Japanese Journal of Physiology, 51, 511–521, 2001

Physiological and Pharmacological Characteristics of -Induced K؉-Current Response in the Ganglion Cells of Aplysia

Shingo KIMURA, Satoshi KAWASAKI, Koichiro TAKASHIMA, and Kazuhiko SASAKI

Department of Physiology and Advanced Medical Science Research Center, School of Medicine, Iwate Medical University, Morioka, 020–8505 Japan

Abstract: The extracellular application of ei- the application of either kainate or AMPA, ago- ther quisqualic acid (QA) or Phe-Met-Arg-Phe- nists for non-NMDA receptors, produced no type NH2 (FMRFamide) induces an outward current in of response in the same neurons. The QA-in- identified neurons of Aplysia ganglion under volt- duced Kϩ-current response was not depressed age clamp. The time course of the QA-induced at all by an intracellular injection of either gua- response is significantly slower than that induced nosine 5Ј-O-(2-thiodiphosphate) (GDP-␤S) or by FMRFamide. The reversal potential for both guanosine 5Ј-O-(3-thiotriphosphate) (GTP-␥S), responses was Ϫ92 mV and was shifted 17 mV but the FMRFamide-induced response was in a positive direction for a twofold increase in markedly blocked by both GDP-␤S and GTP-␥S the extracellular Kϩ concentration. The QA-in- in the same cell. Furthermore, the QA- and FMR- duced response was markedly depressed in the Famide-induced Kϩ-current responses were both presence of Ba2ϩ, a blocker of inward rectifier decreased markedly when the temperature was Kϩ-channel, whereas TEA, a Ca2ϩ-activated Kϩ- lowered to 15¡C, from 23¡C. These results sug- ϩ ϩ channel (BKCa) blocker, or 4-AP, a transient K gested that the QA-induced K -current response (A)-channel blocker, had no effect on the re- is produced by an activation of a novel type of sponse. The QA-induced Kϩ-current was signifi- QA-receptor and that this response is not pro- cantly suppressed by CNQX and GYKI52466, duced by an activation of the G protein. [Japa- antagonists of non-NMDA receptors. However, nese Journal of Physiology, 51, 511Ð521, 2001]

Key words: quisqualic acid, , molluscan neuron, potassium channel.

Glutamate is a major neurotransmitter for excitatory tic potentials are produced by the activation of synaptic transmission in the central nervous system ionotropic glutamate receptors (iGluRs), which con- (CNS) of the vertebrate and the invertebrate. Besides sist of receptor–cation channel complexes as revealed producing the fast excitatory synaptic potentials in the by cDNA cloning. The iGluRs have been subdivided CNS, the synapse plays an important into the NMDA-, AMPA-, and kainate-receptor sub- role in the formation of synaptic plasticity, such as types by their pharmacological profiles [1]. Glutamate learning and memory. An excessive release of gluta- can also activate metabotropic glutamate receptors mate from the synaptic terminal is thought to be one (mGluRs), which are coupled to G proteins and a sub- cause for apoptosis, neuronal death. Glutamatergic sequent intracellular messenger system, such as phos- synaptic transmission is also important for the under- phatidyl inositol (PI) turnover. Usually an activation standing of various neuronal diseases. of the mGluRs produces slow synaptic potentials and In glutamatergic synapse, fast excitatory postsynap- primarily mediates the modulatory function of the cell

Received on February 21, 2001; accepted on June 11, 2001 Correspondence should be addressed to: Shingo Kimura, Department of Physiology and Advanced Medical Science Research Center, School. of Medicine, Iwate Medical University, Uchimaru 19–1, Morioka, 020–8505 Japan. Tel: ϩ81–19–651–5111 ext 3344, Fax: ϩ81–19– 625–3447, E-mail: [email protected]

Japanese Journal of Physiology Vol. 51, No. 4, 2001 511 S. KIMURA et al. response. At least eight subtypes of mGluRs in mam- (QA)-receptor that is quite different from those of malian CNS, termed as mGluR1–mGluR8, have been mammalian neurons. Furthermore, this response is identified by cDNA cloning [2]. produced by an increase in Kϩ permeability without In invertebrate neurons, the application of gluta- activation of the G protein, but it is possibly regulated mate induces an excitatory response similar to the ver- by a certain enzymatic process. tebrate neurons [3, 4]. Furthermore, inhibitory re- sponses are also induced by the stimulation of another METHODS type of glutamate receptor in molluscan neurons (see review by Cleland [5]). For instance, in the neurons of Preparation and perfusion media. The buc- Planorbarius corneus and Aplysia californica, an ap- cal ganglion or abdominal ganglion was isolated from plication of glutamate produces a hyperpolarizing re- Aplysia kurodai and fixed in a perfusing chamber. The sponse because of the opening of the ClϪ-channel [6, connective tissues covering the caudal surface of the 7]. Furthermore, Bolshakov et al. [8], Katz and Levi- buccal ganglion or the dorsal surface of the abdominal tan [9], Kehoe [10], and Watanabe et al. [11] have re- ganglion were carefully removed under a dissecting ported other types of hyperpolarizing responses pro- microscope to expose the cells to the perfusing duced by an increase in Kϩ conductance after receptor medium. Neurons of the buccal ganglia responded to stimulation in the neurons of Planorbarius corneus, either QA or FMRFamide, but not to GABA, whereas Aplysia californica, and Euhadra peliomphala. neurons of the abdominal ganglia responded to GABA We previously hypothesized that most of the recep- and FMRFamide, but not to QA. We therefore used tor-induced Kϩ-current responses are produced by an the buccal ganglia in most experiments, except the activation of the common G protein, Gi or Go, regard- temperature dependence of the GABA-induced re- less of their receptor types, based on experimental evi- sponse. The artificial Aplysia serum solution con- dence [12]. Actually, recent studies on invertebrates tained (in mM) NaCl, 587; MgCl2, 52; KCl, 11.5; ϩ neurons have demonstrated that a slow K -current re- CaCl2, 14; and Tris, 50.1 (pH 7.4) [15]. The cells were sponse induced by glutamate is mediated by pertussis continuously perfused at a constant flow rate of toxin (PTX)-sensitive G protein activation [8]. Studies 5 ml/min and at a temperature of 23Ϯ2°C. The effec- of some vertebrate neurons have also shown that glu- tive perfusing volume of the chamber was 0.2 ml. tamate evokes a hyperpolarizing response as a result Electrical measurements. Two microelec- of the activation of PTX-sensitive G protein. For ex- trodes filled with 1.8 M potassium citrate with a resis- ample, Holmes et al. have shown that the activation of tance of 5 to 7 M⍀ were inserted into a single cell. Ei- 2ϩ ϩ Ca -activated K channels (BKCa channel) through a ther the current- or voltage-clamp method was used to group II mGluR is linked to a PTX-sensitive G pro- evaluate the response of the neurons to the bath-ap- tein in basolateral amygdala neurons of rat [13]. plied agonists. The membrane potential was usually Knoflach and Kemp have also reported that a group II held at Ϫ65 mV, which was close to the resting poten- mGluR-induced hyperpolarizing response produced tial of the cells studied. To evaluate the change in by an opening of the inward rectifying Kϩ channel is membrane slope conductance, repetitive 1 s, 5 to coupled to G protein in rat cerebellum neuron, though 10 nA current pulses or 3 to 6 mV voltage pulses at the type of G protein involved in this response is not 0.2 Hz were applied across the cell membrane in the yet identified [14]. current- or voltage-clamp methods, respectively. On the other hand, Kehoe [10] and Watanabe et al. Drug application. Drugs used were L-glutamic [11] demonstrated in molluscan neurons that gluta- acid (Kanto Chemical); quisqualic acid (QA, Sigma); ϩ mate produces an outward K -current response, Phe-Met-Arg-Phe-NH2 (FMRFamide, Peptide Insti- which does not appear to be mediated by G protein. tute); ␥-aminobutyric acid (GABA, Wako); N-methyl- However, pharmacological properties and the mecha- D- (NMDA, RBI); (Ϯ)-␣-amino-3-hy- nism involved in ionic channel gating in this type of droxy-5-methylisoxazole-4-propionic acid (AMPA, glutamate receptor response have not been studied in RBI); 2-carboxy-4-(1-methyl-ethenyl)-3-pyrrolidin- detail. acetic acid (kainate, RBI); trans-(Ϯ)-1-amino- In this paper, we studied physiological and pharma- (1S,3R)-cyclopentanedicarboxylic acid (transACPD, cological characteristics of the slow hyperpolarizing RBI); trans-2,4-azetidine-dicarboxylic acid (tADA, response to stimulation of the glutamate receptor in RBI); L(ϩ)-2-amino-4-phosphonobutyric acid (L- the identified neurons of the buccal ganglion of AP4, RBI); (Ϯ)-2-amino-5-phosphonopentanoic acid Aplysia. We report that this type of glutamate receptor (AP-5, RBI); 6-cyano-7-nitroquinoxaline-2,3-dione: belongs to a novel subtype of quisqualic acid HBC complex (CNQX, RBI); 1-(4-aminophenyl)-4-

512 Japanese Journal of Physiology Vol. 51, No. 4, 2001 Quisqualic Acid–Induced Kϩ-Current methyl-7,8-methylenedioxy-5H-2,3-benzodiazepine RESULTS hydrochloride (GYKI52466, RBI); (Ϯ)-␣-amino- 4-carboxy-␣-methyl-benzeneacetic acid ((Ϯ)MCPG, Characteristics of quisqualic acidÐinduced RBI); R(ϩ)-2-amino-3-phosphonopropanoic acid 3- response phosphono-L- (L(ϩ)-AP-3, RBI); guano- In current-clamp experiments, the application of ei- sine 5Ј-O-(2-thiodiphosphate) (GDP-␤S, Boehringer ther quisqualic acid (QA) or Phe-Met-Arg-Phe-NH2 Mannheim); guanosine 5Ј-O-(3-thiotriphosphate) (FMRFamide) induces a slow hyperpolarizing re- (GTP-␥S, Boehringer Mannheim); ethyleneglycol-bis- sponse associated with a decrease in membrane resis- (␤-aminoethylether)N,N,NЈ,NЈ-tetraacetic acid (EGTA, tance in both BR6 and BL6 cells [16, 17] of the buc- Dojin); tetraethylammonium (TEA, Kanto Chemical); cal ganglion of Aplysia (Fig. 1A). The current re- and 4-aminopyridine (4-AP, Wako). sponse of the same cell to either QA or FMRFamide Glutamate, QA, FMRFamide, GABA, NMDA, were recorded under voltage clamp at resting potential AMPA, kainate, transACPD, tADA, L-AP4, AP-5, (Ϫ65 mV). An application of QA produced a slowly CNQX, (Ϯ)MCPG, and L(ϩ)-AP-3 were dissolved developing outward current response associated with into the normal artificial Aplysia medium and extra- an increase in membrane slope conductance (Fig. 1B). cellularly applied to the neuron. GYKI52466 was dis- The rise and decay times of the QA-induced response solved in methanol as a stock solution and diluted were significantly slower when compared with those with normal perfusing medium. After each application of the FMRFamide-induced response. It has been of the agonists, the cells were washed with the normal demonstrated that the FMRFamide-induced outward Aplysia medium for at least 15 to 20 min. GDP-␤S current in identified neurons of the abdominal gan- and GTP-␥S were dissolved in the solution containing glion of Aplysia is mediated by the Kϩ channel open- 200 mM KCl. We filled the GDP-␤S or GTP-␥S solu- ing as a result of the activation of a PTX-sensitive G tion into a third glass microelectrode and intracellu- protein [18, 19]. larly applied to the cells by an N2 pressure pulse with An application of glutamate to the same cell in- a duration of 100 to 300 ms and a pressure of 1 to duced a similar outward current response. The rise 3 kg/cm2. time of the glutamate-induced response resembled Intracellular concentrations of GDP-␤S and GTP- that of the FMRFamide-induced response; the decay ␥S were estimated by measuring the size of a droplet time of the glutamate-induced response resembled ejected from the electrode in the air and by calculating that of the QA-induced response. This response the volume ratio of the droplet to the cell examined. seemed to consist of fast and slow outward current The volume ratio described above is an underestima- components, since the fast component inverted its po- tion because the surface tension of the droplet in the larity at Ϫ70 mV, the slow component remaining to be air is larger than it is in water. Thus we used a correc- outward current as Kehoe had already suggested in tion factor of 1.5 as the volume ratio, which was de- other neurons [10]. The remaining component of the termined by measuring the sizes of the droplets glutamate-induced response at Ϫ70 mV was almost ejected from the same electrode in the air and in the completely occluded when examined during the QA- paraffin oil. induced outward current response. Figure 2A shows the dose-dependence of the out-

Fig. 1. FMRFamide-, glutamate-, and qui- squalic acid (QA)-induced responses of the same neuron in the buccal ganglion of Aplysia. A: Hyperpolarizing responses (V) under current clamp. Periodic downward de- flections in the voltage record show the electro- tonic potentials induced by constant 1 s inward current pulses (I) at every 5 s, indicating the changes in effective membrane resistance. B: Outward current responses (I) to the same ago- nists recorded from the same cell under voltage clamp at Ϫ65 mV. Periodic downward deflec- tions in the current record indicate the changes in slope conductance. Upward is outward in the current recording.

Japanese Journal of Physiology Vol. 51, No. 4, 2001 513 S. KIMURA et al.

Fig. 2. A: Dose-dependence of the QA-in- duced outward current response. The concen- tration of QA is shown in mM on the left of each trace. B: A doseÐresponse curve for QA-in- duced outward current. The current responses to given concentrations of QA were normalized to the peak response induced by 30 mM QA. Each point with a vertical bar represents the mean value with standard error (SE) obtained from 5 cells. Cell membrane was voltage-clamped at Ϫ65 mV.

Fig. 3. Effects of the changes in the extra- cellular K؉ concentration on the QA-induced response. A: Current–voltage relationships ob- tained with varying concentrations of extracellu- ϩ ϩ lar K ([K ]o). Peak current responses (I) in- duced by 300 ␮M QA were plotted as a function of the membrane potentials (V). Reversal poten- tials of the QA-induced responses were ϩ Ϫ92 mV in 12 mM [K ]o (᭹), Ϫ75 mV in 24 mM ϩ ϩ [K ]o (᭿), and Ϫ66 mV in 36 mM [K ]o (᭡). Each point represents the meanϮSE of the re- sponses recorded from 5 different cells. B: De- pendence of the reversal potential (Vrev) on the ϩ [K ]o. The abscissa is indicated as a logarith- ϩ mic scale for [K ]o. The straight line is drawn from the Nernst equation for Kϩ. ward current responses induced by various concentra- were determined. A twofold and threefold increase in + tions of QA. The QA-induced current became de- [K ]o shifted the reversal potential in a positive direc- tectable at a concentration of 30 ␮M and increased tion by 17Ϯ4 (nϭ5) and 26Ϯ3 mV (nϭ5), respec- sigmoidally with an increase in the QA concentration. tively. These values were approximately equal to those The relationship between the peak current amplitude predicted by the Nernst equation at 23°C, i.e., 17 and and the QA concentration is shown in Fig. 2B. The 25.5 mV, respectively. Therefore the relationship be- ϩ half-maximum value of the effective concentration tween Vrev and [K ]o on the logarithmic scale was lin- (EC50) was 0.9Ϯ0.3 mM (nϭ5), and the maximal re- ear, and the data points fell on the straight line drawn ϩ sponse was produced at a concentration of 30 mM QA. from the Nernst equation for K . These findings sug- gested that the QA-induced hyperpolarization might Reversal potential be produced by an increase in the permeability of the Figure 3A shows the current (I)–voltage (V) curve membrane exclusively toward Kϩ. of the 300 ␮M QA-induced response measured at holding potentials from Ϫ95 to Ϫ65 mV. The I–V Effects of various ionic channel blockers and curve had a slight outward rectification at this poten- altered ionic environment on the QA-induced tial range. The reversal potential for the QA-induced response response (Vrev) estimated from the intersection on the Since ionic species producing the QA-induced re- voltage axis was Ϫ92Ϯ3 mV (nϭ5) in normal perfus- sponse were suggested to be Kϩ channel, we further ing medium. This value was close to the Kϩ equilib- surveyed to identify possible subclasses of the Kϩ rium potential (EK) of Ϫ89 mV calculated from the channel and to examine the involvement of ions other Nernst equation by Gruol and Weinreich [20]. To fur- than Kϩ that affect the generation of the QA-induced ϩ ther confirm the involvement of an increase in the K response. As shown in Fig. 4, neither 10 mM tetraethy- permeability in the QA-induced response, I–V curves lammonium (TEA), an inhibitor of Ca2ϩ-activated Kϩ ϩ ϩ ϩ at different concentrations of extracellular K ([K ]o) channel (BKCa channel) and voltage-dependent K

514 Japanese Journal of Physiology Vol. 51, No. 4, 2001 Quisqualic Acid–Induced Kϩ-Current

Fig. 4. Effects of TEA (A), 4-AP (B), -and Ba2؉ (C) on the QA-induced K؉ current response. Controls are on the left, test responses in each inhibitor are in the middle, and recovery re- sponses examined 30 min after the in- hibitors are on the right. Records shown in A, B, and C were obtained from three different cells under a volt- age clamp at Ϫ65 mV. Note the marked depressing action of Ba2ϩ on the QA- induced response, and the lack of ef- fect of both TEA and 4-AP.

channel, nor 10 mM 4-aminopyridine (4-AP), an in- nists for glutamate receptors. hibitor of A channel, depressed the QA-induced Kϩ- The application of either NMDA, AMPA, or current response. In contrast, the application of 5 mM kainate, agonists for iGluR, even at a concentration of ϩ Ba2 , which is known to block the inward rectifying 10 mM, evoked no significant current on the resting Kϩ, M, and S channels, markedly depressed the QA- membrane (data not shown, nϭ3). The QA-induced ϩ induced response reversibly. However, 100 ␮M gliben- K -current response was not affected by the applica- clamide or 1 mM tolbutamide, blockers of ATP-sensi- tion of 300 ␮M AP-5, a selective antagonist for the ϩ tive K channel (KAT P channel), did not inhibit the NMDA receptor (data not shown, nϭ5). However, the QA-induced response (data not shown). QA-induced response was completely depressed in ϩ Extracellular perfusion of the cells with Na -free the presence of 300 ␮M CNQX, a competitive antago- solution completely eliminated the QA-induced re- nist for non-NMDA receptors (Fig. 5A, nϭ6). A simi- sponse, the effect being reversible by washing with lar depressing effect was observed in the presence of normal perfusing medium (data not shown). Further- 300 ␮M GYKI52466, a noncompetitive antagonist for more, extracellular perfusion of Ca2ϩ-free solution non-NMDA receptors, as shown in Fig. 5B. The QA- significantly and reversibly augmented the QA-in- induced response was decreased to 62.2Ϯ5.4% (nϭ5) duced response up to 173.1Ϯ28.2% (nϭ5) of control of the control by the 300 ␮M GYKI52466. The de- (data not shown). Nevertheless, the intracellular injec- pressing effects of CNQX and GYKI52466 were re- tion of EGTA, a Ca2ϩ chelator, did not affect the QA- versible in either case. induced response, though it significantly depressed On the other hand, the application of either the depolarization-induced outward current (data not transACPD, an agonist for mGluR, tADA, a selective shown). agonist for mGluR1,5, or L-AP4, a selective agonist for mGluR4,6, induced no response even at a concen- Pharmacological characteristics of receptor tration of 1 mM (data not shown). Furthermore, the producing the QA-induced current QA-induced Kϩ-current response was not affected by A variety of glutamate receptor types are character- the application of either 1 mM L(ϩ)-AP-3 or 1 mM ized by differences in their pharmacological properties (Ϯ)MCPG, antagonists for mGluR (data not shown). and amino acid sequences [21–23]. QA is an agonist for both the AMPA/ and group I Intracellular application of GDP-␤S and GTP- metabotropic glutamate receptor (mGluR1,5) in the ␥S vertebrate neurons [2, 22]. Therefore we next exam- In general, most of the receptor-induced Kϩ-current ined the type of receptor that was activated by QA on responses are produced by the activation of the com- the B6 neuron by using various agonists and antago- mon G protein, Gi or Go, regardless of their receptor

Japanese Journal of Physiology Vol. 51, No. 4, 2001 515 S. KIMURA et al.

Fig. 5. Effects of CNQX, a com- petitive antagonist for non-NMDA receptor (A), and GYKI52466, a noncompetitive antagonist for non-NMDA receptor (B), on the .QA-induced K؉-current responses A and B were recorded from differ- ent cells under a voltage clamp at Ϫ65 mV. The perfusion of each an- tagonist started 5 min before the test with QA.

Fig. 6. A: Effects of GDP-␤S on the FMRFamide- and QA-induced responses of the same cell. Control responses are on the left, and test re- sponses examined 50 and 65 min after an intracellular injection of GDP- ␤S are on the right. Note the lack of any effect of GDP-␤S on the QA-in- duced response, but it almost com- pletely blocked the FMRFamide-in- duced response. The intracellular concentration of GDP-␤S was esti- mated to be 750 ␮M. B: Effects of in- tracellularly applied GTP-␥S on the responses of the same cell that were induced by FMRFamide (filled arrowhead) and QA (open arrow- head). GTP-␥S, injected into the cell at the time shown by the arrow, pro- duced a gradually increasing Kϩ-cur- rent response. The intracellular con- centration of GTP-␥S was estimated to be 150 ␮M. The resting membrane was clamped at Ϫ65 mV in all the ex- periments. types [12]. GDP-␤S is a GDP analog known to com- pletely within 20 min (Fig. 6A). The blocking effect of pete with GTP for a common binding site at G pro- GDP-␤S was irreversible. Contrary to our expecta- teins and inhibits the GTP-dependent activation of G tions, however, the QA-induced Kϩ-current response proteins [24]. To investigate whether G protein medi- recorded from the same cell was not depressed at all ates the production of the Kϩ-current response in- by GDP-␤S. The size of the QA-induced response duced by QA, we applied GDP-␤S intracellularly. An was 98.9Ϯ8.2% (nϭ5) of the control after the injec- intracellular injection of GDP-␤S caused no change in tion of GDP-␤S. either the slope conductance or the holding current of GTP-␥S is a nonhydrolyzable analog of GTP and the resting membrane. However, an injection of GDP- known to cause irreversible activation of G proteins ␤S gradually depressed the FMRFamide-induced Kϩ- [25]. The intracellular application of GTP-␥S irre- current response and finally blocked it almost com- versibly produces an outward current response associ-

516 Japanese Journal of Physiology Vol. 51, No. 4, 2001 Quisqualic Acid–Induced Kϩ-Current

Fig. 7. Effects of lowering the temperature on the GABA- (A), FM- RFamide- (B), and QA- (C) induced responses. Lowering the tempera- ture significantly increased the GABA-induced ClϪ-current response, but it markedly decreased both the FMRFamide- and QA-induced Kϩ- current responses. The records in A are from a cell in the RC cluster of the abdominal ganglion; those in B and C are from the same single cell in the buccal ganglion. Membrane poten- tials were clamped at Ϫ60 mV in A and at Ϫ65 mV in B and C.

ated with an increase in membrane conductance to- DISCUSSION ward Kϩ [12, 26]. The Kϩ current induced by GTP-␥S gradually developed to reach a plateau level (Fig. 6B). Electrophysiological and pharmacological -As already reported, the FMRFamide-induced Kϩ- characteristics of K؉ channel involved in QA current response was irreversibly and almost com- induced outward current. The present study pletely suppressed when examined 20 min after GTP- demonstrated that the application of QA to the B6 ␥S injection. In contrast, the QA-induced Kϩ-current neuron of the buccal ganglia of Aplysia produces an response of the same cell was not affected by GTP- outward current response associated with a slowly de- ␥S. The QA-induced response after the GTP-␥S-in- veloping and decaying time course. The I–V curve of jection was 103.7Ϯ11.6% (nϭ5) of the control. These the QA-induced response exhibited a weak outward results indicate that the QA-induced Kϩ-current re- rectification. Reversal potential of the QA-induced re- sponse is not mediated by activation of any type of G sponse was predicted by the Nernst equation for Kϩ, protein. indicating that the response is produced exclusively by the opening of the Kϩ channel. The QA-induced Kϩ- Modulation of the receptor-induced responses current response was not depressed by 10 mM TEA, by lowering the temperature 10 mM 4-AP, or 100 ␮M glibenclamide. But it was ϩ In the neuron of the RC cluster in the abdominal markedly depressed by 5 mM Ba2 (Fig. 4). Further- ganglion of Aplysia [27], the GABAA receptor-in- more, the response was not depressed by intracellu- duced ClϪ-current response was significantly in- larly injected EGTA, but it was augmented as much as creased when the temperature was lowered to 15°C, 173% of the control in Ca2ϩ-free medium, and it com- from 23°C (Fig. 7A). In contrast, the FMRFamide-in- pletely disappeared in Naϩ-free medium. duced Kϩ-current response induced by metabotropic The lack of depressing effects of TEA, EGTA, and receptor markedly decreased in the B6 cell of the buc- Ca2ϩ-free medium on the QA-induced response indi- cal ganglion (Fig. 7B). Similar to the FMRFamide re- cates that this response is not produced by the opening sponse, QA-induced Kϩ-current response was of the Ca2ϩ-activated Kϩ channel. Similar QA- or glu- markedly decreased by lowering the temperature (Fig. tamate analog-induced Kϩ-currents have been ob- 7C). The responses to GABA, FMRFamide, and QA served in GH-1, -2, and -3 neurons of the suboe- recovered to the control levels when the temperature sophageal ganglion of Euhadra peliomphala [11], v- was returned to 23°C. The values of Q10 from 15 to LCDN neuron of African gland snail Achatinaflulica 23°C for the GABA-, FMRFamide-, and QA-induced ferussac [28], and B1 and B2 neurons of buccal gan- responses were 0.7, 3.5, and 3.1, respectively. glia of Aplysia californica [9, 10], respectively. All these authors also postulated that the QA-induced Kϩ-

Japanese Journal of Physiology Vol. 51, No. 4, 2001 517 S. KIMURA et al. current response is not due to the opening of the response to QA, did not respond to non-NMDA ago- Ca2ϩ-activated Kϩ channel in accordance with our nists, AMPA, and kainate, but it did respond to an conclusion. mGluR agonist, ACPD. Using the same neuron, how- Our new finding of the marked depressing effect of ever, Kehoe demonstrated noninvolvement of the G Ba2ϩ on the QA-induced Kϩ-current reminds us of the protein in the QA-induced Kϩ-current response [10]. involvement of the ATP-dependent inward-rectifying Thus the type of receptor producing the QA-induced ϩ ϩ K channel (KAT P channel). However, both gliben- K -current response seems to remain controversial clamide and tolbutamide were without the effect on even among mollusk neurons. This discrepancy may the QA-induced Kϩ-current. It is known that Ba2ϩ can be due to differences in subspecies among the mol- ϩ also block the inward-rectifying K channel (KIR lusks. It should also be noted that the type of receptor channel), S channel, and M channel. However, the responsible for QA with Kϩ current is totally different outward rectification of the I–V curve for QA-induced from that of a non-NMDA receptor of mammalian current excludes the possibility to be the KIR channel. neurons. Furthermore, the possibility of being the S or M chan- Lack of G-protein activation in generation nel is also excluded because these channels are usu- of QA-induced response. In the pedal ganglion ally mediated by the G protein, unlike the QA-induced neuron of Planorbarius corneus, Bolshakov et al. [8] response of this study. Thus characteristics of the Kϩ have demonstrated that stimulation of glutamate re- channel producing the QA-induced response corre- ceptor leads to the Kϩ channel opening mediated by spond to no type of the previously reported Kϩ chan- the PTX-sensitive G protein. Katz and Levitan [9] nels. previously reported that glutamate- and QA-induced A major point characterizing the Kϩ channel pro- Kϩ-current responses involve G-protein activation be- ducing the QA-induced response is high sensitivity to cause ACPD, an mGluR agonist, produces the same Ba2ϩ. We speculate that this Kϩ channel may have a Kϩ-current response as glutamate and QA. Further- binding site for divalent cation in the extracellular more, they showed that a QA-induced Kϩ-current re- side, and this site may exhibit Ba2ϩ sensitivity because sponse substantially occludes the glutamate-evoked the QA-induced response is augmented by a Ca2ϩ-free currents. Thus we expected in our experiments that medium despite the response not being influenced by the QA-induced Kϩ-current response might be medi- intracellular EGTA. ated by G protein activation. As shown in the results, Pharmacological properties of the receptor however, the intracellular application of either GDP- for QA. The QA-induced response started to ap- ␤S or GTP-␥S did not affect the QA-induced re- pear from 30 ␮M and saturated at 30 mM, showing an sponse, but it completely blocked the FMRFamide-in- ϩ EC50 value of 0.9 mM when analyzed on the dose–re- duced K -current response of the same cell. More- sponse curve. Similar values for EC50 have been re- over, although both QA and FMRFamide produced ported in B1 and B2 neurons of Aplysia californica the Kϩ-current responses, there was no cross desensi- (EC50: 1 mM) [9] and in the suboesophageal ganglion tization between them (unpublished observation). neuron of Euhadra (EC50: 0.1–0.2 mM) [11]. Further- These facts strongly suggested that the activation of a more, the QA-induced response was markedly de- G protein might not mediate the QA-induced Kϩ-cur- pressed by antagonists for the non-NMDA receptor, rent response in this specimen. Kehoe [10] has also though it was not activated by agonists for the non- reported that the activation of glutamate receptor by NMDA receptor. Furthermore, either the agonists or the application of QA evokes a Kϩ-current response antagonists of the NMDA receptor were without the independently of G protein activation, and this current effect on the same neuron responding to QA. Neither response did not occlude the Kϩ-current response in- the agonist nor the antagonist of mGluR had an effect duced by the activation of muscarinic receptors in B1 on the neuron. These pharmacological results suggest and B2 neurons of Aplysia californica. Furthermore, that the receptor type for a QA-induced response may both glutamate and QA stimulate the same receptor have that of an AMPA receptor of mammalian neu- and evoke a Kϩ-current response in Euhadra neuron, rons. the response of which was not depressed by the intra- Watanabe et al. had reported that the QA-induced cellular injection of either pertussis toxin or cholera Kϩ-current response in Euhadra neuron was not de- toxin (CTX) [11]. pressed by a non-NMDA antagonist, CNQX, or an Possible coupling mechanism. The QA-in- NMDA antagonist, AP-5 [11]. On the other hand, duced Kϩ-current response had a significantly slow Katz and Levitan [9] reported that B1 and B2 neurons time course when compared with the FMRFamide-in- of Aplysia californica, which produced the Kϩ-current duced Kϩ-current response of the same neuron. Fur-

518 Japanese Journal of Physiology Vol. 51, No. 4, 2001 Quisqualic Acid–Induced Kϩ-Current thermore, this response was markedly increased by NMDA or QA of rat hippocampal neuron were not af- raising the temperature to 23°C, from 15°C, with Q10 fected significantly by a rise in temperature (unpub- of 3.1 in a fashion similar to the FMRFamide-induced lished observation). It is presumed that these types of response. The Q10 value of 3.1 is comparable with the receptors might increase the single-channel conduc- value of Q10 3.5 of the FMRFamide-induced response. tance in response to a temperature increase, despite A similar increase in receptor-induced current re- the simultaneous decrease in the mean open time [36]. sponses with high Q10 has been reported in responses In summarizing these possibilities, it is suggested produced by D1- and D2-receptors in Aplysia neuron that the QA-induced response in the present study [29], ␤-receptor in guinea pig cardiac muscle [30], may be mediated or alternatively regulated by a cer- 5HT4 receptor in rat vagus nerve [31], and M2-recep- tain unidentified enzymatic reaction, regardless of the tor in Aplysia neuron [32], the natures of which are ionotropic receptor or the enzyme-mediated receptor. common to the receptor responses mediated by enzy- Furthermore, an ATP-deficient experiment made by matic molecules such as G proteins. These results Kehoe would not completely exclude the enzymatic therefore strongly suggest that the production of the process in generation or regulation of the QA-induced QA-induced Kϩ-current response may be mediated or response; she herself considers that further experi- regulated by a certain enzymatic process. ments are necessary to test the possible involvement On the other hand, it is evident from the discussion of protein tyrosine kinase, guanylate kinase, and vari- in the previous section that G protein does not seem to ous types of phosphatase [10]. Recently, it has been mediate the generation of the QA-induced response. reported that stimulation of the mGluR1 receptor pro- Kehoe also concluded a noninvolvement of G protein duces the opening of a cationic channel directly by ac- in the QA-induced response. From her experiment of tivation of Src-type of protein tyrosine kinase without intracellular dialysis by whole-cell recording, she fur- activation of the trimeric G protein [37]. Furthermore, ther postulated that cytosolic soluble mediator would stimulation of a certain type of cytokine receptor fam- not be involved in the generation of the QA-induced ily, such as leptin receptor, can activate a protein phos- response [10]. Our results from pharmacological tests phatase to open the KAT P channel without the activa- of the receptor type also only partly suggest that the tion of trimeric G protein [38]. QA-induced Kϩ-current response may be produced by Physiological significance of the QA-in- an activation of ionotropic glutamate receptor, espe- duced response. It is well established that all the cially the AMPA-type. QA-induced responses reported in vertebrate neurons Recently, Chen et al. found an ionotropic glutamate are fast excitatory responses produced by the opening receptor GluR0 that selectively opens the Kϩ channel of the cationic channels formed by ionotropic recep- in a prokaryote Synechocystis PCC6803 [33]. The tors, i.e., the AMPA type. In contrast, QA-induced re- amino acid sequence of this receptor channel complex sponses reported in mollusk CNS are slow inhibitory is closely related to the eukaryotic glutamate receptors responses with quite long time courses. Besides these and the Kϩ channels. Therefore, the receptor involved characteristics reported previously, we demonstrated in the QA-induced response may also have a structure in the present study that the QA-induced response in similar to that of GluR0. Aplysia neuron is highly sensitive to Ba2ϩ, Ca2ϩ-free, Again back to the temperature dependence, how- and Naϩ-free solutions and that the response is in- ever, if we provide that the QA-induced response is creased by raising the temperature with high Q10 mediated by ionotropic receptors, the QA-induced re- value. These newly identified characteristics suggest sponse should show a fast time course and a decrease that this type of QA-receptor may alter its property of in amplitude under raised temperature, as seen in the the response in a flexible way in responding to GABA-induced ClϪ-current response. In fact, it has changes in extracellular ionic environment and tem- been reported that ionotropic-receptor responses are perature. These types of characteristics, resembling decreased by raising the temperature, presumably be- those of the metabotropic receptor responses mediated cause of a decrease in mean open time at a single by trimeric G protein, may be adequate to play a mod- channel level. Such examples include a fast Naϩ-cur- ulatory role of the synaptic transmission instead of the rent response of dopamine receptor in Aplysia neuron fast transmission of information in CNS. Further- Ϫ [29], Cl -current responses of GABAA receptor in rat more, a certain enzymatic process, possibly involved astrocyte [34] and of nicotinic receptor in pond snail in the generation or regulation of this type of QA-in- CNS neurons [35], and a cationic-current response of duced response, suggested in the previous section, 5HT3-receptor in rat vagus nerve [31]. As a few ex- may easily cause a cross-talk with another type of sig- ceptional examples, ionotropic receptor-responses to naling pathway in the same cell. The B6 neuron used

Japanese Journal of Physiology Vol. 51, No. 4, 2001 519 S. KIMURA et al. for present study is known to innervate the intrinsic dent of G proteins. Neuron 13: 691–702, 1994 muscle of the buccal mass, and it produces EJP by re- 11. Watanabe K, Yamamoto T, Ozono S, Nagasaki S, and Onozuka M: A novel glutamate-mediated inhibitory leasing glutamate from its nerve ending [16, 39]. Be- ϩ mechanism linked with Ca2 /calmodulin–dependent cause Aplysia often has a long pause in feeding behav- protein kinase II in identified Euhadra neurons. J Neu- ior in response to sudden environmental change, this robiol 32: 139–149, 1997 type of QA-receptor in the B6 neuron may play a 12. Sasaki K and Sato M: A single GTP-binding protein ϩ functional role in producing this pause. regulates K -channels coupled with dopamine, hista- We now have no ideal model that interprets the gen- mine and acetylcholine receptors. Nature 325: ϩ 259–262, 1987 eration and regulation of the QA-induced K -current 13. Holmes KH, Keele NB, Arvanov VL, and Shinnick-Gal- response. Nevertheless, we speculate that this type of lagher P: Metabotropic glutamate receptor agonist-in- receptor may exist in the mammalian CNS and play a duced hyperpolarizations in rat basolateral amygdala functional role in the mammalian brain. Further ex- neurons: receptor characterization and ion channels. J periments are necessary to clarify the intracellular Neurophysiol 76: 3059–3069, 1996 mechanism involved in the generation and regulation 14. Knoflach F and Kemp JA: Metabotropic glutamate group II receptors activate a G protein–coupled in- of the QA-induced response. wardly rectifying Kϩ current in neurones of the rat cere- bellum. J Physiol (Lond) 509: 347–354, 1998 The authors thank Mr. Housaku Ito and Mr. Kiyonori 15. Sato M, Austin G, Yai H, and Maruhashi J: The ionic Ichikawa for a constant supply of Aplysia, and Dr. Paul permeability changes during acetylcholine-induced re- Langman for his critical reading of the manuscript. Part of sponses of Aplysia ganglion cells. J Gen Physiol 51: this study has already been reported at the 75th Annual 321–345, 1968 Meeting of the Physiological Society of Japan. This study 16. Gardner D: Bilateral symmetry and interneuronal orga- has been partly supported by Grants-in-Aid for Advanced nization in the buccal ganglia of Aplysia. Science 173: Medical Science Research, and Scientific Research from 550–553, 1971 the Ministry of Education, Culture, Sports, Science and 17. Gardner D and Kandel ER: Physiological and kinetic Technology of Japan. properties of cholinergic receptors activated by multi- action interneurons in buccal ganglia of Aplysia. J REFERENCES Neurophysiol 40: 333–348, 1977 18. Sasaki K, Takahashi J, Matsumoto M, Takashima K, 1. Seeburg PH: The TINS/TiPS Lecture. The molecular bi- Hakozaki S, and Sato M: Islet activating protein-sensi- ology of mammalian glutamate receptor channels. tive guanosine triphosphate-binding protein regulates TINS 16: 359–365, 1993 Kϩ-channels coupled with FMRFamide receptors. Jpn 2. Schoepp DD, Jane DE, and Monn JA: Pharmacological J Physiol 37: 551–557, 1987 agents acting at subtypes of metabotropic glutamate 19. Brezina V, Eckert R, and Erxleben C: Modulation of receptors. Neuropharmacology 38: 1431–1476, 1999 potassium conductances by an endogenous neu- 3. Henry CD, Leslie J, and Kulovich S: Circulating free ropeptide in neurones of Aplysia californica. J Physiol amino acids in Aplysia californica. Comp Biochem (Lond) 382: 267–290, 1987 Physiol 100: 629–632, 1991 20. Gruol DL and Weinreich D: Two pharmacologically dis- 4. Ichinose M, Sawada M, and Maeno T: Excitatory effect tinct histamine receptors mediating membrane hyper- of amino acids on identified neuron R14 of Aplysia. I. polarization on identified neurons of Aplysia californica. -induced depolarization and its ionic mecha- Brain Res 162: 281–301, 1979 nism. J Neurosci Res 14: 129–143, 1985 21. Nakanishi S: Molecular diversity of glutamate receptors 5. Cleland TA: Inhibitory glutamate receptor channels. and implications for brain function. Science 258: Mol Neurobiol 13: 97–136, 1996 597–603, 1992 6. Sawada M, Hara N, Ito I, and Maeno T: Ionic mecha- 22. Hollmann M and Heinemann S: Cloned glutamate re- nism of a hyperpolarizing glutamate effect on two iden- ceptors. Annu Rev Neurosci 17: 31–108, 1994 tified neurons in the buccal ganglion of Aplysia. J Neu- 23. Dingledine R, Borges K, Bowie D, and Traynelis SF: rosci Res 11: 91–103, 1984 The glutamate receptor ion channels. Pharmacol Rev 7. Kehoe J: Transformation by concanavalin A of the re- 51: 7–61, 1999 sponse of molluscan neurones to L-glutamate. Nature 24. Eckstein F, Cassel D, Levkovitz H, Lowe M, and 274: 866–869, 1978 Selinger Z: Guanosine 5Ј-O-(2-thiodiphosphate). An in- 8. Bolshakov V, Gapon SA, and Magazanik LG: Different hibitor of adenylate cyclase stimulation by guanine nu- types of glutamate receptors in isolated and identified cleotides and fluoride ions. J Biol Chem 254: neurones of the mollusc Planorbarius corneus. J Phys- 9829–9834, 1979 iol (Lond) 439: 15–35, 1991 25. Citri Y and Schramm M: Probing of the coupling site of 9. Katz PS and Levitan IB: Quisqualate and ACPD are ag- the beta-adrenergic receptor. Competition between dif- onists for a glutamate-activated current in identified ferent forms of the guanyl nucleotide binding protein Aplysia neurons. J Neurophysiol 69: 143–150, 1993 for interaction with the receptor. J Biol Chem 257: 10. Kehoe J: Glutamate activates a Kϩ conductance in- 13257–13262, 1982 crease in Aplysia neurons that appears to be indepen- 26. Brezina V: Guanosine 5Ј-triphosphate analogue acti-

520 Japanese Journal of Physiology Vol. 51, No. 4, 2001 Quisqualic Acid–Induced Kϩ-Current

vates potassium current modulated by neurotransmit- Brain Res 495: 298–308, 1989 ters in Aplysia neurones. J Physiol (Lond) 407: 15–40, 33. Chen GQ, Cui C, Mayer ML, and Gouaux E: Functional 1988 characterization of a potassium-selective prokaryotic 27. Frazier WT, Kandel ER, Kupfermann I, Waziri R, and glutamate receptor. Nature 402: 817–821, 1999 Coggeshall RE: Morphological and functional proper- 34. Kettenmann H, Backus KH, and Schachner M: ␥- ties of identified neurons in the abdominal ganglion of Aminobutyric acid opens Cl-channels in cultured astro- Aplysia californica. J Neurophysiol 30: 1288–1351, cytes. Brain Res 404: 1–9, 1987 1967 35. Dickinson R, Lieb WR, and Franks NP: The effects of 28. Zhang W, Takeuchi H, Kurono M, and Emaduddin M: temperature on the interactions between volatile gen- Ouabain-sensitive Kϩ-dependent outward current eral anaesthetics and a neuronal nicotinic acetyl- caused by threo-beta-hydroxy-L- on a choline receptor. Br J Pharmacol 116: 2949–2956, snail neuron. Gen Pharmacol 29: 625–632, 1997 1995 29. Matsumoto M, Hakozaki S, Takahashi J, Sasaki K, and 36. McLarnon JG and Curry K: Single channel properties Sato M: Temperature effects on dopamine-induced in- of the N-methyl-D-aspartate receptor channel using hibitory and excitatory responses recorded from NMDA and NMDA agonists: on-cell recordings. Exp Aplysia ganglion cells. Jpn J Physiol 38: 109–113, Brain Res 82: 82–88, 1990 1988 37. Heuss C, Scanziani M, Gahwiler BH, and Gerber U: 30. Walsh KB, Begenisich TB, and Kass RS: Beta-adrener- G-protein–independent signaling mediated by gic modulation of cardiac ion channels. Differential metabotropic glutamate receptors. Nature Neurosci 2: temperature sensitivity of potassium and calcium cur- 1070–1077, 1999 rents. J Gen Physiol 93: 841–854, 1989 38. Harvey J and Ashford ML: Role of tyrosine phosphory- 31. Bley KR, Eglen RM, and Wong EH: Characterization of lation in leptin activation of ATP-sensitive Kϩ channels 5-hydroxytryptamine–induced depolarizations in rat in the rat insulinoma cell line CRI-G1. J Physiol (Lond) isolated vagus nerve. Eur J Pharmacol 260: 139–147, 510: 47–61, 1998 1994 39. Fox LE and Lloyd PE: Glutamate is a fast excitatory 32. Salanki J, Evans ML, and Carpenter DO: Desensitiza- transmitter at some buccal neuromuscular synapses in tion kinetics of a Kϩ acetylcholine response in Aplysia. Aplysia. J Neurophysiol 82: 1477–1488, 1999

Japanese Journal of Physiology Vol. 51, No. 4, 2001 521