Functional Characterizations of RIG-I to GCRV and Viral/Bacterial PAMPs in Grass Carp Ctenopharyngodon idella

Background RIG-I (retinoic acid inducible gene-I) is one of the key cytosolic pattern recognition receptors (PRRs) for detecting nucleotide pathogen associated molecular patterns (PAMPs) and mediating the induction of type I interferon and inflammatory cytokines in innate immune response. Though the mechanism is well characterized in mammals, the study of the accurate function of RIG-I in teleosts is still in its infancy. Methodology/Principal Findings To clarify the functional characterizations of RIG-I in grass carp Ctenopharyngodon idella (CiRIG-I), six representative overexpression plasmids were constructed and transfected into C. idella kidney (CIK) cell lines to obtain stably expressing recombinant proteins, respectively. A virus titer test and 96-well plate staining assay showed that all constructs exhibited the antiviral activity somewhat. The quantitative real-time RT-PCR (qRT-PCR) demonstrated that mRNA expressions of CiIPS-1, CiIFN-I and CiMx2 were regulated by not only virus (GCRV) or viral PAMP (poly(IC)) challenge but also bacterial PAMPs (LPS and PGN) stimulation in the steadily transfected cells. The results showed that the full-length CiRIG-I played a key role in RLR pathway. The repressor domain (RD) exerted an inhibitory function of the signaling channel under all utilized challenges. Caspase activation and recruitment domains (CARDs) showed a positive role in GCRV and poly(I:C) challenge. Helicase motifs were crucial for the signaling pathway upon LPS and PGN stimulation. Interestingly, ΔCARDs (CARDs deleted) showed postive modulation in RIG-I signal transduction. Conclusions/Significance The results provided some novel insights into RIG-I sensing with a strikingly broad regulation in teleosts, responding not only to the dsRNA virus or synthetic dsRNA but also bacterial PAMPs.


Introduction
The innate immune system serves as the first line of protection against invading microbial pathogens through a limited number of germ line-encoded pattern recognition receptors (PRRs) [1]. The PRRs recognize different but overlapping pathogen-associated molecular patterns (PAMPs), and trigger innate immune responses and subsequent adaptive immunity [2]. Currently, four major classes of PRRs have been identified and classified into transmembrane proteins such as C-type lectin receptors (CLRs) and Toll-like receptors (TLRs), and cytoplasmic proteins such as NOD-like receptors (NLRs) and RIG-I-like receptors (RLRs) [3]. The innate immune system recognizes viral nucleic acids and microbial pathogens mainly by the latter three families [4]. TLRs are expressed in cell surfaces or in endosomes, and they recognize virus, bacteria, fungi and protozoa [5]. NLRs mainly detect pathogenic bacteria [6]. In contrast, RLRs primarily sense virusderived RNA molecules in the cytoplasm [7,8]. Upon activation, TLRs and RLRs trigger type I interferon (IFN-I), leading to an enhanced antiviral state of host cells [9].
RLRs are composed of three members: retinoic acid-inducible gene I (RIG-I, also known as DDX58), melanoma differentiation-associated gene 5 (MDA5, also known as IFIH1 or Helicard), and laboratory of genetics and physiology 2 (LGP2, also known as DHX58). RLRs evolve from a common ancestor encoding different core functional domains between mammals and teleosts [10,11]. Recently, the RLR family has become a major focus on the research of innate immunity [8,12]. RIG-I is a particular sensor in nucleotide recognition, and the regulatory functions of RIG-I are broad, not only in detection of RNA and DNA viruses but in recognition of bacterial component (such as lipopolysaccharide (LPS)) as well [13,14].
LPS is an outer-membrane component of gram-negative bacteria that can be recognized by TLR4 [30], and PGN is a major element of gram-positive bacterial cell walls that can be recognized by TLR2, NOD1 and NOD2 [31]. Since eukaryotic organisms do not contain LPS and PGN in their cellular structures, LPS and PGN are ideal target molecules for detecting bacterial invasion. It has been evidenced that RIG-I is induced in endothelial cells and macrophages post LPS stimulation [14,32].
Grass carp (Ctenopharyngodon idella) is employed as a model for antiviral immune studies because it is a crucial aquaculture species in China and is susceptible to grass carp reovirus (GCRV), a double-stranded RNA (dsRNA) virus [33]. Better understanding of the immune defense mechanisms may be conducive to the development of management strategies for disease control and comprehensive study on innate immune system evolution in teleosts [10,34].
To research functional characterizations of RIG-I in grass carp (CiRIG-I), six overexpression vectors were constructed, including full-length CiRIG-I and a series of domain containers, and they were transfected into C. idella kidney (CIK) cell line to obtain steadily expressing recombinant proteins. The mRNA expressions of downstream genes (CiIPS-1, CiIFN-I and CiMx2) of CiRIG-I were examined post GCRV and poly(I:C) (polyinosine-polycytidylic acid, a synthetic analog of dsRNA) challenges, as well as LPS and PGN stimulation.

Antiviral Activity of CiRIG-I and its Domains
To clarify the functional characterizations of CiRIG-I, six representative overexpression plasmids were constructed ( Fig. 1) for stably expressing recombinant protein in CIK cells, respectively. The pCMV (Fig. 2) was employed as a control. Measurement of the viral titer showed that overexpression of CiRIG-I and its variants decreased the viral titer more or less compared to that in control cells (Fig. 3A). Overexpression of CiRIG-I decreased the viral titers of 9-fold and 20-fold at 12 h and 48 h, respectively. Similarly, the viral titers in cells transfected with pDRD reduced 14-fold at 12 h and 31-fold at 48 h. The viral titers in cells transfectected with pDCARDs, pCARDs-RD and pCARDs also declined 4-5 folds at 12 h and 7-9 folds and at 48 h, respectively. As for the viral titers in cells transfected with pRD, they showed slight descent. The consistent results were obtained by antiviral activity assay (Fig. 3B). The cells transfected with pDRD and pRIG-I exhibited the powerful antiviral activity. The cells transfected with pDCARDs, pCARDs-RD and pCARDs displayed moderate roles of resistance to GCRV. The cells transfected with pRD owned little inhibitory effect (Fig. 3B).
CARDs and Helicase of CiRIG-I Participate in the Signaling Cascade after GCRV Infection, Whereas RD Inhibits the Activation The mRNA expression patterns of CiIPS-1 (Fig. 4A), CiIFN-I (Fig. 4B) and CiMx2 (Fig. 4C) were revealed by quantitative realtime RT-PCR (qRT-PCR) in steadily transfected cells after GCRV infection. According to the expression profiles of CiIPS-1 (Fig. 4A), the relative values in cells transfected with pRIG-I were up-regulated, reached 3.91 folds at 24 h and 4.57 folds at 48 h. The relative quantities in cells transfected with pDCARDs were slightly increased at 2 h (2.85 folds) and recovered the control level at 24 h, then were slightly enhanced at 48 h (2.46 folds). The relative values in cells transfected with pDRD were enhanced at 24 h (3.02 folds) and 48 h (3.96 folds). The relative folds in cells transfected with pCARDs-RD were slightly increased at 24 h (2.10 folds) and 48 h (2.81 folds). The relative quantities in cells transfected with pCARDs were enhanced at 24 h (3.29 folds) and 48 h (5.25 folds).The relative values in cells transfected with pRD were slightly increased at 24 h (2.10 folds), then decreased at 48 h (0.36 fold).
As for the expression profiles of CiIFN-I ( These results indicated that CARDs and helicase were crucial for antiviral activity, and RD inhibited the activation of RIG-I signaling cascade. Unexpectedly, the DCARDs significantly enhanced the signaling channel to inhibit GCRV replication.    The relative quantities in cells transfected with pDCARDs were slightly increased at 2 h (2.62 folds) and 24 h (3.03 folds), then sharply enhanced at 72 h (28.30 folds). The relative expressions in cells transfected with pDRD were no significant differences at 2 h, and rapidly increased at 24 h (4.62 folds) and 72 h (17.82 folds). Similar to the transfected pRIG-I cells, the relative quantities in cells transfected with pCARDs-RD were significantly enhanced at 2 h (3.15 folds), 24 h (5.00 folds) and 72 h (12.41 folds). Like the pattern in transfected pCARDs-RD cells, the fold changes in cells transfected with pCARDs were increased at 2 h (2.89 folds), 24 h (4.08 folds) and at 72 h (12.57 folds), respectively. The relative quantities in cells transfected with pRD were slightly increased at 2 h (3.00 folds) and 24 h (2.00 folds), then retrieved the control level at 72 h.

Relative Quantities of GCRV Decreased in pDCARDs Transfected Cells
Next, the temporary expression patterns of CiMx2  These results showed the CARDs were pivotal for signaling cascade upon poly(I:C) stimulation, and the RD played a negative role. As for helicase, it was not significant for RIG-I signaling pathway.

Helicase Domain of CiRIG-I Mediates Signaling Function upon LPS Stimulation and CARDs Play an Assistant Role
To further understanding on the mechanism of innate immune systems of CiRIG-I and its domains, the mRNA expression profiles of CiIPS-1 (Fig. 7A), CiIFN-I (Fig. 7B) and CiMx2 (Fig. 7C) were tested at 0, 2, 24 and 72 h after LPS stimulation. According to the expression patterns of CiIPS-1 (Fig. 7A), the relative values in cells transfected with pRIG-I were increased at 2 h (4.28 folds) and retrieved the control level at 24 h, then rapidly enhanced at 72 h (10.38 folds). The relative quantities in cells transfected with pDCARDs began to increase significantly at 2 h (5.62 folds) and declined to 2.40 folds at 24 h, then recovered the control level at 72 h. The expressions in cells transfected with pDRD were no significant differences at 2 h and 24 h, then enhanced at 72 h (4.25 folds). As for the temporal levels in cells transfected with pCARDs-RD, they were slightly increased at 2 h (2.10 folds), then recovered the control level at 24 h and 72 h. The relative expressions in cells transfected with pCARDs were no significant differences at 2 h, 24 h and 72 h. The relative quantities in cells transfected with pRD were slightly increased at 2 h (2.85 folds) and recovered the control level at 24 h, then decreased at 72 h (0.39 fold).
As for the expression profiles of CiIFN-I (Fig. 7B), the relative values in cells transfected with pRIG-I were rapidly increased at 2 h (13.81 folds) and decreased at 24 h (4.43 folds), then were sharply enhanced at 72 h (26.68 folds). The relative quantities in cells transfected with pDCARDs were significantly increased at 2 h (8.89 folds), and recovered the control level at 24 h and 72 h. The relative expressions in cells transfected with pDRD were moderately increased, and the relative folds were 3.51 at 2 h, 3.97 at 24 h and 6.32 at 72 h. Similar to the pDCARDs transfected cells, the relative values in cells transfected with pCARDs-RD were slightly increased at 2 h (2.35 folds), and retrieved the control level at 24 h and 72 h. The expressions in cells transfected with pCARDs were no significant differences at 2 h, 24 h and 72 h. The relative quantities in cells transfected with pRD were slightly enhanced at 2 h (3.11 folds) and 24 h (2.67 folds), then recovered the control level at 72 h.
Next, the temporary expression patterns of CiMx2 (Fig. 7C) were checked. The relative quantities in cells transfected with pRIG-I were increased at 2 h (7.17 folds) and recovered the control level at 24 h, then sharply enhanced at 72 h (11.86 folds). The relative values in cells transfected with pDCARDs began to increase significantly at 2 h (12.32 folds), and slightly enhanced at 24 h (2.42 folds), then recovered the control level at 72 h. The relative quantities in cells transfected with pDRD were no significant differences at 2 h and 24 h, then rapidly enhanced at 72 h (12.60 folds). Furthermore, the relative expressions in cells transfected with pCARDs-RD were slightly increased at 2 h (2.24 folds) and 24 h (3.39 folds), then retrieved the control level at 72 h. The relative values in cells transfected with pCARDs were no significant differences at 2 h, and slightly increased at 24 h (2.50 folds), then retrieved the control level at 72 h. The folds in cells transfected with pRD were 3.93 at 2 h and 3.04 at 24 h, then sharply decreased at 72 h (0.47 fold).
These results demonstrated that helicase domain was pivotal for RIG-I signaling cascade and CARDs played an assistant role. In contrast, RD inhibited the activation slightly. Additionally, CARDs and RD alone had little influence upon LPS stimulation.

Helicase of CiRIG-I Significantly Elicits Signaling Cascade Post PGN Stimulation and CARDs Moderately Strengthen the Function
For further clarifying on the mechanisms of CiRIG-I and its variants in immune system, the mRNA expressions of CiIPS-1 (Fig. 8A), CiIFN-I (Fig. 8B) and CiMx2 (Fig. 8C)  As for the expression profiles of CiIFN-I (Fig. 8B), the relative quantities in cells transfected with pRIG-I were up-regulated, reached 3.51 folds at 2 h, 4.47 folds at 24 h and 5.53 folds at 72 h compared to the control. The relative values in cells transfected with pDCARDs were significantly increased at 2 h (23.32 folds) and rapidly declined at 24 h (2.75 folds), then recovered the control level at 72 h. The relative folds in cells transfected with pDRD were increased at 2 h (5.98 folds) and 24 h (9.38 folds), and significantly enhanced at 72 h (55.09 folds). The relative quantities in cells transfected with pCARDs-RD were no significant differences at 2 h, and slightly increased at 24 h (2.21 folds) and 72 h (2.00 folds). Similar to the transfected pCARDs-RD cells, the relative values in cells transfected with pCARDs were no significant differences at 2 h, and enhanced at 24 h (2.44 folds) and 72 h (2.35 folds). The relative expressions in cells transfected with pRD were significantly increased at 2 h (5.92 folds), and recovered the control level at 24 h and 72 h.
Next, the temporary expression patterns of CiMx2 (Fig. 8C)   transfected with pCARDs were slightly enhanced at 24 h (2.50 folds) and 72 h (3.17 folds). Finally, the relative quantities in cells transfected with pRD began to increase significantly at 2 h (6.83 folds), and declined to 3.37 folds at 24 h, then recovered at 72 h (0.49 fold).
These results showed helicase played a key function on RIG-I signaling cascade and CARDs act as a positive role, whereas RD inhibited the signaling transduction. Additionally, CARDs and RD alone had few effects on PGN stimulation.

Discussion
RIG-I gene is discovered by a group in Shanghai Institute of Hematology in China in 1997. Until recently, RIG-I is demonstrated as an essential cytoplasmic component of the sensor for the recognition of intracellular dsRNA analogues (such as poly(I:C)) and various RNA/DNA viruses [13,35,36]. In mammals, Nterminal two CARDs are responsible for interacting with downstream signaling pathways that mediate dsRNA inducing IFN-I production [37]. The central helicase domain is responsible for dsRNA recognition and binding, which leads to the dimerization and structural alterations of RIG-I. Subsequently, RIG-I exposes the CARDs to interact with IPS-1. The C-terminal RD is tightly auto-regulated, which is mediated by intramolecular interactions with the CARDs. The RD inhibits RIG-I signaling in the resting state, however, recent report indicates that RD is responsible for RNA binding during virus infection [22].
Antiviral activity assay showed that apoptosis in overexpression cells of CiRIG-I and its domains was delayed more or less. Cells transfected with pDRD and pRIG-I exhibited a powerful antiviral effect against GCRV (Fig. 3A and Fig. 3B). The relative expressions of downstream genes in cells transfected with pDRD were up-regulated, and exhibited high levels especially at 48 h including 3.96 folds in CiIPS-1, 16.67 folds in CiIFN-I and 23.57 folds in CiMx2. The cells transfected with pRIG-I and pCARDs also showed a strong mRNA expression of CiIPS-1, CiIFN-I and CiMx2 (Fig. 4). In EPC (Epithelioma papulosum cyprini) cells, overexpression of the full length of RIG-I or CARDs induces the powerful antiviral activity [24]. Similarly, the innate immune response to viral infection in human cells is modified by a functional polymorphism in the CARDs of RIG-I, and CARDs alone can activate the signaling cascade leading to IFN-I expression in mammal [35,[38][39][40]. Collectively, it implied that CARDs of RIG-I were important for antiviral signaling activation in both mammals and teleosts. The cells overexpressing CiRIG-I induced a strong antiviral signaling function and transfected pCARDs-RD cells exhibited little regulation post GCRV infection (Fig. 4), which indicated that helicase of CiRIG-I played an important role in antiviral immune response in grass carp. Mutation of the ATP-binding site (K270A) inactivates RIG-I to trigger antiviral signaling, and the helicase is evidenced to elicit binding of ATPase and RIG-I signaling pathway in mammals [15,35]. In contrast, the functional characterization about helicase of RIG-I was not researched in teleosts. Our study provided the evidence that helicase of CiRIG-I also played an important role in antiviral signaling channel, which was in consistent with the mammals. Taken together, these results demonstrated CARDs and helicase motifs were crucial for antiviral activity, whereas RD inhibited the activation of signaling cascade in teleosts.
The flag-RIG-I constructs encoding helicase and RD form an inhibitory complex to prevent RIG-I signaling pathway in response to Sendai virus in mammals [15], furthermore, RIG-I-DCARDs act as a dominant-negative inhibitor of RIG-I to bind viral RNA [41]. Unexpectedly, the relative values of CiIPS-1, CiIFN-I and CiMx2 mRNA expressions were increased and GCRV quantities were dramatically decreased in transfected pDCARDs cells (Fig. 4 and Fig. 5). Furthermore, the antiviral activity also showed that DCARDs of CiRIG-I have ability to delay the cytopathic effect (CPE) (Fig. 3A and Fig. 3B). These data demonstrated that DCARDs of CiRIG-I were resistant to GCRV replication indeed. It is supposed that there are other mechanisms which elicit an antiviral immune in response to inhibit GCRV replication in teleosts. One possible hypothesis is that the deletion of CARDs may play a role similar to the LGP2. Firstly, LGP2 is induced by GCRV infection in grass carp [42]. Secondly, it is not alone but with a good prototype in Japanese flounder, which demonstrates that mRNA levels of IFN-I and ISGs (Mx and ISG15) in the hirame natural embryo cells overexpressing LGP2 are increased by viral infection [43]. In addition, overexpressing LGP2 of transfected cells acts as a positive regulator for IFN-I production in rainbow trout [44]. The accurate mechanism that DCARDs of CiRIG-I can elicit an antiviral immune effect to inhibit virus replication needs to be clarified in future.
In accordance with the results of GCRV infection, the cells transfected with pRIG-I and pDRD showed powerful mRNA expressions of CiIPS-1, CiIFN-I and CiMx2 post poly(I:C) stimulation (Fig. 6). The mRNA expressions of CiIPS-1, CiIFN-I and CiMx2 increased intermediately in pCARDs transfected cells. The data proved that consensus mechanism for poly(I:C) or GCRV was shown in RIG-I signaling pathway. CARDs were crucial for signaling cascade after poly(I:C) stimulation, and the results were uniform in both mammals and teleosts [28,41]. According to mRNA expressions of CiIPS-1, CiIFN-I and CiMx2 in transfected pRD, pDRD and pRIG-I cells, we concluded that RD played a negative role of the signaling channel. The result was in line with that reported by Hausmann et al. [41]. In mammals, DCARDs of RIG-I can not induce IFN-I effectively and the helicase participates in poly(I:C) inducing IFN-I activation [41,45]. In contrast to the mammals, it was not significant for signaling pathways of CiRIG-I helicase, because the relative mRNA expressions of CiIPS-1, CiIFN-I and CiMx2 were nearly the same in pCARDs and pDRD transfected cells. Unexpectedly, the helicase and RD domains (DCARDs) triggered a powerful IPS-1, IFN-I and Mx2 activation in grass carp. The structure analysis exhibits low homology of helicase and RD domains between grass carp and human [29]. The variant structure of RIG-I implies the discrepancy of signaling transduction between mammals and teleosts in the innate immunity [10,29,46]. Collectively, these results proved CARDs were crucial for signaling cascade upon poly(I:C) stimulation, and RD played a negative role in the signaling channel. Additionally, helicase and RD had a cooperative effect upon poly(I:C) stimulation, which made up the imperfect theory for the comprehensive mechanism on dsRNA inducing RIG-I signaling pathway in vertebrate.
In order to define the regulation of RIG-I and its domains to LPS stimulation in grass carp, the temporal expressions of CiIPS-1, CiIFN-I and CiMx2 in transfected cells were examined by qRT-PCR (Fig. 7). Based on the common traits, it revealed that helicase motif was pivotal for RIG-I signaling transduction and CARDs played an assistant role. In mammals, helicase motif was evidenced for enzymatic activity, ATP hydrolyze, interferon induction, and antiviral signaling [21]. The results indicated that helicase showed a formidable signaling cascade in grass carp coupled with mammals. What puzzled us was that why mRNA levels of CiIPS-1, CiIFN-I and CiMx2 sharply decreased at 24 h. To date, LPS is capable of eliciting a wide variety of septic shock, cell injury and lethality [47]. It has long been established that lower vertebrates, most notably fish and amphibians, are resistant to the toxic effect of LPS [48]. So we supposed that LPS just acted as a simple bacterial PAMP at the beginning, as a time-dependent check, the endotoxic effect of LPS was provoked at 24 h, subsequently, some representative domains of CiRIG-I could exhibit higher mRNA expression to inhibit the pathogenetic analogue. Although the mechanism of RIG-I domain to LPS in signaling transduction was unclear, our results provided a foundation for precise regulation in fish. In combination with mRNA expression profiles in pDRD and pRD transfected cells, our findings indicated that RD played a suppressive role slightly. Furthermore, RD exhibited a rapid regulation, whereas CARDs played a sustained role in the signaling activation (Fig. 7). In mammals, overexpression of RIG-I increases the expression of IL-1b, IL-6 and IL-8 in gingival fibroblasts after LPS stimulation [49]. Compared with mammals, the lack of TLR4 ortholog in some fish species leads us to hypothesize that mechanism of LPS recognition in fish may be different from mammals, and zebrafish TLR4 orthologs negatively regulates the MyD88-dependent signaling pathway post LPS stimulation [48]. We also provided the evidence that the signaling pathway of overexpression RIG-I was motivated upon LPS stimulation in grass carp. For accurate domain recognition, it was demonstrated that helicase domain was pivotal for signaling cascade and CARDs played an assistant role after LPS stimulation. Interestingly, CARDs and RD alone had little influence post LPS stimulation. Collectively, these results provide the possibility that CiRIG-I can function as a pivotal PRR for recognizing gram-negative bacteria pathogens in couple with mammals.
To further clarify functional characterizations of CiRIG-I upon PGN stimulation, the mRNA expressions of CiIPS-1, CiIFN-I and CiMx2 were detected by qRT-PCR (Fig. 8). According to the common tendency, it proved that helicase was essential for RIG-I signaling transduction. Interestingly, the relative expression levels in pRD and pDCARDs transfected cells reached the maximum at 2 h, then declined sharply. The phenomenon implied that RD excited a rapid regulation upon PGN stimulation. Relatively, the expression patterns of CiIPS-1, CiIFN-I and CiMx2 were continuously up-regulated in pRIG-I, pDRD and pCARDs transfected cells, which indicated that CARDs played positive and persistent roles in RIG-I signaling cascade. In the present study, the results clearly suggested that the signaling pathway after RIG-I overexpression was activated by PGN stimulation. The data evidenced that gram-positive or gram-negative bacterial analogues could induce RIG-I signaling pathway, in which helicase was crucial for signaling cascade. CARDs played a positive role, whereas RD inhibited the activation. Furthermore, RD exhibited a fast regulation and CARDs showed a continuous function. In addition, stronger mRNA levels of CiIPS-1, CiIFN-I and CiMx2 were observed post PGN stimulation than those after LPS challenge.
In summary, RIG-I exhibits extraordinary broad roles in innate immune responses in teleosts, responding to not only dsRNA virus or synthetic dsRNA but also bacterial PAMPs. The CARDs play a positive function whereas RD shows a repressive role. Helicase domain has pivotal roles in response to dsRNA virus and bacterial PAMPs (LPS and PGN). Interestingly, DCARDs display a positive role in RIG-I signaling cascade after both virus and PAMPs stimulation.

Construction of Overexpression Plasmids
The full-length cDNA sequence of CiRIG-I was reported previously [29]. The full-length coding sequence of CiRIG-I was amplified using LA Taq TM DNA polymerase (TaKaRa, Japan) by primers RF322a and RR323a (Table 1). To generate the corresponding domain vectors (Fig. 1), CiRIG-I-DCARDs (lack of CARDs), CiRIG-I-DRD (RD deleted), CiRIG-I-CARDs-RD (CiRIG-I-CARDs plus CiRIG-I-RD with a communal restriction enzyme site, or helicase domain removed), CiRIG-I-CARDs (CARDs remained) and CiRIG-I-RD (RD preserved) were amplified by PCR. The primer sequences of the variants were listed in Table 1. The corresponding PCR products were ligated into pMD18-T easy vector (TaKaRa), transformed into the competent cells Escherichia coli TOP10, and plated on the LB-agar petri-dish containing ampicillin for selection. Colony PCR was used to screen positive colonies. Three of them were picked up and sent to a commercial company (Nanjing Genscript Biotechnology Co., Ltd, China) for sequencing to validate the insert sequences without mutations. The plasmid with correct insert in pMD18-T easy vector was extracted by TIANpure Midi Plasmid Kit (Beijing TIANGEN Biotech, China) and digested with the enzymes of EcoRI (Fermentas, Canada) and BamHI (Fermentas), meanwhile, the plasmid of pCMV-EGFP-CMV-SV40 (Fig. 2) was digested with the same enzymes [50]. The target fragments were purified, ligated with T4 DNA ligase (Fermentas), and named as pRIG-I, pDCARDs, pDRD, pCARDs-RD, pCARDs and pRD, respectively (the full names were shown in Table 2). As for the construction of pCARDs-RD, the brief procedures as follow: CARDs of CiRIG-I were firstly amplified by primers RF322a with EcoRI and RR480a with SalI (Fermentas) and BamHI (Table 1) and pCMV-EGFP-CMV-SV40 was digested with EcoRI and BamHI, then ligated with T4 DNA ligase. Thus the enzyme cutting site of SalI was introduced. The constructed plasmid above was digested with SalI and BamHI, and RD of CiRIG-I was amplified by primers RF479a and RR323a (Table 1), then they were digested with the same enzymes and ligated with T4 DNA ligase. Finally, the plasmid of pCARDs-RD was obtained and validated for sequencing.

Cells Culture, Plasmids Transfection, Virus Infection and PAMPs Stimulation
CIK cell line, provided by China Center for Type Culture Collection, was grown in DMEM-F12 (Invitrogen, USA) supplemented with 10% fetal bovine serum (FBS; Biosource, USA), 100 IU/ml of penicillin (Sigma, USA) and 100 mg/ml of streptomycin (Sigma) [29]. Cells were incubated at 28uC in a 5% CO 2 humidified atmosphere. CIK cells were transfected in 6-well plates at a density of 225610 6 cells/ml with 0.5 mg of purified plasmids by FuGENEH HD Transfection Reagent (Roche, Switzerland) according to the manufacturer's instruction. After 48 h, cells were cultured in Medium 199 (Sigma) and supplied For virus infection, steadily transfected cells including pCMV, pRIG-I, pDCARDs, pDRD, pCARDs-RD, pCARDs and pRD were cultured in 24-well plates, they were washed and counted using a hemocytometer, then resuspended into a final concentration of 6610 5 cells/ml supplemented with FBS. After 24 h incubation, the cells were washed with phosphate buffer solution (PBS) for three times, and cultured in Medium 199 without FBS. Then they were infected with GCRV (097 strain, 3.63610 7 TCID 50 /ml) at a multiplicity of infection (MOI) of 1 [29]. For the time-dependent expression profiles, cells were harvested at 1000 rpm for 8 minutes at 0, 2, 24 and 48 h post infection. RNA was extracted and reverse transcribed.
For PAMPs stimulation, poly(I:C) (Sigma) was dissolved in PBS, then heated to 55uC for 5 minutes and allowed to cool at room temperature. LPS (Sigma) and PGN (Sigma) were also dissolved in PBS following instructions. In accordance with the GCRV infection, the steadily transfected cells were incubated in 24-well plates, and they were treated with 5 mg/ml (terminal concentration) of poly(I:C), 10 mg/ml (terminal concentration) of LPS and PGN, respectively. For studies on mRNA expressions, cells were gathered at 0, 2, 24 and 72 h post stimulation at 1000 rpm for 8 minutes, then RNA was isolated and transcribed.

Antiviral Activity Assay
The stably transfected cells were seeded into 96-well plates at the density of 4610 5 cells/ml, containing 100 ml/well Medium 199 with FBS overnight, then infected with 2-fold-diluted GCRV at the indicated titers in duplicate (Fig. 3B). After 60 h postinfection, cells were fixed with 10% paraformaldehyde for 10 min at room temperature and stained with 0.05% (wt/vol) crystal violet (Sigma) for 30 min. Washed with water and drained, then the plates were photographed under a light box (Bio-Rad, USA). For virus titration test, stably transfected cells were infected with GCRV, and supernatants were harvested at 12 and 48 h post infection, and used for virus titer assay. The virus titer was tested as previous report [51].

The Temporal Expression Profiles of CiIPS-1, CiIFN-I and CiMx2 after GCRV Challenges
To check the effects of viral infection on CiRIG-I signaling pathways in stably transfected cells, qRT-PCR method was established to quantify the mRNA expressions of CiIPS-1 (accession No., GQ483645), CiIFN-I (accession No., DQ357216) and CiMx2 (accession No., AY395698) post GCRV infection using CFX96 Multicolor Real-time PCR Detection System (Bio-Rad). Three parallel samples from each group were harvested at 0, 2, 24 and 48 h post infection. EF1a was utilized as an internal control for cDNA normalization [52]. The primers of EF1a, CiIPS-1, CiIFN-I and CiMx2 were listed in Table 1 for qRT-PCR. The qRT-PCR and data analysis carried out as previous report [29].

Comparing the Relative Virus Yields between Transfected pDCARDs and pCMV Cells Post GCRV Infection
For further research on the antiviral function of pDCARDs transfected cells, we utilized qRT-PCR to directly quantify the virus yields. The viral expression profiles of VP4 (segment 6 of GCRV, outer capsid protein, accession No., GQ469997) were examined in pDCARDs and pCMV transfected cells at 2, 24 and 48 h post GCRV infection. The procedures were reported in previous study [29]. The GCRV yields in pDCARDs tranfected cells were relative to those in pCMV transfected cells (control). The EF1a gene was used as an internal control. The data analysis was referenced as above description.
The mRNA Expression Patterns of CiIPS-1, CiIFN-I and CiMx2 Post Poly(I:C), LPS or PGN Stimulation To examine the effects of PAMPs (poly(I:C), LPS or PGN) stimulation on RIG-I signaling pathway in stably transfected cells, qRT-PCR was employed to quantify the mRNA expressions of CiIPS-1, CiIFN-I and CiMx2. Three independent cell samples from each group were collected at 0, 2, 24 and 72 h post stimulation. The following protocols and data analyses were referenced as above.