ZONs standards (ZON, α-ZOL, β-ZOL, α-ZAL and β-ZAL) and the internal standard, Zearalanone (ZAN), were purchased from Sigma Aldrich (Germany). Acetonitrile (HPLC grade), methanol (HPLC grade) and ammonium acetate (analytical grade) for the mobile phase, and methanol (analytical grade) for sample preparation were from Kanto Chemical (Japan). Potassium chloride (analytical grade) was from Wako Chemical (Japan). The SPE cartridge, Supelclean ENVI-18 6 ml (1 g) was from Supelco (USA). The glass microfiber filter GF/A was from Whatman (UK). The PTFE filter (pore size 0.2 μm) was from Toyo Roshi kaisha (Japan). The control two-raw malt was from Canada Malting (Canada) and the control beer was from Asahi Breweries (Japan), which were used for the method’s validation test.

The analysis was performed with an Alliance 2960 system (Waters, USA) for liquid chromatography and a Quattro Ultima triple stage quadrupole instrument equipped with an ESI interface (Waters) for mass spectrometry. The chromatographic separation was achieved on Nova-Pak phenyl column (150 × 2.0 mm i.d., 4 μm, Waters) at 40 °C, using a linear gradient program. Solvents A, B, and C were methanol, water, and 10 mM ammonium acetate, respectively. Elution was carried out at a flow rate of 0.2 mL applying the following gradient program: 0 min 20% A; 1 min 20% A; 20 min 60% A; 30 min 60% A; 31 min 20% A, and 5% C constant. The injection volume was 20 μL. For the mass spectrometry, the following settings were used in a negative ionization mode: capillary voltage 2.5 kV; cone voltage 90 V; source temperature 120 °C; desolvation temperature 300 °C. Table 1 shows the settings for a multiple reaction monitoring (MRM) mode.

The malt sample was prepared with reference to a previously described method [12]. Twenty-five grams of finely ground malt with 2.5 g of potassium chloride added was extracted with 100 mL of acetonitrile-water (75:25, v/v) by mixing with a magnetic stirrer for 15 minutes. After filtration through a glass microfiber filter, 10 mL of the extract was diluted with 90 mL water and was adjusted to pH 4.0 with acetic acid. 40 mL of this solution was applied to the SPE cartridge, Supelclean ENVI-18 (1 mg), preconditioned with 5 mL of methanol and 5 mL of water. After the passage of the sample solution through the cartridge, 5 mL of water was used for washing the cartridge. Then, ZONs were eluted with 5 mL methanol. The eluate was evaporated under a stream of nitrogen and the residue was well redissolved in 3 mL of methanol-water (75:25, v/v). After filtration through the PTFE filter, the sample solution was analyzed by LC/MS/MS. For the analysis of beer, samples were degassed by sonication for 45 minutes and a volume of 50 mL was applied directly to the SPE cartridge. The subsequent steps were carried out as described for the aforementioned malt analysis. For accurate quantification, the internal standard (ZAN) [12,13] was added before the sample preparation at a concentration of 10 ng/g. The standards of ZON and its metabolites were added to samples artificially contaminated with them before the sample preparation at levels of 2, 5, 10, 20, 50, 100, 200 ng/g (ppb) for the validation test and to have the calibration curve.

To characterize the metabolism of ZON in brewing fermentation, small-scale fermentation testing was conducted in our laboratory. Two liters of wort produced from uncontaminated malt was spiked with ZON at a concentration of 100 ng/g and was fermented at 12 °C for 18 days with 10 g of brewers’ yeast (Saccharomyces pastrianus), which is currently used for the production in our brewery. S. pastrianus is classified as a bottom-fermented yeast, which can sink to the bottom of the wort after fermentation. Meanwhile, another type of yeast used in beer production, S. cerevisia, is a top fermented yeast. Fourteen sampling points in the 18-day trial period were established as shown in Table 2. ZON and its metabolites’ levels were determined at each point, based on the beer analysis procedure in Section 2.2: “LC/MS/MS conditions” and Section 2.3: “Sample preparation”.

Consistent with the findings of previous studies [14], ZON and its metabolites showed higher sensitivity in the negative ionization mode when the tuning of MS/MS conditions were performed. Thus, the deprotonated molecule [M − H]⁻ was selected for each precursor ion. The product ion, most abundantly found in the fragmentation spectra of the standard, was selected as the quantification ion. The product ions second and third in abundance were also selected as qualification ions (Table 1). Nova-Pak phenyl column showed a good performance in separating ZONs on the chromatogram because ZONs naturally have benzene rings in their structures (Figure 2). The π-π interaction of the benzene ring between the phenyl column and ZONs is likely to be a contributing factor of the sharpened and well separated peak of ZONs.

Validation tests were performed for malt and beer. Table 3 shows the results for the malt analysis and Table 4 for the beer analysis. Good recoveries of 101.1%–119.7% were obtained with the spiked samples at a concentration of 10 ng/g for ZONs, with sufficient repeatability of RSD (repeatability standard deviation) = 2.2%–8.1% (n = 5). The calibration curves obtained with the spiked samples were linear at levels of 2, 5, 10, 20, 50, 100, 200 ng/g (ppb), showing correlation coefficients (r) of 0.9997–1.0000. The lowest level of LOQ (limit of quantification) in ZON and its metabolites was 2 ng/g (ppb) for beer and malt, which was the minimum point of the analytical line.

Table 2 shows temporal changes in concentration of ZON and its metabolites during the small-scale fermentation test using commercial wort and brewers’ yeast (Saccharomyces pastrianus). The concentration of ZON decreased from 100ng/g to 9.3 ng/g in 18 days. On the other hand, the concentration of β-ZOL, a ZON metabolite, increased throughout the fermentation process and eventually reached up to 85.9 ng/g in the last day of the trial. A slight amount of α-ZOL, an isomer of β-ZOL, was also detected from day 4 onward, reaching 2.0 ng/g in day 18 in the trial period. Furthermore, α-ZAL or β-ZAL was not detected in this fermentation trial. Figure 3 also shows that the level of β-ZOL exceeded that of ZON, 96 hours after the trial started. These results suggest that ZON is mainly metabolized to β-ZOL during brewing fermentation and that β-ZOL is not further metabolized to β-ZAL by brewers’ yeast unlike the way it is metabolized in the liver [7,8]. The reason for this is likely that a carbonyl group is reduced by brewers’ yeast (Figure 1) [15]. On the other hand, reduction of the double bond to a single bond was not observed when the reduction is made inside the liver.

According to previous reports, β-ZOL has an approximately 10 to 20-fold weaker binding affinity for ER than ZON and some 100-fold weaker than α-ZOL [9,16]. Increase in the concentration of β-ZOL suggests that the health risk posed by ZON contamination in beer is maintained at an extremely low level because β-ZOL exhibits lower estrogenic activity than ZON does.