GALT (human) — gene review notes

UniProt: P07902. Gene: GALT (galactose-1-phosphate uridylyltransferase). Chr 9p13.

Core biology (verified)

GALT catalyses the third, central step of the Leloir pathway of galactose
catabolism:

alpha-D-galactose 1-phosphate + UDP-alpha-D-glucose <=> alpha-D-glucose 1-phosphate + UDP-alpha-D-galactose
(Rhea:RHEA:13989, EC 2.7.7.12)

• MF = GO:0008108 UDP-glucose:hexose-1-phosphate uridylyltransferase activity.
  UniProt CATALYTIC ACTIVITY block cites ECO:0000269|PubMed:22461411 and PubMed:27005423.
• Double-displacement ("ping-pong") mechanism through a covalent uridylyl-enzyme
  intermediate: UMP is transferred onto active-site His186 (HPH motif, His-Pro-His,
  residues 184–186), forming a phospho-His–UMP intermediate with release of glucose-1-P;
  the intermediate then transfers UMP to gal-1-P to form UDP-Gal.
  PMID:27005423
  PMID:22461411
• Homodimer, obligate; each active site formed by residues from BOTH subunits.
  PMID:27005423
  UniProt SUBUNIT: "Homodimer. {ECO:0000269|PubMed:27005423}."
• Zinc binding = structural, not catalytic. hGALT crystal structure (PDB 5IN3) shows
  ONE divalent metal site per monomer (Glu202, His301, His319, His321), ~20 Å from the
  active site, preferring Zn2+; Zn2+ stabilises the protein and prevents aggregation.
  UniProt COFACTOR annotates "Binds 2 zinc ions per subunit" (ECO:0000305|PubMed:27005423).
  PMID:27005423
  Note: the E. coli enzyme has separate Zn (catalytic-stabilising) and Fe sites; those
  residues are NOT all conserved in human GALT — so zinc here is structural.
• Localization: cytosol (GO:0005829). Cytosolic metabolic enzyme (Reactome TAS;
  IBA cytoplasm).

Disease (dismech Galactosemia.yaml + literature)

GALT deficiency causes classic galactosemia (type I; OMIM 230400) — autosomal
recessive, ~1/50,000 newborns (US screening). Most severe galactosemia form:
- Neonatal toxicity on milk (lactose->galactose): jaundice, hepatomegaly/liver failure,
cataracts, renal failure, bleeding diathesis, E. coli sepsis, death within days if
untreated.
PMID:22461411
- Driven by accumulation of galactose-1-phosphate (and galactitol); reduced UDP-hexoses
and disturbed glycosylation.
- Treatment = dietary galactose restriction; but long-term complications persist despite
diet: cognitive/IQ deficits, speech dyspraxia, ataxia, premature ovarian insufficiency.
PMID:22461411
- >300 disease mutations; ~60% missense. Most common p.Gln188Arg (Q188R) — active-site
variant, ~10% residual activity, aggregation-prone. p.Ser135Leu common in Africans;
p.Lys285Asn common in Europeans. [PMID:1897530; PMID:27005423]

Annotation-by-annotation reasoning

• GO:0008108 (MF, uridylyltransferase): CORE. Supported by IBA, IEA(EC/RHEA), IDA
  (PMID:27005423 covalent intermediate in structure), EXP (PMID:1897530, PMID:22461411
  kinetics), TAS(Reactome). ACCEPT all instances.
• GO:0033499 (BP, Leloir catabolism): CORE. IBA + TAS(Reactome). ACCEPT.
  (Note: current ontology primary label is "galactose catabolic process via UDP-galactose,
  Leloir pathway"; GOA/stub carries older "beta-D-galactose ..." label — keep as-is,
  existing-annotation ids are trusted.)
• GO:0006012 (BP, galactose metabolic process): parent of the Leloir catabolic term.
  IEA + IDA(PMID:27005423) + TAS(PMID:1427861). Correct but less specific than GO:0033499.
  ACCEPT the direct/experimental ones; broad but not wrong.
• GO:0006011 (BP, UDP-alpha-D-glucose metabolic process): IDA PMID:27005423. GALT consumes
  UDP-glucose as the uridylyl donor -> ACCEPT (accurate; the substrate is UDP-Glc).
• GO:0008270 (MF, zinc ion binding): IEA + IDA(PMID:27005423). Real (2 Zn2+/subunit,
  structural). ACCEPT / keep as non-core structural function.
• GO:0005829 (cytosol) TAS x2, GO:0005737 (cytoplasm) IBA: ACCEPT (correct localization).
• GO:0005794 (Golgi apparatus) IDA PMID:20605918: SUSPECT. PMID:20605918 is about Lyn
  kinase / ACSL3 Golgi export — GALT is not in the abstract; this is a spurious/
  mis-propagated CC. GALT is a soluble cytosolic Leloir-pathway enzyme with no established
  Golgi role. -> REMOVE (contradicts established cytosolic localization; no independent
  support). This is an IEA/IDA CC error, appropriate to remove per guidelines.
• GO:0005515 (protein binding) IPI x (many): all from large-scale interactome maps
  (PMID:16189514, 25416956, 25910212, 26871637, 28514442, 32296183, 33961781). Bare,
  uninformative; no specific functional partner established. Per guidelines avoid
  'protein binding'. -> MARK_AS_OVER_ANNOTATED / keep non-core; do NOT remove (IntAct
  detections are real) but not a core function.

Deep research

falcon deep-research file did NOT land within the 8-minute poll window; review grounded
in UniProt, GOA, dismech Galactosemia.yaml, and cached PMIDs (structure PMID:27005423,
kinetics PMID:22461411, mutation PMID:1897530, gene PMID:1427861).

2026-11-12: completion audit (supersedes the provisional calls above)

All rows re-audited against cached publications and project rules. Final decisions that
differ from the earlier provisional notes:

• GO:0005515 protein binding IPI (7 review rows covering 10 GOA rows; PMID:16189514,
  25416956, 25910212, 26871637, 28514442, 32296183, 33961781): REMOVE under the
  protein-binding policy (MARK_AS_OVER_ANNOTATED is not permitted for GO:0005515). None of
  the screens supports a more informative MF; removal does not deny the interactions.
• GO:0005794 Golgi apparatus IDA PMID:20605918: UNDECIDED (not REMOVE / not
  over-annotated). The cache is abstract-only and the publisher full text returned 403;
  the abstract concerns Lyn/ACSL3 Golgi export and does not mention GALT. A "GalT"
  Golgi marker (beta-1,4-galactosyltransferase) mapped to GALT is a plausible explanation
  but cannot be confirmed, so no mis-attribution or source defect is asserted.
• GO:0005737 cytoplasm IBA: MODIFY to GO:0005829 cytosol (correct but general; cytosol is
  already on the gene from Reactome), with propagation_review on PTN000235304.
• GO:0006012 galactose metabolic process (IEA, IDA PMID:27005423, TAS PMID:1427861):
  MODIFY to GO:0033499 (correct but general; specific descendant already on the gene).
  GO:0006012 dropped from core_functions.
• GO:0006011 UDP-alpha-D-glucose metabolic process IDA: KEEP_AS_NON_CORE (true; a
  consequence of the core Leloir step, not separate core role).
• GO:0033499 label updated to the current ontology label in core_functions.

Deep research: the falcon file has since landed. It is used only as a retrieval aid. Note
that it names His166 as the catalytic nucleophile, which is the E. coli numbering; the
human structure paper identifies His186 ["confirming that His186 is the site of
uridylylation" PMID:27005423], which the review uses.

Status set to COMPLETE.

Update 2026-11-12: Golgi IDA resolved

• GO:0005794 Golgi apparatus IDA PMID:20605918: now REMOVE, with a WRONG_GENE_PRODUCT
  source defect (GALT P07902 → B4GALT1 P15291, cause SYMBOL_COLLISION, resource UNIPROT).
  The publisher full text (https://journals.biologists.com/jcs/article/123/15/2649/31141/The-Lyn-kinase-C-lobe-mediates-Golgi-export-of-Lyn)
  defines the paper's GalT as the Golgi marker: "ACSL3 was seen with the Golgi caveolin and
  β-1,4-galactosyltransferase (GalT) where Lyn accumulated", and uses it with GM130 ("knockdown
  of ACSL3 did not affect Golgi localization of GalT and GM130"). Galactose-1-phosphate
  uridylyltransferase is not studied. The quotations are from the full text, which is not cached,
  so they are recorded here and in the row's reason rather than as supporting_text.