CHST14

Summary: Retain Golgi localization with topology and disease-model limits.

Reason: The subcellular-location mapping is corroborated by endogenous human MCF10a Golgi colocalization and human D4ST1 fluorescent-fusion overlap with GM130 in COS-7 cells (PMID:29976758). Membrane topology remains sequence-based rather than established by protease protection.

Summary: Retain Golgi localization with topology and disease-model limits.

Reason: Reactome reaction 2022063 places Q8NCH0 entity 1971536 in the Golgi membrane. Independent human-cell and human-construct microscopy corroborates the compartment (PMID:29976758); type-II topology remains sequence-based.

Summary: Retain Golgi localization with topology and disease-model limits.

Reason: Reactome disease event 3636919 places a CHST14 mutant set in the Golgi membrane and links to normal reaction 2022063. Retain the modeled compartment with independent WT localization support (PMID:29976758), without claiming correct delivery or activity for every mutant.

Summary: Retain directly demonstrated dermatan 4-sulfotransferase activity.

Reason: PMID:11470797 expresses the sequence-verified human AF401222 376-residue ORF in CHO/Tag cells. Radiolabeled PAPS transfer in extracts and conditioned medium, with mock and C4ST controls and product digestion/chromatography, establishes C4 sulfation of GalNAc adjacent to iduronic acid. These are not homogeneous purified-enzyme assays. PMID:29976758 independently reconstitutes the reaction with affinity-purified human residues 61–376 and DS-epi1.

Summary: Retain the specific sulfate-transfer reaction.

Reason: EC:2.8.2.35 and RHEA:48052 match the authenticated reaction and GO:0001537. Human recombinant sulfate-transfer assays establish GalNAc C4 modification adjacent to iduronic acid (PMID:11470797), independently corroborated by purified human residues 61–376 with DS-epi1 (PMID:29976758). The original CHO/Tag extracts and medium were not homogeneous enzyme preparations.

Summary: Retain the explicitly modeled wild-type catalytic step.

Reason: Reactome 2022063 identifies catalyst 2022068/Q8NCH0 entity 1971536 and the PAPS-to-PAP/dermatan-sulfation reaction. Direct human recombinant biochemistry independently supports this normal catalytic step (PMID:11470797); the original experiment used extracts and medium, not homogeneous purified enzyme.

Summary: Retain normal gene-product activity while separating the defective-reaction source.

Reason: Reactome 3636919 is a failed CHST14-mutant reaction with no productive output, linked to normalReaction 2022063. The unqualified gene-level capability is retained from WT primary biochemistry (PMID:11470797); this does not assign activity to the defective variants. Patient-cell evidence supplies disease context (PMID:20004762).

Summary: Specify binding to the verified CFTR transmembrane transporter.

Reason: Three exact MINT records connect human CHST14/Q8NCH0 to CFTR/P13569: a traditional MaMTH record attributed to Figure EV5 and WT/F508del MaMTH-HTS records attributed to Dataset EV2 (PMID:35156780). HEK293 host, fusion tags and the CFTR position-508 deletion remain explicit. GO:0044325 requires binding to a transmembrane transporter, an established CFTR class; it does not require CHST14 to regulate transport. Retain this useful assay-level refinement at non-core prominence. A proposed native Golgi/channel mechanism is neither required by the term nor demonstrated, and CHST14-specific co-IP or purified binary binding is not claimed. Supported generic NC decisions elsewhere do not prohibit a definition-compatible class refinement here.

Proposed replacements: transmembrane transporter binding

Summary: Retain Golgi localization with topology and disease-model limits.

Reason: ARBA00028708 includes IPR005331 plus Primates, matching the human protein. Endogenous human-cell and human-construct microscopy independently supports Golgi localization (PMID:29976758); this corroborates the location without upgrading sequence-based topology to a direct protection assay.

Summary: Refine the target activity while retaining the correct broad source premise.

Reason: The original GO:0008146 assertion is correct: CHST14 transfers sulfate from PAPS to a carbohydrate hydroxyl. The proposed GO:0001537 specifies the target reaction directly demonstrated by human recombinant assays (PMID:11470797; PMID:29976758). This is target-level curation under the guideline to refine overly broad terms, not a claim that dermatan specificity arose at PTN000951429 or that its historical IBD placement is wrong. Family membership and WITH/FROM composition do not reconstruct that node. Existing specific annotations corroborate the reaction; this replacement adds no NEW annotation and changes no source evidence or qualifier.

Proposed replacements: dermatan 4-sulfotransferase activity

Summary: Refine the target activity while retaining the correct broad source premise.

Reason: The authentic ARBA00028338 condition set uses IPR005331/IPR018011 with Eukaryota, and those InterPro entries map the broad, correct sulfotransferase capability. Human CHST14 assays establish the dermatan-specific reaction (PMID:11470797; PMID:29976758), supporting target-level refinement to GO:0001537. The recommendation does not rewrite the rule output or claim every family member shares dermatan specificity. It refines an existing target annotation; the already accepted specific rows are corroborating coverage, not a new assertion.

Proposed replacements: dermatan 4-sulfotransferase activity

Summary: Retain broad membrane localization at its stated evidence scope.

Reason: IPR005331/IPR018011 support the broad membrane assignment. The predicted type-II precursor and independent Golgi microscopy are compatible with that location (PMID:11470797; PMID:29976758). Retaining a correct compartment here does not prohibit an informative target-specific catalytic refinement in a separate MF review.

Summary: Retain broad membrane localization at its stated evidence scope.

Reason: The original paper predicts an N-terminal membrane segment and type-II topology. Most recombinant activity is in medium; suggested proteolytic shedding remains untested and does not negate the full-length membrane precursor (PMID:11470797). Independent Golgi microscopy supports membrane localization (PMID:29976758), without establishing native topology by protease protection.

Summary: Specify dermatan sulfate proteoglycan biosynthesis.

Reason: The IPR018011 mapping correctly captures carbohydrate biosynthesis. Human CHST14 itself performs the PAPS-dependent sulfation of the dermatan chain, so GO:0050651 is a more informative process than generic carbohydrate synthesis. It does not assign backbone polymerization or epimerase catalysis to CHST14. This refines an existing source assertion and duplicates no new process proposal.

Proposed replacements: dermatan sulfate proteoglycan biosynthetic process

Summary: The cited PAPS-motif evidence does not establish separate phosphate-ion binding.

Reason: PMID:11470797 identifies putative PAPS 5-prime-phosphosulfate and 3-prime-phosphate contact regions by sequence comparison; its Discussion calls those assignments hypotheses. GO:0042301 names phosphate-ion binding. Contacts with phosphate groups covalently contained in PAPS do not establish that separate ligand activity. Retain MARK_AS_OVER_ANNOTATED for this extrapolation: the source neither assays free phosphate-ion binding nor provides a negative result that would establish its biochemical impossibility. The demonstrated sulfate-transfer reaction is unaffected.

Summary: Retain the experimentally established biosynthetic sulfation step.

Reason: Human recombinant CHST14 transfers sulfate onto the dermatan chain, directly performing a biosynthetic chemical step (PMID:11470797); later purified-component work corroborates that role (PMID:29976758). The original acts_upstream_of_or_within qualifier is broad enough to include participation within the process, so it is not contradicted by this evidence. It understates the more precise involved_in interpretation used in the core and sibling source, but remains preserved as the historical qualifier. Patient fibroblast changes corroborate dermatan-sulfate deficiency (PMID:20004762), without assigning every developmental consequence to the enzyme.

Summary: Retain participation in the explicitly modeled dermatan biosynthetic pathway.

Reason: Reactome 2022923 contains the target reaction 2022063 alongside epimerization, other sulfation and secretion events. CHST14 directly catalyzes its own sulfate-transfer step; this is more specific than mere membership in the pathway and does not transfer other enzymes or transporters activities to CHST14.

Summary: Refine dermatan metabolism to the demonstrated biosynthetic direction.

Reason: The current PTN002637560 assertion correctly records dermatan metabolism. Human biochemical evidence supports the more informative biosynthetic direction, GO:0050651 (PMID:11470797; PMID:29976758). Q8NCH0 self-inclusion is legitimate experimental grounding, not circularity. Its being the sole listed extant source does not establish the breadth or placement of the historical ancestral node; neither the tree nor MSA was reconstructed. The replacement is a target-level refinement, not a revised historical IBD assertion.

Proposed replacements: dermatan sulfate proteoglycan biosynthetic process

Summary: Refine dermatan metabolism to the demonstrated biosynthetic direction.

Reason: The original IDA source directly assays sulfate transfer in the biosynthetic direction, not dermatan degradation. The narrower GO:0050651 is supported by the performed chemical step and human reconstitution/cellular evidence. The existing metabolic assertion is sound but less informative; its source fields remain unchanged.

Proposed replacements: dermatan sulfate proteoglycan biosynthetic process

Summary: Retain original parotid exosome-fraction detection as non-core.

Reason: The authentic publisher Table S2 row 389 identifies CHST14/IPI00044326.3 with two sequences, two spectra and 9.3% sequence coverage. Table S1 rows 3817–3818 supply the two peptide sequences, both matching the current Q8NCH0 sequence. Actual Methods describe a single human donor, a 200,000g exosome-rich pellet and MudPIT protein-group analysis; this is not CHST14-specific vesicle imaging or an enzyme-activity assay. Retain the original HDA location without inferring exosome biogenesis or a defining extracellular catalytic role. The full normal-fetch variant and original tables were inspected separately from the unchanged, less complete main cache.