Mitochondrion Targeting Sequence Binding — Obsoletion & Replacement

Bottom line: Most mitochondrial proteins are imported by receptors of the
TOM and TIM complexes that recognise an N-terminal targeting presequence. GO
has obsoleted the generic binding term GO:0030943 mitochondrion targeting
sequence binding "in favor of more specific molecular functions", and the
receptor term the project was waiting for now exists as GO:0140436
mitochondrial signal sequence receptor activity (OLS, checked 2026-09-26).
We listed the 18 curated annotations on the old term and sorted them into
classes, because only the TOM receptors (and probably TIM50) are true
presequence receptors: the TIM23 and TIM22 channels, a plant phosphatase and
the TIM23 complex records need individual decisions, so upstream ruled out a
blanket replaced_by. In this repo, 10 reviews touch GO:0030943, five of them in
core_functions. The MODIFY pass was done in #3234: receptors
move to GO:0140436 and the TOM40 and TIM22 channels to GO:0008320 protein
transmembrane transporter activity; the per-row outcomes are in Impact on
this repo and Status below. The upstream triage of plant PAP2 and the TIM23 complex records is
not repo work.

Overview

A GO obsoletion proposal will obsolete the molecular-function term
GO:0030943 mitochondrion targeting sequence binding (defined as "Binding to a
mitochondrion targeting sequence, a specific peptide sequence that acts as a
signal to localize the protein within the mitochondrion"). The upstream
rationale is that the curated content is better captured by a non-binding
receptor activity term rather than a generic "binding" term: the ontology
ticket proposes a New Term Request (NTR) for
mitochondrial signal sequence receptor activity as the replacement,
mirroring the existing nuclear/vacuolar pattern
(GO:0061608 nuclear import signal receptor activity,
GO:0005049 nuclear export signal receptor activity,
GO:0010209 vacuolar sorting signal receptor activity).

Crucially, the go-annotation curators note that a simple replaced_by
cannot be applied, because the existing annotations are not only to the
receptor — they span TOM cytosolic receptors, inner-membrane TIM
channel/receptor components, the TIM23 holo-complex, and at least one
non-canonical plant protein. Each annotation therefore needs individual review
to decide whether the new receptor MF is appropriate or whether a different
term (or removal) is the right outcome.

This is part of the broader mitochondrial-import GO-CAM reorganization
(go-ontology#31711) and is a sibling of the "signal sequence binding and
children" review (go-ontology#31419, which also drives the
GO:0008139 nuclear localization sequence binding obsoletion in
go-annotation#6435). It complements — but does not overlap with — the
BP-focused [[MITOCHONDRIAL_IMPORT_PATHWAYS]] project, which covers the import
pathway terms rather than this MF term.

Upstream tickets

• Annotation tracker: geneontology/go-annotation#6437
• Ontology ticket (NTR + obsoletion): geneontology/go-ontology#32142
• Parent reorganization: geneontology/go-ontology#31711 (CLOSED — "Reorganization of mitochondrial import pathways based on GO-CAM modelling")
• Sibling "signal sequence binding" review: geneontology/go-ontology#31419

Obsoletion plan (per upstream)

Term labels verified in OLS on 2026-05-28:

• GO:0030943 (mitochondrion targeting sequence binding) — live, slated for
  obsoletion. Parent is GO:0005048 signal sequence binding. Synonym:
  "mitochondrial targeting sequence binding".
• GO:0005048 (signal sequence binding) — live parent MF.
• GO:0061608 (nuclear import signal receptor activity) — live; the model
  the NTR is patterned on.
• GO:0010209 (vacuolar sorting signal receptor activity) — live; analogous
  receptor MF.
• mitochondrial signal sequence receptor activity — the proposed NTR was
  not yet present in OLS as of 2026-05-28 (no GO ID minted). This is the
  key open dependency: nothing can be remapped until the new MF is created.
  Update 2026-09-26: resolved. The NTR was minted as GO:0140436
  (mitochondrial signal sequence receptor activity, live in OLS) and
  GO:0030943 is now obsolete; see Status.

Affected experimental / curated annotations (18)

Retrieved from the QuickGO annotation API on 2026-05-28
(goId=GO:0030943, manual / experimental + ComplexPortal NAS). Matches the
upstream group tally (ComplexPortal 4, FlyBase 2, HGNC-UCL 1, RGD 2, SGD 7,
TAIR 2 = 18).

Note: rows 11–12 carry assignedBy=FlyBase on human accessions — confirmed
directly from QuickGO, presumably a cross-organism assertion; flagged here for
the curator's awareness.

On top of these 18 curated records, GO:0030943 currently has ~12,091 total
annotations (QuickGO, 2026-05-28), overwhelmingly IEA (TreeGrafter,
GO_REF:0000118) and IBA (GO_REF:0000033). These will be retired/redirected
automatically once the term is obsoleted, but the volume illustrates how far a
small set of curated TOM-receptor annotations has propagated.

Why a blanket replaced_by does not work

The annotations fall into biologically distinct classes, only some of which are
true presequence receptors:

• Cytosolic TOM receptors — TOMM20/TOM20 (human/rat/yeast), TOMM22,
  rat Tomm40l, yeast TOM70. These directly recognize the amphipathic
  N-terminal presequence on the cytosolic face. The NTR
  mitochondrial signal sequence receptor activity is a clean fit here.
• Trans-side (IMS) receptor — yeast TIM50 hands the presequence from TOM
  to the TIM23 channel; receptor-like, the NTR likely fits.
• Channel components — yeast TIM23 (presequence translocation channel).
  Whether a "receptor activity" MF is the right home, versus modelling TIM23 as
  a transporter that is an input to the matrix-import BP, needs curator input.
• Carrier-pathway channel — yeast TIM22. Carrier substrates (e.g.,
  metabolite carriers) use internal targeting signals, not cleavable
  N-terminal presequences. Annotating TIM22 to "mitochondrion targeting
  sequence binding" is already noted as a loose fit in this repo's
  genes/yeast/TIM22 review; the new presequence-receptor MF may not be
  the correct replacement for TIM22.
• Non-canonical / plant — Arabidopsis PAP2 (purple acid phosphatase 2,
  Q9LMG7), an IPI annotation (PMID:26304849). Needs individual review to decide
  whether a receptor MF applies at all.
• Complex-level — ComplexPortal TIM23 holo-complex entries (CPX-539,
  CPX-6127, CPX-6129, CPX-6130). Map to the complex's receptor/transporter
  activity once the NTR exists.

Impact on this repo

Several affected gene products — or their human orthologs — already have
*-ai-review.yaml files that annotate GO:0030943 (re-verified 2026-09-27):

The per-row remapping above was made in #3234; its Status entry below
summarises it.

The receptor/channel split follows the classes above: GO:0140436 only for
the TOM presequence receptors and TIM50, GO:0008320 for the TOM40 and TIM22
channels.

Scope

• GO branch: Molecular Function (single term obsoletion + one NTR).
• Organisms: Human, rat, Saccharomyces cerevisiae, Arabidopsis
  thaliana (curated set); plus the large IEA/IBA tail across all eukaryotes.
• Gene set: TOM complex receptors (TOM20/TOMM20, TOMM22, TOM70/TOMM70,
  Tomm40l), TIM23-complex components (TIM23, TIM50), the TIM22 carrier channel,
  the TIM23 holo-complexes (ComplexPortal), and Arabidopsis PAP2.
• Type of fix: Curation hygiene / refactor. The biology is well established
  (TOM20/TOM22 are the canonical presequence receptors); the work is about
  moving from a generic "binding" MF to a precise "receptor activity" MF and
  triaging the annotations that are not really receptor functions.

Candidate genes for initial review

Listed in priority order. The first three are the clean, high-value receptor
cases that already have reviews in this repo and would directly exercise the
new MF once minted.

1. TOMM20 (human, Q15388) — canonical N-terminal presequence receptor;
   directly affected by two of the 18 annotations; review already
   ACCEPTs GO:0030943 as the core MF. Best positive control for the new
   mitochondrial signal sequence receptor activity.
2. TOMM22 (human, Q9NS69) — TOM central/co-receptor; directly affected.
3. TOMM70 (human) — receptor for carrier/hydrophobic precursors; ortholog
   of the affected yeast TOM70 (P07213, IMP).
4. TIM50 / TIMM50 — trans-side presequence handoff receptor; affected yeast
   TIM50 (Q02776) plus the human ortholog review here.
5. TIM22 (yeast, Q12328) — special case: carrier pathway uses internal
   signals; decide whether the new receptor MF applies or whether the
   annotation should be removed/replaced with a carrier-import term.
6. PAP2 (Arabidopsis, Q9LMG7) — non-canonical IPI annotation; case-by-case.
7. TIM23 complex (ComplexPortal CPX-539/6127/6129/6130) — complex-level
   handling once the NTR exists.

Proposed approach

1. Wait for the NTR + obsoletion to land. (Done 2026-09-26: minted as
   GO:0140436.) The replacement MF
   (mitochondrial signal sequence receptor activity) was not yet minted in GO
   as of 2026-05-28.
2. Pre-stage MODIFY proposals on the existing repo reviews (TOMM20, TOMM22,
   TOMM70, TIMM50, TIMM22, TOMM40), changing GO:0030943 → the new receptor MF
   for the genuine cytosolic/trans-side receptors.
3. Triage the non-receptor cases explicitly: TIM22 (internal-signal carrier
   channel), PAP2 (plant IPI), and the TIM23 holo-complex records. These are
   the reason upstream avoided a blanket replaced_by.
4. Note the IBA/IEA fallout (~12k annotations): once GO:0030943 is
   obsoleted, the GO_Central IBA and TreeGrafter IEA pipelines will need to be
   reseeded against the new MF. The yeast ACL4 review (now UNDECIDED, with
   no replacement proposed) is a concrete example of an IBA that should not be
   carried over to the new receptor term without first resolving the PAINT
   node.
5. Coordinate with [[MITOCHONDRIAL_IMPORT_PATHWAYS]] so MF remapping and the
   BP pathway model stay consistent for shared TOM/TIM genes.

Priority

Medium. Only 18 curated annotations, and several of the key genes already have
reviews here, so the marginal curation effort is low. The receptor biology is
uncontroversial, so the main blocker is external (the NTR GO ID is not yet
minted). The large IEA/IBA tail makes this more impactful than the very small
obsoletions, but no curator group is blocked waiting on AI Gene Review.

Status

• 2026-05-28 — Project file created. Tracking go-annotation#6437 (opened
  2026-05-27) and go-ontology#32142 (closed; NTR + obsoletion request). The 18
  affected curated annotations were retrieved from QuickGO and reconciled
  against the upstream group tally. GO:0030943 confirmed live in OLS (parent
  GO:0005048); the proposed replacement MF
  mitochondrial signal sequence receptor activity is not yet in OLS — the
  key open dependency. Eight existing repo reviews already touch GO:0030943 and
  will need a MODIFY pass once the new term exists. No InterPro2GO / UniRule /
  UniProt-Keyword mappings to GO:0030943 were listed by upstream.
• 2026-09-26 — OLS lists GO:0030943 as obsolete, and the replacement
  GO:0140436 mitochondrial signal sequence receptor activity is live. Ten
  repo reviews touch GO:0030943: human TOMM20, TOMM22, TOMM40, TOMM70, TIMM50,
  TIMM22; yeast TIM22, TOM22, ACL4; worm tomm-22. Five list it in
  core_functions (TOMM20, TOMM22, TIMM50, tomm-22, TOM22). The local
  cache/ontologies/go.tsv still records GO:0030943 as live, so validation
  does not flag these yet. None of the reviews uses GO:0140436. TOMM20 also
  uses GO:0030943 as the proposed_replacement_terms target of its obsolete
  GO:0051082 (unfolded protein binding) row, so that replacement must also
  move to GO:0140436.
• 2026-09-26 — GO:0030943 is obsolete and the NTR is minted as GO:0140436
mitochondrial signal sequence receptor activity. PR #3234 remapped the
  repo reviews. Receptors got MODIFY → GO:0140436: human TOMM20 (IDA, IBA),
  TOMM22 (IDA) and TOMM70 (IBA, ISS). Receptors with no GO:0030943 row got a
  NEW GO:0140436 row backing their core MF: human TIMM50, yeast TOM22 and worm
  tomm-22. Channels got MODIFY → GO:0008320 protein transmembrane
  transporter activity, folded into the GO:0008320 annotation each gene already
  carries: human TOMM40 (IBA), yeast TIM22 (IDA, IBA) and human TIMM22 (IBA,
  same PTN000364156 node as yeast TIM22). Yeast ACL4 stays UNDECIDED with no
  replacement. This corrects the earlier "already REMOVE" description of ACL4.
  The PAP2 and ComplexPortal TIM23 cases remain open.

Slides

• Slides (Marp source: MITOCHONDRION_TARGETING_SEQUENCE_BINDING_OBSOLETION-slides.md) — AI generated