Genuine fetch-gene human EXT2 with explicit Q93063 seeded 60 PENDING contexts. Normal publication caching and the default Falcon command with perplexity-lite fallback overlapped. Both publication commands succeeded. Provider exit 1 reflects actually observed Falcon 403 and Perplexity 401 failures; no provider report or authored substitute was created. Raw credential-capable logs remain private with mode 0600, and bounded marker results are recorded separately.
All 42 normal publication/Reactome/PAINT files are preserved exactly in archive and private runtime. Machine metadata reporting full text is not a claim that this intake reviewer read the originals. ROOT's separately sealed EXT1 structural-original packet is pinned only for future EXT2-specific evidence reading. No EXT1 annotation decision or product conclusion is inherited.
The three source-declared products Q93063-1, Q93063-2 and Q93063-3 are 718, 728 and 751 residues. VSP_001798 replaces canonical residues 392–411 with a 30-residue segment; VSP_046053 replaces the first methionine with a 34-residue N-terminal segment. Exact reconstruction matches each independently fetched official FASTA. Separate RefSeq and Ensembl records are source observations, without an activity, abundance, processing or equivalence claim.
All 60 imported contexts were manually reviewed across the direct experimental, structural and database source assessments: 20 ACCEPT, 23 MODIFY and 17 KEEP_AS_NON_CORE. No annotation decision or source field changed during integration. The negated human soluble-EXT2 homodimer context remains explicitly negated and preparation-specific; full-length mouse self-association does not establish an opposite human stoichiometry. No new annotation or publication quotation was added, and status/description are still the seeded values pending separate core synthesis.
The 33 references preserve all 29 seed identities and findings, with three genuine donor additions (PMID:10639137, PMID:9756849, PMID:11256613) and the verified ComplexPortal record. Five shared assessments were combined explicitly: mouse/bovine versus human constructs, low human soluble GlcA capacity versus predominant complex-site assignment, fluorescent localization and actual access scopes. The modern GlcNAc mechanism does not erase the historical positive low GlcA result. Historical experimental ER contexts remain non-core without declaring ER localization false; conserved ER membrane evidence and human Golgi evidence are distinguished.
See the source guide for actual Methods/Results, target-table, peptide and figure-access scopes, plus public original links. Publication extraction flags remain properties of immutable normal caches, not a substitute for the original-source read ledger. The three natural products remain distinct from engineered truncations and tagged constructs. Supported generic associations remain non-core under the campaign instructions.
For shared-source delivery, PMID:10639137 uses the exact already-published EXT1 abstract-only normal cache. The genuine fuller EXT2 fetch remains preserved privately. This substantive extraction difference does not change the verified original read scope or any annotation judgment. Other shared source bytes match exactly.
One catalytic core describes the principal EXT2 GlcNAc-transfer step within the Golgi EXT1–EXT2 complex. Current GO:0050508 identifies the alpha-1,4 transfer; GO:0015012 supplies the single directly involved biosynthetic process, GO:0000139 the functional membrane location and GO:0120504 the complex. The enzyme performs a measured sugar-transfer step, so this process attribution is supported by chemistry as well as loss/rescue phenotypes. No broad catalytic or process ancestors are added.
Human-complex active-site substitutions and product assays in PMID:36402845 and PMID:36593275 establish EXT2's catalytic contribution. The older positive soluble-human preparation in PMID:12907669 remains distinct: its low GlcA capacity is retained in the accepted contextual annotations and is not promoted to a second core or dismissed as impossible. Engineered soluble proteins are not equated with the three natural products. The two suggested experiments address this preparation/activity distinction and product-specific physiology.
All 60 accepted annotation reviews, 33 references, three product records, the NOT annotation and all normal-source bytes remain unchanged. The status remains INITIALIZED pending later canonical checks. The literal shared-source selection for PMID:10639137 is carried unchanged; its public abstract-only cache does not retract the independently read full original. No new publication quotation or annotation was added. See the source guide for actual assay and access boundaries.
The revised assessment has 60 source contexts: 19 ACCEPT, 27 MODIFY, 13 KEEP_AS_NON_CORE and 1 UNDECIDED. Earlier journal totals describe their dated checkpoints. The source assertions, three natural product records, negative homodimerization record, 33 reference identities and single core unit remain intact. No new annotation or publication quotation was added. See the source guide for the evidence and access distinctions.
EXT2 supplies the principal GlcNAc-transfer site in the human EXT1–EXT2 heterodimer; EXT1 supplies the principal GlcA-transfer site. EXT2’s participation in the assembled GlcA-active complex is now represented in the existing core with contributes_to, separately from the older low positive GlcA capacity of soluble EXT2 (PMID:12907669(https://pubmed.ncbi.nlm.nih.gov/12907669/), PMID:36402845(https://doi.org/10.1038/s41467-022-34882-6), PMID:36593275(https://doi.org/10.1038/s41589-022-01220-2)). The five generic EXT1-partner binding contexts are refined to heterodimerization using these structures; the cellular screens do not themselves establish stoichiometry. The two supported NDST1 associations remain non-core under the explicit campaign instruction.
Conserved ER-membrane residence is retained as contextual rather than a second catalytic compartment. Mouse EXT2-GFP retention and EXT1-dependent redistribution (PMID:10639137(https://pubmed.ncbi.nlm.nih.gov/10639137/)) remain distinct from human fluorescent EXT2 Golgi localization (PMID:35137078(https://pubmed.ncbi.nlm.nih.gov/35137078/)). The unresolved YTS membrane-proteomics record (PMID:19946888(https://pubmed.ncbi.nlm.nih.gov/19946888/)) is now UNDECIDED because independent Golgi evidence cannot verify its uninspected target assignment.
Six reference flags now describe the committed abstract-only caches: PMID:12907669, PMID:17761672, PMID:9326317, PMID:9756849, PMID:11256613 and PMID:10639137. Their VERIFIED assessments retain the actual external-original reading scope. No cache was changed, and the independently preserved fuller PMID:10639137 extraction is not represented as the delivered cache.