Advertisement

Analytical and Bioanalytical Chemistry

, Volume 410, Issue 22, pp 5583–5592 | Cite as

Identification of acetylated derivatives of zearalenone as novel plant metabolites by high-resolution mass spectrometry

  • Laura Righetti
  • Luca Dellafiora
  • Daniele Cavanna
  • Enrico Rolli
  • Gianni Galaverna
  • Renato Bruni
  • Michele Suman
  • Chiara Dall’Asta
Research Paper
Part of the following topical collections:
  1. Food Safety Analysis

Abstract

Zearalenone (ZEN) major biotransformation pathways described so far are based on glycosylation and sulfation, although acetylation of trichothecenes has been reported as well. We investigated herein the ZEN acetylation metabolism route in micropropagated durum wheat leaf, artificially contaminated with ZEN. We report the first experimental evidence of the formation of novel ZEN acetylated forms in wheat, attached both to the aglycone backbone as well as on the glucose moiety. Thanks to the advantages provided by high-resolution mass spectrometry, identification and structure annotation of 20 metabolites was achieved. In addition, a preliminary assessment of the toxicity of the annotated metabolites was performed in silico focusing on the toxicodynamic of ZEN group toxicity. All the metabolites showed a worse fitting within the estrogen receptor pocket in comparison with ZEN. Nevertheless, possible hydrolysis to the respective parent compounds (i.e., ZEN) may raise concern from the health perspective because these are well-known xenoestrogens. These results further enrich the biotransformation profile of ZEN, providing a helpful reference for assessing the risks to animals and humans.

Graphical abstract

Keywords

Food safety Masked mycotoxins Acetylation Plant biotransformation 

Introduction

Zearalenone (ZEN) is a secondary metabolite produced by several Fusarium species during infection of edible plants. It occurs predominantly in cereal grains, such as maize, wheat, rice, oats, and barley [1, 2].

It has been proven [3, 4, 5, 6, 7, 8] that plants employ different phase I and phase II detoxification mechanisms to cope with the adverse effect of xenobiotics, such as mycotoxins. In case of mycotoxins, multiple enzymatic pathways may lead to the formation of modified forms that, once they enter the food and feed production chain, may significantly contribute to the overall toxic load related to mycotoxin exposure [9].

Understanding the plant metabolism of mycotoxins and thus the toxicological role of resulting modified forms is becoming increasingly important for risk assessment. ZEN has low acute toxicity for animals and humans, but represents a public health concern in case of chronic exposure since it is considered to be an endocrine disruptor with estrogen-like properties [10].

Recently, the European Food Safety Authority CONTAM Panel set a group tolerable daily intake (group TDI) based on relative potency factors, calculated for ZEN and its modified forms [9]. Therefore, ZEN contamination should be expressed as the overall amount of the parent compound and its known modified forms.

Exposure to ZEN and its modified forms is primarily due to the consumption of contaminated cereal grains, mainly whole grains and bran-enriched products. Recent studies reported the co-occurrence of ZEN, α- and β-zearalenol (ZEL), and their gluco-conjugated derivatives in cereal-based products at significant concentration levels [11, 12].

ZEN modified forms originating from biotransformation in plants have been previously described, including both phase I and phase II metabolites [3, 7, 13]. Berthiller and co-workers [3] described more than 10 glycosylated metabolites of ZEN, α-ZEL, and β-ZEL in the model plant Arabidopsis thaliana, including malonylglucosides, dihexosides, and pentosylhexoside derivatives. Few years later, the identification of a barley UDP-glucosyltransferase enabled the characterization of a new glucose conjugate, named zearalenone-16-O-β-glucoside [7]. Moreover, an extensive study of the metabolism of ZEN in wheat was recently performed by our group [8], leading to the annotation of novel reductive metabolites as well as sulfate conjugates of α and β-ZEL.

The major biotransformation pathways investigated and described so far for Fusarium mycotoxins are based on glycosylation and sulfation. Acetylation of some trichothecenes has been described as well. For instance, deoxynivalenol (DON) mono- and di-acetylated formation and fate have been recently investigated in isolated wheat cells [14]. The same authors identified for the first time 15-acetyl-DON-3-O-β-D-glucoside (15-AcDON3Glc) as biotransformation product of DON in wheat. However, since 15-AcDON was not detected, it was assumed that the acetylation (phase I reaction) might also take place after the conjugation (phase II reaction) of glucose to DON [4]. Also for T2 toxins, both acetylation on the parent compound (3-Acetyl-T2) and on their phase II metabolites (i.e., 15-Acetyl-T-2-tetraol-glucoside) were detected in barley and oats using high resolution mass spectrometry (HRMS) [5, 6].

Although 14-acetyl ZEN and 14-acetyl-cis-ZEN were isolated from a Fusarium graminearum strain cultured on maize [15], their occurrence in wheat has not been reported so far.

Regarding the toxicological relevance of the above mentioned acetylated forms, experimental data are available for 3- and 15-AcDON [16], but not for T2 or ZEN metabolites.

Apart from their own toxicity, modified mycotoxins are of concern in consideration of their possible cleavage during digestion in mammals [17, 18, 19]. Similarly, deacetylation reaction has been described [20]. According to the authors, oral administration of 3-AcDON in pigs resulted in the presence of DON and DON-glucuronide in blood, but no 3-AcDON, suggesting that AcDONs were rapidly deacetylated by intestinal lipases [20].

Although relevant for an accurate risk assessment, the toxicological evaluation of such metabolites is strongly hindered by the unavailability of commercial reference compounds. In addition, the high costs of synthesis or purification of these compounds make the analysis of toxicity via canonical experimental trials essentially unaffordable. However, toxicity assessment in silico proved to be a reliable first step in the hazard identification process identifying those compounds that are worthy of more detailed investigations [21, 22, 23].

Herein, we present the first experimental evidence of formation of novel ZEN acetylated forms in wheat. In addition, a preliminary assessment of the toxicity of the annotated metabolites has been performed in silico, focusing on the toxicodynamic of ZEN group toxicity.

These results further enrich the biotransformation profile of ZEN, providing a helpful reference for assessing the risks to animals and humans.

Materials and methods

Chemicals and reagents

Analytical standards of ZEN (100 μg mL-1 in acetonitrile), α-ZEL (solution in acetonitrile 10 μg mL-1), and β-ZEL (solution in acetonitrile 10 μg mL-1) were obtained from Sigma-Aldrich (Taufkirchen, Germany). HPLC-grade methanol, acetonitrile, and acetic acid, as well as dimethylsulfoxide (DMSO) were purchased from Sigma-Aldrich (Taufkirchen, Germany); bidistilled water was obtained using a Milli-Q System (Millipore, Bedford, MA, USA). MS-grade formic acid from Fisher Chemical (Thermo Fisher Scientific Inc., San Jose, CA, USA) and ammonium acetate (Fluka, Chemika-Biochemika, Basil, Switzerland) were also used.

Plant material and ZEN administration

Aseptyc cultures of durum wheat (Triticum durum Desf.) were obtained from micropropagated material following a previously described protocol [8]. Briefly, leaves (3–5 cm in length) were placed in 50 mL tubes filled with liquid medium containing ZEN solution. To allow constant exposure of the emerging organ in liquid medium with ZEN, leaves were anchored to the bottom of tubes by immersion of the basal part in a fine layer of solid medium. Liquid medium without mycotoxin was used in all experiments as a control. At the end of the experiment, leave cultures exposed to 100 μg/L ZEN showed no visible degradation.

Sample extraction and analysis

Sample extraction and analysis were performed following the procedure previously described [8]. Leaf samples were freeze dried for 24 h (lyophylizator LIO-5PDGT, 5Pascal s.r.l., Trezzano sul Naviglio, Milan, Italy) and then milled.

Leaf samples (50 mg of homogenized material) were extracted with a mixture of acetonitrile/water/formic acid (79:20:1, v/v), and stirred for 90 min at 200 strokes/min in a shaker. The extract was centrifuged for 10 min at 14,000 rpm at room temperature, then 500 μL of supernatant was evaporated to dryness under nitrogen and finally reconstructed with 500 μL of water/methanol (80:20, v/v) prior to LC-MS analysis.

UHPLC Dionex Ultimate 3000 separation system coupled to a Q-Exactive high resolution mass spectrometer (Thermo Scientific, Bremen, Germany) equipped with an electrospray source (ESI) was employed. Chromatographic separation was performed as previously described [8]. The Q-Exactive mass analyzer was operated in the full MS/data-dependent MS/MS mode (full MS–dd-MS/MS) at following parameters: sheath and auxiliary gas flow rates 32 and 7 arbitrary units, respectively; spray voltage 3.3 kV; heater temperature 220 °C; capillary temperature 250 °C, and S-lens rf level 60. Following parameters were used in full MS mode: resolution 70,000 FWHM (defined for m/z 200; 3 Hz), scan range 100–1000 m/z, automatic gain control (AGC) target 3e6, maximum inject time (IT) 200 ms. Parameters for dd-MS/MS mode: intensity threshold 1e4, resolution 17,500 FWHM (defined for m/z 200; 12 Hz), scan range 50 – fragmented mass m/z (m/z +25), AGC target 2e5, maximum IT 50 ms, normalized collision energy (NCE) 35% with ±25% step.

The annotation process was performed following three criteria; (1) fitting of measured and theoretical accurate mass of [M-H]¯ with a mass tolerance set at ±5 ppm, (2) experimental and theoretical isotopic pattern comparison, (3) MS-MS spectra: product ion of intact ZEN (m/z 317.1389) and ZEL (m/z 319.1550) and/or comparison of the fragments obtained with the fragmentation pathway of ZEN or other mycotoxins metabolites formerly found [3, 4, 5, 6, 7, 8]. Only in few cases, fragmentation spectra could not be collected, due to parent ion abundance below the threshold. In this case, a tentative annotation based on accurate mass and elemental formula was performed, as already proposed by other authors [3, 4, 5, 6, 7, 8].

Assessment of toxicology in silico

The toxicity of all the acetylated metabolites of ZEN and ZELs was assessed using an already published and validated in silico framework [21]. Briefly, the potential xenoestrogenicity is deduced on the basis of the capability of ligands to fit the pocket of the estrogen receptors where modifications preventing a proper ligand-pocket arrangement may correlate with a reduced capability to trigger the toxic stimulus [22, 23].

In more detail, the 3D model for the alpha estrogen receptor (ERα) was derived from the Protein Data Bank (http://www.rcsb.org) structure having PDB code 2YJA [24]. The receptor structure and all ligands were processed according to Dellafiora and co-workers [25]. In particular, the 3D structures of ZEN and ZELs were retrieved from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/), whereas all the other metabolites were edited using the respective parent compounds with the software Sybyl, ver. 8.1 (www.certara.com). All atoms were checked for atom- and bond-type assignments. Amino- and carboxyl-terminal groups were set as protonated and deprotonated, respectively. Hydrogen atoms were computationally added to the protein and energy-minimized using the Powell algorithm with a coverage gradient of ≤0.5 kcal (mol Å)−1 and a maximum of 1500 cycles.

The toxin binding site was defined by using the Flapsite tool of the FLAP (Fingerprint for Ligand and Protein) software (Molecular Discovery Ltd; http://www.moldiscovery.com) [26], whereas the corresponding pharmacophoric space was investigated by using the GRID algorithm [27]. Specifically, the DRY probe was used to describe the hydrophobic environment, whereas the sp2 carbonyl oxygen (O) and the neutral flat amino (N1) probes were used to describe the hydrogen bond acceptor and donor capacity of the pocket, respectively.

The interaction between all the acetylated metabolites of ZEN and ZELs and the αER was assessed using docking simulations coupled with a rescoring procedure to provide a more precise and reliable estimation of the ligand-receptor interaction. In particular, the integrated use of the docking software GOLD (Genetic Optimization for Ligand Docking) and the HINT (Hydropathic INTeraction) rescoring function [28] was chosen on the basis of previous studies demonstrating the high reliability in estimating the ligand binding free energies and in evaluating the bioactivity of small ligands, including xenoestrogenicity [29]. In addition, the HINT scores provide quantitative evaluation of protein-ligand interaction being proportionally linked to the thermodynamic favors of protein-ligand complex formation [30]. Software setting and rescoring procedures reported by Ehrlich et al. were used [23]. Briefly, 50 poses for each ligand have been generated and rescored with HINT. Only the best scored pose for each ligand was taken into consideration. Additionally, taking into account that GOLD implements a Lamarckian genetic algorithm that may introduce variability in the results, all the analyses were done in triplicate to avoid a non-causative scores assignment. Therefore, each score is represented as mean value ± standard deviation. All the positive scores were statistically analyzed using IBM SPSS Statistics for Windows, ver. 19 (IBM Corp., Armonk, NY, USA) with one-way ANOVA (α = 0.05), followed by post hoc Bonferroni test (α = 0.05). All images were obtained using the software PyMol ver. 1.7 (http://www.pymol.org).

Results and discussion

Structure elucidation by HRMS/MS

After 14 days of incubation with a ZEN-enriched medium, leaves were analyzed by UHPLC-HRMS. A qualitative screening of biotransformation metabolites was performed, focused on the acetylation route. Experimentally acquired HRMS/MS data were compared with in silico fragmentation obtained uploading in-house generated SDF files into MetFrag [31] (data not shown) (Fig. 1).
Fig. 1

Chemical structures of ZEN and its acetylated metabolites

As a result, a total of 20 acetylated derivatives of ZEN and ZELs were putatively identified. The representative EICs chromatogram (with mass window of 5 ppm) and HRMS/MS spectra of ZEN metabolites are shown in Figs. 2, 3, and 4, and in Figs. S1 and S2 in the Electronic Supplementary Material (ESM). Detailed information about the accurate mass and retention time of each metabolite is listed in Table 1. The relative mass deviations between theoretical and measured precursor and fragment ions did not exceed 4 ppm.
Fig. 2

Extracted ion chromatograms (EICs) of accurate mass traces (±5 ppm) of ZEN and its acetylated metabolites. Metabolites marked as #4, #6, #7, and #8 were detected in the Ac-ZEN EIC (m/z 359.1502) since they correspond to in-source fragmentation of Ac-ZEN conjugated forms described in the following paragraphs

Fig. 3

EICs and fragmentation pattern of ZEN acetyl metabolites: Ac-ZEN-Glc (A), and ZEN-diGlc-Ac (B). LC-HRMS/MS spectrum of deprotonated adduct, showing the loss of different aglycone corresponding to ZEN (C), and Ac-ZEN (D)

Fig. 4

EICs of ZEL acetyl metabolites: AC-ZEL-Glc (A), and ZEL-diGlc-Ac (B)

Table 1

Exact mass values (calculated and observed) of target compounds and relative ions at negative polarity

   

[M-H]-

 

[M+CH3COO]-

No.

Putative ZEN metabolite

Elemental formula

Theoretical m/z a

Experimental m/z b

Mass deviation ppm

RT

Theoretical m/z

Experimental m/z

Mass deviation ppm

 

ZEN

C18H22O5

317.1389

317.13904

-1.283

12.14

377.1595

N.D.

N.D.

1

Ac-ZEN

C20H24O6

359.1489

359.1502

0.469

15.09

419.1700

N.D.

N.D.

2

Ac-ZENGlc

C26H34O11

521.2017

521.20203

-1.544

4.63

581.2229

581.2229

-1.831

3

521.2022

-1.199

5.18

581.2232

-1.298

4

521.2021

-1.429

5.62

N.D.

N.D.

8

Ac-ZEN-diGlc

C32H44O16

683.2546

683.2542

-2.12

5.13

743.2757

N.D.

N.D.

9

683.2550

-0.963

5.49

N.D.

N.D.

5

ZEN-AcGlc

C26H34O11

521.2017

521.20227

-1.084

8.8

581.2229

581.22205

-3.294

6

521.2018

-1.909

9.31

581.2230

-1.625

7

521.2020

-1.544

9.75

581.2217

-3.827

10

ZEN-AcdiGlc

C32H44O16

683.2546

683.2549

-1.051

9.03

743.2757

743.2761

-0.979

11

683.2546

-1.578

9.46

743.2755

-1.719

 

β-ZEL

C18H24O5

319.1550

319.1548

-0.085

10.39

379.1751

N.D.

N.D.

 

α-ZEL

319.1541

-2.404

11.94

N.D.

N.D.

12

ZEL-Glc-Ac

C26H36O11

523.2174

523.19556

-3.432

7.06

583.2385

583.2387

-1.551

13

523.2175

-1.825

7.54

583.2387

-1.653

14

523.2177

-1.481

8.7

583.2383

-2.288

15

523.2175

-1.94

9.29

583.2386

-1.756

16

523.2172

-2.418

9.59

583.2391

-0.933

17

ZEL-diGlc-Ac

C32H46O16

685.2702

685.2705

-1.253

7.21

745.2913

745.2909

-2.077

18

685.2716

0.353

7.38

745.2927

0.298

19

685.2706

-0.99

8.97

745.2919

-0.748

20

685.2699

-2.055

9.29

745.2925

0.057

aMass value was calculated based on elemental formula

bMass value detected by full scan event

N.D. = not detected

For completeness, an online database containing 3D structures, chemical formula, molecular weight, SMILE code, and LogP values of the putative ZEN biotransformation products was generated. The database is freely available for download at https://sites.google.com/site/zearalenonesdf/home, with new molecules being added monthly.

Detection and identification of acetylated forms of ZEN

Acetylation of zearalenone

Within the full scan mass spectrum, a deprotonated ion at m/z 359.1502 was detected. It fitted with the sum formula C20H24O6 (0.469 ppm mass deviation), consistent with the hypothesis of the acetylation of zearalenone (Ac-ZEN). It must be underlined that ZEN backbone carries two hydroxyl groups at positions C-14 and C-16, and therefore two acetylated isomers were expected. However, extracted ion chromatogram (EIC) gave rise to at least five signals (see Fig. 2).

Considering the gradient elution method used, peaks at low retention time were likely due to in source fragmentation of acetylated forms of more polar ZEN conjugates. Owing to the retention time, it was postulated that the small EIC-peak (m/z 359.1502) at 15.08 min (peak #1) could presumably be attributed to Ac-ZEN. Unfortunately, the low signal intensity did not allow the collection of useful product ion spectra. To support our annotation, we compared calculated LogP values of Ac-ZEN (4.19) and ZEN (2.91). The increased hydrophobicity of Ac-ZEN compared with the parent form may be indeed consistent with the 2.5 min shift between the two compounds.

Based on the retrieved information, it cannot be defined whether the acetylation occurs at position C-14 or C16. However, taking into consideration that position C-14 is less sterically hindered than position C-16, 14-Ac-ZEN may be more likely formed than 16-Ac-ZEN.

Acetylation of ZEN glucosides

The detection of several peaks in the EIC of Ac-ZEN (Fig. 2) indicated the possible in source-fragmentation of polar acetylated modified forms of ZEN. A further conjugation of acetylated ZEN with glucose may indeed occur, together with the acetylation of the glucosidic moiety of ZEN-Glc. The acetylation of ZEN-Glc was actually proven by the presence, in the full scan mass spectrum, of the mass m/z 521.2021, which corresponds to the [M-H]- ion (Fig. 3). In addition, also the acetate adduct was detected and showed perfect coelution with deprotonated ions (see Fig. 3), confirming the putative formula C26H34O11 with a maximum mass deviation <2 ppm. The earlier elution (4.6–9.7 min) of such forms from the C18 column compared with ZEN (12.14 min) and Ac-ZEN (15.09 min), were consistent with their putative identification.

To verify whether the acetylation occurred at the sugar moiety or at the phenolic ring of the parent compound, the fragmentation pattern was further considered. In particular, product ion spectra of [M-H]- and [M+CH3COO]- adducts for peaks #5, #6, and #7 clearly reported the loss of the glucosyl moiety, with the formation of the quasi-molecular ion of ZEN at m/z 317.1390 (see Fig. 3B). This is consistent with the acetylation of the sugar moiety instead of the parent compound. Therefore, compounds #5, #6, and #7 were putatively identified as ZEN-Ac-Glc, with the acetyl group located at the hexoside moiety. The observation of different isomeric peaks may result from the conjugation of acetyl moiety to different hydroxyl groups of glucose. At the same time, glucose may be conjugated to the position C-14 or C-16 of the ZEN backbone.

Similar to what was observed for ZEN-acetyl-monoglucosides, LC-HRMS analysis allowed us to annotate the presence of acetylated diglucosidic forms. [M-H]- and [M + CH3COO]- were consistent with the putative formula C32H44O16 with a mass deviation <2 ppm for both adducts, corresponding to the addition of glucose moiety (C6H10O5). Interestingly, two abundant signals (#10, #11) were detected with m/z 683.2549 at retention times 9.03 and 9.45 min (see Fig. 3), with a shift of 30 s from the putatively identified ZEN-Ac-Glc isomers (detected at 9.30 and 9.75 min). It is assumed that ZEN-Ac-diGlc was obtained from ZEN-Ac-Glc by the conjugation of a second glucose molecule via 1,4- or 1,6-glycosidic linkage, thus resulting in structural isomers. Unfortunately, HRMS analysis was not useful in discriminating between the two possible isomers since only the cleavage of the glycosidic linkage was observed, giving rise to the mass signal m/z 317.1390 (see ESM Fig. S1).

Regarding the existence of poly-glucosides, several signals were detected when searching for ZEN-Ac-triGlc forms, with a calculated mass of [ZEN-triGlc-Ac–H]- m/z 845.30591 (see ESM Fig. S2). However, there was insufficient evidence to support the existence of such hexosides under the applied conditions.

Glucosylation of Acetyl-ZEN

Besides the annotation of ZEN-Ac-Glc signals (peaks #5, #6, #7), also peak #4 (rt 5.62 min) was consistent with m/z 521.2021 and m/z 683.2549 corresponding to [Ac-ZEN-Glc – H]- and [Ac-ZEN-Glc + CH3COO]-), respectively. However, its fragmentation pattern of peak #4 reported the signal at m/z 359.1500, corresponding to the loss of a hexoside moiety, and not the signal at m/z 317.1390, corresponding to the loss of an acetyl-hexoside moiety. According to such fragmentation, the aglycone was annotated as Ac-ZEN (C20H23O6) and not ZEN. Similarly to peak #1, it was not possible to further discriminate between acetylation at position C-14 or C-16.

The annotation of peaks #2 and #3 was unfortunately not possible because of the low intensity and the consequent loss of information in the product ion spectra.

Dihexose conjugate signals (peaks #8, #9) were detected at retention time close to the putative identified monohexoside (#peak 4). However, the structure elucidation was not possible due to the low intensity of the signals, since product ion spectra could not be acquired.

Detection and identification of acetylated forms of ZELs

As for ZEN, the acetylated forms of phase I metabolites α- and β-ZEL were putatively identified for the first time in wheat.

First, it must be underlined that Ac-ZEL signals with a calculated ion mass of m/z 361.1645 (C20H26O6) were detected only at the same retention time of larger metabolites that originated it for in-source fragmentation. The fact that no acetyl-ZEL was detected may suggest that acetylation takes place after the conjugation of glucose to the ZEL backbone, consistent with the hypothesis reported by Meng-Reiterer et al. [6] T2 and HT2 toxins. It may suggest as well that rather apolar acetyl-ZELs are rapidly and quantitatively conjugated to more polar compounds in plants to minimize their toxic behavior and quickly facilitate their segregation in the apoplast or in the vacuole.

Calculated mass of m/z 523.2174 and m/z 685.2702, consistent with the putative formula C26H36O11, were detected at the same retention time, as reported in Fig. 4 and Table 1.

The quasimolecular ion of m/z 523.2174 gave rise to fragment ion of 319.1550, corresponding to the ZEL aglycone resulting from the loss of the acetylated hexoside moiety, cleaved from the molecules after low energy CID. No fragment ion at m/z 361.1645 (C20H26O6) was detected, suggesting that acetyl group was not directly linked to the ZEL backbone.

Based on the data described above, compounds detected at 7.3– 9.6 min (peaks #12, #13, #14, #15, #16) were annotated as ZEL-Ac-Glc, without any further elucidation of isomeric forms.

When seeking for ZEL-Ac-diGlc forms, four signals (rt 7.2–9.2 min) with one major peak at rt 8.97 min were detected, with a calculated mass of 685.2702 [M-H]- and 745.2913 [M + CH3COO]- (Fig. 4). For the most intense peak, also HRMS/MS spectrum was acquired, confirming the presence of the fragment ion of m/z 319.1544 that corresponds to ZEL backbone (see ESM Fig. S1).

Assessing the toxic potential of acetylated forms of ZEN and ZELs

Acetylation is a phase I reaction that may lead to rather apolar metabolites, which can be better absorbed into the gastrointestinal tract of mammals compared with other modified forms. Therefore, although occurring at very low concentration levels, the potential toxicity of such metabolites has to be carefully evaluated. In particular, taking into consideration the relative potency factor of 60 attributed to α-ZEL and computed in the TDI proposed by EFSA, its acetylated forms cannot be disregarded.

Considering the lack of commercial availability and the high costs of synthesis or purification, toxicity of the metabolites presented herein was assessed through an in silico approach. To assess the fit-for-purpose reliability over the presented case study, the procedure has been challenged, checking the capability to properly predict the binding architectures and the interactions between ERα and the well-known estrogenic compounds 17-β-estradiol, ZEN and αZAL. Detailed information is made available in the ESM.

As reported in Table S2 (see ESM), all the metabolites considered herein showed scores lower than ZEN, pointing to a worse fitting into the ER pocket. This indicated a possible reduced capability to trigger an estrogenic stimulus per se. However, it should be noted that nothing is known so far about the possible metabolic transformation – and thus subsequent bioactivation/deactivation – of acetylated forms following ingestion. From a mechanistic point of view, the presence of glycosides and/or acetyl groups was abolished in all the combinations tested for the capability to interact with Glu353 (Fig. 5A, B, and C). Notably, the interaction with Glu353, together with that with Arg394, is crucial for keeping estrogenicity [32]. A similar mechanism was proposed previously by Ehrlich and co-workers to explain the lack of activity of hypothemycin as the phenol methylation disrupted the interaction with Glu353 and Arg394 [23]. The possible reduced activity of the acetylated metabolites is in agreement with the weak estrogenic activity previously found for 14Ac-ZEN by Shier and coworkers [33]. Nevertheless, they were positive even if low scores were recorded for 16Ac-αZEL and 16Ac-βZEL. They showed binding geometries in agreement with the crystallographic data of ZEN and αZAL. In particular, the acetylation in position 16 and not in position 14 was found complying with a proper pocket fitting allowing the contact of Glu353 (Fig. 5D). Therefore, for 16Ac-αZEL and 16Ac-βZEL, a degree of estrogenicity higher than the other metabolites considered herein can be expected. Concerning the glucosilation, in agreement with the data recently shown for ZEN glucosides and glucuronides [22], the presence of glycoside groups was found not complying with the pocket as they exceeded the volume available for ligands, as well as they caused hydrophobic/hydrophilic interferences with the lipophilic environment of the binding site (Fig. 5C).
Fig. 5

The crystallographic pose of ZEN (A) [31] is shown in comparison with the computed poses of 14Ac-αZEL (B), ZEN14Ac-diGlc (C), and 16Ac-αZEL (D). The protein is represented in cartoon, the binding site contour is represented in mesh, ligands and the protein residues involved in polar contacts are represented in sticks, whereas polar contacts are indicated by yellow dashed lines. Black boxes indicate the unfavorable placement of acetyl (B) or acetyl-glycoside groups (C). ZEN14Ac-diGlc (C) is shown in stick and surface for a better understanding of the volume excess and the acyl-glycoside moiety is colored in blue

Notably, not all the acetylated aglycones (i.e., 14Ac- and 16Ac-αZEL, and 14Ac- and 16Ac-βZEL) were found among the occurring metabolites. However, they were considered worthy of toxicological assessment as the release of aglycones from ZEN glycosides may happen, as previously reported [14, 17, 25]. In addition, deacetylation reactions might occur in humans as suggested by a recent study [34] that reported strong deacetylation of 3/15ADON by in vitro and ex vivo approaches.

On this basis, further data concerning the stability of the acetylated metabolites shall be collected with priority to define their toxicological relevance, along with assessing further their xenoestrogenic potential.

Conclusion and outlook

The acetylation metabolism route was investigated in wheat leaves using an in vitro plant model that excludes the contribution of Fusarium spp. Indeed, the presence of acetylated forms of ZEN and ZEL, attached to the aglycone backbone as well as on the glucose moiety, was annotated in wheat samples. Thanks to the advantages provided by HRMS, identification and structure annotation of 20 metabolites was achieved. With regard to the toxicological assessment in silico, all the metabolites considered herein showed a worse fitting within the ER pocket in comparison to ZEN. Accordingly, a reduced capability to trigger an estrogenic stimulus per se can be expected. Nevertheless, the possible hydrolysis to the respective parent compounds (i.e., ZEN and ZEL isomers) may raise concern from the health perspective, as these are well known xenoestrogens.

All the metabolites studied herein should be considered as potential hazards mainly because of the possible hydrolysis rather than on account of the intrinsic estrogenic potential. As an exception, 16Ac-αZEL and 16Ac-βZEL may retain a degree of xenoestrogenicity per se. Therefore, in order to better understand the toxicological concern of the set of ZEN metabolites considered herein, their bioaccessibility, bioavailability, as well as the chemical stability under gastrointestinal conditions should be further assessed with urgency, aside from collecting further data on their effective co-occurrence and abundance in grains. Overall, data presented in this study strongly indicate the need of further analytical efforts to monitor the full profile of ZEN modified forms occurring in cereals under open field conditions. For this purpose, a wide range of plant metabolites of mycotoxins should be made available on the market as analytical standards or as reference materials.

Notes

Acknowledgements

The authors acknowledge with gratitude Mr. Dante Catellani from Advanced Laboratory Research (Barilla G.R. F.lli SpA) for his technical assistance.

Compliance with ethical standards

Conflict of interest

Michele Suman is employee of Barilla G.R. F.lli SpA. Daniele Cavanna has received a PhD grant by Barilla G.R. F.lli SpA. All the other authors declare that they have no conflict of interest.

Supplementary material

216_2018_1066_MOESM1_ESM.pdf (338 kb)
ESM 1 (PDF 338 kb)

References

  1. 1.
    Berthiller F, Crews C, Dall'Asta C, Saeger SD, Haesaert G, Karlovsky P, et al. Masked mycotoxins: a review. Mol Nutr Food Res. 2013;57:165–86.CrossRefPubMedGoogle Scholar
  2. 2.
    EFSA Panel on Contaminants in the Food Chain (CONTAM). Risks for animal health related to the presence of zearalenone and its modified forms in feed, p. 123. EFSA J. 2017;15(7):4851.Google Scholar
  3. 3.
    Berthiller F, Werner U, Sulyok M, Krska R, Hauser MT, Schuhmacher R. Liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) determination of phase II metabolites of the mycotoxin zearalenone in the model plant Arabidopsis thaliana. Food Addit Contam. 2006;23:1194–200.CrossRefPubMedPubMedCentralGoogle Scholar
  4. 4.
    Kluger B, Bueschl C, Lemmens M, Michlmayr H, Malachova A, Koutnik A, et al. Biotransformation of the mycotoxin deoxynivalenol in fusarium resistant and susceptible near isogenic wheat lines. Plos One. 2015;10(3):e0119656.CrossRefPubMedPubMedCentralGoogle Scholar
  5. 5.
    Meng-Reiterer J, Bueschl C, Rechthaler J, Berthiller F, Lemmens M, Schuhmacher R. Metabolism of HT-2 Toxin and T-2 toxin in oats. Toxins. 2016;8:364.CrossRefPubMedCentralGoogle Scholar
  6. 6.
    Meng-Reiterer J, Varga E, Nathanail AV, Bueschl C, Rechthaler J, McCormick SP, et al. Tracing the metabolism of HT-2 toxin and T-2 toxin in barley by isotope-assisted untargeted screening and quantitative LC-HRMS analysis. Anal Bioanal Chem. 2015;407:8019–33.CrossRefPubMedPubMedCentralGoogle Scholar
  7. 7.
    Kovalsky Paris MP, Schweiger W, Hametner C, StuÈckler R, Muehlbauer GJ, Varga E, et al. Zearalenone-16-O-glucoside: a new masked mycotoxin. J Agric Food Chem. 2014;62:1188–9.CrossRefGoogle Scholar
  8. 8.
    Righetti L, Rolli E, Galaverna G, Suman M, Bruni R, Dall’Asta C. Plant organ cultures as masked mycotoxin biofactories: deciphering the fate of zearalenone in micropropagated durum wheat roots and leaves. Plos One. 2017;12(11):e0187247.CrossRefPubMedPubMedCentralGoogle Scholar
  9. 9.
    EFSA Panel on Contaminants in the Food Chain (CONTAM). Appropriateness to set a group health-based guidance value for zearalenone and its modified forms. EFSA J. 2016;14(4):4425.Google Scholar
  10. 10.
    Kowalska K, Habrowska-Górczyńska DE, Piastowska-Ciesielska AW. Zearalenone as an endocrine disruptor in humans. Environ. Toxicol. Pharmacol. 2016;48:141–9.CrossRefPubMedGoogle Scholar
  11. 11.
    De Boevre M, Jacxsens L, Lachat C, Eeckhout M, Di Mavungu JD, Audenaert K, et al. Human exposure to mycotoxins and their masked forms through cereal-based foods in Belgium. Toxicol Lett. 2013;218:281–92.CrossRefPubMedGoogle Scholar
  12. 12.
    Nathanail AV, Syvahuoko J, Malachova A, Jestoi M, Varga E, Michlmayr H, et al. Simultaneous determination of major type A and B trichothecenes, zearalenone and certain modified metabolites in Finnish cereal grains with a novel liquid chromatography-tandem mass spectrometric method. Anal Bioanal Chem. 2015;407:4745–55.CrossRefPubMedPubMedCentralGoogle Scholar
  13. 13.
    De Boevre M, Vanheule A, Audenaert K, Bekaert B, Diana Di Mavungu J, Werbrouk S, et al. Detached leaf in vitro model for masked mycotoxin biosynthesis and subsequent analysis of unknown conjugates. World Mycotox J. 2014;7:305–12.CrossRefGoogle Scholar
  14. 14.
    Schmeitzl C, Warth B, Fruhmann P, Michlmayr H, Malachová A, Berthiller F, et al. The metabolic fate of deoxynivalenol and its acetylated derivatives in a wheat suspension culture: identification and detection of DON-15-O-glucoside, 15-acetyl-DON-3-O-glucoside, and 15-acetyl-DON-3-sulfate. Toxins. 2015;7:3112–26.CrossRefPubMedPubMedCentralGoogle Scholar
  15. 15.
    Muñoz L, Castro JL, Cardelle M, Castedo L, Riguera R. Acetylated mycotoxins from Fusarium graminearum. Phytochemistry. 1989;28(1):83–5.CrossRefGoogle Scholar
  16. 16.
    Pinton P, Tsybulskyy D, Lucioli J, Laffitte J, Callu P, Lyazhri F, et al. Toxicity of deoxynivalenol and its acetylated derivatives on the intestine: differential effects on morphology, barrier function, tight junction proteins, and mitogen-activated protein kinases. Toxicol Sci. 2012;130:180–90.CrossRefPubMedGoogle Scholar
  17. 17.
    Dall'Erta A, Cirlini M, Dall'Asta M, Del Rio D, Galaverna G, Dall'Asta C. Masked mycotoxins are efficiently hydrolyzed by human colonic microbiota releasing their aglycones. Chem Res Toxicol. 2013;26(3):305–12.CrossRefPubMedGoogle Scholar
  18. 18.
    Gratz SW, Dinesh R, Yoshinari T, Holtrop G, Richardson AJ, Duncan G, et al. Masked trichothecene and zearalenone mycotoxins withstand digestion and absorption in the upper GI tract but are efficiently hydrolyzed by human gut microbiota in vitro. Mol Nutr Food Res. 2017;61(4):1600680–90.CrossRefGoogle Scholar
  19. 19.
    Gratz SW. Do plant-bound masked mycotoxins contribute to toxicity? Toxins. 2017;9:85.CrossRefPubMedCentralGoogle Scholar
  20. 20.
    Eriksen GS, Pettersson H, Lindberg JE. Absorption, metabolism, and excretion of 3-acetyl DON in pigs. Arch Anim Nutr. 2003;57:335–45.CrossRefGoogle Scholar
  21. 21.
    Dellafiora L, Ruotolo R, Perotti A, Cirlini M, Galaverna G, Cozzini P, et al. Molecular insights on xenoestrogenic potential of zearalenone-14-glucoside through a mixed in vitro/in silico approach. Food Chem Toxicol. 2017;108:257–66.CrossRefPubMedGoogle Scholar
  22. 22.
    Dellafiora L, Galaverna G, Dall'Asta C. An in silico perspective on the toxicodynamic of tetrodotoxin and analogues – a tool for supporting the hazard identification. Toxicon. 2017;138:107–18.CrossRefPubMedGoogle Scholar
  23. 23.
    Ehrlich VA, Dellafiora L, Mollergues J, Dall'Asta C, Serrant P, Marin-Kuan M, et al. Hazard assessment through hybrid in vitro/in silico approach: the case of zearalenone. ALTEX. 2015;32(4):275–86.PubMedGoogle Scholar
  24. 24.
    Phillips C, Roberts LR, Schade M, Bazin R, Bent A, Davies NL, et al. Design and structure of stapled peptides binding to estrogen receptors. J Am Chem Soc. 2011;133(25):9696–9.CrossRefPubMedGoogle Scholar
  25. 25.
    Dellafiora L, Galaverna G, Righi F, Cozzini P, Dall'Asta C. Assessing the hydrolytic fate of the masked mycotoxin zearalenone-14-glucoside – a warning light for the need to look at the “maskedome”. Food Chem Toxicol. 2017;99:9–16.CrossRefPubMedGoogle Scholar
  26. 26.
    Baroni M, Cruciani G, Sciabola S, Perruccio F, Mason JS. A common reference framework for analyzing/comparing proteins and ligands. Fingerprints for Ligands and Proteins (FLAP): theory and application. J Chem Inf Model. 2007;47(2):279–94.CrossRefPubMedGoogle Scholar
  27. 27.
    Carosati E, Sciabola S, Cruciani G. Hydrogen bonding interactions of covalently bonded fluorine atoms: from crystallographic data to a new angular function in the GRID force field. J Med Chem. 2004;47(21):5114–25.CrossRefPubMedGoogle Scholar
  28. 28.
    Kellogg EG, Abraham DJ. Hydrophobicity: is LogP(o/w) more than the sum of its parts? Eur J Med Chem. 2000;37(7/8):651–61.CrossRefGoogle Scholar
  29. 29.
    Cozzini P, Dellafiora L. In silico approach to evaluate molecular interaction between mycotoxins and the estrogen receptors ligand binding domain: a case study on zearalenone and its metabolites. Toxicol Lett. 2012;214(1):81–5.CrossRefPubMedGoogle Scholar
  30. 30.
    Cozzini P, Fornabaio M, Marabotti A, Abraham DJ, Kellogg GE, Mozzarelli A. Simple, intuitive calculations of free energy of binding for protein-ligand complexes. 1. Models without explicit constrained water. J Med Chem. 2002;45(12):2469–83.CrossRefPubMedGoogle Scholar
  31. 31.
    Ruttkies C, Schymanski EL, Wolf S, Hollender J, Neumann S. MetFrag relaunched: incorporating strategies beyond in silico fragmentation. J Cheminform. 2016;8:3.CrossRefPubMedPubMedCentralGoogle Scholar
  32. 32.
    Katzenellenbogen JA. The 2010 Philip S. Portoghese Medicinal Chemistry Lectureship: addressing the “core issue” in the design of estrogen receptor ligands. J Med Chem. 2011;54(15):5271–82.CrossRefPubMedPubMedCentralGoogle Scholar
  33. 33.
    Shier WT, Shier AC, Xie W, Mirocha CJ. Structure–activity relationships for human estrogenic activity in zearalenone mycotoxins. Toxicon. 2001;39(9):1435–8.CrossRefPubMedGoogle Scholar
  34. 34.
    Ajandouz EH, Berdah S, Moutardier V, Bege T, Birnbaum DJ, Perrier J, et al. Hydrolytic fate of 3/15-acetyldeoxynivalenol in humans: specific deacetylation by the small intestine and liver revealed using in vitro and ex vivo approaches. Toxins. 2016;8:232.CrossRefPubMedCentralGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany, part of Springer Nature 2018

Authors and Affiliations

  • Laura Righetti
    • 1
  • Luca Dellafiora
    • 1
  • Daniele Cavanna
    • 1
    • 2
  • Enrico Rolli
    • 3
  • Gianni Galaverna
    • 1
  • Renato Bruni
    • 1
  • Michele Suman
    • 2
  • Chiara Dall’Asta
    • 1
  1. 1.Department of Food and DrugUniversity of ParmaParmaItaly
  2. 2.Barilla G.R. F.lli SpA, Advanced Laboratory ResearchParmaItaly
  3. 3.Deparment of Chemistry, Life Sciences and Environmental SustainabilityUniversity of ParmaParmaItaly

Personalised recommendations