CRX is a nuclear homeodomain transcription factor that helps establish and maintain photoreceptor gene expression. Purified human CRX binds OTX regulatory sequences with competition specificity, and complete human CRX constructs activate arrestin, IRBP and OTX-containing reporter promoters. These experiments are separate from the RX localization and developmental experiments in the same paper. The original Methods and Results are available in the indexed JBC reprint, especially Methods on human CRX constructs and normalized CAT assays and Results on OTX binding and promoter activation (PMID:10625658(https://doi.org/10.1074/jbc.275.2.1152)). The repository publication cache contains its abstract; that does not negate the separately accessed original text.
The recent human organoid study supports an integrated transcription-regulation role. Its endogenous P2A-mCherry reporter marks CRX expression without making a CRX-mCherry fusion. Chromatin occupancy and RNA sequencing of sorted cell populations identify candidate targets; those population differences are not equivalent to an isogenic perturbation. Individual HEK293T promoter tests show RP1L1 activation and PCDH8 repression, while CRX-knockout organoids supply a separate causal context for impaired photoreceptor maturation and PROX1 derepression. For RP1L1, “this regulatory response was dependent on the dose of CRX”. Not all candidate targets respond in the heterologous reporter system, and a proposed intronic or exonic chromatin mechanism remains speculative. These distinctions support one RNA polymerase II DNA-binding transcription-factor core, while retaining the positive activation annotations (PMID:41418782(https://pmc.ncbi.nlm.nih.gov/articles/PMC12925966/), Results on target promoters and knockout organoids; Methods on reporter construction, CUT&Tag and luciferase assays).
The mapped NRL interaction is strong direct domain-level evidence: “the leucine zipper of NRL can physically interact with CRX.” The actual constructs are human NRL and bovine CRX, with two-hybrid controls and reciprocal deletion/pull-down experiments. This supports leucine-zipper binding without asserting an endogenous all-human complex or fixed enhanceosome stoichiometry (PMID:10887186(https://websites.umich.edu/~retina/publications/Mitton2000.pdf), printed pages 29795–29797, Methods and Results).
The FIZ1 study supplies a distinct interaction: “GST pull-down assays confirmed the direct interaction of FIZ1 and CRX”. Its bovine constructs and native retinal nuclear complexes, mouse promoter occupancy and heterologous reporter assays must remain separate. It demonstrates binding and coactivation, but does not test isolated sequence-specific FIZ1 DNA binding. Current FIZ1 DNA-binding class annotations are phylogenetic and remain unrefuted; NRL provides independent support for the transcription-factor-binding class. The FIZ1 source is not silently replaced with NRL, and an exclusion of direct DNA binding is not inferred merely to choose a coregulator-binding term (PMID:18854042(https://pmc.ncbi.nlm.nih.gov/articles/PMC2571102/), Methods on constructs and Results on the GST pull-down, nuclear association and promoter assays).
Mouse donor evidence includes Crx–ataxin-7 chromatin co-occupancy (PMID:14613968(https://pubmed.ncbi.nlm.nih.gov/14613968/)), Crx–RORbeta regulation of blue cone opsin (PMID:16574740(https://pubmed.ncbi.nlm.nih.gov/16574740/)), and Nr2e3–Crx interaction with promoter-dependent activation and repression (PMID:15689355(https://pubmed.ncbi.nlm.nih.gov/15689355/)). The latter two are different papers and different nuclear receptors. Their complete abstracts and exact current mouse annotation records were inspected; full original Methods were not recovered. Native promoter-complex evidence supports transcription-regulator-complex membership without implying one invariant assembly.
The interaction evidence contains 116 generic protein-binding assertions and a separate NRL domain-binding assertion. Exact accession and publication joins recover positive records for every generic pair. All 42 MITAB fields are retained, including positive/negative flags, assay methods, tags, fragment coordinates and partner isoform identifiers. The used records involve 98 distinct literal protein products; all 654 participant sequence checks match their recorded ROGIDs. These checks authenticate product identity, not the physiological existence of every pair or equivalence of all splice products.
The screen Methods distinguish reference interactions from mutant-dependent disruption. Repeated assay configurations and orthogonal validation of subsets do not establish independent native-complex validation for every CRX pair. Source-specific records cover the early map, HTLV interactions, alternative isoforms, HI-II-14, disease variants, alternative ORFs, developmental variants, population variants and HuRI. The original author correction to the alternative-isoform study changes an author's name only (correction notice). Fuller retrieved extraction variants were used for specified Methods while existing repository publication caches were preserved.
Where an independently supported partner class is informative, the binding term is refined: examples include sequence-specific transcription factors, cyclin C, receptor-regulated SMAD3, kinases, proteases and ubiquitin ligases. This classifies what CRX binds; it does not assign the partner's catalytic or regulatory function to CRX. PRKAB2, PPP1R16B and DELE1 are regulators of enzymes rather than the corresponding catalytic enzymes, and CA8 lacks carbonic anhydrase activity. Those distinctions prevent misleading enzyme-class refinements. Other supported generic associations remain non-core under the project's standing retention instruction. Neither an unexpected cellular compartment nor a high-throughput assay label alone invalidates a positive association.
The current PTHR45793 tree and alignment contain the exact 299-residue human CRX sequence beneath the four cited PTN000677464 ancestral assertions, without a mapped loss on its path. Complete current donor records and the three matching ARBA rules provide source provenance; they do not reconstruct the historical curator session. A donor count is not a measure of phylogenetic support, and experimentally grounded target self-evidence is legitimate. Specificity refinements preserve the positive source assertions.
The official displayed human protein is 299 residues. No numbered alternative protein product is authenticated by the selected official record. The transcript study reports seven human forms with different untranslated regions and promoter usage, which does not establish seven different proteins (PMID:12408971(https://pubmed.ncbi.nlm.nih.gov/12408971/), complete abstract). Absence of an official alternative-product block is not evidence that alternative transcripts do not exist.
The original clinical paper links CRX variants to retinal disease and maintenance: “CRX is essential for the
maintenance of mammalian photoreceptors.” Its full Methods were not recovered, so the exact organ-morphogenesis annotation remains undecided rather than being rejected from the abstract (PMID:9390563(https://pubmed.ncbi.nlm.nih.gov/9390563/), complete abstract). Visual perception is retained as a non-core transcriptional-program contribution, not photon capture by CRX. The general SELEX source is available, but the CRX-specific supplementary model/construct row was not authenticated; no methylation preference or full-length tested construct is inferred from its general Methods (PMID:28473536(https://pmc.ncbi.nlm.nih.gov/articles/PMC8009048/)). Independently demonstrated human sequence-specific DNA binding remains positive.
The source evidence index records the exact annotation identities, source extraction distinctions, portable original-source locators, complete interaction and donor fields, sequence checks, current tree/rule routes and evidence limitations.
The earlier undecided organ-morphogenesis judgment is superseded by a source-explicit refinement to eye photoreceptor cell development (GO:0042462). This term lies below both organ morphogenesis and photoreceptor differentiation in the inspected ontology. Human CRX does transcriptional work in the developing cells: retinal-organoid CUT&Tag maps occupancy, separate HEK293T promoter assays demonstrate RP1L1 activation and PCDH8 repression, and CRX-knockout organoids lose maturation markers and express PROX1 ectopically in photoreceptor cells. The combination supports a developmental role beyond knockout necessity alone. Sorted-population RNA differences and occupancy do not validate every target, and no claim of complete morphological maturation is made (PMID:41418782(https://doi.org/10.1016/j.stemcr.2025.102747), Results on promoter assays and knockout organoids, Figure 6 caption and main Methods; original PMC text). The original 9390563 full text remains unavailable; its historical source is preserved, with the refinement supported independently. No redundant differentiation or maintenance NEW is added.
A current complete human/mouse/rat annotation query retrieved 790 records across these candidate processes and their API-selected descendants. The same-role comparison includes experimental mouse Nr2e3 eye-photoreceptor development, Nrl rod development and Otx2 retinal-cone differentiation. The query also returns Crx positive-regulation records through regulatory relationships; those are not misrepresented as direct developmental annotations. The comparison addresses convention, not whether every comparator paper has been independently reread. Exact records and query scope are in the source evidence index.
Of the 116 original generic-binding assertions, 46 were previously refined and 70 retained non-core; the reported total of 51 modifications also included five other annotations. After this reassessment, 45 generic assertions retain informative class refinements and 71 remain non-core. The OR6B1 interaction stays positive but non-core because the isolated screen does not establish an informative olfactory-receptor role for CRX. The olfactory-binding definition does not require receptor activation, so lack of agonism is not used to negate binding. The other refinements use two explicit premises: the measured CRX–partner association and independent evidence for the bound partner's functional class. They do not claim that CRX has the partner's activity, regulates it, or acts with it in a native retinal complex. GCM2's DNA-binding regulator class, SUFU's corepressor class and the qualified ROR2 receptor-kinase class remain informative at that limited scope; fragments, alternative sequences, NF-Y component status and phylogenetic class evidence retain their recorded limits. Supported generic associations are retained under the explicit project instruction, not removed merely for being generic.
NRL cooperation is now explicit within the single transcription-factor core. It is not a second core activity: the mapped human-NRL/bovine-CRX interaction explains one mechanism of the same transcriptional role. The stable article identifier is 10.1074/jbc.M003658200, with the actual published Methods and Results available in the author reprint, printed pages 29795–29797. The human CRX sequence-binding and reporter experiments likewise remain traceable to 10.1074/jbc.275.2.1152, original Experimental Procedures on GST constructs and CAT normalization and Results on OTX competition and promoter activation. The previously cited indexed original is genuine scientific text even though the immutable repository cache is abstract-only. No source file or cache flag was changed.
The displayed official product remains the 299-residue O43186 sequence; no alternative-products block has been invented. The existing history record is retained unchanged. Its changed-section label does not establish an alternative protein product; the new session record can distinguish review fields from the separate notes file.