Research Article

BKCa-Cav Channel Complexes Mediate Rapid and Localized Ca2+-Activated K+ Signaling

See allHide authors and affiliations

Science  27 Oct 2006:
Vol. 314, Issue 5799, pp. 615-620
DOI: 10.1126/science.1132915

Abstract

Large-conductance calcium- and voltage-activated potassium channels (BKCa) are dually activated by membrane depolarization and elevation of cytosolic calcium ions (Ca2+). Under normal cellular conditions, BKCa channel activation requires Ca2+ concentrations that typically occur in close proximity to Ca2+ sources. We show that BKCa channels affinity-purified from rat brain are assembled into macromolecular complexes with the voltage-gated calcium channels Cav1.2 (L-type), Cav2.1 (P/Q-type), and Cav2.2 (N-type). Heterologously expressed BKCa-Cav complexes reconstitute a functional “Ca2+ nanodomain” where Ca2+ influx through the Cav channel activates BKCa in the physiological voltage range with submillisecond kinetics. Complex formation with distinct Cav channels enables BKCa-mediated membrane hyperpolarization that controls neuronal firing pattern and release of hormones and transmitters in the central nervous system.

Large-conductance Ca2+- and voltage-activated K+ channels (BKCa or KCa1.1) are fundamental modulators of neuronal signaling (1, 2) by contributing to action potential repolarization (3, 4), mediating the fast phase of afterhyperpolarization (3, 58), controlling dendritic Ca2+ spikes (9), and establishing a feedback loop between membrane potential and cytosolic Ca2+ that regulates release of hormones and transmitters (1013).

The physiological functions of BKCa channels arise from their unique allosteric activation by two distinct stimuli, membrane depolarization and cytosolic Ca2+ ions (1416). Increasing Ca2+ concentrations ([Ca2+]i) shift the depolarization required for channel opening into the physiological voltage range. In fact, [Ca2+]i of ≥ 10 μMare usually required for activating BKCa channels at membrane potentials around 0 mV (17). In central nervous system (CNS) neurons, such high levels of [Ca2+]i are tightly restricted in time and space to local “Ca2+-signaling domains” centered around voltage-activated Ca2+ (Cav) channels (18, 19). In these domains, speed and magnitude of Ca2+ signals are inversely related to the distance from the Ca2+ source and are assessed experimentally by the distinct properties of the Ca2+ chelators EGTA and BAPTA. Thus, Ca2+-sensitive processes affected by millimolar concentrations of BAPTA but not EGTA are assumed to be placed within ∼20 nm from the Cav channels (nanodomain), while processes with an equal BAPTA/EGTA sensitivity are located between 20 and 200 nm (microdomain) or even further away from the Ca2+ source (18).

Functional characterization in various types of neurons provided two hallmarks for the activation of BKCa under normal conditions. First, BKCa channels reside in close spatial proximity to Cav channels, as they were robustly activated by Ca2+ influx through the Cav channels in the presence of EGTA, whereas BAPTA at millimolar concentrations largely attenuated or abolished the functional channel-channel coupling (3, 13, 2022). Second, BKCa channels appear to be selectively activated by a subset of Cav channels with distinct functional properties and subcellular distribution. Thus, P/Q-, N- and L-type Cav channels activate BKCa either selectively or concertedly in nerve terminals, dendrites, or somata of various types of CNS neurons (3, 6, 13, 20, 23, 24).

Despite its fundamental importance for the physiology of BKCa channels, the mechanism underlying the intimate and selective association between BKCa and Cav channels is as yet unknown, and selective coupling between BKCa and Cav channels in heterologous expression systems has not been demonstrated.

Affinity Purification of BKCa Channel Complexes from Rat Brain

We used affinity purifications (AP) with two different BKCaα subunit-specific antibodies (anti-BKα and anti-BKα* Abs) on solubilized plasma membrane–enriched protein fractions prepared from total rat brain (25). Separation by blue native and subsequent denaturing gel electrophoresis showed that these protein fractions contained high-molecular-weight complexes of BKCa channels [Fig. 1A, (26)]. Total eluates obtained in APs with the two anti-BKα Abs and with several immunoglobulin G (IgG) pools (preimmunization IgGs and antibodies unrelated to BKα) serving as a control were subjected to analysis by nanoflow liquid chromatography tandem mass spectrometry (nano-LC MS/MS) (Fig. 1A). This approach identified the α subunit of BKCa channels (BKα) by retrieving ≥ 66 different peptide fragments (for each anti-BKα Ab) covering ∼75% of the BKα primary sequence (Fig. 1, B and C and Table 1, top). In addition, MS/MS spectra from the anti-BKα eluates unambiguously identified the two BKβ subunits (BKβ2 and BKβ4) expressed in the CNS (17), as well as several Cav channel α and β subunits (Table 1, top). The Cavα subunits specifically retained by both anti-BKα Abs were Cav1.2, Cav2.1 and Cav2.2 (Table 1 and fig. S1) (26) encoding the pore-forming subunits of the L-, P/Q- and N-type Cav channels, respectively (27). In fact, Cav2.1 was the protein most abundantly copurified with BKα; all together, MS/MS analyses detected 43 different peptides covering ∼44% of the Cav2.1 amino acid sequence. Similar sequence coverage was obtained for the specifically copurified Cavβ subunits Cavβ1b, Cavβ2, and Cavβ3 (Table 1 and fig. S1). In contrast, Cav2.3, R-type Cav channels (27), and the Cavβ4 subunit were detected in the eluates from both anti-BKα Abs and control IgGs with similar abundance (Table 1 and fig. S1).

Fig. 1.

Affinity purification of BKCa channel complexes from CNS plasma membranes. (A) (Top) High-molecular-weight complexes of BKCa channels in solubilized plasma membrane–enriched protein fractions from rat brain visualized by two-dimensional gel electrophoresis (26). Solubilized protein complexes were separated by blue native polyacrylamide gel electrophoresis (BN-PAGE, 1. dimension) and denaturing SDS-PAGE (2. dimension), and were subsequently Westernprobed with anti-BKα (42). BKCa channels (4BKα + 4BKβ) indicated by filled arrow-head, main fraction of high-molecular weight complexes of BKCa channels denoted by open arrowhead. (Bottom) Eluates of APs with two different anti-BKα Abs (anti-BKα and anti-BKα*) or a pool of pre-immunization IgGs. Total eluates were shortly run into a SDS-gel, in-gel trypsinized, and analyzed by nano-LC MS/MS spectrometry. (B) (Top) High-performance liquid chromatography (HPLC) chromatogram of peptides obtained by trypsinization of anti-BKα eluates in the mass range of 50 to 150 kD. (Middle and bottom) MS and MS/MS spectra of a peptide unique for BKα. The peptide had a mass/charge (m/z) ratio of 552.76154 and was eluted at the time point indicated. In the MS/MS spectrum, the complete y+-ion series is indicated, and the amino acid sequence derived from the mass differences is given in carboxy-to-amino-terminal direction. (C) Coverage of the BKα amino acid sequence by the peptides identified with nano-LC MS/MS. Peptides identified by mass spectrometry are in red, those accessible to but not identified in MS/MS analyses are in black, and peptides not accessible to the MS/MS analyses used are in gray. Lines denote hydrophobic segments S0 to S10; the blue box highlights the BKα peptide shown in (B).

Table 1.

(Top) BKCa and Cav channel subunits affinity-purified with anti-BKα from CNS plasma membranes and identified by nano-LC MS/MS. (Bottom) BKCa and Cav channel subunits affinity-purified with anti-Cav1.2 from CNS plasma membranes. Procedures used for affinity-purification and mass spectrometry, as well as the criteria for protein identification, are detailed in (26). remSC is relative exponentially modified sequence coverage. rPQ-Score is relative protein query score. Values for remSC > 5 and rPQ-Scores > 4 indicate specific purification by anti-BKα or anti-Cav1.2 over control IgG pools. * indicates lower estimates of the rPQ score with no matching peptide fragments in the controls.

Protein IDremSCrPQ-Score
BKCa subunits BKα (KCa1.1,KCNMA1) 48.1 322.3
BKβ2 (KCNMB2) 16.0*
BKβ4 (KCNMB4) 112.0*
Cav subunits Cav2.1 (α1A) 496.0*
Cav1.2 (α1C) 40.0*
Cav2.2 (α1B) 5.5 4.9
Cav2.3 (α1E) 3.8 2.4
Cav β1b 124.0*
Cav β2 100.0*
Cav β3 6.3 4.1
Cav β4 2.3 2.5
Cav subunits Cav1.2 (α1C) 216.0*
Cav β1b 128.0*
Cav β2 104.0*
Cav β3 104.0*
Cav β4 6.3 5.0*
BKCa subunits BKα 24.0*

Coassembly with Cav1.2 channels was confirmed by subsequent reverse purification using an antibody specific for the Cav1.2 subunit (anti-Cav1.2) and suitable for AP from rat brain plasma membranes. As illustrated in Fig. 2A by the ion chromatogram (left) and the MS/MS spectrum (right) of one out of the eight unique peptides obtained, BKα was copurified by anti-Cav1.2 but not by the control IgG pools (Table 1, bottom).

Fig. 2.

Co-assembly of BKCa and Cav channels in the CNS and heterologous expression systems. (A) Identification of BKCa in an AP from CNS plasma membranes with anti-Cav1.2. Ion chromatogram (left panel, reflecting the signal generated by the m/z ratio of 552.76154 in the mass spectrometer during elution from the HPLC) and corresponding MS/MS-spectrum (right panel) of one of the eight BKα-specific peptides obtained (the same peptide as in Fig. 1B). The m/z ratio of 552.76154 (±10 parts per million) was not detected in elutions from APs with preimmunization IgG pools serving as a control. (B) Copurification of BKCa and Cav channels from culture cells expressing BKCa (BKα and BKβ4) and Cav1.2 (Cav1.2α, Cavβ1b, and α2δ) channels. Eluates from APs with anti-BKα (left) or anti-Cav1.2 (right) Abs were separated by SDS-PAGE and Western-probed as indicated. Specificity of APs with anti-BKα and anti-Cav1.2 were verified in purifications with cells expressing Cav1.2 (left) or BKCa (right) channels alone.

Coassembly of Heterologously Expressed BKCa and Cav Channels

The copurification of BKCa with specific Cav channel subtypes from rat brain plasma membranes was reproduced by APs from culture cells that heterologously expressed BKCa channels and either Cav1.2 or Cav2.1 channels. For these experiments, the respective channel subunits BKα and BKβ4, as well as Cav1.2 or Cav2.1, Cavβ1b or Cavβ3, and α2δ (28), were transfected into culture cells or injected as cRNAs into Xenopus oocytes (26). Figure 2B and fig. S2A illustrate the results of coimmunoprecipitations using anti-BKα and anti-Cav1.2 on the BKCa-Cav1.2 coexpressions in culture cells. Thus, anti-BKα effectively and specifically retained the Cav1.2 subunit, and anti-Cav1.2 coprecipitated the BKα subunit with similar efficiency. An equivalent result was obtained from an AP using anti-BKα on Xenopus oocytes coexpressing BKCa and Cav2.1 channels. In this experiment, copurification of the Cav channel was verified by MS/MS analysis that retrieved 31 peptides for BKα and 9 and 11 peptides specific for the Cav2.1 and Cavβ3 subunits, respectively (fig. S2B).

Functional Characteristics of BKCa-Cav Complexes

Next, the functional properties of BKCa-Cav2.1 channel complexes were investigated in giant inside-out patches (26) excised from Xenopus oocytes coexpressing BKCa (BKα and BKβ4) and run-down deprived Cav2.1 channels [Cav2.1(I1520H), Cavβ3, and α2δ (29)]. Figure 3A shows typical current traces recorded in response to step depolarizations in the physiological voltage range. Thus, for voltage steps above the activation threshold of Cav2.1 channels, current responses were biphasic (in all 135 patches tested): an initial Ca2+ inward current that was followed by an outward K+ current, as would be expected for activation of BKCa channels by influx of Ca2+ ions through the Cav2.1 channels (Fig. 3A). The coupling between both channels was mandatory for activation of BKCa in the physiological voltage range as shown by control experiments lacking expression of the Cav channels (Fig. 3B). The time course of the Cav2.1-BKCa coupling reflected by the interval between the onsets of Ca2+ and K+ currents (26) was voltage-dependent and markedly reduced by membrane depolarization. At potentials positive to 0 mV, this time interval was less than one millisecond (Fig. 3A, inset); mean values for the duration (n = 15 experiments) were 0.75 ± 0.24 ms (at 20 mV) and 0.95 ± 0.18 ms (at 10 mV). The amplitude of the Cav channel–activated BKCa current was strongly voltage-dependent and exhibited a bell-shaped steady-state current-voltage (I-V) relation with a peak amplitude at about 20 mV (Fig. 3B). Both the shape of the I-Vrelation and the time course of the Cav2.1-BKCa coupling are a reflection of several factors, including the voltage-dependent gating properties of Cav2.1 (Fig. 3C and fig. S3A) and BKCa channels, the amplitude of the Ca2+ current (Fig. 3C), and the Ca2+ sensitivity of BKCa channels (Fig. 3D). In particular, increase in open probability and faster activation kinetics of the Cav channels promote accelerated coupling (Fig. 3A) and increase in the BKCa currents (Fig. 3B), whereas the decrease in Ca2+ current amplitude (due to a reduced driving force at voltages approaching the Ca2+ reversal potential) leads to reduction and cessation of BKCa currents at voltages >20 mV (Fig. 3B). Similar results for coupling and I-V relation were obtained with Cav2.2 (Cav2.2α,Cavβ3, and α2δ) channels (fig. S3B).

Fig. 3.

Functional coupling of heterologously expressed BKCa and Cav2.1 channels. (A) Current response to the indicated voltage steps (–50 to 20 mV at 10 mV increments, holding –80 mV) recorded under physiological ion conditions(1.3 mM extracellular Ca2+) in a giant inside-out (i-o) patch excised from an oocyte coexpressing BKCa (BKα and BKβ4) and Cav2.1[Cav2.1(I1520H), Cavβ3, and α2δ] channels. Cytoplasmic solution was buffered with 5 mM EGTA. Current scale is 1 nA. (Inset) Current traces in black (–50, –20, –10, 0, and 20 mV) at expanded time scale. (B) Normalized (outward) currents through BKCa channels as a function of membrane potential recorded in excised i-o patches from oocytes expressing BKCa and Cav2.1 channels (filled symbols) or BKCa channels alone (open symbols). Data points are mean ± SD of 10 experiments [gray triangles are from the experiment in (A)]. (C) Ca2+ (inward) currents normalized to maximum and recorded under conditions as in (A) in excised i-o patches from oocytes expressing Cav2.1 channels as in (A). (Inset) Representative experiment, traces at –10 and 20 mV are in black. (D) Steady-state activation curves of BKCa channels recorded at the [Ca2+]i indicated in giant i-o patches from oocytes. Data points are mean ± SD of 6 experiments. Gray bar denotes the voltage range of BKCa channel activation by coexpressed Cav2.1 channels from (A) and (B). Continuous lines are fits of Boltzmann functions to the data with values for V½ and slope factor of 197.8 mV and 27.1 mV (0[Ca2+]i), 123.2 mV and 32.1 mV (1 μM[Ca2+]i), –19.9 mV and17.1 mV (10 μM[Ca2+]i), and –58.1 mV and 17.5 mV (100 μM[Ca2+]i).

Comparison of BKCa currents evoked by Cav2.1 and those recorded in excised patches from oocytes expressing only BKCa with defined [Ca2+]i (26) was used to estimate the Ca2+ concentration delivered to BKCa via the Cav channels. The robust BKCa activity observed at membrane potentials ≤0 mV (Fig. 3B) suggested that Cav2.1 channels might provide Ca2+ concentrations of ≥10 μM (Fig. 3D). This value was confirmed by the time constants obtained from monoexponential fits to the activation time course (τactivation) of BKCa currents. The values determined for τactivation of Cav2.1-evoked BKCa currents were similar to the results obtained from BKCa currents at a [Ca2+]i of 10 μM across the entire voltage range tested (Fig. 4A).

Fig. 4.

Localization of BKCa and Cav2.1 channels within Ca2+ nanodomains. (A) Activation time constants (τactivation) of BKCa channels activated either by coexpressed Cav2.1 channels (red circles; experiments as in Fig. 3A) or by [Ca2+]i of 10 and 100 μM (black circles and squares, respectively) together with τactivation of Cav2.1 channels. Data points are mean ± SD of 12 (BKCa/Cav2.1), 6 (BKCa), and 5 (Cav2.1) experiments. (B) Normalized currents through BKCa channels recorded at 20 mV with the indicated buffers present on the cytoplasmic side of i-o patches from oocytes coexpressing BKCa and Cav2.1. Data points are mean ± SD of 9 experiments. (Inset) Representative current traces with 10 mM EGTA (black), 5 mM BAPTA (blue), or 10 mM BAPTA (red). (C) Steady-state Ca2+ concentration profiles at the cytoplasmic opening of a single Cav channel. The profiles were determined with the CalC software v. 5.4.0 (43) (single channel conductance of1.7 pS, driving force of 60 mV). The gray bar represents the range fitting the experimental data shown in Figs. 3 and 4.

The BKCa-Cav2.1 coupling was further probed for its sensitivity to Ca2+-buffers. Excised patches were successively perfused with intracellular solutions containing EGTA and BAPTA in 5 mM and 10 mM concentration. Cav channel–activated BKCa currents (at 20 mV) were unaffected by 10 mM EGTA (with respect to the standard 5 mM EGTA), whereas 5 mM and 10 mM BAPTA reduced their amplitude to 28 ± 5% (mean ± SD of 9 patches) and 10 ± 4%, respectively (Fig. 4B). This reduction resultedfromadecreased[Ca2+]i at the BKCa channels, as indicated by their markedly slowed activation time course (Fig. 4B, inset, and fig. S3C). The τactivation values determined for a membrane potential of 20 mV were 13.4 ± 2.1 ms (mean ± SD, n = 9) and 19.9 ± 2.1 ms (n =9) for 5 mM and 10 mM BAPTA, whereas in EGTA the respective value was 6.9 ± 1.1 ms (n = 12; all values were significantly different from each other, with P < 0.0005, pairwise Student's t test). Other properties of Cav2.1-activated BKCa currents, including the bell-shaped steady-state I-V, were similar in BAPTA- and EGTA-buffered intracellular solutions (fig. S3).

The distinct effects of EGTA and BAPTA on the Cav-activated BKCa currents place the channels within a “local nonequilibrium Ca2+ domain,” a steep Ca2+ concentration gradient around a Cav channel that rapidly builds up after opening of the channel pore in the presence of mobile buffers (18, 19). Figure 4C depicts such Ca2+ concentration profiles simulated for a Cav channel with a single-channel conductance of 1.7 pS [(30), for 1.3 mM external Ca2+] and EGTA or BAPTA at concentrations of 5 mM and 10 mM. Accordingly, the distance between BKCa and Cav2. 1 channels fitting the data on both amplitude and activation time course of BKCa currents may be estimated to ∼10 to 15 nm.

Specificity of BKCa-Cav Channel Complex Formation

Cav1.2, Cav2.1, and Cav2.2 were specifically co-purified with BKCa channels from rat brain, whereas Cav2.3 was not (Table 1 and fig. S1). We, therefore, investigated the specificity of BKCa-Cav coassembly using coexpression of either Cav1.2 or Cav2.3 (plus Cavβ1b and α2δ) with BKCa (BKα and BKβ4) in Chinese hamster ovary (CHO) cells. For functional recordings, the patch-clamp technique was used in whole-cell configuration (26). Figure 5A shows representative current responses to depolarizing voltage steps recorded from coexpression of BKCa with either of the two Cav subtypes after equilibration of the intracellular milieu with the pipette solution. As indicated by the biphasic current response and the bell-shaped I-V relation, Cav1.2 channels effectively activated the coexpressed BKCa channels [Fig. 5, A (top) and B] (n of 48 cells), similar to the Cav2.1-BKCa coexpression in oocytes. Again, the Ca2+ provided through the Cav channels was mandatory for the BKCa currents as reflected by their complete decay paralleling the pronounced inactivation of the Cav1.2 channels (Fig. 5A and fig. S4A). The Cav1.2-BKCa coupling could not be disrupted with EGTA but was reversibly disrupted after replacing EGTA in the pipette solution with BAPTA (n = 13 cells) (fig. S4C).

Fig. 5.

BKCa-Cav coassembly is subtype-specific and determined by the α subunits. (A) Current response to the indicated voltage steps (–50 to 30 mV at 10 mV increments) recorded under physiological ion conditions in whole CHO cells coexpressing BKCa (BKα and BKβ4) and either Cav1.2 (top) or Cav2.3 (bottom) channels. Cav channels were assembled from Cav1.2α/Cav2.3α,Cavβ1b, and α2δ. Traces recorded upon a voltage step to 20 mV are highlighted in red and superimposed in the inset with the response of the respective Cav subtype to the same voltage step. Time constants (mean ± SD) for the decay of BKCa and Cav1.2 currents obtained from monoexponential fits were 28.4 ± 9.9 ms and 27.6 ± 11.7 ms, respectively. (B) I-V relation of the experiments in (A); symbols as indicated. (C) Activation of BKCa by Cav2.3 channels is abolished by 5 mM EGTA washed into CHO cells after establishing whole-cell configuration. Current responses recorded upon voltage steps to 30 mV0.25 min (light) and 3 min (dark) after whole-cell break-in into cells coexpressing BKCa channels with either Cav2.3 (red) or Cav1.2 (black) channels. (D) Failure of anti-BKα to purify the Cav2.3α subunit from culture cells coexpressing BKCa (BKα and BKβ4) and Cav2.3 channels (Cav2.3α,Cavβ1b, and α2δ). Load and anti-BKα eluate were separated by SDS-PAGE and Western-probed by anti-BKα and an antibody against the Cav2.3α subunit (anti-Cav2.3). (E) Copurification of BKCa and Cav channel α subunits from culture cells expressing BKα and Cav1.2α. Eluates from APs with anti-BKα (left) and anti-Cav1.2 (right) Abs were separated by SDS-PAGE and Western-probed as indicated.

In marked contrast to Cav1.2, Cav2.3 channels failed to promote activity of the coexpressed BKCa channels under standard conditions (n = 9 cells) (Fig. 5, A and B). This failure was not due to an inefficient expression of BKCa but rather resulted from equilibration of the cytoplasm with EGTA as monitored in a series of experiments applying step-depolarizations every 30 to 45 s (Fig. 5C, D). Thus, immediately after establishing whole-cell configuration, Ca2+ influx through Cav2.3 channels elicited robust BKCa currents that vanished over the following 3-min period required for diffusion of EGTA into the CHO cell (Fig. 5D) (n = 8 cells). The Ca2+ currents through Cav2.3 channels were unaffected by EGTA (Fig. 5A and fig. S4B), as were both the Ca2+ and K+ currents in the Cav1.2-BKCa coexpressions used as a control (n = 16 cells) (Fig. 5A). In line with the distinct effect of EGTA, immunoprecipitation with the anti-BKα Ab failed to copurify the Cav2.3 protein from Cav2.3-BKCa coexpressing cells (Fig. 5D).

The molecular basis of this subtype specificity is encoded by the pore-forming α subunits. Thus, when cells coexpressing BKα and Cav1.2α in the absence of the respective auxiliary subunits were used for coimmunoprecipitation, both the anti-BKα and the anti-Cav1.2 Abs effectively purified both channel α subunits (Fig. 5E), although to a somewhat lesser extent compared to the previous experiment with cells expressing all BKCa and Cav channel subunits (Fig. 2).

Properties of “Native” BKCa-Cav Channel Complexes

Finally, the functional properties obtained for heterologously reconstituted BKCa-Cav channel complexes were compared to those of their native counterparts. We used chromaffin cells as a model system because of their well-known coupling between Q- and L-type Cav and BKCa channels (31) and their suitability for electrophysiology and efficient intracellular dialysis.

Figure 6A shows a typical sequence of Cav-mediated inward and BKCa-mediated outward currents recorded in response to a step-depolarization in the presence of 5 mM EGTA (n = 32 cells). Coupling of the tetraethylammonium (TEA)–sensitive BKCa currents to the Ca2+ influx is indicated by their deactivation after interruption of the Ca2+ influx by a voltage step to the Ca2+ reversal potential (Fig. 6A) (31). In addition, BKCa currents could be eliminated by application of nifedipine and ω-agatoxin IVA, specific blockers of Cav1.2 and Cav2.1 that are expressed in chromaffin cells (Fig. 6B). The spatiotemporal dynamics of the Cav-BKCa coupling was probed by replacing EGTA in the recording pipette with 5 mM BAPTA. The respective current transients exhibited similar overall properties as with EGTA, although the amplitude of the BKCa currents was decreased by roughly 80% (ratio of mean currents), and the deactivation at the Ca2+ reversal potential was markedly accelerated, as expected for a lower [Ca2+]i at the BKCa channels (Fig. 6A).

Fig. 6.

Recombinant BKCa-Cav channel complexes match the characteristics of their native counterparts. (A) (Left) Current responses to the indicated voltage protocol [adapted from (22)] recorded in chromaffin cells with either 5 mM EGTA (gray) or 5 mM BAPTA (black) in the whole-cell pipette; the trace in red shows block of the BKCa current by 5 mM extracellular TEA (at 5 mM intracellular EGTA). Current scale is 0.5 nA. (Inset) Current traces at expanded time scale. (Right) Mean ± SD of currents through BKCa channels (n =5) from experiments as on the left. (B) Identification of L-type (Cav1.2) and P/Q-type (Cav2.1) channels as the Cav channels coupling to BKCa channels in chromaffin cells. (Left) Mean ± SD of BKCa-mediated currents (n = 5) before and after addition of the L-type Cav channel blocker nifedipine (5 μM) and the P/Q-type channel blocker ω-agatoxin IVA (1 μM). (Right) PCR amplification of transcripts coding for BKα and the Cavα subunits indicated from chromaffin cells; control reactions (without reverse transcription) are referred to as –RT.

Discussion

The central finding of this work is that two distinct classes of ion channels, BK-type Ca2+-activated K+ channels and voltage-gated Ca2+ channels, may be assembled into macromolecular channel-channel complexes in the CNS. Functionally, these complexes reconstitute Ca2+ nanodomains, where Ca2+ influx through the Cav channels provides the [Ca2+]i required for rapid and robust activation of BKCa channels in the physiologically relevant voltage range.

Formation of BKCa-Cav channel complexes. For characterization of the molecular environment of BKCa channels, we started out from proteomic analysis combining APs of appropriately solubilized proteins with nano-LC MS/MS analysis of total eluates. When applied to plasma membrane preparations from total rat brain (25), this approach isolated BKCa channels with high efficiency and provided information on the BKα protein [sequence coverage of ∼75%, splice variations] and on proteins associated with BKCa channels. Respective analyses by nano-LC MS/MS revealed two striking results with respect to the mechanism of native BKCa channel activation. First, the proteins most efficiently copurified with BKα were members of the Cav-channel family (Table 1), with particular abundance of Cav2.1. Second, proteins with a similar peptide yield and proposed scaffolding function were not identified, nor did mass spectrometry retrieve molecules suggested to link BKCa and Cav channels (32). For the Cav channel α subunits identified, quantitative comparisons (between eluates of the anti-BKα Abs and control IgGs) indicated specific copurification for Cav1.2, Cav2.1, and Cav2.2, whereas Cav2.3 was dubbed nonspecific by our specificity scores (Table 1 and fig. S1).

Analyses using biochemistry and electrophysiology on heterologously coexpressed Cav and BKCa channels confirmed the subtype-specific assembly suggested by the proteomic approach. Thus, Cav1.2/Cav2.1 and BKCa channels were effectively copurified from culture cells and Xenopus oocytes without a requirement for additional exogenous partners (Figs. 2 and 5, and fig. S2). The functional properties of the Cav-BKCa coupling fully matched the criteria of Ca2+ nanodomains, with an estimated distance between channels of ∼10 nm (18) (Fig. 4 and fig. S4), a value very similar to the 9.5 nm recently determined for the diameter of the voltage-gated K+ channel Kv1.2 in its crystallized form (33). Mechanistically, channel-channel assembly appears to be determined by the α subunits of BKCa and Cav channels, although a ubiquitously expressed partner protein that escaped our MS/MS analyses cannot be completely ruled out.

Relevance of BKCa-Cav channel complexes. Formation of stable macromolecular complexes with Cav channels affects the physiology of BKCa channels in several ways. First, complex formation provides a simple molecular solution to the issue of how BKCa channels may be supplied with micromolar [Ca2+]i without affecting other Ca2+-dependent metabolic processes. Second, complex formation puts the activity of BKCa channels under tight control of their Cav partners. In the context of an excitable cell, this tight coupling ensures that activation of BKCa channels occurs fast enough to shape the action potential by contributing to its repolarization (4, 7, 8) and to generate the fast afterhyperpolarization following single Na+ or Ca2+ spikes in various types of CNS neurons (1, 3, 6, 7, 34).

BKCa signaling via coassembled Cav channels will be shaped by the distinct distribution of Cav channels to particular types of cells or sub-cellular compartments (Table 1) (35, 36). In fact, all Cav channel subtypes identified were found as partners of somatic BKCa channels in distinct types of CNS neurons (3, 6, 13, 20, 23, 24, 37). In their preferred subcellular localization (38, 39), the presynaptic compartment, however, BKCa channels appear to be fueled by P/Q- and N-type Cav channels (6, 40), in line with our efficient copurification of the Cav subunits Cav2.1 and Cav2.2. Functionally, presynaptic BKCa channels were shown to control transmitter release by narrowing the action potential and reducing Ca2+ influx into the presynaptic elements (6, 12, 40) and to operate as an “emergency brake” that prevents cell damage in the case of globally increased [Ca2+]i (41). Both functions may be related to the molecular arrangement of BKCa channels: Control of transmitter release would well fit with the properties of BKCa-Cav complexes, whereas emergency braking may be attributed to uncomplexed BKCa channels.

The BKCa-Cav channel complexes represent a molecular unit providing effective and precisely timed hyperpolarization of the membrane potential in response to local Ca2+ influx.

Supporting Online Material

www.sciencemag.org/cgi/content/full/314/5799/615/DC1

Materials and Methods

Figs. S1 to S4

References

References and Notes

View Abstract

Navigate This Article