{"gene":"ARHGAP31","run_date":"2026-06-09T22:02:44","timeline":{"discoveries":[{"year":1998,"finding":"CdGAP (ARHGAP31) was identified as a novel GTPase-activating protein with in vitro GAP activity toward both Cdc42 and Rac1 but not RhoA. Microinjection of CdGAP into serum-starved fibroblasts inhibited PDGF-induced lamellipodia (Rac-mediated) and bradykinin-induced filopodia (Cdc42-mediated), but had no effect on LPA-induced stress fiber formation. The C-terminus contains potential PKC phosphorylation sites and five SH3 binding motifs.","method":"Yeast two-hybrid screen, in vitro GAP assay, microinjection into fibroblasts","journal":"The Journal of biological chemistry","confidence":"High","confidence_rationale":"Tier 1 / Strong — in vitro GAP assay with substrate specificity established, functional microinjection with specific phenotypic readouts, foundational characterization paper","pmids":["9786927"],"is_preprint":false},{"year":2001,"finding":"The endocytic protein intersectin interacts with CdGAP through a subset of its SH3 domains and inhibits CdGAP's GAP activity toward Rac1. In PDGF-stimulated Swiss 3T3 cells, intersectin co-localizes with CdGAP. The central domain of CdGAP (not the proline-rich domain) is required for the CdGAP–intersectin interaction, but the C-terminal proline-rich domain is required for intersectin-mediated inhibition of GAP activity, suggesting a conformational change mechanism.","method":"Co-immunoprecipitation, in vitro GAP assay with intersectin-SH3 domains, co-localization in PDGF-stimulated cells, deletion mutant analysis","journal":"The Journal of biological chemistry","confidence":"High","confidence_rationale":"Tier 2 / Strong — reciprocal interaction mapping, in vitro GAP inhibition assay, domain dissection with multiple mutants, in vivo co-localization","pmids":["11744688"],"is_preprint":false},{"year":2005,"finding":"CdGAP is phosphorylated in vivo on serine and threonine residues downstream of the MEK-ERK pathway in response to serum or PDGF. ERK1 and RSK-1 phosphorylate CdGAP in vitro. A DEF domain (docking for ERK, FXFP motif) in the proline-rich region is required for efficient ERK1/2 binding and phosphorylation. Thr776 was identified as an in vivo ERK1/2 target site that serves as an important regulatory site of CdGAP GAP activity.","method":"In vitro kinase assay, site-directed mutagenesis, phosphorylation mapping, co-immunoprecipitation, cell stimulation with serum/PDGF","journal":"Molecular and cellular biology","confidence":"High","confidence_rationale":"Tier 1 / Moderate — in vitro kinase assay with mutagenesis, in vivo phosphorylation confirmed, regulatory site functionally characterized","pmids":["16024771"],"is_preprint":false},{"year":2006,"finding":"CdGAP localizes to focal adhesions through a direct interaction with the N-terminus of actopaxin (α-parvin), a paxillin and actin binding protein. CdGAP GAP activity is regulated in an adhesion-dependent manner. Both overexpression and RNAi knockdown of CdGAP impaired normal cell spreading, polarized lamellipodia formation, and cell migration. An actopaxin mutant defective for CdGAP binding attenuated these effects.","method":"Co-immunoprecipitation/pulldown, RNAi knockdown, overexpression of GAP-deficient mutant, fluorescence localization, cell spreading/migration assays","journal":"Current biology : CB","confidence":"High","confidence_rationale":"Tier 2 / Moderate — direct binding demonstrated, loss-of-function with defined cellular phenotypes, mutant rescue experiments, focal adhesion localization linked to function","pmids":["16860736"],"is_preprint":false},{"year":2006,"finding":"Human CdGAP (KIAA1204) shares 76% sequence identity with mouse CdGAP-long isoform and retains in vitro and in vivo GAP activity toward Cdc42 and Rac1 but not RhoA, confirming substrate specificity is conserved. Human CdGAP is phosphorylated in vivo on serine and threonine residues. Unlike mouse CdGAP, human CdGAP interacts with ERK1/2 through a region not involving a canonical DEF domain. CdGAP expression caused membrane blebbing in COS-7 cells.","method":"In vitro GAP assay, in vivo phosphorylation analysis, co-immunoprecipitation with ERK1/2, cell morphology assay","journal":"Biology of the cell","confidence":"Medium","confidence_rationale":"Tier 2 / Moderate — GAP activity confirmed in vitro and in vivo, interaction with ERK1/2 by co-IP, single lab study","pmids":["16519628"],"is_preprint":false},{"year":2006,"finding":"GSK-3α was identified as a CdGAP binding partner. GSK-3α and GSK-3β both interact with CdGAP in mammalian cells and phosphorylate CdGAP in vitro and in vivo at Thr776, the same site previously identified as an ERK1/2 target. GSK-3 activity is necessary for serum-stimulated upregulation of CdGAP protein levels (but not mRNA), revealing a novel mechanism for controlling Cdc42/Rac1 signaling.","method":"Yeast two-hybrid, co-immunoprecipitation, in vitro kinase assay, in vivo phosphorylation, GSK-3 inhibitor treatment","journal":"The Journal of biological chemistry","confidence":"High","confidence_rationale":"Tier 1 / Moderate — yeast two-hybrid confirmed by co-IP, in vitro and in vivo phosphorylation at specific site, pharmacological inhibitor validates functional role","pmids":["17158447"],"is_preprint":false},{"year":2010,"finding":"CdGAP is required for TGF-β-induced cell motility and invasion of Neu/ErbB-2-expressing mammary tumor cells. The proline-rich domain (PRD) but not the GAP domain of CdGAP is essential for TGF-β-mediated motility and invasion. CdGAP depletion increased E-cadherin expression and prevented complete E-cadherin loss during TGF-β-induced EMT. TGF-β also induces CdGAP expression and phosphorylation in NMuMG cells.","method":"siRNA knockdown, overexpression of deletion mutants (rescue analysis), cell motility/invasion assays, immunoblotting for E-cadherin","journal":"Oncogene","confidence":"Medium","confidence_rationale":"Tier 2 / Moderate — siRNA knockdown with defined phenotype, domain-specific rescue experiments, single lab","pmids":["21042277"],"is_preprint":false},{"year":2011,"finding":"Gain-of-function truncating mutations in the terminal exon of ARHGAP31 cause Adams-Oliver syndrome. Mutant transcripts are stable and increase ARHGAP31 GAP activity in vitro. Constitutively active ARHGAP31 mutations result in loss of available active Cdc42 and disruption of actin cytoskeletal structures. Mouse Arhgap31 expression is restricted to terminal limb buds and craniofacial processes during early development.","method":"In vitro GAP activity assay of mutant proteins, genome-wide linkage analysis, exome sequencing, Rho GTPase activity assay (active Cdc42 measurement), mouse expression analysis","journal":"American journal of human genetics","confidence":"High","confidence_rationale":"Tier 1 / Strong — gain-of-function mechanism confirmed by in vitro GAP assay, loss of active Cdc42 measured in cells, replicated in multiple independent families","pmids":["21565291"],"is_preprint":false},{"year":2011,"finding":"The interaction between intersectin's SH3D domain and CdGAP is mediated through a novel basic-rich motif in the central domain of CdGAP containing a conserved xKx(K/R)K sequence, not a canonical proline-rich SH3-binding motif.","method":"Pulldown assay, deletion and point mutagenesis of CdGAP central domain, direct binding characterization","journal":"FEBS letters","confidence":"Medium","confidence_rationale":"Tier 2 / Moderate — direct binding mapped by pulldown and mutagenesis, single lab","pmids":["21349274"],"is_preprint":false},{"year":2012,"finding":"A polybasic region (PBR) preceding the RhoGAP domain of CdGAP mediates specific binding to PI(3,4,5)P3. In vitro reconstitution with membrane vesicles loaded with prenylated Rac1 showed that the PBR is required for full CdGAP activation in the presence of PI(3,4,5)P3. In fibroblasts, PBR mutants showed reduced ability to induce cell rounding and to negatively regulate cell spreading; the PBR is also required for in vivo Rac1 inactivation.","method":"In vitro lipid binding assay, membrane vesicle reconstitution with prenylated Rac1 and PI(3,4,5)P3, site-directed mutagenesis, cell spreading assays, Rac1 activity assay","journal":"The Journal of biological chemistry","confidence":"High","confidence_rationale":"Tier 1 / Moderate — reconstitution in vitro with prenylated substrate and lipid, mutagenesis, in vivo validation, multiple orthogonal methods","pmids":["22518840"],"is_preprint":false},{"year":2012,"finding":"CdGAP negatively regulates directed and random cell migration by controlling adhesion maturation and dynamics (both assembly and disassembly). CdGAP also localizes to adhesions in 3D matrix environments and CdGAP depletion promotes cancer cell migration and invasion through 3D matrices.","method":"siRNA knockdown, overexpression, live-cell imaging of adhesion dynamics, 3D matrix invasion assays","journal":"Cytoskeleton (Hoboken, N.J.)","confidence":"Medium","confidence_rationale":"Tier 2 / Moderate — loss-of-function with defined adhesion and migration phenotypes, 3D environment validation, single lab","pmids":["22907917"],"is_preprint":false},{"year":2014,"finding":"CdGAP is necessary for U2OS osteosarcoma cells to sense extracellular matrix stiffness and undergo durotaxis. CdGAP regulated rigidity-dependent motility by controlling membrane protrusion, adhesion dynamics, and Rac1 activity. CdGAP depletion abolished directed migration toward stiffer substrates.","method":"siRNA knockdown, PDMS gels with varying stiffness, Rac1 activity assay, adhesion dynamics imaging, durotaxis assay","journal":"PloS one","confidence":"Medium","confidence_rationale":"Tier 2 / Moderate — loss-of-function with mechanosensing phenotype, Rac1 activity linked to function, single lab","pmids":["24632816"],"is_preprint":false},{"year":2016,"finding":"CdGAP is required for embryonic vascular development in mice; CdGAP-deficient embryos show impaired vascular development at E15.5, superficial vessel defects, subcutaneous edema, and 44% embryonic/perinatal lethality. CdGAP associates with VEGF receptor-2 and controls VEGF-dependent signaling; CdGAP depletion impairs VEGF-mediated Rac1 activation and reduces phosphorylation of Gab1, Akt, PLCγ, and SHP2.","method":"CdGAP knockout mice, co-immunoprecipitation (CdGAP–VEGFR2), Rac1 activity assay, phosphorylation analysis by immunoblot, aortic ring sprouting assay, endothelial cell migration assay","journal":"Scientific reports","confidence":"High","confidence_rationale":"Tier 2 / Strong — knockout mouse model with defined vascular phenotype, direct binding to VEGFR2, downstream signaling analyzed, multiple orthogonal methods","pmids":["27270835"],"is_preprint":false},{"year":2017,"finding":"CdGAP uses its proline-rich domain to form a functional complex with Zeb2 to mediate transcriptional repression of E-cadherin (CDH1) in ErbB2-transformed breast cancer cells, acting as a nuclear transcriptional co-repressor in a GAP-independent manner. CdGAP knockdown decreased Snail1 and Zeb2 levels, increased E-cadherin, and restored cell-cell junctions. In vivo, CdGAP loss impaired tumor growth and lung metastasis.","method":"Co-immunoprecipitation (CdGAP–Zeb2), gene expression profiling, reporter assays for E-cadherin promoter, deletion mutant analysis, in vivo mouse tumor/metastasis model","journal":"Oncogene","confidence":"High","confidence_rationale":"Tier 2 / Strong — direct complex identified by co-IP, promoter assay with domain mapping, in vivo validation, multiple orthogonal methods","pmids":["28135249"],"is_preprint":false},{"year":2017,"finding":"Ajuba scaffold protein interacts with CdGAP at cell-cell contacts and inhibits CdGAP GAP activity. CdGAP recruitment to junctions does not require Ajuba; rather Ajuba controls CdGAP residence at cell-cell contacts. Ajuba binds Rac1 and CdGAP via distinct domains and can bring them into proximity to regulate Rac1 activity. Gain-of-function AOS CdGAP mutants strongly destabilize cell-cell contacts.","method":"Co-immunoprecipitation, GAP activity assay, fluorescence localization, overexpression of Ajuba/CdGAP mutants, junction integrity assays","journal":"Scientific reports","confidence":"Medium","confidence_rationale":"Tier 2 / Moderate — binding confirmed by co-IP, GAP inhibition demonstrated, domain mapping, single lab","pmids":["28835688"],"is_preprint":false},{"year":2018,"finding":"RSK phosphorylates CdGAP at Ser1093 and Ser1163 in response to phorbol ester, creating docking sites for 14-3-3 adaptor proteins. Binding of 14-3-3β to CdGAP inhibits CdGAP GAP activity, sequesters CdGAP in the cytoplasm, inhibits nucleocytoplasmic shuttling, abolishes CdGAP-induced cell rounding, inhibits CdGAP-mediated E-cadherin promoter repression, and inhibits CdGAP-induced cell migration. AOS-related CdGAP mutant proteins lacking these phospho-residues are not regulated by 14-3-3β.","method":"In vitro kinase assay (RSK), phospho-site mutagenesis, co-immunoprecipitation (CdGAP–14-3-3β), GAP activity assay, subcellular fractionation, cell migration assay, E-cadherin promoter reporter assay","journal":"Oncotarget","confidence":"High","confidence_rationale":"Tier 1 / Moderate — in vitro phosphorylation with mutagenesis, binding mapped to specific phospho-sites, multiple functional consequences tested, AOS mutant validation","pmids":["29545927"],"is_preprint":false},{"year":2022,"finding":"CdGAP interacts with β-PIX (ARHGEF7) through its basic region in podocytes. Upon EGF stimulation, both proteins translocate to the plasma membrane together. CdGAP depletion results in altered podocyte motility, increased basal Rac1 and Cdc42 activities, impaired β-PIX membrane translocation and tyrosine phosphorylation, and reduced activities of Src kinase, FAK, and paxillin. Podocyte-specific CdGAP knockout mice develop mild proteinuria exacerbated by Adriamycin.","method":"Proximity-based ligation assay, co-immunoprecipitation, siRNA knockdown, live imaging, GTPase activity assay, podocyte-specific knockout mice","journal":"Scientific reports","confidence":"Medium","confidence_rationale":"Tier 2 / Moderate — proximity ligation and co-IP identify interaction, knockout mouse with phenotype, signaling downstream effects measured, single lab","pmids":["36333327"],"is_preprint":false},{"year":2023,"finding":"DLG1 transcriptionally regulates ARHGAP31 expression, and this DLG1-ARHGAP31-CDC42 axis is essential for intestinal stem cell (ISC) survival in the context of increased Wnt signaling. ARHGAP31 acts downstream of DLG1 to deactivate CDC42, an effector of the non-canonical Wnt pathway, thereby modulating ISC response to fluctuating niche Wnt levels.","method":"RNA sequencing, genetic mouse models (conditional knockouts), epistasis analysis, intestinal organoids","journal":"Cell stem cell","confidence":"Medium","confidence_rationale":"Tier 2 / Moderate — genetic epistasis in mouse models with defined ISC phenotype, RNA-seq supports transcriptional regulation, single lab","pmids":["36640764"],"is_preprint":false},{"year":2023,"finding":"CdGAP interacts with the adaptor protein talin to modulate focal adhesion dynamics and integrin activation. CdGAP is positively regulated by TGF-β signaling during EMT and promotes tumor formation and metastasis in HER2+ breast cancer. CdGAP depletion mediates crosstalk with a DLC1-RhoA pathway in HER2+ primary tumors and is associated with a TGF-β-induced EMT transcriptional signature.","method":"Co-immunoprecipitation (CdGAP–talin), focal adhesion dynamics imaging, in vivo HER2+ mouse breast cancer model, gene expression profiling, intravasation/extravasation assays","journal":"Cell reports","confidence":"Medium","confidence_rationale":"Tier 2 / Moderate — direct binding to talin by co-IP, in vivo tumor model, crosstalk pathway identified, single lab","pmids":["37552602"],"is_preprint":false}],"current_model":"ARHGAP31/CdGAP is a serine/proline-rich RhoGAP that catalytically inactivates Rac1 and Cdc42 (but not RhoA) through its N-terminal GAP domain; its activity is regulated by phosphorylation (ERK1/2, GSK-3, RSK at Thr776, Ser1093, Ser1163), lipid binding (PI(3,4,5)P3 via a polybasic region), and protein–protein interactions (intersectin inhibits GAP activity; actopaxin/α-parvin and talin anchor it to focal adhesions; 14-3-3β sequesters it in the cytoplasm after RSK phosphorylation; Ajuba suppresses its activity at cell–cell junctions); independently, CdGAP uses its proline-rich domain in a GAP-independent nuclear function as a transcriptional co-repressor of E-cadherin together with Zeb2, and is positioned downstream of DLG1 and upstream of CDC42 in a Wnt-responsive intestinal stem cell axis."},"narrative":{"mechanistic_narrative":"ARHGAP31 (CdGAP) is a serine/proline-rich RhoGAP that catalytically inactivates Rac1 and Cdc42 but not RhoA, thereby controlling actin-dependent processes including lamellipodia and filopodia formation, cell spreading, adhesion dynamics, and directed migration [PMID:9786927, PMID:16860736, PMID:22907917]. Its GAP activity is integrated with multiple regulatory inputs: a polybasic region preceding the GAP domain binds PI(3,4,5)P3 to drive full activation at membranes [PMID:22518840]; phosphorylation downstream of the MEK-ERK pathway at Thr776 (also a GSK-3 target site that stabilizes CdGAP protein) tunes its activity [PMID:16024771, PMID:17158447]; and RSK phosphorylation at Ser1093/Ser1163 recruits 14-3-3β, which inhibits GAP activity, blocks nucleocytoplasmic shuttling, and sequesters CdGAP in the cytoplasm [PMID:29545927]. Protein interactions further localize and gate the enzyme: intersectin binds the central domain and inhibits GAP activity toward Rac1 [PMID:11744688, PMID:21349274], actopaxin (α-parvin) and talin anchor it to focal adhesions where it controls adhesion maturation, mechanosensing, and durotaxis [PMID:16860736, PMID:24632816, PMID:37552602], and Ajuba restrains its activity at cell-cell junctions [PMID:28835688]. Independently of its catalytic function, CdGAP acts as a nuclear transcriptional co-repressor of E-cadherin, using its proline-rich domain to form a complex with Zeb2 and promote TGF-β-driven EMT, invasion, and metastasis in ErbB2/HER2+ breast cancer [PMID:21042277, PMID:28135249, PMID:37552602]. CdGAP also operates in defined developmental and physiological axes, controlling VEGFR2-dependent Rac1 signaling in embryonic vascular development [PMID:27270835], β-PIX-coupled signaling in podocytes [PMID:36333327], and a DLG1-ARHGAP31-CDC42 axis governing intestinal stem cell survival under Wnt signaling [PMID:36640764]. Gain-of-function truncating mutations in the terminal exon of ARHGAP31 increase its GAP activity, deplete active Cdc42, and cause Adams-Oliver syndrome [PMID:21565291].","teleology":[{"year":1998,"claim":"Established CdGAP as a RhoGAP with defined substrate specificity, answering which GTPases it regulates and to which cytoskeletal events that connects.","evidence":"In vitro GAP assay and microinjection into serum-starved fibroblasts with PDGF/bradykinin/LPA stimulation","pmids":["9786927"],"confidence":"High","gaps":["No structural basis for Rac1/Cdc42 vs RhoA discrimination","Regulatory inputs to GAP activity not yet defined","Function of SH3-binding motifs unknown"]},{"year":2001,"claim":"Identified intersectin as a binding partner that inhibits CdGAP GAP activity, revealing the first protein-level regulatory control over the enzyme.","evidence":"Co-immunoprecipitation, in vitro GAP inhibition with intersectin SH3 domains, deletion mutant analysis in PDGF-stimulated cells","pmids":["11744688"],"confidence":"High","gaps":["Proposed conformational change mechanism not structurally demonstrated","Physiological context of inhibition unresolved"]},{"year":2005,"claim":"Connected CdGAP to MEK-ERK signaling, showing serum/PDGF-induced phosphorylation at Thr776 regulates its GAP activity.","evidence":"In vitro kinase assay (ERK1, RSK-1), phospho-site mapping, DEF-domain mutagenesis, cell stimulation","pmids":["16024771"],"confidence":"High","gaps":["Quantitative effect of Thr776 phosphorylation on catalysis not measured","Functional consequence in cell behavior not fully resolved"]},{"year":2006,"claim":"Defined CdGAP's focal adhesion localization and its requirement for spreading and migration, placing it within adhesion machinery.","evidence":"Co-IP/pulldown with actopaxin (α-parvin), RNAi knockdown, GAP-deficient mutant overexpression, migration assays","pmids":["16860736"],"confidence":"High","gaps":["Adhesion-dependent regulatory mechanism not molecularly defined","Whether catalytic vs scaffold function dominates at adhesions unclear"]},{"year":2006,"claim":"Showed GSK-3 phosphorylates CdGAP at Thr776 and controls its protein stability, adding a second kinase input converging on the same site.","evidence":"Yeast two-hybrid, co-IP, in vitro/in vivo phosphorylation, GSK-3 inhibitor treatment","pmids":["17158447"],"confidence":"High","gaps":["Mechanism linking Thr776 phosphorylation to protein-level upregulation unknown","Interplay between ERK and GSK-3 at Thr776 not resolved"]},{"year":2006,"claim":"Confirmed substrate specificity and ERK1/2 interaction are conserved in human CdGAP, validating mouse findings in human protein.","evidence":"In vitro/in vivo GAP assay, in vivo phosphorylation, co-IP with ERK1/2, cell morphology","pmids":["16519628"],"confidence":"Medium","gaps":["Human ERK1/2 docking region not mapped (non-canonical DEF)","Single lab study"]},{"year":2010,"claim":"Revealed a GAP-independent, proline-rich-domain function in TGF-β-induced EMT and E-cadherin loss, separating catalytic from non-catalytic roles.","evidence":"siRNA knockdown, domain-specific rescue, motility/invasion assays, E-cadherin immunoblotting in mammary tumor cells","pmids":["21042277"],"confidence":"Medium","gaps":["Molecular basis of E-cadherin regulation not yet identified","Single lab"]},{"year":2011,"claim":"Demonstrated gain-of-function truncating ARHGAP31 mutations cause Adams-Oliver syndrome, linking elevated GAP activity and Cdc42 depletion to a developmental disease.","evidence":"In vitro GAP assay of mutants, linkage/exome sequencing in families, active Cdc42 measurement, mouse expression analysis","pmids":["21565291"],"confidence":"High","gaps":["Tissue-specific developmental targets of excess GAP activity unresolved","Mechanism of transcript stabilization for truncated mutants unclear"]},{"year":2011,"claim":"Mapped the intersectin interaction to a novel basic-rich motif in the central domain, refining the earlier interaction model.","evidence":"Pulldown and point mutagenesis of CdGAP central domain","pmids":["21349274"],"confidence":"Medium","gaps":["Structural detail of the basic-motif/SH3D interface not determined","Single lab"]},{"year":2012,"claim":"Identified PI(3,4,5)P3 binding via a polybasic region as a membrane-localized activation mechanism for CdGAP.","evidence":"In vitro lipid binding, vesicle reconstitution with prenylated Rac1 and PI(3,4,5)P3, mutagenesis, Rac1 activity assay","pmids":["22518840"],"confidence":"High","gaps":["Spatial integration of lipid signal with phosphorylation inputs not defined","Upstream PI3K coupling not established"]},{"year":2012,"claim":"Established that CdGAP negatively regulates migration by controlling adhesion assembly/disassembly, including in 3D matrices.","evidence":"siRNA, live-cell imaging of adhesion dynamics, 3D invasion assays","pmids":["22907917"],"confidence":"Medium","gaps":["Direct molecular link between GAP activity and adhesion turnover unresolved","Single lab"]},{"year":2014,"claim":"Showed CdGAP is required for matrix-stiffness sensing and durotaxis, extending its role to mechanotransduction.","evidence":"siRNA, variable-stiffness PDMS gels, Rac1 activity assay, durotaxis assay in U2OS cells","pmids":["24632816"],"confidence":"Medium","gaps":["Mechanosensor coupling to CdGAP activity not defined","Single lab"]},{"year":2016,"claim":"Defined a physiological role in embryonic vascular development through association with VEGFR2 and control of VEGF-dependent Rac1 signaling.","evidence":"CdGAP knockout mice, co-IP with VEGFR2, Rac1 activity, downstream phospho-analysis, aortic ring/migration assays","pmids":["27270835"],"confidence":"High","gaps":["Whether VEGFR2 association is direct or scaffolded unclear","Mechanism of CdGAP-dependent Gab1/Akt/PLCγ/SHP2 phosphorylation unresolved"]},{"year":2017,"claim":"Identified CdGAP as a nuclear transcriptional co-repressor of E-cadherin via a proline-rich-domain complex with Zeb2, giving molecular form to its GAP-independent EMT role.","evidence":"Co-IP with Zeb2, E-cadherin promoter reporter, expression profiling, domain mapping, in vivo tumor/metastasis model","pmids":["28135249"],"confidence":"High","gaps":["Mechanism of CdGAP nuclear import/regulation not fully defined","Direct DNA contact vs Zeb2-mediated recruitment not distinguished"]},{"year":2017,"claim":"Showed Ajuba binds CdGAP at cell-cell contacts and inhibits its GAP activity, adding a junctional layer of regulation.","evidence":"Co-IP, GAP activity assay, fluorescence localization, junction integrity assays with AOS mutants","pmids":["28835688"],"confidence":"Medium","gaps":["What recruits CdGAP to junctions remains unknown","Single lab"]},{"year":2018,"claim":"Demonstrated RSK-mediated phosphorylation at Ser1093/Ser1163 creates 14-3-3β docking sites that inhibit GAP activity and block nuclear shuttling, coupling kinase signaling to both catalytic and transcriptional functions.","evidence":"In vitro RSK kinase assay, phospho-site mutagenesis, co-IP with 14-3-3β, fractionation, migration and promoter reporter assays","pmids":["29545927"],"confidence":"High","gaps":["Stimulus-specificity of RSK input beyond phorbol ester unclear","Quantitative partitioning between cytoplasmic and nuclear pools not measured"]},{"year":2022,"claim":"Identified β-PIX (ARHGEF7) as a partner in podocytes, linking CdGAP to Src/FAK/paxillin signaling and renal physiology.","evidence":"Proximity ligation, co-IP, siRNA, live imaging, GTPase assays, podocyte-specific knockout mice","pmids":["36333327"],"confidence":"Medium","gaps":["Mechanism coupling CdGAP loss to β-PIX phosphorylation defects unclear","Single lab"]},{"year":2023,"claim":"Placed ARHGAP31 in a DLG1-ARHGAP31-CDC42 axis controlling intestinal stem cell survival under Wnt signaling, identifying transcriptional control of its expression.","evidence":"RNA-seq, conditional knockout mice, epistasis analysis, intestinal organoids","pmids":["36640764"],"confidence":"Medium","gaps":["Mechanism of DLG1-driven ARHGAP31 transcription not defined","Single lab"]},{"year":2023,"claim":"Identified talin as a focal-adhesion partner modulating integrin activation and tied CdGAP to a DLC1-RhoA crosstalk in HER2+ breast cancer.","evidence":"Co-IP with talin, adhesion dynamics imaging, in vivo HER2+ mouse model, expression profiling, intravasation/extravasation assays","pmids":["37552602"],"confidence":"Medium","gaps":["Molecular basis of CdGAP-talin-integrin coupling unresolved","Single lab"]},{"year":null,"claim":"How CdGAP's many regulatory inputs (lipid binding, ERK/GSK-3/RSK phosphorylation, and partner interactions) are spatiotemporally integrated to switch between its catalytic GAP and nuclear co-repressor functions remains unresolved.","evidence":"","pmids":[],"confidence":"Medium","gaps":["No structural model of full-length CdGAP","Mechanism governing cytoplasmic-vs-nuclear partitioning incompletely defined","Relative contribution of catalytic vs scaffold/transcriptional roles in vivo unclear"]}],"mechanism_profile":{"molecular_activity":[{"term_id":"GO:0098772","term_label":"molecular function regulator activity","supporting_discovery_ids":[0,1,7,9]},{"term_id":"GO:0140110","term_label":"transcription regulator activity","supporting_discovery_ids":[6,13]},{"term_id":"GO:0008289","term_label":"lipid binding","supporting_discovery_ids":[9]},{"term_id":"GO:0008092","term_label":"cytoskeletal protein binding","supporting_discovery_ids":[3,18]}],"localization":[{"term_id":"GO:0005829","term_label":"cytosol","supporting_discovery_ids":[15]},{"term_id":"GO:0005634","term_label":"nucleus","supporting_discovery_ids":[13,15]},{"term_id":"GO:0005886","term_label":"plasma membrane","supporting_discovery_ids":[16]},{"term_id":"GO:0005856","term_label":"cytoskeleton","supporting_discovery_ids":[3,18]}],"pathway":[{"term_id":"R-HSA-162582","term_label":"Signal Transduction","supporting_discovery_ids":[12,16]},{"term_id":"R-HSA-1266738","term_label":"Developmental Biology","supporting_discovery_ids":[7,12,17]},{"term_id":"R-HSA-74160","term_label":"Gene expression (Transcription)","supporting_discovery_ids":[13,17]},{"term_id":"R-HSA-1643685","term_label":"Disease","supporting_discovery_ids":[7,13,18]}],"complexes":[],"partners":["ITSN1","PARVA","ZEB2","VEGFR2","ARHGEF7","TLN1","AJUBA","YWHAB"],"other_free_text":[]}},"prefetch_data":{"uniprot":{"accession":"Q2M1Z3","full_name":"Rho GTPase-activating protein 31","aliases":["Cdc42 GTPase-activating protein"],"length_aa":1444,"mass_kda":157.0,"function":"Functions as a GTPase-activating protein (GAP) for RAC1 and CDC42. Required for cell spreading, polarized lamellipodia formation and cell migration","subcellular_location":"Cell projection, lamellipodium; Cell junction, focal adhesion","url":"https://www.uniprot.org/uniprotkb/Q2M1Z3/entry"},"depmap":{"release":"DepMap","has_data":true,"is_common_essential":false,"resolved_as":"","url":"https://depmap.org/portal/gene/ARHGAP31","classification":"Not Classified","n_dependent_lines":0,"n_total_lines":1208,"dependency_fraction":0.0},"opencell":{"profiled":false,"resolved_as":"","ensg_id":"","cell_line_id":"","localizations":[],"interactors":[],"url":"https://opencell.sf.czbiohub.org/search/ARHGAP31","total_profiled":1310},"omim":[{"mim_id":"616028","title":"ADAMS-OLIVER SYNDROME 5; AOS5","url":"https://www.omim.org/entry/616028"},{"mim_id":"614264","title":"RHO GTPase-ACTIVATING PROTEIN 30; ARHGAP30","url":"https://www.omim.org/entry/614264"},{"mim_id":"610911","title":"RHO GTPase-ACTIVATING PROTEIN 31; ARHGAP31","url":"https://www.omim.org/entry/610911"},{"mim_id":"100300","title":"ADAMS-OLIVER SYNDROME 1; AOS1","url":"https://www.omim.org/entry/100300"}],"hpa":{"profiled":true,"resolved_as":"","reliability":"","locations":[],"tissue_specificity":"Low tissue specificity","tissue_distribution":"Detected in many","driving_tissues":[],"url":"https://www.proteinatlas.org/search/ARHGAP31"},"hgnc":{"alias_symbol":["CDGAP"],"prev_symbol":[]},"alphafold":{"accession":"Q2M1Z3","domains":[{"cath_id":"1.10.555.10","chopping":"24-218","consensus_level":"medium","plddt":92.5096,"start":24,"end":218}],"viewer_url":"https://alphafold.ebi.ac.uk/entry/Q2M1Z3","model_url":"https://alphafold.ebi.ac.uk/files/AF-Q2M1Z3-F1-model_v6.cif","pae_url":"https://alphafold.ebi.ac.uk/files/AF-Q2M1Z3-F1-predicted_aligned_error_v6.png","plddt_mean":42.88},"mouse_models":{"mgi_url":"https://www.informatics.jax.org/marker/summary?nomen=ARHGAP31","jax_strain_url":"https://www.jax.org/strain/search?query=ARHGAP31"},"sequence":{"accession":"Q2M1Z3","fasta_url":"https://rest.uniprot.org/uniprotkb/Q2M1Z3.fasta","uniprot_url":"https://www.uniprot.org/uniprotkb/Q2M1Z3/entry","alphafold_viewer_url":"https://alphafold.ebi.ac.uk/entry/Q2M1Z3"}},"corpus_meta":[{"pmid":"21565291","id":"PMC_21565291","title":"Gain-of-function mutations of ARHGAP31, a Cdc42/Rac1 GTPase regulator, cause syndromic cutis aplasia and limb anomalies.","date":"2011","source":"American journal of human genetics","url":"https://pubmed.ncbi.nlm.nih.gov/21565291","citation_count":90,"is_preprint":false},{"pmid":"9786927","id":"PMC_9786927","title":"CdGAP, a novel proline-rich GTPase-activating protein for Cdc42 and Rac.","date":"1998","source":"The Journal of biological chemistry","url":"https://pubmed.ncbi.nlm.nih.gov/9786927","citation_count":71,"is_preprint":false},{"pmid":"11744688","id":"PMC_11744688","title":"The activity of the GTPase-activating protein CdGAP is regulated by the endocytic protein intersectin.","date":"2001","source":"The Journal of biological chemistry","url":"https://pubmed.ncbi.nlm.nih.gov/11744688","citation_count":67,"is_preprint":false},{"pmid":"16860736","id":"PMC_16860736","title":"CdGAP associates with actopaxin to regulate integrin-dependent changes in cell morphology and motility.","date":"2006","source":"Current biology : CB","url":"https://pubmed.ncbi.nlm.nih.gov/16860736","citation_count":49,"is_preprint":false},{"pmid":"24632816","id":"PMC_24632816","title":"The focal adhesion-localized CdGAP regulates matrix rigidity sensing and durotaxis.","date":"2014","source":"PloS one","url":"https://pubmed.ncbi.nlm.nih.gov/24632816","citation_count":45,"is_preprint":false},{"pmid":"28135249","id":"PMC_28135249","title":"The Cdc42/Rac1 regulator CdGAP is a novel E-cadherin transcriptional co-repressor with Zeb2 in breast cancer.","date":"2017","source":"Oncogene","url":"https://pubmed.ncbi.nlm.nih.gov/28135249","citation_count":33,"is_preprint":false},{"pmid":"16024771","id":"PMC_16024771","title":"Extracellular signal-regulated kinase 1 interacts with and phosphorylates CdGAP at an important regulatory site.","date":"2005","source":"Molecular and cellular biology","url":"https://pubmed.ncbi.nlm.nih.gov/16024771","citation_count":29,"is_preprint":false},{"pmid":"21042277","id":"PMC_21042277","title":"CdGAP is required for transforming growth factor β- and Neu/ErbB-2-induced breast cancer cell motility and invasion.","date":"2010","source":"Oncogene","url":"https://pubmed.ncbi.nlm.nih.gov/21042277","citation_count":29,"is_preprint":false},{"pmid":"27270835","id":"PMC_27270835","title":"CdGAP/ARHGAP31, a Cdc42/Rac1 GTPase regulator, is critical for vascular development and VEGF-mediated angiogenesis.","date":"2016","source":"Scientific reports","url":"https://pubmed.ncbi.nlm.nih.gov/27270835","citation_count":27,"is_preprint":false},{"pmid":"16519628","id":"PMC_16519628","title":"The human orthologue of CdGAP is a phosphoprotein and a GTPase-activating protein for Cdc42 and Rac1 but not RhoA.","date":"2006","source":"Biology of the cell","url":"https://pubmed.ncbi.nlm.nih.gov/16519628","citation_count":25,"is_preprint":false},{"pmid":"24668619","id":"PMC_24668619","title":"Isolated terminal limb reduction defects: extending the clinical spectrum of Adams-Oliver syndrome and ARHGAP31 mutations.","date":"2014","source":"American journal of medical genetics. Part A","url":"https://pubmed.ncbi.nlm.nih.gov/24668619","citation_count":18,"is_preprint":false},{"pmid":"22907917","id":"PMC_22907917","title":"CdGAP regulates cell migration and adhesion dynamics in two-and three-dimensional matrix environments.","date":"2012","source":"Cytoskeleton (Hoboken, N.J.)","url":"https://pubmed.ncbi.nlm.nih.gov/22907917","citation_count":17,"is_preprint":false},{"pmid":"36640764","id":"PMC_36640764","title":"A DLG1-ARHGAP31-CDC42 axis is essential for the intestinal stem cell response to fluctuating niche Wnt signaling.","date":"2023","source":"Cell stem cell","url":"https://pubmed.ncbi.nlm.nih.gov/36640764","citation_count":17,"is_preprint":false},{"pmid":"22518840","id":"PMC_22518840","title":"A stretch of polybasic residues mediates Cdc42 GTPase-activating protein (CdGAP) binding to phosphatidylinositol 3,4,5-trisphosphate and regulates its GAP activity.","date":"2012","source":"The Journal of biological chemistry","url":"https://pubmed.ncbi.nlm.nih.gov/22518840","citation_count":15,"is_preprint":false},{"pmid":"29545927","id":"PMC_29545927","title":"CdGAP/ARHGAP31 is regulated by RSK phosphorylation and binding to 14-3-3β adaptor protein.","date":"2018","source":"Oncotarget","url":"https://pubmed.ncbi.nlm.nih.gov/29545927","citation_count":12,"is_preprint":false},{"pmid":"34493786","id":"PMC_34493786","title":"CdGAP promotes prostate cancer metastasis by regulating epithelial-to-mesenchymal transition, cell cycle progression, and apoptosis.","date":"2021","source":"Communications biology","url":"https://pubmed.ncbi.nlm.nih.gov/34493786","citation_count":12,"is_preprint":false},{"pmid":"28835688","id":"PMC_28835688","title":"The scaffold protein Ajuba suppresses CdGAP activity in epithelia to maintain stable cell-cell contacts.","date":"2017","source":"Scientific reports","url":"https://pubmed.ncbi.nlm.nih.gov/28835688","citation_count":11,"is_preprint":false},{"pmid":"21349274","id":"PMC_21349274","title":"Cdc42 GTPase-activating protein (CdGAP) interacts with the SH3D domain of Intersectin through a novel basic-rich motif.","date":"2011","source":"FEBS letters","url":"https://pubmed.ncbi.nlm.nih.gov/21349274","citation_count":11,"is_preprint":false},{"pmid":"17158447","id":"PMC_17158447","title":"Glycogen synthase kinase-3 phosphorylates CdGAP at a consensus ERK 1 regulatory site.","date":"2006","source":"The Journal of biological chemistry","url":"https://pubmed.ncbi.nlm.nih.gov/17158447","citation_count":9,"is_preprint":false},{"pmid":"36333327","id":"PMC_36333327","title":"CdGAP maintains podocyte function and modulates focal adhesions in a Src kinase-dependent manner.","date":"2022","source":"Scientific reports","url":"https://pubmed.ncbi.nlm.nih.gov/36333327","citation_count":7,"is_preprint":false},{"pmid":"37552602","id":"PMC_37552602","title":"CdGAP is a talin-binding protein and a target of TGF-β signaling that promotes HER2-positive breast cancer growth and metastasis.","date":"2023","source":"Cell reports","url":"https://pubmed.ncbi.nlm.nih.gov/37552602","citation_count":7,"is_preprint":false},{"pmid":"36176297","id":"PMC_36176297","title":"Synergistic effects of rare variants of ARHGAP31 and FBLN1 in vitro in terminal transverse limb defects.","date":"2022","source":"Frontiers in genetics","url":"https://pubmed.ncbi.nlm.nih.gov/36176297","citation_count":2,"is_preprint":false},{"pmid":"38790165","id":"PMC_38790165","title":"Characterization of a New Variant in ARHGAP31 Probably Involved in Adams-Oliver Syndrome in a Family with a Variable Phenotypic Spectrum.","date":"2024","source":"Genes","url":"https://pubmed.ncbi.nlm.nih.gov/38790165","citation_count":2,"is_preprint":false},{"pmid":null,"id":"bio_10.1101_2024.11.20.624451","title":"Rho GTPases signaling mediates aggressiveness and differentiation in neuroblastoma tumors","date":"2024-11-21","source":"bioRxiv","url":"https://doi.org/10.1101/2024.11.20.624451","citation_count":0,"is_preprint":true}],"cost":{"stage1":{"model":"claude-sonnet-4-6","input_tokens":14085,"output_tokens":5567,"usd":0.06288,"stage1_stop_reason":"end_turn"},"stage2":{"model":"claude-opus-4-8","input_tokens":13954,"output_tokens":5636,"usd":0.105335,"stage2_stop_reason":"end_turn"},"total_usd":0.168215,"stage1_batch_id":"msgbatch_0179Z2pxRYBsXi2xADneafRw","stage2_batch_id":"msgbatch_01LaY9ZaVdgjWArcZo9eqQE5","note":"batch pricing = 50% of standard"},"stage1_raw":"```json\n{\n  \"discoveries\": [\n    {\n      \"year\": 1998,\n      \"finding\": \"CdGAP (ARHGAP31) was identified as a novel GTPase-activating protein with in vitro GAP activity toward both Cdc42 and Rac1 but not RhoA. Microinjection of CdGAP into serum-starved fibroblasts inhibited PDGF-induced lamellipodia (Rac-mediated) and bradykinin-induced filopodia (Cdc42-mediated), but had no effect on LPA-induced stress fiber formation. The C-terminus contains potential PKC phosphorylation sites and five SH3 binding motifs.\",\n      \"method\": \"Yeast two-hybrid screen, in vitro GAP assay, microinjection into fibroblasts\",\n      \"journal\": \"The Journal of biological chemistry\",\n      \"confidence\": \"High\",\n      \"confidence_rationale\": \"Tier 1 / Strong — in vitro GAP assay with substrate specificity established, functional microinjection with specific phenotypic readouts, foundational characterization paper\",\n      \"pmids\": [\"9786927\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2001,\n      \"finding\": \"The endocytic protein intersectin interacts with CdGAP through a subset of its SH3 domains and inhibits CdGAP's GAP activity toward Rac1. In PDGF-stimulated Swiss 3T3 cells, intersectin co-localizes with CdGAP. The central domain of CdGAP (not the proline-rich domain) is required for the CdGAP–intersectin interaction, but the C-terminal proline-rich domain is required for intersectin-mediated inhibition of GAP activity, suggesting a conformational change mechanism.\",\n      \"method\": \"Co-immunoprecipitation, in vitro GAP assay with intersectin-SH3 domains, co-localization in PDGF-stimulated cells, deletion mutant analysis\",\n      \"journal\": \"The Journal of biological chemistry\",\n      \"confidence\": \"High\",\n      \"confidence_rationale\": \"Tier 2 / Strong — reciprocal interaction mapping, in vitro GAP inhibition assay, domain dissection with multiple mutants, in vivo co-localization\",\n      \"pmids\": [\"11744688\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2005,\n      \"finding\": \"CdGAP is phosphorylated in vivo on serine and threonine residues downstream of the MEK-ERK pathway in response to serum or PDGF. ERK1 and RSK-1 phosphorylate CdGAP in vitro. A DEF domain (docking for ERK, FXFP motif) in the proline-rich region is required for efficient ERK1/2 binding and phosphorylation. Thr776 was identified as an in vivo ERK1/2 target site that serves as an important regulatory site of CdGAP GAP activity.\",\n      \"method\": \"In vitro kinase assay, site-directed mutagenesis, phosphorylation mapping, co-immunoprecipitation, cell stimulation with serum/PDGF\",\n      \"journal\": \"Molecular and cellular biology\",\n      \"confidence\": \"High\",\n      \"confidence_rationale\": \"Tier 1 / Moderate — in vitro kinase assay with mutagenesis, in vivo phosphorylation confirmed, regulatory site functionally characterized\",\n      \"pmids\": [\"16024771\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2006,\n      \"finding\": \"CdGAP localizes to focal adhesions through a direct interaction with the N-terminus of actopaxin (α-parvin), a paxillin and actin binding protein. CdGAP GAP activity is regulated in an adhesion-dependent manner. Both overexpression and RNAi knockdown of CdGAP impaired normal cell spreading, polarized lamellipodia formation, and cell migration. An actopaxin mutant defective for CdGAP binding attenuated these effects.\",\n      \"method\": \"Co-immunoprecipitation/pulldown, RNAi knockdown, overexpression of GAP-deficient mutant, fluorescence localization, cell spreading/migration assays\",\n      \"journal\": \"Current biology : CB\",\n      \"confidence\": \"High\",\n      \"confidence_rationale\": \"Tier 2 / Moderate — direct binding demonstrated, loss-of-function with defined cellular phenotypes, mutant rescue experiments, focal adhesion localization linked to function\",\n      \"pmids\": [\"16860736\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2006,\n      \"finding\": \"Human CdGAP (KIAA1204) shares 76% sequence identity with mouse CdGAP-long isoform and retains in vitro and in vivo GAP activity toward Cdc42 and Rac1 but not RhoA, confirming substrate specificity is conserved. Human CdGAP is phosphorylated in vivo on serine and threonine residues. Unlike mouse CdGAP, human CdGAP interacts with ERK1/2 through a region not involving a canonical DEF domain. CdGAP expression caused membrane blebbing in COS-7 cells.\",\n      \"method\": \"In vitro GAP assay, in vivo phosphorylation analysis, co-immunoprecipitation with ERK1/2, cell morphology assay\",\n      \"journal\": \"Biology of the cell\",\n      \"confidence\": \"Medium\",\n      \"confidence_rationale\": \"Tier 2 / Moderate — GAP activity confirmed in vitro and in vivo, interaction with ERK1/2 by co-IP, single lab study\",\n      \"pmids\": [\"16519628\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2006,\n      \"finding\": \"GSK-3α was identified as a CdGAP binding partner. GSK-3α and GSK-3β both interact with CdGAP in mammalian cells and phosphorylate CdGAP in vitro and in vivo at Thr776, the same site previously identified as an ERK1/2 target. GSK-3 activity is necessary for serum-stimulated upregulation of CdGAP protein levels (but not mRNA), revealing a novel mechanism for controlling Cdc42/Rac1 signaling.\",\n      \"method\": \"Yeast two-hybrid, co-immunoprecipitation, in vitro kinase assay, in vivo phosphorylation, GSK-3 inhibitor treatment\",\n      \"journal\": \"The Journal of biological chemistry\",\n      \"confidence\": \"High\",\n      \"confidence_rationale\": \"Tier 1 / Moderate — yeast two-hybrid confirmed by co-IP, in vitro and in vivo phosphorylation at specific site, pharmacological inhibitor validates functional role\",\n      \"pmids\": [\"17158447\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2010,\n      \"finding\": \"CdGAP is required for TGF-β-induced cell motility and invasion of Neu/ErbB-2-expressing mammary tumor cells. The proline-rich domain (PRD) but not the GAP domain of CdGAP is essential for TGF-β-mediated motility and invasion. CdGAP depletion increased E-cadherin expression and prevented complete E-cadherin loss during TGF-β-induced EMT. TGF-β also induces CdGAP expression and phosphorylation in NMuMG cells.\",\n      \"method\": \"siRNA knockdown, overexpression of deletion mutants (rescue analysis), cell motility/invasion assays, immunoblotting for E-cadherin\",\n      \"journal\": \"Oncogene\",\n      \"confidence\": \"Medium\",\n      \"confidence_rationale\": \"Tier 2 / Moderate — siRNA knockdown with defined phenotype, domain-specific rescue experiments, single lab\",\n      \"pmids\": [\"21042277\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2011,\n      \"finding\": \"Gain-of-function truncating mutations in the terminal exon of ARHGAP31 cause Adams-Oliver syndrome. Mutant transcripts are stable and increase ARHGAP31 GAP activity in vitro. Constitutively active ARHGAP31 mutations result in loss of available active Cdc42 and disruption of actin cytoskeletal structures. Mouse Arhgap31 expression is restricted to terminal limb buds and craniofacial processes during early development.\",\n      \"method\": \"In vitro GAP activity assay of mutant proteins, genome-wide linkage analysis, exome sequencing, Rho GTPase activity assay (active Cdc42 measurement), mouse expression analysis\",\n      \"journal\": \"American journal of human genetics\",\n      \"confidence\": \"High\",\n      \"confidence_rationale\": \"Tier 1 / Strong — gain-of-function mechanism confirmed by in vitro GAP assay, loss of active Cdc42 measured in cells, replicated in multiple independent families\",\n      \"pmids\": [\"21565291\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2011,\n      \"finding\": \"The interaction between intersectin's SH3D domain and CdGAP is mediated through a novel basic-rich motif in the central domain of CdGAP containing a conserved xKx(K/R)K sequence, not a canonical proline-rich SH3-binding motif.\",\n      \"method\": \"Pulldown assay, deletion and point mutagenesis of CdGAP central domain, direct binding characterization\",\n      \"journal\": \"FEBS letters\",\n      \"confidence\": \"Medium\",\n      \"confidence_rationale\": \"Tier 2 / Moderate — direct binding mapped by pulldown and mutagenesis, single lab\",\n      \"pmids\": [\"21349274\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2012,\n      \"finding\": \"A polybasic region (PBR) preceding the RhoGAP domain of CdGAP mediates specific binding to PI(3,4,5)P3. In vitro reconstitution with membrane vesicles loaded with prenylated Rac1 showed that the PBR is required for full CdGAP activation in the presence of PI(3,4,5)P3. In fibroblasts, PBR mutants showed reduced ability to induce cell rounding and to negatively regulate cell spreading; the PBR is also required for in vivo Rac1 inactivation.\",\n      \"method\": \"In vitro lipid binding assay, membrane vesicle reconstitution with prenylated Rac1 and PI(3,4,5)P3, site-directed mutagenesis, cell spreading assays, Rac1 activity assay\",\n      \"journal\": \"The Journal of biological chemistry\",\n      \"confidence\": \"High\",\n      \"confidence_rationale\": \"Tier 1 / Moderate — reconstitution in vitro with prenylated substrate and lipid, mutagenesis, in vivo validation, multiple orthogonal methods\",\n      \"pmids\": [\"22518840\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2012,\n      \"finding\": \"CdGAP negatively regulates directed and random cell migration by controlling adhesion maturation and dynamics (both assembly and disassembly). CdGAP also localizes to adhesions in 3D matrix environments and CdGAP depletion promotes cancer cell migration and invasion through 3D matrices.\",\n      \"method\": \"siRNA knockdown, overexpression, live-cell imaging of adhesion dynamics, 3D matrix invasion assays\",\n      \"journal\": \"Cytoskeleton (Hoboken, N.J.)\",\n      \"confidence\": \"Medium\",\n      \"confidence_rationale\": \"Tier 2 / Moderate — loss-of-function with defined adhesion and migration phenotypes, 3D environment validation, single lab\",\n      \"pmids\": [\"22907917\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2014,\n      \"finding\": \"CdGAP is necessary for U2OS osteosarcoma cells to sense extracellular matrix stiffness and undergo durotaxis. CdGAP regulated rigidity-dependent motility by controlling membrane protrusion, adhesion dynamics, and Rac1 activity. CdGAP depletion abolished directed migration toward stiffer substrates.\",\n      \"method\": \"siRNA knockdown, PDMS gels with varying stiffness, Rac1 activity assay, adhesion dynamics imaging, durotaxis assay\",\n      \"journal\": \"PloS one\",\n      \"confidence\": \"Medium\",\n      \"confidence_rationale\": \"Tier 2 / Moderate — loss-of-function with mechanosensing phenotype, Rac1 activity linked to function, single lab\",\n      \"pmids\": [\"24632816\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2016,\n      \"finding\": \"CdGAP is required for embryonic vascular development in mice; CdGAP-deficient embryos show impaired vascular development at E15.5, superficial vessel defects, subcutaneous edema, and 44% embryonic/perinatal lethality. CdGAP associates with VEGF receptor-2 and controls VEGF-dependent signaling; CdGAP depletion impairs VEGF-mediated Rac1 activation and reduces phosphorylation of Gab1, Akt, PLCγ, and SHP2.\",\n      \"method\": \"CdGAP knockout mice, co-immunoprecipitation (CdGAP–VEGFR2), Rac1 activity assay, phosphorylation analysis by immunoblot, aortic ring sprouting assay, endothelial cell migration assay\",\n      \"journal\": \"Scientific reports\",\n      \"confidence\": \"High\",\n      \"confidence_rationale\": \"Tier 2 / Strong — knockout mouse model with defined vascular phenotype, direct binding to VEGFR2, downstream signaling analyzed, multiple orthogonal methods\",\n      \"pmids\": [\"27270835\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2017,\n      \"finding\": \"CdGAP uses its proline-rich domain to form a functional complex with Zeb2 to mediate transcriptional repression of E-cadherin (CDH1) in ErbB2-transformed breast cancer cells, acting as a nuclear transcriptional co-repressor in a GAP-independent manner. CdGAP knockdown decreased Snail1 and Zeb2 levels, increased E-cadherin, and restored cell-cell junctions. In vivo, CdGAP loss impaired tumor growth and lung metastasis.\",\n      \"method\": \"Co-immunoprecipitation (CdGAP–Zeb2), gene expression profiling, reporter assays for E-cadherin promoter, deletion mutant analysis, in vivo mouse tumor/metastasis model\",\n      \"journal\": \"Oncogene\",\n      \"confidence\": \"High\",\n      \"confidence_rationale\": \"Tier 2 / Strong — direct complex identified by co-IP, promoter assay with domain mapping, in vivo validation, multiple orthogonal methods\",\n      \"pmids\": [\"28135249\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2017,\n      \"finding\": \"Ajuba scaffold protein interacts with CdGAP at cell-cell contacts and inhibits CdGAP GAP activity. CdGAP recruitment to junctions does not require Ajuba; rather Ajuba controls CdGAP residence at cell-cell contacts. Ajuba binds Rac1 and CdGAP via distinct domains and can bring them into proximity to regulate Rac1 activity. Gain-of-function AOS CdGAP mutants strongly destabilize cell-cell contacts.\",\n      \"method\": \"Co-immunoprecipitation, GAP activity assay, fluorescence localization, overexpression of Ajuba/CdGAP mutants, junction integrity assays\",\n      \"journal\": \"Scientific reports\",\n      \"confidence\": \"Medium\",\n      \"confidence_rationale\": \"Tier 2 / Moderate — binding confirmed by co-IP, GAP inhibition demonstrated, domain mapping, single lab\",\n      \"pmids\": [\"28835688\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2018,\n      \"finding\": \"RSK phosphorylates CdGAP at Ser1093 and Ser1163 in response to phorbol ester, creating docking sites for 14-3-3 adaptor proteins. Binding of 14-3-3β to CdGAP inhibits CdGAP GAP activity, sequesters CdGAP in the cytoplasm, inhibits nucleocytoplasmic shuttling, abolishes CdGAP-induced cell rounding, inhibits CdGAP-mediated E-cadherin promoter repression, and inhibits CdGAP-induced cell migration. AOS-related CdGAP mutant proteins lacking these phospho-residues are not regulated by 14-3-3β.\",\n      \"method\": \"In vitro kinase assay (RSK), phospho-site mutagenesis, co-immunoprecipitation (CdGAP–14-3-3β), GAP activity assay, subcellular fractionation, cell migration assay, E-cadherin promoter reporter assay\",\n      \"journal\": \"Oncotarget\",\n      \"confidence\": \"High\",\n      \"confidence_rationale\": \"Tier 1 / Moderate — in vitro phosphorylation with mutagenesis, binding mapped to specific phospho-sites, multiple functional consequences tested, AOS mutant validation\",\n      \"pmids\": [\"29545927\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2022,\n      \"finding\": \"CdGAP interacts with β-PIX (ARHGEF7) through its basic region in podocytes. Upon EGF stimulation, both proteins translocate to the plasma membrane together. CdGAP depletion results in altered podocyte motility, increased basal Rac1 and Cdc42 activities, impaired β-PIX membrane translocation and tyrosine phosphorylation, and reduced activities of Src kinase, FAK, and paxillin. Podocyte-specific CdGAP knockout mice develop mild proteinuria exacerbated by Adriamycin.\",\n      \"method\": \"Proximity-based ligation assay, co-immunoprecipitation, siRNA knockdown, live imaging, GTPase activity assay, podocyte-specific knockout mice\",\n      \"journal\": \"Scientific reports\",\n      \"confidence\": \"Medium\",\n      \"confidence_rationale\": \"Tier 2 / Moderate — proximity ligation and co-IP identify interaction, knockout mouse with phenotype, signaling downstream effects measured, single lab\",\n      \"pmids\": [\"36333327\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2023,\n      \"finding\": \"DLG1 transcriptionally regulates ARHGAP31 expression, and this DLG1-ARHGAP31-CDC42 axis is essential for intestinal stem cell (ISC) survival in the context of increased Wnt signaling. ARHGAP31 acts downstream of DLG1 to deactivate CDC42, an effector of the non-canonical Wnt pathway, thereby modulating ISC response to fluctuating niche Wnt levels.\",\n      \"method\": \"RNA sequencing, genetic mouse models (conditional knockouts), epistasis analysis, intestinal organoids\",\n      \"journal\": \"Cell stem cell\",\n      \"confidence\": \"Medium\",\n      \"confidence_rationale\": \"Tier 2 / Moderate — genetic epistasis in mouse models with defined ISC phenotype, RNA-seq supports transcriptional regulation, single lab\",\n      \"pmids\": [\"36640764\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2023,\n      \"finding\": \"CdGAP interacts with the adaptor protein talin to modulate focal adhesion dynamics and integrin activation. CdGAP is positively regulated by TGF-β signaling during EMT and promotes tumor formation and metastasis in HER2+ breast cancer. CdGAP depletion mediates crosstalk with a DLC1-RhoA pathway in HER2+ primary tumors and is associated with a TGF-β-induced EMT transcriptional signature.\",\n      \"method\": \"Co-immunoprecipitation (CdGAP–talin), focal adhesion dynamics imaging, in vivo HER2+ mouse breast cancer model, gene expression profiling, intravasation/extravasation assays\",\n      \"journal\": \"Cell reports\",\n      \"confidence\": \"Medium\",\n      \"confidence_rationale\": \"Tier 2 / Moderate — direct binding to talin by co-IP, in vivo tumor model, crosstalk pathway identified, single lab\",\n      \"pmids\": [\"37552602\"],\n      \"is_preprint\": false\n    }\n  ],\n  \"current_model\": \"ARHGAP31/CdGAP is a serine/proline-rich RhoGAP that catalytically inactivates Rac1 and Cdc42 (but not RhoA) through its N-terminal GAP domain; its activity is regulated by phosphorylation (ERK1/2, GSK-3, RSK at Thr776, Ser1093, Ser1163), lipid binding (PI(3,4,5)P3 via a polybasic region), and protein–protein interactions (intersectin inhibits GAP activity; actopaxin/α-parvin and talin anchor it to focal adhesions; 14-3-3β sequesters it in the cytoplasm after RSK phosphorylation; Ajuba suppresses its activity at cell–cell junctions); independently, CdGAP uses its proline-rich domain in a GAP-independent nuclear function as a transcriptional co-repressor of E-cadherin together with Zeb2, and is positioned downstream of DLG1 and upstream of CDC42 in a Wnt-responsive intestinal stem cell axis.\"\n}\n```","stage2_raw":"{\n  \"mechanistic_narrative\": \"ARHGAP31 (CdGAP) is a serine/proline-rich RhoGAP that catalytically inactivates Rac1 and Cdc42 but not RhoA, thereby controlling actin-dependent processes including lamellipodia and filopodia formation, cell spreading, adhesion dynamics, and directed migration [#0, #3, #10]. Its GAP activity is integrated with multiple regulatory inputs: a polybasic region preceding the GAP domain binds PI(3,4,5)P3 to drive full activation at membranes [#9]; phosphorylation downstream of the MEK-ERK pathway at Thr776 (also a GSK-3 target site that stabilizes CdGAP protein) tunes its activity [#2, #5]; and RSK phosphorylation at Ser1093/Ser1163 recruits 14-3-3\\u03b2, which inhibits GAP activity, blocks nucleocytoplasmic shuttling, and sequesters CdGAP in the cytoplasm [#15]. Protein interactions further localize and gate the enzyme: intersectin binds the central domain and inhibits GAP activity toward Rac1 [#1, #8], actopaxin (\\u03b1-parvin) and talin anchor it to focal adhesions where it controls adhesion maturation, mechanosensing, and durotaxis [#3, #11, #18], and Ajuba restrains its activity at cell-cell junctions [#14]. Independently of its catalytic function, CdGAP acts as a nuclear transcriptional co-repressor of E-cadherin, using its proline-rich domain to form a complex with Zeb2 and promote TGF-\\u03b2-driven EMT, invasion, and metastasis in ErbB2/HER2+ breast cancer [#6, #13, #18]. CdGAP also operates in defined developmental and physiological axes, controlling VEGFR2-dependent Rac1 signaling in embryonic vascular development [#12], \\u03b2-PIX-coupled signaling in podocytes [#16], and a DLG1-ARHGAP31-CDC42 axis governing intestinal stem cell survival under Wnt signaling [#17]. Gain-of-function truncating mutations in the terminal exon of ARHGAP31 increase its GAP activity, deplete active Cdc42, and cause Adams-Oliver syndrome [#7].\",\n  \"teleology\": [\n    {\n      \"year\": 1998,\n      \"claim\": \"Established CdGAP as a RhoGAP with defined substrate specificity, answering which GTPases it regulates and to which cytoskeletal events that connects.\",\n      \"evidence\": \"In vitro GAP assay and microinjection into serum-starved fibroblasts with PDGF/bradykinin/LPA stimulation\",\n      \"pmids\": [\"9786927\"],\n      \"confidence\": \"High\",\n      \"gaps\": [\"No structural basis for Rac1/Cdc42 vs RhoA discrimination\", \"Regulatory inputs to GAP activity not yet defined\", \"Function of SH3-binding motifs unknown\"]\n    },\n    {\n      \"year\": 2001,\n      \"claim\": \"Identified intersectin as a binding partner that inhibits CdGAP GAP activity, revealing the first protein-level regulatory control over the enzyme.\",\n      \"evidence\": \"Co-immunoprecipitation, in vitro GAP inhibition with intersectin SH3 domains, deletion mutant analysis in PDGF-stimulated cells\",\n      \"pmids\": [\"11744688\"],\n      \"confidence\": \"High\",\n      \"gaps\": [\"Proposed conformational change mechanism not structurally demonstrated\", \"Physiological context of inhibition unresolved\"]\n    },\n    {\n      \"year\": 2005,\n      \"claim\": \"Connected CdGAP to MEK-ERK signaling, showing serum/PDGF-induced phosphorylation at Thr776 regulates its GAP activity.\",\n      \"evidence\": \"In vitro kinase assay (ERK1, RSK-1), phospho-site mapping, DEF-domain mutagenesis, cell stimulation\",\n      \"pmids\": [\"16024771\"],\n      \"confidence\": \"High\",\n      \"gaps\": [\"Quantitative effect of Thr776 phosphorylation on catalysis not measured\", \"Functional consequence in cell behavior not fully resolved\"]\n    },\n    {\n      \"year\": 2006,\n      \"claim\": \"Defined CdGAP's focal adhesion localization and its requirement for spreading and migration, placing it within adhesion machinery.\",\n      \"evidence\": \"Co-IP/pulldown with actopaxin (\\u03b1-parvin), RNAi knockdown, GAP-deficient mutant overexpression, migration assays\",\n      \"pmids\": [\"16860736\"],\n      \"confidence\": \"High\",\n      \"gaps\": [\"Adhesion-dependent regulatory mechanism not molecularly defined\", \"Whether catalytic vs scaffold function dominates at adhesions unclear\"]\n    },\n    {\n      \"year\": 2006,\n      \"claim\": \"Showed GSK-3 phosphorylates CdGAP at Thr776 and controls its protein stability, adding a second kinase input converging on the same site.\",\n      \"evidence\": \"Yeast two-hybrid, co-IP, in vitro/in vivo phosphorylation, GSK-3 inhibitor treatment\",\n      \"pmids\": [\"17158447\"],\n      \"confidence\": \"High\",\n      \"gaps\": [\"Mechanism linking Thr776 phosphorylation to protein-level upregulation unknown\", \"Interplay between ERK and GSK-3 at Thr776 not resolved\"]\n    },\n    {\n      \"year\": 2006,\n      \"claim\": \"Confirmed substrate specificity and ERK1/2 interaction are conserved in human CdGAP, validating mouse findings in human protein.\",\n      \"evidence\": \"In vitro/in vivo GAP assay, in vivo phosphorylation, co-IP with ERK1/2, cell morphology\",\n      \"pmids\": [\"16519628\"],\n      \"confidence\": \"Medium\",\n      \"gaps\": [\"Human ERK1/2 docking region not mapped (non-canonical DEF)\", \"Single lab study\"]\n    },\n    {\n      \"year\": 2010,\n      \"claim\": \"Revealed a GAP-independent, proline-rich-domain function in TGF-\\u03b2-induced EMT and E-cadherin loss, separating catalytic from non-catalytic roles.\",\n      \"evidence\": \"siRNA knockdown, domain-specific rescue, motility/invasion assays, E-cadherin immunoblotting in mammary tumor cells\",\n      \"pmids\": [\"21042277\"],\n      \"confidence\": \"Medium\",\n      \"gaps\": [\"Molecular basis of E-cadherin regulation not yet identified\", \"Single lab\"]\n    },\n    {\n      \"year\": 2011,\n      \"claim\": \"Demonstrated gain-of-function truncating ARHGAP31 mutations cause Adams-Oliver syndrome, linking elevated GAP activity and Cdc42 depletion to a developmental disease.\",\n      \"evidence\": \"In vitro GAP assay of mutants, linkage/exome sequencing in families, active Cdc42 measurement, mouse expression analysis\",\n      \"pmids\": [\"21565291\"],\n      \"confidence\": \"High\",\n      \"gaps\": [\"Tissue-specific developmental targets of excess GAP activity unresolved\", \"Mechanism of transcript stabilization for truncated mutants unclear\"]\n    },\n    {\n      \"year\": 2011,\n      \"claim\": \"Mapped the intersectin interaction to a novel basic-rich motif in the central domain, refining the earlier interaction model.\",\n      \"evidence\": \"Pulldown and point mutagenesis of CdGAP central domain\",\n      \"pmids\": [\"21349274\"],\n      \"confidence\": \"Medium\",\n      \"gaps\": [\"Structural detail of the basic-motif/SH3D interface not determined\", \"Single lab\"]\n    },\n    {\n      \"year\": 2012,\n      \"claim\": \"Identified PI(3,4,5)P3 binding via a polybasic region as a membrane-localized activation mechanism for CdGAP.\",\n      \"evidence\": \"In vitro lipid binding, vesicle reconstitution with prenylated Rac1 and PI(3,4,5)P3, mutagenesis, Rac1 activity assay\",\n      \"pmids\": [\"22518840\"],\n      \"confidence\": \"High\",\n      \"gaps\": [\"Spatial integration of lipid signal with phosphorylation inputs not defined\", \"Upstream PI3K coupling not established\"]\n    },\n    {\n      \"year\": 2012,\n      \"claim\": \"Established that CdGAP negatively regulates migration by controlling adhesion assembly/disassembly, including in 3D matrices.\",\n      \"evidence\": \"siRNA, live-cell imaging of adhesion dynamics, 3D invasion assays\",\n      \"pmids\": [\"22907917\"],\n      \"confidence\": \"Medium\",\n      \"gaps\": [\"Direct molecular link between GAP activity and adhesion turnover unresolved\", \"Single lab\"]\n    },\n    {\n      \"year\": 2014,\n      \"claim\": \"Showed CdGAP is required for matrix-stiffness sensing and durotaxis, extending its role to mechanotransduction.\",\n      \"evidence\": \"siRNA, variable-stiffness PDMS gels, Rac1 activity assay, durotaxis assay in U2OS cells\",\n      \"pmids\": [\"24632816\"],\n      \"confidence\": \"Medium\",\n      \"gaps\": [\"Mechanosensor coupling to CdGAP activity not defined\", \"Single lab\"]\n    },\n    {\n      \"year\": 2016,\n      \"claim\": \"Defined a physiological role in embryonic vascular development through association with VEGFR2 and control of VEGF-dependent Rac1 signaling.\",\n      \"evidence\": \"CdGAP knockout mice, co-IP with VEGFR2, Rac1 activity, downstream phospho-analysis, aortic ring/migration assays\",\n      \"pmids\": [\"27270835\"],\n      \"confidence\": \"High\",\n      \"gaps\": [\"Whether VEGFR2 association is direct or scaffolded unclear\", \"Mechanism of CdGAP-dependent Gab1/Akt/PLC\\u03b3/SHP2 phosphorylation unresolved\"]\n    },\n    {\n      \"year\": 2017,\n      \"claim\": \"Identified CdGAP as a nuclear transcriptional co-repressor of E-cadherin via a proline-rich-domain complex with Zeb2, giving molecular form to its GAP-independent EMT role.\",\n      \"evidence\": \"Co-IP with Zeb2, E-cadherin promoter reporter, expression profiling, domain mapping, in vivo tumor/metastasis model\",\n      \"pmids\": [\"28135249\"],\n      \"confidence\": \"High\",\n      \"gaps\": [\"Mechanism of CdGAP nuclear import/regulation not fully defined\", \"Direct DNA contact vs Zeb2-mediated recruitment not distinguished\"]\n    },\n    {\n      \"year\": 2017,\n      \"claim\": \"Showed Ajuba binds CdGAP at cell-cell contacts and inhibits its GAP activity, adding a junctional layer of regulation.\",\n      \"evidence\": \"Co-IP, GAP activity assay, fluorescence localization, junction integrity assays with AOS mutants\",\n      \"pmids\": [\"28835688\"],\n      \"confidence\": \"Medium\",\n      \"gaps\": [\"What recruits CdGAP to junctions remains unknown\", \"Single lab\"]\n    },\n    {\n      \"year\": 2018,\n      \"claim\": \"Demonstrated RSK-mediated phosphorylation at Ser1093/Ser1163 creates 14-3-3\\u03b2 docking sites that inhibit GAP activity and block nuclear shuttling, coupling kinase signaling to both catalytic and transcriptional functions.\",\n      \"evidence\": \"In vitro RSK kinase assay, phospho-site mutagenesis, co-IP with 14-3-3\\u03b2, fractionation, migration and promoter reporter assays\",\n      \"pmids\": [\"29545927\"],\n      \"confidence\": \"High\",\n      \"gaps\": [\"Stimulus-specificity of RSK input beyond phorbol ester unclear\", \"Quantitative partitioning between cytoplasmic and nuclear pools not measured\"]\n    },\n    {\n      \"year\": 2022,\n      \"claim\": \"Identified \\u03b2-PIX (ARHGEF7) as a partner in podocytes, linking CdGAP to Src/FAK/paxillin signaling and renal physiology.\",\n      \"evidence\": \"Proximity ligation, co-IP, siRNA, live imaging, GTPase assays, podocyte-specific knockout mice\",\n      \"pmids\": [\"36333327\"],\n      \"confidence\": \"Medium\",\n      \"gaps\": [\"Mechanism coupling CdGAP loss to \\u03b2-PIX phosphorylation defects unclear\", \"Single lab\"]\n    },\n    {\n      \"year\": 2023,\n      \"claim\": \"Placed ARHGAP31 in a DLG1-ARHGAP31-CDC42 axis controlling intestinal stem cell survival under Wnt signaling, identifying transcriptional control of its expression.\",\n      \"evidence\": \"RNA-seq, conditional knockout mice, epistasis analysis, intestinal organoids\",\n      \"pmids\": [\"36640764\"],\n      \"confidence\": \"Medium\",\n      \"gaps\": [\"Mechanism of DLG1-driven ARHGAP31 transcription not defined\", \"Single lab\"]\n    },\n    {\n      \"year\": 2023,\n      \"claim\": \"Identified talin as a focal-adhesion partner modulating integrin activation and tied CdGAP to a DLC1-RhoA crosstalk in HER2+ breast cancer.\",\n      \"evidence\": \"Co-IP with talin, adhesion dynamics imaging, in vivo HER2+ mouse model, expression profiling, intravasation/extravasation assays\",\n      \"pmids\": [\"37552602\"],\n      \"confidence\": \"Medium\",\n      \"gaps\": [\"Molecular basis of CdGAP-talin-integrin coupling unresolved\", \"Single lab\"]\n    },\n    {\n      \"year\": null,\n      \"claim\": \"How CdGAP's many regulatory inputs (lipid binding, ERK/GSK-3/RSK phosphorylation, and partner interactions) are spatiotemporally integrated to switch between its catalytic GAP and nuclear co-repressor functions remains unresolved.\",\n      \"evidence\": \"\",\n      \"pmids\": [],\n      \"confidence\": \"Medium\",\n      \"gaps\": [\"No structural model of full-length CdGAP\", \"Mechanism governing cytoplasmic-vs-nuclear partitioning incompletely defined\", \"Relative contribution of catalytic vs scaffold/transcriptional roles in vivo unclear\"]\n    }\n  ],\n  \"mechanism_profile\": {\n    \"molecular_activity\": [\n      {\"term_id\": \"GO:0098772\", \"supporting_discovery_ids\": [0, 1, 7, 9]},\n      {\"term_id\": \"GO:0140110\", \"supporting_discovery_ids\": [6, 13]},\n      {\"term_id\": \"GO:0008289\", \"supporting_discovery_ids\": [9]},\n      {\"term_id\": \"GO:0008092\", \"supporting_discovery_ids\": [3, 18]}\n    ],\n    \"localization\": [\n      {\"term_id\": \"GO:0005829\", \"supporting_discovery_ids\": [15]},\n      {\"term_id\": \"GO:0005634\", \"supporting_discovery_ids\": [13, 15]},\n      {\"term_id\": \"GO:0005886\", \"supporting_discovery_ids\": [16]},\n      {\"term_id\": \"GO:0005856\", \"supporting_discovery_ids\": [3, 18]}\n    ],\n    \"pathway\": [\n      {\"term_id\": \"R-HSA-162582\", \"supporting_discovery_ids\": [12, 16]},\n      {\"term_id\": \"R-HSA-1266738\", \"supporting_discovery_ids\": [7, 12, 17]},\n      {\"term_id\": \"R-HSA-74160\", \"supporting_discovery_ids\": [13, 17]},\n      {\"term_id\": \"R-HSA-1643685\", \"supporting_discovery_ids\": [7, 13, 18]}\n    ],\n    \"complexes\": [],\n    \"partners\": [\"ITSN1\", \"PARVA\", \"ZEB2\", \"VEGFR2\", \"ARHGEF7\", \"TLN1\", \"AJUBA\", \"YWHAB\"],\n    \"other_free_text\": []\n  }\n}","audit_flag":null,"evaluation":{"pairwise":"win","faith_supported":5,"faith_total":6,"faith_pct":83.33333333333333}}