Read every source annotation and relevant evidence, including primary full texts for disputed reactions and the exact PAINT target path. See AKR1D1-primary-source-checks.md for assay context, access limits, short exact excerpts, and pending focused adjudication. All original source fields are preserved; no NEW annotation is added. Broad true chemistry and localization are retained separately from exact substrate/reaction conflicts.
This entry supersedes earlier functional judgments where they differ, including the early broad-process over-annotation claims, the description of aldose reduction as merely a less-specific synonym of steroid reduction, and the earlier provider-access description. Historical notes, the September 20 source audit, PAINT lineage and genuine OpenScientist artifacts remain unchanged as provenance.
Identity: the authoritative HGNC snapshot has approved AKR1D1, HGNC:388, previous symbol SRD5B1. The live UniProt P51857 record confirms the human protein and synonym. Canonical/historical directories and open PR searches found no overlap. Parent verified all 11 gene files against main c7078166039c9abd5c62704489283403eb520007; an older local HTML rendering was backed up and restored to exact main before authoring. The baseline was INITIALIZED, with 41 reviewed source rows, 25 references and three isoforms. Every original annotation source object and reference identity/title is preserved, including the unresolved Reactome title placeholder. No NEW annotation is added.
Cached full human studies PMID:21255593 and PMID:18407998 were independently read. The kinetic study uses purified human recombinant enzyme, TLC product comparison with authentic standards, cofactor fluorescence titration, substrate kinetics and inhibition assays. It directly identifies 5beta products for the tested C18-C27 substrates except aldosterone; aldosterone turnover is observed but its product is inferred because an authentic standard was unavailable. Substrate inhibition varies by structure, and the authors discuss how single-concentration/cell preparations contributed to older discrepancies. These results support direct participation in androgen and C21 hormone metabolism alongside bile acid synthesis, without measuring each substrate's share of flux in vivo. PMID:21255593(https://pmc.ncbi.nlm.nih.gov/articles/PMC3056882/), "5β-Reduced products were identified directly with all the C18-C27 steroid substrates except for aldosterone."
PMID:18407998 structures resolve NADP+/steroid complexes and the Tyr58/Glu120 environment; mutant assays use testosterone reduction. Steroid binding is retained as integral substrate recognition. These experiments explain double-bond reduction while retaining the 3-oxo group, not a universal negative result for every sugar or alcohol-forming substrate. PMID:18407998(https://pmc.ncbi.nlm.nih.gov/articles/PMC2423251/), "Each steroid carbonyl accepts hydrogen bonds from catalytic residues Tyr 58 and Glu 120 ."
The original human cDNA/COS-cell study PMID:7508385(https://pubmed.ncbi.nlm.nih.gov/7508385/) remains abstract-only and its full assays were not recovered. Positive bile-acid-intermediate, cortisol and testosterone reactions are retained, while its progesterone/androstenedione negatives are explicitly preparation/condition-specific. Original localization assertions retain curator deference plus independent UniProt, HPA and Reactome corroboration; no fractionation/imaging method is invented from the abstract. Digestion remains UNDECIDED: GO:0007586 concerns breakdown of ingested nutrients, whereas the exposed assay measures bile-acid-precursor reduction. The complete source could provide additional context.
The PMID:11342103 primary abstract explicitly uses human AKR1D1 stably expressed in HEK293 cells. UniProt maps RHEA:53484, androst-4-ene-3,11,17-trione to the C17-hydroxy product, to this paper. Exact product/controls were not exposed in the abstract and the original full body could not be recovered. Both this 17beta-HSD mapping and the ARBA alcohol-dehydrogenase assertion remain UNDECIDED. A possible endogenous cell-enzyme contribution is a question, not an established contaminant or refutation. Citation identity and the study's positive broad scope are VERIFIED; the particular side reaction remains unresolved.
The PMID:21232532 primary abstract describes expression/IHC of progesterone-metabolizing enzymes, explicitly including AKR1D1. Its title/low expression finding cannot establish an absent monooxygenase assay elsewhere in the full paper. Publisher/full-source recovery failed. The monooxygenase IBA remains UNDECIDED, as does ancestral aldose reduction. A bounded independent annotation_aars1 consultation checked both abstracts/UniProt and attempted full-source recovery, supporting these access-qualified judgments.
Eight existing Reactome caches were read, separating pathway summaries from reaction equations. Five cached reduction events retain a 3-oxo steroid carbonyl and therefore support GO:0047787 rather than GO:0008106. A new indexed primary retrieval recovered the sixth human record Reactome:R-HSA-193755, including the complete triol-steroid equation, NADPH, AKR1D1 catalyst, cytosol and PMID:7508385. It is likewise C4-C5 reduction, so the sixth MF row now receives the same source-specific MODIFY. Direct web retrieval returned 404 and the standard cache function failed DNS; the indexed human record supplies positive evidence without inventing a replacement identifier or cache. Raw evidence: /tmp/AKR1D1-193755-primary-evidence.json. The original title placeholder remains pending normal machine cache recovery.
GO definitions/parents were checked through local OAK. GO:0016229 means steroid-substrate redox chemistry and is retained as an informative family-level assertion, without requiring carbonyl/alcohol conversion. Generic GO:0016491 is refined to the experimentally resolved double-bond reductase. Broad cytoplasmic localization is retained at source resolution. GO:0006707 and monocarboxylic-acid metabolism encompass pathways in which AKR1D1 directly performs an intermediate step; it need not consume cholesterol or a free acid itself. Live AmiGO requests timed out; no unverified label/obsoletion change is claimed.
The preserved PAINT lineage was read in full: PTN000198921 grounds aldose reduction/cytosol; PTN000199026 grounds steroid dehydrogenase/monooxygenase; PTN000199134 grounds androgen metabolism. The target P51857 path and IBD rows are explicit. Its own androgen/cytosol experiments are legitimate descendant grounding, not circularity. No new tree/MSA reconstruction or target-specific loss is claimed. Propagation blocks trace all IBA/IEA proximate sources; unread ARBA predicates remain UNRESOLVED even where independent human biology supports the annotation. UniProt Rhea equations were checked individually, including physiological reduction when a reaction is written in reverse. No matching AKR1D1 GO-CAM index entry was found.
BioPlex PMID:28514442 and PMID:33961781 were read at cached methods/full-section scope. The seeded partner is Q04828/AKR1C1; the exact pair supplement was not independently extracted. Generic binding is removed because no informative specific MF is established, without asserting noninteraction. The latter cache has partial full-text sections despite a true availability flag, which does not imply all tables/results were examined.
Two prior activity-identical cores are consolidated into one cytosolic NADPH-dependent steroid 5beta-reductase core covering bile-acid synthesis, cholesterol breakdown and hormone metabolism. Steroid recognition is incorporated in its description. No extra process, cofactor-binding annotation, isoform specificity, transport function or redundant NEW is manufactured.
The genuine prior OpenScientist report and PDF/HTML/citation artifacts were read and preserved. Its conclusion that steroid specialization excludes other chemistry is stronger than its acknowledged lack of direct negative assays/full donor papers, and is not adopted. Its sequence alignment was not independently rerun. A fresh normal Falcon request and perplexity-lite fallback used isolated /tmp/AKR1D1-fresh-research/AKR1D1/ output. A launcher import typo failed before any invocation and was corrected. Both genuine provider commands then exited 2 during uvx dependency retrieval because PyPI DNS failed; no provider ran and no new report exists (/tmp/AKR1D1-fresh-research.log).
Parallel normal caching found all seven existing review PMIDs cached. Six missing records in preserved provider artifacts were requested normally: PMID:26418565, PMID:30254413, PMID:31337596, PMID:36739965, PMID:38034430 and PMID:41387259. All six attempts failed DNS and produced no cache (/tmp/AKR1D1-fetch-provider-citations.log). Several identities were independently verified through primary PubMed; these clinical/mutant leads are not substituted for the decisive substrate assays. Reactome:R-HSA-193755 remains absent after standard cache_reactome_pathway retrieval (/tmp/AKR1D1-fetch-reactome.log). DRAFT is retained for these notes/provider-inclusive source gates. Local full-text flags reflect cache metadata/content, not external reading. Final checks and independent review are recorded in the handoff manifest.
Final independent parent review read all 41 decisions, 25 reference assessments, the integrated core, questions and saved human Reactome equation; no biological change was requested. Targeted validation, history validation and rendering pass. The sole advisory reflects the deliberate GO:0008106 distinction between six resolved Reactome mapping corrections and the unresolved independent ARBA side activity. All 57 supporting snippets match their cached sources exactly after case-sensitive whitespace normalization. All source objects, three isoforms, reference identities and eight protected artifacts are unchanged. The final citation census separates the six notes/provider gaps from unused bibliography-only entries in the immutable UniProt record.
The live PR remained draft at 4e543bbf4191a0a39ee6edfccd2d042ad1c6f6dd before
editing. All 11 canonical gene/artifact files and the published history matched
that exact head. The 41 source assertions, three isoforms, original reference
identities, all action decisions and the single core are preserved. Seven
reference assessments are added: six previously cited provider records and the
reviewer's already-cached mechanistic source, PMID:20522910.
The normal source5 recovery run
ran at 60c5e96f8317dd1e7d8325242d8037b81e272c59. Artifact 10926007634 has ZIP
SHA256 cd022c3e045b00798f8e29ff86904be885a4ed9e8b539ba11d44f6b954149400.
The independently verified import copied these records unchanged:
736edb02fc30da0e99d9100edf54a014ebe0c64d5eb41d08ab45ef212bf6dc81.e0277829914bf18588b0e7db2953dac76696540a2a59ca3760e1a404ecd575dd.ef5c0c8295e51bb7d45637a47bbddae5e2769e69ad198e2cdb40813efd153594.17dc234a6a4ba2fcb89f459ecc52c35e1a62544f2f3c5e41573b7de40ccef74c.171a5fed7afadd595fa550d1e580c25ab375d33f05c0087531d149f83187d28e.27012c7018585e3c846c00dbbdc4f00cfb73cbc4b73e487449fd8d71d85d076a.PMID:26418565 has extracted full Methods, Results and Discussion. Human WT and
P133R proteins were expressed in E. coli and purified using affinity plus Blue
chromatography. Cortisone and 7alpha-hydroxycholest-4-en-3-one assays, cofactor
fluorescence/chase measurements and steroid-binding titrations show diminished
P133R chemistry and cofactor affinity. The numerical effects depend on buffer,
substrate and cofactor occupancy. This source explicitly corrects the earlier
PMID:20522910 conclusion of unchanged cofactor affinity: substantial bound
NADP+ in the older WT preparation obscured the comparison. The later cofactor-
free preparations reveal the affinity difference. The narrow historical finding
is now machine-readably OVERTURNED with exact later-source support. The older
study still supports reduced mutant protein/activity; purified P133R results
are not generalized to every mutant. Neither study resolves the aldose,
monooxygenase or C17-side-activity source gaps.
PMID:30254413 has extracted full clinical case/discussion text. Genetic findings
and urinary profiles support disease context, but the R307C structural model
is predictive and the first urine profile was obtained after treatment. It
provides no purified variant-specific catalytic measurement. PMID:38034430 has
full case-series Methods/Results/Discussion: three of five patients lacked liver
dysfunction despite variants/atypical bile acids. The proposed MRP3 compensation
is explicitly a hypothesis; short follow-up and absent follow-up biopsy limit
clinical conclusions. Neither observation negates AKR1D1's established reaction.
The other three recovered records are abstract-only. PMID:31337596 distinguishes
three AKR1D1 cases from three CYP7B1 cases; the CYP7B1 allele-frequency and CDCA
response results are not assigned to AKR1D1. PMID:36739965 supports an RNA/splicing
mechanism from liver RNA and a minigene experiment, not a purified catalytic
assay. PMID:41387259 reports one fatal infant case with biochemical/genetic
support; its screening suggestion is an author proposal. The three full-body
records have full_text_unavailable: false; these three abstracts and the
existing abstract-only PMID:20522910 have full_text_unavailable: true.
Review comment 5853621026
approved the biology and offered optional refinements. PMID:20522910 is now
explicitly assessed, with the later cofactor correction recorded rather than
repeated as a current fact. The aldose reason explicitly acknowledges the
C5-oriented hydride-transfer geometry in full PMID:18407998. That positive
structural explanation supports steroid specialization but does not replace a
substrate-specific negative assay or resolve the inherited donor evidence.
Likewise, the known NADPH reductase mechanism does not independently adjudicate
the unread oxygen-insertion experiment. UNDECIDED is retained for those rows.
The steroid-dehydrogenase IBA retains its informative ancestral steroid-substrate
class; the generic oxidoreductase signature lacks that substrate restriction.
The exact 5beta reaction remains separately represented. Monocarboxylic-acid
metabolism remains core pathway participation: AKR1D1 performs a chemical step
in making monocarboxylic bile acids, and its immediate substrate need not already
carry the product's acid group. The number of non-core decisions in other gene
reviews is not biological evidence for changing this process judgment.
Reactome:R-HSA-193755 was not recovered: the local attempt failed DNS, and the
hosted source5 normal call returned 1 without an output record. A direct page
404 does not prove that the event never existed or establish a replacement ID.
The original GOA identifier and machine-fetched title placeholder are retained.
The earlier indexed primary human page explicitly identifies the reaction
of 4-cholesten-7alpha,12alpha,24(S)-triol-3-one with NADPH to yield the corresponding
5beta steroid and NADP+, and identifies AKR1D1 in cytosol. This retained-carbonyl
equation supports the existing reaction-specific MODIFY; its indexed primary
verification is separate from cache availability. The absent cache remains a
draft gate. No authored Reactome substitute is created.
The recursive census includes YAML, both source-note files, PAINT JSON and all
unchanged OpenScientist Markdown/citation/HTML/PDF artifacts. URL/HTML decoding,
Markdown escape normalization and PDF text extraction find no additional
DOI-only work in those retained artifacts. With PMID:20522910, all 14 cited
PMIDs are cached; eight of the nine Reactome records are cached, leaving only
R-HSA-193755. The immutable UniProt bibliography is distinguished from the
review/provider citations. The preserved hypothesis report's stronger chemical
exclusions are not adopted, and none of its sequence analysis is represented
as newly rerun.
Final checks for this follow-up passed: targeted validation exited 0 with the
existing evidence-specific GO:0008106 action advisory; history validation and
rendering passed. The independent integrity check confirms all 41 source
objects/actions, three isoforms, the core, 25 original reference identities,
eight protected artifacts and both published history records unchanged. All
60 quoted snippets match actual cached text after case-sensitive whitespace
normalization. The PDF census additionally examines all 33 annotation URI
links and both pypdf and pdftotext extraction, and the HTML scan includes raw
link targets. Exact current-main cache metadata and blob checks distinguish
the six new source5 records from already published sources. These checks do
not resolve the remaining Reactome cache gate or the source-specific biological
uncertainties retained in the annotations.