UNTARGETED METABOLOMICS REVEALS MAIZE METABOLITES INVOLVED IN RESISTANCE AGAINST FUSARIUM INFECTION

Fusarium verticillioides (FV) is the causal agent of Fusarium ear rot (FER) in maize and\r\ncontaminates the grain with fumonisins, carcinogen mycotoxins. A metabolomic approach was\r\napplied in order to reveal the mechanism of host resistance towards this pathogen. Plants of resistant\r\n(R) and susceptible (S) genotypes were grown in glasshouse in a randomized block design. The\r\nmetabolic profiles of untreated, mock-and FV-inoculated maize kernels, sampled at 72 hours after\r\ninoculation (hpi), were investigated with liquid chromatography in combination with high\r\nresolution mass spectrometry. Five biological replicates were taken for each condition. The use of\r\nstable isotopic labelling (SIL) combined to MetExtract algorithm allowed the automatic detection\r\nand prediction of carbon atoms of only compounds of biological origin. Features pairs were\r\ngrouped according to retention time, m/z and combination of both ionization polarities. The most\r\nintense ion species was selected for each feature group and 534 feature groups, corresponding to\r\nlikewise metabolites, were issued. Several databases were screened for the metabolite search of the\r\nhighest m/z within each feature group restricted by the count of the carbon atoms. At least one\r\ndatabase entry was assigned for 119 metabolites. Moreover, 11 metabolites could be identified\r\nthrough the screening of 129 standard compounds, including key amino acids for the secondary\r\nmetabolism, a flavonoid, N-hydroxynnamic acid amides, intermediates in synthesis of lignin and\r\nlignans, plant hormones. Multivariate statistics was used on the whole data set of 534 imputed\r\nrange-scaled metabolites. In principal component analyses FV- samples were grouped separately\r\nfrom other samples, with PC1 explaining 42.7% of the total metabolic variation, PC2 (discriminates\r\ngenotypes) 28.9% and PC3 7.5%. Also in Heat Map, FV samples appeared clearly separated from\r\nthe others, as well as mock-and control samples were grouped together according to the genotype.\r\nDifferent clusters of metabolite could be recognized depending on the abundance levels in the\r\nsamples. Several Student t test were applied to the data matrix of 534 metabolites, focusing on\r\ncomparison between the untreated samples (metabolites related to “constitutive resistance”) and\r\ncomparisons between the mock and FV-inoculated samples (“FV responsive” metabolites). A total\r\nof 222 metabolites resulted significantly affected by FV treatment. Annotation revealed that a part\r\nwas involved in the synthesis of oxygenated fatty acids (oxylipins), in the phenylpropanoid\r\npathway, flavonoids, benzoxizanoid biosynthesis, hydroxycinnamic acid amides and the\r\nmetabolism of the aromatic amino acids. 90 metabolites resulted significantly different between the\r\nuntreated samples of the two genotypes, belonging to the pathways of phenylpropanoids,\r\nbenzoxizanoids, oxylipins, suggesting that there is a different abundance of these metabolites\r\nalready in the untreated samples and their altered level is kept and enhanced at 72 hpi.