Chlorpromazine inhibits EAG1 channels by altering the interdomain coupling

Figures

• 

  Figure 1. Structural and functional properties of EAG1 channels. (A) A linear representation of EAG1 channel subunit topology with residue positions at the border of the domains indicated for clarity. The PAS domain is shown in green, S1-S4 transmembrane helices forming the VSD in light gray, S5-S6 transmembrane helices forming the PD in dark gray, the C-linker connecting the CNBH domain to the S6 helix in cyan and the CNBH domain in blue. (B) A ribbon representation of the full-length cryo-EM structure of rat EAG1 channels (PDB: 5K7L) viewed from the side and down the pore of the channel. The domain color coding is the same as for (A). (C) Voltage protocol used to elicit Cole-Moore shift in EAG1 channels. The protocol consists of 7-s voltage prepulses ranging from −110 to +40 mV in 10-mV increments from a holding potential of −80 mV, followed by a 0.5-s voltage pulse to +40 mV. (D) Representative currents from WT EAG1 channels elicited with the voltage protocol shown in (C). Current traces corresponding to the voltage prepulse at −110 and +40 mV are colored in red and blue, respectively. (I) Currents encompassed by the dashed square in (D) shown in an expanded time resolution.

• 

  Figure 2. RMSD and RMSF profile for apo and CPZ-bound EAG1 channels throughout 1 μs MD simulations. (A) An atomistic model of the full-length mEAG1 channel embedded in a membrane composed of POPC lipids. The PAS domain is shown in green, S1-S4 transmembrane helices forming the VSD in light gray, S5-S6 transmembrane helices forming the PD in dark gray, the C-linker connecting the CNBH domain to the S6 helix in cyan, and the CNBH domain in blue. Root mean-square deviations (RMSD) throughout 1 μs MD simulations for different domains and whole protein (WP) backbone of an EAG1 subunit from three replicates and averaged RMSD for three replicates and all four subunits in apo (B) and holo states (C). The RMSD is calculated for the domains aligned to the starting structure of the whole protein to show domain movement and for WP the same alignment is used to show overall structural changes. The color coding of the domains is the same as in (A) and red for WP. For the averaged RMSD plots the standard deviation (SD) is shown in the semitransparent color of the corresponding domain or WP. (D) Root mean-square fluctuations (RMSF) for an EAG1 subunit from three replicates and averaged RMSF for three…

• 

  Figure 3. CPZ alters interactions between the PAS and CNBH domains in full-length EAG1 channels. H-bonds and salt bridges between the interacting PAS (green) and CNBH (blue) domains for representative poses in an apo (A) and CPZ-bound (B) states with the interacting residues labeled and shown in sticks. Residues that form stronger interactions in the apo state are shown in gray in (A) and residues that form stronger interactions in the CPZ-bound state are shown in red in (B). Parts of the structure in (B) are shown in a semitransparent representation to reveal the residues that otherwise will not be visible in the selected view. (C) Histograms of the bond distances for the indicated residue pairs located in the PAS and CNBH domains from the adjacent EAG1 subunits calculated in the apo (blue) and holo (red) states. Residue pairs interactions between which are strengthened by CPZ binding are shown in red, the ones that are weakened in gray. The dashed line denotes 3.3 Å.

• 

  Figure 4. CPZ strengthens interactions between the isolated PAS and CNBH domains. (A) SPR sensorgrams for the EAG1 PAS domains applied at the indicated concentrations to the EAG1 CNBH domains immobilized on the CM5 sensor chip in the absence and presence of the indicated CPZ concentrations. Gray lines represent fits of the data with the two-state reaction binding model using the Biaevaluation software. Kd values were 2.9, 1.8, 1.6, and 0.5 μM in the presence of 0, 10, 30, and 50 μM CPZ, respectively. (B) Plots of the averaged PAS/CNBH domain interaction affinities versus CPZ concentration. The averaged affinities for different SPR chips are shown in different colors and the means of the affinities ±SD are shown as black lines. n = 3. ∗p < 0.05 by ANOVA. The corresponding association and dissociation rate constants for each of the three experiments and the averaged Kd values can be found in Table S2.

• 

  Figure 5. Network analysis of MD simulations indicates that CPZ binding alters allosteric coupling between the PAS/CNBH domain ring, VSD and PD. (A) Information flow (IF) profile averaged for three replicates and four subunits of EAG1 for the apo (red) and CPZ-bound (blue) states mapped to the amino acid sequence of EAG1 channels. Sequence regions corresponding to the PAS, VSD, PD, C-linker, and CNBH are shaded in green, light gray, dark gray, cyan, and blue, respectively. Error bars correspond to SD. (B) The difference between the averaged information flow profiles of CPZ-bound and free states (ΔIF) mapped to the amino acid sequence of EAG1 channels. Dashed lines denote the 95% CIs. Residues with significantly altered ΔIF are indicated as black dots. (C) Residues with significantly altered ΔIF from (B) projected onto the structure of EAG1 channels. Blue spheres indicate residues with significantly decreased ΔIF, while red indicates those with significantly increased ΔIF.

• 

  Figure 6. CPZ binding to the PAS domain does not alter Cole-Moore shift in EAG1 channels. Representative currents from WT EAG1 (A) and Y71G mutant (D) channels recorded at +40 mV after prepulses from −110 to −40 mV in the absence (black) and presence (red) of 30 μM CPZ. The voltage protocol is illustrated in Fig. 1C and currents in (A) and (D) correspond to the region encompassed by the dashed square in Fig. 1D. The representative currents in the absence and presence of CPZ were recorded from the same oocyte expressing WT (A) or Y71G (D) mutant channels. The dashed lines correspond to zero current levels. Plots of the averaged normalized currents for WT EAG1 channels recorded at +40 mV plotted against the prepulse potential in the absence (black symbols) and presence (red symbols) of 30 μM CPZ for WT EAG1 (B) and Y71G mutant (E) channels. The lines represent fits with the Boltzmann equation with the V1/2 of −74.5 ± 0.7 mV in the absence and −72.4 ± 0.7 in the presence of CPZ for WT channels, and V1/2 of −68.9 ± 0.5 mV in the absence and −66.9 ± 0.5 in the presence of CPZ for Y71G mutant channels. n = 6 for WT channels, and n = 9 for Y71G mutant channels. Plots of the averaged rise time to…

• 

  Figure 7. A model of EAG1 current inhibition by CPZ binding to the PAS domain. Depolarization causes a rotation of the intracellular tetrameric ring formed by the PAS and CNBH domains around the pore formed by PDs of the four channel forming subunits. This leads to the opening of the channel pore (top right). PAS domain residues act as a source and PD residues act as a sink. Binding of CPZ, to the PAS domains alters the PAS/CNBH domain interface decreasing the coupling between the source and sink residues (bottom left). This hinders the rotation of the tetrameric ring, inhibiting current from EAG1 channels (bottom right). PAS domains are shown in green, CNBH domains in blue, PD domains in gray, and CPZ in orange.