MicroRNA-200c-targeted contactin 1 facilitates the replication of influenza A virus by accelerating the degradation of MAVS

Influenza A viruses (IAVs) continuously challenge the poultry industry and human health. Elucidation of the host factors that modulate the IAV lifecycle is vital for developing antiviral drugs and vaccines. In this study, we infected A549 cells with IAVs and found that host protein contactin-1 (CNTN1), a member of the immunoglobulin superfamily, enhanced viral replication. Bioinformatic prediction and experimental validation indicated that the expression of CNTN1 was reduced by microRNA-200c (miR-200c) through directly targeting. We further showed that CNTN1-modulated viral replication in A549 cells is dependent on type I interferon signaling. Co-immunoprecipitation experiments revealed that CNTN1 specifically interacts with MAVS and promotes its proteasomal degradation by removing its K63-linked ubiquitination. Moreover, we discovered that the deubiquitinase USP25 is recruited by CNTN1 to catalyze the deubiquitination of K63-linked MAVS. Consequently, the CNTN1-induced degradation cascade of MAVS blocked RIG-I-MAVS-mediated interferon signaling, leading to enhanced viral replication. Taken together, our data reveal novel roles of CNTN1 in the type I interferon pathway and regulatory mechanism of IAV replication.

Introduction Influenza A viruses (IAVs) are segmented, single-stranded, negative-sense RNA viruses. They are divided into 18 HA and 11 NA subtypes based on the antigenicity of their HA and NA surface glycoproteins. IAVs continuously challenge the poultry industry and human health due to antigenic shift and drift. In the 20th century, H1N1, H2N2, and H3N2 viruses caused four influenza pandemics in humans, resulting in widespread disease and severe loss of human life [1,2]. Since 2003, H5N1 IAVs have infected more than 800 individuals across 16 countries, with an overall case fatality rate of 53% [3,4]. Other H5Nx viruses, such as H5N6 and H5N8, have been detected in wild birds and domestic poultry in many countries, and caused human infections and even deaths [5][6][7]. The H7N9 low pathogenic avian influenza viruses identified in 2013, and the subsequently mutated H7N9 highly pathogenic avian influenza viruses in 2017, led to 1568 human infections, including 615 fatal cases [8,9]. Therefore, studies on the factors and mechanisms underlying the host-virus interaction remain essential for developing novel anti-influenza therapies.
Innate immunity is the first line of host defense. Interferons, as vital part of the innate immune system, play important roles against invading pathogens during the early stages of infection. Upon IAV infection, the host pattern recognition receptors (PRRs) sense the invading viral RNA and initiate a cascade of signaling responses [10]. Influenza viral RNA is generally sensed by retinoic acid-inducible gene-I (RIG-I) and melanoma differentiation-associated gene 5 (MDA5). RIG-I undergoes conformational changes and translocation to the mitochondrial membrane, where it interacts with the mitochondrially localized adaptor protein MAVS (also known as VISA, IPS-1, and Cardif), thereby recruiting downstream effector molecules and activating IRF3, IRF7, and nuclear factor κB (NF-κB) to produce type I interferons (IFN-I), ultimately creating an antiviral state [11,12].
In this study, using transcriptome deep-sequencing and data analysis, we discovered that host factor contactin-1 (CNTN1) expression was reduced upon infection with various subtypes of IAV. Further, we found that CNTN1 facilitates the replication of IAV through negative regulation of the RIG-I-mediated IFN-I pathway. Mechanistically, CNTN1 recruits USP25 to deubiquitinate K63-linked MAVS, which promotes the proteasomal degradation of MAVS. Our data reveal a novel mechanism by which CNTN1 negatively regulates RIG-I-mediated IFN-I pathway and potentiates IAV replication.

CNTN1 is downregulated by IAV infection and promotes the replication of IAVs in A549 cells
To screen for host factors involved in the host-virus interaction, we profiled the mRNA expression in A549 cells infected with or without H5N6 virus. We identified groups of genes that were upregulated or downregulated (Log 2 Fold-change >3, P-value <0.05) after H5N6 virus infection (S1 Table). We noticed that one of the IAV-downregulated genes, CNTN1, has not previously been linked to IAVs. We confirmed that CNTN1 was dramatically downregulated by H5N6 virus infection (Fig 1A and 1B). To investigate the effect of virus-downregulated CNTN1 on IAV replication, we used CRISPR/Cas9 to knockout the endogenous CNTN1 in A549 cells (S1A Fig). Viral growth kinetics assays indicated that overexpressed CNTN1 significantly facilitated H5N6 replication, whereas CNTN1-deficiency significantly inhibited viral replication (Fig 1C and 1D). Furthermore, CNTN1 levels were observably decreased in A549 cells following infection with H1N1 or H7N9 viruses

IAV-induced miR-200c negatively regulates the expression of CNTN1
Within the sequencing data, we found that several miRNAs were significantly up-or downregulated following H5N6 virus infection (S1 Table). We therefore asked whether there was an endogenous miRNA that can specifically suppresses the expression of CNTN1 and promote the replication of H5N6 virus. To screen the miRNAs that target CNTN1, we used MicroRNA. org [13], TargetScan [14] and StarBase [15] software to predict the miRNA binding sites on the 3'-untranslated region (3'-UTR) of human CNTN1. Based on the predicted results, we screened out miR-200c, which appeared to target the 3'-UTR of CNTN1 mRNA. Of note, miR-200c expression in A549 cells was increased following H5N6 virus infection (Fig 2A). To verify the regulatory effect of miR-200c on CNTN1, miR-200c mimic and inhibitor were used to simulate the expression of miR-200c. As shown in S1B Fig, transfection of the miR-200c mimic and inhibitor had no significant impact on A549 cell viability. However, treatment with the miR-200c mimic markedly reduced CNTN1 expression (Fig 2B). Quantitative PCR (qPCR) revealed that the amount of CNTN1 mRNA was significantly reduced following transfection of the cells with the miR-200c mimic ( Fig 2C).
To confirm that the mRNA of CNTN1 directly associated with miR-200c, we first performed an RNA immunoprecipitation (RIP) assay using Flag-Ago2 (a component of the miRNA-silencing complex), and found that the mRNA of CNTN1 was considerably enriched within the miR-200c group ( Fig 2D). In addition, we cloned the region of the CNTN1 3'-UTR that contained the predicted target site of miR-200c into the dual-luciferase reporter vector pmirGLO; we also constructed the 3'-UTR region with mutations in the binding site and cloned it into the pmirGLO vector to obtain a mutant plasmid (Fig 2E). HEK293 cells were transfected the wild-type CNTN1 3'-UTR plasmid or mutant plasmid in the presence of the miR-200c mimic or miRNA control for 24 h. The miR-200c mimic significantly inhibited the luciferase activity of the wild-type CNTN1 3'-UTR plasmid-transfected cells, whereas transfection with the mutant CNTN1 3'-UTR abolished the inhibitory effect of the miR-200c mimic on the luciferase activity (Fig 2E), suggesting that miR-200c directly interacts with CNTN1 3'-UTR.
We next addressed the effect of miR-200c on the replication of H5N6 virus. At 24 h after miR-200c mimic transfection, A549 cells were infected with H5N6 virus (MOI = 3 or 0.01). The viral titers of the cell supernatants were measured by TCID 50 assay, and the results showed that the miR-200c mimic significantly inhibited viral replication ( Fig 2F). We also investigated the effect of the miR-200c inhibitor on IAV replication, and found that treatment with the miR-200c inhibitor led to a marked increase in titers ( Fig 2G). Additionally, using CRISPR/ Cas9 to knockout the endogenous miR-200c in A549 cells, we found that miR-200c-deficiency significantly promoted viral replication, while miR-200c mimic restored the inhibitory effect (Figs S1C and 2H). Furthermore, the inhibitory effect of the miR-200c mimic on viral replication was restored by CNTN1 (Figs 2I and S1D).
Together, our findings indicate that virus-induced miR-200c downregulates CNTN1 expression through direct interaction, thereby inhibiting H5N6 virus replication.

CNTN1 inhibits type I interferon signaling pathway
Upon IAV infection, the RIG-I-mediated IFN-I signaling pathway is the main innate immune response [16]. To investigate the mechanism of CNTN1-promoted IAV replication, we first tested whether CNTN1 affects the activity of the IFN-I signaling pathway. As shown in Fig 3A, overexpressed CNTN1 strongly inhibited Sendai virus (SeV)-triggered activation of the IFN-β promoter and interferon sequence response element (ISRE). When we used siRNA to knock down CNTN1, the cell viability was not markedly influenced and the activation of the IFN-β The level of CNTN1 mRNA was quantified by using a qPCR assay. (E) Predicted miR-200c binding sites in the 3'-UTR of the CNTN1 mRNA. Perfect matches in seed regions are indicated by lines. Mutations (indicated by arrows) were generated in the binding sites of the 3'-UTR for the reporter gene assay. HEK293 cells were transfected with the wild-type or mutant of the CNTN1 3'-UTR reporter plasmid and the miR-200c mimic or mimic control for 24 h prior to the luciferase assays. (F) A549 cells were transfected with the miR-200c mimic or mimic control, followed by infection with H5N6 virus at an MOI of 0.01 or 3. At the indicated times post-infection, the supernatant containing viral particles was assessed in the TCID 50 assay. (G) A549 cells were transfected with the miR-200c inhibitor or inhibitor control, followed by infection with H5N6 virus at an MOI of 0.01 or 3. At the indicated times post-infection, the supernatant containing viral particles was assessed in the TCID 50 assay. (H) miR-200c +/+ or miR-200c -/cells were transfected with the miR-200c mimic or mimic control. Twenty-four hours later, the cells were infected with H5N6 virus at an MOI of 0.01 or 3. At the indicated times post-infection, the supernatant containing viral particles was assessed in the TCID 50 assay. The data collected from miR-200c -/cells are shown as dotted lines. The significant difference between miR-200c +/+ -mimic-Ctrl and miR-200c +/+ -200c-mimic is labeled in blue; the significant difference between miR-200c +/+ -mimic-Ctrl and miR-200c -/--mimic-Ctrl is labeled in green; the significant difference between miR-200c -/--mimic-Ctrl and miR-200c -/--200c-mimic is labeled in red. (I) A549 cells were transfected with the miR-200c mimic or mimic control, and the CNTN1 expression plasmid or Vec. Twenty-four hours later, the cells were infected with H5N6 virus at an promoter and ISRE was significantly elevated (Figs S2A and S2B and 3B). In addition, CNTN1 inhibited the activation of the IFN-β promoter in a dose-dependent manner (Fig 3C). In contrast, CNTN1 did not affect the activation of IRF1 or STAT1 triggered by IFN-β (Figs 3D and S2C). Our qPCR analysis indicated that CNTN1 significantly inhibited SeV-triggered transcription of the Interferon beta 1 (IFNB1), IFN-stimulated gene 15 (ISG15), IFN-stimulated gene 56 (ISG56), Regulated upon activation normal T cell expressed and secreted factor (RANTES), C-X-C motif chemokine ligand 10 (CXCL10) and Oligoadenylate synthetase-like protein (OASL) genes in A549 cells (Fig 3E). CNTN1 knockdown or knockout significantly promoted the expression of IFNB1 and ISG genes (Figs 3F and S2D). Meanwhile, CNTN1 deficiency significantly promoted IAV-induced expression of ISG15 proteins in A549 cells (S2E Fig). SeV-induced phosphorylation of TBK1 and IRF3 were markedly inhibited by CNTN1, suggesting that CNTN1 blocks the activation of TBK1 and IRF3 ( Fig 3G). Consistent with this finding, CNTN1 significantly inhibited the induction of IFN-β triggered by RIG-I-CARD (RIG-IN) ( Fig 3H). To confirm that CNTN1-promoted viral replication was dependent on the IFN-I pathway, we used CRISPR/Cas9 to knock out endogenous interferon alpha/beta receptor 1 (IFNAR1) in A549 cells (S2F Fig). Viral growth kinetics assays showed that IFNAR1-deficiency abolished CNTN1-promoted viral replication, suggesting that CNTN1 promotes viral replication in an interferon-dependent manner ( Fig 3I). Since previous studies demonstrated that CNTN1 participates in multiple signaling pathways [17][18][19][20], we further investigated the role of CNTN1 in the TLR signaling and in defense against DNA viral infection. As shown in S2I and S2J Fig, overexpression of CNTN1 inhibited poly(I:C)-triggered activation of the IFN-β promoter in TLR3-expressed HEK293 cells, and promoted the replication of herpes simplex virus 1 (HSV-1), which suggested that CNTN1 participates in the regulation of TLR signaling and innate immune response against DNA virus. Since the RIG-I is considered to be the main viral RNA sensor and readily activated upon IAV infection [16], the present study focused on the role of CNTN1 in the RIG-I-mediated IFN-I signaling pathway.

CNTN1 specifically interacts with and degrades MAVS
We next investigated how CNTN1 inhibits the IFN-I signaling pathway. To determine the molecular target of CNTN1 in the regulation of virus-triggered IFN-I activation, we transfected plasmids encoding RIG-IN, MDA5, MAVS, or TBK1 together with the IFN-β promoter in the presence or absence of CNTN1. CNTN1 inhibited the activation of the IFN-β promoter triggered by the expression of RIG-IN, MDA5, and MAVS in a dose-dependent manner, but not TBK1 (Fig 4A). Co-immunoprecipitation (Co-IP) experiments indicated that MAVS, but MOI of 0.01 or 3. At the indicated times post-infection, the supernatant containing viral particles was assessed in the TCID 50 assay. The significant difference between Vec-mimic-Ctrl and Vec-200c-mimic is labeled in blue; the significant difference between Vec-mimic-Ctrl and CNTN1-mimic-Ctrl is labeled in green; the significant difference between Vec-mimic-Ctrl and CNTN1-200c-mimic is labeled in red. The data shown represent three independent experiments; bars represent the mean ± SD of the three independent experiments (n = 3). [P< 0.05 ( � ), P < 0.01 ( �� ), P < 0.001 ( ��� ), P < 0.0001 ( ���� ); 'ns' indicates no significant difference].
https://doi.org/10.1371/journal.ppat.1010299.g002 Luciferase reporter plasmids (IFN-β-Luc or ISRE-Luc) and the pRL-TK plasmid were co-transfected into HEK293 cells, along with Flag-CNTN1 or Vec (A) or CNTN1 siRNA or scrambled siRNA (NC) (B). At twenty-four hours after transfection, the cells were left uninfected or were infected with SeV for 12 h before reporter assays. (C) Luciferase reporter plasmids (IFN-β-Luc or ISRE-Luc) and the pRL-TK plasmid were co-transfected into HEK293 cells, along with different concentrations of CNTN1 expression plasmid. At twentyfour hours after transfection, cells were left uninfected or were infected with SeV for 12 h before reporter assays. (D) Luciferase reporter plasmids (IRF1-Luc) and the pRL-TK plasmid were co-transfected into HEK293 cells, along with Flag-CNTN1 or Vec or CNTN1 siRNA or NC. At twenty-four hours after transfection, the cells were left untreated or were treated with IFN-β for 12 h before reporter assays. (E-F) A549 cells were transfected with Flag-CNTN1 or Vec (E) or CNTN1 siRNA or NC (F), and subsequently left uninfected or infected with SeV for 12 h before qPCR analysis. (G) CNTN1 inhibits virus-induced phosphorylation of TBK1 and IRF3. A459 cells were transfected with GFP-CNTN1 plasmid or Vec. At twenty-four hours after transfection, the cells were infected with SeV for the indicated times. Then, immunoblotting analysis were performed with the indicated antibodies. (H) HEK293 cells were transfected with CNTN1 or Vec, and induced by RIG-IN transfection. The concentration of IFN-β in the supernatants was detected by an ELISA. (I) IFNAR1 +/+ or IFNAR1 -/-A549 cells were transfected with the CNTN1 expression plasmid or Vec. Twenty-four hours later, the cells were infected with H5N6 virus at an MOI of 0.01 or 3. At the indicated times post-infection, the supernatant containing viral particles was assessed in the not RIG-I, specifically interacted with CNTN1 ( Fig 4B). Pull-down assays showed that GST-CNTN1 directly interacted with His-MAVS in vitro ( Fig 4C). Endogenous Co-IP experiments further verified that endogenous CNTN1 directly associates with MAVS ( Fig 4D). Confocal microscopy confirmed that MAVS was mostly localized on mitochondria and CNTN1 dispersed throughout the cytoplasm in cells without IAV infection, which showed little colocalization with MAVS. While CNTN1 protein was downregulated and congregated in the cytoplasm in the IAV-infected cells, and colocalized with partial MAVS outside of mitochondria ( Fig 4E). The data suggest that both the Ig-like domains and FNIII domains are the critical domains for the interaction with MAVS.
Because MAVS expression was reduced in CNTN1-overexpressing cells (Fig 4B), we hypothesized that CNTN1 might regulates the expression or stability of MAVS. To test this hypothesis, we co-transfected CNTN1 together with HA-tagged MAVS and IRF3 and performed immunoblotting analysis. CNTN1 specifically downregulated the expression of MAVS, but not that of IRF3 ( Fig 4F). Next, we assessed the transcription of MAVS in A549 cells, and found that CNTN1 did not affect the transcription of MAVS ( Fig 4G). Of note, overexpressed CNTN1 downregulated the expression of endogenous MAVS, but not IRF3 ( Fig  4H). Additionally, the transfection experiments in the wild-type and CNTN1-deficient cells treated with CHX indicated that CNTN1 deficiency prolonged the half-life of endogenous MAVS ( Fig 4I).
Collectively, our results reveal that CNTN1 specifically interacts with and induces the degradation of MAVS.

CNTN1 promotes the degradation of MAVS through a ubiquitinproteasomal pathway
The ubiquitin-proteasome pathway and the autolysosome pathway are the major intracellular protein degradation pathways [22]. To determine which of these pathways is responsible for CNTN1-mediated MAVS degradation, we added the proteasome inhibitor MG132, the caspase inhibitor ZVAD, and the autophagy inhibitors 3-methylademine (3-MA), chloroquine (CQ), and bafilomycin A1 (Baf A1), or the lysosome inhibitor ammonium chloride (NH 4 Cl) to cells co-transfected with CNTN1 and MAVS. CNTN1-mediated MAVS degradation was mostly restored by treatment with MG132 or ZVAD, suggesting that CNTN1 mediates the degradation of MAVS in a proteasome-dependent manner ( Fig 5A).
https://doi.org/10.1371/journal.ppat.1010299.g003 (G) CNTN1 has no effect on the expression of MAVS mRNA. HEK293 cells were transfected with Vec or the CNTN1 plasmid for 24 h before qPCR analysis. (H) CNTN1 decreases endogenous MAVS expression. HEK293 cells were transfected with increasing amounts of Flag-CNTN1 for 24 h before immunoblotting analysis. (I) CNTN1 +/+ and CNTN1 -/cells were treated with CHX (100 μg/mL). Cells were harvested for total protein extraction at 0, 3, 6, and 9 h after treatment, then immunoblotting analysis was with or without SeV stimulation (Fig 5C). It was reported that K63 ubiquitination contributed to the prion-like aggregates of MAVS [24]. The semi-denaturing detergent agarose-gel electrophoresis (SDD-AGE) assay indicated that the aggregation of MAVS was higher in CNTN1 −/− cells than that in CNTN1 +/+ cells (Fig 5D).

CNTN1 potentiates USP25-mediated proteasomal degradation of MAVS
CNTN1 is neither an E3-ubiquitin ligase nor a deubiquitination enzyme (DUB). Therefore, we hypothesized that CNTN1 might functions as a scaffold to link DUBs to MAVS for deubiquitination and degradation. Previous studies have identified several DUBs that participate in the antiviral response, including A20, USP10, USP25, CYLD, OTUB1, and OTUB2 [26][27][28]. Co-IP experiments revealed that CNTN1 specifically interacts with USP25, but not other DUBs ( Fig 6A). Flag-USP25 specifically interacted with HA-MAVS, and endogenous USP25 interacted with Flag-CNTN1 (Fig 6B and 6C). Pull-down analysis showed that His-MAVS directly interacts with GST-USP25, and His-USP25 directly interacts with GST-CNTN1 (Fig 6D and  6E). As expected, endogenous USP25 specifically interacted with MAVS and CNTN1 (Fig 6F). Domain mapping experiments suggest that the Ig-like domains of CNTN1 are critical for the interaction with USP25 (S3D Fig).
To determine the colocalization of CNTN1, USP25, and MAVS in infected or uninfected cells, the immunofluorescent experiment was conducted. Confocal microscopy showed that CNTN1 and USP25 colocalized and dispersed throughout the cytoplasm without virus infection, which had a partial association with MAVS. While USP25 protein was downregulated and congregated with CNTN1 in the cytoplasm of the IAV-infected cells, and colocalized with partial of MAVS that did not associate with mitochondria (Fig 6G and 6H). Moreover, the interaction between USP25 and MAVS was significantly impaired in CNTN1-deficient cells (Fig 6I), and CNTN1-USP25 interaction was markedly abrogated in MAVS-deficient cells ( Fig  6J).
It has been reported that miR-200c directly binds to the 3'-UTR of USP25, reducing both the mRNA and protein levels of USP25 [29]. As verification, we found that treatment with the miR-200c mimic markedly reduced USP25 expression (S4A  Next, we determined whether and how USP25 mediates the K63-linked deubiquitination and degradation of MAVS. Using the ubiquitination assay, we found that USP25 catalyzed the K63-linked deubiquitination of MAVS (Fig 7A). The K63-linked deubiquitination of MAVS was markedly increased in the USP25-deficient cells infected or uninfected with SeV (Figs 7B and  S4D). The SDD-AGE assay indicated that USP25-deficiency increased the aggregation of MAVS, while the CNTN1-induced K63-linked deubiquitination of MAVS was markedly reverted in the USP25-knockdown cells (Fig 7C and 7D). Furthermore, the USP25-catalyzed K63-linked deubiquitination of MAVS was mostly restored in the CNTN1-deficient cells (Fig 7E). These data suggest that CNTN1 recruits USP25 to mediate K63-linked deubiquitination of MAVS. We further revealed that knockdown of USP25 inhibits CNTN1-mediated degradation of MAVS ( Fig 7F).

Discussion
In this study, we demonstrated that endogenous CNTN1 is downregulated by virus-induced miR-200c. CNTN1 acts as a host suppressor of innate immunity by mediating the degradation of MAVS. CNTN1 collaborates with the deubiquitinating enzyme USP25 to mediate K63-linked deubiquitination of MAVS for proteasomal degradation. We have thus uncovered a novel mechanism through which the host factor CNTN1 promotes the replication of influenza A viruses through suppression of innate immune responses.
CNTN1 is the first identified member of the contactin family within the immunoglobulin superfamily, and a GPI-anchored membrane protein that facilitates cell adhesion. CNTN1 and other members of the CNTN family share a common structure of six C2 Ig-like repeats, four FNIII domains, and a GPI-linkage at the C-terminus that anchors the protein to the extracellular plasma membrane [21]. CNTN1 is highly expressed in the human brain and neuronal tissues, and plays an essential role in nervous system development, including neural cell adhesion, myelination, neurite growth, axonal elongation, and fasciculation [30][31][32]. Recent reports have revealed that CNTN1 is also involved in carcinogenesis and cancer progression [17][18][19]. For example, CNTN1 serves as a downstream effector of the vascular endothelial growth factor C (VEGFC)/FLT4 signaling pathway, which activates SRC/p38 mitogen-activated protein kinase (MAPK)-CEBPA signaling [19]. CNTN1 also promotes epithelial-mesenchymal transition, by activating the PI3K/AKT signaling pathway [20]. Nevertheless, the molecular mechanism of CNTN1 involved host-virus interaction has never been demonstrated, although one study described that CNTN1 potentially interacted with classical swine fever virus p7 [33]. The present study discovered that CNTN1 is downregulated upon infection with H1N1, H5N6, or H7N9 IAV subtypes. In addition, CNTN1 facilitated the replication of these viruses (Fig 1). CNTN1 thus participates in the regulatory mechanism of IAV replication.
The innate immune response is the first line of defense against viral infection, and RIG-Imediated type I interferon signaling is one of the most important innate immune responses were transfected with a Flag-USP25 plasmid for 24 h, and subsequently left uninfected or infected with H5N6 virus for 12 h before staining with mito-tracker. In the zoomed images, the yellow color indicated the colocalization of Flag-USP25 and MAVS, the purple color indicated the colocalization of mitochondria and MAVS. (I) CNTN1-deficiency inhibits the interaction between USP25 and MAVS. CNTN1 +/+ or CNTN1 -/-HEK293 cells were transfected with the indicated plasmids for 24 h before co-immunoprecipitation and immunoblotting analysis. (J) MAVS-deficiency inhibits the interaction between USP25 and CNTN1. MAVS +/+ or MAVS -/-HEK293 cells were transfected with the indicated plasmids for 24 h before co-immunoprecipitation and immunoblotting analysis. against infection with RNA viruses, such as influenza A virus [34]. Upon influenza A virus infection, type I IFNs are secreted and bind to the cognate receptor, type I IFN receptor (IFNAR), to initiate a signaling cascade that leads to the transcriptional induction of various ISGs [35]. Our results showed that IFNAR1-knockout abolished the CNTN1-mediated facilitation of IAV replication (Fig 3), demonstrating that CNTN1 facilitates IAV replication by suppressing the innate immune responses. We further revealed that CNTN1 inhibits the RIG-Imediated IFN-I signaling pathway and decreases the expression of IFN-β and IFN-stimulated genes, therefore promotes the replication of IAVs (Fig 3). Additionally, we confirmed that CNTN1 plays roles in the TLR signaling pathway and in defense against HSV-1 infection, which should be further investigated (S2I and S2J Fig).
As a mitochondrial membrane protein, MAVS functions as a critical adaptor of the RLR signaling pathway that links upstream recognition of viral RNA to downstream signal activation during antiviral responses [12]. Mounting evidence suggests that MAVS may be a common target for the regulation of the IFN-I signaling pathway. For viruses, certain viral proteins exert negative effects of innate immunity by targeting MAVS for immune evasion [36][37][38][39][40]. For example, Li et al. revealed that the nonstructural protein 3 of Zika virus interacts with and induces degradation of MAVS to evade antiviral responses of host cells [41]. One of our previous studies revealed that the PB1 protein of H7N9 virus recruits the selective autophagic receptor NBR1 to associate with MAVS for autophagic degradation, resulting in disrupted innate signaling and enhanced viral replication [42]. For host cells, the activity of MAVS should be finely tuned to exert sufficient protective immune responses while avoiding excessive harmful immune pathology. Numerous factors have been reported to be involved in this fine-tuning. RNF115 constitutively interacts with MAVS and induces proteasomal degradation of MAVS in uninfected cells [43]. The Golgi protein 73 (GP73) interacts with MAVS and promotes its degradation to negatively regulate innate immunity [44]. Here, we revealed that CNTN1 inhibits IFN-I induction by specifically mediating the degradation of MAVS (Fig 4).
Ubiquitination is strategically positioned to fine-tune the magnitude and duration of immune signaling cascades and, therefore, is also an attractive target for viral subversion mechanisms [45]. Several studies have demonstrated that MAVS could be modified by ubiquitination during viral infection [25,[46][47][48][49]. Generally, the stability of MAVS is regulated via K48-linked ubiquitination by several ubiquitin E3 ligases (e.g., RNF5, RNF125, AIP4/ITCH, Smurf1/2, and MARCH5), which may control excessive immune responses or be exploited by viruses to block the activation of innate signaling pathways [50][51][52][53]. We previously reported that the PB1 protein of H7N9 virus markedly enhances the K27-linked ubiquitination of MAVS to autophagic degradation [42]. However, in the present study, CNTN1 markedly reduced the K63-linked ubiquitination of MAVS for proteasomal degradation (Fig 5). Consistent with this latter finding, Cheung et al. demonstrated that the viral protein PB1-F2 of avian influenza virus suppresses K63-polyubiquitination and aggregation of MAVS, and directs unaggregated MAVS for degradation [54].
Following viral infection, MAVS undergoes a conformational switch that leads to the formation of prion-like functional aggregates on the mitochondrion. It is now known that TRIM31-mediated K63-polyubiquitination of MAVS is essential for the aggregation and activation [25]. However, persistent MAVS signaling leads to host immunopathology, and it transfected with Myc-MAVS and HA-ubiquitin (K63O), together with a control or USP25 plasmids. The cells were then subjected to denature-IP and immunoblotting analysis with the indicated antibodies. (F) USP25-knockdown inhibits CNTN1-mediated degradation of MAVS. HEK293 cells were transfected with the indicated plasmids and USP25 siRNA or NC for 24 h before immunoblotting analysis. https://doi.org/10.1371/journal.ppat.1010299.g007

PLOS PATHOGENS
CNTN1 -a novel negative regulator of MAVS remains largely unknown how these MAVS aggregates are resolved. RACK1 has been reported to enhances K48-linked ubiquitination of MAVS, attenuates its K63-linked ubiquitination, and decreases MAVS-mediated antiviral signal transduction [55]. Recently, Liu et al. found that the deubiquitinase YOD1 cleaves the K63-linked ubiquitination of MAVS and negatively regulates antiviral innate immunity [56]. In the present study, we discovered that the deubiquitinase USP25 is recruited and interacts with CNTN1 to remove K63-linked ubiquitination of MAVS for degradation (Fig 7).
There is accumulating evidence that noncoding RNAs, including circRNAs, lncRNAs, and miRNAs that are key post-transcriptional regulators, and play important roles in virus-host interactions [57][58][59]. MiR-200c, a well-established tumor suppressor miRNA, participates in the responses against multiple viruses, including hepatitis B virus and hepatitis C virus, by targeting various transcripts [60,61]. One study indicates that H5N1 avian influenza virus upregulates the expression of miR-200c, which reduces the level of Angiotensin-converting enzyme 2 (ACE2), and eventually leads to increased angiotensin II levels and subsequent lung injury [62]. In this study, we showed that miR-200c is upregulated in response to H5N6 virus infection and suppresses viral replication by directly targeting CNTN1 and USP25 to enhance IFN-I signaling (Figs 2 and S4). Of note, the data in this study showed that miR-200c-deficiency did not completely restore the virus-induced degradation of CNTN1 (S4C Fig). Additionally, virus infection led to degradation of GFP-CNTN1 which did not have a miR-200ctargeted 3'UTR (Figs 3G and S4E). Therefore, it's reasonable to speculate that another certain degradation mechanism, but not miR-200c targeted, was induced to downregulate CNTN1 protein upon virus infection. The underpinned mechanism should be further investigated.
Based on our data, we propose a working model for the regulatory role of CNTN1 in the RIG-I-mediated interferon signaling pathway (Fig 8). In non-infected cells, the CNTN1 protein acts as a negative regulator of the IFN-I pathway, by recruiting USP25 to deubiquitinate the K63-linked MAVS and promoting its degradation. Upon infection by influenza A viruses, the RIG-I-mediated signaling pathway is activated by the viral RNA. To fulfill the needs of type I interferon production, miR-200c is upregulated to reduce CNTN1 and USP25 expression and inhibit MAVS degradation. Finally, the RIG-I-mediated innate signaling is elevated, leading to the promotion of type I IFN responses and suppression of viral replication. Our findings clarify the role of CNTN1 in immune function against IAVs and provide valuable insights into the MAVS regulation network.

Biosafety and ethical statements
This study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the Ministry of Science and Technology of the People's Republic of China. Studies with highly pathogenic H5N6 avian influenza viruses were carried out in a biosecurity level 3 laboratory approved for such use by the Chinese Ministry of Agriculture and Rural Affairs. The protocol was approved by the Committee on the Ethics of Animal Experiments of the Harbin Veterinary Research Institute (HVRI) of the Chinese Academy of Agricultural Sciences (CAAS). The details of the facility and the biosafety and biosecurity measures used have been previously reported [63].

Reagents and antibodies
The antibodies used in this study were as follows: poly (

RNA isolation and quantitative PCR
Total RNA from cells was extracted with TRIzol following the manufacturer's instructions. For mRNAs, total RNA was subsequently transcribed into cDNA using M-MLV Reverse Transcriptase, according to the manufacturer's protocol (Promega). For miRNAs, total RNA was reversetranscribed according to the protocol provided with the All-in-One miRNA First-Strand cDNA Synthesis Kit (Genecopoeia, China). RNU6 and GAPDH were used as invariant controls for miRNAs and mRNAs, respectively. Primers for miR-200c were designed by Genecopoeia Co. (China). Real-time PCR was carried out using the ABI 7500 detection System (Applied Biosystems, USA). The RNA level of each gene is shown as fold of induction (2 -ΔΔCT ) in the graphs. The sequences of the gene-specific primers used for qPCR are provided in S2 Table.

Viral infection
All cells were seeded at the desired density in culture plates as per the requirements for different experiments. Viruses were inoculated into cells at a specific multiplicity of infection (MOI) for various experiments. One hour after inoculation, the medium was replaced with fresh OPTI-MEM and the cells were incubated at 37˚C. Virus-containing culture supernatants were collected at the indicated time points for titration.

Virus titration
Viral titers of virus stocks and cell culture supernatants were determined by end-point titration in MDCK cells. Ten-fold serial dilutions of each sample were inoculated into MDCK cells. Two days after inoculation, supernatant from the inoculated cells was collected and tested for the ability to agglutinate chicken erythrocytes as an indicator of viral replication. Infectious viral titers are reported as log 10 TCID 50 /mL, and were calculated from 3 replicates by using the method of Reed-Muench [66].

CRISPR-Cas9 knockout
Genome engineering was performed using the CRISPR-Cas9 system [67,68]. Double-stranded oligonucleotides corresponding to the target sequences were cloned into the pX459-V2.0-SpCas9-HF1 vector. Two micrograms of each pX459 plasmid containing one of the targeting sequences were simultaneously transfected into A549 cells. The transfected cells were selected with puromycin (1 μg/ml) for at least 7 days to obtain knockout cell pools. The following sequences were targeted for human CNTN1, IFNAR1, USP25, TRIM31 or miR-200c genes:

Prediction of miRNA targeting the 3'-UTR of the CNTN1 mRNA
The miRNA targeting the 3'-UTR of CNTN1 mRNA was predicted and selected using MicroRNA.org [13], TargetScan [14] and starBase [15]. All analysis was performed by considering the best optimum scores of each program.

Dual-luciferase reporter assays
To validate the miRNA targeting CNTN1, luciferase reporter vectors containing WT CNTN1 3'-UTR or the mutant were co-transfected with the control mimic or miR-200c mimic into HEK293 cells. At twenty-four hours after transfection, the transfected cell lysates were analyzed by using the dual luciferase assay kit (Promega). All obtained luciferase values were normalized against those of the Renilla luciferase control.
To detect activation of the IFN-I pathway, HEK293 cells grown in 24-well plates were cotransfected with Luciferase reporter plasmids (IFN-β-Luc, ISRE-Luc, or STAT1-Luc) and the pRL-TK plasmid, along with the indicated amount of empty vector or plasmids expressing CNTN1 or other molecules. At 24 h post-transfection, the cells were left untreated or were treated with SeV, IFN-β, or IAV for an additional 12 h. Cell lysates were prepared and analyzed for firefly and Renilla luciferase activities by using the dual luciferase assay kit (Promega).

Western blotting
Cells were lysed in RIPA buffer (Beyotime, China). Proteins were separated by 10% SDS-PAGE and transferred to a nitrocellulose membrane (Bio-Rad). The membrane was blocked for 1 h in TBST containing 5% milk and subsequently incubated with primary antibodies overnight at 4˚C. After a 1-h incubation with HRP-conjugated secondary antibody, the immunoreactive bands were visualized by using an ECL system (GE Healthcare). The intensities of the target bands were quantified by using the Image J program (NIH, USA).

GST pull-down assay
GST pull-down assays were conducted as previously described with slight modifications [39]. Briefly, the encoded GST-or His-tagged fusion proteins and the control GST proteins were expressed in BL21 cells after induction with 0.1 mmol/L IPTG overnight at 18˚C. Centrifuged cells were resuspended in lysis buffer (1 × PBS, 0.2 mM PMSF, 1% Triton X-100) and sonicated for 15 min. After centrifugation, the supernatant was applied to a Glutathione-Sepharose 4B bead column (GE Healthcare) or ProteinIso Ni-NTA Resin (TransGen Biotech, China), in accordance with the manufacturers' instructions. Purified GST/His-tagged fusion proteins were diluted with 1× PBS and filtered through Amicon Ultra 0.5 ml filters (Millipore). Then, 1 μg of purified GST protein or GST fusion protein was captured by the Glutathione-Sepharose 4B beads (GE Healthcare), and His-tagged fusion protein was added for incubation overnight at 4˚C. The beads were then washed three times with ice-cold PBS. The supernatant was loaded onto gels, followed by immunoblotting analysis.

Co-immunoprecipitation
HEK293 cells or A549 cells were co-transfected with the indicated plasmids with or without virus infection for 24 h. The transfected cells were then harvested and lysed in NP-40 lysis buffer [20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1% NP-40, 1 mM EDTA with protease inhibitor cocktails]. For each immunoprecipitation, 1 ml of lysate was incubated for 4 h at 4˚C with 0.5 μg of the indicated antibody or control IgG and 30 μL of protein A/G-Sepharose (Sigma). The beads were washed three times with 1 ml of lysis buffer containing 500 mM NaCl. The precipitates were analyzed by using standard immunoblotting procedures.

Confocal microscopy
Confocal microscopy was performed as previously described [39]. Cells were seeded in 12-well plates (5 × 10 5 cells/well) on coverslips. At twenty-four hours after transfection, the cells were left uninfected or were infected with IAV (MOI = 1) for 12 h before staining with mito-Tracker (Beyotime). Cells were then fixed with 4% paraformaldehyde for 20 min at room temperature, and washed three times with PBS. Cells were permeabilized with 0.1% Triton X-100 in PBS for 10 min and blocked with 5% skimmed milk for 1 h. Then, the cells were incubated with the indicated primary and secondary antibodies and DAPI. The stained cells were observed with a Leica microscope (TCS SP8) with a 100× oil objective.

Detection of ubiquitin-modified proteins
The experiments were performed as previously described [70]. Briefly, the cells were lysed in lysis buffer containing 1% SDS and denatured by heating at 95˚C for 10 min. After centrifugation, the supernatants were diluted with NP-40 lysis buffer until the concentration of SDS was 0.1%, and were then co-immunoprecipitated with the indicated antibodies. Ubiquitin-modified proteins were detected by immunoblotting with the indicated antibodies.

SDD-AGE assay
SDD-AGE was performed as previously described [24,54]. Briefly, cells were lysed with cold SDD-AGE lysis buffer (20 mM Tris-HCl pH 8.0, 137 mM NaCl, 10% glycerol, 1% NP-40, 2 mM EDTA and 1× protease inhibitor) for 30 min. Appropriate amount of 5× SDD-AGE sample buffer (2.5× TBE, 2.5% SDS, 25% glycerol and 0.25% bromophenol blue) was then added to SDD-AGE lysate to reach 1×. The crude sample was passed through 25G needle 10 times at 4˚C to break viscosity. The SDD-AGE protein sample was then separated by 2% agarose gel and transferred onto PVDF Membrane (BioRad). The membrane was blocked with 5% skimmed milk and then incubated with primary and secondary antibodies. The immunoreactive bands were visualized by using an ECL system (GE Healthcare).