ybbO (b0493, P0AFP4) — Escherichia coli K-12 — curation notes

2026-10-30 session

• Deep research providers (falcon, perplexity) are not configured in this environment;
  literature found by PubMed search (ybbO[tiab], YbbO AND coli) and cached.
• Classical SDR, PTHR43313 (SDR family 9C) subfamily SF1. One-to-one orthologue of
  Salmonella STM0505 (Q7CR11): reciprocal best hit, conserved ybbN/cnoX-ybbO-tesA-ybbA
  neighbourhood, 90.6% identity (file:SALTY/ybbO/ybbO-bioinformatics/RESULTS.md).
  E. coli numbering is +13 relative to Salmonella because P0AFP4 carries a 13-residue
  N-terminal extension (MTHKATEILTGKV); UniProt ACT_SITE Tyr159 = Salmonella Tyr146.

Activity evidence (all from metabolic-engineering contexts)

• Short/medium-chain aldehydes, in vitro, purified enzyme:
  PMID:25108218; "all five appear to be strictly
  NADPH dependent with the substrates tested"; about 10-fold lower than DkgA/YahK.
• Long-chain fatty aldehydes: PMID:27134112; kinetics C6-C18, best affinity C18, best efficiency C16 (abstract only).
• Retinal: PMID:25726762; overexpression raises the retinol fraction. Independently,
  PMID:34361955 and PMID:41093029. YbbO is used heterologously as the retinal
  reductase for retinol production in yeast (PMID:38647788, PMID:39307776).
  Retinal reduction fits the SDR9C family, whose vertebrate members (RDH16, RDH5,
  HSD17B6, DHRS9) are retinol / hydroxysteroid dehydrogenases.
• Negative: ybbO deletion had no effect on glutaraldehyde survival
  PMID:34248896.

Open issues

• Physiological substrate unknown: E. coli makes neither retinoids nor much free
  long-chain fatty aldehyde. Neighbour tesA (acyl-CoA thioesterase / lysophospholipase)
  hints at lipid metabolism but is not evidence.
• The 13-residue N-terminal extension is shared with Shigella, Klebsiella and
  E. marmotae entries but absent from most Enterobacterales orthologues (256-261 aa); it
  may reflect an upstream start call. Not recorded as a gene_model_issue: no proteomic or
  N-terminal sequencing evidence was examined.

Decisions

• GO:0008106 IDA x2: ACCEPT.
• GO:0016491 IBA: MODIFY -> GO:0008106 (same as Salmonella review).
• NEW GO:0052650 all-trans-retinol dehydrogenase (NADP+) activity was proposed and then
  withdrawn (see below); recorded as a suggested question instead.

Revision after OpenScientist (2026-10-30)

The blinded OpenScientist run on the Salmonella orthologue
(file:SALTY/ybbO/ybbO-hypotheses/seq-only-blinded-activity/openscientist.md) judged retinal
reduction promiscuous and non-core, citing PMID:34850377. Together with the lack of any purified-enzyme retinal assay, no measured
cofactor, and a substrate absent from E. coli, this does not clear the bar for a NEW
annotation. The retinal evidence stays in the description, references and a suggested
question.

Correction after PR review (2026-10-30)

The retinal evidence was one-sided. PMID:39307776 (95.6% retinal-to-retinol
conversion among three alcohol dehydrogenases screened in yeast) contradicts the
PMID:34850377 result. Yeast efficiency is therefore disputed, not established as low.
The decision to withhold GO:0052650 rests on the other grounds: no purified-enzyme assay,
no measured cofactor, and a substrate absent from E. coli. The reference, the conflict, a
SUSPECTED WRONG_START_SITE gene_model_issues entry, and description fixes (dropped the
unsupported "cytoplasmic" and the CDD-derived "17beta-HSD-like" family gloss) are now in
the review.