Food Sc'i. Technol Res., 10 (3), 328-3_ 3_ _3, 2004

Technical paper

Effect of Heat Treatment on the Lipid Peroxide Content and Aokusami (Beany Flavor) of Soymilk

Hiroshi END0,1 Masahiro OHN0,2 Katsuo TANJI,3 Shinji SHIMADA4 and Kentaro KANEKO

IFukushima Technology Centre Aizuwakamatsu Technical Support Centre, 88-], Shimoyanagihara, Tsuruga, Ikki-machi, Aizuwakamatsu-shi, Fu.kushima 965-0006 Japan 2Fukushima Technology Centre, 1-12, Machiikedai, Kooriyama-shi, Fukushima 963-0215 Japan -?Fukushima Prefecture Agricultural Experiment Station, 20, 14,akamiyamae, Tomita-machi, Kooriyama-shi, Fukushima 963-8041 Japan 4National [nstitute of Crop Science, 2-/-]8 Kannondai, Tsukuba, Ibaraki 305-8518, Japan 5Faculty ofApplied Life Science, Nippon veterinary and Animal Science University, 2-27-5 Sakai, Musashino-shi, Tokyo 180-8602 Japan

Received February 23, 2004; Accepted June 4. 2004.

With the objective of minimizing aokusami (beany flavor), which is an undesirable aspect of soymilk flavor, we studied the effects of heating on lipid peroxide content, a factor contributing to the beany flavor. We also prepared various containing soymilk using this heating process, and evaluated the effects of heating by sensory test. The lipid peroxide content of soymilk prepared from soaked and swollen steamed at 75'C in a relative humidity of 80-90% for 10 minutes was substantially lower than that in soymilk prepared from soybeans which lacked lipoxygenase and was 14% or less that of non-heat-treated soybeans. Additionally, the lipid peroxide content of soymilk prepared by blanching soaked and swollen soybeans in boiling water for 30 seconds was comparable to that found in soymilk prepared from lipoxygenase-lacking soybeans. The beany flavor of custard , Bavarian cream, and Annin prepared with soymilk obtained from heat-treated soybeans was significantly improved.

Keywords: soy bean, soymilk, Iipid peroxide, beany flavor.

Soybeans have long played a pivotal role in the food blanching method in hot water at 99.3'C (Seth & Nath, culture of Japanese people as a foodstuff with a high nutri- 1 988); and the microwave heating method (Wang & Toledo, tive value. Recent research has shown that protein 1987). However, all these methods share the problems of has a cholesterol-lowering effect (Anderson et al., 1995), heating-induced insolubilization of protein and an associ- that soybean saponin has anticancer activity (Kennedy, ated decrease in protein extraction rate (Ediriweera et al., 1995), and that soybean isoflavones have inhibitory effects 1 987). Omura and Takechi ( 1990) proposed heating soaked on breast cancer and prostate cancer (Peterson & Bernes, and swollen, peeled soybeans in hot water at 70'C for 5 l 99 1 ; Peterson G & Bernes S, 1 993) as wel] as a preven- min. However, because the peeling process is associated tative effect on osteoporosis (Tsuchida et al., 1999). Given with detachment of the hypocotyls (Endo et al., 2003a), the above, processed foods made from soybeans are valued this process is unsuitable for the preparation of raw mate- as a source of isoflavones. Soybeans are used in numerous rials for high-isoflavone soymilk. foods, including tofu, natto, miso, soy sauce, and soymilk. The present research was performed with the objective Soymilk is consumed as a beverage, and is also used as an of minimizing aokusami (beany flavor), which is an unde- ingredient in the making of jellies, custard , and sirable aspect of the flavor of soymilk and other soymilk- other desserts. However, the distinctive aokusami (beany related drinks. We studied the effects of heating on lipid flavor) produced by lipoxygenase has a major impact on peroxide content, which contributes to the beany flavor, and consumer preference. Minimizing the beany flavor is there- we prepared desserts made with soymilk processed in this fore a key challenge in widening the use of soymilk. manner to investigate its effects on flavor. Several methods have been proposed for deactivating lipoxygenase: the hot water-added grinding method (Echigo Materials and Methods et al., 1991 ), in which soybeans are soaked in hot water at Test soybeans The following soybean varieties tested 70'C and then homogenized in hot water at 95'C; the were grown, screened, and dried at the Fukushima Prefecture Agricultural Experiment Station: Suzuyutaka, Tohoku 126 (high-isoflavone soybean, Endo et al., 2003b), Tohoku 1 35

E- mail : [email protected], jp (lipoxygenase-lacking soybean; hereafter, Iipoxygenase- Effect of Heat Treatment on the Lipid Peroxide Content and Aokusami of Soymilk 329

lacking soybean), Tohoku 139 (low-allergen soybean strain; The Suzuyutaka variety of soybeans was used for this test. hereafter, Iow-allergen soybean), Shinanokuro (black Soybeans that had been soaked and swollen for 16 h at 20*C soybean), Aomame 4 (blue soybean), and commercially- were blanched in boiling water for l0-40 s, and soymilk available lchihime (Tochigi Prefecture-origin; Iipoxyge- was then prepared using the heat-squeezing method with a nase-lacking soybean). 6-fold volume of water added. Soybean heat-treatment method and soymilk prepara- Measurement of lipid peroxide content To 0.5 g of tion Suzuyutaka was used as the raw material. 32-mesh or finer soybean powder ground in a compact Control group: Soybeans were soaked and swollen in milling machine (AS One Corporation; Wonder Blender water at 20'C for 16 h, and soymilk was then prepared using WB- I ), soymilk, or each of the soymilk desserts was added the heat-squeezing method with the addition of water (6: I , 10 mL of distilled water, and the mixture was ice-cooled water weight: soybean weight) (Endo et al., 2003a, simi- while being homogenized for I min using a Polytron homog- larly hereafter). enizer. The homogenate was allowed to stand for 60 min Test- I : Soybeans were air-dried for 10 min at 75'C (DK- at 4'C, after which the lipid peroxide content was assayed 600T incubator; Hitachi Corp.), rinsed with water, and by the DETBA (1,3-diethyl-2-thiobarbituric acid) method soymilk was then prepared by the same method as that (Suda et al., 1994). described for the control group. Preparation and sensory evaluation of soymilk desserts Test-2: Soybeans were steamed for 10 min at 75'C in a Desserts were made using soymilk prepared from Tohoku relative humidity of 80-90% (MTH-4400 constant tem- 126 soybeans, and a sensory evaluation of the beany flavor perature and humidity chamber; Sanyo) and rinsed with was undertaken. Using the heat-squeezing method with a water. Soymilk was then prepared by the same method as 6-fold volume of water added, soymilk was prepared from that described for the control group. soybeans blanched in boiling water (1000C) for 30 s after Test-3: Soybeans that had been soaked and swollen in being soaked and swollen in water at 20'C for 16 h. Desserts water for 16 h at 20'C were steamed for 10 min at 75'C in were prepared using standard recipes, with milk partly a relative humidity of 80-90% and rinsed with water, after replaced with soymilk (Table 1). Pudding was made using which soymilk was prepared using the heat-squeezing pudding mix (Nisshin Foods), Bavarian cream using method with the addition of water (6:1, water weight Bavarian cream mix (Nisshin Foods), and using soybean weight). Annin tofu mix (Heichinsei) (Table 1). Test-4: Soybeans were heated for 60 min in hot water at Sensory evaluation was performed using a five-point 750C and rinsed with water, after which soymilk was rating scale (1: Strong; 2: Moderately strong; 3: Weak; 4: prepared using the heat-squeezing method with the addition Almost none; 5: None whatsoever) to score the beany flavor of water (6: 1, water weight: soybean weight). of soymilk (Furukawa, 1994). Panelists comprised 15 staff Test-5: Soybeans that had been soaked and swollen for members from the Aizu-Wakamatsu Technical Support 16 h in water at 20'C were blanched for 2 min in hot water Center (ages 25=56 years; 12 men and 3 women). Sensory (10:1, water weight: soybean weight) at 750C, and then evaluation was repeated twice and the mean scores were rinsed with water, after which soymilk was prepared using determined. Differences between samples were tested for the heat-squeezing method with the addition of water (6: I , significance using two-way ANOVA (Furukawa, 1994). water weight : soybean weight). Measurement of extraction rates ofprotein and solid Test-6: Soymilk was prepared using the same method as matter The extraction rates of protein and solid matter in that described for the control group, except that the soybeans soymilk were measured according to a previously described were homogenized with hot water at 95'C. method (Endo et al., 2003a). Blanching in boiling water and preparation of soymilk

Table 1. Material composition of various desserts prepared by mixing the soymilk. Soymilk and Milk (g)

Soymilk Mix*1(g) Yolk (g) Water (g) Soymilk'2 Processed soynulk 3 Milk

Custard pudding 1) 400 2) 60 200 200 3) 200 200

Bavarian cream 1) 400 2) 70 18 200 200 3) 200 200

Annin tofu 1) 300 2) 84 300 150 l 50 3) 150 150

* I : Mix is an essential factor in the making of each dessert which is composed of sugar, polysaccharides and other ingredients. *2: Soymilk was prepared by the heat squeezing method in which soybean seeds were homogenized by adding 6.0 times more water to them after soaking in water for 16 hrs. *3: The soybean seeds were heated for 30 sec in 100'C hot water after soaking in water and then prepared in soymilk the same way as *2. 330 H. ENDO et al Results and Discussion proportional to that of lipid peroxide (Furuta et al., 1996), Lipidperoxide content ofsoybeans and soymilk Table and that the beany flavor of soymilk prepared from all 2 presents data on the lipid peroxide content of 7 varieties lipoxygenase-lacking soybeans is less than that of soymilk of soybean and soymilk. made from normal soybeans (Torres-Penaranda et al., 1998). The lipid peroxide content of normal soybeans (soybeans Given these results, it was concluded that a soymilk with with lipoxygenase activity) ranged from 63.5 to 83.6 nmol, minimal beany flavor can be prepared if the lipid peroxide with a mean value of 76.2 nmol. The soybean variety with content is approximately 5 nmol/g. the lowest content was Shinanokuro and that with the Heat treatment and lipid peroxide content Table 3 shows data on the contents of lipid peroxide, protein, solid Table 2. Content of Lipid peroxide in soybean powders and soymilks matter, and soymilk yield in soymilk prepared from heat- prepared from seven soybean varieties. treated soybeans.

Soybean seed Variety Lipid peroxide content (n mol/g) i. Lipid peroxide content The lipid peroxide content per gram of soymilk in the Seed powder*1 Soymilk 2 control group was 21.3 n mol. The lipid peroxide content Normal seed Suzuyutaka 83.6 21.3 of heat-treated soybeans was appreciably lower, and differed Tohoku 126 77,3 20.9 according to treatment method. Soymilks with the lowest Tohoku 139 75.5 14.4 lipid peroxide content were those prepared from soybeans Shinanokuro 63.5 14.7 Aomame 4 80.9 20.5 that had been steamed for 10 min at 75'C in a relative Average 76.2 18.4 humidity of 80-90(~o after being soaked and swollen in water (Test-3, 2.8 n mol), and from soybeans that had been heated Soybean seeds of Tohoku 135 23.6 4. 1 lipoxygenase lchihime 29.8 5.3 for 2 min in hot water at 75'C after being soaked and lacking variety Average 26.7 4.7 swollen in water (Test-5, 3.3 n mol). The lipid peroxide content in both groups was appreciably lower than that in * I : Seed powder was prepared from several soybeans by crushing with a grinder. soymilk prepared from lipoxygenase-1acking soybeans *2: Soymilk was prepared by the heat-squeezing method in which soybean (Table 2, mean content 4.7 n mol). Soymilk with the next seeds were homogenized by adding 6 times additional water to the beans. lowest lipid peroxide content was that prepared from soybeans that had been soaked for 60 min in hot water at 75'C without prior soaking and swelling (Test-4, 8.0 n highest content was Suzuyutaka. For lipoxygenase-1acking mol), and soybeans that had been air-dried for 10 min at varieties, there was very little difference in lipid peroxide 75'C then soaked and swollen (Test-1 , 8.3 n mol). Soymilk content; 23.6 nmol and 29.8 nmol for Tohoku 135 and with the highest lipid peroxide content was prepared from lchihime, respectively. The mean content for these varieties soybeans that had been homogenized with hot water at was 26.7 nmol, approximately 35% of the content measured 95'C after being soaked and swollen (Test-6, 13.1 n mol), in normal soybeans. and soybeans that were soaked and swollen after being The lipid peroxide content of soymilk showed the dif- steamed for 10 min at 75'C in a relative humidity of 80-90% ferences of a similar trend as for unprocessed soybeans. (Test-2, 10.5 n mol). These findings showed that the lipid The lipid peroxide content of soymilk produced from normal peroxide content of soybeans is reduced by heat treatment soybeans ranged from 14.4 to 2 1 .3 nmol, with a mean value at 75'C or 95'C. Because the activity of L2, the factor most of 18.4 nmol. The mean lipid peroxide content of soymilk responsible for producing the unpleasant odor of lipoxy- produced from lipoxygenase-1acking soybeans was 4.7 genase, is eliminated by soaking and swelling soybeans in nmol, about 26% of that of normal soybeans. These results hot water at 70'C (Matoba et al., 1985), it appears that heat are almost identical to those obtained in previous research treatment at 75'C or 95'C in this experiment deactivated (Furuta et al., 1996). or inhibited this enzyme. In Test-3 and Test-5, which yielded It has been reported that the content of n-hexanal, the soymilk with the lowest lipid peroxide content, the soybeans major constituent of the beany flavor of soymilk, is directly were heated after being soaked and swollen in water, in

Table 3. Contents in lipid peroxide, protein and solid matter of soymilk prepared from soybean anticipated by various heating treatments.

Experimental Lipid peroxide Protein Solid matter Amount of soymilk division (n mol/g) Content Extractive rate Content Extractive rate (g) (~1~.) (g/100g) (%) (g/100g) (%) Control 21.3 5 60 100.0 5.3 78.5 10.5 67. 1 Test- 1 8.3 547 97.7 5.0 72.4 9.9 61.8 Test-2 10.5 553 98.8 4.9 7 1 .7 9.4 59.3 Test-3 2.8 546 97.5 4.9 70.8 10.0 62.3 Test-4 8.0 407 72.7 4.3 46.3 7.9 36.7 Test-5 3.3 500 89.3 4.8 69.8 9.8 60.4 Test-6 13.1 392 70.0 3.9 40.4 8.5 38.0 Various soymilks were prepared by heat squeezing method described in Table I in which the beans were heated in several ways. Effect of Heat Treatment on the Lipid Peroxide Content and Aokusami of Soymilk 331

Test-1 and Test-2, which yielded soymilk with higher lipid humid conditions, it is possible to achieve practical appli- peroxide content, soybeans were soaked and swollen after cation with steam sterilizers or tunnel-type steam heating being heated, and in Test-4, the swelling process was under- systems. taken during heating. It thus appears that heating soaked Blanching in boiling water and lipid peroxide content and swollen soybeans allowed the heat to be more readily Soymilk was prepared after blanching soaked and swollen disseminated, thereby more readily deactivating or inhibit- Suzuyutaka soybeans for O~LO s in boiling water (98'C), ing lipoxygenase. In addition, because the soybeans were followed by cooling in water, and the lipid peroxide content homogenized in hot water after being soaked and swollen and protein and solid matter extractive rates were deter- in Test-6, which yielded soymilk with the highest lipid mined (Figure I ). peroxide content, it appears that enzyme activity is elevated The lipid peroxide content in I g of soymilk prepared by contact with air at an appropriate temperature when without blanching in boiling water was 2 1 .O n mol, but when grinding. blanched in boiling water, the content decreased steadily On the basis of the above findings, it appears that the up to 20 s. The lipid peroxide content in soymilk prepared beany flavor of soymilk prepared from soybeans that were from soybeans blanched for 30 s decreased sharply to about either blanched for 2 min in hot water at 75'C after soaking 23~;~¥ of that for unprocessed soybeans. However, there and swelling in water or steamed for 10 min at 75'C in a appeared to be no further changes. These findings suggest relative humidity of 80-90%, was equivalent to or lower that the lipoxygenase enzymes in soybeans are inhibited or than that of soymilk prepared from lipoxygenase-lacking deactivated when blanched in boiling water after being soybeans. ii. Soymilk yield, and protein and solid matter content 25 90 Heat treatment produced an appreciable decrease in -1H Lipid peroxide -~- Extractive rate of protein soymilk yield. In particular, soymilk yield from soybeans HIH Extractive rate of solid matter homogenized in hot water after soaking and swelling (Test- 20 6) was extremely low at 70.0% of that for the control group, 80 and that from soybeans homogenized after being heated in E hot water at 75'C for 60 min was 72.7% (Test-4). Soymilk ~ )~:oo ~E 15 yield from soybeans heated for 2 min in hot water at 75'C ~ c ~o after being soaked and swollen (Test-5) was somewhat low ~ ~IDo 70 (D .;~ at 89.3%. The yields for other soybeans were slightly :i=~ 9 o reduced, ranging between 97.5% and 98.8%. Because the o:~ 10 ~ ~>< protein content (40.4% and 46.3%, respectively) and solid 12 Lu c~ matter extraction rate (38.0% and 36.7%, respectively) for :i 60 Test-6 and Test-4, for which soymilk yield was extremely 5 10w, were also substantially lower, it appears that protein denaturation and associated insolubilization occurred, thus leading to decreased yield. The yield for Test-5 may have O 50 been reduced by similar causes. For Test-3, in which lipoxy- o 40 genase activity was inhibited to the greatest extent, the 1 OSoaking 20 time (sec) 30 soymilk yield, protein content, and solid matter content were largely the same as those in the control group, but for Test- Fig. 1. Changes of lipid peroxide content and extractive rate of protein and solid matter during soaking treatment in boilling water. 5, there was a tendency for the values to be somewhat lower. Accordingly, for soymilk prepared from soaked and swollen soybeans steamed for 10 min at 75'C in a relative humidity soaked and swollen in water. The extraction rates of protein of 80-90%, Iipoxygenase action is substantially inhibited, and solid matter decreased in almost the same way as lipid and the beany flavor is minimized. peroxide content. The extraction rate of protein was 78.7% Asano et al., ( 1989) discovered that among lipoxygenase for non-heated soybeans, but 75. 1% after blanching for 20 isozymes, the activity of L2 (Matoba et al., 1985), the most s and 70.2% after blanching for 30 s. Similarly, the extrac- important cause of unpleasant odors, is almost completely tive rate of solid matter was 67.5% for non-blanched eliminated by soaking and swelling soybeans in hot water soybeans, and 64.8% and 61.0(~o after blanching for 20 s at 70'C. Omura and Takechi ( 1990) reported that heating and 30 s, respectively. soaked and swollen peeled soybeans in hot water at 70'C Based on the above results, it can be inferred that if normal for 5 min suppresses the production of soybean odor, and soybeans are blanched for 30 s in boiling water after being the loss of soymilk solid content is repressed by about 10% . soaked and swollen, the lipid peroxide content can be The results of the present study are consistent with these reduced without markedly diminishing the content of protein reports. These hot water soaking methods can be readily or solid matter, thus reducing the beany flavor. implemented in small and medium-sized factories because The lipid peroxide content of soymilk prepared from they do not require specialized equipment. Because the soybeans blanched in boiling water for 30 s was approxi- method in the present study involved heat treatment under mately the same as that for soymilk prepared from all 332 H. ENDO et al lipoxygenase-lacking soybeans (4.7 n mol, Table 2), but soymilk can be reduced by blanching soaked and swollen greater than that obtained from soybeans steamed at 75'C soybeans for 30 s in boiling water. in a relative humidity of 80-90% (2.8 nmol, Table 3). However, the protein extractive rate (70.2%) and solid matter extractive rate (61 .O%) were almost identical (Table Ref erences 3). The protein extractive rate was 78.5% and the solid Anderson, J.W., Johnston, B.M. and Cook-Newell, M.E. (1995). Meta- Analysis of the Effects of Soy Protein Intake on Serum Lipids, New matter extractive rate was 67. 1% for soymilk produced by Engl. J. Med., 333, 276-282. the normal method (Table 3), thus confirming that steaming Asano, M., Okubo, K. and Yamauchi, F. (1989). Effect of immersing reduced the protein extractive rate in soymilk by about 8% temperature on the behavior of exuding components from soybean. and the solid matter extractive rate by about 5-6%. Nippon Shokuhin Kogyo Gakkaishi, 36, 636-642 (in Japanese). It is therefore evident that both these processing methods Echigo, T., Takenaka, T., Yatsunami, K. and Yamamoto, H. (1991). Effect of heating treatment in immersing and grinding process on allow the lipid peroxide content of normal soybeans to be the quality of soymilk. Bull. Fac. Agric., Tamagawa Univ., 31, 17-28 reduced to approximately the same level or lower than that (in Japanese). obtained for lipoxygenase-1acking soybeans. However, we Ediriweera, N., Akiyama, Y. and Saio, K. (1987). Inactivation of also determined that the extraction rates of protein and lipoxygenase in soybeans with retention of protein solubility. J. Food solid matter were reduced slightly. Taking the lipid peroxide Sci., 52, 685-690. Endo, H., Ohno, M., Tanji, K., Shimada, S. and Kaneko, K. (2003a). content into account, steaming soaked and swollen soybeans Effect of variety of soybean seeds and processing of soy milk on for 10 min at 75'C in a relative humidity of 80-90% is a the isoflavone content of tofu. Food Preser Sci., 29, 165-172 (in superior method to simply blanching in boiling water. Ja panese). One report notes that when soaked and swollen peeled Endo, H., Ohno, M., Tanji, K., Shimada, S. and Kaneko, K. (2003b). soybeans are heated for 5 min in hot water at 70'C, the loss Isoflavone contents and processing suitability for tofu of the soybeans differing in the variety and planting place. Food Preser of soymilk solid matter is inhibited to about 10% (Omura Sci., 29, (4) 221-228 (in Japanese). & Takechi, 1990), but the result in the present research Furukawa, H. (1994). Scoring method. In "Oishisa o hakaru," Saiwai revealed a slightly lower value. Because this difference may syobo, Tokyo, pp. 29-49 (in Japanese). be attributable to the heating method and duration of heating, Furuta, S., Nishiba, Y., Hajika, M., Igita, K. and Suda, I. (1996). it may be possible to boost the extraction rate by adjusting DETBA value and hexanal production with the combination of unsaturated fatty acids and extracts prepared from soybean seeds the processing conditions. lacking two or three lipoxygenase isozymes. J. Agric. Food Chem., Beanyflavor ofdesserts containing soy'milk A sensory 44, 236-239. evaluation was conducted to assess the beany flavor of Kennedy, A.R. (1995). The evidence for soybean products as cancer custard pudding, Bavarian cream, and Annin tofu made with preventive agents, J. Nutr., 125, 733-743. soymilk (Table 4). The beany flavor of the desserts con- Matoba, T., Hidaka, H., Narita, H., Kitamura, K., Kaizuma, N, and Kito, M. (1985). Lipoxygenase-2 isozyme is responsible for gen- taining soymilk were largely rated as slightly Strong eration of n-hexanal in soybean homogenate. J. Agric. Food Chem., (2. 1 3-2.67). The rating of beany flavor of desserts prepared 33, 852-855. using soymilk prepared from heat-treated soybeans Omura, Y. and Takechi, H. (1990). Effect of hot water treatment on improved significantly to Weak-Almost none (3.47-3.80). flavor of soymilk. Nippon Shokuhin Kogyo Gakkaishi, 37, 278-280 The above results demonstrated that the beany flavor of (in Japanese).

Table 4. Sensory evaluation of beany flavour in various desserts 'l prepared by mixing with soymilk.

Ratio of milk to soymilk*= Soymilk dessert Soymilk O% Processed soymilk 50% Soymilk 50% Milk 100% Milk 50(~o Milk 50%

4--*-H) Custard pudding 4.80 ~-*3.47 2.13

Bavarian cream 4,67 +-*3.80 2.33

~--*-H~ ~---* Annin tofu 4.73 3.67 2.67

*1 and *2: See Table 1 Sensory test was performed by scoring methed (1 - 5) using 15 panelists. *: Significant at p<0.05 Effect of Heat Treatment on the Lipid Peroxide Content and Aokusami of Soyrnilk 333

Peterson, G, and Bernes, S. (1993). Genistein and biochanin A inhibit as an index of lipid peroxidation in plant materials. Biosci. Biotech. the growth of human prostate cancer cells but not epidermal growth Biochem., 58, 14-17. factor receptor tyrosine autophosphorylation, Prostate., 22, 335-345. Torres-Penaranda, A.V., Reitmeier, C.A., Wilson, L.A., Fehr, W.R. Peterson, G. and Bernes, S. ( 1 99 1). Genistein inhibition of the growth and Narvel, J.M. (1998). Sensory characteristics of soymilk and of human breast cancer cells: Independence from estrogen recep- tofu made from lipoxygenase-free and normal soybeans. J. Food tors and the multi-drug resistance Gene, Biochem. Biophys. Res. Sci., 63, 1084-1087. Commun., 179, 661-667. Tsuchida, K., Mizushima, S., Toba, M, and Soda. K. (1999). Dietary Seth, K.K. and Nath, N. (1988). A simple blanch-dry process for lipoxy- soybeans intake and bone mineral density among 995 middle-aged genase inactivation in soybean cotyledons. Int. J. Food Sci. Technol., women in Yokohama, J. Epidemiol., 9, 14-19. 23, 275-279. Wang, S.H, and Toledo, M.C.F. ( 1 987). Inactivation of soybean lipoxy- Suda. I., Furuta, S, and Nishiba, Y. (1994). Fluorometric determina- genase by microwave Heating: Effect of moisture content and tion of a I ,3-diethyl-2-thiobarbituric acid - malondialdehide adduct exposure time. J. Food Sci., 52, 1344-1347.