–G). Overexpression of CP suppressed IFNh promoter activation in FHM cells (Fig. 1H). Meanwhile, CP had a dose-dependent inhibitory effect on poly(I:C)-induced IFNh promoter activation (Fig. 1I). These results suggest that CP may promote RGNNV infection by antagonizing the production of IFN and its downstream ISGs.

To determine the key domain of CP responsible for its IFN antagonistic activity, we monitored IFNh promoter activity in the presence of five truncation mutants of CP (CP-ΔARM, CP-Δarm, CP-ΔS, CP-ΔL, and CP-ΔP) using a luciferase reporter assay. As shown in (Fig. 1J, overexpression of CP-ΔARM, CP-ΔS, and CP-ΔP had no significant inhibitory effect on IFNh promoter compared with wild type CP, indicating the ARM, S, and P domains of CP were vital for the suppression of IFNh production.

Our previous study showed that RGNNV infection could induce IFN production through activating the sea perch RLR signaling pathway (2, 24–26). To investigate whether CP suppresses the IFN response by blocking RLR signaling, we performed co-IP assays to analyze the interaction of CP with the key components of the RLR signaling pathway. pCMV-Myc-CP and Flag-tagged RLR factors, including LjMAVS, LjTRAF3, LjTBK1, and LjIRF3, were cotransfected into HEK293T cells. The results of co-IP assays showed that CP interacted with LjTRAF3 and LjIRF3 (Fig. 2A, 2B), but not LjMAVS and LjTBK1 (Fig. 2C, 2D). Confocal microscopy assays showed that CP protein (red) overlapped with LjTRAF3 and LjIRF3 (green) in the cytoplasm of FHM cells, respectively (Fig. 2E). Furthermore, pull-down assay showed that interaction occurred between CP and LjIRF3 (Fig. 2G), but not LjTRAF3 (Fig. 2F), indicating CP has a direct interaction with LjIRF3, but an indirect association with LjTRAF3. Meanwhile, our data showed that CP interacted with endogenous LjIRF3 in LJB cells infected with RGNNV (Supplemental Fig. 1).

To determine the domains of CP required for its association with LjTRAF3 or LjIRF3, we coexpressed truncated mutants of CP and LjTRAF3 or LjIRF3 in HEK293T cells, then subjected them to co-IP assays. The results showed that the deletion of the S domain of CP abrogated the interaction between CP and LjTRAF3 or LjIRF3 (Fig. 3A, 3B). To dissect domains of LjTRAF3 or LjIRF3 responsible for their interaction with CP, plasmids expressing mutants of LjTRAF3 (Fig. 3C) or LjIRF3 (Fig. 3D) and CP were cotransfected into HEK293T cells, followed by immunoprecipitation. As shown in (Fig 3E, ZF1 and CC domains of LjTRAF3 were responsible for its interactions with CP. The transcriptional activation domain (TAD) and regulatory domain (RD) of LjIRF3, but not the DNA binding domain, interacted with CP (Fig. 3F), indicating both TAD and RD domains of LjIRF3 are sufficient for its interaction with CP.

To investigate the effect of CP on LjTRAF3 and LjIRF3, we detected the expression of LjTRAF3 and LjIRF3 in CP-overexpressing HEK293T cells. As shown in (Fig. 4A and 4B, the significant degradation of LjTRAF3 and LjIRF3 was detected in the presence of CP in a dose-dependent manner. In eukaryotes, cellular proteins were mainly degraded through ubiquitin-proteasome and lysosome-autophagy systems. To further clarify whether either the proteasomes or lysosomes pathway was involved in CP-induced LjTRAF3 and LjIRF3 degradation, HEK293T cells cotransfected with pCMV-Flag/MYC-CP, pCMV-Flag-LjTRAF3, or pCMV-Flag-LjIRF3 plasmids were treated with or without the proteasome inhibitor MG132 or lysosome inhibitor NH₄Cl. In the presence of MG132, CP-induced LjTRAF3 and LjIRF3 degradation was reversed with the increasing concentration of MG132 (Fig. 4C, 4E), but NH₄Cl had no effect on CP-induced LjTRAF3 and LjIRF3 degradation (Fig. 4D, 4E). We further monitored levels of ubiquitination in LJB cells with or without ectopic expression of recombinant CP protein. As expected, the overall ubiquitination level was dramatically increased after ectopic CP protein expression (Fig. 4F). All these results indicate that CP promotes LjTRAF3 and LjIRF3 degradation through the ubiquitin-proteasome pathway.

CP does not, by itself, possess the capacity of degrading the substrates (27). Given that our previous study demonstrated that LjRNF114 interacted with LjTRAF3 and promoted LjTRAF3 ubiquitination and degradation (3), we speculated that LjRNF114 might be involved in CP-induced LjTRAF3 degradation. To test this hypothesis, we investigated the relationship between CP and LjRNF114. As shown in (Fig. 5A and 5B, CP promoted the endogenous expression of LjRNF114 at the mRNA and protein levels. Knockdown of LjRNF114 impaired CP-induced inhibition of IFNh expression (Fig. 5C). To confirm the interaction between CP and LjRNF114, we first performed a co-IP experiment, and the results showed that CP interacted with LjRNF114 (Fig. 5D). Next, the localization of CP and LjRNF114 in cells was analyzed. As shown in (Fig. 5E, the obvious colocalization between LjRNF114 and CP in the cytoplasm of FHM cells was observed. Furthermore, His tag pull-down assays confirmed the direct interaction of CP with LjRNF114 (Fig. 5F). Reciprocal domain mapping experiments were conducted with CP deletion mutants and LjRNF114. As shown in (Fig. 5G, the deletion of the P domain of CP completely abrogated the interaction between CP and LjRNF114, indicating the P domain of CP was responsible for CP–LjRNF114 interaction.

To further confirm that LjRNF114 was responsible for CP-induced LjTRAF3 degradation, we detected LjTRAF3 expression in LJB cells cotransfected with siNC or siLjRNF114 and pCMV-Myc-CP or pCMV-Myc plasmids. We found that ectopic expression of CP reduced the expression of endogenous LjTRAF3 in LJB cells, and cotransfection of siLjRNF114 significantly counteracted CP-induced downregulation of LjTRAF3 (Fig. 6A). Our previous study demonstrated that LjRNF114 promoted LjTRAF3 degradation by facilitating K27- and K48-linked polyubiquitination (3). In this study, we further assessed the effect of CP on LjRNF114-mediated K27- and K48-linked ubiquitination of LjTRAF3. The results showed that CP promoted LjRNF114-mediated K48-linked, but not K27-linked, ubiquitination of LjTRAF3 (Fig. 6B). Overall, these results indicated that CP facilitates the ubiquitination degradation of LjTRAF3 by directly interacting with LjRNF114.

To further confirm that CP mediated the ubiquitination of LjIRF3, we cotransfected HA-Ub and pCMV-Myc-CP or empty vector plasmids into HEK293T cells, followed by MG132 treatment at 24 h posttransfection. Co-IP assays showed that CP promoted the ubiquitination of LjIRF3 (Fig 7A). To illustrate which type of ubiquitin was involved in CP-induced LjIRF3 degradation, we transfected HEK293T cells with K27-, K48-, or K63-linked ubiquitin plasmids in the presence of MG132, respectively. As shown in (Fig. 7B, overexpression of CP promoted the K48-linked, but not K27- and K63-linked, ubiquitination of LjIRF3. IRF3 phosphorylation and its translocation to the nucleus were essential for regulating type I IFN gene expression. Thus, the phosphorylation level and subcellular localization of LjIRF3 were detected in LJB cells on treatment with poly(I:C) via immunoblotting. As expected, ectopic expression of CP reduced the expression of endogenous LjIRF3 and p-LjIRF3 in the nuclear fraction and cytoplasmic fraction in LJB cells on poly(I:C) treatment (Fig. 7C). All together, these data demonstrate that CP promotes K48-linked ubiquitination and degradation of LjIRF3, leading to decreased LjIRF3 phosphorylation and nuclear translocation (Fig. 8).

The IFN system plays an important role in the innate immune response against viruses in fish and mammals (28, 29). However, many viruses have evolved various strategies to evade the host's antiviral innate immune system for their survival. For instance, SARS coronavirus 2 nsp12 inhibits type I IFN production by inhibiting IRF3 nuclear translocation (6). Siniperca chuatsi rhabdovirus negatively regulates virus-triggered type I IFN production by inducing the expression of miR-210, a suppressor of IFN signaling factor STING (30). Grass carp reovirus VP41 targets STING and decreases its phosphorylation to antagonize IFN response (31). NNV is an aquatic virus with strong transmission ability and high mortality, which is at least partly due to its immune evasion or suppression of IFN activation. However, the exact mechanism is not yet known. In this article, we reported that CP of RGNNV suppressed IFN production to promote its replication in the way of inhibiting the RLR signaling pathway by regulating the ubiquitination and degradation of LjTRAF3 and LjIRF3.

To fight against the huge host with a sophisticated immune system, certain viral proteins play multiple functions during the viral life cycle. For example, the nonstructural protein of influenza A viruses regulates not only viral replication but also host innate immune responses (32). The genome of NNV encodes only four proteins, including CP, RDRP, B1, and B2. Among them, the structural protein CP was involved in the induction of apoptosis and autophagy of host cells besides viral particle assembly (10, 33). In this study, we found that CP could antagonize the IFN response, and its ARM, S, and P domains were critical for the anti-IFN activity. A previous study showed that the ARM domain of CP contained a nucleolus localization signal and was rich in basic amino acids (nine arginine and six lysine), which was assumed to participate in binding the RNA genome to the internal capsid wall (27). Besides, the amino acids 247 and 270 in the CP P domain might be involved in the infectivity of the reassorted Betanodavirus (27). Our study extends the biological functions of CP in innate immune regulation.

The RLR signaling pathway is a vital way to induce IFN production. Thus, it is not surprising that viruses target the sensors or regulatory proteins critical for the RLR response to antagonize IFN production. For example, spring viremia of carp virus N and P proteins repress IFN production by degrading MAVS or inhibiting the kinase activity of TBK1, respectively (34, 35). SARS coronavirus orf-9b inhibits the innate immunity through targeting mitochondria and the MAVS/TRAF3/TRAF6 signalosome (36). Our previous studies indicated that RLRs could sense RGNNV infection, then activate the IFN production (2, 24–26, 37). Hence we investigated the physical interaction between CP and the key components of RLR signaling to determine whether CP targets the RLR signaling pathway to inhibit IFN activation. Our results showed that CP indirectly and directly interacted with LjTRAF3 and LjIRF3, respectively. Specifically, the S domain of CP was vital for the interaction between CP and LjTRAF3 or LjIRF3. These results indicate that CP targets LjTRAF3 and LjIRF3 through its S domain to block RLR-induced IFN production. It is well documented that the ubiquitin-proteasome and autophagy-lysosome pathways are mainly routes for some viruses degrading host immune-related proteins (38). For instance, foot-and-mouth disease virus structural protein VP1 promotes ubiquitination and degradation of TPL2 to evade host antiviral immunity (39). Foot-and-mouth disease virus degrades Rab27a via the autophagy-lysosome pathway to suppress the exosome-mediated antiviral immune response (40). In this study, our further investigation showed that protein levels of LjTRAF3 and LjIRF3 were downregulated by CP in a dose-dependent manner and were rescued by MG132, but not NH₄Cl, indicating that CP degraded LjTRAF3 and LjIRF3 by the ubiquitin-proteasome system.

We also found that ectopic expression of CP significantly upregulated the overall ubiquitination level in LJB cells, which further confirmed that CP participated in the regulation of protein ubiquitination. CP is not an E3 ubiquitin ligase; therefore, the additional E3 ubiquitin ligase should be involved in the process. Our earlier study demonstrated that E3 ubiquitin ligase LjRNF114 negatively regulated IFN response via promoting K27- and K48-linked ubiquitination and degradation of LjTRAF3 (3). Thus, we presumed that CP might induce the ubiquitination and degradation of LjTRAF3 by the aid of LjRNF114. To confirm this, we investigated the relationship between CP and LjRNF114. We found that overexpression of CP significantly increased LjRNF114 expression at both transcript and protein levels. Meanwhile, CP was confirmed to directly interact with LjRNF114, and the P domain of CP was necessary for its interaction with LjRNF114. Considering that LjTRAF3 binds to the S domain of CP, we speculated that CP, LjRNF114, and LjTRAF3 form a complex. Furthermore, we found that CP promoted the K48-linked ubiquitination and proteasomal degradation of LjTRAF3 in a LjRNF114-dependent manner. Taken together, CP recruits LjRNF114 to induce the ubiquitination and proteasomal degradation of LjTRAF3.

IRF3 is the key factor in the RLR signaling pathway for IFN production. IRF3 phosphorylation and nuclear translocation are critical for IFN transcription. Several studies have demonstrated that certain viruses suppress IRF3 activation by dephosphorylation and polyubiquitination to attenuate IFN production. For example, the membrane protein of middle east respiratory syndrome coronavirus suppresses the type I IFN response via blocking IRF3 phosphorylation (41). UL13 protein of pseudorabies virus promotes IRF3 ubiquitination and proteasome-mediated degradation to inhibit the IFN signaling pathway (42). Our results in this study demonstrated that CP promoted LjIRF3 K48 ubiquitination degradation, which attenuated IRF3 activation by decreasing LjIRF3 phosphorylation and nuclear translocation. It was well known that TBK1 is responsible for IRF3 phosphorylation. Although our results showed that CP did not interact with TBK1, we could not entirely exclude the possibility that CP might use other factors to affect the roles of TBK1, which in turn influenced IRF3 phosphorylation. Thus, the degradation of IRF3 might occur before or after its phosphorylation. Previous studies demonstrated that several E3 ubiquitin ligases were recruited by viral proteins to degrade the substrate. For example, newcastle disease virus V protein recruits RNF5 to polyubiquitinate and degrade MAVS (5). Due to CP not being an E3 ubiquitin ligase, we speculated that additional E3 ubiquitin ligase might participate in CP-induced LjIRF3 ubiquitination degradation, but it is regrettable that the E3 ubiquitin ligase was not identified in this study. Similar with our results, certain virus proteins, without E3 ubiquitin ligase activity, recruit unsuspected E3 ubiquitin ligase to regulate type I IFN production. For instance, SVCV N and Zika virus NS3 facilitated ubiquitination and proteasome-dependent degradation of the key factors of RLR signaling, although they were not described as E3 ubiquitin ligases yet (34, 43). We also found that knockdown of LjRNF114 had no influence on CP-induced LjIRF3 degradation (Supplemental Fig. 2), indicating that LjRNF114 was not involved in CP-induced LjIRF3 degradation. Increasing evidence has shown that some E3 ubiquitin ligases mediated ubiquitination of IRF3. For instance, E3 ligase MID1 mediated ubiquitination and degradation of IRF3 (44). E3 ubiquitin ligase Casitas B-lineage lymphoma interacted with IRF3 and promoted its K48-linked polyubiquitination and degradation (45). Further study will be needed to verify whether these proteins are involved in CP-induced LjIRF3 polyubiquitination and degradation.

In conclusion, we showed that CP attenuated type I IFN activation by a dual strategy to target two components of the RLR signaling pathway, LjTRAF3 and LjIRF3. Mechanistically, CP potentiates the ubiquitination and proteasomal degradation of LjTRAF3 via directly interacting with LjRNF114. CP interacted with LjIRF3 and promoted its K48 ubiquitination degradation, leading to decreased LjIRF3 phosphorylation and nuclear translocation (Fig. 8). Our study delineates a novel mechanism for NNV evading type I IFN antiviral response.

We thank Dr. Lingbing Zeng (Yangtze River Fisheries Research Institute, Chinese Academy of Fishery Sciences, Wuhan, China) for providing FHM cells.

The authors have no financial conflicts of interest.