1. Introduction
Interferons (IFNs) are virus inducible cytokines that share antiviral properties. IFNs are classified into three subfamilies, type I, II and III, on the basis of the cognate receptors they interact with and the subsequent immune responses they initiate. They differ also in genomic structure and the cell sources producing them.
In tetrapods, both type I and III families contain multiple members whilst type II IFN is encoded by a single gene. However, in teleost fish, a second type II IFN gene is also present.
Type I and III IFNs trigger specific signalling pathways that lead to activation of innate immune defences against viral infection. In contrast, type II IFNs primarily promote cell mediated immunity.
Significant progress has been made in understanding the fish IFN system in recent years. Both the type I and II IFN system are now known to be pivotal to antiviral defence in both innate and adaptive immunity against viral infection in fish. The IFN ligands, receptors and the signalling pathways have been extensively studied, providing important clues to understanding the evolution of the antiviral defence system in vertebrates. This review attempts to summarise recent findings of the interferon system in teleost fish.
2. Type I interferon family
2.1. Discovery of type I IFNs
Although IFN activity has long been known in fish, it was only in 2003 that the first fish type I IFN was cloned by three independent groups in zebrafish (Danio rerio) (Altmann et al., 2003), Atlantic salmon (Salmo salar) (Robertsen et al., 2003) and pufferfish (Lutfalla et al., 2003).
To date, the IFN genes have been reported in a range of teleost species including catfish (Ictalurus punctatus) (Long et al., 2006, 2004), common carp (Cyprinus carpio) (Kitao et al., 2009), rainbow trout (Oncorhynchus mykiss) (Zou et al., 2007; Purcell et al., 2009; Chang et al., 2009), sea bass (Dicentrarchus labrax) (Casani et al., 2009) and three spined stickleback (Gasterosteus aculeatus) (Zou et al., 2007) (Table 1).
In the genome of elephant shark (Callorhinchus milii), a cartilaginous fish, type I IFN genes are also present (Chang et al., 2009). Like mammals, fish possess multiple copies of the IFN genes which are linked in the genome. The gene copy number varies depending on the species (Table 1). Salmonids and cyprinids appear to have more copies than species in the Acanthopterygian superorder, such as pufferfish and three spined stickleback. In Atlantic salmon, 11 IFN genes have been identified in three bacterial artificial chromosome (BAC) clones and it is likely that more copies will be discovered (Sun et al., 2009).
Eight IFN genes are shown to reside closely to the gene encoding growth hormone that has a structure similar to the alpha helical cytokines. Initial analysis of the zebrafish genome identified two copies of the IFN genes in addition to the reported cDNA sequence and these genes were shown to be tandemly linked in a single chromosome (Altmann et al., 2003; Zou et al., 2007).
A fourth copy was later discovered at a different chromosome (Stein et al., 2007; Aggad et al., 2009). These genes were named as IFN1-3 and IFN1-3 by different research groups (Altmann et al., 2003; Zou et al., 2007; Aggad et al., 2009). It must be noted that two more cytokines named as IFN Φ 5 and IFN Φ 6 are not IFN homologues but the homologues of interleukin (IL)-22 and IL-26 (Stein et al., 2007). Based on phylogenetic tree analysis, the zebrafish type I IFNs are now classified as IFN-a1 for IFN1/IFN Φ 1, IFN-c1 for IFN2/IFN Φ 2 and IFN-c2 for IFN3/IFN Φ 3 (Chang et al., 2009; Sun et al., 2009).
One of the characteristic features of fish type I IFN genes is the unique genomic organisation, consisting of five exons and four introns, which is in stark contrast with the single exon IFN genes in reptiles, birds and mammals (Sun et al., 2009; Robertsen et al., 2006). Introns are also present in the IFN genes of cartilaginous fish, suggesting the type I IFN genes evolved from an introncontaining ancestor Chang et al., 2009).
A recent study has shown that the five exon/four intron organisation is also conserved in amphibian IFN genes (Zou et al., 2007; Qi et al., 2010). Interestingly, all four introns separating the IFN coding region among fish species are phase 0 introns. Several reports have suggested that there may exist more introns in the 5 untranslated region in salmonids and cyprinids (Robertsen et al., 2003; Long et al., 2004; Purcell et al., 2009).
The presence of multiple AT rich mRNA instability motifs (ATTTA) in the 3' untranslated region (UTR) reflects the highly inducible nature of these antiviral cytokines. Based on the cysteine patterns in the mature peptide, type I IFNs from vertebrates can be divided into three groups (Zou et al., 2007; Sun et al., 2009; Aggad et al., 2009).
The 4 cysteine (4C) containing group is present in all vertebrate orders from cartilaginous fish to mammals. The two 2 cysteine (2C) containing groups include the teleost specific 2C group with C1 and C3 forming a putative disulphide bond and the mammalian 2C group (eg IFN- β and ε ) with C2 and C4 forming a disulphide bond (Zou et al., 2007).
The two groups of fish IFNs are referred to as group I type I IFNs containing two cysteines (C1 and C3) and group II type I IFNs containing 4 cysteines. To date, group II type I IFNs have been found only in salmonids and cyprinids (Table 1). Phylogenetic analyses reveal two further sub groups within each of the salmonid groups, giving rise to four subgroups, namely IFN-a, b, c and d (Chang et al., 2009; Sun et al., 2009). The recently identified zebrafish zfIFN Φ 4 forms a clade with the IFN-d subgroup and hence we term IFN-d1 (unpublished analysis).
2.2. Transcript variants of type I IFNs
In general, most of the fish IFN proteins deduced from the nucleotide sequences consist of a putative signal peptide and a mature region, indicating that they are secreted via a classic protein secretion pathway. Strikingly, in salmonids and cyprinids, IFN mRNA transcripts coding for variants lacking a signal peptide have been reported (Robertsen et al., 2003; Long et al., 2006; Purcell et al., 2009).
These variants are transcribed from a single gene that generates a transcript encoding the protein with a signal peptide. The mechanism by which these transcript variants are produced is not fully understood. Analysis of the promoter region suggests that the variants, referred to as the long and short transcripts, result from alternative transcription that is under the control of two distinct promoters (Purcell et al., 2009; Bergan et al., 2006). These mRNA alternative spliced variants may have a role in diversifying IFN functions. It is the long transcripts found in Atlantic salmon, rainbow trout and catfish, that encode a putative peptide lacking a signal peptide, but whether they are pseudotranscripts or translate into functional proteins is currently unclear. Constitutive expression of both long and short transcripts can be detected in most tissues and are induced after viral infection, suggesting the transcript variants may play a role in antiviral defence (Purcell et al., 2009).
2.3. Transcriptional regulation of type I IFN expression
Expression studies have shown different patterns between the group I and II type I IFNs and among themembers within the groups, suggesting that they are regulated differentially.
In rainbow trout, the group I type I IFNs appear to be produced widely in cell types and tissues after stimulation with ligands for pattern recognition receptors (PRRs) whilst group II IFN genes are expressed in primary leucocytes derived from the head kidney but not in fibroblast like RTG-2 cells and monocyte/macrophage like RTS-11 cells (Zou et al., 2007).
In Atlantic salmon, expression of group II type I IFNs were primarily expressed in freshly isolated leucocytes and one subtype of group II type I IFNs (IFN-b) were up-regulated by a TLR7 ligand but weakly by a TLR3 ligand, polyI:C (Sun et al., 2009). Such data demonstrate a complex spacial and temporal regulation of IFN production in response to viral infection and various stimuli that may differentially regulate the type I IFN gene promoters through transcription factors. To date, the promoters of the group I IFN genes have been studied in Altantic salmon and zebrafish (Bergan et al., 2006; Chen et al., 2005).
Bioinformatics analysis of a 700 bp fragment of the Atlantic salmon IFN-a1 and -a2 predicted dual putative promoters (named PR-I and PR-II) which are suggested to drive transcription of two distinct transcripts, the so-called short and long transcripts encoding proteins with or without a signal peptide (Purcell et al., 2009; Bergan et al., 2006). The PR-I contains an atypical TATA box (CATAA), two putative interferon regulatory factor (IRF) binding sites, and one nuclear factor kappa B (NF- κ B) site, whilst PR-II lacks a TATA box and contains at least three IRF binding sites and an ATF2/c-Jun binding site. Conversely, the rainbow trout IFN5 promoter contains a predicted TATA box and multiple transcription factors for IRF, STAT, NF- κ B and BLIMP (Chang et al., 2009).
The Zebrafish IFN-a1 (IFN1/IFNФ1) promoter harbours a typical TATA box (Chen et al., 2005). However, it must be noted that the initiation start for IFN transcription has not been determined to date in such studies. Interestingly, putative IRF binding sites are conserved within a short distance of the TATA box in zebrafish, rainbow trout, chicken and humans, suggesting that IFN expression is tightly controlled by IRFs throughout the vertebrates (Chang et al., 2009; Chen et al., 2005). Furthermore, binding motifs for other transcription factors such as Cdx, Sry, HNF and HSF are predicted in the promoter region of the zebrafish IFN-a1 (IFN1/IFNФ1) gene and polyI:C responsive elements have been confirmed to be present within a 2.2 kb region using GFP as a reporter (Chen et al., 2005). In this study, the region between −1 and −202 bp has also been confirmed to be responsive to polyI:C stimulation in a luciferase promoter reporter assay.
2.4. Type I IFN receptors
Despite their high sequence divergence, mammalian type I IFNs all mediate antiviral responses by binding to a common ternary receptor complex containing two chains of IFNAR1 and IFNAR2 (Samuel, 2001; Li et al., 2008; Jaks et al., 2007). The IFNAR1 and IFNAR2 belong to the class II cytokine receptor family (CRFB) that lacks a WSXWS motif that is conserved in the extracellular region of the class I cytokine receptors.
The CRFB family also comprises receptors for IL-10 related α -helical cytokines. IFNAR2 has much higher affinity to bind IFN ligands than IFNAR1 and it is suggested that the signalling pathways leading to differential activities are mainly controlled by IFNAR1 (Jaks et al., 2007).
The genes encoding IFNAR1 and IFNAR2 are clustered with the IL-10R2 and IFN- R2 genes in the human genome (Krause and Pestka, 2005; Langer et al., 2004). Type III IFNs activate a different receptor complex containing IL-28R1 and IL-10R2, the latter a signalling chain shared by several IL-10 related cytokines such as IL-10, IL-22 and IL-26 (Langer et al., 2004; Kotenko et al., 2003; Sheppard et al., 2003).
The CRFB family has been identified by extensive survey of the genomes of zebrafish and pufferfish (Lutfalla et al., 2003; Stein et al., 2007). In general, fish have more CRFB members than humans, with 17 members known in teleost fish (Lutfalla et al., 2003; Aggad et al., 2010).
Most of the pairing relationships with their cytokine ligands have not been determined in fish, except for the receptors for IFNs (Aggad et al., 2009, 2010; Gao et al., 2009; Grayfer and Belosevic, 2009). The chromosomal loci equivalent to that containing IFNAR2, IL-10R2, IFNAR1 and IFN- R2 in humans was first located in the tetraodon genome, and contains six related receptors (named CRFB1-6) (Lutfalla et al., 2003).
This locus has also been confirmed in the genomes of Fugu (Takifugu rubripes), suggesting it is likely to be conserved among teleosts (Stein et al., 2007). However, CRFB2 has not been found in Fugu and CRFB3 has apparently been lost in zebrafish. Unlike in tetraodon and Fugu, CRFB6 in zebrafish is located in chromosome 5 that is different from that (Chromosome 5) containing CRFB1, 2, and 4( zebrafish genome version ZV9). In addition, a CRFB15 member similar to CRFB4 has been identified in zebrafish whilst its genome location is still unknown (Stein et al., 2007). Teleost group I and II type I IFNs appear to bind to two distinct receptor complexes (Table 2).
Using a gain-and loss-of function approach, zebrafish CRFB1 and CRFB5 have been shown to be responsible for the antiviral functions elicited by IFNF1 and F4, both of which are group I type I IFNs containing 2 cysteines. Conversely, CRFB2 and CRFB5 are required for the antiviral activity of IFNF2 and IFNF3 that belong to the 4 cysteine containing group II type I IFNs (Aggad et al., 2009; Levraud et al., 2007). Based on the data from gene synteny analysis and functional studies, it is possible that zebrafish CRFB1 and 2 with specificity to the two groups of IFNs might be the equivalent homologues to the human IFNAR2 that is the high affinity binding chain, whilst CRFB5 could serve as the low affinity binding chain that primarily controls the differential activities of type I IFNs (Fig. 1).
Fig. 1. Interaction between IFNs and receptors in teleost fish. The two groups of type I IFNs, group I containing 2 conserved cysteines and group II containing 4 conserved cysteines, bind to receptor complexes with a common low affinity chain (IFNAR1/CRFB5) but two distinct high affinity chains, IFNAR2-1/CRFB1 and IFNAR2-2/CRFB2. The tetrametric receptor complexes for the two members of type II IFNs share a common low affinity binding chain (IFN- R2/CRFB6) but contain distinct high affinity receptors consisting of either two receptor homodimers or heterodimers, IFN- R1-2/CRFB13, IFN- R1-1/CRFB17 or IFN- R1-2/CRFB13 plus IFN- R1-1/CRFB17. In the tetraodon and Fugu genome, the genes encoding IFNAR1-1/CRFB1, IFNAR1-2/CRFB2, IFNAR1/CRFB5 and IFN- R2/CRFB6 are located in the same chromosome whilst in the zebrafish genome the former three are in chromosome 9 and the IFN- R2/CRFB6 gene is in chromosome 5.
In fact, this type of receptor usage is not unique since it is known that members of the same cytokine family with different functions interact with specific receptors for binding but share a common receptor for signalling (Langer et al., 2004). For example, IL-10R2 is a common signalling receptor for IL-10, 22, 26 and type III IFNs. Future experiments will need to determine the binding affinity between the IFN ligands and their cognate receptors.
2.5. Bioactivity of type I IFNs
The bioactivities of fish type I IFNs have been well documented in teleost species (Altmann et al., 2003; Robertsen et al., 2003; Long et al., 2004; Zou et al., 2007; Aggad et al., 2009; Levraud et al., 2007; Ooi et al., 2008; Lopez-Munoz et al., 2009; Martin et al., 2007a; Rokenes et al., 2007). To date, most studies have focused on teleost group I IFNs. Fish group I type I IFNs induce expression of a wide range of the IFN stimulated genes (ISG) including Mx, viperin, ISG15, PKR, leading to an enhanced antiviral state. Conditioned media from cells transfected with IFN expression plasmids have been shown to elevate expression of Mx, a member of the GPTase family with antiviral functions (Altmann et al., 2003; Robertsen et al., 2003; Berg et al., 2009).
Likewise, recombinant IFN proteins purified from bacteria have similar inducible effects upon the Mx gene, indicating glycosylation is not required for bioactivity (Zou et al., 2007; Ooi et al., 2008; Berg et al., 2009). Cross-species activity has been noted in salmonids, notably between Atlantic salmon and rainbow trout (Martin et al., 2007a; Ooi et al., 2008). A recent microarray analysis of the global gene expression in salmon SHK cells treated with trout IFN has revealed a global increase in many immune genes including antiviral factors, immunoglobulins, cytokines and chemokines (Martin et al., 2007a).
Fish IFNs have a positive feedback on their own expression and other members of the IFN family, possibly through activation of IRF3 and IRF7 (Chang et al., 2009; Ooi et al., 2008; Holland et al., 2008). Stimulation of RTG-2 cells with type I IFN for 6 h leads to induction of both IRF3 and IRF7 expression at the transcript level (Holland et al., 2008). Immune functions of group I type I IFNs have also been studied in vivo. In most cases, recombinant IFNs or IFN expression plasmids were injected into adult fish, larvae and embryos, to investigate their roles in antiviral defence (Aggad et al., 2009; Levraud et al., 2007; Lopez-Munoz et al., 2009, 2010). Delivery of recombinant IFNs by the oral and immersion route is ineffective and fails to induce Mx expression (Ooi et al., 2008).
Injection of recombinant IFN-a2 into adult salmon at a dose of 1 μ g/g body weight led to a marked increase of gene expression of key antiviral factors and MHC-I which is required for antigen-specific cytotoxic killing of infected cells. Direct injection of recombinant IFN-a1 (IFN1/IFNФ1) into zebrafish embryos at 0 or 6 h prior to challenge with infectious spleen and kidney necrosis virus (ISKNV) significantly increased the embryo survival rate (Li et al., 2010). Indeed, approximately 90% of fish larvae injected with as little as 100 pg of recombinant IFN1 at 54 h post-fertilisation survived to 72 h after IHNV infection compared with 100% mortality of the control group. Interestingly, another member of the group I type I IFNs in zebrafish, zfIFN4, failed to provide any protection against IHNV (Aggad et al., 2009).
In vivo administration of a recombinant group I type I IFN protein stimulated MHC-I expression in grass carp and enhanced the cytotoxic killing activity of CD8+ T cells to eradicate infected cells (Chen et al., 2010). However, zebrafish IFN-a1 (IFN1/IFNФ1) conferred strong protection against Streptococcus iniae infection although the exact mechanism by which IFN exerts antibacterial activity is unclear (Lopez-Munoz et al., 2009). Emerging evidence indicates that group II type I IFNs possess antiviral activities, similar to that of group I type I IFNs and thus complementing their actions, although it is possible they may have other functions (see below). Like group I type I IFNs, recombinant trout IFN3, a group II type I IFN, is able to induce Mx gene expression in RTG-2 cells (Zou et al., 2007). Overexpression of zebrafish IFN-a1 (IFN1/IFNФ1) (group I), 2 and 3 (group II) in zebrafish embryos led to remarkable elevation of viperin expression and effective protection against IHNV infection (Aggad et al., 2009; Levraud et al., 2007). The extent of protection displayed by the zebrafish IFN2 appeared to be weaker than that elicited by IFN1. In a more recent study, a rapid and transient induction of expression of antiviral genes by zebrafish group II type I IFNs was detected but a delayed response was apparent in response to group I type I IFN, suggesting the two IFN groups may have complementary antiviral roles in different cells and at different stages of infection (Lopez-Munoz et al., 2009). Members of the group II IFN family may also differ functionally. For example, zebrafish IFN2 but not IFN3 induces IL-1 expression (Lopez-Munoz et al., 2009).
3. Type II interferon family
3.1. Discovery of type II IFNs
In mammals, type II IFN, also named IFN gamma (IFN- γ ), is encoded by a single copy gene which contains four exons and three introns. It is a typical Th1 cytokine produced primarily by CD4+ Th1 cells and NK cells and is instrumental in promoting Th1 responses and activating macrophages (Boehm et al., 1997).
Th1 immunity is essential for host immune defence against intracellular pathogens such as viruses and intracellular bacteria. Type II IFN also induces apoptosis, especially during viral infection, and inhibits cell proliferation. The IFN- γ s in mammals and birds show low sequence homology and thus do not show cross-species bioactivity. The first fish IFN- γ gene was identified by gene synteny analysis of the Fugu genome using the protein sequences of human dualspecificity tyrosine phosphorylation-regulated kinase 2 and mouse nuclear protein double minute 1 (Zou et al., 2004).
The Fugu IFN- γ locus contains two other cytokine genes, IL-22 and IL-26. Later, Igawa et al. (2006) discovered a gene that neighbours the previously identified gene in both the zebrafish and Fugu genome and named it IFN- γ 1 and the previously identified one, believed to be equivalent to mammalian IFN- γ , as IFN- γ 2 (Igawa et al., 2006). This has caused considerable confusion, particularly for the nomenclature of the receptors that they bind. Although the so called IFN- γ 1 has moderate sequence similarity with fish IFN- γ , it has little sequence homology with IFN- γ s in higher vertebrates. Therefore, it is perhaps more appropriate to name it as an IFN related (IFN-rel) gene. In fact, the authors who discovered this gene have now adopted the term “IFN- γ related molecule” for the so-called IFN- γ 1 in teleosts (Savan et al., 2009), in line with other recent studies (Chang et al., 2009). The IFN- molecules have now been sequenced from Atlantic cod (Gadus morhua) (Furnes et al., 2009), Atlantic salmon (Robertsen, 2006), catfish (Milev-Milovanovic et al., 2006), common carp (Stolte et al., 2008), goldfish (Carassius auratus) (Grayfer and Belosevic, 2009), and rainbow trout (Zou et al., 2005) (Fig. 2). Despite sequence divergence, the genomic organisation of teleost IFN- γ genes is well conserved. Moreover, the size of each exon is remarkably comparable among different species (Igawa et al., 2006; Zou et al., 2005). The trout IFN- γ gene is highly polymorphic due to the presence of microsatellite repeats in the first intron (Zou et al., 2005). Two duplicated copies of IFN-s are found in Atlantic salmon, common carp, catfish and rainbow trout (Table 1), which are common for other cytokines in salmonids, resulting in overlapping functions (Zou et al., 2002; Pleguezuelos et al., 2000; Morrison et al., 2007). The IFN- γ rel molecule appears to be a teleost specific gene. It is not a clear homologue of mammalian IFN- but is rather a gene that is believed to have duplicated from the IFN- γ gene (Igawa et al., 2006; Savan et al., 2009). Therefore, it is generally accepted as a second member of the type II IFN family (Fig. 2). To date, the IFN- γ rel gene has been identified mostly from the cyprinid family including catfish (Milev-Milovanovic et al., 2006), common carp (Stolte et al., 2008), goldfish (Grayfer and Belosevic, 2009), grass carp (Chen et al., 2010) and zebrafish (Igawa et al., 2006), and the pufferfish (Igawa et al., 2006). Remarkably, it lacks a C terminal nuclear translocation signal (NLS) motif consisting of four consecutive cationic residues of lysines and arginines, which is conserved in the IFN- γ molecules. The NLS motif is shown to be essential for biological activities of rainbow trout IFN- γ since deletion of the NLS site abolished the ability of IFN- γ to induce expression of an IFN- γ induced chemokine (Zou et al., 2005).
3.2. Type II IFN receptors
IFN- γ activates cellular responses through a tetrameric receptor complex consisting of two receptors, one primarily for ligand binding (referred to as IFN- R1) and the other for signalling (referred to as IFN- γ R2) (Boehm et al., 1997; Pestka et al., 1997). Upon activation by the IFN- γ homodimer, the IFN- γ R1 and IFN- γ R2 recruit Janus Kinase (Jak) 1 and Jak2 to the cytoplasmic region respectively, leading to phosphosphorylation and activation of signal transducers and activators of transcription (Stat) 1.
The phosphorylated Stat1 homodimers migrate into the nucleus and bind to the gamma IFN activation sites (GAS) in the promoter region of IFN- γ responsive genes to initiate gene expression (Darnell et al., 1994). By bioinformatics analysis, the presence of the IFN- γ R1 has been predicted in the genome of zebrafish, pufferfish and three spined stickleback (Table 2). The IFN- γ R1 gene loci in these fish species display a good level of gene synteny compared with the corresponding locus in mammals (Lutfalla et al., 2003; Gao et al., 2009; Grayfer and Belosevic, 2009).
The IFN- γ R1 (named IFN- γ R1- 2 in goldfish and CRFB13 in zebrafish) is located between the oligo3 gene and a gene encoding the IL-22 binding protein (named CRFB9 in zebrafish) in both zebrafish and three spined stickleback. The fish IFN- γ R1 protein consists of a signal peptide, an extracellular region, a transmembrane region and a cytoplasmic region. A signature motif (PxxxL) similar to the binding site for the human Jak1 (LPKSL) is conserved in the cytoplasmic region of fish IFN- γ R1. Furthermore, the Stat1 binding site (YDxxPK(H)) is also present near the C terminus. Direct interaction between the fish IFN- γ and IFN- γ R1 proteins has been confirmed by both immunoprecipitation and pulldown approaches (Gao et al., 2009; Grayfer and Belosevic, 2009). In goldfish, two IFN- R1 isoforms have been characterised; each specifically binds to one of the type II IFN molecules (Grayfer and Belosevic, 2009) (Fig. 1). Functionally, knock-down expression of these receptors in zebrafish embryos results in inhibition of IFN- induced gene expression (Aggad et al., 2010). In addition to IFN- R1/IFN- γ R1-2/CRFB13, a second binding receptor specific for type II IFNs exists in goldfish and zebrafish, named IFN- γ R1-1 and CRFB17, respectively (Aggad et al., 2010; Grayfer and Belosevic, 2009) (Fig. 1). Goldfish IFN- R1-1 preferentially binds to IFN-rel (IFN- γ 1). However, temporary inhibition of zebrafish CRFB17 expression in embryos mitigates IFN- γ (IFN- γ 2) induced IFN-regulated GTPase F1 (IRGF1) expression, confirming it is involved in formation of the receptor complex required for IFN- signalling. The signalling pathways of the IFN-rel molecule have not been characterised although one of the two isoforms of Jak proteins, Jak2a, is suggested to be involved in IFN- γ cellular signalling (Aggad et al., 2010). The signalling chain of IFN- γ , IFN- γ R2, is situated within a locus harbouring three other class II cytokine receptors, namely IFNAR2, IL-10R2 and IFNAR1, in chromosome 21 in the human genome (Soh et al., 1994). Compared with the binding chain, IFN- γ R2 comprises a much shorter cytoplasmic tail containing a well conserved PxxP motif for recruiting the adaptor protein Jak2. In the Fugu and tetraodon genome, the IFN- γ R2 (also referred to as CRFB6) resides in the same chromosome with other CRFB members including the characterised receptors for type I IFNs (eg CRFB1, CRFB2 and CRFB5) (Lutfalla et al., 2003; Stein et al., 2007). This receptor locus is also identifiable in the stickleback genome although the gene copy numbers for the cytokine receptors differ (Gao et al., 2009). However, zebrafish IFN- γ R2 (CRFB6) is located in a different chromosome (chrosome 6, zebrafish genome version ZV9) to that containing CRB1, 2 and 5 (chromosome 9, zebrafish genome version ZV9). The type II IFN receptors are widely expressed in tissues in rainbow trout and zebrafish (Aggad et al., 2010; Gao et al., 2009). Not surprisingly, basal expression of the putative high affinity binding receptor IFN- R1 is significantly higher in macrophages than other cell types (Gao et al., 2009). In contrast to IFN- γ R1, the constitutive transcript level of the IFN- γ R2 appears to be much lower. However, IFN- γ R2 but not IFN- γ R1 is inducible by the ligand, suggesting IFN- γ -triggered signalling is primarily mediated by the signalling receptor in fish (Gao et al., 2009).
3.3. Bioactivity of type II IFNs
Several fish IFN- γ s have been produced in bacteria as recombinant proteins and the bioactivities tested. When treated with IFN- γ , trout macrophages exhibit a remarkable increase of the MHC-I and MHC-II molecules at both the transcript and protein level, suggesting it has an important role in enhancing antigen presentation (Martin et al., 2007a,b,c; Zou et al., 2005). In addition, key components forming the proteasome that degrades pathogenic proteins for antigen presentation via MHC-I pathway, including LMP-2, LMP-7, MECL-1 and tapasin, are significantly elevated in cells exposed to trout IFN- γ (Martin et al., 2007a). Some of these genes have been shown to contain multiple GAS (TTNCNNNAA) and interferon stimulated response element (IRSE) (GAAAN(N)GAAA) motifs in their promoters, the primary targets for the phosphorylated transcription factors activated by IFN (Castro et al., 2008, 2010; Dijkstra et al., 2003; Jorgensen et al., 2007) (Fig. 3). One of these transcription factors, Stat1, has been shown to be rapidly phosphorylated after stimulation with IFN- γ and subsequently migrates into the nucleus (Skjesol et al., 2010). Some of the IRFs such as IRF1, 2, 8 and 9 are also upregulated in goldfish monocytes shortly after stimulation with IFN- (Grayfer et al., 2010). Fish IFN- γ strongly enhances the phagocytic and nitric oxide responses of phagocytes (Grayfer and Belosevic, 2009; Zou et al., 2005; Grayfer et al., 2010; Arts et al., 2010). Head kidney leucocytes freshly isolated from trout, common carp and goldfish exhibit elevated respiratory burst activity when incubated with recombinant IFN- γ . Nitric oxide is synthesised by nitric oxide synthase (NOS) in monocytes, macrophages and neutrophils as a free radical to help destroy bacteria. Stimulation of goldfish macrophages with IFN- leads to transcript accumulation of the two inducible NOS genes and an increased level of nitric oxide (Grayfer et al., 2010). Fish IFN- γ induces gene expression of many ISGs that also respond to type I IFNs, suggesting cross-activation of the innate antiviral responses elicited by type I and II IFNs. These ISGs include the antiviral proteinMx (Jorgensen et al., 2007), ISG15 (Martin et al., 2007a), a salmon homologue to the carp hemorrhagic virus (GCHV)- induced gene 2 (Gig2) (Martin et al., 2007a), viperin (Grayfer and Belosevic, 2009) and guanylate-binding protein (Robertsen et al., 2006). However, the potency of IFN- is much weaker compared to type I IFNs and a high dose of IFN- is required to confer virus resistance (Xu et al., 2010). Furthermore, IFN- enhances host surveillance against viruses and some bacteria by up-regulation of pattern recognition receptors such as TLR9 (Skjaeveland et al., 2008). Fish IFN- γ also modulates cytokine and chemokine expression. It is a potent inducer of proinflammatory cytokines such as IL-1, IL-6, IL-12 and tumor necrosis factor (TNF) (Grayfer et al., 2010; Arts et al., 2010). IFN- γ induces expression of a CXC chemokine (- IFN inducible protein, CXCL9/10/11), that primarily acts on T cells, and CC chemokine ligand 1 (CCL1) but interestingly has no effect on IL-8 expression (Grayfer and Belosevic, 2009; Zou et al., 2005; Grayfer et al., 2010; Arts et al., 2010). The functions of IFN- γ rel have not been fully determined. Evidence is starting to emerge that functional differences may exist between the two members of type II IFNs in fish. In terms of expression, common carp IFN- γ rel is predominantly produced by IgM+ cells whilst IFN- γ is detected in IgM-cells, suggesting that fish IFN- γ rel may primarily regulate humoral immunity (Stolte et al., 2008; Arts et al., 2010). Moreover, common carp IFN- γ rel does not have any stimulatory effect on oxygen and nitrogen radical production in head kidney phagocytes, a classical action of IFN- γ in mammals (Arts et al., 2010). Grayfer et al. (2010) recently performed extensive studies of the functional properties of the two goldfish type II IFN proteins in macrophages (Grayfer et al., 2010). Compared with the long lasting effect of IFN- γ on the priming of the production of reactive oxygen intermediates (ROI), the goldfish IFN- γ rel molecule exhibits much shorter-lived effects and subsequently down-regulates ROI production induced by IFN- γ and TNF- α 1 in macrophages. However, IFN- γ rel is more potent than IFN- γ in activation of iNOS gene expression and the nitric oxide response. The profiles of the genes regulated by IFN- γ and IFN- γ rel also differ. Most significantly, only IFN- γ facilitates Stat1 nuclear translocation although both molecules are able to induce phosphorylation, implying different signalling pathways (Grayfer et al., 2010). However, a recent study suggests that the two type II IFNs in zebrafish may have some overlapping functions, since both are able to stimulate the same set of ISGs when overexpressed in embryos (Sieger et al., 2009).