Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.enzyme.chem.msu.ru/ryabov/pdf/JBIC_99.pdf Дата изменения: Tue Nov 2 16:11:16 1999 Дата индексирования: Mon Oct 1 21:45:43 2012 Кодировка:

JBIC (1999) 4 : 175­182

Q SBIC 1999

ORIGINAL ARTICLE Alexander D. Ryabov 7 Yulia N. Firsova Aleksei Yu. Ershov 7 Ilia A. Dementiev

Spectrophotometric kinetic study and analytical implications of the glucose oxidase-catalyzed reduction of [M III(LL)2Cl2] c complexes by D-glucose (MpOs and Ru, LLp2,2b-bipyridine and 1,10-phenanthroline type ligands)
Received: 1 July 1998 / Accepted: 13 January 1999

Abstract Glucose oxidase-catalyzed reduction of cis[M III (LL)2Cl2] c (MpOs and Ru) complexes to cis[M II (LL)2Cl2] (LLp2,2b-bipyridine and 1,10-phenanthroline type ligands) by D-glucose is a first-order process in the complex and the enzyme in aqueous buffered solution. The reaction follows MichaelisMenten kinetics in D-glucose and the rate is independent of D-glucose concentration above 0.03 M. The reactivity decreases in the series [Ru(bpy)2Cl2] c 1 [Os(phen)2Cl2] c 1 [Os(4,4b-Me2bpy)2Cl2] c 1 [Os(4,7Me2phen)2Cl2] c. The measured second-order rate constant for the oxidation of reduced glucose oxidase by [Os(phen)2Cl2] c in air equals 1.2!10 5 M P1 s P1 at pH 6.7, [D-glucose] 0.05 M, and 25 7C, which is ca. 20% less than that when the reaction solutions are purged with argon. In the case of [Ru(bpy)2Cl2] c the rate constant equals 1.8!10 5 M P1 s P1 under similar conditions in air, showing higher reactivity of Ru complexes compared with Os ones. The reduction is pH-dependent with a maximum around 7. Added for solubilization of poorly soluble metal complexes, surfactants decrease the rates of the enzymatic reaction. The retardation effect increases in the series: cetyltrimethylammonium bromide ~ Triton X-100 P sodium dodecyl sulfate, i.e. on going from positively charged to neutral and then to negatively charged surfactants. The behavior of the Os III and Ru III complexes toward reduced glucose oxidase contrasts to that of recently studied ferricenium cations. As opposed to the latter, the former do not show kinetically meaningful binding with the enzyme,
A. D. Ryabov (Y) Department of Chemistry, M. V. Lomonosov Moscow State University, 119899, Moscow, Russia e-mail: ryabov@enzyme.chem.msu.su, Fax: c7-95-9395417 A. D. Ryabov 7 Y. N. Firsova Division of Chemistry, G. V. Plekhanov Russian Economic Academy, Stremyanny per. 28, 113054, Moscow, Russia A. Yu. Ershov 7 I. A. Dementiev Department of Chemistry, St. Petersburg State University 198904, St. Petersburg, Russia

and the Michaelis kinetics typical of the ferricenium case is not realized for the Os III, and Ru III species. The systems Os III- or Ru III-glucose oxidase are convenient for routine "one pot" spectrophotometric monitoring of the D-glucose content in samples, since the metal reduction to M II is accompanied by a strong increase in absorbance in the visible spectral region. Key words Glucose oxidase 7 Osmium complexes 7 Ruthenium complexes 7 Reduction 7 D-Glucose Abbreviations GO glucose oxidase 7 SDS sodium dodecyl sulfate 7 CTAB cetyltrimethylammonium bromide 7 cmc critical micelle concentration 7 HFc ferrocene

Introduction
The electron transfer from reduced oxidases at osmium(III) complexes is a crucial step in the functioning of modern miniature and efficient amperometric biosensors in which usually both the enzymatic reaction 1 between oxidase and substrate, as well as step 2 involving an Os III complex as an electron acceptor, occur within a polymer hydrogel [1­6]. E(ox)cS ] E(red)cP E(red)c2M
III

(1)
II

] E(ox)c2M

(2)

Here, E(ox) and E(red) are the oxidized and reduced forms of the enzyme, and S and P are substrate and product, respectively. From the standpoint of bioinorganic chemistry, of essential interest is step 2. This involves the interaction of a reduced enzyme with osmium(III) coordination compounds, and therefore such features as ligand environment, speciation, and acidbase equilibria can strongly affect the reactivity. Since the mechanistic features of the interaction between inorganic and organometallic species with oxidoreductases, in general, and flavine-adenine dinucleotide (FAD)-dependent oxidases, in particular, are mecha-


176

nistically poorly understood [7­9], kinetic and mechanistic investigations of step 2 seem to be of substantial importance both from theoretical and practical points of view. Therefore, in this work we (1) present the results of a kinetic study of the interaction between osmium(III) and ruthenium(III) complexes 1 and 2 shown in Scheme 1 with reduced glucose oxidase (GO) generated from the native enzyme in the presence of an excess of D-glucose, (2) report on the stoichiometry of the process and its analytical potential, and (3) describe the effects of potentially interfering agents on reaction 2. The data obtained are compared with that presented recently on the oxidation of GO(red) by ferricenium cations [10­12].

Table 1 Spectral data for osmium and ruthenium compounds used in this work in aqueous solution (pH 6.7, buffer 0.01 M phosphate, [Triton X-100] 2%) Complex [Os(4,7-Me2bpy)2Cl2]Cl (1a) [Os(4,7-Me2phen)2Cl2] [Os(phen)2Cl2]Cl (1b) [Os(phen)2Cl2] [Os(4,7-Me2phen)2Cl2]Cl (1c) [Ru(bpy)2Cl2]Cl (2) [Ru(bpy)2Cl2]

l

max

(nm)

(M

P1

cm

P1

)

401 536 393 533 390 377 490

6047B308.7 4900B100 7090B65 11 440B730 4700B130 5119B82 7196B71

Kinetic measurements Kinetic and spectrophotometric measurements were carried out on a Shimadzu UV-160A spectrophotometer equipped with a CPS-240A cell positioner/temperature controller and a Hitachi 150-20 spectrophotometer equipped with a thermostated cell holder. Solutions of Os III compounds were prepared in a different manner, depending on their solubility in water. The very soluble compound 1b was dissolved in a phosphate buffer to afford a 3.26!10 P3 M stock solution. Less soluble 1c (2.5 mg) was suspended in 7 mL of water and 1 mL of Triton X-100 was introduced. The suspension was stirred for 1 h and then kept for a few days to achieve complete homogeneity. The final concentration of Os III in the stock solution was 4.36!10 P4 M at [Triton X100]p0.19 M. As mentioned above, the solution of the Ru III compound was prepared from the precursor Ru II complex by the peroxidase-catalyzed oxidation with hydrogen peroxide. The reaction was initiated by addition of 30 mL of the HRP solution (1.07!10 P5 M) to the mixture containing 300 mL of Ru II (2.6!10 P3 M), 60 ml of H2O2 (0.01 M), and 2.21 mL of 0.01 M phosphate buffer in a 1-cm quartz cell. The oxidation can be followed spectrophotometrically at 490 nm by observing the fading of [Ru(bpy)2Cl2]. Solutions of GO were prepared by weighing; the enzyme was dissolved in 0.1 M phosphate buffer and kept frozen. Solutions of D-glucose were allowed to stand at least for 24 h before use to achieve equilibration between the a- and b-anomers. The enzymatic reactions were monitored by following an increase in the absorbance due to the formation of Os II or Ru II complexes at wavelengths of their maxima (Table 1). The reactions were initiated by addition of the stock solution of GO to buffered or unbuffered solutions of Os III or Ru III and D-glucose. It was, however, checked that the order of mixing does not affect the reaction kinetics. The pseudo-first-order rate constants kobs were evaluated from the absorbance versus time plots by fitting the kinetic data to the equation ApAec(A0PAe)e Pkt. The rate constants reported throughout are mean values of at least three determinations. The initial rates (n0pDA / tl, where DA is a change of the optical density during time t, is extinction coefficients of the corresponding Os II and Ru II complexes reported in Table 1, and lp1 cm) were calculated from the initial linear portions of the kinetic curves. Electrochemical measurements were carried out as described elsewhere [11]. All calculations were carried out using a SigmaPlot 2.01 package.

Materials and methods
Reagents Glucose oxidase from Aspergillus niger (EC 1.1.3.4) with the activity 220 U/mg was purchased from Serva and used as received. The diimine ligands bpy and phen were obtained from Reanal and Reakhim, respectively. Horseradish peroxidase (HRP) (R/Zp2.8) was obtained from Dia-M and used as received. Methyl-substituted bpy and phen ligands were obtained from Aldrich. Sodium dithionite (80%) was a Sigma preparation. L-Ascorbic and L-glutamic acids, L-histidine hydrochloride, inorganic salts, and the components of buffer solutions were high purity Reakhim reagents. Cetyltrimethylammonium bromide (CTAB) and sodium dodecyl sulfate (SDS) were purchased from Fluka, and Triton X-100 was a Sigma reagent. Osmium complexes used in this work were prepared using the procedures described in the literature starting from OsO4 (Aldrich), which was initially converted into (NH4)2OsCl6. Dichlorobis(4,4b-dimethyl-2,2b-bipyridine)osmium(III) chloride (1a), dichlorobis(1,10-phenanthroline)osmium(III) chloride (1b), and dichlorobis(4,7-dimethyl-1,10phenanthroline)osmium(III) chloride (1c) were prepared in more than 90% yield by following the procedure of Kober et al. [13]. Dichlorobis(2,2b-bipyridine)ruthenium(III) cation (2) was prepared in situ from bis(2,2b-bipyridine)dichlororuthenium(II) (Aldrich) by the peroxidase-catalyzed oxidation with hydrogen peroxide (Y. N. Firsova, A. V. Eliseev, A. D. Ryabov, unpublished results). Spectral data for all Os III and Ru III complexes are summarized in Table 1 together with the data for related Os II and Ru II species. The latter were generated in situ from their Os III precursors by reduction with a saturated freshly prepared solution of sodium dithionite [13].

Results
General observations Significant progress in the utilization of osmium compounds as electron transfer mediators in amperometric biosensors suggests that reaction 2 must be a clear-cut

Scheme 1


177

process which is not accompanied by the formation of inorganic side-products. Since Os II and Os III complexes, as well as Ru II and Ru III ones, with nitrogen diimine ligands absorb light differently in the visible spectral region [13], reaction 2 is easy to follow spectrophotometrically. The course of the enzymatic reduction of Os III to Os II is shown in Fig. 1 by the example of complex 1b. As seen, the conversion is characterized by a strong absorbance change. The isosbestic point observed at ca. 412 nm demonstrates the presence of only two species in the system, viz. Os II and Os III. Stoichiometry and analytical virtue In excess of 2 compared to D-glucose, reaction 2 is convenient for routine "one pot" spectrophotometric monitoring of D-glucose concentrations in samples. Addition of an aliquot containing D-glucose results in a change in absorbance which is complete in a matter of 1­2 min (Fig. 2). When the absorbance levels off, the next sample of D-glucose can be introduced into the same solution and the increase in absorbance can be monitored again. As seen from Fig. 2, the procedure can be carried out several times until Ru III is completely reduced to Ru II. Figure 3 demonstrates that the same operation is possible with an Os III complex. The total amount of Os II or Ru II formed is directly proportional to the total concentration of D-glucose added with zero intercept (see insets in Figs. 2 and 3). Slopes of the plots were equal to 1.89B0.04 and 1.8B0.04 for [Os(4,4b-Me2bpy)2Cl2] c and [Ru(bpy)2

Fig. 2 Absorbance changes at 490 nm due to reduction of Ru III to Ru II that follow addition of 2!10 P8 mol D-glucose to a solution containing GO (6.25!10 P5 M) and 2 (1.45!10 P3 M); total volume 2.5 mL, pH 6.7, 25 7C. Inset Linear dependence between the amount of D-glucose added and Ru II formed

Fig. 3 Absorbance changes at 533 nm due to reduction of Os III to Os II that follow addition of 2!10 P8 mol D-glucose to a solution containing GO (6.25!10 P5 M) and 1b (2!10 P3 M); total volume 2.5 mL, pH 6.7, 25 7C. Inset Linear dependence between the amount of D-glucose added and Os II formed

Fig. 1 Glucose oxidase (GO)-catalyzed reduction of Os III to Os II by D-glucose accompanied by a strong absorbance change; [GO] 1.7!10 P8 M, [D-glucose] 0.05 M, [1b] 1!10 P4 M, 0.01 M phosphate, pH 6.7, 25 7C, time intervals 2 min

Cl2] c, respectively, i.e. close to 2, as should be expected since the oxidation of glucose to gluconolactone is a two-electron process. The observed slopes are lower than two and this could, in principle, result from the competitive M III oxidation of GO(red) by dioxygen dissolved in the aqueous buffer solution (see below). However, a good linearity seen in the insets of Figs. 2 and 3 suggests that the affect of O2 is minor because in


178

the opposite case the dioxygen content in solution must diminish after every addition of glucose [14]. The results shown indicate also that the Ru III and Os III complexes can be used for "one-pot" spectrophotometric analyses of the glucose content in different samples. Several determinations are possible until a trivalent Ru or Os complex is completely reduced by D-glucose. It is important to note that a similar approach has been recently introduced with ferricenium compounds [10], but the higher reactivity of Ru III species makes this variant more attractive because of higher rates and more distinct end points. Formal kinetics The reaction progress was monitored by following the formation of Os II or Ru II species under the conditions [D-glucose]p[Os III] or [Ru III]. The kobs values were found to be practically independent of the initial concentrations of the M III species (Table 2), suggesting a first-order dependence in the oxidant. The same conclusion was reached when we measured the initial rates of formation of Os II or Ru II species, which were linear functions of [Os III] or [Ru III] passing through the origin. A first-order behavior in 1 and 2 is a feature that distinguishes the osmium and ruthenium systems from the recently studied ferricenium one, where a Michaelis-type dependence was observed on the concentration of the latter [12]. Whereas the ferricenium case provides kinetic evidence for the interinediate binding between the organometallic substrate and reduced GO, no such evidence was detected in the case of the osmium(III) or ruthenium(III) complexes. It should however be noted that the range of concentrations of the osmium complexes used is limited by their solubility. Linear dependencies of kobs against GO concentration in the range (0.25­1.50)!10 P7 M pass through the origin for all complexes studied and Fig. 4 demonstrates that the values of kobs level off at higher glucose concentrations, in accord with the Michaelis-Menten equation: k
obs

Fig. 4 Dependence of kobs against D-glucose concentration in the case of 1b (}), 1c (L), 1a (G), and 2 ([). Conditions: 25 7C, 0.01 M phosphate, pH 6.7, [Triton X-100] 2%, [GO] 1.5!10 P7 M

p

k3 [GO] [glucose] KMc[glucose]

(3)

The corresponding parameters of Eq. 3, viz. k3 and KM, are summarized in Table 2. The values of the effective Michaelis constants KM obtained are very close to the lower limit of the range (0.01­0.03 M) usually found for GO-catalyzed reactions [15]. Somewhat lower values of KM may, in principle, result from weak interactions between the metal species and GO in the vicinity of the active site. The effect of dioxygen should always be taken into account when reactions with O2-dependent oxidases are under investigation. We checked the influence of dioxygen on the reaction involving 1b. Since this complex is water soluble even in the absence of surfactants, argon could be bubbled through the reaction solutions before kinetic runs. Extensive bubbling resulted in an increase in kobs by ca. 20%. A similar effect of dioxygen was observed in the case of monitoring this interaction by cyclic voltammetry [4]. Since the effect was not dra-

Table 2 Pseudo-first-order rate constants for glucose oxidase-catalyzed reduction of M III complexes by D-glucose and apparent parameters of the Michaelis-Menten equation (25 7C, 0.01 M phosphate, pH 6.7, [glucose] 0.05 M, [GO] 1.5!10 P7 M) Complex 1a 1b 1c 2
a a a

[M III] (M) 4!10 P5 8!10 P5 1.2!10 P4 4.15!10 P 8.3!10 P5 1.25!10 P 5.4!10 P5 8.9!10 P5 1.7!10 P4

k

obs

(s

P1

)
P3 P3 P3

KM (M) (3.5B0.2)!10 (6.5B1.2)!10 (5.5B0.6)!10 (6.6B0.7)!10
P3

k3 (M

P1

s

P1

)
4

5

4

(8.1B0.2)!10 (8.8B0.2)!10 (8.7B0.2)!10 (1.57B0.03)! (1.80B0.04)! (1.62B0.03)! (4.8B0.1)!10 (4.30B0.02)! (2.7B0.7)!10

(3.80B0.04)!10 (1.21B0.05)!10 (3.20B0.09)!10 (1.80B0.05)!10

10 10 10 10

P2 P2 P2

P3

5

P3 P3

P3

4

P2

P3

5

Measured in the presence of 2% Triton X-100


179

matic and taking into account the impossibility of applying the same procedure to the solutions containing surfactants, all other kinetic runs were carried out in air. Effects of pH and Cl

Surfactant effects Two osmium complexes tested in this work, viz. 1a and 1c, are not soluble enough in aqueous solutions. As before [18], this problem was solved by using differently charged surfactants (CTAB, SDS, and Triton X-100) which do not noticeably affect the catalytic activity of GO [14]. The insensitivity of the rate of dioxygen consumption in the presence of D-glucose and GO to the micellar media was verified in the course of our recent study of the micellar effects on the kinetics of GO-catalyzed reduction of ferricenium and n-butylferricenium cations by D-glucose [14]. It was also demonstrated that as far as various ferricenium cations were concerned, the micellar effect was insignificant in the case of the unsubstituted ferricenium cation, but substantial for the more hydrophobic n-butylferricenium one. The reaction rate decreased with increasing surfactant concentration, the retardation effect increasing in the series CTAB~Triton X-100PSDS. As seen from Fig. 6, the same tendency is observed for the osmium complex 1b. The highest reactivity is observed in cationic CTAB micelles; it is slightly lower in the case of neutral micelles of Triton X-100, and there is no activity in the negatively charged SDS micelles. As before, the data were analyzed in terms of the Berezin pseudo-phase model [19, 20] according to which the second-order rate constant k3 is given by Eq. 4: k3 p km PGO POs CVckw (1PCV ) {1c(PGOP1) CV }{1c(POsP1) CV } (4)

P

The pH profile of kobs is shown in Fig. 5, which contains two sets of data, viz. obtained in buffered and bufferfree solutions. As can be seen, both pH profiles are bell-shaped with a rather sharp maximum around 7. Interestingly, the rate constants are somewhat higher in unbuffered solutions and the effect is more pronounced below pH 7, whereas the difference disappears in more basic solutions. This effect can qualitatively be rationalized by the specific effect of phosphate on the catalytic activity of GO, rather than by a change in the coordinative environment of the Os III complex taking into account its inertness toward the water exchange and, hence, substitution [16]. In accord with this is the lack of any dependence of kobs on Cl P concentration in the range 0.0005­0.012 M. The same is true for the UV-vis spectra and redox potentials of 1b measured by cyclic voltammetry in the indicated concentration range. It is known [15] that the shape of the activity versus pH profile in GO catalysis is strongly influenced by the nature of a second substrate, i.e. other than D-glucose. This makes the basis for the classification of substrates of GO. Figure 5 is typical of the so-called substrates of group II, among which are methylene blue, toluidine blue, tetrathiafulvalene, tetracyanoquinodimethane, and benzylviologen [15], as well as ferricenium cations [12]. A sharp maximum at pH 7 is indicative of the involvement of L-histidine residues which are located in the vicinity of FAD [17].

where km and kw are the second-order rate constants in the "micellar" pseudo-phase and the aqueous phase, respectively, PGO and POs are the partition coefficients

Fig. 5 pH-dependence of kobs for the GO-catalyzed oxidation of 1b by D-glucose in 0.01 M NaClO4 (L) and 0.01 M phosphate (}), 25 7C, [Triton X-100] 2%, [GO] 1.5!10 P7 M, [D-glucose] 0.05 M

Fig. 6 Dependence of k3 on concentration of CTAB (L), Triton X-100 (}), and SDS (G) for 1b; 25 7C, 0.01 M phosphate, pH 6.7, [D-glucose] 0.05 M, [GO] 1.5!10 P7 M


180

for GO and osmium complex, respectively, between the micellar and aqueous phases (PAp[A]m / [A]w, ApGO or Os III), C is the total surfactant concentration without cmc (Cp[surfactant]tPcmc), and V is the molar volume of the micelles. The data in Fig. 6 were fitted to the simplified rate equation 5 which formally indicates that when CVP1 (low surfactant concentrations), the hydrophilic enzyme is localized predominantly in the aqueous phase, whereas the charged osmium complex having a rather hydrophobic periphery is distributed between the micellar and aqueous pseudo-phases. k3 p km PGO POs CVck 1cPOs CV
w

conolactone concentration of 0.125 M the reaction rate drops 10-fold compared to the highest one at its concentration of 2!10 P4 M. A decrease in activity at higher concentrations of gluconolactone has been noted by several groups of workers and is attributable to slow inactivation of GO by gluconolactone [15, 21]. -Glutamic acid and L-histidine

L

(5)

The fitting of the data in Fig. 6 to Eq. 5 obtained in the three micellar media on assumptions made previously [14] that the medium-independent kw equals (1.08B0.01)!10 5 M P1 s P1 and V;0.3 cm 3 mol P1 [20] gave the following numerical values: (3.2B0.1)!10 4, and km PGOp(7.50B0.02)!10 4, (1.45B0.33)!10 M P1 s P1, and POsp430B100, 250B100, and 11 000B1000 in CTAB, Triton X-100, and SDS micelles, respectively. The value of POs in the SDS medium emphasizes a strong binding between the positively charged Os III complex and the negatively charged micelle. It is not, of course, surprising because there are two types of favorable interactions, viz. hydrophobic and electrostatic, which both increase the stability of the complex-micelle aggregate. The lowest value of km PGO in this case is easy to understand taking into account the pI value of GO which is close to 4 [15]. Hence, the enzyme is negatively charged at pH around 7, ruling out effective interaction between two identically charged macromolecular species. The electrostatic effect accounts for the highest values of km in the case of positively charged CTAB micelles.

L-Glutamic acid and L-histidine were chosen as effectors because of their ability to form N,O- and N,N-chelate rings that could affect the reactivity of the osmium complexes. Data in Fig. 7 show that both the effectors diminish the rate of reduction of the Os III complex. It was first assumed that the origin of the effect is complex formation between 1a and the amino acids as shown by Eq. 6. This explanation was rejected, however, since no supporting evidence for Eq. 6 was obtained by either UV-vis spectroscopy or cyclic voltammetry. The spectra of 1a remained unchanged in the presence of L-glutamic acid or L-histidine. No significant change in the formal potential E 0b of 1a was detected in the presence of the two amino acids.

[Os(LL)2Cl2] cchis P a [Os(LL)2(his)]

2c

c 2 Cl

P

(6)

It had to be concluded that the retardation effect should be ascribed to the lowering of the enzymatic activity of GO by these two effectors. To prove this, a similar series of experiments was carried out with a nbutylferricenium salt, [n-BuFc]PF6, since this coordinatively saturated organometallic dye is principally una-

Effects of interfering agents General remark The initial rate method was used to study the influence of several additives, viz. the reaction product and several potential osmium ligands, on the rate of reduction of 1 by GO(red). It was assumed that the initial rates are in better correspondence with a steady-state current measured by amperometric biosensors which are based on the Os II/III couple as an electron-transfer shuttle between electrode and oxidase.

D

-Gluconolactone
Fig. 7 Effects of L-glutainic acid (}, g) and L-histidine (L, l) on the reactivity of complex 1a (closed symbols, right vertical axis) and [n-BuFc] cPFP (open symbols, left vertical axis) in the pres6 ence of D-glucose and GO at 25 7C. Conditions: [1a] 1.21!10 P4 M, [BuFc c] 1!10 P3 M, [D-glucose] 5!10 P2 M, [GO] 1.5!10 P7 M, pH 6.7 (0.01 M MES)

The effect of the product of enzymatic oxidation of glucose studied in the concentration range 0.0­0.125 M by the example of 1b shows that the reaction rate first slightly increases and then strongly decreases. At glu-


181

ble are the be

to coordinate amino acids. The corresponding data shown in Fig. 7. As seen, similar dependencies as in case of 1a are observed. L-Glutamic acid seems to a stronger inhibitor of GO in all cases studied.

L

-Ascorbic acid

The effect of the latter was studied in the pH range 3­8 by the example of complexes 1a­c. It was found that under all the conditions investigated, ascorbic acid (0.0001­0.1 M) reduced the osmium(III) species even in the absence of D-glucose and GO, indicative of the fact that ascorbic acid is a strongly undesirable interfering agent.

Discussion
Ferrocene derivatives and osmium or ruthenium complexes are nowadays mostly frequently used as mediators of electron transport between the electrode and oxidoreductase. Therefore, it is interesting to compare their kinetic features with respect to a such widespread enzyme as GO. The key difference between ferrocenes, on the one hand, and Os and Ru compounds, on the other, is that no evidence was obtained for the kinetically meaningful complex formation between the oxidase and osmium(III) or ruthenium(III) complexes. The reduction of Os III to Os II, as well as Ru III to Ru II, follows first-order kinetics and the initial rates of formation of Os II and Ru II, accordingly, depend linearly on [Os III] or [Ru III]. Similar behavior was recently observed in the case of HRP-catalyzed oxidation of nalkylferrocenes by hydrogen peroxide where strict firstorder kinetics was observed in RFc [18]. This suggests that the Os III, or Ru III complexes do not stay long in the active site of GO and leave it as Os II or Ru II species after the electron transfer. In contrast, ferricenium cations have a much higher affinity to the GO active site, accounting for both the Michaelis-Menten kinetics and deep steady-state portions on the kinetic curves of the GO-catalyzed reduction of ferricenium cations [12]. The above feature of Os III and Ru III complexes seems to be a favorable one for catalysis, since the mediator should not affect the main task of the enzyme, i.e. to convert its natural substrate, by staying in the active site. Another attractive feature worth considering is that even when the potentially coordinatively unsaturated Os III, i.e. with two chloro ligands, complexes are used, there is no fast ligation by other ligands and, hence, the reactivity of the complexes themselves is not affected by the interfering ligands. However, the latter are capable of inhibiting the enzyme and this should always be taken into account. A common feature for Os III and RFc c is a surfactant effect. The kinetic behavior reported here is very close to what we observed for the enzymatic reduction of n-butylferricenium in the presence of CTAB, Triton

X-100, and SDS micelles [14]. In both cases, the positively charged CTAB micelles have the weakest effect on the reactivity, while the negatively charged micelles of SDS abolish the reactivity very efficiently. The absolute rates of the electron transfer from GO(red) to Os III, Ru III, and ferricenium cations [12] do not differ appreciably, if the most reactive representatives are considered. According to the data obtained from UV-vis measurements, the second-order rate constants for Os III complexes, as well as for the ferricenium cation [12], are around 1!10 5 M P1 s P1. Ruthenium complex 2 is more reactive and the second-order rate constant equals 1.8!10 5 M P1 s P1. It should be taken into account that this is an approximate estimate in the ferricenium case (kcat/KM value), since the reductions of M III and ferricenium cations follow different rate laws. The higher reactivity of Ru III compounds compared with Os III ones is a remarkable observation, taking into account the relative cost of Os and Ru compounds. This makes ruthenium derivatives better candidates for loading into various analytical biosensor devices with mediated electron transfer. In conclusion, from the kinetic point view, using ruthenium compounds in amperometric biosensors is more advantageous compared with osmium complexes and ferrocene derivatives. Even the simplest ruthenium compound 2 displays higher reactivity toward reduced GO. This automatically diminishes the effect of dioxygen on the performance of analytical devices which are based on oxidases and mediated electron transfer. In contrast to ferricenium cations, which display Michaelis kinetics toward GO, Os III or Ru III compounds show first-order reactivity toward GO(red), again providing a simple approach to enhance the rate of reoxidation of the reduced enzyme by increasing the metal complex concentration.
Acknowledgements The research described was made possible in part by financial support from the Russian Foundation for Fundamental Research (Grant No 96-03-34328a) and INTAS (Project 1432). We thank A. Tsvetkov for experimental assistance.

References
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. Ohara TJ (1995) Platinum Met Rev 39 : 54­62 Degani Y, Heller A (1989) J Am Chem Soc 111 : 2357­2358 Gregg BA, Heller A (1990) Anal Chem 62 : 258­263 Gregg BA, Heller A (1991) J Phys Chem 95 : 5976­5980 Pishko MV, Michael AC, Heller A (1991) Anal Chem 63 : 2268­2272 Ohara TJ, Rajagopalan R, Heller A (1994) Anal Chem 66 : 2451­2457 Zakeeruddin SM, Fraser DM, Nazeeruddin M-K, GrДtzel M (1992) J Electroanal Chem 337 : 253­283 Fraser DM, Zakeeruddin SM, GrДtzel M (1993) J Electroanal Chem 359 : 125­139 Yang L, Coury LA Jr, Murray RW (1993) J Phys Chem 97 : 1694­1700 Goral VN, Nelen' MI, Ryabov AD (1995) Anal Lett 28 : 2139­2149 Ryabov AD, Amon A, Gorbatova RK, Ryabova ES, Gnedenko BB (1995) J Phys Chem 99 : 14 072­14 077


182 12. Ryabov AD, Firsova YN, Nelen' MI (1996) Appl Biochem Biotechnol 61 : 25­37 13. Kober EM, Caspar JV, Sullivan BP, Meyer TJ (1988) Inorg Chem 27 : 4587­4598 14. Firsova YN, Ryabov AD (1997) Izv RAN, Ser Khim 1795­ 1801 15. Wilson R, Turner APF (1992) Biosensors Bioelectronics 7 : 165­185 16. Cusanelli A, Frey U, Richens DT, Merbach AE (1996) J Am Chem Soc 118 : 5265­5271 17. Hecht HJ, Kalisz HM, Hendle J, Schmid RD, Schomburg D (1993) J Mol Biol 229 : 153­172 18. Ryabov AD, Goral VN (1997) JBIC 2 : 182­190 19. Martinek K, Yatsimirsky AK, Levashov AV, Berezin IV (1977) In: Mittal KL (ed) Micellization, solubilization, and microemulsions, vol 2. Plenum Press, New York, pp 489­507 20. Berezin IV, Yatsimirsky AK, Martinek K (1973) Usp Khim 42 : 1729­1760 21. Nakamura T, Ogura Y (1962) J Biochem 52 : 214­220