CD2AP

Existing Annotations Review

GO Term	Evidence	Action	Reason
GO:0016477 cell migration	IBA GO_REF:0000033	KEEP AS NON CORE	Summary: CD2AP contributes to cell migration in a context-dependent manner. Reason: The PAINT node PTN008554480 records an inherited process, not a pairwise donor vote. Human CMS expression changes cell extension, whereas CD2AP stabilizes junctional actin and can restrain migration in the MDCK system. Retain the broad migration process as a contextual consequence of its cytoskeletal role; do not impose one universal direction. The original IBD tree and alignment were not reconstructed.
GO:0030674 protein-macromolecule adaptor activity	IBA GO_REF:0000033	ACCEPT	Summary: Multivalent protein-macromolecule adaptor activity is a central CD2AP function. Reason: The curated PAINT assertion agrees with target-specific assembly evidence: distinct SH3 and proline-rich regions recognize partners, and CD2AP helps organize junctional SH3BP1/CapZ complexes. This is physical partner organization, not catalytic GAP, kinase or ubiquitin-ligase activity. Uninspected IBD metadata is not evidence of a defective node.
GO:0005737 cytoplasm	IBA GO_REF:0000033	KEEP AS NON CORE	Summary: Cytoplasm is a valid broad location for the adaptor. Reason: Human CMS/CD2AP was directly characterized as cytoplasmic, with enrichment at actin-rich peripheral structures. Retain the inherited broad compartment, while the core synthesis uses the better-resolved cortical and junctional locations. No loss of the inherited localization is established.
GO:0007015 actin filament organization	IBA GO_REF:0000033	MODIFY	Summary: Direct human reconstitution supports the specific actin-organizing mechanism of barbed-end capping. Reason: The PAINT actin-organization assertion is sound; target self-inclusion in WITH/FROM is legitimate descendant evidence. Independent full-length human CD2AP reconstitution resolves a specific mechanism: washed single filaments and oriented Limulus actin seeds resist barbed-end subunit addition, and dilution/sequestration assays show reduced subunit loss. Refine the broad process to barbed-end actin filament capping using that target experiment, without claiming that the uninspected ancestral IBD asserted this narrower mechanism or that every organizational role is capping. The 351–639 construct retains filament binding but lacks full-length capping. Proposed replacements: barbed-end actin filament capping Supporting Evidence: PMID:24322428
GO:0031982 vesicle	IBA GO_REF:0000033	KEEP AS NON CORE	Summary: Vesicular localization is supported in trafficking contexts. Reason: The inherited broad vesicle compartment is compatible with the RIN3 recruitment experiments. The inspected source distinguishes HEK293T interaction/recruitment from rodent neuronal localization and identifies Rab5-positive early endosomes under its conditions. It does not establish every endosomal subtype in every cell.
GO:0005886 plasma membrane	IBA GO_REF:0000033	ACCEPT	Summary: CD2AP acts at the plasma-membrane-associated cortical cytoskeleton. Reason: The PAINT location is consistent with human CMS enrichment in membrane ruffles and leading edges and with human epithelial junctional complexes. This concerns a peripheral cytoplasmic adaptor, not an integral transmembrane protein.
GO:0031252 cell leading edge	IBA GO_REF:0000033	ACCEPT	Summary: Leading-edge enrichment is experimentally grounded for human CMS/CD2AP. Reason: The original human CMS construct colocalized with F-actin and p130Cas at leading edges in COS-7 cells. Preserve that construct and host-cell distinction and retain the corresponding inherited localization; no intrinsic motility motor activity follows.
GO:0001726 ruffle	IEA GO_REF:0000044	ACCEPT	Summary: The UniProt ruffle mapping agrees with direct human-construct imaging. Reason: GO_REF:0000044 maps the Swiss-Prot subcellular-location assertion SL-0300. PMID10339567 visualizes human CMS in actin-rich ruffles in COS-7 cells, providing target-specific support for the mapped location.
GO:0005856 cytoskeleton	IEA GO_REF:0000044	KEEP AS NON CORE	Summary: Cytoskeleton is correct but broader than the resolved actin-associated locations. Reason: The SL-0090 vocabulary mapping is consistent with the target imaging and direct F-actin experiments. Retain this broad component without converting it into a claim of binding every cytoskeletal polymer.
GO:0032991 protein-containing complex	IEA GO_REF:0000117	KEEP AS NON CORE	Summary: Participation in protein-containing complexes is well supported but nonspecific. Reason: ARBA00028902 supplies a broad electronic component assertion. CD2AP participates in several directly studied partner assemblies, so the general location is reasonable. The individual ARBA rule logic was not inspected, and no single obligatory native stoichiometry is inferred.
GO:0070161 anchoring junction	IEA GO_REF:0000044	ACCEPT	Summary: Anchoring-junction association is a central epithelial context. Reason: The SL-0038 mapping is supported by human epithelial junctional CD2AP complexes and by the separate MDCK adherens-junction actin study. These experiments resolve junctional organization without asserting that CD2AP is itself a transmembrane adhesion receptor.
GO:0005515 protein binding	IPI PMID:17020880 Atypical polyproline recognition by the CMS N-terminal Src h...	KEEP AS NON CORE	Summary: CD2 binding is supported by CD2AP SH3-domain biophysics. Reason: The cited structural study and independent NMR/ITC/SAXS work support recognition of CD2 proline-rich sequences by the CD2AP N-terminal SH3 domain. Retain the qualitative partner interaction under the project binding policy. CD2AP carrying an SH3 domain does not make this assertion SH3-domain binding in the reverse direction; fragment stoichiometries are not universal full-length stoichiometries.
GO:0005515 protein binding	IPI PMID:17020880 Atypical polyproline recognition by the CMS N-terminal Src h...	MODIFY	Summary: The CBL interaction can be represented as ubiquitin protein ligase binding. Reason: The exact P22681 edge is corroborated by the reviewed UniProt target record and its independently curated CBL interaction. PMID17020880 discusses the Cbl interaction context, but the complete original CBL-specific assay was not recovered from its abstract-only cache. The class refinement identifies the established E3-ligase partner; it neither assigns ligase catalysis to CD2AP nor claims that this original experiment has been independently reconstructed. Proposed replacements: ubiquitin protein ligase binding
GO:0005515 protein binding	IPI PMID:17020880 Atypical polyproline recognition by the CMS N-terminal Src h...	MODIFY	Summary: Direct CBLB recognition supports ubiquitin protein ligase binding. Reason: The target structural work and independent CD2AP-SH3A/Cbl-b biophysical experiments establish the CBLB interaction. The replacement identifies the partner as a ubiquitin ligase while leaving its catalytic activity with CBLB. The measurements use SH3 domains and partner peptides, not a fixed native full-length complex. Proposed replacements: ubiquitin protein ligase binding Supporting Evidence: PMID:23663663 both type I and type II dimeric complexes with Cbl-b
GO:0005515 protein binding	IPI PMID:17474147 Systematic identification of SH3 domain-mediated human prote...	MODIFY	Summary: FYN association supports protein tyrosine kinase binding. Reason: The original human SH3/peptide screen target table was not independently recovered. The exact FYN edge is curated in UniProt, and the independently read original human CMS study directly tests Fyn SH3 recognition and confirms endogenous Fyn coassociation in 293T cells. This supports the kinase-partner class, not catalytic activity or a kinase-activation function of CD2AP. Proposed replacements: protein tyrosine kinase binding Supporting Evidence: PMID:10339567 the tyrosine kinases Fyn, Src, and Yes
GO:0005515 protein binding	IPI PMID:17853893 Human ESCRT and ALIX proteins interact with proteins of the ...	KEEP AS NON CORE	Summary: The ALIX association is retained with a source-specific access boundary. Reason: The exact Q8WUM4 interaction is independently curated for CD2AP in UniProt and is consistent with the ESCRT/cytokinesis study. The canonical source is abstract-only, so its target-level interaction panel was not reread. The inspected 2012 correction changes the ROCK1 two-hybrid construct to residues 1–123 and does not state a CD2AP–ALIX correction. Retain the qualitative association without inferring an intrinsic ESCRT enzymatic activity.
GO:0005515 protein binding	IPI PMID:17853893 Human ESCRT and ALIX proteins interact with proteins of the ...	KEEP AS NON CORE	Summary: The TSG101 association is supported as a qualitative partner edge. Reason: The reviewed target record explicitly corroborates the exact Q99816 interaction from this source. Its abstract-only cache does not expose the complete target assay. Retain the experimentally curated edge under the standing binding policy, without equating association with membership in every ESCRT-I complex or with a direct isolated binary interface.
GO:0005515 protein binding	IPI PMID:19380743 Charting the molecular network of the drug target Bcr-Abl.	KEEP AS NON CORE	Summary: GRB2 coassociation in the BCR-ABL network is retained. Reason: This source is the BCR-ABL interaction-network study. The exact GRB2 edge is independently corroborated in UniProt; the original network target table remains uninspected. PMID10339567 demonstrates GRB2 N-terminal SH3 recognition of CMS in another experiment, but that mechanism is not automatically assigned to every network context.
GO:0005515 protein binding	IPI PMID:21706016 Selected reaction monitoring mass spectrometry reveals the d...	KEEP AS NON CORE	Summary: The GRB2 association is retained in its AP-SRM profiling context. Reason: PMID21706016 profiles GRB2 signaling complexes by affinity purification and selected-reaction-monitoring mass spectrometry. It is not an APS-recruitment experiment. The exact pair has independent UniProt and primary corroboration, while the original CD2AP-target transition/table was not inspected. Retain generic coassociation rather than asserting that this assay resolved the SH3 interface.
GO:0005515 protein binding	IPI PMID:21822214 The B-cell antigen receptor signals through a preformed tran...	KEEP AS NON CORE	Summary: BLNK/SLP65 association is supported by target-specific pulldown and mutation experiments. Reason: The inspected primary Results include CD2AP–SLP65 coassociation and recognition involving CD2AP SH3 domains and SLP65 PXXXPR motifs, with human and mouse experiments distinguished. The purified binary example for CIN85 and chicken DT40 redundancy experiments are not re-labelled as uniquely indispensable human CD2AP activity. Retain the BLNK interaction without assigning BLNK a domain merely because CD2AP contains one.
GO:0005515 protein binding	IPI PMID:21988832 Toward an understanding of the protein interaction network o...	KEEP AS NON CORE	Summary: The GRB2 interaction is retained despite incomplete original network-table access. Reason: The liver-network source and exact reviewed UniProt pair support qualitative retention; the specific original target table was not inspected. Its 2017 correction changes an author-name spelling only. Independent human CMS/GRB2 SH3 experiments corroborate the same edge, but no claim is made that this network independently resolves that interface.
GO:0005515 protein binding	IPI PMID:22662192 IQGAP1 interacts with components of the slit diaphragm compl...	KEEP AS NON CORE	Summary: CD2AP and IQGAP1 coassociate in human podocytes. Reason: The primary study includes target coimmunoprecipitation and proximity-ligation signals, with IQGAP1 depletion controls. These establish a cellular association; they do not by themselves prove purified binary binding or transfer IQGAP1 perturbation phenotypes to CD2AP. The exact P46940 partner remains unchanged.
GO:0005515 protein binding	IPI PMID:23663663 Multimeric and differential binding of CIN85/CD2AP with two ...	KEEP AS NON CORE	Summary: The CD2 interaction is directly supported by domain-level biophysical experiments. Reason: NMR, ITC and SAXS identify a type-II dimeric CD2AP-SH3A/CD2-peptide complex. This provides direct recognition evidence for the partner edge. Retain it as non-core binding rather than equating peptide/domain stoichiometry with the invariant organization of a native receptor complex.
GO:0005515 protein binding	IPI PMID:23663663 Multimeric and differential binding of CIN85/CD2AP with two ...	MODIFY	Summary: The CBLB interaction supports ubiquitin protein ligase binding. Reason: The original source directly distinguishes CD2AP-SH3A/Cbl-b type-I and type-II dimeric complexes from the related CIN85/Cbl-b trimer. It supports recognition of the E3 ligase partner, not ubiquitin transfer by CD2AP. Proposed replacements: ubiquitin protein ligase binding Supporting Evidence: PMID:23663663 both type I and type II dimeric complexes with Cbl-b
GO:0005515 protein binding	IPI PMID:25036637 A quantitative chaperone interaction network reveals the arc...	MODIFY	Summary: The CBL partner supports ubiquitin protein ligase binding. Reason: The chaperone-network cache is a partial HTML extraction and its exact target table was not inspected. Nonetheless, the experimentally curated P22681 edge is independently supported by the reviewed CD2AP interaction record and the target CBL literature identified there. Refine the partner class while retaining the original source-access caveat; this does not assign chaperone or ubiquitin-ligase activity to CD2AP. Proposed replacements: ubiquitin protein ligase binding
GO:0005515 protein binding	IPI PMID:28514442 Architecture of the human interactome defines protein commun...	KEEP AS NON CORE	Summary: The CIN85/SH3KBP1 interaction is retained as a supported network association. Reason: The exact Q96B97 edge is independently corroborated by the reviewed target record. BioPlex 2 is an affinity-purification interaction resource; the specific CD2AP–CIN85 target table was not independently inspected. No new domain-specific interface or catalytic role is inferred from the network edge.
GO:0005515 protein binding	IPI PMID:31413325 HENA, heterogeneous network-based data set for Alzheimer's d...	MODIFY	Summary: The FYN interaction supports protein tyrosine kinase binding. Reason: HENA is a heterogeneous integrated Alzheimer-disease network, not automatically a new independent binding assay. The exact FYN edge is curated in UniProt and directly supported by the earlier human CMS/Fyn SH3 and coassociation experiments. Refine the partner class using that independent evidence, preserving the original HENA attribution and its unresolved source-table provenance. Proposed replacements: protein tyrosine kinase binding
GO:0005515 protein binding	IPI PMID:31413325 HENA, heterogeneous network-based data set for Alzheimer's d...	KEEP AS NON CORE	Summary: The ALIX edge is retained with the integrated-network provenance explicit. Reason: HENA integrates heterogeneous source information; the original assay behind this exact edge was not reconstructed. The reviewed CD2AP record independently supports the Q8WUM4 pair and cites the ESCRT study. Retain qualitative binding, without claiming HENA itself newly demonstrated a binary interaction or extending the ROCK1 correction to CD2AP.
GO:0005515 protein binding	IPI PMID:31413325 HENA, heterogeneous network-based data set for Alzheimer's d...	KEEP AS NON CORE	Summary: The CIN85 edge is retained without treating HENA as a new physical assay. Reason: The exact Q96B97 partner is supported in the reviewed target interaction record. The integrated HENA row does not expose a newly inspected pair-level experiment here. Retain the supported association and preserve its source fields; no CIN85 activity or universal complex stoichiometry is transferred.
GO:0005515 protein binding	IPI PMID:31980649 Extensive rewiring of the EGFR network in colorectal cancer ...	KEEP AS NON CORE	Summary: The CIN85 association is retained in the EGFR-network context. Reason: This study examines network changes in KRAS(G13D)-expressing colorectal cells. Its target-specific CD2AP–CIN85 table was not inspected, while the exact pair is independently curated in UniProt. Retention represents the qualitative interaction, not proof that every KRAS-dependent signaling phenotype is a direct CD2AP function.
GO:0005515 protein binding	IPI PMID:32552912 Upregulation of RIN3 induces endosomal dysfunction in Alzhei...	KEEP AS NON CORE	Summary: RIN3–CD2AP association and recruitment are experimentally supported. Reason: The source includes target interaction assays and RIN3-dependent recruitment of CD2AP into Rab5 early-endosomal compartments. HEK293T transfection, mouse neuronal imaging and rat PC12 experiments are distinct. RIN3, not CD2AP, supplies the Rab5 exchange activity; partner association does not confer a GEF function on CD2AP.
GO:0005515 protein binding	IPI PMID:33961781 Dual proteome-scale networks reveal cell-specific remodeling...	KEEP AS NON CORE	Summary: The CIN85 edge is retained with the partial BioPlex cache limit. Reason: The canonical BioPlex 3 record is a partial HTML extraction despite its full-text flag; the specific target table was not independently inspected. Independent exact-pair UniProt corroboration supports qualitative retention under the project policy, without asserting a newly resolved interface.
GO:0005515 protein binding	IPI PMID:40205054 Multimodal cell maps as a foundation for structural and func...	KEEP AS NON CORE	Summary: The CIN85 interaction remains a supported qualitative association. Reason: The multimodal cell-map study contains broad interaction and localization resources, but the exact CD2AP–Q96B97 target assay/table was not inspected. Independent reviewed UniProt evidence supports the same pair. This does not transfer all functions of the partner or of the mapped community to CD2AP.
GO:0042802 identical protein binding	IPI PMID:17020880 Atypical polyproline recognition by the CMS N-terminal Src h...	KEEP AS NON CORE	Summary: Homotypic CD2AP association is supported but not a separate catalytic function. Reason: The cited domain-complex study and the original human CMS coiled-coil experiments support multimerization. Domain/peptide complexes and homotypic full-protein association are distinct observations; neither warrants one universal native stoichiometry in all receptor and junctional assemblies.
GO:0042802 identical protein binding	IPI PMID:31413325 HENA, heterogeneous network-based data set for Alzheimer's d...	KEEP AS NON CORE	Summary: The self-association assertion is retained with HENA provenance qualified. Reason: The reviewed target record and original CMS coiled-coil experiment independently support homotypic association. HENA is an integrated resource, and the original assay underlying its self-edge was not re-established. Retain self-binding without calling it a new independent demonstration or a universal dimeric architecture.
GO:0002102 podosome	IEA GO_REF:0000107	UNDECIDED	Summary: The specific podosome donor experiment remains unresolved. Reason: The electronic row explicitly transfers from mouse Q9JLQ0/ENSMUSP00000024709. Ruffles, leading edges and junctional actin are not sufficient to establish the distinct podosome structure. The exact donor localization assay was not recovered; retain uncertainty rather than infer that a cytoskeletal adaptor cannot occur there.
GO:0003779 actin binding	IEA GO_REF:0000107	MODIFY	Summary: Direct filament binding supports a more specific actin-binding term. Reason: The mouse-derived electronic assertion is biologically corroborated by target biochemical evidence. Purified CD2AP cosediments with F-actin, and full-length human recombinant CD2AP binds and caps filaments in the later reconstitution. Specify actin filament binding without interpreting partner-mediated CP regulation as the sole evidence. This converges on the separately retained GO:0051015 assertion as a collapse/refinement signal; it does not propose an additional NEW actin-binding annotation. Proposed replacements: actin filament binding Supporting Evidence: PMID:12217865 CD2AP directly associates with filamentous actin
GO:0005641 nuclear envelope lumen	IEA GO_REF:0000107	MARK AS OVER ANNOTATED	Summary: The electronic inference over-specifies a nuclear-envelope lumen compartment. Reason: The term denotes the space between the inner and outer nuclear membranes, continuous with the ER lumen, rather than nearby cytoplasm. CD2AP is characterized as a cytoplasmic adaptor; its reviewed sequence has neither an annotated signal peptide nor a transmembrane segment. The inspected perinuclear/nuclear-envelope distribution does not resolve luminal topology. Together these observations do not support transferring the donor annotation at this compartmental precision. This is an over-specific electronic inference, not a claim that absence of an annotated signal proves all luminal access impossible, nor a rejection of the underlying perinuclear observation. The exact original mouse mapping remains unverified.
GO:0005737 cytoplasm	IEA GO_REF:0000107	KEEP AS NON CORE	Summary: Broad cytoplasmic localization is supported independently of the donor transfer. Reason: The source is the mouse ortholog transfer. Human CMS/CD2AP is experimentally cytoplasmic and enriched in specialized peripheral structures, so the broad location is retained without requiring an identical cell-state distribution between species.
GO:0005770 late endosome	IEA GO_REF:0000107	UNDECIDED	Summary: The late-endosome-specific transfer needs its actual donor assay. Reason: The inspected RIN3 study supports Rab5 early-endosomal recruitment and distinguishes Rab7 compartments under its conditions. That selective experiment does not rule out late-endosomal CD2AP elsewhere, but it also does not verify this mouse late-endosome assertion. The original donor experiment remains uninspected.
GO:0005884 actin filament	IEA GO_REF:0000107	ACCEPT	Summary: Association with actin filaments is experimentally grounded. Reason: The mouse-derived component assertion is consistent with human CMS colocalization and purified CD2AP filament-binding/capping evidence. The human recombinant assay establishes contact with filaments, while cellular localization remains context-specific.
GO:0005886 plasma membrane	IEA GO_REF:0000107	ACCEPT	Summary: Plasma-membrane-associated cortical localization is conserved and target-supported. Reason: The mouse transfer is corroborated by human CMS enrichment at membrane ruffles/leading edges and by human epithelial junctional complexes. This is peripheral association on the cytoplasmic side; no membrane-spanning domain is implied.
GO:0005911 cell-cell junction	IEA GO_REF:0000107	ACCEPT	Summary: CD2AP contributes to cell-cell junctional complexes. Reason: The ortholog-derived location agrees with human epithelial coassociation/localization and with the separate MDCK adherens-junction actin study. These target and cellular data support the broad cell-cell junction component without assigning every junction subtype.
GO:0030276 clathrin binding	IEA GO_REF:0000107	UNDECIDED	Summary: The direct clathrin-binding donor evidence remains unresolved. Reason: Trafficking and clathrin-associated vesicle phenotypes do not identify which molecule directly binds clathrin. The exact mouse source assay behind this transfer was not recovered, and CD2AP adaptor architecture alone does not resolve it. Do not infer absence of binding or assign another partner’s interface to CD2AP.
GO:0030424 axon	IEA GO_REF:0000107	KEEP AS NON CORE	Summary: Axonal localization is retained as a neuronal context supported in rodents. Reason: The mouse-derived transfer is concordant with endogenous neuronal localization in the RIN3 study. This is a cross-species location claim with explicit mouse/rat evidence, not a direct human axon imaging assay or a universal neuronal distribution.
GO:0030425 dendrite	IEA GO_REF:0000107	KEEP AS NON CORE	Summary: Dendritic localization is supported as a rodent neuronal context. Reason: The inspected source includes endogenous CD2AP in cultured mouse neuronal processes and distinguishes the neuronal experiments from HEK293T interaction assays. Retain the ortholog-derived location as non-core; no claim of direct human dendritic imaging is added.
GO:0030426 growth cone	IEA GO_REF:0000107	UNDECIDED	Summary: The growth-cone-specific localization remains unverified. Reason: The mouse ortholog transfer names a distinct structure beyond a generic axon or dendrite. The exact growth-cone donor assay was not recovered. Broad neuronal processes and actin-rich cell leading edges do not alone settle this more specific location.
GO:0031594 neuromuscular junction	IEA GO_REF:0000107	UNDECIDED	Summary: The mouse neuromuscular-junction mapping requires further source resolution. Reason: The dated MGI graph links this donor context to PMID31412248. The inspected external primary passages distinguish Drosophila Cindr neuromuscular-junction experiments from mouse whole-brain proteostasis measurements. The complete original mouse annotation mapping and relevant supplementary evidence remain unresolved; neither the title nor this partial species separation proves miscuration.
GO:0031982 vesicle	IEA GO_REF:0000107	KEEP AS NON CORE	Summary: The broad vesicle component is retained in trafficking contexts. Reason: The ortholog transfer agrees with target-associated vesicular recruitment in the RIN3 study. Preserve the broad location without collapsing early, late and trans-Golgi compartments into an experimentally demonstrated uniform distribution.
GO:0032588 trans-Golgi network membrane	IEA GO_REF:0000107	UNDECIDED	Summary: The trans-Golgi-network-membrane donor assertion needs its target localization control. Reason: Mouse podocyte GLUT4/GGA2 trafficking is a relevant donor lead, but the exact experiment underlying this GO row was not inspected. A role in cargo return or endosomal sorting does not itself demonstrate CD2AP residence on the trans-Golgi membrane. Retain uncertainty rather than remove the location.
GO:0036057 slit diaphragm	IEA GO_REF:0000107	KEEP AS NON CORE	Summary: Slit-diaphragm association is retained as a specialized podocyte context. Reason: The exact electronic source is mouse, and reviewed UniProt likewise marks slit-diaphragm localization by similarity. Human podocyte coassociation with slit-diaphragm partners and F-actin binding are consistent with this role, but the inspected human IQGAP1 study does not independently resolve every aspect of native slit-diaphragm ultrastructure. Retain the established ortholog localization without claiming new human ultrastructural proof.
GO:0036312 phosphatidylinositol 3-kinase regulatory subunit binding	IEA GO_REF:0000107	KEEP AS NON CORE	Summary: Binding to a PI3K regulatory-subunit domain has direct target support. Reason: The original human CMS study tests the p85 SH3 domain as a binding partner for CMS proline-rich sequences. This independently supports the molecular class in the mouse-derived row. It does not assign PI3K catalytic activity, universal pathway activation, or an interface mediated by CD2AP’s own SH3 domains.
GO:0043005 neuron projection	IEA GO_REF:0000107	KEEP AS NON CORE	Summary: The broad neuron-projection location is supported in rodent cultures. Reason: The RIN3 study provides endogenous neuronal localization consistent with this ortholog transfer. Retain the broad contextual component without treating its exact human localization or a specific synapse subtype as directly imaged.
GO:0050808 synapse organization	IEA GO_REF:0000107	UNDECIDED	Summary: The synapse-organization donor step remains insufficiently resolved. Reason: The dated donor graph points to PMID31412248, whose inspected experiments distinguish fly synaptic morphology/function from mouse brain protein/proteasome measurements. The exact mouse annotation provenance and any additional relevant evidence remain unresolved. Proteostasis changes alone do not show which synapse-assembly step CD2AP performs, but incomplete access does not justify a removal.
GO:0051015 actin filament binding	IEA GO_REF:0000107	ACCEPT	Summary: Actin filament binding is directly demonstrated for CD2AP. Reason: Purified-protein F-actin cosedimentation and the later full-length human recombinant reconstitution support the mouse-derived molecular function on the human target. Filament binding persists in a C-terminal truncation that lacks the full protein’s capping activity, so these two readouts are not interchangeable. Supporting Evidence: PMID:12217865 CD2AP directly associates with filamentous actin
GO:0071944 cell periphery	IEA GO_REF:0000107	KEEP AS NON CORE	Summary: Cell periphery is a valid broad location. Reason: Human CMS enrichment at ruffles and leading edges and epithelial junctional association corroborate the broad mouse-derived component. Retain the umbrella location while using more informative components in the core synthesis.
GO:0005886 plasma membrane	IDA GO_REF:0000052	ACCEPT	Summary: The plasma-membrane immunofluorescence annotation agrees with independent target evidence. Reason: GO_REF:0000052 denotes curated immunofluorescence. The original image/antibody panel was not independently reread, but membrane-associated human CMS imaging and epithelial junction evidence support the same location. Retain with curator deference, without asserting integral membrane topology.
GO:0070161 anchoring junction	EXP PMID:22891260 Epithelial junction formation requires confinement of Cdc42 ...	ACCEPT	Summary: Human epithelial junctional experiments support anchoring-junction association. Reason: The original source shows CD2AP in a SH3BP1/JACOP/CapZ-associated junctional system in human epithelial cells. Coassociation and depletion are interpreted at their tested cellular level; they do not make CD2AP the catalytic GAP or establish direct contact with every complex member. Supporting Evidence: PMID:22891260
GO:0030424 axon	ISS PMID:32552912 Upregulation of RIN3 induces endosomal dysfunction in Alzhei...	KEEP AS NON CORE	Summary: The axonal ISS is retained with the rodent experimental boundary explicit. Reason: The cited RIN3 study includes endogenous CD2AP localization in rodent neuronal models. The source row explicitly uses mouse Q9JLQ0, which is appropriate to preserve rather than convert into a human IDA claim. This is a contextual location rather than the defining molecular activity.
GO:0030425 dendrite	ISS PMID:32552912 Upregulation of RIN3 induces endosomal dysfunction in Alzhei...	KEEP AS NON CORE	Summary: The dendritic ISS is consistent with the cited rodent localization experiments. Reason: Retain the mouse donor and original ISS provenance. Neuronal process imaging in the source is distinct from transfected HEK293T biochemical assays; the latter do not themselves demonstrate a human dendritic compartment.
GO:0050714 positive regulation of protein secretion	IMP PMID:27044754 FRMD4A-cytohesin signaling modulates the cellular release of...	UNDECIDED	Summary: The CD2AP-specific positive tau-secretion experiment remains inaccessible. Reason: The complete cached abstract describes a ten-gene screen and foregrounds FRMD4A/cytohesin signaling, but does not expose the CD2AP perturbation panel or its direction and controls. The title is not evidence that CD2AP was untested. Without that target experiment, neither retention as a positive secretion mechanism nor removal for wrong-gene attribution is justified.
GO:0045296 cadherin binding	HDA PMID:25468996 E-cadherin interactome complexity and robustness resolved by...	UNDECIDED	Summary: The exact CD2AP cadherin-association evidence remains unresolved. Reason: The canonical HTML cache is incomplete and combines final bibliographic/abstract material with an earlier manuscript narrative; its target table and load-bearing controls were not inspected. Proximity proteomics can support a cellular association without purified direct contact, so assay type alone does not prove over-annotation. Retain uncertainty about this specific binding assertion.
GO:0005515 protein binding	IPI PMID:22891260 Epithelial junction formation requires confinement of Cdc42 ...	MODIFY	Summary: CD2AP directly inhibits the actin-capping-protein heterodimer through its CPI region. Reason: The original human epithelial coimmunoprecipitation supports association with a CAPZA1-containing complex. Independent experiments establish a regulatory function beyond that association: human CD2AP CD23 peptide (485–507) binds the CP heterodimer and partially inhibits its capping activity; the extended CD30 peptide is more effective. The crystal and functional experiments use chicken CP, and the peptide interface is principally on CP beta, so neither these experiments nor the original co-IP establishes inhibition of isolated human CAPZA1. GO:0140678 describes direct noncovalent inhibition of a target activity without assigning enzyme catalysis to CP. This refinement concerns the intact CP complex and the demonstrated human CPI region; it does not assert that full-length CD2AP universally suppresses capping in cells or negate its independent direct filament-capping activity. Proposed replacements: molecular function inhibitor activity Supporting Evidence: PMID:20625546 CD23 and CK23 moderately inhibited barbed end capping by CP
GO:0007015 actin filament organization	IMP PMID:22891260 Epithelial junction formation requires confinement of Cdc42 ...	MODIFY	Summary: Human biochemical evidence identifies barbed-end capping as a specific component of CD2AP-mediated actin organization. Reason: The original human epithelial depletion and complex-assembly study supports actin organization but does not itself measure CD2AP capping. Independent full-length human recombinant CD2AP in PMID:24322428 directly caps barbed ends. This mechanistic reassessment uses that additional direct experiment to propose the specific process; it preserves the original IMP attribution and does not recast its knockdown phenotype as a capping assay. Partner assembly remains a separate adaptor core, and SH3BP1 retains its own GAP activity. Proposed replacements: barbed-end actin filament capping Supporting Evidence: PMID:22891260 PMID:24322428
GO:0051058 negative regulation of small GTPase mediated signal transduction	IMP PMID:22891260 Epithelial junction formation requires confinement of Cdc42 ...	KEEP AS NON CORE	Summary: CD2AP helps confine small-GTPase signaling through a recruited junctional complex. Reason: The source supports a contribution to the authors’ spatial-termination model through the CD2AP-associated SH3BP1 complex, not intrinsic GAP activity. CD2AP depletion reduces bulk Cdc42 activity and partially disrupts junctional recruitment of SH3BP1. Selected passages do not demonstrate CD2AP-specific spatial Cdc42 activity imaging. Retain the curated negative-regulation assertion as a contextual organizing contribution to confinement, not as a universal claim that CD2AP lowers total GTPase activation. The defining core activity is adaptor-mediated organization.
GO:0005515 protein binding	IPI PMID:22891260 Epithelial junction formation requires confinement of Cdc42 ...	KEEP AS NON CORE	Summary: The SH3BP1 association is directly supported in the junctional study. Reason: Target coimmunoprecipitation and domain-region pulldown establish CD2AP–SH3BP1 coassociation. SH3BP1 can bridge CD2AP to JACOP; the evidence does not require CD2AP to supply GAP catalysis or direct binary contact to every protein in that assembly.
GO:0005515 protein binding	IPI PMID:18641129 Differential requirements for Alix and ESCRT-III in cytokine...	KEEP AS NON CORE	Summary: ALIX association is retained with the original assay-access limit. Reason: The source concerns ALIX/ESCRT requirements in cytokinesis and HIV release. Its CD2AP-specific panel was not recovered from the abstract-only cache, but the same exact ALIX edge is independently supported by reviewed UniProt and the related primary ESCRT study. Retain qualitative binding without inferring a distinct intrinsic membrane-scission activity.
GO:0070062 extracellular exosome	HDA PMID:23533145 In-depth proteomic analyses of exosomes isolated from expres...	UNDECIDED	Summary: The target exosome-proteomics identification remains unverified. Reason: The source is an expressed-prostatic-secretion urine exosome proteomics study. The specific CD2AP peptide/protein entry and isolation controls were not inspected in the available extraction. Lack of access is not evidence that CD2AP is absent from extracellular vesicles; leave the HDA unresolved.
GO:0070062 extracellular exosome	HDA PMID:19056867 Large-scale proteomics and phosphoproteomics of urinary exos...	UNDECIDED	Summary: The urinary-exosome target identification needs its original table. Reason: The legacy canonical extraction is abstract-only despite its PMC identifier. The original CD2AP identification and fraction-purity information were not read. Do not reject extracellular-vesicle detection simply because the protein’s dominant function is intracellular, and do not accept a specific cargo identification from the paper title alone.
GO:0005515 protein binding	IPI PMID:10339567 CMS: an adapter molecule involved in cytoskeletal rearrangem...	MODIFY	Summary: BCAR1/p130Cas SH3 recognition supports a more informative binding term. Reason: The original human CMS study demonstrates direct binding of CMS proline-rich motifs to the partner p130Cas SH3 domain. Official NCBI BCAR1 identity mapping independently resolves Q6P5Z4 at the gene level; the raw accession is preserved, and no unverified historical isoform equivalence is asserted. This is CD2AP binding a partner SH3 domain, not an inference from CD2AP having SH3 domains. Proposed replacements: SH3 domain binding Supporting Evidence: PMID:10339567 this interaction is direct
GO:0017124 SH3 domain binding	IPI PMID:10339567 CMS: an adapter molecule involved in cytoskeletal rearrangem...	KEEP AS NON CORE	Summary: Recognition of the partner BCAR1 SH3 domain is directly demonstrated. Reason: The original target paper tests the p130Cas SH3 domain against CMS proline-rich sequences, including direct far-Western evidence. This is the correct interaction direction. It supplies a concrete recognition mechanism within the broader CD2AP adaptor function, without assigning SH3 recognition to every partner. Retain this supported partner-specific recognition as non-core; the broader defining activity is organizing multiple partners as an adaptor, rather than SH3-domain recognition alone. Supporting Evidence: PMID:10339567 the SH3 domain of p130Cas
GO:0001726 ruffle	IDA PMID:10339567 CMS: an adapter molecule involved in cytoskeletal rearrangem...	ACCEPT	Summary: Ruffle enrichment is observed for the human CMS construct. Reason: The cited primary study images human CMS in COS-7 cells and reports colocalization with F-actin and p130Cas at membrane ruffles. Preserve the heterologous host-cell context rather than call this endogenous human-cell microscopy. Supporting Evidence: PMID:10339567 membrane ruffles and leading edges
GO:0031941 filamentous actin	IDA PMID:10339567 CMS: an adapter molecule involved in cytoskeletal rearrangem...	ACCEPT	Summary: The filamentous-actin component is supported by cellular and purified-protein evidence. Reason: The original CMS imaging shows colocalization with filamentous actin. Later purified cosedimentation and human recombinant capping experiments independently establish physical filament association, so this component is not being justified solely by colocalization. Supporting Evidence: PMID:10339567
GO:0007165 signal transduction	NAS PMID:10339567 CMS: an adapter molecule involved in cytoskeletal rearrangem...	KEEP AS NON CORE	Summary: Signal transduction is retained as a broad contextual process. Reason: The original CMS study establishes an adaptor with multiple signaling partners and phosphorylation responses. This supports the broad NAS in a non-core role; it does not establish intrinsic kinase activity, a single obligatory pathway, or every downstream phenotype as a separate core function.
GO:0005200 structural constituent of cytoskeleton	TAS PMID:10339567 CMS: an adapter molecule involved in cytoskeletal rearrangem...	KEEP AS NON CORE	Summary: The broad cytoskeletal structural contribution is retained as non-core. Reason: The curated TAS is compatible with a contribution to cytoskeletal integrity, and the term is not restricted to polymer subunits. Direct human filament binding and capping provide a physical stabilization mechanism, whereas reduced junctional actin and weaker stressed MDCK junctions after depletion establish necessity in that cell system, not a separate structural molecular activity by themselves. Retain the broad curated contribution as non-core while representing the experimentally resolved filament binding/capping and adaptor mechanisms explicitly in the core functions. Supporting Evidence: PMID:24322428 knockdown of FSGS3/CD2AP compromised adhesive strength
GO:0005737 cytoplasm	TAS PMID:10339567 CMS: an adapter molecule involved in cytoskeletal rearrangem...	KEEP AS NON CORE	Summary: The broad cytoplasm TAS is consistent with the primary characterization. Reason: The cited source identifies CMS as cytoplasmic and enriched at ruffles and leading edges. Retain the broad compartment as valid but less informative than the resolved cortical locations.
GO:0006930 substrate-dependent cell migration, cell extension	TAS PMID:10339567 CMS: an adapter molecule involved in cytoskeletal rearrangem...	KEEP AS NON CORE	Summary: Cell extension and substrate-dependent migration are contextual CD2AP outcomes. Reason: The original human CMS expression study changes cytoskeletal arrangement and cell extension. Later junctional experiments show that CD2AP can restrain migration by stabilizing adhesion. Retain this broad TAS without imposing uniformly promigratory action or duplicating migration ancestors in the core synthesis.
GO:0015629 actin cytoskeleton	TAS PMID:10339567 CMS: an adapter molecule involved in cytoskeletal rearrangem...	ACCEPT	Summary: Actin-cytoskeleton association is a central location for CD2AP. Reason: The original colocalization evidence is corroborated by direct filament binding and human recombinant capping. This component identifies where the molecular functions act rather than inferring a motor, polymerase or GAP activity. Supporting Evidence: PMID:10339567
GO:0065003 protein-containing complex assembly	TAS PMID:10339567 CMS: an adapter molecule involved in cytoskeletal rearrangem...	ACCEPT	Summary: CD2AP performs physical work in protein-complex assembly. Reason: The original paper demonstrates distinct partner-binding regions and homotypic interactions, while the human epithelial study directly identifies CD2AP-associated junctional assemblies. CD2AP supplies the organizing contacts rather than merely being a required downstream substrate. No additional redundant NEW assembly process is needed. Supporting Evidence: PMID:10339567

Gene Description

Proposed New Ontology Terms

actin filament capping protein inhibitor activity

Existing Annotations Review

Core Functions

References

Suggested Questions for Experts

Suggested Experiments

Deep Research

Falcon

📚 Additional Documentation

Notes

📄 View Raw YAML