J. Agric. Food Chem. 1992, 40, 458-461
458
Application of Nuclear Magnetic Resonance Spectroscopy to the Identification and Quantitation of Pesticide Residues in Soil Michael E. Krolski,' Laura L. Bosnak, and John J. Murphy Agricultural Chemicals Division, Mobay Corporation, Mobay Research Park, 17745 South Metcalf Avenue, Stilwell, Kansas 66085
The application of high-resolution Fourier-transform nuclear magnetic resonance (NMR) spectroscopy to the identification of metabolites of the phosphorodithioate ester pesticide sulprofos (Bolstar) was investigated. While differentiation between the various metabolites using either 'H or '3C NMR was unsuccessful, 31PNMR proved to be an exquisitely sensitive probe for these compounds. Treatment of soil samples with sulprofos at a concentration of 7.4 ppm, followed by incubation for up to 90 days, led to the formation of two major metabolites. The residues were extracted, characterized, and quantitated using 31PNMR. Use of 14Cradiolabeled material and analysis of extracts by high-performance liquid chromatography and liquid scintillation counting confirmed the results obtained by 31P NMR.
INTRODUCTION
The application of 31PNMR techniques to the identification and characterization of organic compounds has grown extensively since the initial experiments of Dickenson (1951) and Gutowsky and McCall (1951). Monographs by Gorenstein (1984) and Verkade and Quin (1987) illustrate the diversity of the application of this technique to chemical and biochemical research. The use of 3lP NMR to characterize pesticides has seen limited application. Ross and Biros (1970) and Miyata et al. (1988)have characterized phosphorus-containing pesticides using NMR. Lu and co-workers (Lu et al., 1985) correlated 31PNMR data to toxicity values for a number of pesticides. Wayne, et al. (1983) used 31PNMR as a probe to monitor the manufacture and stability of pesticides. The application of 3lP NMR to the quantitation of pesticides was carried out by Gurly and Ritchey (19761, and determination of enantiomeric ratios of chiral phosphate ester pesticides was done by Lu and co-workers (Lu et al., 1986). An interesting application of 31PNMR t o the analysis of pesticide residues in human body fluids has recently been reported by Nihira and co-workers (Nihara et al., 1990). There appears to be no application to date, however, of the 31PNMR experiment to t h e analysis of pesticide metabolites i n environmental matrices. We report our initial findings on the application of 31PNMR spectroscopy to t h e analysis of pesticides in soil. EXPERIMENTAL PROCEDURES Chemicals, The [phenyl-UL-14C]sulprofos [Bolstar, 0-ethyl 0-[4-(methylthio)phenyl]S-propyl phosphorodithioate], along with all of the nonradioactive reference standards used in this study, was synthesized in the laboratories of Mobay Corp. Deuterated solvents were obtained from Aldrich Chemical Co. (Milwaukee,WI). Acetonitrile and methanol were of HPLC grade (Burdick and Jackson, Muskegon, MI). NMR Spectroscopy. NMR data were recorded at ambient temperature for solutions of compounds in CDC13,CD30D, perdeuteroacetone, or perdeuterodimethyl sulfoxide. All spectra were recorded using a Varian Model XL-300 (Palo Alto, CA) NMR spectrometer. Observation frequencies were 300.0MHz for 'H, 75.43 MHz for I3C, and 121.42MHz for 31P. Chemical shifts ( 6 ) are reported as parts per million downfield from internal tetramethylsilane (TMS) for 'H and l3C spectra and as parts per million downfield from external phosphoric acid (H3P04)for 31P spectra. 31P NMR spectra of soil extract concentrates were 0021-8561/92/1440-0458$03.00/0
obtained on approximately 500-pg samples using a total spectral acquisition time of 20 h. High-Performance Liquid Chromatography (HPLC). HPLC was performed on representative samples of soil extract concentrates. Samples (100pL) were amended with metabolite standards and chromatographed on a YMC AQ-303S5 120A ODS column (4.6mm i.d. X 25 cm) using varying proportions of 0.1% aqueous acetic acid (solvent A) and acetonitrile (solvent B) as the mobile phase at a constant flow rate of 1.0 mL/min. The solvent program consisted of eluting for 10 min with a solution of 20% solvent B in solvent A, a linear ramp to 75% solvent B in 5 min, a ramp to 80% solvent B in 30 min, a ramp to 100% solvent B in 10 min, and holding at 100% solvent B for 5 min. The column was equilibrated with 15 mL of 20% solvent B in solvent A immediately prior to each run. The chromatographic system used for all analyses was a Beckman Model 345 consisting of a 112 solvent delivery module, a 340 organizer, a 341 controller, and a 165 variable-wavelength detector. The HPLC system was connected to an IN/US radioactivity monitor (Model Raytest Ramona). All data were compiled and processed using Raytest Ramona data reduction software. Quantitation of radioactive residues was accomplished by integration of the peak areas within each individual HPLC run and confirmed by liquid scintillation counting (LSC) data for recovered HPLC eluates. The reported results are the average of duplicate determinations. Radioassay. Liquid samples (0.1-1.0mL) were radioassayed using a Beckman LS 9000 liquid scintillation counter and its associated data reduction software. Aliquots of solid samples were oxidized using a Packard Tri-Carb Model 306 oxidizer, and the total radioactivity was determined by LSC of the evolved 14C02. Quench curves for both liquid and oxidizer samples were determined and used as prescribed by the instrument manufacturer. Soil Metabolism Study. The soil used in this study was obtained from Mobay's Howe Research Farm located near Howe, IN. Textural analysis showed it to be sandy loam (66% sand, 22 % silt, 12% clay, 2.2% organic matter,pH6.3,cation-exchange capacity 20 mequivig, particle density 2.6 g/mL). The soil was shown to be biologically active by incubating the soil with a known concentration of [14C]glucoseand measuring the amount of I4CO2 evolved (Anderson and Domsch, 1978) using the apparatus described by Anderson (1975). Immediately prior to sulprofos treatment, the soil was sieved to a maximum particle size of 2 mm and divided into 100-g (dry weight) aliquots. Aliquots were placed in 250-mL Erlenmeyer flasks with ground glass joints and amended with deionized water sufficient to bring each sample to a moisture level of 75% of 0.33 bar (8.5g/flask). Each soil sample was treated with a 100-pLaliquot of a solution containing 0.740mg of [phenyl-UL-14C]sulprofos(specific activity 25.35 mCi/mmol, 174 600 dpmipg). The soil was thoroughly mixed by tumbling each flask for 2-3 min. The sample flasks 0 1992 American Chemical Society
J. Agric. Food Chem., Vol. 40, No. 3, 1992
NMR Identification of Pesticide Residues
Table I.
compd 1 2 3 4 5 6 7 8
1JC
459
Chemical Shifts for the Aryl Portion of Sulprofos Standards
X S S S 0 0 0
Y S
S
S
R Et Et Et Et Et Et H H
so
so2 S
so
so2
1 24.0 44.34 44.375 24.739 44.338 44.396 36.202 33.650
2 136.5 144.85 (3.1) 139.302 (2.0) 136.0 (5.6) 144.612 (1.6) 139.016 (0.9) C
3 128.5 125.87 (1.7) 130.162 128.773 126.076 (1.2) 130.401 128.624 128.771
4 123.0 123.12 (5.0) 123.087 (5.0) 122.016 (5.0) 122.274 122.162 (5.4) 123.242 (0.8) 122.216
5 149.0 153.09 (8.8) 153.09 (8.5) 144.5 (5.0) 153.555 (7.5) 155.198 (12.0) 132.401 (4.8) 131.45
128.85 0 S Chemical shifts (6) are expressed as parts per million downfield from TMS. Coupling constants (4are expressed in hertz. The indicated resonance was not observed due to extremely slow relaxation.
Table 11. 'JC Chemical Shifts for the Phosphate Ester Portion of Sulprofos Standards x 6 7 11 O-CH,-CH,
-
S -CH,-CH,-CH, 8 9
C H 3 - Y e O - P
