Possible Interactions of Extracellular Loop IVP2-S6 With Voltage-Sensing Domain III in Cardiac Sodium Channel

Figures

• 

  FIGURE 1. In cryo-EM structures of six eukaryotic sodium channels, extracellular loop IVP2-S6 hangs over VSD-III. (A, B). Intra-membrane (A) and extracellular (B) views of the cryo-EM structures superimposed by minimizing the root mean square deviations of Cα atoms in the P1 helices from the crystal structure of the channel Kv1.2-Kv2.1. Note a conserved geometry of the P1 and P2 helices and rather conserved geometry of the transmembrane helices. In all the structures, loop IVP2-S6 hangs over VSD-III implying the functional significance of interdomain interactions. (B1). Disulfide bonds that fasten loop IVP2-S6. (B2). Ionizable residues in loop IVP2-S6. (B3). Acidic residues in loop IIIS3-S4. (C). Sequence alignment of loops IIIS1-S2, IIIS3-S4, and IVP2-S6. Residues, which are homologous to E1295 and R1739 in the hNav1.5 channel, are boxed.

• 

  FIGURE 2. Extracellular loops in two cryo-EM structures of channel hNav1.5. (A, B). Extracellular (A) and side (B) views of hNav1.5 in the apo-form (green; PDB ID: 7dtc) and in complex with quinidine (gray, PDB ID: 6lqa). Note different conformations of loops IIIS3-S4 and IVP2-D6 and very different distances between E1295 and R1739. (C). Extracellular view of the outer pore. Note different conformations of the outer carboxylates (EEDD ring). (D) Cryo-EM structures of the rNav1.5 channel in the apo form (green; PDB ID: 6uz3) and in complex with the deathstalker scorpion toxin (red; PDB ID: 7k18). The toxin shifts down IVS4; the distance between Cβ atoms of arginine R3 in the two states is 11.9 Å. Loop IVS3-S4 moves much closer to loop IS5-P1 so that the distance between Cβ atoms of K1614 and N328 decreases from 24.7 to 8.0 Å (hNav1.5 numbering). At the same time, helices IVS1, IVS2, and IVS3 undergo only small shifts upon binding of the toxin: on average, Cα atoms of these helices are shifted by ∼ 1 Å (not shown). In the MC-minimized model with rigid backbones, K1614 and N328 are H-bonded (inlet).

• 

  FIGURE 3. Interface between VSD-III and IVP2-S6 is a hotspot of mutations associated with Brugada syndrome. (A, B) Extracellular and side views of the cryo-EM structure of quinidine-bound channel hNav1.5 (6lqa). Enlargements of panel (B) show residues whose disease variants are reported in ClinVar. Variants associated with BrS and LQTS are indicated by red and green letters, respectively. (C, D) Side and extracellular views of the cryo-EM structure of hNav1.5 in the apo-state (7dtc).

• 

  FIGURE 4. Mutations have a small impact on the voltage dependence of sodium current density. (See Supplementary Table S1 for numerical data.) (A) Current density of mutant E1295R insignificantly increased vs. the WT channel. (B, C) Current density of mutant R1739E and double mutants E1295R/R1739E is similar to that of WT channel. (D) Current density of mutant N1736R insignificantly increased vs. the WT channel. (E) Examples of current traces for the WT and mutant channels.

• 

  FIGURE 5. Voltage dependence of activation and steady-state inactivation. The least-square fits to the Boltzmann function for the WT channel (black) and mutant channels (red) are shown. (A–D) Voltage dependence of steady-state activation. (A) Substitution E1295R impaired the activation. (B) Mutation R1739E did not affect the activation. (C,D) Double mutation E1295R/R1739E (C) and mutation N1736R (D) insignificantly changed the activation. (E–H) Steady-state inactivation. (E) Mutation E1295R impaired the inactivation. (F) Mutation R1739E slightly facilitated the inactivation. (G,H) Double mutation E1295R/R1739E (G) and mutation N1736R (H) insignificantly impaired the inactivation.

• 

  FIGURE 6. Voltage dependence of the steady-state fast and slow inactivation. (A–D) Voltage dependence of the steady-state fast inactivation. Shown are least-square fits to the Boltzmann function. (A) Mutation E1295R impaired the inactivation. (B) Mutation R1739E slightly enhanced the inactivation. (C) The voltage dependence of inactivation in the double mutant E1295R/R1739E is similar to that in the WT channel. (D) Mutations N1736R caused a positive shift of the inactivation curve. (E–H) Voltage dependence of the steady-state slow inactivation. Comparison was performed using INa at 20 mV. Characteristics of the WT channel: INa/Imax = 0.32 ± 0.02, N = 11. (E) Mutation E1295R slightly enhanced the steady-state slow inactivation (INa/Imax = 0.23 ± 0.02; p = 0.0074, N = 12). (F) Mutation R1739E did not significantly affect the steady-state slow inactivation (INa/Imax = 0.37 ± 0.02; p = 0.15, N = 8). (G) Double mutation E1295R/R1739E slightly enhanced the steady-state slow inactivation (INa/Imax = 0.28 ± 0.03; p = 0.36, N = 11). (H) Mutation N1736R slightly retarded steady-state slow inactivation (INa/Imax = 0.18 ± 0.05; p = 0.037, N = 7).

• 

  FIGURE 7. In silico deactivation of VSD-III. (A) Intra-membrane view of the channel. (B) Enlarged view of the constraint-biased salt bridge R1739-E1295. E1295 in linker IIIS3-S4 moves significantly with IIIS4 and R1739 follows it. (C) Interaction energy of IIIS4 with the rest of the channel. (D) Interaction energy (Etotal) and its components between loop IVP2-S6 (residues 1722–1743) and VSD-III. The energy of flanking residues in the loop stems is not counted.

• 

  FIGURE 8. In silico deactivation of VSD-III. (A) The downshift of helix IIIS4 significantly moves loop IIIS3-S4 and loop IVP2-S6 due to the constraint-biased salt bridge R1295-R1739. Helices IVP2 and IVS6 at the N- and C-termini of the loop are also disturbed, although less significantly than the loop per se. (B) The downshift of helix IIIS4 moves helix IVS5. As IIIS4 slides over IVS5, five contacts between IIIS4 and IVS5 are switched, but due to the hydrophobic nature of residues in the interface, the interdomain friction would be small. (C) The downshift of the N-terminal part of linker-helix IIIS4-S5 moves the inter-repeat linker III-IV and affects contacts of hydrophobic residues in helix IIIS4-S5 with the fast inactivation tripeptide IFM. (D) Motions of IVS5, IVS6, and IIIS3-S4 affect their contacts with the IFM fast inactivation tripeptide. (E) Motions of IVS5, IVS6, and IIIS3-S4 affect contacts between these helices. (F). Side chains of the outer carboxylates (EEDD) shift more than those of the selectivity filter DEKA ring.

• 

  FIGURE 9. Extracellular loop IVP2-S6 hangs over VSD-III. (A, B) Orthogonal views of VSD-III and PD-IV. Loop IVP2-S6 (yellow with blue R1739 and cyan N1736) approaches linkers IIIS1-S2 and IIIS3-S4 but does not establish contacts with E1295 (red). (C) Extracellular view at the map of electrostatic potential plotted on the solvent-accessible surface of VSD-III in the cryo-EM structure (PDB ID: 6lqa). The positive potential (blue) at the extracellular surface of VSD-III is due to basic residues in IIIS4. (D) Side view at the map of electrostatic potential plotted at the solvent-accessible surface of PD-IV. Note a positively charged R1739 on otherwise negatively charged loop IVP2-S6. (E) Extracellular view at the map of electrostatic potential plotted on the solvent-accessible surface of the in silico deactivated VSD-III. The shift of IIIS4 in the intracellular direction reduced the positive potential at the extracellular side of VSD-III, making it more attractive for R1739. (F) Side view of the electrostatic potential map plotted at the solvent-accessible surface of the activated VSD-III bound to PD-IV. Due to the proximity of VSD-III and PD-IV, their positive and negative potentials are…

• 

  FIGURE 10. Ionizable residues and N1736 in the cryo-EM structures of hNav1.5 (A, B) and MC-minimized models of mutants (C–E). Only parts of VSD-III and PD-IV are shown for clarity. (A, B) Cryo-EM structures of the channel hNav1.5 mutant E1784K (PDB ID: 7dtc) and hNav1.5 with flecainide (PDB ID: 6lqa). Different conformations of polar residues (including E1295, R1739, and N1736) indicate their high flexibility. (C) Mutation E1295R decreased electrostatic attraction of R1739 to VSD-III and caused a loss-of-function phenotype. The mutation facilitated slow inactivation and retarded activation but facilitated fast inactivation (Figures 5, 6). This is consistent with the data that various mutations that decrease the electrostatic attraction between VSD-III and IVP2-S6 are associated with BrS (Figure 3). (D) Mutation R1739E, which increases the attraction of IVP2-S6 to VSD-III (note salt bridge K1-R1739), retards the slow inactivation, the effect opposite to that of mutation E1295R (Figure 6). (E). In the double mutant channel, R1739E is salt-bridged to K1, while E1295R turns in the intracellular direction. The changed contacts are likely responsible for the rescue effect of the double mutation…

• 

  FIGURE 11. Proposed pathways of allosteric signal transmission from VSD-III to the fast and slow inactivation gates. Repeats I, II, III, and IV are yellow, pink, green, and gray, respectively. (A). Extracellular view of the hNav1.5 channel cryo-EM structure (PDB ID: 6lqa), which is MC-minimized with Cα atoms pinned to the experimental positions. (B) Enlarged extracellular view of the interface between VSD-III and IVP2-S6. Salt bridge R1739-E1295 was biased by a distance constraint. This decreased the distance between NηR1739 and OεE1295 from 9.3 Å in the cryo-EM structure to 3.0 Å in the MC-minimized structure. However, not a single Cα atom deviated by more than 1 Å from the experimental position. (C). Intra-membrane view of the channel with some parts removed for clarity. Helix IVP2 and the N-terminal part of linker IVP2-S6, which are proposed to transmit allosteric signals from VSD-III to the selectivity filter gate, are cyan. Glutamate (E), lysine (K), and alanine (A) residues in the selectivity filter DEKA ring are shown by sticks. The C-terminal parts of linker IVP2-S6 and helix IVS6, which are proposed to contribute to the motion transmission from VSD-III to the fast inactivation…