- Email: [email protected]
Mediated amperometric determination of xylose and glucose with an immobilized aldose dehydrogenase electrode
Biosensors & Bioelectronics 7 (1992) 637-643
Mediated amperometric determination of xylose and glucose with an immobilized aldose dehydrogenase electrode* M. Smolandert V’TT,Biotechnical
Laboratory,
81 H.-L. Livio
PO Box 202, SF-02151 Espoo, Finland
L. R&hen VTT, Chemical Laboratory, PO Box 204, SF-02151 Espoo, Finland (Received 28 May 1992; revised version received 20 July 1992: accepted 21 July 1992)
Abstract: An enzyme electrode was constructed
for amperometric determination of xylose and glucose. The electrode is based on the PQQ-dependent membrane-bound aldose dehydrogenase (ALDH) from Gluconobacter o.xyduns. ALDH was covalently immobilized on a graphite electrode. Immobilized dimethylferrocene, soluble ferrocene carboxylic acid and phenazine methosulphate were used as electron transfer mediators. When xylose was measured electrochemically using an electrode modified with ALDH and dimethylferrocene. the linear measurement range extended to 100 mM. For glucose
measurement the linear measurement range was about one-tenth of that for xylose. The electrode showed fairly good stability: 50% of the original electrode response was still obtained after 5 days of intermittent use. The effect ofpossible leakage of adsorbed mediator was determined by measuring the response of an electrode with soluble mediator as a function of time. The reproducibility of the electrode was good. the standard deviation of the electrode response in ten measurements with the same electrode being only 2.7%. Keywords: biosensor. D-xylose, D-glucose.
enzyme
electrode,
of xylose
is important
in many
fields of research and industry, especially within biomass utilization. At present xylose can be analysed using HPLC or calorimetric methods. *Based on the poster presented at Biosensors Geneva, Switzerland, 20-22 May 1992. +To whom correspondence should be addressed.
‘92.
0956-5663/92/$05.00 @ 1992 Elsevier Science Publishers
dehydrogenase,
However, these methods have certain drawbacks such as high apparatus costs or unspeciticity of the method. Electrochemical methods for enzymatic xylose determination, based on the detection of NADH formed or oxygen consumed in an enzymatic reaction, have been described by Dominguez er al. (1988) and Olsson ef al. (1990). respectively. Ikeda er al. (1990) constructed a mediated carbohydrate sensor, which was sensitive also to xylose.
INTRODUCTION Determination
PQQ-dependent
Ltd.
637
M. Smolander et al.
PQQ-dependent dehydrogenases represent a new class of enzymes with great potential in biosensor technology (Davidson, 1990). The use of PQQ-dependent dehydrogenases in the determination of aldoses has been studied previously by Turner and co-workers @‘Costa et al., 1986; Turner et al., 1987) who used the PQQ-dependent glucose dehydrogenase of Acinetobacter calcoaceticus for the construction of a glucose biosensor. The use of PQQ-dependent fructose dehydrogenase in biosensors was studied by Ikeda et al. (1991). Ferrocene and its derivatives are widely studied mediators in enzyme electrode research @‘Costa et al., 1986; Degani & Heller, 1987; Foulds & Lowe, 1988; Schuhmann et al., 1991). They are non-toxic and react readily with a wide variety of redox enzymes, for example with PQQdependent dehydrogenases (Cass et al., 1985). Phenazine methosulphate (PMS) has been used as a mediator for the oxidation ofNADH (Yabuki et al., 1991) and flavoprotein glucose oxidase (Jbnsson & Gorton, 1985). PMS is also a very for efficient electron acceptor soluble quinoprotein ALDH (Buchert, 1991) and it was previously shown that PMS is among the most efficient electron transfer mediators for ALDH immobilized on the electrode surface (Smolander ef al., in press). The aim of this work was to construct a biosensor for amperometric xylose determiThis xylose- and glucose-sensitive nation. electrode is based on membrane-bound aldose dehydrogenase (ALDH) from Gluconobacter oxydans immobilized on a graphite surface. The glucose-oxidizing properties of the enzyme have been studied, e.g. by Ameyama et al. (1981) but its xylose oxidizing properties were first studied by Buchert (1991). The aldose dehydrogenase has a tightly bound coenzyme PQQ, which is reduced during_ _the _~__ catalysis of substrate oxidation. _.._ Reduced PQQ can be electrochemically regenerated with the aid of an electrochemical redox mediator. We report here the electrochemical measurement of xylose and glucose based on the oxidation of the aldoses with ALDH.
EXPERIMENTAL Gluconobacter oxydans subsp. suboxydans ATCC
621 was cultivated as described by Buchert(l991).
Biosensors& Bioelectronics
A simple, large-scale applicable method was developed for purification of the aldose dehydrogenase (ALDH) (Smolander et al., in press). ALDH was extracted from the membranes of Gluconobacter oxydans and purified chromatographically with two ionexchange steps and one gel filtration step. The electrodes were prepared by immobilizing ALDH covalently on the surface of graphite electrodes activated with 1-ethyl-3-(dimethylaminopropyl)carbodi-imide. In the immobilization procedure the ALDH solution was applied on the surface of the inverted electrode. Graphite electrodes were constructed from carbon discs (diameter 6 mm) cut out from a graphite sheet (Johnson Matthey). Some of the electrodes were modified prior to the enzyme immobilization by adsorbing 0.2 mg dimethylferrocene from toluene solution onto the electrode material. If dimethylferrocene was not added, soluble ferrocene carboxylic acid or phenazine methosulphate was used as a mediator. The concentration of soluble mediator was 1OOpM unless otherwise stated. Electrochemical experiments were performed with an EG&G 273 potentiostat in a threeelectrode cell with an Ag/AgCl reference electrode. Potassium phosphate (50 mM, pH 6.5) was used as a base electrolyte in all the experiments. Addition of aldose sugars was performed by injecting small aliquots of solutions into concentrated aldose the electrochemical cell.
RESULTS AND DISCUSSION Measurement of xylose and glucose with the soluble mediators ferrocene carboxylic acid and phenazine methosulphate When aldose dehydrogenase was covalently immobilized onto the graphite electrode surface, no current was detected after the addition of aldoses in the absence of a mediator (Fig. 1). However, in certain conditions this might be possible, since some reports of direct electron transfer between PQQ-dependent fructose dehydrogenase and a carbon paste or solid electrodes have previously been published by Ikeda et al. (1991) and Khan et al. (1991). We studied the effect of the concentration of soluble ferrocene carboxylic acid on xylose
Biosensors & Bioelectronics
Mediated amperometric determination of xylose and glucose
relative I (%) 1
I
I
1
I
1
100 200 300 400 500 ferrocenecarboxylic acid (pM) Fig. 1. Eflect of the concentration of soluble ferrocene carboxylic acid on xylose measurement with the ALDHmodified electrode.
measurement with the ALDH electrode. The maximum current response to a xylose concentration of 27 mM was obtained with a ferrocene carboxylic acid concentration of 1OOpM (Fig. 1). This was chosen as the soluble ferrocene carboxylic acid concentration for use in standard measurements. With higher concentrations of the mediator, the background current was also higher. A working electrode potential of 400 mV vs. Ag/AgCl was used, since this was the oxidation potential for soluble ferrocene carboxylic acid on the graphite electrodes used in this study. The calibration curves for xylose and glucose measurement with soluble ferrocene carboxylic acid are presented in Fig. 2. With soluble ferrocene carboxylic acid the linear measurement ranges for xylose and glucose measurement extended to about 40 mM and 3 mM, respectively. Phenazine methosulphate was also studied as a mediator. The oxidation potential of PMS on a graphite electrode is lower than that of ferrocene carboxylic acid, and hence less non-specific oxidation would be expected when PMS is used as mediator in the electrochemical detection of xylose, although it could be possible that the mediator is able to electrocatalytically oxidize some interfering compounds. In this study a working potential of 200 mV was used when PMS was the mediator. The pH optimum for xylose measurement with soluble PMS was 7 (results not shown). However, a pH of 6.5 was used in routine measurements, since the stability of the enzyme increases considerably below pH 7. The linear
10
20
30
40
50
aldose (mM) Fig. 2. Measurement of xylose and glucose with an electrode modified withALDH using soluble mediators. The concentration of ferrocene carboxylic acid (FC-COOH) and phenazine methosulphate (PMS) was 100 pM and the potential ofthe working electrode was 400 m V(FC-COOH) or 200 m V (PMS) vs. Ag/AgCl.
ranges for the aldoses were very similar to those obtained with soluble ferrocene carboxylic acid as a mediator (Fig. 2). Measurement of xylose and glucose with the dimethylferrocene-modified electrode In addition to soluble mediators, immobilized dimethylferrocene was also studied as a mediator. Dimethylferrocene can be deposited on the electrode surface from toluene solution. The use of adsorbed dimethylferrocene for the electron transfer from Acinetobacter quinoprotein glucose dehydrogenase to an electrode has been reported by D’Costa et al. (1986). The effect of the amount of immobilized dimethylferrocene on xylose measurement was almost insignificant and the amount of adsorbed dimethylferrocene in routine experiments was chosen as 0.2 mg per electrode. With higher dimethylferrocene loadings the background current was more susceptible to electrical disturbances. When glucose was measured with the dimethylferrocene-modified electrode, the effect of the amount of immobilized dimethylferrocene on the current response was stronger (Fig. 3). This could be due to the high reaction rate when glucose is used as a substrate for ALDH. The optimum working potential for an electrode modified with dimethylferrocene and 639
M. Smolander et al.
Mediated amperometric determination of xylose and glucose
relative I (%)
1(CIA)
t
8-
loo-
-A
80 -
8-
80 40 -
I
=
I
I
I
I
0,2 0,4 0,8 0,8 dlmethylferrocene (mg)
1
20
40 80 80 100 120 xylose (mM)
Fig. 3. The effect of adsorbed dimethylferrocene on glucose measurement with the ALDH electrode. The potential of the working electrode was 250 m V vs. AdAgCl.
ALDH was 250 mVvs. AgjAgCl (Fig. 4). This is in accordance with the cyclic voltammogram measured for the dimethylferrocene-modified graphite electrode. The measurement ranges for xylose and glucose were higher with dimethylferrocenemodified electrodes than those obtained using soluble mediators, the measurement range for xylose extending to 100 mM (Fig. 5). When the data of current vs. substrate concentration were plotted as a Hanes plot the maximum currents of xylose and glucose measurements could be calculated from the y-intercept of the plots (Bartlett, 1990). When the measurement data from Fig. 5 were
relatlve I (%) ,
t
1
I
I
100
200
300
E (mv) Fig. 4. The effect of thepotential of the working electrode on xylose measurement with the dimethylferrocene-modt$ed ALDH electrode.
5
10 15 glucose (mM)
20
Fig. 5. Measurement of (A) xylose and (B) glucose with the dimethylfenvcene-mod&xi ALDH electrode.
plotted as a Hanes plot the maximum currents for xylose and glucose were 19.6pA and 12*4pA, respectively (Fig. 6a). The rate-limiting step of the electrode reaction was estimated by further processing the measured data. The horizontal line obtained indicated that the rate-limiting step of the ALDH electrode was the enzyme reaction (Fig. 6b). This was as expected, since no external diffusion barrier was introduced. The Michaelis constants of the electrode were also determined from the same plot. An increase in apparent Michaelis constant due to the immobilization was observed, as could be expected because the immobilization probably causes diffusional hindrances for the substrates. The KME values of the electrode were 230 mM and 11 mM for xylose and glucose, respectively. For soluble ALDH the
Mediated amperometric determination of xylose and glucose
Biosensors & Bioelectronics
without this pretreatment. The different ~~~ values for xylose and glucose can be exploited in the glucose measurement. If the sample contains approximately equal amounts of xylose and glucose, the glucose concentration can be measured with adequate accuracy for most purposes, since the current caused by the presence of xylose in the concentration range which is linear for glucose is very small compared with that caused by glucose. In addition to xylose and glucose, the PQQdependent dehydrogenase of G. oxydans also oxidizes some other monosaccharides (Buchert, 1991; Smolander er al., in press). Hence the of biotechnical processes, e.g. control fermentations on defined aldose-containing media, would be possible with the electrode described here. Calibration curves for galactose, arabinose and mannose with the dimethylferrocene-modified electrode used are presented in Fig. 7. The calibration curve for xylose is shown as a reference.
10
5 gluiose :I I
I
20 63)
I
I
40
1
60
80
1
100
I 120
8 (mW
Functional properties of the dimethylferrocenemodified ALDH electrode
I
0,06t
The storage stability of the dimethylferrocenemodified ALDH electrode was fairly good. More than 50% of the original electrode response was still obtained after 5 days of intermittent use (Fig. 8). The response of the electrode as a function of time was also measured with a soluble mediator. When soluble PMS was used as an electron transfer mediator, the storage stability of
Mp glup= I
I
0,2 (b)
I
I-
0,4
0,6
-I..-
0,8
1
r
Fig. 6. Kinetic analysis of the data shown in Fig. 5. (a): Hanes plot of the measured data. (b): r = [i/s]/[i/s],,; [i/s],, was calculated from the y-intercept of Fig 6a. y = (r-r - 1)/s.
Michaelis constants were 44 mM for xylose and 0.7 mM for glucose. When the sample of interest contains a mixture of glucose and xylose, glucose can be removed from the sample by pretreatment with glucose oxidase (Schelleretal., 1987). However. ifxylose is used as a raw material in biotechnical processes and no glucose is present in the medium, the sensor can be used for xylose measurement
aldose
(mM)
Fig. 7. Calibration curves for galactose, arabinose and mannose. The electrode was the same as in Fig. 5. 641
M. Smolander
Biosensors & Bioelectronics
et al.
relative I
(%)
I
immobilizgd mediator
2
4
6
6
10 12 time (d)
14
16
16
20
Fig. 8. Stability of the ALDH electrode.
the electrode was similar to that of the electrode with immobilized dimethylferrocene, indicating that the amount of mediator leakage during storage was negligible. This can be expected, since dimethylferrocene has been reported to be soluble only in the oxidized form when the electrode is poised at the oxidation potential of ferrocene (Schuhmann et al., 1991). Hence ferrocene can leak out from the electrode material only during continuous measurement of aldoses. It can be concluded that small differences in the amount of the mediator caused by the intermittent measurement will have little effect on the electrode response. On the other hand, in continuous measurement, considerably faster leakage of dimethylferrocene could be expected. The reproducibility of the xylose measurement with ind&idual ALDH electrodes was good. The standard deviation of the electrode response to a xylose concentration of 20 mM in ten individual measurements was 2.7%. The response time of the ALDH electrode in xylose measurements was typically 60-120 s and in glucose measurements 20-30 s, the difference probably resulting from the different affinities between the substrates and the enzyme.
CONCLUSIONS Xylose and glucose can be measured electrochemically with the ALDH electrode. The electron transfer between the cofactor PQQ and the electrode can be mediated by dimethylferrocene 642
adsorbed on the electrode or by soluble ferrocenecarboxylic acid or phenazine methosulphate. It is possible to determine xylose and glucose concentrations up to about 100 mM and 10 mM, respectively, with the ALDH-modified electrode. If glucose and xylose are present in approximately equal amounts, the current caused by xylose oxidation is neglible compared with that caused by glucose oxidation and glucose can be measured without removing xylose. When mixtures of xylose and glucose are analysed, it is possible to remove glucose by an enzymatic treatment (Scheller et al., 1987). However, for optimized xylose determination a specific xylose dehydrogenase would be necessary. The ALDH electrode is also sensitive to galactose and arabinose and can be used for the analysis of these sugars. It has previously been observed with the soluble enzyme that the substrate specificity of ALDH depends on the electron acceptor used (Buchert, 1991). This phenomenon could possibly be utilized in a sensor construction consisting of several combinations of ALDH with different electron acceptors. The stability of the ALDH electrode was fairly good and could probably be increased if the immobilization was carried out by a method that took into account the hydrophobic nature of the membrane protein ALDH. In the conditions used during this study direct electron transfer between the graphite electrode and dehydrogenase could not be confirmed, although direct electron transfer between an electrode and a PQQ-dependent fructose de-
Biosensors & Bioelectronics
hydrogenase has been reported (1991) and Ileda et al. (1991).
Mediated amperometric determination of xylose and glucose
by Khan et al.
ACKNOWLEDGEMENTS The authors thank Anne Haanperi for her skilful technical assistance. The study was supported by the Foundation of Biotechnical and Fermentation Industry.
REFERENCES Ameyama, M., Shinagawa, E., Matsushita, K & Adachi, 0. (1981). D-glucose dehydrogenase of Gluconobacter suboxydans: solubilization, purification and characterization. Agric. Biol. Chem., 45, 851-61. Bartlett, P. N. (1990). Conducting organic salt electrodes. In: Biosensors, a Practical Approach (ed. Cass, A E. G.), Oxford University Press, 47-95. Buchert, J. (1991). A xylose-oxidizing membranebound aldose dehydrogenase of Gluconobacter oxydans, ATCC 621. J Biotech., 18, 103-14. Cass, A E. G., Davis, G., Green, M. J. & Hill, H. k 0. (1985). Ferricinium ion as an electron acceptor for oxido-reductases. J Electroanal. Chem., 190, 117-27. Davidson, V. L. (1990). Quinoproteins: a new class of enzymes with potential use as biosensors. Amer. Biotechnol. Lab., 8(2), 32-4. D’Costa, E. J., Higgins, I. J. & Turner, A. P. F. (1986).
Quinoprotein glucose dehydrogenase and its application in an amperometric glucose sensor. Biosensors, 2, 71-87.
Degani, Y. & Heller, A. (1987). Direct electrical communication between chemically modified enzymes and metal electrodes. 1. Electron transfer from glucose oxidase to metal electrodes via electron relays, bound covalently to the enzyme. J Phys. Chem., 91, 1285-9. Dominguez, E., Hahn-Hagerdahl. B., Margo-Varga, G. & Gorton, L. (1988). A flow-injection system for the amperometric determination of xylose and xylulose with co-immobilized enzymes and a modified electrode. Anal. Chem. Acta, 213, 139-50. Foulds, N. C. & Lowe, C. R. (1988). Immobilization of
glucose oxidase in ferrocene-modified pyrrole polymers. Anal. Chem., 60, 2473-8. Ikeda, T., Shibata, T., Todoriki, T. & Senda. M. (1990). Amperometric response to reducing carbohydrates of an enzyme electrode based on oligosaccharide dehydrogenase. Detection of lactose and a-amylase. Anal. Chim. Acta, 230, 75-82.
Ikeda, T., Matsushita, F. & Senda, M. (1991). Amperometric fructose sensor based on direct bioelectrocatalysis. Biosensors & Bioelectronics, 6, 299-304. Jonsson, G. & Gorton, L. (1985). An amperometric glucose sensor made by modification of a graphite electrode surface with immobilized glucose oxidase and adsorbed mediator. Biosensors, 1, 355-68. Khan, G. L., Shinohara, H., Ilcariyama, Y. & Aizawa, M. (1991). Electrochemical behaviour of monolayer quinoprotein adsorbed on the electrode surface. J Electroanal. Chem., 315, 263-73. Olsson, L., Mandenius, C. F. & Volt, J. (1990). Determination of monosaccharides in cellulosic hydrolyzates using immobilized pyronose oxidase in a continuous amperometric analyzer. Anal. Chem.. 62, 2688-91. Scheller, F. W., Pfeiffer, D., Schubert, F., Renneberg, R. & Kirstein, D. (1987). Application of enzymebased amperometric biosensors to the analysis of ‘real’ samples. In: Biosensors, Fundamentals and Applications (ed. Turner, A. F. P., Karube, I. & Wilson, G. S.), Oxford University Press, p. 337. Schuhmann. W., Loffler, U., Wohlschltiger, H., Lammert, R. Schmidt, H.-L., Wiemhofer, H.-D. & Gopel, W. (1990). Leaching of dimethylferrocene, a redox mediator in amperometric biosensors, Sens. Actuators, Bl, 571-5. Schuhmann, W., Ohara, T. J., Schmidt, H.-L. & Heller, A. (1991). Electron transfer between glucose oxidase and electrodes via redox mediators bound with flexible chains to the enzyme surface. J. Amer. Chem. Sot., 113, 1394-7. Smolander, M., Buchert, J. & Viikari, L. Large-scale applicable purification and characterization of a membrane-bound PQQ-dependent aldose dehydrogenase. A Biotech (in press). Turner, k P. F., D’Costa, E. J. & Higgins, I. J. (1987). Enzymatic analysis using quinoprotein dehydrogenases. Ann. New York Acad. Sci., 501, 283-7. Yabuki, S., Mizutani, F. & Asai, M. (1991). Preparation and characterization of an electroconductive membrane containing glutamate dehydrogenase, NADP, and mediator. Biosensors & Bioelectronics, 6, 311-5.
643
Mediated amperometric determination of xylose and glucose with an immobilized aldose dehydrogenase electrode* M. Smolandert V’TT,Biotechnical
Laboratory,
81 H.-L. Livio
PO Box 202, SF-02151 Espoo, Finland
L. R&hen VTT, Chemical Laboratory, PO Box 204, SF-02151 Espoo, Finland (Received 28 May 1992; revised version received 20 July 1992: accepted 21 July 1992)
Abstract: An enzyme electrode was constructed
for amperometric determination of xylose and glucose. The electrode is based on the PQQ-dependent membrane-bound aldose dehydrogenase (ALDH) from Gluconobacter o.xyduns. ALDH was covalently immobilized on a graphite electrode. Immobilized dimethylferrocene, soluble ferrocene carboxylic acid and phenazine methosulphate were used as electron transfer mediators. When xylose was measured electrochemically using an electrode modified with ALDH and dimethylferrocene. the linear measurement range extended to 100 mM. For glucose
measurement the linear measurement range was about one-tenth of that for xylose. The electrode showed fairly good stability: 50% of the original electrode response was still obtained after 5 days of intermittent use. The effect ofpossible leakage of adsorbed mediator was determined by measuring the response of an electrode with soluble mediator as a function of time. The reproducibility of the electrode was good. the standard deviation of the electrode response in ten measurements with the same electrode being only 2.7%. Keywords: biosensor. D-xylose, D-glucose.
enzyme
electrode,
of xylose
is important
in many
fields of research and industry, especially within biomass utilization. At present xylose can be analysed using HPLC or calorimetric methods. *Based on the poster presented at Biosensors Geneva, Switzerland, 20-22 May 1992. +To whom correspondence should be addressed.
‘92.
0956-5663/92/$05.00 @ 1992 Elsevier Science Publishers
dehydrogenase,
However, these methods have certain drawbacks such as high apparatus costs or unspeciticity of the method. Electrochemical methods for enzymatic xylose determination, based on the detection of NADH formed or oxygen consumed in an enzymatic reaction, have been described by Dominguez er al. (1988) and Olsson ef al. (1990). respectively. Ikeda er al. (1990) constructed a mediated carbohydrate sensor, which was sensitive also to xylose.
INTRODUCTION Determination
PQQ-dependent
Ltd.
637
M. Smolander et al.
PQQ-dependent dehydrogenases represent a new class of enzymes with great potential in biosensor technology (Davidson, 1990). The use of PQQ-dependent dehydrogenases in the determination of aldoses has been studied previously by Turner and co-workers @‘Costa et al., 1986; Turner et al., 1987) who used the PQQ-dependent glucose dehydrogenase of Acinetobacter calcoaceticus for the construction of a glucose biosensor. The use of PQQ-dependent fructose dehydrogenase in biosensors was studied by Ikeda et al. (1991). Ferrocene and its derivatives are widely studied mediators in enzyme electrode research @‘Costa et al., 1986; Degani & Heller, 1987; Foulds & Lowe, 1988; Schuhmann et al., 1991). They are non-toxic and react readily with a wide variety of redox enzymes, for example with PQQdependent dehydrogenases (Cass et al., 1985). Phenazine methosulphate (PMS) has been used as a mediator for the oxidation ofNADH (Yabuki et al., 1991) and flavoprotein glucose oxidase (Jbnsson & Gorton, 1985). PMS is also a very for efficient electron acceptor soluble quinoprotein ALDH (Buchert, 1991) and it was previously shown that PMS is among the most efficient electron transfer mediators for ALDH immobilized on the electrode surface (Smolander ef al., in press). The aim of this work was to construct a biosensor for amperometric xylose determiThis xylose- and glucose-sensitive nation. electrode is based on membrane-bound aldose dehydrogenase (ALDH) from Gluconobacter oxydans immobilized on a graphite surface. The glucose-oxidizing properties of the enzyme have been studied, e.g. by Ameyama et al. (1981) but its xylose oxidizing properties were first studied by Buchert (1991). The aldose dehydrogenase has a tightly bound coenzyme PQQ, which is reduced during_ _the _~__ catalysis of substrate oxidation. _.._ Reduced PQQ can be electrochemically regenerated with the aid of an electrochemical redox mediator. We report here the electrochemical measurement of xylose and glucose based on the oxidation of the aldoses with ALDH.
EXPERIMENTAL Gluconobacter oxydans subsp. suboxydans ATCC
621 was cultivated as described by Buchert(l991).
Biosensors& Bioelectronics
A simple, large-scale applicable method was developed for purification of the aldose dehydrogenase (ALDH) (Smolander et al., in press). ALDH was extracted from the membranes of Gluconobacter oxydans and purified chromatographically with two ionexchange steps and one gel filtration step. The electrodes were prepared by immobilizing ALDH covalently on the surface of graphite electrodes activated with 1-ethyl-3-(dimethylaminopropyl)carbodi-imide. In the immobilization procedure the ALDH solution was applied on the surface of the inverted electrode. Graphite electrodes were constructed from carbon discs (diameter 6 mm) cut out from a graphite sheet (Johnson Matthey). Some of the electrodes were modified prior to the enzyme immobilization by adsorbing 0.2 mg dimethylferrocene from toluene solution onto the electrode material. If dimethylferrocene was not added, soluble ferrocene carboxylic acid or phenazine methosulphate was used as a mediator. The concentration of soluble mediator was 1OOpM unless otherwise stated. Electrochemical experiments were performed with an EG&G 273 potentiostat in a threeelectrode cell with an Ag/AgCl reference electrode. Potassium phosphate (50 mM, pH 6.5) was used as a base electrolyte in all the experiments. Addition of aldose sugars was performed by injecting small aliquots of solutions into concentrated aldose the electrochemical cell.
RESULTS AND DISCUSSION Measurement of xylose and glucose with the soluble mediators ferrocene carboxylic acid and phenazine methosulphate When aldose dehydrogenase was covalently immobilized onto the graphite electrode surface, no current was detected after the addition of aldoses in the absence of a mediator (Fig. 1). However, in certain conditions this might be possible, since some reports of direct electron transfer between PQQ-dependent fructose dehydrogenase and a carbon paste or solid electrodes have previously been published by Ikeda et al. (1991) and Khan et al. (1991). We studied the effect of the concentration of soluble ferrocene carboxylic acid on xylose
Biosensors & Bioelectronics
Mediated amperometric determination of xylose and glucose
relative I (%) 1
I
I
1
I
1
100 200 300 400 500 ferrocenecarboxylic acid (pM) Fig. 1. Eflect of the concentration of soluble ferrocene carboxylic acid on xylose measurement with the ALDHmodified electrode.
measurement with the ALDH electrode. The maximum current response to a xylose concentration of 27 mM was obtained with a ferrocene carboxylic acid concentration of 1OOpM (Fig. 1). This was chosen as the soluble ferrocene carboxylic acid concentration for use in standard measurements. With higher concentrations of the mediator, the background current was also higher. A working electrode potential of 400 mV vs. Ag/AgCl was used, since this was the oxidation potential for soluble ferrocene carboxylic acid on the graphite electrodes used in this study. The calibration curves for xylose and glucose measurement with soluble ferrocene carboxylic acid are presented in Fig. 2. With soluble ferrocene carboxylic acid the linear measurement ranges for xylose and glucose measurement extended to about 40 mM and 3 mM, respectively. Phenazine methosulphate was also studied as a mediator. The oxidation potential of PMS on a graphite electrode is lower than that of ferrocene carboxylic acid, and hence less non-specific oxidation would be expected when PMS is used as mediator in the electrochemical detection of xylose, although it could be possible that the mediator is able to electrocatalytically oxidize some interfering compounds. In this study a working potential of 200 mV was used when PMS was the mediator. The pH optimum for xylose measurement with soluble PMS was 7 (results not shown). However, a pH of 6.5 was used in routine measurements, since the stability of the enzyme increases considerably below pH 7. The linear
10
20
30
40
50
aldose (mM) Fig. 2. Measurement of xylose and glucose with an electrode modified withALDH using soluble mediators. The concentration of ferrocene carboxylic acid (FC-COOH) and phenazine methosulphate (PMS) was 100 pM and the potential ofthe working electrode was 400 m V(FC-COOH) or 200 m V (PMS) vs. Ag/AgCl.
ranges for the aldoses were very similar to those obtained with soluble ferrocene carboxylic acid as a mediator (Fig. 2). Measurement of xylose and glucose with the dimethylferrocene-modified electrode In addition to soluble mediators, immobilized dimethylferrocene was also studied as a mediator. Dimethylferrocene can be deposited on the electrode surface from toluene solution. The use of adsorbed dimethylferrocene for the electron transfer from Acinetobacter quinoprotein glucose dehydrogenase to an electrode has been reported by D’Costa et al. (1986). The effect of the amount of immobilized dimethylferrocene on xylose measurement was almost insignificant and the amount of adsorbed dimethylferrocene in routine experiments was chosen as 0.2 mg per electrode. With higher dimethylferrocene loadings the background current was more susceptible to electrical disturbances. When glucose was measured with the dimethylferrocene-modified electrode, the effect of the amount of immobilized dimethylferrocene on the current response was stronger (Fig. 3). This could be due to the high reaction rate when glucose is used as a substrate for ALDH. The optimum working potential for an electrode modified with dimethylferrocene and 639
M. Smolander et al.
Mediated amperometric determination of xylose and glucose
relative I (%)
1(CIA)
t
8-
loo-
-A
80 -
8-
80 40 -
I
=
I
I
I
I
0,2 0,4 0,8 0,8 dlmethylferrocene (mg)
1
20
40 80 80 100 120 xylose (mM)
Fig. 3. The effect of adsorbed dimethylferrocene on glucose measurement with the ALDH electrode. The potential of the working electrode was 250 m V vs. AdAgCl.
ALDH was 250 mVvs. AgjAgCl (Fig. 4). This is in accordance with the cyclic voltammogram measured for the dimethylferrocene-modified graphite electrode. The measurement ranges for xylose and glucose were higher with dimethylferrocenemodified electrodes than those obtained using soluble mediators, the measurement range for xylose extending to 100 mM (Fig. 5). When the data of current vs. substrate concentration were plotted as a Hanes plot the maximum currents of xylose and glucose measurements could be calculated from the y-intercept of the plots (Bartlett, 1990). When the measurement data from Fig. 5 were
relatlve I (%) ,
t
1
I
I
100
200
300
E (mv) Fig. 4. The effect of thepotential of the working electrode on xylose measurement with the dimethylferrocene-modt$ed ALDH electrode.
5
10 15 glucose (mM)
20
Fig. 5. Measurement of (A) xylose and (B) glucose with the dimethylfenvcene-mod&xi ALDH electrode.
plotted as a Hanes plot the maximum currents for xylose and glucose were 19.6pA and 12*4pA, respectively (Fig. 6a). The rate-limiting step of the electrode reaction was estimated by further processing the measured data. The horizontal line obtained indicated that the rate-limiting step of the ALDH electrode was the enzyme reaction (Fig. 6b). This was as expected, since no external diffusion barrier was introduced. The Michaelis constants of the electrode were also determined from the same plot. An increase in apparent Michaelis constant due to the immobilization was observed, as could be expected because the immobilization probably causes diffusional hindrances for the substrates. The KME values of the electrode were 230 mM and 11 mM for xylose and glucose, respectively. For soluble ALDH the
Mediated amperometric determination of xylose and glucose
Biosensors & Bioelectronics
without this pretreatment. The different ~~~ values for xylose and glucose can be exploited in the glucose measurement. If the sample contains approximately equal amounts of xylose and glucose, the glucose concentration can be measured with adequate accuracy for most purposes, since the current caused by the presence of xylose in the concentration range which is linear for glucose is very small compared with that caused by glucose. In addition to xylose and glucose, the PQQdependent dehydrogenase of G. oxydans also oxidizes some other monosaccharides (Buchert, 1991; Smolander er al., in press). Hence the of biotechnical processes, e.g. control fermentations on defined aldose-containing media, would be possible with the electrode described here. Calibration curves for galactose, arabinose and mannose with the dimethylferrocene-modified electrode used are presented in Fig. 7. The calibration curve for xylose is shown as a reference.
10
5 gluiose :I I
I
20 63)
I
I
40
1
60
80
1
100
I 120
8 (mW
Functional properties of the dimethylferrocenemodified ALDH electrode
I
0,06t
The storage stability of the dimethylferrocenemodified ALDH electrode was fairly good. More than 50% of the original electrode response was still obtained after 5 days of intermittent use (Fig. 8). The response of the electrode as a function of time was also measured with a soluble mediator. When soluble PMS was used as an electron transfer mediator, the storage stability of
Mp glup= I
I
0,2 (b)
I
I-
0,4
0,6
-I..-
0,8
1
r
Fig. 6. Kinetic analysis of the data shown in Fig. 5. (a): Hanes plot of the measured data. (b): r = [i/s]/[i/s],,; [i/s],, was calculated from the y-intercept of Fig 6a. y = (r-r - 1)/s.
Michaelis constants were 44 mM for xylose and 0.7 mM for glucose. When the sample of interest contains a mixture of glucose and xylose, glucose can be removed from the sample by pretreatment with glucose oxidase (Schelleretal., 1987). However. ifxylose is used as a raw material in biotechnical processes and no glucose is present in the medium, the sensor can be used for xylose measurement
aldose
(mM)
Fig. 7. Calibration curves for galactose, arabinose and mannose. The electrode was the same as in Fig. 5. 641
M. Smolander
Biosensors & Bioelectronics
et al.
relative I
(%)
I
immobilizgd mediator
2
4
6
6
10 12 time (d)
14
16
16
20
Fig. 8. Stability of the ALDH electrode.
the electrode was similar to that of the electrode with immobilized dimethylferrocene, indicating that the amount of mediator leakage during storage was negligible. This can be expected, since dimethylferrocene has been reported to be soluble only in the oxidized form when the electrode is poised at the oxidation potential of ferrocene (Schuhmann et al., 1991). Hence ferrocene can leak out from the electrode material only during continuous measurement of aldoses. It can be concluded that small differences in the amount of the mediator caused by the intermittent measurement will have little effect on the electrode response. On the other hand, in continuous measurement, considerably faster leakage of dimethylferrocene could be expected. The reproducibility of the xylose measurement with ind&idual ALDH electrodes was good. The standard deviation of the electrode response to a xylose concentration of 20 mM in ten individual measurements was 2.7%. The response time of the ALDH electrode in xylose measurements was typically 60-120 s and in glucose measurements 20-30 s, the difference probably resulting from the different affinities between the substrates and the enzyme.
CONCLUSIONS Xylose and glucose can be measured electrochemically with the ALDH electrode. The electron transfer between the cofactor PQQ and the electrode can be mediated by dimethylferrocene 642
adsorbed on the electrode or by soluble ferrocenecarboxylic acid or phenazine methosulphate. It is possible to determine xylose and glucose concentrations up to about 100 mM and 10 mM, respectively, with the ALDH-modified electrode. If glucose and xylose are present in approximately equal amounts, the current caused by xylose oxidation is neglible compared with that caused by glucose oxidation and glucose can be measured without removing xylose. When mixtures of xylose and glucose are analysed, it is possible to remove glucose by an enzymatic treatment (Scheller et al., 1987). However, for optimized xylose determination a specific xylose dehydrogenase would be necessary. The ALDH electrode is also sensitive to galactose and arabinose and can be used for the analysis of these sugars. It has previously been observed with the soluble enzyme that the substrate specificity of ALDH depends on the electron acceptor used (Buchert, 1991). This phenomenon could possibly be utilized in a sensor construction consisting of several combinations of ALDH with different electron acceptors. The stability of the ALDH electrode was fairly good and could probably be increased if the immobilization was carried out by a method that took into account the hydrophobic nature of the membrane protein ALDH. In the conditions used during this study direct electron transfer between the graphite electrode and dehydrogenase could not be confirmed, although direct electron transfer between an electrode and a PQQ-dependent fructose de-
Biosensors & Bioelectronics
hydrogenase has been reported (1991) and Ileda et al. (1991).
Mediated amperometric determination of xylose and glucose
by Khan et al.
ACKNOWLEDGEMENTS The authors thank Anne Haanperi for her skilful technical assistance. The study was supported by the Foundation of Biotechnical and Fermentation Industry.
REFERENCES Ameyama, M., Shinagawa, E., Matsushita, K & Adachi, 0. (1981). D-glucose dehydrogenase of Gluconobacter suboxydans: solubilization, purification and characterization. Agric. Biol. Chem., 45, 851-61. Bartlett, P. N. (1990). Conducting organic salt electrodes. In: Biosensors, a Practical Approach (ed. Cass, A E. G.), Oxford University Press, 47-95. Buchert, J. (1991). A xylose-oxidizing membranebound aldose dehydrogenase of Gluconobacter oxydans, ATCC 621. J Biotech., 18, 103-14. Cass, A E. G., Davis, G., Green, M. J. & Hill, H. k 0. (1985). Ferricinium ion as an electron acceptor for oxido-reductases. J Electroanal. Chem., 190, 117-27. Davidson, V. L. (1990). Quinoproteins: a new class of enzymes with potential use as biosensors. Amer. Biotechnol. Lab., 8(2), 32-4. D’Costa, E. J., Higgins, I. J. & Turner, A. P. F. (1986).
Quinoprotein glucose dehydrogenase and its application in an amperometric glucose sensor. Biosensors, 2, 71-87.
Degani, Y. & Heller, A. (1987). Direct electrical communication between chemically modified enzymes and metal electrodes. 1. Electron transfer from glucose oxidase to metal electrodes via electron relays, bound covalently to the enzyme. J Phys. Chem., 91, 1285-9. Dominguez, E., Hahn-Hagerdahl. B., Margo-Varga, G. & Gorton, L. (1988). A flow-injection system for the amperometric determination of xylose and xylulose with co-immobilized enzymes and a modified electrode. Anal. Chem. Acta, 213, 139-50. Foulds, N. C. & Lowe, C. R. (1988). Immobilization of
glucose oxidase in ferrocene-modified pyrrole polymers. Anal. Chem., 60, 2473-8. Ikeda, T., Shibata, T., Todoriki, T. & Senda. M. (1990). Amperometric response to reducing carbohydrates of an enzyme electrode based on oligosaccharide dehydrogenase. Detection of lactose and a-amylase. Anal. Chim. Acta, 230, 75-82.
Ikeda, T., Matsushita, F. & Senda, M. (1991). Amperometric fructose sensor based on direct bioelectrocatalysis. Biosensors & Bioelectronics, 6, 299-304. Jonsson, G. & Gorton, L. (1985). An amperometric glucose sensor made by modification of a graphite electrode surface with immobilized glucose oxidase and adsorbed mediator. Biosensors, 1, 355-68. Khan, G. L., Shinohara, H., Ilcariyama, Y. & Aizawa, M. (1991). Electrochemical behaviour of monolayer quinoprotein adsorbed on the electrode surface. J Electroanal. Chem., 315, 263-73. Olsson, L., Mandenius, C. F. & Volt, J. (1990). Determination of monosaccharides in cellulosic hydrolyzates using immobilized pyronose oxidase in a continuous amperometric analyzer. Anal. Chem.. 62, 2688-91. Scheller, F. W., Pfeiffer, D., Schubert, F., Renneberg, R. & Kirstein, D. (1987). Application of enzymebased amperometric biosensors to the analysis of ‘real’ samples. In: Biosensors, Fundamentals and Applications (ed. Turner, A. F. P., Karube, I. & Wilson, G. S.), Oxford University Press, p. 337. Schuhmann. W., Loffler, U., Wohlschltiger, H., Lammert, R. Schmidt, H.-L., Wiemhofer, H.-D. & Gopel, W. (1990). Leaching of dimethylferrocene, a redox mediator in amperometric biosensors, Sens. Actuators, Bl, 571-5. Schuhmann, W., Ohara, T. J., Schmidt, H.-L. & Heller, A. (1991). Electron transfer between glucose oxidase and electrodes via redox mediators bound with flexible chains to the enzyme surface. J. Amer. Chem. Sot., 113, 1394-7. Smolander, M., Buchert, J. & Viikari, L. Large-scale applicable purification and characterization of a membrane-bound PQQ-dependent aldose dehydrogenase. A Biotech (in press). Turner, k P. F., D’Costa, E. J. & Higgins, I. J. (1987). Enzymatic analysis using quinoprotein dehydrogenases. Ann. New York Acad. Sci., 501, 283-7. Yabuki, S., Mizutani, F. & Asai, M. (1991). Preparation and characterization of an electroconductive membrane containing glutamate dehydrogenase, NADP, and mediator. Biosensors & Bioelectronics, 6, 311-5.
643