Highlighting glycosylation ways in Caryophyllaceae saponins by simplex simulation approach

Glycosylation mechanisms in saponins of Caryophyllaceae plant family were subjected to simulation by statistically exploring variability of 231 chemical structures belonging to four different aglycones: gypsogenin (Gyp), quillaic acid (QA), gypsogenic acid (GA), 16-OH-gypsogenic acid (16-OH-GA). Saponins based on different aglycones were initially characterized by relative glycosylation levels of different carbons. Simulation was initialized by combining the four saponin groups using Scheffé’s mixture design which provides a complete set of N gradual weightings of groups. Combined saponins were randomly and iteratively sampled from different groups by bootstrap technique. For a same combination, saponins were averaged leading to barycentric glycosylation profile. Iterations of the N barycentric profiles and averaging provided a final response matrix of N smoothed glycosylation profiles from which regulation mechanisms of carbons were highlighted in different aglycone-based saponins. Glucose (Glc) was revealed to be widely favored in GA and 16-OH-GA with more target aspect of 28-Glc in 16-OH-GA and relatively shared distribution between C28 (mainly) C3 and C23 in GA. Strong competition for galactose (Gal) was highlighted between C3 and C28 with target aspects to 28-Gal in GA and 3-Gal in (Gyp, QA). Gyp and QA showed higher regulations of pentoses (xylose, Xyl; arabinose, Ara) with more affinity of GA for (3-Ara, 28-Xyl) and 16-OH-GA for (3-Xyl, 28-Ara). These results call for further investments in simulations of glycosylation mechanisms helping for better understanding metabolic aspects of saponins, and encouraging future analytic experiments in the field.