Current review · original contexts and source evidence · dated journal
Human GALK1 catalyzes ATP-dependent phosphorylation of galactose at carbon 1, committing the sugar to the Leloir pathway. Human cDNA expression in COS cells and purified recombinant-human enzyme kinetics independently support this reaction (PMID:7542884(https://pubmed.ncbi.nlm.nih.gov/7542884/), PMID:12694189(https://febs.onlinelibrary.wiley.com/doi/full/10.1046/j.1432-1033.2003.03538.x)). The latter original Methods and Results were read, including the human IMAGE3501788 construct, affinity tag, product-inhibition analysis and mutant kinetics. NADH is the coupled assay reporter; it is not a GALK1 cofactor.
Selected original Results, Figure 3 caption and Methods of PMID:14596685(https://pmc.ncbi.nlm.nih.gov/articles/PMC280648/) show robust in-vitro 2-deoxygalactose phosphorylation, with kinetic parameters within fourfold of those measured for galactose. No phosphorylation was detected for glucose, fucose, arabinose or N-acetylgalactosamine under the reported assay conditions. This supports substrate discrimination without claiming absolute galactose exclusivity or established physiological use of 2-deoxygalactose. The paper's introductory structural image is from bacterial L. lactis; the separately read human complex provides the direct human structural evidence. Km is not an equilibrium binding constant.
The single core describes the cytosolic galactokinase reaction and Leloir process. Broad kinase and galactose-metabolism refinements use verified typed GO paths. Carbohydrate phosphorylation remains a valid original claim: no current is_a/part_of path makes the Leloir term its descendant. The inherited NEW galactose-catabolism term is retired because it is an ancestor of the Leloir term already present in two original contexts. The complete historical object and both quotations remain at the evidence supplement’s immutable Git locator; no original GOA context is removed.
Official UniProt JSON and direct FASTA establish one canonical P51570 sequence of 392 residues and no declared alternative products. The sequence equals the inherited record and the actual PAINT alignment. Historical HuRI identifier P51570-1 is retained literally as dataset nomenclature; it does not create a natural product declaration.
The human structure paper (PMID:15590630(https://pubmed.ncbi.nlm.nih.gov/15590630/)) was read from the author-institution PDF, including selected expression, assembly and active-site Results. Official 1WUU has 399 residues, exactly the current sequence with AHHHHHH inserted after Met1. The bound nucleotide is Mg-AMPPNP. Crystal dimers contain Cys391 disulfides, but the authors leave physiological assembly unresolved; the earlier human kinetic study reports a monomer by gel filtration, with data not shown. The current UniProt homodimer statement is therefore recorded as a probable overstatement in a structured OTHER/CHECK source defect, without asserting a corrected physiological monomer state. Deposited mutation counts, crystal packing and affinity tags are not natural isoforms or established physiological assembly states.
The current membrane IDA context from PMID:8908517(https://genome.cshlp.org/content/6/10/980.full.pdf) is explicitly NOT located in membrane. The indexed publisher-original Results/Figure 3 caption and expression Methods show human GALK1 translated with canine microsomes remaining protease-accessible and in the supernatant after membrane pelleting at neutral and carbonate pH. This supports the retained negative assertion and soluble cytosolic setting in the tested system. The source does not exclude all conditional peripheral associations. Direct PDF download and screenshot access failed; an error body was not counted as an inspected PDF, and figure pixels were not independently interpreted.
The YTS membrane-proteomics annotation (PMID:19946888(https://pubmed.ncbi.nlm.nih.gov/19946888/)) is a separate positive context. Its abstract describes treatments reported or attempted to remove peripheral proteins and explicitly includes integral and stably associated proteins. Complete removal is not established, and absence of a transmembrane segment does not identify GALK1’s measured fraction. The target table, peptide and full assay context remain unadjudicated. The positive generic membrane row is now REMOVE: a bulk-fraction detection of a soluble enzyme with no membrane anchor does not establish membrane residence, the survey's own abstract says most non-integral hits were transiently associated species, and the targeted GALK1 experiment found the protein cytosolic and not membrane-associated. Removal does not deny that some GALK1 co-purified with the YTS fraction.
The urinary-vesicle context has positive target evidence. PMID:19056867(https://pmc.ncbi.nlm.nih.gov/articles/PMC2637050/) Table 1 names GALK1, and the authentic expanded NHLBI table identifies GALK1/NP_000145 with one reported peptide and reference 2. Selected isolation and identification Methods were read. This supports limited occurrence in the isolated fraction as non-core; it does not establish encapsulation, membrane topology or a vesicle function. No target peptide sequence was recovered. The original part_of UBERON:0001088 urine atom is preserved literally and is retained as a flat BFO:0000050 extension on the original urinary context. Imported GO columns in the table are not independent experimental evidence.
The authentic HuRI dataset described in the original paper contains three assay records for one GALK1–Q9NPJ4 pair, using human ORFs in yeast. The exact clone identifiers, source suffixes and pooled/array/validated two-hybrid methods remain in the archive. The paper's Methods and data-availability context were read (PMID:32296183(https://pmc.ncbi.nlm.nih.gov/articles/PMC7169983/)). Removing uninformative generic protein binding does not deny this positive interaction or establish a different binding function.
The actual PTHR10457 tree places P51570 at AN489/PTN002470832. All three IBA assertions originate at AN0/PTN000048421; every donor identifier was resolved in the actual tree, and the only retained activity-loss node is off the human lineage. The canonical sequence equals the ungapped human MSA. Human self-donation is legitimate experimental grounding. Full book/MSA files remain in retained provenance rather than being duplicated here.
Official Reactome objects establish normal reaction R-HSA-70355 → catalyst 70354 → physical entity 70353 → P51570, residues 1–392, in cytosol. Pathway R-HSA-70370 includes that reaction. The original R-HSA-5610026 is a FailedReaction of a mutant set with no output and an explicit normalReaction link. That model is not a positive wild-type turnover experiment; normal activity is independently supported by human primary studies. No model or normal cache is rewritten.
The review contains short, exact primary excerpts in structured findings, with explicit links from the activity, binding, localization and refined-process rows. Shared evidence is quoted once. The complete negative-membrane sentence replaces the clipped fragment; its 22 words are not duplicated in the core. The urinary-vesicle occurrence is supported by the actual target table, not by the unrelated cytosol excerpt formerly associated with that row.
All 21 PMID identities and 29 reference assessments remain. The assessments state biological relevance and reading limits, with terse provenance-only notes for sequence/proteome surveys. Cache availability remains distinct from original sections actually read. Selected external originals were read for PMID:12694189, PMID:15590630, PMID:8908517 and PMID:19056867 even though their normal caches are abstract-only. Selected mapping/variant/activity passages were read for PMID:10521295; availability of the genomic/proteome papers does not imply inspection of their GALK1-specific tables.
The compact evidence supplement retains all 26 literal raw records as a column-keyed table, the actual row joins, NOT and urine extension, canonical sequence, PAINT ancestry, ontology paths, Reactome target chains, interaction and urinary-table evidence, and all 28 source hashes. These fields include provenance absent from the 16-column TSV. The full earlier 27-object review, journal, 23 historical quotation associations and detailed acquisition/access records remain linked at immutable commit 9fc29996a93144e2c5f117273e688ab35f08cf25. No old history or source cache is rewritten.
All 26 machine contexts remain unchanged (one action, the YTS membrane HDA row, changed from UNDECIDED to REMOVE in the later audit); no NEW annotation is added. Eight current supporting-text excerpts and the one-word UniProt source assertion stay within 25 words per source. A later audit resolved the three former validation advisories (stale GO:0033499 label, missing IBA propagation metadata, membrane action consistency). Strict GOA matching passed for all 26 original contexts, and the review status is now COMPLETE. Native EDIT history, private render/link checks and scoped whitespace checks passed. All 28 scientific source bytes and the derived term caches remain unchanged.