Mining terroir influence on bioactive polyphenols from grape stems: A correlation-network-driven approach to spatialize metabolomics data

Abstract

In viticulture, the concept of terroir is often used to enlighten the environmental-based typicity of grapevines grown in a local area however its scientific basis remains under debate. Grape polyphenols as key player of the plant defense system enables adaptation to environmental changes and so far, form a unique metabolic component to investigate the terroir influence. Using correlation-based networks and spatial metabolomics we investigated how continuous changes of soil properties may impact the polyphenol composition within a single grape clone. In a vineyard parcel covering four different geological layers and planted with a unique clone of Cabernet Franc, the soil texture was analyzed at 30 georeferenced points with a spatial coverage sampling strategy. Grape stems were harvested at corresponding positions over 3 consecutive years followed by UPLC-DAD-MS-based metabolomic analysis targeted on 43 metabolites including flavonoids, phenolic acids, procyanidins and stilbenoids. Principal component analyses on intra-vintage data presented good reproducibility. A correlation-driven approach was used to select co-varying metabolites before using Geographic Information System (GIS). As results, flavonoids and stilbenoid DP4 were spatialized according to soil granulometry, with stilbenoid DP4 over-accumulating in loamy-silty soils and flavonoids in sandy soils. The present study highlights soil-based terroir influence on polyphenols in a continuous space. Spatial metabolomics driven by correlation-based networks represents a powerful approach to spatialize field-omics data and may serve as new field-phenotyping tool in precision agriculture.