EOGT adds O-linked N-acetylglucosamine to extracellular EGF-like domains as secretory-pathway proteins pass through the endoplasmic reticulum. It uses UDP-GlcNAc as donor and is distinct from intracellular OGT. Direct human activity was demonstrated for the 527-residue KC347596.1 clone on Notch EGF-repeat reporters in fly S2 cells (PMID:23671640, Figure 1 and Methods). Fly purified-enzyme and targeting experiments establish folded-EGF substrate specificity and ER retrieval (PMID:22158438, Figures 2–3). Mouse purified-enzyme and cellular assays provide additional ortholog evidence (PMID:25488668, Methods and Results associated with Figures 1–2). Original-paper links and exact reading limits are in the source guide.
Biallelic EOGT variants segregate with Adams-Oliver syndrome (PMID:23522784). Functional follow-up of corresponding substitutions used mouse EOGT, including when HEK293T cells were the host. Mouse W207S is unstable and prone to proteasomal degradation; mouse R377Q retains ER localization and acceptor binding but has impaired UDP-GlcNAc binding (PMID:25488668, Results associated with Figures 5–8). These engineered disease-substitution constructs and the truncation mimic do not establish the behavior of natural human splice products. Human genetics and fly rescue do not by themselves identify the endogenous human substrate or tissue mechanism responsible for the syndrome.
The three natural products are preserved separately: 527, 208 and 443 residues. Reconstruction and independent FASTA comparison are documented in the product analysis. Only the canonical product exactly matches the human clone tested in PMID:23671640. The shorter products lack the canonical DYD region, and the 208-residue product also lacks HDEL. Retained signal sequences or HDEL motifs alone do not establish localization or catalytic competence. Reduced activity of the engineered DYD-to-AYA mutant is not a direct assay of either natural shorter product.
Both PAINT assertions trace to PTN001841612, with the human target below that node in the current version-19 tree. The exact canonical sequence matches its ungapped MSA leaf, and no matching NOT record occurs on that path in the pinned IBD snapshot. The mouse and fly source identities and direct annotation citations were checked without treating donor count as evidential strength.
All eight imported annotations are retained as separate source assertions. Seven are accepted; the broad InterPro molecular-function annotation is refined to the existing protein O-acetylglucosaminyltransferase activity. No new process annotation is proposed. The parent protein O-linked glycosylation process is retained because the current GlcNAc-specific child excludes elongation, whereas mammalian secretory-pathway O-GlcNAc can be galactosylated (PMID:22310717, abstract). This assigns EOGT the initiating transfer rather than downstream galactosylation. The older mouse paper's full Methods were unavailable; independent human and mouse originals establish the catalytic function.
Acquisition produced eight raw annotations, eight normal TSV records and eight seeded rows. All six initial publication caches and three explicit subsequent additions remain unchanged; only the five cited publications are included as new public caches. Falcon research and its Perplexity fallback failed with provider access/quota errors, so no provider report was generated. The review uses primary evidence and official database records. Cache full-text flags, imported source fields and natural products are preserved, with no added publication quotations.
One core function summarizes protein O-acetylglucosaminyltransferase activity in the ER lumen and its direct contribution to protein O-linked glycosylation. Human activity is established for the canonical 527-residue product; the two shorter products remain functionally uncharacterized. The core introduces no additional process assertion.
Whole-gene validation passed with no errors or curation warnings. The normal status computation reports COMPLETE. All eight source annotations, three natural products and nine normal publication caches remain preserved.
PR feedback follow-up (2026-10-24): fresh QuickGO and OLS records independently confirm that GO:7770074 is active and was created in May 2026. The GO:0006493 parent and current descendant definitions were checked; the child restriction against glycan extension remains the reason to retain the parent. Added an inspectable ontology-response receipt and explicit recovered-publisher-PDF access in the PMID:22158438 reference review. PMID:25488668 reference review and ER reasons now distinguish mouse EOGT localization from the calnexin-overlapping glycosylation-product signal. The folded EGF acceptor context was added to the biological description, and a fragile positional row reference was replaced with evidence-code names. All eight annotation actions, core functions, imported source fields, three natural products and normal cache metadata are unchanged; no publication quotations were added.