Mycotoxins are toxic secondary metabolites produced by different fungi, with different chemical structures. Mycotoxins contaminate food, feed, or raw materials used in their production and cause diseases and disorders in humans and livestock. Because of their great variety of toxic effects and their extreme heat resistance, the presence of mycotoxins in food and feed is considered a high risk to human and animal health. In order to ensure food quality and health consumers, European legislation has set maximum contents of some mycotoxins in different matrices. However, there are still some food commodities susceptible to fungal contamination, which were not contemplated in this legislation. In this context, we have developed new analytical techniques for the multiclass determination of mycotoxins in a great variety of food commodities (some of them scarcely studied), such as cereals, pseudocereals, cereal syrups, nuts, edible seeds, and botanicals. Considering the latest technical developments, ultrahigh performance liquid chromatography coupled to tandem mass spectrometry has been chosen as an efficient, fast, and selective powerful analytical technique. In addition, alternative sample treatments based on emerging methodologies, such as dispersive liquid-liquid microextraction and QuEChERS, have been developed, which allow an increased efficiency and sample throughput, as well as reducing contaminant waste.

1. Introduction

Mycotoxins are toxic natural secondary metabolites produced by several species of fungi (as Fusarium, Aspergillus, and Penicillium genera) on agricultural commodities. The presence of mycotoxins in food and feed may affect human and animal health, as they may cause many different adverse effects such as estrogenic, gastrointestinal, and kidney disorders, induction of cancer, and mutagenicity. Furthermore, some mycotoxins are also immunosuppressive and reduce resistance to infectious diseases [1, 2]. Mycotoxins grow under a wide range of climatic conditions and the Food and Agriculture Organization (FAO) has estimated that they affect 25% of the world crops. On the other hand, mycotoxins are the hazard category with the highest number of border rejections reported by the Rapid Alert System for Food and Feed (RASFF) [3]; therefore their impact on economy is evident.

Hundreds of mycotoxins have been recognized with diverse chemical structures, different toxicity, and biological effects. The most relevant groups of mycotoxins found in food are aflatoxins (aflatoxin B1 (AFB1, included in group 1 of carcinogenic to humans by the International Agency for Research on Cancer (IARC) [4]), aflatoxin B2 (AFB2), aflatoxin G1 (AFG1), aflatoxin G2 (AFG2), and aflatoxin M1 (AFM1, metabolite of AFB1, excreted in the milk of mammals)); ochratoxin A (OTA); trichothecenes (HT-2 and T-2 toxin and deoxynivalenol (DON)); zearalenone (ZEN); fumonisins B1 and B2 (FB1 and FB2); citrinin (CIT); patulin (PAT) and ergot alkaloids [2, 5]. Figure 1 shows the structures of some common mycotoxins and Table 1 includes the most important toxins, their main producing fungi, and typical food commodities that may be contaminated by them.

In order to protect consumer health, international institutions and organizations have proposed regulatory limits for some mycotoxins. Thus, the European Commission (EC) establishes maximum permitted levels for most mycotoxins in foods by means of the Commission Regulation (EC) number 1881/2006 [6] (or recommended levels for other mycotoxins [7]), as well as methods of sampling and analysis for their control by Commission Regulation (EC) number 401/2006 [8], which have been subsequently amended. Regulations are also established by the US Food and Drug Administration (FDA). FDA mycotoxin compliance programs provide introductory information about mycotoxins, products prone to contamination, and analytical methods [9].

In this context, the use of robust analytical methodologies for sampling, sample treatment, and identification/quantification of mycotoxins in food and feed is mandatory, in order to protect the consumer health [1012].

2. Current Analytical Methods for Determination of Mycotoxins in Food

Different analytical methods have been proposed for mycotoxin determination in food, such as thin layer chromatography (TLC) [13], ELISA [14], gas chromatography (GC) [15] or capillary electrophoresis (CE) [16]. However, the most popular technique is high performance liquid chromatography (HPLC) with UV/Vis, fluorescence (FL) [1719], or mass spectrometry (MS) detection [2022]. Recently, ultrahigh performance liquid chromatography (UHPLC) coupled with tandem mass spectrometry (MS/MS) has become very popular, especially for multiclass determination of mycotoxins and for multiresidue determination with other contaminants [2326].

Because of the complexity of food matrices, an extraction and clean-up purification step is usually required before analysis. Different approaches have been proposed. The most common methodology implies solid-liquid extraction (SLE) followed by solid phase extraction (SPE) with immunoaffinity columns (IACs), which contain specific antibodies to the analyte of interest [10]. Several reviews present an overview of the different methodologies proposed for the determination of mycotoxins in food including the most frequent sample treatments [12, 2730].

However, IACs are expensive and complex purification systems which suffer from low recoveries for some mycotoxins and their use in multiclass analysis is limited because of their high selectivity. As a consequence, simpler, more efficient, multiclass, and environmentally friendly extraction systems are demanded. Among the different proposals, the so-called QuEChERS (quick, easy, cheap, effective, rugged, and safe) and dispersive liquid-liquid microextraction (DLLME) are becoming increasingly popular treatments.

DLLME is based on the use of a ternary component solvent system; an appropriate mixture of a few microliters of an organic extraction solvent and a small volume of a disperser solvent (miscible with the extraction solvent and with water) is rapidly injected into an aqueous medium, with the result of a stable emulsion. The organic analytes present in the aqueous medium rapidly migrate to the extraction solvent, because of the large contact surface between the organic and the aqueous phases. After phases separation, the organic phase with the analytes of interest is collected and analysed by an appropriate technique [3133]. DLLME has been applied for the determination of OTA in wine by HPLC-MS [34], cereals by HPLC-FL [35], and patulin in apple juices by CE-UV [36]. Also, we have developed two DLLME methods (one of them using an ionic liquid as extraction solvent) for the determination of OTA in wine by capillary-HPLC with laser induced fluorescence detection (LIF) with excellent results [18, 19].

On the other hand, QuEChERS is a fast and inexpensive method widely used in the last years, mainly for the extraction of pesticides and presents some advantages such as its simplicity, minimum steps, and effectiveness for cleaning-up complex samples [37, 38]. It comprises two steps: (i) an extraction based on partitioning via salting-out, involving the equilibrium between an aqueous and an organic layer; (ii) a dispersive SPE (dSPE) for further clean-up using combinations of MgSO4 and different sorbents, such as C18 or primary and secondary amine (PSA). QuEChERS-based methods have been recently reported for the extraction of different mycotoxins in cereal products [3941], bread [42], eggs [43], or spices [44] and in the multiresidue extraction of different contaminants (including mycotoxins) in foods [45, 46]. Also, we proposed this methodology for the determination of OTA in wine samples by capillary HPLC-LIF [18].

3. New Proposals for Determination of Mycotoxins in Different Foods

Considering the above described advances in sample treatments, we have proposed a multiclass method for the determination of mycotoxins in different food commodities. Taking advantage of UHPLC-MS/MS characteristics, we optimised a separation method that allows the determination of 15 mycotoxins in only four minutes. The studied mycotoxins are included in Regulation (EC) number 1881/2006 or considered as dangerous by the IARC [6, 47].

Moreover, in order to propose alternative methods for multiclass mycotoxins determination in scarcely investigated matrices and considering previous results obtained for the determination of OTA, we have tried to explore the advantages of the above mentioned sample treatments (DLLME and QuEChERS) as green, easy, and simple alternatives to other well-established methodologies, as IAC or SPE.

Below, we will explain the UHPLC-MS/MS conditions (common to all the methods) and then we will focus on the studied samples: cereals and pseudocereals, cereal syrups, edible nuts and seeds, and milk thistle. In all cases, a validation was performed in order to assess the compliance with the current requirements for mycotoxin determination in foods [8]. The validation included matrix effect study, establishment of matrix-matched calibrations, limits of detection (LODs) and quantification (LOQs), and intraday and intermediate precision. Moreover, recovery studies at three different concentration levels were carried out by comparison of the signal obtained for a sample spiked with a known concentration of mycotoxins before the sample treatment with the signal of an spiked extract obtained after the sample treatment. As a summary, the most significant analytical characteristics of the developed methods are shown in Table 2.

3.1. Chromatographic and MS Conditions

UHPLC separations were performed on an Agilent 1290 Infinity LC under the conditions summarised in Table 2.

The triple quadrupole mass spectrometer API 3200 (AB Sciex) worked with electrospray ionization in positive mode (ESI+), under multiple reaction monitoring (MRM) conditions shown in Table 3. The ionization source parameters were source temperature 500°C; curtain gas (nitrogen) 30 psi; ion spray voltage 5000 V; GAS 1 and GAS 2 (both of them nitrogen) 50 psi. Moreover, a precursor ion and two product ions (the most abundant for quantification and the other one for confirmation) were selected, obtaining four identification points, fulfilling the requirements established by European Union (EU) for confirmation of contaminants in foodstuff [48].

3.2. Analysis of Cereals and Pseudocereals

Cereals are a commodity of great interest, highly prone to microbial contamination because of their chemical composition. Rice is one of the most consumed cereals in the world. Moreover, brown rice and red rice (obtained by the fermentation of rice with Monascus fungi [49], that can produce CIT [50]) are increasingly chosen by customers because of their health benefits. Other cereal of interest is spelt. Its nutritional properties, high resistance in unfavourable environments, and low fertilization requirements made it increasingly valuable for food product manufacturers and consumers [51]. Matrices of concern are also pseudocereals, such as amaranth, quinoa, and buckwheat. Though botanically they are not true cereal grains, they produce starch-rich seeds consumed like cereals. Pseudocereals are also susceptible to fungal growth and therefore to mycotoxin contamination. However, this issue has received little attention in literature.

Taking into account the interest and the scarce data about the determination of mycotoxins in some of the above mentioned matrices, we developed and validated an analytical method for the simultaneous identification and quantification of 15 mycotoxins (AFB1, AFB2, AFG1, AFG2, OTA, FB1, FB2, T-2, HT-2, CIT, STE, F-X, NIV, DON, and ZEN) in pseudocereals, spelt, and white, red, and brown rice. As a sample treatment we proposed a simple salting out assisted solid-liquid extraction (i.e., a QuEChERS-based extraction, see Figure 2). No further clean-up was required, although matrix effect was higher than for some mycotoxins (aflatoxins, DON, and NIV). Thus matrix-matched calibration was applied. A typical chromatogram corresponding to a spiked white rice sample submitted to the proposed method is shown in Figure 3(a). This methodology has proved to be a suitable and efficient choice for multiclass mycotoxin determination in these matrices, with LOQs below the contents currently regulated. It provides good recoveries (between 60.0% and 103.5%) and precision (RSD lower than 12% in all cases), allows extraction time reduction, and is environmentally friendly. Among all the samples analysed, a red rice sample was positive for AFB1 (Figure 4(a)). The result was confirmed by comparison with a standard method [52].

3.3. Analysis of Cereal Syrups

Cereal syrups are obtained by isolation of starch after wet milling of grains, hydrolysis, and further purification and are widely used in food and pharmaceutical industry. Mycotoxins may also be found in cereal syrups as result of using contaminated raw material, or because of contamination of the final manufactured product by microorganisms during storage [53]. However, there are very few methods for the determination of mycotoxins in these matrices.

Trying to fill this gap, we modified the sample treatment previously described for cereals and pseudocereals (Figure 2) to make it suitable for the determination of 10 mycotoxins (OTA, FB1, FB2, T-2, HT-2, CIT, STE, F-X, ZEN, and DON) in wheat, barley, and rice syrup. Unfortunately, very low recoveries were obtained for aflatoxins and NIV and their quantification was impossible in these matrices. Good recoveries were obtained (between 62.8% and 100.6%), with RSD lower than 11.5% [54]. A chromatogram corresponding to a spiked barley syrup sample submitted to the proposed method is shown in Figure 3(b). It must be pointed out that there is no specific legislation for this kind of matrices, although the low LOQs obtained allowed the determination of these mycotoxins at concentrations lower than the maximum contents usually established by current legislation in different foodstuffs.

3.4. Analysis of Edible Nuts and Seeds

It is well known that nuts and seeds are susceptible to mould growth and consequently to mycotoxin contamination [5557]. Insect feeding-damage is the principal factor leading to preharvest fungal infection and subsequent mycotoxin contamination, but infection may also occur after harvesting and storage. Current EU food safety legislation only regulates the content of aflatoxins in these matrices, with maximum permitted levels which depend on the kind of nut or seed for direct human consumption [6, 58].

Considering the good results obtained with the previously described matrices, the same approach was attempted. In this case, the QuEChERS-based extraction allowed the determination of OTA, T-2, HT-2, STE, CIT, ZEN, FB1, FB2, DON, and F-X. However, a further purification of the extracts was required for subsequent determination of aflatoxins. Thus, an additional clean-up step based on DLLME was proposed in order to reduce matrix effect, allowing the determination of aflatoxins. A flow chart of the whole procedure is shown in Figure 5. This methodology was applied for the determination of 14 mycotoxins in different nuts and seeds (almonds, peanuts, sunflower seeds, pumpkin seeds, walnuts, macadamia nuts, pistachios, hazelnuts, and pine nuts). The low LOQs obtained for aflatoxins, the only mycotoxins regulated in nut and seed matrices, allowed their quantification at concentrations lower than the maximum level established by current legislation. RSD was lower than 11% and recoveries ranged from 60.7% to 104.3%. Among the samples analysed, a sunflower seed sample showed a high content of STE (Figure 4(b)), a walnut sample of ZEN and DON (Figure 4(c)), and a macadamia nut sample of F-X (Figure 4(d)) [59].

3.5. Analysis of Herbal Products: Milk Thistle

The consumption of products with specific nutritional and/or functional characteristics has significantly increased during the last decade. Among them, there are many food supplements containing herbal products and/or their derivatives as ingredients. Previous studies have demonstrated that these materials can suffer from fungi and mycotoxin contamination [60, 61]. However, maximum levels for mycotoxins in food supplements are not established; only, the European Pharmacopoeia sets a maximum level for AFB1 (2 μg kg−1) and for the sum of AFB1, AFB2, AFG1, and AFG2 (4 μg kg−1) for herbal products used as drug ingredients [62].

Most of the scarce methods proposed for the determination of mycotoxins in herbal products use SLE and IAC for clean-up, including the method recommended by the Pharmacopoeia for the determination of AFB1 [62].

In this context, we proposed a method for the multiclass determination of mycotoxins in milk thistle (Silybum marianum), a botanical consumed as food supplement because of its protective effects on the liver. Moreover, although analytical methods for studying the occurrence of mycotoxins in other herbal products (as tea, ginseng, or ginger) have been previously reported, milk thistle has been scarcely studied and only for aflatoxin content [63].

The sample treatment optimized for this matrix is similar to that proposed for nuts and seeds, with some modifications (Figure 5). After QuEChERS-based extraction, FB1, FB2, NIV, DON, and F-X were quantified. However, a second clean-up step based on DLLME was needed for the determination of AFB1, AFB2, AFG1, AFG2, OTA, T-2, HT-2, STE, CIT, and ZEN. The method allowed the quantification of aflatoxins at concentrations lower than their maximum level established in botanicals by Pharmacopoeia. The rest of mycotoxins were determined at concentrations lower than their usual established limits in different foodstuff. Good recoveries were obtained (between 62.3% and 98.9%, except for ZEN in seed samples and CIT in extract). Among the different commercial samples of milk thistle (seeds and natural extract), two were contaminated with T-2 and HT-2 (Figure 4(e)) and ZEN was detected in one of them [64].

Currently, mycotoxins analysis presents several challenges that still need to be addressed and overcome. On the first place, carryover of mycotoxins from contaminated feeds to animal tissues and biological fluids and eventually to products intended for human consumption (meat, milk, and eggs) is a matter of concern. In some animals, mycotoxins undergo metabolic processes and are transformed into other compounds with different toxicity. For instance, most mammals metabolize AFB1 into AFM1 that is transferred to the milk, while poultries metabolize AFB1 into toxic hydroxylated metabolites that may pass to eggs [65]. Animal feed is the first link in food chain and, therefore, the production of safe food depends not only on the manufacturers compliance with current legislation, but also on the use of safe feed by the farmers. As a solution, the use of detoxifying agents (a new group of feed additives) has been proposed in order to decrease the effect of mycotoxin contamination in feeds. These products are mainly adsorbent agents (i.e., activated charcoal, clays, silicates, and some synthetic polymers) that reduce the absortion of mycotoxins in the gastrointestinal tract of animal. Also, some enzymes and microorganisms are used as agents capable of transforming the mycotoxins by modification of their structure, although their use is more limited [66]. However, because of the different properties of mycotoxins, an adsorbent may be effective against a mycotoxin while ineffective against others. Recently, some published studies deal with various natural extracts, such as derivatives of honey [67] and organosulfur compounds derived from allium, like garlic [68], which could be natural alternatives to the usual binders, opening a very interesting field of research.

Another issue of great importance is the evaluation of occurrence of the so-called “masked mycotoxins” in food and feed. The metabolism of some plants (which have natural detoxification mechanisms) can generate conjugated compounds (masked mycotoxins), with different chemical behaviors than the mycotoxins of origin. Thus, some plants are able to transform the relatively nonpolar trichothecenes and ZEN into more polar derivatives by conjugation with sugars, amino acids, or sulfate groups, which are then isolated into the vacuoles. However, these forms can be hydrolyzed to their precursors in the animal digestive tract, thus showing similar toxicity than free mycotoxins. Although this phenomenon has been studied mainly on Fusarium toxins (trichothecenes, zearalenone, and fumonisins), it has also been described for other mycotoxins. In addition, some technological food processings play an important role in the mechanisms of masking, mainly in cereal products, as they can induce reactions with macromolecules such as sugars, proteins, or lipids and inversely release the native forms of mycotoxins by decomposition of masked derivatives. Nowadays, toxicological data on masked mycotoxins are scarce, although several studies highlight the potential threat of these compounds for consumer safety. In particular, the possible hydrolysis of masked mycotoxin (generating the initial mycotoxin) during mammalian digestion would be a risk factor to be considered. For instance, products with an apparent low mycotoxin contamination have induced toxic effects due to the presence of “occult” fumonisins liberated upon hydrolysis and not detected in routine analysis. In this way, masked mycotoxins can quantitatively contribute to the total amount of mycotoxins, especially in cereals [69, 70]. Consequently, the development of analytical methods for multiclass analysis of mycotoxins including their transformation products is a challenge, in order to assess their real risk to the health of consumers.

Finally, the study of the so-called emerging mycotoxins derived from Fusarium fungi (as fusaproliferin, moniliformin, beauvericin, and enniatins), present prevalently in foods from northern Europe and Mediterranean countries, must also be highlighted. Unlike other better studied mycotoxins, permitted maximum levels have not yet been established for these mycotoxins. This is mainly because of the scarce data related to their presence in food, level of contamination, and toxicity. Although no cases of mycotoxicosis by the intake of these mycotoxins have been described, some studies (most in vitro) revealed the possible toxicity of these compounds, which could be increased as a result of the interaction of several mycotoxins present in food [71, 72]. For this reason, more studies are mandatory in order to evaluate the risk of these emerging toxins. Once again, reliable analytical methods are urgently required.

5. Conclusions

Alternative UHPLC-MS/MS analytical methods for multiclass determination of mycotoxins based on QuEChERS and DLLME for sample treatment have been developed. The proposed methods have been evaluated in diverse food commodities, most of them scarcely investigated. They showed as general advantages their efficacy, simplicity, versatility, and accuracy, as well as their low impact on the environment, shorter analysis time, and the relatively low-cost, compared with conventional IAC. Thus they fulfill the current requirements of analytical methods for the determination of contaminants. However, there are some aspects concerning mycotoxin determination that still are a challenge for the scientific community, as the development of new analytical methods including the determination of masked or emerging mycotoxins. The proposed methodologies could be applied also to these analytes, opening new perspectives that, combined with powerful analytical techniques, such as UHPLC-MS/MS, offer interesting perspectives in this field.

Conflict of Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.