Anal. Chem. 2000, 72, 4914-4920
Laser Interference Pattern Ablation of a Carbon Fiber Microelectrode: Biosensor Signal Enhancement after Enzyme Attachment Steven E. Rosenwald, Wilbur B. Nowall, Narasaiah Dontha, and Werner G. Kuhr*
Department of Chemistry, University of California, Riverside, California 92521
Fluorescence microscopy was used to visualize the accumulated fluorescent product of the enzyme alkaline phosphatase to indicate where active covalently bound enzyme remained on the surface after application of a Nd: YAG laser interference pattern to a surface that was first globally derivatized with the covalently bound enzyme. The electrochemical kinetics of the same carbon fiber surface were examined through the electrogenerated chemiluminescence of Ru(bpy)32+ to determine that electron-transfer sites were indeed segregated from the enzyme-binding sites. The enzyme-derivatized areas are determined to be separate and distinct from the areas of enhanced electron transfer. Two other enzymes, glucose oxidase and malic dehydrogenase, were then covalently bound to carbon fiber microelectrode surfaces in order to verify the change in detection limit of their respective cofactors, NADH or H2O2, under a variety of surface conditions. The S/N of an enzyme-modified electrode after laser interference pattern photoablation and electrocatalytic treatment is improved by more than 1 order of magnitude over that observed at an electrode that is globally enzyme modified. Sensitive and reliable discrimination of the analytical signal ascribed to one specific chemical species in a real-life sample, for example, in vivo detection of neurotransmitters or monitoring of glucose in the blood, is invariably complicated by the complex matrix of the systems monitored. Similarly, in situ detection of an analyte without prior treatment (e.g., separation) to remove interfering species continues to be one of the most formidable challenges facing analytical chemists today. There are relatively few analytical techniques that have the requisite sensitivity, selectivity, probe size and, more importantly, temporal resolution to make physiologically relevant measurements in vivo.1 In fact, the only analytical techniques that have found widespread use for the measurement of neurotransmitter dynamics in vivo involve the use of rapid electrochemical techniques and microelectrode sensors (electrodes of micrometer dimensions). Amperometric biosensors, where the current produced during the oxidation or reduction of the product or reactant from the enzyme-catalyzed reaction is measured,2 can also be used for such measurements if they possess these properties. The necessity for a fast time response coupled with the need for durability and stability after (1) Stamford, J. A.; Justice, J. B. Anal. Chem. 1996, 68, A359-A363. (2) Pantano, P.; Kuhr, W. G. Electroanalysis 1995, 7, 405-416.
4914 Analytical Chemistry, Vol. 72, No. 20, October 15, 2000
implantation requires that the enzyme be covalently immobilized near the electrode surface. However, the usefulness of the immobilized enzyme electrode depends on factors such as the method of immobilization, the chemical and physical conditions of use (pH, temperature, ionic strength of the sample, long-term stability of the biocomponent, interfering species, etc.), the activity and stability of the enzyme once immobilized, the stability of the electrochemical sensor, and the response time and the storage conditions required. We describe here modifications to the basic design of an amperometric biosensor (enzyme-modified electrode) to enhance both efficiency and speed of measurement. Biotin/avidin chemistry has been used extensively for the immobilization of proteins at electrode surfaces.2 The tetravalency of avidin for biotin allows the construction of a “molecular sandwich” that forms a biotin/avidin/biotin tether. Avidin can be coupled to a biotinylated enzyme that has the appropriate characteristics (i.e., substrate consumption and product formation) needed for chemical selectivity. This approach worked extremely well for oxidase enzymes and allowed the production of sensors for peroxide and glucose with response times less than 200 ms.3,4 Carbon fiber microelectrodes can also be covalently modified with dehydrogenase enzymes (including glutamate, alcohol, and glyceraldehyde 6-phosphate dehydrogenase), and these sensors also show subsecond response times.5,6 Unfortunately, the procedures used for the derivatization of the electrode surface reduce the rate of electron transfer for the redox mediator NADH,6-8 decreasing the sensitivity of the measurement dramatically. Thus, at least two principal factors influence the response of an enzymemodified electrode: the amount of enzyme present near the electrode surface and the availability of electron-transfer sites. Increasing the amount of enzyme attached to the surface will increase the amount of electrochemically detectable cofactor produced in the enzymatic reaction. This enzymatically produced cofactor then needs an available electron-transfer site for the electrochemical reaction to take place. (3) Pantano, P.; Morton, T. H.; Kuhr, W. G. J. Am. Chem. Soc. 1991, 113, 18321833. (4) Pantano, P.; Kuhr, W. G. In Monitoring Molecules in Neuroscience; Rollema, H., Westerink, B., Drijfhout, W. J., Eds.; Krips Repro: Mepples, The Netherlands, 1991; pp 117-120. (5) Pantano, P.; Kuhr, W. G. Proc. Electrochem. Soc. 1992, 92-1, 829-830. (6) Pantano, P.; Kuhr, W. G. Anal. Chem. 1993, 65, 2452-2458. (7) Hayes, M. A.; Kuhr, W. G. Anal. Chem. 1999, 71, 1720-1727. (8) Rosenwald, S. E.; Dontha, N.; Kuhr, W. G. Anal. Chem. 1998, 70, 11331140. 10.1021/ac000442t CCC: $19.00
© 2000 American Chemical Society Published on Web 09/08/2000
In general, it is necessary to immerse the electrode surface in a solution containing a redox enzyme at some point in any procedure used for the derivatization of an enzyme-modified electrode surface. In addition to any covalent chemistry performed, this also results in adsorption of the protein to the surface. It has been shown that this protein adsorption causes a reduction in the rate of electron transfer for the redox mediator NADH, thus decreasing the efficiency of the measurement.6,8 As has been demonstrated, photoablation of a carbon fiber electrode in a spatially defined fashion by use of the interference pattern generated with a Nd:YAG laser is effective in restoring facile electron transfer to a carbon fiber microelectrode that has been fouled by adsorption of a protein to its surface.8 The remaining enzyme is spatially segregated from and directly adjacent to active electron-transfer sites on the same electrode surface. The intensity of the laser must be large enough that the electron transfer is completely restored in the areas of interference maximums and small enough in the areas of interference minimums to ensure that it is below the deactivation threshold for the enzyme. In this work, the interference pattern generated with a Nd:YAG laser is used to reactivate electron transfer at a carbon fiber electrode after it has undergone global covalent modification with any of three different classes of enzymes (phosphatase, oxidase, or dehydrogenase). The derivatized carbon electrode surface was characterized with respect to electron-transfer properties and enzyme activity utilizing analytical methodology that has both high spatial resolution and high sensitivity. Fluorescence microscopy with CCD detection was used to examine the distribution of the immobilized protein by visualizing the distribution of a fluorescent product that is produced only where active enzyme (alkaline phosphatase) exists on the surface.9,10 The spatial distribution of active electron-transfer sites was visualized with electrochemically generated chemiluminescence (ECL)6,11 of the same surface to gain a more complete understanding of the spatial relationship between electron-transfer sites and enzymederivatized sites. Once the surface was characterized, two other classes of enzyme commonly employed in the fabrication of electrochemical biosensors, a dehydrogenase and an oxidase, were immobilized to verify the utility of the laser ablation technique for biosensor signal enhancement. A robust enzyme from each family, glucose oxidase (GOX) and malic dehydrogenase (MDH), were covalently bound to the carbon fiber microelectrode surface. Flow injection fast-scan cyclic voltammetry was then used to demonstrate the subsecond response due to the presence of the covalently bound enzyme. The efficiency of electron transfer at the patterned electrode was determined by measuring the difference in detection limit of their respective cofactors (H2O2 or NADH) upon injection of the corresponding substrate (glucose or malic acid), both after global modification and after application of the laser pattern. Further enhancement is then demonstrated following treatment to form an “electrocatalytic surface”, a process that is possible only after ablation of the globally derivatized surface with the Nd:YAG laser interference pattern. (9) Dontha, N.; Nowall, W. B.; Kuhr, W. G. Anal. Chem. 1997, 69, 2619-2625. (10) Huang, Z. J.; Terpetschnig, E.; You, W. M.; Haugland, R. P. Anal. Biochem. 1992, 207, 32-39. (11) Hopper, P.; Kuhr, W. G. Anal. Chem. 1994, 66, 1996-2004.
EXPERIMENTAL SECTION Chemicals. Biotinylated alkaline phosphatase was obtained from Vector (Burlingame, CA). Tris(2,2′-bipyridyl)dichlororuthenium(II) hexahydrate and tripropylamine were obtained from Aldrich (Milwaukee, WI). Biotinamidocaproic acid 3-sulfo-Nhydroxysuccinimide ester (NHS-LC-biotin), 1-ethyl-3-[(dimethylamino)propyl]carbodiimide (EDC), ExtrAvidin, mitochondrial L-malic dehydrogenase from porcine heart (MDH), L-malic acid, and β-nicotinamide adenine dinucleotide, reduced form (NADH) were obtained from Sigma (St. Louis, MO). Avidinylated glucose oxidase (A-GOX) was purchased from Pierce (Rockford, IL) and used as received. The enzyme-linked fluorescence detection kit (ELF-97) was obtained from Molecular Probes (Eugene, OR). Hydrogen peroxide was obtained from Acros Organics (Pittsburgh, PA). Succinic acid was obtained from J. T. Baker Chemical Co. (Phillipsburg, NJ), and sodium chloride, hydrochloric acid, and sodium hydroxide were obtained from Fisher (Fairlawn, NJ). Carbon fiber (32-µm diameter) was obtained from Textron Specialty Materials (Lowell, MA). Jeffamine ED-2003 was obtained from Huntsman (Conroe, TX). All chemicals were used as received unless otherwise noted. Microelectrode Fabrication. Briefly, fabrication of a carbon fiber microelectrode is accomplished through insertion of a 32µm carbon fiber into a borosilicate glass capillary (World Precision Instruments, Inc., Sarasota, FL). This assembly is pulled to a sharp point with a Narishige PE-2 micropipet puller, and the sealed glass tip is cut to release back pressure. Next, the capillary is backfilled with Epo-Tek 828 epoxy (Epoxy Technology, Billerica, MA), a copper wire is inserted to create an electrical connection by direct contact, and the epoxy is cured at 60 °C. The microelectrode is then cleaved, polished, and beveled at 30 °C on a coarse diamond abrasive plate (Sutter Instrument Co., Novato, CA) or 2000-grit sandpaper affixed to a standard 12-V fan motor used to form a rotating platform. Electrochemical pretreatment (ECP) of microelectrodes consists of a 50-Hz triangular wave in the potential range -0.2 to +2 V for 9 s. Preparation of Biotinylated MDH. Phosphate buffer (0.15 M NaCl, 0.1 M Na2HPO4, pH 7.4) was prepared and titrated to the correct pH with concentrated HCl before cooling to 4 °C. A 300-µL aliquot of 15 mM NAD+ and 300 µL of 30 mM malic acid in pH 7.4 phosphate buffer are added to 150 µL of MDH suspension (used as received) and the resultant mixture was gently rocked for 10 min at 4 °C. A solution of 2 mg of NHS-LCbiotin dissolved in 250 µL of pH 7.4 phosphate buffer was prepared just prior to its addition to the MDH solution. The resulting solution was allowed to react for 1 h and then the reaction was quenched with the addition of 12 mg of lysine, which was allowed to react for 10 min prior to dialysis. This reaction solution was transferred to dialysis tubing [Spectra/Por CE (cellulose ester) membrane MWCO 12 000-14 000 (Spectrum Medical Industries, Inc., Houston, TX)] and placed in a 1-L bath of pH 7.4 phosphate buffer with gentle stirring. This buffer is changed 3 times (after 1.5 and 7 h, the final buffer is left overnight, ∼12 h). The dialysate is then transferred to a plastic sealed microcentrifuge tube for storage at 4 °C. Enzyme Derivatization of the Microelectrode Surface. Covalent attachment of enzymes utilizing biotin/avidin technology Analytical Chemistry, Vol. 72, No. 20, October 15, 2000
4915
has been described previously.3,12 In this work, a Nd:YAG laser pretreatment of the carbon surface was performed, or the electrodes were electrochemically pretreated as above, to increase the density of surface carboxylate groups.13 These carboxyl groups are then activated by reaction with 30 mg/mL EDC and 5 mg/ mL NHS in phosphate buffer (0.15 M NaCl, 0.10 M Na2HPO4, pH 6.5) for 1 h at 60 °C. The electrodes are next dipped in a solution of 86 mg/mL Jeffamine ED-2003 in phosphate buffer (0.15 M NaCl, 0.1 M Na2HPO4, pH 7.4) for 1-2 h at 60 °C. This attaches a hydrophilic spacer arm to the electrode surface via an amide bond. Following a deionized water wash, the electrodes are placed in a 5 mg/mL solution of NHS-LC-biotin in pH 7.4 phosphate buffer for 2 h at room temperature to attach biotin to the free terminal amines on the surface-bound Jeffamines. After another deionized water wash, the electrodes are placed in a 0.7 mg/mL solution of Extravidin in pH 7.4 phosphate buffer for at least 2 h on a rocker at 4 °C. Following another deionized water wash, the electrodes are placed in a biotinylated enzyme solution (0.5 units/ mL for alkaline phosphatase and 2 mg/mL for MDH) on a rocker at 4 °C overnight (10-12 h). A-GOX was attached to the biotinylated electrode surface by incubation of the electrode with 0.5 unit/mL A-GOX on a rocker at 4 °C overnight (10-12 h). Laser Interference Patterned Ablation (LIPA). A lasergenerated interference pattern was used for ablation of the microelectrode surface using a Nd:YAG laser (Laser Photonics model MYL-100) operating at 19 mJ/pulse at 1064 nm for 7 ns. The generation of a laser interference pattern has been described before so only a brief description is given here.8,9 Near-IR light (1064 nm) was sent through a beam splitter (50% transmittance at 1064 nm at 45°, from Melles Griot, Carlsbad, CA) and reflected by a mirror to make two nearly parallel identical beams. The placement of the beams is fine-tuned to allow complete overlap of the mode structure of the laser spot. This produces a welldefined pattern of lines across the electrode surface, where the spacing between points of positive interference (D) can be approximated by the Bragg equation: nλ ) 2D sin(θ/2), where λ is the wavelength, θ is the angle between the beams, and n is order. The recombination of the two beams in this manner generates an interference pattern with ∼2-µm spacing when the angle between the 2 beams is 15°. Attenuation of the beam was accomplished by inserting one or more 48% transmittance (at 1064 nm) neutral density filters (Hoya Optics, San Jose, CA) before the beam splitter. The interference pattern is easily and repeatably focused at the tip of a microelectrode through the use of a target and micropositioner system. The tip is beveled on the polishing wheel at a predetermined angle of 30°, and the microelectrode is then placed in a holder (built in-house). This holder is placed in a clamp that is indexed at 30° to the plane of the crossed laser beams and attached to a three-axis micropositioner (Newport). A microscope slide is coated with a thin layer of soot from a butane lighter (Bic, Milford, CT) on one side. A laboratory clamp holds the slide at the focal point of the crossed beams with the blackened side away from the laser source. A 7-ns pulse from the laser is triggered, and the beam removes the soot from the cover slip where it is struck by the beam. The micropositioner is then used to bring
the tip directly behind the “hole” that has been removed from the soot to correctly position it in the crossed beams. Next, the cover slip is removed, and a pulse is applied directly to the tip. Safety Considerations: Pulsed Nd:YAG laser energy is dangerous and standard laser laboratory safety equipment should be used, including protective eyewear. Cyclic Voltammetry. Cyclic staircase voltammetry was performed with an EI-400 potentiostat (Ensman Instruments, Bloomington, IN) where all waveforms were generated and currents acquired via an 80586 personal computer using an A/D-D/A interface (Labmaster DMA, Scientific Solutions, Solon, OH). A flow injection analysis (FIA) system consisted of a pneumatic actuator (Rheodyne, model 5701) controlled via a solenoid valve (Rheodyne kit, model 7163) as described previously. FIA facilitates the generation of a “background-subtracted” voltammogram, generated by subtracting a background voltammogram from one containing the signal of interest.14 All voltammograms shown were performed at a scan rate of 100 V/s and were backgroundsubtracted. Optical Instrumentation. The beveled surfaces of the carbon fiber microelectrodes were positioned face-up toward an Epiplan Neofluar 40X (water immersion) objective. Electrode surfaces were imaged with an epifluorescence microscope (Zeiss Axioskop; Thornwood, NY) equipped with a 100-W Hg arc lamp for epi-illumination and a 50-W halogen lamp for transmitted illumination. All images were collected in a darkened room with a cooled SpectraSource MCD-600 CCD. Images were collected with SpectraSource software and saved on a personal computer (PC). Image data processing was done with Spyglass Transform 2D (Spyglass Software, Champaign, IL) imaging software. Enzyme-Linked Fluorescence. Fluorescence images of the fluorescence of a commercially available phosphatase substrate (ELF-97, Molecular Probes) were obtained with the optical instrumentation described above. This substrate produces an insoluble fluorescent product upon hydrolysis and precipitates at the site of formation.9 Electrode surfaces were derivatized with biotin-conjugated alkaline phosphatase as described and LIPA was performed at 25 MW/cm2. The ELF-97 solution was made according to the supplier’s protocol immediately before the electrode tips were immersed to allow product formation. The microelectrode was placed in a microelectrode holder (built inhouse) and the assembly was placed in a home-built stage to hold it at a 30° angle to the plane of illumination. The entire assembly was rotated until the microelectrode surface was oriented with its face toward the objective. Microscopic surface features were brought into focus with light from the Hg arc lamp after it passed through two 50% neutral density filters with a 1-ms CCD acquisition time and a camera gain setting of 1. Fluorescent images of modified carbon surfaces were obtained by passing the light from the Hg arc lamp through an excitation filter specific for the DAPI absorption band (355 ( 40 nm) and collecting all fluorescence at 450 ( 65 nm with a 2-s camera collection time. All fluorescence images were acquired in the central zone of the CCD with camera gain setting of 1. Electrochemically Generated Chemiluminescence. ECL images of the oxidation of Ru(bpy)32+ was monitored at a laser
(12) Pantano, P.; Kuhr, W. G. Anal. Chem. 1993, 65, 623-630. (13) Pantano, P.; Kuhr, W. G. Anal. Chem. 1991, 63, 1413-1418.
(14) Kristensen, E. W.; Kuhr, W. G.; Wightman, R. M. Anal. Chem. 1987, 59, 1752-1757.
4916
Analytical Chemistry, Vol. 72, No. 20, October 15, 2000
photopatterned carbon fiber microelectrode surface with the system described above with these specific changes. The patterned microelectrode was placed in a microelectrode holder (built in-house) with an Ag/AgCl reference electrode loop attached so as to surround the microelectrode tip and hold a drop of the appropriate immersion fluid. This microelectrode holder was placed in a home-built stage to hold it at a 30° angle to the plane of the beams, and the entire assembly was rotated until the microelectrode surface was oriented with its face toward the objective. The CCD was first focused on the surface using white light and then all filters were moved out of the path of the CCD and all light sources were shut down. A 1-s pulse of the appropriate voltage was applied to the electrode, with a Stanford Research waveform generator (SRS Sunnyvale, CA) through an EI-400 potentiostat (Ensman Instruments, Bloomington, IN) operating in a two-electrode configuration using the “B” channel, in the middle of the 3-s camera integration time. The CCD had a gain of 10 for this experiment. Formation of an Electrocatalytic Surface. A previously published procedure15 was modified in that LIPA produced the NADH prewave, which is necessary for activation and formation of an electrocatalytic surface. Enzyme-derivatized carbon surfaces were freshly patterned and then examined for the presence of an adsorptive prewave (∼530 mV) for the oxidation of NADH. The presence of this wave was determined by cyclic voltammetry from 0 to 1100 mV (100 V/s) and by FIA of 100 µM NADH. Next, the electrode was subjected to a sine wave excitation (1 V(p-p), 50 Hz, 600 mV bias) from a function generator (Stanford Research Systems; Sunnyvale, CA) for 7-10, 5-s injections of 100 µM NADH at 38 °C, with a 1 min pause between each injection. Finally, the electrode was exposed to 1 mM peroxide, readjusted to pH 7.4, for between 7 and 10 injections under the same conditions. The electrode was then retested with 100 µM peroxide under CV conditions to confirm the production of the electrocatalytic wave for the oxidation of NADH at ∼474 mV and for the oxidation of hydrogen peroxide at ∼1050 mV. RESULTS AND DISCUSSION The dehydrogenase family of enzymes is attractive for amperometric biosensors, since their activity is linked to the electroactive cofactor NADH. In operation, L-malic acid is enzymatically oxidized to oxaloacete as NAD+ is concurrently reduced to NADH by malic dehydrogenase. Dehydrogenase-modified electrodes (including glutamate, glyceraldehyde 6-phosphate, and alcohol dehydrogenase) also show subsecond response times,5,12 but the procedures used for the derivatization of the electrode surface dramatically affect the sensitivity for the redox mediator NADH.6,11 As observed previously, the rate of electron transfer for NADH at carbon surfaces can be quite fast and voltammetry at fast scan rates (i.e., 100 V/s) is very well defined at bare and electrochemically pretreated carbon fibers.16 Consistent and reproducible voltammetry of NADH was observed at these scan rates (100 V/s), but the reversibility of electron transfer (as manifested by increased overpotential and decreased current) deteriorated significantly after derivatization.7 This was attributed to the loss (15) Nowall, W. B., and Kuhr, W. G. Electroanalysis 1997, 9, 102-109. (16) Kuhr, W. G.; Barrett, V. L.; Gagnon, M. R.; Hopper, P.; Pantano, P. Anal. Chem. 1993, 65, 617-622.
Figure 1. Visualization of active enzyme on a surface through enzyme-linked fluorescence. The ELF-97 substrate is soluble and only weakly fluorescent when phosphorylated. The covalently bound enzyme APase dephosphorylates the substrate to the corresponding alcohol, which is insoluble in aqueous solution and strongly fluorescent. By allowing this insoluble enzyme product to accumulate,9 it is possible to indirectly visualize the location of active APase by fluorescence microscopy.
Figure 2. Three images of the same carbon fiber microelectrode. The surface was globally modified with covalently bound alkaline phosphatase before being exposed to the Nd:YAG laser interference pattern at 50% power. After the laser shot, the electrode was dipped in an ELF-97 solution to produce the fluorescent alcohol product. As shown in Figure 1, only areas that contain active alkaline phosphatase are fluorescent. (A) Fluorescence image of the reaction product of immobilized alkaline phosphatase on the surface of a 32-µm carbon fiber microelectrode. As expected, the fluorescent product is seen in alternating stripes unaffected by the Nd:YAG laser interference pattern. (B) Electrogenerated chemiluminescence of Ru(bpy)2+ at 1000 mV (vs Ag/AgCl) displays the fast electron-transfer expected in areas ablated by the Nd:YAG laser interference pattern. (C) Composite image combines left and right images to demonstrate the perfect anticorrelation of ECL (fast electron transfer) and fluorescence (immobilized active enzyme).
of electron-transfer sites due to overloading the surface with protein.12 The duration of the initial step of the derivatization procedure, activation of surface carboxylates with EDC, had the most impact on sensor response and on the voltammetry of NADH. A 14-fold decrease in the EDC reaction time resulted in a 10-fold increase in sensitivity.5 This resulted in reduced coverage of the biotinylated Jeffamine and ExtrAvidin across the electrode surface, which in turn reduced the number of sites of attachment of the biotinylated enzyme but increased the availability of electroactive sites for the oxidation of NADH. Rather than rely Analytical Chemistry, Vol. 72, No. 20, October 15, 2000
4917
Figure 3. Enzymatically generated NADH response at various stages of electrode fabrication. (A) Response of a globally derivatized electrode to injections of 3 M malic and succinic acid. (B) Response of the same microelectrode after LIPA treatment to injections of malic and succinic acid. After application of one laser shot at 50% power the signal-to-noise ratio rises to 10:1. (C) Comparison of responses to injection of 3 mM malic acid after global derivatization (dotted line), after LIPA treatment (solid line), and after application of electrocatalytic modification to LIPAtreated surface (dashed line). (D) Demonstration of enzymatic selectivity for L-malate via injection of 3 mM succinic acid. No change in current response is detected at any stage of electrode fabrication. The response to 3 mM succinic acid at one electrode that was globally derivatized (dotted line), then LIPA-treated (solid line), and a different electrode that was derivatized, LIPA-treated and electrocatalytically modified are shown.
on the random nature of this bulk process it would be preferable to be able to activate specific parts of the electrode surface (the freshly cleaned surface would be available for facile electron transfer), while leaving other regions of the derivatized surface untouched (leaving covalently derivatized enzyme in place.) Location and Verification of Active Enzyme on a Surface through Enzyme-Linked Fluorescence. As pictured in Figure 1, the ELF-97 substrate (Molecular Probes) is soluble and only weakly fluorescent when phosphorylated. The enzyme alkaline phosphatase (APase) dephosphorylates the substrate to the corresponding alcohol, which is insoluble in aqueous solution and strongly fluorescent. By allowing this insoluble enzyme product to accumulate, it is possible to indirectly visualize the location of active APase by fluorescence microscopy, since the product collects only where the active enzyme is present.9, 10 A carbon fiber microelectrode surface was first derivatized with APase by the EDC/Jeffamine/biotin/avidin/biotinylated alkaline phosphatase method that has been described and then was exposed to one pulse from the Nd:YAG laser interference pattern at a power density of 25 MW/cm2. After the laser pulse, the tip was immersed in a solution of ELF-97 for 30 min at 4 °C to allow accumulation of enzyme product only where active enzyme 4918 Analytical Chemistry, Vol. 72, No. 20, October 15, 2000
remained on the surface. The fluorescence image seen in Figure 2A clearly shows that active enzyme remains on the surface after the laser pulse in ∼1-µm-wide bands running parallel to each other across the carbon surface. Electrochemical Kinetics at Patterned Microelectrodes. Engstrom and co-workers have shown that the spatial distribution of the luminol ECL generated at an electrode surface reflects the spatial distribution of the current density of the ECL-initiating reaction.17 Microscopic visualization of the light emitted by the ECL reaction allows the discrimination of electron-transfer rates based on the overpotential associated with a specific microenviroment. A potential is applied to a LIPA-treated electrode such that it will cause the photochemical reaction of the lumophore to occur at the substrate, but not at the protein-covered section (which has sluggish electron-transfer properties and a higher overpotential). Thus, an image of the electron-transfer kinetics of the micropatterned surface can be obtained by looking at the bright and dark areas of the carbon electrode. The ECL intensity obtained in this manner can be correlated on a submicrometer (17) Vitt, J. E.; Johnson, D. C.; Engstrom, R. C. J. Electrochem. Soc. 1991, 138, 1637-1643.
scale to the spatial distribution of the previously imaged fluorescent components present across the carbon surface.11 A notable increase in the rate of electron transfer for Ru(bpy)32+ is observed for microelectrodes that have been treated with LIPA at power levels of 25 MW/cm2 or greater. The ECL image shown in Figure 2C provides evidence for enhanced electron transfer at the laser-activated surface at 900 mV vs Ag/AgCl on the same carbon fiber microelectrode surface as that seen in Figure 2A. LIPA also seems to remove enzyme from the carbon surface in the same region that electron-transfer kinetics are enhanced. The left half of the fluorescence image (Figure 2A, accumulated enzyme product) is combined with the right half of the ECL image (Figure 2C, enhanced electron transfer) of the same electrode surface to form the composite image seen in Figure 2B. As shown in the figure, wherever there is high fluorescence intensity (i.e., accumulation of enzyme product), there is very low intensity for luminescence (i.e., very sluggish electron transfer at those sites). Conversely, wherever there is low fluorescence (i.e., freshly laserablated carbon), there is faster electron transfer at the surface. This indicates that areas of enhanced electron transfer due to laser activation are interspersed between unaffected areas of active covalently bound enzyme. We have previously shown that LIPA can remove up to 100-200 nm of carbon from the electrode surface,8 so it seems reasonable to assume that the alkaline phosphatase has been completely removed from the surface in this process. The spatial segregation of these two processes, electron transfer and enzyme attachment, can greatly facilitate the design of fast enzyme-based sensors. Electrochemical Detection of Enzyme-Generated NADH. The entire surface of carbon fiber microelectrodes was derivatized with MDH using the EDC/Jeffamine/biotin/avidin/biotin/enzyme protocol described previously. Fast-scan cyclic staircase voltammetry (100 V/s) was used for the indirect detection of malic acid by monitoring the oxidation of NADH at three different MDH globally modified surfaces: unpatterned, LIPA-treated, and LIPAtreated/electrocatalytically modified carbon fiber microelectrodes (Figure 3). The response of the globally derivatized electrode to a 3 mM injection of malic acid (vs the response to a 3 mM control injection of succinic acid, which is not oxidized by MDH) is shown in Figure 3A. The malic acid response has a S/N of 5 at this concentration and is only slightly above the LOD (1.8 mM). Application of LIPA to the same electrode (1 pulse, 50 MW/cm2) dramatically improves the electrochemical reversibility of NADH oxidation and improves the S/N for enzyme-generated NADH, producing a S/N of 10 for the same concentration of malic acid (Figure 3B). Since it is necessary to apply the electrocatalytic surface immediately after laser patterning, it is not possible to obtain voltammetry under all three conditions (globally derivatized with enzyme, derivatized/patterned, and derivatized/patterned/ EC surface) at the same electrode. The cyclic voltammetric response of an electrocatalytic, LIPA-treated electrode toward enzyme-generated NADH produces a surface that has a LOD of 180 µM for malic acid (Figure 3C). This is over 1 order of magnitude better than that observed at the globally derivatized surface. At the same time, the control response (i.e., 3 mM succinic acid) remains virtually unchanged. Figure 3D demonstrates the similarity in response to injections of 3 mM succinic acid at one electrode that was globally derivatized (dotted line)
Figure 4. Fast-scan cyclic voltammograms (100 V/s) for the indirect detection of glucose by monitoring the oxidation of hydrogen peroxide at patterned and unpatterned glucose oxidase enzyme-modified carbon fiber microelectrodes. (A) Response to 3 mM glucose injection at a carbon fiber microelectrode with no pattern but with an electrocatalytic surface which was then modified over the entire surface with enzyme. (B) Response to 100-µM injection of hydrogen peroxide at an enzyme-modified, LIPA-treated, electrocatalytically modified carbon fiber microelectrode. (C) Response to 3 mM glucose injection at the same electrode shown in (B).
or laser patterned (solid line) and a different electrode that was derivatized, patterned, and electrocatalytically modified. Electrochemical Detection of Enzymatically Generated Peroxide. The entire surface of a carbon fiber electrode was derivatized with glucose oxidase using the EDC/Jeffamine/biotin/ avidin/biotin/enzyme protocol described previously. Fast-scan cyclic voltammetry (100 V/s) was used for the indirect detection of glucose by monitoring the oxidation of hydrogen peroxide at patterned and unpatterned glucose oxidase enzyme-modified carbon fiber microelectrodes (Figure 4). The response to 3 mM glucose injection at an enzyme-modified carbon fiber microelectrode with no pattern but with an electrocatalytic surface is shown in Figure 4A. The voltammogram is very nondescript and has little to distinguish it from background current signals. Fast-scan staircase voltammetry (100 V/s; 0-1100 mV vs Ag/ AgCl) of the flow injection of 1 mM peroxide at a freshly polished and electrochemically pretreated bare carbon fiber microelectrode shows a nondescript voltammogram for peroxide.15 The response to 100-µM injection of hydrogen peroxide at an enzyme-modified carbon fiber microelectrode, but with a photoablation pattern and an electrocatalytic surface, is shown in Figure 4B. This shows the dramatic impact on the voltammetric response that occurs once the surface has been patterned and that laser-ablated regions have been electrocatalytically activated. A response for peroxide was determined to be linear over 3 orders of magnitude (regression line: slope 64 405 nA/M, Y-intercept 0.0388 nA, and R ) 0.999). The detection limits for peroxide on these enzyme modified sensors were experimentally determined to be 10 µM. The activity of the glucose oxidase that was left in the “stripes” between the photoablated regions was demonstrated by monitoring the oxidation of hydrogen peroxide at the electrocatalytic region once glucose was injected. The response to 3 mM glucose injected at the same enzyme-modified carbon fiber microelectrode Analytical Chemistry, Vol. 72, No. 20, October 15, 2000
4919
(shown in Figure 4B) is shown in Figure 4C. This current corresponds to the oxidation of 140 µM hydrogen peroxide; a distinctive voltammogram characteristic of peroxide oxidation at this type of electrocatalytic surface is easily discernible. The limit of detection for glucose at this surface is ∼200 µM, demonstrating the increased efficiency of detection at the LIPA-treated, electrocatalytically modified surface. CONCLUSIONS The surface of a microelectrode that was first globally modified by the covalent attachment of an enzyme can be reactivated to electron transfer in a spatially ordered fashion by application of the interference pattern from a Nd:YAG laser. The interference pattern has sufficient contrast between areas of constructive versus destructive interference to allow reactivation of electron transfer in areas immediately adjacent to stripes of covalently bound, active enzyme. Comparison of the response of microelectrodes that have been globally modified with two different enzyme families commonly used in the fabrication of biosensors, dehydrogenase and
4920
Analytical Chemistry, Vol. 72, No. 20, October 15, 2000
oxidase, demonstrate the effectiveness of this technique. Electrocatalytic modification of the LIPA patterned surface further increases sensitivity by more than 1 order of magnitude over a globally derivatized electrode. Similar combinations of these techniques can be used to fabricate micrometer-sized arrays of a wide variety of proteins. ACKNOWLEDGMENT This work was supported by the National Institutes of Health (Grant GM44112-01A1). We recognize Matt McCormick, UC Riverside machine shop, for all his assistance in fabrication of inhouse equipment. We also thank Eric W. Kristensen for the latest revision of his voltammetry software.
Received for review April 18, 2000. Accepted August 1, 2000. AC000442T
