Abstract: The purpose of the current study was the optimization of the solvent extraction of trypsin inhibitor from pea seeds with ultrasonic assistance. A sample granularity of 80 mesh, an ultrasonic frequency of 40 kHz, a radiation time of 13 min, an extraction temperature of 40 ℃, a solid-to-solvent ratio of 1:80 (g/mL), and an ultrasonic power of 500 W were determined to be optimal using one-factor-at-a-time method. Ultrasonic treatment time, extraction temperature, solid-to-solvent and ultrasonic power were taken as independent variables for further optimization by establishing a quadratic polynomial regression model to predict trypsin inhibitor activity using a Box-Behnken design and analyzing the model by response surface methodology. The final optimized extraction conditions were found to be 80 mesh, 40 kHz, 16 min, 50 ℃, 1:100 (g/mL) and 500 W for sample granularity, and ultrasonic frequency, ultrasonic treatment time, extraction temperature, 50 ℃ and solid-to-solvent ratio, respectively. Under these conditions, the maximum trypsin inhibitor activity of 574.54 TIU/g was obtained experimentally, showing a relative error compared to the predicted value of 1.42%. The developed model exhibited an adjusted coefficient of determination (R2Adj) of 0.886 4 and a coefficient of determination (R2) of 0.943 2, indicating?an?excellent goodness of fit. The model is of great guidance significance.

Key words: pea, trypsin inhibitor, extraction, response surface methodology, ultrasonic