Cost-Effective Optimization of Bacterial Urease Activity Using a Hybrid Method Based on Response Surface Methodology and Artificial Neural Networks

Microbially-induced carbonate precipitation (MICP) based on urea hydrolysis presents a promising novel method in the field of geotechnical and geoenvironmental engineering applications. To gain economics of the large-scale bio-cementation through MICP application, this study has been conducted to evaluate corn steep liquor (CSL), an inexpensive agro-industrial byproduct, and a defined medium for culturing the urease-positive Sporosarcina pasteurii, which is the costliest and most important stage of the MICP process. For this purpose, the effects and interactions of CSL, urea concentration, incubation time, and nickel supplementation were assessed on the production of urease in a central composite design (CCD) approach. In addition, an artificial neural network (ANN) was developed to model and predict urease yield, and thereafter, the performance of the two models were compared. Based on the statistical indices obtained, the ANN model exhibited relatively higher predictive ability and accuracy than CCD in describing nonlinear behavior of the biological process. A maximum activity of 3.6 mM urea min−1 was obtained for 5% v/v CSL, 4.75 g L−1 urea concentration, 60 h incubation time, and 0 μM nickel supplementation at optimal conditions. To lower the operating cost, further analysis showed that the produced enzyme activity of the non-sterile medium suffered a steep decline (34%) compared to the optimum enzyme activity. The obtained results make clear that the CSL-urea medium not only provides comparable ureolysis efficiency but also suggests an extremely economic advantage compared to commonly utilized yeast extract nutrient medium.