Aldose dehydrogenase-modified carbon paste electrodes as amperometric aldose sensors

ANALmcA CHIMICA

ACTA ELSEVIER

Analytica Chimica Acta 302 (1995) 233-240

Aldose dehydrogenase-modified carbon paste electrodes as amperometric aldose sensors Maria Smolander

aT*, Gyijrgy Marko-Varga

b, Lo Gorton b

a VT& Biotechnology and Food Research, Biologinkuja 1, P.O. Box 1500, FIN-02044 Vrr, Espoo, Finland b Department ofAnalytical Chemistry, University oflund, Received

17 May 1994; revised manuscript

P.O. Box 124, S-221 00 Lund, Sweden

received 21 September

1994

Abstract A biosensor using pyrroloquinoline quinone-dependent aldose dehydrogenase (ALDH) as a biological component was developed and used for the measurement of the aldose sugars xylose and glucose. Different immobilization methods for ALDH in carbon paste were studied. The best electrode performance was obtained when ALDH was adsorbed on the surface of a carbon paste electrode. Several mediator compounds were mixed into the carbon paste. The lowest working potential and highest catalytic current were obtained with dimethylferrocene as a mediator. Both storage and operational stability of the ALDH electrodes could be improved by the application of a membrane consisting of a poly(ester-sulfonic acid) cation-exchanger, Eastman AQ-29D. Application of the membrane reduced the non-specific oxidation of fermentation samples on the electrode surface. Keywords: Biosensors; Carbon paste electrode; PQQ-dependent dehydrogenase; D-xylose; D-glucose; Dimethylferrocenc

1. Introduction Pyrroloquinoline quinone (PQQ)-dependent enzymes, quinoproteins, have certain properties which make them especially suitable for biosensor applications [1,2]. PQQ and its apoenzyme form a stable complex and oxygen does not affect the catalytic activity of bacterial quinoproteins. Biosensors based on PQQ-dependent glucose dehydrogenase [3-71, fructose dehydrogenase [g-10] and alcohol dehydrogenase [ll, 121 have already been developed. PQQ-dependent aldose dehydrogenase (ALDH) from Gluconobacter oxydans is able to oxidize al-

* Corresponding author. 0003-2670/95/$09.50 0 1995 Elsevier Science B.V. All rights reserved SSDI 0003-2670(94)00469-2

dose sugars such as glucose, xylose, galactose, mannose and arabinose [13]. It can be easily purified with a simple large-scale applicable method [14]. During aldose oxidation the cofactor PQQ is reduced and can be electrochemically reoxidized via an electrochemically active mediator. ALDH has been applied for the amperometric determination of aldose sugars having the enzyme covalently immobilized either on a solid graphite electrode [IS] or in an immobilized enzyme reactor [ 161. Recently carbon paste electrodes have been widely investigated in biosensor technology. The entire electrode material can be chemically modified (“bulk modified electrodes”). In contrast to solid electrodes, the enzyme in carbon pastes faces an organic hydrophobic environment [17]. The carbon paste can

234

hf. Smolander et al. /Analytica

be easily modified by mixing enzymes, cofactors and/or electron transfer mediators into the paste [18-201. The electrode surface of the carbon paste electrode can be easily and reproducibly regenerated. Biological samples often contain compounds which are easily oxidized on the electrode surface causing an interfering electrochemical signal. These interferences can be minimized if a measurement -200 and 0 mV vs. standard potential between calomel electrode (SCE) can be used [21,22]. Other approaches to reduce the non-specific oxidation are preoxidation of the sample [23] and the application of a protecting membrane or polymer on the electrode surface [24,25]. The anionic polymer resin Eastman AQ-29D, which is a poly(ester-sulfonic acid), has been used to cover the electrode [19,26,27]. This polymer can easily be applied on the electrode surface as an aqueous dispersion which is converted into a water-insoluble form when dried. In this work the preparation of an ALDH-modified carbon paste electrodes was optimized and the electrodes were applied for the measurement of aldose sugars, especially glucose and xylose in a flow-injection system. The effect of the application of Eastman AQ-29D anionic polymer on the electrode surface was also studied.

2. Experimental Aldose dehydrogenase (ALDH) was purified [14] from Gluconobacter oxydans subsp. suboxydans ATCC 621 cells which were cultivated as previously described [ 131. Carbon paste electrodes were prepared as described previously [20]. First, plain graphite-oil paste (100 mg graphite powder (Fluka, cat. No. 50870) and 40 ~1 paraffin oil (Fluka)) was filled into a plastic syringe holder (1.0 ml syringe, o.d. of the tip 4.0 mm, i.d. of the tip 1.7 mm) leaving 2-3 mm of the top empty. A platinum wire was inserted in contact with the paste to bring about electrical contact between the paste and the potentiostat. The empty space was filled with modified paste prepared by mixing 100 mg heat-treated (15 s at 700°C in Muffle furnace) graphite powder (Fluka, cat. No. 50870) with 40 ~1 of paraffin oil (Fluka) in an agate mortar. Most of the electrodes were modified with an

Chimica Acta 302 (1995) 233-240

electron transfer mediator by adding the mediator compound to the graphite as dispersed solid before adding the pasting liquid. The amount of the mediator was 2 mg per 100 mg graphite. Ferrocene (Fluka), tert. -pentylferrocene, butylferrocene, dimethylferrocene, ferrocene acetic acid, ferrocene carboxaldehyde, dimethylferrocene dicarboxylate, tetracyanoquinodimethane, duroquinone and nickelocene (Aldrich), ferrocene carboxylic acid (Sigma), ferrocene dicarboxylic acid (Janssen) and hydroxymethylferrocene (Strem Chemicals) were tested as mediators. After filling the top with modified paste the electrode was rubbed on glass or smooth paper to produce a flat shining surface. ALDH solution in 10 mM sodium acetate, pH 5.0 containing 0.1% Triton X-100 was applied on the surface of a carbon paste electrode and allowed to dry at room temperature for 20 min with the electrode surface facing up. Some electrodes were prepared by mixing ALDH solution with the graphite and the mediator and allowing the enzyme to adsorb. Subsequently the mixture was allowed to dry in a desiccator above drying silica and then mixed with the pasting liquid and used as the modified paste to fill the top of the electrode as described above. After adsorption of the enzyme, the Eastman AQ29D membrane was applied on the electrode surface by dipping the electrode into a 0.5% solution of the polymer and allowing the electrode to dry for 1 h at room temperature with the electrode surface facing down. The 0.5% solution of the polymer was prepared by heating the shaken 30% commercial solution (Eastman catalogue number PM 10101, kindly obtained as a gift from Dr. W. Waeny, Eastman, Zug, Switzerland) to 90°C and diluting it with water to a final concentration of 0.5%. The amperometric measurements were performed in a flow-injection (FI) system equipped with a pneumatically controlled injection valve (Cheminert, type SVA). The detailed structure of the wall-jet type flow cell has been previously described [28]. Sodium phosphate buffer (50 mM, pH 6.5) was used as the carrier buffer throughout the work with a flow rate of 0.7 ml/min. The same carrier buffer was also used for dilution of the xylose and glucose standards and samples. The standard stock solutions of the aldoses were prepared at least 24 h before use in order to allow the aldoses to reach mutarotational

M. Smolander et al. /Analyrica

equilibrium. The potential of the working electrode was controlled with EG&G Princeton Applied Research Model 273 (USA) or Zita Electronics (Lund) potentiostats. Ag/AgCl electrodes, either homemade or from Radiometer (Model K801) and a platinum wire were used as Ihe reference and counter electrode, respectively. Unless otherwise stated the electrodes were stored in 50 mM sodium phosphate, pH 6.5 at +4”C. Samples from a fermentor cultivation of yeast on xylose were kindly provided by H. Ojamo and G. Sewell WIT, Biotechnical Laboratory). Xylose was analysed by liquid chromatography (LC) (Hewlett-Packard 1090 LC, 1037 A, Shimadzu C-R4AX Chromatopac integrator) with a Pb column (Pb’+ form, Aminex HPX-87P, 300 X 7.8 mm, BioRad) using Mini-Q water as eluent.

3. Results and discussion

3.1. Enqmatic

a

C

_

235

50nA

I

500nA

0,,1R,,i” Fig. 1. Typical peaks arising from the injection of 5 mM glucose into the ALDH FI system, working potential 200 mV vs. Ag/AgCl. (a) ALDH adsorbed on the carbon paste, no mediator. (b) ALDH mixed into the carbon paste containing 2% ferrocene carboxylic acid. cc) ALDH adsorbed on the carbon paste containing 2% ferrocene carboxylic acid.

reaction

ALDH oxidizes aldoses, whereby the corresponding lactone and the reduced form of ALDH are produced. The reaction exemplified by the oxidation of glucose is: P-D-glucose

Chimica Acta 302 (1995) 233-240

+ ALDH-PQQ

-+ b-gluconolactone

+ ALDH-PQQH,

(1)

where ALDH-PQQ and ALDH-PQQH, denote the fully oxidized and reduced forms of ALDH, respectively. To reoxidize ALDH-PQQH,, two electrons and two protons need to be donated to an acceptor. It is known that several of the most commonly used mediators in conjunction with the redox enzymes can be used [14-16,291. In some earlier papers direct electron transfer between redox enzymes and electrodes has been reported. For example direct electron transfer has previously been described for PQQ-dependent fructose dehydrogenase adsorbed on a carbon paste electrode [20]. In this study some direct electron transfer between ALDH and the carbon paste electrode could be measured when the enzyme was adsorbed on the unmodified carbon paste surface (Fig. la). The electron transfer was measured by

injecting glucose to the FI system having an ALDHmodified carbon paste electrode with no mediator in the paste. The slow increase of the current peak on the injection indicates that the reaction between the enzyme and the electrode is very slow and kinetically controlled. In order to obtain a measurable signal a potential of + 200 mV vs. Ag/AgCl had to be applied between the working and the reference electrodes. It was checked that at this potential no direct oxidation of glucose on carbon paste electrodes takes place. A dependence of both the peak height and the peak area on the glucose concentration was obtained. In comparison with fructose dehydrogenase (FDH) which can be re-oxidized already at a potential of 0 mV vs. Ag/AgCl, the direct electron transfer seemed to be much slower in the case of ALDH. This could be due to the different location of PQQ in the enzyme molecule and a greater distance for the electrons to travel to the electrode from ALDH than from FDH. The distance is very critical for the electron transfer rate: according to Degani and Heller [30] the electron transfer decreases by a factor of lo4 when the distance increases from 8 to 17 A. Different re-oxidation mechanisms of the two enzymes can also result in

M. Smolander et al. /Analytica

236

their different behaviour on the electrode surface. ALDH does not contain the heme group, whereas FDH is a heme-containing quinoprotein which may signifigantly affect the oxidation mechanism of FDH

Chimica Acta 302 (1995) 233-240 catdytlo aJRant (I-IA) so0

(al 400.

h

[311. In comparison with the direct electron transfer, the catalytic current of the ALDH electrodes on the glucose injection could be increased lOO-fold by adding a mediator compound into the carbon paste (Fig. la and c). Ferrocene derivatives have previously been used as mediators for ALDH [ 14-161 and in this work several ferrocene compounds, nickelocene, tetracyanoquinodimethane and duroquinone were investigated as mediators for the ALDH-modified carbon paste electrodes. Most of these mediator compounds were able to function as mediators at least to some extent. However nickelocene, which has previously been successfully used as a mediator for the flavoprotein glucose oxidase [22], did not seem to be a suitable mediator for the quinoprotein ALDH. Previously PQQ-dependent FDH was also found to be unable to utilize nickelocene as a mediator [32]. This could result from the incompatible standard redox potentials of PQQ and nickelocene which are + 90 mV and - 118 mV [22]. For the other mediators the catalytic response was recorded as a function of the applied potential (Fig. 2). Typically the response decreased at high working potentials, which we believe to result from the inactivation of ALDH. When dimethylferrocene-modified carbon paste was used as the electrode material the catalytic currents were high and over 90% of the maximum catalytic current could be obtained already at a relatively low working potential of +200 mV vs. Ag/AgCl. In principle + 150 mV vs. Ag/AgCl could well be used as the working potential, since 70% of the response can be obtained using this potential. Other mediators yielding high catalytic current at moderate potentials were butylferrocene and tert. -pentylferrocene. However, as dimethylferrocene was found to give the highest currents all further results reported here were obtained using this mediator. 3.2. Anomer

specificity

Previous investigations on ALDH [13-161 have not focused on the anomeric specificity of the en-

p&~ntlal (mVVI. Ag/AgCt) catalyuc currant (I-IA) 200

(b) 150 -

lW-

0

0

100

200 potenUal

300

400

SW

ml

(mV vs. AB/AgCI)

Fig. 2. Catalytic response of ALDH-modified carbon paste electrodes containing 2% of electron transfer mediator as a function of measurement potential. ALDH was adsorbed on the electrode surface and the electrodes were covered with one layer of Eastman AQ-29D polymer. The injected sample was 1 mM glucose and the injection volume was 100 ~1. (a) Dimethylferrocene (0 ), ferrocene carboxylic acid (A 1, hydroxymethylferrocene (01, butylferrocene ( * ), rerf.-pentylferrocene ( n ), ferrocene ( A ). (b) Ferrocene acetic acid (0 ), ferrocene dicarboxylic acid (A ), ferrocene carboxaldehyde CO), dimethylferrocene carboxylate (* ), TCNQ ( n ) duroquinone ( A ).

zyme. The most commonly used glucose oxidizing enzymes in biosensor applications, glucose oxidase and glucose dehydrogenase, are known only to be active on the fl-anomeric form. To get a true picture of the total glucose content in a sample, it is often necessary to add an additional enzyme, mutarotase, catalyzing the interconversion of the two anomeric forms to the sensing system. Therefore, the anomeric specificity of ALDH was studied by continuously injecting freshly dissolved solutions of (Y- and pglucose to the FI system with carbon paste electrode

231

M. Smolander et al. /Analytica Chimica Acta 302 (1995) 233-240

modified with ALDH and dimethylferrocene. After 20 min the response for 5 mM a-glucose was over 300% of the original. On the other hand the response for /?-glucose decreased to 82% of the original during the same period (data not shown). What is happening is that the dissolved pure c-u-glucose rapidly mutarotates to form the p-form and vice versa. As the phosphate buffer used for dissolving the pure anomeric forms of the sugar and the carrier in the flow system is known to catalyze the interconversion of the two anomeric forms, the process is rather rapid. These results clearly indicate that ALDH is active towards the p-anomer of glucose similarly as glucose oxidase and glucose dehydrogenase. 3.3. Immobilization Previous reports with various redox enzymes from different groups have shown that several of these enzymes can be immobilized into the organic phase of carbon paste [18-201. It was therefore investigated whether immobilization of ALDH into the carbon paste or adsorption of ALDH on the carbon paste electrode affected the response of the ALDH FI system to glucose. The highest catalytic currents were obtained when ALDH was immobilized by adsorption on the electrode surface (Fig. lb and c). The current obtained with the electrodes in which the enzyme was incorporated in the paste was only 6.7% of the highest values with ALDH on the electrode even when an approximately %-fold amount of ALDH per electrode was used for the immobilization by incorporation. A similar behaviour has also previously been reported for PQQ-dependent fructose dehydrogenase (FDH) electrodes, which gave no signal at all when the enzyme was incorporated into the carbon paste [20]. ALDH and FDH are membranebound proteins and they are both inactivated relatively easily under unfavourable conditions. The amino acid sequence in ALDH indicated that it has five hydrophobic regions, probably forming the anchor which attaches the enzyme to the membrane [33]. The rest of the protein molecule is water soluble. Hence it seems logical that the part of the enzyme involved in adsorption is in contact with the highly hydrophobic surface while the rest should be in contact with the aqueous environment for optimal activity.

3.4. Stability The stability of dimethylferrocene-modified graphite electrodes has previously been shown to be rather limited [16,34]. In this investigation the stability was studied by continuously injecting glucose in the ALDH FI system and recording the peak height during a long series of injections. The catalytic current of the dimethylferrocene-modified electrodes began to decrease immediately when the electrode was taken for use to a FI system (Fig. 3). This decrease is due either to inactivation of the enzyme or desorption of the mediator or the enzyme. The operational stability of the electrodes modified with dimethylferrocene and ALDH could be considerably improved by applying a layer of Eastman AQ-29D polymer on the electrode surface (Fig. 3). Improved electrode stability after the application of Eastman AQ-29D was previously reported by Gorton et al. [26], who used a polymer-bound mediator which was unlikely to leak. In that case the membrane probably improved the stability by preventing leakage of the enzyme which was mixed into the paste. In our case, however. the stability may also be increased for another reason. The polymer membrane is anionic and likely to interact with the dimethylferrocene mediator in its positively charged oxidized state (ferrocinium). Dimethylferrocene is normally watersoluble in its oxidized state, but electrostatic interac-

response (% from orlginal)

-iL--_ 0

5

10

15 assay number

20

25

Fig. 3. Operational stability of different dimethylfcrrocene-modified ALDH electrodes. Uncovered electrodes CO), electrodes covered with one layer of Eastman AQ-29D polymer ( A ).

238

hf. Smolander et al. /Analytica

tions between the molecule and the membrane may be able to decrease the solubilization of the mediator. In addition to the operational stability, the storage stability of the ALDH electrodes was also improved by application of the AQ-29D membrane. The storage stability was studied by recording the catalytic current of covered and uncovered electrodes at the time of preparation and over-night storage in dry state or in buffer. After covering the electrode with membrane its storage stability both in the dry state and in buffer was increased approximately 3-fold (Fig. 4). When the electrodes are stored at + 4°C it is preferable for their stability to keep them in buffer solution.

Chimica Acta 302 (1995) 233-240 oxidation cunent (nA) 100

3.5. Interferences The Eastman AQ-29D polymer has an additional advantage when the ALDH electrode is used for the measurement of real samples. It was found that the non-specific oxidation could be substantially reduced by applying only a single layer of the polymer (Fig. 5). The non-specific oxidation was investigated by

0

200 mV

9OOmV

400mV

Fig. 5. Effect of the application of Eastman AQ-29D polymer on dimethylferrocene-modified carbon paste on non-specific interferences caused by real samples. Samples from a yeast fermentation on xylose were diluted 1:lO with 50 mM sodium phosphate buffer, pH 6.5. The results for two equivalent electrodes. are shown.

injecting fermentation broth from yeast fermentation into the FI system incorporating a dimethylferrocene-modified electrode not containing ALDH (Fig. 5). Th e no n-specific oxidation taking place on the electrode depends very much on the applied potential; it increased considerably at potentials higher than + 200 mV vs. Ag/AgCl. 3.6. Measurement of aldoses

- Ea&nan AQ29D

+ Eantman AQ-29D

Fig. 4. Storage stability of dimethylferrocene-modified with and without polymer layer.

electrodes

The catalytic current as a function of aldose (xylose or glucose) concentration of the injected sample for dimethylferrocene-modified ALDH electrodes is shown in Fig. 6. The apparent MichaelisMenten constant (Kzr) and the maximum current (i,,X) were determined from the corresponding Hanes plots [35] (Fig. 7). The KipP values for xylose and glucose were 71 mM and 1.2 mM, respectively. Corresponding i,,, values were 163 nA and 361 nA. The KzP values were closer to those of the

M. Stnolander et al. /Analytica

xybae (mM) / ataiytb

ii/;, 0

( 20

40

60

80

, 100

1 120

W4

VW oatalytb

,

curmnt(nA) I

239

Chimica Acta 302 (1995) 233-240 cur18nt (nA)

t OL

0

L

I

I

I

,

20

40

60

80

100

’ 0

I 5

10

12

xyk=

I(rnaq - loa IlA K(m) - 71 mM

glucwa (mM)/catnlytb -

0-M

cutmnt (nA)

0,035

0

I 2

0

I 4

, 6 gl-

0

2

4

6 N--

0

10

12

(mM)

Fig. 6. Calibration curves for an ALDH-modified carbon electrode containing 2% dimethylferrocene as a mediator.

paste

soluble enzyme (44 mM for xylose and 0.7 mM for glucose) than the values obtained previously in other FI applications of ALDH, both when immobilized in a reactor or on solid graphite 1161. The calibration curves in Fig. 6 were also used to estimate the consumption of xylose in a yeast fermentation on xylose (Fig. 8). The results obtained with the ALDH electrode were compared with results from LC measurements with standard procedures, and a good correlation between the ALDH method and LC was obtained with samples taken up to 70 h of fermentation. Towards the end of the fermentation the results obtained with the ALDH method were slightly higher than those from LC. This was probably because the xylose concentrations of these samples were low, especially in comparison with the amount of interfering compounds which is almost certainly increased during the fermentation.

1

(mM)

I(moQ - 381 nA K(m) = 1.2 mM

Fig. 7. Hanes plots from the data in Fig. 6 for (a) xylose and (b) glucose.

memurod xybw

wncantmtbn

(Q/II

80-

60 -

40-

20 Oi

0

I

20

I

I

8

40 80 60 farmantatbn time (h)

I

loo

1

to

Fig. 8. Measurement of samples from a yeast fermentation on xylose with the ALDH-modified carbon paste electrode, ALDH electrode (0) and by LC ( A ).

240

M. Smolander et al. /Analytica

4. Conclusions In this study it was demonstrated that carbon paste electrodes can be modified with PQQ-dependent aldose dehydrogenase (ALDH). In the preparation of these electrodes, adsorption is preferred to incorporation of the enzyme into the paste. Electrons can be transferred via a mediator or directly between PQQ in the active centre of the enzyme and the electrode, the mediated electron transfer being considerably more efficient. The performance of the electrode can be improved by application of an Eastman AQ-29D polymer on the electrode surface, which improves the stability and decreases nonspecific oxidation. The electrodes appear to be useful for the determination of aldose sugars in fermentation samples.

Acknowledgements The authors thank Anne HaanpeA for skillful technical assistance and Majukka Perttula for performing the LC analyses. The financial support of Acta Chemica Scandinavica, the Foundation of Biotechnical and Fermentation Industry and the Academy of Finland, the Swedish Natural Science Research Council (NFR) and the Swedish Board for Technical and Industrial Development (NUTEK) is gratefully acknowledged.

References [II V.L. [21V.L.

Davidson, Am. Biotechnol. Lab., 8 (1990) 32. Davidson (Ed.), Principles and Applications of Quinoproteins, Dekker, New York, 1993. [31W.H. Mullen, S.J. Churchouse and P.M. Vadgama, Analyst, 110 (1985) 925. [41E.J. D’Costa, I.J. Higgins and A.P.F. Turner, Biosensors, 2 (1986) 71. 151A.P.F Turner, E.J. D’Costa and I.J. Higgins, Ann. N.Y. Acad. Sci., 501 (1987) 283. [6] K. Yokoyama, K. Sode, E. Tamiya and 1. Karube, Anal. Chim. Acta, 218 (1989) 137. [7] L. Ye, M. Himmerle, A.J. Olsthoom, W. Schuhmann, H.-L. Schmidt, J.A. Duine and A. Heller, Anal. Chem., 65 (1993) 238. [S] K. Matsumoto, 0. Hamada, H. Ukeda and Y. Osajima, Anal. Chem., 58 (1986) 2732.

Chimica Acta 30.2 (1995) 233-240 [9] T. lkeda, F. Matsushita and M. Senda, Biosensors Bioelectron., 6 (1991) 299. [lo] G.L. Khan, H. Shinohara, Y. Ikariyama and M. Aizawa, J. Electroanal. Chem., 315 (1991) 263. [ll] G. Davis, H.A.O. Hill, W.J. Aston, I.J. Higgins and A.F.P. Turner, Enzyme Microb. Technol., 5 (1983) 383. [12] Y. Kitagawa, K. Kitabatake, M. Suds, H. Muramatsu, T. Ataka, A. Mori, E. Tamiya and 1. Karube, Anal. Chem., 63 (1991) 2391. [13] J. Buchert, J. Biotechnol., 18 (1991) 103. [14] M. Smolander, J. Buchert and L. Viikari, J. Biotechnol., 29 (1993) 287-297. [15] M. Smolander, H.-L. Livio and L. RLslnen, Biosensors Bioelectron., 7 (1992) 637. [16] M. Smolander, J. Cooper, W. Schuhmann, M. HImmerle and H.-L. Schmidt, Anal. Chim. Acta, 280 (1993) 119. 1171 S. Saini, G.F. Hall, M.E.A. Downs and A.P.F. Turner, Anal. Chim. Acta, 249 (1991) 1. [18] T. lkeda, F. Matsushita and M. Senda, Agric. BioI. Chem., 54 (1990) 2919. [19] G. Bremle, B. Persson and L. Gorton, Electroanalysis, 3 (1991) 77. [20] L. Gorton, E. Dominguez, G. Marko-Varga, B. Persson, G. Jiinsson-Pettersson, E. Csaregi, K. Johansson, D. Narasaiah, S. Ghobaldi, V. Kacaniklic, T. Skotheim, P. Hale, Y. Okamoto and H.-S. Lan, in E. Pungor (Ed.), Bioelectroanalysis 2, 2nd Symposium held at Matrifiired, Hungary, Oct. 11-15, 1992, AkadCmiai Kiad6, Budapest, pp. 33-52. [21] K. Johansson, G. JGnsson-Pettersson, L. Gorton, G. MarkoVarga and E. CsGregi, J. Biotechnol., 31 (1993) 301. [22] P. Atanasov, A. Kaisheva, 1. lliev, V. Razumas and J. Kulys, Biosensors Bioelectron., 7 (1992) 361. [23] T. lkeda, 1. Katasho and M. Senda, Anal. Sci., 1 (1985) 455. [24] J.C. Cooper, M. Hiimmerle, W. Schuhmann and H.-L. Schmidt, Biosensors Bioelectron., 8 (1993) 65. [25] R.J. Geise, J.M. Adams, N.J. Barone and A.M. Yacynych, Biosensors Bioelectron., 6 (1991) 151. (261 L. Gorton, H.I. Karan, P.D. Hale, T. lnagaki, Y. Okamoto and T.A. Skotheim, Anal. Chim. Acta, 228 (1990) 23. [27] J. Wang, in P.G. Edelman and J. Wang (Eds.), Biosensors and Chemical Sensors, ACS Symposium Series, Vol. 487, ACS, Washington, DC, 1992, pp. 125-132. (281 R. Appelqvist, G. Marko-Varga, L. Gorton, A. Torstensson and G. Johansson, Anal. Chim. Acta, 169 (1985) 237. [29] P.N. Bartlett, P. Tebbutt and R.G. Whitaker, Prog. React. Kinet., 16 (1991) 55. [30] Y. Degani and A. HeIIer, J. Phys. Chem., 91 (1987) 1285. [31] J. Marcinkeviciene and G. Johansson, FEBS Lett., 318 (1993) 23. [32] L. Gorton and E. Dominguez, (1993) unpublished results. [33] A.-M. Cleton-Jansen, S. Dekker, P. van de Putte and N. Goosen, Mol. Gen. Genet., 229 (1991) 206. [34] W. Schuhmann, U. Liiffler, H. Wohlschliger, R. Lammert, H.-L. Schmidt, H.-D. Wiemh(ifer and W. Giipel, Sensors Actuators, Bl (1990) 571. [35] P.N. Bartlett, in A.E.G. Cass (Ed.), Biosensors, a Practical Approach, Oxford University Press, 47-95.