Many bacteria especially several species belonging to Bacillus are known to produce variety of extracellular enzymes and they have a wide range of industrial applications [1]. Most commercial amylases are produced from a small subgroup of Bacillus species such as B. amyloliquefaciens, B. subtilis, and B. cereus [1–4].Various bacterial protease have been widely used such as B. subtilis, B. licheniformis, and B. mojavensis [5–7]. Xylanase from bacteria, such as B. pumilus, are being applied in textile processing [8]. Numerous bacteria belonging to Bacillus such as B. licheniformis, B. subtilis and B. subtilis natto have been suggested as promising microbial rennet producers [9–12].

Wheat bran, an agro-industrial residue, a cheap source of energy has high potential in the area of fermentation for the production of enzymes. Wheat bran contains cellulose material, starch, crude protein, trace elements and other certain ingredients, which can be used as carbon and nitrogen sources to promote the growth of microorganisms and enzyme production. There are several reports describing wheat bran as potent substrate for enzyme production [12, 13]. We recently reported several bacteria producing milk-clotting enzyme isolated from the yak grazing soil in the north-eastern Tibetan Plateau. Among these bacteria, B. amyloliquefaciens D4 possessed high rennet-producing capacity in wheat bran juice [14, 15]. The objective of this work was to attempt to optimize the culture conditions to increase milk-clotting enzyme production from wheat bran by B. amyloliquefaciens D4.

The producer strain was B. amyloliquefaciens D4(CGMCC 3290), which was deposited in the China General Microbiological Culture Collection Center (Beijing, China). The strain was propagated at 37 °C on lysogeny broth (LB) agar slants (1.0 % (w/v) peptone, 1.0 % (w/v) beef extract, 0.5 % (w/v) NaCl, and 2.0 % (w/v) agar, pH 7.2), and subcultured every 30 days.

The strains obtained from the LB agar slants were inoculated into 5 mL of seed culture medium, composed of beef extract (10 g/L), peptone (3 g/L), and NaCl (5 g/L) in a test tube, and incubated at 37 °C at pH 7.2 for 24 h with shaking at 170 rpm. After incubation, a 1-mL aliquot of the bacteria was inoculated into 100 mL of fermentation medium in a 250-mL flask and incubated for 24 h at 37 °C at pH 7.2 with shaking at 170 rpm. The culture broth was used as seed culture for the later experiments. The fermentation medium was prepared as follows: 100 g of wheat bran in 1,000 mL of distilled water were boiled for 10 min, and filtered through gauze.

All experiments were conducted in 250-mL Erlenmeyer flasks containing fermentation medium. The initial culture conditions were as follows: initial pH 6.0, agitation speed 140 rpm, inoculum ages 4 h, fermentation time 36 h, medium volume in flask 40 mL, temperature 32 °C, inoculum size 3 %. The other different culture conditions were tested according to the experimental statistical design. After fermentation, the crude enzyme solution was obtained by centrifugation at 8000×g for 10 min. All fermentations were carried out in triplicate and the results represent the average of the three trials.

The milk-clotting activity was investigated using the method of Arima and expressed in terms of the Soxhlet unit (SU), which is defined as the amount of enzyme required to clot 1 mL of a substrate solution (10 % skim milk in 10 mM of CaCl2) in 40 min at 35 °C. Enzyme solution (0.5 mL) was added to 5 mL of the substrate solution containing 10 % skim milk powder and 10 mM of calcium chloride and incubated at 35 °C for 5 min. The mixture was mixed well and the formation time of the curd fragment was measured.

The culture conditions having the most significant effect on milk-clotting activity were identified using a two-level Plackett–Burman design. The regression analysis of the Plackett–Burman design shows that low levels of X5 (medium volume in flask) and X7 (inoculum size) enhanced milk-clotting production, whereas high levels of X1 (initial pH), X2 (agitation speed), X3 (inoculum ages), X4 (fermentation time), and X6 (temperature) resulted in high milk-clotting activity. Initial pH, agitation speed, and fermentation time were found to be statistically significant medium components with high confidence levels, but inoculum ages, medium volume in flask, temperature, and inoculum size were not.

Although the most significant variables affecting milk-clotting activity were screened by Plackett–Burman design, it was unable to predict the optimum levels of each variable. In such circumstances, we want to move rapidly to the general vicinity of the optimum levels of the variables. The method of steepest ascent was employed to find the proper direction to change the variables by increasing the initial pH, agitation speed and fermentation time to improve enzyme production. It was found that the maximum value of milk-clotting activity was reached at the third step. Then, these variables were chosen for further optimization.

Three significant independent variables [X1(initial pH), X2(agitation speed), and X4(fermentation time)] were selected and further optimized using the Box–Behnken design to determine their optimal levels based on the above results. In the Box–Behnken design, X5 (medium volume in flask) and X7 (inoculum size) were set at its low levels of 40 mL and 3 %, but X3 (inoculum ages) and X6 (temperature) were set at their high levels of 7 h and 37 °C, respectively. Table 1 shows the Box-Behnken experimental design and the obtained milk-clotting activity of the individual variables. Via multiple regression analysis on the experimental data using Minitab 14.11 software, the following second-order polynomial equation was obtained:

Table 1 Design and results of Box–Benhnken design
Full size table
$$ \begin{gathered} {\text{Y}} = 1 9 6 6. 6- 1 8.0 9 {\text{X}}_{ 1} - 3 3. 5 8 {\text{X}}_{ 2} + 1 6 2. 5 6 {\text{X}}_{ 4} - 1 8 7. 8 8 {\text{X}}_{ 1}^{ 2} - 3 4 3. 3 {\text{X}}_{ 2}^{ 2} - 2 5 3. 6 3 {\text{X}}_{ 4}^{ 2} + 1 1. 9 3 {\text{X}}_{ 1} {\text{X}}_{ 2} - 9. 8 5\hfill \\ {\text{X}}_{ 1} {\text{X}}_{ 4} + 1 7 9. 8 3 {\text{X}}_{ 2} {\text{X}}_{ 4} \hfill \\ \end{gathered} $$
(1)

where Y is the predicted response and X1, X2, and X4 are the coded values of initial pH, agitation speed, and fermentation time, respectively.

The regression coefficients and the analysis of the variance presented in Table 2 indicate the high significance of the model. Our results revealed that linear and quadratic terms of fermentation time had a significant effect on the milk-clotting activity production (p < 0.01). Simultaneously, the square of initial pH, agitation speed, fermentation time and interactive terms of agitation speed and fermentation time were also significant. From equations derived by differentiation of Eq. 1, the optimal values of X1, X2, and X4 in the coded units were found to be −0.0557, 0.0378 and 0.3349, respectively. Correspondingly, we obtained the maximum point of the model, which was initial pH 7.57, agitation speed 241 rpm, fermentation time 53.3 h, respectively. The maximum predicted value of milk-clotting activity was 1994.1 SU/mL.

Table 2 Analysis of variance (ANOVA) for the quadratic model
Full size table

In order to confirm the predicted results of the model, we repeated the experiments under optimal conditions. The milk-clotting enzyme activity of 1996.9 ± 3.1 SU/mL in the statistically optimized conditions was achieved, which was 2.92-fold higher than that of the initial conditions and was only 0.14 % higher than the predicted value of 1994.1 SU/mL according to Eq. 1. The good correlation between these two results validates the model and the existence of an optimal value.

Herein, we investigated the characterizations and applications of the milk-clotting enzyme and found that the enzyme was a metalloprotease with a molecular weight of 58.2 kDa and was completely inactivated by heating at 55 °C for 20 min. The optimum temperature and optimum pH were 65 °C and 5.5, respectively [15]. Milk-clotting enzyme production by B. amyloliquefaciens has been successfully applied in the preparation of cheese (Fig. 1). Considering these properties, B. amyloliquefaciens D4 is a promising producer of microbial rennet.

Fig. 1
figure 1

Cheese made with milk-clotting enzyme produced by Bacillus amyloliquefaciens D4

Full size image