The effects of glutamate in the nervous system are mediated by two main classes of receptors: the ionotropic glutamate receptors, ligand-gated channels classically permeable to cations, and the metabotropic glutamate receptors, G protein-coupled receptors with seven membrane spanning domains. The main objectives of the project were the analysis of the relation between structure and function in known glutamate receptors, the search for new agonists and antagonists and the cloning of new receptors.
Comparison of the structure of glutamate and nicotinic receptors.
At the start of the project it was assumed that the structure of glutamate ionotropic receptors resembled closely that of the prototypic ionotropic receptor, the nicotinic receptor. We obtained a series of evidences indicating that this is not the case, In particular for the NMDA receptor. In a study of this receptor combining site-directed mutagenesis, electrophysiology calcium imaging and cysteine-labelling, it was found that the asparagines of the M2 sequences of both the NR1 and the NR2 subunits all contribute to the narrow constriction of the pore responsible for the ionic selectivity of the channel, but are not located at homologous positions: the asparagine at the N site of NR1 faces the asparagine next to the N-site of NR2. The analysis of the role of these asparagines in the Mg block revealed that the two asparagines of NR2 (the N-site asparagine and its neighbor) are essential in the block of the channel by external Mg, but have little effect on the block by internal Mg. In contrast, the NR1 N-site asparagine represents a primary blocking site for the internal Mg block, which can be also suppressed by mutations downstream the N-site of NR1. These data confirm the hypothesis that the M2 region forms a loop reintering the cytoplasm at its C-terminal end that the structure of the NMDA receptors, but more generally of the vertebrate ionotropic glutamate receptors, is fundamentally different from the structure of the nicotinic, GABA and glycine receptors.
The structure of the nicotinic receptor was simultaneously studied with an increased resolution. The projection structure was obtained at 7.5 A resolution, and the performance of the electron microscope used to analyze the two-dimensional arrays of receptors was dramatically improved by making use of a special high voltage microscope equipped with a liqiquid-helium cooled stage
Finally a new ionotropic glutamate receptor, cloned from the invertebrate Lymnea stagnalis, was found to constitute a bridge between the two families of ionotropic receptors. This receptor differs from all other known ionotropic glutamate receptors by its sequence, which resembles that of the nicotinic receptor rather than that of vertebrate glutamate receptors. Functionally, it mediates a conductance to Cl ions instead of cations.
The Ca permeability of the recombinant kainate receptors constituted by homomeric assemblies of GLUR6 had been considered as rather large after the observation that in high extracellular Ca the reversal potential was close to 0 mV. Microfluorometry, however, did not reveal any Ca influx. This discrepancy was resolved when it was shown that the channel was permeable not only to cations but also to Cl- ions. Cl- permeability was measured using reversal potentials in asymmetric conditions and was found to be comparable to that of Cs (PCe/PCs = 0.8). These observations impose a reevaluation of the pore size of kainate receptors.
The subunit composition of native AMPA receptors has been investigated by combining RT-PCR with electrophysiological analysis. In cortical neurons a correlation was established between the kinetics of AMPA responses and the presence of specific AMPA receptor subunits. It was found in particular that the striking difference in desensitization between pyramidal and non-pyramidal neurons was well explained by the fact that `flop' variants dominate in the non-pyramidal neurons, while `flip' variants dominate in pyramidal neurons. Thus alternative splicing of AMPA receptors seems to play an important role in regulating synaptic function in a cell-type specific manner.
Interaction of metabotropic glutamate receptors (mGluRs) with G-proteins. In the search for the domains of metabotropic receptors involved in activation of G proteins, it was discovered that metabotropic receptors splice variants which have a long C-terminal tail (about 500 residues) (mGluR1a, mGluR5a, mGluR5b) have an intrinsic activity (coupling to G protein in the absence of agonist) which is not found in those having a shorter C-terminal tail (mGluR1b and mGluR1c). Despite their absence of homology with others G coupled receptors, mGluRs interact with the same domains of G proteins and in particular with their C-terminal domains. It was demonstrated that mGluRs which inhibit the adenylylcyclase (via Gi) can stimulate phospholipase C when they are co-expressed with a Gq (a G-protein activating phospholipase C) in which the last 4 residues had been replaced by the corresponding residues from Gi. This engineered cells constitute a good model for screening drugs acting on mGluRs having different coupling.
Pharmacological characterization of glutamate receptors.
The pharmacological characterization of native kainate receptors was studied in cultured neurons expressing GluR 5/6/7 subunits. By analyzing the effect of compounds like the `AMPA antagonists' NS 394 or NS 102 as well as cyclothiazides, it was possible to establish that the low affinity kainate binding sites previously characterized in binding experiments supported very rapidly desensitizing responses which, contrarily to the desensitizing AMPA responses, lack a cyclothiazide modulatory site AMPA receptors.
The comparative study of the effects of a large series of AMPA agonists and antagonists has shown that the structural requirements for activation and blockade of AMPA receptors are distinctly different, and supported the hypothesis that the AMPA receptors exist in an agonist and an antagonist conformation, probably in equilibrium. This study has also led to a particular potent AMPA antagonist (ATPO) which is expected to be of pharmaceutical importance. The analysis of the enantiomeric structure of the various compounds has also led to the concept of `functional partial agonism' in which a partial agonist action (e.g. that of APPA) is shown to be due to the opposite effects of two enantiomers ((S)-APPA is an agonist, (R)-APPA an antagonist). The use of enantiomerically pure EAA receptor ligands has also allowed to reevaluate receptor phenomena like desensitization and agonist enhancement via priming effects.
The recent explosion of new types of metabotropic glutamate receptors contrasts with the pacicity of compounds acting on these receptors. In the frame of the search for such new ligands, it was found that homo-AMPA is a highly selective agonist of mGluR6, a receptor which plays a key role in the retina.
The last two years have permitted us to reach a situation where the design of NMDA and AMPA receptor ligands will be possible on a rational basis. The data have given rise to a patent. Based on structure activity data available, we are in the process of developing a computer-based AMPA agonist pharmacophore model. This work is carried out in collaboration with the drug company, H. Lundbeck A/S/, Copenhagen.
Funding SchemeCSC - Cost-sharing contracts
CB2 2QH Cambridge