AP3S1 encodes sigma3A, the small subunit of adaptor protein complex 3 (AP-3), the heterotetrameric coat adaptor that selects transmembrane cargo out of tubular sorting endosomes and delivers it to lysosomes and to lysosome-related organelles such as melanosomes, platelet dense granules and synaptic vesicles. AP-3 is built from two large chains (delta, encoded by AP3D1, plus beta3A or the neuron-enriched beta3B), a medium chain (mu3A or mu3B) and a small chain (sigma3A or sigma3B), so a cell can assemble several isoform combinations; sigma3A and sigma3B are both expressed in a wide range of tissues. The complex is recruited to membranes by Arf1-GTP through its delta subunit, associates with clathrin through the beta3 appendage, and is released again after the vesicle buds. Within the complex, sigma3A is not an inert filler. Cargo is selected through two classes of cytosolic sorting signal that bind different parts of AP-3: tyrosine-based YxxO motifs bind the mu3 subunit, while acidic dileucine motifs bind no single subunit but a composite site formed by the delta and sigma3 chains together. sigma3A supplies the hydrophobic pockets of that site and one of the two basic residues that grip the acidic position of the motif; substituting the pocket residues abolishes signal binding, and sigma3A is one of the four chains of the solved human AP-3 core, where the same pocket is found occupied by the N-terminal extension of beta3 in the cargo-free state. sigma3A pairs with delta as a stable hemicomplex that survives loss of the other two subunits. Cargoes routed this way include the lysosomal membrane proteins LAMP-1, LAMP-2, CD63 and LIMP-II, the melanosomal enzymes tyrosinase and OCA2, the zinc transporter ZnT3 and the lipid kinase PI4KIIalpha; HIV-1 Nef subverts the same site. Loss of AP-3 function is a disease mechanism in its own right. Mutations in the beta3A subunit cause Hermansky-Pudlak syndrome type 2, with oculocutaneous albinism, a bleeding diathesis from platelet dense-granule deficiency and immune defects, and the mouse mutants mocha and pearl, null for the delta and beta3A subunits, combine coat-colour dilution, platelet storage-pool deficiency and neurological deficits. No human disease has yet been attributed to AP3S1 itself, and the functional division of labour between sigma3A and sigma3B remains unresolved.
Definition: Binding to an acidic dileucine sorting signal, a short peptide motif of the form acidic residue - X - X - X - Leu - Leu/Ile in the cytosolic tail of a transmembrane protein, which directs that protein into coated carriers at the trans-Golgi network, endosomes or the plasma membrane.
Justification: GO expresses the tyrosine-based branch of adaptor cargo recognition through GO:0089710 endocytic targeting sequence binding, whose definition requires an essential tyrosine and internalization by clathrin-coated pits, but it has no counterpart for the dileucine branch. Checked independently for this review: a QuickGO ontology search for 'dileucine' returns zero hits; searches for 'sorting signal binding' and 'targeting sequence binding' return only the tyrosine, signal- and transit-peptide receptor terms; and GO:0140312 cargo adaptor activity has exactly one child, GO:0035615. The consequence is that a well-characterised, subunit-attributable activity -- demonstrated for sigma3A by yeast three-hybrid, GST pull-down and point mutagenesis (V94D, L109S) across four independent signals -- can currently only be annotated through a complex-level coat-binding term. The term would apply to the sigma and large subunits of AP-1, AP-2 and AP-3 across eukaryotes, with contributes_to, since the site straddles the two subunits. On the choice of parent: GO:0005515 protein binding is normally a term to avoid and is used here deliberately, because GO:0089710 -- the tyrosine counterpart this proposal mirrors -- is itself is_a GO:0005515 (is_a ancestors GO:0005515, GO:0005488, GO:0003674, checked on QuickGO), so mirroring it places the two motif-binding siblings in one genus. The same term was proposed independently in this repository's AP1S1 review; this is the AP-3 half of the same gap, not a second request.
Parent term: protein binding
Supporting Evidence:
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0005515 protein binding | IPI PMID:16162817 Functions of adaptor protein (AP)-3 and AP-1 in tyrosinase s... | MODIFY | Summary: Bare 'protein binding' recording the interaction with UniProtKB:P14679, human tyrosinase. The partner resolves, and the interaction has a specific, well-characterised molecular content that the term discards. Reason: P14679 is TYR, human tyrosinase (UniProt, checked). What the cited paper shows is not a generic protein-protein contact but cargo-signal recognition: the acidic dileucine motif in the tyrosinase cytosolic tail is bound by the delta-sigma3 hemicomplex of AP-3, in a yeast three-hybrid assay that is negative for the corresponding AP-2 and AP-4 hemicomplexes, and GST pull-downs show the same interaction with native AP-3 and its dependence on the two leucines of the motif. 'protein binding' is exactly the term the curation guidelines say to replace, and there is a real molecular function underneath it. The replacement is GO:0140312 cargo adaptor activity rather than its only child GO:0035615 clathrin-cargo adaptor activity, because GO:0035615 is defined as 'Bringing together a cargo protein with clathrin, responsible for the formation of endocytic vesicles' and AP-3 is not an endocytic coat -- it acts on exit from tubular sorting endosomes. I checked the QuickGO children endpoint: GO:0035615 is GO:0140312's only child, so the parent is the most specific term in that lineage that does not assert endocytosis. The activity belongs to the hemicomplex rather than to sigma3A alone -- these signals bind no single AP subunit -- but the pockets that accept the leucines are on sigma3A itself, and substituting them (V94D, L109S) abolishes binding, which is why the matching core function uses contributes_to_molecular_function. GO has no term for the precise activity, dileucine sorting-signal binding; that is recorded as a proposed new term and an ONTOLOGY knowledge gap. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: GRANULARITY MISMATCH Sources checked: UniProtKB:P14679 · TYR, Homo sapiens tyrosinase, the interaction partner SUPPORTS TRANSFER The interaction partner, resolved in UniProt to TYRO_HUMAN, tyrosinase, gene TYR, Swiss-Prot, 529 aa. It is not an incidental interactor: its cytosolic tail carries an acidic dileucine sorting signal whose two leucines the cited paper shows are required for binding native AP-3, and the same paper localises the binding to the delta-sigma3 hemicomplexes. This block is attached to a direct IPI rather than to a propagated row -- the schema places residue_claims inside propagation_review, and this is the row whose action rests on the residue argument, so the claims are recorded here in checkable form together with the resolved partner. Proposed replacements: cargo adaptor activity Supporting Evidence: PMID:16162817 Using this assay, the tyrosinase cytoplasmic domain interacts with hemicomplexes from AP-1 and AP-3 but not with those of AP-2 and AP-4 PMID:16162817 we show for the first time that the cytoplasmic dileucine-based sorting signal of tyrosinase binds not only to the δ-σ3 hemicomplexes of AP-3, confirming an AP-3 interaction PMID:16162817 GST pull-down assays confirmed the interactions of the tyrosinase cytoplasmic domain with the native AP-1 and AP-3 complexes and the dependence of these interactions on leucine residues 517 and 518 but not those at 527 and 528 PMID:16162817 Both adaptors recognize the tyrosinase dileucine-based melanosome sorting signal, and tyrosinase largely colocalizes with each adaptor on endosomes. PMID:14691137 signals from the human immunodeficiency virus negative factor protein and the lysosomal integral membrane protein II interact with combinations of the gamma and sigma1 subunits of AP-1 and the delta and sigma3 subunits of AP-3, but not the analogous combinations of AP-2 and AP-4 subunits PMID:21097499 This is evidenced by the loss of signal binding by the σ2 V88D or L103S substitutions and the homologous σ1A V88D and I103S and σ3A V94D and L109S substitutions. PMID:21097499 signals, on the other hand, do not bind to any single AP subunit but to combinations of γ-σ1, α-σ2, and δ-σ3 subunits, as demonstrated by the use of yeast three-hybrid (Y3H) and in vitro binding assays file:human/AP3S1/AP3S1-bioinformatics/RESULTS.md All five verify. Q92572 is at sequence version 1 and is 193 aa, matching the length asserted by the script. file:human/AP3S1/AP3S1-bioinformatics/RESULTS.md Identity at this pocket therefore does not separate binders from the non-binder, and the conservation argument for sigma3A carries no weight on its own. |
| GO:0005769 early endosome | NAS PMID:23247405 Cell type-specific Rab32 and Rab38 cooperate with the ubiqui... | ACCEPT | Summary: Early endosome, from a ComplexPortal curator statement citing a Rab32/Rab38 review. The compartment is right and is where AP-3 does its work. Reason: The evidence code is weak and the reference is a review projected across many gene products, but the location is the one the primary literature converged on. Immuno-electron microscopy places AP-3 on budding profiles of a tubular endosomal compartment, quantitative EM in melanocytes places it on endosome-derived membranes carrying tyrosinase, and the cited review itself describes AP-3 acting with Rab32 and Rab38 on early/recycling endosome tubules. This is where AP-3 selects cargo, so the term is core rather than merely defensible. Retained despite the NAS code, because the claim is true and duplicate support across evidence codes is legitimate. Supporting Evidence: PMID:23247405 Rab32 and Rab38 interact with AP-1, AP-3 and BLOC-2 on early/recycling endosome tubules, where cargo such as tyrosinase and Tyrp1 are loaded into vesicles or transport intermediates. PMID:15051738 Here, we show by immuno-electron microscopy that AP-3 is associated with budding profiles evolving from a tubular endosomal compartment that also exhibits budding profiles positive for AP-1. PMID:15051738 we propose that AP-3 defines a novel pathway by which lysosomal membrane proteins are transported from tubular sorting endosomes to lysosomes. PMID:16162817 The results indicate that tyrosinase traffics through the early endocytic system and that mammalian AP-3 functions on early endosomes to divert tyrosinase away from a pathway leading to multivesicular endosomes and toward one leading to melanosomes. PMID:16162817 These data show that tyrosinase is concentrated in AP-3-coated membranes essentially of endosomal origin. |
| GO:0005794 Golgi apparatus | IEA GO_REF:0000044 | KEEP AS NON CORE | Summary: Golgi apparatus, transferred from the UniProt subcellular-location vocabulary. Real, but it is the 1997 immunofluorescence compartment rather than the compartment where AP-3 was later shown to act. Reason: The vocabulary step itself is exact -- SL-0132 is UniProt's 'Golgi apparatus' term and its GO mapping is GO:0005794, checked against the UniProt locations API -- and the underlying UniProt statement is genuine. What makes this non-core rather than core is the resolution of the evidence behind it. The Golgi assignment comes from immunofluorescence in the two 1997 papers that identified AP-3, which describe the complex in 'the Golgi region of the cell as well as with more peripheral structures' and note that those peripheral structures colocalise only weakly with endosomal markers; the same authors were explicit that the peripheral compartment was unidentified. Immuno-EM later resolved it: AP-3 buds from tubular sorting endosomes, and in melanocytes tyrosinase is concentrated in AP-3-coated membranes 'essentially of endosomal origin'. Juxtanuclear staining at IF resolution does not separate the TGN from perinuclear endosomes. The UniProt CC line carries no ECO evidence tag at all, so there is no experimental attribution to defer to either. Kept, because a Golgi-region pool is reported and not contradicted; marked non-core because the informative locations are the endosome and the coated-vesicle membrane. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB-SubCell:SL-0132 · UniProt subcellular location SL-0132, Golgi apparatus SUPPORTS TRANSFER Verified against the UniProt locations API: SL-0132 is 'Golgi apparatus' and maps to GO:0005794 exactly, so the vocabulary step is faithful. The AP3S1 record states this location with no ECO evidence tag, i.e. it is an uncited UniProt assertion rather than a transferred experimental one, which is why the row is graded on the underlying literature rather than deferred to. Supporting Evidence: file:human/AP3S1/AP3S1-uniprot.txt CC -!- SUBCELLULAR LOCATION: Golgi apparatus. Cytoplasmic vesicle membrane; PMID:9151686 Immunofluorescence using anti-delta antibodies reveals that the AP-3 complex is associated with the Golgi region of the cell as well as with more peripheral structures. PMID:9151686 These peripheral structures show only limited colocalization with endosomal markers and may correspond to a postTGN biosynthetic compartment. PMID:9118953 Immunofluorescence microscopy analyses show that the sigma3-containing complex is present both in the area of the TGN and in peripheral structures, some of which contain the transferrin receptor. PMID:15051738 Here, we show by immuno-electron microscopy that AP-3 is associated with budding profiles evolving from a tubular endosomal compartment that also exhibits budding profiles positive for AP-1. PMID:16162817 These data show that tyrosinase is concentrated in AP-3-coated membranes essentially of endosomal origin. |
| GO:0006886 intracellular protein transport | IEA GO_REF:0000002 | ACCEPT | Summary: InterPro2GO mapping from the clathrin adaptor small-chain conserved site. Reason: IPR000804 is the conserved site shared by clathrin-adaptor small chains, and its complete GO mapping set is GO:0006886, GO:0016192 and GO:0030117 -- all three true of any AP sigma subunit. Moving transmembrane proteins between endosomes, lysosomes and lysosome-related organelles is intracellular protein transport by definition. The term is general, and generality is correct here: the signature is shared across the AP-1, AP-2, AP-3 and AP-4 sigmas, which act in different compartments, so a more specific term at the level of this signature would over-reach. Propagation Review Root cause: NO FAILURE CORE Sources checked: InterPro:IPR000804 · InterPro IPR000804, Clathrin adaptor complex, small chain (conserved site) SUPPORTS TRANSFER Confirmed against the InterPro API: a conserved_site entry, 'Clathrin adaptor complex, small chain', whose complete GO mapping set is GO:0006886, GO:0016192 and GO:0030117. AP3S1 carries the signature (PROSITE PS00989 CLAT_ADAPTOR_S on its UniProt record), and all three mapped terms are true of every AP sigma subunit, so the mapping holds across the whole signature rather than being merely permitted for this member. Supporting Evidence: file:human/AP3S1/AP3S1-uniprot.txt DR PROSITE; PS00989; CLAT_ADAPTOR_S; 1. PMID:9545220 A heterotetrameric complex termed AP-3 is involved in signal-mediated protein sorting to endosomal-lysosomal organelles. PMID:39705307 Adaptor protein complex-3 (AP-3) mediates cargo sorting from endosomes to lysosomes and lysosome-related organelles. |
| GO:0006896 Golgi to vacuole transport | IEA GO_REF:0000002 | MODIFY | Summary: Golgi to vacuole transport, from the AP-3-sigma-specific InterPro family entry. The process is right but the donor compartment is the fungal one. Reason: IPR027155 is the AP-3-sigma family entry ('AP-3 complex subunit sigma'), which AP3S1 genuinely matches, and its GO mappings are GO:0006896 and GO:0030123. The second is right. The first names a route that AP-3 sigma performs in fungi: both AP-3-clade seeds of this family's PAINT node carry GO:0006896 by IMP on their own records (S. cerevisiae Aps3, PMID:9335339; C. albicans Aps3, PMID:20870878, QuickGO checked live), which is the ALP pathway from the Golgi to the vacuole. The term is not fungus-only -- GO:0090160 Golgi to lysosome transport is its child and lysosome is_a vacuole, so nothing excludes mammals a priori -- but the mammalian route was resolved to a different donor compartment. Immuno-EM shows AP-3 budding from a tubular endosomal compartment and concludes that lysosomal membrane proteins are carried 'from tubular sorting endosomes to lysosomes'; in melanocytes AP-3 acts on early endosomes, and the AP-3 cargo it diverts is concentrated on membranes of endosomal origin. So the essence transfers and the compartment does not. Modified to GO:0008333 endosome to lysosome transport, which was checked on QuickGO: biological_process, not obsolete, is_a GO:0007041 and GO:0016192, so it sits under the same vacuolar-transport parent the original row had. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: COMPARTMENT OR COMPLEX MISMATCH Sources checked: InterPro:IPR027155 · InterPro IPR027155, AP-3 complex subunit sigma (family) SUPPORTS TRANSFER Confirmed against the InterPro API: a family entry, 'AP-3 complex subunit sigma', with exactly two GO mappings, GO:0006896 and GO:0030123. AP3S1 is a true member (the signature is on its UniProt record). The family spans fungal and animal AP-3 sigmas, and the experimental grounding for the Golgi-to-vacuole mapping sits on the fungal members, so the mapping carries their compartment rather than the mammalian one. Proposed replacements: endosome to lysosome transport Supporting Evidence: file:human/AP3S1/AP3S1-uniprot.txt DR InterPro; IPR027155; APS3. PMID:15051738 we propose that AP-3 defines a novel pathway by which lysosomal membrane proteins are transported from tubular sorting endosomes to lysosomes. PMID:15051738 Here, we show by immuno-electron microscopy that AP-3 is associated with budding profiles evolving from a tubular endosomal compartment that also exhibits budding profiles positive for AP-1. PMID:15051738 Moreover, recycling of endocytosed LAMP-1 and CD63 back to the cell surface is greatly increased in AP-3-deficient cells. PMID:16162817 The results indicate that tyrosinase traffics through the early endocytic system and that mammalian AP-3 functions on early endosomes to divert tyrosinase away from a pathway leading to multivesicular endosomes and toward one leading to melanosomes. PMID:10024875 The AP-3 deficiency results in increased surface expression of the lysosomal membrane proteins CD63, lamp-1, and lamp-2, but not of nonlysosomal proteins. |
| GO:0008089 anterograde axonal transport | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: Anterograde axonal transport, transferred by Ensembl Compara from the mouse ortholog. Real AP-3 biology, but neuronal-context and complex-level. Reason: The orthology is as safe as it gets: the donor is mouse Ap3s1, the 1:1 ortholog, and the two proteins are identical over all 193 residues. The mouse source is an IMP from PMID:21998198, which shows that BLOC-1 and AP-3 are required to move PI4KIIalpha out of neuronal cell bodies into neurites and nerve terminals, and which concludes that this is 'a novel vesicle transport mechanism requiring BLOC-1 and AP-3 complexes for cargo sorting from neuronal cell bodies to neurites and nerve terminals'. That is a genuine finding and the term fits it. It is not core for this gene: the perturbation is a whole-complex null in a single neuronal context, AP3S1 is ubiquitously expressed and its defining job is cargo selection on endosomes in every cell type, and this term describes the downstream microtubule-based movement of the carriers rather than anything sigma3A does. Kept as non-core. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q9DCR2 · Ap3s1, Mus musculus sigma3A, the 1:1 ortholog of the target SUPPORTS TRANSFER One donor under two identifiers, not two: UniProt's Ensembl cross-reference on Q9DCR2 is ENSMUST00000025357.9 / ENSMUSP00000025357.8, so the accession and the Ensembl protein id name the same mouse protein. Q9DCR2 is AP3S1_MOUSE, gene Ap3s1, Swiss-Prot, 193 aa -- the 1:1 ortholog and the same subunit, not the sigma3B paralog (which is Q8BSZ2 and is the donor for AP3S2's matching rows). The two sequences are byte-identical over all 193 residues, so the transfer has no divergence to bridge. The mouse source row is IMP from PMID:21998198, and the perturbation in that paper is the mocha allele Ap3d1(mh/mh) -- a delta-subunit null that removes the whole AP-3 complex -- with sigma3 appearing as an immunoblot readout rather than as the mutated gene. Annotating an obligate-complex subunit from a complex-null phenotype is standard GO practice and is not being called a defect; it is why the claim is complex-level and neuronal-context, hence non-core for a subunit of a ubiquitously expressed adaptor. ensembl:ENSMUSP00000025357 · Ensembl mouse translation ENSMUSP00000025357 SUPPORTS TRANSFER The Ensembl protein identifier for the same mouse entity as the UniProtKB:Q9DCR2 entry above, per that record's Ensembl cross-reference (ENSMUST00000025357.9 / ENSMUSP00000025357.8). It is the Compara handle for the transfer, not an independent second donor. Supporting Evidence: PMID:21998198 Our findings indicate a novel vesicle transport mechanism requiring BLOC-1 and AP-3 complexes for cargo sorting from neuronal cell bodies to neurites and nerve terminals. PMID:21998198 PI4KIIα targeting to neuronal processes is impaired in AP-3–deficient mocha (Ap3d1mh/mh) neurons. PMID:21998198 Synaptosome fractions from control brains (lanes 1–8) and AP-3–deficient mocha (Ap3d1mh/mh) brains (lanes 1′–8′) were resolved on SDS–PAGE, and the contents were analyzed by immunoblot with antibodies against synaptic vesicle markers (SV2, synaptophysin), AP-3–dependent synaptic vesicle cargoes (PI4KIIα, VAMP7, ZnT3), and AP-3 σ3 subunit. file:human/AP3S1/AP3S1-bioinformatics/RESULTS.md there is no divergence anywhere in the protein for the transfer to have to bridge. |
| GO:0008089 anterograde axonal transport | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: The same anterograde-axonal-transport claim as the Compara row, restated as a curator ISS from the same mouse ortholog. Reason: Same donor, same conclusion, same grade. A UniProt curator made this transfer by sequence similarity under GO_REF:0000024 while the Ensembl Compara pipeline made it under GO_REF:0000107; duplicate terms across evidence codes are legitimate and neither is redundant with the other in GOA's accounting. The transfer is safe because human AP3S1 and mouse Ap3s1 are identical over all 193 residues, and it is non-core for the same reason as the Compara row: the underlying mouse experiment is an AP-3 complex null assayed in brain. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q9DCR2 · Ap3s1, Mus musculus sigma3A, the 1:1 ortholog of the target SUPPORTS TRANSFER One donor under two identifiers, not two: UniProt's Ensembl cross-reference on Q9DCR2 is ENSMUST00000025357.9 / ENSMUSP00000025357.8, so the accession and the Ensembl protein id name the same mouse protein. Q9DCR2 is AP3S1_MOUSE, gene Ap3s1, Swiss-Prot, 193 aa -- the 1:1 ortholog and the same subunit, not the sigma3B paralog (which is Q8BSZ2 and is the donor for AP3S2's matching rows). The two sequences are byte-identical over all 193 residues, so the transfer has no divergence to bridge. The mouse source row is IMP from PMID:21998198, and the perturbation in that paper is the mocha allele Ap3d1(mh/mh) -- a delta-subunit null that removes the whole AP-3 complex -- with sigma3 appearing as an immunoblot readout rather than as the mutated gene. Annotating an obligate-complex subunit from a complex-null phenotype is standard GO practice and is not being called a defect; it is why the claim is complex-level and neuronal-context, hence non-core for a subunit of a ubiquitously expressed adaptor. Supporting Evidence: PMID:21998198 Our findings indicate a novel vesicle transport mechanism requiring BLOC-1 and AP-3 complexes for cargo sorting from neuronal cell bodies to neurites and nerve terminals. PMID:21998198 Synaptosome fractions from control brains (lanes 1–8) and AP-3–deficient mocha (Ap3d1mh/mh) brains (lanes 1′–8′) were resolved on SDS–PAGE, and the contents were analyzed by immunoblot with antibodies against synaptic vesicle markers (SV2, synaptophysin), AP-3–dependent synaptic vesicle cargoes (PI4KIIα, VAMP7, ZnT3), and AP-3 σ3 subunit. file:human/AP3S1/AP3S1-bioinformatics/RESULTS.md there is no divergence anywhere in the protein for the transfer to have to bridge. |
| GO:0008286 insulin receptor signaling pathway | TAS PMID:9792713 Interaction of insulin receptor substrate-1 with the sigma3A... | MARK AS OVER ANNOTATED | Summary: Insulin receptor signalling, from a 1998 traceable author statement. The paper reports a binding observation and closes on a hypothesis; no step of the pathway is shown to be executed by sigma3A. Reason: The finding is real as far as it goes: a 3T3-L1/3T3-F442A adipocyte expression library screened with IRS-1 pulled out a sigma3A cDNA, and GST-sigma3A released IRS-1 from low-density membranes. What the paper does not show is any signalling step. The term's definition is 'The series of molecular signals generated as a consequence of the insulin receptor binding to insulin', and the authors themselves state only that the results are 'consistent with the hypothesis that sigma3A serves as an IRS-1 receptor'. Nothing has confirmed it since: PubMed's cited-in link set for this PMID returns 5 papers, against 119 for PMID:14691137 and 70 for PMID:21097499 by the same query; a PubMed search for 'sigma3A adaptor' returns two papers in total; and UniProt's SUBUNIT line for Q92572 lists AGAP1 and BLOC-1 but not IRS1. TAS is a traceable author statement, not an experimental code, so the deference owed to a curator who read an assay does not apply here. Marked as an over-annotation rather than removed: the interaction itself is not contradicted, it simply does not support a pathway-participation claim. Supporting Evidence: PMID:9792713 A cDNA clone that encodes sigma3A, a small subunit of the AP-3 adaptor protein complex, was demonstrated to bind IRS-1 utilizing this cloning strategy. PMID:9792713 These results are consistent with the hypothesis that sigma3A serves as an IRS-1 receptor that may dictate the subcellular localization and the signaling functions of IRS-1. |
| GO:0015031 protein transport | IEA GO_REF:0000002 | ACCEPT | Summary: Protein transport, from the pan-AP sigma-subunit InterPro family entry. Reason: IPR016635 is the family entry for AP-complex sigma subunits and its only GO mapping is GO:0015031. AP3S1 is a member and AP-3 moves transmembrane proteins between compartments, so the mapping is correct. The term is broad, but it is broad in the right way: the signature covers sigma subunits of all four AP complexes, and this is what they have in common. Propagation Review Root cause: NO FAILURE CORE Sources checked: InterPro:IPR016635 · InterPro IPR016635, Adaptor protein complex, sigma subunit (family) SUPPORTS TRANSFER Confirmed against the InterPro API: a family entry, 'Adaptor protein complex, sigma subunit', whose only GO mapping is GO:0015031 protein transport. A pan-AP-sigma family entry mapped to a pan-AP-sigma process; nothing compartment-specific is smuggled in. Supporting Evidence: file:human/AP3S1/AP3S1-uniprot.txt DR PANTHER; PTHR11753; ADAPTOR COMPLEXES SMALL SUBUNIT FAMILY; 1. PMID:9545220 A heterotetrameric complex termed AP-3 is involved in signal-mediated protein sorting to endosomal-lysosomal organelles. PMID:10024875 Our results suggest that AP-3 functions in protein sorting to lysosomes |
| GO:0016020 membrane | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: Membrane, transferred by Ensembl Compara from a CACAO IMP on the mouse ortholog. True of a peripheral membrane coat, and uninformative. Reason: AP-3 is a peripheral membrane coat that cycles on and off membranes -- UniProt records AP3S1 as a peripheral membrane protein on the cytoplasmic side of cytoplasmic vesicles -- so a membrane association is genuine. The term is the root-level compartment and says nothing about which membrane, which is the whole content of an AP-3 localisation. The mouse source is a CACAO IMP citing PMID:16837549, a study of BLOC-1 interacting with AP-3 on early-endosome-associated tubules; that paper's own localisation result is far more specific than the term transferred from it. Kept, because it is not wrong, and marked non-core because the cytoplasmic vesicle membrane and early endosome rows carry the same information at a useful resolution. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q9DCR2 · Ap3s1, Mus musculus sigma3A, the 1:1 ortholog of the target SUPPORTS TRANSFER One donor under two identifiers, not two: UniProt's Ensembl cross-reference on Q9DCR2 is ENSMUST00000025357.9 / ENSMUSP00000025357.8, so the accession and the Ensembl protein id name the same mouse protein. Q9DCR2 is AP3S1_MOUSE, gene Ap3s1, Swiss-Prot, 193 aa -- the 1:1 ortholog and the same subunit, not the sigma3B paralog (which is Q8BSZ2 and is the donor for AP3S2's matching rows). The two sequences are byte-identical over all 193 residues, so the transfer has no divergence to bridge. The mouse source row here is a CACAO IMP (PMID:16837549) rather than the PMID:21998198 rows that feed the axonal/synaptic terms. ensembl:ENSMUSP00000025357 · Ensembl mouse translation ENSMUSP00000025357 SUPPORTS TRANSFER The Ensembl protein identifier for the same mouse entity as the UniProtKB:Q9DCR2 entry above; the Compara handle, not a second donor. Supporting Evidence: file:human/AP3S1/AP3S1-uniprot.txt CC -!- SUBCELLULAR LOCATION: Golgi apparatus. Cytoplasmic vesicle membrane; PMID:16837549 Both BLOC-1 and -2 localize mainly to early endosome-associated tubules as determined by immunoelectron microscopy. PMID:16837549 Here, we report that BLOC-1 interacts physically and functionally with AP-3 to facilitate the trafficking of a known AP-3 cargo, CD63, and of tyrosinase-related protein 1 (Tyrp1), a melanosomal membrane protein previously thought to traffic only independently of AP-3. |
| GO:0016183 synaptic vesicle coating | NAS PMID:15537701 Genetic analysis of the neuronal and ubiquitous AP-3 adaptor... | MODIFY | Summary: Synaptic vesicle coating, from ComplexPortal. AP-3 does coat synaptic vesicles, but this term's definition describes coated-pit formation at the presynaptic plasma membrane, which is AP-2 territory. Reason: Decided from the definition rather than the label. GO:0016183 is defined as 'The formation of clathrin coated pits in the presynaptic membrane endocytic zone, triggered by the presence of high concentrations of synaptic vesicle components' -- a plasma-membrane endocytic event. AP-3 is not a plasma-membrane adaptor: it acts on exit from tubular sorting endosomes, and its synaptic role in the cited paper is the delivery of ZnT3 and ClC-3 into synaptic vesicles, not the nucleation of endocytic pits. GO already has the right term: GO:0016182 synaptic vesicle budding from endosome, 'Budding of synaptic vesicles during the formation of constitutive recycling vesicles from early endosomes' (QuickGO, checked; biological_process, not obsolete). That is exactly the compartment and the step. The essence of the ComplexPortal claim -- AP-3 coats synaptic vesicles -- is kept; only the compartment-bearing term is changed. Note the knock-on: the GO:0098793 presynapse row is derived from this one by inter-ontology inference, and is graded on its own merits there rather than on this term surviving. Proposed replacements: synaptic vesicle budding from endosome Supporting Evidence: PMID:15537701 Consistently, beta3B deficiency compromised synaptic zinc stores assessed by Timm's staining and the synaptic vesicle targeting of membrane proteins involved in zinc uptake (ZnT3 and ClC-3). PMID:15537701 Our results suggest that concerted nonredundant functions of neuronal and ubiquitous AP-3 provide a mechanism to control the levels of selected membrane proteins in synaptic vesicles. PMID:15537701 beta3B-containing AP-3 complexes were preferentially targeted to neuronal processes. PMID:15051738 we propose that AP-3 defines a novel pathway by which lysosomal membrane proteins are transported from tubular sorting endosomes to lysosomes. |
| GO:0016192 vesicle-mediated transport | IBA GO_REF:0000033 | ACCEPT | Summary: Phylogenetic inference of vesicle-mediated transport from PTN000204281, the pan-AP-sigma ancestral node of PANTHER family PTHR11753. Correct, and correctly general. Reason: The tempting review is that vesicle-mediated transport is vague when AP-3 does something far more specific, and that the row should be modified to an endosome-to-lysosome term. That would be wrong, and the reason is visible once the node is resolved rather than the donor count read. Each of the eleven gene-level donors was resolved individually, and they are not AP-3 orthologs: they are AP-3 sigmas (budding-yeast Aps3, C. albicans Aps3), AP-1 sigmas (mouse Ap1s1 and Ap1s2, fly AP-1sigma, budding-yeast Aps1, fission-yeast vas2) and AP-2 sigmas (human AP2S1, rat Ap2s1, fly AP-2sigma, worm aps-2). Those subunits work in different compartments and different pathways -- endosome-to-lysosome and LRO delivery, TGN/endosome sorting, and clathrin-mediated endocytosis at the plasma membrane. The donors therefore do not agree on a pathway, and vesicle-mediated transport is precisely their least common ancestor. A compartment-specific term at this node would propagate AP-3 biology onto every AP-1 and AP-2 sigma in the family, which is the failure the general term avoids. GRANULARITY_MISMATCH is explicitly not applicable: it requires that the donors agree and that a more specific term was available, and neither holds. Separately, the target is unambiguously inside the inheriting clade -- two AP-3-clade seeds are in the donor list, AP3S1 is a full-length 193-residue sigma with an intact longin-like fold, and it is one of the four chains of a solved human AP-3 core (PDB 9C58/9C59/9C5B/9C5C, chain S mapped to Q92572 on the UniProt record) -- and there is no target-specific evidence of loss or divergence. The term is genuinely core rather than merely defensible: moving cargo in coated vesicles is what this protein exists to do. Propagation Review Root cause: NO FAILURE CORE Sources checked: CGD:CAL0000182525 · APS3, Candida albicans AP-3 sigma subunit (UniProtKB:Q59QC5) SUPPORTS TRANSFER Resolved by UniProt xref lookup to Q59QC5, the C. albicans AP-3 sigma (Swiss-Prot, 175 aa); the GO API bioentity record for the CGD id returns 'Component of the adaptor complex AP-3, which is involved in vacuolar protein sorting'. Its own GO record carries GO:0006896 Golgi to vacuole transport by IMP (PMID:20870878), so it supplies independent experimental grounding from the AP-3 clade -- the target's own subfamily. FB:FBgn0039132 · AP-1sigma, Drosophila melanogaster AP-1 sigma subunit SUPPORTS TRANSFER Resolved by UniProt xref lookup to the D. melanogaster AP-1sigma gene (four TrEMBL entries, all 'AP complex subunit sigma'). An AP-1-clade donor, i.e. one of the members whose pathway differs from the target's -- which is why the node's term is the shared parent. FB:FBgn0043012 · AP-2sigma, Drosophila melanogaster AP-2 sigma subunit SUPPORTS TRANSFER Resolved by UniProt xref lookup to the D. melanogaster AP-2sigma gene (two TrEMBL entries). An AP-2-clade donor. Present in this row's WITH/FROM; it appears in the node's GO:0043231 seed list in the fetched PAINT slice but not in that slice's GO:0016192 seed list, which has 10 seeds. The slice's GO:0016192 row is dated 20260828 and the GOA row 20250902, so this is a snapshot difference between releases rather than a defect. MGI:MGI:1098244 · Ap1s1, Mus musculus sigma1A (UniProtKB:P61967) SUPPORTS TRANSFER Resolved by UniProt xref lookup to P61967, AP1S1_MOUSE, the mouse sigma1A. An AP-1-clade donor, not the target's ortholog -- worth stating explicitly, because the target's own mouse ortholog (Q9DCR2) is NOT in this list and neither is the target itself. MGI:MGI:1889383 · Ap1s2, Mus musculus sigma1B (UniProtKB:Q9DB50) SUPPORTS TRANSFER Resolved by UniProt xref lookup to Q9DB50, the mouse sigma1B. It carries GO:0016192 by IMP and GO:0016182 synaptic vesicle budding from endosome by IDA on its own record -- experimental grounding inside the AP-1 sigma clade. PANTHER:PTN000204281 · PTHR11753 ADAPTOR COMPLEXES SMALL SUBUNIT FAMILY, ancestral IBD node SUPPORTS TRANSFER The node, not a donor. The PAINT slice in this repository shows PTN000204281 carrying two IBD assertions: GO:0016192 (this one, 10 seeds in the slice) and GO:0043231 (13 seeds). UniProtKB:Q92572 is in neither seed list, so this IBA is not self-referential; and no GO:0043231 IBA row exists in this gene's GOA, so there is nothing to review for the other assertion. Because the seeds span AP-1, AP-2 and AP-3 sigma subunits across fungi, nematode, fly, rodent and human, the node is the deep pan-AP-sigma ancestor and vesicle-mediated transport is its correct least common ancestor term. AP3S1 descends from it, and QuickGO shows the identical 12-entry WITH/FROM on AP1S1, AP1S3, AP2S1, AP3S2 and AP4S1 as well -- the row is a family-wide consequence of one node placement, as an IBA should be. PomBase:SPAP27G11.06c · vas2, Schizosaccharomyces pombe AP-1 sigma-1 (UniProtKB:Q9P7N2) SUPPORTS TRANSFER Resolved by UniProt xref lookup to Q9P7N2, the fission-yeast AP-1 sigma vas2. Independently grounded: GO:0042147, GO:0099638, GO:0005768, GO:0005794 and GO:0030121 all by IDA (PMID:19624755). Its presence makes the node at least as old as the animal/fungal divergence. RGD:620188 · Ap2s1, Rattus norvegicus AP-2 sigma (UniProtKB:P62744) SUPPORTS TRANSFER Resolved by UniProt xref lookup to P62744, the rat AP-2 sigma (five hits returned, one Swiss-Prot and four TrEMBL fragments/isoforms of the same gene). Independently grounded by IDA/IMP/EXP rows for GO:0098884 and GO:0008021. An AP-2-clade donor. SGD:S000003561 · APS3, Saccharomyces cerevisiae AP-3 sigma (UniProtKB:P47064) SUPPORTS TRANSFER Resolved by UniProt xref lookup to P47064, the budding-yeast AP-3 sigma Aps3, with GO:0006896 by IMP (PMID:9335339) and GO:0030123 by IMP and IPI (PMID:9250663) on its own record. The strongest AP-3-clade grounding in the list: the same subunit of the same complex, experimentally placed in it. SGD:S000004160 · APS1, Saccharomyces cerevisiae AP-1 sigma-1 (UniProtKB:P35181) SUPPORTS TRANSFER Resolved by UniProt xref lookup to P35181, the budding-yeast AP-1 sigma Aps1, with GO:0006896 by IMP (PMID:17003107) and GO:0030121 by IDA (PMID:10564262). An AP-1-clade donor from the same yeast lineage as the AP-3 seed above, which is what shows the two pathways already existed separately at that node. UniProtKB:P53680 · AP2S1, Homo sapiens AP-2 sigma SUPPORTS TRANSFER The human AP-2 sigma paralog; found directly in interpro/panther/PTHR11753/PTHR11753-entries.csv. Its inclusion is what makes the node pan-AP rather than AP-3-specific, and it is the structural anchor used in this review's bioinformatics analysis, since the dileucine pocket was defined on AP-2. Note that P53680 is both a donor here and a recipient of the same IBA in its own GOA -- self-referential for AP2S1, and legitimately so; not the case for AP3S1. WB:WBGene00000157 · aps-2, Caenorhabditis elegans AP-2 sigma (F02E8.3, UniProtKB:Q19123) SUPPORTS TRANSFER Not resolvable through a UniProt xref search on the WBGene id. Recovered via the GO API bioentity endpoint, which returns 'AP complex subunit sigma', NCBITaxon:6239 and the synonyms CELE_F02E8.3 / F02E8.3; the UniProt cross-reference for F02E8.3 then returns Q19123, gene aps-2. An AP-2-clade nematode donor, grounded by IMP rows including GO:0072583 clathrin-dependent endocytosis. Supporting Evidence: file:human/AP3S1/AP3S1-uniprot.txt DR PANTHER; PTHR11753; ADAPTOR COMPLEXES SMALL SUBUNIT FAMILY; 1. file:human/AP3S1/AP3S1-uniprot.txt DR PDB; 9C59; EM; 4.30 A; S/s=1-193. PMID:39705307 Adaptor protein complex-3 (AP-3) mediates cargo sorting from endosomes to lysosomes and lysosome-related organelles. PMID:9545220 A heterotetrameric complex termed AP-3 is involved in signal-mediated protein sorting to endosomal-lysosomal organelles. PMID:21097499 The clathrin-associated, heterotetrameric adaptor protein (AP) complexes, AP-1, AP-2, and AP-3, recognize signals in the cytosolic domains of transmembrane proteins, leading to their sorting to endosomes, lysosomes, lysosome-related organelles, and/or the basolateral membrane of polarized epithelial cells. file:human/AP3S1/AP3S1-bioinformatics/RESULTS.md All five verify. Q92572 is at sequence version 1 and is 193 aa, matching the length asserted by the script. |
| GO:0016192 vesicle-mediated transport | IEA GO_REF:0000002 | ACCEPT | Summary: The same process reached independently by InterPro2GO from the small-chain conserved site. Reason: GO:0016192 is one of IPR000804's three GO mappings, the protein carries the signature, and the conclusion is identical to the IBA's and correct for the same reasons. Duplication of a term across evidence codes is expected and is not a defect; here it is mild corroboration, since the two routes -- a phylogenetic node and a sequence signature -- are independent of each other. Propagation Review Root cause: NO FAILURE CORE Sources checked: InterPro:IPR000804 · InterPro IPR000804, Clathrin adaptor complex, small chain (conserved site) SUPPORTS TRANSFER Confirmed against the InterPro API: a conserved_site entry, 'Clathrin adaptor complex, small chain', whose complete GO mapping set is GO:0006886, GO:0016192 and GO:0030117. AP3S1 carries the signature (PROSITE PS00989 CLAT_ADAPTOR_S on its UniProt record), and all three mapped terms are true of every AP sigma subunit, so the mapping holds across the whole signature rather than being merely permitted for this member. Supporting Evidence: file:human/AP3S1/AP3S1-uniprot.txt DR PROSITE; PS00989; CLAT_ADAPTOR_S; 1. PMID:39705307 Adaptor protein complex-3 (AP-3) mediates cargo sorting from endosomes to lysosomes and lysosome-related organelles. |
| GO:0016192 vesicle-mediated transport | NAS PMID:23247405 Cell type-specific Rab32 and Rab38 cooperate with the ubiqui... | ACCEPT | Summary: Statement-level restatement of AP-3's transport role, from a Rab32/Rab38 review that discusses AP-3 as one of the machineries delivering cargo to lysosome-related organelles. Reason: The term is correct and core for AP3S1 -- this is what the protein does -- so the row is accepted even though its evidence is weak and complex-level. It adds nothing the IBA does not already carry with a considered phylogenetic judgment behind it, and the cited review says nothing about sigma3A specifically; a QuickGO query by this reference returns 178 annotations over 48 entities, so it is one statement projected across AP-1, AP-3 and BLOC-2. Kept rather than removed because duplicate terms across evidence codes are legitimate and the claim is true. Supporting Evidence: PMID:23247405 BLOC-2, AP-3, and AP-1 proteins function in concert with Rab38 and Rab32 proteins to mediate protein trafficking to lysosome-related organelles. PMID:23247405 Rab32 and Rab38 interact with AP-1, AP-3 and BLOC-2 on early/recycling endosome tubules, where cargo such as tyrosinase and Tyrp1 are loaded into vesicles or transport intermediates. |
| GO:0030117 membrane coat | IEA GO_REF:0000002 | ACCEPT | Summary: Membrane coat, from the clathrin-adaptor small-chain conserved site. Reason: GO:0030117 is defined to include coats 'formed by clathrin plus an adaptor complex, the COPI and COPII complexes, and possibly others', and AP-3 is one of those coats -- it associates with clathrin through the beta3 appendage, and it is itself the coat on the vesicles it forms. sigma3A is a constituent of that coat, not merely near it: it is one of the four chains in the reconstituted, cryo-EM-solved human AP-3 core, and UniProt places the protein in the coat surrounding the cytoplasmic face of coated vesicles. Correct and core. Propagation Review Root cause: NO FAILURE CORE Sources checked: InterPro:IPR000804 · InterPro IPR000804, Clathrin adaptor complex, small chain (conserved site) SUPPORTS TRANSFER Confirmed against the InterPro API: a conserved_site entry, 'Clathrin adaptor complex, small chain', whose complete GO mapping set is GO:0006886, GO:0016192 and GO:0030117. AP3S1 carries the signature (PROSITE PS00989 CLAT_ADAPTOR_S on its UniProt record), and all three mapped terms are true of every AP sigma subunit, so the mapping holds across the whole signature rather than being merely permitted for this member. Supporting Evidence: PMID:9545220 In vitro binding assays showed that mammalian AP-3 did associate with clathrin by interaction of the appendage domain of its beta3 subunit with the amino-terminal domain of the clathrin heavy chain. PMID:9545220 Thus, AP-3 function in protein sorting may depend on clathrin. PMID:39705307 To understand the mechanism of AP-3 membrane recruitment and activation, we reconstituted human AP-3 and determined multiple structures in the soluble and membrane-bound states using electron cryo-microscopy. file:human/AP3S1/AP3S1-uniprot.txt DR PDB; 9C59; EM; 4.30 A; S/s=1-193. |
| GO:0030119 AP-type membrane coat adaptor complex | TAS PMID:8697810 Cloning, expression pattern and mapping to 12p 13.2 --> p13.... | ACCEPT | Summary: AP-type membrane coat adaptor complex, from the 1996 cloning paper that named CLAPS3. Correct, and appropriately general for what that paper established. Reason: GO:0030119 is the parent of the four specific AP complexes, defined as the heterotetramers that link clathrin or another coat-forming molecule to a membrane, with two large, one medium and one small subunit. The 1996 paper cloned a 193-residue clathrin-adaptor small chain expressed as a 1.35 kb transcript in every tissue tested; it did not identify which AP complex the chain belonged to, which was established the following year. The general term is therefore the right level for this reference, and nothing is lost, because the specific child GO:0030123 is separately annotated three times from the papers that did establish it. Worth a footnote: this paper's title asserts a 12p13.2-p13.1 FISH assignment, which is superseded -- HGNC places AP3S1 at 5q22.3-q23.1 and UniProt files it under Chromosome 5 with CCDS4123.1. The locus is stale; the protein-level content the annotation rests on is not. Supporting Evidence: PMID:8697810 The cDNA, designated CLAPS3 (clathrin-associated/assembly/adaptor protein, small 3, 22 kDa), contained an open reading frame of 579 nucleotides encoding 193 amino acids. PMID:8697810 Northern-blot analysis revealed expression of a 1.35-kb transcript in all human tissues examined. PMID:8697810 This gene was mapped to chromosome bands 12p13.2 --> p13.1 by FISH. file:human/AP3S1/AP3S1-uniprot.txt DR Proteomes; UP000005640; Chromosome 5. file:human/AP3S1/AP3S1-deep-research-affinage.md AP3S1 (CLAPS3) encodes a small clathrin-adaptor chain that functions as a subunit of the AP-3 adaptor complex in intracellular vesicular trafficking, and is expressed ubiquitously across human tissues |
| GO:0030123 AP-3 adaptor complex | IDA PMID:9118953 AP-3: an adaptor-like protein complex with ubiquitous expres... | ACCEPT | Summary: Direct demonstration that sigma3A is a subunit of AP-3, from the paper that identified both sigma3 isoforms and named the complex. Reason: The strongest and most specific row on the gene. The cited paper identified sigma3A and sigma3B, showed by immunoprecipitation and immunoblot that they are components of a large complex with the 47, 140 and 160 kDa chains that became mu3, beta3 and delta, and named it AP-3. The GO term definition names the subunit and both isoforms explicitly, so the term and the protein match without interpretation. This reference is not a broad projection either: a QuickGO query by PMID:9118953 returns 3 annotations over 2 entities, Q92572 and P59780 -- precisely the two proteins the paper identified. Core. Supporting Evidence: PMID:9118953 We have identified two closely related human proteins (sigma3A and sigma3B) that are homologous to the small chains, sigma1 and sigma2, of clathrin-associated adaptor complexes. PMID:9118953 sigma3A and sigma3B are components of a large complex, named AP-3, that also contains proteins of apparent molecular masses of 47, 140 and 160 kDa. PMID:9118953 Northern and Western blot analyses demonstrate that the products of both the sigma3A and sigma3B genes are expressed in a wide variety of tissues and cell lines. file:human/AP3S1/AP3S1-uniprot.txt DR GO; GO:0030123; C:AP-3 adaptor complex; IDA:FlyBase. |
| GO:0030123 AP-3 adaptor complex | IEA GO_REF:0000120 | ACCEPT | Summary: AP-3 complex membership reached automatically by two independent routes, an ARBA rule and the AP-3-sigma InterPro family entry. Reason: Both routes were checked rather than assumed. The ARBA rule was fetched (https://rest.uniprot.org/arba/ARBA00033921) and has a single condition set: FunFam id 3.30.450.60:FF:000001 AND taxon Primates, annotating GO:0030123. A FunFam plus a clade looks like a loose condition for a complex-membership call, so I tested whether that FunFam actually separates AP-3 sigmas from the rest of the family. It does: of the seven human sigma paralogs, only AP3S1 and AP3S2 carry 3.30.450.60:FF:000001 (AP1S1 and AP1S3 carry FF:000005, AP1S2 FF:000009, AP2S1 FF:000004, AP4S1 FF:000010; UniProt xref_funfam, queried live). The second route, IPR027155, is the AP-3-sigma-specific family entry and maps to GO:0030123 directly. Neither route distinguishes sigma3A from sigma3B, and neither needs to -- the term definition admits both isoforms. Correct and core. Propagation Review Root cause: NO FAILURE CORE Sources checked: ARBA:ARBA00033921 · ARBA00033921, automatic annotation rule SUPPORTS TRANSFER Fetched and read: one condition set, 'FunFam id = 3.30.450.60:FF:000001' AND 'taxon = Primates', producing GO:0030123. AP3S1 satisfies both (the FunFam is on its UniProt record and it is human). The FunFam condition is discriminative in practice -- among the seven human sigma paralogs only the two AP-3 sigmas carry it -- so the rule is not a family-wide over-call, though it cannot and does not separate sigma3A from sigma3B. InterPro:IPR027155 · InterPro IPR027155, AP-3 complex subunit sigma (family) SUPPORTS TRANSFER Confirmed against the InterPro API: family entry 'AP-3 complex subunit sigma', GO mappings GO:0006896 and GO:0030123. This is the AP-3-sigma-specific signature, matched by AP3S1, so for this term it is a paralog-appropriate route and not the family-wide generalisation that the same entry's other mapping turns out to be (see the GO:0006896 row). Supporting Evidence: file:human/AP3S1/AP3S1-uniprot.txt DR FunFam; 3.30.450.60:FF:000001; AP complex subunit sigma; 1. file:human/AP3S1/AP3S1-uniprot.txt DR InterPro; IPR027155; APS3. file:human/AP3S1/AP3S1-uniprot.txt DR CDD; cd14834; AP3_sigma; 1. PMID:14691137 Together, these experiments demonstrate that the LIMP-II cytosolic tail binds to the AP-1 and AP-3 complexes through γ1–σ1A and δ–σ3 (A or B isoforms). |
| GO:0030123 AP-3 adaptor complex | NAS PMID:9151686 Characterization of the adaptor-related protein complex, AP-... | ACCEPT | Summary: AP-3 complex membership as a ComplexPortal curator statement, citing the companion 1997 paper that raised antibodies against recombinant delta and sigma3. Reason: Weak evidence code, correct and core claim, and the cited paper is a genuine primary demonstration rather than a review: antibodies against recombinant delta and sigma3 identified them as the remaining two subunits of the adaptor-like complex. A QuickGO query by this reference returns 17 annotations over 14 entities -- the AP-3 subunits in human and mouse -- so it is a complex-level projection, but the projection is onto exactly the set of proteins the paper is about. Duplicate with the IDA row and none the worse for it. Supporting Evidence: PMID:9151686 Antibodies raised against recombinant delta and sigma3 show that they are the other two subunits of the adaptor-like complex. PMID:9118953 sigma3A and sigma3B are components of a large complex, named AP-3, that also contains proteins of apparent molecular masses of 47, 140 and 160 kDa. |
| GO:0030133 transport vesicle | TAS PMID:9118953 AP-3: an adaptor-like protein complex with ubiquitous expres... | KEEP AS NON CORE | Summary: Transport vesicle, a 2003 traceable author statement from the AP-3 identification paper. Defensible but off-compartment for what AP-3 was later shown to coat. Reason: GO:0030133 is defined as 'Any of the vesicles of the constitutive secretory pathway, which carry cargo from the endoplasmic reticulum to the Golgi, between Golgi cisternae, from the Golgi to the ER (retrograde transport) or to destinations within or outside the cell' -- a secretory-pathway term. The cited paper localised the sigma3-containing complex to the TGN area and to peripheral structures that it could not identify, and the companion paper explicitly left them open as possibly a post-TGN biosynthetic compartment. Immuno-EM subsequently identified AP-3-coated carriers as endosome-derived. So the row is not contradicted -- the final clause of the definition is broad enough to accommodate a carrier heading for the lysosome -- but it is a low-resolution statement from before the compartment was resolved. Kept as non-core; the informative cellular-component rows are the AP-3 complex itself, the cytoplasmic vesicle membrane and the early endosome. Supporting Evidence: PMID:9118953 Immunofluorescence microscopy analyses show that the sigma3-containing complex is present both in the area of the TGN and in peripheral structures, some of which contain the transferrin receptor. PMID:9151686 These peripheral structures show only limited colocalization with endosomal markers and may correspond to a postTGN biosynthetic compartment. PMID:15051738 Here, we show by immuno-electron microscopy that AP-3 is associated with budding profiles evolving from a tubular endosomal compartment that also exhibits budding profiles positive for AP-1. PMID:16162817 These data show that tyrosinase is concentrated in AP-3-coated membranes essentially of endosomal origin. |
| GO:0030659 cytoplasmic vesicle membrane | IEA GO_REF:0000044 | ACCEPT | Summary: Cytoplasmic vesicle membrane, from the UniProt subcellular-location vocabulary. Reason: SL-0089 is UniProt's 'Cytoplasmic vesicle membrane' term and maps to GO:0030659 exactly (UniProt locations API, checked). This is where an adaptor coat sits: UniProt records AP3S1 as a peripheral membrane protein on the cytoplasmic side and as a component of the coat surrounding the cytoplasmic face of coated vesicles, and the cryo-EM work reconstitutes exactly that geometry, building Arf1-AP-3 complexes on lipid nanodiscs. Correct and core. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB-SubCell:SL-0089 · UniProt subcellular location SL-0089, Cytoplasmic vesicle membrane SUPPORTS TRANSFER Verified against the UniProt locations API: SL-0089 is 'Cytoplasmic vesicle membrane' and maps to GO:0030659. Unlike the Golgi row, this location is the one the structural and EM work independently confirms, so the vocabulary step delivers a well-supported statement. Supporting Evidence: file:human/AP3S1/AP3S1-uniprot.txt CC -!- SUBCELLULAR LOCATION: Golgi apparatus. Cytoplasmic vesicle membrane; PMID:39705307 To understand the mechanism of AP-3 membrane recruitment and activation, we reconstituted human AP-3 and determined multiple structures in the soluble and membrane-bound states using electron cryo-microscopy. PMID:15051738 Here, we show by immuno-electron microscopy that AP-3 is associated with budding profiles evolving from a tubular endosomal compartment that also exhibits budding profiles positive for AP-1. |
| GO:0035651 AP-3 adaptor complex binding | IEA GO_REF:0000107 | REMOVE | Summary: Asserts that AP3S1 binds the AP-3 adaptor complex. AP3S1 is a constitutive subunit of that complex, and GO already records the relation, three times, as GO:0030123 part_of. Reason: The term is defined as 'Binding to an AP-3 adaptor complex', with a definition that goes on to enumerate the beta3, delta, mu3 and sigma3 subunits -- one of which is the annotated protein. Typing a constitutive subunit as a ligand of its own complex is a role conflation; the relation GO provides for a subunit is part_of, and this gene carries it on GO:0030123 by IDA, by NAS and by IEA. This is not an isolated slip: a QuickGO query for GO:0035651 restricted to human returns 29 annotations over 28 entities, and all seven human AP-3 subunit genes are among them -- AP3B1, AP3B2, AP3D1, AP3M1, AP3M2, AP3S1, AP3S2 -- every one by IEA, alongside the entities that do have experimental evidence for binding AP-3 from outside it (PI4K2A by IDA, RAB32 and RAB38 by IPI). The complex's own subunits have been swept in with its ligands. The proximate source is the mouse Ap3s1 IDA on PMID:19010779, a study in which AP-3 was the cross-linking bait and BLOC-1, BLOC-2, HOPS, clathrin and PI4KIIalpha were the associated proteins; that cache is abstract-only, so this argument deliberately does not rest on anything inside the assay. It rests on the relation: no experiment can make a constitutive subunit a ligand of the complex it constitutes. Note also that this is the gene's only molecular-function row in GOA, which is why the review adds a subunit-level MF rather than leaving the aspect empty. The merged sibling reviews of AP3D1 and AP3M1 grade this same IEA row MARK_AS_OVER_ANNOTATED rather than REMOVE, on the ground that the mouse IDA behind it is a real observation read in full by MGI curators. REMOVE is preferred here because the row being removed is the human IEA projection, not the mouse IDA; because the family-wide sweep (all seven human AP-3 subunits, every one by IEA) shows the defect is in the relation type, not in any one experiment; and because a subunit that coassembles with delta into the complex core cannot be a ligand of that complex under the definition of the term. The sibling verdicts are noted so the family carries one argument with an explicit choice of action, and harmonising them to REMOVE is recommended. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: ROLE CONFLATION Sources checked: UniProtKB:Q9DCR2 · Ap3s1, Mus musculus sigma3A, the 1:1 ortholog of the target SUPPORTS TRANSFER One donor under two identifiers, not two: UniProt's Ensembl cross-reference on Q9DCR2 is ENSMUST00000025357.9 / ENSMUSP00000025357.8, so the accession and the Ensembl protein id name the same mouse protein. Q9DCR2 is AP3S1_MOUSE, gene Ap3s1, Swiss-Prot, 193 aa -- the 1:1 ortholog and the same subunit, not the sigma3B paralog (which is Q8BSZ2 and is the donor for AP3S2's matching rows). The two sequences are byte-identical over all 193 residues, so the transfer has no divergence to bridge. The mouse source row for this term is GO:0035651 IDA from PMID:19010779 (QuickGO). The transfer is faithful and the orthology is perfect; the problem is upstream, in typing a complex subunit as a binder of its own complex, and it is inherited unchanged. ensembl:ENSMUSP00000025357 · Ensembl mouse translation ENSMUSP00000025357 SUPPORTS TRANSFER The Ensembl protein identifier for the same mouse entity as the UniProtKB:Q9DCR2 entry above; the Compara handle, not a second donor. It adds no independent support and cannot repair the relation error. Supporting Evidence: PMID:19010779 AP-3 was co-isolated with BLOC-1, BLOC-2, and homotypic fusion and vacuole protein sorting complex subunits; clathrin; and phosphatidylinositol-4-kinase type II alpha (PI4KIIalpha). PMID:19010779 To identify molecular associations of these complexes, we used in vivo cross-linking followed by purification of cross-linked AP-3 complexes and mass spectrometric identification of associated proteins. file:human/AP3S1/AP3S1-uniprot.txt DR GO; GO:0030123; C:AP-3 adaptor complex; IDA:FlyBase. PMID:11807095 In pearl cells, the delta and sigma3 subunits coassemble into a heterodimer, whereas mu3 gets destroyed. PMID:39705307 To understand the mechanism of AP-3 membrane recruitment and activation, we reconstituted human AP-3 and determined multiple structures in the soluble and membrane-bound states using electron cryo-microscopy. |
| GO:0035654 clathrin-coated vesicle cargo loading, AP-3-mediated | NAS PMID:9545220 Association of the AP-3 adaptor complex with clathrin. | ACCEPT | Summary: AP-3-mediated cargo loading into coated vesicles, from ComplexPortal. This is the process term for exactly what sigma3A does. Reason: The definition is 'Formation of a macromolecular complex between proteins of the AP-3 adaptor complex and proteins and/or lipoproteins that are going to be transported by a clathrin-coated vesicle', with the clathrin clause hedged ('in some organisms'). Cargo capture is the step sigma3A performs: the delta-sigma3A hemicomplex binds the acidic dileucine signals of Nef, LIMP-II, tyrosinase and others, and substituting the sigma3A pocket residues abolishes it. The clathrin part of the definition is also satisfied for mammalian AP-3, which is what the cited paper established. Core, and the process counterpart of the molecular function proposed on the GO:0005515 row. Supporting Evidence: PMID:9545220 In vitro binding assays showed that mammalian AP-3 did associate with clathrin by interaction of the appendage domain of its beta3 subunit with the amino-terminal domain of the clathrin heavy chain. PMID:9545220 Thus, AP-3 function in protein sorting may depend on clathrin. PMID:14691137 signals from the human immunodeficiency virus negative factor protein and the lysosomal integral membrane protein II interact with combinations of the gamma and sigma1 subunits of AP-1 and the delta and sigma3 subunits of AP-3, but not the analogous combinations of AP-2 and AP-4 subunits PMID:21097499 This is evidenced by the loss of signal binding by the σ2 V88D or L103S substitutions and the homologous σ1A V88D and I103S and σ3A V94D and L109S substitutions. PMID:16162817 Using this assay, the tyrosinase cytoplasmic domain interacts with hemicomplexes from AP-1 and AP-3 but not with those of AP-2 and AP-4 |
| GO:0036465 synaptic vesicle recycling | NAS PMID:15537701 Genetic analysis of the neuronal and ubiquitous AP-3 adaptor... | KEEP AS NON CORE | Summary: Synaptic vesicle recycling, from ComplexPortal, citing the mouse beta3A/beta3B genetic comparison. Reason: Defensible and cell-type-restricted. The cited paper shows that neuronal (beta3B-containing) and ubiquitous (beta3A-containing) AP-3 control the synaptic vesicle content of ZnT3 and ClC-3 in opposite directions, and concludes that the two AP-3 variants control the levels of selected membrane proteins in synaptic vesicles. sigma3A is a subunit of both variants -- UniProt lists it in both the ubiquitous and the neuronal ComplexPortal entries -- so the claim reaches this gene. It is not core: AP3S1 is ubiquitously expressed, the experiments are mouse brain and the perturbations are of the beta3 subunits, and the defining function of sigma3A is cargo selection on endosomes in all cell types. Not modified to the more specific child GO:0036466 synaptic vesicle recycling via endosome, because that term requires vesicles endocytosed via clathrin-coated pits passing through an endosomal intermediate, and the cited paper measures vesicle composition rather than the recycling route. Supporting Evidence: PMID:15537701 Neurons express adaptor (AP)-3 complexes assembled with either ubiquitous (beta3A) or neuronal-specific (beta3B) beta3 isoforms. PMID:15537701 Our results suggest that concerted nonredundant functions of neuronal and ubiquitous AP-3 provide a mechanism to control the levels of selected membrane proteins in synaptic vesicles. PMID:15537701 Consistently, beta3B deficiency compromised synaptic zinc stores assessed by Timm's staining and the synaptic vesicle targeting of membrane proteins involved in zinc uptake (ZnT3 and ClC-3). file:human/AP3S1/AP3S1-uniprot.txt DR ComplexPortal; CPX-5055; Neuronal AP-3 Adaptor complex, sigma3a variant. |
| GO:0048490 anterograde synaptic vesicle transport | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: Anterograde synaptic vesicle transport, transferred by Ensembl Compara from the mouse ortholog. Same source experiment as the axonal-transport row. Reason: GO:0048490 is the child of GO:0008089 specific to synaptic vesicles, and both come from the same mouse IMP on PMID:21998198, so the two rows stand or fall together. The transfer is safe -- human and mouse sigma3A are identical over 193 residues -- and the biology is real: AP-3 with BLOC-1 is required to export PI4KIIalpha from neuronal cell bodies to neurites and terminals. It is non-core for the same reasons: a whole-complex null, in brain, describing the movement of carriers rather than the cargo-selection step sigma3A performs. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q9DCR2 · Ap3s1, Mus musculus sigma3A, the 1:1 ortholog of the target SUPPORTS TRANSFER One donor under two identifiers, not two: UniProt's Ensembl cross-reference on Q9DCR2 is ENSMUST00000025357.9 / ENSMUSP00000025357.8, so the accession and the Ensembl protein id name the same mouse protein. Q9DCR2 is AP3S1_MOUSE, gene Ap3s1, Swiss-Prot, 193 aa -- the 1:1 ortholog and the same subunit, not the sigma3B paralog (which is Q8BSZ2 and is the donor for AP3S2's matching rows). The two sequences are byte-identical over all 193 residues, so the transfer has no divergence to bridge. The mouse source row is IMP from PMID:21998198, and the perturbation in that paper is the mocha allele Ap3d1(mh/mh) -- a delta-subunit null that removes the whole AP-3 complex -- with sigma3 appearing as an immunoblot readout rather than as the mutated gene. Annotating an obligate-complex subunit from a complex-null phenotype is standard GO practice and is not being called a defect; it is why the claim is complex-level and neuronal-context, hence non-core for a subunit of a ubiquitously expressed adaptor. ensembl:ENSMUSP00000025357 · Ensembl mouse translation ENSMUSP00000025357 SUPPORTS TRANSFER The Ensembl protein identifier for the same mouse entity as the UniProtKB:Q9DCR2 entry above; the Compara handle, not a second donor. Supporting Evidence: PMID:21998198 Our findings indicate a novel vesicle transport mechanism requiring BLOC-1 and AP-3 complexes for cargo sorting from neuronal cell bodies to neurites and nerve terminals. PMID:21998198 PI4KIIα targeting to neuronal processes is impaired in AP-3–deficient mocha (Ap3d1mh/mh) neurons. PMID:21998198 Synaptosome fractions from control brains (lanes 1–8) and AP-3–deficient mocha (Ap3d1mh/mh) brains (lanes 1′–8′) were resolved on SDS–PAGE, and the contents were analyzed by immunoblot with antibodies against synaptic vesicle markers (SV2, synaptophysin), AP-3–dependent synaptic vesicle cargoes (PI4KIIα, VAMP7, ZnT3), and AP-3 σ3 subunit. file:human/AP3S1/AP3S1-bioinformatics/RESULTS.md there is no divergence anywhere in the protein for the transfer to have to bridge. |
| GO:0048490 anterograde synaptic vesicle transport | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: The same synaptic-vesicle-transport claim restated as a curator ISS from the same mouse ortholog. Reason: Same donor, same source experiment, same grade as the Compara row. The curator route (GO_REF:0000024) and the pipeline route (GO_REF:0000107) reach the same conclusion from the same mouse IMP; that is duplication, not corroboration, and it is legitimate in GOA. Non-core because the underlying perturbation is an AP-3 complex null in mouse brain. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q9DCR2 · Ap3s1, Mus musculus sigma3A, the 1:1 ortholog of the target SUPPORTS TRANSFER One donor under two identifiers, not two: UniProt's Ensembl cross-reference on Q9DCR2 is ENSMUST00000025357.9 / ENSMUSP00000025357.8, so the accession and the Ensembl protein id name the same mouse protein. Q9DCR2 is AP3S1_MOUSE, gene Ap3s1, Swiss-Prot, 193 aa -- the 1:1 ortholog and the same subunit, not the sigma3B paralog (which is Q8BSZ2 and is the donor for AP3S2's matching rows). The two sequences are byte-identical over all 193 residues, so the transfer has no divergence to bridge. The mouse source row is IMP from PMID:21998198, and the perturbation in that paper is the mocha allele Ap3d1(mh/mh) -- a delta-subunit null that removes the whole AP-3 complex -- with sigma3 appearing as an immunoblot readout rather than as the mutated gene. Annotating an obligate-complex subunit from a complex-null phenotype is standard GO practice and is not being called a defect; it is why the claim is complex-level and neuronal-context, hence non-core for a subunit of a ubiquitously expressed adaptor. Supporting Evidence: PMID:21998198 Our findings indicate a novel vesicle transport mechanism requiring BLOC-1 and AP-3 complexes for cargo sorting from neuronal cell bodies to neurites and nerve terminals. PMID:21998198 Synaptosome fractions from control brains (lanes 1–8) and AP-3–deficient mocha (Ap3d1mh/mh) brains (lanes 1′–8′) were resolved on SDS–PAGE, and the contents were analyzed by immunoblot with antibodies against synaptic vesicle markers (SV2, synaptophysin), AP-3–dependent synaptic vesicle cargoes (PI4KIIα, VAMP7, ZnT3), and AP-3 σ3 subunit. file:human/AP3S1/AP3S1-bioinformatics/RESULTS.md there is no divergence anywhere in the protein for the transfer to have to bridge. |
| GO:0060155 platelet dense granule organization | NAS PMID:23247405 Cell type-specific Rab32 and Rab38 cooperate with the ubiqui... | KEEP AS NON CORE | Summary: Platelet dense granule organization, from ComplexPortal, one of four terms it attaches to AP3S1 from the same Rab32/Rab38 review. Reason: The biology is solid at the complex level and is the historical reason AP-3 matters: the mocha mouse, a delta-subunit null, has defective platelets, and the paper concludes that AP-3 is responsible for cargo selection to lysosome-related organelles including platelet dense granules; mutations in AP-3 subunits cause Hermansky-Pudlak syndrome, whose bleeding diathesis is dense-granule deficiency. The row is non-core for AP3S1 because it is cell-type-restricted and because the evidence attaches to the complex rather than to this subunit: a QuickGO query by PMID:23247405 returns 178 annotations over 48 entities spanning AP-1, AP-3 and BLOC-2 subunits in human and mouse, so this is one statement projected broadly. Supporting Evidence: PMID:23247405 Mutations in subunits of AP-3, BLOC-1, BLOC-2 and BLOC-3 underlie many forms of HPS PMID:9697856 Our results demonstrate that the AP-3 complex is responsible for cargo selection to lysosome-related organelles such as melanosomes and platelet dense granules as well as to neurotransmitter vesicles. PMID:9697856 Here, we show that mocha is a null allele of the delta subunit of the adaptor-like protein complex AP-3, which is associated with coated vesicles budding from the trans-Golgi network, and that AP-3 is missing in mocha tissues. |
| GO:0098793 presynapse | IEA GO_REF:0000108 | KEEP AS NON CORE | Summary: Presynapse, inferred by the inter-ontology pipeline from this gene's GO:0016183 synaptic vesicle coating row. Reason: The inference step is sound given its input: GO_REF:0000108 derives a location from a process that occurs there, and the source process term does name the presynaptic membrane endocytic zone. The input itself is the row this review modifies (the GO:0016183 term names a plasma-membrane coated-pit process that belongs to AP-2), so the derivation would not survive the replacement -- GO:0016182 occurs on endosomes. The location is nonetheless kept on its own merits rather than removed: sigma3A is a subunit of the neuronal AP-3 variant as well as the ubiquitous one (UniProt lists ComplexPortal CPX-5055, 'Neuronal AP-3 Adaptor complex, sigma3a variant'), the neuronal complex is preferentially targeted to neuronal processes, and AP-3-dependent cargoes are depleted from synaptosomes of AP-3-null brain. Non-core: a neuronal-context localisation of a ubiquitously expressed adaptor subunit. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: GO:0016183 · GO:0016183 synaptic vesicle coating, the source term of the inter-ontology link SUPPORTS TRANSFER The source is a GO term on this same gene, not a gene product: the inter-ontology link from GO:0016183 synaptic vesicle coating, whose definition places it in the presynaptic membrane endocytic zone. The mechanical step is correct. The source row is itself MODIFYed in this review to GO:0016182 synaptic vesicle budding from endosome, which occurs on endosomes rather than at the presynapse, so if that change were adopted this derived row would no longer follow from it and would need to rest on the direct evidence cited here instead. Supporting Evidence: file:human/AP3S1/AP3S1-uniprot.txt DR ComplexPortal; CPX-5055; Neuronal AP-3 Adaptor complex, sigma3a variant. PMID:15537701 beta3B-containing AP-3 complexes were preferentially targeted to neuronal processes. PMID:21998198 Synaptosome fractions from control brains (lanes 1–8) and AP-3–deficient mocha (Ap3d1mh/mh) brains (lanes 1′–8′) were resolved on SDS–PAGE, and the contents were analyzed by immunoblot with antibodies against synaptic vesicle markers (SV2, synaptophysin), AP-3–dependent synaptic vesicle cargoes (PI4KIIα, VAMP7, ZnT3), and AP-3 σ3 subunit. PMID:15537701 Neurons express adaptor (AP)-3 complexes assembled with either ubiquitous (beta3A) or neuronal-specific (beta3B) beta3 isoforms. |
| GO:1903232 melanosome assembly | NAS PMID:23247405 Cell type-specific Rab32 and Rab38 cooperate with the ubiqui... | KEEP AS NON CORE | Summary: Melanosome assembly, from ComplexPortal, citing the Rab32/Rab38 review. True of AP-3 in pigment cells; cell-type-restricted for this subunit. Reason: Better supported than most complex-level projections, because the cargo link is mechanistic and runs through sigma3A itself: tyrosinase carries an acidic dileucine signal, that signal binds the delta-sigma3 hemicomplex, and in AP-3-deficient melanocytes tyrosinase accumulates in vacuolar and multivesicular endosomes instead of reaching melanosomes. The cited review adds that Rab32 and Rab38 recruit AP-1, AP-3 and BLOC-2 on endosome tubules for exactly this cargo. There is even a gene-level in vivo perturbation of the ap3s1 orthologue: partial knockdown in zebrafish sensitises developing melanocytes to hypopigmentation under low copper. Still non-core for a ubiquitously expressed subunit: melanosomes exist in one cell type, AP-3 and AP-1 provide partially redundant routes to them, and the reference is projected across 48 entities. The general function -- delivering dileucine cargo out of sorting endosomes -- is what is core; melanosome assembly is one of its cell-type-specific consequences. Supporting Evidence: PMID:16162817 In AP-3-deficient melanocytes, tyrosinase accumulates inappropriately in vacuolar and multivesicular endosomes. PMID:16162817 Both adaptors recognize the tyrosinase dileucine-based melanosome sorting signal, and tyrosinase largely colocalizes with each adaptor on endosomes. PMID:23247405 Rab32 and Rab38 interact with AP-1, AP-3 and BLOC-2 on early/recycling endosome tubules, where cargo such as tyrosinase and Tyrp1 are loaded into vesicles or transport intermediates. PMID:23247405 Furthermore, AP-3 and BLOC-2 define parallel pathways for melanosome biogenesis, thus, deficiency of both complexes causes a more severe defect in Tyrp1 transport and overall pigmentation than either single deficiency PMID:20713646 partial knockdown of the analogous Ap3s1 and Ap1s1 trafficking components in zebrafish sensitized developing melanocytes to hypopigmentation in low-copper environmental conditions |
| GO:1904115 axon cytoplasm | IEA GO_REF:0000108 | KEEP AS NON CORE | Summary: Axon cytoplasm, inferred by the inter-ontology pipeline from this gene's GO:0008089 anterograde axonal transport rows. Reason: A mechanical consequence of the axonal-transport annotation, and it inherits that row's grade exactly. The underlying evidence is a mouse AP-3 complex null assayed in brain, so the location is real in that context -- AP-3-dependent cargoes are depleted from AP-3-null synaptosomes and fail to reach neurites -- but it is a neuronal-context statement about a ubiquitously expressed adaptor subunit, and it carries no information beyond the process row it is derived from. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: GO:0008089 · GO:0008089 anterograde axonal transport, the source term of the inter-ontology link SUPPORTS TRANSFER The source is a GO term on this same gene, not a gene product: the inter-ontology link from GO:0008089 anterograde axonal transport, whose definition places the movement in nerve cell axons. The step is mechanically correct and inherits the grade of its input, which this review keeps as non-core because the mouse experiment behind it is an AP-3 complex null. Supporting Evidence: PMID:21998198 PI4KIIα targeting to neuronal processes is impaired in AP-3–deficient mocha (Ap3d1mh/mh) neurons. PMID:21998198 Our findings indicate a novel vesicle transport mechanism requiring BLOC-1 and AP-3 complexes for cargo sorting from neuronal cell bodies to neurites and nerve terminals. PMID:21998198 Synaptosome fractions from control brains (lanes 1–8) and AP-3–deficient mocha (Ap3d1mh/mh) brains (lanes 1′–8′) were resolved on SDS–PAGE, and the contents were analyzed by immunoblot with antibodies against synaptic vesicle markers (SV2, synaptophysin), AP-3–dependent synaptic vesicle cargoes (PI4KIIα, VAMP7, ZnT3), and AP-3 σ3 subunit. |
| GO:0005198 structural molecule activity | IDA PMID:39705307 A structure-based mechanism for initiation of AP-3 coated ve... | NEW | Summary: sigma3A is an obligate structural chain of the AP-3 core: the human complex was reconstituted from delta, beta3, mu3 and sigma3 and solved by cryo-EM with sigma3A as one of the four chains, and the delta-sigma3 pair behaves as a stable hemicomplex. Reason: GOA gives this gene exactly one molecular-function row, GO:0035651, which this review removes as a role conflation, leaving the MF aspect empty. The subunit's own contribution -- holding the complex together -- is annotated nowhere, and it is directly demonstrated in human protein. Human AP-3 was reconstituted from its four subunits including sigma3A and multiple structures were determined by electron cryo-microscopy; UniProt maps four of those PDB entries to this protein with chain S covering residues 1-193, so the solved core contains full-length AP3S1. In pull-downs of separated halves of the complex, the delta-sigma3 hemicomplex behaves as a unit, binding Arf1 nearly as well as the intact complex, while beta3-mu3 does not -- note that the Arf1 site itself is on delta, so this is evidence that sigma3 is structurally integral to that half, not that sigma3A binds Arf1. Mouse genetics corroborates the pairing from the other direction: in delta-null (mocha) cells sigma3 is left monomeric, while in beta3A-null (pearl) cells delta and sigma3 still coassemble. Coded IDA rather than IMP because the reconstitution and the structures are direct assays of the human protein; the mouse hemicomplex genetics is cited as corroboration, not as the evidence. The more interesting subunit-level function -- binding the acidic dileucine sorting signal -- gets no NEW row here because GO has no term for it; that is the proposed new term and the ONTOLOGY knowledge gap. Supporting Evidence: PMID:39705307 To understand the mechanism of AP-3 membrane recruitment and activation, we reconstituted human AP-3 and determined multiple structures in the soluble and membrane-bound states using electron cryo-microscopy. PMID:39705307 was coexpressed in Escherichia coli BL21 Star cells (Invitrogen) with HRV-cleavable glutahtion S-transferase tags on the C termini of δ and β3. PMID:39705307 Comparing the binding of AP-3 hemicomplexes, it is apparent that the δ-σ3 complex binds nearly as well as the full complex, with binding of the β3-μ3 hemicomplex barely above background levels in the Arf1GTP state. file:human/AP3S1/AP3S1-uniprot.txt DR PDB; 9C59; EM; 4.30 A; S/s=1-193. PMID:11807095 In mocha cells, the beta3 and mu3 subunits coassemble into a heterodimer, whereas the sigma3 subunit remains monomeric. PMID:11807095 In pearl cells, the delta and sigma3 subunits coassemble into a heterodimer, whereas mu3 gets destroyed. |
Loading supporting content…
Download this section (compressed HTML)Q: Is there any cargo, tissue or trafficking step that requires sigma3A specifically and cannot be served by sigma3B? Both isoforms are broadly expressed, both delta-sigma3A and delta-sigma3B bind the LIMP-II dileucine tail, and the two proteins are identical at every residue named for the cargo-signal site, so nothing currently distinguishes them functionally.
Suggested experts: Juan S. Bonifacino, Esteban C. Dell'Angelica
Q: In the cargo-free human AP-3 structure the sigma3 dileucine pocket is occupied by the N-terminal extension of beta3. Is that an autoinhibited state that a dileucine cargo has to displace, and is displacement coupled to the tyrosine-motif site on mu3?
Suggested experts: Richard W. Baker
Q: The delta ear binds sigma3 and blocks Arf binding without affecting dileucine-signal binding. Is that interaction regulated in cells -- by phosphorylation, by cargo occupancy, or by a partner -- and does the reported phosphoserine at position 191 of sigma3A play any part in AP-3 recruitment?
Suggested experts: Juan S. Bonifacino
Q: Given that AP3S1 is ubiquitously expressed and essential to a complex whose loss causes Hermansky-Pudlak syndrome through other subunits, why has no human AP3S1 disease allele been reported? Is the gene intolerant of loss of function, or is sigma3B able to compensate?
Experiment: Generate single and double knockouts of AP3S1 and AP3S2 in a pigmented human cell line (MNT-1) and in a non-pigmented line, and score AP-3 complex assembly (co-IP of delta, beta3 and mu3), membrane recruitment of delta by immunofluorescence, and steady-state sorting of tyrosinase, LAMP-1 and CD63. Then rescue the double knockout with sigma3A, with sigma3B, and with sigma3A carrying V94D or L109S. The prediction under the hypothesis is that either isoform rescues fully, the pocket mutants rescue assembly but not cargo sorting, and only the double knockout has a phenotype.
Hypothesis: sigma3A and sigma3B are functionally interchangeable for dileucine cargo selection, and any difference between them lies outside the cargo-signal site.
Type: genetic knockout with isoform swap rescue
Experiment: Reconstitute human AP-3 with a beta3 lacking the N-terminal extension and compare it with wild type for dileucine-peptide binding by ITC or fluorescence anisotropy, for Arf1-dependent recruitment to nanodiscs, and for the conformational state by cryo-EM. If the extension is autoinhibitory, its removal should raise dileucine-peptide affinity and may uncouple cargo binding from the second Arf1 site.
Hypothesis: The beta3 N-terminal extension occupies the sigma3 dileucine pocket as an autoinhibitory element that dileucine cargo must displace.
Type: in vitro reconstitution and structural analysis
Experiment: Knock in V94D or L109S at the endogenous AP3S1 locus in a cell line also lacking AP3S2, and measure: AP-3 assembly and membrane recruitment (which should be unaffected, since these residues are in the cargo site rather than the delta interface), and the sorting of a panel of dileucine cargoes (tyrosinase, LIMP-II, OCA2) versus tyrosine-motif cargoes (LAMP-1). A clean separation would establish the pocket's role in vivo and would give the proposed dileucine-signal-binding term an in-cell anchor.
Hypothesis: sigma3A contributes the cargo-signal pocket in cells, not only in yeast three-hybrid assays.
Type: knock-in point mutagenesis with cargo sorting readout
Experiment: Map the delta-ear binding surface on sigma3A by cross-linking mass spectrometry or cryo-EM of the delta-ear/sigma3 complex, then mutate that surface on sigma3A without touching the cargo pocket and ask whether AP-3 becomes constitutively membrane-bound in cells. Combine with a phospho-null and phospho-mimetic pair at sigma3A Ser191 to test whether the reported phosphorylation modulates the switch.
Hypothesis: The delta-ear/sigma3 interaction is a regulated switch for AP-3 membrane recruitment.
Type: structural mapping with separation-of-function mutants
What is not known — curated, literature-grounded statements of the open unknowns (the inverse of core functions).
Gap: Nothing is known that functionally separates sigma3A from sigma3B. The two paralogs are both broadly expressed, both delta-sigma3A and delta-sigma3B bind the LIMP-II dileucine tail, and no cargo, tissue or trafficking step has been shown to require one and not the other.
OPEN BIOLOGY RESIDUAL_SUBGAP
What is known: The shared properties are firmly established: both were identified together as AP-3 subunits and both are expressed in a wide variety of tissues and cell lines; the GO definition of the complex admits either; and the yeast three-hybrid comparison that tested them side by side found both hemicomplexes binding the same signal. The bioinformatics analysis in this review adds that the two proteins, 83.9% identical overall, are identical at every residue the literature names for the cargo-signal site. What has never been done is a side-by-side loss-of-function comparison in cells, or a test of sigma3B against the Nef signal, which does discriminate between the AP-1 and AP-3 large subunits.
Significance: GOA gives AP3S1 and AP3S2 near-identical annotation sets from the same references. That is currently an honest reflection of the evidence rather than a curation shortcut, but it means neither gene's annotation can be improved without new experiments, and any isoform-specific role -- including a possible explanation for why no human AP3S1 disease allele has been reported -- is invisible.
What would resolve it: Single and double AP3S1/AP3S2 knockouts in the same cell background, with isoform swap rescue, scored on complex assembly, membrane recruitment and cargo sorting.
Provenance (the field's own admissions):
Gap: It is not known what displaces the beta3 N-terminal extension from the sigma3 dileucine pocket, or whether that occupancy is a genuine autoinhibited state of human AP-3 rather than a feature of the reconstituted, cargo-free complex.
OPEN BIOLOGY MF_DARK
What is known: The structural observation is solid and unexpected: in the cargo-engaged human AP-3 structure the tyrosine motif is in the mu3 pocket while the dileucine pocket on sigma3 is filled by the N-terminal extension of beta3, and the authors offer the alternative that full activation requires ejection of that tail. AP-1 and AP-2 have the equivalent extensions and release them on activation. What is missing is any measurement of dileucine-peptide binding to the assembled complex with and without the beta3 extension, and any test of whether the two cargo sites are coupled.
Significance: If the beta3 tail is autoinhibitory, then sigma3A's cargo site is a regulated switch rather than a constitutively available pocket, and AP-3 cargo selection would be ordered rather than simultaneous. That changes what a contributes_to cargo-adaptor-activity annotation on sigma3A actually means in the assembled complex.
What would resolve it: Quantitative dileucine-peptide binding to reconstituted AP-3 with and without the beta3 N-terminal extension, plus a structure of the complex with a dileucine cargo bound.
Provenance (the field's own admissions):
Gap: Whether AP3S1 has any role outside AP-3 is undetermined. The one published claim of an independent activity -- that sigma3A acts as a membrane receptor for IRS-1 in adipocytes -- has never been tested again.
OPEN BIOLOGYCURATION BP_DARK
What is known: The original observation is specific: a sigma3A cDNA came out of an expression-library screen with IRS-1, GST-sigma3A bound IRS-1 in vitro, and adding GST-sigma3A released IRS-1 from low-density membranes. The authors framed the functional interpretation as a hypothesis. Since then the paper has accumulated 5 PMC-indexed citations, UniProt's SUBUNIT line for Q92572 records AGAP1 and BLOC-1 but not IRS1, and no AP-3-independent pool of sigma3A has been described. Against that, the assembly genetics show that sigma3 is left monomeric when delta is absent, so a free monomeric pool does exist in at least one genetic background.
Significance: GOA carries an insulin-receptor-signalling annotation on this gene that rests entirely on that single study. Either it should be supported by modern evidence or it should be retired; at present it is the only claim on the gene that does not connect to AP-3.
What would resolve it: Test IRS-1 association with endogenous sigma3A in adipocytes by proximity labelling or co-IP, in wild type and in AP3D1-null cells where sigma3 is monomeric, and ask whether insulin signalling output changes on AP3S1 knockout.
Provenance (the field's own admissions):
Gap: What decides whether a given acidic dileucine signal is read by the AP-1 gamma1-sigma1A hemicomplex or by the AP-3 delta-sigma3 hemicomplex is not understood, and the sigma-side contribution to that choice has not been isolated.
OPEN BIOLOGY MF_DARK
What is known: The phenomenon is well documented. Several cargoes bind both adaptors -- tyrosinase binds gamma1-sigma1A and delta-sigma3 -- while single-residue changes in an OCA2 signal can be engineered to favour one adaptor over the other, and only AP-3 binding gives steady-state melanosome localisation. Discrimination also runs in the other direction: of five Nef variants tested, all bound gamma1-sigma1A but only two bound delta-sigma3A. Mutagenesis has mapped the shared site onto all three complexes and shown which residues matter for which signal, but the pockets are nearly identical between sigma1A and sigma3A, so the determinant is most likely in the large subunit or in accessibility rather than in the sigma pocket itself. That has not been tested with chimeric hemicomplexes.
Significance: Adaptor choice is what routes a membrane protein to a lysosome rather than to a melanosome or back to the surface. Until the determinant is known, the specificity of a sigma subunit cannot be predicted from sequence -- which is exactly the inference an automated pipeline would attempt -- and AP-1 and AP-3 cargo assignments have to be made one cargo at a time.
What would resolve it: Test chimeric hemicomplexes pairing gamma1 with sigma3A and delta with sigma1A against a panel of signals with known adaptor preference, and ask which subunit carries the preference.
Provenance (the field's own admissions):
Gap: The delta-ear/sigma3 interaction that gates AP-3 membrane recruitment has no known regulator: it is not known what relieves it in cells, nor where on sigma3A the ear binds.
OPEN BIOLOGY RESIDUAL_SUBGAP
What is known: The interaction itself is characterised and is unusually specific in its effect: the ear domain of delta binds sigma3, this blocks AP-3 binding to Arf while leaving dileucine-signal binding intact, and the isolated delta ear inhibits AP-3 recruitment to membranes in vitro and in vivo and impairs lysosomal membrane protein sorting. So the switch acts on the recruitment step, not on cargo recognition. What is missing is the binding surface on sigma3A, any physiological input that toggles it, and whether the phosphoserine UniProt records at sigma3A position 191 is involved.
Significance: If this is a regulated switch, sigma3A is not only a cargo-binding and structural subunit but also the target of AP-3's recruitment control, which would be a third, currently unannotatable role for the protein. It also matters for interpreting AP-3's constitutively open conformation: an open complex still has to be kept off membranes somehow.
What would resolve it: Map the delta-ear footprint on sigma3A, then make separation-of-function sigma3A mutants that lose ear binding but keep the cargo pocket, and test whether AP-3 becomes constitutively membrane-associated.
Provenance (the field's own admissions):
Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)