Molecular dynamics of the pathogenic KCNQ2 variant G256W reveal mechanisms of channel dysfunction in epileptic encephalopathy

Figures

• 

  Figure 1.. G256W lies in the extracellular turret region of the pore domain near the protein-lipid interface A. Extracellular up, side view of a KCNQ2 tetramer based on the cryoEM model (PDB: 7CR3). One subunit, labeled in color, shows the modifications added in this study, i.e., the extracellular loop between S3-S4, and, near the top of the channel’s extracellular domain, the G256W substitution. Two subunits are in transparent gray; one subunit has been partially removed for clarity. B. Extracellular view of an MD simulation representative frame depicting a tetramer embedded in the membrane. Intracellular N and C terminal regions were omitted for clarity. Red arrows point to G256 locations, highlighting their proximity to lipid. C. Extracellular view of the tetrameric pore domain (PDB: 7CR3) extracellular region and transmembrane regions near the selectivity filter. The four substituted W256 sidechains are shown in green. Helix and segment colors are as shown for the colored subunit in A. Abbreviations: VSD, voltage-sensor domain; HA, HB, and HC, the cytoplasmic helices HA, HB, and HC; CaM, calmodulin.

• 

  Figure 2.. Root mean square deviation of the WT and G256W production simulations show structure stable structures and greater deviation for G256W Root mean square deviation averaged across all subunits for A. WT and B. G256W 500 ns production simulations. Lines for replicates one (black), two (blue), and three (green) illustrate full structure RMSD traces (faint) and smoothed average lines (bolded) per simulation. Legend displays the calculated RMSD average and standard deviation for each replicate simulation.

• 

  Figure 3.. Preferential interaction between HC helix and one calmodulin subunit differing between WT and G256W indicates G256W may have long range structural effects A. Backbone atom root mean square fluctuation averaged over all of WT (black) and G256W (blue) subunits. Arrows point to local peaks at S3-S4 loop and turret regions. A KCNQ2 domain map is shown below (map colors as in Figure 1A). The vertical grey line at residue 366 (asterisk in the domain diagram) indicates a jump in sequence to residue 536 corresponding to a poorly resolved HA-HB loop. B. Surface representation of the channel pore domain and HC helix. The more peripheral VSD and HA-HB-CaM components of KCNQ2 are transparent gray ribbons. A pink line highlights the asymmetric orientation of the tetrameric HC coiled-coil (corresponding to the pink box enclosed region in A).

• 

  Figure 4.. Root mean square fluctuation analysis reveals G256W shows reduced turret flexibility compared to the WT Per residue root mean square fluctuation averaged over all subunits including side chains. Schematic representations of the KCNQ2 domains are shown below each graph with colors corresponding to Figure 1A. A. Comparison of the WT (black) and G256W (blue) structures of all modeled KCNQ2 residues. The grey line marks the point of omission of residues 367 to 535 (i.e., the HA-HB loop, unresolved in cryoEM structures). The green line marks G256W. B. A zoom-in on the pore domain region boxed in pink in panel A.

• 

  Figure 5.. W256 side chain migration and burial into the lipid membrane Superimposition of 150 images of single subunits from G256 and W256 simulations. A. G256 and B. W256 are colored green; samples are taken every 10 ns intervals over 1.5 μs. WT G256 is largely at or above the membrane surface, but W256 sidechain is often buried in the lipid (grey). Scale with tick marks are approximately 1 Å. C. Running average plots of the distance between membrane center and the average positions of G256 (black) or W256 (blue) backbone atoms, or selected subsets of W256 sidechain atoms. D. Trp stick model. Atom numbering corresponds to line plots in C and published literature.38,39 Distance to membrane center Average ± SD were 23.1 ± 1.1 Å (WT 11–13), 23.1 ± 0.9 Å (G256W 11–13), 22.2 ± 0.9 Å (G256W 3–5), 21.7 ± 0.9 Å (G256W 1–9), and 21.3 ± 0.9 Å (G256W 1–2, 9). E. W256 side chain atom subset distances are significantly smaller than the G256 mainchain distance.

• 

  Figure 6.. WT and G256W simulations reveal differences in turret network hydrogen bonding Representative structure frames of A. WT and B. G256W turret regions are labeled for percentage of time spent hydrogen bonded to nearby residues across all simulations. G256 and W256 are shown in solid green; other amino acid sidechains are colored by heteroatom. S5, S6, and pore helices are shown in blue; the turret is pink, and the selectivity filter (including SF bridge) is shown in yellow. Prominent interactions and those captured in the selected still images are labeled.

• 

  Figure 7.. Network diagram of hydrogen bonding frequencies depict differences in WT and variant simulations Hydrogen bonding network diagrams for the A. WT (black) and G256W (red) 1.5 μs of equilibrated simulations, and B. N258Y, C. H260P, and D. K283M. For B-D, percentages are averaged from the 200 ns simulations. Open circles are colored as pink (turret), teal (pore helix), yellow (selectivity filter bridge), and blue (helix S6). Labels on lines connecting residues display the percentage of frames showing interaction. The line color key in B (upper right) applies to C and D. Black lines represent hydrogen bonds within 5% of the WT simulations, blue and red lines indicate at least a 5% increase or decrease, respectively, compared to the WT. Dotted lines represent interactions in the variant simulations that are not present (less than 15%) in the WT.

• 

  Figure 8.. K255 has access to more conformational states in the WT structure K255 (purple) positions are superimposed every 10 ns for all simulations (1.5 μs) of A. WT and B. G256W channels. Extracellular view of static representative G256 and W256 pore domains with S5 and S6 helices restricted to the extracellular halves and intervening p-loop are shown in C and D. K255 locations shown in Figure 5 of all subunits superimposed every 10 ns for all simulations (1.5 μs) of C. WT and D. G256W. The protein ribbon is shown in blue with a yellow selectivity filter and pink turret. A W256 skeletal structure is shown in green in panel B and D.

• 

  Figure 9.. Representative images illustrate altered pore profile and asymmetry that occur with G256W A. WT and B. G256W representative structures of the transmembrane domains (blue). Extracellular view of pore radii using the canonical selectivity filter sequence carbonyls (pink) embedded in a membrane (dark grey) superimposed across all simulations (1.5 μs) for every 1 ns. Intracellular N and C terminal regions are omitted for clarity.