Dynamical Nonequilibrium Molecular Dynamics Simulations Reveal Atomistic Steps of C-Type Inactivation in Cardiac hERG Channels

Figures

• 

  1. Computational model and protocol used to study the hERG channel. (A) hERG channel embedded in a POPC lipid bilayer; the Voltage-Sensor Domain is shown in blue, the Pore Domain in red, and the Carboxy-terminal Domain in orange. (B) SF configurations used to start the equilibrium simulations. K+ ions are colored yellow and displayed in vdW representation, like the water molecules. The residues of the SF and the corresponding occupancy sites (S0 to S4) are reported. (C) Scheme of the procedure used to set up the D-NEMD simulations. From equilibrium simulations of the systems started with 2 K+ at 0 mV with the CHARMM36 force field, 20 independent simulations were performed, applying the removal of the K+ ions from the SF as a perturbation. For each D-NEMD replica, equivalent time points were combined to measure the Cα deviation of the protein amino acids from their equilibrium.

• 

  2. Structural response of hERG to the removal of K+ ions from the SF. (A) Cα displacement of the protein’s amino acids computed for each residue following K+ ions removal from the SF. Ensemble averages are taken for each of the shown instants over the 20 D-NEMD replicas and over the four protein subunits. (B) Contact map showing intrasubunit interactions computed from D-NEMD simulations, where a contact fraction equal to 1 corresponds to the presence of the interaction between pairs of residues, while equal to 0 corresponds to its complete absence.

• 

  3. Proposed three-step-based pathway for hERG C-type inactivation. The first step involves the breaking of the N633–W585 and S631–F617 contacts. The second step involves the rotation of V625 in the SF, which allows/leads to the constriction of the filter at the level of G626 (third step). For simplicity, only one subunit is shown.

• 

  4. Alanine mutations at W585 and N633 abolish C-type inactivation in hERG channels. Current traces for WT (A), W585A (B), and N633A (C), respectively. Inset in (A) shows the voltage protocol used; because W585A and N633A are right-shifted and display no inactivation, the related protocols were slightly different from WT. For WT, the voltage pulses reached 80 mV, and for the mutants, 120 mV. The hyperpolarizing returning pulse was −20 mV for WT and 20 mV for the mutants. (D) The peak current normalized by its maximum, plotted against the membrane voltage. The current shown in (D) was corrected for the leak. The raw normalized current without leak subtraction is shown in Figure S12D. (E) Normalized G–V curves plotted against the membrane voltage for the WT and the mutants. They were taken from the tail currents: −20 mV for WT and +20 mV for the mutants. Data are presented as mean ± SEM. WT n = 5, W585A n = 5, N633A n = 4, where n is the number of biologically independent cells. For visualization purposes, the current traces were decimated to 20 μs.