CRB1

Summary: Retain the supported product-dependent outer-segment location as a non-core context.

Reason: The old mixed mapping includes a mouse IDA record citing PMID:15316081, whose actual three-antibody and knockout-controlled experiment instead places the recognized Crb1 form in the subapical inner segment and explicitly disputes earlier cone outer-segment staining. That source conflict is real. However, PMID:32620864 subsequently authenticates a distinct CRB1-B antibody with a delB knockout and detects mouse CRB1-B in both outer and inner segments by tangential-section Western blotting after its antibody proved unsuitable for immunohistochemistry. The paper also identifies the conserved, abundant human retinal B transcript. This later positive isoform evidence supports retaining the existing gene-level orthology inference; it is not direct human protein localization, and the deposited human B sequence is not any of the five official products. Do not apply the older canonical-tail staining result to exclude all CRB1 products from outer segments. This product-dependent location is retained as non-core relative to the established canonical-tail apical-polarity function; the classification does not negate the mouse localization or the supported gene-level transfer.

Summary: Rat-to-human inner-segment transfer agrees with human retinal localization.

Reason: The combined mapping names rat D3ZZL8 and SL-0457. Its target-term donor records include PMID:25878282, whose abstract describes altered CRB1 localization at the photoreceptor–Müller junction in a mutant rat. The original rat main text was not recovered, so its detailed staining is not reconstructed. Independent human fetal retina and organoid experiments in PMID:30956116 locate CRB1 at photoreceptor inner segments and apical Müller villi and support retaining the location.

Summary: The curated mouse inner-segment transfer is corroborated on human CRB1.

Reason: Q8VHS2 carries the relevant experimental assertion citing PMID:11850625. Its available abstract explicitly identifies the inner segment of mammalian photoreceptors; no full-paper assay claim is made from that abstract. Actual later mouse/human localization in PMID:15316081 and human tissue/organoids in PMID:30956116 independently supports the transferred location.

Summary: Calcium-binding EGF repeats support the conservative domain-based assignment.

Reason: The exact InterPro sources IPR001881 and IPR018097 designate calcium-binding EGF-like domains present in CRB1. Retain this established domain inference without treating it as a measured affinity or stoichiometry for purified full-length human CRB1, and without elevating it to a separate core activity.

Summary: Human CRB1 binds the PDZ domain of PALS1.

Reason: PMID:11927608 maps CRB1 C-terminal binding to the PALS1 PDZ domain and uses domain deletion controls. The literal MINT record joins human CRB1 and human PALS1 with sequence-authenticated participants. The specific partner-domain term describes the demonstrated interaction; it does not assert equivalent binding by tail-altered products.

Proposed replacements: PDZ domain binding

Summary: The positive PATJ association is mediated through PALS1.

Reason: The original PMID:11927608 deletion experiments show that PALS1 links CRB1 and PATJ. The two literal MINT records include the complex/spoke representation. Preserve the positive contextual association without assigning CRB1 direct binding to a PATJ PDZ domain or treating the spoke as an isolated binary interface.

Summary: Bovine PALS1 provides a supported PDZ-binding partner for the CRB1 tail.

Reason: PMID:15914641 uses species-distinguished human and bovine CRB1-tail and MPP5 constructs; A0A3Q1LY00 is the verified bovine PALS1/MPP5 partner. The original experiments establish PDZ-mediated tail recognition. Current bounded IntAct/MINT retrieval did not return a literal record for this PMID, so no reconstructed interaction-record match is claimed; the original paper and official partner identity support the refinement.

Proposed replacements: PDZ domain binding

Summary: Human MPP5/PALS1 PDZ binding is more informative than generic binding.

Reason: PMID:15914641 directly tests human CRB1-tail and human MPP5 PDZ constructs. The initial human-library screen result does not negate the subsequent directed interaction. Preserve the original human partner and refine to the demonstrated domain-binding activity; current interaction downloads yielded no record for this PMID.

Proposed replacements: PDZ domain binding

Summary: CRB1 and MPP4 associate through the PALS1 bridge.

Reason: PMID:15914641 reports negative direct interaction with the MPP4 PDZ domain and recruitment through MPP5/PALS1. Retain the experimentally supported association at its complex context, without turning the MPP4 partner class into an unsupported direct PDZ-binding claim.

Summary: The bovine MPP5 association is consistent with direct PDZ recognition.

Reason: PMID:17920587 pulls MPP5 from bovine retinal material with the human CRB1 tail and independently maps the ERLI-dependent interaction to the MPP5 PDZ domain. A0A3Q1LY00 is bovine PALS1. The original species context is retained; current PMID-specific interaction responses were empty and are not presented as literal assay records.

Proposed replacements: PDZ domain binding

Summary: Bovine EPB41L5 association is supported by a mapped FERM-domain interface.

Reason: PMID:17920587 combines bovine retinal-lysate pulldown with recombinant CRB1-tail and EPB41L5 domain experiments. Q58CU2 is the bovine partner; the human retinal-cDNA constructs establish that interaction requires the EPB41L5 FERM domain and CRB1 GTY motif. GO:0019904 captures this demonstrated domain specificity without labeling the FERM interaction as PDZ binding. No direct actin linkage or membrane-adaptor activity follows from the domain name alone.

Proposed replacements: protein domain specific binding

Summary: Human PALS1 binding depends on the CRB1 PDZ-binding tail.

Reason: PMID:17920587 maps the human MPP5 interaction to its PDZ domain and shows loss of that interaction after removing CRB1 ERLI. GO:0030165 captures the demonstrated activity while the original term and participant remain unchanged.

Proposed replacements: PDZ domain binding

Summary: The aggregate PALS1–EPB41L5 source supports protein domain specific binding.

Reason: The original source names both Q8N3R9 and Q9HCM4. PMID:17920587 separates ERLI-dependent PALS1 PDZ binding from GTY-dependent EPB41L5 FERM binding by domain and tail-deletion assays. GO:0019904 encompasses both experimentally mapped domain interfaces while retaining the aggregate source. It does not assign a PDZ interaction to EPB41L5 or imply fixed ternary stoichiometry.

Proposed replacements: protein domain specific binding

Summary: Human EPB41L5 binding is mapped to its FERM domain and the CRB1 GTY motif.

Reason: PMID:17920587 uses human retinal-cDNA constructs to map CRB1-tail binding to the EPB41L5 FERM domain; loss of the CRB1 GTY motif or the FERM BC portion disrupts the interaction. GO:0019904 expresses the demonstrated domain specificity. The authors did not identify an actin-binding domain in EPB41L5, and their calcium-switch perturbation concerns EPB41L5, so no additional CRB1 actin-anchoring or adaptor activity is inferred.

Proposed replacements: protein domain specific binding

Summary: The human CRB1 C-terminal peptide binds purified human PALS1 PDZ.

Reason: PMID:33758649 directly tests a biotinylated human CRB1 1393–1406 peptide against purified PALS1 PDZ, with concentration-dependent SPR and a PALS1 F318W comparison. The literal IntAct record and both sequence ROGIDs match the official participants. This CRB1 control experiment supports PDZ domain binding; the paper’s viral-protein emphasis does not make it a wrong-gene citation.

Proposed replacements: PDZ domain binding

Summary: The secretion prediction remains unresolved at the product and assay level.

Reason: The Swiss-Prot SL-0243 mapping includes predicted secreted CRB1 products. The initially described 1376-residue form differs from the 1406-residue transmembrane form, but PMID:11734541 also discussed a possible alternative membrane segment in the shorter form. Reconstructed products differ substantially, and loss of the canonical transmembrane segment is not proof that all membrane association is absent. No direct secretion assay for the relevant official product was established in the sources read; retain uncertainty without using the extracellular domain of a membrane protein as proof of secretion.

Summary: The ancestral plasma-membrane assertion is retained.

Reason: The current PTHR24049 tree places the sequence-authenticated human target below PTN001526177 and the matching plasma-membrane IBD, with no intervening loss annotation. Mouse Crb1, Crb2 and zebrafish Crb2a are descendant context, not a donor-count test. Direct human CRB1 membrane localization independently agrees. The separately fetched PANTHER v19 PTHR12916 tree has 111 nodes and lacks the target and PTN001526177. In contrast, the protected PTHR12916 entries cache contains P82279/CRB1 in SF4; that cache is preserved rather than silently repointed.

Summary: The mouse plasma-membrane inference agrees with the target’s transmembrane form.

Reason: The literal Q8VHS2 source includes experimental retinal localization in PMID:16885194. Its full article remains inaccessible here, but the abstract describes immuno-electron microscopy and CRB1 subapical localization. Human CRB1 experiments independently support plasma-membrane localization. Retain the broad positive assertion without assigning every alternative product the canonical topology.

Summary: Apical Müller-cell villi provide a supported microvillus context.

Reason: The mixed rat and mouse mapping is retained with its literal donor records. PMID:16885194 and PMID:25878282 have limited original access, but the independently read human fetal retina and organoid study PMID:30956116 localizes CRB1 to apical Müller villi. This supports the location without treating all photoreceptor CRB1 as microvillar or mistaking neighboring apical compartments for one structure.

Summary: Human CRB1 is in the apical junctional region above the adherens junction.

Reason: The actual human retinal images and text in PMID:15914641 place CRB1 apical to the beta-catenin-positive adherens junction, with MPP5 in the subapical region. GO:0043296 preserves this junctional localization without placing CRB1 inside the cadherin–catenin junction itself. This source-scoped generalization intentionally converges with existing apical-junction-complex assertions.

Proposed replacements: apical junction complex

Summary: The matched rule and mouse source require the same regional distinction.

Reason: Current ARBA00028426 branch 5 matches the official target, but applicability alone does not settle location precision. The literal mouse IDA source PMID:15316081 explicitly separates CRB1-positive subapical region from the beta-catenin adherens junction in mouse and human. Refine to apical junction complex, preserving the mixed rule/orthology mapping and recognizing that the historical rule execution was not replayed.

Proposed replacements: apical junction complex

Summary: Broad cell communication includes the supported contact and adhesion role.

Reason: ARBA00028945 branch 316 currently matches CRB1. GO:0007154 includes interactions through attachment, not only signal transduction. The conserved adhesion/polarity contribution and actual fish extracellular-domain adhesion experiments are compatible with this broad existing term. This does not settle the narrower cell-cell signaling assertion.

Summary: Retain the supported inherited adhesion activity as a non-core context.

Reason: The matching adhesion IBD lies on the actual human target path at PTN001526177 without a recorded intervening loss. PMID:22579223 tests zebrafish Crb2a/Crb2b extracellular-fusion interactions by pairwise Sf9 coexpression followed by lysate pulldown, and extracellular-domain-dependent aggregation of transfected HEK cells. These are positive adhesion experiments, not purified binary-affinity measurements or human CRB1 constructs. Human CRB1–CRB2 association in PMID:38570189 is compatible but does not establish a trans-cellular interface: cis/trans geometry remains unresolved. Retain the supported ancestral assertion as a non-core adhesion context, while the compact core describes canonical-tail PDZ binding and apical polarity. No target-specific loss is asserted.

Summary: The polarity rule can be refined to epithelial apicobasal polarity.

Reason: Current ARBA00029058 branch 7 matches the target. Direct tail-binding, the human-tail/fly chimera experiments and human retinal organoid polarity defects support the more precise epithelial apicobasal process. This is a refinement of an existing source assertion and intentionally converges with the IBA and NAS rows; it is not a new redundant annotation.

Proposed replacements: establishment or maintenance of epithelial cell apical/basal polarity

Summary: Later experiments refine the original polarity proposal.

Reason: PMID:10508521 identified human CRB1 disease variants and proposed polarity-related function by homology. The specific epithelial apicobasal role is supported by later work, including a human cytoplasmic tail grafted into a fly Crumbs signal/transmembrane construct in PMID:11734541 and human retinal organoids in PMID:30956116. Do not describe the chimera as rescue by full-length human CRB1. The source-specific refinement intentionally converges with existing GO:0045197 rows.

Proposed replacements: establishment or maintenance of epithelial cell apical/basal polarity

Summary: The original cell-cell signaling inference exceeds the demonstrated source mechanism.

Reason: PMID:10508521 does not demonstrate information transfer between cells or an induced response in a receiving cell. The later CRB1–CRB2 association study leaves cis/trans geometry unresolved, and epithelial polarity or barrier maintenance is not by itself such a signaling assay. Mouse Pttg1-expression changes in PMID:17234588 are real but mechanistically unresolved and do not close this specific intercellular claim. This is a judgment on the narrow source-scoped assertion, not a claim that CRB1 can never influence signaling.

Summary: Apical plasma-membrane mapping is supported for the established membrane form.

Reason: The exact SL-0015 mapping agrees with human CRB1 localization in transfected epithelial cells and human retinal tissue. Retain it at gene level without asserting that all five reconstructed products share the same topology or tissue distribution.

Summary: The curated mouse apical-membrane transfer is independently corroborated in human tissue.

Reason: The mouse source retains its literal PMID:11850625 grounding and its available abstract describes the mammalian photoreceptor apical compartment. Human CRB1 localization in PMID:15914641 and PMID:30956116 supplies independent target evidence. The unavailable earlier full article is not treated as a contradiction.

Summary: The review’s apical-membrane summary agrees with primary evidence.

Reason: PMID:28424578 is a secondary synthesis. Its apical CRB1 account is supported by the original human retinal and epithelial studies and modern human organoids; it is not treated as a new independent experiment or as an exhaustive current isoform catalogue.

Summary: Current rule applicability is clear but the performed signaling role remains unresolved.

Reason: ARBA00026303 branch 313 matches the official target identifiers and taxonomy. Actual mouse experiments in PMID:17234588 show light-dependent Pttg1 transcript changes after Crb1 mutation, but not corresponding Pttg1 protein detection, and leave direct versus indirect mechanism open. These positive observations must not be dismissed as absent signaling evidence, yet the specific input-to-response work assigned to CRB1 by GO:0023052 is unresolved. Retain uncertainty about the biological scope, not about whether the current rule matches.

Summary: The ancestral protein-complex assignment fits the characterized polarity assembly.

Reason: The complex IBD is on the actual target path and includes legitimate human target evidence among descendant sources; that self-with/from is not circular. Direct CRB1–PALS1 binding, mapped bridging interactions, retinal localization and functional polarity context support the established assembly holistically. A lone co-immunoprecipitation would not satisfy the current complex definition. No dedicated Crumbs-complex identifier or fixed native stoichiometry is invented.

Summary: CRB1 participates in the experimentally characterized polarity assembly.

Reason: PMID:15914641 combines direct tail/PDZ interaction, MPP5-dependent recruitment of MPP4 and retinal colocalization. Together with the other polarity studies this supports the existing general complex assertion. It does not make MPP4 a direct CRB1 ligand or establish a unique stoichiometry or all-product equivalence.

Summary: The current complex-membership rule is biologically consistent with the known assembly.

Reason: ARBA00028902 branch 165 matches the official target. The original CRB1–PALS1 and retinal polarity-complex experiments support the general existing term; current rule matching is not represented as an independent complex assay or a replay of the historical inference.

Summary: A vertebrate subapical region should not be equated with the defined invertebrate subapical complex.

Reason: GO:0035003 explicitly defines the apical lateral region of an invertebrate epithelial cell. The mouse donor work PMID:16885194 and full PMID:17234588 concerns a retinal subapical region, and human tissue independently shows the corresponding apical junctional region. Refine to GO:0043296 rather than asserting that the donor’s positive localization is false. This source-specific replacement intentionally converges with existing apical-junction-complex rows.

Proposed replacements: apical junction complex

Summary: The broad apical-junction-complex transfer preserves the actual regional localization.

Reason: The literal mouse records include PMID:15316081, whose original distinguishes subapical CRB1 from the more basal adherens junction. GO:0043296 describes the broader functional junctional unit and is compatible with this localization and the human counterpart. Retain it without requiring CRB1 to be a cadherin–catenin adherens-junction subunit.

Summary: The review supports the appropriately broad apical junctional region.

Reason: PMID:28424578 summarizes retinal Crumbs-complex localization. Actual human retinal images and epithelial targeting experiments corroborate the regional claim; the secondary source is not substituted for those primary assays.

Summary: Apical localization is retained with the donor antibody’s paralog limitation explicit.

Reason: The mouse donor PMID:25636444 used a pan-Crb antibody recognizing a sequence shared by Crb1, Crb2 and Crb3, so its apical staining is not independently CRB1-specific. Direct human CRB1 tissue and organoid localization nevertheless supports this broad positive target location. Preserve the original transfer and its limitation rather than relabeling pan-Crb staining as an isoform-specific experiment.

Summary: The conserved apicobasal-polarity role is supported by both phylogeny and target biology.

Reason: The polarity IBD lies on the authenticated human path below PTN001526177, with no intervening recorded loss. Fish and mouse paralog experiments provide ancestral context; human CRB1 tail interactions, the carefully delimited human-tail/fly rescue and human retinal organoid defects independently support structural participation in polarity. This is an actual inherited-role assessment, not a transfer justified by donor number.

Summary: Epithelial apicobasal polarity is a well-supported central role.

Reason: The secondary synthesis PMID:28424578 agrees with primary CRB1 interaction and localization studies and human retinal organoid phenotypes. CRB1 supplies part of the membrane-associated polarity architecture; the assertion does not depend only on downstream disease or generic necessity.

Summary: Retinal architecture is a supported developmental and tissue-maintenance context.

Reason: Mouse Crb1 loss and rd8 studies PMID:12915475 and PMID:25147295 show retinal disorganization with strong age and genetic-background dependence. Human retinal organoids independently show disrupted outer-limiting-membrane organization. Retain this existing developmental annotation as non-core alongside the molecular polarity role. The later correction PMID:41403208 discloses reused control images and clarifies Methods; it is not a retraction, and the editors state that conclusions are unchanged.

Summary: Human Müller glial projections support the location despite a malformed donor citation.

Reason: The literal mouse donor record cites PMID:129154, which is not silently rewritten to PMID:12915475. This is a traceability gap in that source record. Independently read human fetal retina and organoid experiments in PMID:30956116 localize CRB1 to apical Müller villi, supporting the target location without claiming the malformed citation was resolved.