Biosensors based on novel plant peroxidases: a comparative study

Electrochimica Acta 46 (2000) 255– 264 www.elsevier.nl/locate/electacta

Biosensors based on novel plant peroxidases: a comparative study Szilveszter Gaspar a,1, Ionel Catalin Popescu1 a, Irina G. Gazaryan b, A. Gerardo Bautista c, Ivan Yu. Sakharov c,2, Bo Mattiasson a, Elisabeth Cso¨regi a,* b

a Lund Uni6ersity, Department of Biotechnology, P.O. Box 124, SE-22100 Lund, Sweden Moscow State Uni6ersity, Chemical Faculty, Department of Chemical Enzymology, Moscow 119899, Russia c Santander Industrial Uni6ersity, School of Chemistry, Bucaramanga, A.A.678, Colombia

Received 6 September 1999; received in revised form 1 February 2000

Abstract Amperometric biosensors for hydrogen peroxide detection have been constructed using horseradish peroxidase (HRP) and two newly purified peroxidases extracted from tobacco (TOP) and sweet potato (SPP). The peroxidases were cross-linked to a redox polymer [poly(vinylimidazole) complexed with Os(4,4%dimethylbipyridine)2Cl] using poly(ethylene glycol) diglycidyl ether as the cross-linker. A comparative study with regard to their bioelectrochemical characteristics showed that, irrespective of peroxidase, the biosensors sensitivity was strongly influenced by hydrogel composition, curing procedure, film thickness and applied potential. The electrostatic interaction between the cationic redox polymer and the negatively charged peroxidases (TOP and SPP) enhanced the hydrogen peroxide signal. When operated in a FI system, the optimized SPP biosensor (48% redox polymer, 23% cross-linker and 29% enzyme, w/w %) displayed the highest sensitivity for H2O2 (3.2 A M − 1cm − 2), a linear range up to 220 mM, a detection limit of 25 nM (calculated as 2S/N) and a response time of about 2 min. © 2000 Elsevier Science Ltd. All rights reserved. Keywords: Hydrogen peroxide; Tobacco and sweet potato peroxidases; Osmium redox polymer; Redox hydrogel; Amperometric biosensor; Flow injection

1. Introduction Among the systems employing enzyme-catalyzed electron transfer, generic called ‘molecular transducers’ [1], those using peroxidases are deserving increased interest, justified, at least, by the following reasons: (i) * Corresponding author. Tel.: +46-46-2224274; fax: + 4646-2224713. E-mail address: elisabeth.cso¨[email protected] (E. Cso¨regi). 1 On leave from University ‘‘Babes-Bolyai’’, Department of Physical Chemistry, RO-3400 Cluj-Napoca, Romania 2 Moscow State University, Chemical Faculty, Department of Chemical Enzymology, Moscow 119899, Russia

the hydrogen peroxide itself is an important analyte; (ii) the detection of H2O2 is used in more than 65% of all biosensor designs [2]; and (iii) HRP is one of the most often used oxidoreductase in the field of bioelectrocatalysis applied for biosensor development [3]. HRP has a very broad specificity for reducing substrates [4] and although it exhibits a relatively good stability at room temperature, displays an unsatisfactory one at more elevated temperatures [5]. These drawbacks were highly motivating the search for new, more selective, sensitive and thermostable peroxidases (POD). Thus, biosensors with microperoxidase [6 – 9], cytochrome c peroxidase [10], fungal peroxidase from Arthromyces ramosus (ARP) [11– 13], peroxidase from

0013-4686/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved. PII: S 0 0 1 3 - 4 6 8 6 ( 0 0 ) 0 0 5 8 0 - 6

256

S. Gaspar et al. / Electrochimica Acta 46 (2000) 255–264

bovine milk [6,7,13], soybean peroxidase [5,13–15], chloroperoxidase [16] and more recently, tobacco peroxidase and peanut peroxidase [17] have been already reported. Most of these peroxidase modified electrodes are based on a direct electron transfer (DET) between the active site of the enzyme and an electrode [1,18,19], but electrode configurations based on various mediated electron transfer (MET) pathways using either diffusional mediators [3,19,20], immobilized molecular redox couples [3,20] or redox polymers [5,13,14,21] have been also described. However, enzyme electrodes based on various redox enzymes entrapped in redox hydrogels were previously shown to display high sensitivities and improved stabilities [22–25]. Therefore, the present work targeted the development of H2O2 sensing bioelectrodes cross-linking a hydrophilic redox polymer, [poly(vinylimidazole) complexed with Os(4,4%dimethylbipyridine)2Cl (PVI7dme-Os)], with two newly purified peroxidases, tobacco peroxidase TOP and sweet potato peroxidase SPP, using poly(ethylene glycol) diglycidyl ether (PEGDGE) as the cross-linker. The hydrogel-modified graphite electrodes were optimized with regard to the curing procedure, the hydrogel’s composition and its thickness. The optimized biosensors were thoroughly characterized considering their reproducibility, sensitivity, detection limit, apparent Michaelis constant K app M , response time, storage and operational stability. An overall comparison considering their bioelectrochemical characteristics, relative to those corresponding to similarly build HRP based ones, as well as to those containing the two new enzymes, but based on a DET approach, was performed, in order to select the most efficient molecular transducer design.

published protocol [27]. The subscript 7 indicates the number of vinylimidazole units per vinylimidazole unit complexed to an osmium centre. Phosphate buffer (PB) solution was prepared using Na2HPO4.2H2O and NaH2PO4.H2O from Merck (Darmstadt, Germany). Hydrogen peroxide standard solutions were prepared daily from 35% H2O2 solution (Cat. no. 20246-0010) purchased from Acros Organics (Geel, Belgium). Its concentration was checked spectrophotometrically (molar extinction coefficient at 230 nm of 72.7 M − 1cm − 1 [28]). All solutions were prepared using tridistilled water produced in a Milli-Q system (Millipore, Bedford, MA) if not otherwise stated.

2.2. Methods 2.2.1. Sweet potato peroxidase purification The peel (200 g) of tubers of sweet potato (Imopotea batatas) was homogenized in 0.1 M phosphate buffer, pH 7.0, containing 28 mM ascorbic acid, 5 mM EDTA and 5% NaCl and incubated at ambient temperature over night. After filtration, the extract was treated with (NH4)2SO4 (85% saturation). The obtained precipitate was separated by centrifugation and dissolved in water. Then, solid (NH4)2SO4 was added to a final concentration of 1.3 M. The solution was applied on a PhenylSepharose column (Pharmacia Fine Chemicals, Uppsala, Sweden) (1.5 × 18 cm) equilibrated with 0.1 M phosphate buffer, pH 6.2, containing 1.3 M (NH4)2SO4. The enzyme was eluted by decreasing the concentration of the salt. Active fractions were dialyzed against 5 mM Tris-HCl, pH 8.3, and applied on a DEAE-Toyopearl 650M column (Toyo Soda MFG Co. Ltd., Tokyo, Japan) (0.9 × 9 cm) equilibrated with 5 mM Tris-HCl buffer, pH 8.3. The enzyme was eluted with a 0 to 100 mM gradient of NaCl. Active fractions were concentrated by ultrafiltration and stored at 4°C.

2. Experimental

2.1. Chemicals Horseradish peroxidase with specific activity of 1100 U/mg solid (HRP; EC 1.11.1.7, Cat no. P6782), bovine serum albumin (BSA, Cat no. A4503), Triton-X100 (Cat. no. X100) and 2,2%-azino-bis(ethylbenzothiazoline-6-sulfonic acid) (ABTS, Cat. no. A1888) were purchased from Sigma-Aldrich (Tyreso¨, Sweden). Tobacco peroxidase was isolated from Nicotiana syl6estris as described earlier [26]. Sweet potato peroxidase was purified from sweet potato peels using hydrophobic and anion-exchange chromatography as briefly described under Methods. Poly(ethylene glycol) (400) diglycidyl ether (PEGDGE, Cat. no. 08210) was purchased from Polysciences, (Warrington, PA). Poly(vinylimidazole) complexed with Os(4,4%dimethylbipyridine)2Cl (PVI7dme-Os) was synthesized according to a previously

2.2.2. Electrode preparation Graphite electrodes (i.d. 0.305 cm, type RW 001, Ringsdorff Werke, Bonn, Germany) were wet polished on emery paper (Tufbak Durite type P1200, Allar, Sterling Heights, MI) and thoroughly washed with deionized water, before modification. DET based, type I electrodes were prepared by placing different volumes of stock peroxidase solutions on the electrode surface, all electrodes being modified by the same amount of enzyme (20 mg). Stock solutions of the studied peroxidases were 4.78 mg ml − 1 of TOP in 0.15 M Tris-HCl, pH 6.0; 5 mg ml − 1 of HRP in 0.1 M PB pH 7.0; 1.86 mg ml − 1 of SPP in 0.005 M Tris-HCl, pH 8.1, respectively. Finally, modified electrodes were cured at room temperature for 2.5 h. MET based electrodes were prepared as follows: first, different volumes of the investigated peroxidase solutions, redox polymer and cross-linker were mixed and

S. Gaspar et al. / Electrochimica Acta 46 (2000) 255–264

placed on the top of the polished graphite. The premixed hydrogels were made using the same peroxidase stock solutions (see type I electrodes), 3.3 mg ml − 1 PVI7-dme-Os and 2.5 mg ml − 1 PEGDGE (freshly prepared and used within 15 min). Next, the electrodes were cured in three different ways: “ type II, at room temperature for 20 h in desiccated conditions (in a beaker covered with a perfored sealing film). “ type III, under the same conditions as before, but the beaker not being covered. “ type IV, at room temperature for 25 min, followed by a curing at 50°C for 10 min. Prior to use, all modified electrodes were thoroughly washed with deionized water. All presented results are the mean of at least three identically prepared electrodes, if not otherwise mentioned.

2.3. Enzyme acti6ity assay The specific activities of the studied peroxidases were determined using the procedure recommended by Sigma for HRP [29]. The reaction mixture contained 96 mM potassium phosphate, 8.9 mM ABTS, 0.1% (w/w) H2O2, 0.004% (w/v) BSA, 0.008% (v/v) Triton X-100

257

and 0.01– 0.04 unit peroxidase. The millimolar absorption of ABTS oxidation product at 405 nm was taken equal to 36.8 mM − 1cm − 1. The experiments were run at 25°C using a light path of 1 cm. The specific activities observed were: TOP 6170 9 56 U/mg solid, SPP 4174939 U/mg solid and HRP 1226 95 U/mg solid, the latter value being in a good agreement with its stated value (1100 U/mg solid).

2.4. Instrumentation Cyclic voltammetry experiments were performed using a BAS potentiostat (Model CV-50W, Bioanalytical Systems, West Lafayette, IN) and a three-electrode cell, having a Pt wire as the counter and a Ag/AgCl, 0.1 M KCl as the reference electrode, respectively. Amperometric measurements were done in a home built flow through wall-jet cell [30] inserted in a single line flow injection (FI) system. The experimental set-up also contained a peristaltic pump (type U4-MIDI, Alitea AB, Stockholm, Sweden) and a 100 mL manual injector (Vici AG-Valco Europe, Switzerland). The output signal was recorded on a strip chart recorder (Model BD 111, Kipp and Zonen, Delft, The Netherlands). All FI experiments were done at −50 mV vs. Ag/AgCl (0.1 M KCl) using a low current potentiostat (Za¨ta-Elektronik, Ho¨o¨r, Sweden). Operational stability experiments were performed with an Automated Sample Injection Analyzer (Ismatec, Glattburg– Zu¨rich, Switzerland) by injecting 50 ml of 20 mM H2O2, with a sample throughput of 25 injections h − 1. Throughout this work a deaerated 0.1 M PB solution at pH 7.0 was used as the supporting electrolyte and all experiments were carried out at room temperature.

3. Results and discussion

Fig. 1. Mechanism of direct (A) and mediated (B) bioelectrocatalytic reduction of hydrogen peroxide at peroxidase modified electrodes. P+ is a cation localised on the porphyrin ring or the polypeptide chain. OsII and OsIII are the reduced and oxidised forms of the Os mediator, respectively.

An amperometric biosensor based on immobilized PODs uses the electrical communication between the enzyme’s redox center and the electrode via a direct or a mediated electron transfer pathway. In the first one (DET, see Fig. 1A) the enzyme redox center (oxidized by H2O2) accepts electrons directly from the electrode, which substitutes the electron donor substrates. This approach has the advantage of being simple and to be operable at a potential closer to the enzyme’s redox potential [31]. However, a proper orientation of the protein at the electrode surface is extremely important [19,32], electrostatic interactions influencing this process. Thus, the isoelectric point, determined to be 7.2 for HRP [33] and 3.5 for TOP [26], and SPP [34], is an important enzyme parameter. Glycosylation is another key characteristic, since it can hinder a DET due to the increased distance between an electrode and the heme cofactor [35]. The access to the redox center was ob-

258

S. Gaspar et al. / Electrochimica Acta 46 (2000) 255–264

served to be blocked (due to the presence of the negatively charged Glu-141 at the entrance to the active center) by Ca2 + ions in the particular case of TOP [36]. The observation of this blockage in the case of TOP adsorbed on the graphite surface indicates a conformational flexibility of the immobilized enzyme [37]. However, slow electron transfer kinetics represents the main disadvantage of DET approach. In a mediated electron transfer (see Fig. 1B) a supplementary redox couple shuttles the electrons from/to electrode to/from the oxidized forms of the peroxidase (compounds I and II). The main advantage of the mediated electron transfer is a more effective electron transfer mechanism, and especially redox hydrogels have been shown to be very advantageous [22–25]. In these hydrogels the enzyme is electrically wired to the electrode surface by the three-dimensional, redox polymer network. The hydrogels assure high permeability, high sensitivities for many substrates and improved stability in an environment with high water content [22,23]. Among different redox centers used in hydrogels, osmium complexes have been proved to be very successful [38]. The redox polymer has to complex the enzyme, electrically connect the redox center of the enzyme to the electrode and physically attach the enzyme to the electrode surface [23]. Due to electrostatic interactions between the polycationic redox polymer and the anionic TOP and partly to the cross-linking reaction, it was expected that the above-mentioned conformational flexibility observed for the immobilized enzyme in a DET approach will be limited when using a hydrogel-based electrode design. Indeed, injecting Ca2 + ions in the presence of hydrogen peroxide no current decrease (due to the blocking of the TOP’s active center) was observed between 5–500 mM Ca2 + . Cyclic voltammograms recorded for HRP, TOP and SPP modified type II electrodes, showed a pair of peaks corresponding to the Os redox couple with a formal potential of 158 96 mV vs. Ag/AgCl (0.1 M KCl). This formal potential is somewhat higher than that previously reported for a glucose oxidase (GOx) and PVI15dme-Os based hydrogel (95 mV vs. SCE [27]). However, a true comparison is difficult, since different electrode materials, redox polymers with different Os loading, different supporting electrolyte concentration and hydrogels of different structure were used. It also has been previously observed, that increasing the concentration of the cross-linker in a film without enzyme, the apparent standard potential of the osmium couple shifted to a more positive value [39]. Moreover, perturbations of the standard potential of redox ions bound to charged polymers are described in the literature; e.g. the potential of Fe(CN)36 − /4 − system is shifted with a few hundreds of mVs towards negative values when the redox couple is entrapped in polymeric films (alkyl ammonium and pyridinium containing polypyrrole and alkyl-

aminesiloxane polymers, respectively [40– 42]). These perturbations are attributed to electrostatic and charge transfer interactions. Furthermore, isoelectric focusing experiments also proved the formation of complexes between the polycationic redox polymer and enzymes [23]. In present work, the newly purified peroxidases have an isoelectric point of 3.5 being negatively charged at the working pH of 7.0, while HRP is slightly positive at this pH. Accordingly, one could expect that these bulky, charged biomolecules could have a similar influence on the formal potential of Os3 + /2 + redox couple as the above mentioned polymeric matrixes. However, cyclic voltammetry measurements did not reveal any difference between the redox responses of hydrogels containing the positively charged HRP and the negatively charged TOP or SPP, respectively. The current signal recorded for hydrogels is known to be greatly influenced by two parameters: the redox polymer/enzyme ratio and the flexibility of the created redox hydrogel (determined by the amount of crosslinker). Therefore, the optimisation strategy of MET, Type II electrodes started considering first the enzyme/ polymer ratio, since this is the main factor determining the efficiency of the electron transfer process. Next, the amount of cross-linker was optimised in order to find the redox hydrogel of appropriate flexibility, thus assuring an efficient transport of substrate, product and electrons through the hydrogel and an adequate stability of the film. Practically, first the redox polymer/enzyme ratio was varied at constant cross-linker content followed by varying the cross-linker content at constant redox polymer/enzyme ratio kept at its optimal value. Finally, the electrodes were also optimised with regard to hydrogel’s thickness and its curing procedure.

3.1. Redox polymer/enzyme ratio Hydrogels with different polymer/enzyme ratio but constant amount of cross-linker have been prepared and used to obtain bioelectrodes with constant hydrogel loading (22 mg). The hydrogels gave maximum current responses for 15 – 25% TOP, 20 – 30% HRP and 30 – 40% SPP enzyme content (see Fig. 2), well in agreement with the reported values of 20 – 45% for HRP and ARP, 35 – 50% for lactoperoxidase and 40 – 60% for soybean peroxidase [13,14]. Higher enzyme contents start to disturb the electron relaying capacity of the films due to the presence of the insulating, bulky enzyme molecules [43,44]. In the rising part of the curve, the current was limited by the enzyme activity in the film (the number of catalytic sites), while in the declining part it was limited by the rate of electron transfer to and/or through the polymer in a similar way as previously observed for other enzymes entrapped in hydrogels [13,24,25,27,43– 47]. It is thus, clear that the polymer/enzyme ratio has to be kept at a sufficiently

S. Gaspar et al. / Electrochimica Acta 46 (2000) 255–264

259

3.2. Cross-linker content

Fig. 2. Dependence of the relative sensitivity on the amount of TOP ( ), HRP ( ) and SPP (*), for electrodes prepared with 11.33% PEGDGE. Experimental conditions: 0.1 M phosphate buffer, pH 7.0; flow rate, 0.5 mL min − 1; applied potential, − 50 mV vs. Ag/AgCl (0.1 M KCl); hydrogel loading, 0.3 mg cm − 2.

high value to wire all of enzyme molecules and the capacity of the network to carry current must equal or exceed the capacity of the entrapped enzyme molecules to deliver electrons [23].

When the polymer/enzyme ratio has been kept at optimum values and the cross-linker PEGDGE content was varied in the range of 5 – 35 w/w %, for a constant hydrogel loading (22 mg), the current response first increased with increasing cross-linker contents to a maximum, followed by a consequent decrease (see Fig. 3). At increasing PEGDGE contents the current increase was attributed to a higher amount of crosslinked enzyme and a better mechanical stability of the film. At high amounts of cross-linker the sensitivity tends to decrease due to an increasing rigidity of the film. This phenomenon was most obvious for SPP, but HRP also showed a similar tendency. Our results suggest that both HRP and SPP might have more accessible redox centers and thereby the communication with the electrode surface can occur properly even in a less flexible redox polymer network. Hydrogels containing SPP and HRP displayed an optimum current response for 20 – 25% and 10 – 15% PEGDGE, respectively. The relatively high optimal cross linker content of the SPP hydrogel was surprising. Considering the optimum enzyme content of about 29% observed for these electrodes results in a small percentage of redox polymer inside this film (48%). This rather small amount of redox centers seems to be enough to wire most of the SPP molecules and the capacity of the network to carry current seem to equal the capacity of the incorporated enzyme molecules to deliver electrons. This observation supports additionally the assumption of a good accessibility of the redox center of SPP by the Os complexes.

3.3. Hydrogel thickness

Fig. 3. Dependence of the relative current response on the PEGDGE content for TOP ( ), HRP ( ) and SPP (*) electrodes with optimal enzyme/redox polymer ratio. Experimental conditions: 0.1 M phosphate buffer containing 1 mM H2O2, pH 7.0; other conditions as mentioned for Fig. 2.

Keeping the percentage of redox polymer, PODs, and PEGDGE in the film at previously determined optimal values the amount of applied hydrogel was increased from 68 to 945 mg/cm2. The catalytic efficiency of the bioelectrodes was expected to increase with increasing amounts of deposited redox hydrogel. This effect is, however, counteracted by limitations imposed by the rates of substrate diffusion and charge propagation across thick coatings [48]. Indeed, the H2O2 sensitivity of the POD electrodes first increased with increasing loading (see Fig. 4), followed by a decrease at a loading higher than 22 mg, similarly to earlier reported glucose and lactate sensors [27]. However, SPP showed a broader optimal range (22– 37 mg/electrode), indicating once again a better accessibility of its redox center, the recorded current signals being not as accentuatedly influenced by diffusional limitations as the ones recorded for TOP and HRP.

260

S. Gaspar et al. / Electrochimica Acta 46 (2000) 255–264

Fig. 4. Dependence of the relative sensitivity on the hydrogel loading for optimal TOP ( ), HRP ( ) and SPP (*) electrodes. For optimal electrode configuration see text, other experimental conditions as mentioned for Fig. 2.

3.4. Working potential The optimal working potential was established for the biosensors of optimal composition for 1 mM H2O2. As shown in Fig. 5, the bioelectrocatalytic reduction of H2O2 starts at slightly more positive potentials for TOP

Fig. 6. Influence of the hydrogel curing procedure on the relative current response of TOP, HRP and SPP electrodes. Electrodes were prepared equally, containing 22% enzyme, 67% PVI7-Os and 11% PEGDGE (w/w %), other conditions as in Fig. 2.

and SPP than for HRP, showing a leveling off tendency at 0, − 50 and around −100 mV for TOP, SPP and HRP, respectively. Cyclic voltammetry studies showed that the standard potential of the hydrogel did not change with the nature of the used POD. The different half-wave potential values of the different POD hydrogels observed in hydrodynamic voltammetry could be explained by the different nature of the responses recorded for the two compared cases. In hydrodynamic conditions, the H2O2 signal is the result of a bioelectrochemical reaction, where a good communication between the enzyme’s redox center and Os centers is determinant. Therefore, as mentioned above the reduction of H2O2 starts at more positive potentials for TOP and SPP electrodes, attributed to an attractive interaction between the negatively charged enzymes and the polycationic redox polymer [13,38,49]. Contrarily, the cyclic voltammetry response being electrochemically controlled was practically not influenced by the nature of the studied enzymes.

3.5. Curing procedures

Fig. 5. Hydrodynamic voltammograms recorded in a FI system for TOP ( ), HRP ( ) and SPP (*), electrodes of optimal configuration. Experimental conditions: 0.1 M phosphate buffer containing 2 mM H2O2, pH 7.0, other conditions as in Fig. 2.

Hydrogels containing different PODs but of same composition, were cured in different conditions in order to obtain reproducible and stable electrodes. Type II electrodes yielded in general higher current responses than type III ones (about twice higher for HRP and TOP, and 0.7 times higher for SPP). Type IV electrodes showed current responses 4 to 10 times smaller than type II (Fig. 6) ones. Obviously, the sensitivity of the biosensors is strongly influenced by the rate of the water evaporation and, taking into account that the cross-linking reaction takes place just after the solvent evaporation [39], the extent

S. Gaspar et al. / Electrochimica Acta 46 (2000) 255–264

of cross-linking also affects the sensor characteristics. Thus, the first hydrogel curing procedure (type II electrodes) assures a less extent of cross-linking and, consequently, a more flexible hydrogel than the other two procedures. Electron transfer in such kind of hydrogels is occurring by diffusion between the relaying redox centers along the wire and occasionally crossing between undulating segments of the wires [22,23]. On the other hand, these electrodes displayed unsatisfactory mechanical stability and worse reproducibility (see Fig. 6). The relative standard deviation for HRP hydrogels decreased from 16.5% (type II) to 7.1% and 1.6% for type III and IV electrodes, respectively. Unfortunately, type IV biosensors, displaying the best reproducibility, yielded current responses to H2O2 even smaller than DET-type I ones (TOP and HRP hydrogels). The different behavior observed for type I and type IV SPP electrodes was attributed to substrate inhibition observed for type I electrodes at hydrogen peroxide concentrations higher than 1–2 mM. However, SPP electrodes seem to be less affected by the increasing rigidity of the film, probably due to a higher accessibility of the redox centers, as previously mentioned. The low responses generally observed for type IV electrodes might be caused by either a bad communication between enzyme and electrode or a thermal denaturation of the PODs during the curing procedure. The 20% decrease in current response observed for TOP type I electrodes exposed at 50°C for 10 min, clearly

261

showed, that the low sensitivity obtained for TOP type IV electrodes is mostly due to the rigidity of the polymer matrix, and in a less extent to the denaturation of the enzyme.

3.6. Influence of electrostatic interactions Strong electrostatic interactions between the polymer and PODs result in a tight coupling of the enzyme and polymer, and thus, in a shorter average distance of the electron transfer [13,38,49]. Previous work showed that comparing the peroxidases from the fungus A. ramosus, horseradish and bovine milk, the first one yielded the highest currents which was attributed to the fact that the enzyme was more negatively charged than the other two ones at the working pH of 7.0 [13]. Also, using HRP oxidized by periodate, when the isolelectric point of the enzyme is lowered and thereby the electrostatic binding to the polycationic redox polymer enhanced, an improvement of the sensitivity was found relative to unoxidized HRP [5]. Comparing the currents obtained for the investigated PODs upon addition of 2 mM H2O2 (almost in the saturation domain), it seems that higher TOP and SPP amounts could be entrapped than HRP (excluding type II electrodes when HRP has given higher currents than SPP). This assumption is considering the pI values of the studied PODs (3.5 for TOP and SPP and 7.2 for HRP), the working pH (7.0) and the positively charged redox polymer. Accordingly, the

Table 1 Bioelectrochemical characteristics of peroxidase modified electrodes Electrode type

Enzyme

Imax (mA)a

a K app M (mM)

Type I

HRP

12.6 90.2

TOP

Type II

Type III

a

Sensitivity (mA mM−1)b

Linear range (mM)

5109 24

24.7 91.6

27.7 90.5

1316 961

21.09 1.3

SPP

5.89 0.2

1059 18

56 9 12

HRP

6791.6

399 926

169 915

TOP

9695.7

693 993

139 927

SPP

1309 3.7

7739 67

168 9 19

HRP

3491.8

615 9100

55 912

TOP

8894.9

2371 9264

37 96.2

SPP

9193.4

386 947

0.5–130 R =0.9939 0.5–240 R =0.9936 0.5–60 R =0.9998 0.5–70 R =0.9977 0.5–200 R =0.9991 0.5–100 R =0.9999 0.5–130 R =0.9981 0.5–470 R =0.9958 0.5–220 R =0.9988

236 938

Obtained by fitting the calibration curves with the Michaelis–Menten equation. Calculated as the ratio between Imax and K app M . c Estimated for a signal-to-noise ratio value of 2. b

Detection limit (nM)c

71 933 136 910 10.5 9 0.4 38 9 11 81 9 15 19 93.3 35 9 15 81 968 25 9 10

262

S. Gaspar et al. / Electrochimica Acta 46 (2000) 255–264

electrostatic interactions should be strongest for TOP and SPP. It can be concluded that type II electrodes assured the highest currents, but considering practical applications, where a good stability and reproducibility are also needed, type III ones might be more favorable.

3.7. Biosensor characteristics The optimized MET type electrodes as well as DET type electrodes have been characterized recording the current responses to H2O2 up to the saturation concentrations. Since these experiments involved high peroxide concentration the substrate inhibition should not be neglected. A time dependent, decreasing current due to substrate inhibition was previously reported for HRP bioelectrodes of similar design when the H2O2 concentration exceeded 0.25 mM [40]. Nevertheless, excepting SPP type I electrodes no substrate inhibition was noticed even for higher hydrogen peroxide concentrations (up to 5 mM).

The main bioelectrochemical parameters of type I – III bioelectrodes (maximum current, Imax, apparent Michaelis-Menten constant, K app M , sensitivity, linear range and detection limit) are presented in Table 1. The highest Imax values were observed for SPP and TOP type II electrodes, indicating a benefical electrostatic interaction between these enzymes and the redox polymer. However, irrespective of the curing procedure, SPP and HRP electrodes displayed higher sensitivities than TOP ones. The different biosensor designs showed a decreasing sensitivity tendency as follows: type II \ type III \type I. A remarkable exception was noticed for SPP type III electrodes, which yielded the highest sensitivity (3.2 90.5 AM − 1cm − 2). TOP based bioelectrodes exhibited the largest linear range, and for all PODs it increased with the increase of the hydrogel cross-linking degree (type III \ type II \ type I). The lowest detection limit, estimated as 2S/N, was found for SPP type I electrodes (see Table 1). However, all investigated electrodes are detecting submicromolar concentrations of H2O2. MET type electrodes showed not only higher sensitivities than DET based ones, but also shorter response times ( B 1 min), attributed both to an improved electron transfer kinetics and to the good permeability of hydrogels. Type II biosensors of optimal composition have been characterized with regard to storage and operational stability, the latter carried out by recording the current responses to repetitive injections of 20 mM H2O2, during a predetermined period of time (see Fig. 7A). After 6 h of continuous use, the current response decreased from its initial value with 20, 31 and 8% for HRP, TOP and SPP electrodes, respectively. The storage stability test, performed by recording discontinuously the bioelectrodes response for 5 – 6 injections of 20 mM H2O2 during one month, showed a current response decrease of 22, 38 and 19% for HRP, TOP and SPP electrodes, respectively (see Fig. 7B). The superior storage and operational stabilities observed for SPP based electrodes were attributed to the higher cross-linker content of the hydrogel (23%) compared with that used for HRP and TOP hydrogels (11%).

4. Conclusions

Fig. 7. Operational (A) and storage stability (B) of TOP ( ), HRP ( ) and SPP (*) electrodes of optimal composition. Experimental conditions: 0.1 M phosphate buffer containing 20 mM H2O2, injected volume 50 mL (A) and 100 mL (B); other conditions as in Fig. 2.

Wiring electrically the redox center of various peroxidases to the electrode surface has been proved again to be a very efficient molecular transducer design to construct sensitive H2O2 biosensors. The current response given by hydrogels containing HRP, TOP or SPP was highly dependent on the cross-linking degree of the three-dimensional redox matrix and on the accessibility of the enzyme redox center. The electrostatic interaction between the polycationic redox polymer and the protein surface seems also to play

S. Gaspar et al. / Electrochimica Acta 46 (2000) 255–264

an important role. Stronger electrostatic interactions are assumed to result in a tighter coupling and increased current response. Selecting the optimal curing conditions and hydrogel composition, the sensitivity and the linear range of the POD electrodes can be controlled. The most sensitive H2O2 biosensor was obtained for type III bioelectrodes integrating the newly purified sweet potato peroxidase (3.2 AM − 1cm − 2). This value is about three times higher than that reported for HRP immobilized in a similar matrix (1 AM − 1cm−2 [43]), being one of the highest ever reported for bioelectrocatalytical H2O2 detection. These high sensitivity values are, however, rather attributed to the particular bioelectrode design than to an enhanced bioelectrocatalytic activity of the investigated PODs. SPP has also been found to be the most efficient for H2O2 detection when immobilized by simple adsorption on graphite electrode (type I electrodes), suggesting that its redox center is the most accessible. The newly isolated and characterized enzymes are thus very promising candidates for further use in complex, coupled enzyme based systems. Such experiments are currently in progress.

Acknowledgements The European Commission (Contract IC15CT961008), the Swedish Council for Forestry and Agricultural Research (SJFR), the Swedish National Board for Industrial and Technical Development and the Swedish Institute (S.G and I.C.P.) are acknowledged for financial support.

References [1] A.L. Ghindilis, P. Atanasov, E. Wilkins, Electroanalysis 9 (1997) 661. [2] J. Woodward, in: A. Mulchandani, K.R. Rogers (Eds.), Enzyme and Microbial Biosensors. Techniques and Protocols, Humana Press Inc., Totowa, New Jersey, 1998, pp. 67 – 79. [3] T. Ruzgas, E. Cso¨regi, J. Emne´us, L. Gorton, G. MarkoVarga, Anal. Chim. Acta 30 (1996) 123. [4] H.B. Dunford, in: J. Everse, K.E. Everse, M.B. Grisham (Eds.), Peroxidase in Chemistry and Biology, vol. 2, CRC Press, Boca Raton, 1991, pp. 2–24. [5] M.S. Vreeke, K.T. Yong, A. Heller, Anal. Chem. 67 (1995) 4247. [6] E. Cso¨regi, G. Jo¨nsson-Petersson, L. Gorton, J. Biotechnol. 30 (1993) 315. [7] L. Gorton, G. Bremle, E. Cso¨regi, G. Jo¨nsson-Petersson, B. Persson, Anal. Chim. Acta 43 (1991) 249. [8] V. Razumas, J. Kazlauskaite, T. Ruzgas, J. Kulys, Bioelectrochem. Bioenerg. 28 (1992) 159. [9] T. Tatsuma, T. Watanabe, Anal. Chem. 63 (1991) 1580.

263

[10] R.M. Paddock, E.F. Bowden, J. Electroanal. Chem. 260 (1989) 487. [11] E. Cso¨regi, L. Gorton, G. Marko-Varga, A.-J. Tu¨do¨s, W.T. Kok, Anal. Chem. 66 (1994) 3604. [12] J. Kulys, R.D. Schmid, Bioelectrochem. Bioenerg. 24 (1990) 305. [13] M.S. Vreeke, A. Heller, in: A.M. Usmani, N. Akmal (Eds.), Diagnostic Biosensor Polymers. In: A.C.S. Symposium Series, vol. 556, American Chemical Society, Washington, 1994, pp. 180– 192. [14] G. Kenausis, Q. Chen, A. Heller, Anal. Chem. 69 (1997) 1054. [15] A. Lindgren, J. Emne´us, T. Ruzgas, L. Gorton, G. Marko-Varga, Anal. Chim. Acta 347 (1997) 51. [16] T. Ruzgas, L. Gorton, J. Emne´us, E. Cso¨regi, G. MarkoVarga, Anal. Proceed. 6 (1995) 207. [17] F.-D. Munteanu, A. Lindgren, J. Emne´us, L. Gorton, T. Ruzgas, E. Cso¨regi, A. Ciucu, R.B. van Huystee, I.G. Gazaryan, L.M. Lagrimini, Anal. Chem. 70 (1998) 2596. [18] L. Gorton, A. Lindgren, T. Larsson, F.-D. Munteanu, T. Ruzgas, I. Gazaryan, Anal. Chim. Acta 400 (1999) 91. [19] A. Lindgren, F.-D. Munteanu, I.G. Gazaryan, T. Ruzgas, L. Gorton, J. Electroanal. Chem. 485 (1998) 113. [20] L. Gorton, E. Cso¨regi, T. Ruzgas, I. Gazaryan, G. Marko-Varga, in: A. Mulchandani, K.R. Rogers (Eds.), Enzyme and Microbial Biosensors. Techniques and Protocols, Humana Press, Totowa, New Jersey, 1998, pp. 93 – 120. [21] V. Rajendran, E. Cso¨regi, Y. Okamoto, L. Gorton, Anal. Chim. Acta 373 (1998) 241. [22] A. Heller, Acc. Chem. Res. 23 (1990) 23. [23] A. Heller, J. Phys. Chem. 96 (1992) 3579. [24] E. Cso¨regi, D. Schmidtke, A. Heller, Anal. Chem. 67 (1995) 1240. [25] N. Larsson, T. Ruzgas, L. Gorton, P. Kissinger, E. Cso¨regi, Electrochim. Acta 43 (1998) 3541. [26] I.G. Gazaryan, L.M. Lagrimi, Phytochemistry 41 (1996) 1029. [27] T.J. Ohara, R. Rajagopalan, A. Heller, Anal. Chem. 66 (1994) 2451. [28] A.C. Maethly, Methods Enzymol. 2 (1953) 801. [29] Enzymatic Assay of Peroxidase, Sigma Quality Control Test Procedure, Sigma-Aldrich, Inc., (1994). [30] R. Appelqvist, G. Marko-Varga, L. Gorton, A. Torstensson, G. Johansson, Anal. Chim. Acta 169 (1985) 237. [31] L. Gorton, Electroanalysis 7 (1995) 23. [32] T. Ruzgas, L. Gorton, J. Emne´us, G. Marko-Varga, J. Electroanal. Chem. 391 (1995) 41. [33] Worthington, B.C., (Ed.), 1988. Worthington Enzyme Manual Worthington Biochemical Corporation, Freehold, p.254. [34] I.Y. Sakharov, A.G. Bautista, I.V. Sakharova, A. Rojas, O.Y. Pletyuschkina, Revista Colombiana de Quimica (2000) in press. [35] A. Lindgren, M. Tanaka, T. Ruzgas, L. Gorton, I. Gazaryan, K. Ishimori, I. Morishima, Electrochem. Commun. 1 (1999) 171. [36] I.G. Gazaryan, L.M. Lagrimi, S.J. George, R.N.F. Thorneley, Biochem. J. 320 (1996) 369.

264

S. Gaspar et al. / Electrochimica Acta 46 (2000) 255–264 [43] M. Vreeke, R. Maidan, A. Heller, Anal. Chem. 64 (1992) 3084. [44] M. Elmgren, S.-E. Lindquist, M. Sharp, J. Electroanal. Chem. 362 (1993) 227. [45] N.A. Surridge, E.R. Diebold, J. Chang, G.W. Neudeck, in: A.M. Usmani, N. Akmal (Eds.), Diagnostic Biosensor Polymers. In: A.C.S. Symposium Series, vol. 556, American Chemical Society, Washington, 1994, pp. 47 – 70. [46] T.J. Ohara, R. Rajagopalan, A. Heller, Anal. Chem. 65 (1993) 3512. [47] B.A. Gregg, A. Heller, J. Phys. Chem. 95 (1991) 5976. [48] C.P. Andrieux, O. Haas, J.-M. Save´ant, J. Am. Soc. 108 (1986) 8175. [49] I. Katakis, L. Ye, A. Heller, J. Am. Chem. Soc. 116 (1994) 3617.

[37] F.-D. Munteanu, L. Gorton, T. Ruzgas, J. Emne´us, E. Cso¨regi, I.G. Gazaryan, I.V. Ouporov, E.A. Mareeva, A. Giucu, L.M. Lagrimini, Appl. Biochem. Biotechnol. (2000) in press. [38] I. Katakis, A. Heller, in: F.W. Scheller, F. Schuhbert, J. Fedrowitzk (Eds.), Frontiers in Biosensoric I. Fundamental Aspects, Birkha¨user Verlag, Basel, 1997, pp. 229–241. [39] B.A. Gregg, A. Heller, J. Phys. Chem. 95 (1991) 5970. [40] S. Cosnier, A. Deronzier, J.-C. Moutet, J.F. Roland, J. Electroanal. Chem. 271 (1989) 69. [41] K.-N. Kuo, R.W. Murray, J. Electroanal. Chem. 131 (1982) 37. [42] I.C. Popescu, S. Cosnier, P. Labbe, Electroanalysis 9 (1997) 998.

.