- Email: [email protected]
Bienzyme strip-type glucose sensor
Biosensors & Bioelectronics 8 (1993) 2W212
Bienzyme strip-type
glucose sensor
Jovita Marcinkeviciene Institute
of Biochemistry,
Lithuanian
& Juozas Kulys*
Academy of Sciences, Vilnius-MTP,
Mokslininku
12, Lithuania
Abstract: A strip-type glucose biosensor, prepared using screen-printing technology and comprising glucose oxidase (E.C.1.1.3.4.), peroxidase (E.C.1.1.3.13.) and ferrocyanide as mediator incorporated into graphitehydroxyethyl cellulose matrices is described. The sensor acted at 0.0 V vs Ag/AgCl electrode, and the response time was SO-60 s. The calibration was linear up to 25 mM of glucose. The sensor response was constant in the range of pH 7XW5. At 25°C the biosensor temperature coefficient was 2*7%‘C-‘. The sensor was insensitive to a physiological level of ascorbic acid (40 PM) and was used for glucose determination in whole blood. Keywords: glucose, graphite, peroxidase,
1. INTRODUCTION
In recent years new technologies for thin- and thick-film electrodes have been applied to glucose biosensors (Matthews et al., 1987; Mann-Buxbaum et al., 1990; Bilitewski et al., 1991; Scholze et al., 1991). Screen-printing, an example of such a technology, is well suited to mass production. The Exactech device developed by Medisense for blood glucose testing can be considered as the most successful application of these technologies (Matthews et al., 1987). The electrochemical mediator technology has been adapted for screenprinted glucose biosensors that have functioned at a high potential at which other active blood components, including ascorbic acid and uric acid, may also be oxidized (Scheller et al., 1991). This interference can be eliminated only by decreasing the electrode potential. With that end in view, a bienzyme system comprising peroxidase and glucose oxidase has been suggested and
* To whom correspondence
should be addressed.
glucose oxidase, ferrocyanide.
either different mediators or mediatorless systems have been investigated (Kulys et al., 1981; Kulys & Schmid, 1990,199l; Kulys et al., 1991; Scheller et al., 1991; Jonsson-Pettersson, 1991). The bienzyme mediatorless electrodes are based on graphite and response to glucose at low potential (Kulys & Schmid, 1991; Kulys et al., 1991; Jonsson-Pettersson, 1991). However, these electrodes have not been applied to the preparation of screen-printed glucose sensors because of the complex electrode matrix used and the low peroxidase-catalysed electron transfer rate in these matrices. The aim of our work was the preparation of a screen-printed mediator-type bienzyme glucose sensor and the investigation of parameters of such a sensor. 2. EXPERIMENTAL 2.1. Materials and methods Glucose oxidase (GO) (1.1.3.4.) isolated from Penicillium vitale (Ukraine) and horse radish
0956-5663/93/$06.00 @ 1993 Elsevier Science Publishers Ltd.
209
J. Marcinkeviciene
& J. Kulys
peroxidase (PO) (1.1.3.13.) (Reanal, Hungary) had activities of 40 U mgg’ and 320 U mg-r respectively. Ferrocyanide and glucose (Reachim, Russia) were of reagent grade. Hydroxyethylcellulose, ethylene glycol (Aldrich Chemical Co Ltd) and graphite powder (Fluka) were used as received. A foil of 1-O mm thickness was prepared from ‘Ketjenblack’ hydrophobized carbon black (ACKO Chemie, Amsterdam, The Netherlands) with Teflon (19.7% w/w). The carbon black was a gift from Dr I. Iliev (Sofia, Bulgaria). 2.2. Biosensor preparation The chemically modified graphite (CMG) was prepared in the following manner. 200 mg of graphite powder was mixed with 2 ml of 5 mM K,[Fe(CN),] solution in water and the solvent was allowed to evaporate in air. The enzyme solution was prepared by mixing 50 mg of PO and 10 mg of GO with O-5 ml of 2% (w/v) hydroxyethylcellulose solution in water containing 6% (w/v) ethylene glycol (Kulys & D’Costa, 1991). Then 30 mg of mediator-covered graphite powder was thoroughly mixed with O-2 ml of enzyme solution. A thin layer of chemically modified graphite was put onto the foil, using a steel screen of O-1 mm thickness, and was dried. A piece (diameter 2 mm) of modified foil was glued to the end of a glass tube, using an epoxy compound. A copper wire was glued inside the tube by silver epoxy compound to provide the electrical contact. The prepared working electrode was rinsed with 0.01 M phosphate buffer, containing O-1 M KCl, pH 7.2, and stored in a refrigerator in a dry state.
Biosensors & Bioelectronics
to be approximately 1 nA (33 nA cm-‘). On introduction of glucose, the cathodic current of the biosensor increased. The response time (to 90% stationary state) was 5W s, depending on glucose concentration. Linear calibration was achieved up to 25 mM of glucose, and the sensitivity of the biosensor was O-4 p,A cm-* rnM_l (Fig. 1). After the first calibration graph had been plotted, the biosensor was washed with buffer, kept for 10 min in buffer and recalibrated again. The sensitivity of the biosensor decreased, on average, by 32% of the first calibration. The third calibration was only a little different from the second one. To standardize the conditions of pH, temperature, ionic strength and selectivity measurement, a new biosensor was used every time. The biosensor current changed about 16% in the range of pH 5.2-7.2, but was practically constant between pH 7.2 and 8.5 (Fig. 2). At pH 7.0, the biosensor current increased slightly while the temperature increased up to a maximum of 35°C. Above this temperature, irreversible thermal inactivation of the biosensor occurred. The energy of activation, calculated according to the Arrhenius equation, was 19.7 kJ mol-t K-l, and the temperature coefficient at 25°C was 2.7% ‘C-l. The biosensor response was not sensitive to the change of ionic strength, which was achieved by varying the concentration of KC1 from 0.01 to O-2 M. Nor did the biosensor respond to a physiological level of ascorbic acid (40 PM). The biosensor appeared to be sufficiently stable
2.3. Biosensor operation The measurements were done in a two-electrode circuit using a thermostatted glass cell (10 ml) and polarograph OH-105 (Radelkis, Hungary). The buffer solution used was 0.01 M phosphate with O-1 M KCl, pH 7.2. The bioelectrode current was detected by applying a potential of O-0 V vs. Ag/AgCl, allowing the background current to decay to a steady-state value.
o! 0
3. RESULTS
AND DISCUSSION
The residual current of the GO/PO biosensor, based on CMG at O-0 V at pH 7.2, was estimated 210
10
20
30 glucose,
40 mM
Fig. 1. Dependence of bienzyme sensor response on glucose concentration. Electrode potential 0.0 V vs AglAgCl electrode, 0.01 M phosphate buffer comprising 0.1 M KCl, pH 7.2 at 25°C.
Bienzyme strip-type glucose sensor
Biosensors & Bioelectronics
.~---
0
/
/
a 3.5 I5
6
7
8
PH
9
Fig. 2. Dependence of bienzyme sensor response on solution PH. 0.01 M phosphate buffer comprising 0.1 M KU, glucose concentration 10 mM at 25°C.
good correlation between glucose concentration determined by the biosensor in whole blood and by Eksan-G in ten-fold diluted blood samples was observed (r = O-998, slope 0.983 at n = 7). The low potential of biosensors gives some advantages to such types of biosensor for eliminating the influence of interfering species in real glucose assays. A good correlation of two compared methods for the determination of glucose in blood enables us to conclude that no other including ascorbic acid, sample components, interact with the bienzyme sensor, and that the latter provides competitive limits of detection with respect to other methods, permitting production of inexpensive bioprobes.
ACKNOWLEDGEMENTS when kept dry in the refrigerator for more than 2 months. The non-reproducibility of the graphs was presumably caused by the loss of the mediator. The similarity of the second and third calibration graphs suggests that the rate of mediator loss decreased in time because the remaining mediator was more strongly adsorbed deep in the graphite matrices.
The authors wish to thank Dr I. Iliev for providing carbon black foil, Dr V. Razumas for discussions and Dr J. Piksilingaite for reading the manuscript.
REFERENCES Bilitewski, U., Ruger, P. & Schrnid, R.D. (1991). Glucose biosensors based on thick film technology. Biosens. Bioelectron., 6, 369-73.
4. EXPLANATION ACTION
Jonsson-Pettersson, G. (1991). Reagentless hydrogen peroxide and glucose sensors based on peroxidase immobilized on graphite electrodes. Electroanal.,
FOR THE BIOSENSOR
The biosensor response was in the form of cathodic current, which appeared as a result of the mediator reduction on the electrode. The process occurs according to the scheme: Glucose
+ O2 z
H202 + 2Fe(CN)$2Fe(CN)zD Fe(CN)z- + e
&gluconolactone
+
+ 2H+ PO\ 2H,O + Fe(CN)i_
A large linear calibration area (much higher than K,,, = 5.9 mM of GO as referred to in Kulys, 1989) of the biosensor, and a slight dependence of its response on the temperature and pH, permit us to conclude that the sensor acts under the diffusion-controlled conditions. The biosensor was used for a glucose determination in whole blood, using Eksan-G as a reference method. The concentration of glucose in blood samples varied from 3.2 to 23.2 mM. A
3, 741-50.
Kulys, J. (1989). Amperometric enzyme electrodes in analytical chemistry. Fresenius 2. Anal. Chem., 335, 86-91.
Kulys, J. & D’Costa, E. (1991). Printed arnperometric sensor based on TCNQ and cholinesterase. Biosens. Bioelectron., 6, 109-15.
Kulys. J. & Schmid, R.D. (1990). Mediatorless peroxidase electrode and preparation of bienzyme sensors. Bioelectrochem. Bioenerg., 24, 305-l 1.
Kulys, J. & Schmid, R.D. (1991). Bienzyme sensors based on chemically modified electrodes. Biosens. Bioelectron., 6, 438.
Kulys, J., Piesliakiene, M.V. & Samalius, A.S. (1981). The development of bienzyme glucose electrodes. Bioelectrochem. Bioenerg., 8, 81-8.
Kulys, J., Bilitewski, U. & Schmid, R.D. (1991). Robust graphite-based bienzyme sensors. Sens. Actuat. B, 3, 227-34.
Mann-Buxbaum, E., Pittner, F., Schalkhammer, T., Jachimowicz, A., Jobst, G., Olcaytug, F. & Urban, G. (1990). New microminiaturized glucose sensors using covalent immobilization Sens. Actuat. B, 1, 518-22.
techniques.
211
J. Marcinkeviciene
Matthews, D.R., Bown, E., Watson, A., Holman, R.R., Steemson, J., Hughes, S. & Scott, D. (1987). Pen-sized digital 30-second blood glucose meter. Lancet, 4, 778-9. Scheller, F.W., Schubert, F., Neumann, B., Pfeiffer, D., Hintsche, R., Dransfeld, I., Wollenberger, U., Renneberg, R., Warsinke, A., Johansson,
212
Biosensors & Bioelectronics
& .I. Kulys
G., Skoog, M., Yang, X., Bogdanovskaya, V., Buckmann, A. & Zaitsev, S. Yu. (1991). Second generation biosensors. Biosens. Bioelectron., 6, 245-53.
Scholze, J., Hampp, N. & Brauchle, C. (1991). Enzymatic hybrid biosensors. Sens. Actuat. B, 4, 211-5.
Bienzyme strip-type
glucose sensor
Jovita Marcinkeviciene Institute
of Biochemistry,
Lithuanian
& Juozas Kulys*
Academy of Sciences, Vilnius-MTP,
Mokslininku
12, Lithuania
Abstract: A strip-type glucose biosensor, prepared using screen-printing technology and comprising glucose oxidase (E.C.1.1.3.4.), peroxidase (E.C.1.1.3.13.) and ferrocyanide as mediator incorporated into graphitehydroxyethyl cellulose matrices is described. The sensor acted at 0.0 V vs Ag/AgCl electrode, and the response time was SO-60 s. The calibration was linear up to 25 mM of glucose. The sensor response was constant in the range of pH 7XW5. At 25°C the biosensor temperature coefficient was 2*7%‘C-‘. The sensor was insensitive to a physiological level of ascorbic acid (40 PM) and was used for glucose determination in whole blood. Keywords: glucose, graphite, peroxidase,
1. INTRODUCTION
In recent years new technologies for thin- and thick-film electrodes have been applied to glucose biosensors (Matthews et al., 1987; Mann-Buxbaum et al., 1990; Bilitewski et al., 1991; Scholze et al., 1991). Screen-printing, an example of such a technology, is well suited to mass production. The Exactech device developed by Medisense for blood glucose testing can be considered as the most successful application of these technologies (Matthews et al., 1987). The electrochemical mediator technology has been adapted for screenprinted glucose biosensors that have functioned at a high potential at which other active blood components, including ascorbic acid and uric acid, may also be oxidized (Scheller et al., 1991). This interference can be eliminated only by decreasing the electrode potential. With that end in view, a bienzyme system comprising peroxidase and glucose oxidase has been suggested and
* To whom correspondence
should be addressed.
glucose oxidase, ferrocyanide.
either different mediators or mediatorless systems have been investigated (Kulys et al., 1981; Kulys & Schmid, 1990,199l; Kulys et al., 1991; Scheller et al., 1991; Jonsson-Pettersson, 1991). The bienzyme mediatorless electrodes are based on graphite and response to glucose at low potential (Kulys & Schmid, 1991; Kulys et al., 1991; Jonsson-Pettersson, 1991). However, these electrodes have not been applied to the preparation of screen-printed glucose sensors because of the complex electrode matrix used and the low peroxidase-catalysed electron transfer rate in these matrices. The aim of our work was the preparation of a screen-printed mediator-type bienzyme glucose sensor and the investigation of parameters of such a sensor. 2. EXPERIMENTAL 2.1. Materials and methods Glucose oxidase (GO) (1.1.3.4.) isolated from Penicillium vitale (Ukraine) and horse radish
0956-5663/93/$06.00 @ 1993 Elsevier Science Publishers Ltd.
209
J. Marcinkeviciene
& J. Kulys
peroxidase (PO) (1.1.3.13.) (Reanal, Hungary) had activities of 40 U mgg’ and 320 U mg-r respectively. Ferrocyanide and glucose (Reachim, Russia) were of reagent grade. Hydroxyethylcellulose, ethylene glycol (Aldrich Chemical Co Ltd) and graphite powder (Fluka) were used as received. A foil of 1-O mm thickness was prepared from ‘Ketjenblack’ hydrophobized carbon black (ACKO Chemie, Amsterdam, The Netherlands) with Teflon (19.7% w/w). The carbon black was a gift from Dr I. Iliev (Sofia, Bulgaria). 2.2. Biosensor preparation The chemically modified graphite (CMG) was prepared in the following manner. 200 mg of graphite powder was mixed with 2 ml of 5 mM K,[Fe(CN),] solution in water and the solvent was allowed to evaporate in air. The enzyme solution was prepared by mixing 50 mg of PO and 10 mg of GO with O-5 ml of 2% (w/v) hydroxyethylcellulose solution in water containing 6% (w/v) ethylene glycol (Kulys & D’Costa, 1991). Then 30 mg of mediator-covered graphite powder was thoroughly mixed with O-2 ml of enzyme solution. A thin layer of chemically modified graphite was put onto the foil, using a steel screen of O-1 mm thickness, and was dried. A piece (diameter 2 mm) of modified foil was glued to the end of a glass tube, using an epoxy compound. A copper wire was glued inside the tube by silver epoxy compound to provide the electrical contact. The prepared working electrode was rinsed with 0.01 M phosphate buffer, containing O-1 M KCl, pH 7.2, and stored in a refrigerator in a dry state.
Biosensors & Bioelectronics
to be approximately 1 nA (33 nA cm-‘). On introduction of glucose, the cathodic current of the biosensor increased. The response time (to 90% stationary state) was 5W s, depending on glucose concentration. Linear calibration was achieved up to 25 mM of glucose, and the sensitivity of the biosensor was O-4 p,A cm-* rnM_l (Fig. 1). After the first calibration graph had been plotted, the biosensor was washed with buffer, kept for 10 min in buffer and recalibrated again. The sensitivity of the biosensor decreased, on average, by 32% of the first calibration. The third calibration was only a little different from the second one. To standardize the conditions of pH, temperature, ionic strength and selectivity measurement, a new biosensor was used every time. The biosensor current changed about 16% in the range of pH 5.2-7.2, but was practically constant between pH 7.2 and 8.5 (Fig. 2). At pH 7.0, the biosensor current increased slightly while the temperature increased up to a maximum of 35°C. Above this temperature, irreversible thermal inactivation of the biosensor occurred. The energy of activation, calculated according to the Arrhenius equation, was 19.7 kJ mol-t K-l, and the temperature coefficient at 25°C was 2.7% ‘C-l. The biosensor response was not sensitive to the change of ionic strength, which was achieved by varying the concentration of KC1 from 0.01 to O-2 M. Nor did the biosensor respond to a physiological level of ascorbic acid (40 PM). The biosensor appeared to be sufficiently stable
2.3. Biosensor operation The measurements were done in a two-electrode circuit using a thermostatted glass cell (10 ml) and polarograph OH-105 (Radelkis, Hungary). The buffer solution used was 0.01 M phosphate with O-1 M KCl, pH 7.2. The bioelectrode current was detected by applying a potential of O-0 V vs. Ag/AgCl, allowing the background current to decay to a steady-state value.
o! 0
3. RESULTS
AND DISCUSSION
The residual current of the GO/PO biosensor, based on CMG at O-0 V at pH 7.2, was estimated 210
10
20
30 glucose,
40 mM
Fig. 1. Dependence of bienzyme sensor response on glucose concentration. Electrode potential 0.0 V vs AglAgCl electrode, 0.01 M phosphate buffer comprising 0.1 M KCl, pH 7.2 at 25°C.
Bienzyme strip-type glucose sensor
Biosensors & Bioelectronics
.~---
0
/
/
a 3.5 I5
6
7
8
PH
9
Fig. 2. Dependence of bienzyme sensor response on solution PH. 0.01 M phosphate buffer comprising 0.1 M KU, glucose concentration 10 mM at 25°C.
good correlation between glucose concentration determined by the biosensor in whole blood and by Eksan-G in ten-fold diluted blood samples was observed (r = O-998, slope 0.983 at n = 7). The low potential of biosensors gives some advantages to such types of biosensor for eliminating the influence of interfering species in real glucose assays. A good correlation of two compared methods for the determination of glucose in blood enables us to conclude that no other including ascorbic acid, sample components, interact with the bienzyme sensor, and that the latter provides competitive limits of detection with respect to other methods, permitting production of inexpensive bioprobes.
ACKNOWLEDGEMENTS when kept dry in the refrigerator for more than 2 months. The non-reproducibility of the graphs was presumably caused by the loss of the mediator. The similarity of the second and third calibration graphs suggests that the rate of mediator loss decreased in time because the remaining mediator was more strongly adsorbed deep in the graphite matrices.
The authors wish to thank Dr I. Iliev for providing carbon black foil, Dr V. Razumas for discussions and Dr J. Piksilingaite for reading the manuscript.
REFERENCES Bilitewski, U., Ruger, P. & Schrnid, R.D. (1991). Glucose biosensors based on thick film technology. Biosens. Bioelectron., 6, 369-73.
4. EXPLANATION ACTION
Jonsson-Pettersson, G. (1991). Reagentless hydrogen peroxide and glucose sensors based on peroxidase immobilized on graphite electrodes. Electroanal.,
FOR THE BIOSENSOR
The biosensor response was in the form of cathodic current, which appeared as a result of the mediator reduction on the electrode. The process occurs according to the scheme: Glucose
+ O2 z
H202 + 2Fe(CN)$2Fe(CN)zD Fe(CN)z- + e
&gluconolactone
+
+ 2H+ PO\ 2H,O + Fe(CN)i_
A large linear calibration area (much higher than K,,, = 5.9 mM of GO as referred to in Kulys, 1989) of the biosensor, and a slight dependence of its response on the temperature and pH, permit us to conclude that the sensor acts under the diffusion-controlled conditions. The biosensor was used for a glucose determination in whole blood, using Eksan-G as a reference method. The concentration of glucose in blood samples varied from 3.2 to 23.2 mM. A
3, 741-50.
Kulys, J. (1989). Amperometric enzyme electrodes in analytical chemistry. Fresenius 2. Anal. Chem., 335, 86-91.
Kulys, J. & D’Costa, E. (1991). Printed arnperometric sensor based on TCNQ and cholinesterase. Biosens. Bioelectron., 6, 109-15.
Kulys. J. & Schmid, R.D. (1990). Mediatorless peroxidase electrode and preparation of bienzyme sensors. Bioelectrochem. Bioenerg., 24, 305-l 1.
Kulys, J. & Schmid, R.D. (1991). Bienzyme sensors based on chemically modified electrodes. Biosens. Bioelectron., 6, 438.
Kulys, J., Piesliakiene, M.V. & Samalius, A.S. (1981). The development of bienzyme glucose electrodes. Bioelectrochem. Bioenerg., 8, 81-8.
Kulys, J., Bilitewski, U. & Schmid, R.D. (1991). Robust graphite-based bienzyme sensors. Sens. Actuat. B, 3, 227-34.
Mann-Buxbaum, E., Pittner, F., Schalkhammer, T., Jachimowicz, A., Jobst, G., Olcaytug, F. & Urban, G. (1990). New microminiaturized glucose sensors using covalent immobilization Sens. Actuat. B, 1, 518-22.
techniques.
211
J. Marcinkeviciene
Matthews, D.R., Bown, E., Watson, A., Holman, R.R., Steemson, J., Hughes, S. & Scott, D. (1987). Pen-sized digital 30-second blood glucose meter. Lancet, 4, 778-9. Scheller, F.W., Schubert, F., Neumann, B., Pfeiffer, D., Hintsche, R., Dransfeld, I., Wollenberger, U., Renneberg, R., Warsinke, A., Johansson,
212
Biosensors & Bioelectronics
& .I. Kulys
G., Skoog, M., Yang, X., Bogdanovskaya, V., Buckmann, A. & Zaitsev, S. Yu. (1991). Second generation biosensors. Biosens. Bioelectron., 6, 245-53.
Scholze, J., Hampp, N. & Brauchle, C. (1991). Enzymatic hybrid biosensors. Sens. Actuat. B, 4, 211-5.