β₄-Subunit increases Slo responsiveness to physiological Ca²⁺ concentrations and together with β₁ reduces surface expression of Slo in hair cells

Abstract

Changing kinetics of large-conductance potassium (BK) channels in hair cells of nonmammalian vertebrates, including the chick, plays a critical role in electrical tuning, a mechanism used by these cells to discriminate between different frequencies of sound. BK currents are less abundant in low-frequency hair cells and show large openings in response to a rise in intracellular Ca²⁺ at a hair cell's operating voltage range (spanning −40 to −60 mV). Although the molecular underpinnings of its function in hair cells are poorly understood, it is established that BK channels consist of a pore-forming α-subunit (Slo) and a number of accessory subunits. Currents from the α (Slo)-subunit alone do not show dramatic increases in response to changes in Ca²⁺ concentrations at −50 mV. We have cloned the chick β₄- and β₁-subunits and show that these subunits are preferentially expressed in low-frequency hair cells, where they decrease Slo surface expression. The β₄-subunit in particular is responsible for the BK channel's increased responsiveness to Ca²⁺ at a hair cell's operating voltage. In contrast, however, the increases in relaxation times induced by both β-subunits suggest additional mechanisms responsible for BK channel function in hair cells.

large-conductance potassium (BK) channels are ubiquitously distributed, and this is reflected in their role in numerous disease states including deafness, hypertension, epilepsy, and movement disorders (1, 10, 11, 13, 38, 47). These channels are unusual in that they have a large potassium conductance and are regulated by both membrane voltage and intracellular Ca²⁺ (29, 30). In hair cells of the auditory system of nonmammalian vertebrates, these channels play a critical role in electrical tuning, a mechanism that enables hair cells to discriminate between different frequencies of sound (14, 21).

Hair cells from the turtle auditory epithelium have an intrinsic oscillation in membrane potential in response to a small injected depolarizing current that sharply correlates with the frequency of sound to which they best respond (characteristic frequency) (9, 16). This electrical tuning, although first described in the turtle auditory epithelium, has been confirmed in other species, including the chick auditory and frog vestibular epithelium (9, 16, 22, 28, 41). Hair cells in the auditory epithelium are arranged tonotopically and show a gradual change in oscillation frequency that corresponds to the cell's tonotopic location (9, 16). The oscillation in membrane potential is brought about by an inward Ca²⁺ current and an outward Ca²⁺-activated K⁺ current (BK) (3, 22, 28). It should be noted that the principal method of activating BK channels in these cells is by increases in intracellular Ca²⁺. Native BK channels show minimal sensitivity to voltage changes operational in hair cells (estimated to vary between −50 mV and −30 mV in the turtle) at an estimated resting Ca²⁺ concentration of 4 μM (4). Owing to the high concentration of Ca²⁺ buffer present in these cells that limits diffusion of Ca²⁺, both Ca²⁺ channels and BK channels were predicted to lie in close physical apposition, and this has been borne out empirically (23, 45, 48, 50–52). A detailed analysis of these channels' kinetics has shown that the primary determinant of changing oscillation frequency is the changing kinetic parameters of the BK channel (2–4). Specifically, data from the turtle have shown that a hair cell's characteristic frequency relates closely to the relaxation times of BK channels from that hair cell. These data have since been broadly replicated in the chick auditory epithelium (12).

The molecular underpinnings of the changing kinetics of BK channels along the tonotopic spectrum have yet to be established with certainty. BK channels contain a pore-forming α-subunit encoded by a single gene (Slo; KCNMA1) that shows extensive alternative splicing (5, 7). The α-subunit alone can form a channel (5, 7). Early work showed evidence of multiple splice variants at individual splice sites that changed systematically along the tonotopic axis, raising the possibility that these splice variants could cause changes in the kinetic properties of the channel (25, 36, 46). However, subsequent work in the chick has shown that these splice variants possess a limited range of relaxation time constants, the parameter that correlated best with characteristic frequency along the tonotopic axis (12, 43, 44). Moreover, these α-subunits also have other features in their kinetics that are at variance with BK channels native to hair cells. Discordance in Ca²⁺ sensitivity is one such example (43, 44). In the chick, different splice variants expressed in HEK cells have K_d (concentration required for half-activation) values at 0 mV that vary from 5 to 40 μM, in contrast to native cells that have a range of values from 1 to 20 μM (43, 44). Similarly, in the turtle the K_d at −50 mV for channels expressed in oocytes varied from 6 μM to >200 μM, in contrast to native hair cells that had a mean value of 12 μM (4, 24, 25, 27). More importantly, turtle native channels showed maximal opening in response to an increase in Ca²⁺ concentration approaching 35–50 μM at −50 mV (4). Similarly, chick native channels held at −50 mV were shown to increase open probability (P_o) from 0.1 to 0.8 in response to an increase in Ca²⁺ concentration from 1 μM to 20 μM (12). These data contrast with different α-subunits expressed in heterologous systems that when held at −50 mV show smaller increases in P_o in response to increases in physiological concentrations of Ca²⁺ (5–50 μM) (24, 25, 27, 43, 44).

In this study, we identify and characterize the β₄-subunit of Slo in the chick and compare its effects on Slo kinetics with the those of the β₁-subunit. While the β₄-subunit has effects on deactivation times similar to those of the β₁-subunit, we show that β₄ is preferentially expressed at the apical end of the basilar papilla and alters the α-subunit's Ca²⁺ responsiveness at the hair cell's operating voltage closer to that observed in native hair cells. We also establish, using small interfering RNA (siRNA) knockdown, that the β₁- and β₄-subunits decrease surface expression of Slo in hair cells. Overall, our data suggest β₄-subunit interactions with Slo as one of many mechanisms that help bring about the complex behaviors of this channel in hair cells.

MATERIALS AND METHODS

DNA and RNA synthesis (constructs information).

For RNA injection, chicken Slo (cSlo) (36), β₁, and β₄ cDNAs were inserted into pGH 19 vector. Plasmids were linearized with NotI and cRNA transcribed with the mMESSAGE mMACHINE kit (Ambion, Austin, TX).

For fluorescence-activated cell sorting (FACS) analysis, cSlo cDNA tagged with FLAG and yellow fluorescent protein (YFP) was subcloned into pcDNA 6 vector; β₁ and β₄ cDNAs were subcloned in-frame into the pECFP N1 vector.

Oocyte preparation and RNA injection.

Oocyte preparation was as previously described (32). All experimental procedures involving animals were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC), Yale University School of Medicine. In brief, oocyte sacks were harvested from anesthetized Xenopus laevis and then incubated in 2 mg/ml collagenase type Ia (Sigma, St. Louis, MO) in nominally Ca²⁺-free ND96 (in mM: 96 NaCl, 2 KCl, 1 MgCl₂, 5 HEPES, pH 7.4) for 20 min with continuous agitation. After several washes in nominally Ca²⁺-free ND96, oocytes were incubated again in 2 mg/ml collagenase type Ia in nominally Ca²⁺-free ND96 for 15 min. The oocytes were washed again in nominally Ca²⁺-free ND96 with agitation until they were well separated from each other. They were then repeatedly washed in ND96 with 1.8 mM CaCl₂ to remove collagenase and complete defolliculation. Stage V and VI oocytes were then sorted and stored at 18°C in OR-3 medium [50% Leibovitz L-15 medium (GIBCO BRL, Grand Island, NY), 5 mM HEPES, 10 ml of 10,000 U/ml of penicillin, and 10,000 μg/ml of streptomycin (Sigma), pH 7.4, 200 mosmol/kgH₂O].

RNA at a volume of ∼46 nl was injected into each oocyte on the following day. Concentrations of RNA varied depending upon the design of the experiment. For macroscopic current recording, cSlo RNA at a final concentration of 0.3 mg/ml was injected with or without β₁/β₄ RNA. For single-channel recording, to ensure one channel per patch, we injected cSlo RNA at a final concentration of 0.05 mg/ml with or without β-subunits. We generally injected cSlo and β-subunits at 1-to-3 ratios by weight. Recordings were performed 2–6 days after injection.

Macroscopic current recording.

Macroscopic currents were recorded in inside-out patches from injected oocytes by a whole cell voltage-clamp technique (17) at room temperature with an Axon 200B amplifier (Axon Instruments, Sunnyvale, CA). Command delivery and data collections were carried out with a Windows-based whole cell voltage-clamp program, jClamp (Scisoft, Ridgefield, CT), using a Digidata 1322A interface (Axon Instruments). Currents were digitized at 100 kHz and filtered at 5–10 kHz. Data analysis was accomplished with jClamp or Clampfit (Molecular Devices, Union City, CA). Capacitance and leak currents were subtracted with a P/-5 leak subtraction protocol. Conductance-voltage (G-V) curves were obtained by measuring the amplitude of tail currents 100–200 μs after repolarization to −100 mV from the various test voltages. Each G-V curve was fitted with a Boltzmann function:

$$G(V)=\frac{G_{\max }}{1+{\exp }^{\frac{-zF(v-v_h)}{\mathrm{RT}}}}$$ (1)

where G_max is the fitted value for maximal conductance, V_h is the voltage of half-maximal activation of conductance, z reflects the net charge moved across the membrane during the transition from the closed to the open state, and F, R, and T are the Faraday constant, the gas constant, and temperature, respectively.

To characterize relaxation kinetics, tail currents were collected at various relaxation voltages after a 100-ms (for cSlo only or cSlo with β₁) or 200-ms (for cSlo with β₄) depolarization to 180 mV, to maximize the channel opening at specific Ca²⁺ concentrations. The time constants were obtained by fitting the tail currents to a single exponential function.

Pipettes were pulled from thin-walled borosilicate glass (TW 150, World Precision Instruments, Sarasota, FL) with impedance of ∼1 MΩ. The standard pipette/extracellular solution contained (in mM) 140 K-methanesulfonate, 20 KOH, 10 HEPES, and 2 MgCl₂, pH 7.2. The composition of the bath/intracellular solution was (in mM) 140 K-methanesulfonate, 20 KOH, 5 mM HEDTA, and 10 HEPES, pH 7.2, and Ca-methanesulfonate₂ was added to reach the appropriate free Ca²⁺ concentration. No Ca²⁺ chelator was used in the solution containing 100 μM free Ca²⁺. The amount of total Ca-methanesulfonate₂ needed to obtain the desired free Ca²⁺ concentration was calculated with Max Chelator (6), which was downloaded from http://www.stanford.edu/∼cpatton/webmaxc.htm. Final free Ca²⁺ concentration was measured with a Ca²⁺ electrode (Thermo Electron, Beverly, MA). The bath/intracellular solutions were delivered with the ALA QMM micromanifold perfusion system (ALA Scientific Instrument, Westbury, NY).

Single-channel recording.

Single-channel recordings were carried out with excised inside-out patches from injected oocytes at room temperature (17). Command delivery, data collection, and analysis were as described for macrocurrent recordings, except that we used AxoScope (Axon Instruments) for data collection. Patch pipettes with impedance of ∼15–25 MΩ were pulled from thick-walled borosilicate glass (TW 150, World Precision Instruments) and coated with Sylgard. The components of pipette/extracellular and bath/intracellular solutions were identical to those used in macrocurrent recordings. The recordings were low-pass filtered with an eight-pole Bessel filter at between ∼2 and 5 kHz and digitized at 20 kHz. Open- and closed-state transitions were defined with half-amplitude threshold analysis (8). Channel amplitudes were determined by fitting an amplitude histogram of data points. Duration histograms were constructed and fitted by the maximum likelihood method (33, 49).

To identify bursts of single-channel opening, critical gaps were defined in the interevent (closed) interval distribution (31). In brief, after detection of single-channel activities, the distributions of interevent interval durations were first fitted with appropriate exponential functions, generally, the sum of two exponential components. A critical gap length was then calculated from the distribution to give equal numbers of misclassified closed events. This analysis was performed on the recordings with P_o generally <0.5, since it was hard to define gaps between bursts with higher P_o.

Quantitative PCR.

Chick basilar papillae were microdissected and divided into four quadrants [low-frequency quadrant, middle-frequency (MF1 and MF2) quadrants, and high-frequency quadrant]. RNA was isolated from each segment with the RNAqueous kit (Ambion) according to the manufacturer's instructions. cDNA was synthesized with oligo(dT) primers and random hexamers as previously described (36). Quantitative (q)PCR amplification with cDNA from each quadrant of the papillae were performed with the IQ SYBR Green supermix (Bio-Rad, Hercules, CA). The reaction mixtures were set up in 96-well thin-wall plates (Bio-Rad) and amplified with a C1000 thermal cycler containing a CFX96 optical reaction module (Bio-Rad). Amplification data were analyzed with CFX Manager software (Bio-Rad) and normalized to 18S RNA.

The primer sequences were as follows: 18S forward (F): AGAAACGGCTCCACATCCAAG, 18S reverse (R): TCAAAGTCCCTCCAATGGTCC; cSlo F: TGGACTGTCTGCTGCCACTG, cSlo R: CCCATCTGCCCTGGTATCTC; β₁ F: CCCCGGCAACTCCGAGAACG, β₁ R: ACATGAGCGTGGGCCAAAGGA; β₄ F: CCAACCCCAAGTGTTCCTACATC, β₄ R: GCCTCAAGTGCTGGTTAAAATAGC.

FACS analysis.

Chick Slo fused to YFP (at its COOH terminus) and FLAG tag (at its NH₂ terminus) with or without chick β₁-cyan fluorescent protein (CFP) and, separately, chick β₄-CFP were transfected into HEK cells. Forty-eight hours after transfection, live cells were harvested and incubated with anti-flag (M2) antibody (Sigma) in phosphate-buffered saline, 1% bovine serum albumin at 4°C. The primary antibody was detected by a secondary anti-mouse antibody conjugated to Alexa 647. The cells were fixed briefly in 1% paraformaldehyde before analysis. FACS analysis was carried out with FlowJo software (Tree Star, Ashland, OR) as previously described (35).

RNA interference experiments.

RNA interference (RNAi) experiments in chick cochlea were performed with a protocol previously described (15). Two sets of siRNA duplexes each were used for β₁ and, separately, β₄ knockdown. They had the following sequences: β₁ ccaaagaagugaaagcaaauu/uuugcuuucacuucuuugguu (sense/antisense) and gcucagaagaugaagauauuu/auaucuucaucuucugagcuu; β₄ cagaaauacuggaaagagauu/ucucuuuccaguauuucuguu and acacggaggcggaggacaauu/uuguccuccgccuccguguuu.

Cochlea were fixed 3 days after transfection and Slo detected with confocal microscopy as previously described (48). Cells were scanned in the z-plane from the base to the apex of each cell with a Zeiss 510 confocal microscope. Fluorescence intensity in a 50 × 50-μm area was ascertained after background subtraction with Zeiss LSM Image Analyser software. We had identified membrane clustering of Slo with this methodology and demonstrated Slo membrane clusters to have increased fluorescence. The average fluorescence intensity of Slo membrane clusters was 400 arbitrary units/pixel and was used as a threshold to demarcate surface clusters of Slo.

Data analysis.

All results are given as means ± SE. Where appropriate, ANOVA was used to test for significance in differences.

RESULTS

β₄- and β₁-subunits are expressed differentially along tonotopic axis.

We identified and cloned the chick β₄- and β₁-subunits of the Slo channel by screening expressed sequence tag (EST) databases. Overall, β₄- and β₁-subunits have 64% and 45% identity, respectively, with the corresponding human subunits (Fig. 1). Notably, the chick β₁-subunit has 94% identity with the previously identified quail subunit. Interestingly, while there was great sequence divergence in the overall sequence of both proteins, the NH₂ terminus of these proteins showed much greater sequence conservation compared with the human sequence. Thus there was 86% identity in the first 30 amino acids of the chick and human β₁-subunits. Similarly, there was a 92% identity in the first 30 amino acids of the chick and human β₄-subunits. Previous work has shown the NH₂ terminus of the β-subunits as key in influencing Slo channel kinetics (42).

Fig. 1.

Conservation of the β₁ and β₄ amino acid sequences is particularly notable at their NH₂ termini. The chick β₁ (A)- and β₄ (B)-subunits aligned (with the BLAST algorithm) with their corresponding human orthologs are shown. Identical amino acids and conservative substitutions (+) are indicated. β₁ shows 45% identity across the entire sequence, although the first 20 amino acids show much greater sequence identity (17/20 or 85%). In contrast, the β₄-subunits show greater (61%) identity across the entire sequence and a degree of identity similar to the β₁ orthologs at the NH₂ terminus (17/21 or 81%).

We determined the distribution of these two subunits along the tonotopic axis with qPCR. As shown in Fig. 2 the β₄- and β₁-subunits were expressed predominantly in the low-frequency quadrant of the basilar papilla. Similarly, and unexpectedly, Slo transcripts were also expressed in greater abundance at the low-frequency end of the papilla. There was a steep and gradual decrease in the expression of these transcripts. Compared with the lowest-frequency quadrant, β₄ was expressed at levels of 78%, 68%, and 43% in the successive higher-frequency quadrants, respectively. Similarly, the β₁-subunit was expressed at 63%, 65%, and 35% in the successive higher-frequency segments. We were unable to confirm whether these differences in transcript levels corresponded to levels in protein expression. Presently there are no commercial antibodies against the chick β-subunits. We were unable to detect chick β₁ and β₄ proteins with several commercial antibodies against the mammalian ortholog. The Slo transcripts were expressed at 58%, 57%, and 26% in successive higher-frequency quadrants compared with the lowest-frequency quadrant. These data stand in contrast to the demonstrated increase in Slo protein expressed in the membrane as determined by both electrophysiological and antibody labeling experiments (2–4, 48).

Fig. 2.

β₁- and β₄-subunits are expressed in a gradient along the tonotopic axis of the basilar papilla. Transcript levels of β₁- and β₄-subunits expressed in successive quadrants of the tonotopic axis normalized to expression in the low frequency (LF) quadrant are shown. There is a decreasing amount of both transcripts from the LF to high-frequency (HF) ends of the papilla. The expression of both transcripts is also minimally changed across both middle-frequency (MF) quadrants (MF1 and MF2). The differences in expression between LF and both MF quadrants, MF and HF quadrants, and LF and HF quadrants are statistically significant (P < 0.01, ANOVA). Error bars are SD (n = 3).

β₄-Subunit has effects on macroscopic currents that are at variance with those of β₁.

We tested the effects of the β₄-subunit on the kinetics of the Slo channel. Oocytes were injected with cRNA from the β₄-subunit and cSlo. We chose for these experiments a cSlo transcript that when translated lacks extraneous amino acids inserted into its COOH terminus as a result of alternative splicing. The Slo open reading frame has a large number of extraneous amino acids inserted into its sequence as a result of alternative splicing. RNA encoding the very COOH terminus of the protein is also alternatively spliced, and the sequence we chose for these experiments encoded the amino acids KQKYVQEDRL at its COOH terminus. This variant is the most abundantly expressed form in the chick cochlea (34, 36).

Using tail current measurement from macropatches in an inside-out configuration, we noted that cβ₄ had opposing effects on the G-V curves that depended on the concentration of Ca²⁺. In 0, 1, and 10 μM Ca²⁺ G-V curves were shifted in a depolarizing (rightward) direction compared with cSlo alone. This effect was most pronounced in nominally Ca²⁺-free solutions (Fig. 3A). At higher concentrations of Ca²⁺ (>25 μM) G-V function was shifted in a hyperpolarizing (left) direction. This effect on G-V function was in contrast to the effect of the β₁-subunit, which showed a shift in a depolarizing (right) direction in the nominally Ca²⁺-free solutions but a hyperpolarizing (left) shift in 1 μM Ca²⁺ and above.

Fig. 3.

β₁ and β₄ produce differential effects on conductance-voltage (G-V) relationships that are dependent on Ca²⁺ concentration. A: G-V relationships of inside-out macropatches from oocytes injected with Slo alone, Slo with β₁, and, separately, Slo with β₄, in different concentrations of “intracellular” Ca²⁺. Both β-subunits cause a shift in G-V function in a depolarizing direction in nominally Ca²⁺-free solutions. In contrast to β₁, which causes a shift in a hyperpolarizing direction at all concentrations of Ca²⁺ (1 μM and greater), β₄ causes a shift in a depolarizing direction in G-V relationships except when the concentration of intracellular Ca²⁺ exceeds 25 μM (and above). cSlo, n = 11; cSlo + β₁, n = 12; cSlo + β₄, n = 15. B: β₄ increases Ca²⁺ sensitivity of BK channels at voltages that span a hair cell's operating voltage range (−50 mV ± 5 mV). The effects of increasing Ca²⁺ concentration on G/G_max at both −60 mV and −40 mV reveal β₄ to sensitize Slo to physiological concentrations of Ca²⁺ (5–50 μM). There is a dramatic rise in current in the presence of the β₄-subunit. In contrast, the β₁-subunit confers increased channel opening at <10 μM Ca²⁺, at these voltages (cSlo, n = 11; cSlo + β₁, n = 12; cSlo + β₄, n = 15).

Calcium sensitivity is an important parameter of hair cell BK channels. We determined effects on Ca²⁺ sensitivity of the BK channel imparted by the β₄- and β₁-subunits. We plotted the Ca²⁺ concentration needed to bring about half-maximal opening of the channel at a particular voltage in the presence and absence of the two β-subunits (Fig. 3B). The value for the half-saturating Ca²⁺ concentration at 0 mV interpolated from these data confirmed the increased Ca²⁺ sensitivity brought about by the β₁-subunit; K_0.5 (0 mV) was 20.9 μM for cSlo and 2.9 μM for cSlo and β₁. The corresponding value for Slo and β₄ was 14.8 μM Ca²⁺. However, β₄ had a sharper rise in current at higher (more physiological) concentrations of Ca²⁺. This is depicted in Fig. 3B, where G/G_max is plotted against Ca²⁺ concentration at −40 mV and −60 mV. This voltage range spans the presumed operating voltage range in hair cells (−50 mV ± 5 mV). As is evident, β₄ has the ability to act as a binary molecular switch at these voltages since it dramatically raises current with increasing concentrations of Ca²⁺ in the physiological range (estimated to lie between 5 and 50 μM Ca²⁺ in hair cells) (4).

Using the Hill equation, we determined that at +60 mV Slo alone had a K_d of 2.3 (n = 1.32). In contrast, at +60 mV the β₁- and β₄-subunits gave K_d of 0.35 μM (n = 0.9) and 3.4 μM (n = 1.45), respectively. In turtle hair cells, where comparable values have been published, at +50 mV the Hill equation gave a K_d of 2.9 (n = 3.5), a value closest to Slo and β₄.

The effects of the β₄- and β₁-subunits on activation times are depicted in the macrocurrent traces in Fig. 4A. Activation times of Slo were affected by both voltage and Ca²⁺ concentration and are shown in Fig. 4B. We first subjected excised inside-out patches to a step protocol holding at −100 mV and stepping to a range of voltages from −80 to +180 mV. β₄ shortened activation times of cSlo akin to the effects of the β₁-subunit, albeit more pronounced (Fig. 4). This effect was seen at all voltages. Changing the concentration of Ca²⁺ perfusing the “intracellular” surface of the patch at a given voltage command also resulted in faster rates of activation.

Fig. 4.

Activation of Slo, dependent on voltage and Ca²⁺, is slowed by β₄ and β₁. A: current traces of inside-out macropatches from oocytes injected with cSlo, cSlo and β₁, and cSlo and β₄ cRNA subjected to a step voltage protocol. Patches were held at −100 mV and stepped to a range of voltages varying from −80 to +180 mV before stepping back to −100 mV (1 μM intracellular Ca²⁺). β₄ slows activation more than β₁ (cSlo, n = 15; cSlo + β₁, n = 12; cSlo + β₄, n = 20). B: activation time constants (τ_activation) as a function of Ca²⁺ (with voltage stepped from −100 to +60 mV, top) and voltage (with intracellular Ca²⁺ held at 1 μM, bottom). As is evident, higher voltage and Ca²⁺ both result in shortened times of activation.

Similarly, both β-subunits increased relaxation times after a step depolarization. In these experiments we held the voltage at a value that resulted in maximal opening of channels in the patch and then changed its voltage to a range of values and determined relaxation of the channels. As with activation times, relaxation times were dependent on the concentration of Ca²⁺ and the voltage used for the step depolarization (Fig. 5). Lower concentrations of Ca²⁺ and more hyperpolarizing voltages to which the channel was stepped both resulted in shorter relaxation times. As with activation times, the addition of either the β₁- or β₄-subunit slowed relaxation. In contrast to activation times, relaxation times were most affected by the β₁-subunit. Notably, addition of either β₁- or β₄-subunit does not affect the ability of voltage and Ca²⁺ to influence relaxation times, although they influence the set point.

Fig. 5.

Relaxation times, dependent on both Ca²⁺ and voltage, are shortened by β₁- and β₄-subunits. A: current traces from macropatches subjected to a depolarizing voltage (+180 mV) and stepped to a range of voltages (−100 to 0 mV) (intracellular Ca²⁺ concentration of 1 μM). The β₁-subunit and, to a lesser extent, the β₄-subunit shorten relaxation times. B: both decreasing Ca²⁺ (stepped from +100 mV to −100 mV, top) and hyperpolarizing voltage (intracellular Ca²⁺ held at 1 μM, bottom) shorten relaxation times. The relative differences in relaxation time induced by both β-subunits are maintained in different voltages and Ca²⁺ concentrations (cSlo, n = 11; cSlo + β₁, n = 11; cSlo + β₄, n = 11). τ_relaxation, relaxation time constant.

In previous experiments in turtle and chick hair cells relaxation times were shown to correlate with characteristic frequency of individual hair cells. In these experiments relaxation times were determined with a fixed-voltage protocol and Ca²⁺ concentration. Two voltage commands and Ca²⁺ concentrations were used in the two species [a step from +50 mV to −50 mV with 4 μM Ca²⁺ in the turtle (4, 24, 25) and a step from +100 to −100 mV with 5 μM Ca²⁺ in the chick by the Fuchs group (12, 44)]. We subjected patches to a similar protocol (+100 to −100 mV in 5 μM Ca²⁺) to enable comparison with BK channels native to the hair cell. As shown in Table 1 the β-subunits both increased relaxation times compared with cSlo alone at these voltage and Ca²⁺ parameters. However, relaxation times were less prolonged by β₄ than by β₁.

Table 1.

Relaxation times obtained from macropatches hyperpolarized from +100 mV to −100 mV in 5 μM Ca²⁺

	cSlo (n = 10)	β1 + cSlo (n = 9)	β4 + cSlo (n = 11)
τrelaxation, ms	1.06	7.71	5.32
SE	0.19	0.60	0.58

Values are means and SE for n macropatches. cSlo, chicken Slo; τ_relaxation, relaxation time constant.

β₄-Subunit has effects on single-channel kinetics that are intermediate between those of β₁ and Slo alone.

We recorded single-channel tracings from inside-out patches of oocytes expressing cSlo with and without the β₄- and (separately) β₁-subunits. We limited our analysis to patches containing one single channel. We determined P_o of these channels but noted significant variability in P_o-V relationships between individual patches. A corollary of this finding is that only a minority of patches had P_o-V relationships that coincided with the G-V curves obtained from macrocurrent recordings.

We determined that the β₄- and β₁-subunits retain the channel in a bursting state, increasing channel burst duration (Fig. 6). This effect was previously characterized in detail in the mammalian β₁-subunit (39, 40). The effect of the β₄-subunit was less than that of the β₁-subunit. Notably, both of these subunits increased burst duration in nominally Ca²⁺-free solutions. In addition to their effects on burst duration, both β-subunits also increased interburst duration. These effects are depicted in Fig. 7, which shows β₁ and β₄ to increase burst duration and interburst duration at a given P_o. The increase in burst duration and interburst duration was independent of the mechanism of channel activation, whether by an increase in concentration of Ca²⁺ or in voltage. Figure 8 shows representative traces depicting the effects of increasing Ca²⁺ and voltage on P_o and burst duration of Slo in the absence and presence of β₁ and, separately, β₄. These data are in agreement with an exhaustive study by Nimigean and Magleby (39, 40), who noted similar results in the mammalian (mouse) β₁-subunit.

Fig. 6.

Single-channel recordings confirm that both β-subunits prolong bursting, even in the absence of Ca²⁺. Single-channel tracings from excised inside-out patches expressing Slo, Slo and β₄, and Slo and β₁ in 0 μM (A) and 1 μM (B) Ca²⁺ are shown. As is evident, both β-subunits increase bursting of the channel even in nominally Ca²⁺-free solutions, with β₁ having a more marked effect. Open (o) and closed (c) states are indicated.

Fig. 7.

Both β-subunits increase burst duration and interburst duration, with the β₁-subunit having a greater effect on both and β₄ having an intermediate effect. Plot of open probability (P_o) vs. burst duration (A) and interburst duration (B) are shown. Filled symbols designate data in nominally Ca²⁺-free solutions with varying voltage. Open symbols designate data from traces at +60 mV with varying concentrations of Ca²⁺. Increasing P_o, whether from increasing Ca²⁺ or voltage, changes the burst and interburst duration in a parallel fashion in all 3 experimental groups (cSlo alone, cSlo and β₁, and cSlo and β₄).

Fig. 8.

Effects of increasing Ca²⁺ and voltage on P_o and burst duration. Representative traces depicting effects on burst duration and P_o of increasing voltage (in nominally 0 μM Ca²⁺; A) and increasing Ca²⁺ (with constant voltage; B), which both increase burst duration and P_o are shown. Traces from cSlo, cSlo and β₁, and cSlo and β₄ are shown. Burst duration induced by β₁ is greater than that induced by β₄.

We additionally determined a correlation between burst duration and relaxation times after a voltage step (Fig. 9). For patches expressing Slo alone, Slo with β₁, or Slo with β₄, there was a linear correlation between the burst duration at a given concentration of Ca²⁺ and the relaxation times after a step depolarization (slope = 8.1, regression coefficient = 0.8). A better correlation was obtained between mean open times and relaxation times (slope = 1.2, regression coefficient = 0.9). In assessing these correlations we limited our analysis to single channels that had P_o-V relationships that matched G-V relationships from macropatch recordings.

Fig. 9.

Relaxation times correlated with both mean open time and burst duration. Relaxation times from macropatches subjected to a step depolarization from +60 mV to −100 mV (1 and 10 μM Ca²⁺) or from +120 mV to −100 mV (nominally 0 μM Ca²⁺) were plotted against mean burst duration (left axis) or mean open times (right axis). To make valid comparisons, the values for both mean open times and mean burst times were obtained at the corresponding voltage and Ca²⁺ concentrations (+60 mV with 1 or 10 μM Ca²⁺ and +120 mV with nominally 0 μM Ca²⁺). Filled and open symbols indicate the correlation between burst duration and relaxation times and between mean open times and relaxation times, respectively. Fits in both correlations are shown. Relaxation time constants correlated better with mean open times (with a regression coefficient of 0.9 and slope 1.2) compared with burst duration and relaxation times (with a regression coefficient of 0.80 and slope 8.1).

Knockdown of both subunits by siRNA increases surface expression of Slo.

Analysis of macrocurrent data showed that addition of both β-subunits significantly decreased the total current within a macropatch (note that macropatches were of the same dimension, ∼0.5 μm²). Previous reports have shown contradictory effects on the surface expression of Slo by the β₁-subunit. To test the effects on surface expression we used a Slo construct that included a FLAG tag at its extracellular NH₂ terminus. Using FACS analysis, we observed a significant decrease in the surface expression of Slo in HEK cells that were cotransfected with either β₁- or β₄-subunits, consistent with our macropatch data. This decrease in surface expression was achieved without affecting total amounts of intracellular Slo (Fig. 10).

Fig. 10.

Both β-subunits decrease surface expression of Slo in HEK cells. A: histogram of fluorescence intensity detected by fluorescence-activated cell sorting (FACS) of HEK cells expressing FLAG-cSlo-yellow fluorescent protein (YFP) cotransfected with cyan fluorescent protein (CFP), β₁-CFP, and, separately, β₄-CFP shows that surface expression of Slo, detected 48 h later, is decreased. Surface expression of Slo was detected with anti-FLAG antibody labeling of live cells. Anti-mouse Alexa 647 was used as a secondary antibody. Slo-expressing cells were selected with identical YFP gating. Note that fluorescence intensity is denoted on a log scale. B: in contrast, total expression of Slo detected by YFP fluorescence (in the identical experiment shown in A) confirms that total Slo expression was minimally affected by coexpression of the β-subunits.

We tested whether this decrease in Slo surface expression induced by the β-subunits is physiologically important. Chick cochlea were kept in culture in the presence of siRNA to β₁- and, separately, β₄-subunits (Fig. 11). We had established separately that these siRNAs were able to knock down expression of these respective β-subunits in HEK cells transfected with β₁- and β₄-subunits fused to CFP (data not shown). Slo was identified with fluorescent antibody labeling. Slo expressed on the surface of hair cells is clustered and consequently has a markedly higher fluorescence intensity. As is evident, there was an increase in the expression of Slo on the surface of hair cells treated with siRNA to both subunits. The effect of β₄ siRNA was similar to that of β₁, with a 330% increase in surface expression of Slo in hair cells with β₄ siRNA and a 290% increase in surface expression with β₁ siRNA.

Fig. 11.

Slo surface expression in hair cells is increased by small interference RNA (siRNA)-mediated knockdown of the β₁- and β₄-subunits. Confocal images in the x-y plane of hair cells derived from chick cochlea kept in culture and transfected with a nontargeting siRNA (A, D, G) or siRNA to β₁ (B, E, H) and β₄ (C, F, I) are shown. Slo immunofluorescence (A–C) at the basal pole of hair cells detected under low magnification is contrasted with Alexa 647 phalloidin staining of stereocilliary actin bundles (D–F) 2 μm above in the z-axis. Also shown are the corresponding high-magnification confocal images in the x-y plane of Slo immunofluorescence (G–I). As is evident, there was an increase in Slo immunofluorescence in hair cells after knockdown of both β₁- and β₄-subunits. Moreover, there was also an increase in the clustering of Slo in hair cells treated with β₁ and β₄ siRNA. Scale bars are indicated. J: volumetric analysis of image intensity of Slo in a 2,500-μm² area along the z-plane from the base to apex of hair cells along the neural edge shows an increase in Slo labeling after treatment with β₁ and β₄ siRNA. The mode in fluorescence intensity increased from 160 in control cochlea to 276 in cochlea treated with β₁ siRNA and 256 in cochlea treated with β₄ siRNA. Increased clustering of the channel in cochlea transfected with siRNA to the β₁- and β₄-subunits was evidenced. The threshold in fluorescence intensity for demarcating clusters is shown. There is a 290% and 330% increase in the number of clusters in cochlea treated with siRNA to the β₁- and β₄-subunits, respectively, compared with control cochlea.

DISCUSSION

In this article we describe the distribution of the β₄-subunit of Slo in the basilar papilla and its physiological effects, comparing it to the β₁-subunit. Our data confirm that the chick β-subunits have effects on Slo macroscopic and single-channel kinetics that are broadly similar to their mammalian counterparts.

Since there is sequence preservation between β-subunits from different phyla that is most notable at their NH₂ terminus, these data confirm a role for the NH₂ terminus of the protein in determining its effects on Slo kinetics. Prior work has confirmed the NH₂ terminus of β-subunits as a region most critical for determining their effects on the kinetic properties of Slo (42).

We noted a significant variability in single-channel recordings, a finding that has been established previously with mammalian Slo channels (18–20). Macropatch recordings, in contrast, show considerably less variability. Our macropatches were estimated to have a 100-fold greater density of channels. It is possible that the lower variability in macropatches merely represents statistical averaging.

In a side-by-side comparison the β₄-subunit has effects on Slo kinetics that parallel those of the β₁-subunit, although not as pronounced. Thus the β₁-subunit increased apparent Ca²⁺ sensitivity by shifting G-V curves in the presence of Ca²⁺ to more hyperpolarizing voltages. This effect was seen with the β₄-subunit only at higher concentrations of Ca²⁺. In contrast, in nominally Ca²⁺-free solutions β₁ shifts Slo G-V function to more depolarizing voltages than β₄ alone. Similarly, both β-subunits slow relaxation times, with the β₁-subunit having more pronounced effects. The singular exception to the rule that β₁ had more pronounced effects was the effect on activation, where the β₄-subunit prolonged activation times more than β₁.

Single-channel recordings showed that both subunits prolong bursting, with the β₁-subunit having a more pronounced effect. The effects on bursting were seen in nominally Ca²⁺-free solutions. These data are consistent with prior work by Nimigean and Magleby (39, 40). We observed that the duration of bursting correlated with relaxation times. Together with data showing slowed activation and relaxation times with the β-subunits, these data suggest that the β-subunits slow the transition from open to closed states and vice versa.

While we have evidence for transcription of the β₁ and β₄ genes in the basilar papilla, the exact physiological role of these proteins in hair cells has to be established. The data presented here are indirectly supportive of the role of the β-subunits in Slo kinetics in hair cells. Moreover, we have direct evidence that they play a part in Slo localization on the surface of hair cells.

BK channels in chick hair cells have a critical role in electrical tuning, a mechanism that allows frequency selectivity. Both empirical and modeling data have shown hair cell resting membrane potential to fluctuate 2–5 mV from approximately −50 mV, with Ca²⁺ concentration estimated to fluctuate from 5 to 50 μM (2, 4, 9). These data argue for β₄ to play an important and integral role in electrical tuning. The β₄-subunit changes Slo kinetics, with steeper opening of the channel in the 5–35 μM range in Ca²⁺ concentrations at voltages (−40 mV to −60 mV) that span a hair cell's operating voltage (−50 mV). In contrast, the Slo α-subunit alone has a more shallow opening in response to changes in Ca²⁺ concentrations at these voltages. Similarly, the β₁-subunit results in Slo channels that are close to maximally open even at lower physiological concentrations of Ca²⁺ (5 μM) at −40 mV and −60 mV. Taken together, these data give theoretical credence that β₄ plays a critical role in electrical tuning. The prolonged bursting behavior induced by both β-subunits is seen in low-frequency hair cells in the turtle, in keeping with the increased expression of both β-subunits in low-frequency hair cells. The relaxation time constants induced by either subunit, however, are in excess of that observed in native hair cells, suggesting that additional mechanisms are responsible for BK channel kinetics. It is plausible that other modifiers including tonotopically determined phosphorylation or association with other subunits selectively affect specific aspects of Slo kinetics (bursting/relaxation times), keeping other kinetic properties (Ca²⁺ sensitivity) constant.

Our data showing increasing amounts of Slo clusters in the surface membrane of hair cells treated with siRNA to the β₄- and β₁-subunits suggest that the β₄- and β₁-subunits are expressed in hair cells and that they play a meaningful physiological role. Our data in explanted chick basilar papillae replicate data in HEK cells, where these subunits reduce surface expression of Slo. While they suggest a common mechanism between HEK cells and hair cells, it remains to be determined how these subunits normally reduce surface expression of Slo. These data also help reconcile the discordance between the increased amounts of Slo transcript in low-frequency hair cells compared with high-frequency hair cells, which contrasts with the lower channel density in the membrane of these cells. Since β₁ and β₄ are expressed at higher levels in low-frequency hair cells, our data suggest that these proteins preferentially prevent Slo expression on the surface of low-frequency hair cells.

GRANTS

This study is supported by National Institute on Deafness and Other Communication Disorders Grant R01-DC-007894.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).