QuEChERS LC–MS/MS Screening Method for Mycotoxin Detection in Cereal Products and Spices

We developed and validated a screening method for mycotoxin analysis in cereal products and spices. Ultra-high-performance liquid chromatography coupled with tandem mass spectrometry (UHPLC–MS/MS) was used for the analysis. Dispersive solid-phase extractions (d-SPEs) were used for the extraction of samples. Ochratoxin A (OTA), zearalenone (ZEA), aflatoxins (AFLA; AFB1, AFB2, AFG1, AFG2), deoxynivalenol (DON), fumonisin (FUMO; FB1, FB2, FB3), T2, and HT2 were validated in maize. AFLA and DON were validated in black pepper. The method satisfies the requirements of Commission Regulation (EC) no. 401/2006 and (EC) no. 1881/2006. The screening target concentration (STC) was under maximum permitted levels (MLs) for all mycotoxins validated. The method’s performance was assessed by two different proficiencies and tested with 100 real samples.

Mycotoxicosis is a human and animal disease caused by ingestion, inhalation, or skin contact of mycotoxins [7]. The symptoms, target organ of systemic toxicity, and clinical outcome depend on several parameters such as type of mycotoxins, intake levels, time, and route of exposures [8]. Mycotoxicosis can be acute or chronic, and different symptoms characterize these two forms. AFB 1 , AFB 2 , AFG 1 , and AFG 2 are produced by fungi of the genus Aspergillus. The most representatives fungi that produce AFLA are Aspergillus flavus and Aspergillus parasiticus [9,10]. AFLA are characterized by a lipophilic structure ( Figure 1) derived from the same precursor, versiconal hemiacetal acetate [11].
The acute ingestion of AFLA (aflatoxicosis) can lead to several symptoms that include gastrointestinal problems (diarrhea, abdominal pain), nervous system dysfunctions (epilepsy, coma), liver damage (jaundice, hepatitis), and even death [12,13]. Chronic exposure to AFLA is associated with multiple-organ cancer, immunosuppression, and other diseases [14]. AFB 1 , AFB 2 , AFG 1 , and AFG 2 are classified as group 1 (carcinogenic to humans) by the Agency of Research on Cancer (IARC) and have mutagenic and teratogenic effects in humans [15]. Once ingested, AFLA are converted by cytochrome P450 into high reactive epoxides that can create adducts with nucleobases [16]. Hepatocellular carcinoma (HCC) is strictly correlated with dietary exposure to AFB 1 and adducts excreted in urine [17,18]. Int  The acute ingestion of AFLA (aflatoxicosis) can lead to several symptoms that include gastrointestinal problems (diarrhea, abdominal pain), nervous system dysfunctions (epilepsy, coma), liver damage (jaundice, hepatitis), and even death [12,13]. Chronic exposure to AFLA is associated with multiple-organ cancer, immunosuppression, and other diseases [14]. AFB1, AFB2, AFG1, and AFG2 are classified as group 1 (carcinogenic to humans) by the Agency of Research on Cancer (IARC) and have mutagenic and teratogenic effects in humans [15]. Once ingested, AFLA are converted by cytochrome P450 into high reactive epoxides that can create adducts with nucleobases [16]. Hepatocellular carcinoma (HCC) is strictly correlated with dietary exposure to AFB1 and adducts excreted in urine [17,18].
FUMO (FB1, FB2, FB3) are produced by fungi of the genus Fusarium [19]. FB1 contamination is common in cereals, and it is the most toxic FUMO [20]. Acute ingestion of FUMO can cause gastrointestinal problems, and they are considered possibly carcinogenic to humans (group 2B) by IARC [15,21]. FUMO can interfere with folic acid metabolism (teratogenic effects), cause inhibition of sphingolipid biosynthesis, and have carcinogenic effects [11,21]. They are polar compounds and are not soluble in non-polar solvents ( Figure 1) [21]. Chronic exposure to AFB1 and FUMO can lead to liver cancer (sum of carcinogenic effect) [22].
Fusarium species also produce DON, which is one of the most common mycotoxins in cereals [23]. It is considered not classifiable as to carcinogenicity to humans (group 3) [15]. The acute toxicity is mainly gastrointestinal, with nausea, diarrhea, and abdominal pain [24]. DON is also called vomitoxin since it can induce emesis [25]. It can also cause dysfunctions of the immune, neuroendocrine, and cardiovascular systems [26]. DON is a polar molecule that can resist at high temperatures, and it is soluble in polar organic solvents [27,28]. It is classified as non-macrocyclic trichothecenes [29].
Non-macrocyclic trichothecenes also include T2 and HT2 (C-4 deacetylated form of T2, Figure 1) produced from Fusarium species [30]. The name derived from trichothecin, FUMO (FB 1 , FB 2 , FB 3 ) are produced by fungi of the genus Fusarium [19]. FB 1 contamination is common in cereals, and it is the most toxic FUMO [20]. Acute ingestion of FUMO can cause gastrointestinal problems, and they are considered possibly carcinogenic to humans (group 2B) by IARC [15,21]. FUMO can interfere with folic acid metabolism (teratogenic effects), cause inhibition of sphingolipid biosynthesis, and have carcinogenic effects [11,21]. They are polar compounds and are not soluble in non-polar solvents ( Figure 1) [21]. Chronic exposure to AFB 1 and FUMO can lead to liver cancer (sum of carcinogenic effect) [22].
Fusarium species also produce DON, which is one of the most common mycotoxins in cereals [23]. It is considered not classifiable as to carcinogenicity to humans (group 3) [15]. The acute toxicity is mainly gastrointestinal, with nausea, diarrhea, and abdominal pain [24]. DON is also called vomitoxin since it can induce emesis [25]. It can also cause dysfunctions of the immune, neuroendocrine, and cardiovascular systems [26]. DON is a polar molecule that can resist at high temperatures, and it is soluble in polar organic solvents [27,28]. It is classified as non-macrocyclic trichothecenes [29].
Ochratoxin A (OTA) is the most important and toxic mycotoxin among ochratoxins [36]. It is an isocumaric derivate with a β-phenylalanine ( Figure 1) [11]. Aspergillus and Penicillium species can produce OTA; Aspergillus ochraceus and Penicillium verrucosum are the most common [37]. It is located in group 2B in the IARC classification [15]. Its toxicological activity can affect neuronal, renal, and immune systems [36]. OTA has a high affinity for serum albumin and, therefore, it is characterized by a high plasma half-life [38]. It has been found in several foods such as milk, coffee, wine, and vegetables [38]. It also has been reported in baby food [39]. OTA is soluble in polar organic solvents [40].
Zearalenone (ZEA) is a resorcylic acid lactone ( Figure 1) produced by the Fusarium genus species [8]. It is classified as group 3 by IARC [15]. It is associated with endocrine alteration with clinical manifestations of hyperestrogenism [41]. Swine are the most sensitive species to the toxic effects of ZEA; clinical signs of acute mycotoxicosis are swelling of the vulva, prolonged estrus, pseudopregnancy, and infertility [37]. ZEA can be found in crops, wheat, maize, and snacks [42].
In developing countries, there have been relatively recent episodes of mycotoxicosis. One-hundred and twenty-five people died in Kenya (Africa) in 2004 [50], and there were 100 victims in 1974 in India [51]. There were alimentary toxic aleukia (ATA) cases in the USSR in the 1950s associated with consuming trichothecenes in wheat [37].
One of the biggest problems related to the cereal contaminations of mycotoxin is the carry-over effect. Mycotoxins consumed in feeding stuff are absorbed and metabolized by animals and can be found in meat, eggs, and milk [63].
Different cereal-based products for human consumptions are mixed with black pepper for flavoring purposes or preservatives, leading to possible increases in mycotoxin contents [60]. Spices are usually dried on the ground in the open air in poor hygienic conditions that promote the growth of molds and production of mycotoxins [64,65]. The Rapid Alert System for Food and Feed (RASFF) reported that the third most common matrices with AFLA in Europe are herbs and spices [66].
The Commission Regulation (EC) no. 401/2006 established MLs of mycotoxins for different food and feed [67]. AFLA (AFB 1 , the sum of AFB 1 + AFB 2 + AFG 1 + AFG 2 , AFM 1 ), OTA, patulin, DON, ZEA, FUMO (FB 1 + FB 2 ), T-2 and HT-2, citrinin, ergot sclerotia, and ergot alkaloids were regulated in Europe. The regulation does not consider the assessment of exposure to the sum of mycotoxins [33] with potentially higher exposure. Furthermore, the presence of masked mycotoxins can lead to an underestimation of their concentration in a sample [68].
The QuEChERS (Quick Easy Cheap Effective Rugged and Safe) method is becoming one of the most used dispersive solid-phase extraction (d-SPE) methods in food safety [69,[85][86][87][88]. According to Web of Science, more than 1200 scientific papers cover aspects or use QuEChERS for extraction procedures. The common steps of QuEChERS methods are LLE between an organic and an inorganic phase, addition of salts, agitation of the sample, removal of the supernatant, and d-SPE between supernatant and sorbents [89]. Competent authorities already use QuEChERS extraction to analyze various molecules such as pesticides and antibiotics [90,91].
Multiclass multi-residue methods (MRMs) with QuEChERS d-SPE are preferred over other cleanup solutions requiring numerous steps and cost more time. Fast cleanup reduces the costs, time, and space required with solid-phase cartridges [92].
QuEChERS does not require high solvents volume and can be performed with basic laboratory devices [93] and with an extraction procedure that allows for the analysis of 20-30 samples in 1 h.
The majority of methods are focused on increasing mycotoxins analyzed in a single matrix [94][95][96][97], especially in maize and in wheat [98].There are already QuEChERS mycotoxin procedures in the literature [99][100][101]. However, they are validated in a single matrix or use high-resolution mass spectrometry (HRMS), which is more expensive than an LC-MS/MS.
The present work aimed to validate QuEChERS extraction together with a fast and reliable LC-MS/MS method to detect 12 mycotoxins in cereals and 5 mycotoxins in black pepper, following the (EC) Regulation limits 1881/2006 and (EC) Regulation 401/2006.

Working Solutions
Standards were mixed to obtain the following working solution: OTA, AFLA (AFB 1 , AFB 2 , AFG 1 , AFG 2 ), FUMO (FB 1 , FB 2 , FB 3 ), ZEA, DON, MIX T2 (T2, HT2). Acetonitrile was used as solvent, except for SMix1 (low level) and Smix2 (high level), where methanol was used as solvent. A working solution was used to fortify the blank matrix sample, and low level and high level were used as calibration solutions (Table 1).

Sample Preparation
The method was applied to screen mycotoxins in Sicily (Southern Italy). About 10 kg of maize and 5 kg of black pepper were collected from 2 different local vendors in Palermo (Sicily) and used for validation procedures. Samples were grounded using a Mixer B-400 laboratory mill by BÜCHI (Cornaredo, Italy) at ambient temperature with knives' rotation speed of 9000 rpm. Samples grounded were stored at −10 • C until analysis.

Sample Extraction
About 5.0 ± 0.1 g of the sample was weighted in a falcon tube of 50 mL. A total of 150 µL of OTA-d5 with a concentration of 100 µg/L was added to all samples (3.0 µg/Kg, 7.5 µg/L). One sample for each matrix was fortified with the working solution as described above (Table 1). After 10 minutes, 10 mL of bidistilled water and 10 mL of an acetonitrile/formic acid solution (80:20 v/v) were added to the sample. The sample was vortexed for 15 min and left to rest for 15 minutes at −20 • C. A mixture of salt (4 g MgSO 4 , 1 g NaCl, 0.5 g sodium citrate dibasic sesquihydrate, 1 g sodium citrate tribasic dihydrate) was added, handle shacking occurred for about 1 minute, and the mixture was centrifugated for 10 min at 5000 rpm. The supernatant was transferred into a mix of salt, 900 mg MgSO 4 , and 150 mg Supelclean PSA. After 1 minute of handle shacking and 5 minutes of centrifugationat 5000 rpm, 3 mL of the supernatant was evaporated (40 • C) and dissolved in 600 µL of methanol/water (50/50 v/v). The sample was ready for the injection.

Instrumentation
The analysis was performed on a Thermo Fischer Ultra High Performance Liquid Chromatography(UHPLC ) system (Thermo Fisher Scientific, California, CA, USA) consisting of an ACCELA 1250 quaternary pump and an ACCELA autosampler. A Thermo Scientific Hypersil Gold reversed-phase UHPLC column (50 mm, 2.1 mm ID, 1.9 µm) was used for mycotoxin analysis.
The mobile phase (Table 2) was a time-programmed gradient using water (eluent A) and methanol (eluent B). Both contained 2.50 mM of ammonium formate and 0.1% formic acid. The chromatographic run started with 100% of A with a variation of 20% in 0.5 min. The conditions were maintained for 1 min, and then A decreased until a percentage of 40% in 0.1 min. Linear decrease of A occurred with a total percentage of B of 100% in 2.6 min. Conditions were maintained for 0.7 min, and the system returned to 100% A and 0% of B in 0.1 min for 1 min.  The ESI parameters were set as follows: capillary temperature 310 • C, vaporizer temperature 300 • C, sheath gas pressure 40 psi, auxiliary gas pressure 30 psi, capillary voltage 4.8 kV. Collision gas, peak resolution, scan time, and scan width parameters were set as described in [102]. Trace Finder version 4.1 from Thermo Fisher (Kandel, Germany) was used to record and elaborate data. Results obtained from the analyte's direct infusion (parent, product 1, product 2, CE) and chromatography runs can be seen in Table 3.  The semi-quantification of analytes was made by extrapolation of data obtained in the linear regression between low level and high level (Table 1), and the concentration in the sample was obtained with the following formula: C c is the concentration in the matrix (µg/Kg), C s is solution (µg/L), and D is the dilution factor.
Matrix effect (MEs) in maize and spice were calculated as described by Juan Sun et al. [99]: A bm is the area in the blank matrix and A s is the area of mycotoxin standard in solvent (n = 3).
Accuracy was evaluated with the extraction of 20 fortified samples for each matrix.

Real Samples
After validation, the method was tested with real samples collected during an inspection in Palermo (Sicily). Maize (n = 25) and wheat (n = 25) were collected from the same local vendors; black pepper (n = 25) and coffee (n = 25) were from two different local supermarkets in Palermo (Sicily). All samples were transported daily in the lab and, if needed, grounded as described in Section 2.4. The extractions were performed as described in Section 2.5.

Extraction Solvent and Cleanup
The extraction method developed allows for extracting 12 mycotoxins in cereals and 5 in black pepper with a cheap and fast procedure. The best performance was attained using water and a mixture of acetonitrile/formic acid (80:20 v/v) Other ratios (60:40, 40:60, 20:80, 50:50) were examined to reach this conclusion (Figure 2). Acetonitrile and formic acid enhance analytical performance, as previously reported in the literature [104]. Acetonitrile/water extraction (in different percentages) is one of the most common mixtures used for mycotoxin analysis [105]. Acetonitrile can reduce the extraction of lipophilic materials such as fats and has a high capacity to extract molecules characterized by different polarities [106]. All mycotoxins analyzed are soluble in acetonitrile, and a higher percentage of acetonitrile can improve analytes' extraction. OTA, AFLA, and ZEA are soluble in polar organic solvents such as methanol and acetonitrile [40,107,108]. FUMO are hydrophilic mycotoxins ( Figure 1) and are soluble in the same solvents and water [109]. Regarding trichothecenes, T2, HT2, and DON are low soluble in water but highly soluble in ethanol and organic solvents [110]. For these reasons, acetonitrile, methanol, and water were used for the extraction procedure. Furthermore, water is added in high starch or low water matrices to reduce the interaction between them and analytes [111].
Formic acid and citrate salts decrease pH and contrast PSA's effect of increasing pH in the second step [112,113]. MgSO4 and NaCl increase the recovery of polar analytes, and MSO4 with PSA performs better and increases mycotoxins' recovery. [106,114].
The ratio between MgSO4 and PSA in the sorbent d-SPE step is always greater than 1 in all the literature, even if the quantity can change [106]. However, 900 MgSO4 and 150 mg of Supelclean PSA is already used in different analytical techniques that use QuEChERS methods for the extraction of analytes such as pesticides [115,116], likewise for the 4:1 w/w ratio between MgSO4 and NaCl as salt added in the extraction procedure [69,113,117,118]. Regarding trichothecenes, T2, HT2, and DON are low soluble in water but highly soluble in ethanol and organic solvents [110]. For these reasons, acetonitrile, methanol, and water were used for the extraction procedure. Furthermore, water is added in high starch or low water matrices to reduce the interaction between them and analytes [111].

Matrix Effect in Mycotoxin Analysis
Formic acid and citrate salts decrease pH and contrast PSA's effect of increasing pH in the second step [112,113]. MgSO 4 and NaCl increase the recovery of polar analytes, and MSO 4 with PSA performs better and increases mycotoxins' recovery. [106,114].
The ratio between MgSO 4 and PSA in the sorbent d-SPE step is always greater than 1 in all the literature, even if the quantity can change [106]. However, 900 MgSO 4 and 150 mg of Supelclean PSA is already used in different analytical techniques that use QuEChERS methods for the extraction of analytes such as pesticides [115,116], likewise for the 4:1 w/w ratio between MgSO 4 and NaCl as salt added in the extraction procedure [69,113,117,118].

Matrix Effect in Mycotoxin Analysis
LC-MS/MS is susceptible to MEs, a common and unpredictable problem that can influence the validation process [119]. The ESI ionization is subject to ion suppression more than other atmospheric pressure ionization (API) techniques such as atmospheric pressure chemical ionization (APCI) [120].
Isotope-labeled internal standards can be useful to handle with MEs and are considered the "gold standard" approach [121]. This approach is useful for predicting the strong MEs in the analytical signal [122]. However, isotope-labeled internal standards are very expensive and do not exist for all mycotoxins, and recovery is not required in the screening method [123]. For these reasons, the use of a fortified sample for each analyte session seems more reasonable. Nevertheless, OTA-d5 was used in validation procedure and in real samples to evaluate the entire process.
In the present work, a strong ME was observed for AFLA with a range of (+43.74, −20.18) in maize and (−26.17, −0.89) in black pepper (Table 4). MEs have been reported several times in the literature, but with discordant data, enhancing or suppressing black pepper [122,124,125] and maize [122]. FUMO was strongly enhanced with an overall increase of 48.51%, as already reported by Beltrán et al. (2009) [104]. Curiously, OTA suppression in black pepper and maize was similar and had been reported before for maize [126].
The strong MEs in the mycotoxin analysis was caused by the lipid/water/protein content of matrices analyzed [127]. For these reasons, the analysis of different matrices should be validated according to (EC) no. 401/2006 that divides food into commodity groups to validate screening methods [67]. In this case, the research validated the commodity group of difficult or unique commodities (black pepper) that include cocoa beans and products thereof; copra and products thereof; coffee, tea, and licorice and the commodity group cereal grain and products thereof (maize), which include wheat, rye, barley, rice, oats, wholemeal bread, white bread, crackers, breakfast cereals, and pasta. Confirmatory methods must be validated in each matrix and for these reason screening methods, increasing the chance to discover new incidents and protect the consumers from high mycotoxin exposure [67].

Validation Parameters
The analytical parameters of the methods used for mycotoxin analysis are regulated in Europe by the Commission Regulation (EC) no. 401/2006 that defines the criteria of analysis for the official control of the levels of mycotoxins in foodstuffs [128]. The analysis was performed under RSDRi, and all cut-off levels were under STC. The linear correlation of level tested in the range of linearity was acceptable (r 2 > 0.99).

Instrumental Method
A cheap and simple screening method of 12 (cereal) and 5 (black pepper) mycotoxins was validated. Ammonium formate and formic acid were used to form ammonium adduct and protonated precursor ion, respectively. The ammonium adduct was selected as precursor ion only for T2 and HT2 (Table 3). All mycotoxins were detected within 4 min. The adequate resolution was obtained between ions with the same m/z, which would be indistinguishable from the mass spectrometer if they coalesced. LC-MS/MS is used frequently to analyze molecules regulated by EU Legislation [129][130][131][132], and it is possible to perform semi-quantitative analysis [133]. For this reason, LC-MS/MS was preferred over other screening analytical techniques such as enzyme-linked immunosorbent assay (ELISA) [134,135].
Several parameters can influence the performance of mycotoxin analysis. Regarding instrument setting, the positive ion mode (ESI + ) was chosen because there is a better response for the overall of mycotoxin [104,127], especially AFLA that among mycotoxins are more regulated in the EU Legislation [67].
Water and methanol as mobile phases provided the best performance for peak resolution and run time for the chromatographic run. The same result was reported in the literature [84,99,127]. Ammonium acetate and formic acid addiction in the mobile phase increase the analytical performance [94,104,136,137]. The best results were achieved with 0.1% of formic acid and 2.5 mM of NH 4 COOH, as already reported by other authors [123].
With a total chromatography run of 6 min and an extraction procedure that takes approximately 1 hour, the method developed is faster than other methods already reported in the literature [138][139][140][141][142] and is useful for quick screening.
AFLA are not sensitive to heat treatment and can increase during the food storage period. A quick screening before storage can be useful to have some data on mycotoxins' presence [48,143,144].
Black pepper is a less studied matrix, and increased data on mycotoxins presence can help in the risk assessment of mycotoxins exposure. This is especially the case because there is a possible co-occurrence of mycotoxin due to the multiple fungal infections [145] and because they are used as flavor-enhancers in convenience foods.
However, this result has not been reached without compromises. Masked mycotoxins such as 3-acetyl deoxynivalenol (3-AcDON) and 15-acetyl-deoxynivalenol (15-AcDON) that present different toxicities [146] are not currently regulated in EU Legislation and were not analyzed. Some peaks are moderately overlapped, such as for FB 2 (Table 5).

Conclusions
A new method for detecting 12 mycotoxins in cereals and 5 mycotoxins in spices (black pepper) was developed and validated according to (EC) no. 401/2006. QuEChERS extraction was used effectively. The best performances were obtained with acetonitrile/formic acid (80:20 v/v) as extraction solvent. Strong MEs were observed in all the FUMO analyzed in maize, while AFLA had enhancing or suppressing effects. In black pepper, there was a suppression of signals for all mycotoxins analyzed. Six PTs were developed to evaluate the performance of the method. The method was applied to 100 real samples (25 maize, 25 wheat, 25 black pepper, and 25 coffee). Two samples had a detectable amount of mycotoxin, maize (OTA, 2.53 µg/Kg), and black pepper (OTA, 1.85 µg/Kg, and AFB 2 , 0.358 µg/Kg). The method proposed is suitable for screening and routine analysis to monitor mycotoxins content in foodstuff following European Regulamentation. Further studies are needed to increase the number of mycotoxins analyzed and increase commodity groups analyzed with the same method.