Eighty-two stored wheat samples produced in Iran were collected from supermarkets (Table 1). Fusarium counts were determined by dilution plating and direct plating according to Ghiasian et al. [16] with minor modifications. Briefly, ground samples (10 g each) were mixed with 90 mL of sterile solution containing 0.1% peptone for the 10⁻¹ dilution. Further serial dilutions to the 10⁻⁵ were made with 0.1% peptone solution. Aliquots (1.0 mL) of each dilution were then transferred to Petri dishes containing potato dextrose agar (PDA) under sterile conditions. The Petri dishes were incubated at 25 °C for 7 days. Colonies that developed on plates were counted at end of the incubation period and recorded as Colony Forming Units per gram (CFU/g). Fusarium incidence was also evaluated by direct plating. Briefly, kernels from each sample were surface disinfected for 1 min with 1% sodium hypochlorite solution, rinsed twice in sterile distilled water, and dried under sterile conditions. One hundred kernels per sample were plated onto pentachloronitrobenzene (PCNB) agar plates [23]. The plates were incubated at 28 °C for 5–7 days. The resulting Fusarium colonies recovered in both methods were identified based on morphological characters according to Nelson et al. [24] and Leslie and Summerell [25].

Fumonisins were extracted and estimated according to the method of Sydenham et al. [10]. Briefly, 20 g of homogenized samples were mixed with 100 mL of 70% methanol and blended for 3 minutes. After centrifugation at 1,500 g for 10 min at 4 °C, the mixture was filtered through a Macherey-Nagel 617 filter paper. The pH was adjusted to 6.0 with 1 M NaOH and an aliquot (5 mL) was applied at a flow rate of less than 2 mL/min to a strong anion exchange (SAX) solid phase extraction cartridge (Bond-Elut, Varian, Harbor City, CA, USA) containing 500 mg sorbent. The cartridge was conditioned by successive passage of methanol (5 mL) and methanol/water (3:1, v/v, 5 mL). The columns were pre-washed with 5 mL of 100% methanol followed by 5 mL of methanol/water (75:25). Fumonisins were eluted from the columns with 1% acetic acid in 10 mL of methanol, dried in an evaporator, and used for HPLC analysis.

Fumonisin standards were obtained from Sigma, St Louis, USA. Standard solutions were prepared containing FB1 (245 µg/mL), FB2 (200 µg/mL) and FB3 (285 µg/mL) in acetonitrile/water (1:1, v/v) [16] and determined by HPLC analysis using fluorescence detection was conducted as described by Sydenham et al. [10]. An aliquot (25 µL) of purified extract was derivatized with 225 µL O-phthaldialdehyde solution prepared by dissolving 40 mg of O-phthaldialdehyde in 1 mL of 100% methanol, diluting into 25 mL of distilled water containing 5 mL of 0.1 M disodium tetraborate, and mixing with 50 µL of 2-mercaptoethanol. The fumonisin O-phthaldialdehyde derivatives were analyzed using HPLC system consisting of a Waters model 590 pump (Milford, MA) connected to a Waters 470 scanning fluorescence detector. The separations were performed on a reversed phase stainless steel Ultracarb 5 ODS column (150 mm × 4.6 mm id, Phenomenex, Torrance, CA). The mobile phase was methanol/0.1 M sodium dihydrogen phosphate (75:25, v/v) adjusted to pH 3.35 with ortho-phosphoric acid and pumped at a flow rate of 1 mL/min. Fumonisins were quantified by peak area measurement in comparison with fumonisin reference standard using Borwin integration software (JMBS Developments, LE Fontanil, France). Recoveries were determined in uncontaminated wheat samples spiked at levels of 100, 500 and 1000 μg/kg. Mean recoveries with SDs were 82.1 ± 2.02, 83.5 ± 1.95 and 86.9 ± 3.12 for FB1, FB2 and FB3, respectively. The detection limit was 10 μg/kg.

Eighty-two wheat samples produced in Iran were analyzed for the occurrence of Fusarium spp. and fumonisins. All samples were found positive for Fusarium species. A total of 386 strains were isolated and identified using morphological characters. All these strains belonged to four Fusarium spp.; F. graminearum, F. culmorum, F. proliferatum and F. verticillioides (Table 1). F. graminearum was the most prevalent species in all samples followed by F. verticillioides, F. proliferatum and F. culmorum. F. culmorum was not observed in the samples collected from Eilam, Gilan Gharb, Qasr Shirn and Sarpol Zohab (Table 1).

F. graminearum is one of the most frequently found Fusarium species on European cereals, where it is more common in the wet and warm climate of central and Southern Europe [26]. A species commonly found in wheat grown in cooler regions is F. culmorum [27,28,29,30,31,32]. In this study, we also identified F. graminearum isolates in all samples from the warm regions, mainly Kermanshah, Hamadan, Kurdistan and Eilam provinces, and F. culmorum isolates were only found in samples from moderate to cold regions of Iran (Table 1).

The occurrence of mycotoxins produced by Fusarium spp. in small cereal grains, particularly in wheat, is of great concern worldwide, because their presence in processed feeds and foods seems unavoidable. Consequently, they have been associated with chronic or acute mycotoxicoses in livestock and, to a lesser extent, in humans [2,33]. In this survey, the members of Gibberella fujikuroi complex, especially F. verticillioides and F. proliferatum, were detected at the highest frequencies after F. graminearum (Table 1). Our results are in agreement with other studies in the USA, Canada, Argentina [34,35] and Europe [2,11].

Of the 82 samples analyzed, 56 (68.2%) were found positive for FB1 contamination ranging from 15–155 μg/kg. The greatest FB1 contamination was detected in samples collected from Eilam ranging from 22–155 μg/kg followed by samples from Kangavar ranging from 65–142 μg/kg (Table 2). All samples (100%) from Asad Abad and Gilan Gharb were contaminated with FB1 ranging from 25–85 μg/kg. FB2 was detected in 35 (42.6%) samples ranging from 12–86 μg/kg. FB2 was not detected in samples collected from Kangavar (Table 2). FB3 could be detected in 26 (31.7%) samples ranging from 13–64 μg/kg. Highest FB2 and FB3 were detected in samples collected from Gilan Gharb ranging from 38–86 μg/kg and 25–64 μg/kg, respectively. FB3 was not detected in samples collected from Eilam, Kangavar, Sarpol Zohab and Sonqor.

In a survey of ethnic foods in the United Kingdom, one of four wheat based noodles was reported to contain 26 µg/kg fumonisins [36]. A survey of fumonisin contamination in cereals conducted in Spain reported FB1 in 8 of 17 wheat samples in the range of 0.2–8.8 mg/kg and FB2 in one sample (0.2 mg/kg) [37]. Our results show that 68.2% of wheat samples were contaminated with FB1 ranging from 15 to 155 ug/kg, 42.6% of samples were contaminated with FB2 ranging from 12 to 86 μg/kg, and 31.7% samples for FB3 ranging from 13–64 μg/kg. Of the 42 white wheat flour samples analyzed in Italy, five were contaminated at levels below 100 µg/kg; of the 214 white wheat flour samples from France, 76 were reported to be contaminated, mostly at levels below 100 µg/kg [26]. However, until today, no attempts have been made to elucidate the fumonisin levels in wheat grains and its products in Iran.