Optimization of panose production by enzymatic synthesis using neural networks

Panose is a trisaccharide constituted by a maltose molecule bonded to a glucose molecule by an α-1,6-glycosidic bond. This trisaccharide has the potential to be used in the food industry as a non-cariogenic sweetener and is considered a pre-biotic carbohydrate. Despite the industrial potential applications of panose, a large-scale process for its production is still not available. In this paper, the production of panose by enzymatic synthesis in a batch and a fed-batch reactor was optimized by means of the use of neural network. The optimization of panose has shown that the maximum productivity is obtained when the process is conducted in fed-batch with initial sucrose concentration S0 = 100 mmol/L, a maltose to sucrose ratio of 5.58 and a feed flow rate of 3000 L/h, yielding 35.3 g/L h of panose, which is 74% higher than the yield that can be obtained through the batch process.