NDI1 (YML120C, P32340) review notes

Provenance / process

• Data fetched with just fetch-gene yeast NDI1; publications cached with just fetch-gene-pmids yeast NDI1.
• Additional papers cached with ai-gene-review fetch-pmid: PMID:9689052 (Seo et al. 1998, full text),
  PMID:11479321 (Bai et al. 2001, abstract), PMID:9733747 (Luttik et al. 1998, abstract),
  PMID:11152939 (Bakker et al. 2001 review, abstract).
• Deep research: just deep-research-falcon yeast NDI1 --fallback perplexity-lite was launched
  (see end of file for outcome). The review below is based on my own reading of the cached
  publications and PubMed searches, not on a deep-research file.
• Context: chosen from projects/MITOTOL.md, based on Chen et al. 2026 (PMID:42822426), a
  comparative mitochondrial proteome study. That paper is not annotation evidence for NDI1 function.

Identity and biochemistry

• Single-subunit, ~53 kDa, FAD-containing, rotenone-insensitive NADH:ubiquinone oxidoreductase
  (type II NADH dehydrogenase, NDH-2), EC 1.6.5.9 [PMID:3138118 "The purified NADH dehydrogenase consists of a
  single subunit with molecular mass of 53 kDa"; "The enzyme contains FAD, non-covalently linked, as the sole"].
• Specific for NADH, inhibited by flavone but not rotenone or piericidin
  PMID:3138118; uses several quinones,
  with natural UQ6 turnover ~500 s-1.
• Note: de Vries & Grivell (1988) initially suggested the purified enzyme was the external enzyme;
  the gene cloning and disruption paper (Marres et al. 1991) established that the same enzyme is the
  internal one, oxidising NADH generated inside the mitochondrion
  PMID:1900238.
  So the SGD IDA for GO:0120555 from PMID:3138118 is correctly attached to NDI1.
• No proton pumping: transfers electrons from NADH via FAD to quinone without proton translocation
  PMID:22949654;
  consistent with reduced P:O in NDI1-expressing human cells PMID:11479321.
• Hence the correct MF is GO:0120555 NADH dehydrogenase (ubiquinone) (non-electrogenic) activity
  (child of GO:0050136, which is a child of GO:0003954). GO:0008137 (NADH dehydrogenase (ubiquinone)
  activity) is defined with proton translocation (4 H+ out) and is the complex I term; it is NOT
  annotated to NDI1 (checked GOA file), which is correct. No complex I CC term (GO:0045271) is
  annotated to NDI1 either - also correct.

Location / topology

• Nuclear-encoded, 26-residue mitochondrial presequence (UniProt TRANSIT 1..26).
• Monotopic/peripheral inner-membrane protein, catalytic site on the matrix side
  PMID:22949654;
  PMID:9733747.
• Same matrix-facing topology when expressed in hamster/human cells
  PMID:9689052.
• Found in a native-gel supramolecular dehydrogenase assembly with external dehydrogenases and TCA enzymes
  PMID:11502169; this is a co-migration observation, not an obligate complex.
• Cytosol (RCA from YeastPathways PWY3O-188 GO-CAM conversion) is wrong: the GO-CAM import
  (gocams/index.tsv row gomodel:RXN3O-165) places the activity in cytosol, which contradicts the
  established matrix-side topology.

Structure

• Crystal structures (Iwata et al. 2012; Feng et al. 2012): dimer, amphipathic membrane anchor formed by dimer
  packing, overlapping NAD+ and quinone binding sites PMID:22949654; homodimerization through the C-terminal
  domain is critical for activity and membrane targeting; two UQ sites PMID:23086143.
• The IPI "identical protein binding" rows are best expressed as protein homodimerization activity (GO:0042803).

Physiology

• Main entry point of matrix NADH into the respiratory chain in yeast, which lacks complex I
  PMID:9689052.
• ndi1 null: growth on glucose and ethanol unaffected, lactate/pyruvate/acetate impaired PMID:1900238.
• Part of a set of parallel NADH reoxidation routes (internal Ndi1 for matrix NADH; external Nde1/Nde2 and
  Gut2 shuttle for cytosolic NADH) [PMID:11152939; PMID:9733747].
• Complements complex I deficiency in mammalian cells and animals [PMID:9689052; PMID:11479321; PMID:16543240 (not cached)].

Apoptosis / ROS (Li et al. 2006, PMID:16436509)

• Overexpression of NDI1 (not NDE1) causes apoptosis-like death, associated with mitochondrial ROS, suppressed by
  respiration on glucose-limited media and by interrupting ETC components; deletion lowers ROS and extends
  chronological lifespan. These are gain-of-function / pleiotropic effects consistent with Ndi1 being a
  flavin-containing NADH oxidising enzyme that can leak electrons; there is no evidence that Ndi1 acts in a
  regulated apoptotic signalling step. The GO:0043065 IMP annotation is therefore treated as over-annotation
  (not removed: experimental, full text not cached).

Evolutionary context (Chen et al. 2026, PMID:42822426)

• Comparative analysis reconstructs NDH2 in LOCA/LECA mitochondria; retained in yeast, lost in humans
  PMID:42822426.
• NDH2 recovered as a candidate pathogen drug target absent from human mitochondria. Background only.
• Human AIFM1/AIFM2 are distantly related flavoproteins (NDH-2/AIF superfamily); human has no orthologous NDH-2.

Decisions summary

• ACCEPT: MF terms GO:0120555 (x3), GO:0050136, GO:0003954, GO:0016491 (x2, general); BP GO:0006120,
  GO:0019646, GO:0015980; CC mitochondrion (x5), inner membrane, matrix (x2).
• MODIFY: GO:0042802 identical protein binding (x2) -> GO:0042803 protein homodimerization activity.
• MARK_AS_OVER_ANNOTATED: GO:0043065 positive regulation of apoptotic process.
• REMOVE: cytosol (RCA, YeastPathways GO-CAM conversion artifact).
• NEW: GO:0071949 FAD binding; GO:0099617 matrix side of mitochondrial inner membrane.

Deep research outcome (2026-10-09)

timeout 1500 just deep-research-falcon yeast NDI1 --fallback perplexity-lite FAILED:
falcon timed out at the wrapper's 600 s limit, and the perplexity-lite fallback failed with
"Provider 'perplexity' not available. Available: falcon, asta, openscientist". No deep-research
file was produced (the orphaned falcon client process was killed). The review relies on the cached
publications listed above plus the reviewer's own PubMed searches (PubMed MCP).