Comparative gain-of-function effects of the KCNMA1-N999S mutation on human BK channel properties

Abstract

KCNMA1, encoding the voltage- and calcium-activated potassium channel, has a pivotal role in brain physiology. Mutations in KCNMA1 are associated with epilepsy and/or dyskinesia (PNKD3). Two KCNMA1 mutations correlated with these phenotypes, D434G and N999S, were previously identified as producing gain-of-function (GOF) effects on BK channel activity. Three new patients have been reported harboring N999S, one carrying a second mutation, R1128W, but the effects of these mutations have not yet been reported under physiological K⁺ conditions or compared to D434G. In this study, we characterize N999S, the novel N999S/R1128W double mutation, and D434G in a brain BK channel splice variant, comparing the effects on BK current properties under a physiological K⁺ gradient with action potential voltage commands. N999S, N999S/R1128W, and D434G cDNAs were expressed in HEK293T cells and characterized by patch-clamp electrophysiology. N999S BK currents were shifted to negative potentials, with faster activation and slower deactivation compared with wild type (WT) and D434G. The double mutation N999S/R1128W did not show any additional changes in current properties compared with N999S alone. The antiepileptic drug acetazolamide was assessed for its ability to directly modulate WT and N999S channels. Neither the WT nor N999S channels were sensitive to the antiepileptic drug acetazolamide, but both were sensitive to the inhibitor paxilline. We conclude that N999S is a strong GOF mutation that surpasses the D434G phenotype, without mitigation by R1128W. Acetazolamide has no direct modulatory action on either WT or N999S channels, indicating that its use may not be contraindicated in patients harboring GOF KCNMA1 mutations.NEW & NOTEWORTHYKCNMA1-linked channelopathy is a new neurological disorder characterized by mutations in the BK voltage- and calcium-activated potassium channel. The epilepsy- and dyskinesia-associated gain-of-function mutations N999S and D434G comprise the largest number of patients in the cohort. This study provides the first direct comparison between D434G and N999S BK channel properties as well as a novel double mutation, N999S/R1128W, from another patient, defining the functional effects during an action potential stimulus.

Keywords: BK channel; calcium-activated potassium channel; channelopathy; human mutations; seizure.

Fig. 1.

Wild-type (WT) and N999S BK channel properties in symmetrical K⁺ and 1 μM Ca²⁺. A: schematic representation of 1 BKα subunit, with the extracellular NH₂ terminus, 7 transmembrane segments (TM0–TM6), and the intracellular COOH terminus containing the Ca²⁺ binding sites. Black, voltage sensor domain (VSD); gray, RCK1 and RCK2 domains comprising the gating ring. Dark gray circles correspond to the calcium binding sites, Gray dotted line between TM6 and RCK1 calcium binding site corresponds to AC region. Relative positions for the mutations D434G [13 known patients (Bailey et al. 2019; Du et al. 2005)], N999S [7 known patients (Bailey et al. 2019; Li et al. 2018; Zhang et al. 2015)], and R1128W [1 patient known (Heim et al. 2019)] are shown. The only patient carrying the mutation R1128W additionally harbors the N999S mutation. B: representative macroscopic currents at 20-mV increments with the voltage-step protocol applied. C: conductance-voltage (G-V) relationship for WT (n = 7) and N999S (n = 6) channels. G was normalized to the highest conductance calculated (G_max). The G-V curves fit with a Boltzmann function (solid line). D: half-maximal voltage of activation (V_1/2) values for WT and N999S channels (****P < 0.0001, 1-way ANOVA, Tukey post hoc test). E: activation time constant (τ_act) vs. voltage for WT (n = 7) vs. N999S (n = 6). Insets show representative activation traces of both channels at +160 and +240 mV (the scale is the same for both). N999S shows significantly faster activation between +120 and +170 mV (P < 0.05, 2-way repeated-measures ANOVAs with Bonferroni post hoc) but not between −180 and +250. F: deactivation time constant (τ_deact) for WT (n = 4) and N999S (n = 5) channels. N999S shows significantly slower deactivation kinetics compared with WT (P < 0.05 at −60 to −100, 2-way repeated-measures ANOVAs with Bonferroni post hoc). Inset illustrates representative portion of the tail currents at −60 mV (2 s of the total 20-s deactivation protocol).

Fig. 2.

Wild-type (WT) and N999S channel properties in physiological K⁺ and 10 μM Ca²⁺. A: representative macroscopic currents at 10-mV increments are shown together with the voltage-step protocol applied. B: conductance-voltage (G-V) curves of WT (n = 9) and N999S (n = 11) channels. G was normalized to the highest conductance calculated (G_max). C: half-maximal voltage of activation (V_1/2) individual values for WT and N999S channels (P < 0.0001, 1-way ANOVA, Tukey post hoc test). D: activation time constant (τ_act) vs. voltage for WT (n = 9) vs. N999S (n = 7). Inset shows representative activation traces of both channels at +100 mV. The N999S channels open significantly faster than the WT in a range between 0 and 100 mV (P < 0.0001, 2-way ANOVA with Bonferroni post hoc). E: comparison of WT (n = 8) and N999S (n = 6) deactivation time constants (τ_deact). N999S exhibits slower deactivation kinetics compared with WT in the range of −80 to −60 mV (P < 0.05, 2-way ANOVAs with Bonferroni post hoc). Inset illustrates representative portion of the tail currents at −60 mV (2 s of the total 20-s deactivation protocol). F: action potential (AP) waveform voltage command (inset) and representative traces for WT and N999S AP-evoked currents. AP_peak, AP peak current. G: normalized AP_peak level to the maximum steady-state current evoked by a standard voltage activation protocol (I_max) (WT: 0.025, n = 9 and N999S: 0.465, n = 7; P < 0.0001, 2-way ANOVA, Tukey post hoc test). H: normalized subthreshold current I_Subthr level at −50 mV to I_max (WT: 0.002 ± 0.0007 and N999S: 0.07 ± 0.0008); ****P < 0.0001, 1-way ANOVA, Tukey post hoc test.

Fig. 3.

Comparison between N999S and D434G channel properties in physiological K⁺ and 10 μM Ca²⁺. A: representative macroscopic currents with the voltage-step protocol. B: conductance-voltage (G-V) relationships of N999S (n = 11) and D434G (n = 12) channels. The G-V curves are fitted by a Boltzmann function (solid line) with a half-maximal voltage of activation (V_1/2) of −8 ± 4.2 mV for N999S and +8 ± 3.2 mV for D434G mutations. G was normalized to the highest conductance calculated (G_max). C: individual V_1/2 values for N999S and D434G are significantly different (*P = 0.01, 1-way ANOVA with Tukey post hoc). D: activation kinetics [activation time constant (τ_act)] of N999S (n = 7) and D434G (n = 12) mutations. Inset corresponds to representative activation traces of both channels at +100 mV. The N999S channels open faster than D434G in the voltage range between 0 and 100 mV (P < 0.0001, 2-way ANOVA with Bonferroni post hoc). E: deactivation kinetics [deactivation time constant (τ_deact)] of N999S (n = 6) vs. D434G (n = 10). The N999S mutation has slower deactivation kinetics than the D434G mutation only at −60 mV (P < 0.0001, 2-way ANOVA with Bonferroni post hoc). Inset illustrates a representative portion of the tail currents at −60 mV (2 s of the total 20-s deactivation protocol). F: voltage protocol and representative traces of N999S and D434G action potential (AP)-evoked current. AP_peak, AP peak current. G: normalized AP_peak values to the maximum steady-state current evoked by a standard voltage activation protocol (I_max) for both mutations. The amplitude of N999S (n = 7) current is 2 times larger than that of D434G (n = 12) (****P < 0.0001, 1-way ANOVA with Tukey post hoc test).

Fig. 4.

Effect of the R1128W mutation with N999S. A: representative macroscopic currents at +20-mV increments for 1 μM Ca²⁺ or +10 mV with 10 μM Ca²⁺ are depicted. B: conductance-voltage (G-V) relationships of N999S (n = 11) and N999S/R1128W (n = 7) channels in symmetrical K⁺ and 1 μM Ca²⁺ and physiological K⁺ and 10 μM Ca²⁺. G was normalized to the highest conductance calculated (G_max). The N999S/R1128W half-maximal voltage of activation (V_1/2) is +79 ± 1 mV in symmetrical K⁺-1 μM Ca²⁺ and −4 ± 6 mV in physiological K⁺-10 μM Ca²⁺. C: individual V_1/2 values for N999S and N999S/R1128W are not statistically different [P > 0.05 (n.s.), 1-way ANOVA with Tukey post hoc test]. D and E: activation [D; activation time constant (τ_act) N999S/R1128W n = 5 and N999S n = 7] and deactivation [E, deactivation time constant (τ_deact) N999S/R1128W n = 5 and N999S n = 7] kinetics of N999S and N999S/R1128W mutations. There is no significant difference in both kinetics in the entire range of voltages evaluated (P > 0.05, 2-way ANOVA with Bonferroni post hoc). F: voltage protocol (inset) and representative traces of N999S and N999S/R1128W action potential (AP)-evoked current. AP_peak, AP peak current. G: normalized AP_peak values for both mutations (N999S/R1128W n = 5 and N999S n = 7) show the same current amplitude, [P > 0.05 (n.s.), 1-way ANOVA with Tukey post hoc test].

Fig. 5.

Effect of acetazolamide (ACTZ) and paxilline on wild-type (WT) and N999S channels. A and B: representative experimental protocol (left) and current (I) level vs. time for WT (A) and N999S (B) channels. The black bar denotes application of 50 μM ACTZ to the patches, and the gray bar corresponds to application of 100 nM paxilline. C and E: individual current levels from each patch before ACTZ (30 s), after ACTZ (120 s), and after paxilline (Pax; 240 s) for WT (n = 6) and N999S (n = 9) channels. D and F: normalized data presented as % of the initial current (I_initial). No significant differences were found for WT or N999S channels after 50 μM ACTZ [P > 0.05 (n.s.), paired t test]. The reduction of the current due to paxilline application was significant (P < 0.05, paired t test).