Monitoring d-Serine Dynamics in the Rat Brain Using Online

Pierre Pernot , Caroline Maucler , Yannick Tholance , Natalia Vasylieva , Gabriel Debilly , Loredano Pollegioni ... Maura Perry , Qiang Li , Robert T...
1 downloads 0 Views 212KB Size
Anal. Chem. 2004, 76, 6582-6587

Monitoring D-Serine Dynamics in the Rat Brain Using Online Microdialysis-Capillary Electrophoresis Chanda M. Ciriacks and Michael T. Bowser*

Department of Chemistry, University of Minnesota, 207 Pleasant Street SE, Minneapolis, Minnesota 55455

D-Serine

was detected in dialysate collected from the rat striatum using an online microdialysis-CE-LIF instrument. Dialysate can be analyzed every 12.5 s using the online instrument, giving much better temporal resolution than previously possible for D-serine. Basal concentrations of D-serine (8 ( 2 µM), glutamate (0.8 ( 0.2 µM), GABA (0.11 ( 0.04 µM), and L-serine (23 ( 4 µM) were measured. Increases in the concentrations of these neurochemicals induced by the introduction of high-K+ aCSF were quantitated. Notably, an increase in D-serine concentration in response to high-K+ aCSF was observed for the first time. The identity of the D-serine peak was confirmed unambiguously using D-amino acid oxidase to selectively remove D-serine from a dialysate sample. The microdialysis-CE-LIF instrument was able to monitor this enzymatic reaction as it proceeded over a period of 60 min, demonstrating that online microdialysis-CE-LIF is not only useful in monitoring in vivo dynamics but can also be used to monitor other chemical systems.

D-Serine was first identified in the mammalian brain in 1992.1 This was surprising considering that at the time it was thought that higher organisms used L-amino acids exclusively. Upon further examination, it was noted that the distribution of D-serine throughout the brain was heterogeneous. More specifically, D-serine histology matched that of the glutamate NMDA receptor.2 The NMDA receptor is a major class of receptor for glutamate, the major excitatory neurotransmitter in the central nervous system, and has been implicated in a range of functions such as memory, learning,3 and pain4 as well as dysfunctions including stroke,5,6 alchoholism,7 and epilepsy.8 The current hypothesis for

* Corresponding author. E-mail: [email protected]. (1) Hashimoto, A.; Nishikawa, T.; Hayashi, T.; Fujii, N.; Harada, K.; Oka, T.; Takahashi, K. FEBS Lett. 1992, 296, 33-36. (2) Schell, M. J.; Molliver, M. E.; Snyder, S. H. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 3948-3952. (3) Reidel: G.; Platt, B.; Micheau, J. Behav. Brain Res. 2003, 140, 1-47. (4) Petrenko, A. B.; Yamakura, T.; Baba, H.; Shimoji, K. Anesth. Analg. 2003, 97, 1108-1116. (5) Hoyte, L.; Barber, P. A.; Buchan, A. M.; Hill, M. D. Curr. Mol. Med. 2004, 4, 131-136. (6) Tranquillini, M. E.; Reggiani, A. Expert Opin. Invest. Drugs 1999, 8, 18371848. (7) Hoffman, P. L. Int. Rev. Neurobiol. 2003, 56, 35-82. (8) Manocha, A. Ind. J. Pharmacol. 1998, 30, 277-298.

6582 Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

D-serine’s

role in the central nervous system is that D-serine coactivates the NMDA receptor through the “glycine” binding site.9,10 Glycine was originally thought to be the agonist of this site, but it has since been found that D-serine binds the glycine site more strongly than glycine itself.11 If true, this pathway represents a novel mechanism for intercellular communication in the brain. It would also offer a new pharmacological target for modifying NMDA receptor activity during stroke, epilepsy, or other periods of hyperactivity. In support of this hypothesis, enzymes capable of synthesizing and removing D-serine have been identified in the mammalian brain;9,10 it has been shown that application of D-serine potentiates synaptic action potentials and application of D-amino acid oxidase (D-AAO) decreases synaptic action potentials.12 The only remaining piece to complete the hypothesis is to demonstrate that D-serine is actually released into the synapse. Measurement of D-serine release is not trivial. D-Serine is not electroactive, ruling out electrochemical methods. There is no natural chromophore present limiting the effectiveness of spectroscopic approaches. The analysis is further complicated by the enantiomeric selectivity required to resolve D-serine from L-serine. Microdialysis has become the method of choice for monitoring chemical dynamics in vivo.13-15 Briefly, a probe incorporating a short region of hollow fiber dialysis tubing is implanted into the tissue of interest. The probe is continuously perfused with a salt solution that mimics the environment of the probe. Small molecules diffuse across the membrane and are carried in the perfusate out of the probe. Fractions can be collected and analyzed to monitor changes in analyte concentration over time. The rate that fractions can be collected is often determined by the mass sensitivity of the assay used to analyze the fractions. Typically 10-20 µL is required for analysis, giving temporal resolutions on

(9) Baranano, D. E.; Ferris, C., D.; Snyder, S., H. Trends. Neurosci. 2001, 24, 99-106. (10) Snyder, S., H.; Ferris, C., D. Am. J. Psychiatry 2000, 157, 1738-1751. (11) Matsui, T.-a.; Sekigushi, M.; Hashimoto, A.; Tomita, U.; Nishikawa, T.; Wada, K. J. Neurochem. 1995, 65, 454-458. (12) Mothet, J.-P.; Parent, A., T.; Wolosker, H.; Brady, R., O., Jr.; Linden, D., J.; Ferris, C., D.; Rogawski, M., A.; Snyder, S. H. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 4926-4931. (13) Robinson, T. E.; Justice, J. B. Microdialysis in the Neurosciences; Elsevier: Amsterdam, 1991. (14) Adell, A.; Artigas, F. In In Vivo Neuromethods; Boulton, A. A., Baker, G. B., Bateson, A. N., Eds.; Humana Press: Totowa, NJ, 1998. (15) Davies, M. I. Anal. Chim. Acta 1999, 379, 227-249. 10.1021/ac0490651 CCC: $27.50

© 2004 American Chemical Society Published on Web 10/16/2004

the order of 5-20 min since the perfusion rate through the probe is limited to 1-2 µL/min. The development of online microdialysis-CE has overcome this limitation.16-18 The high mass sensitivity of laser-induced fluorescence (LIF) and electrochemical CE detectors makes the analysis of nanoliter-scale samples possible. The online nature of the instrument eliminates the troublesome task of collecting, storing, derivatizing, and analyzing submicroliter samples. Temporal resolutions as low as 3 s have been demonstrated for glutamate and aspartate.19 Recent improvements in sensitivity and efficiency have allowed greater numbers of compounds to be analyzed simultaneously.20 We have previously described an online microdialysis-capillary electrophoresis assay capable of measuring a number of primary amine neuromessengers, including glutamate, GABA, taurine, and D-serine, in both homogenized21 and continuously perfused intact retinas.22 Unfortunately, little D-serine was detected in homogenates and no release was detected from intact retinas. This online microdialysis assay is well suited for performing in vivo measurements in the mammalian brain, where D-serine concentrations are expected to be higher.23 In the current paper, we demonstrate the use of an enantioselective, high-temporal resolution microdialysis assay to monitor in vivo neuromessenger dynamics for the first time. EXPERIMENTAL SECTION Chemicals. Unless otherwise noted, chemicals were obtained from Mallinckrodt (Paris, KY). All buffers were made in deionized water (Milli-Q, 18.2 MΩ; Millipore, Bedford, MA) and filtered (0.2 µm). Artificial cerebral spinal fluid (aCSF) consisted of 145 mM NaCl, 2.7 mM KCl, 1.0 mM MgSO4, and 1.2 mM CaCl2. High-K+ aCSF contained 45 mM NaCl, 102.7 mM KCl, 1.0 mM MgSO4, and 1.2 mM CaCl2. Derivatization solution contained 43 mM o-phthaldialdehyde (Sigma-Aldrich, St. Louis, MO) dissolved in 9:11 methanol:75 mM borate (pH 10.5, Fisher Scientific, Fair Lawn, NJ). The internal standard (5 µM L-2-aminoadipic acid (Sigma-Aldrich)) was dissolved in 114 mM β-mercaptoethanol (Sigma-Aldrich) in aCSF. CE separation buffer consisted of 100 mM borate/20 mM hydroxypropyl-β-cyclodextrin (HP-β-CD, Cerestar USA Inc., Hammond, IN) adjusted to pH 10.5. A solution of 50 mM borate (pH 10.5) was used as the sheath-flow buffer. D-AAO (Sigma-Aldrich, Lot 88H72505) was prepared in cold aCSF (5.2 units in 1 mL). Catalase (Sigma-Aldrich, Lot 88H72505) was dissolved in cold deionized water (16 300 units/mL). D-AAO and catalase were prepared immediately before use and kept on ice. Potassium phosphate (monobasic, 765 mM, pH 8.3 (J.T. Baker, Phillipsburg, NJ)) was used as the enzymatic reaction buffer. Oxygen was bubbled through the potassium phosphate for 5 min prior to use. (16) Hogan, B. L.; Lunte, S. M.; Stobaugh, J. F.; Lunte, C. E. Anal. Chem. 1994, 66, 596-602. (17) Zhou, S. Y.; Zuo, H.; Stobaugh, J. F.; Lunte, C. E.; Lunte, S. M. Anal. Chem. 1995, 67, 594-599. (18) Lada, M. W.; Schaller, G.; Carriger, M. H.; Vickroy, T. W.; Kennedy, R. T. Anal. Chim. Acta 1995, 307, 217-225. (19) Lada, M. W.; Vickroy, T. W.; Kennedy, R. T. Anal. Chem. 1997, 69. (20) Bowser, M. T.; Kennedy, R. T. Electrophoresis 2001, 22, 3668-3676. (21) O’Brien, K. B.; Esquerra, M.; Klug, C. T.; Miller, R. F.; Bowser, M. T. Electrophoresis 2003, 24, 1227-1235. (22) O’Brien, K. B.; Esquerra, M.; Miller, R. F.; Bowser, M. T. Anal. Chem. 2004, 76, 5069-5074. (23) Hashimoto, A.; Oka, T.; Nishikawa, T. Neuroscience 1995, 66, 635-643.

Figure 1. Schematic representations of the online microdialysisCE-LIF instrument (A), the reaction cross, flow gate, and sheath-flow detector (B), and the microdialysis probe (C).

Microdialysis-CE-LIF. A schematic of the online microdialysis-CE-LIF instrument is shown in Figure 1A. The online microdialysis instrument has been described previously.21,22 Briefly, side-by-side probes containing a 3-mm-long, 200-µmouter diameter regenerated cellulose (18 000 MWCO, Spectrum Laboratories, Inc., Rancho Dominguez, CA) sampling region were fabricated in-house (see Figure 1C). The microdialysis probe was perfused with aCSF at 20 µL/h (0.33 µL/min). Dialysate was mixed online with the derivatization solution (10 µL/h) and internal standard solution (10 µL/h) in a 0.25-mm-i.d. reaction cross (Valco Instruments Co. Inc., Houston, TX). Derivatization occurred as the mixed solution traveled through the reaction capillary (150-µm inner diameter, 8.5-9.5 cm long, 2.3-2.5-min reaction time) to the flow gate interface (see Figure 1B). The reaction capillary and the CE separation capillary were held 50 µm apart in the flow gate. Separation buffer flowed through this gap at 40 mL/h. To perform an injection the cross-flow was stopped for 1-2 s. An injection voltage of -3 kV was applied at the outlet of the separation capillary for 0.5-1 s. After the injection, the cross-flow was resumed and the separation voltage was raised to -20-24 kV over 500 ms. CE separation took place in a 7.6-7.8-cm-long, 5-µm-i.d., 150-µm-o.d. fused-silica capillary. LIF detection was performed in a high-sensitivity sheath-flow cell. Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

6583

The 351-nm line of an argon ion laser (20-mW multiline UV, Enterprise II, Coherent Laser Group, Santa Clara, CA) was used for excitation. Fluorescence was collected at 90° through a 60×, 0.7 NA long working distance objective (Universe Kogaku, Oyster Bay, NY), passed through spatial and band-pass filters and collected on a photomultiplier tube (PMT R1477, Hamamatsu Corp., Bridgewater, NJ). Signal from the PMT was amplified, filtered, and recorded using a data acquisition card (National Instruments Corp., Austin, TX). In Vitro Characterization. The in vitro recoveries of dialysis probes were measured by comparing the peak areas for analytes collected through the probe with peak areas observed when the same analyte solution was pumped directly into the reaction cross. Temporal resolution of the probe was measured by moving the probe from aCSF to a solution of 8 µM D-serine in aCSF for 3 min and then back again. The temporal resolution of the valve used for reverse microdialysis stimulations was measured by switching the probe perfusate from aCSF to aCSF spiked with 8 µM D-serine for 3 min and then back to aCSF. The probe was placed in 2.6 µM D-serine for the entire time. In Vivo Monitoring. All animal experiments were performed in strict accordance with protocols approved by the Institutional Animal Care and Use Committee at the University of Minnesota. Male Sprague-Dawley rats (250-350 g) were anesthetized with halothane U.S.P. (2% in 1 L/min industrial grade oxygen, Halocarbon Laboratories, River Edge, NJ) in an induction chamber (Harvard Apparatus, Hollison, MA). The animal was then transferred to a stereotaxic instrument mounted with an anesthesia mask (David Kopf Instruments, Tujunga, CA). After the animal no longer exhibited limb reflex, ear bars were secured in place and halothane was reduced to 1.375-1.5% in industrial grade oxygen to maintain surgical anesthesia. Surgery was performed to expose the brain region of interest. The stereotaxic instrument was used to implant a microdialysis probe in the left striatum (+0.2 mm AP, +3.0 mm ML, -6.5 mm DL from bregma24). The probe was implanted slowly, over a period of 15 min to minimize tissue damage. Once in place, the probe was allowed to equilibrate for 45 min before measurements were made. Potassium stimulations were performed immediately following the 45-min equilibration time. To perform a potassium stimulation, a four-port, manual turn valve was used to switch the microdialysis perfusion solution from aCSF to high-K+ aCSF for 3 min. A second potassium stimulation was performed 30 min after the end of the first potassium stimulation (i.e., 30 min after the valve was switched back to aCSF). D-Amino Acid Oxidase Digestion. D-AAO was used to confirm the identity of the D-serine peak in dialysate. A 20-µL aliquot of dialysate was collected in a 500-µL Eppendorf tube from the outlet of a microdialysis probe implanted in the rat striatum. A 20-µL aliquot of potassium phosphate (final concentration 153 mM), 58 µL of D-AAO (final activity 0.3016 unit), and 2 µL of catalase (final activity 32.6 units) were added to the 20 µL of dialysate to give a total reaction volume of 100 µL. A fresh microdialysis probe was placed directly into the reaction mix. Analyte concentrations in the reaction mix were measured (24) Paxinos, G.; Watson, C. The rat brain in stereotaxic coordinates, 4 ed.; Academic Press: San Diego, 1998.

6584 Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

Figure 2. Electropherograms from the online CE-LIF analysis of dialysate collected from the striatum of an anesthetized rat. (A) shows normal basal concentrations while (B) was observed during perfusion with high-K+ aCSF. Peak identifications: (1) aminoadipic acid (internal standard), (2) glutamate, (3) aspartate, (4) GABA, (5) taurine, (6) D-serine, and (7) L-serine. Capillary, 5 µm i.d. × 7.8 cm long; separation buffer, 100 mM borate/20 mM HP-β-CD, pH 10.5; voltage, -20kV; LIF detection, λex ) 351 nm, λem ) 450 nm.

every 12.5 s over ∼1 h using the online microdialysis-CE-LIF instrument. RESULTS AND DISCUSSION Figure 2A shows an electropherogram of dialysate that was sampled from the striatum of an anesthetized rat and analyzed online. The high electric field and short separation distances give rise to high separation efficiencies (∼400 000 plates) and fast analysis times (∼9.5 s). Overlapping injections22 allow separations to be performed every 12.5-14 s, including time for the injection sequence. A number of important neuromessengers can be identified including glutamate, GABA, taurine, and aspartate. Dand L-serine are also clearly resolved. Table 1 shows the basal concentrations measured for each of these neurochemicals. In all cases, the basal levels are significantly higher than the limits of detection. The values are similar to those reported elsewhere using conventional microdialysis techniques.23 One major strength of the online microdialysis-CE approach is the ability to monitor chemical dynamics over time. Figure 3 illustrates the temporal response of the instrument. Figure 3A shows the instrument response to the introduction of a step change in D-serine concentration at the surface of the microdialysis probe. The probe was moved from a solution of aCSF

Table 1. Basal Concentration, Limit of Detection, and in Vitro Relative Recovery of Glutamate, GABA, and D- and L-Serinea

glutamate GABA D-serine L-serine

basal concn (µM)

limit of detection (µM)

% recovery

0.8 ( 0.2 0.11 ( 0.04 8(2 23 ( 4

0.14 ( 0.03 0.05 ( 0.01 0.06 ( 0.01 0.06 ( 0.01

68 ( 8b 63 ( 8 60 ( 4 67 ( 7

a Confidence intervals are the standard error based on measurements made in the striatum of four Sprague-Dawley rats. b n ) 3.

Figure 3. (A) Temporal response of the microdialysis probe. The microdialysis probe was moved from aCSF to aCSF spiked with 8 µM D-serine for 3 min (denoted by the black bar) and back again. (B) Time profile of a 3-min stimulus introduced by the reverse microdialysis valve. Microdialysis perfusate was switched from aCSF to aCSF spiked with 8 µM D-serine for 3 min (denoted by the black bar) and then back to aCSF. The probe was placed in a 2.6 µM D-serine solution throughout this experiment. Electropherograms were recorded every 13.1 s in (A) and every 12.7 s in (B).

to a D-serine standard solution for 3 min and back again. The signal reached a plateau in approximately two to three separations. Separations were performed every 13.1 s. Therefore, dynamics at the surface of the probe as fast as 26 s can be measured. Figure 3B shows the time profile of the valve used to introduce agents into the brain using reverse microdialysis. This is important when performing pharmacological experiments since it is usually the time course of the drug that determines the shape of the response more than the temporal resolution of the instrument. We have used a zero dead volume four-port

Table 2. Maximum Percent Change Observed during High-K+ aCSF Stimulationa max % increase

glutamate GABA D-serine L-serine

first stimulation

second stimulation

min detectable % change

600 ( 100 8000 ( 2000 24 ( 5 22 ( 4

700 ( 100 8000 ( 2000 27 ( 4 28 ( 5

40 ( 10 70 ( 40 7(5 7(2

a Confidence intervals are the standard error based on measurements made in the striatum of four Sprague-Dawley rats.

valve to minimize broadening. The time profile of this valve is on a similar order of magnitude as the temporal response of the microdialysis-CE instrument. This is important when attempting to detect the immediate effects of a pharmacological agent or make an accurate measurement of the recovery time after the stimulus. Figures 2B and 4 illustrate a simple pharmacological experiment. Reverse microdialysis is used to expose the rat striatum to high-K+ aCSF for 3 min (denoted by the solid line). High-K+ concentration depolarizes neurons, promoting exocytotic release of neuromessengers. Careful examination of Figure 2 shows relatively large increases for glutamate, GABA, taurine, and aspartate. Other peaks increase in intensity, but to a much lesser extent. Figure 4 shows the time course of glutamate, GABA, D-serine, and L-serine concentrations during the stimulation. Large increases for glutamate and especially GABA are observed. Smaller increases are observed for D- and L-serine. To our knowledge, this is the first time that D-serine release has been observed in response to K+ stimulation. This may be due in part to the relatively small change in D-serine concentration induced by the high-K+ aCSF. Small changes are difficult to detect using conventional microdialysis because the low number of data points limits the statistical significance. The high sampling rate of the online microdialysis-CE-LIF instrument allows multiple points to be recorded both before and during the stimulation, making it easier to reliably identify small increases. Table 2 lists the minimum detectable change that could be observed for glutamate, GABA, and D- and L-serine. This value is analogous to the limit of detection where a change of three standard deviations from baseline is considered detectable. The open circles in Figure 4 show a control experiment where a probe was placed in a standard amino acid solution and the perfusate was changed from aCSF to high-K+ aCSF and back as in the in vivo experiment. No deviation from the baseline was observed, demonstrating that high-K+ aCSF does not change the recovery of the analytes and switching the four-port valve to introduce the stimulus does not generate artifacts in the signal. Table 2 also lists the maximum percent change observed for glutamate, GABA, D-serine, and L-serine during each K+ stimulation. These data are broken down between the first application of high-K+ aCSF and a second dose applied 30 min later. There is no statisitical difference between the first and second stimulations for any of the neurochemicals listed, suggesting that the recovery time between experiments was sufficient. Decreased release due to exhaustion of neurotransmitter or Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

6585

Figure 4. Relative change in concentration during a 3-min stimulation with high-K+ aCSF (solid markers). The black bars denote the time that the probe is perfused with high-K+ aCSF. Note that this has been corrected to account for the time required for the high-K+ aCSF to travel from the valve, through the probe, and to the CE interface. Traces with open markers are a control experiment where the same experiment was performed with the probe placed in a standard amino acid solution of 1 µM glutamate, 2 µM GABA, 8 µM D-serine, and 8 µM L-serine in aCSF. Electropherograms were recorded every 12.7 s for the control experiment and 12.5 s for the in vivo potassium stimulation experiment. The control traces in (C) and (D) have been offset for clarity.

energy stores is often observed if the recovery time is not long enough. It should be noted that the observed increases and decreases occur on the same time scale as the temporal resolution of the instrument, suggesting that these concentration changes occur very quickly. Large increases in glutamate and GABA concentration were expected during K+ stimulation since these are well-known neurotransmitters that are localized in synaptic vesicles. The observed increase in D-serine concentration was somewhat surprising considering that D-serine is not located in neurons. Instead, it has been shown that D-serine is localized in glia,2 which are not excitable by K+. A possible explanation is that significant release of glutamate during high-K+ perfusion activates a glial receptor, which in turn induces D-serine release. This result should not be overstated since a similar increase was observed for L-serine, which is not typically considered a neuromessenger. Unambiguously identifying D-serine was a serious concern, especially considering the peak density of the electropherogram between 7 and 9 s. Standards were used to locate where D-serine should migrate. However, it was possible that another compound could comigrate with D-serine at this same time. To address this, we digested a dialysate sample with D-AAO. D-AAO selectively oxidizes D-amino acids. To date D-serine and D-aspartate are the only D-amino acids that have been identified in the mammalian 6586 Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

brain at significant concentrations.25 D-Aspartate was not detected in the current experiments, which is not unexpected since D-aspartate concentrations in the mammalian brain decrease drastically as the animal matures. D-AAO therefore provides a way of selectively removing D-serine from a dialysate sample. To confirm the identity of the D-serine peak, we collected 20 µL of dialysate from a probe implanted in the rat striatum. D-AAO and catalase were added to the dialysate, and the reaction was monitored using the online microdialysis-CE-LIF instrument by placing a new probe into the reaction mixture. There are several advantages to this “double-dialysis” approach. No fluid is removed from the reaction mixture, allowing frequent analyses to be made on a small-volume solution. The total volume of the reaction mixture used here was only 100 µL. Microdialysis sampling also excludes proteins from the dialysate, minimizing the effect of D-AAO and catalase on the CE separation. Figure 5A shows an electropherogram recorded immediately after D-AAO was added to the dialysate. A large D-serine peak is observed at ∼9.1 s. Figure 5B shows an electropherogram of the same solution after 56 min of D-AAO digestion. The peak identified as D-serine is completely eliminated, confirming our hypothesis that this was indeed D-serine and that there are no comigrating compounds present in the dialysate that would interfere with the (25) Hashimoto, A.; Oka, T. Prog. Neurobiol. 1997, 52, 325-353.

depletion of the 100-µL reaction solution by the microdialysis probe. Surprisingly, a new peak was generated at ∼8 s. At this time, we are not certain of the identity of this peak. No new primary amines are expected from the direct reaction of D-serine with D-AAO. It should be remembered that brain dialysate is a very complex matrix and other side reactions may be possible. Figure 5C shows the normalized peak areas for D-serine and L-serine plotted over the 56-min reaction. Electropherograms were taken every 12.5 s giving rise to 270 measurements. The curve for D-serine follows the first-order decay expected for Michaelis-Menten kinetics when the concentration of the substrate is much lower than the dissociation constant (KM). The KM for D-serine with D-AAO has been reported to be 41 mM,26 clearly much higher than the substrate concentration used in the current experiments. An exponential regression of the curve gives an apparent first-order reaction constant (Vmax/KM) of 0.088 ( 0.002 min.-1 with an R2 of 0.9895. The decrease observed for L-serine is much smaller. Again, this confirms the identity of the D-serine peak and rules out the presence of any comigrating interferences.

Figure 5. D-AAO digestion of dialysate collected from the rat striatum. (A) shows an electropherogram immediately after D-AAO was added to the dialysate. Peak identifications: (1) D-serine and (2) L-serine. The peak labeled with an asterisk (/) is an unknown side product of the reaction. (B) shows an electropherogram of the same reaction mixture after 56 min of digestion. (C) is a plot of the normalized peak areas for D-serine (O) and L-serine (b) over the time course of the reaction. Electropherograms were recorded every 12.5 s.

D-serine analysis. Other peaks in the electropherogram decreased only slightly over the 56-min reaction time. This may be due to

(26) Vanoni, M. A.; Cosma, A.; Mazzeo, D.; Mattevi, A.; Todone, F.; Curti, B. Biochemistry 1997, 36, 5624-5632.

CONCLUDING REMARKS This paper has demonstrated a high temporal resolution microdialysis assay for D-serine. D-Serine dynamics have been measured on the time scale of seconds in the rat brain for the first time. This assay will be an important tool in determining what role D-serine plays in the mammalian central nervous system. Pharmacological experiments similar to the simple K+ stimulation shown here can now be performed to identify which receptors and transporters are involved in determining D-serine concentrations. The high data collection rate of the online microdialysis-CE-LIF instrument allows changes in D-serine concentration as small as 7% to be detected, making it possible to observe subtle changes that may be missed using conventional techniques. The D-AAO experiment also showcases the potential of the online microdialysis-CE-LIF instrument for other chemical monitoring applications. To our knowledge, this is the first example of high temporal resolution monitoring of an enzymatic reaction using an online CE instrument. Microdialysis is well suited for this application. Since no solution is removed, very small volumes can be monitored frequently over long periods of time. This is important considering the cost and limited availability of many enzymes. Microdialysis also excludes large molecules, minimizing the effect of the enzyme on the CE assay, making the technique widely applicable. ACKNOWLEDGMENT Funding for this research was provided by the National Institutes of Health (NS 043304) and the University of Minnesota.

Received for review June 25, 2004. Accepted August 15, 2004. AC0490651

Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

6587