Journalof Analytical Toxicology,Vol. 28, January/February2004
Identificationand Quantitation by High-Performance LiquidChromatographyof Mancozeb Following Derivatization by 1,2-Benzenedithiol* Ikram Debbarh, Karine Titier, Evelyne Deridet, and Nicholas Moore ~ Departement of Pharmacology, EA 525, Universit~ Victor 5egalen 33076 Bordeaux, France
Abstract] Ethylenebisdithiocarbamate (EBDC) fungicides are the most important class of organic fungicides and exhibit a high degree of carcinogenicity, mutagenicity, and neurotoxicity. For that reason, the safe application of these fungicides in practice requires a convenient method for their determination, applicable to biological fluids. We describe a high.performance liquid chromatography (HPLC) assay. After elimination of the metal which defines the product (maneb, mancozeb, zineb...) with EDTA, the resulting EBDC is derivatized with 1,2-benzenedithiol to yield a cyclocondensation product, 1,3.benzodithiole-2.thione, which can then be quantitated by reversed-phase HPLC at 365 nm using a pBondapak Cr8 column. The mobile phase was methanol/H20 (70:30, v/v). The assay was linear from 0.25 to 100 Fg/mL. Withinand belween.day precision and accuracy for this assay were better than 9% and 6%, respectively. The lower limits of detection and quantitation were estimated to be 0.1 and 0.25 pg/mL, respectively. This simple new method has been applied to determine mancozeb concentration in rat urine samples from urinary excretion studies.
Introduction Ethylenebisdithiocarbamates(EBDC)are the fungicidesmost widelyused to protect fruits, vegetables,and field crops against a widespectrum of fungal disease. The class includes (Figure 1) zineb (zinc EBDC), maneb (manganese EBDC), mancozeb (manganese and zinc complex EBDC), metiram (zinc ethylenebisdithiocarbarnate-polyethylenebisthiuramdisulfide), and nabam (disodium EBDC). These compounds are generally considered to have a low short-term mammalian toxicity. A major toxicological concern, however, is ethylenethiourea (ETU), an industrial contaminant and breakdown product of EBDC (1). On the other This work was supported by Universit~ Victor Segalen EA 525. ' Author to whom correspondence should be addressed. E-mail:
[email protected].
hand, long-term effects can present some hazards. EBDCproduces significant toxicological effects on thyroid, gonads, and chromosomes (2,3) after repeated exposure and is not devoidof teratogenicity (4) and carcinogenicity (5,6) in experimental animals. The neurotoxic effect of these widely used fungicides has been reported, and EBDC are known to provoke a wide range of neurobehavioraleffectsand neuropathologicalchanges in the brain (7). Additionally,maneb has been reported to induce permanent extrapyramidalsyndromes resembling Parkinsonism (8--10). EBDCs have been determined in crops by high-performance liquid chromatography (HPLC) after conversion into nabam (11); by liquid chromatography with postcolumn acid hydrolysis to form ethylenediamine, which is fluorogenically labeled with o-phthalaldehyde-mercaptoethanol (12); or with ion-pair methylation (13). In biologicalfluids, ETU is commonly used as a biological indicator of exposure to EBDCs (14,15). No direct and generic method for quantitating EBDCsin biologicalfluids has yet been described. Thus, there is a strong need to develop suitable methods to monitor exposure to these compounds. In this paper, we describe an HPLC assay for the determination of EBDC. This assay is based on a reaction of these compounds that have two highly electrophiliccentral carbon atoms, with excess 1,2-benzenedithiol to yield the compound 1,3-benzodithiole-2-thione,which then can be quantitated by reversedphase HPLC with UV detection (Figure 2). Similar methods
CH2- MI- C- S
~2-1~-C1-$
S
n
[OIZ-NH- 'C-
9
Figure 1. Chemical structureof someethylenebisdithiocarbamates.
Reproduction(photocopying)of editorialcontentof thisjournalis prohibitedwithoutpublisher'spermission.
41
Journal of Analytical Toxicology, Vol. 28, January/Februa~ 2004
have been developed to quantitate isothiocyanates (R-N=C=S) (16). The chemical specifity of this cyclocondensation reaction is, however, not restricted to isothiocyanates (17), making it useful for the analysis of structurally related thiocarbonyl compounds, including dithiocarbamates.
Materials and Methods Chemicals 1,2-Benzenedithiol was purchased from Aldrich (SigmaAldrich, St. Louis, MO), mancozeb (80%) was purchased from Rohm and Haas (Paris, France), monobasic potassium phosphate and ethylenediaminetetraacetic acid (EDTA) were obtained from Sigma (Sigma-Aldrich), and isopropyl alcohol and methyl alcohol were obtained from Carlo-Erba (Val de Rueil, France). The 80% were taken into account in all concentration calculations and for rat administration. The remaining 20%, which are undetermined but are not mancozeb, do not produce CS2 (from the manufacturer certificate) and therefore will not react to our derivatization method.
Em'A I
M n "oebe
I~n~ny
5
$~
4"
5
5H 5H NH2 + 1.2 Ih'n:em'dithlol ],3.Bcn:~,dlthudc2 Ihtotlc[~2) Formula of cvclocondensation reaction with 1,2-benzenedithiol.
Figure 2.
Standard solution and sample EBDCs are insoluble in water and almost all organic solvents, however a stock solution of mancozeb was prepared at 1 mg/mL in an alkaline solution of EDTA(25raM) adjusted to pH 9.25 with NaOH (11). Standard solutions (0.25, 0.5, 1, 2.5, 5, 10, 25, 50, and 100 pg/mL) were made by futher dilution of the stock solution with appropriate volumes of human plasma or urine. Standard and stock solutions of mancozeb were stored at -20~ Urine collection Male rats (200-250 g) were purchased from Charles River (l~Arbresle, France). Uponarrival, the rats were housed at 22~ with a 12 h light/dark cycle, for one week before use. The rats were then transferred to individual metabolic cages and were given intraperitoneal doses of 0 (blank), 1, 10, 50, or 100 mg/kg mancozeb suspended in 0.25% carboxymethylcellulose (4 rats per dose). They had free access to foodand water. Urine was collected for 24 h and stored at -20~ until analysis. Daily urine volume was 2-3 mL. Cyclocondensation reaction 1,3-Benzodithiole-2-thione resulted from the reaction of EBDC with excess 1,2-benzenedithiol. The 1-mL reaction mixture consisted of 100 pL of EDTA (25mM) at pH 9.25 (11), 300 pL of 100raM potassium phosphate at pH 8.5, 100 pL of sample, and 500 pL of isopropanol containing 1,2-benzenedithiol (to give a final concentration of 10mM). The reaction was carried out in 7-mL screw-top glass vials equipped with tight-fitting plastic caps. The vials were flushed with nitrogen and incubated for 2 h at 65~ After cooling to room temperature and centrifugation (2000 rpm, 10 rain), 10- or 100-pL aliquots of the reaction mixtures were injected into the chromatographic column.
Apparatus and chromatographic conditions The apparatus used for HPLC analysis were Millipore 501 pumps (Waters, Millford, MA), a Millipore WISP 710 B autosampler, and a Spectra system UV 100 detector (Spectra Physics, Franklin, MA) set at 365 nm. The chromatographic column was a pBondapack C]8 (300 x 3.9 ram, Waters). The mobile phase was methanol/water (70:30, v/v).The riow rate was 1 mL/min, and the injection volume was 10-100 pL. The column was rinsed for 10 min between successive sample injections. Validation Assay performance of the present method was assessed for linearity, precision, accuracy, limit of quantitation (LOQ), limit of detection (LOD), and stability.
A
B
G
Figure 3. Chromatograms resulting from analysis of blank sample (A), human urine spiked with 50 pg/mL mancozeb (B), and rat urine after intraperitoneal administration of 10 mg/kg mancozeb (C).
42
Results Expressionof results The compound that was actually assayed is 1,3-benzodithiole2-thione (BDT). However, because all standards are prepared
Journal of Analytical Toxicology, Vol. 28, January/February 2004
Linearity The relationship between peak area and concentration in urine or plasma was linear within the range of 0.25-100 IJg/mL of added mancozeb with correlation coefficientsof 0.9997 and 0.9999 (n = 6), respectively.The equations were y = -0.057 + 1.512x for urine and y = -0.029 + 1.096x in plasma. The linearity of the method was confirmed using the linear leastsquares regression procedure by comparing the slopes and the intercepts of calibration curves with zero and correlation coefficients with 1. The g-intercepts were not different from zero. For each calibration curve, the slope was statistically different from 0, as determined by a Fischer test (TableI).
from mancozeb, the results are expressed as micrograms per milliliter mancozeb or millimoles per liter EBDC. Under the chromatographic conditions used, BDT had retention times of approximately 4.5 rain, exact values depended on column usage. Typical chromatograms are given in Figure 3. The effects of 1,2-benzenedithiol concentrations of 5, 10, and 20raM were tested with 5, 10, and 100 IJg/mLmancozeb, using the procedure described, in triplicate at all points. A 5-mM concentration resulted in peak heights that were approximately half those obtained with 10 or 20mM 1,2benzenedithiol, which were the same. At the highest concentration of 1,2-benzenedithiol, an extra peak appeared approximately 0.3 rain before the main peak. Ten millimoles was therefore chosen as the optimal 1,2-benzenedithiolconcentration for the cyciocondensation reaction. There was no interaction with CS2, ethylene thiourea (the common metabolite of EBDC), or with human urinary isothiocyanates (data not shown).
LOQ and LOD The LOQ is definedas the lowest drug concentration that can be determined with an accuracy of 80-120% and precision better than 20%. The LOQ was 0.250 IJg/mLfor samples prepared in 100 IlL of urine or plasma. The LOQ samples were analyzed six times, with coefficients of variation of, respectively, 18.9% and 14.8% in plasma and urine. The LOD, the calibrator concentration below that Table I. Accuracy and Correlation of the Calibration Curves Obtained from which qualified as LOQ if the peak was more Human Plasma and Urine Spiked with Mancozeb from 0.25 to 100 pg/mL than three times the baseline noise, was 0.1 IJg/mL. Matrix Day 1 Day 2 Day 3 Plasma (n = 30) Slope Intercept Correlation coefficient Comparison of slope to zero (Fischer test, F) Comparison of intercept to zero (Fischer test, F)
1.449 --0.031 0.9996 8.3 t 1.36*
1.51 -0.046 0.9998 9.011 3.28*
1.58 -0.095 0.9997 10.91 2.2*
1.12 -0.04 0.9998 43.3"I"I 1.42t
1.075 -0.02 0.9999 51.31 2.54 t
1.095 -0.028 0.9999 62.91 1.721
Urine (n = 30) Slope Intercept Correlation coefficient Comparison of slope to zero (Fischer test, F) Comparison of Intercept to zero (Fischer test, F)
9 Abbreviations:theoreticalvalues, [F(0.0S;!; 28) = 4.26] and n = total numberof data points. t Significantvalues. 9 Non-significantvaluesat the 0.05 level.
Table II. Intra- and Interday Coefficient of Variation and Accuracy for Determination of Mancozeb in Human Plasma and Urine Between.DayVariability(n = 6) Theoretical Concentration Concentration found(pg/mt)
Concentration Accuracy found(pg/mt) (mean • SD)
CV (%)
Accuracy
(%)
8.1
96.0
0.48 • 0.02
4.2
96.0
5.20 • 0.25 51.8 • 1.8
4.8 3.7
104.0 103.6
5.10 + 0.15 51.7 • 1.5
3.0 3.0
102.0 103.4
0.52 • 0.02 5.26 • 0.08 48.8 • 1.09
4.2 1.5 2.2
104.0 105.2 97.6
0.50 • 0.03 5.10 • 0.13 49.4 • 0.58
5.2 2.5 1.2
100.0 102.0 98.8
(mean :1:SD)
CV (%)
0.5
0,48 + 0,04
5 50
(pg/mt)
Within.DayVariability(n = 6)
(%)
Plasma
Urine 0.5 5 50
Precision and accuracy The precision of the method was tested for within- and between-day reproducibility by 6 repeated analyses in urine and plasma at concentrations of 0.5, 5, and 50 Iag/mL, respectively (TableII). Precisionwas better than or at 5% for all concentrations but one (0.5 pg/mL in plasma, which was still better than 10%). Analytical accuracy was evaluated by measuring the variation between the added concentration and the measured concentration. This was within 5% for all concentrations (Table II). Recovery Recovery was measured at concentrations of 0.5, 5, and 50 I~g/mL;four samples of each concentration as a standard solution; or in plasma and urine, followed by the cyclocondensation reaction. The recovery was calculated by comparing the peak heights in plasma and urine samples to those in the EDTAmatrix, obtained after derivatization. The recoveries are shown in Table III.
Stability Stability studies were conducted in blank plasma and urine spiked with 1 and 5 IJg/mL mancozeb, respectively,and stored at -20~ Four samples were analyzed on the day of sample preparation (control samples) and then weeklyover two months (Figure 4). 1~o urine samples from rats treated with mancozebwere
43
Journal of Analytical Toxicology, Vol. 28, January/February 2004
reanalyzed after three-weeks storage at-20~ No indication of drug instability was found in blank plasma and blank urine spiked with mancozeb. Analyte concentrations in samples analyzed during the two-month period did not differ statistically from control samples. Similar results were obtained upon reanalyzing experimental samples.
lected over 24 h, and concentration of the compound was determined. The results are shown in Figure 5. No products were found in the control (blank) group. The amount of EBDC excreted in urine over 24 h was 29.8% of each administered dose. This method had adequate sensitivity to measure urinary excretion after intraperitoneal injection of mancozeb to rats.
Application This method was applied to the quantitation of EBDC excreted in rat urine over 24 h. Four rats each were given 0, 1, 10, 50, or 100 mg/kg mancozeb intraperitoneally. Urine was col-
Discussion
Table III. Recovery of Mancozeb from Human Plasma and Urine (n = 4) Recovery (%) (mean • SD)
Concentration(m~mt)
~ma
0.5 5 50
Urine
96.6• 97.3• 102.0•
105.0• 103.5• 96.83•
6 $
I1"
qp
-"
.~"
T
T
T
6
7
8
i' 2
I
!
Z
~
4
5
Week
Figure4. Stability of mancozeb in urine stored at -20~ over 8 weeks at a concentration of 5 pglmL (upper curve) and I pglml. (lower curve) (mean, SD, n = 4). 4O
~. 3s O
i
30
The current study reports a new method for the quantitative determination of mancozeb in biological fluid samples. Mancozeb and other EBDC are difficult to dissolve in the aqueous phase of biological samples. In addition, the EBDC moiety does not readily absorb UV light, making spectroscopic quantitation difficult. To avoid the problem of mancozeb insolubility, a reaction with EDTA solution was used. Addition of EDTA to aqueous solutions of mancozeb causes dissolution. The EBDC polymers are simplified by breaking into smaller units, and the metal is substituted by sodium, resulting in the formation of the soluble nabarn, a step common to most dosage methods for EBDC (11). Nabam then reacts quantitatively with 1,2-benzenedithiol to give rise to 1,3-benzodithiole-2-thione, which has UVabsorbance characteristics appropriate for spectroscopic quantitation. Because it is the metal ion that mainly differentiates the EBDC products, this method, which first separates the ion, is applicable to most EBDC derivatives.Although what is assayed is not mancozeb per se, but 1,3-benzodithiole-2-thione, which may also result from some of the EBDC metabolites, its concentration is determined with reference to mancozeb, and, for the sake of commodity, we equate its concentrations with those of mancozeb, or rather EBDC. This method was successfully applied to the determination of the concentration of EBDC in urine after intraperitoneal administration of mancozeb. EBDC in the urine appeared to be dependent of dose when single intraperitoneal doses of mancozeb were administred. The excretion of EBDC in urine of four rats by dose was 29.8 • 6.0%. The use of this method for the detection of EBDC in human urine remains to be tested. If confirmed, this may facilitate the identification of exposures in excess to the World Health Organization recommendations (18).
0
References 0
~
s O.
/ 10 50 Mancozeb dose (mg/kg)
100
Figure 5. Twenty-four hour urinary excretion of EBDC after single intraperitoneal administration of 1, 10, 50, and 100 mg/kg mancozeb to rats (mean, SE, n = 4). 44
1. R Houeto, G. Bindoula, and J.R. Hoffman. Ethylenebisdithiocarbamates and ethylenethiourea: possible human health hazards. Environ. Health Perspect. 103(6): 568-573 (1995). 2. R. Kackar, M.K. Srivastava, and R.B. Raizada. Induction of gonadal toxicity to male rats after chronic exposure to mancozeb. Ind. Health 35('1): 104-111 (1997). 3. A.I. Kurinnyi. Approaches to genetic control for the use of pesticides. Tsitol Genet. 20(6): 463-477 (1986). 4. L. Ivanova-Chemishanska, D.V. Markov, S. Milanov, D.D. Strashimirov, G.I. Dashev, and G.A. Chemishanski. Effect of subacute oral administration of zinc ethylenebis(dithiocarbamate) on the thyroid galnd and adenohypophysis of the rat. Food Cosrnet. Toxicol. 13(4): 115-117 (1975).
Journalof Analytical Toxicology,Vol. 28, January/February2004 5. O.V. Chernov, II. Khistenko. Blastomogenic properties of derivatives of dithiocarbamiques acid. Vopr. Onkol. 15:71-74 (1969). 6. RN. Balin. Experimental data on the blastomogenic activity of the fungicides mancozeb. Vrach. Delo. 4:21-24 (1970). 7. D.B. Miller. Neurotoxicity of pesticidal carbamates. Neurobehav. Toxicol. Teratol. 4:779-787 (1982). 8. T.U. Hoogenraad. Dithiocarbamates and Parkinson's disease. Lancet 1(8588): 767 (1988). 9. H.B. Ferraz, RH. Bertolucci, J.S. Pereira, J.G. Lima, and L.A. Andrade. Chronic exposure to the fungicide maneb may produce symptoms and signs of CNS manganese intoxication. Neurology 38(4)" 550-553 (1988). 10. G. Meco, V. Bonifati, N. Vanacore, and E. Fabrizio. Parkinsonism after chronic exposure to the fungicide maneb (manganese ethylene-bis-dithiocarbamate). Scand. J. Work Environ. Health 20(4)" 301-305 (1994). 11. J. Lehotay and D. Kisov~. HPLC study of mancozeb degradation on leaves. J. Liquid Chrom. 16(5): 1015-1022 (1993). 12. C.J. Milles and M. Zhou. Determination of nabam fungicide in crops by liquid chromatography with postcolumn reactiondetection. J. Assoc. Off. Anal. Chem. 74(2): 384-388 (1991 ). 13. N. Kibune, K. Higashisaka, M. Nakamura, and Y. Maekawa. Rapid analysis method of dithiocarbamate pesticides in agricultural products by high performance liquid chromatography. J. Food Hy-
genic. 36(2):244-251 (1995). 14. P. Kurttio and K. Savolainen. Ethylenethiourea in air and in urine as an indicator of exposure to ethylenebisdithiocarbamate fungicides. Scand. J. Work Environ. Health 16(3): 203-207 (I 990). 15. P. Kurttio, K. Savolainen, A. Naukkarinen, V.M. Kosma, L. Tuomisto, and I. Penttila. Urinary excretion of ethylenethiourea and kidney morphology in rats after continuous oral exposure to nabam or ethylenethiourea. Arch. Toxicol. 65(5): 381-385 (I 991 ). 16. Y. Zhang, K.L. Wade, T. Prestera, and P. Talalay. Quantitative determination of isothiocyanates, dithiocarbamates, carbon disulfide, and related thiocarbonyl compounds by cyclocondensation with 1,2-benzenedithiol. Anal. Biochem. 239(2): 160-I 67 (I 996). 17. Y. Zhan8, C.G. Cho, G.H. Posner, and R Tzlalay. Spectroscopic quantitation of organic isothiocyanates by ctclocondentation with vicinal dithiols. Anal. Biochem. 205(1): 100-I 07 (I 992). 18. G. Vettorazi, W.F. Almeida, G.J. Burin, R.B. Jaeger, F.R. Puga, A.F. Rahde, F.G. Reyes, and S. Schvarsman. Internationl safety assessmentof pesticides: dithiocarbamate pesticides ETU and PTU-a review and update. Teratogenesis, Carcinogenisis, and Mutagenesis 15:313-337 (1995). Manuscript received May 15, 2001; revision received June 4, 2003.
45
