Untargeted metabolomics approach enabling the separate investigation of the attack of FUSARIUM GRAMINEARUM during Fusarium Head Blight on wheat - plant microbe interaction.

Metabolomics studies of plant-microbe interactions are largely hampered by the intimate physical association and potential commonality of metabolites of the respective microorganism and its host. Here, we present a novel liquid chromatography-high resolution mass spectrometry (LC-HRMS) approach for the untargeted screening of microbe-derived metabolites in colonized plant tissue. To this end, a reference list of metabolic features is generated with the software MetExtractII [1] from mixtures of globally 13C labelled and native culture samples of the plant pathogenic fungus Fusarium graminearum (Fg) as well as mixtures of 13C labelled and unlabelled wheat plant samples. Next, LC-HRMS data of the biological plant-microbe experiment under investigation are matched against the reference features with the tool MetMatch [2] and filtered with the aim to assign all microbe-derived metabolites. To exemplify the developed workflow, flowering wheat ears were treated with spores of Fg in a time series experiment. Among the 1485 detected low molecular weight biochemical constituents, 100 of these were suggested as Fg-derived metabolites in plant- fungus interaction based on the applied matching- and filtering criteria.
With the aim of confirming the annotation of these assumed metabolites, a more detailed classification process was established. First, the number of potential metabolites was further narrowed down by literature research, keeping only unique metabolite entries, which were then verified with the help of LC-HRMS/MS data of the biological plant-microbe experiment. As indicators of quality, the mass accuracy of the precursor ions (m/z values) from the LC-HRMS measurements, the respective peak shapes, as well as retention time shifts and signal intensities within certain thresholds were taken in consideration. Finally, fragment ions of metabolite entries were compared with standards and reference data from databases, such as MassBank [3] among others. Through this process it was possible to confirm the detection of known Fg metabolites like trichothecenes, culmorins and others in the plant-fungus interaction.