tlr22 (zebrafish, UniProt A0A2R8RTN4) — curation notes

These notes rest on the UniProt record, the GOA rows and the cached publications;
the Falcon deep-research report arrived after the review was written and is
cross-checked at the end of this file. This is the thinnest evidence base of the
eight genes reviewed in this batch, and the review is correspondingly conservative.

Entry identity (TrEMBL, unreviewed)

A0A2R8RTN4_DANRE, 955 aa, submitter name "Toll-like receptor 22 precursor", gene
name tlr22 (synonym TLR21.2), cross-referenced to RefSeq NP_001122147, GeneID 403135
and ZFIN ZDB-GENE-040220-5. NCBI gene 403135 is "tlr22 toll-like receptor 22" on
chromosome 21, so this is the right gene. Features: signal peptide 1..30,
transmembrane 749..774, TIR domain 801..942, leucine-rich repeats with a
cysteine-rich flanking region. Its PANTHER subfamily assignment is
PTHR24365:SF522 "LOW QUALITY PROTEIN: TOLL-LIKE RECEPTOR 13-RELATED", which is
consistent with the phylogenetic placement of tlr22 in the TLR21/TLR22/TLR13 branch
and is a family name, not a statement about this protein's quality.

GOA holds only four rows: two electronic, and two IDA rows from ZFIN citing
expression studies.

Established biology

For zebrafish tlr22 specifically, almost everything available is transcript-level:

• PMID:22729906
  PMID:22729906, with
  PMID:22729906
  PMID:22729906
  lipopolysaccharide, peptidoglycan or poly I:C; expression is mainly in spleen and
  kidney. The same study found
  PMID:22729906
  PMID:22729906, which the authors read as
  functional diversification of ligand recognition.
• PMID:17804254
  PMID:17804254
  PMID:17804254 after
  Listonella anguillarum inoculation. The cached record is abstract-only and the
  abstract does not state what tlr22 did; the genes it names as changing are others.
• In another teleost the function is known:
  PMID:18714020 and
  signals through the TICAM-1 (TRIF) adaptor to induce interferon, protecting cells
  from aquatic dsRNA viruses; the authors propose it as a functional substitute for
  cell-surface TLR3 in fish. That is fugu, not zebrafish, so it informs expectations
  without supporting an annotation here.
• The zebrafish TLR21 study notes in passing that zebTLR21 is most closely related to
  zebTLR22, zebTLR5b and zebTLR3, and that poly(I:C), described there as
  "the TLR3 and TLR22 ligand", did not activate TLR21 — an indirect indication that
  dsRNA recognition is attributed to tlr22 in this species too, though it was not
  tested on tlr22 in that work.

The project's question: which pathway term fits TLR22?

GO has numbered pathway terms for TLR1-13, 15 and 21 but none for TLR22. The answer
this review reaches is that GO:0002224 toll-like receptor signaling pathway is
sufficient and no new numbered term should be requested yet. The reasoning:

1. The numbered terms exist where a receptor's ligand and signalling route are
   established, which is what makes the term more informative than its parent. For
   zebrafish tlr22 neither is established: the ligand is inferred from fugu, and no
   signalling assay has been reported for the zebrafish protein.
2. If a term were created now it would be defined by receptor name alone, and the
   name is unstable: this gene's own synonym is TLR21.2, the teleost tlr21/tlr22/tlr23
   genes have been renamed and re-partitioned between species, and the PANTHER
   subfamily groups it with TLR13.
3. A term would be worth requesting once the zebrafish receptor's ligand and adaptor
   are demonstrated — and at that point the useful shape would be a term modelled on
   GO:0035682, with the fugu route (TICAM-1-dependent interferon induction in response
   to long dsRNA) as the template if it proves conserved.

Recorded in suggested_questions rather than in proposed_new_terms, deliberately:
proposing a term now would assert a receptor-specific pathway that nobody has
demonstrated for this species.

Annotation decisions (summary)

• GO:0002221 pattern recognition receptor signaling pathway (IDA, PMID:22729906)
  modified to GO:0002224 toll-like receptor signaling pathway. The gene is an
  unambiguous member of the family by domain architecture and phylogeny, so the
  TLR-specific child of the annotated term is the right level; this is also the
  concrete answer to the project's "which pathway term" question.
• GO:0009617 response to bacterium (IDA, PMID:17804254) left UNDECIDED. The cached
  record is abstract-only; the abstract lists tlr22 among genes measured but does not
  report how it behaved, and the genes it singles out as induced are others. The
  ZFIN curator may well have read a figure showing induction, so the row is not
  removed; but the assertion cannot be verified from what is available, which is the
  circumstance the UNDECIDED action exists for.
• GO:0007165 kept as non-core; GO:0016020 membrane accepted (signal peptide plus a
  predicted transmembrane segment) but deliberately not refined to a specific
  compartment.
• No NEW annotations. In particular no dsRNA-binding, interferon-pathway or
  cell-surface annotation: all three would rest on the fugu orthologue, and
  cross-species transfer of a ligand specificity is exactly what this project set out
  to scrutinise. They appear instead as the leading suggested experiments.

Deep-research cross-check (2026-09-30)

Compared tlr22-deep-research-falcon.md against the finished review. Right gene, and it
independently confirms the review's central judgment, in almost the same terms: the
functional picture for TLR22 comes from other teleosts, and
[file:DANRE/tlr22/tlr22-deep-research-falcon.md "the zebrafish TLR22 protein itself (A0A2R8RTN4) has not been directly tested for ligand binding"].
It lists as unresolved exactly the four things the review declined to annotate: ligand
specificity of the zebrafish protein, subcellular localisation, adaptor usage and downstream
pathway. Added as a reference and cited on the generic membrane row for that reason.

One refinement taken up:

• The cross-species localisation transfer the review declined would not even have had a
  single target. The report shows the comparative literature divided, with the fugu work
  placing the receptor at the cell surface while recent reviews of fish Toll-like receptors
  describe TLR22 as endosomal and lysosomal, possibly a genuine species difference. That
  strengthens the decision to keep GO:0016020 generic, and is now recorded in that row's reason.

Considered and not taken up:

• Mycobacterium-induced expression of zebrafish tlr22, which the report offers, traces to a
  2024 review rather than to a primary zebrafish paper that could be quoted; searches for a
  primary source found none. It would in any case be another transcript-level observation, so it
  does not resolve the UNDECIDED GO:0009617 row, which turns on what PMID:17804254's figures
  show. The row stays UNDECIDED.
• Interferon induction, dsRNA recognition, MyD88 or TICAM-1 dependence, ERK/JNK/p38 effects,
  dendritic-cell cross-presentation and the reactive-oxygen-species-limiting role are all from
  fugu, grouper, trout or reviews. None is annotated, consistent with the review's original
  position that cross-species ligand and pathway transfer is what this project is scrutinising.
• The report notes one comparative genomic study reporting TLR22 absent from zebrafish
  assemblies, which conflicts with the existence of this gene; the entry's RefSeq, GeneID and
  ZFIN cross-references settle that, and nothing in the review depends on it.
• Structural details (LRR counts of 15 or 26 in different analyses, flattened horseshoe
  ectodomain) are predictions, and the report itself flags the disagreement.

Other change: the references list contained PMID:18714020 and PMID:24282308 twice each; the
duplicates were removed.

Actions: every existing annotation unchanged, including the UNDECIDED GO:0009617 row, and no
new annotations. Changes are one reason field, one new supported_by quote, one new reference
(the Falcon file) and the reference de-duplication.