# Production of invertase and inulinase

**Published:** **Last Edited:**

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

## Abstract

A newly isolated yeast Candida guilliermondii TISTR 5844 was shown to be a good producer of invertase and inulinase in an inexpensive 48-hour batch fermentation that used inulin as the sole carbon source. Ammonium chloride was used as the sole nitrogen source after screening several other sources. Fermentation conditions (pH, temperature, inoculum size, inulin concentration) were optimized for producing the two enzymes. Production of invertase was most sensitive to the culture pH and the temperature. In contrast, none of the aforementioned factors had a significant influence on the production of inulinase within the experimental space. The optimized medium for producing invertase consisted of 5% inulin, 4.8% NH4Cl and 0.6% Mg2SO4ïƒ-7H2O at a pH value of 6. Inulinase production was maximized by using a medium of 1% inulin, 2.4% NH4Cl and 1.2% Mg2SO4ïƒ-7H2O at a pH value of 5. The optimal temperature was 30 ï‚°C for both enzymes.

Key words: invertase, inulinase, Candida guilliermondii, optimization, response surface methodology, Taguchi method

## List of symbols

CCD Central composite design

CI Confidence interval defined by Eq. 3

CL Confidence level

DOFFactor Degree of freedom of factors

ei The random error in the ith experiment

FÎ± (1, fe) F-ratio at a confidence level of (1âˆ’Î±) for a given degree of freedom (DOF); 1, fe is

the error in DOF

Fratio F-ratio

MS Mean squares

n, N Number of experiments

QP Volumetric production rate of enzymes (g Lï€1 hï€1)

RSM Response surface methodology

SI Severity index

S/N Signal-to-noise ratio

S The sample size for confirmation test

S' The pure sum of squares

SSError Sum of squares of error

SSFactor Sum of squares of factors

t Time (s)

Grand average of signal-to-noise ratio defined by Eq. 6

ve Error variance

VFactor Variance of a factor

Xi The fixed effect of the level combination used in the ith experiment

Yexpected The expected values of enzyme activity, yield or productivity (U Lï€1, U gï€1, U Lï€1 hï€1, respectively) (=Yexpected,Obs in the explanation, which were converted from Yexpected,S/N)

yi The comparison variable in experiment

YP/S Enzyme yield based on substrate (U enzyme gï€1 substrate)

Greek letters

ï Specific growth rate (hï€1)

ï¢ The overall mean of the performance value (Eq. 2)

## 1. Introduction

This work deals with the production of invertase and inulinase using a newly isolated yeast Candida guilliermondii TISTR 5844 (Sirisansaneeyakul et al., 2007). Invertase catalyzes the hydrolysis of terminal non-reducing ï¢-fructofuranoside residues in ï¢-fructofuranosides. In commercial processes, invertase is used to convert sucrose to fructose and glucose. Fructose is sweeter than sucrose and does not crystallize as easily. It is therefore preferred to sucrose in the confectionery industry (Rubio and Navarro, 2006). The second enzyme of interest here is inulinase or ï¢-fructan fructanohydrolase, a commercial catalyst for the hydrolysis of inulin. (Inulin is a ï¢-(2,1)-linked polysaccharide of fructose with a terminal glucose residue in the polysaccharide chain.) Inulinase hydrolyzes inulin to fructose and fructo-oligosaccharides, both of which are commonly used in the food and pharmaceutical industry (Vandamme and Derycke, 1983).

Invertase and inulinase are generally produced using yeasts and filamentous fungi. The yeast Candida guilliermondii TISTR 5844 recently isolated from Jerusalem artichoke tubers is a novel producer of invertase and inulinase (Sirisansaneeyakul et al., 2007). This strain is capable of producing copious quantities of invertase and lesser levels of inulinase when grown on inulin and has the potential for commercial use in producing these enzymes.

Bulk enzymes such as invertase and inulinase are relatively low-value products that must be produced inexpensively. This typically necessitates a batch fermentation, ideally without the control of the pH. Production of many enzymes by fermentation is highly susceptible to the composition of the culture medium and factors such as feedback inhibition by the carbon source. Therefore, identifying the medium composition to maximize the enzyme titer in a minimally controlled batch fermentation is important.

Traditionally, fermentation processes have been optimized by changing one independent variable or factor at a time while keeping the others at some fixed values. This single dimensional search is slow and laborious, especially if a large number of independent variables are involved. Consequently, statistical methods are increasingly preferred for fermentation optimization because they reduce the total number of experiments needed and provide a better understanding of the interactions among factors on the outcome of the fermentation (Revankar and Lele, 2006). Statistical techniques such as the response surface methodology (RSM) and Taguchi method have gained broad acceptance in fermentation optimization.

Taguchi's method of orthogonal array experimental design involves the study of a given system by changing the values (or levels) of a set of independent variables (or factors) over the range of interest. For a given number of independent variables tested at a given number of levels, the Taguchi's method specifies orthogonal arrays for combining the various variables and levels in a minimum acceptable number of experimental trials. This method determines the optimal levels of the important controllable factors based on the concept of robustness and "signal-to-noise" (S/N) ratio (Roy, 1990). The desired design is sought by selecting the best performance under conditions that produce a consistent performance (Roy, 2001). The conclusions drawn from the experiments are valid over the entire experimental space spanned by levels of the controlled factors (Phadke and Dehnad, 1988). Whereas the traditional experiment design focuses on the average process performance characteristics, Taguchi's method concentrates on the effect of variation on the process characteristics. In addition, Taguchi's approach facilitates the identification of the influence of individual factors and interactive effects of factors on performance with a few well-defined experimental sets (Prasad and Mohan, 2005). Taguchi's method combined with the RSM is a powerful tool for production optimization. It not only saves time but allows to rapidly construct accurate models for optimization.

Here the focus is on optimizing the production of the enzymes invertase and inulinase by C. guilliermondii TISTR 5844 using a low-cost minimally controlled batch fermentation process. The fermentation conditions and the medium composition are optimized using a combination of the RSM, the one factor at a time method and the Taguchi method.

## 2. Materials and Methods

## 2.1 Microorganism and enzyme assay

The yeast Candida guilliermondii TISTR 5844 (Sirisansaneeyakul et al., 2007) was used to simultaneously produce invertase and inulinase in a batch fermentation process. The activities of invertase and inulinase were measured as previously specified (Sirisansaneeyakul et al., 2007). One unit of invertase activity was defined as the quantity of the enzyme that liberated 1 ïmol of fructose in 1 min in a 0.5% w/v (g/100 mL) solution of sucrose in 0.5 M McIlvaine buffer at pH 5.0 and 40 Â°C. One unit of inulinase activity was defined as the quantity of the enzyme that liberated 1 ïmol of fructose in 1 min in a 0.5% w/v solution of inulin in 0.5 M McIlvaine buffer at pH 5.0 and 40 Â°C.

## 2.2 Optimization of invertase and inulinase fermentation by RSM

All fermentations were performed in accordance with a central composite design (CCD) in 500 mL Erlenmeyer flasks filled with 200 mL of the culture medium. The latter contained 1.2% w/v yeast extract and 0.2% MgSO4Â·7H2O in 0.1 M McIlvaine buffer, pH 5. Inulin at concentrations of 1-2% was the carbon source. After inoculation, the flasks were incubated for 48 h and assayed for enzyme activity. The RSM was used for the investigation of the effects of independent variables: pH, temperature, inoculum size and inulin concentration on the responses of invertase and inulinase activities. Using a CCD for 4 factors, 30 treatment combinations were generated. To set up a statistical model, five levels for each variable were chosen. The upper and lower limits of each variable were chosen to encompass the range in the literature and therefore to be consistent with the actual practice for this fermentation.

Data were analyzed using SPSS 12 software (SPSS Inc., Chicago, IL, USA) for Windows (Microsoft Corporation, Redmond, WA, USA) to obtain the relevant regression equations, regression coefficients and analysis of variance.

## 2.3 Optimization by one factor at a time method and Taguchi method

## 2.3.1 One factor at a time method

To investigate the effects of carbon source on the production of invertase and inulinase, in separate experiments inulin was supplemented with an equal mass of glucose, fructose, sucrose and starch. To establish the effects of nitrogen source on enzyme production, yeast extract was replaced with beef extract, peptone, NH4Cl, (NH4)2SO4 and NH4NO3 in separate experiments.

## 2.3.2 Taguchi methodology

The Taguchi method uses various types of signal-to-noise (S/N) ratios to measure the variability around the target performance (Engin et al., 2008). A high value of a S/N implies that the signal is much higher than the random effects of the noise factors. The noise is usually due to the uncontrollable factors that often cannot be completely eliminated. From the point of view of the performance of a process, three possible performance attributes are: (1) the-smaller-the-better; (2) the nominal smaller-the-better; and (3) the larger-the-better (Yang et al., 2007). In this study, the performance attribute of "the larger-the-better" was used to define the optimum conditions. The S/N for "the larger-the-better" performance attribute was estimated using the following equation:

(1)

where yi is the combination variable in experiment i for a certain combination of controlled factor levels and n is the number of experiments performed for that combination.

Some times, no optimal conditions can be identified for a process within the entire experimental space selected for the study. In such cases, the balanced characteristics of the orthogonal experimental array can help in predicting the performance value corresponding to the optimum operation conditions, using the following additive model:

(2)

In Equation (2) ï¢ is the overall mean of the performance value, Xi is the fixed effect of the quantity level combination used in the ith experiment, and ei is the random error in the ith experiment. Because Equation (2) is a point estimation that is used to determine whether the results of the confirmation experiments are meaningful, the confidence interval (CI) of Yi must be calculated. This is done using the following equation at the selected level of confidence:

(3)

where Fï¡(1, fe) is F-ratio at a confidence level of (1âˆ’Î±) for a given degree of freedom (DOF), (1, fe) is the error DOF, neff is (N/{1+ total DOF associated with the estimate of the mean}), N is total number of results, S is the larger-the-better sample size for confirmation test, ve is error variance, and CI is confidence interval (Engin et al., 2008).

In this study, four factors at three levels of variations (Table 1) were used in the experiments. The factors optimized included the concentrations of inulin, NH4Cl, MgSO4ïƒ-7H2O and H+ (i.e. the pH). The various combinations of factors and levels were in accordance with Taguchi's L9 orthogonal array. The factor level combinations for all the experiments are shown in Table 2. Submerged batch fermentations were conducted using various media formulations (1, 3 and 5% inulin; 1.2, 2.4 and 4.8% NH4Cl; 0.2, 0.6 and 1.2% MgSO4ïƒ-7H2O) in 0.1 M McIlvaine buffer at pH values of 4, 5 and 6. Enzyme activities were assayed after 48 h of incubation.

## 3. Results and Discussion

## 3.1 Optimization by RSM

In keeping with the selected experimental design, the experimental data and the response correlated with the following second order general polynomial regression model:

(4)

where Y is the final activity of invertase or inulinase (U Lï€1) in the broth, bi's are the linear coefficients, bii's are the quadratic coefficients, bij's are the cross product coefficients and ï™ is the model constant.

The specific RSM models obtained for the two enzymes and the corresponding R2-values are shown in Table 3. (R2-value for a model is a measure of its goodness of fit at a given level of significance.) Analysis of variance (ANOVA) suggested that the invertase model (Table 3) was highly significant (p < 0.05). The R2-value indicated that 82% of the variability in the data could be explained by the model. However, the R2-value for the model of inulinase at 0.538 (p < 0.05) indicated that the factors studied did not significantly influence its production at a given level of significance, but poorly for use in the prediction; nevertheless, it could be used as a basis for Taguchi optimization. *please check the revised statement again. This is the only sense I can make of what you wrote * (See additional sheet 1 of explanation)

The model in Table 3 suggested that the optimal conditions for producing invertase were in the operational range of pH of 5-8, temperature of 25-40 Â°C, inoculum size of 6-10% (v/v) and the initial inulin concentration of >3% (g/100 mL). Similarly, based on the model for inulinase (Table 3), the optimal conditions for producing it were likely in the operational range of pH 2-6, temperature 30-40 Â°C, inoculum size 6-10% and an initial substrate concentration of >3%.

Further optimization using the Taguchi method was performed within the above ranges, but using the initial inulin concentration in the range of 1-5%, a narrower pH range of 4-6, a fixed temperature of 30 ï‚°C and a fixed inoculum size of 10%, as specified in Table 4. The expected range of attainable activities within this experimental space was 220-1229 U Lï€1 for invertase and 335-3287 U Lï€1 for inulinase (Table 4).

## 3.2 Optimization by one factor at a time method and Taguchi methodology

## 3.2.1 One factor at a time method

The invertase and inulinase activities were highest when only inulin was used as the carbon source. These activities were 1,094 and 370 U Lï€1 for invertase and inulinase, respectively. Hence, inulin was selected as the optimal carbon source for further study. Use of inulin as a sole carbon source is known to enhance production of inulinase by Kluyveromyces marxianus YS-1 (Singh and Bhermi, 2008) and Pichia guilliermondii (Gong et al., 2007). Apparently, production of inulinase is induced by inulin (Selvakumar and Pandey, 1999; Looten et al., 1987; Singh and Bhermi, 2008) and actually repressed by some of the other carbon sources (Looten et al., 1987).

Inorganic nitrogen sources were superior to organic nitrogen sources for the production of the two enzymes. Ammonium chloride afforded the maximum invertase and inulinase activities at 488 and 473 U Lï€1, respectively, and was selected as the nitrogen source for all further work.

## 3.2.2 Taguchi methodology

The experimental data was processed using the Qualitek-4 software (Nutek, Inc., Bloomfield Hills, MI, USA), with the larger-the-better attribute selected for establishing the optimum composition of the fermentation medium and identifying the individual factors that influenced enzyme production. The percentage contributions of the factors to production of invertase and inulinase are shown in Table 5 and Table 6, respectively. The carbon source (i.e. inulin, factor A) and nitrogen source (i.e. NH4Cl, factor B) were significant factors for the production of invertase (Table 5). The carbon source and the concentration of magnesium sulfate (factor C) appeared to strongly influence the production of inulinase (Table 6). As the degree of freedom of the error was zero (Table 5, Table 6), information regarding the sum of the squares could not be determined for the error and the F-ratios for the factors could not be calculated. To complete the analysis, smaller factorial effects were added together, or pooled, to obtain a nonzero estimate of the error term (Table 5, Table 6). Pooling was done until the degrees of freedom (DOF) of the error term became close to half of the total of DOF. Thus MgSO4ïƒ-7H2O (factor C) and pH (factor D) were pooled. Based on the pooled data, initial concentration of inulin and NH4Cl had a clear substantial influence on the results. However, when interactions of different factors were calculated (Table 7), factors that had low influence individually (e.g. factor C, MgSO4ïƒ-7H2O; factor D, pH) had interactively the strongest influence on invertase production. For example, the interactive effect of the factors C and D had the highest percent severity index (Table 7). Similarly, the percent severity index for the interaction of pH (factor D, the least important factor individually) with inulin (factor A, the most influential factor individually) was only 0.2%. For inulinase, the interactive effect of the concentration of ammonium chloride (factor B) and the concentration of magnesium sulfate (factor C) had the strongest influence on activity (Table 8). These results suggest that the influence of one factor on enzyme production depended on the value of the other factors in the production process.

The above mentioned pooling was used to identify the optimum fermentation conditions shown in Table 9. The relevant main effect plots are shown in Figure 1. A main effect plot reveals how the changes in a factor level affect the response of the fermentation process. For four factors (A-D) each at three levels (1-3), only one of the levels maximized the value of the mean S/N ratio (Figure 1). Figure 1 suggests that the conditions for attaining a high activity, yield and productivity of any of the two enzymes are different.

The equations in Table 9 were used to estimate the expected activity (Yexpected) of the enzymes under various conditions. The highest predicted inulinase activity of (174 U Lï€1) was for the conditions CON-1 (i.e. 10 g Lï€1 inulin, 24 g Lï€1 NH4Cl, 12 g Lï€1 MgSO4ïƒ-7H2O, pH 5, 30 ï‚°C). This set of conditions was therefore selected for further experiments. In contrast, two different sets of conditions were revealed in relation to invertase. One set of conditions, CON-2, maximized the predicted final activity of invertase. These conditions are (CON-2): 50 g Lï€1 inulin, 48 g Lï€1 NH4Cl, 6 g Lï€1 MgSO4ïƒ-7H2O, pH 6 and 30 ï‚°C. A different set of conditions (CON-3) maximized both the predicted invertase yield on inulin and invertase productivity. This set of conditions was as follows (CON-3): 10 g Lï€1 inulin, 24 g Lï€1 NH4Cl, 6 g Lï€1 MgSO4ïƒ-7H2O, pH 6 and 30 ï‚°C.

Under the optimal conditions identified by the Taguchi method for maximizing the invertase activity, i.e. CON-2, the predicted value of the response parameter, i.e. the invertase activity could be calculated using the following equations:

(5)

where was calculated using the following equation:

(6)

For the pooled ANOVA (factor C, MgSO4ïƒ-7H2O, and D, pH, pooled) for invertase activity, Eq. (5) was revised to the following:

(7)

From Eq. (6) and (7), we have .

The problem that I pointed out here has not been resolved. Unless I get an absolutely clear explanation, I can do noting about this issue

(8)

Therefore, The value of Yexpected for invertase was calculated from the predicted optimal values in Table 10; thus,

where and therefore Yexpected = 750.4.

(See additional sheet 2 of explanation)

This additional explanation is not satisfactory! Just below Equation (7) you claim that Yexpected is 57.506. How does the same Yexpected become 750.4? Yexpected cannot have two different values.

(See additional sheet 2 of explanation)

Where exactly to site Table 10?

(See additional sheet 2 of explanation)

## 3.3.3 Confirmation under optimal conditions

Confirmation testing is a necessary requirement of the Taguchi method. A single confirmation test was conducted for enzyme production using the above identified optimum settings of the process parameters. The 90% confidence interval (CI) of the confirmation test was calculated using Eq. (3) and found to be ï‚±3.562. Therefore, the predicted optimal range was estimated as: (57.506 ï€ 3.562) < Yexpected < (57.506 + 3.562), or 53.944 < Yexpected < 61.068.

The confirmation test results for the set of conditions CON 1ï€3 are shown in Table 11. For both enzymes, the measured activity and productivity were within the expected ranges that were consistent with the predictions. The specific growth rate of the yeast at all three conditions (CON 1-3) was similar at 0.1136, 0.1035 and 0.1199 hï€1 (Table 11), but enzyme production varied greatly with the specific conditions used.

Table 11 reveals comparable levels of invertase activity and productivity for the conditions CON-2 and CON-3. In contrast, the confirmation data in Figure 2 showed the highest inulinase activity (261.59 U Lï€1) and productivity (1.98 U Lï€1 hï€1) under CON-2 (Figure 2b). CON-1 and CON-3 favored the preferential production of invertase. CON-2 maximized the production of both invertase and inulinase.

Under CON-2, maximal production of the two enzymes occurred during the exponential phase of growth within the first 24 h (Figure 2b). Inulinase production declined continuously with the approach of the stationary phase of growth (Figure 2b). In contrast, under CON-1, the level of inulinase production was relative low but was not adversely affected by the onset of the stationary phase of growth (Figure 2a). The high level of inulinase attained within the first 24 h under CON-2 (Figure 2) compares well with the results for the other producer microorganisms where the inulinase activity generally peaks much later in the fermentation. For example, for Aspergillus niger-245, Cruz et al. (1998) reported peak production by after 48-60 h of fermentation. Similarly, using K. marxianus YS-1, Singh and Bhermi (2008) observed the highest inulinase activity after 60 h of incubation.

## 4. Conclusion

The yeast C. guilliermondii TISTR 5844 was found to be a good producer of invertase and inulinase. The yields of the enzyme were enhanced substantially by optimization of the culture conditions and the medium composition. Conditions were established for preferential production of only invertase, but a different set of fermentation conditions simultaneously afforded high titers of both invertase and inulinase. Concentrations of the carbon source (inulin) and the nitrogen source (NH4Cl) were found to have the most impact on the production of the enzymes. Under optimal conditions, the titers of both the enzymes peaked much earlier than has been generally attained in fermentations with other microorganisms. In view of its ability to rapidly produce both invertase and inulinase in a minimally controlled batch fermentation, C. guilliermondii TISTR 5844 is potentially useful for producing a commercial enzyme cocktail for making sweeteners from inulin.

## 5. Acknowledgements

The authors would like to thank Kasetsart University Research and Development Institute (KURDI), and the Center for Advanced Studies in Tropical Natural Resources, Kasetsart University Institute for Advanced Studies (KUIAS) for financial support.