Spectral trees as a robust annotation tool in LC-MS based metabolomics

The identification of large series of metabolites detectable by mass spectrometry (MS) in crude extracts is a challenging task. In order to test and apply the so-called multistage mass spectrometry (MS (n) ) spectral tree approach as tool in metabolite identification in complex sample extracts, we firstly performed liquid chromatography (LC) with online electrospray ionization (ESI)-MS (n) , using crude extracts from both tomato fruit and Arabidopsis leaf. Secondly, the extracts were automatically fractionated by a NanoMate LC-fraction collector/injection robot (Advion) and selected LC-fractions were subsequently analyzed using nanospray-direct infusion to generate offline in-depth MS (n) spectral trees at high mass resolution. Characterization and subsequent annotation of metabolites was achieved by detailed analysis of the MS (n) spectral trees, thereby focusing on two major plant secondary metabolite classes: phenolics and glucosinolates. Following this approach, we were able to discriminate all selected flavonoid glycosides, based on their unique MS (n) fragmentation patterns in either negative or positive ionization mode. As a proof of principle, we report here 127 annotated metabolites in the tomato and Arabidopsis extracts, including 21 novel metabolites. Our results indicate that online LC-MS (n) fragmentation in combination with databases of in-depth spectral trees generated offline can provide a fast and reliable characterization and annotation of metabolites present in complex crude extracts such as those from plants.