In vertebrates, the interferon (IFN) response, characterized by induction of IFNs and the subsequent establishment of the cellular antiviral state, is the hallmark of antiviral immunity. IFNs are a group of secreted cytokines with activities to inhibit viral replication and regulate the function of immune cells1,2. In mammals, three types of IFNs (type I, II and III IFNs) have been identified, all exhibiting significant antiviral activities3,4. Activation of type I and III IFNs, occurring in various cells in response to viral infection, is considered to be central to the antiviral innate immunity in vertebrates3,5,6.
Among the IRF family, IRF-3 and IRF-7 are essential for the regulated expression of IFNs8,9. In mammals, IRF-3 is constitutively expressed, while IRF7 is low-expressed in most cells and can be strongly induced by type I IFN via the JAK/STAT pathway and thus itself is an IFN-stimulated gene (ISG)12,13.
On infection of virus, host pattern recognition receptors (PRRs) sense viral pathogen-associated molecular patterns (PAMPs) to initiate immune responses.
To date, numerous PRRs that specifically recognize foreign nucleic acids have been identified in mammals, such as Toll-like receptor 3, 8, 9 (TLR3, TLR8, and TLR9), DNA-dependent activator of IRFs (DAI), interferon-gamma-inducible protein 16 (IFI16), RIG-I-like receptors (RLRs) and Leucine-rich repeat flightless-interacting protein 1 (LRRFIP1)14–19. These virus-activated PRRs trigger signaling cascades leading to activation of TANK-binding kinase 1 (TBK1) and inhibitor of NF-kB kinase ε (IKK-ε), which in turn phosphorylate IRF-3 and IRF-720,21.
The phosphorylation mediates the formation of IRF3 homodimers, IRF7 homodimers, or IRF3/IRF7 heterodimers, which translocate into the nucleus to bind the virus responsive element (VRE)/ISRE region within the promoters of IFNs to activate their expression22,23. The secreted IFNs bind to IFN receptors to activate expression of hundreds of ISGs through the JAK/STAT pathway. These processes lead to the activation of the IFN system and determine the establishment of the antiviral state in vertebrate cells.
The origin and evolution of the IFN system have attracted increasing attention in recent years. Since initially discovered in human cells in the 1950s, multiple homologous subgroups of the IFN family have been identified in vertebrates from fish to mammals24. The origin of IFN protein with conserved sequence could be evolutionarily derived from teleosts25,26. The fish IFN genes show similarities with those of mammals and play important role in antiviral immunity27,28. Besides, a total of eleven IRF family members have been identified in fish to date, among which IRF-1, -3, and -7 have been evidenced to play vital roles in IFN responses29–31.
As the IFN homologous gene has not been found in invertebrate genomes so far, it had been thought that the IFN signaling pathway was absent from invertebrates. However, recent studies have suggested that invertebrates possess nucleic acid-induced antiviral immunity, which may be similar to the IFN responses of mammals32–35.
The JAK-STAT pathway as well as many ISG-homologous genes and nucleic acid-recognizing PRRs have also been identified in invertebrates and proved to be essential for the antiviral responses36–39. Moreover, a number of IRF-like genes have been explored in genomes and expressed sequence tag (EST) databases of many invertebrates, covering all principal metazoan groups except Nematoda and Hexapoda40–42. More importantly, the novel identified Vago gene from arthropods, encoding a viral-activated secreted peptide that restricts virus infection in infected and neighboring cells by activating the JAK-STAT pathway, is considered to be an arthropod cytokine similar to vertebrate IFN in function (not in sequence)43,44.
These offer a new insight into the invertebrate antiviral immunity and lead us reconsider the question whether invertebrates have the IFN system. Activation of Vago can be induced independent of an RNAi response through recognition of viral double-stranded RNA (dsRNA) by Dicer-2, a DExD/H-box helicase with similarity to vertebrate RIG-I– like receptors44. A recent study revealed that the TRAF-Rel2 signaling pathway is involved in the activation of Vago after viral infection45. However, the conclusion that Vago is a bona fide functional homolog of interferon needs further support. Moreover, the function of invertebrate IRFs remains largely unknown and the functionally evolutionary relation between invertebrate IRFs and vertebrate IFN pathway is still unclear.
In this study, we identified an IRF-like gene from Pacific white shrimp, Litopenaeus vannamei, which is the first studied IRF gene in crustaceans. We demonstrated that the L. vannamei IRF plays a role in the context of host defense against white spot syndrome virus (WSSV) infection and can activate the IFN response to induce an antiviral state in mammalian cells.
Moving forward, we showed that L. vannamei IRF can directly bind the Vago promoter to regulate its transcription, which suggests that the regulatory mechanism of Vago induction could be similar to that of IFN in vertebrates, strongly supporting the hypothesis that Vago is an IFN-like molecule in invertebrates. Furthermore, these exhibited that the crustacean immunity could have an IRF-Vago-JAK/STAT pathway regulatory axis that shows similarity to the IRF-IFN-JAK/STAT pathway axis of vertebrates, suggesting that invertebrates might possess an IFN system-like antiviral mechanism.
Results
Bioinformatics and expression analysis of L. vannamei IRF. The full length of the L. vannamei IRF transcript is 1416bp, comprising a 202bp 5′-untranslated region (5′-UTR), a 125bp 3′-UTR and a 1089bp ORF encoding a 362 amino acids protein (Fig. S1A). The L. vannamei IRF has a DNA-binding domain in the amino-terminal regions that is homologous to those of vertebrate IRFs (Fig. S1B). The putative DNA-binding domain of L. vannamei IRF exhibits 43%, 43% and 42% similarities to those of fishes Paralichthys olivaceus, Takifugu rubripes, and Epinephelus coioides, respectively, and also shows 24% identity and 45% similarity to human IRF2, suggesting a evolutionary relation between crustacean and vertebrates IRFs.
The mRNA and protein levels of L. vannamei IRF in tissues were analyzed using RT-PCR and western-blot, which provided consistent results. Expression of IRF was detected at high level in hepatopancreas and intestines, weakly in pyloric caecus, but not in other detected tissues (Fig. 1A,B). The results were validated by real-time PCR, which demonstrated that the expression of IRF in intestines, hepatopancreas and pyloric caecus was 18317-, 11141- and 438-fold over that in hemocytes, respectively (Fig. 1C). The expression profile of IRF in hepatopancreas upon immune stimulation was also detected using real-time PCR. After WSSV or poly (I:C) injection, the expression of IRF was significantly up-regulated making a periodical shape of expression curve, suggesting IRF could be activated by virus infection or foreign dsRNA challenge (Fig. 1D,E).
Dimmerization of IRF.
We analyzed the IRF protein in shrimp tissues by western-blot using native-PAGE with β-actin as control (Fig. 2A). The bands corresponding to the IRF dimmers and monomers could be observed in intestines and hepatopancreas, but not in the control hemocytes, which had been shown above not to express IRF. The interaction between L. vannamei IRF molecules was further analyzed by co-immunoprecipitation in S2 cells co-expressing V5-tagged and GFP-tagged IRFs (Fig. 2B).
The result demonstrated that IRF could interact with itself but not the control GFP protein, confirming that the IRF protein could form dimmers. Given that WSSV infection and poly (I:C) challenge can up-regulate the expression of IRF, we investigated their activation effects on IRF functions. The subcellular localization of GFP-tagged IRF was detected using confocal laser scanning microscopy (Fig. 2C).
In PBS mock treated cells, IRF was mainly present in the cytoplasm, while after WSSV or poly (I:C) treatment, IRF mainly located in the nucleus, suggesting WSSV and poly (I:C) could stimulate the translocation of IRF from the cytoplasm into the nucleus. We further detected the formation of IRF dimmers in WSSV or poly (I:C)-treated cells using native-PAGE and western-blot (Fig. 2D). In both the WSSV and poly (I:C) treated groups, following the prolongation of treatment, more IRF dimmers were formed with the highest ratio of dimmer to monomer observed at 24h post treatment, suggesting that WSSV or poly (I:C) stimulation could promote the formation of IRF dimmers.
Transcription factor activity of IRF.
The L. vannamei IRF showed a similarity to vertebrate IRFs and could translocate into the nucleus in response to immune stimulation, indicating L. vannamei IRF could function as a transcription factor. As L. vannamei IRF contains a DNA-binding domain similar to those of mammalian IRFs, we investigated the effect of L. vannamei IRF on a promoter containing ISRE through dual luciferase reporter gene assays (Fig. 3A). The results showed that as the level of the transfected IRF-expressing plasmid increased from 20 to 30 and 50ng, the activity of the promoter was also up-regulated from 14.3- to 24.0- and 50.6-fold compared with the control, respectively. This suggested that L. vannamei IRF could have transcriptional regulation activity on promoters containing mammalian ISRE sequence. As it has been known that promoters of many human type I and III IFNs can be regulated by human IRFs, we investigated the effects of L. vannamei IRF on promoters of human type I and III IFNs (Fig. 3B). The L. vannamei IRF could significantly enhance the activities of the promoters of type I IFNs, IFN-α and –β, by 2.0- and 2.1-fold compared with the control, respectively, whereas it had no effect on the promoters of type I IFN IFN–ω and type III IFNs IL28 and IL29. Mx1 is an interferon-induced GTP-binding protein with antiviral activity46. Using western-blot, we investigated the expression of Mx1 in L. vannamei IRF-expressing cells. After IRF transfection, with the increase of IRF expression, the level of Mx1 was up-regulated, while the internal control tubulin protein remained unchanged (Fig. 3C). This confirmed that L. vannamei IRF could function in mammalian cells and be involved in the IFN response. Tiger frog virus (TFV) is a member of the genus Ranavirus, family Iridoviridae and can infect a wide range of cell lines including mammalian cells47,48. HEK293T cells were transfected with pcDNA-IRF or original pcDNA3.1 plasmid (as control) and then experimentally infected with TFV. We observed that at 24h, 48, and 72h post infection, the cytopathic signs of L. vannamei IRF-expressing cells were obviously milder than the control cells (Fig. 3D, left panel). The titers of the released TFV in the cell supernatants were analyzed using TCID50 method (Fig. 3D, right panel), which demonstrated that compared with the control, the TFV titers in the supernatants of IRF-expressing cells were significantly decreased at each time point. These suggested that as a transcription factor, L. vannamei IRF could trigger an antiviral state in mammalian cells.
Regulation of Vago promoter by IRF.
We analyzed the promoter regions of shrimp Vago genes49, and observed that an ISRE sequence can be predicted in the 5′-UTR of Vago4 but not in the other four Vago isoforms. To verify the transcriptional regulation effect of L. vannamei IRF on Vago promoters, dual luciferase reporter gene assays were performed on Drosophila S2 cells using luciferase-expressing vectors containing promoters of Vago isoforms (Fig. 4A). The results demonstrated that IRF could significantly up-regulate expression of Vago4 and Vago5 but not Vago1-3. We further detected the transcriptional regulation activity of IRF on promoters of Vago4 and Vago5 through improving the level of IRF expression in the dual luciferase reporter gene detection system (Fig. 4B). With the increase of the transfected IRF-expressing vector, the activities of the promoters of Vago4 and Vago5 were also enhanced, confirming that L. vannamei IRF could regulate the expression of Vago4 and Vago5. It has also been reported that in insect Vago could be activated by viral nucleic acids43,44. We introduced poly (I:C) and WSSV treatments into the dual luciferase reporter gene detection system in order to investigate the role of L. vannamei IRF in activation of Vago by foreign nucleic acids. The Vago4 promoter was representatively detected. We observed that both poly (I:C) and WSSV could up-regulate the activity of Vago4 promoter (Fig. 4C,D). Without IRF expression, compared with the untreated control, the activity of Vago4 promoter was enhanced 6-fold after 6h of WSSV treatment and 4-fold after 0.5ug/ mL poly (I:C) treatment. In the 30ng IRF-expressing vector-transfected groups, the activity of Vago4 promoter was up-regulated 19-fold in the untreated cells, 40-fold in the WSSV treated cells, and 48-fold in the poly (I:C) treated cells. Obviously, compared with the IRF nonexpressing groups, overexpression of the L. vannamei IRF could enhance the effects of WSSV and poly (I:C) on the Vago4 promoter. To further investigate the interaction between L. vannamei IRF and the Vago4 promoter, electrophoretic mobility shift assay (EMSA) was performed using purified 6His-tagged IRF protein expressed in E. coli cells together with wild-type or ISRE-mutated (as control) Vago4 promoters (Fig. 4E). IRF could retard the mobility of the wild-type Vago4 promoter DNA, but not the mutant one, and overloading of the mutant promoter DNA by 2- or 5- fold didn’t affect the DNA level in the protein/DNA complex of IRF/wild-type Vago4 promoter, suggesting that IRF could bind the IRSE sequence in Vago4 promoter. Moreover, the anti-IRF or anti-6his antibody was added into the EMSA system to bind the 6his-tagged IRF protein. We observed that these antibodies could further retard the mobility of the wild-type Vago4 promoter DNA to generate super shift bands, confirming the interaction between L. vannamei IRF and Vago4 promoter
Involvement of IRF in antiviral responses.
To investigate the role of L. vannamei IRF in the antiviral response, we knockdown the expression of IRF in living shrimps using RNAi strategy through injection of IRF-specific dsRNA. The knockdown efficiency was detected using RT-PCR and western-blot, which demonstrated that the IRF-specific dsRNA could significantly reduce the mRNA and protein levels of IRF in hepatopancreas and intestine (Fig. 5A,B). Interestingly, the control GFP dsRNA could obviously increase the expression of IRF, confirming that IRF could be activated by foreign nucleic acids. After dsRNA or PBS treatment, shrimps were experimentally infected with WSSV. We detected the expression of WSSV functional genes, including immediate early genes wsv051, wsv069 and wsv249, and structural protein genes VP28, VP26 and VP24 using real-time RT-PCR (Fig. 5C). Expression of all these genes during 24–96hpi was up-regulated in the IRF-dsRNA treated group but down-regulated in the GFP-dsRNA treated group, with the exception of the structural protein genes VP28, VP26 and VP24 in the GFP-dsRNA treated group at 96hpi, which demonstrated higher levels than those in the IRF-dsRNA treated group. Expression of the three immediate early genes wsv051, wsv069 and wsv249 in the IRF-dsRNA, GFP-dsRNA and the control PBS treated groups were all peaked at 48hpi, while for the three structural protein genes, compared with the PBS control group, their expression peaks were advanced in the IRF-dsRNA group and postponed in the GFP-dsRNA group. These may suggest the antiviral effects derived by IRF on immediate early genes and structural protein genes of WSSV could be different, which needs further investigation. We further detected the expression of Vago4 and Vago5 after WSSV infection (Fig. 5D). In the PBS control group, expressions of Vago4 and Vago5 were up-regulated and both peaked at 72hpi with 4.0- and 4.1-fold increase at 72hpi compared with 0hpi, respectively. In contrast, the expressions of Vago4 and Vago5 in the IRF-dsRNA group were significantly inhibited and peaked at 72h with only 2.8- and 3.0-fold increase, respectively, while in the GFP-dsRNA group, they were significantly enhanced with expression peaks advanced to 48hpi and reached 6.8- and 5.6-fold values, respectively. These confirmed that Vago4/5 expression could be induced by heterogeneous nucleic acids and be regulated by IRF. A parallel experiment was also performed to record the mortality rates of IRF-knockdown shrimps after WSSV injection (Fig. 6A). Compared with the PBS control group, the cumulative mortality was significantly increased in the IRF-dsRNA treated group. Interestingly, we also observed that the mortality was decreased in the GFP-dsRNA treated group. Given that GFP-dsRNA could enhance the expression of IRF, this confirmed that IRF could be involved in the nucleic acid-induced antiviral immune responses of shrimp. We further investigated the virus load in shrimp muscle tissues using absolute quantative real-time PCR and observed that the WSSV DNA copies were significantly increased after IRF-dsRNA treatment and decreased after GFP-dsRNA treatment, consistent with the mortality result (Fig. 6B). In addition, to verify the role of Vago4/5 in antiviral responses, we knockdown the expression of Vago 4/5 in shrimps using RNAi strategy. The results demonstrated that suppression of Vago4/5 significantly increased the mortality of shrimps caused by WSSV infection and reduced the WSSV copies in the shrimp tissues, confirming that Vago plays important role in shrimp antiviral immunity (Fig. 6C,D).