The voltage-gated sodium channel pore exhibits conformational flexibility during slow inactivation

Abstract

Slow inactivation in voltage-gated sodium channels (Na_Vs) directly regulates the excitability of neurons, cardiac myocytes, and skeletal muscles. Although Na_V slow inactivation appears to be conserved across phylogenies-from bacteria to humans-the structural basis for this mechanism remains unclear. Here, using site-directed labeling and EPR spectroscopic measurements of membrane-reconstituted prokaryotic Na_V homologues, we characterize the conformational dynamics of the selectivity filter region in the conductive and slow-inactivated states to determine the molecular events underlying Na_V gating. Our findings reveal profound conformational flexibility of the pore in the slow-inactivated state. We find that the P1 and P2 pore helices undergo opposing movements with respect to the pore axis. These movements result in changes in volume of both the central and intersubunit cavities, which form pathways for lipophilic drugs that modulate slow inactivation. Our findings therefore provide novel insight into the molecular basis for state-dependent effects of lipophilic drugs on channel function.

Figure 1.

Na_VSp1-P pore-domain adopts a conductive conformation. (A) A schematic representation of the full-length Na_VSp1 with the VSD and the PD marked. Only two subunits are shown for clarity. An exemplar recording of Na_VSp1 whole-cell Na⁺ currents. The activation protocol consisted of 500-ms depolarization from −80 to +40 mV in 5-mV increments from a holding potential of −120 mV with a sweep-to-sweep interval of 4 s. The inset shows a typical response to 0-mV depolarization pulse, where the steady-state open probability is less than 5% of the peak value (left). Na_VSp1 currents elicited by pulse-protocol (shown above). At the end of a 0-mV prepulse, the test pulse to 10 mV evokes minimal inward current, indicating that the channels are inactivated at the end of the prepulse (right). (B) Single-channel recordings of Na_VSp1-P reconstituted in liposomes (a schematic of Na_VSp1-P shown above) show functionally active channels that stochastically open and close. The trace within the gray box is shown in an expanded scale. C and O denote the closed and open states. The dotted line marks the zero-current level. (C) Two diagonal subunits of closed-gate Na_VAb crystal structure (PDB ID 3RVY) with the position at the bundle crossing (I215) shown as black spheres. The boxed regions highlight the activation gate and the selectivity filter region in the channel. Spin-normalized CW-EPR spectra of Na_VSp1-P I215C in asolectin membrane (bottom). Background-corrected Q-band DEER echo intensity is plotted against evolution time for I215R1 in DM (middle panel) and fit using model-free Tikhonov regularization. The Tikhonov L-curve is shown in the inset. The corresponding interspin distance distribution with regularization parameter (α = 100) is shown in the right panel, with two distributions corresponding to the distances between the adjacent and nonadjacent subunits. For the position I215C, interspin distance distributions were simulated with Na_V structures (PDB ID 3RVY for closed, orange; and PDB ID 3ZJZ for open, green) based on analysis of spin label rotamers using the MMM software package (Polyhach et al., 2007, 2011). Rotamer library calculations were conducted at 83 K.

Figure 2.

Characterization of individual cysteine mutants. (A) Elution profiles from size exclusion chromatography for WT and representative spin-labeled cysteine mutants before reconstitution. (B) Whole-cell current traces for representative Na_VSp1 cysteine mutants recorded from HEK-293 cells in response to an activation protocol consisting of 500-ms depolarizations from −80 to +100 mV in 5-mV increments from a holding potential of −120 mV with a sweep-to-sweep interval of 4 s. The bottom panel is the voltage dependence of activation for WT and cysteine mutants. The V_1/2 of activation is WT (28.9 ± 0.5 mV, n = 3); P156C (26.1 ± 0.5 mV, n = 3); D157C (25.5 ± 0.7 mV, n = 3); F170C (21.2 ± 1.1 mV, n = 4); V186C (29.0 ± 0.4 mV, n = 4); and V189C (29.0 ± 0.4 mV, n = 3). The vertical lines show SD from “n” individual measurements.

Figure 3.

Conformational differences in the selectivity filter region of Na_VSp1 and Na_VSp1-P. Spin-normalized CW-EPR spectra for positions along the turret loop, P1 helix, and P2 helix in Na_VSp1 (red) and Na_VSp1-P (black). These regions are color-coded on the Na_VAb structure (PDB ID 3RVY). Dotted lines marked as “i” and “m” represent the immobile and mobile components of the spectra, respectively. The two components may arise from two different rotameric orientations of the spin labels and/or from two conformational states of the protein in equilibrium. Arrows highlight positions with the most prominent differences in line shapes in the two channels.

Figure 4.

A close-up view of the intersubunit cavity in the Na_VAb structure. Residues that show a prominent increase in mobility in Na_VSp1 in comparison to Na_VSp1-P are shown in red, and those that show a decrease in this parameter are represented in blue. The residues with higher mobility values are found to line the intersubunit cavity, which is at the interface of P1 and P2 helices from two adjacent subunits.

Figure 5.

EPR environmental parameters reveal conformational differences in the membrane-embedded Na_VSp1 and Na_VSp1-P. (A) A plot of residue environmental parameters for Na_VSp1-P (shown in gray) and Na_VSp1 (shown in color). ΔH_o⁻¹ parameter (top); O₂ accessibility, ΠO₂ (middle); water accessibility, ΠNiEDDA (bottom). The EPR parameters for continuous positions are linked by solid lines. Residues that were not studied (because of sub-optimal expression/oligomeric stability) appear as gaps. Positions revealing prominent changes are highlighted within blue circles. (B) Difference in ΠO₂ values between the Na_VSp1 and Na_VSp1-P are mapped on the Na_VAb structure (PDB ID 3RVY) and color-coded, red denoting an increase and blue representing a decrease in the environmental parameter. (C) A plot of the average lipid accessibility (ΠO₂) as a function of the average mobility parameter ΔH_o⁻¹ for the turret loop (yellow), P1 helix (green), and P2 helix (magenta). The light gray boxes represent Na_VSp1-P, and the black boxes denote Na_VSp1. The dotted lines connecting the light and black boxes for each region signify the magnitude of difference between the EPR environmental parameters in the two channels.

Figure 6.

Extent of conformational change measured by DEER spectroscopy. Na_VAb structure showing the positions investigated by DEER (PDB ID 4EKW). Fit to the background subtracted DEER echo intensity is plotted against evolution time. The corresponding interspin distance distribution (right) for the Na_VSp1-P (black) and Na_VSp1 (red) for different spin-labeled positions. The two expected distances (C_β–C_β) from the adjacent and nonadjacent subunits in Na_VAb crystal structure are shown in black dotted lines. The shorter distance is from the adjacent subunits and the longer from the diagonal subunits. The dotted gray lines are distance distributions from rotamer simulation for respective positions on PDB ID 4EKW (potentially slow-inactivated Na_VAb conformation) using MMM. The arrows highlight the direction of change. Asterisk denotes distance peaks that may arise from intermolecular spin interactions between individual tetramers.

Figure 7.

Spectral changes at the SF region underlying transitions between the putative resting/conductive and slow-inactivated states. CW spectra for Na_VSp1 at representative positions in the P2 helix (positions marked on the Na_VAb structure) reconstituted in asolectin (red) and DOTAP (cyan) and compared with Na_VSp1-P reconstituted in asolectin (black). A similarity of Na_VSp1-DOTAP and Na_VSp1-P–asolectin spectra is highlighted in the gray box.

Figure 8.

Electrophysiological characterization of slow inactivation in a mutant channel Na_VSp1-170C. (A) Use-dependent development of slow inactivation: depolarizations from a holding potential of −120 mV to 20 mV were applied at 0.2 Hz (left, for 200 ms in duration) and 1 Hz (right, for 150 ms in duration) frequencies, and the normalized peak current for each pulse was plotted as a function of the pulse number for Na_VSp1-WT (gray squares) and Na_VSp1-F170C (blue squares). The error bars denote SDs from n = 4 for Na_VSp1-WT and n = 5 for Na_VSp1-F170C. (B) Voltage dependence of inactivation: peak currents were measured during test pulses to 20 mV after a 2-s inactivating pulse to the indicated potentials. Values from individual experiments were normalized to the maximum test pulse currents. Inactivation curves were fit with a Boltzmann equation, 1/1+exp[(V−V_h)/k_h], where V_h is the half-inactivation voltage and k_h is the slope factor. For Na_VSp1-WT, mean V_h was −30.3 ± 1.7 mV (n = 4); for Na_VSp1-F170C (n = 4), V_h was −35.7 ± 1.3 mV. (C) Time constants of inactivation versus voltage curves for Na_VSp1-WT (gray squares, n = 4) and Na_VSp1-F170C (blue squares, n = 4) obtained by fitting a single exponential to the current decay during depolarization to shown potentials from a holding potential of −120 mV.

Figure 9.

An EPR-based model for the mechanism of slow inactivation at the SF region. A cartoon representation of putative changes in the P1 and P2 helices during slow inactivation. These positional differences are likely to change intersubunit cavity volumes and hence, the accessibility of lipids and lipophilic molecules into these vestibules. The region in the inset is shown from the top (extracellular end).