Statistical and neural network modeling of β-glucanase production by Streptomyces albogriseolus (PQ002238), and immobilization on chitosan-coated magnetic microparticles

Abstract

β-Glucanases are a series of glycoside hydrolases (GHs) that are of special interest for various medical and biotechnological applications. Numerous β-glucanases were produced by different types of microorganisms. Particularly, bacterial β-glucanases have the privilege of being stable, easily produced, and suitable for large-scale production. This study aimed for finding potent β-glucanase producing bacterial strains and optimizing its production. Soil samples from Egyptian governorates were screened for such strains, and 96 isolates were collected. The β-glucanase activity was qualitatively assessed and quantitatively measured using 3,5-dinitrosalicylic acid (DNS) method. The highest β-glucanase producing strain (0.74 U/ml) was identified as Streptomyces albogriseolus S13-1. The optimum incubation period and temperature, determined one-variable at a time, were estimated as 4 d and 45 ͦ C, respectively. Similarly, yeast β-glucan and beef extract were selected as the best carbon and nitrogen sources, with enzymatic activities of 0.74 and 1.12 U/ml, respectively. Other fermentation conditions were optimized through response surface methodology (RSM); D-optimal design (DOD) with a total of 28 runs. The maximum experimental β-glucanase activity (1.3 U/ml) was obtained with pH 6.5, inoculum volume of 0.5% v/v, agitation speed of 100 rpm, carbon concentration of 1% w/v, and nitrogen concentration of 0.11% w/v. This was 1.76-fold higher compared to unoptimized conditions. Using the same experimental matrix, an artificial neural network (ANN) was built to predict β-glucanase production by the isolated strain. Predicted β-glucanase levels by RSM and ANN were 1.79 and 1.32 U/ml, respectively. Both models slightly over-estimated production levels, but ANN showed higher predictivity and better performance metrics. The enzyme was partially purified through acetone precipitation, characterized, and immobilized on chitosan-coated iron oxide microparticles. The optimal pH and temperature for enzyme activity were 5 and 50 °C, respectively. The immobilized enzyme showed superior characters such as higher stability at temperatures 50, 60, and 70 °C compared to the free enzyme, and satisfactory reusability, losing only 30% of activity after 6 cycles of reuse.

Keywords: Characterization; Immobilization; Neural network; Optimization; Response surface methodology; β-Glucanase.

Fig. 1

Phylogenetic tree showing the position of isolate S13-1 relative to its closest phylogenetic neighbors. The strain was identified as S. albogriseolus S13-1

Fig. 2

3D surface plots showing interactions between: A pH and beef extract concentration. B Yeast β-glucan and beef extract concentrations. C Agitation speed and beef extract concentration. D pH and yeast β-glucan concentration

Fig. 3

The Actual versus predicted plot of ANN and RSM models, constructed by JMP Pro 17.0 model comparison dialogue

Fig. 4

FTIR spectrum of A chitosan coated iron oxide microparticles, B glutaraldehyde-activated iron oxide microparticles, C enzyme-linked microparticles

Fig. 5

SEM images of A chitosan coated iron oxide microparticles, B glutaraldehyde-activated iron oxide microparticles, C enzyme-linked microparticles

Fig. 6

Reuse of β-glucanase enzyme from S. albogriseolus S13-1 over six cycles. Vertical bars are displayed as mean ± standard deviation

Fig. 7

A Effect of pH on β-glucanase activity, B effect of temperature on β-glucanase activity, C thermal stability of free β-glucanase, D thermal stability of immobilized β-glucanase