Structural insights into the function of P2X4: an ATP-gated cation channel of neuroendocrine cells

Abstract

The P2X4 receptor (P2X4R) is a member of a family of ATP-gated cation channels that are composed of three subunits. Each subunit has two transmembrane (TM) domains linked by a large extracellular loop and intracellularly located N- and C-termini. The receptors are expressed in excitable and non-excitable cells and have been implicated in the modulation of membrane excitability, calcium signaling, neurotransmitter and hormone release, and pain physiology. P2X4Rs activate rapidly and desensitize within the seconds of agonist application, both with the rates dependent on ATP concentrations, and deactivate rapidly and independently of ATP concentration. Disruption of conserved cysteine ectodomain residues affects ATP binding and gating. Several ectodomain residues of P2X4R were identified as critical for ATP binding, including K67, K313, and R295. Ectodomain residues also account for the allosteric regulation of P2X4R; H140 is responsible for copper binding and H286 regulates receptor functions with protons. Ivermectin sensitized receptors, amplified the current amplitude, and slowed receptor deactivation by binding in the TM region. Scanning mutagenesis of TMs revealed the helical topology of both domains, and suggested that receptor function is critically dependent on the conserved Y42 residue. In this brief article, we summarize this study and re-interpret it using a model based on crystallization of the zebrafish P2X4.1 receptor.

Fig. 1

Attaching GFP to C-terminus did not affect the expression and gating of P2X4R. a Confocal images of GFP fluorescence in transfected HEK293 cells. Distribution of GFP without P2X4 fusion (left) and P2X4-fused GFP (right). b Typical patterns of ATP-induced currents in HEK293 cells expressing P2X4R (left) and GFP-tagged P2X4R (right). The traces shown are from different cells. The numbers below the traces show mean values ± SEM for the time constants of receptor desensitization (τ_des) and peak amplitude (I). c The averaged ATP-induced Ca²⁺ responses in GT1 neurons expressing P2X4R (left) and P2X4R-GFP (right). The numbers above the traces indicate the time constants of calcium signal desensitization

Fig. 2

Characterization of P2X4R gating. a The 10–90% deactivation time values are independent of ATP concentration. Horizontal line illustrates the mean value for deactivation time. b The inverse relationship between 10 and 90% activation time and ATP concentration. c, d main panels, The inverse relationship between the values for 10% (c) and 90% (d) desensitization time and ATP concentration. Insets: the correlation between 10 and 90% activation time and 10% (c) or 90% (d) desensitization time. R, coefficient of correlation

Fig. 3

Structural model of rat P2X4R with bound ATP. The predicted ATP binding site is located between two neighboring subunits (Kawate et al. 2009). Residues predicted to be involved in ATP binding are shown as sticks. In addition, five disulfide bridges (SS1–SS5) are also shown as sticks. The three-dimensional model of rat P2X4R (sequence Arg33–Val355) was generated using the Modeller 9v7 package (Sali and Blundell 1993) and the crystal structure of zebra fish P2X4R solved at 3.5 Å resolution (PDB access code 3I5D) (Kawate et al. 2009). Missing side-chains were built using the DeepView/Swiss-PdbViewer v4.0.1 program (Guex and Peitsch 1997). The final model was validated by PROCHECK and bad contacts were corrected manually. The resulting model was energy-minimized using the DeepView/Swiss-PdbViewer with the GROMOS96 43B1 parameters set. The ATP docking started with manual docking of the ATP molecule into the predicted binding site of P2X4R (Kawate et al. 2009). Subsequently, AutoDock v4.2 (Morris et al. 2009) was used to predict the structure of the rat P2X4-ATP complex. This figure as well as Figs. 4 and 5 were generated using Pymol v0.99 (http://www.pymol.org)

Fig. 4

View of the rat P2X4R model parallel to the molecular three-fold axis from the intracellular side of the membrane. Residues that have been shown to alter the IVM effects on rat P2X4R are shown in yellow/light gray (Jelinkova et al. 2008). Only IVM-sensitive residues of one subunit are shown as sticks (Color figure online)

Fig. 5

Cartoon representation of the transmembrane domains of rat P2X4R viewed parallel to the membrane plane. Residues in the vicinity of the conserved Y42 residue of one subunit are shown as sticks