Abstract: Twelve kinds of culinary medicinal plant matrices were used as new substrates for making koji. The effect of Rhizopus inoculum size, matrix moisture content, cultivation time and temperature on glucoamylase and liquefaction enzyme activity was examined. Optimization of the process conditions was carried out by the combined use of single factor method and the Box-Behnken design. Glucoamylase activity was used as the response value. The results showed that the optimum process conditions were determined as follows: inoculum size 0.3%, culture time 40 h, matrix water content 65% and culture temperature 28 ℃. Under these conditions, the predicted value of glucoamylase activity was 778.65 mg/(g·h), while the average of three replicate experiments was (779.233 ± 0.577) mg/(g·h). This good agreement indicates that this scheme is feasible. The glucoamylase and liquefaction enzyme activity measured were about 1.3 and 2.8 times higher than that of traditional koji, respectively. A total of 33 volatile components were detected from the two rice wines. Isoamyl alcohol (10.065 versus 109.394 μg/100 mL) and ethyl butyrate (0.867 versus 187.552 μg/100 mL), acetic acid (135.240 versus 560.939 μg/100 mL) and phenol (0.156 versus 0.858 μg/100 mL) showed the greatest differences between the two rice wines. Based on the above results, the new koji had good fermentation performance and the brewed rice wine had a richer taste, showing that this koji to be a promising starter culture for rice wine.

Key words: culinary medicinal plant matrices, koji, response surface methodology, glucoamylase, volatile components