512  Schaner & Hickes: Journal of AOAC International Vol. 98, No. 2, 2015 RESIDUES AND TRACE ELEMENTS

Quantitative Analysis of Paraquat in Vegetation by Stable Isotope Dilution and Liquid Chromatography/Electrospray Ionization-Mass Spectrometry Angela Schaner and Heidi Hickes1

Montana Department of Agriculture, Montana State University, McCall Hall, Bozeman, MT 59717

The method presented is a suitable approach for routine testing of paraquat in difficult sample types, such as winter wheat and alfalfa plant tissue, typically found with accidental spray drift. Hydrophilic interaction chromatography and ultra-performance LC is utilized with tandem quadrupole MS in the positive electrospray ionization mode. Three precursor-product ion transitions are measured in the multiple reaction monitoring mode, and paraquat d8 is added as an internal standard at the beginning of the extraction procedure to correct for losses in recovery and/or matrix effects in instrument response. A 5 g portion is digested with 6 M HCl in a 100°C water bath for 1 h. An aliquot is removed and adjusted to pH 7–8 prior to loading on a mixed mode weak cation-exchange SPE cartridge, and paraquat is eluted with formic acid– acetonitrile (10 + 90, v/v). Average recoveries of paraquat fortified at 0.020–0.080 ppm in winter wheat and alfalfa ranged from 80 to 114% (RSD 12–30%). Result data from naturally incurred paraquat (0.027–0.51 ppm) in composite garden plants, potato leaves, tree leaves, and alfalfa are presented. The LOQ is 0.020 ppm.

P

araquat is one of the most widely used herbicides in the world. It is used in approximately 90 countries to control a wide range of weeds and grasses in more than 100 crops and is effective in all growing climates. It has proved very useful to farmers using low-tillage methods because of its deactivation upon contact with soil, and it is has become a popular choice as an alternative broad spectrum weed control in situations where crops have become tolerant to glyphosate (1). First manufactured and sold by Imperial Chemical Industries in 1962, paraquat continues to be a force in agriculture as no other herbicide works in the same manner or as effectively (1). Whenever accidental drift or exposure of paraquat is suspected it is imperative to have sound, reliable testing methods in place. However, paraquat’s unique chemistry has hindered advances in method design, and today there are few laboratories with the capability to offer dependable routine testing for investigative samples. Determining optimum conditions for a complete extraction of paraquat has been one of the biggest challenges as it binds Received February 21, 2014. Accepted by AK March 18, 2014. 1   Corresponding author’s e-mail: [email protected] DOI: 10.5740/jaoacint.14-043

strongly to glass, soil, and plant tissue. Past methods used a harsh acidic extraction/digestion with refluxing or heat to fully extract paraquat from plant tissue (2–6). Recent publications have incorporated adaptations to the extraction for a limited number of sample types including reductions in acid concentration and the addition of an organic fraction  (7–9). For difficult sample types, such as garden, wheat, tree, and alfalfa leafy tissue, extraction efficiency is critical and has not been investigated with less vigorous extraction approaches. Many approaches have been developed to adapt paraquat determinations to LC/MS/MS technology. Early work relied on volatile ion paring reagents (trifluoroacetic acid, hepta fluorobutyric acid) in the mobile phase to increase retention of paraquat, but these reagents caused problems with a reduction in ionization and added extra contaminants to the MS system (10–14). Recent advances in column technology have demonstrated the success of LC/MS/MS paraquat determinations in the absence of ion pairing. The use of liquid separation cell technology with the Obelisc column to separate paraquat and diquat in combination with an MS friendly non-ion pairing MS mobile phase has been reported (7, 8). Hydrophilic interaction chromatography (HILIC), a technique designed for the retention and separation of polar–ionic compounds such as paraquat, has been described in the literature frequently (9, 15–21). It is easily coupled to MS with compatible mobile phases and does not need ion pair reagents. A variety of SPE chemistries have been incorporated into paraquat methods as an efficient and selective technique to reduce sample matrix prior to MS determination. Recent work has illustrated the usefulness of weak cation exchange SPE for paraquat determination. Application notes from Waters Corp. (Milford, MA) and Makihata et al. (19) used the OasisTM WCX SPE weak cation cartridge for water (16–19), and Wang et al. used it for biological samples (20). Whitehead et al. used a weak cation exchange cartridge for urine (21). In an effort to save time and labor, another approach used no SPE cleanup but instead used isotopically labeled analogs of paraquat and diquat as internal standards (ISTDs) to correct for loss of recovery as well as matrix affects (7, 8). The study involved a limited number of food sample types such as potatoes and lentils and did not include leafy or grassy samples such as alfalfa or winter wheat that may be encountered in a drift investigation. The method presented is the result of an effort to find a suitable approach to measure paraquat residues reliably at low levels in sample types typically found with suspected drift. The method takes advantage of new column technology coupled with LC/MS/MS and is designed to handle a wide range of challenging sample types. A rigorous acid digestion ensures a thorough extraction, and the inclusion of both an SPE cleanup

Schaner & Hickes: Journal of AOAC International Vol. 98, No. 2, 2015  513 Table  1.  MRM conditions Compound name

Precursor ion, m/z

Product ions, m/z

Paraquat

186.0

170.9

a

 

171.0

155.0

b

 

171.0

77.0

Paraquat d8

194.0

179.0a

a

  Quantitative ion.

b

  Confirmation ion.

Cone voltage, V

Collision energy, eV

Dwell time, s

Retention time, min

30

15

0.080

1.55

60

30

0.080

1.55

60

30

0.080

1.55

20

18

0.080

1.55

b

step to reduce interferences and an ISTD to correct for recovery and/or matrix MS effects provides for maximum confidence in results. Experimental Standards and Reagents (a) Standards.—Paraquat dichloride was obtained from The U.S. Environmental Protection Agency Pesticide Standard Repository (Ft. Meade, MD), and paraquat d8 dichloride was obtained from CDN Isotopes (Pointe-Claire, Quebec, Canada). (b) Reagents.—All reagents were purchased from Fisher Scientific (Pittsburgh, PA). Acetonitrile (ACN), ammonium formate, and formic acid used to prepare the mobile phase were LC/MS grade. Methanol (MeOH) was OptimaTM grade. Formic acid, ammonium formate, ammonium hydroxide, and hydrochloric acid (HCl) used to digest samples and for SPE cleanup were ACS grade. LC grade water was deionized and purified with a NANOpure water system (Barnstead) from VWR (Wayne, PA). Preparation of Standard Solutions Standard solutions of paraquat and paraquat d8 were stored at 4°C in polypropylene (PP) bottles or PP autosampler vials. Prepare 100 ppm paraquat and paraquat d8 stock standards as cations in MeOH. Prepare 1.0 ppm paraquat and paraquat d8 intermediate standards in LC water. Use paraquat 1.0 ppm to fortify spikes and prepare calibration standards. Use paraquat d8 1.0 ppm as an ISTD to be added to all samples and calibration standards. Prepare fresh calibration standards daily by diluting an aliquot of the 1.0 ppm standards with ACN–150 mmol/L ammonium formate with 0.5% formic acid (6 + 4, v/v). Prepare calibration standards to range from 1.0 to 5.0 ppb paraquat and each standard to contain 2.0 ppb paraquat d8. Sample Preparation and Extraction Use a GM 300 knife mill (Retsch, Newtown, PA) and dry ice to grind winter wheat and alfalfa. Weigh 5 g into a 50 mL PP tube. Fortify spiked samples to contain 0.020 and 0.080 µg/g paraquat. Add paraquat d8 as an ISTD to all samples to equal approximately 0.040 µg/g. Add 14.8 mL 6 M HCl. Cap securely and shake gently by hand to mix the sample. Heat samples in a 100°C water bath for 1 h. Allow samples to cool. Carefully open the cap on the tubes to vent. Cap tubes, shake briefly by hand, and vent again. Centrifuge 5 min. Transfer a 0.075 mL aliquot to a 50 mL PP beaker containing 10 mL 25 mmol/L ammonium

formate, pH 8.0. Use a pH meter to adjust the sample pH to 7–8 with diluted ammonium hydroxide or formic acid. SPE Column Cleanup Use a VacMaster-20 vacuum manifold with PTFE stopcocks (Isolute, Hengoed, MidGlamorgan, UK) to condition OasisTM WCX cartridges (3 cc, 60 mg; Waters Corp.) with 1 column volume MeOH followed by 1 column volume 25 mmol/L ammonium formate, pH 8.0. Load sample extracts onto cartridges in a dropwise fashion. Wash cartridges with 3 mL each of the following: 25 mmol/L ammonium formate pH 8.0, HPLC water, and MeOH. Dry SPE cartridges under full vacuum for 5 min. Place 15 mL PP tubes in the manifold and elute paraquat and paraquat d8 with 2.5 mL formic acid–ACN (10 + 90, v/v). Evaporate extracts to dryness under a gentle stream of nitrogen using a 50°C water bath. Add 0.5 mL ACN–150 mmol/L ammonium formate with 0.5% formic acid (60 + 40, v/v) to the tubes. Vortex mix and fill PP vials for LC/MS/MS analysis. Ultra-Performance LC (UPLC) Conditions An ACQUITY UPLCTM system (Waters Corp.) was used, and separations were achieved using an ACQUITY UPLC bridged ethyl hybrid (BEH) HILIC column (100 × 2.1 mm id, 1.7 µm particle size) from Waters Corp. The flow rate was 0.3 mL/min, injection volume was 5 μL, and column temperature was 30°C. Strong and weak needle washes were MeOH/LC water (50 + 50, v/v) and seal wash MeOH/ LC water (10 + 90, v/v). Mobile phase A was 150 mmol/L ammonium formate with 0.5% formic acid; mobile phase B was ACN. The gradient program started at 40% A and ramped to 20% B over the course of 0.1 min, where it was held until 3.5 min. A subsequent re-equilibration for 2.4 min of 40% A was performed before the next injection. MS/MS Conditions MS/MS detection was performed using a Waters Corp. Xevo TQ-S tandem quadrupole mass spectrometer. The instrument was operated using an electrospray ionization (ESI) source in positive mode. The ESI source was set as follows: capillary 0.80 kV, desolvation temperature 450°C, source temperature 150°C, cone gas 150 L/h, and desolvation gas 700 L/h. Both gases were nitrogen (>95% purity). Collision-induced dissociation was performed using argon (99.9999%) as a collision gas at a pressure of 3.5 × 10–3 mbar in the collision cell.

514  Schaner & Hickes: Journal of AOAC International Vol. 98, No. 2, 2015 choice for conditioning the cartridge, we preferred more volatile ammonium formate. For the final elution, 10% formic acid in ACN was used to avoid any aqueous fraction and allow an efficient evaporation step prior to LC/MS/MS. These adaptations worked well with the WCX SPE and produced an effective cleanup for the matrixes in our study. ISTD

Figure  1.  Chromatogram of paraquat LOQ standard (1.0 ppb).

Multiple reaction monitoring (MRM) parameters for paraquat and paraquat d8 are shown in Table 1. Quantitation and Confirmation Prior to sample analysis, system suitability was obtained by establishing a linear three-point ISTD calibration curve with a correlation coefficient (r) of at least 0.995 and a chromatogram equivalent to that in Figure 1. Retention times were stable and consistent. Ion ratios of the confirmation ions to the quantitation ion were within 10% (absolute). Results and Discussion Extraction Developing an efficient extraction of paraquat for difficult matrixes such as vegetable, wheat, tree, and alfalfa leafy tissue typical of what may be encountered in pesticide misuse cases turned out to be the most critical component of this project. Our investigation began with recent publications that used less rigorous extraction procedures (7–9) than had been previously reported (2–6). While these publications demonstrated successful extraction with specific matrixes, when applied to the plant tissue types in our study we were unable to recover paraquat d8. With a more rigorous extraction of 1 h at 100°C with 6 M HCl on the same plant, tissue an almost complete recovery of paraquat d8 (fortified at 0.040 µg/g) was obtained: 71 ± 4.5% for broadleaf composite matrix and 84 ± 12% for garden plant composite matrix. These acid digestion conditions were previously published by Chichila et al. (3, 4) and had been used successfully by our laboratory in the past. It was evident that the acid concentration was crucial. Incorporating a more rigorous acid digestion in the method was necessary to obtain an effective extraction with the more difficult sample types. SPE Column Cleanup Utilizing SPE to reduce the amount of matrix introduced into the MS system was a key element in our method design. The intent was to develop an efficient SPE cleanup that produced an MS friendly extract. We evaluated the Oasis WCX SPE cartridge that was used successfully for paraquat in several studies (16–20). Although phosphate buffer was a popular

We routinely incorporate ISTDs into our test methods whenever practicable to improve accuracy and consistency as it is not possible to eliminate all matrix affects in the samples we typically test. This approach eliminates the need for matrixmatched standards, which adds extra steps to the method and can be time consuming. Paraquat d8 is used as the ISTD and was effectively extracted over a range of challenging sample types, providing useful correction for recovery rates and MS matrix effects. Throughout the course of this study with different data sets run on different days, paraquat d8 recovery averaged 62 ± 25% (n = 43). Instrumentation Our plan was to investigate new options in chromatography and MS friendly mobile phases for LC/MS/MS determinations and to avoid ion-pairing reagents. Two recent publications described novel chromatography that worked well for paraquat separations in LC/MS/MS. The first approach used an Obelisc R 150 × 2.1 mm, 5 µm particle size LC column (7, 8). When used in our UPLC/MS system this column provided good separation, but the 20 min run times were not efficient. Problems were encountered with significant paraquat carryover that interfered with our ability to meet the LOQ in blank samples and contributed to nonreproducible standard responses. Utilizing two blank injections between samples and standards helped but was considered too time consuming for routine work. A decision was made to try the ACQUITY UPLC BEH HILIC 100 × 2.1 mm id, 1.7 µm particle size column (18, 20). Run times were considerable shorter at 6 min, and standard injections were consistent and reproducible. Paraquat carryover was still present but at a reduced amount that did not interfere with the LOQ. The column and mobile phase produced good peak shape and sensitivity and was constant and reproducible through all the data sets. Method Validation Method validation was performed using alfalfa and winter wheat leafy tissue. Each matrix was run as a single sample set Table  2.  Recoveries from winter wheat and alfalfa Matrix

Mean, %

Range, %

RSD, %

0.020 ppm fortification level (n = 3) Winter wheat Alfalfa

105

75–127

26

114

88–141

23

0.080 ppm fortification level (n = 3) Winter wheat

93

71–124

30

Alfalfa

80

71–90

12

Schaner & Hickes: Journal of AOAC International Vol. 98, No. 2, 2015  515

Figure  2.  Chromatogram of reagent blank, winter wheat, winter wheat spiked at LOQ (0.020 ppm), and winter wheat spiked at four times LOQ (0.080 ppm).

Table  3.  Reproducibility of results for incurred residues Matrix

Date extracted

Result, ppm

Potato leaves No. 1

12/6/2013

0.044

12/10/2013

0.027

Broadleaf composite

Alfalfa

Potato leaves No. 2

Garden plant composite

Tree leaves

12/6/2013

0.083

12/10/2013

0.080

12/10/2013

0.12

12/10/2013

0.074

12/6/2013

0.063

12/6/2013

0.063

12/10/2013

0.051

12/6/2013

0.34

12/10/2013

0.37

1/16/2014

0.26

12/6/2013

0.25

12/10/2013

0.16

12/10/2013

0.22

12/10/2013

0.25

1/16/2014

0.25

12/6/2013

0.51

12/10/2013

0.35

1/16/2014

0.39

1/16/2014

0.40

Avg., ppm

SD, ppm

0.089

0.021

0.059

0.0069

0.32

0.057

0.23

0.039

0.41

0.068

and included three replicates fortified at the reporting limit of 0.020 ppm, three replicates at 0.080 ppm, one matrix blank, and one reagent blank (Figure 2). Recovery was calculated using the average relative response of bracketing standards equal in concentration to the expected response. Average recoveries ranged from 105 to 114% (RSD 23–26%) at 0.020 ppm fortification level and 80 to 94% (RSD 12–30%) at 0.080 ppm fortification level (Table 2). In addition, garden plant composite, alfalfa, tree leaves, and potato leaves with incurred paraquat residues were analyzed on separate days by a different analyst. Consistent results demonstrated method reproducibility (Table 3). Retention times were stable and consistent, varying less than 1% from injection to injection. Confirmation of ion ratios passed the criteria established in the section Quantitation and Confirmation above. Although trace paraquat residues were noted in all reagent and matrix blanks (estimated to be about 0.0040 to 0.010 ppm), these levels did not interfere with the ability to measure at the LOQ (Figure 2). It is unknown where the residues originated or how to eliminate them completely from the system. Injections of LC water showed minimal contamination from the UPLC or MS system and did not account for all of the residues. The LOQ of 0.020 ppm was established as a compromise between how low the instrument could measure and the trace carryover effects. This level was comparable to what other methods have established and is sufficient to determine drift. Acknowledgments We thankfully acknowledge the efforts of Mark Winslow who assisted with the reproducibility study and the Montana Department of Agriculture for their open support of this project.

516  Schaner & Hickes: Journal of AOAC International Vol. 98, No. 2, 2015 References   (1) http://www.paraquat.com/ (accessed on February 27, 2014)   (2) Anderson, L., & Boseley, A.D. (1997) ZENECA Agrochemicals SOP RAM 272/02, London, UK   (3) Chichila, T.M., & Walters, S.M. (1991) J. Assoc. Off. Anal. Chem. 74, 961–967   (4) Chichila, T.M., & Gilvydis, D.M. (1993) J. AOAC Int. 76, 1323–1328   (5) Guijarro, E.C., Yanez-Sedeno, P., Polo Diez, L.M. (1987) Anal. Chim. Acta 199, 203–208. http://dx.doi.org/10.1016/S00032670(00)82816-0   (6) Worobey, B.L. (1987) Pestic. Sci. 18, 245–257. http://dx.doi. org/10.1002/ps.2780180405   (7) Anastassiades, M., Kolberg, D.I., Mack, D., Wildgrube, C., Sigalov, I., & Dork, D. (2013) EU Reference Laboratories for Residues of Pesticides Single Residue Methods (EURL-SRM) QuPPE- Method version 7.1, Stuttgart, Germany   (8) Kolberg, D.I., Mack, D., Anastassiades, M., Hetmanski, M.T., Fussell, R.J.,Meijer, T., & Mol, H.G. (2012) Anal. Bioanal. Chem 404, 2465–2474. http://dx.doi.org/10.1007/s00216-0126340-9   (9) Young, M.S., Shia, J.C., & Tran, K.V. (2013) Application Note 720004851en, Waters Corp., Milford, MA (10) Grey, L., Nguyen, B., & Yang, P., (2002) J. Chromatogr. A 958, 25–33. http://dx.doi.org/10.1016/S0021-9673(02)00400-4 (11) Wang, K.G., Chen, S.M., Hsu, J.F., Cheng, S.G., & Lee,

C.K. (2008) J. Chromatogr. B 876, 211–218. http://dx.doi. org/10.1016/j.jchromb.2008.10.042 (12) Arffin, M.M., & Anderson, R.A. (2006) J. Chromatogr. B 842, 91–97. http://dx.doi.org/10.1016/j.jchromb.2006.03.051 (13) Taguchi, V.Y., Jenkins, S.W.D., Crozier, P.W., & Wang, D.T. (1998) J. Am. Soc. Mass Spectrom. 9, 830–839. http://dx.doi. org/10.1016/S1044-0305(98)00043-9 (14) Paterio-Moure, M., Martinez-Carballo, E., Aris-Estevez, M., & Simal-Grandara, J., (2008) J. Chromatogr. A 1196-1197, 110–116. http://dx.doi.org/10.1016/j.chroma.2008.03.081 (15) Wunnapuk, K., Medley, G.A., Liu, X., Grice, J.E., Jayasinghe, S., Gawarammana, I., Buckley, N.A., & Roberts, M.S. (2011) J. Chromatogr. B 879, 3047–3052. http://dx.doi.org/10.1016/j. jchromb.2011.09.008 (16) Young, M.S., & Jenkins, K.M. (2004) LC GC North America, 44–45 (17) Shia, J.C., Yabu, M., Tran, K.V., & Young, M.S. (2013) Application Note 720004732en, Waters Corp., Milford, MA (18) Tran, K.V., Shia, J.C., & Young, M.S. (2012) Application Note 720004220en, Waters Corp., Milford, MA (19) Makihata, N., Yamasaki, T., & Eiho, J. (2007) Bunseki Kagaku 56, 579–585. http://dx.doi.org/10.2116/bunsekikagaku.56.579 (20) Wang, Z., Wang, Z., & Xing, J. (2011) J. Anal. Toxicol. 35, 23–27. http://dx.doi.org/10.1093/anatox/35.1.23 (21) Whitehead, R.D., Montesano, M.A., Jayatilaka, N.K., Buckley, B., Winnik, B., Needham, L.L., & Barr, D.B. (2010) J. Chromatogr. B 878, 2548–2553. http://dx.doi.org/10.1016/j. jchromb.2009.09.029

Copyright of Journal of AOAC International is the property of AOAC International and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.

electrospray ionization-mass spectrometry.

The method presented is a suitable approach for routine testing of paraquat in difficult sample types, such as winter wheat and alfalfa plant tissue, ...
497KB Sizes 17 Downloads 12 Views