The antiviral immune responses were triggered by the innate immune recognition of viral infection. The type I IFNs (IFN-β and IFN-α) are the key cytokines produced upon viral infection and consequently link innate immunity with adaptive immunity. A main antiviral system in mammals is TRIF-dependent TLRs pathway, but the TRIF-independent RIG-I pathway, has also been discovered recently. In this manuscript, our study focuses on the functional characterization of zebrafish TRIF based on the comparison of its sequence and functional evolution from zebrafish to mammals. Our experimental results show that the full length cDNA of zebrafish TRIF cloned by RACE-PCR approach encodes a protein of 556 amino acids. Luciferase reporter assay confirms that zebrafish TRIF is able to induce the IFN promoter as well as activate NF-κB response promoter. The IFN induction function of zebrafish TRIF is abolished when Ala359 is mutated to Pro or His. Laser confocal microscopy shows that zebrafish TRIF is colocalized with a Golgi apparatus marker, implying its unique subcellular localization in Golgi apparatus. In zebrafish, the mRNA expression of molecules participating in RIG-I pathway are much more sensitive and specific to polyinosine-polycytidylic acid induction compared with those in TRIF-dependent antiviral pathway. The TRIF-dependent TLR4 IFN induction signaling appears not to be functional in zebrafish, since IFN expression cannot be up-regulated by LPS. These two striking findings from de novo ligand induction experiments suggest a novel antiviral mechanism in zebrafish.
In immune systems of mammals, the TLRs play a crucial role in recognition of pathogen-associated molecular patterns and are the first lines of defense against bacterial and viral infection (1). Ten TLRs (TLR1–10) and five intracellular adaptors (MyD88, TIR domain-containing adaptor inducing IFN-β (TRIF),4 TIR domain-containing adapter protein (TIRAP), TRAM, and SARM) have been identified in humans (2). TIRAP is specifically involved in TLR2- and TLR4-mediated signaling pathways (3). MyD88 is responsible for inflammatory cytokine production to all TLR ligands, except for the TLR3 ligand (4, 5). TRIF is unique to TLR3 and TLR4 signaling pathways leading to IFN regulatory factors (IRFs)-3/7 activation and IFN-β production (6). IFN-β and IFN-α Type I IFNs are the key cytokines produced upon viral infection and can link innate immunity with adaptive immunity (7). In the TLR3-mediated pathway, TRIF directly binds to TLR3, activates IRF3/7 (8, 9), and promotes IFN-β promoter depending on the TLR3 ligand (poly(I:C)) stimuli (10). In the TLR4-mediated pathway, TRIF induces IFN-β production via the TLR4 signaling in response to LPS. The unique ability of TRIF to induce IFN-β production in TLR4-mediated pathway is attributed to TRAM (11, 12). SARM (13) has been recently identified to negatively regulate TRIF-dependent TLR signaling (14).
In contrast to the critical function of TRIF in IFN-β induction, an antiviral system independent of TRIF and TLRs, the retinoic acid inducible gene I (RIG-I) pathway, is also identified in mammals. The RIG-I pathway is an intracellular antiviral system different from the extracellular TRIF-dependent TLRs antiviral system. RIG-I is a RNA helicase and functions as an intracellular receptor recognizing the intracellular dsRNA (15, 16). The molecule mitochondrial antiviral signaling (MAVS) is used as an adapter in RIG-I pathway (17, 18, 19, 20). The upstream receptor and adaptor are unique for the RIG-I pathway but not for the TRIF-dependent TLRs system. However, these two systems converge at the downstream signaling involving the IRF3/7 activation and IFN-β induction. TLRs and RNA helicases are two parallel ways to trigger antiviral responses (21).
Except for research on the functional mechanism of receptors and adaptors in the signal transduction of innate immunity, many studies have been devoted to find relationships between their intracellular localizations and their immune functions. As reported, TLR3 is localized in the endoplasmic reticulum of unstimulated cells, then moves to dsRNA-containing endosomes in response to dsRNA, and colocalizes with c-Src on endosomes containing dsRNA in the lumen (22). MAVS is localized to the mitochondria when directed by its C-terminal sequence, and mislocalization of MAVS impairs its antiviral function (19). All this experimental evidence indicates to us that the specific intracellular localization is essential for these receptors and adaptors to function properly in signal pathways.
Zebrafish, an ideal model for developmental research, has now emerged as a valuable tool for immunological study. In 2004, two independent teams simultaneously published their identification of TLRs family and TIR domain containing adaptors in the zebrafish genome (23, 24). They reported the presence of TRIF gene in zebrafish, particularly Meijer et al. (24) found an expressed sequence tag (EST) (accession number: AY389465) corresponding to the partial TIR domain of this gene, and their work demonstrated that zebrafish TRIF was not responsive to Mycobacterium infection. However, they were not able to obtain its complete DNA sequence. Due to the vital function of TRIF-dependent pathway in IFN-β induction in mammals, it is of particular interest to characterize this pathway in fish to further understand this important pathway in mammals. The first fish TRIF gene was cloned from channel catfish (25). Then, Sullivan et al. (26) reported the zebrafish TRIF gene sequence and conducted certain functional research. Our current report on the characters of zebrafish TRIF is consistent with Sullivan and coworker’s findings on some aspects. In contrast, our study also presents new experimental results regarding this important gene. We have cloned the full length cDNA sequence of zebrafish TRIF, which have 5′ and 3′ untranslated regions (UTR) and the coding sequence. Our results indicate that zebrafish TRIF is colocalized with a Golgi apparatus marker gene, and is able to induce the promoter of IFN and activate NF-κB response promoter. Furthermore, the IFN expression cannot be up-regulated by LPS in zebrafish, implying that the TRIF-dependent TLR4 IFN induction signal pathway may not function in this species. All these findings suggested a novel antiviral mechanism in zebrafish.
Materials and Methods
HEK293T cells were grown in 90% DMEM (Invitrogen) supplemented with 10% FCS and antibiotics. The NIH3T3 cells were maintained in a culture medium containing 90% RPMI 1640 supplemented with 10% FCS and antibiotics. Both cells were cultured in the atmosphere of 5% CO2 at 37°C. Zebrafish embryo fibroblast-like cells (Zf4; ATCC CRL-2050) (27) were grown in the 1:1 mixture of DMEM medium and F12 medium in the atmosphere of 5% CO2 at 28°C.
Cloning of zebrafish TRIF
The cDNA ready for RACE-PCR was prepared by the BD SMART RACE cDNA Amplification kit (Clontech). TRIF gene specific primers for PCR amplification were designed according to an EST (Gene Bank accession number: AY389465) homologous to its mammalian counterpart. The gene specific primer for 5′-RACE and its nested primer were 5′-TCTGTTGGCGTGTGGAGGCTCATC-3′ and 5′-GGAGGTAATGGATCGGGGTTTATGTC-3′, respectively. The primer for 3′-RACE and the nested primer were 5′-CGTTCTCCGAAGACTTTGCC-3′ and 5′-GTTCGCGTTACTACTGCTCACC-3′, respectively. There was an overlay between target sequences of 5′- and 3′-RACE PCR for latter reconstructing TRIF full length cDNA sequence. The open reading frame of zebrafish TRIF was amplified with the upstream primer 5′-ATGGCAGAAGGTGGAATGAAGCCT-3′ and the downstream primer 5′-CTACGACTCTTCGGCTGAGCTTTTAG-3′ using zebrafish cDNA as amplification templates. The PCR products were cloned into pGEM-T easy vector (Promega) for sequencing with an ABI 3730 DNA sequencer (Applied Biosystems).
Total RNAs from developing embryos and various tissues of adult zebrafish as indicated in Fig. 2 were extracted by Trizol reagent (Invitrogen) according to the manufacturer’s instructions. Traces of genomic contamination were removed by RNase-free DNAse I (Takara Shuzo) treatment, followed by phenol/chloroform extraction and ethanol precipitation. RT reaction was performed with the ReverTraAce cDNA synthesis kit (Toyobo). A pair of primers for PCR was the same as that used for gene cloning into the pGEM-T easy vector. The PCR cycling parameters were initial denaturation at 94°C for 3 min, 35 cycles of denaturation at 94°C for 15 s, annealing at 60°C for 30 s, and extension at 72°C for 2 min, followed by a final extension at 72°C for 10 min. The PCR products were separated by 1% agarose gel electrophoresis and detected by UV imaging system.
Using the pGEM-T easy vector cloned with zebrafish TRIF sequence as the template, PCR was performed with primers listed in Table I for various expression vectors construction. The pcDNA3.0 and pEGFP-N1 vectors inserted with zebrafish TRIF sequence were constructed by cloning the PCR products into the KpnI and BamHI double digested sites of these two vectors, respectively. The construct for the luciferase reporter assay of zebrafish IFN promoter was constructed by cloning 1355 bp of the genomic fragment upstream of the transcription start site of a zebrafish IFN gene (accession number: NM_207640) into SacI and HindIII cloning sites of the pGL3-Basic vector, a luciferase reporter vector. Single point mutation at A359 of TRIF was introduced by primer extension and fusion PCR with fusion primer pairs F1/R1 and F2/R2. Truncated forms of TRIF including the N terminus (1–311 amino acids), the TIR domain (304–460 amino acids), the C terminus (453–566 amino acids), ΔC (1–460 amino acids), and ΔN (304–566 amino acids) were subcloned into pEGFP-N1 vector, respectively. All constructs were verified by DNA sequencing. The NF-κB luciferase vector was kindly provided by Dr. Ping Wang at the University of London (London, U.K.).
|Expression Vector .||Forward Primer .||Reverse Primer .|
|pEGFP-N1-TRIF. N terminus||ACCCCCAAGCTTATGGCAGAAGGTGGAATGAAGC||AGACGCGGATCCCGGAGTGTAGGGTGAAAACTGTTATC|
|pEGFP-N1-TRIF. C terminus||ACCCCCAAGCTTATGCAGGATGCTGTGGAAAAACAAAG||AGACGCGGATCCCGCGACTCTTCGGCTGAGCTTTTAG|
|Expression Vector .||Forward Primer .||Reverse Primer .|
|pEGFP-N1-TRIF. N terminus||ACCCCCAAGCTTATGGCAGAAGGTGGAATGAAGC||AGACGCGGATCCCGGAGTGTAGGGTGAAAACTGTTATC|
|pEGFP-N1-TRIF. C terminus||ACCCCCAAGCTTATGCAGGATGCTGTGGAAAAACAAAG||AGACGCGGATCCCGCGACTCTTCGGCTGAGCTTTTAG|
Quantitative real-time PCR
Gene-specific primers for quantitative real-time PCR were designed to specifically amplify 140–150 bp fragments. Zebrafish β-actin or GADPH primers were used as internal control to normalize the starting quantity of RNA. Each gene was assayed in triplicate for each sample time point. Quantitative real-time PCR was performed on an ABI Prism 7900 Sequence detection system (Applied Biosystems) as two-step RT-PCR using SYBR Premix Ex TaqTM (Takara Shuzo) according to the manufacturer’s instructions. Reactions were performed in a volume of 20 μl containing 10 μl of 2× SYBR Green PCR Mast Mix, 1 μl of 10 μM primers, 8 μl of nuclease water, and 1 μl of cDNA template. Cycling parameters were 94°C for 5 min, followed by 40 cycles of 94°C for 15 s and 60°C for 1 min. The threshold cycles and relative fold inductions were calculated by the ABI PRISIM 7900HD SDS software.
Transient transfection was conducted with the lipofectine2000 (Invitrogen) according to the manufacturer’s instructions. HEK293T cells or ZF4 cells were transfected in 24-well plates with a total of 1-μg plasmid DNA in a serum-free culture medium. After 4–6 h, the medium was replaced by complete medium with 10% FBS and antibiotics. The transfected 293T cells were lysed for luciferase assay or fixed for laser confocal imaging at 24-h posttransfection. ZF4 cells transfected with fluorescent vectors were directly observed under fluorescent microscopy.
Luciferase reporter assay
The analysis on IFN promoter induction and NF-κB response promoter activation were conducted using a luciferase reporter gene. HEK293T cells were transiently transfected with the IFN promoter luciferase reporter plasmid or NF-κB response promoter luciferase reporter plasmid, together with the indicated TRIF expression vectors using lipofectamine2000 reagent (Invitrogen). Luciferase activity of total cell lysates was measured using luciferase reporter assay system (Promega). The β-galactosidase gene was used as an internal control. Luciferase activity was normalized to β-galactosidase activity and expressed as the fold stimulation relative to transfected empty vector. Values were expressed as mean relative stimulations for a representative experiment from three separate experiments with each performed in duplicate.
Laser confocal imaging
HEK293T cells were seeded onto the coverslips (10 mm × 10 mm) in 24-well plates. After allowing the cells to adhere for 24 h, cells were transfected with the pEGFP-N1-TRIF vector or various truncation vectors. At 24 h of posttransfection, cells on the coverslips were washed once with PBS either directly for confocal imaging or fixed in 70% ethanol in PBS for 30 min at 4°C and stained with PI (propidium iodide) for 15 min at 37°C. Imaging of the cells was conducted using Leica laser scanning confocal microscopy. The 293T cells were also cotransfected with pEGFP-N1-TRIF vector and pDsRed-mono-Golgi vector (An expression vector encoding the membrane-anchoring signal peptide of human β 1,4-galactosyltransferase that targets the red fluorescent protein DsRed monomeric to the trans-medial region of the Golgi apparatus; Clontech Laboratories). At 24 h after transfection, laser scanning confocal imaging was conducted as previously described.
TLRs ligand induction
Adult zebrafish were i.p. injected with one of five groups of TLRs ligands, including polyinosine-polycytidylic acid (poly(I:C)) (100 mugg/g), LTA (100 mugg/g), PGN (50 mugg/g), Zymosan (100 mugg/g), and Glucon (50 mugg/g) from Sigma-Aldrich at five time points (2, 6, 18, 24, and 42h). Each group of LPS (10 and 100 mugg/g, respectively; Sigma-Aldrich) was also injected into adult zebrafish at five time points (2, 6, 18, 24, and 48h). The fish were treated with PBS as external control. For each group divided by ligands vs time points, three fish were treated and collected to eliminate individual variations. At each time point, the visceral tissues dissected from these treated zebrafish were mixed and grinded in liquid nitrogen. After all the samples were collected, total RNA was extracted by Trizol reagent (Invitrogen). The extracted total RNA was then treated with RNase-free DNase (Takara Shuzo) and reverse-transcribed into cDNA using ReverTraAce cDNA synthesize kit (Toyobo). Quantitative real-time PCR was conducted following the method mentioned above. Primer sequences and the products sizes were listed in Table II.
|Gene .||Accession Number .||Forward Primer .||Reverse Primer .||Amplicon size(bp) .|
|Gene .||Accession Number .||Forward Primer .||Reverse Primer .||Amplicon size(bp) .|
The real-time PCR primers for zebrafish IL-1β were referenced to Bin Lin et al (45 ).
GenBank accession number
Zebrafish TRIF cDNA sequence was deposited in the National Center for Biotechnology Information database with the accession number of EF204937.
Isolation and expression pattern of zebrafish TRIF gene
We isolated a full-length cDNA of zebrafish TRIF based on an EST (accession number AY389465) by RACE-PCR approach. The cDNA with two poly(A) signals (ATTAAA and AATAAA) in the 3′-UTR encoded a protein of 556 amino acids (Fig. 1,A). The bioinformatics analysis demonstrated that the protein had 32% identity with its human homologue, with a TIR domain close to its C terminus. The evolutionary tree constructed based on the TIR domains of TRIF showed that the zebrafish TRIF was clustered with catfish TRIF, and the evolutionary trend of TIR domains was in accordance with that of the species (Fig. 1 B). The alignment of zebrafish TRIF cDNA sequence with its genomic DNA sequence indicated that there were two exons and one intron with the second exon corresponding to the partial 5′-UTR, the open reading frame, and the entire 3′-UTR.
The alignment of zebrafish TRIF with human TRIF demonstrated that zebrafish TRIF had striking differences from its mammalian homologue in three aspects. First, zebrafish TRIF lacked both N-terminal and the C-terminal proline-rich domains. Second, the proline in box2 (28) conserved in mammalian TRIF, TIRAP, and MyD88 was substituted to alanine at position 359 in zebrafish TRIF (Fig. 1,C). It was reported that the conserved proline in box2 was essential for human TRIF and Myd88 mediated signaling as reported (10, 28). Finally, a NS3/4A cleavage site of the HCV polyprotein was missing in zebrafish TRIF and TLR3 (Fig. 1 D), but human TRIF was cleaved at this site by NS3/4A protease in Hepatitis C virus (HCV) infection (29).
Since TIR domain is a structural basis for signal transduction by TIR domain containing molecules, we analyzed TIR domains of TRIF from human, mouse, monkey, dog, cow, chicken, fugu, zebrafish, and catfish, MyD88 from human and zebrafish, and TIRAP from human (Fig. 1 C). By comparison with MyD88 and TIRAP, TRIF from zebrafish also lacked the conserved sequences for (F/Y) D in box1, RD in box2, and FW in box3 (28), like its human counterpart (10). The similarities between zebrafish and catfish in box1 and box3 were much higher than those between fish and mammals. It was interestingly noticed that the conserved proline, in box2 of TRIF in mammals, fugu and catfish, was replaced by alanine in zebrafish and chicken in the corresponding site, suggesting that the alanine in box2 was not general feature for fish.
RT-PCR and real-time PCR were conducted to analyze the gene expression profiles in different adult tissues and embryonic development stages of zebrafish. The results indicated that PCR products of zebrafish TRIF could be amplified by RT-PCR for embryonic stages from 6 to 48 h (Fig. 2,A) and for adult tissues (Fig. 2,B), which revealed that TRIF was ubiquitously expressed in terms of time and tissue regions. In addition, the results from real-time PCR demonstrated that the liver showed the highest expression among those tissues (Fig. 2 C). For TRIF tissue expression pattern, Oshiumi et al. (10) reported that human TRIF expression was ubiquitous, while Yamamoto et al. (30) indicated that this gene was expressed ubiquitously and showed strong expression in liver. It appeared that the expression pattern of zebrafish TRIF was consistent with the finding by Yamamoto and coworkers (30). It was likely that the tissue expression pattern of this gene was conserved from fish to human in accordance with their functional conservation in these two species. However, components of this signaling pathway did not show this preference based on our previously experimental result. Zebrafish IRF7 showed a ubiquitous expression in most of tissues with low abundance in intestine and heart (data not shown). Quantitative analysis by Phelan et al. (31) demonstrated that zebrafish TLR3, fIRAK-4, and fTRAF6 mRNA showed ubiquitous expression, albeit at varying levels of all three target genes. TLR3 showed the most expression in gill and heart tissue, zfIRAK-4 mRNA expression was highest in the spleen and zfTRAF6 showed maximum expression in the gill tissue (31). Therefore, the liver preferential expression pattern of zebrafish TRIF may reflect a relevance to its functional character in innate immune response, but it appeared that it was not a common mechanism for other related components of this signaling pathway.
TRIF activates IFN promoter and the NF-κB response promoter
Using transfection and luciferase report assay, we demonstrated that zebrafish TRIF was able to activate the promoter of IFN in HEK293T cells in a dose-dependent manner (Fig. 3,A). However, zebrafish TRIF significantly induced the NF-κB response promoter in a different way, by which the NF-κB response was down-regulated after initial up-regulation by 200 ng of TRIF plasmid (Fig. 3,B). The constructs with full length TRIF or two single mutants of A359P and A359H were then cotransfected with IFN promoter luciferase vector into HEK293T cells. The results from these experiments indicated that the full length TRIF could activate IFN promoter and none of the two mutated forms was able to activate IFN promoter (Fig. 4), further substantiating that A359 was vital for TRIF to induce IFN promoter.
TRIF was localized to the Golgi apparatus
The fluorescent imaging of TRIF-GFP fusion protein analyzed under confocal microscopy indicated that the green fluorescence was localized to the unique cytoplasmic site near to the nuclear membrane in HEK293T cells, the site very likely to be the location of Golgi apparatus (Fig. 5, A and B). The same result was observed in NIH3T3 cells (Fig. 5,C). However, the truncated TRIF segments have different localizations. Two truncated segments with only N terminal or TIR domain were both localized to the entire cell. The segment spanning N terminal and TIR domain was localized to the cytoplasm exclusively. Although, the segment spanning TIR domain and C terminus showed the unique star-like subcellular localization as full length did (Fig. 6). To preclude the possibility of an experimental artifact in mammalian cells, ZF4 cells (zebrafish fibroblast cells) were also transfected with GFP-fused zebrafish TRIF expression vectors for fluorescent microscopy analysis. Results confirmed that zebrafish TRIF was exactly localized to the same intracellular site comparable to the location of mammalian cells except with much lower transfection efficiency in ZF4 cells (Fig. 7). To further investigate whether zebrafish TRIF was exactly localized to the Golgi apparatus, cotransfection and laser confocal imaging were conducted. Results demonstrated that TRIF gene was colocalized with the red fluorescent monomeric protein DsRed targeted to the Golgi apparatus, as seen from the yellow fluorescent overlay (Fig. 8). All these results revealed that zebrafish TRIF was a Golgi-localized protein, and the sequence spanning TIR domain and C terminus might contribute to its unique subcellular localization. Unfortunately, in this sequence, there was no specific transmembrane domain or signal peptide that could be predicted as a specific anchor signal for the Golgi apparatus localization of this protein except one with a low theoretical PI of 6.02. This suggested the presence of a potential new anchor sequence for Golgi apparatus localization of this protein.
Up-regulation of anti-viral genes mRNA upon TLRs ligand induction
By real-time PCR, we analyzed the induced mRNA expression of TRIF and TLR3 for TRIF-dependent IFN induction pathway and RIG-I and MAVS for TRIF-independent pathway upon TLRs ligand induction. Our results revealed three important findings. In the TRIF-dependent TLRs pathway, TLR3 had a notable response to poly(I:C) at 2 and 6-h post induction. The response was not very specific, since TLR3 can also be induced by PGN, Zymozan, and Glucon (Fig. 9,A). TRIF nearly had no obvious up-regulation in response to poly(I:C), but was able to be induced by PGN, Zymosan, and Glucon to ∼10-fold up-regulation (Fig. 9,B). In the TRIF-independent pathway, RIG-I expression was found to have a strong and specific response to poly(I:C) at 6, 18, and 42-h post induction, especially at 6-h postinduction (>80-fold induction; Fig. 9,C). MAVS was shown to be specifically up-regulated in response to poly(I:C) at 42-h postinduction (Fig. 9,D). The up-regulations of downstream signal molecules in anti-viral pathway were also analyzed. The mRNA level of IRF-7 was significantly up-regulated at 18 and 42-h postinduction (Fig. 9,E). The expression of IFN was strongly induced by poly(I:C) and weakly induced by LTA at 18-h postinduction (Fig. 9,F). In the in vivo LPS i.p. injection experiment, two concentrations of LPS (10 and 100 mugg/g) were both able to induce the up-regulation of IL-1β expression, a cytokine produced in inflammatory reactions. However, neither two concentrations of LPS were able to induce IFN expression (Fig. 10). It appeared that the lower concentration of LPS yielded greater IL-1β expression than higher concentration of LPS. However, we thought there was no significant fold stimulation difference (∼1.5-fold stimulation difference) in terms of 10-fold concentration difference between these two groups based on the relative quantification via real-time PCR. As a matter of fact, a similar result (not shown in this manuscript) was also obtained for another acute phase immune response gene which could be significantly induced by bacterial infection (our previously unpublished observation). The potential reason for such unexpected observation might be that the lower concentration of LPS may have reached the threshold value of immune induction for zebrafish and the higher concentration could have reached saturation to cause no significant difference of induction. Detailed mechanisms for such observations will be interesting for future study.
In the current study, we report the cloning of zebrafish TRIF full-length cDNA, which has 32% identity with its human counterpart. By comparison with mammalian counterparts, zebrafish TRIF obviously has different amino acid character from its mammalian homologues. It does not have N-terminal and C-terminal proline-rich domains, the proline is not conserved in box2 of the TIR domain, and the NS4B/5A cleavage site is lost. It is of particular interest to know if the amino acid sequence difference will result in a functional difference between zebrafish and mammalian TRIF.
To answer this question, we did a functional analysis using a luciferase reporter assay system. The results from the luciferase report assay revealed that zebrafish TRIF activated the IFN promoter as well as the NF-κB response promoter but in a different way. We supposed that there was a negative feedback effect of zebrafish TRIF-mediated NF-κB pathway. The proline conserved in box2 of TIR domain was essential for human TRIF and MyD88. However there was a substitution from Pro to Ala in the corresponding site of zebrafish TRIF. To reveal the functional importance of A359 for zebrafish TRIF, we designed two single-point mutations A359P and A359H. The results showed that these two mutations lost their IFN promoter induction function, even though one of them was mutated to the conserved proline. So it was indicated that the A359 was vital for zebrafish TRIF to activate the IFN promoter rather than the P384 in human TRIF, as reported previously (10). The reason for this observation might be that the mutated amino acid changed the 3D structure of TIR domain, which was the basis for the recognition and interaction between signaling molecules. Another possible reason might be that the lack of signaling for the murine TLR4 mutant of P712H was not due to disruption of the TIR domain structure itself, but rather to the disruption of a direct point of contact with other molecule(s), specifically other TIR domains (28). Though P384 and its corresponding sites in other adaptors were thought to contribute to their functional conservation, it was noticeable that the Glycine adjacent to the Proline was much more conserved between all adaptors from mammals and fish based on our alignment (Fig. 1 C). It was of particular interest to study if the conserved Gly was a pure coincidence or had an important biological meaning.
In addition, the ectopic expression of the active NS3/4A protease inhibited IFN-β promoter activation by poly(I:C) in the cultured SL1c cells via a mechanism that NS3/4A protease inhibits activation of IRF-3 through the TLR3 pathway by mediating proteolysis of TRIF, which might contribute to HCV persistence (29). However, the result from amino acid alignment showed that zebrafish TRIF had no NS4B/5A cleavage site in its sequence. The physiological significance of this sequence variation needs to be elucidated thoroughly in the future.
Much evidence indicated that TLRs and adaptors had specific localizations in cells for their functions performed. To determine which sequence motif of zebrafish TRIF was responsible for its localization, we constructed TRIF-GFP fusion vectors and its truncated forms. The laser confocal microscopy showed that zebrafish TRIF was localized to a very unique site near to the nuclear membrane, a site most likely to be Golgi apparatus. It was of interest for us to further study the localization of TRIF in Golgi apparatus. Thus in the following experiments, we conducted cotransfection and laser scanning confocal imaging. The results indicated that zebrafish TRIF was able to colocalize with the marker gene targeted to the trans-medial region of the Golgi apparatus. Thus, it was revealed that zebrafish TRIF was a Golgi apparatus localized protein. Results from many recent reports suggested a tight connection between the intracellular membrane organs and innate immunity adaptors as well as receptors in mammals (19, 22, 32). It was very interesting that the similar phenomenon was also observed in zebrafish in our study. In addition, it appeared that the phosphorylation and/or myristylation of these molecules at specific sites were a common mechanism in helping their intracellular membrane anchoring and the signaling transduction (22, 32, 33, 34, 35, 36). There were several Ser or Tyr amino acids in zebrafish TRIF protein sequence predicted to be the potential phosphorylation sites, and it was very interesting that zebrafish TRIF also contained a putative inner myristylation sequence Met-Gly-Arg-Lys-Pro-Thr analogous to the consensus sequence Met-Gly-Xaa-Xaa-Xaa-Ser/Thr and Lys/Arg, which was found in the N-terninal myristoylation sequence of TRAM (32). TRIF and TRAM showed sequence similarity and appeared to evolve from the same ancestor as demonstrated (Fig. 11). Thus, it was suspected that zebrafish TRIF might also be phosphorylated and/or myristylated at certain sites. This posttranslational modification was suspected to be vital for its Golgi apparatus membrane binding and signal transduction function.
There are two antiviral signaling pathways in mammals, a TRIF-dependent pathway and TRIF-independent RIG-I pathway. In zebrafish, we have cloned the cDNA sequences of TLR3, MAVS, and TRIF. An EST with high similarity with mammalian RIG-I was also found in zebrafish database. All this information suggested that the key elements of these two antiviral signaling pathways were evolutionarily conserved in zebrafish and, thus, led us to propose that both two pathways could also work in this species. However, in zebrafish we did not know yet which pathway had a more important and specific response to viral infection or whether these two pathways had the same contribution to the antiviral signal transduction. To investigate the difference between the TRIF-dependent TLRs antiviral pathway and the TRIF-independent RIG-I pathway in response to ligand inductions, and to mimic infections by virus, bacterial, fungi, and yeast, we conducted i.p. injection tests in living adult zebrafish. Because it was reported that the downstream signal transduction in both two pathway converged with IRF3/5 phosphorylation and the IFN induction, we were not expecting to find the difference just focused on the downstream signaling molecules. Considering the studies on the mRNA expression of TLRs up-regulated or down-regulated by LPS (37, 38, 39), we analyzed the induced expression of TLR3, TRIF, RIG-I, and MAVS themselves upon poly(I:C), LTA, PGN, Zymosan, and Glucon induction and expected to investigate their different induction patterns in response to TLRs ligands challenge. Interestingly, we found that IRF7 was up-regulated by poly(I:C) in addition to being activated by phosphorylation in signaling transduction. We also observed that the expressions of RIG-I and MAVS were up-regulated specifically by poly(I:C) but the expression of TLR3 and TRIF had no obvious change. This suggested that there were different responses to virus infection between RIG-I/MAVS and TLR3/TRIF signal pathways and it appeared that RIG-I/MAVS pathway was more sensitive and specific to poly(I:C) induction than TLR3/TRIF. All these findings revealed an unique response to poly(I:C) for zebrafish.
It was demonstrated that TRAM, a fourth adaptor identified in TLR signaling pathways, could function in LPS-TLR4 signal pathway of IFN induction through interacting with TRIF (11). The amino acid sequence of mammalian TRAM had a much higher similarity with mammalian TRIF than with that of other TIR domain containing molecules. The evolutionary analysis based on sequence comparison also suggested that TRAM and TRIF evolved from the same ancestor and developed into two branches in higher vertebrates (Fig. 11). Since TRIF and TRAM had similar connection in sequence and function, we expected to find TRAM coding sequence from zebrafish in addition to the discovery of TRIF. Unfortunately, we could not find an EST or a genomic sequence homologous to mammalian TRAM in zebrafish DNA database. Iliev and his collogues (40) also reported the absence of TRAM in any of the fish genomic databases. We were wondering whether the absence of TRAM in zebrafish would have any physiological consequences, and were also curious about whether TRIF might be able to substitute for TRAM in zebrafish because of their close relationship. Results from us and Iliev et al. (40) suggested that in teleost fish, IFN was not induced by LPS; in contrast, inflammatory molecules (such as IL-1β) were significantly induced by the same ligand, which was similar for the TRAM-deficient mice (41). These results might suggest that zebrafish TLR4 or TLR4-receptor complex could respond to LPS, and the following downstream signaling pathways for inflammatory reactions were workable. In addition, the TRIF-mediated IFN induction system could also be functional as indicated by luciferase reporter assay and QPCR assay on poly(I:C) induction. All these key molecules were in the position of signal transduction pathway in zebrafish except for TRAM. Therefore, we speculated that in zebrafish, the absence of TRAM might be one of the potent reasons why IFN induction in response to LPS was blocked and consequently resulted in low responsiveness to endotoxin shock in zebrafish unlike in mammals (42, 43, 44).
In short, in this study, we reported the identification of the full cDNA sequence of zebrafish TRIF with certain unique characters in contrary to its mammalian counterpart. This zebrafish TRIF was able to activate IFN promoter and NF-κB response promoter, suggesting an evolutionary conservation of these two pathways from fish to mammals. In contrast, the unique sequence, subcellular localization, and functional characters of zebrafish TRIF and the striking findings from TLRs ligand induction test revealed a real difference in innate immune system between zebrafish and mammals. These observations may help our further understanding of innate immune system in mammals, particularly de novo innate immune responses to various pathogenic stimuli by using the zebrafish as a research model for studying the evolution of immune systems.
We thank Professor Jianguo He and Associate Professor Shaoping Wong’s permissions for ZF4 cells culture and transfection in their lab. We thank Ms. Qiuyun Liang for valuable experimental guidance in ZF4 cells culture. We thank Xiaoju Dong and Can Peng for helpful discussions. We also thank Haizhen Yan, Qingyu Yan, and Chi Ma for excellent technical assistances.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by Project 2007CB815800 of the National Basic Research Program (973) and Project 2006AA09Z433 of the State High-Tech Development Project (863) from the Ministry of Science and Technology of China, Project 2007DFA30840 of International S&T Cooperation Program of China, Key Project (0107) from the Ministry of Education, and Key Projects of Commission of Science and Technology of Guangdong Province and Guangzhou City. X.A. is a recipient of Outstanding Young Scientist Award of National Natural Science Foundation of China.
Abbreviations used in this paper: TRIF, TIR domain-containing adaptor inducing IFN-β; RIG-I, retinoic acid inducible gene I; HCV, Hepatitis C virus; TIRAP, TIR domain-containing adapter protein; IRF, IFN regulatory factor; poly(I:C), polyinosine-polycytidylic acid; EST, expressed sequence tag; UTR, untranslated regions; PI, propidium iodide; LTA, lipoteichoic acid.