3.3. Optimization of fermentation conditions for enhanced 2-PE production from L-phe
Based on the above results using the method of “one time one factor”, further statistical experimental design methodology was applied to evaluate the interaction and determine the optimal level of three influencing factors including temperature, initial glucose and initial L-phe levels. For the response surface analysis, 17 experiments with triplicates were conducted according to the RSM design, where Y is 2-PE production (g/L), X1 denotes temperature (℃), X2 denotes initial glucose level (g/L) and X3 denotes initial L-phe level (g/L) (Table 3). According to the response values obtained from the experimental results, a second order regression equation was generated for the response surface, as follows:
Y = 2.28 + 0.013X1 − 0.12X2 + 0.072X3 − 0.12X1X2 + 0.032X1X3 + 0.0016X2X3 – 0.34X12 − 0.87X22 – 0.50X32 (Eq. 2, [R2 = 0.9710; R2(Adj) = 0.9337; R2(Pred) = 0.5362]). The regression model (Eq. (1)) gave a high R2 value of 0.9710 (Table 4), which indicated aptness of the model (Isar, Agarwal, Saran, & Saxena, 2006). The R2 value is between 0 and 1. When R2 is close to 1.0, the model will better predict the response (Aghaie, Pazouki, Hosseini, Ranjbar, & Ghavipanjeh, 2009). Notably, the model F value of 26.06 and values of probability (P) > F (0.0001) indicated that the model terms were significant (Bezerra, Santelli, Oliveira, Villar, & Escaleira, 2008). The three-dimensional response surface plot is generally used to demonstrate relationships between the response and experimental levels of each variable (Fig. 3). The highest point on the contour profiles in Fig. 3 indicates the optimal parameter values of the highest 2-PE production. As shown, there was obvious interaction between each pair of variables, and the interaction between the selected three variables was significant. The optimal 2-PE production value predicted from the response surface model was 2.37 g/L, when the temperature was controlled at about 24.7℃ and initial glucose and L-phe concentrations were 63.5 and 10.7 g/L, respectively.
The following validation experiment in 3.0 L bioreactor was carried out under the optimized conditions with 63.54 g/L of initial glucose, 10.70 of L-phe g/L at 24.7℃. The 2-PE production was finally improved up to 2.55 g/L after 96 h, which is 7.8% higher than the predicted level, representing the highest 2-PE production from L-phe by using M. guilliermondii (Karolina et al., 2017). Hence, the models developed were considered to be accurate and reliable for predicting 2-PE production from L-phe by using strain YLG18.