FMO3 is an ER-membrane, FAD-containing monooxygenase whose defining adult hepatic activity is trimethylamine oxygenation. This review retains all 23 original annotation contexts: nine ACCEPT, nine KEEP_AS_NON_CORE and five MODIFY. There are no newly proposed annotations, publication quotations or natural alternative products. The individual source fields, original partner accessions and source order remain unchanged.
Original study 9776311 supplies authentic abstract evidence for recombinant human FMO3 and adult human liver microsomal TMAO production. Original study 17531949 directly measures radiolabeled TMA and TMAO by HPLC. Its human C-terminally His-tagged protein was expressed in insect cells, and kinetic measurements used membrane fractions because purification was limited. Separate FAD and immunoreactive-protein measurements distinguish holoenzyme assembly from protein abundance. Neither the host organism nor disease substitutions define a different natural FMO3 product. The paper's human structural model uses a yeast template.
Original study 29116146 studies purified human WT and N61S protein obtained from E. coli membranes. ITC directly supports NADP+ binding; stopped-flow spectroscopy follows the bound flavin's reduction and reoxygenation. TMA turnover is inferred from substrate-dependent NADPH consumption in this paper, whereas its benzydamine/tamoxifen assays quantify product by HPLC. These readouts are not interchanged. NADP+ retention stabilizes the oxygenating intermediate; the N61S defect is not attributed to normal FMO3.
Original study 10759686 directly demonstrates recombinant human FMO3 microsomal albendazole sulfoxide production. The human-liver analysis favors a larger CYP contribution, particularly CYP3A4, and does not establish an exclusive FMO3 route. The source's nonsaturating, solubility-limited kinetic estimates are not promoted to exact physiological fractions. This activity remains contextual.
Original study 32156684 directly tests human FMO3 baculosomes at 2.5 mM hypotaurine. Figure 3B quantifies taurine by NMR peak integration (n=2, half-range); FMO3 produces substantially less than FMO1 under the tested conditions. That is actual catalytic participation and supports the existing hypotaurine and taurine-synthesis annotations as non-core. The displayed FMO1 spectra, FMO1 high-resolution product confirmation and kinetic constants, and Fmo1/2/4-null mouse experiment are not FMO3-specific assays. Selected Methods, Results and figure legend were read; ROOT and independent peer also inspected the retained Figure 3 image. No whole-supplement audit is claimed.
GO:0004499 has a specific N,N-dimethylaniline reaction definition even though its synonyms include broad historical FMO language. Original study 9536088 has verified identity and an authentic abstract, but its exact dimethylaniline assay remains unverified because full Methods were not recovered. Existing curated source support, the reviewed RHEA:24468 mapping and a verified positive ancestral placement justify contextual retention without claiming a directly inspected substrate experiment. GO and UniProt share the paper and are not independent experiments. TMA assays do not themselves demonstrate dimethylaniline oxidation; no ancestral loss or wrong-gene inference follows from the access gap.
The five positive source contexts receive partner-class refinements, without acquiring partner catalytic or regulatory activities. Exact human-protein IntAct records use a yeast host. Three records for PMID:25910212 and one for PMID:31515488 preserve CREB3 O43889-2; the latter is the displayed product named isoform 1/LZIP, not the product named small LZIP. The distinct unsuffixed record is not substituted. DNA-binding transcription factor binding (GO:0140297) is supported. Binding specifically to processed or DNA-bound CREB3, endogenous transcription regulation and pair-specific GPCA/PCA validation are not established. GO:0008140 remains an unselected narrower option because its CREB3 family scope was not resolved by the bounded ontology/usage check.
For PMID:32296183, three positive source records each support the original CLEC10A Q8IUN9, SGSM3 Q96HU1 and SACM1L Q9NTJ5 pairs. Their reviewed partner classes support signaling receptor binding, GTPase activating protein binding and phosphatase binding, respectively. SGSM3 is a GAP, not a small GTPase; SACM1L is a lipid phosphatase, so protein phosphatase binding would be too narrow. The CLEC10A association is not evidence that FMO3 is an agonist, glycan ligand or antigen-presentation regulator. Dataset-level MAPPIT/GPCA benchmarks do not establish an orthogonal assay for every target pair. Exact records and original method URLs are listed in the source guide.
The current PMID:25910212 normal cache lacks the relevant Methods despite its full-text flag. Those selected original passages were read from an older genuine, preserved fetch, not by editing the normal cache. Original supplementary target cells and complete insert sequences were not reconstructed. Multiple records from one study are not independent publications.
Reviewed human P31513 explicitly transfers ER/microsome localization from verified rabbit FMO3 P32417. The rabbit donor's microsomal purification/sequencing provenance supports conserved localization, not a newly performed human imaging experiment or measured membrane-side/topology assignment. The two older rabbit abstracts were used only to trace the curated donor; they were not added as new normal-source dependencies.
The original Reactome graph for R-HSA-139970 resolves through catalyst 114712, FMO3:FAD complex 217276 and protein 54929 to human P31513 explicitly in GO:0005789. Substrate ER-lumen labels are not assigned to the enzyme. R-HSA-5602966 is a FailedReaction through catalyst 9631786, complex 9631787 and set 5603019. Its five human P31513 mutant members retain the ER-membrane compartment; failed catalysis does not mean negated localization or normal abundance/activity for every allele.
The complete 1488-node PTHR23023 tree/MSA includes the exact 532-residue P31513 target. Original IBD data place GO:0004499 positively at PTN000545689 (Deuterostomia) and GO:0042412 at PTN000545693 (Tetrapoda), both on the target path. Four genuine family records were derived using the repository parser/writer from the pinned original IBD file. Tree/MSA and IBD retrievals are separately pinned; a shared release identity is not assumed. The target's appearance among descendant evidence is expected and not circular. No donor-count argument, speculative loss or bulk leaf-GAF acquisition is used. Unchecked prose in inherited family metadata was not used as primary evidence.
This is manual primary-source and database research, with independent enzyme/binding consultation and a ROOT check of ER/PAINT provenance. No provider-named research file was fabricated. Source access and normal-cache completeness are separate properties. The published campaign instructions govern reference-only support without new quotations; informative binding refinements are used where supported. The single TMA core does not add a new GO biological-process assertion or a second generic monooxygenase core.
The CLEC10A refinement describes the receptor class of the measured binding partner. The official GO:0005102 term distinguishes receptor binding from the additional agonist activity that may be appropriate in other cases. Exact positive yeast two-hybrid records establish the reported interaction capacity, while native encounter between hepatic/ER FMO3 and myeloid cell-surface/endolysosomal CLEC10A remains unestablished. Distinct reported contexts do not alone prove that the experimental interaction is false.
The TMA and hypotaurine terms have no is_a path to GO:0004499 in the checked official hierarchy. The dimethylaniline rows therefore remain contextual specific activity assertions; common historical FMO terminology does not justify treating them as redundant ancestors. FAD/NADP binding supports the already stated FAD-containing, NADPH-dependent catalytic core. The nicotine reaction in reviewed P31513 is recorded as a question for later primary-assay assessment, without adding a source or annotation.