YueD (Bacillus cereus, UniProt Q8RJB2) — Native Substrate & Pocket Hypothesis

Focus: native-substrate-pocket-alphafold (free-text)
Seed hypothesis: The physiological substrate of YueD is not benzil but a naturally occurring aromatic carbonyl compound, most plausibly a quinone, and the AlphaFold substrate pocket should discriminate between candidate substrate classes.

Note on this version: Iteration 3 retrieved the actual Km/kcat constants from UniProt's BIOPHYSICOCHEMICAL PROPERTIES block (ECO:0000269, PMID 11796169). These data correct an iteration‑2 misreading in which the UniProt FUNCTION‑field phrase "in decreasing order" was taken as a substrate‑preference ranking. The quantitative kinetics actually support the seed's premise that 1,4‑naphthoquinone has a far lower Km than benzil. The judgment below reflects the corrected analysis.

Executive Judgment

Verdict: Partially supported. The seed's structural/kinetic reasoning holds up better than a first pass suggested, but the specific identification of the native substrate as "a quinone" remains unresolved.

• Supported — "not benzil": Benzil is a wholly synthetic screening substrate (PMID 11745140) and the weakest-binding diketone tested (Km = 768 µM). GO's obsoletion of the benzil-specific term is justified.
• Supported — pocket/kinetics alignment: The measured Km(1,4-naphthoquinone) = 27.6 µM is ~28× lower than Km(benzil) = 768 µM (PMID 11796169). The AlphaFold pocket is enclosed and aromatic/hydrophobic, sterically disfavouring the bulky, twisted benzil (10 Å, plane-RMSD 0.39 Å) relative to a flat, compact aromatic such as naphthoquinone. The pocket geometry I measured is consistent with the stated Km ordering — i.e., the pocket does appear to discriminate on steric grounds, exactly as the seed proposed.
• Unresolved — "most plausibly a quinone": By catalytic efficiency (kcat/Km), the single best in-vitro substrate is 1-phenyl-1,2-propanedione (3.93 min⁻¹µM⁻¹), ~15× better than 1,4-naphthoquinone (0.268) and ~47× better than benzil (0.084) — and it is also a synthetic aryl-alkyl diketone, not a quinone. Methylglyoxal and sepiapterin were never assayed. So the kinetics favour small/flat aromatic carbonyls in general, consistent with (but not uniquely pointing to) a quinone.
• Competing signal — structural homology: Foldseek's nearest characterized neighbours are sepiapterin reductases and SDR ketoreductases, not quinone reductases. So evolutionary context points to pterin/keto-reductase chemistry rather than a dedicated quinone reductase.
• Mechanistic caveat: In the homologous human SPR, quinone turnover can be NADPH-dependent redox cycling at the cofactor site rather than substrate-pocket reduction (PMID 23640889). For YueD, the defined Km/kcat for 1,4-NQ implies genuine substrate-pocket turnover, but the mechanism was not explicitly dissected for this enzyme.
• Over-annotation flag: sepiapterin reductase activity (GO:0004757) and tetrahydrobiopterin biosynthetic process (GO:0006729) on Q8RJB2 are IEA:TreeGrafter inferences from PANTHER PTHR44085, with no experimental support and no known BH4 pathway in B. cereus.

Bottom line for the curator: The AlphaFold model is high-confidence and the pocket does appear to discriminate sterically in a way consistent with the real kinetics (tight flat aromatics, weak bulky benzil). This supports treating benzil as non-physiological and favouring a small/flat aromatic carbonyl substrate. It does not, by itself, prove the native substrate is a quinone — a quinone is one plausible member of the favoured class, alongside other flat aromatic carbonyls and (untested) pterins. The most defensible annotation remains a general NADPH-dependent aromatic carbonyl reductase of unknown physiological substrate.

Tools & Database Versions Actually Invoked (provenance)

Honest limitations: The AlphaFold model is apo (no NADP/substrate), so pocket metrics are geometric, not energetic; no docking/MD was run. Km/kcat were taken from UniProt's curated ECO:0000269 annotation of PMID 11796169 (full text not retrieved); methylglyoxal and sepiapterin were not among the tested substrates. Foldseek hits are remote (21–28% id), reflecting the common SDR fold.

Kinetics (PMID 11796169, via UniProt Q8RJB2; NADPH, 37 °C, pH 6.5)

Interpretation: Lowest Km → 1,4-naphthoquinone (27.6 µM), tightest binder; benzil is the weakest binder (768 µM). This is exactly the ordering the seed cited and is consistent with the enclosed aromatic pocket favouring flat compact aromatics over bulky twisted benzil. Highest catalytic efficiency, however, is a non-natural aryl-alkyl diketone, so "best substrate" ≠ "quinone." (methylglyoxal, sepiapterin not tested.)

Evidence Matrix

GO Curation Implications (leads — require curator verification)

Net curation lead: Keep a general NADPH carbonyl/aryl-ketone reductase MF + cytoplasm CC; treat sepiapterin-reductase MF and BH4-biosynthesis BP as non-core phylogenetic over-annotations; annotate the physiological substrate/process as unknown. The structural+kinetic evidence supports "benzil is non-physiological" and "pocket favours flat/compact aromatic carbonyls," but does not justify naming a quinone as the core function.

Mechanistic Scope

Direct activity: NADPH-dependent reduction of an aromatic carbonyl (aryl α-diketone/aldehyde/quinone) to the alcohol, by a cytoplasmic SDR using the Ser140–Tyr154–Lys158 relay with Asn86 and an NADP(H) Rossmann site. All in-vitro substrates assayed are synthetic/xenobiotic; "quinone reductase," "sepiapterin reductase," and "tetrahydrobiopterin biosynthesis" are inferences, not demonstrated physiological roles. The pocket's steric preference for flat compact aromatics (low Km for 1,4-NQ) is a property of the enzyme; which natural flat aromatic carbonyl it acts on in the cell is unknown.

Conflicts and Alternatives

1. Efficiency vs affinity split: 1,4-NQ has the lowest Km but modest kcat; the most efficient substrate (1-phenyl-1,2-propanedione) is a non-natural diketone — so tight binding of a quinone does not by itself make a quinone the physiological substrate.
2. Family evidence cuts toward pterin/keto-reductase: PANTHER PTHR44085 (sepiapterin reductase) and Foldseek neighbours argue for pterin/carbonyl chemistry, not a dedicated quinone reductase.
3. Mechanistic category caveat: homologous SPR handles quinones by cofactor-site redox cycling (PMID 23640889); YueD's defined Km/kcat suggest true pocket turnover but this was not dissected.
4. Bacterial quinone size mismatch: the natural bacterial quinone is menaquinone (large lipophilic isoprenoid), which would not fit this small soluble pocket — so "quinone" in vivo is unlikely to be a naphthoquinone-shaped molecule.
5. Apo-model caveat: no NADP/substrate in the model; pocket metrics are geometric approximations.
6. Database carry-over: sepiapterin/BH4 terms likely propagated across the PANTHER clade (also seen on orthologs).

Knowledge Gaps

Discriminating Tests (most efficient first)

1. ΔyueD metabolomics / stress phenotyping — best route to the native substrate and BP.
2. Direct sepiapterin and methylglyoxal assays with recombinant YueD + NADPH — close the two untested candidate classes and settle GO:0004757.
3. Quinone mechanism assay (ROS/O2 consumption vs clean 2-electron reduction for 1,4-NQ) — tests whether the "quinone" activity is pocket turnover or redox cycling.
4. NADP+-bound model + docking — enables a real steric/electronic discrimination test rather than apo geometry.
5. Operon/genomic-context analysis in B. cereus — may reveal the served pathway.

Curation Leads (require curator verification)

• Reference snippets to verify
• PMID 11796169 (kinetics, via UniProt ECO:0000269): Km = 768 µM (benzil) vs 27.6 µM (1,4-naphthoquinone); kcat/Km best for 1-phenyl-1,2-propanedione → confirms NQ binds tighter than benzil but is not the most efficient substrate.
• PMID 11796169: "Green fluorescent protein-tagged B. cereus benzil reductase distributed in the bipolar cytoplasm in B. cereus cells." → GO:0005737.
• PMID 23640889: "Site-directed mutagenesis of the SPR C-terminal substrate-binding site (D257H) completely inhibited sepiapterin reduction but had minimal effects on redox cycling." → quinone activity need not be substrate-pocket-based.
• Candidate GO actions: retain GO:0004090, GO:0016616 (MF), GO:0005737 (CC); downgrade/flag GO:0004757 (MF) and GO:0006729 (BP) as IEA:TreeGrafter over-annotations; do not elevate any quinone-reductase term to core function (in-vitro only).
• Action change vs seed: accept "benzil non-physiological" and "pocket discriminates / favours flat aromatic carbonyls"; keep "quinone" as a plausible but unproven member of the favoured class, not the established native substrate.
• Suggested curator question: Does any source place a natural flat aromatic carbonyl (e.g., an o-quinone, isatin, or pterin) as a B. cereus metabolite that YueD could act on in vivo?
• Suggested experiment: ΔyueD vs WT metabolomics + targeted activity screen incl. sepiapterin and methylglyoxal.

Supported vs Refuted (summary)

• Supported: Benzil is non-physiological (weakest binder, Km 768 µM, synthetic). YueD is a high-confidence cytoplasmic NADPH SDR carbonyl reductase. Km(1,4-NQ) ≪ Km(benzil), and the enclosed aromatic pocket is consistent with that ordering — the pocket does discriminate sterically.
• Partially supported: "Native substrate is a naturally occurring aromatic carbonyl" — the enzyme clearly prefers flat/compact aromatic carbonyls, but no natural substrate has been identified.
• Unresolved / not uniquely supported: "Most plausibly a quinone" — a quinone fits the favoured class and binds tightly, but the most efficient substrate is a non-natural diketone, structural neighbours are pterin/keto-reductases, and the natural bacterial quinone (menaquinone) is too large. "Pocket discriminates between classes" — supported in the sense that bulky benzil binds poorly and flat aromatics bind well, but an apo model cannot by itself name the physiological substrate.

Limitations

Apo AlphaFold model; geometric (not energetic) sterics; kinetics taken from UniProt's curated annotation (full text not retrieved); methylglyoxal/sepiapterin untested; remote Foldseek hits; single characterization study underlies most wet-lab claims.

Artifacts

• OpenScientist candidate steric comparison
• OpenScientist final report
• OpenScientist final report
• OpenScientist go decision table
• OpenScientist kinetics table
• OpenScientist structural metrics