The Interaction of Mandarin Fish DDX41 with STING Evokes type I Interferon Responses Inhibiting Ranavirus Replication

DDX41 is an intracellular DNA sensor that evokes type I interferon (IFN-I) production via the adaptor stimulator of interferon gene (STING), triggering innate immune responses against viral infection. However, the regulatory mechanism of the DDX41-STING pathway in teleost fish remains unclear. The mandarin fish (Siniperca chuatsi) is a cultured freshwater fish species that is popular in China because of its high market value. With the development of a high-density cultural mode in mandarin fish, viral diseases have increased and seriously restricted the development of aquaculture, such as ranavirus and rhabdovirus. Herein, the role of mandarin fish DDX41 (scDDX41) and its DEAD and HELIC domains in the antiviral innate immune response were investigated. The level of scDDX41 expression was up-regulated following treatment with poly(dA:dT) or Mandarin fish ranavirus (MRV), suggesting that scDDX41 might be involved in fish innate immunity. The overexpression of scDDX41 significantly increased the expression levels of IFN-I, ISGs, and pro-inflammatory cytokine genes. Co-immunoprecipitation and pull-down assays showed that the DEAD domain of scDDX41 recognized the IFN stimulatory DNA and interacted with STING to activate IFN-I signaling pathway. Interestingly, the HELIC domain of scDDX41 could directly interact with the N-terminal of STING to induce the expression levels of IFN-I and ISGs genes. Furthermore, the scDDX41 could enhance the scSTING-induced IFN-I immune response and significantly inhibit MRV replication. Our work would be beneficial to understand the roles of teleost fish DDX41 in the antiviral innate immune response.


Introduction
Asp-Glu-Ala-Asp (DEAD) box 41 (DDX41) is a crucial cytosolic DNA sensor and plays a crucial role in antiviral innate immune response via the adaptor stimulator of interferon gene (STING) to induce type I interferon (IFN-I) response. DDX41 consists of two conserved RecA-like domains (DEAD and HELIC domains), which contain various functional motifs involved in nucleic acid recognition, ATP binding, ATP hydrolysis, and RNA unwinding [1][2][3][4]. In the mammalian DDX41-STING signaling pathway, Bruton's tyrosine kinase phosphorylates the DDX41 HELIC domain at Tyr414, and the activated DDX41 recognizes double-stranded DNAs (dsDNAs) and bacterial cyclic dinucleotides (CDNs), such as cyclic di-AMP and cyclic di-GMP via its DEAD domain [5][6][7]. DDX41 binds to STING located on the endoplasmic reticulum membrane through its DEAD domain to activate IκB kinase-β and TANK-binding kinase 1 (TBK1); the activated kinases phosphorylate IκB and interferon regulatory factor (IRF) -3/-7 trigger the production of proinflammatory cytokines and IFNs [7][8][9]. The HELIC domain of DDX41 regulates the DEAD domain to bind

Fish, Cells, and Virus
Fifty mandarin fish (body weight of 75-100 g) were purchased from a farm in Guangdong province and maintained at 27 • C in a laboratory recirculating freshwater system to acclimatize for 2 weeks before being used. Six fish were randomly selected for pathogen detection, following which, five fish were selected for tissue expression analysis and ten fish for viral infection. For tissue sampling, fish were anaesthetized with MS-222 (40 mg/L, Sigma-Aldrich, St. Louis, MO, USA). All animal experiments were performed in accordance with the regulation for animal experimentation of Guangdong Province, China, and approved by the Ethics Committee of Sun Yat-sen University (No. 2019121705). The MRV strain NH-1609 (GenBank: MG941005.1) was originally isolated from disease-infected mandarin fish and preserved in our laboratory [18]. The mandarin fish fry (MFF-1) cell line was cultured in Dulbecco's modified Eagle medium (DMEM; Gibco, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (FBS; Gibco, Grand Island, NY, USA) at 27 • C under a humidified atmosphere containing 5% CO 2 [19].

Molecular Cloning of scDDX41 cDNA
Total RNAs were extracted from MFF-1 cells, and cDNAs were synthesized using previously described methods [17]. The full-length cDNA sequence of scDDX41 was obtained by rapid amplification of cDNA-end (RACE) reactions using a SMARTer RACE cDNA Amplification Kit (Clontech, Tokyo, Japan) in accordance with the manufacturer's instruction using the primers listed in Table 1. The PCR products were purified, cloned into the pMD19-T vector (Clontech, Tokyo, Japan) and sequenced (Thermo Fisher Scientific, Waltham, MA, USA).

Sequence Analysis
Homology sequences were obtained using the BLAST program at the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/blast). The deduced amino acid sequence of scDDX41 was analyzed using the Simple Modular Architecture Research Tool (SMART) program (http://smart.emblheidelberg.de/). A phylogenetic tree was constructed using the Bootstrap Neighbor-Joining method of the Molecular Evolutionary Genetics Analysis (MEGA) software v10.0 program. Bootstrap sampling was reiterated 1000 times.

Tissue Expression Profiles of scDDX41
Total RNAs were extracted from the gill, fin, spleen, intestine, brain, head kidney, hind kidney, middle kidney, blood, fat, heart, liver, and muscles using the SV Total RNA Isolation Kit (Promega, Madison, Wisconsin, USA) in accordance with manufacturer's instructions to detect the tissue distribution pattern of scDDX41 in healthy fish. The expression of scDDX41 in various tissues was determined by quantitative reverse-transcriptase PCR (RT-qPCR) using the primers in Table 2.

Dual-Luciferase Reporter Assays
Cells were cultured in 24-well plates for 24 h and co-transfected with the IFN-β-luc, or NF-κB-Luc plasmid (0.4 µg), tested plasmid (0.4 µg), and pRL-TK (40 ng) plasmid. After 2 h, the transfection mixture was replaced with 500 µL of DMEM with 10% FBS. After 36 h of transfection, cells were harvested and lysed in 200 µL of Passive Lysis Buffer (Promega, Madison, Wisconsin, USA). Total cell lysates were subjected to a Dual-Luciferase Reporter Gene Assay Kit (Promega, Madison, Wisconsin, USA) in accordance with the manufacturer's instructions. Luciferase activities were measured using Glomax (Promega, Madison, Wisconsin, USA). Experiments were performed in triplicate, and data represent the mean of at least three independent experiments. All experiments were performed in at least three independent trials with three technical replicates.

RT-qPCR
The RT-qPCR reactions were performed with a SYBR premix ExTaqTM (Takara, Tokyo, Japan) on a LightCycler ® 480 instrument (Roche Diagnostics, Switzerland) as previously described [20]. The primers for RT-qPCR were designed using Primer premier software v5.0 program (Applied Biosystems, Table 2). The expression of each transcript was normalized to the expression of the β-actin gene. All data of the target genes were analyzed using the Q-gene statistics add-in, followed by an unpaired sample t-test. Statistical significance was accepted at p<0.05, and high significance was accepted at p<0.01. All data are expressed as mean ± standard deviation (SD).

Absolute Quantitative Real-time PCR (qPCR) and TaqMan Probe
Virus load was determined by viral genomic copies using qPCR. DNA was extracted using the DNA isolation Mini Kit (Vazyme, Nanjing, China). The levels of viral genomic copies were quantified by the amount of viral major capsid protein (mcp) gene copies, which were determined with the method of standard curve. Specific primers and probes were designed from the partial MRV mcp sequence of NH-1609 using Primer Express v3.0 program. The forward primer 5 -GCCAAAAACCTTGTCCTTCCT -3 , reverse primer 5 -TCGTTGTAAGGCAGGGTGACT-3 , and the probe 5 -FAM-CCCTTCTTTTTCGGCAGAGA-BHQ1-3 were identified. Each PCR mixture in 10 µL contained 5 µL 2× Real PCR Easy TM Mix-Taqman (FOREGENE, Chengdu, China), 400 nM primer each, 200 nM probe, and 100 ng total DNA template. The program started with an initial denaturation at 94 • C for 3 min, followed by 40 cycles at 94 • C for 10 s and 60 • C for 30 s. The amplification and date analysis were carried out in Roche LightCycler ® 480 (Roche Diagnostics, Switzerland).

Determinations of Virus Titer
Cells were seeded into 96-well dishes. After growing overnight, cell density exceeded 80%. Briefly, cells cultivated in 96-well plates were infected with serial 10-fold dilutions of MRV-containing samples, with eight wells per dilution, and 100 µL of culture medium without virus was placed in eight wells as negative control. Then, the 96-well dishes were incubated at 27 • C. After 7 days of incubation, the number of positive and negative wells were recorded; the TCID 50 was calculated using the Spearman-Karber method as previously described [20].
2.11. Co-Immunoprecipitation (Co-IP) and Western Blot (WB) Analysis The interaction between scDDX41 and scSTING was analyzed via Co-IP assay. Cells seeded on 60 mm dishes (1 × 10 7 cells/dish) were transfected with 8 µg of empty plasmid or various expression plasmids including pCMV-myc-scDDX41 plus pCMV-flag-scSTING (at a ratio of 1:1), pCMV-myc-scDDX41 plus pCMV-flag (at a ratio of 1:1), and pCMV-myc plus pCMV-flag-scSTING (at a ratio of 1:1). At 24 h post-transfection, the cells were placed in an ice-cold cell lysis buffer (Beyotime, Shanghai, China) containing a cocktail protease inhibitor (Merck Millipore, Billerica, Massachusetts, USA) for 30 min at 4 • C. The cell lysates were centrifuged at 12,000 g for 15 min at 4 • C, precipitated with mouse anti-myc mAb or mouse anti-flag mAb at 37 • C for 60 min, and incubated with 50 µL of protein G-magnetic beads (Thermo Fisher Scientific, Waltham, MA, USA) for 60 min. The beads were washed four times with lysis buffer and eluted with an SDS loading buffer by boiling for 10 min. The transfected cells were harvested at 36 h post-transfection and lysed for WB assay as previously described [17]. Protein bands were visualized using a High-sig chemiluminescence Western blot substrate kit (Tanon, Shanghai, China).

Electrophoretic Mobility Shift Assay (EMSA)
Cells were transfected with pCMV-myc-scDDX41, pCMV-myc-GFP, or pCMV-myc. After 24 h, the cells were collected, and nuclear proteins were extracted using NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Fisher Scientific, Waltham, MA, USA). The 5 biotin-labeled IFN stimulatory DNA (ISD) probe (5 -TACAGATCTACTAGTGATCTATGA CTGATCTGTACATGATCTACA-3 ) was synthesized by Thermo Fisher Scientific (Waltham, MA, USA). In the competitive binding experiment, the unbiotinylated probes at 10-or 100-fold molar excess over the labeled probes were used to challenge the complexes of wild-type probes and proteins. EMSA was performed using a LightShift Chemiluminescent EMSA kit (Thermo Fisher Scientific, Waltham, MA, USA). The nuclear proteins (10 µg) were incubated with 20 fmol probes for binding reactions between the probes and proteins, separated by 5% native PAGE, transferred to positively charged nylon membranes (Roche Diagnostics, Switzerland), and cross-linked by UV light. The biotin-labeled DNA on the membrane was detected by chemiluminescence and developed on X-ray films as mentioned earlier [21].

Pull-Down Assay
The 5 biotinylated sense strand of ISD was mixed with its complementary unlabeled DNA and annealed to make double-stranded bio-ISD. For negative controls and competition assays, unlabeled ISD and Bio-ISD, each at a final concentration of 1 µM, were incubated with Dynabeads M-280 Streptavidin (Thermo Fisher Scientific, Waltham, MA, USA) for 30 min at 4 • C in binding buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10% glycerol, 0.5 mM EDTA, 1 mM 2-mercaptoethanol, and 0.5% Nonidet P-40). The beads were washed with binding buffer three times and incubated with the proteins for 1 h at 4 • C in binding buffer. The beads were washed with binding buffer four times. The bound proteins were eluted by boiling for 10 min at 100 • C using an SDS-PAGE sample buffer [6].

Molecular Characteristics of scDDX41
The cDNA fragment of scDDX41 was obtained from transcriptome data (unpublished). The RACE method was used to clone the full-length cDNA of scDDX41.  Figure 1A). The expression levels of scDDX41 in various tissues were examined by RT-qPCR. The transcriptions levels of scDDX41 were constitutively detected in all the selected tissues (including gill, hind kidney, fin, blood, spleen, middle kidney, head kidney, intestine, fat, liver, heart, brain, and muscles) under healthy conditions. The expression of scDDX41 in blood was higher than those in the other tissues and was approximately 42 times that in the muscles ( Figure 1B). Phylogenetic analysis showed that scDDX41 was clustered with DDX41 proteins from other fish to form an exclusive group and was merged with mammalian DDX41 proteins into a large group with high bootstrap probability, which differed from the other DDX41 family members, including DDX1, DDX3X, and DDX5 ( Figure 1C). published). The RACE method was used to clone the full-length cDNA of scDDX41. The full-length cDNA of scDDX41 (MN482034) is 2250 bp and includes a 5′-untranslated region (UTR) of 133 bp, a 3′-terminal UTR of 272 bp, and an 1845 bp open reading frame encoding a protein of 614 aas. SMART analysis and amino acid alignment showed that scDDX41 contains a coiled coil region (78-114 aas), a DEAD domain (192-403 aas), a helicase superfamily C-terminal (HELIC) domain (438-519 aas), and a ZnF_C2HC domain (573-589 aas) ( Figure 1A). The expression levels of scDDX41 in various tissues were examined by RT-qPCR. The transcriptions levels of scDDX41 were constitutively detected in all the selected tissues (including gill, hind kidney, fin, blood, spleen, middle kidney, head kidney, intestine, fat, liver, heart, brain, and muscles) under healthy conditions. The expression of scDDX41 in blood was higher than those in the other tissues and was approximately 42 times that in the muscles ( Figure 1B). Phylogenetic analysis showed that scDDX41 was clustered with DDX41 proteins from other fish to form an exclusive group and was merged with mammalian DDX41 proteins into a large group with high bootstrap probability, which differed from the other DDX41 family members, including DDX1, DDX3X, and DDX5 ( Figure 1C).

Expressions of scDDX41 in Response to Immune Stimulations
The expression profiles of scDDX41 in response to immune stimulation were evaluated. Cells were treated with double-stranded RNA (dsRNA) analogue poly(I:C), DNA analogue poly(dA:dT) or MRV, and then the transcription levels of scDDX41 were detected at indicated time by RT-qPCR. The levels of scDDX41 transcription were not changed by poly(I:C) stimulation at concentrations of 0.1-1 µg/mL ( Figure 1D), whereas the levels were significantly increased by poly(dA:dT) stimulation at concentrations of 0.5-1 µg/mL ( Figure 1E) or upon infection with MRV ( Figure 1F). The levels of type I IFN (IFN-h) transcription in the controls were significantly increased by poly(I:C) and poly(dA:dT) stimulations ( Figure 1D,E). These observations suggested poly(dA:dT) stimulation or infection with MRV induced the expression of scDDX41, indicating that scDDX41 might be involved in innate immune responses.

ScDDX41 and Its Domains Evoked IFN-I and Inflammatory Response
To investigate the roles of scDDX41 in innate immune response, the IFN-I and NF-κB signaling pathway were detected. As shown in Figure 2A,B, overexpression of scDDX41 induced the activities of IFN-β-luc and NF-κB-luc in a dose-dependent manner, compared with the overexpression of pCMV-myc (negative control). The activities were also significantly increased in cells treated with poly (I:C), which was set as positive control. These results suggested that scDDX41 was capable of activating IFN-β and NF-κB promoters. According to SMART analysis, scDDX41 possessed two important domains, namely, Nterminal DEAD domain (181-413 aas) and a C-terminal HELIC domain (428-529 aas) ( Figure 2C). To investigate the roles of DEAD and HELIC domains in the IFN-I and NF-κB signaling pathway, various mutants of scDDX41 were constructed. As shown in Figure 2D,E, the activities of IFN-β-luc and NF-κB-luc were significantly increased by the DEAD domain compared with those in the full length of scDDX41, which was similar to previous studies [6,14]. Interestingly, the HELIC domain could strongly induce the activities of IFN-β-luc and NF-κB-luc.
To further elucidate the roles of scDDX41 and its domains in innate immune responses, the effects on the expressions of innate immune-related genes were investigated using RT-qPCR. As shown in Figure 3, the expression levels of IFN-I signaling-related genes (scIFN-h, scMx, scISG15, and scViperin genes; Figure 3B-E) and pro-inflammatory factor (scTNF-α gene; Figure 3F) in the cells transfected with scDDX41-myc were significantly higher than those in the cells transfected with the control vector (pCMV-myc plasmid) at 24 h post-transfection (p<0.01). These observations suggested that scDDX41 evoked IFN and inflammatory responses. Moreover, the DEAD and HELIC domains strongly enhanced the expression levels of scMx and scISG15 compared with the cells transfected with control and scDDX41 plasmids ( Figure 3G,H), suggesting that the DEAD and HELIC domains of scDDX41 were important in IFN-I immune responses. signaling pathway, various mutants of scDDX41 were constructed. As shown in Figure  2D-E, the activities of IFN-β-luc and NF-κB-luc were significantly increased by the DEAD domain compared with those in the full length of scDDX41, which was similar to previous studies [6,14]. Interestingly, the HELIC domain could strongly induce the activities of IFNβ-luc and NF-κB-luc.  Luciferase assays were performed 36 h after the transfection. All luciferase assays were repeated at least three times, and data are means ±SD (n = 3) from single representative experiments. * p < 0.05, ** p < 0.01 between normal cells and stimulated cells.

ScDDX41 Interacted with scSTING
To investigate whether the function of scDDX41 in IFN-I expression depends on scSTING, scDDX41 and its domain interactions with scSTING were determined by Co-IP assay. The results showed that anti-flag mAb precipitated myc-tagged-scDDX41 in cells co-expressed with flag-tagged-scSTING and anti-myc mAb precipitated flag-tagged-scSTING in cells co-expressed with myc-tagged-scDDX41 ( Figure 4A). Co-IP experiments were conducted on the interaction between DEAD and HELIC with scSTING, respectively. Importantly, in contrast to mammalian DDX41, the HELIC domain of scDDX41 also directly interacted with scSTING ( Figure 4B) [3,6]. Furthermore, the interactions of the scSTING domain, scSTING N-terminal domain (NTD), and scSTING C-terminal domain (CTD) with the DEAD and HELIC domain of scDDX41 were explored. As shown in Figure 4C,D, the DEAD domain of scDDX41 was bound to both scSTING NTD and CTD, whereas the HELIC domain of scDDX41 was bound to scSTING NTD but not to CTD. mid) at 24 h post-transfection (p<0.01). These observations suggested that scD evoked IFN and inflammatory responses. Moreover, the DEAD and HELIC do strongly enhanced the expression levels of scMx and scISG15 compared with th transfected with control and scDDX41 plasmids ( Figure 3G-H), suggesting that the and HELIC domains of scDDX41 were important in IFN-I immune responses.

ScDDX41 Interacted with scSTING
To investigate whether the function of scDDX41 in IFN-I expression depen scSTING, scDDX41 and its domain interactions with scSTING were determined by assay. The results showed that anti-flag mAb precipitated myc-tagged-scDDX41 i co-expressed with flag-tagged-scSTING and anti-myc mAb precipitated flag-ta scSTING in cells co-expressed with myc-tagged-scDDX41 ( Figure 4A). Co-IP exper were conducted on the interaction between DEAD and HELIC with scSTING, r tively. Importantly, in contrast to mammalian DDX41, the HELIC domain of scDDX directly interacted with scSTING ( Figure 4B) [3,6]. Furthermore, the interactions To further investigate whether scDDX41 could enhance scSTING-induced IFN-I immune response, the expression levels of downstream genes in the cells co-transfected with scDDX41 and scSTING were examined. As shown in Figure 5A-E, scDDX41 significantly enhanced the activities of IFN-β-luc and the expression levels of scIFN-h, scMx, scISG15, and scViperin genes in the cells co-transfected with scSTING. These results suggested that scDDX41 enhanced the scSTING-induced IFN-I immune response.

ScDDX41 Recognizes dsDNA through the DEAD Domain
To verify whether teleost fish DDX41 could recognize dsDNAs, the interaction between scDDX1 and interferon stimulatory DNA (ISD) was evaluated using EMSA. As shown in Figure 5F, scDDX41 could bind to the labeled probe of ISD to form a retarded band, whereas the retarded bands of the DNA-protein complex were eliminated when 10× and 100× unlabeled probes were added to competitively bind to the scDDX41 protein. By contrast, no retarded band of the DNA-protein complex was observed in the analysis of the control group performed using nuclear proteins extracted from cells transfected with pCMV-myc-GFP or pCMV-myc empty plasmid. These findings confirmed that scDDX41 could bind to ISD. To further determine whether the DEAD domains of scDDX41 could directly bind to dsDNA, the dsDNA-binding abilities of myc-tagged scDDX41 and myctagged scDDX41∆DEAD were assessed by pull-down assays using 5 -biotinylated ISD (bio-ISD). The myc-tagged human DDX41 (hsDDX41) was used as positive control. As shown in Figure 5G, avidin beads plus bio-ISD immunoprecipitated hsDDX41 and scDDX41 but not scDDX41∆DEAD due to the absence of the DEAD domain. Hence, scDDX41 could bind to DNA through the DEAD domain. These results indicated that scDDX41 recognized dsDNA through the DEAD domain.  To further investigate whether scDDX41 could enhance scSTING-induced IFN-I immune response, the expression levels of downstream genes in the cells co-transfected with scDDX41 and scSTING were examined. As shown in Figure 5A-E, scDDX41 significantly enhanced the activities of IFN-β-luc and the expression levels of scIFN-h, scMx, scISG15, and scViperin genes in the cells co-transfected with scSTING. These results suggested that scDDX41 enhanced the scSTING-induced IFN-I immune response.

ScDDX41 Recognizes dsDNA through the DEAD Domain
To verify whether teleost fish DDX41 could recognize dsDNAs, the interaction b tween scDDX1 and interferon stimulatory DNA (ISD) was evaluated using EMSA. A shown in Figure 5F, scDDX41 could bind to the labeled probe of ISD to form a retarde band, whereas the retarded bands of the DNA-protein complex were eliminated whe 10× and 100× unlabeled probes were added to competitively bind to the scDDX41 protei By contrast, no retarded band of the DNA-protein complex was observed in the analys of the control group performed using nuclear proteins extracted from cells transfecte with pCMV-myc-GFP or pCMV-myc empty plasmid. These findings confirmed th scDDX41 could bind to ISD. To further determine whether the DEAD domains of scDDX4 could directly bind to dsDNA, the dsDNA-binding abilities of myc-tagged scDDX41 an myc-tagged scDDX41ΔDEAD were assessed by pull-down assays using 5ʹ-biotinylate ISD (bio-ISD). The myc-tagged human DDX41 (hsDDX41) was used as positive contro As shown in Figure 5G, avidin beads plus bio-ISD immunoprecipitated hsDDX41 an Figure 5. Involvement of scDDX41 in scSTING-mediated IFN expression, and scDDX41 recognizes dsDNA through the DEAD domain. (A) Cells transfected with an IFN-β-luc 400 ng/well and TK (40 ng/well) plus 400 ng/well) of the expression vectors for pCMV-myc, pCMV-myc and scDDX41myc, pCMV-myc and scSTING-myc, scDDX41-myc, and scSTING-myc. Vertical bars represent ±SD (n = 3). Statistical significance is indicated by asterisks, with ** referring to p < 0.01. scDDX41 enhanced the IFN-I response induced by STING. (B-E) Cells seeded in 6-well plates were transfected or co-transfected with scDDX41 (400 ng/well), and scSTING (400 ng/well). The cells transfected with pCMV-myc acted as negative control. The expression levels of interferon signaling molecules including scIFN-h, scMx, scISG15, and scViperin were examined using RT-qPCR. The β-actin gene served as the internal control to calibrate the cDNA template for all samples. Vertical bars represent ±SD (n = 3). Statistical significance is indicated by asterisks, with ** referring to p < 0.01. (F) scDDX41 recognizes ISD through the DEAD domain using EMSA with biotin-labeled (Bio-) or unlabeled (Unbio-) probes ISD. The black lines indicate where parts of the image were joined. (G) Immunoblot analysis of the immunoprecipitated purified myc-tagged hsDDX41, scDDX41, or scDDX41∆DEAD recombinant proteins incubated individually with biotinylated ISD and probed with anti-myc antibodies.

ScDDX41 Inhibits the Replication of Ranavirus
To investigate the potential antiviral roles of scDDX41, the transcription level of MRV viral major capsid protein (mcp), ICP-18 and DNA polymerase (DNA pol) genes were evaluated in scDDX41-overexpressing cells. As shown in Figure 6A-D, the transcriptions of MRV mcp, ICP-18, and DNA pol genes were significantly inhibited in scDDX41-overexpressing cells (A) compared with that in the cells transfected with the control vector at 24 and 48 h post-infection (p < 0.01). The relative expression levels of scIFN-h (E), scMx (F), scISG15 (G), scViperin (H), and scTNF-α (I) genes in the cells transfected with scDDX41-myc were significantly higher than those in the cells transfected with pCMV-myc at 24 and 48 h (p < 0.01) post-MRV infection. Moreover, the level of virus load and virus titers (TCID 50 ) was also suppressed by scDDX41 at 24 and 48 h post-infection, compared to those in the control group ( Figure 6J,K). These findings suggested that scDDX41 exerted antiviral activity. myc were significantly higher than those in the cells transfected with pCMV-myc at 24 and 48 h (p<0.01) post-MRV infection. Moreover, the level of virus load and virus titers (TCID50) was also suppressed by scDDX41 at 24 and 48 h post-infection, compared to those in the control group ( Figure 6J-K). These findings suggested that scDDX41 exerted antiviral activity. β-actin gene served as the internal control to calibrate the cDNA template for all samples. Vertical bars represent ±SD (n=3). Statistical significance is indicated by asterisks, with . Teleost fish DDX41 also interacts with STING via the DEAD domain; however, the HELIC domain of D. rerio DDX41 might obstruct the DDX41-STING signaling pathways [11]. Although the DDX41 of fish is highly homologous with the DDX41 from other species, there are differences between the STING of fish and those of other species [17,23]. The N-terminal TM domain of fish STING presents diversity; for example, grass carp STING has three TM structure domains [22], while grouper STING contains four TM domains [24], and the scSTING contains five TM domains [17]. The TM domains of STING

Discussion
DDX41, a member of the DExD/H-box helicase superfamily, is involved in intracellular DNA recognition and triggers innate immune responses against viral infection [3, 6,11]. Current knowledge about DDX41 mainly originates from humans and mice [6]. To fully understand the biological function of DDX41 in innate immunity, particularly of lower vertebrates such as teleost fish, scholars have cloned and studied DDX41 from olive flounder (Paralichthys olivaceus), zebrafish (Danio rerio), orange spotted grouper (Epinephelus coioides), Nile tilapia (Oreochromis niloticus), and grass carp (Ctenopharyngodon idella) for the past few years [10][11][12][13][14]. P. olivaceus DDX41 directly sensed cyclic di-GMP and triggered an IFN immune response to enhance the antiviral effect [21]. D. rerio DDX41 activated NF-κB and IFN-I signaling pathways through the interaction of the DEAD domain with STING [11]. E. coioides DDX41 enhanced MAVS-and TBK1-induced IFN and inflammatory immune responses [12]. C. idella DDX41 interacted with STING to induce the phosphorylation and nuclear translocation of IRF7 and then initiated IFN-I and ISG15 expressions [13]. However, no direct evidence is available to support the interaction of DDX41 with DNA (dsDNA, CDNs, or ISD) in teleost fish. In the present study, the DEAD domain of scDDX41 is a key region for binding to DNA (ISD) and interaction with STING to trigger IFN and inflammatory immune responses. Hence, scDDX41 is an important DNA patternrecognition receptor for STING-dependent type I IFN responses in teleost fish. DDX41 can directly bind to the STING-N-segment transmembrane region through the DEAD domain and subsequently promotes the activation of TBK1 and IRF3, leading to IFN-I production; meanwhile, the HELIC domain of DDX41 regulates the binding ability of the DEAD domain to dsDNA but does not interact with STING [6,22]. Teleost fish DDX41 also interacts with STING via the DEAD domain; however, the HELIC domain of D. rerio DDX41 might obstruct the DDX41-STING signaling pathways [11]. Although the DDX41 of fish is highly homologous with the DDX41 from other species, there are differences between the STING of fish and those of other species [17,23]. The N-terminal TM domain of fish STING presents diversity; for example, grass carp STING has three TM structure domains [22], while grouper STING contains four TM domains [24], and the scSTING contains five TM domains [17]. The TM domains of STING in fishes may lead to different mechanisms from those of mammals due to their high degree of unconserved sequences. The signal activation of STING depends on the IRF3 and NFkB. Different from the mammalian STING, the zebrafish STING can significantly stimulate a downstream NF-κB signal but the IRF3-IFN signal is weaker, which is due to the difference of the C terminal of STING [23,25]. Those findings indicated the difference in the activation of the DDX41-STING pathway between mammals and teleost fish. Interestingly, the DEAD domain of scDDX41 could directly interact with not only scSTING-NTD but also scSTING-CTD, and the HELIC domain of scDDX41 could also directly interact with scSTING-NTD. Hence, the DEAD and HELIC domains of scDDX41 contributed to the interaction of scDDX41 with scSTING to induce STING-dependent IFN-I and inflammatory immune responses. As such, the activation of the DDX41-STING pathway might differ between mammals and teleost fish. The detailed mechanisms of the DDX41 HELIC domains in synergistically regulating the STING-mediated innate immune response in fish should be further studied.
In summary, our study demonstrated that the scDDX41 activated the IFN-I responses by interacting with scSTING though the DEAD and HELIC domain, and the scDDX41mediated STING pathway possessed antiviral activity against MRV infection. This work will help us to understand the roles of teleost fish DDX41 and its HELIC domain in innate immune responses.