Effect of Processing on Fumonisin Concentration in Corn Flakes

Transcription

1 701 Journal of Food Protection, Vol. 64, No. 5, 2001, Pages Copyright q, International Association for Food Protection Effect of Processing on Fumonisin Concentration in Corn Flakes ANNALISA DE GIROLAMO, MICHELE SOLFRIZZO, AND ANGELO VISCONTI* Istituto Tossine e Micotossine da Parassiti Vegetali, CNR, Viale Einaudi, 51, Bari, Italy MS : Received 24 July 2000/Accepted 26 November 2000 ABSTRACT The stability of fumonisin B 1 and fumonisin B 2 during processing of corn akes was investigated with three different methods for analysis of the naturally contaminated raw material (corn our), intermediate product (extruded, but not roasted corn akes), and nal product (roasted corn akes). Only one method, using immunoaf nity column clean-up, provided reliable results in the determination of fumonisins in corn ake samples at the intermediate and nal steps of processing. About 60 to 70% of the initial amount of fumonisins were lost during the entire cycle of corn ake processing, with less than 30% losses occurring during the intermediate extrusion step (70 to 1708C for 2 to 5 min). The effect of different additives commonly present in commercial products (sodium chloride, sucrose, and ferrous sulfate heptahydrate) on the reliability of fumonisin analysis has also been investigated. The presence of sodium chloride strongly reduced fumonisin recovery when strong anionexchange (SAX) columns were used for the clean-up step, whereas the other additives appeared to have little or no effect on the accuracy of fumonisin analysis. The use of reliable analytical methods that are effective for both raw materials and processed products is of paramount relevance for studying the effect of food processing on mycotoxin-contaminated commodities. Despite the fact that some effective fumonisin decontamination occurring during corn ake processing has been shown, more work is needed to identify the thermal breakdown products of fumonisins and their relevant toxicity. Fumonisins B 1 (FB 1 ) and B 2 (FB 2 ) are mycotoxins mainly produced by Fusarium verticillioides (5Fusarium moniliforme Sheld) and Fusarium proliferatum that are frequently associated with corn worldwide (10). Fumonisins have been implicated in several animal diseases including leukoencephalomalacia in horses (1, 22), pulmonary edema in pigs (6), and hepatocarcinoma in rats (5). A possible correlation between the consumption of fumonisin-contaminated corn foods and the high incidence of esophageal cancer in southern Africa has been suggested (16). Clear evidence of carcinogenic activity of FB 1 in male rats and female mice has been recently shown based on the increased incidence of renal tubule neoplasms and hepatocellularneoplasms, respectively (18). Several reports on the occurrence of fumonisins in corn and corn-based food products have shown high fumonisin contents ( 1,000 ng/g) in corn kernels and minimally processed corn, such as polenta, corn grits, and corn our samples, but low ( 100 ng/g) or no detectable fumonisin levels in processed corn products, such as corn akes, corn tortillas, and infant mixed cereals, has been reported (2, 3, 12, 15, 19, 20). In particular, for corn akes, with the exception of a recent nding of fumonisin concentrations up to 1,300 ng/g, most of the literature reports do not exceed 90 ng/g. This was attributed, to a certain extent, to the low recoveries achievable with the current analytical methods mainly based on strong anion-exchange (SAX) clean-up (13). Several investigations on the stability of fumonisins in * Author for correspondence. Tel: ;Fax: ; visconti@area.ba.cnr.it. thermally processed corn products have shown that FB 1 is stable during cooking of polenta for 20 to 30 min in boiling water (9); the rate of fumonisin decomposition in aqueous solutions increases with processing temperature, from,27% at 1258C to.80% at 1758C, for 60 min (7, 8); substantial losses of FB 1 occurred only after dry corn was heated at 1508C for 50 min (4); losses of FB 1 and FB 2 exceeding 70% occur in dry corn meal heated at 1908C for 60 min and complete loss at 2208C for 25 min (11); fumonisins in spiked and naturally contaminated corn meal were unstable under roasting conditions (2188C for 15 min) but were stable under canning (1218C up to 87 min) and baking conditions (204 to 2328C for 20 min), probably because the canned and baked products reached lower internal temperatures than the roasted products (2). The objective of this study was to determine the effect of processing, i.e., extrusion and roasting, on the stability of fumonisins in naturally contaminated corn akes by using different analytical methods. The effect of several additives (sodium chloride, sucrose, and ferrous sulfate heptahydrate) commonly present in commercial corn akes on the analytical performance of the Association of Of cial Analytical Chemists (AOAC) of cial method (17) has also been investigated. MATERIALS AND METHODS Samples. Portions (400 g) of unprocessed corn our (sample A 1 ) and roasted corn akes (sample A 2 ) from an industrial processing plant were provided by a corn ake-producing company (company A). Aliquots (800 g) of unprocessed corn our mixed with NaCl (2%) and sucrose (8%) (sample B 1 ), dried corn ake

2 702 DE GIROLAMO ET AL. J. Food Prot., Vol. 64, No. 5 samples at the intermediate extrusion step (sample B 2 ), and the nal roasted corn akes (sample B 3 ) were obtained from another processing plant (company B) together with the procedure for industrial processing, brie y described below. Processing of corn akes. The rst step consisted of cleaning and hulling of corn kernels, followed by the grinding process. The resulting corn our, having a reduced moisture content with respect to the starting material, was stocked in stainless steel cells at 208C for 10 min, then mixed for 10 min with a mixture of NaCl ( 2%) and sucrose ( 8%), and for 15 min with liquid malt (4% dissolved in hot water). A mixture, with a moisture content of ;27% was obtained. The following extrusion process was performed in a slot hot extruder at temperatures of 70 to 1058C for 5 min. Flakes of 0.6 to 0.7 mm size with a moisture content of ;20% were obtained, dehydrated, and roasted at 170 to 2208C for 50 s, obtaining the nal product ready for packaging with a moisture content 5%. Extraction and clean-up. Unprocessed corn our, corn our mixture, and corn ake samples were analyzed by using the following three methods that differ in the extraction and clean-up procedures: (i) the AOAC of cial method (17) using methanol: water (3:1, vol/vol) as extraction mixture and SAX columns for the clean-up of the extracts; (ii) the Borax method (11) (slightly modi ed), consisting of the following: extraction of 20-g samples with 50 ml of methanol:boric acid (0.4 M H 3 BO 4, adjusted to ph 9.2 with 1 M NaOH) (3:1, vol/vol) by shaking for 30 min using an orbital shaker; clean-up of 10 ml of ltered extract through a SAX column; elution of the absorbed toxins with 10 ml of methanol:acetic acid (99:1, vol/vol), and evaporation to dryness under a nitrogen stream at 608C; (iii) the standards measurements and testing (SMT) method (14) developed in our laboratory with the support of the European Commission s SMT Program, which is brie y reported below: 20 g of sample was extracted with 50 ml methanol:acetonitrile:water (25:25:50, vol/vol/vol) by shaking for 20 min using an orbital shaker, then centrifuged for 10 min at 2,500 3 g, and ltered. The remaining solid material was extracted again with 50 ml of methanol:acetonitrile:water (25:25:50, vol/ vol/vol), centrifuged, and ltered as above. Ten milliliters of the combined extracts was mixed with 40 ml phosphate-buffered saline solution (136 mm NaCl, 8.5 mm Na 2 HPO 4, 1.5 mm KH 2 PO 4, 2.7 mm KCl, adjusted at ph 7 with concentrated HCl) and ltered through glass micro ber. A 10-ml volume (equivalent to 0.4 g test portion) was passed through a FumoniTest (Vicam, Watertown, Mass.) immunoaf nity column, and after washing the column with 10 ml PBS solution, fumonisins were eluted with 1.5 ml MeOH and collected in a 4-ml clean vial. The eluted extract was evaporated to dryness under nitrogen stream at 608C and retained at ;48C until high-performance liquid chromatography (HPLC) analysis. Experiments to test the effect of different additives on the analytical performances of the AOAC method (see below) were performed on corn our samples of company A mixed with different additives. In particular, 2% salt (NaCl), 8% sugar (sucrose), 2% salt 1 8% sugar, 7% iron (added as ferrous sulfate heptahydrate) were tested, corresponding to the concentrations frequently used for commercial production. FIGURE 1. Effect of industrial corn ake processing on fumonisin (FB 1 and FB 2 ) content determined by different analytical methods (AOAC, n 5 9; Borax, n 5 14; and SMT, n 5 3). Derivatization and HPLC analysis. The puri ed sample was redissolved in 200 ml acetonitrile:water (1:1, vol/vol) and 50 ml of extract were derivatized with 200 ml of o-phthaldialdehyde solution as reported by Sydenham et al. (17), or with 50 ml o- phthaldialdehyde for the SMT method (14), and mixed for 30 s with a vortex mixer. Exactly 3 min after the addition of o-phthaldialdehyde, 20 ml of derivatized solutions was injected into the HPLC system. All reagents were obtained from Sigma-Aldrich s.r.l. (Milan, Italy). Apparatus. The HPLC apparatus consisted of a Waters 625 LC quaternary pump (Waters, Milford, Mass.) equipped with a Rheodyne model 9125 injection valve (Rheodyne, Cotati, Calif.), a Perkin Elmer LC 240 or Jasco FP 1520 uorometer detectors, and a Turbochrom 4.0 data system (Perkin Elmer, Norwalk, Conn.). The analytical column was a reversed-phase Discovery C 18 (15 cm mm, 5 mm particles) (Supelco, Bellefonte, Pa.) preceded by a Rheodyne guard lter (0.5 mm). The mobile phase was a mixture of methanol:0.1 M sodium dihydrogen phosphate (15.6 g NaH 2 PO 4 2H 2 O in 1 liter HPLC-grade water) (77:23), adjusted to ph 3.35 with o-phosphoric acid, and ltered through a 0.45-mm lter membrane. Flow rate was 1.0 ml/min. The uorescence detector was set at 335 nm excitation and 440 nm emission. RESULTS AND DISCUSSION The results of the analysis of corn our and corn akes performed with the three methods are graphically summarized in Figures 1 and 2. Figure 1 shows that similar results were obtained with the three methods tested on the starting material (corn our, A 1 ) while quite different fumonisin levels were detected with the three methods from the relevant corn ake sample (A 2 ). In particular, 40 and 27% of the original fumonisins were recovered with the SMT and Borax method, respectively, while very low levels of fumonisins (,5% with respect to the original content) were detected with the AOAC method. Similar results were obtained in the analysis of corn akes ( nal product, B 3 ) provided by company B, as shown in Figure 2. The different extraction ef ciency of the three methods was con rmed by standard addition experiments performed on the sample A 2 (n 5 3), spiked with 150 ng/g FB 1 and 75 ng/g FB 2 and on sample B 3 (n 5 6) spiked with 1,400 ng/g FB 1 and

3 J. Food Prot., Vol. 64, No. 5 FUMONISIN CONCENTRATION IN CORN FLAKES 703 TABLE 1. Effect of additives on fumonisins extraction from corn our by using the AOAC method Fumonisin content (n 5 3) Additive None 2% NaCl 8% sucrose 2% NaCl 6 8% sucrose Ferrous sulfate heptahydrate a, not detected (,5 ng/g). FB SD (ng/g) a FB SD (ng/g) FIGURE 2. Fumonisin (FB 1 and FB 2 ) content of intermediate and nal products of corn ake industrial processing, as determined by different analytical methods (AOAC, n 5 3; Borax, n 5 3; and SMT, n 5 6). 700 ng/g FB 2. The best results were obtained with the SMT method, giving recoveries of 74 and 76% for FB 1 and FB 2, respectively in sample A 2 and 91 and 87% for FB 1 and FB 2, respectively, in sample B 3. The analysis performed on corn akes at an intermediate step (sample B 2 ) (Fig. 2) showed that only with the SMT method was it possible to extract FB 1 and FB 2 ef ciently from this matrix. This result was supported by standard addition experiments (n 5 6) performed on the same sample B 2, spiked with 2,600 ng/g FB 1 and 1,300 ng/g FB 2, giving recoveries of 99% for FB 1 and 94% for FB 2. Finally, the results of the analyses of samples B 1 (Fig. 2) showed that problems with the AOAC method already occur at the initial step of processing, i.e., when the additives are mixed with raw corn our, while the Borax method is still acceptable. Concerning the effect of processing on fumonisin decontamination, it can be concluded that about a 60 to 70% decrease of the fumonisin amounts present in the starting materials occurs during the entire cycle of corn ake processing. The low fumonisin recoveries obtained when naturally contaminated corn akes are analyzed by the AOAC method had been previously attributed to the presence of reduced iron (Fe 21 ) in some commercial corn akes brands (11, 13). Nevertheless, results of the present investigation (summarized in Table 1) indicate that, even in the absence of iron, a reduction of fumonisin levels is observed in the presence of other additives, when the AOAC method is used. The HPLC analysis showed that addition of sodium chloride (with or without sucrose) led to the complete loss of fumonisins that are not retained by the resin; this was con rmed by collecting the eluates from the SAX clean-up and analyzing them following the immunoaf nity clean-up of the SMT method. On the contrary, addition of sugar or ferrous sulfate heptahydrate (7% Fe 21 ) produced a less signi cant loss of the fumonisin content (Table 1). A possible explanation for the low fumonisin recovery in the speci c case of NaCl could be related to the ion exchange between the SAX resin and the Cl 2 ions, reducing the resin capacity toward fumonisins. From the comparison of different analytical methods it can be concluded that: (i) only the SMT method (based on double extraction with acetonitrile:methanol:water and immunoaf nity column clean-up) provides good results for the determination of fumonisins in unprocessed corn our as well as in the intermediate and nal products of corn ake industrial processing; (ii) the AOAC of cial method for corn is not applicable to the analysis of processed products containing additives such as salts or sugar, due to the failure of SAX columns to retain fumonisins during the clean-up step; (iii) physical parameters such as processing time and temperature are critical factors that affect fumonisin stability during corn ake production. Table 2 summarizes the results of fumonisin analyses of corn akes at different stages of processing obtained by two different producing companies. In particular, losses of 58 and 62% of FB 1 and FB 2, respectively, were observed after processing of corn akes by company A (matrix A). The results relevant to matrix B indicate that when the starting corn our (B 1 ) is heated at the extrusion temperature (70 to 1058C for 5 min), some fumonisin decontamination occurs, with losses of about 26% FB 1 and 32% FB 2. When dried corn akes (B 2 ) were brought to temperatures 2008C, i.e., during the roasting process, additional losses of fumonisins occurred, with total losses of 71% for FB 1 and 72% for FB 2 with respect to the amount present in the starting corn our. In summary, about 30 to 40% of fumonisins were recovered after the whole process of corn ake production by the two processing companies considered in the present study. However, corn ake processing may vary considerably from plant to plant depending on the time and temperature of cooking, the kind and amount of additives (salts, iron, vitamins, sugars, etc.), and on the quality of the raw material (corn variety, with or without germ and bran layers, etc.). These different parameters, while they are determinants for the quality of the nal product, may also affect the degree of fumonisin reduction during processing. This work has shown that the use of reliable analytical methods that are effective for both raw materials and processed products is of paramount relevance for studying the effect of food processing on mycotoxin-contam-

4 704 DE GIROLAMO ET AL. J. Food Prot., Vol. 64, No. 5 TABLE 2. Stability of fumonisins during different steps of corn ake industrial processing FB 1 (ng/g) a FB 1 recovery b (%) FB 2 (ng/g) a FB 2 recovery b (%) FB 1 1 FB 2 recovery b (%) Matrix A Corn our A 1 (starting material) Corn akes A 2 ( nal product) Matrix B Corn our B 1 (1NaC1 1 sucrose) 3, , Dried corn akes B 2 (not roasted) 2, Corn akes B 3 ( nal product) 1, a Mean fumonisin concentration 6 1 SD (Matrix A, n 5 3; Matrix B, n 5 6). b Mean fumonisins recovered after corn ake processing with respect to starting amount. inated commodities. Some effective fumonisin decontamination occurring during corn ake processing has been shown, although more work is needed to identify the thermal breakdown products of fumonisins and the relevant toxicity. ACKNOWLEDGMENTS This work was supported in part by the European Commission, contract QLK1-CT The valuable technical assistance of Giuseppe Panzarini is acknowledged. REFERENCES 1. Caramelli, M., A. Dondo, G. Cantini Cortellazzi, A. Visconti, F. Minervini, M. B. Doko, and F. Guarda Leucoencefalomalacia nell equino da fumonisine: prima segnalazione in Italia. Ippologia 4: Castelo, M. M., S. S. Sumner, and L. B. Bullerman Occurrence of fumonisins in corn based food products. J. Food Prot. 61: Doko, M. S., and A. Visconti Occurrence of fumonisins B 1 and B 2 in corn and corn-based foodstuffs in Italy. Food Addit. Contam. 11: Dupuy, J., P. Le Bars, H. Boudra, and J. Le Bars Thermostability of fumonisin B1, a mycotoxin from Fusarium moniliforme. Appl. Environ. Microbiol. 56: Gelderblom, W. C. A., N. P. J. Kriek, W. F. O. Marasas, and P. G. Thiel Toxicity and carcinogenicity of the Fusarium moniliforme metabolite, fumonisin B 1, in rats. Carcinogenesis 12: Harrison, L. R., B. M. Colvin, J. T. Greene, L. E. Newman, and J. R. Cole Pulmonary edema and hydrothorax in swine produced by fumonisin B 1, a toxic metabolite of Fusarium moniliforme. J. Vet. Diagn. Invest. 2: Jackson, L. S., J. J. Hlywka, K. R. Senthil, and L. B. Bullerman Effects of thermal processing on the stability of fumonisins, p In L. Jackson, J. De Vries, and L. B. Bullerman (ed.), Fumonisins in food. Plenum Press, New York. 8. Jackson, L. S., J. J. Hlywka, K. R. Senthil, B. L. Bullerman, S. M. Musser Effects of time, temperature, and ph on the stability of fumonisin B 1 in an aqueous model system. J. Agric. Food Chem. 44: Pascale, M., M. B. Doko, and A. Visconti Determinazione di fumonisine nella polenta mediante cromatogra a HPLC. Atti del II Congresso Nazionale di Chimica degli Alimenti. Giardini Naxos, Maggio Ross, P. F., L. G. Rice, R. D. Plattner, G. D. Osweiler, T. M. Wilson, P. E. Nelson, and J. L. Richard Concentration of fumonisin B 1 in feeds associated with animal health problems. Mycopathologia 114: Scott, P. M., and G. A. Lawrence Stability and problems in recovery of fumonisins added to corn-based foods. J. Assoc. Off. Anal. Chem. Int. 77: Shephard, G. S., P. G. Thiel, S. Stockenstrom, and E. W. Sydenham Worldwide survey of fumonisin contamination of corn and corn-based products. J. Assoc. Off. Anal. Chem. Int. 79: Solfrizzo, M., A. De Girolamo, and A. Visconti Determination of fumonisins B 1 and B 2 in corn akes by high performance liquid chromatography and immunoaf nity clean-up. Food Addit. Contam. 18: Solfrizzo, M., A. De Girolamo, A. Visconti, E. Sizoo, and H. P. van Egmond Determination of fumonisins B 1 and B 2 in corn based foodstuffs by high performance liquid chromatography and immunoaf nity clean-up in house and interlaboratory validation. Proceedings of the X International IUPAC Symposium on mycotoxins and phycotoxins, São Paulo, Brasil, May 2000, in press. 15. Stack, M. E., and R. M. Eppley Liquid chromatographic determination of fumonisins B 1 and B 2 in corn and corn products. J. Assoc. Off. Anal. Chem. Int. 75: Sydenham, E. W., G. S. Shepard, W. F. O. Marasas, D. J. Van Schalkwyk, K. R. Koch, and P. G. Thiel Natural occurrence of some Fusarium mycotoxins in corn from low and high esophageal cancer prevalence areas of Transkei, Southern Africa. J. Agric. Food Chem. 38: Sydenham, E. W., G. S. Shephard, P. G. Thiel, S. Stockenstrom, P. W. Snijman, and D. J. Van Schalkwyk Liquid chromatographic determination of fumonisins B 1, B 2, and B 3 in corn: AOAC- IUPAC collaborative study. J. Assoc. Off. Anal. Chem. Int. 79: U.S. National Toxicology Program NTP technical report on the toxicology and carcinogenesis studies of fumonisin B 1 (CAS No ) in F344/N rats and B6C3F 1 mice (feed studies), p NTP TR 496, National Institutes of Health Publication No (U.S. Department of Health and Human Services). 19. Visconti, A Fumonisins in maize genotypes grown in various geographic areas, p In L. Jackson, J. De Vries, and L. B. Bullerman (ed.), Fumonisins in food. Plenum Press, New York. 20. Visconti, A., W. F. O. Marasas, J. D. Miller, and R. Riley Mycotoxins of growing interest: fumonisins. Third Joint FAO/WHO/UNEP International Conference on Mycotoxins, Tunis [Internet, WWW]. Ad-

5 J. Food Prot., Vol. 64, No. 5 FUMONISIN CONCENTRATION IN CORN FLAKES 705 dress: mycoto/papers/myco5a.pdf. 21. Visconti A., M. Solfrizzo, A. De Girolamo, and A. Boenke Development/optimization of analytical methods for the measurement of fumonisins at levels of interest for future EU legislation. EU Report EN (European Commission, DG Science, Research and Development, Brussels, Belgium). 22. Wilson, T. M., P. R. Ross, D. L. Owens, L. G. Rice, S. A. Green, S. J. Jenkins, H. A. Nelson Experimental reproduction of ELEM. Mycopathologia 117: