New molecular evidence of wine yeast-bacteria interaction unraveled by non-targeted exometabolomic profiling

• Published in: Metabolomics 2016-04, Vol.12 (4), p.1-16, Article 69
• Main Authors: Liu, Youzhong, Forcisi, Sara, Harir, Mourad, Deleris-Bou, Magali, Krieger-Weber, Sibylle, Lucio, Marianna, Longin, Cédric, Degueurce, Claudine, Gougeon, Régis D., Schmitt-Kopplin, Philippe, Alexandre, Hervé
• Format: Article
• Language: English
• Subjects: Biochemistry; Biomedical and Life Sciences; Biomedicine; Cell Biology; Developmental Biology; Food engineering; Life Sciences; Molecular Medicine; Original Article; Vitaceae
• ISSN: 1573-3882; 1573-3890
• DOI: 10.1007/s11306-016-1001-1

Introduction Bacterial malolactic fermentation (MLF) has a considerable impact on wine quality. The yeast strain used for primary fermentation can systematically stimulate ( MLF + phenotype) or inhibit ( MLF −) bacteria and the MLF process as a function of numerous winemaking practices, but the underlying molecular evidence still remains a mystery. Objectives The goal of the study was to elucidate such evidence by the direct comparison of extracellular metabolic profiles of MLF + and MLF − phenotypes. Methods We have applied a non-targeted metabolomic approach combining ultrahigh-resolution FT-ICR-MS analysis, powerful statistical tools and a comprehensive wine metabolite database. Results We discovered around 2500 unknown masses and 800 putative biomarkers involved in phenotypic distinction. For the putative biomarkers, we also developed a biomarker identification workflow and elucidated the exact structure (by UPLC-Q-ToF–MS 2 ) and/or exact physiological impact (by in vitro tests) of several novel biomarkers, such as D-gluconic acid, citric acid, trehalose and tripeptide Pro-Phe-Val. In addition to valid biomarkers, molecular evidence was reflected by unprecedented chemical diversity (around 3000 discriminant masses) that characterized both the yeast phenotypes. While distinct chemical families such as phenolic compounds, carbohydrates, amino acids and peptides characterize the extracellular metabolic profiles of the MLF + phenotype, the MLF − phenotype is associated with sulphur-containing peptides. Conclusion The non-targeted approach used in this study played an important role in finding new and unexpected molecular evidence.