MIM1 (YOL026C / TOM13), Saccharomyces cerevisiae — research notes

UniProt: Q08176 · SGD: S000005386 · 113 aa, 12.8 kDa · single predicted TM helix (residues 39–62)
Pfam PF08219 (TOM13) · InterPro IPR013262 (OMP_MIM1/TOM13_mt) · PANTHER PTHR28241 / PTHR28241:SF1
ComplexPortal CPX-2281 (MIM mitochondrial import complex)

Provenance note. All automated deep-research providers failed in this environment
(falcon → HTTP 402 Payment Required; perplexity → provider not configured; openai →
invalid API key), so no MIM1-deep-research-<provider>.md file exists. The notes below
were assembled by hand from the cached publications in publications/ and from the
UniProt record. Every assertion carries an inline citation with verbatim supporting text.

1. Identity and discovery

YOL026C was independently named twice in 2004–2005:

• Tom13, in a screen for essential outer-membrane proteins of unknown function
  PMID:15326197.
• Mim1 (mitochondrial import 1), in a proteomic characterisation of the Neurospora
  crassa outer membrane whose 14.2 kDa hit was traced to the S. cerevisiae orthologue
  PMID:15608614.

Mim1 is the accepted name; Tom13 survives as a synonym and as the Pfam family name. Note
that despite the "Tom13" name, Mim1 is not a TOM-complex subunit — see §3.

2. Localisation and topology

Mitochondrial outer membrane, integral, with the N-terminus facing the cytosol.

• Protease-protection and alkaline-carbonate extraction of isolated mitochondria place
  Tom13/Mim1 as an integral outer-membrane protein
  PMID:15326197.
• Independently confirmed by protease shaving of mitochondria carrying an N-terminally
  His-tagged allele: "Mim1 was degraded and could no longer be immunodecorated with
  antibodies against the His tag ( Fig 1E ). Hence, the N terminus of the protein is
  exposed to the cytosol." PMID:15608614.
• The topology is cytosolic-N / IMS-C with one central α-helical TM segment
  PMID:28916712.

UniProt records Mitochondrion outer membrane with ECO:0000269 from both PMID:15326197
and PMID:15608614. This is one of the best-supported facts about the protein.

3. The MIM complex

Mim1 self-associates into an oligomer that is the bulk of a ~400–450 kDa complex, distinct
from both TOM and TOB/SAM:

• PMID:15608614 and
  "Therefore, Mim1 is not a Tom or Tob component."
• PMID:28916712.

The second, low-abundance subunit is Mim2 (Q3E798, ~10 kDa, same topology, 1–2 copies
per complex):

• PMID:22467864 — this is the reference behind both the GO:0140595 MIM complex
  IPI annotations and the GO:0005515 protein binding IPI (WITH UniProtKB:Q3E798).
• PMID:28916712

GO:0140595 MIM complex is definitionally about exactly this: "A protein complex located
in the mitochondrial outer membrane that functions as an insertase, mediating the insertion
of alpha-helical proteins from the cytosol into the outer membrane."

4. Biological role — insertase for α-helical outer-membrane proteins

The MIM complex is the outer-membrane insertase for α-helically anchored precursors. Client
classes established over ~15 years:

• PMID:22467864
• PMID:32348752 and "We report that the yeast MIM\ncomplex promotes
  the insertion of proteins with N-terminal (signal-anchored) or\nC-terminal
  (tail-anchored) membrane anchors."
• The 2020 study concludes "the MIM complex\nis a major and versatile protein translocase
  of the mitochondrial outer\nmembrane" PMID:32348752.

MIM also cooperates with the TOM receptor Tom70 to accept precursors — PMID:32348752 — and
UniProt records a direct Mim1–Tom70 interaction from PMID:21825073.

5. Role in TOM complex assembly

This is the phenotype through which Mim1 was discovered, and it is largely a consequence
of the insertase activity acting on Tom20/Tom70 plus a late, less well-defined step in
Tom40 maturation.

• Depletion abrogates TOM assembly: PMID:15608614; the block is downstream of TOB/SAM-mediated Tom40 insertion — "The
  steps in the assembly of the TOM complex requiring Mim1 are located after the
  TOB-complex-dependent insertion of Tom40 into the outer membrane."
• Independently: PMID:15326197
• Receptor specificity: PMID:17974559 — Tom22, being
  tail-anchored and requiring pre-existing Tom receptors, is Mim1-independent here.
• Mim1 physically joins SAM: PMID:17974559, echoed in 2020 as "coupling of MIM and SAM\npromotes early assembly
  steps of TOM subunits" PMID:32348752.
• Disruption phenotype (UniProt, ECO:0000269|PubMed:17974559): "Leads to reduced levels of
  TOM complex but retains the respiratory chain."

Importantly, the specificity is real, not a generic import collapse: β-barrel machineries
other than TOM are unaffected — PMID:15608614, and depletion does not impair import of matrix or inner
membrane clients (mtHsp60, AAC) PMID:15326197.

6. The channel activity (basis of GO:0022832) — read carefully

PMID:28916712 screened yeast outer-membrane proteins for channel activity and reported
Mim1 as one of four new channels. The relevant measurements, verbatim:

• "Upon reconstitution into planar lipid bilayers, Mim1 exhibited a characteristic channel
  activity that was inhibited by Mim1-specific antibodies"
• "We observed a main conductance state of ΔG¯main=580 pS that closed upon application of
  high positive or negative voltages"
• "The Mim1 channel was cation selective with a ratio of PK+/PCl− of 23.5:1 based on a
  positive reversal potential of 53 mV"
• Co-reconstitution with Mim2 reduced selectivity to PK+/PCl− = 11:1 — the channel is
  MIM-complex-relevant, not an artefact of Mim1 alone.

Assessment for curation:

1. The observation is solid. Recombinant protein, antibody-inhibited, reproduced with a
   second expression system ("Mim1 synthesized in wheat germ lysate exhibited the same
   channel activity"), modulated by its physiological partner Mim2. SGD's IDA is warranted.
2. The GO term is imprecise in two ways. GO:0022832 voltage-gated channel activity is
   an unspecified-substrate voltage-gated channel term. The measured behaviour is (a)
   strongly cation-selective (23.5:1) and (b) voltage-dependent only in the sense that
   the pore closes at high |Vm| — the classic gating description for a protein-conducting
   channel (Tom40, Sam50, Mdm10 all behave this way), not for a signalling voltage sensor.
   GO:0022843 voltage-gated monoatomic cation channel activity is the substrate-correct
   child if an ion-channel framing is kept.
3. The biological reading is protein conduction, not ion conduction. The authors' own
   framing is "The mitochondrial import component Mim1 forms a channel that is predicted to
   have an α-helical structure for protein import." No physiological ion flux through Mim1
   has been demonstrated; the cation preference is shared with every known protein-import
   channel of this membrane, whose in vivo cargo is polypeptide. Treating GO:0022832 as
   MIM1's molecular function would misstate what the protein does.

Downstream consequence: the GO:0055085 transmembrane transport IEA (GO_REF:0000108) is
generated logically from GO:0022832 via an MF→BP inter-ontology link (WITH/FROM field is
literally GO:0022832). It inherits whatever imprecision GO:0022832 carries, and adds
nothing — "transmembrane transport" is far too general to describe MIM1 and does not capture
protein insertion.

7. What is missing from GOA

No molecular function term describes what Mim1 actually does. The only MF annotations
are GO:0005515 protein binding (uninformative by project convention) and GO:0022832
(imprecise, see §6). GO:0032977 membrane insertase activity — "Binds transmembrane
domain-containing proteins and mediates their integration into a membrane" — is an exact
description of the MIM complex's activity and is supported by PMID:32348752,
PMID:21825073 and PMID:22467864. It should be proposed.

Because insertase activity is delivered by the Mim1–Mim2 heterocomplex rather than Mim1
alone, contributes_to is the appropriate qualifier at the gene-product level.

8. Orthology / PANTHER

PANTHER PTHR28241 "MITOCHONDRIAL IMPORT PROTEIN 1", 1248 proteins across 2583 taxa, one
subfamily (SF1). The IBA annotations (GO_REF:0000033) for GO:0140595 and GO:0045040
derive from node PANTHER:PTN002000670, and MIM1's own SGD annotation
(SGD:S000005386) appears in the WITH/FROM — i.e. S. cerevisiae Mim1 is itself one of
the experimentally characterised descendants the PAINT curator used to place the IBD. That
is expected and is a marker of experimental grounding, not circularity. The family is
fungal/eukaryote-restricted; the functional counterpart in higher eukaryotes (MTCH1/2)
lacks sequence homology and is a case of convergent evolution, which is why the family is
narrow.

9. Summary of curation judgements

10. Addendum — reviewer adjudication (annotation-reviewer pass)

The provisional table in §9 was independently re-checked against the cached full texts.
Two amendments, plus two corrections of detail.

Amendment 1 — GO:0022832 is MARK_AS_OVER_ANNOTATED, not MODIFY → GO:0022843.
The substrate-correctness argument in §6.2 is right as far as it goes, but MODIFY to the
cation-specific child moves the annotation in the wrong direction overall. GO:0022843
asserts that MIM1's molecular function is to conduct monoatomic cations; GO:0022832 leaves
the substrate open. Since §6.3 concludes that the physiological permeant is polypeptide, the
vaguer parent is the less false of the two. There is also a mechanical consequence:
GO:0022832 already generates the spurious GO:0055085 BP by inter-ontology inference
(GO_REF:0000108), and GO:0022843 would generate a more specific and more definitely wrong
cation-transport process in its place. KEEP_AS_NON_CORE was also considered and rejected —
it implies a genuine secondary activity, whereas the measured pore is the same conduit that
performs the insertase function, observed with an electrode.

Amendment 2 — GO:0032977 is entered as a NEW annotation, not only as a proposal.
contributes_to is retained as the qualifier, per §7. In addition, a genuinely new ontology
term is proposed: protein-conducting channel activity (parent GO:0015267 channel activity). No such term exists — checked against OLS — and its absence is exactly why the
electrophysiology of Tom40, Sam50, Mdm10 and Mim1 keeps being curated with ion-channel terms.

Correction A — quotation error in §4. The sentence quoted as
"MIM interacts with TOM to accept precursor proteins from\nthe receptor Tom70." is not a
verbatim substring of publications/PMID_32348752.md: the cached text has a non-breaking
space, to\xa0accept. Use
"Free MIM complexes insert single-spanning proteins that are\nimported in a Tom70-independent manner."
instead, which verifies.

Correction B — topology is not fully settled. PMID:15326197, using a C-terminally
tagged allele, states "These results indicate that Tom13 and Tom38 are an integral membrane
protein and a peripheral membrane protein, respectively, of the mitochondrial outer membrane
and expose at least their COOH termini to the cytosol" — i.e. C-terminus cytosolic, the
opposite of the later consensus recorded in §2. The outer-membrane assignment is unaffected
(both papers agree), but the discrepancy is real and is now recorded as a suggested question.
Note also that the "residues 39–62" TM span in the header is UniProt FT TRANSMEM 39..62
with ECO:0000255 — a sequence-analysis prediction, not experimental.

Final actions (14 GOA-derived entries + 1 NEW). ACCEPT ×11 (GO:0140595 IBA; GO:0045040
IBA; GO:0005741 IEA; GO:0005741 IDA ×3; GO:0045040 IMP; GO:0045040 IDA; GO:0140595 IPI;
GO:0070096 IMP ×2) · MARK_AS_OVER_ANNOTATED ×2 (GO:0005515 IPI; GO:0022832 IDA) ·
REMOVE ×1 (GO:0055085 IEA) · NEW ×1 (GO:0032977, contributes_to).

11. Addendum 2 — five UniProt-cited references folded in

fetch-gene seeds references: from the GOA only, so five papers that UniProt cites with
ECO:0000269 on Q08176 were absent from the first pass. None is in the GOA, so none
supports an existing annotation, but two change what the review can say. All are now cached
and cited.

PMID:18177669 (Popov-Čeleketić, Waizenegger & Rapaport 2008) — the most valuable of
the five. Two contributions:

1. Settles the topology. It states the arrangement directly — "The protein is composed
   of an N-terminal cytosolic domain, a\ncentral putative transmembrane segment (TMS) and a
   C-terminal domain facing the\nintermembrane space." This is the primary source that
   PMID:28916712 was merely restating. Combined with PMID:15608614's note that "a
   C-terminally tagged Mim1 had compromised function", it explains why the C-tagged
   construct in PMID:15326197 reported a cytosolic C-terminus. Correction B in §10 is
   therefore resolved, not open — the suggested question has been removed and the
   resolution recorded in the PMID:15326197 reference_review.
2. Supplies the oligomerisation mechanism. "We show that Mim1 forms homo-oligomeric
   structures via\nits TMS, which contains two helix-dimerization GXXXG motifs. Mim1 with
   mutated\nGXXXG motifs did not form oligomeric structures and was inactive", and "Thus,
   the TMS of Mim1 is the minimal functional\ndomain of the protein." This is the
   explanation for the odd conservation pattern Waizenegger flagged in 2005 (conservation
   concentrated in the TM segment, unusual for a membrane protein) and it is now cited in
   core_functions.

PMID:21825073 (Becker 2011) — full text available. Strengthens the proposed
GO:0032977 from inference to direct observation: "The critical component for the
subsequent import into the outer membrane is Mim1, and the precursor proteins directly
interact with Mim1." GO:0032977 is defined as binding TM-domain-containing proteins and
mediating their integration; before this, only the integration half was evidenced and the
binding half was inferred from depletion phenotypes.

It also carries a mechanistic correction to §5: "a recent study showed that Mim1 does
not directly promote the biogenesis of Tom40 but functions via the import of small Tom
proteins that are needed for Tom40 assembly (Becker et al., 2010)". So the β-barrel row in
the §4 client table is indirect in a stronger sense than "late step" conveys — Mim1 never
handles Tom40 itself. This does not change the GO:0070096 action (involvement in TOM
assembly is not in doubt) but is now recorded in that annotation's reason.

PMID:21825074 (Papić 2011) — companion paper, independent group, same conclusion for
Ugo1. Its value is that the multispan substrate class rests on two laboratories, not one.

PMID:18187149 (Hulett 2008) — third-laboratory corroboration of the Tom20 dependency,
via the Tom20 TM segment. Uses the Tom13 synonym.

PMID:15242642 (Mnaimneh 2004) — the promoter-shutoff screen that gave MIM1 its name.
Historical provenance only; a genome-scale phenotypic inference, marked relevance: LOW.

Organism caveat. The abstracts of PMID:18177669 and PMID:18187149 do not name the
organism, and the cached records are abstract-only. The yeast attribution rests on UniProt's
ECO:0000269 links to Q08176 and on PMID:28916712 citing Popov-Čeleketić for yeast Mim1
topology — not on anything the abstracts state. This is recorded in both
reference_review.review_notes rather than being silently assumed. By contrast
PMID:21825073 ("in the model organism Saccharomyces cerevisiae") and PMID:21825074 ("using
yeast mitochondria") state it outright.

Not done: the N. crassa orthologue. Q8X0G8 (MIM1_NEUCR, NCU01101) carries three
annotations, none experimental: GO:0140595 and GO:0045040 are IBAs propagating from
PTN002000670 with SGD:S000005386 — yeast MIM1 — in the WITH/FROM, and GO:0005741
is an InterPro IEA. UniProt's FUNCTION line is ECO:0000250 (by similarity). PubMed returns
one hit for Mim1 + Neurospora, and it is a Rapaport-lab study the field cites for yeast
Mim1. A review would therefore have no independent evidence to weigh; every judgement would
resolve to "correct, because yeast MIM1 is correct". The one real gain available is
upgrading GO:0005741 from IEA to experimental on the strength of the N. crassa outer
membrane proteomics in PMID:15608614, which identified NCU01101.1 in that fraction. Judged
not worth a review file on its own.

12. Addendum 3 — full texts of the two 2008 J Mol Biol papers

Publisher PDFs for PMID:18177669 and PMID:18187149 were supplied and extracted into
publications/ (full_text_extraction_method: pdf). Both had been abstract-only. This
resolves one open caveat and adds several findings that the abstracts do not carry.

Organism caveat closed. §11 flagged that neither abstract names its organism, so the
yeast attribution rested on UniProt's ECO:0000269 links. Both full texts state it
outright: PMID:18177669 and
PMID:18187149 (also W303). The indirect attribution is no longer needed.

mim1Δ is viable. Both papers build and use null strains — PMID:18177669 replaced the
ORF with a HIS3 cassette. MIM1 has been carried as "essential" since the deletion-project
database and the promoter-shutoff screen PMID:15242642, and PMID:18187149 addresses this
directly: "Mim1 was first described as an essential gene required for mitochondrial import
by Mnaimneh et al. ... on the grounds that down regulation of MIM1 expression with a TetR
promoter caused a loss of growth on nonfermentable carbon sources". The null is sick, not
dead. Nothing in the review asserted essentiality, but it is worth having on record.

PMID:18177669 — what the abstract omits

• GXXXG motifs are mapped: two consecutive GXXXG/A motifs at residues 57–61 and 63–67.
  The single G61L change is tolerated; the triple mutant (G61L + G63I + A67I, "Mim1-LII")
  "was not able to complement the function of native Mim1 under all conditions tested",
  forms no dimer, and yields no detectable Mim1 complex on BN-PAGE — while phenocopying
  mim1Δ for reduced Tom20 and the low-mass Tom40 species.
• Oligomerisation is necessary, not merely correlated: "Collectively, these results
  demonstrate that the ability of Mim1 to form oligomers is essential for the function of
  the protein in the biogenesis of the TOM complex."
• Stoichiometry: cross-linking plus dual-tag (7His / 3HA) co-isolation show the
  "Mim1-containing complex harbours at least two copies of Mim1."
• Cross-species complementation — directly relevant to the N. crassa question in §11.
  PMID:18177669, with expression verified by Western blot so the
  weak rescue is not a level artefact. By contrast, "cells harbouring S. pombe Mim1 did not
  present any growth phenotype under all tested conditions" — full complementation. The
  authors say plainly they cannot explain the difference. This is the only direct
  functional evidence on the N. crassa protein, and it argues mildly against treating the
  IBA propagation to that orthologue as functional equivalence. It reinforces rather than
  reverses the §11 decision not to curate NEUCR MIM1: the one experiment that exists says
  the orthologue behaves differently, which is a reason not to copy the yeast review across.
• TMS boundary discrepancy: the paper places the TMS at residues 34–79 (deletions were
  1–34 and 76–113); UniProt annotates TRANSMEM 39..62 with ECO:0000255 (prediction).
  Not consequential for any GO term, but the UniProt span is narrower than the functional
  segment this paper defines.

PMID:18187149 — the more valuable of the two, on re-reading

• Mim1 acts catalytically. Using a cysteine-free Mim1(3S) allele that is fully
  functional, the authors show "Mim1 is not one of the prominent cross-linked partners of
  Tom20" — the cross-link fingerprint of docked Tom20 is unchanged. So Mim1 drives the
  assembly reaction and is released: "or else Mim1 catalyzes Tom20 assembly with known
  subunits of the TOM complex". This is the cleanest experimental warrant for an
  insertase molecular function rather than a structural-subunit role, and it is now cited
  on the proposed GO:0032977 entry and in core_functions. It is also the experimental
  substance behind Waizenegger's 2005 speculation that Mim1 "seems to act catalytically".
• Independent, earlier evidence that the Tom40 requirement is indirect. §11 recorded
  this correction from Becker 2010 via PMID:21825073. This 2008 paper got there first by a
  different route: loss of Mim1 reduces the amount of Tom40 imported "but does not diminish
  the overall rate of Tom40 assembly", because "This is consistent with the known role of
  Tom20 as a receptor for Tom40 import:" — less functional Tom20 means less Tom40 arrives,
  though what arrives assembles normally through the 250K and 100K intermediates.
  Two different indirect mechanisms are proposed (via Tom20 as receptor here; via the
  small Tom proteins in Becker 2010), and both are indirect. Note this specific observation
  is in tension with PMID:15608614, which reported intermediate II reduced on Mim1
  depletion. The disagreement is about how much of the Tom40 defect is import versus
  assembly, not about whether Mim1 is required for TOM biogenesis, so no GO action changes.
• Third value for the complex size: ~300 kDa by BN-PAGE, against 400–450 kDa by gel
  filtration PMID:15608614 and ~180 kDa by glycerol gradient. The paper also notes the
  Mim1 complex needs 0.3% digitonin to solubilise where TOM needs 0.1–0.2%, which is a
  plausible methodological source of the spread. The existing suggested question has been
  updated with all three values and this hint.
• Specificity control: "porin is assembled into the outer membrane of mitochondria
  without need of Mim1", confirming TOM and SAM are functional in these mitochondria.

Cache provenance

The two cached records now carry full_text_available: true /
full_text_extraction_method: pdf, written in the same format the ETL produces. Text is the
publisher PDF's, not authored here; PDF extraction leaves minor artefacts (one residual
"homooligomer" where a hyphenated line break was rejoined, and some GXXXG occurrences render
as "G XXXG" in the source layout). Quotes used as supporting_text avoid those spans. All
114 supporting_text values in the review verify as verbatim substrings.

13. Addendum 4 — systematic literature sweep and final pass

Searched PubMed on Mim1 AND mitochondri* (32 hits), Mim2 AND mitochondri*, Tom13 AND mitochondri*, "MIM complex" AND (outer membrane) and YOL026C (0 hits — the systematic
name is not indexed). Eight further papers were cached and cited, taking the reference list
from 14 to 22.

Name collisions to be aware of. Several PubMed hits for "MIM1" are not this gene:
MIM1 the Mcl-1 BH3-mimetic small molecule (PMID:40892149, PMID:31410885, PMID:30796196,
PMID:30134184) and the chicken c-Myb target gene mim-1 (PMID:37752090). Anyone re-running
a naive MIM1 search will hit these; none is relevant.

The significant find — a second function (PMID:41748941, Nat Cell Biol 2026)

"Here we report that the MIM complex performs a second major function in\nlipid-droplet
homeostasis." The lipid metabolism enzyme Ayr1 captures MIM and nucleates
mitochondria–lipid droplet contact sites; MIM and Ayr1 together set cellular lipid
droplet number. The mechanism is the elegant part: "Ayr1 binds to MIM via its single
hydrophobic segment in a\nsubstrate-mimicry mechanism but remains bound and is not released
into the outer\nmembrane." So this is a captured insertase, not a second catalytic
activity — and the two jobs are done by distinct populations, "MIM-Ayr1 for recruiting lipid
droplets and MIM-preprotein for protein insertion".

A pleasing loop closes here: Ayr1 was the other new channel identified alongside Mim1 in
the same 2017 bilayer screen PMID:28916712. The two proteins were already known to co-occur
in the outer-membrane fraction.

Curation consequences:
- New annotation proposed: GO:0034389 lipid droplet organization (IMP, involved_in),
flagged provisional — one 2026 study, no replication, abstract-only in cache.
- Ontology gap. GO has GO:0160259 (ER–lipid droplet contact site) and GO:0170007
(ER–lipid droplet tether activity), and GO:0044233 for ER–mitochondrion, but nothing
for mitochondrion–lipid droplet. Two terms proposed, and the gap is recorded as a
knowledge_gaps entry (ONTOLOGY / CC_DARK) on a second core_functions block.

A genuine disagreement in the literature (PMID:19345216)

Lueder & Lithgow 2009, "The three domains of Mim1 have discrete functions", contradicts
PMID:18177669 on the soluble domains: it assigns the N-terminal domain a role regulating the
SAM-mediated early reaction of Tom40 assembly, where Popov-Čeleketić found ΔN and ΔC
variants phenotypically normal and concluded the TMS is the minimal functional domain. The
assays differ (in vitro multi-step assembly kinetics vs in vivo complementation), so a
quantitative contribution invisible to a growth assay would reconcile them. Marked
correctness: DISPUTED for the domain-function claim only — its topology report is not
disputed and in fact makes the cytosolic-N/IMS-C assignment two-laboratory. No GO action
turns on the disputed point. Recorded as a suggested question and experiment.

Tom40 indirectness — primary source, and a second route

• PMID:20026336 (Thornton/Becker 2010) is the actual "Becker et al., 2010" that
  PMID:21825073 cites and that the review had been relying on second-hand. Mim1 inserts the
  small α-helical subunit Tom6, which SAM-Tom5/Tom40 then acts on. It also explains the
  Tom22 exception from PMID:17974559: Tom22 is handled by the parallel SAM-Mdm10 module.
• So there are now two proposed indirect routes to the Tom40 phenotype — via Tom20 as
  import receptor PMID:18187149 and via the small Toms PMID:20026336 — plus the unresolved
  disagreement about whether assembly rate is actually reduced. The suggested question has
  been rewritten to ask which contributes and in what proportion, rather than the now-answered
  "is it direct or indirect".

Substrate scope is narrower than "the" insertase implies

• PMID:31945731 — others are MIM-independent. Also independently establishes
  tail-anchored proteins as MIM substrates.
• PMID:35262629 — even a multi-span client
  is not absolutely MIM-dependent.
• Against this, PMID:41748941 credits MIM with >90% of integral outer-membrane precursors.
  Both are now in the description: main translocase, but substrate-specific rather than
  obligatory.

Orthology put on an experimental footing (PMID:29923829)

The convergent-evolution story previously rested on the GO:7770059 term definition. Vitali
et al. 2018 demonstrate it: Mim1/Mim2 is "fungi-specific", and trypanosome pATOM36 is a
functional analogue "even though these proteins show neither sequence nor topological
similarity", shown by reciprocal complementation. This explains why PTHR28241 is fungal
and why phylogenetic propagation of MIM function should stop at the fungal clade. The
description now says restricted to fungi rather than highly conserved among fungi.
PMID:40782022 (the GO:7770059 xref) adds MTCH1/MTCH2 as the mammalian solution.

Also added

• PMID:23959800 — Djp1 and Tom70 are needed for the biogenesis of Mim1 itself. Recorded
  mainly as a confusability warning: the Mim1–Tom70 relationship appears in this review in
  the opposite direction (Tom70 handing precursors to MIM), and the two are easy to conflate.

Final-pass checks

• 146 supporting_text values verified verbatim against the cache (one wrap error caught and
  fixed during the sweep).
• No reference declared-but-uncited or cited-but-undeclared; every reference has both
  findings and reference_review; all full_text_unavailable flags agree with the cache.
• Two suggested questions that the new evidence had answered were rewritten rather than left
  standing (the Tom40 direct/indirect question, and the non-homologous-counterpart question).
• Final tallies: 16 annotations (ACCEPT ×11, MARK_AS_OVER_ANNOTATED ×2, REMOVE ×1, NEW ×2),
  22 references (HIGH 13 / MEDIUM 5 / LOW 4; VERIFIED 21 / DISPUTED 1), 2 core functions,
  3 proposed new terms, 2 knowledge gaps, 10 questions, 10 experiments.
• Schema, term-branch, GOA-consistency and compliance checks all pass; status: COMPLETE.

14. Addendum 5 — disputes and knowledge gaps as first-class objects

The repo already models both, following the monarch-initiative/dismech pattern documented in
projects/FUNCTION_KNOWLEDGE_GAPS.md. Earlier addenda recorded these as prose in
review_notes; they are now promoted to structured objects.

Disputes: FindingReview (finding-level, not reference-level)

FindingReview carries finding_status (CURRENT / CORROBORATED / DISPUTED /
OVERTURNED / UNVERIFIED), superseded_by, and review_notes. It is explicitly
finer-grained than reference_review, whose docstring notes that "a paper may contain some
findings that stand and others that are overturned" — exactly the MIM1 situation. Five
findings are now adjudicated:

Two points of method worth recording:

• The 19345216 pair is the reason finding-level matters. One paper, two findings, opposite
  verdicts: its topology result is corroborated and is now one of the two pillars of the
  accepted topology, while its domain-function claim is contested by PMID:18177669. A
  reference-level DISPUTED alone would have tainted the good finding; a reference-level
  VERIFIED alone would have hidden the contested one. Both levels are now set, so a summary
  view still flags the paper while the detail says which claim.
• OVERTURNED vs DISPUTED was applied deliberately. Only the 15326197 topology claim is
  OVERTURNED — three later studies contradict it and PMID:15608614 supplies the mechanism
  of the error (the C-terminal tag "had compromised function"). The Tom40 claims are
  DISPUTED, not overturned: the two superseding papers propose different indirect routes,
  and PMID:18187149's assembly-rate result is itself in tension with PMID:15608614. Where the
  supersession is not clean, the schema's own guidance is to curate with caution rather than
  declare a winner.

Knowledge gaps: KnowledgeGap

Five gaps now recorded — three at gene level (GeneReview.knowledge_gaps) and two on the
lipid-droplet core function:

1. Mechanism of helix integration (BIOLOGY / MF_DARK / NARROWING). The path a substrate
   helix takes is undetermined. Provenance is the field's own admissions: PMID:15608614 and
   PMID:28916712. This is the gap that makes the
   central curation call hard — whether GO:0022832 is over-annotation or an
   under-specified description of the real conduit turns on it.
2. What determines MIM dependence (BIOLOGY / BP_DARK / OPEN). No sequence or biophysical
   rule predicts which α-helically anchored clients need MIM; dependence is demonstrably
   partial (PMID:31945731, PMID:35262629).
3. Fungal-clade uniformity (BIOLOGY+CURATION / RESIDUAL_SUBGAP / OPEN). Why N. crassa
   Mim1 only partially complements while S. pombe Mim1 complements fully is unexplained —
   PMID:18177669. Bears directly on the safety of the IBA propagation from
   PANTHER:PTN002000670 to Q8X0G8, and closes the loop on the §11/§12 decision not to curate
   the Neurospora orthologue.
4. No ontology term for the mitochondrion–lipid droplet contact site (ONTOLOGY / CC_DARK),
   paired with the two proposed_new_terms.
5. Which partner supplies the lipid droplet-facing determinant (BIOLOGY / MF_DARK).

All gap provenance quotes are verbatim-checked by the reference validator, the same discipline
as positive claims.

OpenScientist jobs launched

Two hypotheses from this review are computationally testable, and both were scoped to a
single decisive question per the openscientist-hypothesis skill (broad "characterize this
gene" prompts hit the 7200s ceiling). Run with max_iterations=3, timeout=7200,
--timeout-seconds 8100:

• mim1-tms-gxxxg-oligomeric-pore — does the GXXXG-mediated TMS oligomer enclose a
  continuous aqueous lumen, or only a helix bundle with a lipid-exposed groove? This is the
  computational proxy for gap 1 and bears directly on the GO:0022832 decision and on the
  proposed "protein-conducting channel activity" term. A negative (no enclosed lumen) would
  strengthen the MARK_AS_OVER_ANNOTATED call and weaken the proposed channel term.
• mim1-patom36-mtch-fold-convergence — do Mim1, pATOM36 and MTCH1/MTCH2 share a fold, or
  is the convergence purely functional? Foldseek-scoped. A negative is informative and was
  requested explicitly in the prompt.

Per the skill, no local *-bioinformatics/RESULTS.md was fed to either job (there is none for
this gene yet); verdicts must be compared against held-out evidence after the run, and a
refuted proposal is deleted rather than set to REMOVE.

15. Addendum 6 — GO issues #31711 and #32108

Two GO ontology tickets change how this gene should be annotated.

Access caveat. geneontology/go-ontology is outside this session's GitHub scope. Anonymous
git read is available but the GitHub API is not (403 for unattached repos), there is no gh
CLI, and the sanctioned add_repo --access push route was blocked. WebFetch retrieves issue
bodies but GitHub now renders comment timelines client-side, so
#32108's comment 4449178932 was not read. Everything below comes from the two issue
bodies plus the live ontology, which was checked directly.

#31711 — mitochondrial import reorganization (CLOSED, largely landed)

ValWood's GO-CAM-driven restructuring. The new terms are live in OLS/the GO API:

GO:0045040 now has exactly two children: GO:7770059 (alpha-helical) and GO:7770063
(beta-barrel). The proposed obsoletions have not landed — GO:0070585, GO:0051204,
GO:0072656 and the GO:19036xx/19037xx regulation terms are all still live.

#32108 — obsolete GO:0070096 (OPEN)

Opened May 2026, 21 EXP annotations affected. Rationale, quoted from #31711:

"Assembly of TOM isn't a separate assembly process, assembly is only affected by import
pathways so the correct way to annotate this would be to add TOM subunits as targets of the
appropriate import pathways."

This is the same argument this review reached independently from the primary literature —
that the TOM-assembly phenotype is a downstream consequence of the insertase acting on Tom20
and Tom70, with the Tom40 arm indirect besides. PMID:17974559 is the neatest illustration: its
substrate-resolved result (Tom20 and Tom70 yes, Tom22 no) names exactly the "TOM subunits as
targets" that #31711 wants annotated instead.

Changes made

1. Three GO:0045040 rows → MODIFY, proposed_replacement_terms: GO:7770059. MIM1 is
   unambiguously alpha-helical; picking the child encodes the Tom40-is-indirect conclusion
   rather than obscuring it — MIM1 belongs on GO:7770059 and explicitly not on GO:7770063.
   The IBA row also carries a propagation_review (TERM_SCOPING_PROBLEM /
   GRANULARITY_MISMATCH) recording that the node placement is sound and the re-term should
   move with the node, since the whole fungal clade is alpha-helical-specific.
2. Two GO:0070096 rows → KEEP_AS_NON_CORE. Deliberately not REMOVE or MODIFY: the term
   is still live and non-obsolete, the loss-of-function evidence is sound, and the obsoletion
   request is open, not decided. Pre-empting a GO decision that has not been taken would be
   worse than recording the position.
3. New annotation proposed for GO:7770059 (IDA, PMID:32348752), recording a plain GOA
   gap: Q08176 has no GO:7770059 annotation although the term's definition is written around
   the Mim1–Mim2 complex. Of ~286 annotations to the term only seven are manual — PomBase IMPs
   on S. pombe mim1 (Q9C1W7) and mim2 (G2TRP0) from PMID:33138913, and FlyBase IMPs on
   Drosophila Mtch and human MTCH1/MTCH2 — the rest Ensembl orthology IEA. The budding-yeast
   protein the definition describes is the conspicuous absence.
4. core_functions.directly_involved_in: GO:0070096 dropped (demoted to non-core).

The one unresolved item, and why

core_functions.directly_involved_in still holds GO:0045040, not GO:7770059. This is a
tooling limit, not a judgement:

• core_functions term ids are hard-validated (branch reachability + label resolution).
• GO:7770059 is live in OLS and the GO API but absent from the semantic-sql GO build that
  linkml-term-validator resolves labels against. Verified by deleting the local 1.7 GB
  go.db and re-downloading a fresh one — the new build lacks it too, so the lag is
  upstream, not a stale local cache.
• A hand-patch adding the five terms to cache/enums/gobiologicalprocessenum_*.csv cleared
  the reachability check but not the label check, and was reverted: forcing a term the pinned
  ontology genuinely lacks into a committed derived artifact is the wrong fix.

Consequence: one standing validation warning ("Process term GO:0045040 not reflected in
existing_annotations"), which is accurate — core_functions and existing_annotations do
disagree, precisely because of the upstream lag. It is left visible rather than suppressed,
and status is therefore DRAFT, not COMPLETE. Swap GO:7770059 into that slot and restore
COMPLETE once semantic-sql catches up.

Flagged for GO, unrelated to this gene

ComplexPortal:CPX-336 ("clb3-cdc28_yeast-1", a cyclin–CDK complex) carries an IDA to
GO:7770059 from PMID:25378463. That looks like a data error rather than curation.