BK channel inactivation gates daytime excitability in the circadian clock

Abstract

Inactivation is an intrinsic property of several voltage-dependent ion channels, closing the conduction pathway during membrane depolarization and dynamically regulating neuronal activity. BK K(+) channels undergo N-type inactivation via their β2 subunit, but the physiological significance is not clear. Here, we report that inactivating BK currents predominate during the day in the suprachiasmatic nucleus, the brain's intrinsic clock circuit, reducing steady-state current levels. At night inactivation is diminished, resulting in larger BK currents. Loss of β2 eliminates inactivation, abolishing the diurnal variation in both BK current magnitude and SCN firing, and disrupting behavioural rhythmicity. Selective restoration of inactivation via the β2 N-terminal 'ball-and-chain' domain rescues BK current levels and firing rate, unexpectedly contributing to the subthreshold membrane properties that shift SCN neurons into the daytime 'upstate'. Our study reveals the clock employs inactivation gating as a biophysical switch to set the diurnal variation in suprachiasmatic nucleus excitability that underlies circadian rhythm.

Figure 1. The β2 subunit is required for SCN neuronal firing rhythmicity and circadian behaviour.

(a) Representative spontaneous action potential activity recorded for 3 days on a multielectrode array from WT SCNs. Firing rate shows a robust peak-to-trough difference. (b) β2 KO SCN activity. Firing amplitude and rhythmicity were reduced. (c) The percentage of recordings within the SCN exhibiting rhythmic firing is decreased in β2 KO compared with WT and β4 KO SCNs (n=8, 11 and 10 SCN slices, respectively). (d) χ² periodogram analysis of action potential activity. χ² circadian peak amplitudes were reduced in β2 KO compared with WT and β4 KO SCNs. (e) Multiunit firing frequency from the peak and trough of rhythmic recordings, or from arrhythmic recordings. β2 KO firing is reduced during the peak. (f) Locomotor wheel running activity from a representative WT mouse. (g) β2 KO actogram. (h) χ² periodogram analysis of wheel behaviour. Dotted line denotes 3,000 (amplitude). (i) β2 KO mice re-entrained to a 6 h phase advance of the light–dark cycle faster than WT. (j) Exposure to a light pulse at CT16 caused a greater phase delay in β2 KO compared with WT. Representative actograms for i and j are in Supplementary Fig. 1. All values are mean±s.e.m. *P<0.05, Bonferroni post hoc (c–e) or t-test (i–j).

Figure 2. BK_i and BK_s currents in SCN neurons.

(a) Representative BK current traces and macroscopic decay time constants (τ_inact) for BK_i and BK_s currents from WT SCN neurons. BK currents were isolated by subtraction with the antagonist paxilline (Methods). (b) Fractional BK current (I_30ms/I_peak at +90 mV) was reduced during the day compared to night in WT neurons, D, daytime recording; N, night time recording. (c) τ_inact (at +90 mV) is lower for daytime BK_i currents compared to night BK_s currents in WT neurons. (d) Proportion of neurons with BK_i and BK_s currents. WT neurons have more BK_i currents during the day compared with night. (e) Current density versus voltage relationship for day and night BK currents from WT SCN neurons. The average daytime current magnitude is reduced compared with night. (f) BK_i current density is lower than BK_s, during the day or at night. All values are mean±s.e.m. n values: WT, day BK_i (18), BK_s (9) and night BK_i (3) BK_s (19). *P<0.05, Fisher's exact test (d) or Bonferroni post hoc (e,f).

Figure 3. Expression of BK α and β2 in WT SCNs.

(a) BK channel complexes were immunoprecipitated with an anti-BK α subunit antibody from WT SCNs at the indicated time points. Western blot analysis was performed for BK α (top panel), β2 (middle) or α-tubulin (bottom). α-Tubulin westerns (loading control) were obtained by running an equivalent volume of supernatant as was used for the immunoprecipitation. Protein was also harvested from α+β2 subunits co-expressed in HEK293 cells (positive control) or from β2 KO SCNs (negative control). ZT, zeitgeber time. Images have been cropped for presentation. Full size images are presented in Supplementary Fig. 3. (b) BK α and β2 band intensities normalized to α-tubulin. α expression increases at night, while β2 expression does not change. Data are the average from four independent timed SCN collections (four SCNs at each timepoint). (c) Ratio of β2:α expression in WT SCNs across the circadian cycle. All values are mean±s.e.m. *P<0.05, Bonferroni post hoc.

Figure 4. The β2 subunit is required for BK_i current decay and the diurnal difference in BK current magnitude in SCN neurons.

(a) Representative BK current traces and macroscopic decay time constants (τ_inact) for BK_s currents from β2 KO neurons. Voltage protocol same as in Figure 2. (b) All currents are BK_s from β2 KO SCNs, day or night. (c) β2 KO neurons do not show a day–night difference in BK current density. The β2 KO daytime BK current magnitude is larger than WT, comparable to WT levels at night. (d) The fractional BK current (I_30ms/I_peak) was similar during the day compared with night in β2 KO neurons. (e) τ_inact values from β2 KO BK_s currents are all >100 ms. All values are mean±s.e.m. n values: β2 KO, day (20) and night (20).

Figure 5. The β2N terminus (β2N) causes inactivation of BK currents from α-only channels expressed in heterologous cells.

(a) Representative macroscopic current traces from HEK293 cells expressing BK_SRKR channels, a daytime BK variant previously cloned from SCN (ref. 60). β2N (0–100 μM), corresponding to the first 45 amino acids of the N terminus, was dissolved in recording solution and applied to the intracellular side via the patch pipette. (b) Representative currents in the presence of 50 μM β2N^ΔFIW, a peptide mutating the three residues required for inactivation. There was no macroscopic current decay with β2N^ΔFIW. (c) Representative traces from the +90 mV voltage step at each concentration of β2N. The current peaks were scaled to illustrate the dose-dependent speeding of inactivation. (d) τ_inact values were calculated from the current elicited at +90 mV and plotted as a function of β2N concentration. τ_inact was plotted for BK_i and β2 KO SCN BK currents for cross-comparison. (e) τ_inact versus voltage for BK channels co-expressed the β2 subunit, or with 50 μM β2N. There is no difference in the voltage dependence of activation using the isolated β2N compared with the intact β2 subunit. All values are mean±s.e.m. For currents from HEK293 patches, n=4 at each concentration of β2N or β2N^ΔFIW peptide. For currents from daytime SCN neurons, BK_i (n=18) and β2 KO+β2N (n=10).

Figure 6. β2N rescues inactivation and restores daytime BK current levels in β2 KO SCN neurons.

(a) Representative traces showing rescue of the macroscopic BK current decay in β2 KO neurons with 50 μM β2N applied intracellularly. 50 μM β2N^ΔFIW did not rescue the current decay. Voltage protocol same as in Fig. 2. (b)Application of β2N reduced the BK current density in β2 KO neurons to levels comparable to WT. β2N^ΔFIW had no effect on BK current levels. WT and β2 KO I–Vs re-plotted from Figs 2 and 4 for cross-comparison. n values: WT (27); β2 KO (20); β2 KO+β2N^ΔFIW (18); β2 KO+β2N (20). All values are mean±s.e.m. *P<0.05, Bonferroni post hoc.

Figure 7. Loss of β2 eliminates the diurnal difference in firing rate, and rescue of inactivation with β2N restores daytime firing rates in SCN neurons.

(a) Spontaneous action potential activity from representative day (BK_i) and night (BK_s) WT neurons. Dotted line (a,c,d) denotes −50 mV. (b) In WT SCNs, BK_i neurons fired at higher frequencies than BK_s, similar to the average day–night difference in firing. β2 KO neurons did not exhibit a diurnal difference in frequency, and during the day, fired at levels similar to WT night. Application of β2N^ΔFIW to daytime β2 KO neurons had no effect on frequency, but β2N increased firing rate to WT levels. (c) Day (BK_s) and night (BK_s) β2 KO neurons. (d) Day β2 KO neurons with 50 μM β2N (BK_i) or 50 μM β2N^ΔFIW (BK_s). All values are mean±s.e.m. n values: WT: BK_i (17), BK_s (10), day (17), night (20); β2 KO: day (19), night (19), β2N^ΔFIW (20), β2N (19), β2N/pax (8), and pax (8).

Figure 8. β2N can confer inactivation to WT neurons at night.

(a) Representative macroscopic traces from WT neurons at night showing a typical BK_s current, and a BK_i current resulting from application of 50 μM β2N. Voltage protocol same as in Fig. 2. (b) β2N reduced the night time current density in WT neurons to daytime levels. WT day and night data re-plotted from Figs 2 and 4 for cross-comparison. (c) The proportion of BK_i currents increased significantly with β2N (P=0.0001, Fisher's exact test). (d) β2N increased night time firing to daytime levels. WT day and night data re-plotted from Fig. 7 for cross-comparison. All values are mean±s.e.m. n Values: WT, day (27); WT, night (22); and WT+β2N, night (20). *P<0.05, Bonferroni post hoc.

Figure 9. BK current inactivation alters membrane potentials and input resistance in SCN neurons.

(a,b) Resting membrane potentials (a) and input resistance (b) measured in 1 μM tetrodotoxin to block action potentials. Conditions with inactivation (BK_i, β2 KO/β2N, and WT day overall) had more depolarized V_m and higher R_i than BK_s, night, or β2 KO. All values are mean±s.e.m. n values: WT: day/BK_i (9), day/BK_s (5), day (25), night (18) and β2 KO: β2N^ΔFIW/day (11), β2N^ΔFIW/night (10), β2N/day (12). *P<0.05, Bonferroni post hoc.

Figure 10. Steady-state and action potential-evoked BK_i currents are reduced compared to BK_s in SCN neurons.

(a) Representative macroscopic BK current traces from daytime WT neurons in response to a subthreshold voltage protocol (−60 to −30 mV in 5 mV steps). BK currents were identified as BK_i or BK_s from a maximally activating step to +90 mV (as in Fig. 2). (b) Current–voltage relationship showing significantly more activation of BK_s current above −60 mV compared to BK_i. (c) Pre-recorded action potential commands (top) were used to elicit BK currents from BK_i or BK_s neurons. From a holding potential of −150 mV to remove inactivation completely, cells were stepped to the inter-spike potential (−48 mV) with a sequence of three action potential commands as depicted (Peak, 8 mV; t_1/2, 5.5 ms; and AHP/antipeak, −54 mV). Arrow, the subthreshold current level in d was taken just before the second action potential command. (d,e) Average subthreshold BK current density from the inter-spike interval (d) or at the peak of the action potential (e). n values for (a–d): WT: day/BK_i (9), day/BK_s (5). (f) BK current as a function of conditioning potential in SCN neurons. Representative macroscopic BK current traces from β2 KO neurons during the day with 50 μM β2N^ΔFIW or β2N. Neurons were held at −90 mV for 150 ms to remove inactivation, and then stepped to a conditioning potential for 100 ms to allow channels to transition into the inactivated state, followed by a maximally activating step to +90 mV. (g) The peak BK current elicited from the +90 mV step was plotted as a function of the conditioning potential. β2N causes a reduction in current, but no reduction is observed in the absence of inactivation with β2N^ΔFIW. n=5 for each condition. All values are mean±s.e.m. *P<0.05, Bonferroni post hoc.