TOXIGENIC FUNGAL AND MYCOTOXIN CONTAMINATION OF MAIZE SAMPLES FROM DIFFERENT DISTRICTS IN SERBIA

: This study was carried out in order to investigate the natural occurrence of toxigenic fungi and levels of zearalenone (ZEA), deoxynivalenol (DON) and aflatoxin B 1 (AFB 1 ) in the maize stored immediately after harvesting in 2016 and used for animal feed in Serbia. A total of 22 maize samples were collected from four different districts across the country: City of Belgrade (nine samples), Šumadija (eight samples), Podunavlje (four samples) and Kolubara (one sample). Toxigenic fungi were identified according to the morphological characteristics whereas the mycotoxins contamination were detected using biochemistry enzyme-linked immuno-sorbent (ELISA) assay. The tested samples were mostly infected with Aspergillus, Fusarium and Penicillium spp., except that one sample originated from Kolubara was not contaminated with Aspergillus species. Fusarium graminearum was the most common species in the maize sample from Kolubara district (60%), F. verticillioides in the maize samples from Podunavlje (43.75%) and City of Belgrade samples were established, respectively. Generally, remarkable infection of all the tested samples with toxigenic fungal species from Aspergillus, Fusarium and Penicillium genera were recorded. In addition, high contamination with mycotoxins ZEA, DON and AFB 1 were also recorded; nevertheless, only in one sample the level of DON exceeded the allowed legal limit (1750 μg kg -1 ) according to Regulation for unprocessed maize. Therefore, permanent mycological and mycotoxicological analyses of maize grain are necessary for risk assessment of fungal and mycotoxin contamination throughout the food chain. genus, was determined in the maize samples, depending on the district tested. F. graminearum was the most common species (60%) in the maize sample from Kolubara district, followed by F. verticillioides in the maize samples from Podunavlje (43.75%) and City of Belgrade districts (22.4%) and Penicillium spp. in the maize samples from Šumadija district (26.38%). In regard to the tested mycotoxins, ZEA, DON and AFB 1 , only DON exceeded the allowed limit (1750 μg kg -1 ; EC, 2007) in the maize sample from Kolubara district. This was expected due to the high incidence of F. graminearum in the maize sample from this district and a positive correlation between incidence of F. graminearum and DON level. These studies have confirmed the potential danger and risk from toxigenic species, primarily Fusarium species and their mycotoxins in the production of maize in agro-ecological conditions in Serbia. The obtained results justify the need for constant fungal and mycotoxin analyses of maize grain and other types of feeds, in order to find preventive measures for reducing these contaminants in the food chain. Future research should focus on the examination of the incidence and analysis the greater number of the samples from a number of localities, as well as a more detailed examination of the dependence of fungal and mycotoxin contamination from climatic factors in order to more accurately assess the effect of the locality (district) on the natural occurrence of these contaminants in the production of maize.


Introduction
Maize as the most important crop in diet for humans and animals is grown on an area of about 1.03 million hectares in Serbia (Statistical Yearbook of Serbia, 2017). Maize grain as a rich source of nutrients represents a very good substrate for the development of toxic fungi (moulds) from the genera Aspergillus, Fusarium and Penicillium. These moulds are producers of secondary metabolites (mycotoxins). The most commonly detected mycotoxins in maize grain are fumonisins, deoxynivalenol (DON), zearalenone (ZEA), and aflatoxins (Covarelli et al., 2011). Among aflatoxins, aflatoxin B 1 (AFB 1 ) is a potent hepatotoxin and carcinogen that is a common contaminant of cereals and feeds. Aflatoxin M 1 (AFM 1 ) is a 4-hydroxylated metabolite of AFB 1 , which is excreted into milk through diet for dairy cows and represents a potential human carcinogen (Britzi et al., 2013). Zearalenone (ZEA) and deoxynivalenol (DON) are produced mainly by F. graminearum, and aflatoxin B 1 (AFB 1 ) produced by Aspergillus flavus and A. parasiticus (Nuryono et al., 2005;Zain, 2011). The harmful effects of aflatoxins, ZEA and trichothecenes on human and animal health are globally known (Khatoon et al., 2012).
Temperature, humidity and light are the key factors for Fusarium infection (Doohan et al., 2003), while the main factor for Aspergillus infection is the presence of primary inoculum at the time of maize ripening (Tédihou et al., 2012). In addition to field grain infestation, the development of these toxic species can be continued even during the storage period. Frequent adverse abiotic (high humidity and temperature) and biotic factors, including microorganisms, insects, mites, rodents and birds, can greatly contribute to increase contamination of maize grain with moulds and mycotoxins during storage conditions (Santin et al., 2005).
The occurrence of moulds and their mycotoxins in food is unpredictable and therefore can sometimes lead to adverse effects (mycotoxicoses) when consuming mouldy food (Nugmanov et al., 2018). Mycotoxicoses can be with acute and chronic symptoms. Farm animals, such as cattle, sheep, pigs and poultry Toxigenic fungal and mycotoxin contamination … 241 are very sensitive to increased mycotoxin concentrations in food. Intoxication with ZEA, DON and AFB 1 leads to disorders of reproductive functions and functions in the gastrointestinal tract in animals (Biagi, 2009;Liew and Mohd-Redzwan, 2018).
Due to the inevitable fungal and mycotoxin contamination of maize grain, it is necessary to propose preventive measures in the field in order to increase food safety. The application of maize hybrids less sensitive to fungal infection is one of the ways to reduce mycotoxin level in grain (Iglesias et al., 2010).
The aim of this research was to establish the presence of toxigenic fungal species and the level of some mycotoxins (ZEA, DON and AFB 1 ) in maize grain samples which were used for animal feed and to assess the risk of possible harmful effects of these contaminants in four districts of Serbia.

Materials and Methods
A total of 22 maize samples were collected from the maize stored immediately after harvesting, during November and December in 2016, from four different districts of Serbia, City of Belgrade (nine samples), Šumadija (eight samples), Podunavlje (four samples) and Kolubara (one sample). Most maize samples (eight samples from the City of Belgrade, four samples from Šumadija and Podunavlje districts, each, and one maize sample from Kolubara district) were collected from ventilated maize cribs, while fewer samples (one sample from City of Belgrade and four samples from Šumadija district) were collected from closed concrete warehouses in which the temperature and relative humidity conditions are not controlled. In both types of maize warehouses, maize was dried naturally. Maize grains were harvested manually from cob samples which were collected from maize cribs. The samples of maize grains of about 1 kg were stored in the paper bags in a refrigerator at 4°C prior to fungal and mycotoxin analyses.
Mycological analyses of maize grain samples were conducted according to the previously described methods by Krnjaja et al. (2015). Based on morphological properties (colony and spore appearance), toxigenic species have been identified according to fungal keys of Burgess et al. (1994) andSingh et al. (1991). The incidence of toxigenic species was calculated per sample according to Lević et al. (2012).
In order to determine the moisture content and the level of mycotoxins, the tested maize samples were first ground in an analytical mill (IKA A11, Germany). The moisture content was determined in laboratory conditions using a moisture analyser (OHAUS MB35, USA). Prior to the mycotoxicological analysis, maize samples were dried at 60°C for 72h and then the mycotoxin level was determined by a competitive ELISA method. The ELISA assay was done according to the manufacturer's instructions Celer Tecna® ELISA kits, at a wavelength of 450 nm.
The Pearson correlation coefficients between investigated variables (moisture content, incidence of toxigenic fungal species and the level of mycotoxins) were done in Excel 2010.
Mycological analyses confirmed the presence of Aspergillus, Fusarium and Penicillium spp. in the maize grain samples from all investigated districts, except in the maize grain sample from Kolubara district, in which Aspergillus species were not identified. Considering the average values, the most frequent species were F. graminearum in the maize grain sample from Kolubara district (60%) and F. verticillioides (43.75%) in the samples of maize grain from the Podunavlje district. Penicillium species were most prevalent in maize grain samples from Šumadija district (26.38%). Aspergillus species were present from 0 to 1.78% in the tested maize grain samples. Among the identified Aspergillus species, the species A. flavus was the most prevalent in maize samples from the City of Belgrade district (1.78%), while A. niger was most prevalent in samples of maize grain from the Podunavlje district (1%) and A. parasiticus was equally represented in maize samples from all districts (0.22-0.25%) except for Kolubara district (0%) ( Table 1). In mycotoxicological analyses, a high percentage of ZEA positive samples of maize grain originating from all districts was established (88.89-100%). DON was detected in 100% of the samples of maize grain from Šumadija and Kolubara Toxigenic fungal and mycotoxin contamination … 243 districts and in 22.2% of the samples of maize grain from the City of Belgrade district. AFB 1 was detected in 50, 55.6, and 87.5% of maize grain samples from Podunavlje, City of Belgrade and Šumadija districts, respectively (Table 2).
In the tested samples of maize grain, the mean level of ZEA was from 16.82 (Podunavlje district) to 26.97 μg kg -1 (Šumadija district), DON from 445 (City of Belgrade district) to 1977 μg kg -1 (Kolubara district) and AFB 1 of 1.3 (Podunavlje district) to 1.39 μg kg -1 (City of Belgrade district). DON and AFB 1 were not detected in maize grain samples from Podunavlje and Kolubara districts. In all tested samples, the levels of ZEA, DON, and AFB 1 were not above the allowed limits of 350, 1750, and 5 μg kg -1 , respectively, adopted by the European Commission (EC, 2007(EC, , 2010, except for DON level in maize grain sample from Kolubara district (1977 μg kg -1 ) ( Table 3).  Examination of the correlation ratios considered for a total of 22 tested maize samples showed positive correlations between moisture content and incidence of F. graminearum (r=0.28), F. proliferatum (r=0.12) and Penicillium spp. (r=0.58) and levels of ZEA (r=0.56), DON (r=0.49) and AFB 1 (r=0.16). Also, a positive correlation was established between the incidence of F. graminearum and DON (r=0.13) and between the incidence of A. parasiticus and AFB 1 (r=0.52). Likewise, positive correlations were established between the levels of ZEA and DON (r=0.52) and the level of DON and AFB 1 (r=0.42). Negative correlations were found between moisture content and incidence of A. flavus (r=-0.01), A. parasiticus (r=-0.16), A. niger (r=-0.01), F. verticillioides (r=-0.11) and F subglutinans (r=-0.09) and between the incidence of A. flavus and AFB 1 (r=-0.20) and incidence of F. graminearum and ZEA (r=-0.09).

Discussion
In the present study, the presence of toxigenic fungal species from Aspergillus, Fusarium and Penicillium genera and mycotoxins such as ZEA, DON and AFB 1 was confirmed in most of the maize samples originated from four Serbian districts. In all investigated districts, among Fusarium species, F. verticillioides was the most prevalent, except in maize sample from Kolubara district where the most dominant species was F. graminearum (Table 1). Similarly, in Argentina, in mycological studies including 52 maize samples, Pacin et al. (2001) have found that in all examined departments F. verticillioides was the prevalent toxigenic species, whereas the incidence of F. graminearum was low. In the present study, F. graminearum highly infected the maize sample from Kolubara district with incidence of 60%, while relatively high incidence were recorded (8.5-12.33%) in the samples from other investigated districts. These results may be explained with high average values of grain moisture content (13.57-15.45%). Moisture content is consider as one of the most important physiological factors for successful and safe storage of maize. The recommended moisture content for the safe maize storage is around 13% and below (Alptekin et al., 2009). The higher values of moisture content promote the favourable condition for development and proliferation of fungi and the appearance of insects leading to the storage problems (Weinberg et al., 2007). Also, Logrieco et al. (2002) reported that in some geographic regions the incidence of F. graminearum varies considerably from the investigated years and locality, which is highly connected with abiotic and biotic conditions. Among toxigenic fungal genera, Covarelli et al. (2011) found out the dominance of Fusarium species in the tested maize samples, followed by Aspergillus species of Flavi and Nigri sections and Penicillium spp., which is similar to our findings. In contrast, Alptekin et al. (2009) demonstrated a significantly higher incidence of Penicillium spp. in the maize samples collected from various counties in Turkey during the 2005-2006 growing season, relative to the species from Fusarium and Aspergillus genera.
Incidence of positive maize samples for all investigated mycotoxins was relatively high (50-100%) in all examined districts, except for DON in maize samples from City of Belgrade district (22.2%) and Podunavlje district (0%) and for AFB 1 in maize samples from Kolubara district (0%). Mean levels of ZEA, DON and AFB 1 were not above the allowed limits of 350, 1750 and 5 μg kg -1 , respectively, prescribed by the European Commission (EC, 2007(EC, , 2010 for unprocessed maize, except for DON level in maize sample from Kolubara district (1977 μg kg -1 ). Similar results were reported by Covarelli et al. (2011) in Italy, with DON and AFB 1 levels in some samples of maize grain being very high, 14,000 μg kg -1 and 820 μg kg -1 , respectively. Czembor et al. (2015) detected the incidence of DON and ZEA of 66.67% and 43.33%, respectively in the 30 tested maize samples originated from Poland, with the mean levels for positive samples of 50.77 μg kg -1 and 18.39 μg kg -1 , for DON and ZEA, respectively. In Turkey, Alptekin et al. (2009) detected AFB 1 in 72.4% of maize samples, with a concentration in the range of 0.63-108.86 μg kg -1 , with 43% of maize samples having an AFB 1 concentration above the permitted limit (5 μg kg -1 ). In Argentina, Pacin et al. (2001) have not detected ZEA and DON, while AFB 1 was detected in only one maize sample at a concentration of 16.8 μg kg -1 . In North Korea, Kim et al. (1993) have established 30% DON and 8% ZEA positive maize samples with an average concentration of 310 and 151 μg kg -1 , respectively. In Indonesia, Nuryono et al. (2005) have established a low percentage (3%) of ZEA positive maize samples for feed with an average concentration of 25.5 μg kg -1 .
Considering the correlation values for the total number of 22 samples examined, medium positive correlations between moisture content and incidence of Penicillium spp. were found (r=0.58) and levels of ZEA (r=0.56) and DON (r=0.49), also between the incidence of A. parasiticus and AFB 1 (r=0.52), ZEA and DON levels (r=0.52). Slight positive correlations were found between moisture content and incidence of F. graminearum (r=0.28) and F. proliferatum (r=0.12) and levels of AFB 1 (r=0.16), and between incidence of F. graminearum and DON levels (r=0.13). Slight negative correlations were found between moisture content and incidence of A. flavus (r=-0.01), A. parasiticus (r=-0.16), A. niger (r=-0.01), F. verticillioides (-0.11) and F. subglutinans (-0.09) and between incidence of F. graminearum and ZEA (r = -0.09) and incidence of A. flavus and AFB 1 levels (r = -0.20). In similar studies, Alptekin et al. (2009) have established slight positive correlations between relative humidity (RH) and fungal count (r=0.378) and AFB 1 level (r=0.258) and slight negative correlations between fungal count and AFB 1 level (r=-0.249). In contrast, Gourama and Bullerman (1995) have established a positive correlation between fungal growth and AFB 1 production.

Conclusion
Based on the obtained results, it can be concluded that the considerable presence of certain potentially toxic fungi species, especially from the Fusarium genus, was determined in the maize samples, depending on the district tested. F. graminearum was the most common species (60%) in the maize sample from Kolubara district, followed by F. verticillioides in the maize samples from Podunavlje (43.75%) and City of Belgrade districts (22.4%) and Penicillium spp. in the maize samples from Šumadija district (26.38%). In regard to the tested mycotoxins, ZEA, DON and AFB 1 , only DON exceeded the allowed limit (1750 μg kg -1 ; EC, 2007) in the maize sample from Kolubara district. This was expected due to the high incidence of F. graminearum in the maize sample from this district and a positive correlation between incidence of F. graminearum and DON level.