This paper is often read as evidence that Gsta2 detoxifies bromobenzene metabolites. It shows the
opposite direction of causality: GSTA1/GSTA2 are targets of arylation, not catalysts acting on
bromobenzene metabolites. The GST here was simply an abundant hepatic protein recovered on a
GSH-agarose column.
- PMID:17112229
- PMID:17112229
Neither UniProtKB:P04903 nor this paper establishes GSTA2-catalysed conjugation of bromobenzene-derived
epoxides or quinones, so description and core_functions[0] state the adduction finding rather than a
substrate claim.
The five propagation_review blocks argue that RGD's mouse-to-rat ISO transfers here are cluster-level
rather than 1:1. That claim is verifiable from files in this repo, without relying on MGI subunit
nomenclature (Ya/Yc naming for the mouse donors is not sourced in-repo and is deliberately not used as
evidence):
- genes/rat/Gsta2/Gsta2-goa.tsv — rat Gsta2 receives ISO from two mouse donors, MGI:MGI:1095417
and MGI:MGI:95863. Two mouse genes seeding one rat gene is already a many-to-one mapping, not 1:1.
- genes/rat/Gsta1/Gsta1-goa.tsv — MGI:MGI:95863 seeds ISO onto rat Gsta1 as well, for the identical
five terms it seeds onto rat Gsta2: GO:0005739, GO:0005829, GO:0004364, GO:0009617,
GO:0035634. One mouse gene, two rat genes, the same five terms.
Correction (2026-08-07): an earlier revision of these notes claimed rat Gsta1 "receives ISO from nearly the
same mouse donor set (MGI:95856, 95861, 95862, 95863, 95864, 95865, 1095417)". That was wrong
— those seven identifiers are the WITH/FROM set of rat Gsta1's IBA rows under GO_REF:0000033, not
ISO donors. Rat Gsta1's only mouse ISO donor is MGI:MGI:95863, and MGI:MGI:1095417 appears in no rat
Gsta1 ISO row at all. Verify with:
awk -F'\t' '$9=="ISO"{print $5"\t"$11}' genes/rat/Gsta1/Gsta1-goa.tsv.
Donor identities (checked against the Alliance of Genome Resources API, https://www.alliancegenome.org/api/gene/MGI:1095417
and .../MGI:95863, retrieved 2026-08-07): MGI:1095417 = mouse Gsta1, "glutathione S-transferase, alpha
1 (Ya)"; MGI:95863 = mouse Gsta2, "glutathione S-transferase, alpha 2 (Yc2)". The Ya/Yc2 subunit naming
in those full names is recorded here only as retrieved provenance; the propagation reviews deliberately do
not rest on it, and argue from the GOA donor overlap above instead.
Cluster-level transfer matters asymmetrically by term type, which is why the file's actions differ on the
same donors: the class-defining catalytic term GO:0004364 transfers safely (every alpha-class member has
that activity), whereas compartment (GO:0005739) and stimulus-response (GO:0009617, GO:0035634,
GO:0061771) terms are member-specific and do not. Note that the target's own GO:0004364 IDA row
(PMID:17112229, CAFA) is not independent confirmation of catalysis — that paper assays adduction of the
protein and states the effect on activity "is not known".
GOA changes. Eight new seeded rows. Five are donor splits from mouse Gsta2 (MGI:MGI:95863) of terms previously reviewed with mouse Gsta1 (MGI:MGI:1095417) as donor: glutathione transferase activity, mitochondrion, cytosol, response to bacterium, response to stilbenoid (all ISO, GO_REF:0000121). One is a further cytosol ISO split from human GSTA2 (UniProtKB:P09210). Two are new terms: phospholipid-hydroperoxide glutathione peroxidase activity (GO:0047066, ISS from human GSTA2) and cellular oxidant detoxification (GO:0098869, IEA GO_REF:0000108, inferred from GO:0047066). No retired rows. Total 27 rows.
Actions.
- Mouse Gsta2-donor rows: PENDING -> same action as their mouse Gsta1-donor siblings (GST activity ACCEPT; cytosol KEEP_AS_NON_CORE; mitochondrion, response to bacterium, response to stilbenoid MARK_AS_OVER_ANNOTATED). The MGI:MGI:95863 propagation_review.source_entities entries that had been written on the sibling rows were moved to the new rows, so each row now carries only its own donor.
- Cytosol ISO from human GSTA2: PENDING -> KEEP_AS_NON_CORE (location confirmed for rat GSTA2 PMID:17112229).
- GO:0047066 ISS from human GSTA2: PENDING -> KEEP_AS_NON_CORE. Donor activity is measured PMID:10395737; not measured on rat GSTA2, and secondary to conjugation. PMID:10395737 was fetched and added to references.
- GO:0098869 IEA: PENDING -> KEEP_AS_NON_CORE, following GO:0047066 PMID:10395737.
- No action changed on existing rows. Donors are now named on the cytosol, GST ISO and GST ISS (mouse Gsta2, P10648) siblings; IDA rows from PMID:17112229 now quote the Cys-111/cytosolic A1/A2 sentence rather than a background sentence; UniProt FUNCTION / SUBCELLULAR LOCATION lines added as positive support.
Open questions. Has selenium-independent peroxidase activity toward phospholipid hydroperoxides been measured for purified rat GST A2-2 (Ya2 homodimer)? Which mouse alpha-class gene is the true one-to-one orthologue of rat Gsta2 (both mouse Gsta1 and Gsta2 donate ISO rows)?
Review follow-up 2026-10-17 (PR #4616 review). Finding 1: the GO:0004364 and GO:0042178 ACCEPT rows citing the PMID:17112229 Cys-111 arylation quote now carry positive UniProt support (CATALYTIC ACTIVITY RX+glutathione, FUNCTION conjugation line) with the reason noting retention by deferral to the RGD curator. Finding 2: all fifteen DR GO; cross-reference quotes on ACCEPT/KEEP rows replaced with the relevant UniProt CC line, and the circular DR GO; entries removed from the caloric-restriction/mitochondrion/response-to-bacterium/epithelial/stilbenoid MARK_AS_OVER_ANNOTATED rows. Validates clean; 0 stale.