Calcium-dependent inactivation of NMDA channels was examined on cultured rat hippocampal neurons using whole-cell voltage-clamp and cell-attached single-channel recording. An ATP regeneration solution was included in the patch pipette to retard current "rundown." In normal [Ca2+]o (1-2 mM) and 10 microM glycine, macroscopic currents evoked by 15 sec applications of NMDA (10 microM) inactivated slowly following an initial peak. At -50 mV in cells buffered to [Ca2+]i < 10(-8) M with 10 mM EGTA, the inactivation time constant (tau inact) was approximately 5 sec. Inactivation did not occur at membrane potentials of +40 mV and was absent at [Ca2+]o < or = 0.2 mM, suggesting that inactivation resulted from transmembrane calcium influx. The percentage inactivation and tau inact were dependent on [Ca2+]o. The tau inact was also longer with BAPTA in the whole-cell pipette compared to EGTA, suggesting that tau inact reflects primarily the rate of accumulation of intracellular calcium. Inactivation was incomplete, reaching a steady state level of 40-50% of the peak current. At steady state, block of open NMDA channels with MK-801 ((+)-5-methyl-10,11-dihydro-5H- dibenzo[a,d]cyclohepten-5,10-imine) completely blocked subsequent responses to NMDA, suggesting that "inactivated" channels can reopen at steady state. Inactivation was fully reversible in the presence of ATP but was not blocked by inhibiting phosphatases or proteases. In cell-attached patches, transient increases in [Ca2+]i following cell depolarization also resulted in inactivation of NMDA channels without altering the single-channel conductance. This suggests that Ca(2+)-dependent inactivation occurs in intact cells and can be triggered by calcium entry through nearby voltage-gated calcium channels, although calcium entry through NMDA channels was more effective. We suggest that [Ca2+]i transients induce NMDA channel inactivation by binding to either the channel or a nearby regulatory protein to alter channel gating. This mechanism may play a role in downregulation of postsynaptic calcium entry during sustained synaptic activity.

SUMMARY1. Responses to the excitatory amino acids kainate, quisqualate, N-methyl-Daspartate (NMDA), L-glutamate and L-aspartate were recorded in mouse hippocampal neurones in cell culture, using the whole-cell configuration of the patch clamp technique. Agonists were applied rapidly from an array of flow pipes each of 250 ,um diameter, positioned within 100 ,um of the nerve cell body.2 control, while responses to kainate and quisqualate were increased to 1-08 and 1 15 times control. With 1 mM-cadmium responses to NMDA were reduced to 0 04 times control while responses to kainate and quisqualate were reduced to 0 79 and 0-60 times control. Mercury was neurotoxic and increased the leakage current; however, no reduction of the response to NMDA was produced by 5/M-mercury. 4. The equilibrium dissociation constant (Kd) for zinc antagonism of responses to NMDA, estimated from fit of a single binding site adsorption isotherm, was 13 ,UM; cadmium was about 4 times less potent than zinc. These effects of zinc and cadmium were nearly voltage independent. In contrast the antagonism of responses to NMDA by 150 /tm-magnesium was highly voltage dependent, such that the Kd for magnesium increased e-fold per 17-6 mV depolarization.

The metabotropic glutamate receptor (mGluR) cDNAs were originally cloned from rat, except for the mouse cDNA clone encoding mGluR8. Mouse mGluR8 couples weakly to the inhibition of adenylate cyclase, thus hindering the characterization of its pharmacological properties. We isolated a rat mGluR8 cDNA that encodes a protein of 908 amino acids. In situ hybridization revealed prominent mGluR8 mRNA expression in olfactory bulb, pontine gray, lateral reticular nucleus of the thalamus, and piriform cortex. Less abundant expression was detected in cerebral cortex, hippocampus, cerebellum, and mammillary body. Glutamate evoked pertussis toxin-sensitive potassium currents in Xenopus laevis oocytes coexpressing mGluR8 and G protein-coupled inwardly rectifying potassium channels. mGluR8 was also activated by the group III-specific agonist L-2-amino-4-phosphonobutyric acid; (2(S), 1'(S), 2'(S)]- 2-(carboxycyclopropyl)glycine, which has been frequently used as a selective group II agonist; and the nonselective agonist (1(S), 3(R)]-1-aminocyclopentane-1,3-dicarboxylic acid but not by the group I-specific agonist 3,5-dihydroxyphenylglycine or the group II-specific agonist [2(S), 1'(R), 2(R), 3'(R)]-2-(2, 3-dicarboxycyclopropyl)glycine. The agonist profile in order of potency was [2(S), 1'(S), 2'(S)]-2-(carboxycyclopropyl)glycine approximately L-2-amino-4-phosphonobutyric acid> glutamate > > [1(S), 3(R)]-1-aminocyclopentane-1, 3-dicarboxylic acid, with EC50 values of 0.63, 0.67, 2.5, and 47 microM, respectively. Both the group I/II-specific antagonist (R,S)-alpha-methyl-4-carboxyphenylglycine and the group III-specific antagonist alpha-methyl-amino-phosphonobutyrate inhibited mGluR8. The pharmacological profile of mGluR8 is distinct among mGluRs but closely matches that of presynaptic inhibition in some central nervous system pathways. Thus, cellular responses mediated by both group II and III agonists may in some cases reflect activation of mGluR8 rather than multiple mGluR subtypes.