ETFDH (human, UniProtKB:Q16134) — review notes

Summary of gene function

ETFDH encodes electron transfer flavoprotein-ubiquinone oxidoreductase (ETF-QO / ETF
dehydrogenase; EC 1.5.5.1)
, a monotopic iron-sulfur flavoprotein of the mitochondrial
inner membrane. It accepts electrons from reduced electron-transfer flavoprotein (ETF,
the matrix-soluble ETFA/ETFB heterodimer) and transfers them to ubiquinone in the
respiratory chain. This is the terminal step of an electron-transfer relay that couples
~10-11 mitochondrial FAD-dependent flavoprotein dehydrogenases (fatty-acid beta-oxidation
acyl-CoA dehydrogenases; amino-acid and choline catabolism dehydrogenases) to oxidative
phosphorylation.

Cofactors: one FAD and one [4Fe-4S] cluster; both redox centers are in a 64-kDa
mature monomer.

Key provenance

Disease

Biallelic loss-of-function → glutaric acidemia type II / multiple acyl-CoA dehydrogenase
deficiency (MADD)
, a disorder of fatty acid, amino acid and choline metabolism; the
late-onset myopathic form is frequently riboflavin-responsive and is associated with a
secondary (myopathic) CoQ10 deficiency (PMID:17412732). Corroborated by the disorder KB
(dismech/kb/disorders/Multiple_Acyl-CoA_Dehydrogenase_Deficiency.yaml), which also notes an
ETFDH-CIII-COQ2 metabolon that routes lipid-derived electrons into the respiratory chain.

Annotation review reasoning

Deep research (falcon) file was NOT produced within the 8-min poll window; grounding is from
UniProt Q16134, seeded GOA, the cached experimental publications above, and the MADD disorder KB.

Annotation integration, 24 October 2026

All 37 current contexts have now been reviewed: 16 ACCEPT, 12 MODIFY, seven KEEP_AS_NON_CORE and two UNDECIDED. The original 31 positions remain fixed. The previous single HuRI context expands to seven exact partner contexts: the first fresh context, MYH7B, occupies original position 8 and the other six are appended at positions 31–36. The inherited paper-title quotation is citation context only and is not evidence for any particular interaction or exact clone identity. The actual HuRI pair/method records support the contextual interactions. The archived MYH7B clone does not match the current short A7E2Y1-2 sequence, so no current-product equivalence is asserted.

Earlier statements above that dismiss HuRI hits as artifacts because of their expected compartment, or describe a mouse ETFDH knockout oxidative-stress phenotype, are historical and superseded by this evidence assessment. The seven supported interactions are retained as non-core; expected localization does not erase a measured screen interaction. The two oxidative-stress contexts remain UNDECIDED because their actual mouse source, PMID:17023274, reports oxidative modification and the complete response-role experiment remains inaccessible. The inherited GO_REF quotations describe transfer methods and do not resolve that biological role. This applies the campaign's supported-binding instruction and the standing instruction to add no new publication quotations.

Specificity refinements use independently established ETFDH dehydrogenase activity, respiratory electron transport, inner mitochondrial membrane and 4Fe–4S binding. Broader quinone binding remains supported by actually tested alternative quinones. The revised assessments distinguish human protein expressed in yeast or Sf9 cells, native porcine structure, clinical evidence, protein-group proteomics and database participant models. PMID:8306995 remains marked full-text-unavailable because only the specified original pages were recovered; that limit does not negate those operative experiments.

All 42 inherited quotation paths, text and reference associations are preserved. No new publication quotation was added. The two natural product records and five ordered legacy core objects remain unchanged at this annotation checkpoint; descriptions and cores will be assessed in the subsequent synthesis phase. This is an annotation integration checkpoint, not a computed final status or publication claim. Normal sources are preserved, including the genuine additive PMID:17023274 record. Direct URLs, assay locators and access limits are collected in ETFDH-source-guide.md.

Core synthesis, 24 October 2026

The five ordered core slots are retained as facets of one ETFDH redox enzyme: ETF-to-ubiquinone catalysis, ubiquinone binding, FAD binding, 4Fe–4S binding and the fatty-acid oxidation context of the same catalytic activity. This ordering preserves the seven inherited core quotation anchors; it does not assert five independent enzymes. The fatty-acid branch now also cites the actual human MCAD/ETF/ETFDH coupled assay (PMID:12049629) as bare supporting evidence for executed electron-transfer work. Clinical PMID:17412732 provides physiological context rather than, alone, establishing that chemical step.

Human full-cDNA ETFDH was expressed in yeast or Sf9 cells in the original experiments; the atomic structure is native porcine enzyme. The paired cofactors and quinone chemistry are well supported, while a compulsory internal electron-transfer order, universal native dimerization and physiological use of every tested alternative quinone are not asserted. The processed canonical chain and the alternative 570-residue product remain distinct. Product processing and the role of oxidative modification are retained as explicit questions.

All 37 accepted annotation/source/review objects and nineteen reference identities, findings and access judgments remain exact. All 42 inherited supporting-text paths and reference associations remain exact; new support consists only of bare reference IDs. The two response-to-oxidative-stress annotations remain UNDECIDED. The normal publication caches are immutable. Separately captured genuine GOA/UniProt refreshes are proposed machine inputs, with both preimages and diffs retained; their use is not represented as unchanged machine bytes. Three genuine PAINT auxiliary normal records are also included as source-chain dependencies alongside the eleven citation records.

Final canonical validation passes all six checks, with all 82 authored narratives present in the private rendering. Computed DRAFT retains 7 advisories. All 37 current source contexts have reviewed decisions, including two explicitly unresolved oxidative-stress transfers. All 42 inherited quotation paths, texts and reference associations, two natural products and five ordered core facets remain preserved. The 14 selected publication/Reactome/PAINT normal files and two genuinely refreshed machine inputs remain exact to the accepted selections; all 27 private normal dependency files and eight unrelated baseline paths remain unchanged.