The single official human EIF3F product is O00303, 357 amino acids. No alternative-product block or artificial suffix was added. Historical fragments, processed chain 2–357, mouse proteins and the MSH4 paper’s 361-residue notation are kept separate from the natural reference product.
The main role is participation in cytosolic eIF3. Native human F:H, F:M and F:H:M subcomplexes establish specific assembly contacts; F-containing recombinant human complexes support 48S initiation. The five-subunit control lacks C, while F-minus material was insufficiently purified for a matched activity comparison. Direct RNA crosslinks assigned to A/B/D/G do not become F-specific RNA binding. The core therefore combines F-specific initiation-factor binding with contribution to the assembled complex’s initiation-factor activity.
Reported deubiquitinase activity is retained with its limits. Human GST-F and MPN preparations cleaved Ub-GFP in bacterial extracts; independent fly preparations cleaved Ub-rhodamine. These observations do not establish purified-human kinetics, intrinsic activity free of copurifying enzymes, physiological human substrate specificity, metal dependence or a catalytic thiol. The broader GO:0101005 is used without inventing a downstream process or importing the initiation location into the DUB core. Genuine PAINT placement and human/fly experimental grounding remain intact.
FBXO32 and processed CDK11 experiments establish EIF3F as an interacting substrate, not a ligase or kinase. Other supported screen bindings remain non-core unless independent mechanism or established complex evidence supports a subtype. This follows the campaign’s task-specific retention of supported generic binding; partner class alone was not treated as a new native molecular function. Mouse synaptic detection and viral-IRES observations remain contextual. The YTS membrane-proteomics assertion remains unresolved because its target table and operative fractionation were not recovered; cytosolic activity is not evidence that another location is impossible.
All 119 imported source dictionaries and accepted reviews are preserved: 30 ACCEPT, 36 MODIFY, 52 KEEP_AS_NON_CORE and 1 UNDECIDED. The 121 raw GOA records reconcile to 119 full contexts with two explicit duplicate groups. There are no NEW rows or publication quotations. All 48 original normal publication caches remain exact; five additional genuine caches and the existing correction notice are separately qualified. The provider attempt failed with no output; no manual prose is attributed to a provider. Workflow status is left for canonical validation.
Primary-source consultations independently checked PMID17581632, PMID21124883 and the fly donor PMID38279019; the source guide records their precise assay and image scopes. The 41039152 publisher correction concerns a duplicated loss-function equation and variable labels, not stated interaction-edge changes. Normal cache availability flags and independent original access are reported separately.
Source access and assay guide and reproducible database results give the exact source routes, hashes, raw-data joins and limits. Database replay verifies those joins; it does not reproduce the experiments or mechanically choose annotation actions.
The DUB summary now explicitly includes the positive activated-NOTCH1 and DTX1 findings curated by UniProt from PMID:21124883. The earlier blanket wording about physiological substrate specificity was too broad: activated NOTCH1 is a supported cellular substrate context. Partial depletion of endogenous human EIF3F increased modified activated Notch constructs, with rescue by murine wild-type EIF3F but not its MPN mutant (Results and Figure 2). DTX1 enabled the observed association between eIF3f and activated Notch constructs (Results and Figures 5–6). Notch reporter cocultures provide additional functional evidence (Figure 7). These cellular experiments, their murine constructs and the separate human GST-EIF3F bacterial cleavage assay remain distinguished.
The same original Discussion explicitly leaves the precise monoubiquitinated Notch substrate form unsettled and acknowledges a possible associated DUB (sections A DUB Acting during Notch Activation and eIF3f Is a DUB). Those limits qualify the mechanism; they do not erase the positive Notch findings or establish inactivity. The broader GO:0101005 decision remains, because the inspected experiments do not settle human metal dependence or thiol chemistry. No new Notch-process annotation, DUB location or complex is inferred. The second core retains its term and references while describing this evidence more precisely.
The PMID:17581632 cache contains the Abstract, Introduction and Discussion. The separate original consultation supplied Results, Methods and figure panels; the access description now names those missing sections explicitly. No cache or earlier image-inspection claim was changed. Additional publication quotations were not added under the standing instruction; reference-only anchors and existing public source locators remain. All 119 source objects, all annotation dispositions, the single 357-aa natural product and both core term structures are unchanged.