ISIS, University of Strasbourg (France)
Local Project ID:
HPC Platform used:
SuperMUC of LRZ
Ligand-gated ion channels (LGIC) play a central role in intercellular communication in the central and peripheral nervous systems as well as in non neuronal cells. Understanding their function at an atomic level of detail will be beneficial for the development of drug therapies against a range of diseases including Alzheimer's disease, schizophrenia, pain, and depression. By capitalizing on the increasing availability of high-resolution structures of both pentameric and trimeric LGICs we aim at elucidating the molecular mechanism underlying activation/deactivation by atomistic Molecular Dynamics (MD) simulations, which is essential to rationalize the design of potent allosteric modulators.
Two independent projects have been conducted using the 20 million core-hours granted on HPC system SuperMUC through the Partnership for Advanced Computing in Europe (PRACE). The first project focused on the spontaneous deactivation of the pentameric glutamate-gated chloride channel (GluCl), which led to the publication of a general mechanism of allostery in pLGICs . The second project explored ion gating in trimeric ion channels activated by extracellular ATP (P2X receptors) and demonstrated that the crystal structure of the active state is non-native. Based on the simulation analysis, new models were produced for both active and resting states of the channel, which are being tested by our experimental collaborators (T. Grutter).
Pentameric ligand-gated ion channels (pLGICs) such as GluCl mediate intercellular communication in the brain by converting a chemical signal into an ion flux through the postsynaptic membrane. The opening of the transmembrane ion channel is controlled by the binding of a neurotransmitter to the extracellular EC domain (Fig. 1, left). The molecular mechanism of gating ions has remained elusive despite the recent availability of high-resolution structures of pLGICs.
Using the compute resources of the Leibniz Supercomputing Centre in Garching, the scientists performed sub-µs atomistic molecular dynamics (MD) simulations of the eukaryotic GluCl and the prokaryotic channels ELIC and GLIC with an explicit treatment of the solvent and membrane environment. The GluCl channel was simulated with and without the allosteric agonist ivermectin (IVM), which stabilizes the open-pore (active) state. The relaxation of GluCl induced by the removal of IVM captured a sequence of spontaneous events, which elucidate the link between neurotransmitter unbinding and ion-channel cosing. A general mechanism of pLGIC deactivation has been proposed . The mechanism involves four sequential events (Fig. 1, right). A global quaternary twisting of the ion channel is the first step of the transition, which activates the structural communication between the neurotransmitter-binding site and the ion pore. Then, outward tilting of the β-sandwiches of the EC domain, which is directly controlled by neurotransmitter unbinding, allows for a reorientation or untilting of the pore-lining helices M2 to shut the ion pore. The proposed mechanism shows that the structural communication between these distant sites is mediated by the interaction of highly conserved residues at the EC/TM domains interface.
P2X receptors are trimeric ligand-gated ion channels activated by extracellular ATP (Fig 2A). They are ubiquitously distributed in humans and involved in fundamental physiological processes as diverse as synaptic transmission, response to inflammation, rescue of ischemic heart failure, and pain perception . Similarly to pLGICs, the binding of ATP to the extracellular domain is allosterically coupled to a large conformational change that opens the ion channel, thus allowing cations to flow through the transmembrane pore. Despite the obvious pharmaceutical interest, little is known about the molecular mechanism of P2X activation and there is no drug presently available on the market . The availability of high-resolution structures for both the closed and open channel states [4, 5] is however a fundamental breakthrough.
Using the SuperMUC computing resources, the scientists aimed at characterizing the functional dynamics of these receptors starting from the crystal structures of the zebrafish P2X4 by atomistic MD simulations. The researchers found that the TM domain of the active state is unstable and rapidly (< 10 ns) shuts even with bound ATP. The spontaneous closure of the ion pore in simulation, which is consistent with recent observations by others , is due to the straightening of the pore-lining helices TM2 that destabilizes the interface with TM1.
Based on the simulations results, improved models for both the active and resting states of the P2X receptor were produced. Several independent simulations were used to test the structural stability of the new models. Remarkably, the new model of the resting state, which includes a modified TM1/TM2 interface in all subunits, is stable over 1 µs of MD simulation. Finally, a 2 µs long MD simulation started with the crystal structure of the resting state shows a spontaneous rearrangement of the TM1/TM2 interface in two over three subunits (Fig 2B), which is consistent with the new model. The improved models are currently investigated by site-directed mutagenesis and electrophysiology essays.
Benefits to the community
The gating mechanism emerging from the pLGICs simulations assigns a primary role to receptor twisting, which mediates the structural communication between the orthosteric site and the ion pore. The global twisting is propsed to contribute to the allosteric coupling by “locking” the ion channel into an open-pore conformation. Further, the researchers have shed light onto the structural coupling between the M2 and M3 helices, providing an interpretation for the “locally-closed” structure of GLIC . Finally, the large reorientation of the β-sandwiches has been recently confirmed by the X-ray structure of GLIC at pH=7 which is supposed to represent a closed state . On the P2X receptors, the simulations pointed at deficiencies of the available X-ray structures and support improved models, which will hopefully open up to a molecular understanding of the P2X receptors’ function and aid on the quest for small-molecule modulators.
Benefits from running on Tier-0 facilities
All MD simulations were conducted on atomic representations of the full-length pLGIC pentamers (GluCl with and without ivermectin, ELIC, GLIC) and P2X trimers embedded in a lipid bilayer and solvated in a water box, for a total of 150,000 to 200,000 atoms each. Typical simulation rns of 50-60 ns/day could be attained using 2048 CPU cores, which reduces by one order of magnitude the cost of these simulations in terms of human time. Moreover, given the major role of quaternary twisting in the proposed model of gating in pLGICs, the available compute resources allowed for sampling the spontaneous deactivation of the GluCl ion channel over multiple independent MD simulations on the µs timescale, which was very important to test the robustness of the scientists’ conclusions. Similarly, proposing improved models of P2X that are stable over the µs timescale required amounts of sampling that would not be possible without the level of performance and scalability offered by the Tier-0 facilities.
Figure 1. (left) Topology of pLGICs as visualized by the crystal structure of GluCl. Side view of the homopentamer with the five-fold pseudosymmetry axis (gray spheres) that runs perpendicular to the membrane (yellow box). The front A and B subunits are shown in cartoon representations (light and dark gray). The endogenous agonist L-glutamate (green spheres) and the allosteric agonist ivermectin (IVM, magenta sticks) bind at the subunits interface in the EC and TM domain, respectively.
(right) Model of the allosteric mechanism for gating ions in pLGICs. The β-sandwiches in the EC domain are depicted as a gray bucket. The simulation of GluCl with IVM removed shows that the closing of the ion pore (red cross) involves four sequential events (large red numbers): (1) a global transition to a twisted conformation of the receptor; (2) the spontaneous unbinding of L-Glu from the orthosteric site; (3) the outward tilting of the β-sandwiches in the EC domain, which lifts up the β1-β2 loop at the EC/TM domains interface (pink); (4) passage of the bulky side chain of a conserved residue (P, blue) along with the untilting of the M2-M3 helices in the direction of the ion pore.
 Calimet N. et al. (2013) Proc. Natl. Acad. Sci. USA. 110(42):E3987-96.
 Khakh B.S. & North R.A (2006) Nature 442:527-532.
 Zhiyuan L. et al. (2008) Assay and drug development technologies, 6(2):277-284.
 Toshimitsu K. et al. (2009) Nature 460(7255):592-598.
 Hattori M. & Gouaux E. (2012) Nature 485(7397):207-212.
 Heymann et al. (2013) Proc. Natl. Acad. Sci. USA. 110(42):E4045-E4054.
 Prevost M.S. et al. (2012) Nature Structural & Molecular Biology 19, 642-649.
 Sauguet L. et al. (2014) Proc. Natl. Acad. Sci. USA. 111(3):966-71.
Project team and scientific contact:
Dr. Marco CECCHINI (PI), Dr. Nicolas CALIMET, Dr. Jeremy ESQUE, Nicolas MERSTORF, Siddharth MALIK, Florian BLANC (all: ISIS, University of Strasbourg)
ISIS, University of Strasbourg
Laboratoire d'Ingénierie des Fonctions Moléculaires
8, allee Gaspard Monge, 67000 Strasbourg/France
e-mail: mcecchini [@] unistra.fr
LRZ project ID: pr89te