- Email: [email protected]
Development of an enzyme-linked immunosorbent assay for determination of pretilachlor in water and soil
Ecotoxicology and Environmental Safety 74 (2011) 1595–1599
Contents lists available at ScienceDirect
Ecotoxicology and Environmental Safety journal homepage: www.elsevier.com/locate/ecoenv
Development of an enzyme-linked immunosorbent assay for determination of pretilachlor in water and soil Zhen-Jiang Liu, Peng-Min Yu, Song Fang, Jia-qin Fan, Ming-Hua Wang n Department of Pesticide Science, College of Plant Protection, Nanjing Agricultural University, Jiangsu Key Laboratory of Pesticide Science, Key Laboratory of Integrated Management of Crop Diseases and Pests (Nanjing Agricultural University), Ministry of Education, Nanjing 210095, PR China
a r t i c l e i n f o
abstract
Article history: Received 2 April 2010 Received in revised form 29 March 2011 Accepted 17 April 2011 Available online 4 May 2011
An indirect competitive enzyme-linked immunosorbent assay (ic-ELISA) has been developed for detection of pretilachlor in water and soil. An immunogen was prepared from haptens of pretilachlor conjugated to bovine serum albumin(BSA). The specific polyclonal antibodies were obtained by immunizing New Zealand white rabbits. The influence of parameters including concentrations of methanol, ionic strength and pH values were optimized to improve the sensitivity of the assay. The optimized ELISA was shown to have a high sensitivity and specificity for pretilachlor. Under optimal conditions, the ELISA has demonstrated an 50% inhibitory concentration (IC50) value of 0.0359 mg/L with a limit of detection (LOD, IC10) of 6.9 ng/L. The cross-reactivities to some analogs of pretilachlor (acetochlor, butachlor, metazachlor and metalaxyl) were below 1.5%. The average recoveries of pretilachlor from distilled water, tap water, paddy water and soil were in the range of 77.0–115.2% between 0.005 and 5.0 mg/L. The results of ELISA for spiked samples were confirmed by GC–ECD with a high correlation coefficient of 0.9950(n ¼ 11). Thus, the ELISA proven to be a sensitive, specific, inexpensive and quantitative tool for detection of pretilachlor from four kinds of spiked samples. & 2011 Elsevier Inc. All rights reserved.
Keywords: Enzyme-linked immunosorbent assay Pretilachlor Hapten Polyclonal antibody
1. Introduction Pretilachlor [2-chloro-N-(2-ethyl-6-ethylphenyl)-N-(2-ethoxy1-methylethyl) acetamide] is a selective pre-emergent herbicide, which is widely applied to control grasses in rice. Pretilachlor has a low toxicity to human and mammals, however it is highly toxic to fishes (Liu, 2002). Due to its extensive application, pretilachlor could have an adverse impact on the environment especially on water body. It is highly desirable to develop a specific and sensitive method to monitor residues of pretilachlor in environmental samples. Several methods for the determination of pretilachlor have been described, such as gas chromatography (Huang, 1997) and gas chromatography–mass spectrometry (Tanabe et al., 1996; Chu et al., 2005; Akiko et al., 1996). They are characterized by low limits of detection and high precision and sensitivity. These regular methods, however, generally require specialized technicians and instrumentation, and their procedure is rather cumbersome and expensive, therefore unsuitable for screening large numbers of environmental samples. The immunoassay has received more and more concerns as a simple, sensitive and cost-effective technique and has became a reliable
n
Corresponding author. Fax: þ86 25 84395479. E-mail address: [email protected] (M.-H. Wang).
0147-6513/$ - see front matter & 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.ecoenv.2011.04.026
analytical tool for rapid screening analysis in recent years (Hennion and Barcelo, 1998). Since the first application of enzyme-linked immunosorbent assay (ELISA) for detection of alachlor in water has been reported by Feng et al. (1990), immunoassay techniques have been developed for determination of several chloroacetanilide herbicides from various matrices. In order to develop an immunoassay with high sensitivity and selectivity, the first key step is the design of immunogen. The general strategy to obtain the immunogen for chloroacetanilides relies on the modification of N-chloroacetyl moiety. Such modification preserves N-alkyl side chain, which is the unique structural feature of chloroacetanilide herbicide. Protein thiolating agents such as S-acetylmercaptosuccinic anhydride and N-Acetylhomocysteinethiolactone were utilized as cross-linkers to develop an ELISA of alachlor (Feng et al. 1990; Feng et al. 1994; Gabaldon et al. 2002), amidochlor, butachlor and metolachlor (Feng et al. 1992). In recent years, the strategy has been improved by utilizing 3-mercaptopropionic acid (3-MPA) as a cross-linker. Such improvement has been proved successfully in rapid immunoassays of acetochlor (Yakovleva et al. 2002; Hegedus et al. 2002) and butachlor (Yakovleva et al. 2003). So far, there were not any ELISAs for the detection of pretilachlor has been reported. In this paper, a sensitive ELISA was developed for detection of pretilachlor residues in water and soil samples based on polyclonal antibodies. Furthermore, the ELISA performance was evaluated with conventional GC–ECD in terms of precision and accuracy using fortified samples.
1596
Z.-J. Liu et al. / Ecotoxicology and Environmental Safety 74 (2011) 1595–1599
2. Materials and methods
2.6. Preparation of polyclonal antibody
2.1. Regents
Male New Zealand white rabbits weighing about 2 kg were immunized with PMPA-BSA according to the immunization protocol reported (Zeng et al., 2006). The rabbit had free access to drinking water and commercial standard laboratory diet (CZZ, Nanjing, China). It was housed according to the EEC 609/86 Directives regulating the welfare of experimental animals. The rabbits were bled after 8 days of the last injection. The blood was coagulated for 1 h at 37 1C and 12 h at 4 1C. The antiserum was centrifuged and then purified via the caprylic acid–ammonium sulfate precipitation method and freeze-dried.
All regents and solvents used were of analytical grade. Pretilachlor was obtained from the Changqing Agrochemical Co., Ltd. (Jiangsu, China). 1,3-dicyclohexylcarbo- diimide (DCC), N-hydroxysuccinimide (NHS), isobutylchlorocarbonate (TBA) and tri-n-butylamine (TEA) were acquired from Sinopharm Chemical Reagent Co., Ltd. (Shanghai. China). Bovine serum albumin (BSA), Ovalbumin (OVA), Freund’s complete and incomplete adjuvants, goat anti-rabbit IgG-horseradish peroxidase, o-phenylenediamine (OPD) and polyoxyethylene sorbitan monolaurate (Tween-20) were purchased from Sigma Chemical Co. (St. Louis, USA). 3-mercaptopropionic acid (3-MPA) was from A Johnson Matthey Company.
2.2. Buffers and solutions Phosphate buffered saline (PBS, 0.01 mol/L, pH 7.4). Carbonate-bicarbonate buffer saline (CBS, 0.05 mol/L, pH 9.6). Phosphate buffered saline contained 0.05% Tween-20 (PBST). The substrate solution contained 0.025 mol/L of citrate and 0.062 mol/L of sodium phosphate, pH 5.4. The OPD solution contained 0.4 mg/mL OPD and 0.012% H2O2 in substrate solution.
2.3. Instrumentation and equipment 1 H NMR spectrum was obtained via a DRX500 spectrometer (Bruker, Germany). MS spectrum was measured using a LC-MSQDECA (Finnigan, USA). IR spectrum was obtained with a Vertex 70FT-IR (Bruker, Germany). UV spectra were recorded on a DU 800 spectrophotometer (Beckman Coulter, USA). 96-Well Polystyrene microplates (Maxisorp) were purchased from Nunc (Roskilde, Denmark). ELISA plates were washed with a Wellwash Plus (Thermo, USA). Absorbances were read with an Infinite M200 microtiter plate reader (Tecan, Switzerland) at 490 nm. Aglient 7890A gas chromatography (Aglient, USA).
2.7. Protocol of ELISA Microplates were coated 2 h at 37 1C with 100 mL well 1 of the coating antigen diluted 1:6000 (v/v)in CBS. The plates were washed three times with PBST. 200 mL of PBS containing 1% (w/v) OVA was added and incubated for 30 min at 37 1C and the plates were washed again. For competition, 50 mL of sample or standard and 50 mL of diluted antiserum solution (0.5 mg/L) were incubated together at 37 1C. After incubation for 1 h, the plates were washed again. Then 100 mL/well of a diluted (1:3000) goat anti-rabbit IgG-horseradish peroxidase was added and incubated for 1 h at 37 1C. After another washing step, 100 mL/well of the OPD solution was added and incubated for 15 min at 37 1C. Finally 2 mol/L of sulfuric acid (50 mL/well) was added and the absorbance was measured at 490 nm. The standard curve for pretilachlor was obtained by napierian logarithm of plotting percent binding (logit (B/B0)) versus the logarithm of the pretilachlor concentration (log C). %(B/B0) was calculated by the following equation: %ðB=B0 Þ ¼ ½ðAx Amin Þ=ðAmax Amin Þ 100 where Ax is the absorbance of the sample, Amax is the absorbance in the absence of analyte and Amin is the absorbance of the background. IC50 and IC10 values were calculated.
2.8. Optimization of ELISA 2.4. Synthesis of hapten The hapten, pretilachlor-3-mercaptopropionic acid (PMPA), was synthesized according to the method described previously (Yakovleva et al. 2002, 2003). The synthesis route was illustrated Fig. 1. A mixture of 0.32 g (1.0 mmol) of pretilachlor, 0.15 g (1.4 mmol) of 3-MPA and 0.15 g (2.7 mmol) of potassium hydroxide in 20 ml of ethanol was refluxed for 6 h. The reaction mixture was filtered, and the filtrate was concentrated to a white solid. The solid was dissolved with 10 ml of 5% NaHCO3 and filtered. The solution was acidified to pH 3.0 with 6 mol/L HCl and extracted three times with 15 ml of ethyl acetate. The organic phase was washed with water and dried with Na2SO4. The PMPA was purified by TLC using a mixture of benzene: dioxane (5:1, v/v) as an eluent. The product (Rf ¼ 0.6) was extracted by ethanol and evaporated to obtain a yellow oily PMPA. The product was characterized by 1H NMR, IR and MS. 1H NMR (CDCl3), d7.15–7.31(m, 3H, ArH), 3.74–3.76(t, 2H, CH2O), 3.58–3.60(t, 2H, OCH2), 3.30– 3.32(t, 2H, NCH2), 2.92(s, 2H, COCH2S), 2.86–2.89(t, 2H, SCH2), 2.62–2.64(t, 2H, CH2COOH), 2.55–2.62(m, 4H, 2ArCH2), 1.48–1.52(m, 2H, CH2), 1.23–1.26(t, 6H, 2CH3), 0.83–0.86(t, 3H, CH3). IR (KBr, Vcm 1): 3500–3000 (v, O–H), 1725 (vs, carboxylic acid C ¼ O). MS, ESI m/z: 381 [M–H þ ], M ¼ 382.
The influence of some of the parameters such as organic solvent, ionic strength and pH were evaluated to improve the sensitivity of the ELISA. Competitive curves were carried out for pretilachlor standards in PBS containing 10–40% (v/v) of methanol, PBS buffer with concentrations of Na þ (0.1, 0.14, 0.2, 0.3, 0.4, 0.5 mol/L) and PBS buffers with pH values 4.5, 5.5, 6.5, 7.5, 8.5 and 9.5 to determine the effects of solvent, ionic strength and pH values.
2.9. Cross-reactivity determination Cross-reactivities (CR) were determined to evaluate the selectivity of the optimized ELISA. The cross-reactivities of four analogs of pretilachlor (acetochlor, butachlor, metazachlor and metalaxyl) were determined. CR was calculated by the following equation: CRð%Þ ¼ ðIC50 of pretilachlor=IC50 of analoguesÞ 100:
2.10. Evaluation of the ELISA with spiked samples 2.5. Preparation of immunogen and coating antigen The PMPA was coupled to BSA by active ester method (Zhang et al., 2008) and OVA via the mixed anhydride reaction (Zeng et al., 2006). The conjugates were dialyzed against PBS over 72 h at 4 1C and stored at 20 1C. PMPA-BSA and PMPAOVA were used as immunogen and coating antigen, respectively. The conjugated formation was confirmed spectrophotometrically. The absorbances of hapten, carrier protein and protein conjugate solutions (0.1 mg/L) at 280 nm were measured and hapten density was calculated directly by the mole absorbance e (Lu et al., 2009): Hapten density ¼ ðeconjugation eprotein Þ=ehapten
ELISA was applied for the determination of pretilachlor in water and soil samples. Water samples were collected from three sources: distilled water, tap water and paddy water. The water was spiked pretilachlor at various levels (0.05, 0.1, 0.5 and 5 mg/L). To reduce the matrix effect, water samples were diluted 10 times with PBS containing 10% methanol. 10 g of soil sample spiked with pretilachlor at 0.10, 0.50 and 1.00 mg/kg was mixed with 30 ml of acetonitrile and ultrasonic extraction for 10 min; the suspension was filtrated. The filtrate was added to 10 ml saturated sodium chloride solution and was vigorously shaken. The organic phase was concentrated and the remainder was dissolved with 10 ml PBS containing 10% methanol. The samples were analyzed with ELISA method by triplicate. Samples without pretilachlor were used as blanks in all case.
C2H5
C2H5 CH2CH2OC3H7
N COCH2Cl
KOH + HSCH2CH2COOH C2H5OH
C2H5
CH2CH2OC3H7 N COCH2SCH2CH2COOH C2H5
Fig. 1. Synthetic route of the hapten.
Z.-J. Liu et al. / Ecotoxicology and Environmental Safety 74 (2011) 1595–1599
1597
1.2
2.11. Evaluate of the assay with GC–ECD GC was performed according to the procedure of Wang et al. (2005). Samples included distilled water, tap water and paddy water and soil were spiked pretilachlor at various levels (0.01–1 mg/kg). Water samples were extracted with dichloromethane. Soil samples were extracted with a mixture of petroleum ether– acetone (1:1 v). The pretilachlor were measured by GC–ECD simultaneously and the measured results were compared with ELISA.
Logit (B/B0)
0.6
3. Result 3.1. Identification of conjugations In the UV–vis spectrometry obtained from continuous wavelength scanning, there were qualitative differences between the spectra of the conjugate and the corresponding carrier protein, especially near 280 nm and showed that the carrier protein and hapten had been coupled successfully. The hapten density was 11.9:1 and 8.5:1 for PMPA-BSA and PMPA-OVA, respectively.
0
-0.6
-1.2 -3.2
-2.2
-1.2 Log[pretilachlor] (mg/L)
-0.2
Fig. 2. Standard curves by ic-ELISA for pretilachlor.
3.2. Optimization of assay conditions IC50 is an important feature, which reflects the sensitivity of the immunoassay. The IC50 values and B0 of the pretilachlor of the optimization are shown in Table 1. It was observed that immunoassay is more sensitive as buffer salt concentration increased. However, the maximum signal sharply diminished when salt concentration more than 0.14 M. A most common water miscible solvent, methanol, was tested for its effect on assay performance. The best sensitivity was observed when the methanol concentration in PBS was 10%. The influence of pH on both the maximum signal and sensitivity was studied. It was found that the pH did not have a notable effect on the assay. The result seemed to indicate that electrostatic interactions are not the dominant forces in the recognition of antibody and the antigen in this assay.
Table 2 Cross-reactivity of pretilachlor and some of its analogs. Compounds
Structure
IC50(mg/L)
CR (%)
Pretilachlor
0.042
100
Acetochlor
7.6
0.55
3.3. Sensitivity of the assay
Butachlor
2.8
1.5
Under optimum conditions, the ic-ELISA procedures were conducted in triplicates with a series of concentration of pretilachlor. The standard curve was represented in Fig. 2. It was observed that the immunoassay had a linear relationship from 0.001 to 1.000 mg/L between logit(B/B0) and the logarithm of the pretilachlor concentration. An equation was obtained: logit (B/B0)¼ 0.591 Log[C] 0.8539, R2 ¼0.9909. The method with an IC50 of 0.0359 mg/L, a limit of detection (IC10) of 6.9 ng/L and a linear working range from 0.001 to 1.000 mg/L.
Metazachlor
4100
o0.1
Metalaxyl
4100
o0.1
Table 1 Effect of ionic strength, solvent, pH on the ELISA system. Ionic strength
Methanol concentration
Ionic strength (mol/L)
IC50 B0 (mg/L)
0.10 0.14 0.20 0.30 0.40 0.50
0.141 0.102 0.108 0.087 0.085 0.048
1.02 1.20 0.93 0.74 0.62 0.58
pH
Methanol concentration (v/v, %)
IC50 B0 (mg/L)
10 20 30 40 50
0.041 0.086 0.097 0.197 0.682
pH IC50 B0 (mg/L)
3.4. Specificity of the assay 1.18 1.13 1.16 1.07 1.17
4.5 5.5 6.5 7.5 8.5 9.5
0.026 0.036 0.045 0.042 0.041 0.040
1.34 1.28 1.15 1.21 1.22 1.30
The values of cross-reactivity of the antibody with pretilachlor and related compounds were listed in Table 2. It was found to be less than 1.5% of cross-reactivity for acetochlor, butachlor, metazachlor and metalaxyl. Despite the high similarity of the chloroacetanilide compounds, the ELISA has high specificity to
1598
Z.-J. Liu et al. / Ecotoxicology and Environmental Safety 74 (2011) 1595–1599
pretilachlor, which demonstrated that alkyl radicals moiety in the chloroacetanilide are the most important antigenic determinants. Besides, replacement of methyl in benzene ring resulted in lower of antibody reactivity. 3.5. Accuracy and precision of the ELISA The spiked recoveries and the relative standard deviation (RSD) were calculated to evaluate the accuracy and precision of the ELISA. As illustrated in Table 3, the recoveries were in the range of 77.0–115.2% and the RSD were found to be 2.3–9.4%. These data suggested that the matrix interference of water samples can be minimized by diluting the samples 10 times with PBS. The results demonstrated that this assay is suitable for the quantitative detection of pretilachlor at trace levels in water and soil. 3.6. Correlation between the ELISA and GC As shown in Fig. 3, good correlation was obtained between ELISA (x) and GC (Y) with the linear regression equation of Y ¼0.9487x þ0.001 (R2 ¼0.9950, n ¼11). These results suggested that pretilachlor in the samples could be simply, rapidly and accurately detected by ELISA. Table 3 Recoveries of pretilachlor in spiked water and soil samples. Samples
Mean detected Spiked concentration (mg/L) (mg/L, mg/kg)
Average recoveries (%)
RSD (%) (n¼3)
Distilled water
0.05 0.50 5.0
0.058 0.435 3.859
115.2 74.9 87.0 76.7 77.0 76.1
4.3 7.8 7.9
Tap water
0.05 0.10 0.50 1.00
0.044 0.109 0.523 1.077
87.5 78.2 109.3 73.1 104.6 72.4 107.7 78.5
9.4 2.9 2.3 7.9
Paddy water
0.50 1.00
0.392 0.823
78.4 74.2 82.3 76.6
5.4 8.1
Soil
0.10 0.50 1.00
0.082 0.462 0.791
82.3 73.4 92.3 77.9 79.1 76.3
4.0 8.6 7.9
GC (mg/kg)
0.6
0.4
0.2
0 0.9
Fig. 3. Correlation between the ELISA and GC for spiked samples.
An indirect competitive ELISA for pretilachlor was developed and successfully detected pretilachlor residue in water and soil samples. The assay was shown to have a high sensitivity, with a LOD of 6.9 ng/L and a working range from 0.001 to 1.000 mg/L, and a good correlation between ELISA and GC–ECD in the fortified samples. The accuracy and precision were well within the requirements of residue analysis. The maximum residue limits (MRL) of rice is 0.1 mg/kg in China, so the ic-ELISA may provide an alternative valuable analytical method for the determination of pretilachlor in environmental samples.
References
0.8
0.6 ELISA (mg/kg)
5. Conclusion
This work was co-supported by the Special Fund for Agroscientific Research in the Public Interest (No. 200903033).
1
0.3
The design of the hapten is a key step for obtaining high quality antibodies against small molecules. The optimum hapten for a selected target molecule should be a near perfect mimic of that molecule in structure and geometry, in electronic and hydrogenbonding capabilities and in its hydrophobic properties (Goodrow and Hammock, 1998). Herein this study, 3-MPA was coupled with the pretilachlor, and the carboxylic group is far away from the antigenic determinants(alkyl moiety and aromatic ring in the chloroacetanilide). Sulfur is resembled with chlorine in electronic structure and atom size according to the bioisosteric replacement. Such approach could ensure maximal sensitivity to pretilachlor. Hydrogen bonding, electrostatic interactions, Van der Waal’s forces and hydrophobic interactions participate in the recognition of antigen by antibody (Hennion and Barcelo, 1998). The sensitivity of the immunoassay is affected by some of the physicochemical parameters. Therefore, it is important to optimize assay condition such as organic solvent, ionic strength and pH. The sensitivity was improved 3.4-fold after optimization.
Acknowledgment
1.2
0
4. Discussion
1.2
Akiko, T., Hideko, M., Kuniaki, K., Masaaki, S., 1996. Monitoring of herbicides in river water by gas chromatography-mass spectrometry and solid-phase extraction. J. Chromatogr. A 754, 159–168. Chu, X.G., Hu, X.Z., Yao, H.Y., 2005. Determination of 266 pesticide residues in apple juice by matrix solid-phase dispersion and gas chromatography–mass selective detection. J. Chromatogr. A 1063, 201–210. Feng, P.C.C., Horton, S.R., Sharp., C.R., 1992. A general method for developing immunoassays to chloroacetanilide herbicides. J. Agric. Food Chem. 40, 211–214. Feng, P.C.C., Sharp, C.R., Horton, S.R., 1994. Quantitation of alachlor residues in monkey urine by ELISA. J. Agric. Food Chem. 42, 316–319. Feng, P.C.C., Wratten, S.J., Horton, S.R., Sharp, C.R., Logusch, E.W., 1990. Development of an enzyme-linked immunosorbent assay for alachlor and its application to the analysis of environmental water samples. J. Agric. Food Chem. 38, 159–163. Gabaldon, J.A., Cascales, J.M., Maquieira, A., Puchades, R., 2002. Rapid trace analysis of alachlor in water and vegetable samples. J. Chromatogr. A 963, 125–136. Goodrow, M.H., Hammock, B.D., 1998. Hapten design for compound-selective antibodies: ELISAS for environmentally deleterious small molecules. Anal. Chim. Acta 376, 83–91. Huang, B.H., 1997. Gas chromatographic determination of pretilachlor. Pesticides 36, 5–6. Hegedus, G., Krikunova, V., Belai, I., Eremin, S.A., Szekacs, A., 2002. An enzymelinked immunosorbent assay (ELISA) for the detection of acetochlor. Int. J. Environ. Anal. Chem. 82, 879–891. Hennion, M.C., Barcelo, D., 1998. Strengths and limitations of immunoassays for effective and efficient use for pesticide analysis in water samples: a review. Anal. Chim. Acta 362, 3–34. Liu, C.L., 2002. World Pesticide Volume: Herbicide, 1st ed. Chemical Industry Press, Beijing.
Z.-J. Liu et al. / Ecotoxicology and Environmental Safety 74 (2011) 1595–1599
Lu, X.Q., Zhang, B.H., Shi, H.Y., Wang, M.H., 2009. The research on enzyme-linked immunosorbent assay for fenvalerate. Food Agric. Immunol. 20, 147–154. Tanabe, A., Mitobe, H., Kawata, K., Sakai, M., 1996. Monitoring of herbicides in river water by gas chromatography–mass spectrometry and solid-phase extraction. J. Chromatogr. A 754, 159–168. Wang, S.L., Jin, S.H., Liu, F.M., 2005. Analytical methodology for propisochlor pesticide residues in the environment. Pesticides 44, 174–175. Yakovleva, J., Zherdev, A.V., Popova, V.A., Eremin, S.A., Dzantiev, B.B., 2003. Production of antibodies and development of enzyme-linked immunosorbent assays for the herbicide butachlor. Anal. Chim. Acta 491, 1–13.
1599
Yakovleva, J.N., Lobanona, A.I., Panchenko, O.A., Eermin, S.A., 2002. Production of antibodies and development of specific polarization fluoroimmunoassay for acetochlor. Int. J. Environ. Anal. Chem. 82, 851–863. Zeng, D.Y., Shi, H.Y., Li, B., Wang, M.H., Song, B.A., 2006. Development of an enzyme-linked immunosorbent assay for quantitative determination of quizalofop-p-ethyl. J. Agric. Food Chem. 54, 8682–8687. Zhang, Q., Wu, Y.R., Wang, L.M., Hu, B.S., Li, P.W., Liu, F.Q., 2008. Effect of hapten structures on specific and sensitive enzyme-linked immunosorbent assays for N-methylcarbamate insecticide metolcarb. Anal. Chim. Acta 625, 87–94.
Contents lists available at ScienceDirect
Ecotoxicology and Environmental Safety journal homepage: www.elsevier.com/locate/ecoenv
Development of an enzyme-linked immunosorbent assay for determination of pretilachlor in water and soil Zhen-Jiang Liu, Peng-Min Yu, Song Fang, Jia-qin Fan, Ming-Hua Wang n Department of Pesticide Science, College of Plant Protection, Nanjing Agricultural University, Jiangsu Key Laboratory of Pesticide Science, Key Laboratory of Integrated Management of Crop Diseases and Pests (Nanjing Agricultural University), Ministry of Education, Nanjing 210095, PR China
a r t i c l e i n f o
abstract
Article history: Received 2 April 2010 Received in revised form 29 March 2011 Accepted 17 April 2011 Available online 4 May 2011
An indirect competitive enzyme-linked immunosorbent assay (ic-ELISA) has been developed for detection of pretilachlor in water and soil. An immunogen was prepared from haptens of pretilachlor conjugated to bovine serum albumin(BSA). The specific polyclonal antibodies were obtained by immunizing New Zealand white rabbits. The influence of parameters including concentrations of methanol, ionic strength and pH values were optimized to improve the sensitivity of the assay. The optimized ELISA was shown to have a high sensitivity and specificity for pretilachlor. Under optimal conditions, the ELISA has demonstrated an 50% inhibitory concentration (IC50) value of 0.0359 mg/L with a limit of detection (LOD, IC10) of 6.9 ng/L. The cross-reactivities to some analogs of pretilachlor (acetochlor, butachlor, metazachlor and metalaxyl) were below 1.5%. The average recoveries of pretilachlor from distilled water, tap water, paddy water and soil were in the range of 77.0–115.2% between 0.005 and 5.0 mg/L. The results of ELISA for spiked samples were confirmed by GC–ECD with a high correlation coefficient of 0.9950(n ¼ 11). Thus, the ELISA proven to be a sensitive, specific, inexpensive and quantitative tool for detection of pretilachlor from four kinds of spiked samples. & 2011 Elsevier Inc. All rights reserved.
Keywords: Enzyme-linked immunosorbent assay Pretilachlor Hapten Polyclonal antibody
1. Introduction Pretilachlor [2-chloro-N-(2-ethyl-6-ethylphenyl)-N-(2-ethoxy1-methylethyl) acetamide] is a selective pre-emergent herbicide, which is widely applied to control grasses in rice. Pretilachlor has a low toxicity to human and mammals, however it is highly toxic to fishes (Liu, 2002). Due to its extensive application, pretilachlor could have an adverse impact on the environment especially on water body. It is highly desirable to develop a specific and sensitive method to monitor residues of pretilachlor in environmental samples. Several methods for the determination of pretilachlor have been described, such as gas chromatography (Huang, 1997) and gas chromatography–mass spectrometry (Tanabe et al., 1996; Chu et al., 2005; Akiko et al., 1996). They are characterized by low limits of detection and high precision and sensitivity. These regular methods, however, generally require specialized technicians and instrumentation, and their procedure is rather cumbersome and expensive, therefore unsuitable for screening large numbers of environmental samples. The immunoassay has received more and more concerns as a simple, sensitive and cost-effective technique and has became a reliable
n
Corresponding author. Fax: þ86 25 84395479. E-mail address: [email protected] (M.-H. Wang).
0147-6513/$ - see front matter & 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.ecoenv.2011.04.026
analytical tool for rapid screening analysis in recent years (Hennion and Barcelo, 1998). Since the first application of enzyme-linked immunosorbent assay (ELISA) for detection of alachlor in water has been reported by Feng et al. (1990), immunoassay techniques have been developed for determination of several chloroacetanilide herbicides from various matrices. In order to develop an immunoassay with high sensitivity and selectivity, the first key step is the design of immunogen. The general strategy to obtain the immunogen for chloroacetanilides relies on the modification of N-chloroacetyl moiety. Such modification preserves N-alkyl side chain, which is the unique structural feature of chloroacetanilide herbicide. Protein thiolating agents such as S-acetylmercaptosuccinic anhydride and N-Acetylhomocysteinethiolactone were utilized as cross-linkers to develop an ELISA of alachlor (Feng et al. 1990; Feng et al. 1994; Gabaldon et al. 2002), amidochlor, butachlor and metolachlor (Feng et al. 1992). In recent years, the strategy has been improved by utilizing 3-mercaptopropionic acid (3-MPA) as a cross-linker. Such improvement has been proved successfully in rapid immunoassays of acetochlor (Yakovleva et al. 2002; Hegedus et al. 2002) and butachlor (Yakovleva et al. 2003). So far, there were not any ELISAs for the detection of pretilachlor has been reported. In this paper, a sensitive ELISA was developed for detection of pretilachlor residues in water and soil samples based on polyclonal antibodies. Furthermore, the ELISA performance was evaluated with conventional GC–ECD in terms of precision and accuracy using fortified samples.
1596
Z.-J. Liu et al. / Ecotoxicology and Environmental Safety 74 (2011) 1595–1599
2. Materials and methods
2.6. Preparation of polyclonal antibody
2.1. Regents
Male New Zealand white rabbits weighing about 2 kg were immunized with PMPA-BSA according to the immunization protocol reported (Zeng et al., 2006). The rabbit had free access to drinking water and commercial standard laboratory diet (CZZ, Nanjing, China). It was housed according to the EEC 609/86 Directives regulating the welfare of experimental animals. The rabbits were bled after 8 days of the last injection. The blood was coagulated for 1 h at 37 1C and 12 h at 4 1C. The antiserum was centrifuged and then purified via the caprylic acid–ammonium sulfate precipitation method and freeze-dried.
All regents and solvents used were of analytical grade. Pretilachlor was obtained from the Changqing Agrochemical Co., Ltd. (Jiangsu, China). 1,3-dicyclohexylcarbo- diimide (DCC), N-hydroxysuccinimide (NHS), isobutylchlorocarbonate (TBA) and tri-n-butylamine (TEA) were acquired from Sinopharm Chemical Reagent Co., Ltd. (Shanghai. China). Bovine serum albumin (BSA), Ovalbumin (OVA), Freund’s complete and incomplete adjuvants, goat anti-rabbit IgG-horseradish peroxidase, o-phenylenediamine (OPD) and polyoxyethylene sorbitan monolaurate (Tween-20) were purchased from Sigma Chemical Co. (St. Louis, USA). 3-mercaptopropionic acid (3-MPA) was from A Johnson Matthey Company.
2.2. Buffers and solutions Phosphate buffered saline (PBS, 0.01 mol/L, pH 7.4). Carbonate-bicarbonate buffer saline (CBS, 0.05 mol/L, pH 9.6). Phosphate buffered saline contained 0.05% Tween-20 (PBST). The substrate solution contained 0.025 mol/L of citrate and 0.062 mol/L of sodium phosphate, pH 5.4. The OPD solution contained 0.4 mg/mL OPD and 0.012% H2O2 in substrate solution.
2.3. Instrumentation and equipment 1 H NMR spectrum was obtained via a DRX500 spectrometer (Bruker, Germany). MS spectrum was measured using a LC-MSQDECA (Finnigan, USA). IR spectrum was obtained with a Vertex 70FT-IR (Bruker, Germany). UV spectra were recorded on a DU 800 spectrophotometer (Beckman Coulter, USA). 96-Well Polystyrene microplates (Maxisorp) were purchased from Nunc (Roskilde, Denmark). ELISA plates were washed with a Wellwash Plus (Thermo, USA). Absorbances were read with an Infinite M200 microtiter plate reader (Tecan, Switzerland) at 490 nm. Aglient 7890A gas chromatography (Aglient, USA).
2.7. Protocol of ELISA Microplates were coated 2 h at 37 1C with 100 mL well 1 of the coating antigen diluted 1:6000 (v/v)in CBS. The plates were washed three times with PBST. 200 mL of PBS containing 1% (w/v) OVA was added and incubated for 30 min at 37 1C and the plates were washed again. For competition, 50 mL of sample or standard and 50 mL of diluted antiserum solution (0.5 mg/L) were incubated together at 37 1C. After incubation for 1 h, the plates were washed again. Then 100 mL/well of a diluted (1:3000) goat anti-rabbit IgG-horseradish peroxidase was added and incubated for 1 h at 37 1C. After another washing step, 100 mL/well of the OPD solution was added and incubated for 15 min at 37 1C. Finally 2 mol/L of sulfuric acid (50 mL/well) was added and the absorbance was measured at 490 nm. The standard curve for pretilachlor was obtained by napierian logarithm of plotting percent binding (logit (B/B0)) versus the logarithm of the pretilachlor concentration (log C). %(B/B0) was calculated by the following equation: %ðB=B0 Þ ¼ ½ðAx Amin Þ=ðAmax Amin Þ 100 where Ax is the absorbance of the sample, Amax is the absorbance in the absence of analyte and Amin is the absorbance of the background. IC50 and IC10 values were calculated.
2.8. Optimization of ELISA 2.4. Synthesis of hapten The hapten, pretilachlor-3-mercaptopropionic acid (PMPA), was synthesized according to the method described previously (Yakovleva et al. 2002, 2003). The synthesis route was illustrated Fig. 1. A mixture of 0.32 g (1.0 mmol) of pretilachlor, 0.15 g (1.4 mmol) of 3-MPA and 0.15 g (2.7 mmol) of potassium hydroxide in 20 ml of ethanol was refluxed for 6 h. The reaction mixture was filtered, and the filtrate was concentrated to a white solid. The solid was dissolved with 10 ml of 5% NaHCO3 and filtered. The solution was acidified to pH 3.0 with 6 mol/L HCl and extracted three times with 15 ml of ethyl acetate. The organic phase was washed with water and dried with Na2SO4. The PMPA was purified by TLC using a mixture of benzene: dioxane (5:1, v/v) as an eluent. The product (Rf ¼ 0.6) was extracted by ethanol and evaporated to obtain a yellow oily PMPA. The product was characterized by 1H NMR, IR and MS. 1H NMR (CDCl3), d7.15–7.31(m, 3H, ArH), 3.74–3.76(t, 2H, CH2O), 3.58–3.60(t, 2H, OCH2), 3.30– 3.32(t, 2H, NCH2), 2.92(s, 2H, COCH2S), 2.86–2.89(t, 2H, SCH2), 2.62–2.64(t, 2H, CH2COOH), 2.55–2.62(m, 4H, 2ArCH2), 1.48–1.52(m, 2H, CH2), 1.23–1.26(t, 6H, 2CH3), 0.83–0.86(t, 3H, CH3). IR (KBr, Vcm 1): 3500–3000 (v, O–H), 1725 (vs, carboxylic acid C ¼ O). MS, ESI m/z: 381 [M–H þ ], M ¼ 382.
The influence of some of the parameters such as organic solvent, ionic strength and pH were evaluated to improve the sensitivity of the ELISA. Competitive curves were carried out for pretilachlor standards in PBS containing 10–40% (v/v) of methanol, PBS buffer with concentrations of Na þ (0.1, 0.14, 0.2, 0.3, 0.4, 0.5 mol/L) and PBS buffers with pH values 4.5, 5.5, 6.5, 7.5, 8.5 and 9.5 to determine the effects of solvent, ionic strength and pH values.
2.9. Cross-reactivity determination Cross-reactivities (CR) were determined to evaluate the selectivity of the optimized ELISA. The cross-reactivities of four analogs of pretilachlor (acetochlor, butachlor, metazachlor and metalaxyl) were determined. CR was calculated by the following equation: CRð%Þ ¼ ðIC50 of pretilachlor=IC50 of analoguesÞ 100:
2.10. Evaluation of the ELISA with spiked samples 2.5. Preparation of immunogen and coating antigen The PMPA was coupled to BSA by active ester method (Zhang et al., 2008) and OVA via the mixed anhydride reaction (Zeng et al., 2006). The conjugates were dialyzed against PBS over 72 h at 4 1C and stored at 20 1C. PMPA-BSA and PMPAOVA were used as immunogen and coating antigen, respectively. The conjugated formation was confirmed spectrophotometrically. The absorbances of hapten, carrier protein and protein conjugate solutions (0.1 mg/L) at 280 nm were measured and hapten density was calculated directly by the mole absorbance e (Lu et al., 2009): Hapten density ¼ ðeconjugation eprotein Þ=ehapten
ELISA was applied for the determination of pretilachlor in water and soil samples. Water samples were collected from three sources: distilled water, tap water and paddy water. The water was spiked pretilachlor at various levels (0.05, 0.1, 0.5 and 5 mg/L). To reduce the matrix effect, water samples were diluted 10 times with PBS containing 10% methanol. 10 g of soil sample spiked with pretilachlor at 0.10, 0.50 and 1.00 mg/kg was mixed with 30 ml of acetonitrile and ultrasonic extraction for 10 min; the suspension was filtrated. The filtrate was added to 10 ml saturated sodium chloride solution and was vigorously shaken. The organic phase was concentrated and the remainder was dissolved with 10 ml PBS containing 10% methanol. The samples were analyzed with ELISA method by triplicate. Samples without pretilachlor were used as blanks in all case.
C2H5
C2H5 CH2CH2OC3H7
N COCH2Cl
KOH + HSCH2CH2COOH C2H5OH
C2H5
CH2CH2OC3H7 N COCH2SCH2CH2COOH C2H5
Fig. 1. Synthetic route of the hapten.
Z.-J. Liu et al. / Ecotoxicology and Environmental Safety 74 (2011) 1595–1599
1597
1.2
2.11. Evaluate of the assay with GC–ECD GC was performed according to the procedure of Wang et al. (2005). Samples included distilled water, tap water and paddy water and soil were spiked pretilachlor at various levels (0.01–1 mg/kg). Water samples were extracted with dichloromethane. Soil samples were extracted with a mixture of petroleum ether– acetone (1:1 v). The pretilachlor were measured by GC–ECD simultaneously and the measured results were compared with ELISA.
Logit (B/B0)
0.6
3. Result 3.1. Identification of conjugations In the UV–vis spectrometry obtained from continuous wavelength scanning, there were qualitative differences between the spectra of the conjugate and the corresponding carrier protein, especially near 280 nm and showed that the carrier protein and hapten had been coupled successfully. The hapten density was 11.9:1 and 8.5:1 for PMPA-BSA and PMPA-OVA, respectively.
0
-0.6
-1.2 -3.2
-2.2
-1.2 Log[pretilachlor] (mg/L)
-0.2
Fig. 2. Standard curves by ic-ELISA for pretilachlor.
3.2. Optimization of assay conditions IC50 is an important feature, which reflects the sensitivity of the immunoassay. The IC50 values and B0 of the pretilachlor of the optimization are shown in Table 1. It was observed that immunoassay is more sensitive as buffer salt concentration increased. However, the maximum signal sharply diminished when salt concentration more than 0.14 M. A most common water miscible solvent, methanol, was tested for its effect on assay performance. The best sensitivity was observed when the methanol concentration in PBS was 10%. The influence of pH on both the maximum signal and sensitivity was studied. It was found that the pH did not have a notable effect on the assay. The result seemed to indicate that electrostatic interactions are not the dominant forces in the recognition of antibody and the antigen in this assay.
Table 2 Cross-reactivity of pretilachlor and some of its analogs. Compounds
Structure
IC50(mg/L)
CR (%)
Pretilachlor
0.042
100
Acetochlor
7.6
0.55
3.3. Sensitivity of the assay
Butachlor
2.8
1.5
Under optimum conditions, the ic-ELISA procedures were conducted in triplicates with a series of concentration of pretilachlor. The standard curve was represented in Fig. 2. It was observed that the immunoassay had a linear relationship from 0.001 to 1.000 mg/L between logit(B/B0) and the logarithm of the pretilachlor concentration. An equation was obtained: logit (B/B0)¼ 0.591 Log[C] 0.8539, R2 ¼0.9909. The method with an IC50 of 0.0359 mg/L, a limit of detection (IC10) of 6.9 ng/L and a linear working range from 0.001 to 1.000 mg/L.
Metazachlor
4100
o0.1
Metalaxyl
4100
o0.1
Table 1 Effect of ionic strength, solvent, pH on the ELISA system. Ionic strength
Methanol concentration
Ionic strength (mol/L)
IC50 B0 (mg/L)
0.10 0.14 0.20 0.30 0.40 0.50
0.141 0.102 0.108 0.087 0.085 0.048
1.02 1.20 0.93 0.74 0.62 0.58
pH
Methanol concentration (v/v, %)
IC50 B0 (mg/L)
10 20 30 40 50
0.041 0.086 0.097 0.197 0.682
pH IC50 B0 (mg/L)
3.4. Specificity of the assay 1.18 1.13 1.16 1.07 1.17
4.5 5.5 6.5 7.5 8.5 9.5
0.026 0.036 0.045 0.042 0.041 0.040
1.34 1.28 1.15 1.21 1.22 1.30
The values of cross-reactivity of the antibody with pretilachlor and related compounds were listed in Table 2. It was found to be less than 1.5% of cross-reactivity for acetochlor, butachlor, metazachlor and metalaxyl. Despite the high similarity of the chloroacetanilide compounds, the ELISA has high specificity to
1598
Z.-J. Liu et al. / Ecotoxicology and Environmental Safety 74 (2011) 1595–1599
pretilachlor, which demonstrated that alkyl radicals moiety in the chloroacetanilide are the most important antigenic determinants. Besides, replacement of methyl in benzene ring resulted in lower of antibody reactivity. 3.5. Accuracy and precision of the ELISA The spiked recoveries and the relative standard deviation (RSD) were calculated to evaluate the accuracy and precision of the ELISA. As illustrated in Table 3, the recoveries were in the range of 77.0–115.2% and the RSD were found to be 2.3–9.4%. These data suggested that the matrix interference of water samples can be minimized by diluting the samples 10 times with PBS. The results demonstrated that this assay is suitable for the quantitative detection of pretilachlor at trace levels in water and soil. 3.6. Correlation between the ELISA and GC As shown in Fig. 3, good correlation was obtained between ELISA (x) and GC (Y) with the linear regression equation of Y ¼0.9487x þ0.001 (R2 ¼0.9950, n ¼11). These results suggested that pretilachlor in the samples could be simply, rapidly and accurately detected by ELISA. Table 3 Recoveries of pretilachlor in spiked water and soil samples. Samples
Mean detected Spiked concentration (mg/L) (mg/L, mg/kg)
Average recoveries (%)
RSD (%) (n¼3)
Distilled water
0.05 0.50 5.0
0.058 0.435 3.859
115.2 74.9 87.0 76.7 77.0 76.1
4.3 7.8 7.9
Tap water
0.05 0.10 0.50 1.00
0.044 0.109 0.523 1.077
87.5 78.2 109.3 73.1 104.6 72.4 107.7 78.5
9.4 2.9 2.3 7.9
Paddy water
0.50 1.00
0.392 0.823
78.4 74.2 82.3 76.6
5.4 8.1
Soil
0.10 0.50 1.00
0.082 0.462 0.791
82.3 73.4 92.3 77.9 79.1 76.3
4.0 8.6 7.9
GC (mg/kg)
0.6
0.4
0.2
0 0.9
Fig. 3. Correlation between the ELISA and GC for spiked samples.
An indirect competitive ELISA for pretilachlor was developed and successfully detected pretilachlor residue in water and soil samples. The assay was shown to have a high sensitivity, with a LOD of 6.9 ng/L and a working range from 0.001 to 1.000 mg/L, and a good correlation between ELISA and GC–ECD in the fortified samples. The accuracy and precision were well within the requirements of residue analysis. The maximum residue limits (MRL) of rice is 0.1 mg/kg in China, so the ic-ELISA may provide an alternative valuable analytical method for the determination of pretilachlor in environmental samples.
References
0.8
0.6 ELISA (mg/kg)
5. Conclusion
This work was co-supported by the Special Fund for Agroscientific Research in the Public Interest (No. 200903033).
1
0.3
The design of the hapten is a key step for obtaining high quality antibodies against small molecules. The optimum hapten for a selected target molecule should be a near perfect mimic of that molecule in structure and geometry, in electronic and hydrogenbonding capabilities and in its hydrophobic properties (Goodrow and Hammock, 1998). Herein this study, 3-MPA was coupled with the pretilachlor, and the carboxylic group is far away from the antigenic determinants(alkyl moiety and aromatic ring in the chloroacetanilide). Sulfur is resembled with chlorine in electronic structure and atom size according to the bioisosteric replacement. Such approach could ensure maximal sensitivity to pretilachlor. Hydrogen bonding, electrostatic interactions, Van der Waal’s forces and hydrophobic interactions participate in the recognition of antigen by antibody (Hennion and Barcelo, 1998). The sensitivity of the immunoassay is affected by some of the physicochemical parameters. Therefore, it is important to optimize assay condition such as organic solvent, ionic strength and pH. The sensitivity was improved 3.4-fold after optimization.
Acknowledgment
1.2
0
4. Discussion
1.2
Akiko, T., Hideko, M., Kuniaki, K., Masaaki, S., 1996. Monitoring of herbicides in river water by gas chromatography-mass spectrometry and solid-phase extraction. J. Chromatogr. A 754, 159–168. Chu, X.G., Hu, X.Z., Yao, H.Y., 2005. Determination of 266 pesticide residues in apple juice by matrix solid-phase dispersion and gas chromatography–mass selective detection. J. Chromatogr. A 1063, 201–210. Feng, P.C.C., Horton, S.R., Sharp., C.R., 1992. A general method for developing immunoassays to chloroacetanilide herbicides. J. Agric. Food Chem. 40, 211–214. Feng, P.C.C., Sharp, C.R., Horton, S.R., 1994. Quantitation of alachlor residues in monkey urine by ELISA. J. Agric. Food Chem. 42, 316–319. Feng, P.C.C., Wratten, S.J., Horton, S.R., Sharp, C.R., Logusch, E.W., 1990. Development of an enzyme-linked immunosorbent assay for alachlor and its application to the analysis of environmental water samples. J. Agric. Food Chem. 38, 159–163. Gabaldon, J.A., Cascales, J.M., Maquieira, A., Puchades, R., 2002. Rapid trace analysis of alachlor in water and vegetable samples. J. Chromatogr. A 963, 125–136. Goodrow, M.H., Hammock, B.D., 1998. Hapten design for compound-selective antibodies: ELISAS for environmentally deleterious small molecules. Anal. Chim. Acta 376, 83–91. Huang, B.H., 1997. Gas chromatographic determination of pretilachlor. Pesticides 36, 5–6. Hegedus, G., Krikunova, V., Belai, I., Eremin, S.A., Szekacs, A., 2002. An enzymelinked immunosorbent assay (ELISA) for the detection of acetochlor. Int. J. Environ. Anal. Chem. 82, 879–891. Hennion, M.C., Barcelo, D., 1998. Strengths and limitations of immunoassays for effective and efficient use for pesticide analysis in water samples: a review. Anal. Chim. Acta 362, 3–34. Liu, C.L., 2002. World Pesticide Volume: Herbicide, 1st ed. Chemical Industry Press, Beijing.
Z.-J. Liu et al. / Ecotoxicology and Environmental Safety 74 (2011) 1595–1599
Lu, X.Q., Zhang, B.H., Shi, H.Y., Wang, M.H., 2009. The research on enzyme-linked immunosorbent assay for fenvalerate. Food Agric. Immunol. 20, 147–154. Tanabe, A., Mitobe, H., Kawata, K., Sakai, M., 1996. Monitoring of herbicides in river water by gas chromatography–mass spectrometry and solid-phase extraction. J. Chromatogr. A 754, 159–168. Wang, S.L., Jin, S.H., Liu, F.M., 2005. Analytical methodology for propisochlor pesticide residues in the environment. Pesticides 44, 174–175. Yakovleva, J., Zherdev, A.V., Popova, V.A., Eremin, S.A., Dzantiev, B.B., 2003. Production of antibodies and development of enzyme-linked immunosorbent assays for the herbicide butachlor. Anal. Chim. Acta 491, 1–13.
1599
Yakovleva, J.N., Lobanona, A.I., Panchenko, O.A., Eermin, S.A., 2002. Production of antibodies and development of specific polarization fluoroimmunoassay for acetochlor. Int. J. Environ. Anal. Chem. 82, 851–863. Zeng, D.Y., Shi, H.Y., Li, B., Wang, M.H., Song, B.A., 2006. Development of an enzyme-linked immunosorbent assay for quantitative determination of quizalofop-p-ethyl. J. Agric. Food Chem. 54, 8682–8687. Zhang, Q., Wu, Y.R., Wang, L.M., Hu, B.S., Li, P.W., Liu, F.Q., 2008. Effect of hapten structures on specific and sensitive enzyme-linked immunosorbent assays for N-methylcarbamate insecticide metolcarb. Anal. Chim. Acta 625, 87–94.