Natural occurrence of Fusarium toxins in soy food marketed in Germany

International Journal of Food Microbiology 113 (2007) 142 – 146 www.elsevier.com/locate/ijfoodmicro

Natural occurrence of Fusarium toxins in soy food marketed in Germany Margit Schollenberger ⁎, H.-M. Müller, Melanie Rüf le, Helga Terry-Jara, Sybille Suchy, Susanne Plank, W. Drochner Institute of Animal Nutrition, Hohenheim University, Emil-Wolff-Str. 10, D-70599 Stuttgart, Germany Received 27 December 2005; received in revised form 23 May 2006; accepted 2 June 2006

Abstract A total of 45 samples of soy food including whole beans, roasted soy nuts, flour and flakes, textured soy protein, tofu, proteinisolate including infant formulas and fermented products (soy sauce) were randomly collected in food and health food stores and analysed for Fusarium toxins. A spectrum of 13 trichothecenes of the A-type as well as of the B-type were determined by gas chromatography/mass spectrometry, zearalenone (ZEA), α- and β-zearalenol (α- and β-ZOL) by high performance liquid chromatography (HPLC) with fluorescence and UV-detection. Detection limits ranged between 1 and 19 μg/kg. At least one of the toxins investigated was detected in 11 out of a total of 45 samples of soy food belonging to different commodities. Scirpentriol (SCIRP), 15-monoacetoxyscirpenol, 4,15-diacetoxyscirpenol, T-2 tetraol, HT-2 toxin, deoxynivalenol (DON), 15- and 3-acetyldeoxynivalenol, ZEA, α- and β-ZOL were detected in at least one sample, T-2 triol, T-2, NEO, NIV and FUS-X were not detected in any sample. Five out of 11 samples were positive for one toxin, one sample for two, three, six or seven toxins, two samples for 5 toxins, demonstrating the possibility of a contamination of soy food with a spectrum of Fusarium toxins. SCIRP, DON and ZEA were found up to 108, 260 and 214 μg/kg, the other toxins did not exceed 61 μg/kg. A first insight into the contamination of soy food with a broad spectrum of Fusarium toxins is provided. © 2006 Elsevier B.V. All rights reserved. Keywords: Trichothecenes; A-type; B-type; Zearalenone; Soy food

1. Introduction A global incidence of Fusarium toxins has been reported for cereals (Placinta et al., 1999) and a frequent contamination of grain-based foods was confirmed (Scott, 1997; Schollenberger et al., 1999, 2002, 2005; Lombaert et al., 2003). As these toxins are capable of producing a wide range of toxic effects, advisory levels of 500 μg/kg of DON and 50 μg/kg of ZEA were passed by the European Union (European Commission, 2005) for grain-based foodstuffs such as bread, cakes, biscuits, cerealbased snacks and breakfast cereals. Soybeans (Glycine max) have been grown in the Far East since early times and have become of supreme importance as a source of oil and protein throughout the world during the 20th century (Hepperly, 1985). Soya meal is a main protein source in animal nutrition (Jeroch et al., 1999), other soy products, flour, textured

⁎ Corresponding author. Tel.: +49 711 459 2407; fax: +49 711 459 2421. E-mail address: [email protected] (M. Schollenberger). 0168-1605/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.ijfoodmicro.2006.06.022

products, tofu, fermented products amongst others, are used as foodstuffs. Such foods are consumed in particular by healthconscious people (Keinan et al., 2002), represent a significant part of vegetarian diets and frequently are used also for infant food formulas (Divi et al., 1997). Fusarium rot of soybeans is described in the literature and a variety of Fusarium species have been isolated from this commodity (Pitt and Hocking, 1999). These strains are known to produce a broad spectrum of toxins including zearalenone (ZEA) and trichothecenes of the A- and B-type (DeNijs et al., 1996). Thus the in vitro formation of HT-2 toxin (HT-2), T-2 toxin (T-2), T-2 tetraol, neosolaniol (NEO) and ZEA by Fusarium strains with soybeans as substrate was reported by Richardson et al. (1985). These authors suggested soybean products to present a mycotoxic hazard which warrants attention. In native beans as well as in some products used in agricultural practice trichothecenes and ZEA were detected by Clear et al. (1989), Jacobsen et al. (1995) and Rafaj et al. (2000). Schollenberger et al. (2006) analysed soya meal for a total of 16 Fusarium toxins and detected eight of these toxins at levels up to 240 μg/kg.

M. Schollenberger et al. / International Journal of Food Microbiology 113 (2007) 142–146

Little information exists about the occurrence of Fusarium toxins in soy-based foodstuffs. Those items so far have been analysed only for deoxynivalenol (DON) and ZEA. Both toxins were detected in infant soy-based cereals, ZEA was not found in other soy-based infant foods (Lombaert et al., 2003); Scott (1997) reported the presence of ZEA in soybeans and soy food. In the present study a variety of soya derived foodstuffs marketed in Germany were analysed for a total of 16 Fusarium toxins including type A trichothecenes (scirpentriol (SCIRP), 15-monoacetoxyscirpenol (MAS), 4,15-diacetoxyscirpenol (DAS), T-2 tetraol, T-2 triol, HT-2, T-2, NEO), type B trichothecenes (DON, 3- and 15-acetyl-DON (3-, 15-ADON), nivalenol (NIV) and fusarenon-X (FUS-X)), as well as ZEA, α- and β-zearalenole (α- and β-ZOL). Trichothecenes were analysed by gas chromatography/mass spectrometry (GC/MS), ZEA and its derivatives by high performance liquid chromatography (HPLC) as described previously (Schollenberger et al., 1998, 1999, 2005, 2006). 2. Experimental 2.1. Sample materials A total of 45 samples of soy food was purchased randomly in commercially available size from food stores and health food shops in Southwest Germany. The objective was to include a spectrum of different commodities of soy food into the survey, belonging to the established product line of food stores and health food shops. Provenances of basing soybeans were Europe, Brazil and USA for 7, 1 and 1 samples, respectively; for other foodstuffs provenance was not mentioned. Soy food samples analysed were produced by firms located in different regions of Germany and in European countries, which are distributing their products in food stores and health food stores at national or international level. The samples collected did not contain any grain components to allow for statements about toxin origin with the exception of fermented products, which in general are produced using wheat or rice as ingredients. The following commodities were collected: Whole beans (n = 6); soy nuts, roasted (n = 5); flour and flakes, not defatted (n = 10); flakes, partially defatted and products (crisp) (n = 5); soy protein, textured (n = 5); tofu (n = 5); proteinisolate including infant formulas (n = 5); possible other ingredients: glucose syrup, plant fat, glucose, vitamins, minerals, amino acids; fermented products (soy sauce) (n = 4), possible other ingredients: salt, wheat, corn, rice. Samples were milled (particle size about 1.5 mm) if necessary and stored at −18 °C prior to analyses. Samples of tofu were freeze-dried prior to milling. 2.2. Mycotoxin analyses 2.2.1. Trichothecenes All standard substances were bought from Sigma (Deisenhofen, Germany). Extraction, clean-up and derivatisation for trichothecene analysis in samples other than fermented products was carried out as described in detail previously by Schollenber-

143

ger et al. (1998, 2005). In brief, for all samples except fermented products extraction was performed with a mixture of acetonitrile and water followed by liquid/liquid extraction with hexane. Clean-up was carried out by solid phase extraction using a florisil and a cation exchange cartridge. Derivatisation was performed with trifluoraceticanhydride, verrucarol was added to control derivatisation efficiency. Separation and quantitation was by GCMS consisting of a Magnum Ion Trap system in the chemical ionisation mode with isobutane as reactant gas. A DB-5 MS phase (30 m × 0.25 mm, film thickness 0.25 μm) (J&W Scientific, Folsom, CA, USA) was used as a capillary column. The carrier gas was helium 5.0, the temperature of the injection port was 260 °C, the injection volume 1 μl. The temperature program of the gas chromatograph was: 140 °C (2 min)–7 °C/min–275 °C (2 min)–30 °C/min–290 °C (5 min). The transfer line was heated to 270 °C, the mass spectrometer to 190 °C. Detection limits as assessed with pure reference substances at a signal to noise ratio of 3:1 were 7, 9, 7, 14, 19, 4, 3, 7, 5, 8, 3, 14 and 6 μg/kg for DON, 3and 15-ADON, NIV, FUS-X, T-2, HT-2, T-2 tetraol, T-2 triol, SCIRP, MAS, DAS and NEO, respectively. Quantitation limits were at a signal to noise ratio of 6:1. Repeatability and recovery were determined by spiking 10 g of soy flour matrix with toxin standard solution at a level of 200 μg/kg prior to addition of solvent and extraction. Recovery rates were between 74 and 111%, standard deviations (n = 4) were between 1.6 and 12.5%, respectively (Table 1). Performing the sample preparation described above for fermented products a matrix layer not soluble in methanol was obtained when evaporating extract to dryness prior to florisil cartridge clean-up, preventing complete solubilisation of trichothecenes in methanol. Therefore a method using liquid–liquid extraction with extrelut material was developed. Ten ml of soy sauce were mixed with 5 ml of bidestilled water and applied to an extrelut cartridge (sample volume 5–20 ml) (VWR, Darmstadt, Germany). After 15 min 70 ml of a mixture of ethylacetate and methanol (85/15 v/v) were passed through the column, eluted into a pear shape flask and evaporated to dryness. The residue was dissolved with 5 ml of methanol, 0.5 ml of extract were Table 1 Recoveries and standard deviations (n = 4) of trichothecene toxins from soybean and soy sauce spiked at a level of 200 μg/kg each Toxin

NIV FUS-X DON 15-ADON 3-ADON HT-2 T-2 T-2 triol T-2 tetraol SCIRP MAS DAS NEO a

Not quantifiable.

Recoveries (%) and standard deviations (%) Soybean

Soy sauce

81 ± 3.2 86 ± 1.6 74 ± 2.3 103 ± 8.4 107 ± 3.2 84 ± 5.2 81 ± 7.2 82 ± 5.7 75 ± 7.0 79 ± 5.3 76 ± 2.9 111 ± 12.5 105 ± 9.2

65 ± 3.8 81 ± 8.0 65 ± 7.3 92 ± 12.6 61 ± 5.5 56 ± 6.4 81 ± 11.0 40 ± 6.8 41 ± 7.7 65 ± 7.3 65 ± 9.6 55 ± 5.5

a

144

M. Schollenberger et al. / International Journal of Food Microbiology 113 (2007) 142–146

samples, toxin contents of the other samples are summarised in Table 2. Positive samples contained at least one of the toxins, SCIRP, MAS, DAS, T-2 tetraol, HT-2, DON, 15-ADON, 3-ADON, ZEA, α- and β-ZOL, whereas T-2 triol, T-2, NEO, NIVand FUS-X were not detected in any of these samples. ZEA, DON, α-ZOL, MAS and HT-2 were found in 7, 6, 5, 4, and 3 samples, respectively, both SCIRP and β-ZOL in two samples, DAS, T-2 tetraol, 15- and 3-ADON in one sample. Contents ranged at 11- 260 μg/kg for DON, 2- 214 μg/kg for ZEA, 25- 108 μg/kg for SCIRP, and at 2 and 61 μg/kg for the other toxins. Five out of eleven samples were positive for only one toxin, one sample for two, three, six or seven toxins, two samples for five toxins. Toxin combinations mostly comprised at least one toxin out of the three groups investigated, i.e. of the A- and B-type trichothecenes and of the group of the estrogenic active toxins ZEA, α- and β-ZOL. Thus a soy flour sample was positive for SCIRP, MAS, HT-2, DON, ZEA, α- and β ZOL, a crisp sample for SCIRP, MAS, T-2 tetraol, DON, ZEA and α-ZOL, a textured product for MAS, HT-2, DON, ZEA and α-ZOL, and a further crisp sample for MAS, HT-2, DON, ZEA and α-ZOL. Among the estrogenic active toxins, ZEA was found together with α-ZOL in 5 samples, and two of these samples additionally contained the β-isomer.

derivatised prior to trichothecene analysis as described above. Repeatability and recovery were determined using soy sauce as matrix and spiking level of 200 μg/kg. Recovery rates of trichothecenes are given in Table 1. The quantitation of NIV was impeded by coeluting matrix components. 2.2.2. Zearalenone and derivatives Determination of ZEA, α- and β-ZOL was carried out as described previously by Schollenberger et al. (1999, 2006). In brief, after extraction with a mixture of acetonitrile and water, sample clean-up was carried out using an immunoaffinity column. Identification and quantitation of ZEA, α- and β-ZOL was carried out by HPLC. Fluorescence detection was performed at 235 nm (excitation) and 450 nm (emission), detection limits at a signal to noise ratio of 3:1 were at 1, 1 and 8 μg/kg for ZEA, α-ZOL and β-ZOL respectively. Additionally UV-detection at 235/280 nm or diode array detection, respectively, was used to control toxin identity for toxin levels. Detection limits for ZEA and its derivatives at 235 nm were at 2 μg/kg, respectively. As compared to ZEA and α-ZOL fluorescence intensity of β-ZOL was relatively low, for this toxin signal of UV 235 nm was used for quantitation. Quantitation limits were at a signal to noise ratio of 6:1. Toxin contents between detection and quantitation limits were calculated as the average. Recovery experiment was conducted as described above using spiking levels of 10 μg/kg for ZEA and α-ZOL and 20 μg/kg for β-ZOL, respectively. Mean recovery rates out of soy flour were between 85 and 106%, standard deviations between 5 and 8% respectively. Results of toxins investigated were not corrected for recovery.

4. Discussion In the present study soya based foodstuffs were analysed for the first time for a broad spectrum of 16 Fusarium toxins and 11 of these toxins were detected. In soybeans for feed use T-2, DON, NIV and ZEA were found at a considerable rate of contamination (Rafaj et al., 2000). According to Jacobsen et al. (1995) DAS, DON, HT-2 and ZEA were present in mold damaged soybeans in Midwest USA. Scott (1997) found ZEA in soybeans and soy foods at an incidence of 6.2% and contents between 5 and 39 μg/kg. Lombaert et al. (2003) reported an incidence of 100% DON and of 46% ZEA for soy-based infant

3. Results In the present study a total of 45 soy food samples were analysed for trichothecenes of the A- and B-type as well as for ZEA and its alcohols. None of these toxins was found in 34

Table 2 Fusarium toxins in positive samples of soy food in μg/kg Kernel roasted Kernel roasted Crisp Crisp Textured product Textured product Soy flour Tofu Protein concentrate Soy sauce Soy sauce SCIRP MAS DAS T-2 tetraol T-2 triol HT-2 T-2 NEO DON 15-ADON 3-ADON NIV FUS-X ZEA α-ZOL β-ZOL

– – – – – – – – – 11 – – – – –

– – 21 – – – – – – – – – – – – –

– not detected Each item is represented by one sample.

108 34 – 32 – – – – 260 – – – – 36 5 –

– 5 – – – 5

32 – – – – 8 2 –

– – – – – – – – – – – – – 50 3 5

– 6 – – – 5 – – 61 – – – – 17 2 –

25 6 – – – 11 – – 11 – – – 214 11 5

– – – – – – – – – – – – – 2 – –

– – – – – – – – – – – – – 2 – –

– – – – – – – – 25 – 14 – – – – –

– – – – – – – – 14 – – – – – – –

M. Schollenberger et al. / International Journal of Food Microbiology 113 (2007) 142–146

food. This food type usually contains corn, and the contribution of this fraction to toxin contamination was not clear. In the present study ZEA as well as α- and β-ZOL were detected in soy food samples, with the combinations ZEA/α-ZOL and ZEA/α-ZOL/β-ZOL in 3 and 2 samples, respectively. Jacobsen et al. (1995) detected ZEA in whole soybeans, hulls, meal and oil, but zearalenol only in hull and meal. In 13 samples of soymeal α- and β-ZOL were found in 4 and 2 samples, respectively (Schollenberger et al., 2006). In wheat, barley and oat samples collected during several years in an area of southwest Germany however, the two zearalenols could be only found in 0.33% out of 1504 samples, whereas 23.5% of these samples were positive for ZEA (Müller et al., 1997a,b, 1998, unpublished). Zearalenols were not detected in any of 60 samples of wheat flour collected in Germany (Schollenberger et al., 2002). The presence of these estrogenic active toxins and also of the other Fusarium toxins may depend on the Fusarium profile of the original substrate (cereals, soybeans), the toxigenicity of the strains present (Richardson et al., 1985; DeNijs et al., 1996), and/or on the influence of substrate and other environmental conditions on toxin formation (Barath et al., 1997). Further research must be awaited to clarify the occurrence of Fusarium toxins in soybeans and their transfer into soya products. The present study demonstrates the possibility of a multitoxin contamination of soy food, with up to 7 toxins co-occurring in the same sample (Table 2). The co-occurrence of up to 5 toxins has been described previously for soya meal (Schollenberger et al., 2006). Advisory levels of 500 μg/kg of DON and 50 μg/kg of ZEA were passed by the European Union (European Commission, 2005) for bread, cakes, biscuits, cereal-based snacks and breakfast cereals. For the soy food samples analysed in the present study the advisory DON level was not equaled in any sample, that of ZEA was equaled in one and exceeded in a second sample. For estimating the toxin uptake by the consumer a mean content of foodstuffs must be determined, considering both positive and negative samples. Eriksen and Alexander (1998) proposed the use of a corrected mean content based on a value of half the detection limit in negative samples. For 45 soy food samples analysed in the present study the corrected mean DON and ZEA contents can be calculated at 12 and 8 μg/kg, based on half a detection limit for DON and ZEA of 3.5 and 0.5 μg/kg, respectively. For 60 samples of wheat flour analysed previously (Schollenberger et al., 2002) the corrected mean DON and ZEA contents are at 287 and 2 μg/kg, based on the same detection limits. This suggests a significant difference between the DON contamination of the two commodities. However, amongst others, the limited number of samples analysed must be considered. The corrected mean content of ZEA was higher in soy food compared to wheat flour. When calculating the exposure of the consumer to ZEA great differences in average daily intake of grain-based foods and soy food must be considered. Thus the average per caput supply of the European consumer amounts to 226 g/day for cereals (WHO, 1998), but only at 0.8 and 1.7 g/day for men and women for soy products (Keinan et al., 2002). The widespread use of soy products in infant food formulas and the significant consumption of soy products by people consuming a

145

vegetarian diet (Divi et al., 1997) should be kept in mind. Thus Keinan et al. (2002) reported a daily intake of soy products by European people consuming a health-conscious diet at 144 g for men and 89 g for women. Therefore for consumers with high soy food intake the contribution of this item to total exposure of ZEA seems not to be negligible. A regular control of ZEA content in soy food might be recommendable.

References Barath, A., Sawinsky, J., Halasz, A., Borbiro, G., 1997. Substrate influence on mycotoxin production of Fusarium species and their analytical detection. Cereal Research Communicaton 25, 353–354. Clear, R.M., Nowocki, T.W., Daun, J.K., 1989. Soybean seed discoloration by Alternaria spp. and Fusarium spp., effects on quality and production of fusariotoxins. Canadian Journal of Plant Pathology 11, 308–312. DeNijs, M., Rombouts, F., Notermans, S., 1996. Fusarium molds and their mycotoxins. Journal of Food Safety 16, 15–58. Divi, R.L., Chang, H.C., Doerge, D.R., 1997. Anti-thyroid isoflavones from soybean. Biochemical Pharmacology 54, 1087–1096. Eriksen, G.S., Alexander, J., 1998. Fusarium toxins in cereals — a risk assessment. Tema Nord 1998, 502. European Commission, 2005. Commission Regulation 856/2005/EC of 6. June 1998. Official Journal of the European Communities L143/3. Hepperly, P.R., 1985. Fusarium species and their association with soybean seed under humid tropical conditions in Puerto Rico. Journal of Agriculture of the University of Puerto Rico 79, 25–33. Jacobsen, B.J., Harlin, K.S., Swanson, S.P., Lambert, R.J., Beasley, V.R., Sinclair, J.B., Wei, L.S., 1995. Occurrence of fungi and mycotoxins associated with field mold damaged soybeans in the midwest. Plant Disease 79, 86–89. Jeroch, H., Drochner, W., Simon, O., 1999. Ernährung landwirtschaftlicher Nutztiere. Verlag Eugen Ulmer, Stuttgart, p. 229. Keinan, B.L., Peeters, P.H.M., Mulligan, A.A., Navarro, C., Slimani, N., 2002. Consumption of soy products among European consumers of a healthconscious diet. IARC Scientific Publications 156, 109–112. Lombaert, G.A., Pellaers, P., Roscoe, V., Mankotia, M., Neil, R., Scott, P.M., 2003. Mycotoxins in infant cereal foods from the Canadian retail market. Food Additives and Contaminants 20, 494–504. Müller, H.-M., Reimann, J., Schumacher, U., Schwadorf, K., 1997a. Fusarium toxins in wheat harvested during six years in an area of southwest Germany. Natural Toxins 5, 24–30. Müller, H.M., Reimann, J., Schumacher, U., Schwadorf, K., 1997b. Natural occurrence of Fusarium toxins in barley harvested during five years in an area of southwest Germany. Mycopathologia 137, 185–192. Müller, H.M., Reimann, J., Schumacher, U., Schwadorf, K., 1998. Natural occurrence of Fusarium toxins in oats harvested during five years in an area of southwest Germany. Food Additives and Contaminants 15, 801–806. Pitt, J.I., Hocking, A.D., 1999. Fungi and Food Spoilage. Aspen Publ. Inc., Gaithersburg, Maryland. Placinta, C.M., D'Mello, J.P.F., Macdonald, A.M.C., 1999. A review of worldwide contamination of cereal grains and animal feed with Fusarium mycotoxins. Animal Feed Science and Technology 78, 21–37. Rafaj, P., Bata, A., Jakab, L., Vanyi, A., 2000. Evaluation of mycotoxincontaminated cereals for their use in animal feeds in Hungary. Food Additives and Contaminants 17, 799–808. Richardson, K.E., Hagler, W.M., Haney, C.A., Hamilton, P.B., 1985. Zearalenone and trichothecene production in soybeans by toxigenic Fusarium. Journal of Food Protection 48, 240–243. Schollenberger, M., Lauber, U., Terry Jara, H., Suchy, S., Drochner, W., Müller, H.-M., 1998. Determination of eight trichothecenes by gas chromatography mass-spectrometry after sample clean-up by a two-stage solid phase extraction. Journal of Chromatography A 815, 123–132. Schollenberger, M., Terry Jara, H., Suchy, S., Drochner, W., Müller, H.-M., 1999. A survey of Fusarium toxins in cereal-based foods marketed in an area of southwest Germany. Mycopathologia 147, 49–57.

146

M. Schollenberger et al. / International Journal of Food Microbiology 113 (2007) 142–146

Schollenberger, M., Terry Jara, H., Suchy, S., Drochner, W., Müller, H.-M., 2002. Fusarium toxins in wheat flour collected in an area in southwest Germany. International Journal of Food Microbiology 72, 85–89. Schollenberger, M., Müller, H.-M., Rüfle, R., Suchy, S., Planck, S., Drochner, W., 2005. Survey of Fusarium toxins in foodstuffs of plant origin marketed in Germany. International Journal of Food Microbiology 97, 317–326. Schollenberger, M., Müller, H.-M., Rüfle, R., Suchy, S., Plank, S., Drochner, W., 2006. Natural occurrence of 16 Fusarium toxins in grains and feedstuffs of plant origin from Germany. Mycopathologia 161, 43–52.

Scott, P.M., 1997. Multi-year monitoring of Canadian grains and grain-based foods for trichothecenes and zearalenone. Food Additives and Contaminants 14, 333–339. WHO, 1998. Global Environment Monitoring System/Food Contamination Monitoring and Assessment Programme (GEMS/Food), Regional Diets. WHO/FSF/FOS/98.3.