Indian Journal of Biotechnology Vol 4, January 2005, pp 127-133 Kinetic study of a low molecular weight protease from newly isolated Pseudomonas sp. using artificial neural network Jayati Ray Dutta 1 , Pranab Kumar Dutta 2 and Rintu Banerjee 1 * 1 Microbial Biotechnology and Downstream Processing Laboratory, Agricultural and Food Engineering Department and 2 Department of Electrical Engineering, Indian Institute of Technology, Kharagpur 721 302, India Received 27 June 2003; revised 19 January 2004; accepted 28 January 2004 A soil isolate, identified as Pseudomonas sp. produced an extra-cellular protease enzyme of 14.4 kDa molecular weight. The kinetic properties of the purified fraction of the bacterial protease were studied experimentally and the rate of casein hydrolysis was predicted by a model based on artificial neural network. The various kinetic factors studied were in- cubation time, initial enzyme concentration, initial substrate concentration, pH and temperature. The prediction error in simulating casein hydrolysis was less than 1%. Keywords: Pseudomonas sp., protease, kinetics, simulation, artificial neural network IPC Code: Int. Cl. 7 A 10 N 63/02; C 12 N 9/52 Introduction Proteases find enormous industrial applications in detergents, leather, baking, brewery, dairy, photo- graphic industries, etc 1 . Purification and characteriza- tion of protease, therefore, play an important role in the success of its commercial production. There are many types of conventional purification techniques like gel-filtration, ion exchange chromatography, af- finity chromatography, high performance liquid chromatography, ultrafiltration and electrophoresis used for purification 2 . In most cases satisfactory puri- fication is obtained by using many of these techniques in combination. Kinetic study of purified enzyme constitutes the in- formation pertaining to rates of activation and inacti- vation of enzymes and actually gives the rate at which a process occurs 3 . The kinetic parameters controlling the rates of enzyme catalyzed reaction are: (i) incuba- tion time, (ii) initial substrate concentration, (iii) ini- tial enzyme concentration, (iv) pH, and (v) reaction temperature. The dependence of the reaction rate on these parameters is in general non-linear in nature and there can be considerable amount of cross-coupling between the different parameters. Modeling of this integrated behaviour is difficult to capture within a set of empirical equations. Artificial neural network (ANN) can solve such problems efficiently. There is a report on ANN based computations be- ing applied for developing predictive models as a re- sult of combined effects of three environmental fac- tors (temperature, pH and water activity) on thermal inactivation rate of Escherichia coli 4 . In this paper, the modelling ability of ANN algorithm compared to that of response surface methodology (RSM) method showed the superiority of the accurate prediction by ANN-based approach. Similarly, ANN and RSM have been used to build up a predictive model of the joint effect of NaCl concentration, pH level and storage temperature on kinetic parameters of the growth curve of Lactobacillus plantarum. In the present study also ANN showed much better results than RSM tech- nique 5 . Static and dynamic neural networks were also used for on-line estimation of biomass concentration during batch cultures of Streptococcus salivarius ssp. thermophilus 404 and Lactobacillus delbrueckii ssp. bulgaricus 398 at controlled pH and temperature 6 . This paper reports the purification and kinetic study of extra-cellular protease enzyme from a newly isolated Pseudomonas sp. RAJR 044 along with its parametric behaviour. The results were integrated into an ANN based model that can be used for extensive simulation studies. Materials and Methods Microorganism A new soil isolate from Indian Institute of Tech- nology (IIT) Kharagpur Campus, identified as Pseu- —————— *Author for correspondence: Tel: 91-3222-283104; Fax: 91-3222-282244 E-mail: rb@agfe.iitkgp.ernet.in