Gating and modulation of an inward-rectifier potassium channel

Figures

• 

  Figure 1.. Important structural features of Kir channels. (A) Diagonally opposite subunits of membrane-embedded cKir2.2 (taken from Sim. 2; t = 0). Side chains that are in the helix-bundle crossing (activation gate), involved in rectification, or are part of the lipid-binding sites, are represented as sticks. PIP2 molecules bound to the channel are represented as semi-transparent sticks. Cα atoms of the mutated G178 residue are shown as orange spheres. SF residues 144–148 are represented as sticks, with SF-bound ions individually colored. E139 and R149, which are located behind the SF and highly conserved across the Kir family, are represented as red and blue sticks, respectively. (B) Sequence alignment of key functional regions in Kir channels that are strongly rectifying (cKir2.2, hKir2.2, and hKir2.1) or weakly rectifying (hKir1.1, hKir3.2, and hKir6.2), gated by PIP2 (cKir2.2, hKir2.2, and hKir2.1), ATP (hKir6.2), pH (hKir1.1), or G-protein (hKir3.2). Red, cyan, and blue highlighted residues are all known to be crucial for gating and/or rectification.

• 

  Figure 2.. Activation gate opening and permeation. (A) Activation gate closure (WT; Sim. 2) and opening (G178R; Sim. 3) measured by the Cα–Cα diagonal distance at gate residues I177 and M181. Insets: Intracellular view of the TM pore with I177 residues indicated as spheres and the TM region in grey. (B) Movement of the inner TM helices illustrating the main activation gate conformational change with only two diagonally opposed subunits shown. (C) Pore opening viewed from the intracellular side with the same representations as in B; double-headed arrow indicates S174–I177 positional interchange. (D) Ion permeation events obtained under depolarizing voltages of different magnitude (Sims. 8–12). (E) Density profiles of ion (solid lines) and water (dashed lines) permeating across the SF (Sim. 7, black; Sim. 8, red; both at 310 mV) with a simulation snapshot of the ion-occupied SF (residues 144–148) and the rectification controller D173, located in the pore cavity, with two of the four D173 residues (de)protonated; ions are individually colored, and water molecules appear in red and white. (F) Position of the incoming ion vs. the centroid of the two SF-bound ions located above it (left), and…

• 

  Figure S1.. Pore radius profiles. (A) Pore hydration (transparent cyan color) in the closed and open pore conformations; only three protein subunits are shown (grey surface), and permeating potassium ions appear as green spheres. (B) Pore radius profiles of the PIP2-agonized crystal structure (PDB ID 3SPI), from a snapshot 2 µs into Sim. 1, and from a snapshot 5 µs into Sim. 2. Together, these profiles suggest a transition from the open, non-conducting conformation to a preactivated, semi-open pore conformation, and subsequently to a fully closed pore conformation. (C) Pore radius profile of the open-pore conformation obtained in simulations of the R178G activation gate mutant (Sim. 4) compared with the crystal structures in the presence (PDB ID 3SPI) and absence (PDB ID 3JYC) of PIP2, with preactivated, semi-open, and closed pore conformations. The pore radius profiles were generated using HOLE (Smart et al., 1996).

• 

  Figure S2.. Lipid interactions with the R/K-rich region, movement of the CTD, and open state stability. (A) Interactions between negatively charged PIP2 and positively charged R/K residues are preserved during channel activation and pore opening (Sim. 3); PIP2 atoms within 4 Å of an R/K residue were defined as interacting. (B) Per-subunit (i.e., A–D) distribution of POPS:P distances to K62:Cζ/R65:Nζ (IH), R78:Nζ/R80:Nζ (TM1), and R219:Nζ/K220:Cζ (CTD). (C) Snapshot of one TM subunit (gray) and part of the CTD of an adjacent subunit (green) highlighting PIP2 and POPS-interacting (R/K) residues. SF-bound ions are shown as colored spheres. (D) CTD-SF center of mass separation during pore opening in simulations of the activation gate G178R mutant (Sim. 3) and upon back mutation of this mutant to the WT (Sim. 4). Green dotted lines are reference distances from the apo (PDB ID 3JYC) and PIP2-bound (PDB ID 3SPI) crystal structures (Tao et al., 2009; Hansen et al., 2011). Inset: Two crystal structures and a simulation snapshot of the open state obtained in Sim. 4; dotted red line indicates the CTD movement toward the TM domains (all structures were TM-aligned). (E) Stability of the open pore…

• 

  Figure 3.. CTD charge neutralization affects permeation. (A) Structure of the full-length channel with only two diagonally opposed subunits shown, highlighting various regions within the extended pore. Ions in the SF are individually colored. (B) Ion-density profiles (each normalized to 1) with potassium and chloride shown as solid and dashed lines, respectively, across the extended pore in the WT (Sims. 7–8, black) and in the charge density-reduced E225A/E300A and E225A/R261Q double mutants (Sims. 26–27, red; Sims. 28–29, green). The average potassium ion occupancies across the CTD, pore cavity, and the SF are listed for each channel construct. (C) Total potassium ion occupancy across the extended pore. (D) Ion permeation events as a function of simulation time. (E) Ion residence time within the CTD, pore cavity, and SF during permeation in the three channel constructs, with color coding as in B.

• 

  Figure S3.. Permeation mechanism in the E225A/E300A double mutant. Position of the incoming ion vs. the centroid of the two SF-bound ions located above it (left), and position of the leaving ion vs. the centroid of the two SF-bound ions located below it (right). Color bar is in units of log(ρ[x,y]); ρ[x,y] is the two-dimensional histogram of ion positions averaged over the entire simulation time of Sim. 26 and Sim. 27. Blue minima represent the four predominant three-ion and four-ion configurations in the knock-on permeation mechanism. All observed configurations are similar to those observed during permeation across the WT (see Fig. 1 E), indicating that the predominant permeation mechanism in this mutant is unchanged from that observed in the WT.

• 

  Figure 4.. Channel deactivation in the absence of PIP2. (A) Channel deactivation and activation gate closure measured by the Cα–Cα diagonal distance at gate residues I177 and M181 (Sim. 13, top), and number of pore cavity water molecules and CTD-SF center of mass separation (Sim. 13, bottom). (B) Intracellular view of the inner helix, TM2, and IH in the presence and absence of PIP2 (blue, Sim. 4; yellow, Sim. 13), superposed with the PIP2-depleted, apo Kir2.2 crystal structure (red, PDB ID 3JYC), with an SF-bound potassium ion in magenta and the C, N, and O atoms of the SF residues shown in gray, blue, and red, respectively. (C) Closest atomic distances (per monomer) between any of the TM2 residues of the R/K-rich region (R186:Nζ, K188:Cζ, K189:Cζ, or R190: Nζ) and each of the following residues: (i) E304:Cδ of the G-loop, (ii) D76:Cγ, and (iii) D69:Cγ of the IH. The distance in each monomer is colored individually and the conformation of the channel is indicated by Open and Closed. (D) Simulation snapshots revealing proximity between TM2 residues R186, K188, K189, and R190 (gray), and PIP2 in the open state (left), and between IH residues D69 and D76 and the R/K-rich region in the absence…

• 

  Figure 5.. SPM binding and pore block. (A) Ion permeation events per unit time under depolarizing (Sims. 31–46, blue) and hyperpolarizing (Sims. 51–58, red) conditions, and in the presence of SPM. Inset: Molecular structure of SPM. (B) Ion transitions across the SF (top) and SPM position relative to the D173 center of mass (Cα atoms; bottom) at three different depolarizing voltages (Sims. 48, 39, and 32). (C) Simulation snapshots (Sim. 39) revealing sequential SPM (in cyan) binding resulting in pore block, with potassium ions individually colored. (D) Voltage-dependent SPM binding (Fig. S4 A) and SPM residence times at the pore cavity-SF region (bottom; Sims. 31–46). (E) Ion permeation events as a function of simulation time in the presence (red, Sim. 47) and absence (black, Sim. 12) of SPM at low voltage (155 mV), where SPM is only weakly bound, and mostly to the CTD.

• 

  Figure S4.. SPM-binding kinetics. (A) Voltage-dependent SPM binding times (Sims. 31, 32, 35, 36, 39–41, and 43–45), with binding defined as when the distance from SPM to the D173 center of mass (Cα atoms) was 10 Å or less (30-ns running medians). (B) SPM binding (at 310 mV) to the CTD in the E225A/E300A CTD double mutant (Sims. 91 and 92), in WT Kir2.2 (Sims. 31 and 32), and in ROMK (Sims. 153 and 154). Binding was defined as when the distance from SPM to the E225/E300 center of mass (Cα atoms) was 10 Å or less (30-ns running medians). (C) SPM snapshots (cyan) during binding (Sim. 35) to the extended pore spanning the entire protein (gray cut-through surface). SPM center of mass number density along the (extended) pore-axis in WT with two or four rectification controller (D173) residues deprotonated, and in the E225A/E300A CTD double mutant. (D) Ion permeation events per unit time in the presence and absence of SPM (Sims. 36–46, red; Sims. 9–11, black) at different depolarizing voltages. (E) Voltage-dependent SPM binding (Fig. S4 A) and SPM residence times at the pore cavity-SF region with two and with four D173 residues deprotonated (Sims. 31–42, 150–152).

• 

  Figure 6.. SPM unbinding under hyperpolarizing conditions. (A) Ion transitions across SF (top) and SPM position relative to the D173 center of mass (Cα atoms; bottom) at three different hyperpolarizing voltages (Sims. 63, 69, and 90). (B) Simulation snapshots (Sim. 69) revealing sequential SPM (cyan) unbinding, with the individual residues represented as in Fig. 5 C. (C) Voltage-dependent SPM unbinding times (Sims. 61–90), with unbinding identified when the distance between SPM and the D173 center of mass (Cα atoms) exceeded 20 Å (30-ns running medians). (D) Current vs. hyperpolarizing voltage with SPM present in bulk solution (Sims. 51, 52, 55–59; black), and after SPM unbinding from the pore cavity-SF region (Sims. 61–81, 83–90; red). (E) SPM-unbinding times in WT (Sims. 61–81, 83–90; black) and in the negative charge density-reduced double mutant E225A/E300A (Sims. 93–120, red) and in E225A/R261Q (Sims. 121–148, green).

• 

  Figure S5.. Impact of the CTD charge density on SPM unbinding. Fraction of simulations relieved from SPM pore block at different voltages observed in simulations of the WT (Sims. 61–81, 83–90) and in the E225A/E300A and E225A/R261Q CTD double mutants (Sims. 93–120 and Sims. 121–148, respectively). SPM unbinding times were identified when the distance from SPM to the D173 center of mass (Cα atoms) exceeded 20 Å (30 ns-running medians).

• 

  Figure 7.. Model of lipid-dependent activation mechanism and SPM pore block. Closed conformation: In the absence of negatively charged lipids, positively charged R/K residues near the activation gate engage in salt bridge interactions with negatively charged IH residues and TM1 backbone residues, thereby stabilizing the closed, deactivated channel conformation. Open conformation: In the presence of negatively charged PIP2 and POPS lipid molecules in the membrane, three key sequential structural changes lead to pore opening and channel activation: (1) the R/K-rich region loses interaction with negatively charged IH residues in exchange for interactions with PIP2 molecules at the primary lipid-binding site, which leads to disengagement of the IH from the pore domain, allowing the IH to move radially away from the pore axis; (2) the CTD then moves towards the pore domain and CTD residues R219 and K220 of the secondary lipid-binding site interact with POPS lipid molecules, permitting G-loop residue E304 to interact with the R/K-rich region; and (3) slight rotation and splaying of inner TM helices occurs, leading to pore opening. Blocked conformation: Under depolarizing conditions (V > 0), SPM…