To improve dextransucrase production from Leuconostoc mesenteroides NRRL B-640 culture medium was screened and optimized using the statistical design techniques of Plackett-Burman and response surface methodology (RSM). Plackett-Burman design with six variables viz. sucrose, yeast extract, K2HPO4, peptone, beef extract and Tween 80 was performed to screen the nutrients that were significantly affecting dextransucrase production. The variables sucrose, K2HPO4, yeast extract and beef extract showed above 90% confidence levels for dextransucrase production and were considered as significant factors for optimization using response surface methodology. 2(4)-central composite design was used for RSM optimization. The experimental results were fitted to a second-order polynomial model which gave a coefficient of determination R-2 = 0.95. The optimized composition of 30 g/l sucrose, 18.9 g/l yeast extract, 19.4 g/l K2HPO4 and 15 g/l beef extract gave an experimental value of dextransucrase activity of 10.7 U/ml which corresponded well with the predicted value of 10.9 U/ml by the model. (c) 2008 Elsevier Ltd. All rights reserved.