Mediatorless biosensor for H2O2 based on recombinant forms of horseradish peroxidase directly...

Biosensors & Bioelectronics 16 (2001) 147–157

Mediatorless biosensor for H2O2 based on recombinant forms ofhorseradish peroxidase directly adsorbed on polycrystalline gold

Elena E. Ferapontova a,*, Vitaly G. Grigorenko a, Alexey M. Egorov a,Torsten Borchers b, Tautgirdas Ruzgas c, Lo Gorton c

a Department of Chemical Enzymology, Chemical Faculty, Moscow State Uni�ersity, Vorobe’�y gory, Moscow, 119899 Russiab Institute of Chemical and Biochemical Sensors Research, Mendelstr. 7, Muenster, D-48149 Germany

c Department of Analytical Chemistry, Uni�ersity of Lund, PO Box 124, SE 221 00 Lund, Sweden

Received 25 May 2000; received in revised form 18 December 2000; accepted 1 March 2001

Abstract

Four forms of horseradish peroxidase (HRP) have been used to prepare peroxidase-modified gold electrodes for mediatorlessdetection of peroxide: native HRP, wild type recombinant HRP, and two recombinant forms containing six-His tag at theC-terminus and at the N-terminus, respectively. The adsorption of the enzyme molecules on gold was studied by direct massmeasurements with electrochemical quartz crystal microbalance. All the forms of HRP formed a monolayer coverage of theenzyme on the gold surface. However, only gold electrodes with adsorbed recombinant HRP forms exhibited high and stablecurrent response to H2O2 due to its bioelectrocatalytic reduction based on direct electron transfer between gold and HRP. Thesensitivity of the gold electrodes modified with recombinant HRPs was in the range of 1.4–1.5 A M−1 cm−2 at −50 mV versusAg�AgCl. The response to H2O2 in the concentration range 0.1–40 �M was not dependent on the presence of a mediator (i.e.catechol) giving strong evidence that the electrode currents are diffusion limited. Lower detection limit for H2O2 detection was 10nM at the electrodes modified with recombinant HRPs. © 2001 Elsevier Science B.V. All rights reserved.

Keywords: Recombinant horseradish peroxidase; Direct electron transfer; Gold; Electrochemical quartz crystal microbalance

www.elsevier.com/locate/bios

1. Introduction

Horseradish peroxidase (HRP, hydrogen peroxideoxidoreductase, EC-1.11.1.7) is a heme- and Ca2+-containing glycoprotein, a member of the plant perox-idase superfamily, and is able to utilize hydrogen perox-ide to catalyze one electron oxidation of a wide varietyof organic and inorganic substrates (Welinder, 1992;Dunford, 1999). The high catalytic activity of HRP inthe reaction of H2O2 reduction makes it attractive touse HRP as a highly effective bioelectrocatalyst if elec-tron transfer (ET) between the surface of the electrodeand the active center of the enzyme can be established.

The most simple electrode for the detection of H2O2

represents a monolayer of HRP molecules adsorbed on

the electrode surface. When applying a correct potentialto the HRP-modified electrode the registered reductioncurrent will be proportional to the peroxide concentra-tion. The following reaction is expected to proceed(Ruzgas et al., 1995):

HRP(Fe3+)+H2O2�k1

E1+H2O (1)

E1+2e−+2H+�ks

HRP(Fe3+)+H2O (2)

Here, HRP(Fe3+) describes the initial oxidation stateof HRP immobilized on the electrode surface, and E1 isthe compound representing oxidized HRP (oxidationstate +5), consisting of oxyferryl iron (Fe4+�O) and aporphyrin � cation radical. The first reaction involves atwo-electron oxidation of the ferriheme prostheticgroup of the peroxidase by peroxide. This mediatorlessbioelectrocatalytic reaction has been observed for HRPadsorbed on carbon black (Yaropolov et al., 1978),carbon and graphite (Razumas et al., 1984; Bog-danovskaya et al., 1988; Jonsson and Gorton, 1989;

* Corresponding author. Present address: Department of Analyti-cal Chemistry, Faculty of Pharmacy, University of Alcala, 28871Alcala de Henares, Madrid, Spain. Fax: +34-918854666.

E-mail address: [email protected] (E.E. Ferapontova).

0956-5663/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.PII: S0956-5663(01)00134-8

E.E. Feraponto�a et al. / Biosensors & Bioelectronics 16 (2001) 147–157148

Bogdanovskaya, 1993; Ruzgas et al., 1995; Lindgren etal., 1998, 1999), gold (Yaropolov et al., 1978; Zhao etal., 1992; Presnova et al., 2000), gold modified withviologens (Razumas et al., 1986), and platinum (Durliatet al., 1989). The direct electrochemical reduction ofcompound E1 has been shown to be kinetically slow ona majority of the electrode materials with values of theET rate, ks, generally, below one electron per s (Ruzgaset al., 1996).

To overcome a slow direct ET of reaction (2) smallredox active molecules with high ET rates (mediators)are often used, such as phenols, aromatic amines, io-dide, ferrocenes. They are able to diffuse in and reactwith the active site of the enzyme, then diffuse out andreact with the electrode surface, shuttling the electronsbetween the electrode and the active site of the enzyme(Ruzgas et al., 1995, 1996). When an electron donorother than the electrode is present in the HRP-electrodesystem compound E1 can be reduced by direct ET orby this electron donor substrate S (mediated ET) (Ruz-gas et al., 1995):

E1+S+H+�k2

E2+S� (3)

E2+S+H+�k3

HRP(Fe3+)+S�+H2O (4)

In reaction (3) compound E1 is reduced upon one-electron reduction by the mediator S and forms com-pound E2 (oxidation state +4). The latter reduces inthe next step (4) to the initial peroxidase state,ferriperoxidase.

The formed oxidized donor S� is then electrochemi-cally reduced by the electrode:

S�+e−�ks,m

S (5)

Mediated ET is known to be more efficient in com-parison with direct ET (Zhao et al., 1992; Ruzgas et al.,1995). Disadvantages of mediated ET are the drop ofthe reaction selectivity as redox mediators are rathergeneral catalysts that facilitate not only ET between theelectrode and the enzyme but also some interferingreactions.

HRP immobilized at the electrode surface and capa-ble of direct ET without the addition of any mediatorgives the opportunity for creation of mediatorlessbiosensors for peroxide and other metabolites if com-bined with oxidase enzymes, e.g. glucose oxidase, lysineoxidase (Ruzgas et al., 1996). However, direct ET in thesystem gold electrode-native HRP in the absence ofmediator is known to be a low efficient one (Yaropolovet al., 1978; Zhao et al., 1992). The reasons for that arelikely to be the long ET distance between the electrodesurface and the active center of the enzyme, possibly incombination with unfavorable orientation of themolecule at the electrode or low adsorption of theglycosylated native enzyme leading to a low surfaceconcentration of the catalyst, i.e. native HRP.

Electrochemical estimations of the surface concentra-tion of immobilized biomolecules on electrode surfacesalways involve certain obstacles since only the biologi-cally or electrochemically active fraction of the adsor-bate can be determined. The total surface concentrationin this case can be estimated only with a certain degreeof approximation due to minute amounts of the sub-stance adsorbed. Advances in piezocrystal (PZ)methodology in the last decades now allow measuringminute mass changes (less than 10−9 g cm−2) at theelectrode interface in liquid media, making possible toexploit this principal in chemical and biological sensordevelopment (Ward and Buttry, 1990). The microgravi-metric assay is based on measuring small mass changesof a quartz PZ crystal, which are due to the adsorptionof a reactant on a specially modified surface. The masschange is related to the variation of the resonant fre-quency �f of the PZ crystal. The relationship betweenthe change of surface mass �m (g) and �f (Hz) forAT-cut piezoelectric crystals is given by the Sauerbreyequation (Sauerbrey, 1959).

�ff0

2 =−2�m

A(�q�q)1/2 (6)

where A is the surface area covered by adsorbed mate-rial (cm2), �q is the shear quartz modulus (�q=2.947×1011 dyne cm−2), �q is the quartz density (�q=2.648 gcm−3) and f0 is the fundamental resonant frequency ofthe unloaded crystal oscillator. Eq. (6) describes thefrequency changes in vacuum due to mass changes ofdeposits uniformly spread over the active surface of theelectrode. It can be used under conditions of the liq-uid�electrode interface with a certain degree of approxi-mation (Kanazawa and Gordon, 1985). In a greatnumber of reports, comparisons between the calculatedmass change using Eq. (6) and the mass value deter-mined from potentiostatic, galvanostatic, and cyclicvoltammetry measurements appeared to be fairly good(Varineau and Buttry, 1987; Deakin et al., 1988; Bakerand Reynolds, 1988; Ward, 1989; Koh et al., 1993).

It was earlier shown (Lindgren et al., 1999) thatrecombinant HRPs immobilized directly on the surfaceof graphite electrodes demonstrated a higher percentageof properly oriented molecules for direct ET moleculesthan the native enzyme, this fact being ascribed to theabsence of oligosaccharide chains in the recombinantenzymes expressed in E. coli. Thus, it was supposedthat for highly hydrophobic gold electrodes modifiedwith recombinant HRPs the percentage of the enzymemolecules active in direct ET, as well as the efficiency ofdirect ET should be even higher. The aim of the presentwork is to study the effect of deglycosylation of nativeHRP on the electrochemical sensor characteristics ofperoxidase-modified gold electrodes. This was done bycomparing the electrochemical behavior of non-glycosy-lated wild type recombinant versus native HRP. Intro-

E.E. Feraponto�a et al. / Biosensors & Bioelectronics 16 (2001) 147–157 149

duction of hexahistidine sequences at the N- and C-termini of the enzyme by genetic engineering of non-glycosylated recombinant HRP was performed to facili-tate the adsorption and possibly the orientation of theenzyme on polycrystalline gold electrodes.

2. Experimental

2.1. Instrumentation

Piezoelectric (PZ) quartz crystals coated with gold(OMIG Co, Poland) served as gold electrodes. The PZquartz crystals were AT-cut, plano-plano; with a basicresonant frequency of 10 MHz; the crystal diameterwas 14 mm with a gold coating diameter of 5 mm. Thesensitivity was specified to be equal to 4.5 ng Hz−1

cm−2. Electrochemical Quartz Crystal Microbalance,EQCM (type 5510, the Institute of the Physical Chem-istry, Polish Academy of Sciences), was used to measurethe frequency of the vibrating crystal. Amperometricmeasurements with the PZ gold electrodes were carriedout with a three-electrode potentiostat �AUTOLAB(Type II, Eco Chemie B.V., Netherlands) equipped witha General Purpose Electrochemical System GPES 4.7software (Eco Chemie). The quartz crystal coated goldelectrodes were inserted into a Kel-F®/PTFE wall-jetcell (the Institute of the Physical Chemistry, PolishAcademy of Sciences). The PZ quartz crystal coveredwith gold represented one of the cell walls and served asthe working electrode. A silver ring, pressed in Teflon,was used as a pseudo-reference electrode (the support-ing electrolyte always contained 0.15 M NaCl), and aplatinum ring of a larger diameter, also pressed inTeflon, served as the auxiliary electrode.

Amperometric measurements with polycrystallinegold disk electrodes (BAS, West Lafayette, IN) ofdiameter 0.15 cm were performed in a standard three-electrode flow-injection cell (Lindgren et al., 1998). Thecell contained a Ag�AgCl�0.1M KCl reference electrode.The auxiliary electrode was a platinum wire. The dis-tance between the nozzle and the working electrode wasabout 0.8 mm. The flow of the solutions was main-tained by a peristaltic pump MINIPULS 2 (Gilson,Villiers-le-Bel, France). Pump pulsations were attenu-ated by a 0.5 m length Teflon tubing (0.61 mm diame-ter) placed between the cell and the pump.

2.2. Materials

Native horseradish peroxidase (native HRP) andother reagents from Sigma Chemical Co.(St. Louis,USA) were used and were of analytical grade or ofultra-high purity (NaCl). All solutions were preparedwith deionized Milli-Q water (Millipore, Bedford, MA,USA).

2.2.1. Vectors construction for HRP expression in E.coli

Recombinant wild type HRP (rec-HRP), recombi-nant HRP containing six histidine tags at C-terminus ofthe enzyme (CHisrec-HRP), and recombinant HRP con-taining 6 histidine tag at N-terminus of the enzyme(NHisrec-HRP) have been produced in E. coli strainBL21(DE3)pLysS transformed with the appropriatepET based expression vectors. The procedure is coveredin detail elsewhere (Egorov et al., 1994; Grigorenko etal., 1999).

To produce wild type recombinant enzyme, the HRPgene was amplified by PCR techniques using the proof-reading polymerase Pfu (Stratagene) where forward andreverse primers were: 5�-AACATATGCAGTTAACGC-CGACTTTCTACG–3� and 5�-TTCGGCCGTCAT-GAGTTCGAGTTTACGACTCGGCAGTTC–3�,respectively, with pETHRPhis (Grigorenko et al., 1999)as a template. Bold italic letters indicate the introducedrestriction sites (NdeI and XmaIII); stop codon intro-duced just after the Ser308 is underlined. The amplifiedPCR fragment was first ligated into pCRScript vector(Stratagene) and further recloned into pET20 (No-vagen) cut with NdeI and XmaIII. The final constructpETHRP, corresponding to recombinant wild typeHRP (rec-HRP) was confirmed with sequencing.

To produce recombinant HRP with 6 histidine tag atN-terminus (NHisrec-HRP) pETHRP has been digestedwith NdeI and PstI; 2 fragments NdeI-PstI and PstI-PstI (one PstI site in HRP gene and one in bla gene inthe vector) were purified from agarose gel and ligatedinto pET15 (Novagen) cut with NdeI and PstI andtransformed into E. coli DH5� cells. Several colonies(only those with correct orientation of PstI-PstI frag-ment were able to grow on ampicilin plates) werepicked and analyzed by restriction with NdeI and PstIand further confirmed by sequencing.

2.3. Immobilization and measurement procedure

All QCM and electrochemical experiments were per-formed at ambient temperature (22�1°C).

2.3.1. Measurements with EQCMPretreatment of the PZ electrodes used for micro-

weighing measurements was the following (procedure1): the electrodes were cleaned by keeping in 3 MNaOH for 10 min, then in concentrated sulfuric acidfor 10 min, rinsed with de-ionized water and dried in anargon flux. Then the dried gold electrodes were posi-tioned into the cell with the cell part including auxiliaryand reference electrodes removed and leaving the work-ing surface of the crystal exposed to air. Then, 60 �l ofa freshly prepared hot mixture of concentrated sulfuricacid and 30% hydrogen peroxide (1:1) were droppedonto the surface of the PZ electrode for 2 min (NB:


Piranha solution is highly hazardous one and reactsviolently with organic materials; precautions should betaken at all times when handled). Immediately after theelectrode surface was rinsed with de-ionized water leav-ing a thin layer of H2O in the cell compartment toprevent any direct contact of the cleaned gold surfacewith air. After that 200 �l of a 0.01 mg ml−1 HRPsolution in 0.01 M phosphate buffer containing 0.15 MNaCl (PBS), pH 7.4, were pipetted into the cell com-partment. By immediate setting the frequency counterto zero the EQCM measurements of the adsorption ofperoxidase on gold was started. After the measurementof the adsorption process, i.e. when the EQCM fre-quency was stabilized, the measurement cell was assem-bled by carefully placing the cell part containing theinjection port and the two, i.e. reference and counter,electrodes on the cell compartment containing theworking gold electrode. Buffer solution was aspiratedthrough the cell and appropriate electrode potentialswere applied for further combined electrochemical andEQMC measurements.

In order not to break off the frequency signals fromthe steady values obtained in the buffer flow before theaddition of the peroxidase solution the PZ electrodesurface was treated according to procedure 2 describedbelow. Prior to measurements the PZ electrodes werecleaned by keeping in 3 M NaOH for 10 min, then inconcentrated sulfuric acid for 10 min, rinsed with de-ionized water, treated in the hot mixture of concen-trated sulfuric acid and 30% hydrogen peroxide for 5min, rinsed with de-ionized water and dried in an argonflux. After that the dried electrodes were positioned intothe cell leaving only one surface of the crystal incontact with the solution, with the opposite side beingexposed to air. Then the electrodes were polarized at 0V in a flow of PBS. After a steady signal was reached(the frequency response in the flow of PBS being con-stant) 150 �l of a 0.01 mg ml−1 native or recombinantforms of HRP solution in PBS pH 7.4, were injectedinto the cell. The content of the cell was left still for 3h to achieve binding of HRP onto the gold surface. Theadsorption process was registered with EQCM in thecourse of the adsorption process. After this procedurethe electrodes were thoroughly rinsed in a flow of PBSin situ for 0.5–1 h until the frequency value wasstabilized. Further simultaneous microweighing andelectrochemical measurements were performed.

2.3.2. Electrochemical measurements in the wall-jetsystem

The surface of the gold disk electrode (BAS) waspolished on fine emery paper (Tufbak Durite, P2000)and washed with de-ionized water. After that the elec-trode was rinsed with water and polished to a mirrorluster with alumina suspension (size of the particles was0.1 �m) in water, and finally rinsed with de-ionized

water. Then the electrode was immersed for 2 min in afreshly prepared hot mixture of concentrated sulfuricacid and 30% hydrogen peroxide (1:1) and then rinsedthoroughly with de-ionized water. Immediately afterthat for adsorption of HRP, the electrode was im-mersed in a 0.01 mg ml−1 HRP solution in PBS, pH6.0, for 2 h. After that the electrode was rinsed in PBS,inserted into the flow-through electrochemical cell, andused in the following measurements. The HRP solutionfor the immobilization was prepared directly before thisprocedure from stock solutions with a concentration ofHRP higher than 0.5 mg ml−1.

During the amperometric measurements in the wall-jet cell, a buffer solution pH 7.4, was pumped throughthe cell with a flow rate of 900 �l min−1 and asteady-state baseline current was registered at −50 mVversus the Ag�AgCl reference electrode. For the mea-surement of the response of the peroxidase-modifiedgold electrode to peroxide, the flow-carrier buffer solu-tion containing hydrogen peroxide was used. The signaldifference between the background current and thesteady state current in the presence of hydrogen perox-ide was taken as the response signal for the peroxidasemodified electrodes. For the experiments on mediatedET, the flow-carrier buffer solution containing hydro-gen peroxide and 5×10−4 M catechol was used.

The reproducibility of the data was verified by mea-surements with at least three equivalently preparedelectrodes.

3. Results and discussion

3.1. Immobilization of HRPs at the gold piezocrystalelectrodes

The immobilization experiments were performed withnative HRP, rec-HRP, CHisrec-HRP, and NHisrec-HRP.All recombinant peroxidases were non-glycosylated, thelast two forms additionally containing a histidine tag ateither the C- or the N-terminus of the enzyme. Thestructural peculiarities of rec-HRP of wild type and ofthe other two recombinant forms can be followed fromFig. 1. The heme, the C-terminus and the N-terminusof the enzyme are colored with black. As can be seen,the introduction of a histidine tag at the differenttermini of the enzyme may result in different orienta-tions of the molecule on the surface of the electrode (insupposition of the adsorption through the histidineresidues). Thus, any difference in the properties be-tween these four peroxidases was assumed to result indifferent arrangements and orientations of the HRPmolecules on the surface of the gold electrode and thusin different characteristic of direct ET including sensi-tivity for H2O2.


The course of immobilization of the peroxidases wasstudied by quartz microbalance measurements. Figs. 2and 3 present the raw data as a resonant frequency shiftwith time due to the adsorption of CHisrec-HRP (Fig. 2)and native HRP (Fig. 3) on the gold/quartz crystal

electrodes. The curves denoted 1 in Figs. 2 and 3correspond to the pretreatment of the electrode surfaceaccording to procedure 1. The general shape of thecurves and frequency shifts for the other recombinantforms under study were very close to those presented in

Fig. 1. Three-dimensional structure of recombinant horseradish peroxidase also indicating the location of the N- and C-termini.

Fig. 2. Variation of resonant frequency shift for gold quartz PZ crystal electrodes during the immobilization of CHisrec-HRP from 0.01 mg ml−1

solution in PBS, pH 7.4, at 23°C. The surface of gold was pretreated according to (1) procedure 1 and to (2) procedure 2. Curve 2 was speciallysmoothed for better data presentation.


Fig. 3. Variation of resonant frequency shift for gold quartz PZ crystal electrodes during the immobilization of native HRP from 0.01 mg ml−1

solution in PBS, pH 7.4, at 23°C. The surface of gold was pretreated according to (1) procedure 1 and to (2) procedure 2. Curve 2 was speciallysmoothed for better data presentation.

Table 1Adsorption of native and recombinant HRPs at gold PZ crystal electrodes, surface pretreatment with procedure 1

HRP surface concentration (pmoles cm−2)Frequency shiftMolecular weightPeroxidase type HRP surface concentration (ng(kDa) cm−2)(Hz)

Native HRP 44 −230�21 1016�93 23.1�2.1−230�20 1016�8834 29.9�2.6rec-HRP−240�40 1060�177 31.2�5.2CHisrec-HRP 34−250�30 1104�133 31.5�3.934NHisrec-HRP

Figs. 2 and 3 (curves 1), practically undistinguishablewithin experimental error. The decrease of the reso-nance frequency of the gold-quartz electrodes presentedin Figs. 2 and 3 is due to the enzyme binding to thesurface of the gold. From a net frequency variation thecorresponding amount of the molecules bound to thesurface of gold/quartz electrode was determined in ac-cordance with Eq. (1) using the difference between thetwo steady state frequency values related to the initialand final time of the immobilization experiment. Theresults obtained with gold after its pretreatment withprocedure 1 are tabulated in Table 1. The amount ofrecombinant HRP, i.e. rec-HRP, NHisrec-HRP, orCHisrec-HRP, adsorbed on gold is virtually the same(Table 1). The amount of adsorbed native HRP inmoles is about 30% lower than that for the recombinantforms as a result of a higher molecular weight of nativeHRP due to glycosylation. From the measurements ofthe mass of adsorbed HRP, the geometrical sizes ofHRP molecules were calculated. As a result, the follow-ing values of the diameter were calculated for nativeHRP 3 nm and for the recombinant forms 2.6–2.7 nm,assuming that the adsorbed enzyme molecules have aglobular form. The resulting numbers are in good

agreement with the crystallographic dimensions of HRP(Gajhede et al., 1997).

Equal mass measurements of adsorbed HRP havebeen performed using gold quartz electrodes pretreatedaccording to procedure 2. Examples of raw data arepresented in Figs. 2 and 3 (curves 2). From these datait is obvious that the amounts of adsorbed enzyme are50–70% lower than in the case of the gold pretreatedwith procedure 1; the amount of CHisrec-HRP is nearlytwice as large as that for immobilized native HRP andthe binding proceeds more than twice as rapidly. Thesurface concentrations calculated from the microbal-ance measurements on gold pretreated according toprocedure 2 are presented in Table 2. On further elec-trochemical measurements with EQCM in a flow of10−4 M H2O2 (polarization at −50 mV) native HRP,immobilized according to procedure 2, desorbed fromthe surface of the gold PZ electrode contrary to recom-binant HRP, which desorbed only partially (Table 2).This fact implies that the binding of native HRP withthe electrode surface is less strong compared with thatof the recombinant forms.

If one assumes that the non-glycosylated free recom-binant peroxidases are more hydrophobic if compared


with glycosylated native HRP one may speculate thatthe adsorption of recombinant enzyme molecules oc-curs due to hydrophobic interactions between the sur-face of gold and the hydrophobic surface clusters of therec-HRP molecule. The His tags present on the surfaceof CHisrec-HRP and NHisrec-HRP might additionallyincrease the strength of binding interaction between theenzyme and gold. This is supported by the fact thathistidine adsorbs irreversibly on gold (Safronov et al.,1983; Khudyakova et al., 1997). The adsorption processresults in a monolayer coverage of HRP. The adsorbedamounts of the enzyme are equal for the differentrecombinant forms of HRP. The microbalance mea-surements demonstrate that the amount of the ad-sorbed enzyme and the strength of the binding with thegold surface are effected by the procedure of the pre-treatment of the gold surface.

3.2. Amperometric detection of H2O2 in the case ofdirect and mediated ET

Amperometric detection of H2O2 was carried outwith gold disk electrodes modified with the variousforms of peroxidase enzymes. The electrodes were in-serted into a wall-jet cell and the electrochemical reduc-tion of H2O2 at −50 mV (Ruzgas et al., 1995) wasstudied. For comparison, the experiments were alsoperformed using bare gold electrodes. The current re-sponse due to the electroreduction of 10−4 M H2O2 onthe unmodified gold surface was 100 times lower thanthat due to the bioelectrocatalytic reduction on theelectrodes modified with native HRP and was negligiblewhen compared with that observed using electrodesmodified with the recombinant forms of the enzyme.

Calibration curves, i.e. dependence of the currentresponse on the concentration of H2O2, for electrodesmodified with native and recombinant peroxidases arepresented in Fig. 4. CHisrec-HRP and NHisrec-HRPgave the highest response with a current density of110–113 �A cm−2 for 10−4 M H2O2 at −50 mV dueto direct ET. Rec-HRP gave a lower current response(85 �A cm−2). These values are 8–11 times higher thanthat for native HRP (Table 3). The standard deviationfor the measured current densities was within 10%.Since the measured surface concentrations of the re-combinant forms under study are close (Table 1) andonly 30% higher than that for the native form, the

higher bioelectrocatalytic activity of the recombinantHRPs (and as a result the sensitivity of the electrodes)in the reaction of H2O2 reduction is connected with amore favorable arrangement of the enzyme moleculeson the surface of the gold electrode for establishingdirect ET in the enzyme-electrode system.

As rec-HRP modified electrodes exhibited lower cur-rent response for hydrogen peroxide reduction it can besupposed that the adsorption of CHisrec-HRP andNHisrec-HRP occurs not only due to hydrophobic inter-actions between gold and HRP molecules but throughthe histidine tags. Histidine adsorbs irreversibly ongold, has the potential of maximal adsorption at pH 6.0in the region of −0.25– +0.25 V versus the standardhydrogen electrode (approximately −0.45– +0.05 ver-sus Ag�AgCl electrode) (Safronov et al., 1983;Khudyakova et al., 1997) and was specially introducedin non-glycosylated rec-HRP to improve the adsorptionprocess. As in neutral solutions a maximum coverage isachieved at those potentials that are close to the pointof zero charge, then the adsorption of His may beconnected with His chemisorption at the gold electrodesurface through the nitrogen atom of the imidazol ring(Khudyakova et al., 1997). The same adsorption mech-anism may be proposed for rec-HRP with histidinegroups at the C- and N-termini.

The electrodes modified with recombinant HRPs ex-hibit high sensitivity for H2O2, determined from theinitial slopes of the calibration curves (see insert in Fig.4). The theoretical maximal sensitivity of an HRPmodified electrode (Ruzgas et al., 1996) defined by thevariation of the current response with the variation ofthe substrate concentration, is limited by the masstransfer of peroxide to the electrode surface and can beestimated using the general relationship between thecurrent density, j, and the bulk concentration of H2O2,c,

j=nFkDc (7)

where n=2 is the number of the electrons participatingin the reaction, F is the Faraday constant, and kD is themass transfer constant. From Eq. (7) it is clear that theupper limit for the sensitivity for H2O2 is dependent onthe mass transfer constant, and the approximate esti-mated of the maximal sensitivity could be close to 1 AM−1 cm−2 (Ruzgas et al., 1996). The obtained valuesof the sensitivity for H2O2 for electrodes modified with

Table 2desorption of native and CHisrec-HRP at the gold PZ crystal electrodes, surface pretreatment with procedure 2

Peroxidase type HRP surface concentration HRP surface concentration HRP amount after desorption (pmolesFrequency shift(ng cm−2) cm−2)(Hz) (pmoles cm−2)

0Native HRP 10.7�0.5471�23−107�510.1�0.421�1.7715�57CHisrec-HRP −162�13


Fig. 4. Dependence of the steady-state current density on the H2O2 concentration determined with gold disk electrodes modified with (1) nativeHRP; (2) rec-HRP; (3) NHisrec-HRP; and (4) CHisrec-HRP. The electrodes were placed into a wall-jet cell: flow rate of the carrier (PBS containing0.15 M NaCl, pH 7.4) 0.9 ml min−1. Applied potential was −50 mV versus Ag�AgCl in 0.1 M KCl.

Table 3The activity of peroxidases and characteristics of peroxidase-modified electrodesa

Activity towards ABTS (UHRP type Sensitivity (A cm−2 M−1), Sensitivity with catechol (A cm−2jmax with 10−4 M H2O2

M−1)(�A cm−2)direct ETmg−1)

0.17�0.02 10.20�1.12Native 1.25�0.101500�651.36�0.14 84.89�8.341400�94 1.36�0.16rec-HRP

CHisrec-HRP 1580�60 1.41�0.12 113.19�9.62 1.41�0.131.47�0.15 110.36�10.94 1.47�0.15950�40NHisrec-HRP

a Measurements at −50 mV versus Ag�AgCl electrode in a flow of PBS pH 7.4 flow, rate 900 �l min−1.

recombinant HRPs are presented in Table 3. It can beseen that the sensitivity values are close to 1.4 A M−1

cm−2. The sensitivity for the gold electrode modifiedwith recombinant forms of HRP is about 8–9 timeshigher than that for native HRP and is the best amongthe existing HRP modified electrodes described in theliterature based on direct ET (Kulys and Schmid, 1990;Csoregi et al., 1993) and the best obtained so far forgold electrodes.

The sensitivity of gold electrodes modified with per-

oxidases under study increases in the presence of asaturating concentration of catechol (5×10−4 M) onlyfor native HRP modified electrodes (Table 3). In thepresence of catechol acting as a mediator in bioelectro-chemical reduction of H2O2 the increasing responsecurrent is observed at the rec-HRP and histidine HRPmodified electrodes only for concentrations higher than10−4 M H2O2 (101 and 125 �A cm−2 for 10−4 MH2O2 at −50 mV, respectively). In contrast, for nativeHRP modified electrodes mediated ET is more efficient


than direct one for all the involved concentrations ofH2O2 (25 �A cm−2 at −50 mV for 10−4 M H2O2). Itwas earlier assumed (Ruzgas et al., 1995; Lindgren etal., 1998) that in the presence of a saturating concentra-tion of catechol, the total amount of the enzymemolecules adsorbed on the gold surface should be ac-tive in the electrode process. As there was no increasingcurrent response of the electrodes modified with therecombinant forms of HRP observed in the presence ofcatechol up to 100 �M, it can be considered that eitherthe total amount of the adsorbed enzyme is active indirect ET or the efficiency of direct ET is so high thatthe model put forward in (Ruzgas et al., 1995; Lindgrenet al., 1998) is no longer valid. This fact implies that therates of direct ET between recombinant peroxidasesand gold are so high, that the current response forH2O2 is determined only by the mass transfer of H2O2

to the electrode surface. The introduction of the media-tor in the system results in an increase of the responsecurrent over the entire investigated concentration rangeof H2O2 only for the native HRP modified electrodesdue to the fact that the rate of direct ET is low betweenadsorbed native HRP and gold. In this case, in thepresence of the mediator a 2–7-fold increase in thecurrent response is registered, depending on the concen-tration of hydrogen peroxide.

The lower detection limit of the biosensor based onthe recombinant and native forms of HRP was deter-mined to be 10−8 M H2O2. The detection limit wasdetermined as the value of the H2O2 concentration thatgave a signal in the region of the linear dependence ofcalibration plots. In fact, the lowest H2O2 concentrationthat gave a current response corresponding to a signalto noise ratio of three, was below 10−9 M H2O2 for therecombinant forms. Considering other methods forH2O2 detection, among the most sensitive can be men-tioned, for comparison, the determination of H2O2 byflow injection analysis in homogeneous phase based onenhanced chemiluminescent reaction (detection limit forH2O2 is 10−6 M) (Eremin et al., 1989), chemilumines-cent biosensors based on porous membrane supportwith immobilized HRP (detection limit for H2O2 is10−6 M) (Rubtsova et al., 1998). Among biosensorelectrodes for H2O2 determination the most sensitiveare redox mediator modified graphite electrodes (detec-tion limit for H2O2 is 7×10−9 M) (Wendzinski et al.,1997).

Electrodes modified with the recombinant forms ofHRP exhibited good stability compared with electrodesmodified with native HRP operated daily for 2 h andstored in buffer at 4°C between days. As can be seenfrom Fig. 5a, for CHisHRP the current signal remainedthe same during the first 24 h of storage, similar datawere obtained for NHisHRP modified electrodes. Thehalf-life time for both histidine HRPs was 72–96 h,that for rec-HRP in the order of 40–72 h, depending on

the H2O2 concentration (Fig. 5). For H2O2 concentra-tions lower than 1 �M 100% stability of the signal wasretained, in fact, during the initial 24 h of testing for allthe recombinant forms. Electrodes modified with nativeHRP demonstrated constantly decreasing sensitivity asa result of desorption of the enzyme (see curve 7 in theinsert of Fig. 5a). The conclusion is based on theexperimental fact that the signal was constantly de-creasing also in the presence of catechol in solution.

In summary it can be stated that the highest currentresponses to H2O2 (see Fig. 5 and Table 3) were ob-tained for gold electrodes modified with the recombi-nant forms of HRP. The sensitivity values are muchhigher than those for native HRP immobilized ontogold. Taking into account that the total amount of thenative and recombinant enzyme adsorbed on the sur-face of the electrode is similar (the microbalance mea-surements), the highest current response to H2O2 ofgold electrodes modified with the recombinant HRPsresults, in fact, from the high rates of direct ET. Theresults thus undoubtedly demonstrate that deglycosyla-tion of the native enzyme favors direct ET between thesurface of gold electrode and the active site of HRP. Asthe rec-HRP modified electrodes exhibit lower currentresponse for H2O2 and lower stability in comparisonwith the electrodes modified with the histidine forms, itcan be assumed that the higher direct ET rates in thelast case are connected also with the strong binding ofthe histidine residues with the electrode surface. Thus,the high direct ET rates can be considered as a result ofa shorter ET distance and/or correct orientation of theenzyme on the surface. The specific impact of these tworeasons onto the overall direct ET process remains stillundetermined. Further experiments are to be made withrotating disk electrodes to obtain precise data on thereaction rates and to estimate the fraction of adsorbedperoxidase active in direct ET (Ferapontova et al.,2001).

4. Conclusion

Different forms of HRP, native, rec-HRP, CHisrec-HRP and NHisrec-HRP, were directly adsorbed on baregold surfaces and the electrodes were studied as amper-ometric enzyme electrodes for the detection of hydro-gen peroxide. It was shown that the gold electrodes,modified with the recombinant forms of HRP, exhib-ited high and stable during 24-h current response toH2O2 biocatalytic reduction. The response was due todirect ET between HRP and the gold electrode. Dataon microbalance adsorption gave strong evidence thatthe high current responses of the gold electrodes,modified with recombinant HRPs, are due to strongbinding of these HRP forms with the electrode surface,


Fig. 5. Stability of the current signal with time for gold electrodes modified with (a) CHisrec-HRP and (b) rec-HRP. The concentration of H2O2:(1) 0.1, (2) 1, (3) 10, (4) 20, (5) 40, and (6) 100 �M. In the insert data normalized to the initial current response. The insert of (a) shows data fornative HRP in the concentration range of H2O2 from 10−5 to 4×10−5 M (curve 7). Other conditions as in Fig. 4.


especially in the case of the histidine forms. This fa-vored direct ET between the enzyme molecules and thesurface of gold. Since gold electrodes modified withrecombinant HRPs exhibit high stability and high sensi-tivity to H2O2 the recombinant peroxidases may beconsidered to be promising for the mediatorless biosen-sor development.

Acknowledgements

This work was supported by the European Commis-sion (grant No. ERB IC15-CT96-1002). A.E. acknowl-edges funding by grant No. 4-59 of the Russian StateScientific and Technological Program �AdvancedMethods of Bioengineering� and L.G.-funding by theSwedish Natural Science Research Council (NFR).

References

Baker, C.K., Reynolds, J.R., 1988. A quartz microbalance study of theelectrosynthesis of polypyrrole. J. Electroanal. Chem. 251, 307–322.

Bogdanovskaya, V.A., 1993. Bioelectrocatalysis: problems and per-spectives. Elektrokhimiya 29, 441–447.

Bogdanovskaya, V.A., Tarasevich, M.R., Hintsche, R., Scheller, F.,1988. Electrochemical transformations of proteins adsorbed atcarbon electrodes. Bioelectrochem. Bioenerg. 19, 581–584.

Csoregi, E., Jonsson-Pettersson, G., Gorton, L., 1993. Mediatorlesselectrocatalytic reduction of hydrogen peroxide at graphite elec-trodes chemically modified with peroxidases. J. Biotechnol. 30,315–337.

Deakin, M.R., Li, T.L., Melroy, O.R., 1988. A study of the electrosorp-tion of bromide and iodide ions on gold using the quartz crystalmicrobalance. J. Electroanal. Chem. 243, 343–351.

Dunford, H.B., 1999. Heme Peroxidase Nomenclature. Plant Perox-idase Newslett. 13, 65–74.

Durliat, H., Courteix, A., Comtat, M., 1989. Reactions of horseradishperoxidase on a platinum cathode. Bioelectrochem. Bioenerg. 22,197–209.

Egorov, A.M., Gazaryan, I.G., Kim, B.B., Doseeva, V.V., Kapeliuch,J.L., Veryovkin, A.N., Fechina, V.A., 1994. Horseradish peroxidaseisozyme C. A comparative study of native and recombinant enzymeproduced by E. coli transformations. Ann. NY Acad. Sci. 721,73–81.

Eremin, S.A., Vlasenko, S.B., Osipov, A.P., Eremina, I.D., Egorov,A.M., 1989. Determination of hydrogen peroxide concentrations byflow injection analysis based on the enhanced chemiluminescentreaction using peroxidase. Anal. Lett. 22 (9), 2037–2050.

Ferapontova, E.E., Grigorenko, V.G., Egorov, A.M., Borcher, T.,Ruzgas, T., Gorton, L., 2001. Direct electron transfer in the systemgold electrode-recombinant horseardish peroxidases. J. Elec-troanal. Chem,. In press.

Gajhede, M., Schuller, D.J., Henriksen, A., Smith, A.T., Poulos, T.L.,1997. Crystal structure of horesradish peroxidase C at 2.15 Aresolution. Nat. Struct. Biol. 4 (12), 1032–1038.

Grigorenko, V., Chubar, T., Kapeliuch, Y., Borchers, T., Spener, F.,Egorov, A., 1999. New approaches for functional expression ofrecombinant horseradish peroxidase c in Escherichia coli. Biocatal-ysis and Biotransformation 17 (5), 359–379.

Jonsson, G., Gorton, L., 1989. An electrochemical sensor for hydrogen

peroxide based on peroxidase adsorbed on a spectrographicgraphite electrode. Electroanalysis 1, 465–468.

Kanazawa, K.K., Gordon, J.G. II, 1985. Frequency of a quartzMicrobalance in contact with liquid. Anal. Chem. 57, 1770–1771.

Koh, W., Kutner, W., Jones, M.T., Kadish, K.M., 1993. An improvedholder for the electrochemical quartz crystal microbalance and itscyclic voltammetry characteristics. Electroanalysis 5, 209–214.

Khudyakova, R.V., Soloshko, S.V., Safronov, A.Yu., 1997. Electro-chemical behavior of histidine on gold. Elektrokhimiya 33, 1165–1171.

Kulys, J., Schmid, R.D., 1990. Mediatorless peroxidase electrode andpreparation of bienzyme sensors. Bioelectrochem. Bioenerg. 24,305–311.

Lindgren, A., Munteanu, F., Gazaryan, I., Ruzgas, T., Gorton, L.,1998. Comparison of rotating disk electrode and wall-jet electrodesystems for studying the kinetics of direct and mediated electrontransfer for horseradish peroxidase on a graphite electrode. J.Electroanal. Chem. 458, 113–120.

Lindgren, A., Tanaka, M., Ruzgas, T., Gorton, L., Gazaryan, I.,Ishimori, K., Morishima, I., 1999. Direct electron transfer catalysedby recombinant forms of horseradish peroxidase: insight into themechanism. Electrochem. Commun. 1, 171–175.

Presnova, G., Egorov, A., Ruzgas, T., Lindgren, A., Gorton, L.,Borchers, T., 2000. Direct heterogeneous electron transfer ofrecombinant horseradish peroxidase on gold. Faraday Discuss. 116,281–289.

Razumas, V., Jasaitis, J., Kulys, J., 1984. Electrocatalysis on enzyme-modified carbon materials. Bioelectrochem. Bioenerg. 12, 297–322.

Razumas, V.J., Gudavicius, A.V., Kulys, J.J., 1986. Kinetics ofperoxidase redox conversion on viologen-modified gold electrodes.J. Electroanal. Chem. 198, 81–87.

Rubtsova, M., Kovba, G., Egorov, A., 1998. Chemiluminescentbiosensors based on porous supports with immobilized peroxidase.Biosens. Bioelectron. 13, 75–85.

Ruzgas, T., Gorton, L., Emneus, J., Marko-Varga, G., 1995. Kineticmodels of horseradish peroxidase action on a graphite electrode.J. Electroanal. Chem. 391, 41–49.

Ruzgas, T., Csoregi, E., Emneus, J., Gorton, L., Marko-Varga, G.,1996. Peroxidase-modified electrodes: fundamentals and applica-tion. Anal. Chim. Acta 330, 123–138.

Safronov, A.Yu., Tarasevich, M.R., Bogdanovskaya, V.A., Chernyak,A.S., 1983. Electrooxidation of glycilglycine, cystein and histidineon gold electrode. Elektrokhimiya 19, 421–424.

Sauerbrey, G.Z., 1959. Use of quartz crystal vibrator for wieghting thinfilms on a microbalance. Z. Physik. 155, 206–210.

Varineau, P.T., Buttry, D.A., 1987. Applications of the quartz crystalmicrobalance to electrochemistry. Measurement of ion and solventpopulations in thin films of poly(vinylferrocene) as functions ofredox state. J. Phys. Chem. 91, 1292–1295.

Ward, M.D., 1989. Probing electrocrystallization of charge transfersalts with the quartz crystal microbalance. J. Electroanal. Chem.273, 79–105.

Ward, M.D., Buttry, D.A., 1990. In situ interfacial mass detection withpiezoelectric transducers. Science 249, 1000–1007.

Welinder, K.G., 1992. Superfamily of plant, fungal and bacterialperoxidases. Curr. Opin. Struc. Biol. 2, 388–393.

Wendzinski, F., Grundig, B., Renneberg, R., Spener, F., 1997. Highlysensitive determination of hydrogen peroxide and peroxidase withtetrathiafulvalene-based electrodes and the application in im-munosensing. Biosens. Bioelectron. 12, 43–52.

Yaropolov, A.I., Tarasevich, M.R., Varfolomeev, S.D., 1978. Electro-chemical properties of peroxidase. Bioelectrochem. Bioenerg. 5,18–24.

Zhao, J., Henkens, R.W., Stonehuerner, J., O’Daly, J.P., Crumbliss,A.L., 1992. Direct electron transfer at horseradish peroxidase-col-loidal gold modified electrode. J. Electroanal. Chem. 327, 109–119.

.

Mediatorless biosensor for H2O2 based on recombinant forms of horseradish peroxidase directly...

Documents

Transcript of Mediatorless biosensor for H2O2 based on recombinant forms of horseradish peroxidase directly...