Talanta 53 (2000) 379–389

Amperometric biosensors based on nafion coatedscreen-printed electrodes for the determination of

cholinesterase inhibitors

E.V. Gogol a,b, G.A. Evtugyn a,*, J.-L. Marty b, H.C. Budnikov a, V.G. Winter a

a Ecology Faculty, Kazan State Uni6ersity, Kremle6skaya 18, 420008 Kazan, Russiab Centre de Phytopharmacie URA CNRS 461, Uni6ersite de Perpignan, Perpignan, France

Received 10 March 2000; received in revised form 28 June 2000; accepted 5 July 2000

Abstract

Screen-printed electrodes coated with the nafion layer have been investigated for cholinesterase biosensor design.The butyrylcholinesterase (ChE) from horse serum was immobilised onto the nafion layer by cross-linking withglutaraldehyde vapours. The biosensors obtained showed better long-term stability and lower working potential incomparison to those obtained with no nafion coating. The sensitivity of a biosensor toward organophosphatepesticides is not affected by the nafion coating. The detection limits were found to be 3.5×10−7 M for trichlorfonand 1.5×10−7 M for coumaphos. © 2000 Elsevier Science B.V. All rights reserved.

Keywords: Biosensor; Cholinesterase; Pesticide determination; Screen-printed

www.elsevier.com/locate/talanta

1. Introduction

Organophosphorus and carbamic pesticides arewidely used in agriculture as insecticides due totheir high efficiency and moderate toxicity forwarm-blooded living organisms. Nevertheless, acertain amount of pesticides when transferred intothe environment can cause various toxic effects onhuman beings within several weeks. Paraoxonresidues, considered to be toxic for agricultureworkers, were found on the citrus peel 28 days

after the pesticide application [1]. In model exper-iments, trace amounts of organophosphates weredetermined in surface runoff 19 days after thesimulated rainfall applied [2]. Standard chro-matography analysis commonly used for pesticidedetection is very reliable but there is a need of fastand inexpensive testing devices, especially for fieldapplication.

The toxic effect of organophosphorus and car-bamic pesticides is mainly related to their abilityto irreversibly bind acetylcholinesterase and henceto prevent the transmission of nerve impulses [3].The anticholinesterase effect, being very specific,dictates the limited threshold values of pesticidecontent in surface, ground and drinking waters,

* Corresponding author. Tel.: +7-8432-315569; fax: +7-8432-380412.

E-mail address: gennady.evtugyn@ksu.ru (G.A. Evtugyn).

0039-9140/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved.

PII: S0039-9140(00)00507-5

E.V. Gogol et al. / Talanta 53 (2000) 379–389380

food and soil on the ppm–ppb level [4]. For thesame reason, cholinesterase isolated from varioussources is often used for the detection and semi-quantitative determination of organophosphoruspesticides based on their inhibitory effect. Part ofthe recent results related to cholinesterase biosen-sor development is summarised in reviews [5–7].

For the electrochemical detection ofcholinesterase activity various potentiometric [8–11], amperometric [12–14] and conductometric[15,16] sensors have been proposed. A two-enzymesystem based on acetylcholinesterase and cholineoxidase either immobilised or present in solutionhas been developed for the determination of acetyl-choline and cholinesterase inhibitors [17–19]. Ofother enzymatic methods for organophosphatedetection, phosphotriesterase [20,21], acid [22] andalkaline phosphatases [23,24], tyrosinase [25] werealso used for biosensor development. Althoughthese biosensors showed sufficient sensitivity to-ward pesticides, the detection limits were foundhigher in comparison to the results obtained withthe cholinesterase biosensors.

As applied to inhibitor determination, the prac-tical application of immobilised enzymes has asignificant limitation. The inhibition results in thedecay of the enzyme activity so that the number ofconsecutive measurements with the same biosensoris limited. This makes level the main advantage ofbiosensors, i.e. the repeated use of immobilisedenzymes. The necessity of enzyme reloading orreactivation after each measurement complicatesthe operation of an inhibitor biosensor, especiallyin field conditions, and impairs the metrologicalcharacteristics of inhibitor determination.

For this reason, during the past decade specialattention was paid to the investigations of dispos-able cholinesterase biosensors. They have beendeveloped on the basis of the sensors manufacturedwith the screen-printing techniques (planar sensorson plastic support [18,24,26–29]) or semiconductormicromachining techniques (pH-sensitive field ef-fect transistors, FETs [15,30–33]) and coveredwith an ultra thin film of an enzyme directlybonded to the sensor surface. The microfabricationtechniques and small amounts of the enzyme usedmake the measurement cheap enough for a singleuse of the biosensor with no reactivation of the

inhibited cholinesterase.Carbon-based composites were found most suit-

able for the development of working electrodesbecause of the variety of characteristics attained bythe incorporation of small amounts of appropriatemodifiers. Thus, ruthenium [18], cobalt phtalocya-nine [26,28] and 7,7,8,8-tetracyanoquinodimethane[29] were implemented in the carbon paste todecrease the working potential and to avoid theinterfering influence of electroactive impurities.

Nevertheless, the long-term stability of biosen-sor response as well as the reproducibility ofoperational characteristics in the series producedremains a weak point of disposable biosensors. Inthis study, the influence of the additional nafionlayer on the operational and analytical characteris-tics of carbon screen-printed cholinesterase biosen-sors has been investigated.

2. Experimental

2.1. Reagents

Butyrylcholinesterase (ChE) from horse serum(EC 3.1.1.8), specific activity 500 U mg−1 ofprotein, and S-butyrylthiocholine iodide were pur-chased from Sigma Chemical Co. (St. Louis,USA). Nafion (perfluorinated ion-exchange resin,5% w/v solution in lower alcohols/water) wasobtained from Aldrich (Steinheim, Germany). Theenzyme was immobilised by cross-linking withglutaraldehyde (Sigma).

Organophosphorus pesticides trichlorfon([2,2,2-trichloro-1-hydroxyethyl]-phosphonic aciddimethyl ester) and coumaphos (phosphorothionicacid O-(3-chloro-4-methyl-2-oxo-2H-1-benzopy-ran-7-yl)O,O-diethyl ester) were purchased fromRiedel-de-Haen (Seelze, Germany). Chemicalstructures of the pesticides used are presented in 1.

(1)

E.V. Gogol et al. / Talanta 53 (2000) 379–389 381

All the other reagents used were of analyticalgrade (Reakhim, Russia, Fluka, Neu-Ulm,Switzerland, and Prolabo, Fontenay S. Bois,France).

2.2. Screen-printed electrode manufacture

The electrodes for biosensor design were manu-factured with screen-printing techniques on aDEK 248 printer (DEK, Weimouth, UK). Threetypes of inkblots, i.e. silver inkblot (ElektrodagPF 410), graphite inkblot (Elektrodag 423 SS),silver/silver chloride inkblot (Elektrodag 6037 SS)were supplied by Acheson (Scheemda, TheNetherlands). Two types of graphite, i.e. TimrexT 15 graphite provided by Timcal (Bodio,Switzerland) and graphite supplied by Le CarboneLorraine (Gennevilliers, France), were used forcoating the working area of electrodes.

The electrode sets were prepared on the PVCpad with the screen-printing techniques. Eachelectrode set contains a pair of working and refer-ence electrodes (Fig. 1). In one manufacture cycle,32 sets were prepared simultaneously on a singleplastic plate. The manufacture cycle involves theconsecutive deposition of five layers using differ-ent moulds:

silver conducting connections;graphite layer covering the ends of the silverconnectors to isolate them from direct contactwith an analyte solution;Ag/AgCl layer on the area of the referenceelectrode;insulation film of Astral commercial paint;graphite paste on the area of a workingelectrode.The graphite paste was prepared by mixing 23

ml of 3% w/v hydroxyethyl cellulose (mediumviscosity, Fluka) in deionised water with 3.5 g ofgraphite powder. The working surface of the elec-trode is rectangular (8.5×2 mm). The overall sizeof the electrode set is 60×10 mm.

2.3. Biosensor de6elopment

Prior to enzyme immobilisation, the workingarea of an electrode was covered with the nafioncoating. For this purpose, 2 ml of the 0.5 or 0.05%ethanol solution of nafion were carefully spreadon the active surface and dried during 24 h atroom temperature. After that, 5 ml of the ChEsolution corresponding to the specific activity 15U cm−2 were placed on the electrode surface,dried at room temperature for 20 min and thentreated with glutaraldehyde vapours as describedin [11]. For this purpose, the electrodes were fixed1 cm above the surface of a 3% glutaraldehyde inan isolated glass vessel and left for 2 min underthe vacuum produced by a water-jet pump. Next,the biosensor was washed with deionised waterand the phosphate buffer 2×10−3 M, pH 7.8.The biosensors prepared were stored at 4°C underdry conditions for several weeks.

2.4. Procedures

2.4.1. The ChE acti6ityThe activity of free enzyme was photometrically

evaluated by the Ellman method [32] in the 2×10−3 M phosphate buffer, pH 7.8, at 37°C. Thekinetic parameters of immobilised enzyme weredetermined in Lineweaver–Burk plots assumingthe oxidation current as a measure of the rate ofenzymatic reaction.

Fig. 1. Four electrode sets on the PVC pad. Each set containsone working electrode (right one, black) and one referenceelectrode (left one, grey). Middle grey area refers to thecapsulation paint layer.

E.V. Gogol et al. / Talanta 53 (2000) 379–389382

Fig. 2. I–E curve of 5×10−4 M of S-butyrylthiocholineiodide, 2×10−3 M phosphate buffer+1×10−1 M Na2SO4,pH 7.8, 37°C, (1) screen-printed electrode with no coating; (2)nafion modified electrode; (3) I–E curve of buffer solution,electrode with no coating.

tive measurements performed with the samebiosensor during a day did not exceed 3%.

2.4.3. Inhibition measurementsImmediately before inhibition measurement,

coumaphos was electrochemically oxidised to itsoxygen analogue in accordance with Evtugyn et al.[34]. The 5-min electrolysis of an appropriateamount of the pesticide in 10% acetonitrile solutionwas performed in the presence of 1×10−1 M NaClon Pt electrodes at 1 mA cm−2. Next, the excessof chlorine generated at the Pt electrode wasremoved by addition of 100 ml of 2% v/v formic acidfollowed by the pH correction with 10% NaOH.

Trichlorfon was allowed to stay for 4 h forspontaneous conversion to DDVP [O,O-dimethyl-O-(2,2-dichlorvinyl)-phosphate] in accordancewith 2.

(2)

The biosensor was incubated in the pesticidesolution for 10 min, then it was washed with theworking solution and the current of thiocholineoxidation was measured in standard conditions asdescribed earlier. The degree of inhibition, I(%),was calculated as a relative decrease in the biosen-sor response after the contact of the biosensor withan inhibitor solution Eq. (3).

I(%)=�(I0−It)

It

n×100 (3)

with I0 and It as the current of thiocholine oxidationbefore and after the incubation procedure, respec-tively. After the measurement, the inhibited ChEwas reactivated by treatment with 2-pyridineal-doxime methiodide (2-PAM) in a 0.1% solution for10 min. After that, the biosensor was washed withdistilled water and then in the working buffersolution. Each biosensor allowed at least ten mea-surements, if the degree of inhibition did not exceed40%.

For the measurement of fluoride inhibition, theNaF solution was first mixed with the substratesolution and then an appropriate volume of the

2.4.2. Biosensor responseThe oxidation current was measured at +400

mV versus Ag/AgCl in a three-electrode cell witha Pt wire as an auxiliary electrode at 37°C, usingthe scanning potentiostat model 362 (EG&GPrinceton Applied Research, USA) connected toLY 1400 x–y plotter (LINSEIS, West Germany).All the solutions were stirred with the magneticagitator Bioblock Scientific, AM 3000D.

First, the range of working potentials has beendetermined in the scanning regime at 5 mV s−1.Typical voltammograms of thiocholine oxidationobtained with non-modified and nafion-coatedbiosensors are presented in Fig. 2.

The following measurements of biosensor re-sponse were performed at fixed potential as follows.The biosensor and Pt auxiliary electrode wereimmersed in a 2×10−3 M phosphate buffer solu-tion, pH 7.8, containing 1×10−1 M sodium sul-phate, at 37°C. After the biosensor reached a stablebackground current, an appropriate volume ofbutyrylthiocholine iodide solution was injected tothe final concentration 1×10−3 M and the oxida-tion current was recorded at the fixed potential of+200 to +400 mV versus Ag/AgCl 5 min after thesubstrate injection. The maximum shift of thecurrent was measured as the biosensor response.The S.D. of the response toward 1×10−3 MS-butyrylthiocholine calculated from six consecu-

E.V. Gogol et al. / Talanta 53 (2000) 379–389 383

mixture was injected into the working solution toobtain the substrate concentration of 5×10−4 M.After response measurement, the initial enzymeactivity was restored by washing out the biosensorin the working buffer solution for 15 min.

3. Results and discussion

3.1. The conditions of response measurements

The amperometric response of the ChE biosen-sor on substrate injection is based on the anodicoxidation of the thiocholine formed in the enzy-matic hydrolysis of S-butyrylthiocholine iodide(Eq. (4)).

(CH3)3N+CH2CH2SC(O)C3H7+H2O �

(CH3)3N+CH2CH2SH+C3H7COOH (4)

The mechanism of electrode reaction involves thetransfer of one electron from thiocholine followedby its dimerisation to disulphide [35,36] Eq. (5).

(CH3)3N+CH2CH2SH−e−−H+�

(CH3)3N+CH2CH2S��dimer (5)

This principal reaction scheme (Eqs. (4) and (5))can be followed by a strong sorption of the reactanton the electrode. For metal electrodes, this resultsin the increase of the oxidation potential and inelectrode passivation. To avoid the sorption ofsulphur containing species, electrochemical clean-ing at higher potentials up to +800 mV versusAg/AgCl or mechanical polishing of an electrodesurface were recommended [35–37]. Besides, thesimultaneous oxidation of both thiocholine andiodide and the participation of the iodide/iodinepair in the mediation of electron transfer fromthiocholine were observed. All of these reactionscomplicate the response measurement and increasethe influence of operation conditions on the biosen-sor response.

3.1.1. The influence of nafion coatingPreviously we found that the use of epoxy-car-

bon composites as electrode materials makes itpossible to separate the electrode oxidation ofthiocholine and iodide so that the current peak at

+580 mV corresponds only to the thiocholineoxidation [14]. The further decrease of the workingpotential can be reached if the working surface ofa carbon electrode is covered with a thin layer ofnafion.

Nafion deposition on the working area of asensor followed by ChE immobilisation results inthe decrease of the potential of thiocholine oxida-tion to +400 mV compared with +550 to +650mV for the same electrode with no nafion coating.Although the appropriate current wave on voltam-mograms is flattened, the oxidation current valuerecorded at a fixed potential is proportional to theS-butyrylthiocholine concentration in the range2.5×10−5–8×10−4 M and does not depend onthe nature of the counter ion of the substrate, i.e.iodide or chloride.

The operational characteristics of the biosensordepend on the amount of the nafion deposited onthe working area of the electrode. The increase ofeither the nafion solution volume (Fig. 3a) or itsconcentration (Fig. 3b) results in the decrease of thebiosensor response toward S-butyrylthiocholine.Probably, this effect is determined by the degree ofthe surface covering as well as by the diffusionlimitation caused by the additional polymer layer.The maximum amounts of nafion which do notalter the thiocholine oxidation current correspondto the application of 2 ml of 0.05% solution per oneelectrode.

The decrease of both the working potential andthe thiocholine oxidation current observed fornafion modified sensors can result from the jointinfluence of the ion exchange and diffusion limita-tion. Due to its own negative charge, nafion am-plifies the deceleration of the transfer of iodideanions to the electrode surface. Meanwhile, for thepositively charged thiocholine, the diffusional lim-itation of the transfer can to some extent becompensated by the ion exchange features of thepolymer used. Although thiocholine is ionised inthe bulk buffer solution (pH 7.8), its formation inthe membrane body is followed by the release ofequal amounts of butyric acids in accordance withEq. (4). Previously for the ChE membranes of thesame content it was shown, that the addition of2×10−3 butyrylcholine results in the pH shift inthe membrane body to 1.5 pH units [38]. Hence,

E.V. Gogol et al. / Talanta 53 (2000) 379–389384

most part of the thiocholine at the inner part of themembrane next to the nafion layer is formed in thenon-ionised cationic form Eq. (6).

(CH3)3N+CH2CH2SHmembrane body: pHB7

X

(CH3)3N+CH2CH2S−+H+

buffer solution: pH 7.8(6)

A similar effect of the additional nafion mem-branes has previously been reported for FETs withan immobilised urease examined for the detectionof heavy metal cations. The deposition of nafionresulted in a 5-fold increase of the inhibitory effectobserved for the 1×10−5 M CuCl2 solution [29].The influence of nafion is also related to its promo-tion of the Cu(II) cation transfer to the enzymeimmobilised on the FET surface.

Fig. 4. The stability of the biosensor response toward thesubstrate (a) nafion coated electrode. (1) E+400 mV, S-bu-tyrylcthiocholione iodide 1×10−3 M; (2) E+150 mV, S-bu-tyrylthiocholine iodide 2×10−4 M. The biosensors stored inbuffer solution at room temperature, the average values of sixconsecutive measurements; (b) non-modified electrode, E+400 mV, S-butyrylthiocholine 2×10−4 M. Each measurementwas made with a novel electrode set manufactured on the samePVC pad. The average from three measurements was deduced.

Fig. 3. The influence of nafion quantities on the biosensorresponse. (a) The dependence of the response on the volume ofthe 0.05% nafion solution used for coating. S-Butyrylthio-choline iodide 1×10−4 M, E+400 mV vs. Ag/AgCl. (b) Thecalibration curves of S-butyrylthicholine. Electrodes aretreated with 2 ml nafion solution. E+400 mV vs. Ag/AgCl.

All the biosensors show good reproducibilityboth for the background current and the responseto the low concentrations of S-butyrylthiocholine.The variation coefficient was found to be 6% forsix biosensors manufactured on the same PVCplate and the S-butyrylthiocholine concentrationvarying from 5×10−4 to 1×10−3 M. The varia-tion of the biosensor response with storage time ispresented in Fig. 4. The lower the working poten-tial and substrate concentration, the less the decayof enzymatic activity. Probably, this is related tothe accumulation of the passivating intermediatesof thiol oxidation. Thus, when the substrate con-centration exceeds 1×10−3 M the response de-creases by about 10–15% a day.

The same biosensors with no nafion modifica-tion show a fast decay of the enzyme activity

E.V. Gogol et al. / Talanta 53 (2000) 379–389 385

irrespective of whether they are stored in dryconditions or in buffer solution (Fig. 4b). Per-haps, nafion provides better adhesion of the enzy-matic layer to the electrode surface due to theelectrostatic interaction with the protein molecule.It can also diminish the washout of carbon parti-cles of the electrode material from the hydrox-yethyl cellulose matrix. The stabilising effect ofnafion on the response and its long-term stabilitywas previously found for glucose, lysine biosen-sors [39,40] and for the carbon-fiber choline sen-sor [41]. Besides nafion, other polymer additivesprovide for the better operational stability of am-perometric sensors as shown for lysine and hypox-antine sensors with the enzymes incorporatedtogether with teflon into the carbon paste [42,43].

ChE immobilisation does not significantlychange the kinetic parameters of the enzymaticreaction. Thus, Michaelis constant values calcu-lated from the substrate calibration curves pre-sented in double inverse coordinates I−1–CS

−1

were found to be 2.0×10−4 M for the freeenzyme and 1.7×10−4 M for the immobilisedChE. It should be mentioned that the latter couldbe overestimated due to the limitation of thereverse thiocholine diffusion from the membraneto the bulk solution. This can result in its extra-accumulation in the membrane body in compari-son to the amounts of the substrate hydrolysed inthe enzymatic reaction Eq. (4). The above corre-sponds to the fact that the increase in the amountof nafion is followed by an insignificant decreaseof the Michaelis constant, probably due to thepromotion of the thiocholine access to the elec-trode surface coated with an ion exchange layer.

3.1.2. The influence of sil6er connectorsIn several cases a sharp small current peak with

the maximum at +200 mV appeared in the I–Ecurves during the biosensor exploitation on therising crest of the wave of thiocholine oxidation.Probably, this peak is related to the oxidation ofthe silver conductor due to the break of theintegrity of the carbon paste layer. In the presenceof the substrate the peak current grows and shiftsto the lower potential due to the formation ofinsoluble salt promoting the electron transfer Eq.(7).

Ag+ (CH3)3N+CH2CH2SH−e−

�AgSCH2CH2N+(CH3)3+H+ (7)

The electrode reaction Eq. (7) does not interferewith recording the direct electron transfer fromthiocholine Eq. (5) occurring at higher potentials(E= +350 to +400 mV). To avoid the possibleerrors of current measurement due to this sideeffect the following inhibition measurements wereperformed at the poised potential of +400 mV.

3.2. Inhibitor determination

All the inhibition measurements were per-formed with the concentration of the substrate5×10−4 M. This corresponds to the saturationof the enzymatic layer with the substrate andhence to the maximum sensitivity of the biosensortoward inhibitors. For the lower substrate concen-tration, the actual decay of the enzyme activity inthe inhibition is partly compensated for by itsinvolvement in the reaction of free active sites. Asa result, the experimental decrease of the biosen-sor response will be lower than the actual decreaseof the enzyme activity due to the inhibition. Thisphenomenon which Traven called figuratively the‘Zulu effect’ [44] is typical for immobilised en-zymes irrespective of the mechanism of inhibitionor the system used for the detection of enzymeactivity.

For irreversible inhibitors, the immobilised en-zyme contacts the inhibitor without any substrateso that the latter cannot compete for the bindingsite with an inhibitor. For fluoride determination,the sensitivity toward the inhibitor increases withthe decrease of the substrate concentration. Forthese experiments the working concentration ofS-butyrylthiocholine was chosen to be 5×10−4

M for reasons of measurement accuracy.

3.2.1. TrichlorfonThe inhibitory effect of the freshly prepared

trichlorfon solution left at room temperature in-creases within the first 3–4 h. The HPLC chro-matography of pesticide solution showed at leastfour products of trichlorfon conversion, one ofwhich was identified as DDVP (see 2). The maxi-mum of DDVP accumulation coincides with the

E.V. Gogol et al. / Talanta 53 (2000) 379–389386

greatest inhibitory effect of the solution on theimmobilised ChE. The decay of the biosensorresponse linearly depends on the incubation timein the range 5–20 min. This corresponds to theirreversible mechanism of inhibition. The resultsof the trichlorfon determination in optimal condi-tions are presented in Table 1. The sensitivity ofbiosensors toward trichlorfon is in good agree-ment with the results reported for various ChEbiosensors. Thus, the detection limit of 1×10−7

M was reported for FETs and conductometricmicrocells with the ChE immobilised directly onthe electrode surface [15]. A higher detection limit(8×10−7 M of DDVP) was reported for biosens-ing device consisting of a polyurethane pad withimmobilised ChE and Pt net electrodes [37]. In thelatter case the incubation time was chosen to be 3min in comparison to 10 min in this study. Inthese cases, the conditions of mass transfer con-tribute to the kinetic regime of the enzyme-in-hibitor interaction. For other cases, if theenzymatic layer is not saturated with the sub-strate, the results of inhibitor detection could befar from those in optimal conditions. Thus, forthe potentiometric biosensor based on the plasti-cised PVC membrane the detection limit oftrichlorfon was only about 1×10−6 M [45].

The treatment of the biosensor after inhibitionmeasurement with the 2-PAM solution makes itpossible to restore the initial response of biosen-sor. The repeated application of the biosensor forthe determination of the same trichlorfon concen-

tration gives a similar value of inhibition thoughthe subsequent reactivation is not as efficient thanthe first one. After 10 min of incubation of thebiosensor in the 8×10−7 M trichlorfon solutionthe efficiency of the first reactivation is about 87%whereas after repeated incubation it is only 75%.The biosensor gives reproducible values of thedegree of inhibition for at least ten consecutivecycles of inhibition–reactivation even though theinitial response of the biosensor decreases duringthis series by 30–50%.

3.2.2. CoumaphosContrary to trichlorfon, coumaphos does not

affect cholinesterase activity. In insects, this pesti-cide is pre-oxidised in vivo with mixed functionoxidases to its oxygen analogue, which is realirreversible inhibitor of the ChE. In this study theelectrochemical pre-treatment of the stock solu-tion proposed earlier [14] was used for the samepurposes. In comparison to the other procedures,i.e. application of bromine water or nitric acid,electrolysis makes it possible to avoid the dilutionof the sample tested. Because of the low solubilityof coumaphos in water, all the measurementswere performed in 10% acetonitrile. Neither theresponse of the biosensor nor the sensitivity ofinhibitor determination was affected by the addi-tion of such amounts of organic solvent.

The oxidised pesticide solution exhibits stronglyirreversible inhibitory effect on immobilised ChEin the range of concentrations 1.5×10−7–5.5×10−6 M. This corresponds well with the analyticalcharacteristics of potentiometric biosensors basedon Sb electrode and ChE immobilised with thesame technique (detection limit 9×10−7 M) [38].The immobilisation of the ChE on cellulose ni-trate membrane increases the sensitivity ofcoumaphos determination by 15–30% due to thepre-concentration of hydrophobic pesticide on theenzyme carrier in the incubation stage [46].

The reactivation of inhibited ChE makes itpossible to restore up to 90–100% of enzymeactivity in the whole range of pesticide concentra-tions examined. The reactivation decreases boththe sensitivity of immobilised ChE toward pesti-cide and the biosensor response toward the sub-strate by 10–15% in repeated measurements.

Table 1Coefficients of calibration curves I, %=a+b×log(CI, M),detection limits and range of pesticide concentration to bedetermined with the screen-printed cholinesterase biosensor

CoumaphosTrichlorfon

Slope (b) 219926 4794 (3797)(200930)14109170 326924Intercept (a)(12009300) (400950)0.9785 0.9834Correlation

coefficient3.5×10−7Detection limit (M) 1.5×10−7

(5×10−7)(9×10−7)4.0×10−7– 2.0×10−7–Concentration range

(M) 8×10−7 5.5×10−6

E.V. Gogol et al. / Talanta 53 (2000) 379–389 387

Fig. 5. Inhibitory effect of sodium fluoride. S-butyrylthio-choline iodide, 5×10−4 M.

facture pool, the non-modified biosensors arehardly convenient for practical applications.

3.2.4. Fluoride detectionThe reversible inhibition of sodium fluoride has

been investigated in the measurement conditionsdifferent from those used for pesticides determina-tion. A 1:1 mixture (v/v) of the substrate andfluoride solutions was injected into the workingcell so that both the inhibitor and substrate simul-taneously interact with immobilised enzyme (incu-bation time equal 0, one stage measurement). Forthese working conditions, the biosensor preparedwithout additional nafion coating showed the re-versible decay of the response in the range offluoride concentration 1×10−5–1×10−2 M,with linear calibration curve in the plots I (%) orIi–Ci. The appropriate line passes through theorigin of coordinated and the inhibition does notdepend on the duration of the contact of biosen-sor with fluoride solution. This proves the re-versible mechanism of inhibition.

The nafion deposition results in significant de-cay of the sensitivity toward fluoride ions, espe-cially at high concentrations of the inhibitor (Fig.5). The detection limits increases from 1×10−5

to 1.7×10−4 M. In the range of concentrations2×10−4–2×10−3 M the inhibition can be pre-sented as linear function of log CF in accordancewith Eq. (7).

I,%= (97910)+ (2394)× log(CF, M)

n=7, r=0.9742 (8)

Possibly, the decrease of the sensitivity offluoride determination is due to the electrostatichindrance to the access of negatively chargedfluoride to the enzyme layer bordering on nafion.In spite of this, the fluoride inhibition effect isvery stable and after 1 week of storage the result-ing decay of the response does not differ fromthat obtained with a freshly prepared biosensor.

4. Conclusions

Previously the application of additional coatinglayers in the biosensor assembly has been dis-cussed mainly from the viewpoint of substrate

3.2.3. Nafion influenceThe application of screen-printed biosensors

without nafion coatings for pesticide detectionleads problems related to the reproducibility ofthe response. The response of the biosensor to-ward the substrate decreases by 10–15% witheach repeated measurement. The pesticide detec-tion assumes two consecutive measurements sepa-rated by incubation stage. This complicates theestimation of actual inhibitory effect on the back-ground of the drop of the activity of immobilisedenzyme. For this reason, each inhibition measure-ment was performed with a new biosensor. Incomparison, the whole calibration curve of thepesticide (six concentrations of the inhibitor) canbe obtained with a single nafion-modifiedbiosensor.

The estimation of analytical characteristics ofnon-modified ChE biosensors toward trichlorfonand coumaphos is presented in the Table 1 (theslope and intercept values in brackets). As shown,the sensitivity of the response toward both pesti-cides does not differ from that established fornafion-coated biosensors. The appropriate detec-tion limits are about three times higher. These aredetermined from the minimal detectable value ofinhibition. The last was determined from the dou-ble error of the response measurement equal to25% against 10% for modified biosensors. For thisreason, the mentioned above values of the slopeand intercept of pesticide calibration curves areconsidered to be semi-quantitative. Taken intoaccount the instability of the response and thevariation of biosensor characteristics in the manu-

E.V. Gogol et al. / Talanta 53 (2000) 379–389388

determination. These membranes were placedonto the enzymatic layer to diminish its relativesaturation with the substrate and hence to en-hance the linear range of the substrate concentra-tions to be determined. When applied to inhibitordetermination, this approach does not show anyremarkable advantages because the maximumsensitivity toward an inhibitor requires the satura-tion of the membrane with the substrate (kineticregime of response formation). The only excep-tion found is when the inhibitor is pre-concen-trated on the enzyme support due to the ionexchange [15] or non-specific sorption [7,10,38].

In this study, a thin nafion layer was placedin-between the working surface of the sensor andthe enzymatic layer. This biosensor assembly canbe considered as a compromise between the re-quirements for the operational and analyticalcharacteristics of the biosensor. On the one hand,the protective nafion layer decreases the responseof a biosensor. On the other hand, the modifica-tion provides for a better stability of the carbon-paste composition and better stability of the ChEimmobilised on this ultra-thin conductive layer.After electrode modification, the immobilisedChE remains accessible for the potential in-hibitors present in the sample tested. Hence, thenafion coating does not change the sensitivity ofthe biosensor response toward organophosphatepesticides. In comparison to the mediated biosens-ing systems [8,12,28], see also reviews [5,6],nafion-coated biosensors seem more appropriatefor the testing of the samples containing potentialmediators, i.e. dyes, heavy metal ions, phenoliccompounds etc.

Acknowledgements

Financial support of French–Russian Pro-gramme ‘Biotechnology for Agriculture’ is grate-fully acknowledged.

References

[1] R. Basanta, A. Nunez, E. Lopez, M. Fernandez, F.Diaz-Fierros, Int. J. Environ. Stud. 48 (1995) 211.

[2] R.C. Spear, W.J. Poppendorf, W.F. Spencer, T.N. Milby,J. Occup. Med. 19 (1977) 411.

[3] J.R. Corbett, The Biochemical Mode of Action of Pesti-cides, Academic Press, London, 1974.

[4] Index Phytosanitaire, ACTA — association de coordina-tion technique Agricole, Paris, 1993.

[5] P. Skladal, Food Technol. Biotechnol. 34 (1996) 43.[6] J.-L. Marty, D. Garcia, R. Rouillon, Trends Anal. Chem.

14 (1995) 329.[7] G.A. Evtugyn, H.C. Budnikov, E.B. Nikolskaya, Talanta

46 (1998) 465.[8] J. Diehl-Faxon, A.L. Ghindilis, P. Atanasov, Wilkins,

Sensors Actuators B 36 (1996) 448.[9] Y.A. Cho, H.S. Lee, G.S. Cha, Y.T. Lee, Biosens.

Bioelectron. 14 (1999) 387.[10] H.C. Budnikov, G.A. Evtugyn, NATO ASI series, in:

D.P. Nikolelis, U.J. Krull, J. Wang, M. Mascini (Eds.),Biosensors for Direct Monitoring of Environmental Pol-lutants in Field, vol. 38, Kluwer Academic Publishers,Dordrecht, The Netherlands, 1997, p. 239.

[11] R.E. Gyurcsanyi, Z. Vagfoldi, K. Toth, G. Nagy, Electro-analysis 11 (1999) 712.

[12] M. Trojanowicz, M.L. Hitchman, Trends Anal. Chem. 15(1996) 38.

[13] M. Stoycheva, Electroanalysis 7 (1995) 660.[14] G.A. Evtugyn, A.N. Ivanov, E.V. Gogol, J.L. Marty,

H.C. Budnikov, Anal. Chim. Acta 385 (1999) 13.[15] N. Jaffrezic-Renault, C. Martelet, P. Clechet, A.M.N.

Hendji, A.A. Shul’ga, S.V. Dzyadevitch, L.I. Nechiporuk,A.P. Soldatkin, Sens. Mater. 8 (1996) 161.

[16] T.C. Rodrigues, M. Tubino, O.E.S. Godinho, G.D. Neto,Anal. Sci. 13 (1997) 423.

[17] C. Cremisini, S. Di Sario, J. Mela, R. Pilloton, G.Palleschi, Anal. Chim. Acta 311 (1995) 273.

[18] I. Palchetti, A. Cagnini, M. Del Carlo, C. Coppi, M.Mascini, A.P.F. Turner, Anal. Chim. Acta 337 (1997)315.

[19] L. Campanella, S. De Luca, M.P. Sammartino, M.Tomassetti, Anal. Chim. Acta 385 (1999) 59.

[20] A. Mulchandani, P. Mulchandani, S. Chauhan, I.Kaneva, W. Chen, Electroanalysis 10 (1998) 733.

[21] A. Mulchandani, S. Pan, W. Chen, Biotechnol. Prog. 15(1999) 130.

[22] F. Mazzei, F. Botre, C. Botre, Anal. Chim. Acta 336(1996) 67.

[23] M.S. Ayyagari, S. Kamtekar, R. Pande, K.A. Marx, J.Kumar, S.K. Tripathy, J. Akkara, D.L. Kaplan, Biotech-nol. Prog. 11 (1995) 699.

[24] Y.S. Su, A. Cagnini, M. Mascini, Chem. Amalityczna 40(1995) 579.

[25] W.R. Everett, G.A. Rechnitz, Anal. Chem. 70 (1998) 807.[26] P. Skladal, M. Fiala, J. Krejci, Int. J. Environ. Anal.

Chem. 65 (1996) 139.[27] R. Koncki, M. Mascini, NATO ASI series, in: D.P.

Nikolelis, U.J. Krull, J. Wang, M. Mascini (Eds.), Biosen-sors for Direct Monitoring of Environmental Pollutantsin Field, vol. 38, Kluwer Academic Publishers, Dordrecht,The Netherlands, 1997, p. 139.

E.V. Gogol et al. / Talanta 53 (2000) 379–389 389

[28] J.J. Rippeth, T.D. Gibson, J.P. Hart, I.C. Hartley, G.Nelson, Analyst 122 (1997) 1425.

[29] T.T. Bachmann, R.D. Schmid, Anal. Chim. Acta 401(1999) 95.

[30] V. Volotovsky, N. Jaffrezic-Renault, C. Martelet, A.P.Soldatkin, Proceedings of the 11th European Conferenceon Solid State Transducers, EROSENSORS XI, 1997, p.1201.

[31] C. Dumschat, H. Muller, K. Stein, G. Schwedt, Anal.Chim. Acta 252 (1991) 7.

[32] N.F. Starodub, N.I. Kanjuk, A.L. Kukla, Y.M. Shirshov,Anal. Chim. Acta 385 (1999) 461.

[33] G.L. Ellman, K.D. Courtney, V. Andres, R.M. Feather-stone, Biochem. Pharmacol. 7 (1961) 88.

[34] G.A. Evtugyn, E.P. Rizaeva, E.E. Stoikova, H.C. Bud-nikov, Electroanalysis 9 (1997) 1124.

[35] G.G. Guilbault, D.N. Kramer, P.L. Cannon, Anal.Chem. 34 (1962) 1437.

[36] D.Z. Suznjevic, D.S. Veselinovic, N.C. Vukelic, D.Z.Pavlovic, A.V. Nikolic, J. Serb. Chem. Soc. 50 (1985) 83.

[37] L.H. Goodson, W.B. Jacobs, Enzyme Eng. 2 (1974) 55.[38] A.N. Ivanov, G.A. Evtugyn, R.E. Gyurcsanyi, K. Toth,

H.C. Budnikov, Anal. Chim. Acta 404 (2000) 55.[39] C.P. Andrieux, P. Audebert, B. Divisiablohorn, S. Lin-

quettemaillet, J. Electroanal. Chem. 353 (1993) 289.[40] H.Y. Liu, J.Q. Deng, Electrochim. Acta 40 (1995) 1845.[41] E.N. Navera, M. Suzuki, T. Takeuchi, E. Tamiya, I.

Karube, Electroanalysis 5 (1993) 65.[42] G. Cayuela, N. Pena, A.J. Reviejo, J.M. Pingarron, Ana-

lyst 123 (1998) 371.[43] B. Serra, A.J. Reviejo, C. Parrado, J.M. Pingarron,

Biosens. Bioelectron. 31 (1999) 505.[44] M.D. Trevan, Immobilized Enzymes: An Introduction

and Applications in Biotechnology, Wiley, Chichester,1980.

[45] T. Imato, N. Ishibashi, Biosens. Bioelectron. 19 (1995)435.

[46] H.C. Budnikov, G.A. Evtugyn, E.P. Rizaeva, A.N.Ivanov, V.Z. Latipova, Zh. Anal. Khim. 54 (1999) 973.

.