Succinate Dehydrogenase (Complex II) — gp2term Relation Review (enables → contributes_to)

Bottom line: succinate dehydrogenase (respiratory complex II) is a
four-subunit enzyme, and no single subunit carries out the whole
succinate-to-quinone reaction (GO:0008177). GO has agreed
(go-annotation#6414) that each subunit should link to that activity with
contributes_to rather than enables. The first repo audit found five
reviews carrying GO:0008177: human SDHA, SDHB, SDHC, SDHD and one 9POAL
plant iron-sulfur-subunit ortholog, NCGR_LOCUS67308. Those five all argue for
contributes_to in prose. At that 2026-09-26 audit, only SDHC and one SDHA row
carried the structured qualifier: contributes_to field, and in both cases
that qualifier comes from GOA itself, not from a curation edit.
PR #3223 (merged
2026-09-27) fixed SDHB and SDHD by marking the GOA enables GO:0008177 rows
MODIFY and pairing them with one NEW GO:0008177
qualifier: contributes_to row per gene, the relation-only pattern the
validator accepts. Still open in
#582: the SDHA
GO:0000104 half-reaction decision, the plant SDH2 ortholog
(NCGR_LOCUS67308), and PSEPK sdhA/sdhB, which were added later and still
carry GOA enables rows on GO:0008177.

The fix matters because a qualifier that lives only in prose is invisible to
any tool that reads the YAML, and SDHA needs a real judgement: its
flavoprotein subunit can run the succinate half-reaction alone, so enables
on the parent GO:0000104 may be correct for it.

Overview

As part of the GO-CAM protein-complex annotation policy, the GO
consortium has agreed that the individual subunits of respiratory
chain complex II (succinate dehydrogenase) should all be annotated
to the molecular-function term GO:0008177 succinate dehydrogenase
(quinone) activity (and to its parent GO:0000104 succinate
dehydrogenase activity) using the contributes_to qualifier
rather than enables. No individual subunit carries out the full
succinate → quinone reaction on its own, so enables (which asserts
the gene product has the activity independently) is the wrong
gene-product-to-term relation; contributes_to (gene product
contributes to a complex-level activity) is correct.

The biology is unchanged — this is a gp2term relation
(qualifier) correction, not a term change.

Upstream tickets

• Annotation tracker: geneontology/go-annotation#6414 — Review gp2term relations for succinate dehydrogenase activity terms (opened 2026-05-11; still open as of 2026-10-05).
• Repo tracker: ai4curation/ai-gene-review#582 — tracks the repo-side Complex II qualifier cleanup.
• Curator-maintained review spreadsheet: https://docs.google.com/spreadsheets/d/1h0_fUrGRlDKUyy9o1S3gMuN8XiVJW_3vHwIe5aaP250
• Upstream progress (issue comments): CGD done (@jlewsmith), TAIR done (@lreiser), MGI done (@LiNiMGI).

The two MF terms (verified via OLS, 2026-05-19)

Both terms are active (not obsolete) in the GO API / OLS as of
2026-05-19. This is purely a qualifier/relation fix; no obsoletion is
involved.

Complex II subunits

Respiratory chain complex II has four subunits:

Nuance for SDHA: the upstream blanket "contributes_to for
GO:0008177 and GO:0000104" applies cleanly to SDHB/SDHC/SDHD. For
SDHA, the complex-level GO:0008177 (quinone) is unambiguously
contributes_to, but the parent GO:0000104 (succinate +
generic acceptor → fumarate) describes the half-reaction that the Fp
subunit can catalyse via its FAD with an artificial acceptor — so
enables on GO:0000104 for SDHA specifically is defensible. Reviews
touching SDHA should flag this rather than blindly applying
contributes_to to GO:0000104. (This is exactly the kind of
relation judgement this repo is designed to surface.)

Impact on this repo

Seven review files carry an existing annotation to GO:0008177.
Checked genes/**/*-ai-review.yaml for GO:0008177 / GO:0000104:

genes/PSEPK/sdhC and genes/PSEPK/sdhD mention GO:0008177 only under
core_functions.contributes_to_molecular_function, not as existing GOA rows;
they are already modelled correctly for this issue. genes/DESVH/Q72DT2
matches GO:0000104 because TreeGrafter propagated succinate dehydrogenase
activity to the wrong SdhA/FrdA-family protein, but that row is already
REMOVE; it is a false positive, not an unresolved SDH qualifier case.

The substantive finding: the qualifier intent was discussed in prose
in the original five human and 9POAL reviews but the structured
qualifier: contributes_to field was applied inconsistently (present
on SDHC and one SDHA entry; missing on SDHB and SDHD until #3223).
The follow-up is a small, well-defined consistency pass, not new
biology.

Scope

• Organisms in this repo: human (SDHA/SDHB/SDHC/SDHD) plus one
  plant SDH2 ortholog (9POAL) and the Pseudomonas putida KT2440
  SdhABCD reviews that now exist under genes/PSEPK/. Other MOD
  orthologs (CGD/TAIR/MGI/yeast SDH1–4/E. coli sdhCDAB) are handled
  by their respective groups upstream and are not reviewed in
  this repo.
• GO branch: MF only — GO:0008177 and its parent GO:0000104.
• Type of fix: gp2term relation/qualifier (enables →
  contributes_to). No term changes, no obsoletion.
• Classification: Type B (systematic, gene-family-wide
  relation correction).

Candidate genes for follow-up review

All scoped genes are already in the repo — this is a refresh/consistency
pass, not new reviews. Re-pull only if upstream qualifier changes
have propagated to GOA: just fetch-gene human SDHB etc.

Tier 1 — concrete consistency fixes (clear, low-judgement)

1. SDHB (human, P21912) — ✅ done in PR #3223 (merged 2026-09-27): the three GOA
   enables GO:0008177 rows (one IEA, two IMP) are MODIFY with
   GO:0008177 as the proposed replacement, and a NEW GO:0008177 row
   carries qualifier: contributes_to (IDA, PMID:37098072).
2. SDHD (human, O14521) — ✅ done in #3223, same pattern for its
   one IEA row.

Note: qualifier on an existing annotation mirrors the GOA row
(SDHC-goa.tsv itself ships contributes_to; SDHB-goa.tsv and
SDHD-goa.tsv ship enables), so on the SDHB/SDHD GOA rows the
qualifier is written to match GOA (enables), not overridden;
an override would be overwritten by just fetch-gene. The intended
relation is recorded as a MODIFY on each enables row plus a paired
NEW contributes_to row, which is the combination the validator
accepts. Once GOA reflects #6414, re-fetch and the paired rows collapse
into the GOA rows.

Tier 2 — verify / nuance

1. SDHA (human, P31040) — re-examine the GO:0000104-vs-GO:0008177
   split per the SDHA nuance above. Confirm GO:0008177 is
   contributes_to; decide explicitly whether GO:0000104 (FAD
   half-reaction) is enables for the Fp subunit and document the
   reasoning.
2. NCGR_LOCUS67308 (9POAL SDH2) — decide between two curation
   strategies: keep the current MODIFY → GO:0009055 electron transfer
   activity replacement, or revise the existing-annotation review to
   retain GO:0008177 with structured qualifier: contributes_to. Do
   not simply add a qualifier if the MODIFY-to-GO:0009055 strategy
   remains the intended review outcome.
3. PSEPK sdhA and sdhB — update the GOA enables
   GO:0008177 rows to the #6414 relation-only pattern. sdhC and
   sdhD do not have GOA GO:0008177 rows and already capture their
   Complex II contribution in core_functions.

Reference exemplar (no action)

1. SDHC (human, Q99643) — already correctly uses qualifier: contributes_to with action: ACCEPT. Use as the template for
   the others.

Proposed approach

1. Hold as a tracking project. Upstream #6414 is being worked
   group-by-group (CGD/TAIR/MGI done as of 2026-05-18); the GOA
   qualifier changes for human SDH subunits may not yet have
   propagated. Doing the consistency pass now risks re-pulling GOA
   that still shows enables.
2. Tier 1 on SDHB and SDHD is done (#3223, MODIFY + paired
   NEW contributes_to rows). When GOA reflects the upstream change,
   re-fetch both and re-run just validate human SDHB /
   just validate human SDHD.
3. Do the Tier 2 SDHA/9POAL/PSEPK checks, explicitly documenting
   the SDHA GO:0000104 half-reaction nuance and applying the paired
   MODIFY + NEW pattern to PSEPK sdhA and sdhB.
4. Do not create reviews for non-repo MOD orthologs — those are
   owned by the upstream groups.

Priority

Low–Medium. Small, well-bounded scope (6 genes needing a
structured-field consistency edit or explicit relation decision, of
which SDHB and SDHD are done in #3223; SDHC is already correct). No new
biology and no obsoletion — the substantive value is making the
repo's gp2term qualifiers internally consistent and aligned with the
upstream Complex II policy decision. Good low-risk follow-up once the
upstream qualifier change propagates.

Status

• 2026-05-19 — Project file created. Upstream go-annotation#6414
  open (last updated 2026-05-18; CGD/TAIR/MGI done upstream).
  GO:0008177 and GO:0000104 both verified active (not obsolete) via
  OLS. Five repo genes carry GO:0008177; structured contributes_to
  qualifier present on SDHC (and one SDHA entry) but missing on
  SDHB/SDHD — held as a tracking project with a defined Tier 1/Tier 2
  consistency pass. No gene reviews edited yet.
• 2026-09-27 — SDHB and SDHD edited in
  PR #3223
  (merged 2026-09-27): GOA enables GO:0008177 rows → MODIFY, each gene gains a paired
  NEW GO:0008177 contributes_to row (IDA, PMID:37098072). Open:
  SDHA GO:0000104, 9POAL NCGR_LOCUS67308, PSEPK sdhA/sdhB.
• 2026-10-05 — Re-checked all current GO:0008177 / GO:0000104 review
  hits. Confirmed PSEPK sdhA and sdhB are unresolved GOA-row cases,
  while PSEPK sdhC/sdhD only mention GO:0008177 in core functions
  and DESVH aprA is a false-positive GO:0000104 propagation that the
  review already removes.