ARFGEF2 encodes BIG2 (brefeldin A-inhibited guanine nucleotide-exchange protein 2), a ~202 kDa guanine-nucleotide exchange factor for ADP-ribosylation factor GTPases. Its central Sec7 domain catalyses GDP-to-GTP exchange on class I ARFs, principally ARF1 and ARF3, and this activity is blocked by the fungal metabolite brefeldin A. BIG2 works at two membrane sites: at the trans-Golgi network, where ARF activation recruits the clathrin adaptors AP-1 and GGA - but not the COPI coat - to drive traffic towards endosomes and the cell surface; and on recycling endosomes, where it is required to maintain compartment integrity and to return internalised cargo such as the transferrin receptor and integrin beta1 to the plasma membrane. It is recruited from a large cytosolic pool onto membranes, and its exchange activity is damped by PKA phosphorylation and restored by protein phosphatase 1. Beyond catalysis, BIG2 is an A-kinase anchoring protein: three short amphipathic regions in its N-terminal half bind PKA regulatory subunits, assembling a local cAMP module that also contains phosphodiesterase PDE3A, and a separate non-catalytic C-terminal region binds non-muscle myosin IIA and restrains myosin light-chain phosphorylation, stress fibre formation and cell migration. BIG2 forms homodimers through its DCB and HUS regions and is found in large complexes with its close paralog BIG1, with which it shares trans-Golgi functions while retaining distinct roles - BIG1 maintains Golgi morphology, BIG2 maintains endosomal integrity. It is expressed in placenta, lung, heart, brain, kidney and pancreas, with the Human Protein Atlas reporting tissue-enhanced expression in parathyroid, and in neurons it is concentrated in the trans-Golgi network and in dendritic vesicles, where it contributes to GABA(A) receptor export from the endoplasmic reticulum. Biallelic loss-of-function mutations in ARFGEF2 cause autosomal recessive periventricular nodular heterotopia with microcephaly (PVNH2), in which cortical neurons fail to migrate away from the periventricular proliferative zone.
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0000139 Golgi membrane | IDA PMID:10716990 Identification and localization of two brefeldin A-inhibited... | ACCEPT | Summary: BIG2 is recovered with Golgi membrane markers by density-gradient fractionation and colocalises with Golgi 58K protein and gamma-adaptin by immunofluorescence in human HeLa S3 and HepG2 cells. Golgi membrane is one of BIG2's two core sites of action. Reason: Direct fractionation and imaging evidence in human cell lines. The co-immunoprecipitation half of this paper used bovine brain cytosol, but the localisation experiments cited for this term are the human-cell immunofluorescence and the microsomal gradient, so the human IDA is sound. Supporting Evidence: PMID:10716990 After density gradient centrifugation of a microsomal fraction, BIG1 and BIG2 were recovered in the same fraction as beta-COP, a marker for Golgi membranes. PMID:10716990 endogenous BIG1 and coexpressed BIG2 were distributed in a punctate pattern throughout the cytosol, and also concentrated in the perinuclear region, where endogenous BIG1 and BIG2 each partially colocalized with Golgi-specific 58K protein and gamma-adaptin |
| GO:0000139 Golgi membrane | IDA PMID:12571360 Protein kinase A-anchoring (AKAP) domains in brefeldin A-inh... | ACCEPT | Summary: Independent subcellular-fractionation evidence for BIG2 at Golgi membranes, here as the cAMP-regulated endpoint: raising cAMP moves BIG2 from cytosol onto Golgi and other membranes. Reason: A second, mechanistically different route to the same location (regulated translocation rather than steady-state distribution). Duplicate GO ids from different references are expected and useful here. Supporting Evidence: PMID:12571360 Western blot analysis of subcellular fractions demonstrated translocation of BIG2 (and BIG1) from cytosol to the Golgi and other membrane structures after incubation of cells with 8-Br-cAMP or forskolin. |
| GO:0001881 receptor recycling | IDA PMID:16477018 Association of brefeldin A-inhibited guanine nucleotide-exch... | ACCEPT | Summary: BIG2, but not BIG1, accumulates with recycling-endosome markers during transferrin uptake, and BIG2 knockdown causes perinuclear accumulation of the transferrin receptor and slows transferrin release. Receptor recycling is a core BIG2 process. Reason: The functional core of the paper and one of the best-supported BIG2 processes, with a clean BIG1 control that fails to reproduce the phenotype. Two evidence-code notes for the curator, neither of which changes the term: the decisive experiments are siRNA knockdowns, which is IMP rather than IDA; and the cells are COS7, with siRNAs designed against the human BIG1 and BIG2 sequences, so the reagents rather than the cell line carry the human specificity. Supporting Evidence: PMID:16477018 Treatment of cells with BIG2 small interfering RNA (siRNA), but not BIG1 or control siRNAs, decreased BIG2 protein >90% without affecting BIG1, ARF, or actin content, whereas TfnR was significantly increased as was its accumulation in perinuclear recycling endosomes. PMID:16477018 Tfn release appeared unaffected by BIG1 siRNA but was significantly slowed from cells treated with BIG2 siRNA alone or plus BIG1 siRNA. PMID:16477018 COS7 cells (used because they are more favorable than HepG2 cells for immunofluorescence microscopy) |
| GO:0005085 guanyl-nucleotide exchange factor activity | IBA GO_REF:0000033 | ACCEPT | Summary: Pan-eukaryotic PAINT inference that BIG2 is a guanine-nucleotide exchange factor, placed at a node spanning the whole Sec7 superfamily. Correct and core. Reason: The donor set resolves to GNOM (Arabidopsis), Sec71 and garz (Drosophila), cytohesin-1 and PSD3 (mouse), GEA1/GEA2/SEC7/MON2/SYT1 (budding yeast), an S. pombe PSD-like protein, and human PSD, CYTH1, CYTH3, CYTH4, IQSEC2 and ARFGEF1. That clade mixes BIG-type, GBF-type, cytohesin-type and MON2-type GEFs acting on different ARF and ARF-like substrates, so the general term is the least common ancestor of the donors, not a curator failing to be specific; GRANULARITY_MISMATCH does not apply. ARFGEF2 itself (UniProtKB:Q9Y6D5) appears among the sources, which is the expected marker that the target carries its own experimental grounding for the term and the function is inherited, not a circularity. Separately, there is no more specific term available to propose: GO:0005086 ARF guanyl-nucleotide exchange factor activity and eight sibling substrate-specific GEF terms were merged into GO:0005085 and survive only as its secondaryIds, so GO:0005085 is already maximal and the ARF1/ARF3 preference has to be recorded machine-readably instead. Propagation Review Root cause: NO FAILURE CORE Sources checked: PANTHER:PTN008950430 · Sec7-family ancestral node SUPPORTS TRANSFER PAINT ancestral node; not a protein. Reach spans BIG, GBF, cytohesin, PSD, IQSEC and MON2 clades. UniProtKB:Q9Y6D5 · ARFGEF2 (self) SUPPORTS TRANSFER Self-referential: the target's own IDA is one of the descendant evidences behind the IBD. Valid, not circular. UniProtKB:Q9Y6D6 · ARFGEF1 SUPPORTS TRANSFER Human paralog; holds GO:0005085 by IDA. AGI_LocusCode:AT1G13980 · GNOM SUPPORTS TRANSFER Arabidopsis large ArfGEF, Swiss-Prot Q42510. FB:FBgn0028538 · Sec71 SUPPORTS TRANSFER Drosophila single BIG1/BIG2 ortholog; resolves only to an unreviewed (TrEMBL) entry Q9VJW1, so its protein NAME is automatic even though its annotations may be experimental. FB:FBgn0264560 · garz SUPPORTS TRANSFER Drosophila GBF1 ortholog; unreviewed (TrEMBL), four accessions for the same gene. MGI:MGI:1334257 · Cyth1 SUPPORTS TRANSFER Mouse cytohesin-1, a small BFA-insensitive ArfGEF. MGI:MGI:1918215 · Psd3 SUPPORTS TRANSFER Mouse PH/Sec7 protein 3. PomBase:SPAC11E3.11c · SPAC11E3.11c SUPPORTS TRANSFER S. pombe PH and Sec7 domain protein O13690. SGD:S000000748 · GEA2 SUPPORTS TRANSFER Budding yeast GBF/GEA-type ArfGEF. SGD:S000002577 · SEC7 SUPPORTS TRANSFER Budding yeast BIG-type ArfGEF, the family's namesake. SGD:S000003792 · GEA1 SUPPORTS TRANSFER Budding yeast GBF/GEA-type ArfGEF. SGD:S000005241 · MON2 SUPPORTS TRANSFER Verified against UniProt rather than asserted: MON2 (P48563) has NO SEC7 domain feature and no PF01369 Pfam match, while yeast SEC7 (P11075) and human ARFGEF2 both do. MON2 keeps the family's DCB (PF16213) and HUS (PF12783) modules but not the catalytic domain, which is a concrete reason the node-level term stays general. It does still hold GO:0005085 itself, by ISS/IGI/IPI from PMID:12052896 - an observation about MON2's own record, not an argument about ARFGEF2. SGD:S000006299 · SYT1 SUPPORTS TRANSFER Budding yeast Arf GEF. UniProtKB:A5PKW4 · PSD SUPPORTS TRANSFER Human PH and Sec7 domain protein 1. UniProtKB:O43739 · CYTH3 SUPPORTS TRANSFER Human cytohesin-3. UniProtKB:Q15438 · CYTH1 SUPPORTS TRANSFER Human cytohesin-1. UniProtKB:Q5JU85 · IQSEC2 SUPPORTS TRANSFER Human IQ motif and Sec7 domain protein 2, an ARF6 GEF. UniProtKB:Q9UIA0 · CYTH4 SUPPORTS TRANSFER Human cytohesin-4. Supporting Evidence: file:human/ARFGEF2/ARFGEF2-bioinformatics/RESULTS.md Because that clade mixes BIG-type, GBF-type, cytohesin-type and MON2-type GEFs file:human/ARFGEF2/ARFGEF2-bioinformatics/RESULTS.md Independently: **`GO:0005086` ARF guanyl-nucleotide exchange factor activity no |
| GO:0005085 guanyl-nucleotide exchange factor activity | IDA PMID:10212200 Purification and cloning of a brefeldin A-inhibited guanine ... | ACCEPT | Summary: The founding biochemistry: recombinant BIG2 made in Sf9 cells accelerates GTPgammaS binding by ARF1, ARF5 and ARF6, works best on native ARF1/ARF3, and is inhibited by brefeldin A but not by the inactive analog B17. Reason: A direct in vitro exchange assay on purified recombinant protein with a specificity control (B17) - the strongest single piece of evidence in the whole annotation set, and the core molecular function. The substrates actually assayed here are recorded as has_input extensions, since GO no longer has a substrate-specific ARF GEF term to carry them. Supporting Evidence: PMID:10212200 BIG2, synthesized as a His6 fusion protein in Sf9 cells, accelerated guanosine 5'-3-O-(thio)triphosphate binding by recombinant ARF1, ARF5, and ARF6. PMID:10212200 It activated native ARF (mixture of ARF1 and ARF3) more effectively than it did any of the nonmyristoylated recombinant ARFs. PMID:10212200 BIG2 activity was inhibited by BFA in a concentration-dependent manner but not by B17, a structural analog without effects on Golgi function. file:human/ARFGEF2/ARFGEF2-uniprot.txt Inhibited by brefeldin A. |
| GO:0005085 guanyl-nucleotide exchange factor activity | IDA PMID:15385626 BIG2, a guanine nucleotide exchange factor for ADP-ribosylat... | ACCEPT | Summary: In-cell demonstration that BIG2 exchanges nucleotide on the class I ARFs ARF1 and ARF3, using the catalytically dead E738K Sec7 mutant and ARF inactivation as reciprocal controls. Reason: Complements the in vitro assay by establishing the physiological substrate preference. Class I ARFs are recorded as has_input extensions because the substrate-specific term GO:0005086 was merged into GO:0005085 and cannot be used. Supporting Evidence: PMID:15385626 We also have shown that BIG2 has an exchange activity toward class I ARFs (ARF1 and ARF3) in vivo and inactivation of either ARF exaggerates the BIG2(E738K)-induced tubulation of endosomal membranes. |
| GO:0005085 guanyl-nucleotide exchange factor activity | IEA GO_REF:0000120 | ACCEPT | Summary: Automatic assignment of GEF activity from the Sec7 domain signatures IPR000904 and IPR035999 plus an ARBA rule. Reason: A domain-derived catalytic call that is independently confirmed by direct assay on this very protein, so it is not a name-implied activity: UniProt annotates the Sec7 domain at 654-785 and the purified protein has measured exchange activity. The pipeline and the experiment agree. Supporting Evidence: PMID:10212200 BIG2, synthesized as a His6 fusion protein in Sf9 cells, accelerated guanosine 5'-3-O-(thio)triphosphate binding by recombinant ARF1, ARF5, and ARF6. file:human/ARFGEF2/ARFGEF2-uniprot.txt FT DOMAIN 654..785 |
| GO:0005085 guanyl-nucleotide exchange factor activity | TAS PMID:10212200 Purification and cloning of a brefeldin A-inhibited guanine ... | ACCEPT | Summary: Author-statement duplicate of the IDA made from the same paper. Reason: Redundant with the IDA on the identical term and reference rather than wrong. Duplicate GO ids are acceptable and there is no reason to prefer one evidence code over the other; the underlying claim is directly measured in the same paper. Supporting Evidence: PMID:10212200 BIG2, synthesized as a His6 fusion protein in Sf9 cells, accelerated guanosine 5'-3-O-(thio)triphosphate binding by recombinant ARF1, ARF5, and ARF6. |
| GO:0005515 protein binding | IPI PMID:10716990 Identification and localization of two brefeldin A-inhibited... | REMOVE | Summary: BIG2 interacting with its close paralog ARFGEF1/BIG1. The two proteins reciprocally co-immunoprecipitate, each pulling down most of the other, and co-purify in a >670 kDa complex. Reason: Generic protein binding carries no functional information about BIG2. Removing it does not assert that the interaction is false - it is not - only that the term should be replaced by something informative or dropped. The BIG1-BIG2 association is real and quantitatively strong, but the right place to record it is complex membership, which this review does in the gene description and the review notes, not a bare molecular-function term. Note also that these co-immunoprecipitations were performed on bovine brain cytosol; the human-cell work in the same paper is the imaging. Supporting Evidence: PMID:10716990 approximately 70% of BIG2 was immunoprecipitated along with 100% of BIG1 by the anti-BIG1 IgG, and approximately 75% of BIG1 was coprecipitated with 100% of BIG2 by the anti-BIG2 IgG PMID:10716990 All observations were consistent with the conclusion that significant fractions of BIG1 and BIG2 exist as components of the same macromolecular complexes in bovine brain cytosol and are similarly localized in cultured cells. |
| GO:0005515 protein binding | IPI PMID:15705715 Interaction of BIG2, a brefeldin A-inhibited guanine nucleot... | REMOVE | Summary: BIG2 binding the exocyst subunit EXOC7/Exo70, mapped to BIG2 residues 1-643 by yeast two-hybrid and confirmed by co-immunoprecipitation of in vitro-translated fragments. Reason: Generic protein binding carries no functional information about BIG2. Removing it does not assert that the interaction is false - it is not - only that the term should be replaced by something informative or dropped. This interaction is well mapped and is the mechanistic link between BIG2 and the exocyst proposed for transferrin-receptor return to the surface, but GO has no informative molecular-function term for exocyst binding, so there is nothing to modify the annotation to. The functional consequence is already captured by GO:0001881 receptor recycling and the GO:0005815 MTOC localisation from the same paper. Supporting Evidence: PMID:15705715 Clone p2-5-3, encoding a form of human exocyst protein Exo70, interacted with BIG2 amino acids 1-643 and 1-832, but not 644-832, which was confirmed by coimmunoprecipitation of in vitro-translated BIG2 N-terminal segments and 2-5-3. |
| GO:0005515 protein binding | IPI PMID:16866877 AMY-1 (associate of Myc-1) localization to the trans-Golgi n... | REMOVE | Summary: BIG2 binding MYCBP/AMY-1. The targeted study showing the interaction also shows that BIG2, and not BIG1, is what holds AMY-1 at the trans-Golgi network. Reason: Generic protein binding carries no functional information about BIG2. Removing it does not assert that the interaction is false - it is not - only that the term should be replaced by something informative or dropped. MYCBP is the most heavily replicated ARFGEF2 partner in this set - this targeted study plus three orthogonal proteome-scale datasets - so the interaction is not in doubt. The informative content is the RNAi result that AMY-1 needs BIG2 to reach the TGN, which would support GO:0034067 protein localization to Golgi apparatus; I did not add that as a NEW annotation because the cached abstract does not name the cell system and the full text is unavailable, so I could not state the species the experiment was done in. Raised as a suggested question instead. Supporting Evidence: PMID:16866877 we have found that AMY-1 interacts with BIG2 and BIG1, both of which are high molecular weight guanine-nucleotide exchange factors for ADP-ribosylation factors (ARFs) and mainly localize to the TGN PMID:16866877 we have demonstrated that AMY-1 is associated with the TGN through interacting with BIG2 but not with BIG1 using an RNA interference approach, although AMY-1 can interact with both BIG1 and BIG2 in vitro |
| GO:0005515 protein binding | IPI PMID:19332778 Interaction of phosphodiesterase 3A with brefeldin A-inhibit... | REMOVE | Summary: BIG2 binding phosphodiesterase PDE3A, the cAMP-degrading enzyme of the BIG2 AKAP module. PDE3A depletion strips BIG1 and BIG2 off membranes and lowers ARF1-GTP. Reason: Generic protein binding carries no functional information about BIG2. Removing it does not assert that the interaction is false - it is not - only that the term should be replaced by something informative or dropped. The functional content of this interaction - that BIG2 scaffolds a local cAMP signalling module - is already carried informatively by GO:0034237 protein kinase A regulatory subunit binding, so nothing is lost by dropping the generic term. Supporting Evidence: PMID:19332778 Specific depletion of HeLa cell PDE3A with small interfering RNA significantly decreased membrane-associated BIG1 and BIG2, which by confocal immunofluorescence microscopy were widely dispersed from an initial perinuclear Golgi concentration. |
| GO:0005515 protein binding | IPI PMID:19332778 Interaction of phosphodiesterase 3A with brefeldin A-inhibit... | REMOVE | Summary: A second row from the same paper, recording BIG2 with ARFGEF1/BIG1 in the PDE3A-containing complexes. Reason: Generic protein binding carries no functional information about BIG2. Removing it does not assert that the interaction is false - it is not - only that the term should be replaced by something informative or dropped. Same partner as row 9 from a different reference; the BIG1-BIG2 association is recorded in the gene description and the review notes. Supporting Evidence: PMID:19332778 Specific depletion of HeLa cell PDE3A with small interfering RNA significantly decreased membrane-associated BIG1 and BIG2, which by confocal immunofluorescence microscopy were widely dispersed from an initial perinuclear Golgi concentration. |
| GO:0005515 protein binding | IPI PMID:22084092 Effects of brefeldin A-inhibited guanine nucleotide-exchange... | REMOVE | Summary: ARFGEF2 with ARFGEF1 again, here incidental to a BIG1-centred study of KANK1 in wound-healing directed migration. Reason: Generic protein binding carries no functional information about BIG2. Removing it does not assert that the interaction is false - it is not - only that the term should be replaced by something informative or dropped. Querying GOA by this reference shows it produced 14 annotations across 4 entities, 7 of them on ARFGEF1 and just this one on ARFGEF2 - the paper is about BIG1, and ARFGEF2's only stake in it is the same paralog association already recorded elsewhere. Supporting Evidence: file:human/ARFGEF2/ARFGEF2-bioinformatics/RESULTS.md | `PMID:22084092` | 4 | 14 | 1 | 7 | BIG1-centred; ARFGEF2 gets one IPI row | |
| GO:0005515 protein binding | IPI PMID:33961781 Dual proteome-scale networks reveal cell-specific remodeling... | REMOVE | Summary: MYCBP recovered with ARFGEF2 in the BioPlex affinity-purification mass-spectrometry interactome. Reason: Generic protein binding carries no functional information about BIG2. Removing it does not assert that the interaction is false - it is not - only that the term should be replaced by something informative or dropped. This is one of three independent proteome-scale datasets that recover the same partner, which is why the MYCBP interaction is treated as solid; but a high-throughput AP-MS hit expressed as bare protein binding tells a curator nothing about what BIG2 does. The reference is too large for QuickGO to page (9514 annotations), so its entity count is reported as unavailable rather than guessed - the per-gene count is exact and is 1. |
| GO:0005515 protein binding | IPI PMID:35271311 OpenCell: Endogenous tagging for the cartography of human ce... | REMOVE | Summary: MYCBP recovered with ARFGEF2 by endogenous tagging and immunoprecipitation mass spectrometry in OpenCell. Reason: Generic protein binding carries no functional information about BIG2. Removing it does not assert that the interaction is false - it is not - only that the term should be replaced by something informative or dropped. Independent methodology from BioPlex (endogenous split-mNeonGreen tagging rather than overexpressed bait), which is what makes the MYCBP interaction credible; the term itself is still uninformative. |
| GO:0005515 protein binding | IPI PMID:35271311 OpenCell: Endogenous tagging for the cartography of human ce... | REMOVE | Summary: ARFGEF1 recovered with ARFGEF2 in the same OpenCell dataset. Reason: Generic protein binding carries no functional information about BIG2. Removing it does not assert that the interaction is false - it is not - only that the term should be replaced by something informative or dropped. Corroborates the BIG1-BIG2 complex at endogenous expression levels, which is worth noting because the classical evidence was from overexpression and from bovine brain; the fact is recorded in the gene description and the review notes. |
| GO:0005515 protein binding | IPI PMID:40205054 Multimodal cell maps as a foundation for structural and func... | REMOVE | Summary: MYCBP recovered with ARFGEF2 in the multimodal cell-map interactome. Reason: Generic protein binding carries no functional information about BIG2. Removing it does not assert that the interaction is false - it is not - only that the term should be replaced by something informative or dropped. Third orthogonal screen recovering MYCBP. All four MYCBP rows are removed as uninformative rather than as unreliable. |
| GO:0005737 cytoplasm | IEA GO_REF:0000044 | ACCEPT | Summary: Cytoplasm, from the UniProt subcellular-location vocabulary. BIG2 is predominantly cytosolic at rest and moves onto membranes when cAMP rises. Reason: Broad but correct, and the resting cytosolic pool is functionally meaningful for this protein rather than incidental - membrane recruitment is the regulated step. Supporting Evidence: PMID:12571360 Western blot analysis of subcellular fractions demonstrated translocation of BIG2 (and BIG1) from cytosol to the Golgi and other membrane structures after incubation of cells with 8-Br-cAMP or forskolin. PMID:10716990 On Western blot analysis, both BIG1 and BIG2 were clearly more abundant in the cytosol than in the microsomal fractions. |
| GO:0005768 endosome | IEA GO_REF:0000044 | ACCEPT | Summary: Endosome, from the UniProt subcellular-location vocabulary; the parent of the experimentally established recycling-endosome localisation. Reason: Correct and supported by direct human evidence one level down (GO:0055037 recycling endosome, IDA). Keeping the parent alongside the child is harmless. Supporting Evidence: PMID:15385626 we have revealed that another population of BIG2 is associated with the recycling endosome and found that expression of a catalytically inactive BIG2 mutant, E738K, selectively induces membrane tubules from this compartment |
| GO:0005794 Golgi apparatus | IDA GO_REF:0000052 | ACCEPT | Summary: Golgi apparatus by Human Protein Atlas immunofluorescence. Reason: Independent antibody-based imaging that agrees with the targeted literature. The more informative child GO:0000139 Golgi membrane is separately annotated by IDA. Supporting Evidence: PMID:10716990 endogenous BIG1 and coexpressed BIG2 were distributed in a punctate pattern throughout the cytosol, and also concentrated in the perinuclear region, where endogenous BIG1 and BIG2 each partially colocalized with Golgi-specific 58K protein and gamma-adaptin |
| GO:0005794 Golgi apparatus | IEA GO_REF:0000120 | ACCEPT | Summary: Golgi apparatus from combined automated annotation (ARBA plus the UniProt subcellular location vocabulary). Reason: Agrees with the IDA rows; no conflict to resolve. |
| GO:0005802 trans-Golgi network | IBA GO_REF:0000033 | ACCEPT | Summary: PAINT inference that BIG2 acts at the trans-Golgi network, from a node whose donors are genuine orthologs: Drosophila Sec71 (the single fly BIG1/BIG2 ortholog), rat Arfgef1 and rat Arfgef2. This is BIG2's principal site of action. Reason: An ortholog-restricted node with an is_active_in qualifier, which is the right relation for a GEF that must be on the TGN membrane to act. The donor set is orthologous rather than paralogous drift, and human BIG2 has its own TGN evidence. Propagation Review Root cause: NO FAILURE CORE Sources checked: PANTHER:PTN002240728 · BIG1/BIG2 ancestral node SUPPORTS TRANSFER PAINT ancestral node, not a protein; reach is the large BFA-sensitive ArfGEF clade. FB:FBgn0028538 · Sec71 SUPPORTS TRANSFER Drosophila single BIG1/BIG2 ortholog; unreviewed (TrEMBL) entry, so its name is automatic. RGD:1560793 · Arfgef1 (rat BIG1) SUPPORTS TRANSFER Rat paralog; TGN localisation is shared by both BIGs. RGD:631430 · Arfgef2 (rat BIG2) SUPPORTS TRANSFER True rat ortholog, Swiss-Prot Q7TSU1, with its own TGN IDA from two independent papers. Supporting Evidence: PMID:12051703 These observations indicate that BIG2 is involved in trafficking from the TGN by regulating membrane association of AP-1 and GGA through activating ARF. |
| GO:0005802 trans-Golgi network | IEA GO_REF:0000107 | ACCEPT | Summary: trans-Golgi network transferred from rat Arfgef2 by Ensembl Compara. The rat annotation is itself an IDA supported by two independent papers. Reason: One of the few rows in the 20-row rat projection block whose term is independently established for the human protein as well, so the projection adds no risk. Unlike the synaptic terms from the same donor, this one is corroborated by human-cell work. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q7TSU1 · Arfgef2 (rat BIG2) SUPPORTS TRANSFER True ortholog; holds GO:0005802 by IDA from two separate references (PMID:11809827 and PMID:15198677), and the term is independently true of human BIG2. Supporting Evidence: PMID:15198677 Double label immunofluorescence of cultured hippocampal neurons and electron microscopy immunocytochemistry of rat brain tissue show that BIG2 concentrates in the trans-Golgi network. |
| GO:0005802 trans-Golgi network | ISS GO_REF:0000024 | ACCEPT | Summary: The same rat-derived trans-Golgi network call, arriving a second time as a curator-judged ISS transfer. Reason: Correct term, redundant route. It is worth recording that UniProt ISS (GO_REF:0000024) and Ensembl Compara (GO_REF:0000107) independently project the same rat annotations into this entry, which is why 20 of the 66 rows trace to one donor - here the duplication is harmless because the term is right. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:Q7TSU1 · Arfgef2 (rat BIG2) SUPPORTS TRANSFER Same rat donor as the Compara row; duplicate pipeline, same underlying IDA. Supporting Evidence: file:human/ARFGEF2/ARFGEF2-bioinformatics/RESULTS.md **20 of the 66 rows**, covering **13 distinct GO terms**, carry rat Arfgef2 |
| GO:0005813 centrosome | IEA GO_REF:0000044 | KEEP AS NON CORE | Summary: Centrosome, from the UniProt subcellular-location vocabulary, resting on the finding that BIG2 and Exo70 are present in purified human centrosomes. Reason: Real but peripheral. Centrosomal BIG2 has never been given a function; the protein's characterised activity is at the TGN and recycling endosome. Keep as a secondary location. Supporting Evidence: PMID:15705715 By immunofluorescence microscopy, endogenous BIG2 and Exo70 in HepG2 cells were visualized at Golgi membranes and apparently at the microtubule-organizing center (MTOC). Both were identified in purified centrosomes. |
| GO:0005815 microtubule organizing center | IDA PMID:15705715 Interaction of BIG2, a brefeldin A-inhibited guanine nucleot... | KEEP AS NON CORE | Summary: BIG2 at the microtubule-organizing center, seen by confocal microscopy overlapping gamma-tubulin and recovered in purified centrosome fractions from human HepG2 cells. Reason: A genuine direct observation in human cells, but no BIG2 function has been attached to the MTOC and the localisation survives brefeldin A treatment while the Golgi pool disperses, suggesting a distinct pool rather than the working one. Keep, non-core. Supporting Evidence: PMID:15705715 By immunofluorescence microscopy, endogenous BIG2 and Exo70 in HepG2 cells were visualized at Golgi membranes and apparently at the microtubule-organizing center (MTOC). Both were identified in purified centrosomes. |
| GO:0005829 cytosol | IDA PMID:12571360 Protein kinase A-anchoring (AKAP) domains in brefeldin A-inh... | ACCEPT | Summary: Cytosol, established by subcellular fractionation as the compartment BIG2 leaves when cAMP rises. Reason: The cytosolic pool is the resting state of this protein and the starting point of its regulated recruitment, so the term is informative rather than a default localisation. Supporting Evidence: PMID:12571360 Western blot analysis of subcellular fractions demonstrated translocation of BIG2 (and BIG1) from cytosol to the Golgi and other membrane structures after incubation of cells with 8-Br-cAMP or forskolin. PMID:10716990 endogenous BIG1 and coexpressed BIG2 were distributed in a punctate pattern throughout the cytosol, and also concentrated in the perinuclear region, where endogenous BIG1 and BIG2 each partially colocalized with Golgi-specific 58K protein and gamma-adaptin PMID:10716990 On Western blot analysis, both BIG1 and BIG2 were clearly more abundant in the cytosol than in the microsomal fractions. |
| GO:0005856 cytoskeleton | IEA GO_REF:0000044 | KEEP AS NON CORE | Summary: Cytoskeleton, from the UniProt subcellular-location vocabulary. Reason: This generic term is the honest level for what has been observed: BIG2 co-purifies with microtubule preparations and centrosomes, and the rat entry's own curated subcellular location says only 'Cytoplasm, cytoskeleton'. Keeping the general term is correct here, and is precisely the level at which the axonemal-microtubule child overreaches. Supporting Evidence: PMID:15705715 By immunofluorescence microscopy, endogenous BIG2 and Exo70 in HepG2 cells were visualized at Golgi membranes and apparently at the microtubule-organizing center (MTOC). Both were identified in purified centrosomes. PMID:16477018 Endogenous BIG2 and Exo70, a component of the exocyst complex required for exocytosis, were present in preparations of microtubules purified from human cells and with γ-tubulin in centrosome fractions from the same cells |
| GO:0005879 axonemal microtubule | IEA GO_REF:0000107 | MARK AS OVER ANNOTATED | Summary: Axonemal microtubule, transferred from rat Arfgef2 by Ensembl Compara. It rests on a single rat IDA from a synaptic-trafficking paper with no ciliary content, and conflicts with the generic 'cytoskeleton' that the same rat record states in prose. Reason: This is the one row in the rat projection block I cannot reconcile. The term claims a cilium or flagellum: GO:0005879 is defined as a microtubule in the axoneme of a eukaryotic cilium or flagellum. Three observations bear on it. First, the donor paper is a study of GABA(A) receptor trafficking whose systems are cultured hippocampal neurons, HEK293 cells and electron-microscopic immunocytochemistry of rat brain; its abstract contains no ciliary or flagellar content at all. Second, the rat UniProt entry's own SUBCELLULAR LOCATION section, curated from that same paper, records the generic 'Cytoplasm, cytoskeleton' and lists no cilium, axoneme or flagellum. Third, a census over the four human large ArfGEFs (ARFGEF1, ARFGEF2, ARFGEF3, GBF1, accessions derived by gene-name lookup rather than written by hand) plus the nine proteins resolvable from this gene's own WITH/FROM column - 13 accessions in all - finds exactly one cilium-compartment holder among the large ArfGEFs, and it is ARFGEF2 itself by this very projection; the cytoskeletal association that does exist for human BIG2 is with cytoplasmic microtubules and centrosomes. The census also returned a result that cuts the other way and is reported rather than suppressed: EXOC7/Exo70, BIG2's own exocyst partner, carries GO:0036064 ciliary basal body, so a ciliary context near BIG2 is not absurd - which is another reason to flag this row for measurement rather than delete it. I have NOT concluded that the rat IDA is wrong - the full text is subscription-only (Europe PMC reports isOpenAccess: N, inEPMC: N, no PMC identifier) and I could not see the figure the curator saw. What I am judging is the non-experimental transfer of an unreproduced, compartment-specific localisation into human, where the general parent GO:0005856 cytoskeleton is already annotated and covers everything that is actually established. On the choice of action: MARK_AS_OVER_ANNOTATED rather than REMOVE is deliberate. REMOVE has to be earned by a positive argument, and the one thing that would earn it here - reading the figure the rat IDA rests on - is exactly what the paywall prevents; the census that does the work is a negative result about the family, not a demonstration that BIG2 is absent from cilia, and the EXOC7 ciliary-basal-body hit shows the surrounding biology is not hostile to the claim. UNDECIDED would discard a measurement that was made. The term is kept visible and flagged, which is what a curator can act on. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: COMPARTMENT OR COMPLEX MISMATCH GRANULARITY MISMATCH Sources checked: UniProtKB:Q7TSU1 · Arfgef2 (rat BIG2) SUPPORTS SOURCE BUT NOT TARGET Single rat IDA from PMID:15198677, assigned by UniProt. The same UniProt record's prose subcellular location says only 'Cytoplasm, cytoskeleton'. Donor full text is paywalled, so the rat row is flagged for re-examination rather than contradicted. Supporting Evidence: PMID:15198677 Double label immunofluorescence of cultured hippocampal neurons and electron microscopy immunocytochemistry of rat brain tissue show that BIG2 concentrates in the trans-Golgi network. PMID:15198677 BIG2 is also present in vesicle-like structures in the dendritic cytoplasm, sometimes colocalizing with GABA(A)Rs. PMID:15705715 By immunofluorescence microscopy, endogenous BIG2 and Exo70 in HepG2 cells were visualized at Golgi membranes and apparently at the microtubule-organizing center (MTOC). Both were identified in purified centrosomes. file:human/ARFGEF2/ARFGEF2-bioinformatics/RESULTS.md | `Q9Y6D5` | ARFGEF2 | `GO:0005879` | this projection only | |
| GO:0005879 axonemal microtubule | ISS GO_REF:0000024 | MARK AS OVER ANNOTATED | Summary: The same axonemal-microtubule claim arriving a second time, as a curator-judged ISS transfer from the same single rat IDA. Reason: This is the one row in the rat projection block I cannot reconcile. The term claims a cilium or flagellum: GO:0005879 is defined as a microtubule in the axoneme of a eukaryotic cilium or flagellum. Three observations bear on it. First, the donor paper is a study of GABA(A) receptor trafficking whose systems are cultured hippocampal neurons, HEK293 cells and electron-microscopic immunocytochemistry of rat brain; its abstract contains no ciliary or flagellar content at all. Second, the rat UniProt entry's own SUBCELLULAR LOCATION section, curated from that same paper, records the generic 'Cytoplasm, cytoskeleton' and lists no cilium, axoneme or flagellum. Third, a census over the four human large ArfGEFs (ARFGEF1, ARFGEF2, ARFGEF3, GBF1, accessions derived by gene-name lookup rather than written by hand) plus the nine proteins resolvable from this gene's own WITH/FROM column - 13 accessions in all - finds exactly one cilium-compartment holder among the large ArfGEFs, and it is ARFGEF2 itself by this very projection; the cytoskeletal association that does exist for human BIG2 is with cytoplasmic microtubules and centrosomes. The census also returned a result that cuts the other way and is reported rather than suppressed: EXOC7/Exo70, BIG2's own exocyst partner, carries GO:0036064 ciliary basal body, so a ciliary context near BIG2 is not absurd - which is another reason to flag this row for measurement rather than delete it. I have NOT concluded that the rat IDA is wrong - the full text is subscription-only (Europe PMC reports isOpenAccess: N, inEPMC: N, no PMC identifier) and I could not see the figure the curator saw. What I am judging is the non-experimental transfer of an unreproduced, compartment-specific localisation into human, where the general parent GO:0005856 cytoskeleton is already annotated and covers everything that is actually established. This row and the Compara row are the same rat datum counted twice by two independent pipelines, so removing or re-scoping one without the other would leave the claim standing. On the choice of action: MARK_AS_OVER_ANNOTATED rather than REMOVE is deliberate. REMOVE has to be earned by a positive argument, and the one thing that would earn it here - reading the figure the rat IDA rests on - is exactly what the paywall prevents; the census that does the work is a negative result about the family, not a demonstration that BIG2 is absent from cilia, and the EXOC7 ciliary-basal-body hit shows the surrounding biology is not hostile to the claim. UNDECIDED would discard a measurement that was made. The term is kept visible and flagged, which is what a curator can act on. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: COMPARTMENT OR COMPLEX MISMATCH GRANULARITY MISMATCH Sources checked: UniProtKB:Q7TSU1 · Arfgef2 (rat BIG2) SUPPORTS SOURCE BUT NOT TARGET Same single rat IDA as the Compara row; duplicate transfer by a second pipeline. Supporting Evidence: file:human/ARFGEF2/ARFGEF2-bioinformatics/RESULTS.md **20 of the 66 rows**, covering **13 distinct GO terms**, carry rat Arfgef2 file:human/ARFGEF2/ARFGEF2-bioinformatics/RESULTS.md | `Q9Y6D5` | ARFGEF2 | `GO:0005879` | this projection only | |
| GO:0006887 exocytosis | IEA GO_REF:0000120 | KEEP AS NON CORE | Summary: Exocytosis from combined automated annotation, carrying across the rat IMP that BIG2 promotes GABA(A) receptor exit from the ER and delivery to the cell surface. Reason: Defensible: BIG2 does act on the exocytic route, both in the rat GABA(A) receptor work and in the human TNFR1 exosome-release study. It is not the core activity - that is ARF exchange - and the more specific human processes are separately annotated, so keep it as a contextual term. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q7TSU1 · Arfgef2 (rat BIG2) SUPPORTS TRANSFER Rat IMP from PMID:15198677; the ER-exit experiment in that paper was done in human HEK293 cells, so the human relevance is direct. Supporting Evidence: PMID:15198677 In transfected human embryonic kidney cell line 293 cells, BIG2 promotes the exit of GABA(A)Rs from endoplasmic reticulum. |
| GO:0006887 exocytosis | TAS PMID:10212200 Purification and cloning of a brefeldin A-inhibited guanine ... | KEEP AS NON CORE | Summary: A legacy author-statement annotation of exocytosis taken from the 1999 cloning and purification paper, which measures nucleotide exchange and brefeldin A sensitivity and does not assay exocytosis. Reason: Kept, non-core. The term is right for BIG2 - it is separately supported by the rat GABA(A) receptor work and by the human TNFR1 exosome-release study - and the cited paper does describe brefeldin A as a metabolite that blocks protein secretion, so the author statement is not unfounded. What is worth recording is that this row is statement-level: the 1999 paper measures nucleotide exchange and brefeldin A sensitivity, not exocytosis, and the same reference already supplies a directly measured GO:0005085 IDA which is what it actually establishes. Harmonised with the GO:0006887 IEA row, which carries the rat IMP. Supporting Evidence: PMID:10212200 BIG2, synthesized as a His6 fusion protein in Sf9 cells, accelerated guanosine 5'-3-O-(thio)triphosphate binding by recombinant ARF1, ARF5, and ARF6. |
| GO:0006893 Golgi to plasma membrane transport | IBA GO_REF:0000033 | ACCEPT | Summary: PAINT inference that BIG2 is involved in Golgi-to-plasma-membrane transport. The WITH/FROM column cites ARFGEF2 itself alongside the ancestral node. Reason: The self-reference is the expected signature that the target's own experimental annotation was one of the descendant evidences used to place the ancestral inference; it is a marker of experimental grounding, not circularity. The term is well supported for human BIG2 by the E-cadherin/beta-catenin and Filamin A cargo studies, neither of which has produced any GO annotation. Propagation Review Root cause: NO FAILURE CORE Sources checked: PANTHER:PTN002580165 · BIG2 ancestral node SUPPORTS TRANSFER PAINT ancestral node, not a protein. UniProtKB:Q9Y6D5 · ARFGEF2 (self) SUPPORTS TRANSFER Self-referential and valid: the target's own IMP is one of the evidences behind the IBD. Supporting Evidence: PMID:14647276 ARFGEF2 encodes the large (>200 kDa) brefeldin A (BFA)-inhibited GEF2 protein (BIG2), which is required for vesicle and membrane trafficking from the trans-Golgi network (TGN). PMID:14647276 Inhibition of BIG2 also disturbed the intracellular localization of such molecules as E-cadherin and beta-catenin by preventing their transport from the Golgi apparatus to the cell surface. |
| GO:0006893 Golgi to plasma membrane transport | IMP PMID:17276987 The brefeldin A-inhibited guanine nucleotide-exchange protei... | ACCEPT | Summary: BIG2 knockdown in human umbilical vein endothelial cells reduces constitutive release of TNFR1 in exosome-like vesicles, in an ARF1- and ARF3-dependent way, with BIG1 as a negative control. Reason: The term is right and core for BIG2 and I am accepting it rather than modifying it, but it is worth recording that the experiment cited supports it only obliquely: exosome-like vesicle release is a multivesicular-body-to-plasma-membrane route rather than the trans-Golgi-to-plasma-membrane route the term names. The strongest evidence for Golgi-to-surface transport by BIG2 is the E-cadherin/beta-catenin result and the Filamin A result, neither of which carries any GO annotation - so the fix is to add references, not to change the term. Supporting Evidence: PMID:17276987 Effects of specific RNA interference (RNAi) showed that BIG2, but not BIG1, regulated the release of TNFR1 exosome-like vesicles, whereas neither BIG2 nor BIG1 was required for the IL-1beta-induced proteolytic cleavage of TNFR1 ectodomains. PMID:14647276 ARFGEF2 encodes the large (>200 kDa) brefeldin A (BFA)-inhibited GEF2 protein (BIG2), which is required for vesicle and membrane trafficking from the trans-Golgi network (TGN). PMID:14647276 Inhibition of BIG2 also disturbed the intracellular localization of such molecules as E-cadherin and beta-catenin by preventing their transport from the Golgi apparatus to the cell surface. |
| GO:0007032 endosome organization | IBA GO_REF:0000033 | ACCEPT | Summary: PAINT inference that BIG2 organises the endosomal compartment, again citing ARFGEF2 itself among the sources. Reason: Endosome organisation is the function that most cleanly distinguishes BIG2 from BIG1, and the human IMP backing it is a direct comparative knockdown. The self-referential source is the expected marker of experimental grounding on the target. Propagation Review Root cause: NO FAILURE CORE Sources checked: PANTHER:PTN002580165 · BIG2 ancestral node SUPPORTS TRANSFER PAINT ancestral node, not a protein. UniProtKB:Q9Y6D5 · ARFGEF2 (self) SUPPORTS TRANSFER Self-referential and valid: the target's own IMP (PMID:20360857) is a descendant evidence for the IBD. Supporting Evidence: PMID:20360857 BIG1 is required to maintain the normal morphology of the Golgi; BIG2 is important for endosomal compartment integrity and cannot replace the function of BIG1 in Golgi organization. |
| GO:0007032 endosome organization | IMP PMID:20360857 Specific functions of BIG1 and BIG2 in endomembrane organiza... | ACCEPT | Summary: siRNA depletion of BIG2, but not BIG1, in human cells tubulates the recycling endosomal compartment; BIG1 depletion instead fragments the Golgi. A direct, controlled dissection of the two paralogs. Reason: The single most discriminating experiment in the BIG2 literature and the clearest statement of a non-redundant BIG2 function. Core. Supporting Evidence: PMID:20360857 Our results show that depletion of BIG2 but not of BIG1 induces a tubulation of the recycling endosomal compartment, consistent with a specific role for BIG2 here. PMID:20360857 BIG1 is required to maintain the normal morphology of the Golgi; BIG2 is important for endosomal compartment integrity and cannot replace the function of BIG1 in Golgi organization. PMID:20360857 Here we have used small-interfering RNA on human cells and a combination of fixed and live-cell imaging to investigate the differential functions of BIG1 and BIG2 in endomembrane organization and function. |
| GO:0010256 endomembrane system organization | IEA GO_REF:0000117 | ACCEPT | Summary: Endomembrane system organization from an ARBA machine-learning rule; the parent of the experimentally established endosome-organisation role. Reason: Correct, and the specific child is separately annotated by IMP from the same body of work. An IEA sitting above a curated child is not an over-annotation. Supporting Evidence: PMID:20360857 BIG1 is required to maintain the normal morphology of the Golgi; BIG2 is important for endosomal compartment integrity and cannot replace the function of BIG1 in Golgi organization. |
| GO:0010256 endomembrane system organization | IMP PMID:20360857 Specific functions of BIG1 and BIG2 in endomembrane organiza... | ACCEPT | Summary: The same human knockdown study read at the level of the endomembrane system as a whole. Reason: Supported by the same controlled comparison as GO:0007032; keeping both the general and the specific term is standard and harmless. Supporting Evidence: PMID:20360857 Our results show that depletion of BIG2 but not of BIG1 induces a tubulation of the recycling endosomal compartment, consistent with a specific role for BIG2 here. |
| GO:0016020 membrane | HDA PMID:19946888 Defining the membrane proteome of NK cells. | KEEP AS NON CORE | Summary: Membrane, from a high-throughput mass-spectrometry survey of the NK-like cell line YTS membrane proteome. Reason: The term is true of BIG2 - it is a peripheral membrane protein recruited from cytosol - but this particular row is a proteome sweep rather than a BIG2 result: the study identified 1843 proteins and the reference assigns GO:0016020 to 1142 distinct entities across GOA, with the authors themselves noting that much of what they detected is only transiently membrane-associated. Keep as non-core; the informative membrane annotations are the IDA rows. Supporting Evidence: PMID:19946888 Mass spectrometric analysis identified 1843 proteins with high confidence scores. PMID:19946888 The remaining species were largely involved in cellular processes and molecular functions that could be predicted to be transiently associated with membranes. file:human/ARFGEF2/ARFGEF2-bioinformatics/RESULTS.md `PMID:19946888` (NK-cell membrane proteome) assigns `GO:0016020` membrane to |
| GO:0016020 membrane | IDA PMID:12571360 Protein kinase A-anchoring (AKAP) domains in brefeldin A-inh... | KEEP AS NON CORE | Summary: Membrane association demonstrated directly by subcellular fractionation, as the cAMP-inducible endpoint of BIG2 translocation. Reason: Membrane association demonstrated directly by subcellular fractionation, as the cAMP-inducible endpoint of BIG2 translocation, so the evidence is good. Non-core only because GO:0016020 is the uninformative parent: the compartments that matter for this protein - Golgi membrane, trans-Golgi network and recycling endosome - are separately annotated with their own direct evidence. Harmonised with the other two membrane rows. Supporting Evidence: PMID:12571360 Western blot analysis of subcellular fractions demonstrated translocation of BIG2 (and BIG1) from cytosol to the Golgi and other membrane structures after incubation of cells with 8-Br-cAMP or forskolin. |
| GO:0016020 membrane | IEA GO_REF:0000044 | KEEP AS NON CORE | Summary: Membrane, from the UniProt subcellular-location vocabulary. Reason: Membrane, from the UniProt subcellular-location vocabulary. Correct and consistent with the direct fractionation evidence, but the general parent rather than an informative location; harmonised with the IDA and HDA rows on the same term. |
| GO:0017022 myosin binding | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: Myosin binding, transferred from rat Arfgef2 by Ensembl Compara. The rat row comes from a paper about the paralog BIG1 - but the term is nevertheless independently true of human BIG2 from a different study, so the defect is the evidence chain, not the term. Reason: The donor reference PMID:15644318 is titled 'BIG1 is a binding partner of myosin IXb', its yeast two-hybrid screen recovered BIG1, and its co-immunoprecipitations used anti-BIG1 antibodies. Querying GOA by that reference returns 18 annotations across 7 entities: 6 of them on human ARFGEF1, none on human ARFGEF2, and exactly one row on any BIG2-family protein anywhere - GO:0017022 on rat Arfgef2. Human ARFGEF2's myosin-binding annotation is the Compara projection of that single rat row, while human ARFGEF1 holds the same term directly by IPI from the same paper. I have NOT concluded that the rat curator's IPI is wrong: PMID:15644318's full text is not in the cache, so the by-reference census establishes what curators did, not what the full text contains - the same posture taken on the GO:0005879 rat row. The propagation_review is therefore classified on what is verifiable, EVIDENCE_CIRCULAR_OR_REDUNDANT: human BIG2 has stronger direct evidence for this term than the projection carries. I am keeping the term rather than removing it because it is independently established for BIG2 by PMID:23918382, which shows reciprocal co-immunoprecipitation of endogenous BIG2 and non-muscle myosin heavy chain IIA in HeLa cells and direct binding between separately in vitro-synthesised BIG2 and NMHC IIA. The actionable recommendation is to re-reference this row to PMID:23918382 with human evidence. It is non-core: myosin binding is a non-catalytic scaffolding contact, and the same paper shows BIG2 does not share BIG1's direct binding to MYPT1 and PP1c-delta. Propagation Review Root cause: EVIDENCE CIRCULAR OR REDUNDANT Sources checked: UniProtKB:Q7TSU1 · Arfgef2 (rat BIG2) UNRESOLVED Rat Arfgef2's GO:0017022 IPI comes from PMID:15644318, whose full text is NOT in the cache (full_text_available: false), so the rat curator's call is NOT contradicted here. What is established: across GOA that reference placed 6 annotations on human ARFGEF1 and 0 on human ARFGEF2, and the rat row is the only BIG2-family row it produced anywhere. That shows what curators did, not what the full text contains. The human row is classified redundant rather than defective because human BIG2 has stronger direct evidence for the term (PMID:23918382, read in full). Supporting Evidence: PMID:15644318 Through the yeast two-hybrid screening using the tail domain of myosin IXb as bait we found BIG1, a guanine nucleotide exchange factor for ADP-ribosylation factor (Arf1), as a potential binding partner for myosin IXb. PMID:15644318 The interaction between myosin IXb and BIG1 was demonstrated by co-immunoprecipitation of endogenous myosin IXb and BIG1 with anti-BIG1 antibodies in normal rat kidney cells. PMID:23918382 Reciprocal coimmunoprecipitation of endogenous HeLa cell BIG1 and BIG2 with myosin IIA was demonstrably independent of Arf guanine nucleotide-exchange factor activity PMID:23918382 After incubation of in vitro-synthesized BIG2 and NMHC IIA, BIG2 IP collected approximately 1% of added NMHC IIA PMID:23918382 Despite >70% sequence identity of the two C fragments, no interaction of BIG2-C with MYPT1 or PP1cδ was detected file:human/ARFGEF2/ARFGEF2-bioinformatics/RESULTS.md - **0 are on human ARFGEF2** (`Q9Y6D5`). |
| GO:0030425 dendrite | IEA GO_REF:0000044 | KEEP AS NON CORE | Summary: Dendrite, from the UniProt subcellular-location vocabulary, resting on the rat brain immuno-electron-microscopy study. Reason: A genuine neuronal observation - BIG2 is in vesicle-like structures in dendritic cytoplasm - but a cell-type-specific location rather than part of the core molecular story. Supporting Evidence: PMID:15198677 BIG2 is also present in vesicle-like structures in the dendritic cytoplasm, sometimes colocalizing with GABA(A)Rs. |
| GO:0031410 cytoplasmic vesicle | IEA GO_REF:0000120 | KEEP AS NON CORE | Summary: Cytoplasmic vesicle, from combined automated annotation carrying the rat IDA. Reason: Broad and correct for a trafficking protein, and directly observed in the rat study. Non-core because the informative compartments (TGN, recycling endosome, Golgi membrane) are separately annotated with human evidence. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q7TSU1 · Arfgef2 (rat BIG2) SUPPORTS TRANSFER Rat IDA from PMID:15198677; a broad compartment term that transfers safely. Supporting Evidence: PMID:15198677 BIG2 is also present in vesicle-like structures in the dendritic cytoplasm, sometimes colocalizing with GABA(A)Rs. |
| GO:0031410 cytoplasmic vesicle | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: The same rat-derived cytoplasmic-vesicle call arriving again by the UniProt ISS route. Reason: Duplicate of the Compara row from the same rat IDA. Correct but redundant; part of the 20-row double projection from this one donor. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q7TSU1 · Arfgef2 (rat BIG2) SUPPORTS TRANSFER Same rat IDA as the Compara row; duplicate pipeline. Supporting Evidence: file:human/ARFGEF2/ARFGEF2-bioinformatics/RESULTS.md **20 of the 66 rows**, covering **13 distinct GO terms**, carry rat Arfgef2 |
| GO:0032012 regulation of ARF protein signal transduction | IEA GO_REF:0000002 | ACCEPT | Summary: Regulation of ARF protein signal transduction, mapped automatically from the Sec7 domain signatures. This is the process counterpart of BIG2's core molecular function. Reason: An InterPro2GO mapping that is independently confirmed by direct assay on this protein: BIG2 loads GTP onto ARF1/ARF3 and the catalytically dead E738K mutant produces the expected coat-recruitment and membrane-tubulation phenotypes. Core. Supporting Evidence: PMID:15385626 We also have shown that BIG2 has an exchange activity toward class I ARFs (ARF1 and ARF3) in vivo and inactivation of either ARF exaggerates the BIG2(E738K)-induced tubulation of endosomal membranes. PMID:12051703 These observations indicate that BIG2 is involved in trafficking from the TGN by regulating membrane association of AP-1 and GGA through activating ARF. |
| GO:0032279 asymmetric synapse | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: Localisation of BIG2 to asymmetric (excitatory) synapses, transferred from rat Arfgef2 by Ensembl Compara. The rat evidence is immuno-electron microscopy of rat brain tissue in a study of GABA(A) receptor trafficking. Reason: Part of a block of 20 rows, covering 13 GO terms, that Ensembl Compara and UniProt ISS both project from rat Arfgef2; 17 of those rows rest solely on one paper, PMID:15198677. Two things are worth separating. The projection discriminator comes back NEGATIVE: that reference annotates only 2 entities in all of GOA - rat Arfgef2 and its partner Gabrb3 - so this is genuine per-gene curation at figure granularity by SynGO and UniProt, not a complex-membership phenotype distributed across subunits. The annotation is therefore kept. But it is not core: the observation is rat brain immuno-electron microscopy of a broadly expressed trafficking protein, ARFGEF2 is expressed in placenta, lung, heart, kidney and pancreas as well as brain, and the synaptic presence has never been given a synapse-specific mechanism. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q7TSU1 · Arfgef2 (rat BIG2) SUPPORTS TRANSFER Rat IDA from PMID:15198677, which annotates only 2 entities in all of GOA, so it is per-gene curation rather than a distributed projection. Transfers safely but is cell-type-specific. Supporting Evidence: PMID:15198677 BIG2 is present in both inhibitory GABAergic synapses that contain GABA(A)Rs and in asymmetric excitatory synapses. PMID:15198677 Double label immunofluorescence of cultured hippocampal neurons and electron microscopy immunocytochemistry of rat brain tissue show that BIG2 concentrates in the trans-Golgi network. file:human/ARFGEF2/ARFGEF2-bioinformatics/RESULTS.md `PMID:15198677` annotates only **2 entities** in the whole of GOA — rat Arfgef2 |
| GO:0032279 asymmetric synapse | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: Localisation of BIG2 to asymmetric (excitatory) synapses, transferred from rat Arfgef2 by UniProt curator-judged ISS. The rat evidence is immuno-electron microscopy of rat brain tissue in a study of GABA(A) receptor trafficking. Reason: Part of a block of 20 rows, covering 13 GO terms, that Ensembl Compara and UniProt ISS both project from rat Arfgef2; 17 of those rows rest solely on one paper, PMID:15198677. Two things are worth separating. The projection discriminator comes back NEGATIVE: that reference annotates only 2 entities in all of GOA - rat Arfgef2 and its partner Gabrb3 - so this is genuine per-gene curation at figure granularity by SynGO and UniProt, not a complex-membership phenotype distributed across subunits. The annotation is therefore kept. But it is not core: the observation is rat brain immuno-electron microscopy of a broadly expressed trafficking protein, ARFGEF2 is expressed in placenta, lung, heart, kidney and pancreas as well as brain, and the synaptic presence has never been given a synapse-specific mechanism. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q7TSU1 · Arfgef2 (rat BIG2) SUPPORTS TRANSFER Rat IDA from PMID:15198677, which annotates only 2 entities in all of GOA, so it is per-gene curation rather than a distributed projection. Transfers safely but is cell-type-specific. Supporting Evidence: PMID:15198677 BIG2 is present in both inhibitory GABAergic synapses that contain GABA(A)Rs and in asymmetric excitatory synapses. PMID:15198677 Double label immunofluorescence of cultured hippocampal neurons and electron microscopy immunocytochemistry of rat brain tissue show that BIG2 concentrates in the trans-Golgi network. file:human/ARFGEF2/ARFGEF2-bioinformatics/RESULTS.md `PMID:15198677` annotates only **2 entities** in the whole of GOA — rat Arfgef2 |
| GO:0032280 symmetric synapse | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: Localisation of BIG2 to symmetric (inhibitory) synapses, transferred from rat Arfgef2 by Ensembl Compara. The rat evidence is immuno-electron microscopy of rat brain tissue in a study of GABA(A) receptor trafficking. Reason: Part of a block of 20 rows, covering 13 GO terms, that Ensembl Compara and UniProt ISS both project from rat Arfgef2; 17 of those rows rest solely on one paper, PMID:15198677. Two things are worth separating. The projection discriminator comes back NEGATIVE: that reference annotates only 2 entities in all of GOA - rat Arfgef2 and its partner Gabrb3 - so this is genuine per-gene curation at figure granularity by SynGO and UniProt, not a complex-membership phenotype distributed across subunits. The annotation is therefore kept. But it is not core: the observation is rat brain immuno-electron microscopy of a broadly expressed trafficking protein, ARFGEF2 is expressed in placenta, lung, heart, kidney and pancreas as well as brain, and the synaptic presence has never been given a synapse-specific mechanism. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q7TSU1 · Arfgef2 (rat BIG2) SUPPORTS TRANSFER Rat IDA from PMID:15198677, which annotates only 2 entities in all of GOA, so it is per-gene curation rather than a distributed projection. Transfers safely but is cell-type-specific. Supporting Evidence: PMID:15198677 BIG2 is present in both inhibitory GABAergic synapses that contain GABA(A)Rs and in asymmetric excitatory synapses. PMID:15198677 Double label immunofluorescence of cultured hippocampal neurons and electron microscopy immunocytochemistry of rat brain tissue show that BIG2 concentrates in the trans-Golgi network. file:human/ARFGEF2/ARFGEF2-bioinformatics/RESULTS.md `PMID:15198677` annotates only **2 entities** in the whole of GOA — rat Arfgef2 |
| GO:0032280 symmetric synapse | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: Localisation of BIG2 to symmetric (inhibitory) synapses, transferred from rat Arfgef2 by UniProt curator-judged ISS. The rat evidence is immuno-electron microscopy of rat brain tissue in a study of GABA(A) receptor trafficking. Reason: Part of a block of 20 rows, covering 13 GO terms, that Ensembl Compara and UniProt ISS both project from rat Arfgef2; 17 of those rows rest solely on one paper, PMID:15198677. Two things are worth separating. The projection discriminator comes back NEGATIVE: that reference annotates only 2 entities in all of GOA - rat Arfgef2 and its partner Gabrb3 - so this is genuine per-gene curation at figure granularity by SynGO and UniProt, not a complex-membership phenotype distributed across subunits. The annotation is therefore kept. But it is not core: the observation is rat brain immuno-electron microscopy of a broadly expressed trafficking protein, ARFGEF2 is expressed in placenta, lung, heart, kidney and pancreas as well as brain, and the synaptic presence has never been given a synapse-specific mechanism. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q7TSU1 · Arfgef2 (rat BIG2) SUPPORTS TRANSFER Rat IDA from PMID:15198677, which annotates only 2 entities in all of GOA, so it is per-gene curation rather than a distributed projection. Transfers safely but is cell-type-specific. Supporting Evidence: PMID:15198677 BIG2 is present in both inhibitory GABAergic synapses that contain GABA(A)Rs and in asymmetric excitatory synapses. PMID:15198677 Double label immunofluorescence of cultured hippocampal neurons and electron microscopy immunocytochemistry of rat brain tissue show that BIG2 concentrates in the trans-Golgi network. file:human/ARFGEF2/ARFGEF2-bioinformatics/RESULTS.md `PMID:15198677` annotates only **2 entities** in the whole of GOA — rat Arfgef2 |
| GO:0032760 positive regulation of tumor necrosis factor production | IMP PMID:17276987 The brefeldin A-inhibited guanine nucleotide-exchange protei... | MODIFY | Summary: Positive regulation of tumour necrosis factor production, from a paper that measures release of the TNF *receptor* TNFR1 in exosome-like vesicles and never measures production of the TNF cytokine. A receptor/ligand conflation. Reason: GO:0032760 is about the appearance of tumour necrosis factor itself - its parent GO:0032680 regulates TNF production. What this paper shows is that BIG2 knockdown reduces constitutive release of full-length TNFR1 in exosome-like vesicles from human endothelial cells, and explicitly not the proteolytic route: neither BIG2 nor BIG1 was required for IL-1beta-induced cleavage of TNFR1 ectodomains. The direction is doubtful too: the paper frames released TNFR1 as a soluble species that binds and modulates TNF bioactivity, i.e. a decoy receptor, so if anything more shed TNFR1 antagonises TNF rather than increasing its production. Querying GOA by this reference returns exactly 2 annotations on 1 entity, both on ARFGEF2, so this is one term choice on one gene, not a propagated error. GO:1903543 positive regulation of exosomal secretion is what the experiment measures. Proposed replacements: positive regulation of exosomal secretion Supporting Evidence: PMID:17276987 Effects of specific RNA interference (RNAi) showed that BIG2, but not BIG1, regulated the release of TNFR1 exosome-like vesicles, whereas neither BIG2 nor BIG1 was required for the IL-1beta-induced proteolytic cleavage of TNFR1 ectodomains. PMID:17276987 We conclude that the association between BIG2 and TNFR1 selectively regulates the extracellular release of TNFR1 exosome-like vesicles from human vascular endothelial cells via an ARF1- and ARF3-dependent mechanism. PMID:17276987 The type I, 55-kDa tumor necrosis factor receptor (TNFR1) is released from cells to the extracellular space where it can bind and modulate TNF bioactivity. |
| GO:0034237 protein kinase A regulatory subunit binding | IDA PMID:12571360 Protein kinase A-anchoring (AKAP) domains in brefeldin A-inh... | ACCEPT | Summary: BIG2 is an A-kinase anchoring protein: three short amphipathic regions in its N-terminal half bind the RI and RII regulatory subunits of PKA, confirmed by co-immunoprecipitation of endogenous proteins from human HepG2 cytosol. Reason: A second, genuinely distinct molecular function of this protein, mapped to residues and verified on endogenous proteins. Note that the three AKAP regions (27-48, 284-301, 517-538) lie in the DCB/HUS half of the molecule, well N-terminal of the Sec7 domain at 654-785, so the anchoring function is not a property of the catalytic module. Core. Supporting Evidence: PMID:12571360 Residues 27-48 (domain A) interacted with RI alpha and RI beta, 284-301 (domain B) interacted with RII alpha and RII beta, and 517-538 (domain C) interacted with RI alpha, RII alpha, and RII beta. PMID:12571360 Coimmunoprecipitation experiments confirmed interaction of in vitro translated BIG2 and RI alpha, as well as of the endogenous proteins in cytosol of cultured HepG2 cells. |
| GO:0034237 protein kinase A regulatory subunit binding | IEA GO_REF:0000117 | ACCEPT | Summary: The same PKA regulatory subunit binding, from an ARBA machine-learning rule. Reason: An automated call that coincides with a residue-mapped IDA on the same protein. Supporting Evidence: PMID:12571360 Residues 27-48 (domain A) interacted with RI alpha and RI beta, 284-301 (domain B) interacted with RII alpha and RII beta, and 517-538 (domain C) interacted with RI alpha, RII alpha, and RII beta. |
| GO:0035556 intracellular signal transduction | IDA PMID:12571360 Protein kinase A-anchoring (AKAP) domains in brefeldin A-inh... | MARK AS OVER ANNOTATED | Summary: Intracellular signal transduction, taken from the AKAP-domain-mapping paper. Reason: What that paper measures is binding - three PKA regulatory-subunit interaction regions mapped with 28 deletion mutants - plus cAMP-induced redistribution of BIG2 from cytosol to membranes. It does not show BIG2 transducing a signal; the authors' own conclusion is hedged as a role that 'could coordinate cAMP and ARF regulatory pathways'. The term is also close to contentless: almost any cytoplasmic regulator is under it. Everything informative that this paper establishes is already carried by GO:0034237 and the GO:0016020/GO:0000139 membrane rows from the same reference. Flagged rather than removed because BIG2 genuinely sits in a cAMP signalling module. Supporting Evidence: PMID:12571360 All findings are consistent with a role for BIG2 as an A kinase-anchoring protein (or AKAP) that could coordinate cAMP and ARF regulatory pathways. PMID:12571360 Residues 27-48 (domain A) interacted with RI alpha and RI beta, 284-301 (domain B) interacted with RII alpha and RII beta, and 517-538 (domain C) interacted with RI alpha, RII alpha, and RII beta. PMID:12571360 Using 28 deletion mutants, we found three regions of BIG2 that interacted with R subunits of PKA. |
| GO:0043197 dendritic spine | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: Localisation of BIG2 to dendritic spines, transferred from rat Arfgef2 by Ensembl Compara. The rat evidence is immuno-electron microscopy of rat brain tissue in a study of GABA(A) receptor trafficking. Reason: Part of a block of 20 rows, covering 13 GO terms, that Ensembl Compara and UniProt ISS both project from rat Arfgef2; 17 of those rows rest solely on one paper, PMID:15198677. Two things are worth separating. The projection discriminator comes back NEGATIVE: that reference annotates only 2 entities in all of GOA - rat Arfgef2 and its partner Gabrb3 - so this is genuine per-gene curation at figure granularity by SynGO and UniProt, not a complex-membership phenotype distributed across subunits. The annotation is therefore kept. But it is not core: the observation is rat brain immuno-electron microscopy of a broadly expressed trafficking protein, ARFGEF2 is expressed in placenta, lung, heart, kidney and pancreas as well as brain, and the synaptic presence has never been given a synapse-specific mechanism. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q7TSU1 · Arfgef2 (rat BIG2) SUPPORTS TRANSFER Rat IDA from PMID:15198677, which annotates only 2 entities in all of GOA, so it is per-gene curation rather than a distributed projection. Transfers safely but is cell-type-specific. Supporting Evidence: PMID:15198677 BIG2 is present in both inhibitory GABAergic synapses that contain GABA(A)Rs and in asymmetric excitatory synapses. PMID:15198677 Double label immunofluorescence of cultured hippocampal neurons and electron microscopy immunocytochemistry of rat brain tissue show that BIG2 concentrates in the trans-Golgi network. file:human/ARFGEF2/ARFGEF2-bioinformatics/RESULTS.md `PMID:15198677` annotates only **2 entities** in the whole of GOA — rat Arfgef2 |
| GO:0043197 dendritic spine | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: Localisation of BIG2 to dendritic spines, transferred from rat Arfgef2 by UniProt curator-judged ISS. The rat evidence is immuno-electron microscopy of rat brain tissue in a study of GABA(A) receptor trafficking. Reason: Part of a block of 20 rows, covering 13 GO terms, that Ensembl Compara and UniProt ISS both project from rat Arfgef2; 17 of those rows rest solely on one paper, PMID:15198677. Two things are worth separating. The projection discriminator comes back NEGATIVE: that reference annotates only 2 entities in all of GOA - rat Arfgef2 and its partner Gabrb3 - so this is genuine per-gene curation at figure granularity by SynGO and UniProt, not a complex-membership phenotype distributed across subunits. The annotation is therefore kept. But it is not core: the observation is rat brain immuno-electron microscopy of a broadly expressed trafficking protein, ARFGEF2 is expressed in placenta, lung, heart, kidney and pancreas as well as brain, and the synaptic presence has never been given a synapse-specific mechanism. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q7TSU1 · Arfgef2 (rat BIG2) SUPPORTS TRANSFER Rat IDA from PMID:15198677, which annotates only 2 entities in all of GOA, so it is per-gene curation rather than a distributed projection. Transfers safely but is cell-type-specific. Supporting Evidence: PMID:15198677 BIG2 is present in both inhibitory GABAergic synapses that contain GABA(A)Rs and in asymmetric excitatory synapses. PMID:15198677 Double label immunofluorescence of cultured hippocampal neurons and electron microscopy immunocytochemistry of rat brain tissue show that BIG2 concentrates in the trans-Golgi network. file:human/ARFGEF2/ARFGEF2-bioinformatics/RESULTS.md `PMID:15198677` annotates only **2 entities** in the whole of GOA — rat Arfgef2 |
| GO:0045202 synapse | IEA GO_REF:0000044 | KEEP AS NON CORE | Summary: Synapse, from the UniProt subcellular-location vocabulary; the general parent of the four SynGO-derived rat synaptic terms. Reason: Same basis as the specific synaptic terms - rat brain immuno-EM in a GABA(A) receptor study. Correct at this level and worth keeping as the general statement, but cell-type-specific rather than core. Supporting Evidence: PMID:15198677 BIG2 is present in both inhibitory GABAergic synapses that contain GABA(A)Rs and in asymmetric excitatory synapses. |
| GO:0048471 perinuclear region of cytoplasm | IEA GO_REF:0000044 | ACCEPT | Summary: Perinuclear region of cytoplasm, from the UniProt subcellular-location vocabulary. Directly observed: BIG2 concentrates perinuclearly with Golgi markers, and BIG2 depletion drives transferrin receptor into perinuclear recycling endosomes. Reason: Better supported than a typical vocabulary-derived row - the perinuclear concentration is where the TGN and recycling-endosome pools of BIG2 sit and is reported in several independent human-cell studies. Supporting Evidence: PMID:10716990 endogenous BIG1 and coexpressed BIG2 were distributed in a punctate pattern throughout the cytosol, and also concentrated in the perinuclear region, where endogenous BIG1 and BIG2 each partially colocalized with Golgi-specific 58K protein and gamma-adaptin PMID:16477018 Treatment of cells with BIG2 small interfering RNA (siRNA), but not BIG1 or control siRNAs, decreased BIG2 protein >90% without affecting BIG1, ARF, or actin content, whereas TfnR was significantly increased as was its accumulation in perinuclear recycling endosomes. |
| GO:0050811 GABA receptor binding | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: GABA receptor binding, transferred from rat Arfgef2 by Ensembl Compara. The rat row is an IMP from the paper that identified the BIG2-GABA(A) receptor beta subunit interaction. Reason: Well founded despite being a transfer. The rat evidence is a yeast two-hybrid interaction between BIG2's C-terminal 110 residues and the large intracellular loop of the GABA(A) receptor beta subunits, plus reciprocal co-immunoprecipitation of the native proteins from brain. Critically, the functional experiment in the same paper - that BIG2 promotes GABA(A) receptor exit from the ER - was done in human HEK293 cells, so the human relevance does not depend on the projection. It is non-core: this is one cargo among many for a general trafficking GEF, and GO:0050811 describes a cargo-recognition contact rather than the protein's catalytic function. UniProt records the corresponding human interactions with GABRB1, GABRB2 and GABRB3 as By similarity. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q7TSU1 · Arfgef2 (rat BIG2) SUPPORTS TRANSFER Rat IMP from PMID:15198677. The binding itself was shown by Y2H and brain co-IP, and the functional assay in the same paper used human HEK293 cells. Supporting Evidence: PMID:15198677 The interaction between the 110 amino acid C-terminal fragment of BIG2 and the large intracellular loop of the GABA(A)R beta subunits was revealed with a yeast two-hybrid assay. PMID:15198677 In transfected human embryonic kidney cell line 293 cells, BIG2 promotes the exit of GABA(A)Rs from endoplasmic reticulum. |
| GO:0050811 GABA receptor binding | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: The same GABA receptor binding call arriving again by the UniProt ISS route. Reason: Well founded despite being a transfer. The rat evidence is a yeast two-hybrid interaction between BIG2's C-terminal 110 residues and the large intracellular loop of the GABA(A) receptor beta subunits, plus reciprocal co-immunoprecipitation of the native proteins from brain. Critically, the functional experiment in the same paper - that BIG2 promotes GABA(A) receptor exit from the ER - was done in human HEK293 cells, so the human relevance does not depend on the projection. It is non-core: this is one cargo among many for a general trafficking GEF, and GO:0050811 describes a cargo-recognition contact rather than the protein's catalytic function. UniProt records the corresponding human interactions with GABRB1, GABRB2 and GABRB3 as By similarity. This is the ISS half of the duplicate projection; the Compara row asserts exactly the same thing from the same rat datum. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q7TSU1 · Arfgef2 (rat BIG2) SUPPORTS TRANSFER Same rat IMP as the Compara row; duplicate pipeline. Supporting Evidence: PMID:15198677 The interaction between the 110 amino acid C-terminal fragment of BIG2 and the large intracellular loop of the GABA(A)R beta subunits was revealed with a yeast two-hybrid assay. file:human/ARFGEF2/ARFGEF2-bioinformatics/RESULTS.md **20 of the 66 rows**, covering **13 distinct GO terms**, carry rat Arfgef2 |
| GO:0055037 recycling endosome | IDA PMID:15385626 BIG2, a guanine nucleotide exchange factor for ADP-ribosylat... | ACCEPT | Summary: BIG2 on recycling endosomes, shown directly and reinforced by the catalytically dead E738K mutant selectively tubulating that compartment. Reason: The location that distinguishes BIG2 from BIG1 and the site of its non-redundant function. Core, and independently reproduced by a different group using siRNA. Supporting Evidence: PMID:15385626 we have revealed that another population of BIG2 is associated with the recycling endosome and found that expression of a catalytically inactive BIG2 mutant, E738K, selectively induces membrane tubules from this compartment PMID:20360857 Our results show that depletion of BIG2 but not of BIG1 induces a tubulation of the recycling endosomal compartment, consistent with a specific role for BIG2 here. |
| GO:0098793 presynapse | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: Localisation of BIG2 to the presynaptic compartment, transferred from rat Arfgef2 by Ensembl Compara. The rat evidence is immuno-electron microscopy of rat brain tissue in a study of GABA(A) receptor trafficking. Reason: Part of a block of 20 rows, covering 13 GO terms, that Ensembl Compara and UniProt ISS both project from rat Arfgef2; 17 of those rows rest solely on one paper, PMID:15198677. Two things are worth separating. The projection discriminator comes back NEGATIVE: that reference annotates only 2 entities in all of GOA - rat Arfgef2 and its partner Gabrb3 - so this is genuine per-gene curation at figure granularity by SynGO and UniProt, not a complex-membership phenotype distributed across subunits. The annotation is therefore kept. But it is not core: the observation is rat brain immuno-electron microscopy of a broadly expressed trafficking protein, ARFGEF2 is expressed in placenta, lung, heart, kidney and pancreas as well as brain, and the synaptic presence has never been given a synapse-specific mechanism. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q7TSU1 · Arfgef2 (rat BIG2) SUPPORTS TRANSFER Rat IDA from PMID:15198677, which annotates only 2 entities in all of GOA, so it is per-gene curation rather than a distributed projection. Transfers safely but is cell-type-specific. Supporting Evidence: PMID:15198677 BIG2 is present in both inhibitory GABAergic synapses that contain GABA(A)Rs and in asymmetric excitatory synapses. PMID:15198677 Double label immunofluorescence of cultured hippocampal neurons and electron microscopy immunocytochemistry of rat brain tissue show that BIG2 concentrates in the trans-Golgi network. file:human/ARFGEF2/ARFGEF2-bioinformatics/RESULTS.md `PMID:15198677` annotates only **2 entities** in the whole of GOA — rat Arfgef2 |
| GO:0098794 postsynapse | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: Localisation of BIG2 to the postsynaptic compartment, transferred from rat Arfgef2 by Ensembl Compara. The rat evidence is immuno-electron microscopy of rat brain tissue in a study of GABA(A) receptor trafficking. Reason: Part of a block of 20 rows, covering 13 GO terms, that Ensembl Compara and UniProt ISS both project from rat Arfgef2; 17 of those rows rest solely on one paper, PMID:15198677. Two things are worth separating. The projection discriminator comes back NEGATIVE: that reference annotates only 2 entities in all of GOA - rat Arfgef2 and its partner Gabrb3 - so this is genuine per-gene curation at figure granularity by SynGO and UniProt, not a complex-membership phenotype distributed across subunits. The annotation is therefore kept. But it is not core: the observation is rat brain immuno-electron microscopy of a broadly expressed trafficking protein, ARFGEF2 is expressed in placenta, lung, heart, kidney and pancreas as well as brain, and the synaptic presence has never been given a synapse-specific mechanism. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q7TSU1 · Arfgef2 (rat BIG2) SUPPORTS TRANSFER Rat IDA from PMID:15198677, which annotates only 2 entities in all of GOA, so it is per-gene curation rather than a distributed projection. Transfers safely but is cell-type-specific. Supporting Evidence: PMID:15198677 BIG2 is present in both inhibitory GABAergic synapses that contain GABA(A)Rs and in asymmetric excitatory synapses. PMID:15198677 Double label immunofluorescence of cultured hippocampal neurons and electron microscopy immunocytochemistry of rat brain tissue show that BIG2 concentrates in the trans-Golgi network. file:human/ARFGEF2/ARFGEF2-bioinformatics/RESULTS.md `PMID:15198677` annotates only **2 entities** in the whole of GOA — rat Arfgef2 |
| GO:0098978 glutamatergic synapse | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: Localisation of BIG2 to glutamatergic synapses, transferred from rat Arfgef2 by Ensembl Compara. The rat evidence is immuno-electron microscopy of rat brain tissue in a study of GABA(A) receptor trafficking. Reason: Part of a block of 20 rows, covering 13 GO terms, that Ensembl Compara and UniProt ISS both project from rat Arfgef2; 17 of those rows rest solely on one paper, PMID:15198677. Two things are worth separating. The projection discriminator comes back NEGATIVE: that reference annotates only 2 entities in all of GOA - rat Arfgef2 and its partner Gabrb3 - so this is genuine per-gene curation at figure granularity by SynGO and UniProt, not a complex-membership phenotype distributed across subunits. The annotation is therefore kept. But it is not core: the observation is rat brain immuno-electron microscopy of a broadly expressed trafficking protein, ARFGEF2 is expressed in placenta, lung, heart, kidney and pancreas as well as brain, and the synaptic presence has never been given a synapse-specific mechanism. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q7TSU1 · Arfgef2 (rat BIG2) SUPPORTS TRANSFER Rat IDA from PMID:15198677, which annotates only 2 entities in all of GOA, so it is per-gene curation rather than a distributed projection. Transfers safely but is cell-type-specific. Supporting Evidence: PMID:15198677 BIG2 is present in both inhibitory GABAergic synapses that contain GABA(A)Rs and in asymmetric excitatory synapses. PMID:15198677 Double label immunofluorescence of cultured hippocampal neurons and electron microscopy immunocytochemistry of rat brain tissue show that BIG2 concentrates in the trans-Golgi network. file:human/ARFGEF2/ARFGEF2-bioinformatics/RESULTS.md `PMID:15198677` annotates only **2 entities** in the whole of GOA — rat Arfgef2 |
| GO:0098982 GABA-ergic synapse | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: Localisation of BIG2 to GABAergic synapses, transferred from rat Arfgef2 by Ensembl Compara. The rat evidence is immuno-electron microscopy of rat brain tissue in a study of GABA(A) receptor trafficking. Reason: Part of a block of 20 rows, covering 13 GO terms, that Ensembl Compara and UniProt ISS both project from rat Arfgef2; 17 of those rows rest solely on one paper, PMID:15198677. Two things are worth separating. The projection discriminator comes back NEGATIVE: that reference annotates only 2 entities in all of GOA - rat Arfgef2 and its partner Gabrb3 - so this is genuine per-gene curation at figure granularity by SynGO and UniProt, not a complex-membership phenotype distributed across subunits. The annotation is therefore kept. But it is not core: the observation is rat brain immuno-electron microscopy of a broadly expressed trafficking protein, ARFGEF2 is expressed in placenta, lung, heart, kidney and pancreas as well as brain, and the synaptic presence has never been given a synapse-specific mechanism. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q7TSU1 · Arfgef2 (rat BIG2) SUPPORTS TRANSFER Rat IDA from PMID:15198677, which annotates only 2 entities in all of GOA, so it is per-gene curation rather than a distributed projection. Transfers safely but is cell-type-specific. Supporting Evidence: PMID:15198677 BIG2 is present in both inhibitory GABAergic synapses that contain GABA(A)Rs and in asymmetric excitatory synapses. PMID:15198677 Double label immunofluorescence of cultured hippocampal neurons and electron microscopy immunocytochemistry of rat brain tissue show that BIG2 concentrates in the trans-Golgi network. file:human/ARFGEF2/ARFGEF2-bioinformatics/RESULTS.md `PMID:15198677` annotates only **2 entities** in the whole of GOA — rat Arfgef2 |
| GO:0051497 negative regulation of stress fiber assembly | IMP PMID:23918382 Arf guanine nucleotide-exchange factors BIG1 and BIG2 regula... | NEW | Summary: siRNA depletion of BIG2 in human HeLa cells more than doubles phosphorylation of myosin regulatory light chain at T18/S19 and makes actin stress fibres significantly more prominent; re-expressing BIG2, or its C-terminal fragment alone, reverses both effects. BIG2 therefore restrains stress fibre assembly. Reason: A well-controlled human loss-of-function result with a rescue, currently carrying no GO annotation of any kind - querying GOA by this reference returns zero annotations on any gene. The experiments are in HeLa cells with human-targeted siRNA, so IMP is the correct evidence code. The mechanism is non-catalytic: the rescuing C-terminal fragment does not contain the Sec7 domain, so this is a scaffolding function separable from ARF exchange. This paper was surfaced by the affinage deep-research record, whose dated-findings table summarises it accurately; the claim here nonetheless rests on the primary full text, which was read, and the provider sentence is cited only as corroboration. Supporting Evidence: PMID:23918382 After BIG1 or BIG2 depletion ( Fig. 2 A ), phosphorylation of RLC T18 and S19 was more than twice that in cells treated with the NT siRNA or vehicle alone. PMID:23918382 Stress fibers were significantly more prominent after BIG1 or BIG2 depletion than in control cells PMID:23918382 Motility of HeLa cells transfected with BIG1 or BIG2 siRNA was significantly impaired relative to that of cells transfected with NT siRNA, and overexpression of BIG1 or BIG2, full-length or C fragment, but not N or S fragments, reversed those effects PMID:23918382 Reciprocal coimmunoprecipitation of endogenous HeLa cell BIG1 and BIG2 with myosin IIA was demonstrably independent of Arf guanine nucleotide-exchange factor activity file:human/ARFGEF2/ARFGEF2-deep-research-affinage.md depletion of BIG2 (or BIG1) enhances phosphorylation of myosin regulatory light chain (T18/S19) and increases F-actin content, impairing cell migration |
| GO:0030335 positive regulation of cell migration | IMP PMID:23918382 Arf guanine nucleotide-exchange factors BIG1 and BIG2 regula... | NEW | Summary: BIG2 depletion significantly impairs HeLa cell motility in Transwell assays, and the defect is reversed by re-expressing BIG2 or its C-terminal fragment. A second, independent study reaches the same conclusion through integrin beta1 recycling and actin remodelling. Reason: Two independent human HeLa studies with consistent loss-of-function phenotypes, one with a rescue control, and neither has produced a single GO annotation anywhere. Both used human cells and human-targeted siRNA, so IMP is correct. This links BIG2's recycling function to a cell-level output and is the process most plausibly connecting BIG2 to the neuronal migration failure seen in periventricular nodular heterotopia, although that connection is not itself demonstrated and is deliberately not annotated. Supporting Evidence: PMID:23918382 Motility of HeLa cells transfected with BIG1 or BIG2 siRNA was significantly impaired relative to that of cells transfected with NT siRNA, and overexpression of BIG1 or BIG2, full-length or C fragment, but not N or S fragments, reversed those effects PMID:22908276 Treatment of HeLa cells with BIG2 siRNA resulted in perinuclear accumulation of integrin β1 and its delayed return to the cell surface. Motility of BIG2-depleted cells was simultaneously decreased, as were actin-based membrane protrusions |
Loading supporting content…
Download this section (compressed HTML)Q: Should the rat Arfgef2 GO:0005879 axonemal microtubule annotation (IDA, PMID:15198677, assigned by UniProt) be re-examined? The cited paper is a study of GABA(A) receptor trafficking in hippocampal neurons and rat brain with no ciliary content in its abstract, and the same UniProt record's own prose subcellular location says only 'Cytoplasm, cytoskeleton'. The rat row is the sole origin of two human rows, via Ensembl Compara and UniProt ISS. I could not read the full text (subscription only) and am therefore flagging rather than contradicting it.
Q: Should human ARFGEF2's GO:0017022 myosin binding row be re-referenced? It is currently a Compara projection of a rat row made from PMID:15644318, a paper about the paralog BIG1 that put six annotations on human ARFGEF1 and none on human ARFGEF2. Direct human evidence for BIG2 exists in PMID:23918382 and would replace a paralog-derived chain with a species- and gene-correct one.
Q: Was BIG2's role in holding MYCBP/AMY-1 at the trans-Golgi network (PMID:16866877, RNAi) performed in human cells? If so, GO:0034067 protein localization to Golgi apparatus is annotatable by IMP. I did not add it because the cached abstract does not name the cell system and the full text was unavailable, so I could not state the species the experiment was done in.
Q: Human ARFGEF1 carries GO:0005096 GTPase activator activity by IDA from PMID:15644318, a paper whose result is that BIG1 *inhibits* the RhoGAP activity of myosin IXb - and the same paper also gave ARFGEF1 GO:0034260 negative regulation of GTPase activity, which reads the opposite way. This is ARFGEF1's row rather than this gene's and is offered as a question, not a finding; it was noticed while tracing the myosin-binding chain and has not been checked against the full text.
Q: Twenty-six of the 43 papers in ARFGEF2's own literature produce no GO annotation anywhere, including the founding disease paper (PMID:14647276), the AP-1/GGA dominant-negative papers (PMID:11777925, PMID:12051703), the PP1-gamma and RII-beta AKAP papers, the DCB/HUS homodimer paper and the myosin-phosphatase paper. Is this gene a candidate for a targeted curation pass rather than further pruning?
Experiment: Test directly whether BIG2 is ever ciliary. Image endogenous BIG2 in ciliated human cells (RPE-1 serum-starved, or multiciliated ependymal cultures) against acetylated tubulin and ARL13B, and score axonemal versus basal-body signal. This is the cheapest way to settle the GO:0005879 rows in either direction, and a positive result would be interesting given the centrosomal pool already reported. The census run for this review adds a concrete reason to do it: EXOC7/Exo70, the exocyst subunit that BIG2 binds and co-localises with at the MTOC, itself carries GO:0036064 ciliary basal body, so BIG2 has a documented partner at the ciliary base even though BIG2 has never been looked for there.
Experiment: Separate the catalytic and scaffolding functions in neurons. Compare wild-type BIG2, the Sec7-dead E738K mutant, and a C-terminal fragment lacking Sec7 for rescue of neuronal migration and of myosin light-chain phosphorylation in an ARFGEF2-null background. The myosin-IIA work shows the C-terminus suffices for the cytoskeletal phenotype in HeLa cells; whether the migration failure in PVNH2 is a GEF defect or a scaffolding defect is unresolved.
Experiment: Measure which cargoes actually depend on BIG2 at the TGN in a human neural progenitor system, using degron-mediated acute depletion plus surface proteomics, rather than brefeldin A. Every current cargo claim (E-cadherin, beta-catenin, Filamin A) rests on brefeldin A or dominant-negative overexpression, neither of which distinguishes BIG2 from BIG1 or GBF1.
Experiment: Test whether the released TNFR1 species is functionally agonistic or antagonistic toward TNF signalling after BIG2 depletion. The existing GO annotation asserts positive regulation of TNF production, which the paper does not measure; a simple TNF bioactivity assay on conditioned medium would establish the sign.
Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)