Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

L-Glutamate acts as the principal excitatory neurotransmitter in the mammalian central nervous system via ionotropic and metabotropic glutamate receptors. Ionotropic glutamate receptors (iGluRs) are classified into the NMDA and non-NMDA subtypes, and non-NMDA receptors are further divided into AMPA and KA subtypes. Molecular cloning studies demonstrated that iGluRs are encoded by at least six NMDA (NR1, NR2A-2D, and NR3A), four AMPA (GluR1-4), and five KA (GluR5-7, KA1, and KA2) receptor genes (Nakanishi, 1992; Seeburg, 1993; Hollmann and Heinemann, 1994). Metabotropic receptors (mGluRs) are encoded by eight distinct genes classified into three groups: group I receptors (mGluR1, 5) are coupled to phosphoinositide (PI) hydrolysis and calcium mobilization signal transduction pathways, whereas group II (mGluR2, 3) and group III (mGluR4, 6-8) receptors are negatively linked to cAMP formation [reviewed by Conn and Pin (1997)]. Ionotropic and metabotropic GluRs are widely distributed in the vertebrate brain and play integral roles in excitatory neurotransmission. In addition, these receptors are thought to be involved in higher brain mechanisms such as memory formation, learning, pain transmission, and several neuronal disorders [reviewed by Conn and Pin (1997) and Bleakman and Lodge (1998)].

Understanding the complex roles that iGluRs play in physiological and pathological processes in the brain has been facilitated by the isolation of selective pharmacological compounds. Differentiation of NMDA and non-NMDA (AMPA and KA) receptors has been possible for a number of years because of the divergent pharmacological profiles of these distinct receptors families. However, pharmacological separation of neuronal currents mediated by AMPA and KA receptors has been more difficult. Recent development of AMPA and KA receptor-selective agonists and antagonists have aided in the detection and characterization of neuronal KA receptors, which appear to play roles in synaptic transmission distinct from those of AMPA and NMDA receptors [reviewed by Chittajallu et al. (1999)].

Several important classes of GluR ligands have been isolated from marine organisms (Laycock et al., 1989). The non-NMDA receptor agonists KA and domoic acid (DOM) have been of particular interest because of their potent epileptogenic properties. KA-induced seizures have long been used as a model for human temporal lobe epilepsy (Ben-Ari, 1985). DOM, a causative agent of amnesiac shellfish poisoning (Perl et al., 1990), is the most potent excitatory amino acid (EAA) yet characterized (Stewart et al., 1990). The seizurogenic activity of these compounds has suggested that EAAs may be useful not only as pharmacological tools but also as lead compounds for therapeutic agents of neurological disorders (Krogsgaard-Larsen and Hansen, 1992; Bleakman and Lodge, 1998). We therefore searched for EAAs in marine organisms and isolated a novel di-amino, di-acid, dysiherbaine (DH, Fig. 1), from the sponge Dysidea herbacea (Sakai et al., 1997).

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Structures of dysiherbaine and other KA receptor agonists.

In the present study, we have explored the pharmacological specificity of DH in radioligand binding and electrophysiological experiments. We have also carried out a more detailed characterization of the behavioral responses to DH injection in mice. These studies demonstrate that DH is an agonist for both ionotropic and metabotropic GluRs and has strikingly potent seizurogenic activity.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Convulsant Action in Mice

Male ddY mice (16-25 g, 4-6 weeks old; Owada experimental animals, Iwate, Japan) were kept in cages with free access to their standard diet until use. Five or more animals were tested for each group. All efforts were made to minimize both suffering and the number of animals used. Animals received i.p. or i.c.v. injections of drug solution or vehicle. i.c.v. administrations were performed according to a method described previously (Laursen and Belknap, 1986) with some modifications. The sample solution (20 µl) was injected slowly over 5 s at 2 mm lateral to the midline of the skull, 3 mm rostral to a line down through the anterior base of the ears, at a depth of 3.5 mm. Normal saline was used as a control vehicle and did not induce any notable change in animal behaviors.

The behavior of animals was observed following drug administration for 3 or 6 h continuously. Severity of seizures was classified based on a behavioral scale developed to assess toxicity of DOM (Tasker et al., 1991), with some modifications into seven grades as follows: grade 0, normal; grade 1, hypoactivity (20-40 s of immobility); grade 2, occasional scratching and/or 40 to 60 s of immobility without rigidity; grade 3, frequent scratching, vigilant staring, hyperactivity, and/or head bobbing; grade 4, loss of balance, and/or sudden running or jumping accompanied by frequent scratching; grade 5, forepaw or neck clonus and/or rearing with complete loss of balance; grade 6, rigid and splayed hind limbs with forepaw clonus and neck clonus plus rearing; grade 7, tonic convulsion followed by death. The highest score observed in each 10-min interval was averaged for 0 to 1, 2 to 3, and 5 to 6 h, respectively. The score from 1-h observation was probit-transformed by setting a score of 7.0 to be 100%. Two distinct ED₅₀ values (doses that cause 50% of maximum seizure score) were calculated by linear regression of probit-transformed data in this study. The ED₅₀(max) value was calculated using the highest seizure score observed within the initial 1-h observation period after injection. The ED₅₀(mean) values of DH and DOM were calculated using averaged (as opposed to the highest) seizure scores obtained in six observation periods (10 min each, total 1 h). Linear regression of the data was performed with StatView software (SAS Institute Inc., Cary, NC).

Receptor Binding Studies

Synaptic membranes were prepared as previously described (Murphy et al., 1987). Briefly, male Sprague-Dawley rats (200-220 g) were decapitated and forebrains were homogenized with a Teflon-glass homogenizer in 10-fold volumes of 0.32 M sucrose solution. The homogenate was centrifuged at 1000g for 10 min. After centrifugation of the supernatant (17,000g for 20 min), the resulting pellet was suspended in distilled water. The suspension was centrifuged at 8000g for 20 min. The supernatants and soft, buffy uppercoat layers of the pellets were collected and centrifuged (48,000g for 20 min). After washing with a suitable buffer for each assay, the membrane pellets were stocked at 78°C until use. On the day of the assay, the frozen pellet was thawed at room temperature and incubated with 0.04% Triton X-100 in the assay buffer at 37°C for 30 min followed by centrifugation (48,000g for 10 min). The detergent was removed by washing twice with the assay buffer. The conditions for each binding assay were as follows (ligand, ligand concentration, incubation temperature, incubation time, buffers): [³H]KA, 1 nM, 4°C, 1 h, 100 mM Tris-HCl (pH 7.1); [³H]AMPA, 5 nM, 4°C, 1 h, 50 mM Tris-HCl (pH 7.4) and 100 mM KSCN; [³H]CGS-19755, 10 nM, 4°C, 1 h, 50 mM Tris-HCl (pH 8.0) (London and Coyle, 1979; Murphy et al., 1987, 1988).

Binding studies with recombinant GluR5 and GluR6 KA receptor subunits in HEK293 cells were carried out as described previously (Swanson et al., 1997). [³H]KA displacement assays were performed as described previously (Swanson et al., 1997). The [³H]KA concentrations used in this study were 20 to 27 and 13 to 20 nM for GluR5 and GluR6 assays, respectively. Nonspecific binding was defined in the presence of 1 mM glutamate. The GluR5 and GluR6 cDNAs were generously donated by Dr. Peter Seeburg (Max-Planck-Institute, Heidelberg, Germany) and Dr. Stephen Heinemann (The Salk Institute, La Jolla, CA).

Electrophysiological Studies

Culture Preparation. Hippocampal neurons were prepared from newborn rat pups and grown in microdot cultures as described previously (Bekkers and Stevens, 1991). Cultures were maintained in Dulbecco's modified minimal essential medium supplemented with 10% heat-inactivated horse serum, 20 mM glucose, 1% N2 supplement, and penicillin/streptomycin. After 4 to 5 days in vitro, non-neuronal cell division was inhibited by exposure to 35 µM fluorodeoxyuridine and 75 µM uridine for 1 to 3 days. Recordings were obtained from cells that had been cultured for 2 to 3 weeks.

Electrophysiology. Whole-cell voltage-clamp recordings were made from neurons using standard patch-clamp techniques. Glass fragments of coverslips with adherent cells were placed in a perfusion chamber and rinsed with a buffer of composition (in mM): NaCl, 150; KCl, 2.8; CaCl₂, 1.8; MgCl₂, 1; HEPES, 10; tetrodotoxin, 0.001 (pH was adjusted to 7.3 with NaOH). Experiments were performed at room temperature (22-25°C) and recorded on an Axopatch 200B amplifier using pClamp8 software (Axon Instruments, Foster City, CA). Patch pipettes had initial resistances of 4 to 5 M when filled with an internal solution composed of (in mM): CsF, 95; CsCl, 25; Cs-HEPES, 10; Cs-EGTA, 10; NaCl, 2; Mg-ATP, 2; QX-314, 10; TEA-Cl, 5; 4-aminopyridine, 5 (pH adjusted to 7.3 with CsOH). Drugs were applied through three-barrel glass tubing attached to a piezo bimorph controlled by pClamp8 software, which allowed rapid exchange of solutions to the neurons. To test for activation of GABA_A receptors, DH (50 µM) was applied to cultured hippocampal interneurons in the presence of GYKI 53655 (100 µM), CNQX (50 µM), DL-2-amino-5-phosphonovaleric acid (50 µM), and CPCCOEt (100 µM) to block AMPA, kainate, NMDA, mGlu receptors, respectively. GYKI 53655 (100 µM) was added to perfusion solutions to isolate KA receptor currents. Dose-response curves were fitted to the Hill equation using Origin software (MicroCal Software, Inc., Northampton, MA).

Ca²⁺ Imaging of Recombinant Group 1 Metabotropic Glutamate Receptors (mGluR1 and mGluR5)

Standard calcium phosphate precipitation techniques were used to transiently transfect HEK293 cells with plasmids containing rat mGluR1 or mGluR5 cDNAs. Two to three days after transfection, cells were incubated in 3 µM Fura-2 AM (Molecular Probes, Eugene, OR) for 20 to 40 min at room temperature in the presence of 0.02% pluronic F-127. The cells were then transferred to a recording chamber and continuously perfused with HEPES-buffered saline containing (in mM): NaCl 135; KCl 5; CaCl₂ 2; sucrose 20; glucose 10; HEPES 5 (pH 7.4). Ratiometric Ca²⁺ imaging experiments were performed on an inverted microscope (Axiovert 100 TV; Zeiss, Thornwood, NY) equipped for epifluorescence microscopy. Samples were excited with a xenon lamp and the excitation wavelengths (350 or 380 nm) were selected using a polychromatic illumination system (TILL Photonics, Planegg, Germany). Images were acquired using a cooled charge-coupled device camera (MicroMAX, Roper Scientific, Trenton, NJ) controlled by Imaging Workbench software from Axon Instruments. 100 µM (S)-DHPG was used to elicit calcium signals in mGluR-expressing cells.

Uptake Assay on Glutamate Transporters

Madin-Darby canine kidney (MDCK) cells stably expressing subtypes of human excitatory amino acid transporters (EAAT1, EAAT2, or EAAT3) were seeded onto 96-well plates and cultured in Dulbecco's modified minimal essential medium containing 10% dialyzed fetal bovine serum and antibiotics (G418) for 2 days before the uptake assay. The uptake assay was performed as previously described (Shimamoto et al., 1998). The relative specific uptake of [¹⁴C]glutamate was determined from three different experiments.

Drugs

Dysiherbaine was isolated and purified according to a previously described procedure (Sakai et al., 1997). GYKI 53655 was custom synthesized by Research Biochemicals International (Natick, MA). All other chemicals were commercially available at highest purity from Nakarai Tesque (Kyoto, Japan), Sigma Chemical Co. (St. Louis, MO), Tocris Cookson (Bristol, UK), Research Biochemicals International, and Diagnostic Chemicals Ltd. (West Royalty, Canada). Radiolabeled compounds were purchased from PerkinElmer Life Sciences (Boston, MA).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

DH Is a Potent Convulsant. We first compared the behavioral responses to i.c.v. injection of DH and other representative excitatory amino acids. The rank order of potency for the tested EAAs was determined from the ED₅₀(max) values (see Materials and Methods), which were estimated as the doses that gave 50% of the maximum seizure score in the initial 1-h observation period after injection of the drugs. The entire 1-h period was used for comparative purposes, because some compounds induced only short-lasting behavioral changes. As reported previously, behavioral changes induced after injection of NMDA- and non-NMDA-type EAAs were clearly different in terms of duration of convulsant behavior (Chiamulera et al., 1992).

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Epileptogenic actions induced by DH and DOM in mice. Behavioral changes, induced by administration of DH and DOM i.c.v. (A) or i.p. (B), were graded using a seven-point scale (see Materials and Methods). Values are the mean scores ± S.E.M. Numbers in parentheses represent the observation period expressed by the time (h) after administration of drugs. More than five mice were used for each dose group. Note that behavioral changes induced by DH were observed for 3 h (i.c.v.) or 6 h (i.p.) after the administration.

DH Displaces Radiolabeled Ligands from Non-NMDA Receptors. The affinity of DH for ionotropic glutamate receptors was determined by radioligand binding techniques using rat brain synaptic membrane preparations. DH and other agonists were used to displace [³H]CGS-19755, [³H]DL-AMPA, and [³H]KA from NMDA and non-NMDA glutamate receptors (Table 2 and Fig. 3). Table 2 summarizes K_i values and Hill coefficients estimated from the displacement curve for each drug. DH inhibited [³H]KA binding and [³H]AMPA binding with K_i values of 26 ± 4.0 and 153 ± 10 nM, respectively. These values were higher than those for unlabeled KA and AMPA displacement of their tritiated counterparts (1.8 and 3.9 nM, respectively). The Hill coefficient determined from DH displacement of [³H]KA binding was much smaller (0.52) than those of other ligands (Table 2), suggesting that DH was displacing [³H]KA from a heterogeneous population of binding sites. Additionally, some data points for DH displacement of [³H]KA both in low (<0.2 nM) and high (>100 nM) concentration range fitted rather poorly to the Hill-type model. A multiphasic model for [³H]KA displacement fitted better to the curve; however, such an analysis was not rigorously made, because we could not define the points of the inflection unambiguously. Therefore, the apparent K_i value for DH displacement of [³H]KA binding likely represents an average value from several binding sites. DH did not inhibit [³H]CGS-19755 binding at 10 µM, suggesting that the compound does not interact with NMDA receptors.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Inhibition of [3H]KA (A) and [3H]AMPA (B) binding to rat brain membranes by DH. The results summarized in the figures represent at least three experiments for each compound; the concentrations of [3H]AMPA and [3H]KA were 5 nM and 1 nM throughout, respectively.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Displacement of [3H]kainate from GluR5 and GluR6 receptors by dysiherbaine. Representative displacement curves are shown for GluR5 (closed triangles) and GluR6 (open squares). For these examples, the estimated Ki values were 0.74 nM for GluR5 and 1.2 nM for GluR6.

DH Evokes Currents from Non-NMDA Receptors in Hippocampal Neurons. Because DH selectively displaced non-NMDA ligands from synaptic membranes, we next tested whether DH had agonist or antagonist activity on neuronal non-NMDA receptor currents. DH evoked currents from glutamate receptors in cultured hippocampal neurons. Measurable inward current responses were observed at 0.1 µM DH, and the responses were saturating at 100 to 300 µM (Fig. 5A). Steady-state currents evoked by 100 µM DH had a mean amplitude of 4.1 ± 0.6 nA (n = 7), which was approximately 2-fold larger than the amplitude of steady-state currents evoked by 1 mM glutamate in the same neurons (1.9 ± 0.7 nA, n = 7). Indeed, the predominantly nondesensitizing DH currents were comparable in amplitude to peak glutamate currents (4.0 ± 0.9 nA). The majority of this current was mediated by AMPA receptors, as opposed to KA receptors, because the currents were inhibited to a large degree by the AMPA receptor antagonist GYKI 53655 (see below). These data therefore demonstrate that the activity of DH on hippocampal AMPA receptors is similar to that of KA, although of somewhat higher potency.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   DH-evoked inward currents in cultured rat hippocampal neurons. Whole-cell AMPA/KA receptor currents evoked by increasing concentrations of DH in cultured rat hippocampal neurons. Drugs were applied for the periods indicated by bars above traces (A). KA receptor currents evoked by DH in the presence of GYKI 53655 (100 µM) (B), which were suppressed by coapplication of CNQX (100 µM) (C). Concentration-response curves for AMPA/KA or KA receptor steady-state currents evoked by DH in the absence or presence of GYKI 53655, respectively. KA receptor currents (+GYKI) were normalized to current at 1 µM DH, and AMPA/KA receptor currents (GYKI) were normalized to current amplitudes evoked by 100 µM DH. The smooth lines are Hill-type fits to the data and gave EC50 values of 9.68 and 0.21 µM for the AMPA/KA and KA receptor components, respectively (D).

DH Activates mGluR5 Receptors but Not mGluR1 Receptors. Because activation of mGluR1 metabotropic glutamate receptors can induce convulsant actions in vivo (Schoepp and Conn, 1993), we next examined the effect of DH on group I metabotropic glutamate receptors. We measured mGluR-mediated increases in intracellular calcium using the membrane-permeable Ca²⁺ indicator Fura-2 AM in ratiometric calcium imaging experiments. HEK293 cells expressing mGluR1 receptors gave a increased Ca²⁺ signal when 100 µM (S)-DHPG was applied, but no signal was observed upon application of 100 µM DH (Fig. 6A). In contrast, cells expressing mGluR5 receptors gave a robust signal with (S)-DHPG and also showed elevated Ca²⁺ levels when DH was applied (Fig. 6B). The calcium signal elicited by DH in mGluR5-expressing cells was comparable with that evoked by (S)-DHPG. These data demonstrate that DH is a subtype-selective group I mGluR agonist.

View larger version (13K): [in this window] [in a new window]   Fig. 6.   DH activates mGluR5 receptors but not mGluR1 receptors. Ratiometric imaging of intracellular calcium using the indicator Fura-2 AM in HEK 293 cells transfected with either mGluR1 (n = 10, A) or mGluR5 (n = 16, B) metabotropic receptors upon application of DHPG (100 µM, closed squares) or DH (100 µM, open triangles). Drugs were applied for 1.5 min, and images were acquired at 10-s intervals.

DH Does Not Interact with Glutamate Transporters. Glutamate transporters play an important role in maintaining the extracellular concentration of glutamate below neurotoxic levels. To examine the activity of DH for transporters, inhibition of [¹⁴C]Glu uptake in MDCK cells permanently expressing EAAT1, EAAT2, or EAAT3 was measured. DH (1 mM) did not inhibit glutamate uptake, whereas DOM showed weak inhibition (IC₅₀ = 500 ± 150 µM) for EAAT2, a KA-sensitive subtype. The relative specific uptakes of [¹⁴C]Glu to control were 101 ± 3% (EAAT1), 101 ± 4% (EAAT2), and 106 ± 9% (EAAT3) in the presence of DH (1 mM).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In the present study, we demonstrate that dysiherbaine, a novel amino acid isolated from Micronesian sponge D. herbacea, is a non-NMDA receptor agonist possessing the most potent convulsant activity among the known EAAs (Stewart et al., 1990; Chiamulera et al., 1992). Injection of higher concentrations of DH produced status epilepticus, which was previously characterized as a typical reaction to non-NMDA glutamate receptor agonists (Chiamulera et al., 1992), and in preliminary ligand binding studies DH displaced non-NMDA receptor ligands (Sakai et al., 1997); these observations suggested to us that DH could act as an agonist at non-NMDA receptors, which may in part underlie its epileptogenic action. This hypothesis was supported by subsequent radioligand binding and physiology experiments.

In radioligand binding studies using rat synaptic membranes, DH selectively interacted with non-NMDA receptors. Interestingly, despite the strong convulsant activity, DH displaced [³H]KA and [³H]AMPA with only moderate affinities (K_i values = 26 and 153 nM, respectively, Table 2) compared with the unlabeled counterparts of these radioligands. However, the Hill coefficient determined from DH displacement of [³H]KA binding (0.52) was particularly low compared with those of other ligands (or compared with DH displacement of [³H]AMPA). These results suggest that DH has discrete affinities for several KA-binding sites in rat brain, possibly comprised of both AMPA and KA receptor populations. If this is indeed the case, it is possible that the higher affinity DH binding sites correspond to those KA receptors involved in generation of seizure behavior (Mulle et al., 1998). This interpretation was supported by subsequent studies using recombinant kainate receptor subunits expressed in HEK293 cells. A high affinity of DH for KA receptors was clearly demonstrated by displacement of [³H]KA from recombinant KA receptors GluR5 and GluR6 (K_i = 0.48 and 1.28 nM for GluR5 and -6, respectively). The affinity of DH to GluR6 is 8 times higher than that of KA-selective ligand 4-methylglutamate (Zhou et al., 1997).

We next verified that DH is an agonist for non-NMDA ionotropic glutamate receptors with KA selectivity in electrophysiological studies using cultured hippocampal neurons. A large proportion of the DH-evoked current at higher concentrations was mediated by AMPA receptors, because GYKI 53655, a selective noncompetitive antagonist of AMPA receptors, substantially reduced mean current amplitudes. DH elicited currents in a concentration-dependent manner with an EC₅₀ of 9.68 µM and a mean current amplitude of about 4 nA at a concentration of 100 µM DH. We also observed GYKI-resistant current when DH was applied to the neurons in the presence of the antagonist. The GYKI-resistant current was of relatively high affinity, with an EC₅₀ of 0.21 µM. DH appears to be approximately 50-fold more potent for KA receptors as compared with AMPA receptors in hippocampal neurons, a degree of selectivity similar to that described for SYM 2081 (Jones et al., 1997; Zhou et al., 1997). The steady-state KA receptor current saturated at approximately 3 µM and had maximum amplitude of about 300 pA. High concentrations of the non-NMDA receptor antagonist CNQX suppressed GYKI-resistant DH currents; the lower sensitivity of KA receptors to blockade by CNQX compared with AMPA receptors is consistent with a previous report (Paternain et al., 1996). Finally, our observation that KA receptors are present at much lower densities compared with AMPA receptors in cultured neurons is also consistent with previous data (Paternain et al., 1996; Wilding and Huettner, 1997).

DH also stimulated calcium mobilization by heterologously expressed mGluR5 receptors, which belong to the group I family of receptors coupled to phospholipase C activity and phosphoinositide hydrolysis. Surprisingly, we found that this agonist activity was subtype-selective, because DH showed no agonist activity on mGluR1 receptors. This selective activity is relatively rare; only two agonists and three antagonists have been characterized with a similar preference for mGluR5 receptors (Doherty et al., 1997; Gasparini et al., 1999; Mannaioni et al., 1999; Varney et al., 1999). It is possible that mGluRs contribute to the epileptogenic activity of DH, although the different behavioral profiles of DH and trans-ACPD suggest that the selective action of DH on mGluR5 receptors does not account solely for the potent epileptogenic properties of DH. Further characterization of the pharmacological activity of DH on mGluRs, including group II and III receptors, is necessary to explore direct or indirect relationships of DH-induced seizure and mGluRs.

The striking potency that DH demonstrates for neuronal and recombinant KA receptors suggests that this activity predominantly underlies its efficacy for generation of seizures. KA receptors have been clearly implicated in the induction of seizures in the hippocampus (Ben-Ari, 1985; Sander et al., 1997; Mulle et al., 1998). Firing properties of CA3 pyramidal neurons are acutely sensitive to modulation by KA receptor agonists (Robinson and Deadwyler, 1981), and KA-induced seizurogenesis is commonly used as a model for the pathologies observed in human temporal lobe epilepsy (Nadler, 1981; Ben-Ari, 1985). In addition, susceptibility to KA-induced seizures was reduced in gene-targeted mice that lacked the GluR6 KA receptor subunit (Mulle et al., 1998).

The behavioral profile elicited by DH injection suggests that this compound may prove useful both as a new model for epileptogenesis studies and as a tool for characterizing the contribution of KA receptors to seizure behavior. Central or peripheral administration of DH in mice produced behavioral seizures that resembled those caused by DOM (Tryphonas et al., 1990a,b; Tasker et al., 1991) and to some extent with those by KA (Sperk, 1994). These similarities include preconvulsive behaviors and clonic convulsions followed by status epilepticus at the higher doses. However, the behavioral changes elicited by DH lasted much longer than those of DOM or other compounds. Furthermore, the ED₅₀ values we calculated for DH, 13 pmol/mouse i.c.v., and 0.97 mg/kg, i.p. were approximately 7-fold lower than those calculated for DOM in parallel experiments. Two behaviors induced by DHscratching at lower doses and long-lasting status epilepticuswere produced only by DOM among the other drugs we tested, suggesting that DOM and DH seizurogenesis may arise from activation of similar (or overlapping) receptor groups in the brain. Previously, it was proposed that DOM and KA activate different receptor populations to produce their seizurogenic behaviors, because in vivo excitotoxicity of DOM was significantly attenuated by NS 102, whereas that of KA was less sensitive to this drug (Tasker et al., 1996). Finally, it is worth noting that an intriguing difference emerged between the behaviors produced by DH and DOM: DH induced much longer lasting status epilepticus than other EAAs (including DOM). This difference could just be attributed to differences in the metabolic rate of the two drugs, or it could arise from a distinct spectrum of activities on ionotropic and metabotropic glutamate receptors.

Interestingly, in preliminary experiments DH was shown to activate homomeric recombinant GluR5 and GluR6 KA receptors at exceptionally low concentrations, whereas heteromeric KA receptors were significantly less sensitive to DH (G. T. Swanson, R. Sakai, A. Contractor, T. Green, H. Kamiya, and S. F. Heinemann, poster, 29th annual meeting, Society for Neuroscience, Miami Beach, FL, October 23-28, 1999). These results along with the potent binding affinity of DH to recombinant KA receptors suggest that DH exhibits a unique receptor subunit selectivity that may underlie its distinctive seizurogenic profile. Further characterization of the action and kinetics of DH on recombinant KA receptors is in progress.

In summary, we find that the novel amino acid DH is a non-NMDA agonist that possesses the most potent epileptogenic action among the known amino acids. Further investigations are necessary to fully characterize the mechanism of DH-evoked epileptogenic action. Nevertheless, unique binding, electrophysiological, and behavioral profiles, as well as the distinctive chemical structure of DH, warrant classification of this compound as a new type of EAA. We are now conducting a structure-activity relationship study of DH (Sasaki et al., 1999). With this information in hand, we anticipate that DH or its analogs may serve as useful pharmacological agents for investigating glutamate receptor function as well as lead compounds to develop therapeutic agents for central nervous system disorders.