Molecular architecture of the Gαi-bound TRPC5 ion channel

Abstract

G-protein coupled receptors (GPCRs) and ion channels serve as key molecular switches through which extracellular stimuli are transformed into intracellular effects, and it has long been postulated that ion channels are direct effector molecules of the alpha subunit of G-proteins (Gα). However, no complete structural evidence supporting the direct interaction between Gα and ion channels is available. Here, we present the cryo-electron microscopy structures of the human transient receptor potential canonical 5 (TRPC5)-Gα_i3 complexes with a 4:4 stoichiometry in lipid nanodiscs. Remarkably, Gα_i3 binds to the ankyrin repeat edge of TRPC5 ~ 50 Å away from the cell membrane. Electrophysiological analysis shows that Gα_i3 increases the sensitivity of TRPC5 to phosphatidylinositol 4,5-bisphosphate (PIP₂), thereby rendering TRPC5 more easily opened in the cell membrane, where the concentration of PIP₂ is physiologically regulated. Our results demonstrate that ion channels are one of the direct effector molecules of Gα proteins triggered by GPCR activation-providing a structural framework for unraveling the crosstalk between two major classes of transmembrane proteins: GPCRs and ion channels.

Fig. 1. Overall architecture of the TRPC5-Gα_i3 complex.

a Composite map of the TRPC5-Gα_i3 complex with four bound Gα_i3 proteins. TRPC5 is colored in sky blue and dark blue by two subunits in the opposite tetrameric assembly. Gα_i3 is colored yellow. The density for the nanodisc is shown in gray. Side view (b) and bottom view (c) of the atomic model of TRPC5-Gα_i3. d Bottom views of the cryo-EM densities of the TRPC5-Gα_i3 with various stoichiometries. All possible binding cases were observed: one, two in trans, two in cis, three, and four Gα_i3 proteins (left to right) on a single TRPC5 channel.

Fig. 2. Functional evidence for direct activation of the TRPC5 channel by Gα_i3.

a Schematic drawing of a G_i protein-coupled receptor (G_iPCR) activation coupled to TRPC5 channels. Once activated by agonizts, GPCR permits the Gα_i to exchange its nucleotide from GDP to GTP. The nucleotide exchange immediately releases Gα_i from Gαβγ. The GTP-bound Gα_i then interacts with effector molecules. Given that [Ca²⁺]_i is in the submicromolar range and that a physiological amount of phosphatidylinositol 4,5-bisphosphate (PIP₂) is provided near the channel, the binding of Gα_i directly activates the TRPC5 channels. b Fold increase in the open probability (P_o) with respect to each intracellular determinant (PIP₂, PIP₂ + Gα_i3^Q204L, or PIP₂ + denatured Gα_i3^Q204L*). The mean open probability at each intracellular condition was divided by the mean open probability at which the patch was exposed to bath solution whose calcium concentration was buffered to 500 nM ([Ca²⁺]_i = 500 nM bath solution, see Methods). c Representative current trace for excised inside/out recordings of TRPC5 channels. The intracellular side of the same patch was exposed to either PIP₂, PIP₂ + Gα_i3^Q204L, or PIP₂ + denatured Gα_i3^Q204L*. Slashes indicate 3 min of wash-out using [Ca²⁺]_I = 500 nM bath solution. d–f Current traces with an expanded time scale under corresponding conditions. g Open probability trace with respect to each intracellular condition. Colored bars above the trace indicate the corresponding intracellular conditions as in (c). h Dose-dependent action of Gα_i3 proteins. Normalized open probability (P_o/P_o^max) from six independent recordings was fitted into Hill’s equation. K_d and n represent the apparent dissociation constant and Hill coefficient, respectively. Symbols and bars represent the mean ± s.e.m. i (left) Current traces at different membrane potentials. Unitary currents (i) at 0 mV and +20 mV were indistinguishable from the baseline. (right) The i–V curve showed a doubly-rectifying shape. Slope conductances at both negative driving force and positive driving force were calculated from least-squared fit to linear equation. Symbols and bars represent the mean ± s.e.m. (n = 3).

Fig. 3. Two different conformations of the TRPC5-Gα_i3 complexes.

Atomic models of TRPC5_Class1-Gα_i3 (a) and TRPC5_Class2-Gα_i3 (b) viewed from the side. Conformational changes in the coiled-coil domain [(c) and (d)], ankyrin repeat domain (c) and the connecting helix (d) between TRPC5_Class1-Gα_i3 and TRPC5_Class2-Gα_i3. The superimposed models were viewed from the bottom (c) and the side (d). Gα_i3 subunits are omitted for clarification. ARD, ankyrin repeat domain; CCD, coiled-coil domain; CH, connecting helix. Cartoon views of CCD and ARD of TRPC5_Class1-Gα_i3 (f), TRPC5_Class2-Gα_i3 (g), and their superposition (e).

Fig. 4. Molecular determinants of Gα_i3 binding.

a Surface representation of the TRPC5-Gα_i3 complex. b Side view of an atomic model focused on the ankyrin repeat domain of TRPC5 and Gα_i3. Boxes indicate the TRPC5-Gα_i3 interfaces expanded in (c) and (d). c, d Interfaces between TRPC5 and Gα_i3 with residues involved in TRPC5-Gα_i3 binding. ARD, ankyrin repeat domain; α₂H, α₂ helix; α₃H, α₃ helix. e (Left) Summarized current amplitudes from both wild-type TRPC5_EM and mutant TRPC5_EM^IYY/AAA. When the IYY motif was changed to three consecutive alanine residues, the coexpression of Gα_i3 could not render significant activation of the channel (n = 11), whereas wild-type channels were readily activated (n = 8). Both channels were similarly activated by 100 nM extracellular (−)-Englerin A, a highly potent and specific TRPC4 or 5 channel activator. Squares and triangles represent whole-cell current at +60 mV and −60 mV, respectively. (Right) Representative current traces are also shown. f In an excised patch from cells expressing TRPC5_EM^IYY/AAA channels, exposure of Gα_i3 protein could not induce a significant increase in open probability of the channel. The response to PIP₂, however, was similar in both wild-type and mutant channels (n = 3, wild-type; n = 6, mutant). g Representative current trace from the excised inside/out patch (top) and corresponding open probability trace (bottom). h Fluorescence signals from Gα_i3^Q204L (EYFP), wild-type TRPC5 channel (TRPC5_EM, ECFP) or a mutant channel (TRPC5_EM^IYY/AAA, ECFP). EYFP was fused between Ala114 and Glu115 of the Gα_i3 protein, and ECFP was tagged on the C-termini of TRPC5 channels. i Epifluorescence images, overlay images, and FRET images from each expression pair. Colored E_EFF indicators on the right side of FRET images were set to cover from 0% (dark blue) to 100% (red) linearly. Ex excitation, Em emission. j Summary of the E_EFF from each expression pair (n = 24, wild-type; n = 15, mutant). Bars represent the mean ± s.e.m.

Fig. 5. Putative PIP₂ binding site and physiological role of Gα_i3 in TRPC5 activation.

a Representative structural snapshot of PIP₂-bound TRPC5 by MD simulation. b Zoomed-in view of a putative PIP₂ binding pocket in (a). c The RMSD of PIP₂ plotted as a function of simulation time. The distance between center of mass of PIP₂ inositol ring and center of mass of side chain of TRPC5 residues K228 (d), K232 (e), K299 (f), R512 (g), and K645 (h) plotted as a function of simulation time. i A summarized PIP₂-induced current of wild-type and mutant channels (n = 5, wild-type; n = 4, K228A; n = 4, K232A; n = 4, K299A; n = 6, R512A; n = 6, K645A). Thick and thin lines represent the mean and s.e.m. I–V curves at current peaks from excised inside/out patches were also drawn. j Normalized open probability (P_o) (black spheres and lines, mean ± s.e.m., n = 6–8 excised patches) is graphed as a function of Gα_i3 and PIP₂ concentrations. Lines were drawn from mathematical fit to the Hill equation. k Data points in (j) as a collection of curves (line projections onto the Normalized P_o-[diC8-PIP₂] plane) according to each Gα_i3 concentration. Dotted lines and numbers at corresponding vertical intersections were drawn to explain how the fold increase in P_o by Gα_i3 in (l) was obtained. l A Gα_i3 amplification curve is shown as the purple curve ([Gα_i3] = 14.6 μM) divided by the black curve ([Gα_i3] = 0.07 μM) in (k). Three circles at [PIP₂] = 5, 16, and 50 μM were calculated from the dotted lines in (k), but other data points in (k) and the curves encompassing them were deliberately omitted from the analysis to avoid the numerical singularity. In both (k) and (l), a tightly regulated physiological [PIP₂] range is shown in green shadings.

Fig. 6. Possible mechanism for activation of the TRPC5 channels by Gα_i3 and the cofactors Ca²⁺ and PIP₂.

Schematic drawing for a possible activation mechanism of TRPC5 channels by Ca²⁺, PIP₂, and Gα_i3. From electrophysiological recordings, we found that in the presence of intracellular Ca²⁺, the application of PIP₂ at physiological concentrations (4–10 μM) onto the intracellular side of the patch increased the open probability of the channel four-fold on average. The binding of Gα_i3 greatly increased the open probability approximately seven-fold with increased sensitivity of TRPC5 against PIP₂. Representative current traces from the excised patch at each condition are attached. Only two diagonally opposed subunits of TRPC5, Gα_i3, Ca²⁺ ions, and PIP₂ molecules are shown for clarity.