Chemical tuning reveals a cation–π gating bridge between the voltage-sensor and pore domains in the K v 7.1 potassium channel

Figures

• 

  Fig. 1.. Assessing the role of side-chain aromaticity in voltage gating of Kv7.1 cardiac potassium channels coexpressed with the E1 subunit. (A) Structural depiction of local aromatic residues of interest (6uzz, PDB) (8). (B) G/V curves of WT Kv7.1 (black) and penta-F-Phe mutant (green) of other Phe residues in the channel. None of the tested mutants show an effect like the one at the F232 site, indicating that the F232 fluoroPhe effect is unique to the site. Of particular note is that F279 penta-F-Phe shows no change in the G/V compared to WT, indicating no electrostatic effect at that site. This is important because it shows that despite the proposed steric interaction with F232, F279 does not participate in a π-stacking interaction with F232. (C) Western blot of Q1 F232 TAG + E1 and D6 synthetase, grown in HEK cells. There is no expression of Q1 channel in the absence of supplemented fluoroPhe unnatural amino acid to the cell culture media. (D) G/V curves of Kv7.1 channels with F232 fluoroPhe mutations. There is a significant, step-wise left-shift of the G/V curve as serial F232 fluorination increases from 2 to 5 fluorine atoms on the F232 ring, strongly suggesting a π-interaction is…

• 

  Fig. 2.. Identifying basic residues in putative interactions with F232. (A) Location of each cationic residue tested via charge neutralization (6uzz, PDB). (B) Lysine and arginine amino acids within a 30 Å radius of F232 were individually mutated to a neutral Gln and the ΔG was calculated from the Boltzmann fit of the G/V curve of each mutant. ΔΔG was calculated as ΔGmut – ΔGWT = ΔΔG. Cations that showed ΔΔG more similar to F232 penta-F-Phe than to WT were further investigated. R231Q and R237Q are not shown as a G/V fit could not be found. (C) Combined effect of charge neutralization on the background of F232 penta-F-Phe. ΔΔG values were calculated as ΔGR/K→Q, F232 penta-F-Phe - ΔGR/K→Q = ΔΔG. A cluster of five cations was identified that showed greatly reduced effect of F232 fluorination compared to WT: R174, K196, K218, R228, and K285 (Table 2). (D) Representative ionic currents for mutants of interest. (E) G/V curves of the R/K → Q mutants of R174, K196, K218, and R228. R174Q shows a very different phenotype than either the WT Q1/E1 or F232 penta-F-Phe Q1/E1 channel, and R228Q shows a visible G/V shift upon F232 fluorination that precludes channel closing and accurate G/V curve…

• 

  Fig. 3.. Roles of the E1 or E3 subunits in F232 fluorination effects on channel gating. (A) Representative traces of KCNQ1 (Q1) with F232 WT, tri2,3,6-F-Phe, or penta-F-Phe. The top row shows Q1 channel with no KCNE, bottom row shows Q1 +KCNE3 (E3). Mild phenotypic changes can be seen with progressive fluorination, but the changes are subtle. (B) G/V curves of Q1 alone (no coexpressed) KCNE and Q1 + E3 with progressive fluorination of F232. The Q1 channel alone shows mild left-shift upon penta-F-Phe F232 fluorination (P = 0.014), and the stepwise trend upon serial fluorination is lost. By contrast, the Q1 + E3 channel shows a step-wise rightward shift, in a reversal of the trend shown by the Q1 + E1 channel. (C) ΔΔG (kcal) linear trends of the Q1 channel coexpressed with either E1, E3, or no KCNE. E1 shows the clearest linear trend (R2 = 0.96), but notably, the E3 has reversed gating polarity yet retains a linear trend (R2 = 0.86). (D) Structure of KCNQ1 in the presence and absence of E1 shows distance between F232 and K285.

• 

  Fig. 4.. Structural analysis of Kv7.1 channels with an encoded penta-F-Phe at F232. (A) SEC elution of Kv7.1 penta-F-Phe purification. The 13.9 mL peak (labeled with red arrow) was isolated and concentrated and used for cryo-EM structural resolution. (B) Example high-resolution micrograph of Kv7.1 F232 penta-F-Phe channel. (C) Representative 2D classes of Kv7.1 F232 penta-F-Phe channel, using the Krios 300 kV TEM. (D) GSFSC resolution plot after autotightening, showing an approximately 6.1 Å resolution of the full channel. (E) 3D volume of the Kv7.1 F232 penta-F-Phe pore domain, highlighting the regions of highest helical resolution. The Q1 pore domain is visualized, but the VSD is not. (F) z-level of the Q1 pore, shown in cross section in G. (G) Estimated volumes of the structure were made at 6 Å resolution in ChimeraX and compared to the resolved volume in cross-section at the pore. The resolved volume shows better agreement with the pore open state. (H) z-level of the Q1 HC domain, shown in cross section in I. (I) Estimated volumes compared to the resolved volume HC helix, showing that the resolved helical domains are in the positions expected of an open state channel.