Toll-like receptor of the fish-specific TLR22 group, found in aquatic vertebrates and absent from tetrapods. The protein has the canonical architecture of the family: a signal peptide, a leucine-rich-repeat ectodomain with a cysteine-rich flanking region, a single transmembrane segment and a cytoplasmic TIR domain, and it belongs phylogenetically to the branch containing TLR21 and mouse TLR13 rather than to any mammalian receptor's orthology group. In zebrafish its transcript is found mainly in spleen and kidney and is raised by exposure to unrelated microbial products including lipopolysaccharide, peptidoglycan and double-stranded RNA, and its leucine-rich repeats carry signatures of positive selection at several sites, consistent with ongoing diversification of ligand recognition. The function of the receptor is best characterised in another teleost, fugu, where TLR22 detects long double-stranded RNA at the cell surface and signals through the TICAM-1 adaptor to induce interferon and confer resistance to aquatic RNA viruses, a role proposed to substitute for cell-surface TLR3 in fish. Whether the zebrafish protein has the same ligand and adaptor has not been tested.
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0002221 pattern recognition receptor signaling pathway | IDA PMID:22729906 Positive selection pressure within teleost Toll-like recepto... | MODIFY | Summary: The gene is an unambiguous Toll-like receptor, so the TLR-specific child of this term is the right level; no receptor-numbered term exists for TLR22. Reason: GO:0002221 is correct but is the parent shared with the NOD-like and RIG-I-like receptor pathways, and nothing about this gene suggests either of those: the domain architecture (leucine-rich repeats, transmembrane segment, TIR domain), the UniProt family assignment and the phylogenetic placement beside zebrafish tlr21, tlr5b and tlr3 all make it a Toll-like receptor. GO:0002224 is therefore the informative level. This is also the answer to the project's question about which pathway term fits TLR22: GO has numbered terms for TLR1-13, 15 and 21 but none for TLR22, and none should be requested yet, because for the zebrafish protein neither ligand nor adaptor has been demonstrated and the gene's naming is unstable (its own synonym is TLR21.2). The evidence behind this row is transcript modulation rather than signalling, so the annotation is retained at family level rather than promoted further. Proposed replacements: toll-like receptor signaling pathway Supporting Evidence: PMID:22729906 Immunostimulation experiments revealed that expression of zebrafish tlr22 is PMID:22729906 Up to a 3-fold increase in tlr21 and tlr22 PMID:24282308 zebTLR21 is most closely related to zebTLR22, zebTLR5b, and zebTLR3 |
| GO:0007165 signal transduction | IEA GO_REF:0000002 | KEEP AS NON CORE | Summary: True but generic. Reason: A sound TIR-domain inference from InterPro, and superseded by the proposed GO:0002224. |
| GO:0009617 response to bacterium | IDA PMID:17804254 Innate immune gene expression in individual zebrafish after ... | UNDECIDED | Summary: Cannot be verified from the available record: the cited study measured this transcript among many after bacterial inoculation, but the cached abstract does not report what it did. Reason: The cached publication is abstract-only. The abstract states that this gene was among the transcripts monitored by real-time RT-PCR after Listonella anguillarum inoculation, and singles out other genes as induced without saying anything about this one; it also emphasises wide interindividual variation in the measurements. The ZFIN curator will have had the figures and may have seen clear induction, so the annotation is not removed or marked as over-annotated on the strength of an abstract. Two further considerations argue for leaving it open rather than accepting it: transcript induction alone is weak evidence for participation in a response process, and the only characterised ligand of a TLR22 protein in any species is double-stranded RNA, which would make a viral rather than a bacterial role the expected one. A curator with the full text should settle it. Supporting Evidence: PMID:17804254 response (il1b, cebpb, tfa, mpx, tnfa, nitr9, tlr22, hsc70, cp, mrlp1, c3b and PMID:18714020 fgTLR22 recognizes long-sized dsRNA on the cell surface |
| GO:0016020 membrane | IEA GO_REF:0000120 | ACCEPT | Summary: Correct at the generic level: signal peptide plus a predicted transmembrane segment. Reason: Well founded for this family and for this entry's features, and deliberately left generic. The fugu orthologue acts at the cell surface, but no localisation has been reported for the zebrafish protein, so a plasma-membrane annotation would be a cross-species transfer of exactly the kind this review avoids elsewhere. The transfer would in any case be ambiguous, because the comparative literature is itself divided. The fugu work places the receptor at the cell surface, whereas recent reviews of fish Toll-like receptors describe TLR22 as endosomal and lysosomal, and the difference may be genuinely species-specific. Supporting Evidence: PMID:18714020 fgTLR22 recognizes long-sized dsRNA on the cell surface file:DANRE/tlr22/tlr22-deep-research-falcon.md the zebrafish TLR22 protein itself (A0A2R8RTN4) has not been directly tested for ligand binding |
Loading supporting content…
Download this section (compressed HTML)Q: Should GO create a toll-like receptor 22 signaling pathway term? This review says not yet: GO:0002224 is the right level while the zebrafish receptor's ligand and adaptor are undemonstrated, the fish tlr21/tlr22/tlr23 names are unstable (this gene's own synonym is TLR21.2), and a term defined by receptor name alone would add nothing. Once ligand and adaptor are established, a term modelled on GO:0035682 would be worthwhile.
Q: Does zebrafish tlr22 recognise long double-stranded RNA, as the fugu protein does? This is the single measurement that would change this gene's annotation the most, converting family-level statements into a specific recognition activity and an interferon-directed process.
Q: Which adaptor does it use? The fugu receptor signals through TICAM-1 (TRIF) rather than MyD88, which is unusual in this branch of the family and would need confirming in zebrafish before any MyD88-dependent pathway term were considered.
Q: What does the GO:0009617 row actually rest on? The cited study is abstract-only in this repository and the abstract does not report this transcript's behaviour; a curator with the figures could confirm, correct or retire the annotation.
Q: Is the positive selection detected in the leucine-rich repeats accompanied by functional divergence between tlr22 alleles or between species? If so, ligand specificity may not be safely transferable even between teleosts, which would bear directly on how far any TLR22 annotation should propagate.
Experiment: Test ligand specificity of the zebrafish receptor directly: express it in zebrafish and mammalian cells, with and without zebrafish UNC93B1, and stimulate with long and short double-stranded RNA, poly(I:C), lipopolysaccharide, peptidoglycan, flagellin and CpG oligodeoxynucleotides, reading NF-kappa-B and interferon-stimulated response element reporters.
Hypothesis: Zebrafish tlr22 recognises long double-stranded RNA and signals to interferon, as its fugu orthologue does, which would justify specific recognition and pathway annotations that cannot be made now.
Type: ligand specificity assay
Experiment: Determine the adaptor requirement and localisation: co-express the receptor with zebrafish ticam1 and myd88, test loss of each for effect on the dsRNA response, and image a tagged receptor for surface versus intracellular distribution.
Hypothesis: The receptor acts at the cell surface through TICAM-1, unlike the MyD88-dependent intracellular receptors of the same branch.
Type: adaptor dependence and localisation
Experiment: Generate zebrafish lacking tlr22 and challenge them with an aquatic RNA virus and with a Gram-negative bacterial pathogen, measuring interferon and inflammatory responses and survival, so that the viral and bacterial hypotheses are tested in the same animals.
Hypothesis: Loss of the receptor impairs antiviral rather than antibacterial defence, which would also indicate whether the existing response-to-bacterium annotation should stand.
Type: loss-of-function genetics with infection challenge
Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)