Testing the extraction of 12 mycotoxins from aqueous solutions by insoluble beta-cyclodextrin bead polymer

Mycotoxins are toxic metabolites of filamentous fungi; they are common contaminants in numerous foods and beverages. Cyclodextrins are ring-shaped oligosaccharides, which can form host-guest type complexes with certain mycotoxins. Insoluble beta-cyclodextrin bead polymer (BBP) extracted successfully some mycotoxins (e.g., alternariol and zearalenone) from aqueous solutions, including beverages. Therefore, in this study, we aimed to examine the ability of BBP to remove other 12 mycotoxins (including aflatoxin B1, aflatoxin M1, citrinin, dihydrocitrinone, cyclopiazonic acid, deoxynivalenol, ochratoxin A, patulin, sterigmatocystin, zearalanone, α-zearalanol, and β-zearalanol) from different buffers (pH 3.0, 5.0, and 7.0). Our results showed that BBP can effectively extract citrinin, dihydrocitrinone, sterigmatocystin, zearalanone, α-zearalanol, and β-zearalanol at each pH tested. However, for the removal of ochratoxin A, BBP was far the most effective at pH 3.0. Based on these observations, BBP may be a suitable mycotoxin binder to extract certain mycotoxins from aqueous solutions for decontamination and/or for analytical purposes.


Introduction
Mycotoxins, the toxic secondary metabolites of filamentous fungi, are common food contaminants (Bennett and Klich 2003;da Rocha et al. 2014). Aflatoxins are mainly produced by Aspergillus flavus and Aspergillus parasiticus. They were isolated in the 1960s, after the death of more than 100,000 turkeys in "turkey X" disease, due to the consumption of aflatoxin-contaminated peanut meal (Bennett and Klich 2003). Aflatoxins appear in nuts, cereals, figs, vegetables, meat, and spices, possessing primarily hepatotoxic, mutagenic, and carcinogenic effects (Bennett and Klich 2003;da Rocha et al. 2014;Klingelhöfer et al. 2018). The International Agency for Research on Cancer (IARC) classified aflatoxins as Group 1 carcinogens (IARC 2012). Aflatoxin B1 (AFB1) is the most frequent and the most toxic member of this group, while aflatoxin M1 (AFM1) is a metabolite of AFB1 which is a common contaminant in milk ( Fig. 1) (Bennett and Klich 2003;Klingelhöfer et al. 2018;Smith and Groopman 2019). Sterigmatocystin (STC; Fig. 1) is a precursor in the biosynthesis of aflatoxins; it exerts mutagenic, carcinogenic, and teratogenic effects and classified as a possible carcinogen (Group 2B) by the IARC (Veršilovskis and De Saeger 2010). STC contaminates typically rapeseed, peanut, spices, and cereals (e.g., wheat, barley, and rice); furthermore, it has also been detected in beer, cocoa, and coffee beans (Veršilovskis and De Saeger 2010). Cyclopiazonic acid (CPA; Fig. 1) was isolated from Penicillium cyclopium; nevertheless, several Penicillium and Aspergillus molds can produce CPA (Bennett and Klich 2003). It appears as a contaminant in oilseeds, cereals, nuts, maize, meat, milk, egg, and peanut (Ostry et al. 2018). The acute toxicity of CPA is low; however, based on animal studies, the chronic exposure to the mycotoxin may cause degenerative changes in the gastrointestinal tract, kidney, liver, and central nervous system (Ostry et al. 2018). Ochratoxin A (OTA) and citrinin (CIT), produced by Aspergillus, Penicillium, and/or Monascus species, are nephrotoxic mycotoxins ( Fig. 1) (EFSA 2012;. CIT frequently appears as a contaminant in grains (e.g., wheat, barley, oat, and rye), rice, beans, peas, spices, nuts, and fruits (EFSA 2012), while OTA occurs for example in cereals, fruits, meat, spices, cacao, chocolate, coffee, tea, beer, and wine (EFSA 2020). IARC classified OTA as a possible human carcinogen (Group 2B) (EFSA 2020). Dihydrocitrinone (DHC; Fig. 1) is the major urinary metabolite of CIT, which is less toxic and more hydrophilic than the parent mycotoxin Degen et al. 2018). DHC is not a food contaminant; however, we also examined its extraction from buffers, because the cyclodextrin polymer tested may also be suitable for analytical sample preparation regarding body fluids. Patulin (PAT; Fig. 1) is formed by Aspergillus and Penicillium species. It occurs in different fruits (especially in apple and pear) and in the corresponding products (e.g., fruit juices) (Vidal et al. 2019). Acute PAT intoxication causes gastrointestinal disturbances (e.g., nausea, vomiting, ulceration, and lesions), while its mutagenic, neurotoxic, immunotoxic, genotoxic, teratogenic, and carcinogenic effects have also been reported as a result of the chronic exposure (Puel et al. 2010;Vidal et al. 2019). Deoxynivalenol (DON or vomitoxin; Fig. 1) is a trichothecene mycotoxin produced by Fusarium species (e.g., Fusarium graminearum and Fusarium culmorum) (Ji et al. 2019). DON is one of the most common mycotoxin contaminants in cereals; the exposure can cause gastrointestinal disorders and weight loss as well as the teratogenic and immunotoxic effects of this mycotoxin have also been reported (Ji et al. 2019). Zearalenone is a Fusarium-derived mycotoxin; it commonly appears in cereals (e.g., maize, barley, oat, and wheat) and related products (e.g., beer) (EFSA 2017). Despite its non-steroidal structure, zearalenone (and some of its metabolites) binds to estrogen receptors and consequently exerts xenoestrogenic effects; while other harmful (e.g., immunotoxic, nephrotoxic, hepatotoxic, and hematotoxic) impacts are also attributed to this mycotoxin (Ji et al. 2019). The phase I metabolites of zearalenone are αzearalenol, β-zearalenol, zearalanone (ZAN; Fig. 1), αzearalanol (α-ZAL; Fig. 1), and β-zearalanol (β-ZAL; Fig. 1) (EFSA 2017). The presence of ZAN and ZALs has been reported in maize products, rice, and soy meal (Ji et al. 2019); while α-ZAL is applied as a growth promoter in certain farm animals in non-EU countries (EFSA 2017). Some zearalenone derivatives, including αzearalenol and α-ZAL, exert significantly higher estrogenic action than the parent mycotoxin (EFSA 2017).
In the current explorative study, we aimed to investigate the extraction of 12 mycotoxins, namely AFB1, AFM1, CIT, DHC, CPA, DON, OTA, PAT, STC, ZAN, α-ZAL, and β-ZAL ( Fig. 1), from different buffers (pH 3.0, 5.0, and 7.0) by insoluble water-swellable β-CD bead polymer (BBP). Our results demonstrate which mycotoxins can be effectively removed from aqueous solution and give a good starting point for the planning of further and deeper investigation regarding the extraction of these mycotoxins from different solutions (including beverages) for decontamination or analytical purposes.
To examine the impact of the environmental pH on mycotoxin removal, the same experiments were performed at pH 3.0 (0.05 M sodium phosphate) and pH 7.0 (0.05 M sodium phosphate), applying 0.0, 1.67, and 6.67 mg/mL final BBP concentrations. However, these buffers interfered with the efficiency of the HPLC method applied for the analyses of ZAN and ZALs. Therefore, the latter mycotoxins were incubated in sodium tartrate (0.05 M, pH 3.0) and TRIS-HCl (0.05 M, pH 7.0) buffers.
Most of the supernatants were directly injected into the HPLC after the sedimentation of BBP. Nevertheless, pH adjustment of certain samples was reasonable for the appropriate conditions of HPLC analyses. The 500 μL aliquots of these supernatants were acidified or alkalinized based on the followings. The pH 3.0 AFM1 supernatants were alkalinized with 3 μL of 1 M NaOH. The pH 5.0 and pH 7.0 CIT samples were acidified with 8 and 10 μL of 1.5 M HCl, respectively. Similarly, pH 7.0 DHC supernatants were acidified with 10 μL of 1.5 M HCl. The pH 3.0 CPA samples were alkalinized with 5 μL of 0.5 M NaOH. The pH 7.0 DON supernatants were acidified with 8 μL of 1.5 M HCl. OTA samples were alkalinized with 7 μL of 3 M NaOH (pH 3.0 samples) or 7 μL of 1 M NaOH (pH 5.0 and pH 7.0 samples). The pH 3.0 STC supernatants were alkalinized with 3 μL of 1 M NaOH, while pH 7.0 STC samples were acidified with 8 μL of 1.5 HCl.

HPLC analyses
CIT, DHC, and OTA were analyzed by an integrated HPLC system (Jasco, Tokyo, Japan), which included an autosampler (AS-4050), a binary pump (PU-4180), and a fluorescence detector (FP-920). Chromatograms were evaluated using ChromNAV software (Jasco, Tokyo, Japan). Furthermore, AFB1, AFM1, CPA, DON, PAT, STC, ZAN, and ZALs were analyzed by an integrated HPLC system built up from a Waters 510 HPLC pump (Milford, MA, USA), a Rheodyne 7125 injector (Berkeley, CA, USA) with a 20-μL sample loop, and a Waters 486 UV detector (Milford, MA, USA). Chromatograms were evaluated employing Millennium Chromatography Manager software (Waters, Milford, MA, USA). Each HPLC analysis was performed with isocratic elution using 1.0 mL/min flow rate at room temperature, and 20 μL volume of samples was injected.

Statistics
Data represent mean ± SEM values at least from three independent measurements. One-way ANOVA with Tukey's post hoc test was applied to establish the statistical significance (p < 0.01), employing SPSS Statistics software (version 24; IBM, Armonk, NY, USA).

Results
Extraction of mycotoxins from sodium acetate buffer (pH 5.0) by BBP To test the mycotoxin binding of BBP, increasing amounts of the polymer were added to standard concentration of mycotoxins (each 2 μM in 1.5 mL volume) in sodium acetate buffer (pH 5.0). Fig. 2a demonstrates the mycotoxins which were extracted with less than 75% efficacy by 13.3 mg/mL (or 20.0 mg/1.5 mL) BBP. The bead polymer barely affected DON and PAT contents of the solutions. Furthermore, approximately 28 and 35% decreases in the concentrations of AFM1 and OTA were caused by 13.3 mg/mL BBP, respectively. In addition, more than 50% of AFB1 and CPA were removed by the same amount of the polymer. Fig. 2b represents the mycotoxins which were extracted with 75% or even better efficacy by 13.3 mg/mL BBP. Among these mycotoxins, approximately 75% of CIT and DHC were extracted, followed by STC (80%). Interestingly, the lower concentrations of BBP (0.67 to 3.33 mg/mL) induced a much steeper decrease in the STC content of the solution compared to CIT and DHC (Fig. 2b). Moreover, BBP proved to be the strongest binder of ZAN and ZALs, removing approximately 90-95% of these mycotoxins at 13.3 mg/mL concentration.

Testing the pH dependence of mycotoxin extraction
The pH dependence regarding the mycotoxin binding of BBP was also examined. Since the pH of beverages is typically in the acidic or neutral range (Feldman and Barnett 1995), we tested the mycotoxin extraction between pH 3.0 and pH 7.0 (see details in "Mycotoxin Extraction" section). Fig. 3 illustrates the removal of mycotoxins at pH 3.0, pH 5.0, and pH 7.0 by 1.67 and 6.67 mg/mL BBP. Under the applied conditions, we did not find significant differences regarding AFM1, DON, DHC, PAT, and ZAN (Fig. 3). CIT was the only mycotoxin where a little bit higher mycotoxin removal was observed at pH 7.0 vs. pH 5.0; however, only the higher BBP concentration caused statistically significant difference (Fig.  3c). Furthermore, the slightly lower removal of AFB1 (6.67 mg/mL BBP), STC (6.67 mg/mL BBP), α-ZAL (6.67 mg/mL BBP), and β-ZAL (1.67 and 6.67 mg/mL BBP) was noticed at pH 3.0 than at pH 5.0. In the presence of 1.67 mg/mL BBP, the decrease in CPA content was the largest at pH 3.0; however, we did not observe pH-dependent differences when 6.67 mg/mL polymer concentration was applied. Despite the above-listed statistically significant differences regarding the mycotoxin removal in different buffers, the only relevant pH effect was demonstrated by OTA. The decrease in the pH to 3.0 considerably enhanced the removal of OTA by BBP compared to both pH 5.0 and pH 7.0 (Fig. 3g). Therefore, the extraction of OTA at pH 3.0 was also tested with each BBP concentration applied in "Mycotoxin Extraction" section. As it is demonstrated in Fig. 4, the lower pH favors the interaction of OTA with BBP, leading to the strong decrease in the mycotoxin content at pH 3.0 and resulting in more than 80% removal of OTA by 13.3 mg/mL BBP.

Discussion
Few studies demonstrated that BBP may be a promising candidate for the removal of some mycotoxins from aqueous solutions: For example, alternariol and zearalenone have been successfully extracted from buffers and from certain beverages (wine and beer, respectively) (Poór et al. 2018;Fliszár-Nyúl et al. 2019. Therefore, in the current explorative study, we aimed to examine the ability of BBP to extract other 12 mycotoxins (AFB1, AFM1, CIT, CPA, DON, DHC, OTA, PAT, STC, ZAN, α-ZAL, and β-ZAL) from aqueous buffers. Mycotoxins which were extracted with less than 75% efficacy by 13.3 mg/mL (or 20.0 mg/1.5 mL) BBP; b: mycotoxins which were extracted with 75% or even better efficacy by 13.3 mg/mL BBP Mycotoxins are common contaminants in food and beverages (e.g., milk, coffee, beer, wine, and fruit juices) (Bennett and Klich 2003;Veršilovskis and De Saeger 2010). Since the pH of these drinks are typically acidic or neutral (Feldman and Barnett 1995), our experiments were performed between pH 3.0 and pH 7.0.
Only limited data are available regarding the interaction of mycotoxins with CD polymers. The successful extraction of OTA and PAT by polyurethane-β-CD polymer has been reported from aqueous solutions, including wine and apple juice, respectively Jackson 2010, 2012). In addition, BBP considerably decreased the mycotoxin (alternariol, zearalenone, α-zearalenol, β-zearalenol, zearalenone-14-glucoside, and zearalenone-14-sulfate) content of aqueous solutions and effectively removed alternariol and zearalenone from wine and from beer samples, respectively (Poór et al. 2018;Faisal et al. 2019aFaisal et al. , 2020Fliszár-Nyúl et al. 2019. At pH 5.0, BBP produced the highest removal of ZAN and ZALs; however, it seems to be a suitable binder of CIT, DHC, and STC as well (Fig. 2b). No data are available regarding the interactions of ZAN, ZALs, and STC with CDs; however, zearalenone forms stable complexes with β-CD (K ≈ 10 4 L/mol) and was successfully removed from aqueous solutions by BBP (Poór et al. 2018). Nevertheless, the binding constants of CIT-β-CD and DHC-β-CD complexes are low (K ≈ 10 2 L/mol) (Zhou et al. 2012;Poór et al. 2016;Faisal et al. 2019b); therefore, the removal of these mycotoxins by BBP is unexpectedly high. Similar phenomenon was observed at pH 3.0 with OTA (Fig. 4), despite the fact that the mycotoxin forms poorly stable complexes with β-CD (K ≈ 10 2 L/mol) (Poór et al. 2015a). In addition, Verrone et al. reported the highest affinity of β-CD towards dianionic OTA (both carboxyl and phenolic hydroxyl groups are deprotonated), followed by the nonionized and the monoanionic (only the carboxyl group is deprotonated) forms (Verrone et al. 2007). These results indicate the cooperative interactions of CD rings in BBP with CIT, DHC, and OTA, as it has also been observed regarding mycotoxin alternariol and some other compounds (Harada et al. 1976;Saenger 1980;Fliszár-Nyúl et al. 2019). Under the applied conditions, BBP (13.3 mg/mL) removed approximately 50% of AFB1 and CPA; however, the polymer only slightly decreased the concentrations of AFM1, PAT, and DON (Fig. 2a).  The pH dependence of mycotoxin extraction was tested in the pH range 3.0 to 7.0. No or only slight changes were observed in the extraction of mycotoxins tested, except OTA (Fig. 3). BBP strongly decreased the OTA content of the solution at pH 3.0 (Figs. 3g and 4.). These data suggest that BBP mainly interacts with the nonionic form of OTA, which is also supported by the effective removal of the mycotoxin from red wine by polyurethane-β-CD polymer (Appell and Jackson 2012). Amadasi et al. suggest the inclusion of the phenyl ring of L-phenylalanine in the CD cavity (Amadasi et al. 2007). Furthermore, the protruding parts of OTA (the carboxyl group and the isocoumarin moiety) form hydrogen bonds with the outer hydroxyl groups of the CD, which can further stabilize the inclusion (Amadasi et al. 2007). The carboxyl and phenolic hydroxyl groups of OTA can be ionized (acid dissociation constants are 4.2-4.4 and 7.0-7.3, respectively) (Perry et al. 2003). The deprotonation of the carboxyl and/or the phenolic hydroxyl group(s) at higher pH (e.g., pH 5.0 and pH 7.0) may explain the lower efficacy of BBP regarding OTA removal.
The Langmuir and Freundlich sorption isotherms are suitable for the quantitative evaluation of mycotoxin-BBP interactions (Appell and Jackson 2012;Faisal et al. 2019a;Fliszár-Nyúl et al. 2019). The Langmuir model typically characterizes a strictly homogenous monolayer adsorption, while the Freundlich isotherm does not need this restriction (Ayawei et al. 2017). Based on the Freundlich model, the heterogeneity index (n) was close to one, indicating the relatively homogenous sorption of OTA by BBP. In our previous studies, the extraction of zearalenone (Poór et al. 2018) and alternariol (Fliszár-Nyúl et al. 2019) was also tested from aqueous buffers by BBP. Regarding this polymer, the Q 0 values of OTA and zearalenone were similar, while it was significantly higher for alternariol. However, both the Langmuir equilibrium constant (K L ) and the adsorptive capacity (K F , determined applying the Freundlich model) demonstrate the weaker interaction of BBP with OTA (K L = 0.12 L/mg; K F = 0.49 (mg/g) × (L/mg) 1/n ) compared to zearalenone (K L = 0.60 L/ mg; K F = 1.16 (mg/g) × (L/mg) 1/n ) and alternariol (K L = 0.16 L/mg; K F = 5.52 (mg/g) × (L/mg) 1/n ). These data are also in agreement with our observations that the removal of OTA by BBP is less effective vs. zearalenone or alternariol (Poór et al. 2018;Fliszár-Nyúl et al. 2019).
For comparison of the mycotoxin binding ability of BBP with other CD polymers, adsorbents, and nanoparticles, our results were combined with previously reported data in Table 1, including the mycotoxin binder used, the mycotoxin extracted, the environmental conditions, and the toxin removal.
In conclusion, the extraction of 12 mycotoxins by BBP was tested in different buffers (pH 3.0, 5.0, and 7.0). BBP induced the concentration-dependent decrease in the mycotoxin content and proved to be an effective binder of CIT, DHC, OTA, STC, ZAN, and ZALs. Among the mycotoxins tested, only the extraction of OTA showed considerable pH dependence: its removal by BBP was far the most effective at pH 3.0. Our results suggest that BBP may be a suitable mycotoxin binder to extract certain mycotoxins from aqueous solutions for decontamination and/ or for analytical purposes.
Acknowledgements The authors thank Katalin Fábián for her excellent assistance in the experimental work.
Availability of data and materials The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Author contribution MP and LS conceived the study. MP, VM, and ZF wrote the paper. VM and ZF performed mycotoxin extraction experiments. EF-N performed HPLC analyses. All authors read and approved the final manuscript.

Declarations
Ethics approval and consent to participate Not applicable.
Consent for publication Not applicable.

Competing interests
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.