Ntrk1 (TrkA) — Heterocephalus glaber (naked mole-rat), UniProt A0AAX6QC09

Research journal for the GO annotation review. Every assertion below carries provenance.
Quotes are verbatim substrings of the cited cached publication or of the UniProt record in
this folder.

1. What the entry itself is

• TrEMBL entry A0AAX6QC09_HETGA, 798 aa, PE 3: Inferred from homology, derived from RefSeq
  XP_004871305.1. Gene name Ntrk1. No experimental UniProt evidence of any kind.
• UniProt names it generically Tyrosine-protein kinase receptor with
  [file:HETGA/Ntrk1/Ntrk1-uniprot.txt "EC=2.7.10.1 {ECO:0000256|RuleBase:RU000312};"], and
  places it in
  [file:HETGA/Ntrk1/Ntrk1-uniprot.txt "Belongs to the protein kinase superfamily. Tyr protein"]
  kinase family, insulin receptor subfamily.
• The NTRK1-specific identity is nonetheless unambiguous from the domain signatures:
  [file:HETGA/Ntrk1/Ntrk1-uniprot.txt "DR InterPro; IPR020461; NTRK1."] and
  [file:HETGA/Ntrk1/Ntrk1-uniprot.txt "DR PANTHER; PTHR24416:SF370; HIGH AFFINITY NERVE GROWTH FACTOR RECEPTOR; 1."].
• Architecture from the feature table: signal peptide 1–33, LRR + two Ig-like domains
  extracellularly, a single TM helix
  [file:HETGA/Ntrk1/Ntrk1-uniprot.txt "FT TRANSMEM 418..441"], and an intracellular
  protein-kinase domain
  [file:HETGA/Ntrk1/Ntrk1-uniprot.txt "FT DOMAIN 512..783"].

Sequence check of the catalytic and docking machinery

Method: residue positions read directly off the SQ sequence of
genes/HETGA/Ntrk1/Ntrk1-uniprot.txt (798 aa) and cross-checked against the FT lines of the
same file. No alignment was performed; no position is asserted from memory.

Two conclusions follow. (i) The complete catalytic apparatus of a receptor tyrosine kinase —
ATP site, β3 lysine, HRD aspartate, DFG motif, activation-loop tyrosine pair — and both
principal phosphotyrosine docking sites are intact in the naked mole-rat protein. (ii) The
cysteine at 774 is exactly what the primary literature reports for this species (see §2), so
this UniProt entry is the hypofunctional allele, not some divergent assembly artefact.

2. What the naked-mole-rat literature establishes about THIS protein

The decisive paper is Omerbašić et al. 2016, and we have its full text.

2a. The organismal phenotype it explains

Naked mole-rats do not develop NGF- or inflammation-induced heat hyperalgesia
PMID:18232734. Critically, the 2008 phenotype paper already ruled out simple receptor
loss: PMID:18232734, with the effect seen at
PMID:18232734.

This is the single most important sentence in the whole dossier for GO purposes: NGF
binds and signals through naked mole-rat TrkA well enough to drive neurite outgrowth in the
animal's own sensory neurons.

2b. The receptor is expressed on naked mole-rat nociceptors

PMID:27732851; of cultured naked mole-rat sensory neurons the authors
PMID:27732851. Independently confirmed in
2025: PMID:40705105.

2c. The receptor was cloned; sequence divergence maps to the kinase domain

PMID:27732851. Alignment across 27 mammals showed
PMID:27732851 and, importantly,
PMID:27732851

The species-specific change: PMID:27732851 — matching the C at position 774 found in A0AAX6QC09 (§1). The overall claim is
PMID:27732851

Note the numbering caveat. The paper states position 774 with rat numbering as the
reference. That the naked mole-rat UniProt entry also carries a Cys at its own position 774 is
a coincidence of near-identical lengths (798 vs 799 aa), not something the paper asserts. Both
facts are recorded separately above; neither is inferred from the other.

2d. What "hypofunctional" was actually measured on

This is the crux for curation, so the assays are itemised precisely:

1. Native nociceptors. PMID:27732851 (100 ng/mL NGF, 5 min). This is an absence of TRPV1
   sensitisation, not an absence of receptor function.
2. Full-length nmrTrkA in oocytes at 100 ng/mL NGF. PMID:27732851, with equal receptor protein
   PMID:27732851. So: smaller, not absent.
3. Dose–response — the deficit is surmountable. PMID:27732851, the authors concluding
   PMID:27732851. A right-shifted dose–response curve
   that reaches the wild-type maximum is the textbook signature of reduced efficiency, not
   lost function.
4. The lesion is intracellular, not in ligand binding. A chimera (rat ectodomain, naked
   mole-rat TM + kinase domain) reproduced the deficit — summarised by the later review as
   PMID:32206859. Consequently the
   possibility that human recombinant NGF simply binds the naked mole-rat ectodomain poorly was
   experimentally excluded as the explanation.
5. Receptor autophosphorylation. PMID:27732851 — but in the same
   experiment PMID:27732851, which the authors
   attribute to overexpression-driven dimerisation. The naked mole-rat kinase domain is
   therefore catalytically competent; what is blunted is the ligand-driven increment.
6. Downstream phospho-proteome. NGF still upregulated a large phosphopeptide set through
   the naked mole-rat intracellular domain, just fewer of them:
   PMID:27732851.
   Specific effectors: PMID:27732851 with rat TrkA;
   PMID:27732851; and
   PMID:27732851.

Every one of these is a reduced, not abolished, output. The naked mole-rat intracellular
domain does couple NGF stimulation to ERK1/2 and p38α — less efficiently than rat.

2e. Development and survival outputs are preserved

PMID:27732851, and
PMID:27732851.

There is a chronic in-vivo consequence, but it is postnatal and correlative:
PMID:27732851, while at P3
PMID:27732851. The authors are explicit
that this is a postulate, and note the confound that the Mashona mole-rat shares two of the
three kinase-domain variants yet retains C-fibers.

2f. NGF signalling is not globally dead in the adult animal

PMID:27732851. Two independent lines confirm the
narrowness of the defect: the naked mole-rat retains a functional purinergic pain pathway
(PMID:32478202) and its sensory neurons are still sensitised by artemin/GFRα3
PMID:40705105 — i.e. other sensitising pathways work, bounding what the TrkA deficit explains.

3. Curation consequences

3a. A quantitative signalling deficit does not abolish a GO molecular function

GO:0004714 transmembrane receptor protein tyrosine kinase activity,
GO:0005030 neurotrophin receptor activity, GO:0043121 neurotrophin binding,
GO:0005524 ATP binding and GO:0038180 nerve growth factor signaling pathway are all
retained in this species and should be ACCEPTed:

• ligand binding is not the lesion (chimera experiment, §2d.4);
• catalytic competence is demonstrated by basal activation-loop autophosphorylation (§2d.5)
  and by an entirely intact catalytic apparatus in the sequence (§1);
• the receptor reaches rat-equivalent output at 1,000 ng/mL NGF (§2d.3);
• it drives neurite outgrowth in the animal's own neurons at 500 ng/mL (§2a).

REMOVE for any of these would be a category error: GO functions are qualitative capability
statements, and every one of these capabilities is positively demonstrated for the naked
mole-rat protein. The right place to record "less efficient at physiological NGF
concentrations" is review.reason, the description, and suggested_experiments — not the
action field.

3b. "No NGF-induced nociceptor sensitisation" is a phenotype, not a GO function

The organism-level observation is the absence of heat hyperalgesia and the absence of NGF
sensitisation of TRPV1. GO annotates what a gene product does, not what an animal fails to do.
There is no defensible GO term for "fails to sensitise TRPV1", and a NOT annotation would be
wrong too, because the receptor does sensitise TRPV1 — at 1,000 ng/mL NGF (§2d.3) — so the
negation is not qualitative. I have therefore written no annotation encoding the pain
phenotype, and instead recorded it as prose in description and as a suggested_question.

3c. The downstream BP terms, one by one

3d. Why GO:0070374 is added as NEW rather than left implicit

The 2016 SILAC/phosphoproteomics and the anti-phospho-ERK western are the only direct
measurements anywhere of what the naked mole-rat TrkA intracellular domain does downstream,
and they show ERK1/2 activation on NGF stimulation (reduced relative to rat, but present, §2d.6).
Leaving that unannotated would discard the one downstream branch this species has actual data
for while keeping the PI3K/AKT branch that it has none for. The caveat that the assay used a
rat-ectodomain chimera in HEK293 cells is recorded in the annotation's reason.

3e. Root-level and redundant parents

GO:0000166 (nucleotide binding), GO:0016740 (transferase activity), GO:0016301 (kinase
activity), GO:0004672 (protein kinase activity), GO:0004713 (protein tyrosine kinase
activity), GO:0016020 (membrane), GO:0005770 (late endosome) and GO:0007169 (cell surface
RTK signaling pathway) are all true, all uninformative, and all have a strictly more specific
sibling already present in the same GOA set. Each is MODIFYed onto that specific term. None
is removed: they are not wrong, merely shallow.

3f. One label discrepancy, deliberately left asymmetric

GO:0043235 appears in the review under two different labels, and this is intentional. GOA
and the live QuickGO record both call it "signaling receptor complex", and that is the
label carried on the existing annotation, whose ids and labels are machine-sourced and must
not be rewritten. The repository's local ontology cache is a version behind and still calls it
"receptor complex"; core_functions term labels are hard-checked against that cache, so
the in_complex slot uses the older label to avoid a spurious validation warning. Same term,
same claim, two ontology releases.

3g. Validation status

just validate HETGA Ntrk1 passes with no errors and one warning, deliberately left:
"No annotations reference available deep research files." Clearing it would require putting
the affinage record into an annotation's supported_by with a quoted sentence, which the
review brief explicitly forbids — a provider sentence is a lead, not evidence, and this
particular record is about the human orthologue. The affinage file is instead cited through
additional_reference_ids on the four annotations whose family-level basis it actually
supplies (early/late/recycling endosome membrane and PI3K/AKT), and assessed in full in
references[].reference_review. The warning is the correct outcome here, not a defect.

4. Assessment of the affinage human-ortholog record

Ntrk1-deep-research-affinage-human-ortholog.md is the Affinage record for human NTRK1
(P04629), deliberately fetched as a conserved-mechanism baseline because Affinage refuses
non-human species.

What it is good for. An excellent, densely cited mechanistic map of the conserved receptor:
the D5/IgC2 NGF-binding hot spot, ordered bi-bi kinetics, Y490/Y785 docking, IRS-1/2 → p85-PI3K
recruitment, N-glycosylation gating surface delivery, STX8 Golgi→PM transport, GGA3/Arf6
recycling sustaining NGF-induced AKT, retrograde MVB-derived signalling endosomes, and
TRAF4/Nedd4-2 ubiquitin control. That corpus is what makes the endosome-membrane and
receptor-complex annotations in this GOA set credible as family-level transfers.

What it decisively missed — and this is the recall observation. The record contains zero
occurrences of "naked mole-rat", "Heterocephalus", or PMID:27732851 across all 34 citations.
It therefore misses, entirely: the hypofunctional allele, the L→C substitution at 774, the
right-shifted NGF dose–response, the reduced ERK1/2 and p38α coupling, and the whole reason
this gene is interesting in this species. This is not a defect of the provider — it is
human-only by construction — but it means the affinage record is useless as the evidential
basis for any species-specific call here, and every naked-mole-rat claim in this review is
anchored to PMID:27732851 / PMID:18232734 / PMID:32206859 / PMID:40705105 instead.

One genuinely useful lead it did surface (recorded as a lead, not as evidence, because the
paper is not in publications/): PMID:24623787, a knock-in mouse deleting the KFG
juxtamembrane element, which raises TrkA levels/activity and produces enhanced thermal
sensitivity and inflammatory pain without changing DRG neuron numbers. That is the exact
mirror image of the naked mole-rat: TrkA output up → thermal hyperalgesia up; TrkA output down
→ thermal hyperalgesia absent. It is a strong independent argument that TrkA signalling
strength is the rate-limiting variable for thermal hyperalgesia, which is precisely the
2016 paper's thesis.

Do not import its mechanism_profile GO ids. They collapse to coarse parents
(GO:0140096, GO:0016740, GO:0140657, GO:0060089, GO:0048018) — and one of them,
GO:0048018 receptor ligand activity, is flatly wrong for a receptor: TrkA is the receptor,
not the ligand. Every term in this review was grounded independently against QuickGO.

5. What I could not resolve

1. NGF binding affinity of the naked mole-rat ectodomain was never measured. The chimera
   experiment shows the ectodomain is not the cause of hypofunction, but no Kd, Scatchard, or
   SPR measurement on naked mole-rat TrkA exists. GO:0043121 neurotrophin binding is accepted
   on functional grounds (NGF-driven neurite outgrowth, and rat-equivalent sensitisation at
   1,000 ng/mL), not on a direct binding assay. Recorded as a suggested_experiment.
2. PI3K/AKT output in naked mole-rat TrkA is entirely unmeasured — see §3c. Kept, not core.
3. The subcellular-trafficking annotations (early/late/recycling endosome membrane, axon,
   receptor complex) have no naked-mole-rat evidence of any kind. They are ARBA / UniProt-SubCell
   / TreeGrafter transfers resting on the human/rat literature summarised in the affinage record.
   Nothing in the naked-mole-rat data contradicts them and the receptor is demonstrably surface-
   expressed and NGF-responsive, so they are kept as non-core rather than removed. If the L774C
   change altered receptor trafficking or turnover this is exactly where it would show, and no
   one has looked.
4. Whether the adult C-fibre loss is apoptotic is not established — the 2016 paper says
   "postulate". So GO:0043524 is accepted on the conserved neurotrophic-survival function plus
   the explicit statement that embryonic survival is supported, not on the C-fibre data.
5. Which of the one-to-three kinase-domain substitutions is causal is unresolved by the
   paper itself; no single-variant reversion experiment was done.