Biosensors & Bioelectronics 7 (1992) 411-419

A piezoelectric immunobiosensor atrazine in drinking water

for

George G. Guilbault a*b, Bet-told Hocka & Rolf Schmidb aDepartment of Botany, Technische Universimt Miinchen-Weihenstephan, D-8050 Freising, Germany bGesellschaft fur Biotechnologische Forschung, Mascheroder Weg, D-3300 Braunschweig, Germany (Received 3 October 1991; revised verson received 6 January 1992; accepted 9 January 1992)

Abstract: A piezoelectric crystal immunobiosensor has been developed for the assay of atrazine herbicides in drinking water. Determinations from O-03-100 pg 1-l (parts per billion) of atrazine can be made with a relative SD of about f8%. Atrazine antibodies (polyclonal from sheep) are layered onto the gold electrode of 10MHz piezoelectric crystals, which are precoated with protein A. The sensor is reversible, being reusable for about eight or nine assays. Keywords: immunobiosensor, piezocrystal, herbicides, atrazine, antibodies.

INTRODUCTION Pesticides are important auxiliaries for achieving high agricultural yields. According to an estimation made in 1988 (Haberer et al., 1988a), about 30 000 tons of these substances were used in agriculture in the Federal Republic of Germany alone. As Table 1 shows, pesticides are usually classified according to their activity into groups comprising herbicides, fungicides, insecticides and pesticides having other actions. As analytical techniques improved, it became clear that traces of pesticide residues and metabolites were entering ground water and drinking water (Haberer et al., 1988b). Since pesticides accumulate in tissue and high concentrations can cause diseases in animal experiments (Datensammlung zur Toxicologic der Herbizide, 1986), there is considerable interest in producing pesticide-free drinking water. This requirement is taken into account by the European Drinking Water Act (EDWA) of 1980, which, in some of its statutory orders that were

enacted in the Federal Republic of Germany on 1 October 1989, requires that drinking water contains less than O.lpgl-’ of an individual pesticide, and less than 0.5pg 1-l of total pesticides. This concentration range can be detected only by sensitive physicochemical methods such as high-performance liquid chromatography (HPLC), gas chromatography and, if unambiguous identitication of a pesticide is required, by gas chromatography/mass spectrometry. Usually these procedures are preceded by enrichment steps which can render analysis cumbersome, especially in the case of hydrophilic pesticides. As a result, the procedures presently available are complex and often beyond the possibilities of the some 3000 drinking water suppliers in the Federal Republic. The search for simpler analytical methods for the determination of pesticides in drinking water is therefore of considerable importance in practice (Quentin et al., 1987). Biosensors-a combination of highly specific biological molecules (enzymes, antibodies, etc.),

0965-5663/92/$05.00 Q 1992 Elsevier Science Publishers Ltd.

411

Biosensors & Bioelectronics

G. G. Guilbault et al. TABLE 1 Estimated consumption of pesticides Germany, 1988 (Haberer et al., 1988a)

in the Federal

Republic

of

Total: 30 053 tons in 1839 formulations Herbicides Phenoxyacetic acid derivatives Triazinones Triazines Trichloroacetic acid Phenylurea derivatives Fungicides Phthalimides

8 475 tons (28%) 331 formulations

Insecticides Arylcarbamates Organophosphonates

1 563 tons (7%) 405 formulations

Others

2 615 tons (9%) 386 formulations

with electronic

or optical signal transducers optrodes, piezocrystals, etc.), have come onto the market in the last two decades (Schmid & Karube, 1988; Scheller & Schubert, (electrodes,

1989). A number of biosensor formats have also already been proposed for biological oxygen demand (BOD) and pesticide analysis (Kindervater et al., 1990). In connection with the problems described above, enzyme- or antibodybased flow injection analysis (FIA) (Schmid & Kunnecke, 1990) is therefore of significant interest, since it allows virtually continuous monitoring and is therefore a suitable concept for establishing ‘alarm systems’. We have been involved in the development of such FIA systems for analysis of pesticides for several years. Since the mid-1980s pesticide-specific antibodies have been developed (Haberer & Kramer, 1988) and rendered pesticide-specific FIA possible. However, the first report to our knowledge of the suitability of this test format for pesticide analysis originates only from 1989 (Kramer et al., 1989). Based on various polyclonal antibodies specific for atrazine* and on a competitive assay format with peroxidaselabelled haptens, a sensitivity of less than O-1part per billion (0.1 ,ug 1-l) could be achieved for various triazines (see Table 2). Further efforts can thus be directed towards development of a fully automatic system and then to field experiments. A membrane reactor which *All pesticide-specific antibodies used in these studies were kindly supplied by Professor B. Hock and his group at Technical University Munchen-Weihenstephan. 412

17 430 tons (58%), 717 formulations

allows automated removal of used antibody has already been developed with this aim in mind (Kramer & Schmid, 1991). Unfortunately, the FIA system is a one-time assay, and replacement of the antibody cartridge is necessary after use (Kramer & Schmid, 1990, 1991). The origin and theoretical foundation of piezoelectricity was first postulated by Raleigh in 1885. However, the use of piezoelectric (PZ) devices as potential chemical sensors was realized only after Sauerbrey derived the following equation describing the frequency-tomass relationship (Sauerbrey, 1959): AF = -2.3 X lo+ F2 AM/A where AF is the change in fundamental frequency of the coated crystal, F is the resonant frequency of the crystal, A is the area coated and AA4 is the mass deposited. The first analytical application of a PZ detector was reported by King in 1964. For the next two decades, intensive research was directed toward developing organic and inorganic coatings for the detection and TABLE 2 Comparison of the detection sensitivity of various triazines by FIA and ELISA (Kramer & Schmid, 1990, 1991)

Atrazine Propazine Simazine “50% inhibition

FIA OIg 1-l)

ELISA 018 1-l)

0~06”/0~01-1ob 0~02/0JlO2-10 0*5/0~1-10

0.08”/0.01-lb 0~02/0~002-1 0~5/0~01-30

valueslbmeasurement

ranges.

Biosensors & Bioelectronics

determination of various toxic agents in the environment and the workplace area (Guilbault & Jordan, 1988; McCallum, 1989). Since biologically active materials such as antibodies, enzymes and antigens are highly specific, their use as active coatings can be exploited, leading to a new class of PZ biosensors in liquid and gasphase analysis (Ngeh-Ngwainbi et al., 1990). The use of an antigen as a PZ crystal coating was first demonstrated by Shons et al. (1972). The crystals were precoated with a low-surface-energy plastic coating, nyebar C (30%) solution in 1,3di(trifluoromethyl)benzene, providing a layer capable of forming hydrophobic bonds with proteins. Crystals were then coated with bovine serum albumin (BSA), and subsequently exposed to a solution containing anti-BSA to determine the antibody activity. The sensitivity of this technique was equal to or better than that of the conventional passive agglutination method. An improved indirect assay method for the determination of antigens was disclosed by Oliveira & Silver (1980). PZ crystals were coated with the antigen or a protein mixture containing a predetermined amount of antibody and free antigen. The decrease in frequency is inversely related to the antigen concentration in the sample. A wide range of antigenic materials ranging from low-molecular-weight compounds to large macromolecules could be determined using this technique, with sensitivities as low as nanograms per millilitre for many significant small molecules. A method for determining both the type of antibody subclass and the concentration of antibody present was also patented by Rice (1980). The first assay in solution was reported by Roederer & Bastiaans (1983). A PZ quartz surface acoustic wave (SAW) device was developed to detect human IgG in solution. The surface of an ST-cut crystal was modified with glycidoxypropyltrimethoxy-silane (GGPS), a saline derivative with high reactivity to proteins. An antibody, goat antihuman IgG, was then immobilized on the crystal and tested for the detection of IgG. The detection limit of the technique was found to be 13pg ml-’ and the calibration curve was linear over an IgG concentration range of 0,023-2.25 mg ml-‘. above-mentioned findings, Unlike the Muramatsu et al. (1987) demonstrated that an immobilized layer of protein A placed onto the

A piezoelectric immunobiosensor for atrazine

surface of crystals modified with (y-aminopropyl)triethoxy-silane could be successfully utilized to determine the concentration of IgG and its subclass in solution. A continuous liquid-phase PZ biosensor for kinetic immunoassay was developed by Davis & Leary (1989). The proposed system was used to continuously monitor the frequency during protein-protein interactions, facilitating the observation of both the extent and the kinetics of these interactions. A frequency change of approximately 1 Hz for each 10 ng of added immunoglobulin was observed. In a previous study (Kindervater et al., 1990) we described an immunological atrazine determination using a piezosensor, but the sensitivity was quite poor (55 Hz mg-‘) because of the rabbit antibodies used. In this publication we demonstrate the use of atrazine antibodies coated with protein A on a PZ crystal detector for the direct label-less assay of atrazine in water samples. The assay is sensitive (sub-p.p.b.) and highly specific.

EXPERIMENTAL Anti-atrazine

antibody S84

S84 is an anti-atrazine

antibody obtained from sheep. An atrazine derivative with an aminohexanoic acid residue in position 4 was provided from Riedel-de Haen AG, Seelze. This derivative was covalently linked with cyclohexylmorpholinocarbodiimide (CMC) to the amino groups of keyhole limpet haemocyanin (KLH): 50 mg KLH were dissolved in 10 ml distilled water, followed by 17 mg CMC. A solution of 12 mg atrazine derivative in 1 ml dimethylformamide was added dropwise and the whole solution was incubated for 16 h at 20°C. Two sheep were immunized with this conjugate on a weekly schedule: (1) three intramuscular injections with 0.5 mg conjugate in 0.5 ml 0.9% NaCl and 0.5 ml Freund’s adjuvant incomplete, mixed with 10 ~1 Bordeteh pertussis suspension; (2) intravenous booster injections with 0.5 mg conjugate in 0.5 ml 0.9% NaCl. The first blood was collected four weeks after the last injection and the crude antisera were handled as described before (Huber & Hock, 1986). After a clean-up step with ammonium 413

Biosensors & Bioelectronics

G. G. Guilbault et al.

sulphate the IgG fraction was lyophilized for storage. Only one of the two sheep (no. 84) showed an immunoresponse. Avid antibodies were obtained from the bleedings of the tenth week after immunization (Table 3). An enzyme immunoassay performed with a peroxidase-atrazine-tracer showed a detection limit for atrazine of 30 ngl-‘, and a range of O-03-3 pug1-l using ELISA. The protein content of this antibody is 28 mg ml-‘. Antibody dilutions were made using phosphate buffer, 0.1 M, pH 7.

IMMOBILIZATION

METHODS

Immobilization via polyetbyleneimine (PEI) The PZ crystal was dipped for 15 s in a 2% (in methanol) solution of PEI (Aldrich); the crystal was air dried, then washed with methanol to remove excess unbound material. The crystal was then immersed in a 2.5% glutaraldehyde (pH 7) solution for 14 h. Following washing, 3 ~1 of a 5 mg ml-’ antibody solution was placed on each side of the crystal with a syringe. To block all unreacted aldehyde groups the crystal was then immersed in a solution of 0.1 M glycine in 20 mM phosphate-buffered saline (pH 7.0). The crystal was washed with distilled water and dried. Silanization method

TABLE 3

Atrazine Propazine Deethylatrazine Prometryn Simazine Terbutylazine Simetryn Terbutryn Ametryn Aziprotryn

414

Protein A-gold method The crystals were prepared using the previously reported protein A-gold method (Davis & Leary, 1989). The 10 MHz quartz crystals were soaked in 1.2 N NaOH for 30 min, rinsed with distilled water, then immersed in 1.2 N HCl for 5 min. Then 50~1 of concentrated HCl were placed on the gold electrodes (both sides) of the crystal for 2 min, with special care to keep the acid from the electrode leads. The crystals were rinsed with water, and dried in air for 30 min. Then 3~1 of protein A solution (1 mg dissolved in 1 a0 ml of 0.1 M phosphate buffer, pH 7.2 and 1.0 ml of O-1 M acetate buffer, pH 4.5, final pH -5.5) were placed on each side of the PZ crystals on the entire surface of the gold electrodes. After 1 h drying, 3 ~1 of atrazine antibody solution (diluted with phosphate buffer, pH 7.2, 1 : 25 1.1 mg ml-‘) was added to each surface, dried, then washed. The final AF (PI - Fo, Fig. 1) was measured after addition of protein A and then antibody to assure a good coating. Pesticides

The PZ crystal was immersed in a 5% (acetone) solution of y-aminopropyltriethoxy-silane (APTES) (Sigma) for 1; h at room temperature. The crystals were dried at 100°C in an oven, washed with acetone, then dipped in 1%

Substance

glutaraldehyde (pH 7) for 1.5 h. Following washing with distilled water, 3 ~1 of a 5 mg ml-’ antibody solution was spread on both sides of the crystal. After drying for 30 min the crystals were washed with PBS and then with distilled water.

S84, 10th week 100~0 115.5 12-9 10.1 1.1 1.2 0.4

A piezoelectric immunobiosensor for atrazine in drinking water.

A piezoelectric crystal immunobiosensor has been developed for the assay of atrazine herbicides in drinking water. Determinations from 0.03-100 microg...
868KB Sizes 0 Downloads 0 Views