β-Secretase BACE1 regulates hippocampal and reconstituted M-currents in a β-subunit-like fashion

Abstract

The β-secretase BACE1 is widely known for its pivotal role in the amyloidogenic pathway leading to Alzheimer's disease, but how its action on transmembrane proteins other than the amyloid precursor protein affects the nervous system is only beginning to be understood. We report here that BACE1 regulates neuronal excitability through an unorthodox, nonenzymatic interaction with members of the KCNQ (Kv7) family that give rise to the M-current, a noninactivating potassium current with slow kinetics. In hippocampal neurons from BACE1(-/-) mice, loss of M-current enhanced neuronal excitability. We relate the diminished M-current to the previously reported epileptic phenotype of BACE1-deficient mice. In HEK293T cells, BACE1 amplified reconstituted M-currents, altered their voltage dependence, accelerated activation, and slowed deactivation. Biochemical evidence strongly suggested that BACE1 physically associates with channel proteins in a β-subunit-like fashion. Our results establish BACE1 as a physiologically essential constituent of regular M-current function and elucidate a striking new feature of how BACE1 impacts on neuronal activity in the intact and diseased brain.

Keywords: Alzheimer's disease; BACE1; KCNQ; M-current; epilepsy; hippocampus.

Figure 1.

Reduced M-current (I_M) accounts for enhanced excitability of hippocampal CA1 pyramidal neurons from BACE1^−/− mice. Whole-cell current-clamp recordings from CA1 pyramidal neurons in hippocampal slices demonstrated that the characteristic frequency adaptation that gradually slows repetitive firing during sustained depolarization under normal conditions (A) is strongly attenuated in the absence of BACE1 (B). For comparison, membrane potentials were set to −70 mV before depolarizing current of equal size (50 pA) was injected. Histograms quantify number of APs (C) and time to first AP (D) during depolarizing steps of increasing amplitude for wt neurons (white columns, n = 33) and BACE1-deficient neurons (red columns, n = 25). *p < 0.05 (unpaired t test). Whereas XE991 (10 μm) reliably impaired frequency adaptation in wt neurons (E), the blocker failed to excite BACE1-deficient neurons (F). Again, membrane potential was set to −70 mV before firing. The amplitude of injected current was adjusted to elicit 4–5 APs during a 300 ms pulse before XE991 application. Histograms quantify number of APs (G) and time to first AP (H) during step depolarization for wt neurons (n = 8) and BACE1-deficient neurons (n = 5). *p < 0.05 (paired t test).

Figure 2.

Acutely isolated CA1 pyramidal cells of BACE1^−/− mice show reduced I_M. A, The significant difference in zero-current potential between wt and BACE1-deficient neurons disappeared in the presence of 10 μm XE991. ΔV, Drug-induced voltage shifts. Inset, Acutely isolated neuron. B, Small hyperpolarizing voltage steps from −30 mV (inset) were used to interrogate I_M of wt (black traces) and BACE1^−/− neurons (red traces) in the absence and presence of 10 μm XE991 (blue traces). Traces above the latter represent XE991-subtracted current at −45 mV with fitted curve superimposed. At all voltages examined, the XE991-sensitive current of BACE1^−/− neurons (red data points) had smaller amplitudes (C) and faster deactivation kinetics (D) than that of wt neurons (black data points). Wt, n = 11; BACE1^−/−, n = 12. *p < 0.05, **p < 0.01, ***p < 0.01, paired or unpaired t test, as applicable.

Figure 3.

Coexpression of BACE1 increases whole-cell current through heteromeric KCNQ2/Q3 channels in voltage-clamped HEK293T cells. Typical KCNQ2/Q3 current responses to depolarizing voltage steps of increasing amplitude (see insets) were obtained in the absence (A) or presence of BACE1 (B). C, Using the set of experiments illustrated in A, B, current–voltage (I–V) relationships were determined in which current amplitudes at the end of each step were plotted as function of test potentials in the absence (black data points) and presence of BACE1 (red data points) or its catalytically inactive variant BACE1 D289N (green data points). D, I-V curves of C were transformed to activation curves, in which normalized conductance was plotted as a function of voltage (see Materials and Methods). Half-activation voltages (V_mid) were −22.3 ± 0.3 mV for KCNQ2/Q3 alone (black curve), −26.1 ± 0.5 mV when BACE1 was coexpressed (red curve), and −25.3 ± 0.9 mV when BACE1 D289N was coexpressed (green curve). As indicated by the like-colored asterisks, both BACE1 and BACE1 D289N caused a small, but significant, leftward shift of current activation. Deactivation of KCNQ2/Q3 currents in the absence (E) and presence of coexpressed BACE1 (F) was determined using a stepwise repolarization protocol (see inset) in high external K⁺ solution. G, Current decay was fitted using a mono-exponential function to determine voltage-dependent deactivation kinetics of KCNQ2/Q3 current alone (black data points) and in the presence of coexpressed BACE1 (red data points) or BACE1 D289N (green data points). Like-colored asterisks indicate significant slowing of KCNQ2/Q3 current deactivation when either BACE1 or its inactive variant was present. H, BACE1 and BACE1 D289N accelerated current activation. Activation time constants were obtained by fitting a mono-exponential time course to the rising phase of the current trace in activation experiments of A, B. KCNQ2/Q3, n = 153 (C, D, H), n = 138 (G); +BACE1, n = 116 (C, D, H), n = 77 (G); +BACE1 D289N, n = 32 (C, D, G, H). *p < 0.05, **p < 0.01, ***p < 0.01, pairwise Mann–Whitney test with Bonferroni correction.

Figure 4.

Coexpression of BACE2 in lieu of BACE1 does not alter KCNQ2/Q3 currents of HEK293T cells. All parameters that were affected by BACE1 in previous recordings (red data points) and that are reproduced here from Figure 3 for comparison together with the control data (black data points) remained unchanged in the presence of BACE2 (purple data points), including I-V relationship (A), voltage dependence of activation (B), time constants of activation (C), and deactivation (D). KCNQ2/Q3 + BACE2, n = 27 (A–C), n = 26 (D). *p < 0.05, **p < 0.01, ***p < 0.01, pairwise Mann–Whitney test with Bonferroni correction.

Figure 5.

BACE1 amplifies activation of heterologously expressed KCNQ2/Q3 currents during simulated firing patterns. Voltage trajectories obtained from spontaneous (A) or evoked firing (B) of CA1 pyramidal cells were used as voltage commands in transfected HEK293T cells. Coexpression of BACE1 strongly enhanced KCNQ2/Q3 currents in a frequency-dependent fashion. Note much slower time scale in A than in B. For quantification, current amplitudes were averaged over the last 50 ms of stimulation (A, B, arrows) from recordings performed at room temperature (C) or at 32°C (D). White columns represent KCNQ2/Q3 alone, n = 19 (room temperature), n = 23 (32°C); red columns represent KCNQ2/Q3 and BACE1, n = 17 (room temperature), n = 36 (32°C). **p < 0.01, ***p < 0.01, Mann–Whitney test.

Figure 6.

Effects of BACE1 on surface and total levels of KCNQ2 and KCNQ3. A, Western blot of a surface biotinylation of HEK293T cells transfected with KCNQ2 and KCNQ3 with coexpression of BACE1 as indicated above the blot. EGFP and MaxiK served as transfection marker and control, respectively. B, The surface levels of the channel proteins were normalized to the corresponding surface levels of Na⁺/K⁺-ATPase or cadherin, with the normalized levels determined in the absence of BACE1 set to 1. The histogram shows channel surface levels with BACE1 relative to lanes without BACE1, using antibodies against KCNQ2 and KCNQ3 as indicated above the columns: white column represents transfection with KCNQ2 and KCNQ3; red column represents transfection with KCNQ2, KCNQ3, and BACE1. C, Total protein bands were quantified and normalized to β-actin. n = 10. **p < 0.01, one-sample t test.

Figure 7.

Effects of BACE1 coexpression on activation (A, B, left) and deactivation (A, B, right) of currents mediated by homomeric KCNQ2 (A) or KCNQ3 channels (B) in HEK293T cells. From activation protocols, we constructed I-V relationships (C) and activation curves (D). For comparison, I-V curves and activation curves also include previous results from heteromeric KCNQ2/Q3 (black data points and traces). Half-activation voltages (V_mid) were −14.7 ± 1.2 mV for KCNQ2 alone, −17.8 ± 1.2 mV for KCNQ2 and BACE1, −31.1 ± 1.5 mV for KCNQ3 alone, and −32.7 ± 1.2 mV for KCNQ3 and BACE1. KCNQ2, n = 33; KCNQ3, n = 29; KCNQ2 and BACE1, n = 38; KCNQ3 and BACE1, n = 34. *p < 0.05, **p < 0.01, ***p < 0.01, pairwise Mann–Whitney test with Bonferroni correction. Coexpression of BACE1 strongly upregulated currents through homomeric KCNQ4 (E) and homomeric KCNQ5 (F) channels in HEK293T cells. KCNQ4, n = 34; KCNQ4 +BACE1, n = 38; KCNQ5, n = 24; KCNQ5 +BACE1, n = 29. ***p < 0.01, Mann–Whitney test. n.s., Not significant.

Figure 8.

Effects of PIP₂ and retigabine on KCNQ2/Q3 in the absence and presence of BACE1. A, To deplete PIP₂ in a controlled fashion, the voltage-dependent phosphatase DR-VSP was coexpressed with KCNQ2/Q3 alone or together with BACE1 in HEK293T cells. KCNQ2/Q3 currents were measured before (I₁) and after the phosphatase was activated by strong depolarization of variable length (I₂). B, Time course of changes in I₂/I₁ ratio is indicative of current run-down as phosphatase was activated (n = 15, blue data points and lines). Coexpression of BACE1 did not attenuate phosphatase-mediated current decline (n = 19, purple data points and lines). No changes in I₂/I₁ ratio were observed for currents mediated by KCNQ2/Q3 alone (n = 15, black data points and lines), by KCNQ2/Q3 and BACE1 (n = 20, red data points and lines), by KCNQ2/Q3 and the inactive phosphatase DR-VSP-C302S (n = 15, gray data points and lines), or by KCNQ2/Q3, BACE1, and DR-VSP-C302S (n = 17, green data points and lines). C, Responsiveness of KCNQ2/Q3 currents to retigabine (10 μm) is preserved in the presence of BACE1 as indicated by the characteristic drug-induced leftward shift of the activation curves. Maximum current was not enhanced by retigabine (KCNQ2/Q3, n = 40; +BACE1, n = 43). D, The boosting effect of BACE1 on KCNQ2 current is not impaired when BACE1 is coexpressed with the retigabine-insensitive mutant W236L (KCNQ2 W236L, n = 26; +BACE1, n = 30). *p < 0.05, **p < 0.01, ***p < 0.01, Mann–Whitney test. n.s., Not significant.

Figure 9.

Direct interaction of BACE1 with KCNQ2 and KCNQ3 in HEK293T cells. A, Western blot of a coimmunoprecipitation of KCNQ2 and BACE1 (n = 3). B, Western blot of a coimmunoprecipitation of KCNQ3 with BACE1 (n = 5). C–E, PLA was performed on HEK293T cells expressing KCNQ2-V5 and BACE1 (C) or ENaC-V5 and BACE1 (D). Left, Unprocessed data (grayscale, 8 bits). Protein–protein proximity is indicated by bright fluorescent spots. Images were automatically analyzed using the CellProfiler program (right). Nuclei were counterstained with DAPI (gray channel). In a first step, nuclei (white outlines) were identified. The “cell borders” (green outlines) were expanded beyond corresponding PLA signals (orange blobs), if present. In-focus PLA signals were detected (yellow outlines) and counted. Adjacent PLA signals were separated using local intensity maxima. Scale bar, 5 μm. E, Number of PLA signals per cell. n = 293 (ENaC-V5 + BACE1), n = 288 (KCNQ2-V5 + BACE1), n = 307 (KCNQ2-V5 + KCNQ3-HA), n = 312 (KCNQ2-V5 + ENaC-HA) from two independent transfections each. ***p < 0.01, Kruskal–Wallis followed by Mann–Whitney test with Bonferroni correction.