APOO encodes MIC26 (also called MIC23), a 22 kDa subunit of the mitochondrial contact site and cristae organising system (MICOS), the inner-membrane protein complex that builds and maintains crista junctions - the narrow tubular necks connecting each crista to the inner boundary membrane. MIC26 is an integral inner-membrane protein that belongs, with its paralogue MIC27, with MIC10 and with MIC13, to the MIC10 subcomplex of MICOS; the other subcomplex is built around MIC60, which reaches across the intermembrane space to SAMM50 and the metaxins in the outer membrane so that MICOS forms the core of the ~2 MDa mitochondrial intermembrane space bridging (MIB) complex. MIC26 and MIC27 are confined to opisthokonts and are the most recently acquired MICOS subunits; they add onto an otherwise complete complex late in assembly, and neither is needed for the stability or incorporation of the other subunits. Loss of MIC26 is nonetheless not silent: crista junction number falls, the mitochondrial network fragments, and cellular cardiolipin drops, while loss of both MIC26 and MIC27 collapses cristae into concentric onion-like whorls and destabilises respiratory chain supercomplexes and the F1Fo-ATP synthase, a defect that restoring cardiolipin reverses. The two paralogues also set each other's abundance post-transcriptionally, so removing one raises the other. In humans, loss-of-function variants in the X-linked APOO gene cause mitochondrial disease through cristae disruption, presenting either as myopathy with lactic acidosis, cognitive impairment and autistic features or, for a C-terminal truncation, as a lethal progeria-like syndrome of infancy. Despite the gene name, which was given when a 55 kDa serum glycoprotein was attributed to this locus, MIC26 is not a circulating apolipoprotein: that species persists in APOO knockout cells and contains no MIC26 peptides, and the protein is found only in mitochondria.
Definition: The action of a molecule that contributes to the structural integrity of the mitochondrial contact site and cristae organizing system (MICOS) complex, the inner mitochondrial membrane complex that shapes and maintains crista junctions.
Justification: GO:0005198 structural molecule activity is currently the only molecular function assignable to MIC26, and it is uninformative: every structural subunit of every complex in the ontology receives the same term. GO already provides complex-specific children of this pattern - structural constituent of ribosome, of chromatin, of cytoskeleton, of proteasome, of synapse - and MICOS, with seven mammalian subunits of which at least two (MIC26 and MIC27) have no demonstrated activity beyond a structural contribution, is the same situation. A MICOS-specific child would let the MF slot carry the information that currently exists only in the CC slot.
Parent term: structural molecule activity
Supporting Evidence:
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0000139 Golgi membrane | IDA PMID:25764979 The non-glycosylated isoform of MIC26 is a constituent of th... | REMOVE | Summary: The Golgi-membrane pool is the second of the three forms proposed by Koob et al., and is the same 55 kDa glycosylated species as the secreted form, caught in transit. Reason: This row traces to the 55 kDa glycosylated species that was assigned to MIC26 before knockout controls existed. PMID:37279200, from the same laboratory that reported the three-form model, showed that species is not MIC26: it is unchanged by three independent siRNAs, unchanged by CRISPR knockout in HEK293, HepG2, HeLa and HAP1 cells, not generated by GFP- or myc-tagged MIC26, unaffected by mutagenesis of the predicted glycosylation sites, and devoid of MIC26 peptides when the 47-63 kDa gel region is excised and sequenced. With the 55 kDa band gone, nothing supports a secretory-pathway pool of MIC26. The glycosylation that motivated a Golgi itinerary is itself unsupported: mutating the predicted O-glycosylation serines of MIC26 does not change the 55 kDa band, and tagged MIC26 never produces a high-molecular-weight species. Supporting Evidence: PMID:37279200 In these KOs, we used four anti-MIC26 antibodies and consistently detected the loss of mitochondrial MIC26 (22 kDa) and MIC27 (30 kDa) but not the loss of intracellular or secreted 55 kDa protein. Thus, the protein assigned earlier as 55 kDa MIC26 is nonspecific. PMID:37279200 we conclude that both MIC26 and MIC27 are exclusively localized in mitochondria and that the observed phenotypes reported previously are exclusively due to their mitochondrial function PMID:37279200 we did not find MIC26-specific peptides in the MS analysis in any of the three samples analysed which included WT, MIC26 and MIC27 KOs PMID:37279200 Mutagenesis of predicted glycosylation sites in MIC26 also did not affect the detection of the 55 kDa protein band. |
| GO:0000139 Golgi membrane | IEA GO_REF:0000044 | REMOVE | Summary: Mechanical mapping of the UniProt subcellular-location term 'Golgi apparatus membrane' on the Q9BUR5 record. Reason: This row traces to the 55 kDa glycosylated species that was assigned to MIC26 before knockout controls existed. PMID:37279200, from the same laboratory that reported the three-form model, showed that species is not MIC26: it is unchanged by three independent siRNAs, unchanged by CRISPR knockout in HEK293, HepG2, HeLa and HAP1 cells, not generated by GFP- or myc-tagged MIC26, unaffected by mutagenesis of the predicted glycosylation sites, and devoid of MIC26 peptides when the 47-63 kDa gel region is excised and sequenced. With the 55 kDa band gone, nothing supports a secretory-pathway pool of MIC26. The electronic row inherits the defect of the UniProt statement it is generated from. Propagation Review Root cause: SOURCE BAD Failure modes: COMPARTMENT OR COMPLEX MISMATCH Sources checked: UniProtKB-SubCell:SL-0134 · Golgi apparatus membrane (UniProt subcellular location vocabulary) SOURCE BAD This IEA is generated mechanically from a UniProt SUBCELLULAR LOCATION line on Q9BUR5 that rests entirely on the 55 kDa immunoreactive species once assigned to MIC26. PMID:37279200 showed that species to be non-specific: it survives siRNA knockdown and CRISPR knockout in four human cell lines, is not reproduced by GFP- or myc-tagged MIC26, is unaffected by mutating the predicted glycosylation sites, and yields no MIC26 peptides by mass spectrometry. The UniProt line has not been retracted, so this row will regenerate until it is. Supporting Evidence: PMID:37279200 In these KOs, we used four anti-MIC26 antibodies and consistently detected the loss of mitochondrial MIC26 (22 kDa) and MIC27 (30 kDa) but not the loss of intracellular or secreted 55 kDa protein. Thus, the protein assigned earlier as 55 kDa MIC26 is nonspecific. PMID:37279200 we conclude that both MIC26 and MIC27 are exclusively localized in mitochondria and that the observed phenotypes reported previously are exclusively due to their mitochondrial function PMID:37279200 we did not find MIC26-specific peptides in the MS analysis in any of the three samples analysed which included WT, MIC26 and MIC27 KOs |
| GO:0001401 SAM complex | HDA PMID:26477565 Evolution and structural organization of the mitochondrial c... | REMOVE | Summary: MIC26 is assigned to the sorting and assembly machinery complex of the mitochondrial OUTER membrane on the basis of one complexome profile. Querying QuickGO by this reference and this term returns 12 entities - MIC10, MIC13, MIC19, MIC25, MIC26, MIC27, MIC60, SAMM50, MTX1, MTX2, MTX3 and HSPA9 - which is the whole MIB complex, not the SAM complex. Everything that co-migrated received the SAM term. Reason: A single projection over a co-migrating assembly, applied to a protein in the wrong membrane. GO:0001401 is defined as 'A large complex of the mitochondrial outer membrane'; MIC26 is a single-pass INNER-membrane protein whose only transmembrane helix is annotated at residues 108-128 and which interacts with the inner-membrane subunits MIC60, MIC27 and MIC10. The genuine relationship between MIC26 and SAMM50/MTX1/MTX2 is bridging across the intermembrane space, and that is exactly what GO:0140275 MIB complex says - a term this same reference already supplies for this gene. Nothing is lost by removing the SAM row and a false compartment claim is. Supporting Evidence: PMID:25781180 In human mitochondria, the inner membrane MICOS complex interacts with the outer membrane sorting and assembly machinery (SAM) complex, to form the mitochondrial intermembrane space bridging complex (MIB). PMID:25764979 The latter isoform spans the mitochondrial inner membrane and physically interacts with several MICOS complex subunits such as MIC60, MIC27, and MIC10. file:human/APOO/APOO-uniprot.txt FT TRANSMEM 108..128 |
| GO:0005515 protein binding | IPI PMID:25764979 The non-glycosylated isoform of MIC26 is a constituent of th... | MODIFY | Summary: Co-immunoprecipitation of MIC26 with IMMT/MIC60 (Q16891, 758 aa), the central scaffold subunit of MICOS and the subunit that connects the complex through MIC19 to SAMM50 in the outer membrane. The partner was resolved against UniProt; it is a MICOS subunit, so this is an intra-complex interaction. Reason: Bare 'protein binding' says nothing about what MIC26 does. What these co-immunoprecipitations establish is complex membership - already captured, and better, by the GO:0061617 MICOS complex rows from this same reference - plus the thing membership implies about this subunit's activity: MIC26 contributes to the structural integrity of MICOS. That is GO:0005198, whose definition ('the action of a molecule that contributes to the structural integrity of a complex') is exactly the claim the genetics supports, since removing MIC26 measurably costs crista junctions without abolishing the complex. Two notes for whoever acts on this. First, all three GO:0005515 rows on this gene collapse onto the same replacement term, so adopting them yields one MF annotation rather than three; the partner identities are not lost, because each stays in its row's supporting_entities and the interactions themselves are recorded as complex membership by the GO:0061617 rows. Second, the evidence code has to change with the term: IPI belongs to the binding itself, whereas structural molecule activity inferred from co-immunoprecipitation with the rest of the complex is IC - or IDA if taken from the knockout ultrastructure - and should not be carried across as IPI. Proposed replacements: structural molecule activity Supporting Evidence: PMID:25764979 The latter isoform spans the mitochondrial inner membrane and physically interacts with several MICOS complex subunits such as MIC60, MIC27, and MIC10. PMID:32788226 SKOs and DKO cells have significant reduction in CJs per mitochondrial section. |
| GO:0005515 protein binding | IPI PMID:25764979 The non-glycosylated isoform of MIC26 is a constituent of th... | MODIFY | Summary: Co-immunoprecipitation of MIC26 with MICOS10/MIC10 (Q5TGZ0, 78 aa), the membrane-bending core of the MIC10 subcomplex to which MIC26 belongs. The partner was resolved against UniProt; it is a MICOS subunit, so this is an intra-complex interaction. Reason: Bare 'protein binding' says nothing about what MIC26 does. What these co-immunoprecipitations establish is complex membership - already captured, and better, by the GO:0061617 MICOS complex rows from this same reference - plus the thing membership implies about this subunit's activity: MIC26 contributes to the structural integrity of MICOS. That is GO:0005198, whose definition ('the action of a molecule that contributes to the structural integrity of a complex') is exactly the claim the genetics supports, since removing MIC26 measurably costs crista junctions without abolishing the complex. Two notes for whoever acts on this. First, all three GO:0005515 rows on this gene collapse onto the same replacement term, so adopting them yields one MF annotation rather than three; the partner identities are not lost, because each stays in its row's supporting_entities and the interactions themselves are recorded as complex membership by the GO:0061617 rows. Second, the evidence code has to change with the term: IPI belongs to the binding itself, whereas structural molecule activity inferred from co-immunoprecipitation with the rest of the complex is IC - or IDA if taken from the knockout ultrastructure - and should not be carried across as IPI. Proposed replacements: structural molecule activity Supporting Evidence: PMID:25764979 The latter isoform spans the mitochondrial inner membrane and physically interacts with several MICOS complex subunits such as MIC60, MIC27, and MIC10. PMID:32788226 SKOs and DKO cells have significant reduction in CJs per mitochondrial section. |
| GO:0005515 protein binding | IPI PMID:25764979 The non-glycosylated isoform of MIC26 is a constituent of th... | MODIFY | Summary: Co-immunoprecipitation of MIC26 with APOOL/MIC27 (Q6UXV4, 268 aa), MIC26's paralogue and partner in the same subcomplex, with which it shares the MIC10-subcomplex role and reciprocal post-transcriptional regulation of protein levels. The partner was resolved against UniProt; it is a MICOS subunit, so this is an intra-complex interaction. Reason: Bare 'protein binding' says nothing about what MIC26 does. What these co-immunoprecipitations establish is complex membership - already captured, and better, by the GO:0061617 MICOS complex rows from this same reference - plus the thing membership implies about this subunit's activity: MIC26 contributes to the structural integrity of MICOS. That is GO:0005198, whose definition ('the action of a molecule that contributes to the structural integrity of a complex') is exactly the claim the genetics supports, since removing MIC26 measurably costs crista junctions without abolishing the complex. Two notes for whoever acts on this. First, all three GO:0005515 rows on this gene collapse onto the same replacement term, so adopting them yields one MF annotation rather than three; the partner identities are not lost, because each stays in its row's supporting_entities and the interactions themselves are recorded as complex membership by the GO:0061617 rows. Second, the evidence code has to change with the term: IPI belongs to the binding itself, whereas structural molecule activity inferred from co-immunoprecipitation with the rest of the complex is IC - or IDA if taken from the knockout ultrastructure - and should not be carried across as IPI. One further thing this interaction does NOT license: in Saccharomyces cerevisiae, Mic26 destabilises Mic10 oligomers and so acts antagonistically to Mic27 (PMID:29733859). That is a yeast result. What has been shown for the human pair is reciprocal control of steady-state protein levels, post-transcriptionally, which is a different claim, so no regulatory molecular function is proposed from it here. Proposed replacements: structural molecule activity Supporting Evidence: PMID:25764979 The latter isoform spans the mitochondrial inner membrane and physically interacts with several MICOS complex subunits such as MIC60, MIC27, and MIC10. PMID:32788226 SKOs and DKO cells have significant reduction in CJs per mitochondrial section. |
| GO:0005576 extracellular region | EXP PMID:16956892 ApoO, a novel apolipoprotein, is an original glycoprotein up... | REMOVE | Summary: The original 2006 characterisation of 'apolipoprotein O' as a secreted chondroitin sulfate proteoglycan recovered from HDL, LDL and VLDL, and promoting cholesterol efflux from macrophages. Every one of those observations was read from a 55 kDa immunoreactive band, against a recombinant protein of 22 kDa. Reason: This row traces to the 55 kDa glycosylated species that was assigned to MIC26 before knockout controls existed. PMID:37279200, from the same laboratory that reported the three-form model, showed that species is not MIC26: it is unchanged by three independent siRNAs, unchanged by CRISPR knockout in HEK293, HepG2, HeLa and HAP1 cells, not generated by GFP- or myc-tagged MIC26, unaffected by mutagenesis of the predicted glycosylation sites, and devoid of MIC26 peptides when the 47-63 kDa gel region is excised and sequenced. With the 55 kDa band gone, nothing supports a secretory-pathway pool of MIC26. The independent checks all point the same way: MIC26 is absent from HDL proteomes in which the major apolipoproteins are detected, and this review's own sequence analysis finds that the N-terminal segment UniProt calls a SIGNAL peptide (ECO:0000255, prediction only) lacks the hydrophobic core that every member of a ten-protein secretory control panel has, while matching mitochondrial presequences in charge and amphipathicity - the paralogue MIC27's equivalent N-terminus is annotated by UniProt as a TRANSIT peptide. The 2006 paper remains a valid record of a 55 kDa serum proteoglycan; what has failed is its assignment to this gene. The machine-generated deep research record is listed as an additional reference because it is what first pointed to PMID:37279200; nothing here is supported by a sentence from it, and every claim is quoted from the primary papers or from UniProt. Supporting Evidence: PMID:37279200 In these KOs, we used four anti-MIC26 antibodies and consistently detected the loss of mitochondrial MIC26 (22 kDa) and MIC27 (30 kDa) but not the loss of intracellular or secreted 55 kDa protein. Thus, the protein assigned earlier as 55 kDa MIC26 is nonspecific. PMID:37279200 we conclude that both MIC26 and MIC27 are exclusively localized in mitochondria and that the observed phenotypes reported previously are exclusively due to their mitochondrial function PMID:37279200 we did not find MIC26-specific peptides in the MS analysis in any of the three samples analysed which included WT, MIC26 and MIC27 KOs PMID:16956892 We detected this new protein in the following lipoproteins: high density lipoprotein, low density lipoprotein, and very low density lipoprotein. PMID:37279200 MIC26 was not found to be a part of high-density lipoprotein (HDL)-associated proteins in a proteomic study where other major apolipoproteins could be detected file:human/APOO/APOO-bioinformatics/RESULTS.md On 4 of the 5 metrics the MIC26 1–25 segment sits closer to the mitochondrial-presequence panel than to the secretory-signal-peptide panel. |
| GO:0005576 extracellular region | IDA PMID:25764979 The non-glycosylated isoform of MIC26 is a constituent of th... | REMOVE | Summary: The secreted 55 kDa form, the first of the three forms proposed by Koob et al. Its levels are insensitive to MIC26 depletion, a fact visible in the companion data paper of the same year and unremarked at the time. Reason: This row traces to the 55 kDa glycosylated species that was assigned to MIC26 before knockout controls existed. PMID:37279200, from the same laboratory that reported the three-form model, showed that species is not MIC26: it is unchanged by three independent siRNAs, unchanged by CRISPR knockout in HEK293, HepG2, HeLa and HAP1 cells, not generated by GFP- or myc-tagged MIC26, unaffected by mutagenesis of the predicted glycosylation sites, and devoid of MIC26 peptides when the 47-63 kDa gel region is excised and sequenced. With the 55 kDa band gone, nothing supports a secretory-pathway pool of MIC26. The diagnostic was available in 2015: an antibody signal that does not fall when the gene is knocked down is not that gene's product. Supporting Evidence: PMID:37279200 In these KOs, we used four anti-MIC26 antibodies and consistently detected the loss of mitochondrial MIC26 (22 kDa) and MIC27 (30 kDa) but not the loss of intracellular or secreted 55 kDa protein. Thus, the protein assigned earlier as 55 kDa MIC26 is nonspecific. PMID:37279200 we conclude that both MIC26 and MIC27 are exclusively localized in mitochondria and that the observed phenotypes reported previously are exclusively due to their mitochondrial function PMID:37279200 we did not find MIC26-specific peptides in the MS analysis in any of the three samples analysed which included WT, MIC26 and MIC27 KOs PMID:26217777 Our data further demonstrate that depletion of MIC26 primarily affects the level of the 22 kDa mitochondrial isoform of MIC26 but not the amount of the secreted 55 kDa isoform of MIC26. PMID:37279200 Overall, only the mitochondrial MIC2622kDa protein was responsive while the MIC2655kDa was resistant to MIC26 downregulation leading us to question the specificity of the MIC2655kDa glycosylated protein. |
| GO:0005576 extracellular region | IEA GO_REF:0000044 | REMOVE | Summary: Mechanical mapping of the UniProt subcellular-location term 'Secreted' on the Q9BUR5 record, which cites PMID:16956892 and PMID:25764979. Reason: This row traces to the 55 kDa glycosylated species that was assigned to MIC26 before knockout controls existed. PMID:37279200, from the same laboratory that reported the three-form model, showed that species is not MIC26: it is unchanged by three independent siRNAs, unchanged by CRISPR knockout in HEK293, HepG2, HeLa and HAP1 cells, not generated by GFP- or myc-tagged MIC26, unaffected by mutagenesis of the predicted glycosylation sites, and devoid of MIC26 peptides when the 47-63 kDa gel region is excised and sequenced. With the 55 kDa band gone, nothing supports a secretory-pathway pool of MIC26. The electronic row inherits the defect of the UniProt statement it is generated from. Propagation Review Root cause: SOURCE BAD Failure modes: COMPARTMENT OR COMPLEX MISMATCH Sources checked: UniProtKB-SubCell:SL-0243 · Secreted (UniProt subcellular location vocabulary) SOURCE BAD This IEA is generated mechanically from a UniProt SUBCELLULAR LOCATION line on Q9BUR5 that rests entirely on the 55 kDa immunoreactive species once assigned to MIC26. PMID:37279200 showed that species to be non-specific: it survives siRNA knockdown and CRISPR knockout in four human cell lines, is not reproduced by GFP- or myc-tagged MIC26, is unaffected by mutating the predicted glycosylation sites, and yields no MIC26 peptides by mass spectrometry. The UniProt line has not been retracted, so this row will regenerate until it is. Supporting Evidence: PMID:37279200 In these KOs, we used four anti-MIC26 antibodies and consistently detected the loss of mitochondrial MIC26 (22 kDa) and MIC27 (30 kDa) but not the loss of intracellular or secreted 55 kDa protein. Thus, the protein assigned earlier as 55 kDa MIC26 is nonspecific. PMID:37279200 we conclude that both MIC26 and MIC27 are exclusively localized in mitochondria and that the observed phenotypes reported previously are exclusively due to their mitochondrial function PMID:37279200 we did not find MIC26-specific peptides in the MS analysis in any of the three samples analysed which included WT, MIC26 and MIC27 KOs |
| GO:0005576 extracellular region | ISS PMID:22261194 Proteomics analysis of cardiac extracellular matrix remodeli... | REMOVE | Summary: Sequence-similarity transfer from pig APOO (F1SPZ9). The donor's own GO:0005576 is an HDA from this same reference: a single detection of pig APOO in a sequentially extracted extracellular-matrix fraction of ischaemic/reperfused myocardium. Reason: The donor evidence is one high-throughput identification in a tissue-fraction preparation, from an organ whose cells are packed with mitochondria, and abundant mitochondrial proteins are a routine contaminant of such fractions. There is no orthogonal support for a secreted pig APOO - F1SPZ9's only other annotations are electronic, and they place it in the mitochondrial inner membrane. On the acceptor side, the human protein is now shown by knockout, tagging and mass spectrometry to be exclusively mitochondrial, so the transfer imports a compartment the target does not occupy. Propagation Review Root cause: SOURCE WEAK OR INFERRED Failure modes: COMPARTMENT OR COMPLEX MISMATCH Sources checked: UniProtKB:F1SPZ9 · APOO (Sus scrofa) SOURCE WEAK OR INFERRED Verified in UniProt as an unreviewed (TrEMBL) 166-residue pig MICOS complex subunit, shorter than the 198-residue human protein. QuickGO shows its GO:0005576 is an HDA from this same PMID:22261194 ECM proteomics run, with no independent support; its four remaining annotations are all electronic and place it in mitochondria and the mitochondrial inner membrane. A lone high-throughput detection in an enriched tissue fraction is too weak a basis for a cross-species compartment transfer, and the acceptor's own knockout data now contradict it. Supporting Evidence: PMID:37279200 In these KOs, we used four anti-MIC26 antibodies and consistently detected the loss of mitochondrial MIC26 (22 kDa) and MIC27 (30 kDa) but not the loss of intracellular or secreted 55 kDa protein. Thus, the protein assigned earlier as 55 kDa MIC26 is nonspecific. PMID:37279200 we conclude that both MIC26 and MIC27 are exclusively localized in mitochondria and that the observed phenotypes reported previously are exclusively due to their mitochondrial function PMID:37279200 we did not find MIC26-specific peptides in the MS analysis in any of the three samples analysed which included WT, MIC26 and MIC27 KOs PMID:22261194 ECM proteins were sequentially extracted and identified by liquid chromatography tandem mass spectrometry. PMID:22261194 We have applied a novel proteomic method for the analysis of ECM in cardiovascular tissues to a porcine model of ischemia/reperfusion injury. |
| GO:0005739 mitochondrion | HDA PMID:20833797 Phosphoproteome analysis of functional mitochondria isolated... | ACCEPT | Summary: MIC26 was identified in a phosphoproteomic survey of functional mitochondria purified from resting human skeletal muscle. Reason: Detection in purified human mitochondria, consistent with every other line of evidence. Less informative than the GO:0005743 rows on the same gene, but the assay resolves the organelle rather than the submitochondrial compartment, so the term matches the evidence. Supporting Evidence: PMID:20833797 We performed a phosphoproteomics study of functional mitochondria isolated from human muscle biopsies with the aim to obtain a comprehensive overview of mitochondrial phosphoproteins. |
| GO:0005739 mitochondrion | HTP PMID:34800366 Quantitative high-confidence human mitochondrial proteome an... | ACCEPT | Summary: MIC26 is a member of MitoCoP, the high-confidence human mitochondrial proteome defined by quantitative proteomics across mitochondrial preparations. Reason: MitoCoP is explicitly built to separate genuine mitochondrial proteins from co-purifying contaminants, so inclusion is stronger than a bare detection. Less informative than the GO:0005743 rows on the same gene, but the assay resolves the organelle rather than the submitochondrial compartment, so the term matches the evidence. Supporting Evidence: PMID:34800366 We classified >8,000 proteins in mitochondrial preparations of human cells and defined a mitochondrial high-confidence proteome of >1,100 proteins (MitoCoP). |
| GO:0005739 mitochondrion | IDA GO_REF:0000052 | ACCEPT | Summary: Human Protein Atlas immunofluorescence assigns APOO to the mitochondrion. Reason: Independent imaging evidence in intact cells, agreeing with biochemical fractionation and with purified-mitochondria proteomics. Less informative than the GO:0005743 rows on the same gene, but the assay resolves the organelle rather than the submitochondrial compartment, so the term matches the evidence. Supporting Evidence: PMID:24743151 In vivo fluorescent labeling and subcellular fractionation revealed that APOO localizes with mitochondria. |
| GO:0005739 mitochondrion | IDA PMID:24743151 Apolipoprotein O is mitochondrial and promotes lipotoxicity ... | ACCEPT | Summary: Turkieh et al. localised APOO to mitochondria by in vivo fluorescent labelling and subcellular fractionation, in the study that first moved the protein from the secretory pathway into the organelle. Reason: Two orthogonal methods in the same study, and the historically decisive observation for this gene. Less informative than the GO:0005743 rows on the same gene, but the assay resolves the organelle rather than the submitochondrial compartment, so the term matches the evidence. Supporting Evidence: PMID:24743151 In vivo fluorescent labeling and subcellular fractionation revealed that APOO localizes with mitochondria. PMID:24743151 myocardial sections from APOO-Tg mice revealed mitochondrial degenerative changes |
| GO:0005739 mitochondrion | IDA PMID:25781180 Detailed analysis of the human mitochondrial contact site co... | ACCEPT | Summary: FLAG-tagged Mic23/ApoO expressed in HeLa cells colocalised with MitoTracker, and the authors concluded that it is a mitochondrial protein. Reason: The imaging here uses an ectopically expressed tagged construct, which on its own would be weak; it is accepted because the same paper independently imports radiolabelled ApoO into isolated mitochondria and recovers it in the MIB complex, and because four other references reach the same localisation. Less informative than the GO:0005743 rows on the same gene, but the assay resolves the organelle rather than the submitochondrial compartment, so the term matches the evidence. Supporting Evidence: PMID:25781180 We observed that all the proteins, including Mic23/ApoO-FLAG (Fig. 1F), localized to mitochondria upon expression PMID:25781180 We conclude that Mic23/ApoO is a mitochondrial protein |
| GO:0005739 mitochondrion | IEA GO_REF:0000120 | ACCEPT | Summary: Combined automatic annotation: an ARBA rule keyed on the MIC26/MIC27 InterPro domain in eukaryotes, an Ensembl Compara projection from the 1:1 mouse orthologue, and the UniProt subcellular-location vocabulary. All three routes converge on the same, correct compartment. Reason: Every component of this composite inference checks out, and the conclusion is confirmed by five experimental rows on the same gene. Less informative than the GO:0005743 rows on the same gene, but the assay resolves the organelle rather than the submitochondrial compartment, so the term matches the evidence. Propagation Review Root cause: NO FAILURE CORE Sources checked: ARBA:ARBA00026962 · ARBA rule asserting GO:0005739 SUPPORTS TRANSFER Fetched from rest.uniprot.org/arba/ARBA00026962: 380 condition sets, a single asserted annotation (GO:0005739). One of those sets is exactly 'InterPro id IPR019166' plus 'taxon Eukaryota', both of which APOO satisfies, so the rule fires for the right reason rather than by a coincidental match. UniProtKB:Q9DCZ4 · Mic26 (Mus musculus) SUPPORTS TRANSFER The 1:1 mouse orthologue, 198 aa like the human protein and reviewed in Swiss-Prot as MIC26_MOUSE. Its UniProt record lists only 'Mitochondrion inner membrane' as subcellular location - the mouse entry has already been corrected in the way the human entry has not. ensembl:ENSMUSP00000109531 · Ensembl mouse Apoo translation SUPPORTS TRANSFER Verified against the Ensembl REST lookup as a 198-residue Mus musculus translation, cross-referenced by UniProt to Q9DCZ4. The same orthologue as the preceding donor, reached by the Ensembl Compara route rather than the Swiss-Prot one: two identifiers, one donor. UniProtKB-SubCell:SL-0173 · Mitochondrion (UniProt subcellular location vocabulary) SUPPORTS TRANSFER The organelle-level location term on the Q9BUR5 record, itself derived from the experimentally supported mitochondrial inner-membrane statement. Supporting Evidence: PMID:34800366 We classified >8,000 proteins in mitochondrial preparations of human cells and defined a mitochondrial high-confidence proteome of >1,100 proteins (MitoCoP). |
| GO:0005743 mitochondrial inner membrane | IDA PMID:25764979 The non-glycosylated isoform of MIC26 is a constituent of th... | ACCEPT | Summary: The 22 kDa form of MIC26 spans the mitochondrial inner membrane, consistent with its single predicted transmembrane helix at residues 108-128 and with its interactions with the inner-membrane MICOS subunits MIC60, MIC27 and MIC10. Reason: The most informative and best-supported location term on this gene, and the one subcellular claim from this reference that later knockout work confirmed rather than withdrew. Supporting Evidence: PMID:25764979 The latter isoform spans the mitochondrial inner membrane and physically interacts with several MICOS complex subunits such as MIC60, MIC27, and MIC10. file:human/APOO/APOO-uniprot.txt FT TRANSMEM 108..128 |
| GO:0005743 mitochondrial inner membrane | IEA GO_REF:0000044 | ACCEPT | Summary: Mapped mechanically from the UniProt subcellular-location vocabulary term 'Mitochondrion inner membrane' on the Q9BUR5 record. Reason: Unlike the three other GO_REF:0000044 rows on this gene, this one derives from the half of the UniProt location record that is experimentally supported and has not been withdrawn. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB-SubCell:SL-0168 · Mitochondrion inner membrane (UniProt subcellular location vocabulary) SUPPORTS TRANSFER Derived from the Q9BUR5 SUBCELLULAR LOCATION line 'Mitochondrion inner membrane {ECO:0000269|PubMed:25764979}', the experimentally supported and still-current half of that record. Supporting Evidence: PMID:25764979 The latter isoform spans the mitochondrial inner membrane and physically interacts with several MICOS complex subunits such as MIC60, MIC27, and MIC10. |
| GO:0005743 mitochondrial inner membrane | NAS PMID:25997101 QIL1 is a novel mitochondrial protein required for MICOS com... | ACCEPT | Summary: ComplexPortal's statement that the MICOS complex resides in the inner membrane. Reason: Correct and consistent with the IDA row; a complex-level statement that transfers safely to an integral subunit. Supporting Evidence: PMID:25781180 In human mitochondria, the inner membrane MICOS complex interacts with the outer membrane sorting and assembly machinery (SAM) complex, to form the mitochondrial intermembrane space bridging complex (MIB). |
| GO:0005789 endoplasmic reticulum membrane | IDA PMID:25764979 The non-glycosylated isoform of MIC26 is a constituent of th... | REMOVE | Summary: Koob et al. assigned an ER/Golgi-resident pool to MIC26 as part of a three-form model (secreted 55 kDa glycoprotein, ER/Golgi-resident form of the same species, and the 22 kDa mitochondrial protein). The ER/Golgi form is the intracellular precursor of the 55 kDa band, so it stands or falls with it. Reason: This row traces to the 55 kDa glycosylated species that was assigned to MIC26 before knockout controls existed. PMID:37279200, from the same laboratory that reported the three-form model, showed that species is not MIC26: it is unchanged by three independent siRNAs, unchanged by CRISPR knockout in HEK293, HepG2, HeLa and HAP1 cells, not generated by GFP- or myc-tagged MIC26, unaffected by mutagenesis of the predicted glycosylation sites, and devoid of MIC26 peptides when the 47-63 kDa gel region is excised and sequenced. With the 55 kDa band gone, nothing supports a secretory-pathway pool of MIC26. Removing this is not overruling a curator from an abstract: the authors of the annotated paper are co-authors of the refutation and state the corrected conclusion themselves. The mitochondrial claims of PMID:25764979 are accepted in full elsewhere in this review; only the forms that rest on the 55 kDa band are removed. Supporting Evidence: PMID:37279200 In these KOs, we used four anti-MIC26 antibodies and consistently detected the loss of mitochondrial MIC26 (22 kDa) and MIC27 (30 kDa) but not the loss of intracellular or secreted 55 kDa protein. Thus, the protein assigned earlier as 55 kDa MIC26 is nonspecific. PMID:37279200 we conclude that both MIC26 and MIC27 are exclusively localized in mitochondria and that the observed phenotypes reported previously are exclusively due to their mitochondrial function PMID:37279200 we did not find MIC26-specific peptides in the MS analysis in any of the three samples analysed which included WT, MIC26 and MIC27 KOs PMID:25764979 We demonstrate that human MIC26 exists in three distinct forms: (1) a glycosylated and secreted 55kDa protein, (2) an ER/Golgi-resident form thereof, and (3) a non-glycosylated 22kDa mitochondrial protein. |
| GO:0005789 endoplasmic reticulum membrane | IEA GO_REF:0000044 | REMOVE | Summary: Mechanical mapping of the UniProt subcellular-location term 'Endoplasmic reticulum membrane' on the Q9BUR5 record, which cites PMID:25764979. Reason: This row traces to the 55 kDa glycosylated species that was assigned to MIC26 before knockout controls existed. PMID:37279200, from the same laboratory that reported the three-form model, showed that species is not MIC26: it is unchanged by three independent siRNAs, unchanged by CRISPR knockout in HEK293, HepG2, HeLa and HAP1 cells, not generated by GFP- or myc-tagged MIC26, unaffected by mutagenesis of the predicted glycosylation sites, and devoid of MIC26 peptides when the 47-63 kDa gel region is excised and sequenced. With the 55 kDa band gone, nothing supports a secretory-pathway pool of MIC26. The electronic row inherits the defect of the UniProt statement it is generated from. Note that the UniProt entry for the mouse orthologue Q9DCZ4 already lists only 'Mitochondrion inner membrane'; the human entry has not been brought into line, and until it is this IEA will regenerate. Propagation Review Root cause: SOURCE BAD Failure modes: COMPARTMENT OR COMPLEX MISMATCH Sources checked: UniProtKB-SubCell:SL-0097 · Endoplasmic reticulum membrane (UniProt subcellular location vocabulary) SOURCE BAD This IEA is generated mechanically from a UniProt SUBCELLULAR LOCATION line on Q9BUR5 that rests entirely on the 55 kDa immunoreactive species once assigned to MIC26. PMID:37279200 showed that species to be non-specific: it survives siRNA knockdown and CRISPR knockout in four human cell lines, is not reproduced by GFP- or myc-tagged MIC26, is unaffected by mutating the predicted glycosylation sites, and yields no MIC26 peptides by mass spectrometry. The UniProt line has not been retracted, so this row will regenerate until it is. Supporting Evidence: PMID:37279200 In these KOs, we used four anti-MIC26 antibodies and consistently detected the loss of mitochondrial MIC26 (22 kDa) and MIC27 (30 kDa) but not the loss of intracellular or secreted 55 kDa protein. Thus, the protein assigned earlier as 55 kDa MIC26 is nonspecific. PMID:37279200 we conclude that both MIC26 and MIC27 are exclusively localized in mitochondria and that the observed phenotypes reported previously are exclusively due to their mitochondrial function PMID:37279200 we did not find MIC26-specific peptides in the MS analysis in any of the three samples analysed which included WT, MIC26 and MIC27 KOs |
| GO:0007007 inner mitochondrial membrane organization | IC PMID:26477565 Evolution and structural organization of the mitochondrial c... | MODIFY | Summary: Curator inference from MIB complex membership to a role in organising the inner mitochondrial membrane. The premise is sound, but the conclusion is drawn at the wrong level: GO:0042407 cristae formation is a descendant of GO:0007007 and is already supported on this gene by IMP, IBA, IEA and NAS evidence. Reason: The parent term is true but uninformative here. MIC26's demonstrated contribution to inner-membrane organisation is specifically the formation of crista junctions - the single-knockout phenotype, the disease-variant phenotype and the family inference all name that process - so the specific child should carry the annotation. The contact-site aspect that GO:0007007 might otherwise have covered is captured structurally by the GO:0140275 MIB complex row rather than by a process term, and there is no evidence that MIC26 itself drives contact-site formation: MIC60, MIC19 and SAM50 are the subunits whose depletion destabilises the bridge. Note what this MODIFY does in practice: GO:0042407 is already on this gene four times over, so the replacement duplicates an existing row and the net effect is retirement of GO:0007007, not the addition of anything new. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: GRANULARITY MISMATCH Sources checked: GO:0140275 · MIB complex (the premise of the curator inference) SUPPORTS TRANSFER The premise is correct and independently evidenced: MIC26 really is in the MIB complex, by complexome profiling in this reference and by radiolabelled import in PMID:25781180. The issue is not the source but the granularity of the term inferred from it. Proposed replacements: cristae formation Supporting Evidence: PMID:25764979 MIC26 depletion led to alterations in mitochondrial ultrastructure and caused a significant reduction in the number of crista junctions. PMID:25781180 Mic25/CHCHD6, Mic27/ApoOL and Mic23/ApoO appear to be periphery subunits of the MICOS complex, because their depletion does not affect cristae morphology or stability of other components. PMID:25781180 We show that the most important subunits of the MIB complex in human mitochondria are Mic60/Mitofilin, Mic19/CHCHD3 and an outer membrane component Sam50. |
| GO:0042407 cristae formation | IBA GO_REF:0000033 | ACCEPT | Summary: Family inference from the single PTHR14564 PAINT node, whose GO:0042407 IBD is seeded by human APOO itself. The family is small and functionally homogeneous: its characterised members - C. elegans moma-1, Drosophila Mic26-27, human MIC26 and human MIC27 - all shape cristae, and none has been shown to do anything else. Reason: Core, and the phylogenetic placement is sound. The one-gene WITH/FROM is short because the clade is small and the seeding evidence is on the target itself, not because the inference is thin; the assertion is where the function arose, and no family member argues against it. Propagation Review Root cause: NO FAILURE CORE Sources checked: PANTHER:PTN001803267 · PTHR14564 MIC26/MIC27 IBD node SUPPORTS TRANSFER The same node as the MICOS-complex IBA. The committed PAINT slice shows a second IBD on it, GO:0042407, dated 2018-05-10 and seeded by human APOO alone. A single gene-level donor is not weak evidence here: the assertion is about where the function arose, and the family's characterised members (C. elegans moma-1, Drosophila Mic26-27, human MIC26 and human MIC27) all shape cristae, with none contradicting it. UniProtKB:Q9BUR5 · APOO / MIC26 (Homo sapiens, the target) SUPPORTS TRANSFER The target itself. Its IMP from PMID:25764979 (crista junction number falls on MIC26 depletion) is the descendant evidence behind this IBD, later reinforced by CRISPR single-knockout data (PMID:32788226). Expected, not circular. Supporting Evidence: PMID:25764979 MIC26 depletion led to alterations in mitochondrial ultrastructure and caused a significant reduction in the number of crista junctions. PMID:32788226 SKOs and DKO cells have significant reduction in CJs per mitochondrial section. PMID:25781180 Yet another protein involved in cristae maintenance, MOMA1, has been reported in Caenorabditis elegans PMID:25781180 Its human homologues are apolipoprotein O (ApoO) and ApoO like (ApoOL) proteins. |
| GO:0042407 cristae formation | IEA GO_REF:0000002 | ACCEPT | Summary: InterPro2GO mapping from the two MIC26/MIC27 entries, which map to cristae formation and MICOS complex and to nothing else. Reason: Correct family, correct term, and confirmed independently on this gene by IMP, IBA and NAS evidence. Propagation Review Root cause: NO FAILURE CORE Sources checked: InterPro:IPR019166 · MICOS complex subunit MIC26/MIC27 SUPPORTS TRANSFER InterPro family entry whose InterPro2GO mapping is exactly GO:0042407 plus GO:0061617, and whose description covers fungal and animal MIC26 together with the MIC27 paralogue. APOO is a core member, matched by its Pfam PF09769 (ApoO) domain. InterPro:IPR033182 · MICOS complex subunit MIC26/MIC27, animal SUPPORTS TRANSFER The animal-restricted child entry, same two GO mappings. APOO satisfies its taxonomic scope, so the narrower entry adds nothing incorrect. Supporting Evidence: file:human/APOO/APOO-uniprot.txt DR Reactome; R-HSA-8949613; Cristae formation. |
| GO:0042407 cristae formation | IMP PMID:25764979 The non-glycosylated isoform of MIC26 is a constituent of th... | ACCEPT | Summary: Depletion of MIC26 in human cells significantly reduced the number of crista junctions and altered mitochondrial ultrastructure. CRISPR knockout in HAP1 cells later reproduced this for the MIC26 single knockout and showed the MIC26/MIC27 double knockout to be worse still, while human APOO disease variants disrupt cristae by the same route. Reason: This is the core biological process of the gene and the one place where MIC26's contribution can be separated from the complex it belongs to: the single-gene perturbation has a measurable phenotype. A contemporaneous knockdown study reported no cristae effect and called MIC26 a periphery subunit (PMID:25781180); the later knockout data break the tie in favour of a real but partly MIC27-redundant contribution, since MIC26 and MIC27 are dispensable for the assembly of the rest of MICOS yet their individual loss still costs crista junctions. The same phenotype appears outside cultured cells: adipocyte-specific Apoo knockout mice show disrupted mitochondrial structure in brown adipocytes with impaired oxidative phosphorylation (PMID:37088120), which is mouse and tissue-specific and so is cited as corroboration rather than annotated on the human gene. Supporting Evidence: PMID:25764979 MIC26 depletion led to alterations in mitochondrial ultrastructure and caused a significant reduction in the number of crista junctions. PMID:32788226 SKOs and DKO cells have significant reduction in CJs per mitochondrial section. PMID:32788226 we found that MIC26 and MIC27 are dispensable for the stability and integration of the remaining MICOS subunits into the complex suggesting that they assemble late into the MICOS complex PMID:32439808 The mutation caused impaired processing of the protein during import and faulty insertion into the IMM. This was associated with altered MICOS assembly and cristae junction disruption. |
| GO:0042407 cristae formation | NAS PMID:25997101 QIL1 is a novel mitochondrial protein required for MICOS com... | ACCEPT | Summary: ComplexPortal's statement that the MICOS complex, of which MIC26 is a subunit, acts in cristae formation. Reason: A complex-level author statement that happens also to be true of this subunit individually, as the single-knockout crista junction phenotype shows. Redundant with the IMP row rather than additive, but correct. Supporting Evidence: PMID:25997101 loss of QIL1 resulted in MICOS disassembly with the accumulation of a MIC60-MIC19-MIC25 sub-complex and degradation of MIC10, MIC26, and MIC27 |
| GO:0044284 mitochondrial crista junction | NAS PMID:25997101 QIL1 is a novel mitochondrial protein required for MICOS com... | ACCEPT | Summary: ComplexPortal places the MICOS complex, and so its subunit MIC26, at the crista junction - the tubular neck connecting a crista to the inner boundary membrane. Reason: The crista junction is where MICOS works, and MIC26's own loss-of-function phenotype is a reduction in the number of these structures, so the sublocation is not merely inherited from the complex but matches this subunit's own genetics. Supporting Evidence: PMID:25764979 MIC26 depletion led to alterations in mitochondrial ultrastructure and caused a significant reduction in the number of crista junctions. PMID:32788226 SKOs and DKO cells have significant reduction in CJs per mitochondrial section. |
| GO:0061617 MICOS complex | HDA PMID:26477565 Evolution and structural organization of the mitochondrial c... | ACCEPT | Summary: Complexome profiling of purified mammalian mitochondria places MIC26 inside the MICOS complex and resolves the assembly ladder from MICOS through the membrane-bridging subcomplex to the complete MIB complex. Four independent IDA rows on this gene say the same thing. Reason: MICOS membership is the defining, core fact about this protein and is supported here by a high-throughput method whose conclusion is independently confirmed by co-immunoprecipitation, radiolabelled import and MICOS-directed proteomics. Supporting Evidence: PMID:26477565 We find three main forms of the complex: A) The MICOS complex, containing all the MICOS proteins, B) a membrane bridging subcomplex, containing in addition SAMM50, MTX2 and the previously uncharacterized MTX3, and C) the complete MIB complex containing in addition DNAJC11 and MTX1. PMID:25764979 we show that the human apolipoprotein MIC26 is a bona fide subunit of the MICOS complex and that MIC26 is linked to cardiolipin metabolism and promotes crista junction formation |
| GO:0061617 MICOS complex | IBA GO_REF:0000033 | ACCEPT | Summary: Correctly placed family inference. PTHR14564 is a small, opisthokont-restricted family named 'MICOS COMPLEX SUBUNIT MIC26 / MIC27 FAMILY MEMBER', every one of whose nine UniProt-reviewed members in the committed entries index is a MIC26 or MIC27 orthologue, and its single PAINT node PTN001803267 carries GO:0061617 as an IBD seeded by Drosophila Mic26-27, human MIC27 and human MIC26 itself. All three seeds have experimental support for the term, and the target is one of them. Reason: The node sits ancestral to the MIC26/MIC27 duplication, APOO is inside the clade that inherited the property, and APOO additionally carries four direct experimental annotations to the same term. There is nothing to argue with in the node placement. Propagation Review Root cause: NO FAILURE CORE Sources checked: FB:FBgn0038400 · Mic26-27 (Drosophila melanogaster) SUPPORTS TRANSFER Resolves through UniProt to Q9VEY5, the single pre-duplication MIC26/MIC27 gene of Drosophila (CG5903, dApoO). QuickGO shows Q9VEY5 carrying GO:0061617 by IMP (PMID:32439808) and IGI (PMID:38253420), so this donor is experimentally grounded on the term being transferred. PANTHER:PTN001803267 · PTHR14564 MIC26/MIC27 IBD node SUPPORTS TRANSFER The only PAINT node in this family. The committed slice interpro/panther/PTHR14564/PTHR14564-paint.tsv shows it carrying GO:0061617 as an IBD dated 2024-07-30, seeded by the fly gene plus both human paralogues. The fly gene and both human paralogues all receive IBA from this node (checked in QuickGO), so it sits at or near the family root, ancestral to the MIC26/MIC27 duplication; APOO is inside the inheriting clade and there is no IRD or IKR anywhere in the slice. UniProtKB:Q6UXV4 · APOOL / MIC27 (Homo sapiens, paralogue) SUPPORTS TRANSFER The human paralogue, verified in UniProt as MIC27_HUMAN. QuickGO shows it carrying two independent IDA GO:0061617 annotations (PMID:25781180, PMID:25997101). UniProtKB:Q9BUR5 · APOO / MIC26 (Homo sapiens, the target) SUPPORTS TRANSFER The target's own accession. Its four direct IDA/HDA GO:0061617 rows are among the descendant evidences the PAINT curator weighed in placing this IBD; a gene appearing in the WITH/FROM of its own IBA is expected and marks experimental grounding on the target, not circularity. Supporting Evidence: PMID:25781180 We provide additional proof that ApoO indeed is a subunit of the MICOS and MIB complexes and propose the name Mic23 for this protein. PMID:26477565 Mic23 and Mic27 appear to be the youngest MICOS proteins, as they only occur in opisthokonts. |
| GO:0061617 MICOS complex | IDA PMID:25764979 The non-glycosylated isoform of MIC26 is a constituent of th... | ACCEPT | Summary: Koob et al. showed that the 22 kDa non-glycosylated form of MIC26 spans the inner membrane and physically associates with MIC60, MIC27 and MIC10, establishing it as a constituent of the mammalian MICOS complex. Reason: Direct interaction evidence with three MICOS subunits, from the study that named the protein MIC26. Core. Note that only this paper's mitochondrial conclusions survive: the same laboratory later withdrew its glycosylated/secreted-form claims (PMID:37279200), which is why the ER, Golgi and extracellular rows from this reference are removed while this one is accepted. Supporting Evidence: PMID:25764979 The latter isoform spans the mitochondrial inner membrane and physically interacts with several MICOS complex subunits such as MIC60, MIC27, and MIC10. |
| GO:0061617 MICOS complex | IDA PMID:25781180 Detailed analysis of the human mitochondrial contact site co... | ACCEPT | Summary: Ott et al. detected ApoO in the >1 MDa MIB complex after radiolabelled import into isolated mitochondria and showed co-immunoprecipitation with MIC60/Mitofilin, concluding that it is a genuine MICOS subunit and proposing the name Mic23. Reason: Independent biochemical confirmation of complex membership by a second laboratory, using import into native mitochondria rather than steady-state co-migration. Supporting Evidence: PMID:25781180 we show that ApoO is detected in the >1 MDa MIB complex after radiolabelled import into mitochondria and co-immunoprecipitates with Mic60/Mitofilin PMID:25781180 We provide additional proof that ApoO indeed is a subunit of the MICOS and MIB complexes and propose the name Mic23 for this protein. |
| GO:0061617 MICOS complex | IDA PMID:25997101 QIL1 is a novel mitochondrial protein required for MICOS com... | ACCEPT | Summary: Guarani et al. recovered MIC26 in systematic proteomics of human MICOS, and showed that MIC26 is degraded along with MIC10 and MIC27 when the complex disassembles on loss of MIC13/QIL1 - the behaviour of a subunit, not of a co-purifying contaminant. Reason: A third independent laboratory, and the strongest kind of membership evidence short of structure: MIC26's stability depends on the integrity of the complex it is assigned to. Supporting Evidence: PMID:25997101 Through systematic proteomic analysis of human MICOS, we identified QIL1 (C19orf70) as a novel conserved MICOS subunit. PMID:25997101 loss of QIL1 resulted in MICOS disassembly with the accumulation of a MIC60-MIC19-MIC25 sub-complex and degradation of MIC10, MIC26, and MIC27 |
| GO:0061617 MICOS complex | IEA GO_REF:0000002 | ACCEPT | Summary: InterPro2GO mapping from the two MIC26/MIC27 family entries, both of which map to exactly this term plus cristae formation. APOO is a founding member of both entries. Reason: The domain-to-term mapping is precise rather than promiscuous: these InterPro entries contain only MIC26/MIC27 sequences, and every characterised member is a MICOS subunit. The electronic row agrees with four experimental rows on the same gene. Propagation Review Root cause: NO FAILURE CORE Sources checked: InterPro:IPR019166 · MICOS complex subunit MIC26/MIC27 SUPPORTS TRANSFER InterPro family entry whose InterPro2GO mapping is exactly GO:0042407 plus GO:0061617, and whose description covers fungal and animal MIC26 together with the MIC27 paralogue. APOO is a core member, matched by its Pfam PF09769 (ApoO) domain. InterPro:IPR033182 · MICOS complex subunit MIC26/MIC27, animal SUPPORTS TRANSFER The animal-restricted child entry, same two GO mappings. APOO satisfies its taxonomic scope, so the narrower entry adds nothing incorrect. Supporting Evidence: file:human/APOO/APOO-uniprot.txt DR PANTHER; PTHR14564; MICOS COMPLEX SUBUNIT MIC26 / MIC27 FAMILY MEMBER; 1. |
| GO:0061617 MICOS complex | NAS PMID:25997101 QIL1 is a novel mitochondrial protein required for MICOS com... | ACCEPT | Summary: ComplexPortal's curated statement of MICOS membership (CPX-6141), attributed to the Guarani MICOS proteomics paper. Reason: A database author statement that restates a claim experimentally established in the same and several other references. Redundant with the IDA rows but not wrong, and the ComplexPortal cross-reference is carried on the UniProt record itself. Supporting Evidence: file:human/APOO/APOO-uniprot.txt DR ComplexPortal; CPX-6141; MICOS mitochondrial contact site and cristae organizing system complex. |
| GO:0140275 MIB complex | HDA PMID:26477565 Evolution and structural organization of the mitochondrial c... | ACCEPT | Summary: Complexome profiling recovers MIC26 in the complete MIB complex, the ~2 MDa assembly in which MICOS in the inner membrane is bridged through MIC60 to SAMM50, the metaxins and DNAJC11 in the outer membrane. Reason: The definition of GO:0140275 explicitly includes MICOS components, and MIC26's presence in the >1 MDa MIB assembly was independently shown by radiolabelled import (PMID:25781180). This is the term that correctly captures what the complexome profile shows about MIC26 and outer-membrane proteins - unlike the SAM complex row from the same reference, which is removed. Supporting Evidence: PMID:26477565 We find three main forms of the complex: A) The MICOS complex, containing all the MICOS proteins, B) a membrane bridging subcomplex, containing in addition SAMM50, MTX2 and the previously uncharacterized MTX3, and C) the complete MIB complex containing in addition DNAJC11 and MTX1. PMID:25781180 we show that ApoO is detected in the >1 MDa MIB complex after radiolabelled import into mitochondria and co-immunoprecipitates with Mic60/Mitofilin |
| GO:1900208 regulation of cardiolipin metabolic process | IMP PMID:32788226 MIC26 and MIC27 cooperate to regulate cardiolipin levels and... | NEW | Summary: Mass-spectrometric lipidomics of CRISPR knockout human HAP1 cells shows that cellular cardiolipin falls significantly when MIC26 alone is deleted, and also in the MIC26/MIC27 double knockout, while MIC27 single knockout leaves it normal. The reduction affects all cardiolipin molecular species rather than a subset, so it reflects the amount of cardiolipin rather than a remodelling defect. That cardiolipin pool is functionally consequential: restoring it by overexpressing cardiolipin synthase rescues respiratory chain supercomplex stability in the double knockout. An independent link had already been noted, MIC26 levels correlating with those of tafazzin, the cardiolipin-remodelling acyltransferase (PMID:25764979). Reason: A cardiolipin phenotype that is specific to MIC26 among the two paralogues, measured in a human single-gene knockout, and shown to be causally upstream of the OXPHOS defect by a synthase-overexpression rescue. The obvious objection - that low cardiolipin is a passive consequence of disrupted cristae - is answered inside the same experiment: the MIC27 single knockout also loses crista junctions, yet its cardiolipin is normal. Crista junction loss therefore does not by itself lower cardiolipin, and the MIC26 requirement is separable from the architectural phenotype. GO:1900208 is used rather than the positive-regulation child because reduced steady-state cardiolipin does not by itself distinguish decreased synthesis from increased turnover, and the study does not resolve which. This is deliberately NOT annotated as GO:1901612 cardiolipin binding: direct cardiolipin binding has been demonstrated in vitro for the paralogue APOOL/MIC27 (PMID:23704930), not for MIC26, and the only claim of cardiolipin recruitment by MIC26 itself comes from coarse-grained molecular dynamics on an AlphaFold model (PMID:42647630). Supporting Evidence: PMID:32788226 We found significantly reduced levels of cardiolipin in DKO and MIC26 KO cells, whereas they remain normal in MIC27 KO cells PMID:32788226 SKOs and DKO cells have significant reduction in CJs per mitochondrial section. PMID:32788226 While cardiolipin was reduced in DKO cells, overexpression of cardiolipin synthase in DKO restores the stability of RCs/SC. PMID:25764979 we show that the human apolipoprotein MIC26 is a bona fide subunit of the MICOS complex and that MIC26 is linked to cardiolipin metabolism and promotes crista junction formation PMID:25764979 Both proteins are positively correlated with the levels of MIC10 as well as tafazzin, an enzyme required for cardiolipin remodeling. PMID:23704930 It specifically binds to cardiolipin in vitro but not to the precursor lipid phosphatidylglycerol. |
Loading supporting content…
Download this section (compressed HTML)Q: Does purified MIC26 bind cardiolipin, as its paralogue MIC27 does, and if so is that binding what sets the cellular cardiolipin level in MIC26 knockout cells?
Suggested experts: Ruchika Anand, Andreas S. Reichert, Arun Kumar Kondadi
Q: What distinguishes MIC26 from MIC27 such that only MIC26 knockout fragments the mitochondrial network, given that the two are reciprocally regulated, occupy the same subcomplex, and are redundant for crista junction formation?
Suggested experts: Ruchika Anand, Andreas S. Reichert
Q: Is the MIC26 N-terminal segment cleaved on import, and where is the mature N terminus? The answer decides whether UniProt's SIGNAL 1..25 should be reannotated as a transit peptide, matching the paralogue.
Suggested experts: Vera Kozjak-Pavlovic, Christian Ott
Q: Does MIC26 have one transmembrane helix or two, and are its amphipathic helices exposed to the intermembrane space?
Experiment: Express and purify recombinant human MIC26 and MIC27 in detergent, and assay cardiolipin binding by liposome flotation and native mass spectrometry against phosphatidylglycerol and phosphatidylcholine controls, replicating the MIC27 result as the positive control in the same hands. Then charge-reverse the basic loop residues predicted to contact cardiolipin, and test whether the mutant (i) loses lipid binding in vitro and (ii) fails to restore cardiolipin levels and crista junction number when re-expressed in MIC26 knockout HAP1 cells. A mutant that separates the two would show the cardiolipin phenotype is not simply a consequence of disrupted cristae.
Hypothesis: MIC26 binds cardiolipin directly through the basic loop between its transmembrane segments, and that binding is what maintains the cellular cardiolipin pool.
Type: in vitro lipid binding plus structure-guided mutagenesis rescue
Experiment: Immunopurify MIC26 from human mitochondria and determine its N terminus by N-terminal-oriented proteomics (TMPP or TAILS labelling), comparing the mature N terminus with the predicted 1-25 boundary. In parallel, import radiolabelled MIC26 into isolated mitochondria with and without membrane potential and with MPP inhibition, and test whether the precursor is processed. Substituting the MIC26 N-terminal segment for that of a canonical presequence protein, and vice versa, would test sufficiency.
Hypothesis: Residues 1-25 of MIC26 are a mitochondrial targeting presequence, not a secretory signal peptide, and are removed during import.
Type: N-terminomics plus in vitro mitochondrial import
Experiment: Map MIC26 topology by protease protection of intact mitochondria versus mitoplasts versus solubilised membranes, using epitope tags at the N terminus, the C terminus and the inter-helix loop, combined with cysteine-accessibility scanning using membrane-permeant and membrane-impermeant maleimides on single-cysteine variants in a cysteine-free background.
Hypothesis: MIC26 spans the inner membrane twice, with both amphipathic helices facing the intermembrane space.
Type: membrane topology mapping
Experiment: Measure fission and fusion events directly in MIC26 knockout, MIC27 knockout and double knockout HAP1 cells by live-cell imaging of matrix-targeted photoactivatable GFP and by scoring DRP1 puncta at constriction sites, with MIC26 re-expression as the specificity control. Then test epistasis by deleting DNM1L in the MIC26 knockout background: if fragmentation is fission-driven it should be abolished, whereas cristae disorder should persist.
Hypothesis: The fragmented network of MIC26 knockout cells reflects an increased fission rate rather than a passive consequence of disordered cristae.
Type: live-cell mitochondrial dynamics with epistasis
Experiment: Perform pulse-chase lipidomics with stable-isotope-labelled precursors in wild-type and MIC26 knockout cells to separate the synthesis and degradation arms, measuring CRLS1, PTPMT1, TAZ and PLA2 activities and levels alongside. Re-expression of MIC26 should restore whichever arm is affected.
Hypothesis: Cardiolipin falls in MIC26 knockout cells because synthesis decreases, not because turnover increases.
Type: stable-isotope pulse-chase lipidomics
What is not known — curated, literature-grounded statements of the open unknowns (the inverse of core functions).
Gap: MIC26 has no demonstrated molecular activity of its own. No biochemical assay has been performed on purified MIC26, so it is undetermined whether it contributes to crista junction formation by deforming the bilayer through its amphipathic helices, by binding a lipid, or purely as a scaffold that positions MIC10 and MIC27 - and consequently whether the membrane-bending activity of the MIC10 subcomplex requires MIC26 or merely tolerates it. The one positive result on that question, that a Mic10-Mic26 dimer bends the bilayer, is coarse-grained molecular dynamics on an AlphaFold 3 model, with no measurement behind it.
OPEN BIOLOGY MF_DARK
What is known: Firmly established: MIC26 is an integral inner-membrane protein, a constituent of the MIC10 subcomplex of MICOS, physically associated with MIC60, MIC27 and MIC10; its deletion in human cells significantly reduces crista junction number; and both known human pathogenic variants disrupt cristae. What is missing is the step between membership and phenotype.
Significance: MIC26 is one of two MICOS subunits whose loss causes human mitochondrial disease without abolishing the complex, so what it adds to the complex is exactly what the disease removes.
What would resolve it: In vitro reconstitution with purified MIC26: liposome tubulation or GUV deformation assays with MIC26 alone and with MIC10, and cryo-EM or cryo-ET of the MIC10 subcomplex with and without MIC26.
Provenance (the field's own admissions):
Gap: Whether MIC26 binds cardiolipin directly is unknown, and so is the mechanism by which its loss lowers the cellular cardiolipin pool - whether synthesis falls, turnover rises, or the lipid is destabilised by the loss of a binding site.
NARROWING BIOLOGY MF_DARK
What is known: Established: cardiolipin falls significantly in MIC26 single-knockout human cells and not in MIC27 single-knockout cells, the loss affects all cardiolipin molecular species rather than a subset so is not a remodelling defect, and restoring cardiolipin by cardiolipin-synthase overexpression rescues respiratory supercomplex stability. Established for the paralogue, not for MIC26: direct in vitro cardiolipin binding.
Significance: It decides whether GO:1901612 cardiolipin binding belongs on APOO at all, and it is the reason that term is deliberately not annotated here: the only positive claim for MIC26 comes from simulation, and the only experimental demonstration in this family is on APOOL.
What would resolve it: Lipid-binding assays (liposome flotation, lipid overlay, native MS) on purified recombinant MIC26 side by side with MIC27, testing cardiolipin against phosphatidylglycerol, plus charge-reversal mutants of the predicted cardiolipin-contacting loop.
Provenance (the field's own admissions):
Gap: Why MIC26, but not its paralogue MIC27, is required for a tubular mitochondrial network is unknown. Loss of MIC26 fragments mitochondria in human cells without changing OPA1 levels or the balance of its long and short forms, so the phenotype is not explained by the fusion machinery, and no connection to the fission machinery has been established in human cells.
OPEN BIOLOGY BP_DARK
What is known: Established: MIC26 knockout and MIC26/MIC27 double knockout HAP1 cells show comparable, significant increases in fragmented mitochondria while MIC27 knockout cells are indistinguishable from wild type; the defect is therefore MIC26-specific and not compensated by MIC27. A mouse macrophage knockout reaches an increased-fission state by lowering OPA1, which the human cells do not do.
Significance: It is the one MIC26 phenotype that MIC27 cannot cover, so it is the best available handle on what distinguishes the two paralogues; it is also why no mitochondrial-fission GO term is proposed here.
What would resolve it: Live-cell fission and fusion rate measurements (mito-PAGFP fusion assay, DRP1 recruitment imaging) in MIC26 versus MIC27 knockout cells with MIC26 re-expression as the specificity control, to establish whether the steady-state fragmentation reflects a change in fission rate at all.
Provenance (the field's own admissions):
Gap: The membrane topology of MIC26 is not experimentally determined. UniProt annotates one transmembrane helix (108-128); the paralogue MIC27 is annotated with two, and a 2026 structure-prediction and simulation study models two short transmembrane helices in both. Which is correct, and on which side of the membrane the amphipathic helices sit, is unresolved.
OPEN BIOLOGY CC_DARK
What is known: Established: MIC26 is an integral inner-membrane protein that is imported into mitochondria and inserted into the inner membrane, and a patient missense substitution at residue 117 - inside the annotated helix - impairs that insertion.
Significance: The topology determines which face of the inner membrane the amphipathic helices deform, and therefore any mechanistic model of how MIC26 contributes to crista junction curvature.
What would resolve it: Protease protection of intact mitochondria and mitoplasts combined with cysteine-accessibility scanning of tagged MIC26, or a structure of the MIC10 subcomplex.
Provenance (the field's own admissions):
Gap: The Q9BUR5 UniProt record still asserts Secreted, Golgi apparatus membrane and endoplasmic reticulum membrane subcellular locations and a chondroitin-sulfate O-glycosylation, all of which trace to a protein species shown not to be MIC26; and it annotates residues 1-25 as a SIGNAL peptide by prediction only, where the paralogue Q6UXV4 carries a TRANSIT peptide. This is an un-propagated curation update rather than a biological uncertainty: the mouse orthologue record Q9DCZ4 already lists only the mitochondrial inner membrane.
OPEN CURATION CC_DARK
What is known: Established: the 55 kDa species persists in APOO knockout cells in four human cell lines, is not produced by tagged MIC26, is unaffected by glycosylation-site mutagenesis, and yields no MIC26 peptides by mass spectrometry.
Significance: While those lines stand, GO_REF:0000044 will keep regenerating the three subcellular-location IEA rows removed in this review, and the gene will keep reading as an apolipoprotein to anything that consumes the UniProt record.
What would resolve it: UniProt removal of the Secreted, Golgi apparatus membrane and endoplasmic reticulum membrane locations, of the three-form Note, and of the chondroitin-sulfate PTM line, citing PMID:37279200; and reassessment of SIGNAL 1..25 against the paralogue's TRANSIT annotation.
Provenance (the field's own admissions):
Gap: There is no molecular function term in GO for a structural constituent of the MICOS complex, so the only MF assignable to MIC26 is the generic GO:0005198 structural molecule activity, which conveys nothing specific to this complex.
OPEN ONTOLOGY MF_DARK
What is known: Established: MIC26 contributes to the structural integrity of MICOS - it is a subunit, it is degraded when the complex disassembles, and its removal costs crista junctions without abolishing the complex. The biology is known; the vocabulary to state it in the MF aspect is not.
Significance: MIC26 and MIC27 both lack any demonstrated activity beyond a structural contribution, so they collapse onto the same uninformative MF term, and the MF aspect carries almost no information for the complex as a whole.
What would resolve it: Creation of a 'structural constituent of MICOS complex' child of GO:0005198, following the existing pattern of complex-specific structural-constituent terms.
Provenance (the field's own admissions):
Proposed term (ontology gap):
Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)