APPL2 (DCC-interacting protein 13-beta, DIP13B) is a 664-residue cytoplasmic adaptor protein built from an N-terminal BAR domain, a central pleckstrin-homology domain and a C-terminal phosphotyrosine-binding domain, with no catalytic activity of its own. The BAR domain drives self-association into crescent-shaped dimers and also heterodimerisation with its paralog APPL1, while the PH and PTB domains bind phosphoinositides; together these give the protein a curvature-sensing, lipid-binding scaffold. APPL2 is an effector of small GTPases of the Rab5 branch, binding GTP-loaded Rab5, Rab22A, Rab24 and Rab31 - the last being the only one whose affinity has been measured, a dissociation constant of 140 nM - and this is what recruits it to membranes. On a peripheral subpopulation of early endosomes, called APPL endosomes for the two adaptors that mark them, APPL2 acts together with Rab5 and annexin A2 as part of a signalling platform that links internalised receptors to the nucleus; a pool of the protein shuttles into the nucleus and associates with the NuRD/MeCP1 chromatin-remodelling complex and with a Reptin-beta-catenin-HDAC repressive complex, where it promotes beta-catenin/TCF-dependent transcription. In muscle, APPL2 is a brake on metabolic signalling: it binds the adiponectin receptors AdipoR1 and AdipoR2 in competition with APPL1 and sequesters APPL1 away from the adiponectin and insulin pathways, and it binds phosphorylated TBC1D1 to block insulin-evoked GLUT4 translocation and glucose uptake, so losing APPL2 improves insulin sensitivity and glucose tolerance. In macrophages the protein instead occupies surface ruffles, macropinosomes and early phagosomes, where Rab31 recruits it to promote phagocytic cup closure and Fc-gamma-receptor-driven PI3K/Akt signalling, while downstream of TLR4 it restrains Akt and NF-kappaB activation and limits cytokine secretion. Further roles have been described in hypothalamic control of adipose beiging and cold tolerance, in hepatocyte-growth-factor-dependent fibroblast migration, and in tuning glucocorticoid-receptor sensitivity during adult neurogenesis. Across these settings APPL2 repeatedly behaves as the counterweight to APPL1 rather than its duplicate: the two proteins are only about 54% identical, bind different Rabs, and have opposite effects on adiponectin signalling and on Akt activation downstream of several receptors.
Definition: A process that stops, prevents or reduces the frequency, rate or extent of the adiponectin-activated signaling pathway. GO currently has GO:0033211 adiponectin-activated signaling pathway but no signed children: a QuickGO ontology search for 'regulation of adiponectin' returns only the secretion branch (GO:0070163, GO:0070164, GO:0070165) and GO:0055100 adiponectin binding. APPL2's best-characterised metabolic role is specifically to antagonise this pathway - it competes with APPL1 for AdipoR1/AdipoR2 and sequesters APPL1 - and that direction currently cannot be expressed, so the annotation has to fall back on the neutral parent.
Justification: Without a signed child, an adiponectin antagonist and an adiponectin potentiator receive the same GO term. APPL1 and APPL2 are exactly that pair, and both are annotated to GO:0033211.
Parent term: adiponectin-activated signaling pathway
Supporting Evidence:
Definition: A subpopulation of early endosomes marked by the Rab5 effectors APPL1 and APPL2, distinguishable from canonical EEA1-positive early endosomes by limited EEA1 colocalisation, a distinct migration pattern in continuous density gradients, a peripheral distribution beneath the plasma membrane, and dependence on annexin A2 as well as Rab5 for membrane recruitment. Would be a child of GO:0005769 early endosome.
Justification: The compartment has been defined biochemically and by imaging since 2004 and is referred to by name throughout the field, but the most specific GO term available for APPL2 is GO:0031901 early endosome membrane, which does not distinguish it from the EEA1-positive population the authors explicitly separate it from.
Parent term: early endosome
Supporting Evidence:
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0001726 ruffle | IEA GO_REF:0000120 | KEEP AS NON CORE | Summary: APPL2 is on cell-surface ruffles in macrophages; the ruffle call is transferred from mouse Appl2. Reason: Endogenous and GFP-tagged APPL2 were immunolabelled on cell-surface ruffles in mouse primary macrophages and RAW264.7 cells (PMID:25568335), which is the observation behind the mouse annotation. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. It is kept as non-core because it is a macrophage-specific membrane domain seen during phagocytosis, not where APPL2 sits in the cell types where its signalling functions were characterised. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse observation is immunofluorescence of endogenous and tagged APPL2 on macrophage surface ruffles (PMID:25568335); it transfers, but only as a cell-type-specific location. ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. UniProtKB-SubCell:SL-0300 · UniProt subcellular-location keyword: Ruffle SUPPORTS TRANSFER Projection of UniProt's own curated subcellular-location keyword, so this row re-states the Swiss-Prot annotation rather than adding independent evidence. The mouse observation is immunofluorescence of endogenous and tagged APPL2 on macrophage surface ruffles (PMID:25568335); it transfers, but only as a cell-type-specific location. Supporting Evidence: PMID:25568335 APPL2 was immunolabeled in the cytoplasm, on cell surface ruffles, and on early, actin-rich phagosomes file:human/APPL2/APPL2-bioinformatics/RESULTS.md Human and mouse APPL2 are 92.7% identical over a full-length global alignment |
| GO:0001726 ruffle | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: APPL2 is on cell-surface ruffles in macrophages; the ruffle call is transferred from mouse Appl2. Reason: Endogenous and GFP-tagged APPL2 were immunolabelled on cell-surface ruffles in mouse primary macrophages and RAW264.7 cells (PMID:25568335), which is the observation behind the mouse annotation. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. It is kept as non-core because it is a macrophage-specific membrane domain seen during phagocytosis, not where APPL2 sits in the cell types where its signalling functions were characterised. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse observation is immunofluorescence of endogenous and tagged APPL2 on macrophage surface ruffles (PMID:25568335); it transfers, but only as a cell-type-specific location. Supporting Evidence: PMID:25568335 APPL2 was immunolabeled in the cytoplasm, on cell surface ruffles, and on early, actin-rich phagosomes file:human/APPL2/APPL2-bioinformatics/RESULTS.md Human and mouse APPL2 are 92.7% identical over a full-length global alignment |
| GO:0001786 phosphatidylserine binding | IDA PMID:18034774 Membrane targeting by APPL1 and APPL2: dynamic scaffolds tha... | KEEP AS NON CORE | Summary: Phosphatidylserine binding was seen only for the isolated PTB domain on a lipid strip, not for full-length APPL2. Reason: On PIP Strips the isolated APPL PTB domains bound phosphatidylserine, but the full-length proteins did not, and the authors explicitly flag the discrepancy. This is a real property of the domain in vitro and it is consistent with the PTB domain's role in membrane targeting, but it has not been shown for the intact protein and no cellular consequence has been attached to it. Kept, not removed, and marked non-core: the phosphoinositide binding of the full-length protein (GO:0035091) is the better-supported lipid-binding claim. Supporting Evidence: PMID:18034774 APPL isolated PTB domains also bound to membrane-immobilized phosphatidylserine (PS, spot 15). PMID:18034774 It is possible that APPL-APPL interactions and/or the overall conformation of the full-length proteins may inhibit PTB domain-mediated PS binding. |
| GO:0001786 phosphatidylserine binding | IEA GO_REF:0000117 | KEEP AS NON CORE | Summary: ARBA rule reproducing the isolated-domain phosphatidylserine-binding result. Reason: ARBA00089669 is a narrow rule (5 condition sets over FunFam signatures in primates and fungi) that assigns the same term the IDA row carries. It adds no evidence beyond the IDA, and inherits the same caveat: the binding was demonstrated for the isolated PTB domain, not full-length APPL2. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: ARBA:ARBA00089669 · ARBA rule ARBA00089669 SUPPORTS TRANSFER ARBA rule for phosphatidylserine binding; 5 condition sets over FunFam signatures in primates and fungi. Narrow, and it reproduces a conclusion the IDA row already carries. Supporting Evidence: PMID:18034774 APPL isolated PTB domains also bound to membrane-immobilized phosphatidylserine (PS, spot 15). |
| GO:0002024 diet induced thermogenesis | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: diet induced thermogenesis rests on a neuron-restricted Appl2 knockout in mouse hypothalamus. Reason: Deleting APPL2 in ventromedial-hypothalamic RIP-Cre neurons reduced beiging of subcutaneous white adipose tissue and produced cold intolerance and obesity, acting through AMPK and sympathetic outflow (PMID:29467283). The perturbation is APPL2-specific and conditional. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. This is nonetheless an organism-level physiological outcome of APPL2 acting in one neuronal population, several steps removed from what the protein does molecularly, so it is peripheral rather than core. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse evidence is a conditional, neuron-restricted APPL2 knockout with a whole-animal thermogenic phenotype (PMID:29467283) - APPL2-specific, but organismal. ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. Supporting Evidence: PMID:29467283 Genetic ablation of APPL2 in RIP-Cre neurons diminishes beiging in sWAT without affecting BAT, leading to cold intolerance and obesity in mice. PMID:29467283 Hypothalamic APPL2 enhances neuronal activation in VMH RIP-Cre neurons and raphe pallidus, thereby eliciting SNS outflow to sWAT and subsequent beiging. |
| GO:0002024 diet induced thermogenesis | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: diet induced thermogenesis rests on a neuron-restricted Appl2 knockout in mouse hypothalamus. Reason: Deleting APPL2 in ventromedial-hypothalamic RIP-Cre neurons reduced beiging of subcutaneous white adipose tissue and produced cold intolerance and obesity, acting through AMPK and sympathetic outflow (PMID:29467283). The perturbation is APPL2-specific and conditional. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. This is nonetheless an organism-level physiological outcome of APPL2 acting in one neuronal population, several steps removed from what the protein does molecularly, so it is peripheral rather than core. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse evidence is a conditional, neuron-restricted APPL2 knockout with a whole-animal thermogenic phenotype (PMID:29467283) - APPL2-specific, but organismal. Supporting Evidence: PMID:29467283 Genetic ablation of APPL2 in RIP-Cre neurons diminishes beiging in sWAT without affecting BAT, leading to cold intolerance and obesity in mice. PMID:29467283 Hypothalamic APPL2 enhances neuronal activation in VMH RIP-Cre neurons and raphe pallidus, thereby eliciting SNS outflow to sWAT and subsequent beiging. |
| GO:0005515 protein binding | IPI PMID:15016378 APPL proteins link Rab5 to nuclear signal transduction via a... | MODIFY | Summary: Bare protein binding for a Rab GTPase partner (RAB5A); the informative term is small GTPase binding. Reason: APPL2 is a Rab effector: it binds GTP-loaded Rab5, Rab22A, Rab24 and Rab31 in a screen against 46 Rabs, binds purified Rab31 directly with a preference for the GTP form, and the APPL2 BAR-PH dimer binds two Rab31 molecules. 'Protein binding' says none of this. GO:0031267 small GTPase binding is the current term for it - the former GO:0017137 'Rab GTPase binding' is listed among GO:0031267's secondary ids, i.e. it was merged into small GTPase binding rather than obsoleted, so GO:0031267 is the id to use. Proposed replacements: small GTPase binding Supporting Evidence: PMID:15016378 This pathway operates via APPL1 and APPL2, two Rab5 effectors, which reside on a subpopulation of endosomes. PMID:15016378 Both APPL1 and APPL2 are essential for cell proliferation and their function requires Rab5 binding. |
| GO:0005515 protein binding | IPI PMID:15016378 APPL proteins link Rab5 to nuclear signal transduction via a... | MODIFY | Summary: Bare protein binding for the APPL1 partner; the interaction is a defined BAR-domain heterodimer. Reason: APPL1 is by far the most frequently reported APPL2 partner in this GOA record, and the interaction is not a generic one: the minimal BAR domains are necessary and sufficient for APPL1-APPL2 heterodimerisation in yeast two-hybrid and the proteins reciprocally coimmunoprecipitate. GO:0046982 protein heterodimerization activity states that, where 'protein binding' does not. The heterodimer is also the mechanistic substrate of the adiponectin antagonism, in which APPL2 sequesters APPL1. Proposed replacements: protein heterodimerization activity Supporting Evidence: PMID:18034774 the APPL minimal BAR domains were necessary and sufficient for mediating APPL1-APPL1, APPL2-APPL2 and APPL1-APPL2 interactions in the yeast two-hybrid system PMID:18034774 APPL2-V5-His specifically coimmunoprecipitated with APPL1-YFP and vice versa PMID:15016378 This pathway operates via APPL1 and APPL2, two Rab5 effectors, which reside on a subpopulation of endosomes. |
| GO:0005515 protein binding | IPI PMID:16189514 Towards a proteome-scale map of the human protein-protein in... | MODIFY | Summary: Bare protein binding for the APPL1 partner; the interaction is a defined BAR-domain heterodimer. Reason: APPL1 is by far the most frequently reported APPL2 partner in this GOA record, and the interaction is not a generic one: the minimal BAR domains are necessary and sufficient for APPL1-APPL2 heterodimerisation in yeast two-hybrid and the proteins reciprocally coimmunoprecipitate. GO:0046982 protein heterodimerization activity states that, where 'protein binding' does not. The heterodimer is also the mechanistic substrate of the adiponectin antagonism, in which APPL2 sequesters APPL1. Proposed replacements: protein heterodimerization activity Supporting Evidence: PMID:18034774 the APPL minimal BAR domains were necessary and sufficient for mediating APPL1-APPL1, APPL2-APPL2 and APPL1-APPL2 interactions in the yeast two-hybrid system PMID:16189514 Using a stringent, high-throughput yeast two-hybrid system, we tested pairwise interactions |
| GO:0005515 protein binding | IPI PMID:16189514 Towards a proteome-scale map of the human protein-protein in... | MODIFY | Summary: Bare protein binding for a Rab GTPase partner (RAB22A); the informative term is small GTPase binding. Reason: APPL2 is a Rab effector: it binds GTP-loaded Rab5, Rab22A, Rab24 and Rab31 in a screen against 46 Rabs, binds purified Rab31 directly with a preference for the GTP form, and the APPL2 BAR-PH dimer binds two Rab31 molecules. 'Protein binding' says none of this. GO:0031267 small GTPase binding is the current term for it - the former GO:0017137 'Rab GTPase binding' is listed among GO:0031267's secondary ids, i.e. it was merged into small GTPase binding rather than obsoleted, so GO:0031267 is the id to use. Proposed replacements: small GTPase binding Supporting Evidence: PMID:23055524 For APPL2, the yeast two-hybrid screen revealed interactions with Rabs 5, 22a, 24, and 31 PMID:16189514 Using a stringent, high-throughput yeast two-hybrid system, we tested pairwise interactions |
| GO:0005515 protein binding | IPI PMID:17030088 APPL1, APPL2, Akt2 and FOXO1a interact with FSHR in a potent... | KEEP AS NON CORE | Summary: Interaction with the FSH receptor, reported alongside APPL1 in a receptor-proximal complex. Reason: FSHR was shown to interact with APPL2 as well as APPL1, independently of FSH stimulation, in a putative FSHR signalling complex. The same paper reports the one clean functional distinction between the paralogs - APPL1 binds Akt2 and APPL2 does not - which is why this reference matters beyond the interaction itself. No APPL2-specific functional consequence of the FSHR interaction has been established, and GO has no more informative molecular-function term for it, so the row is kept as a peripheral protein-binding record rather than re-termed. Supporting Evidence: PMID:17030088 This report shows that, in addition to APPL1, FSHR interacts with FOXO1a and APPL2. PMID:17030088 APPL1 associates with Akt2, whereas APPL2 does not. This is the first documented difference in function between APPL1 and APPL2. |
| GO:0005515 protein binding | IPI PMID:17030088 APPL1, APPL2, Akt2 and FOXO1a interact with FSHR in a potent... | MODIFY | Summary: Bare protein binding for the APPL1 partner; the interaction is a defined BAR-domain heterodimer. Reason: APPL1 is by far the most frequently reported APPL2 partner in this GOA record, and the interaction is not a generic one: the minimal BAR domains are necessary and sufficient for APPL1-APPL2 heterodimerisation in yeast two-hybrid and the proteins reciprocally coimmunoprecipitate. GO:0046982 protein heterodimerization activity states that, where 'protein binding' does not. The heterodimer is also the mechanistic substrate of the adiponectin antagonism, in which APPL2 sequesters APPL1. Proposed replacements: protein heterodimerization activity Supporting Evidence: PMID:17030088 APPL1 and APPL2 associate with one another via the N-terminus of APPL1, presumably via the Bin-Amphiphysin-Rvs (BAR) domain. PMID:17030088 APPL1 associates with Akt2, whereas APPL2 does not. This is the first documented difference in function between APPL1 and APPL2. |
| GO:0005515 protein binding | IPI PMID:18034774 Membrane targeting by APPL1 and APPL2: dynamic scaffolds tha... | MODIFY | Summary: Bare protein binding for the APPL1 partner; the interaction is a defined BAR-domain heterodimer. Reason: APPL1 is by far the most frequently reported APPL2 partner in this GOA record, and the interaction is not a generic one: the minimal BAR domains are necessary and sufficient for APPL1-APPL2 heterodimerisation in yeast two-hybrid and the proteins reciprocally coimmunoprecipitate. GO:0046982 protein heterodimerization activity states that, where 'protein binding' does not. The heterodimer is also the mechanistic substrate of the adiponectin antagonism, in which APPL2 sequesters APPL1. Proposed replacements: protein heterodimerization activity Supporting Evidence: PMID:18034774 the APPL minimal BAR domains were necessary and sufficient for mediating APPL1-APPL1, APPL2-APPL2 and APPL1-APPL2 interactions in the yeast two-hybrid system PMID:18034774 APPL2-V5-His specifically coimmunoprecipitated with APPL1-YFP and vice versa |
| GO:0005515 protein binding | IPI PMID:19433865 Endosomal adaptor proteins APPL1 and APPL2 are novel activat... | KEEP AS NON CORE | Summary: Membership of an endogenous Reptin/beta-catenin/HDAC1/HDAC2 complex together with APPL1. Reason: This row's WITH/FROM is the set CTNNB1, HDAC1, HDAC2 and APPL1 - it records co-membership of a nuclear repressive complex rather than a pairwise interaction. The complex is endogenous and the functional consequence is real: overexpressing APPL2 reduces the HDAC1/HDAC2 and beta-catenin associated with Reptin. It is non-core because the APPL2 evidence in this paper is entirely overexpression-based - APPL2 is practically undetectable in the HEK293 cells used, so no loss-of-function control was possible. Supporting Evidence: PMID:19433865 APPL proteins are present in an endogenous complex containing Reptin, beta-catenin, HDAC1, and HDAC2. PMID:19433865 overexpression of either APPL protein clearly reduced the levels of HDAC1 and HDAC2 associated with Reptin PMID:19433865 Since APPL2 is practically undetectable in HEK293 cells, we could not test the effects of its knockdown on transcription. |
| GO:0005515 protein binding | IPI PMID:19433865 Endosomal adaptor proteins APPL1 and APPL2 are novel activat... | KEEP AS NON CORE | Summary: Direct interaction with RUVBL2/Reptin, the mechanistic basis of APPL2's effect on beta-catenin/TCF transcription. Reason: Both APPL proteins bind Reptin directly, shown by in vitro binding of purified recombinant proteins as well as coimmunoprecipitation and GST pull-down, and the interaction relieves Reptin-mediated repression of beta-catenin/TCF transcription. This is one of the few APPL2 protein-binding rows with a mapped downstream consequence. It stays as protein binding because GO offers no molecular-function term for 'binds a transcriptional corepressor and displaces its partners'; the functional content is captured instead by the NEW GO:0090263 row. Supporting Evidence: PMID:19433865 Both APPL proteins interact directly with Reptin, a transcriptional repressor binding to beta-catenin and HDAC1 (histone deacetylase 1) PMID:19433865 this repression was relieved in a dose-dependent manner by adding increasing amounts of APPL1 or APPL2 PMID:19433865 overexpression of either APPL protein caused an increased association of β-catenin with the cyclin D1 and Axin2 promoters with respect to mock-transfected cells |
| GO:0005515 protein binding | IPI PMID:21645192 Biochemical characterization of APPL endosomes: the role of ... | KEEP AS NON CORE | Summary: Interaction with annexin A2, which is required to keep APPL2 on endosomal membranes. Reason: Annexin A2 was identified as an APPL1 and APPL2 partner by in vivo biotinylation and confirmed by GST pull-down, and silencing it solubilises APPL2 from endosomes and can substitute for Rab5 in recruiting APPL proteins to membranes. The interaction is therefore functionally consequential - it is part of how APPL2 gets to the compartment - but the annotation itself is a binding record, and the compartment claim is carried by the GO:0005768 and GO:0010008 rows from the same paper. Supporting Evidence: PMID:21645192 identified Annexin A2 as an interacting partner of both APPL1 and APPL2 PMID:21645192 silencing of its expression causes solubilization of APPL2 from endosomes PMID:21645192 Annexin A2 is essential for targeting APPL2 to endosomes and can compensate for Rab5 deficiency in mediating APPL membrane recruitment |
| GO:0005515 protein binding | IPI PMID:23055524 Membrane curvature protein exhibits interdomain flexibility ... | MODIFY | Summary: Bare protein binding for a Rab GTPase partner (RAB31); the informative term is small GTPase binding. Reason: APPL2 is a Rab effector: it binds GTP-loaded Rab5, Rab22A, Rab24 and Rab31 in a screen against 46 Rabs, binds purified Rab31 directly with a preference for the GTP form, and the APPL2 BAR-PH dimer binds two Rab31 molecules. 'Protein binding' says none of this. GO:0031267 small GTPase binding is the current term for it - the former GO:0017137 'Rab GTPase binding' is listed among GO:0031267's secondary ids, i.e. it was merged into small GTPase binding rather than obsoleted, so GO:0031267 is the id to use. Proposed replacements: small GTPase binding Supporting Evidence: PMID:25568335 Using purified APPL2 after cleavage of GST, together with GST-Rab31, we show direct binding of these two proteins. PMID:23055524 two Rab31 molecules bind one hAPPL2 BARPH dimer |
| GO:0005515 protein binding | IPI PMID:23414517 A human skeletal muscle interactome centered on proteins inv... | MODIFY | Summary: Bare protein binding for the APPL1 partner; the interaction is a defined BAR-domain heterodimer. Reason: APPL1 is by far the most frequently reported APPL2 partner in this GOA record, and the interaction is not a generic one: the minimal BAR domains are necessary and sufficient for APPL1-APPL2 heterodimerisation in yeast two-hybrid and the proteins reciprocally coimmunoprecipitate. GO:0046982 protein heterodimerization activity states that, where 'protein binding' does not. The heterodimer is also the mechanistic substrate of the adiponectin antagonism, in which APPL2 sequesters APPL1. Proposed replacements: protein heterodimerization activity Supporting Evidence: PMID:18034774 the APPL minimal BAR domains were necessary and sufficient for mediating APPL1-APPL1, APPL2-APPL2 and APPL1-APPL2 interactions in the yeast two-hybrid system PMID:23414517 We undertook a large-scale study using two-hybrid screens and a human skeletal-muscle cDNA library to establish a proteome-scale map of protein-protein interactions centered on proteins involved in limb-girdle muscular dystrophies (LGMD). |
| GO:0005515 protein binding | IPI PMID:23455924 A Y2H-seq approach defines the human protein methyltransfera... | MARK AS OVER ANNOTATED | Summary: Systematic-screen interaction with SUV39H2, with no functional follow-up for APPL2. Reason: This row comes from a proteome-scale screen. The SUV39H2 interaction has never been followed up for APPL2: there is no localisation, no mapped binding surface, no perturbation and no pathway placement connecting it to anything APPL2 is known to do. Bare GO:0005515 rows of this kind are reproducible observations rather than statements about function, and marking them over-annotated is the accurate call - it does not dispute that the screen detected the pair. Note also that IntAct's experiment counts for APPL2 pool replicate experiments from the same studies and are not counts of independent publications. Supporting Evidence: PMID:23455924 we developed a yeast two-hybrid interaction screening approach involving short-read second-generation sequencing (Y2H-seq) with improved sensitivity and a quantitative scoring readout allowing rapid interaction validation |
| GO:0005515 protein binding | IPI PMID:24879834 The adaptor protein APPL2 inhibits insulin-stimulated glucos... | KEEP AS NON CORE | Summary: Phospho-dependent interaction with mouse Tbc1d1 (Q60949), the mechanism behind APPL2's inhibition of insulin-stimulated glucose uptake. Reason: TBC1D1 was identified as an APPL2 partner and the interaction was mapped: insulin-driven phosphorylation of TBC1D1 Ser-235 enhances binding to the APPL2 BAR domain, and that binding suppresses TBC1D1 Thr-596 phosphorylation, blocking GLUT4 translocation. This is the best-characterised APPL2 interaction of all. It is kept rather than re-termed because the functional content is already carried, at the right level, by the GO:0046325 IMP and GO:0042593 IMP rows; 'protein binding' here is the interaction record underneath them. Supporting Evidence: PMID:24879834 We identified the Rab-GTPase-activating protein TBC1D1 as an interacting partner of APPL2. PMID:24879834 Insulin stimulates TBC1D1 phosphorylation on serine 235, leading to enhanced interaction with the BAR domain of APPL2, which in turn suppresses insulin-evoked TBC1D1 phosphorylation on threonine 596 in cultured myotubes and skeletal muscle. PMID:24879834 the APPL2-TBC1D1 interaction is a key step to fine tune insulin-stimulated glucose uptake by regulating the membrane recruitment of GLUT4 in skeletal muscle. |
| GO:0005515 protein binding | IPI PMID:24879834 The adaptor protein APPL2 inhibits insulin-stimulated glucos... | KEEP AS NON CORE | Summary: Phospho-dependent interaction with human TBC1D1 (Q86TI0), the mechanism behind APPL2's inhibition of insulin-stimulated glucose uptake. Reason: TBC1D1 was identified as an APPL2 partner and the interaction was mapped: insulin-driven phosphorylation of TBC1D1 Ser-235 enhances binding to the APPL2 BAR domain, and that binding suppresses TBC1D1 Thr-596 phosphorylation, blocking GLUT4 translocation. This is the best-characterised APPL2 interaction of all. It is kept rather than re-termed because the functional content is already carried, at the right level, by the GO:0046325 IMP and GO:0042593 IMP rows; 'protein binding' here is the interaction record underneath them. Supporting Evidence: PMID:24879834 We identified the Rab-GTPase-activating protein TBC1D1 as an interacting partner of APPL2. PMID:24879834 Insulin stimulates TBC1D1 phosphorylation on serine 235, leading to enhanced interaction with the BAR domain of APPL2, which in turn suppresses insulin-evoked TBC1D1 phosphorylation on threonine 596 in cultured myotubes and skeletal muscle. PMID:24879834 the APPL2-TBC1D1 interaction is a key step to fine tune insulin-stimulated glucose uptake by regulating the membrane recruitment of GLUT4 in skeletal muscle. |
| GO:0005515 protein binding | IPI PMID:24879834 The adaptor protein APPL2 inhibits insulin-stimulated glucos... | MODIFY | Summary: Bare protein binding for the APPL1 partner; the interaction is a defined BAR-domain heterodimer. Reason: APPL1 is by far the most frequently reported APPL2 partner in this GOA record, and the interaction is not a generic one: the minimal BAR domains are necessary and sufficient for APPL1-APPL2 heterodimerisation in yeast two-hybrid and the proteins reciprocally coimmunoprecipitate. GO:0046982 protein heterodimerization activity states that, where 'protein binding' does not. The heterodimer is also the mechanistic substrate of the adiponectin antagonism, in which APPL2 sequesters APPL1. Proposed replacements: protein heterodimerization activity Supporting Evidence: PMID:24879834 We identified the Rab-GTPase-activating protein TBC1D1 as an interacting partner of APPL2. PMID:18034774 the APPL minimal BAR domains were necessary and sufficient for mediating APPL1-APPL1, APPL2-APPL2 and APPL1-APPL2 interactions in the yeast two-hybrid system |
| GO:0005515 protein binding | IPI PMID:25416956 A proteome-scale map of the human interactome network. | MODIFY | Summary: Bare protein binding for a Rab GTPase partner (RAB5C); the informative term is small GTPase binding. Reason: APPL2 is a Rab effector: it binds GTP-loaded Rab5, Rab22A, Rab24 and Rab31 in a screen against 46 Rabs, binds purified Rab31 directly with a preference for the GTP form, and the APPL2 BAR-PH dimer binds two Rab31 molecules. 'Protein binding' says none of this. GO:0031267 small GTPase binding is the current term for it - the former GO:0017137 'Rab GTPase binding' is listed among GO:0031267's secondary ids, i.e. it was merged into small GTPase binding rather than obsoleted, so GO:0031267 is the id to use. Proposed replacements: small GTPase binding Supporting Evidence: PMID:23055524 For APPL2, the yeast two-hybrid screen revealed interactions with Rabs 5, 22a, 24, and 31 PMID:25416956 Here, we describe a systematic map of ?14,000 high-quality human binary protein-protein interactions. |
| GO:0005515 protein binding | IPI PMID:25416956 A proteome-scale map of the human interactome network. | MARK AS OVER ANNOTATED | Summary: Systematic-screen interaction with CRADD, with no functional follow-up for APPL2. Reason: This row comes from a proteome-scale screen. The CRADD interaction has never been followed up for APPL2: there is no localisation, no mapped binding surface, no perturbation and no pathway placement connecting it to anything APPL2 is known to do. Bare GO:0005515 rows of this kind are reproducible observations rather than statements about function, and marking them over-annotated is the accurate call - it does not dispute that the screen detected the pair. Note also that IntAct's experiment counts for APPL2 pool replicate experiments from the same studies and are not counts of independent publications. Supporting Evidence: PMID:25416956 Here, we describe a systematic map of ?14,000 high-quality human binary protein-protein interactions. |
| GO:0005515 protein binding | IPI PMID:25416956 A proteome-scale map of the human interactome network. | MARK AS OVER ANNOTATED | Summary: Systematic-screen interaction with SUV39H2, with no functional follow-up for APPL2. Reason: This row comes from a proteome-scale screen. The SUV39H2 interaction has never been followed up for APPL2: there is no localisation, no mapped binding surface, no perturbation and no pathway placement connecting it to anything APPL2 is known to do. Bare GO:0005515 rows of this kind are reproducible observations rather than statements about function, and marking them over-annotated is the accurate call - it does not dispute that the screen detected the pair. Note also that IntAct's experiment counts for APPL2 pool replicate experiments from the same studies and are not counts of independent publications. Supporting Evidence: PMID:25416956 Here, we describe a systematic map of ?14,000 high-quality human binary protein-protein interactions. |
| GO:0005515 protein binding | IPI PMID:25416956 A proteome-scale map of the human interactome network. | MODIFY | Summary: Bare protein binding for the APPL1 partner; the interaction is a defined BAR-domain heterodimer. Reason: APPL1 is by far the most frequently reported APPL2 partner in this GOA record, and the interaction is not a generic one: the minimal BAR domains are necessary and sufficient for APPL1-APPL2 heterodimerisation in yeast two-hybrid and the proteins reciprocally coimmunoprecipitate. GO:0046982 protein heterodimerization activity states that, where 'protein binding' does not. The heterodimer is also the mechanistic substrate of the adiponectin antagonism, in which APPL2 sequesters APPL1. Proposed replacements: protein heterodimerization activity Supporting Evidence: PMID:18034774 the APPL minimal BAR domains were necessary and sufficient for mediating APPL1-APPL1, APPL2-APPL2 and APPL1-APPL2 interactions in the yeast two-hybrid system PMID:25416956 Here, we describe a systematic map of ?14,000 high-quality human binary protein-protein interactions. |
| GO:0005515 protein binding | IPI PMID:25416956 A proteome-scale map of the human interactome network. | MODIFY | Summary: Bare protein binding for a Rab GTPase partner (RAB22A); the informative term is small GTPase binding. Reason: APPL2 is a Rab effector: it binds GTP-loaded Rab5, Rab22A, Rab24 and Rab31 in a screen against 46 Rabs, binds purified Rab31 directly with a preference for the GTP form, and the APPL2 BAR-PH dimer binds two Rab31 molecules. 'Protein binding' says none of this. GO:0031267 small GTPase binding is the current term for it - the former GO:0017137 'Rab GTPase binding' is listed among GO:0031267's secondary ids, i.e. it was merged into small GTPase binding rather than obsoleted, so GO:0031267 is the id to use. Proposed replacements: small GTPase binding Supporting Evidence: PMID:23055524 For APPL2, the yeast two-hybrid screen revealed interactions with Rabs 5, 22a, 24, and 31 PMID:25416956 Here, we describe a systematic map of ?14,000 high-quality human binary protein-protein interactions. |
| GO:0005515 protein binding | IPI PMID:25814554 Phospho-tyrosine dependent protein-protein interaction netwo... | MARK AS OVER ANNOTATED | Summary: Systematic-screen interaction with MAPRE3, with no functional follow-up for APPL2. Reason: This row comes from a proteome-scale screen. The MAPRE3 interaction has never been followed up for APPL2: there is no localisation, no mapped binding surface, no perturbation and no pathway placement connecting it to anything APPL2 is known to do. Bare GO:0005515 rows of this kind are reproducible observations rather than statements about function, and marking them over-annotated is the accurate call - it does not dispute that the screen detected the pair. Note also that IntAct's experiment counts for APPL2 pool replicate experiments from the same studies and are not counts of independent publications. Supporting Evidence: PMID:25814554 We extended an established yeast two-hybrid system employing human protein kinases for the analyses of phospho-tyrosine (pY)-dependent PPIs in a direct experimental, large-scale approach. |
| GO:0005515 protein binding | IPI PMID:28514442 Architecture of the human interactome defines protein commun... | MODIFY | Summary: Bare protein binding for the APPL1 partner; the interaction is a defined BAR-domain heterodimer. Reason: APPL1 is by far the most frequently reported APPL2 partner in this GOA record, and the interaction is not a generic one: the minimal BAR domains are necessary and sufficient for APPL1-APPL2 heterodimerisation in yeast two-hybrid and the proteins reciprocally coimmunoprecipitate. GO:0046982 protein heterodimerization activity states that, where 'protein binding' does not. The heterodimer is also the mechanistic substrate of the adiponectin antagonism, in which APPL2 sequesters APPL1. One caveat specific to this row: its own reference is an affinity-purification mass-spectrometry screen, which establishes co-complex membership rather than a 1:1 dimer. The stoichiometry that justifies GO:0046982 comes from PMID:18034774 - yeast two-hybrid with the minimal BAR domains plus reciprocal coimmunoprecipitation - not from the AP-MS experiment itself. The re-term is still the right call here, because the partner is identified and the nature of the APPL1-APPL2 interaction is independently established; that is what distinguishes this row from the other partners in the same screens, which have no such follow-up and are marked over-annotated. Proposed replacements: protein heterodimerization activity Supporting Evidence: PMID:18034774 the APPL minimal BAR domains were necessary and sufficient for mediating APPL1-APPL1, APPL2-APPL2 and APPL1-APPL2 interactions in the yeast two-hybrid system PMID:28514442 Here we present BioPlex 2.0 (Biophysical Interactions of ORFeome-derived complexes), which uses robust affinity purification-mass spectrometry methodology to elucidate protein interaction networks and co-complexes nucleated by more than |
| GO:0005515 protein binding | IPI PMID:31515488 Extensive disruption of protein interactions by genetic vari... | MODIFY | Summary: Bare protein binding for a Rab GTPase partner (RAB5C); the informative term is small GTPase binding. Reason: APPL2 is a Rab effector: it binds GTP-loaded Rab5, Rab22A, Rab24 and Rab31 in a screen against 46 Rabs, binds purified Rab31 directly with a preference for the GTP form, and the APPL2 BAR-PH dimer binds two Rab31 molecules. 'Protein binding' says none of this. GO:0031267 small GTPase binding is the current term for it - the former GO:0017137 'Rab GTPase binding' is listed among GO:0031267's secondary ids, i.e. it was merged into small GTPase binding rather than obsoleted, so GO:0031267 is the id to use. Proposed replacements: small GTPase binding Supporting Evidence: PMID:23055524 For APPL2, the yeast two-hybrid screen revealed interactions with Rabs 5, 22a, 24, and 31 PMID:31515488 Each human genome carries tens of thousands of coding variants. |
| GO:0005515 protein binding | IPI PMID:31515488 Extensive disruption of protein interactions by genetic vari... | MARK AS OVER ANNOTATED | Summary: Systematic-screen interaction with CRADD, with no functional follow-up for APPL2. Reason: This row comes from a proteome-scale screen. The CRADD interaction has never been followed up for APPL2: there is no localisation, no mapped binding surface, no perturbation and no pathway placement connecting it to anything APPL2 is known to do. Bare GO:0005515 rows of this kind are reproducible observations rather than statements about function, and marking them over-annotated is the accurate call - it does not dispute that the screen detected the pair. Note also that IntAct's experiment counts for APPL2 pool replicate experiments from the same studies and are not counts of independent publications. Supporting Evidence: PMID:31515488 Each human genome carries tens of thousands of coding variants. |
| GO:0005515 protein binding | IPI PMID:31515488 Extensive disruption of protein interactions by genetic vari... | MARK AS OVER ANNOTATED | Summary: Systematic-screen interaction with SUV39H2, with no functional follow-up for APPL2. Reason: This row comes from a proteome-scale screen. The SUV39H2 interaction has never been followed up for APPL2: there is no localisation, no mapped binding surface, no perturbation and no pathway placement connecting it to anything APPL2 is known to do. Bare GO:0005515 rows of this kind are reproducible observations rather than statements about function, and marking them over-annotated is the accurate call - it does not dispute that the screen detected the pair. Note also that IntAct's experiment counts for APPL2 pool replicate experiments from the same studies and are not counts of independent publications. Supporting Evidence: PMID:31515488 Each human genome carries tens of thousands of coding variants. |
| GO:0005515 protein binding | IPI PMID:31515488 Extensive disruption of protein interactions by genetic vari... | MODIFY | Summary: Bare protein binding for the APPL1 partner; the interaction is a defined BAR-domain heterodimer. Reason: APPL1 is by far the most frequently reported APPL2 partner in this GOA record, and the interaction is not a generic one: the minimal BAR domains are necessary and sufficient for APPL1-APPL2 heterodimerisation in yeast two-hybrid and the proteins reciprocally coimmunoprecipitate. GO:0046982 protein heterodimerization activity states that, where 'protein binding' does not. The heterodimer is also the mechanistic substrate of the adiponectin antagonism, in which APPL2 sequesters APPL1. Proposed replacements: protein heterodimerization activity Supporting Evidence: PMID:18034774 the APPL minimal BAR domains were necessary and sufficient for mediating APPL1-APPL1, APPL2-APPL2 and APPL1-APPL2 interactions in the yeast two-hybrid system PMID:31515488 Each human genome carries tens of thousands of coding variants. |
| GO:0005515 protein binding | IPI PMID:31515488 Extensive disruption of protein interactions by genetic vari... | MODIFY | Summary: Bare protein binding for a Rab GTPase partner (RAB22A); the informative term is small GTPase binding. Reason: APPL2 is a Rab effector: it binds GTP-loaded Rab5, Rab22A, Rab24 and Rab31 in a screen against 46 Rabs, binds purified Rab31 directly with a preference for the GTP form, and the APPL2 BAR-PH dimer binds two Rab31 molecules. 'Protein binding' says none of this. GO:0031267 small GTPase binding is the current term for it - the former GO:0017137 'Rab GTPase binding' is listed among GO:0031267's secondary ids, i.e. it was merged into small GTPase binding rather than obsoleted, so GO:0031267 is the id to use. Proposed replacements: small GTPase binding Supporting Evidence: PMID:23055524 For APPL2, the yeast two-hybrid screen revealed interactions with Rabs 5, 22a, 24, and 31 PMID:31515488 Each human genome carries tens of thousands of coding variants. |
| GO:0005515 protein binding | IPI PMID:32296183 A reference map of the human binary protein interactome. | MARK AS OVER ANNOTATED | Summary: Systematic-screen interaction with PRR35, with no functional follow-up for APPL2. Reason: This row comes from a proteome-scale screen. The PRR35 interaction has never been followed up for APPL2: there is no localisation, no mapped binding surface, no perturbation and no pathway placement connecting it to anything APPL2 is known to do. Bare GO:0005515 rows of this kind are reproducible observations rather than statements about function, and marking them over-annotated is the accurate call - it does not dispute that the screen detected the pair. Note also that IntAct's experiment counts for APPL2 pool replicate experiments from the same studies and are not counts of independent publications. Supporting Evidence: PMID:32296183 Here we present a human 'all-by-all' reference interactome map of human binary protein interactions, or 'HuRI'. |
| GO:0005515 protein binding | IPI PMID:32296183 A reference map of the human binary protein interactome. | MODIFY | Summary: Bare protein binding for a Rab GTPase partner (RAB5C); the informative term is small GTPase binding. Reason: APPL2 is a Rab effector: it binds GTP-loaded Rab5, Rab22A, Rab24 and Rab31 in a screen against 46 Rabs, binds purified Rab31 directly with a preference for the GTP form, and the APPL2 BAR-PH dimer binds two Rab31 molecules. 'Protein binding' says none of this. GO:0031267 small GTPase binding is the current term for it - the former GO:0017137 'Rab GTPase binding' is listed among GO:0031267's secondary ids, i.e. it was merged into small GTPase binding rather than obsoleted, so GO:0031267 is the id to use. Proposed replacements: small GTPase binding Supporting Evidence: PMID:23055524 For APPL2, the yeast two-hybrid screen revealed interactions with Rabs 5, 22a, 24, and 31 PMID:32296183 Here we present a human 'all-by-all' reference interactome map of human binary protein interactions, or 'HuRI'. |
| GO:0005515 protein binding | IPI PMID:32296183 A reference map of the human binary protein interactome. | MARK AS OVER ANNOTATED | Summary: Systematic-screen interaction with EPM2AIP1, with no functional follow-up for APPL2. Reason: This row comes from a proteome-scale screen. The EPM2AIP1 interaction has never been followed up for APPL2: there is no localisation, no mapped binding surface, no perturbation and no pathway placement connecting it to anything APPL2 is known to do. Bare GO:0005515 rows of this kind are reproducible observations rather than statements about function, and marking them over-annotated is the accurate call - it does not dispute that the screen detected the pair. Note also that IntAct's experiment counts for APPL2 pool replicate experiments from the same studies and are not counts of independent publications. Supporting Evidence: PMID:32296183 Here we present a human 'all-by-all' reference interactome map of human binary protein interactions, or 'HuRI'. |
| GO:0005515 protein binding | IPI PMID:32296183 A reference map of the human binary protein interactome. | MARK AS OVER ANNOTATED | Summary: Systematic-screen interaction with KIFC3 (isoform 5), with no functional follow-up for APPL2. Reason: This row comes from a proteome-scale screen. The KIFC3 (isoform 5) interaction has never been followed up for APPL2: there is no localisation, no mapped binding surface, no perturbation and no pathway placement connecting it to anything APPL2 is known to do. Bare GO:0005515 rows of this kind are reproducible observations rather than statements about function, and marking them over-annotated is the accurate call - it does not dispute that the screen detected the pair. Note also that IntAct's experiment counts for APPL2 pool replicate experiments from the same studies and are not counts of independent publications. Supporting Evidence: PMID:32296183 Here we present a human 'all-by-all' reference interactome map of human binary protein interactions, or 'HuRI'. |
| GO:0005515 protein binding | IPI PMID:32296183 A reference map of the human binary protein interactome. | MARK AS OVER ANNOTATED | Summary: Systematic-screen interaction with SUV39H2 (isoform 2), with no functional follow-up for APPL2. Reason: This row comes from a proteome-scale screen. The SUV39H2 (isoform 2) interaction has never been followed up for APPL2: there is no localisation, no mapped binding surface, no perturbation and no pathway placement connecting it to anything APPL2 is known to do. Bare GO:0005515 rows of this kind are reproducible observations rather than statements about function, and marking them over-annotated is the accurate call - it does not dispute that the screen detected the pair. Note also that IntAct's experiment counts for APPL2 pool replicate experiments from the same studies and are not counts of independent publications. Supporting Evidence: PMID:32296183 Here we present a human 'all-by-all' reference interactome map of human binary protein interactions, or 'HuRI'. |
| GO:0005515 protein binding | IPI PMID:32296183 A reference map of the human binary protein interactome. | MODIFY | Summary: Bare protein binding for the APPL1 partner; the interaction is a defined BAR-domain heterodimer. Reason: APPL1 is by far the most frequently reported APPL2 partner in this GOA record, and the interaction is not a generic one: the minimal BAR domains are necessary and sufficient for APPL1-APPL2 heterodimerisation in yeast two-hybrid and the proteins reciprocally coimmunoprecipitate. GO:0046982 protein heterodimerization activity states that, where 'protein binding' does not. The heterodimer is also the mechanistic substrate of the adiponectin antagonism, in which APPL2 sequesters APPL1. Proposed replacements: protein heterodimerization activity Supporting Evidence: PMID:18034774 the APPL minimal BAR domains were necessary and sufficient for mediating APPL1-APPL1, APPL2-APPL2 and APPL1-APPL2 interactions in the yeast two-hybrid system PMID:32296183 Here we present a human 'all-by-all' reference interactome map of human binary protein interactions, or 'HuRI'. |
| GO:0005515 protein binding | IPI PMID:32296183 A reference map of the human binary protein interactome. | MODIFY | Summary: Bare protein binding for a Rab GTPase partner (RAB22A); the informative term is small GTPase binding. Reason: APPL2 is a Rab effector: it binds GTP-loaded Rab5, Rab22A, Rab24 and Rab31 in a screen against 46 Rabs, binds purified Rab31 directly with a preference for the GTP form, and the APPL2 BAR-PH dimer binds two Rab31 molecules. 'Protein binding' says none of this. GO:0031267 small GTPase binding is the current term for it - the former GO:0017137 'Rab GTPase binding' is listed among GO:0031267's secondary ids, i.e. it was merged into small GTPase binding rather than obsoleted, so GO:0031267 is the id to use. Proposed replacements: small GTPase binding Supporting Evidence: PMID:23055524 For APPL2, the yeast two-hybrid screen revealed interactions with Rabs 5, 22a, 24, and 31 PMID:32296183 Here we present a human 'all-by-all' reference interactome map of human binary protein interactions, or 'HuRI'. |
| GO:0005515 protein binding | IPI PMID:33961781 Dual proteome-scale networks reveal cell-specific remodeling... | MODIFY | Summary: Bare protein binding for the APPL1 partner; the interaction is a defined BAR-domain heterodimer. Reason: APPL1 is by far the most frequently reported APPL2 partner in this GOA record, and the interaction is not a generic one: the minimal BAR domains are necessary and sufficient for APPL1-APPL2 heterodimerisation in yeast two-hybrid and the proteins reciprocally coimmunoprecipitate. GO:0046982 protein heterodimerization activity states that, where 'protein binding' does not. The heterodimer is also the mechanistic substrate of the adiponectin antagonism, in which APPL2 sequesters APPL1. One caveat specific to this row: its own reference is an affinity-purification mass-spectrometry screen, which establishes co-complex membership rather than a 1:1 dimer. The stoichiometry that justifies GO:0046982 comes from PMID:18034774 - yeast two-hybrid with the minimal BAR domains plus reciprocal coimmunoprecipitation - not from the AP-MS experiment itself. The re-term is still the right call here, because the partner is identified and the nature of the APPL1-APPL2 interaction is independently established; that is what distinguishes this row from the other partners in the same screens, which have no such follow-up and are marked over-annotated. Proposed replacements: protein heterodimerization activity Supporting Evidence: PMID:18034774 the APPL minimal BAR domains were necessary and sufficient for mediating APPL1-APPL1, APPL2-APPL2 and APPL1-APPL2 interactions in the yeast two-hybrid system PMID:33961781 Through affinity-purification mass spectrometry, we have created two proteome-scale, cell-line-specific interaction networks. |
| GO:0005515 protein binding | IPI PMID:35271311 OpenCell: Endogenous tagging for the cartography of human ce... | MODIFY | Summary: Bare protein binding for the APPL1 partner; the interaction is a defined BAR-domain heterodimer. Reason: APPL1 is by far the most frequently reported APPL2 partner in this GOA record, and the interaction is not a generic one: the minimal BAR domains are necessary and sufficient for APPL1-APPL2 heterodimerisation in yeast two-hybrid and the proteins reciprocally coimmunoprecipitate. GO:0046982 protein heterodimerization activity states that, where 'protein binding' does not. The heterodimer is also the mechanistic substrate of the adiponectin antagonism, in which APPL2 sequesters APPL1. One caveat specific to this row: its own reference is an affinity-purification mass-spectrometry screen, which establishes co-complex membership rather than a 1:1 dimer. The stoichiometry that justifies GO:0046982 comes from PMID:18034774 - yeast two-hybrid with the minimal BAR domains plus reciprocal coimmunoprecipitation - not from the AP-MS experiment itself. The re-term is still the right call here, because the partner is identified and the nature of the APPL1-APPL2 interaction is independently established; that is what distinguishes this row from the other partners in the same screens, which have no such follow-up and are marked over-annotated. Proposed replacements: protein heterodimerization activity Supporting Evidence: PMID:18034774 the APPL minimal BAR domains were necessary and sufficient for mediating APPL1-APPL1, APPL2-APPL2 and APPL1-APPL2 interactions in the yeast two-hybrid system PMID:35271311 We combined genome engineering, confocal live-cell imaging, mass spectrometry, and data science to systematically map the localization and interactions of human proteins. |
| GO:0005634 nucleus | IDA PMID:15016378 APPL proteins link Rab5 to nuclear signal transduction via a... | ACCEPT | Summary: APPL2 shuttles from endosomal membranes to the nucleus. Reason: The paper that defined APPL endosomes also established the nuclear pool of APPL proteins and the stimulus-dependent translocation that links the two, and UniProt records Nucleus for Q8NEU8 with experimental evidence from this paper. A nuclear APPL2 pool is corroborated independently: both APPL1 and APPL2 translocate to the nucleus on EGF stimulation, APPL2 carries a BAR-loop basic motif absent from APPL1, and the Reptin/HDAC work places APPL2 in a nuclear complex. Supporting Evidence: PMID:15016378 APPL1 translocates from the membranes to the nucleus where it interacts with the nucleosome remodeling and histone deacetylase multiprotein complex NuRD/MeCP1 PMID:15016378 Our findings identify an endosomal compartment bearing Rab5 and APPL proteins as an intermediate in signaling between the plasma membrane and the nucleus. PMID:26583432 both APPL1 and APPL2 translocate into the nucleus from the cytosol in response to EGF stimulation, to regulate cell proliferation |
| GO:0005634 nucleus | IDA PMID:18034774 Membrane targeting by APPL1 and APPL2: dynamic scaffolds tha... | ACCEPT | Summary: Imaging of APPL2 fusion proteins shows a nuclear pool alongside the membrane pool. Reason: Coexpressed full-length APPL fusion proteins localised to cytosolic membranes and the nucleus, and the isolated PH domains reached distinct nuclear and perinuclear structures. This is tagged-protein imaging rather than endogenous staining, so it is weaker than the compartment rows, but the nuclear pool is independently supported by the Rab5/NuRD and Reptin literature and by the APPL2-specific basic motif at 151-157, so the annotation stands. Supporting Evidence: PMID:18034774 moderate overexpression did not disrupt APPL localization to cytosolic membranes and the nucleus PMID:18034774 We found that APPL1 and APPL2 isolated PH domains localized to the plasma membrane, cytosolic vesicles and distinct nuclear and perinuclear structures file:human/APPL2/APPL2-bioinformatics/RESULTS.md also present verbatim, sits inside the BAR domain (loop 2), and is conserved in mouse Appl2 |
| GO:0005634 nucleus | IEA GO_REF:0000044 | ACCEPT | Summary: Projection of UniProt's curated Nucleus location keyword. Reason: This row restates UniProt's own Nucleus annotation for Q8NEU8, which is itself supported by two experimental papers. It is redundant with the two IDA rows rather than independent of them, but it is correct. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB-SubCell:SL-0191 · UniProt subcellular-location keyword: Nucleus SUPPORTS TRANSFER Projection of UniProt's own curated subcellular-location keyword, so this row re-states the Swiss-Prot annotation rather than adding independent evidence. Supporting Evidence: PMID:15016378 APPL1 translocates from the membranes to the nucleus where it interacts with the nucleosome remodeling and histone deacetylase multiprotein complex NuRD/MeCP1 PMID:18034774 moderate overexpression did not disrupt APPL localization to cytosolic membranes and the nucleus |
| GO:0005737 cytoplasm | IEA GO_REF:0000120 | KEEP AS NON CORE | Summary: A cytosolic pool exists, but 'cytoplasm' is too general to be informative for a peripheral membrane adaptor. Reason: APPL2 is a peripheral membrane protein with a genuine soluble pool - it was immunolabelled in the cytoplasm of macrophages, and annexin A2 knockdown solubilises it from endosomes, which only makes sense if a cytosolic pool exists. So the term is not wrong. It is non-core because the informative statements about where APPL2 acts are the endosome, early-endosome-membrane and phagosome rows; 'cytoplasm' is their uninformative parent. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse annotation reflects cytoplasmic immunolabelling of endogenous APPL2 in macrophages (PMID:25568335). True, but generic. ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. InterPro:IPR004148 · InterPro BAR domain signature (IPR004148) SUPPORTS TRANSFER A domain-presence signature. APPL2 does carry a BAR domain (UniProt FT DOMAIN 3..268), but domain presence is not evidence about where the protein is. UniProtKB-SubCell:SL-0086 · UniProt subcellular-location keyword: Cytoplasm SUPPORTS TRANSFER Projection of UniProt's own curated subcellular-location keyword, so this row re-states the Swiss-Prot annotation rather than adding independent evidence. The mouse annotation reflects cytoplasmic immunolabelling of endogenous APPL2 in macrophages (PMID:25568335). True, but generic. Supporting Evidence: PMID:25568335 APPL2 was immunolabeled in the cytoplasm, on cell surface ruffles, and on early, actin-rich phagosomes PMID:21645192 silencing of its expression causes solubilization of APPL2 from endosomes |
| GO:0005737 cytoplasm | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: A cytosolic pool exists, but 'cytoplasm' is too general to be informative for a peripheral membrane adaptor. Reason: APPL2 is a peripheral membrane protein with a genuine soluble pool - it was immunolabelled in the cytoplasm of macrophages, and annexin A2 knockdown solubilises it from endosomes, which only makes sense if a cytosolic pool exists. So the term is not wrong. It is non-core because the informative statements about where APPL2 acts are the endosome, early-endosome-membrane and phagosome rows; 'cytoplasm' is their uninformative parent. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse annotation reflects cytoplasmic immunolabelling of endogenous APPL2 in macrophages (PMID:25568335). True, but generic. Supporting Evidence: PMID:25568335 APPL2 was immunolabeled in the cytoplasm, on cell surface ruffles, and on early, actin-rich phagosomes PMID:21645192 silencing of its expression causes solubilization of APPL2 from endosomes |
| GO:0005768 endosome | IDA PMID:21645192 Biochemical characterization of APPL endosomes: the role of ... | ACCEPT | Summary: Endogenous APPL2 marks a biochemically distinct subpopulation of early endosomes. Reason: APPL endosomes were characterised here by membrane fractionation and by imaging of endogenous APPL2 in human fibroblasts, and they behave as a discriminable compartment: a distinct migration pattern in continuous density gradients and only limited colocalisation with EEA1. This is core - membrane recruitment through Rab and annexin A2 binding is what APPL2 is for. GO has no 'APPL endosome' term, which is recorded under proposed_new_terms; GO:0005768 and the more specific GO:0031901 are the available and correct approximations. Supporting Evidence: PMID:21645192 APPL endosomes are a recently identified subpopulation of early endosomes characterized by the presence of two homologous Rab5 effector proteins APPL1 and APPL2. PMID:21645192 They exhibit only limited colocalization with EEA1, another Rab5 effector and a marker of the canonical early endosomes. PMID:21645192 Annexin A2 is essential for targeting APPL2 to endosomes and can compensate for Rab5 deficiency in mediating APPL membrane recruitment |
| GO:0005768 endosome | IEA GO_REF:0000117 | ACCEPT | Summary: ARBA rule reaching the same compartment the IDA row establishes. Reason: ARBA00028568 is a broad rule (97 condition sets across eukaryotes), so on its own it would be weak evidence; here it happens to agree with a well-supported IDA row from endogenous-protein fractionation and imaging, so the annotation is correct even though the rule is not what establishes it. Propagation Review Root cause: NO FAILURE CORE Sources checked: ARBA:ARBA00028568 · ARBA rule ARBA00028568 SUPPORTS TRANSFER ARBA rule for endosome; 97 condition sets across eukaryotes. Broad, but the compartment is independently established for APPL2 by IDA. Supporting Evidence: PMID:21645192 APPL endosomes are a recently identified subpopulation of early endosomes characterized by the presence of two homologous Rab5 effector proteins APPL1 and APPL2. |
| GO:0005886 plasma membrane | IDA PMID:18034774 Membrane targeting by APPL1 and APPL2: dynamic scaffolds tha... | KEEP AS NON CORE | Summary: The plasma-membrane signal comes from overexpressed isolated PH domains, not the intact protein. Reason: The plasma-membrane observation in this paper is of isolated APPL PH domains expressed as CFP fusions; the full-length APPL2 fusion protein localised to cytosolic membrane structures, not the plasma membrane. An ectopically expressed fragment reaching the plasma membrane is a statement about the domain's lipid affinity rather than about where APPL2 resides. There is a defensible membrane-proximal pool - APPL endosomes accumulate beneath the plasma membrane and APPL2 is on surface ruffles in macrophages - so the term is kept, but as non-core. Supporting Evidence: PMID:18034774 We found that APPL1 and APPL2 isolated PH domains localized to the plasma membrane, cytosolic vesicles and distinct nuclear and perinuclear structures PMID:18034774 Both APPL1-YFP and APPL2-YFP showed dynamic associations with cytosolic membrane structures that underwent movement, fusion and fission events PMID:25568335 APPL2 was immunolabeled in the cytoplasm, on cell surface ruffles, and on early, actin-rich phagosomes |
| GO:0005886 plasma membrane | IEA GO_REF:0000120 | KEEP AS NON CORE | Summary: Driven by the most promiscuous ARBA rule on the gene, combined with a mouse keyword projection. Reason: This row combines ARBA00027801 with the mouse ortholog and a Cell membrane keyword. ARBA00027801 has 686 condition sets spanning bacteria, plants, fungi and metazoa - a rule that broad firing on APPL2 conveys no information about APPL2, so none of this row's confidence should come from the rule. The term is nonetheless true: APPL2 sits on surface ruffles and phagocytic cups in macrophages, APPL endosomes accumulate beneath the plasma membrane, and UniProt records Cell membrane for Q8NEU8 with experimental evidence. Kept as non-core, consistent with the IDA and ISS rows for the same term, because ruffle, ruffle membrane and early phagosome state the same localisation far more precisely. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: ARBA:ARBA00027801 · ARBA rule ARBA00027801 SUPPORTS TRANSFER ARBA rule for plasma membrane; 686 condition sets spanning bacteria, plants, fungi and metazoa. This is the most promiscuous rule on the gene and it is what drives the plasma-membrane call. UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse donor's plasma-membrane association is really the ruffle and phagocytic-cup pool, which the ruffle, ruffle-membrane and early-phagosome rows already state more precisely. ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. UniProtKB-SubCell:SL-0039 · UniProt subcellular-location keyword: Cell membrane SUPPORTS TRANSFER Projection of UniProt's own curated subcellular-location keyword, so this row re-states the Swiss-Prot annotation rather than adding independent evidence. The mouse donor's plasma-membrane association is really the ruffle and phagocytic-cup pool, which the ruffle, ruffle-membrane and early-phagosome rows already state more precisely. Supporting Evidence: PMID:25568335 APPL2 was immunolabeled in the cytoplasm, on cell surface ruffles, and on early, actin-rich phagosomes PMID:25568335 This contrasts with the predominant and stronger labeling of mCherry-APPL2 on phagosomes and its general absence from endosomes |
| GO:0005886 plasma membrane | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: Orthology transfer of the mouse Cell membrane location. Reason: Mouse Appl2 is annotated to the cell membrane on the strength of its ruffle and phagocytic-cup localisation in macrophages (PMID:25568335). The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. The transfer is sound, but the term is a generic parent of the ruffle and ruffle-membrane rows that carry the same observation more precisely. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse cell-membrane call reflects ruffle and phagocytic-cup localisation in macrophages (PMID:25568335), already stated more precisely by the ruffle rows. Supporting Evidence: PMID:25568335 APPL2 was immunolabeled in the cytoplasm, on cell surface ruffles, and on early, actin-rich phagosomes |
| GO:0006606 protein import into nucleus | IDA PMID:26583432 APPL proteins promote TGFβ-induced nuclear transport of the ... | KEEP AS NON CORE | Summary: APPL proteins are required for nuclear delivery of the TGF-beta receptor intracellular domain - but only a double knockdown was done. Reason: Silencing APPL1 and APPL2 abolished nuclear accumulation of endogenous and tagged TbetaRI-ICD after TGF-beta stimulation, so the gene product is genuinely required for this import event and the annotation is not wrong. Every functional experiment in the paper, however, silences both adaptors simultaneously, and the physical association with TbetaRI and PKCzeta was mapped to APPL1 alone - so APPL2's individual contribution is unresolved. Given 54% paralog identity, a double knockdown cannot be read as evidence about either gene singly. Kept as a real but peripheral role rather than removed, since curators reading the full text had the same design available to them and the requirement itself is not in doubt. Note the cargo imported is TbetaRI-ICD, not APPL2. Supporting Evidence: PMID:26583432 the endocytic adaptor molecules APPL1 and APPL2 are required for TGFβ-induced nuclear translocation of TβRI-ICD and for cancer cell invasiveness of human prostate and breast cancer cell lines PMID:26583432 we identify APPL1 as a TβRI- and PKCζ-associated protein and show that APPL1 and APPL2 are required for the nuclear translocation of TβRI-ICD PMID:26583432 We observed that nuclear accumulation of endogenous TβRI-ICD decreased after siRNA-mediated silencing of APPL1 and APPL2 file:human/APPL2/APPL2-bioinformatics/RESULTS.md The paralog identity, 53.9%, reproduces the 52% that King et al. report by ClustalW |
| GO:0006606 protein import into nucleus | IEA GO_REF:0000117 | KEEP AS NON CORE | Summary: Narrow ARBA rule reproducing the same claim as the IDA row. Reason: ARBA00028627 has 14 condition sets over FunFam and InterPro signatures in metazoa and fungi. It reaches the same term the IDA row carries and inherits the same caveat - the only perturbation behind that term is a simultaneous knockdown of APPL1 and APPL2. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: ARBA:ARBA00028627 · ARBA rule ARBA00028627 SUPPORTS TRANSFER ARBA rule for protein import into nucleus; 14 condition sets over FunFam/InterPro signatures in metazoa and fungi. Narrow, and it reproduces the IDA row's claim. Supporting Evidence: PMID:26583432 we identify APPL1 as a TβRI- and PKCζ-associated protein and show that APPL1 and APPL2 are required for the nuclear translocation of TβRI-ICD |
| GO:0007165 signal transduction | TAS PMID:15016378 APPL proteins link Rab5 to nuclear signal transduction via a... | KEEP AS NON CORE | Summary: A true but uninformative parent term, taken as an author statement from the paper that defined APPL endosomes. Reason: APPL2 does act in signal transduction - the founding paper links Rab5 to signalling and mitogenesis through an endosomal intermediate carrying APPL1 and APPL2. But GO:0007165 is a high-level parent with no discriminating content, and APPL2's specific signalling roles are stated far better by the adiponectin, TLR4, Fc-gamma-receptor and insulin rows. A projection check confirms this reference is not a bulk source (20 annotations over 4 gene products in QuickGO), so the row is a genuine author statement, just a general one. Supporting Evidence: PMID:15016378 We report the identification of a pathway directly linking the small GTPase Rab5, a key regulator of endocytosis, to signal transduction and mitogenesis. PMID:15016378 Our findings identify an endosomal compartment bearing Rab5 and APPL proteins as an intermediate in signaling between the plasma membrane and the nucleus. |
| GO:0007179 transforming growth factor beta receptor signaling pathway | IEA GO_REF:0000117 | KEEP AS NON CORE | Summary: Narrow ARBA rule agreeing with the IMP row. Reason: ARBA00085361 has 14 metazoan condition sets and assigns the same term as the IMP row. It reproduces rather than corroborates that row, and inherits its double-knockdown caveat. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: ARBA:ARBA00085361 · ARBA rule ARBA00085361 SUPPORTS TRANSFER ARBA rule for TGF-beta receptor signalling; 14 condition sets, metazoan. Narrow, and it reproduces the IMP row's claim. Supporting Evidence: PMID:26583432 the endocytic adaptor molecules APPL1 and APPL2 are required for TGFβ-induced nuclear translocation of TβRI-ICD and for cancer cell invasiveness of human prostate and breast cancer cell lines |
| GO:0007179 transforming growth factor beta receptor signaling pathway | IMP PMID:26583432 APPL proteins promote TGFβ-induced nuclear transport of the ... | KEEP AS NON CORE | Summary: APPL1/APPL2 depletion impairs TGF-beta-driven TbetaRI-ICD transport, invasion and MMP transcription. Reason: The functional readouts are real and multiple - loss of nuclear TbetaRI-ICD, reduced TGF-beta-induced invasion of prostate and breast cancer lines, and reduced MMP2/MMP9 transcription. The design, though, is a combined APPL1 + APPL2 knockdown throughout, with the receptor interaction mapped to APPL1. So APPL2 is required in combination, which justifies keeping the annotation, but the evidence does not make TGF-beta signalling a core APPL2 function, and it was obtained in cancer cell lines. Supporting Evidence: PMID:26583432 the endocytic adaptor molecules APPL1 and APPL2 are required for TGFβ-induced nuclear translocation of TβRI-ICD and for cancer cell invasiveness of human prostate and breast cancer cell lines PMID:26583432 we identify APPL1 as a TβRI- and PKCζ-associated protein and show that APPL1 and APPL2 are required for the nuclear translocation of TβRI-ICD file:human/APPL2/APPL2-bioinformatics/RESULTS.md The paralog identity, 53.9%, reproduces the 52% that King et al. report by ClustalW |
| GO:0009631 cold acclimation | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: cold acclimation rests on a neuron-restricted Appl2 knockout in mouse hypothalamus. Reason: Deleting APPL2 in ventromedial-hypothalamic RIP-Cre neurons reduced beiging of subcutaneous white adipose tissue and produced cold intolerance and obesity, acting through AMPK and sympathetic outflow (PMID:29467283). The perturbation is APPL2-specific and conditional. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. This is nonetheless an organism-level physiological outcome of APPL2 acting in one neuronal population, several steps removed from what the protein does molecularly, so it is peripheral rather than core. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse evidence is a conditional, neuron-restricted APPL2 knockout with a whole-animal thermogenic phenotype (PMID:29467283) - APPL2-specific, but organismal. ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. Supporting Evidence: PMID:29467283 Genetic ablation of APPL2 in RIP-Cre neurons diminishes beiging in sWAT without affecting BAT, leading to cold intolerance and obesity in mice. PMID:29467283 Hypothalamic APPL2 enhances neuronal activation in VMH RIP-Cre neurons and raphe pallidus, thereby eliciting SNS outflow to sWAT and subsequent beiging. |
| GO:0009631 cold acclimation | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: cold acclimation rests on a neuron-restricted Appl2 knockout in mouse hypothalamus. Reason: Deleting APPL2 in ventromedial-hypothalamic RIP-Cre neurons reduced beiging of subcutaneous white adipose tissue and produced cold intolerance and obesity, acting through AMPK and sympathetic outflow (PMID:29467283). The perturbation is APPL2-specific and conditional. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. This is nonetheless an organism-level physiological outcome of APPL2 acting in one neuronal population, several steps removed from what the protein does molecularly, so it is peripheral rather than core. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse evidence is a conditional, neuron-restricted APPL2 knockout with a whole-animal thermogenic phenotype (PMID:29467283) - APPL2-specific, but organismal. Supporting Evidence: PMID:29467283 Genetic ablation of APPL2 in RIP-Cre neurons diminishes beiging in sWAT without affecting BAT, leading to cold intolerance and obesity in mice. PMID:29467283 Hypothalamic APPL2 enhances neuronal activation in VMH RIP-Cre neurons and raphe pallidus, thereby eliciting SNS outflow to sWAT and subsequent beiging. |
| GO:0010008 endosome membrane | IBA GO_REF:0000033 | ACCEPT | Summary: Family-level endosomal-membrane inference placed at the Bilateria node ancestral to both APPL paralogs; correct and core for APPL2. Reason: The IBD behind this IBA sits on PTN000572460 at taxon 33213 (Bilateria), above the APPL1/APPL2 duplication, and was seeded by mouse Appl1, human APPL1 and human APPL2 itself. Membrane recruitment through Rab binding is the defining property of both paralogs, so placing it at their last common ancestor is the right depth, and APPL2 sits squarely inside the inheriting clade (PANTHER subfamily PTHR46415:SF1). The target's own accession in the WITH/FROM is expected: APPL2 carries two independent IDA rows for this exact term plus an EXP row for the more specific early endosome membrane, and those descendant evidences are part of what the curator used to place the node. There is no evidence of loss or divergence - APPL2 binds four Rabs, is recruited by annexin A2, and is solubilised from endosomes when that recruitment fails. Propagation Review Root cause: NO FAILURE CORE Sources checked: MGI:MGI:1920243 · Appl1 (Mus musculus) SUPPORTS TRANSFER Mouse Appl1, resolved to UniProtKB:Q8K3H0. A Rab5 effector on the same APPL endosome compartment; the localisation is correct for it and transfers correctly to APPL2, which was shown on the same structures in the original characterisation. PANTHER:PTN000572460 · PTHR46415 Bilateria (taxon 33213) IBD node ancestral to both APPL1 (subfamily SF3) and APPL2 (SF1) SUPPORTS TRANSFER Bilateria-level IBD node of PTHR46415, seeded by mouse Appl1, human APPL1 and human APPL2. The family has only two nodes and five reviewed members, and this node is ancestral to both subfamilies (SF1 = APPL2, SF3 = APPL1), so endosomal-membrane localisation is asserted for the whole family. APPL2 is inside the inheriting clade. UniProtKB:Q8NEU8 · APPL2 (Homo sapiens), the target SUPPORTS TRANSFER The target's own accession, which is correct and expected here rather than circular: APPL2 has two IDA rows for GO:0010008 (PMID:15016378, PMID:21645192) and an EXP row for early endosome membrane, and those are among the descendant evidences the PAINT curator used to place the IBD. Their presence means the IBA adds a claim of inheritance on top of experimental grounding that already exists on the target. UniProtKB:Q9UKG1 · APPL1 (Homo sapiens) SUPPORTS TRANSFER Human APPL1, the paralog. APPL1 and APPL2 were identified together as the two Rab5 effectors defining APPL endosomes, so its endosomal-membrane evidence is directly applicable to the node. Supporting Evidence: PMID:21645192 APPL endosomes are a recently identified subpopulation of early endosomes characterized by the presence of two homologous Rab5 effector proteins APPL1 and APPL2. PMID:15016378 This pathway operates via APPL1 and APPL2, two Rab5 effectors, which reside on a subpopulation of endosomes. PMID:21645192 Annexin A2 is essential for targeting APPL2 to endosomes and can compensate for Rab5 deficiency in mediating APPL membrane recruitment |
| GO:0010008 endosome membrane | IDA PMID:15016378 APPL proteins link Rab5 to nuclear signal transduction via a... | ACCEPT | Summary: APPL2 is a Rab5 effector on endosomal membranes - the founding observation for this compartment. Reason: APPL1 and APPL2 were identified as the two Rab5 effectors residing on a distinct endosome subpopulation, with Rab5 binding required for their membrane recruitment and their role in proliferation. UniProt records Early endosome membrane and Peripheral membrane protein for Q8NEU8 from this paper. This is core. Supporting Evidence: PMID:15016378 This pathway operates via APPL1 and APPL2, two Rab5 effectors, which reside on a subpopulation of endosomes. PMID:15016378 Both APPL1 and APPL2 are essential for cell proliferation and their function requires Rab5 binding. PMID:15016378 Our findings identify an endosomal compartment bearing Rab5 and APPL proteins as an intermediate in signaling between the plasma membrane and the nucleus. |
| GO:0010008 endosome membrane | IDA PMID:21645192 Biochemical characterization of APPL endosomes: the role of ... | ACCEPT | Summary: Endogenous APPL2 on endosomal membranes, with the recruitment mechanism demonstrated. Reason: Membrane fractionation plus imaging of endogenous APPL2 places it on a discriminable subpopulation of early endosomes, and knocking down annexin A2 solubilises APPL2 from those membranes - a loss-of-function control on the localisation itself, which few compartment annotations have. Core. Supporting Evidence: PMID:21645192 APPL endosomes are a recently identified subpopulation of early endosomes characterized by the presence of two homologous Rab5 effector proteins APPL1 and APPL2. PMID:21645192 silencing of its expression causes solubilization of APPL2 from endosomes PMID:21645192 Annexin A2 is essential for targeting APPL2 to endosomes and can compensate for Rab5 deficiency in mediating APPL membrane recruitment |
| GO:0010008 endosome membrane | IEA GO_REF:0000044 | ACCEPT | Summary: Projection of UniProt's Endosome membrane keyword. Reason: Restates UniProt's curated location, which derives from the same experimental papers as the two IDA rows. Redundant with them, and correct. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB-SubCell:SL-0100 · UniProt subcellular-location keyword: Endosome membrane SUPPORTS TRANSFER Projection of UniProt's own curated subcellular-location keyword, so this row re-states the Swiss-Prot annotation rather than adding independent evidence. Supporting Evidence: PMID:21645192 APPL endosomes are a recently identified subpopulation of early endosomes characterized by the presence of two homologous Rab5 effector proteins APPL1 and APPL2. |
| GO:0010762 regulation of fibroblast migration | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: regulation of fibroblast migration, from mouse embryonic fibroblasts where a single Appl2 knockout has the phenotype. Reason: The title of the source paper reads as a double-knockout result, and I checked whether it is: it is not. Single Appl1 KO, single Appl2 KO and the double KO all showed decreased HGF-induced Akt activation, so there is an APPL2-specific loss-of-function anchor here and the rows do not suffer the paralog conflation that the TGF-beta rows do. The migration and invasion defects are reported for Appl-deficient MEFs collectively, so the migration row is slightly weaker than the HGF-response row. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Non-core because the effect is growth-factor-selective - EGF, insulin and serum activate Akt normally - and Appl2-null mice are viable and grow normally. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse annotation has a single-Appl2-knockout anchor (decreased HGF-induced Akt activation in Appl2 KO MEFs), not only the double knockout the title implies. ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. Supporting Evidence: PMID:26445298 Appl1 KO, Appl2 KO, and especially Appl1/2 DKO MEFs showed consistently decreased Akt activation upon stimulation with HGF PMID:26445298 Appl1/2-null mouse embryonic fibroblasts exhibited defects in HGF-induced Akt activation, migration, and invasion. PMID:26445298 ubiquitous Appl2 knockout (Appl2-/-) mice, much like Appl1-/- mice, are viable and grow normally to adulthood |
| GO:0010762 regulation of fibroblast migration | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: regulation of fibroblast migration, from mouse embryonic fibroblasts where a single Appl2 knockout has the phenotype. Reason: The title of the source paper reads as a double-knockout result, and I checked whether it is: it is not. Single Appl1 KO, single Appl2 KO and the double KO all showed decreased HGF-induced Akt activation, so there is an APPL2-specific loss-of-function anchor here and the rows do not suffer the paralog conflation that the TGF-beta rows do. The migration and invasion defects are reported for Appl-deficient MEFs collectively, so the migration row is slightly weaker than the HGF-response row. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Non-core because the effect is growth-factor-selective - EGF, insulin and serum activate Akt normally - and Appl2-null mice are viable and grow normally. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse annotation has a single-Appl2-knockout anchor (decreased HGF-induced Akt activation in Appl2 KO MEFs), not only the double knockout the title implies. Supporting Evidence: PMID:26445298 Appl1 KO, Appl2 KO, and especially Appl1/2 DKO MEFs showed consistently decreased Akt activation upon stimulation with HGF PMID:26445298 Appl1/2-null mouse embryonic fibroblasts exhibited defects in HGF-induced Akt activation, migration, and invasion. PMID:26445298 ubiquitous Appl2 knockout (Appl2-/-) mice, much like Appl1-/- mice, are viable and grow normally to adulthood |
| GO:0016020 membrane | IDA PMID:18034774 Membrane targeting by APPL1 and APPL2: dynamic scaffolds tha... | KEEP AS NON CORE | Summary: Live-cell imaging of APPL2 on dynamic cytosolic membrane structures; the term is the root of the compartment branch. Reason: Full-length APPL2-YFP associated with cytosolic membrane structures that moved, fused and underwent fission, and recruited endogenous RAB5. The observation is sound and it is the direct basis for calling APPL2 a peripheral membrane protein, but 'membrane' is the least specific term available and the same paper's cytoplasmic-vesicle row plus the endosome rows state it properly. Supporting Evidence: PMID:18034774 Both APPL1-YFP and APPL2-YFP showed dynamic associations with cytosolic membrane structures that underwent movement, fusion and fission events PMID:18034774 APPL2-YFP colocalized strongly with RAB5 to smaller cytosolic vesicular structures |
| GO:0023052 signaling | IBA GO_REF:0000033 | KEEP AS NON CORE | Summary: A deliberately generic family-level term: the node's seeds are APPL1 and APPL2, whose signalling roles are opposite in sign. Reason: The IBD sits on PTN000572460 at Bilateria, seeded by mouse Appl2, human APPL1 and human APPL2. The instinct is to call such a high-level term a granularity mismatch, and that would be wrong here. The two seeds disagree in sign: APPL1 potentiates adiponectin and Akt signalling while APPL2 antagonises it, competing for AdipoR1/R2 and sequestering APPL1; and in the same macrophage system APPL1 and APPL2 have 'separate and opposing functions' in Akt and MAPK signalling. Any child term specific enough to be informative would be false for one seed, so the uninformative parent is the honest least common ancestor and there is no propagation failure. It is non-core only in the sense that the term conveys nothing; the underlying claim - that APPL2 is a signalling adaptor - is correct and is stated properly by the pathway-specific rows. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: MGI:MGI:2384914 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, resolved to UniProtKB:Q8K3G9 - the target's own 1:1 ortholog, 92.7% identical. Its signalling annotations come from Appl2 knockout mice and Appl2-specific knockdowns. PANTHER:PTN000572460 · PTHR46415 Bilateria (taxon 33213) IBD node ancestral to both APPL1 (subfamily SF3) and APPL2 (SF1) SUPPORTS TRANSFER Bilateria-level IBD node of PTHR46415, seeded by mouse Appl2, human APPL1 and human APPL2. Because the node is ancestral to both paralogs and the paralogs' signalling roles are opposite in sign, GO:0023052 is the correct least common ancestor rather than an under-specified choice. The same curator placed the AKT-binding and insulin-receptor -signalling IBDs one node down, at PTN008708200 (Tetrapoda, APPL1 subfamily only), where APPL2 does not inherit them - which is corroborated by direct human evidence that APPL1 binds Akt2 and APPL2 does not. UniProtKB:Q8NEU8 · APPL2 (Homo sapiens), the target SUPPORTS TRANSFER The target's own accession among the seeds, which is expected rather than circular: APPL2 has its own experimental signalling annotations (the TAS row from PMID:15016378 and the IMP rows from PMID:24879834 and PMID:26583432), and those descendant evidences are part of what the PAINT curator used to place the node. UniProtKB:Q9UKG1 · APPL1 (Homo sapiens) SUPPORTS TRANSFER Human APPL1, the paralog. It supports the generic parent term; it would not support a signed child, since its effect on adiponectin/Akt signalling is the opposite of APPL2's. Supporting Evidence: PMID:19661063 APPL2, an isoform of APPL1 that forms a dimer with APPL1, can interacts with both AdipoR1 and AdipoR2 and acts as a negative regulator of adiponectin signaling in muscle cells. PMID:27219021 By depleting cells of each adaptor respectively we show separate and opposing functions for APPL1 and 2 in Akt and MAPK signaling. PMID:19661063 APPL2 also suppresses adiponectin and insulin signaling by sequestrating APPL1 from these two pathways. |
| GO:0030659 cytoplasmic vesicle membrane | IEA GO_REF:0000117 | KEEP AS NON CORE | Summary: Broad ARBA rule giving a generic parent of the vesicle compartments APPL2 actually occupies. Reason: ARBA00026540 has 72 condition sets across eukaryotes. The term is true - APPL2 is on endosomal, phagosomal and macropinosomal membranes - but it is the uninformative parent of rows that already say which vesicle. Kept rather than marked over-annotated because, unlike the plasma-membrane and bounding-membrane rules, this one lands on a compartment class APPL2 genuinely belongs to. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: ARBA:ARBA00026540 · ARBA rule ARBA00026540 SUPPORTS TRANSFER ARBA rule for cytoplasmic vesicle membrane; 72 condition sets across eukaryotes. Broad, and the term is a generic parent of the compartments APPL2 is actually on. Supporting Evidence: PMID:21645192 APPL endosomes are a recently identified subpopulation of early endosomes characterized by the presence of two homologous Rab5 effector proteins APPL1 and APPL2. PMID:25568335 APPL2 was immunolabeled in the cytoplasm, on cell surface ruffles, and on early, actin-rich phagosomes |
| GO:0030670 phagocytic vesicle membrane | IEA GO_REF:0000044 | KEEP AS NON CORE | Summary: Phagosome membrane keyword projection, ultimately from mouse macrophage imaging. Reason: UniProt's Phagosome membrane keyword for APPL2 is a by-similarity annotation from mouse Appl2, which was imaged on early, actin-rich phagosomes in primary macrophages and RAW264.7 cells, recruited there by GTP-Rab31. The location is real but macrophage-specific and, in macrophages, APPL2 is notably absent from endosomes - so this is a distinct cellular context rather than an extension of the endosomal pool. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB-SubCell:SL-0205 · UniProt subcellular-location keyword: Phagosome membrane SUPPORTS TRANSFER Projection of UniProt's own curated subcellular-location keyword, so this row re-states the Swiss-Prot annotation rather than adding independent evidence. The underlying observation is APPL2 on early phagosomes in mouse macrophages (PMID:25568335), recruited by GTP-Rab31. Supporting Evidence: PMID:25568335 APPL2 was immunolabeled in the cytoplasm, on cell surface ruffles, and on early, actin-rich phagosomes PMID:25568335 This is the first evidence that APPL2 is associated with early phagosomes. PMID:25568335 This contrasts with the predominant and stronger labeling of mCherry-APPL2 on phagosomes and its general absence from endosomes |
| GO:0031410 cytoplasmic vesicle | IDA PMID:18034774 Membrane targeting by APPL1 and APPL2: dynamic scaffolds tha... | KEEP AS NON CORE | Summary: APPL2 fusion proteins on dynamic cytosolic vesicles. Reason: Live imaging showed APPL2-YFP on cytosolic membrane structures undergoing movement, fusion and fission and recruiting endogenous RAB5 - the vesicles are the APPL endosomes. The term is correct but is the generic parent of the endosome rows that identify the vesicle. Supporting Evidence: PMID:18034774 Both APPL1-YFP and APPL2-YFP showed dynamic associations with cytosolic membrane structures that underwent movement, fusion and fission events PMID:18034774 APPL2-YFP colocalized strongly with RAB5 to smaller cytosolic vesicular structures |
| GO:0031901 early endosome membrane | EXP PMID:15016378 APPL proteins link Rab5 to nuclear signal transduction via a... | ACCEPT | Summary: The most specific correct compartment term for APPL2, from the founding characterisation. Reason: APPL endosomes are, by the authors' own definition, a subpopulation of early endosomes that shows only limited colocalisation with EEA1. UniProt records Early endosome membrane with experimental evidence from this paper. Since GO has no 'APPL endosome' term, this is the most specific true statement available and it is core. Supporting Evidence: PMID:15016378 This pathway operates via APPL1 and APPL2, two Rab5 effectors, which reside on a subpopulation of endosomes. PMID:21645192 APPL endosomes are a recently identified subpopulation of early endosomes characterized by the presence of two homologous Rab5 effector proteins APPL1 and APPL2. PMID:21645192 They exhibit only limited colocalization with EEA1, another Rab5 effector and a marker of the canonical early endosomes. |
| GO:0031901 early endosome membrane | IEA GO_REF:0000044 | ACCEPT | Summary: Projection of UniProt's Early endosome membrane keyword. Reason: Restates the curated UniProt location behind the EXP row. Redundant, correct. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB-SubCell:SL-0093 · UniProt subcellular-location keyword: Early endosome membrane SUPPORTS TRANSFER Projection of UniProt's own curated subcellular-location keyword, so this row re-states the Swiss-Prot annotation rather than adding independent evidence. Supporting Evidence: PMID:21645192 APPL endosomes are a recently identified subpopulation of early endosomes characterized by the presence of two homologous Rab5 effector proteins APPL1 and APPL2. |
| GO:0031982 vesicle | IDA PMID:21645192 Biochemical characterization of APPL endosomes: the role of ... | KEEP AS NON CORE | Summary: Generic vesicle term from the APPL-endosome fractionation study. Reason: The same study that established the endosome and endosome-membrane rows. 'Vesicle' is the top-level parent and adds nothing to them, but it is not wrong. Supporting Evidence: PMID:21645192 APPL endosomes are a recently identified subpopulation of early endosomes characterized by the presence of two homologous Rab5 effector proteins APPL1 and APPL2. PMID:21645192 They exhibit only limited colocalization with EEA1, another Rab5 effector and a marker of the canonical early endosomes. |
| GO:0032009 early phagosome | IEA GO_REF:0000107 | ACCEPT | Summary: APPL2 on early phagosomes in macrophages, transferred from mouse. Reason: Rab31-GTP recruits APPL2 to early-stage phagosomes in mouse macrophages, where the two colocalise around the phagosome circumference; the authors call this the first evidence that APPL2 is associated with early phagosomes. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Accepted as core rather than peripheral: this is precisely where APPL2 performs the phagocytic-cup-closure and Fc-gamma-receptor signalling functions that are themselves accepted as core, and a location is not made peripheral by being cell-type-restricted when the activity carried out there is central. Note that in macrophages this pool replaces rather than supplements the endosomal one - APPL2 is largely absent from endosomes in that cell type. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse observation is GTP-Rab31-dependent recruitment of APPL2 to early phagosomes in primary macrophages and RAW264.7 cells (PMID:25568335). ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. Supporting Evidence: PMID:25568335 This is the first evidence that APPL2 is associated with early phagosomes. PMID:25568335 During early phagocytosis, we find that Rab31 recruits the signaling adaptor APPL2. |
| GO:0032009 early phagosome | ISS GO_REF:0000024 | ACCEPT | Summary: APPL2 on early phagosomes in macrophages, transferred from mouse. Reason: Rab31-GTP recruits APPL2 to early-stage phagosomes in mouse macrophages, where the two colocalise around the phagosome circumference; the authors call this the first evidence that APPL2 is associated with early phagosomes. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Accepted as core rather than peripheral: this is precisely where APPL2 performs the phagocytic-cup-closure and Fc-gamma-receptor signalling functions that are themselves accepted as core, and a location is not made peripheral by being cell-type-restricted when the activity carried out there is central. Note that in macrophages this pool replaces rather than supplements the endosomal one - APPL2 is largely absent from endosomes in that cell type. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse observation is GTP-Rab31-dependent recruitment of APPL2 to early phagosomes in primary macrophages and RAW264.7 cells (PMID:25568335). Supporting Evidence: PMID:25568335 This is the first evidence that APPL2 is associated with early phagosomes. PMID:25568335 During early phagocytosis, we find that Rab31 recruits the signaling adaptor APPL2. |
| GO:0032587 ruffle membrane | IEA GO_REF:0000120 | KEEP AS NON CORE | Summary: APPL2 is on cell-surface ruffles in macrophages; the ruffle membrane call is transferred from mouse Appl2. Reason: Endogenous and GFP-tagged APPL2 were immunolabelled on cell-surface ruffles in mouse primary macrophages and RAW264.7 cells (PMID:25568335), which is the observation behind the mouse annotation. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. It is kept as non-core because it is a macrophage-specific membrane domain seen during phagocytosis, not where APPL2 sits in the cell types where its signalling functions were characterised. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse observation is immunofluorescence of endogenous and tagged APPL2 on macrophage surface ruffles (PMID:25568335); it transfers, but only as a cell-type-specific location. ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. UniProtKB-SubCell:SL-0301 · UniProt subcellular-location keyword: Ruffle membrane SUPPORTS TRANSFER Projection of UniProt's own curated subcellular-location keyword, so this row re-states the Swiss-Prot annotation rather than adding independent evidence. The mouse observation is immunofluorescence of endogenous and tagged APPL2 on macrophage surface ruffles (PMID:25568335); it transfers, but only as a cell-type-specific location. Supporting Evidence: PMID:25568335 APPL2 was immunolabeled in the cytoplasm, on cell surface ruffles, and on early, actin-rich phagosomes file:human/APPL2/APPL2-bioinformatics/RESULTS.md Human and mouse APPL2 are 92.7% identical over a full-length global alignment |
| GO:0032587 ruffle membrane | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: APPL2 is on cell-surface ruffles in macrophages; the ruffle membrane call is transferred from mouse Appl2. Reason: Endogenous and GFP-tagged APPL2 were immunolabelled on cell-surface ruffles in mouse primary macrophages and RAW264.7 cells (PMID:25568335), which is the observation behind the mouse annotation. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. It is kept as non-core because it is a macrophage-specific membrane domain seen during phagocytosis, not where APPL2 sits in the cell types where its signalling functions were characterised. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse observation is immunofluorescence of endogenous and tagged APPL2 on macrophage surface ruffles (PMID:25568335); it transfers, but only as a cell-type-specific location. Supporting Evidence: PMID:25568335 APPL2 was immunolabeled in the cytoplasm, on cell surface ruffles, and on early, actin-rich phagosomes file:human/APPL2/APPL2-bioinformatics/RESULTS.md Human and mouse APPL2 are 92.7% identical over a full-length global alignment |
| GO:0033211 adiponectin-activated signaling pathway | IEA GO_REF:0000107 | ACCEPT | Summary: APPL2 acts in the adiponectin-receptor pathway as APPL1's antagonist; the GO term is sign-neutral and therefore correct. Reason: APPL2 binds AdipoR1 and AdipoR2 through its BAR domain, competes with APPL1 for the receptors and sequesters APPL1 away from the pathway; knocking APPL2 down enhances adiponectin-stimulated glucose uptake and fatty acid oxidation, and overexpressing it does the reverse. This is a core function. The sign-inversion risk here is real but does not bite: GO:0033211 is defined purely as the series of signals initiated by adiponectin binding its receptor, with no direction, so involvement is all that is asserted and involvement is exactly what the mouse data show. A QuickGO search finds no signed child of this term - there is no 'negative regulation of adiponectin-activated signaling pathway' - which is recorded under proposed_new_terms. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse evidence (PMID:19661063) is APPL2-specific in both directions (RNAi and overexpression in C2C12 myotubes) and the transferred term carries no regulatory sign, so the negative direction of APPL2's effect cannot be inverted by this transfer. ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. Supporting Evidence: PMID:19661063 APPL2, an isoform of APPL1 that forms a dimer with APPL1, can interacts with both AdipoR1 and AdipoR2 and acts as a negative regulator of adiponectin signaling in muscle cells. PMID:19661063 suppressing APPL2 expression by RNAi significantly enhances adiponectin-stimulated glucose uptake and fatty acid oxidation. PMID:19661063 APPL2 also suppresses adiponectin and insulin signaling by sequestrating APPL1 from these two pathways. |
| GO:0033211 adiponectin-activated signaling pathway | ISS GO_REF:0000024 | ACCEPT | Summary: APPL2 acts in the adiponectin-receptor pathway as APPL1's antagonist; the GO term is sign-neutral and therefore correct. Reason: APPL2 binds AdipoR1 and AdipoR2 through its BAR domain, competes with APPL1 for the receptors and sequesters APPL1 away from the pathway; knocking APPL2 down enhances adiponectin-stimulated glucose uptake and fatty acid oxidation, and overexpressing it does the reverse. This is a core function. The sign-inversion risk here is real but does not bite: GO:0033211 is defined purely as the series of signals initiated by adiponectin binding its receptor, with no direction, so involvement is all that is asserted and involvement is exactly what the mouse data show. A QuickGO search finds no signed child of this term - there is no 'negative regulation of adiponectin-activated signaling pathway' - which is recorded under proposed_new_terms. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse evidence (PMID:19661063) is APPL2-specific in both directions (RNAi and overexpression in C2C12 myotubes) and the transferred term carries no regulatory sign, so the negative direction of APPL2's effect cannot be inverted by this transfer. Supporting Evidence: PMID:19661063 APPL2, an isoform of APPL1 that forms a dimer with APPL1, can interacts with both AdipoR1 and AdipoR2 and acts as a negative regulator of adiponectin signaling in muscle cells. PMID:19661063 suppressing APPL2 expression by RNAi significantly enhances adiponectin-stimulated glucose uptake and fatty acid oxidation. PMID:19661063 APPL2 also suppresses adiponectin and insulin signaling by sequestrating APPL1 from these two pathways. |
| GO:0034143 regulation of toll-like receptor 4 signaling pathway | IEA GO_REF:0000107 | ACCEPT | Summary: APPL2 regulates TLR4 signalling in macrophages, with each adaptor depleted separately. Reason: Depleting APPL1 and APPL2 individually revealed separate and opposing functions in Akt and MAPK signalling downstream of LPS/TLR4, with APPL2 dominant in NF-kappaB p65 nuclear translocation and in constraining cytokine secretion. Crucially the adaptors were depleted 'respectively', so this is an APPL2-specific result, not a pooled one. Reinforced by the Appl2 knockout mouse, which is hypersensitive to LPS. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Core. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. Mouse macrophage evidence with each adaptor depleted separately (PMID:27219021), plus an Appl2 knockout mouse phenotype (PMID:25328665). ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. Supporting Evidence: PMID:27219021 By depleting cells of each adaptor respectively we show separate and opposing functions for APPL1 and 2 in Akt and MAPK signaling. PMID:27219021 APPL2 has a dominant role in nuclear translocation of NF-KB p65 and it serves to constrain the secretion of pro- and anti-inflammatory cytokines. PMID:25328665 When challenged with lipopolysaccharides (LPS), Appl2 KO mice exhibited more severe symptoms of endotoxin shock, accompanied by increased production of proinflammatory cytokines. |
| GO:0034143 regulation of toll-like receptor 4 signaling pathway | ISS GO_REF:0000024 | ACCEPT | Summary: APPL2 regulates TLR4 signalling in macrophages, with each adaptor depleted separately. Reason: Depleting APPL1 and APPL2 individually revealed separate and opposing functions in Akt and MAPK signalling downstream of LPS/TLR4, with APPL2 dominant in NF-kappaB p65 nuclear translocation and in constraining cytokine secretion. Crucially the adaptors were depleted 'respectively', so this is an APPL2-specific result, not a pooled one. Reinforced by the Appl2 knockout mouse, which is hypersensitive to LPS. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Core. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. Mouse macrophage evidence with each adaptor depleted separately (PMID:27219021), plus an Appl2 knockout mouse phenotype (PMID:25328665). Supporting Evidence: PMID:27219021 By depleting cells of each adaptor respectively we show separate and opposing functions for APPL1 and 2 in Akt and MAPK signaling. PMID:27219021 APPL2 has a dominant role in nuclear translocation of NF-KB p65 and it serves to constrain the secretion of pro- and anti-inflammatory cytokines. PMID:25328665 When challenged with lipopolysaccharides (LPS), Appl2 KO mice exhibited more severe symptoms of endotoxin shock, accompanied by increased production of proinflammatory cytokines. |
| GO:0035091 phosphatidylinositol binding | IDA PMID:18034774 Membrane targeting by APPL1 and APPL2: dynamic scaffolds tha... | ACCEPT | Summary: Full-length APPL2 and its isolated PH and PTB domains bind phosphoinositides in vitro. Reason: Lipid-strip overlays with affinity-purified protein show binding of full-length APPL2 - not just isolated domains - to PtdIns(3)P, PtdIns(4)P, PtdIns(5)P, PtdIns(3,4)P2 and PtdIns(3,5)P2, and UniProt records that the PH and PID domains mediate phosphoinositide binding. The structural work nominates a PH-domain contact surface, and I verified that those residues are present at the stated positions in Q8NEU8 (R287, K289, W297, all inside the PH span 277-375) and conserved in mouse Appl2. Core: lipid binding together with Rab binding is what puts APPL2 on membranes. Supporting Evidence: PMID:18034774 APPL1 and APPL2 full-length proteins, isolated PH domains and isolated PTB domains all bound to membrane-immobilized PtdIns(3)P, PtdIns(4)P, PtdIns(5)P, PtdIns(3,4)P 2 and PtdIns(3,5)P 2 PMID:23055524 hAPPL2 BARPH has Lys-289, Arg-287, and Trp-297 file:human/APPL2/APPL2-bioinformatics/RESULTS.md The patch is intact in human APPL2 and conserved in mouse Appl2 |
| GO:0035091 phosphatidylinositol binding | IEA GO_REF:0000117 | ACCEPT | Summary: Narrow ARBA rule agreeing with the IDA row. Reason: ARBA00026346 has only 7 condition sets and assigns the term the IDA row already establishes for full-length protein. Redundant but correct. Propagation Review Root cause: NO FAILURE CORE Sources checked: ARBA:ARBA00026346 · ARBA rule ARBA00026346 SUPPORTS TRANSFER ARBA rule for phosphatidylinositol binding; 7 condition sets across eukaryotes. Narrow, and it reproduces the IDA row's claim. Supporting Evidence: PMID:18034774 APPL1 and APPL2 full-length proteins, isolated PH domains and isolated PTB domains all bound to membrane-immobilized PtdIns(3)P, PtdIns(4)P, PtdIns(5)P, PtdIns(3,4)P 2 and PtdIns(3,5)P 2 |
| GO:0035729 cellular response to hepatocyte growth factor stimulus | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: cellular response to hepatocyte growth factor stimulus, from mouse embryonic fibroblasts where a single Appl2 knockout has the phenotype. Reason: The title of the source paper reads as a double-knockout result, and I checked whether it is: it is not. Single Appl1 KO, single Appl2 KO and the double KO all showed decreased HGF-induced Akt activation, so there is an APPL2-specific loss-of-function anchor here and the rows do not suffer the paralog conflation that the TGF-beta rows do. The migration and invasion defects are reported for Appl-deficient MEFs collectively, so the migration row is slightly weaker than the HGF-response row. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Non-core because the effect is growth-factor-selective - EGF, insulin and serum activate Akt normally - and Appl2-null mice are viable and grow normally. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse annotation has a single-Appl2-knockout anchor (decreased HGF-induced Akt activation in Appl2 KO MEFs), not only the double knockout the title implies. ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. Supporting Evidence: PMID:26445298 Appl1 KO, Appl2 KO, and especially Appl1/2 DKO MEFs showed consistently decreased Akt activation upon stimulation with HGF PMID:26445298 Appl1/2-null mouse embryonic fibroblasts exhibited defects in HGF-induced Akt activation, migration, and invasion. PMID:26445298 ubiquitous Appl2 knockout (Appl2-/-) mice, much like Appl1-/- mice, are viable and grow normally to adulthood |
| GO:0035729 cellular response to hepatocyte growth factor stimulus | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: cellular response to hepatocyte growth factor stimulus, from mouse embryonic fibroblasts where a single Appl2 knockout has the phenotype. Reason: The title of the source paper reads as a double-knockout result, and I checked whether it is: it is not. Single Appl1 KO, single Appl2 KO and the double KO all showed decreased HGF-induced Akt activation, so there is an APPL2-specific loss-of-function anchor here and the rows do not suffer the paralog conflation that the TGF-beta rows do. The migration and invasion defects are reported for Appl-deficient MEFs collectively, so the migration row is slightly weaker than the HGF-response row. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Non-core because the effect is growth-factor-selective - EGF, insulin and serum activate Akt normally - and Appl2-null mice are viable and grow normally. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse annotation has a single-Appl2-knockout anchor (decreased HGF-induced Akt activation in Appl2 KO MEFs), not only the double knockout the title implies. Supporting Evidence: PMID:26445298 Appl1 KO, Appl2 KO, and especially Appl1/2 DKO MEFs showed consistently decreased Akt activation upon stimulation with HGF PMID:26445298 Appl1/2-null mouse embryonic fibroblasts exhibited defects in HGF-induced Akt activation, migration, and invasion. PMID:26445298 ubiquitous Appl2 knockout (Appl2-/-) mice, much like Appl1-/- mice, are viable and grow normally to adulthood |
| GO:0036186 early phagosome membrane | IEA GO_REF:0000107 | ACCEPT | Summary: APPL2 on early phagosome membranes in macrophages, transferred from mouse. Reason: Rab31-GTP recruits APPL2 to early-stage phagosomes in mouse macrophages, where the two colocalise around the phagosome circumference; the authors call this the first evidence that APPL2 is associated with early phagosomes. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Accepted as core rather than peripheral: this is precisely where APPL2 performs the phagocytic-cup-closure and Fc-gamma-receptor signalling functions that are themselves accepted as core, and a location is not made peripheral by being cell-type-restricted when the activity carried out there is central. Note that in macrophages this pool replaces rather than supplements the endosomal one - APPL2 is largely absent from endosomes in that cell type. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse observation is GTP-Rab31-dependent recruitment of APPL2 to early phagosomes in primary macrophages and RAW264.7 cells (PMID:25568335). ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. Supporting Evidence: PMID:25568335 This is the first evidence that APPL2 is associated with early phagosomes. PMID:25568335 APPL2 was immunolabeled in the cytoplasm, on cell surface ruffles, and on early, actin-rich phagosomes |
| GO:0036186 early phagosome membrane | ISS GO_REF:0000024 | ACCEPT | Summary: APPL2 on early phagosome membranes in macrophages, transferred from mouse. Reason: Rab31-GTP recruits APPL2 to early-stage phagosomes in mouse macrophages, where the two colocalise around the phagosome circumference; the authors call this the first evidence that APPL2 is associated with early phagosomes. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Accepted as core rather than peripheral: this is precisely where APPL2 performs the phagocytic-cup-closure and Fc-gamma-receptor signalling functions that are themselves accepted as core, and a location is not made peripheral by being cell-type-restricted when the activity carried out there is central. Note that in macrophages this pool replaces rather than supplements the endosomal one - APPL2 is largely absent from endosomes in that cell type. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse observation is GTP-Rab31-dependent recruitment of APPL2 to early phagosomes in primary macrophages and RAW264.7 cells (PMID:25568335). Supporting Evidence: PMID:25568335 This is the first evidence that APPL2 is associated with early phagosomes. PMID:25568335 APPL2 was immunolabeled in the cytoplasm, on cell surface ruffles, and on early, actin-rich phagosomes |
| GO:0042592 homeostatic process | IEA GO_REF:0000117 | MARK AS OVER ANNOTATED | Summary: 'Homeostatic process' is a content-free term reached by a rule spanning bacteria to plants. Reason: ARBA00027526 has 135 condition sets covering bacteria, fungi, plants and metazoa. APPL2 does act in homeostasis - glucose homeostasis and adaptive thermogenesis are both annotated - but those specific rows are what carry the claim. A rule this promiscuous attached to a term this general is the clearest over-annotation in the record. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: GRANULARITY MISMATCH Sources checked: ARBA:ARBA00027526 · ARBA rule ARBA00027526 SUPPORTS TRANSFER ARBA rule for homeostatic process; 135 condition sets spanning bacteria, plants, fungi and metazoa. A rule this broad attached to a term this general carries no information about APPL2. Supporting Evidence: PMID:24879834 conditional deletion of APPL2 in skeletal muscles enhances insulin sensitivity, leading to an improvement in glucose tolerance. PMID:29467283 Genetic ablation of APPL2 in RIP-Cre neurons diminishes beiging in sWAT without affecting BAT, leading to cold intolerance and obesity in mice. |
| GO:0042593 glucose homeostasis | IMP PMID:24879834 The adaptor protein APPL2 inhibits insulin-stimulated glucos... | ACCEPT | Summary: Muscle-specific APPL2 deletion improves glucose tolerance in mice. Reason: Conditional deletion of APPL2 in skeletal muscle enhanced insulin sensitivity and improved glucose tolerance, and the cellular phenotype runs in both directions - overexpression impairs and knockdown enhances insulin-evoked GLUT4 recruitment and glucose uptake. The mechanism is mapped to a phospho-Ser235-dependent TBC1D1 interaction. This is the best-evidenced APPL2 function in the record and is core. Supporting Evidence: PMID:24879834 conditional deletion of APPL2 in skeletal muscles enhances insulin sensitivity, leading to an improvement in glucose tolerance. PMID:24879834 insulin-evoked plasma membrane recruitment of GLUT4 and glucose uptake are impaired by APPL2 overexpression but enhanced by APPL2 knockdown. PMID:24879834 Insulin stimulates TBC1D1 phosphorylation on serine 235, leading to enhanced interaction with the BAR domain of APPL2, which in turn suppresses insulin-evoked TBC1D1 phosphorylation on threonine 596 in cultured myotubes and skeletal muscle. |
| GO:0042802 identical protein binding | IPI PMID:18034774 Membrane targeting by APPL1 and APPL2: dynamic scaffolds tha... | ACCEPT | Summary: APPL2 self-association, demonstrated by reciprocal coimmunoprecipitation of differently tagged APPL2. Reason: Differently tagged APPL2 constructs reciprocally coimmunoprecipitated, and yeast two-hybrid mapped the requirement to the minimal BAR domain. The WITH/FROM is APPL2's own accession, which is what an identical-protein-binding row should carry. The BAR dimer is the structural unit of the protein - the crystal structure shows the crescent-shaped antiparallel BAR dimer and solution methods agree - so this is core. Supporting Evidence: PMID:18034774 APPL2-V5-His specifically coimmunoprecipitated with APPL2-YFP and vice versa PMID:18034774 the APPL minimal BAR domains were necessary and sufficient for mediating APPL1-APPL1, APPL2-APPL2 and APPL1-APPL2 interactions in the yeast two-hybrid system PMID:23055524 Each dimer is crescent-shaped and formed by an anti-parallel arrangement of the helical BAR domains of each hAPPL2 BARPH monomer. |
| GO:0042803 protein homodimerization activity | IDA PMID:18034774 Membrane targeting by APPL1 and APPL2: dynamic scaffolds tha... | ACCEPT | Summary: BAR-domain-mediated homodimerization is the structural basis of APPL2 function. Reason: The minimal BAR domains are necessary and sufficient for APPL2-APPL2 interaction, the crystal structure shows a crescent-shaped antiparallel BAR dimer, and independent solution measurements - chemical cross-linking and multi-angle light scattering - both give a dimer. The dimer is also the functional unit for Rab binding, since two Rab31 molecules bind one APPL2 BAR-PH dimer. Core. Supporting Evidence: PMID:18034774 the APPL minimal BAR domains were necessary and sufficient for mediating APPL1-APPL1, APPL2-APPL2 and APPL1-APPL2 interactions in the yeast two-hybrid system PMID:23055524 Each dimer is crescent-shaped and formed by an anti-parallel arrangement of the helical BAR domains of each hAPPL2 BARPH monomer. PMID:23055524 hAPPL2 BARPH is predominantly a dimer in solution |
| GO:0044354 macropinosome | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: Macropinosome localisation in LPS-activated mouse macrophages. Reason: APPL1 and APPL2 are differentially localised to distinct signalling-competent membrane domains on the surface and in endocytic compartments of LPS-activated macrophages, which is the basis for the mouse macropinosome annotation. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Non-core: it is one membrane domain in one activated cell type. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse annotation rests on differential localisation of APPL1 and APPL2 to distinct membrane domains in LPS-activated macrophages (PMID:27219021). ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. Supporting Evidence: PMID:27219021 APPL1 and 2 are differentially localized to distinct signaling-competent membrane domains on the surface and in endocytic compartments of LPS-activated macrophages. PMID:25568335 APPL2 was immunolabeled in the cytoplasm, on cell surface ruffles, and on early, actin-rich phagosomes |
| GO:0044354 macropinosome | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: Macropinosome localisation in LPS-activated mouse macrophages. Reason: APPL1 and APPL2 are differentially localised to distinct signalling-competent membrane domains on the surface and in endocytic compartments of LPS-activated macrophages, which is the basis for the mouse macropinosome annotation. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Non-core: it is one membrane domain in one activated cell type. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse annotation rests on differential localisation of APPL1 and APPL2 to distinct membrane domains in LPS-activated macrophages (PMID:27219021). Supporting Evidence: PMID:27219021 APPL1 and 2 are differentially localized to distinct signaling-competent membrane domains on the surface and in endocytic compartments of LPS-activated macrophages. PMID:25568335 APPL2 was immunolabeled in the cytoplasm, on cell surface ruffles, and on early, actin-rich phagosomes |
| GO:0044877 protein-containing complex binding | IPI PMID:15016378 APPL proteins link Rab5 to nuclear signal transduction via a... | ACCEPT | Summary: Binding to the MBD2/NuRD nucleosome remodelling and deacetylase complex (ComplexPortal CPX-880). Reason: The WITH/FROM resolves to CPX-880, the MBD2/NuRD complex, and UniProt records that APPL2 binds subunits of the NuRD/MeCP1 complex with experimental evidence from this paper on the Q8NEU8 entry. The cached record for this reference is abstract-only and the abstract foregrounds APPL1's nuclear translocation, but the curator and the UniProt annotator read the full text, and the nuclear APPL2 pool is independently supported. Unlike a bare protein-binding row this one names a defined complex, which makes it informative, and it connects the endosomal and nuclear halves of APPL2 biology. Supporting Evidence: PMID:15016378 APPL1 translocates from the membranes to the nucleus where it interacts with the nucleosome remodeling and histone deacetylase multiprotein complex NuRD/MeCP1 file:human/APPL2/APPL2-uniprot.txt CC subunits of the NuRD/MeCP1 complex (PubMed:15016378). Interacts with PMID:26583432 both APPL1 and APPL2 translocate into the nucleus from the cytosol in response to EGF stimulation, to regulate cell proliferation |
| GO:0045088 regulation of innate immune response | IEA GO_REF:0000107 | ACCEPT | Summary: Appl2 knockout mice are hypersensitive to endotoxin; APPL2 restrains the innate immune response. Reason: A whole-animal Appl2 knockout gave more severe endotoxin shock with increased proinflammatory cytokines, and Appl2-null primary macrophages produced more TNF-alpha and IL-1beta with enhanced Akt and NF-kappaB phosphorylation; the authors trace this to an Appl2-Appl1-p85alpha complex restraining PI3K/Akt. Clean, APPL2-specific loss of function. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Core. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The donor annotation rests on an Appl2 knockout mouse and Appl2-null primary macrophages (PMID:25328665) - an APPL2-specific loss of function, not a paralog-pooled one. ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. Supporting Evidence: PMID:25328665 When challenged with lipopolysaccharides (LPS), Appl2 KO mice exhibited more severe symptoms of endotoxin shock, accompanied by increased production of proinflammatory cytokines. PMID:25328665 deletion of Appl2 led to higher levels of TNF-α and IL-1β in primary macrophages PMID:25328665 Appl2 is a critical negative regulator of innate immune response via inhibition of PI3K/Akt/NF-κB signaling pathway by forming a complex with Appl1 and PI3K. |
| GO:0045088 regulation of innate immune response | ISS GO_REF:0000024 | ACCEPT | Summary: Appl2 knockout mice are hypersensitive to endotoxin; APPL2 restrains the innate immune response. Reason: A whole-animal Appl2 knockout gave more severe endotoxin shock with increased proinflammatory cytokines, and Appl2-null primary macrophages produced more TNF-alpha and IL-1beta with enhanced Akt and NF-kappaB phosphorylation; the authors trace this to an Appl2-Appl1-p85alpha complex restraining PI3K/Akt. Clean, APPL2-specific loss of function. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Core. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The donor annotation rests on an Appl2 knockout mouse and Appl2-null primary macrophages (PMID:25328665) - an APPL2-specific loss of function, not a paralog-pooled one. Supporting Evidence: PMID:25328665 When challenged with lipopolysaccharides (LPS), Appl2 KO mice exhibited more severe symptoms of endotoxin shock, accompanied by increased production of proinflammatory cytokines. PMID:25328665 deletion of Appl2 led to higher levels of TNF-α and IL-1β in primary macrophages PMID:25328665 Appl2 is a critical negative regulator of innate immune response via inhibition of PI3K/Akt/NF-κB signaling pathway by forming a complex with Appl1 and PI3K. |
| GO:0045335 phagocytic vesicle | IEA GO_REF:0000044 | KEEP AS NON CORE | Summary: Phagocytic vesicle keyword projection, same macrophage origin as the phagosome-membrane row. Reason: Restates UniProt's Phagosome keyword, whose experimental basis is APPL2 on early phagosomes in mouse macrophages. Correct, cell-type-specific, and redundant with GO:0032009 early phagosome, which is the more precise term for the same observation. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB-SubCell:SL-0206 · UniProt subcellular-location keyword: Phagosome SUPPORTS TRANSFER Projection of UniProt's own curated subcellular-location keyword, so this row re-states the Swiss-Prot annotation rather than adding independent evidence. Same source as the phagosome-membrane keyword row; GO:0032009 early phagosome states it more precisely. Supporting Evidence: PMID:25568335 This is the first evidence that APPL2 is associated with early phagosomes. PMID:25568335 During early phagocytosis, we find that Rab31 recruits the signaling adaptor APPL2. |
| GO:0046322 negative regulation of fatty acid oxidation | IEA GO_REF:0000107 | ACCEPT | Summary: APPL2 restrains adiponectin-stimulated fatty acid oxidation in muscle cells. Reason: Suppressing APPL2 by RNAi significantly enhanced adiponectin-stimulated fatty acid oxidation, which is the loss-of-function direction that establishes the negative sign; overexpression produced the converse effect on adiponectin signalling. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Core, as part of the adiponectin antagonism. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. APPL2-specific RNAi in C2C12 myotubes gives the loss-of-function direction (PMID:19661063), so the negative sign is established rather than assumed. ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. Supporting Evidence: PMID:19661063 suppressing APPL2 expression by RNAi significantly enhances adiponectin-stimulated glucose uptake and fatty acid oxidation. PMID:19661063 APPL2, an isoform of APPL1 that forms a dimer with APPL1, can interacts with both AdipoR1 and AdipoR2 and acts as a negative regulator of adiponectin signaling in muscle cells. PMID:19661063 Overexpression of APPL2 inhibits the interaction between APPL1 and AdipoR1, leading to down-regulation of adiponectin signaling in C2C12 myotubes. |
| GO:0046322 negative regulation of fatty acid oxidation | ISS GO_REF:0000024 | ACCEPT | Summary: APPL2 restrains adiponectin-stimulated fatty acid oxidation in muscle cells. Reason: Suppressing APPL2 by RNAi significantly enhanced adiponectin-stimulated fatty acid oxidation, which is the loss-of-function direction that establishes the negative sign; overexpression produced the converse effect on adiponectin signalling. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Core, as part of the adiponectin antagonism. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. APPL2-specific RNAi in C2C12 myotubes gives the loss-of-function direction (PMID:19661063), so the negative sign is established rather than assumed. Supporting Evidence: PMID:19661063 suppressing APPL2 expression by RNAi significantly enhances adiponectin-stimulated glucose uptake and fatty acid oxidation. PMID:19661063 APPL2, an isoform of APPL1 that forms a dimer with APPL1, can interacts with both AdipoR1 and AdipoR2 and acts as a negative regulator of adiponectin signaling in muscle cells. PMID:19661063 Overexpression of APPL2 inhibits the interaction between APPL1 and AdipoR1, leading to down-regulation of adiponectin signaling in C2C12 myotubes. |
| GO:0046325 negative regulation of D-glucose import across plasma membrane | IEA GO_REF:0000120 | ACCEPT | Summary: Negative regulation of glucose import, agreeing with the human/mouse IMP row from the same study. Reason: APPL2 suppresses insulin-evoked GLUT4 translocation to the plasma membrane and glucose uptake, shown by overexpression, knockdown and a muscle-specific knockout. The signed term is correctly signed. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Core, and not merely redundant with the IMP row: it is the same conclusion reached through the ortholog. Propagation Review Root cause: NO FAILURE CORE Sources checked: ARBA:ARBA00093152 · ARBA rule ARBA00093152 SUPPORTS TRANSFER ARBA rule for negative regulation of D-glucose import; only 4 condition sets. Narrow, and it agrees with the IMP row. UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse donor annotation comes from the same study as the human IMP row, with bidirectional APPL2 perturbation and a muscle-specific knockout. ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. Supporting Evidence: PMID:24879834 insulin-evoked plasma membrane recruitment of GLUT4 and glucose uptake are impaired by APPL2 overexpression but enhanced by APPL2 knockdown. PMID:24879834 conditional deletion of APPL2 in skeletal muscles enhances insulin sensitivity, leading to an improvement in glucose tolerance. PMID:24879834 the APPL2-TBC1D1 interaction is a key step to fine tune insulin-stimulated glucose uptake by regulating the membrane recruitment of GLUT4 in skeletal muscle. |
| GO:0046325 negative regulation of D-glucose import across plasma membrane | IMP PMID:24879834 The adaptor protein APPL2 inhibits insulin-stimulated glucos... | ACCEPT | Summary: APPL2 overexpression impairs and knockdown enhances insulin-stimulated glucose uptake and GLUT4 recruitment. Reason: Both perturbation directions were done, in cultured myotubes and in vivo, and the mechanism is mapped to the BAR-domain interaction with phospho-Ser235 TBC1D1 that blocks TBC1D1 Thr-596 phosphorylation. Core function. Supporting Evidence: PMID:24879834 insulin-evoked plasma membrane recruitment of GLUT4 and glucose uptake are impaired by APPL2 overexpression but enhanced by APPL2 knockdown. PMID:24879834 Insulin stimulates TBC1D1 phosphorylation on serine 235, leading to enhanced interaction with the BAR domain of APPL2, which in turn suppresses insulin-evoked TBC1D1 phosphorylation on threonine 596 in cultured myotubes and skeletal muscle. PMID:24879834 conditional deletion of APPL2 in skeletal muscles enhances insulin sensitivity, leading to an improvement in glucose tolerance. |
| GO:0046325 negative regulation of D-glucose import across plasma membrane | ISS GO_REF:0000024 | ACCEPT | Summary: Negative regulation of glucose import, agreeing with the human/mouse IMP row from the same study. Reason: APPL2 suppresses insulin-evoked GLUT4 translocation to the plasma membrane and glucose uptake, shown by overexpression, knockdown and a muscle-specific knockout. The signed term is correctly signed. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Core, and not merely redundant with the IMP row: it is the same conclusion reached through the ortholog. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse donor annotation comes from the same study as the human IMP row, with bidirectional APPL2 perturbation and a muscle-specific knockout. Supporting Evidence: PMID:24879834 insulin-evoked plasma membrane recruitment of GLUT4 and glucose uptake are impaired by APPL2 overexpression but enhanced by APPL2 knockdown. PMID:24879834 conditional deletion of APPL2 in skeletal muscles enhances insulin sensitivity, leading to an improvement in glucose tolerance. PMID:24879834 the APPL2-TBC1D1 interaction is a key step to fine tune insulin-stimulated glucose uptake by regulating the membrane recruitment of GLUT4 in skeletal muscle. |
| GO:0050768 negative regulation of neurogenesis | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: negative regulation of neurogenesis, driven mainly by APPL2 transgenic overexpression in mouse. Reason: The in vivo evidence is an APPL2 transgenic mouse: APPL2 Tg animals have reduced hippocampal and olfactory neurogenesis and a higher glial fraction, reversible by a glucocorticoid-receptor antagonist. A matching loss-of-function exists only in cultured neural stem cells, where APPL2 knockdown promoted neurogenesis. So both directions exist and the negative sign is right, but the animal-level claim rests on gain of function, where an overexpressed adaptor can titrate partners non-physiologically. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Non-core: this is a tissue-specific developmental consequence of APPL2 tuning glucocorticoid-receptor sensitivity, several steps from the molecular function. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The donor evidence is chiefly an APPL2 transgenic overexpression mouse (PMID:28965332, PMID:32468397), with the loss-of-function direction shown only in vitro. ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. Supporting Evidence: PMID:28965332 APPL2 Tg mice had decreased hippocampal neurogenesis that was reversed by GR antagonist RU486. PMID:32468397 APPL2 overexpression resulted in NSCs switching from neuronal differentiation to gliogenesis while APPL2 knockdown promoted neurogenesis. PMID:32468397 APPL2 Tg mice had a higher population of glial cells and dampened neuronal production in the olfactory system |
| GO:0050768 negative regulation of neurogenesis | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: negative regulation of neurogenesis, driven mainly by APPL2 transgenic overexpression in mouse. Reason: The in vivo evidence is an APPL2 transgenic mouse: APPL2 Tg animals have reduced hippocampal and olfactory neurogenesis and a higher glial fraction, reversible by a glucocorticoid-receptor antagonist. A matching loss-of-function exists only in cultured neural stem cells, where APPL2 knockdown promoted neurogenesis. So both directions exist and the negative sign is right, but the animal-level claim rests on gain of function, where an overexpressed adaptor can titrate partners non-physiologically. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Non-core: this is a tissue-specific developmental consequence of APPL2 tuning glucocorticoid-receptor sensitivity, several steps from the molecular function. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The donor evidence is chiefly an APPL2 transgenic overexpression mouse (PMID:28965332, PMID:32468397), with the loss-of-function direction shown only in vitro. Supporting Evidence: PMID:28965332 APPL2 Tg mice had decreased hippocampal neurogenesis that was reversed by GR antagonist RU486. PMID:32468397 APPL2 overexpression resulted in NSCs switching from neuronal differentiation to gliogenesis while APPL2 knockdown promoted neurogenesis. PMID:32468397 APPL2 Tg mice had a higher population of glial cells and dampened neuronal production in the olfactory system |
| GO:0051289 protein homotetramerization | IDA PMID:23055524 Membrane curvature protein exhibits interdomain flexibility ... | MODIFY | Summary: The cited paper reports a dimer by every solution method; the tetramer is a reading of the crystallographic asymmetric unit. Reason: The word 'tetramer' does not occur anywhere in the full text of this paper. What it reports is that the APPL2 BAR-PH construct crystallised with two dimers in the asymmetric unit - four chains, which is a statement about crystal packing, not about the assembly state in solution. Every solution measurement in the same paper gives a dimer: chemical cross-linking, and multi-angle light scattering returning 95.1 kDa against a theoretical dimer mass of 92.1 kDa. The only tetramer described is a heterotetramer with the Rab (an APPL2 BAR-PH dimer plus two Rab31 monomers), which is not homotetramerization either. UniProt's 'Homotetramer (PubMed:23055524)' appears to rest on the same misreading. I propose GO:0051260 protein homooligomerization rather than deletion: the BAR domain does self-associate, and the defined oligomer the paper supports is the dimer, already annotated as GO:0042803 and GO:0042802. I am overruling a curator here only because the cached full text is complete and the decisive sentences are quotable verbatim. Proposed replacements: protein homooligomerization Supporting Evidence: PMID:23055524 hAPPL2 BARPH crystallized in the P 2 1 2 1 2 1 space group with two dimers in the asymmetric unit. PMID:23055524 hAPPL2 BARPH is predominantly a dimer in solution PMID:23055524 95.1 kDa ± 2.0% for hAPPL2 BARPH (theoretical mass of 92.1 for dimer) PMID:23055524 we conclude the stoichiometry of the complex is 2:2. |
| GO:0060100 positive regulation of phagocytosis, engulfment | IEA GO_REF:0000107 | ACCEPT | Summary: APPL2 depletion reduces Fc-gamma-receptor-mediated phagocytosis and delays phagocytic cup closure. Reason: This is one of the better-evidenced APPL2-specific functions: siRNA against APPL2 alone reduced internalisation of IgG-coated particles, delayed the PI(3,4,5)P3 transition and delayed phagocytic cup closure, with APPL2 physically on the closing cup and bound to GTP-Rab31 there. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Accepted as a core function - promoting phagosome closure is a direct membrane-remodelling role for a BAR-domain Rab effector, not a downstream physiological effect. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse evidence is an APPL2-only siRNA depletion with a quantified phagocytosis defect and a live-imaging cup-closure delay (PMID:25568335). ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. Supporting Evidence: PMID:25568335 siRNA depletion of either Rab31 or APPL2 reduces FcγR-mediated phagocytosis. PMID:25568335 Mechanistically, this corresponds with a delay in the transition to PI(3,4,5)P3 and phagocytic cup closure. PMID:25568335 During early phagocytosis, we find that Rab31 recruits the signaling adaptor APPL2. |
| GO:0060100 positive regulation of phagocytosis, engulfment | ISS GO_REF:0000024 | ACCEPT | Summary: APPL2 depletion reduces Fc-gamma-receptor-mediated phagocytosis and delays phagocytic cup closure. Reason: This is one of the better-evidenced APPL2-specific functions: siRNA against APPL2 alone reduced internalisation of IgG-coated particles, delayed the PI(3,4,5)P3 transition and delayed phagocytic cup closure, with APPL2 physically on the closing cup and bound to GTP-Rab31 there. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Accepted as a core function - promoting phagosome closure is a direct membrane-remodelling role for a BAR-domain Rab effector, not a downstream physiological effect. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse evidence is an APPL2-only siRNA depletion with a quantified phagocytosis defect and a live-imaging cup-closure delay (PMID:25568335). Supporting Evidence: PMID:25568335 siRNA depletion of either Rab31 or APPL2 reduces FcγR-mediated phagocytosis. PMID:25568335 Mechanistically, this corresponds with a delay in the transition to PI(3,4,5)P3 and phagocytic cup closure. PMID:25568335 During early phagocytosis, we find that Rab31 recruits the signaling adaptor APPL2. |
| GO:0070062 extracellular exosome | HDA PMID:19056867 Large-scale proteomics and phosphoproteomics of urinary exos... | KEEP AS NON CORE | Summary: Detected in an extracellular-vesicle proteome. The protein is plausible EV content, but this is not where it functions. Reason: Peripheral early-endosome adaptor. Cytosolic protein with lipid or membrane-protein binding ability; MISEV2018 (PMID:30637094) category 2a lists such proteins (ESCRT components, ALIX, annexins, HSPA8 and others) as expected EV contents. True, but not its core location. (Manual re-review for OMICS_EVIDENCE, 2026-10-10; previous action: MARK_AS_OVER_ANNOTATED.) Supporting Evidence: PMID:19056867 we used LC-MS/MS to profile the proteome of human urinary exosomes PMID:19056867 the analysis identified 1132 proteins unambiguously |
| GO:0098588 bounding membrane of organelle | IEA GO_REF:0000117 | MARK AS OVER ANNOTATED | Summary: Very broad ARBA rule assigning a term with no discriminating content. Reason: ARBA00027281 has 171 condition sets spanning bacteria, plants, fungi and metazoa. 'Bounding membrane of organelle' is a structural-class term that applies to any peripheral membrane protein on any organelle; combined with a rule that broad it conveys nothing about APPL2 that the endosome-membrane and phagosome-membrane rows do not say better. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: GRANULARITY MISMATCH Sources checked: ARBA:ARBA00027281 · ARBA rule ARBA00027281 SUPPORTS TRANSFER ARBA rule for bounding membrane of organelle; 171 condition sets spanning bacteria to plants. Too broad and too generic to say anything about APPL2. Supporting Evidence: PMID:21645192 APPL endosomes are a recently identified subpopulation of early endosomes characterized by the presence of two homologous Rab5 effector proteins APPL1 and APPL2. |
| GO:0120162 positive regulation of cold-induced thermogenesis | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: positive regulation of cold-induced thermogenesis rests on a neuron-restricted Appl2 knockout in mouse hypothalamus. Reason: Deleting APPL2 in ventromedial-hypothalamic RIP-Cre neurons reduced beiging of subcutaneous white adipose tissue and produced cold intolerance and obesity, acting through AMPK and sympathetic outflow (PMID:29467283). The perturbation is APPL2-specific and conditional. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. This is nonetheless an organism-level physiological outcome of APPL2 acting in one neuronal population, several steps removed from what the protein does molecularly, so it is peripheral rather than core. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse evidence is a conditional, neuron-restricted APPL2 knockout with a whole-animal thermogenic phenotype (PMID:29467283) - APPL2-specific, but organismal. ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. Supporting Evidence: PMID:29467283 Genetic ablation of APPL2 in RIP-Cre neurons diminishes beiging in sWAT without affecting BAT, leading to cold intolerance and obesity in mice. PMID:29467283 Hypothalamic APPL2 enhances neuronal activation in VMH RIP-Cre neurons and raphe pallidus, thereby eliciting SNS outflow to sWAT and subsequent beiging. |
| GO:0120162 positive regulation of cold-induced thermogenesis | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: positive regulation of cold-induced thermogenesis rests on a neuron-restricted Appl2 knockout in mouse hypothalamus. Reason: Deleting APPL2 in ventromedial-hypothalamic RIP-Cre neurons reduced beiging of subcutaneous white adipose tissue and produced cold intolerance and obesity, acting through AMPK and sympathetic outflow (PMID:29467283). The perturbation is APPL2-specific and conditional. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. This is nonetheless an organism-level physiological outcome of APPL2 acting in one neuronal population, several steps removed from what the protein does molecularly, so it is peripheral rather than core. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The mouse evidence is a conditional, neuron-restricted APPL2 knockout with a whole-animal thermogenic phenotype (PMID:29467283) - APPL2-specific, but organismal. Supporting Evidence: PMID:29467283 Genetic ablation of APPL2 in RIP-Cre neurons diminishes beiging in sWAT without affecting BAT, leading to cold intolerance and obesity in mice. PMID:29467283 Hypothalamic APPL2 enhances neuronal activation in VMH RIP-Cre neurons and raphe pallidus, thereby eliciting SNS outflow to sWAT and subsequent beiging. |
| GO:1900016 negative regulation of cytokine production involved in inflammatory response | IEA GO_REF:0000107 | ACCEPT | Summary: APPL2 constrains cytokine secretion in the immune response; the negative sign is the one the data support. Reason: Two independent mouse studies agree on the direction: Appl2 deletion raised TNF-alpha and IL-1beta from primary macrophages, and APPL2 depletion in macrophages showed it 'serves to constrain the secretion of pro- and anti-inflammatory cytokines'. The signed term is therefore correctly signed, which is the thing to check on a transfer like this. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. Two independent APPL2-specific mouse perturbations (PMID:25328665 knockout, PMID:27219021 depletion) agree on the negative direction, so the signed term transfers safely. ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. Supporting Evidence: PMID:25328665 deletion of Appl2 led to higher levels of TNF-α and IL-1β in primary macrophages PMID:27219021 APPL2 has a dominant role in nuclear translocation of NF-KB p65 and it serves to constrain the secretion of pro- and anti-inflammatory cytokines. PMID:25328665 When challenged with lipopolysaccharides (LPS), Appl2 KO mice exhibited more severe symptoms of endotoxin shock, accompanied by increased production of proinflammatory cytokines. |
| GO:1900016 negative regulation of cytokine production involved in inflammatory response | ISS GO_REF:0000024 | ACCEPT | Summary: APPL2 constrains cytokine secretion in the immune response; the negative sign is the one the data support. Reason: Two independent mouse studies agree on the direction: Appl2 deletion raised TNF-alpha and IL-1beta from primary macrophages, and APPL2 depletion in macrophages showed it 'serves to constrain the secretion of pro- and anti-inflammatory cytokines'. The signed term is therefore correctly signed, which is the thing to check on a transfer like this. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. Two independent APPL2-specific mouse perturbations (PMID:25328665 knockout, PMID:27219021 depletion) agree on the negative direction, so the signed term transfers safely. Supporting Evidence: PMID:25328665 deletion of Appl2 led to higher levels of TNF-α and IL-1β in primary macrophages PMID:27219021 APPL2 has a dominant role in nuclear translocation of NF-KB p65 and it serves to constrain the secretion of pro- and anti-inflammatory cytokines. PMID:25328665 When challenged with lipopolysaccharides (LPS), Appl2 KO mice exhibited more severe symptoms of endotoxin shock, accompanied by increased production of proinflammatory cytokines. |
| GO:1900077 negative regulation of cellular response to insulin stimulus | IEA GO_REF:0000107 | ACCEPT | Summary: APPL2 negatively regulates the cellular insulin response, by sequestering APPL1 and by acting on TBC1D1. Reason: Two independent mechanisms converge on the same sign: APPL2 sequesters APPL1 away from the insulin pathway in muscle cells, and it binds phospho-TBC1D1 to block GLUT4 translocation. The signed term matches both. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Core. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. Two independent APPL2-specific mouse mechanisms (PMID:19661063 sequestration, PMID:24879834 TBC1D1) agree on the negative sign. ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. Supporting Evidence: PMID:19661063 APPL2 also suppresses adiponectin and insulin signaling by sequestrating APPL1 from these two pathways. PMID:24879834 insulin-evoked plasma membrane recruitment of GLUT4 and glucose uptake are impaired by APPL2 overexpression but enhanced by APPL2 knockdown. PMID:19661063 APPL2, an isoform of APPL1 that forms a dimer with APPL1, can interacts with both AdipoR1 and AdipoR2 and acts as a negative regulator of adiponectin signaling in muscle cells. |
| GO:1900077 negative regulation of cellular response to insulin stimulus | ISS GO_REF:0000024 | ACCEPT | Summary: APPL2 negatively regulates the cellular insulin response, by sequestering APPL1 and by acting on TBC1D1. Reason: Two independent mechanisms converge on the same sign: APPL2 sequesters APPL1 away from the insulin pathway in muscle cells, and it binds phospho-TBC1D1 to block GLUT4 translocation. The signed term matches both. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Core. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. Two independent APPL2-specific mouse mechanisms (PMID:19661063 sequestration, PMID:24879834 TBC1D1) agree on the negative sign. Supporting Evidence: PMID:19661063 APPL2 also suppresses adiponectin and insulin signaling by sequestrating APPL1 from these two pathways. PMID:24879834 insulin-evoked plasma membrane recruitment of GLUT4 and glucose uptake are impaired by APPL2 overexpression but enhanced by APPL2 knockdown. PMID:19661063 APPL2, an isoform of APPL1 that forms a dimer with APPL1, can interacts with both AdipoR1 and AdipoR2 and acts as a negative regulator of adiponectin signaling in muscle cells. |
| GO:1905303 positive regulation of macropinocytosis | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: Positive regulation of macropinocytosis, from mouse macrophage work. Reason: APPL2 is recruited to surface ruffles and then to macropinosomes in LPS-activated macrophages, and the adaptors were depleted individually in that study, so this is an APPL2-specific observation rather than a paralog-pooled one. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. It is non-core because it is restricted to activated macrophages and because the closely related phagocytosis rows are better supported - there the depletion phenotype was measured directly. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. Mouse macrophage evidence (PMID:27219021) in which each adaptor was depleted separately. ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. Supporting Evidence: PMID:27219021 APPL1 and 2 are differentially localized to distinct signaling-competent membrane domains on the surface and in endocytic compartments of LPS-activated macrophages. PMID:25568335 APPL2 was immunolabeled in the cytoplasm, on cell surface ruffles, and on early, actin-rich phagosomes |
| GO:1905303 positive regulation of macropinocytosis | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: Positive regulation of macropinocytosis, from mouse macrophage work. Reason: APPL2 is recruited to surface ruffles and then to macropinosomes in LPS-activated macrophages, and the adaptors were depleted individually in that study, so this is an APPL2-specific observation rather than a paralog-pooled one. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. It is non-core because it is restricted to activated macrophages and because the closely related phagocytosis rows are better supported - there the depletion phenotype was measured directly. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. Mouse macrophage evidence (PMID:27219021) in which each adaptor was depleted separately. Supporting Evidence: PMID:27219021 APPL1 and 2 are differentially localized to distinct signaling-competent membrane domains on the surface and in endocytic compartments of LPS-activated macrophages. PMID:25568335 APPL2 was immunolabeled in the cytoplasm, on cell surface ruffles, and on early, actin-rich phagosomes |
| GO:1905451 positive regulation of Fc-gamma receptor signaling pathway involved in phagocytosis | IEA GO_REF:0000107 | ACCEPT | Summary: APPL2 promotes PI3K/Akt signalling downstream of the Fc-gamma receptor during phagocytosis. Reason: APPL2 depletion reduced PI3K/Akt signalling and enhanced p38 signalling from the Fc-gamma receptor, and reduced recruitment of Akt to phagosomal membranes. The sign is positive for the Akt arm, which is what this term asserts. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Accepted as core, with one caveat recorded in the notes: in the same cell type but downstream of TLR4 rather than FcgammaR, APPL2 restrains Akt signalling - the sign of APPL2's effect on this kinase is receptor-dependent, and no single annotation can express that. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. Mouse macrophage evidence from an APPL2-only depletion (PMID:25568335); the opposite-sign TLR4 result (PMID:25328665) is a different receptor, not a contradiction of this term. ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. Supporting Evidence: PMID:25568335 APPL2 depletion also reduced PI3K/Akt signaling and enhanced p38 signaling from FcγR. PMID:25568335 siRNA depletion of either Rab31 or APPL2 reduces FcγR-mediated phagocytosis. PMID:25328665 phosphorylation of Akt and its downstream effector NF-κB was significantly enhanced. |
| GO:1905451 positive regulation of Fc-gamma receptor signaling pathway involved in phagocytosis | ISS GO_REF:0000024 | ACCEPT | Summary: APPL2 promotes PI3K/Akt signalling downstream of the Fc-gamma receptor during phagocytosis. Reason: APPL2 depletion reduced PI3K/Akt signalling and enhanced p38 signalling from the Fc-gamma receptor, and reduced recruitment of Akt to phagosomal membranes. The sign is positive for the Akt arm, which is what this term asserts. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Accepted as core, with one caveat recorded in the notes: in the same cell type but downstream of TLR4 rather than FcgammaR, APPL2 restrains Akt signalling - the sign of APPL2's effect on this kinase is receptor-dependent, and no single annotation can express that. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. Mouse macrophage evidence from an APPL2-only depletion (PMID:25568335); the opposite-sign TLR4 result (PMID:25328665) is a different receptor, not a contradiction of this term. Supporting Evidence: PMID:25568335 APPL2 depletion also reduced PI3K/Akt signaling and enhanced p38 signaling from FcγR. PMID:25568335 siRNA depletion of either Rab31 or APPL2 reduces FcγR-mediated phagocytosis. PMID:25328665 phosphorylation of Akt and its downstream effector NF-κB was significantly enhanced. |
| GO:2000045 regulation of G1/S transition of mitotic cell cycle | IDA PMID:15016378 APPL proteins link Rab5 to nuclear signal transduction via a... | KEEP AS NON CORE | Summary: APPL2 is required for cell proliferation in a Rab5-binding-dependent manner. Reason: Both APPL1 and APPL2 are essential for cell proliferation and the requirement depends on Rab5 binding, which is the observation behind the cell-cycle term; the pathway is proposed to run from the APPL endosome to the nucleus and the NuRD/MeCP1 complex. It is kept as non-core because the specific G1/S claim is a downstream consequence of the signalling adaptor role rather than something APPL2 does directly, and because no APPL2-specific cell-cycle mechanism has been mapped since. Supporting Evidence: PMID:15016378 Both APPL1 and APPL2 are essential for cell proliferation and their function requires Rab5 binding. PMID:15016378 Our findings identify an endosomal compartment bearing Rab5 and APPL proteins as an intermediate in signaling between the plasma membrane and the nucleus. PMID:15016378 APPL1 translocates from the membranes to the nucleus where it interacts with the nucleosome remodeling and histone deacetylase multiprotein complex NuRD/MeCP1 |
| GO:2000045 regulation of G1/S transition of mitotic cell cycle | IEA GO_REF:0000117 | KEEP AS NON CORE | Summary: Narrow ARBA rule reaching the same cell-cycle term as the IDA row. Reason: ARBA00033889 has 7 condition sets over primates, rodents and yeast. It restates the IDA row's claim; the yeast condition set is irrelevant to APPL2 but does not make the human call wrong. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: ARBA:ARBA00033889 · ARBA rule ARBA00033889 SUPPORTS TRANSFER ARBA rule for regulation of the G1/S transition; 7 condition sets over primates, rodents and yeast. Narrow, and it reproduces the IDA row's claim. Supporting Evidence: PMID:15016378 Both APPL1 and APPL2 are essential for cell proliferation and their function requires Rab5 binding. |
| GO:2000178 negative regulation of neural precursor cell proliferation | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: negative regulation of neural precursor cell proliferation, driven mainly by APPL2 transgenic overexpression in mouse. Reason: The in vivo evidence is an APPL2 transgenic mouse: APPL2 Tg animals have reduced hippocampal and olfactory neurogenesis and a higher glial fraction, reversible by a glucocorticoid-receptor antagonist. A matching loss-of-function exists only in cultured neural stem cells, where APPL2 knockdown promoted neurogenesis. So both directions exist and the negative sign is right, but the animal-level claim rests on gain of function, where an overexpressed adaptor can titrate partners non-physiologically. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Non-core: this is a tissue-specific developmental consequence of APPL2 tuning glucocorticoid-receptor sensitivity, several steps from the molecular function. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The donor evidence is chiefly an APPL2 transgenic overexpression mouse (PMID:28965332, PMID:32468397), with the loss-of-function direction shown only in vitro. ensembl:ENSMUSP00000020500 · Appl2 (Mus musculus), Ensembl protein SUPPORTS TRANSFER Ensembl protein accession for the same mouse Appl2 entry (it resolves to UniProtKB:Q8K3G9), so this row carries one donor under two identifiers, not two independent donors. Supporting Evidence: PMID:28965332 APPL2 Tg mice had decreased hippocampal neurogenesis that was reversed by GR antagonist RU486. PMID:32468397 APPL2 overexpression resulted in NSCs switching from neuronal differentiation to gliogenesis while APPL2 knockdown promoted neurogenesis. PMID:32468397 APPL2 Tg mice had a higher population of glial cells and dampened neuronal production in the olfactory system |
| GO:2000178 negative regulation of neural precursor cell proliferation | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: negative regulation of neural precursor cell proliferation, driven mainly by APPL2 transgenic overexpression in mouse. Reason: The in vivo evidence is an APPL2 transgenic mouse: APPL2 Tg animals have reduced hippocampal and olfactory neurogenesis and a higher glial fraction, reversible by a glucocorticoid-receptor antagonist. A matching loss-of-function exists only in cultured neural stem cells, where APPL2 knockdown promoted neurogenesis. So both directions exist and the negative sign is right, but the animal-level claim rests on gain of function, where an overexpressed adaptor can titrate partners non-physiologically. The donor is mouse Appl2 (Q8K3G9), a 1:1 ortholog 92.7% identical to the human protein, so the orthology transfer carries no ortholog-assignment risk. Non-core: this is a tissue-specific developmental consequence of APPL2 tuning glucocorticoid-receptor sensitivity, several steps from the molecular function. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target and the only gene-level donor on this row; human and mouse APPL2 are 92.7% identical over a full-length global alignment (APPL2-bioinformatics/RESULTS.md), so the transfer itself is safe. The donor evidence is chiefly an APPL2 transgenic overexpression mouse (PMID:28965332, PMID:32468397), with the loss-of-function direction shown only in vitro. Supporting Evidence: PMID:28965332 APPL2 Tg mice had decreased hippocampal neurogenesis that was reversed by GR antagonist RU486. PMID:32468397 APPL2 overexpression resulted in NSCs switching from neuronal differentiation to gliogenesis while APPL2 knockdown promoted neurogenesis. PMID:32468397 APPL2 Tg mice had a higher population of glial cells and dampened neuronal production in the olfactory system |
| GO:0090263 positive regulation of canonical Wnt signaling pathway | IMP PMID:19433865 Endosomal adaptor proteins APPL1 and APPL2 are novel activat... | NEW | Summary: APPL2 relieves Reptin-mediated repression of beta-catenin/TCF-dependent transcription in human cells, a function with no corresponding GOA row. Reason: GOA carries the APPL2-RUVBL2/Reptin and APPL2-CTNNB1/HDAC1/HDAC2 interactions as bare protein-binding rows but no biological-process row for what those interactions do. In human HEK293 cells, APPL2 overexpression increased beta-catenin-stimulated SuperTOPflash activity dose-dependently with no effect on the mutated-site control reporter, increased Wnt3a-induced expression of endogenous Wnt target genes, relieved Reptin-mediated repression, reduced HDAC1/HDAC2 and beta-catenin bound to Reptin, and increased beta-catenin occupancy at the cyclin D1 and Axin2 promoters while reducing Reptin and HDAC1 there. UniProt annotates the same role on Q8NEU8 with experimental evidence from this paper. Coded IMP rather than IDA because the perturbation is overexpression; the caveat that this is a gain-of-function assay is genuine - APPL2 is practically undetectable in HEK293 cells, so no knockdown control was possible, and APPL2's effect was consistently smaller than APPL1's. Supporting Evidence: PMID:19433865 overexpression of either APPL1 or APPL2 increased the levels of β-catenin-stimulated reporter activity in a dose-dependent manner PMID:19433865 in cells treated with Wnt3a-conditioned medium, overexpression of either APPL1 or APPL2 proteins caused higher expression of the Wnt target genes PMID:19433865 this repression was relieved in a dose-dependent manner by adding increasing amounts of APPL1 or APPL2 PMID:19433865 overexpression of either APPL protein clearly reduced the levels of HDAC1 and HDAC2 associated with Reptin PMID:19433865 Since APPL2 is practically undetectable in HEK293 cells, we could not test the effects of its knockdown on transcription. |
| GO:0140311 protein sequestering activity | ISS PMID:19661063 Yin-Yang regulation of adiponectin signaling by APPL isoform... | NEW | Summary: APPL2 sequesters APPL1 away from the adiponectin and insulin pathways - the mechanism behind its antagonism, and a molecular function GOA does not record. Reason: Every functional row on this gene describes what APPL2 does to a pathway; none names the molecular activity by which it does it. The mechanism is explicit in the source: APPL2 binds AdipoR1 and AdipoR2 through its BAR domain, competes with APPL1 for the receptors, and additionally suppresses both adiponectin and insulin signalling by sequestrating APPL1 from those pathways; adiponectin and metformin both trigger APPL1-APPL2 dissociation, so the sequestration is regulated rather than constitutive. GO:0140311 is defined as binding a protein to prevent it interacting with other partners or to inhibit its localisation to where it is active, which is exactly this. Coded ISS with mouse Appl2 (Q8K3G9) as the donor because the experiments are in C2C12 mouse myotubes; the human protein is 92.7% identical and the APPL1-APPL2 heterodimer that the activity depends on is established directly for the human proteins. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q8K3G9 · Appl2 (Mus musculus) SUPPORTS TRANSFER Mouse Appl2, the 1:1 ortholog of the target (92.7% identity over a full-length global alignment, APPL2-bioinformatics/RESULTS.md). The sequestration was shown in mouse C2C12 myotubes by overexpression and RNAi, and the heterodimer it depends on is independently established for the human proteins by yeast two-hybrid and reciprocal coimmunoprecipitation, so the transfer is safe. Supporting Evidence: PMID:19661063 APPL2 also suppresses adiponectin and insulin signaling by sequestrating APPL1 from these two pathways. PMID:19661063 APPL2, an isoform of APPL1 that forms a dimer with APPL1, can interacts with both AdipoR1 and AdipoR2 and acts as a negative regulator of adiponectin signaling in muscle cells. PMID:19661063 Overexpression of APPL2 inhibits the interaction between APPL1 and AdipoR1, leading to down-regulation of adiponectin signaling in C2C12 myotubes. file:human/APPL2/APPL2-bioinformatics/RESULTS.md Human and mouse APPL2 are 92.7% identical over a full-length global alignment |
Loading supporting content…
Download this section (compressed HTML)Q: Is the receptor-dependent sign of APPL2's effect on Akt - restraining after TLR4, promoting after Fc-gamma receptor engagement - set by which Rab recruits it (Rab31 on phagosomes versus the macropinosomal pool), or by which PI3K is available at each membrane domain?
Q: Is sequestration the molecular activity of APPL2 rather than a side effect of scaffolding? Everything mapped so far is competitive occupancy - of AdipoR1/R2 against APPL1, of phospho-Ser235 TBC1D1, of Reptin.
Q: Given that APPL2 and APPL1 heterodimerise through their BAR domains yet oppose each other functionally, is the APPL1-APPL2 heterodimer an inactive reservoir, or does it have activities neither homodimer has?
Q: GO cannot currently express negative regulation of the adiponectin-activated signaling pathway. Should such a term be created, and should APPL1 and APPL2 then be re-annotated to the corresponding positive and negative children?
Q: Should GO represent the APPL endosome as a distinct compartment, given that it is separable from EEA1-positive early endosomes biochemically, by density-gradient migration, by peripheral distribution and by its dependence on annexin A2?
Q: Does the homotetramer recorded for APPL2 in UniProt and GOA have any experimental basis beyond two dimers in a crystallographic asymmetric unit?
Experiment: Rescue APPL1/APPL2 double-knockdown cells with APPL2 alone, APPL1 alone, or a BAR-interface mutant that cannot heterodimerise, and score TbetaRI-ICD nuclear accumulation and TGF-beta-induced invasion. This is the control the existing literature lacks and is what would convert the GO:0007179 and GO:0006606 rows from a paralog-pooled result into an APPL2-specific one.
Hypothesis: APPL2 contributes to TbetaRI-ICD nuclear transport independently of APPL1, rather than only as the redundant half of a pair.
Type: genetic rescue / single-adaptor complementation
Experiment: Mutate 151-157 to alanine and ask whether the nuclear pool of APPL2 is lost and whether the Reptin/HDAC association and the beta-catenin/TCF reporter response go with it; in parallel test whether a PH-domain-deletion construct, which should unmask the motif, shows increased nuclear accumulation. The motif is APPL2-specific - the aligned span in APPL1 is SKKREND - so a positive result would also explain a paralog difference.
Hypothesis: The basic cluster at residues 151-157 (PKKKENE) is a functional nuclear localisation signal that the PH domain masks.
Type: site-directed mutagenesis with imaging and biochemical readouts
Experiment: Measure the oligomeric state of full-length recombinant APPL2 by mass photometry and SEC-MALS, and compare with the BAR-PH fragment; attempt a structure or prediction of the full-length protein to determine whether the PTB domain supports higher-order assembly.
Hypothesis: Full-length APPL2 is a dimer in solution, and the homotetramer in the current annotations is a crystal-packing artefact.
Type: biophysical characterisation
Experiment: Compare Akt phosphorylation, PI(3,4,5)P3 kinetics and PI3K-subunit recruitment side by side in APPL2-depleted macrophages stimulated through TLR4 versus the Fc-gamma receptor, with Rab31 and Rab5 co-depletion arms, using live probes for the phosphoinositide transitions.
Hypothesis: The sign of APPL2's effect on Akt tracks the Rab that recruits it to the membrane.
Type: comparative perturbation with live-cell imaging
Experiment: Express in Appl2-null myotubes an APPL2 mutant that binds AdipoR1 but not APPL1 and one that binds APPL1 but not AdipoR1, and measure adiponectin-stimulated glucose uptake and fatty acid oxidation for each. The result would also determine whether GO:0140311 protein sequestering activity is the right molecular function or only part of it.
Hypothesis: APPL2 antagonises adiponectin signalling by two separable mechanisms - receptor competition and APPL1 sequestration - either of which is sufficient.
Type: separation-of-function mutants in a null background
What is not known — curated, literature-grounded statements of the open unknowns (the inverse of core functions).
Gap: It is not known what determines the sign of APPL2's effect on PI3K/Akt signalling: in macrophages APPL2 restrains Akt downstream of TLR4 but promotes it downstream of the Fc-gamma receptor.
OPEN BIOLOGY
What is known: Known: Appl2 knockout raises Akt and NF-kappaB phosphorylation after LPS, while APPL2 depletion reduces Akt recruitment and phosphorylation during FcgammaR-mediated phagocytosis, both in macrophages. Unknown: whether the reversal is set by the recruiting Rab, the membrane domain, or a different PI3K at each site.
Significance: Two GO annotations on this gene assert opposite directions of Akt-pathway regulation, and both are correct in context. No single annotation can express a receptor-dependent sign reversal, so the ontology forces an apparent contradiction onto a real biological phenomenon.
Provenance (the field's own admissions):
Gap: Whether APPL2 recruits or regulates a PI3K directly has never been tested, even though the proposed mechanism for its role in phagocytosis depends on it.
OPEN BIOLOGY
What is known: Known: APPL2 depletion delays the PI(4,5)P2-to-PI(3,4,5)P3 transition at the phagocytic cup, and mouse Appl2 is found in a complex with the PI3K regulatory subunit p85alpha together with Appl1. Unknown: whether APPL2 binds any PI3K subunit directly, and whether that binding is what produces the phosphoinositide phenotype.
Significance: It is the difference between APPL2 being a lipid-kinase-recruiting adaptor and being a passive curvature sensor whose loss perturbs the membrane indirectly.
Provenance (the field's own admissions):
Gap: How the nuclear pool of APPL2 is generated is unknown: the proposed nuclear localisation signal at residues 151-157 has never been mutated.
OPEN BIOLOGY
What is known: Known: the motif PKKKENE is present verbatim at 151-157 of Q8NEU8, lies inside the BAR domain, is conserved in mouse Appl2 and diverges in APPL1 (SKKREND), and the crystal structure places it on a face masked by the PH domain. A nuclear APPL2 pool is well documented. Unknown: whether this motif is the signal, and whether PH-domain rotation gates it.
Significance: The nucleus rows and the NuRD and Reptin complex-binding rows all rest on a nuclear pool whose import mechanism is unexplained. It is also one of the very few sequence features that distinguishes APPL2 from APPL1.
Provenance (the field's own admissions):
Gap: APPL2's individual contribution to TGF-beta-driven TbetaRI-ICD nuclear transport is unresolved because only simultaneous APPL1 + APPL2 knockdowns were performed.
OPEN CURATION
What is known: Known: depleting both adaptors abolishes nuclear TbetaRI-ICD accumulation and reduces TGF-beta-induced invasion and MMP2/MMP9 transcription. Unknown: whether APPL2 alone is required; the TbetaRI and PKCzeta associations were mapped to APPL1 only.
Significance: Two GOA rows (GO:0007179 IMP and GO:0006606 IDA) rest entirely on this design. At ~54% paralog identity and with documented opposite-sign behaviour elsewhere, a double knockdown is not evidence about either gene singly.
Provenance (the field's own admissions):
Gap: The assembly state of full-length APPL2 is unestablished; all structural work is on the BAR-PH fragment and the PTB domain's contribution is unknown.
OPEN BIOLOGY
What is known: Known: the BAR-PH module (residues 1-384, PDB 4H8S and 5C5B) forms a crescent-shaped dimer, and cross-linking and light scattering both give a dimer in solution. Unknown: whether the full-length protein behaves the same way; no full-length structure and no APPL2-Rab complex structure exists.
Significance: A GOA row asserts protein homotetramerization on the strength of two dimers in a crystallographic asymmetric unit. Settling the solution assembly state of the intact protein would close that question directly.
Provenance (the field's own admissions):
Gap: Most of what is annotated for human APPL2 - adiponectin antagonism, endotoxin hypersensitivity, phagocytosis, adipose beiging, neurogenesis - has only been demonstrated in mouse.
OPEN CURATION
What is known: Known: human and mouse APPL2 are 92.7% identical over a full-length alignment, so the orthology transfer is as safe as such transfers get. Unknown: whether the human protein reproduces these phenotypes; the only human-cell functional data are the glucose-uptake, Wnt-reporter and TGF-beta experiments.
Significance: Forty-three of the 124 rows on this gene (24 ISS plus 19 orthology-based IEA) come from the single mouse donor Q8K3G9.
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)