{"gene":"EPN2","run_date":"2026-04-28T17:46:03","timeline":{"discoveries":[{"year":1999,"finding":"Epsin 2 (EPN2) was cloned from human and rat brain cDNA libraries and characterized as a paralog of epsin 1. EPN2 contains an ENTH domain, interacts with the clathrin adaptor AP-2 through its central DPW-containing region, binds clathrin, and binds Eps15 through its C-terminal NPF motifs. EPN2 is enriched in brain, is present in clathrin-coated vesicle fractions, localizes to the peri-Golgi region and cell periphery, and high-level overexpression mislocalizes clathrin coat components and inhibits clathrin-mediated endocytosis.","method":"cDNA cloning, co-immunoprecipitation, pulldown assays, subcellular fractionation (clathrin-coated vesicle fraction), GFP-fusion live-cell imaging, endocytosis inhibition assay","journal":"The Journal of biological chemistry","confidence":"High","confidence_rationale":"Tier 2 — multiple orthogonal methods (Co-IP, fractionation, imaging, functional assay) in a single foundational study","pmids":["10567358"],"is_preprint":false},{"year":2009,"finding":"Combined genetic inactivation of Epn1 and Epn2 in mice causes embryonic lethality at E9.5–E10 with developmental defects recapitulating global impairment of Notch signaling. Expression of Notch primary target genes is severely reduced in double-knockout embryos, establishing EPN2 (together with EPN1) as a specialized endocytic adaptor required for Notch signaling activation in mammals. Housekeeping clathrin-mediated endocytosis is not impaired in double-knockout cells.","method":"Epn1/Epn2 double-knockout mouse generation, embryo phenotyping, Notch target gene expression analysis, endocytosis assays in knockout-derived cells","journal":"Proceedings of the National Academy of Sciences of the United States of America","confidence":"High","confidence_rationale":"Tier 2 — genetic epistasis with rigorous in vivo phenotypic readout and molecular validation, replicated findings consistent with Drosophila epsin-Notch studies","pmids":["19666558"],"is_preprint":false},{"year":2013,"finding":"Endothelial-specific deletion of EPN1 and EPN2 causes elevated VEGFR2 levels and aberrant angiogenesis during embryonic development and in adult mice. Genetic reduction of VEGFR2 by 50% (heterozygosity) rescues normal VEGF signaling, EC proliferation, EC migration, and EC network formation in vitro, and restores normal wound healing, inflammatory angiogenesis, and tumor angiogenesis in vivo, providing the first genetic demonstration that epsins function specifically to downregulate VEGFR2 by mediating activated VEGFR2 internalization and degradation.","method":"Conditional endothelial-specific Epn1/Epn2 double-knockout mice, VEGFR2 genetic reduction (Flk1 haploinsufficiency), in vitro EC proliferation/migration/tube-formation assays, in vivo wound healing/tumor angiogenesis assays","journal":"Arteriosclerosis, thrombosis, and vascular biology","confidence":"High","confidence_rationale":"Tier 2 — genetic epistasis with multiple orthogonal in vitro and in vivo functional readouts","pmids":["24311377"],"is_preprint":false},{"year":2017,"finding":"EPN2 is a direct target of the mammal-specific microRNA miR-1224. Knockdown of EPN2 stimulates capillary-like tube formation in human umbilical vein endothelial cells, while overexpression of EPN2 suppresses miR-1224-mediated stimulation of tube formation, enhanced VEGF signaling, and repressed NOTCH signaling, confirming EPN2 as a suppressor of angiogenesis acting downstream of miR-1224.","method":"Luciferase reporter assay (direct miR-1224 target validation), siRNA knockdown, EPN2 overexpression, tube-formation assay on Matrigel, VEGF/NOTCH signaling readouts","journal":"Scientific reports","confidence":"Medium","confidence_rationale":"Tier 2 — direct target validated by reporter assay with functional rescue, single lab","pmids":["28717225"],"is_preprint":false},{"year":2020,"finding":"Podocyte-specific triple knockout of Epn1, Epn2, and Epn3 in mice results in increased albuminuria and foot process effacement. Primary podocytes from knockout mice show defective cell adhesion and spreading attributable to reduced activation of Cdc42 and downstream serum response factor (SRF), leading to diminished β1 integrin expression. Loss of SRF alone phenocopies epsin loss, placing epsins upstream of the Cdc42–SRF–β1 integrin axis in podocyte function.","method":"Podocyte-specific triple-knockout mouse generation, albumin/creatinine ratio measurement, electron microscopy (foot process effacement), primary podocyte isolation, cell adhesion/spreading assays, Cdc42 activation assay, SRF and β1 integrin expression analysis, podocyte-specific Srf knockout","journal":"Journal of the American Society of Nephrology : JASN","confidence":"High","confidence_rationale":"Tier 2 — genetic epistasis (triple KO + Srf KO) with multiple orthogonal in vivo and in vitro readouts establishing pathway placement","pmids":["33051360"],"is_preprint":false},{"year":2006,"finding":"Mass spectrometry-based phosphoproteomic analysis of EGF-stimulated HeLa cells identified phosphorylation sites on EPN2, indicating that EPN2 undergoes dynamic phosphorylation in response to EGF signaling, and that phosphorylation is modulated at least 2-fold by EGF stimulation.","method":"Large-scale quantitative phosphoproteomics (SILAC-MS) of EGF-stimulated HeLa cells","journal":"Cell","confidence":"Low","confidence_rationale":"Tier 3 — mass spectrometry detection of PTM without identification of writer/eraser or functional validation of specific EPN2 sites","pmids":["17081983"],"is_preprint":false}],"current_model":"EPN2 (epsin 2) is a clathrin-mediated endocytic adaptor that binds AP-2, clathrin, and Eps15 (via NPF motifs) and localizes to clathrin-coated vesicles; together with EPN1 it acts as a specialized adaptor required for Notch receptor activation and VEGFR2 internalization/degradation, and in podocytes it sustains Cdc42–SRF–β1 integrin signaling, while its expression is post-transcriptionally suppressed by miR-1224 to promote angiogenesis."},"narrative":{"teleology":[{"year":1999,"claim":"Identification and molecular characterization of EPN2 established it as a brain-enriched epsin-1 paralog that bridges cargo to the clathrin coat through its ENTH domain, AP-2-binding DPW motifs, clathrin-binding capacity, and Eps15-binding NPF motifs, answering whether mammals possess more than one epsin isoform and defining its interaction repertoire.","evidence":"cDNA cloning, co-immunoprecipitation, pulldown assays, subcellular fractionation, GFP-fusion imaging, and endocytosis inhibition assay in mammalian cells","pmids":["10567358"],"confidence":"High","gaps":["No assessment of whether EPN2 has non-redundant functions compared with EPN1","No structural information on the ENTH domain or cargo-recognition interface","Physiological cargo specificity unresolved"]},{"year":2009,"claim":"Genetic ablation of both Epn1 and Epn2 in mice revealed that epsins are dispensable for housekeeping endocytosis but essential for ligand-dependent Notch receptor activation, resolving the long-standing question of whether mammalian epsins have a cargo-selective signaling role analogous to Drosophila liquid facets.","evidence":"Epn1/Epn2 double-knockout mouse embryos with Notch target gene expression analysis and endocytosis assays in knockout-derived cells","pmids":["19666558"],"confidence":"High","gaps":["Whether EPN2 alone is sufficient for Notch activation remains untested","Mechanism by which epsins promote Notch ligand endocytosis versus receptor processing is not defined","Tissue-specific requirements beyond embryonic development not addressed"]},{"year":2013,"claim":"Endothelial-specific deletion of EPN1/EPN2 demonstrated that epsins specifically mediate activated VEGFR2 internalization and degradation, and that aberrant VEGFR2 accumulation is the causal driver of the resulting pathological angiogenesis — establishing a second major cargo-selective endocytic role for epsins.","evidence":"Conditional endothelial Epn1/Epn2 double-knockout mice with VEGFR2 haploinsufficiency rescue, in vitro EC assays, and in vivo wound healing and tumor angiogenesis models","pmids":["24311377"],"confidence":"High","gaps":["Whether EPN2 binds VEGFR2 directly or through an intermediary adaptor is not resolved","Relative contribution of EPN2 versus EPN1 in endothelial cells unknown","Fate of internalized VEGFR2 (lysosomal vs. proteasomal degradation) not mechanistically dissected"]},{"year":2017,"claim":"EPN2 was identified as a direct target of miR-1224, linking post-transcriptional regulation of EPN2 levels to the balance between VEGF and Notch signaling outputs and providing a mechanism by which EPN2 expression tunes angiogenic responses.","evidence":"Luciferase reporter assay for direct miR-1224 targeting, siRNA knockdown, EPN2 overexpression rescue, and tube-formation assays in HUVECs","pmids":["28717225"],"confidence":"Medium","gaps":["Single-lab finding; independent confirmation of miR-1224–EPN2 axis in vivo is lacking","Quantitative relationship between EPN2 protein levels and angiogenic output not defined","Whether miR-1224 also targets EPN1 was not assessed"]},{"year":2020,"claim":"Podocyte-specific loss of all three epsins (including EPN2) revealed an unexpected role for epsins in maintaining cell adhesion through a Cdc42–SRF–β1 integrin signaling axis, extending epsin function beyond classical endocytic cargo sorting to adhesion-dependent signaling in specialized epithelial cells.","evidence":"Triple-knockout mice with albuminuria, electron microscopy, primary podocyte adhesion/spreading assays, Cdc42 activation measurement, and phenocopy by podocyte-specific Srf deletion","pmids":["33051360"],"confidence":"High","gaps":["Individual contributions of EPN1, EPN2, and EPN3 in podocytes not dissected","Mechanism linking epsin endocytic activity to Cdc42 activation remains undefined","Whether the Cdc42–SRF axis operates in non-podocyte cell types downstream of epsins is unknown"]},{"year":null,"claim":"Open question: how EPN2 distinguishes cargo-selective endocytic events (Notch ligand, VEGFR2) from housekeeping endocytosis at the molecular level, and what non-redundant functions EPN2 has versus EPN1 and EPN3, remain unresolved.","evidence":"","pmids":[],"confidence":"High","gaps":["No structural basis for cargo selectivity of EPN2","Isoform-specific knockout phenotypes for EPN2 alone are lacking","Role of EGF-stimulated phosphorylation of EPN2 is functionally uncharacterized"]}],"mechanism_profile":{"molecular_activity":[{"term_id":"GO:0060090","term_label":"molecular adaptor activity","supporting_discovery_ids":[0,1,2]},{"term_id":"GO:0038024","term_label":"cargo receptor activity","supporting_discovery_ids":[0,2]}],"localization":[{"term_id":"GO:0031410","term_label":"cytoplasmic vesicle","supporting_discovery_ids":[0]},{"term_id":"GO:0005794","term_label":"Golgi apparatus","supporting_discovery_ids":[0]}],"pathway":[{"term_id":"R-HSA-5653656","term_label":"Vesicle-mediated transport","supporting_discovery_ids":[0,1,2]},{"term_id":"R-HSA-162582","term_label":"Signal Transduction","supporting_discovery_ids":[1,2,3]}],"complexes":[],"partners":["EPN1","AP2A1","CLTA","EPS15","KDR"],"other_free_text":[]},"mechanistic_narrative":"EPN2 (epsin-2) is a clathrin-mediated endocytic adaptor that, together with its paralog EPN1, functions as a specialized sorting factor linking ubiquitinated cargo to the clathrin coat machinery during receptor-mediated endocytosis. EPN2 contains an N-terminal ENTH domain and interacts with AP-2 via DPW motifs, with clathrin directly, and with Eps15 through C-terminal NPF motifs; it localizes to clathrin-coated vesicle fractions and the peri-Golgi region [PMID:10567358]. Combined loss of EPN1 and EPN2 in mice causes embryonic lethality due to impaired Notch signaling activation without disrupting housekeeping endocytosis [PMID:19666558], while endothelial-specific deletion elevates VEGFR2 levels and drives aberrant angiogenesis that is rescued by genetic reduction of VEGFR2 [PMID:24311377]. In podocytes, epsins including EPN2 sustain cell adhesion through a Cdc42–SRF–β1 integrin signaling axis, and their loss leads to foot process effacement and albuminuria [PMID:33051360]."},"prefetch_data":{"uniprot":{"accession":"O95208","full_name":"Epsin-2","aliases":["EPS-15-interacting protein 2"],"length_aa":641,"mass_kda":68.5,"function":"Plays a role in the formation of clathrin-coated invaginations and endocytosis","subcellular_location":"Cytoplasm; Cytoplasmic vesicle, clathrin-coated vesicle","url":"https://www.uniprot.org/uniprotkb/O95208/entry"},"depmap":{"release":"DepMap","has_data":true,"is_common_essential":false,"resolved_as":"","url":"https://depmap.org/portal/gene/EPN2","classification":"Not Classified","n_dependent_lines":15,"n_total_lines":1208,"dependency_fraction":0.012417218543046357},"opencell":{"profiled":true,"resolved_as":"","ensg_id":"ENSG00000072134","cell_line_id":"CID000526","localizations":[{"compartment":"cytoplasmic","grade":3},{"compartment":"vesicles","grade":2},{"compartment":"membrane","grade":1}],"interactors":[{"gene":"ARHGAP18","stoichiometry":0.2},{"gene":"CLTA","stoichiometry":0.2},{"gene":"HIST2H2AA3;HIST2H2AC","stoichiometry":0.2},{"gene":"RPS27A;UBC;UBB;UBA52","stoichiometry":0.2},{"gene":"HIST1H2BN;HIST1H2BM;HIST1H2BH;HIST2H2BF;HIST1H2BC;HIST1H2BD;HIST1H2BK;H2BFS","stoichiometry":0.2},{"gene":"H1FX","stoichiometry":0.2},{"gene":"RAD50","stoichiometry":0.2},{"gene":"SMARCA1","stoichiometry":0.2},{"gene":"PARP1","stoichiometry":0.2},{"gene":"WHSC1","stoichiometry":0.2}],"url":"https://opencell.sf.czbiohub.org/target/CID000526","total_profiled":1310},"omim":[{"mim_id":"607264","title":"EPSIN 3; EPN3","url":"https://www.omim.org/entry/607264"},{"mim_id":"607263","title":"EPSIN 2; EPN2","url":"https://www.omim.org/entry/607263"}],"hpa":{"profiled":true,"resolved_as":"","reliability":"Supported","locations":[{"location":"Vesicles","reliability":"Supported"}],"tissue_specificity":"Low tissue specificity","tissue_distribution":"Detected in all","driving_tissues":[],"url":"https://www.proteinatlas.org/search/EPN2"},"hgnc":{"alias_symbol":["KIAA1065","EHB21"],"prev_symbol":[]},"alphafold":{"accession":"O95208","domains":[{"cath_id":"1.25.40.90","chopping":"4-158","consensus_level":"high","plddt":92.9832,"start":4,"end":158}],"viewer_url":"https://alphafold.ebi.ac.uk/entry/O95208","model_url":"https://alphafold.ebi.ac.uk/files/AF-O95208-F1-model_v6.cif","pae_url":"https://alphafold.ebi.ac.uk/files/AF-O95208-F1-predicted_aligned_error_v6.png","plddt_mean":59.09},"mouse_models":{"mgi_url":"https://www.informatics.jax.org/marker/summary?nomen=EPN2","jax_strain_url":"https://www.jax.org/strain/search?query=EPN2"},"sequence":{"accession":"O95208","fasta_url":"https://rest.uniprot.org/uniprotkb/O95208.fasta","uniprot_url":"https://www.uniprot.org/uniprotkb/O95208/entry","alphafold_viewer_url":"https://alphafold.ebi.ac.uk/entry/O95208"}},"corpus_meta":[{"pmid":"19666558","id":"PMC_19666558","title":"Embryonic arrest at midgestation and disruption of Notch signaling produced by the absence of both epsin 1 and epsin 2 in mice.","date":"2009","source":"Proceedings of the National Academy of Sciences of the United States of America","url":"https://pubmed.ncbi.nlm.nih.gov/19666558","citation_count":96,"is_preprint":false,"source_track":"pubmed_title"},{"pmid":"24311377","id":"PMC_24311377","title":"Genetic reduction of vascular endothelial growth factor receptor 2 rescues aberrant angiogenesis caused by epsin deficiency.","date":"2013","source":"Arteriosclerosis, thrombosis, and vascular biology","url":"https://pubmed.ncbi.nlm.nih.gov/24311377","citation_count":42,"is_preprint":false,"source_track":"pubmed_title"},{"pmid":"31562977","id":"PMC_31562977","title":"LINC00473 promotes hepatocellular carcinoma progression via acting as a ceRNA for microRNA-195 and increasing HMGA2 expression.","date":"2019","source":"Biomedicine & pharmacotherapy = Biomedecine & pharmacotherapie","url":"https://pubmed.ncbi.nlm.nih.gov/31562977","citation_count":25,"is_preprint":false,"source_track":"pubmed_title"},{"pmid":"28717225","id":"PMC_28717225","title":"A mammalian mirtron miR-1224 promotes tube-formation of human primary endothelial cells by targeting anti-angiogenic factor epsin2.","date":"2017","source":"Scientific reports","url":"https://pubmed.ncbi.nlm.nih.gov/28717225","citation_count":16,"is_preprint":false,"source_track":"pubmed_title"},{"pmid":"35867415","id":"PMC_35867415","title":"Proteomic Identification of Potential Target Proteins of Cathepsin W for Its Development as a Drug Target for Influenza.","date":"2022","source":"Microbiology spectrum","url":"https://pubmed.ncbi.nlm.nih.gov/35867415","citation_count":12,"is_preprint":false,"source_track":"pubmed_title"},{"pmid":"33051360","id":"PMC_33051360","title":"Murine Epsins Play an Integral Role in Podocyte Function.","date":"2020","source":"Journal of the American Society of Nephrology : JASN","url":"https://pubmed.ncbi.nlm.nih.gov/33051360","citation_count":4,"is_preprint":false,"source_track":"pubmed_title"},{"pmid":"37444635","id":"PMC_37444635","title":"Dissecting the Methylomes of EGFR-Amplified Glioblastoma Reveals Altered DNA Replication and Packaging, and Chromatin and Gene Silencing Pathways.","date":"2023","source":"Cancers","url":"https://pubmed.ncbi.nlm.nih.gov/37444635","citation_count":2,"is_preprint":false,"source_track":"pubmed_title"},{"pmid":"40513938","id":"PMC_40513938","title":"Mendelian randomization analysis with the GEO database: Exploring the molecular mechanism underlying insulin therapy for perioperative neurocognitive disorders.","date":"2025","source":"European journal of pharmacology","url":"https://pubmed.ncbi.nlm.nih.gov/40513938","citation_count":1,"is_preprint":false,"source_track":"pubmed_title"},{"pmid":"17081983","id":"PMC_17081983","title":"Global, in vivo, and site-specific phosphorylation dynamics in signaling networks.","date":"2006","source":"Cell","url":"https://pubmed.ncbi.nlm.nih.gov/17081983","citation_count":2861,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"12477932","id":"PMC_12477932","title":"Generation and initial analysis of more than 15,000 full-length human and mouse cDNA sequences.","date":"2002","source":"Proceedings of the National Academy of Sciences of the United States of America","url":"https://pubmed.ncbi.nlm.nih.gov/12477932","citation_count":1479,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"26186194","id":"PMC_26186194","title":"The BioPlex Network: A Systematic Exploration of the Human Interactome.","date":"2015","source":"Cell","url":"https://pubmed.ncbi.nlm.nih.gov/26186194","citation_count":1118,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"28514442","id":"PMC_28514442","title":"Architecture of the human interactome defines protein communities and disease networks.","date":"2017","source":"Nature","url":"https://pubmed.ncbi.nlm.nih.gov/28514442","citation_count":1085,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"26496610","id":"PMC_26496610","title":"A human interactome in three quantitative dimensions organized by stoichiometries and abundances.","date":"2015","source":"Cell","url":"https://pubmed.ncbi.nlm.nih.gov/26496610","citation_count":1015,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"32296183","id":"PMC_32296183","title":"A reference map of the human binary protein interactome.","date":"2020","source":"Nature","url":"https://pubmed.ncbi.nlm.nih.gov/32296183","citation_count":849,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"12353027","id":"PMC_12353027","title":"Curvature of clathrin-coated pits driven by epsin.","date":"2002","source":"Nature","url":"https://pubmed.ncbi.nlm.nih.gov/12353027","citation_count":806,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"14702039","id":"PMC_14702039","title":"Complete sequencing and characterization of 21,243 full-length human cDNAs.","date":"2003","source":"Nature genetics","url":"https://pubmed.ncbi.nlm.nih.gov/14702039","citation_count":754,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"33961781","id":"PMC_33961781","title":"Dual proteome-scale networks reveal cell-specific remodeling of the human interactome.","date":"2021","source":"Cell","url":"https://pubmed.ncbi.nlm.nih.gov/33961781","citation_count":705,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"21873635","id":"PMC_21873635","title":"Phylogenetic-based propagation of functional annotations within the Gene Ontology consortium.","date":"2011","source":"Briefings in bioinformatics","url":"https://pubmed.ncbi.nlm.nih.gov/21873635","citation_count":656,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"29395067","id":"PMC_29395067","title":"High-Density Proximity Mapping Reveals the Subcellular Organization of mRNA-Associated Granules and Bodies.","date":"2018","source":"Molecular cell","url":"https://pubmed.ncbi.nlm.nih.gov/29395067","citation_count":580,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"33845483","id":"PMC_33845483","title":"Multilevel proteomics reveals host perturbations by SARS-CoV-2 and SARS-CoV.","date":"2021","source":"Nature","url":"https://pubmed.ncbi.nlm.nih.gov/33845483","citation_count":532,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"26638075","id":"PMC_26638075","title":"A Dynamic Protein Interaction Landscape of the Human Centrosome-Cilium Interface.","date":"2015","source":"Cell","url":"https://pubmed.ncbi.nlm.nih.gov/26638075","citation_count":433,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"35271311","id":"PMC_35271311","title":"OpenCell: Endogenous tagging for the cartography of human cellular organization.","date":"2022","source":"Science (New York, N.Y.)","url":"https://pubmed.ncbi.nlm.nih.gov/35271311","citation_count":432,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"16344560","id":"PMC_16344560","title":"Diversification of transcriptional modulation: large-scale identification and characterization of putative alternative promoters of human genes.","date":"2005","source":"Genome research","url":"https://pubmed.ncbi.nlm.nih.gov/16344560","citation_count":409,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"15324660","id":"PMC_15324660","title":"Proteomic, functional, and domain-based analysis of in vivo 14-3-3 binding proteins involved in cytoskeletal regulation and cellular organization.","date":"2004","source":"Current biology : CB","url":"https://pubmed.ncbi.nlm.nih.gov/15324660","citation_count":386,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"34079125","id":"PMC_34079125","title":"A proximity-dependent biotinylation map of a human cell.","date":"2021","source":"Nature","url":"https://pubmed.ncbi.nlm.nih.gov/34079125","citation_count":339,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"28388416","id":"PMC_28388416","title":"An Approach to Spatiotemporally Resolve Protein Interaction Networks in Living Cells.","date":"2017","source":"Cell","url":"https://pubmed.ncbi.nlm.nih.gov/28388416","citation_count":328,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"15345747","id":"PMC_15345747","title":"Phosphoproteomic analysis of the developing mouse brain.","date":"2004","source":"Molecular & cellular proteomics : MCP","url":"https://pubmed.ncbi.nlm.nih.gov/15345747","citation_count":291,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"9303539","id":"PMC_9303539","title":"Binding specificity and in vivo targets of the EH domain, a novel protein-protein interaction module.","date":"1997","source":"Genes & development","url":"https://pubmed.ncbi.nlm.nih.gov/9303539","citation_count":290,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"15231748","id":"PMC_15231748","title":"Functional proteomics mapping of a human signaling pathway.","date":"2004","source":"Genome research","url":"https://pubmed.ncbi.nlm.nih.gov/15231748","citation_count":247,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"9813051","id":"PMC_9813051","title":"Intersectin, a novel adaptor protein with two Eps15 homology and five Src homology 3 domains.","date":"1998","source":"The Journal of biological chemistry","url":"https://pubmed.ncbi.nlm.nih.gov/9813051","citation_count":239,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"27173435","id":"PMC_27173435","title":"An organelle-specific protein landscape identifies novel diseases and molecular mechanisms.","date":"2016","source":"Nature communications","url":"https://pubmed.ncbi.nlm.nih.gov/27173435","citation_count":211,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"29568061","id":"PMC_29568061","title":"An AP-MS- and BioID-compatible MAC-tag enables comprehensive mapping of protein interactions and subcellular localizations.","date":"2018","source":"Nature communications","url":"https://pubmed.ncbi.nlm.nih.gov/29568061","citation_count":201,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"10470851","id":"PMC_10470851","title":"Prediction of the coding sequences of unidentified human genes. XIV. The complete sequences of 100 new cDNA clones from brain which code for large proteins in vitro.","date":"1999","source":"DNA research : an international journal for rapid publication of reports on genes and genomes","url":"https://pubmed.ncbi.nlm.nih.gov/10470851","citation_count":162,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"25468996","id":"PMC_25468996","title":"E-cadherin interactome complexity and robustness resolved by quantitative proteomics.","date":"2014","source":"Science signaling","url":"https://pubmed.ncbi.nlm.nih.gov/25468996","citation_count":162,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"10567358","id":"PMC_10567358","title":"The epsins define a family of proteins that interact with components of the clathrin coat and contain a new protein module.","date":"1999","source":"The Journal of biological chemistry","url":"https://pubmed.ncbi.nlm.nih.gov/10567358","citation_count":154,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"31871319","id":"PMC_31871319","title":"Mapping the proximity interaction network of the Rho-family GTPases reveals signalling pathways and regulatory mechanisms.","date":"2019","source":"Nature cell biology","url":"https://pubmed.ncbi.nlm.nih.gov/31871319","citation_count":137,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"30639242","id":"PMC_30639242","title":"The Functional Proximal Proteome of Oncogenic Ras Includes mTORC2.","date":"2019","source":"Molecular cell","url":"https://pubmed.ncbi.nlm.nih.gov/30639242","citation_count":124,"is_preprint":false,"source_track":"gene2pubmed"},{"pmid":"28190767","id":"PMC_28190767","title":"An Interaction Landscape of Ubiquitin Signaling.","date":"2017","source":"Molecular cell","url":"https://pubmed.ncbi.nlm.nih.gov/28190767","citation_count":119,"is_preprint":false,"source_track":"gene2pubmed"}],"cost":{"stage1":{"model":"claude-sonnet-4-6","input_tokens":5757,"output_tokens":1182,"usd":0.0175},"stage2":{"model":"claude-opus-4-6","input_tokens":4392,"output_tokens":1758,"usd":0.098865},"total_usd":0.281027,"stage1_batch_id":"msgbatch_0118RScg3WGaGkaZRtWZbtce","stage2_batch_id":"msgbatch_01Cpfzvtbj42aqp5oYZ5Te2s","note":"batch pricing = 50% of standard","round2_stage1":{"model":"claude-sonnet-4-6","input_tokens":17900,"output_tokens":2808,"usd":0.04791},"round2_rules_fired":"R3","round2_stage2":{"model":"claude-opus-4-6","input_tokens":5422,"output_tokens":2029,"usd":0.116752}},"stage1_raw":"```json\n{\n  \"discoveries\": [\n    {\n      \"year\": 2009,\n      \"finding\": \"Epsin 1 and epsin 2 (EPN1/EPN2) function as specialized endocytic adaptors required for Notch signaling activation in mammals; combined knockout of both genes causes embryonic lethality at E9.5-E10 with phenotypes recapitulating global Notch signaling impairment and severely reduced Notch primary target gene expression, while housekeeping clathrin-mediated endocytosis remains intact in DKO cells.\",\n      \"method\": \"Genetic double knockout in mice, embryonic phenotyping, gene expression analysis of Notch target genes, endocytosis assays in DKO-derived cells\",\n      \"journal\": \"Proceedings of the National Academy of Sciences of the United States of America\",\n      \"confidence\": \"High\",\n      \"confidence_rationale\": \"Tier 2 — clean double-KO with defined developmental and molecular phenotype, replicated in Drosophila context, multiple orthogonal readouts\",\n      \"pmids\": [\"19666558\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2013,\n      \"finding\": \"EPN1 and EPN2 function specifically to mediate internalization and degradation of activated VEGFR2 in endothelial cells; endothelial-specific deletion of both epsins causes elevated VEGFR2 levels, enhanced VEGF signaling, and aberrant angiogenesis, all of which are rescued by genetic reduction of VEGFR2.\",\n      \"method\": \"Conditional endothelial-specific double knockout mice, genetic epistasis with heterozygous VEGFR2 deletion, in vitro angiogenesis assays, in vivo wound healing and tumor angiogenesis assays\",\n      \"journal\": \"Arteriosclerosis, thrombosis, and vascular biology\",\n      \"confidence\": \"High\",\n      \"confidence_rationale\": \"Tier 2 — genetic epistasis demonstrated by VEGFR2 reduction rescuing all epsin-loss phenotypes, multiple in vitro and in vivo assays\",\n      \"pmids\": [\"24311377\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2017,\n      \"finding\": \"EPN2 acts as a suppressor of angiogenesis downstream of miR-1224; miR-1224 directly targets EPN2 mRNA, knockdown of EPN2 stimulates tube formation and enhances VEGF signaling while repressing NOTCH signaling, and EPN2 overexpression rescues miR-1224-mediated tube formation stimulation.\",\n      \"method\": \"miRNA target reporter assays, EPN2 knockdown and overexpression in HUVECs, tube formation assays on Matrigel, VEGF and Notch signaling readouts\",\n      \"journal\": \"Scientific reports\",\n      \"confidence\": \"Medium\",\n      \"confidence_rationale\": \"Tier 2-3 — direct target validation with gain/loss-of-function rescue, single lab\",\n      \"pmids\": [\"28717225\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2020,\n      \"finding\": \"Epsins (including EPN2) in podocytes regulate Cdc42 activation and downstream SRF-dependent β1 integrin expression; podocyte-specific triple knockout of Epn1/Epn2/Epn3 causes albuminuria, foot process effacement, and defects in cell adhesion and spreading attributable to reduced Cdc42-SRF-β1 integrin signaling.\",\n      \"method\": \"Podocyte-specific triple-knockout mice, albuminuria measurement, electron microscopy of foot processes, primary podocyte cell adhesion/spreading assays, Cdc42 activation assays, SRF and β1 integrin expression analysis\",\n      \"journal\": \"Journal of the American Society of Nephrology : JASN\",\n      \"confidence\": \"Medium\",\n      \"confidence_rationale\": \"Tier 2 — clean tissue-specific KO with defined molecular pathway, but EPN2-specific contribution not separated from EPN1/EPN3\",\n      \"pmids\": [\"33051360\"],\n      \"is_preprint\": false\n    }\n  ],\n  \"current_model\": \"EPN2 (epsin 2) is a clathrin-mediated endocytic adaptor that functions as a specialized cargo-selective internalization factor: it is required for Notch receptor activation by mediating endocytosis critical to Notch signaling, for VEGFR2 internalization and degradation in endothelial cells (thereby suppressing angiogenesis), and in podocytes it regulates Cdc42-SRF-β1 integrin signaling through endocytic control of upstream receptors.\"\n}\n```\n\n--- ROUND 2 (R2/R3, augmented corpus) ---\n\nI'll classify each paper before extracting discoveries.\n\n**Classification:**\n\nRetrieved papers:\n1. PMID:19666558 — EPN2 double-KO mouse study → **KEEP**\n2. PMID:24311377 — Endothelial epsin (EPN1/EPN2) and VEGFR2 → **KEEP**\n3. PMID:31562977 — LINC00473/miR-195/HMGA2 in HCC; EPN2 mentioned only as an \"anti-angiogenic factor\" induced in vivo with no mechanistic study of EPN2 protein → **EXCLUDE** (expression correlation, no mechanistic EPN2 experiment)\n4. PMID:28717225 — miR-1224 targets epsin2 to regulate angiogenesis → **KEEP**\n5. PMID:35867415 — Cathepsin W substrates/influenza; EPN2 may appear in dataset but paper is about CTSW → **EXCLUDE**\n6. PMID:33051360 — Podocyte-specific Epn1/2/3 triple-KO → **KEEP**\n7. PMID:37444635 — Glioblastoma methylome; EPN2-AS1 (antisense RNA) mentioned → **EXCLUDE** (alt-locus product, case B)\n8. PMID:40513938 — Mendelian randomization/PND; EPN2 identified as hub gene computationally → **EXCLUDE** (Tier 4, computational only)\n\nGene2pubmed papers:\n- PMID:10567358 — Characterization of epsin 2 → **KEEP**\n- PMID:12353027 — Epsin 1 membrane curvature (epsin 1 focused, but establishes ENTH domain mechanism relevant to epsin family; epsin 2 not directly studied) → boundary case; paper is about epsin 1, not EPN2 specifically → **EXCLUDE** (paralog-specific)\n- PMID:17081983 — Phosphoproteomics mass spec; EPN2 phosphorylation sites detected → **KEEP** (PTM finding)\n- PMID:15345747 — Phosphoproteomics mouse brain; EPN2 phosphorylation → borderline; no specific EPN2 mechanistic finding described → **EXCLUDE**\n- PMID:9303539 — EH domain binding to NPF motifs; Eps15/Eps15R, mentions epsin family context → **KEEP** (foundational for EPN2 interaction via NPF/EH domain)\n- PMID:9813051 — Intersectin; identifies NPF-motif binding proteins → relevant to EPN2 interaction landscape → borderline; EPN2 not specifically studied → **EXCLUDE**\n- PMID:15324660 — 14-3-3 binding proteomics; EPN2 may appear → **EXCLUDE** (large proteomics, EPN2 not specifically studied)\n- PMID:26186194, 28514442, 33961781, 32296183, 26496610 — Large interactome maps; EPN2 interactions inferred by AP-MS → Low tier but include EPN2 interaction data → **KEEP** for interaction data\n- PMID:29395067, 34079125, 26638075, 35271311, 29568061, 27173435 — BioID/proximity maps; EPN2 proximity interactions → **KEEP** for interaction data\n- Others (cDNA sequencing papers, GO annotation, etc.) → **EXCLUDE** (no mechanistic findings)\n\n```json\n{\n  \"discoveries\": [\n    {\n      \"year\": 1999,\n      \"finding\": \"Epsin 2 (EPN2) was cloned from human and rat brain cDNA libraries and characterized as a paralog of epsin 1. EPN2 contains an ENTH domain, interacts with the clathrin adaptor AP-2 through its central DPW-containing region, binds clathrin, and binds Eps15 through its C-terminal NPF motifs. EPN2 is enriched in brain, is present in clathrin-coated vesicle fractions, localizes to the peri-Golgi region and cell periphery, and high-level overexpression mislocalizes clathrin coat components and inhibits clathrin-mediated endocytosis.\",\n      \"method\": \"cDNA cloning, co-immunoprecipitation, pulldown assays, subcellular fractionation (clathrin-coated vesicle fraction), GFP-fusion live-cell imaging, endocytosis inhibition assay\",\n      \"journal\": \"The Journal of biological chemistry\",\n      \"confidence\": \"High\",\n      \"confidence_rationale\": \"Tier 2 — multiple orthogonal methods (Co-IP, fractionation, imaging, functional assay) in a single foundational study\",\n      \"pmids\": [\"10567358\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2009,\n      \"finding\": \"Combined genetic inactivation of Epn1 and Epn2 in mice causes embryonic lethality at E9.5–E10 with developmental defects recapitulating global impairment of Notch signaling. Expression of Notch primary target genes is severely reduced in double-knockout embryos, establishing EPN2 (together with EPN1) as a specialized endocytic adaptor required for Notch signaling activation in mammals. Housekeeping clathrin-mediated endocytosis is not impaired in double-knockout cells.\",\n      \"method\": \"Epn1/Epn2 double-knockout mouse generation, embryo phenotyping, Notch target gene expression analysis, endocytosis assays in knockout-derived cells\",\n      \"journal\": \"Proceedings of the National Academy of Sciences of the United States of America\",\n      \"confidence\": \"High\",\n      \"confidence_rationale\": \"Tier 2 — genetic epistasis with rigorous in vivo phenotypic readout and molecular validation, replicated findings consistent with Drosophila epsin-Notch studies\",\n      \"pmids\": [\"19666558\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2013,\n      \"finding\": \"Endothelial-specific deletion of EPN1 and EPN2 causes elevated VEGFR2 levels and aberrant angiogenesis during embryonic development and in adult mice. Genetic reduction of VEGFR2 by 50% (heterozygosity) rescues normal VEGF signaling, EC proliferation, EC migration, and EC network formation in vitro, and restores normal wound healing, inflammatory angiogenesis, and tumor angiogenesis in vivo, providing the first genetic demonstration that epsins function specifically to downregulate VEGFR2 by mediating activated VEGFR2 internalization and degradation.\",\n      \"method\": \"Conditional endothelial-specific Epn1/Epn2 double-knockout mice, VEGFR2 genetic reduction (Flk1 haploinsufficiency), in vitro EC proliferation/migration/tube-formation assays, in vivo wound healing/tumor angiogenesis assays\",\n      \"journal\": \"Arteriosclerosis, thrombosis, and vascular biology\",\n      \"confidence\": \"High\",\n      \"confidence_rationale\": \"Tier 2 — genetic epistasis with multiple orthogonal in vitro and in vivo functional readouts\",\n      \"pmids\": [\"24311377\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2017,\n      \"finding\": \"EPN2 is a direct target of the mammal-specific microRNA miR-1224. Knockdown of EPN2 stimulates capillary-like tube formation in human umbilical vein endothelial cells, while overexpression of EPN2 suppresses miR-1224-mediated stimulation of tube formation, enhanced VEGF signaling, and repressed NOTCH signaling, confirming EPN2 as a suppressor of angiogenesis acting downstream of miR-1224.\",\n      \"method\": \"Luciferase reporter assay (direct miR-1224 target validation), siRNA knockdown, EPN2 overexpression, tube-formation assay on Matrigel, VEGF/NOTCH signaling readouts\",\n      \"journal\": \"Scientific reports\",\n      \"confidence\": \"Medium\",\n      \"confidence_rationale\": \"Tier 2 — direct target validated by reporter assay with functional rescue, single lab\",\n      \"pmids\": [\"28717225\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2020,\n      \"finding\": \"Podocyte-specific triple knockout of Epn1, Epn2, and Epn3 in mice results in increased albuminuria and foot process effacement. Primary podocytes from knockout mice show defective cell adhesion and spreading attributable to reduced activation of Cdc42 and downstream serum response factor (SRF), leading to diminished β1 integrin expression. Loss of SRF alone phenocopies epsin loss, placing epsins upstream of the Cdc42–SRF–β1 integrin axis in podocyte function.\",\n      \"method\": \"Podocyte-specific triple-knockout mouse generation, albumin/creatinine ratio measurement, electron microscopy (foot process effacement), primary podocyte isolation, cell adhesion/spreading assays, Cdc42 activation assay, SRF and β1 integrin expression analysis, podocyte-specific Srf knockout\",\n      \"journal\": \"Journal of the American Society of Nephrology : JASN\",\n      \"confidence\": \"High\",\n      \"confidence_rationale\": \"Tier 2 — genetic epistasis (triple KO + Srf KO) with multiple orthogonal in vivo and in vitro readouts establishing pathway placement\",\n      \"pmids\": [\"33051360\"],\n      \"is_preprint\": false\n    },\n    {\n      \"year\": 2006,\n      \"finding\": \"Mass spectrometry-based phosphoproteomic analysis of EGF-stimulated HeLa cells identified phosphorylation sites on EPN2, indicating that EPN2 undergoes dynamic phosphorylation in response to EGF signaling, and that phosphorylation is modulated at least 2-fold by EGF stimulation.\",\n      \"method\": \"Large-scale quantitative phosphoproteomics (SILAC-MS) of EGF-stimulated HeLa cells\",\n      \"journal\": \"Cell\",\n      \"confidence\": \"Low\",\n      \"confidence_rationale\": \"Tier 3 — mass spectrometry detection of PTM without identification of writer/eraser or functional validation of specific EPN2 sites\",\n      \"pmids\": [\"17081983\"],\n      \"is_preprint\": false\n    }\n  ],\n  \"current_model\": \"EPN2 (epsin 2) is a clathrin-mediated endocytic adaptor that binds AP-2, clathrin, and Eps15 (via NPF motifs) and localizes to clathrin-coated vesicles; together with EPN1 it acts as a specialized adaptor required for Notch receptor activation and VEGFR2 internalization/degradation, and in podocytes it sustains Cdc42–SRF–β1 integrin signaling, while its expression is post-transcriptionally suppressed by miR-1224 to promote angiogenesis.\"\n}\n```","stage2_raw":"```json\n{\n  \"mechanistic_narrative\": \"EPN2 (epsin 2) is a clathrin-associated endocytic adaptor that functions redundantly with epsin 1 (and epsin 3 in certain tissues) to mediate cargo-selective receptor internalization essential for developmental signaling. Combined loss of EPN1 and EPN2 in mice causes embryonic lethality with global Notch signaling impairment while leaving housekeeping endocytosis intact, establishing epsins as specialized adaptors for Notch ligand/receptor trafficking [PMID:19666558]. In endothelial cells, EPN1/EPN2 are required for internalization and degradation of activated VEGFR2; their conditional deletion elevates VEGFR2 levels and produces aberrant angiogenesis that is rescued by genetic reduction of VEGFR2, positioning epsins as negative regulators of VEGF-driven angiogenesis [PMID:24311377, PMID:28717225]. In podocytes, epsin loss disrupts Cdc42–SRF–β1 integrin signaling, causing foot process effacement and albuminuria, linking epsin-dependent endocytosis to glomerular filtration barrier integrity [PMID:33051360].\",\n  \"teleology\": [\n    {\n      \"year\": 2009,\n      \"claim\": \"The fundamental question of whether mammalian epsins are required for general endocytosis or for specific signaling pathways was resolved: EPN1/EPN2 double knockout demonstrated that epsins are dispensable for housekeeping clathrin-mediated endocytosis but essential for Notch signaling, establishing them as cargo-selective endocytic adaptors.\",\n      \"evidence\": \"Genetic double knockout in mice with embryonic phenotyping, Notch target gene expression, and endocytosis assays in DKO cells\",\n      \"pmids\": [\"19666558\"],\n      \"confidence\": \"High\",\n      \"gaps\": [\n        \"Whether EPN1 and EPN2 have non-redundant individual functions remains unresolved\",\n        \"The molecular mechanism by which epsins specifically recognize Notch pathway cargo is not defined\",\n        \"Whether epsins act on the signal-sending or signal-receiving cell for Notch activation is unclear\"\n      ]\n    },\n    {\n      \"year\": 2013,\n      \"claim\": \"Beyond Notch, the cargo repertoire of epsins was expanded to include VEGFR2: endothelial-specific EPN1/EPN2 deletion showed that epsins mediate VEGFR2 internalization and degradation, and genetic epistasis with VEGFR2 heterozygosity proved that the angiogenic phenotype is a direct consequence of VEGFR2 accumulation.\",\n      \"evidence\": \"Conditional endothelial-specific double knockout mice, genetic epistasis with heterozygous VEGFR2 deletion, in vivo angiogenesis assays\",\n      \"pmids\": [\"24311377\"],\n      \"confidence\": \"High\",\n      \"gaps\": [\n        \"The structural basis for epsin recognition of ubiquitinated VEGFR2 versus other receptors is not characterized\",\n        \"Whether epsins route VEGFR2 to lysosomal degradation or also modulate recycling is not distinguished\"\n      ]\n    },\n    {\n      \"year\": 2017,\n      \"claim\": \"EPN2 was placed downstream of miR-1224 as a direct transcriptional target that suppresses angiogenesis; EPN2 knockdown phenocopied miR-1224 overexpression by enhancing VEGF signaling and reducing Notch signaling, and EPN2 overexpression rescued miR-1224-induced tube formation.\",\n      \"evidence\": \"miRNA target reporter assays, EPN2 knockdown and overexpression in HUVECs, Matrigel tube formation assays\",\n      \"pmids\": [\"28717225\"],\n      \"confidence\": \"Medium\",\n      \"gaps\": [\n        \"Single-lab study without independent replication\",\n        \"Relative contribution of VEGFR2 versus Notch pathway modulation to the tube formation phenotype is not dissected\",\n        \"In vivo relevance of miR-1224–EPN2 axis in physiological or pathological angiogenesis is not tested\"\n      ]\n    },\n    {\n      \"year\": 2020,\n      \"claim\": \"Epsin function was extended to a non-vascular context: podocyte-specific triple knockout (Epn1/Epn2/Epn3) revealed that epsins regulate Cdc42 activation and downstream SRF–β1 integrin expression, linking endocytic adaptor function to cytoskeletal organization and glomerular filtration barrier integrity.\",\n      \"evidence\": \"Podocyte-specific triple-knockout mice, Cdc42 activation assays, β1 integrin expression analysis, electron microscopy of foot processes\",\n      \"pmids\": [\"33051360\"],\n      \"confidence\": \"Medium\",\n      \"gaps\": [\n        \"The specific contribution of EPN2 versus EPN1 and EPN3 in podocytes is not separated\",\n        \"The upstream receptor whose endocytosis controls Cdc42 in podocytes is not identified\",\n        \"Whether the Cdc42–SRF pathway is also relevant in endothelial or other epsin-dependent contexts is unknown\"\n      ]\n    },\n    {\n      \"year\": null,\n      \"claim\": \"The individual non-redundant functions of EPN2 (distinct from EPN1), the full cargo repertoire it internalizes across tissues, and the structural determinants of its cargo selectivity remain uncharacterized.\",\n      \"evidence\": \"\",\n      \"pmids\": [],\n      \"confidence\": \"Low\",\n      \"gaps\": [\n        \"No EPN2 single-knockout phenotype has been described that separates it from EPN1\",\n        \"Structural basis for cargo selectivity (e.g., ubiquitin-interacting motif specificity) is not resolved\",\n        \"Comprehensive identification of EPN2-specific cargoes beyond VEGFR2 and Notch pathway components is lacking\"\n      ]\n    }\n  ],\n  \"mechanism_profile\": {\n    \"molecular_activity\": [\n      {\"term_id\": \"GO:0060090\", \"supporting_discovery_ids\": [0, 1]},\n      {\"term_id\": \"GO:0038024\", \"supporting_discovery_ids\": [0, 1]}\n    ],\n    \"localization\": [\n      {\"term_id\": \"GO:0005886\", \"supporting_discovery_ids\": [0, 1]},\n      {\"term_id\": \"GO:0031410\", \"supporting_discovery_ids\": [1]}\n    ],\n    \"pathway\": [\n      {\"term_id\": \"R-HSA-162582\", \"supporting_discovery_ids\": [0, 1, 2]},\n      {\"term_id\": \"R-HSA-5653656\", \"supporting_discovery_ids\": [0, 1]}\n    ],\n    \"complexes\": [],\n    \"partners\": [\n      \"EPN1\",\n      \"VEGFR2\",\n      \"EPN3\"\n    ],\n    \"other_free_text\": []\n  }\n}\n```\n\n--- ROUND 2 (R2/R3, augmented corpus) ---\n\n```json\n{\n  \"mechanistic_narrative\": \"EPN2 (epsin-2) is a clathrin-mediated endocytic adaptor that, together with its paralog EPN1, functions as a specialized sorting factor linking ubiquitinated cargo to the clathrin coat machinery during receptor-mediated endocytosis. EPN2 contains an N-terminal ENTH domain and interacts with AP-2 via DPW motifs, with clathrin directly, and with Eps15 through C-terminal NPF motifs; it localizes to clathrin-coated vesicle fractions and the peri-Golgi region [PMID:10567358]. Combined loss of EPN1 and EPN2 in mice causes embryonic lethality due to impaired Notch signaling activation without disrupting housekeeping endocytosis [PMID:19666558], while endothelial-specific deletion elevates VEGFR2 levels and drives aberrant angiogenesis that is rescued by genetic reduction of VEGFR2 [PMID:24311377]. In podocytes, epsins including EPN2 sustain cell adhesion through a Cdc42–SRF–β1 integrin signaling axis, and their loss leads to foot process effacement and albuminuria [PMID:33051360].\",\n  \"teleology\": [\n    {\n      \"year\": 1999,\n      \"claim\": \"Identification and molecular characterization of EPN2 established it as a brain-enriched epsin-1 paralog that bridges cargo to the clathrin coat through its ENTH domain, AP-2-binding DPW motifs, clathrin-binding capacity, and Eps15-binding NPF motifs, answering whether mammals possess more than one epsin isoform and defining its interaction repertoire.\",\n      \"evidence\": \"cDNA cloning, co-immunoprecipitation, pulldown assays, subcellular fractionation, GFP-fusion imaging, and endocytosis inhibition assay in mammalian cells\",\n      \"pmids\": [\"10567358\"],\n      \"confidence\": \"High\",\n      \"gaps\": [\n        \"No assessment of whether EPN2 has non-redundant functions compared with EPN1\",\n        \"No structural information on the ENTH domain or cargo-recognition interface\",\n        \"Physiological cargo specificity unresolved\"\n      ]\n    },\n    {\n      \"year\": 2009,\n      \"claim\": \"Genetic ablation of both Epn1 and Epn2 in mice revealed that epsins are dispensable for housekeeping endocytosis but essential for ligand-dependent Notch receptor activation, resolving the long-standing question of whether mammalian epsins have a cargo-selective signaling role analogous to Drosophila liquid facets.\",\n      \"evidence\": \"Epn1/Epn2 double-knockout mouse embryos with Notch target gene expression analysis and endocytosis assays in knockout-derived cells\",\n      \"pmids\": [\"19666558\"],\n      \"confidence\": \"High\",\n      \"gaps\": [\n        \"Whether EPN2 alone is sufficient for Notch activation remains untested\",\n        \"Mechanism by which epsins promote Notch ligand endocytosis versus receptor processing is not defined\",\n        \"Tissue-specific requirements beyond embryonic development not addressed\"\n      ]\n    },\n    {\n      \"year\": 2013,\n      \"claim\": \"Endothelial-specific deletion of EPN1/EPN2 demonstrated that epsins specifically mediate activated VEGFR2 internalization and degradation, and that aberrant VEGFR2 accumulation is the causal driver of the resulting pathological angiogenesis — establishing a second major cargo-selective endocytic role for epsins.\",\n      \"evidence\": \"Conditional endothelial Epn1/Epn2 double-knockout mice with VEGFR2 haploinsufficiency rescue, in vitro EC assays, and in vivo wound healing and tumor angiogenesis models\",\n      \"pmids\": [\"24311377\"],\n      \"confidence\": \"High\",\n      \"gaps\": [\n        \"Whether EPN2 binds VEGFR2 directly or through an intermediary adaptor is not resolved\",\n        \"Relative contribution of EPN2 versus EPN1 in endothelial cells unknown\",\n        \"Fate of internalized VEGFR2 (lysosomal vs. proteasomal degradation) not mechanistically dissected\"\n      ]\n    },\n    {\n      \"year\": 2017,\n      \"claim\": \"EPN2 was identified as a direct target of miR-1224, linking post-transcriptional regulation of EPN2 levels to the balance between VEGF and Notch signaling outputs and providing a mechanism by which EPN2 expression tunes angiogenic responses.\",\n      \"evidence\": \"Luciferase reporter assay for direct miR-1224 targeting, siRNA knockdown, EPN2 overexpression rescue, and tube-formation assays in HUVECs\",\n      \"pmids\": [\"28717225\"],\n      \"confidence\": \"Medium\",\n      \"gaps\": [\n        \"Single-lab finding; independent confirmation of miR-1224–EPN2 axis in vivo is lacking\",\n        \"Quantitative relationship between EPN2 protein levels and angiogenic output not defined\",\n        \"Whether miR-1224 also targets EPN1 was not assessed\"\n      ]\n    },\n    {\n      \"year\": 2020,\n      \"claim\": \"Podocyte-specific loss of all three epsins (including EPN2) revealed an unexpected role for epsins in maintaining cell adhesion through a Cdc42–SRF–β1 integrin signaling axis, extending epsin function beyond classical endocytic cargo sorting to adhesion-dependent signaling in specialized epithelial cells.\",\n      \"evidence\": \"Triple-knockout mice with albuminuria, electron microscopy, primary podocyte adhesion/spreading assays, Cdc42 activation measurement, and phenocopy by podocyte-specific Srf deletion\",\n      \"pmids\": [\"33051360\"],\n      \"confidence\": \"High\",\n      \"gaps\": [\n        \"Individual contributions of EPN1, EPN2, and EPN3 in podocytes not dissected\",\n        \"Mechanism linking epsin endocytic activity to Cdc42 activation remains undefined\",\n        \"Whether the Cdc42–SRF axis operates in non-podocyte cell types downstream of epsins is unknown\"\n      ]\n    },\n    {\n      \"year\": null,\n      \"claim\": \"Open question: how EPN2 distinguishes cargo-selective endocytic events (Notch ligand, VEGFR2) from housekeeping endocytosis at the molecular level, and what non-redundant functions EPN2 has versus EPN1 and EPN3, remain unresolved.\",\n      \"evidence\": \"\",\n      \"pmids\": [],\n      \"confidence\": \"High\",\n      \"gaps\": [\n        \"No structural basis for cargo selectivity of EPN2\",\n        \"Isoform-specific knockout phenotypes for EPN2 alone are lacking\",\n        \"Role of EGF-stimulated phosphorylation of EPN2 is functionally uncharacterized\"\n      ]\n    }\n  ],\n  \"mechanism_profile\": {\n    \"molecular_activity\": [\n      {\"term_id\": \"GO:0060090\", \"supporting_discovery_ids\": [0, 1, 2]},\n      {\"term_id\": \"GO:0038024\", \"supporting_discovery_ids\": [0, 2]}\n    ],\n    \"localization\": [\n      {\"term_id\": \"GO:0031410\", \"supporting_discovery_ids\": [0]},\n      {\"term_id\": \"GO:0005794\", \"supporting_discovery_ids\": [0]}\n    ],\n    \"pathway\": [\n      {\"term_id\": \"R-HSA-5653656\", \"supporting_discovery_ids\": [0, 1, 2]},\n      {\"term_id\": \"R-HSA-162582\", \"supporting_discovery_ids\": [1, 2, 3]}\n    ],\n    \"complexes\": [],\n    \"partners\": [\"EPN1\", \"AP2A1\", \"CLTA\", \"EPS15\", \"KDR\"],\n    \"other_free_text\": []\n  }\n}\n```"}