Regulated Mycotoxin Occurrence and Co-Occurrence in Croatian Cereals

A total of 209 samples of various cereal crops (maize, wheat, barley, rye and oats) grown in Croatian fields during 2016 and 2017 were collected to analyze and determine the occurrence and co-occurrence of EU regulated mycotoxins in cereals (AFB1, AFB2, AFG1, AFG2, DON, FB1, FB2, ZEA, T-2, HT-2 and OTA). The analysis, performed by a validated confirmatory LC-MS/MS method based on a dilute and shoot principle, highlighted Fusarium mycotoxins as the main contaminants, often co-occurring in samples from both years (50.0% in 2016 and 33.7% in 2017). DON was found to be the most frequent mycotoxin, present in 72.5% of the 2016 samples and 32.6% of the 2017 samples, while maize proved to be the most contaminated cereal type of both years with FUM as the most abundant mycotoxins, with an average concentration of 1180 µg/kg. Moderate temperatures with periods of high humidity favored the accumulation of DON in wheat samples instead of other Fusarium mycotoxins, while similar conditions favored maize contamination with FUM. A total of 8.3% of all the 2016 harvest samples and 7.9% of the 2017 harvest samples were assessed as non-compliant, containing mycotoxins in concentrations higher than the levels set by the EU legislation for food.


Introduction
Mycotoxins, secondary fungal metabolites of low-dose toxicity, are unambiguous food and feed contaminants, able to enter the food chain at any point from the field to fork chain. Cereal crop contamination with toxigenic fungi and consequently mycotoxins can occur in fields, before or after harvest, during crop manipulation and storage [1,2], thus endangering the world's still most important source of food [2]. According to the data of the Croatian Bureau of Statistics (https://www.dzs.hr, accessed on 20 January 2022), cereals represent about 60% of all sown land in Croatia. A total cereal production, including the 2 most represented agricultural crops, wheat and maize, amounts for over 3,000,000 t, giving a yield of up to 6.0 and 9.0 t/ha, respectively. Other cereal crop types, such as barley, oats and ins proved to be the most common contaminants, found in all the samples in concentrations above LOD. In both analyzed harvest years, 2016 and 2017, DON was found to be the most frequent mycotoxin, appearing in 72.5% and 32.6% of all cereal samples, respectively. In 2016, the highest concentration of DON was found in maize amounting to 4902 µg/kg, while the average value of all positive cereal samples was 595 µg/kg, similar to the average value of the 2017 samples when the mean value was 510 µg/kg. However, the highest measured concentration for the 2017 samples was 2408 µg/kg, found in wheat. The concentrations of EU regulated mycotoxins in all the cereal samples from 2016 and all the cereal samples from the 2017 harvest obtained in this study using the dilute and shoot LC-MS/MS method are presented in Table 1. The obtained mycotoxin occurrence data on the individual crop types from 2016 and from 2017 are shown in Tables 2 and 3.     [13]. One sample (maize) from 2016 and two samples (wheat and oats) from 2017 did not even comply with the permitted values for animal feed set by the Commission Recommendation 2006/576/EC [15]. An overview of non-compliant samples, with the stated counties of origin, is given in Table 4. The most contaminated cereal type from 2016 and 2017 was maize, with FB1 being the most common mycotoxin (contaminating 85.7% of all the maize samples), followed by HT-2 (73.8%), DON (65.5%), FB2 (64.3%), ZEA (54.8%), T-2 (28.6%), AFB1 (2.4%), and AFB2 and AFG1 (1.2%). Mycotoxin co-occurrence was found in 50.0% of all the 2016 samples (98.4% of all the maize samples), containing 2 or more mycotoxins. The highest co-occurrence was observed for Fusarium mycotoxins, DON, FB1, FB2, ZEA, T-2 and HT-2, which were contained in 26.2% of all the the maize samples. FB1 and FB2 in maize samples were the most co-occurring mycotoxins in the 2017 samples, followed by DON and HT-2 in wheat samples, giving a total of 33.7% of all the analyzed samples containing 2 or more mycotoxins. An overview of all the co-occurring mycotoxin combinations found in cereals from the 2016 and 2017 harvests is shown in Table 5.
Mycotoxin occurrence data analysis revealed the statistically significant difference in the AFT (p = 0), FB1 (p = 0.042), FUM (p = 0.041)), T-2 (p = 0.038) and OTA (p = 0) distribution between the harvest years 2016 and 2017, unlike for other tested mycotoxins, for which no statistically significant difference (DON (p = 0.740), FB2 (p = 0.445), ZEA (p = 0.578) and HT-2 (p = 0.139)) was found. A comparison of the variability of mycotoxin contamination between the individual crops showed a statistically significant difference in the occurrence of DON (p = 0.002) and HT-2 (p = 0.038) between maize and wheat, as well as a statistically significant difference in the HT-2 concentration (p = 0.002) between wheat and oats, which was expected, given the focus of legislation [14] on the HT-2 toxin (T-2/HT-2 sum) in oats, distinguishing this type of cereal from others. The obtained statistical analysis data are summarized and presented in box-plot diagrams for both years and the analyzed mycotoxins in Appendix A ( Figures A1-A5).
Data on mycotoxin occurrence and co-occurrence in Croatian cereals reported by other authors for harvest seasons 2016 and 2017 are limited; however, the data that are available for the relevant years agree with the data of the study conducted in this work. Kovač et al. (2021) [2] conducted a multicontaminant analysis of maize and wheat grown in Croatian fields in 2016, including regulated mycotoxin occurrence determination. Fusarium mycotoxins were found to be the most significant secondary metabolites, with DON being the most frequent, present in 73.7% of samples, while AFT and OTA were not detected. The analysis also revealed 8.5% of cereal samples non-compliant to the EU legislation for food regarding mycotoxins. Sunic et al. (2021) [34] investigated Fusarium mycotoxins levels, including regulated DON and ZEA, in the unprocessed wheat from the seasons in 2019/2020 from    Mycotoxin incidence in Croatian cereals of earlier harvest seasons was better investigated. Pleadin [31]. Maize was proved to be the most contaminated cereal type, with DON as the most occurring mycotoxin (in 52.5% of samples), followed by ZEA (40.5%), FUM (37.5%) and T-2 toxin (33.0%). The authors associated the relatively low mycotoxin contamination with the very warm and dry weather during the observation period, which is supported by the studies [29,30] conducted on cereals from the year 2010, which abounded in rain and cold weather, revealing the significant contamination of the analyzed samples with Fusarium mycotoxins.
Furthermore, high air temperature and low amounts of precipitation were correlated to the high occurrence of AFB1 in cereal samples harvested in the years 2009-2013, and in products thereof [32,33], emphasizing the importance of the geographical origin of the crops, i.e., the weather conditions.
The obtained data also conform to the mycotoxin occurrence and co-occurrence studies from nearby/Balkan countries, with a climate similar to Croatia. In the Serbian study conducted by Radić et al. (2021) [36], it was shown that the growing season had an impact on Fusarium mycotoxin occurrence in maize; generally higher levels of certain Fusarium mycotoxins were detected in 2016, compared to the 2017 samples, due to increased humidity. Among the analyzed Fusarium metabolites in maize, FUM were found to be the most frequent. Stanciu et al. (2019) [37] performed a study on trichothecene and ZEA presence in Romanian wheat from the 2015 harvest season. The highest overall frequency was registered for DON, and the authors correlated the mycotoxin concentration with the precipitation and temperature values during anthesis and the preharvest period. Cereal samples from Bosnia and Herzegovina were investigated by Pleadin et al. (2017b) [38] studying the occurrence of certain Fusarium mycotoxins. Among the analyzed unprocessed cereals (maize, wheat, barley), maize was the most contaminated, containing DON in 85.2%, ZEA in 73.0% and FUM in 66.9% of the samples. Higher mycotoxin levels determined from the 2014 samples, compared to the samples from 2013 and 2015, were linked to extreme humidity observed during the cereal growth and harvesting periods [37].

Weather Conditions and Mycotoxin Concentrations in Cereals
Given the above, mycotoxin biosynthesis is greatly influenced by climatic conditions, humidity and temperature, contributing to the mycotoxigenic fungi colonization and mycotoxin production [1,7,10,39]. The plant is generally most susceptible to contamination during the flowering period; in the case of maize, the plant is most susceptible to contamination in late July and early August (the anthesis and the silking period), when Aspergillus fungi and consequently AFT contamination is more likely to occur if extreme drought takes place. In addition, the rainy periods during September and October can cause the contamination of maize with Fusarium fungi and mycotoxins. In wheat and other spring crops, the end of May and the beginning of June are critical, since extremely rainy periods favor Fusarium fungi contamination [26,27]. Figure 1 shows the sowing, contamination and harvest periods of certain cereal crops. midity. Among the analyzed Fusarium metabolites in maize, FUM were found to be the most frequent. Stanciu et al. (2019) [37] performed a study on trichothecene and ZEA presence in Romanian wheat from the 2015 harvest season. The highest overall frequency was registered for DON, and the authors correlated the mycotoxin concentration with the precipitation and temperature values during anthesis and the preharvest period. Cereal samples from Bosnia and Herzegovina were investigated by Pleadin et al. (2017b) [38] studying the occurrence of certain Fusarium mycotoxins. Among the analyzed unprocessed cereals (maize, wheat, barley), maize was the most contaminated, containing DON in 85.2%, ZEA in 73.0% and FUM in 66.9% of the samples. Higher mycotoxin levels determined from the 2014 samples, compared to the samples from 2013 and 2015, were linked to extreme humidity observed during the cereal growth and harvesting periods [37].

Weather Conditions and Mycotoxin Concentrations in Cereals
Given the above, mycotoxin biosynthesis is greatly influenced by climatic conditions, humidity and temperature, contributing to the mycotoxigenic fungi colonization and mycotoxin production [1,7,10,39]. The plant is generally most susceptible to contamination during the flowering period; in the case of maize, the plant is most susceptible to contamination in late July and early August (the anthesis and the silking period), when Aspergillus fungi and consequently AFT contamination is more likely to occur if extreme drought takes place. In addition, the rainy periods during September and October can cause the contamination of maize with Fusarium fungi and mycotoxins. In wheat and other spring crops, the end of May and the beginning of June are critical, since extremely rainy periods favor Fusarium fungi contamination [26,27]. Figure 1 shows the sowing, contamination and harvest periods of certain cereal crops.  Relatively low average levels or the absence of found mycotoxins in wheat and other spring crops from both 2016 and 2017 can be explained by the unfavorable weather conditions during crop flowering, generally characterized by relatively mild temperatures and low amounts of precipitation. According to the publicly available meteorological data, temperature conditions in May and June of both years were normal to very warm, while precipitation conditions were normal to wet, both not deviating significantly from the multiyear average. The incidence of DON in wheat samples (68.4% in 2016 and 42.6% in 2017) instead of other Fusarium mycotoxins, is likely to be caused by moderate temperatures with periods of high humidity that helped certain Fusarium fungi contamination and DON accumulation [40], which is confirmed by the general absence or low contamination of cereals by ZEA. The sporadic high concentration of ZEA found in the oat sample (5155 µg/kg, non-compliant sample) from Osijek-Baranja County was probably caused by somewhat colder (normal) temperature conditions during May 2017, compared to the rest of the country, as well as by improper storage, since other samples from the same county were within the EU permitted levels. Moreover, the average monthly air temperatures, mostly below the multi-year average in October 2016, may be the cause of the significant ZEA production of 61-9344 µg/kg and FB2 49-2442 µg/kg), compared to 2017 (FB1 52-7350 µg/kg and FB2 47-2139 µg/kg). The high concentration of AFT found in the 2017 maize sample from the Split-Dalmatia County (AFB1 9.7 µg/kg and AFB2 0.5 µg/kg) can be explained by the high temperatures during the silking period, i.e., a very warm July and extremely warm August throughout the country and drought in the southern regions, which probably favored the development of aflatoxigenic fungi and the production of AFT. For comparison, in the 2016 harvest samples, AFT was not detected in any of the analyzed samples, and according to the available meteorological data, August 2016 is characterized by normal precipitation amounts in most of the country and normal to warm temperatures that did not favor the occurrence of these secondary metabolites. Significant contamination of a wheat sample originating from Šibenik-Knin County with AFT (AFB1 16.2 µg/kg, AFB2 4.4 µg/kg, AFG1 20.9 µg/kg and AFG2 2.2 µg/kg), but also the unusually high OTA presence (614 µg/kg) in the same sample, cannot be explained by weather conditions at the certain micro location, thus it can be assumed that is probably a consequence of accidental contamination due to contaminated storage space. Additionally, the presence of G group AFT (AFG1, AFG2) indicates crop contamination with A. parasiticus, which is not a common resident of cereal fields [5,11,41].
To review the cereal contamination levels of regulated mycotoxins from each Croatian county, the obtained data were used to create heat maps of mycotoxin occurrence (AFT Croatia (Osijek-Baranja and Požega-Slavonia counties), similar to HT-2, which was found in the highest concentrations in Požega-Slavonia County. In 2017, T-2 and HT-2 most frequently appeared in the Karlovac and Koprivnica-Križevci counties, followed by Varaždin and Osijek-Baranja counties. AFT were found in only a few samples of the analyzed cereals from the 2017 harvest year, with the highest concentrations originating from Southern Croatia (Šibenik-Knin and Split-Dalmatia counties), which also contained a significant concentration of OTA.   County. The occurrence of T-2 toxin in 2016 was the most significant in the eastern part of Croatia (Osijek-Baranja and Požega-Slavonia counties), similar to HT-2, which was found in the highest concentrations in Požega-Slavonia County. In 2017, T-2 and HT-2 most frequently appeared in the Karlovac and Koprivnica-Križevci counties, followed by Varaždin and Osijek-Baranja counties. AFT were found in only a few samples of the analyzed cereals from the 2017 harvest year, with the highest concentrations originating from Southern Croatia (Šibenik-Knin and Split-Dalmatia counties), which also contained a significant concentration of OTA.

Conclusions
The research conducted in the current study provides the first comprehensive occurrence and co-occurrence data on all mycotoxins regulated in cereals by the EU legislation (AFB1, AFB2, AFG1, AFG2, DON, FB1, FB2, ZEA, T-2, HT-2 and OTA), presenting a clearer insight on the profiles of mycotoxin contamination of samples from all Croatian counties, together with the impact of weather conditions. The results highlighted Fusarium

Conclusions
The research conducted in the current study provides the first comprehensive occurrence and co-occurrence data on all mycotoxins regulated in cereals by the EU legislation (AFB1, AFB2, AFG1, AFG2, DON, FB1, FB2, ZEA, T-2, HT-2 and OTA), presenting a clearer insight on the profiles of mycotoxin contamination of samples from all Croatian counties, together with the impact of weather conditions. The results highlighted Fusarium mycotoxins as the main contaminants, with DON being the most frequent in all the sampled cereal crops, while maize proved to be the most contaminated cereal type for both years. A total of 8.3% of all the harvest samples from 2016, and 7.9% of the 2017 harvest samples were non-compliant regarding the permitted levels given in the EU legislation for food. Mycotoxin co-occurrence was confirmed in 50.0% and 33.7% of cereal samples from 2016 and 2017, respectively, with Fusarium mycotoxin combination being the most frequent. Furthermore, the found mycotoxins and their concentrations are shown to be related to meteorological conditions during the flowering period of the plants, and the reason for the more significant contamination level of the 2016 cereals rather than 2017. Even though a relatively small percentage of the analyzed cereal samples was assessed as non-compliant to the EU legislation for food, sporadically high concentrations of certain mycotoxins, as well as a large number of co-occurring mycotoxin combinations, point out the necessity of the constant and continuous monitoring of mycotoxin levels and gathering enough data for proper toxicological studies, to ensure food safety in a world exposed to climate change.

Sample Collection
A total of 209 samples of unprocessed cereals (maize 84, wheat 104, barley 9, rye 6 and oats 6) grown in the Croatian fields in the years 2016 and 2017, were selected to analyze and determine the occurrence and co-occurrence of EU regulated mycotoxins. The samples were collected from all Croatian counties, at least one sample from each county and the rest according to the cereal type representation. The sampling was conducted in warehouses and silos in accordance with the Commission Regulation (EC) No 401/2006 [42]. To obtain a homogeneous particle size, samples prepared in test portions (10 kg per sample) were ground using a Romer RAS mill (Romer Labs, Tulln, Austria) with a sieve pore size 0.5 mm, reduced to a laboratory subsample of 500 g using a Retsch Sample Splitter RT 12.5 (Retsch, Haan, Germany) and afterwards stored in a freezer at −18 • C pending analysis.

Chemicals, Reagents and Standards
All certified mycotoxin standards were supplied by Romer Labs Biopure (Romer Labs, Tulln, Austria), stored according to the manufacturer's instructions and tempered to room temperature before use. LC-MS grade acetonitrile and methanol were obtained from J.T. Baker (J.T. Baker, Deventer, The Netherlands), while LC-MS grade formic acid and ammonium format were supplied by Sigma-Aldrich (Sigma-Aldrich, St. Louis, MO, USA). Ultrapure water was generated by the Niro VV system (Nirosta d.o.o., Osijek, Croatia).

Mycotoxins Determination
The determination of all the 11 mycotoxins regulated in cereals by the above-mentioned EU legislation (AFB1, AFB2, AFG1, AFG2, DON, FB1, FB2, ZEA, T-2, HT-2 and OTA) was performed by a confirmatory LC-MS/MS method, based on a dilute and shoot principle, previously described by Kovač et al. (2021) [2]. The method was in-house validated according to the relevant EU legislation [42,43] to fulfil the performance criteria and other requirements set for the analytical methods. For identification purposes, the relative retention time of the certain analyte had to correspond to that of the calibration solution at a tolerance of ±2.5%, and the relative ion intensities (quantification and confirmation ion response ratio) had to match an amount of ±30% of the ion ratio in the analyte standard from the same measurement sequence, thus satisfying the criteria of the Commission Decision (EC) No. 657/2002 [43]. The external matrix-matched calibration for each cereal type was used for quantification purposes, while internal quality control was conducted by performing the recovery experiments by spiking within each sample batch, correcting the measured concentrations for recovery when outside the allowed range (90-110%) for mycotoxins by the Commission Regulation (EC) No. 401/2006 [42]. An additional method control was carried out via participation in proficiency testing schemes in both years, in which measurements were taken and evaluated using the obtained z-scores (presented in Appendix B, Table A1), which were considered as satisfactory when within range |z| ≤ ±2.

Statistical Analyses
Statistical software packages, Statistica 13.3 (TIBCO Software Inc., Palo Alto, CA, USA) and Tableau 2019.4.5 (3 Excel 2016 (Microsoft, Redmond, WA, USA), were used for statistical data processing. In results processing, the methods of descriptive statistical analysis were used: measures of central tendency and measures of variability (range, standard deviation) calculated from the obtained measurements. The normality of the data distribution was tested by the Shapiro-Wilk W test, the homogeneity of the Levene s test variance, and a comparison between the two individual variables by the Mann-Whitney U test in data not normally distributed and by the t-test in normally distributed data. The Spearman s correlation coefficient was used to calculate the correlation; data with p < 0.05 values were considered statistically significantly different.

Meteorological Data/Heat Maps of Mycotoxin Incidence
A comparison of the type and concentration of found mycotoxins with weather conditions at individual micro locations was performed, using the data obtained from the Croatian Meteorological and Hydrological Service (publicly available at https://meteo.hr, accessed on 22 December 2021) on weather conditions (average monthly air temperature, average monthly amount of precipitation) during the cereal growth and harvest season in the years 2016 and 2017. Heat maps of mycotoxin incidence throughout the country (representing each county) were designed using Tableau 2019.4.5 (Tableau Software, Seattle, WA, USA).

Sample Compliance
The compliance of the analyzed samples regarding the mycotoxins was proved by comparing the found mycotoxin concentrations with the permitted/indicative levels for mycotoxins stated in the EU legislation given by the Commission Regulation (EC) No 1881/2006 [13] and the Commission Recommendation 2013/165/EU [14] for food, and the Commission Recommendation 2006/576/EC [15] and Directive 2002/32/EC [16] for feed.