COX10 farnesylates protoheme IX to produce heme O, the substrate for the next heme A biosynthetic step. This supports one catalytic core: protoheme IX farnesyltransferase activity at the mitochondrial inner membrane, directly involved in heme A biosynthesis. The physiological consequence is the supply of heme A required for respiratory complex IV. Heme synthesis, heme insertion and structural membership of complex IV are distinct claims.
The original human cDNA study tested respiratory complementation in a yeast cox10 mutant, with segregation, retransformation and subcloning controls. It identified a 443-residue protein homologous to yeast/bacterial heme farnesyltransferases; it did not measure purified human-enzyme kinetics (PMID:8078902(https://pubmed.ncbi.nlm.nih.gov/8078902/)). Direct bacterial CyoE membrane assays convert protoheme and farnesyl diphosphate to heme O, providing experimental grounding for the conserved reaction rather than a human IDA assay (PMID:8253713(https://pubmed.ncbi.nlm.nih.gov/8253713/)).
Two historical GO:0004311 assertions need reaction correction. That ID currently denotes geranylgeranyl diphosphate synthase activity, while its historical record label describes another reaction. Neither describes heme O synthesis. The original IDs and labels are retained as source history, with GO:0008495 proposed as the replacement. Current complete source objects and the two label discrepancies are inspectable in the evidence index.
In patient fibroblasts, human COX10 expression restores COX activity and raises heme A toward control levels. The original HPLC Results report about 107% and 81% of control in the two transduced patient lines, compared with about 48% and 20% before treatment. This supports the heme A pathway contribution. The source measures respiratory-subunit abundance and holoenzyme assembly, not COX10 membership in the respiratory enzyme (PMID:12928484(https://pubmed.ncbi.nlm.nih.gov/12928484/), original Results: Heme A biosynthesis and Table 2).
The separate patient-subassembly study supports retaining the existing experimental complex IV assembly context. Its native/2D immunoblots show altered early subassemblies in COX10-deficient fibroblasts. The authors infer the order of heme incorporation from those patterns; they do not directly assay COX10 as an insertase. The inherited reference findings should be read with that distinction: a proposed assembly model is not a demonstrated insertion reaction (PMID:14607829(https://pubmed.ncbi.nlm.nih.gov/14607829/)). The more recent mouse knockout/transduction abstract likewise supports cofactor-dependent assembly physiology, without establishing a second catalytic function (PMID:16782876(https://pubmed.ncbi.nlm.nih.gov/16782876/)).
The current Human Protein Atlas record identifies COX10 mitochondria as the main supported ICC/IF location, with additional nucleolar/cytosolic observations. This assessment uses the returned experimental summary; images were not independently interpreted. Reactome's actual catalyst entity resolves to Q12887 at the mitochondrial inner membrane. The reaction's matrix compartment also describes other participants and must not be transferred mechanically to the protein (R-HSA-2995330).
The current electronic matrix transfer resolves to mouse Q8CFY5 and PMID:16103131. The complete normal Methods/Results/Discussion text describes muscle-specific Cox10 deletion, mitochondrial enzyme assays and respiratory-subunit immunoblots. The authors explicitly report unsuccessful Cox10p-specific antibody generation and use Cox1p as a functional proxy. That source does not establish COX10 localization in the gel-like matrix lumen. The source-specific matrix annotation is therefore over-specific, without implying that inner-membrane COX10 cannot have matrix-facing regions (PMID:16103131(https://pubmed.ncbi.nlm.nih.gov/16103131/), Discussion and immunoblot Methods).
The MitoCoP article supplies the experimental mitochondrial proteome context and explicitly mentions COX10. Its final inventory combines several evidence pools, including a literature-supported pool. Selected original Methods and Results were read, but the exact COX10 supplementary Table S1 row was not recovered. No target abundance or individual pool assignment is inferred. The broad mitochondrial HTP assertion remains compatible with independent human localization and established function; incomplete supplement access does not invalidate the curator's evidence (PMID:34800366(https://pmc.ncbi.nlm.nih.gov/articles/PMC8664129/)).
The current PTHR43448 tree places the exact human reference sequence below PTN000114243 and PTN000114245. The ungapped human alignment sequence matches the official 443-residue sequence. Current mouse, yeast, fission-yeast and bacterial annotations provide distinct experimental and inferred grounding. Current topology and donor records do not reproduce every historical narrow-term IBD decision; a target listed among its own descendants is legitimate evidence, not circularity. The actual three ARBA rules match the target's official InterPro/FunFam and taxonomic features. These records and literal donor evidence are retained in the evidence index.
The two official products remain distinct. Q12887-1 matches the 443-residue reference. VSP_056867 replaces residues 1–15, and VSP_056868 deletes 16–207; applying both reconstructs the 251-residue Q12887-2 FASTA exactly. The shorter product lacks the canonical N-terminal 207 residues, including one predicted transmembrane helix, while retaining six downstream predicted helices. Sequence retention alone does not establish import, topology or equal catalysis for that product. Historical nine-helix predictions and the current seven predicted helices are different prediction sets, not experimentally reconciled topology.
Original author-deposited text was recovered for the 1994 complementation and 2003/2004 patient studies; their protected normal caches remain abstract-only. Public original URLs and precise indexed line ranges are in the evidence index. The 16103131 normal source contains full relevant text, while 16782876 and 8253713 remain abstract-only. No source-access flag is inferred solely from whether the normal cache or an external original supplied the evidence.
The active labels of the two GO:0004311 rows now use its current name, geranylgeranyl diphosphate synthase activity. Their complete historical source objects, including the earlier labels, remain verbatim in the evidence index under source_reconciliation.rows[17/23].historical_source; displayed_label_corrections records the two changes. This supersedes the earlier note about retaining those labels in the active rows. The source-specific replacements correct existing assertions to terms already present on COX10; they add no NEW annotation.
The mouse matrix donor is already preserved at donor_evidence.joins[12/13].records[0]: UniProtKB:Q8CFY5!447474109, MGI IDA/ECO:0000314, PMID:16103131, taxon10090. The evidence index now exposes the dated QuickGO query and exact record locators. The antibody-failure anchor in the review quotes the relevant result directly; matrix-facing topology is not established by the returned qualifier.
The HPLC figures above are from the externally read author-deposited original, Results “Heme A biosynthesis” and Table2 on printed page2696, rather than the abstract-only cache. Indexed lines682–690 and788–834 distinguish untreated and COX10-transduced patient fibroblasts and the heme A/B normalization.
GO:0070069 includes cytochrome-containing protein complexes beyond respiratory complexes. The source-specific membership assessment does not assert that narrower definition. The inherited MitoCoP finding is now explicitly qualified: its title excerpt does not establish an individual COX10 inner-membrane measurement, and the exact Table S1 pool remains unverified. The genuine Falcon report is cited only as secondary synthesis for independently corroborated reaction and location, not as an additional experiment.