Figure 1. Activation of HCN1 via multi-microsecond long MD simulations.

(A) (Top) Conformational changes of the HCN1 voltage sensor triggered by an applied electric field. Representative snapshots of one voltage sensor along the trajectory are shown. The numbers correspond to the timepoints marked on the trajectory plot below. S4 is shown in blue, S3 in gray, S1 and S2 are transparent. (Bottom) Time-dependent displacement of the gating charges along the membrane normal in a representative voltage sensor measured relative to the charge transfer center (F186). (B) Overlay of representative structures of six activated voltage sensors from two independent ~20 µs simulation runs. (C) Overall displacement of gating charges (S4 basic residues) along the electric field direction. The box plot shows the median, 25-75% (box), 1-99% (bars) of the data collected from the six voltage sensors that underwent activation. (D) A comparison of the bend angles of lower S4 sub-helix with the principal axis of the VSDs of HCN1 and TAX4. HCN1 Up shows the angle between the S4 C-terminus of the HCN1 (PDB 5U6O) and the principal axis of its VSD; HCN1 Down shows the angle between the S4 C-terminus of the simulated structures at the end of the run and the principal axis of its VSD; TAX4 shows the angle between the S4 C-terminus of the TAX4 channel (PDB 5H3O, Li et al., 2017) and the principal axis of its VSD. The box plot shows the median, 25-75% (box), 1-99% (bars) of the data collected from the six voltage sensors that underwent activation. (E) Comparison of the VSDs from the HCN1 activated and resting states, the Kv1.2/2.1 activated and resting state extracted from long timescale simulations (Jensen et al., 2012), and the TAX4 open structure (PDB 5H3O, Li et al., 2017). Small cyan arrows show the displacement of the C$\alpha$ atom of R3 along the applied electric field vector. (F) Per-residue gating charge computed per HCN1 subunit. Positive residues are shown in blue, negative in red and non-charged in orange. Green lines show the cumulative gating charge (thick) and the standard error (thin). (G) Coupling function corresponding to the resting (cyan) and activated (red) states. The dashed lines depict the boundaries of the transmembrane part of the voltage sensor. (H) Hydration of the HCN1 voltage sensor in the resting (left) and activated (right) states. The shaded regions show 25-75 (dark) and 10-90 (light) percentiles of the water molecule number collected for the six voltage sensors that underwent activation.

Figure 1—figure supplement 1. Design and model of an HCN1 mutant with faster activation kinetics.

(A) Consensus sequences from multiple sequence alignments for the voltage sensor of the EAG and HCN families shown as sequence logos (Crooks et al., 2004). The height of each residue is proportional to its frequency, while the height of the overall stack of residues is inversely proportional to Shannon entropy. Residues facing the interior of the voltage sensor are highlighted with gray rectangles. Candidate residues for mutations were chosen according to two criteria: they were hydrophobic and facing the VSD hydrated lumen in HCN1, and substitution to hydrophilic residues were tolerated in the EAG family. The mutated residues are marked with stars. (B) Z-position (along the membrane normal) of R270 relative to the charge transfer center residue (F186) along simulation time in the HCN1 wild type (cyan) and mutant (red). Each subunit is shown as a separate trace. (C) Cartoon representation of the voltage sensor with the two mutated residues (red). R270 and the hydrophobic plug (F186) used to monitor activation kinetics (see panel B) are shown in green. (D) Top: Example traces of the M153T/I160V mutant (right) compared to wild type HCN1 (left). Black traces represent current responses to depolarizing pulses whereas are red ones depict current responses elicited by hyperpolarizing potential pulses. Test pulses range from −150 mV to 50 mV from a holding potential of −10 mV. Scale bars show 2 µA (vertical) and 200 ms (horizontal). Bottom: Relative open probability vs. voltage relationships for the wild type (blue line) and mutant channel (red) with error bars (hidden within the symbols) showing standard deviation from three independent measurements.

Figure 1—figure supplement 2. Activation of the six voltage sensor domains observed in two independent MD simulations.

(A) Z-position of gating charges (K261, R267, R270, R273, R276) and their negative counterparts (D183, D189, D233, D255) with respect to the hydrophobic plug (F186) along simulations time. (B) Representative conformations of the activated voltage sensor domains highlighting the localization of the residues presented in panel A.