ACAP1 is a multidomain, phosphoinositide-regulated GTPase-activating protein for the small GTPase ARF6 and a component of the coat that sorts cargo out of recycling endosomes. Its four domains divide the work unconventionally: an N-terminal BAR domain that oligomerises ACAP1 into a lattice on the membrane rather than binding membrane itself, a pleckstrin-homology domain that binds phosphoinositides and performs the membrane binding and curvature generation a BAR domain would normally perform, a C4 zinc-finger Arf-GAP catalytic domain that accelerates GTP hydrolysis on ARF6 in preference to ARF1 and ARF5, and three ankyrin repeats. GAP activity is stimulated by phosphatidylinositol 4,5-bisphosphate and phosphatidic acid, so catalysis is coupled to lipid composition. ACAP1 is largely cytosolic at rest and is recruited by ARF6-GTP both to tubular recycling endosomes and to the cell surface during ruffling. On the endosome it forms part of an ARF6-regulated clathrin coat that packages cargo for return to the plasma membrane: it binds clathrin heavy chain directly and captures the cytoplasmic tails of integrin beta-1 and of the glucose transporter GLUT4/SLC2A4, thereby supporting integrin-dependent cell migration and insulin-stimulated glucose uptake. Cargo capture is a switch rather than a constitutive activity. Akt phosphorylation of Ser-554 relieves a local autoinhibition and licenses binding to a sorting signal in the integrin beta-1 tail, which couples growth-factor signalling directly to the recycling step. Because ARF6 also governs peripheral actin, ACAP1 acting on ARF6 dampens membrane protrusions, dorsal ruffles and focal adhesions; the PTB-domain adaptor GULP1 opposes this by sequestering ACAP1 in a complex with GDP-bound ARF6. ACAP1 is expressed most strongly in haematopoietic and lymphoid tissue, and is one of three paralogous human ACAPs together with ACAP2 and ACAP3.
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0005886 plasma membrane | IBA GO_REF:0000033 | KEEP AS NON CORE | Summary: Phylogenetic inference that ACAP1 is active at the plasma membrane. Retained, and unusually for an IBA on a location, it can be checked directly against human data rather than argued from the family. Jackson et al. imaged ACAP1 itself and found a genuine surface pool that is ARF6-dependent: ACAP1 localises to dorsal ruffles and to the edge of protrusive cells, and recruitment to both the endosome and the plasma membrane fails when ARF6 is locked in its GDP form. So this is not a compartment mismatch imported from other species; the human protein does go to the plasma membrane, under ARF6-GTP control. Two details argue that it is not the principal site of catalysis, which is why this is kept as non-core rather than accepted as core. First, ACAP1's steady-state location in UniProt is the recycling endosome membrane, peripheral, on the cytoplasmic side. Second, the GAP-dead R448Q mutant loses the tubular-endosome pool but is still found on plasma-membrane protrusions - surface recruitment and catalysis are separable, and it is the endosomal pool that tracks with GAP function. The sorting and coat work that defines ACAP1 happens on the endosome; the surface pool is where its substrate ARF6 is activated. Reason: Correct and independently supported by human imaging, but peripheral to the gene's core activity. ACAP1's coat and cargo-sorting function operates on the tubular recycling endosome, and the catalytically dead mutant retains the surface pool while losing the endosomal one. Note on the relation: is_active_in asserts that the function is carried out here, and that is the part of this annotation least well supported. Because R448Q keeps the surface pool while losing the endosomal one, the plasma-membrane pool may be recruitment without catalysis, in which case located_in would be the better relation. Retained as is_active_in because ARF6 is activated at the plasma membrane and no experiment has yet localised the hydrolysis event; a curator revisiting this row should treat the relation, not the term, as the open question. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q15057 · ACAP2 (human paralog) SUPPORTS TRANSFER Human ACAP2, a paralog rather than an ortholog, with plasma membrane by IDA. Within-family support of an ancestral node is legitimate PAINT practice, and plasma-membrane residence is ancestral across the ACAP clade, so this is not a bad transfer. WB:WBGene00000565 · cnt-1 (C. elegans) SUPPORTS TRANSFER Carries plasma membrane by EXP plus cytoplasmic side of plasma membrane, apical and basolateral plasma membrane by IDA - the most finely localised member of the family. PomBase:SPBC17G9.08c · csx2 (S. pombe) SUPPORTS TRANSFER Plasma membrane by IDA. dictyBase:DDB_G0279649 · Dictyostelium Q54WI0 SUPPORTS TRANSFER A full-architecture ACAP (BAR + PH + Arf-GAP by UniProt domain annotation) with plasma membrane by IDA. AGI_LocusCode:AT5G61980 · Arabidopsis AGD1 SUPPORTS TRANSFER ADP-ribosylation factor GTPase-activating protein AGD1. AGI_LocusCode:AT5G13300 · Arabidopsis AGD3 SUPPORTS TRANSFER ADP-ribosylation factor GTPase-activating protein AGD3. PANTHER:PTN001142372 · PTHR23180 ancestral node SUPPORTS TRANSFER The CENTAURIN/ARF family node from which the assignment was made. Supporting Evidence: PMID:11062263 During PDGF stimulation of NIH 3T3 cells, ACAP1, ACAP2, and ASAP1 colocalized in dorsal ruffles and at the edge of resting and protrusive HeLa cells. PMID:11062263 ACAP1 remained cytosolic and was not recruited to the endosome or the plasma membrane in cells expressing the dominant-negative Arf6, T27N, even in the presence of AlF4 PMID:11062263 A point mutant of ACAP1 that lacked GAP activity was not observed on the tubular endosomes, but rather was associated with the plasma membrane protrusions file:human/ACAP1/ACAP1-uniprot.txt CC -!- SUBCELLULAR LOCATION: Recycling endosome membrane |
| GO:0005096 GTPase activator activity | IBA GO_REF:0000033 | ACCEPT | Summary: The core molecular function, and correct - but arrived at by the wrong route, which is the central finding of this review. ACAP1's GAP activity was measured directly, on ACAP1, in the paper that named the family: recombinant ACAP1 preferred ARF6 over ARF1 and ARF5, the activity required phosphatidylinositol 4,5-bisphosphate, and the arginine-finger mutant R448Q was catalytically dead. That is textbook IDA evidence, and UniProt records it with ECO:0000269|PubMed:11062263. GO has none of it. A QuickGO query for ACAP1 annotations from PMID:11062263 returns zero. The same paper produced three IDAs on the paralog ACAP2 (GO:0005096 GTPase activator activity, GO:0030029 actin filament-based process, GO:0001726 ruffle). ACAP1 - first-named in the title, assayed in the same figures, and described in the text as the more potent of the two at inhibiting protrusions - received nothing. The consequence is an inversion of the usual paralog hazard: ACAP1's GAP activity now reaches ACAP1 only as a phylogenetic inference whose WITH/FROM cites ACAP2, for an activity that was measured on ACAP1 itself in the source paper. Accepted as core, because the function is right and the IBA is well founded. What is needed is an IDA on ACAP1 from PMID:11062263, which would make this IBA redundant rather than load-bearing. One ontology note. Substrate specificity cannot be expressed in the term. GO:0008060 ARF GTPase activator activity is now a secondary id of GO:0005096, merged along with the Rab, Ras and Rho GAP terms, so the best-established fact about this protein - that it is an ARF6 GAP with a measured preference over ARF1 and ARF5 - has to be carried by an annotation extension. Any experimental annotation of this term on ACAP1 should carry has_input UniProtKB:P62330 (ARF6). Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q15057 · ACAP2 (human paralog) SUPPORTS TRANSFER Human ACAP2, with GTPase activator activity by IDA from PMID:11062263 - the same paper and the same experiment in which ACAP1's own GAP activity was measured. A legitimate source for the ancestral node, but its presence here records that ACAP1's direct evidence was captured on the paralog and not on ACAP1. MGI:MGI:2153589 · Acap3 (mouse paralog) SUPPORTS TRANSFER Mouse Acap3, a paralog. Legitimate for an ancestral GAP-activity node. Note that the mouse Acap3-specific annotations (neuron migration, regulation of neuron projection development) hang off a different PANTHER node, PTN002754173, and correctly did not reach ACAP1. SGD:S000002932 · AGE1 (S. cerevisiae YDR524C) SUPPORTS TRANSFER Yeast Arf-GAP AGE1, showing the activity is ancestral well beyond the metazoa. dictyBase:DDB_G0279649 · Dictyostelium Q54WI0 SUPPORTS TRANSFER Full-architecture ACAP with GTPase activator activity by IDA. AGI_LocusCode:AT5G13300 · Arabidopsis AGD3 SUPPORTS TRANSFER ADP-ribosylation factor GTPase-activating protein AGD3. PANTHER:PTN001142372 · PTHR23180 ancestral node SUPPORTS TRANSFER The CENTAURIN/ARF node. Arf-GAP catalytic activity is the defining feature of the family, so the ancestral assignment is secure. Supporting Evidence: PMID:11062263 In vitro, ACAP1 and ACAP2 preferred Arf6 as a substrate, rather than Arf1 and Arf5, more so than did ASAP1. PMID:11062263 [R448Q]ACAP1 and [R442Q]ACAP2 had no detectable activity using Arf6 as a substrate PMID:11062263 ACAP1 and ACAP2 are preferentially activated by phosphatidylinositol 4,5-bisphosphate [PIP2; not shown] and do contain GAP activity. file:human/ACAP1/ACAP1-uniprot.txt CC -!- FUNCTION: GTPase-activating protein (GAP) for ADP ribosylation factor 6 |
| GO:0010008 endosome membrane | IBA GO_REF:0000033 | ACCEPT | Summary: Correct, and the site where ACAP1's core activity happens. Accepted. The human evidence is direct: ACAP1 is recruited from the cytosol onto tubular portions of the ARF6 endosomal compartment, and Chen et al. place it specifically on a tubular recycling endosome distinct from the ARAP2/APPL1-positive compartment. The GAP-dead mutant is absent from these tubules, tying the endosomal pool to catalysis. A more specific term exists and is already annotated on this gene by a different route: GO:0055038 recycling endosome membrane, from the UniProt SubCell mapping. This IBA is not modified to that term, because an IBA states what the ancestral node supports and the family is not uniformly recycling-endosomal - cnt-1 carries early endosome membrane by IDA as well as recycling endosome membrane. Endosome membrane is the right generality for a phylogenetic inference. The specificity belongs on the experimental annotation that is missing. Propagation Review Root cause: NO FAILURE CORE Sources checked: RGD:1562939 · Acap2 (rat paralog) SUPPORTS TRANSFER Rat Acap2, a paralog. Its RGD record describes involvement in endocytic recycling and location in endosome membrane, consistent with the transfer. WB:WBGene00000565 · cnt-1 (C. elegans) SUPPORTS TRANSFER Carries both early endosome membrane and recycling endosome membrane by IDA, which is why the generic endosome-membrane term is the correct level for the node. PANTHER:PTN001142372 · PTHR23180 ancestral node SUPPORTS TRANSFER The CENTAURIN/ARF family node. Supporting Evidence: PMID:11062263 In untreated cells, although ACAP1 and ACAP2 were primarily cytosolic, they could be observed on tubular portions of the Arf6 endosomal compartment PMID:25225293 ARAP2 formed a complex with APPL1 and colocalized with Arf6 and APPL in a compartment distinct from the Arf6/ACAP1 tubular recycling endosome. PMID:11062263 A point mutant of ACAP1 that lacked GAP activity was not observed on the tubular endosomes, but rather was associated with the plasma membrane protrusions |
| GO:0030036 actin cytoskeleton organization | IBA GO_REF:0000033 | KEEP AS NON CORE | Summary: Real but indirect, and downstream of the GAP activity rather than a function in its own right, so kept as non-core. For human ACAP1 the actin readouts exist: overexpression blocks ARF6-dependent protrusions in HeLa cells, and reduces focal adhesions. But these are consequences of removing active ARF6, which is itself a well-known regulator of peripheral actin. ACAP1 does not act on actin; it switches off the GTPase that does. Both experiments are overexpression, and the knockdown phenotype that ACAP1 does have is a traffic phenotype - accelerated integrin beta-1 internalisation - not an actin one. The WITH/FROM merits a note, because one source is a poor match for this gene even though the annotation survives it. FB:FBgn0004133 is Drosophila blown fuse (blow), which PANTHER places in PTHR23180 subfamily SF399, but blow has only a PH domain: UniProt annotates no BAR, no Arf-GAP and no ankyrin repeats on any of its isoforms (A1Z714, P91678, E1JGZ4, Q8MSU1). Its actin annotation is a myoblast-fusion IMP, a lineage-specific developmental phenotype. A domain-truncated family member is exactly the kind of source that produces a bad transfer when it is the only one. Here it is not: the two Dictyostelium sources are genuine ACAPs, and DDB_G0279649 in particular carries actin cytoskeleton organization by both IMP and IGI, plus regulation of filopodium assembly by IMP, filamentous actin by IDA and cell cortex by IMP. The ancestral inference is therefore sound; it is the role, not the evidence, that makes this non-core. Reason: A genuine but indirect consequence of ARF6 inactivation, shown by overexpression. ACAP1's own activity is GAP catalysis and cargo sorting on the recycling endosome; actin remodelling is what ARF6 does when ACAP1 is not acting on it. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: FB:FBgn0004133 · blow (blown fuse, Drosophila) SUPPORTS SOURCE BUT NOT TARGET Placed in PTHR23180 subfamily SF399 but has only a PH domain - no BAR, no Arf-GAP, no ANK repeats - so it is not an ACAP and cannot act through ARF6 GAP catalysis. Its GO:0030036 is an IMP tied to Drosophila myoblast fusion. Supports an actin role for blow, but should not on its own support one for a human Arf-GAP. dictyBase:DDB_G0279649 · Dictyostelium Q54WI0 SUPPORTS TRANSFER Full-architecture ACAP (BAR + PH + Arf-GAP) with actin cytoskeleton organization by IMP and IGI, regulation of filopodium assembly by IMP, filamentous actin by IDA and cell cortex by IMP. This is what actually carries the ancestral inference. dictyBase:DDB_G0276395 · Dictyostelium Q551Q8 SUPPORTS TRANSFER PH + Arf-GAP architecture, with actin cytoskeleton organization by IGI. PANTHER:PTN001142372 · PTHR23180 ancestral node SUPPORTS TRANSFER The CENTAURIN/ARF family node. Supporting Evidence: PMID:11062263 In HeLa cells, overexpression of either ACAP blocked the formation of Arf6-dependent protrusions. PMID:25225293 ARAP2 overexpression promoted large FAs, but ACAP1 overexpression reduced FAs. PMID:25225293 ARAP2 knockdown slowed, whereas ACAP1 knockdown accelerated, integrin β1 internalization. |
| GO:0005096 GTPase activator activity | IEA GO_REF:0000002 | ACCEPT | Summary: InterPro2GO mapping from IPR001164 (Arf-GAP domain) and IPR045258 (ACAP1/2/3-like). Both signatures match ACAP1 correctly - UniProt annotates the Arf-GAP domain at 405-527 with a C4-type zinc finger at 420-443 - and the inferred activity is the one that was measured directly on this protein. Accepted. Worth recording that IPR045258 is the ACAP1/2/3-specific signature, so this electronic annotation is more discriminating than the family-level Arf-GAP one and is not a case of a domain match over-reaching. It is also, together with the IBA above, the only thing standing between this gene and having no molecular function in GO at all. Supporting Evidence: file:human/ACAP1/ACAP1-uniprot.txt DR InterPro; IPR045258; ACAP1/2/3-like. file:human/ACAP1/ACAP1-uniprot.txt FT /note="Arf-GAP" PMID:11062263 In vitro, ACAP1 and ACAP2 preferred Arf6 as a substrate, rather than Arf1 and Arf5, more so than did ASAP1. |
| GO:0005737 cytoplasm | IEA GO_REF:0000002 | KEEP AS NON CORE | Summary: InterPro2GO mapping from IPR004148, the BAR domain, to cytoplasm. The mapping itself is uninformative - a BAR domain implies peripheral membrane binding and deformation, not a compartment - but the conclusion is true of this protein for a reason the mapping does not know: ACAP1 is genuinely a soluble cytosolic protein at rest, and only becomes membrane-bound when ARF6-GTP recruits it. Jackson et al. describe it as primarily cytosolic in untreated cells, and it stays cytosolic when ARF6 is locked off. Retained rather than modified. A more specific term would be recycling endosome membrane, which GOA already carries from the SubCell route, but replacing cytoplasm with it would discard a pool of the protein that is real and, for a GAP recruited on demand, functionally meaningful. The two annotations describe different states of the same protein rather than competing for one answer. Kept as non-core rather than accepted: the cytosolic pool is where ACAP1 waits, not where it works. None of the core functions below lists cytoplasm as a location - all four are sited on the recycling endosome membrane - so marking this core would overstate it, and it would be inconsistent to grade the generic term above the more informative plasma-membrane annotation, which is also non-core. Reason: True and experimentally corroborated - ACAP1 is primarily cytosolic until ARF6-GTP recruits it - but it records the resting state rather than a site of action, and it derives from an InterPro2GO mapping (BAR domain to cytoplasm) that carries no compartment information. Supporting Evidence: PMID:11062263 In untreated cells, although ACAP1 and ACAP2 were primarily cytosolic, they could be observed on tubular portions of the Arf6 endosomal compartment PMID:11062263 ACAP1 remained cytosolic and was not recruited to the endosome or the plasma membrane in cells expressing the dominant-negative Arf6, T27N, even in the presence of AlF4 |
| GO:0055038 recycling endosome membrane | IEA GO_REF:0000044 | ACCEPT | Summary: The best and most informative location on the gene, and the one place where the experimental literature did reach GO - though only by accident of routing. This row is an electronic mapping of UniProt SubCell SL-0231. But UniProt's underlying statement is not a prediction: it is ECO:0000269|PubMed:16256741, a curator's reading of a primary experiment, and it specifies the topology as well as the compartment - peripheral membrane protein, cytoplasmic side. So the evidence code understates the evidence. The same compartment is independently supported by the ARF6-dependent tubular endosome imaging and by the tubular recycling endosome assignment in Chen et al. Accepted as core. It should be an experimental annotation from PMID:16256741 rather than an IEA, which is the same gap as for the molecular function. Supporting Evidence: file:human/ACAP1/ACAP1-uniprot.txt CC -!- SUBCELLULAR LOCATION: Recycling endosome membrane file:human/ACAP1/ACAP1-uniprot.txt CC {ECO:0000269|PubMed:16256741}; Peripheral membrane protein PMID:25225293 ARAP2 formed a complex with APPL1 and colocalized with Arf6 and APPL in a compartment distinct from the Arf6/ACAP1 tubular recycling endosome. |
| GO:0005515 protein binding | IPI PMID:17474147 Systematic identification of SH3 domain-mediated human prote... | MARK AS OVER ANNOTATED | Summary: Partner resolved: UniProtKB:P62993 = GRB2. The assay is a peptide-array screen, in which 12 SH3 domains from eight proteins were run against 1536 immobilised peptides and yielded 921 interactions. Judged against ACAP1's own biology, an SH3 hit is chemically plausible - ACAP1 has a proline-rich stretch at 538-549 within its disordered 525-581 region, the kind of PxxP context an SH3 domain reads. But an array of synthetic peptides tests a motif, not a protein, and this particular pair was not among the ones the authors carried forward to co-immunoprecipitation. No functional consequence of an ACAP1-GRB2 association has been reported since. Marked over-annotated for what the term records rather than for the partner being impossible. Bare protein binding conveys nothing about ACAP1's function, and there is no informative molecular-function term to modify it to, because nothing is known about what the interaction would do. GRB2 is nonetheless the one screen partner that recurs across two orthogonal assays (see PMID:25814554 below), which makes it the one worth testing directly. Reason: Uninformative bare protein binding from a synthetic-peptide array screen that was not validated for this pair. Retained in the record as a lead on GRB2 rather than as a statement of function. Supporting Evidence: PMID:17474147 A group of 12 Src homology (SH) 3 domains from eight human proteins (Swiss-Prot ID: SRC, PLCG1, P85A, NCK1, GRB2, FYN, CRK) were used to screen a peptide target array composed of 1536 potential ligands, which led to the identification of 921 binary interactions between these proteins and 284 targets. file:human/ACAP1/ACAP1-uniprot.txt FT /note="Pro residues" |
| GO:0005515 protein binding | IPI PMID:25814554 Phospho-tyrosine dependent protein-protein interaction netwo... | MARK AS OVER ANNOTATED | Summary: Partner resolved: UniProtKB:P62993 = GRB2 again, this time from a phosphotyrosine-dependent yeast two-hybrid network in which human kinases were co-expressed to generate pY sites, yielding 292 pY-dependent interactions. GRB2 recurring in a second, mechanistically different assay is the strongest signal in ACAP1's whole interaction record. The difficulty is that the assay premise does not fit what is known about the protein: a pY-dependent GRB2 interaction implies a phosphotyrosine on ACAP1 for the GRB2 SH2 domain, and no tyrosine phosphorylation of ACAP1 is on record in UniProt. The only annotated tyrosine modification is 3'-nitrotyrosine at 485, from a pituitary-adenoma nitroproteome, and Y712 was mutated to F with no effect on cargo binding. So the two GRB2 hits may not even be reporting the same binding mode - one an SH3-PxxP contact, the other nominally SH2-pY. Marked over-annotated on the same grounds as the peptide-array row: bare protein binding, no functional follow-up, and nothing specific enough to modify it to. Reason: Uninformative bare protein binding from a large-scale pY-dependent Y2H screen whose premise - a phosphotyrosine on ACAP1 - has no independent support in the UniProt record. Supporting Evidence: PMID:25814554 We extended an established yeast two-hybrid system employing human protein kinases for the analyses of phospho-tyrosine (pY)-dependent PPIs in a direct experimental, large-scale approach. PMID:25814554 We identified 292 mostly novel pY-dependent PPIs file:human/ACAP1/ACAP1-uniprot.txt FT /note="3'-nitrotyrosine" |
| GO:0005515 protein binding | IPI PMID:32296183 A reference map of the human binary protein interactome. | MARK AS OVER ANNOTATED | Summary: Three partners from HuRI, the systematic binary interactome, all resolved and judged separately. UniProtKB:P14927 = UQCRB, cytochrome b-c1 complex subunit 7. A nuclear-encoded subunit of mitochondrial respiratory complex III, resident in the inner membrane. There is no route by which a cytosolic protein recruited to recycling-endosome membranes by ARF6-GTP encounters it. Compartment-implausible. UniProtKB:O76041 = NEBL, nebulette, an actin-binding Z-disc protein of cardiac muscle. ACAP1 is expressed most strongly in bone marrow, intestine and lymphoid tissue, and low in heart. A shared actin connection is superficially suggestive, but ACAP1 acts on actin only through ARF6, and no cardiac function has been reported for it. UniProtKB:O94868-3 = FCHSD2 isoform 3, an F-BAR and double-SH3 protein that functions in clathrin-mediated endocytosis. This is the one hit whose shape makes sense: two membrane-bending proteins operating on the clathrin pathway, and FCHSD2's SH3 domains would have the ACAP1 proline-rich region available to them. It remains a single-assay Y2H hit, and the HuRI authors note that most of their interactions were found in only one screen and that the cellular function of most individual PPIs is still unknown. Marked over-annotated as a row: bare protein binding, with the two implausible partners recorded alongside the plausible one. The three could be split into separate entries - the format allows one entry per partner, and GOA does carry three separate rows here - but all three reach the same action, because an unreplicated single-assay screen hit yields an uninformative protein binding annotation whether or not the partner is chemically plausible. Splitting would change the bookkeeping and not the verdict, so the partners are distinguished in prose and in supporting_entities instead. FCHSD2 is the one worth pursuing experimentally. Reason: Uninformative bare protein binding from single-assay Y2H. Two of the three partners are compartment-implausible for a cytosolic ARF6-recruited endosomal protein; the third, FCHSD2, is mechanistically sensible but unreplicated. Supporting Evidence: PMID:32296183 the majority of PPIs in HuRI were found in only one screen PMID:32296183 While HuRI displays highly significant overlap with known functional relationships, the cellular function of most individual PPIs remains to be elucidated. file:human/ACAP1/ACAP1-uniprot.txt DR HPA; ENSG00000072818; Tissue enhanced (bone marrow, intestine, lymphoid tissue). |
| GO:0005515 protein binding | IPI PMID:32814053 Interactome Mapping Provides a Network of Neurodegenerative ... | MARK AS OVER ANNOTATED | Summary: Six partners from a neurodegeneration-focused Y2H interactome that screened about 500 ND-related proteins against roughly 5000 human proteins to produce about 30000 candidate interactions. ACAP1 appears as prey, not as one of the disease proteins of interest. All six resolved. Compartment-implausible: UniProtKB:P22695 = UQCRC2, mitochondrial respiratory complex III core protein 2 - a second complex III subunit, after UQCRB in the HuRI row, and the recurrence of complex III subunits across unrelated screens is itself a signature of sticky-prey behaviour rather than biology. UniProtKB:Q15047-2 = SETDB1 isoform 2, a nuclear histone H3K9 methyltransferase. UniProtKB:Q92993 = KAT5, a nuclear histone acetyltransferase. UniProtKB:Q8TAP4-4 = LMO3 isoform 4, a nuclear LIM-only protein. Not implausible, but unsupported: UniProtKB:P17252 = PRKCA and UniProtKB:P61981 = YWHAG. Both are cytosolic and both are generic hub proteins in Y2H data. A kinase and a 14-3-3 protein are the sort of partners a phosphoprotein might genuinely have - ACAP1 is phosphorylated on Ser-554 by PKB - but 14-3-3 binding requires a specific phosphomotif that has not been mapped on ACAP1, and no PKC phosphorylation of ACAP1 is on record. Marked over-annotated. Nothing here is informative about ACAP1's function, and four of the six partners are in compartments the protein does not reach. Reason: Uninformative bare protein binding from a large-scale Y2H screen in which ACAP1 was incidental prey. Four of six partners are mitochondrial or nuclear and unreachable by a cytosolic ARF6-recruited endosomal protein; the two cytosolic ones are generic Y2H hubs with no mapped interaction surface on ACAP1. Supporting Evidence: PMID:32814053 connects ∼5,000 human proteins via ∼30,000 candidate interactions and is generated by systematic yeast two-hybrid interaction screening of ∼500 ND-related proteins and integration of literature interactions file:human/ACAP1/ACAP1-uniprot.txt FT /note="Phosphoserine; by PKB" |
| GO:0016020 membrane | HDA PMID:19946888 Defining the membrane proteome of NK cells. | MARK AS OVER ANNOTATED | Summary: ACAP1 was identified in the membrane fraction of the YTS natural-killer cell line by two-dimensional LC-MS/MS, one of 1843 proteins. The native context is right - ACAP1 is tissue-enhanced in bone marrow and lymphoid tissue, so an NK-like line is a reasonable place to find it, and detection at protein level is not in doubt. What the experiment cannot do is say which membrane, and in this instance it argues against itself. The authors treated the isolated membranes specifically to strip peripheral proteins, and describe the surviving non-integral fraction as proteins that would be predicted to be transiently associated with membranes. ACAP1 is precisely such a protein: peripheral, cytoplasmic-side, recruited on demand by ARF6-GTP and otherwise soluble. So it is in this dataset as a member of the class the method was trying to remove, and the term it produced - an unspecified bilayer - conveys none of what is actually known about where the protein sits. Marked over-annotated for lack of specificity, not because the detection is doubted. GO:0055038 recycling endosome membrane, already on the gene, is the statement this row should have been. Reason: Unspecified membrane from a proteomic fractionation that by design cannot resolve which membrane, and that was attempting to deplete exactly the peripheral class ACAP1 belongs to. A more specific and better-supported compartment term is already present. Supporting Evidence: PMID:19946888 Isolated membranes were treated with reagents that have been reported to remove peripheral membrane proteins. PMID:19946888 The remaining species were largely involved in cellular processes and molecular functions that could be predicted to be transiently associated with membranes. file:human/ACAP1/ACAP1-uniprot.txt CC {ECO:0000269|PubMed:16256741}; Peripheral membrane protein |
| GO:0032456 endocytic recycling | IMP PMID:17664335 An ACAP1-containing clathrin coat complex for endocytic recy... | NEW | Summary: Proposed new annotation. ACAP1 has no biological process annotation describing what it actually does, despite the title of PMID:17664335 being "An ACAP1-containing clathrin coat complex for endocytic recycling". That paper shows ACAP1 is required for two distinct recycling events - stimulation-dependent integrin recycling and insulin-stimulated GLUT4 recycling - and PMID:25225293 supplies an independent loss-of-function confirmation from a different laboratory, in which ACAP1 knockdown accelerates integrin beta-1 internalisation. Evidence code IMP rather than IDA: the cellular evidence rests on knockdown and on interfering constructs, not on a direct assay of the protein at native levels. Reason: The gene's principal biological process, established in two independent laboratories, and currently absent from GO entirely. Supporting Evidence: PMID:17664335 We find that ACAP1, a GTPase-activating protein (GAP) for ADP-ribosylation factor (ARF) 6, is part of a novel clathrin coat complex that is regulated by ARF6 for endocytic recycling in two key physiological settings, stimulation-dependent recycling of integrin that is critical for cell migration and insulin-stimulated recycling of glucose transporter type 4 (Glut4), which is required for glucose homeostasis. PMID:25225293 ARAP2 knockdown slowed, whereas ACAP1 knockdown accelerated, integrin β1 internalization. |
| GO:0140312 cargo adaptor activity | IDA PMID:22645133 Mechanistic insights into regulated cargo binding by ACAP1 p... | NEW | Summary: Proposed new annotation, and the term that captures what ACAP1 does that a GAP alone would not. The definition of GO:0140312 requires binding the structural scaffolding of a vesicle coat and bridging membrane, cargo and deformation machinery. ACAP1 does all three: it binds clathrin heavy chain directly, it recognises a defined sorting sequence in the integrin beta-1 cytoplasmic tail - co-crystallised with residues 378-740 in PDB 3T9K and 4F1P - and it deforms the membrane through its own PH domain. IDA is appropriate because the cargo-signal recognition was demonstrated by direct binding and structure determination with purified protein. Only ITGB1 (UniProtKB:P05556) is listed as cargo, because that is the cargo this paper assayed; SLC2A4/GLUT4 (UniProtKB:P14672) is a second cargo, bound via its third cytoplasmic loop, but that comes from UniProt SUBUNIT and PMID:17664335, so it would belong on a separate annotation. This is offered in preference to the generic GO:0060090 molecular adaptor activity or GO:0005198 structural molecule activity, either of which would lose the coat and cargo context. Reason: Records ACAP1's coat-adaptor function, which is distinct from and additional to its GAP activity, at the correct level of specificity. Supporting Evidence: PMID:22645133 We initially defined a critical sequence in the cytoplasmic domain of integrin β1 recognized by ACAP1 and showed that this sequence acts as a recycling sorting signal. PMID:22645133 ACAP1 (ArfGAP with coiled-coil, ankyrin repeat, and PH domains protein 1) is an ARF6 GAP that also acts as a key component of a recently defined clathrin complex for endocytic recycling. file:human/ACAP1/ACAP1-uniprot.txt CC bound ARF6 and GULP1. Interacts with ITGB1; required for ITGB1 file:human/ACAP1/ACAP1-uniprot.txt CC Interacts with CLTC. Interacts with GULP1. Forms a complex with GDP- PMID:17664335 We find that ACAP1, a GTPase-activating protein (GAP) for ADP-ribosylation factor (ARF) 6, is part of a novel clathrin coat complex that is regulated by ARF6 for endocytic recycling in two key physiological settings, stimulation-dependent recycling of integrin that is critical for cell migration and insulin-stimulated recycling of glucose transporter type 4 (Glut4), which is required for glucose homeostasis. |
| GO:0030118 clathrin coat | IDA PMID:17664335 An ACAP1-containing clathrin coat complex for endocytic recy... | NEW | Summary: Proposed new annotation. ACAP1 is a subunit of an ARF6-regulated clathrin coat isolated in PMID:17664335, and its direct interaction with clathrin heavy chain is recorded in UniProt SUBUNIT with ECO:0000269 for that paper. Complex membership belongs in a part_of cellular component annotation, and GO:0030118 is a GO:0032991 descendant, so this is the correct slot for it. One caveat for whoever curates it: GO:0030118 is defined around coated pits and coated vesicles linked to the membrane through an AP adaptor complex, whereas the ACAP1 coat operates on recycling endosomes and is regulated by ARF6. The term is the closest available but is silent on which trafficking step the coat serves. Reason: Records the complex membership demonstrated in PMID:17664335, which no current annotation conveys. Placed in a part_of cellular component annotation rather than as a location. Supporting Evidence: PMID:17664335 We find that ACAP1, a GTPase-activating protein (GAP) for ADP-ribosylation factor (ARF) 6, is part of a novel clathrin coat complex that is regulated by ARF6 for endocytic recycling in two key physiological settings, stimulation-dependent recycling of integrin that is critical for cell migration and insulin-stimulated recycling of glucose transporter type 4 (Glut4), which is required for glucose homeostasis. file:human/ACAP1/ACAP1-uniprot.txt CC Interacts with CLTC. Interacts with GULP1. Forms a complex with GDP- |
| GO:0180020 membrane bending activity | IDA PMID:25284369 A PH domain in ACAP1 possesses key features of the BAR domai... | NEW | Summary: Proposed new annotation. Purified ACAP1 binds and deforms liposomes, and Pang et al. localised the activity precisely: the BAR domain can do neither on its own and the neighbouring PH domain performs both membrane binding and curvature generation, with the BAR domain instead oligomerising ACAP1 into the lattice that bends the bilayer. The F280 mutagenesis series in UniProt scores the activity residue by residue - F280A reduces membrane binding and liposome tubulation, F280E almost abolishes binding, F280W preserves both. IDA: liposome tubulation with purified protein is a direct in vitro assay of the activity. The cellular counterpart in PMID:17010122 is weaker evidence because it depends on co-overexpression of PIP5K1C. The domain attribution is the part of this row most likely to be got wrong, so it is worth stating twice and in the negative. The BAR domain is not the bending element. In the liposome assay the isolated BAR domain shows little affinity for liposomes of any size, and in the cryoEM reconstruction of coated tubules there is no significant interaction between the BAR domain and the underlying membrane at all. The bilayer contact and the curvature both come from the PH domain, and specifically from Loop1, whose distal aromatic F280 inserts into one leaflet: F280A reduces both binding and tubulation, F280E effectively abolishes binding, and the conservative F280W preserves both. What the BAR domain does instead is bind the BAR domains of neighbouring ACAP1 molecules. So "the BAR-PH module tubulates membrane" is the right answer for the fragment and the wrong answer for the mechanism, and "its BAR domain bends membrane" - the phrasing a reader would default to for any other BAR protein - is simply false for ACAP1. GO:0180020 was created in June 2023, after all of ACAP1's primary literature, which is a likely reason the activity is unannotated. It is also still very sparsely used: a QuickGO query for the term and its descendants in human returns four annotations over three gene products - CHMP2A and CHMP3 (ESCRT-III, PMID:36604498) and OPA1 (PMID:32228866, PMID:37612504) - all IDA, and none of them a BAR-domain protein. ACAP1 would be the first BAR-domain entry on the term, which is a reason to state the PH-domain attribution explicitly rather than let the term be read as a routine BAR-protein annotation. Reason: A distinct, directly assayed molecular function with an unusual mechanism worth recording, and a term that did not exist when the work was published. The activity belongs to the PH domain: the BAR domain neither binds membrane nor imparts curvature, and contributes only the BAR-BAR contacts that cluster ACAP1 on the bilayer. Supporting Evidence: PMID:25284369 Here, we show that this BAR domain can neither bind membrane nor impart curvature, but instead requires a neighboring PH (Pleckstrin Homology) domain to achieve these functions. PMID:25284369 Specific residues within the PH domain are responsible for both membrane binding and curvature generation. PMID:25284369 We next found that ACAP1BAR-PH also induces membrane curvature, as reflected by the tubulation of liposomes, which was visualized by negative-stain electron microscopy (EM) PMID:25284369 the BAR domain of ACAP1 also showed little affinity to the generated liposomes, regardless of their size PMID:25284369 Second, there is no significant interaction between the BAR domain and the underlying membrane. PMID:25284369 In particular, mutation of this residue to alanine (F280A) reduced the ability of ACAP1BAR-PH to induce liposome tubulation, while a more conservative mutation (F280W) preserved the ability to tubulate liposomes. file:human/ACAP1/ACAP1-uniprot.txt FT /note="F->A: Reduced membrane binding and ability to induce |
| GO:0042803 protein homodimerization activity | IDA PMID:25284369 A PH domain in ACAP1 possesses key features of the BAR domai... | NEW | Summary: Proposed new annotation, and the other half of the domain division of labour recorded on GO:0180020 above. The one function ACAP1's BAR domain does have is self-association: it binds the BAR domain of another ACAP1 molecule. UniProt states this as a domain property with experimental evidence for this paper - the BAR domain mediates homodimerization - and its SUBUNIT line describes the protein as a banana-shaped homodimer. The evidence resolves stoichiometry, which is what makes the term annotatable rather than merely suggestive. Three orthogonal readouts agree on a dimer: the 2.2 angstrom crystal structure contains a banana-shaped dimer in one asymmetric unit, with a 4100 square-angstrom and largely hydrophobic interface; size exclusion chromatography with light scattering confirms dimerisation in solution; and quantitative gel filtration puts the dissociation constant below 5 micromolar, tighter than the published value for the N-BAR protein endophilin A2. This is the distinction that matters for a self-association term - an affinity-capture or co-immunoprecipitation experiment reports co-complex membership and cannot report stoichiometry, whereas crystallography, light scattering and a measured Kd do. Two things are deliberately not claimed. First, the higher-order assembly. The paper also describes lateral packing of dimers into tetramers and helical packing of those tetramers on the tubule, and calls the BAR-BAR function clustering at the membrane - but the tetramer and the lattice come from fitting the crystal structure into cryoEM maps of roughly 14 and 17 angstrom resolution, and clustering of an unspecified number of subunits is not a dimer. So the defensible dimer term is used and the lattice is left in prose, rather than reaching for an oligomerisation or complex-assembly claim the resolution does not support. Second, the assays used the BAR-PH fragment - residues 1-377 in the cryoEM work, 1-364 modelled in the crystal - and not full-length ACAP1, so this is a domain-attributed statement about ACAP1 rather than a demonstration on the intact protein. This is kept out of core_functions deliberately. Homodimerisation is the structural precondition for the curved scaffold rather than an output of the protein, and it is already described inside the membrane-bending core function; promoting it would give a prerequisite the same weight as the GAP activity and the cargo-sorting activity. Reason: The only demonstrated function of ACAP1's BAR domain, and the fact that makes the PH-domain attribution of GO:0180020 legible rather than surprising. Annotated at the dimer level because dimer stoichiometry is what the crystallography, the light scattering and the measured dissociation constant establish; the tetramer and the helical lattice are cryoEM-fitted models and are left in prose. Not proposed as a core function. Supporting Evidence: PMID:25284369 Dimerization of ACAP1BAR-PH was confirmed from size exclusion chromatography and light scattering PMID:25284369 The dimer interface has a large area (4100 Å2), and contains mainly hydrophobic residues that likely mediate dimerization. PMID:25284369 The in vitro disassociation constant of the dimerization was calculated to be below 5 µM, according to the quantitative gel-filtration experiment PMID:25284369 The BAR domain adjacent to the PH domain instead interacts with the BAR domains of neighboring ACAP1 proteins to enable clustering at the membrane. file:human/ACAP1/ACAP1-uniprot.txt CC -!- DOMAIN: The BAR domain mediates homodimerization, it can neither bind file:human/ACAP1/ACAP1-uniprot.txt CC -!- SUBUNIT: Banana-shaped homodimer laterally assembling into tetramers, |
| GO:0005546 phosphatidylinositol-4,5-bisphosphate binding | IDA PMID:17010122 Cooperation of phosphoinositides and BAR domain proteins in ... | NEW | Summary: Proposed new annotation. The ACAP1 PH domain binds phosphoinositides directly, and the single substitution K274N abolishes binding to both PIP2 and PIP3 and with it the association with endosomal tubules. Independently, PIP2 is what stimulates ACAP1's GAP activity in vitro, so this is not an incidental lipid affinity but the input that gates catalysis. Only the PIP2 term is proposed here. UniProt also lists phosphatidic acid, PI3P, PI(3,5)P2 and PIP3 for this domain, and GO:0005547 phosphatidylinositol-3,4,5-trisphosphate binding would be equally defensible from the same K274N result; PIP2 is chosen because it is the species with corroborating functional evidence from the GAP assay. Reason: A directly demonstrated molecular function with a point mutant that separates it from the rest of the protein, and the regulatory input that couples ACAP1's catalysis to membrane lipid composition. Supporting Evidence: file:human/ACAP1/ACAP1-uniprot.txt FT /note="K->N: Loss of binding to PIP2 and PIP3. Loss of file:human/ACAP1/ACAP1-uniprot.txt CC -!- DOMAIN: PH domain binds phospholipids including phosphatidic acid, PMID:11062263 ACAP1 and ACAP2 are preferentially activated by phosphatidylinositol 4,5-bisphosphate [PIP2; not shown] and do contain GAP activity. |
| GO:0072659 protein localization to plasma membrane | IMP PMID:17664335 An ACAP1-containing clathrin coat complex for endocytic recy... | NEW | Summary: Proposed new annotation. This is the outcome of ACAP1's sorting activity stated as a process: the cargoes it captures on the endosome - integrin beta-1 and GLUT4 - are delivered to the cell surface, and interfering with ACAP1 blocks that delivery. Akt phosphorylation of Ser-554 is required, and the S554A mutant loses both ITGB1 binding and integrin-dependent migration. Proposed alongside GO:0032456 endocytic recycling rather than instead of it, because the two describe different aspects: the route taken and the destination reached. IMP, since the evidence is from knockdown, interfering constructs and phosphosite mutants. Reason: Records the functional output of ACAP1-dependent sorting - arrival of integrin beta-1 and GLUT4 at the cell surface - which no existing annotation captures. Supporting Evidence: PMID:17664335 We find that ACAP1, a GTPase-activating protein (GAP) for ADP-ribosylation factor (ARF) 6, is part of a novel clathrin coat complex that is regulated by ARF6 for endocytic recycling in two key physiological settings, stimulation-dependent recycling of integrin that is critical for cell migration and insulin-stimulated recycling of glucose transporter type 4 (Glut4), which is required for glucose homeostasis. PMID:16256741 Disrupting the activities of either ACAP1 or Akt, or their assembly with endosomal beta1, inhibits beta1 recycling and cell migration. file:human/ACAP1/ACAP1-uniprot.txt CC ITGB1-dependent cell migration. {ECO:0000269|PubMed:16256741}. |
Loading supporting content…
Download this section (compressed HTML)Q: Why does ACAP1 have no experimental GO annotation? All seven of its primary papers - PMID:11062263, PMID:16256741, PMID:17398097, PMID:17664335, PMID:22645133, PMID:25284369 and PMID:25225293 - return zero ACAP1 annotations on QuickGO, so the ARF6 GAP activity, the integrin and GLUT4 recycling, the clathrin coat membership and the membrane-bending mechanism are all absent. UniProt has read every one of them, with ECO:0000269 throughout. The annotations most clearly warranted are GO:0005096 IDA with has_input ARF6 from PMID:11062263, GO:0055038 IDA from PMID:16256741, and GO:0032456 endocytic recycling IMP from PMID:17664335 or PMID:25225293.
Q: Should the ACAP2 IDAs from PMID:11062263 be mirrored onto ACAP1? That paper produced three IDAs on ACAP2 - GTPase activator activity, actin filament-based process and ruffle - and none on ACAP1, although ACAP1 is first in the title, was assayed in the same figures, and is described as the more effective of the two at inhibiting protrusions. As it stands, ACAP1's GAP activity reaches ACAP1 only as an IBA whose WITH/FROM cites ACAP2, which inverts the normal direction of paralog inference.
Q: How should ARF6 substrate specificity be recorded now that GO:0008060 ARF GTPase activator activity has been merged into GO:0005096? The merge was deliberate and applied to the Rab, Ras and Rho GAP terms too, so the intended mechanism is presumably an annotation extension. If so, has_input UniProtKB:P62330 should be treated as required rather than optional on GAP annotations, otherwise the distinction between an ARF6 GAP, an ARF1 GAP and a Rho GAP is lost at the point where it matters most.
Q: Is there a GO representation for a coat that acts on the recycling route rather than at endocytosis? ACAP1's clathrin coat is regulated by ARF6 and operates on recycling endosomes, which is not the clathrin-coated pit of endocytosis. GO:0030118 clathrin coat is the only available complex term and is silent on which step it serves.
Q: Which ACAP1 interactions deserve annotation? Not one of its validated partners - ARF6, ITGB1, CLTC, GULP1, SLC2A4 - appears anywhere in the GO record, while eleven protein binding rows record screen hits including two mitochondrial complex III subunits and three nuclear chromatin or LIM proteins. Among the screen hits only FCHSD2, an F-BAR protein of the clathrin pathway, and GRB2, which recurs in two orthogonal assays, look worth pursuing.
Q: Does ACAP1 have a haematopoietic function? Expression is enhanced in bone marrow, intestine and lymphoid tissue, and the original survey found the highest levels in lung and spleen, yet every mechanistic experiment has been done in HeLa, COS-7, NIH 3T3 or adipocytes. Integrin recycling matters a great deal in leukocytes, so the native setting may be where the phenotype is.
Experiment: Knock out ACAP1 and ACAP2 singly and together in a cell line expressing both, and measure integrin beta-1 and GLUT4 recycling kinetics with surface-biotinylation and antibody-uptake assays, then attempt cross-rescue with each paralog including the GAP-dead R448Q equivalents. A clean cross-rescue would justify shared annotation; failure to rescue would mean the two genes need annotating separately and would make the current reliance on ACAP2 in ACAP1's IBA WITH/FROM inappropriate.
Hypothesis: ACAP1 and ACAP2 are not interchangeable on the recycling route, and the annotation asymmetry between them reflects which was assayed rather than a biological difference.
Type: genetics
Experiment: Combine live-cell imaging of endogenously tagged ACAP1 with an ARF6 activity biosensor, and compare wild type with R448Q, which is reported to lose the tubular endosome pool while retaining plasma-membrane protrusion association. Localising the actual ARF6 hydrolysis event would settle whether is_active_in plasma membrane is warranted or whether located_in is the correct qualifier there.
Hypothesis: ACAP1's catalytic activity acts on the endosome while its plasma-membrane pool is non-catalytic.
Type: imaging
Experiment: Perform proximity labelling from endogenously tagged ACAP1 in a cell type where it is natively abundant, such as a myeloid or lymphoid line, before and after Akt activation, then test candidate transmembrane cargoes for the phospho-Ser-554-dependent binding that defines a genuine ACAP1 sorting substrate. This distinguishes cargo from bystander far better than Y2H and would populate the interaction record with partners that mean something.
Hypothesis: The cargo range of the ACAP1 coat extends beyond integrin beta-1 and GLUT4, as the M4 muscarinic receptor result suggests.
Type: proteomics
Experiment: Test the HuRI hit directly, by co-immunoprecipitation of endogenous proteins, by mapping which FCHSD2 SH3 domain engages the ACAP1 proline-rich 538-549 region, and by asking whether FCHSD2 depletion changes ACAP1-dependent endosomal tubulation. This is the one screen interaction whose mechanism is predictable, so it is the one whose confirmation would convert a bare protein binding row into an informative annotation.
Hypothesis: FCHSD2 and ACAP1 collaborate on membrane deformation in the clathrin pathway.
Type: biochemistry
Experiment: Introduce S554A and S554D at the endogenous ACAP1 locus by knock-in rather than by overexpression, then measure integrin beta-1 surface delivery, focal adhesion turnover and migration with and without growth-factor stimulation. All the existing evidence for the switch comes from transfected constructs, so a knock-in would establish whether the autoinhibition operates at native expression levels.
Hypothesis: Akt-phosphorylated Ser-554 is the switch that couples growth-factor signalling to integrin recycling in a physiological setting.
Type: genetics
Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)