GLS2KO vs WT mouse hepatocytes
Ontology highlight
ABSTRACT: Test the effect of GLS2 knockout in primary mouse hepatocytes. We isolated hepatocytes from GLS2 knockout and wild-type mice, and briefly applied media lacking L-glutamine (2 hours). Sixty minutes after resupplying Gln, metabolites were extracted and analyzed with the Mixed Mode method. Data were processed through XCMS, and features were filtered for p<0.01, fold change >2, and a minimum intensity of 1x10^6. Sorting by smallest p value, the first extracted ion chromatogram (EIC) with good chromatographic peak shape corresponded to 188.0567m/z at 24.41 min. Six putative IDs were within 3ppm of the experimentally observed m/z, representing two chemical formulas, none of which had documented retention times in training or test sets. Amongst these potential IDs, the Message Passing Neural Network (MPNN) model correctly predicted N-acetyl-L-glutamic acid as the most likely candidate, as verified by injection of purchased standards. The next most significant difference between GLS2KO vs WT was 117.0196m/z observed at 20.07 minutes. The model had been trained on 2/6 of the putative IDs. Despite the four additional isomers suggested, the model correctly selected succinate as reduced by GLS2 KO (Figure 5B). The third most significant hit corresponds to 171.0068m/z at 23.11 min. Although glycerol 1-P and 2-P are both potential hits, almost indistinguishable by the model, the large gap in retention times between these top hits and the Cl- adducts of threonate (+ isomers) is apparent, further supporting the correct identification as glycerol mono-phosphate. Expansion of the list to include p<0.05 leads to the identification of Glutamine, Glutamate, and other downstream metabolites known to be altered by GLS2 KO.
ORGANISM(S): Mouse Mus Musculus
TISSUE(S): Liver
SUBMITTER: Michelle Clasquin
PROVIDER: ST001952 | MetabolomicsWorkbench | Fri Oct 22 00:00:00 BST 2021
REPOSITORIES: MetabolomicsWorkbench
ACCESS DATA