Predictive Modeling of Riboflavin Production in Lactiplantibacillus plantarum MTCC 25432 Using Fuzzy Inference System

Abstract

Riboflavin (Vitamin B2) is an essential vitamin and a microbial metabolite produced by some lactic acid bacteria (LAB). This investigation aims to study the overproduction of riboflavin in selected Lactiplantibacillus plantarum strain by using the one factor at a time (OFAT) tool coupled with the Fuzzy Inference System (FIS) and its validation through fermentative production in semi-defined media. Out of three Lactiplantibacillus strains used in this study, the maximum riboflavin producing strain was selected based on its ability to grow and produce higher levels of riboflavin. In results, Lactiplantibacillus plantarum strain MTCC 25432 was able to produce 346 µg/L riboflavin in riboflavin deficient assay medium and was investigated further. By using the OFAT–fuzzy FIS system, casamino acid in the range of 5–20 g/L, GTP 0.01–0.04 g/L, sodium acetate 5–15 g/L, and glycine 5–15 g/L were used to predict their effect on riboflavin production. The conditions optimized with modeling showed a 24% increment in riboflavin production (429 µg/L) by Lactiplantibacillus plantarum MTCC 25432 vis-a-vis the unoptimized counterpart (346 µg/L). In conclusion, an FIS-based predictive model was effectively implemented to estimate the riboflavin within an acceptable limit of 3.4%. Riboflavin production enhancing effects observed with various levels of sodium acetate, casamino acid, and GTP could be useful to re-design matrices for riboflavin production.