Abstracts

Rationale: Many antiepileptic drugs are known to inhibit voltage gated sodium channels in a state-dependent manner. Phenytoin (PHT) acts in this fashion, and it has been proposed that it binds to a residue on the sixth transmembrane helix of domain IV in eukaryotic sodium channels (DIVS6). It has also been proposed that it primarily modulates the fast inactivated state, although some studies have implicated separate slow inactivation processes. Finally, studies on antiarrythmic drugs, which are thought to have a similar binding site and mechanism of action, have suggested that the apparent inactivation-state preference is secondary to effects on voltage-gating. We attempted to clarify these issues by performing extensive unbiased atomic scale simulations of PHT and the local anaesthetic benzocaine (BZC) binding to the bacterial Na_vAb channel structure derived from crystallography data. Methods: Newly derived CHARMM models of PHT and BZC were used to explore interactions with lipid and the Na_vAb channel using the Anton supercomputer over a time scale of microseconds. The bacterial NavAb channel shows close homology with the mammalian sodium channel. Results: PHT displayed relatively strong binding at the membrane interface with a minimum of -4.1±0.1 kcal/mol allowing both aromatic rings to reside in the lipid region. It was found to cross the membrane on the sub-millisecond scale (35±3ms^-1), but about 2 orders of magnitude more slowly than BZC, and in a much more restricted region of the channel compared to BZC. PHT did not appear to interact significantly with the homologue of the D4S6 eukaryotic channel consensus binding site, but had several apparent sites, including the P loop region involved in slow inactivation, as well as a charged residue in the S4 voltage sensing domain (R108). Binding regions were often composed of links between spatially disparate amino acids, rather than just single residues. Intriguingly, PHT appeared to form a complex with the channel, locking helices in place as well as blocking the activation gate from both opening and collapse, associated with slow inactivation (N211, D219). Surprisingly high affinity binding of PHT was found to the pore domain and voltage sensing domain interface. Conclusions: These simulations provide potential explanations for a wide range of experimental observations, including the role of slow inactivation and voltage-sensing regions for sodium channel modulators such as phenytoin, They also provide very high resolution data about potential binding sites that will inform further pharmacological studies as well as drug design.