Screening Test for Rapid Food Safety Evaluation by Menadione-Catalysed

Food Chemistry 138 (2013) 2146–2151

Contents lists available at SciVerse ScienceDirect

Food Chemistry

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

Analytical Methods Screening test for rapid food safety evaluation by menadione-catalysed chemiluminescent assay ⇑ Shiro Yamashoji , Naoko Yoshikawa, Masayuki Kirihara, Toshihiro Tsuneyoshi

Shizuoka Institute of Science and Technology, 2200-2 Toyosawa, Fukuroi, Shizuoka 437-8555, Japan article info abstract

Article history: The chemiluminescent assay of menadione-catalysed H2O2 production by living mammalian cells was Received 2 November 2012 proposed to be useful for rapid food safety evaluation. The tested foods were extracted with water, eth- Received in revised form 10 December 2012 anol and dimethylsulfoxide, and each extract was incubated with NIH3T3, Neuro-2a and HepG2 cells for Accepted 13 December 2012 4 h. Menadione-catalysed H O production by living mammalian cells exposed to each extract was deter- Available online 29 December 2012 2 2 mined by the chemiluminescent assay requiring only 10 min, and the viability of the cells was estimated

as percentage based on H2O2 production by intact cells. In this study the cytotoxicity of food was rated in Keywords: order of inhibitory effect on H O production by intact cells. The well known natural toxins such as Fusar- Menadione 2 2 ium mycotoxin, tomato toxin tomatine, potato toxin solanine and marine toxins terodotoxin and breve- Chemiluminescent assay Viable cell number toxin could be detected by the above chemiluminescent assay. Food safety evaluation Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction nescent assay (Yamashoji, Yoshikawa, Kirihara, & Tsuneyoshi, 2012) was applicable to the determination of the cytotoxicity of Food safety evaluation has traditionally relied on toxicological (1) grain foods, (2) nuts and seeds, (3) tubers, (4) sugar, (5) confec- data that have been obtained through animal experiments (Hug- tionery, (6) fat and oil, (7) bean, (8) fruits, (9) green and yellow veg- get, Schilter, Roberfroid, Antignac, & Koeman, 1996). Single sub- etables, (10) other vegetables, (11) mushroom, (12) seaweed, (13) stances, such as food additives, show clear toxicological data, but drink and seasoning, (14) fish and shellfish, (15) meat, egg and the toxicity of whole foods is difficult to measure because the var- milk, and (16) others (spice). We discuss the application of the ious components may induce additive, synergistic or antagonistic chemiluminescent assay to food safety evaluation and the detec- effects (Konemann & Pieters, 1996). Chemical analysis is a conven- tion of well-known natural toxins. This paper is a first study to tional method for quality control of foods but is not applicable to demonstrate the cytotoxicity of the above different foods by using the evaluation of food safety. Even if chemical analysis is con- a screening test with mammalian cells. ducted for food safety, chemical analysis may require a long time to identify the toxic compounds and may miss causative agents, 2. Materials and methods which are not the target of analysis. Cytotoxicity testing cannot identify toxic compounds, but is ex- 2.1. Organisms and growth conditions pected to detect the presence of harmful compounds in foods. Con- ventional cytotoxicity testing has been conducted by colorimetric Contact-inhibited NIH/3T3 Swiss mouse embryo cells, ATCC CRL WST-1 (Ishiyama et al., 1995) or MTT (Mosmann, 1983) assay, 1658 (NIH/3T3 cells) and HepG2 cells were cultured in DMEM sup- and these assays take a few hours to show enough absorbance. plemented with 10% foetal bovine serum. Neuro-2a cells were cul- However, these traditional assays have been little used as screen- tured in RPMI1640 supplemented with 10% foetal bovine serum. ing test for safety evaluation of different foods other than essential The cells were grown in a Falcon tissue flask in a humidified atmo- oils (Huang & Luo, 2009). sphere of 95% air and 5% CO2 at 37 °C. Exponentially growing stock We demonstrated the usefulness of the traditional chemilumi- cells were trypsinised and diluted with the above culture medium nescent assay (Yamashoji & Isshiki, 1998) for the detection of to adjust 400,000 cells/ml. One hundred microlitres of the diluted various food additives and natural toxins in culture medium, but cells were added to each well on a 96-well plate, and cultivated we had no result for food safety evaluation of different foods. In for 18 h, and each well was used for cytotoxicity testing. this study, we demonstrated that the latest developed chemilumi- 2.2. Determination of living cell number by chemiluminescent assay

⇑ Corresponding author. Fax: +81 538 45 0110. Menadione-catalysed H2O2 production by living mammalian E-mail address: [email protected] (S. Yamashoji). cells was determined by chemiluminescent assay (Yamashoji

0308-8146/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.foodchem.2012.12.037 S. Yamashoji et al. / Food Chemistry 138 (2013) 2146–2151 2147 et al., 2012) and the viability of cells exposed to the extract was Hungerford, & Wekell, 1993). The viability of the treated cells estimated on the basis of the production of H2O2 by cells exposed was determined by chemiluminescent assay. to extract solvent. Mammalian cells were incubated with food extract in a humid- 2.7. Foods and chemicals ified atmosphere of 95% air and 5% CO2 at 37 °C for 4 h, and the cells in each well were washed two times with 100 ll of MEM Foods were purchased from a local supermarket. Wakame, hijiki without phenol red (Medium A), and 100 ll of Medium A were and kelp were not crude, and dried after a wash with water. Chem- added to each well. After that 100 ll of Medium A containing icals were purchased from Sigma and Wako Pure Chemical 0.2 mM menadione were added to each well, and the cells were Industries. incubated for 10 min in a humidified atmosphere of 95% air and 5% CO2 at 37 °C. After the incubation 150 ll of chemiluminescent 3. Results reagent containing 50 mg of TCPO [bis(2,4,6-trichlorophenyl)oxa- late] and 0.14 mg of rhodamine B in 50 ml of acetonitrile were 3.1. Principle of chemiluminescent assay added to each well, and the chemiluminescence intensity was counted for 5 s. The principle of chemiluminescent assay is shown in Fig. 1 (Yamashoji, Nishimoto, Usuda, Kubota, & Isshiki, 1992). Living 2.3. Cytotoxicity test of foods mammalian cells are expected to reduce extracellular menadione by the combination of NAD(P)H and NAD(P)H:quinone reductase, Foods were homogenised with distilled water, ethanol or DMSO and the resulting menadiol may reduce dissolved oxygen to pro- at the ratio of 1–4 (weight/g to volume/ml), and liquid foods were duce H2O2.H2O2 can be specifically determined by chemilumines- mixed with culture medium at the ratio of 1–4 (v/v). After these cent assay with TCPO and fluorescent material, and the mixtures were shaken at 50 °C for 5 h, a supernatant was obtained. concentration of H2O2 is proportional to living cell number. The as- Two microlitres of the supernatant were mixed with 98 ll of cell say with 96-well micro plate was conducted within 15 min. culture medium, and incubated for 4 h under the culture conditions. As negative control water was mixed with extracting solvent 3.2. Extracting solvent at the ratio of 1–4 (v/v), and 2 ll of the mixture were mixed with 98 ll of cell culture medium, and incubated for 4 h under the cul- Hydrophobic solvent such as hexane and benzene was not used ture conditions. After incubation the cells in each well were as extracting solvent because hydrophobic solvent is separated washed three times with Medium A in order to remove interfering from culture medium and cells can be little exposed to the ex- substances or H2O2 produced by autoxidation of foods, and the via- tracted compounds. In this test water, ethanol and DMSO were ble cell number was estimated by chemiluminescent assay. The used as extracting solvent to make provide complete mixing of ex- viability was estimated by the following equation: tract and culture medium. In the case of the extraction with DMSO, the mixture was completely dispersed during 4-h incubation even Viability ð%Þ¼ðViable cell number of the cells exposed to if turbidity was observed. The final volume percent of the homog- food extractÞ=ðViable cell number of the cells exposed to enate was 2% in culture medium, in order to minimise the cytotox- extracting solventÞ100: icity of solvent. The viability of cells exposed to the extract from When it was difficult to get 2 ll from the high-viscosity mix- food with extracting solvent was calculated on the basis of the pro- ture, 200 ll of the mixture were mixed with 9800 ll of culture duction of H2O2 by the cells exposed to extracting solvent alone as medium, and 100 ll were added to each well. The final volume described in Section 2. percent of the tested foods was 0.4% in the culture medium, and the toxic compounds in foods were estimated to be diluted 250 3.3. Cytotoxic effects of salt, acetate and ethanol included in foods times in the culture medium. Incubation time of food extract with cells was adjusted to 4 h Salt, acetate and ethanol are contained in various processed because cytotoxic effect of brevetoxin, tomatine, isobutyl p- foods, and we should know the cytotoxicity of these compounds hydroxybenzoate and dibutylhydroxytoluene has been quantita- because these cytotoxic effects increase in a dose-dependent man- tively determined for 4 h-incubation (Yamashoji & Isshiki, 1998). ner (Yamashoji & Isshiki, 1997). As sodium chloride and acetate showed cytotoxic effects on cells at above 0.1%, we have to pay 2.4. Cytotoxicity test of Fusarium mycotoxins

NIH3T3cells (4 103cells/100 ll/well) were cultivated in the presence of nivalenol (NV), deoxynivalenol (DON) or fusarenon-X (FX) for 48 h. After washing cells with Medium A, the viability was determined by chemiluminescent assay.

2.5. Cytotoxicity test of solanine and tomatine

Solanine and tomatine were incubated with Neuro-2a and NIH/ 3T3 cells, respectively, for 4 h under culture conditions. After washing cells with Medium A, viability was determined by chemiluminescent assay.

2.6. Cytotoxicity test of marine toxins

Cytotoxicity testing for tetrodotoxin and brevetoxin were Fig. 1. Proposed mechanism for menadione-catalysed H2O2 production by mam- conducted under the conditions proposed by (Manger, Leja, Lee, malian cells. 2148 S. Yamashoji et al. / Food Chemistry 138 (2013) 2146–2151

Table 1 Table 2 Cytotoxic effect of various foods having viability of more than 95% in three different Cytotoxic effect of various foods having viability from 50% to 95% in three different fractions extracted with water, ethanol and DMSO. fractions extracted with water, ethanol and DMSO.

Food group Name Food group Name (1) Grain foods Rice, wheat meal, corn (A) Fraction extracted with water (2) Nuts and seeds Peanut, walnut, chestnut (2) Nuts and seeds Tapioca (B) (3) Tubers Sweet potato, Chinese yam, potato, devil’s tongue (8) Fruits Orange peel (C), lemon peel (B), citrus sudachi (A), (4) Sugar Strawberry jam, maple syrup plum (C) (5) Confectionery Cracker, fried rice cracker, pocky, jelly, chocolate, (10) Other vegetable Garlic (A) cocoa cookie (13) Drink and Lemonade (A, C), soy source (A, B, and C) (6) Fat and oil Mayonnaise seasoning (7) Beans Soybean paste, douban jiang, tou jiang, soy bean, (14) Fish and Salmon (A) green peas, linsen soup, kidney bean shellfish (8) Fruits Dried plum, kiwifruit, orange, cherry, banana, grape, (16) Spice Sage (C), anise (C), caraway (B and C), red pepper (B strawberry, apple and C), Horse radish (B) (9) Green and yellow Gumbo (okra), carrot, Spanish paprika, spinach (B) Fraction extracted with ethanol vegetables pumpkin, Japanese parsley, tomato, broccoli (2) Nuts and seeds Sesame (B) (10) Other Eggplant, cucumber, onion, welsh onion, bamboo (4) Sugar Blueberry jam (B) vegetables shoot, Chinese cabbage, Chinese radish (10) Other Ginger (B), cabbage (C) (11) Mushrooms King trumpet mushroom, hen of the woods nameko vegetables mushroom, Jew’s ear, shiitake mushroom, forest (16) Spice Mustard (B), laurel (A), nut meg (A), chili pepper (A), mushroom japanese hackberry red pepper (B), cumin (B), dill (A), cardamon (B), anise (12) Seaweed Wakame, hijiki, kelp (A) (13) Beverages and Barley tea, cocoa, milk tea, meat sauce bouillon, seasoning tomato sauce, wine, rice wine vinegar, Japanese sweet (C) Fraction extracted with DMSO rice wine for cooking (2) Nuts and seeds Sesame (C), tapioca (B), almond (B) (14) Fish and Scallop, northern shrimp, yellow fin tuna bonito, (4) Sugar Honey (C) shellfish oyster, filefish, canned tuna sea cucumber, little clam, (7) Bean Red bean (C) fresh-water clam flatfish, sea bream, octopus, crab, (8) Fruits Apple peel (A and B), citrus sudachi (A), orange peel cuttlefish sardine (A,C) (15) Meat, egg and Pork, beef, poultry, sausage, hen egg, omelette, (9) Green and yellow Pumpkin peel (B) Milk yoghurt, cow milk, dry skim milk, cheese vegetables (10) Other Ginger (C) vegetables (14) Fish and Salmon (A and B) attention to those final concentrations in the culture medium. shellfish Median lethal concentration of ethanol was 10% (v/v), and distilled (16) Spice Red pepper (A, B, and C), Chili pepper (A and C) Cardamon (B) Anise (A and C) water had no cytotoxic effect at 10% (v/v). When we use ethanol as extraction solvent, the final volume of ethanol in the culture med- (A) NIH/3T3, (B) Neuro-2a and (C) HepG2 show the cells damaged by foods. ium should be below 2% in order to minimise its cytotoxicity. In this test, the final volume percent of extract was 2% in the culture medium. Table 3 3.4. Cytotoxic effects of different foods Cytotoxic effect of various foods having viability of less than 50% in three different fractions extracted with water, ethanol and DMSO.

Tables 1–3 show the order of cytotoxicity of foods in groups cat- (A) Fraction extracted with water egorised from (1) to (16) as described in the introduction. (13) Drink and Black tea (A), coffee (A, B, and C), vinegar seasoning (16) Spice Oregano (A and B), cinnamon (C), mustard (A and C) 3.4.1. Cytotoxic effect of different foods having viability of more than (B) Fraction extracted with ethanol 95% (13) Drink and Black tea (B), lemonade (C), green tea (B) Table 1 shows the non-cytotoxic foods as follows. In group (1), seasoning rice, wheat meal and corn had little cytotoxic effect. These grain (16) Spice Dill (B), mace (A, B, and C), back pepper (A, B and C) foods mixed with water showed high viscosity due to soluble Fennel (A and B), cinnamon (A, B, and C), clove (A, B, and C) starch, and the preparation of supernatant required centrifugation Sage (A, B and C), coriander (A), anise (B), ginger (A) at 10,000g for 10 min. Wheat, corn and potato infected by Fusarium Caraway (A, B and C), celery (A, B and C), wasabi (A) species and contaminated by mycotoxin such as NV, DON and FX Laurel (B and C), nutmeg (A, B and C), oregano (A, B are well known to cause food poisoning. In the tests, rice, wheat and C) Japanese pepper (A, B and C), turmeric (A, B and C) meal and corn were estimated to be little contaminated by the Mustard (A and C) above mycotoxins because the cytotoxic effects of NV, DON and (C) Fraction extracted with DMSO FX were not observed under the high sensitivity detection condi- Fruits Lemon peel (B) tions requiring 48 h-incubation with mammalian cells as shown (13) Drink and Black tea (A, B and C), lemonade (C), green tea (A and in Fig. 2. seasoning B) In group (3), Chinese yam, potato and devil’s tongue had few (16) Spice Dill (A and B), mace (A, B and C), black pepper (A) cytotoxic effects as shown in Table 1. Though potatoes are well Fennel (A), cinnamon (A, B and C), clove (A, B and C) Sage (A), coriander (A), anise (A, B and C), ginger (A) known to contain toxic glycoalkaloids, of which the most prevalent Caraway (A, B and C), celery (A and C), wasabi (A, B are solanine and chaconine (Adnan, Yildirim, San, Fatma, & Yildi- and C) rim, 2010), the cytotoxicity effect of these glycoalkaloids was not Laurel (A, B and C), nutmeg (A and C), oregano (A, B detected by bioassay. and C) Japanese pepper (A and C), turmeric (A, B and C), In group (5), cracker, fried rice, cracker, Pocky (a Japanese snack Cumin (B), mace (A, B and C), mustard (A and C) food), jelly, chocolate and cocoa cookie had little cytotoxic effects. If confectionery shown in Table 1 contains antioxidants and preser- (A), NIH/3T3; (B) Neuro-2a; and (C) HepG2 show the cells damaged by foods. S. Yamashoji et al. / Food Chemistry 138 (2013) 2146–2151 2149

tetrodotoxin (puffer poison) and brevetoxin are included in foods, these toxins can be detected under the conditions of this study as described later.

3.4.3. Cytotoxic effect of different foods having viability of less than 50% Spice, herb and tea showed viability of less than 50% in fractions extracted with three different solutions as shown in Table 3. Vari- ous alkaloids, terpenes, ﬂavones and pigments in spice, herb and

tea might contribute to inhibition of menadione-catalysed H2O2 production. The biological activity of spices and herb is very com- plex, e.g., turmeric has anti-inﬂammatory, hypocholesterolaemic, choleratic, antimicrobial, antirheumatic, antibacterial, antiviral, cytotoxic, spasmolytic, hypersensitive, antidiabetic and antihepa- totoxic effects (Peter, 2001). Turmeric is also credited with antican- Fig. 2. Cytotoxic effects of Fusarium mycotoxins Nivalenol (d), deoxynivalenol (N), cerous properties (Kuttan, Bhanumathy, Nirmala, & George, 1985; fusarenon-X (j). Each symbol represents the mean and the standard deviation Rao, Riverson, Simi, & Reddy, 1995). A few compounds such as cap- (small bar) of six determinations. saicin are known to be irritants. In a 2005 study capsaicin showed possible neurotoxic effects on cultivated human neuroblastoma (cancer) cells (Lee et al., 2005). Though spices and herbs have var- vatives at a few percent, cytoxicity might be detected under the ious biological effects, these have been used habitually without conditions of this study because IC50 of these food additives deter- health hazard for a long time. The addition of small quantities of mined by menadione-catalysed H2O2 production assay was spices and herbs might not cause extreme biological effects. approximately 0.02% (0.2 g/l) (Yamashoji & Isshiki, 2001). In this test 50% inhibition concentration (IC50) of the above In group (6), mayonnaise alone was tested, because it could be foods was not determined. As described in the introduction, single dispersed in the culture medium. Mayonnaise adjusted to pH 7.2 substances show a clear IC50, but it is difﬁcult to determine IC50 had little cytotoxic effect as shown in Table 1. Other fats and oils values of whole foods (data not shown). The cytotoxicity of dried were not tested because of complete separation between fat or spice, herb and tea might depend on degradation of substances oil and culture medium. during storage.

3.5. Determination of Fusarium mycotoxins 3.4.2. Cytotoxic effect of different foods having viability from 50% to 95% Fusarium mycotoxins showed little cytotoxic effect after 4-h In group (2), tapioca, sesame and almond showed viability from cultivation (data not shown), but showed obvious effect after 48- 50% to 95%, as shown in Table 2. Tapioca and almond are well h cultivation (Fig. 2), suggesting the inhibition of cell proliferation. known to contain harmful alkaloids, such as amygdalin which is DON might inhibit the synthesis of DNA and RNA and protein syn- a cyanogen compound (Norton, 1996). For example, amygdalin is thesis at the ribosomal level (Kouadio, Dano, Moukha, Mobio, & present in bitter almond at a few percent (Adnan et al., 2010), Creppy, 2007). As the detection limit of DON ranged from 106 to and was expected to be determined under the conditions of this 105%, DON present at 103% might be detected under the condi- study, because IC of amigdalin determined by menadione-cata- 50 tions of this study. lysed H2O2 production assay was about 0.05% (0.5 g/l) (Yamashoji & Isshiki, 2001). In group (7), fraction extracted from red bean with DMSO 3.6. Determination of solanine showed viability from 50% to 95%, suggesting the presence of hydrophobic toxic compounds. Toxic phytohaemagglutinin (kid- For food safety purposes, an upper limit for solanine content of ney bean lectin) is known to be found in many species of beans, 20 mg per 100 g of potato is generally accepted. The average con- and it is in highest concentration in red kidney beans (Phaseolus tent of solanine in the peel can vary from 3 to more than vulgaris). Red bean had more harmful effect than kidney bean in 100 mg/100 g (0.1%) (Bushway, Bureau, & McGann, 1983), and this test. The cytotoxic compounds of red bean will be clarified in the combination of solanine with chaconine shows synergistic the near future. cytotoxic effect (Kenkens et al., 1995; Yamashoji & Isshiki, 2001). In group (8), citrus sudachi and peel of orange, lemon and apple As the detectable concentration of solanine by chemiluminescent showed viability from 50% to 95%. Peal of some fruits is well known assay was about 0.003% (Fig. 3), solanin of peel at more than to be contaminated by pesticides. If the concentration of pesticides 100 mg/100 g (0.1%) is expected to be detected by the chemiluminescent assay when potato is homogenised without solvent and in peel is at the level of LD50 (g/kg), it should be determined by the final volume percentage of homogenate is adjusted to 10% in menadione-catalysed H2O2 production assay at lower level than culture medium. LD50 (Yamashoji & Isshiki, 2001). In group (9), pumpkin peel showed cytotoxic effects in a fraction extracted with DMSO as shown in Table 2, suggesting the 3.7. Determination of tomatine presence of hydrophobic toxic compounds such as alkaloids. In group (13), lemonade and soy source showed viability from The detectable concentration of tomatine by the chemilumines- 50% to 95%. As ethanol, acetic acid and salt showed cytotoxicity cent assay was 0.001% (Fig. 4). As tomatine in tomato that is not in a dose-dependent manner, we should pay the attention to the fully ripened is estimated to be present at 0.03%, tomatine in concentration of these compounds in foods. immature tomato is expected to be determined in a range from In group (14), salmon showed a cytotoxic effect in fractions ex- 0.003% to 0.03% when tomato is homogenised without solvent tracted with ethanol and DMSO, suggesting the presence of hydro- and the final volume percentage of homogenate is adjusted to phobic toxic compounds, such as red pigment astaxanthin. If 10% in culture medium. 2150 S. Yamashoji et al. / Food Chemistry 138 (2013) 2146–2151

Fig. 5. Cytotoxicity effect of tetrodotoxin. Each symbol represents the mean and the Fig. 3. Cytotoxic effect of solanine. Each symbol represents the mean and the standard deviation (small bar) of six determinations. standard deviation (small bar) of six determinations.

Fig. 6. Cytotoxicity effect of brevetoxin. Each symbol represents the mean and the standard deviation (small bar) of six determinations.

Fig. 4. Cytotoxic effect of tomatine. Each symbol represents the mean and the standard deviation (small bar) of six determinations. natural toxins such as Fusarium mycotoxin, solanine, tomatine, 3.8. Determination of tetrodotoxin and brevetoxin tetrodotoxin and brevetoxin were demonstrated to be determined by the chemiluminescent assay. Antibiotics, heavy metals and pes- Tetradotoxin could rescue cells by blocking the sodium efflux ticides are also expected to be detected by the chemiluminescent promoted by sodium channel activator and an inhibitor of Na+/K+- assay because these compounds were determined by colorimetric ATPase (Manger et al., 1993). According to this principle, tetrodo- assay of menadione-catalysed H2O2 production by mammalian toxin showed survival advantage in a dose-dependent manner as cells (Yamashoji & Isshiki, 2001). These facts suggest that the shown in Fig. 5. Tetrodotoxin is present from 0.1 to 70 lg/g in puf- chemiluminescent assay is useful as a screening test for the iden- fer fish, which is expected to be determined by chemiluminescent tification of the above known toxic compounds. assay, because the detectable concentration of tetrodotoxin was On the other hand, some foods showed viability of from 50% to approximately 0.1 ng/well (Fig. 5). On the other hand, brevetoxin 95% in fractions extracted with water, ethanol or DMSO as shown could enhance the sodium channel to induce cell death and show in Table 2. Letters in brackets indicated those cells sensitive to cyto- a cytotoxic effect in a dose-dependent manner (Fig. 6). The detec- toxic effects from the corresponding foods. Though peels of fruits tion limit of brevetoxin by the chemiluminescent assay was 0.5 ng, and vegetables had cytotoxic effects, these peels might be little used and was much lower than that (0.5 lg) by HPLC (Ishida, Murama- for cooking. Green tea, black tea, coffee and spice showed viability tsu, Nukaya, Kosuge, & Tsuji, 1996). The cytotoxicity test is ex- of less than 50% in fractions extracted with water, ethanol or DMSO pected to be applicable to the rapid detection of these toxins. as shown in Table 3. Strong cytotoxicity was observed in fractions extracted with ethanol and DMSO rather than water, suggesting the presence of hydrophobic toxic compounds in green tea, black 4. Discussion tea, coffee and spice. These main hydrophobic compounds might be polyphenols, alkaloids and/or pigments (Gloro et al., 2005; Most nuts and seeds, tubers, sugar, confectionery, mayonnaise, Kim, Ham, Shigenaga, & Han, 2008; Ma, Kim, & Han, 2010). The cyto- beans, fruits, green and yellow vegetables, other vegetables, mush- toxic foods shown in Tables 2 and 3 may not be harmful. However, it rooms, seaweed, beverages and seasonings, fish and shellfish, is important to identify the cytotoxic compounds in foods listed in meat, eggs and milk had little cytotoxic effects on three different Tables 2 and 3 because some foods might contain naturally- mammalian cells as shown in Table 1. If these foods show cytotoxic occurring toxic compounds such as lipid peroxides formed during effects, they may have been contaminated by toxic compounds storage. However, it is difficult to differentiate the naturally- during transportation or processing. In this study, the well-known occurring toxic compounds and existing cytotoxic compounds. S. Yamashoji et al. / Food Chemistry 138 (2013) 2146–2151 2151

Therefore it is reasonable that the proposed screening testing is extract of green tea. European Journal of Gastroenterology and Hepatology, 17, applied to the detection of contamination by toxic compounds in 1135–1137. Huang, F., & Luo, M. (2009). Evaluation of methyl thiazolyl tetrazolium (MTT) assay the foods shown in Table 1. Furthermore, the intrinsic toxic com- for screening antitumor essential oils. Food Science (Chinese), 30, 75–79. pounds, such as harmful tomatine in tomato and solanine in pota- Hugget, A. C., Schilter, B., Roberfroid, M., Antignac, E., & Koeman, J. H. (1996). to, can be also detected by the proposed screening testing. If Comparative methods of toxicity testing. Food and Chemical Toxicology, 34, 183–192. antibiotics, heavy metals, pesticides, plant toxins, fish and shell fish Ishida, H., Muramatsu, N., Nukaya, H., Kosuge, T., & Tsuji, K. (1996). Study on poisons are contained in the foods listed in Table 1, these com- neurotoxic shellfish poisoning involving the oyster crassostrea giogas in New pounds are expected to be detected by the proposed screening Zealand. Toxicon, 34, 1050–1053. Ishiyama, M., Tominaga, H., Shiga, M., Sasamoto, K., Ohkuma, Y., Ueno, K., et al. testing (Yamashoji & Isshiki, 2001). Though the proposed screening (1995). Novel cell proliferation and cytotoxicity assays using tetrazolium salt testing is useful for the detection of toxic compounds in foods, the that produces a water-soluble formazan dye. In Vitro Toxicology, 8, 187–193. identification of toxic compounds should be conducted by other Kenkens, E. A. J., de Vrije, T., van den Boom, C., de Waard, P., Plasman, H. H., & Thiel, F. (1995). Molecular basis of glycoalkaloid-induced membrane disruption. analytical methods. By way of exception, tetrodotoxin in blowfish Biochimica et Biophysica Acta, 1240, 216–228. and brevetoxin in fish or shells contaminated by Ptychodiscus brevis Kim, E. Y., Ham, S. K., Shigenaga, M. K., & Han, O. (2008). Bioactive dietary were specifically determined by the cytotoxicity testing with nerve polyphenolic compounds reduce nonheme iron transport across human cells. intestinal cell monolayers. Journal of Nutrition, 138, 1647–1651. Konemann, W. H., & Pieters, M. N. (1996). Confusion of concepts in mixture In this study, NIH3T3, Neuro2a and HepG2 cells were used be- toxicology. Food and Chemical Toxicology, 34, 1025–1031. cause of their high activity of menadione-catalysed H2O2 produc- Kouadio, J. H., Dano, S. D., Moukha, S., Mobio, T. A., & Creppy, E. E. (2007). Effects of tion. NIH3T3 cells are normal cells and show contact inhibition. combinations of Fusarium mycotoxins on the inhibition of macromolecular synthesis, malondialdehyde levels, DNA methylation and fragmentation, and Neuro-2a cells are nerve cells and have been used for the detection viability in Caco-2 cells. Toxicon, 49, 306–317. of tetrodotoxin and brevetoxin. HepG2 cells are tumour cells of li- Kuttan, R., Bhanumathy, P., Nirmala, K., & George, M. C. (1985). Potential anti-cancer ver. Three different cells having different sensitivity to cytotoxicity activity of turmeric (C. longa L.). Cancer Letters, 29, 197–202. Lee, B. K., Kim, J. H., Jung, J. W., Choi, J. W., Han, E. S., Lee, S. H., et al. (2005). of foods are expected to detect the various cytotoxic compounds. Myristicin-induced neurotoxicity in human neuroblastoma SK-N-SH cells. From the above results and discussion, the above screening as- Toxicology Letters, 157, 49–56. say is expected to be applicable to the detection of contaminating Ma, Q., Kim, E. Y., & Han, O. (2010). Bioactive dietary polyphenols decrease heme iron absorption by decreasing basolateral iron release in human intestinal Caco- cytotoxic compounds as well as the inherent cytotoxic compounds, 2 cells. Journal of Nutrition, 140, 1117–1121. such as tomatine, solanine, tetrodotoxin and brevetoxin. Manger, R. L., Leja, L. S., Lee, S. Y., Hungerford, J. M., & Wekell, M. M. (1993). Tetrazolium-based cell bioassay for neurotoxins active on voltage-sensitive sodium channels: semiautomated assay for saxitoxins, breveroxinsm and Acknowledgements ciguatoxins. Analytical Biochemistry, 214, 190–194. Mosmann, T. (1983). Rapid colorimetric assay for cellular drowrh and survival: Application to proliferationand cytotoxicity assay. Journal of Immunological This study was supported in part by a Grant from the Strategic Methods, 65, 55–63. Research Foundation Grant-aided Project for Private Universities Norton, S. (1996). Toxic effects of plants. In C. D. Klaassen, M. O. Amdur, & J. Doull from Ministry of Education, Culture, Sport, Science and Technology, (Eds.), Casarett & Doull’s toxicology (pp. 849). New York: McGraw-Hill. Peter, K. V. (2001). Handbook of herbs and spices. New York: CRC Press. Japan (MEXT), 2010–2014 (S1001032). We thank Ms. Takako Mat- Rao, C. V., Riverson, A., Simi, B., & Reddy, B. S. (1995). Chemoprevention of colon suda for her technical assistance in the cultivation of mammalian carcinogenesis by dietary curcumin, naturally colouring plant phenolic cells and the performance of various assays. compound. Cancer Research, 55, 259–266. Yamashoji, S., Yoshikawa, N., Kirihara, M., & Tsuneyoshi, T. (2012). Advantage of menadione-caytalyzed chemiluminescent assay for the determination of viable References mammalian cell number. Analytical Biochemistry, 421, 428–432. Yamashoji, S., & Isshiki, K. (2001). Rapid detection of cytotoxicity of food additives

and contaminants by a novel cytotoxicity test, menadione-catalyzed H2O2 Adnan, N., Yildirim, A. N., San, B., Fatma, K. F., & Yildirim, F. (2010). Variability of production assay. Cytotechnology, 37, 171–178. phenolics, a-tocopherol and amygdalin contents of selected almond (Prunus Yamashoji, S., Nishimoto, F., Usuda, M., Kubota, H., & Isshiki, K. (1992). Application amygdalus Batsch.) genotypes. Journal of Food Agriculture & Environment, 8, of the chemiluminescent assay to cytotoxicity test: detection of menadione- 76–79. catalyzed H2O2 production by viable cells. Analytical Biochemistry, 207, 255–260. Bushway, R. J., Bureau, J. L., & McGann, D. F. (1983). Alpha-chaconine and alpha- Yamashoji, S., & Isshiki, K. (1997). Chemiluminescent cytotoxicity test for food stuff solanine content of potato peels and potato peel products. Journal of Food with vitamin K3. Nippon Shokuhin Kagaku Kogaku Kaishi, 44, 424–429. Science, 48, 84–86. Yamashoji, S., & Isshiki, K. (1998). Cytotoxicity testing for evaluating food safety. Gloro, R., Hourmand-Ollivier, I., Mosquet, B., Mosquet, L., Rousselot, P., Salamé, E., Alternatives to Animal Testing Experiments, 5, 146–161. et al. (2005). Fulminant hepatitis during self-medication with hydroalcoholic