Carlos Bueno1, Aidan Harrison2, Gabriel Ruiz3, Lupi Martínez3 Jianru Stahl-Zeng4
SCIEX, Spain1; SCIEX UK2; Eurofins Ecosur3; SCIEX Germany4
Quantitation of mycotoxins and tropane alkaloids in food using SCIEX 7500+ system and QTRAP technology
Abstract
This technical note demonstrates a method for analysis of 23 mycotoxins and 2 tropane alkaloids in four food matrices achieving limits of quantitation (LOQs) 10-250 fold lower than the European Union (EU) Maximum Residue Limits (MRL), see Figure 1. Using the SCIEX 7500+ system and QuEChERS extraction, LOQs as low as 0.01 µg/kg were shown using a standard MRM acquisition method. These LOQs surpassed the MRLs according to regulation EU 2023/9151 and EU 2024/10382 and recommendation EU 2022/5533 . In addition, MRM3 scans using the 7500+ QTRAP technology selectively removed endogenous matrix interferences for some mycotoxin compounds such as Aflatoxin B2. The enhanced selectivity of the MRM3 scans resulted in LOQ values as low as 0.0025 µg/kg
Figure 1: Extracted ion chromatograms (XICs) of four mycotoxins in a matrix blank (first row) and obtained LOQ (second row) in pre-spiked baby food. Quantifer ion is shown in blue, first qualifier ion is shown in pink. Using the SCIEX 7500+ QTRAP system, LOQ values of 0.01 µg/kg were achieved for Aflatoxin, 0.0625 µg/kg for Aflatoxin B2, 0.2 µg/kg for HT-2 toxin and 0.03 µg/kg for Ochratoxin A. These values are 10-25 times lower than the MRL values required in regulation EC 2023/915 for these specific mycotoxins in baby food matrix.
Methods
Standards and reagents: Standards were purchased from Sigma Alrich, Biopure and Dr Ehrenstorfer.
Sample preparation: A simplified QuEChERS method was used for all food matrices tested. Briefly, a 5 g sample of the dried foodstuff or wine/juice was weighed into a 50 mL plastic centrifuge tube and spiked with a mixed standard. If dried foodstuff was tested, 10 mL of water was added, and the mixture was left for 10 minutes to hydrate the foodstuff sample. 10 mL of acetonitrile with 2% formic acid (v/v) was added, followed by mixing for 2 minutes. The QuEChERS sachet of magnesium sulfate and sodium chloride salt was added, and the sample was vortexed before being centrifuged for 4 minutes at 4000 rpm. 2 mL of supernatant was removed and transferred to a 15 mL plastic centrifuge tube. Magnesium sulfate and C18 powder was added, and the tube agitated before centrifugation for 10 minutes at 9500 rpm. 1 mL of the supernatant was filtered using a 0.22 µm nylon filter into an LC-MS vial for injection.
Chromatography: Chromatographic separation (Figure 2) was performed using a SCIEX Exion AE LC system with a Phenomenex Luna Omega Polar C18 (2.1 x 100 mm, 1.6 µm, 100 Å) column. A 15-minute gradient was run at a flow rate of 0.3 mL/min using ultra-pure water with 0.02% (v/v) formic acid, 2mM ammonium formate and 0.05mM ammonium fluoride as mobile phase A, and methanol with 0.02% (v/v) formic acid, 2mM ammonium formate and 0.05mM ammonium fluoride as mobile phase B (Table 1). The column temperature was maintained at 40°C and the injection volume was 1.5 μL. A mixture of 3:3:3:1 (v/v/v/v) acetonitrile/methanol/water/isopropyl alcohol with 0.1% (v/v) formic acid was used as a needle wash solvent.
Table 1: Chromatographic gradient for analysis of mycotoxins in foodstuffs.
Mass spectrometry: Samples were analyzed using the SCIEX 7500+ QTRAP system operated in electrospray ionization (ESI) mode with polarity switching. Data was acquired using the Scheduled MRM algorithm with at least two selective MRM transitions per analyte. Q0D optimization was performed and operated in simple mode for the analysis. MRM3 experiments used a 1000 Da/s scan rate and 25 ms excitation time, with dynamic fill time.
Data processing: Processing was performed using SCIEX OS software (v 3.4). The peak-to-peak signal-to-noise and MQ4 integration algorithm were used
Figure 2: Extracted ion chromatogram (XIC) of all compounds overlaid. The XIC above highlights the chromatographic separation achieved in a wine matrix. All compounds eluted between 4 and 10 minutes.
Quantitative performance in various food matrices
The LOQ values reported here (Table 2) were below the required EU MRL for all four matrices tested (baby food, almond, grape juice and wine), with the LOQ defined by S/N of the quantifier ion >10, as well as an ion ratio of ±20% of the average, and accuracy 80-120%. These LOQ values ranged from 0.01 to 10 µg/kg in baby food, the most challenging of the four matrices tested.
Linearity (1/x weighting, r2 > 0.99) was achieved for the calibration curves of all mycotoxins in all four matrices without the need for labeled internal standards (Figure 3 shows the calibration curve of Aflatoxin B1 in baby food) with a linear dynamic range of 2-3 orders of magnitude for most compounds in all four matrices.
Average accuracy and precision were assessed for 5 replicate injections of the extracted baby food LOQ standard and were within the acceptable validation requirements (accuracy ±20% and peak area %CV < 15%). Figure 4 shows the blank and five LOQ replicates for Aflatoxin B1 (0.01 µg/kg) in baby food.
As baby food has the lowest EU MRL requirements for most compounds, Table 3 summarizes the precision, linearity and accuracy data for the calibration curve prepared in this matrix.
Figure 3: Calibration curve for Aflatoxin B1 in baby food. The figure above shows a linear range of 0.01 – 5 µg/kg for Aflatoxin B1 highlighting an r2 value of 0.999
Table 2: LOQ values achieved for mycotoxins in 4 food matrices using standard MRM workflow, compared with the lowest MRL value required by Regulation EU 2023/9151, EU 2024/10382 and Recommendation EU 2022/5533
Figure 4: Extracted ion chromatograms (XICs) of the matrix blank and 5 replicates of the Ochratoxin A LOQ spike (0.03 µg/kg) in baby food.. Quantifier ion is shown in blue, first qualifier ion is shown in pink.
Table 3. Calibration range, LOQ in baby food (or almond where noted), peak-to-peak (p2p) S/N, peak area %CV at the LOQ (n=5) and r2 value
Achieve EU MRL without lyophilization
Using a large injection volume is one strategy to meet low MRL values such as those for the EU mycotoxins levels in food. However, large injection volumes can also lead to higher levels of matrix interference during the analysis of foodstuffs, impacting overall method sensitivity. Using the 7500+ system, a 1.5 µL injection volume was sufficient to achieve LOQ values that were 10-250 times less than the lowest required MRL for a variety of foodstuffs. Table 4 shows the LOQ values achieved, alongside the minimum baby food EU limits.1,2,3 Due to the small injection volume, very low carry-over was observed. Further, an injection of Fumonisin B1 at the ULOQ of 100 µg/kg resulted in just 0.08% carry-over in the following blank injection.
This high-sensitivity quantitation of the regulated mycotoxins at levels below the required MRL values may allow for sample preparation that does not require a lengthy lyophilization step. This is because if the MRL applies to the dry matter content (as is the case with baby food1,2,3), and the LOQ can be shown to be sufficiently lower than the required MRL, the method can be validated on the wet-weight if shown to meet the criteria for the equivalent weight of the dry matter content. As the LOQ values achieved were 10-250 fold lower than the MRL requirements for baby food (Table 4), if the average ratio percent between dry and wet weight is <10%, a method without lyophilization is likely possible for all mycotoxins and plant toxins described in this technical note.
Table 4: LOQ value achieved for baby food compared with minimum baby food EU limits (MRL)
MRM3 scan for interference removal
Certain mycotoxins such as Aflatoxin B2, Ochratoxin A and Scopolamine typically show an interference peak in matrix blank samples in baby food and certain other food matrices, even when using small injection volumes. The presence of these interference peaks results in a high LOQ value., as compared to samples with a clean background. MRM3 scans allow for greater selectivity through the fragmentation of a secondary ion into a tertiary ion. This second-generation fragment can be used for quantification, with lower LOQs typically achieved as compared to MRM (Table 5). Figure 5 shows a comparison between the low-level quantitation of Aflatoxin B2 by MRM and MRM3 scan.
The SCIEX 7500+ system allows for ultra-fast scanning with pause times as low as 2 ms, and dwell times of 1 ms when using sMRM and polarity switching.6 This ultra-fast scanning means that very large analyte panel methods – which monitor a high number of MRMs – can possess low cycle times and sufficient data points across the chromatographic peaks. This ensures good data quality while maintaining sensitivity. Therefore, a single MRM3 experiment can be implemented with a scan time of only 60 ms so that up to two MRM3 scans can be acquired alongside the full mycotoxin screen using positive and negative MRMs without a significant increase in cycle time and preserving ultra-trace level sensitivity.
Table 5: LOQ value achieved for selected mycotoxins and plant toxins in baby food matrix using MRM3
Figure 5: Comparison of MRM and MRM3 scan XICs of blank and low-level standards for Aflatoxin B2 in baby food matrix. The XICs show that the MRM scan for Aflatoxin B2 in baby food contains an interference which is seen in the blank, and results in accurate quantitation not being possible below 0.0625 µg/kg (see Table 3). The MRM3 scan is more selective, and no peak in the blank is seen, meaning that Aflatoxin B2 can be quantified as low as 0.0025 µg/kg.