TMC1 review notes

Why this gene was selected

Contested/newly-assigned molecular function. Two questions are live at once:

1. Is TMC1 an ion channel at all, or only a component of the transduction apparatus? This was a
   genuine long-running dispute, driven by the failure of TMC1 to reach the plasma membrane in
   heterologous cells. The 2024/2025 Neuron paper that largely settles it says so in its own abstract:
   PMID:39674179
2. Is TMC1 also a lipid scramblase? Structurally the same question as TMEM63A/TMEM63B, which a
   sibling review in this batch handled, and which the TMC field itself frames the same way, because
   TMC1 was modelled on the TMEM16 fold:
   PMID:30063209

Position taken

Channel: core, and no longer contested. Scramblase: process asserted, activity withheld.

The asymmetry is deliberate and is the point of the review.

The channel assignment

Genetic necessity, with normal hair-bundle structure so the phenotype is not developmental:
PMID:22105175 and rescued by either paralogue
PMID:22105175

Permeation properties tracked to TMC subunit identity:
PMID:23871232 and PMID:23871232

Pore mapping by cysteine modification in hair cells:
PMID:30138589

Six pore-region deafness mutations, all altering Ca2+ permeability (PMID:36191207).

The remaining gap - mechanical gating of the human protein in a naive cell - was closed by forcing
TMC1 to the plasma membrane with a CRISPRi screen:
PMID:39674179, with a pore-identity argument
PMID:39674179 and the conclusion
PMID:39674179

The scramblase assignment - why it is handled differently from TMEM63B

The TMEM63B review in this batch made both activities core, on the grounds that four independent
peer-reviewed groups using four different assay systems reported scrambling, and that disease variants
could dissociate the two activities. That evidence base does not exist for TMC1. What exists is:

• Peer-reviewed genetics showing necessity, not agency.
  PMID:40073458 and PMID:40073458 The title is exact about this - "necessary for scramblase activity" -
  and the authors decline to go further:
  PMID:40073458
  Note also that the Tmie mutant result cuts both ways: hair cells that still express TMC1/TMC2 but
  mislocalise them show no PS externalisation, so the requirement is for functional, correctly placed
  TMCs - which is what you would expect either of a direct scramblase or of an upstream permissive role.
• One direct translocation assay, in a preprint.
  PMID:40631239 and
  PMID:40631239
  This is bioRxiv 2025.07.03.663083 and was still a preprint when this review was written (no journal
  version in PubMed as of September 2026). It is flagged as such in reference_review.

So the curation call is: assert GO:0017121 plasma membrane phospholipid scrambling involved_in
(NEW, IMP, PMID:40073458) because necessity in hair cells is established in peer-reviewed work; do
not assert GO:0017128 phospholipid scramblase activity enables, because the only evidence
that TMC1 performs the translocation is a preprint and the peer-reviewed authors explicitly leave the
mechanism open. "Necessary for" is not "performs". Dual function remains a legitimate possible outcome -
the TMEM16 and TMEM63/OSCA precedents show a single groove can pass both ions and lipids - it is simply
not yet established here.

A further caveat recorded in suggested_questions: in hair cells PS externalisation is triggered by
blocking transduction and by deafness alleles, so it may report hair-cell distress rather than a normal
physiological process. If so, even the BP annotation describes a disease mechanism.

Term-choice problem found in the existing annotation set

GO:0005245 voltage-gated calcium channel activity - wrong gating stimulus

TMC1 carries this term twice (IBA GO_REF:0000033; IEA GO_REF:0000107 from mouse). TMC1 is gated by
tension through the PCDH15 tip link and has no voltage sensor. Tracing the propagation:

• Mouse Tmc1 (MGI:MGI:2151016 / UniProtKB:Q8R4P5) carries enables GO:0005245 with IMP and IGI from
  PMID:23871232, a paper that measured calcium permeability and single-channel conductance of
  the mechanotransduction channel. It makes no voltage-gating claim.
• The same mouse record also carries enables GO:0005262 calcium channel activity from the same paper,
  which is the term the data actually support.
• GO:0005245 is a child of GO:0005262 that adds a gating mechanism. So this is a GRANULARITY_MISMATCH
  under TERM_SCOPING_PROBLEM: the child overstates specificity in a way that is not just uninformative
  but factually wrong about the mechanism.

Action: MODIFY both rows to GO:0140135 mechanosensitive monoatomic cation channel activity +
GO:0005262 calcium channel activity, with structured propagation_review on the IBA.

This is the same shape of error the TMEM63B review found (GO:0005227 calcium-activated cation
channel activity for a stretch-gated, calcium-permeable channel): a permeation property written
into a gating term.

GO:0008381 left as ACCEPT rather than tightened

GO:0140135 (cation) would fit the mammalian MET channel better than GO:0008381 (any ion). But the
IBA sits at a node that also spans invertebrate TMCs, and per project rules an IBA encodes a
phylogenetic curator's judgement about where a function arose; tightening the term at the human leaf
would misrepresent what the node asserts. Left as ACCEPT with the point recorded in reason.

GO:0009897 external side of plasma membrane

Orthology transfer of a mouse IDA (PMID:16455951). TMC1 is a ten-TM protein with cytoplasmic termini,
so the term sits oddly, but this is an experimental curator call whose full text was not read here, so
per project rules it is KEEP_AS_NON_CORE, not REMOVE.

No complex term at all

Two MF rows use contributes_to, which presupposes a complex, yet GOA gives TMC1 no complex membership.
Added GO:0034703 cation channel complex as a NEW part_of annotation (IDA, PMID:34089643, which
defines CIB2/CIB3 as MET channel auxiliary subunits), and proposed a dedicated
mechanoelectrical transduction channel complex term, since GO:0034703 does not distinguish this
assembly from any other cation channel.

PMIDs verified

All PMIDs cited were checked against PubMed metadata before use. Three initial guesses were wrong
and were discarded rather than cited: 29804837 (a V(D)J recombination paper, not Pan et al. 2018),
25683721 (phylogenetic profiling) and 35768512 (enteric viruses). The correct Pan et al. 2018 Neuron
pore paper is PMID:30138589.