Constitutive sodium permeability in a C. elegans two-pore domain potassium channel

Abstract

Two-pore domain potassium (K2P) channels play a central role in modulating cellular excitability and neuronal function. The unique structure of the selectivity filter in K2P and other potassium channels determines their ability to allow the selective passage of potassium ions across cell membranes. The nematode C. elegans has one of the largest K2P families, with 47 subunit-coding genes. This remarkable expansion has been accompanied by the evolution of atypical selectivity filter sequences that diverge from the canonical TxGYG motif. Whether and how this sequence variation may impact the function of K2P channels has not been investigated so far. Here, we show that the UNC-58 K2P channel is constitutively permeable to sodium ions and that a cysteine residue in its selectivity filter is responsible for this atypical behavior. Indeed, by performing in vivo electrophysiological recordings and Ca²⁺ imaging experiments, we demonstrate that UNC-58 has a depolarizing effect in muscles and sensory neurons. Consistently, unc-58 gain-of-function mutants are hypercontracted, unlike the relaxed phenotype observed in hyperactive mutants of many neuromuscular K2P channels. Finally, by combining molecular dynamics simulations with functional studies in Xenopus laevis oocytes, we show that the atypical cysteine residue plays a key role in the unconventional sodium permeability of UNC-58. As predicting the consequences of selectivity filter sequence variations in silico remains a major challenge, our study illustrates how functional experiments are essential to determine the contribution of such unusual potassium channels to the electrical profile of excitable cells.

Keywords: C. elegans; electrophysiology; ion selectivity; molecular dynamics; two-pore domain potassium channel.

Fig. 1.

The two-pore domain potassium (K2P) channel UNC-58 is expressed in neurons and muscle cells in C. elegans. (A) Schematic transmembrane topology of a two-pore domain (K2P) potassium channel subunit. Transmembrane segments are numbered “1” through “4.” The two selectivity filter regions (SF1 and SF2) precede the second and fourth transmembrane helix. A pore helix (light gray rectangles) precedes each selectivity filter. Functional channels are formed by the dimerization of two K2P subunits. (B) Schematic representation of a prototypical potassium channel selectivity filter. Four evenly spaced ion binding sites (S1–S4) are formed by the carbonyl oxygens (red dots) of the TxGYG selectivity filter residues. Two ions are positioned in S2 and S4 (purple). (C) unc-58 gene structure, position of transmembrane domains (TM1 through TM4, above exons), point mutations (e665, bln205, n495, e415) and deletion (bln223, red line) are shown, as well as the transcriptional (bln259) and translational N-terminal knock in alleles (bln323, bln324). Amino acid positions in reference to unc-58 isoform b (transcript T06H11.1b.1). (D) unc-58 transcriptional reporter. Left panel, anterior body wall muscles (BWM). Middle panel, ventral nerve cord (VNC) and cell bodies of ventral nerve cord motoneurons (MN). Right panel, ALM, commissures (C, arrowhead), and dorsal nerve cord (DNC). (Scale bar, 10 µm.) (E) unc-58 translational reporter. Left panel, UNC-58A isoform, mNeonGreen was fused in-frame upstream of exon 1a. Right panel, UNC-58B isoform, mNeonGreen was fused in-frame upstream of exon 1b. Asterisk, neuronal cell bodies. Arrows indicate body wall muscle plasma membranes. (Scale bar, 10 µm.) (F) Displacement of worms over 50 s on nematode growth media. Depletion of UNC-58(gof) in neurons and muscle improves locomotion compared to depletion of UNC-58(gof) in neurons alone. Inset, tenfold-magnified view of unc-58(gof) tracks. (Scale bar, 50 µm.) (G) Average speed of unc-58(gof) (n = 49), neuronally (n = 39), and neuronally and muscle-depleted (n = 41) unc-58(gof) animals. Bar, mean; whiskers, SD. Each point represents the speed of an animal. One-way ANOVA, followed by Holm–Šídák’s multiple comparisons test, ***P < 0.001 and ****P < 0.0001.

Fig. 2.

unc-58 gain-of-function mutation increases neuronal and muscle excitability. (A) Average traces of Ca²⁺ responses of ALM neurons stimulated using a “short press” touch protocol in control (n = 37), gain-of-function unc-58(e665) (n = 29) and loss-of-function unc-58(bln223) (n = 36) animals expressing the ratiometric calcium sensor YC3.60 in mechanosensory neurons. Orange bars indicate duration of physical stimulus. Black trace, mean CFP/YFP ratio change, ΔR (%). Gray trace, SD. (B) Touch response of ALM mechanosensory neurons is modulated by mutation of unc-58. Left, proportion of ALM neurons responding to short press mechanical stimulus in wild-type (n = 37), gain-of-function mutant unc-58(e665) (n = 29), and loss-of-function mutant unc-58(bln223) (n = 36). P = 0.00369 [wild type vs. unc-58(e665)] and 0.0137 [wild type vs. unc-58(bln223)], chi-squared test. Right, ratio change, ΔR (%) of calcium response only for responder cells. P = 0.0111 and P = 0.3443, respectively, Mann–Whitney test. Each data point represents the response of one ALM neuron. Nonresponders, empty circles. (C) Resting membrane potential of C. elegans body wall muscles from wild-type (n = 7) and gain-of-function mutant unc-58(bln205) (n = 7). Line, median; box, IQR. Mann–Whitney test, **P < 0.005.

Fig. 3.

UNC-58 is an unconventional sodium-permeable K2P channel. (A) Current–voltage relationships obtained from X. laevis oocytes after injection of cRNA encoding wild-type (UNC-58, n = 12), gain-of-function mutant (UNC-58 L428F, n = 6), chimeric UNC-58 channels (UNC-58_ctTWK-18, n = 15), and noninjected oocytes (n = 12). Recordings were performed 72 h after injection. (B) Current–voltage relationships from X. laevis oocytes expressing UNC-58 F294N (n = 5) in physiological extracellular solution (96 mM Na⁺, solid cyan square) or after ionic substitution of extracellular sodium by NMDG (0 mM Na⁺, open cyan square). The inset shows NMDG condition at a reduced scale. Recordings were performed 24 h after injection. (C) Current–voltage relationships from X. laevis oocytes expressing UNC-58 L428F (n = 7) in physiological solution (96 mM Na⁺, solid red square) or after ionic substitution of extracellular sodium by NMDG (0 mM Na⁺, open red square). Recordings were performed 48 h after injection. (D) Current–voltage relationships from X. laevis oocytes expressing the UNC-58_ctTWK-18 chimera (n = 9) in physiological extracellular solution (96 mM Na⁺, solid blue square) or after ionic substitution of extracellular sodium by NMDG (0 mM Na⁺, open blue square). Recordings were performed 24 h after injection. Each point represents the mean ± SD. For clarity, only one side of the SD is shown in panel A. Curves were drawn for illustrative purposes only.

Fig. 4.

Ion coordination at the selectivity filter of the UNC-58 channel. (A) Alignment of SF1 and SF2 selectivity filter sequences from selected human and C. elegans K2P channels. The SF1 and SF2 residues are labeled in red and blue, respectively. UNC-58 is the only channel containing a cysteine residue (in bold and underlined) at the second position of the TxGYG SF1 motif. (B) Structural model of the UNC-58 dimer, modeled with AlphaFold and used for molecular dynamics simulations. Individual subunits are labeled in orange and purple. Transmembrane helix M3 and SF2 have been omitted for clarity. The inset represents the SF1 region and shows opposing selectivity filter loops belonging two each subunit. K⁺ ions (green spheres) are coordinated at position S2 and S4 within the selectivity filter, while positions S1 and S3 are occupied by water molecules. (C–F) Upper panels indicate the distance between the S2 ion and the oxygen atom of the C266 carbonyl backbone during molecular dynamics (MD) simulations for wild-type (WT) UNC-58 with K⁺ (C) or Na⁺ (D) and for the UNC-58 C266I mutant with K⁺ (E) or Na⁺ (F). Oxygen atoms within a confinement radius of 3.5 Å around the ion (red dotted lines) are counted as coordinating. Lower panels show representative licorice structures at the beginning (0 ns) and at the end (200 ns) of MD simulation. The residues of the SF1 selectivity filter of chain A and B are shown in orange and purple, respectively. K⁺ and Na⁺ ions are shown as green and cyan spheres, respectively.

Fig. 5.

Cysteine 266 contributes to the sodium permeability of UNC-58. (A) Current–voltage relationship obtained from X. laevis oocytes injected with water (gray square, n = 9) or cRNA encoding UNC-58 F294N (cyan square, n = 16) or UNC-58 C266I F294N channels (burgundy square, n = 16). Each point represents the mean ± SD. Curves were drawn for illustrative purposes only. Recordings were performed 48 h after injection. (B) Reversal potential of the current recorded from X. laevis oocytes expressing UNC-58 F294N and UNC-58 C266I F294N mutant channels. Current–voltage relationships in A were fitted with a linear fit from −60 to −10 mV for oocytes injected by water (gray square), from −40 to +10 mV for oocytes expressing UNC-58 F294N (cyan square), and from −20 to +20 mV for oocytes expressing UNC-58 C266I F294N (burgundy square). Line, median; box, IQR. Kruskal–Wallis test, P = 0.0031, followed by Dunn’s post hoc test (**P < 0.005 and ***P < 0.0005). (C) Current–voltage relationships obtained in X. laevis oocytes expressing UNC-58 F294N (cyan square, n = 16) or UNC-58 C266I F294N channels (burgundy square, n = 16) in physiological extracellular solution (96 mM Na⁺, solid square) or after ionic substitution of extracellular sodium by choline (0 mM Na⁺, open square). Current values for oocytes expressing UNC-58 F294N should be read with the scale on the Left of y-axis; current values for oocytes expressing UNC-58 C266I F294N should be read with the scale on the Right of y-axis. Each point represents the mean ± SD. Curves were drawn for illustrative purposes only. Recordings were performed 24 h after injection. (D) Reversal potential of the current recorded from X. laevis oocytes expressing UNC-58 F294N and UNC-58 C266I F294N mutant channels. Current–voltage relationships in C were fitted with a linear fit from −40 to +10 mV for oocytes expressing UNC-58 F294N in physiological solution (solid cyan square), from −60 to −20 mV for oocytes expressing UNC-58 F294N in 0 mM Na⁺ solution (open cyan square), from −20 to 20 mV for oocytes expressing UNC-58 C266I F294N in physiological solution (solid burgundy square), and from −60 to −20 mV for oocytes expressing UNC-58 C266I F294N in 0 mM Na⁺ solution (open burgundy square). Line, median; box, IQR. Kruskal–Wallis test, P < 0.0001, followed by Dunn’s posttests (NS: nonsignificant, *P < 0.05, **P < 0.005, and ***P < 0.0005).