Feature MS fragments-based method for identification of toxic furanoids in biological samples

Numerous furan-containing compounds have been reported to be toxic. The toxicity may be attributed to the metabolic activation of the furan ring to cis-enediones. Identification of unknown furans that undergo bioactivation is challenging. Here, we present a novel approach that enables non-targeted profiling of bioactivation of unknown furanoids both in vitro and in vivo. Cyclic pyrrole-glutathione conjugate was the predominant product of cis-enediones with glutathione. The shared glutathione substructure of conjugates was capable of generating four constant and signature fragments under collision-induced dissociation (CID) in the mass spectrometer, including neutral loss fragments 103.0269 Da and 146.0691 Da and product ions at m/z 130.0499 and 177.0328. The unique structure and high abundance of conjugates in combination with the consistency and specificity of CID fragmentation brought extraordinarily high selectivity and reliability for the four fragments as a fingerprint of bioactivated furanoids. The bioactivated furanoids can be identified by screening the four fragments in high-resolution MS/MS datasets using the neutral loss filtering and diagnostic fragmentation filtering of data post-acquisition software MZmine. The simultaneous formation of four individual signal points in the filtering channel with the same precursor ion and retention time was assigned to be furanoids. The method has been rigorously validated. In the pooled urine samples from nine model furanoids-treated mice, nine cis-enediones from the parent furanoids and two from furanoid metabolites were accurately detected and identified. The method showed great performance in non-targeted profiling bioactivated furanoids and their metabolites in urine samples of herbal extract-treated mice.