- Email: [email protected]
Direct electrochemical determination of glucose oxidase in biological samples
178,427-430
ANALYTICALBIOCHEMISTRY
(1989)
Direct Electrochemical Determination Oxidase in Biological Samples Anne Aubrhe-Lecat, Christiane Marie-H&ne Remy’ Laboratoire de Technologie BP 649,60206 CompiGgne
Received
December
Enzymatique, Cbdex, France
Hervagault,
Anne
Delacour,
U.A. No. 523 du CNRS,
Universit&
Pascale
Beaude,
Christian
Bourdillon,
and
de Compi&ze,
7,1988
An electrochemical method for the quantitation of glucose oxidase in murine plasma and tissues has been developed. Instead of oxygen, this method uses benzoquinone as an artificial cosubstrate of glucose oxidase. The quantitative detection of the enzymatically produced hydroquinone by controlled-potential amperometry allows measurement of glucose oxidase concentrations in biological samples. The use of an internal standard corrects for all possible interfering effects. We demonstrated a lo-fold increase in sensitivity, as well as the ability to work in turbid media, in comparison to spectrophotometric methods. o 1989 Academic PWS, IUC.
The potential use of enzymes for treatment of inherited metabolic diseases has attracted much interest (1,2). Replacement therapy of lysosomal storage diseasesimplies the targeting of enzymes to the correct cell type and to the specific intracellular site where the substrate accumulates as a consequence of the enzyme deficiency. Drug targeting could also be of great interest in the treatment of cancer (3,4) because the toxicity of the antimitotics:immunotoxins is much studied around the world (5) and because metastasis imaging with radiolabeled tumor-specific antibodies is a challenge in cancer diagnosis. Drug targeting would also be useful to concentrate antibiotics (6,7) in mononuclear phagocytes; many bacteria survive and even multiply after phagocytosis and are thereby protected since several classesof antibiotics, p-lactamines, and aminosides, for instance, are unable to cross cellular membranes. Recent research has attempted to improve the treatment of parasitic (8-10) and viral (11) diseasesby using this concept. The physiological exchange of macromolecules, natural or artificial, ’ To whom
of Glucose
correspondence
should
be addressed.
0003-2697/89 $3.00 Copyright 0 1989 by Academic Press. Inc. All rights of reproduction in any form reserved.
across capillary walls (12) remains a research area with many unanswered questions and has in common with the previous domains the necessity to detect and track a macromolecule inside an organism. To study the usefulness of different molecules as drug vectors, we chose to use an enzyme and not an isotope as a marker, since isotopes indicate only where the radioactive nucleus is and not the state of the molecule to which it originally belonged. Glucose oxidase was elected as a model enzyme (13) because its activity is resistant even after chemical modification, it is easily quantified in plasma, and it can be detected histoenzymatically since this activity is absent in mammals; in addition, it has been proposed as an anticancer drug (14,15). Several spectrophotometric and fluorometric techniques are based on the reaction of the hydrogen peroxide formed by glucose oxidase with a suitable chromogen (16). These methods allow measurement of glucose oxidase activity in plasma but not in tissues because of the turbidity of these media. Glucose oxidase activity can also be detected by following oxygen consumption (17) or acid production since the gluconolactone produced by the enzyme hydrolyzes spontaneously (more rapidly at alkaline pH) to give gluconic acid (18). To fulfill our requirements, we needed a method more sensitive than any currently available and that was able to quantify low amounts of glucose oxidase activity in both plasma and crushed tissues, i.e., in turbid media. In this paper, we describe an electrochemical method that meets this description. Although glucose oxidase is extremely specific with respect to the sugar, i.e., /?-D-glUCOSe, it is well known that oxygen, the usual cosubstrate, can be replaced by other electron acceptors, of which benzoquinone is very convenient (19,20). The detection of the reduced artificial cosubstrate by controlled-potential amperometry allows determination of the enzymatic reaction rate, and the knowledge of kinetic constants leads to an absolute de427
428
AUBRkE-LECAT
termination of the number of enzymatic molecules in the sample. After injection of glucose oxidase, we used this technique to study the evolution of plasma glucose oxidase concentration and the enzyme content of different organs. MATERIALS
AND
METHODS
Materials Glucose oxidase (EC 1.1.3.4) from Aspergillus niger was acquired from Boehringer (grade I). All the reactants used were of analytical grade. For safety reasons, benzoquinone was weighed inside an aspirating hood. Biological Model Swiss mice (25-40 g), provided with water and laboratory chow ad Zibitum, were used.
ET
AL.
tion was kept under a flow of argon and was thermostated at 20 + 0.5”C. (c) Measurements. The glucose and benzoquinone concentrations in the electrochemical cell were respectively 0.5 M and 5 mM. In order to buffer the acid produced during the oxidation of hydroquinone and the spontaneous hydrolysis of gluconolactone, 0.1 M phosphate buffer at the optimum pH of glucose oxidase (pH 6.5) was used. One hundred microliters of the sample solution was added to the bulk. The difference in the slopes with and without glucose oxidase is directly proportional to the enzyme concentration. After each kinetic measurement, a known amount of hydroquinone was added as an internal standard in order to calibrate each measurement with the actual conditions prevailing in the electrochemical cell. (II) Spectrofluorometric Measurement
Measurement of Glucose Oxidase Concentration (1) Electrochemical Measurement (a) Enzymatic and electrochemical reactions. The enzymatic reaction, oxidation of glucose into gluconolactone and reduction of benzoquinone to give hydroquinone, occurs in the bulk. The second reaction, which is electrochemical, hydroquinone + benzoquinone + 2H+ + 2e-, proceeds only at the surface of the working electrode. Under controlled-potential conditions, the current due to the electrons released at the working electrode is proportional to the hydroquinone concentration in the bulk and is consequently related to the enzyme activity. (b) Electrochemical apparatus. The apparatus is an anaerobic electrochemical cell including a classical three-electrode system: a calomel-saturated KC1 electrode (WE)’ as a reference electrode; an auxiliary electrode (platinum foil) separated from the main compartment by a glass frit; and the working electrode, which is a disk electrode rotated at 1250 rpm by a motor (ED1 from Tacussel, Neuilly-Plaisance, France). The disk electrode was made from a 3-mm-diameter glassy carbon cylinder (Tokai GC 20) glued in a plexiglass sleeve. The surface was polished with alumina powders of decreasing size down to 50 nm and cleaned with distilled water. This treatment gave a suitable and reproducible surface. Between assays, the surface of the electrode was cleaned with sandpaper in order to eliminate proteins adsorbed on the surface. Experiments were performed with an amperometric unit (PRG-DEL from Tacussel). The working potential for hydroquinone detection was 350 mV/SCE. The solu’ Abbreviations used: SCE, calomel-saturated electrode; HPPA, (p-hydroxyphenyl)propionic acid; HRP, horseradish peroxidase.
3-
Horseradish peroxidase (HRP) catalyzes the oxidation of a wide variety of hydrogen-donating substrates with hydrogen peroxide. Among them, 3-(p-hydroxyphenyl)propionic acid (HPPA) was described by Zaitsu and Ohkura (21) as the best fluorogenic substrate for providing sensitive and rapid assay procedures of hydrogen peroxide. Their protocol was adapted for the measurement of glucose oxidase. Light intensities were measured at 20°C with a Perkin-Elmer LS 5 spectrofluorometer using quartz cells of 1 X l-cm optical path lengths. The fluorescence intensity was measured at 404 t 5 nm with the excitation at 320 + 5 nm. The assay procedure was as follows: the glucose oxidase test solution was added to a mixture of 20 ~1 of 10 mM HPPA solution, 2.9 ml of 42 mM glucose solution in 0.1 M Tris-HCl buffer, pH 8.5, and 20 ~1 of 20 UI ml-’ HRP solution. To test the interaction with serum, 5 to 50 ~1 of serum was added to the cuvette before the introduction of glucose oxidase. In Vivo Experiments Two hundred fifty micrograms (75 IU) of enzyme dissolved in 0.1 ml of physiologic serum containing 3 IU heparin (to facilitate sampling) was injected in a lateral tail vein. At predetermined times, 30 to 40 /*l of blood was sampled and centrifuged in Eppendorf microtubes. Ten microliters of plasma was taken, diluted in 0.1 ml of distilled water, and kept in an ice bath. At the time of measurement, the sample was introduced into the electrochemical cell. At the end of each experiment, the mouse was killed by spinal elongation and the sampled organs were immediately immersed in refrigerated physiologic serum. Each organ was then dried and weighed. One hundred
ELECTROCHEMICAL
DETERMINATION
OF
GLUCOSE
429
OXIDASE
milligrams was crushed by an Ultraturax (four times for 2 s) in 2 ml of distilled water and then centrifuged. The supernatant was used for kinetic assays, which were performed with the same protocol as that for plasma assays. RESULTS
AND
DISCUSSION
For pharmacological studies described elsewhere (13), we chose glucose oxidase as a marker and therefore needed to quantify low activities in biological samples. The replacement of oxygen as a cosubstrate of glucose oxidase by an artificial electron acceptor, benzoquinone, which is electroreactive in its reduced form, allows the detection of enzymatic activity by controlled-potential amperometry. Other potential artificial cosubstrates are not convenient because they are not stable (e.g., ferricinium) or because they exhibit first-order kinetics over a large concentration range (e.g., ferricyanide). Controlled-potential amperometry offers the advantage over cyclic voltamperometry of a sensitivity increase of at least lo-fold. In the final procedure, the benzoquinone concentration (5 mM or 3K,) is a compromise between a high value, in order to approximate zero-order enzyme kinetics and to avoid rate variation due to cosubstrate consumption, and a low value, in order to minimize background noise due to spontaneous benzoquinone evolution. The chosen glucose concentration is 5K, (0.5 M). At higher values, one observes an excess glucose inhibition. As can be seen in Fig. 1, a slow redox reaction occurs between glucose and benzoquinone. For the chosen glucose and benzoquinone concentrations, this reaction is not negligible and gives a current of about 60 nA min-‘. In order to quantify the hydroquinone produced by the enzyme activity, the measurement should be carried out in several steps (Fig. 1). First, the electrode is stabilized in the benzoquinone solution. Glucose is then added. After a few minutes of chemical reaction, the enzyme sample to be tested is introduced into the cell. The slope difference is the consequence of the enzyme activity. In order to calibrate each measurement with the conditions prevailing in the electrochemical cell at that time, a known amount of hydroquinone is added as an internal standard. The step increase in the current produced by the hydroquinone addition is proportional to the hydroquinone concentration increase in the cell under conditions identical to those prevailing during the measurement. Since the chemical reaction gives a current of about 60 nA min-I, we considered as significant only those results that were 10% above this value. A current of 6 nA min-’ is given by 0.4 ng ml-’ of native enzyme or by 10 ~1 of plasma containing 0.2 fig ml-l of glucose oxidase. The sensitivity is 10 times that which we obtained spectrophotometrically using iodonitrotetrazolium. For 10 mea-
I
I
5
I,
10 15
20
25
30 TIME
(minute)
FIG. 1. Experimental curve of current (in @A) collected at the working electrode as a function of time during actual measurement of a sample. (1) The measurement is begun by introducing 1.4 ml of 0.1 M phosphate buffer (pH 6.5) and 2.5 ml of 10 mM benzoquinone dissolved in the same buffer into the electrochemical cell. (2) After current stabilization, 1 ml of 2.5 M glucose dissolved in the same phosphate buffer is added. The signal variation is a consequence of the dilution produced by the glucose solution. (3) When a response sufficient to determine the slope of the reaction between glucose and benzoquinone is obtained, the glucose oxidase sample (100 ~1) is introduced into the electrochemical cell. (4) Fifty microliters of 10 mM hydroquinone is added to calibrate the kinetics. This addition produces a step increase in the current.
surements, the coefficient of variation was 5.4% for a glucose oxidase concentration of 0.11.18 ml-’ in the electrochemical cell, which gives an initial rate of 1.6 PA min-l. The injection of glucose oxidase into the cell when one of the substrates is absent, or the injection of centrifuged hemolyzed blood or of crushed tissues, does not give any electrochemically detectable activity. The sensitivity of the spectrofluorometric measurement is about the same as that of the electrochemical method for transparent solutions. However, quenching occurs after serum addition to the cuvette and there is background activity with hemolyzed blood, probably due to the peroxidase activity of hemoglobin. Moreover, the reaction begins only after a delay proportional to the serum concentration when the glucose oxidase concentration is kept constant. For 10 ~1 of serum in 3 ml of a 3.37 ng ml-l enzyme solution, this delay is 38 min. The method described here has been developed to quantify glucose oxidase activity in plasma and crushed tissues. The obtainment with this method of a lo-fold increase in sensitivity has allowed us to follow for a longer time the plasma concentration of glucose oxidase after intravenous injection of a smaller amount than has previously been reported (13). Quantification of enzyme concentrations after intravenous injection was performed in liver, spleen, gut, and muscle, and the results
430
AUBREE-LECAT
confirmed those that (13). Electrochemical plementary techniques; ization and a very high ical detection permits
previously obtained by histology detection and histology are comhistology allows a precise localsensitivity, whereas electrochemquantitation of the results.
ACKNOWLEDGMENTS This work was supported by External Research Grant 861006 from the Institut National de la Sante et de la Recherche Mddicale (INSERM) and by financial support from the Comiti de I’Oise de la Ligue Nationale de Lutte contre le Cancer. The authors thank Dominique Domurado for helping with this paper and Catherine Briasco for reviewing the manuscript.
REFERENCES 1. Desnick, R. J., Thorpe, Rev. 56,57-99. 2. Brady,
R. 0. (1983)
3. Mathe, G., Tran, 246,1626-1628.
S. R., and Fiddler,
Pharmacol.
Ther.
B. L., and
Bernard,
S., and Pihl,
6. Vladimirsky, 375-377.
A. (1981)
J. (1958)
Phurmacol.
M. A., and Ladigina,
7. Bakker-Woudenberg, Regts, D., and Michel,
Physiol.
19,327-336.
4. Monsigny, M., Roche, A.-C., and Bailly, phys. Res. Commun. 121.579-584. 5. Olsnes,
M. B. (1976)
P. (1984) Ther.
G. A. (1982)
C. R. Acad.
Sci.
Biochem.
Bio-
15.355-381. Biomedicine
36,
I. A. J. M., Lokerse, A. F., Roerdink, F. H., M. F. (1985) J. Infect. Dis. 151,917-924.
ET
AL.
8. Jadin, J. M., Trouet, A., Van Hoof, F., Bioul-Marchand, M., Maldague, P., and Jadin-Nyssens, M. (1977) Ann. Sot. Be&s Med. Trop. 57,525-530. 9. Alving, C. R., and Steek, E. A. (1979) Trends Biochem. Sci. 4, N175-N177. 10. Trouet, A., Pirson, P., Steiger, R., Masquelier, M., Baurain, R., and Gillet, J. (1981) Bull. W.H.O. 59,449-458. 11. Fiume, L., Busi, C., Mattioli, A., Balboni, P. G., and BarbantiBrodano, G. (1981) FEBS Lett. 129,261-264. 12. Taylor, A. E., and Ganger, D. N. (1984) in Handbook of Physiology, Section 2, The Cardiovascular System, Vol. IV: The Microcirculation (Renkin, E. M., and Michel, C. C., Eds.), pp. 467-520, Amer. Physiol. Sot., Washington, DC. 13. Demignot, S., and Domurado, D. (1987) Drug Des. Deb 1,333348. 14. Nakane, P. K., and Watanabe, T. (1984) J. H&o&em. Cytochem. 32,894-898. 15. Muzykantov, V. R., Sakharov, D. V., Sinitsyn, V. V., Domogatsky, S. P., Goncharov, N. V., and Danilov, S. MT (1988) Anal. B&he&
169,383-389. 16. Clark, C. A., Downs, E. C., and Primus, F. J. (1982) J. Histochem. Cytochem. 30,27-34. 17. Weibel, M. K., and Bright, H. J. (1971) J. Biol. Chem. 245,27342744. 18. Lecoq, D. (1972) Influence de la Structure sur le Comportement Cinitique de Systemes Multienzymatiques Siquentiels, p. 55, Thesis (3rd cycle), Rouen. 19. Schlapfer, P., Minot, W., and Racine, P. (1974) Clin. Chim. Acta 57,283-289. 20. Bourdillon, C., Hervagault, C., and Thomas, D. (1985) Biotechnol. Bioeng. 27,1619-1622. 21. Zaitsu, K., and Ohkura, Y. (1980) Anal. Biochem. 109.109-113.
ANALYTICALBIOCHEMISTRY
(1989)
Direct Electrochemical Determination Oxidase in Biological Samples Anne Aubrhe-Lecat, Christiane Marie-H&ne Remy’ Laboratoire de Technologie BP 649,60206 CompiGgne
Received
December
Enzymatique, Cbdex, France
Hervagault,
Anne
Delacour,
U.A. No. 523 du CNRS,
Universit&
Pascale
Beaude,
Christian
Bourdillon,
and
de Compi&ze,
7,1988
An electrochemical method for the quantitation of glucose oxidase in murine plasma and tissues has been developed. Instead of oxygen, this method uses benzoquinone as an artificial cosubstrate of glucose oxidase. The quantitative detection of the enzymatically produced hydroquinone by controlled-potential amperometry allows measurement of glucose oxidase concentrations in biological samples. The use of an internal standard corrects for all possible interfering effects. We demonstrated a lo-fold increase in sensitivity, as well as the ability to work in turbid media, in comparison to spectrophotometric methods. o 1989 Academic PWS, IUC.
The potential use of enzymes for treatment of inherited metabolic diseases has attracted much interest (1,2). Replacement therapy of lysosomal storage diseasesimplies the targeting of enzymes to the correct cell type and to the specific intracellular site where the substrate accumulates as a consequence of the enzyme deficiency. Drug targeting could also be of great interest in the treatment of cancer (3,4) because the toxicity of the antimitotics:immunotoxins is much studied around the world (5) and because metastasis imaging with radiolabeled tumor-specific antibodies is a challenge in cancer diagnosis. Drug targeting would also be useful to concentrate antibiotics (6,7) in mononuclear phagocytes; many bacteria survive and even multiply after phagocytosis and are thereby protected since several classesof antibiotics, p-lactamines, and aminosides, for instance, are unable to cross cellular membranes. Recent research has attempted to improve the treatment of parasitic (8-10) and viral (11) diseasesby using this concept. The physiological exchange of macromolecules, natural or artificial, ’ To whom
of Glucose
correspondence
should
be addressed.
0003-2697/89 $3.00 Copyright 0 1989 by Academic Press. Inc. All rights of reproduction in any form reserved.
across capillary walls (12) remains a research area with many unanswered questions and has in common with the previous domains the necessity to detect and track a macromolecule inside an organism. To study the usefulness of different molecules as drug vectors, we chose to use an enzyme and not an isotope as a marker, since isotopes indicate only where the radioactive nucleus is and not the state of the molecule to which it originally belonged. Glucose oxidase was elected as a model enzyme (13) because its activity is resistant even after chemical modification, it is easily quantified in plasma, and it can be detected histoenzymatically since this activity is absent in mammals; in addition, it has been proposed as an anticancer drug (14,15). Several spectrophotometric and fluorometric techniques are based on the reaction of the hydrogen peroxide formed by glucose oxidase with a suitable chromogen (16). These methods allow measurement of glucose oxidase activity in plasma but not in tissues because of the turbidity of these media. Glucose oxidase activity can also be detected by following oxygen consumption (17) or acid production since the gluconolactone produced by the enzyme hydrolyzes spontaneously (more rapidly at alkaline pH) to give gluconic acid (18). To fulfill our requirements, we needed a method more sensitive than any currently available and that was able to quantify low amounts of glucose oxidase activity in both plasma and crushed tissues, i.e., in turbid media. In this paper, we describe an electrochemical method that meets this description. Although glucose oxidase is extremely specific with respect to the sugar, i.e., /?-D-glUCOSe, it is well known that oxygen, the usual cosubstrate, can be replaced by other electron acceptors, of which benzoquinone is very convenient (19,20). The detection of the reduced artificial cosubstrate by controlled-potential amperometry allows determination of the enzymatic reaction rate, and the knowledge of kinetic constants leads to an absolute de427
428
AUBRkE-LECAT
termination of the number of enzymatic molecules in the sample. After injection of glucose oxidase, we used this technique to study the evolution of plasma glucose oxidase concentration and the enzyme content of different organs. MATERIALS
AND
METHODS
Materials Glucose oxidase (EC 1.1.3.4) from Aspergillus niger was acquired from Boehringer (grade I). All the reactants used were of analytical grade. For safety reasons, benzoquinone was weighed inside an aspirating hood. Biological Model Swiss mice (25-40 g), provided with water and laboratory chow ad Zibitum, were used.
ET
AL.
tion was kept under a flow of argon and was thermostated at 20 + 0.5”C. (c) Measurements. The glucose and benzoquinone concentrations in the electrochemical cell were respectively 0.5 M and 5 mM. In order to buffer the acid produced during the oxidation of hydroquinone and the spontaneous hydrolysis of gluconolactone, 0.1 M phosphate buffer at the optimum pH of glucose oxidase (pH 6.5) was used. One hundred microliters of the sample solution was added to the bulk. The difference in the slopes with and without glucose oxidase is directly proportional to the enzyme concentration. After each kinetic measurement, a known amount of hydroquinone was added as an internal standard in order to calibrate each measurement with the actual conditions prevailing in the electrochemical cell. (II) Spectrofluorometric Measurement
Measurement of Glucose Oxidase Concentration (1) Electrochemical Measurement (a) Enzymatic and electrochemical reactions. The enzymatic reaction, oxidation of glucose into gluconolactone and reduction of benzoquinone to give hydroquinone, occurs in the bulk. The second reaction, which is electrochemical, hydroquinone + benzoquinone + 2H+ + 2e-, proceeds only at the surface of the working electrode. Under controlled-potential conditions, the current due to the electrons released at the working electrode is proportional to the hydroquinone concentration in the bulk and is consequently related to the enzyme activity. (b) Electrochemical apparatus. The apparatus is an anaerobic electrochemical cell including a classical three-electrode system: a calomel-saturated KC1 electrode (WE)’ as a reference electrode; an auxiliary electrode (platinum foil) separated from the main compartment by a glass frit; and the working electrode, which is a disk electrode rotated at 1250 rpm by a motor (ED1 from Tacussel, Neuilly-Plaisance, France). The disk electrode was made from a 3-mm-diameter glassy carbon cylinder (Tokai GC 20) glued in a plexiglass sleeve. The surface was polished with alumina powders of decreasing size down to 50 nm and cleaned with distilled water. This treatment gave a suitable and reproducible surface. Between assays, the surface of the electrode was cleaned with sandpaper in order to eliminate proteins adsorbed on the surface. Experiments were performed with an amperometric unit (PRG-DEL from Tacussel). The working potential for hydroquinone detection was 350 mV/SCE. The solu’ Abbreviations used: SCE, calomel-saturated electrode; HPPA, (p-hydroxyphenyl)propionic acid; HRP, horseradish peroxidase.
3-
Horseradish peroxidase (HRP) catalyzes the oxidation of a wide variety of hydrogen-donating substrates with hydrogen peroxide. Among them, 3-(p-hydroxyphenyl)propionic acid (HPPA) was described by Zaitsu and Ohkura (21) as the best fluorogenic substrate for providing sensitive and rapid assay procedures of hydrogen peroxide. Their protocol was adapted for the measurement of glucose oxidase. Light intensities were measured at 20°C with a Perkin-Elmer LS 5 spectrofluorometer using quartz cells of 1 X l-cm optical path lengths. The fluorescence intensity was measured at 404 t 5 nm with the excitation at 320 + 5 nm. The assay procedure was as follows: the glucose oxidase test solution was added to a mixture of 20 ~1 of 10 mM HPPA solution, 2.9 ml of 42 mM glucose solution in 0.1 M Tris-HCl buffer, pH 8.5, and 20 ~1 of 20 UI ml-’ HRP solution. To test the interaction with serum, 5 to 50 ~1 of serum was added to the cuvette before the introduction of glucose oxidase. In Vivo Experiments Two hundred fifty micrograms (75 IU) of enzyme dissolved in 0.1 ml of physiologic serum containing 3 IU heparin (to facilitate sampling) was injected in a lateral tail vein. At predetermined times, 30 to 40 /*l of blood was sampled and centrifuged in Eppendorf microtubes. Ten microliters of plasma was taken, diluted in 0.1 ml of distilled water, and kept in an ice bath. At the time of measurement, the sample was introduced into the electrochemical cell. At the end of each experiment, the mouse was killed by spinal elongation and the sampled organs were immediately immersed in refrigerated physiologic serum. Each organ was then dried and weighed. One hundred
ELECTROCHEMICAL
DETERMINATION
OF
GLUCOSE
429
OXIDASE
milligrams was crushed by an Ultraturax (four times for 2 s) in 2 ml of distilled water and then centrifuged. The supernatant was used for kinetic assays, which were performed with the same protocol as that for plasma assays. RESULTS
AND
DISCUSSION
For pharmacological studies described elsewhere (13), we chose glucose oxidase as a marker and therefore needed to quantify low activities in biological samples. The replacement of oxygen as a cosubstrate of glucose oxidase by an artificial electron acceptor, benzoquinone, which is electroreactive in its reduced form, allows the detection of enzymatic activity by controlled-potential amperometry. Other potential artificial cosubstrates are not convenient because they are not stable (e.g., ferricinium) or because they exhibit first-order kinetics over a large concentration range (e.g., ferricyanide). Controlled-potential amperometry offers the advantage over cyclic voltamperometry of a sensitivity increase of at least lo-fold. In the final procedure, the benzoquinone concentration (5 mM or 3K,) is a compromise between a high value, in order to approximate zero-order enzyme kinetics and to avoid rate variation due to cosubstrate consumption, and a low value, in order to minimize background noise due to spontaneous benzoquinone evolution. The chosen glucose concentration is 5K, (0.5 M). At higher values, one observes an excess glucose inhibition. As can be seen in Fig. 1, a slow redox reaction occurs between glucose and benzoquinone. For the chosen glucose and benzoquinone concentrations, this reaction is not negligible and gives a current of about 60 nA min-‘. In order to quantify the hydroquinone produced by the enzyme activity, the measurement should be carried out in several steps (Fig. 1). First, the electrode is stabilized in the benzoquinone solution. Glucose is then added. After a few minutes of chemical reaction, the enzyme sample to be tested is introduced into the cell. The slope difference is the consequence of the enzyme activity. In order to calibrate each measurement with the conditions prevailing in the electrochemical cell at that time, a known amount of hydroquinone is added as an internal standard. The step increase in the current produced by the hydroquinone addition is proportional to the hydroquinone concentration increase in the cell under conditions identical to those prevailing during the measurement. Since the chemical reaction gives a current of about 60 nA min-I, we considered as significant only those results that were 10% above this value. A current of 6 nA min-’ is given by 0.4 ng ml-’ of native enzyme or by 10 ~1 of plasma containing 0.2 fig ml-l of glucose oxidase. The sensitivity is 10 times that which we obtained spectrophotometrically using iodonitrotetrazolium. For 10 mea-
I
I
5
I,
10 15
20
25
30 TIME
(minute)
FIG. 1. Experimental curve of current (in @A) collected at the working electrode as a function of time during actual measurement of a sample. (1) The measurement is begun by introducing 1.4 ml of 0.1 M phosphate buffer (pH 6.5) and 2.5 ml of 10 mM benzoquinone dissolved in the same buffer into the electrochemical cell. (2) After current stabilization, 1 ml of 2.5 M glucose dissolved in the same phosphate buffer is added. The signal variation is a consequence of the dilution produced by the glucose solution. (3) When a response sufficient to determine the slope of the reaction between glucose and benzoquinone is obtained, the glucose oxidase sample (100 ~1) is introduced into the electrochemical cell. (4) Fifty microliters of 10 mM hydroquinone is added to calibrate the kinetics. This addition produces a step increase in the current.
surements, the coefficient of variation was 5.4% for a glucose oxidase concentration of 0.11.18 ml-’ in the electrochemical cell, which gives an initial rate of 1.6 PA min-l. The injection of glucose oxidase into the cell when one of the substrates is absent, or the injection of centrifuged hemolyzed blood or of crushed tissues, does not give any electrochemically detectable activity. The sensitivity of the spectrofluorometric measurement is about the same as that of the electrochemical method for transparent solutions. However, quenching occurs after serum addition to the cuvette and there is background activity with hemolyzed blood, probably due to the peroxidase activity of hemoglobin. Moreover, the reaction begins only after a delay proportional to the serum concentration when the glucose oxidase concentration is kept constant. For 10 ~1 of serum in 3 ml of a 3.37 ng ml-l enzyme solution, this delay is 38 min. The method described here has been developed to quantify glucose oxidase activity in plasma and crushed tissues. The obtainment with this method of a lo-fold increase in sensitivity has allowed us to follow for a longer time the plasma concentration of glucose oxidase after intravenous injection of a smaller amount than has previously been reported (13). Quantification of enzyme concentrations after intravenous injection was performed in liver, spleen, gut, and muscle, and the results
430
AUBREE-LECAT
confirmed those that (13). Electrochemical plementary techniques; ization and a very high ical detection permits
previously obtained by histology detection and histology are comhistology allows a precise localsensitivity, whereas electrochemquantitation of the results.
ACKNOWLEDGMENTS This work was supported by External Research Grant 861006 from the Institut National de la Sante et de la Recherche Mddicale (INSERM) and by financial support from the Comiti de I’Oise de la Ligue Nationale de Lutte contre le Cancer. The authors thank Dominique Domurado for helping with this paper and Catherine Briasco for reviewing the manuscript.
REFERENCES 1. Desnick, R. J., Thorpe, Rev. 56,57-99. 2. Brady,
R. 0. (1983)
3. Mathe, G., Tran, 246,1626-1628.
S. R., and Fiddler,
Pharmacol.
Ther.
B. L., and
Bernard,
S., and Pihl,
6. Vladimirsky, 375-377.
A. (1981)
J. (1958)
Phurmacol.
M. A., and Ladigina,
7. Bakker-Woudenberg, Regts, D., and Michel,
Physiol.
19,327-336.
4. Monsigny, M., Roche, A.-C., and Bailly, phys. Res. Commun. 121.579-584. 5. Olsnes,
M. B. (1976)
P. (1984) Ther.
G. A. (1982)
C. R. Acad.
Sci.
Biochem.
Bio-
15.355-381. Biomedicine
36,
I. A. J. M., Lokerse, A. F., Roerdink, F. H., M. F. (1985) J. Infect. Dis. 151,917-924.
ET
AL.
8. Jadin, J. M., Trouet, A., Van Hoof, F., Bioul-Marchand, M., Maldague, P., and Jadin-Nyssens, M. (1977) Ann. Sot. Be&s Med. Trop. 57,525-530. 9. Alving, C. R., and Steek, E. A. (1979) Trends Biochem. Sci. 4, N175-N177. 10. Trouet, A., Pirson, P., Steiger, R., Masquelier, M., Baurain, R., and Gillet, J. (1981) Bull. W.H.O. 59,449-458. 11. Fiume, L., Busi, C., Mattioli, A., Balboni, P. G., and BarbantiBrodano, G. (1981) FEBS Lett. 129,261-264. 12. Taylor, A. E., and Ganger, D. N. (1984) in Handbook of Physiology, Section 2, The Cardiovascular System, Vol. IV: The Microcirculation (Renkin, E. M., and Michel, C. C., Eds.), pp. 467-520, Amer. Physiol. Sot., Washington, DC. 13. Demignot, S., and Domurado, D. (1987) Drug Des. Deb 1,333348. 14. Nakane, P. K., and Watanabe, T. (1984) J. H&o&em. Cytochem. 32,894-898. 15. Muzykantov, V. R., Sakharov, D. V., Sinitsyn, V. V., Domogatsky, S. P., Goncharov, N. V., and Danilov, S. MT (1988) Anal. B&he&
169,383-389. 16. Clark, C. A., Downs, E. C., and Primus, F. J. (1982) J. Histochem. Cytochem. 30,27-34. 17. Weibel, M. K., and Bright, H. J. (1971) J. Biol. Chem. 245,27342744. 18. Lecoq, D. (1972) Influence de la Structure sur le Comportement Cinitique de Systemes Multienzymatiques Siquentiels, p. 55, Thesis (3rd cycle), Rouen. 19. Schlapfer, P., Minot, W., and Racine, P. (1974) Clin. Chim. Acta 57,283-289. 20. Bourdillon, C., Hervagault, C., and Thomas, D. (1985) Biotechnol. Bioeng. 27,1619-1622. 21. Zaitsu, K., and Ohkura, Y. (1980) Anal. Biochem. 109.109-113.