The PAINT record for PTHR43173 has exactly one annotated node, PTN005148758, carrying exactly three terms, all seeded from SGD:S000004243 (yeast MCP2/Cqd2). The node's reach is 97 entities, identical for all three terms. No molecular-function term is asserted at the node, so PAINT did not propagate any kinase activity into this family.

Maps UniProt SUBCELLULAR LOCATION SL-0173 "Mitochondrion" to GO:0005739. For ADCK1 the underlying UniProt statement is experimental (ECO:0000269|PubMed:33988507), so the resulting IEA is grounded rather than predicted.

Human HeLa siRNA knockdown of ADCK1, assayed by transmission electron microscopy, increased mitochondrial length and decreased cristae number - the direct human basis for both IMP rows, despite the paper's Drosophila-focused title.

Co-expression of ADCK1 with OPA1 in HEK293T cells increased cleavage of long-form OPA1, quantified across four replicates - the biochemical step underlying the morphological phenotypes, and the basis of the proposed GO:0010954 annotation.

Kinase-dead substitutions at the phosphotransfer residues did not abolish the phenotypes caused by ADCK1 over-expression. The scope of this result is gain-of-function: it shows the over-expression phenotype does not require these residues, not that ADCK1 lacks activity.

The authors state the prior explicitly rather than claiming a negative: at the time of writing no report confirmed that ADCK1 is a kinase. No in vitro assay of purified ADCK1 has been published since.

Knockdown of dYME1L1 rescued the pupal lethality of dADCK1 over-expression, placing ADCK1 genetically upstream of YME1L1. The relationship is epistatic; no physical ADCK1-YME1L1 interaction is reported.

Localisation was determined by imaging epitope-tagged kinases ectopically expressed in HeLa cells. This is the assay behind the GO:0005739 IDA and behind UniProt's SUBCELLULAR LOCATION line; the over-expression caveat is recorded, and two independent measurements agree with the result.

A purpose-built high-confidence human mitochondrial proteome; ADCK1 is among the proteins for which the paper lists a validating antibody, so its inclusion is not a pure MS assignment. The reference carries 1235 GO annotations, but no functional or phenotype term spreads with the localisation term, so this is per-protein measurement rather than a projection.

Mouse ADCK1 was localised directly to the mitochondrial inner membrane in primary brain endothelial cells - independent mammalian corroboration of the compartment that human ADCK1 holds only by IBA.

ADCK1 co-immunoprecipitates with IMMT, the MICOS core subunit that forms crista junctions, in primary brain endothelial cells.

ADCK1 over-expression increases and Adck1 knockdown decreases cleavage of long-form OPA1 in mouse primary brain endothelial cells - a bidirectional replication, in a second organism and a second laboratory, of the OPA1-processing result first reported in HEK293T cells.

Defines the UbiB-specific structural features from the ADCK3/COQ8A crystal structure, including the alanine-rich loop that selects ADP over ATP. ADCK1's A164 occupies the same loop position as the ADCK3 A339 mutated here, which is why the ATP ligand assigned to ADCK1's binding sites by ProRule is questionable.

The single ADCK3 A339G substitution reverses coenzyme selectivity and permits autophosphorylation, establishing that the alanine-rich loop is a functional determinant rather than an incidental sequence feature.

COQ8 lacks canonical protein-kinase activity in trans but has ATPase activity and binds lipid coenzyme-Q intermediates. This is the experimental basis of the GO:0004672 NOT|enables and GO:0006468 NOT|involved_in IDA rows on both COQ8A and COQ8B. It is the main basis for doubting the serine/threonine-kinase keyword on ADCK1, though not on its own: an ancestrally reconstructed COQ8B was later shown to phosphorylate a protein substrate in trans (PMID:38425362), so the family-level claim that survives is the narrower one that no UbiB protein has been typed as a Ser/Thr kinase.

The single demonstration of protein-kinase activity anywhere in the UbiB family, and the source of COQ8B's non-negated GO:0004672 IDA: COQ8B phosphorylates COQ3 in trans. Two limits on how far it reaches - the enzyme is an ancestral reconstruction rather than the extant human protein, and the residue class was not determined, so it does not establish Ser/Thr specificity for any UbiB protein. It also sits alongside COQ8B's own NOT|enables GO:0004672 row from PMID:27499294, which the two annotations leave unreconciled.

The COQ8B used throughout the study is an ancestrally reconstructed protein, purified because ancestral COQ8A gave poor yields. The paper does not state that it reverted to extant human COQ8B for the phosphorylation experiment.

The same work excludes small-molecule kinase activity for COQ8B, so its phosphotransfer output is protein-directed rather than directed at the CoQ intermediates it binds.

Renames yeast Mcp2/Ylr253w as Cqd2 and shows that its loss shifts coenzyme Q distribution in the direction opposite to loss of Cqd1. Cqd2 is the sole protein donor behind all three of ADCK1's IBA rows, so this is the best-characterised activity of the source of those transfers.

The phenotype is one of distribution, not synthesis - total cellular CoQ is unchanged. This is why no GO:0006744 ubiquinone-biosynthesis annotation is proposed for ADCK1.

Cqd2 function requires intact canonical protein-kinase-like and UbiB-specific residues. This is the direct counterweight to the kinase-independence result reported for ADCK1 in animals, and the reason this review declines to call ADCK1 catalytically dead.

The source publication for all three of MCP2's own experimental annotations (GO:0005743 IDA, GO:0007005 IMP and IGI, GO:0055088 IGI) and therefore the ultimate evidence behind ADCK1's three IBA rows. The lipid-homeostasis result comes from high-copy suppression of an mdm10 deletion, which is why that row is kept as non-core rather than as core.

ADCK1 depletion in osteosarcoma cells lowers mitochondrial membrane potential and ATP and raises reactive oxygen species, consistent with the mitochondrial role established elsewhere. Adds a disease context but no new function, and produces no GO annotation.

Reports an endogenous co-immunoprecipitation of ADCK1 with the transcription factor TCF4 in colon cancer cells. The paper reports no experiment placing ADCK1 outside mitochondria, while three independent studies place it inside them and the protein has a mitochondrial transit peptide. Single laboratory, unreplicated; no GO annotation derives from it and none is proposed.

Machine-generated provider record, gates_passed True, five citations, all numeric PMIDs with no bioRxiv identifiers in PMID-shaped fields. Its contribution to this review was bibliographic: it surfaced PMID:40884816, PMID:36371387, PMID:33824271 and PMID:31175694, none of which is cited by any GOA row, and the first of those turned out to carry the only independent mammalian localisation and OPA1-processing data for ADCK1.

The entry asserts EC=2.7.-.-, the keywords Kinase, Serine/threonine-protein kinase, ATP-binding and Transferase, and ATP-ligand BINDING sites at 161..169 and 183 - all from PROSITE-ProRule:PRU00159 by ECO:0000255 - while its FUNCTION comment states that it is not known whether ADCK1 has protein kinase activity or what class of residue it would phosphorylate. The keyword-derived GO:0004674 and GO:0005524 cross-references are no longer present in GOA.

The three mutagenesis entries record the same gain-of-function result reported by Yoon et al.: substituting the phosphotransfer residues has no effect on the protein's role in maintaining mitochondrial structure and function.

SUBCELLULAR LOCATION is experimentally grounded but resolves only to the organelle, which is what GO_REF:0000044 then maps to GO:0005739.

ADCK1 retains all four core phosphotransfer residues (K183, D315, N320, D338), conserved in 8 of 8 UbiB proteins examined, so a pseudokinase reading based on lost catalytic residues is refused by measurement.

ADCK1 carries the two UbiB diagnostics: an alanine-rich phosphate-binding loop retaining 1 of 3 canonical glycines, with A164 in the same column as the COQ8A ADP/ATP-selectivity determinant A339, and no catalytic-loop arginine in 0 of 8 UbiB proteins against the PKA control.