Figure 1.
Drug positions in unbiased simulations.
20 Equally spaced snapshots from three unbiased simulations of benzocaine (A and C) and phenytoin (B and D) in the closed (A and B) and inactivated (C and D) NavAb. Each colour represents snapshots from a different simulation, while a single protein conformation is shown in each case.
Figure 2.
Free energy surfaces for (A,C) benzocaine and (B,D) phenytoin in the NavAb central cavity viewed along the pore axis (A,C) and from the membrane (B,D) obtained from metadynamics simulations.
Contours are shown at 1/mol intervals.
Figure 3.
Snapshots of the most commonly sampled binding poses.
The drug and surrounding residues are shown. The residues with the strongest interactions with the drug are named in bold. (A) benzocaine in the activation gate with amine pointing at the central cavity. (B) benzocaine in the activation gate with amine pointing down. (C) benzocaine in a fenestration with amine pointing to the central cavity. (D) Benzocaine in a fenestration with amine pointing toward the lipid. (E) Phenytoin in the activation gate. (F) Phenytoin in a fenestration.
Figure 4.
Drug-protein interactions in each site.
The interaction energies for (A) benzocaine and (B) phenytoin with residues lining the channel lumen when the drug is in one of the commonly occupied clusters. Four significantly different cluster are shown for benzocaine, corresponding to those pictured in Fig. 1 & 3. Two clusters are shown for phenytoin corresponding to binding at the activation gate (green) and fenestration (blue). Residues from regions not having significant interactions with the drugs are omitted.
Table 1.
Free energy of binding, and dissociation constants relative to bulk water for phenytoin and benzocaine at two sites in the NavAb central cavity.
Figure 5.
Potentials of mean force for benzocaine (red) and phenytoin (blue) moving from the pore axis (0 Å) to the lipid () through one of the hydrophobic fenestration (10–20 Å).
Snapshots of the drugs at positions along the fenestration are shown in Fig. 7.
Figure 6.
Lipid and water in the fenestrations as a function of drug position.
(A) A representative snapshot showing lipid occupying the lateral fenestrations while benzocaine sits at its minimum energy position in the activation gate. The position of F203 (orange) and T138 (yellow surface) are also shown. (B) The extent to which lipid penetrates into the fenestration is plotted as a function of the position of each drug. Low values indicate extension further into the fenestration. (C) A snapshot showing an extreme example of a water chain extending from the channel lumen to phenytoin on the exterior surface of the protein. In most cases the water chain does not extend this far. (D) The probability that a continuous water chain extends from each drug back to the channel lumen as a function of drug position (solid lines). Also shown is the probability that a water chain extends from the drug directly to bulk water (dashed lines). In B and D the data for individual windows are shown in points and a moving average of 5 data points is indicated by the line.
Figure 7.
Snapshots from the umbrella simulations are shown that represent important points in the PMF for benzocaine (A–D) and phenytoin (E–H).
The global minimum for each drug is at the activation gate (A,E), and the drug positions here replicate the binding poses seen in the equilibrium simulations. The same is true for the second minimum (B, F), which shows binding in the fenestration. A further hydrophobic pocket supports the drugs in the outer fenestration (C,G). At the external entrance to the fenestrations, the drugs have to pass the bulky phenylalanine residue (orange) and are at their most dehydrated creating the largest barrier in the pathway (D,H). Example snapshots show that even at this point water chains extending to the channel lumen are sometimes present.