Peroxidase enzyme electrodes as nitric oxide biosensors

Analytica Chimica Acta 403 (2000) 1–9

Peroxidase enzyme electrodes as nitric oxide biosensors E. Casero, M. Darder, F. Pariente, E. Lorenzo ∗ Departamento de Qu´ımica Anal´ıtica y Análisis Instrumental, Universidad Autónoma de Madrid, Cantoblanco 28049, Madrid, Spain Received 17 March 1999; received in revised form 16 July 1999; accepted 25 July 1999

Abstract An inexpensive and simple nitric oxide biosensor is developed with a peroxidase amperometric electrode using hydrogen peroxide as substrate and p-benzoquinone as mediator. The enzyme has been immobilized on a glassy carbon electrode or on a nylon mesh which is attached to the electrode. The influence of the immobilization method on the biosensor response has been studied. Peroxidase reduces hydrogen peroxide to water and oxidizes the mediator which is regenerated in a cathodic reaction. Cyclic voltammetry has been employed to assess the ability of certain redox-active couples to act as mediator. The activity of horseradish peroxidase is inhibited in the presence of nitric oxide. Thus, the decrease in activity of the enzyme monitored can be correlated to the concentration of nitric oxide present in solution. This biosensor responds to 2.7 × 10−6 −1.1 × 10−5 M nitric oxide. A detection limit of 2.0 × 10−6 M for nitric oxide was found. ©2000 Elsevier Science B.V. All rights reserved. Keywords: Biosensor; Peroxidase; NO; Enzyme electrode

1. Introduction The importance of nitric oxide (NO) determination in clinical and environmental analysis has resulted in an increasing interest in the development of methods for detecting and monitoring this compound, which is involved in a number of diverse physiological processes. The ability to travel rapidly between cells and the short half-life time make NO ideal in intercell communication as well as for other important functions. Among others, although there is no complete understanding of NO functions, the NO radical clearly accounts for the activity of the endothelium-derived relaxing factor (EDRF) [1], acts as a neurotransmitter [2], prevents platelet aggregation [3] and plays an im∗ Corresponding author. Tel.: +34-1-3974488; fax: +34-1-3974931 E-mail address: [email protected] (E. Lorenzo)

portant role in the immune system against tumor cells [4] and intracellular parasites [5]. Despite the great interest in the determination of NO production, most of the methods used for the detection of NO are indirect, relying on measurements of secondary species, such as, breakdown products of NO metabolism (nitrite and nitrate). Other bioassays have been proposed, based on the physiological effects of NO, such as the stimulation of guanylate cyclase, or inhibition of platelet aggregation [6]. Recent developments in probe technology have allowed the direct detection of NO which are very attractive because they offer the highest specificity. The two most direct approaches are the detection of the electric current produced when NO is oxidized and the detection of light produced when NO reacts with ozone (chemiluminescence) [7]. In order to obtain direct evidence for NO functions in vivo Kojima et al. [8] have synthesized diamino fluoresceins as fluorescent indicators for NO.

0003-2670/00/$ – see front matter ©2000 Elsevier Science B.V. All rights reserved. PII: S 0 0 0 3 - 2 6 7 0 ( 9 9 ) 0 0 5 5 5 - 3

2

E. Casero et al. / Analytica Chimica Acta 403 (2000) 1–9

The interaction of nitric oxide within biological systems provides a key to the construction of optical biosensors. Fiber-optical biosensor that are selective to nitric oxide have been prepared with cytochrome c0 [9]. Due to their inherent sensitivity, electrochemical techniques are particularly well suited for the development of analytical methods for the determination of NO, and the use of microelectrodes can allow for in vivo applications. Thus, there have been some studies performed aiming at the development of electrochemically based sensors for the determination of NO [10,11]. We have also described a sensor for the direct determination of NO based on its oxidation at platinum electrodes modified with Nafion and cellulose acetate [12]. The purpose of the Nafion film was to exclude anionic species such as nitrite, which, like ascorbate, is a severe interferent. Enzyme-based biosensors are highly selective devices which rely on the specific binding of the target analyte (the substrate) to the active-site regions of the enzyme. Amperometric signals resulting from this biorecognition process have led to many useful enzyme electrodes [13,14]. The response of these devices is often affected by the presence of inhibitors, which combine with the free enzyme in a manner that prevents substrate binding having great influence on the velocity of the biocatalytic reaction [15]. Indeed, such inhibitory effects have been exploited for indirect assays of the inhibitor. Horseradish peroxidase (HRP) is a redox hemoglycoprotein. The sixth coordination position of Fe(III) in the Fe(III) protoporphyrin IX prosthetic group of the enzyme molecule is the active site in HRP-catalyzed reactions [16]. In a single two-electrons process the HRP [Fe(III)] is oxidized by hydrogen peroxide to form water and an oxidized form of HRP, denoted HRP I. The reduction of HRP I back into the HRP [Fe(III)] occurs in two separate 1e− step. The complete process can be represented by the following reaction sequence [16–18]: HRP [Fe(III)] + H2 O2 → HRP I + H2 O −

(1)

HRP I + 1e → HRP II

(2)

HRP II + 1e− → HRP [Fe(III)]

(3)

HRP is the most widely studied of all the peroxidases as a biological catalyst in the development of

Scheme 1. Schematic representation of the assay system. HRPox and HRPred are the redox forms of HRP and Medox and Medred are the redox forms of the mediator of electron transfer.

enzyme-based amperometric biosensors. For an ideal devices the enzyme regeneration, Step (3), must take place at the electrode surface. However, rapid electron transfer between electrodes and metalloproteins is, in general, difficult to achieve. Thus, many electrochemical studies of redox proteins have made use of small-molecule electrode-active mediators, either free in solution [19,20] or bound to the electrode surface [21,22] to enhance the rate of electron transfer between electrode and protein. This is the approach we have adopted in the design of a peroxidase based electrochemical assay and a peroxidase based biosensor for H2 O2 . The essential features of the analytical system are shown schematically in Scheme 1. The fundamental reaction is the two-electron oxidation of peroxidase in the presence of substrate (H2 O2 ), the latter being reduced to water. The reduction of the enzyme back to its original oxidation state is carried out by electron transfer from a suitable mediator in its reduced form. Thus, in the presence of hydrogen peroxide an external current will flow, the magnitude of which is a function of the peroxide concentration. A mediator is required because it is true that redox conversion of peroxidase has been observed on modified or pretreated electrodes [23–26] on bare electrodes, but electrochemical activity is, in general, a slower process [27]. Thus, prior to the biosensor development, the activity of a number of redox-active compounds to act as mediator was assessed by cyclic voltammetry. The characteristics of assays based on different mediator–peroxidase combinations were then evaluated in terms of sensibility and reliability. As previously described in the literature [28] nitric oxide can react directly with Fe(III) protoporphyrin, [Fe(III) (PP)]+ , yielding the porphyrin nitrosyl complex 2 according to the reaction:

E. Casero et al. / Analytica Chimica Acta 403 (2000) 1–9

[Fe(III) (PP)]+ + NO → [Fe(II) (NO+ ) (PP)]+ 1

3

2.2. Instrumentation

2

The formation of the iron nitrosyl porphyrin complex is related to the partial inhibition of the Reaction (1). In a similar way, we have observed a decrease in the activity of HRP in presence of NO. In addition, the decrease in the current intensity associated to the electrocatalytic reduction of hydrogen peroxide is correlated to the concentration of NO present in solution. Thus, based on these results, in this work a novel amperometric biosensor for the determination of NO based on its inhibitory effect on HRP activity immobilized on glassy carbon electrodes is described.

2. Experimental section 2.1. Materials High-purity nitric oxide (NO) was purchased from Air Liquide. Standard saturated NO solutions were prepared by bubbling NO gas through oxygen-free 0.1 M phosphate buffer solutions for 30 min. Aliquots of this saturated solution were used to prepare solutions of known NO concentration, using a value of 1.9 mM for its concentration at saturation [29]. Peroxidase Type I and II (EC 1.11.1.7) from horseradish were obtained from Sigma as a powder containing 120 and 200 units per mg of solid, respectively and were stored as received at −20 ◦ C. Glutaraldehyde (Grade I, 25% aqueous solution) was obtained from Sigma and stored below 0◦ C. Nylon filter meshes of 150 ␮m pore size were purchased from Nytal. Hydrogen peroxide solution was purchased from Carlo Erba. The concentration of the hydrogen peroxide solution was determined from its absorbance at 240 nm using an extinction coefficient of 39.41 mmol−1 cm−1 [22]. p-Benzoquinone was obtained from Aldrich Chemical Co. and used as received. Methyl viologen was obtained from Sigma and 1,4-naphtoquinone from Merck. [Ru(phendione)2 bpy](PF6 )2 was prepared as described previously [30]. Sodium phosphate was used in the preparation of supporting electrolyte and buffer solutions. Water was purified with a Millipore Milli-Q system.

Cyclic voltammetric and chronoamperometric studies were carried out with a BAS CV-27 potentiostat connected to a BAS X-Y recorder. The electrochemical experiments were carried out in a three-compartment electrochemical cell with standard taper joints so that all compartments could be hermetically sealed with PTFE adapters. This is important so as to ensure that there is no gas leakage to or from the electrochemical cell. A PTFE covered glassy carbon (GC) disk electrode (geometric area, 0.07 cm2 ) was used as working electrode. It was polished prior to use with 1 ␮m diamond paste (Buehler) and rinsed thoroughly with water. A coiled platinum wire (5 cm) served as auxiliary electrode. All potentials are reported against a sodium-saturated calomel electrode (SSCE) without taking into account the liquid junction. Pre-purified nitrogen gas was used to deaerate all solutions before use and flowed over the solutions during experiments, in order to minimize the reaction of NO with oxygen. All measurements were carried at room temperature.

2.3. Enzyme immobilization and biosensor construction 2.3.1. Chemical cross-linking Prior to immobilization, enzyme stock solutions were prepared by dissolving 4 mg of the enzyme in 24 ␮l of 0.1 M phosphate buffer. Nylon meshes were cut into 6 mm diameter disks. The disks were dipped in methanol, rinsed with water and dried in air prior to use. For enzyme immobilization, the following solutions were added to each disk: 2.0 ␮l of glutaraldehyde (2.5% v/v), 2.0 ␮l of bovine serum albumin (BSA) (10% w/v) and 2.5 ␮l of the peroxidase stock solution (50 U). The mixture was homogenized on the surface of the disk and allowed to dry at room temperature for 30 min. Prior to use, the remaining unreacted carboxaldehyde groups were inactivated by immersing the disks in 0.1 M phosphate buffer (pH 7.0) containing 0.1 M glycine for 20 min at room temperature. When not in use, the membranes were stored in 0.1 M phosphate buffer solution (pH 7.0) at 4◦ C.

4

E. Casero et al. / Analytica Chimica Acta 403 (2000) 1–9

The biosensor was assembled by securing, with a perforated cap, an enzyme-modified nylon mesh disk (prepared as described above) over a glassy carbon electrode previously polished and activated by placing it in 1.0 M NaOH solution and holding the potential at +1.20 V for 5 min, as we have previously described [31]. The assembled biosensor was placed in buffer solution for ca. 5 min prior to use to ensure solvent equilibration. 2.3.2. Electrochemically induced cross-linking HRP was immobilized by electrochemically induced cross-linking [32]. 10 mg of peroxidase , 10 mg of BSA (10% w/v) and 100 ␮l of glutaraldehyde (2.5% v/v) were added to 2 ml of a pH 7.0 solution containing: 5.3 mM potassium phosphate, 2.5 mM potassium chloride and 0.15 M sodium chloride. The solution was carefully stirred to dissolve the proteins and transferred to the electrochemical cell. A pulse train composed of 0.2 s pulses at 10 ␮A separated by 2 s at open circuit potential for 30 min was applied. 2.4. Procedures The ability of the different redox-active couples to act as a mediator was assessed by cyclic voltammetry. Solutions containing the mediator in 0.1 M phosphate buffer pH 6.5 were degassed with a steam of nitrogen gas. The potential of working electrode was poised sufficiently negative with regard to the formal potential for the mediator to generate the reduced form. The current response was monitored on addition of peroxidase and a small volume of a solution containing hydrogen peroxide. For the determination of NO using a peroxidase based biosensor for hydrogen peroxide, the biosensor was placed in 5 ml of 0.1 M N2 -saturated phosphate buffer solution (pH 7.0) containing 1.0 mM of p-benzoquinone at an applied potential of +0.2 V for 10 s, then the electrode potential was stepped to E2 = −0.3 V and the background current was allowed to decay to a steady-state value. After the addition of 5.0 mM of H2 O2 and stirring for 30 s the resulting current from the reduction of 1,4-benzoquinone reached a steady state (Iss) within 1 min. Aliquots of a saturated NO solution were subsequently added with a gas-tight syringe and the resulting steady-state

current decrease was measured. The percentage of inhibition was calculated according to I (%) =

100(Iss1 − Iss2 ) Iss1

where Iss1 is the initial steady-state current and Iss2 the steady-state current reached after the addition of a determinate amount of inhibitor (NO).

3. Results and discussion 3.1. Evaluation of different mediators by cyclic voltammetry Different compounds were examined with regard to their ability to mediate electron transfer between peroxidase and a glassy carbon electrode: 1,4-naphthoquinone, 1,4-benzoquinone, ruthenium phendione bipyridine complex ([Ru (phendione)2 bpy] (PF6 )2 ) and methyl viologen.A series of cyclic voltammograms was recorded at scan rates of 2–100 mV s−1 . As is expected, under the experimental conditions used in this study, 1,4-naphthoquinone, benzoquinone, and ruthenium phendione complex gave voltammograms consistent with a reversible two-electron/two-proton redox agent, ascribed to the oxidation/reduction of the hydroquinone/quinone group. In aqueous media [30], which is the focus of the present work, the phen-dione as well as its metal complexes exhibited pH-dependent redox responses based on the quinones moieties of the phen-dione ligands. Methyl viologen gave voltammograms consistent with a reversible one-electron redox agent (1Ep ≈ 60 mV; ip /v1/2 = constant). Introduction of horseradish peroxidase had no discernible effect upon their electrochemistry. However, upon addition of hydrogen peroxide to the system, the cyclic voltammograms exhibited a dramatic enhancement of the cathodic peak current with a decrease in the anodic peak current. Both observations were particularly apparent at slower scan rates, implying a catalytic regeneration of oxidized mediator by peroxidase in the presence of hydrogen peroxide. These effects are illustrated in Fig. 1 for the specific example of the Ruthenium–phendione complex (A) and methyl viologen (B). They are, however, representative of the type of results observed with the other compounds studied.

E. Casero et al. / Analytica Chimica Acta 403 (2000) 1–9

5

Fig. 1. Cyclic voltammograms obtained in 0.1 M phosphate buffer at pH 6.5 at different scan rates in the presence of 2.0 ␮M of HRP and 0.5 mM of hydrogen peroxide for 0.3 mM of ruthenium complex (A) and 1.0 mM of methyl viologen (B).

Cyclic voltammetry is a useful technique for studying the rate of electron transfer from a mediator to the enzyme. Quantitative kinetic data for the homogeneous reaction between peroxidase and the mentioned compounds were obtained under conditions of substrate excess using the theory developed by Nicholson and Shain [33] for catalytically coupled reversible electron transfer reactions of the type O + n e− ⇔ R R+Z →O

(E)

(C)

In our system, O and R are the redox forms of the mediator and Z the oxidized peroxidase. The theory allows separation of the diffusion id , and kinetically ik controlled components of the cathodic current.

Data were analyzed according to the theory described in [32] which correlates the experimentally derived parameter, ik /id , to the kinetic parameter (kf /a)1/2 in a similar way to that described by Frew et al. [34]. A scan rate-independent pseudo-first-order rate constant corresponding to a given peroxidase concentration was obtained from the slope of a plot of kf /a against 1/v, where a = nFv/RT and v is the scan rate. Fig. 2 shows the data obtained using 1,4-benzoquinone as mediator. By varying the peroxidase concentration the second-order homogeneous rate constant (k) for the reaction between the enzyme and mediator was obtained from the slope of the plot of kf versus the concentration of HRP (Fig. 3). Values of the rate constants obtained for the different mediators assayed are presented in Table 1. As can be

6

E. Casero et al. / Analytica Chimica Acta 403 (2000) 1–9

Fig. 2. kf calculated for the HRP-1,4-benzoquinone system in the following conditions: 0.16 mM of 1,4-benzoquinone, 0.23 mM of hydrogen peroxide and 0.04 ␮M (䊐), 0.09 ␮M (䊊) and 0.17 ␮M (4) of horseradish peroxidase.

seen electron transfer seems to be more efficient with quinone derivatives. Based on the above results, an analytical system employing 1,4-benzo or naphthoquinone as mediator with peroxidase should be used for greater efficiency. The performance of assays with the different mediators was evaluated in terms of sensitivity and reliability. This involved the construction of calibration graphs using standard aqueous solutions of hydrogen peroxide and the protocol described in procedures in the experimental section. The response is clearly dependent upon the concentration of hydrogen peroxide introduced into the system. Data of this type were processed as follows: the maximum catalytic current (ik ) was calculated from the peak height. The peak current due to the mediator was taken as the background (id ) and all measurements were normalized at the peak current of the mediator, that is ik −id /id . The normalized current was then plotted against hydrogen per oxide concentration. The response to hydrogen peroxide was found to be linear over the concentration range assayed (upper limit 3.5 × 10−4 M) for systems employing the mediators listed in Table 2. As an example, we present the results using 1,4-benzoquinone (0.16 mM) as mediator and 0.17 ␮M of HRP. Linear calibration plots (y = −0.01 + 15.41 x; r = 0.9991) were obtained over the range of 0.5 × 10−4 − 0.5 × 10−3 M. 3.2. Biosensor response

Fig. 3. k obtained for the HRP-1,4-benzoquinone system from slopes of Fig. 2.

Table 1 kf and k obtained for the studied mediators in the following conditions: [H2 O2 ]/[Mediator] = 1.4 Mediator

kf (s−1 )

k (M−1 s−1 )

1,4-Benzoquinonea 1,4-Naphtoquinonea Methyl viologena [Ru(phendione)2 bpy](PF6 )2 b

0.19 0.21 0.11 0.02

2.9 × 106 3.3 × 106 5.3 × 105 9.4 × 104

a [HRP]: b [HRP]:

0.09 ␮M. 0.34 ␮M.

One of the objectives of these investigations was the development of biosensors based on peroxidase. For this purpose HRP was immobilized on the electrode surface. As described in Section 2, two different enzyme immobilization methods were employed and these involved chemical or electrochemical induced cross-linking with glutaraldehyde, respectively. 1,4-benzoquinone was used as a soluble hydrogen donor mediator and hydrogen peroxide as substrate. The cyclic voltammetric response of the biosensor in the absence and presence of hydrogen peroxide was used to assess the activity of the biosensor. Fig. 4a, trace a) shows the cyclic voltammetric response at 5 mV s−1 for a biosensor with HRP immobilized onto glassy carbon electrode by electrochemical induced cross-linking and in contact with a pH 7.0 phosphate buffer solution con-

E. Casero et al. / Analytica Chimica Acta 403 (2000) 1–9

7

Table 2 Linear range and detection limit reached in the NO determination with different peroxidase electrodes Linear range 1.0 × 10−5 –3.0 × 10−5

Chemical cross-linking immobilization Electrochemical induced cross-linking immobilization

2.7 × 10−6 –1.1 × 10−5

Detection limit M M

7.9 × 10−6 M 2.0 × 10−6 M

addition of hydrogen peroxide to a concentration of 1.0 mM, an enhancement of the cathodic current (Fig. 4, traces b), is clearly noted. Additional increases in the concentration of hydrogen peroxide to 2.0 (c) and 3.0 mM (d) resulted in concomitant increases in the cathodic peak current. In addition no current is observed in the return (anodic) wave, consistent with a very strong electrocatalytic effect. Moreover, the biosensor response is linear (y = −0.003 + 1.37 x; r = 0.9999) to hydrogen peroxide concentration in all the range studied as can be seen in Fig. 4B. Similar results are obtained if the biosensor is assembled by securing, with a perforated cap, a HRP-modified nylon mesh disk (prepared as described above) over a glassy carbon electrode. 3.3. Determination of NO

Fig. 4. (A) Cyclic voltammograms (5 mV s−1 ) obtained in 0.1 M phosphate buffer at pH 7.0 for a HRP biosensor (a) 1.0 mM p-benzoquinone, and (b–d) 1.0 mM p-benzoquinone and 1.0 mM, 2.0 mM and 3.0 mM of hydrogen peroxide, respectively; (B) Dependence of the catalytic reduction current on hydrogen peroxide concentration.

taining 1.0 mM 1,4-benzoquinone but in the sence of hydrogen peroxide. The characteristic well-behaved redox response ascribed to the versible two-electron/two-proton process of the droquinone/quinone system is observed. Upon

aband rehythe

As we have previously mentioned, Fe(III) form associated to the HRP has a high affinity for NO. Thus, one would expect that in the presence of NO the formation of the nitroxyl complex takes place and as a result, a decrease in the enzymatic activity would be expected. In order to ascertain this, the voltammetric response of a biosensor, with 50 U of HRP in the nylon mesh, was assayed to 5.0 mM of H2 O2 in the absence and presence of increasing concentrations of NO, and the data are presented in Fig. 5. In the absence of NO (Fig. 5, trace b) the characteristic catalytic current due to HRP biosensor response is readily apparent. Upon the addition of NO (trace c–e) a large decrease in the catalytic current was observed. These results clearly demonstrate an inhibitory effect on the electrocatalytic reduction of H2 O2 . In order to determine the most effective method for enzyme immobilization, comparisons were made in terms of the detection limit and reproducibility reached in NO determination with the resulting biosensor. This involved measuring the biosensor response against a determinate concentration of substrate (5.0 mM H2 O2 ) in pH 7.0 phosphate buffer

8

E. Casero et al. / Analytica Chimica Acta 403 (2000) 1–9

Fig. 6. Relationship between % inhibition of peroxidase electrode and NO concentration. Fig. 5. Cyclic voltammograms (5 mV s−1 ) obtained in 0.1 M phosphate buffer at pH 7.0 for a HRP biosensor with 50 U in the membrane (a) 1.0 mM p-benzoquinone, (b) 1.0 mM p-benzoquinone and 5.0 mM hydrogen peroxide, (c–e) after addition of 3.8, 7.6, and 9.5 ␮M of NO. Inset shows the dependence of the catalytic reduction current (at −0.3 V) on nitric oxide concentration.

containing 1.0 mM 1,4-benzoquinone at a potential of −0.3 V in the absence (control experiment) and in the presence of successive additions of aliquots of a saturated NO solution. Either the HRP electrode was prepared by immobilizing the enzyme onto a nylon mesh (chemical immobilization) or directly onto the GC electrode (electrochemical immobilization), the presence of NO results in a linear decay in the steady state current , that is, a linear decrease in the biosensor response. Data for a biosensor, with 50 U of HRP in the nylon mesh, is presented in the inset of Fig. 5. Each data point represents the mean value of three replicate measurements, and the scatter of individual measurements from this value is indicated. A calibration graph of % inhibition versus [NO] for a biosensor assembled by securing a HRP-modified nylon mesh disk over a glassy carbon electrode is shown in Fig. 6. Each data point represents the mean of three replicate measurements, and the scatter of individual measurements from this value is indicated. The response was linear over the entire range studied (Table 2). The response was rapid (less than 30 s) and reproducible with a detection limit of 7.9 × 10−6 M (defined as the concentration of NO required to ob-

tain 5% of inhibition) and a sensitivity of 1.70 × 106 % I M−1 . Similar results, except with more sensitivity (2.68 × 106 % I M−1 ) and lower detection limit for NO determination (2.0 × 10−6 M, see also Table 2), are obtained if the biosensor is prepared by immobilizing HRP directly onto glassy carbon electrode by electrochemically induced cross-linking. This method also provides for a very easy and convenient way to immobilize the enzyme to microelectrodes. The used biosensor was stored in phosphate buffer under refrigeration overnight, and then the measurement was conducted again. The biosensor response to 5.0 mM of H2 O2 in the absence of NO was about 30% of the initial value and inhibition of the biosensor response by NO was not observed. Thus, we conclude that the developed HRP electrode can serve as a disposable biosensor only. In conclusion we have demonstrated that immobilized HRP in conjunction with a mediator can be used to monitor nitric oxide with high sensitivity and short response time. We are currently investigating the preparation of HRP-modified carbon fiber ultramicroelectrodes and their potential utility for the in vivo determination of NO. Acknowledgements This work was supported by the Comunidad Autónoma de Madrid through the grant 06M/044/96 and by the DGICYT of Spain through the Grant BIO

E. Casero et al. / Analytica Chimica Acta 403 (2000) 1–9

96-1016-C02-02. E.C. and M.D. also acknowledges support by a Comunidad Autonóma de Madrid fellowship. We thank Prof. Abruña (Cornell University) for generously providing the Ru complex. References [1] R.M.J. Palmer, A.G. Ferrige, S. Moncada, Nature 327 (1987) 526. [2] T.J. O’Dell, R.D. Hawkins, E.R. Kandel, O. Arancio, Proc. Natl. Acad. Sci. U.S.A. 88 (1991) 11285. [3] M.W.J. Radomski, R.M. Palmer, S. Moncada, Proc. Natl. Acad. Sci. U.S.A. 87 (1990) 5193. [4] J.B. Hibbs Jr., R.R. Tainton, Z. Vavrin, E. Mrachlin, Biochem. Biophys. Res. Commun. 157 (1988) 87. [5] J.B. Hibbs Jr., Z. Vavrin, R.R. Tainton, J. Immunol. 138 (1987) 550. [6] A.M. Leone, P. Rhodes, V. Furst, S. Moncada, in: D.A. Kendall, S.J. Hill (Eds.), Methods in Molecular Biology, vol. 41: Signal transduction Protocols, Humana Press, Totowa, NJ, 1995. [7] O.C. Zafiriou, M. McFarland, Anal. Chem. 52 (1980) 1662. [8] H. Kojima, N. Nakatsuko, K. Kikuchi, S. Kawahara, Y. Kirino, H. Nagoshi, Y. Hirata, T. Nagano, Anal. Chem. 70 (1998) 2446. [9] S.L.R. Barker, R. Kopelman, T.E. Meyer, M.A. Cusanovich, Anal. Chem. 70 (1998) 971. [10] K. Shibuki, Neurosci. Res. 9 (1990) 69. [11] D.A. Wink, D. Chistodoulou, M. Ho, M.C. Krishna, J.A. Cook, H. Haut, J.K. Randolph, M. Sullivan, G. Coia, R. Murray, T. Meyer, Methods: A Compenion to Methods in Enzymology, vol. 7, Academic Press, San Diego, 1995, pp. 71–77. [12] F. Pariente, J.L. Alonso, H.D. Abruña, J. Electroanal. Chem. 379 (1994) 191. [13] A.P. Turner, I. Karube, G.S. Wilson (Eds.), Biosensors: Fundamentals and Applications, Oxford Press, Oxford, UK, 1987. [14] J.E. Frew, H.A.O. Hill, Anal. Chem. 59 (1987) 933A. [15] M. Dixon, E. Webb, Enzymes, Academic Press, New York, 1979, Chap. VIII.

9

[16] H.B. Dunford, in: J. Everse, K. Everse, M. Grisham (Eds.), Peroxidases in Chemistry and Biology, vol. II, CRC Press, New York, 1991, pp. 1–24. [17] B. Chance, L. Powers, Y. Ching, T. Poulos, G.R. Schonbaum, I. Yamazaki, K.G. Paul, Arch. Biochem. Biophys. 235 (1984) 596. [18] L. Yang, R.W. Murray, Anal. Chem. 66 (1994) 2710. [19] S. Dasgupta, M.D. Ryan, Bioelectrochem. Bioenerg. 7 (1980) 587. [20] A.E.G. Cass, G. Davis, M.J. Green, H.A.O. Hill, J. Electroanal. Chem. 190 (1985) 117. [21] U. Wollenberger, A. Drungiliene, W. Stocklein, J.J. Kulys, F.W. Scheller, Anal. Chim. Acta 329 (1996) 231. [22] C. Lei, J. Deng, Anal. Chem. 68 (1996) 3344. [23] V.J. Razumas, A.V. Gudavicius, J.J. Kulys, J. Electroanal. Chem. 151 (1983) 311. [24] G. Jonsson, L. Gorton, Electroanalysis 1 (1989) 465. [25] J.J. Kulys, R.D. Schmid, Bioelectrochem. Bioenerg. 24 (1990) 305. [26] L. Gorton, M. Bardheim, G. Bremle, E. Csoregi, B. Persson, G. Petersson, Amperometric biosensors based on immobilized enzymes and chemically modified electrodes, in: R.D. Schmid (Ed.), Flow Injection Analysis (FIA)-Based on Enzymes or Antibodies, vol. 14, Gesellschaft fur Biotechnologische Forschung (GBF), Monographs, VCH Weinheim, 1991, pp. 305–314. [27] T. Ruzgas, E. Csoregi, J. Emneus, L. Gorton, G. Marko-Varga, Anal. Chim. Acta 330 (1996) 123. [28] J.N. Younathan, K.S. Wood, T. Meyer, J. Inorg. Chem. 31 (1992) 3280. [29] W. Gerrard, Gas Solubilies Widespread Applications, Pergamon Press, Oxford, 1980. [30] C.A. Goss, H.D. Abruña, Inorg. Chem. 24 (1985) 4263. [31] F. Pariente, E. Lorenzo, H.D. Abruña, Anal. Chem. 66 (1994) 4337. [32] D.J. Strike, N.F. de Rooij, M. Koudelka-Hep, Electrochemical-based Immobilization of Enzymes, in: G.F. Bickerstaff (Ed.), Methods in Biotechnology, vol. 1: Immobilization of Enzymes and Cells, Humana Press, Totowa NJ, 1997. [33] R.S. Nicholson, I. Shain, Anal. Chem. 36 (1964) 706. [34] J.E. Frew, M.A. Harmer, H.A.O. Hill, S.I. Libor, J. Electroanal. Chem. 201 (1986) 1.