LWT - Food Science and Technology 42 (2009) 327331

Contents lists available at ScienceDirect

LWT - Food Science and Technology

journal homepage: www.elsevier.com/locate/lwt

Optimization of natural fermentative medium for selenium-enriched yeast by D-optimal mixture design
Hongfei Yin, Zhigang Chen, Zhenxin Gu*, Yongbin Han
College of Food Science & Technology, Nanjing Agricultural University, Nanjing 210095, PR China

a r t i c l e i n f o
Article history: Received 13 September 2007 Received in revised form 7 April 2008 Accepted 9 April 2008 Keywords: Yeast Se Medium optimization D-optimal mixture design

a b s t r a c t
Natural fermentative medium for selenium-enriched yeast (Saccharomyces cerevisiae) culture was investigated using response surface methodology (RSM). The medium with a concentration of 15 mg/mL Na2SeO3, contained various ratios of juices from germinated brown rice (0.40 w 0.80, 12 Brix), beerwort (0.10 w 0.50, 12 Brix) and soybean sprout (0.10 w 0.50, 12 Brix), which were optimized by applying Doptimal mixture design (DMD). The effects of their ratios on biomass yield and total selenium (Se) yield were analyzed. The results showed that when the mixed ratio of the components was 4:4:2 (v:v:v), the maximum value of biomass yield and total Se yield were 8.5 g/L and 3.53 mg/L, respectively. Verication experimental trials were performed for validating the models, and it indicated that the above mixture of these natural materials can be used as proper medium for the growth of selenium-enriched yeast and accumulation of Se. 2008 Swiss Society of Food Science and Technology. Published by Elsevier Ltd. All rights reserved.

1. Introduction The trace element selenium (Se) is an essential nutrient for the growth of animals and humans. This has become increasingly obvious as new researches have shown a hitherto unsuspected role for this element to human health (Neve, 1998; Zheng & Ouyang, 2001). It is shown that Se has a cancer protective effect (Chen, Zhang, Hou, & Chai, 1999; Clark et al., 1996; Combs, 1997; Klein et al., 2003) and a profound effect on survival of HIV infected patients (Bologna et al., 1994). It is generally believed that the ingestion of organic Se compounds is better and safer than that contained inorganic Se. A variety of selenium-enriched biological derivatives, such as garlic (Yang, Wang, & Li, 1993) and broccoli (Davis, Zeng, & Finley, 2002) have been developed as diet supplements. It is known that some microorganisms (Chasteen & Bentley, 2003), mainly yeast, can utilize soluble sugars and organic acids, producing biomass with high protein content and meanwhile transform inorganic Se (low bioavailability, potentially toxic) into organic form (safer and highly bioactive). The Se supplementation using microorganisms has received much attention in recent years. Previous works have shown that yeast was a good carrier for Se biotransformation (Chanda & Chakrabatri, 1996; Choi et al., 2002). Hence, one of the most economic sources of organic Se is yeast grown in selenium-enriched medium. However, most previous

* Corresponding author. Tel./fax: 86 25 8439 6293. E-mail address: guzx@njau.edu.cn (Z. Gu).

works on Se enrichment of yeast were carried out in biochemical medium and simplex nutrition inhibited growth of the cell. Thus low yield of biomass and conversion rate of inorganic Se were major obstacles to the successful commercialization of this organism in the health food industry (Demirci, Pometto, & Cox, 1999; Suhajda, Hegoczki, Janzso, Pais, & Vereczkey, 2000). Moreover, the proportion of organic Se in cells was still little explored while total Se content was more concerned. Response surface methodology (RSM) is a collection of mathematical and statistical techniques useful for designing experiments, building models and analyzing the effects of several independent factors (Garrido-Vidal, Pizarro, & Gonzalez-Saiz, 2003; Li & Fu, 2005). The main advantage of RSM is the reduced number of experimental trials needed to evaluate multiple factors and their interactions. Also, study of the individual and interactive effects of these factors will be helpful in effort to nd the target value. Hence, RSM provides an effective tool for investigating the aspects affecting the desired response if there are many factors and interactions in the experiment. To optimize the process, RSM can be employed to determine a suitable polynomial equation for describing the response surface. D-optimal mixture design (DMD) is an effective technique of RSM for optimizing complex processes. It is widely used in opti mizing the culture medium (Caroline, Marina, Ricardo, & Hector, 2007). The basic theoretical and fundamental aspects of DMD have been reviewed (Cornell, 1990; Kowalsky, Cornell, & Vining, 2000, 2002). In this design, the total amount of the components is held constant when proportions of the mixture components change. Several approaches have been presented to deal with problems that

0023-6438/$34.00 2008 Swiss Society of Food Science and Technology. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.lwt.2008.04.002

328

H. Yin et al. / LWT - Food Science and Technology 42 (2009) 327331 Table 2 Instrumental operating conditions for HG-AFS Parameters High voltage of PMT (V) Lamp current (mA) Atomizer temperature ( C) Atomizer height (mm) Gas type Carrier gas ow rate (mL/min) Shield gas ow rate (mL/min) Injection volume (mL) Dwell time (s) Read time (s) Read method Measurement method Value 300 80 200 8 Argon 400 900 2 1 10 Peak area Stand curve

involve mixtures (Dingstad, Egelandsdal, & Naes, 2003; Omer & Sukru, 2006), and mixtures of two or more major components (categories or factors), each of which may be a mixture of one or more minor components (Piepel, 1999). In this study, germinated brown rice juice, beerwort and sprout soybean juice were used as natural medium components to produce selenium-enriched yeast. The DMD was applied to optimize the composition of culture medium, nd the most signicant factor affecting biomass yield and total Se yield, and establish a quadratic polynomial regression model, so that to provide scientic basis for industrialization. 2. Materials and methods 2.1. Materials Yeast (Saccharomyces cerevisiae) was obtained from fermentation with a concentration of 15 mg/mL Na2SeO3 by comparing the selenium-enrichment ability of 17 strains, which were chosen by resistance screening with a concentration of 50 mg/mL Na2SeO3. Na2SeO3 purchased from Sigma was of analytical grade. Na2SeO3 standard solution was purchased from Jiangsu center for disease control and prevention. All other chemicals used in this work were also from commercial sources. Germinated brown rice, malt and soybean sprout were bought from the local market. 2.2. Fermentative medium preparation 2.2.1. Preparation of germinated brown rice juice Germinated brown rice was pulverized and then added with deionized water to reach a ratio of 1:4 (w/v). Initial pH value was adjusted to 4.0 by citric acid. The mixture was gelatinized (90  C, 30 min), liqueed (60  C, 1 h) by adding glucoamylase (1.25 g/L, 100,000 U/g) and then homogenized through the colloid mill twice (30 MPa). For further saccharication, the mixture was incubated at 60  C for 2 h by adding the same amount of glucoamylase. Germinated brown rice juice was obtained by centrifugation (1900g, 5 min) and the soluble solid was adjusted to 12 Brix. 2.2.2. Preparation of beerwort Malt was pulverized and four volumes of deionized water were added to the powder. The mixture was kept at 37  C for 30 min, at 50  C for 1 h and at 65  C for 3 h in turn. The saccharied substance was boiled for 5 min and then refrigerated at 4  C for 12 h. Beerwort was obtained by centrifugation (1900g, 5 min) and the soluble solid was adjusted to 12 Brix. 2.2.3. Preparation of soybean sprout juice Soybean sprout was boiled with deionized water at a ratio of 1:4 (w/v) to deactivate its enzymes, and then wet-milled twice (30 MPa). Soybean sprout juice was obtained by centrifugation (1900g, 5 min) and the soluble solid was adjusted to 2 Brix. Several main nutritional characteristics of three different culture components were shown in Table 1.

2.3. Fermentation conditions Yeast cells were pre-grown aerobically for 24 h in PDB (potato, dextrose, broth) nutrient medium. Fermentative medium was inoculated with 10% (v/v) seed liquid and then cultivated for 48 h on a rotary shaker. The starter and fermentative conditions of volume, temperature and shaking speed were xed at 50 mL/250 mL, 28 0.5  C and 160 rpm, respectively. Inorganic Se was added to the sterile medium before the start of yeast cultivation as a solution of Na2SeO3, at a concentration of 10 mg/mL. Initial Na2SeO3 concentrations in medium were 0 mg/mL in control and 15 mg/mL in experimental treatments. 2.4. Determination of total Se, inorganic Se and organic Se yield The yeast cells were obtained by centrifugation (1900g, 10 min). To remove unbound Se, the cells were washed with deionized water thoroughly and then centrifugated at 1900g for 10 min. The cells were dried under vacuum to a constant weight. The Se determination was carried out according to the hydride generation atomic uorescence spectrometry (HG-AFS) method described by Wu, Jin, Shi, and Bi (2007) with some modications. The instrumental operating conditions were given in Table 2. About 0.2 g dried samples were digested (170  C, 0.5 h) with 5 mL of a mixture of concentrated HNO3, HClO4 and H2SO4 (v:v:v 10:4:5) in a digestive ask. Se (6) was reduced to Se (4) by addition of 1 mL concentrated HCl. To avoid the volatilization loss of Se, a reux equipment was employed during digestion process. After digesting, the solution was cooled and made a constant volume with ultra-pure water. The blank was digested in the same way. The digested product was used for total Se determination. For inorganic Se determination, the biomass mixed with ultra-pure water was extracted in boiling bath for 1 h and made a constant volume. Then the mixture was centrifugated at 8300g for 15 min. The supernatant liquor was ltrated and the ltrate could be analyzed directly. Organic Se yield was calculated from the difference between the total Se yield and inorganic Se yield.

Table 1 Main nutritional characteristics of different culture components Nutritional characteristics Total sugar (g/L) Reducing sugar (g/L) Soluble protein (mg/mL) Free amino acid (mg/mL) Germinated brown rice juice 112.5 0.2 105.5 1.8 1646.6 43.7 94.1 3.3 Beerwort 105.0 1.4 70.4 0.1 923.1 21.4 449.3 17.6 Soybean sprout juice 2.6 0.1 2.0 0.1 1472.3 28.6 350.4 42.7

Table 3 Independent variables and their coded and actual levels used for optimization in Doptimal mixture design experiment Variable factors Symbol Proportional level 1 Germinated brown rice juice Beerwort Soybean sprout juice X1 X2 X3 0.80 0.50 0.50 2/3 0.67 0.37 0.37 1/2 0.60 0.30 0.30 1/3 0.53 0.23 0.23 1/6 0.47 0.17 0.17 0 0.40 0.10 0.10

H. Yin et al. / LWT - Food Science and Technology 42 (2009) 327331 Table 4 Observed and predicted values of biomass yield and total Se yield in D-optimal mixture design experiment Trials Level X1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 0.40 0.60 0.40 0.40 0.60 0.80 0.47 0.67 0.53 0.47 0.40 0.60 0.40 0.80 X2 0.50 0.10 0.30 0.10 0.30 0.10 0.37 0.17 0.23 0.17 0.50 0.10 0.10 0.10 X3 0.10 0.30 0.30 0.50 0.10 0.10 0.17 0.17 0.23 0.37 0.10 0.30 0.50 0.10 Biomass yield (g/L) Observed value 8.35 0.02 7.59 0.01 8.49 0.01 7.57 0.01 7.59 0.01 6.19 0.01 8.13 0.02 7.21 0.00 8.09 0.00 7.85 0.02 8.35 0.02 7.59 0.01 7.57 0.01 6.19 0.01 Predicted value 8.34 7.58 8.44 7.56 7.57 6.19 8.24 7.24 7.95 7.98 8.34 7.58 7.56 6.19 Total Se yield (mg/L) Observed value 3.53 0.08 3.08 0.09 3.51 0.11 2.78 0.13 3.28 0.04 2.60 0.09 3.25 0.05 2.84 0.07 3.37 0.06 3.08 0.06 3.53 0.08 3.08 0.09 2.78 0.13 2.60 0.09 Predicted value 3.51 3.07 3.49 2.78 3.22 2.59 3.42 2.98 3.20 3.13 3.51 3.07 2.78 2.59

329

Proportion of organic Se

0.90 0.89 0.91 0.91 0.90 0.92 0.87 0.89 0.93 0.87 0.90 0.89 0.91 0.91

2.5. Experimental design DMD was used to determine the optimum mixture ratio for yeast growth and Se accumulation. The experimental design and statistical analysis were performed using the software of Design Expert version 6.0.10 (Stat-Ease, Inc.). A six-level three-factor design was chosen to evaluate the combined effects of three independent variables, germinated brown rice juice, beerwort and soybean sprout juice, coded as X1, X2 and X3, respectively. Based on the preliminary trials, the proportion of the above three materials was 0.40 X1 0.80, 0.10 X2 0.50, and 0.10 X3 0.50, respectively (Table 3). The response values were yeast biomass yield and total Se yield. The design consisted of 14 combinations including four replicates (trials 11 w 14) (Table 4). The responses function (Y) of the special cubic model was partitioned into linear, interactive and cubic components:

2.6. Statistical analysis All trials were carried out in triplicate and all the data were reported as means SD (standard deviation). The statistics significance was evaluated using Students t-test and P < 0.05 or 0.01 was taken as signicant.

3. Results and discussion 3.1. Analysis of D-optimal mixture experiment tting the models The data shown in Table 4 were converted into two secondorder polynomial equations with three independent variables. Consequently, the polynomial models describing the correlation between responses (biomass yield & total Se yield) and the variables were presented as follows:

n X j1

bj x j

X
h<j

bhj xh xj 8 > < xj ! 0 P j 1; 2; 3; .; n > n 1 xj 1 : j

X
h<j<k

Ybiomass yield 3:99X1 7:89X2 1:80X3 6:82X1 X2 17:13X1 X3 11:42X2 X3 With adj R-squared 0:99; adeq precision 38:87 Ytotal Se yield 1:71X1 3:46X2 0:78X3 2:54X1 X2 8:60X1 X3 6:59X2 X3 With adj R-squared 0:88; adeq precision 12:18

(1)

bhjk xh xj xk ;

(2)

where bj, bhj and bhjk are regression coefcients calculated from the experimental data by multiple regression, Xh, Xj and Xk are levels of the independent variables while n equals to the number of the tested factors (n 3). The analysis of variance (ANOVA) tables were generated. The effect and regression coefcients of individual linear, interactive and cubic terms were determined. The signicances of all terms in the polynomial were judged statistically by computing the F-value at a probability (P) of 0.05 or 0.01. The regression coefcients were then used to make statistical calculations to generate contour maps from the regression models.

where X1, X2 and X3 denoted germinated brown rice juice, beerwort and soybean sprout juice, respectively. The value of adjusted R2, a measurement for tness of the regressed model 1 (Eq. (1)), is 0.99. It indicated that the experimental data were in a good agreement with predicted values. The obtained model 1 is very signicant (P < 0.01) according to ANOVA given in Table 5. t-Test results and Pvalues show the interactions between X1 and X2, X2 and X3, X1 and X3 are signicant at P < 0.05, P < 0.01 and P < 0.01, respectively. Model 2 (Eq. (2)) clearly demonstrated that X1 and X2 had positive effects on total Se yield, while X3 had a positive effect on it.
Table 6 Analysis of variance (ANOVA) for regression model 2 Source Model Linear mixture X1X2 X2X3 X1X3 Residual Cor total Sum of squares Degree of freedom Mean square F-value P-value 1.33 1.07 0.01 0.18 0.07 0.11 1.43 5 2 1 1 1 8 13 0.27 0.53 0.01 0.18 0.07 0.01 19.80 39.59 0.79 13.10 5.35 0.0003 <0.0001 0.3988 0.0068 0.0495

Table 5 Analysis of variance (ANOVA) for regression model 1 Source Model Linear mixture X1X2 X2X3 X1X3 Residual Cor total Sum of squares Degree of freedom Mean square F-value P-value 6.56 5.57 0.08 0.67 0.22 0.06 6.62 5 2 1 1 1 8 13 1.31 2.79 0.08 0.70 0.22 7.56E-003 173.69 <0.0001 368.77 <0.0001 10.17 0.0128 92.20 <0.0001 28.55 0.007

330

H. Yin et al. / LWT - Food Science and Technology 42 (2009) 327331

The absolute value for the coefcient of X2 is larger than those of X1 and X3. This indicated that the linear term inuence of X2 was more signicant than those of X1 and X3. The function was not linear and indicated the importance of interactions among X1, X2 and X3. The coefcients of X1X3 and X2X3 are obviously larger than that of X1X2. This meant that the effects of former two (X1X3, X2X3) combinations were main factors inuencing Se productivity. According to ANOVA shown in Table 6, the obtained model 2 is very signicant (P < 0.01). On the basis of t-test and P-values, the interaction between X2 and X3 is signicant (P < 0.01) as well as that between X1 and X3 (P < 0.05). However, the interaction between X1 and X2 is not signicant (P > 0.05). 3.2. Response surfaces analyses To obtain more detailed interrelations and interactions, contour maps are shown in Figs. 1 and 2 for the effects of the independent variables on biomass yield and total Se yield. The diagrams describe the variation on the responses as a function of the mixture composition. Giving an overall curvilinear effect, biomass yield and total Se yield enhance mainly by an increased ratio of beerwort. A lower quantity of germinated brown rice juice is better for the growth of yeast. Although it was noted that the optimum conditions for accumulation of biomass and total Se were slightly different, the correlation of their observed values (Table 4) is positive and the correlation coefcient is 0.89. Hence, on the basis of medium optimization evaluated from the model 1 and model 2, the predicted largest biomass yield and total Se yield were 8.5 g/L and 3.53 mg/L, when the proportion of germinated brown rice juice, beerwort and soybean sprout juice were 0.4, 0.4 and 0.2, respectively. The optimum medium of mixed juices could provide more nutrition for cells growing and Se producing than a single kind of medium and synthetic medium. Firstly, the medium had ample carbon and nitrogen sources. Secondly, there were plenty of vitamins in the optimum medium which were necessary for the growth
0.10

X1: germinated brown rice juice

0.80

2.75

2.91

3.07 0.10 3.22

3.38 2.91

0.50 X2: beerwort

0.40

0.50 X3: soybean sprout juice

total Se yield(mg/L)
Fig. 2. Contour plot for the effects of germinated brown rice juice (X1), beerwort (X2) and soybean sprout juice (X3) on total Se yield.

X1: germinated brown rice juice

0.80 6.6

7.0

of yeast. Brown rice contains more vitamin E and B than the ordinary polished rice (Kenichi, Keitaro, Yuji, & Takafumi, 2005). Malt extract had various vitamins, such as biotin, pantothenate, thiamine, pyridoxine, riboavin, folic acid, nicotinic acid, etc. Soybean was rich in ascorbic acid and phytic acid (Mohamed & Rangappa, 1992). Moreover, the medium could supply various minerals. The content of Ca, K, P, Mg, Fe, Mn and Zn in germinated brown rice is higher than in polished. It was showed that the minerals were benecial not only to the growth of yeast but also to the synthesis of organic Se in yeast. As Se was a biomass related product, the increase of biomass enhance the activity of the most important selenic enzyme Glutathione Peroxidase (GSH-Px) (Abdel Rahim, 2005; Bayse, Baker, & Ortwine, 2005), which is critical for survival and function of aerobic organisms because it can effectively inhibit oxidation and prolong life of the organism. 3.3. Verication of model Within the scopes of the variables investigated in experimental design, additional experiments with different conditions were carried out in order to assess the validity of model 1 and model 2. The arrangement and results of the conrmatory trials were shown in Table 7. The experimental values were found to be reasonably close to the predicted ones, which conrmed the validity and adequacy of the predicted models. Our observed yield of biomass (9.86 g/L) and total Se (3.86 mg/L) under the optimal conditions was higher than that under all the other conditions. 4. Conclusion

7.3

0.10

7.7

0.10

8.1

7.7

0.50 X2: beerwort

0.40

0.50 X3: soybean sprout juice

biomass yield(g/L)
Fig. 1. Contour plot for the effects of germinated brown rice juice (X1), beerwort (X2) and soybean sprout juice (X3) on biomass yield.

The statistical methodology, D-optimal mixture design is demonstrated to be effective and reliable in nding the optimal ratio of components in fermentative medium for selenium-enriched yeast production. It was showed that when the mixed ratio of germinated brown rice juice, beerwort and soybean sprout juice was 4:4:2, the

H. Yin et al. / LWT - Food Science and Technology 42 (2009) 327331 Table 7 Arrangement and result of conrmatory trials Trials Proportion of components X1 Optimum medium Random medium 1 Random medium 2 Random medium 3 0.4 0.4 1 0 X2 0.4 0.5 0 1 X3 0.2 0.1 0 0 Biomass yield (g/L) Observed value 9.86 0.06a 9.86 0.02a 8.49 0.33b 8.52 0.17b Predicted value 8.49 8.34 Total Se yield (mg/L) Observed value 3.86 0.02a 3.40 0.05b 3.17 0.12c 2.99 0.06d Predicted value 3.89 3.19

331

Proportion of organic Se 0.91 0.90 0.89 0.91

Within the same column, values followed by different letters (a, b, c.) differ signicantly at P < 0.05.

maximum biomass yield and total Se yield were 8.50 g/L and 3.53 mg/L respectively. Conrmatory trials suggested that there be good t degree between the observed values in experiment and the values predicted by equation. All of these illuminated that the method and results were feasible and efcacious. The optimum medium obtained in this paper would be expected to be useful on further development for functional food and pharmaceuticals. References
Abdel Rahim, A. G. (2005). The relationship between whole blood selenium (Se) concentration and the activity of the seleno-enzyme, glutathione peroxidase (GSH-Px, E.C.I.11.1.9) in camel (Camelus dromedarius). Journal of Arid Environments, 62, 359362. Bayse, C. A., Baker, R. A., & Ortwine, K. N. (2005). Relative strengths of Se.N, O interactions: implications for glutathione peroxidase activity. Inorganica Chimica Acta, 358, 38493854. Bologna, R., Indacochea, F., Shor Posner, G., Mantero Atienza, E., Grazziutii, M., Sotomayor, M. C., et al. (1994). Selenium and immunity in HIV-1 infected pediatric patients. Journal of Nutritional Immunology, 3, 4149. Caroline, D., Marina, E., Ricardo, K., & Hector, C. G. (2007). Crossed mixture design and multiple response analysis for developing complex culture medium used in recombinant protein production. Chemometrics and Intelligent Laboratory Systems, 86, 19. Chanda, S., & Chakrabatri, S. (1996). Plant origin liquid waste: a resource for single cell protein production by yeast. Bioresource Technology, 57, 5154. Chasteen, T. G., & Bentley, R. (2003). Biomethylation of selenium and tellurium: microorganisms and plants. Chemical Reviews, 103, 125. Chen, C. Y., Zhang, P. Q., Hou, X. L., & Chai, Z. F. (1999). Subcellular distribution of selenium and Se-containing proteins in human liver. Biochimica et Biophysica Acta, 1427, 205215. Choi, M. H., Ji, G. E., Koh, K. H., Ryu, Y. W., Jo, D. H., & Park, Y. H. (2002). Use of waste Chinese cabbage as a substrate for yeast biomass production. Bioresource Technology, 83, 251253. Clark, L. C., Combs, G. F., Jr., Turnbull, B. W., Slate, E. H., Chalker, D. K., Chow, J., et al. (1996). Effects of selenium supplementation for cancer prevention in patients with carcinoma of the skin: a randomized controlled trial. Journal of the American Medical Association, 276, 19571963. Combs, G. F., Jr. (1997). Selenium as a cancer-protective agent. Bulletin of SeleniumTellurium Development Association 14. Cornell, J. A. (1990). Experiments with mixtures, designs, models and the analysis of mixture data. New York. Davis, C. D., Zeng, H. W., & Finley, J. W. (2002). Selenium-enriched broccoli decreases intestinal tumorigenesis in multiple intestinal neoplasia mice. Journal of Nutrition, 132, 307309.

Demirci, A., Pometto, A. L., & Cox, D. J. (1999). Enhanced organically bound selenium yeast production by fed-batch fermentation. Journal of Agricultural and Food Chemistry, 47, 24962500. Dingstad, G., Egelandsdal, B., & Naes, T. (2003). Modeling methods for crossed mixture experiments a case study from sausage production. Chemometrics and Intelligent Laboratory Systems, 66, 175190. Garrido-Vidal, D., Pizarro, C., & Gonzalez-Saiz, J. M. (2003). Study of process variables in industrial acetic fermentation by a continuous pilot fermentor and response surfaces. Biotechnology Progress, 19, 14681479. Kenichi, O., Keitaro, S., Yuji, Y., & Takafumi, K. (2005). Bio-functional components in the processed pre-germinated brown rice by a twin-screw extruder. Journal of Food Composition and Analysis, 18, 303316. Klein, E. A., Lippman, S. M., Thompson, I. M., Goodman, P. J., Albanes, D., Taylor, P. R., et al. (2003). The selenium and vitamin E cancer prevention trial. World Journal of Urology, 21, 2127. Kowalsky, S. M., Cornell, J. A., & Vining, G. G. (2000). A new model and class of designs for mixture experiments with process variables. Communications in Statistics Theory and Methods, 29, 22552280. Kowalsky, S. M., Cornell, J. A., & Vining, G. G. (2002). Split-plot designs and estimation methods for mixture experiments with process variables. Technometrics, 44, 7279. Li, Q. H., & Fu, C. L. (2005). Application of response surface methodology for extraction optimization of germinant pumpkin seeds protein. Food Chemistry, 92, 701706. Mohamed, A. I., & Rangappa, M. (1992). Nutrient composition and anti-nutritional factors in vegetable soybean: II. Oil, fatty acids, sterols, and lipoxygenase activity. Food Chemistry, 44, 277282. Neve, J. (1998)Metal ions in biology and medicine, Vol. 5. Paris: John Libbey Eurotext. pp. 242247. Omer, Z., & Sukru, K. (2006). Optimization of emulsion characteristics of beef, chicken and turkey meat mixtures in model system using mixture design. Meat Science, 73, 611618. Piepel, G. F. (1999). Modeling methods for mixture-of-mixture experiments applied to a tablet formulation problem. Pharmaceutical Development and Technology, 4, 593606. Suhajda, A., Hegoczki, J., Janzso, B., Pais, I., & Vereczkey, G. (2000). Preparation of selenium yeasts I. Preparation of selenium-enriched Saccharomyces cerevisiae. Journal of Trace Elements in Medicine and Biology, 14, 4347. Wu, H., Jin, Y., Shi, Y. Q., & Bi, S. P. (2007). On-line organoselenium interference removal for inorganic selenium species by ow injection coprecipitation preconcentration coupled with hydride generation atomic uorescence spectrometry. Talanta, 71, 17621768. Yang, M., Wang, K., & Li, B. F. L. (1993). Biological activities and structural predictions of Se-containing protein from garlic. Journal of Inorganic Biochemistry, 51, 422. Zheng, W. J., & Ouyang, Z. (2001). Organic selenium compounds from plants: Their chemistry and applications in medicine. Guangzhou, China: Jinan University Press.