Development of Polyclonal Antibodies Against Domoic Acid for Their ...

MARCEL DEKKER, INC. • 270 MADISON AVENUE • NEW YORK, NY 10016 ©2003 Marcel Dekker, Inc. All rights reserved. This material may not be used or reproduced in any form without the express written permission of Marcel Dekker, Inc.

ANALYTICAL LETTERS Vol. 36, No. 9, pp. 1851–1863, 2003

Development of Polyclonal Antibodies Against Domoic Acid for Their Use in Electrochemical Biosensors Matthias Kania,2 Mark Kreuzer,1 Eric Moore,1 Miloslav Pravda,1 Bertold Hock,2 and George Guilbault1,* 1

Sensor Development Group, Department of Chemistry, University College Cork, Ireland 2 Department of Plant Sciences, Centre of Life Sciences, Technische Universitaet Muenchen, Alte Akademie, Freising, Germany

ABSTRACT The current work describes the raising of polyclonal antibodies (pAbs) against domoic acid (DA), an algal toxin produced by the diatom Pseudonitzschia pungens. They were screened for sensitivity and selectivity using a competitive enzyme-linked immunosorbent assay (ELISA). The antiserum produced against a keyhole limpet haemocyanin (DA-KLH) conjugate displayed a high affinity for free DA. The optimized horseradish peroxidase (HRP) ELISA had a detection limit of 0.6 ng mL1 (ppb) and a working range of

*Correspondence: George Guilbault, Department of Chemistry, University College Cork, Western Road, Cork, Ireland; E-mail: [email protected]. 1851 DOI: 10.1081/AL-120023618 Copyright & 2003 by Marcel Dekker, Inc.

0003-2719 (Print); 1532-236X (Online) www.dekker.com


1852

Kania et al. 0.8–300 ppb DA applying a streptavidin-biotin amplification system (ABC system). Furthermore this antiserum did not cross-react with similar chemical structures and algal toxins such as kainic acid, aspartic acid, glutamic acid, geranic acid and 2-methyl-3-butenoic acid. When the ELISA was compared using an alkaline phosphatase (AP) label we found the system to behave in a similar manner to the optimized HRP system but the linear range was smaller in the high DA concentration range. These pAbs were then used in the optimization of a screen-printed electrode (SPE) system for measurement of DA. A disposable screen-printed carbon electrode coupled with amperometric detection of p-aminophenol at þ300 mV vs. Ag/AgCl, produced by the enzyme AP, was used for signal measurement. The sensor incorporates a relevant range for toxin detection, by which humans become ill (Iverson, F.; Truelove, J. Toxicology and seafood toxins: domoic acid. Natural Toxins 1994, 2, 334–339.) with detection limits achieved at SPE to the order of ppb. The SPE system is simple and cost-effective due to its disposable nature, and analysis time is complete in 30 min. In addition, recovery experiments on DA for both ELISA and SPE highlighted the functionality of these systems yielding a 12% deviation for the true value for the ELISA using AP and 25% for the sensor. Key Words: Domoic acid; Polyclonal antibodies; ELISA; Screenprinted electrode.

INTRODUCTION A large variety of poisoning arise, in seafood as with terrestrial based foods, after ingestion of certain low molecular weight marine toxins. Toxin-producing bacteria and toxic phytoplankton are typically implicated in seafood-borne poisonings. An example of such poisoning is Amnesic Shellfish Poisoning (ASP) due to DA. DA is a potent neuroexcitatory toxins, that interferes with neurotransmission in the central nervous system.[2] Since the first outbreak of food poisoning occurred at Prince Edward Island, Canada in 1987, DA has been found to accumulate in edible shellfish. Upon ingestion, it causes the human illness syndrome ASP. The source of the toxin was shown to be the marine diatom Pseudonitzschia pungens forma multiseries.[3] The incidence of Pseudonitzschia blooms is increasing world-wide, and toxic blooms have been reported in many localities.[4,5] For protection against ASP, it is necessary to develop practical and sensitive methods of rapid detection able to measure trace levels of DA. Analysis of DA is currently performed by mouse bioassay,[6–8] HPLC,[9,10] thin-layer


Development of Polyclonal Antibodies

1853

chromatography,[11] and immunochemical methods.[12–14] The ideal assay for detection of marine toxins should be simple, cost effective, highly sensitive and specific. For these reasons, methods involving immunological analyses have become popular in recent years. Our groups have developed/produced antibodies, chosen specifically for their impact on the environment and seafood industry and used them to develop an immunosensor, which can rapidly assess, with high accuracy and selectivity, trace levels of DA.

EXPERIMENTAL Reagents Plastic microtitre plates, Falcon Pro-bind, were purchased from Becton Dickinson (New Jersey, USA). Bio-Tek Instruments (Vermont, USA) supplied the microplate washer (model ELP-40) and reader (model EL311). Incubations at elevated temperatures were carried out in a thermostated oven supplied by Heraeus Instruments. Electrodes were prepared using a DEK (model 247) screen-printer (Dorset, UK). Electrochemical measurements were performed using the BAS 100 W electrochemical workstation supplied by BioAnalytical Systems (West Lafayette, Indiana). Toxins and antibodies used in electrochemical work, including the following; The National Research Centre (NS, Canada) supplied domoic acid calibration solution (DACS-1C). Bovine serum albumin (BSA), biotinylated goat a-rabbit antibody, streptavindinHRP conjugate, tris(hydroxymethyl)aminomethane base, 1-ethyl-3-(3dimethylaminopropyl) carbodiimide (EDC), N-hydroxy succinimide (NHS), HRP-labeled goat anti-rabbit IgG, AP-labeled anti-rabbit IgG, and Tween 20 (polyoxyethylenesorbitan monolaurate) were obtained from Sigma (Dublin, Ireland). p-Aminophenyl phosphate ( p-APP) was synthesised in-house with slight variation from the original synthetic route[15] and is now available from Universal Sensors, Inc. The inks used in screen-printing (Electrodag B-0851, 423-SS and 451-SS) were purchased from Acheson (Plymouth, UK). All other chemicals were of analytical grade or better and doubly distilled water was used throughout.

Buffers and Solutions Conjugation buffer for conjugate synthesis consisted of 0.1 M 2-[Nmorpholino] ethanesulphonic acid (MES) þ0.9 M NaCl. Carbonate


1854

Kania et al.

buffer (50 mM l NaHCO3 þ 50 mM Na2CO3, pH 9.6) for coating. Phosphate-buffered saline (PBS, 40 mM Na2HPO4 þ 40 mM NaH2PO4, pH 7.2 including 0.15 M NaCl); PBS washing buffer 4 mM, pH 7.2 including 15 mM NaCl and 0.05% Tween; 3% skim milk powder in PBS (PBS-SMA) as blocking buffer. Substrate for HRP was prepared by mixing two parts of phosphate buffer (140 mM, pH 5.0) containing 3 mmol urea peroxide and one part of 1.2 mM 3,30 ,5,50 -tetramethylbenzidine (TMB) in 8 mM phosphoric acid containing 10% dimethylsulfoxide (DMSO) and 12 mg L1 penicillin G. The latter was used to prevent contamination and facilitate long term storage. Stopping solution for HRP: 1 M H2SO4. Dilutions of DA, a-DA pAbs and HRP-labelled goat anti-rabbit antibody were dissolved in PBS. Zero point five percent BSA in PBS were used for the dilution of biotinylated goat a-rabbit antibody (gt-rb-biotin) and streptavindin-HRP conjugate.

METHODS Domoic Acid Conjugation and Antisera Production Domoic acid was conjugated to keyhole limpet hemocyanin (KLH) according to Smith and Kitts.[16] For KLH conjugation, 1 mg of DA was dissolved in 520 mL conjugation buffer and added to 2 mg of KLH dissolved in 200 mL of distilled water following by adding of 50 mL of an EDC solution (10 mg mL1) under stirring. For OVA/BSA conjugation 40 mg of DA were dissolved in 520 mL conjugation buffer and added to 2 mg of protein (coupling ratio 3:1). All subsequent steps were performed as described for KLH-conjugation. Both preparations were incubated for 2 h at room temperature. The precipitates from both conjugates were removed by centrifugation (800 g, 10 min) and free DA was separated by dialysis against distilled water for 3 days. The coupling of the coating conjugates were validated by MALDI-TOF. Rabbits were immunized with a DA-KLH conjugate. Two hundred (200)mL DA-KLH (1 mg mL1) were mixed with the same volume of Freund’s adjuvant and applied 3 times subcutaneous in 1 week intervals. Two weeks later intramuscularly booster injections were each given, spaced 4 weeks apart, containing the same doses as the first 3 injections. Freund’s adjuvant complete was used for the first injection and for all subsequent injections. One and two weeks after each booster injection, blood was taken from the ear vein and the antisera were checked for specific antibodies by ELISA. Four months after the first immunization, 30 mL blood was taken from the rabbit that showed the highest antibody



1855

titre. The separation of serum from the blood clot proceeded at room temperature, followed by decantation and centrifugation. The immunoglobulin fractions were lyophilized and stored at 20 C.

Optimization of DA ELISA Microwell plates were coated overnight at 4 C with 5 mg mL1 DAOVA dissolved in carbonate buffer (200 mL). Wells treated with carbonate buffer were used as a control. After a washing step (three times with 350 mL PBS washing buffer) the remaining active sites were blocked by addition of 3% PBS-SMA (300 mL) for 1h at room temperature. The blocking solution was discharged, omitting a further washing step. Then 100 mL standards of DA and PBS as control were added to each well together with 100 mL pAb, dissolved 1:1500 in PBS. After incubation time of 1.5 h at room temperature the plates were washed and 200 mL of a goat a-rabbit IgG labeled with biotin was added in a dilution of 1:20,000 dissolved in 0.5% BSA-PBS. After 1 h the plates were washed again and 200 mL streptavidin-HRP conjugate was added (1:50,000; 0.5% BSAPBS). After 30 min, 200 mL of HRP substrate was added after another washing step. The enzyme substrate reaction proceeded for 20–30 min and was stopped prior to reading at 450 nm. The data were normalized by the % B/Bo transformation according to Dudley et al.[17]

Indirect Competitive ELISA for DA with HRP The coating and blocking steps were carried out as described above. For the competitive assay 100 mL of DA standard dilutions (0.1–10,000 mg L1) and PBS as a control were added to the plates together with the antiserum at 1:1500 dilution. After incubation for 1.5 h the plates were washed again three times with washing buffer. The specific pAbs bound to DA-OVA coating conjugate were detected by a goat a-rabbit antibody labeled with HRP. After 1 h incubation plates were washed and enzyme substrate was added. The enzyme reaction was stopped and measured as described previously.

Indirect Competitive ELISA for DA with AP When DA-BSA was used as coating conjugate, 50 mL of a stock solution (10 mg mL1) in carbonate buffer, pH 9.6 was dispensed into


1856

Kania et al.

each well and incubated for 1 h at 37 C. Blocking followed with 1% TrisBSA, 100 mL, again for 1 h at 37 C. Competition was facilitated by using a-DA pAb (6.7 mg mL1) and a dilution range of DACS-1 calibration solution made up in PBS buffer, pH 7.4. 1 h at 37 C. The secondary arabbit IgG-AP (1:2000) followed in 100 mL of 1% Tris-BSA buffer, for 1 h at 37 C. Color was developed by the addition of 100 mL of a 1mg mL1 stock of p-nitrophenyl phosphate substrate solution in carbonate buffer, pH 9.8 added to each well.

Determination of Cross-Reactivity The cross-reactivity of the pAbs was tested in the optimized HRP ELISA as described above. To check the specificity of the pAbs, ELISA was carried out with kainic acid, an algal toxin produced by the red algae Digenea simplex, aspartic acid, glutamic acid, geranic acid and 2-methyl3-butenoic acid. These compounds share common structures with domoic acid (Fig. 1). Therefore standards of DA and the compounds described above (0.1–10,000 ppb) were applied together with the pAbs (1:1500) in the assay. All subsequent steps were performed as in the optimized ELISA.

SCREEN PRINTED ELECTRODES Screen printed electrodes were prepared in advance and stored desiccated at room temperature. Primary protein incubation was facilitated by the addition of 5 mL of stock solution for 1 h at 37 C. Carbonate buffer (50 mM), pH 9.6 was used for primary protein immobilization. Tris buffer (50 mM), 100 mM NaCl, 0.05% Tween 20, pH 7.4 were the constituents of the wash buffer and the blocking solution contained 1% BSA. Blocking was attained by incubating each electrode into a volume of 150 mL placed into preblocked microtitre wells, again for 1 h at 37 C. Competition reactions involved 75 mL of each species under the same time and temperature conditions previously used in the optimized AP ELISA. Signal detection was realized by placing the SPE into a stirred electrochemical batch cell of 5 mL volume containing substrate buffer. The substrate buffer for electrochemical detection with optimal Michaelis-Menten characteristics was found to be 50 mM Tris, 100 mM NaCl, 1 mM MgCl2, 0.1 mM ZnCl2 and adjusted to pH 9.0.[18] The batch cell was connected to a potentiostat and injecting p-APP to a final concentration of 1 mM facilitated signal detection. Current was


Development of Polyclonal Antibodies CH3 H

a)

1857

b)

COOH

CH2 CH2COOH

H3C N H

COOH

N H

COOH

c)

CH2COOH

H3C

d) O

HO

O

OH O

NH2

OH

e)

f) O

O

O

HO

OH

OH

NH2

Figure 1. Chemical structure of (a) domoic acid and closely related compounds; (b) kainic acid; (c) aspartic acid; (d) geranic acid; (e) glutamic acid; and (f ) 2methyl-3-butenoic acid, tested for cross-reactivity using ELISA and HRP as the enzyme label.

monitored at þ300 mV vs. Ag/AgCl. All data points were carried out in duplicate and each value was the average of two readings unless otherwise stated. Washing of SPE occurred after each step.

RESULTS AND DISCUSSION Domoic acid was coupled to OVA and KLH. Rabbits were immunized with DA-KLH in regular time intervals and were bled 7 and 14 days after each booster injection. Sera were harvested and screened for specific antibodies. After 6 bleedings immunization was stopped because the antisera indicated a high concentration of antibodies specific to DA. To check the sensitivity of the pAbs a competitive ELISA was applied. In the nonoptimized assay a detection limit of 5 ppb with an EC50 value of 200 ppb could be obtained in the competitive ELISA described above (Fig. 2). EC50 is the effective concentration of analyte to decrease the


1858

Kania et al.

non-optimised ELISA (HRP) optimised ELISA (HRP) optimised ELISA (AP)

Figure 2. Optimized (f), nonoptimized ELISA (g) for domoic acid using HRP and the indirect competitive ELISA using AP (œ). B/Bo values were calculated by dividing the mean signal of a given set of replicates containing analyte (B) by the mean signal of all control wells containing no analyte (Bo).

response by 50%. A streptavidin biotin amplification system was used for the optimization of the ELISA based on immunoenzymatic techniques developed by Avrameas.[19] With this assay a detection limit of 0.6 ppb DA with a corresponding EC50 of 15 ppb DA was achieved (Fig. 2). These ELISAs using HRP were compared to an AP indirect competitive system. On this occasion the system showed a smaller linear range (0.4–200 ppb) than the optimized HRP system but with similar detection limits (0.2 ppb). Once the linear range had been established and validated, recovery experiments were performed using this AP ELISA system and values returned were 12% (Table 1). To further assess the specificity of these a-DA pAbs, experiments with the optimized HRP EIA format were performed using various inhibitors (kainic acid, aspartic acid, glutamic acid, geranic acid and 2methyl-3-butenoic acid (Fig. 1)). These substances were added together with the pAbs in microwells coated with DA-OVA. The high specificity of the pAbs was clearly demonstrated as all chosen haptens were essentially unrecognized in the concentration range 0–10,000 ppb (Fig. 3). At this points the a-DA pAbs were used to develop a disposable immunosensor. A typical indirect competitive curve for DA



1859

Table 1. Recovery experiments performed on both ELISA and SPE for the indirect competitive immunoassay using AP as the enzyme label. DA1 denotes the recovery value for the unknown sample #1 measured using the calibration curve. Format ELISA (%) SENSOR (%)

DA1

DA2

DA3

DA4

99 91

112 121

103 125

106 —

Figure 3. Cross-reactivities of a-DA pAbs to similarly structured analogues. Each column represents the competition between the competitor, at 2 concentrations (0 and 10 mg L1).

immunosensor can be seen in Fig. 4. Linear range analysis for the immunosensor assay showed a working range between 0.2 and 160 ppb DA. Errors associated with each standard were generally all below 7% (n ¼ 2). Repeated analysis of the linear range showed that the exact range did fall within the values quoted above and the associated regression and errors (0.995 0.006; n ¼ 4) were seen. This system was comparable with the ELISA developed. The DA immunosensor returned an EC50 value of 15 ppb and an associated detection limit of 0.1 ppb. For protection against ASP, it is very important to closely monitor DA. At present high performance liquid chromatography (HPLC) is most widely used for monitoring DA.[20,21] However the HPLC methods are unsuitable for rapidly screening large numbers of samples, since they


1860

Kania et al.

100 % B/Bo 80

60

40

20

0 0

0 .1

10

1000

domoic acid [ µg/l]

Figure 4. Typical competitive immunoassay for DA using SPE with amperometric detection at þ300 mV vs. Ag/AgCl. B/Bo values as described in Fig. 2 legend.

require time-consuming cleanup of sample prior analysis. In the recent years several polyclonal[12–14,16,22] and monoclonal antibodies[23,24] have been produced against DA. These ELISA procedures could be carried out with little pretreatment and results obtained within 2 h. The present study was targeted at the development of pAbs against DA for their use in disposable biosensors. The necessity is clear for such devices considering their use on-site for simplistic analysis, low-cost instrumentation, fast response times and minimum sample pretreatment. To check the sensitivity and selectivity of the obtained pAbs, ELISA were developed using both the enzymes HRP and AP. In the first approach this pAb allowed a broad working range of 5–3800 ppb. In an attempt to improve the sensitivity of the DA ELISA reducing the concentration of the pAbs was tried. However, with low antibody concentrations only low signals can be obtained in the ELISA. To increase the signal an ABC amplification system was applied.[19] By using this assay format a working range of 0.8–300 ppb could be obtained. To simplify the system further the AP ELISA for DA highlighted a similar range (0.4–200 ppb). These latter systems were sensitive enough to measure DA concentrations that cause the onset of illness.



1861

The disadvantage of measuring real samples is that they are often contaminated with many more than one or two toxins. Sato et al.[25] described that analogous compounds, such as kainic acid, are sometimes found in algae blooms. However, no cross-reactivity towards kainic acid, aspartic acid, glutamic acid, geranic acid and 2-methyl-3-butenoic acid were detected with our a-DA pAbs. After examining the sensitivity and selectivity of these pAbs, they were used in electrochemical immunosensors. Once the system had been tested, numerous repetitive assays were performed with quite good errors (