Reelin signaling — RELN binding to VLDLR/LRP8, SRC-family phosphorylation of DAB1, PI3K and LIS1 effector branches, and CRL5(SOCS7) degradation of phospho-DAB1; RELN/VLDLR/LRP8/DAB1/FYN/SRC/PIK3R1/PAFAH1B1/SOCS7/CUL5/RNF7

Reelin (RELN) is a large secreted glycoprotein made by Cajal-Retzius cells in the marginal zone of the developing cortex and by other neurons later in life. It controls where radially migrating neurons stop and so sets up the layers of the cerebral cortex, hippocampus and cerebellum. Reelin binds the ectodomains of two lipoprotein receptors of the LDL receptor family, the very-low-density lipoprotein receptor (VLDLR) and apolipoprotein E receptor 2 (ApoER2, LRP8), and clusters them. The cytoplasmic tails of both receptors carry an NPxY motif bound by the phosphotyrosine-binding domain of the adaptor Disabled-1 (DAB1). Receptor clustering brings DAB1 molecules together at the inner face of the plasma membrane, where the Src-family kinases FYN and SRC phosphorylate DAB1 on tyrosines; phosphorylated DAB1 in turn activates more Src-family kinase, giving a positive feedback loop. Tyrosine-phosphorylated DAB1 is a docking hub for downstream effectors. It binds the p85-alpha regulatory subunit (PIK3R1) of class IA phosphoinositide 3-kinase, which activates AKT and inhibits GSK3B; and it binds LIS1 (PAFAH1B1), the dynein regulator mutated in lissencephaly, linking the Reelin signal to the microtubule machinery of neuronal migration. Phosphorylated DAB1 is also the substrate of the signal's own off-switch. The substrate adaptor SOCS7 recognises phospho-DAB1 and recruits it to a cullin-5 RING ubiquitin ligase (CUL5 with the RING protein RNF7/RBX2 and the Elongin B/C adaptor), which polyubiquitinates DAB1 for proteasomal degradation and so stops neurons at their target layer. In the adult brain the same receptor-DAB1-kinase core modulates NMDA and AMPA receptor function at excitatory synapses and enhances long-term potentiation. Loss of RELN, of DAB1, or of both receptors produces the reeler phenotype in mice, and RELN mutations cause lissencephaly with cerebellar hypoplasia in humans.

Derived from the RGD Pathway Ontology diagram PW:0000388 through the RGD_PATHWAY_ONTOLOGY project's diagram->module route. RGD curates 15 human genes to PW:0000388 (DAB1 FYN GRIA1 GRIN2A GRIN2B ITGA3 ITGB1 LRP8 PAFAH1B1 PAFAH1B2 PAFAH1B3 PIK3R1 RELN SRC VLDLR). Nine of them are modelled as participants here (RELN, VLDLR, LRP8, DAB1, FYN, SRC, PIK3R1, PAFAH1B1, plus the downstream AKT step by reference), and three gene products not on the RGD list (SOCS7, CUL5, RNF7) are added for signal termination. Molecular functions were chosen from each protein's experimental GO annotations in QuickGO (October 2026) and, where a gene review exists, from its core functions: LRP8 reelin receptor activity GO:0038025 (core function in the LRP8 review); VLDLR reelin receptor activity GO:0038025 (mouse Vldlr IDA/IGI/IMP, human ISS; VLDLR has no gene review yet); FYN and SRC non-membrane spanning protein tyrosine kinase activity GO:0004715 (IDA for both; core function in the SRC review; the GO-CAMs use the parent GO:0004713); DAB1 signaling adaptor activity GO:0035591 (ISS/IEA only, as in both GO-CAMs; DAB1 has no experimental MF row in human GOA and no gene review); PIK3R1 1-phosphatidylinositol-3-kinase regulator activity GO:0046935 (core in the PIK3R1 review); PAFAH1B1 cytoskeletal motor regulator activity GO:0140659 (core in the PAFAH1B1 review); SOCS7 ubiquitin-like ligase-substrate adaptor activity GO:1990756 (ISS, as in the GO-CAM); CUL5 ubiquitin ligase complex scaffold activity GO:0160072 (IDA); RNF7 ubiquitin protein ligase activity GO:0061630 (IDA). RELN carries receptor ligand activity GO:0048018, as in both GO-CAMs and accepted in the RELN review, whose core function is the sibling-branch lipoprotein particle receptor binding GO:0070325; the ligand role is the activity that matters to this module, so the GO-CAM term is used and the core-coverage warning this produces is expected. Comparator check for the boundary (CLAUDE.md, "do not add what curators deliberately declined to add"). QuickGO annotations to GO:0038026 or its child GO:0038025 in human, mouse and rat (October 2026) are carried by RELN, VLDLR, LRP8, DAB1, FYN, PAFAH1B1, CUL5, RNF7, CRK, CRKL, FKRP and STAT5A. None is carried by GRIN2A, GRIN2B, GRIA1, ITGA3, ITGB1, PAFAH1B2 or PAFAH1B3, and, notably, none by SRC or PIK3R1 either. Boundary decisions. (1) SRC is kept as a participant despite lacking GO:0038026, because the double Fyn/Src knockout almost abolishes Dab1 tyrosine phosphorylation and phenocopies dab1 mutants (PMID:16162939), and SRC itself performs the step (phosphorylating DAB1), which is participation, not substrate status. FYN is the main kinase and SRC the redundant one. The SRC GO:0038026 absence looks like a curation gap, not a convention, given that FYN carries the term by IDA in mouse. (2) PIK3R1 is kept for the same reason: Reelin triggers DAB1 binding to p85-alpha, and PI3K activity is required for Reelin-dependent cortical plate formation (PMID:12882964). Its PI3K -> PIP3 -> AKT continuation is not re-modelled; it is included by reference (module_ref) to pi3k_akt_mtor#akt_activation, the AKT step of the generic PI3K-AKT-mTOR module, which is also how RGD draws it (Pik3r1 --+> the PI3K-Akt subpathway PW:0000232). PW:0000232 is therefore mentioned here but not listed as an evidence source, since this module does not cover that pathway. (3) PAFAH1B1 (LIS1) is a participant: it binds phospho-DAB1 in a Reelin-induced manner and genetically interacts with the pathway (PMID:14578885), and it carries GO:0038026 by IMP in mouse. Its activity in this module is its regulation of dynein (GO:0140659), the activity that moves the nucleus during migration; the nucleokinesis module (modules/nucleokinesis.yaml) is the downstream neighbour that models that machinery and treats Reelin signalling as upstream. (4) PAFAH1B2 and PAFAH1B3 are excluded. They are the catalytic alpha2/alpha1 subunits of the platelet-activating factor acetylhydrolase IB heterotrimer (PAF-AH, EC 3.1.1.47), LIS1's partners in that enzyme. RGD draws them only as associated with Vldlr (Zhang et al., PLoS One 2007, an RGD reference). They carry no reelin-pathway annotation in any of the three species, they have no gene reviews, and the reported link is a single binding study. No step of the Reelin signal needs their phospholipase activity. They are recorded as candidate modulators to revisit if receptor binding is confirmed. (5) ITGA3/ITGB1 (alpha3beta1 integrin) are context, not participants. Reelin was reported to bind alpha3beta1, and loss of alpha3beta1 lowers Dab1 levels (PMID:10939329), but later work established VLDLR and ApoER2 as the receptors that are necessary and sufficient for Dab1 phosphorylation (the double receptor knockout is reeler-like, PMID:10380922), and Benhayon et al. explicitly list integrin and protocadherin as proposed alternative receptors (PMID:12670700). Curators have not annotated either integrin to GO:0038026, and neither GO-CAM includes them. RGD's integrin edges point at the separate integrin-mediated signaling subpathway PW:0000286, which is modelled in modules/integrin_fak_src_signaling.yaml; PW:0000286 is therefore mentioned here but not listed as an evidence source. (6) GRIN2A, GRIN2B and GRIA1 are excluded as downstream targets. In the adult, Reelin acting through ApoER2, DAB1 and Src-family kinases causes tyrosine phosphorylation of NMDA receptor subunits and potentiates Ca2+ influx through them (PMID:16148228, PMID:16102539); RGD also draws a PI3K route to AMPA receptors (GRIA1; Qiu et al., J Neurosci 2006, an RGD reference not cached here). In both cases the glutamate receptor is the object the pathway acts on (the kinase substrate, or the channel whose activity or trafficking is modulated), not a protein that carries the Reelin signal: it binds no Reelin, recruits no DAB1 and phosphorylates nothing. Applying the substrate-vs-participant rule, they are outputs of the module, as SREBP target genes are outputs of the SREBP module. The comparator agrees: none of the three carries GO:0038026. The adult synaptic arm is recorded in prose and as a knowledge gap, and should become a separate module (Reelin modulation of glutamatergic synaptic plasticity) built on the ionotropic glutamate receptor modules once those exist. (7) The termination arm is added from the human GO-CAM gomodel:66187e4700002704 and the primary papers (PMID:14645539, PMID:17974915, PMID:24210661). ELOB and ELOC, the Elongin B/C adaptor that links SOCS box proteins to CUL5, are obligate members of this ECS complex but are not modelled as separate annotons, following the GO-CAM, which also omits them; both have gene reviews. (8) Other branches not on the RGD diagram are left out of this draft: CRK/CRKL binding to phospho-DAB1 and activation of RAPGEF1 (C3G)/RAP1, which are annotated to GO:0038026 by IGI/IMP in mouse and are a strong candidate for a further effector part; the cadherin-related neuronal receptors (protocadherins), which the RGD text mentions as alternative Reelin receptors; and the Reelin-regulated transcription factor STAT5A. Competing receptor ligands (APOE, LRPAP1/RAP), Reelin proteolytic processing, and transcriptional control of RELN by TBR1, CASK and thyroid hormone are upstream regulation, not parts. GO-CAM and GOA observations for upstream review. Neither human GO-CAM includes VLDLR, although VLDLR is a co-equal Reelin receptor with GO:0038025 annotations. Both models enable LRP8 activity with isoform UniProtKB:Q14114-4 but take LRP8 (canonical Q14114) as RELN's input; this module grounds LRP8 on the canonical accession. In gomodel:66187e4700002704 the FYN activity and its causal edges cite PMID:16979591, a study of Src-family phosphorylation of Nogo-A that does not address Reelin or DAB1, and CUL5's process and causal edge cite PMID:34518685, a CUL5-ARIH2 structure paper that also does not address DAB1; the Reelin-specific evidence is PMID:17974915 and PMID:24210661. In GOA, CUL5 and RNF7 carry GO:0038026 but SOCS7 does not. Gene reviews exist for RELN, LRP8, SRC, PIK3R1 and PAFAH1B1; VLDLR, DAB1, FYN, SOCS7, CUL5 and RNF7 have none, and PANTHER family ids and PTNs are deliberately not asserted.