Introduction

Top Summary Introduction Procedures Results Discussion References

A wide range of organic compounds, including ethanol and volatile anesthetics, produce a plethora of behavioral effects, such as impairment of learning, memory, and motor coordination and, at higher doses, loss of consciousness and surgical immobility. Despite the wide use of these drugs, their mechanism of action remains obscure. There is evidence that certain neurotransmitter receptors, including ligand-gated ion channels and G protein-coupled (metabotropic) receptors, are affected by ethanol and volatile anesthetics (Franks and Lieb, 1994; Harris et al., 1995). In particular, the function of muscarinic cholinergic and 5-HT₂ receptors is inhibited by these compounds (Durieux, 1995; Minami et al., 1997; Sanna et al., 1994). However, not all metabotropic receptors are affected by anesthetics; halothane does not inhibit the function of the AT_1A angiotensin II receptor, a receptor that uses the same phosphatidylinositol signaling system as the m1 muscarinic receptor, which is sensitive to halothane (Durieux, 1995). One goal of the current study was to determine whether the actions of ethanol and volatile anesthetics extend to two mGluRs. The mGluRs are distinct from the other metabotropic receptors in that they are much larger proteins and show little sequence similarity to most members of the G protein-coupled receptor family, although there is appreciable homology with the -aminobutyric acid_B receptors (Kaupmann et al., 1997). As discussed below, mGluRs are important modulators of synaptic transmission in the mammalian central nervous system and are believed to play a role in processes such as memory and learning; therefore, it was of interest to determine whether the function of these receptors is affected by ethanol or anesthetics.

Previous studies suggest that ethanol and anesthetics inhibit metabotropic receptor function by increasing PKC-dependent desensitization of the receptor (Minami et al., 1997; Sanna et al., 1994) and it was of interest to determine whether this mechanism could account for the actions of these drugs on mGluRs. PKC phosphorylates mGluR5, and this phosphorylation is important in producing oscillations in intracellular Ca²⁺ signaling (Kawabata et al., 1996). In addition, recent data suggest that specific serine residues in mGluR5 are important for desensitization of this receptor, a process that is mediated by PKC (Gereau and Heinemann, in press). Thus, mutation of these sites provides a test of the hypothesis that ethanol and anesthetics inhibit receptor function because they enhance this endogenous PKC-mediated desensitization.

Several new halocarbon compounds may also provide insight regarding the role of metabotropic receptors in anesthetic action. Although most halogenated hydrocarbons produce anesthesia, and it is generally assumed that potency of general anesthetics is determined solely by lipid solubility, recent studies show a more subtle structure-activity relationship for anesthesia (Koblin et al., 1994). In particular, a novel halogenated compound (F3) is an anesthetic, whereas its congener (F6) does not produce anesthesia, and another halocarbon, F8, also is nonanesthetic, despite their high lipid solubilities (Koblin et al., 1994). Although F6 does not produce anesthesia (or, more accurately, surgical immobility), it does interfere with learning and memory, and this property is shared with traditional anesthetic compounds (Gonsowski et al., 1995; Kandel et al., 1996). Thus, different sites of action may be responsible for the immobilizing and amnestic actions of anesthetics; the former would be sensitive to F3 but not F6, whereas the latter would be affected by both compounds. Indeed, several ligand-gated ion channels are affected by F3 but not F6, whereas a metabotropic receptor (5-HT_2A) is sensitive to both (Harris et al., 1995; Minami et al., 1997). One goal of the current study was to determine whether mGluRs are affected by both anesthetic and nonanesthetic halocarbons.

Part of the rationale for studying the mGluRs is their importance in learning and memory. As mentioned above, ethanol, anesthetics, and at least one nonanesthetic interfere with learning and memory (Dwyer et al., 1992; Givens and McMahon, 1997; Gonsowski et al., 1995; Kandel et al., 1996). A neurotransmitter that has a major role in synaptic excitation in the central nervous system and is critical for information storage in memory and learning is glutamate (Hollman and Heinemann, 1994). The ionotropic glutamate receptors are influenced by anesthetics. For example, volatile anesthetics inhibit N-methyl-D-aspartate and AMPA receptor function but enhance kainate (GluR6) receptor responses (Harris et al., 1995). However, these actions are not shared by the nonanesthetic F6, even though it interferes with learning and memory (Harris et al., 1995). Thus, the ionotropic glutamate receptors are unlikely to account for all the amnestic actions produced by halocarbons.

The mGluRs form a family of receptors; eight different subtypes have been described (mGluR1-8; Abe et al., 1992; Conn and Pin, 1997). Based on their pharmacology, second messenger coupling, and sequence differences, these receptors can be divided into class I (mGluR1 and mGluR5), class II (mGluR2 and mGluR3), and class III (mGluR4 and mGluR6-8) (Riedel, 1996). The class I receptors are linked with learning and memory (Riedel, 1996); this is based on pharmacological studies showing that an agonist of mGluR1 and mGluR5 enhances memory and that mutant mice lacking mGluR1 show deficits in learning and memory and reduced hippocampal LTP (Aiba et al., 1994; Conquet et al., 1994; Riedel, 1996). Mice lacking the mGluR1 also display poor motor coordination. More recently, it was reported that mice lacking mGluR5 show impaired learning and reduced CA1 LTP but normal CA3 LTP (Lu et al., 1997). In view of the effects of ethanol and anesthetics on learning and memory as well as motor function, it is of interest to consider the effects of these drugs on class I mGluRs.

There are few studies of the effects of alcohols or anesthetic agents on mGluRs. Ethanol inhibits quisqualate-induced burst activity in rat cultured cerebellar Purkinje neurons (mediated by mGluRs) (Netzeband and Gruol, 1995) and quisqualate-induced currents in oocytes expressing mRNA from rat cerebellum (Sanna et al., 1994) but does not affect glutamate-stimulated phospholipase C activity of brain astrocytes (Smith, 1994). These results suggest that some, but not all, subtypes of mGluRs are inhibited by ethanol and encouraged us to test the effects of ethanol and anesthetics on specific subtypes of mGluRs expressed in Xenopus laevis oocytes.

The X. laevis oocyte expression system has been used to express a multiplicity of brain receptors from cDNAs or cRNAs with pharmacological properties that mimic those of native brain receptors (Harris et al., 1995; Snutch, 1988). Activation of mGluR1 or mGluR5 receptors results in activation of phospholipase C, mobilization of calcium stores, and activation of an endogenous Ca²⁺-dependent Cl current in oocytes (Abe et al., 1992; Masu et al., 1991). This system has been well characterized for the study of the effects of anesthetics and ethanol on G protein-coupled receptors.

We used the oocyte expression system to study the effects of ethanol, halothane, F3, F6, and F8 on glutamate-induced current via mGluR1 and mGluR5 (class I). Moreover, we analyzed the effects of these compounds in the presence of a PKC inhibitor and a protein phosphatase inhibitor and studied receptors with mutations in PKC phosphorylation sites to investigate the role of PKC in the actions of these drugs.

	Experimental Procedures

Top Summary Introduction Procedures Results Discussion References

Materials. Adult X. laevis female frogs were purchased from Xenopus I (Ann Arbor, MI). Glutamate, dimethylsulfoxide, phorbol-12-myristate-13-acetate, and L-glutamate were purchased from Sigma Chemical (St. Louis, MO). Ethanol was purchased from Aaper Alcohol and Chemical (Shelbyville, KY). Calyculin A was from LC Laboratories (Woburn, MA). Halothane was from Halocarbons Laboratories (River Edge, NJ). F3, F6, and F8 were obtained from PCR Inc. (Gainesville, FL). Ultracomp Escherichia coli transformation kit was from InVitrogen (San Diego, CA). A kit from Qiagen (Chatworth, CA) was used for purification of plasmid cDNA. mGluR1 and mGluR5 cRNAs were prepared using mCAP mRNA capping kit (Stratagene, La Jolla, CA). GF109203X and chelerythrine were from Calbiochem (La Jolla, CA). mGluR1 and mGluR5 cDNAs were kindly provided by Dr. S. Nakanishi (Kyoto University, Kyoto, Japan).

Metabotropic glutamate cRNA preparation. The cDNA for mGluR1 was inserted into the pGEM vector, and the cDNA for mGluR5 was inserted into the pBluescript SK vector. The mGluR cDNAs were linearized with NotI, phenol-chloroform extracted, and ethanol precipitated with sodium acetate and cRNA prepared using the Stratagene transcription kit. These cRNAs were extracted using phenol-chloroform and precipitated with ethanol and sodium acetate. Site-directed mutagenesis was performed according to the procedure for the Quik-Change site-directed mutagenesis kit (Stratagene) (Gereau and Heinemann, in press). These mutants were inserted into the pBluescript SK vector; the cDNAs were linearized with XbaI, phenol-chloroform extracted, and ethanol precipitated with ethanol and sodium acetate.

Whole-cell voltage-clamp of injected oocytes. Isolation and microinjection of X. laevis oocytes were performed as described by Sanna et al. (1994). X. laevis oocytes were injected with 25-50 ng of cRNA coding for the mGluRs. Oocytes were placed in a 100-µl recording chamber and perfused with MBS (containing 88 mM NaCl, 1 mM KCl, 2.4 mM NaHCO₃, 10 mM HEPES, 0.82 mM MgSO₄, 0.33 mM Ca(NO₃)₂, 0.91 mM CaCl₂, pH 7.5) at rate of 1.8 ml/min at room temperature. Recording and clamping electrodes (1-5 M) were pulled from 1.2-mm o.d. capillary tubing and filled with 3 M KCl. A recording electrode was impaled into the animal pole; once the resting membrane potential stabilized, a clamping electrode was inserted with the resting membrane potential allowed to restabilize. Warner oocyte clamp OC 725-B (Hampden, CT) was used for voltage-clamping of each oocyte at 70 mV.

C_m measurements. Capacitative transients elicited by the voltage jumps from the holding potential of 30 to 60 mV were measured (Vasilets et al., 1990). Three voltage pulses of 30 mV were applied from the resting potential. The signal, which was averaged from the three pulses, was integrated using the Strathclyde Electrophysiology Software program (courtesy of John Dempster, Glasgow, UK), Whole Cell Electrophysiology Program V1.2e, and C_m was determined from the slope of the regression of the area-versus-voltage relationships.

Statistical analysis. Results are expressed as percentages of control responses due to variability in oocyte expression. The control responses were measured before and after each drug application to take into account possible shifts in the control currents as recording proceeded. However, in the experiments to study the effects of GF109203X or PMA, we used the glutamate-induced currents before GF109203X or PMA application as a control. Each experiment was carried out with oocytes from at least two different frogs. Statistical analyses were performed using either a t test or one-way ANOVA. Curve fitting and estimation of EC₅₀ values for concentration-response curves were performed using Inplot (GraphPAD Software, San Diego, CA).

	Results

Top Summary Introduction Procedures Results Discussion References

Effects of ethanol, anesthetics, and nonanesthetics on currents activated by glutamate in oocytes expressing mGluRs. Glutamate concentration-response curves were determined in X. laevis oocytes expressing mGluR1 and mGluR5 (Fig. 1). Nonlinear regression analysis of these curves of mGluR1 yielded an EC₅₀ value for glutamate of 1 mM and a Hill coefficient of 1.1. Maximal currents were observed at 10 mM. X. laevis oocytes expressing mGluR5 yielded an EC₅₀ value for glutamate of 0.4 mM and a Hill coefficient of 1.4. 

View larger version (22K): [in this window] [in a new window]   Fig. 1.   A, Concentration-response curves for glutamate-activated Ca2+-dependent Cl current in X. laevis oocytes expressing mGluR1 and mGluR5. Oocytes were voltage-clamped at 70 mV; glutamate was applied for 20 sec, and the peak current was measured. Nonlinear regression analysis of the concentration-response curves of mGluR1 yielded an EC50 value for glutamate of 1 mM and a Hill coefficient of 1.1. Maximal currents were observed at 10 mM. Concentration-response curves of mGluR5 yielded an EC50 value for glutamate of 0.4 mM and a Hill coefficient of 1.4. Maximal currents were observed at 3 mM. Values are mean ± standard error from six to eight oocytes. B and C, Effects of anesthetics [halothane (Hal.) and F3], nonanesthetics (F6 and F8), and ethanol (EtOH) on the current evoked by 1 mM glutamate in oocytes expressing mGluR1 (B) and on the current evoked by 100 µM glutamate in oocytes expressing mGluR5 (C). Halothane (0.06-2 mM), F3 (0.05-0.8 mM), F6 (0.06-17.8 µM), F8 (0.06-8.8 µM), and ethanol (12.5-200 mM) were preapplied for 2 min before being coapplied with glutamate for 20 sec. Values are mean ± standard error of 9-14 oocytes.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   A and B, Tracings were obtained from a single oocyte and show the effects of halothane (0.25 mM) and ethanol (200 mM) on currents evoked by 100 µM glutamate (Glu.) in oocytes expressing mGluR5s. The control responses were measured 20 min before and 20 min after each drug application. C, Tracings were obtained from a single oocyte and show the effect of GF109203X on currents evoked by 100 µM glutamate in oocytes expressing mGluR5s. Oocytes expressing mGluR5s were incubated with GF109203X (200 nM) for 2 hr.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   A, Effects of anesthetic (MAC) concentrations of halothane, F3, F6, F8, and ethanol on currents evoked by 1 mM glutamate in oocytes expressing mGluR1s. The MAC concentrations of ethanol (EtOH), halothane (Hal.), and F3 are 190, 0.25, and 0.8 mM, respectively. The predicted MAC concentrations of F6 and F8 are 17.8 and 8.8 µM, respectively. Values are mean ± standard error of five oocytes. *, p < 0.05 versus control response with 1 mM glutamate (paired t test). B, Effects of anesthetic (MAC) concentrations of halothane, F3, F6, F8, and ethanol on currents evoked by 100 µM glutamate in oocytes expressing mGluR5. Values are mean ± standard error of five or six oocytes. **, p < 0.01 versus control response obtained with 100 µM glutamate (paired t test).

Effects of a PKC inhibitors on mGluR function. Because PKC plays an important role in regulating some G protein-coupled receptors (Kato et al., 1988; Manzoni et al., 1990; Moran and Dascal, 1989; Sanna et al., 1994) and has been shown to regulate mGluR function (Catania et al., 1991; Schoepp and Johnson, 1988), we studied glutamate responses in X. laevis oocytes that were pretreated with the PKC inhibitor GF109203X (200 nM) (Toullec et al., 1991). GF109203X had no effects on currents activated by glutamate in the oocytes expressing mGluR1 but markedly (213 ± 55%) enhanced glutamate-induced currents in oocytes expressing mGluR5 (Fig. 4, A and B) without altering the EC₅₀ value for glutamate (0.2 mM) or the Hill coefficient (1.5). We also studied the effects of the PKC activator PMA on glutamate-induced currents in oocytes expressing mGluR5. We measured the 100 µM glutamate-induced currents as a control and perfused the oocytes with MBS for 20 min. We then treated the oocytes with 50 and 10 nM PMA for 5 min and tested them with 100 µM glutamate. The 5-min application of the PKC activator PMA inhibited 100 µM glutamate-evoked currents to 6 ± 4% (six oocytes) and 11 ± 4% (six oocytes) of control response at 50 and 10 nM PMA, respectively. However, PMA (50 nM for 5 min) had no effect on mGluR1 function (control, 759 ± 157 nA, 11 oocytes; PMA treatment, 747 ± 224 nA, 10 oocytes). Enhancement by GF109203X was observed with maximal (3 mM) and submaximal (300 µM) concentrations of glutamate in oocytes expressing mGluR5. Treatment with GF109203X enhanced the 3 mM glutamate-induced currents to 245 ± 41% of initial currents and 300 µM glutamate-induced currents to 192 ± 39% of initial currents.

View larger version (25K): [in this window] [in a new window]   Fig. 4.   A, Oocytes expressing mGluR1s were incubated with GF109203X (200 nM) for 2 hr. Glutamate (1 mM) was applied at 5, 20, 40, 60, and 120 min during the treatment with GF109203X or at 20, 40, 80, and 120 min in oocytes not treated with GF109203X (control). Values are mean ± standard error of nine oocytes. B, Oocytes expressing mGluR5s were incubated with GF109203X (200 nM) for 2 hr. Glutamate (100 µM) was applied at 5, 20, 40, 60, and 120 min in oocytes treated with GF109203X or at 20, 40, 80, and 120 min in oocytes not treated with GF109203X (control). Values are mean ± standard error of 14-20 oocytes. C, GF109203X (GF) blocked the action of anesthetics [halothane (Hal.) and F3] and ethanol but not the nonanesthetic F6. Inhibition of glutamate-induced responses was measured in oocytes expressing mGluR5s. Oocytes were incubated with 200 nM GF109203X for 2 hr. Halothane (0.25-2 mM) or F3 (0.2-0.8 mM) was preapplied for 2 min before being coapplied with glutamate (100 µM) for 20 sec. The anesthetics, nonanesthetics, and ethanol (EtOH) were preapplied for 2 min before being coapplied with glutamate (100 µM) for 20 sec. Values are mean ± standard error of five separate determinations. **, p < 0.01 versus control response obtained with 100 µM glutamate (paired t test).

View larger version (22K): [in this window] [in a new window]   Fig. 5.   A, Effects of halothane and ethanol (EtOH) on mGluR5 function in oocytes treated with the protein phosphatase inhibitor calyculin A (Cal A). The control current produced by glutamate (Glu.) (100 µM) was measured 20 min before injection of 30 nl of calyculin A (2 µM) or 30 nl of water. At 5 min after injection, halothane (0.25 mM) or ethanol (200 mM) was applied for 2 min. After the application of these compounds, oocytes were washed with buffer for 3 min, and the current was remeasured. Bars, effects of injection with water on control responses (A), effects of ethanol (200 mM) on oocytes injected with water (B), effects of halothane (0.25 mM) on oocytes injected with water (C), effects of injection with calyculin A (D), effects of ethanol (200 mM) on oocytes injected with calyculin A (E), and effects of halothane (0.25 mM) on oocytes injected with calyculin A (F). Values are mean ± standard error of five or six oocytes. Statistical analyses were performed using a paired t test. *, p < 0.05 versus control response.

View larger version (11K): [in this window] [in a new window]   Fig. 6.   Tracings were obtained from a single oocyte and showed the effects of ethanol (EtOH) (200 mM) and halothane (Hal.) (0.25 mM) on oocytes injected with calyculin A (CalA). The control current produced by glutamate (100 µM) was measured 20 min before the injection of 30 nl of calyculin A (2 µM). Five minutes after injection, halothane (0.25 mM) or ethanol (200 mM) was applied for 2 min. After the application of these compounds, oocytes was washed with buffer for 3 min, and the current was remeasured.

Effects of ethanol, anesthetics, and nonanesthetics on mutant mGluR5s. Abe et al. (1992) suggested several possible phosphorylation sites in mGluR5, some of which represent consensus phosphorylation sites for PKC. One recent study has shown that desensitization of mGluR5 expressed in oocytes is mediated by PKC (Gereau and Heinemann, in press). Furthermore, this study showed that at least two of the PKC consensus sites in mGluR5 (Ser613 and Ser890) are important for this PKC-mediated desensitization. In the current study, these two sites were mutated on mGluR5, and the glutamate concentration-response curves were determined in X. laevis oocytes expressing mGluR5(S890G) and mGluR5(S613G) receptors (Fig. 7). Analysis of (S613G) yielded an EC₅₀ value for glutamate of 2 mM and a Hill coefficient of 1.6, and mGluR5(S890G) yielded an EC₅₀ value for glutamate of 1.5 mM and a Hill coefficient of 1.6. The magnitude of the current produced by maximally effective concentrations of glutamate was greater for the S890G mutation than for the wild-type or S613G mutation. The glutamate (100 µM)-induced currents on the wild-type, mGluR5(S613G), and mGluR5(S890G) were 512 ± 129 nA (27 oocytes), 1470 ± 303 nA (17 oocytes, p < 0.05), and 2116 ± 432 nA (18 oocytes, p < 0.01) (unpaired t test and Dunnett's correction), respectively.

View larger version (15K): [in this window] [in a new window]   Fig. 7.   Responses of mutant receptors to glutamate and anesthetic agents. A, Concentration-response curves for glutamate-activated Ca2+-dependent Cl current in X. laevis oocyte expressing mGluR5, mGluR5(S613G), or mGluR5(S890G). Oocytes were voltage-clamped at 70 mV; glutamate was applied for 20 sec, and the peak current was measured. Values are mean ± standard error from five oocytes. B, Representative tracings of currents induced by 100 µM glutamate in oocytes expressing mGluR5, mGluR5(S613G), or mGluR5(S890G). Glutamate was applied for 20 sec. Calibration bars, 10 sec and 200 nA.

View larger version (17K): [in this window] [in a new window]   Fig. 8.   Effects of ethanol, halothane, F3, or F6 on currents evoked by glutamate in oocytes expressing mGluR5, mGluR5(S613G), or mGluR5(S890G). Anesthetic or predicted anesthetic (MAC) concentrations were used (see Fig. 3); the glutamate concentration was 100 µM. Values are mean ± standard error of five or six oocytes. Statistical analyses were performed using unpaired t test and Dunnett correction. *, p < 0.05 and **, p < 0.01 versus wild-type (W-T).

View larger version (15K): [in this window] [in a new window]   Fig. 9.   Tracings obtained from a single oocyte show the effects of ethanol (200 mM) and halothane (0.25 mM) in oocytes expressing mGluR5(S890G). The control responses were measured 20 min before and 20 min after each drug application. Glu., glutamate.

	Discussion

Top Summary Introduction Procedures Results Discussion References

Our results demonstrate that ethanol, halothane, and F3 inhibited mGluR5 but had little effect on mGluR1 function. The amino acid sequence of mGluR5 is homologous to that of the other members of the mGluR family and most closely related to that of mGluR1 (60% sequence identity) (Abe et al., 1992). The question arises as to how these anesthetics and ethanol inhibit mGluR5 but not mGluR1 function. Several mechanisms of anesthetic action on G protein-coupled receptors have been suggested. Durieux (1995) proposed that the site of interaction between halothane and muscarinic receptors could be a hydrophobic domain in the receptor protein. However, we suggested that anesthetics and ethanol do not act directly on metabotropic receptors but rather enhance receptor phosphorylation by PKC and thereby inhibit receptor function (Minami et al., 1997; Sanna et al., 1994). The current study provides substantial evidence to support this hypothesis; the findings can be summarized as follows: 1) The function of the mGluR5 is reduced by an activator of PKC and enhanced by a selective PKC inhibitor, indicating that receptor function may be modulated in both directions by changes in PKC activity. 2) A selective PKC inhibitor completely abolished ethanol-, halothane-, and F3-induced inhibition of mGluR5 function. (It should be noted that the effects of anesthetics and ethanol in the absence or presence of PKC inhibitor were tested at the same EC value; thus, the reduction in the anesthetic effects in the presence of the PKC inhibitors is not due to an inhibitor-induced shift in the glutamate concentration-response curve.) 3) A protein phosphatase inhibitor was able to sustain the actions of ethanol and anesthetics even after they were removed from the bath, suggesting that these drugs do not act directly on the receptor (and do not act on these protein phosphatases). 4) Mutation of a putative PKC phosphorylation site (Ser890) enhanced receptor function and prevented the action of ethanol and anesthetics. It is important to note that mGluR1 does not have a sequence corresponding to Ser890 and its function was not affected by an activator or inhibitor of PKC or by ethanol or anesthetics. However, both mGluR1 and mGluR5 contain identical PKC consensus sequences at Ser613, and we found that this amino acid is not critical for modulation of receptor function by PKC activators/inhibitors but may influence the action of some anesthetics. Although Ser890 is clearly the dominant site, Ser613 may play some role in mediating the effects of ethanol and anesthetics. Thus, our data suggest that the sensitivity of mGluR5 to ethanol and anesthetics could be due to a consensus PKC phosphorylation site at Ser890. Taken together, our results provide evidence that ethanol and anesthetics indirectly inhibit metabotropic receptor function by activation of PKC. It should be noted that the studies of effects of anesthetics on PKC activity in vitro have had mixed results. For example, halothane stimulates PKC activity in brain synaptosomes (Hemmings and Adamo, 1996, 1997), brain cytosol (Tsuchiya et al., 1988), and PC12 cells (Tas and Koeschel, 1991). Studies of purified PKC are less consistent, and both inhibition and enhancement of PKC activity have been obtained with halothane and ethanol (Hemmings et al., 1995; Hemmings and Adamo, 1994; Slater et al., 1993, 1997). Further studies are needed to determine whether ethanol and anesthetics increase the phosphorylation of mGluR5 and whether Ser890 or Ser613 is indeed phosphorylated by PKC. Moreover, additional experiments will be necessary to investigate whether anesthetics modulate PKC access to its phosphorylation site in the receptor. It is possible that the Ser890 site is important for the coupling of the receptor with a specific G protein. It also is possible that the Ser890 site is important for the coupling of the receptor with a specific G protein. Another possibility is raised by the recent report that the protein "Homer" regulates the metabotropic glutamate signaling by an interaction with the carboxyl-terminal domain of mGluR1 and mGluR5 (Brakeman et al., 1997). This alternative mechanism proposes that activation of PKC by alcohols and anesthetics inhibits receptor function by phosphorylation of receptor-associated proteins rather than the receptor. Further studies are required to distinguish among these hypotheses.

It is of interest to consider the functional importance of mGluR5 and possible consequences of inhibition of receptor function by ethanol and anesthetics. In situ hybridization found prominent expression of mGluR5 mRNA in cerebral cortex, nucleus accumbens, striatum, hippocampal CA1-4 and dentate gyrus regions, lateral septum and cerebellar Golgi, and internal granule cells (Abe et al., 1992). This localization raises the possibility of roles in cognition, learning and memory, reinforcement, and motor control. There is recent evidence that inhibition of mGluR1 and/or mGluR5 can impair learning and memory and reduce motor coordination and anxiety (Aiba et al., 1994; Chojnacka-Wojcik et al., 1997; Conquet et al., 1994; Lu et al., 1997; Riedel, 1996).

The nonanesthetic F6 inhibited function of both mGluR1 and mGluR5 equally, and this inhibition was not affected by the PKC inhibitor or by mutation of Ser890. This is consistent with our earlier finding that F6 inhibits the function of 5-HT_2A receptors by a mechanism that does not require PKC (Minami et al., 1997). The inhibition of mGluR5 receptor function by both F3 and F6 (albeit by different mechanisms) is of interest in view of the report (Kandel et al., 1996) that F6 suppresses learning at doses of 0.5-1 times the predicted MAC. In earlier studies, concentrations of F6 corresponding to these doses had no effect on several ligand-gated ion channels affected by F3 (Dildy-Mayfield et al., 1996; Mascia et al., 1996; Mihic et al., 1994), making it unlikely that these channels are important for the amnestic actions of F6 (and perhaps of anesthetics). However, the metabotropic receptors may be responsible for some actions, such as amnesia, that are shared by both F3 and F6. Conversely, it is unlikely that inhibition of metabotropic receptor function by anesthetics is responsible for immobility because F6 does not produce this component of anesthesia, yet it inhibits metabotropic receptor function (Minami et al., 1997).

In conclusion, our results show that ethanol and volatile anesthetics inhibited mGluR5 function but had little effect on mGluR1. This receptor specificity is likely due to the presence of a site (Ser890) in mGluR5 that is absent in mGluR1. These findings, together with those of others (Minami et al., 1997; Sanna et al., 1994), suggest that alcohols and anesthetics amplify an endogenous regulatory mechanism that uses PKC phosphorylation to reduce receptor function.