Distribution of fumonisins in dry-milled corn fractions in Argentina

Transcription

1 Q Taylor & Francis Food Additives and Contaminants, 2002, Vol. 19, No. 5, ~ Taylor&franci,Group Distribution of fumonisins in dry-milled corn fractions in Argentina L. E. Broggit, S. L. Resnik *,A. M. Pacin ~, H. H. L. Gonzálezll**,G. Cano ~ and D. Taglieri ~ t Facultad de Bromatología, UniversidadNacional de Entre Ríos, Argentina; :j:comisión de Investigaciones Científicas de la Provincia de Buenos Aires, Argentina; Facultad de Ciencias Exactas y Naturales and IIFacultad de Ingeniería, Universidad de Buenos Aires, Argentina;,- Centro de Investigación en Micotoxinas, Universidad Nacional de Luján, Argentina; ** Consejo Nacional de Investigaciones Científicas y Técnicas, Argentina (Received 20 July 2001; revised 24 September 2001; accepted 9 Octo ber 2001) Corn samples and different dry-milled fractions collected from an industrial mill in Argentina were analysed. Average contaminations were FB flgkg-1, FB2 716 flgkg-1 and FB] 152 flgkg-1 in whole corn; FB1 135 flgkr1, FB flgkg-1 and FB] 10.2 flgkr1 in corn grits; FB1 358 flgkg-1, FB2 122 flgkg-1 and FB] 45.9 flgkg-1 in 'C' fiour; FB1 148 flgkg-1, FB flgkr1 and FB] 28.3 flgkr1 in corn meal; and FB flgkr1, FB flgkg-1 and FB] 447 flgkg-1 in germ and bran together. The fumonisin contamination level was approximately three times higher in germ and bran than in whole corn, 13 times higher than in 'C' fiour and 29 times higher than in corn meal and corn grits. Taking into account the distribution of fumonisins in commercial dry-milled corn fractions and corn meal consumption in Argentina, a theoretical whole corn level of 6640 flgkg-1 maximum of total fumonisins could be processed to obtain products considered safe for human health. Keywords: fumonisins, corn, dry-mill, mycotoxins risk exposure *To whom correspondence should be addressed at Departamento de Industrias, FCEyN-UBA, (1428) Ciudad Universitaria, Buenos Aires, Argentina. resnik@di.[cen.uba.ar Introduction Cereal corn represents in Argentina about 30% of total cereal and oilseed production. Corn exportation is about 50%. Eighty per cent of the remaining corn production (6 x 106 tonnes per year) is destined for animal feeds. The natural occurrence of mycotoxins such as fumonisins (FB), zearalenone (ZEN), deoxynivalenol (DON) and aflatoxins (AFL), in Argentinian corn, corn meal and corn-based products has been reported (Chulze et al. 1989, Sydenham et al. 1993, Resnik et al. 1996, González et al. 1999, Solovey et al. 1999). Most of the corn me al for human consumption is processed in dry-mills. In the Argentinian Province of Entre Ríos, corn is the main cereal grain produced and the industrial fractions obtained are: corn grits (fractions that do not pass through the tm screen), 'C' flour (all these fractions pass through the tmscreen), germ and bran together (fractions that include germ, bran and part of the starchy endosperm), and corn me al (fractions that pass tmbut do not through the tmscreen). In the literature, there is only one study on distribution of fumonisins in corn in the USA that includes both commercial and experimental dry-milling processes (Katta et al. 1997). In comparison with the Argentinean commercial process, significant differences can be observed in the fractions obtained that are not technologically defined. In Argentina, germ and bran are considered one fraction, as mentioned above, and in the USA, germ and bran are two separated fractions. The distribution of fumonisin was analysed in stored whole corn and its fractions obtained from an Argentinean commercial dry-mill. The objective was to study what was the maximum fumonisin content in the whole corn that could be processed to achieve an acceptable final fumonisin contamination in the fractions destined for human consumption with avoidance of health risks. Food Additives and Contaminants ISSN X printjissn Taylor & Francis Ltd jwww.tandf.co. ukjjournals DO!: j

2 466 Materials and methods L. E. Broggi et al. Statistical analysis Commercial dry-milled corn During the crop season 1999, over 9 months (March- November), 75 samp1es of who1e kerne1 corn (with storage time of 15 days to 8 months), corn grits, germ and bran, 'C' flour and corn me al were obtained from an Entre Ríos commercial dry-mill. The corn fractions were chosen for analyses beca use they were the major parts of the milled corno Randomized samples were withdrawn in accordance with Apro et al. (1987). Subsamples of about 2 kg each were ground in a laboratory mill (Romer Labs, Inc., Union, MO, USA). Approximately 100 g of each subsample were dried to 13.5% moisture content at 60 C and maintained at -18 C prior to mycotoxin analysis. The sto red whole corn used was select to cover between 60 and gkg-, and the fractions were linked to each initial whole corn sample. Fumonisin analysis Subsamples were analysed to determine the presence of FB, FB2 and FB3 according to the method of AOAC International (2000). Briefly, aqueous methanol extracts were prepared for each of the samples, the ph was adjusted before they were applied to strong anion exchange (SAX) solid-phase extraction cartridges. Available SAX columns were assayed before application. The columns tested were Merck Adsorbex cat. no and J. & W. Scientists Fisons Accubond #1601. The cartridges were washed to remove other compounds, the fumo nisin s selectively eluted with an acetic acid-methanol solution, and the eluates were collected and dried. o-pthaldialdehyde (OPA) derivatives of each purified sample extract were prepared and analysed by reversed-phase HPLC using fluorescence detection with a stainless steel LC column, 25 cm x 4 mm i.d., packed with Lichrosorb 5 m C 8 reversed-phase material (Merck Co., Darmstadt, Germany). Identification and quantification of the fumonisins were by comparison of the retention times and peak areas in the samples with those observed for fumonisins standards. Detection limits were: FB, 10; FB2, 10; and FB3 longg- (signal-noize ratio = 3:1). Statistix for Windows (1996) software was used to compare results FB, FB2 and FB3 in corn samples and its fractions. Results and discussion Whole corn, corn grits, corn me al and 'C' flour presented an average recovery (in triplicate) for FB and FB3 > 94% at 100.1gkg- of spiked sample and for FB2 > 90% at the same leve!. Because no negative samples of germ and bran were available, two of the low contaminated samples (sample 'a': FB, 327.1gkg- ; FB2, l78.1gkg- ; no detected FB3; sample 'b': FB, 440.1gkg- ; FB2, 160.1gkg- ; no detected FB3) were chosen for recovery studies. Sample 'a' was spiked (in triplicate) with standard equivalent of 240.1gkg- FB and 75.1gkg- FB2. Sample 'b' was spiked with 190.1gkg- FB and 90.1gkg- FB2, also in triplicate. For both mycotoxins recoveries were >70%. The spiking levels chosen were approximately 50% of the initial contamination. Use of contaminated samples could affect the precision and their use is not normally recommended. Scott and Lawrence (1994) found matrix problems with the recovery of corn bran flour when the extraction solvent was methanol-water (3:1) and they proposed methanol- OAM borate buffer (ph 9.2) (3:1). In the present work, the ph of the extracts was adjusted before the SAX columns were used. Table 1 gives the different recoveries for germ and bran extracts as the average of three replicates for each column assayed. The toxin recoveries depend on the column type and ph. Thus, Merck Adsorbex, ph 5, allowed a better recovery of FB, FB2 and FB3. On the other hand, J. & W. Scientists' Fisons column ph 7 was optimal for FB and FB2 while for FB3, the best recovery corresponded to ph 5. The Merck Adsorbex column was selected in order to use only one column in each analysis. Table 2 shows the mean and range of fumonisins levels in whole corn and well as its dry-milled fractions. FB was detected in all samples of whole corn and its fractions; the germ and bran were the most

3 Distribution offumonisins in dry-milled corn fractions in Argentina 467 Table l. Fumonisins recoveries (f-lgkg-1) from corn germ and branfor Merck Adsorbex (M) and J. & W. Accubond (A) columns versus ph. FB1 FB2 FB3 ph A M A M A M Table 2. Fumonisin levels (f-lgkg- ) in whole corn and its dry-milled fractions. FB1 FB2 FB3 No. of Product samp1es Mean Min Max Mean Min Max Mean Min Max Who1e corn 'C' flour Corn meal ND ND 271 Corn grits ND ND 37 Germ and bran ND, none detected. contaminated fractions followed by whole corn and 'C' flour. Corn meal and corn grits were the least contaminated fractions. The same results were observed when the median fumonisin levels were considered (figure 1). FB2 and FBJ (table 1) were detected in all samples of whole corn, 'C' flour and germ and bran and in seven of 12 samples of corn grits. In corn meal, FB2 was detected in 11 of 14 samples and FBJ in 10 of 14 samples. Some samples of the germ and bran were highly contaminated with fumonisins and could be a serious animal health risk because this fraction is used in feeds. The FB, FB2 and FBJ levels in corn meal obtained in Entre Ríos were lower in comparison with those observed elsewhere (Bullerman and Tsai 1994, Doko and Visconti 1994, Doko et al. 1996, Katta et al. 1997, Patel et al. 1997, Castelo et al. 1998) and by Solovey et al. (1999) in corn meal fram commercial outlets in Buenos Aires. The FB levels observed in the corn grits obtained in the present study were higher than those of Sanchis et al. (1994) in corn grits fram the Spanish market and similar to those reported by Pittet et al. (1992) in Switzerland. Significant (p < 0.05) and highly significant (p < 0.01) differences in the comparison between FB, FB2, FBJ and total fumo ni sins levels in whole corn and its drymilled fractions and between the different fractions were calculated by using a non-parametric statistical method. Based on Wilks-ShapirajRankit Plots for FB, FB2, FBJ and total fumonisins levels in whole corn and its fractions, a Normal distribution was not observed. A non-parametric Sign test was then applied for comparison. Taking into account the highly significant differences between the most contaminated fraction (germ and bran) and the other fractions (p < 0.01), the relationship between total fumonisin means was calculated. Total fumonisin means were approximately three times higher in germ and bran than in whole corn, 13 times higher than in 'C' flour, and 29 times higher than in corn me al and corn grits. At the last meeting of the Joint FAOjWHO Expert Committee on Food Additives (FAOjWHO 2001), a provisional maximum tolerable daily intake (PMTDI) offumonisins B, B2 and BJ, alone or in combination, of 2l-1gkg-1 of body weightday- was suggested. Corn me al consumption in the country was estimated as 22.8 gday- per adult by Pacin et al. (1998) and with the fumonisins distribution obtained in the present study, a theoretical whole corn level of

4 468 L. E. Broggi et al tg/kg ' ~"'o,,0 0'<:- G,~é,CJ ~ ~0 0'<:- C! CJo~ is;"'~ <",<:- /:)'" ",'" 00'~ Figure l. Median offe], FE2 and FE3 levels in whole corn and its dry-milledfractions tgkg-1 maximum of total fumonisins could be processed to obtain products considered safe for human health. On the other hand, germ and bran contamination obtained in that production would be of high risk for animal consumption. Acknowledgments The authors acknowledge the Comisión de Investigaciones Cientificas de la Provincia de Buenos Aires, Consejo Nacional de Investigaciones Científicas y Técnicas, Universidad de Buenos Aires, Universidad Nacional de Entre Ríos, Universidad Nacional de Luján and Agencia Nacional de Promoción Científica for financial support. References AOAC INTERNATIONAL, 2000, Ofjicial Methods of Analysis, vol. 2 (Gaithersburg: AOAC International), pp APRO, N., RESNIK, S. L., and FERRO FONTÁN, c., 1987, Representative samples for mycotoxin analysis. Anales de la Asociación Química Argentina, 75, BULLERMAN, L. B., and TSAI, W. Y. J., 1994, Incidence and levels of Fusarium moniliforme, Fusarium proliferatum and fumonisins in corn and corn-based foods and feeds. Journal of Food Protection, 57, CASTELO, M. M., SUMNER, S. S., and BULLERMAN, L. B., 1998, Occurrence of fumonisins in corn-based food products. Journal of Food Protection, 61, CHULZE, S., BERTINETTI, c., DALCERO, A., ETcHEVERRY, M., FARNOCHI C, TORREs, A., RIZZO, 1., and VARSAVSKY, E., 1989, Incideuce of aflatoxin, zearalenone, and deoxynivalenol on corn in Argentina. Mycotoxin Research, 5, DOKO, M. B., CANET, c., BROWN, N., SYDENHAM, E. W., MpUCHANE, S., and SIAME, B. A., 1996, Natural co-occurrence offumonisins and zearalenone in cereals and cereal-based foods from eastern and southern Africa. Journal ol Agricultural and Food Chemistry, 44, DOKO, M. B., and VISCONTI, A., 1994, Occurrence of fumonisins Bl and B2 in corn and corn-based human foodstuff in Italy. Food Additives and Contaminants, 11, FAO/WHO, 2001, Joint FAO/WHO Expert Committee on Food Additives. Summary ami Conclusions, Fifty-sixth Meeting, Geneva, 6-15 February. GONZÁLEZ, H. H. L., MARTÍNEZ, E. J., PAClN, A. M., RESNIK, S. L., and SYDENHAM, E. W., 1999, Natural co-occurrence of fumonisins, deoxynivalenol, zearalenone and aflatoxins in field trial corn in Argentina. Food Additives and Contaminants, 16, KATTA, S. K., CAGAMPANG A. E., JACKSON, L. S., and BULLERMAN, L. B., 1997, Distribution of Fusarium molds and fumonisins in dry-milled corn fractions. Cereal Chemistry, 74, PACIN, A., MARTINEZ, M. E., PORTELA, M. L., and NEIRA, M. S., 1998, Consumo de alimentos en la población de la Universidad Nacional de Luján. Aporte energético y proteico. La Alimentación Latinoamericana, 221, PATEL, S., HAZ LE, C. M., WINTERTON, A. G., and GLEADLE, A. E., 1997, Surveillance of fumonisins in UK maize foods and other cereals. Food Additives and Contaminants, 14, PITTET, A., PARISOD, V., and SCHELLENBERG,M., 1992, Occurrence of fumonisins BI and B2 in corn-based products from the Swiss market. Journal ol Agricultural and Food Chemistry, 40, RESNIK, S. L., NEIRA, M. S., PACIN, A., MARTINEZ, E., APRO, N., and LATREITE, S., 1996, A survey ofnatural occurrence of aflatoxins and zearalenone in Argentine field corn: Food Additives and Contaminants, 3,

5 Distribution oi jumonisins in dry-milled corn fractions in Argentina 469 SANCHIS, V., ABADIAS, M., ONCINS, L., SALA, N., VIÑAS, l., and CANELA, R., 1994, Occurrence of fumonisins B and B2 in corn-based products from the Spanish market. Applied and Environme/1/al Microbiology, 60, SCOTT, P. M., and LAWRENCE, G. A., 1994, Stability and problems in recovery of fumonisins added to corn-based foods. Journal of AOAC International, 77, SOLOVEY,M. M. S., SOMOZA, c., CANO, G., PACIN, A., and RESNIK, S., 1999, A survey of fumonisins, deoxynivalenol, zearalenone and aflatoxins contamination in corn-based food products in Argentina. Food Additives and Contaminants, 16, STATlSTlX FOR WINDOWS, 1996, User's Manual. Copyright Analytical Software (Borland International, lnc.). SYDENHAM,E. W., SHEPARD, G. S., THIEL, P. G., MARASAS, W. F. O., RHEEDER, J. P., PERALTA SANHUEZA, C. E., GONZÁLEZ, H. H. L., and RESNIK, S. L., 1993, Fumonisins in Argentinian fieldtrial corno Journal of Agricultural and Food Chemistry, 41,