PNPO

Summary: Core catalytic molecular function of PNPO. The IBA propagation across the pyridoxamine 5'-phosphate oxidase family is correct and matches the experimentally established human activity (EC 1.4.3.5). The GO term definition explicitly covers oxidation of both pyridoxamine 5'-phosphate and pyridoxine 5'-phosphate to PLP.

Summary: Mitochondrial localization is not supported for human PNPO, which is a soluble cytosolic enzyme (Reactome places both reactions in the cytosol). This IBA call is driven by a subset of family members with divergent localization and represents an over-annotation for the human protein. Not removed (IBA can reflect genuine biology in some orthologs), but flagged. ORGANELLES re-check (2026-11-03) revises this: a minor mitochondrial pool of human PNPO has two independent proteomic lines, so the row is kept as non-core.

Reason: ORGANELLES check (2026-11-03). PAINT (interpro/panther/PTHR10851/PTHR10851-paint.tsv) places GO:0005739 at PTN000091367 (taxon:2759), seeded by Arabidopsis AT5G49970 (TAIR IDA to mitochondrion, PMID:24599492) and yeast Pdx3/SGD:S000000239 (SGD IDA to GO:0005758 intermembrane space, PMID:22984289), both checked in QuickGO; the node placement is sound and human PNPO descends from it. For the human protein, the MitoCoP high-confidence proteome (PMID:34800366, the HTP row) and MitoCarta3.0 (Human.MitoCarta3.0.xls, Broad; evidence "APEX_matrix, yeast mito homolog++, coexpression, MS/MS", SubMitoLocalization "Matrix") both list PNPO, so the earlier "cytosolic contamination" reading is not supported. The Human Protein Atlas immunofluorescence (cytosol and nucleoplasm, as summarised in MitoCarta3.0) and Reactome place the bulk of the enzyme in the cytosol, so the mitochondrial pool is minor and non-core (DTYMK precedent), not over-annotated. is_active_in is not shown: no study measures PNPO activity inside human mitochondria, so the review supports presence (located_in) rather than activity there. Stay at the root: the sub-mitochondrial lines disagree (MitoCarta APEX matrix call versus the yeast seed in the intermembrane space), PNPO has no annotated transit peptide or TM segment, and MitoCarta3.0 gives a TargetP score of 5.

Summary: Correct core biological process. PNPO catalyzes the terminal step of PLP (active vitamin B6) formation. Consistent with the UniProt PLP-salvage pathway annotation and with the experimental IDA annotation of the same term.

Summary: Electronic (ARBA/InterPro/EC/RHEA) assignment of the correct core catalytic activity. Redundant with the experimental IDA/EXP annotations of the same term, which are the stronger evidence.

Summary: This InterPro2GO/keyword-derived term mislabels the process. PNPO does not synthesize pyridoxine (vitamin B6); it oxidizes the already-phosphorylated B6 vitamers PNP and PMP to the active cofactor PLP. The specific, correct process term for this reaction is pyridoxal 5'-phosphate biosynthetic process. Suggest replacing with GO:0042823.

Proposed replacements: pyridoxal 5'-phosphate biosynthetic process

Summary: Correct and core. PNPO is a flavoprotein that binds one FMN per subunit as its redox cofactor. Supported experimentally (IDA, PMID:12824491) and by the crystal structure. Redundant with the IDA annotation of the same term.

Summary: Correct but overly general parent-type oxidoreductase term (relevant to the pyridoxamine-phosphate half-reaction, which acts on the CH-NH2 group of PMP). The specific child term GO:0004733 (pyridoxamine phosphate oxidase activity) is already annotated with experimental evidence and fully captures the activity, so this general term is redundant/over-general.

Summary: Bare "protein binding" from a proteome-scale interactome screen (partner AGTRAP, Q6RW13-2). Non-informative for PNPO function - it identifies no specific molecular activity or biologically meaningful complex for this metabolic enzyme.

Reason: High-throughput binary interactome screens report no functional context for this pair, and bare protein binding carries no information about PNPO activity; removal is about informativeness and does not assert that the interaction is false.

Summary: Bare "protein binding" from a high-throughput interactome map (partner MTERF1, Q99551). Uninformative for the enzyme's function.

Reason: High-throughput binary interactome screens report no functional context for this pair, and bare protein binding carries no information about PNPO activity; removal is about informativeness and does not assert that the interaction is false.

Summary: Bare "protein binding" from a binary interactome reference map (partners AGTRAP Q6RW13-2 and LIME1 Q9H400). Uninformative.

Reason: High-throughput binary interactome screens report no functional context for this pair, and bare protein binding carries no information about PNPO activity; removal is about informativeness and does not assert that the interaction is false.

Summary: Bare "protein binding" from a proteome-scale interaction network (partner MTERF1, Q99551). Uninformative for PNPO function.

Reason: High-throughput binary interactome screens report no functional context for this pair, and bare protein binding carries no information about PNPO activity; removal is about informativeness and does not assert that the interaction is false.

Summary: Direct experimental support for the core catalytic activity. Human brain PNPO was expressed in E. coli and its pyridox(am)ine 5'-phosphate oxidase activity assayed, including via N-/C-terminal truncation mutants. Core function.

Summary: Mitochondrial location from a high-throughput proteome dataset. PNPO is a canonical soluble cytosolic enzyme (Reactome localizes both catalyzed reactions to the cytosol), and the cited high-throughput mitochondrial proteome study does not specifically establish PNPO as a mitochondrial resident. Likely reflects cytosolic contamination of the mitochondrial preparation. Retained (HTP experimental-derived) but flagged as an over-annotation. ORGANELLES re-check (2026-11-03) revises this to non-core.

Reason: ORGANELLES check (2026-11-03). MitoCoP is not a single crude fraction but a high-confidence set built from subtractive proteomics, spatial proteomics and importomics; the per-protein calls are in its supplementary tables, which are not in the cached text, so the curator's HTP row is not overruled. The earlier notes called this a mouse proteome; it is human. MitoCarta3.0 independently lists PNPO (APEX matrix labelling in human cells, MS/MS in two mouse tissues). With HPA and Reactome putting the bulk of the enzyme in the cytosol, this is a minor, non-core mitochondrial pool (as for DTYMK in batch 3). The assay cannot resolve the sub-compartment and the available sub-mitochondrial lines disagree, so no NEW row.

Summary: Direct experimental support for the core activity. Human PNPO variants expressed in CHO cells; the disease variant R229W markedly reduced pyridox(am)ine phosphate oxidase activity while splice/stop mutations were null, directly demonstrating the enzyme's catalytic function. Core function.

Summary: Direct experimental support for the core catalytic activity. Recombinant human PNPO was characterized kinetically (low turnover, low-micromolar KM for both PNP and PMP) and structurally. Core function.

Summary: Correct and core. The crystal structure and biochemistry establish that human PNPO tightly binds one FMN per subunit as its flavin cofactor.

Summary: Experimentally supported - each subunit binds one molecule of the product pyridoxal 5'-phosphate tightly (product inhibition). This is a genuine ligand-binding property but is a consequence of the catalytic mechanism (PLP is the reaction product), not an independent core function; kept as non-core supporting evidence.

Summary: Experimentally supported - the purified human enzyme is a homodimer, the catalytically active quaternary form. A real structural property but supportive rather than the enzyme's core catalytic function.

Summary: Correct core biological process, experimentally supported. PNPO catalyzes the terminal step in PLP synthesis.

Summary: Correct subcellular localization. PNPO is a soluble cytosolic enzyme; Reactome localizes the PNP-to-PLP oxidation reaction to the cytosol. Core location.

Summary: Correct subcellular localization (Reactome reaction for the PMP-to-PLP oxidation, cytosol). Consistent with the companion cytosol annotation.