Conduction through a narrow inward-rectifier K+ channel pore

Abstract

Inwardly rectifying potassium (Kir) channels play a key role in controlling membrane potentials in excitable and unexcitable cells, thereby regulating a plethora of physiological processes. G-protein-gated Kir channels control heart rate and neuronal excitability via small hyperpolarizing outward K⁺ currents near the resting membrane potential. Despite recent breakthroughs in x-ray crystallography and cryo-EM, the gating and conduction mechanisms of these channels are poorly understood. MD simulations have provided unprecedented details concerning the gating and conduction mechanisms of voltage-gated K⁺ and Na⁺ channels. Here, we use multi-microsecond-timescale MD simulations based on the crystal structures of GIRK2 (Kir3.2) bound to phosphatidylinositol-4,5-bisphosphate to provide detailed insights into the channel's gating dynamics, including insights into the behavior of the G-loop gate. The simulations also elucidate the elementary steps that underlie the movement of K⁺ ions through an inward-rectifier K⁺ channel under an applied electric field. Our simulations suggest that K⁺ permeation might occur via direct knock-on, similar to the mechanism recently shown for K_v channels.

Figure 1.

Overview of GIRK2 structure. (A) Two opposing subunits of the GIRK2 structure (PDB accession no. 3SYA) are shown; subunits in the front and back are hidden for clarity. The approximated membrane boundaries between the TMD and CTD are indicated as gray bars. (B) PIP₂-binding site details: carbon atoms of the short-chain PIP₂ are colored cyan, dashed lines indicate interactions with different basic amino acids; the F192 side chain, forming the narrowest part of the HBC gate, is also shown. (C) Details of the selectivity filter region with classical K⁺ ion–binding sites (S0–S4) shown as sticks; “bowstring” interactions between E150 and R160 are indicated with dashed lines. (D) HBC and G-loop gate regions formed by F192 and G318/M319 residues are shown as sticks, respectively. (E) Simulation box after equilibration: protein embedded in POPC membrane (cyan lines); water is shown in red (represented as lines); K⁺ ions (purple) and PIP₂ (cyan lipid tail) are indicated as spheres.

Figure 2.

Solvation and spontaneous K⁺ flux through the HBC gate. Side view of the HBC gate region, with water molecules shown as sticks and a K⁺ ion (pink sphere) forming cation-π interactions. (A) Starting state of the simulation after equilibration of the simulation system. (B) Ion permeating the wetted HBC gate. (C) End state after 0.2 µs.

Figure 3.

Wetting of the G-loop gate and inner cavity. (A) Left: Two subunits of the gate region in the crystal structure (PDB accession no. 3SYA); middle: water occupancy along the pore over simulation time; every blue dot represents a water molecule; for orientation, the centers of mass of the pore-facing hydrophobic residues M319, F192, and V188 are plotted in black. Right: Average occupancy (per 1 Å) of water molecules along the pore axis during the same 0.2-µs run (200ns_run8). (B) Example run (1µs_run1) with applied electric field; the left side shows the conformation of the gate region after 1 µs.

Figure 4.

Comparison of HBC gate distances of PIP₂-bound and PIP₂-depleted control simulations. Combined F192 distance data of control simulations with (10 × 0.2 µs, orange color) or without (10 × 0.2 µs, black color) PIP₂ bound. Table S1 gives an overview of all control simulations. (A) Distances between opposing Cα atoms of the HBC-gate–forming residue F192. The distance observed in the crystal structure (15.3 Å) is indicated by a red cross. (B) Minimum distances between opposing F192 residues. In both panels, only the narrower subunit pair was considered at each time step of all control runs.

Figure 5.

G-loop gate conformations and minimum distances. (A) Representative snapshots of the G-loop area at different time steps showing M319 and G318 as spheres and water molecules within 4 Å as sticks; a K⁺ ion is shown as a pink sphere. (B) Histograms with minimum distances between opposing subunits at the HBC gate (upper plot) and the G-loop gate (lower plot); residues A316 to T320 were included for analysis of the G-loop gate. The G-loop gate was wider than the average pore diameter during ion permeation (6.4 Å; see Table 2) in 55.4% of all simulation steps, as indicated by the hatched area. 11-µs data were included (unrestrained runs) for both plots.

Figure 6.

Conduction mechanism in PIP₂–GIRK2. (A) Left: Cartoon representation of a single Kir3.2 subunit, with key residues shown as sticks. Middle: K⁺ ion movement along the pore axis (z axis) as a function of time, represented by differently colored lines. Right: Ion occupancy, calculated every 0.5 Å along the pore axis. (B) Representative snapshots of an ion traversing the SF of Kir3.2 with the permeating ion colored in dark purple. (C) Cumulative ion flux over simulation time; 12 runs with different simulation conditions included (290 or 580 mV, restrained or unrestrained G-loop); each run is colored differently. (D) PMF^EB derived from the K⁺ occupancy along the pore (unrestrained runs), with black error bars (SD).

Figure 7.

Conformational analysis of the SF region. (A) Distances between opposing Cα atoms of SF-forming amino acids (combined data of all 14 × 1-µs runs). (B) Left plot: Comparison of distance data of opposing Y157 Cα atoms between clusters of MD runs with high (magenta, 1µs_run1 + 1µs_run5 [40 mV nm⁻¹]; orange, 1µs_run12 + 1µs_run13 [20 mV nm⁻¹]) or low (green, 1µs_run2 + 1µs_run4 [40 mV nm⁻¹]; blue, 1µs_run11 + 1µs_run14 [20 mV nm⁻¹]) conductance. Right plot: To analyze temporary conductance fluctuations, snippets of phases with high (orange) or low (blue) conductance of the MD runs 1µs_run5 (0.2–0.5 µs = high; 0.7–1 µs = low) and 1µs_run8 (0–0.4 µs = high; 0.6–1 µs = low) are clustered. See also Fig. S9. (C) Distances between opposing Cα atoms of residues at the HBC gate (F192 and V193). The runs were clustered and colored the same way as in the left plot of B. (D) Analysis of hydrogen bonds in the SF area (combined data of all 14 × 1-µs runs). (E) Visualization of the hydrogen bonds in the SF area. The front subunit is hidden for better visibility. (F) Psi angles of residue Y157 as a function of time for two MD runs with phases of high and low conductance. Each plot shows the psi angles of all four subunits (colored in black, blue, green, and orange).