Optimization of productivity of pullulan by means of multivariable linear regression analysis

Abstract Biosynthesis of pullulan by a mutant clone of Pullularia pullulans (producing polysaccharide not contaminated with melanins) was optimized by means of finding numerically, using multivariable linear regression analysis, the maximum of an arbitrarily chosen second-order polynomial model function of several variables which describe the dependence of the pullulan productivity (expressed in wt. % of the fermentation broth) on medium component concentrations and pH. The determined optimum medium composition for a 4-day fermentation at 28°C: sucrose, 35.3; NH 4 NO 3 , 0.031; KNO 3 , 0.092; MgSO 4 · 7H 2 O, 0.023; K 2 HPO 4 , 0.069; FeSO 4 · 7H 2 O, 0.0035; yeast extract, 0.031 (all in wt. %), initial pH 5.56, allowed to obtain 12.2 wt. % of pullulan (3.5 times higher than before optimization).