Anal. Chem. 1998, 70, 3685-3694
An Enzyme Switch Employing Direct Electrochemical Communication between Horseradish Peroxidase and a Poly(aniline) Film Philip N. Bartlett,* Peter R. Birkin, and Jin Hai Wang
Department of Chemistry, University of Southampton, Southampton, SO17 1BJ, U.K. Francesco Palmisano
Department of Chemistry, University of Bari, 70126 Bari, Italy Giuseppe De Benedetto
Department of Chemistry, University of Basilicata, 85100 Potenza, Italy
An enzyme switch, or microelectrochemical enzyme transistor, responsive to hydrogen peroxide was made by connecting two carbon band electrodes (∼10 µm wide, 4.5 mm long separated by a 20-µm gap) with an anodically grown film of poly(aniline). Horseradish peroxidase (EC 1.11.1.7) was either adsorbed onto the poly(aniline) film or immobilized in an insulating poly(1,2-diaminobenzene) polymer grown electrochemically on top of the poly(aniline) film to complete the device. In the completed device, the conductivity of the poly(aniline) film changes from conducting (between - 0.05 and + 0.3 V vs SCE at pH 5) to insulating (>+0.3 V vs SCE at pH 5) on addition of hydrogen peroxide. The change in conductivity is brought about by oxidation of the poly(aniline) film by direct electrochemical communication between the enzyme and the conducting polymer. This was confirmed by measuring the potential of the poly(aniline) film during switching of the conductivity in the presence of hydrogen peroxide. The devices can be reused by rereducing the poly(aniline) electrochemically to a potential below +0.3 V vs SCE. A blind test showed that the device can be used to determine unknown concentrations of H2O2 in solution and that, when used with hydrogen peroxide concentrations below 0.5 mmol dm-3, the same device maybe reused several times. The possible development of devices of this type for use in applications requiring the measurement of low levels of hydrogen peroxide or horseradish peroxidase is discussed. Hydrogen peroxide is produced by many enzyme-catalyzed redox reactions. For example, the enzymes glucose oxidase (EC 1.1.3.4), choline oxidase (EC 1.1.3.17), L-amino acid oxidase (EC 1.4.3.2), D-amino acid oxidase (EC 1.4.3.3), and galactose oxidase (EC 1.1.3.9) all catalyze reactions between their respective substrates and oxygen which produce hydrogen peroxide as one S0003-2700(97)01088-3 CCC: $15.00 Published on Web 08/05/1998
© 1998 American Chemical Society
of the products.1 As a consequence, the electrochemical detection of hydrogen peroxide has long been utilized in amperometric biosensors for a range of substrates, and in particular glucose.2 Nevertheless, hydrogen peroxide is not the ideal species to detect in an electrochemical biosensor because the electrode kinetics for the oxidation or reduction of hydrogen peroxide are slow on many electrode materials. This is because the use of high overpotentials to drive the reaction of hydrogen peroxide at the electrode often leads to problems of interference from reactions of other molecules in the sample and because the reaction of hydrogen peroxide is often easily poisoned by organic compounds present in the sample. These problems have led to a variety of efforts to develop electrodes for hydrogen peroxide detection including the use of the enzyme horseradish peroxidase (HRP; EC 1.11.1.7) to catalyze the electrochemical reduction of hydrogen peroxide either coupled with the use of a redox mediator3-5 or using the direct electrochemistry of the enzyme.6-8 At the same time, horseradish peroxidase has been widely used as an enzyme label in immunoassays9 and to label DNA probes10 because it is reasonably stable, it is readily available commercially, and it shows rapid kinetics for the reduction of hydrogen peroxide with a range of mediators. In these applications, the analyte, either the antigen or target oligonucleotide sequence, is present in low concentration (typically below 10-6 mol dm-3) so that direct detection is difficult or impossible and the enzyme label is used to amplify the signal (1) Dixon, M.; Webb, E. C. Enzymes, 3rd ed.; Longman: London, 1979. (2) Hall, E. A. H. Biosensors; Open University Press: Milton Keynes, 1990, 216267. (3) Tatsuma, T.; Okawa, Y.; Watanabe, T. Anal. Chem. 1989, 61, 2352-2355. (4) Liu, H. Y.; Ying, T. L.; Sun, K.; Li, H. H.; Qi, D. Y. Anal. Chem. 1997, 344, 187-199. (5) Vreeke, M.; Maidan, R.; Heller, A. Anal. Chem. 1992, 64, 3084-3090. (6) Jonsson, G.; Gorton, L. Electroanalysis 1989, 1, 465. (7) Wollenberger, U.; Wang, J.; Ozsos, M.; Gonzalez-Romero, E.; Scheller, F. Bioelectrochemistry Bioenerg. 1991, 26, 287-296. (8) Gorton, L.; Jonsson-Peterson, G.; Csoregi, E.; Johansson, K.; Dominguez, E.; Marko-Varga, G. Analyst 1992, 117, 1235-1241. (9) Vreeke, M.; Rocca, P.; Heller, A. Anal. Chem. 1995, 67, 303-306. (10) Park, J. S.; Kurman, R. J.; Kessis, T. D.; Shah, K. V. Mod. Pathol. 1991, 4, 81-85.
Analytical Chemistry, Vol. 70, No. 17, September 1, 1998 3685
by catalyzing the generation of a larger number of molecules for detection from each labeled molecule.11 This type of approach is attractive for electrochemical immunoassays and DNA probes because the concentrations of analyte are too low for simpler direct amperometric detection, and yet, one would like to be able to exploit the advantages of electrochemical methods in terms of cost and ease of measurement in developing inexpensive disposable medical diagnostic devices based on the assay of these analytes. One way to do this is to develop approaches to the lowlevel electrochemical detection of hydrogen peroxide and/or horseradish peroxidase since this would be central to the operation of any functioning device. Amperometric detection depends on the direct measurement of the current generated by the reaction of the detected species at the electrode. In general, this approach is restricted to concentrations greater than ∼10-6 mol dm-3 because of limitations of the signal-to-noise ratio caused by the background current due to interferences and nonfaradaic processes. For potentiometric devices, the detection limits are similar. To extend electrochemical detection below this limit, the most successful approaches have been based on the use of some form of stripping voltammetry. The essential feature of these techniques is that the analyte is allowed to accumulate over a known period of time at the electrode surface and is then subsequently detected. This accumulation period achieves an integration of the signal and hence a significant improvement in the detection limit.12,13 Recently we14,15 and others16,17 described enzyme switches (also referred to microelectrochemical enzyme transistors) which achieve a similar integration of the analyte signal and hence could be an attractive approach toward the detection of low analyte concentrations using simple, disposable structures coupled with simple detection electronics. In the enzyme switch, an enzyme-catalyzed reaction is used to change the oxidation state of a conducting polymer film deposited across the gap between two electrodes. This oxidation or reduction of the conducting polymer causes a change in the conductivity of the device which is sensed by measuring the current flowing between the two electrodes through the polymer film. Similar structures, but without the use of the enzyme to provide selectivity in the device response, were first described by Wrighton and colleagues.18-20 Devices of this type are distinct and different from the more common amperometric or potentiometric approaches to electrochemical detection. Thus, no reference electrode is required to make the measurement, the measurement circuitry is very simple, although low concentrations of analytes are detected the currents to be measured are not necessarily small, and the devices can be readily miniaturized. To date, the most fully studied system has been an enzyme switch (11) Ghindilis, A. L.; Makower, A.; Bauer, C. G.; Bier, E. F.; Scheller, F. W. Anal. Chim. Acta 1995, 304, 25-31. (12) Martin, C. R.; Dollard, K. A. J. Electroanal. Chem. 1983, 159, 127-135. (13) Bard, A. J.; Faulkner, L. R. Electrochemical Methods; John Wiley & Sons: New York, 1980. (14) Bartlett, P. N.; Birkin, P. R. Anal. Chem. 1993, 65, 1118-1119. (15) Bartlett, P. N.; Birkin, P. R. Anal. Chem. 1994, 66, 1552-1559. (16) Matsue, T.; Nishizawa, M.; Sawaguchi, T.; Uchida, I. J. Chem. Soc., Chem. Commun. 1991, 1029. (17) Matsue, T.; Nishizawa, M.; Uchida, I. Anal. Chem. 1992, 64, 1029. (18) Paul, E. W.; Ricco, A. J.; Wrighton, M. S. J. Phys. Chem. 1985, 89, 14411447. (19) Thackeray, J. W.; Wrighton, M. S. J. Phys. Chem. 1985, 89, 5133-5140. (20) Thackeray, J. W.; Wrighton, M. S. J. Phys. Chem. 1986, 90, 6674-6679.
3686 Analytical Chemistry, Vol. 70, No. 17, September 1, 1998
responsive to glucose and based on glucose oxidase immobilized on a poly(aniline) film. In this device, we have shown that the oxidation of glucose by the immobilized glucose oxidase leads, through the use of a redox mediator, to the reduction of the poly(aniline) film from the insulating, pernigraniline form to the conducting, emeraldine form. This change in conductivity of the poly(aniline) causes a significant change in resistance of the enzyme switch (typically from >1 MΩ to ∼2 kΩ), and this large change in resistance is readily detected as a significant change in the current flowing through the polymer film. We were also able to demonstrate that these devices could be electrochemically reoxidized to the pernigraniline state and reused to assay glucose in solution both at pH 5 and, by changing from poly(aniline) to a composite of poly(aniline) and poly(styrenesulfonate), at pH 7.21 Since the resistance of the enzyme switch depends on the total number of analyte molecules that have reacted with the immobilized enzyme, the device is able to achieve accumulation or integration of the signal in a manner that is in some ways analogous to that occuring in stripping voltammetry. Therefore, this type of device should be suitable for detection of low levels of the analyte. We have explored this possibility for the detection of low concentrations of glucose using the glucose oxidase-based enzyme switch.22 In that work, we demonstrated that we could reproducibly fabricate enzyme switches responsive to glucose and that using these devices we were able to detect glucose down to 2 × 10-6 mol dm-3 in 1 cm3 of air-saturated buffer solution. This detection limit was shown to be ∼40 times lower that the corresponding detection limit for the same device operated as an amperometric sensor under the same conditions. It is therefore of interest to see whether similar devices might be constructed to detect either hydrogen peroxide or the enzyme horseradish peroxidase since these then might be developed to form the basis of simple, inexpensive disposable systems for electrochemical immunoassay or DNA probes. In principle, it should be possible to extend our previous work on enzyme switches to use horseradish peroxidase in place of glucose oxidase because, as we have recently demonstrated, poly(aniline) is a very efficient electrode for the reduction of horseradish peroxidase and the enzyme is readily adsorbed onto the poly(aniline). Thus, at pH 5 horseradish peroxidase efficiently catalyses the reduction of hydrogen peroxide at a poly(aniline)coated electrode at +0.05 V vs SCE without the need for any added mediator.23 Further work has shown that this same system can be used for the quantitative reduction of hydrogen peroxide at a reticulated vitreous carbon electrode for extend periods with high current densities.24 This works both with adsorbed enzyme and with immobilized enzyme. Horseradish peroxidase reduces hydrogen peroxide through the following set of reactions8 (21) Bartlett, P. N.; Wang, J.-H. J. Chem. Soc., Faraday Trans. 1996, 92, 41374143. (22) Bartlett, P. N.; James, W.; Wang, J. H. Analyst 1998, 123, 387-392. (23) Bartlett, P. N.; Birkin, P. R.; Palmisano, F.; De Benedetto, G. J. Faraday Trans. 1996, 92, 3123-3130. (24) Bartlett, P. N.; Pletcher, D.; Zeng, J. J. Electrochem. Soc. 1997, 144, 37053710.
H2O2 + HRP(red) f HRP (compound I) + 2H2O HRP (compound I) + M + H+ f HRP (compound II) + M+ HRP (compound II) + M + H+ f HRP(red) + M+
where HRP(red) is the ferric form of the enzyme, HRP (compound I) is the oxidized form of the enzyme, HRP (compound II) is the intermediate form of the enzyme, and M and M+ are the reduced and oxidized forms of the electron donor, respectively. HRP is unspecific in its choice of electron donor, M in the scheme above, and in the present work the role of the electron donor is taken by the poly(aniline) film. Although we can still use poly(aniline) as the conducting polymer in the enzyme switch, there will be some significant differences in the functioning of enzyme switches responsive to glucose and to hydrogen peroxide because they will function in opposite ways. Thus, whereas glucose leads to the reduction of poly(aniline) from the fully oxidized, insulating state to the partially oxidized, conducting state leading to an “off” to “on” switching of the device as the polymer becomes conducting, the hydrogen peroxide switch will operate in the opposite sense as the partially oxidized poly(aniline) is converted to the fully oxidized, insulating state by the hydrogen peroxide leading to a switching of the device from “on” to “off”. This has some important consequences for the operation of the device as we show below. In this paper, we present the first description of an enzyme switch based on horseradish peroxidase and responsive to hydrogen peroxide. We demonstrate that the device operates by the chemical oxidation of the conducting polymer by the enzymecatalyzed reaction. We investigate the consequences of the switch operating in the “on” to “off” direction and the stability and reuse of the devices by resetting them to the “on” state by electrochemical reduction. Finally, we discuss the possible development of this type of device for the low-level detection of hydrogen peroxide or horseradish peroxidase. EXPERIMENTAL SECTION Chemicals. Aniline (Aldrich 99.7%) was distilled once before use and stored in the dark at 0-4 °C; 1,2-diaminobenzene (Aldrich 99%) and 1,4-diaminobenzene (Aldrich 99%) were purified by sublimation before use and stored in the dark. NaHSO4 (Aldrich 99%), Na2SO4 (Aldrich 99%), H2SO4 (Aldrich 99%), citric acid (Aldrich 99.5%), disodium hydrogen phosphate (Fisons A. R. 99.5%), and hydrogen peroxide (Sigma, 30 vol %) were used without further purification. Stock solutions of hydrogen peroxide were freshly prepared each day. Hydrogen peroxide concentrations were standardized by titration with potassium permanganate.25 Horseradish peroxidase (EC 1.11.1.7) was obtained from two sources, the first (Sigma) had an assayed26 activity of 275 Sigma units mg-1 at 26 °C, the second had an assayed activity of 317 Sigma units mg-1 (at 20 °C) and was a gift from MediSense UK. Both samples were used without further purification. All aqueous solutions were prepared with deionized water purified (25) Vogel, A. L. A Text of Quantitative Inorganic Analysis, 3rd ed.; Longmans: Harlow, Essex, U.K., 1962. (26) Sigma Data Sheet supplied with product.
by passing through a Whatman RO 50 and a Whatman Stillplus water purification systems. All glassware was soaked overnight in a 3% Decon 90 solution and washed thoroughly with deionized water. Equipment. The electrochemical work station consisted of an Oxford Electrodes bipotentiostat, a Gould series 60000 X-Y-t chart recorder and/or a Pharmacia REC 102 Y-Y′-t chart recorder, and a Keithley 175A digital multimeter. Solutions were pumped through the flow injection cell by a Pharmacia P-3 peristaltic pump. A standard three-electrode system was employed consisting of a working electrode, platinum gauze or stainless steel counter electrode (FIA experiments), and reference electrode. All potential measurements were made with reference to a saturated calomel or silver/silver chloride reference electrode. The exact type of reference electrode is stated in the appropriate figure legend. In experiments where polymer potentials and conductivity were measured simultaneously, data were recorded through an ADC card to a PC. Electrode Construction. A dual carbon microband electrode was employed as the working electrode. The construction of these electrodes has been reported elsewhere.14,15 The electrodes are 10 µm wide, 4.5 mm long and separated by a 20-µm gap. Electrical connections were made to the electrodes by attaching two wires to the contact pads with silver paint (RS) and securing the joint with fast-setting epoxy (RS). The bonded electrode was then sealed in a screw-threaded Teflon mount with casting epoxy (Ciba Geigy, Araldite). To avoid the formation of bubbles in the cast, the unset cast was placed in an operating ultrasonic bath for ∼5 min at 50 °C. The epoxy was then allowed to cure fully overnight at room temperature before polishing the electrode to expose the pair of carbon microband electrodes. Polishing was performed on “wet and dry” paper followed by 25-, 5-, 1-, and 0.3-µm alumina (Buehler) powder/water slurry on velvet polishing pads (Buehler, Microcloth). Once the electrode had been prepared for the first time, subsequent polishing was performed with 1- and 0.3-µm alumina. Electrode Modification. The poly(aniline) films were grown in two ways. In the first procedure, poly(aniline) was deposited onto the pair of microband electrodes by stepping the potential of the electrode pair from 0 to +0.9 V vs SCE (or +0.94 V vs Ag/AgCl) in a solution containing 0.2 cm3 of aniline in 5 cm3 of 0.5 mol dm-3 NaHSO4 acidified to pH ∼0 with 0.5 cm3 of concentrated H2SO4. In the second procedure, poly(aniline) films were grown by cycling the electrode pair in the same growth solution from -0.2 to +0.9 vs SCE or from -0.2 to +0.9 vs SCE for the first cycle and from -0.2 to +0.78 V vs SCE for each subsequent cycle. In some of the poly(aniline) growth solutions, 1,4-diaminobenzene was added. The exact conditions of deposition are shown in the appropriate figure legends. After deposition, the poly(aniline)-modified electrodes were washed and stored in a pH 5 citrate/phosphate buffer solution (McIlvane)27 containing 0.5 mol dm-3 Na2SO4. This same buffer solution was used for all peroxide measurements. Two methods were used to immobilize HRP onto the poly(aniline)-coated dual carbon microband electrodes. In the first, (27) Dawson, R. M. C., Elliott, D. C., Elliott, W. H., Jones, K. M., Eds. Data for Biochemical Research, 3rd ed.; Oxford Science Publications: Oxford, U.K., 1986; p 427.
Analytical Chemistry, Vol. 70, No. 17, September 1, 1998
3687
Figure 1. Cyclic voltammogram of a poly(aniline)-modified glassy carbon electrode (area 0.333 cm2) after full reduction of the film (held at -150 mV vs SCE) to the leucoemeraldine state and then reoxidation in 2 mol dm-3 H2SO4. The experiment was performed at room temperature in a pH 5 citrate/phosphate buffer solution containing 0.5 mol dm-3 Na2SO4. The cyclic voltammogram was recorded at 50 mV s-1. The film was grown under potentiostatic control at +900 mV vs SCE for 10 s (growth charge 54.32 mC). The aqueous growth solution was composed of 5 cm3 of 0.5 mol dm-3 NaHSO4, 200 µL of aniline, and 0.5 cm3 of concentrated H2SO4.
poly(1,2-diaminobenzene)28 films containing HRP were deposited electrochemically on top of the poly(aniline) from a solution containing 25 mmol dm-3 1,2-diaminobenzene and ∼120 units cm-3 HRP. The poly(aniline)-coated electrodes were held at open circuit for 10 min in the 1,2-diaminobenzene/HRP growth solution to allow adsorption of the enzyme onto the poly(aniline) surface before deposition of the 1,2-diaminobenzene film under potentiostatic control at +0.4 V vs SCE for a period of 4 min or until 1 mC of charge had been passed. In the second, HRP was absorbed onto the surface from a buffer solution containing ∼120 units cm-3 of the enzyme. In both cases, the electrode was then removed from the solution and placed in a stirred buffer solution for several minutes to wash the device. The blind calibration experiments where performed by using a single poly(aniline)/poly(1,2-diaminobenzene)/HRP device which was calibrated with a 0.47 mmol dm-3 H2O2 solution. This singlepoint calibration was used to calculate the concentrations of the unknown solutions by assuming a linear relationship between switching rate and peroxide concentration. Between measurements, the device was reset electrochemically. RESULTS AND DISCUSSION Switching Function. The electrochemistry of poly(aniline) is known to be pH sensitive with conflicting literature reports on the possible pH range within which it may remain conductive.29-34 (28) Centonze, D.; Guerroeri, A.; Malitesta, C.; Palmisano, F.; Zambonin, P. G. Ann. Chim. 1992, 82, 219-234. (29) Vuki, M.; Kalaji, M.; Nyholm, L.; Peter, L. M. Synth. Met. 1993, 55-57, 1515-1520. (30) Nyholm, L.; Peter, L. M. Synth. Met. 1993, 55-57, 1509-1514. (31) McManus, P. M.; Cushman, R. J.; Yang, S. C. J. Phys Chem. 1987, 91, 744747.
3688 Analytical Chemistry, Vol. 70, No. 17, September 1, 1998
Figure 2. Resistances of two typical devices as a function of their redox potential. The two films were grown potentiostatically from a growth solution containing 200 µL of aniline, 0.5 cm3 of concentrated H2SO4, and 5 cm3 of 0.5 mol dm-3 NaHSO4 at +900 mV vs SCE for ∼20 s. The charges passed during the growth of these poly(aniline) films were 2.57 (O) and 2.30 mC (b). The resistance was measured with a bias of 50 mV applied across the two carbon electrodes as they were cycled between 0 and 0.5 V vs SCE at a sweep rate of 50 mV s-1. The vertical line indicates the transition potential between the insulating, pernigraniline, state and the conducting, emeraldine, state. Arrow A represents the transition required to produce an “off” to “on” transient. Arrow B represents the transition required to produce a “on” to “off” transient.
Figure 1 shows a cyclic voltammogram of a poly(aniline)-modified glassy carbon electrode cycled in a pH 5 buffer solution. It can be seen from the figure that the film, although electrochemically active, shows single broad redox processes in contrast to the two well-separated redox waves characteristic of the electrochemistry of the polymer at low pH.29-34 Figure 2 shows the resistance of two poly(aniline)-coated dual carbon microband electrodes as a function of the potential of the polymer in a pH 5 solution. It can be seen that there is a transition in the conductivity of the polymer which occurs around +280 mV vs SCE, the polymer being conducting below this potential. This corresponds to a transition between the conducting, emeraldine, and insulating, pernigraniline, states of the polymer. Under these conditions, the transition between the insulating and conducting forms of the polymer has been shown14,15 to be fast and reversible so that a poly(aniline) film at pH 5, under the correct conditions, exhibits suitable characteristics for employment in an enzyme switch. Many reports can be found in the literature that discuss the interaction of HRP with modified electrodes, and it has been demonstrated that direct electrochemical communication between the redox center of this enzyme and an electrode surface is possible.6,7 In the reaction with hydrogen peroxide, the heme active site of the enzyme is oxidized so that it must then be reduced by electrons transferred from the electrode to complete the redox cycle. When HRP is used with poly(aniline) in an (32) Asturias, G. E.; Jang, G. W.; MacDiarmid, A. G.; Doblhofer, K.; Zhong, C. Bunsen-Ges. Phys. Chem. 1991, 95, 1381-1384. (33) Chartier, P.; Mattes, B.; Reiss, H. J. Phys. Chem. 1992, 96, 3556-3560. (34) Mafe, S.; Manzanares, J. A.; Reiss, H. J. Phys. Chem. 1993, 98, 2408-2410.
Figure 3. Schematic showing the cycle of events performed during the microelectrochemical enzyme transistor experiments. The conducting state is represented as black, the insulating state as white.
enzyme switch, reaction of the enzyme with hydrogen peroxide is followed by oxidation of the poly(aniline) film by the oxidized form of the enzyme so that the polymer is switched from its conducting, emeraldine form to its insulating, pernigraniline form: this corresponds to an “on” to “off” transition as shown by the arrow labeled B in Figure 2. This is in contrast to the operation of the glucose oxidase-based enzyme switch14 in which the poly(aniline) film is reduced enzymatically from the insulating, pernigraniline form to the conducting, emeraldine state producing an “off” to “on” transition. Figure 3 shows a schematic representation of the repetitive operation for the HRP-based enzyme switch operating from “on” to “off”. In this system, the device is electrochemically reset by potentiostatic reduction of the polymer from the pernigraniline state (>+280 mV vs SCE, pH 5) to the emeraldine state of the polymer (
