Eur Food Res Technol (2002) 214:138–142 DOI 10.1007/s00217-001-0431-8

O R I G I N A L PA P E R

Holger Hrenn · Carsten Jahn · Wolfgang Schwack

Formation of protein-bound residues of the fungicide chlorothalonil in tomatoes (Lycopersicon esculentum Mill.)

Received: 17 July 2001 / Revised version: 7 September 2001 / Published online: 18 December 2001 © Springer-Verlag 2001

Abstract Adding the fungicide chlorothalonil to homogenized tomatoes leads to formation of bound residues in a pH dependent reaction. Recoveries of the fungicide ranged from 84% in pulps with natural pH (4.2) to 7% in pulps with alkaline pH (10.0). Incubation experiments with different model constituents, e.g. carbohydrates, organic acids, and proteins, showed that only the latter were responsible for the very low recoveries of the fungicide. Accordingly, in spiked protein solutions and in spiked tomato pulp, protein-bound residues of chlorothalonil were detected by a dot-blot immunoassay as a structure-specific method. Keywords Chlorothalonil · Bound residues · Tomato · Dot-blot immunoassay

Introduction Bound pesticide residues were first mentioned in the literature by Bailey and White in 1964 [1]. The term bound residue refers to residues which are associated with one or more classes of macromolecules [2]. The main topics of early studies, mainly obtained by using radiolabeled pesticides, were related to the fate of pesticides in soil. Recently, the focus has changed to the formation of bound residues in food. In addition to radiotracing studies, structure-specific methods for detection of bound residues (e.g. ELISA) [3, 4, 5, 6] are also available. Chlorothalonil [2,4,5,6-tetrachloroisophthalonitrile (1)] is a widely used non-systemic fungicide with protective action. According to a report on monitoring for pesticide residues [7] 1 was often found on food crops in many countries. This fungicide is frequently applied in tomato cultivation; a maximum residue level was set by the EU for this fruit crop at 2 mg/kg [8]. After applicaH. Hrenn · C. Jahn · W. Schwack (✉) Institut für Lebensmittelchemie, Universität Hohenheim, Garbenstrasse 28, 70593 Stuttgart, Germany e-mail: [email protected] Tel.: +49-711-4593979, Fax: +49-711-4594096

tion, 1 remains on the surface of plants, where it is exposed to environmental influences. During photochemical studies, we found dechlorination reactions of 1, photoaddition reactions with unsaturated lipids, and also formation of plant cuticle bound residues [4, 9]. The problem of low recoveries in residue analysis of chlorothalonil (1) is well known [10]. Gilsbach et al., who reported the results of an inter-laboratory pesticide study, observed a very low recovery of 1 in extraction from milk powder [11]. Tomatoes are industrially processed to a large extent. One of the major industrial treatments is peeling the fruits by steam or lye [12,13]. The latter procedure causes an increase in pH of up to 10 for at least 5 min, possibly resulting in unknown reactions between the fungicide and fruit constituents. Therefore, it is important to know if low recoveries can also be expected in processed tomatoes, e.g. tomato puree and if this is a result of decomposition of the active substance or of reactions with food constituents. Thus, the intention of the present study was to get a deeper insight into the reactions of 1 in foods. Degradation studies in the presence of several model substances (carbohydrates, organic acids, phenolic compounds, and proteins) and also in tomato pulps should reveal the reasons for low recoveries. For the detection of bound and conjugated residues, a dot-blot immunoassay had to be developed using monoclonal antibodies, which have previously been successfully used for the analysis of bound residues of 1 in isolated fruit cuticles [3, 4].

Materials and methods Materials Tomatoes (Lycopersicon esculentum Mill.; cultivars are not known) were obtained from our own cultivations to ensure untreated fruits. To obtain homogenous samples, 1.5 kg (fresh weight) were peeled and minced. Since the composition of the fruits varies with the provenance and the conditions of growth [14], the same batch was used for all investigations in this study.

139 Chemicals Chlorothalonil (1) was isolated from Daconil (ISK Biotech, Mentor, USA) by extraction with acetone (100 ml/g), and the crude product was re-crystallized from acetone {purity: >99.5% (HPLCDAD, λ=234 nm), melting point 251–253 °C (250–251 °C [15])}. Water was purified by a Milli-Q 185 plus system (Millipore, Bedford, USA). Tween 20, 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium solution (BCIP/NBT), a premixed electrophoresis reagent, and bovine serum albumin (BSA) (fraction V) were obtained from Sigma (Steinheim, Germany). Nitrocellulose membrane (90×105 mm sheets, 0.45 µm) was obtained from Pharmacia Biotech (Freiburg, Germany) and cut into pieces of 15×35 mm before use. All other reagents and solvents (distilled before use) were of analytical grade, purchased either from Fluka (Steinheim, Germany) or Merck (Darmstadt, Germany). Incubation buffer [sodium chloride (120 mmol, 7.013 g), potassium chloride (2.7 mmol, 201 mg), potassium dihydrogen phosphate (1.97 mmol, 268 mg), disodium hydrogen phosphate (8.03 mmol, 1.43 g), sodium azide (3.08 mmol, 200 mg), BSA (2 g) and Tween 20 (1 ml)] were filled up with water in a 100 ml volumetric flask. The pH was adjusted to pH 7.4 with 0.1 M sodium hydroxide solution and the solution was diluted by water (1+9, v+v) before use. The blocking buffer was 250 µl of 1% BSA (w/v) in incubation buffer, and the washing buffer, phosphate buffered saline (pH 7.4) containing 0.05% (v/v) Tween 20. Hybridoma supernatants of mAb chl. 4/11 [3] were diluted with glycerol (1+1, v+v) and stored at –18 °C until use. Secondary antibody [alkaline phosphatase-conjugated affinity purified goat anti-mouse IgG (H+L), minimal crossreaction to human, bovine and horse serum proteins] was purchased from Jackson ImmunoResearch (West Grove, USA). Equipment HPLC analysis. The analytical HPLC system comprised an HP1100 autosampler, HP1100 gradient pump, and HP1100 DAD module (Hewlett-Packard, Waldbronn, Germany); DAD detection wavelength 234 nm, spectral band width (SBW) 8 nm, reference 500 nm (SBW 100 nm). For data acquisition and processing, HP ChemStation (Rev. A.04.02) software was used. Chromatographic separations were performed at 25 °C on a 5 µm Eurospher 100– C18 column (250×4 mm) including a pre-column (Nucleosil 5 µm C18, 5×4 mm) (Knauer, Berlin, Germany). The mobile phase consisted of methanol (A) and 20 mM phosphate buffer (pH 4.0) (B). The gradient program was: 60% A held for 5 min, followed by an increase to 80% A over 10 min, followed by an increase to 100% A over 3 min. A reduction to 60% A over the next 2 min followed, and this was held for 5 min; the flow rate was 0.8 ml/min. The injection volume was 20 µl. Incubation steps and color development of the dot-blot immunoassay were performed on a M-1000 Microplate Shaker (MedTec, USA). Methods Recovery of chlorothalonil depending on the pH of the pulp. The pH of tomato pulp (20 g) was adjusted by adding solid dipotassium hydrogen phosphate and potassium dihydrogen phosphate in varying amounts until pH 5.5, 6.0, 6.5, 7.0, 7.5, 8.0, 9.0, or 10.0 was reached. The amount of each salt necessary for the desired pH was calculated from a buffer table [16]. If necessary, the pH values were finally corrected by adding 1 M sodium hydroxide solution. To each batch 0.5 ml of a solution of 1 in methanol (200 mg/l) was pipetted and the pulp was stirred for 2 h at room temperature. Acetone (50 ml) was added to the pulp, homogenized and filtered through a Buchner funnel. The filtrate was extracted three times with tert-butyl methyl ether. The organic phase was dried with anhydrous sodium sulfate, the solvent removed in vac-

uo and the residue dissolved in methanol (5 ml). The resulting solution was filtered (0.45 µm) and analyzed by HPLC. Recovery of chlorothalonil depending on the incubation time. Amounts of 2.5 ml of a solution of 1 in methanol (200 mg/l) or 2.5 ml of a suspension of Daconil in water (268 mg/l), respectively, were added to tomato pulp (100 g) and the mixtures stirred at room temperature. At regular intervals, aliquots (10 g) were taken and analyzed as described above. Recovery of chlorothalonil from model incubations. Aqueous solutions of fructose (1%, w/v), starch (1%, w/v), BSA (1%, w/v), citric acid (5%, w/v), and phenolic compounds (suspensions, concentrations about 0.1 %, w/v) were prepared. To obtain neutral solutions the pH was corrected by adding 0.1 M sodium hydroxide solution, if necessary. The acidic solutions (pH 4.5) were made of citrate phosphate buffer [citric acid monohydrate (23.5 g) and disodium hydrogen phosphate dihydrate (31.4 g) dissolved in water (1 l)] with the same concentrations as the neutral solutions. Each solution (50 ml) was spiked with 500 µl of a methanolic solution of 1 (0.5 mg/ml), stirred for 2 h, and analyzed as described above. Dot-blot immunoassay. Tomato pulp (20 g) was spiked with 1 (1 and 10 mg/kg) and stirred for 2 h. Five microliters of each sample (tomato pulp and BSA solution from model incubations with 1) were placed on a nitrocellulose sheet (15×35 mm) and the spots allowed to desiccate. After a blocking step (10 ml blocking buffer, 1 h) a primary incubation was performed (10 ml incubation buffer, 200 µl of 1+1 diluted hybridoma supernatant of mAb chl. 4/11, 1 h). An incubation step with a secondary antibody labeled with alkaline phosphatase (10 ml incubation buffer, 5 µl secondary antibody solution, 1 h) and, finally, a color development step followed (10 ml BCIP/NBT solution, 10 min). After each incubation step, rinsing was performed with buffer solutions (3×15 ml washing buffer, 10 min each). Additionally, the dot-blot immunoassay was performed adding 100 µg of 1 (100 µl of 1 g/l of 1 in methanol) during the primary incubation.

Results and discussion Recovery of chlorothalonil from tomato pulps Incubation of raw tomato pulp samples (pH 4.2), spiked with 1 resulted in a recovery decrease to about 80% after only 30 min. Thereafter for up to 2 h, recoveries were found to be constant, indicating that no further reactions occurred. Hence, the results proved an incubation time of 2 h to be sufficient for further studies. In terms of degradation rates and kinetics, tomato pulps spiked with the commercial formulation Daconil gave comparable results. Thus, there are no or negligible influences of formulation additives on the reaction rate and turnover. For the determination of pH dependent recovery rates, tomato pulps were set to different pH values (pH 4.2–10.0) and incubated with 1 for 2 h. While the experiments performed under gentle acidic conditions showed recoveries of about 85% (Fig. 1), a remarkable loss of 1 was noticed above pH 6.5, which reached recoveries of only 7% at pH 10.0. Recovery from model incubations The observed decrease in recovery with increasing pH values indicates a base-catalyzed mechanism in the for-

140

Fig. 1 Recovery of chlorothalonil (1) from tomato pulp at different pH values

mation of non-extractable residues. To prove this assumption, different food constituents, representing inherent components of many fruits and vegetables (carbohydrates: fructose, starch; proteins: BSA as model protein; organic acids: citric acid; phenolic compounds: rutin,

Fig. 2 Scheme of the dot-blot immunoassay

quercetin) were incubated in aqueous solutions (pH 4.5 and pH 7.0) in the presence of 1; the chosen concentrations of model food components were comparable to those found in tomatoes. The determined recoveries of 1 were close to 100% in all cases except for the incubation with the model protein BSA at pH 7.0. In this case only 11% of 1 could be extracted after an incubation time of 2 h. Therefore, proteins seem to be mainly responsible for the very low recovery of the fungicide under neutral conditions. This result can be explained as follows: 1, owning an electrophilic aromatic core, can easily react with nucleophiles, e.g. thiols or primary amines as functional groups present in side chains of proteins [17]. At pH 4.5 the majority of amino groups are protonated, resulting in low reactivity and in high recovery rates. At pH>7.0 partially deprotonated amino groups, e.g. those of lysine side chains, can react with 1 resulting in the formation of conjugates. Thus, the results obtained from model incubations showed that proteins could also be re-

141

sponsible for decreasing recoveries found in tomato pulps. Dot-blot immunoassay The monoclonal antibody mAb chl. 4/11 was obtained from an immunization of mice with hemocyanin-(2,4,5trichloro-6-(2-hydroxyethoxy)isophthalonitrile)-conjugate as immunogen [3]. Since mAb chl. 4/11 has proven its high affinity toward derivatives of 1 with amines or thiols (for cross-reactivities see [3, 4]), this antibody was employed to detect protein-bound residues of 1 in BSA solutions (pH 7.0) and tomato pulps (natural pH). For this purpose, a dot-blot immunoassay was developed, which is shown schematically in Fig. 2. It is a non-competitive antibody method based on the adsorptive binding of proteins to a nitrocellulose membrane. If, during a first incubation step, an antigen-antibody complex is formed between a residue covalently bound to adsorbed protein and an added primary antibody (mouse IgG), this reaction will be detectable by a color development step after a further incubation step with an enzyme-labeled secondary antibody (anti-mouse IgG). The phosphate group of BCIP is cleaved by the enzyme conjugated to the secondary antibody. The BCIP tautomerizes forming a ketone and, under alkaline conditions, dimerization occurs. While dimerizing, NBT is reduced and precipitates forming an intense blue deposition of diformazan [18]. In the case of a positive result, a blue dye is formed from a substrate, which is detected visually. Being non-competitive, dot-blots are not disturbed by small amounts of extractable residues. When this type of immunoassay was applied to the above-mentioned incubation attempts, blue colored spots developed on the nitrocellulose membrane indicating protein-bound residues both in spiked BSA solutions and spiked tomato pulps (Fig. 3). For the BSA experiments, incubation solutions had the same protein concentrations as found in tomatoes, and the spiking level of 1 was 5 mg/l in both cases. As discussed above, incubations at neutral pH resulted in smaller recoveries of 1 compared to the experiments at acidic pH (4.5) due to the formation of bound residues. Accordingly, the blue spots in Fig. 3a show a pH dependent increase in color intensity (pH 4.5 vs. pH 7.0) corresponding to the quantity of BSA-bound residues and correlating with the decreased recovery of 1. Using spiked tomato pulps of natural pH, the color intensity increased with elevated spiking levels (Fig. 3b; 1 mg/kg vs. 10 mg/kg). For nonspiked samples used as controls, the assay was generally negative (Fig. 3a, b). By adding an excess of 1 to the incubation step with the primary antibody (mAb chl. 4/11), the color development was suppressed, indicating that bound residues and the free fungicide react with the same antigen-binding site of the antibody. The results unequivocally demonstrate that the formation of protein conjugates is responsible for low recoveries of 1 from tomatoes and other food plant samples.

Fig. 3a, b Detection of bound chlorothalonil residues by dot-blot immunoassay: a in aqueous bovine serum albumin solutions (1%, w/v). i Without spiking, pH 7.0; ii spiking 5 µg/ml, pH 4.5; iii spiking 5 µg/ml, pH 7.0; and b in tomato pulp (natural pH: 4.2). i Without spiking; ii spiking 1 mg/kg; iii spiking 10 mg/kg

On the basis of the obtained results it seems justifiable to expect that the reaction of 1 with proteins may be important during the industrial lye peeling procedure of tomatoes. As this process is performed at high pH values and high temperature, the formation of protein-bound residues very likely becomes a favored reaction. Furthermore, low recoveries of the fungicide in residue analysis could also be explained by reactions with proteins during sample preparation. This becomes more evident by the instruction in official methods of residue analysis to neutralize acidic plant samples before work-up [19], enhancing the formation of bound residues. Further research is necessary to elucidate the structures of the protein conjugates of 1. Acknowledgements We are grateful to A. Sprick for sample preparation during her research course in food chemistry and Dr. Arnd Petersen (Forschungszentrum Borstel, Abteilung Biochemische und Molekulare Allergologie, Germany) for providing mAb chl. 4/11. The financial support of the “Deutsche Forschungsgemeinschaft” is gratefully acknowledged.

References 1. Bailey GW, White JL (1964) J Agric Food Chem 12:324–332 2. Skidmore MW, Paulson GD, Kuiper HA, Ohlin B, Reynolds S (1998) Pure Appl Chem 70:1423–1447 3. Jahn C, Zorn H, Petersen A, Schwack W (1999) Pestic Sci 55:1167–1176 4. Jahn C, Schwack W (2001) J Agric Food Chem 49:1233–1238 5. Giersch T, Sohn G, Hock B (1993) Anal Lett 26:1831–1845 6. Weller MG, Zeck A, Pfortner P, Simon E, Niessner R (1999) Int J Environ Anal Chem 75:201–215 7. European Commission Directorate General (1996) Consumer policy and consumer health protection: monitoring for pesticide residues in the European Union and Norway – report 8. Commission Directive 2001/35/EC of 11 May 2001 (2001) Official Journal of the European Communities, L 136, 18.5.2001 pp. 42–48

142 9. Zorn H (1997) Verhalten und Reaktivität des Fungicids Chlorothalonil auf Fruchtoberflächen unter besonderer Beachtung der photochemischen Bildung gebundener Rückstände. PhD thesis, University of Hohenheim, Shaker Verlag, Aachen, Germany 10. El-Bidaoui M, Jarju OP, Maestroni B, Phakaeiw Y, Ambrus A (2000) Spec Publ R Soc Chem 256:75–88 11. Gilsbach W, Thier H-P (1998) Lebensmittelchemie 52:30–32 12. Schade JE, Schultz WG, Neumann HJ, Morgan JP, Finley JW (1980) J Agric Food Chem 28:512–515 13. Gould W (1992) Tomato production, processing, and technology. CTI, Baltimore, Md. USA, pp 164–177 14. Herrmann K (1979) Z Lebensm Unters Forsch 169:179–200 15. Tomlin C (1994) The pesticide manual. 10th edn. BCPC, Binfield, UK

16. Rauscher K, Voigt J, Wilke I, Wilke KT (1986) Chemische Tabellen und Rechentafeln für die analytische Praxis. 8. Auflage VEB Deutscher Verlag für Grundstoffindustrie, Leipzig, Germany, pp 162–168 17. Jahn C (2000) Analytik cuticulagebundener Rückstände des Fungicids Chlorthalonil unter besonderer Berücksichtigung immunchemischer Nachweisverfahren. PhD thesis, University of Hohenheim, Shaker Verlag Aachen, Germany 18. McGadey J (1970) Histochemie 23:180–184 19. Amtliche Sammlung von Untersuchungsverfahren nach §35 LMBG(1999) BgVV (ed) L 00.00–34, Behrs Verlag, Hamburg, Germany