Fumonisins and their masked forms in maize products

Food Control 59 (2016) 619e627

Contents lists available at ScienceDirect

Food Control journal homepage: www.elsevier.com/locate/foodcont

Fumonisins and their masked forms in maize products Marcin Bryła a, b, *, Marek Roszko a, Krystyna Szymczyk a, Renata Je˛ drzejczak a,
ski b Mieczysław W. Obiedzin a b

Institute of Agricultural and Food Biotechnology, Department of Food Analysis, Rakowiecka 36, 02-532, Warsaw, Poland Warsaw University of Life Sciences, Faculty of Food Sciences, Nowoursynowska 159, 02-776, Warsaw, Poland

a r t i c l e i n f o

a b s t r a c t

Article history: Received 18 December 2014 Accepted 16 June 2015 Available online 20 June 2015

In recent years many papers on masked mycotoxins in maize-based products appeared, including reports on fumonisins capable to form non-covalent bonds with food macro constituents. Such so-called “hidden fumonisins” are frequently present in food at quantities higher than the free forms. The aim of this work was to assess levels of free and total (free þ hidden) fumonisins (B1, B2 and B3) in 88 maize products available on the Polish retail market. Isotope dilution ion trap mass spectrometery coupled to a high performance liquid chromatography was used. 57% of all tested samples contained free fumonisins at concentrations above our limit of quantification LOQ (mean 390 ± 676 mg/kg). More than 77% of the samples contained free þ hidden fumonisins at concentrations above LOQ (mean 574 ± 1177 mg/kg). The highest mean fumonisins concentrations 1006 ± 1131 mg/kg, 1651 ± 2317 mg/kg, respectively for free and free þ hidden forms were observed in the group of maize snacks. The lowest fumonisin concentrations were found in maize-based starch concentrate products. None of the tested products within that group had free fumonisins concentrations above LOQ, while mean concentration of free þ hidden fumonisins was as low as 82 ± 42 mg/kg. In thermally processed products like corn flakes and various snacks the hidden-to-free fumonisin concentration ratio was higher than in unprocessed products like flour, groats or raw popcorn grains. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Fumonisins Hidden fumonisins Masked mycotoxins Maize products

1. Introduction Fumonisins are fungal secondary metabolites produced mainly by Fusarium verticilloides and Fusarium proliferatum, species
czuk, commonly found on maize (Pascale, Visconti, Pron Wi
sniewska, & Chełkowski, 1997; Warfield & Gilchrist, 1999; €rner, 2001). They belong to mycotoxins that might Weidenbo cause adverse health effects in mammals, including hepatotoxicity, nephrotoxicity, or cytotoxicity. It was proven that chronic exposure to high doses of fumonisin B1 may results in an increased risk of esophagus cancer development (Munkvold & Desjardins, 1997; Clements, Kleinschmidt, Maragos, Pataky, & White, 2003). In this respect IARC has classified FB1 to toxicity class 2B (IARC, 2002). Contamination of maize with fumonisins is commonly reported in southern Europe. This is expected to be related with the seasonal climate fluctuations (Pietri, Bertuzzi, Pallaroni, & Piva, 2004). In this

* Corresponding author. Institute of Agricultural and Food Biotechnology, Department of Food Analysis, Rakowiecka 36, 02-532, Warsaw, Poland. E-mail address: [email protected] (M. Bryła). http://dx.doi.org/10.1016/j.foodcont.2015.06.032 0956-7135/© 2015 Elsevier Ltd. All rights reserved.

respect European Union has set the maximum allowable levels of fumonisins (sum of FB1 and FB2) regarding unprocessed maize grain and some maize products (EC No 1126/2007). Maize grain intended for food purposes in Poland is mainly imported. However, in recent years also maize grain produced in the country was more and more frequently processed (CSO 2013). It was used to produce flour, groats, corn flakes, popcorn, snacks, gluten free bread, and infants formulas. Maize becomes infected by Fusarium during the growing season. Suitable temperature and humidity increase the infestation rate and result in larger production of fumonisins (Lazzaro et al., 2012). Grain processing (including cleaning) might decrease fumonisin concentration (Pascale et al., 1997; Humpf & Voss, 2004; FAO/WHO 2012; Bryła, Roszko et al., 2013). However, maize processing at elevated temperatures might lead to formation of masked fumonisins. Alkylated derivatives of fumonisins like N-carboxymethyl fumonisin B1 or N-deoxy-D-fructose-1-yl fumonisin B1 were among the first discovered compounds. Those are formed in reactions with reductive carbohydrates at elevated temperatures (e.g. during extrusion of maize grits at 160e180  C) (Seefelder, Hartl, & Humpf, 2001). In recent years some reports on formation of

620

M. Bryła et al. / Food Control 59 (2016) 619e627

fumonisins esterified with fatty acids were also published. Those were identified in Fusarium vericilloides in vitro cultures grown on
k, Sze
csi, Juha
sz, Barto
k, & Mesterha
zy, 2013; rice and maize (Barto
k, To € lgyesi, Mesterha
zy, Barto
k, & Sze
csi, 2010; Falavigna, Barto Cirlini, Galaverna, & Dall'Asta, 2012; Falavigna, Cirlini, Galaverna, Sforza, et al., 2012). N-acylated derivatives of FB1 were also identified in Fusarium cultures grown on rice. Concentrations of the derivatives were however significantly lower than concentration of
k et al., 2013, 2010). native forms (Barto Fumonisins non-covalently bound to food macro-constituents like proteins or carbohydrates are also regarded masked fumonisins. Such compounds were identified in maize grain and maize products. Contrary to fumonisins bound to other substances with covalent bonds these are called “hidden fumonisins” (Dall'Asta et al., 2009a; Dall’Asta, Falavigna, Galaverna, Dossen, & Marchelli, 2010; Falavigna, Cirlini, Galaverna, & Dall'Asta, 2012). Concentration of these compounds might well exceed concentration of the free forms, and significantly contribute to profile of masked fumonisins (Dall’Asta, Mangia et al., 2009, Dall’Asta, Galaverna et al., 2009, Dall’Asta et al., 2010; Di Mavungu & De Saeger, 2011; Falavigna, Cirlini, Galaverna, & Dall'Asta, 2012). In addition it is believed that hidden fumonisins might be released inside the human gastrointestinal tract, therefore it seems important to assess their actual levels in food and feed for toxicological reasons (Falavigna, Cirlini, Galaverna, & Dall'Asta, 2012). In this study concentrations of free and masked fumonisins have been evaluated in various group of low- and medium-processed maize products available on the Polish retail market. Free fumonisin forms, hidden esterified fumonisins, as well as fumonisins covalently bound via carboxyl group to other substances present in food easily hydrolyze in alkaline media, releasing its side chain off the native fumonisin molecule. Such reaction is depicted in Fig. 1. 2. Materials and methods 2.1. Chemicals and reagents Only solvents of HPLC grade supplied by Rathburn (Walkerburn, UK) were used in this study. Formic acid, acetic acid, potassium hydroxide of prior to analysis grade were supplied by POCH (Gliwice, Poland). Molecularly imprinted polymeric SPE cartridges (FumoZON AFFINIMIP) were provided by Polyintell (Val de Reuil, France). 2.2. Samples Fumonisins were determined in 88 samples of commercially available maize-based food products from a variety of manufacturers bought in local supermarkets in 2013, including groats (n ¼ 15), starch concentrates (n ¼ 6), noodles (n ¼ 14), flour

(n ¼ 20), corn flakes (n ¼ 19) and maize snacks (n ¼ 14). Sampling was performed on a random basis. Prior to analysis the samples were stored at room temperature except for bread samples which were stored frozen below minus 18  C. 2.3. Standards and reference materials Native standards of FB1, FB2 & FB3 (50 mg/mL), labeled 13C34-FB1 (25 mg/mL), 13C34-FB2, 13C34-FB3 (10 mg/mL), and hydrolyzed HFB1 (25 mg/mL) were supplied by Biopure (Tulln, Austria). HFB2 and HFB3 standards were synthetized in a laboratory scale by alkaline hydrolysis of native FB2 and FB3's. Briefly, FBs standard solutions were transferred into a test tube with a ground glass joint. Solvent was evaporated under a gentle stream of nitrogen, then 2 M KOH aqueous solution was added. Digestion was performed for 24 h at room temperature. The solution was quantitatively transferred with acetonitrile into a separator funnel. Aqueous layer was extracted two more times with acetonitrile, organic extracts were combined evaporated to dryness using a rotary evaporator operated at 40  C, then re-dissolved in acetonitrile. For further calculations a complete hydrolysis of fumonisins was assumed. Biopure (BRM 003017) certified reference material of maize was used in validation experiments. 2.4. Analytical procedures Free and hidden fumonisins were determined using our previously published method without any major modifications (Bryła, Je˛ drzejczak et al., 2013; Bryła et al., 2014; Bryła, Szymczyk,
ski, 2015). Je˛ drzejczak, & Obiedzin 2.4.1. Free FBs Briefly, samples were ground using a laboratory grinder. 2.5 g of sample was transferred into a glass beaker, spiked with 10 ml of the solution with 13C34-FB1, 13C34-FB2 and 13C34-FB3 labeled internal standards and homogenized with 10 mL of methanol: acetonitrile: water solution 25:25:50 (v/v/v) for 3 min. The solution was centrifuged at 10,730  g, 5 mL of the supernatant was mixed with 5 mL of water and used for analysis. Extracts were purified using polymeric SPE's. Cartridges were conditioned with 2 mL of acetonitrile and 2 mL of water at 3e4 droplets per second flow rate. Two 4 mL portions of the diluted extract were loaded on the cartridge. Cartridge was washed with 6 mL of acetonitrile: water mixture 40:60 (V/V) and subsequently eluted with 4 mL of formic acid: methanol solution 2:98 (V/V) into a 25 mL round bottom flask. Solution was evaporated to dryness using a rotary evaporator and re-dissolved in 1 mL of methanol: water: acetic acid mixture (10:89.9:0.1) and sonicated in an ultrasound bath to improve transfer of the analyte and sample dissolution. Samples were

Fig. 1. Degradation of fumonisin molecules in alkaline media.

M. Bryła et al. / Food Control 59 (2016) 619e627

filtered through a 0.2 mm nylon syringe filter, transferred into glass vials, and analyzed in an LC-MS/MS system.

2.4.2. Total FBs 0.5 g of a well-ground sample was weighted into a PP centrifuge tube and spiked with 10 ml of labeled FB1, FB2 and FB3 internal standard solution. Subsequently 10 mL of 2 M KOH solution was added and the sample was hydrolyzed for 24 h at room temperature. The solution was shaken with 12.5 mL of dichloromethane and centrifuged at 10,730  g. 5 mL of the organic extract was transferred into a round bottom flask and evaporated to dryness using a rotary evaporator operated at 40  C. Dry residues were dissolved in 1 mL of methanol: water: acetic acid solution (30:69.9:0.1) (v/v/v), filtered through a 0.2 mm nylon syringe filter, transferred into glass vials and analyzed in an LC-MS/MS system.

2.4.3. Free HFBs 5 g of a well-ground sample was homogenized with 20 mL of methanol: acetonitrile: water solution (25:25:50 v/v/v) for 3 min. Solution was transferred into a 50 mL PP centrifuge tube and centrifuged at 10,730  g. Supernatant was transferred into a 50 mL separator funnel and mixed with 20 mL of n-hexane. The solution was shaken vigorously several times to extract the lipid fraction and allowed to rest until the phases separated completely. 0.8 mL of the lipid-free aqueous layer was transferred into a round bottom flask. Solvent was evaporated to dryness using rotary evaporator and the dry residue was re-dissolved in 1 mL of methanol: water: acetic acid (30:69.9:0.1) mixture. Dry residues were dissolved in 1 mL of methanol: water: acetic acid solution (30:69.9:0.1) (v/v/v), filtered through a 0.2 mm nylon syringe filter, transferred into glass vials and analyzed in an LC-MS/MS system. Free HFB's were quantified using the external calibration method. All determinations were performed in triplicates.

2.5. LC-MS/MS High performance liquid chromatograph Accela ThermoFinnigan (Austin, TX, USA) equipped with an auto injector and coupled to an ion trap mass spectrometer LTQ Thermo-Finnigan (Austin, TX, USA) were used for the purpose of this study. Chromatographic separations of free fumonisins were performed on a Kinetex PFP 100 mm  2.1 mm, 2.6 mm (Phenomenex, Torrance, CA, USA) column. Mobile phase was composed of water (A), methanol (B) and 2% aqueous acetic acid solution (C). The following mobile phase gradient was applied: 0e4 min 60% A, 30% B, 10% C, 14e28 min 90% B, 10% C, 32e55 min 60% A, 30% B, 10% C. Hydrolyzed fumonisins were separated using Luna C-18 150 mm  2.0 mm, 3 mm (Phenomenex, Torrance, CA, USA) column. Mobile phase was composed of water (A) and methanol (B) both containing 0.1% of acetic acid. The following mobile phase gradient was applied: 0e3 min 70% A, 7e20 min 35% A, 21e31 min 100% B, 33e48 min 70% A. Mobile phase flow rate in all cases was set at 0.15 cm3 min. Electro spray ion source was used in the ion trap mass spectrometer for the analyte ionization. Nitrogen was used as the nebulizing gas with flow rate set at 20 a.u., auxiliary gas flow rate was set at 10 a.u. Helium was used as the ion trap dumping gas. The m/z values corresponding to the fumonisins pseudo-molecular ions [MþH]þ were used as precursors in collision-induced ionization (CID) experiments. The detailed ion trap operating conditions and the optimized ionization parameters are show in Table 1. Xcalibur 1.2 computer data system was used for the instrument control, data acquisition and analysis.

621

2.6. Data analysis Statistical evaluations were performed using the Statgraphics 4.1 software package (Graphics Software System, STCC, Inc., Rockville, MD, USA). One-way analysis of variance was used to assess the significance of the differences in the determined fumonisins concentrations. Tuckey's test at a ¼ 0.05 was used in paired tests. 2.7. Validation experiments Laboratory-made spiked samples and the certified reference materials were used to assess the method statistical parameters. Concentrations corresponding to the lowest calibration points were used as the method limit of quantification (LOQ). Signal-to-noise ratio of the analyte chromatographic peaks calculated for the concentration equal to LOQ was not less than 10. 2.7.1. Native FBs calibration solutions Matrix-matched calibration was used for quantification. Standard solutions used to plot the calibration curve were prepared by dilution of FB1, FB2 and FB3 stock (5 mg/mL) standards. The diluted native standard were then added to the samples previously cleaned up using MIP solid phase extraction cartridges and evaporated to dryness. Labeled internal standards 13C34-FB1, 13C34-FB2, 13C34-FB3 were also added to the samples. Final concentration of the standards was the same as concentration in other samples. After standard addition the residual solvent was evaporated to dryness using a gentle stream of nitrogen and re-dissolved in 1 mL of the methanol: water: acetic acid mixture (10:89.9:0.1). Calibration curves were plotted as relations of native fumonisin concentrations and the corresponding ratios of internal standard peak: fumonisin peak area. 2.7.2. Calibration solutions used to determine total fumonisins The matrix-matched calibration method was used in the determinations. Calibration solutions of HFB1, HFB2 and HFB3 were prepared from the fumonisin stock (5 mg/mL) standard solution. Samples were spiked with known amount of native fumonisin standard and the corresponding labeled internal standard. Samples were subsequently hydrolyzed as described previously. Matrix (maize, flour, flakes or groats) free of significant amount of fumonisins was used for the standard preparation. Samples were further processed as described previously. The quotient of molar masses was used as a conversion factor for calculation of analyte concentration before the hydrolysis. 2.7.3. Calibration solutions used to determine free hydrolyzed fumonisins The matrix-matched calibration method was used to determine free HFBs. Known amount of the standard solution was added to the dry extract previously evaporated using a rotary evaporator. Sample extract used for the calibration has no significant quantities of free hydrolyzed fumonisins determined. Calibration curves were plotted as simple relations of the HFBs concentrations against analyte peak areas. Calculated linear fits to calibration curves are shown in Table 2. 2.7.4. Analytical method performance Laboratory-made spiked samples were used to assess statistical parameters (recovery and repeatability) of the method. Ground maize grain and maize flours with no detectable amounts of the analytes were used in the validation experiments. Spiked samples were prepared at three concentration levels. All samples were analyzed in triplicates. Native fumonisin concentrations in the spiked samples were 200, 500 and 800 mg/kg (each of FB1, FB2 and

622

M. Bryła et al. / Food Control 59 (2016) 619e627

Table 1 ESI-MS/MS acquisition parameters. Compound

FB1 FB2 FB3 13 C-FB1 13 C-FB2 13 C-FB3 HFB1 HFB2 HFB3 13 CeHFB1 13 CeHFB2 13 CeHFB3

Precursor ion

Daughter ions

Capillary voltage

Spray voltage

Isolation width

Excitation time

Excitation voltage

[m/z]

[m/z]

[V]

[kV]

[m/z]

[ms]

[%]

722 706 706 756 739 739 406 390 390 428 412 412

704 688 688 738 722 722 388 372 372 410 394 394

19 19 19 19 19 19 4 4 4 4 4 4

5 5 5 5 5 5 5 5 5 5 5 5

±4 ±4 ±4 ±4 ±4 ±4 ±3 ±3 ±3 ±3 ±3 ±3

30 30 30 30 30 30 30 30 30 30 30 30

30 30 30 30 30 30 29 30 30 29 30 30

þ þ þ þ þ þ þ þ þ þ þ þ

546 530 530 574 558 558 370 354 354 392 376 376

Table 2 Linear fits to fumonisins calibrations curves.

FBs

HFBs*(A)

HFBs*(B)

Compound

Calibration curve equation

FB1 FB2 FB3 HFB1 HFB2 HFB3 HFB1 HFB2 HFB3

Y Y Y Y Y Y Y Y Y

¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼

Determination coefficient [R2]

0.000819896X þ 0.126409 0.00107953X þ 0.122665 0.00098847X þ 0.123744 0.437126X  46.392 1.28504X  150.325 1.13707X  153.558 335522X 189517X  285874 181698X  459088

0.9911 0.9920 0.9953 0.9941 0.9950 0.9956 0.9855 0.9973 0.9844

*HFBs (A) e hydrolyzed fumonisins released after alkaline digestion. HFBs (B) e free hydrolyzed fumonisins present in the extract.

FB3). Concentrations of hydrolyzed fumonisins in samples evaluated after alkaline digestion were 280, 560, 1120 mg/kg(HFB1) and 275, 550, 1100 mg/kg (HFB2 & HFB3). Samples used to determine free hydrolyzed fumonisins were spiked at 125, 500, 1000 mg/kg (HFB1) and 15, 56, 112 mg/kg (HFB2 & HFB3). Mean determined

analyte recovery values (R), recovery repeatability expressed as recovery relative standard deviation (RSD), and calculated limit of quantification (LOQ) are shown in Table 3. Validation experiments have proved that statistical parameters of the method were satisfactory. Recovery rates of FB1 and

Table 3 Recovery (R), recovery relative standard deviation (RSD) and limit of quantification (LOQ) evaluated in validation experiments.

FBs

Compound

Fortification level [mg/kg]

Counts n

Recovery R [%]

RSD [%]

Mean recovery R [%]

Mean RSD [%]

LOQ [mg/kg]

FB1

200 500 800 200 500 800 200 500 800 280 560 1120 275 550 1100 275 550 1100 125 500 1000 15 56 112 15 56 112

4 4 4 4 4 4 4 4 4 5 5 5 5 5 5 5 5 5 3 3 3 3 3 3 3 3 3

109 98 99 111 102 105 94 102 96 94 98 89 97 95 106 95 100 98 74 107 88 90 83 88 86 82 90

12.0 10.7 11.8 15.1 11.7 10.2 17.4 9.4 7.7 13.1 10.2 9.3 13.0 11.3 10.3 17.2 11.6 7.2 19.5 12.2 11.1 21.1 18.0 15.3 26.3 18,1 17,2

102

11.5

12.5

106

12.3

12.5

97

11.5

12.5

94

10.9

22.0

99

11.5

22.0

98

12.0

22.0

90

14.3

12.5

87

18.1

14.0

86

20.5

14.0

FB2

FB3

HFBs*(A)

HFB1

HFB2

HFB3

HFBs*(B)

HFB1

HFB2

HFB3

*HFBs (A) e hydrolyzed fumonisins released after alkaline digestion. HFBs (B) e free hydrolyzed fumonisins present in the extract.

M. Bryła et al. / Food Control 59 (2016) 619e627

FB2 were close to 100% at recovery relative standard deviations RSD between 11.5% and 12.3%. Recovery rates of hydrolyzed fumonisins were in the 86e99% range at RSD between 10.9% and 20.5%; those latter parameters are also considered satisfactory. 2.7.5. Method trueness 2.7.5.1. Certified reference material. Trueness of the developed analytical method was assessed using a commercially available certified reference material (CRM). The results are shown in Table 4. Extended uncertainty of the results obtained in this work was calculated according to the EURACHEM/CITAC (Ellison & Williams, 2012) guidelines. Differences between the certified and determined concentrations were lower than the combined uncertainties of both analyzed variables. On that basis one can conclude that the differences were statistically insignificant. The certified reference material sample was also assessed in respect to the total fumonisins contents. The calculated concentration value was also compared with the certified free fumonisins contents. It was found that the concentrations of FB2 and FB3 calculated after an alkaline digestion were not different from the certified concentrations. In case of FB1 the calculated difference between the certified and calculated concentrations was significantly higher than the combined uncertainty of both variables. It was assumed that the concentration of the certified concentration was significantly different from the determined concentration. 2.7.5.2. Proficiency testing. Participation in a proficiency testing programme was part of the validation study. Evaluation was conducted between September and October 2013 within the Food Analysis Performance Assessment Scheme (FAPAS). A ground maize grain sample was tested. FB1 and FB2 concentrations determined in the tested sample were 706 and 299 mg/kg, respectively. The calculated z-score values were in the 0.3 to 0.6 range (considered satisfactory). 3. Results and discussion Summary data on concentrations of free and total FBs found in the tested samples are shown in Table 5 and Figs. 2e7. Either free or total FBs were observed in all tested groups of maize-based food products. Frequency of occurrence of free FBs ranged from 0 to 86% of samples in various product groups, on the average amounted to 57%. Frequency of occurrence of total FBs ranged from 50 to 100%, on the average amounted to 77%. Mean summary concentration of total FBs in the tested samples was 574 ± 1177 mg/kg, a value statistically higher than 390 ± 676 mg/ kg, mean summary concentration of free FBs determined for the same sample set. The highest mean concentrations were found in maize snacks: 1006 ± 1131 mg/kg (free FBs) and 1651 ± 2317 mg/kg (total FBs).

623

Samples with the highest absolute concentrations of the studied compounds (3297 mg/kg free FBs, 7331 mg/kg total FBs) were also found in that group of products. On the other hand, the lowest FBs concentrations were observed in starch concentrates. Concentration of free FBs above LOQ was found in none of the samples of that group, mean concentration of total FBs was 82 ± 42 mg/kg. Allowable limits of free FB1 þ FB2 concentrations set by the EC 1126/2007 regulation were not exceeded in any of the tested samples. In any case the limit was also exceeded when FB3 was included in calculations. However, when hidden fumonisins and FB3 were taken into account, the summary concentration exceeded the limit in six samples (see Figs. 2e7). Significantly higher concentrations of total FBs were found in only 2 out of 7 tested popcorn samples, see Fig. 7. No significant concentrations of hydrolyzed free fumonisins (HFBs) have been found. Total FBs concentrations were in the LOQ450 mg/kg range in majority (49) of the tested samples. Distribution of total fumonisins concentrations found in the tested samples is shown in Fig. 8 as a frequency histogram. Kim, Scott, and Lau (2003) and Park, Scott, Lau, and Lewis (2004) have noticed that FBs concentrations determined in alkali-digested food products are significantly higher than those determined using traditional extraction-only based methods. Originally it was explained by some interactions occurring between fumonisins' TCA groups and proteins/carbohydrates in food matrix formed during high temperature food processing. The formed bonds could limit the possibility of fumonisin extraction from the matrix. In recent years it was however found that FBs concentration determined in alkali-digested raw maize grain is also higher than that determined after extrac only. It indicates that fumonisins in complex food matrices are bound to macro molecules (like starch or proteins) by means of some non-covalent interactions rather than by means of some covalent bonds (Dall’Asta, Mangia et al., 2009, Dall’Asta, Galaverna et al., 2009, Dall’Asta et al., 2010). Despite the fact that hidden fumonisins have been found also in unprocessed maize, a possibility of formation of covalent bonds between fumonisins' TCA groups and proteins/starch molecules during heat treatment should not be ignored. Seefelder, Knecht, and Humpf (2003) have concluded (on the basis of some model experiments and some acquired ESI-MS and NMR spectra) that such covalent interactions are indeed possible. Up to now there are not many pieces of data on intensity of such reactions. However it was found that fumonisin affinity to starch is significantly higher than to proteins. 68 samples altogether were fumonisin-positive. However, the hidden-to-free FBs ratios could have been calculated only for those 50 samples, in which both free and hidden fumonisins were found above the LOQ level (doubly positive samples; in other 18 samples only total FBs could have been quantified but not free FBs, see Table 5). The ratio as large as 25.85 was found in one extreme case, but majority (39) of the samples showed ratios within the 0e2.8

Table 4 Results of determination of a fumonisin CRM. Compound

FB1 FB2 FB3

Reference concentration

Determined concentration

Free FBs

Free FBs

Total (free þ hidden FBs)

CCRM [mg/kg]

UCRM [mg/kg]

n

C [mg/kg]

U [mg/kg]

n

C [mg/kg]

U [mg/kg]

n

2630 690 310

740 340 210

20 20 20

2554 574 330

231 53 30

7 7 7

3618 894 382

266 71 30

7 7 7

CCRM e certified concentration. C e determined concentration. UCRM e extended combined uncertainty of the certified concentration (k ¼ 2). U e calculated extended combined uncertainty (k ¼ 2). n e repetitions count.

624

M. Bryła et al. / Food Control 59 (2016) 619e627

Table 5 Free and total (free þ hidden) fumonisins in various tested maize-based food products: mean concentrations, standard deviations, medians, min/max values, hiddeneto-free ratios. Product

Sample count Free SFBs Positive samples

a

8 (53%) 0 (0%) 12 (60%) 12 (86%) 9 (47%) 9 (64%) 50 (57%)

309 e 428 175 85 1006 390

Mean SD

Median Min Max

Positive samples

a

260 e 58 83 18 523 130

8 (53%) 3 (50%) 17 (85%) 14 (100%) 15 (79%) 11 (79%) 68 (77%)

743 82 428 312 204 1651 574

mg/kg Groats Starch concentrates Flour Noodles Corn flakes Maize snacks Total a b c

15 6 20 14 19 14 88

b,c

Total (free þ hidden) SFBs Mean SD

Hidden-to-free ratio

Median Min Max

mg/kg 236 e 628 245 84 1131 676

39 e 13 13 13 13 13

776 e 1688 759 248 3297 3297

811 42 715 378 215 2317 1177

498 109 85 175 126 481 147

42 22 22 26 26 27 22

2766 114 2543 1278 745 7331 7331

1.47 e 1.95 3.93 5.95 1.87 3.06

(0.08÷4.62) (0.06÷8.08) (0.10÷25.85) (1.29÷14.00) (0.34÷8.85) (0.06÷25.85)

Mean concentration of fumonisins among positive samples only. Concentration of hidden SFBs calculated as the difference between determined concentration of total and determined concentration of free SFBs. Ratio of means among doubly positive samples only. Min/max ratios shown in brackets.

Fig. 2. Concentrations (±SD) of free and total (free þ hidden) FBs determined in the tested maize groats samples (letters mark statistically homologous groups).

Fig. 3. Concentrations (±SD) of free and total (free þ hidden) FBs determined in the tested maize starch concentrates samples (letters mark statistically homologous groups).

Fig. 4. Concentrations (±SD) of free and total (free þ hidden) FBs determined in the tested maize noodle samples (letters mark statistically homologous groups).

M. Bryła et al. / Food Control 59 (2016) 619e627

625

Fig. 5. Concentrations (±SD) of free and total (free þ hidden) FBs determined in the tested maize flour samples (letters mark statistically homologous groups).

Fig. 6. Concentrations (±SD) of free and total (free þ hidden) FBs determined in the tested corn flake samples (letters mark statistically homologous groups).

range. The mean for the entire population was 3.06. Mean calculated for 39 samples (excluding 11 extreme values) was equal to 1.80. Distribution of the ratios is shown in Fig. 9. Higher levels of fumonisins were observed in majority of the tested samples of groats, starch concentrates, and maize flour if samples were alkali-digested, at least in 10 cases twice as high as in the non-digested samples. For products processed solely by means of grain milling (like flour of groats) it is not possible to conclude whether the bound fumonisins liberated by the digestion were formed during grain processing or else if they were already present in raw grain (the latter seems to be more likely). Hidden FBs in raw maize grain were reported by Dall’Asta et al. (2010). They found statistically significantly higher concentrations of hidden fumonisins in 65% out of investigated 31 maize samples from different regions of Italy. However, they reported total concentrations within

the 1135e40,821 mg/kg range, i.e. significantly higher values than those reported in this study. The hidden-to-free FBs ratios were in the 0e1.40 range. Similar data on hidden fumonisins in unprocessed maize were also reported by Dall’Asta et al. (2009a, 2009b) and Mangia (2009). Literature data on hidden fumonisins in milled maize products are still limited. Maize grain intended for food is subjected to dry or wet cleaning. The processing could reduce concentration of fumonisins in grain (Pascale et al., 1997; Humpf & Voss, 2004; FAO/WHO 2012). It is very likely that free and hidden fumonisins are reduced at some different rates, so significant changes in the hidden-to-free FBs ratio may be expected. The noted very low concentrations of free FBs (below LOQ) and low (22e114 mg/kg) concentrations of hidden FBs in starch concentrates may result from specific nature of maize starch

Fig. 7. Concentrations (±SD) of free and total (free þ hidden) FBs determined in the tested maize snack samples (letters mark statistically homologous groups). Maximum concentrations allowed by EU regulations (4000 and 800 mg/kg for popcorn and other products, respectively) are indicated.

626

M. Bryła et al. / Food Control 59 (2016) 619e627

Fig. 8. Frequency histogram of total FBs found in the tested 88 samples.

production process, which involves not only dry cleaning of the grain, but also soaking it in some acidic aqueous solution to eliminate soluble grain constituents. Such process may eliminate a large part of water-soluble fumonisins from the grain (SCF 2000). Available literature data on hidden fumonisins in corn flakes are still limited. Meister (2001) reported higher FBs concentrations in alkali-digested corn flake samples. Kim et al. (2003) investigated hidden/bound FB1 in 20 corn flake samples and reported rather low levels (101 mg/kg), however the hidden-to-free FBs ratio was 2.6 on the average. It should be emphasized that FB1 was determined in proteins isolated from the samples by means of extraction with 1% sodium dodecyl sulphate aqueous solution. It is possible that initial FBs concentration in the samples was higher due to possible interactions of FBs with other food macro constituents, including starch. This hypothesis is supported by model studies done by Seefelder et al. (2003) indicating that FBs' TCA groups bind to starch hydroxyl groups more efficiently than to proteins. Similar results have also been reported by Park et al. (2004) who evaluated 15 samples of breakfast cornflakes and found that total mean summary concentration of hidden/bound FB1 was twice as high as concentration of FB1 bound to proteins only. FB1 level in the tested samples was low (173 mg/kg), but the hidden-to-free FBs ratio was as high as 1.83. The results reported in this study are in line with the above reports. The hidden-to-free FBs ratios found in this study for unprocessed popcorn agree with literature data, which are around 1.0 (Dall’Asta, Mangia et al., 2009, Dall’Asta, Galaverna et al., 2009, Dall’Asta et al., 2010; Mangia, 2009). Results of this study confirm other authors' suggestions that some FBs masking mechanisms must be at work while maize plants grow in the field. In the maize wafer samples tested in this study the total FBs concentration was

low and did not exceeded 435 mg/kg. Hidden-to-free FBs ratios were high (2.54 and 8.85) as compared to other tested low-processed products (popcorn, flour, groats). Just like maize flakes, maize wafers are thermally processed during production. Elevated FBs ratios were observed for products of this type (mean 5.95 for maize flakes, 5.7 for maize wafers). Lower ratios were observed in products subjected only to dry/wet cleaning (maize groats, starch concentrates, flours, popcorn). A higher hidden to free FBs concentration ratios in thermally processed products comparing to low-processed products have been already reported by the Mangia (2009). Results of some studies aimed to explain this phenomenon have been already reported by Bryła et al. (2014). In the latter work an increase in the FBs concentration ratios in baked gluten-free corn bread was observed. On this basis it has been suggested that such differences were related to the higher thermal stability of the hidden fumonisins comparing to the free forms. 4. Conclusions Classic sample preparation methods based only on extraction might lead to underestimation of levels of fumonisins as determined in food products. It was shown that indirect methods based on alkaline digestion allow to determine the masked (hidden) fumonisins forms as well. Unfortunately such methods are non-selective. Thermally processed maize food products like flakes and some snacks exhibit higher ratio of hidden-to-free fumonisins than low-processed products like flour, groats or unprocessed popcorn grain. In this respect behavior of fumonisins during food processing (especially during heat treatment) should be investigated in some greater detail, since food

Fig. 9. Frequency histogram of hidden-to-free FBs ratios found in 50 doubly-positive samples.

M. Bryła et al. / Food Control 59 (2016) 619e627

processing technological operations involving heat treatment may significantly affect levels of hidden and free fumonisins. Little is known about toxicological profile of fumonisin conjugates formed during food processing and it seems necessary to identify the formed compounds. Some in vitro studies have proven that hidden fumonisins bound to food macromolecules by means of some non-covalent interactions might be released to free FBs within human gastrointestinal tract. Acknowledgment This study was financially supported by the 2012/05/N/NZ9/ 01316 National Science Centre grant. References
Juha
(2013). ESI-MS and MS/
k, T., Sze
csi, A.,
sz, K., Barto
k, M., & Mesterha
zy, A. Barto MS identification of the first ceramide analogues of fumonisin B1 mycotoxin from a Fusarium verticillioides culture following RP-HPLC separation. Food Additives and Contaminants, Part A, 30(9), 1651e1659.
Barto
(2010). Identification
k, T., To €lgyesi, L., Mesterha
zy, A.,
k, M., & Sze
csi, A. Barto of the first fumonisin mycotoxins with three acyl groups by ESI-ITMS and ESITOFMS following RP HPLC separation: palmitoyl, linoleoyl and oleoyl EFB1 fumonisin isomers from a solid culture of Fusarium verticillioides. Food Additives and Contaminants, Part A, 27, 1714e1723.
ski, M. W., Se˛ kul, J., Bryła, M., Je˛ drzejczak, R., Roszko, M., Szymczyk, K., Obiedzin et al. (2013). Application of molecularly imprinted polymers to determine B1, B2, and B3 fumonisins in cereal products. Journal of Separation Science, 36, 578e584.
ski, M. W., & Se˛ kul, J. Bryła, M., Roszko, M., Szymczyk, K., Je˛ drzejczak, R., Obiedzin (2013). Fumonisins in plant-origin food and fodder e a review. Food Additives and Contaminants, Part A, 30(9), 1626e1640.
ski, M. W. Bryła, M., Roszko, M., Szymczyk, K., Je˛ drzejczak, R., Slowik, E., & Obiedzin (2014). Effect of baking on reduction of free and hidden fumonisins in glutenfree bread. Journal of Agricultural and Food Chemistry, 62(42), 10341e10347.
ski, M. W. (2015). Free Bryła, M., Szymczyk, K., Je˛ drzejczak, R., Słowik, E., & Obiedzin and hidden fumonisins in various fractions of maize dry milled under model conditions. LWT e Food Science and Technology, 64, 171e176. Clements, M. J., Kleinschmidt, C. E., Maragos, C. M., Pataky, J. K., & White, D. G. (2003). Evaluation of inoculation techniques for Fusarium ear rot and fumonisin contamination of maize. Plant Disease, 87, 147e153. CSO. (2013). Central statistical office. Results of plant production in 2013. Poland, Warsaw: Agriculture Division. http://stat.gov.pl/obszary-tematyczne/rolnictwo-lesnictwo/ uprawy-rolne-i-ogrodnicze/wyniki-produkcji-roslinnej-w-2013-r-,6,10.html. Dall’Asta, C., Falavigna, C., Galaverna, G., Dossen, A., & Marchelli, R. (2010). In vitro digestion assay for determination of hidden fumonisins in maize. Journal of Agricultural and Food Chemistry, 58, 12042e12047. Dall’Asta, C., Galaverna, G., Mangia, M., Sforza, S., Dossena, A., & Marchelli, R. (2009). Free and bound fumonisins in gluten-free food products. Molecular Nutrition and Food Research, 53, 492e499. Dall’Asta, C., Mangia, M., Berthiller, F., Molinelli, A., Sulyok, M., Schuhmacher, R., et al. (2009). Difficulties in fumonisin determination: the issue of hidden fumonisins. Analytical and Bioanalytical Chemistry, 395, 1335e1345. Di Mavungu, & De Saeger, S. (2011). Masked mycotoxins in food and feed: Challenges and analytical approaches. In Determining mycotoxins and mycotoxigenic fungi in food and feed (Vol. 2, pp. 385e400). Cambridge (UK): Woodhead Publishing. Ellison, S. L. R., & Williams, A. (2012). Eurachem/CITAC guide: Quantifying uncertainty in analytical measurement (3rd ed.) 978-0-948926-30-3.

627

European Commission. (2007). Commission regulation (EC) no 1126/2007 of 28 September 2007 amending regulation (EC) no 1881/2006 setting maximum levels for certain contaminants in foodstuffs as regards Fusarium toxins in maize and maize products. Falavigna, C., Cirlini, M., Galaverna, G., & Dall'Asta, C. (2012). Masked fumonisins in processed food: co-occurrence of hidden and bound forms and their stability under digestive conditions. World Mycotoxin Journal, 5(3), 325e334. Falavigna, C., Cirlini, M., Galaverna, G., Sforza, S., Dossena, A., & Dall'Asta, C.h (2012). LC/ESI-MS/MS analysis outlines the different fumonisin patterns produced by F. verticillioides in culture media and in maize kernels. Journal of Mass Spectrometry, 47, 1170e1176. FAO/WHO. (2012). FAO/WHO-Joint FAO/WHO food standards programme Codex committee on contaminants in foods. 2012. Sixth Session, Discussion Paper on Proposed Draft Maximum Levels for Fumonisins in Maize and Maize-Products and Associated Sampling Plans, Codex Alimentarius Commission. Available from: http://ec.europa.eu/food/fs/ifsi/eupositions/cccf/docs/cccf_6th_agenda_ item_12a.pdf. Humpf, H. U., & Voss, K. A. (2004). Effects of thermal food processing on the chemical structure and toxicity of fumonisin mycotoxins. Molecular Nutrition and Food Research, 48, 255e269. International Agency for Research on Cancer, IARC. (2002). Monographs, 82, 301e366. http://monographs.iarc.fr/ENG/Monographs/vol82/mono82-7B.pdf. Kim, E. K., Scott, P. M., & Lau, B. P. Y. (2003). Hidden fumonisin in maize flakes. Food Additives and Contaminants, 20(2), 161e169. , G., Ritieni, A., Ferracane, R., Marocco, A., et al. (2012). Lazzaro, I., Susca, A., Mule Effects of temperature and water activity on FUM2 and FUM21 gene expression and fumonisin B production in Fusarium verticillioides. European Journal of Plant Pathology, 134, 685e695. Mangia, M. (2009). Free and hidden Fumonisins in maize and gluten-free products. Ph.D. dissertation http://dspace-unipr.cilea.it/bitstream/1889/1449/1/Mattia_ Mangia_PhD.pdf. Meister, U. (2001). Investigations on the change of fumonisin content of maize during hydrothermal treatment of maize. Analysis by means of HPLC methods and ELISA. European Food Research and Technology, 213, 187e193. Munkvold, G. P., & Desjardins, A. E. (1997). Fumonisins in maize. Can we reduce their occurrence? Plant Disease, 81, 556e564. Park, J. W., Scott, P. M., Lau, B. P., & Lewis, D. A. (2004). Analysis of heat-processed maize foods for fumonisins and bound fumonisins. Food Additives and Contaminants, 12, 1168e1178.
czuk, M., Wi
sniewska, H., & Chełkowski, J. (1997). Pascale, M., Visconti, A., Pron Accumulation of fumonisins in maize hybrids inoculated under field conditions with Fusarium moniliforme sheldon. Journal of the Science of Food and Agriculture, 74, 1e6. Pietri, A., Bertuzzi, T., Pallaroni, L., & Piva, G. (2004). Occurrence of mycotoxins and ergosterol in maize harvested over 5 years in Northern Italy. Food Additives and Contaminants, 21(5), 479e487. SCF. (2000). Scientific committee for food; opinion of the scientific committee on food on Fusarium toxins (Expressed on 22 June 2000). Seefelder, W., Hartl, M., & Humpf, H. U. (2001). Determination of N-(carboxymethyl) fumonisin B1 in maize products by liquid chromatography/electrospray ionization-mass spectrometry. Journal of Agricultural and Food Chemistry, 49, 2146e2151. Seefelder, W., Knecht, A., & Humpf, H. U. (2003). Bound fumonisin B1: analysis of fumonisin-B1 glyco and amino acid conjugates by liquid chromatographyelectrospray ionization-Tandem mass spectrometry. Journal of Agricultural and Food Chemistry, 51(18), 5567e5573. Warfield, C. Y., & Gilchrist, D. G. (1999). Influence of kernel age on fumonisin B1 production in maize by Fusarium moniliforme. Applied and Environmental Microbiology, 65, 2853e2856. €rner, M. (2001). Foods and fumonisins. European Food Research and Weidenbo Technology, 212, 262e273.