Intracellular Toll-like receptor of zebrafish that senses unmethylated CpG DNA. It has a signal peptide, a leucine-rich-repeat ectodomain, and a cytoplasmic TIR domain, is expressed from early embryonic stages and, in adults, mainly in organs with innate immune function such as intestine, spleen and kidney. The receptor recognises CpG oligodeoxynucleotides with a sequence preference of its own, favouring GTCGTT motifs, and this specificity resides in its ectodomain; it does not respond to ligands of other Toll-like receptors such as double-stranded RNA, flagellin, lipopolysaccharide, acylated lipopeptides or the imidazoquinolines. Most of the protein is found in the endoplasmic reticulum, it binds the trafficking chaperone UNC93B1 through acidic residues in its juxtamembrane region, and its activation requires endosomal maturation and is blocked by chloroquine, so ligand engagement occurs inside the cell. Signalling depends on conserved residues of the TIR box 2 region. Functionally it works alongside the phylogenetically distant TLR9 of the same species: ligands that engage both receptors induce stronger cytokine responses and give better protection against bacterial infection than those engaging TLR9 alone, making this receptor a substantial part of the CpG-DNA sensing capacity of fish. It is the counterpart of avian TLR21, and the two are functional analogues of mammalian TLR9 rather than its orthologues.
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0002221 pattern recognition receptor signaling pathway | IDA PMID:22729906 Positive selection pressure within teleost Toll-like recepto... | KEEP AS NON CORE | Summary: The generic parent of the better-supported pathway terms, and its cited evidence is transcript induction rather than signalling. Reason: The claim is true: this receptor does initiate a pattern recognition receptor signalling pathway. But GO:0002224 and the proposed GO:0035682 are annotated on the same gene with stronger support, and the cited paper measured expression changes after immunostimulant exposure, which speaks to regulation of the gene rather than to its participation in signalling. Kept as a non-core parent rather than removed. Supporting Evidence: PMID:22729906 Up to a 3-fold increase in tlr21 and tlr22 |
| GO:0002224 toll-like receptor signaling pathway | IBA GO_REF:0000033 | MODIFY | Summary: Correct family-level statement, but the receptor-specific child is supported by direct evidence on this gene product. Reason: The ancestral assertion is sound and the target has its own direct evidence: cell-based activation assays in both human and zebrafish cell backgrounds, ectodomain swap chimeras, and TIR box 2 point mutants that abolish signalling. Since GO:0035682 toll-like receptor 21 signaling pathway names this receptor and is warranted by that evidence, both GO:0002224 rows are moved to it rather than leaving the inherited row at the parent level. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: GRANULARITY MISMATCH Sources checked: PANTHER:PTN002808083 SUPPORTS TRANSFER Family-level pathway assertion is correct; direct evidence on this gene product licenses the receptor-specific child. Proposed replacements: toll-like receptor 21 signaling pathway Supporting Evidence: PMID:24282308 zebTLR21 responded preferentially to CpG-ODN with GTCGTT motifs PMID:24282308 the ectodomains of these two TLRs determine their ligand recognition |
| GO:0002224 toll-like receptor signaling pathway | IDA PMID:24282308 Toll-like receptor 9 and 21 have different ligand recognitio... | MODIFY | Summary: This is the row that should carry the receptor-specific term, since the cited study demonstrated ligand-induced signalling through this receptor. Reason: GO:0035682 toll-like receptor 21 signaling pathway exists for exactly this receptor, and the cited work supplies what the term requires: activation by a defined ligand, specificity determined by this receptor's ectodomain, dependence on its own TIR residues, and downstream output. The receptor identity is unambiguous, so unlike the mouse TLR11/TLR12 pair the numbered term can be used with confidence. The family-level IBA row is retained separately. Proposed replacements: toll-like receptor 21 signaling pathway Supporting Evidence: PMID:24282308 zebTLR21 responded preferentially to CpG-ODN with GTCGTT motifs PMID:24282308 the ectodomains of these two TLRs determine their ligand recognition PMID:24282308 did not activate cells transfected with vector for control, zebTLR9, and zebTLR21 |
| GO:0005783 endoplasmic reticulum | IDA PMID:24282308 Toll-like receptor 9 and 21 have different ligand recognitio... | ACCEPT | Summary: Most of the receptor is in the endoplasmic reticulum, as visualised directly. Reason: Confocal microscopy after immunofluorescence staining placed the majority of the receptor in the endoplasmic reticulum, where it colocalises with UNC93B1 and binds it. This is the resting location; activation additionally requires endosomal maturation, which is recorded in the description and in the questions rather than as a competing location annotation, since no endosomal localisation was imaged directly. Supporting Evidence: PMID:24282308 The majority of that zebTLR9 and zebTLR21 was localized in the endoplasmic reticulum PMID:24282308 both zebTLR9 and zebTLR21 bound with UNC93B1 in immunoprecipitation reactions |
| GO:0005886 plasma membrane | IBA GO_REF:0000033 | MARK AS OVER ANNOTATED | Summary: An intracellular, chloroquine-sensitive receptor; the surface location was inherited from cell-surface TLR donors. Reason: The ancestral node is seeded by surface receptors, and the property does not transfer: the receptor was imaged in the endoplasmic reticulum, its activation by CpG ligands occurs inside cells and is abolished by blocking endosomal acidification, and it depends on UNC93B1 like the other nucleic-acid-sensing receptors. Marked as over-annotated rather than removed because some transit through or residence at the membrane system is not excluded and the endosomal compartment of action has not itself been imaged; the generic membrane row already covers what is certain. Propagation Review Root cause: PROPAGATION BAD Failure modes: COMPARTMENT OR COMPLEX MISMATCH Sources checked: PANTHER:PTN002808045 SUPPORTS SOURCE BUT NOT TARGET Donors are cell-surface TLRs and Drosophila Toll; this nucleic-acid-sensing branch acts intracellularly. Supporting Evidence: PMID:24282308 The majority of that zebTLR9 and zebTLR21 was localized in the endoplasmic reticulum PMID:24282308 This agent effectively blocked the activation of zebTLR9 |
| GO:0007165 signal transduction | IEA GO_REF:0000002 | KEEP AS NON CORE | Summary: True but generic. Reason: Sound TIR-domain inference, superseded by the pathway annotations on the same gene. |
| GO:0016020 membrane | IEA GO_REF:0000044 | KEEP AS NON CORE | Summary: Correct at the generic level for a Toll-like receptor. Reason: True and, for this receptor, the only membrane statement that is safe: the specific compartment of action is intracellular but has not been imaged directly. |
| GO:0019731 antibacterial humoral response | IDA PMID:24282308 Toll-like receptor 9 and 21 have different ligand recognitio... | KEEP AS NON CORE | Summary: The receptor mediates a CpG-driven protective response to a bacterial pathogen, but as the sensor upstream of the humoral effectors. Reason: The cited work showed that ligands engaging this receptor induced cytokines and improved survival and bacterial clearance in Edwardsiella tarda infection, so the annotation is evidentially grounded and the acts_upstream_of_or_within relation is apt. It is kept as non-core rather than core because the receptor does not itself perform the humoral antibacterial work: it recognises the ligand and initiates signalling, and the effector molecules are induced downstream. The informative statements about this gene product are the recognition activity and the pathway term. Supporting Evidence: PMID:24282308 These results suggest that zebTLR9 and zebTLR21 cooperatively mediate the antimicrobial PMID:24282308 were more bactericidal in zebrafish |
| GO:0038023 signaling receptor activity | IBA GO_REF:0000033 | MODIFY | Summary: Unmethylated CpG DNA is a pathogen-associated pattern recognised through this receptor's ectodomain, so the specific molecular function applies. Reason: GO:0038023 is correct but generic. Heterologous expression of this receptor confers responsiveness to CpG oligodeoxynucleotides with a distinctive motif preference, chimera experiments show that its ectodomain determines that preference, and ligands of the other Toll-like receptors do not activate it, which together support recognition by this receptor rather than by something it recruits. GO:0038187 is the conventional molecular function for characterised TLRs. GO:0045322 unmethylated CpG binding is not proposed, because no direct binding assay has been done for this receptor and even mammalian TLR9 carries that term only by ISS and IEA. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: GRANULARITY MISMATCH Sources checked: PANTHER:PTN002808045 SUPPORTS TRANSFER The node assertion is right; direct functional data on this gene product license the specific child. Proposed replacements: pattern recognition receptor activity Supporting Evidence: PMID:24282308 zebTLR21 responded preferentially to CpG-ODN with GTCGTT motifs PMID:24282308 the ectodomains of these two TLRs determine their ligand recognition PMID:24282308 did not activate cells transfected with vector for control, zebTLR9, and zebTLR21 |
| GO:0061809 NAD+ nucleosidase activity, cyclic ADP-ribose generating | IEA GO_REF:0000003 | REMOVE | Summary: A spurious enzymatic activity inherited from an EC number on the sequence record; no Toll-like receptor has demonstrated NADase activity. Reason: This row is an electronic mapping of EC:3.2.2.6, which appears on the TrEMBL DE line because the underlying RefSeq record carries it. The claim is almost certainly a mis-generalisation from TIR domains that do have NAD hydrolase activity, namely plant TIR proteins and human SARM1. UniProt states the opposite conclusion for this family explicitly on its reviewed entries: despite the presence of the catalytic aspartate, the isolated TIR domain of human TLR4 lacks NADase activity, and it is therefore unlikely that Toll-like receptors have it. Nothing in the functional literature on any TLR21 reports enzymatic activity; the receptor's characterised activity is CpG-DNA recognition, and a nucleosidase annotation would also imply a cytosolic catalytic role at odds with the ectodomain-mediated recognition demonstrated. This is a demonstrably wrong EC-to-GO mapping, which is the case the REMOVE action exists for. Supporting Evidence: PMID:24282308 zebTLR21 responded preferentially to CpG-ODN with GTCGTT motifs PMID:24282308 did not activate cells transfected with vector for control, zebTLR9, and zebTLR21 file:DANRE/tlr21/tlr21-deep-research-falcon.md The EC number 3.2.2.6 assigned to F1QMN8 in UniProt appears to be an annotation error |
Loading supporting content…
Download this section (compressed HTML)Q: Which adaptor does this receptor use? No zebrafish experiment has tested adaptor recruitment for it, and the comparative literature is divided between a MyD88-associated model and a TRIF-exclusive model proposed for fish TLR21, so no adaptor-dependent process term is annotated here.
Q: Does this receptor bind CpG DNA directly, and at what pH? All the evidence is functional, and a binding assay on the ectodomain would license GO:0045322 unmethylated CpG binding, which mammalian TLR9 itself holds only by similarity.
Q: In which compartment does activation actually occur? The resting receptor was imaged in the endoplasmic reticulum and activation is blocked by inhibitors of endosomal maturation, but no endosomal pool has been visualised, which is why no endosome annotation is proposed here.
Q: Why does GOA carry an EC:3.2.2.6-derived NADase annotation on this entry? The EC number is on the RefSeq record; correcting it at source would prevent the same annotation reappearing on this and other fish TLR entries.
Q: How should GO express the relationship between TLR21 and TLR9, which sense the same ligand class from different branches of the family and, in zebrafish, do so in the same animal with complementary motif preferences? Annotation transfer between them would be wrong, but the functional analogy is informative.
Experiment: Measure binding of defined CpG and control oligodeoxynucleotides to the purified ectodomain of this receptor at acidic and neutral pH, alongside the zebrafish TLR9 ectodomain, to test whether the motif preferences seen in cells reflect differential affinity.
Hypothesis: The GTCGTT preference is a property of direct ligand binding by the ectodomain rather than of downstream events.
Type: direct binding assay
Experiment: Track the receptor at endogenous levels with a tagged knock-in allele before and after CpG stimulation, and test whether UNC93B1 disruption prevents its exit from the endoplasmic reticulum and abolishes the response.
Hypothesis: Activation follows UNC93B1-dependent trafficking from the endoplasmic reticulum to an acidified endosomal compartment, which would justify an endosome annotation currently unsupported.
Type: endogenous localisation and trafficking
Experiment: Generate zebrafish lacking this receptor, its paralogous relative tlr22 and tlr9 singly and in combination, and compare CpG-induced cytokine responses and survival after Edwardsiella tarda challenge.
Hypothesis: The cooperation inferred from ligand-preference comparisons reflects genuinely non-redundant contributions of the two receptors to antibacterial defence.
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)