Semi-Targeted Metabolomics to Validate Biomarkers of Grape Downy Mildew Infection Under Field Conditions

Abstract

Grape downy mildew is a devastating disease worldwide and new molecular phenotyping tools are required to detect metabolic changes associated to plant disease symptoms. In this purpose, we used UPLC-DAD-MS-based semi-targeted metabolomics to screen downy mildew symptomatic leaves that expressed oil spots (6 dpi, days post-infection) and necrotic lesions (15 dpi) under natural infections in the field. Leaf extract analyses enabled the identification of 47 metabolites belonging to the primary metabolism including 6 amino acids and 1 organic acid, as well as an important diversity of specialized metabolites including 9 flavonols, 11 flavan-3-ols, 3 phenolic acids, and stilbenoids with various degree of polymerization (DP) including 4 stilbenoids DP1, 8 stilbenoids DP2, and 4 stilbenoids DP3. Principal component analysis (PCA) was applied as unsupervised multivariate statistical analysis method to reveal metabolic variables that were affected by the infection status. Univariate and multivariate statistics revealed 33 and 27 metabolites as relevant infection biomarkers at 6 and 15 dpi, respectively. Correlation-based networks highlighted a general decrease of flavonoid-related metabolites, whereas stilbenoid DP1 and DP2 concentrations increased upon downy mildew infection. Stilbenoids DP3 were identified only in necrotic lesions representing late biomarkers of downy mildew infection.