MOTS-c (A0A0C5B5G6, UniProt gene symbol MT-RNR1) — curation notes

Folder follows the <HOST>__<ACC> convention for alternative-ORF peptides
(CLAUDE.md, "Alternative-ORF peptides"): MOTS-c is a separate UniProt entry
(A0A0C5B5G6, 16 aa, MOTSC_HUMAN) that UniProt files under the host locus symbol
MT-RNR1, the mitochondrial 12S rRNA gene. It is not an isoform or a cleavage
product of any MT-RNR1 product, so it cannot be modelled as a
functional_isoform of the host.

1. Origin of the peptide, and the one unusual thing about it

The ORF is 51 bp inside the mitochondrial 12S rRNA
PMID:25738459.
Critically, it cannot be made by the mitochondrial ribosome:
PMID:25738459.
The authors therefore posit export of a polyadenylated mitochondrial transcript
to the cytosol, a route they concede is poorly understood. Nuclear NUMT copies
were excluded as the source of the exact peptide
PMID:25738459,
and mtDNA/mtRNA depletion (rho-0 HeLa, actinonin) removed the MOTS-c signal.

So: endogenous existence rests on antibody signal (immunofluorescence,
immunoblot of nuclear extracts), an in-house ELISA, and loss of signal on
mtDNA/mtRNA depletion. No peptide-level mass-spectrometric identification of
endogenous MOTS-c appears in the cached literature. The required cytosolic
translation of a mitochondrially transcribed rRNA-internal ORF remains the
unresolved mechanistic premise of the whole field. Curation consequence:
localisation and molecular-interaction annotations are retained where the assay
was done, but the physiological/metabolic phenotype terms that rest entirely on
injecting or adding synthetic peptide are flagged rather than treated as core.

2. What the peptide is reported to do, by evidence type

Secreted/circulating pool (GO:0005576)

PMID:25738459.
Exercise raises both muscle and plasma levels in human volunteers
PMID:33473109,
[PMID:32182209 — tissue/plasma levels, age-dependent]. This is endogenous-peptide
evidence and is the best-supported location claim.

Mitochondrion / nucleus (GO:0005739, GO:0005634)

Partial mitochondrial co-localisation
PMID:25738459;
PMID:29983246.
Nuclear pool, with endogenous detection and stress dependence
PMID:29983246;
nuclear entry needs the hydrophobic core
PMID:29983246.
Reynolds et al. only tracked exogenous FITC peptide
PMID:33473109,
which is weaker than the Kim 2018 endogenous data but concordant with it.

Nuclear/ARE molecular activities — the strongest molecular claims

• Direct DNA binding by EMSA:
  PMID:29983246.
  Caveat: the probe-binding assays used 0.5–6 µg synthetic peptide, and the
  control mutants (YIFY→AAAA, RKLR→AAAA) remove exactly the hydrophobic and
  cationic residues that would mediate non-specific peptide–nucleic-acid
  interaction. Loss of binding in those mutants therefore does not establish
  sequence specificity; GO:0003677 (DNA binding) is as far as the evidence goes,
  and a sequence-specific/cis-regulatory-region descendant would over-reach.
• Occupancy of ARE-containing promoters in cells (ChIP-qPCR):
  PMID:29983246.
• Transcription-factor binding (GO:0140297): NRF2/NFE2L2
  PMID:29983246;
  ATF1
  PMID:29983246;
  NFE2L1/NRF1 is the weakest of the three
  PMID:29983246.
  All three partners are genuine DNA-binding transcription factors, so
  GO:0140297 is the informative replacement for the bare protein binding row
  that GOA also carries against NFE2L2 from the same paper.

Metabolic effects (Lee 2015) are a cascade, not a molecular activity

Folate-cycle inhibition → AICAR accumulation → AMPK:
PMID:25738459,
PMID:25738459,
PMID:25738459,
PMID:25738459.
No direct molecular target in the folate cycle has been identified; the enzyme
that MOTS-c would inhibit is unknown. Hence:
- GO:0072522 purine-containing compound biosynthetic process (involved_in) is
the wrong direction: MOTS-c blocks de novo purine synthesis. The honest term
is negative regulation (GO:1900372).
- GO:0032147 activation of protein kinase activity describes a two-step
metabolite-mediated consequence (AICAR is the AMPK activator, not MOTS-c), so
it over-states a direct action on the kinase.

Muscle / PTEN–AKT–FOXO1 (Kumagai 2021)

PMID:33554779,
PMID:33554779,
PMID:33554779.
Two problems with the GOA rows from this paper:
1. GO:2001145 names the wrong enzyme chemistry. PTEN is a
phosphatidylinositol-3,4,5-trisphosphate 3-phosphatase (GO:0016314);
GO:2001145 is negative regulation of PIP3 5-phosphatase activity, the
INPP5D/SHIP-family reaction. GO has no "negative regulation of PIP3
3-phosphatase activity" term, so the defensible term is the parent
GO:0010923 negative regulation of phosphatase activity.
2. The muscle-mass effect is marginal and pharmacological:
PMID:33554779,
PMID:33554779.
GO:0048630 skeletal muscle tissue growth therefore rests on a
body-weight-adjusted difference in injected mice — an over-annotation.

Bone (two Eur Rev Med Pharmacol Sci papers)

Both are synthetic-peptide dose-response experiments on cultured cells, one in
rat BMSCs PMID:30468456,
one in the hFOB1.19 line
PMID:31081069.
Neither tests endogenous MOTS-c; GO:0033687 osteoblast proliferation is in
fact a CCK-8 viability readout. Both are flagged as over-annotations: they
describe what the peptide does to cells when applied, not a process the
endogenous peptide is known to participate in.

Antibacterial / host-defence activity — the one well-supported direct function not in GOA

PMID:42611943,
PMID:42611943.
An earlier, independent group reached the same in-vivo conclusion
PMID:29096170.

Participation test (CLAUDE.md): here the peptide itself does the work — it
aggregates and permeabilises bacteria. It is not merely required for, or a
substrate of, someone else's process. Comparator check: the canonical
secreted human antimicrobial peptide CAMP/LL-37 (P49913) carries exactly
GO:0019731 (IDA), GO:0061844 (IDA), GO:0050829 and GO:0050830 — so the
convention for a circulating cationic amphipathic peptide with direct killing
activity is to annotate these terms, and MOTS-c's absence from them is a gap,
not a convention. Proposed as NEW: GO:0061844, GO:0050829, GO:0050830.
Deliberately not proposed: GO:0140911 pore-forming activity — the papers
show aggregation and surface/membrane damage, not transmembrane pore formation.

3. Cross-cutting caveats recorded for the review

• Dose and route. Essentially every functional result uses synthetic peptide
  at 0.5–10 µM in culture (100 µM in the antibacterial assays) or 0.5–5 mg/kg/day
  i.p. in mice. Endogenous-level manipulations do not exist, because mtDNA cannot
  be knocked out gene-specifically; actinonin/rho-0 depletion removes all
  mitochondrial gene expression at once.
• Species. Most in-vivo work is mouse or rat; human data are correlative
  (plasma/muscle levels, the m.1382A>C / K14Q association PMID:33468709).
• Lab concentration. The large majority of mechanistic papers come from the
  Cohen/Lee groups. Independent corroboration exists for the anti-MRSA in-vivo
  effect PMID:29096170 and for exercise/age-related expression
  PMID:32182209, but not for the ARE/NRF2 nuclear mechanism.
• Policy applied. Following CLAUDE.md, no experimental row is REMOVEd for
  being hard to verify. The single REMOVE is the bare protein binding row,
  which the protein-binding policy directs to REMOVE because the same paper and
  the same partner already support the more informative GO:0140297.

4. Action summary (21 GOA rows)

• ACCEPT ×12: GO:0003677; GO:0006357; GO:0140297 ×3 (ATF1, NFE2L1, NFE2L2);
  GO:0005576 ×2 (IDA + the mirroring IEA); GO:0005634 ×3; GO:0005739 ×2.
• MODIFY ×3: GO:0072522 → GO:1900372 (direction reversed); GO:2001145 →
  GO:0010923 (wrong phosphatase chemistry); GO:0032147 → GO:0071902
  (regulation, not activation — AICAR, not MOTS-c, activates AMPK).
• KEEP_AS_NON_CORE ×2: GO:0043610 regulation of carbohydrate utilization;
  GO:0071902 positive regulation of protein ser/thr kinase activity (AKT, via
  the CK2–PTEN–mTORC2 cascade in injected mice).
• MARK_AS_OVER_ANNOTATED ×3: GO:0001649, GO:0033687, GO:0048630.
• REMOVE ×1: GO:0005515 protein binding.
• NEW ×3: GO:0061844, GO:0050829, GO:0050830.

2026-11-03 ORGANELLES batch 4 (mitochondrion rows)

• Rows: GO:0005739 IDA (PMID:29983246, located_in) and IEA (GO_REF:0000044, SL-0173, located_in). Both ACCEPT -> KEEP_AS_NON_CORE.
• Evidence is organelle-level imaging of the endogenous peptide: PMID:29983246; PMID:25738459.
• IEA: SL-0173 rests on the curated UniProt location [file:human/MT-RNR1__A0A0C5B5G6/MT-RNR1__A0A0C5B5G6-uniprot.txt "Mitochondrion {ECO:0000269|PubMed:29983246}"]; mapping sound.
• No transit peptide or TM in the record; no sub-mitochondrial assay. Stays at the root.
• Why non-core: core_functions place the activities in nucleus and extracellular region; nothing shows a MOTS-c activity inside mitochondria. Evidence is about the peptide (antibody/imaging, rho-zero loss), not the host 12S rRNA.