Molecular Modeling of N-Terminal Domains of NMDA-Receptor. Study of Ligand Binding to N-Terminal Domains

Ligands of N-terminal domains of the N-methyl-D aspartic acid (NMDA) receptor are of considerable interest as potential drugs for treatment of neurodegenerative diseases. The search for structures complementary to functionally important sites of receptors provides a rational approach to construction of pharmacological preparations. However, information about three-dimensional structure of ligand-binding site in the target macromolecule is a necessary prerequisite for the search for complementary structures. The goal of this work was to construct a computer model of the spatial structure of N-terminal domains (NR1 and NR2B subunits) of NMDA receptor. This model was designed to provide qualitative and quantitative evaluation of the efficiency of binding of presently known ligands to NMDA receptor. At the first stage of the study, the proteins with amino acid sequences homological to the sites of the Nterminal domains of the NR1 and NR2B subunits of NMDA receptor were searched in the Internet (http://www.sbg.bio.ic.ac.uk/~3dpssm). This search showed that, among the proteins studied by X-ray diffraction structural analysis, the highest degree of homology (18%) and the best scoring functions (taking into account the estimates of correspondence of the secondary structure of the target protein predicted using the program PSI-Pred [1] with the experimental secondary structure of the template and with solvation potential) was observed for the N-terminal domain of type 1 metabotropic glutamate receptor (mGluR1). Spatial coordinates of the mGluR1 [2] (entry code in protein database [3]: IEWV) complex with an antagonist (open conformation of protein) were used as a template for constructing spatial models of N-terminal domains of NR1 and NR2 subunits of NMDA receptor. At the next stage of the study, the amino acid sequences forming N-terminal domains of NR1 and NR2B subunits of NMDA receptor were aligned with the primary sequence of mGluR1 using the software available from the Internet (http://raven.bioc.cam.ac.uk/~kenij) [4] and implementing the threading method. Spatial models of N-terminal domains were constructed using the Syby16.7.2 software package of molecular graphics [5]. The geometry of each model was optimized using the methods of molecular mechanics based on the Tripos force field [5]. At the next stage of construction of this model, spa