CRB1 is best understood as a membrane component of apical polarity assemblies. Its canonical ERLI tail binds the PDZ domain of PALS1/MPP5. PALS1 links CRB1 to PATJ and can recruit MPP4; CRB1 also binds EPB41L5 through a different tail motif. These interactions and their epithelial/retinal context support one PDZ-binding and apicobasal-polarity core. They do not make every associated protein a direct PDZ ligand, establish a fixed native stoichiometry, or demonstrate that CRB1 is itself a generic membrane adaptor.
The original human-tail experiments and the modern human tissue work matter more than a broad family description. In the full PALS1 study (PMID:11927608), the deletion controls resolve the PALS1 bridge to PATJ; a short anchor is “mediating the indirect interaction between CRB1 and PATJ”. The MPP4 study (PMID:15914641) separates direct MPP5 binding from MPP4 recruitment. The EPB41L5 study (PMID:17920587) separates GTY-dependent FERM binding from ERLI-dependent PALS1 binding. Its calcium-switch perturbation is an EPB41L5 experiment, and the authors do not identify an EPB41L5 actin-binding domain. These distinctions preserve positive interactions without adding an untested anchoring mechanism.
The human-tail rescue study (PMID:11734541) used a human cytoplasmic tail grafted into a fly Crumbs construct carrying the fly signal and transmembrane regions. Partial rescue and overexpression phenotypes support a conserved tail function; they are not rescue by full-length human CRB1. The more recent human CRB1 peptide experiment (PMID:33758649) directly measures binding to purified PALS1 PDZ with a mutant-domain comparison. Its viral-protein title does not erase this genuine CRB1 control experiment.
The authentic original PMID:15316081(https://doi.org/10.1242/jcs.01301), Results on knockout generation and Figure 3, uses three antibodies and knockout controls to place Crb1 in the subapical inner segment, explicitly contrasting earlier cone outer-segment staining. The short source anchor is “not in the outer segments of the cones”. This identifies a genuine conflict in the older source, but it does not settle all products. The later PMID:32620864 delB-controlled, isoform-specific antibody and tangential-section Western experiment positively places mouse CRB1-B in outer and inner segments. Together with conserved human retinal B transcripts, that evidence supports retaining the existing gene-level outer-segment inference. It does not directly locate human CRB1-B protein or identify B with any of the five official products. The earlier PMID:11850625(https://pubmed.ncbi.nlm.nih.gov/11850625/) remains abstract-only here; its abstract itself names mammalian inner segments.
Both the mouse/human Figure 1 in PMID:15316081 and the human retina images in PMID:15914641 distinguish CRB1-positive subapical region from the more basal beta-catenin-positive adherens junction. The broader apical junction complex is appropriate for these source rows. In addition, the current definition of GO:0035003 explicitly names an invertebrate structure; a vertebrate retinal region called subapical is not automatically that named complex. No donor organism or positive localization is discarded merely because of this terminology difference.
Human fetal retina and patient-derived organoids (PMID:30956116) show CRB1 at photoreceptor inner segments and apical Müller villi, with altered outer-limiting-membrane organization in patient material. These direct target observations support microvillus and glial-projection locations. The mouse source underlying one broad apical-cell inference, PMID:25636444(https://pubmed.ncbi.nlm.nih.gov/25636444/), used a pan-Crb antibody shared by Crb1, Crb2 and Crb3; that staining is not independently CRB1-specific. Another literal donor record cites PMID:129154, which is retained as written and is not silently repaired to PMID:12915475. Independent human localization supports the target without resolving that citation defect.
All five official sequences were reconstructed exactly from the current displayed sequence and its VSP changes: 1,406, 1,376, 1,294, 942 and 870 residues for P82279-1 through P82279-5. Products 1, 3 and 5 retain the canonical C-terminal membrane/tail region; products 2 and 4 alter or remove that region. The historical 1,376-residue form was discussed as potentially having an alternative membrane segment. Loss of the canonical segment therefore does not by itself prove secretion, and the secretion assertion remains unresolved without a product-specific release assay.
PMID:32620864(https://pmc.ncbi.nlm.nih.gov/articles/PMC7335077/) distinguishes mouse Crb1-A and a photoreceptor Crb1-B using isoform-specific genetics and biochemical/localization evidence. The paper also reports human transcript sequences. Its human A sequence matches the official 1,406-residue product, whereas the deposited human B and C proteins do not exactly match any of the five official products. The study’s B is not automatically UniProt isoform 2. These data support context-dependent CRB1 biology, not equal localization or activity of all official products.
The actual current PAINT tree and alignment place the human sequence below the stated ancestral node for plasma membrane, heterophilic adhesion, complex membership and epithelial polarity. All four corresponding IBDs are on the target path, with no recorded intervening loss. The protected older family cache is retained separately; a current family assignment does not rewrite an older selected artifact. Extant donor evidence, including the target itself in the complex assertion, is legitimate grounding rather than circularity.
The zebrafish Crb2a/Crb2b study (PMID:22579223) directly tests extracellular-domain interactions, cell aggregation and retinal organization with appropriate controls. This provides ancestral/paralog context for the existing adhesion IBA. The 2024 human CRB1–CRB2 study (PMID:38570189) adds human tagged-protein co-immunoprecipitation, human retina/organoid localization and a human CRB2-bait experiment with porcine retinal material. It leaves cis/trans geometry and oligomeric stoichiometry unresolved. Association alone is not a demonstrated human trans-adhesion interface or an intercellular signaling event.
The 1999 cell-cell signaling inference is over-specific for its actual source. By contrast, the broad signaling rule remains unresolved: PMID:17234588(https://pubmed.ncbi.nlm.nih.gov/17234588/) reports real, light-dependent Pttg1 transcript changes after Crb1 mutation, but not a corresponding detected Pttg1 protein and not a resolved direct mechanism. Broad cell communication can include attachment under its actual definition, so it is not adjudicated as if it meant only signal transduction. All five current ARBA rules match the target; a match establishes applicability, not the truth of every biological claim or a replay of the historical execution.
The 2024 bacterial-translocation study (PMID:38412859) supplies mouse evidence for Crb1-dependent retinal and intestinal epithelial barriers, including colon-targeted rescue and microbiota-related interventions. Selected actual indexed Results, Discussion, Methods and captions were read; complete Methods and supplements were not. These findings broaden epithelial context without establishing a human treatment, an autonomous immune activity or a new process solely from necessity. The human organoid paper’s failure to detect CRB1 in its fetal/hiPSC RPE context is also retained without turning it into a universal negative across ages and species.
Retinal architecture is retained as a non-core developmental/tissue-maintenance context. The mouse rd8 study PMID:25147295(https://pubmed.ncbi.nlm.nih.gov/25147295/) has a genuine correction, PMID:41403208, which discloses reused control images and clarifies animal Methods. The editors state that the correction does not alter the conclusions. This is documented rather than treated as a retraction or ignored.
The source evidence index preserves exact seeded assertions, literal donor and interaction records, current tree/rule checks, product reconstruction and portable original-source locators. Source-cache bytes remain unchanged. An original accessed separately can support the review even when the normal cache contains only an abstract; selected original passages are distinguished from full-paper or supplement access. The 12915475, 16885194, 25878282 and 11850625 originals remain incompletely accessed. Positive target claims supported independently are retained with that limitation.
Useful next experiments would distinguish native human product topology and secretion, resolve CRB1–CRB2 cis/trans interactions, and connect the positive Pttg1 transcript response to a performed molecular step. The provider research attempt timed out and its fallback returned an authentication/quota error; the assessment above is manual research, not a synthetic provider report.
The protected PTHR12916-entries.csv contains P82279/CRB1 (1,406 residues, subfamily SF4) on line 12. The absence described in the current phylogeny assessment concerns only the separately retrieved PANTHER v19 PTHR12916 tree with 111 nodes, which lacks both the target and PTN001526177. The independently resolved PTHR24049 target path and on-path IBDs remain unchanged. The two artifacts are distinct observations; the original cache remains intact.
The CRB1–EPB41L5 interaction is mapped to the CRB1 GTY motif and the EPB41L5 FERM domain. Deleting GTY disrupts EPB41L5 association while preserving PALS1 binding; deleting terminal ERLI has the converse specificity. The EPB41L5 assertions and the aggregate PALS1/EPB41L5 assertion therefore use protein domain specific binding, which covers the distinct FERM and PDZ interfaces without implying an actin bridge or a fixed ternary complex. See PMID:17920587, Results and Figures 4–5.
The PATJ and MPP4 associations remain positive, indirect complex associations: PALS1 bridges CRB1 to PATJ and recruits MPP4. The experiments do not demonstrate direct CRB1 contact with either partner's PDZ domain. These two existing generic binding assertions remain non-core under the ClinGen project's explicit binding instruction, which retains supported generic binding when a more informative replacement is not established. This project instruction intentionally differs from the general informational-exclusion recommendation; it does not convert association into a new molecular mechanism. See PMID:11927608, PALS1 deletion and complex-assembly experiments and PMID:15914641, MPP4 recruitment through MPP5.
Outer-segment localization remains a supported product-dependent context. Mouse CRB1-B detection uses a delB-controlled antibody and tangential-section Western blotting; the failure of that antibody in immunohistochemistry does not negate those results. Human CRB1-B transcript conservation is distinct from direct human protein localization and from the five official numbered products. Heterophilic adhesion likewise remains a supported inherited context: zebrafish extracellular fusion proteins were coexpressed in Sf9 cells and recovered by lysate pulldown, and extracellular-domain-dependent aggregation was tested in transfected HEK cells. Human CRB1–CRB2 association does not resolve cis/trans geometry. Both annotations are retained as non-core, with positive propagation support, while the compact core describes canonical-tail PDZ binding and apical polarity. This supersedes the earlier core classification of those two contexts, without retracting their positive evidence. See PMID:32620864, CRB1-B localization and tangential sectioning and PMID:22579223, extracellular-domain interactions, aggregation and Methods.
The five official products now have explicit sequence-based descriptions. Products 1, 3 and 5 retain the canonical transmembrane segment and GTY/ERLI tail, although the extracellular deletions in 3 and 5 preclude assuming identical native function. Product 2 replaces the C-terminal region and loses those canonical motifs; product 4 lacks the canonical membrane segment and tail. Sequence architecture alone does not establish secretion or equivalent activity. The published human CRB1-A sequence matches P82279-1; human CRB1-B and CRB1-C sequences do not match any of the five official products, so lettered and numbered names are not interchanged. See UniProt P82279, PMID:32620864(https://pubmed.ncbi.nlm.nih.gov/32620864/), and the source evidence.