Structure and physiological function of the human KCNQ1 channel voltage sensor intermediate state

Figures

• 

  Figure 1.. NMR spectra of the human KCNQ1 voltage-sensor domain. (A) 1H-15N TROSY-HSQC spectrum recorded at 900 MHz of 2H,13C,15N-Q1-VSD in LMPG micelles. Backbone amide peaks for 140 out of 147 non-proline residues (95%) have been assigned. Only a single set of peaks is observed. (B) 1H-13C HSQC methyl optimized spectrum recorded at 900 MHz of 13C,15N-Q1-VSD in perdeuterated LMPG micelles. Methyl groups for 58 out of 68 (85%) residues were assigned (Ala 10 of10, Thr 6 of 8, Ile 9 of 11, Val 18 of 22, Leu 15 of 17). In addition to the presence of a very limited number of unassigned peaks from the VSD in this spectrum other unassigned peaks likely derive from natural abundance 13C in residually protonated LMPG and also from the fully protonated buffer components TCEP and MES. The chemical shift assignments illustrated for panels A and B shift have been deposited in BioMagResBank (BMRB ID 30517).

• 

  Figure 1—figure supplement 1.. Long range PREs are consistent with the expected KCNQ1 VSD topology in LMPG micelles. The residue positions of the spin-labeled cysteines for each sample are : (A) 144-SL (B) 155-SL (C) 214-SL (D) 224-Sl (E) 121-SL (F) 177-Sl (G) 180-SL (H) 238-SL. Figure 1—figure supplement 1—source data 1.Excel file with numerical data used for Figure 1—figure supplement 1.

• 

  Figure 1—figure supplement 2.. TALOS-N secondary structure analysis of backbone chemical shifts (top panel) and (bottom panels) deviations of the (Osteen et al., 2012) Cα and carbonyl 13C’ chemical shifts from random coil values. The chemical shift data for the KCNQ1 VSD been deposited in BioMagResBank (BMRB ID 30517). Figure 1—figure supplement 2—source data 1.Excel file with numerical data used for Figure 1—figure supplement 2.

• 

  Figure 1—figure supplement 3.. Examples of NOE measurements. Example strip plots showing long-range methyl-methyl and methyl-backbone Hα proton NOEs (red arrows) taken from the methyl-optimized 3D 13C-edited NOESY recorded on a 13C,15N-labeled sample in deuterated LMPG. The chemical shift assignments illustrated in this figure have been deposited in BioMagResBank (BMRB ID 30517).

• 

  Figure 2.. Structure of the KCNQ1 VSD. (A) The human KCNQ1 VSD NMR-determined ensemble after molecular dynamics refinement in a hydrated DMPC bilayer (PDB ID 6MIE, see statistics in Table 1 and also Figure 2—figure supplement 1) (B) Cryo-EM structure of the Xenopus KCNQ1 VSD (PDB ID 5VMS) (Sun and MacKinnon, 2017). (C) Representative low energy NMR structure from PDB 6MIE in a hydrated DMPC bilayer. In panels A-C, the transmembrane helices S1, S2, S3, and S4, are colored bluish green, yellow, vermillion, and sky blue respectively. The S2-S3 linker and S0 helices are colored reddish purple and orange. The approximate position of the membrane-water interfaces is indicated by a pair of black lines in panels A and B.

• 

  Figure 2—figure supplement 1.. KCNQ1-VSD XplorNIH-determined structural ensemble, before and after molecular dynamics refinement. (A) The ensemble is comprised of the top 15 lowest energy structures out of 150 conformations generated by NMR data-restrained XPLOR-NIH calculations (no bilayer). The transmembrane helices S1, S2, S3, and S4, are colored bluish green, yellow, vermillion, and sky blue respectively. The S2-S3 linker and S0 helices are colored orange and reddish purple. The loop between S1 (bluish green) and S2 (yellow) appears disordered due to lack of NMR-derived restraints. This does not necessarily mean this region does not adopt stable secondary structure. The approximate position of the membrane is indicated by a pair of gray lines. (B) The unrestrained molecular dynamics refined ensemble (MD carried out in an explicit hydrated DMPC bilayer) colored as in panel A (corresponding to PDB ID 6MIE containing 10 conformers). (C) The upper panel shows the total number of distance restraints per residue. For NOE-derived restraints that define a distance between two atoms in different residues, each residue receives a count of 0.5 per restraint. Long range, medium, and sequence distance restraints are…

• 

  Figure 3.. Comparison of intermediate and activated KCNQ1 VSD conformations. (A) Intermediate conformation of human KCNQ1 VSD (1st structure in the PDB 6MIE ensemble). (B) Activated conformation of the Xenopus KCNQ1 VSD (PDB 5VMS) (Sun and MacKinnon, 2017). In both panels, the Cα atoms of the S4 polar residues are shown as yellow spheres and the transmembrane helices are labeled in vermillion text. S0 and S1 are not shown to improve clarity of side chain interactions. (C) Overlay of the NMR (sky blue) and cryo-EM (bluish green) structures of the KCNQ1 VSD. All structural elements other than the S4 helix are semi-transparent. The Cα of the human residue G229 and the corresponding Xenopus residue G219 are shown as spheres. (D) Sequence alignments for S2 and S4 in the KCNQ/Kv7 family and select other voltage-gated ion channels.

• 

  Figure 4.. S2—S4 salt bridges/hydrogen bonds in the NMR and cryo-EM structures of the KCNQ1 VSD. (A) Intermediate state conformation of human KCNQ1 VSD (1st structure in the PDB 6MIE ensemble). Of particular note are the ionic interactions of E1-R2 and E2-R4, as well as, the close packing of Q3 and F0. (B) Activated state conformation of the Xenopus KCNQ1 VSD (PDB 5VMS). Note the gating charge residue pairings of E1-R4 and H5-E2. In all panels, the Cα atoms of the S4 polar residues are shown as yellow spheres and the transmembrane helices are labeled in vermillion text. S0 and S1 are not shown to improve clarity of side chain interactions.

• 

  Figure 5.. Schematics and electrophysiology data validating the intermediate and activated KCNQ1 VSD functional states utilizing auxiliary subunit KCNE1 regulation as a probe. Amino acid residue nomenclature: E2 = E170, R4 = R237, F0 = F167, Q3 = Q234, and H5 = H240. Numbering corresponds to the human KCNQ1 sequence. All error bars are ± SEM. All horizontal scale bars correspond to 2 s. (A) A cartoon schematic illustrating key S2, S3, and S4 residues interactions found in the NMR and cryo-EM VSD structures. Positive and polar gating residues on S4 (R1–H5) are colored blue, negative counter charges on S2 (E1, E2) and S3 (D202) are colored red, and the hydrophobic plug on S2 (F0) is colored orange. (B) A cartoon schematic displaying how the S4 charge-reversal mutation (Q3E) disrupts VSD function. The Q3E mutation creates electrostatic repulsion with the negative counter charges (E1, E2) and leads to VSD loss of function. (C) A cartoon schematic showing how the double charge-reversal mutations Q3E/F0R bias the VSD conformation. The double mutations ensure that electrostatic interactions between S2 and S4 are only favorable when the two mutation sites are in alignment (Q3E-F0R). (D) Table…

• 

  Figure 5—figure supplement 1.. Electrophysiology results for KCNQ1 F0R single and double mutants with and without KCNE1 co-expression, and with XE991 exposure. F0 = F167; Q3 = Q234; R4 = R237; H5 = H240. Residue numbers correspond to human KCNQ1. All error bars are ± SEM. (A) Top: Current recordings of KCNQ1 F0R single mutant with and without KCNE1 co-expression. Bottom: Average GV relationships for KCNQ1 F0R (black, n = 3), KCNQ1 F0R + KCNE1 (dark blue, n = 5), KCNQ1 WT (gray), and KCNQ1 WT + KCNE1 (light blue). The single mutant KCNQ1 F0R exhibited dramatic left-shifted voltage-dependent activation compared to KCNQ1 WT, suggesting that the mutation destabilizes the resting state VSD. KCNE1 co-expression restored voltage-dependent activation to near WT levels and potentiated ionic current compared to the α-subunit alone, suggesting that KCNE1 partly rescues the ability of the single mutant F0R to gate normally. (B–E) Top: Current recordings of KCNQ1 F0R double mutants with and without KCNE1 co-expression. Bottom: Average GV relationships for the KCNQ1 double mutant α-subunit only (black), KCNQ1 double mutant + KCNE1 (dark blue), KCNQ1 F0R (gray), and KCNQ1 F0R + KCNE1 (light blue). All n > 3. (B) F0R/D202N…

• 

  Figure 6.. Electrophysiology validating the intermediate and activated KCNQ1 VSD functional states utilizing XE991 pharmacology as a probe. Amino acid residue nomenclature and numbering and error bars are similar as in Figure 5. (A) Table outlining 5 µM XE991 effect on KCNQ1 IO and AO state currents based on prior studies (Zaydman et al., 2014). (B) Top: Exemplar diary plots of E2R/R4E drug studies demonstrating current amplitude over time. Cells were held at −20 mV and pulsed to +40 mV for 4 s and −40 mV for 2 s every 20 s. Each point shows the steady-state current amplitude at the end of the 4 s +40 mV test pulse. Cells were recorded in ND96 solution and the top bars indicate application of 5 μM XE991 (red) and 150 μM chromanol 293B (blue). Scale bar indicates 5 min. Bottom: Current traces for the E2R/R4E mutant in control ND96 solution (black), in solution containing 5 μM XE991 (red), and in solution containing 150 μM chromanol 293B (blue). The arrows in the diary plot indicate respective traces shown. Note that because the holding potential was −20 mV and the mutant channels are constitutively open, non-zero currents were observed before the test pulse. (C) Chromanol-subtracted E2R/R4E…

• 

  Figure 7.. KCNE3 shifts the voltage dependence of the intermediate-open state to render its conductance relevant under physiological conditions. (A) Voltage-clamp fluorometry recordings for pseudo-WT (C214A/G219C/C331A) KCNQ1 (left), KCNQ1+KCNE1 (middle), and KCNQ1+KCNE3 (right). Current (black) and fluorescence (blue) were recorded with voltages from −160 mV to +100 mV in 20 mV increments and then back to −40 mV. The bottom panels are the G–V (black) and F–V (blue) relationships with F1 and F2 components (dotted lines). The F–V relationship for KCNQ1 (gray) is also shown with KCNQ1+KCNE1 and KCNQ1+KCNE3 for comparison. (B) VCF recordings for KCNQ1-F351A in the presence (blue) and absence (gray) of KCNE3. The bottom of this panel shows the G–V and F–V relationships for KCNQ1-F351A (gray) and KCNQ1-F351A+KCNE3 (blue). (C) Left: WT KCNQ1+KCNE3 currents (black) with voltages from −120 mV to +80 mV in 20 mV increments and back to −40 mV to test the tail current. The red traces are total KCNQ1+KCNE3 currents with the instantaneous current subtracted, which show typical time- and voltage-dependent activation. Right: The two steps of KCNQ1+KCNE3 voltage sensor activation, F1 and F2, are shown in…