ADCK1

Existing Annotations Review

GO Term	Evidence	Action	Reason
GO:0005743 mitochondrial inner membrane	IBA GO_REF:0000033	ACCEPT	Summary: Correct, core, and better supported than an IBA usually is. The single protein donor, SGD:S000004243, resolves to Q06567 / MCP2_YEAST (Swiss-Prot, 569 aa, "ABC1 family protein MCP2", also known as Ylr253w and renamed Cqd2), which sits in the same PANTHER subfamily PTHR43173:SF19 as human ADCK1 - an ortholog transfer, not a paralog transfer - and which carries its own IDA to this exact term from PMID:23781023. The propagation therefore lands at the donor's own term rather than above it. Independently, mouse ADCK1 was localised to the mitochondrial inner membrane in primary brain endothelial cells (PMID:40884816), so a mammalian measurement now corroborates what the tree inferred. The is_active_in qualifier is appropriate: ADCK1's demonstrated role, regulating YME1L1-dependent handling of OPA1 and IMMT, is exerted at this membrane whether or not it turns out to be catalytic. Reason: Ortholog-based transfer from a donor that holds the same term by IDA, at the same level of specificity, and now corroborated by direct mammalian localisation. This is the gene's core compartment. Propagation Review Root cause: NO FAILURE CORE Sources checked: SGD:S000004243 · Q06567 / MCP2_YEAST (Swiss-Prot), ABC1 family protein MCP2 = Ylr253w = Cqd2, 569 aa SUPPORTS TRANSFER Resolved through UniProt xref:sgd-S000004243, single unambiguous hit, Swiss-Prot reviewed. QuickGO shows this donor carries its own GO:0005743 IDA from PMID:23781023, so the IBA does not rest on inference alone. PANTHER places it in subfamily PTHR43173:SF19 together with human, mouse, chicken, Xenopus and Drosophila ADCK1, i.e. it is the budding-yeast ortholog rather than a paralog. Corroborating detail from this review's own alignment: ADCK1 and Cqd2 share the identical HCD catalytic loop, distinct from ADCK2/Cqd1 (HAD) and the COQ8 clade (QTD). PANTHER:PTN005148758 · PANTHER internal tree node in PTHR43173 SUPPORTS TRANSFER An ancestral tree node, not a protein. PTHR43173-paint.tsv shows this is the only annotated node in the family and that it carries exactly three terms, all seeded from MCP2 alone. Queried through QuickGO with withFrom=PANTHER:PTN005148758, its reach is 97 entities, identical for all three terms, spanning metazoa, fungi, plants, algae and ciliates. Human ADCK5 (Q3MIX3) belongs to the same family but to subfamily SF28 and receives none of the three, so the node behaves as SF19-scoped, which is the appropriate scope. Supporting Evidence: PMID:23781023 We identified two novel mitochondrial proteins (open reading frames YOR228c and YLR253w) that we named Mdm10 complementing protein (Mcp) 1 and Mcp2. PMID:40884816 We first confirmed that mouse ADCK1 was localized to the mitochondrial inner membrane in primary brain ECs
GO:0007005 mitochondrion organization	IBA GO_REF:0000033	ACCEPT	Summary: Accept. The donor MCP2/Cqd2 holds this term by IMP and IGI from PMID:23781023, so the transfer is from experiment rather than from further inference, and it lands at the donor's own level - the ACRV1-style defect of a propagation coming to rest several levels above its donor was checked for and is absent here. The term is also independently true of human ADCK1 on its own data: siRNA knockdown in HeLa cells lengthens mitochondria and reduces cristae number (PMID:31125351). It is not made redundant by the two IMP rows below, because GO keeps the regulation branch disjoint from the process branch: neither GO:0010637 nor GO:1903852 is a descendant of GO:0007005 (verified against QuickGO ancestors). A more specific child was considered and rejected - the donor set is a single protein holding precisely this term, so there is no more specific term the evidence points to. Reason: Experimentally grounded ortholog transfer that lands at the donor's own term, and separately true of human ADCK1 from human knockdown data. Not redundant with the regulation-branch IMP rows. Propagation Review Root cause: NO FAILURE CORE Sources checked: SGD:S000004243 · Q06567 / MCP2_YEAST (Swiss-Prot), ABC1 family protein MCP2 = Ylr253w = Cqd2, 569 aa SUPPORTS TRANSFER Carries GO:0007005 by both IMP and IGI from PMID:23781023 (the IGI is with SGD:S000000008, MDM10). Same term, not a parent of one, so the transfer neither over-generalises nor under-specifies. PANTHER:PTN005148758 · PANTHER internal tree node in PTHR43173 SUPPORTS TRANSFER Tree node rather than a protein. Same node and same 97-entity reach as the other two IBA rows on this gene. Supporting Evidence: PMID:31125351 As a result, the mitochondrial length was increased and the number of mitochondrial cristae was decreased by ADCK1 knockdown
GO:0055088 lipid homeostasis	IBA GO_REF:0000033	KEEP AS NON CORE	Summary: Keep, but not as a core function of the human gene. The transfer is structurally sound - MCP2 is the ortholog and holds the term by IGI from PMID:23781023 - but the donor evidence is the weakest of the three rows in kind: it comes from a high-copy suppressor screen in which over-expressing Mcp2 restored phospholipid and ergosterol homeostasis in cells lacking Mdm10, which reports what excess Mcp2 can rescue rather than what normal Mcp2 does. Newer work strengthens rather than undermines the lipid connection for the yeast protein: loss of Cqd2 shifts the cellular distribution of coenzyme Q, a prenylated lipid quinone, in the direction opposite to loss of Cqd1, without changing total cellular CoQ (PMID:34362905). But no lipid or coenzyme-Q measurement of any kind has been made on human or Drosophila ADCK1, and every readout that has been measured in animals concerns inner-membrane architecture through YME1L1. Retained as a defensible ortholog inference awaiting a mammalian test, not promoted to core. Reason: A sound ortholog transfer whose donor evidence is real but gain-of-function in kind, and which has never been tested in any animal. Biologically plausible given the yeast ortholog's coenzyme-Q distribution phenotype, so not an over-annotation; but it is not what human ADCK1 is known to do. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: SGD:S000004243 · Q06567 / MCP2_YEAST (Swiss-Prot), ABC1 family protein MCP2 = Ylr253w = Cqd2, 569 aa SUPPORTS TRANSFER Holds GO:0055088 by IGI from PMID:23781023, the genetic interactor being SGD:S000005754. The donor's evidence is genuine experimental work by a curator, so SOURCE_WEAK_OR_INFERRED would be false; what limits it is that the assay is high-copy suppression of an mdm10 deletion rather than a loss-of-function lipid phenotype. Subsequent loss-of-function work on the same protein, renamed Cqd2, did find a lipid (coenzyme Q) distribution phenotype. PANTHER:PTN005148758 · PANTHER internal tree node in PTHR43173 SUPPORTS TRANSFER Tree node rather than a protein. Same node and reach as the other two IBA rows; nothing in the node's composition argues against the transfer. Supporting Evidence: PMID:23781023 Lipid analysis demonstrates that elevated levels of Mcp1 and Mcp2 restore the alterations in mitochondrial phospholipid and ergosterol homeostasis in cells lacking Mdm10. PMID:34362905 Loss of Cqd1 skews cellular CoQ distribution away from mitochondria, resulting in markedly enhanced resistance to oxidative stress caused by exogenous polyunsaturated fatty acids, whereas loss of Cqd2 promotes the opposite effects.
GO:0005739 mitochondrion	IEA GO_REF:0000044	ACCEPT	Summary: Correct. This is the automatic mapping of UniProt's SUBCELLULAR LOCATION line, which is itself an experimental statement (ECO:0000269\|PubMed:33988507) rather than a prediction. The term is less specific than GO:0005743 above, but that is a faithful rendering of its source vocabulary term SL-0173 "Mitochondrion", and refining an IEA past its own controlled-vocabulary source would be wrong. The place to fix the specificity is upstream, in UniProt's SUBCELLULAR LOCATION line, now that mouse ADCK1 has been placed in the inner membrane directly; that is raised in suggested_questions rather than acted on here. Reason: Accurate automatic mapping of an experimentally grounded UniProt subcellular-location statement. General relative to GO:0005743, but faithful to its source term. Supporting Evidence: file:human/ADCK1/ADCK1-uniprot.txt CC -!- SUBCELLULAR LOCATION: Mitochondrion {ECO:0000269\|PubMed:33988507}.
GO:0005739 mitochondrion	HTP PMID:34800366 Quantitative high-confidence human mitochondrial proteome an...	ACCEPT	Summary: Accept. Morgenstern et al. built a quantitative high-confidence human mitochondrial proteome, and ADCK1 is among the proteins for which they list a validating antibody, so its inclusion is not purely an MS assignment. This reference carries 1235 annotations, which is the shape a complex-to-subunit projection also takes, so the discriminator was applied: every annotation sampled is the same localisation term and no functional or phenotype term spreads with it, so this is a proteome survey annotating each protein from its own measurement, not one experiment projected across a set. Note the result is paginated (100 of 1235 returned), so the distinct entity count is unavailable and is deliberately not substituted with the annotation total. The assigner is recorded as FlyBase, which is odd for a human annotation but is a provenance curiosity rather than a reason to doubt the call. Reason: High-throughput but per-protein localisation from a purpose-built high-confidence mitochondrial proteome, concordant with the IDA and IBA localisation rows. Projection test applied and negative. Supporting Evidence: PMID:34800366 Rabbit polyclonal anti-ADCK1Thermo Fisher ScientificCat.#PA5-28685; RRID: AB_2546161
GO:0005739 mitochondrion	IDA PMID:33988507 A subcellular map of the human kinome.	ACCEPT	Summary: Accept, with the assay's nature recorded. The Kinome Atlas imaged epitope-tagged kinases expressed in HeLa cells, so this is a tagged, ectopically expressed construct rather than the endogenous protein; that is standard and accepted practice for a localisation IDA, and it is not being used to support a functional claim. It is also the statement UniProt's SUBCELLULAR LOCATION line rests on. Two independent measurements agree with it - the native mitochondrial-proteome survey above, and inner-membrane localisation of the mouse protein in primary brain endothelial cells - so the over-expression caveat does not put the conclusion in doubt. This reference carries only 6 annotations over 4 entities, so there is no projection concern. Reason: Direct imaging localisation, concordant with two independent lines of evidence. The tagged over-expression caveat is recorded but does not alter the call for a localisation term. Supporting Evidence: PMID:33988507 456 epitope-tagged kinases, representing 85% of the human kinome, were expressed in HeLa cells and imaged by immunofluorescent microscopy
GO:0010637 negative regulation of mitochondrial fusion	IMP PMID:31125351 Drosophila ADCK1 is critical for maintaining mitochondrial s...	ACCEPT	Summary: Accept as core, and note that it rests on human data rather than only on the fly work the paper is titled for. Yoon et al. knocked ADCK1 down with siRNA in HeLa cells and measured by transmission electron microscopy that mitochondrial length increased, the signature of elevated fusion; over-expressing ADCK1 produced the converse, shortened mitochondria. Loss of ADCK1 therefore increases fusion, so ADCK1 negatively regulates it, and the sign of the annotation is right. The mechanism is consistent: ADCK1 acts upstream of YME1L1, whose processing of long-form OPA1 opposes fusion, and over-expressing ADCK1 with OPA1 increases L-OPA1 cleavage. Independently corroborated in mouse primary brain endothelial cells, where Adck1 knockdown alters mitochondrial length and the transcriptional programme that includes Adck1 was enriched for negative regulation of mitochondrial fusion (PMID:40884816). Reason: Directly supported by human loss-of-function imaging in the cited paper, with a coherent mechanism and independent mammalian corroboration. This is one of the gene's two core biological processes. Supporting Evidence: PMID:31125351 As a result, the mitochondrial length was increased and the number of mitochondrial cristae was decreased by ADCK1 knockdown PMID:31125351 The result demonstrated that the over-expression of both OPA1 and ADCK1 led to increased cleavage of L-OPA1
GO:1903852 positive regulation of cristae formation	IMP PMID:31125351 Drosophila ADCK1 is critical for maintaining mitochondrial s...	ACCEPT	Summary: Accept as core. The same human HeLa siRNA experiment counted cristae by transmission electron microscopy and found their number reduced on ADCK1 knockdown, which is what a positive regulator of cristae formation predicts. The direction is confirmed from the other side in a second organism and a second laboratory: in mouse primary brain endothelial cells, Adck1 shRNA disrupts cristae and re-expressing ADCK1 rescues the cristae defect of Foxq1-knockout cells, and ADCK1 co-precipitates with IMMT, the MICOS core subunit that builds crista junctions (PMID:40884816). UniProt's FUNCTION line summarises the same conclusion. One asymmetry the term does not capture, and which no GO term would: the phenotype is non-monotonic. Losing ADCK1 reduces cristae number, but raising it does not increase them - it destroys cristae outright and produces abnormal circular profiles. So the loss-of-function arm reads as positive regulation while the gain-of-function arm reads as disruption, which is the behaviour of a factor that has to be held within a range rather than of a simple activator. GO:1903852 is right for the arm the IMP rests on; the other arm is recorded as a knowledge gap rather than annotated, since no term expresses "required at a set level". Reason: Human loss-of-function ultrastructural measurement, with an independent mouse rescue experiment and a physical link to the MICOS subunit IMMT. The gene's second core biological process. Supporting Evidence: PMID:31125351 As a result, the mitochondrial length was increased and the number of mitochondrial cristae was decreased by ADCK1 knockdown PMID:40884816 Co‐IP assay detecting interaction between ADCK1 and IMMT in primary brain ECs
GO:0010954 positive regulation of protein processing	IMP PMID:31125351 Drosophila ADCK1 is critical for maintaining mitochondrial s...	NEW	Summary: Proposed new annotation. ADCK1's effect on mitochondrial shape is exerted through the proteolytic processing of OPA1 by YME1L1, and that step has itself been measured bidirectionally in two organisms and two laboratories, yet it is captured by no existing GO row. Yoon et al. showed in HEK293T cells that co-expressing ADCK1 with OPA1 increases cleavage of the long form; the FOXQ1 study showed in mouse primary brain endothelial cells that ADCK1 over-expression increases and Adck1 shRNA decreases L-OPA1 cleavage, and that ADCK1 co-precipitates with OPA1. GO:0010954 states what is measured - that ADCK1 raises the rate of a protein-processing event - without asserting that ADCK1 is itself the protease, which it is not; the has_input extension records OPA1 as the substrate machine-readably. IMP is the appropriate code because the evidence is knockdown and over-expression rather than a direct assay of ADCK1 activity. One limit stated plainly, because the two species carry different halves of the argument: the HUMAN evidence is over-expression only - ADCK1 and OPA1 co-expressed in HEK293T - while the loss-of-function arm that makes it directional is MOUSE Adck1 shRNA in primary brain endothelial cells. Neither species has both arms. A human ADCK1 knockdown scored for L-OPA1 cleavage would close that, and is listed under suggested_experiments; until then the row is well founded but rests on a cross-species pairing rather than on one complete human dataset. Reason: A reproducibly measured, directionally controlled effect of ADCK1 on OPA1 processing that the current annotation set does not capture at all. Additive to the two accepted IMP rows, which record the morphological outcomes rather than the biochemical step that produces them. Supporting Evidence: PMID:31125351 The result demonstrated that the over-expression of both OPA1 and ADCK1 led to increased cleavage of L-OPA1 PMID:40884816 Immunoblot images showing the levels of OPA1 in primary bECs infected with OPA1‐Flag and with either the empty vector, ADCK1‐HA, or Adck1‐shRNA lentivirus.

Core Functions

Regulator of mitochondrial inner-membrane architecture acting upstream of the i-AAA protease YME1L1. ADCK1 sits in the inner membrane and restrains mitochondrial fusion while promoting cristae formation: depleting it in human HeLa cells lengthens mitochondria and reduces cristae number, and raising it fragments mitochondria and destroys cristae. The two outputs share one route. Excess ADCK1 increases YME1L1-dependent cleavage of long-form OPA1, and lowers IMMT, the MICOS core subunit that builds crista junctions; knocking down YME1L1, or supplying extra OPA1 or IMMT, rescues the phenotypes that excess ADCK1 causes, placing all three downstream of it. Mouse ADCK1 reproduces this in primary brain endothelial cells, where it localises to the inner membrane, co-precipitates with IMMT and with OPA1, and is required for normal cristae. The relationship to YME1L1 is genetic and epistatic rather than shown to be a direct physical one, so ADCK1 is best described as acting on the protease's output rather than as a component of it.

Directly Involved In:

negative regulation of mitochondrial fusion positive regulation of cristae formation positive regulation of protein processing

Cellular Locations:

mitochondrial inner membrane

Supporting Evidence:

PMID:31125351
As a result, the mitochondrial length was increased and the number of mitochondrial cristae was decreased by ADCK1 knockdown
PMID:31125351
the flies over-expressing dADCK1 with simultaneous dYME1L1 knockdown successfully survived into adulthood
PMID:31125351
The result demonstrated that the over-expression of both OPA1 and ADCK1 led to increased cleavage of L-OPA1
PMID:40884816
We first confirmed that mouse ADCK1 was localized to the mitochondrial inner membrane in primary brain ECs
PMID:40884816
Co‐IP assay detecting interaction between ADCK1 and IMMT in primary brain ECs

Catalytically uncommitted UbiB-family active site. ADCK1 retains all four core phosphotransfer residues - the beta3 lysine K183, the catalytic aspartate D315, the catalytic-loop asparagine N320 and the Mg-binding aspartate D338 - conserved in 8 of 8 UbiB proteins examined (ADCK1, ADCK2, ADCK5, COQ8A, COQ8B, and yeast Cqd1, Cqd2 and Coq8). It also carries the two deviations that mark the family: an alanine-rich phosphate-binding loop, LGTASLAQV against PKA's LGTGSFGRV, retaining 1 of the 3 canonical glycines, and no arginine in the catalytic loop, which reads HCD - an arginine present in 0 of 8 UbiB proteins examined, against the PKA control which has it. ADCK1's A164 occupies the same loop position as COQ8A A339, whose mutation to glycine flips COQ8A from ADP- to ATP-selectivity and enables autophosphorylation. In the characterised relatives these features accompany the absence of canonical protein-kinase activity in trans alongside retained ATPase activity and binding of lipid coenzyme-Q intermediates - with the one exception that an ancestrally reconstructed COQ8B was later shown to phosphorylate COQ3 in trans, at residues that were not typed (PMID:38425362). No molecular function is asserted here: no purified ADCK1 has been assayed, so the site is untested rather than refuted, and this entry exists to record what the residues do and do not license.

Cellular Locations:

mitochondrial inner membrane

Supporting Evidence:

file:human/ADCK1/ADCK1-bioinformatics/RESULTS.md
All four core catalytic residues are canonical in **8/8** of the UbiB
PMID:25498144
We find that multiple UbiB-specific features are poised to inhibit protein kinase activity, including an N-terminal domain that occupies the typical substrate binding pocket and a unique A-rich loop that limits ATP binding by establishing an unusual selectivity for ADP.
PMID:27499294
Although COQ8 was predicted to be a protein kinase, we demonstrate that it lacks canonical protein kinase activity in trans. Instead, COQ8 has ATPase activity and interacts with lipid CoQ intermediates
PMID:31125351
Until now, ADCK1 was predicted to be a kinase, yet there were no available reports to confirm it.

References

Loading supporting content…

Download this section (compressed HTML)

Suggested Experiments

Experiment: Express and purify the ADCK1 kinase-like domain lacking the transit peptide, following the construct strategy that worked for COQ8A, and measure nucleotide binding by differential scanning fluorimetry or isothermal titration calorimetry against ATP and ADP side by side. Test the A164G substitution, which occupies the position at which the equivalent COQ8A A339G reverses coenzyme selectivity, as the discriminating mutant. Assay ATP hydrolysis, and test autophosphorylation in the wild-type and A164G proteins. A clear ADP preference reversed by A164G would place ADCK1 in the same mechanistic class as COQ8A and would show that the ATP ligand currently assigned to ADCK1's binding sites by ProRule is wrong.

Hypothesis: Purified ADCK1 is an ADP-preferring nucleotide-binding enzyme rather than a protein kinase, as its alanine-rich phosphate-binding loop predicts by analogy with COQ8A.

Type: in vitro biochemistry

Experiment: Generate ADCK1-knockout and ADCK1-over-expressing human cell lines, fractionate into pure mitochondrial and non-mitochondrial compartments, and quantify CoQ10 in each by LC-MS alongside total cellular CoQ10. The yeast result predicts a redistribution with no change in the total, so both numbers matter. Include an ADCK2 knockout as the counterpart to yeast Cqd1, since the reciprocal relationship between the two yeast proteins is the most distinctive feature of the phenotype, and test sensitivity to polyunsaturated-fatty-acid-induced lipid peroxidation as the functional readout.

Hypothesis: ADCK1 controls the partitioning of coenzyme Q between mitochondria and other membranes in human cells, as Cqd2 does in yeast.

Type: lipidomics and cell biology

Experiment: Test for a direct interaction between ADCK1 and YME1L1 in human cells at endogenous levels by proximity labelling from an ADCK1 bait and by co-immunoprecipitation from mitochondrial lysates, and check topology by protease protection to establish which face of the inner membrane ADCK1 presents its kinase-like domain to. All existing evidence for the ADCK1-YME1L1 relationship is epistatic, whereas ADCK1's contacts with IMMT and OPA1 have been shown by co-immunoprecipitation; a direct ADCK1-YME1L1 contact would change the mechanism from indirect modulation to protease regulation.

Hypothesis: ADCK1 acts on YME1L1 physically, not only genetically.

Type: interaction and topology mapping

Experiment: The published mutants were all scored for whether they still produce the phenotype of excess ADCK1. Invert the design: knock out ADCK1 in human cells, re-express wild-type or the A164G, K183I, D315A and D338N mutants at near-endogenous levels from a single-copy locus, and score rescue of cristae number, mitochondrial length, membrane potential and L-OPA1 cleavage. This is the assay that was applied to yeast Cqd2, where the residues did matter, and it is the missing control that would reconcile the two literatures.

Hypothesis: The catalytic residues of ADCK1 are required for its function when tested by loss-of-function rescue, even though they are dispensable for its over-expression phenotype.

Type: genetic complementation

Gene Description

Existing Annotations Review

Core Functions

References

Suggested Questions for Experts

Suggested Experiments

Deep Research

Affinage

📚 Additional Documentation

Notes

Bioinformatics Results

📄 View Raw YAML