Lysine and the Na+/K+ Selectivity in Mammalian Voltage-Gated Sodium Channels

Abstract

Voltage-gated sodium (Nav) channels are critical in the generation and transmission of neuronal signals in mammals. The crystal structures of several prokaryotic Nav channels determined in recent years inspire the mechanistic studies on their selection upon the permeable cations (especially between Na+ and K+ ions), a property that is proposed to be mainly determined by residues in the selectivity filter. However, the mechanism of cation selection in mammalian Nav channels lacks direct explanation at atomic level due to the difference in amino acid sequences between mammalian and prokaryotic Nav homologues, especially at the constriction site where the DEKA motif has been identified to determine the Na+/K+ selectivity in mammalian Nav channels but is completely absent in the prokaryotic counterparts. Among the DEKA residues, Lys is of the most importance since its mutation to Arg abolishes the Na+/K+ selectivity. In this work, we modeled the pore domain of mammalian Nav channels by mutating the four residues at the constriction site of a prokaryotic Nav channel (NavRh) to DEKA, and then mechanistically investigated the contribution of Lys in cation selection using molecular dynamics simulations. The DERA mutant was generated as a comparison to understand the loss of ion selectivity caused by the K-to-R mutation. Simulations and free energy calculations on the mutants indicate that Lys facilitates Na+/K+ selection by electrostatically repelling the cation to a highly Na+-selective location sandwiched by the carboxylate groups of Asp and Glu at the constriction site. In contrast, the electrostatic repulsion is substantially weakened when Lys is mutated to Arg, because of two intrinsic properties of the Arg side chain: the planar geometric design and the sparse charge distribution of the guanidine group.

Fig 1. The structural representation and sequence alignment for the SFs of Na_v channels.

In the structural representation (upper panel), the Na_vRh crystal structure is shown as an example, where only two of the four chains (A and C) are represented in cartoons for clarity. In the sequence alignment (lower panel), the sequences for two prokaryotic channels (Na_vRh and Na_vAb) as well as the four domains of one mammalian channel (human Na_v1.4) are compared side-by-side, with the P-loop of the SF region highlighted by a blue frame. Two rings of residues are highly conserved in mammalian Na_v channels: the inner ring (shaded in red) and the outer ring (shaded in blue). The positions of these two rings are labeled by dotted frames colored in red and blue respectively, in the structural representation (upper panel).

Fig 2. Two constraint-free short simulations on the DEKA mutant in NaCl.

Na⁺ ions are represented as yellow spheres. (a, b) In the first simulation (ID 5 in Table B in S1 File), no Na⁺ ions (violet dots) can pass Lys180 side chain (blue line) to reach the interior Site_INT (a), and Lys180 side chain never falls within the interacting distances of any Glu183 residues (distances colored in black, red, green and blue for chain A, B, C and D respectively) (b). (c-e) In the second simulation (ID 6 in Table B in S1 File), the Na⁺ ion (violet dots) can cross Lys180 side chain (blue line) and arrive at Site_INT (c), and Lys180 side chain begins to interact with at least one of the Glu183 residues (distances colored in the same scheme) roughly at the same time. In the last 15 ns of this simulation, steady hydrogen bonds can form between Lys180 and Glu183 of chain A (e).

Fig 3. Ion binding patterns in the SF of the DEKA (a, b) and DERA (c, d) mutants of Na_vRh for Na⁺ (a, c) and K⁺ (b, d) ions in the a-system (IDs 13–16 in Table B in S1 File).

For each map, the upper panel shows a representative structure obtained from the equilibrium simulation. The side chains of residues 180 are shown in the licorice representation for clarity. The SF is shown in the cartoon representation and colored in cyan. Na⁺ and K⁺ ions are represented as yellow and orange spheres respectively. The lower panel exhibits the 2D free energy profile (in the unit of kcal/mol) estimated from the probability density map of cations in SF that is counted from the corresponding 200 ns equilibrium simulation. The vertical axis is the relative distance of the cation to the geometric center of the SF along the z-axis, with labels denoting the binding sites (Ion_EX, Site_HFS, Site_CEN and Site_IN) proposed from the static crystal structure of Na_vAb. Positions of Site_OC are labeled by black arrows. The horizontal axis is the distance between cation and the geometric center of the SF projected in the xy-plane.

Fig 4. Time-dependent ion occupancy in the SF of the DEKA (a, b) and DERA (c, d) mutants of Na_vRh for Na⁺ (a, c) and K⁺ (b, d) ions in the a-system (IDs 13–16 in Table B in S1 File).

The vertical axis is the relative distance of the cation to the geometric center of the SF along the z-axis. Violet and orange dots are used to represent the vertical positions of Na⁺ and K⁺ ions respectively. The red and green lines reflect the center of carboxylate oxygen atoms of Asp180 and Glu180 respectively. The blue line describes the position of the side-chain ζ-nitrogen (NZ) atom of Lys180 (a, b) or the ζ-carbon (CZ) atom of Arg180 (b, d).

Fig 5. Structural comparison on the hydrogen bonding interactions between the side chains of Lys180/Arg180 and other residues.

(a) For the DEKA mutant in the a-system (ID 13 in Table B in S1 File), Lys180 side chain can form only one stable hydrogen bond with Glu180. (b) For the DEKA mutant in the d-system (ID 17 in Table B in S1 File), Lys180 side chain can form concurrent hydrogen bonds with the carboxylate groups of Glu180 and Glu183 in chain D as well as the backbone carbonyl oxygen atom of Ala180. (c) For the DERA mutant in the a-system (ID 15 in Table B in S1 File), Arg180 side chain can form concurrent hydrogen bonds with the carboxylate groups of Glu183 residues in both chain A and chain D. (d) For the DERA mutant in the d-system (ID 19 in Table B in S1 File), Arg180 side chain can form concurrent hydrogen bonds with the carboxylate groups of Asp180 and Glu183 in chain D.

Fig 6. Diagram of the model to explain the difference in ion selectivity between the DEKA and DERA mutants.

Na⁺ and K⁺ ions are represented by yellow and orange spheres respectively. (a) In the DEKA mutant, Lys180 side chain tends to protrude into the center of the SF, repelling the cation to bind at the Na⁺-preferred sub-location sandwiched by at least three carboxylate oxygen atoms of Asp180 and Glu180. (b) In the DERA mutant, because of the stable bifurcate interactions with Glu183 residues (in chain A and D), Arg180 side chain tends to line along the side wall of the SF pore. This conformational difference as well as the lower charge density on the guanidine group jointly weaken the electrostatic repulsion and allow the cation to sample the less Na⁺-selective or even K⁺-preferred sub-locations coordinated by no more than two carboxylate oxygen atoms of Asp180 and Glu180.