CEP85L links centrosomal organization to cortical neuronal migration. The five original GO assertions and all five UniProt products are preserved. The retained neuronal-migration role has mouse knockdown and partial rescue evidence; it is not described as a direct human-neuron perturbation. Tagged human protein localizes to the pericentriolar region, while independent endogenous human-cell experiments connect CEP85L to CDK5 recruitment at the proximal mother centriole.
PMID:32097630(https://pmc.ncbi.nlm.nih.gov/articles/PMC7357395/) Figure 2 follows mouse cortical neurons after embryonic RNAi, including rescue and a persistent postnatal migration defect. Figure 3A–C maps human tagged CEP85L relative to centrin, SAS6 and pericentrin. Figure 3D separately uses mouse cortical expression. The construct Methods identify human NM_001042475 and mouse NM_001204983. Neither the human disease variants nor the mouse perturbations establish a uniform loss-of-function mechanism across all five current protein products. The start-codon variants affect a particular transcript, and the paper itself leaves their precise pathogenic mechanism open.
PMID:32097629(https://pmc.ncbi.nlm.nih.gov/articles/PMC7255387/) Figure 4E reports CEP85L–CDK5 co-immunoprecipitation. Figure 4F–H and supplementary Figure S2K–L connect CEP85L abundance to centrosomal CDK5 and its phosphorylated form. These findings support one new protein-kinase-binding MF and the corresponding core. They do not resolve a purified binary interface, and no substrate-turnover assay was inspected. CDK5 activation remains a mechanistic model supported by phosphorylation and cellular phenotypes. The roscovitine experiment used a CDK5/1/2 inhibitor, so it is not treated as uniquely selective for CDK5. The human expression construct was amplified from HeLa cDNA as ENST00000368491; endogenous co-IP does not distinguish the five products. The perturbation of microtubule organization does not establish CEP85L microtubule binding or polymerase activity.
The original centrosome survey, PMID:21399614(https://pmc.ncbi.nlm.nih.gov/articles/PMC3102290/), explicitly lists C6orf204 in Table 1 and the target Results. Its PCP-SILAC score is 1.9; HPA staining is marked positive, but GFP expression was too weak to confirm centrosome association. The unchanged repository cache is partial despite its full-text flag. The target table and surrounding Results were read through indexed original PMC text, not reconstructed from the paper title or from CEP85, which is a different Table 1 entry. The later primary studies independently strengthen the localization. No original 2011 image pixels were inspected.
The durable source evidence record contains the exact original source objects, the manually inspected target fields and source locators, construct/species boundaries, and hashes of the actual source caches and external access receipt. It is a finite curation extract, not a new bioinformatics analysis. The PAINT node PTN001257853 is preserved without claiming reconstruction of its ancestry. Q6P2H3 was verified as human CEP85. The target's own Q5SZL2 source is valid experimental grounding, not circular evidence; independent CDK5 recruitment supports activity at the centrosome beyond simple residence.
The new MF supplies informative coverage absent from the five source assertions. No biological-process term is added. The current official GO definition of protein kinase binding and its typed is_a/part_of ancestry were checked. The molecular-function decision is distinct from the campaign's stricter NEW-process participation and hierarchy checks. The authenticated GO-CAM index contained no Q5SZL2/CEP85L entry. The published campaign instruction governs supported binding judgments; no generic-binding assertion is introduced here.
Normal fetch seeded all five assertions and products. The genuine Falcon attempt reached its 90-second timeout; its explicit Perplexity-lite fallback returned HTTP 401 for insufficient quota. No provider report was produced or simulated. Manual primary research continued with the authentic normal publication caches. All existing main-cache bytes were retained, and additional publications were fetched through the unchanged normal tool. Short source-specific anchors are in the YAML; this notes section adds no literature quotations. Biological derived status and campaign completion are separate: the latter still requires the approved-head, required-check and merge gates.
This appended assessment supersedes the earlier parent-level MF and no-NEW-process decisions while preserving the five original source assertions, products and reference identities. The authored interaction assertion and core now use GO:0120283 protein serine/threonine kinase binding; its IPI partner is explicitly UniProtKB:Q00535 (human CDK5). Current official UniProt identity/function and GO definitions support this refinement. It remains a co-immunoprecipitation association, without a claimed purified binary interface or autonomous kinase-activator mechanism.
One NEW GO:0071539 protein localization to centrosome assertion is supported by PMID:32097629 Figure 4E–J and the selected human cell/IP Methods. Endogenous HeLa association, loss of centrosomal CDK5 after CEP85L depletion with unchanged total CDK5, and preserved CEP85L localization in CDK5-deficient patient fibroblasts jointly support CEP85L performing recruitment/maintenance work. This reasoning uses the physical association and directional dependency, not only a neuronal-migration necessity phenotype. It does not resolve a direct tether interface or distinguish delivery from retention. The existing migration evidence remains scoped to the mouse perturbations PMID:32097630.
The official process definition includes transport to or maintenance at the centrosome. Typed is_a/part_of parents and both ancestry directions exclude overlap with all original and proposed terms. Fresh same-role comparisons found GO:0071539 for CEP192 (human IMP, IBA and IEA), PCM1 (IMP, ISS, IBA and IEA), and CEP83 (two IMP records with acts_upstream_of_or_within). CEP164 and CEP63 returned no exact/descendant annotation in the bounded current queries; these negative observations are preserved, not treated as missing curation. These are comparator conventions rather than a substitute for the target assays. The exact-head GO-CAM index has no Q5SZL2, CEP85L or C6orf204 match. The durable source record contains the literal comparator evidence and qualifiers, term responses, query URLs/hashes and this performed-step assessment.
The prior PMID:21399614 cache remains byte-exact and partial despite its full-text flag. Its target table evidence still comes from the documented indexed original source; this follow-up neither fabricates missing cache text nor claims a fresh full-cache retrieval. The absent PTHR31075 PAINT companion does not supply an ancestral reconstruction: the original node remains UNRESOLVED and the genuine normal metadata/entries remain unchanged. No provider retry was made because the documented timeout/quota conditions have not changed. The earlier notes section remains an exact historical prefix; this appendix adds one short source anchor.