Mapping the Site of Action of Human P2X7 Receptor Antagonists AZ11645373, Brilliant Blue G, KN-62, Calmidazolium, and ZINC58368839 to the Intersubunit Allosteric Pocket

Abstract

The P2X7 receptor is a trimeric ligand-gated ion channel activated by ATP. It is implicated in the cellular response to trauma/disease and considered to have significant therapeutic potential. Using chimeras and point mutants we have mapped the binding site of the P2X7R-selective antagonist AZ11645373 to the known allosteric binding pocket at the interface between two subunits, in proximity to, but separated from the ATP binding site. Our structural model of AZ11645373 binding is consistent with effects of mutations on antagonist sensitivity, and the proposed binding mode explains variation in antagonist sensitivity between the human and rat P2X7 receptors. We have also determined the site of action for the P2X7R-selective antagonists ZINC58368839, brilliant blue G, KN-62, and calmidazolium. The effect of intersubunit allosteric pocket "signature mutants" F88A, T90V, D92A, F103A, and V312A on antagonist sensitivity suggests that ZINC58368839 comprises a binding mode similar to AZ11645373 and other previously characterized antagonists. For the larger antagonists, brilliant blue G, KN-62, and calmidazolium, our data imply an overlapping but distinct binding mode involving the central upper vestibule of the receptor in addition to the intersubunit allosteric pocket. Our work explains the site of action for a series of P2X7R antagonists and establishes "signature mutants" for P2X7R binding-mode characterization.

Fig. 1.

Chimeric hP2X7-1 receptors identify regions important for action of the P2X7R-selective antagonist AZ1165373. (A) Effect of the antagonist AZ11645373 (100 nM, traces indicated by black circle) on current evoked by an EC₉₀ concentration of ATP (10-second application indicated by black bar) at the P2X7-2Nβ, 112–118, and 89–94 chimeras and P2X1R. Controls are indicated by open circles. (B) Concentration-dependent inhibition by AZ11645373 of responses to an EC₉₀ concentration of ATP for P2X7-2Nβ (black), chimeras 81–88 (firebrick), 89–94 (red), 112–118 (yellow), 279–285 (blue), 295–310 (magenta), and P2X1 (gray). (C) Histogram showing the pIC₅₀ of AZ11645373 at P2X7-2Nβ and chimeric receptors. Blacked dotted lines correspond to a 3-fold change in sensitivity. Exact values for each receptor tested are given in Supplemental Table 1. *P < 0.05; **P < 0.01; ****P < 0.0001, n ≥ 3. (D) Location of chimeras that reduced AZ11645373 action mapped on a pdP2X7R-based homology model; chimeras with no change are shown as gray spheres.

Fig. 2.

Effects of point mutants in the allosteric binding pocket on sensitivity to the antagonist AZ11645373. Effects of point mutations on AZ11645373 sensitivity are reported by their pIC₅₀ value. Black dotted lines correspond to a 3-fold change in sensitivity. Pink residues are variant, green have similar properties, and those in black are conserved between P2X and P2X7Rs. Exact values for each receptor tested are given in Supplemental Table 32. *P < 0.05; ***P < 0.001; ****P < 0.0001, n ≥ 3.

Fig. 3.

Representative binding pose for AZ11645373 in the hP2X7R. (A) View from the top of the extracellular domain along the central axis perpendicular to the membrane. The P2X7R model is shown as cartoon, with the three subunits highlighted in light blue, light pink, and gray; AZ11645373 is shown as spheres. (B) As in (A) but rotated 90°. (C) Zoom into the proposed AZ11645373 binding site, one subunit [light blue in (A) and (B)] is omitted for clarity. Residues K110 (blue, increase), I310, F88, V312, D92, A91, T94, F103, F95, P96, L97 (red, decrease or no inhibition), whose mutations showed the strongest effects on AZ11645373 sensitivity, are shown as sticks (data from Fig. 2). (D) As in (C), F95, V312, and Y108 residues, whose P2X7R rat-to-human and human-to-rat mutations (data shown in Fig. 4) affected AZ11645373 sensitivity, are shown as spheres.

Fig. 4.

Introduction of AZ1165373 sensitivity by point mutation of species-variant residues in the allosteric pocket. (A) Homology model illustrating variant residues in the intersubunit allosteric pocket between hP2X7 and rP2X7R (red spheres) in the left panel, zoom and stick representation in the right panel. (B) AZ11645373 inhibition curves at human-to-rat P2X7R point mutations (red), P2X7-2Nβ (black dotted line), and rP2X7 (black). (C) AZ11645373 inhibition curves at rat to human P2X7R point mutations.

Fig. 5.

Use of intersubunit allosteric pocket “signature mutants” to investigate the site of action for brilliant blue G, KN-62, calmidazolium, ZINC58368839, and PPADS. (A) Concentration-dependent inhibition by P2X7 antagonists of response to an EC₉₀ concentration of ATP for P2X7-2Nβ (black), F88A (firebrick), T90V (orange), D92A (yellow), F103A (green), and V312A (purple). (B) Top panel: P2X7R overview with “signature mutants” residues shown as spheres; colors as in (A). Other panels: View from the top of the extracellular domains along the central axis perpendicular to the membrane. The P2X7R model is shown as cartoon with the three subunits highlighted in light blue, light pink, and gray with the docked pose of the antagonist shown as spheres. (C) Top panel: Surface representation of entrance to allosteric pocket. Other panels: Zoom into the proposed binding site, one subunit [light blue in (B)] is omitted for clarity. Residues F88, T90, D92, F103, and V312 and the respective antagonist are shown as sticks [colors as in (A) and (B)].