LC-MS-Based Metabolomics for the Chemosystematics of Kenyan Jacq (Sapindaceae) Populations.

Jacq (Sapindaceae) is a medicinal plant with a worldwide distribution. The species has undergone enormous taxonomic changes which caused confusion amongst plant users. In Kenya, for example, two varieties are known to exist based on morphology, i.e., var. along the coast, and var. in the Kenyan inland. These two taxa are recognized as distinct species in some reports. This prompted us to apply metabolomics to understand the relationship among naturally occurring populations of in Kenya, and to identify compounds that can assist in taxonomic delineation of the different varieties of from different parts of Kenya. The phytochemical variability of Kenyan var. populations collected from four different geographical regions (Nanyuki, Machakos, Nairobi, and Narok) and one coastal var. (the Gazi) were analyzed by LC-MS using a metabolomics-driven approach. Four known compounds, two diterpenoids (dodonic acid (), hautriwaic acid lactone (), and two flavonoids (5,7,4',5'-tetrahydroxy-3,6,2'-trimethoxyflavone () and catechin ()) were isolated and purified from the Gazi coastal collection. The presence of these compounds and their relative abundance in other populations was determined by LC-MS analyses. Multivariate statistical analyses of LC-MS data was used for the visualization of the patterns of variation and identification of additional compounds. Eleven discriminant compounds responsible for separating chemometric clusters were tentatively identified. In an antimicrobial assay, hautriwaic acid lactone () and catechin () were the most active compounds followed by the extract from the coastal (Gazi) population. The clustering pattern of the five populations of suggested that the metabolite profiles were influenced by geo-environmental conditions and did not support the current classification of based on morphology. This study disputes the current classification of in Kenya and recommends revision using tools such as molecular phylogenetics.