Anal. Chem. 1998, 70, 4170-4176
Modulation of Firefly Luciferase Bioluminescence at Bioelectrochemical Interfaces Roger S. Chittock,*,† Andrew Glidle,‡ Christopher W. Wharton,† Nikolas Berovic,§ T. Derek Beynon,§ and Jonathan M. Cooper‡
The School of Biochemistry and School of Physics and Space Research, The University of Birmingham, Edgbaston, Birmingham, B15 2TT, U.K. and Bioelectronics Research Centre, Department of Electronics and Electrical Engineering, The University of Glasgow, Glasgow, G12 8QQ, U.K.
This paper describes a method by which the activity of an immobilized enzyme can be modulated electrochemically at an electrode. The particular example studied, involving the enzyme firefly luciferase being immobilized in a gelatin film of thickness 1 s to reduction of protons at the electrochemical interface, resulting in a local pH increase within the bioactive membrane. These changes in concentrations of species within the membrane have consequent implications for variations in enzyme activity, as indicated by eqs 1 and 2. Previously, in complementary studies, Karatani et al.22 used a different strategy to modulate the bacterial luciferase bioluminescent reaction in solution, by electrogenerating an appropriate form of the enzyme’s substrate at an electrode surface. Thus FADH2, the cosubstrate for the luciferase from the bacterium Vibrio harveyi, was generated from FAD at an electrode surface in order to initiate bioluminescence. In contrast to the experiments described here, the bacterial luciferase was in the solution phase, and so the experiments were less relevant to the direct spatial control (in the x-y plane) of enzyme activity which may be envisaged as a requirement in certain bioanalytical applications. EXPERIMENTAL SECTION Luciferase and luciferin were obtained in a purified form from Sigma. Luciferase stock solutions were prepared in phosphatebuffered saline (PBS) to give a final concentration of 2 mg mL-1. Multilayer gold-coated glass surfaces consisting of 10, 10, and 100 nm of Ti, Pd, and Au, respectively, were prepared by electron beam evaporation. Gold-on-glass surfaces of approximate dimensions 5 mm × 15 mm were then coated with a layer of 2% (w/v) gelatin + 0.025% (v/v) Triton X-100 in aqueous solution and allowed to dry, creating a robust protein layer 95% of its original value, although at a slower rate than the “switch off” process. During these changes in bioluminescence, the current passing through the cell, shown by the dotted line in Figure 2a, decayed rapidly to a steady state after an initial, instantaneous peak. On stepping the voltage back to 0 V, a fast reverse transient current was observed due to the capacitive effect of the gold electrode and dispersal of the ionic concentration gradients formed in solution as a result of the applied electric field. When a suitably oxidizing potential was applied (e.g., +1.5 V vs Ag|AgCl, Figure 2b), a similar reversible decrease in biolumi-
nescence was observed although the kinetics differed significantly from that shown in Figure 2a. For example, an unrecoverable loss of enzyme activity was observed after each on/off cycle. Consequently, since our investigations are focused on the longterm viability of these luciferase-based electrobioluminescent devices, the experiments discussed below are designed to elucidate the mechanism by which the level of bioluminescence changes during cathodic (reducing) potential steps. Indeed, the cathodic modulation of the immobilized luciferase was repeatable over extended periods of time, although the oxidative charge passed during the step from -1.2 to 0 V was a small fraction of that passed during the reverse step from -1.2 to 0 V, reflecting the longer period of time required for the bioluminescence to return to its maximum value. Thus, the reduced species formed during the step from 0 to -1.2 V (which is postulated as the cause of bioluminescence decrease) is not quantitatively reoxidized at 0 V. Effect of Reducing Potential on Modulation of Bioluminescence. To further investigate the mechanism of the electropotential-induced variation in bioluminescence, a series of potential step experiments were performed in which bioluminescence and electrochemical responses were collected simultaneously (Figure 3). Monitoring the change in photon flux as it approached a
4172 Analytical Chemistry, Vol. 70, No. 19, October 1, 1998
Figure 3. Effect of electrochemical potential on the modulation of bioluminescence of immobilized firefly luciferase: (a) time courses of modulation at a working electrode potential of -0.8 to -1.4 V relative to reference electrode. No modulation was observed at -0.5 V (data not shown). The potential was applied at t ) 0. In each case, bioluminescence returned to >95% of its prepotential value when the applied potential was returned to 0 V. Reaction conditions: 200 HEPES, 5 mM MgCl2, 10 mM DTT, pH 7.8. The steady-state bioluminescence was determined by extrapolation of a doubleexponential function to each time course. (b) Steady-state bioluminescence (filled circles) against working electrode potential measured relative to the reference electrode. Nonlinear fitting of eq 6 to this data set (continuous line) gave values of A ) -11.5 ( 0.5, B ) 7.59 × 10-6 ( 2.05 × 10-5, Imax ) (1.43 ( 0.11) × 105.
steady-state value at a given potential step indicated that the species causing the reduction in bioluminescence was either directly or indirectly reducible at potentials below ∼-0.8 V (Figure 3a). An indication that a redox reaction of a film-bound species (e.g., the enzyme) was not responsible for the bioluminescence decrease comes from Figure 3b, which shows that the decrease in potential required to reduce the steady state bioluminescence from 10 to 90% of its original value was ∼0.4 V. This is considerably greater than the ∼113 mV step generally required to change the fraction of redox species in the either the oxidized or reduced state from 10 to 90% of the total number of electroactive sites within a simple redox-active film (this estimation is based on the assumption that the redox species in the film are in thermodynamic equilibrium permitting the application of the Nernst equation). The above experiments indicate that a solution species is involved in the electromodulation of bioluminescence. Of the species present in solution phase that influence eqs 1 and 2, both protons and oxygen are reducible below -0.5 V and the experiments detailed below were performed in order to elucidate the influence, if any, of these electrochemical reductions on the modulation in bioluminescence. Effect of pH on Modulation of Bioluminescence. Evidence that the mechanism of bioluminescence modulation involves a pH gradient induced by current flow was provided by variation of the pH of the solution phase (Figure 4). While at pH 7.8, the optimum pH for bioluminescence, the application of a reducing potential of -1.2 V gave a decrease in bioluminescence (Figure 2); at pH 6.0 the same potential produced an increase in bioluminescence, reaching a maximum after ∼280 s (Figure 4). Similar results were obtained for pH values below the biological optimum (pH 6.5 and pH 7.0), while at pH 9.0, a reducing potential brought about a decrease in bioluminescence. Interestingly, at pH 7.7 (a value
Figure 4. Effect of pH on the modulation of immobilized luciferase bioluminescence. Reaction conditions were 200 µM ATP, 17 µM luciferin, 12.5 mM HEPES, 12.5 mM MES, 5 mM MgCl2, 10 mM DTT, with pH as specified for individual time courses. Modulation was initiated at t ) 0 by a -1.2 V potential relative to a Ag|AgCl electrode. (a) Bioluminescence time courses normalized to prepotential bioluminescence to show direction of bioluminescence change. (b) Unnormalized data to show absolute bioluminescence values. Note variations in prepotential bioluminescence due to solution pH.
just below the enzyme’s optimum pH), a small increase in bioluminescence was subsequently followed by a decrease in bioluminescence, Figure 4a. In all cases, bioluminescence returned to >95% of its original value when the applied potential was removed. The pH-dependent effects described can be explained in terms of an decrease in the [H+] in the gelatin layer, brought about by the electrochemical process occurring at the cathode surface and giving rise to local changes in the pH. In all cases, if the pH of the bulk medium was less than the optimum (pH 7.8), then the electrochemical reaction that occurred at the electrode shifted the local pH toward the enzyme’s optimum and hence the bioluminescence increased. In general, the more acidic the bulk medium, the longer the time taken for the interfacial pH to reach a maximum bioluminescence value. For an initial solution pH of >7.8, the electrochemical reaction moved the local pH experienced by luciferase away from its optimum and the bioluminescence decreased. However, both H+ and O2 electroreduction would act to increase the pH at the electrode surface and so a consideration of the kinetics of the bioluminescence changes is required to clearly distinguish which effect was responsible for bioluminescence electromodulation. Effect of Buffer Strength on Modulation of Bioluminescence. The dynamic nature of the decrease in bioluminescence on electroreduction noted in Figure 2a is illustrated more clearly Analytical Chemistry, Vol. 70, No. 19, October 1, 1998
4173
Figure 5. Effect of HEPES concentration on the modulation of immobilized luciferase bioluminescence. Reaction conditions were 200 µM ATP, 17 µM luciferin, 5 mM MgCl2, 10 mM DTT, pH 7.8 with a HEPES concentration of 150 (solid line), 125 (dotted line), 100 (dashed line), 75 (dash-dot), 50 (long dash-short dash), and 25 mM (dash-dot-dot-dot). The ionic strength between experiments was kept constant by the addition of an appropriate amount of sodium chloride. In each case, modulation was initiated at t ) 0 by a -1.2 V potential relative to a Ag|AgCl electrode. (a) Time course of modulation with increasing buffer concentration. Bioluminescence is normalized to the prepotential value. (b) Time taken for bioluminescence to reach 20% of its original value as a function of buffer concentration. The decrease in the rate of the slow kinetic phase of electromodulation as the buffer concentration increased is strong evidence that the phenomenon arises from a net decrease in [H+] at the cathode surface.
as the buffer capacity of the supporting electrolyte is substantially increased (Figure 5a). In these experiments, HEPES concentration was varied from 25 to 150 mM (at constant ionic strength), and the luciferase-gelatin film thickness was decreased (compared to the film illustrated in Figure 2) to reduce the time for the luminescence to switch off and so emphasize the bioluminescence electromodulation variations with HEPES concentration. On increasing the buffer strength, while the initial rate of bioluminescence decrease remained constant on the time scale used, the length of time taken for bioluminescence to reach 20% of its initial value substantially increased (Figure 5b). The increased time taken at higher buffer concentrations for the second phase of the bioluminescence response (after 1 s) can be accounted for by the ready supply of protons from the buffer species, counteracting the loss of protons due to electroreduction. Thus, the pH changes in the local film are considerably slowed at high buffer concentrations. The marked slowness of the change in bioluminescence during the second kinetic phase of bioluminescence modulation (t > 1 s, Figures 2, 4, and 5) is consistent with an electrochemically induced depletion of H+ within the film caused by reduction of protons at the gold electrode interface and their subsequent replenishment by the fast dissociation of water within the film together with the buffering capacity of the bulk solution. The fast kinetic phase of electromodulation (
