31. July 2014, 16:00
Ernst-Abbe-Platz 2, seminar room 3423
Solving a metabolomics Sudoku: high-throughput structural annotation by searching candidate substrate-product pairs (CSPP)
Kris Morreel, Yvan Saeys, Oana Dima, Ruben Vanholme, Bartel Vanholme, Wout Boerjan
(VIB Department of Plant Systems Biology, Ghent University)
The main bottleneck in metabolomics is the inability to structurally annotate the large number of unknown metabolites, greatly restricting its value in systems biology. Here, we show that combining mass spectral matching with biochemical knowledge and correlation analysis allows a high-throughput structural annotation in liquid chromatography-mass spectrometry (LC-MS). An algorithm is presented that picks peak pairs having mass and retention time differences corresponding with those expected for the substrates and products of well-known enzymatic reactions. Whenever a peak pair is annotated as a candidate substrate product pair (CSPP), the MS2 product ions and neutral losses of both peaks are matched and their abundances correlated across biological replicates. This strategy was evaluated on Arabidopsis thaliana Col0 leaves via negative ionization LC-MS. The resulting CSPP network showed regions associated with glucosinolate, flavonoid, benzenoid, phenylpropanoid, oligolignol/(neo)lignan, indolic and apocarotenoid metabolism. By knowing the structures of some members of these regions, the adjacent members were structurally characterized leading to the elucidation of 145 from a total of 229 predicted compounds. Based on the Scifinder database, 61 compounds had never been described in plants before and only 54 were previously identified in Arabidopsis. The CSPP-based annotation was confirmed by independent MSn experiments. Besides being high-throughput, this method allows the annotation of minor abundant compounds that are otherwise not amenable for purification. This method will greatly advance our insight into the various, nowadays largely unknown, pathways in plant (secondary) metabolism as well as in the complex interacting pathways between the human metabolome and microbiome; pathways that are further complicated by the nutritional content variety and the nutritional differences between human populations.