Tac1 (Heterocephalus glaber, A0A0P6JY17) — review journal

Protachykinin-1 (PPT / PPT-A), the precursor of substance P and neurokinin A.
UniProt TrEMBL, 130 aa, PE 3: Inferred from homology. Five GOA rows, all IEA,
all electronic; the naked mole rat has essentially no experimental GO annotation.

1. The framing problem this gene sets

The naked mole rat is repeatedly described in the secondary literature — and in the
title of one of the primary papers — as an animal that "naturally lacks substance P".
Taken at face value that would be a strong claim about this gene product, and it
would push a reviewer towards REMOVE on the molecular-function annotation.

It is an over-simplification, and the primary papers are careful in a way that their
own titles are not. Every primary source qualifies the absence by tissue:

And the absence is quantitative rather than absolute even within that compartment:

The decisive sentence, which appears in the discussion of the very paper whose title
says "an animal naturally lacking substance P", is this:

So the accurate statement is: substance P immunoreactivity is absent from cutaneous
sensory fibres and small-diameter DRG/TG somata; it is present elsewhere, including
visceral afferents. Corroborating, in the colon: PMID:31992138. The
peptidergic system is regionally reorganised, not abolished.

The "myths" paper (PMID:34476892) does not list "naked mole-rats lack substance P"
among its 28 myths, so it does not adjudicate this claim directly. It does treat the
adjacent over-statement in exactly the way argued here — Myth 10 is "naked mole-rats
feel no pain", and the response emphasises the word selective: PMID:34476892. It also
records that naked mole-rats PMID:34476892 — diminished, and condition-dependent, not absent.

2. The gene and the protein are intact — sequence evidence

The TrEMBL record annotates no PEPTIDE features at all, only SIGNAL 1..19,
CHAIN 20..130, and two SMART Tachykinin domain calls (58..68, 97..107). It
therefore does not itself say whether substance P is in this isoform. I resolved this
directly from the sequence
(file:HETGA/Tac1/Tac1-bioinformatics/RESULTS.md):

peptide coords naked mole-rat identical to human?
Substance P 58–68 RPKPQQFFGLM yes
Neurokinin A 98–107 HKTDSFVGLM yes
Neuropeptide K 72–107 EADSSIEKQ…FVGLM D72E only
Neuropeptide gamma, 2nd part 89–107 GHGQISHKRHKTDSFVGLM yes
C-terminal-flanking peptide 111–126 ALNSVAYERNAMQNYE S120N only

The two precursors are the same length bar one C-terminal residue, so the comparison
is gap-free: 125/129 identical (96.9%), substitutions A6V, Y35S, D72E, S120N, none of
which touch substance P or neurokinin A. Both tachykinin cores end in the F-x-G-L-M
consensus that the GO definition of GO:0007217 uses to define a tachykinin, and
both are immediately followed by G-K-R — the Gly that donates the obligatory
C-terminal amide plus the dibasic prohormone-convertase site. UniProt's own keywords
agree that processing is expected: [file:HETGA/Tac1/Tac1-uniprot.txt "Cleavage on pair
of basic residues"] and the Amidation keyword.

Three consequences:

  1. This entry has the beta-preprotachykinin architecture — it encodes both
    substance P and neurokinin A, so the answer to "is substance P in this isoform?" is
    yes. (The alpha form would encode substance P only.) I am not claiming which
    splice forms the naked mole rat actually transcribes: the entry declares no
    ALTERNATIVE PRODUCTS, and all three cross-referenced RefSeq proteins carry this one
    sequence. Alpha/beta/gamma/delta usage in this species is unestablished.
  2. The gene is transcribed: [file:HETGA/Tac1/Tac1-uniprot.txt "Expressed in hypothalamus
    and 3 other cell types or tissues."] (Bgee, ENSHGLG00000001397).
  3. Nothing in the protein is broken. The "lack of substance P" is a statement about
    where the peptide is made, not about what the protein is or does.

3. The naked-mole-rat literature positively confirms the projected function

This is the unusual and pleasing part of this gene: the species-specific literature
does not merely fail to contradict the electronic annotations, it actively confirms
them, because researchers have repeatedly added substance P back and watched the
pathway work.

Mechanism, as understood by the authors of the naked-mole-rat work:
PMID:21200438 — this is a general
mechanistic statement citing prior work, not a naked-mole-rat measurement, and I have
treated it as such. It agrees with UniProt for this entry:
[file:HETGA/Tac1/Tac1-uniprot.txt "activation of G(q) and phosphatidylinositol"]
hydrolysis by phospholipase C.

4. The curation question: does a tissue-restricted absence change a GO annotation?

This is the interesting call, so I want the argument on the record.

No, and it should not. GO annotates what a gene product is and does, not where a
cell chooses to express it. Three points make this concrete here:

  1. The annotations under review are about the protein's activity and pathway
    membership.     GO:0031835 (substance P receptor binding), GO:0007217 (tachykinin
    receptor signalling pathway) and GO:0005576 (extracellular region) are claims about
    the encoded peptide. Every one of them survives the naked-mole-rat evidence intact:
    the peptide is human-identical, the processing signals are conserved, the receptor is
    present, and the receptor responds to the peptide with the expected pharmacology.
  2. The naked-mole-rat finding is an expression-pattern finding. "Absent from
    cutaneous C-fibres" is IEP-shaped information about a cell type. GO has no slot for
    it on the gene product's MF, and encoding it as REMOVE would assert something
    false — that this protein cannot bind an NK1 receptor — which the SP add-back
    experiments directly refute.
  3. The rescue experiments are the decisive test. If the annotation were wrong, adding
    the peptide (or the gene) back would do nothing. It does the opposite: it restores
    itch, capsaicin nocifensive behaviour, thermal hyperalgesia and formalin responses,
    all NK1R-dependent. That is a positive confirmation of the projected annotations, in
    this species, by intervention.

Where the naked-mole-rat finding does legitimately bear on curation is on the
organism-level physiology term, GO:0006954 inflammatory response. In other
mammals a large part of the tachykinin contribution to inflammation is neurogenic —
substance P released from peptidergic cutaneous C-fibre terminals
(PMID:32478202) —
and that arm is precisely the one documented to be hypofunctional here
(PMID:32478202, and
the selective inflammatory-pain insensitivity of PMID:18232734). The term is still
defensible — the capability is retained and is unmasked by SP add-back — but it is not
a core function of this protein in this animal. Hence KEEP_AS_NON_CORE, not
REMOVE and not ACCEPT.

The right home for the species-specific finding is therefore: this notes file, the
description, and suggested_questions/suggested_experiments — plus one explicit
sentence in each affected review.reason so a future curator does not re-litigate it.

5. Actions taken

GO term evidence action one-line reason
GO:0005576 extracellular region IEA, GO_REF:0000044 (SubCell SL-0243) ACCEPT signal peptide 1..19, UniProt Secreted; mature tachykinins act on cell-surface GPCRs
GO:0006954 inflammatory response IEA, GO_REF:0000118 (PTN000134172) KEEP_AS_NON_CORE capability retained (SP add-back enhances formalin response) but the cutaneous neurogenic arm is hypofunctional in this species
GO:0007217 tachykinin receptor signalling pathway IEA, GO_REF:0000120 ACCEPT both encoded peptides literally satisfy the term's F-x-G-L-M definition; NK1R present and antagonist-reversible in NMR
GO:0031835 substance P receptor binding IEA, GO_REF:0000118 (PTN000134172) ACCEPT NMR substance P is human-identical; NMR NK1R responds and is blocked by GR-82334
GO:0007204 positive regulation of cytosolic Ca²⁺ IEA, GO_REF:0000120 (from human P20366) KEEP_AS_NON_CORE correct and well-grounded (human IDA + PAINT) but a generic downstream readout of Gq–PLC, not what distinguishes this protein
GO:0031837 substance K receptor binding NEW (ISS) NEW the set covers only the SP/NK1R arm; the precursor also encodes neurokinin A and UniProt states it is a TACR2 ligand
GO:0007209 PLC-activating tachykinin receptor signalling NEW (ISS) NEW mechanistic refinement of GO:0007217 that makes the GO:0007204 calcium claim interpretable; human ortholog carries it

No REMOVE, no MODIFY, no UNDECIDED. Every projected term traces to a PAINT node
(GO_REF:0000033) on human P20366 with experimental grounding beneath it, and
GO:0007204 additionally has a human IDA (PMID:8957234, not cached) — I checked the
human annotation set via QuickGO rather than assuming.

6. Deep-research providers: falcon (species-specific) vs affinage (human ortholog)

Two provider records are present in the folder, and they perform very differently.

6a. The falcon report — good, and it converged independently

Tac1-deep-research-falcon.md is a genuine Heterocephalus glaber report (Edison,
2026-09-02, 23 citations). I read it after forming the argument in §4 from the primary
papers, and it reaches the same conclusion by the same route:

It also handles the identity check properly — it explicitly separates Tac1 from Tac2 and
does not drift onto the C. elegans / Candida / rice homonyms that swamp the affinage
record. And it flags its own orthology-based inferences as inferences, e.g. on the
neurokinin A arm: [file:HETGA/Tac1/Tac1-deep-research-falcon.md "Production of NKA or
extended NKA peptides is biologically plausible from mammalian conservation but should
remain annotated as predicted until naked-mole-rat transcript isoforms and endogenous
processed peptides are established by long-read RNA sequencing and targeted peptidomics."]
That is exactly the standard I applied to the GO:0031837 proposal.

Two leads it supplied that the cached corpus does not contain, recorded here rather than
curated because neither is verifiable from the local cache:

  1. Zhao, Lee, Gryszkiewicz & Park (2025), "Life without substance P: The naked mole
    rat"     (Elsevier, pp. 275-290, doi:10.1016/b978-0-443-22194-1.00013-6) — a book chapter
    devoted to precisely the question this review turns on. Not in publications/, no
    PMID recovered, so nothing here rests on it; it should be fetched before any future
    revision of this gene.
  2. CAP-TAC1 (Wiggenhorn et al. 2023, Nat Commun, doi:10.1038/s41467-023-43857-0) —
    [file:HETGA/Tac1/Tac1-deep-research-falcon.md "an N-terminally pyroglutamylated and
    C-terminally amidated TAC1-derived circulating peptide"] reported as a nanomolar
    agonist at multiple tachykinin receptors in mouse and human plasma. This is a third
    class of product from the TAC1 locus, beyond substance P and the neurokinin A series.
    The report is explicit that [file:HETGA/Tac1/Tac1-deep-research-falcon.md "CAP-TAC1 has
    not been demonstrated in H. glaber"], so it is not annotated here, but it raises
    a real question: an animal described as lacking substance P in skin could still carry a
    systemic tachykinin tone nobody has measured. Added to suggested_questions.

Reservation: the report quotes some numbers loosely (it renders the intrathecal substance
P doses from PMID:21200438 as "1-10 mM" and "100 mM" where the paper says µM), so its
figures should not be copied without checking the source. Nothing in this review depends
on a number taken from it.

7. What the affinage human-ortholog record missed

The provided Tac1-deep-research-affinage-human-ortholog.md is a human-TAC1 record,
used here only as a conserved-mechanism baseline. Two problems worth recording, because
this is how provider recall gets measured:

  1. It contains no naked-mole-rat content whatsoever — which the falcon report, run
    on the same gene, shows was avoidable rather than inherent to the question. Not one of its 36 citations
    is a naked-mole-rat paper, and none of the six decisive papers used above appears in
    it. Everything that actually decides this review — the tissue-restricted absence, the
    visceral SP/CGRP fibres, the intrathecal SP rescues, the HSV-PPT gene add-back — was
    invisible to it. That is expected (affinage is human-only) but it means the record
    contributed nothing to the curation calls, only to background.
  2. Its corpus is contaminated by symbol collisions and it says so itself: the record
    mixes mammalian TAC1 (the neuropeptide precursor) with C. elegans TAC-1 (a TACC
    microtubule-associated protein), Candida albicans/C. parapsilosis Tac1 (a
    zinc-cluster transcription factor) and rice TAC1 (tiller angle control). Roughly a
    third of its "dated findings" table describes proteins unrelated to this gene. It
    flags this in its narrative, but a downstream consumer that imported the table
    wholesale would produce nonsense.
  3. Its mechanism_profile grounding (GO:0048018 receptor ligand activity,
    GO:0005576 extracellular region) is not wrong but is coarser than the existing GOA
    set, and per the brief it was not imported. I re-grounded every term independently
    against QuickGO.

Where it was useful: the mouse Tac1-knockout phenotype literature it collects
(anxiety/depression PMID:12427862, nociceptor mechanical sensitization PMID:31012376,
energy balance PMID:28775376, prolactin secretagogue PMID:20434341) is a good reminder
that this gene is pleiotropic well beyond nociception, which is what motivates the
hypothalamic-expression question in suggested_questions.

8. Unresolved / knowledge gaps