Speaker/Author: W.P. Nxumalo*, ** Co-authors: Y. Naudé**, and M. Fernandes-Whaley*

Transcription

1 Combining Ultrasonic Extraction (USE) and Solid Phase Extraction (SPE) for the Analysis of 11 Mycotoxins in Maize by Isotope Dilution Liquid Chromatography Tandem Mass Spectrometry (SIDA-UHPLC-MS/MS) Speaker/Author: W.P. Nxumalo*, ** Co-authors: Y. Naudé**, and M. Fernandes-Whaley* * National Metrology Institute of South Africa, CSIR Campus, Meiring Naude Road, Brummeria, Pretoria 0182, South Africa ** Department of Chemistry, Faculty of Agricultural and Natural Sciences, University of Pretoria, Lynwood Road, Hatfield, Pretoria 0083, South Africa wnxumalo@nmisa.org Phone: Abstract An analytical method was developed for determination of aflatoxins B 1, B 2, G 1 and G 2, fumonisins B 1, B 2, deoxynivalenol (DON), ochratoxin A (OTA), zearalenone (ZEN), HT-2 and T-2 Toxins in maize. The method consists of solid-liquid extraction and a clean-up step using a Select HLB SPE column, followed by liquid chromatography with electrospray ionisation coupled to tandem mass spectrometry. Mycotoxins were extracted from maize with acidified acetonitrile/water using an ultrasonic bath. Chromatographic separation was achieved on an Acquity BEH C 18 UHPLC column with a solution of methanol/water/formic acid as the mobile phase. Recoveries between % were achieved for AFG 2, ZEN, OTA, AFB 1 and DON whereas enhanced recoveries were observed for FB 1, AFG 1, FB 2, T-2 Toxin, HT-2 Toxin and AFB Introduction Mycotoxins are toxic metabolites of fungi that occur in agricultural commodities such as maize. The occurrence of mycotoxins in food and feed affects human health. Its presence is therefore strictly regulated, resulting in economic challenges for countries where mycotoxin contamination is poorly managed. The Food and Drug Administration (FDA) estimates an annual loss of $ 932 million due to mycotoxin contamination [1], and the World Health Organisation (WHO) recognises the contamination of food and feed by major mycotoxins as a source of food-borne illness [2]. Consequently, the WHO and Food and Agriculture Organisation (FAO) have evaluated the toxicity of the mycotoxins and established food safety measures. Mycotoxin contamination of food can be severe thus accurate mycotoxin identification and quantification are required. The determination of contaminants in food matrices frequently involves extensive sample preparation prior to instrumental analysis. Solid-liquid extraction (SLE) is one of the most used extraction techniques for isolating mycotoxins from grains, cereal food stuffs and other solid food matrices [3]. The choice of solvent depends on the analytes to be analysed; deoxynivalenol (DON), T-2 and HT-2 Toxins are highly soluble in polar solvents whereas zearalenone (ZEN) and ochratoxin A (OTA) are highly soluble in non-polar solvents [4]. Currently, a mixture of acetonitrile (ACN)/water (84/16, v/v) is the most frequently used solvent for multi mycotoxin extraction in maize [5-7]. However, there is no gold-standard

2 solvent mixture that can extract all analytes efficiently. Other solvent mixtures and ratios have also been investigated in literature and include predominantly methanol (MeOH)/water (greener alternatives) [8]. Varga et al. (2012) used a two-step sequential extraction using acidified mixture of acetonitrile and water, varying solvent composition at each step [9]. A methanol/water mixture (70/30, v/v) was used as an extraction solvent using ultrasonic extraction for determination of fumonisins in maize. Also, an acetonitrile/water (80/20, v/v) mixture was used for the determination of OTA in cereal products [10]. The ultrasonic extraction (USE) is used to improve extraction efficiency. This technique allows several extractions to be performed simultaneously and it is relatively inexpensive as it requires no specialised laboratory equipment. A disadvantage of USE is that it is not automated and it is not suitable for volatile analytes. Li et al. (2012) developed a method for determination of fumonisins FB 1 and FB 2 in maize by LC-ESI-MS/MS, omitting the clean-up step and the extraction was done at room temperature for 10 min [11]. USE is limited by selectivity and sample pre-concentration capabilities, thus further clean-up and/or enrichment steps are often required for quantification of trace contaminants in food. An analytical method based on USE and immunoaffinity column clean-up coupled with LC and post column derivatisation-fluorescence was developed for the simultaneous determination of multiple mycotoxins in nutmeg samples [12]. Solid phase extraction (SPE) is an invaluable tool to minimise or possibly eliminate matrix effects associated with mass spectrometric techniques resulting in accurate analyte quantification. It has been used as a clean-up step during multi mycotoxin analysis. Hydrophilic-lipophilic balanced (HLB) SPE columns are increasingly used in multi mycotoxin analysis due to advantages such as relatively low cost and that they can be used for isolating acidic, neutral and basic compounds. A clean-up procedure based on reverse phase SPE Oasis HLB columns was used in the determination of trichothecenes in cereal and cereal based products by LC-MS/MS [13]. The aim of this study was to combine the ultrasonic extraction technique and affordable solid phase extraction clean-up for the determination of multiple mycotoxins in maize analysed using isotope dilution liquid chromatography tandem mass spectrometry (LC-SIDA-MS/MS). Two different solvent combinations were compared during extractions. Two different brands (Supelco Select and Waters Oasis) of HLB SPE clean-up were also compared. 2. Experimental 2.1. Materials and reagents All standards including the unlabelled; aflatoxin B1 (AFB 1 ), aflatoxin B2 (AFB 2 ), aflatoxin G1 (AFG 1 ), aflatoxin G2 (AFG 2 ), fumonisins B1 (FB 1 ), fumonisins B2 (FB 2 ), ochratoxin A (OTA), zearalenone (ZEN), deoxynivalenol (DON), HT-2 and T-2 Toxins, and the U-[ 13 C]- labelled; AFB 1, AFG 1, FB 1, FB 2, OTA, ZEN, DON, mycotoxins were purchased from Romer Labs GmbH (Tulln, Austria) ( 99% purity). All the standards were individual stock solutions in acetonitrile or acetonitrile/water (1/1, v/v). Working solutions, for both the unlabelled and labelled mycotoxins, were prepared gravimetrically from the stock solutions in acetonitrile. The working solution for the unlabelled mycotoxins had the following concentrations: AFB 1, AFG 1, 100 ng/g; AFB 2, AFG 2, 25 ng/g; FB 1, DON, HT-2, T-2 Toxin, 5000 ng/g; FB 2, ZEN, 2500 ng/g; OTA, 250 ng/g. The working solution for the labelled mycotoxins had the following concentrations: AFB 1, AFG 1, 10 ng/g; FB 1, DON, 500ng/g; FB 2, OTA, 250 ng/g; ZEN, 300 ng/g. Individual stocks and the two working solutions were

3 stored at -20 C. Prior to usage, the working solutions were brought to room temperature and mixed thoroughly. Acetonitrile (ACN) and methanol (MeOH), both HPLC grade, were purchased from Merck (Darmstadt, Germany). Formic acid (HCOOH) was obtained from Fluka (Steinheim, Germany). Ultra-pure water was produced from distilled water using a Milli-Q system (Millipore Corp., Bedford, MA, USA). Ammonium formate (NH 4 COOH) ( 99% purity), was obtained from Sigma-Aldrich (Steinheim, Germany). Select HLB and Oasis HLB columns were sourced from Sigma-Aldrich and Waters, respectively. Reference materials (RM) for mycotoxins in maize with well-defined analyte concentrations were sourced from Trilogy and Food Analysis Performance Assessment Scheme (FAPAS). Trilogy maize has the following concentrations: AFB 1, 18.8 ng/g; AFB 2, 0.9 ng/g; AFG 1, 2.4 ng/g; FB 1, 28.3 mg/g; FB 2, 7.1 mg/g; OTA, 5.57 ng/g; ZEN, 352 ng/g; DON, 2.6 mg/g; HT- 2, 523 ng/g; T-2 Toxin 263 ng/g. Whereas FAPAS maize had the following concentrations: AFB 1, 8.01 ng/g; OTA, 4 ng/g; ZEN, 344 ng/g; DON, 1.79 mg/g Sample preparation MeOH/H 2 O mixture: Ground and homogenised maize samples (5.00 ± 0.01 g) were weighed into 50 ml polypropylene tubes. Methanol/water (40 ml, 70/30, v/v) was added. Extraction was performed in an ultrasonic bath (Branson 8800) for 60 min at room temperature, output power was set to 280 W. After extraction the tubes were centrifuged for 5 min (3 500 g). ACN/H 2 O/HCOOH mixture: Ground and homogenised maize samples (5.00 ± 0.01 g) were weighed into 50 ml polypropylene tubes. Double extraction was performed. The first extraction was performed with 20 ml of extraction solvent 1, ACN/H 2 O/HCOOH (80/19.9/0.1, v/v/v), in an ultrasonic bath for 60 min at room temperature. After extraction, the tubes were centrifuged for 5 min (3 500 g) and the extract was decanted into a clean, new polypropylene tube. The residue was extracted for the second time with 20 ml of solvent 2, ACN/H 2 O/HCOOH (20/79.9/0.1, v/v/v), on the ultrasonic bath for 30 min at room temperature. Thereafter, the tubes were centrifuged again for 5 min (3 500 g) and the supernatant was combined with the first extract. The combined extracts were centrifuged again for 5 min (3 500 g). An aliquot of 5 ml of the extract was evaporated to dryness under a gentle stream of Nitrogen gas (N 2 ) at 50 C. The residue was dissolved in 2 ml 5% MeOH in Milli-Q water and loaded into a pre-conditioned Select HLB column (Supelco, Sigma Aldrich). The analytes were eluted with 5 ml ACN/MeOH/HCOOH (50/49.9/0.1, v/v) and collected in a silanized test tube. The eluate was collected and evaporated to dryness. To the dried residue, 250 µl of the [ 13 C]-labelled working solution was added and dried under a gentle stream of N 2 at 50 C. The residue was reconstituted in 250 µl of the starting mobile phase composition (MeOH/H 2 O, 35/65, v/v), filtered through a 0.22 µm nylon syringe filter and then transferred into an HPLC vial. The solution was mixed and a 10µL aliquot was used for LC-MS/MS analysis UHPLC-MS/MS An Acquity Ultra High Performance Liquid Chromatography (UHPLC) system interfaced with a Micromass Quattro Premier XE triple-quadrupole mass spectrometer (Waters Corp, MA, USA) was used. MassLynx software version 4.1 (Waters Corp, MA, USA) was used for

4 Peak Area % Recovery instrumental control, data acquisition and processing. LC separation was achieved using an Acquity UPLC BEH C 18 column (2.1 mm i.d. x 100 mm; 1.7 µm particles) (Waters Corp.). The column and auto sampler tray temperature were controlled at 30 C and 4 C, respectively. The injection volume was 10 µl and the mobile phase flow rate was 0.35 ml min -1. The mobile phase composition was (A) 5 mm NH 4 COOH in MeOH: Milli-Q (95/5, v/v) and (B) 5 mm NH 4 COOH in Milli-Q, ph 3. The MS instrument was operated in electrospray ionisation (ESI) in fast polarity switching mode. Ten of the analytes were analysed in ESI positive (ESI+) mode, ZEN was detected in ESI negative (ESI-) mode. Analyte specific MS/MS parameters were obtained by direct infusion of each standard solution into the ESI source. The two most abundant product ions generated from each precursor ion were chosen as the multiple reaction monitoring (MRM) transitions of each analyte. 3. Results and discussion 3.1. Sample preparation In this study we compared MeOH/H 2 O and sequential ACN/H 2 O/HCOOH extraction mixtures. Of these two solvent mixtures, only the ACN/H 2 O/HCOOH mixture has been used for determination of multiple mycotoxins simultaneously [12]. MeOH/H 2 O gave recoveries between % for aflatoxins, and 78% for DON while ACN/H 2 O/HCOOH gave recoveries between % and 120% for aflatoxins and DON respectively in spiked maize samples, Table 2 lists spike concentrations. Apparent recoveries between % has been reported using this sequential extraction [12]. Aflatoxins and DON are polar compounds and their extraction efficiency is affected by the solvent polarity therefore they were chosen to evaluate extraction efficiency of the solvent mixtures. Also, aflatoxins are regulated at low levels [14] of 5 ppb, thus the extraction solvent should be able to extract aflatoxins with good recoveries. ACN/H 2 O/HCOOH was chosen as the extraction solvent for extracting mycotoxins in maize in this study due to high extraction efficiency for most analytes. To optimise the extraction time, different extraction times (10, 15, 30, 45, and 60 min) were tested to extract mycotoxins in certified matrix reference material. For most analytes a plateau was reached between 30 and 60 min, and thus 60 min was chosen for further analysis. Figure 1(A) shows the effect of extraction time on AFB 1, DON and FB 1 (secondary axis) peak areas AFB1 DON FB Time/min A O-HLB S-HLB B Figure 1. (A) Effect of extraction time on analyte peak areas, (B) recovery test of two HLB SPE columns.

5 An optimised clean-up method based on SPE for purification and concentration of maize extract was used with the objective of getting optimal recoveries and minimal matrix effects. Oasis HLB SPE columns have been used in the determination of nine mycotoxins in maize by HPLC-MS/MS [7]. We used Select HLB SPE columns for the determination of 11 mycotoxins in maize. Preliminary experiments were done to evaluate matrix removal and recovery efficiencies for Oasis HLB and Select HLB SPE columns in multi mycotoxin analysis. Figure 1(B) shows recoveries for these two SPE columns, both columns gave relatively comparable results. Select HLB columns were chosen as they are more costeffective relative to Oasis HLB UHPLC-MS/MS Analyte specific MS/MS parameters, including the determination of precursor and product ions and corresponding collision energies, were obtained by direct infusion of single unlabelled and labelled analyte solution into the ESI source. Transitions for all compounds were evaluated in positive and negative mode using the MassLynx software. With the exception of ZEN, for most of the compounds the [M+H] + ion was the most abundant precursor, which had a highest sensitivity when using the [M-H] - ion as the precursor. For T- 2 and HT-2 toxins, the [M+NH 4 ] + and [M+Na] + ions were used respectively as the precursor ions. Fast polarity switching mode was used to include all the compounds in a single analytical run. Table 1 lists MS/MS parameters obtained during infusions. Table 1. MS/MS parameters for mycotoxin detection by the multiple reaction monitoring (MRM) method. Analyte Retention Window m/z precursor m/z product ions time/min ion (CE in V) a AFB 1 [ 13 C 17 ]-AFB [M+H] [M+H] (24), 241 (30) 301 (30), (30) AFB [M+H] (27), (30) AFG 1 [ 13 C 17 ]-AFG [M+H] [M+H] (27), (30) (42), (42) AFG [M+H] (30), 217 (32) DON [ 13 C 15 ]-DON [M+H] [M+H] (10), 231 (12) 263 (17), 245 (17)

6 FB 1 [ 13 C 34 ]-FB 1 FB 2 [ 13 C 34 ]-FB 2 OTA [ 13 C 20 ]-OTA ZEN [ 13 C 18 ]-ZEN [M+H] [M+H] [M+H] [M+H] [M+H] [M+H] [M-H] [M-H] (40), 334 (40) (40), 374 (40) (38), (38) (40), (40) (25), (30) (30), (30) (24), (30) (25), (25) HT-2 Toxin [M+NH 4 ] (22), 263 (27) T-2 Toxin [M+Na] (22), (27) a Values are given in the order quantifier ion, qualifier ion (in parentheses are the corresponding collision energy (CE) settings in volts). Chromatography was optimised to achieve baseline separation for the compounds including aflatoxins. The best chromatographic results in terms of peak shape, signal intensity, and reproducibility, were achieved with MeOH and Milli-Q water both containing 5 mm NH 4 COOH (ph 3). This acidified mobile phase was used to yield stable retention times and better ionization efficiencies for the fumonisins. The use of Acquity BEH C 18 UHPLC column helped to achieve a short run time and to increase chromatographic resolution. Figure 1 shows the extracted ion MRM chromatogram of a matrix matched calibration sample including all 11 unlabelled mycotoxins and 7 labelled mycotoxins. Analytes were identified by their retention times and by two selected MRM transitions (1 quantifier, 1 qualifier). The gradient elution began with 35% MeOH because DON is poorly retained

7 relative to other analytes. All mycotoxins were detected within 12 min without any coelutions providing rapid determination of multiple mycotoxins. Figure 2. Extracted ion chromatogram (XIC) of a blank maize sample spiked after extraction with labelled and unlabelled mycotoxins. Elution order and ions are as per Table Evaluation of recovery and matrix effects For the evaluation of recovery, blank maize samples were spiked in duplicate with a known amount of unlabelled mycotoxins before extraction. Table 2 lists the absolute recoveries determined using matrix matched standards. Obtained method recoveries for ACN/H 2 O/HCOOH extracted samples ranged between 55% and 122% for AFG 2, ZEN, OTA, AFB 1 and DON whereas enhanced values between 148% and 262% were observed for FB 1, AFG 1, FB 1, T-2 Toxin, HT-2 Toxin and AFB 2. Mean recoveries ranging between 68% and 95% were reported by Wang et al. (2013) in spiked maize samples extracted with ACN/H 2 O/CH 3 COOH prior to Oasis HLB SPE clean-up [7]. MeOH/H 2 O extracted samples had optimal recoveries for AFB 1, AFB 2, FB 2, HT-2 and T-2 Toxins, recoveries were between 81% and 105%. However, poor recoveries (11% %REC 67%) were observed for AFG 2, AFG 1, OTA and ZEN whereas FB 1 had enhanced recovery of 131%. Matrix reference material (Trilogy ) was used to evaluate the accuracy of the method. Enhanced recoveries (%REC >120%) for fumonisins, HT-2 and T-2 Toxins were observed when using the sequential extraction using ACN/H 2 O/HCOOH. MeOH/H 2 O extracted samples resulted in poor recoveries for most analytes, below 65%, with the exception of fumonisins. To further evaluate accuracy of the method, FAPAS reference material was used. Mean recoveries for AFB 1 and DON were 106% and 105% respectively for sequential extraction using ACN/H 2 O/HCOOH. However, low recoveries of 37% and 73% for AFB 1 and DON respectively were observed when using the MeOH/H 2 O mixture. Matrix effects were quantified by evaluating signal suppression or enhancement (SSE) during the ionisation step. The signal suppression or enhancement, expressed as a percentage was quantified as the ratio between the slope of the curve obtained for the matrix matched extracts and the slope of the curve for the standard calibration curve. Matrix matching compensates for any loss/gain due to SSE. However, a challenge remains to always find the best matrix

8 for each sample analysed. Most of the analytes were influenced by presence of matrix effects with AFB 2 and ZEN being the most affected by signal suppression (% SSE 80%). This was observed for both samples extracted with the two different solvent combinations. In contrast, AFG 2 and T-2 Toxin showed strong enhancement (%SSE 120%). Table 2. Method performance parameters for ACN/H 2 O/HCOOH sequential extracted maize samples prior SPE clean-up by UHPLC-SIDA-MS/MS. Mycotoxin Spike concentration (ng/g) Recovery (%) ± RSD (%) (n=2) SSE (%) LOD (ng/g) LOQ (ng/g) AFB ± AFB ± AFG ± AFG ± FB ± FB ± OTA ± ZEN ± DON ± HT-2 Toxin ± T-2 Toxin ± Linearity, Limit of detection and Limit of quantification Calibration curves were prepared by spiking unlabelled and labelled mycotoxin standards into maize extracts at five concentrations: AFB 1, AFG 1, ng/g; AFB 2, AFG 2, ng/g; OTA, ng/g; DON, FB 1, T-2, HT-2 Toxin, ng/g, FB 2, ZEN, ng/g. The coefficient of determination (R 2 ) values were between 0.93 and The limit of detection (LOD) and limit of quantification (LOQ) were determined through regression analysis. The LOD and LOQ ranged between ng/g and between ng/g, respectively. 4. Conclusion An analytical method for determination of multiple mycotoxins in maize using ultrasonic extraction, economical HLB SPE clean-up and UHPLC-MS/MS is reported. The aim of this work was to combine the ultrasonic extraction technique and SPE clean-up for simultaneous

9 extraction and determination of 11 mycotoxins in maize. Sample extraction based on two extraction with ACN/H 2 O/HCOOH steps resulted in good recoveries for most mycotoxins. An SPE clean-up procedure based on hydrophilic-lipophilic based sorbents enabled the simultaneous analysis of mycotoxins with different physicochemical properties. 5. Acknowledgements This work was supported by the Department of Trade and Industry and the National Research Foundation. 6. References 1. P. Songsermsakul and E. Razzazi-Fazeli, Journal of Liquid Chromatography & Related Technologies, 2008, 31, J. F. W. E. C. o. F. A. Meeting and W. H. Organization, Safety evaluation of certain mycotoxins in food, Food & Agriculture Organisation, V. L. Pereira, J. O. Fernandes and S. C. Cunha, Trends in Food Science & Technology. 4. A. Daković, M. Tomašević-Čanović, V. Dondur, G. E. Rottinghaus, V. Medaković and S. Zarić, Colloids and Surfaces B: Biointerfaces, 2005, 46, C. Cavaliere, P. Foglia, C. Guarino, M. Nazzari, R. Samperi and A. Laganà, Rapid Communications in Mass Spectrometry, 2007, 21, V. M. T. Lattanzio, M. Solfrizzo, S. Powers and A. Visconti, Rapid Communications in Mass Spectrometry, 2007, 21, Y. Wang, C. Xiao, J. Guo, Y. Yuan, J. Wang, L. Liu and T. Yue, Journal of Food Science, 2013, 78, M1752-M A. L. Capriotti, G. Caruso, C. Cavaliere, P. Foglia, R. Samperi and A. Laganà, Mass Spectrometry Reviews, 2012, 31, E. Varga, T. Glauner, R. Köppen, K. Mayer, M. Sulyok, R. Schuhmacher, R. Krska and F. Berthiller, Analytical and Bioanalytical Chemistry, 2012, 402, N. Kieu Thi Ngoc and R. Y. U. Dojin, Journal of AOAC International, 2014, 97, C. Li, Y.-L. Wu, T. Yang and W.-G. Huang-Fu, Journal of Chromatographic Science, 2012, 50, W.-J. Kong, S.-Y. Liu, F. Qiu, X.-H. Xiao and M.-H. Yang, Analyst, 2013, 138, V. M. T. Lattanzio, M. Solfrizzo and A. Visconti, Food Additives & Contaminants: Part A, 2008, 25, E. Commission, Official. Journal of the European. Union, 2010, 35, 7-8.