Optimization of the Medium Components by Statistical Experimental Methods to Enhance Nattokinase Activity

Optimization of the Medium Components by Statistical Experimental Methods to Enhance Nattokinase Activity

Fooyin J Health Sci 2009;1(1):21−27 O R IGINAL ART IC L E Optimization of the Medium Components by Statistical Experimental Methods to Enhance Natto...

164KB Sizes 0 Downloads 7 Views

Fooyin J Health Sci 2009;1(1):21−27

O R IGINAL ART IC L E

Optimization of the Medium Components by Statistical Experimental Methods to Enhance Nattokinase Activity Jau-Kai Wang*, Hua-Hsien Chiu, Ching-Shieh Hsieh Technology Development Center of Molecular Diagnosis and Fermentation Engineering, Fooyin University, Kaohsiung, Taiwan Received: March 27, 2009

Revised: June 1, 2009 Accepted: June 2, 2009

Natto, a traditional Japanese soy food, contains nattokinase, a potent fibrinolytic enzyme with fibrinolytic properties. Statistical experimental methods were used to find the optimal conditions of nattokinase production by Bacillus natto. Effects of various nitrogen, carbon, and inorganic salt sources were investigated to select the optimal medium components. In the variable screening, a 25−1 fractional factorial design was used to examine the experimental variables. Screening experiments showed that glucose, KH2PO4, and MgSO4 significantly affected the enzyme activity of nattokinase. Based on these results, the Box-Behnken design and response surface methodology were applied to find the optimal conditions. A regression model was built by fitting the experimental results with a second-order polynomial. The regression model was statistically significant since the coefficient of determination (R2) was 0.9352. By analyzing the response surface plots and their corresponding contour plots with the regression model, the optimal variable conditions were obtained as follows: glucose, 0.065%; KH2PO4, 0.0016%; and MgSO4, 0.0016%. The corresponding maximal activity of nattokinase was 12.34 FU/mL. Confirmation experiments were performed and the results showed that the difference between the predicted and experimental values of nattokinase activity were within 15%. A theoretical approach calculated from numerical calculation is in agreement with the experimental data. Key Words: Box-Behnken design; medium; nattokinase; optimization; regression; statistical experimental method

Introduction Natto, a traditional Japanese soy food, is produced by fermenting steamed soybeans with the bacterium Bacillus subtilis (formerly termed B. natto).1 Nattokinase, a potent fibrinolytic enzyme, was primarily isolated from Natto by Sumi et al.2 Recent research has shown that nattokinase has fibrinolytic properties, and can be used to treat thrombosis.3 As result, the use of Natto in food products has increased in Japan.4

To date, research has focused on the isolation and screening of microorganisms with high fibrinolytic activity for later use in enzyme production,5 or on purifying and characterizing newly found enzymes.6 In contrast, few reports focus on culture medium optimization.7,8 In biotechnology-based industrial processes, the formulation of culture media is of critical importance—its composition affects the product concentration, yield, and volumetric productivity. Cost reduction is another important aspect of medium optimization. The complexities

*Corresponding author. Technology Development Center of Molecular Diagnosis and Fermentation Engineering, Fooyin University, 151 Chin-Hsueh Road, Kaohsiung 831, Taiwan. E-mail: [email protected] ©2009 Fooyin University

22 and uncertainties associated with large-scale fermentation usually arise from a lack of knowledge regarding the sophisticated interactions among the different variables.9 Conventional practice, by single variable optimization (while others are kept constant), does not elucidate the combined effects of all the variables involved.10 The limitations of single variable optimization can be eliminated by optimizing all the affecting parameters collectively using a statistical experimental method with a Box-Behnken design and response surface methodology.11,12 This is designed by using statistical methods to yield the most information in the minimum number of experiments. It has been successfully applied to the optimization of medium composition,13 conditions of enzymatic hydrolysis,14 parameters of food preservation,15 and fermentation processes.16 The aim of this study was to optimize the fermentation conditions to maximize nattokinase activity. Stepwise optimization was performed as follows: 1) selection of the optimal carbon and nitrogen source via the one-variable-at-a-time method; 2) screening of the significant variables that affect nattokinase activity using a 25−1 fractional factorial design; 3) searching the optimal conditions of the significant variables using the Box-Behnken design with response surface methodology; and 4) a confirmation experiment to verify the optimal conditions.

Materials and Methods

J.K. Wang et al MgSO4, 0.5; pH 7.0−7.2]. The subsequent cultivation lasted for 12 hours at 37ºC in an orbital shaker to obtain seed culture. Next, a 5% (v/v) seed culture was added to the production medium (consisting of ingredients dictated by the experimental design). Throughout the work, liquid cultures were incubated at 37ºC and in an orbital shake water bath at 250 rpm. After 72 hours of fermentation, cells were removed by centrifugation, and the supernatant was used for assaying the nattokinase activity assay.8

Nattokinase activity assay The fibrinolytic activity of nattokinase was measured by the hydrolysis of fibrin.17 The incubation mixture contained 2.5 mL of 12 g/L fibrin solution (pH 7.8), 6.5 mL of 0.1 M Tris−HCl buffer (containing 10 mM CaCl2, pH 7.8), and 1 mL enzyme solution in a suitable dilution. The incubation was carried out at 37ºC for 15 minutes and ended by adding 5 mL of 0.11 M trichloroacetic acid, containing 0.22 M sodium acetate and 0.33 M acetic acid. The absorbency at 275 nm of the supernatant was determined after centrifugation. A fibrinolytic unit (FU) was defined as the amount of enzyme that resulted in an increase in absorbency at 275 nm—equivalent to 1 mg of tyrosine per minute at 37ºC.

Methodology and experimental design In this study, Design Expert package (version 6.0.10; Stat-Ease Inc., Minneapolis, MN, USA, 2003) was used for experimental designs and analysis of experimental results.

Microorganism De-hulled yellow-seeded soybeans were soaked overnight in tap water at 5ºC to avoid fermentative acidification. The water was discarded and the beans were cooked in fresh tap water for 30 minutes (ratio of beans:water = 1:3). They were then cooled and superficially dried at room temperature. One hundred grams of cooked soybeans were then transferred to a glass flask and auto-cleaved at 120ºC for 30 minutes, cooled, and inoculated with 5 mL of diluted culture. After mixing, the beans were fermented at 37ºC for 72 hours. B. natto was isolated and maintained as spores suspended in 50% glycerol, stored at −18ºC.4 The chemical reagents used were extra pure grade of Hayashi Pure Chemical Industries Ltd. (Osaka, Japan).

Cultivation and production of nattokinase A 5% (v/v) spore suspension was added to seed 200 mL of medium [medium constituents and properties (g/L): glucose, 10; yeast extract, 10; K2HPO4, 1;

Selecting the optimal nitrogen, carbon and inorganic sources Five sources of nitrogen, six sources of carbon, and eight inorganic sources were investigated using a one-at-a-time strategy. The nitrogen sources used were: soy meal, 20 g/L; soy peptone, 10 g/L; sodium glutamate, 21 g/L; ammonium phosphate, 5.0 g/L; and yeast extract broth, 10 g/L. All nitrogen sources had equivalent nitrogen element contents. The carbon sources used were: crystal sugar, maltose, sucrose, glycerol, soybean meal, and glucose. The concentration of all carbon sources was 20 g/L. The inorganic sources used were: CaCl2, NaCl, KCl2, MgSO4, CuSO4, MnSO4, and K2HPO4; in all cases, the concentration was 0.05%. To investigate nitrogen sources, growth was allowed in the minimal synthetic medium: K2HPO4, 1 g/L; and MgSO4, 0.5 g/L. This was supplemented with 10 g/L of glucose and the nitrogen source to be investigated. To screen carbon sources, fermentation was carried out in the same minimal synthetic medium, supplemented with the optimal nitrogen source and the carbon source

Optimization medium for nattokinase activity

23

to be investigated. Each observation was done in triplicate. Screening the significant variables Based on the preliminary study, we chose the following five variables as the experimental variables: MgSO4 (X1), KH2PO4 (X2), yeast extract (X3), glucose (X4), and maltose (X5). A 25−1 resolution V fractional factorial design was used to determine the significant variables that affected nattokinase activity. Each variable had two levels to be examined—a high (+1) and low level (−1). The high and low levels selected represented the extremes of normal operating ranges. Table 1 shows the variables and levels in detail. The screening, experimental design, and experimental results are shown in Table 2. The experimental variables and interactions were tested with analysis of variance (Table 3). Optimization by response surface methodology Response surface methodology is an empirical modeling technique used to evaluate the relationship between a set of the controlled experimental variables and the measured responses. A prior knowledge and understanding of the process and process variables under investigation are necessary to create a more realistic model. Based on the results of variable screening, the model examined the experimental variables required for optimal nattokinase activity using the Box-Behnken design13 and response surface Table 1

Variables and their levels in a 25−1 fractional factorial design Code level

Variables

Range (%)

X1: MgSO4 X2: KH2PO4 X3: yeast extract X4: glucose X5: maltose

0.01−0.3 0.01−0.3 0.5−1.5 0.1−1.0 0.1−1.0

−1

+1

0.01 0.01 0.5 0.1 0.1

0.3 0.3 1.5 1.0 1.0

methodology.14 In developing the regression model, the experimental variables were coded according to the following equation: xi =

Xi − X0 ⌬Xi

(1)

where xi is the coded value of the variable Xi, X0 is the value of Xi at the center point, and ⌬Xi is the step change value. The range and the levels of the experimental variables investigated in this study are shown in Table 4. The Box-Behnken design matrix is shown in Table 5. Once the experiments were performed, the regression model was constructed by fitting the experimental results with a second-order polynomial. The optimal conditions of selected variables were found using the regression model and by analysis of the response surface plots.

Table 2

25−1 fractional factorial design and experimental results

Trial number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Variables X1

X2

X3

X4

X5

−1 1 −1 1 −1 1 −1 1 −1 1 −1 1 −1 1 −1 1

−1 −1 1 1 −1 −1 1 1 −1 −1 1 1 −1 −1 1 1

−1 −1 −1 −1 1 1 1 1 −1 −1 −1 −1 1 1 1 1

−1 −1 −1 −1 −1 −1 −1 −1 1 1 1 1 1 1 1 1

1 −1 −1 1 −1 1 1 −1 −1 1 1 −1 1 −1 −1 1

Nattokinase activity (FU/mL) 6.2 8.4 9.5 11.9 7.5 9.9 10.1 11.9 10.7 12.0 11.6 13.0 9.4 11.4 10.4 13.0

Table 3 Analysis of variance for screening experiments Source X1 X2 X3 X4 X5 X2X4 Residual Cor total Cor: correlation.

Sum of squares

Degrees of freedom

Mean square

F value

Prob. > F

Effect

16.20 15.80 0.01 16.20 0.11 2.98 4.01 55.29

1 1 1 1 1 1 9 15

16.20 15.80 0.01 16.20 0.11 2.98 0.45

36.40 35.50 0.01 36.40 0.24 6.69

0.0002 0.0002 0.9130 0.0002 0.6378 0.0294

2.01 1.99 0.038 2.01 0.16 −0.86

10 5

0 0 0 0 −1 −1 1 1 −1 −1 1 1 0 0 0 0 0

8.40 10.05 9.61 10.00 6.58 11.00 10.62 11.44 7.70 11.33 11.36 12.23 12.14 11.84 11.58 11.89 12.11

So yb

ic am

25 20 15 10 5 0

in So yb m ean ea l Gl uc os e

−1 −1 1 1 0 0 0 0 −1 1 −1 1 0 0 0 0 0

yc er

−1 1 −1 1 −1 1 −1 1 0 0 0 0 0 0 0 0 0

Figure 1 Effect of various nitrogen sources on production of nattokinase in the medium solution: glucose, 10 g/L; K2HPO4, 1 g/L; MgSO4, 0.5 g/L; pH 7.0 ± 0.2, 37ºC.

Gl

X4

e

X2

ea

n 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

X1

Nattokinase activity (FU/mL)

Cr su ysta ga l r M al to se

Variables Trial number

Nattokinase activity (FU/mL)

Table 5 Box-Behnken design and experimental results

sa lt Am ph mo os ni ph um at e Ye as te xt br rac ot t h

0 e

0.0031 0.0031 0.011

ut

0.0016 0.0016 0.006

Gl

0.0001 0.0001 0.001

15

cr os

+1

to n

0

Su

0.0001−0.003 0.0001−0.003 0.001−0.011

−1

m

X1: MgSO4 X2: KH2PO4 X4: glucose

Range (%)

ep

Variables

20

l

Code level

25

ea

Variables and their levels for a Box-Behnken design

So yp

Table 4

J.K. Wang et al Nattokinase activity (FU/mL)

24

Figure 2 Effect of various carbon sources on production of nattokinase in the medium solution: yeast extract, 10 g/L; K2HPO4, 1 g/L; MgSO4, 0.5 g/L; pH 7.0 ± 0.2, 37ºC.

Results In order to screen different nitrogen sources, soybean meals, soy peptones, sodium glutamate, ammonium phosphate, and yeast extract were replaced at the same concentration (1%) with carbon source (glucose) as indicated in the reference media. The complex media was inoculated and cultivated for 3 days in the fermentation conditions outlined above. The nattokinase activity for each nitrogen source is shown in Figure 1. Of all the nitrogen sources investigated, yeast extract was the most promising, and the corresponding nattokinase activity was 20.83 FU/mL. When inorganic nitrogen sources were used, very poor enzyme activities were achieved— much higher activities were obtained with organic nitrogen sources. Soy meal media led to poorer results than the other three organic nitrogen sources, a finding presumably due to its poor solubility. Research has reported that reduced nitrogen (as found in ammonium ions, amino groups, or amide groups) is the form utilized in biosynthesis.18 The organic nitrogen sources in these observations—yeast extract, soy peptone, and soy meal—could extract

ammonium and amino acids by enzyme catalysis, all of which are the preferred nitrogen source of B. subtilis.19 Our results showed that the optimal nitrogen source was yeast extract. Crystal sugar, maltose, sucrose, glycerol, soybean meal, and glucose were screened as main carbon sources, by replacing at the same concentration (1%) with nitrogen source (yeast extract) as indicated in the reference media. The complex media was inoculated and cultivated for 3 days at the fermentation conditions indicated in the experimental section of this paper. B. subtilis uses glucose as the most preferred source for carbon and energy (Figure 2),20,21 and it was applied as reference. The other four carbon sources all gave similarly good results in respect of protease production by strains of B. subtilis. The results showed that, in terms of nattokinase production, maltose and glucose were the optimal carbon source. Considering these results, and the associated lower cost, glucose was chosen as the carbon source for further optimization studies. In order to study the effects of inorganic salts on nattokinase production, the reference media

20 15

Discussion

10 5

e

ric

Cu p

M

So

di

um

ch

lo

lo ch m

siu as

rid

rid

rid lo ch m ciu

e an su ga ne lfate M se ag su ne siu lfat e m su Po lfa Fe ta te rri ss c iu su m lfa di te h ph yd os rog ph en at e

0

Po t

Ca l

25 0.006. The Box-Behnken design and experimental results are shown in Table 5.

25

e

Nattokinase activity (FU/mL)

Optimization medium for nattokinase activity

Figure 3 Effect of various inorganic salts on production of nattokinase in the medium solution: glucose, 10 g/L; yeast extract, 10 g/L; pH 7.0 ± 0.2, 37ºC.

was enriched with CaCl2, NaCl, KCl2, MgSO4, CuSO4, MnSO4, and K2HPO4. The complex media was inoculated and cultivated for 3 days using the standard fermentation conditions. Nattokinase activity and inorganic salt choice are shown in Figure 3. The results show that heavy metal ions—including cupric ion, manganese ion and ferric ion—result in relatively poor fermentation of B. subtilis.22,23 The use of inorganic salts sources—K2HPO4 and MgSO4— led to a high rate of nattokinase production. The optimal inorganic sources were potassium dihydrogen phosphate and magnesium sulfate (both nitrogen sources). As a preliminary step for optimization, we applied a 25−1 fractional factorial design to screen the experimental variables. The experimental design and experimental results are shown Table 2. Depending on the media components, the production of nattokinase varied from 6.2 to 13.0 FU/mL. Results from analysis of variance are shown in Table 3— nattokinase activity was significantly affected by MgSO4 (p = 0.0002), KH2PO4 (p = 0.0002), and glucose (p = 0.0002); there was a significant interaction between glucose and K2HPO4 (p = 0.0294). Maltose and yeast extract did not significantly affect nattokinase activity. MgSO4, KH2PO4, and glucose all increased nattokinase activity, a finding that suggests they could be used to promote activity in future media; these variables were further investigated by a Box-Behnken design. Based on the results of variable screening, we selected MgSO4 (X1), KH2PO4 (X2) and glucose (X4) as experimental variables to find the optimal conditions for higher nattokinase activity. The range and the levels of the experimental variables investigated in this study are shown in Table 4. The central values (zero level) chosen for experimental design were: MgSO4, 0.00155; KH2PO4, 0.00155; and glucose,

The experimental results were analyzed using statistical methods appropriate for the experimental design used. After discovering the significant variables, a full second-order regression model was applied to fit the experimental results as follows: Y = B0 + B1X1 + B2X2 + B4X4 + B11X12 + B22X22 + B44X42 + B12X1X2 + B14X1X4 + B24X2X4

(2)

where Y is the nattokinase activity (FU/mL), Bi is the model coefficients, X1 the coded variable of MgSO4, X2 the coded variable of KH2PO4, and X4 the coded variable of glucose. Regression analysis gave the following second-order regression model: Y = 11.91 + 0.91X1 + 0.71X2 + 1.13X4 − 1.57X12 − 0.83X22 − 0.43X42 − 0.32X1X2 − 0.90X1X4 − 0.69X2X4

(3)

Coefficients and significance level of regression analysis are shown in Table 6. The results showed that three linear terms, two quadratic terms (X1X2), and two-variable interactions (X1X4) had a significant effect (p < 0.05) on nattokinase activity. The R2 (determination coefficient) of the regression equation obtained from analysis of variance (shown in Table 7) was 0.9352 (a value > 0.75 indicates aptness of the model), which indicates that the model explains 93.52% of the variance in the response. The optimal concentrations for the production of nattokinase were: 0.0016 of MgSO4, 0.0016 of K2HPO4, and 0.065 of glucose. The predicted maximum nattokinase activity was calculated to be 12.34 FU/mL. Comparisons of experimental value and predicted value of the regression model in Equation 3 are shown in Table 8—data shows that agreement was satisfactory. In order to confirm the predicted results, experiments using the optimal conditions were performed and a value of 13.0 FU/mL was obtained. The aim of this study was to increase the productivity of Natto cultivation using a statistical experiment design method. Statistical optimization methods overcome the limitations of classic empirical methods and are a powerful tool for the optimization of nattokinase production. Data from this study showed that nattokinase production is dependent mainly on MgSO4, K2HPO4, and yeast extract. The developed medium showed a high nattokinase activity—13 ± 1 FU/mL, six times higher than

26

J.K. Wang et al

Table 6 Coefficients and significance levels of regression analysis Source

Coefficient

Sum of squares

Degrees of freedom

Mean square

F value

Prob. > F

Model X1 X2 X4 X12 X22 X42 X1X2 X1X4 X2X4

0.91 0.71 1.13 −1.57 −0.83 −0.43 −0.32 −0.90 −0.69

41.61 6.62 4.00 10.22 10.39 2.87 0.78 0.40 3.24 1.90

9 1 1 1 1 1 1 1 1 1

4.62 6.62 4.00 10.22 10.39 2.87 0.78 0.40 3.24 1.90

11.23 16.09 9.73 24.81 25.24 6.98 1.90 0.96 7.87 4.63

0.0022 0.0051 0.0169 0.0016 0.0015 0.0334 0.2106 0.3589 0.0263 0.0686

Table 7 Analysis of variance for the regression model Source

Sum of squares

Degrees of freedom

Mean square

F value

Prob. > F

R2

41.61 2.88 44.49

9 7 16

4.62 0.41

11.23

0.0022

0.9352

Regression Residual Cor total Cor: correlation.

Table 8 Trial no. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Comparisons between the predicted and actual value Actual value

Predicted value

Residual

8.40 10.05 9.61 10.00 6.58 11.00 10.62 11.44 7.70 11.33 11.36 12.23 12.14 11.84 11.58 11.89 12.11

7.58 10.03 9.63 10.82 6.97 10.59 11.03 11.05 8.13 10.92 11.77 11.80 11.91 11.91 11.91 11.91 11.91

0.82 0.018 −0.018 −0.82 −0.39 0.41 −0.41 0.39 −0.43 0.41 −0.41 0.43 0.23 −0.072 −0.33 −0.022 0.20

the original medium. On the basis of our findings, a theoretical approach calculated from numerical calculation is in agreement with the experimental data.

Acknowledgments The authors sincerely appreciate the financial support of the National Science Council (NSC93-2622E-242-002-CC3) and Limmer Biotech Co. Ltd., Taiwan, for this work.

References 1.

Ichishima E, Kato M, Wada Y, et al. Spore fatty acid composition in Bacillus natto, a food microorganism fermented soybean food. Food Chem 1982;8:1−9. 2. Sumi H, Hamada H, Nakanishi K. Enhancement of the fibrinolytic activity in plasma by oral administration of nattokinase. Acta Haemotol 1990;84:139−43. 3. Fujita M, Nomura K, Hong K, et al. Purification and characterization of a strong fibrinolytic enzyme (nattokinase) in the vegetable cheese natto, a popular soybean fermented food in Japan. Biochem Biophys Commun 1993;197:1340−7. 4. Chang CT, Fan MH, Kuo FC, et al. Potent fibrinolytic enzyme from a mutant of Bacillus subtilis IMR-NK-1. J Agri Food Chem 2000;48:3210−6. 5. Samy A, Aassar E. Production and properties of fibrinolytic enzyme in solid state cultures of Fusarium pallidoroseum. Biotechnol Lett 1995;17:943−8. 6. Beaumont M. Flavouring composition prepared by fermentation with Bacillus. Intern J Food Microbiol 2002;75: 189−96. 7. Elibol M. Optimization of medium composition for actinorhodin production by Streptomyces coelicolor A3(2) with response surface methodology. Process Biochem 2004;31: 1057−62. 8. Liu J, Xing J, Chang T, et al. Optimization of nutritional conditions for nattokinase production by Bacillus natto NLSSE using statistical experimental methods. Process Biochem 2005;40:2757−62. 9. Montgomery DC. Design and Analysis of Experiments, 3rd edition. Wiley: New York, 1991. 10. Haaland PD. Experimental Design in Biotechnology. Dekker: New York, 1989. 11. Khuri AI, Cornell JA. Response Surfaces Design and Analysis. Dekker: New York, 1987. 12. Adinarayana K, Ellaiah P, Srinivasulu B, et al. Response surface methodological approach to optimize the nutritional parameters for neomycin production by Streptomyces marinensis under solid-state fermentation. Process Biochem 2003;38:1565−72.

Optimization medium for nattokinase activity 13. Dean A, Voss D. Response surface methodology. In: Design and Analysis of Experiments. New York: Springer-Verlag, 1999:547−86. 14. Myers RH, Montgomery DC. Response Surface Methodology: Process and Product Optimization Using Designed Experiments. New York: John Wiley & Sons, 1995. 15. Almeidae S, Lima UA, Taqueda M, et al. Use of response surface methodology for selection of nutrient levels for culturing Paecilomyces variotii in eucalyptus hemicellulosic hydrolyzate. Bioresource Technol 2003;87:45−50. 16. Sato T, Yamada Y, Ohtani Y, et al. Production of menaquinone (vitamin K2)-7 by Bacillus subtilis. J Biosci Bioeng 2001;91:16−20. 17. Gupta R, Beg Q, Khan S, et al. An overview on fermentation, downstream processing and properties of microbial alkaline proteases. App Microbiol Biotechnol 2002;60: 381−95.

27 18. Allagheny N, Obanu ZA, Platt GC, et al. Control of ammonia formation during Bacillus subtilis fermentation of legumes. Food Microbiol 1996;29:321−33. 19. Fisher SH. Control of carbon and nitrogen metabolism in Bacillus subtilis. Annu Rev Microbiol 1991;45:107−35. 20. Stuke J, Hillen W. Regulation of carbon catabolism in Bacillus species. Annu Rev Microbiol 2002;54:849−80. 21. Xu ZF, Affleck R, Wangikar P, et al. Transition state stabilization of subtilisins in organic media. Biotechnol Bioeng 1993;43:515−20. 22. Wei Q, Hall CW, Chang KC. Natto characteristics as affected by steaming time, Bacillus strain, and fermentation time. J Food Sci 2001;66:167−73. 23. Ammar YB, Matsubara T, Ito K, et al. Some properties of levansucrase of Bacillus natto stabilized with periodate oxidized glucomannan. Enzyme Microbiol Technol 2002; 30:875−82.