Anal Bioanal Chem (2010) 396:2235–2249 DOI 10.1007/s00216-009-3443-z

ORIGINAL PAPER

Screening of pesticides in blood with liquid chromatography–linear ion trap mass spectrometry Sylvain Dulaurent & Christian Moesch & Pierre Marquet & Jean-Michel Gaulier & Gérard Lachâtre

Received: 12 November 2009 / Revised: 23 December 2009 / Accepted: 28 December 2009 / Published online: 9 February 2010 # Springer-Verlag 2010

Abstract In clinical or forensic toxicology, general unknown screening procedures are used to identify as many xenobiotics as possible, belonging to numerous chemical classes. We present here a general unknown screening procedure based on liquid chromatography coupled with use of a single linear ion trap mass spectrometer, and focus on the identification of pesticides and/or metabolites in whole blood. After solid-phase extraction (SPE), the compounds of interest were separated using a reversedphase column and identified by the mass spectrometer operated first in the full-scan mass spectrometry (MS) mode, in the positive and negative polarities, followed by MS2 and MS3 scanning of ions selected in data-dependent acquisition. The total scan time was 2.45 s. Two mass spectral libraries (MS2 and MS3), each of 450 spectra, were created for the 320 pesticides and metabolites detected after injection of pure solutions. Robustness of the spectra and matrix effects were studied and were satisfactory for the present application. Detection limits for the 320 compounds were studied by extracting 1 mL spiked blood

S. Dulaurent (*) : C. Moesch : P. Marquet : J.-M. Gaulier : G. Lachâtre Department of Pharmacology and Toxicology, CHU de Limoges, 87042 Limoges, France e-mail: [email protected] C. Moesch : P. Marquet : G. Lachâtre University of Limoges, 87032 Limoges, France P. Marquet INSERM U850, 87025 Limoges, France

at concentrations between 10 µg/L and 10 mg/L. If necessary, it was possible to decrease the detection limits of some compounds by 10–100-fold by scanning MS2 in only one polarity, owing to a shorter total scan time. However, at the same time, the detection specificity decreased as no confirmation could be recorded in the following MS3 scan and no information could be registered in the other polarity. So, in these rare cases, confirmation by another method was required. Keywords General unknown screening . Pesticide . Blood . Linear ion trap

Introduction Clinical and forensic laboratories generally use several screening methods in parallel for the identification of drugs in biological samples: automated immunoassays [1] and separable techniques based on gas chromatography–mass spectrometry (MS) [2], liquid chromatography (LC)– diode-array detection [3], and more recently LC/MS [4–6]. Indeed, LC/MS is described as a suitable method for the identification and quantification of drugs in human samples in the context of clinical and forensic toxicology [7]. Pesticides are responsible for human poisoning and our laboratory frequently receives requests for the determination of pesticides in biological samples [8–10]. LC/MS has been used for the quantitative determination of pesticides in biological samples [9–12]; however, most of the reported multiresidue methods using tandem mass spectrometry in the multiple reaction monitoring mode were dedicated to the determination of pesticides in food or environmental samples and covered only a limited

2236

number of target compounds [13, 14]. The main application of these methods was the determination of target pesticide residues in these last matrices to check whether they fulfilled regulations and/or customers’ requirements. The designation “general unknown screening” (GUS) procedure was previously defined as analytical methods or combinations of analytical methods aimed at detecting and identifying unknown xenobiotics in biological fluids [6]. Numerous GUS or identification methods which have been published concerned drugs and toxic compounds in general [4–6], but no LC/MS screening procedure specifically oriented towards pesticides in biological matrices has been reported so far. We found only one study focused on the GUS of pesticides in human samples using LC, but with UV diode-array detection [15]. It concerned 26 compounds and was applied to pesticide-poisoning cases. A library with UV spectra was developed by these authors. In the few GUS methods published for pesticides in environmental and food matrices [16–18], the authors used a single quadrupole analyzer associated with in-source collision-induced dissociation [16] and a time-of-flight analyzer associated with in-source collision-induced dissociation [17, 18]. The aim of the present study was to develop a GUS method for pesticides in whole blood samples using a single linear ion trap mass spectrometer and to present potential applications in clinical and forensic toxicology.

Experimental Chemicals and reagents Reagents and organic solvents were of analytical grade. Albendazole was selected as the internal standard and was purchased from Dr. Ehrenstorfer (Ausburg, Germany). Methanol and acetonitrile were purchased from Carlo Erba Reagenti (Rodano, Italy), ammonia solution was supplied by VWR (Fontenay-sous-Bois, France), and ammonium formate and formic acid were from Fluka (St Quentin Fallavier, France). Pure water was obtained using a Millipore Direct Q purification system (Saint Quentin en Yvelines, France). Oasis HLB, MAX, and WAX cartridges (3 mL, 60 mg) were purchased from Waters (Saint-Quentin en Yvelines, France). Plexa (3 mL, 60 mg), –NH2 (3 mL, 300 mg) and primary–secondary amine (PSA; 3 mL, 500 mg) cartridges were supplied by Varian (Les Ulis, France). Stock solutions of standard compounds were prepared at 1 g/L, generally in methanol, but for some compounds in water, dichloromethane, or toluene, depending on their solubility.

S. Dulaurent et al.

Liquid chromatography The chromatographic system consisted of a Shimadzu LC-10AD-vp high-pressure pumping system and a SIL HTc autosampler (Shimadzu, Champs sur Marne, France). Chromatographic separation was performed on an Inertsil ODS3 C18, 3µm (100 mm×1-mm inner diameter) column (GL Science, Tokyo, Japan), using a linear gradient of acetonitrile in 10 mM pH 3.0 ammonium formate buffer as the mobile phase (constant flow rate of 50µL/min), programmed as follows: 0–2 min, 20% acetonitrile ; 2–19 min, 20–90% acetonitrile; 19–31 min, 90% acetonitrile; 31–33 min decrease from 90 to 20% acetonitrile; 33–38 min, column equilibration with 20% acetonitrile. All chromatographic solvents were degassed with helium beforehand. Electrospray ionization conditions Mass-spectrometric analyses were conducted using a ThermoFisher Scientific (San Jose, CA, USA) LTQ linear ion trap mass spectrometer equipped with an electrospray ionization source. Source settings were optimized by infusing pesticides at 3µL/min in the LC flow (at 50µL/min) through a “T” coupling system. The pesticides used for optimization [fenpropimorph, carbendazim, pyrimethanil, etofenprox, difenoconazole, dimethylaminosulfotoluidine (dmst), quinalphos, cadusafos, imazalil, atrazine, triforine, abamectin, dimethoate, acetamiprid, diuron, indoxacarb, trifloxystrobin, acibenzolarS-methyl, pirimicarb, and carbofuran] were previously prepared at 10 mg/L in a 10 mM pH 3.0 ammonium formate buffer/acetonitrile mixture (70:30, v/v). These compounds were chosen owing to their wide range of chemical features (molecular weight, class of pesticides, etc.). The main parameter settings in the positive ionization mode were as follows: ion spray voltage, 4,5 kV; sheath gas flow rate, 12; auxiliary gas flow rate, 10; sweep gas flow rate, 12; capillary voltage, 15 V; capillary temperature, 300°C, and tube lens voltage, 65 V. In the negative ionization mode, the ion spray voltage was set to –3 kV, the capillary voltage to –15 V, and tube lens voltage to –65 V. MS conditions The scan mode was chosen with the following conditions: MS/MS2/MS3 in positive ionization mode, followed by positive–negative ion polarity switching, then MS/MS2/ MS3 in negative ionization mode followed by negative– positive ion polarity switching. In these conditions, the total scan time was 2.45 s. The scan mass range was from 100 to 1,000 amu. The normalized collision energy was set at 35, the activation time at 20 ms, the isolation width at 3 amu, and activation Q at 0.26 for the first ion isolation and 0.35 for the second ion isolation. Automatic gain control (AGC)

Screening of pesticides in blood with liquid chromatography–linear ion trap mass spectrometry

was on; therefore, the ion injection time in the ion trap (maximum 300 ms) was varied until approximately 20,000 ions had been trapped. Data-dependent scan In this scan type, the MS2 and MS3 spectra were generated by using the base peak of the parent scan as the precursor ion. This precursor ion was further isolated, excited, and fragmented in the ion trap. So, an MS2 spectrum was generated from the base peak of the MS spectrum and an MS3 spectrum was generated from the base peak of the MS2 spectrum. The intensity threshold was set to 0 and the ion previously selected was excluded for 15 s after one occurrence, entering a 500-ion exclusion list.

2237

spectra registered after an injection with spectra listed in libraries. The results reported by this program are those fulfilling the following criteria: global intensity cutoff, retention time (with defined range), identification criteria (match and reversed match threshold), parent ion and product ion (with defined m/z range), and product ion intensity threshold. These criteria must be chosen by the investigator and manually listed in an Excel sheet, together with the compound name (written in exactly the same way as registered in the NIST library). This application was for the moment used with the MS2 library. If an MS2 spectrum was identified by this automatic method, the corresponding MS3 spectrum was checked manually using NIST MS Search 2.0. Sample preparation

MS2 and MS3 libraries At first, all pesticides were infused individually in the LC/ MS system at 3µL/min by means of an integrated syringe pump to provide a first reference mass spectrum. Then, pesticide reference mass spectra were generated by injecting mixtures of pure solutions of approximately ten compounds at 10 mg/L in 10 mM pH 3.0 ammonium formate buffer/ acetonitrile (70:30, v/v) mixture. The corresponding spectra obtained after chromatographic injection (using the datadependant scan and the electrospray ionization and MS conditions previously described) were compared with these first reference spectra to check their similarities. Then, they were registered in MS2 and MS3 libraries, respectively, using NIST MS Search 2.0. Molecular weight, Chemical Abstracts Service number, as well as the molecular structure, drawn using ISIS Draw 2.4 for each compound, were added manually from Qual Browser (ThermoFisher Scientific) to the library of NIST MS Search 2.0. Retention time was not a parameter available in this library. Library searching Library searching was performed manually using NIST MS Search 2.0. The results were classified according to three parameters: (1) a match factor between the unknown and the library spectra (direct match), (2) a reversed match factor between the unknown and the library spectra ignoring any peaks in the unknown that are not in the library spectrum, and (3) probability. The maximum match or reverse match score is 999. A value of 0 is obtained when spectra have no peaks in common. Automatic spectra comparison ThermoFisher Scientific developed the program ToxID, which allows systematic comparison of all unknown

One milliliter of whole blood (sampled from human patients without known exposure to pesticides and collected from various routine activities in our laboratory) was pipetted into a 10-mL screw-top vial, to which were added sequentially 100µL of a 20 mg/L solution of albendazole in a 10 mM pH 3.0 ammonium formate buffer/acetonitrile (70:30, v/v) mixture and 2 mL of acetonitrile. The mixture was shaken with a vortex for 10 s, then centrifuged at 3,000 rpm for 5 min. The supernatant was then collected in another vial and evaporated at 40°C under a gentle stream of nitrogen, until less than 1 mL remained. The residue was diluted with 1 mL of 0.5 M pH 7 phosphate buffer, and after vortex-mixing the mixture was loaded on an Oasis WAX cartridge, previously conditioned with 2 mL pf methanol and 2 mL of water. For the cleanup step, 1 mL of water was added three times. After drying (20 min), a first elution was carried out with 3 mL of methanol followed by another elution with 3 mL of a 2% ammonia solution in methanol/acetonitrile (20:80, v/v). The two eluates were pooled to obtain only one sample extract and were then evaporated to dryness under a gentle stream of nitrogen (at 40°C). Then the residue was removed using 80 µL of 10 mM pH 3 ammonium formate buffer/acetonitrile (70:30, v/v). Two microliters of this solution was injected into the LC/MS system. Extraction recovery, ion suppression and detection limits Extraction recovery in whole blood was studied for 14 compounds (diethylphosphate, dimethylthiophosphate, formetanate, abamectin, diuron, 2,4-D, carbendazim, atrazine, dimethoate, fluroxypyr, kresoxim-methyl, methomyl, azoxystrobin, and warfarin) by comparing the compound peak area obtained after extraction of spiked whole blood samples (two replicates at 1 or 10 mg/L) with those of compound-free whole blood extracts further spiked with the

2238

compounds. These compounds were chosen owing to their wide range of acidic to basic properties, representing the diversity of our 320 pesticides and metabolites. Peak area in the parent ion chromatogram was used for the determination of extraction recovery. Ion suppression was evaluated by comparing the response of the same 14 pesticides in pure solution with their MS response when spiked into a blank whole blood extract (n=5). The compounds were prepared at appropriate concentrations (1 and 10 mg/L), depending on each compound’s response with the mass spectrometer. Ion suppression was studied following the postextraction spiked method described by Chambers et al. [13]. The detection limits of the scan types described in “MS conditions” were determined for all pesticides after extraction of whole blood spiked at 10µg/L, 100µg/L, 1 mg/L, and 10 mg/L. These limits were defined as the lowest concentration at which the parent ion was automatically selected and further analyzed in the MS2 and MS3 modes (or in the MS2 mode only for those molecules without an MS3 spectrum owing to the low stability of MS2 selected ions), without consideration of the signal-to-noise ratio. Quantification of pesticides in human samples The chromatographic system consisted of Shimadzu LC20-AD pumps and a SIL-20-AC autosampler. A Thermofisher Scientific TSQ Quantum Ultra triple-quadrupole mass spectrometer was used for detection. Ionization was achieved using an electrospray source.

Results An example of a blood sample spiked with pirimicarb at 10 µg/L, then extracted with our method, is given in Fig. 1 (total ion current and parent ion chromatograms; MS, MS2, and MS3 spectra), showing the sensitivity and the selectivity of the present method. The current library comprises 320 compounds (pesticides and metabolites), which represent 450 MS2 spectra and 430 MS3 spectra (positive or negative polarities). Indeed more than one spectrum was registered for some compounds, whether in the positive or the negative polarities, and/or with and without adducts (ammonium, sodium, etc.), and/or with different isotopes (79Br and 81Br, 35Cl and 37Cl), and/ or with daughter ions previously generated in the ion source (accidental ion-source collision-induced dissociation), and/or if an MS2 spectrum presented more than one major ion taken as the base peak at random, which resulted in the same number of MS3 spectra. Our libraries were built and tested according to the recommendations and criteria described by Josephs and

S. Dulaurent et al.

Sanders [19]: determination of standard conditions suitable for a wide range of compounds, spectra robustness, acquisition of spectra in fully automated fashion, efficiency of library searching for closely related compounds, and acquisition of spectra in real-condition concentrations. The libraries fulfilled the criteria proposed by Josephs and Sanders [19]: for example, the robustness of the MS2 spectra was studied during two periods of 5 days over a 5-month period, with five injections per day of a 10 mg/L solution in 10 mM pH 3 ammonium formate buffer/ acetonitrile (70:30, v/v) of six pesticides (benoxacor, bentazone, bupirimate, coumatetralyl, 2,4-MCPA, and phosalone), chosen for their different fragmentation properties. For example, benoxacor and bupirimate had a rich MS2 spectrum, whereas 2,4-MCPA and phosalone had a poor MS2 spectrum; 2,4-MCPA had low m/z ratio fragments, in contrast to coumatetralyl. Statistical results of robustness, calculated using two-way analysis of variance, are presented in Table 1. The intraday coefficient of variation (CV) of the m/z ratio’s relative abundance was always less than 17%, the interday CV was always less than 8%, and the interperiod CV was always less than 23%. Spectrum stability was also tested against concentration with imazalil at 10, 50, 200, 500, 1,000, 2,000, 5,000, and 10,000µg/L, and showed unchanged fragmentation over this range. This pesticide was chosen owing to its large fragmentation (ten largest MS2 peaks: m/z=159, 173, 176, 185, 201, 202, 211, 254, 255, 256). Extraction efficiency was studied using some commercial sorbents packed in different SPE cartridges: Waters Oasis HLB, WAX, and MAX; Varian Plexa, –NH2, and PSA. The extraction recovery yields obtained with these commercial cartridges are presented in Table 2. The Oasis WAX sorbent was the only one which allowed extraction of all the compounds tested and was more powerful than Oasis MAX, another mixed-mode anion-exchange sorbent. Varian PSA and –NH2 were less efficient than Oasis WAX for the extraction of basic compounds (formetanate, carbendazim, atrazine, methomyl), whereas Varian Plexa and Oasis HLB were less efficient than Oasis WAX for acidic compounds (2,4-D, diethylphosphate, dimethylthiophosphate). All sorbents were tested according to their respective generic conditions (pH of the aqueous phase, polarity of the organic phase, etc.). We noticed that among the 14 compounds tested, ion suppression ranged between 1 and 15% (formetanate, abamectin, carbendazim, diuron, atrazine, dimethoate, kresoxim-methyl, methomyl, azoxystrobin, and warfarin) and signal enhancement ranged between 4 and 13% (diethylphosphate, dimethylthiophosphate, 2,4-D, and fluroxypyr) according to the formula reported by Chambers et al. [20]. The detection limits of this GUS method were investigated for the 450 pesticides or metabolites we had as pure

Screening of pesticides in blood with liquid chromatography–linear ion trap mass spectrometry 100

8.43

100

b

NL: 1.37E5 m/z= 238.70-239.70 F: ITMS + c ESI Full ms [100.001000.00] MS

Relative Abundance

0 100

#1158-1176 RT:8.40-8.51AV:4 NL:1.08E6 F:ITMS + c ESI Full ms [100.00-1000.00] 150.14

NL: 5.35E7 TIC MS

a Relative Abundance

Fig. 1 a Top: Total ion current chromatogram of a sample spiked at 10µg/L with pirimicarb. Bottom: Pirimicarb parent ion chromatogram at m/z 239.2 extracted from the top chromatogram. b Mass spectrometry (MS) spectrum that corresponds to the pirimicarb retention time from the chromatogram in a. c MS2 spectrum from the mass spectrum in b. d MS3 spectrum from the mass spectrum in c

pirimicarb parent ion 513.93 321.07

591.35

239.22

0 0

2239

0

35

200

1000

Time (min)

m/z #1166 RT:8.44 AV:1 NL:7.01E3 F:ITMS + c ESI d Full ms3 [email protected] [email protected] [60.00-195.00]

#1165 RT:8.44 AV:1 NL:4.49E4 F:ITMS + c ESI d Full ms2 [email protected] [55.00-250.00]

182.16

100

137.14

100

Relative Abundance

d

Relative Abundance

c

195.17

84.98

150.05

108.98 83.01

71.95

0 60

134.06

166.12

m/z

compounds in the laboratory. With use of our method, some organochloride pesticides (lindane, dicofol, DDT, etc.) were not even detected during direct infusion, some pyrethroids (bifenthrin, deltamethrin, cyfluthrin, etc.) were not detected after on-column injection of pure solutions, whereas other compounds, such as glyphosate and glufosinate, were not detected after extraction. The experimental detection limits of the remaining pesticides (320 compounds), obtained after extraction of whole blood, are presented in Table 3. These limits range from 10µg/L to more than 10 mg/L, with no specific tendency with respect to pesticide classes: for example, pyridaphenthion and fenitrothion, two pesticides belonging to the organophosphate class [21], showed very different limits of detection (10µg/L and more than 10 mg/L, respectively). Using MS/MS2 (without MS3) in only one polarity allowed us to decrease the limits of detection by 10–100-fold for more than half of the compounds tested. The present GUS method was applied for identification of human exposure to pesticides. Two examples are shown in Fig. 2. The first case concerned occupational exposure to fipronil. Fipronil was not detected but fipronil sulfone, its metabolite, was identified by the present GUS method, and was then quantified at 42µg/L in serum using triplequadrupole LC/MS-MS. The second case concerned poisoning with 2,4-D, further quantified at 63 mg/L in whole blood, again using triple-quadrupole LC/MS-MS.

206.04221.21

240

0 60

182.12

167.11

m/z

180

Discussion The LC/MS GUS procedure developed here for pesticides and metabolites in blood, using LC coupled with use of a linear ion trap mass spectrometer, is able to detect and specifically identify more than 320 different pesticides and metabolites. The detection parameters (source, ion-trap settings) were optimized to maximize the signal-to-noise ratio of each compound, but also to find a range of values convenient for most compounds close to these ideal values so we could choose default settings. The duty cycle on the LTQ mass spectrometer in the data-dependent mode (positive–negative switching, MS2, MS3) used here is similar to that presented in a previous paper [22]. The authors stated that this type of experiment is an extremely powerful approach for the identification of unknown compounds in complex samples. They also presented the LTQ system working with alternation of prescans (low resolution, allowing counting of ions to set the ion-trap filling time) and analytical scans (normal resolution). In the operation mode, the scans were performed with AGC automatically set by the system, to limit space charging effects in the ion trap and to obtain reproducible spectra, as in the present study. Indeed, space charge effects decrease the ion trap performance, by limiting the resolution, mass accuracy, sensitivity, and

2240

S. Dulaurent et al.

Table 1 Robustness of the MS2 spectra of benoxacor, bentazone, bupirimate, coumatetralyl, 2,4-MCPA, and phosalone, studied during two periods of 5 days over a 5-month period, with five injections per day of a 10 mg/L solution Compound

m/z (amu)

Mean relative abundance (%)

Intraday CV (%)

Interday CV (%)

Interperiod CV (%)

2,4-MCPA

141 155 120 149 150 188 196 202 214 224 242 132 133

100.0 12.6 16.5 100.0 23.5 12.6 27.6 11.2 13.4 16.5 9.4 9.1 6.6

0.0 6.8 7.3 0.0 5.5 6.6 5.7 7.4 8.3 6.5 9.1 11.2 12.9

0.0 1.8 1.9 0.0 0.4 2.8 1.8 3.4 2.1 2.0 2.6 3.5 3.4

0.0 1.6 1.3 0.0 2.2 1.4 0.6 2.1 0.8 1.3 7.9 1.6 2.0

175 176 197 198 239 150 166 192 208 209 210 237 238 272 135 141 187 219

61.6 7.0 100.0 9.1 16.3 12.1 36.7 14.8 16.7 21.0 37.6 100.0 14.8 45.1 10.1 41.0 13.3 19.2

4.6 15.1 0.0 16.6 8.2 11.0 5.9 10.7 8.3 10.7 7.9 0.0 7.3 7.3 10.2 6.2 10.9 7.6

0.8 3.0 0.0 4.6 2.3 3.5 0.8 2.8 3.6 3.2 3.3 0.0 2.6 2.6 3.9 3.6 2.4 3.0

0.6 3.1 0.0 3.5 22.7 1.4 1.0 1.4 2.3 2.4 2.0 0.0 3.2 3.2 2.5 1.6 2.4 3.9

247 248 291 182 322

100.0 19.7 29.0 26.9 100.0

0.0 9.5 7.1 7.9 0.0

0.0 7.2 1.6 0.5 0.0

0.0 3.2 15.7 5.8 0.0

Benoxacor

Bentazone

Bupirimate

Coumatetralyl

Phosalone

CV coefficient of variation

dynamic range [23]. The AGC function or its equivalent for a different mass spectrometer vendor was previously described as suitable for screening purposes, owing to the variation of compound concentrations and background noise in biological samples [4]. In our study, the scan power was n=3 (to MS3), which, associated with two positive–negative switches generated a total scan time of 2.45 s. In this GUS technique, our aim was to use the shortest scan time possible in order to select

the highest number of parent ions under a chromatographic peak and in this way to obtain the best detection efficiency. Our method allowed the collection of positive and negative information after a single injection, as opposed to the previous GUS method developed in our laboratory using a different type of instrument [4], which had a total scan time of 1.36 s but required two injections to cover both polarities. Moreover, the two positive–negative switches which took 1 s with the LTQ system could be decreased to

Screening of pesticides in blood with liquid chromatography–linear ion trap mass spectrometry

2241

Table 2 Extraction recovery of 14 pesticides and metabolites in whole blood using different commercial solid-phase-extraction cartridges Recovery (%) (n=2) Oasis WAX

Oasis MAX

Oasis HLB

Varian Plexa

Varian–NH2

Varian PSA

2,4-D Abamectin Atrazine Azoxystrobin Carbendazim

42 82 85 100 100

87 49 87 81 97

35 100 100 100 100

7 76 85 77 79

14 24 56 100 100

4 9 14 19 11

Diethylphosphate Dimethoate Dimethylthiophosphate Diuron Fluroxypyr Formetanate Kresoxim-methyl Methomyl Warfarin

4 94 24 100 100 31 100 81 100

0 83 0 87 88 25 68 39 76

0 100 0 100 0 100 78 100 91

0 90 0 78 0 58 67 35 54

26 56 28 67 0 0 80 0 100

11 14 9 18 0 0 22 0 21

0.140 s by using the new LTQ XL system now available. Obtaining MS3 spectra was also time consuming (2× 250 ms), but they allowed false-positive results to be avoided. An example of a false-positive result with n=2, uncovered with n=3, is shown in Fig. 3 for asulam. In this case, the retention times and MS2 spectra of asulam and the unknown compound were identical. The MS3 spectrum was the only way to discriminate the two compounds, improving the method specificity. As described in “Results,” the intraday, interday, and interperiod CVs of the relative intensity results were satisfactory for the present application. We specify that over the 5-month period, the LTQ system was used for the analysis of blood samples as much as for the recording of spectra derived from injections of standard solutions of pesticides. Preventive and curative cleanup were performed once a week to ensure a satisfactory performance. Most of the pesticides ionizable with the electrospray source employed had neutral or acidic properties (carbamates, organophosphates, phenoxyalcanoic acids, pyrethroids metabolites, benzimidazoles, etc.), except for some specific compounds such as atrazine or formetanate, which are basic. For this reason, we focused on mixed-mode anion exchange (Oasis WAX, MAX) and hydrophobic sorbents (Oasis HLB, Varian Plexa) for SPE. Owing to the lack of data concerning extraction of a wide range of pesticides from human biological fluids, we also tested the extraction methods described by Schenk et al. [24] for pesticides in fruits and vegetables. They reported that aminopropyl (–NH2) and PSA phases provided excellent cleanup of fresh fruit and vegetable extracts for multiresidue pesticide analysis in a wide range of polarity

and molecular weight. Among these sorbents, the mixedmode SPE sorbent Oasis WAX, which combined an effective reversed-phase chemistry with ion-exchange sites, was the most efficient, probably owing on the one hand to its reversed-phase mechanism with azoxystrobin, abamectin, warfarin, and kresoxim-methyl, and on the other hand to its anion-exchange mechanism with 2,4-D, fluroxypyr, diethylphosphate, and dimethylthiophosphate. Moreover, we noticed that the present procedure allowed the extraction of other compounds such as the basic pesticides formetanate, diuron, and atrazine, and also other unclassified compounds such as dimethoate, methomyl, and carbendazim. The extraction procedure on the Oasis WAX sorbent was inspired, with slight modifications, by the procedure of Fontanals et al. [25]. The authors presented SPE conditions for weak anion exchange that allowed extraction with recoveries of more then 94% for both an acidic compound (2-naphthalenesulfonic acid; pKa <1) and a basic compound (amitriptyline; pKa 9.4). None of the sorbents were able to extract with efficiency diethylphosphate and dimethylthiophosphate, two dialkylphosphates [10]: it is well known that dimethylphosphate is a hydrophilic compound which is hardly extracted from the aqueous phase, and so with a low recovery [10] and diethylphosphate has similar physicochemical properties. Even though matrix effects might have less influence on the results from qualitative than from quantitative analyses, they were investigated here. The postextraction spiked approach used has the advantage of being quantitative, as compared with the method consisting in the infusion of a working solution and injection of a blank sample in parallel

2242

S. Dulaurent et al.

Table 3 Experimental detection limits of pesticides and metabolites in whole blood Compound

Detection limit in the dualpolarity, MS3 mode

Detectability at 10 µg/L in the singlepolarity, MS2 mode

1-(3,4-Dichlorophenyl)-3methylurea 1-(3,4-Dichlorophenyl)urea 1-(3-Chloro-4-methylphenyl)urea 1-(4-Chlorophenyl)urea 1-Naphthyl acetic acid 2,3,6-Trichlorobenzoic acid 2,4,5-T

100 µg/L

Yes

1 mg/L 100 µg/L 1 mg/L >10 mg/L >10 mg/L 1 mg/L

Yes Yes No ND ND No

2,4-D 2,4-DB 2,4-Dimethylaniline 2,6-Diethylaniline 2-Isopropyl-6-methyl-4-pyrimidinol 3,5,6-Trichloro-2-pyridinol 3-Chloro-p-anisidine 3-Hydroxycarbofuran 3-Methyl-4-nitrophenol 3-Phenoxybenzoïc acid 4-Fluoro-3-phenoxybenzoïc acid 4-Isopropylaniline 5-Hydroxyimidacloprid 6-Chloronicotinic acid Abamectin Acephate Acetamiprid Acibenzolar-S-methyl Aclonifen

1 mg/L 1 mg/L 10 mg/L >10 mg/L 10 mg/L 10 mg/L 10 mg/L 1 mg/L 10 µg/L 1 mg/L 100 µg/L >10 mg/L 100 µg/L 100 µg/L 10 mg/L 1 mg/L 100 µg/L >10 mg/L 10 mg/L

No No ND ND ND ND ND ND ND No Yes ND No No ND No Yes ND ND

Acrinathrin Alachlor Aldicarb Aldicarb sulfone Aldicarb sulfoxide Allethrin Allidochlor Amitraz [expressed in N(2,4-dimethylphenyl) formamide] Asulam Atraton-desisopropyl

10 mg/L 1 mg/L 10 mg/L 100 µg/L 1 mg/L 100 µg/L 1 mg/L 100 µg/L

ND Yes ND No No No No No

1 mg/L 100 µg/L

No No

Atrazine Atrazine-desethyl Atrazine-desethyldesisopropyl Atrazine-desisopropyl Azinphos-ethyl Azinphos-methyl Azoxystrobin Bendiocarb

10 mg/L 10 mg/L >10 mg/L 100 µg/L 100 µg/L 1 mg/L 10 µg/L 10 mg/L

ND ND ND No Yes No Yes ND

Table 3 (continued) Compound

Detection limit in the dualpolarity, MS3 mode

Detectability at 10 µg/L in the singlepolarity, MS2 mode

Benfuracarb (expressed in carbofuran) Benoxacor Bentazone Benzoximate Benzthiazuron Bitertanol

100 µg/L

Yes

1 mg/L 100 µg/L 100 µg/L 100 µg/L 100 µg/L

No Yes No Yes No

Boscalid Bromacil Bromadiolone Bromophos-ethyl Bromophos-methyl Bromopropylate Bromoxynil Bupirimate Buprofezin Cadusafos Carbaryl Carbendazim Carbetamide Carbofuran Carbosulfan (expressed in carbofuran) Carboxin 2-Chloro-N-[2,6-diethylphenyl] acetamide CGA321113 Chloralose Chlorbromuron Chlordecone Chlorfenvinphos

100 µg/L 100 µg/L 1 mg/L >10 mg/L 1 mg/L 10 mg/L 100 µg/L 10 µg/L 10 µg/L 100 µg/L 100 µg/L 100 µg/L 100 µg/L 100 µg/L 100 µg/L

Yes Yes Yes ND No ND No Yes Yes Yes No No Yes Yes Yes

100 µg/L 100 µg/L

Yes Yes

100 µg/L 1 mg/L 100 µg/L 10 mg/L 100 µg/L

Yes No Yes ND Yes

Chlorfluazuron Chlormequat Chlorophacinone Chlorotoluron Chlorpyrifos-ethyl Chlorpyrifos-methyl Chlorsulfuron Chlorthion Chlorthiophos Clethodim Clopyralid Coumachlor Coumaphos Coumatetralyl Crimidine Cyanazine

10 mg/L >10 mg/L 10 mg/L 100 µg/L 10 mg/L 10 mg/L 10 mg/L 10 mg/L 10 mg/L 1 mg/L 10 mg/L 10 µg/L 100 µg/L 10 µg/L 100 µg/L >10 mg/L

ND ND ND No ND ND ND ND ND No ND Yes Yes Yes No ND

Screening of pesticides in blood with liquid chromatography–linear ion trap mass spectrometry Table 3 (continued)

2243

Table 3 (continued)

Compound

Detection limit in the dualpolarity, MS3 mode

Detectability at 10 µg/L in the singlepolarity, MS2 mode

Compound

Detection limit in the dualpolarity, MS3 mode

Detectability at 10 µg/L in the singlepolarity, MS2 mode

Cycloate Cycloheximide Cymiazole Cymoxanil Cyproconazole Cyprodinil Daminozide Demeton-O and demeton-S Demeton-S-methyl Demeton-S-methyl sulfoxide Demeton-S-methylsulfone Desmetryn Dialifos Diazinon

>10 mg/L 1 mg/L 100 µg/L 10 mg/L 100 µg/L 100 µg/L >10 mg/L 1 mg/L >10 mg/L 1 mg/L 100 µg/L 10 µg/L >10 mg/L 10 mg/L

ND No Yes ND No Yes ND No ND Yes Yes Yes ND ND

Endosulfan lactone Epoxiconazole Ethephon Ethidimuron Ethiofencarb Ethion Ethoprophos Ethoxyquin Etofenprox Etoxazole Fenamiphos Fenarimol Fenazaquin Fenbuconazole

1 mg/L 100 µg/L >10 mg/L 1 mg/L 10 mg/L 100 µg/L 1 mg/L 10 mg/L 10 mg/L 100 µg/L 1 mg/L 100 µg/L 10 mg/L 100 µg/L

No Yes ND No ND No No ND ND No ND Yes ND No

Dibutylamine Dicamba Dichlofluanid Dichlorprop Dichlorvos Diethofencarb Diethylphosphate Diethylthiophosphate Difenoconazole Diflubenzuron Diflufenican Dimefuron Dimethenamid Dimethoate Dimethomorph Dimethylaminosulfotoluidine Dimethyldithiophosphate Dimethylphosphate

1 mg/L 100 µg/L >10 mg/L 100 µg/L >10 mg/L 100 µg/L 1 mg/L 1 mg/L 100 µg/L 10 mg/L 100 µg/L 100 µg/L >10 mg/L 100 µg/L 100 µg/L 1 mg/L 1 mg/L >10 mg/L

No No ND No ND Yes No No Yes ND Yes Yes ND Yes Yes No No ND

Fenhexamid Fenitrothion Fenobucarb Fenoprop Fenoxycarb Fenpropimorph Fenpyroximate Fenthion Fentichlor Fenuron Fipronil Fipronil sulfone Flonicamid Fluazinam Fludioxonil Flufenoxuron Fluometuron Fluquinconazole

100 µg/L >10 mg/L 100 µg/L 100 µg/L 1 mg/L 10 µg/L 100 µg/L 100 µg/L 100 µg/L 100 µg/L 10 µg/L 10 µg/L >10 mg/L 10 µg/L 100 µg/L 10 mg/L 10 µg/L >10 mg/L

No ND Yes No No ND No No No Yes Yes Yes ND Yes No ND Yes ND

Dimethylthiophosphate Dinocap Dinoterb Dioxathion Diphacinone Diphenylamine Dithianon Diuron Dodemorph Dodine Endosulfan alcohol Endosulfan-alpha Endosulfan-beta

1 mg/L 100 µg/L 100 µg/L 10 mg/L 10 mg/L 10 mg/L >10 mg/L 100 µg/L 10 µg/L 100 µg/L 10 mg/L 10 mg/L 10 mg/L

ND No No ND ND ND ND Yes Yes No ND ND ND

Flurochloridone Fluroxypyr Fluroxypyr-1-methylheptyl ester Flusilazole Flutriafol Fonofos Formetanate Fosetyl-aluminum Furathiocarb (expressed in carbofuran)

10 mg/L 1 mg/L 10 mg/L 10 µg/L 100 µg/L 10 mg/L 1 mg/L >10 mg/L 10 mg/L

ND No ND Yes No ND No ND ND

Glufosinate Glyphosate Heptenophos

>10 mg/L >10 mg/L 100 µg/L

ND ND No

2244

S. Dulaurent et al.

Table 3 (continued)

Table 3 (continued)

Compound

Detection limit in the dualpolarity, MS3 mode

Detectability at 10 µg/L in the singlepolarity, MS2 mode

Compound

Detection limit in the dualpolarity, MS3 mode

Detectability at 10 µg/L in the singlepolarity, MS2 mode

Hexaconazole Hexaflumuron Hexazinone Hexythiazox Imazalil Imidacloprid Indoxacarb Ioxynil Iprodione Isofenphos Isoprocarb Isoproturon Isoxaben Kresoxim-methyl

100 µg/L 10 mg/L 10 µg/L 10 mg/L 100 µg/L 100 µg/L 10 mg/L 100 µg/L 10 mg/L 100 µg/L 100 µg/L 10 µg/L 100 µg/L 100 µg/L

Yes ND Yes ND Yes No ND Yes ND Yes No Yes Yes Yes

Monolinuron Monuron Myclobutanil N-(2,4-Dimethylphenyl)formamide N-2,4-Dimethylphenyl-N′methylformamide Naled

100 µg/L 100 µg/L 100 µg/L 100 µg/L >10 mg/L

Yes No No No ND

>10 mg/L

ND

Lenacil Linuron Lufenuron Malathion Malathion monocarboxylic acid MCPA MCPB Mecarbam mecoprop Mepanipyrim Merphos Metalaxyl Metaldehyde Metamitron Metazachlor Methacrifos Methamidophos Methfuroxam

100 µg/L 100 µg/L 10 mg/L 100 µg/L 10 mg/L 100 µg/L 10 mg/L 100 µg/L 100 µg/L 100 µg/L 100 µg/L 10 µg/L >10 mg/L 100 µg/L 10 µg/L >10 mg/L >10 mg/L 100 µg/L

Yes Yes ND Yes ND No ND Yes No Yes Yes Yes ND Yes Yes ND ND Yes

Neburon Norflurazon Nuarimol Omethoate Oryzalin Oxadiazon Oxamyl Paranitrophenol Paraoxon-ethyl Paraoxon-methyl Parathion-ethyl Penconazole Pendimethalin Pentachlorophenol Phenmedipham Phenothrin Phorate Phosalone Phosmet

100 µg/L 10 µg/L 100 µg/L 10 mg/L 10 µg/L 10 mg/L 1 mg/L 100 µg/L 100 µg/L 10 mg/L 10 mg/L 10 mg/L 100 µg/L 10 mg/L 100 µg/L >10 mg/L >10 mg/L 100 µg/L >10 mg/L

Yes Yes Yes ND Yes ND No No Yes ND ND ND No ND Yes ND ND No ND

Methidathion Methiocarb Methomyl Methomyl-oxime Methoxyfenozide Methyl-N-(3-hydroxyphenyl)carbamate Metobromuron Metolachlor Metolcarb Metoxuron Metsulfuron-methyl

100 µg/L 100 µg/L 1 mg/L 10 mg/L 10 µg/L 10 µg/L

No Yes No ND Yes Yes

10 µg/L 1 mg/L 100 µg/L 10 mg/L 10 mg/L

Yes No Yes ND ND

Phosphamidon Picloram Piperonyl butoxide Pirimicarb Pirimiphos-ethyl Pirimiphos-methyl Prochloraz Procymidone Promecarb Prometryn Propachlor Propamocarb Propaquizafop Propargite Propazine Propham Propiconazole Propoxur

100 µg/L 10 mg/L 100 µg/L 10 µg/L 10 mg/L 10 mg/L 100 µg/L 1 mg/L 100 µg/L 10 µg/L 100 µg/L 1 mg/L 100 µg/L 10 mg/L 10 µg/L >10 mg/L 100 µg/L 100 µg/L

Yes ND Yes Yes ND ND Yes No Yes Yes Yes No Yes ND Yes ND Yes Yes

Mevinphos Monocrotophos

1 mg/L 100 µg/L

Yes No

Prothiophos pymetrozine

10 mg/L 100 µg/L

ND No

Screening of pesticides in blood with liquid chromatography–linear ion trap mass spectrometry Table 3 (continued)

2245

Table 3 (continued)

Compound

Detection limit in the dualpolarity, MS3 mode

Detectability at 10 µg/L in the singlepolarity, MS2 mode

Compound

Detection limit in the dualpolarity, MS3 mode

Detectability at 10 µg/L in the singlepolarity, MS2 mode

Pyraclostrobin Pyrazophos Pyridaben Pyridaphention Pyridate Pyrifenox Pyrimethanil Quinalphos Quinclorac Quinmerac Quizalofop-ethyl Rotenone Simazine Spinosyn A

100 µg/L 10 µg/L 10 mg/L 10 µg/L 1 mg/L 100 µg/L 100 µg/L 100 µg/L 1 mg/L 1 mg/L 100 µg/L 10 mg/L 100 µg/L 100 µg/L

Yes Yes ND Yes No Yes No Yes No Yes No ND No No

Triflumizole Triflumuron Trifluralin Triforine Vamidothion Warfarin

100 µg/L 10 mg/L >10 mg/L 10 mg/L 100 µg/L 100 µg/L

Yes ND ND ND Yes Yes

Spinosyn D Strychnine Sulfotep Tebuconazole Tebufenozide Tebufenpyrad Teflubenzuron Temephos Terbufos Terbumeton Terbuthylazine Terbutryn Tetraconazole Tetrahydrophtalimide Thiabendazole Thiacloprid Thiamethoxam Thiobencarb

100 µg/L 100 µg/L 10 mg/L 1 mg/L 100 µg/L 10 µg/L 10 mg/L 10 mg/L >10 mg/L 10 µg/L 100 µg/L 10 µg/L 1 mg/L 10 mg/L 100 µg/L 100 µg/L 100 µg/L 100 µg/L

No Yes ND No Yes Yes ND ND ND Yes Yes Yes Yes ND No Yes No No

Thiocyclam Thiodicarb Thiometon Thiophanate-methyl Thiram Triadimefon Triadimenol Triallate Triazamate Trichlorfon Triclopyr Tridemorph Trietazine Trifloxystrobin

10 mg/L >10 mg/L >10 mg/L 100 µg/L >10 mg/L 100 µg/L 100 µg/L 10 mg/L 100 µg/L 100 µg/L 1 mg/L 100 µg/L 100 µg/L 10 µg/L

ND ND ND Yes ND Yes No ND Yes No No No Yes Yes

ND Not determined

with a tee. This last method only allows the identification of chromatographic regions where a compound would be most susceptible to loss or increase of signal. Our matrix effect results were satisfactory for the present application and similar to previously described results concerning drugs (e.g., atenolol, risperidone, pseudoephedrine, imipramine, amitriptyline) [20]. This similarity in the results is more related to the extraction method used than to the type of molecules analyzed, mixed-mode ion-exchange extraction being shown as the most effective sample preparation technique with minimal matrix effects from biological samples and excellent recoveries for a range of polar and nonpolar analytes [20]. Using MS2 in only one polarity scan versus MS2 and MS3 scans in both polarities improved the limits of detection for the majority of the compounds tested. We used this short scan mode for applications requiring the detection of low concentrations of pesticides in an unknown sample if we were able to confirm their presence and identify them using another complementary method (generally in the multiple reaction monitoring mode with a triple-quadrupole instrument). The lowering of the detection limits was obtained here at the cost of a lower specificity, but this lowering should now be possible without sacrificing either the MS3 level or one ion polarity, owing to very recent hardware improvements. Hopefully, the whole +/- MS/MS2/MS3 scheme will be possible within a reasonable time in the future. As previously mentioned, the detection limits were determined as the lowest concentration at which the parent ion was selected, and then were analyzed both in MS2 and in MS3 mode. During our detection limit determination experiments, we noticed that once a parent ion had been selected, an MS2 spectrum was obtained successfully but then an MS3 spectrum was not always obtained, perhaps owing to a too low MS2 base peak signal. In this last case (MS3 spectrum not obtained), the detection limit studied

2246

0 100

19.57

450.89

100

A-2

NL: 6.79E4 m/z= 450.40-451.40 F: ITMS - c ESI Full ms [100.00-1000.00] MS

Relative Abundance

A-1 Relative Abundance

#2693-2750RT:19.32-19.70AV:10 NL:2.04E4 F:ITMS - c ESI Full ms [100.00-1000.00]

NL: 3.79E7 TIC MS

100

452.93

35

0

0

1000

Time (min)

m/z

#2719 RT:19.49 AV:1 NL:1.84E4

#2720 RT:19.50AV:1 NL:1.46E4

F:ITMS - c ESI d Full ms2 [email protected] [115.00-465.00]

100

F:ITMS - c ESI d Full ms3 [email protected] [email protected] [145.00-425.00]

100

414.92

282.00

A-4

Relative Abundance

A-3 Relative Abundance

415.99 0

450

m/z

Relative Abundance

B-1

12.56

100

219.06

161.11

0

#1761 RT:12.49AV:1 NL:6.52E5

1000

m/z

#1762 RT:12.49AV:1 NL:5.03E4

F:ITMS - c ESI d Full ms2 [email protected] [50.00-230.00]

F:ITMS - c ESI d Full ms3 [email protected] [email protected] [50.00-175.00]

161.03

100

Relative Abundance

B-4

161.00

175.03 0 m/z

was rejected and the upper concentration (tenfold factor) was retained as the detection limit. To the best of our knowledge, we found only one study focused on the screening of pesticides in human samples using LC [15], also applied to the investigation of pesticidepoisoning cases. However, this method used UV diodearray detection and concerned 26 compounds only. For this

125.13

100

B-3

Relative Abundance

400

m/z

221.02

35

Time (min)

346.01

B-2

NL: 1.01E7 m/z= 218.50-219.50 F: ITMS - c ESI Full ms [100.00-1000.00] MS

0

0

#1697-1863 RT: 12.06-13.16 AV: 28 NL: 1.50E6 F:ITMS - c ESI Full ms [100.00-1000.00]

NL: 6.38E7 TIC MS

100

0 100

283.05

Relative Abundance

Fig. 2 A-1: Total ion current chromatogram of an unknown serum sample (top) and fipronil sulfone parent ion chromatogram at m/z 450.9 extracted from the top chromatogram (bottom). A-2: MS spectrum that correspond to the fipronil sulfone retention time from the chromatogram in A-1. A-3: MS2 spectrum from the mass spectrum in A-2. A-4: MS3 spectrum from the mass spectrum in A-3. B-1: Total ion current chromatogram of an unknown blood sample (top) and 2,4-D parent ion chromatogram at m/z=219.0 extracted from the top chromatogram (bottom). B-2: MS spectrum that corresponds to the 2,4-D retention time from the chromatogram in B-1. B-3: MS2 spectrum from the mass spectrum in B-2. B-4: MS3 spectrum from the mass spectrum in B-3

S. Dulaurent et al.

89.17 220

0 m/z

160

purpose, a library was developed with UV absorption spectra. The detection limits were estimated after direct injection of diluted stock solutions (10µL of solutions at 1 mg/L) and defined by a signal-to-noise ratio of 5 or more. Owing to the lack of both an extraction step and a biological matrix in the injected extract, these limits were not comparable with those in the present study. The main

Screening of pesticides in blood with liquid chromatography–linear ion trap mass spectrometry NL: 2.78E7 TIC MS

100

NL: 3.50E5 m/z= 247.50248.50 MS

Relative Abundance

A-2

Relative Abundance

A-1

11.76

100

0

0

30

Time (min)

#1579 RT:11.72AV:1 NL:1.19E5 F:ITMS + c ESI d Full ms2 [email protected] [ 60.00-260.00]

230.93

100

30

Time (min)

#1580 RT:11.73AV:1 NL:1.34E5 F:ITMS + c ESI d Full ms3 [email protected] [email protected] [ 55.00-245.00]

155.87

100

A-4 Relative Abundance

A-3 Relative Abundance

Fig. 3 A-1: Total ion current chromatogram of a stock solution of asulam at 10 mg/L in 10 mM pH 3 ammonium formate/acetonitrile (70:30, v/v). A-2: Parent ion chromatogram of asulam (ammonium adduct) at m/z=248 extracted from the chromatogram in A-1 (retention time, 11.76 min). A-3: MS2 spectrum of asulam (ammonium adduct). A-4: MS3 spectrum of asulam (ammonium adduct). B-1: Total ion current chromatogram of an unknown sample. B-2: Parent ion chromatogram at m/z=248 extracted from the chromatogram in A-2 (retention time, 11.64 min). B-3: MS2 spectrum corresponding to the unknown compound eluted at 11.64 min. B-4: MS3 spectrum of the unknown compound eluted at 11.64 min

2247

0

60

260

m/z NL: 1.90E7 TIC MS

100

0

11.64

NL: 8.60E5 m/z= 247.50248.50 MS

B-2

0

30

Time (min)

#1611 RT:11.48AV:1 NL:3.12E4 F:ITMS + c ESI d Full ms3 [email protected] [email protected] [ 55.00-245.00]

#1610 RT:11.47AV:1 NL:7.54E4 F:ITMS + c ESI d Full ms2 [email protected] [ 60.00-260.00]

231.10

100

240

m/z

100

30

Time (min)

153.17

100

B-4

Relative Abundance

B-3

Relative Abundance

60

Relative Abundance

Relative Abundance

B-1

0

135.17 171.15 195.12

0

60

m/z

advantage of MS with UV detection is the greater specificity of a parent mass spectrum as compared with a UV spectrum. During our experiments, we noticed a system limit, owing to the LTQ dynamic exclusion principle: if an ion was excluded in one polarity, ions with the same m/z ratio in the opposite polarity were also excluded. For example,

260

0

60

m/z

240

we encountered a few cases with ions excluded in one polarity owing to an interfering compound, resulting in parent ions of compounds of interest (pesticides or metabolites) in the other polarity not being scanned, even when they were the base peaks in their scan event. Another limit was during injection pesticide mixes at 10 mg/L in 10 mM pH 3 ammonium formate buffer/acetonitrile (70:30, v/v), atrazine

2248

S. Dulaurent et al.

though an inconvenience such as the one previously described can occur.

and dmst, whose parent ion m/z ratios are only separated by 1 amu, were not totally separated in our chromatographic conditions (Fig. 4). The base peak of the dmst MS2 spectrum, at m/z=174, actually belonged to atrazine. As a consequence, the so-called dmst MS3 spectrum was indeed the atrazine MS3 spectrum. This problem occurred down to a 1.5 isolation width value. However, at this value or less, the robustness of several spectra decreased dramatically and was not sufficient for correct library identification. The best compromise for the isolation width was a value of 3, even

The results obtained with the present GUS method are satisfactory in terms of sensitivity, selectivity, and specificity for the identification of unknown pesticides in a complex matrix. For instance, our method is focused on the identification

100

a

NL: 2.76E6 m/z=

214.60-215.60 F: ITMS + c ESI Full ms [ 100.00-1000.00]

Relative Abundance

dmst 11.62

0 100

11.23

b

0

0

NL: 1.29E7 m/z=

215.60-216.60 F:

atrazine

4

8

12

ITMS + c ESI Full ms [ 100.00-1000.00]

16 Time (min)

20

ITMS + c ESI d Full ms2 [email protected] [ 50.00-230.00]

100

24

28

151.13

c 106.11

0 100

dmst

ITMS + c ESI d Full ms2 [email protected] [ 50.00-230.00]

Relative Abundance

Fig. 4 a: Dimethylaminosulfotoluidine (dmst) parent ion chromatogram (m/z 215.1) obtained after injection of a 10 mg/L dmst stock solution. b: Atrazine parent ion chromatogram (m/z 216.1) obtained after injection of a 10 mg/L atrazine stock solution. c: dmst MS2 spectrum after injection of a 10 mg/L dmst stock solution. d: Atrazine MS2 spectrum after injection of a 10 mg/L atrazine stock solution. e: dmst MS2 spectrum after injection of a 10 mg/L stock solution of ten pesticides that contained atrazine and dmst

Conclusions

174.07

d atrazine

216.18 0 100

ITMS + c ESI d Full ms2 [email protected] [ 50.00-230.00]

173.99

e dmst

151.02

105.89 216.05 0

50

100

150 m/z

200

Screening of pesticides in blood with liquid chromatography–linear ion trap mass spectrometry

of pesticides in whole blood, i.e., mainly for forensic (postmortem) applications. However, it could be also compatible with other matrices such as serum, plasma, and urine, allowing clinical applications. Apart from the recent instrument upgrade with decreased polarity switch time, the software improvements that should increase the efficiency of this technique, depending on the mass spectrometer manufacturer, are as follows: (1) dynamic background subtraction, described in previously published works with other instruments [4]; (2) dynamic exclusion list taking into account each polarity separately; and (3) the possibility of combining the information obtained from MS2 and MS3 spectra.

Acknowledgements The authors thank E. Genin and C. Dabadie (ThermoFisher Scientific France) for their technical support.

References 1. Fucci N, De Giovanni N, De Giorgio F, Liddi R, Chiarotti M (2006) Forensic Sci Int 156:102–105 2. Maurer HH (1992) J Chromatogr 580:3–41 3. Schönberg L, Grobosch T, Lampe D, Kloft C (2007) J Anal Toxicol 31:321–327 4. Sauvage FL, Saint-Marcoux F, Duretz B, Deporte D, Lachâtre G, Marquet P (2006) Clin Chem 52:1735–1742 5. Saint-Marcoux F, Lachâtre G, Marquet P (2003) J Am Soc Mass Spectrom 14:14–22 6. Marquet P (2002) Ther Drug Monit 24:125–33 7. Maurer HH (2007) Anal Bioanal Chem 388:1315–1325

2249

8. Dulaurent S, Gaulier JM, Marquet P, Lachâtre G (2006) Acta Clin Belg 61:71–76 9. Dulaurent S, Gaulier JM, Baudel JL, Fardet L, Maury E, Lachâtre G (2008) Forensic Sci Int 176:72–75 10. Dulaurent S, Saint-Marcoux F, Marquet P, Lachâtre G (2006) J Chromatogr B 831:223–229 11. Hernández F, Sancho JV, Pozo OJ (2005) Anal Bioanal Chem 382:934–946 12. Bicker W, Lämmerhofer M, Linder W (2005) J Chromatogr B 822:160–169 13. Lehotay SJ, De Kok A, Hiemstra M, Van Bodegraven P (2005) J AOAC Int 88:595–614 14. Ortelli D, Edder P, Corvi C (2004) Anal Chim Acta 520:33–45 15. Mori H, Sato T, Nagase H, Takada K, Yamazaki F (1998) Jpn J Toxicol Environ Health 44:182–194 16. Schreiber A, Efer J, Engenwald W (2000) J Chromatogr A 869:411–425 17. Garcia-Reyes JF, Molina-Diaz A, Fernandez-Alba AR (2007) Anal Chem 79:307–321 18. Garcia-Reyes JF, Ferrer I, Thurman EM, Molina-Diaz A, Fernandez-Alba AR (2005) Rapid Commun Mass Spectrom 19:2780–2788 19. Josephs JL, Sanders M (2004) Rapid Commun Mass Spectrom 18:743–759 20. Chambers E, Wagrowski-Diehl DM, Lu Z, Mazzeo JR (2007) J Chromatogr B 852:22–34 21. World Health Organization (2005) The WHO recommended classification of pesticides by hazard and guidelines to classification 2004. Word Health Organization, Geneva 22. Dear GJ, James AD, Sarda S (2006) Rapid Commun Mass Spectrom 20:1351–1360 23. Schwartz JC, Senko MW, Syka JEP (2002) J Am Soc Mass Spectrom 13:659–669 24. Schenk FJ, Lehotay SJ, Vega V (2002) J Sep Sci 25:883–890 25. Fontanals N, Trammell BC, Galià M, Marcé RM, Iraneta PC, Borrull F, Neue UD (2006) J Sep Sci 29:1622–1629