AP1S2 encodes sigma1B (also called sigma-2), one of three vertebrate small-subunit isoforms that can occupy the sigma slot of the heterotetrameric AP-1 clathrin adaptor, alongside the large gamma (AP1G1 or AP1G2) and beta1 (AP1B1) adaptins and a medium mu1 subunit. Within that complex the sigma subunit is not a spacer: together with the N-terminal trunk of the gamma adaptin it forms the binding site for (D/E)XXXL(L/I) dileucine sorting signals in the cytosolic tails of transmembrane cargo, and human sigma1B assembles specifically into AP-1 and not into AP-2, AP-3 or AP-4. The complex is recruited to trans-Golgi network and endosomal membranes, links clathrin to those membranes, and buds cargo-laden vesicles that traffic between the TGN, endosomes and lysosomes. Unlike the ubiquitous sigma1A isoform, sigma1B is tissue-restricted and its biology is most penetrant in neurons. In hippocampal neurons the AP-1/sigma1B complex reforms synaptic vesicles from early endosomes after stimulation; without it, vesicle reformation is slow and incomplete and large clathrin-budding endosomal intermediates accumulate in the presynaptic terminal. sigma1B also carries an isoform-private protein interaction that sigma1A lacks: it binds the Rab5 exchange factor Rabex-5, which prevents assembly of the AP-1/sigma1A-ArfGAP1-Rabex-5 complex, lowers endosomal Rabex-5 and thereby restrains Rab5/Vps34-driven maturation of early endosomes into multivesicular late endosomes, balancing recycling of synaptic-vesicle proteins against their degradation. Outside the nervous system it directs a dileucine-motif cargo, sortilin, toward lysosomes in adipocytes. Loss-of-function variants cause an X-linked intellectual disability syndrome with highly variable expressivity that has been described under three separate names - Fried syndrome, Pettigrew syndrome and X-linked intellectual disability with hydrocephalus - typically combining intellectual disability, hypotonia and delayed walking with basal ganglia calcification or iron deposition, hydrocephalus or Dandy-Walker malformation, and seizures. Patient fibroblasts show no detectable defect in AP-1 stability, localization or function, consistent with sigma1A and sigma1C covering the ubiquitous AP-1 functions and the phenotype arising from a brain-restricted sorting defect.
Definition: Binding to a (D/E)XXXL(L/I) dileucine-based sorting signal in the cytosolic domain of a transmembrane protein. Such signals are recognised jointly by two subunits of a heterotetrameric adaptor protein complex - the gamma and sigma1 subunits of AP-1, the alpha and sigma2 subunits of AP-2, or the delta and sigma3 subunits of AP-3 - so a single subunit contributes to, rather than independently enables, this activity.
Justification: The molecular activity of an AP complex sigma subunit is signal recognition, and it has been mapped to individual residues by mutagenesis, yet GO offers no term for it. The only molecular function available for AP1S2 today is GO:0035615 clathrin-cargo adaptor activity, which is a complex-level term and whose definition additionally restricts it to endocytic vesicle formation. A signal-binding term would let the subunit-level knowledge be annotated with contributes_to, and would apply equally to AP1S1, AP1S3, AP2S1, AP3S1 and the corresponding large subunits.
Parent term: protein binding
Supporting Evidence:
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0000139 Golgi membrane | TAS Reactome:R-HSA-421831 | ACCEPT | Summary: Reactome places the AP-1 clathrin coat, including the sigma1b variant complex, on the trans-Golgi network membrane during coat assembly. This is the membrane on which AP-1 is recruited and assembled, and it is a core location for the protein. Reason: Coat assembly at the TGN is the canonical AP-1 step and the Golgi membrane is where it happens. UniProt states the same location independently for this entry, and the GO:0030121 AP-1 adaptor complex term that AP1S2 carries has GO:0005794 among its own ontology ancestors, so Golgi membrane is consistent with everything else asserted about the complex. The reaction-level citation is complex-level rather than sigma-isoform-specific, but the location does not depend on which sigma isoform is present. Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt SUBCELLULAR LOCATION: Golgi apparatus. Cytoplasmic vesicle membrane; |
| GO:0000139 Golgi membrane | TAS Reactome:R-HSA-421833 | ACCEPT | Summary: Same Golgi membrane location, cited to the Reactome reaction in which AP-1 binds VAMP and TGN membrane while capturing cargo. Reason: Cargo capture at the TGN membrane is exactly the step in which the gamma/sigma1 hemicomplex engages dileucine sorting signals, so this location is core rather than incidental. The residue analysis in this review confirms AP1S2 retains the sigma-side residues of that binding site. Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt FUNCTION: Subunit of clathrin-associated adaptor protein complex 1 that |
| GO:0000139 Golgi membrane | TAS Reactome:R-HSA-421835 | ACCEPT | Summary: Golgi membrane during TGN vesicle scission, the step at which the assembled coat pinches off a vesicle. Reason: Scission is still a Golgi-membrane event and the coat is present on it, so the location is correct. It is a duplicate of the other Reactome Golgi-membrane rows, which is acceptable - the same location is legitimately reached through several reactions. Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt SUBCELLULAR LOCATION: Golgi apparatus. Cytoplasmic vesicle membrane; |
| GO:0000139 Golgi membrane | TAS Reactome:R-HSA-432706 | ACCEPT | Summary: Golgi membrane during assembly of the coat on TGN membrane destined for the lysosome. Reason: Same membrane, a different downstream destination. The lysosome-directed branch is the one for which AP1S2 has isoform-specific cargo evidence: sigma1B binds sortilin through a dileucine-type motif and sigma1B-deficient adipocytes deliver less sortilin to lysosomes. Supporting Evidence: PMID:24928897 Ο1B-specific binding of sortilin requires the sortilin DxxD-x12-DSxxxL motif. |
| GO:0000139 Golgi membrane | TAS Reactome:R-HSA-432707 | ACCEPT | Summary: Golgi membrane during scission of the TGN-derived lysosomal vesicle. Reason: Correct for the same reason as the other Golgi-membrane rows; the coat is on the donor membrane until the vesicle separates. Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt SUBCELLULAR LOCATION: Golgi apparatus. Cytoplasmic vesicle membrane; |
| GO:0000139 Golgi membrane | TAS Reactome:R-HSA-432712 | ACCEPT | Summary: Golgi membrane during AP-1 binding and cargo capture on the lysosome-destined Golgi membrane. Reason: Cargo capture on the TGN membrane is the AP-1 step in which the sigma subunit participates directly, through the gamma/sigma1 dileucine-signal binding site. Supporting Evidence: PMID:21097499 all AP-1 hemicomplexes containing Ξ³1 (Ξ³1-Ο1A, Ξ³1-Ο1B, and Ξ³1-Ο1C) interact with similar avidities with the Nef, tyrosinase, and LIMP-II signals |
| GO:0005515 protein binding | IPI PMID:25416956 A proteome-scale map of the human interactome network. | MODIFY | Summary: High-throughput interaction with AP1G1 isoform 2 (gamma-1 adaptin). The partner is sigma1B's direct structural neighbour in the AP-1 core, so the informative content of this row is complex membership, not a generic binding activity. Reason: GO:0005515 states nothing about function. The partner resolves to AP1G1, the large gamma adaptin with which sigma1 forms the gamma/sigma1 hemicomplex, and that pairing is established by targeted work as well as by interactome screens: Takatsu et al. found sigma1B in yeast two-hybrid with both gamma adaptins, and Mattera et al. showed HA-tagged human sigma1B co-precipitating gamma1, beta1 and mu1 and no AP-2, AP-3 or AP-4 subunit. The informative replacement is therefore AP-1 adaptor complex membership. Proposed replacements: AP-1 adaptor complex Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt P56377; O43747-2: AP1G1; NbExp=4; IntAct=EBI-1054374, EBI-10185819; PMID:21097499 all three Ο1-HA isoforms are incorporated into AP-1 complexes containing Ξ³1, Ξ²1, and ΞΌ1 subunits but not into complexes including the Ξ±C subunit of AP-2, the Ξ²3A subunit of AP-3, or the Ο΅ subunit of AP-4 |
| GO:0005515 protein binding | IPI PMID:26496610 A human interactome in three quantitative dimensions organiz... | MODIFY | Summary: Interaction with canonical AP1G1 from a quantitative interactome study. Same partner, same conclusion as the other AP1G1 rows. Reason: The partner is the gamma-1 adaptin, sigma1B's obligate partner in the AP-1 core. A bare protein binding term hides that; AP-1 adaptor complex states it. Note that the cached record for this reference is abstract-only and does not name AP1S2 - the pair is in the study's supplementary data - so the quotable evidence for the interaction itself is the curated UniProt INTERACTION block rather than the paper narrative. Proposed replacements: AP-1 adaptor complex Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt P56377; O43747: AP1G1; NbExp=4; IntAct=EBI-1054374, EBI-447609; |
| GO:0005515 protein binding | IPI PMID:28514442 Architecture of the human interactome defines protein commun... | MODIFY | Summary: Interaction with AP1G1 isoform 2 from the BioPlex-derived interactome architecture study. Reason: Same partner and same reasoning as the other AP1G1 rows: the informative annotation is AP-1 complex membership. Note that IntAct's NbExp counts experiments within IntAct, not independent studies, so the several AP1G1 rows should not be read as several independent confirmations; the independent confirmation comes from the targeted experiments in PMID:9733768 and PMID:21097499. Proposed replacements: AP-1 adaptor complex Supporting Evidence: PMID:9733768 gamma2-adaptin is capable of interacting not only with the sigma1 chain (called as sigma1A in this paper), the small chain of the AP-1 complex, but also with a novel sigma1-like protein, designated as sigma1B |
| GO:0005515 protein binding | IPI PMID:29892012 An interactome perturbation framework prioritizes damaging m... | MODIFY | Summary: Interaction with AP1G1 from an interactome-perturbation study of disease missense variants. Reason: The partner is again the gamma-1 adaptin. This study's framing - testing whether disease-associated missense variants disrupt specific interactions - is relevant to AP1S2 because the reported human variants are predominantly nonsense and splice-site rather than missense, so the AP1G1 interaction is the wild-type baseline rather than a characterised disease mechanism. The informative replacement remains complex membership. Proposed replacements: AP-1 adaptor complex Supporting Evidence: PMID:17186471 we identified two nonsense mutations and one consensus splice-site mutation in the AP1S2 gene on Xp22 in three families |
| GO:0005515 protein binding | IPI PMID:32296183 A reference map of the human binary protein interactome. | MARK AS OVER ANNOTATED | Summary: Binary yeast two-hybrid interaction with MAB21L2 from the HuRI reference interactome. Reproducible within that assay format but functionally uncharacterised. Reason: MAB21L2 is a mab-21-like developmental protein with no described role in membrane traffic, and no follow-up study connects it to AP-1, to clathrin coats, or to any sigma1B function; searching Europe PMC for AP1S2 together with MAB21L2 returns no mechanistic paper. The interaction is curated by UniProt with five IntAct experiments, so it is not an isolated artefact, but a bare protein binding term on an unexplained binary hit adds no functional information and should not be promoted to any specific molecular function. Keeping it flagged as over-annotated rather than removing it leaves the observation available should a function emerge. Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt P56377; Q9Y586: MAB21L2; NbExp=5; IntAct=EBI-1054374, EBI-6659161; |
| GO:0005515 protein binding | IPI PMID:33961781 Dual proteome-scale networks reveal cell-specific remodeling... | MODIFY | Summary: Interaction with AP1G1 from the BioPlex 3.0 dual cell-line interactome. Reason: Same partner as the other AP1G1 rows. The gamma1-sigma1B pairing is the physical basis of the cargo-recognition site the complex uses, so the annotation should say that AP1S2 is part of AP-1 rather than that it binds a protein. Proposed replacements: AP-1 adaptor complex Supporting Evidence: PMID:21097499 the Ξ³1 isoform was incorporated into AP-1 complexes containing either Ο1A, Ο1B, or Ο1C, the Ξ³2 isoform was incorporated into AP-1 complexes containing Ο1A or Ο1B, but not Ο1C |
| GO:0005737 cytoplasm | IEA GO_REF:0000117 | MODIFY | Summary: ARBA machine-learning rule assigning the broadest possible cytoplasmic location. True of a peripheral coat protein on the cytoplasmic face of Golgi and vesicle membranes, but the least informative of the several locations already annotated on this gene. Reason: AP1S2 is cytoplasmic - it has no signal peptide or transmembrane segment and UniProt describes it as a peripheral membrane protein on the cytoplasmic side - so the term is not wrong. It is simply too broad: the informative statement about the non-membrane-bound pool is cytosol, which Reactome already annotates ten times on this gene, and the membrane-associated pool is covered by the Golgi, TGN and vesicle-membrane rows. Tracing the cited rule also failed (see propagation_review), which is a further reason not to lean on this row for anything. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: GRANULARITY MISMATCH Sources checked: ARBA:ARBA00026971 Β· ARBA rule ARBA00026971 (annotates GO:0005737 cytoplasm) SOURCE STALE OR MISSING Fetched live from rest.uniprot.org/arba/ARBA00026971: 2388 condition sets, conditions of type FunFam id, InterPro id, taxon and PANTHER id, and a single annotation, GO:0005737. Scanning every set against AP1S2's actual signature complement - InterPro IPR000804/IPR011012/IPR016635/IPR022775/IPR044733 from the InterPro REST record for P56377, FunFam 3.30.450.60:FF:000009 from the UniProt DR line, PANTHER PTHR11753 - leaves only two sets that mention anything AP1S2 has, and neither can fire: one pairs IPR011012 with taxon Saccharomyces, and the other requires IPR027156 'AP-2 complex subunit sigma' alongside IPR016635 and IPR022775, which AP1S2 does not match (its family signature is the AP-1 one, IPR044733). The rule's FunFam conditions cover 3.30.450.60:FF:000003/4/7/8/10/11 but not FF:000009. So the published rule does not reproduce this row. That is a trace failure, not a reason to drop a term that is biologically true of a peripheral coat protein on the cytoplasmic face of Golgi and vesicle membranes. Proposed replacements: cytosol Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt Peripheral membrane protein; Cytoplasmic side. Membrane, clathrin- |
| GO:0005765 lysosomal membrane | NAS PMID:23247405 Cell type-specific Rab32 and Rab38 cooperate with the ubiqui... | KEEP AS NON CORE | Summary: ComplexPortal projection onto the sigma1b AP-1 complex from a commentary on lysosome-related organelle biogenesis. The AP-1 coat is a transient visitor to the lysosomal membrane at the point of vesicle delivery, not a resident of it. Reason: An AP-1 coat assembles on the donor TGN or endosomal membrane and disassembles before or at delivery - the source itself says the clathrin coat and adaptors disassemble on vesicle budding and return to the cytosolic pool. Being detectable at the lysosomal membrane during that hand-off is real but peripheral, and it is not where the sigma subunit does its work. Kept rather than removed because sigma1B does have a lysosome-directed cargo route in at least one tissue. Supporting Evidence: PMID:23247405 Upon vesicle budding, the clathrin coat and adaptors disassemble and return to the cytosolic pool so they can be reutilized in further rounds of traffic PMID:24928897 Ο1B deficiency does not lead to a block of sortilin transport out of a specific organelle, but the fraction that reaches lysosomes is reduced. |
| GO:0005765 lysosomal membrane | TAS Reactome:R-HSA-432688 | KEEP AS NON CORE | Summary: Reactome models uncoating of the TGN-derived lysosomal vesicle, which it places at the lysosomal membrane; the AP-1 coat is present at that instant. Reason: A modelling convention as much as a localization: the reaction is the coat leaving, so annotating the coat to the acceptor membrane records a hand-off rather than a residence. Defensible within the pathway model, not a core site of action for the sigma subunit. Supporting Evidence: PMID:23247405 Upon vesicle budding, the clathrin coat and adaptors disassemble and return to the cytosolic pool so they can be reutilized in further rounds of traffic |
| GO:0005765 lysosomal membrane | TAS Reactome:R-HSA-432707 | KEEP AS NON CORE | Summary: Lysosomal membrane during scission of the TGN-derived lysosomal vesicle. Reason: Same reasoning as the other lysosomal-membrane rows. Note that the same Reactome reaction also generates a Golgi-membrane row on this gene, which is the more informative of the two compartment assignments for a coat that is being formed on the donor membrane. Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt FUNCTION: Subunit of clathrin-associated adaptor protein complex 1 that |
| GO:0005769 early endosome | NAS PMID:23247405 Cell type-specific Rab32 and Rab38 cooperate with the ubiqui... | ACCEPT | Summary: Early endosome, projected from the same ComplexPortal source as the melanosome and platelet rows - but unlike those, this one is independently correct for sigma1B, and strongly so. Reason: The cited reference is weak in itself (a commentary, projected across 48 gene products), but the term it asserts is confirmed by sigma1B-specific work. The compartment that accumulates in sigma1B-deficient synapses was characterised biochemically as a classic early endosome, and the AP-1/sigma1B complex is shown to act on early endosomes in regulating their maturation. A weak reference and a wrong term are different problems; this is only the former. Supporting Evidence: PMID:25128028 endosomes proved to be classic early endosomes with an increase in the phospholipid phosphatidylinositol 3-phosphate (PI-3-P) PMID:27411398 AP-1/Ο1A and AP-1/Ο1B regulate maturation of these early endosomes into multivesicular body late endosomes PMID:23247405 The localization of Rab32 and Rab38 is likely to specific tubular domains of early/recycling endosomes that contain AP-1, AP-3 or BLOC-2 |
| GO:0005794 Golgi apparatus | IDA GO_REF:0000052 | ACCEPT | Summary: Direct immunofluorescence localization of AP1S2 to the Golgi apparatus, curated by the Human Protein Atlas. This is the only direct experimental localization of the human protein in the annotation set. Reason: An IDA on the human protein for the compartment UniProt lists first in its SUBCELLULAR LOCATION line, and the compartment on which the AP-1 complex is assembled. Core. Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt SUBCELLULAR LOCATION: Golgi apparatus. Cytoplasmic vesicle membrane; |
| GO:0005794 Golgi apparatus | IEA GO_REF:0000044 | ACCEPT | Summary: Keyword-mapped Golgi apparatus location, derived from UniProt's own subcellular location statement for this entry. Reason: The mapping is one-to-one and the underlying UniProt statement is correct; the same term is independently supported by HPA immunofluorescence on this gene. A duplicate of the IDA row, which is fine. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB-SubCell:SL-0132 Β· UniProt subcellular location SL-0132 Golgi apparatus SUPPORTS TRANSFER SL-0132 maps one-to-one to GO:0005794 in the UniProt locations service. The underlying UniProt SUBCELLULAR LOCATION line for P56377 begins 'Golgi apparatus', and the same term is independently annotated IDA from HPA immunofluorescence, so the mapping and the source statement agree. Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt SUBCELLULAR LOCATION: Golgi apparatus. Cytoplasmic vesicle membrane; |
| GO:0005829 cytosol | TAS Reactome:R-HSA-182263 | KEEP AS NON CORE | Summary: Cytosolic pool of AP-1, cited to the Reactome MHC class I degradation reaction in the HIV-1 Nef pathway. Reason: AP complexes cycle between a soluble cytosolic pool and membranes, so a cytosol annotation is true; it is the inactive state rather than the site of action. Relevant background: the HIV-1 Nef dileucine signal used in these Reactome models is one of the signals the gamma1-sigma1B hemicomplex was directly shown to bind. Supporting Evidence: PMID:21097499 all AP-1 hemicomplexes containing Ξ³1 (Ξ³1-Ο1A, Ξ³1-Ο1B, and Ξ³1-Ο1C) interact with similar avidities with the Nef, tyrosinase, and LIMP-II signals |
| GO:0005829 cytosol | TAS Reactome:R-HSA-182279 | KEEP AS NON CORE | Summary: Cytosol, cited to formation of the MHC I:Nef:AP-1:PACS-1 complex. Reason: Same reasoning: the cytosolic pool is where unrecruited AP-1 sits. True, generic, not core. Supporting Evidence: PMID:23247405 Upon vesicle budding, the clathrin coat and adaptors disassemble and return to the cytosolic pool so they can be reutilized in further rounds of traffic |
| GO:0005829 cytosol | TAS Reactome:R-HSA-182286 | KEEP AS NON CORE | Summary: Cytosol, cited to transport of the MHC I:Nef:AP-1:PACS-1 complex. Reason: True but generic, as for the other cytosol rows. Supporting Evidence: PMID:23247405 Upon vesicle budding, the clathrin coat and adaptors disassemble and return to the cytosolic pool so they can be reutilized in further rounds of traffic |
| GO:0005829 cytosol | TAS Reactome:R-HSA-2130619 | KEEP AS NON CORE | Summary: Cytosol, cited to TGN-lysosomal vesicle coat assembly - the point at which cytosolic AP-1 is recruited to the membrane. Reason: The reaction consumes the cytosolic pool, so the annotation records the starting state of the complex. Accurate, uninformative. Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt Peripheral membrane protein; Cytoplasmic side. Membrane, clathrin- |
| GO:0005829 cytosol | TAS Reactome:R-HSA-2213236 | KEEP AS NON CORE | Summary: Cytosol, cited to uncoating of the TGN-lysosome vesicle and release of the nonameric complex. Reason: Uncoating returns AP-1 to the cytosol, so this is the product side of the same cycle. True, not core. Supporting Evidence: PMID:23247405 Upon vesicle budding, the clathrin coat and adaptors disassemble and return to the cytosolic pool so they can be reutilized in further rounds of traffic |
| GO:0005829 cytosol | TAS Reactome:R-HSA-421833 | KEEP AS NON CORE | Summary: Cytosol, cited to VAMP and TGN AP-1 binding coupled with cargo capture. Reason: Same cycle, same verdict. The informative half of this reaction for AP1S2 is the Golgi membrane row it also generates. Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt Peripheral membrane protein; Cytoplasmic side. Membrane, clathrin- |
| GO:0005829 cytosol | TAS Reactome:R-HSA-421836 | KEEP AS NON CORE | Summary: Cytosol, cited to uncoating of the TGN-derived vesicle. Reason: Uncoating releases the adaptor back to the cytosol; true but generic. Supporting Evidence: PMID:23247405 Upon vesicle budding, the clathrin coat and adaptors disassemble and return to the cytosolic pool so they can be reutilized in further rounds of traffic |
| GO:0005829 cytosol | TAS Reactome:R-HSA-432688 | KEEP AS NON CORE | Summary: Cytosol, cited to uncoating of the TGN-derived lysosomal vesicle. Reason: As above. Ten Reactome reactions independently generate this same cytosol row, which illustrates how a single compartment convention can dominate an annotation set without adding information. Supporting Evidence: PMID:23247405 Upon vesicle budding, the clathrin coat and adaptors disassemble and return to the cytosolic pool so they can be reutilized in further rounds of traffic |
| GO:0005829 cytosol | TAS Reactome:R-HSA-432712 | KEEP AS NON CORE | Summary: Cytosol, cited to AP-1 binding and cargo capture on the lysosome-destined Golgi membrane. Reason: Same cycle and same verdict as the other cytosol rows. Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt Peripheral membrane protein; Cytoplasmic side. Membrane, clathrin- |
| GO:0005829 cytosol | TAS Reactome:R-HSA-8951498 | KEEP AS NON CORE | Summary: Cytosol, cited to dissociation of the Arf1:GDP, AP-1 clathrin-coated nonameric complex. Reason: The dissociation step returns AP-1 to the cytosol. True, generic, non-core. Supporting Evidence: PMID:23247405 Upon vesicle budding, the clathrin coat and adaptors disassemble and return to the cytosolic pool so they can be reutilized in further rounds of traffic |
| GO:0005905 clathrin-coated pit | IEA GO_REF:0000044 | KEEP AS NON CORE | Summary: Clathrin-coated pit, mapped from UniProt's subcellular location keyword. Compatible with the GO definition, which covers TGN and endosomal pits as well as plasma-membrane ones, but with no direct evidence placing sigma1B in a pit. Reason: The obvious objection - that coated pits are an AP-2 plasma-membrane structure and an AP-1 subunit does not belong there - does not survive reading the term definition, which explicitly includes pits forming in the trans-Golgi network and on some endosomes. Electron microscopy of sigma1B-deficient synapses does show clathrin-coated buds on the accumulated endosomes, so coated structures at non-plasma-membrane sites are part of this biology. Kept, but non-core: the evidence is a keyword mapping, not an observation of AP1S2 in a pit. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB-SubCell:SL-0069 Β· UniProt subcellular location SL-0069 Clathrin-coated pit SUPPORTS TRANSFER SL-0069 maps one-to-one to GO:0005905, and P56377's SUBCELLULAR LOCATION line does carry 'Membrane, clathrin-coated pit'. The GO definition of GO:0005905 is not restricted to the plasma membrane - it states coated pits form on the plasma membrane, in the trans-Golgi network, and on some endosomes - so the term is compatible with an AP-1 subunit, which is why this is kept rather than removed. Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt Peripheral membrane protein; Cytoplasmic side. Membrane, clathrin- PMID:20203623 Endosome-like structures accumulating in the Ο1B-deficient synapses are coated with clathrin |
| GO:0006886 intracellular protein transport | IEA GO_REF:0000002 | ACCEPT | Summary: InterPro2GO mapping from the clathrin adaptor small-chain conserved site to intracellular protein transport. Correct, and at a usable level of specificity for this gene. Reason: Moving transmembrane cargo between the TGN, endosomes and lysosomes is intracellular protein transport, and it is what the AP-1 complex does. UniProt states the same: sigma1B is a subunit of a complex that plays a role in protein sorting in the late-Golgi/ trans-Golgi network and/or endosomes. This term is the informative member of the protein-transport pair on this gene and the broader GO:0015031 row is modified to point at it. Propagation Review Root cause: NO FAILURE CORE Sources checked: InterPro:IPR000804 Β· InterPro IPR000804 Clathrin adaptor complex, small chain SUPPORTS TRANSFER A conserved-site signature that AP1S2 genuinely matches (InterPro REST for P56377 returns IPR000804 among five entries). Its InterPro2GO mapping is GO:0006886, GO:0016192 and GO:0030117, all of which describe what an AP-complex small chain does. The mapping is family-wide and so says nothing about which AP complex or which compartment; that is a granularity limit, not an error. Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt plays a role in protein sorting in the late-Golgi/trans-Golgi network |
| GO:0015031 protein transport | IEA GO_REF:0000002 | MODIFY | Summary: The pan-sigma family signature maps to the broadest protein-transport term. True but strictly less informative than the intracellular protein transport row the same gene already carries. Reason: GO:0015031 covers directed movement of proteins into, out of or within a cell or between cells, which is broader than anything AP-1 does - the complex works entirely within the cell, at the TGN/endosome interface. The correct specific term, GO:0006886, is already present from a different InterPro signature, so this row should be raised to it rather than kept at the generic level. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: GRANULARITY MISMATCH Sources checked: InterPro:IPR016635 Β· InterPro IPR016635 Adaptor protein complex, sigma subunit SUPPORTS TRANSFER The pan-sigma family signature, matched by P56377, whose entire InterPro2GO mapping is the single term GO:0015031 protein transport. Correct but the broadest of the transport terms on this gene. Proposed replacements: intracellular protein transport Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt plays a role in protein sorting in the late-Golgi/trans-Golgi network |
| GO:0016192 vesicle-mediated transport | IBA GO_REF:0000033 | ACCEPT | Summary: The single phylogenetic annotation on this gene, placed at the PTHR11753 node PTN000204281 whose seeds are sigma subunits of AP-1, AP-2 and AP-3 from fungi, plants, amoebae, flies, worms and mammals. Vesicle-mediated transport is the correct last common ancestor term for that node and AP1S2 sits inside the clade that inherits it. Reason: The node placement is sound and the target is well inside it. The seeds disagree about which AP complex, which membrane and which destination is involved - AP-1 at the TGN, AP-2 at the plasma membrane, AP-3 toward lysosome-related organelles - so GO:0016192 is the correct LCA rather than an over-cautious parent; a more specific child would be wrong for most donors. Nine of the eleven gene-level donors carry their own experimental evidence for the term. The strongest is the direct mouse orthologue Ap1s2, which is 100% identical to human AP1S2 over all 157 human residues and carries IDA and IMP evidence for vesicle-mediated transport, so the inference is not merely deep-homology reasoning. There is no IRD or IKR anywhere in the family PAINT slice and no target-specific evidence of loss: human AP1S2 retains the full sigma-subunit fold, the dileucine-signal binding residues, and the gamma-adaptin interaction. Propagation Review Root cause: NO FAILURE CORE Sources checked: CGD:CAL0000182525 Β· APS3 (Candida albicans AP-3 sigma) SUPPORTS TRANSFER Resolves to UniProtKB:Q59QC5 (APS3_CANAL), IMP for GO:0016192 from PMID:20870878. QuickGO rejects the CGD identifier at the geneProductId endpoint, so the donor was traced through its UniProt xref. FB:FBgn0039132 Β· AP-1sigma (Drosophila) SUPPORTS TRANSFER The FlyBase gene resolves to four TrEMBL accessions (A0A0B4KHB7, B8A403, Q9VCF4, Q9GQM8) for one gene; those accessions carry GO:0016192 only as ISS/IEA, but the FlyBase gene-level record queried through the GO API carries an IMP from PMID:22389401, so the donor is experimentally grounded. FB:FBgn0043012 Β· AP-2sigma (Drosophila) SUPPORTS TRANSFER Resolves to UniProtKB:Q9VDC3 and Q9GQM7. GO:0016192 is supported by IMP PMID:20226669 on the UniProt-keyed record and by NAS PMID:11598180 at the FlyBase gene level. This donor appears in the GOA WITH/FROM but not in the seed column of the current PAINT slice; the slice row is dated 20260828 against the GOA row's 20250902, so the simplest reading is that the seed list was revised after the GOA snapshot. I read nothing further into the difference. MGI:MGI:1098244 Β· Ap1s1 (mouse sigma1A), the paralogous AP-1 sigma isoform SUPPORTS TRANSFER Resolves to UniProtKB:P61967 (AP1S1_MOUSE), which carries IMP for GO:0016192 from PMID:24928897 and TAS from PMID:9714600. A paralogue rather than an orthologue of the target, but an AP-1 sigma subunit, so it supports the AP-complex-wide term at this node. MGI:MGI:1889383 Β· Ap1s2 (mouse sigma1B), the direct orthologue of the target SUPPORTS TRANSFER Resolves by UniProt xref to UniProtKB:Q9DB50 (AP1S2_MOUSE). QuickGO returns IDA and IMP for GO:0016192 on this donor from PMID:20203623 plus an IMP from PMID:24928897. The mouse protein is 100.0% identical to human AP1S2 over all 157 human residues, differing only by a 3-residue insertion at mouse 143-145 (AP1S2-bioinformatics/RESULTS.md), so this donor grounds the transfer as directly as any donor can. PANTHER:PTN000204281 Β· PTHR11753 ADAPTOR COMPLEXES SMALL SUBUNIT FAMILY IBD node (seeded by AP-1, AP-2 and AP-3 sigma subunits of mouse, rat, human, fly, worm, fission and budding yeast and Candida) SUPPORTS TRANSFER The node, not a donor. The family PAINT slice (interpro/panther/PTHR11753/PTHR11753-paint.tsv) carries two IBD rows and both sit on this node; the GO:0016192 IBD lists ten gene-level seeds spanning AP-1, AP-2 and AP-3 sigma subunits from Candida, budding and fission yeast, fly, worm, rat, mouse and human. Because the seeds disagree on compartment and destination, GO:0016192 is the correct last common ancestor term for the node rather than an under-specific parent. AP1S2 is a PTHR11753 member (PTHR11753-entries.csv lists P56377, AP-1 complex subunit sigma-2, AP1S2, 157 aa) and so sits inside the clade that inherits it; there is no IRD or IKR anywhere in the slice. PomBase:SPAP27G11.06c Β· vas2 (fission yeast AP-1 sigma-1) SUPPORTS TRANSFER Resolves to UniProtKB:Q9P7N2 (AP1S1_SCHPO, gene vas2), with IDA for GO:0016192 from PMID:19624755. This is the reference protein of the repository's own PTHR11753 family review, which independently reaches the AP-1 Golgi/endosome assignment for it. RGD:620188 Β· Ap2s1 (rat sigma2) SUPPORTS TRANSFER Resolves to UniProtKB:P62744 (AP2S1_RAT), with IDA and EXP for GO:0016192 from PMID:17289840. Same subfamily as UniProtKB:P53680, so it adds species breadth rather than an independent activity. SGD:S000003561 Β· APS3 (budding yeast AP-3 sigma) SUPPORTS TRANSFER Resolves to UniProtKB:P47064 (AP3S_YEAST), IMP for GO:0016192 from PMID:9335339. An AP-3 subunit, so another donor constraining the node to the AP-complex-wide term. SGD:S000004160 Β· APS1 (budding yeast AP-1 sigma-1) SUPPORTS TRANSFER Resolves to UniProtKB:P35181 (AP1S1_YEAST), IMP for GO:0016192 from PMID:17003107. Places the assertion at least as deep as the fungal-metazoan split. UniProtKB:P53680 Β· AP2S1 (human sigma2) SUPPORTS TRANSFER Human AP-2 sigma subunit, with its own IDA and IMP for GO:0016192 from PMID:11102472. Being an AP-2 rather than an AP-1 subunit, it is one of the donors that forces the node term up to the AP-complex-wide parent. WB:WBGene00000157 Β· aps-2 (C. elegans AP-2 sigma) SOURCE WEAK OR INFERRED Neither a UniProt accession nor resolvable through the accession-keyed PTHR11753-entries.csv, and QuickGO rejects the WB prefix at the geneProductId endpoint; resolved through the GO API bioentity endpoint as aps-2. Its own GO:0016192 evidence is only IEA and IBA, so this donor carries no independent experimental weight. It does not weaken the node, which has nine other donors with experimental support. Supporting Evidence: PMID:20203623 Synaptic vesicle reformation in cultured neurons from Ο1B-deficient mice is reduced upon stimulation, and large endosomal intermediates accumulate. file:human/AP1S2/AP1S2-bioinformatics/RESULTS.md the mouse protein differs only by a 3-residue insertion at mouse positions 143-145 |
| GO:0016192 vesicle-mediated transport | IEA GO_REF:0000002 | ACCEPT | Summary: InterPro2GO route to the same term the IBA asserts, from two signatures AP1S2 genuinely matches. Reason: A duplicate of the IBA at the same granularity, reached independently from sequence signatures. Duplicates across evidence codes are not a problem, and the agreement between a family-signature route and a phylogenetic route is mildly reassuring rather than redundant noise. Propagation Review Root cause: NO FAILURE CORE Sources checked: InterPro:IPR000804 Β· InterPro IPR000804 Clathrin adaptor complex, small chain SUPPORTS TRANSFER Matched by P56377. Its InterPro2GO mapping includes GO:0016192, the same term the phylogenetic node asserts, so this row duplicates the IBA at the same granularity rather than adding anything. InterPro:IPR044733 Β· InterPro IPR044733 AP-1 complex subunit sigma SUPPORTS TRANSFER The AP-1-specific family signature, matched by P56377. Its InterPro2GO mapping is GO:0035615, GO:0016192 and GO:0030121. For GO:0016192 it is the more specific signature of the two cited on this row, and it agrees with the less specific one. Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt FUNCTION: Subunit of clathrin-associated adaptor protein complex 1 that |
| GO:0016192 vesicle-mediated transport | NAS PMID:23247405 Cell type-specific Rab32 and Rab38 cooperate with the ubiqui... | ACCEPT | Summary: Vesicle-mediated transport, projected from the ComplexPortal sigma1b AP-1 complex. The weakest of the three routes to this term on this gene, but the term itself is uncontroversial. Reason: The reference is a commentary whose annotations are spread across 48 gene products, but the claim - that an AP-1 complex participates in vesicle-mediated transport - is independently established by the IBA and by direct mouse experiments on sigma1B itself. There is no reason to contest the term because its citation is weak. Supporting Evidence: PMID:23247405 Upon vesicle budding, the clathrin coat and adaptors disassemble and return to the cytosolic pool so they can be reutilized in further rounds of traffic |
| GO:0030117 membrane coat | IEA GO_REF:0000002 | MODIFY | Summary: Membrane coat, from the small-chain conserved-site signature. True, but a strict ancestor of the AP-1 adaptor complex term this gene already carries. Reason: GO:0030117 is an ancestor of GO:0030121 in the ontology (QuickGO ancestor query on GO:0030121 returns GO:0030117), so this row adds nothing once the AP-1-specific assignment is in place, and the AP-1-specific assignment is well supported - human sigma1B assembles into AP-1 and into no other AP complex. Raising the row to GO:0030121 keeps the structural claim and loses the vagueness. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: GRANULARITY MISMATCH Sources checked: InterPro:IPR000804 Β· InterPro IPR000804 Clathrin adaptor complex, small chain SUPPORTS TRANSFER Matched by P56377; GO:0030117 membrane coat is one of the three terms in this signature's InterPro2GO mapping. AP-1 is a clathrin-plus-adaptor coat, which is explicitly within the GO:0030117 definition, so the parent term is true. It is two levels above what is known for this protein, which already has AP-type and AP-1 complex rows. Proposed replacements: AP-1 adaptor complex Supporting Evidence: PMID:21097499 all three Ο1-HA isoforms are incorporated into AP-1 complexes containing Ξ³1, Ξ²1, and ΞΌ1 subunits but not into complexes including the Ξ±C subunit of AP-2, the Ξ²3A subunit of AP-3, or the Ο΅ subunit of AP-4 |
| GO:0030119 AP-type membrane coat adaptor complex | TAS PMID:9733768 Identification and characterization of novel clathrin adapto... | MODIFY | Summary: The paper that named sigma1B, cited for membership of an AP-type complex. Its evidence is in fact AP-1-specific - yeast two-hybrid interaction with the gamma1 and gamma2 adaptins - so the parent term understates what the reference shows. Reason: Takatsu et al. found sigma1B binding gamma1-adaptin, the AP-1 large subunit, and it is now established that both gamma isoforms assemble into AP-1 complexes and that sigma1B is incorporated into AP-1 and not into AP-2, AP-3 or AP-4. GO:0030119 is a proper ancestor of GO:0030121 in the ontology, so the AP-1-specific child is both available and better supported. This is a generality correction, not a disagreement with the curator. Proposed replacements: AP-1 adaptor complex Supporting Evidence: PMID:9733768 a novel sigma1-like protein, designated as sigma1B, which shows an 87% amino acid identity to sigma1A PMID:21097499 the Ξ³1 isoform was incorporated into AP-1 complexes containing either Ο1A, Ο1B, or Ο1C, the Ξ³2 isoform was incorporated into AP-1 complexes containing Ο1A or Ο1B, but not Ο1C |
| GO:0030121 AP-1 adaptor complex | IEA GO_REF:0000002 | ACCEPT | Summary: AP-1 adaptor complex membership from the AP-1-specific sigma family signature. This is the single most important annotation on the gene and it is confirmed experimentally in human cells. Reason: The signature assignment is correct and independently verified: HA-tagged human sigma1B co-precipitates gamma1, beta1 and mu1 and no AP-2, AP-3 or AP-4 subunit, and it pairs with either gamma isoform. The GO term definition explicitly anticipates this gene, naming sigma1A, sigma1B and sigma1C as alternative subunits of a heterogeneric AP-1. Core. Propagation Review Root cause: NO FAILURE CORE Sources checked: InterPro:IPR044733 Β· InterPro IPR044733 AP-1 complex subunit sigma SUPPORTS TRANSFER The AP-1 sigma family signature, matched by P56377 and mapped by InterPro2GO to GO:0030121. This is the correct AP-1-specific assignment and it is confirmed experimentally in human cells by the sigma1B-HA immunoprecipitation-recapture experiment in PMID:21097499, which recovers gamma1, beta1 and mu1 but no AP-2, AP-3 or AP-4 subunit. Supporting Evidence: PMID:21097499 all three Ο1-HA isoforms are incorporated into AP-1 complexes containing Ξ³1, Ξ²1, and ΞΌ1 subunits but not into complexes including the Ξ±C subunit of AP-2, the Ξ²3A subunit of AP-3, or the Ο΅ subunit of AP-4 file:human/AP1S2/AP1S2-uniprot.txt a medium adaptin (mu-type subunit AP1M1 or AP1M2) and a small |
| GO:0030121 AP-1 adaptor complex | NAS PMID:15377783 Crystal structure of the clathrin adaptor protein 1 core. | ACCEPT | Summary: ComplexPortal projection from the AP-1 core crystal structure. The crystallised small chain was sigma1A rather than sigma1B, but the conclusion transfers on independent grounds. Reason: The structure paper establishes the architecture of the AP-1 core - two large-chain N-terminal fragments plus intact mu1 and sigma1 - and the sigma slot it describes is the slot AP1S2 occupies. The isoform actually crystallised is sigma1A, so this reference is complex-level rather than sigma1B-specific; the AP-1 membership of sigma1B itself rests on the immunoprecipitation evidence in PMID:21097499 and the yeast two-hybrid evidence in PMID:9733768. Term correct, citation generic. Supporting Evidence: PMID:15377783 The AP-1 core comprises N-terminal fragments of the two large chains, beta1 and gamma, and the intact medium and small chains, micro1 and sigma1. PMID:21097499 all three Ο1-HA isoforms are incorporated into AP-1 complexes containing Ξ³1, Ξ²1, and ΞΌ1 subunits but not into complexes including the Ξ±C subunit of AP-2, the Ξ²3A subunit of AP-3, or the Ο΅ subunit of AP-4 |
| GO:0030659 cytoplasmic vesicle membrane | IEA GO_REF:0000044 | ACCEPT | Summary: Cytoplasmic vesicle membrane, mapped from UniProt's subcellular location. This is the membrane an AP-1 coat sits on once a vesicle has formed. Reason: The mapping is one-to-one and the underlying UniProt statement is specific about the topology: peripheral membrane protein on the cytoplasmic side of cytoplasmic vesicle membranes, which is exactly a coat subunit. Core. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB-SubCell:SL-0089 Β· UniProt subcellular location SL-0089 Cytoplasmic vesicle membrane SUPPORTS TRANSFER SL-0089 maps one-to-one to GO:0030659, and P56377's SUBCELLULAR LOCATION line carries 'Cytoplasmic vesicle membrane; Peripheral membrane protein; Cytoplasmic side', which is exactly what an AP-1 coat subunit is. Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt Peripheral membrane protein; Cytoplasmic side. Membrane, clathrin- |
| GO:0030659 cytoplasmic vesicle membrane | TAS Reactome:R-HSA-421835 | ACCEPT | Summary: Cytoplasmic vesicle membrane during TGN vesicle scission, when the coated vesicle separates from the donor membrane. Reason: The coat is on the newly formed vesicle at this step, so the location is exactly right, and it agrees with the independent UniProt-derived row. Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt Note=Component of the coat surrounding the cytoplasmic face |
| GO:0030659 cytoplasmic vesicle membrane | TAS Reactome:R-HSA-421836 | ACCEPT | Summary: Cytoplasmic vesicle membrane during uncoating of the TGN-derived vesicle. Reason: Same membrane, at the moment the coat leaves it. Correct and core, for the same reason as the scission row. Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt Note=Component of the coat surrounding the cytoplasmic face |
| GO:0032588 trans-Golgi network membrane | NAS PMID:15377783 Crystal structure of the clathrin adaptor protein 1 core. | ACCEPT | Summary: TGN membrane, projected from the AP-1 core structure paper through the ComplexPortal sigma1b complex. The term is correct and specific; the citation is complex-level. Reason: The TGN is where AP-1 is recruited, and the structure paper says so directly: it reports the core of the complex which functions in the trans-Golgi network, and shows that TGN recruitment depends on Arf1 and PI-4-P. This is the most specific and most informative of the Golgi-related locations on the gene. Supporting Evidence: PMID:15377783 We report the crystal structure of the core of the AP-1 complex, which functions in the trans-Golgi network (TGN). file:human/AP1S2/AP1S2-uniprot.txt plays a role in protein sorting in the late-Golgi/trans-Golgi network |
| GO:0032588 trans-Golgi network membrane | TAS Reactome:R-HSA-2130641 | ACCEPT | Summary: TGN membrane, cited to translocation of the TGN-lysosome vesicle. Reason: The TGN membrane is the AP-1 recruitment platform and the origin of the vesicles this reaction follows. Core location. Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt plays a role in protein sorting in the late-Golgi/trans-Golgi network |
| GO:0032588 trans-Golgi network membrane | TAS Reactome:R-HSA-2213236 | ACCEPT | Summary: TGN membrane, cited to uncoating of the TGN-lysosome vesicle. Reason: Same location, reached through a different reaction in the same pathway. Duplicates of a correct specific location are harmless. Supporting Evidence: file:human/AP1S2/AP1S2-uniprot.txt plays a role in protein sorting in the late-Golgi/trans-Golgi network |
| GO:0032588 trans-Golgi network membrane | TAS Reactome:R-HSA-5333658 | ACCEPT | Summary: TGN membrane, cited to the reaction in which the clathrin:AP-1:CLVS complex binds PI(3,5)P2. Reason: Phosphoinositide-dependent recruitment to TGN and endosomal membranes is how AP-1 gets onto the membrane in the first place, and the structural work on the AP-1 core showed the gamma-chain corner responsible for the PI-4-P-dependent half of that recruitment. The location is correct; the lipid specificity in this particular Reactome reaction is a property of the complex and its partners, not of the sigma subunit. Supporting Evidence: PMID:15377783 TGN localization of AP-1 depends on the small GTPase, Arf1, and the phosphoinositide, PI-4-P. |
| GO:0032588 trans-Golgi network membrane | TAS Reactome:R-HSA-8951498 | ACCEPT | Summary: TGN membrane, cited to dissociation of the Arf1:GDP, AP-1 clathrin-coated nonameric complex. Reason: Arf1-GTP hydrolysis releases AP-1 from the TGN membrane, so the complex is there immediately before this reaction. Correct and core. Supporting Evidence: PMID:15377783 TGN localization of AP-1 depends on the small GTPase, Arf1, and the phosphoinositide, PI-4-P. |
| GO:0035615 clathrin-cargo adaptor activity | IEA GO_REF:0000002 | ACCEPT | Summary: The only molecular function term on the gene. AP1S2 does not have this activity on its own - it is one of two subunits that jointly form the cargo-signal binding site - but the activity is genuinely executed by the complex it belongs to, and AP1S2 retains all five sigma1A positions PMID:21097499 names in its text, including all three whose substitution abolishes binding. Reason: Direct evidence exists for the sigma1B-containing hemicomplex: gamma1-sigma1B binds the Nef, tyrosinase and LIMP-II (D/E)XXXL(L/I) signals with avidity comparable to gamma1-sigma1A, and the gamma2-containing variant binds a subset of them. The three sigma-subunit substitutions reported to abolish that binding in sigma1A - V88D and I103S for the Nef signal, A63D for the tyrosinase signal - all map to residues AP1S2 retains, at V87, I102 and A62, and the weaker R15 and L101 positions are retained too, at R14 and L100. Tested here rather than assumed. The one caveat worth recording is that the GO definition of this term ends with responsibility for the formation of endocytic vesicles, which fits AP-2 better than AP-1; AP-1 forms TGN- and endosome-derived vesicles, not endocytic ones. That is a limitation of the term definition rather than of this annotation, and the missing subunit-level alternative is recorded under proposed_new_terms. Graded as a contributes_to-level activity in core_functions. Propagation Review Root cause: NO FAILURE CORE Sources checked: InterPro:IPR044733 Β· InterPro IPR044733 AP-1 complex subunit sigma SUPPORTS TRANSFER Matched by P56377 and mapped by InterPro2GO to GO:0035615 clathrin-cargo adaptor activity. The signature-based transfer is corroborated at the residue level: all five sigma1A positions PMID:21097499 names in its text are retained in AP1S2 (sigma1A R15, A63, V88, L101, I103 -> AP1S2 R14, A62, V87, L100, I102), including all three whose substitution abolishes the gamma1-sigma1A interaction outright. The paper also tested the sigma1A counterparts of the sigma2 residues wholesale without naming each outcome; of those, the two positions where AP1S2 differs from sigma2 (sigma2 N92 -> AP1S2 D91 and sigma2 L103 -> AP1S2 I102) are both states sigma1A shares, so neither is a sigma1B-specific loss. This is not a fold-without-function assignment. Supporting Evidence: PMID:21097499 all AP-1 hemicomplexes containing Ξ³1 (Ξ³1-Ο1A, Ξ³1-Ο1B, and Ξ³1-Ο1C) interact with similar avidities with the Nef, tyrosinase, and LIMP-II signals PMID:21097499 the loss of signal binding by the Ο2 V88D or L103S substitutions and the homologous Ο1A V88D and I103S and Ο3A V94D and L109S substitutions PMID:21097499 the interaction with Ο1A was abolished only by V88D and I103S (for Nef) and also by A63D (for tyrosinase) PMID:21097499 the N92A and L101A mutations had no effect on the interaction with the Nef signal but decreased the interaction with the tyrosinase signal file:human/AP1S2/AP1S2-bioinformatics/RESULTS.md There is therefore no residue-level basis for arguing that sigma1B has lost the cargo-signal binding site |
| GO:0060155 platelet dense granule organization | NAS PMID:23247405 Cell type-specific Rab32 and Rab38 cooperate with the ubiqui... | MARK AS OVER ANNOTATED | Summary: ComplexPortal projection of a speculative sentence. The cited commentary contains no platelet experiment, no AP-1 platelet experiment, and no sigma-isoform resolution; its only platelet statement is an explicit prediction about Rab38. Reason: Reading the source settles this. Its platelet content is one forward-looking sentence saying it is likely that Rab38 and the ubiquitous machinery also work in platelet dense granules and lamellar bodies - a hypothesis about a Rab GTPase, not a result about AP-1. The paper's own AP-1 depletion experiment is negative for the mechanism it is cited for. Independently, the human lysosome-related-organelle disorders arise from AP-3, BLOC-1, BLOC-2 and BLOC-3 mutations, not AP-1, and sigma1B is a tissue-restricted isoform with no reported megakaryocyte biology. The reference-projection test makes the mechanism plain: this one commentary generates 178 annotations across 48 distinct gene products from only five distinct terms. Marked as over-annotated rather than removed because the AP-1 complex does participate in cargo traffic to lysosome-related organelles in general, so the term is not flatly contradicted - it is unsupported for this subunit. Supporting Evidence: PMID:23247405 it is likely that the cooperation between Rab38 and the ubiquitous transport machinery uncovered in melanocytes also functions in the biogenesis of other LROs, such as platelet dense granules and lamellar bodies PMID:23247405 Depletion of AP-1 has no effect on either Rab32 or Rab38 membrane association. |
| GO:1903232 melanosome assembly | NAS PMID:23247405 Cell type-specific Rab32 and Rab38 cooperate with the ubiqui... | MARK AS OVER ANNOTATED | Summary: Melanosome assembly, projected from the same commentary. AP-1 involvement in melanosomal cargo traffic is real, but the experiments used pan-AP-1 reagents in melanocytes and say nothing about which sigma isoform is present. Reason: The source's melanocyte experiments use antibodies and RNAi against AP-1 as a whole, so they cannot distinguish an AP-1/sigma1A complex from an AP-1/sigma1B complex - and the distinction matters here, because sigma1B is expressed in a tissue-restricted pattern while sigma1A is ubiquitous, and patient fibroblasts lacking sigma1B show no AP-1 defect at all. Assigning melanosome assembly to the sigma1B variant specifically is the ComplexPortal complex object doing work the evidence does not support. The term is not flatly wrong for AP-1, which is why this is marked over-annotated rather than removed; it is unsupported for this isoform. Supporting Evidence: PMID:23247405 Confocal immunofluorescence microscopy experiments show that a significant percentage of structures labeled with endogenous AP-1, AP-3 and BLOC-2 co-localize with endogenous Rab32 and Rab38 in many locations throughout MNT-1 melanocytes. PMID:20203623 Expression levels of Ο1B and Ο1C are highly variable and tissue specific, with most tissues expressing Ο1A and only one of the other isoforms PMID:18428203 no major alteration of the stability, subcellular localization, and function of the AP-1 complex was observed in fibroblasts derived from a patient carrying an AP1S2 mutation |
| GO:0036466 synaptic vesicle recycling via endosome | ISS PMID:20203623 AP-1/sigma1B-adaptin mediates endosomal synaptic vesicle rec... | NEW | Summary: The central, isoform-specific function of sigma1B, missing entirely from GOA. Loss of sigma1B in mouse hippocampal neurons slows and truncates the reformation of synaptic vesicles through an endosomal intermediate. Reason: Three independent lines from the same laboratory converge: kinetically, only 46.5% of exocytosed vesicles become re-available within 10 s in sigma1B-deficient neurons against 94% in isogenic controls, and reformation plateaus at 70%; ultrastructurally, large endosomal intermediates accumulate in the terminals; and biochemically, those intermediates are classic PI-3-P-rich early endosomes carrying a stored pool of synaptic-vesicle proteins. That is precisely what GO:0036466 describes. Coded ISS rather than IMP because every perturbation was done in mouse; the donor is justified by the sequence identity, which is complete over the human protein. GOA currently records nothing about this gene's neuronal function at all, which is the largest single gap in the annotation set. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q9DB50 Β· Ap1s2 (mouse sigma1B) SUPPORTS TRANSFER The mouse orthologue, 100.0% identical to human AP1S2 over all 157 human residues (AP1S2-bioinformatics/RESULTS.md). Its knockout gives the phenotype this term describes: reduced synaptic vesicle reformation on stimulation with accumulation of large endosomal intermediates (PMID:20203623), and the accumulating compartment is characterised as a classic PI-3-P-rich early endosome (PMID:25128028). Coded ISS rather than IMP because the perturbation was done in mouse. Supporting Evidence: PMID:20203623 Within the first 10 s after depleting stimulation, only 46.5% of the exocytosed vesicles become re-available for release in Ο1B-deficient neurons compared with 94% in the isogenic controls. PMID:20203623 Synaptic vesicle reformation in cultured neurons from Ο1B-deficient mice is reduced upon stimulation, and large endosomal intermediates accumulate. PMID:25128028 The accumulation of early endosomes and endocytotic AP-2 CCV indicates the regulation of SV recycling via early endosomes file:human/AP1S2/AP1S2-bioinformatics/RESULTS.md the mouse protein differs only by a 3-residue insertion at mouse positions 143-145 |
| GO:0016182 synaptic vesicle budding from endosome | ISS PMID:20203623 AP-1/sigma1B-adaptin mediates endosomal synaptic vesicle rec... | NEW | Summary: The specific step within endosomal synaptic-vesicle recycling that fails when sigma1B is absent: vesicles bud less efficiently from the early endosome, leaving clathrin-budding endosomal profiles behind. Reason: Distinct from the recycling row above because it is supported by a different kind of observation. Electron microscopy and three-dimensional reconstruction of sigma1B-deficient synapses show endosome-like structures decorated with clathrin buds, and the vesicle deficit is quantified directly - resting synapses contain about 135 vesicles per square micrometre against about 230 in controls, and docked vesicles fall by 60% after stimulation. The authors exclude a defect earlier in the secretory pathway, noting that targeting and docking of the vesicles that do form is efficient. Coded ISS from the identical mouse orthologue. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q9DB50 Β· Ap1s2 (mouse sigma1B) SUPPORTS TRANSFER Same donor and same sequence identity. This term covers the budding step specifically, which is what the electron-microscopy arm of PMID:20203623 shows: clathrin-coated buds on the endosome-like structures that accumulate when sigma1B is absent, together with a 60% drop in docked vesicles after stimulation. Coded ISS because the evidence is mouse. Supporting Evidence: PMID:20203623 The analysis of EM images revealed clathrin-coated buds on the accumulated endosome-like structures. PMID:20203623 This result cannot be reconciled with a defect in SV biogenesis earlier in the secretory pathway at the TGN. |
| GO:2000642 negative regulation of early endosome to late endosome transport | ISS PMID:27411398 AP-1/Ο1A and AP-1/Ο1B adaptor-proteins differentially regula... | NEW | Summary: The one activity that is private to sigma1B rather than shared with sigma1A. By binding Rabex-5, sigma1B keeps it out of the sigma1A-ArfGAP1-Rabex-5 complex, lowering endosomal Rabex-5 and damping Rab5/Vps34-driven maturation of early endosomes into multivesicular late endosomes. Reason: The sign of the regulation is established directly. In gamma1/sigma1 hemicomplex yeast-three-hybrid assays Rabex-5 binds sigma1B and not sigma1A, and the same is true of the RabGAP5 RUN domain; the authors show that this binding prevents formation of the tripartite activating complex; and proximity-ligation assays in sigma1B-null cells show more AP-1/Rabex-5 and more AP-1/Rab5 colocalisation than in wild type, i.e. removing sigma1B releases the brake. Because the family contains both an activating isoform and an inhibiting one, the sign matters: an unsigned regulation term would blur exactly the distinction the paper draws. Coded ISS from the identical mouse orthologue. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q9DB50 Β· Ap1s2 (mouse sigma1B) SUPPORTS TRANSFER Same donor. PMID:27411398 shows the sign of the regulation directly: sigma1B binds Rabex-5 in gamma1/sigma1 hemicomplex yeast-three-hybrid assays where sigma1A does not, this prevents assembly of the AP-1/sigma1A-ArfGAP1-Rabex-5 complex, and sigma1B-deficient cells show more AP-1/Rabex-5 and more AP-1/Rab5 colocalisation. Coded ISS because the experiments are mouse knockout cells and mouse constructs. Supporting Evidence: PMID:27411398 Unexpectedly, Rabex-5 binds Ο1B, not Ο1A PMID:27411398 Formation of AP-1/Ο1A-ArfGAP1-Rabex-5 complexes is prevented by Ο1B binding of Rabex-5 and the amount of endosomal Rabex-5 is reduced. PMID:27411398 AP-1/Ο1A and AP-1/Ο1B regulate maturation of these early endosomes into multivesicular body late endosomes |
| GO:0061462 protein localization to lysosome | ISS PMID:24928897 Ο1B adaptin regulates adipogenesis by mediating the sorting ... | NEW | Summary: The one well-characterised non-neuronal cargo route for sigma1B: delivery of the sorting receptor sortilin to lysosomes in adipocytes, mediated by a dileucine-type motif in sortilin's tail. Reason: The cargo determinant is defined and it is of exactly the class the gamma1/sigma1B hemicomplex recognises: sigma1B-specific binding of sortilin requires the sortilin DxxD-x12-DSxxxL motif. The functional consequence is a reduced lysosomal fraction of sortilin, with downstream stabilisation of DLK1 and inhibition of adipogenesis. Two caveats are stated rather than hidden: the defect is partial - the authors explicitly exclude a block at any single organelle - and it is tissue-restricted, since sortilin and DLK1 levels are unchanged in the brain of the same animals despite brain having the highest sigma1B expression. Coded ISS from the identical mouse orthologue. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q9DB50 Β· Ap1s2 (mouse sigma1B) SUPPORTS TRANSFER Same donor. PMID:24928897 shows sigma1B binds sortilin through the sortilin DxxD-x12-DSxxxL motif - a dileucine-type signal of the class the gamma1/sigma1B hemicomplex recognises - and that in sigma1B-deficient adipocytes the fraction of sortilin reaching lysosomes is reduced. Coded ISS because the work is in mouse adipose tissue; the effect is partial and tissue-restricted, which is stated in the review reason. Supporting Evidence: PMID:24928897 Ο1B-specific binding of sortilin requires the sortilin DxxD-x12-DSxxxL motif. PMID:24928897 Ο1B deficiency does not lead to a block of sortilin transport out of a specific organelle, but the fraction that reaches lysosomes is reduced. PMID:24928897 DLK1 and sortilin expression are not increased in the brain tissue of Ο1B(-/-) mice, although this is the tissue with the highest expression of Ο1B and sortilin. |
Loading supporting contentβ¦
Download this section (compressed HTML)Q: Does any AP-1/sigma1B complex in the human brain bind a synaptic-vesicle protein through a (D/E)XXXL(L/I) motif, and is that motif required for the protein's recycling?
Suggested experts: Peter Schu, Juan S. Bonifacino
Q: Can sigma1A overexpression rescue the synaptic-vesicle recycling defect of sigma1B-null neurons, which would show the isoforms differ only in expression, or does it fail, which would show they differ in specificity?
Suggested experts: Peter Schu
Q: Do the human AP1S2 variants that truncate the protein and the one splice variant that shortens it without abolishing it (c.426+1G>T, losing 46 residues) produce the same cellular defect, or is there a residual-function allele series?
Suggested experts: Laurent Villard, F. Lucy Raymond
Q: Is the Rabex-5 interaction detectable with the human proteins, and does it require the sigma1 C-terminal extension that distinguishes sigma1 isoforms from sigma2?
Suggested experts: Peter Schu, Dan Cassel
Q: Why do patient fibroblasts show no AP-1 defect while patient brains show a developmental malformation - is the requirement developmental-stage-specific rather than cell-type-specific?
Suggested experts: Guntram Borck, Laurence Colleaux
Experiment: Reconstitute human gamma1/sigma1A and gamma1/sigma1B hemicomplexes and compare their binding to a panel of synaptic-vesicle protein cytosolic tails, to find neuronal cargo that discriminates between the isoforms the way sortilin does in adipocytes.
Hypothesis: Neuronal synaptic-vesicle proteins carrying (D/E)XXXL(L/I) motifs are selected preferentially by gamma1/sigma1B rather than gamma1/sigma1A hemicomplexes.
Type: yeast three-hybrid and in vitro binding
Experiment: Generate isogenic AP1S1 and AP1S2 knockouts in human iPSC-derived neurons and compare synaptic-vesicle recycling kinetics, endosomal proteome and Rabex-5 distribution, so the two isoforms are tested in the same human background.
Hypothesis: The sigma1A and sigma1B isoforms are functionally non-equivalent in human neurons, not merely differently expressed.
Type: isogenic knockout plus synaptopHluorin imaging and proteomics
Experiment: Make a zebrafish ap1s2 loss-of-function line and quantify ventricular volume, hindbrain morphology and basal ganglia mineralisation, to test whether the human brain malformations can be modelled at all.
Hypothesis: Loss of ap1s2 disrupts ventricular and hindbrain development, accounting for the hydrocephalus and Dandy-Walker malformation seen in patients.
Type: zebrafish mutant phenotyping
Experiment: Solve or predict the structure of a gamma1/sigma1B hemicomplex bound to a dileucine signal and compare the pocket with the gamma1/sigma1A structure.
Hypothesis: The R14/V87/L100/I102 residues identified here as retained form the same dileucine-signal pocket in sigma1B as their counterparts do in sigma1A.
Type: structural biology
Experiment: Perform isoform-resolved targeted proteomics of AP-1 immunoprecipitates from melanocytes and megakaryocytes to determine whether sigma1B is present at all in those complexes.
Hypothesis: sigma1B is absent from the AP-1 complexes that operate in melanocytes and megakaryocytes, so the melanosome and platelet dense granule annotations do not apply to it.
Type: targeted proteomics
What is not known β curated, literature-grounded statements of the open unknowns (the inverse of core functions).
Gap: No AP1S2-specific animal model exists outside the mouse. In particular there is no zebrafish ap1s2 mutant or morphant in the published literature, although zebrafish models exist for the AP-1 large subunit ap1g1, so the hydrocephalus and Dandy-Walker features of the human syndrome have no developmental model.
OPEN BIOLOGY BP_DARK
What is known: The mouse sigma1B knockout is viable and fertile and reproduces the cognitive phenotype - reduced motor coordination and severely impaired long-term spatial memory - but it has not been reported to have hydrocephalus, Dandy-Walker malformation or basal ganglia mineral deposition, which are the features that define the human syndrome radiologically.
Significance: The brain malformations are what make the human disorder recognisable, and no model currently explains them.
What would resolve it: A zebrafish ap1s2 loss-of-function line assayed for ventricular volume and hindbrain morphology, or a conditional mouse allele in neuroepithelium and ependyma.
Provenance (the field's own admissions):
Gap: How loss of sigma1B produces a brain-restricted phenotype is unexplained at the mechanistic level. Patient fibroblasts show no AP-1 defect at all, so the disease mechanism cannot be a general failure of AP-1 assembly, but the specific brain process that cannot be covered by sigma1A or sigma1C has not been identified.
OPEN BIOLOGY RESIDUAL_SUBGAP
What is known: It is established that sigma subunits are required for AP complex stability in general, that sigma1B is dispensable for the housekeeping functions of AP-1, and that sigma1A and sigma1B are both highly expressed in brain. The redundancy argument is therefore a reasonable hypothesis rather than a demonstrated mechanism.
Significance: This is the gap between a well-described molecular activity and a well-described clinical syndrome, and closing it would say which neuronal cargo or which developmental window is non-redundant.
What would resolve it: Compare the endosomal and synaptic-vesicle proteomes of sigma1A- and sigma1B-depleted human neurons from the same isogenic background, and test whether sigma1A overexpression rescues the sigma1B-null recycling defect.
Provenance (the field's own admissions):
Gap: Whether AP1S2 has any postsynaptic role is untested. No study has examined dendritic spine number or morphology in sigma1B-deficient neurons, so the possibility that the cognitive phenotype has a postsynaptic component is open rather than excluded.
OPEN BIOLOGY CC_DARK
What is known: The one relevant observation is negative and localisational: AP-1 shows little or no colocalisation with the postsynaptic marker Homer in either wild-type or sigma1B-deficient boutons, supporting a predominantly presynaptic distribution. That is a statement about where the complex is, not about whether spines are normal.
What would resolve it: Spine density and morphology quantification in sigma1B-null hippocampal neurons, with electrophysiology separating pre- from postsynaptic contributions.
Provenance (the field's own admissions):
Gap: The functional meaning of the MAB21L2 interaction is unknown. It is the only non-AP-1 partner in the curated interaction set for human AP1S2 and no study has tested whether it reflects cargo selection, a regulatory interaction, or an assay artefact.
OPEN BIOLOGY MF_DARK
What is known: The interaction is curated by UniProt with five IntAct experiments from a binary yeast-two-hybrid reference map, so it is reproducible within that format; what is missing is any orthogonal validation or functional test.
What would resolve it: Co-immunoprecipitation from a tissue expressing both proteins, and a test of whether MAB21L2 carries a dileucine-type motif that the gamma1/sigma1B hemicomplex binds.
Provenance (the field's own admissions):
Gap: Which sigma1 isoform operates in the melanocyte and platelet AP-1 complexes has never been determined, which is why the melanosome and platelet dense granule annotations on this gene cannot be evaluated on their merits.
OPEN BIOLOGYCURATION BP_DARK
What is known: AP-1 as a complex does participate in cargo traffic from endosomal tubules to maturing melanosomes, and the reagents used were pan-AP-1. Separately, sigma1B expression is tissue-restricted rather than ubiquitous.
Significance: Complex-level annotation projected onto isoform-specific complex objects will keep generating unsupported subunit annotations until the isoform composition of tissue-specific AP-1 complexes is measured.
What would resolve it: Isoform-resolved immunoprecipitation or targeted proteomics of AP-1 from melanocytes and megakaryocytes.
Provenance (the field's own admissions):
Loading supporting contentβ¦
Download this section (compressed HTML)Loading supporting contentβ¦
Download this section (compressed HTML)Loading supporting contentβ¦
Download this section (compressed HTML)Loading supporting contentβ¦
Download this section (compressed HTML)