Anal. Chem. 1997, 69, 3124-3127
Thermal Stabilization of Enzymes Immobilized within Carbon Paste Electrodes Joseph Wang,* Jie Liu, and Gemma Cepra
Department of Chemistry and Biochemistry, New Mexico State University, Las Cruces, New Mexico 88003
In this note we report on the remarkable thermal stabilization of enzymes immobilized in carbon paste electrodes. Amperometric biosensors are shown for the first time to withstand a prolonged high-temperature (>50 °C) stress. Nearly full activity of glucose oxidase is retained over periods of up to 4 months of thermal stress at 6080 °C. Dramatic improvements in the thermostability are observed for polyphenol oxidase, lactate oxidase, alcohol oxidase, horseradish peroxidase, and amino acid oxidase. Such resistance to heat-induced denaturation is attributed to the conformational rigidity of these biocatalysts within the highly hydrophobic (mineral oil or silicone grease) pasting liquid. While no chemical stabilizer is needed for attaining such protective action, it appears that low humidity (i.e., low water content) is essential for minimizing the protein mobility. Besides their implications for electrochemical biosensors, such observations should lead to a new generation of thermoresistant enzyme reactors based on nonpolar semisolid supports. Enzyme-based bioassays have proven to be viable tools in many clinical, biotechnological, and environmental analytical problems demanding high selectivity. However, the lack of long-term stability greatly limits the practical utility of enzymatic assays and probes. Even a “robust” enzyme, such as glucose oxidase (GOx) is denatured in solution above 55 °C,1 and its half-life at 60 °C is only 22 min.2 The immobilization of enzymes onto transducer surfaces can lead to changes in their behavior compared to that observed in homogeneous solutions.3,4 In addition to changes in the kinetic parameters of the enzymatic reaction, the immobilization step may offer improved stability (due to restricted movement of the attached biocatalyst). Still, most immobilization schemes do not protect the enzyme from thermal deactivation at temperatures exceeding 50 °C. For example, GOx entrapped in the common polymeric films Nafion or polypyrrole displays very short half-lives of 30 and 100 min, respectively, at 60 °C.2 In this paper, we report for the first time on the dramatic enhancement of the thermal stability of several enzymes upon immobilization in carbon paste matrixes. The mixed enzyme/ carbon paste strategy has received considerable attention in recent years as it results in versatile and renewable reagentless biosensors.5 Several reports have indicated improvements in the stability (1) Cass, C. G.; Davis, G.; Francis, G.; Hill, H. A.; Higgins, I. J.; Plotkin, E.; Scott, L.; Turner, A. P. Anal. Chem. 1984, 56, 667. (2) Fortier, G.; Vaillancourt, M.; Belanger, D. Electroanalysis 1992, 4, 275. (3) Wheetal, H. Anal. Chem. 1974, 46, 602A. (4) Bowers, L. Anal. Chem. 1986, 58, 513A. (5) Gorton, L. Electroanalysis 1995, 7, 23.
3124 Analytical Chemistry, Vol. 69, No. 15, August 1, 1997
of enzymes in the paste microenvironment. For example, Kulys and Hansen6 reported recently on the high stability of GOxcontaining carbon pastes over 3 months at 35 °C. Amine and Kauffmann7 demonstrated the extended room-temperature stability of the “fragile” enzyme glutamate dehydrogenase. Yet, no reports are available on prolonged activity of amperometric biosensors at temperatures in excess of those that normally denature the native protein. Such improvements are illustrated in the following sections for six different enzymes which display unusually extended lifetimes upon stressing the corresponding carbon paste biosensors at elevated temperatures (60-80 °C) for prolonged periods as long as 4 months. The surprising fact that enzymes retain their biocatalytic activity at temperatures higher than 50 °C within such “hostile” paste environment is related to enzymology in nonaqueous media.8,9 Studies in Klibanov’s laboratory confirmed that the dispersion of various enzymes in hydrophobic organic solvents can greatly enhance their thermal stability.8,9 Such behavior has been attributed to the minimization of the enzyme unfolding in hydrophobic environments, as compared to the conformational flexibility common in aqueous media. As will be demonstrated below, hydrophobic carbon paste matrixes impart a similar conformational rigidity and hence a dramatic enhancement of the thermal stability of the resulting biosensors. EXPERIMENTAL SECTION Apparatus. Batch experiments were carried out with the Bioanalytical Systems (BAS, West Lafayette, IN) Model CV-27 voltammetric analyzer in connection with a BAS X-Y-t recorder. The working electrode, reference electrode (Ag/AgCl, Model RE1, BAS), and platinum wire counter electrode joined the 10-mL cell (BAS, Model VC-2) through holes in its Teflon cover. A magnetic stirrer and bar provided the convective transport. Bare platinum and glassy carbon disk electrodes (BAS Inc.) were employed for monitoring the activity of the soluble enzyme. Some experiments were performed in a flow injection system that consisted of a carrier reservoir, a Rainin Model 5041 sample injection valve (20-µL loop), interconnecting Teflon tubing, an Alitea (Sweden) XV pump, and a homemade large-volume walljet amperometric detector. An EG&G PAR Model 264A voltammetric analyzer, coupled to a Houston Omniscribe strip-chart recorder, was used in these flow experiments. A National Appliance oven (Model 420) was used for high-temperature storage of the various enzyme electrodes. (6) (7) (8) (9)
Kulys, J.; Hansen, H. Anal. Chim. Acta 1995, 303, 285. Amine, A.; Kauffmann, J. M. Bioelectrochem. Bioenerg. 1992, 28, 117. Zaks, A.; Klibanov, A. M. Science 1984, 224, 1249. Zaks, A.; Klibanov, A. M. J. Biol. Chem. 1988, 263, 3194.
S0003-2700(97)00230-8 CCC: $14.00
© 1997 American Chemical Society
Reagents. Glucose oxidase (GOx, EC 1.1.3.4, 181.6 units/ mg, from Asperigilus niger), lactate oxidase (LOx, EC 1.1.3.2, 26 units/mg, from Pediococcus species), alcohol oxidase (AOx, EC 1.1.3.13, 26 units/mg, from Hansenula species), tyrosinase (polyphenol oxidase, PPO, EC 1.14.18.1, 3900 units/mg from mushroom), peroxidase (HRP, EC 1.11.1.7, 80 units/mg, from horseradish), and L-amino acid oxidase (AAOx, EC 1.4.3.2, 0.34 units/mg, from Crotalus admanteous) were purchased from Sigma (St. Louis) and used as received. β-D-(+)-glucose, L-(+)-lactic acid, L-phenylanaline, and glutaraldehyde (grade I, 25% aqueous solution) were also obtained from Sigma. Phenol was received from Baker and ethanol from Quantum Chemical Co. (Tuscolosa, IL). A 0.05 M phosphate buffer solution (pH 7.4) served as the supporting electrolyte and carrier solution. Graphite powder (grade 38) was obtained from Fisher. Rhodium-on-carbon (5% rhodium), mineral oil, resorcinol, 1,3phenylenediamine (99+%), hydrogen peroxide, and potassium peroxide were obtained from Aldrich (Milwaukii, WI). Iridiumon-carbon (Vulcan XC-72) was obtained from E-Tek (Natik, MA) and silicone grease from Dow Corning (Midland, MI). Electrode Preparation. Various bioelectrodes, based on different immobilization schemes and enzymes, were employed. Carbon paste enzyme electrodes were prepared by thoroughly mixing the corresponding enzyme with the carbon paste in the desired composition (as described below). A portion of the resulting paste was packed firmly into the electrode cavity (3mm diameter, 1-mm depth) of a Teflon sleeve. The paste surface was smoothed on a weighing paper. A wide range of carbon paste bioelectrodes were prepared as follows: 1. The GOx/Rh carbon paste was prepared by thoroughly hand-mixing 10 mg of GOx with 100 mg of the metalized carbon paste (40% rhodium-on-carbon and 60% mineral oil, w/w). Some experiments employed a silicone grease at 65% (w/w) (instead of the mineral oil binder). 2. The Rh/LOx and Rh/AOX carbon pastes were prepared in a similar fashion, by mixing 2 mg of lactate oxidase or 8.4 mg of alcohol oxidase, respectively, with the 100 mg of the metalized carbon paste (40% rhodium-on-carbon and 60% mineral oil, w/w). 3. The PPO- and HRP-containing carbon pastes were prepared by mixing 2 mg of tyrosinase or 5 mg of peroxidase, respectively, with 100 mg of nonmetalized carbon paste (60% graphite and 40% mineral oil, w/w). 4. The AAOx/Ir carbon paste was prepared by mixing 10 mg of L-amino acid oxidase with 100 mg of the iridium-containing paste (55% iridium-on-carbon and 45% mineral oil, w/w). A platinum disk electrode (1.5-mm diameter, Model 2013, BAS) was used in connection with the polymeric entrapment of GOx. The platinum surface was first coated with a 10 µL of the enzyme solution (prepared by dissolving 1 mg of GOx in 100 µL of phosphate buffer). The enzyme layer was then covered with a 10-µL aliquot of the 2.5% glutaraldehyde solution and was allowed to dry at room temperature. Electropolymerization was performed subsequently in a phosphate buffer solution (pH 6.0) containing 3 mM 1,3-phenylenediamine and 3 mM resorcinol, by holding the potential of +0.65 V for 10 min. Procedure. The thermal stability of the biosensors was evaluated by storing the different bioelectrodes in the oven at different temperatures and measuring the response for the
Figure 1. Thermodeactivation kinetics at 60 °C of soluble GOx (a) and of GOx entrapped in poly(phenylenediamine) (b) and within a carbon paste matrix (c). The activity was assayed at the times indicated using amperometric monitoring of the liberated hydrogen peroxide at potentials of +0.6 (a, b) and -0.05 V (c). A stirred phosphate buffer solution (pH 7.4) and room temperature were used during the amperometric experiment. A platinum disk electrode (BAS) was used in (a).
corresponding substrate at regular intervals. Measurements were performed at room temperature (upon removing the electrode from the oven), by applying the operating potential and allowing the transient background current to decay prior to the addition (or injection) of the substrate solution. A stirring rate of 300 rpm was employed in the batch experiments, while a carrier flow rate of 1.0 mL/min was used in connection with the flow injection experiments. Batch experiments at the bare disk electrode were used in connection with the soluble GOx or PPO (in the heated stirred phosphate buffer solution). Additions of the glucose or phenol substrates were used for obtaining the amperometric response. RESULTS AND DISCUSSION Thermal stability is a measure of the ability of the biosensor to withstand elevations in temperature, frequently in excess of those that normally denature the native enzyme. The dramatic enhancement in the thermal stability of enzyme within the carbon paste environments is illustrated below for several analytically important “robust” and “fragile” enzymes. Figure 1 displays a comparative study of the thermodeactivation at 60 °C of soluble glucose oxidase (GOx, a), with that of the polymer- (b) and carbon paste- (c) entrapped enzymes. Both the free and polymerentrapped GOx rapidly lose their activity under this thermal stress. For example, 40 and 100% of the activity of the soluble enzyme is lost within 2 and 4 h, respectively. As expected,2,10 a slower decay is observed for the polymer-entrapped GOx. Yet, 80 and 97% activity losses are indicated after 1 and 6 days of incubation at 60 °C. In contrast, using the carbon paste confinement, nearly full activity (95%) of the enzyme is retained even after 10 days of stress. A greatly extended period of 4 months was subsequently employed to examine the long-term stability of the enzyme-based carbon paste upon exposure to 60 °C (Figure 2). An excellent (10) Sasso, S.; Pierce, R.; Walla, R.; Yacynych, A. M. Anal. Chem. 1990, 62, 1111.
Analytical Chemistry, Vol. 69, No. 15, August 1, 1997
3125
Figure 2. Variation of the amperometric response to 6 mM glucose at -0.05 V of the GOx-carbon paste electrode with time upon exposure at 60 °C over a 4-month period. Also shown the actual response tracings for 2 mM glucose additions at the first (A) and 70th (B) days of this study. The same paste surface was used throughout. Other conditions, as in Figure 1c.
Figure 3. Thermal inactivation at 80 °C of soluble PPO (A) and carbon paste-entrapped PPO (B). The activity was assayed at times indicated by monitoring the response of the liberated quinone at -0.10 V in the presence of the phenol substrate. A glassy carbon disk electrode (BAS) was used in (A).
thermoresistance is indicated from the retention of 85% of the original activity over this prolonged period. To the best of our knowledge, these data represent the most stable response reported to date for any enzyme-based biosensor under hightemperature (>50 °C) stress. Even sensors based on new thermostable enzymes display a decay of 2%/h at similar temperatures.11 Apparently, the oil binder increases the structural rigidity of the protein in a manner similar to the protective action of hydrophobic organic solvents.8,9 Also shown are the actual response tracings (used for assessing the stability) at the first (A), and 70th (B) days of this prolonged thermal stress. These data indicate that the dynamic properties are not compromised by such lengthy exposure to 60 °C. Note that the same paste surface was used throughout this study. A similar tolerance to heat-induced denaturation was observed over the same period by (11) Vreeke, M.; Yong, K.; Heller, A. Anal. Chem. 1995, 67, 4247.
3126
Analytical Chemistry, Vol. 69, No. 15, August 1, 1997
replacing the mineral oil binder with silicone grease (not shown). Note also that such remarkable thermostability was achieved without any enzyme stabilizer. At higher temperature (80 °C) stress, the GOx-containing carbon paste lost 20 and 35% of its initial response following 10 and 60 days, respectively (not shown). The inferior thermoresistence reported in ref 6 was attributed in part to a significantly higher humidity (75%) and to the presence (leaching) of mediator. The data of Figure 2 were obtained using a mediatorless (metalized) carbon paste transducer and a low humidity (of ∼10%). The reduction in water content has a similar effect upon the thermostability of lipase in organic solvents.8 Apparently, low humidity and water content have a profound effect upon the protein mobility and its thermal stability. Results with other enzyme electrodes suggest that the paste thermostabilization process may be broadly applied. For example, polyphenol oxidase is known for its effective biocatalytic activity in organic media.12 In water at 80 °C, the soluble enzyme loses its activity almost instantaneously (Figure 3A). In contrast, when PPO is incorporated within the mineral oil-based paste, its thermoresistance increased dramatically (Figure 3B); only negligible loss in activity is observed over this 6 h of thermal stress at 80 °C. Actually, at this elevated temperature, the PPOentrapped carbon paste had a half-life of 3 days (Figure 4c). The hydrophobic paste environment minimizes not only the thermal denaturation of PPO but also its deactivation by reaction products generated in aqueous environments.13 Similarly, carbon pastecontaining horseradish peroxidase and amino acid oxidase retained 50% of their activity after 6-day and 12-h stress, respectively, at 80 °C (Figure 4b and d). Improved thermoresistance was observed also for the immobilization of fragile enzymes such as alcohol oxidase or lactate oxidase. For example, 1 and 5 days of 60 °C stress of the alcohol oxidase-containing carbon paste (12) Wang, J.; Lin, Y.; Chen, Q. Electroanalysis 1993, 5, 23. (13) Hall, G. F.; Best, D.; Turner, A. P. Anal. Chim. Acta 1988, 213, 113.
Figure 4. Stability of GOx- (a), HRP- (b), PPO- (c), and AAO- (d) containing carbon paste electrodes at 80 °C. The stability was evaluated from the response to 6 mM glucose (a), 6 mM hydrogen peroxide (b), 6 mM phenol (c), and 5 mM L-phenylalanine (d). Operating potential, -0.05 (a), 0.00 (b,d) and -0.10 (c) V. Other conditions are as in Figure 1c, except that a flow injection detection was employed in (d); flow rate, 1.0 mL/min.
electrode resulted in retention of 70 and 30%, respectively, of the initial response. Similarly, the lactate oxidase-modified electrode retained 40 and 15% of its sensitivity over the same periods (not shown). In conclusion, we have demonstrated that confinement of various enzymes within hydrophobic carbon paste matrixes can (14) Wang, J.; Chen, Q. Anal. Lett. 1995, 28, 1131. (15) Moussy, F.; Jakeway, S.; Harrison, D. J.; Rajotte, R. Anal. Chem. 1994, 66, 3882.
be used to dramatically enhance the thermal stability and storage life of the resulting biosensors. The extreme thermostability in the organic pasting liquids is attributed to the unusual conformational rigidity in this nonpolar binding environment. Other carbon composite bioelectrodes, including graphite epoxy, carbon wax, or screen-printed ones, may similarly benefit from the inherent thermal stability imparted by their hydrophobic organic binder. Indeed, short-term (1-2 h) heat curing of such electrodes at temperature exceeding 100 °C has been reported.14,15 The finding reported above should have broader implications, beyond the design of thermostable biosensors. It should lead to heat-resistant immobilized enzyme systems (desired in numerous industrial and biotechnological applications involving elevated temperatures). Self-supported semisolid hydrophobic matrixes, e.g., wax or grease (without a conducting carbon) may thus be used for fabricating various enzyme-based bioreactors. A deeper understanding of the protective action of such nonpolar enzyme supports and of the role of low humidity would be required to fully realize these opportunities. ACKNOWLEDGMENT We thank the U.S. EPA and IL Inc. for the generous financial support. G.C. acknowledges a fellowship from Diputacion General de Aragon (Spain). Received for review February 28, 1997. Accepted May 20, 1997.X AC9702305 X
Abstract published in Advance ACS Abstracts, July 1, 1997.
Analytical Chemistry, Vol. 69, No. 15, August 1, 1997
3127
