Concentration of Extracellular l-Glutamate Released from Cultured

Jun 1, 1996 - An online sensor with a low detection limit for l-glutamate was developed in order to monitor the change in the extracellular l-glutamat...
0 downloads 14 Views 198KB Size
Download PDF
Anal. Chem. 1996, 68, 1865-1870

Concentration of Extracellular L-Glutamate Released from Cultured Nerve Cells Measured with a Small-Volume Online Sensor Osamu Niwa,*,† Keiichi Torimitsu,† Masao Morita,† Peter Osborne,‡ and Katsunobu Yamamoto‡

NTT Basic Research Laboratories, 3-1 Morinosato, Wakamiya, Atsugi 243-01, Japan, and Department of R&D, BAS Corporation Ltd., 1-36-4 Oshiage, Sumida, Tokyo 131, Japan

An online sensor with a low detection limit for L-glutamate was developed in order to monitor the change in the extracellular L-glutamate concentration as a result of stimulated release from cultured nerve cells. The sensor consisted of a microdialysis (MD) probe fixed at the manipulator, a small-volume L-glutamate oxidase enzymatic reactor (0.75 mm i.d. and 2.5 cm long), and an electrochemical detector in a thin-layer radial flow cell with an active volume of 70-340 nL. Glassy carbon bulk or carbon film ring-disk electrodes were used as detectors by modifying them with Os poly(vinylpyridine) mediator containing horseradish peroxidase. The overall efficiency of L-glutamate detection with the sensor is 94% under optimum conditions, due to an efficient enzymatic reaction in the reactor and a high conversion efficiency in the radial flow cell. As a result, we achieved a sensitivity of 24.3 nA/µM and a detection limit of 7.2 nM (S/N ) 3). The effect of interferents such as L-ascorbic acid can be minimized effectively by applying a low potential to the electrode for hydrogen peroxide detection (0 mV) and via the ring-disk electrode geometry by using the disk for preoxidation. In the in vitro experiment, an MD probe for sampling was connected to a manipulator that controls distance between the probe and the stimulated cells. The cells were stimulated by KCl in a glass capillary or electrically with microarray film electrodes fabricated on a substrate. By using the sensor, we can monitor L-glutamate concentration changes at the submicromolar level caused by KCl stimulation of a single nerve cell and micromolar L-glutamate concentration increases caused by electrical stimulation of a brain slice. An increase in L-glutamate concentration can also be measured by positioning the probe near the cell that is connected synaptically to the stimulated cell. Various in vivo or in vitro methods have been developed for detecting neurotransmitters because the monitoring of extracellular concentration changes in the transmitters is very important for understanding the function of the nervous system. These methods include liquid chromatography (LC) or flow injection analysis (FIA) combined with microdialysis (MD) sampling,1,2 †

NTT Basic Research Laboratories. BAS Corp. Ltd. (1) Kissinger, P. T. In Microdialysis in Neuroscience; Robinson, T. E., Justice J. B., Jr., Eds.; Elsevier: Amsterdam, 1991; pp 103-115. ‡

S0003-2700(95)01154-1 CCC: $12.00

© 1996 American Chemical Society

carbon fiber microelectrodes,3-5 and online sensors for real-time measurement.6 LC has been widely used to determine various transmitters by combining it with MD sampling and derivatization.1 This method is capable of achieving a low detection limit with a small sample volume and is very useful for determining amino acids quantitatively. However, it requires time for the derivatization and column separation. Recently, real-time measurement of transmitters has attracted much attention because the time resolution is much better than that of LC combined with MD sampling. There are basically two methods for the real-time measurement of L-glutamate. One uses a microelectrode-based in vivo sensor, and the other uses an online flow sensor. Carbon fiber microelectrodes are often used for both the in vivo and in vitro measurement of catecholamine or indolamines.3,7 This technique can also be applied to determine electroinactive transmitters such as amino acids or choline by modifying the electrode with enzymes which convert the transmitters to electroactive species.8-10 The advantages of the enzyme-modified microelectrodes are fast response and the need for only a very small area for measurement. However, it is difficult to quantify the analyte reproducibly because it is hard to make electrodes with a consistent response. An in vivo electrode is less stable than an online sensor since the amount of immobilized enzyme on the electrode is small. On the other hand, online enzyme sensors have been developed consisting of a syringe pump, an MD sampling probe, an enzymatic reactor, and a detector.6,11,12 Although the response of an online system is slower than that of an in vivo modified microelectrode, the former is more quantitative, providing a higher conversion of analytes, a longer lifetime, and ease of calibration. (2) Pettit, H. O.; Justice, J. B. In ref 1, pp 117-151. (3) Wightman, R. N.; May, L. J.; Michael, A. C. Anal. Chem. 1988, 60, 769A779A. (4) Leszczyszyn, D. J.; Jankowski, J. A.; Viveros, O. H.; Diliberto, R. J., Jr.; Near, J. A.; Wightman, R. M. Biol. Chem. 1990, 265, 14736-14737. (5) Chow, R. H.; von Ruden, L.; Neher E. Nature 1992, 356, 60-63. (6) Korf, J.; Boer, J. de; Postema, F.; Venema, K. In ref 1, pp 349-366. (7) Bruns, D.; Jahn, R. Nature 1995, 377, 62-65. (8) Hu, Y.; Mitchell, K. M.; Albahadily, F. N.; Michaelis, E. K.; Wilson, G. S. Brain Res. 1994, 659, 117-125. (9) Gargulio, M. G.; Michael, A. C. J. Am. Chem. Soc. 1993, 115, 12218-12219. (10) Tamiya, E.; Sugiura, Y.; Amou, Y.; Karube, I.; Ajima A.; Kado, R.; Ito, M. Sens. Mater. 1995, 7, 249-59. (11) Albery, W. J.; Boutelle, M. G.; Galley, P. T. J. Chem. Soc., Chem. Commun. 1992, 900-901. (12) Berners, M. O. M.; Boutelle, M. G.; Fillenz, M. Anal. Chem. 1994, 66, 20172021.

Analytical Chemistry, Vol. 68, No. 11, June 1, 1996 1865

Many neurotransmitter studies have focused on L-glutamate in regard to excitatory neurotransmission in the mammalian central nervous system.13,14 L-Glutamate is one of the main neurotransmitters and plays an important role in synaptic plasticity. The long-term changes in synaptic efficacy, such as long-term potentiation (LTP)15,16 and long-term depression (LTD)17-19 of excitatory synaptic transmission, are considered to be the neuronal bases for learning and memory. Therefore, an online sensor combined with an MD probe has been studied for monitoring in vivo L-glutamate.11,12,20,21 L-Glutamate dehydrogenase or oxidase22,23 has been used for measuring L-glutamate. The oxidase has been more widely used than the dehydrogenase because it is more robust and is now commercially available.23 Glutamate oxidase oxidizes L-glutamate very efficiently in the presence of oxygen, and the generated hydrogen peroxide can be detected at a Pt- or mediator-modified electrode. Boutelle et al. developed an MD probe-based online electrochemical sensor and applied it for the in vivo monitoring of L-glutamate by adding L-glutamate oxidase to the perfusion solution.11 L-Glutamate sampled by the probe is enzymatically oxidized, and the generated hydrogen peroxide is then oxidized at a Pt electrode held at 600 mV. Recently, the sensor was improved by immobilizing the oxidase on the platinum tubular electrode in a polymer film in order to reduce the consumption of oxidase during measurement. A detection limit of 0.3 µM was reported in the presence of L-ascorbate determined by a standard solution experiment.12 Another method consisting of the MD sampling of L-glutamate and FIA was also reported by the same group.24 Obrenovitch et al. applied an online sensor consisting of an MD probe, immobilized L-glutamate oxidase, and a Pt electrode to the measurement of L-glutamate in rat brain.21 They reported a detection limit of 0.5 µM and an increase in L-glutamate concentration caused by KCl stimulation. L-Glutamate measurement in cultured nerve cells is also important for understanding the function of this transmitter at the single-cell or synaptic level. Modified carbon electrodes10 or patch sensors25 have generally been used for monitoring L-glutamate in cultured cells. A method was developed for culturing nerve cells on microarray film electrodes which selectively stimulated one or a few cells electrically while monitoring a potential pulse at other arrays.26,27 (13) Curtis, D. R.; Phillis, J. W.; Watkins, J. C. J. Physiol. 1960, 150, 656-682. (14) Johnson, J. L.; Aprison, M. H. Brain Res. 1970, 24, 285-292. (15) Bliss, T. V. P.; Douglas, R. M.; Errington, M. L.; Lynch, M. A. J. Physiol. 1986, 377, 391-408. (16) Bliss, T. V. P.; Collingridge, G. L. Nature, 1993, 361, 31-39. (17) Ito, M. Neurosci. Res. 1986, 3, 531-539. (18) Ito, M.; Sakurai, M.; Tongroach, P. J. Physiol. 1982, 324, 113-34. (19) Lindenm, D. J.; Connor, J. A. Eur. J. Neurosci. 1992, 4, 10-15. (20) Koshy, A.; Zilkha, E.; Obrenovitch, T. P.; Bennetto, H. P.; Richards, D. A.; Symon, L. Anal. Lett. 1993, 26, 831-849. (21) Zilkha, E.; Obrenovitch, T. P.; Koshy, A.; Kusakabe, H.; Bennetto, H. P. J. Neurosci. Methods 1995, 60, 1-9. (22) Blankenstein, G.; Preuschoff, F.; Spohn, U.; Mohr, K. H.; Kula, M. R. Anal. Chim. Acta 1993, 271, 231-271. (23) Kusakabe, H.; Midorikawa Y.; Fujishima, T.; Kuninaka, A.; Yoshino, H. Agric. Biol. Chem. 1983, 47, 1323-1328. (24) Boutelle, M. G.; Fellows, L. K.; Cook, C. Anal. Chem. 1992, 64, 17901794. (25) Maeda, T.; Shimoshige, Y.; Mizutani, K.; Kaneko, S.; Akaike, A.; Satoh, M. Neuron 1995, 15, 253-257. (26) Jimbo, Y.; Kawana, A. Bioelectrochem. Bioenerg. 1992, 29, 193-204.

1866

Analytical Chemistry, Vol. 68, No. 11, June 1, 1996

Therefore, by monitoring the change in L-glutamate concentration not only around the stimulated cells but also at synaptically connected cells, we may be able to understand synaptic plasticity and signal transmission through the neural network. A sensor that can detect small concentration changes and that also has a low detection limit is required for this in vitro application. This is because the nerve cell population in a cultured system is lower than that in the brain, and so it should be possible to detect smaller changes in the transmitter concentration quantitatively. The recovery of the MD probe is also low when the transmitter distribution is measured, because a small probe is needed to improve the resolution. In this paper, we report the real-time monitoring of the stimulated release of L-glutamate from cultured nerve cells with an online L-glutamate sensor. Since the L-glutamate concentration is reduced upon increasing the distance of the sensor from the stimulated cell, the MD probe and KCl capillary were fixed to a manipulator and controlled precisely. Their position was observed with an optical microscope during measurement. To improve the detection limit of the system, we tried to increase the sensitivity and lower the noise level. We made a small-volume enzymatic reactor in order to improve the conversion efficiency of L-glutamate without greatly increasing the dead volume. A thin-layer radial flow cell with an inner volume of 0.0850.34 µL was used to improve the sensitivity since the cell exhibits higher sensitivity and conversion efficiency than a conventional thin-layer cross-flow cell.28-31 The electrode surface was modified with Os poly(vinylpyridine) (Os gel) containing horseradish peroxidase (HRP)32-34 to lower the potential for hydrogen peroxide detection. Since HRP is covalently linked to the Os gel coated on the electrode, its oxidized form can be reduced electrochemically through the redox centers of the polymer network. As a result, the noise level caused by the electroactive interferent should be reduced since the electrode can reduce hydrogen peroxide even at 0 mV vs Ag/AgCl. We also studied the preoxidation of L-ascorbate in one flow cell by using a ring-disk carbon film electrode,30 which is superior to the previously reported series dual-electrode systems because the dead volume can be reduced. EXPERIMENTAL SECTION Materials. L-Glutamate oxidase from Streptomyces sp. X-119-6 was obtained from Yamasa Shoyu (Choshi, Japan). Os poly(vinylpyridine) mediator solution containing HRP34 was obtained from Bioanalytical Systems Inc. (BAS, West Lafayette, IN). Aminopropyl-CPG beads (200/400 mesh) were obtained from Electro Nucleonics Inc. (Fairfield, NJ). Phosphate-buffered saline (PBS) (Life Technologies, Grand Island, NY), laminin- and poly-D-lysinecoated plastic culture dishes (Nunc), Dulbecco’s Modified Eagle’s (27) Robinson, H. P. C.; Kawahara, M.; Jimbo, Y.; Torimitsu, K.; Kuroda, Y.; Kawana, A. J. Neurophysiol. 1993, 70, 1606-1616. (28) Bohs, C. E.; Linhares, M. C.; Kissinger, P. T. Curr. Sep. 1994, 12, 181186. (29) Huang, T.; Kissinger, P. T. Curr. Sep. 1995, 13, 114-118. (30) Niwa, O.; Morita, M.; Solomon, B. P.; Kissinger, P. T. Electroanalysis, in press. (31) Niwa, O.; Morita, M. Anal. Chem. 1996, 68, 355-359. (32) Vreeke, M. S.; Maiden, R.; Heller, A. Anal. Chem. 1992, 64, 3084-3090. (33) Ohara, T. J.; Vreeke, M. S.; Battaglini, F.; Heller, A. Electroanalysis 1993, 5, 825-831. (34) Yang, L.; Janle, E.; Huang, T.; Gitzen, J.; Kissinger P. T.; Vreeke, M.; Heller, A. Anal. Chem. 1995, 67, 1326-1331.

medium (Gibco), and L-glutamate (Funakoshi, Tokyo, Japan) were used as purchased. The external medium used for L-glutamate measurement contains 148 mM NaCl, 2.8 mM KCl, 2 mM CaCl2, 2 mM MgCl2, 10 mM glucose (Wako, Tokyo, Japan), and 10 mM HEPES (Sigma, St. Louis, MO) (pH 7.2). Papain was obtained from Sigma. Brain-derived neurotrophic factor (BDNF) was obtained from Iwaki (Tokyo, Japan). Other chemicals such as L-ascorbic acid and NGF (nerve growth factor) were purchased from Wako. Column and Electrode. L-Glutamate oxidase was immobilized on beads with glutalaldehyde. The enzyme-immobilized beads were packed in poly(ether ether ketone) (PEEK) tubes which were 2.5 cm long with an inner diameter of 0.75 mm. The inner volume of the enzymatic reactor was controlled by changing the length of the column. Glassy carbon (GC) electrodes 3 and 6 mm in diameter were purchased from BAS. The carbon film ring-disk electrode was fabricated by chemical vapor deposition (CVD) of 3,4,9,10-perylenetetracarboxylic dianhydride, photolithography, and the dry-etching technique as described previously.30,31 The diameter of the disk in the ring-disk electrode was 3 mm, and the width of the ring was 1 mm. The disk and ring were separated by a 0.5-mm gap. The Os gel was coated on the GC or ring electrode of the ring-disk carbon film electrode by the casting method. The amount of gel solution applied to the electrode was 7 µL/cm2. The array electrode was fabricated on a substrate by photolithography and the lift-off technique, on which the cells were then cultured.27 Each electrode was plated with platinum black to lower the impedance and increase the adhesion of the cell to the electrode. Online Sensor System. Figure 1 shows a block diagram of the sensor system. A CMA 102 dual-syringe pump (CMA Microdialysis, Stockholm, Sweden) was used to introduce the solutions into the system. Two different solutions can be introduced independently and mixed upstream of the enzymatic reactor. To study the basic performance of the sensor, a PBS blank solution was introduced by syringe 1, and the sample solutions containing L-glutamate and/ or L-ascorbic acid were introduced by syringe 2. The concentration of the sample solution can be changed continuously by controlling the flow from the two syringes. An MD probe which was fixed to the manipulator was inserted between syringe 1 and the merge point of the two solutions for the in vitro measurement of the L-glutamate released from the nerve cells. In this case, both syringes were filled with the PBS solution. The enzymatic reaction was stabilized by saturating the PBS in syringe 2 with oxygen just before introducing the perfused solution into the column. The enzymatic reactor was connected directly to the inlet port of a thin-layer radial flow cell. The structure of the flow cell is shown in Figure 1b. A 12-µm-thick Teflon spacer was sandwiched between the auxiliary block and the working electrodes to form a small-volume flow cell. The active volume of the cell was from 70 to 340 nL, depending on the electrode size. The electrode potentials were controlled with a BAS LC-4C potentiostat. To determine L-glutamate, the potential of the modified electrode was held at 0 mV vs Ag/AgCl. When the ring-disk electrode was used, the disk electrode was held at +600 mV for preoxidation, and the ring electrode modified with Os redox gel was held at 0 mV. The magnitude of the oxidation

Figure 1. Schematic representation of the small-volume online L-glutamate sensor. (a) Total system. (b) Thin-layer radial flow cell and GC and carbon film ring-disk electrodes. The inlet port is on the back side of the auxiliary electrode block.

current for the L-ascorbate at the ring was compared with and without potentiostating the disk electrode. Nerve Cell Preparation. We dissected the nerve cells from the cerebral cortex of 18-day-old rat embryos (Wistar) for dissociated cell preparation and from 2-day-old postnatal rats for slice culture. After dissociation with 0.02% papain, they were cultured on laminin- and poly-D-lysine-coated plastic culture dishes. The cells were maintained for 1-2 weeks in Dulbecco’s Modified Eagle’s medium containing 5% heat-inactivated fetal bovine serum and 5% heat-inactivated horse serum at 37 °C in a water-saturated 10% CO2 atmosphere. A 300-µm slice of cerebral cortex was cultured with the medium described above, where 20 ng/mL 7S NGF and 20 ng/mL BDNF were added to the medium. Measurement of Concentration of Extracellular L-Glutamate from Nerve Cells. Figure 2a shows a schematic representation of the method for real-time measurement of the concentration of extracellular L-glutamate released from cultured nerve cells or brain slice. The substrate on which the nerve cells were cultured had 64 platinum black-plated ITO microarray electrodes. The size of each electrode was 60 × 60 µm2, and it was used for electrical stimulation. Both the glass pipet for KCl stimulation and the MD probe were fixed to manipulators to control their distances from the cells or microarray electrodes to be stimulated. An optical microscope photograph of a KCl capillary and sampling probe is shown in Figure 2b. The perfusion rate was 4 µL/min, and the flow rate was increased to 16 µL/min by adding oxygen-saturated PBS buffer to supply sufficient oxygen for the enzymatic reaction. Variations in the current were measured while stimulating the cells with 100 nL of 100 mM KCl or electrically with one of the Analytical Chemistry, Vol. 68, No. 11, June 1, 1996

1867

Figure 3. Response of the sensor to PBS buffer containing 1 µM L-glutamate. The electrochemical detector is a 6-mm GC electrode modified with Os gel mediator containing HRP. The flow rate is 16 µL/min. The spikes that appear when the solution is changed are due to the pressure changes caused by the liquid switch.

Figure 2. (a) Schematic representation of L-glutamate sampling probe and glass capillary and microarray film electrode for KCl and electrical stimulation. (b) Optical microscope photograph of glass capillary for stimulation and MD probe.

microarray electrodes on the substrate (1-1.5 V for 500 ms, 200 Hz). RESULTS AND DISCUSSION Sensitivity and Detection Limit. PBS containing L-glutamate was introduced into the sensor system in order to study the sensitivity and detection limit. Figure 3 shows the typical timedependent current when 1 µM L-glutamate was introduced. A 6-mm glassy carbon electrode coated with Os gel was used for this measurement. The flow rate was 16 µL/min. The current starts to decrease because of the reduction of the hydrogen peroxide generated in the enzymatic reactor after about 1.0 min caused by switching from the blank PBS solution to the sample solution. Therefore, the dead volume of the sensor from the syringe to the detector is less than 16 µL. The peak height of the first injection of L-glutamate for 7.4 min is the same as that of the second injection. This peak is also stable for the long-term introduction of 1 µM L-glutamate as long as the PBS was saturated with oxygen, indicating that the sensor shows good reproducibility 1868 Analytical Chemistry, Vol. 68, No. 11, June 1, 1996

and stability. The peak height at the optimum condition was -24.3 nA. This value is much higher than that for the previously reported online L-glutamate sensor, which employed an L-glutamatemodified platinum electrode.12 The theoretical limiting current is -25.7 nA when 100% of the 1 µM L-glutamate is oxidized enzymatically, and all the generated H2O2 molecules are reduced at the modified electrode at a flow rate of 16 µL/min. Therefore, the sensor exhibits a conversion efficiency of 94%. This high value is due to the high conversion efficiencies of the enzymatic reactor and the disk electrode in the radial flow cell.30,31 Assuming that the noise level is proportional to the electrode area, a higher current density is a good way to lower the detection limit. Since the current density at a 3-mm electrode is higher than that at a 6-mm electrode, a 3-mm GC electrode was used to estimate the detection limit. Figure 4 shows the L-glutamate calibration curve. The sensor shows a linear range from 10-8 to 10-6 M. A further increase in the L-glutamate concentration above 5 µM causes the current to saturate. A detection limit of 7.6 nM (S/N ) 3) was achieved as a result of the high efficiency of the sensor and the low noise level because of the low potential of the Os gel-modified electrode. In our measurement of L-glutamate from cultured cells, we used a short MD probe (1 mm) in order to determine the distribution of the released L-glutamate with better resolution. Therefore, it is very important that the sensor has a low detection, because the low recovery of the probe reduces the concentration of the L-glutamate sampled in the sensor and the cell density on the substrate is low. The recovery of the sensor is only 1.9% in our system. However, this sensor can still be applied to the measurement of extracellular L-glutamate concentrations at submicromolar levels.

Figure 4. Calibration curve of L-glutamate determined with a 3-mm GC electrode modified with a mediator. The flow rate is 16 µL/min.

Selectivity of Sensor with Ring-Disk Carbon Film Electrode. It has been reported that the major problem regarding L-glutamate measurement is the existence of ascorbate, which has low oxidation potential.11 The operating potential of our electrode was 0 mV because the carbon electrode surface was modified with Os gel mediator containing HRP. Oxidation of the L-ascorbate is still observed since ascorbate is oxidized at 0 mV. Boutelle et al. used two electrochemical cells: one was a platinum tubular electrode for the preoxidation of ascorbate, and the other was an L-glutamate oxidase-modified platinum electrode for H2O2 measurement.12 The cell volume increases when the number of electrochemical cells is increased. We used ringdisk carbon film electrodes arranged in one small-volume radial flow cell. This geometry can preoxidize most of the species which are easily oxidized when the flow rate is low and the flow cell is thin. In L-glutamate measurements, the inner disk electrode only oxidizes the ascorbate, and the outer ring electrode modified with Os gel containing HRP detects H2O2 generated by the enzymatic reaction. Figure 5 shows the flow rate dependence of the L-ascorbate oxidation current at a ring-disk electrode with or without potentiostating the disk electrode. The disk potential is 600 mV, and the ring potential is 0 or 500 mV, which are typical operating potentials of Os gel-modified carbon and bare platinum electrodes for H2O2 detection. PBS buffer solution containing 10 µM ascorbic acid was used for testing the electrochemical removal of L-ascorbic acid. When the disk and ring electrode potentials are 600 and 500 mV, the limiting currents at the disk and ring are 342 and 80 nA, respectively. This indicates that >80% of the ascorbate which flowed into the cell was oxidized at the disk electrode. However, the limiting current of ascorbate is still high. When the ring potential was switched to 0 mV, the magnitude of the limiting current at 16 µL/min was 7.9 nA without using the disk. This indicates that the oxidation of L-ascorbic acid can be suppressed by using an Os gel-HRP mediator. However, the magnitude of the ascorbate oxidation current is still comparable to that of the usual hydrogen peroxide reduction current for L-glutamate measurement (-11.1 nA for 1 µM L-glutamate at the ring electrode

Figure 5. Flow rate dependence of L-ascorbate limiting current at ring-disk electrode. b, Disk current when the disk and ring electrode potentials are 0.6 and 0.5 V, respectively; O, ring current when the disk and ring electrode potentials are 0.6 and 0.5 V, respectively; 4, ring current at 0 mV when the disk electrode is not used for measurement; and 2, ring current at 0 mV when the disk electrode potential is 0.6 V.

modified with Os polymer). When the mediator was coated on the ring and the potential of the disk electrode was 600 mV, the limiting current of the ascorbate was lower than 1.0 nA. This indicates that the effect of the ascorbate can be removed effectively by combining the mediator and the ring-disk geometry in a radial flow cell with a very small active volume (340 nL). The submicromolar level detection limit is estimated by performing a calculation using the limiting currents of the 1 µM L-glutamate and 10 µM L-ascorbate standards. However, the detection limit depends on the concentration of L-ascorbate. Further modificaton of the sensor to increase the selectivity will be needed for in vivo applications. This includes enlargement of the disk electrode or roughing the disk electrode35 to preoxidize the interferent efficiently or modification of the disk electrode with a mediator to accelerate ascorbate oxidation. Change in Extracellular L-Glutamate Concentration Induced by Stimulation. In our measurement of L-glutamate released from cultured cells, we used a glass capillary containing 100 mM KCl for stimulation and kept the MD probe less than 5 µm from the cell. In this experiment, the MD probe and the KCl glass capillary were positioned very close to the cell, as shown in the Figure 2b. However, the KCl capillary was placed farther from the MD probe when we studied synaptic transmission. Since the concentration of reducing agents such as ascorbate in most of our experiment is lower than that for an in vivo system, the measurement can be made with a single disk electrode modified with HRP containing Os gel. Figure 6a shows the variation in L-glutamate concentration near the cell as a result of the stimulation of a single nerve cell in a dissociated culture system, caused by injecting 100 mM KCl from a glass microcapillary. The sampling probe was located near the stimulated cell. The variation in the limiting current caused by (35) Niwa, O.; Horiuchi, T.; Tabei, H. J. Electroanal. Chem. 1994, 367, 265269.

Analytical Chemistry, Vol. 68, No. 11, June 1, 1996

1869

Figure 6. Change in L-glutamate concentration caused by KCl and electrical stimulation. (a) KCl stimulation. (b) Electrical stimulation.

stimulation was calculated for the glutamate concentration by calibrating it with a standard L-glutamate solution sampled with the same MD probe. Without stimulation, the glutamate concentration gradually decreases during measurement. This is considered to be due to the neurotransmitter recycling (uptake) by the nonneural cells. In contrast, the L-glutamate concentration was increased 2 min after KCl injection by about 1 µM by injecting KCl, and this increase was reproduced by injecting the KCl 10 min after the first injection. The concentration change caused by the second stimulation is smaller than that by the first stimulation. This result was reproduced, and this is not caused by a decrease in the sensitivity of the sensor. This is certain because we calibrated the sensor with the glutamate standard before and after measurement, and the sensitivity of the sensor did not change during the measurement. The column and modified electrode in the online sensor can be used for a week. The L-glutamate concentration was greatly reduced upon moving the probe farther from the cell. This indicates that this method can be used to measure the L-glutamate distribution around the cells after stimulation in order to study signal transmission among neural networks.

1870 Analytical Chemistry, Vol. 68, No. 11, June 1, 1996

We also observed a variation in the L-glutamate concentration in this system as a result of electrical stimulation. In this experiment, slices of rat brain were cultured on a substrate on which platinum black-plated 64 microarray film electrodes had been fabricated. Each electrode can be used for both electrical stimulation and observation of impulse transmission. Figure 6b shows the change in L-glutamate concentration of a brain slice cultured on a microarray electrode stimulated by applying a potential to one of the array electrodes. The Lglutamate concentration increases in a way similar to that in the experiment in Figure 6a, but the increase is greater because the nerve cell population of the brain slice is larger than that of the cultured cells. The electrical stimulation could transmit the signal only to a cell which was connected synaptically. The L-glutamate concentration increase measured by the electrical stimulation is observed by positioning the probe near a cell which is connected synaptically to the stimulated cells. These results indicate that the sensor could be applied to various neurophysiological studies, including neural networks and the synaptic efficacy (LTP, LTD) induced by the change in neurotransmitter release. CONCLUSION An online sensor was developed to monitor the extracellular L-glutamate concentration change as a result of stimulated release from cultured nerve cells. A high sensitivity and a detection limit of 7.6 nM were obtained by the standard solution experiment. The effect of L-ascorbate can be reduced effectively by applying a low potential (0 V) for measurement and by preoxidation with a ring-disk dual electrode. The extracellular glutamate concentration change was observed either by stimulating a single nerve cell with KCl or by stimulating a brain slice electrically. An increase in L-glutamate concentration during synaptic transmission can also be measured by positioning the probe near the cell which is connected synaptically to the stimulated cell. This sensor system is potentially a powerful tool with which to study synaptic plasticity, memory mechanisms, and Alzheimer’s diseases.

Received for review November 28, 1995. Accepted March 8, 1996.X AC951154D X

Abstract published in Advance ACS Abstracts, May 1, 1996.