Anal. Chem. 2001, 73, 1037-1042
Technical Notes
Self-Referencing Ceramic-Based Multisite Microelectrodes for the Detection and Elimination of Interferences from the Measurement of L-Glutamate and Other Analytes Jason J. Burmeister and Greg A. Gerhardt*
Departments of Anatomy & Neurobiology and Neurology, Center for Sensor Technology, and Morris K. Udall Parkinson’s Disease Research Center of Excellence, University of Kentucky, Chandler Medical Center, Lexington, Kentucky 40536-0098
A self-referencing technique utilizing two microelectrodes on a ceramic-based multisite array is employed for confirmation and elimination of interferences detected by enzyme-based microelectrodes. The measurement of Lglutamate using glutamate oxidase was the test system; however, other oxidase enzymes such as glucose oxidase can be employed. One recording site was coated with Nafion with L-glutamate oxidase and bovine serum albumin (BSA) cross-linked with glutaraldehyde while the other had Nafion with BSA cross-linked with glutaraldehyde. Differences in the chemistry of the two recording sites allowed for identification and elimination of interfering signals to be removed from the analyte response. The electrode showed low detection limits (LOD ) 0.98 ( 0.09 µM, signal-to-noise ratio of 3), fast response times (T90 ∼1 s), and excellent linearity (R2 ) 0.999 ( 0.000) over the concentration range of 0-200 µM for calibrations of L-glutamate in vitro. The selectivity and dimensions of the multisite electrode allow in vivo glutamate measurements. This electrode has been applied to in vivo measurements of the clearance of locally applied glutamate and release of glutamate in the prefrontal cortex of anesthetized rats. In addition, a similar approach has been applied to the development of a microelectrode for measures of glucose. Perhaps the greatest problem in the use of microelectrodes for the detection of analytes involves the lack of specificity of the recording electrode for a given analyte. Historically, microelectrodes have not had the capabilities of a double-beam spectrophotometer to remove background interferences by using matched cuvettes and Beer’s law. With electrochemical sensors, selectivity is achieved by using coatings on the microelectrode surface that allow certain types of molecules access to the electrode surface while blocking others. The molecules that pass through this coating may be electrochemically detected at the electrode surface. There have been great advances in the design of such * Corresponding author: (phone) (859) 323-3998; (fax) (859) 323-5310. 10.1021/ac0010429 CCC: $20.00 Published on Web 01/26/2001
© 2001 American Chemical Society
sensors. Nafion has been extensively used to minimize anionic interferences from the analyte signal.1,2 Other coatings such as cellulose acetate,2 electropolymerized o-phenylenediamine,3 and lipid coatings of phosphatidylethanolamine or stearic acid4 have been utilized. Enzymes such as ascorbate oxidase have also been used to eliminate interfering contributions from ascorbate or other interferences.2 The major problems with these approaches relate to the thickness of the films that decrease the response time of the microelectrodes and failures of the coatings. In general, such microelectrodes are only selective for certain analytes and not specific for the analytes of interest. The coatings often fail by cracking, pealing, or swelling, allowing interfering species access to the electrode surface that create signals that are interpreted as analyte responses. The electrochemical equivalent of the double-beam spectrophotometer experimental system or a selfreferencing microelectrode is needed to remove interferences from the analytical signal. The possibility of developing a self-referencing microelectrode involves the use of fabricated microelectrode arrays.5-8 We recently developed and characterized ceramic-based multisite microelectrode arrays with platinum recording sites and polyimide insulation that are suitable for measurements of hydrogen peroxide and dopamine with minimal cross-talk between the recording sites (200 µm center to center).8 Instead of matching the background matrix as with double-beam spectrophotometry, similar coatings are applied to two of the recording sites on the microelectrode array. An oxidase enzyme specific to an analyte (1) Pan, S.; Arnold, M. A. Talanta 1996, 43, 1157-1162. (2) Hu, Y.; Mitchell, K. M.; Albahadily, F. N.; Michaelis, E. K.; Wilson, G. S. Brain Res. 1994, 659, 117-125. (3) Alvarez-Crespo, S. L.; Lobo-Castanon, M. J.; Miranda-Ordieres, A. J.; TunonBlanco, P. Biosens. Bioelectron. 1997, 12 (8), 739-747. (4) Ryan, M. R.; Lowry, J. P.; O’Neill, R. O. Analyst 1997, 122, 1419-1424. (5) van Horne, C.; Bement, B.; Hoffer, B. J.; Gerhardt, G. A. Neurosci. Lett. 1990, 120, 249-252. (6) Sreenivas, G.; Ang, S. S.; Fritsch, I.; Brown, W. D.; Gerhardt, G. A.; Woodward, D. J. Anal. Chem. 1996, 68, 1858-1864. (7) Moore, P. A.; Zimmer-Faust, R. K.; BeMent, S. L.; Weissburg, M. J.; Parrish, J. M.; Gerhardt, G. A. Biol. Bull. 1992 183, 138-142. (8) Burmeister, J. J.; Moxon, K.; Gerhardt, G. A. Anal. Chem. 2000, 72, 187192.
Analytical Chemistry, Vol. 73, No. 5, March 1, 2001 1037
of interest is applied to one electrode while a chemically inactive protein (bovine serum albumin) is applied to the other recording site. The enzyme-coated electrode detects the analyte in addition to everything that the inactive protein-coated electrode detects. Subtraction is used to remove the background and interfering signals from the analyte signal. Concentrations of interfering species can also be quantified. L-Glutamate was selected as the test molecule for this study because there have been many reports of glutamate sensors.1-4,9-23 In addition, there is a need for a glutamate microelectrode that has a low detection limit, fast response time, and small recording area and is free from interferences, as L-glutamate is the major excitatory neurotransmitter in the central nervous system (CNS). In this paper, we present the assembly, characterization, and testing of a self-referencing microelectrode array for in vivo brain measurements of L-glutamate. The array consists of a ceramicbased multisite microelectrode with different coatings on the recording sites. This allows interfering agents to be detected and removed from the analyte signal. The size of the self-referencing multisite microelectrode allows placement and measurements in rat brain tissue. To demonstrate that the multisite microelectrodes are compatible with other oxidase enzymes, we also present a glucose sensor that is capable of measuring glucose in the physiological range (1-20 mM). EXPERIMENTAL SECTION Reagents. All chemicals were used as received unless otherwise stated. Nafion (5% in a mixture of aliphatic alcohols and water) and ascorbate (AA) were obtained from Aldrich. Dopamine (DA), norepinephrine (NE), sodium glutamate, uric acid (UA), 3,4-dihydroxyphenylacetic acid (DOPAC), sodium chloride, glutaraldehyde, bovine serum albumin (BSA), dibasic sodium phosphate, monobasic sodium phosphate, glucose oxidase, polyurethane, tetrahydrofuran, and dimethylformamide were obtained from Sigma. L-Glutamate oxidase was purchased from Seikagku America, Inc. All solutions were prepared using distilled water which was deionized using a Barnstead/Thermolyne D8922 ionexchange column. Stock solutions of dopamine and norepinephrine were prepared with 1% perchloric acid to improve its shelf life. Solutions used for intracranial injections were prepared in 0.9% saline, adjusted to pH 7.4, and filtered before use. Glutaraldehyde was stored in a freezer at -4 °C. (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) (23)
Yao, T.; Kobayashi, N.; Wasa, T. Anal. Chim. Acta 1990, 231, 121-124. Chen, C.-Y.; Su, Y.-C. Anal. Chim. Acta 1991, 243, 9-15. Wang, A.; Arnold, M. A. Anal. Chem. 1992, 64, 1051-1055. Botre, F.; Botre, C.; Lorenti, F.; Mazzei, F.; Porcelli, F.; Scibona, G. J. Pharm. Biomed. Anal. 1993. 11, 679-686. Cordek, J.; Wang, X.; Tan, W. Anal. Chem. 1999, 71, 1529-1533. Pantano, P.; Kuhr, W. G. Anal. Chem. 1993, 65, 623-630. Kar, S.; Arnold, M. A. Anal. Chem. 1992, 64, 2438-2443. Tamiya, E.; Sugiura, Y.; Amou, Y.; Karube, I.; Ajima, A.; Kado, R. T.; Ito, M. Sens. Mater. 1995, 7, 249-259. Tamiya, E.; Karube, I. Ann. N. Y. Acad. Sci. 1992, 672, 272-277. Dremel, B. A. A.; Schmid, R. D.; Wolfbeis, O. S. Anal. Chim. Acta 1991, 248, 351-359. Hale, P. D.; Lee, H.-S.; Okamoto, Y.; Skotheim, T. A. Anal. Lett. 1991, 24 (3), 345-356. Wang, A.; Arnold, M. A. Anal. Chem. 1992. 64, 1051-1055. Kar, S.; Arnold, M. A. Anal. Chem. 1992, 64, 2438-2443. Dremel, B. A. A.; Schmid, R. D.; Wolfbeis, O. S. Anal. Chim. Acta 1991, 248, 351-359. Cordek, J.; Wang, X.; Tan, W. Anal. Chem. 1999, 71, 1529-1533.
1038
Analytical Chemistry, Vol. 73, No. 5, March 1, 2001
Electrode Fabrication. The multisite ceramic electrodes were constructed as previously described in conjunction with Thin-Film Technologies, Inc. (Buellton, CA).8 Briefly, a photomask was used to pattern 56 multisite electrodes into photoresist coated onto a cleaned 0.005 ( 0.0005 in. thick, 1-in. square polished Superstrate 996 ceramic substrate (Coors Ceramics Co., Golden, CO). The resulting microarrays were composed of 50 µm × 50 µm recording sites and connecting lines and bonding pads that extended ∼1 cm from the recording sites. The substrates were sputter-coated with an adhesion layer of 500 Å of titanium and then a 1500-Åthick layer of platinum. The recording sites, connecting lines, and bonding pads were exposed by liftoff and the remaining photoresist was removed. Individual microelectrodes were cut from the bulk ceramic wafer and wire bonded to a printed circuit board in conjunction with Hybrid Circuits, Inc. (Sunnyvale, CA). The microelectrode surfaces excluding the recording sites and bonding pads were then coated with polyimide (∼10 µm) using a proprietary brush to insulate the conducting lines. Following curing of the polyimide, an epoxy coating was used on the wire bonding sites to insulate these areas from the printed circuit board substrate. The printed circuit board was connected to instruments through a custom-fabricated connector (Quanteon, L.L.C., Denver, CO) for electrode characterization. Epoxy (5-min Devcon, Danvers, MA) was used to further insulate the interface between the ceramic electrode and the printed circuit board as well as the surface of the printed circuit board for the in vivo experiments. The anionic polymer Nafion was applied onto the electrodes to repel anions such as ascorbate while allowing cations and uncharged species to pass to the electrode surface.24-26 The electrode tips were gently dipped in the Nafion suspension (5% in aliphatic alcohols). The electrodes were then dried at 175 °C for 4 min. Nafion in conjunction with high-temperature drying has been shown to enhance selectivity for cationic neurotransmitters, such as DA over anionic interferences.27,28 Glutamate Oxidase. Glutamate oxidase was immobilized onto the coated-electrode surface using a procedure modified from Hu and Wilson.2 An aqueous solution containing 2% BSA and 0.13% glutaraldehyde was prepared fresh. Glutamate oxidase (∼0.2 mg) was dissolved in 10 µL of the BSA/glutaraldehyde solution. BSA is included in the cross-linked enzyme matrix to protect the glucose oxidase.2 A drop (∼0.5 µL) of the glutamate oxidase/ BSA/glutaraldehyde solution was coated onto one of the microelectrode sites and allowed to dry. A drop of the BSA/ glutaraldehyde solution was coated onto a second microelectrode site. The microelectrodes were stored dry at -5 °C before use and in 0.1 M phosphate-buffered saline (PBS) between measurements. Glucose Oxidase. Glucose oxidase was immobilized onto the coated-electrode surface in the same manner as glutamate oxidase. The linear range of the glucose sensors was adjusted by dip (24) van Horne, C.; Hoffer, B. J.; Stromberg, I.; Gerhardt, G. A. J. Pharmacol. Exp. Therap. 1992, 263, 1285-1292. (25) Sreenivas, G.; Ang, S. S.; Fritsch, I.; Brown, W. D.; Gerhardt, G. A.; Woodward, D. J. Anal. Chem. 1996, 68, 1858-1864. (26) Gerhardt, G. A.; Oke, A. F.; Nagy, G.; Moghaddam, B.; Adams, R. N. Brain Res. 1984, 290, 390-395. (27) Hebert, M. A.; van Horne, C. G.; Hoffer, B. J.; Gerhardt, G. A. J. Pharmacol. Exp. Ther. 1996, 279, 1181-1190. (28) Hoffman, A. F.; Lupica, C. R.; Gerhardt, G. A. J. Pharmacol. Exp. Ther. 1998, 287, 487-496.
coating the sensor tip in a 5% polyurethane solution dissolved into 98% tetrahydrofuran and 2% dimethylformamide.29 Electrode Testing and Characterization. In Vitro Tests. Constant-voltage amperometry was performed using a FAST-12 high-speed electrochemistry instrument (Quanteon, L.L.C.) using prototype software (Fast Analytical Sensor Technology (FAST), Quanteon, L.L.C.) written for simultaneous two-channel recordings. A magnetic stir plate and a thermostated water bath (Fischer Scientific) held at 37 °C were used for in vitro calibrations. The selectivity, sensitivity, detection limit, and linearity were calculated for the multisite microelectrodes. Calibrations were performed by adding aliquots of standard solutions to a beaker containing 40 mL of 0.1 M PBS. The magnetic stir plate rapidly mixed the calibration solution. For glutamate, a 250 µM AA challenge was followed by increasing the glutamate concentration in increments of 10 µM and then four increments of 40 µM to a final glutamate concentration of 200 µM. Increases in concentration of 10 µM DA, NE, DOPAC, and UA were included after the glutamate additions. Signals were normalized by subtracting the background current and then dividing by the electrode response from the addition of DA. For glucose, interferences were added to the calibration solution including 250 µM AA and UA and 20 µM increases in concentration of DA, NE, and DOPAC. After additions of interfering agents, the glucose concentration was stepped to 1, 2, 5, 10, 15, and 20 mM. The selectivity ratio, denoted as the ratio of the signals of a given concentration of analyte to an equal concentration of the interfering agent, was used to compare selectivities. High selectivity ratios (no detected change following interference addition or ratios greater than 1000:1) were included in the average as 1000. The sensitivity of each electrode was described as the slope of the current versus concentration plot in units of pA/µM. The linearity of each electrode’s response was described by the Pearson correlation coefficient (R2). The limit of detection was defined as the concentration that corresponds to a signal-to-noise ratio of 3. Root-mean-square (rms) noise levels were calculated over 25 points of the baseline. Errors are reported as standard error of the mean (SEM). In Vivo Measurements of Glutamate. All animals were handled and cared for using protocols approved by our institutional animal care and use committee. For the in vivo measurements, male Fischer 344 rats were anesthetized with urethane (1.25 g/kg, ip). The rats were prepared as previously described.30,31 A miniature 200-µm-diameter Ag/AgCl reference electrode was implanted into a site that was remote from the recording areas. A Kopf model 730 pipet puller (David Kopf Instruments, Tujunga, CA) was used to pull single-barrel glass capillaries (1.0 mm × 0.58 mm, 6 in., A-M Systems, Inc., Everett, WA). A KCl (120 mM KCl, 2.5 mM CaCl2, 35 mM NaCl, pH 7.4) or glutamate solution (200 mM, 0.9% saline, pH 7.4) was loaded into a single-barrel micropipet with a 10-15-µm-o.d. tip diameter. The micropipets were attached to the multisite recording electrodes with Sticky Wax (Kerr Brand) with a tip separation of 100 µm. The electrode and pipet were inserted into the prefrontal cortex of the rat brain. The coordinates were AP +2.7 mm, ML (1.5 mm, and DV -3.0 mm from bregma using (29) Bindra, D. S.; Zhang, Y.; Wilson, G. S. Anal. Chem. 1991, 63, 1692-1696. (30) Cass, W. A.; Zahniser, N. R.; Flach, K. A.; Gerhardt, G. A. J. Neurochem. 1993, 61, 2269-2278. (31) Friedemann, M. N.; Gerhardt, G. A. Neurobiol. Aging 1992, 13, 325-332.
Figure 1. Schematic diagram of the ceramic-based microelectrode tip. Recording sites 1 and 3 were coated with the glutamate oxidase/ BSA/glutaraldehyde mixture and sites 2 and 4 were coated with the BSA/glutaraldehyde mixture.
the atlas of Paxinos and Watson.32 Volumes of the KCl or glutamate solution were ejected from the pipets using a PPS-2 pressure ejection system (Harvard Apparatus). The resulting glutamate signals were detected at the multisite microelectrodes using constant-potential amperometry carried out using the FAST12 recording system (+0.7 V vs Ag/AgCl). RESULTS AND DISCUSSION Figure 1 is a schematic diagram of the ceramic-based multisite microelectrode tip. Laser cutting was utilized to produce the tapered design depicted in Figure 1. A rectangular design was used for most of the in vitro and some in vivo testing.8 Recording sites 1 and 3 were coated with the oxidase enzyme mixture. Sites with immobilized glutamate oxidase are e designated as glutamatedetecting sites. Sites 2 and 4 were coated with the BSA mixture. Sites without glutamate oxidase are designated as the control sites. A single glutamate-detecting site and a control site were used for the self-referencing glutamate detection. The remaining two sites could be used as an additional recording pair. We are in the process of preparing electrodes with different site-to-site geometries (i.e., side-by-side spacing and/or 50-µm site-to-site distances), as the current design may not be optimal for certain recordings in heterogeneous brain structures. The BSA/glutaraldehyde coating on the control microelectrodes made the response times of those electrodes similar to that of the glutamate oxidase/BSA/glutaraldehyde microelectrodes. Without this protein coating, the control or interferencedetecting electrodes respond more quickly to cationic compounds. Protein layers are necessary on both recording sites to minimize differences between the recording properties of the different microelectrode surfaces. In addition, the pH and temperature dependence of electrodes based on glutamate oxidase has been reported.1,2,16 Sample pH is a concern primarily for enzymatic activity. Calibrations were therefore performed in PBS with the pH adjusted to 7.4 and the temperature was held at 37 °C to approximate physiological conditions. The enzyme used in the fabrication of these sensors, glutamate oxidase from Streptomyces sp. X-119-6, has been shown to be very selective for L-glutamate.33 L-Aspartic acid had the greatest activity with 0.6% of L-glutamic acid’s activity. Glutamate oxidase showed no activity to the other 18 amino acids. A small interference (∼1%) from aspartic acid will be expected that cannot be removed using (32) Paxinos, G.; Watson, C. The Rat Brain Stereotaxic Coordinates; Academic Press: New York, 1986. (33) Kusakabe, H.; Midorikawa, Y.; Fujishima, T.; Kuninaka, A.; Yoshino, H. Agric. Biol. Chem., 1983, 47 (6), 1323-1328.
Analytical Chemistry, Vol. 73, No. 5, March 1, 2001
1039
Table 1. Recording Characteristics of the Microelectrodes for the Measurement of L-Glutamatea analysis method (N ) 9)
LOD (µM)
filtered LOD (µM)
R2
AA
single channel subtracted channels
0.98 ( 0.09 1.93 ( 0.49
0.59 ( 0.13 0.98 ( 0.30
0.999 ( 0.000 0.995 ( 0.002
389 ( 153 636 ( 218
a
selectivity ratios (XX:1) DA NE 0.15 ( 0.01 561( 174
0.43 ( 0.04 675 ( 163
DOPAC
UA
890 ( 110 779 ( 146
780 778
All data are reported as average ( SEM.
Figure 2. Raw current traces for a typical calibration of four 10 µM glutamate concentration increases and four 40 µM concentration increases of glutamate. The inset is a calibration curve for illustrating the excellent linearity of the glutamate microelectrodes.
the self-referencing technique.2,34 If elimination of aspartic acid is critical, an enzyme may be coated onto the control electrode or onto a third electrode to detect and remove the aspartate contribution to the signal. In Vitro Testing. Table 1 lists the recording properties of the microelectrodes and the selectivity ratios for measuring glutamate over AA, DA, NE, DOPAC, and UA interferences. In vitro calibrations were performed on 12 glutamate microelectrodes. Of these, nine electrodes were selected on the basis of their detection limit. Electrodes were included in the average if their LOD was less than 1.5 µM. The mean sensitivity was -1.58 ( 0.24 pA/µM. This is ∼3 times lower (worse) than the sensitivity of Nafioncoated ceramic-based microelectrodes for hydrogen peroxide.8 Diffusional losses of H2O2 and the protein diffusion barrier of the glutamate oxidase-coated electrodes likely account for this difference. For these sensors, the mean limit of detection was 0.98 ( 0.09 µM. This is consistent with the response of these sensors to hydrogen peroxide. Digital filtering of the electrode response lowered detection limits to 0.59 ( 0.13 µM. Figure 2 shows the raw data for a typical calibration with the corresponding calibration curve (inset). The first four steps correspond to 10 µM glutamate steps followed by four increases in concentration of 40 µM glutamate. Excellent linearities were observed between 10 and 40 µM (R2 ) 0.999 ( 0.000) and also between 40 and 200 µM (R2 ) 0.999 ( 0.000) for the nine glutamate microelectrodes. Selectivity. The high-temperature dried Nafion coating repels the majority of the anionic interferences such as AA. This layer alone gives the sensor a selectivity of greater than 300:1 for measuring glutamate over AA. The high concentration of AA in the brain requires this layer to be present. By lowering the (34) Torimitsu, K.; Niwa, O. Neuroreport 1997, 8, 1353-1358.
1040
Analytical Chemistry, Vol. 73, No. 5, March 1, 2001
background current due to AA, smaller changes in glutamate concentration can be detected. No change in baseline was observed for the 10 µM concentration increases of DOPAC and UA. Selectivities for these compounds were ∼800:1. However, very poor selectivities were observed over DA and NE. This is not surprising because these Nafion-coated microelectrodes have been shown to be fairly sensitive to DA.8 Nafion actually concentrates cationic species such as DA and NE. The enzyme layer acts only as a diffusion barrier. A different method is needed to remove contributions to the glutamate signal from DA and NE. Self-Referencing Operation. The array microelectrode has four recording sites that have uniform size and recording properties. Two of the sites may be utilized to remove interfering signals from cationic species or large amounts of anionic compounds. Figure 3A shows the raw response of two microelectrodes to steps in concentration of 250 µM AA, followed by 10 µM concentration steps of glutamate, DA, and glutamate. The lower trace is from a glutamate microelectrode. Small changes are seen after additions of glutamate. The change when DA is added is much larger than the glutamate change. The upper trace is from an adjacent control electrode without immobilized glutamate oxidase. As a result, this control electrode showed no response to glutamate but does respond to DA. Figure 3B shows the signals from the two electrodes when the background was subtracted and the signals were normalized to the DA change. The upper trace is again the control electrode. Figure 3C shows the resulting signal when the normalized control electrode response was subtracted from the glutamate response. The contribution from dopamine has been removed. This illustrates the potential of the subtraction mode for removing interfering signals. In the self-referencing mode, a detection limit of 1.93 ( 0.49 µM was realized. Digital filtering lowered the detection limit to 0.98 ( 0.30 µM. Table 1 lists the selectivity ratios for the test molecules. Significant improvements are seen for the elimination of DA and NE. Subtraction of the channels removed any response that is present in both channels. This is not only useful for chemical interferences but also for certain types of noise. The in vitro detection limit was increased by self-referencing. We believe this is due to the stir plate used in calibration. Current induction by magnetic stir plate can cause increased noise levels. However, in vivo results often show lowering of noise levels when subtraction is employed. The self-referencing electrode is the electrochemical equivalent of a double-beam experiment in spectroscopy. Ideally, the only difference between the two sites is that one responds to glutamate. However, it is very difficult to achieve exactly the coating thickness of Nafion and protein on two different sites. Subsequently, the two electrodes have slightly different relative responses to interfering species. However, these differences are small and can be corrected with calibration methods. In addition, the 200-µm
Figure 4. Microelectrode response in the rat prefrontal cortex to repeated 200-nL ejections of KCl. The arrows indicate ejection times.
Figure 5. Pressure ejection of L-glutamate to study glutamate uptake in the medial prefrontal cortex of the rat brain. Glutamate (200 mM) solution was applied via a micropipet at the arrows. Note the very rapid time course of the glutamate signals.
Figure 3. Electrode response versus time plots illustrating the selfreferencing capabilities of the glutamate sensor. The arrows indicate concentration steps of 250 µM ascorbic acid (AA), 10 µM glutamate (G), and 10 µM dopamine (DA). The upper traces in (A) and (B) are the control electrode responses and the lower traces are the glutamate electrode responses. (A) is the raw current response, (B) is the normalized, baseline-subtracted signal, and (C) is the subtracted signal resulting from glutamate (arbitrary units, A.U.),
spacing of the recording sites is likely not optimal for use in certain layered brain areas. Additional electrode designs are needed with closer spacing between the microelectrodes. Signal Processing and Interpretation. The data from the two microelectrodes can be processed many ways. Not all electroactive species will be present in every brain region or every experiment. Signals from the individual channels can be monitored. If the glutamate-detecting electrode responds and there is no response from the interference-detecting channel or control, the signal can be considered to be due to glutamate. For this situation, the control electrode is not used to quantify glutamate; instead it is used to qualitatively identify if an interfering signal is present. On the other hand, for experiments where neurotransmitters other than glutamate are present, the self-referencing mode may be employed. If the normalized signal from the control electrode is subtracted from the glutamate channel, the resulting signal is mostly due to glutamate. Response Time. The response time of the multisite microelectrode could not be measured with flow injection analysis because of the electrode’s geometry. However, the 90% response times for glutamate estimated from calibrations were on the order
of solution mixing (approximately 1-2 s). This is consistent with previously reported glutamate microelectrodes based on glutamate oxidase.2 More work will need to be done to fully characterize the true response time of the glutamate microelectrodes. In Vivo Measures of Glutamate Uptake and Release. Potassiumevoked release of glutamate was measured using the glutamate microelectrode. Figure 4 shows glutamate detected at the microelectrode that was released from neurons following pressure ejections of KCl from a glass micropipet located 100 µm from the electrode surface. The speed of the glutamate uptake illustrates the need for a fast sensor. Glutamate uptake is the decrease in extracellular concentration by the action of glutamate transporters located on the surface of cells translocating the glutamate into cells. Figure 5 demonstrates glutamate uptake studies in the prefrontal cortex of the anesthetized rat that were also carried out with the microarray. The pressure ejection of L-glutamate solution produced signals that were very similar to those seen from KCl stimulations. Using a first-order fit of the signals, the average K-1 for uptake of glutamate was 0.20 ( 0.01 s-1 (n ) 10) and the average uptake rate was 59.4 ( 18.3 µM/s. This is a very fast clearance of glutamate in the prefrontal cortex as compared to the rates seen for DA in DA-abundant areas such as the rat striatum.35,36 Further studies are needed to investigate the dynamics of glutamate regulation in the CNS. Glucose Measures. Sensors for the detection of glucose were assembled in a fashion similar to that for the glutamate electrodes. (35) May, L. J.; Wightman, R. M. Brain. Res. 1989, 487 (2), 311-320. (36) Garris, P. A.; Wightman, R. M. In Voltammetric Methods in Brain Systems; Boulton, A., Baker, G., Adams, R. N., Eds.; Neuromethods Series 27; Humana Press: Totowa, NJ, 1986; pp 179-220.
Analytical Chemistry, Vol. 73, No. 5, March 1, 2001
1041
Table 2. Recording Characteristics of the Microelectrodes for the Measurement of Glucosea interference bias
a
LOD (mM)
R2
AA
DA
NE
DOPAC
UA
0.64 ( 0.14
0.988 ( 0.008
-0.07 ( 0.26
0.24 ( 0.24
0.52 ( 0.24
0.47 ( 0.27
0.63 ( 0.24
All data are reported as average ( SEM. Bias in mM for 250 µM AA, 20 µM DA, 20 µM NE, 20 µM DOPAC, and 250 µM UA.
Figure 6. Raw current traces for a typical calibration where the glucose concentration is stepped to 1, 2, 5, 10, 15, and 20 mM. The inset is a calibration curve demonstrating the linearity of the glucose microelectrodes.
A mixture of 2% glucose oxidase, 2% BSA, and 0.125% glutaraldehyde was substituted for the glutamate oxidase/BSA/glutaraldehyde mixture. The linear range was adjusted by dip coating in a polyurethane solution. These electrodes responded linearly (R2 ) 0.988 ( 0.008) to glucose over the concentration range between 1 and 20 mM (Figure 6) with a LOD of 0.64 ( 0.14 mM. The mean sensitivity for the eight electrodes tested was 30.2 ( 6.5 pA/mM. Response times were again on the order of solution mixing. These results demonstrate the feasibility of the ceramicbased multisite microelectrodes for the electrochemical measurement of analytes based on oxidase enzymes. Table 2 lists the bias in perceived glucose concentration when 250 µM AA and UA and 20 µM DA, NE, and DOPAC are added. The differences in concentration between glucose and the interfering agents as well as the polyurethane coating contributed to the minimal bias. For most experiments, these errors would be acceptable without the use of the self-referencing technique. CONCLUSIONS We have used a self-referencing technique in combination with ceramic-based multisite electrodes to detect and remove interfer-
1042
Analytical Chemistry, Vol. 73, No. 5, March 1, 2001
ences from the analyte signal. Self-referencing is made possible by the spatially defined, uniform recording sites on the multisite electrodes. A sensor for the test molecule glutamate was prepared by coating one recording site with Nafion with L-glutamate oxidase and BSA cross-linked with glutaraldehyde while another had Nafion with BSA cross-linked with glutaraldehyde. Differences in the chemistry of the sites allowed interfering signals to be identified and removed from the analyte signal. These ceramicbased multisite electrodes may be applied to quantification of other analytes. Enzyme-coated electrodes have been reported for measurements of glucose, lactate, choline, and γ-aminobutyric acid (GABA).37-42 Combinations of these may be applied to the electrode surface to measure some of these analytes simultaneously. Ceramic-based multisite microelectrode are compatible with enzymatic measures of glucose. Future work will concentrate on the development of ceramic sensors with different recording site geometries to allow for recordings in numerous brain areas. ACKNOWLEDGMENT The authors acknowledge Mike Parrish and Steve Robinson for modifying the software necessary to record from the twochannel systems. The authors acknowledge Karen Giardina and Kathy Adams for their expertise in the in vivo experiments. The authors also thank the participants in the electrochemistry course at Woods Hole, MA, May 11-15, 2000 for their input. This work was supported by NSF DBI-9730899 and MH01245.
Received for review August 31, 2000. Accepted December 12, 2000. AC0010429 (37) Albery, W. J. Ciba Found. Symp. 1991, 158, 55-67. (38) Niwa, O.; Kurita, R.; Horiuchi, T.; Torimitsu, K. Anal. Chem. 1998, 70, 89-93. (39) Hu, Y.; Wilson, G. S. J. Neurochem. 1997, 69 (4), 1484-1490. (40) Shram, N. F.; Netchiporouk, L. I.; Martelet, C.; Jaffrezic-Renault, N.; Bonnet, C.; Cespuglio, R. Anal. Chem. 1998, 70, 2618-2633. (41) Netchiporouk, L. I.;. Shram, N. F.; Jaffrezic-Renault, N.; Martlet, C.; Cespuglio, R. Anal. Chem. 1996, 68, 4358-4364. (42) Hu, Y.; Wilson, G. S. J. Neurochem. 1997, 68 (4), 1745-1752.
