All 29 source contexts have reviewed decisions: 16 ACCEPT, 6 KEEP_AS_NON_CORE, 7 MODIFY. Two source-described natural products (601/239 residues) remain separate. Human acetyl-CoA-dependent autoacetylation, combined ESCO1/ESCO2-dependent SMC3 acetylation and mouse Esco2 loss experiments retain their distinct assay scope. The acetyl-phosphate/free-lysine annotation is refined to protein-lysine acetyltransferase chemistry. Nuclear imaging and chromatin fractionation are distinguished. Shared references retain the direct lane's actual figure-access assessment; the source guide also records the independent database cross-check. No new quotation or annotation is introduced. Description and core synthesis follow this annotation checkpoint.
One catalytic core now combines protein-lysine acetyltransferase activity with mitotic sister-chromatid cohesion and positive regulation of DNA replication. Chromatin is the activity location; no permanent cohesin-complex membership is asserted. The measured human autoacetylation fragment, combined human ESCO1/ESCO2 SMC3 experiment, single ESCO2 fork perturbation and mouse Esco2-loss experiment remain distinct. The short natural product remains functionally unestablished. Existing annotation decisions, reference assessments, natural-product records and source fields are unchanged. No new annotation or publication quotation is added.
Final canonical validation records DRAFT with 4 advisories. All 29 source contexts and both natural product records remain exact; all 60 authored narratives appear in the private HTML rendering. No NEW annotation or publication quotation was introduced.
PMID:15958495(https://doi.org/10.1091/mbc.e04-12-1063), RNA-interference Methods and Results/Figure 3, reports cohesion defects after depletion of either human EFO1/ESCO1 or EFO2/ESCO2. This is positive evidence that the proteins are not generally interchangeable. PMID:22101327(https://doi.org/10.1038/emboj.2011.381), Results/Figures 3–4 and 6–7, resolves a nonredundant mouse Esco2 requirement: pericentric enrichment, altered cohesion, and reduced Smc3 K105/K106 acetylation and Sororin association after Esco2 loss. The mouse results are not a direct human short-product assay.
PMID:21111234(https://doi.org/10.1016/j.cell.2010.10.031), printed pp. 741–742 and Figure 3A/B, shows that only combined ESCO1/ESCO2 depletion substantially reduced the measured chromatin acetyl-SMC3 endpoint in that S-phase HeLa assay. This narrower observation is retained without generalizing to all cohesion functions. Its separate purified-enzyme experiment uses ESCO1 and Xenopus chromatin. Human ESCO2-fragment autoacetylation in PMID:15958495 Figure 2A remains separate.
PMID:22699483(https://doi.org/10.1095/biolreprod.112.099978), immunofluorescence Methods and Results/Figures 4–5, supports ESCO2/phospho-H2AFX-positive chromosomal clumps in E18.5 female pachytene oocytes. This differs from the male XY-body experiment in Figure 2. GO:0035861 names a chromosome region, so direct DNA-end binding and repair catalysis are not prerequisites. The specialized localization remains non-core with explicit marker-based resolution; it is not converted into a repair-function assertion. PMID:22614755 remains unavailable beyond metadata/abstract.
The positive HPA nucleoplasmic staining and actual Reactome target compartments are retained as non-core contexts. The principal acetylation/cohesion core remains on chromatin. PMID:19907496(https://doi.org/10.1038/nature08550), Figures 3d/e and 4e, supplies independent human chromatin and replication-fork velocity evidence. GO:0045740 includes regulation of rate; the precise fork-velocity endpoint remains in the biological rationale and open question. The broader cohesion annotation is refined to the already-supported mitotic term GO:0007064.
The public source guide records original access routes, precise figures/sections, species and constructs. PMID:15958495 was recovered as a full main original despite its unchanged abstract-only normal cache. That distinction is recorded in the reference review; no cache or source flag is rewritten to conceal it. The standing campaign instructions explicitly retain supported generic binding as non-core and prohibit adding publication quotations. The four HuRI interaction decisions therefore remain unchanged, with their exact partner/clone limits. Reference-only evidence does not claim a verbatim quotation check.
The revised actions are 12 ACCEPT, 9 KEEP_AS_NON_CORE and 8 MODIFY across the same 29 imported contexts. All original source fields, 15 reference identities and two natural-product records are preserved; no NEW annotation or quotation is added. Earlier count/status paragraphs above are historical snapshots. The 239-residue product remains an explicit open question, rather than a demonstrated equivalent of the 601-residue protein or engineered fragments.
Revised canonical validation passes all six checks, with all 64 authored narratives present in the private rendering. Computed DRAFT retains four supported generic protein-binding advisories under the campaign instruction. The 29 source objects, 15 reference identities, two natural products and 27 selected normal source files remain exact; no NEW annotation or quotation was introduced.