Optimization of biomass and glucoamylase production by Candida famata using response surface methodology

Glucoamylase is among the most important enzymes in biotechnology. The present study aims to determine better conditions for growth and glucoamylase productivity by Candida famata and to reduce the overall cost of the medium using central composite design (CCD) with one central point and response surface methodology. A three-level central composite design (CCD) factorial design based was employed to obtain optimal medium combination of four independent variables such as soluble starch, (NH4)2HPO4, yeast extract, and MgSO4. 25 randomized mediums were incubated in flask on a rotary shaker at 105 rpm for 72 h at 30°C. The production of biomass was found to be starch, (NH4)2HPO4 and yeast extract dependent; maximum production was obtained when the starch concentration was 5 g/L, yeast extract, 5 g/L and (NH4)2HPO4 2 g/L. Positive interaction was observed between (NH4)2HPO4 and both starch and yeast extract. All the variables were highly significant for glucoamylase production according to their p values; maximum production was found at 5 g/L of yeast extract, 7 g/L of starch and 3 g/L of (NH4)2HPO4; furthermore, yeast extract and (NH4)2HPO4 interacted positively. Central composite design used for the analysis of treatment combinations gave a secondorder polynomial regression model with R 2 = 0.99 for biomass and R 2 = 0.98 for glucoamylase. The final biomass and glucoamylase activity obtained was very close to the calculated parameters; the predicted optimal parameters were confirmed and provide a basis for further studies in the valuation of starch waste products.


INTRODUCTION
The microbial amylases are the most used enzymes, due to their reproductivity (Burhan et al., 2003).Natural fermented media (foods, soils and wastes) are sources for isolation of microorganism strains producing amylases.Amylases have been used as industrial enzymes since decades, these include hydrolases which cause starch hydrolysis and are of great importance in biotechnology with broad range of applications in fermentation, food, textile, brewing, bakery and paper industries (Mushtaq et al., 2016).Amylases can be easily obtained from several sources including plants and animals but microbial enzymes are most preferred (Adrio and Demain, 2014).
Generally, two main methods (classical and statistical) are used for the process of optimization.The classical method is based on the "one-factor-at-a-time" method, in which one independent variable is observed, whereas the other factors are kept at a fixed level.However, this method cannot guarantee the determination of optimal conditions and is unable to study the interactions between the factors, thus probably leads to unreliable results and inaccurate conclusions (Hu et al., 2016).The statistical optimization method (response surface methodology) uses the data from a few sets of experiments to determine equations; this method can overcome the limitations of the classical method, it has been proved to be a powerful tool for designing experiments, building models, evaluating the effects of factors and analyzing optimal conditions of factors for desirable responses (Wu et al., 2014).
Such statistical design had already been used in many research works such as in the improvement of biomass production and glucoamylase activity by Candida famata (Mosbah et al., 2015), in the optimization of α-amylase production by Aspergillus niger (Oberoi et al., 2014), Aspergillus oryzae (Naili et al., 2016) and Bacillus sp.(Murat et al., 2015).These designs were also used for the optimization of the culture medium for the production of inulinase by Kluyveromyces S120 (Xiong et al., 2007) in the production of glucoamylase from food waste (Kiran et al., 2014), and in the optimization study of α-amylase activity (Keharom et al., 2016).
As a continuation of a previous screening work (Lagzouli et al., 2007), the present study aims to determine better conditions for growth and glucoamylase productivity by C. famata, response surface methodology using central composite design with one central point to optimize media composition to reduce the overall cost of the medium to provide a basis for further studies in the baking additives.

Microorganism and culture medium used
C. famata was isolated from traditional Moroccan sourdough (Lagzouli et al., 2007) using medium containing soluble starch (5 g/L), KH2PO4 (3g/L), (NH4)2SO4 (1 g/L), MgSO4 (0.5 g/L) and yeast extract (4 g/L).Initial pH was adjusted to 5 with HCl 0.1 M. The medium was solidified by the addition of 1.5% agar, and autoclaved at 121°C for 15 min.Liquid medium was incubated in flask on a rotary shaker set at 105 rpm for 72 h at 30°C.

Cultivation and production of glucoamylase by C. famata
Growth rate was determined after 72 h of incubation by measuring the absorbance of the suspension at 600 nm and then free substrate supernatant was obtained by centrifugation at 7000 rpm for 10 min and used for estimation of enzyme activity.Glucoamylase activity was determined by measuring the reducing sugar formed by the enzymatic hydrolysis of starch using the method of Nelson (1944); 0.25 ml soluble starch (1%), 0.15 ml phosphate buffer (0.1 M) and 0.1ml enzyme solution were mixed and then incubated at 40°C in water bath for 30 min.The reaction was stopped by 2 ml of Somogyi reactive, and 1.5 ml of distilled water, followed by boiling for 15 min to develop blue color.The absorbance was measured at 540 nm with a spectrophotometer against the control in which no enzyme was added.A calibration curve of absorbance and concentration of glucose was established with known amount of glucose.
One unit (µmol/L/min) of amylase activity was defined as the amount of µmole of reducing sugar per liter of enzymes per min, measured as glucose under the conditions of assay.

Optimization studies
A 3-level 4-factor central composite design was adopted to evaluate the effects of soluble starch (X1), yeast extract (X2), (NH4)2HPO4 (X3) and MgSO4 (X4) on the biomass and glucoamylase production by C. famata.In this study, the independent variables were studied at three different levels, namely low (-1), medium (0) and high (+1), providing 25 trials (mediums).The minimum and maximum ranges of independent variables were investigated with respect to their coded values given that the response variable was fitted by a second order model in order to correlate the response variables to the independent variables, the second order polynomial coefficients were calculated and analyzed using the adequate statistical software.The general form of the second-degree polynomial equation is: Where, Y is the predicted response; b0 is the intercept, bi the linear coefficient, bij is the quadratic coefficient, bii is the linear-by-linear interaction between Xi and Xj regression coefficients, and Xi Xj are input variables that influence the response variable Y.
Statistical analysis of the model was performed to evaluate the analysis of variance (ANOVA), this analysis included Fisher"s F test (overall model significance), its associated probability p(F), correlation coefficient R, determination coefficient R 2 which measure the goodness of fit of regression model.The quadratic models were represented as contour plots (2D) and response surface curves (3D) for each variable.The 25 randomized experiments with the coded and real values of the experimental variables are given in Tables 1 and 2.
The response variable was fitted by a second order model in order to correlate the response variables to the independent variables, the second order polynomial coefficients were calculated and analyzed using the adequate statistical software.The general form of the second-degree polynomial equation is: 2 1 1 0 *Corresponding author.E-mail: lagzoulimohamed@gmail.com.Tel: +21237377559.Fax: +21237372770.Where, Y is the predicted response; b0 is the intercept, bi is the linear coefficient, bij is the quadratic coefficient, bii is the linear-by-linear interaction between Xi and Xj regression coefficients, and Xi Xj are input variables that influence the response variable Y.

RESULTS
In earlier studies, various nutrients were screened for the biomass and glucoamylse production using the traditional one-factor-at-a-time technique (Lagzouli et al., 2007).It was found that starch, (NH 4 ) 2 HPO 4 , yeast extract and magnesium sulfate were the most effective in promoting both biomass and extracellular glucoamylase production.
The experimental and calculated results of experiments carried out with the central composite design are given in Table 3.The analysis of variance (ANOVA) was calculated for each response to determine significant parameters then was carried out by Fisher"s statistical test for the analysis of variance.
The F value is a measure of variation of the data on the mean.Generally, the calculated F value should be  Values of Probability P>F indicated that model terms were significant.The ANOVA of quadratic regression model demonstrates that the model is highly significant, as is evident with the two results from the Fisher"s test with a very high probability value (P model > F) <0.001***) (Table 5).
For all the variables tested, soluble starch, yeast extract, (NH 4 ) 2 HPO 4 , and MgSO 4 were very highly significant for glucoamylase production conforming to p values indicating that the model was extremely affected by these variables.The p value of the linear effect of MgSO 4 (b4) was very significant (p=0.0467),but interact negatively with soluble starch on the glucoamylase production model.Furthermore, interaction between soluble (NH 4 ) 2 HPO 4 and both yeast extract (b23) and MgSO 4 (b24) were positively significant, considering p values.
The contour of the effect of soluble starch (b1) and yeast extract (b2) on the biomass production (Figure 1) are not elliptical; maximum biomass production was represented corresponding to the maximum level of soluble starch (7 g/L), which indicated that maximum biomass production would be reached above (7 g/L).Yeast extract showed a linear plot, which means that it effected slightly the biomass production.
Figure 2 shows that interaction between (NH 4 ) 2 HPO 4 and MgSO 4 on the biomass production indicated a hyperbolic shape with positive effect, biomass activity decreased gradually with an increment of (NH 4 ) 2 HPO 4 to reach to a minimum value at 2 g/L, higher than this value, biomass production increased.In the same plot, maximum biomass production was attained at 0.5 g/L of MgSO 4 , beyond this value, biomass production decreased.
Graphical representation of starch and yeast extract on the glucoamylase production reported in Figure 3 shows a net elliptical peak at 5.0 g/L of yeast extract, the shape of the effect of soluble starch indicates that maximum glucoamylase activity could be reached at 5.5 g/L.
It is evident in Figure 4 that the hyperbolic plot indicates the contour plot is not helicoids; maximum glucoamylase activity shown with 2 g/L of (NH 4 ) 2 HPO 4 and 0.75 g/L of MgSO 4 show a higher effect on glucoamylase.

DISCUSSION
The main goal of response surface is hunt efficiency for the optimum values of the variables such that the response is maximized.Each contour curve represents an infinitive number of combinations of two test variables.The maximum predicted value was indicated by the surface confined in the smallest ellipse in the contour diagram.Elliptical contours are obtained whenever there is a perfect interaction between the independent variables (Guo et al., 2010).

Effect of starch
Biomass became more active in relation to the increase in starch concentration from 2 to 7 g/L, plots are note elliptical which indicated that higher biomass production   would be reached at above 7g/L (Figure 1); whereas, maximum glucoamylase production was found at 6.5 g/L (Figure 3); above this concentration, the production decreases.
Literature abounds on instances, which show the prominent role played by starch as a carbon source for high biomass and glucoamylase production.In contrast, maximum α-amylase production by Candida guilliermondii was at 5 g/L of starch (Acourene et al., 2013); whereas, other studies showed maximum glucoamylase activity by thermophilic fungus, Humicola grisea MTCC 352 at 28.41 g/L (Ramesh and Ramachandra, 2014); starch seems to have an "inductive effect" and portrays a significant role in glucoamylase production (Prajapati et al., 2013).

Effect of yeast extract
Yeast extract exhibited the high biomass production at 4 g/L (Figure 1), in addition, maximum glucoamylase activity was attained at 5 g/L of yeast extract (Figure 3); the contour graphics of the effect of yeast extract on glucoamylase production is not elliptical, which suggest that the maximum glucoamylase could be given by higher concentrations.
As in many other studies, yeast extract helps in the development of mycelial structures with a corresponding higher yield of enzymes (Ramesh and Ramachandra, 2014;Kumar and Satyanarayana, 2007;Arnthong et al., 2015).However, some studies showed that the peptone was the best nitrogen source with Thermomyces lanuginosus, Penicillium fellutanum and Bacillus licheniformis (Kunamneni et al., 2005;Kathiresan and Manivannan, 2006).McTigue et al. (1994) reported that the results of enzyme yield are probably due to the excessive amount of the yeast extract, which may inhibit the production of enzyme when concentration of yeast exceeds a critical value (Alam et al., 1989;Pedersen and Nielsen, 2000).

Effect of (NH 4 ) 2 HPO 4
The glucoamylase production was affected by (NH 4 ) 2 HPO 4 as nitrogen source, which confirms the results found with C. famata in a previous work (Lagzouli et al., 2007).Also, the beneficial effects of (NH 4 ) 2 HPO 4 on the production of amylase was reported by many researchers on Aspergillus fumigatus (Akhter et al., 2013) and by Bacillus subtilis (Demirkan, 2011;Aiyer, 2005).

Effect of MgSO 4
Biomass production highly occurred at 0.5 g/L of MgSO 4 , which are comparable to previously reported results (Lagzouli et al., 2007), in same way, effect of MgSO 4 on the production of biomass and glucoamylase by Humicola grisea MTCC 352 (Ramesh and Ramachandra, 2014) and with Thermomucor indicaeseudaticae by Kumar and Satyanarayana (2007).

Conclusion
In the present study, the authors demonstrated the use of statistical design for the rapid identification and optimization of significant media components for the production of biomass and glucoamylase by C. famata.
Biomass production was found to be starch (NH 4 ) 2 HPO 4 and yeast extract dependent, the maximum Lagzouli et al. 245 was obtained when the starch concentration was 5 g/L, yeast extract was 5 g/L and (NH 4 ) 2 HPO 4 was 2g/L; Positive interaction was observed between (NH 4 ) 2 HPO 4 and both starch and yeast extract.All variables were highly significant for glucoamylase production according to their p values, highest production was found at 5 g/L of yeast extract, 7 g/L of starch and 3 g/L of (NH 4 ) 2 HPO 4 ; furthermore, yeast extract and (NH 4 ) 2 HPO 4 interacted positively.
The predicted values were verified experimentally and gave a second-order polynomial regression model, which was in good agreement with experimental results, with R2 = 0.99 for biomass production and R2 = 0.98 for glucoamylase, according to the p values (p< 0.01).These predicted optimal parameters were confirmed in the laboratory and the final biomass and glucoamylase activity obtained was very close to the calculated parameters.
The optimized media composition found in the present investigation might reduce the overall cost of the medium, and provide a basis for further studies, as a potential candidate for application in baking, detergent industry, or in the valuation of starch waste products.

Figure 1 .
Figure 1.Contour and 3D plot of effects of soluble starch and yeast extract on biomass production.

Figure 2 .
Figure 2. Contour and 3D plot of effects of MgSO4 and (NH4)2HPO4 on biomass production.

Figure 3 .
Figure 3. Contour and 3D plot of effects of soluble starch and yeast extract on glucoamylase production.

Table 1 .
Levels of variables chosen for the central composite design optimization experiment.

Table 2 .
Coded levels (in parentheses) and real values of experimental variables.

Table 3 .
Central composite design for the production of biomass and glucoamylase production by Candida famata.

Table 4 .
ANOVA for central composite model results.

Sum of squares Degrees of freedom Mean square Ratio Prob (P) > F
several times greater than the tabulated value if the model is a good prediction of the experimental results and the estimated factor effects are real.The corresponding analysis of variance (ANOVA) is presented in Table4.The value R 2 of Biomass (Y1) was 0.99 and of glucoamylase (Y2) was 0.98, which indicated good correlation between observed and experimental values of amylase activity.The determination coefficient (R) indicated the fitness of model.The high value (close to 1) of R indicates good correlation between predicted and observed values.

Table 5 .
Significance of regression coefficients of biomass and extracellular glucoamylase production model.