2.1. Samples
Maize samples (n = 400) were randomly collected from the main maize producing areas in Northern Serbia (Autonomous Province of Vojvodina, latitude 45°18″ (N), longitude 20°09″ (E), and altitude 111 m) during a period from 2018 to 2021 (
Figure 1). A total of 100 maize samples from each year were taken evenly from the entire region of Northern Serbia. In Croatia 268 maize samples were analysed on DON, 191 samples on FUMs, 382 samples on ZEN, and 262 samples on T-2/HT-2 toxin concentrations. Samples were collected within the period 2018–2021 from different fields situated in four Croatian regions (Eastern, Central, Northern and Western Croatia), with latitude 45°10″ (N), longitude 15°30″ (E), and altitude 142 m (
Figure 1). Since the largest production of maize is in Eastern Croatia, about 70% of the samples originate from that part of the country. Sampling in both countries was performed according to Commission Regulation (EC) No. 401/2006 [
29], while in Serbia the Serbian Regulation [
30] was also taken into account.
With the aim to examine the influence of weather conditions on the contamination of maize samples with Fusarium mycotoxins, and to avoid the possibility of secondary contamination, maize samples from both countries were collected during September and October (2018–2021), immediately after harvest or from dryers, before further storage in silo and distribution.
In Serbia after harvest in each year, at the Institute of Food technology in Novi Sad, representative maize samples were prepared from approximately 10 kg of aggregate samples. Preparation of representative samples included: homogenization (Nauta mixer, model 19387, Nauta patenten, The Netherlands), quartering, milling (KnifetecTM 1095 mill, Foss, Hoganas, Sweden), packing in zip lock bags (150–200 g), and storage at −18 °C. The laboratory samples were removed from freezing at the beginning of 2022, and again homogenized (Rotary laboratory mixer RRM Mini-II, Ludwigshafen, Germany) before analysis on high-performance liquid chromatography-tandem mass spectrometry (LC-MS/MS). Before mycotoxins analysis in maize samples from Croatia by ELISA methods, each year, the prepared test portions were ground into a fine powder having a particle size of 1.0 mm using an analytical mill (Cylotec 1093, Tecator, Sweden), and then stored at 4 °C until analyses of particular mycotoxins within 48 h.
2.2. Fusarium Mycotoxins Analysis
In Serbia LC-MS/MS was used for analysis of the collected maize samples. The following chemicals were used: acetonitrile (ACN) (Fisher Scientific, Geel, Belgium) of HPLC grade and water (Fisher Scientific, Geel, Belgium), methanol (MeOH) (Carlo Erba, Val de Reuil, France), and formic acid (Fluka Analytical, Sigma Aldrich, Steinheim, Germany) of LC-MS/MS grade.
Ultra-pure water was produced by Adrona Crystal EX HPLC Water Purification system (Riga, Latvia). Mycotoxin standards DON (100 μg/mL), FUMs (FB1 and FB2, each 50 μg/mL), HT-2 toxin (100 μg/mL), and ZEN (100 μg/mL) were purchased from Fluka Analytical (Steinheim, Germany), except for T-2 toxin (100 μg/mL) which was purchased from LGC Standards (Wesel, Germany). A mixed standard stock solution of 6 different mycotoxin standards was dissolved in MeOH/water (50:50, v/v). Correspondingly, matrix-matched standards were prepared by diluting an appropriate volume of the appropriate stock standard solution in the blank sample extract with MeOH/water (50:50, v/v), yielding concentration levels from 5 to 150 ng/mL for DON, from 0.5 to 200 ng/mL for FB1 and FB2, from 0.05 to 30 ng/mL for HT-2 toxin and T-2 toxin, and from 1 to 200 ng/mL for ZEN. Both mycotoxin standards and stock standard solutions were stored in the freezer at −18 °C until analysis.
Sample preparation and analysis were performed according to Application Note from Thermo Fisher Scientific [
31]. Briefly, five grams of milled and homogenized sample was weighed into a 50 mL centrifuge tube. Twenty milliliters of ACN/water (80:20,
v/v) extraction solvent was added and the tube was shaken on a horizontal shaker (6 Hz, 60 min). This was followed with centrifugation at room temperature (67 Hz, 5 min). Supernatant (400 μL) was diluted with 600 μL MeOH/water (50:50,
v/v), and shaken on a vortex shaker. The supernatant was filtered through a 0.2 µm PTFE filter into HPLC vials and injected into the LC-MS/MS system.
LC-MS/MS analysis was performed using an HPLC Vanquish Core system (Thermo Fisher Scientific, Waltham, MA, USA) equipped with a TSQ Quantis Triple Quadrupole mass spectrometer equipped with a heated electrospray ionization (HESI) source (Thermo Fisher Scientific, Waltham, MA, USA). A ZORBAX Eclipse Plus C18 column (2.1 × 100 mm, 1.8 µm) (Agilent, Santa Clara, CA, USA) was selected for chromatographic separation. A mobile phase consisting of water with 0.1% formic acid (mobile phase A) and 0.1% formic acid in methanol (mobile phase B) was used. The gradient elution program was as follows: 0 min 95% mobile phase A, 0.5 min 95% mobile phase A, 7 min 30% mobile phase A, 9 min 0% mobile phase A, 12 min 0% mobile phase A, 12.1 min 95% mobile phase A, and 15 min 95% mobile phase A. The mobile phase flow rate was 0.3 mL/min. The temperature of the column was set at 40 °C and the autosampler tray temperature was set at 20 °C. The sample injection volume was 10 µL. The mass spectrometer analyses were carried out using selected reaction monitoring (SRM) mode. The HESI source was operated under positive (3.5 kV) mode, except for ZEN in negative (2.5 kV) mode. The applied parameters were as follows: ion transfer tube, 325 °C; vaporizer temperature, 350 °C; cycle time, 0.5 s; collision induced dissociation gas pressure, 1.5 mTorr; auxiliary gas, 6 Arb; sheath gas, 30 Arb; and sweep gas, 1 Arb. Argon served as collision gas, while nitrogen served as auxiliary, sheath, and sweep gas. Data were acquired and analyzed using Thermo Scientific TraceFinder software TSQ Quantis 3.2 Tune (Thermo Fisher Scientific, Waltham, MA, USA).
The LC-MS/MS method was validated in accordance with the performance criteria outlined in Commission Decision [
32] and Technical Report CEN/TR 16059:2010 [
33]. The method was validated in terms of linearity, limit of quantification (LOQ), trueness, recovery, repeatability, and reproducibility. The validation study was conducted by the analysis of quality control materials (for DON, FB
1, FB
2, ZEN) and spiked uncontaminated maize samples (for T-2 and HT-2 toxin). The quality control materials, maize flours, for the examined mycotoxins were provided by: Trilogy analytical laboratory for FUMs (product code TR-F100; contained 2400 µg/kg of FUMs) and ZEN (product code TR-Z100; contained 59.4 µg/kg of ZEN), and Food Analysis Performance Assessment Scheme (FAPAS) for DON (Food Chemistry Proficiency Test 04378; contained 1200 µg/kg of DON). Matrix effects were accounted by using matrix matched calibration (MMC). The matrix effects were expressed as the signal suppression/enhancement (SSE) and calculated from the slope ratio for MMC and solvent calibration. For each examined
Fusarium mycotoxin, the primary product ion which corresponds to the most abundant product ion was used for quantification, while the other two were used for confirmation: DON (249.0, 231.0, and 175.0), FB
1 (334.1, 352.1, and 704.4), FB
2 (336.1, 354.2, and 688.5), ZEN (175.0, 131.0, and 273.0), T-2 (245.1, 327.0, and 387.1), and HT-2 toxin (345.1, 255.1, and 285.1). The retention times for the examined
Fusarium mycotoxins were the following: DON, 1.05 min; FB
1, 10.85 min; FB
2, 11.50 min; ZEN, 11.66 min; T-2, 11.30 min; and HT-2, 10.86 min.
All obtained validation parameters, for the applied LC-MS/MS method in this study, fulfil the criteria given under the Regulation 2006/401/EC [
29] and the Technical Report [
33]. All mycotoxins were quantified using MMC curves, because for all investigated mycotoxins SSEs were higher than ±20%. For all curves the squared correlation coefficients (R
2) were above 0.998. LOQs for DON, FB
1, FB
2, ZEN, T-2, and HT-2 toxin were 50.0, 5.0, 5.0, 10.0, 0.50, and 0.50 µg/kg, respectively. The analysis of quality control materials yielded values for trueness: DON (104%), FB
1 (83%), FB
2 (81%), ZEN (110%), while the analysis of spiked uncontaminated maize samples yielded recovery: T-2 (96%), HT-2 toxin (88%). Repeatability and reproducibility were calculated as relative standard deviations and none of them exceeded 20%. Therefore, all obtained values for trueness, recovery, repeatability, and reproducibility were in accordance to the criteria specified in used regulations [
29,
33].
In Croatia, concentrations of mycotoxins were determined using competitive ELISA Ridascreen® test kits for DON (Art. No. R5906), ZEN (Art. No. R1401), FUMs (Art. No. R3401), and T-2/HT-2 toxin (Art. No. R3805). Analyses were performed completely as instructed by the kits manufacturer (R-Biopharm, Darmstadt, Germany). Each kit contains: a microtiter plate with 96 wells coated with antibodies; standard solutions containing different concentrations of mycotoxin standards: DON (0, 3.7, 11.1, 33.3, 100 μg/L); ZEN (0, 50, 150, 450, 1350, 4050 ng/L); FUM (0, 0.025, 0.074, 0.222, 0.666, 2 mg/L), and T2-HT2 (0, 1, 3, 6, 12, 36 μg/L); an enzyme conjugate; an anti-antibody; the substrate and chromogen solution (urea peroxide/tetramethylbenzidine); washing and dilution buffers and stop solution. All other chemicals used in the analysis were of analytical grade.
ELISA tests were evaluated using a ChemWell auto-analyzer (Awareness Technology Inc. 2910, Palm City, FL, USA) with the absorbance being measured at 450 nm. In order to determine mycotoxin concentrations in the sample, a standard curve was plotted based on the sample extract dilution factors. Implemented ELISA methods were validated in earlier studies [
34,
35,
36,
37] and final concentrations were calculated per each mycotoxin based on the average recoveries. Further, applied ELISA methods were validated as accredited in agreement with the ISO/IEC 17025 Standard [
38]. LOQ values for DON, ZEN, FUMs, and T-2/HT-2 equaled to 22, 3, 24, and 5 µg/kg, respectively.
For all ELISA tests the validation was performed by determination of recovery and intermediate precision. The recovery rates were determined at three different levels (50, 100, and 200 mg/kg) by spiking the control maize samples with the standard in-house mycotoxin working solution (300 mg/L) corresponding to the assessed content levels. Regarding the determination of intermediate precision, the same steps were repeated on two other occasions by two different analysts and under the same analytical conditions. The mean recovery rates for DON, ZEN, FUMs, and T-2/HT-2 toxins were 98.2%, 89.2%, 75.2%, and 97.6%, while for intermediate precision the mean rates equaled to 95.9%, 86.8%, 72.1%, and 92.9%, respectively.
Quality control was performed by analysis of different reference materials (RMs) in parallel with each batch of the studied samples, so as to check whether the obtained concentration falls within the assigned range. The producer (FAPAS, Sand Hutton, York, UK) assigned ranges of RMs were: T04354QC, maize—1104 μg/kg (756–1452 μg/kg) for DON and 90.7 μg/kg (50.8–130.5 μg/kg) for ZEN; TYG079RM, maize flour—854 μg/kg (791–917 μg/kg) for FUM; and TYG087RM, cereal-based animal feed—523 μg/kg (496–550 μg/kg) for T-2/HT-2. Mycotoxin concentrations obtained with RMs were compared to the assigned values given by the manufacturer and were found to be within the defined ranges.