Abstract
Here, a single sample preparation was used to analyze hundreds of pesticides, veterinary drugs and mycotoxins in different food and feed matrices using the SCIEX 7500 system. A large panel method was used to analyze all contaminants in a single run using time-scheduled acquisition. The high instrument sensitivity permitted the use of a dilution strategy to mitigate matrix effects while maintaining the required quantification limits. This work explores the tradeoffs between dilution and sensitivity requirements with the goal of simplifying sample preparation procedures and streamlining laboratory workflows.
Introduction
Regulations surrounding testing for contaminants, such as pesticides, veterinary drugs and mycotoxins, in foodstuffs are generally determined by analyte-specific guidelines. One of the main challenges of this approach is that food matrices vary widely in their chemical composition and therefore are prone to matrix effects and interferences. For this reason, sample preparation and analysis methods are often tailored to a particular food matrix. This can significantly impact laboratory productivity, throughput and overhead costs because each unique method must be learned by technicians, supplies must be purchased for each preparation and analysis methods must be run independently for each analyte type. Further, as analyte lists continue to grow, their chemical and physical properties are becoming more similar to those of the matrix, which further complicates sample cleanup.
In an ideal scenario, a single sample preparation and analysis method would be employed for all contaminants in all food and feed matrices that must be analyzed by a particular laboratory. This is often not possible due to the matrix effects or sensitivity requirements for each panel.
Key features of the SCIEX 7500 system for contaminant analysis
- High instrument sensitivity enables a simple sample preparation to be used that is based on a new QuEChERSER1,2 procedure with an additional dilution step
- Dilution reduces matrix interferences and makes the procedure usable in many different food and feed matrices
- High sensitivity of the SCIEX 7500 system allows further dilution steps to be used while meeting sensitivity requirements
- Analytes with widely different chemistries can be analyzed together in a single large panel method
Methods
Sample preparation: Egg, hemp pellet, hemp powder, hemp oil and protein powder were prepared using the QuEChERSER1,2 protocol, a new version of the widely used QuEChERS3 protocol. Samples were chosen to mimic high protein food and animal feed products. In this protocol, sample homogenization was followed by extraction with 4:1, acetonitrile/water. After centrifugation, an aliquot was either evaporated to dryness and reconstituted in water or only diluted with water. This helps retain the highly polar analytes that are often part of LC-MS/MS panels. No further cleanup was employed and typically 8-10 µL of sample was directly injected for analysis. To mimic different dilution factors, the injection volumes were varied between 1µL, 2µL, 4µL and 8µL to inject the amount of sample expected from 8-fold, 4-fold, 2-fold and no dilution conditions, respectively. Matrix-matched standards were also used for analysis.
Chromatography: An ExionLC AD system was used with a Waters HSS T3 (1.8 µm, 2.1 mm x 100 mm) column at a flow rate of 500 µL/min. The column temperature was kept at 40°C. Mobile phase A was 5mM ammonium formate in water with 0.1% formic acid. Mobile phase B was 5mM ammonium formate in 50:50, methanol:acetonitrile with 0.1% formic acid. Initial conditions were 0% B held for 0.5 min, increased to 10 B%, ramped to 80% B over 10 min, increased to 100% B in 0.5 min and finally held for 1.5 min before returning to initial conditions.
Mass spectrometry: A SCIEX 7500 system equipped with the OptiFlow Pro ion source (ESI) was used for analysis. Both positive and negative ion modes were combined in a single time-scheduled MRM method containing nearly 1000 transitions, using the Scheduled MRM algorithm.5
Data processing:SCIEX OS software was used for data analysis and processing.
Leveraging sensitivity to simplify workflows
Because it is “Qu ick, E asy, Ch eap, E ffective, R ugged and S afe,” the QuEChERS protocol has been widely adopted for many applications and by many analytical chemists.3 Recently, a new version of this protocol, called QuEChERS ER (E fficient and R obust), was introduced.1,2 This modified protocol can efficiently extract more analytes from a wider variety of sample types than the original QuEChERS procedure. However, suppression and interferences from residual matrix components can still be problematic, depending upon the analyte tested and its originating foodstuff.
The SCIEX 7500 system was developed to enable new levels of quantitative sensitivity across a large suite of sample types and workflows.4 An increased orifice size and the newly redesigned dual stage D Jet ion guide increase ion sampling and enhance ion capturing and focusing. Additionally, the OptiFlow Pro ion source with E Lens probe enables more efficient ionization and desolvation, in addition to improved ion sampling in a simple plug-and-play source without the need for any source optimization.
The improved sensitivity of the SCIEX 7500 system makes it possible to analyze many compounds and achieve lower limits of quantification (LOQs). The maximum residue limits for most contaminants have remained the same over time and the SCIEX 7500 system is capable of detecting and quantifying contaminants within these limits. The sensitivity of the instrument therefore permits using dilution an effective strategy to mitigate matrix effects that can suppress or interfere with analyte signals and accurate quantification. Additionally, the very high sensitivity of the SCIEX 7500 system reduces the need to optimize the chromatography, source conditions and data acquisition parameters for specific assays. Instead, using this system, a single large panel LC-MS/MS method can be used to analyze all contaminants simultaneously.
A universal large panel method
A fast, 15-minute LC-MS/MS run was used for the analysis of approximately 400 contaminants that included veterinary drugs, pesticides and mycotoxins. Many analytes were analyzed with 2 to 3 transitions and nearly 1000 MRM transitions were scheduled for acquisition during the LC-MS/MS run (Figure 1). This highly multiplexed method resulted in as many as 80 transitions that were analyzed concurrently. To ensure enough data points were collected across each peak, cycle times of approximately 500 ms were used (Figure 2). The high ion flux on the SCIEX 7500 system shortened dwell times without significantly affecting signal-to-noise and enabled quantitative accuracy and precision.
Accuracy at the 1X standard level
Figure 3 shows the quantitative accuracy of the large panel method for different compound classes of 1X standards at all injection volumes. Here, the 1X standard concentrations corresponded to the most common MRLs for each compound class and pesticide standards were prepared at 10 ng/g, mycotoxins at 1 ng/g and veterinary drugs between 1-10 ng/g.
As shown in Figure 3, excellent quantitative performance was achieved for pesticides, such as fungicides, herbicides and insecticides, with a median distribution near 100% accuracy and high precision. Slightly higher variability was observed for mycotoxins, as the 1X standard for this compound class was 10x less concentrated. Similarly, veterinary drugs displayed slightly higher variability as many of these standards were prepared at lower concentrations. For most compounds, the peaks showed excellent signal-to-noise even at the 1 µL injection level. Thus, some sensitivity could be sacrificed by tailoring the source conditions to the more challenging compounds, if desired.
Reduced matrix effects with lower injection volumes
This method provided enough sensitivity to quantify standards at the lowest injection volumes. Therefore, the impact of decreasing injection volumes on minimizing matrix effects was evaluated.
Figure 4 shows the influence of different injection volumes on analyte response by plotting the peak area ratio of analytes of interest to solvent standards prepared without evaporation. An area ratio of 1 indicates that the analyte response is equivalent to the response of the solvent standards prepared without evaporation. The solvent standards with evaporation, presented in the top panel of Figure 4, reveal that the evaporation step caused significant losses without matrix present. This phenomenon is commonly observed during evaporation steps when there is nothing in the vial for moderately volatile analytes to bind to, therefore resulting in volatilization-induced losses. As expected, the equivalent responses across different injection volumes (red arrow) indicate that there was no effect of injection volume on signal suppression in the absence of matrix.
In contrast, when matrix was present, the degree of matrix effect depended on the sample injection volume. The egg matrix data presented in the top panel of Figure 4 reveal that decreasing the injection volume from 8 µL to 1 µL decreased median suppression by 50%, with approximately 40% suppression observed for 8 µL injections and 20% suppression for 1 µL injections (red arrows). Decreased matrix suppression was quantified by measuring the increase in analyte signal relative to the solvent standard.
The bottom panel of Figure 4 demonstrates that lowering the injection volume also reduced matrix effects in the hemp samples. The response ratios for hemp powder, hemp pellets, unprocessed hemp plants and a commercially available hemp protein powder showed increased area ratios with decreasing injection volume. The reduction in matrix effects with decreasing injection volume was more pronounced in the hemp products than the egg matrix conditions shown in the top panel of Figure 4. This result demonstrates greater improvement in signal using lower injection volumes for challenging sample types. Minimizing injection volumes reduces the on-column matrix load, thereby reducing the impact of contaminants and suppressants by providing a more “solvent-like” matrix for analytes.
While matrix effects can be mitigated with the use of internal standards, this practice would be prohibitively expensive for the large panel method used here, as it would require a mass-labeled internal standard for each of the nearly 400 analytes. A representative subset of internal standards could be used, however, care must be taken to ensure that the internal standards behave like their designated analytes. Alternatively, by minimizing matrix effects upfront by using a dilution strategy, the reliance on these surrogates can be minimized.
Analyte loss due to evaporation during sample preparation
As expected, the evaporation step did not reduce matrix effects. This is shown in the comparison between the egg matrix datasets prepared with and without evaporation, presented in the top panel of Figure 4. Instead, the evaporation step tended to increase variability and, in some cases, produced further losses.
Figure 5 shows the effects of evaporation for veterinary drugs in 1 µL injections. Solvent extracted samples with evaporation show highly variable recovery and a distinct split in the data. This indicates that some analytes performed better and some performed worse, as compounds could be lost to either volatilization or lack of re-solubilization into aqueous solvent. Extraction from egg matrix with evaporation also increased the variability for the observed area ratios. In contrast, extraction from matrix without an evaporation step resulted in a narrow distribution centered around 0% suppression and nearly complete extraction, with little deviance from the solvent standards.
Only a small subset of drugs (~2%) showed reduced signal intensity following extraction from egg matrix without evaporation specifically the thyreostats, tranquilizers and triphenylmethane dyes (Figure 6). Most drug classes showed excellent recovery with nearly 0% suppression and nearly 0 matrix effects.
Conclusions
A sensitive, streamlined workflow for the analysis of approximately 400 contaminants was demonstrated. Example chromatograms for the 1 µL injection egg matrix spikes without evaporation are shown in Figure 7. The sample preparation method used was based on the QuEChERSER protocol to provide efficient extraction for a wide variety of compounds from multiple matrices. No reagents other than water and acetonitrile were needed and LC-MS/MS data acquisition was performed using a single, large panel, time-scheduled MRM method.
The typical evaporation step was often unnecessary or detrimental to the procedure, resulting in losses and more variability. Good retention of early eluting compounds was observed with up to 25% organic in the vial.
To minimize matrix effects, the high sensitivity of the SCIEX 7500 system was leveraged using a dilution strategy. Dilution resulted in less suppression and improved quantitative accuracy and precision by providing more consistent ionization efficiency from a more “solvent-like” matrix. This also resulted in less reliance on internal standards for correction, saving costs. Dilution resulted in additional benefits, such as reduced sample consumption, and is expected to result in extended column lifetimes and improved instrument uptime.
The workflow presented here improves laboratory efficiency and productivity by minimizing the need for different sample preparation and analysis methods for different classes of samples and analytes. This workflow could potentially be automated to further enhance laboratory productivity.
References
- Michlig, M., Lehotay, S., Lightfield, A.R., Beldomenico, H., Repetti, M.R. (2021) “Validation of a high-throughput method for analysis of pesticide residues in hemp and hemp products”, J. Chromatogr. A., 1645(24) 462097.
- Monteiro, S.H., Lehotay, S., Sapozhnikova, Y., Ninga, E., Lightfield, A.R. (2021) “High-Throughput Mega-Method for the Analysis of Pesticides, Veterinary Drugs, and Environmental Contaminants by Ultra-High-Performance Liquid Chromatography-Tandem Mass Spectrometry Robotic Mini-Solid-Phase Extraction Cleanup + Low-Pressure Chromatography-Tandem Mass Spectrometry, Part 1: Beef”, J. Agric. Food Chem. 69, 1159-1168.
- Anastassiades, M., Lehotay, S., Stajnbaher, D., Schenck, F.J. (2003) "Fast and easy multi-residue method employing acetonitrile extraction/partitioning" and "Dispersive solid-phase extraction for the determination of pesticide residues in produce", J. AOAC Int. 86(2): 412–431.
- Enabling new levels of quantification. SCIEX technical note, RUO-MKT-02-11886-A.
- The Scheduled MRM algorithm enables intelligent use of retention time during Multiple Reaction Monitoring. SCIEX technical note, RUO-MKT-02-8539-A.