Development of a kinetic model of the phenylpropanoid pathway in potato tuber. Step 1: The construction of a comprehensive knowledge base

VALIÑAS M; LANTERI ML; TEN HAVE A; VILLARREAL F; ANDREU AB

Mar del Plata

Congreso; IX Congreso Argentino de Bioinformática y Biología Computacional; 2018

Background: Potato (Solanum tuberosum L.) constitutes a staple food at national and worldwide levels. Even though it has remarkable nutritional properties, commercial potato varieties present low contents of flavonoids and anthocyanins, antioxidants that show impact in the prevention, treatment, and management of chronic disease. However, potato andean varieties present higher levels of these compounds. The metabolic pathways leading to the biosynthesis of phenylpropanoids, including flavonoids and anthocyanins, form a complex network that includes (i) reactions with substrates leading to (or converging from) multiple sub-pathways; (ii) promiscuous enzymes; and (iii) multiple isozymes that derive from gene superfamilies or alternative splicing. This hinders with breeding strategies to produce commercial varieties with higher flavonoid and anthocyanin contents. Objectives: A kinetic model for this pathway can be used to make quantitative predictions regarding the major and minor metabolic flows. Such predictions could identify gene targets to alter the contents of specific compounds, while having a minimal impact on other metabolites and pathways. Here, we present the first step required to construct such a model: the development of a knowledge database specific for potato tuber, containing comprehensive information of all known reactions, metabolites and genes involved in the biosynthesis of phenylpropanoids. Results: Using public databases (KEGG, MetaCyc, ExplorEnz), we developed a first version of the metabolic pathway, written in SBGN using PathVisio. This was expanded based on metabolites reported to be detected in tubers, according to data generated in our lab or from the literature (e.g., coumarins, modified anthocyanidins, isorhamnetin). Then, all the reactions were completed according to reactants, stoichiometry, and enzymes involved. Using databases (BRENDA, Uniprot, GenBank) and literature, we identified (whenever possible) genes with experimentally characterized activities. These genes were used to find potential homologous in potato. Finally, these candidates were filtered for expression in tubers of different varieties, using publicly available RNA-seq data. The final model was written using SBML in CellDesigner. The knowledge base includes 202 reactions, 215 metabolites and 91 enzymes, with more than 300 tuber-specific candidate genes. Conclusion: The knowledge-base is the pillar for the construction of a predictive kinetic model of phenylpropanoid metabolism in potato tuber.