Cell-surface Toll-like receptor that forms a distinct subfamily of the Toll-like receptors, with no mammalian orthologue, and the only known member of the family that is switched on by proteolysis rather than by binding a microbial molecule. It is best known from birds and reptiles, but orthologues also occur in lungfish and in holocephalan cartilaginous fish, so the gene is an ancient jawed-vertebrate lineage that has been lost or pseudogenised in many groups, including all mammals, rather than an avian invention. It has the usual TLR architecture of a leucine-rich-repeat ectodomain, a transmembrane segment and a cytoplasmic TIR domain, and is displayed as a glycoprotein at the cell surface. Secreted proteases from fungi and bacteria cleave its ectodomain, and this cleavage activates the receptor and drives NF-kappa-B-dependent transcription; a truncated form lacking the whole ectodomain is constitutively active, indicating that the leucine-rich repeats hold the receptor in an inhibited state that proteolysis relieves. Consistent with this mechanism, yeast lysates activate the receptor in a manner abolished by heat or by serine-protease inhibition, while purified ligands for the other Toll-like receptors, including acylated lipopeptides, lipopolysaccharide, flagellin, zymosan and CpG DNA, do not. Its transcript is induced in chicken tissues after Salmonella infection. Phylogenetically it sits near the TLR1/TLR2/TLR6/TLR10 group without belonging to it, and its ectodomain has a single-domain architecture rather than the three-domain architecture shared by that group; its activation mechanism resembles that of the protease-activated receptors more than that of a pattern recognition receptor. The cytoplasmic TIR domain has been crystallised as a dimer whose interface differs from those of the TLR1-subfamily TIR domains, and comparison of that structure with adaptor TIR domains led its authors to propose MyD88 as the adaptor.
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0002224 toll-like receptor signaling pathway | IBA GO_REF:0000033 | MODIFY | Summary: Correct, but the receptor-specific child exists and is supported by direct experiment. Reason: Two independent studies show that signalling is initiated at this receptor and ends in NF-kappa-B-dependent transcription, which is what GO:0035681 toll-like receptor 15 signaling pathway describes. Using the numbered child is appropriate here and not merely more specific: TLR15 has no mammalian orthologue, so the pathway it initiates is not a species variant of any other numbered TLR pathway, and its activation mechanism is distinct. An independent study in a chicken cell background adds gain- and loss-of-function support for signalling at this receptor: overexpression in DF-1 fibroblasts augmented, and knockdown reduced, activation of the NF-kappa-B and NLRP3/IL-1-beta pathways on exposure to Eimeria tenella sporozoites. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: GRANULARITY MISMATCH Sources checked: PANTHER:PTN002808083 SUPPORTS TRANSFER Family-level pathway assertion is correct; the target has direct evidence for the receptor-specific child. Proposed replacements: toll-like receptor 15 signaling pathway Supporting Evidence: PMID:21383168 Activation of TLR15 involves proteolytic cleavage of the receptor ectodomain PMID:23066147 TLR15 mediated NF-ΞΊB induction in response to lysates from yeast, but not those PMID:33514434 knockdown of ChTLR15 in DF1 cells showed inverse effects PMID:37079400 These observations allow us to propose that TLR15 signals through MyD88 |
| GO:0005886 plasma membrane | IBA GO_REF:0000033 | ACCEPT | Summary: Surface localisation, unusually for an inherited location call, is confirmed experimentally for this receptor. Reason: The ancestral node's donors are cell-surface TLRs and the property does transfer here: cell-surface biotinylation labelled only the mature receptor band, confocal microscopy showed colocalisation with a surface marker in three cell types, and only the surface-exposed fraction was cleaved by protease. Surface exposure is also mechanistically required, since the activating proteases are extracellular. Supporting Evidence: PMID:21383168 In contrast to TLR9, TLR15 is expressed at the cell surface PMID:21383168 Activation of TLR15 involves proteolytic cleavage of the receptor ectodomain |
| GO:0007165 signal transduction | IEA GO_REF:0000002 | KEEP AS NON CORE | Summary: True but generic. Reason: Sound TIR-domain inference, superseded by the proposed GO:0035681. |
| GO:0016020 membrane | IEA GO_REF:0000044 | KEEP AS NON CORE | Summary: Correct for a single-pass membrane protein. Reason: True but uninformative next to the plasma-membrane annotation, which is experimentally supported. |
| GO:0038023 signaling receptor activity | IBA GO_REF:0000033 | ACCEPT | Summary: The right molecular function for this receptor, and deliberately not upgraded to pattern recognition receptor activity. Reason: TLR15 receives a signal and transmits it, so GO:0038023 is accurate. It is kept as-is rather than replaced with GO:0038187 pattern recognition receptor activity, which every other characterised TLR in this batch takes, because GO:0038187 requires combining with a pathogen-associated molecular pattern and this receptor does not do that: it is activated by cleavage of its own ectodomain by secreted microbial proteases, and purified microbial patterns do not activate it. The proteases are enzymes acting on the receptor, not ligands the receptor binds, so the specific term would misdescribe the mechanism. A 2025 report is consistent with this reading rather than against it: the parasite molecule recovered as a candidate ligand is itself an aspartyl protease, so the one candidate protein ligand proposed for this receptor is again a protease, and the authors present the receptor assignment as provisional. Supporting Evidence: PMID:21383168 Activation of TLR15 involves proteolytic cleavage of the receptor ectodomain PMID:21383168 a high dose of purified microbial TLR ligands did not activate TLR15 PMID:21383168 various types of TLR ligands (LPS, di- and triacylated lipopeptides, zymosan, flagellin, TLR7/8 ligand, CpG DNA, and profilin-like protein) were tested for their ability to activate TLR15 PMID:40639002 an aspartyl protease from Eimeria tenella (EteASP) was identified as a potential TLR15 ligand by co-immunoprecipitation file:CHICK/TLR15/TLR15-deep-research-falcon.md pathogen-associated protease sensing remains its best-supported molecular function |
| GO:0042497 triacyl lipopeptide binding | IBA GO_REF:0000033 | REMOVE | Summary: Ligand specificity transferred from human TLR2; TLR15 is not a TLR2 orthologue and does not respond to acylated lipopeptides. Reason: The donor of this assertion is human TLR2, which has the activity by IDA. The transfer fails on two counts. First, orthology: TLR15 exists only in birds and reptiles, and although it falls near the TLR1/TLR2/TLR6/TLR10 group it shows no close relationship to any member of it, so there is no clade within which the TLR2 ligand specificity could have been inherited. A later phylogenomic study puts this beyond doubt and strengthens it structurally: the gene is a distinct vertebrate subfamily of ancient origin, present in lungfish and holocephalan cartilaginous fish as well as birds and reptiles, and its ectodomain has a single-domain architecture rather than the three-domain architecture shared by all TLR1-family members, which is the architecture in which the lipopeptide-binding site of the TLR1/TLR2 pair is built. Second, direct evidence against: the study that identified the activation mechanism tested di- and triacylated lipopeptides on this receptor, with and without each candidate chicken surface co-receptor, and a high dose of purified microbial TLR ligands did not activate it; a second study found no response to a panel of well-characterised TLR agonists. Removal concerns the ligand claim only and says nothing against the family-level annotations. Propagation Review Root cause: PROPAGATION BAD Failure modes: WRONG ORTHOLOG OR PARALOG FUNCTIONAL DIVERGENCE Sources checked: UniProtKB:O60603 Β· TLR2 SUPPORTS SOURCE BUT NOT TARGET Human TLR2 genuinely has triacyl lipopeptide binding by IDA; the specificity is a property of the TLR1/TLR2 heterodimer and there is no orthology bridge to this avian/reptilian receptor. PANTHER:PTN002808136 SOURCE BAD The node places a TLR2-clade ligand specificity on a branch that includes TLR15, which is not a TLR2 orthologue and is experimentally unresponsive to lipopeptides. Supporting Evidence: PMID:21383168 various types of TLR ligands (LPS, di- and triacylated lipopeptides, zymosan, flagellin, TLR7/8 ligand, CpG DNA, and profilin-like protein) were tested for their ability to activate TLR15 PMID:21383168 a high dose of purified microbial TLR ligands did not activate TLR15 PMID:36605191 an ectodomain with single-domain architecture, instead of the three-domain architecture shared by all TLR1 family members PMID:36605191 Our phylogenetic analyses also show, with good support, that TLR15 forms a distinct subfamily within vertebrate TLRs PMID:23066147 TLR15 can only be identified in avian PMID:16495540 TLR15 groups with high bootstrap support with the recognized TLR1/TLR2/TLR6/TLR10 clade but displays no close relationship to any particular clade member |
| GO:0043235 signaling receptor complex | IBA GO_REF:0000033 | KEEP AS NON CORE | Summary: Inherited from the TLR2/TLR1 heterodimer; defensible only at this generic level, since TLR15's signalling complex is unidentified. Reason: The donor basis is TLR2, whose complex is the TLR1/TLR2 heterodimer, and no heterodimerisation partner has been found for TLR15 β the ligand study expressed it with each candidate chicken surface TLR without effect. What is reported is that the ectodomain-less receptor autoactivates, which the authors interpret as a tendency to self-dimerisation, as for TLR4 and Drosophila Toll. Since TLRs generally signal as dimers, membership of some signalling receptor complex is plausible, so the row is kept rather than removed, but it is not a core statement. The composition is now better constrained than it was, since the cytoplasmic TIR domain has been crystallised as a dimer and mutation of an interface residue abolished the dimer peak in solution, so the complex is most likely a homodimer of this receptor rather than a heterodimer with another TLR. It is still kept as non-core, because the isolated TIR domain is predominantly monomeric in solution and no dimer of the full-length receptor has been shown. Propagation Review Root cause: NO FAILURE NON CORE Failure modes: COMPARTMENT OR COMPLEX MISMATCH Sources checked: UniProtKB:O60603 Β· TLR2 SUPPORTS SOURCE BUT NOT TARGET TLR2's complex is the TLR1/TLR2 heterodimer; the generic parent term survives the transfer, the specific partnership does not. Supporting Evidence: PMID:21383168 various types of TLR ligands (LPS, di- and triacylated lipopeptides, zymosan, flagellin, TLR7/8 ligand, CpG DNA, and profilin-like protein) were tested for their ability to activate TLR15 PMID:37079400 Collectively, our structural and mutational analyses suggest that TLR15TIR employs the BB, Ξ±C2 and DD regions for dimerization |
Loading supporting contentβ¦
Download this section (compressed HTML)Q: Is the Eimeria tenella evidence sufficient for a defence-response process annotation on this receptor? Overexpression and knockdown in a fibroblast line change NF-kappa-B and NLRP3/IL-1-beta activation by sporozoites, a parasite aspartyl protease was recovered as a candidate ligand, and cecal expression correlates negatively with oocyst output, but no loss-of-function animal has been tested and the ligand paper leaves other pattern recognition receptors open, so no process term is proposed here.
Q: Does GO need a molecular function for a receptor activated by proteolysis of its own ectodomain? This receptor is left on the generic GO:0038023 because GO:0038187 pattern recognition receptor activity would assert pattern binding that does not occur; the protease-activated receptors face the same modelling problem.
Q: Is any co-receptor required for TLR15 signalling? Candidate chicken surface TLRs were co-expressed without effect, and the GO:0043235 row inherited from the TLR2/TLR1 heterodimer is kept only as a generic claim; identifying the actual dimer would let it be replaced by something specific.
Q: Do the reported responses to bacterial infection and to yeast lysates both reduce to protease sensing, or is there a second, protease-independent activation route? Yeast-lysate activity was abolished by serine-protease inhibition and zymosan did not activate, which suggests one mechanism, but this has not been tested with protease-deficient pathogens.
Q: Which host proteases, if any, can activate this receptor? Non-microbial proteases were reported to activate it in vitro, which raises the question of whether the receptor also responds to host-derived proteolysis during inflammation, as the protease-activated receptors do.
Experiment: Infect chicken macrophages and epithelial cells with isogenic pairs of bacterial and fungal strains differing only in secreted protease production, and measure receptor cleavage and NF-kappa-B output.
Hypothesis: Protease secretion is necessary and sufficient for pathogen-driven activation of this receptor, which would establish protease sensing rather than pattern recognition as the physiological mechanism.
Type: infection with isogenic protease mutants
Experiment: Map the cleavage site or sites in the ectodomain by mass spectrometry of the cleaved receptor, then test non-cleavable substitutions for loss of activation while confirming surface expression.
Hypothesis: A defined cleavage event, rather than general ectodomain degradation, activates the receptor; the cleaved epitope would also give a reagent for detecting activation in tissue.
Type: cleavage site mapping with mutagenesis
Experiment: Test the lipopeptide question directly in a chicken cell background by expressing this receptor with and without each chicken TLR1-like and TLR2-like paralogue and stimulating with Pam3CSK4 and Pam2CSK4 across a dose range.
Hypothesis: The inherited triacyl lipopeptide binding annotation is wrong; a systematic negative in the homologous species would close the question rather than leaving it resting on a heterologous assay.
Type: ligand specificity assay
Loading supporting contentβ¦
Download this section (compressed HTML)Loading supporting contentβ¦
Download this section (compressed HTML)Loading supporting contentβ¦
Download this section (compressed HTML)