Quantitative and Confirmatory Analysis of Pesticide Residues in Cereal Grains and Legumes by Liquid Chromatography–Quadrupole-Time-of-Flight Mass Spectrometry

For controlling pesticide residues in food and ensuring food safety, multiresidue methods that can monitor a wide range of pesticides in various types of foods are required for regulatory monitoring. In this study, to demonstrate the applicability of liquid chromatography–quadrupole time-of-flight mass spectrometry (LC–QTOF-MS) for quantitative and confirmatory analysis of pesticide residues in cereal grains and legumes, the LC–QTOF-MS method using full-scan acquisition was validated for 151 pesticides in brown rice, soybeans, and peanuts at a spiked level of 0.01 mg/kg. With the exception of 5 out of 151 target pesticides, sufficiently high signal intensities were obtained at 0.005 μg/mL (corresponding to 0.01 mg/kg). Trueness was in the range 70–95%, with intra- and inter-day precisions below 16% and 24%, respectively, with the exception of 7 pesticides in brown rice, 10 pesticides in soybeans, and 9 pesticides in peanuts. No interfering peaks were observed near the retention times of the target pesticides. Furthermore, information on accurate fragment-ion masses obtained by a data-independent acquisition enabled unambiguous confirmation. The results suggest that the LC-QTOF-MS method is suitable for pesticide residues’ analysis of cereal grains and legumes, and can be utilized for regulatory routine analysis.


Introduction
Pesticides are used worldwide to increase crop yields by protecting crops from pests, including insects, rodents, fungi, and weeds; however, the intake of pesticide residues contained in foods may adversely affect human health [1]. To ensure food safety and protect consumer health, international organizations such as the Codex Alimentarius Commission, established by the Food and Agriculture Organization of the United Nations (FAO) and the World Health Organization (WHO), and the European Union (EU), as well as many individual countries, including Japan, have established maximum residue limits (MRLs) to regulate pesticide residue levels in foods. In Japan, MRLs are currently established for various foods with respect to more than 750 agricultural compounds, i.e., pesticides, veterinary drugs, and feed additives. Therefore, the need for multiresidue methods, which detect a wide range of pesticides in various types of foods, is increasing in laboratories concerned with the regulatory monitoring of pesticide residues.
Nowadays, liquid chromatography (LC) and gas chromatography (GC) coupled with triple quadrupole mass spectrometry (MS/MS) operated in selected reaction monitoring (SRM) mode are the most widely used techniques for analyzing pesticide residues in foods. They are highly sensitive and selective, which enables the robust quantification of trace amounts of pesticide residues in complex matrices. In recent years, LC and GC soybeans were ground using a centrifugal mill (Ultra Centrifugal Mill ZM 200; Retsch, Haan, Germany). Peanuts were milled using a laboratory mill (SCM-40A, Shibata, Japan).

Apparatus
LC-QTOF-MS analyses were performed using an Acquity UPLC I-class system (Waters) coupled to a Xevo G2-S QTOF mass spectrometer (Waters). The chromatographic separation was carried out using an Inertsil ODS-4 column (100 × 2.1 mm, 2 µm; GL Sciences). The mobile phases consisted of 5 mmol/L ammonium acetate in water (A) and 5 mmol/L ammonium acetate in methanol (B). The mobile phase was pumped at a flow rate of 0.3 mL/min with the following gradient profile: 5% B followed by increasing B to 95% at 10 min and holding it at this concentration for 3 min, increasing to 100% at 13.01 min and holding for 5 min, and finally, returning to 5% at 18.01 min. The column temperature was set to 40 • C. The injection volume was 3 µL. The retention times of the target pesticides are presented in Table 1.
The QTOF mass spectrometer was operated in resolution mode, providing a resolving power of >30,000 at full width at half maximum (FWHM), at m/z 556.2766. The following MS conditions were used: ionization mode, electrospray ionization in positive mode (ESI(+)); scan range, m/z 50-1000; source temperature, 120 • C; desolvation gas temperature, 450 • C; capillary voltage, 1000 V; cone voltage, 20 V; collision energy, low energy (4 eV) and high energy (ramp from 10 to 40 eV); desolvation gas (nitrogen), 800 L/h; cone gas (nitrogen), 50 L/h; collision gas, argon. Leucine-enkephalin (m/z 556.2766) was used as a reference compound, being introduced from a lock spray probe during analyses. The mass window of ±5 mDa was used for the extraction of chromatograms for each target pesticide. The calculated exact mass and retention time for each pesticide are summarized in Table 1.

Sample Preparation
Samples were prepared according to the official Japanese multiresidue method, namely, "Multi-residue Method I for Agricultural Chemicals by LC-MS (Agricultural Products)," except for the use of a tandem GCB/PSA cartridge instead of a GCB/aminopropylsilyl silica gel (NH 2 ) cartridge for cleanup.
A 10.0 g sample was weighed in a glass tube and water (20 mL) was added; subsequently, it was left to stand for 30 min. Acetonitrile (50 mL) was added to the mixture; then it was homogenized using a homogenizer (Polytron PT 10-35 GT; Kinematica, Lucerne, Switzerland) for 1 min. The homogenate was filtered with suction, and then the residue was rehomogenized with acetonitrile (20 mL) before being filtered with suction. The filtrates were combined, and the resulting volume was adjusted to 100 mL by the addition of acetonitrile.
A 20 mL aliquot of the extract was added to a 50 mL polypropylene (PP) centrifuge tube containing sodium chloride (10 g) and phosphate buffer (pH 7.0, 0.5 mol/L). The mixture was shaken for 5 min by a shaker (SR-2w; Taitec, Saitama, Japan) and centrifuged for 5 min at 3000 rpm (Centrifuge 8100, Kubota, Japan). The resultant acetonitrile layer was loaded onto an ODS cartridge, which was preconditioned with acetonitrile (10 mL), and then eluted with acetonitrile (5 mL). The resultant eluates were combined and concentrated to approximately 0.5 mL by a rotary evaporator (NVC-2100/N-1000, Eyela, Tokyo, Japan) at <40 • C; it was then dried by evaporation under a nitrogen stream. The residue was redissolved in acetonitrile/toluene (3:1, v/v, 2 mL) and loaded onto a GCB/PSA cartridge, which was preconditioned with acetonitrile/toluene (3:1, v/v, 10 mL) and then eluted with acetonitrile/toluene (3:1, v/v, 20 mL). The eluate was concentrated to approximately 0.5 mL by a rotary evaporator at <40 • C and evaporated to dryness under a nitrogen stream; finally, the resultant residue was redissolved in methanol (4 mL) prior to LC-QTOF-MS analysis.

Method Validation
The LC-QTOF-MS method was validated using a nested experimental design for brown rice, soybeans, and peanuts. Samples were spiked in duplicate at a level of 0.01 mg/kg, and the recovery experiments were repeated on five different days. To prepare the spiked samples, a 1 mL aliquot of the 0.1 µg/mL mixed standard solution was added to 10.0 g of sample, and the mixture was allowed to stand for 30 min before proceeding with the subsequent sample preparation steps. The quantification was carried out using six-point calibration curves with solvent-based standard solutions prepared in methanol. The concentrations of the standard solutions used to construct the calibration curves, to allow quantification, were 0.00125, 0.0025, 0.00375, 0.005, 0.00625, and 0.0075 µg/mL. The linearity of each calibration curve over a wider range was examined in the range of 0.002-0.1 µg/mL.
Matrix-matched standards were prepared by evaporating a 100 µL aliquot of blank solution under a nitrogen stream and then redissolving it in 100 µL of the mixed standard solution in methanol. Matrix effects were evaluated by comparing peak areas of the matrixmatched standards with standards in solvents as follows: average peak area (n = 5) of matrix-matched standard/average peak area (n = 5) of standard in solvent.

Optimization of LC-QTOF-MS Conditions
A total of 151 LC-amenable pesticides, which had molecular weights from 189 to 746, were selected as target pesticides for this study. Because most of the target pesticides produced high-intensity signals under positive-mode operation of the instrument (c.f., negative-mode operation), and since the instrument used in this study was unable to simultaneously operate in both the positive and negative modes, LC-QTOF-MS analyses were carried out only in the positive mode, using the MS parameters optimized in a previous study [14]. The calculated exact mass of each pesticide is presented in  4 ] + and none of the compounds were seen at their highest intensity for [M+Na] + . The mass window for extracting the chromatograms of each pesticide was optimized by comparing the repeatability of the peak areas of the target compounds, obtained by replicate analyses (n = 5, 0.01 µg/mL) for the extraction of mass windows of ±2.5, ±5, and ±10 mDa. It should be noted that, in general, a narrow mass window for the extraction of chromatograms will result in low background noise and allow the discrimination of coeluting matrix components. This will increase sensitivity and selectivity; however, the use of a disproportionately narrow window will result in peak shape deterioration and low repeatability. Hence, mass windows of ±5 and ±10 mDa resulted in relative standard deviations (RSDs) of <5% for all the target pesticides; whereas the RSD values for 10 pesticides were >5% with a mass window of ±2.5 mDa. In addition, narrow mass windows produced higher signal-to-noise (S/N) ratios. Therefore, considering these results, the mass window was set to ±5 mDa, as a trade-off between S/N and peak area repeatability.

Method Validation
As mentioned earlier, the samples were prepared according to the official Japanese multiresidue method "Multi-residue Method I for Agricultural Chemicals by LC-MS (Agricultural Products)" prior to analysis by LC-QTOF-MS, except for the modification in the cleanup step. A tandem GCB/PSA cartridge was used instead of a tandem GCB/NH 2 cartridge for cleanup because the PSA sorbent can more effectively remove acidic matrix components, such as organic acids and fatty acids, compared to a NH 2 sorbent. The LC-QTOF-MS method was validated in terms of linearity, matrix effect, trueness, intra-and inter-day precisions, and selectivity for detection of the spiking at a concentration level of 0.01 mg/kg with 151 pesticides of brown rice, soybeans, and peanuts. Quantification was carried out using solvent-based calibration curves in this study.
Injecting a standard solution of 0.005 µg/mL, which corresponds to 0.01 mg/kg, five pesticides, i.e., hexaconazole, isoprocarb, methidathion, pentoxazone, and quizalofop ethyl, exhibited insufficient sensitivities, i.e., S/N < 10. Among them, the low sensitivity of quizalofop ethyl could be a consequence of a high background noise level due to polysiloxane contamination, which has a similar calculated exact mass (m/z 373.0981, [C 10 H 31 Si 4 30 SiO 5 ] + ). Therefore, the validation method was continued for 146, of the original 151, pesticides and achieved the required sensitivity of 0.005 µg/mL (corresponding to 0.01 mg/kg). Furthermore, because ferimzone and tricyclazole were detected at concentrations of 0.02 mg/kg and <0.01 mg/kg, respectively, in the brown rice sample used for the method validation in this study, ferimzone and tricyclazole were also excluded from the target compounds for method validation in brown rice. It should be noted that the residue levels of ferimzone and tricyclazole detected in brown rice were below the MRLs (2 ppm and 3 ppm, respectively) established in Japan.

Confirmation
For discriminating analytes from coeluting matrix components in complex foods at low concentrations, information on the exact mass and retention times of the target analytes may not be sufficient, even when using LC-Orbitrap-MS, which, compared to LC-TOF-MS, provides a high resolving power [19]. The EU guidelines [22] state that two ions, preferably a molecular adduct and at least one fragment ion, are required for accurate mass measurement by high-resolution MS. Hybrid HR-MS, such as QTOF-MS and QOrbitrap-MS, provide fragment-ion information via data dependent acquisition (DDA) and/or DIA [2,18]. In DDA, precursor ions are sequentially selected from full scans based on user-selected criteria (e.g., minimal intensity threshold, m/z values). In contrast, in DIA, all ionized compounds are subjected to fragmentation, and thus, DIA provides information regarding the fragment ions derived from all ions. In a previous study into pesticide residue analyses in vegetables and fruits using LC-QTOF-MS [14], we demonstrated DIA using the MS E technique (Waters) [24], which provided full-scan data on both the molecular adduct (at low collision energy) and fragment (at high collision energy) ions in a single run, without selecting the precursor ion. In the study reported herein, we further The results of method validation revealed that LOQs, defined as the lowest concentration that can be quantified with satisfactory trueness values and precision, are 0.01 mg/kg for most pesticides (Table S1). MRLs of the target pesticides in brown rice, soybeans, and peanuts established in Japan are shown in Table S1. For pesticide/food combinations whose MRLs are not established, a uniform limit of 0.01 mg/kg is applied. As can be seen, MRLs are ≥0.01 mg/kg. Therefore, the proposed method exhibits sufficient sensitivity for regulatory purpose analysis.

Confirmation
For discriminating analytes from coeluting matrix components in complex foods at low concentrations, information on the exact mass and retention times of the target analytes may not be sufficient, even when using LC-Orbitrap-MS, which, compared to LC-TOF-MS, provides a high resolving power [19]. The EU guidelines [22] state that two ions, preferably a molecular adduct and at least one fragment ion, are required for accurate mass measurement by high-resolution MS. Hybrid HR-MS, such as QTOF-MS and QOrbitrap-MS, provide fragment-ion information via data dependent acquisition (DDA) and/or DIA [2,18]. In DDA, precursor ions are sequentially selected from full scans based on user-selected criteria (e.g., minimal intensity threshold, m/z values). In contrast, in DIA, all ionized compounds are subjected to fragmentation, and thus, DIA provides information regarding the fragment ions derived from all ions. In a previous study into pesticide residue analyses in vegetables and fruits using LC-QTOF-MS [14], we demonstrated DIA using the MS E technique (Waters) [24], which provided full-scan data on both the molecular adduct (at low collision energy) and fragment (at high collision energy) ions in a single run, without selecting the precursor ion. In the study reported herein, we further applied the MS E technique to brown rice, soybeans, and peanuts spiked at a level of 0.01 mg/kg. Figure 2 shows extracted ion chromatograms of molecular adduct and fragment ions from the soybean samples; Table 1 shows that the fragment ions could be detected at 0.005 µg/mL in the presence of the matrices. Among the 146 target pesticides, for 126 pesticides, we were able to detect one or more fragment ions; for 84 pesticides, we were able to detect one or more isotopic ions, and for 134 pesticides, we were able to detect one or more fragment ions and/or isotopic ions. Figure 3 shows extracted ion chromatograms of incurred ferimzone residue found in the rice sample used for validation in this study. As can be seen, the [M+H] + (m/z 255.1604) for ferimzone together with its two fragments ions, i.e., [C 9 H 10 N] + (m/z 132.0808) and [C 6 H 10 N 3 ] + (m/z 124.0869), were clearly detected, suggesting that the MS E technique could be a useful tool for obtaining fragment ion information for confirmation purposes. However, because the sensitivities of the fragment ion peaks for several pesticides were low, more sensitive methods, such as LC-MS/MS in SRM mode, may be required for confirmation, especially for the pesticides that were shown to be detected with low sensitivities using the LC-QTOF-MS technique described in this study.
applied the MS E technique to brown rice, soybeans, and peanuts spiked at a level of 0.01 mg/kg. Figure 2 shows extracted ion chromatograms of molecular adduct and fragment ions from the soybean samples; Table 1 shows that the fragment ions could be detected at 0.005 μg/mL in the presence of the matrices. Among the 146 target pesticides, for 126 pesticides, we were able to detect one or more fragment ions; for 84 pesticides, we were able to detect one or more isotopic ions, and for 134 pesticides, we were able to detect one or more fragment ions and/or isotopic ions. Figure 3 shows extracted ion chromatograms of incurred ferimzone residue found in the rice sample used for validation in this study. As can be seen, the [M+H] + (m/z 255.1604) for ferimzone together with its two fragments ions, i.e., [C9H10N] + (m/z 132.0808) and [C6H10N3] + (m/z 124.0869), were clearly detected, suggesting that the MS E technique could be a useful tool for obtaining fragment ion information for confirmation purposes. However, because the sensitivities of the fragment ion peaks for several pesticides were low, more sensitive methods, such as LC-MS/MS in SRM mode, may be required for confirmation, especially for the pesticides that were shown to be detected with low sensitivities using the LC-QTOF-MS technique described in this study.

Conclusions
In this study, the multiresidue method using LC-QTOF-MS in full-scan acquisition mode was validated for the determination of 151 pesticides in cereal grains and legumes. Sufficiently high sensitivities were achieved at 0.005 μg/mL (corresponding to 0.01 mg/kg), with the exception of 5 of the 151 pesticides. Excellent results were obtained in terms of trueness, intra-and inter-day precision, and selectivity for most of the target pesticides at 0.01 mg/kg. The results revealed that the LC-QTOF-MS method offers reliable quantitative analysis of pesticide residues in cereal grains and legumes. In addition, we demonstrated the usefulness of the MS E technique for obtaining information on fragment ions for confirmation. Although we were unable to detect several parent and fragment ions owing to low S/N at 0.01 mg/kg, the LC-QTOF-MS method was shown to be suitable for regulatory-purpose analysis for most of the target pesticides.
Supplementary Materials: The following are available online at www.mdpi.com/xxx/s1, Table S1: MRLs set in Japan and LOQs of the target pesticides. Table S2

Conclusions
In this study, the multiresidue method using LC-QTOF-MS in full-scan acquisition mode was validated for the determination of 151 pesticides in cereal grains and legumes. Sufficiently high sensitivities were achieved at 0.005 µg/mL (corresponding to 0.01 mg/kg), with the exception of 5 of the 151 pesticides. Excellent results were obtained in terms of trueness, intra-and inter-day precision, and selectivity for most of the target pesticides at 0.01 mg/kg. The results revealed that the LC-QTOF-MS method offers reliable quantitative analysis of pesticide residues in cereal grains and legumes. In addition, we demonstrated the usefulness of the MS E technique for obtaining information on fragment ions for confirmation. Although we were unable to detect several parent and fragment ions owing to low S/N at 0.01 mg/kg, the LC-QTOF-MS method was shown to be suitable for regulatorypurpose analysis for most of the target pesticides.

Data Availability Statement:
The data presented in this study are available in the article or supplementary material.

Conflicts of Interest:
The authors declare no conflict of interest.