509 Journal of Food Protection, Vol. 70, No. 2, 2007, Pages 509–513 Copyright ᮊ, International Association for Food Protection Research Note Vomitoxin and Zearalenone Content of Soft Wheat Flour Milled by Different Methods VALENTINO PALPACELLI,1* LUCA BECO,1 AND MAURIZIO CIANI2 1Studio Analisi Alimentari, Via XXV aprile 2, 06014 Montone (PG), Italy; and 2Dipartimento di Scienze degli Alimenti, Universita` Politecnica delle Marche, via Brecce Bianche, 60121 Ancona, Italy Downloaded from http://meridian.allenpress.com/jfp/article-pdf/70/2/509/1678933/0362-028x-70_2_509.pdf by guest on 01 October 2021 MS 06-297: Received 31 May 2006/Accepted 10 August 2006 ABSTRACT Given the prominence and the growing importance of mycotoxins in human and animal health, and particularly of vomitoxin and zearalenone in people who use wheat and wheat products as their staple diet, we investigated two different types of wheat milling. Wheat produced according to good manufacturing practice related to mycotoxin risks (from sowing to harvesting) was used to compare the vomitoxin and zearalenone content of soft wheat flour, following the use of two different types of milling, traditional milling with a stone mill and modern milling with a roller mill. Moreover, the vomitoxin and zearalenone content was also evaluated in commercial stone-milled and roller-milled flours. Our results show that stone milling reduced vomitoxin and zearalenone content in flours, compared with the use of the roller-mill system. Wheat and wheat products form the staple diet of about usual, so it became possible to produce middlings, which half of the world’s population (6), and for this reason, the were then further reduced by a low-grinding phase. nutritional and toxicological properties of wheat are partic- This represented the first step toward the introduction ularly important to the human diet. Mycotoxins are the pri- of roller mills, which supplanted the stone-mill grinding in mary food risk in wheat (9), and vomitoxin and zearalenone Europe, followed the rest of the world (except for some are mycotoxins that are produced by a well-known wheat rural areas of developing countries and in some typical pathogen: the fungi of the Fusarium genus (11). In Western products in Western countries). In this context and in view countries, the risks associated with the mycotoxins appear of the fact that vomitoxin and zearalenone show widespread to be related to chronic exposure rather than to those of diffusion in wheat products (7), we carried out this study acute exposure (as seen in the developing countries) (7). to investigate the effects of the introduction of the roller- It has been shown that vomitoxin is able to superinduce mill method on the content of these mycotoxins in wheat cytokine gene expression and secretion. This superinduc- flours. tion of the cytokines is associated with many allergic con- ditions and autoimmune diseases (7, 8, 13–16, 18), and it MATERIALS AND METHODS appears that chronic exposure to low levels of this toxin Sample preparation. Beginning November 2001, to obtain can produce immunological effects. Zearalenone appears to the necessary wheat an experimental plan and protocol was fol- be immunotoxic and associated with various diseases of the lowed to reduce the potential for Fusarium infection to a mini- reproductive system (1, 12, 13, 16). mum (4): Bolero variety, late seeding, seed dressing, fungicide Vomitoxin and zearalenone are generally found in the treatment at anthesis, correct soil tillage, and reduced nitrogen external layers of the wheat kernels (17, 19); however, un- fertilization. In July 2002, the mechanical harvesting was carried der particular climatic conditions the infection can be sys- out, and ca. 6,000 kg of sound kernels were obtained. These were temic, and the contamination can extend through the whole stored in one semiunderground masonry bin located on a private farm at about 400 m above sea level, under a rooftop. Hermetic kernel (5). Indeed, the milling phase initially included the storage with periodic fumigation was adopted. Between July 2002 elimination of the external layers of the wheat kernels, thus and December 2004, 36 samples of wheat kernels were taken at reducing vomitoxin and zearalenone levels. Before the 20th regular intervals to both an ancient stone mill located in Bevagna century, the milling operation was conducted exclusively (Perugia, Italy), and experimental roller mills. The resulting flour by the stone-mill method (10). However, in the 19th cen- samples were analyzed for their vomitoxin and zearalenone con- tury, with the introduction of hard wheat varieties in Europe tent. that produced lower flour yields with the stone-mill method, In the stone mill, the wheat kernels were loaded into a seed the practice of ‘‘high grinding’’ was developed. This con- box and then passed through: (i) the grain sieve to eliminate the sisted of milling with the stone more widely separated than cereal straw, stones, etc.; (ii) the trimming machine with an as- pirator to clean the surface layer of the grain and eliminate dust * Author for correspondence. Tel: ϩ39-075-9306479; Fax: ϩ39-075- generated by this operation; and (iii) the seed winnower to elim- 9306479; E-mail: [email protected]. inate vetch seed and other nonwheat seeds and to further clean 510 PALPACELLI ET AL. J. Food Prot., Vol. 70, No. 2 the wheat. Finally, the cleaned wheat kernels were passed to the 80%. Data were not corrected for recovery. Analyses were carried mill stones. The millstone grinder comprises two large circular out in triplicate, and the data are given in the tables as means Ϯ stones that are placed one on the top of the other. The lower stone SD. is fixed, whereas the upper stone rotates. The internal surfaces of the stones are scored, and in the middle of the stones there is a (iv) Zearalenone analysis. Samples were analyzed for zear- hole. The wheat entered through this hole and passed between the alenone with a JASCO HPLC system with a fluorimetric detector, stones, where it was ground into smaller and smaller fragments, using an immunoaffinity column (VICAM) according to manu- thus obtaining the ‘‘whole meal’’ that went into a sack. Whole facturer instructions (20). Briefly, 20 g of each flour sample was meal can be sold directly or, as in this case, it was manually put added to2gofNaCl and 50 ml of acetonitrile-water (90:10, vol/ into a separator (a simple system of sieves) to obtain the bran, the vol), and then blended at high speed for ca. 2 min. The samples flour mixed with bran, and the white flour. were then filtered through fluted filter paper, and then filtered White stone-mill flour is typically like semolina, with larger again through a microfiber filter. The filtered extract (5 ml) was granulation than roller-mill flour. then applied to the VICAM IAC column. The column was washed For the experimental roller mills, the wheat kernels were with 5 ml of distilled water, and the zearalenone was eluted with sieved to remove the unwanted material, such as stones, dust, and 1.5 ml of methanol (HPLC grade). The eluate was collected in a Downloaded from http://meridian.allenpress.com/jfp/article-pdf/70/2/509/1678933/0362-028x-70_2_509.pdf by guest on 01 October 2021 weed seeds, and then they were put into a wheat bin, and 3% glass cuvette, and 1.5 ml of water (HPLC grade) was added. The water was added to them (tempering for 48 h to provide 15 to samples were then injected (200 ␮l) into the isocratic HPLC sys- 16% moisture content). The wheat kernels were then milled in the tem (reverse-phase C18 column: mobile phase, acetonitrile-water- experimental roller mills, with the separation of bran, germ, and methanol [46:46:8; vol/vol/vol, degassed]; flow rate, 1.0 ml/min; endosperm. The first set of mills (the breaking rollers) separated loop, 100 ␮l; detection wavelength, 274-nm excitation and 440- the bulk of the bran and the germ from the endosperm. The re- nm emission). The zearalenone standard was 50 ppm in acetoni- sulting product was then passed through the sieves to separate the trile (Supelco). The software was JASCO Borwin HPLC software, various constituents, giving the bran, the wheat germ, the endo- version 1.50. The limit of quantification was 1 ppb, with a mean sperm chunks, and the animal-feed byproducts. The endosperm recovery rate of 85%. Data were not corrected for recovery. Anal- chunks, together with small residual amounts of bran and germ, yses were carried out in triplicate, and the data are given in the were passed through two series of rollers (the reducing rollers) to tables as means Ϯ SD. produce the flour. Statistical analysis. Normality testing was performed using For each kind of milling, we obtained bran, flour mixed with GraphPad InStat, version 3.0a, for Macintosh. None of the sam- bran, and white flour (about 50 kg of wheat flour in 1-kg pack- ples of both of the analyses passed the normality tests aging). Thirty-six commercial flours were analyzed (18 stone-mill (D’Agostino and Pearson omnibus normality test, alpha ϭ 0.05), flours and 18 roller-mill flours), Samples were obtained over 3 so we performed a nonparametric test (Mann-Whitney test) using months from five different sales points. GraphPad Prism, version 4.00, for Macintosh (GraphPad Soft- Sample analysis. (i) Sampling methods. Sampling of flours ware, San Diego, Calif.). For the commercial flour screening, we was carried out according to the official AOAC International assumed a non–Gaussian distribution (due to the small numbers methods (2). of samples) and performed a nonparametric test (Mann-Whitney For each analysis, 10 sacks were sampled, and 100 g was test) using GraphPad Prism, version 4.00, for Macintosh.