Analytica Chimica Acta 528 (2005) 1–8

Development of a tyrosinase biosensor based on goldnanoparticles-modified glassy carbon electrodes

Application to the measurement of a bioelectrochemical polyphenolsindex in wines

V. Carralero Sanz, Ma Luz Mena, A. Gonzalez-Cortes, P. Yanez-Sedeno, J.M. Pingarron∗

Department of Analytical Chemistry, Faculty of Chemistry, University Complutense of Madrid, 28040 Madrid, Spain

Received 1 July 2004; received in revised form 5 October 2004; accepted 5 October 2004Available online 13 November 2004

Abstract

dified withe retains ah ne theb ol, catechol,c sivec ul lifetimeo ts obtainedw nds tested.T he results( Ciocalteaur©

K

1

mncaaas

r-

e-yticalon of

yingsorsfaceicalrnalct asween

0d

The preparation of a tyrosinase biosensor based on the immobilization of the enzyme onto a glassy carbon electrode molectrodeposited gold nanoparticles (Tyr-nAu-GCE) is reported. The enzyme immobilized by cross-linking with glutaraldehydeigh bioactivity on this electrode material. Under the optimized working variables (a Au electrodeposition potential of−200 mV for 60 s, anzyme loading of 457 U, a detection potential of−0.10 V and a 0.1 mol l−1 phosphate buffer solution of pH 7.4 as working medium)iosensor exhibited a rapid response to the changes in the substrate concentration for all the phenolic compounds tested: phenaffeic acid, chlorogenic acid, gallic acid and protocatechualdehyde. A R.S.D. of 3.6% (n= 6) was obtained from the slope values of succesalibration plots for catechol with the same Tyr-nAu-GCE with no need to apply a cleaning procedure to the biosensor. The useff one single biosensor was of at least 18 days, and a R.S.D. of 4.8% was obtained for the slope values of catechol calibration ploith five different biosensors. The kinetic constants and the analytical characteristics were calculated for all the phenolic compouhe Tyr-nAu-GCE was applied for the estimation of the phenolic compounds content in red and white wines. A good correlation of tr = 0.990) was found when they were plotted versus those obtained by using the spectrophotometric method involving the Folin–eagent.

2004 Elsevier B.V. All rights reserved.

eywords:Gold nanocrystal-modified glassy carbon electrodes; Tyrosinase biosensors; Phenolic compounds; Wines

. Introduction

As it is well known, tyrosinase (Tyr) is a monophenolono-oxygenase which catalyzes the oxidation of the phe-ol group too-quinone, thus allowing a variety of phenolicompounds to be used as substrates of this enzyme. Severalmperometric biosensors based on the immobilization of Tyrt different electrode material have been described in the liter-ture. Glassy carbon electrodes modified with polymers[1],ol–gel materials[2,3], self-assembled monolayers (SAMs)

∗ Corresponding author. Tel.: +34 913944297; fax: +34 913944295.E-mail address:pingarro@quim.ucm.es (J.M. Pingarron).

on gold[4], Clark’s electrodes[5–7], reticulated vitreus cabon (RVC)[8], screen-printed[9] carbon paste[10–13]andother composite electrodes[14,15] have been used to prpare Tyr electrochemical biosensors. Some of the analcharacteristics of these bioelectrodes for the determinatiphenolic compounds are summarized inTable 1.

On the other hand, the use of gold nanoparticles is plaan increasing important role for the preparation of biosen[16]. Among other properties, they provide a stable surfor enzyme immobilization, and allow the electrochemsensing to be performed without the need of exteelectron-transfer mediators. Gold nanoparticles can ananoscale electrodes that electrically communicate bet

003-2670/$ – see front matter © 2004 Elsevier B.V. All rights reserved.oi:10.1016/j.aca.2004.10.007

2V.C

arraleroSanzetal./A

nalytica

Chimica

Acta

528(2005)1–8

Table 1Some analytical characteristics of Tyr amperometric biosensors for phenolic compounds

Biosensor Edet/technique KappM (M) Analyte/sample Linear range (×107 M) Slope,A (M−1) LOD (×107 M) Repeatability

R.S.D. (%)Useful lifetime References

Tyr/PEDT-GCE −200 mV – Phenol herbicides – 0.1412a, 0.0429b

(S= 0.07068 cm2)0.5b – 30% activity after

12 days[1]

Tyr-TiO2-sol–gel-GCE −150 mV 2.9× 10−4 Phenol 1.2–2600b 0.103b 1.0 2.8 (n= 9) 80 days [2]Tyr-SiO2-sol–gel-GCE 0 mV – Phenols 1.0–1000a 0.0596a 0.4a 1.8 (n= 8) 110 days [3]Tyr/MPA/AuE −100 mV/FI 3.345× 10−4a,

1.464× 10−4bPhenols/waste waters 2–1000a, 2–2000b 0.034a, 0.014b 1.1a, 0.88b 2.6a (n= 50),

3.6b (n= 50)5 days [4]

Tyr/GDH/ClarkE −600 mV – Catecholamines, phenols 0.1–7 – 0.1a 3a (n= 9) Various months [5]Tyr/polyacrilamide/ClarkE −700 mV – Polyphenols/E. coli – – – 7a (n= 5) 7 days [6]Tyr/ClarkE −600 mV – Polyphenols/wine – 0.073b – – – [7]Tyr-RVCE −0.20 mV/FI – Phenol herbicides 5–300b 0.0082 2.6 2.4 (n= 5) 20 days [8]PCS(Tyr)-HRP-SPCE −50 mV – Phenols/waters 0.25–450 0.0623b 0.025 4.5 60 days [9]Tyr-Au-CPE −150 mV 53.6× 10−6b Phenol, catechol dopamine 40–480 0.023 0.061 3.2 (n= 6) 2 weeks [10]Tyr-PS 086-CPE −200 mV/HPLC – Phenols Up to 1000 0.093 – – – [11]Tyr-Ru-CPE 5.9 <5 3–4 h [12]Tyr-Ru-CPE 150 2.4 (n= 40) – [13]Tyr/C-EPD 0.28a, 0.26b 4.3 (n= 10)b 5 days [14]

Tyr-C-Teflon 1.0a, 1.0b – – [14]

Tyr-C-epoxi – <2 (n= 100) 30 days [15]

MPA: 3-mercaptopropionic : poly(carbamoylsulfonate); HRP: horseradish peroxidase; SPCE: screen-printedcarbon electrode; FI: flow i

a Catechol.b Phenol.

−100 mV – Polyphenols/wine 59–3530 (gallic acid) –0 mV/FI – Phenol Up to 6000 0.00081

−150 mV 5.63× 10−5a,6.26× 10−5b

Phenols 0.5–80a, 0.5–60b 0.66a, 0.62b

−150 mV 4.99× 10−5a,7.76× 10−5b

Phenols 1–150a, 1–250b 0.30a, 0.28b

−100 mV/FI – Phenols 1.0b –

acid; C-EPD: graphite-ethylene propylene diene; PEDT: poly(3,4-ethylenedioxythiophene); PCSnjection.

V. Carralero Sanz et al. / Analytica Chimica Acta 528 (2005) 1–8 3

enzymes and bulk electrode materials. The conductivity prop-erties of such materials allow to design simple, sensitiveand stable electroanalytical procedures based on enzyme im-mobilization[17]. Thus, colloidal gold deposited on SAMsmodified-gold electrodes has been used for the immobiliza-tion of HRP[18]. More recently, colloidal-gold modified car-bon paste electrodes have been employed to develop glucoseoxidase[19] and tyrosinase[10] biosensors. On the otherhand, gold nanoparticles-modified glassy carbon electrodes(nAu-GCEs) have been characterized[20] and used for theformation of self-assembled monolayers[21]. These nAu-GCEs have demonstrated to exhibit high catalytic activitytowards some electrooxidation reactions[22]. Concerningbiosensors based on the use of nAu-GCEs, the only refer-ence found in the literature was the preparation of an elec-trode based on xanthine oxidase adsorbed to colloidal goldand evaporated onto the surface of glassy carbon[23].

In this work, we report on the preparation of a tyrosi-nase biosensor based on the use of a glassy carbon electrodemodified with electrodeposited gold nanoparticles (Tyr-nAu-GCE). The immobilized enzyme retains a high bioactivityon this electrode material, giving rise to fast, stable and sen-sitive responses to various phenolic compounds. Moreover,and considering the nutritional importance of phenolic com-ponents, due to their antioxidant power, in wines, particu-l tot olicc

2

2

AS1 cells CE)a iam-e MF2 ntere entsi po-t ap-p posee HP8 y ther e ofF

2

a fiedG ro 3,4-d oro-

genic acid (Sigma, 99%) were also used. The solutions usedfor the enzyme inmobilization were a 91.4 U�l−1 solutionof tyrosinase (Sigma, EC 1.14.18.1 fromMushroomsp.,2590 U mg−1) prepared in a 0.1 mol l−1 phosphate buffer so-lution of pH 7.4, and a 25% glutaraldehyde (Aldrich) so-lution. All other solvents and chemicals were of analyticalreagent grade. The water used was obtained from a MilliporeMilli-Q system.

2.3. Samples

The samples analysed were commercial bottled wines pur-chased in a local market. Three white wines with alcoholcontents of 11 vol.% (sample 1), 12 vol.% (sample 2) and15 vol.% (sample 3), and three red wines with alcohol con-tents of 12.5 vol.% (sample 4), and 12 vol.% (samples 5 and6) were analyzed.

2.4. Procedures

2.4.1. Preparation of the tyrosinase-goldnanocrystal-modified GCE (Tyr-nAu-GCE)

A 100 mg l−1 HAuCl4 solution was prepared in Milli-Q water previously filtered through a 0.45�m nylon filter(Whatman), and deaerated by passing a N2 stream. The GCEw sr ach,a oldn rodei n-ts trodes e wasi in.T t in0

erep ingt

2 lsi

f a0 asc em-iw om-p d ad-d 0a

yzedb ofF fd gent( d to0 was

arly in red wines[7], we have applied the Tyr-nAu-GCEhe amperometric estimation of the total content of phenompounds in this type of samples.

. Experimental

.1. Apparatus and electrodes

Voltammetric measurements were carried out with a B00B potentiostat provided with a BAS C2 EF-1080tand. A Metrohm 6.084.010 glassy carbon electrode (Gnd a Metrohm 6.1204.140 gold electrode (3.0 mm dter each) were used as working electrodes. A BAS063 Ag/AgCl 3 M reference electrode and a Pt wire coulectrode were also employed. Amperometric measurem

n stirred solutions were carried out using a PGSTAT 12entiostat from Autolab to control the potential valueslied. The electrochemical software was the general purlectrochemical system (GPES) (EcoChemie B.V.). A453 UV–vis spectrophotometer was also used to appleference spectrophotometric method involving the usolin–Ciocalteau reagent[24].

.2. Reagents and solutions

An aqueous 1% HAuCl4·3H2O solution (Sigma, >49%s Au) was used for the preparation of gold modiC electrodes. Stock 1.0× 10−2 mol l−1 solutions in watef catechol (Sigma, 99%), phenol (Scharlab, 99.5%),ihydroxybenzaldehyde, caffeic acid, gallic acid and chl

as polished with 0.3�m alumina for 1 min. Then it wainsed with ethanol and water, alternatively, three times end dried using a nitrogen stream. Modification with ganocrystals was performed by immersion of the elect

nto the 100 mg l−1 HAuCl4 solution and applying a poteial of −200 mV during 1 min. Then, 5�l (457 units) of Tyrolution were deposited on the nAu-GCE. Once the elecurface had dried out at room temperature, the electrod

mmersed in a 25% (v/v) glutaraldehyde solution for 30 mhe bioelectrode was washed with Milli-Q water and kep.1 mol l−1 phosphate buffer of pH 7.4 until using.

Amperometric measurements in stirred solutions werformed by applying the desired potential and allow

he steady-state current to be reached.

.4.2. Measurement of a bioelectrochemical polyphenondex in wines using the Tyr-nAu-GCE biosensor

A 270�l wine aliquot was directly added to 10 ml o.1 mol l−1 phosphate buffer solution of pH 7.4, which wontinuously stirred at a constant rate in the electrochcal cell. Then, amperometric measurements at−100 mVere carried out, and the estimation of the phenolic counds content was performed by applying the standaritions method, which implied the addition of sucessive 2�lliquots of a 2.5× 10−3 mol l−1 phenol stock solution.

For comparison purposes, wines were also analy the spectrophotometric method involving the useolin–Ciocalteau reagent[24]. In this method, 4.2 ml oeionized water and 0.5 ml of Folin–Ciocalteau reaphosphotungstic–phosphomolybdic acid) were adde.5 ml of sample (in the case of red wines a 1:8 dilution

4 V. Carralero Sanz et al. / Analytica Chimica Acta 528 (2005) 1–8

Fig. 1. Cyclic voltammograms for 2.0× 10−4 mol l−1 solutions of catechol (a) and caffeic acid (b), at: (1) Tyr-nAu-GCE; (2) Tyr-GCE; (3) Au-GCE; (4) GCE;v = 25 mV s−1. Supporting electrolyte: 0.05 mol l−1 phosphate buffer (pH 7.4).

carried out). The mixture was stirred for about 1 min, and1.0 ml of an 80% sodium carbonate solution and 4.2 ml ofdeionized water were added. The resulting solution was al-lowed to stand for 2 h at ambient temperature in darkness. Theabsorbance is then read at 730 nm. The same procedure wasused to construct a calibration plot with standard solutions ofcaffeic acid.

3. Results and discussion

Fig. 1 shows a comparison of cyclic voltammograms forcatechol (Fig. 1a) and caffeic acid (Fig. 1b) obtained at aTyr-nAu-GCE, at a biosensor in which Tyr was immobilized,under strictly identical conditions that those used for the Tyr-nAu-GCE, at a bare GCE (Tyr-GCE), and at a conventionalnaked GCE. As expected, no reduction signal was observedat the electrode with no enzyme. Moreover, although a slightreduction wave can be observed at the Tyr-GCE, the phenoliccompound catalytic responses were remarkably higher at theTyr-nAu-GCE, showing fairly well the advantages predictedfor this type of biosensor configuration, commented in Sec-tion 1. Obviously, the enzyme reaction involves the catalyticoxidation of the phenolic compounds to their correspondingo-quinones, at the expense of reducing oxygen to water[25].T sfer-r nitort

3

lec-t , thec i-fi

the optimized previously for the construction of the nAu-GCEs[26].

All other working variables were optimized taking thehighest value of the slope obtained for catechol calibrationplots in the (1.0–5.0)× 10−5 mol l−1 concentration range as acriterion of selection. Thus, the effect of the Au electrodepo-sition potential was checked by batch amperometry in stirredsolutions, using a Tyr-nAu-GCE biosensor constructed byimmobilisation of 457 units tyrosinase.Fig. 2a shows thatthe slope of the calibration graph for catechol increased asthe electrodeposition potential became more negative up to

F b) ontc r(

he electrochemical reduction of these quinones, by traning two electrons and two protons, was employed to mohis reaction.

.1. Optimization of the biosensor preparation

Concerning experimental variables involved in the erodeposition of gold nanoparticles on the GCE surfaceoncentration of the HAuCl4 solution from which the moded electrodes were prepared, was the same, 100 mg l−1, that

ig. 2. Effect of the applied potential (a) and of the tyrosinase loading (he slope of the calibration graph for catechol in the (1.0–5.0)× 10−5 mol l−1

oncentration range. Supporting electrolyte: 0.1 mol l−1 phosphate buffepH 7.4).

V. Carralero Sanz et al. / Analytica Chimica Acta 528 (2005) 1–8 5

Fig. 3. Scanning electronic micrographs of a nAu-GCE (a) and a Tyr-nAu-GCE (b).

−200 mV. However, a decrease in the slope value was ob-served for more negative potential values. This can be at-tributed to the larger amount of gold electrodeposited as thepotential was more negative[20]. As it has been reportedfor a colloidal-gold-modified carbon paste electrode[10], alarge amount of gold can produce a decrease in the catalyticresponse as a consequence of an increase in the resistanceand double layer capacitance of the modified electrode. Us-ing −200 mV as electrodeposition potential, an electrodepo-sition time of 60 s showed to be sufficient for the obtentionof analytically useful calibration graphs for catechol. Con-sequently, Tyr-nAu-GCEs were constructed by using goldnanocrystal-modified GCEs prepared by electrodeposition at−200 mV for 60 s.

Under these conditions, the enzyme loading influence wasalso evaluated over the 160–720 units range (Fig. 2b). Theslope of the catechol calibration graph increased with theenzyme loading up to 457 U of Tyr, after which the slopevalue exhibited a slight decrease which can be attributed toeffects such as an increase of the resistance, making the elec-tron transfer more difficult, and concentration-dependent de-naturation at the interface. Therefore, the above mentionedenzyme loading was selected for further work.

Scanning electronic micrographs of the nAu-GCE andTyr-nAu-GCE are displayed inFig. 3. Gold nanoparticlesd rfacea cor-r clicv0 fero a-l of0 hatt eo

theo age( sur-f ceda sely

trapped with the electrode material, thus improving both theretention of the biomolecule on the electrode surface and itselectrical communication.

3.2. Optimization of variables affecting theamperometric detection at the Tyr-nAu-GCE

The influence of the potential applied to the biosensoron the amperometric responses from different phenolic com-pounds is shown inFig. 4. In general, a similar shape wasobserved for all of them, with an increase of the steady-statecurrent from +0.15 V up to approximately−0.10 V. A de-crease in the current was produced at more negative poten-tials, probably due to the polymerization of the correspondingo-quinone at these negative potentials[8,28]. The behaviourshown inFig. 4 indicates that the 1,2-quinone formed in theenzyme reaction for each of the tested substrates is reducedat similar potential values. Consequently, a potential valueof −0.10 V was chosen for subsequent work. The big differ-ences in sensitivity for the different phenolic compounds, as aconsequence of their different structure, can also be observedin Fig. 4.

F nt for15

eposited had a mean size of around 80 nm. The surea covered by gold was calculated from the chargeesponding to the reduction of gold oxides formed by cyoltammetry at 100 mV s−1 between−0.35 and +1.5 V in.1 mol l−1 H2SO4 [20]. From the theoretical charge transf 482�C cm−2 for the reduction of a monolayer of div

ent oxygen from a gold surface[27], a gold surface area.0074± 0.0003 cm2 was obtained. Taking into account t

he geometric area of the GCE is 0.071 cm2, the percentagf this area covered by gold was of 10.4%.

The enzyme immobilization with glutaraldehyde canbserved by a change in topology to a speckled, grainy imFig. 3b), showing clusters of protein cross-linked on theace of the nAu-GCE. This immobilization method produ

three-dimensional matrix in which the enzyme is clo

ig. 4. Influence of the applied potential on the steady-state curre.0× 10−5 mol l−1 phenol ( ), chlorogenic acid (©), catechol (�), and.0× 10−5 mol l−1 caffeic acid (�) at a Tyr-nAu-GCE.

6 V. Carralero Sanz et al. / Analytica Chimica Acta 528 (2005) 1–8

Fig. 5. Control chart constructed for a Tyr-nAu-GCE. Data shown cor-respond to the mean values of the slopes from three successive calibra-tion pots for catechol in the (1.0–5.0)× 10−5 mol l−1 concentration range.Eapp=−0.10 V; 0.1 mol l−1 phosphate buffer of pH 7.4.

Concerning the effect of pH on the amperometric re-sponse, as expected, the higher steady-state currents wereobtained between pH 6.5 and 7.5, and therefore, a 0.1 mol l−1

phosphate buffer of pH 7.4 was selected as working medium.Under these conditions, the biosensor exhibited a rather

rapid response to the changes in the substrate concentra-tion for all the phenolic compounds tested, the steady-statecurrent being reached in less than 2 min for the pheno-lic compound exhibiting the slowest kinetics (chlorogenicacid).

3.3. Stability of the Tyr-nAu-GCE

Different aspects regarding the stability of the biosensorwere considered. First, the repeatability of the measurementswas evaluated by constructing successive calibration plotsfor catechol in the (1.0–5.0)× 10−5 M concentration rangewith the same Tyr-nAu-GCE. A relative standard deviation(R.S.D.) value of 3.6% (n= 6) was obtained from the slopevalues of such calibration graphs, which indicated a goodrepeatability of the measurements with no need to apply acleaning or pretreatment procedure to the Tyr-nAu-GCE.

The useful lifetime of one single biosensor was checkedby performing repetitive calibration graphs for catechol inthe above mentioned concentration range, the bioelectrodeb 4F eanv inedt r andl tv rec threes alues

TK differen

C inear r

P 0.01–0C 0.005–C 0.02–2C 0.01–2G 0.25–9P 0.06–6

remained within the control limits for at least 18 days, no databeing recorded for larger times.

Finally, the reproducibility of the responses obtained withdifferent Tyr-nAu-GCEs was also evaluated. Results fromfive different biosensors yielded a R.S.D. of 4.8% for theslope values of the corresponding catechol calibration plots.This result demonstrated the reliability of the Tyr-nAu-GCEfabrication procedure, allowing reproducible electroanalyti-cal responses to be obtained with different biosensors con-structed in the same manner.

3.4. Kinetic constants and analytical characteristics

Under the experimental conditions optimized above, thekinetic parameters and the analytical characteristics of thetyrosinase reaction were calculated for the following pheno-lic compounds: phenol, catechol, caffeic acid, chlorogenicacid, gallic acid and 3,4-dihydroxybenzaldehyde (protocat-echualdehyde) (Table 2). The kinetics of the enzyme reac-tion fitted in all cases into a Michaelis–Menten type kinetics,as demonstrated by calculation of the parameterx from theHill’s plot (log[( imax/i) − 1] versus the log of the substrateconcentration), indicating that the immobilization proceduredid not alter the Michaelis–Menten behaviour. Then, calcula-tion of the apparent Michaelis–Menten constants (K

appM ) and

t -p . Asio ilart ord mo-b

ibra-t Thec dingt c-t ardd ents orre-s plot.A sorf dingK tics( heTf erved

eing stored after use in phosphate buffer of pH 7.4 at◦C.ig. 5 shows the control chart constructed, taking the malue of the slopes of 10 successive calibration plots obtahe first day of this study as the central value. The uppeower limits of control were set at±3× S.D. of this targealue (0.113± 3× 0.006). The values shown in this figuorresponded to the mean values from the slopes ofuccessive calibration plots. As can be observed, these v

able 2inetic parameters of the tyrosinase reaction and calibration data for

ompound x KappM (mM) Vm (�A) L

henol 1.01 0.14 15.7atechol 1.01 0.12 28.3affeic acid 0.99 0.22 5.9hlorogenic acid 1.01 0.18 4.9allic acid 0.99 0.44 0.30rotocatechualdehyde 1.10 0.37 0.71

t phenolic compounds obtained at Tyr-nAu-GCEs

ange (×104 M) r Slope,A (M−1) LOD (×107 M)

.4 0.9993 0.082 2.10.5 0.9998 0.107 1.5.0 0.998 0.014 6.6.0 0.9994 0.017 6.2

0.998 2.3× 10−4 700.998 8.3× 10−4 20

he maximum rate values of the reaction (Vm) was accomlished from the corresponding Lineweaver–Burk plots

t was expected, the lowerKappM and the higherVm values were

btained for catechol and phenol. Moreover, they are simo those reported inTable 1for other tyrosinase biosensesigns, thus revealing a good affinity of the enzyme imilized on gold nanoparticles for these substrates.

Table 2also summarizes the characteristics of the calion plots obtained for the phenolic compounds tested.orresponding limits of detection were calculated accoro the 3sb/mcriterion, wheremis the slope value of the respeive calibration graph, andsb was estimated as the standeviation (n= 10) of the amperometric signals from differolutions of the substrates at the concentration level cponding to the lowest concentration of the calibrations it is theoretically predicted, the sensitivity of the biosen

or each phenolic compound is higher as the corresponappM is lower. Moreover, when the analytical characteris

range of linearity, sensitivity, and limit of detection) of tyr-nAu-GCE are compared with those shown inTable 1or other tyrosinase biosensor designs, it can be obs

V. Carralero Sanz et al. / Analytica Chimica Acta 528 (2005) 1–8 7

that they are also similar to those of the bioelectrodes ex-hibiting a better performance for the amperometric detectionof these compounds. It is interesting to compare the perfor-mance of the Tyr-nAu-GCE with that reported for tyrosinasecolloidal-gold-modified carbon paste electrodes[10]. As itcan be deduced fromTables 1 and 2, the linear range forphenol (it is the only phenolic compound for which quanti-tatively data are given in Ref.[10]) is wider, the sensitivityis higher and the useful lifetime longer with the Tyr-nAu-GCE. Moreover, it must be noted that the preparation of goldnanocrystal-GCEs by gold electrodeposition is a much morerapid, simple, reproducible and easy to be controlled processthan the preparation of colloidal-gold-modified carbon pasteelectrodes by mixing a colloidal gold suspension, preparedpreviously by a time-consuming and delicate procedure[10],and carbon paste.

3.5. Measurement of a bioelectrochemical polyphenolsindex in wines using the Tyr-nAu-GCE

The practical usefulness of the Tyr-nAu-GCE was evalu-ated by estimating the “pool” of polyphenols in wines, whichis of interest because of the correlation between wines antiox-idant capacity and their polyphenol content. Red and whitewines were analyzed following the extremely simple proce-d aw tiono stratet lved.T as ap g lo on-s auseo om-p CE.T ticalm ofi g ont

calwt e de-t lysedu e of

TE Tyr-n

W d

WWWRRR

Fig. 6. Correlation between the results obtained for the estimation of thecontent of phenolic compounds in wines by using the Tyr-nAu-GCE and theFolin–Ciocalteau method.

Folin–Ciocalteau reagent[24]. This reagent by reacting withthe phenol OH group, produced a blue coloured complexwhose absorbance is read at 730 nm. The total amount ofpolyphenols is estimated in this method by comparison witha standard solution of caffeic acid. The results obtained arealso shown inTable 3.

In spite of the big differences observed for these twopolyphenol index values, as a consequence of the completelydifferent analytical methodologies employed, a good correla-tion was found (r = 0.990) when the results obtained with thebiosensor were plotted versus the results achieved with theFolin–Ciocalteau method (Fig. 6). This good correlation ob-tained with different types of wines, which is also kept whencaffeic acid was used in the standard additions method in-stead of phenol as the enzyme substrate, allows the use of theTyr-nAu-GCE for a rapid and in situ measurement of a bio-electrochemical polyphenol index able to give an estimationof the content of phenolic compounds in wines. Moreover, asexpected, important differences in the polyphenol content ofred and white wines were found by both methods, which sup-ports the claimed higher antioxidant capacity of red wines.

4. Conclusions

ssyc goldn hiche ata-b reat-m ono etricd arac-t thet ancef Tyr-n ation

ure described in Section2, in which the direct addition ofine aliquot to the electrochemical cell, and the applicaf the standard additions method using phenol as the sub

o perform the successive standard additions, were invohe content of phenolic compounds was then obtainedhenol concentration, which was further expressed in m−1

f caffeic acid[7]. The estimation of this content must be cidered as a bioelectrochemical polyphenol index, becf the different sensitivity observed for each phenolic cound, depending on their structure, with the Tyr-nAu-Ghis is a common approach to that used with other analyethodologies[7], although it is obvious that this type

ndexes will provide different absolute values dependinhe particular analytical methodology applied.

The results obtained with the Tyr-nAu-GCE for three lohite wines and three red wines are summarized inTable 3,

he values given corresponding to the mean value of fiverminations. The same wine samples were also anasing the spectrophotometric method involving the us

able 3stimation of the content of phenolic compounds in wines by using aAu-GCE

ine Caffeic acid(mg l−1)

Folin–Ciocalteau metho(mg l−1 caffeic acid)

hite “Emparrado” 16.1± 0.4 257± 9hite “Sotillo” 18 ± 3 303± 73hite “Alvear, J. Oro” 18± 3 314± 15ed “Anadas de Oro” 47± 4 1790± 61ed “San Asensio” 50± 4 2266± 53ed “Sotillo” 35 ± 2 1112± 78

Immobilization of tyrosinase by cross-linking onto glaarbon electrodes modified with electrodepositedanoparticles permits the construction of biosensors wxhibit a good analytical performance (concerning repeility of the measurements with no need of cleaning pretent, stability with time, reproducibility in the constructif different biosensors, and sensitivity) for the amperometection of phenolic compounds. These analytical ch

eristics are similar to those reported in the literature foryrosinase bioelectrodes with a claimed better performor these compounds. The practical usefulness of theAu-GCEs has been demonstrated by the rapid estim

8 V. Carralero Sanz et al. / Analytica Chimica Acta 528 (2005) 1–8

of the content of phenolic compounds in wines, using an ex-tremely simple procedure involving the direct addition of asample aliquot to the electrochemical cell.

Acknowledgements

Financial support from the Ministerio de Ciencia y Tec-nologıa (Projects BQU 2001-2050 and BQU 2003-00365),and Comunidad de Madrid Project No. 07G/0006/2003 aregratefully acknowledged.

References

[1] C. Vedrine, S. Fabiano, C. Tran-Minh, Talanta 59 (2003) 535.[2] J. Yu, S. Liu, H. Ju, Biosens. Bioelectron. 19 (2003) 509.[3] B. Wang, J. Zhang, S. Dong, Biosens. Bioelectron. 15 (2000) 397.[4] S. Campuzano, B. Serra, M. Pedrero, F.J. Manuel de Villena, J.M.

Pingarron, Anal. Chim. Acta 494 (2003) 187.[5] K. Streffer, E. Vijgenboom, A.W.J.W. Tepper, A. Makower, F.W.

Scheller, G.W. Canters, U. Wollenberger, Anal. Chim. Acta 427(2001) 201.

[6] Y. Hasebe, K. Yokobori, K. Fukasawa, T. Kogure, S. Uchiyama,Anal. Chim. Acta 357 (1997) 51.

[7] L. Campanella, A. Bonanni, E. Finotti, M. Tomassetti, Biosens. Bio-electron. 19 (2004) 641.

[9] S.C. Chang, K. Rawson, C.J. McNeil, Biosens. Bioelectron. 17(2002) 1015.

[10] S. Liu, J. Yu, H. Ju, J. Electroanal. Chem. 540 (2003) 61.[11] K.R. Rogers, J.Y. Becker, J. Cembrano, Electrochim. Acta 45 (2000)

4373.[12] W.T. Jewell, S.E. Ebeler, Am. J. Enol. Vitic. 52 (2001) 219.[13] J. Wang, F. Lu, D. Lopez, Biosens. Bioelectron. 9 (1994) 9.[14] B. Serra, S. Jimenez, L. Mena, J. Reviejo, J.M. Pingarron, Biosens.

Bioelectron. 17 (2002) 217.[15] P. Onnerfjord, J. Emneus, G. Marko-Varga, L. Gorton, F. Ortega, E.

Domınguez, Biosens. Bioelectron. 10 (1995) 607.[16] S. Liu, D. Leech, H. Ju, Anal. Lett. 36 (2003) 1.[17] E. Katz, I. Willner, J. Wang, Electroanalysis 16 (2004) 19.[18] S. Liu, H. Ju, Anal. Biochem. 307 (2002) 110.[19] S. Liu, H. Ju, Biosens. Bioelectron. 19 (2003) 177.[20] M.O. Finot, G.D. Braybrook, M.T. McDermott, J. Electroanal. Chem.

466 (1999) 234.[21] M.O. Finot, M.T. McDermott, J. Electroanal. Chem. 488 (2000)

125.[22] K. Yahikozawa, K. Nishimura, M. Kumazawa, N. Tateishi, Y. Takasu,

K. Yasuda, Y. Matsuda, Electrochim. Acta 37 (1992) 453.[23] J. Zhao, J.P. O’Daly, R.W. Henkens, J. Stonehuerner, A.L. Crumbliss,

Biosens. Bioelectron. 11 (1996) 493.[24] V.L. Singleton, J.A. Rossi, Am. J. Enol. Vitic 16 (1965) 144.[25] C. Nistor, J. Emneus, L. Gorton, A. Ciucu, Anal. Chim. Acta 387

(1999) 309.[26] V. Carralero Sanz, A. Gonzalez-Cortes, P. Yanez-Sedeno, J.M. Pin-

garron, Electroanalysis, in press.[27] U. Oesch, J. Janata, Electrochim. Acta 28 (1983) 1237.[28] B. Serra, E. Mateo, M. Pedrero, A.J. Reviejo, J.M. Pingarron, Anal-

[8] N. Pena, A.J. Reviejo, J.M. Pingarron, Talanta 55 (2001) 179.

ysis 27 (1999) 592.