Fully Flexible Ligand Docking for the P2X7 Receptor Using ROSIE.

The availability of P2X7 receptor structures with allosteric antagonists bound enables us to predict specific interactions between receptor and antagonists at atomistic detail. In this chapter we outline how modern ligand docking techniques can be employed by the nonexpert to predict putative binding modes for known or hypothetical allosteric P2X7 antagonists.

Homology Modeling of P2X Receptors.

Since the X-ray structure of the zebra fish P2X4 receptor in the closed state was published in 2009 homology modeling has been used to generate structural models for P2X receptors. In this chapter, we outline how to use the MODELLER software to generate such structural models for P2X receptors whose structures have not been solved yet.

Organization of ATP-gated P2X1 receptor intracellular termini in apo and desensitized states.

The human P2X1 receptor (hP2X1R) is a trimeric ligand-gated ion channel opened by extracellular ATP. The intracellular amino and carboxyl termini play significant roles in determining the time-course and regulation of channel gating-for example, the C terminus regulates recovery from the desensitized state following agonist washout. This suggests that the intracellular regions of the channel have distinct structural features. Studies on the hP2X3R have shown that the intracellular regions associate to form a cytoplasmic cap in the open state of the channel. However, intracellular features could not be resolved in the agonist-free apo and ATP-bound desensitized structures. Here we investigate the organization of the intracellular regions of hP2X1R in the apo and ATP-bound desensitized states following expression in HEK293 cells. We couple cysteine scanning mutagenesis of residues R25-G30 and H355-R360 with the use of bi-functional cysteine reactive cross-linking compounds of different lengths (MTS-2-MTS, BMB, and BM(PEG)2), which we use as molecular calipers. If two cysteine residues come into close proximity, we predict they will be cross-linked and result in ∼66% of the receptor subunits running on a Western blot as dimers. In the control construct (C349A) that removed the free cysteine C349, and some cysteine-containing mutants, cross-linker treatment does not result in dimerization. However, we detect efficient dimerization for R25C, G30C, P358C, K359C, and R360C. This selective pattern indicates that there is structural organization to these regions in the apo and desensitized states in a native membrane environment. The existence of such precap (apo) and postcap (desensitized) organization of the intracellular domains would facilitate efficient gating of the channel.

Mapping the Allosteric Action of Antagonists A740003 and A438079 Reveals a Role for the Left Flipper in Ligand Sensitivity at P2X7 Receptors.

P2X7 receptor (P2X7R) activation requires ∼100-fold higher concentrations of ATP than other P2X receptor (P2XR) subtypes. Such high levels are found during cellular stress, and P2X7Rs consequently contribute to a range of pathophysiological conditions. We have used chimeric and mutant P2X7Rs, coupled with molecular modeling, to produce a validated model of the binding mode of the subtype-selective antagonist A438079 at an intersubunit allosteric site. Within the allosteric site large effects on antagonist action were found for point mutants of residues F88A, D92A, F95A, and F103A that were conserved or similar between sensitive/insensitive P2XR subtypes, suggesting that these side-chain interactions were not solely responsible for high-affinity antagonist binding. Antagonist sensitivity was increased with mutations that remove the bulk of side chains around the center of the binding pocket, suggesting that the dimensions of the pocket make a significant contribution to selectivity. Chimeric receptors swapping the left flipper (around the orthosteric site) reduced both ATP and antagonist sensitivity. Point mutations within this region highlighted the contribution of a P2X7R-specific aspartic acid residue (D280) that modeling suggests forms a salt bridge with the lower body region of the receptor. The D280A mutant removing this charge increased ATP potency 15-fold providing a new insight into the low ATP sensitivity of the P2X7R. The ortho- and allosteric binding sites form either side of the ß-strand Y291-E301 adjacent to the left flipper. This structural linking may explain the contribution of the left flipper to both agonist and antagonist action.

Unique residues in the ATP gated human P2X7 receptor define a novel allosteric binding pocket for the selective antagonist AZ10606120.

The P2X7 receptor (P2X7R) for ATP is a therapeutic target for pathophysiological states including inflammation, pain management and epilepsy. This is facilitated by the predicted low side effect profile as the high concentrations of ATP required to activate the receptor are usually only found following cell damage/disease and so P2X7Rs respond to a "danger" signal and are not normally active. AZ10606120 is a selective antagonist for P2X7Rs (IC50 of ~10 nM) and ineffective at the P2X1R (at 10 µM). To determine the molecular basis of selectivity we generated a series of P2X7/1R chimeras and mutants. Two regions that are unique to the P2X7R, a loop insertion (residues 73-79) and threonine residues T90 and T94, are required for high affinity antagonist action. Point mutations ruled out an orthosteric antagonist site. Mutations and molecular modelling identified an allosteric binding site that forms at the subunit interface at the apex of the receptor. Molecular dynamics simulations indicated that unique P2X7R features regulate access of AZ10606120 to the allosteric site. The characterisation of the allosteric pocket provides a new and novel target for rational P2X7R drug development.

Mechanistic insights from resolving ligand-dependent kinetics of conformational changes at ATP-gated P2X1R ion channels.

Structural studies of P2X receptors show a novel U shaped ATP orientation following binding. We used voltage clamp fluorometry (VCF) and molecular dynamics (MD) simulations to investigate agonist action. For VCF the P2X1 receptor (P2X1R) K190C mutant (adjacent to the agonist binding pocket) was labelled with the fluorophore MTS-TAMRA and changes in fluorescence on agonist treatment provided a real time measure of conformational changes. Studies with heteromeric channels incorporating a key lysine mutation (K68A) in the ATP binding site demonstrate that normally three molecules of ATP activate the receptor. The time-course of VCF responses to ATP, 2'-deoxy ATP, 3'-deoxy ATP, Ap5A and αßmeATP were agonist dependent. Comparing the properties of the deoxy forms of ATP demonstrated the importance of the 2' hydroxyl group on the ribose ring in determining agonist efficacy consistent with MD simulations showing that it forms a hydrogen bond with the γ-phosphate oxygen stabilizing the U-shaped conformation. Comparison of the recovery of fluorescence on agonist washout, with channel activation to a second agonist application for the partial agonists Ap5A and αßmeATP, showed a complex relationship between conformational change and desensitization. These results highlight that different agonists induce distinct conformational changes, kinetics and recovery from desensitization at P2X1Rs.