Neural network-ant-colony hybrid optimization of a bacterial fruit juice clarifying metallo-neutral-protease production.

Abstract

Neutral proteases are invaluable for food, pharmaceutical, and feed industries because of their mild catalysis. This study reports the production and potential food industry application of a neutral protease by a coastal sediment Bacillus cereus strain. An integrated approach involving response surface methodology (RSM) and artificial neural network-ant-colony hybrid optimization (ANN-ACO) was adopted to develop a suitable bioprocess medium for improved yield. The enzyme was purified, characterized for activity and stability to physicochemical parameters, and investigated for fruit juice clarification. The ANN-ACO model was observed to be superior (predicted r ² = 98.5%, mean squared error [MSE] = 0.0353) to the RSM model (predicted r ² = 86.4%, MSE = 23.85). It recommended a medium containing (gL^- 1) molasses 45.00, urea 9.81, casein 25.45, Ca²⁺ 1.23, Zn²⁺ 0.021, Mn²⁺ 0.020, and 4.45% (vv^- 1) inoculum, for a 6.75-fold increase in protease activity from 76.63 UmL^- 1 in un-optimized medium. The 10.0-fold purified 46.6-kDa-enzyme had maximum activity at pH 6.5, 45-55ºC. The 353 amino acid protein had its HEXXH motif between 173 and 177 residues. Its substantial requirements for Zn²⁺ and Ca²⁺, and significant inhibition by EDTA confirmed its metallo-protease nature. Successful clarification of fruit juices underscores its biotechnological potential.

Supplementary information: The online version contains supplementary material available at 10.1007/s13197-024-06095-w.