Abstract: Fucoidanase was used to degrade fucoidan to obtain fucosylated oligosaccharides. A total of 122 experiments were designed and implemented using uniform design and orthogonal array design. A three-layer BP network model was built to simulate the hydrolysis process, and it was optimized by genetic algorithm using Matlab software. A mouse model of D-galactose-induced aging was used to test the antioxidant activities of fucosylated oligosaccharides and fucoidan. The results showed that the molecular weight of hydrolysates was changed significantly during 4 h hydrolysis at 20–40 ℃ with 0–24 U/mg of fucoidanase. On the basis of this finding, the optimization was carried out and the result showed that an enzyme dosage of 20.7 U/mg, a temperature of 26 ℃ and a hydrolysis time of 2.3 h were found to be the optimal conditions to obtain a hydrolysate with the highest antioxidant activity. The hydrolysate scavenged 85.1% 1,1-diphenyl-2-picrylhydrazyl (DPPH) radical with a molecular weight of 18.2 kDa. The results of animal experiments showed that the antioxidant activity of fucosylated oligosaccharides were significantly higher than that of fucoidan in terms of hydroxyl free radical scavenging activity, catalase (CAT) activity and malonaldehyde (MDA) content (P < 0.05). This paper can provide experimental support for the industrial preparation of fucoidan hydrolyzate with high antioxidant activity.

Key words: fucosylated oligosaccharides, artificial neural network, genetic algorithm, antioxidant activity, enzymatic hydrolysis, molecular weight