The estimated effect for each variable and the interac- tions between the variables were determined and are re- ported in Table 5. Both variables are relevant to tannase production at a 95% confidence level. The correlation analysis has shown that the MgSO4 interaction versus the agitation rate and the MgSO4 quadratic term were not statistically significant. The agitation rate quadratic term and both linear terms are significant, which affects the curvature in the response surface plot. The analyses of the results show that the deviations were low in the de- sirable range where the activity value was high.
Table 6 shows the analysis of variance (ANOVA) for tannase production. The terms with no significance were removed from the ANOVA. The results of the ANOVA show good reproducibility of the data observed with a very low error value. The determination coefficient (R2) explained 97.6% of the variability of the data, and the F-test results demonstrated that the fit of the model was very good (F-regression >3.71).
The regression analysis of the experimental data ob- tained after the ANOVA resulted in the adjusted second order polynomial Equation (2) below:
Tannase activity Y( ) = 2 (2)
0.170−0.036x1 + 0.124x2 + 0.103x2 where Y is the tannase produced as a function of the coded levels of MgSO4•7H2O concentration (x1) and agi- tation rate (x2). The predicted values and the relative de- viation between the real and predicted response for each assay are presented in Table 2, and the response surface and contour curve with the influence of the variables MgSO4 and agitation rate on enzyme activity are pre- sented in Figure 2.
Figure 2 shows that the optimum conditions for tan- nase production are at the highest levels of agitation (between 110 and 120 rpm) and that the variations in the MgSO4 levels do not exhibit a greater influence on the tannase activity. This result indicates that the MgSO4 levels should be between 0.25 and 0.5 g/L. These results are different from the previous ones from the Plackett- Burman design, most likely due to differences in the tested ranges.
In this simplified low cost medium, Aspergillus sp. GM4 tannase demonstrated an activity of 539 mU/mL, representing a 2.66-fold increase compared to the activity obtained before applying the experimental design. The experimental design reduced the number of medium components and slightly increased the tannase produc- tion.