Galactosylceramide 3'-sulfotransferase (cerebroside sulfotransferase, EC 2.8.2.11), a single-pass type II membrane protein of the Golgi apparatus whose catalytic domain faces the Golgi lumen. It transfers sulfate from 3'-phosphoadenosine 5'-phosphosulfate to the 3-hydroxyl of the non-reducing terminal beta-galactose of galactolipids, converting galactosylceramide into sulfatide (3-O-sulfogalactosylceramide) and galactosyl alkylacylglycerol into seminolipid. It also sulfates lactosylceramide and galactosyl diacylglycerol in vitro, preferring beta-galactosides at the non-reducing end of lipid-linked sugar chains. This is the only gene in mammals that makes sulfoglycolipids: mice lacking it have no sulfatide in brain and no seminolipid in testis, develop paranodal junction abnormalities with hindlimb weakness, tremor and progressive ataxia, and males are sterile through a block in spermatogenesis before the first meiotic division. Sulfatide is a major myelin lipid and the epitope of the oligodendrocyte marker O4. In humans the gene is expressed in kidney proximal tubule and gastric mucosa and is strongly upregulated in renal cell carcinoma and other tumours. A 2026 report additionally proposed that GAL3ST1 acts as a histone sulfotransferase modifying tyrosine 99 of cytosolic nascent histone H3, but that modification is itself disputed and the proposed reaction is difficult to reconcile with the enzyme's lumenal Golgi topology.
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0001733 galactosylceramide sulfotransferase activity | IBA GO_REF:0000033 | ACCEPT | Summary: Phylogenetic assertion of the family's defining activity, seeded by mouse Gal3st1 (MGI:1858277) and by human GAL3ST1's own experimental annotation. Reason: Core molecular function. The target appearing in its own WITH/FROM list is expected here: human GAL3ST1 has a direct IDA/EXP annotation for this term, which is one of the descendant evidences behind the IBD node. Supporting Evidence: PMID:8830034 We have purified 3'-phosphoadenosine-5'-phosphosulfate:GalCer sulfotransferase [EC 2.8.2.11] from a human renal cancer cell line SMKT-R3 |
| GO:0006682 galactosylceramide biosynthetic process | IBA GO_REF:0000033 | ACCEPT | Summary: Phylogenetic assertion that the family acts in galactosylceramide-derived lipid biosynthesis, seeded by mouse Gal3st1. Reason: Accepted: the product, sulfatide, is a sulfated galactosylceramide, so the enzyme does act in the biosynthesis of a galactosylceramide species even though it consumes rather than makes the unsulfated parent lipid. The more precise description of the step is sulfolipid biosynthetic process (GO:0046506), which is proposed in core_functions; the IBA node placement itself is not in question and is well grounded in the mouse knockout. Supporting Evidence: PMID:11917099 Cst(-/-) mice lacked sulfatide in brain and seminolipid in testis, proving that a single gene copy is responsible for their biosynthesis. |
| GO:0042552 myelination | IBA GO_REF:0000033 | ACCEPT | Summary: Phylogenetic transfer of the mouse Gal3st1 myelination phenotype to the human orthologue. Reason: Strongly grounded. The donor, mouse Gal3st1 (MGI:1858277), is the knockout that removes sulfatide from brain and produces paranodal junction defects with progressive ataxia; a single well-characterized donor is not weak support when the node judgement rests on evidence of this quality. Sulfatide is a major myelin lipid in humans as well, and GAL3ST1 is the only enzyme that makes it. Supporting Evidence: PMID:11917099 Although compact myelin was preserved, Cst(-/-) mice displayed abnormalities in paranodal junctions. PMID:11917099 These data show a critical role for sulfoglycolipids in myelin function and spermatogenesis. |
| GO:0000139 Golgi membrane | IEA GO_REF:0000044 | ACCEPT | Summary: Transfer of the UniProt SUBCELLULAR LOCATION 'Golgi apparatus membrane; single-pass type II membrane protein'. Reason: Correct and specific. The type II topology predicted from the cDNA places the catalytic domain in the Golgi lumen, where its glycolipid acceptors are. Supporting Evidence: PMID:9030544 The deduced amino acid sequence predicts a type II transmembrane topology and contains two potential N-glycosylation sites. |
| GO:0001733 galactosylceramide sulfotransferase activity | IEA GO_REF:0000120 | ACCEPT | Summary: Automated assignment of EC 2.8.2.11 and the corresponding RHEA reactions. Reason: Agrees with the direct experimental evidence; same term and action as the EXP, IDA, ISS and TAS rows. |
| GO:0009247 glycolipid biosynthetic process | IEA GO_REF:0000002 | MODIFY | Summary: InterPro domain-based assignment of a general glycolipid biosynthesis term. Reason: Correct but under-specific. The distinguishing product of this enzyme is a sulfated glycolipid, so the sulfolipid biosynthetic process term captures what GAL3ST1 uniquely contributes. Proposed replacements: sulfolipid biosynthetic process |
| GO:0016020 membrane | IEA GO_REF:0000002 | MODIFY | Summary: InterPro domain-based assignment of the generic membrane term. Reason: Uninformatively general when the specific compartment is known and already annotated to this gene. Proposed replacements: Golgi membrane |
| GO:0050694 galactose 3-O-sulfotransferase activity | IEA GO_REF:0000117 | ACCEPT | Summary: ARBA assignment of the chemistry-level parent term, 3-O-sulfation of a galactose residue. Reason: Chemically accurate and consistent with the measured acceptor preference: the enzyme sulfates position 3 of non-reducing terminal beta-galactose across several lipid classes, not only galactosylceramide. Retained alongside the more specific GO:0001733. Supporting Evidence: PMID:8830034 These observations suggest that the sulfotransferase prefers beta-glycoside, especially beta-galactoside, at the nonreducing termini of sugar chains attached to a lipid moiety. |
| GO:0006681 galactosylceramide metabolic process | IEA GO_REF:0000107 | ACCEPT | Summary: Orthology transfer from mouse Gal3st1. Reason: Correct; galactosylceramide is the preferred acceptor and is consumed by the reaction. |
| GO:0006688 glycosphingolipid biosynthetic process | TAS Reactome:R-HSA-9840309 | ACCEPT | Summary: Reactome places GAL3ST1 in glycosphingolipid biosynthesis. Reason: Correct; sulfatide is a glycosphingolipid and GAL3ST1 catalyses its terminal biosynthetic step. |
| GO:0006665 sphingolipid metabolic process | IEA GO_REF:0000041 | ACCEPT | Summary: UniPathway mapping (UPA00222, sphingolipid metabolism). Reason: Correct, though general; matches the UniProt PATHWAY line and the direct evidence. |
| GO:0001733 galactosylceramide sulfotransferase activity | TAS Reactome:R-HSA-9844587 | ACCEPT | Summary: Reactome reaction 'GAL3ST1 transfers sulfate to GalCer,LacCer'. Reason: Correct, and the inclusion of lactosylceramide matches the purified-enzyme acceptor panel. Supporting Evidence: PMID:8830034 GalCer was the best acceptor for the purified enzyme. LacCer, GalAAG, and GalDG were also good acceptors. |
| GO:0001733 galactosylceramide sulfotransferase activity | EXP PMID:9030544 Molecular cloning and expression of cDNA encoding human 3'-p... | ACCEPT | Summary: Expression of the cloned cDNA in COS-1 cells raised galactosylceramide sulfotransferase activity 8- to 16-fold and produced sulfatide. Reason: Direct experimental identification of this gene product as the sulfatide synthase; core molecular function. Supporting Evidence: PMID:9030544 When the cDNA was inserted into the expression vector pSVK3 and transfected into COS-1 cells, galactosylceramide sulfotransferase activity in the transfected cells increased from 8- to 16-fold over that of controls, and the enzyme product, sulfatide, was expressed on the transformed cells. |
| GO:0001733 galactosylceramide sulfotransferase activity | ISS GO_REF:0000024 | ACCEPT | Summary: Curator sequence-similarity transfer from mouse Gal3st1 (UniProtKB:Q9JHE4). Reason: Redundant with, but consistent with, the direct human experimental evidence; same term and action as the other GO:0001733 rows. |
| GO:0001733 galactosylceramide sulfotransferase activity | IDA PMID:8830034 Purification and characterization of 3'-phosphoadenosine-5'-... | ACCEPT | Summary: Activity measured on the enzyme purified to homogeneity from human SMKT-R3 renal cancer cells. Reason: The strongest single piece of evidence for the core molecular function - a purified human enzyme with a defined acceptor panel. Supporting Evidence: PMID:8830034 GalCer was the best acceptor for the purified enzyme. LacCer, GalAAG, and GalDG were also good acceptors. |
| GO:0006665 sphingolipid metabolic process | IDA PMID:8830034 Purification and characterization of 3'-phosphoadenosine-5'-... | ACCEPT | Summary: The purified enzyme acts on galactosylceramide and lactosylceramide, both sphingolipids. Reason: Directly supported; same term and action as the UniPathway IEA row. |
| GO:0006681 galactosylceramide metabolic process | ISS GO_REF:0000024 | ACCEPT | Summary: Curator sequence-similarity transfer from mouse Gal3st1. Reason: Consistent with the human experimental data; same term and action as the Ensembl Compara row. |
| GO:0046486 glycerolipid metabolic process | IDA PMID:8830034 Purification and characterization of 3'-phosphoadenosine-5'-... | ACCEPT | Summary: The purified enzyme also sulfates galactosyl alkylacylglycerol (GalAAG) and galactosyl diacylglycerol (GalDG). Reason: Not a stray annotation: GalAAG is the direct precursor of seminolipid, and the mouse knockout shows that the same enzyme makes seminolipid in testis as makes sulfatide in brain. The glycerolipid arm of the activity is therefore physiologically real, not just an in vitro curiosity. Supporting Evidence: PMID:8830034 GalCer was the best acceptor for the purified enzyme. LacCer, GalAAG, and GalDG were also good acceptors. PMID:11917099 Cst(-/-) mice lacked sulfatide in brain and seminolipid in testis, proving that a single gene copy is responsible for their biosynthesis. |
| GO:0000139 Golgi membrane | TAS Reactome:R-HSA-9844587 | ACCEPT | Summary: Reactome places the GAL3ST1 reaction at the Golgi membrane. Reason: Correct and specific; same term and action as the UniProt-derived IEA row. |
| GO:0005886 plasma membrane | TAS PMID:9030544 Molecular cloning and expression of cDNA encoding human 3'-p... | REMOVE | Summary: A 2003 legacy author-statement annotation placing GAL3ST1 at the plasma membrane. Reason: This misreads the cited sentence. PMID:9030544 reports that the enzyme's *product*, sulfatide, was displayed on the surface of transfected cells; it says nothing about the enzyme being at the plasma membrane. The same paper predicts a type II transmembrane topology, and UniProt places the protein in the Golgi apparatus membrane. Removing a TAS row that conflicts with the protein's established compartment, not an experimental localization. Supporting Evidence: PMID:9030544 When the cDNA was inserted into the expression vector pSVK3 and transfected into COS-1 cells, galactosylceramide sulfotransferase activity in the transfected cells increased from 8- to 16-fold over that of controls, and the enzyme product, sulfatide, was expressed on the transformed cells. |
| GO:0006487 protein N-linked glycosylation | TAS PMID:9030544 Molecular cloning and expression of cDNA encoding human 3'-p... | REMOVE | Summary: A 2003 legacy author-statement annotation asserting involvement in protein N-linked glycosylation. Reason: A substrate/enzyme inversion. The only relevant statement in PMID:9030544 is that the deduced sequence contains two potential N-glycosylation sites - i.e. GAL3ST1 is itself a likely acceptor of N-glycans. That does not make it a participant in the N-glycosylation process, and GAL3ST1 has no oligosaccharyltransferase or glycan-processing activity. Supporting Evidence: PMID:9030544 The deduced amino acid sequence predicts a type II transmembrane topology and contains two potential N-glycosylation sites. |
| GO:0008146 sulfotransferase activity | TAS PMID:9030544 Molecular cloning and expression of cDNA encoding human 3'-p... | MODIFY | Summary: Generic sulfotransferase assignment from the cloning paper. Reason: Correct but under-specific; the same paper identifies the enzyme as 3'-phosphoadenylylsulfate:galactosylceramide 3'-sulfotransferase, EC 2.8.2.11. Proposed replacements: galactosylceramide sulfotransferase activity Supporting Evidence: PMID:9030544 We have isolated a cDNA clone encoding human 3'-phosphoadenylylsulfate:galactosylceramide 3'-sulfotransferase (EC 2.8.2.11). |
| GO:0016020 membrane | TAS PMID:9030544 Molecular cloning and expression of cDNA encoding human 3'-p... | MODIFY | Summary: Legacy author-statement annotation that GAL3ST1 is a membrane protein, based on the predicted type II transmembrane topology. Reason: The underlying observation is sound but the term is uninformatively general; the specific compartment is known. Same action as the InterPro-derived GO:0016020 row. Proposed replacements: Golgi membrane Supporting Evidence: PMID:9030544 The deduced amino acid sequence predicts a type II transmembrane topology and contains two potential N-glycosylation sites. |
Loading supporting content…
Download this section (compressed HTML)Q: Does GAL3ST1 sulfate histone H3 Tyr99, as PMID:41686426 (Cancer Res 2026) reports? The underlying modification, H3Y99sulf, was defined in PMID:36805701 and formally challenged in PMID:40890505, which reannotated the original raw mass spectra as a phosphotyrosine peptide and detected no histone sulfation. PMID:41686426 cites the 2023 paper as settled background and does not cite the refutation, and its H3Y99sulf readouts are immunoblot and CUT&RUN with the antibody class the refutation says needs further validation. None of the papers has been retracted.
Q: How would a Golgi type II membrane protein whose catalytic domain faces the lumen sulfate a nascent histone in the cytosol? PMID:41686426 shows the in vitro reaction with purified soluble His-GAL3ST1, which bypasses the topology entirely. Is there a cytosolic pool of GAL3ST1, an alternative translation start or splice form lacking the transmembrane anchor, or is the in vitro activity an artefact of removing the protein from its membrane?
Q: Two different enzymes are now claimed to write the same mark: SULT1B1 (PMID:36805701) and GAL3ST1 (PMID:41686426). These attributions are mutually inconsistent. If the mark exists, a side-by-side assay of both recombinant enzymes on the same histone substrate, read out by sulfate-preserving mass spectrometry, is needed.
Q: GO can express the disputed claim - GO:0008476 protein-tyrosine sulfotransferase activity and GO:0006478 peptidyl-tyrosine sulfation both exist - but in human the activity term is annotated only to TPST1 and TPST2. Should the GO Consortium record contested molecular functions of this kind explicitly, rather than leaving the absence of an annotation to be read as either 'refuted' or 'not yet curated'?
Q: Beyond the histone question, does the strong GAL3ST1 upregulation seen in renal cell carcinoma and gastric cancer act through sulfatide itself - for example through sulfatide-dependent selectin binding or membrane organization - rather than through any protein-sulfating activity?
Experiment: Incubate purified recombinant GAL3ST1 (both the full-length protein reconstituted in a membrane mimetic and the soluble lumenal domain) with recombinant histone H3.2 and PAPS, and analyse by tandem mass spectrometry using electron-transfer rather than collisional dissociation, so that any sulfotyrosine survives fragmentation. Run synthetic Y99-sulfo and Y99-phospho peptides as retention-time and fragmentation standards on the same gradient, include blanks between runs to exclude carryover, and include galactosylceramide as the positive-control acceptor to confirm the enzyme preparation is active.
Hypothesis: GAL3ST1 does not sulfate histone H3 Tyr99.
Type: in vitro enzyme assay with electron-transfer tandem mass spectrometry
Experiment: Test topology directly with a protease-protection assay on isolated Golgi membranes and with N-glycosylation-site mapping, and test for a cytosolic pool by digitonin fractionation and by split-fluorophore complementation using a cytosolically expressed tag. If the catalytic domain is exclusively lumenal and no cytosolic pool exists, the proposed histone reaction cannot occur in cells as described.
Hypothesis: GAL3ST1's catalytic domain is not accessible to cytosolic substrates.
Type: membrane topology and subcellular fractionation
Experiment: Compare GAL3ST1 knockout gastric cancer cells rescued with wild-type GAL3ST1 versus a catalytically dead PAPS-binding mutant, and separately test whether exogenous sulfatide supplementation or CGT (UGT8) co-manipulation reproduces or rescues the migration and EMT phenotypes. Quantify sulfatide by lipidomics in every arm. This separates a lipid-mediated mechanism from the proposed chromatin mechanism without depending on the contested antibody.
Hypothesis: The gastric cancer phenotype attributed to GAL3ST1-dependent histone sulfation is instead driven by sulfatide.
Type: structure-function rescue with lipidomics
Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)