How does binding of agonist ligands control intrinsic molecular dynamics in human NMDA receptors?

Abstract

NMDA-type glutamate receptors (NMDAR) are ligand-gated ion channels that contribute to excitatory neurotransmission in the central nervous system. NMDAR dysfunction has been found to be involved in various neurological disorders. Recent crystallographic and EM studies have shown the static structure of different states of the non-human NMDARs. Here we describe a model of a human NMDA receptor (hNMDAR) and its molecular dynamics (MD) before and after the binding of agonist ligands, glutamate and glycine. It is shown that the binding of ligands promotes a global reduction in molecular flexibility that produces a more tightly packed conformation than the unbound hNMDAR, and a higher cooperative regularity of moving. The ligand-induced synchronization of motion, identified on all structural levels of the modular hNMDA receptor is apparently a fundamental factor in channel gating. Although the time scale of the MD simulations (300 ns) was not sufficient to observe the complete gating event, the obtained data has shown the ligand-induced stabilization of hNMDAR that conforms the “going to be open state”. We propose a mechanistic dynamic model of the ligand-dependent gating mechanism in the hNMDA receptor. At the binding of the ligands, the differently twisted conformations of the highly flexible receptor are stabilized in unique conformation with a linear molecular axis, which is a condition that is optimal for pore development. By searching the receptor surface, we have identified three new pockets, which are different from the pockets described in the literature as the potential and known positive allosteric modulator binding sites. A successful docking of two NMDAR modulators to their binding sites validates the model of a human NMDA receptor as a biological relevant target.

NMDA Receptor Opening and Closing—Transitions of a Molecular Machine Revealed by Molecular Dynamics

“…As expected, the largest difference could be seen for GluN2B LBD in response to the glutamate binding and unbinding. Conformational changes indicating rotational motions of LBD were observed when all-atom explicit solvation MD simulation was used, however, despite 300 ns the timescale was not sufficient to observe complete gating event [37]. Our observation is also in agreement with the “rolling motion” described in the recent study by Esmenjaud et al [60] based on fitting homology models to experimental cryoEM density.…”

“…Molecular simulation has proven useful in describing the dynamics and energetics of ligand binding to the NMDAR [30,31,32], receptor dynamics and channel opening [29,33,34,35,36,37,38], drug action [39,40], role of glycans [41,42] and conformational changes underlying disease-associated mutations in genes encoding NMDAR subunits [43,44,45]. This work concentrated on processes associated with receptor transition from the closed to open conformation, where the so far unknown closed conformation should correspond to the receptor not just when the ion channel is impermeable, but also structurally relaxed as a consequence of the glycine and glutamate ligands not bound nor acting on the receptor.…”

Binding and Dynamics Demonstrate the Destabilization of Ligand Binding for the S688Y Mutation in the NMDA Receptor GluN1 Subunit

“…However, dysfunction of the NMDAR has been associated with complex, multifactorial conditions such as schizophrenia [ 23 , 24 , 25 , 26 ] and Alzheimer’s disease [ 27 , 28 ]. There have been several studies conducted which utilize molecular dynamics simulations to study the function and effects of the NMDAR [ 29 , 30 , 31 , 32 , 33 ].…”

Three-stage multiscale modelling of the NMDA neuroreceptor

“…Therefore, the computational capability to tackle large-scale protein transitions and ion passage, has major potential to study many neurological diseases and design new drugs through medicinal chemistry methods. Some initial studies have indeed tackled the study of NMDA by a brute force MD simulations that, while showing some degree of success, have also highlighted the difficulty of covering the complete times scale of the allosteric and functional response Song et al [3], Dai and Zhou [6], Zheng et al [7], Palmai et al [8].…”