Genuine source acquisition yielded 62 raw/TSV annotations and 61 distinct seeded contexts; one source-normalization collapse differed only in record ID, date and assigning source. The displayed UniProt Q16394 sequence is 746 residues and matches NP_000118.2. Independent transcript models were retained separately, with no invented UniProt products or functional equivalence.
The configured deep-research attempt and fallback failed with observed 403/401 responses. No provider output was fabricated. Research proceeded through actual original articles, authoritative database records and documented access attempts. Normal acquisition files remain unchanged.
All 61 source contexts are reviewed: 33 ACCEPT, 15 MODIFY, 12 KEEP_AS_NON_CORE and 1 UNDECIDED. There are 30 references, no NEW annotations and no publication quotations. Description and status remain at their seeded values; core synthesis is a separate phase. The source guide documents actual original access, experiments, products, source discrepancies and ontology relations.
Human EXT1 participates in HS-chain elongation. The historical soluble enzyme experiments in PMID:12907669 and modern human structures/assays in PMID:36402845/36593275 distinguish robust GlcA transfer, more condition-dependent secondary GlcNAc activity, and the dominant EXT2 GlcNAc site in the heterocomplex. Purified soluble constructs, cellular immunoprecipitates and full-length cell-rescue experiments are not interchangeable. PMID:17761672 supplies a human endogenous-knockdown result separately from mouse full-length overexpression constructs.
Historical tagged ER localization in PMID:9620772 and EXT2-dependent Golgi redistribution in PMID:10639137 remain positive non-core contexts. Independent tagged ER imaging in PMID:33962942, with construct provenance from PMID:27229929, supports accepted ER membrane localization. No general artifact or false-location conclusion is made. Newly recovered originals for PMID:9620772/10639137 supersede earlier abstract-only access notes; twelve review text fields were updated only to identify that access progression. Actions, supporting references and machine source fields are unchanged.
The reference assessments retain current original-access scope: the two modern structural studies use main text/captions without new pixel or supplementary numerical inspection. PMID:35137078 Figure 7 pixels were inspected. BioPlex preserves its literal EXT2 Q93063-2 pair label without bait-direction inference; the separate U2OS network explicitly identifies EXT1 as bait. Generic supported interactions remain non-core under the campaign instructions.
PMID:10679296 remains complete-original unavailable, with independent location/complex support. PMID:7550340 remains unavailable beyond abstract/preview; its source-specific ossification IMP is unresolved, while the independently reviewed electronic developmental context is retained as non-core. This does not reject the skeletal phenotype. Current ComplexPortal term/linkage queries and the unresolved Reactome R339L nomenclature are preserved without editing source records or sending upstream reports.
The dependency inventory preserves the 35 intake files and three genuine additional PMID acquisitions. Twenty-six dependency files plus the two gene machine sources form the selected citation/provenance closure; twelve unused acquisition files stay private. The existing main version of PMID:27229929 is selected for eventual delivery; the separately fetched private version is preserved. Their titles/shared metadata and original body agree after line-end whitespace removal; the private version adds publication-type metadata. No publication source was overwritten, and there is no quotation dependence on the differing bytes.
Two transfer reactions describe the enzyme more precisely than a generic glycosyltransferase or autonomous-polymerase label. EXT1 GT-B supplies the principal glucuronyl-transfer reaction in the Golgi EXT1–EXT2 complex. Positive human EXT1 single-transfer and chain-extension assays also support GlcNAc-transfer capacity. That capacity is retained with an explicit quantitative boundary: the modern complex studies favor EXT2 as the dominant GlcNAc-transferase, differ in the effects of EXT1 GT-A substitutions, and report stability or preparation-integrity limitations. Neither a blanket EXT1 GlcNAc-inactivity claim nor equivalence of isolated EXT1 to the native heterocomplex follows.
The older human soluble EXT1 29–746 preparation in PMID:12907669 was affinity captured from HEK293 conditioned medium. Single-transfer radiochemical assays and product digestion complement its polymerization experiments. The reported isoelectric-focusing assessment of endogenous partner contamination is data not shown, and apparent gel-filtration size does not establish an obligatory native oligomer. The 2022 and 2023 human preparations use different soluble boundaries and assays; low-yield, substantially proteolyzed EXT1 alone in the latter study permits qualitative activity but not a reliable quantitative comparison.
The core location is the Golgi membrane, supported by the accepted topology/localization evidence and native-participant model, not by secretion of engineered soluble constructs. The two catalytic units share this setting and one biosynthetic process; no developmental, generic-binding, ancestor-process or additional complex-activity unit is added. Questions focus on native GT-A contribution, catalytic versus stability effects and coordination of alternating sugar transfer. The accepted source discrepancy queries remain unchanged.
All four ER-membrane assertions now retain the positive localization as non-core, consistent with the Golgi location of the active elongation complex. The mouse donor, tagged Cos7 imaging and historical source-access limits remain explicit. Five EXT2-partner associations refine to protein heterodimerization: the screens and FRET establish association, whereas the purified human structural studies establish the 1:1 assembly. The NDST1 interaction remains non-core.
Both ossification assertions remain unresolved: the clinical PMID:7550340 abstract supports skeletal disease but does not settle the specific bone-formation role. Cryo-EM references were removed from that process assertion's support. Conditional EXT1 self-association remains positive non-core evidence, without replacing the native heterodimer. The secondary GlcNAc activity retains its direct MF slot because separately expressed preparations perform single transfer; contamination, stability and native quantitative-contribution limits remain stated.
Six concise paraphrased findings identify the operative figures/assays in key originals; no quotation is added and no normal cache is changed. Public source locators now identify the released BioPlex table, NDEx network and article sections instead of unexplained private block numbers. All imported source fields, reference identities, two core positions and source-defect queries are preserved. Compilation status is left unchanged for the later canonical validation.