EXT2 is a type II membrane glycosyltransferase that forms the EXT1–EXT2 complex in the Golgi. Its C-terminal GT-A domain supplies the principal alpha-1,4-N-acetylglucosaminyl-transfer reaction, adding N-acetylglucosamine from UDP-GlcNAc to glucuronic-acid-terminated chains. Alternating with the principal glucuronyl-transfer reaction of EXT1, this activity elongates the heparan sulfate backbone on proteoglycans. Disruption of EXT2 impairs heparan sulfate synthesis and causes hereditary multiple exostoses, a disorder characterized by cartilage-capped skeletal outgrowths.
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0000139 Golgi membrane | EXP PMID:10679296 Association of EXT1 and EXT2, hereditary multiple exostoses ... | ACCEPT | Summary: Golgi membrane localization is independently supported for human EXT2. Reason: The PMID:10679296 abstract reports EXT association and Golgi localization, but its complete operative original remains unavailable. Independent PMID:35137078 Figure 7A and Methods show full-length human EXT2 fused to fluorescent protein in COS-7 cells, overlapping GM130; its membrane anchor supports the membrane location. This corroborates the location without claiming to recover the earlier experiment or to assay every natural product. |
| GO:0000139 Golgi membrane | IEA GO_REF:0000044 | ACCEPT | Summary: Retain the supported Golgi-membrane location. Reason: The current SL-0134 mapping is exact. Native human Reactome entity 2022888 identifies Q93063 residues 1–718 with its own Golgi-membrane compartment. PMID:35137078 Figure 7A and construct Methods independently place tagged full-length human EXT2 with GM130 in COS-7 cells; HPA supports Golgi localization in human SiHa cells. The imaging does not itself resolve membrane topology or all three natural products. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB-SubCell:SL-0134 SUPPORTS TRANSFER Actual current Golgi-membrane mapping; native target participant supplies an independent human model. |
| GO:0000139 Golgi membrane | TAS Reactome:R-HSA-2022851 | ACCEPT | Summary: Retain Golgi-membrane localization with the exact native or disease-participant boundary. Reason: The catalyst/complex route 2022851 → 2022845 → 2022878 → 2022888 reaches native human EXT2 2022888, Q93063 residues 1–718, with its own Golgi-membrane compartment. Retain participant localization without treating a complex-level reaction as proof that EXT2 alone catalyzes every sugar-transfer step. Supporting Evidence: Reactome:R-HSA-2022851 |
| GO:0000139 Golgi membrane | TAS Reactome:R-HSA-2022856 | ACCEPT | Summary: Retain Golgi-membrane localization with the exact native or disease-participant boundary. Reason: The catalyst/complex route 2022856 → 2022849 → 2022878 → 2022888 reaches native human EXT2 2022888, Q93063 residues 1–718, with its own Golgi-membrane compartment. Retain participant localization without treating a complex-level reaction as proof that EXT2 alone catalyzes every sugar-transfer step. Supporting Evidence: Reactome:R-HSA-2022856 |
| GO:0000139 Golgi membrane | TAS Reactome:R-HSA-2076392 | ACCEPT | Summary: Retain Golgi-membrane localization with the exact native or disease-participant boundary. Reason: The catalyst/complex route 2076392 → 2022849 → 2022878 → 2022888 reaches native human EXT2 2022888, Q93063 residues 1–718, with its own Golgi-membrane compartment. Retain participant localization without treating a complex-level reaction as proof that EXT2 alone catalyzes every sugar-transfer step. Supporting Evidence: Reactome:R-HSA-2076392 |
| GO:0000139 Golgi membrane | TAS Reactome:R-HSA-3656254 | ACCEPT | Summary: Retain Golgi-membrane localization with the exact native or disease-participant boundary. Reason: The catalyst route 3656254 → 9631891 → 9036278 → 3731111 → 3731083 is a defective-EXT2 model. Its set contains A361Pfs*44, D227N, G172*, Q258*, Y222* and V187Pfs*115 members, each assigned Golgi membrane; these are disease variants, not the three natural splice products. That model alone does not establish wild-type localization. Independently, Native human Reactome entity 2022888 identifies Q93063 residues 1–718 with its own Golgi-membrane compartment. PMID:35137078 Figure 7A and construct Methods independently place tagged full-length human EXT2 with GM130 in COS-7 cells; HPA supports Golgi localization in human SiHa cells. The imaging does not itself resolve membrane topology or all three natural products. Retain the supported native location with this source limitation; no activity loss or localization phenotype is transferred from the mutant set. |
| GO:0000139 Golgi membrane | TAS Reactome:R-HSA-3656257 | ACCEPT | Summary: Retain Golgi-membrane localization with the exact native or disease-participant boundary. Reason: The catalyst/complex route 3656257 → 9631805 → 9036279 → 2022888 reaches native human EXT2 2022888, Q93063 residues 1–718, with its own Golgi-membrane compartment. The defective partner in this event is EXT1; the event does not replace EXT2 with a disease product. Retain participant localization without treating a complex-level reaction as proof that EXT2 alone catalyzes every sugar-transfer step. Supporting Evidence: Reactome:R-HSA-3656257 |
| GO:0000139 Golgi membrane | TAS Reactome:R-HSA-3656261 | ACCEPT | Summary: Retain Golgi-membrane localization with the exact native or disease-participant boundary. Reason: The catalyst/complex route 3656261 → 9631718 → 9036279 → 2022888 reaches native human EXT2 2022888, Q93063 residues 1–718, with its own Golgi-membrane compartment. The defective partner in this event is EXT1; the event does not replace EXT2 with a disease product. Retain participant localization without treating a complex-level reaction as proof that EXT2 alone catalyzes every sugar-transfer step. Supporting Evidence: Reactome:R-HSA-3656261 |
| GO:0000139 Golgi membrane | TAS Reactome:R-HSA-3656267 | ACCEPT | Summary: Retain Golgi-membrane localization with the exact native or disease-participant boundary. Reason: The catalyst route 3656267 → 9631865 → 9036278 → 3731111 → 3731083 is a defective-EXT2 model. Its set contains A361Pfs*44, D227N, G172*, Q258*, Y222* and V187Pfs*115 members, each assigned Golgi membrane; these are disease variants, not the three natural splice products. That model alone does not establish wild-type localization. Independently, Native human Reactome entity 2022888 identifies Q93063 residues 1–718 with its own Golgi-membrane compartment. PMID:35137078 Figure 7A and construct Methods independently place tagged full-length human EXT2 with GM130 in COS-7 cells; HPA supports Golgi localization in human SiHa cells. The imaging does not itself resolve membrane topology or all three natural products. Retain the supported native location with this source limitation; no activity loss or localization phenotype is transferred from the mutant set. |
| GO:0000139 Golgi membrane | TAS Reactome:R-HSA-9036285 | ACCEPT | Summary: Retain Golgi-membrane localization with the exact native or disease-participant boundary. Reason: The catalyst/complex route 9036285 → 9631886 → 9036279 → 2022888 reaches native human EXT2 2022888, Q93063 residues 1–718, with its own Golgi-membrane compartment. The defective partner in this event is EXT1; the event does not replace EXT2 with a disease product. Retain participant localization without treating a complex-level reaction as proof that EXT2 alone catalyzes every sugar-transfer step. Supporting Evidence: Reactome:R-HSA-9036285 |
| GO:0000139 Golgi membrane | TAS Reactome:R-HSA-9036289 | ACCEPT | Summary: Retain Golgi-membrane localization with the exact native or disease-participant boundary. Reason: The catalyst route 9036289 → 9631922 → 9036278 → 3731111 → 3731083 is a defective-EXT2 model. Its set contains A361Pfs*44, D227N, G172*, Q258*, Y222* and V187Pfs*115 members, each assigned Golgi membrane; these are disease variants, not the three natural splice products. That model alone does not establish wild-type localization. Independently, Native human Reactome entity 2022888 identifies Q93063 residues 1–718 with its own Golgi-membrane compartment. PMID:35137078 Figure 7A and construct Methods independently place tagged full-length human EXT2 with GM130 in COS-7 cells; HPA supports Golgi localization in human SiHa cells. The imaging does not itself resolve membrane topology or all three natural products. Retain the supported native location with this source limitation; no activity loss or localization phenotype is transferred from the mutant set. |
| GO:0000271 polysaccharide biosynthetic process | IDA PMID:12907669 In vitro polymerization of heparan sulfate backbone by the E... | MODIFY | Summary: The demonstrated biosynthetic work is heparan-sulfate chain elongation. Reason: PMID:12907669 shows strong polymerization by the soluble human EXT1–EXT2 preparation, while EXT2 alone gives a small single-GlcA extension rather than appreciable chain elongation. Independent human heterocomplex experiments in PMID:36593275 identify EXT2-mediated GlcNAc transfer and directly elongate glypican-1-associated chains. Replace the broad polysaccharide process with the already-carried heparan sulfate proteoglycan process. This is a lateral evidence-based specificity correction, not a claim of current ontology ancestry or that EXT2 synthesizes the core protein and linker. Proposed replacements: heparan sulfate proteoglycan biosynthetic process |
| GO:0001503 ossification | IMP PMID:9326317 Mutation screening of the EXT1 and EXT2 genes in patients wi... | KEEP AS NON CORE | Summary: EXT2 has a supported skeletal-developmental phenotype distinct from its core catalytic reaction. Reason: The complete scanned PMID:9326317 original identifies EXT2 mutations segregating with hereditary multiple exostoses in human families, including truncating variants and D227N, and documents osteocartilaginous outgrowths and variable skeletal severity. Its operative assays are family genotyping and clinical assessment, not a cell-based bone-formation assay. Retain this curator-assigned ossification phenotype as a non-core developmental consequence of EXT2 dysfunction, without claiming that EXT2 directly deposits bone or that the study resolves the affected ossification step. Supporting Evidence: |
| GO:0005515 protein binding | IPI PMID:18337501 Heparan sulfate biosynthesis enzymes EXT1 and EXT2 affect ND... | KEEP AS NON CORE | Summary: EXT2 associates with NDST1, with the source species distinguished. Reason: PMID:18337501 Methods uses mouse EXT2 cDNA in HEK293 cells stably expressing NDST1. Reciprocal co-immunoprecipitation supports association, with preimmune-serum and peptide-blocked EXT2 controls. Independent PMID:35137078 Figure 7B uses full-length human EXT2 and human NDST1 constructs in COS-7 cells and detects FRET. These experiments corroborate the imported P52848 interaction without establishing a purified binary interface, fixed stoichiometry or the proposed transport/competition model. Retain supported generic binding as non-core. |
| GO:0005515 protein binding | IPI PMID:33961781 Dual proteome-scale networks reveal cell-specific remodeling... | MODIFY | Summary: EXT2 binding to EXT1 is more specifically supported as protein heterodimerization. Reason: The actual BioPlex 293T interaction table records Q93063-2 and Q16394 at line 8626, with pInt 0.99473832262035. Its A/B ordering is undirected and does not identify the individual immunoprecipitation that generated the edge, even though both genes occur in the bait inventory. Methods describes C-terminal FLAG–HA human ORFeome constructs and anti-HA capture. Preserve the literal historical product label without equating an unverified insert sequence to all current EXT2 products. The absent HCT116 pair, where neither target is a listed bait, is not negative evidence. Independently purified human EXT1–EXT2 complexes in PMID:36402845 and PMID:36593275 establish a functional heterodimer with this exact gene partner, supporting GO:0046982. This refinement uses the structural evidence; it does not infer dimer stoichiometry or current-product equivalence from the BioPlex edge. Proposed replacements: protein heterodimerization activity |
| GO:0005515 protein binding | IPI PMID:35137078 A dominant negative splice variant of the heparan sulfate bi... | KEEP AS NON CORE | Summary: Human EXT2 and NDST1 show positive cellular FRET. Reason: PMID:35137078 Figure 7B includes EXT2 with NDST1A and NDST1B. Methods specifies full-length human constructs fused to Venus or Cherry and expressed in COS-7 cells, with fused-fluorophore and unrelated glycosylation-enzyme controls. This supports the P52848 association as a non-core interaction, not purified binary binding or transfer of NDST catalytic activity to EXT2. The displayed Figure 7C localization pairs concern EXT1 and are not reassigned to EXT2. Supporting Evidence: |
| GO:0005515 protein binding | IPI PMID:35137078 A dominant negative splice variant of the heparan sulfate bi... | MODIFY | Summary: Human EXT2–EXT1 association supports a specific heterodimerization refinement. Reason: PMID:35137078 Figure 7B tests full-length human fluorescent EXT1 and EXT2 in COS-7 cells. Positive FRET establishes cellular proximity/association. Independently purified and structurally resolved human EXT1–EXT2 complexes in PMID:36402845 and PMID:36593275 establish heterodimerization with the same partner, supporting GO:0046982. The FRET assay alone does not measure purified affinity or dimer stoichiometry. Proposed replacements: protein heterodimerization activity |
| GO:0005515 protein binding | IPI PMID:36402845 Structure of the human heparan sulfate polymerase complex EX... | MODIFY | Summary: Human EXT2 and EXT1 form a purified, active heterodimer. Reason: PMID:36402845 purification Results, Figure 1 and Methods describe soluble human EXT1 residues 29–746 and EXT2 residues 47–718 co-expressed in HEK293-F cells. Purification, subunit stoichiometry, mass photometry and cryo-EM support direct association with the specified EXT1 partner Q16394. Refine generic binding to GO:0046982, which describes binding a different protein to form the experimentally resolved heterodimer. Engineered secretion of the anchor-deleted, tagged proteins does not establish extracellular localization of native EXT2 or equivalence of its three natural products. Proposed replacements: protein heterodimerization activity Supporting Evidence: |
| GO:0005515 protein binding | IPI PMID:36593275 Structural basis for heparan sulfate co-polymerase action by... | MODIFY | Summary: The EXT2–EXT1 association is independently supported by structural and biochemical analysis. Reason: PMID:36593275 expression/purification Methods use human EXT1 residues 28–746 and EXT2 residues 46–718, co-expressed in HEK293-F cells. The purified heterodimer extends heparan-sulfate acceptors and resolves both subunits structurally, directly supporting the Q16394 interaction. Refine the generic term to GO:0046982 protein heterodimerization activity, supported by this exact partner and resolved assembly. Failure to secrete full soluble EXT2 alone in this particular system is an expression result, not an independent demonstration that EXT2 has no catalytic activity or that every possible homodimer is absent. Proposed replacements: protein heterodimerization activity Supporting Evidence: |
| GO:0005515 protein binding | IPI PMID:40205054 Multimodal cell maps as a foundation for structural and func... | MODIFY | Summary: The EXT2–EXT1 APMS association can be refined using independent heterodimer evidence. Reason: The PMID:40205054 NDEx APMS network contains edge 552868 from EXT1 node 548664, explicitly bait=true, to EXT2 node 552866, bait=false. Original Methods describes human ORFeome C-terminal HA–Flag fusions expressed in U2OS cells, anti-HA capture and mass spectrometry. This is an exact target association, not a predicted structural pair: the inspected supplementary assembly and AlphaFold pair tables do not contain this pair. The target-specific historical insert sequence is unverified, so no equivalence among current EXT2 products is asserted. The independently purified human complex in PMID:36402845 and PMID:36593275 establishes heterodimerization with the same EXT1 partner, supporting GO:0046982. This specificity comes from the structural studies, not an inferred stoichiometry of the APMS capture. Proposed replacements: protein heterodimerization activity |
| GO:0005576 extracellular region | IEA GO_REF:0000044 | KEEP AS NON CORE | Summary: Retain extracellular localization as a non-core soluble-form context. Reason: PMID:9756849 purified an approximately 70-kDa protein from bovine serum, sequenced four tryptic peptides and recovered the corresponding 718-residue bovine EXT2 cDNA. This is positive extracellular donor evidence. Loss of the membrane anchor by proteolysis was inferred from size/topology; the cleavage site was not measured. The human annotation is a conserved soluble-form inference, not proof that all human EXT2 products are normally secreted. Engineered secretion leaders in later recombinant studies provide no additional physiological secretion evidence. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:O77783 SUPPORTS TRANSFER Actual bovine serum protein and peptide-to-cDNA evidence in PMID:9756849; ortholog transfer, not a direct human secretion assay. UniProtKB-SubCell:SL-0243 SUPPORTS TRANSFER Exact secreted/extracellular mapping; the underlying soluble-form evidence is bovine. Supporting Evidence: |
| GO:0005576 extracellular region | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: Retain extracellular localization as a non-core soluble-form context. Reason: PMID:9756849 purified an approximately 70-kDa protein from bovine serum, sequenced four tryptic peptides and recovered the corresponding 718-residue bovine EXT2 cDNA. This is positive extracellular donor evidence. Loss of the membrane anchor by proteolysis was inferred from size/topology; the cleavage site was not measured. The human annotation is a conserved soluble-form inference, not proof that all human EXT2 products are normally secreted. Engineered secretion leaders in later recombinant studies provide no additional physiological secretion evidence. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:O77783 SUPPORTS TRANSFER Actual bovine serum protein and peptide-to-cDNA evidence in PMID:9756849; ortholog transfer, not a direct human secretion assay. Supporting Evidence: |
| GO:0005783 endoplasmic reticulum | IDA PMID:10679296 Association of EXT1 and EXT2, hereditary multiple exostoses ... | KEEP AS NON CORE | Summary: The historical ER compartment is retained with conserved-source and access limits. Reason: The complete PMID:10679296 original remains unavailable. Independently read PMID:10639137 Figure 2 and Results show predominantly ER-localized GFP-tagged mouse EXT2 in BHK and sog9 cells, with redistribution toward Golgi after human EXT1 co-expression. This provides conserved positive support for the non-core ER context, not a recovered human experiment from the original source or a reason to dismiss ER residence as an artifact. Human EXT2 imaging in PMID:35137078 independently supports Golgi, not ER. |
| GO:0005783 endoplasmic reticulum | IEA GO_REF:0000107 | MODIFY | Summary: Specify the supported ER-membrane compartment. Reason: Mouse P70428 and ENSMUSP00000028623 have identical current 718-residue sequences. The actual PMID:10639137 Methods and Figure 2 use mouse EXT2-GFP in BHK/sog9 cells: expressed alone it is predominantly ER-associated, whereas coexpression with human EXT1 redistributes the complex to Golgi. The native N-terminal membrane anchor and current SL-0097 record support membrane specificity. This is ortholog/construct evidence, not endogenous human ER imaging or an assay of the 728- or 751-residue products. Propagation Review Root cause: TERM SCOPING PROBLEM Sources checked: UniProtKB:P70428 SUPPORTS TRANSFER Actual mouse donor and positive ER/Golgi imaging read; host and coexpression state retained. ensembl:ENSMUSP00000028623 SUPPORTS TRANSFER Official Ensembl protein sequence exactly matches P70428; not a separate experiment. Proposed replacements: endoplasmic reticulum membrane Supporting Evidence: |
| GO:0005783 endoplasmic reticulum | ISS GO_REF:0000024 | MODIFY | Summary: Specify the supported ER-membrane compartment. Reason: Mouse P70428 and ENSMUSP00000028623 have identical current 718-residue sequences. The actual PMID:10639137 Methods and Figure 2 use mouse EXT2-GFP in BHK/sog9 cells: expressed alone it is predominantly ER-associated, whereas coexpression with human EXT1 redistributes the complex to Golgi. The native N-terminal membrane anchor and current SL-0097 record support membrane specificity. This is ortholog/construct evidence, not endogenous human ER imaging or an assay of the 728- or 751-residue products. Propagation Review Root cause: TERM SCOPING PROBLEM Sources checked: UniProtKB:P70428 SUPPORTS TRANSFER Actual mouse donor and positive ER/Golgi imaging read; host and coexpression state retained. Proposed replacements: endoplasmic reticulum membrane Supporting Evidence: |
| GO:0005789 endoplasmic reticulum membrane | EXP PMID:10679296 Association of EXT1 and EXT2, hereditary multiple exostoses ... | KEEP AS NON CORE | Summary: ER membrane is retained as a contextual location with an explicit evidence boundary. Reason: The exact experiment behind PMID:10679296 remains unread. Independent PMID:10639137 combines ER-pattern microscopy of full-length mouse EXT2, its N-terminal membrane anchor and reported endoglycosidase-H-sensitive glycosylation; the latter supports secretory-pathway topology but is not a membrane-leaflet assay. Retain the conserved ER membrane context as non-core without describing it as directly demonstrated native human ER imaging. Golgi accumulation after EXT1 co-expression is a separate positive observation. |
| GO:0005789 endoplasmic reticulum membrane | IEA GO_REF:0000044 | KEEP AS NON CORE | Summary: ER-membrane residence is retained as a contextual, non-core location. Reason: SL-0097 is the exact ER-membrane mapping. Mouse P70428 and ENSMUSP00000028623 have identical current 718-residue sequences. The actual PMID:10639137 Methods and Figure 2 use mouse EXT2-GFP in BHK/sog9 cells: expressed alone it is predominantly ER-associated, whereas coexpression with human EXT1 redistributes the complex to Golgi. The native N-terminal membrane anchor and current SL-0097 record support membrane specificity. This is ortholog/construct evidence, not endogenous human ER imaging or an assay of the 728- or 751-residue products. ER residence is biologically supported, but these experiments describe expression-dependent retention and redistribution rather than an independently established ER-localized catalytic role. It is therefore retained as non-core alongside the historical ER records; the principal chain-elongation core is localized to Golgi membrane. This does not dismiss ER residence as an artifact. The broad ER annotations refined to ER membrane retain their specificity corrections without establishing a second catalytic location. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB-SubCell:SL-0097 SUPPORTS TRANSFER Actual mapping and membrane topology, interpreted with the independently read ortholog localization assay. Supporting Evidence: |
| GO:0005794 Golgi apparatus | IBA GO_REF:0000033 | MODIFY | Summary: Refine the valid ancestral Golgi annotation to Golgi membrane. Reason: The actual PTHR11062 tree places human EXT2 leaf PTN002476065 below PTN000115158, and its ungapped MSA exactly matches the 718-residue Q93063 sequence. The imported Golgi and glycosyltransferase IBDs are at PTN000115158; separate sugar-transfer IBDs occur at PTN000868988. No target-path NOT was found. Current UniProt PTHR48261 classification is a separate family/version record; its returned tree does not contain this target. Target self-WITH is valid experimental grounding, not circularity. Native human Reactome entity 2022888 identifies Q93063 residues 1–718 with its own Golgi-membrane compartment. PMID:35137078 Figure 7A and construct Methods independently place tagged full-length human EXT2 with GM130 in COS-7 cells; HPA supports Golgi localization in human SiHa cells. The imaging does not itself resolve membrane topology or all three natural products. Propagation Review Root cause: TERM SCOPING PROBLEM Sources checked: PANTHER:PTN000115158 SUPPORTS TRANSFER Verified target descent from the positive Golgi IBD; independent native participant supports membrane refinement. Proposed replacements: Golgi membrane |
| GO:0005794 Golgi apparatus | IDA GO_REF:0000052 | MODIFY | Summary: Refine the supported Golgi location using independent membrane evidence. Reason: Actual HPA HPA078409 ICC/IF in SiHa is rated supported and annotated Golgi apparatus. The selected image was inspected; its perinuclear signal does not independently resolve membrane topology. The antigen maps to Q93063 residues 423–501 and is shared by the three source-described products, so this is not isoform-specific localization. Native human Reactome entity 2022888 identifies Q93063 residues 1–718 with its own Golgi-membrane compartment. PMID:35137078 Figure 7A and construct Methods independently place tagged full-length human EXT2 with GM130 in COS-7 cells; HPA supports Golgi localization in human SiHa cells. The imaging does not itself resolve membrane topology or all three natural products. Proposed replacements: Golgi membrane |
| GO:0005794 Golgi apparatus | IDA PMID:10679296 Association of EXT1 and EXT2, hereditary multiple exostoses ... | MODIFY | Summary: Golgi membrane is the supported specific location. Reason: The PMID:10679296 source has a positive Golgi-localization abstract, with its complete original still unavailable. Independent full-length human EXT2 imaging in PMID:35137078 Figure 7A shows GM130 overlap, and the membrane-anchored protein context supports GO:0000139. This refines Golgi apparatus through the current part_of relationship; fluorescence alone is not claimed to resolve a membrane leaflet or the historical source protocol. Proposed replacements: Golgi membrane |
| GO:0005794 Golgi apparatus | IEA GO_REF:0000120 | MODIFY | Summary: Refine the conserved Golgi location to Golgi membrane. Reason: The positive mouse PMID:10639137 assay establishes EXT2 Golgi redistribution in the coexpressed EXT1–EXT2 complex; it is not a human EXT2 localization experiment. Native human Reactome entity 2022888 identifies Q93063 residues 1–718 with its own Golgi-membrane compartment. PMID:35137078 Figure 7A and construct Methods independently place tagged full-length human EXT2 with GM130 in COS-7 cells; HPA supports Golgi localization in human SiHa cells. The imaging does not itself resolve membrane topology or all three natural products. Propagation Review Root cause: TERM SCOPING PROBLEM Sources checked: UniProtKB:P70428 SUPPORTS TRANSFER Actual mouse EXT2-GFP Golgi redistribution with human EXT1 in BHK cells, PMID:10639137. ensembl:ENSMUSP00000028623 SUPPORTS TRANSFER Current sequence is exactly the mouse donor. UniProtKB-SubCell:SL-0132 SUPPORTS TRANSFER Valid broad Golgi mapping; native target supports membrane refinement. Proposed replacements: Golgi membrane |
| GO:0005794 Golgi apparatus | ISS GO_REF:0000024 | MODIFY | Summary: Refine the conserved Golgi location to Golgi membrane. Reason: The positive mouse PMID:10639137 assay establishes EXT2 Golgi redistribution in the coexpressed EXT1–EXT2 complex; it is not a human EXT2 localization experiment. Native human Reactome entity 2022888 identifies Q93063 residues 1–718 with its own Golgi-membrane compartment. PMID:35137078 Figure 7A and construct Methods independently place tagged full-length human EXT2 with GM130 in COS-7 cells; HPA supports Golgi localization in human SiHa cells. The imaging does not itself resolve membrane topology or all three natural products. Propagation Review Root cause: TERM SCOPING PROBLEM Sources checked: UniProtKB:P70428 SUPPORTS TRANSFER Actual mouse EXT2-GFP Golgi redistribution with human EXT1 in BHK cells, PMID:10639137. Proposed replacements: Golgi membrane |
| GO:0006024 glycosaminoglycan biosynthetic process | IDA PMID:12907669 In vitro polymerization of heparan sulfate backbone by the E... | MODIFY | Summary: Use the specific heparan-sulfate proteoglycan biosynthetic process. Reason: The human EXT1–EXT2 complex in PMID:12907669 elongates free K5-derived HS-backbone acceptors; the isolated EXT2 preparation does not appreciably polymerize them. PMID:36593275 independently resolves EXT2 GlcNAc-transfer chemistry and chain elongation on glypican-1. These establish EXT2 participation in the specific biosynthetic process, not merely necessity. The replacement is a lateral specificity correction, not a current descendant of the broad glycosaminoglycan term or a claim that the older assay synthesized a proteoglycan core. Proposed replacements: heparan sulfate proteoglycan biosynthetic process |
| GO:0008375 acetylglucosaminyltransferase activity | IDA PMID:12907669 In vitro polymerization of heparan sulfate backbone by the E... | MODIFY | Summary: The complex contribution is specifically alpha-1,4-GlcNAc transfer onto HS acceptors. Reason: The source contributes_to qualifier is preserved. PMID:12907669 measures sugar transfer in human EXT preparations, with weak and variable isolated EXT2 GlcNAc activity. Modern human structures and catalytic substitutions in PMID:36402845 and PMID:36593275 independently establish EXT2 GT-A as the principal GlcNAc-transfer site of the heterocomplex. GO:0050508 specifies this reaction and is an is_a descendant of the imported acetylglucosaminyltransferase activity. |
| GO:0008375 acetylglucosaminyltransferase activity | IEA GO_REF:0000107 | MODIFY | Summary: Specify the heparan-chain GlcNAc-transfer reaction. Reason: The exact mouse donor/Ensembl chain has experimental PMID:10639137 grounding. Human PMID:36402845 and PMID:36593275 directly assay N-terminally truncated EXT1–EXT2 heterocomplexes with defined UDP-sugar donors and glycan acceptors. Mutational and structural evidence assigns the major GlcNAc-transfer step to EXT2 GT-A and the major GlcA-transfer step to EXT1 GT-B. These are human complex assays, not experiments on every natural EXT2 product. GO:0050508 captures addition of GlcNAc to the terminal GlcA of the HS precursor, rather than generic acetylglucosaminyl transfer. Propagation Review Root cause: TERM SCOPING PROBLEM Sources checked: UniProtKB:P70428 SUPPORTS TRANSFER Actual donor sequence and experimental sugar-transfer source verified; independent human complex chemistry resolves specificity. ensembl:ENSMUSP00000028623 SUPPORTS TRANSFER Exact mouse product identity. |
| GO:0009101 glycoprotein biosynthetic process | IEA GO_REF:0000002 | MODIFY | Summary: Specify heparan-sulfate proteoglycan biosynthesis. Reason: IPR004263 genuinely matches EXT2 and carries the broad glycoprotein-biosynthesis mapping. Human PMID:36593275 directly assays HS elongation on glypican-1 as well as defined oligosaccharides; the EXT2 GlcNAc-transfer site performs part of that synthesis. GO:0015012 is the supported descendant, not an inference from general glycosyltransferase similarity or knockout necessity alone. Propagation Review Root cause: TERM SCOPING PROBLEM Sources checked: InterPro:IPR004263 SUPPORTS TRANSFER Actual target match and broad mapping verified; specific synthetic participation is independently measured. Proposed replacements: heparan sulfate proteoglycan biosynthetic process |
| GO:0015012 heparan sulfate proteoglycan biosynthetic process | IDA PMID:36593275 Structural basis for heparan sulfate co-polymerase action by... | ACCEPT | Summary: EXT2 directly performs the GlcNAc-addition step in heparan sulfate proteoglycan biosynthesis. Reason: PMID:36593275 Figure 4 and operative Results/Methods show heparan-sulfate extension on glypican-1 and defined oligosaccharide acceptors by purified human EXT1–EXT2. EXT2 GT-A active-site mutants markedly reduce GlcNAc transfer and co-polymerization; UDP-Glo assays are corroborated by glypican mobility shifts and MALDI product analysis. This is direct participation in chain synthesis, beyond a requirement inferred only from knockout. It does not assign synthesis of the proteoglycan core protein or the entire linkage region to EXT2. The preserved raw import contains equivalent UniProt and ComplexPortal assertions represented by one seeded context. Supporting Evidence: |
| GO:0015012 heparan sulfate proteoglycan biosynthetic process | IEA GO_REF:0000117 | ACCEPT | Summary: Retain direct participation in heparan-sulfate proteoglycan synthesis. Reason: Current ARBA00026373 disjuncts 0 and 2 match the target. Human heterocomplex activity, glypican-1 extension and catalytic-site perturbations establish that EXT2 performs a glycan-extension step. This is synthetic participation, not merely a developmental consequence of reduced heparan sulfate. Soluble recombinant constructs and natural product boundaries remain distinct. Propagation Review Root cause: NO FAILURE CORE Sources checked: ARBA:ARBA00026373 SUPPORTS TRANSFER Actual matching current condition sets; independent human catalytic evidence supports the process. |
| GO:0015012 heparan sulfate proteoglycan biosynthetic process | IMP PMID:17761672 Contribution of EXT1, EXT2, and EXTL3 to heparan sulfate cha... | ACCEPT | Summary: Human EXT2 performs a necessary catalytic step in heparan-sulfate chain biosynthesis. Reason: PMID:17761672 uses human-targeted EXT2 siRNA in HEK293 cells, verifies transcript depletion and shows shorter metabolically labeled HS chains and reduced HS/chondroitin-sulfate ratio. Its separate overexpression constructs are mouse brain cDNAs and are not reclassified as human assays. Independent human-complex PMID:36593275 identifies the GlcNAc-transfer step performed by EXT2, so the biosynthetic participation is supported beyond the knockdown phenotype. |
| GO:0015020 glucuronosyltransferase activity | IDA PMID:12907669 In vitro polymerization of heparan sulfate backbone by the E... | MODIFY | Summary: Specify the measured secondary GlcA-transfer reaction without making it the principal EXT2 activity. Reason: PMID:12907669 reports low but readily detected GlcA transfer by affinity-captured soluble human EXT2, including a single-sugar extension of a GlcNAc-terminated K5 acceptor. This supports the more specific beta-1,4-GlcA-transfer term already represented for the gene. Isoelectric-focusing evidence against endogenous EXT1 contamination is reported but not displayed. Modern human-complex studies assign the principal GlcA site to EXT1; the refinement retains the historical measured capacity without asserting equal catalytic contributions or efficient EXT2-alone polymerization. Proposed replacements: N-acetylglucosaminyl-proteoglycan 4-beta-glucuronosyltransferase activity |
| GO:0015020 glucuronosyltransferase activity | IEA GO_REF:0000120 | KEEP AS NON CORE | Summary: Retain contextual GlcA-transfer capacity without making it the principal EXT2 reaction. Reason: ARBA00043286 disjunct 2 matches the target, and the mouse/Ensembl source chain is exact. Older positive evidence must be retained: PMID:12907669 detects low GlcA addition by bead-captured, secreted human EXT2 lacking residues 1–46, although that preparation does not appreciably polymerize chains; PMID:11256613 reports positive GlcA transfer from full-length bovine EXT2 in Pichia lysates with mock controls. PMID:10639137 mouse assays are strongly dependent on the cellular/coexpression context. The 2022/2023 human heterocomplex studies instead identify EXT1 as the major GlcA catalyst and do not require EXT2 GT-B for that reaction. Failed secretion of isolated EXT2 or proteolysis of isolated domains is not proof of universal inactivity. Retain measured transfer capacity as non-core without asserting that autonomous native human EXT2 supplies the physiological GlcA step or that all products are equivalent. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: ARBA:ARBA00043286 SUPPORTS TRANSFER Actual rule matches; biological scope is limited by the older positive assays and modern complex-level catalytic assignment. UniProtKB:P70428 SUPPORTS TRANSFER Experimental PMID:10639137 source is positive but context-dependent; no donor-count argument. ensembl:ENSMUSP00000028623 SUPPORTS TRANSFER Exact mouse product identity. |
| GO:0016020 membrane | HDA PMID:19946888 Defining the membrane proteome of NK cells. | UNDECIDED | Summary: The source-specific membrane-proteomics identification remains unresolved. Reason: PMID:19946888 reports membrane proteomics of human YTS natural-killer cells, but the operative original and target-level supplement were not recovered. The EXT2 identification, peptide assignment and membrane fraction therefore remain unverified. Independent human EXT2 Golgi localization in PMID:35137078 supports other annotations, but does not identify the compartment or validate the EXT2 hit in this screen. Retain the literal HDA source assertion as UNDECIDED; the study title alone does not resolve plasma membrane or Golgi membrane. |
| GO:0016020 membrane | IEA GO_REF:0000002 | MODIFY | Summary: Specify the supported Golgi and ER membrane locations. Reason: IPR015338 genuinely matches EXT2 but its catalytic-domain mapping does not alone identify a compartment. Native human Reactome entity 2022888 and human tagged Golgi imaging support Golgi membrane; the actual mouse EXT2 imaging and conserved membrane anchor support the ER-membrane context. Specify these two locations without claiming that HPA optically resolves a membrane leaflet, or that isolated soluble constructs model native localization. Propagation Review Root cause: TERM SCOPING PROBLEM Sources checked: InterPro:IPR015338 SUPPORTS TRANSFER Actual domain match; compartment specificity derives from independent participant/localization evidence. Proposed replacements: Golgi membrane endoplasmic reticulum membrane |
| GO:0016757 glycosyltransferase activity | IBA GO_REF:0000033 | MODIFY | Summary: Specify the principal HS-chain GlcNAc-transfer activity. Reason: The actual PTHR11062 tree places human EXT2 leaf PTN002476065 below PTN000115158, and its ungapped MSA exactly matches the 718-residue Q93063 sequence. The imported Golgi and glycosyltransferase IBDs are at PTN000115158; separate sugar-transfer IBDs occur at PTN000868988. No target-path NOT was found. Current UniProt PTHR48261 classification is a separate family/version record; its returned tree does not contain this target. Target self-WITH is valid experimental grounding, not circularity. Human PMID:36402845 and PMID:36593275 directly assay N-terminally truncated EXT1–EXT2 heterocomplexes with defined UDP-sugar donors and glycan acceptors. Mutational and structural evidence assigns the major GlcNAc-transfer step to EXT2 GT-A and the major GlcA-transfer step to EXT1 GT-B. These are human complex assays, not experiments on every natural EXT2 product. Refine the broad ancestral activity to GO:0050508. The separately retained GlcA annotations preserve the older contextual capacity; this refinement does not assert its absence. Propagation Review Root cause: TERM SCOPING PROBLEM Sources checked: PANTHER:PTN000115158 SUPPORTS TRANSFER Verified positive glycosyltransferase IBD and exact target MSA; independent human activity supports substrate specificity. |
| GO:0016757 glycosyltransferase activity | IDA PMID:12907669 In vitro polymerization of heparan sulfate backbone by the E... | MODIFY | Summary: Specify the principal GlcNAc-transfer chemistry instead of generic glycosyl transfer. Reason: PMID:12907669 demonstrates sugar transfer and enhanced polymerization in human EXT1–EXT2 preparations. Independent human structural and mutational experiments in PMID:36402845 and PMID:36593275 identify EXT2 GT-A as the principal alpha-1,4-GlcNAc-transfer site. Replace the generic glycosyltransferase term with GO:0050508, an is_a descendant. The separately reviewed historical low GlcA activity is retained with its own scope rather than being denied by this principal-function refinement. |
| GO:0016757 glycosyltransferase activity | IEA GO_REF:0000002 | MODIFY | Summary: Specify the principal HS-chain GlcNAc-transfer activity. Reason: IPR004263 and IPR015338 are genuine target family/domain matches. Human PMID:36402845 and PMID:36593275 directly assay N-terminally truncated EXT1–EXT2 heterocomplexes with defined UDP-sugar donors and glycan acceptors. Mutational and structural evidence assigns the major GlcNAc-transfer step to EXT2 GT-A and the major GlcA-transfer step to EXT1 GT-B. These are human complex assays, not experiments on every natural EXT2 product. GO:0050508 supplies substrate specificity; the separate contextual GlcA rows remain retained rather than being silently discarded. Propagation Review Root cause: TERM SCOPING PROBLEM Sources checked: InterPro:IPR004263 SUPPORTS TRANSFER Actual target family match. InterPro:IPR015338 SUPPORTS TRANSFER Actual target GT64-domain match; catalytic interpretation uses human assays. |
| GO:0042328 heparan sulfate N-acetylglucosaminyltransferase activity | NAS PMID:12907669 In vitro polymerization of heparan sulfate backbone by the E... | MODIFY | Summary: Use the reaction-specific alpha-1,4-GlcNAc-transferase term. Reason: The NAS source PMID:12907669 describes HS-acceptor GlcNAc transfer, with low variable activity in separately expressed EXT2 and stronger activity in the assembled complex. The specific alpha-1,4 reaction is independently resolved for human EXT2 by PMID:36402845 and PMID:36593275. GO:0050508 gives that defined chemistry; this is a reaction-specific lateral correction rather than a claimed current is_a relationship to GO:0042328. |
| GO:0042803 protein homodimerization activity | IDA NOT PMID:12907669 In vitro polymerization of heparan sulfate backbone by the E... | KEEP AS NON CORE | Summary: Retain the negated homodimerization result within the tested soluble-construct context. Reason: The NOT flag remains unchanged. PMID:12907669 reports that human EXT2 lacking its first 46 residues elutes near 80 kDa, consistent with a monomer, whereas the EXT1–EXT2 preparation elutes near 160 kDa. That supports the source-specific negative homodimerization result for this soluble preparation. It is not a universal absence claim: PMID:10639137 reports co-immunoprecipitation of differently tagged full-length mouse EXT2 molecules, which establishes self-association without resolving a homodimer stoichiometry. Different constructs, species and assays prevent a categorical contradiction or generalization to all three natural human products. |
| GO:0043541 UDP-N-acetylglucosamine transferase complex | IDA PMID:12907669 In vitro polymerization of heparan sulfate backbone by the E... | MODIFY | Summary: The source describes the EXT1–EXT2 complex, not the current dolichol-linked complex definition. Reason: PMID:12907669 directly captures and assays soluble human EXT1–EXT2, supported by the later human structures. Current GO:0043541 specifically describes the Alg7/Alg13/Alg14-type complex for dolichyl-linked oligosaccharide synthesis. Replace it with GO:0120504, the current EXT1–EXT2 complex. The positive experimental assembly is retained; this is a demonstrated term-scoping mismatch rather than a rejection of the source experiment. Proposed replacements: EXT1-EXT2 complex |
| GO:0043541 UDP-N-acetylglucosamine transferase complex | IPI PMID:36593275 Structural basis for heparan sulfate co-polymerase action by... | MODIFY | Summary: EXT2 belongs to the EXT1–EXT2 heparan sulfate polymerase complex; the imported component term denotes a different assembly. Reason: GO:0043541 currently describes the Alg7/Alg13/Alg14 complex involved in dolichol-linked oligosaccharide assembly. PMID:36593275 characterizes the distinct human EXT1–EXT2 heterodimer. The captured ComplexPortal CPX-8307 record identifies both correct subunits but retains GO:0043541, also present in the ComplexPortal-assigned EXT2 source row. Replace that component term with current GO:0120504 EXT1-EXT2 complex, whose definition explicitly names this heterodimer and cites both structural studies. This is a documented correction of complex identity, not a claimed descendant refinement or a change to the preserved source qualifier. Proposed replacements: EXT1-EXT2 complex |
| GO:0046982 protein heterodimerization activity | IPI PMID:12907669 In vitro polymerization of heparan sulfate backbone by the E... | ACCEPT | Summary: EXT2 heterodimerizes with EXT1 in the catalytic assembly. Reason: PMID:12907669 co-expresses soluble human FLAG-EXT1 and Myc/His-EXT2, recovers both tags together and detects the complex in approximately 160-kDa gel-filtration fractions. Independent human cryo-EM and mass-photometry experiments in PMID:36402845 and PMID:36593275 resolve the heterodimer. This is specific assembly evidence, not simply generic binding; the engineered soluble constructs are not treated as additional natural products. |
| GO:0050508 glucuronosyl-N-acetylglucosaminyl-proteoglycan 4-alpha-N-acetylglucosaminyltransferase activity | EXP PMID:36402845 Structure of the human heparan sulfate polymerase complex EX... | ACCEPT | Summary: EXT2 has the specific alpha-1,4-N-acetylglucosaminyltransferase activity used in heparan-sulfate elongation. Reason: PMID:36402845 mutational Results and Figure 4 identify a GlcNAc-transferase site in the C-terminal EXT2 GT-A domain. The purified human heterocomplex carrying EXT2 D538N/D540N has reduced GlcNAc transfer in the fluorescent oligosaccharide FACE assay; separately, full-length mutant EXT2 incompletely restores cell-surface HS in EXT2-deficient HeLa cells. These results support EXT2 catalysis within the complex. They do not require isolated EXT2 to function as a complete autonomous copolymerase. The same study also finds an EXT1 GT-A contribution, with a thermal-stability confound explicitly discussed for that partner mutant. Supporting Evidence: |
| GO:0050508 glucuronosyl-N-acetylglucosaminyl-proteoglycan 4-alpha-N-acetylglucosaminyltransferase activity | IDA PMID:36593275 Structural basis for heparan sulfate co-polymerase action by... | ACCEPT | Summary: EXT2 supplies the major alpha-1,4-N-acetylglucosaminyltransferase activity of the human heparan sulfate polymerase. Reason: PMID:36593275 Figure 4 and Results assign most GlcNAc transfer to EXT2 GT-A through substrate-bound structure and mutant heterocomplex assays. D538A and R569A greatly reduce transfer to the GlcA-terminated five-sugar acceptor and impair glypican-1 chain extension, although inactivation is not complete. MALDI analysis independently verifies the expected sugar-addition products of the wild-type complex. This directly supports the specific EXT2 activity within the assembled enzyme. Low EXT1-alone activity and the earlier study’s EXT1 GT-A effects are retained as evidence about the partner contribution, rather than treated as contradictions of EXT2 catalysis. |
| GO:0050508 glucuronosyl-N-acetylglucosaminyl-proteoglycan 4-alpha-N-acetylglucosaminyltransferase activity | IEA GO_REF:0000120 | ACCEPT | Summary: Retain the established HS-chain GlcNAc-transfer reaction. Reason: ARBA00090157 disjunct 0 matches EXT2, and the mouse/Ensembl source chain is verified. Official GO:0050508 and ENZYME 2.4.1.224 specify the same chain-extension reaction; the direct Rhea endpoint returned HTTP 403, so no successful independent Rhea retrieval is claimed. Human PMID:36402845 and PMID:36593275 directly assay N-terminally truncated EXT1–EXT2 heterocomplexes with defined UDP-sugar donors and glycan acceptors. Mutational and structural evidence assigns the major GlcNAc-transfer step to EXT2 GT-A and the major GlcA-transfer step to EXT1 GT-B. These are human complex assays, not experiments on every natural EXT2 product. Propagation Review Root cause: NO FAILURE CORE Sources checked: ARBA:ARBA00090157 SUPPORTS TRANSFER Current matching rule and substrate-specific annotation. UniProtKB:P70428 SUPPORTS TRANSFER Experimental donor source read. ensembl:ENSMUSP00000028623 SUPPORTS TRANSFER Exact mouse product identity. RHEA:16213 UNRESOLVED Reaction cross-reference corroborated by official GO and ENZYME; direct Rhea service unavailable. EC:2.4.1.224 SUPPORTS TRANSFER Actual official ENZYME reaction agrees with GO:0050508. |
| GO:0050509 N-acetylglucosaminyl-proteoglycan 4-beta-glucuronosyltransferase activity | IDA PMID:36593275 Structural basis for heparan sulfate co-polymerase action by... | ACCEPT | Summary: EXT2 contributes to the heterocomplex that transfers glucuronic acid, while the principal catalytic site is in EXT1. Reason: The imported assertion already uses contributes_to. PMID:36593275 assigns beta-1,4-GlcA transfer predominantly to the EXT1 GT-B site through substrate binding and mutant assays; EXT2 remains a physical component of the active human heterodimer. That supports the qualified complex-level activity without asserting that the EXT2 GT-B domain independently catalyzes the reaction. PMID:36402845 likewise finds that EXT2 R266A retains complex GlcA-transfer and rescue activity and describes a pseudo-GlcA-T domain. The 2023 failure to express full soluble EXT2 alone, and trace activities from partially proteolyzed isolated domains, do not justify a blanket claim of inactivity under all conditions. Earlier positive isolated-enzyme reports require their own source-specific assessment. |
| GO:0050509 N-acetylglucosaminyl-proteoglycan 4-beta-glucuronosyltransferase activity | IEA GO_REF:0000107 | KEEP AS NON CORE | Summary: Retain measured HS-substrate GlcA capacity as non-core. Reason: Older positive evidence must be retained: PMID:12907669 detects low GlcA addition by bead-captured, secreted human EXT2 lacking residues 1–46, although that preparation does not appreciably polymerize chains; PMID:11256613 reports positive GlcA transfer from full-length bovine EXT2 in Pichia lysates with mock controls. PMID:10639137 mouse assays are strongly dependent on the cellular/coexpression context. The 2022/2023 human heterocomplex studies instead identify EXT1 as the major GlcA catalyst and do not require EXT2 GT-B for that reaction. Failed secretion of isolated EXT2 or proteolysis of isolated domains is not proof of universal inactivity. Retain measured transfer capacity as non-core without asserting that autonomous native human EXT2 supplies the physiological GlcA step or that all products are equivalent. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:P70428 SUPPORTS TRANSFER Actual donor and original assay context traced; independent bovine and human positive preparations support bounded capacity. ensembl:ENSMUSP00000028623 SUPPORTS TRANSFER Exact mouse product identity, not an independent assay. |
| GO:0050509 N-acetylglucosaminyl-proteoglycan 4-beta-glucuronosyltransferase activity | NAS PMID:12907669 In vitro polymerization of heparan sulfate backbone by the E... | KEEP AS NON CORE | Summary: Historical human EXT2 preparations have positive low GlcA-transfer capacity. Reason: PMID:12907669 explicitly reports low but readily detected GlcA transfer by the separately expressed soluble human EXT2 preparation, with a single-GlcA product shoulder and no appreciable chain polymerization. Its reported isoelectric-focusing purity check is not displayed. Retain the specific capacity as non-core rather than claiming universal inactivity: modern human structures and mutations identify EXT1 as the principal GlcA-transfer site and EXT2 as the principal GlcNAc-transfer site in the heterocomplex. These studies do not establish equivalent efficiency or activity of every natural EXT2 product. |
| GO:0070062 extracellular exosome | HDA PMID:19199708 Proteomic analysis of human parotid gland exosomes by multid... | KEEP AS NON CORE | Summary: EXT2 was detected in a human parotid exosome-rich preparation. Reason: PMID:19199708 actual Supplementary Table 2 row 429 identifies IPI00004047.1 as EXT2 with two sequences, two spectra and 5.3% coverage; Supplementary Table 1 rows 4105–4106 supplies the corresponding peptide records. Methods describes stimulated parotid saliva from one donor, differential centrifugation, an exosome-rich pellet, IPI database searching and a 5% protein false-positive rate. Retain this source-specific extracellular-vesicle detection as non-core. Marker imaging validates the preparation, not EXT2 itself, and the historical isoform-1 database label does not establish product-specific localization or physiological secretion of every EXT2 product. Supporting Evidence: |
| GO:1902494 catalytic complex | IDA PMID:10679296 Association of EXT1 and EXT2, hereditary multiple exostoses ... | MODIFY | Summary: The catalytic assembly is specifically the EXT1–EXT2 complex. Reason: The complete PMID:10679296 original is unavailable, although its abstract positively reports EXT association. Independently read human PMID:12907669 and the two modern structural studies isolate, assay and resolve the EXT1–EXT2 heterodimer. Replace the broad catalytic complex with GO:0120504, whose current is_a ancestors include GO:1902494. This preserves the original source while distinguishing independent corroboration from verification of its unread protocol. Proposed replacements: EXT1-EXT2 complex |
| GO:1902494 catalytic complex | IDA PMID:36593275 Structural basis for heparan sulfate co-polymerase action by... | MODIFY | Summary: The active catalytic assembly is specifically the EXT1–EXT2 complex. Reason: PMID:36593275 resolves and assays the human EXT1–EXT2 heterodimer. GO:0120504 EXT1-EXT2 complex is the current specific descendant of the broad GO:1902494 catalytic complex term and matches the observed assembly. Preserve the source is_active_in qualifier. Anchor-deleted soluble constructs used in purification are not additional natural EXT2 products or evidence of endogenous extracellular localization. Proposed replacements: EXT1-EXT2 complex Supporting Evidence: |
Loading supporting content…
Download this section (compressed HTML)Q: What accounts for the low GlcA-transfer activity measured in the older soluble human EXT2 preparation, and does that capacity contribute measurably to chain extension by an intact human EXT1–EXT2 complex?
Q: How do the three source-described human EXT2 products differ in expression, membrane localization, EXT1 assembly and GlcNAc-transfer activity?
Experiment: Compare sequence-verified human full-length and soluble EXT2 preparations with matched EXT1–EXT2 complexes using the older oligosaccharide acceptors and modern defined primers. Quantify intact enzyme, endogenous EXT1 carryover and assembly before measuring both single-sugar transfers and product-resolved chain extension. Include catalytic substitutions, donor-omission controls and membrane-reconstituted complexes to separate positive capacity from its quantitative contribution in the assembled enzyme.
Hypothesis: Preparation integrity, partner content and acceptor conditions help explain the difference between the older positive soluble-EXT2 GlcA result and the predominant EXT1 GlcA site in modern heterocomplex assays.
Type: Controlled biochemical comparison and product analysis
Experiment: Verify each transcript and protein sequence, then compare matched-expression constructs in an EXT2-deficient human cell background. Measure protein abundance, Golgi and ER distribution, EXT1 association and cell-surface heparan sulfate alongside enzyme assays on intact purified complexes. Normalize activity to stable assembled protein and retain endogenous transcript/protein measurements to distinguish a construct phenotype from physiological product usage.
Hypothesis: Sequence differences among the 718-, 728- and 751-residue products may alter abundance, trafficking or complex activity without having identical effects on each property.
Type: Product-specific cellular complementation and enzymology
Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)