Determination of Penicillium mycotoxins in foods and feeds using liquid chromatography–mass spectrometry
Introduction
Mycotoxins from Penicillium species can be a serious contamination problem in poorly stored foods and feeds. The toxins are naturally produced by a variety of Penicillium species, which can be formed rapidly during transportation and storage. Because of the toxicity of many Penicillium mycotoxins [1], it is desirable that a rapid and reliable analytical procedure is available to monitor mycotoxin levels in foodstuffs. Several methods for analysing Penicillium mycotoxins have been developed. Due to the high polarity of many of the mycotoxins, high-performance liquid chromatography (HPLC) has been the preferred separation technique. When combined with UV or fluorescence detection, many of the important, commonly encountered, Penicillium toxins can been determined [2], [3], [4], [5], [6], [7]. These methods, however, often include time consuming clean up steps and seldom more than one or two mycotoxins can be determined in the same analysis. Multi-toxin methods have also been developed [8], [9], [10], [11], [12], [13], [14], [15], but these methods often suffer from low sensitivity or specificity and are mainly applicable to simple matrices such as fungal isolates grown on agar or rice.
The combination of LC and mass spectrometry (MS) simplifies the development of analytical methods for polar mycotoxins. With electrospray (ESI) and atmospheric pressure chemical ionisation (APCI) interfaces, mainly MH+ ions are formed. Unequivocal identification of target mycotoxins can be achieved, using single quadrupole or ion trap LC–MS systems by monitoring the diagnostic ion fragmentation patterns obtained by collision induced dissociation (CID). Higher selectivity and sensitivity can, however, be achieved using LC–MS–MS, especially when analysing samples which exhibit complex interferences. LC–MS–MS methods can discriminate different analytes by combinations of chromatographic, parent ion (MS) and product ion(s) (MS–MS) data.
The concentrations of Penicillium mycotoxins in both food and feed can span from zero up to high μg/kg or mg/kg levels in cases of mouldy feed intoxication of animals. Levels depend on the species of fungi, growth conditions and storage time. The toxicity of Penicillium mycotoxins necessitates low detection limits. For example, in the European Union (EU), the legally acceptable concentration of ochratoxin A in grain for human consumption is set at 5 ng/g [16].
This paper reports the development of a rapid and highly selective LC–MS and LC–MS–MS methods for the simultaneous determination of roquefortine C, griseofulvin, mycophenolic acid, ochratoxin A, verruculogen, chaetoglobosin B, penitrem A, citrinin, rubratoxin B, cyclopiazonic acid, PR-toxin, patulin and penicillic acid. Detailed validation of the methodologies was performed using roquefortine C, griseofulvin, mycophenolic acid, verruculogen, chaetoglobosin B and penitrem A (Fig. 1).
Section snippets
Chemicals
Methanol, ethyl acetate, dichloromethane, acetonitrile, hexane, ammonium acetate, ammonium formate, acetic acid and formic acid were HPLC or analytical-reagent grade and obtained from Rathburn (Walkerburn, UK). Deuterated T-2 toxin, roquefortine C, griseofulvin, mycophenolic acid, citrinin, rubratoxin B, cyclopiazonic acid, PR-toxin, patulin and penicillic acid, ochratoxin A, verruculogen, chaetoglobosin B and penitrem A were purchased from Sigma (St. Louis, MO, USA).
Sample preparation
A blank food mixture (1 kg)
LC–MS analysis
An investigation of the positive and negative ion mass spectral characteristics of a series of Penicillium mycotoxins including, roquefortine C, griseofulvin, mycophenolic acid, ochratoxin A, verruculogen, chaetoglobosin B, penitrem A, citrinin, rubratoxin B, cyclopiazonic acid, PR-toxin, patulin and penicillic acid, lead to the development and subsequent optimisation and validation of LC–MS and LC–MS–MS methodologies for the determination of Penicillium mycotoxins in a mixed food matrix. The
Conclusions
The methodologies reported here enable the determination of mycophenolic acid, griseofulvin, roquefortine, chaetoglobosin B, verruculogen, penitrem A and other Penicillium mycotoxins in food and feed, down to the same level (5 ng/g) as EU food safety regulations define for ochratoxin A in grain [16]. Full scan MS detection limits for mycophenolic acid, griseofulvin, roquefortine C, chaetoglobosin B, verruculogen and penitrem A were 2–3.5 times higher (typically in the range 10–70 ng/g) than was
Acknowledgements
We thank the Norwegian Research Council for supporting this work and The University of Waikato for the granting study leave to A.L.W. The authors also wish to thank Arne Flaoyen, Silvio Uhlig and Tone Asp for their helpful suggestions during the preparation of this manuscript.
References (19)
- et al.
J. Chromatogr.
(1989) J. Chromatogr.
(1986)- et al.
J. Chromatogr.
(1991) - et al.
J. Chromatogr.
(1990) - et al.
J. Chromatogr.
(1987) - et al.
J. Chromatogr.
(1987) J. Chromatogr.
(1987)- et al.
J. Chromatogr.
(1978) - et al.
J. Chromatogr. A
(1999)
Cited by (91)
Tremorgenic Mycotoxins
2018, Veterinary Toxicology: Basic and Clinical Principles: Third EditionOccurrence of roquefortine C, mycophenolic acid and aflatoxin M1 mycotoxins in blue-veined cheeses
2015, Food ControlCitation Excerpt :The mean recovery rates ranged from 63–66, 85–100 and 100–120% for ROQC, MPA and AFM1, respectively. Considering the chemical complexity of blue-veined cheeses (high protein, fat and salt contents), the obtained recovery percentages can be considered to be acceptable in regards to those previously found for other complex food products (Rundberget & Wilkins, 2002), although recovery rates obtained for ROQC were below the guideline value (−20% for concentrations >10 μg/kg) reported in the European regulation 2002/657/EC. ROQC being more lipophilic (log P = 5.09) than MPA (log P = 2.92) and AFM1 (log P = −0.35), it is likely that a loss occurred during cheese fat removal using hexane.
Development and validation of a reversed phase liquid chromatographic method for analysis of griseofulvin and impurities
2013, Journal of Pharmaceutical and Biomedical AnalysisCitation Excerpt :Recently, an electrospray ionization LC–MS/MS method was validated and adopted for the determination of GF in human plasma [15]. Characterization and determination of GF in some other sample matrices such as moulds, marine strains, food and feeds and bioavailability studies have also been reported [16–22]. In the United States Pharmacopeia, both the drug substance and the dosage forms are assayed by LC using a stationary phase with nitrile groups bonded to porous silica particles [23].
Sampling techniques for the determination of mycotoxins in food matrices
2012, Comprehensive Sampling and Sample Preparation: Analytical Techniques for ScientistsNeurotoxicity of Penicillium crustosum secondary metabolites: Tremorgenic activity of orally administered penitrem A and thomitrem A and E in mice
2012, ToxiconCitation Excerpt :Supernatants were defatted with hexane (1:1, v/v) and 0.5 mL aliquots were centrifuged (0.2 μm nylon filters) at 9300 g through Spin-X microcentrifuge (Costar, Corning, NY, USA) and analyzed by LC–MS. The recovery of the extraction method was considered quantitative, since the method had been previously validated in a variety of matrices (Rundberget and Wilkins, 2002b). The samples analyzed were diluted, when necessary, to approx. 50 ng/g, where not significant matrix effects due to ion suppression have been previously reported (Rundberget and Wilkins, 2002b). Tissue density was set to approximately 1 for the data conversion to organ concentrations (Brown et al., 1997).