Article pubs.acs.org/JAFC
Cite This: J. Agric. Food Chem. 2018, 66, 1264−1269
Detection of the Cytotoxic Penitrems A−F in Cheese from the European Single Market by HPLC-MS/MS Svetlana A. Kalinina,†,‡ Annika Jagels,† Sebastian Hickert,†,‡ Lucas M. Mauriz Marques,§ Benedikt Cramer,† and Hans-Ulrich Humpf*,†,‡ †
Institute of Food Chemistry, Westfälische Wilhelms-Universität Münster, Corrensstraße 45, 48149 Münster, Germany NRW Graduate School of Chemistry, Wilhelm-Klemm-Str. 10, 48149 Münster, Germany § Department of Physics and Chemistry, Faculty of Pharmaceutical Sciences of Ribeirão Preto, University of São Paulo, Ribeirão Preto, São Paulo CEP 14049-900, Brazil ‡
S Supporting Information *
ABSTRACT: Penitrems are fungal indole diterpene-derived tremorgenic secondary metabolites, which are mainly produced by Penicillium spp. Several cases of intoxications with penitrems and subsequent occurrences of penitrem A in foodstuff underline the need for reliable quantitation methods for the detection of these mycotoxins in food. In this study, a simple and fast highperformance liquid chromatography−tandem mass spectrometry (HPLC-MS/MS) method for the quantitative analysis of penitrems A−F in cheese was developed. Therefore, penitrems A−F were isolated from Penicillium crustosum as analytical reference standards. The analysis of 60 cheese samples from the European single market (EU) revealed the occurrence of penitrem A in 10% of the analyzed samples with an average concentration of 28.4 μg/kg and a maximum concentration of 429 μg/kg. In addition to penitrem A, other members of the group of penitrems, namely, penitrems B, C, D, E, and F, were for the first time quantitatively detected in food samples, although in lower concentrations and with lower incidence in comparison to penitrem A. Moreover, we report cytotoxic effects of all penitrems on two cell lines (HepG2 and CCF-STTG1). This clearly underlines their relevance and the importance to analyze food samples in order to get insights into the human exposure toward these mycotoxins. KEYWORDS: penitrems, mass spectrometry, liquid chromatography, cheese
■
INTRODUCTION Penitrems are toxic secondary metabolites produced mainly by Penicillium species.1 Penitrem A, 1 (Figure 1), being a paxillinelike, 8 (Figure 1) indole diterpene, was first isolated by Wilson et al.2 from Penicillium cyclopium. Its absolute configuration and the structures of the other members of the group of penitrems B, C, D, E, and F, 2−6 (Figure 1), were described later in 1983.3 In 2003, a new member of the penitrems was isolated by González et al.4 and named penitrem G, 7 (Figure 1). Intoxications with penitrems have been well-documented in animals and humans. Clinical signs of poisoning comprise severe generalized muscle tremors accompanied by incoordination, seizures, muscle fasciculations, and generalized convulsions, eventually resulting in death.5−8 In 2005, the ingestion of food contaminated with 1 was shown to cause tremor syndromes in humans.9 A comprehensive toxicity profile was also well-described, and the tremorgenic dose of 1 after oral administration (PO) to mice was found to be 0.5 mg/kg of body weight (bw). The LD50 in mice was reported to be 10 mg/kg of bw (PO), whereas the intraperitoneal (IP) LD50 is 1.1 mg/kg.10 Chicken showed neurological syndromes after administration of 5 mg/kg of bw (PO) followed by mortality at higher concentrations.11 In lower doses, 1 exhibits high insecticidal activity against the insects Spodoptera f rugipera and Heliothis zea.12 Insecticidal activities were also reported for penitrems A−D and F.4 Notably, the chlorinated penitrems A, C, and F exhibit the highest acute toxicities. The halogen atom © 2018 American Chemical Society
was proposed to be important for the lethality of treated insects, whereas the epoxy group seems to be responsible for their delayed mortality. In general, relative toxicity ranged in the order 1 > 6 > 3 > 2 > 4.4 Analytical approaches to detect penitrems in food, feed, and physiological samples were focused on 1 in the past. For instance, Braselton and Rumler13 analyzed the stomach contents of dogs, which exhibited neurological disease by gas chromatography coupled with a tandem mass spectrometer (GC-MS/MS), revealing the presence of 1. Penicillium toxins, including 1, were further analyzed in food and feed by using liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS). The limit of detection (LOD) for 1 was 5 μg/ kg.14 In 2005, Tor et al.15 elaborated a rapid LC-MS/MS method for the detection of 1 in serum and urine samples. Furthermore, an LC-MS/MS based multimycotoxin method including 1 with an LOD of 5 μg/kg was reported by Sulyok et al.16 For the quantitative determination of penitrems B−F, only semiquantitative methods have been applied so far.17 Analytical studies have been limited, mostly due to the lack of reference standards. Up to date, there is only one publication regarding Received: Revised: Accepted: Published: 1264
December 21, 2017 January 15, 2018 January 16, 2018 January 16, 2018 DOI: 10.1021/acs.jafc.7b06001 J. Agric. Food Chem. 2018, 66, 1264−1269
Article
Journal of Agricultural and Food Chemistry
Figure 1. Structures of penitrems A−G, 1−7, and paxilline, 8. several European countries (Germany, United Kingdom, The Netherlands, and Belgium). After homogenization a part of each sample was transferred to 50 mL polypropylene tubes and stored at −22 °C until sample preparation. Sample Preparation. Five g (±0.05 g) of cheese was placed into a 50 mL plastic tube, and the exact sample weights were determined. Ten mL of the extraction solvent (MeCN/H2O (80:20, v/v)) was added, and the mixture was vortexed for 30 s. Extraction was carried out on a laboratory shaker for 1 h at 250 rpm. Subsequently, the samples were centrifuged (3.000 × g, 5 min), and ∼1 mL of the supernatant was passed through a 16 mm, 0.45 μm membrane filter (Phenomenex, Aschaffenburg, Germany). The filtered extracts were stored at −22 °C until HPLC-MS/MS analysis. All samples were analyzed in triplicate. HPLC-MS/MS Parameters. A QTrap 5500 mass spectrometer (SCIEX, Darmstadt, Germany) equipped with an LC system (Shimadzu, Kyoto, Japan) was employed for the analysis using electrospray ionization in the positive mode. The source temperature was 500 °C, curtain gas was set to 35 psi, and the ion source gases 1 and 2 were set to 35 and 45 psi, respectively. The ion spray voltage was 4500 V, and the collision gas was set to “high”. A dwell time of 40 ms was used for all multiple reaction monitoring (MRM) transitions. The separation of penitrems was performed on a column 150 mm × 2.0 mm i.d., 3 μm, Reprosil Gold RP-18, with a 5 mm × 2 mm i.d. guard column of the same material (Dr. Maisch GmbH, Ammerbuch, Germany) using a binary gradient consisting of MeCN (solvent A) and H2O (solvent B), both containing 0.1% formic acid. The column oven was operated at 40 °C, and the flow rate was 350 μL/min. Twenty μL of the sample solution was injected. Starting conditions of the gradient were as follows: 0−6 min, 40−100% A; 6−10 min, held at 100% A; followed by reequilibrating the column under starting conditions (40% of A) for 4 min. Method Performance. The limits of detection (LODs) and the limits of quantitation (LOQs) were based on a signal-to-noise ratio (S/N) equal to 3 for LOD and equal to 10 for LOQ. Therefore, cheese matrix obtained from a noncontaminated sample was spiked with penitrems down to concentrations with the respective S/N ratios. Recovery rates for each analyte were examined by spiking 5 g of blank cheese at three different levels within the calibration range (50, 100, and 150 μg/kg to 100, 200, and 300 ng/g) in triplicate, followed by the sample preparation described earlier. Quantitative Determination of Penitrems in Cheese Samples. To analyze the obtained results, Analyst software version 1.6.2 (SCIEX, Darmstadt, Germany) was employed. Calibration curves were generated by plotting the peak area against the concentration for each penitrem. Each calibration solution was measured in duplicate, and the peak areas were averaged. The slope and intercept of these curves were used to calculate the concentrations of each penitrem. The sample weight was taken into account, and the results were not corrected by recovery rates. Results are based on three independent measurements and given as mean ± standard deviation. Cytotoxicity. For cytotoxicity assays, 3.0 mg of 1, 2.0 mg of 5, and 1.0 mg of penitrems B, C, D, and F were dissolved in MeOH. These stock solutions were diluted to the respective concentrations with serum-free culture medium. Human liver cancer cells (HepG2, ACC 180) (DSMZ, Braunschweig, Germany) were cultivated in Dulbecco’s
the presence of penitrems A−F in food waste of private households in a concentration range of 30−7500 μg/kg.17 In addition, penitrems A−D were detected in organ tissues of poisoned dogs in a semiquantitative matter.18 A quantitative method for the simultaneous determination of all penitrems A− F has not been described so far. In our previous study, we reported the optimal incubation parameters for the production of penitrems A−F in high concentrations applying various abiotic factors and stress factors.19 In this study, all penitrems A−F were isolated in high yield and high purity. It was further shown that penitrems were also produced by P. crustosum on a cheese medium in high concentrations, and cheese might thus be a food potentially susceptible to contamination with penitrems. Consequently, the objectives of the presented study were the development of a robust HPLC-MS/MS method for the quantitative determination of penitrems A−F and its application to the analysis of cheese samples purchased from the EU market. In addition, cytotoxic effects on two different cell lines were investigated to get further insights into the toxicity of penitrems A−F.
■
MATERIALS AND METHODS
Chemicals and Reagents. If not otherwise mentioned, all solvents were of gradient grade and purchased from VWR (Darmstadt, Germany). HPLC-MS-grade acetonitrile was purchased from Fisher Scientific (Schwerte, Germany). ASTM type 1 water was obtained with a Purelab Flex 2 system from Veolia Water Technologies (Celle, Germany). Formic acid was from Merck (Darmstadt, Germany). Penitrems A−F (purity ≥96%) were isolated and spectroscopically characterized as reported in our previous study.19 Penitrems A−F were each dissolved in MeCN/H2O (80:20, v/v); the exact concentration was determined using molar absorptivity values obtained on a V-750 series UV spectrometer (Jasco, GrossUmstadt, Germany). Obtained stock solutions were used for the calibration. Calibration. Due to the lack of isotopically labeled standards for penitrems, matrix-matched calibration was used for quantitation. The matrix-matched blank solution was obtained by combining three different blank extracts of the cheese samples. The extraction was carried out after mixing and homogenizing of the cheese samples. Solutions of the penitrems A−F were evaporated under a gentle nitrogen stream at 40 °C and subsequently redissolved in 1 mL of blank cheese matrix extract. Eleven calibration points ranging from 2 to 202 ng/mL of penitrems (A−F) were prepared in blank cheese matrix extract using a solvent mixture of MeCN/H2O (80:20, v/v). Solvent calibration was prepared in the same concentration range as in the matrix-matched calibration. The signal suppression/enhancement factor (SSE) was determined using the following equation: SSE (%) = 100 ×
Slopematrix‐matched calibration Slopesolvent calibration
Samples. Different types of cheese (firm, soft, hard, processed, and supplemented with nuts, fruits, and fungal culture) were purchased in 1265
DOI: 10.1021/acs.jafc.7b06001 J. Agric. Food Chem. 2018, 66, 1264−1269
Article
Journal of Agricultural and Food Chemistry
Figure 2. HPLC-MS/MS chromatogram of spiked cheese matrix with penitrems A−F (150 ng/mL) and the respective MRM transition. modified Eagle medium (DMEM) supplemented with 10 mM N-2hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES buffer), 100 μg/mL streptomycin, 100 U/mL penicillin, 2 mM L-glutamine, and 10% (v/v) fetal calf serum (FCS) using standardized culture conditions (37 °C, 5% CO2, saturated humidified atmosphere). Human astrocyte cells (CCF-STTG1) (ATCC, Manassas, VA) were cultivated in Roswell Park Memorial Institute medium (RPMI 1640) supplemented with 100 μg/mL streptomycin, 100 U/mL penicillin, 2 mM L-glutamine, and 10% (v/v) FCS. Twice a week both culture media were changed, and the cells were subcultivated after trypsination when a microscopic confluence was at the level of 80%. For the evaluation of cytotoxic effects of penitrems on both cell lines, Cell Counting Kit-8 (CCK-8) (Dojindo Laboratories, Tokyo, Japan) was used. The assay was performed according to the manufacturer’s instructions and our previous studies.20−23 Both cell lines were treated with penitrems A−F (1−50 μM) and incubated for 48 h. After toxin application, media were replaced by 100 μL/well of the 10-fold diluted water-soluble tetrazolium dye solution WST-8 (2-(2-methoxy-4nitrophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium monosodium salt) followed by incubation for 70 min at 37 °C. The viability of the cells was assessed by their ability to reduce WST-8 dye to a water-soluble formazan, which was analyzed by measuring the formazan absorbance at a wavelength of 450 nm with a Tecan Infinite 200 PRO microplate reader (Tecan, Salzburg, Austria). Absorption values were subtracted by the average absorption of six cell-free blank wells. The results for toxin-exposed cells were normalized to the values of a solvent control (1% MeOH). Tests were repeated in triplicate for each cell line from three independent passages (n ≥ 9). The data are presented as the mean ± standard deviation (SD). The IC50 values were calculated by log-linear regression, and significance indicated refers to the significance level as compared to the solvent-treated control (1% MeOH) calculated with the OriginPro 2016G (64-bit) Sr2 b9.3.2.303 (SF8T5-3089-7901139) (OriginLab Corporation,
Northampton, U.S.A.). Obtained data were evaluated by analysis of variance (ANOVA) and Student’s t test; * indicates statistically significant (p ≤ 0.01) and ** indicates statistically highly significant (p ≤ 0.001).
■
RESULTS AND DISCUSSION Method Development and Performance. A simple and fast HPLC-MS/MS method for the quantitative determination of penitrems A−F in cheese was developed. The penitrems were analyzed in the positive ion electrospray ionization mode, which had the best relative signal intensity. The separation of penitrems was achieved by HPLC using a C18 column with a binary gradient consisting of MeCN/H2O. A chromatogram displaying the recorded multiple reactions monitoring (MRM) transitions and the observed typical elution order of penitrems A−F is shown in Figure 2. The retention times were as follows: 5, 7.1 min; 1, 7.8 min; 4, 8.6 min; 2, 8.8 min; 3, 9.1 min; and 6, 9.3 min. The signal intensities of penitrems A and E are much higher than those for any other penitrems, due to enhanced ionization, which could be explained by the presence of an additional hydroxy group at C15 of the cyclobutyl unit. The MRM transitions used to quantitate the analytes included m/z 600 → 524 for 5, m/z 634 → 558 for 1, m/z 568 → 550 for 4, m/z 584→ 566 for 2, m/z 602 → 532 for 3, and m/z 618 → 530 for 6 (Figure 2). For quantitation, matrix-matched calibration was used, whereas signal enhancement was observed for all penitrems in cheese matrix. Signal enhancement was as follows: 1, 239%; 2, 228%; 3, 453%; 4, 377%; 5, 348%; and 6, 455%. Further method characteristics are given in Table 1. The limits of detection (LODs) and the limits of quantitation 1266
DOI: 10.1021/acs.jafc.7b06001 J. Agric. Food Chem. 2018, 66, 1264−1269
Article
Journal of Agricultural and Food Chemistry
the LOQ (Table 3 and Figure 3). In general, penitrems in dairy products can originate either from indirect contamination, because of the ingestion of contaminated feed by dairy animals; or from direct contamination, which appears to be due to accidental or intentional molds growing. The predominant flora found on cheese belongs to the genus Penicillium, which is known to be a producer of penitrems, demonstrating a high stability in cheese.24,25 In the context of the penitrem production, direct contamination and therefore accidental growth of molds is the major source. Four of six positive samples contained nuts with various consistencies. Samples N2 and N3 contained whole walnuts placed on the cheese surface. In a cheese supplemented with hazelnuts, a variety of penitrems A−F was observed, which is typical for P. crustosum. Interestingly, the hard cheese supplemented with pistachios was found to be positive for penitrems A and E. By relying on these data, it can be hypothesized that contaminated nuts could be a frequent source of mold. It is assumed that in these cases cheese played the role of medium or substrate for the cultivation of penitrems-producing molds. The results of cultivating P. crustosum on cheese medium in our previous study confirm these findings.19 Also, we suggest that different Penicillium species could be involved in penitrem production regarding the tested samples. In processed cheese with walnuts (sample N3), only penitrem A in a high concentration of 429 μg/kg was detected. Recently, it was reported that P. melanoconidium is able to produce only penitrem A up to mg/kg scale, but no other penitrems.26 However, positive samples were found not only among cheese supplemented with nuts. For instance, in a creamy blue cheese (N3), 1 (1.2 μg/kg) and 5 (
