NQO2 (P16083) — Function-Assignment Hypothesis Review

Hypothesis: NQO2 has NAD(P)H dehydrogenase (quinone) activity (GO:0003955). Focus: function_assignment · existing IBA annotation (GO_REF:0000033) · slug function-hypothesis-go-0003955 Gene: human NQO2 / UniProt P16083

Executive Judgment

Verdict: Over-annotated → recommend REMOVAL of the IBA GO:0003955 (activity real, but this specific term is substrate-incorrect and redundant).

NQO2 is unequivocally a flavin-dependent, two-electron quinone reductase, so the quinone-reductase concept is correct. But the specific term GO:0003955 "NAD(P)H dehydrogenase (quinone) activity" (reaction: NAD(P)H + quinone → NAD(P)⁺ + quinol) makes a cofactor/substrate claim that NQO2 does not satisfy. The definitive enzymology (Wu et al., 1997, 9367528) shows NQO2 "uses dihydronicotinamide riboside (NRH) rather than NAD(P)H as an electron donor." UniProt (P16083) codifies this as EC 1.10.5.1; the NAD(P)H reaction (EC 1.6.5.2) belongs to the paralog NQO1 (P15559).

Two ontology facts (verified via QuickGO this iteration) make the recommendation removal rather than generalization: 1. The biochemically exact term GO:0001512 "dihydronicotinamide riboside quinone reductase activity" (NRH + quinone → nicotinamide riboside + hydroquinone) is already annotated to NQO2 with experimental evidence — IDA, 18254726 (plus IEA GO_REF:0000120). 2. GO:0003955 is NOT an is_a ancestor of GO:0001512. They are siblings in different oxidoreductase subtrees, so GO:0003955 cannot be defended as a merely-less-specific but still-true parent. It is a distinct, incorrect molecular function.

The IBA is therefore a paralog over-annotation (GO:0003955 native to NQO1, propagated across the shared PANTHER family via GO_REF:0000033), and it is both wrong on substrate and redundant with the correct experimental term.

Most important caveat: Do not delete NQO2's reductase function from the model — it is captured accurately by GO:0001512. The action is limited to the mis-specified IBA row.

Evidence Matrix

GO Curation Implications

Lead (requires curator verification):

• Molecular Function — REMOVE the IBA GO:0003955 from NQO2 (or mark NOT / do-not-propagate). It asserts NAD(P)H substrate that NQO2 cannot use and is a paralog carry-over from NQO1.
• Retain the already-present GO:0001512 "dihydronicotinamide riboside quinone reductase activity" (IDA, 18254726) as the accurate leaf MF term. No new term needs to be created.
• Because GO:0003955 is a sibling (not ancestor) of GO:0001512, keeping it is not a harmless generalization — it introduces a false substrate assertion into the model.
• Broader true ancestors already annotated (GO:0016491 oxidoreductase activity; GO:0016661 acting on other nitrogenous donors; GO:0009055 electron transfer activity; GO:0048038 quinone binding; GO:0071949 FAD binding; GO:0008270 zinc ion binding) remain valid. Avoid defaulting to generic "protein binding."

GO decision table

Mechanistic Scope

• Immediate molecular function tested: FAD-mediated two-electron reduction of quinones to hydroquinones, electrons supplied by a dihydronicotinamide riboside (NRH) donor in a ping-pong mechanism; Zn²⁺ is structural.
• Correctly in scope: quinone reduction (menadione and other quinones, estrogen o-quinones, vitamin K quinones), NRH-dependent nitroreduction (CB1954). Captured by GO:0001512.
• Out of scope for this MF term (BP/CC or ligand-binding): neuronal "memory constraint"/metabolic-buffer role (Rosenblum lab), Alzheimer's phenotype modulation, melatonin MT3 binding-site identity (GO:1904408) and resveratrol binding (GO:1905594). These are downstream physiology or ligand properties, not the electron-donor chemistry adjudicated here.

Conflicts and Alternatives

• Paralog confusion (primary): GO:0003955 is correct for NQO1 (EC 1.6.5.2, NADH/NADPH). Its presence on NQO2 is explained by IBA propagation, not NQO2 biochemistry. Structural basis: NQO2 (231 aa) lacks the ~43-residue C-terminal region present in NQO1 (274 aa) that helps form the adenosine/2′-phosphate NAD(P)H-binding site, so NQO2 uses the smaller NRH.
• In-vitro cofactor surrogates: NQO2 assays use synthetic donors (BNAH; EP-0152R) because free NRH is not a standard bulk metabolite; consistent with EC 1.10.5.1, not with NAD(P)H use.
• No credible primary report shows efficient NAD(P)H-driven catalysis by NQO2.
• Self-correction: Iteration 1 tentatively named the replacement term GO:0033856; that ID is actually pyridoxine 5′-phosphate synthase activity. The correct term is GO:0001512, confirmed via QuickGO.

Knowledge Gaps

1. Physiological NRH source in vivo. Checked: literature uses in-vitro NRH/BNAH surrogates. Matters for BP context, not the MF term. Resolve via tissue metabolomics for NRH and NRH-generating enzymes.
2. Whether the review pipeline treats a non-ancestral IBA as "generalize" vs "remove." Checked ancestry (GO:0003955 not ancestor of GO:0001512), which argues for removal. Curator should confirm project policy.
3. Quantitative NAD(P)H vs NRH kinetics. Existing data are qualitative (NRH-preference). A kcat/Km ratio would formally bound the error; confirmatory only.

Discriminating Tests

• Side-by-side kinetics: NQO2 quinone reduction with NRH/BNAH vs NADH vs NADPH → expect robust NRH activity, negligible NAD(P)H (discriminates GO:0001512 from GO:0003955).
• Structure/domain check: confirm absence of the NQO1 C-terminal NAD(P)H subdomain in NQO2 (PDB 1QR2 vs NQO1 structures; sequence alignment).
• Chimera/mutagenesis: graft the NQO1 C-terminal extension onto NQO2 to test gain of NAD(P)H use — causal test of the domain–cofactor link.

Curation Leads (require curator verification)

• Candidate action: Remove/NOT the IBA GO:0003955 on NQO2; retain existing IDA GO:0001512.
• Candidate reference to verify (donor specificity): 9367528 — "NQO2 uses dihydronicotinamide riboside (NRH) rather than NAD(P)H as an electron donor. It catalyzes a two-electron reduction of quinones and oxidation-reduction dyes."
• Candidate reference supporting the correct term (already the IDA source): 18254726 (Calamini et al., 2008, QR2 kinetics + X-ray) — verify it is the IDA basis for GO:0001512.
• Context reference: 10945627 (CB1954 co-substrate-mediated bioactivation) confirms NRH-type co-substrate, not NAD(P)H.
• Database support: UniProt P16083 (EC 1.10.5.1, NRH→quinol reaction, FAD+Zn²⁺); QuickGO ancestry showing GO:0003955 is not an ancestor of GO:0001512.
• Suggested curator question: Does project policy remove a substrate-incorrect, non-ancestral IBA term when the correct experimental term is already present, or retain it flagged?
• Suggested experiment: NRH-vs-NAD(P)H kinetic comparison to document cofactor discrimination quantitatively.

Provenance (executed this review)

• UniProt REST: NQO2 P16083 = 231 aa, EC 1.10.5.1, NRH→quinol, FAD+Zn²⁺; NQO1 P15559 = 274 aa, EC 1.6.5.2, NAD(H)/NADP(H). Confirms paralog cofactor divergence + 43-aa length difference.
• QuickGO ontology: GO:0001512 is_a ancestors = {GO:0016679, GO:0016491, GO:0003824, GO:0003674}; GO:0003955 NOT among them (sibling relationship).
• QuickGO annotation (UniProtKB:P16083, MF): GO:0001512 = IDA (18254726) + IEA; GO:0003955 = IBA only (GO_REF:0000033).
• NCBI eSummary: 18254726 = QR2 melatonin kinetics/X-ray; 10945627 = CB1954 co-substrate bioactivation.
• Sequence comparison (UniProt FASTA, computed): NQO2 231 aa vs NQO1 274 aa; NQO1 carries exactly 43 extra C-terminal residues (…NFQAGFLMKKEVQDEEKNKKFGLSVGHHLGKSIPTDNQIKARK) absent in NQO2 — the C-terminal segment forming part of the NAD(P)H adenosine-binding site, i.e. the structural reason NQO2 cannot use NAD(P)H.
• Literature: 9367528 (definitive), 18996184, 21506232, plus review 18374191.

Artifacts generated

• NQO2_GO_decision_table.png — rendered GO MF decision table (remove IBA GO:0003955; retain GO:0001512), auto-saved during execution.
• Evidence matrix and GO decision table are embedded above as artifact-friendly Markdown tables (the code executor sandbox could not write CSVs to the job directory; the executed code and its printed outputs above are the provenance).