␥-Aminobutyric acid (GABA) activates two qualitatively different inhibitory mechanisms through ionotropic GABA A multisubunit chloride channel receptors and metabotropic GABA B G protein-coupled receptors. Evidence suggests that pharmacologically distinct GABA B receptor subtypes mediate presynaptic inhibition of neurotransmitter release by reducing Ca 2ϩ conductance, and postsynaptic inhibition of neuronal excitability by activating inwardly rectifying K ϩ (Kir) conductance. However, the cloning of GABA B gb1 and gb2 receptor genes and identification of the functional GABA B gb1-gb2 receptor heterodimer have so far failed to substantiate the existence of pharmacologically distinct receptor subtypes. The anticonvulsant, antihyperalgesic, and anxiolytic agent gabapentin (Neurontin) is a 3-alkylated GABA analog with an unknown mechanism of action. Here we report that gabapentin is an agonist at the GABA B gb1a-gb2 heterodimer coupled to Kir 3.1/3.2 inwardly rectifying K ϩ channels in Xenopus laevis oocytes. Gabapentin was practically inactive at the human gb1b-gb2 heterodimer, a novel human gb1c-gb2 heterodimer and did not block GABA agonism at these heterodimer subtypes. Gabapentin was not an agonist at recombinant GABA A receptors as well. In CA1 pyramidal neurons of rat hippocampal slices, gabapentin activated postsynaptic K ϩ currents, probably via the gb1a-gb2 heterodimer coupled to inward rectifiers, but did not presynaptically depress monosynaptic GABA A inhibitory postsynaptic currents. Gabapentin is the first GABA B receptor subtype-selective agonist identified providing proof of pharmacologically and physiologically distinct receptor subtypes. This selective agonism of postsynaptic GABA B receptor subtypes by gabapentin in hippocampal neurons may be its key therapeutic advantage as an anticonvulsant.

Arachidonyl trifluoromethyl ketone (AACOCF3) is a slow- and tight-binding inhibitor of the human cytosolic phospholipase A2 (cPLA2) [Street et al. (1993) Biochemistry 32, 5935]. 19F and 13C NMR experiments have been carried out to elucidate the structure of the cPLA2.AACOCF3 complex. One mole of AACOCF3 per mole of enzyme is tightly bound in the active site while excess molar equivalents of the inhibitor associate loosely and nonspecifically with hydrophobic regions of the protein. Incubation of the cPLA2.AACOCF3 complex with a 10-fold molar excess of a structurally related inhibitor allows the slow dissociation of the enzyme-inhibitor complex to be followed with 19F NMR. These results establish that the bound inhibitor is in slow exchange with the free ligand and that inhibition of the cPLA2 by AACOCF3 is not due to irreversible modification of the protein. AACOCF3 labeled with 13C at the carbonyl position was used to determine the nature of the bound inhibitor species. A comparison of the 13C NMR chemical shift value obtained from labeled enzyme-inhibitor complex (delta c 101.0 ppm) with the chemical shift values obtained from model compounds suggests that the enzyme-bound inhibitor species is a charged hemiketal. These results are very similar to those obtained previously with alpha-chymotrypsin and a peptidyl trifluoromethyl ketone inhibitor [Liang, T.-C., & Abeles, R. H. (1987) Biochemistry 26, 7603] and, by analogy with the serine proteases, a structural model for the cPLA2.AACOCF3 complex is proposed.