Article pubs.acs.org/ac
Multiresidue Pesticide Analysis of Botanical Dietary Supplements Using Salt-out Acetonitrile Extraction, Solid-Phase Extraction Cleanup Column, and Gas Chromatography−Triple Quadrupole Mass Spectrometry Douglas G. Hayward,*,† Jon W. Wong,*,† Feng Shi,∥ Kai Zhang,† Nathaniel S. Lee,‡ Alex L. DiBenedetto,‡ and Mathew J. Hengel§ †
U.S. Food and Drug Administration, Center for Food Safety and Applied Nutrition, 5100 Paint Branch Parkway, HFS-706, College Park, Maryland 20740-3835, United States ‡ Joint Institute for Food Safety and Applied Nutrition, University of Maryland, 1122 Patapsco Building, College Park, Maryland 20742-6730, United States § IR-4 Laboratory, University of California, Department of Environmental Toxicology, One Shields Avenue, Davis, California 95616-5270, United States S Supporting Information *
ABSTRACT: Dietary supplements form an increasing part of the American diet, yet broadly applicable multiresidue pesticide methods have not been evaluated for many of these supplements. A method for the analysis of 310 pesticides, isomers, and pesticide metabolites in dried botanical dietary supplements has been developed and validated. Sample preparation involved acetonitrile:water added to the botanical along with anhydrous magnesium sulfate and sodium chloride for extraction, followed by cleanup with solid-phase extraction using a tandem cartridge consisting of graphitized carbon black (GCB) and primary− secondary amine sorbent (PSA). Pesticides were measured by gas chromatography-tandem mass spectrometry. Accuracy and precision were evaluated through fortifications of 24 botanicals at 10, 25, 100, and 500 μg/kg. Mean pesticide recoveries and relative standard deviations (RSDs) for all botanicals were 97%, 91%, 90%, and 90% and 15%, 10%, 8%, and 6% at 10, 25, 100, and 500 μg/kg, respectively. The method was applied to 21 incurred botanicals. Quinoxyfen was measured in hops (100−620 μg/kg). Tetraconazole (48 μg/kg), tetramethrin (15 μg/kg), methamidophos (50 μg/kg), and chlorpyrifos (93 μg/kg) were measured in licorice, mallow, tea, and tribulus, respectively. Quintozene, its metabolites and contaminants (pentachloroaniline, pentachlorobenzene, pentachloroanisole, and pentachlorothioanisole and hexachlorobenzene and tecnazene, respectively), with hexachlorocyclohexanes and DDT were identified in ginseng sources along with azoxystrobin, diazinon, and dimethomorph between 0.7 and 2800 μg/kg. Validation with these botanicals demonstrated the extent of this method’s applicability for screening 310 pesticides in a wide array of botanical dietary supplements.
A
products.1−3 An interlaboratory comparison study conducted by Kong et al.4 revealed the difficulties and the many different current procedures used for analyzing pesticides in ginseng products. There are a number of methods5−9 based on procedures for fresh plant-derived foods10−16 that describe pesticide screening in dried botanical dietary supplements, spices, medicinal plants, herbals, teas, and phytomedicines. Methods involve organic solvent extraction and cleanup procedures to remove
botanical dietary supplement is a plant-based material primarily valued for its possible medicinal or therapeutic properties as an aphrodisiac, stress reliever, memory enhancer, or for treatment of inflammation or infection. Such botanicals are widely regarded as valuable agricultural crops, and to prevent economic losses, pesticides may be used against mold and insects that may cause plant damage. Despite the usefulness of pesticides in agricultural practices, there are concerns about their improper use and residue levels in plant commodities. Pesticides can remain or accumulate from application during the growing stage of the plant or from postharvest treatment. Previous studies have shown that many organochlorine and organophosphorus pesticides have been detected in botanical © 2013 American Chemical Society
Received: February 1, 2013 Accepted: March 27, 2013 Published: March 27, 2013 4686
dx.doi.org/10.1021/ac400481w | Anal. Chem. 2013, 85, 4686−4693
Analytical Chemistry
Article
number of product ions monitored by modern triple quadrupoles is only limited by the MS/MS scan speed of the quadrupole, approximately 2 ms per transition, and the scan cycle time required to record sufficient data points across the peak width of a given gas chromatographic response. Triple quadrupoles are well-suited for MS/MS of large numbers of pesticides in a single GC acquisition.16 In this study organohalogen, organophosphorus, organonitrogen, and pyrethroid pesticides were measured using GC− triple quadrupole-based mass spectrometry. GC−MS/MS has recently become more affordable than in the past, finding use with many pesticide laboratories for multiresidue screening in fresh food8,13−16,23−25 despite the increased cost over single quadrupole or quadrupole ion traps. This approach has recently been increasingly used for the targeted measurement of pesticides in fresh produce, ginseng, and other plant products.8,13−16 Our objective was to study the performance limits and validate a method covering 310 pesticides in a single GC−MS/MS acquisition, using time scheduled monitoring of two SRMs per pesticide in 24 botanical dietary supplements while using a minimal, but effective, extraction and SPE-based sample clean up.21,22
coextracted interfering components from the matrix. The prepared extracts are measured by various instrumental methods such as capillary gas chromatography (GC) or highperformance liquid chromatography with element-selective, spectrophotometric and/or mass spectrometric detection. While these methods work well with fresh produce, there are serious challenges with dried botanicals, due to the varieties and complexities of botanical matrices and concentrated levels of natural products present in dehydrated products. The complexity and variety of dried botanicals along with the low concentration of many pesticides complicate the detection, identification, and quantification of the residues. We have also evaluated methods for the analysis of pesticides in the dietary supplement ginseng using salt-out extraction with different organic solvents [acetonitrile and 2:1:1 acetone:cyclohexane:ethyl acetate mixture (v:v:v)] or dry extraction with ethyl acetate7 paired with solid-phase dispersive clean up6, the more effective solid-phase extraction techniques8, or gel permeation chromatography/solid-phase extraction cleanup.7 GC-based separation and detection used a combination of either flame photometric detection and single quadrupole mass spectrometry with selective ion monitoring (GC−MS/SIM)6 or GC−high resolution-time-of-flight-mass spectrometry (GCHR-TOF-MS) and GC−MS/SIM7 or GC−tandem (or triple quadrupole) mass spectrometry (GC−MS/MS) and GC−MS/ SIM.8 The latter report8 showed lower limits of detection at or well-below the 10 μg/kg level and improved selectivity for the identification for all 167 targeted pesticides in ginseng products by MS/MS over GC−MS/SIM and GC−HR−TOF−MS, despite a more extensive cleanup.7,8 While GC−MS/SIM provides qualitative and quantitative information on pesticide residues in fresh produce, there are limitations with this approach in complex botanicals. Especially with SIM, due to less selectivity than MS/MS, the SIM ions may lose sufficient specificity due to high noise or interferences. Ion abundances may result from coextractives in the plant matrix and could skew the ratios between the qualifying ions, preventing detection or identification. In addition, the U.S. Federal Drug Administration (FDA) pesticide identification criteria17 using GC−MS/SIM requires the use of four ions found with the proper ion ratios in combination with correct GC retention time and other quality assurance factors to take regulatory action, while GC−MS/MS requires only two precursor/product ion pairs. Yost and Enke first reported MS/MS using a triple quadrupole mass spectrometer in 1979.18 Over the next ten years these instruments became more commercially available but were very expensive compared with single quadrupoles or ion trap instruments available at the same time. Ion traps are also capable of doing MS/MS, but MS/MS in ion traps was not available commercially until 1995. Ion trap MS/MS was then exploited for a variety of applications, including pesticides in foods. Ion traps using MS/MS have been reported to work well for fresh produce.19 Most reports discuss methods with limited numbers of pesticides 50% of the quantification ion or ±25% for ions 20% of the quantification ion, ±30% between 10 and 20%, or ±50% for ions 10% of the spectrum base peak and tried to use the same precursor to produce both product ions where possible, perhaps originating from approaches used with ion traps in MS/MS.19,20 Selecting a molecular ion helps ensure that one will utilize the highest mass m/z ion for the precursor, and it has been suggested that the signal-to-noise will improve.27 Table 1 of the Supporting Information lists the overlapping time-dependent MRM windows used in the method for each transition pair. Most windows were 0.8−1.0 min wide, except where close-eluting isomers required a wider window. The windows illustrated represent a compromise between window widths and scan cycle times, while maximizing dwell times and the total number of pesticides in a single GC−MS/MS acquisition. Narrower windows (0.4−0.6) were found to increase the frequency of missing pesticides due to retention time shifts. Fortification Studies for 24 Botanicals. Recoveries were calculated for all identified pesticides fortified in each botanical at four concentrations in quadruplicate. The desired recovery range was 70−120% at 10 μg/kg fortification with an RSD less than 20%. Table 2 of the Supporting Information provides the mean recoveries and the RSD for each pesticide at each fortification concentration in all 24 botanicals. Table 1 provides the mean recoveries by botanical across all fortified pesticides, measured at each fortifying concentration along with the average RSD and the number of compounds not detected at each concentration for each botanical. Overall, the mean recoveries and RSDs met the criteria at all concentrations for the measured pesticides with an overall mean recovery across all botanicals of 97, 91, 90, and 90% for 10, 25, 100, and 500 μg/kg with mean RSDs of 15, 10, 8, and 6, respectively. Nevertheless, a trend was observed for the number of nondetected pesticides, increasing by as much as 20-fold for certain botanicals when the fortifying concentration reached 10 μg/kg (Table 1). In contrast, the accuracy for measurement of the remaining detected pesticides was retained, with the exceptions of cinnamon (35% of pesticides) and dong quai (60% of pesticides) with fewer pesticides measured with recoveries between 70−120%, at the lowest fortification. The overall accuracy met our criteria for 72−99% of the pesticides across the 24 botanicals and all fortifying concentrations with a few exceptions such as cinnamon at 10 and 100 μg/kg (Table 1). At
Table 2. Percentage of the Total Pesticides (including Isomers and Metabolites, n = 310) Measured with LOQs ≤ 5, 10, 20, or 40 μg/kg by Botanical percent LOQs determined at ≤5, 10, 20, and 40 μg/kg (n = 310) botanical
≤5 μg/kg
≤10 μg/kg
≤20 μg/kg
≤40 μg/kg
astralagus bitter orange peel black cohosh root chamomile cinnamon comfrey root dong quai echinacea fenugreek garlic ginger gingko biloba ginseng green tea hoodia hops jasmine kava kava licorice root milk thistle psyllium saw palmetto st. john’s wort valerian root
23 22 25 18 17 26 15 18 25 27 9 2 19 24 23 2 21 18 18 16 15 32 6 10
45 44 44 37 39 45 33 38 48 48 30 15 34 42 42 9 43 35 33 29 28 59 15 20
71 69 75 60 60 71 54 62 72 73 47 46 65 65 64 20 65 55 56 51 62 75 26 41
84 84 90 73 73 85 68 80 87 87 58 68 85 85 83 37 83 65 70 67 80 82 48 64
detected at or near the MRL (ginger, gingko biloba, hops, kava kava, St. John’s Wort, and valerian root). Although the total number of LOQs estimated were often not that different for some of these botanicals such as gingko biloba (68%), kava kava (65%), and valerian root (64%), the fraction of LOQs measured at or below 10 μg/kg could be quite low, such as in gingko biloba (15%), St John’s Wort (15%), and valerian root (20%). Fewer than half of all the pesticide−botanical 4690
dx.doi.org/10.1021/ac400481w | Anal. Chem. 2013, 85, 4686−4693
Analytical Chemistry
Article
combinations were able to produce LOQs at or below the MRL (Table 2). Difficulties Using MS/MS with Botanical Supplements. For certain botanical extracts, identification criteria was not met at 10 μg/kg for up to 210 of the pesticides for ginger or 233 in hops, while matrices such as black cohosh, ginseng, green tea, licorice root, and psyllium produced only 39, 64, 43, 43, and 39 unidentified pesticides fortified at the MRL, respectively (Table 1). Some pesticides, such as metofluthrin or chlorthion, were problematic with nearly all botanicals. Neither transition was observable due to high concentrations of isobaric background in matrix standards, except the very highest concentrations. For other pesticides, the coextracted components would sometimes shift the pesticide retention times outside the acquisition window preventing the determination (e.g., crotoxphos, diamidophos, fenamiphos sulfoxide, phorate oxon sulfoxide). Chloroneb did not produce a useful matrix calibration curve due to interference in any botanical. Similarly, allethrin and biphenyl did not often produce a useful calibration in many botanicals due to interference. Poorer responses were found for carfentrazone-ethyl, dioxathion, ethoxyquin, fenamiphos, sulfoxide, naled, schraden, tralomethrin, and vernolate. Other pesticides typically demonstrated less stability in the clean up or GC/MS, such as with azinphos methyl captan, chlorthalonil, diamidophos, dioxathion, dichlofluanid, endrin aldehyde, endrin ketone, fosthiazate, iprodione, naled, and tetraidoethylene. The above-mentioned pesticides aside, most pesticide determinations were more profoundly affected by the specific matrix and to a lesser extent the concentration or stability of the fortified pesticide (Table 1) (Table 2 of the Supporting Information). Figure 1 illustrates that several of these matrix-derived difficulties seen with the SRM chromatograms are acquired for difenaconazole through fragmentation of the spectrum base
peak, m/z 265, and a second major ionization-induced fragment at m/z 323. Figure 1 provides six SRM chromatograms for the pesticide difenoconazole that contains the two isomers in solvent-only standards, licorice root or milk thistle standards. Two isomers are present in a solvent standard in an approximately a 60:40 ratio as are the SRMs from milk thistle with nearly matching retention times. The relative response ratios of the two difenoconazole-specific transitions are not the same for the solvent standard and the milk thistle response. The retention time for difenoconazole in licorice root was shifted by 0.3 min relative to the solvent standards. The SRM ratios are also not the same for each isomer, indicating interference was probably present. The earlier eluting difenoconazole isomer could not be identified nor could an LOQ be estimated in licorice root at this concentration (10 μg/ L). The later-eluting isomer was identified with a correct ion ratio for the matrix-matched standard, and the LOQ could be estimated at this concentration. Recent reports have suggested that foods could be efficiently screened for pesticides using very fast chromatography by greatly increasing the temperature programming rate while shortening the GC column and/or widening its internal diameter.15,23,24 These column systems are operated with reduced resolving power and rely on the selectivity of MS/MS to resolve the pesticides from matrix components. When our licorice, milk thistle, or dong quai extracts were chromatographed on a 30 m column using these very fast GC temperature ramping rates, poorer response for difenoconazole and other pesticides was observed, generating false negatives or interferences. Figure 2 illustrates difficulties with the analysis in two botanicals for mirex, a pesticide which normally gives excellent chromatographic performance, rarely with interference, and a sensitive response higher than all others tested. Figure 2 provides the product ion chromatograms for mirex in a typically
Figure 1. Product ion chromatograms from matrix-matched standards equal to 10 μg/L and solvent-only standards (20 μg/L) for difenoconazole isomers in licorice (top two traces), milk thistle (middle two traces), and solvent standard (bottom two traces) showing both transition m/z and ion ratios derived from the sum of the two isomers.
Figure 2. Product ion chromatograms for mirex (tR = 30.92 min) from matrix matched standards at 2 μg/L for astralagus showing strong interference for transition 272 → 237, hops and comfrey root. Hops standard was done on a different, but the same type of GC column (tR = 30.79 min). 4691
dx.doi.org/10.1021/ac400481w | Anal. Chem. 2013, 85, 4686−4693
Analytical Chemistry
Article
We combined a typical fresh food cleanup using much less sample to help maintain recoveries while diluting the extract to one-fourth the amount to reduce matrix difficulties, unlike our previous studies.6−8 Nevertheless, this study demonstrates there are limits to the sensitivity and selectivity for many pesticides in complex botanicals described in this validation. GC-amenable pesticides, including ones most commonly determined by GC/ MS with ease and high sensitivity (e.g., hexachlorobenzene, BHCs, DDT, mirex), were sometimes difficult to measure near the MRL. Botanical supplement test portion preparation has been improved in some recent studies7,29−31 to the point where the comprehensive screening feature of the GC/MS-based method is reduced by an unacceptable degree through the loss of many pesticides.7,31 There is a balance to optimize between efforts to reduce matrix interference and efficiency of a multianalyte screen. Future improvements to comprehensive screening methods may require measuring techniques that go beyond the most commonly used technologies for the mass spectrometric determination of pesticides that use lowresolution quadrupole with electron ionization or electrospray ionization.32−34
nonproblematic matrix, comfrey root, indicating no interference and an LOQ of 1.3 μg/kg. Nevertheless, mirex fortified into astralagus could not be identified below ∼8 μg/kg due to a nearly coeluting interference for m/z 272 → 237. A similar situation was observed for mirex in hops, where the mirex LOQ was ∼7 μg/kg because of high noise accompanied by closely or coeluting interferences at m/z 270 → 235 and m/z 272 → 237. Incurred Botanicals. During the fortification study, twentyone botanicals were prepared in quadruplicate for pesticide screening. Table 4 of the Supporting Information provides a summary of pesticides found in the botanicals with residues identified and measurable. All six sources of ginseng contained reportable pesticides. Four ginsengs were found to contain quintozene (148−2800 μg/kg) and its metabolites and contaminants (PCB, HCB, pentachloroaniline, pentachloroanisole, pentachlorothioanisole, tecnazene, and tetrachloroaniline; 6−120 μg/kg). In addition, three of these ginsengs also contain lindane and α, β, and δ isomers of BHC isomers (29−543 μg/ kg) and two had lower amounts of iprodione (20 and 150 μg/ kg). One ginseng contained p,p′-DDE (576 μg/kg)-p,p′-DDT (121 μg/kg), while a sixth ginseng contained azoxystrobin (31 μg/kg), diazinon (11 μg/kg), and dimethomorph (6 μg/kg). Three sources of hops contained quinoxyfen (100−620 μg/kg). Selected product ion chromatograms for two ginseng and one hop sources are provided in Figure 1 of the Supporting Information. The ginseng sources with identified incurred residues for quintozene or azoxystrobin are shown and quinoxyfen confirmed in a hops source. The hop sources originated from field studies for the application of quinoxyfen on hops, therefore some hops sources were expected to contain quinoxyfen residue.26 The tolerance for quinoxyfen on hops is 3000 μg/kg; none of the found residues approach this concentration. No tolerances exist for the pesticides found on any of the ginsengs in Table 3 of the Supporting Information, except iprodione. One ginseng contained DDT isomers and metabolites. The relatively high amounts of p,p′-DDE indicate that DDT probably originates from historic contamination of the soil. Azoxystrobin, diazinon, and dimethomorph found on one ginseng indicate more recent use of these pesticides as does quintozene on the other four ginsengs. Single residues were found on four other botanicals all below 100 μg/kg. Tribulus and licorice root had the highest and lowest amount with chlorpyrifos at 93 μg/kg and tetramethrin at 15 μg/kg, respectively. Green tea and mallow contained approximately 50 μg/kg each of methamidophos and tetraconazole, respectively. Pesticides were identified in some of the botanicals used in the fortification study. The following pesticides were identified: allethrin (hoodia, jasmine, and valerian root), bifenthrin (green tea), cyhalothrin (green tea), DDT isomers (jasmine and valerian root), fenvalerate, (green tea and jasmine), prallethrin (Milk Thistle, St John’s Wort, and valerian root), and resmethrin (St John’s Wort).
■
ASSOCIATED CONTENT
S Supporting Information *
This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*D.G.H.: e-mail,
[email protected]; tel, (240)-4021654; fax, (240) 402-2332. J.W.W.: e-mail,
[email protected]. gov; tel, (240)-402-2172; fax, (301)-436-2332. Present Address ∥
Givaudan Flavors Corporation, 1199 Edison Drive, Cincinnati, OH 45216−2265, United States. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We appreciate the Office of Dietary Supplements, National Institutes of Health for financial support of this research through an interagency agreement and the United States Environmental Protection Agency, National Pesticide Repository, Ft. Meade, MD for providing a majority of the pesticide standards used in these studies.
■
REFERENCES
(1) Huggett, D. B.; Block, D. S.; Khan, I. A.; Allgood, J. C.; Benson, W. H. Hum. Ecol. Risk Assess. 2000, 6, 767−776. (2) Huggett, D. B.; Khan, I. A.; Allgood, J. C.; Blosck, D. S.; Schlenk, D. Bull. Environ. Contam. Toxicol. 2001, 66, 150−155. (3) Durgnat, J.-M.; Heuser, J.; Andrey, D.; Perrin, C. Food Addit. Contam. 2005, 22 (12), 1224−1230. (4) Kong, M.-F.; Chan, S.; Wong, Y.-C.; Wong, S.-K.; Sin, D. W.-M. J. AOAC Int. 2007, 90 (4), 1133−1141. (5) U.S. Pharmacopeia 24. Articles of Botanical Origin, Pesticide Residue Testing. Sec. 561. (6) Wong, J. W.; Hennessy, M. K.; Hayward, D. G.; Krynitsky, A. J.; Cassias, I.; Schenck, F. J. J. Agric. Food Chem. 2007, 55 (4), 1117− 1128. (7) Hayward, D. G.; Wong, J. W. Anal. Chem. 2009, 81 (14), 5716− 23.
■
CONCLUSIONS Effective and efficient methods are necessary to adapt to changing pesticide usage and to detect illegal use of older pesticides and recent or historic environmental contamination. The method described here would be part of a comprehensive pesticide screen for botanicals, which would also require LC− MS methods that might overlap many of the pesticides determined in the method.28 Tandem mass spectrometry using quadrupoles is fast, sensitive and selective, compact, and easy to use and maintain, with a large dynamic range. 4692
dx.doi.org/10.1021/ac400481w | Anal. Chem. 2013, 85, 4686−4693
Analytical Chemistry
Article
(8) Wong, J. W.; Zhang, K.; Tech, K.; Hayward, D. G.; Krynitsky, A. J.; Cassias, I.; Schenck, F. J.; Banerjee, K.; Dasgupta, S.; Brown, D. J. Agric. Food Chem. 2010, 58, 5884−5896. (9) Tagami, T.; Kajimura, K.; Satsuki, Y.; Nakamura, A.; Okihashi, M.; Takatori, S.; Kakimoto, K.; Obana, H.; Kitagawa, M. J. Nat. Med. 2008, 62, 126−129. (10) U.S. Food and Drug Administration. Pesticide Analytical Manual, 3rd ed.; U.S. Department of Health and Human Services, Public Health Service: Rockville, MD, 1999; Vol. 1. (11) Fillion, J.; Hindle, R.; Lacroix, M.; Selwyn, J. J. AOAC Int. 1995, 78, 1252−1266. (12) U.S. Food and Drug Administration. Pesticide Analytical Manual, 3rd ed.; U.S. Department of Health and Human Services, Public Health Service: Rockville, MD, 1999; Vol. 1. (13) Sannino, A.; Bandini, M.; Bolzoni, L. J. AOAC Int. 2003, 86, 101−108. (14) Pang, G. F.; Fan, C. L.; Liu, Y. M.; Cao, Y. Z.; Zhang, J. J.; Li, X. M.; Li, Z. Y.; Wu, Y. P.; Guo, T. T J. AOAC Int. 2006, 89, 740−771. (15) Walorczyk, S.; Gnusowski, B. J. Chromatogr., A 2006, 1128 (1− 2), 236−243. (16) Okihashi, M.; Takatori, S.; Kitagawa, Y.; Tanaka, Y. J. AOAC Int. 2007, 90 (4), 1165−1179. (17) http://inside.fda.gov:9003/downloads/ORA/ OfficeofRegionalOperations/DivisionofFieldScience/UCM189632.pdf (accessed Month Day, Year). (18) Yost, R. A.; Enke, C. G. Anal. Chem. 1979, 51, 1251A−1264A. (19) Leohtay, S. J. J. AOAC Int. 2000, 83, 680−697. (20) Martínez Vidal, J. L.; Arrebola, F. J.; Garrido Frenich, A.; Martínez Fernández, J.; Mateu-Sanchez, M. Chromatographia 2004, 59, 321−326. (21) Wong, J. W.; Zhang, K.; Tech, K.; Hayward, D. G.; Makovi, C. M.; Krynitsky, A. J.; Schenck, F. J.; Banerjee, K.; Dasgupta, S.; Brown, D. J. Agric. Food Chem. 2010, 58, 5868−5883. (22) Anastassiades, M.; Lehotay, S. J.; Štajnbaher, D.; Schenck, F. J. J. AOAC Int. 2003, 86, 412−431. (23) Frenich, A. G.; González-Rodríquez, M. J.; Arrebola, F. J.; Vidal, J. L. M. Anal. Chem. 2005, 77, 4640−4648. (24) Koesukwiwat, U.; Lehotay, S. J.; Leepipatpibooon, N. J. Chromatogr., A 2011, 1218 (39), 7039−7050. (25) Banerjee, K.; Utture, S.; Dasgupta, S.; Kandaswamy, S. P.; Kulkarni, S.; Adsule, P. J. Chromatogr., A 2012, 1270, 283−295. (26) Hengel, M. J.; Miller, M. J. Agric. Food Chem. 2008, 56, 6851− 6856. (27) Filakov, A. B.; Steiner, U.; Jones, L.; Amirav, A. Int. J. Mass Spectrom. 2006, 251, 47−58. (28) Fillâtre, Y.; Rondeau, D.; Bonnet, B.; Daguin, A.; Jadas-Hécart, A.; Communal, P.-Y. Anal. Chem. 2011, 83, 109−117. (29) Specht, W.; Pelz, S.; Gilsbach, W. Anal. Bioanal. Chem. 1995, 353 (2), 183−190. (30) Mol, H. G.; Rosseboom, A.; van Dam, R.; Roding, M.; Arondeus, K.; Sunarto, S. Anal. Bioanal. Chem. 2007, 389 (6), 1715− 1754. (31) Cajka, T.; Sandy, C.; Bachanova, V.; Drabova, L.; Kalachova, K.; Pulkrabova, J.; Hajslova, J. Anal. Chim. Acta 2012, 743, 51−60. (32) Shi, Y.; Chang, J. S.; Esposito, C. L.; Lafontaine, C.; Berube, M. J.; Fink, J. A.; Espourteille, F. A. Food Addit. Contam. 2011, 28 (10), 1383−1392. (33) Caijka, T.; Riddelova, K.; Zomer, P.; Mol, H.; Hajslova, J. Food Addit. Contam. 2011, 28 (10), 1372−1382. (34) Kaufmann, A.; Butcher, P.; Maden, K.; Walker, S.; Widmer, M. Rapid Commun. Mass Spectrom. 2011, 25, 979−992.
■
NOTE ADDED AFTER ASAP PUBLICATION This paper was published on the Web on April 15, 2013. Table 1 of the Supporting Information was revised, and the corrected version was reposted on April 18, 2013.
4693
dx.doi.org/10.1021/ac400481w | Anal. Chem. 2013, 85, 4686−4693