CSDE1 binds RNA and connects it to regulatory proteins. Direct UNR–PABP assays and coding-region RNA binding support a protein–RNA adaptor role: PMID:15314026(https://pmc.ncbi.nlm.nih.gov/articles/PMC514181/). This does not make CSDE1 a nuclease. The c-FOS mCRD experiments identify CCR4 as the poly(A) nuclease and show that blocking initiation or ribosome transit stabilizes the reporter. Deadenylation-dependent turnover is therefore the appropriate source-specific process; these experiments do not demonstrate surveillance of elongation-stalled transcripts by no-go decay.
The c-FOS mCRD assembly is distinct from the c-MYC coding-region stability complex. Human HeLa complex isolation and mouse NIH3T3 reporter experiments establish complementary aspects of its biology. Overexpression PMID:11051545(https://doi.org/10.1016/S0092-8674(00)00102-1), whereas subsequent experiments show that UNR is required for rapid translation-coupled turnover. These results describe different ribonucleoprotein states. They do not require a universal stabilizer-only or destabilizer-only role.
Positive translation is independently demonstrated by recombinant human-derived UNR in reticulocyte assays, where it PMID:10049359(https://pmc.ncbi.nlm.nih.gov/articles/PMC316477/). The 2000 mCRD paper proposes enhanced initiation in its model and cites earlier work; it is not treated as a new direct translation-rate measurement. HRV-2 RNA mapping identifies PMID:17947529(https://doi.org/10.1099/vir.0.82463-0) in two structured domains, including exposed bulges and loops. This supports structured-RNA recognition without establishing duplex specificity or a solved universal RNA-chaperone mechanism. Viral IRES initiation remains a contextual activity rather than a separate primary core.
CSDE1 is predominantly cytosolic, with cell-state-dependent associations. Human UNR–4E-T coimmunoprecipitation, RNase treatment and motif deletion support the PMID:27342281(https://pmc.ncbi.nlm.nih.gov/articles/PMC5291280/). These assays do not establish purified pairwise contact, and the combined deletion phenotypes leave alternative explanations for repression. The HID1 and FAM9B binary-screen positives are also retained at their measured scope. Partner compartment expectations do not invalidate those experiments; no informative binding class is asserted solely from a partner's general cellular role.
The miRNA study reports that CSDE1 PMID:32161113(https://pubmed.ncbi.nlm.nih.gov/32161113/). Human interaction experiments are distinguished from mouse and fly perturbations. Association and domain dependence do not imply that CSDE1 assembles the core RISC or performs intrinsic decapping. The HCC source identifies CSDE1 as an PMID:36354136(https://pmc.ncbi.nlm.nih.gov/articles/PMC9647857/), ARHGAP5-AS1. Its VIM/RAC1 protein-output results support a translation-regulatory context but are not direct translation-rate measurements. TRIM28 supplies the proposed ubiquitin-ligase activity.
The stress-granule source abstract explicitly names CSDE1 among proteins PMID:29395067(https://pubmed.ncbi.nlm.nih.gov/29395067/) granules. The positive IMP and localization are retained as context-dependent, with full target-level Methods and supplementary data still unavailable. In contrast, the P-body assignment remains unresolved: inspected PMID:32354837 main and supplementary material predominantly localizes 4E-T, while another study reports UNR absence in one condition. Neither observation establishes universal CSDE1 exclusion from P-bodies.
The original cloning study reports PMID:2204029(https://pmc.ncbi.nlm.nih.gov/articles/PMC331975/). Its mouse testis Northern comparison measures expression, not an executed gonad-development function. The broad developmental annotation is over-scoped by that source; the positive expression result is retained.
The current PAINT tree places all four cited IBDs at PTN000985585 on the human lineage. Target self-donation is valid experimental grounding. The c-FOS complex correction changes term scope, not ancestry. The fly protein–RNA-adaptor experiments support conserved RNA/protein coordination without assigning Drosophila dosage compensation to human CSDE1. The RISC-binding rule's Eukaryota/FunFam condition set matches the target; its separate plant branch does not. The existing GO-CAM represents mRNA binding contributing to translation regulation, consistent with the assessed source.
The five official protein sequences are 798, 767, 813, 844 and 622 residues. Their VSP operations and individual FASTAs agree. Historical exon-5 inclusion/deletion constructs, a second-residue cloning change in the 1999 recombinant proteins, and study-specific tags remain explicit. The 2025 study tests two labeled isoforms in human melanocytic cells and mouse keratinocytes; it does not resolve the function of every listed product. Whole-lysate phosphatase treatment does not isolate a CSDE1-specific phosphorylation mechanism, and ribosome association does not establish structural ribosome membership.
The evidence index preserves literal annotation, interaction, product, tree, rule, donor and table records, with public retrieval locators. It is an authored provenance resource, not independent biological evidence. The 1990 original was read as nine scanned pages; the 2000 and 2007 originals were recovered as indexed author-uploaded reprints, with the exact returned windows distinguished from unavailable material. Full main-source passages were assessed for the mCRD, IRES, 4E-T, RISC, lncRNA and 2025 studies. PMID:22681889 retains a target-table access limitation; PMID:29395067 retains a full-Methods limitation. Selected cache flags do not override these actual access boundaries.
Direct mRNA recognition and protein–RNA coordination are represented as separate core activities. Purified UNR binds c-FOS coding-region RNA in mobility-shift and competition assays (PMID:15314026(https://pmc.ncbi.nlm.nih.gov/articles/PMC514181/)); the structured HRV-2 IRES experiments independently map RNA contacts (PMID:17947529(https://doi.org/10.1099/vir.0.82463-0)). The broader mRNA-binding core accommodates coding-region and untranslated-region recognition. It does not make the viral context a separate conserved core or equate stem-loop recognition with double-stranded-RNA specificity.
The c-MYC-complex IBA is removed because the experimentally grounded assembly is the c-FOS mCRD complex. The correct GO:0106002 IBA already exists with the same node and donor. The experimental GO:0070937 row still receives its source-specific correction to GO:0106002. This does not reject PAINT ancestry, self-donation, or multiple annotations with different evidence. The cytoplasm-to-cytosol refinement likewise remains supported alongside the distinct IDA rows.
The positive cytoplasmic-translation refinement relies on independent IRES reconstitution, while the HCC paper's VIM/RAC1 abundance and cited elongation interpretation retain their assay limits (PMID:10049359(https://pmc.ncbi.nlm.nih.gov/articles/PMC316477/); PMID:36354136(https://pmc.ncbi.nlm.nih.gov/articles/PMC9647857/)). The positive regulation term GO:1900153 exists, but GO:0000288 is retained for the c-FOS experiment: CSDE1 contributes RNA recognition and PABP-mediated organization to the decay machinery itself. CCR4 supplies catalysis; CSDE1's molecular contribution establishes participation beyond a necessity-only phenotype. No additional process assertion is proposed.
RISC complex binding remains an accepted part of the established regulatory biology, and its context is now explicit in the adaptor-core description (PMID:32161113(https://pmc.ncbi.nlm.nih.gov/articles/PMC7067469/)). Human HEK293T association and deletion mapping are distinguished from the mouse functional assays. RNA-resistant coimmunoprecipitation does not establish a purified AGO2 contact; the paper does not establish core miRISC assembly or CCR4-NOT recruitment by CSDE1. A third standalone binding core would repeat this regulatory context.
The existing evidence index is retained under its published filename so its links and literal provenance remain stable. It holds source evidence, while final curation decisions are in the review YAML. Failed provider attempts remain documented separately from the manual source assessment; no provider-generated report is implied.