Effects of intracellular stores and extracellular Ca²⁺ on Ca²⁺-activated K⁺ currents in mature mouse inner hair cells

Abstract

Ca²⁺-activated K⁺ currents were studied in inner hair cells (IHCs) of mature mice. I_K,f, the large-conductance Ca²⁺-activated K⁺ current (BK) characteristic of mature IHCs, had a fast activation time constant (0.4 ms at −25 mV at room temperature) and did not inactivate during 170 ms. Its amplitude, measured at −25 mV, and activation time constant were similar between IHCs in the apical and basal regions of the cochlea. I_K,f was selectively blocked by 30 nm IbTx but was unaffected by superfusion of Ca²⁺-free solution, nifedipine or Bay K 8644, excluding the direct involvement of voltage-gated Ca²⁺ channels in I_K,f activation. Increasing the intracellular concentration of the Ca²⁺ chelator BAPTA from 0.1 mm to 30 mm reduced the amplitude of I_K,f at −25 mV and shifted its activation by 37 mV towards more depolarized potentials. A reduction in the size of I_K,f and a depolarizing shift of its activation were also seen when either thapsigargin and caffeine or ryanodine were added intracellularly, suggesting that I_K,f is modulated by voltage-dependent release from intracellular Ca²⁺ stores. Mature IHCs had a small additional Ca²⁺-activated K⁺ current (I_K(Ca)), activated by Ca²⁺ flowing through L-type Ca²⁺ channels. This current was still present during superfusion of either IbTx (60 nm) or apamin (300 nm) but was abolished in Cs⁺-based intracellular solution or during superfusion of 5 mm TEA, suggesting the presence of an additional BK-channel type. Current clamp experiments at body temperature show that I_K,f, but not I_K(Ca), is essential for fast voltage responses of mature IHCs.

Large-conductance Ca²⁺-activated K⁺ channels (BK) are expressed in hair cells of both mammalian and non-mammalian vertebrates (Lewis & Hudspeth, 1983; Art & Fettiplace, 1987; Hudspeth & Lewis, 1988; Fuchs et al. 1988; Kros & Crawford, 1990; Dulon et al. 1995; Kros et al. 1998; Armstrong & Roberts, 2001; Skinner et al. 2003). Non-mammalian auditory hair cells are intrinsically tuned to a particular frequency of sound, which is primarily dictated by the interplay of Ca²⁺ and BK channels (Art & Fettiplace, 1987; Fuchs et al. 1988). Therefore colocalization between Ca²⁺ and BK channels (Hudspeth & Lewis, 1988; Roberts et al. 1990), their relative numbers and kinetic properties (Wu et al. 1995; Jones et al. 1999; Ramanathan et al. 2000) as well as local Ca²⁺ buffering kinetics (Ricci et al. 2000) are important determinants for setting the hair cell's intrinsic resonant frequency.

No evidence for electrical tuning is found in IHCs, the primary sensory receptors of the mammalian cochlea (Kros & Crawford, 1990). The appearance of the fast-activating I_K,f at the onset of hearing (around 10–12 days after birth in mice) abolishes the action potentials characteristic of immature IHCs (Kros et al. 1998) and enables the cells to respond to sustained depolarization with graded receptor potentials that are essential for accurate sound encoding. A number of observations suggest that I_K,f is likely to flow through BK channels: it is highly sensitive to iberiotoxin (Kros et al. 1998; Skinner et al. 2003), a specific blocker of BK channels (Kaczorowski et al. 1996), to charybdotoxin and to submillimolar concentrations of TEA (Kros et al. 1998) and is resistant to high concentrations of extracellular and intracellular 4-AP (Kros & Crawford, 1990). Moreover, the majority of inside-out patches from the basolateral membrane of guinea-pig IHCs contained BK channels with a single channel conductance of about 230 pS with kinetics similar to I_K,f (Appenrodt & Kros, 1997) and currents recorded from inside-out macropatches taken from mouse IHCs were Ca²⁺ dependent and had fast activation kinetics (Oliver et al. 2003). However, removal of extracellular Ca²⁺ did not affect the amplitude or activation kinetics of I_K,f (Kros & Crawford, 1990). It was thus suggested that I_K,f must depend on Ca²⁺ released within the cell, or operate exclusively by means of its intrinsic voltage sensitivity (Kros & Crawford, 1990). The first direct evidence in mature IHCs for a Ca²⁺-activated K⁺ current that was affected by extracellular Ca²⁺ came from a study on the guinea-pig (Dulon et al. 1995). The authors showed that in mature IHCs when Ca²⁺ entry through L-type Ca²⁺ channels was blocked by the superfusion of nifedipine, a reduction of the total outward current occurred. Since the isolated Ca²⁺-activated K⁺ current was partially sensitive to charybdotoxin and insensitive to apamin the authors suggested that it might be carried by BK channels. The relatively small amplitude and slow activation of this current compared with those of I_K,f (Kros & Crawford, 1990) suggests that both currents might be carried by two distinct BK channel types.

To investigate whether more than one population of BK channels is expressed in mouse IHCs we set out experiments in which the sensitivity of the outward K⁺ current of mature cells to intracellular and extracellular Ca²⁺ was tested. To test the basis of the voltage-dependent activation of I_K,f we applied various agents that affect intracellular Ca²⁺ stores, which have been reported in immature IHCs (Kennedy & Meech, 2002). Moreover, we studied whether kinetic variations of the BK channel, like those found in the hair cells of the chick and turtle cochlea, were also present along the tonotopic axis of the mouse cochlea. We show that mouse cochlear IHCs express two distinct Ca²⁺-activated K⁺ currents that are both modulated by increases in cytosolic Ca²⁺ but only one of which is sensitive to extracellular Ca²⁺.

Methods

Tissue preparation

IHCs (n = 224) were studied in acutely dissected organs of Corti from CD-1 mice (Swiss CD-1, Charles Rivers, Margate, UK) from postnatal day 14 (P14) to P27, where the day of birth is P0. To investigate any effects of BAPTA on the delayed-rectifier K⁺ current 15 immature IHCs (P9) were also studied. Mice were killed by cervical dislocation in accordance with UK Home Office regulations. The organs of Corti were dissected and transferred to a microscope chamber and immobilized using a nylon mesh fixed to a stainless steel ring. The chamber (volume 2 ml) was perfused at a flow rate of about 10 ml h⁻¹, from a peristaltic pump, with extracellular solution composed of (mm): 135 NaCl, 5.8 KCl, 1.3 CaCl₂, 0.9 MgCl₂, 0.7 NaH₂PO₄, 5.6 d-glucose, 10 Hepes-NaOH, 2 sodium pyruvate. Amino acids and vitamins for Eagle's minimum essential medium (MEM) were added from concentrates (Invitrogen, Paisley, UK). The pH was adjusted to 7.5 and the osmolality was about 308 mosmol kg⁻¹. The organs of Corti were observed with an upright microscope (Zeiss ACM, Germany or Olympus, Japan) with Nomarski differential interference contrast optics (×40 water immersion objectives). To expose the basolateral surfaces of the cells, a small tear was made in the epithelium with a suction pipette (tip diameter about 2–4 μm) filled with normal extracellular solution. Only cells of healthy appearance (criteria included smooth surface of the cell membrane, absence of vacuoles in the cytoplasm and lack of Brownian motion of mitochondria) and with well-preserved hair bundles were investigated. The position of IHCs along the cochlea was recorded as fractional distance from the extreme apex. Cells were positioned between 0.06 and 0.19 in the apex and between 0.81 and 0.93 in the base, corresponding to approximate frequency ranges of 0.8–3.0 kHz and 49–75 kHz, respectively (using eqn (13) in Ehret, 1975).

Electrical recording

Membrane currents from IHCs under voltage clamp were studied at room temperature (22–25°C) by the whole-cell patch clamp technique using EPC-8 (HEKA, Lambrecht, Germany) or Optopatch (Cairn Research Ltd, Faversham, UK) amplifiers. To obtain realistic voltage responses, all current clamp experiments (Fig. 9) were performed near body temperature (35–37°C). Patch pipettes were pulled from soda glass capillaries (Harvard Apparatus Ltd, Edenbridge, UK) and electrode resistances in extracellular solution were 2–3 MΩ. In order to reduce the electrode capacitance, the shank of the electrode was coated with surf wax (Mr Zogs SexWax, Carpinteria, CA, USA).

Figure 9. Role of I_K,f in shaping voltage responses in mature IHCs.

A and B, voltage responses under current clamp from apical-coil IHCs in the presence of 1 mm EGTA (A, P22) or 30 mm BAPTA (B, P22) in the KCl-based intracellular solution. Current steps were applied between −100 pA and +2000 pA, starting from the resting potential, and for clarity only a few responses are shown. Note the larger and slower voltage responses in 30 mm BAPTA. Recordings of injected current are shown above. EGTA (1 mm): V_m−73 mV; C_m 11.2 pF; R_s 4.6 MΩ; g_leak 12 nS; temperature 37°C. BAPTA (30 mm): V_m−74 mV; C_m 9.5 pF; R_s 4.9 MΩ; g_leak 3.8 nS; temperature 37°C. C, membrane currents in response to depolarizing voltage steps from −84 mV to the various test potentials shown by some of the traces. Cells as in A and B. D, exponential fits to the onset of the voltage responses to 100 pA depolarizing current for the two Ca²⁺ buffers. Cells as in A and B. E, time constants of voltage responses of apical-coil IHCs as a function of current injection with either 1 mm EGTA (n = 11, P22) or 30 mm BAPTA (n = 10, P22) in the intracellular solution. F, non-linear behaviour of the peak and steady-state (40 ms) voltage responses of IHCs, obtained by applying depolarizing and hyperpolarizing current steps from rest. EGTA (1 mm): n = 14, P22; 30 mm BAPTA; n = 14, P22. G, expanded version of panel F, over the range −90 pA to +130 pA.

The pipette filling solution for most of the current recordings contained (mm): 131 KCl, 3 MgCl₂, 5 Na₂ATP, 1 EGTA-KOH, 5 Hepes-KOH, 10 sodium phosphocreatine (pH 7.3, 292 mosmol kg⁻¹). For the experiments designed to determine the specific roles of cytosolic Ca²⁺ on I_K,f, 0.1 mm, 1 mm or 30 mm of the calcium buffer BAPTA (Molecular Probes, Leiden, the Netherlands) was used instead of EGTA. From the access resistance of the electrodes (Oliva et al. 1988), we can estimate that the time constant of diffusion from the patch pipette to the cytoplasm for a substance the size of BAPTA is in the order of 1 min. Therefore, all currents recorded in the presence of BAPTA were obtained between 3 and 14 min from breaking the membrane patch to allow the cytoplasm to equilibrate with the pipette solution. The role of Ca²⁺ released from intracellular stores on activation of I_K,f was assessed by using either 10 mm caffeine and 1 μm thapsigargin or 20 μm or 30 μm ryanodine (Calbiochem, Nottingham, UK) in the above KCl-based intracellular solution. In some cases, including a number of the experiments to assess the effects of caffeine and thapsigargin, a potassium gluconate-based solution was used since its greater viscosity, compared to the normal KCl-based solution, is thought to preserve the cell's intracellular environment more effectively. Its composition was (mm): 110 KOH-gluconate, 16 KCl (11 in the presence of caffeine and thapsigargin), 3 MgCl₂, 5 Na₂ATP, 1 EGTA-KOH, 5 Hepes-KOH, 10 sodium phosphocreatine (pH 7.3, 294 mosmol kg⁻¹).

In some experiments 15 mm 4-AP (Fluka, Gillingham, UK) was added to the KCl-based intracellular solution mentioned above in order to block the delayed rectifier I_K,s (Figs 2C–E and 8C, D and I). When 30 mm BAPTA, 10 mm caffeine or 15 mm 4-AP was used in the intracellular solution, the concentration of KCl was adjusted to keep the osmolality constant. Finally, in the experiments designed to define the nature of I_K(Ca) (Fig. 8H) the following CsCl-based intracellular solution was used (mm): 128 CsCl, 3 MgCl₂, 5 Na₂ATP, 1 EGTA-KOH, 5 Hepes-KOH, 15 4-AP, 10 sodium phosphocreatine (pH 7.3, 292 mosmol kg⁻¹).

Figure 2. Kinetics of I_K,f activation.

A, total control current (I_K,f+I_K,s), current in the presence of 30 nm IbTx (I_K,s) and the IbTx-sensitive current (‘I_K,f’), obtained by subtracting I_K,s from the control current, at a membrane potential near −10 mV. B, isolated I_K,f at −10 mV (dotted trace, with single exponential fit superimposed) obtained by subtracting I_K,s from the total control current (I_K,f+I_K,s) after the current traces were corrected point-by-point for the uncompensated series resistance (see Methods). Recordings in A and B are from the same IHC as in Fig. 1. C, current recordings from an apical-coil IHC (P19) in which I_K,f was isolated by superfusing the cell with 100 μm linopirdine and with 15 mm 4-AP in the KCl-based intracellular solution. D, currents recorded from the same cell as in C but with the additional superfusion of 60 nm IbTx. E, single exponential fit (continuous line) to the activation of I_K,f (dotted trace) at the test potential of −12 mV, from panel C. C_m 10.0 pF; R_s 1.0 MΩ; g_leak 3.5 nS; temperature 24°C. F, activation time constants for I_K,f as a function of membrane potential in apical-coil (Ac: n = 16, P19–P22) and basal-coil (Bc: n = 11, P12–P26) IHCs.

Figure 8. Characterization of the additional Ca²⁺-activated K⁺ current.

A, effect of Ca²⁺-free solution (red traces) on the total outward K⁺ current (black traces) recorded when long voltage steps (4 s) were applied. Membrane currents were elicited in response to depolarizing voltage steps from −94 mV to the various test potentials shown by some of the traces, from the holding potential of −84 mV. B, steady-state I–V curves for the recordings shown in A and the isolated Ca²⁺-sensitive K⁺ current (in blue). P18 apical coil: C_m 9.8 pF; R_s 1.1 MΩ; g_leak 5.4 nS; temperature 24°C. C and D, currents recorded in apical-coil IHCs for voltage steps near −14 mV using 15 mm 4-AP in the intracellular KCl-based solution, and 100 μm linopirdine and 60 nm IbTx in the extracellular solution in order to block I_K,s, I_K,n and I_K,f, respectively. Under these conditions the effects of 50 μm nifedipine (C) and 5 mm Ca²⁺ (D) were investigated on the remaining outward K⁺ current. C: P20; C_m 9.5 pF; R_s 1.2 MΩ; g_leak 1.6 nS; temperature 23°C. D: P20; C_m 10.9 pF; R_s 1.1 MΩ; g_leak 1.5 nS; temperature 22°C. E, amplitude of the isolated Ca²⁺- or nifedipine-sensitive current measured around −14 mV, from a holding potential of −84 mV, under different conditions. F, double exponential fits to the isolated nifedipine-sensitive current shown in C. G, average time constants (τ₁ and τ₂) for six apical-coil IHCs. H, effects of superfusion of Ca²⁺-free solution on currents recorded in an apical-coil IHC using 15 mm 4-AP in the intracellular CsCl solution and 100 μm linopirdine and 60 nm IbTx in the extracellular solution as in C and D. P21: C_m 11.2 pF; R_s 5.8 MΩ; g_leak 1.4 nS; temperature 24°C. I, effect of a Ca²⁺-free solution on the current recorded in a P18 IHC using, in addition to the pharmacological conditions as in C and D, 5 mm TEA in the extracellular solution. C_m 10.0 pF; R_s 1.6 MΩ; g_leak 1.0 nS; temperature 24°C.

Data acquisition was performed using pCLAMP software (Axon Instruments, Union City, CA, USA) connected to a LabMaster DMA or a Digidata 1320A. Data were filtered, depending on the protocols used at 2.5 or 10 kHz (8 pole Bessel), sampled at 5, 50 or 100 kHz and stored on computer for off-line analysis using Origin software (OriginLab, Northampton, MA, USA). Current recordings were corrected offline for leak conductance (g_leak) of 4.2 ± 0.2 nS (n = 224, P14–P27) measured around −90 mV, as the major outward K⁺ currents activate positive to −80 mV. This voltage was chosen as the best compromise to minimize the (small) contribution of I_K,n to the measured outward current size by incorporating about half of it into the leak conductance (Marcotti et al. 2003a). The residual series resistance (R_s) after compensation (50–90%) was 1.7 ± 0.1 MΩ (n = 224, range from 0.5 to 6.7 MΩ) in mature cells resulting in an average voltage-clamp time constant of 17 μs. Membrane potentials for both voltage- and current-clamp experiments were corrected for residual series resistance and for the liquid junction potential measured between pipette and bath solutions. Liquid junction potentials were −4 mV for the KCl-based intracellular solution, −8 mV in the presence of 30 mm BAPTA and −12 mV for the gluconate-based intracellular solution. The majority of voltage clamp protocols are referred to a holding potential (V_h) of −84 mV. When 30 mm BAPTA was used V_h was −88 mV while in the presence of the gluconate-based intracellular solution V_h was either −82 mV or −92 mV. The holding currents were plotted as zero current.

We attempted to isolate I_K,f using three different methods: by blocking most of the other K⁺ currents expressed in mature IHCs (Fig. 2C–E), by measuring the amplitude of I_K,f at a membrane potential of −25 mV at 1.5 ms from the start of the voltage step (Figs 3, 4 and 6) or by subtracting the current in the presence of 30 nm or 60 nm iberiotoxin (IbTx) from the total control current (Figs 1 and 2A and B). The time point of 1.5 ms for the second method was chosen because the fast-activating I_K,f had nearly reached its steady-state (about 98% for a 0.4 ms activation time constant at −25 mV, see Fig. 2F) while the other, slower, outward K⁺ currents were only just beginning to activate. In the last method, due to uncompensated R_s, subtracting large currents with different activation kinetics might produce an artefactual representation of the subtracted current (Fig. 2A) since each recorded current trace may not have been obtained at a constant voltage throughout the duration of the steps. In order to overcome this problem, point-by-point current–voltage curves were plotted and each nominal voltage was corrected for the residual R_s (Fig. 2B). Linear interpolation of the current–voltage (I–V) plots allowed the reconstruction of the corrected current trace at a fixed membrane potential (Santos-Sacchi et al. 1997).

Figure 3. Extracellular Ca²⁺ does not block I_K,f.

A and C, control currents (continuous traces) and currents in the absence of extracellular Ca²⁺ (dotted traces) recorded from an apical-coil P18 (A) and a basal-coil P22 (C) IHC. Membrane currents were elicited in response to depolarizing voltage steps (10 mV increments) from −104 mV to the various test potentials shown by some of the traces, starting from a holding potential of −84 mV. B and D, I–V curves measured at 1.5 ms from the start of the voltage steps, for the recordings shown in A and C, respectively, and for the Ca²⁺-sensitive current obtained by subtracting the current in the absence of extracellular Ca²⁺ from the control current. A and B: C_m 9.8 pF; R_s 1.1 MΩ; g_leak 5.4 nS; temperature 24°C. C and D: C_m 10.9 pF; R_s 1.4 MΩ; g_leak 5.8 nS; temperature 24°C. E–G, amplitude of I_K,f measured at 1.5 ms and at −25 mV in both apical and basal IHCs before and during superfusion of a Ca²⁺-free solution (E, Ac: P18-P22; Bc: P22-P24), 30 μm nifedipine (F, Ac: P17–P22; Bc: P22–P26) or 5 μm Bay K 8644 (G, Ac and Bc: P22). Numbers of cells are indicated above each pair of columns.

Figure 4. The size of I_K,f is affected by intracellular Ca²⁺ buffering.

A and B, typical currents recorded in the presence of 0.1 mm and 30 mm BAPTA, respectively, from basal-coil IHCs. Currents were recorded by applying voltage steps in 10 mV increments starting from the holding potential. C, I–V curves measured at 1.5 ms from the start of the voltage steps for the recordings shown in A and B. BAPTA (0.1 mm), P25: C_m 10.6 pF; R_s 1.5 MΩ; g_leak 6.3 nS; temperature 24°C. BAPTA (30 mm), P26: C_m 10.7 pF; R_s 1.1 MΩ; g_leak 1.7 nS; temperature 24°C. D, amplitude of the isolated I_K,f, measured at −25 mV, in both apical (Ac) and basal (Bc) coil IHCs using 0.1 mm, 1 mm and 30 mm intracellular BAPTA. Numbers of cells are indicated above each column.

Figure 6. Effects of Ca²⁺ stores on I_K,f amplitude and activation.

A and B, typical outward K⁺ currents recorded in basal-coil IHCs using a potassium gluconate based intracellular solution (A) and when 10 mm caffeine and 1 μm thapsigargin were added (B). Membrane currents were elicited in response to depolarizing voltage steps (10 mV increments) from −102 mV to the various test potentials shown by some of the traces, starting from the holding potential of −82 mV. C, outward K⁺ currents in a basal IHC with 20 μm ryanodine in a KCl-based intracellular solution. Holding potential −84 mV, depolarizing voltage steps from −104 mV. A: C_m 10.8 pF; R_s 1.0 MΩ; g_leak 5.8 nS; temperature 24°C. B: C_m 10.9 pF; R_s 0.6 MΩ; g_leak 2.3 nS; temperature 23°C. C: C_m 10.8 pF; R_s 1.1 MΩ; g_leak 3.9 nS; temperature 24°C. D, I–V curves measured 1.5 ms from the start of the voltage steps for the recordings shown in A–C. E, amplitude of the isolated I_K,f, measured at 1.5 ms and −25 mV in both apical and basal IHCs using KCl and potassium gluconate based intracellular solutions. F, average steady-state activation curves, measured at 0.1–0.2 ms from stepping to a 2–3 ms test potential near −44 mV, for the isolated I_K,f in apical-coil IHCs. The continuous lines are fits using eqn (2). Fitting parameters are: Control, P20–P22, •, I_max = 673 pA, V_½ = −46.2 mV, S = 8.5 mV, n = 6; 10 mm caffeine +1 μm thapsigargin: ○, P21−P22, I_max= 735 pA, V_½= −24.3 mV, S = 8.8 mV, n = 3. G, V_½ (left panel) and S (right panel) for the activation curves shown in F. H, half-activation (V_½)voltage of I_K,f as a function of BAPTA concentration. The equivalent BAPTA concentration for 1 mm EGTA with and without caffeine and thapsigargin is estimated. Values for EGTA V_½ from g and BAPTA V_½ from Fig. 5D.

Figure 1. I_K,f is selectively blocked by IbTx.

A and B, total control current (I_K,f+I_K,s) and current in the presence of 30 nm IbTx (I_K,s) recorded from a P19 IHC from the apical coil of the cochlea. Membrane currents were elicited in response to depolarizing voltage steps (10 mV increments) from −104 mV to the various test potentials shown by some of the traces, starting from the holding potential of −84 mV. C, steady-state I–V curves for the cell shown in A and B. C_m 8.5 pF; R_s 2.1 MΩ; g_leak 3.5 nS; temperature 24°C. Current recordings in this and in the following figures are single traces, unless otherwise stated.

Extracellular superfusion

The K⁺ channel blockers IbTx (Tocris, Bristol, UK), apamin (Calbiochem, Nottingham, UK) and linopirdine (RBI, Natick, MA, USA) were extracellularly superfused in order to specifically abolish I_K,f, I_SK and I_K,n, respectively. The role of Ca²⁺ channels in the activation of the outward K⁺ current was investigated by extracellular application of nifedipine (Sigma), Bay K 8644 (Calbiochem) or a Ca²⁺-free solution containing 0.5 mm EGTA, in which MgCl₂ was increased to 3.9 mm to keep membrane charge screening approximately constant (Blaustein & Goldman, 1968). In a few experiments, extracellular TEA (Fluka) was used in order to test its effects on I_K(Ca). When TEA was added to the extracellular solution equimolar substitution of NaCl was used in order to keep the osmolality constant. The extracellular solutions containing drugs were applied through a multibarrelled pipette positioned close to the patched hair cell.

Statistical analysis

Statistical comparisons of means were made by Student's two-tailed t test or, for multiple comparisons, analysis of variance, usually one-way ANOVA followed by Tukey's test (Figs 3E–G, 4D, 5D, 6E, 7G and I, and 8E). Two-way ANOVA, followed by the Bonferroni test, was used to compare current properties between apical and basal IHCs during development (Figs 2F, 5E and F, and 9E and F). P < 0.05 was used as the criterion for statistical significance. Mean values are quoted ±s.e.m. in text and figures. In the bar graphs statistically significant differences are indicated by an asterisk.

Figure 5. Effects of BAPTA concentration on activation and kinetics of I_K,f.

A and B, membrane currents recorded from apical-coil IHCs when 0.1 mm and 30 mm BAPTA were used as the intracellular Ca²⁺ buffer. Currents are averages from six and 10 repetitions, respectively. A: C_m 11.0 pF; R_s 1.2 MΩ; g_leak 6.6 nS; temperature 23°C. B: C_m 10.8 pF; R_s 1.0 MΩ; g_leak 4.6 nS; temperature 23°C. C, average steady-state activation curves for I_K,f in the presence of 0.1 mm, 1 mm and 30 mm BAPTA. The continuous lines are fits using eqn (2). Fitting parameters are: 0.1 mm BAPTA, •, I_max= 1341 pA, V_½= −50.4 mV, S = 11.8 mV, n = 6; 1 mm BAPTA: ▵, I_max= 891 pA, V_½= −30.1 mV, S = 10.8 mV, n = 3; 30 mm BAPTA: ○, I_max= 1148 pA, V_½= −13.1 mV, S = 10.6 mV, n = 9. D, V_½ (left panel) and S (right panel) for the activation curves shown in C. Numbers of cells above each column. E, time constant of activation of I_K,f, fitted using a single exponential, in the presence of different BAPTA concentrations. F, time constants of deactivating tail currents recorded at a membrane potential near −47 mV following a series of 2–3 ms voltage steps from the holding potential. Number of cells in E and F as in D.

Figure 7. An additional Ca²⁺-activated K⁺ current is sensitive to extracellular Ca²⁺.

A and D, effects of superfusing Ca²⁺-free solution (dotted traces) on the total outward K⁺ current (continuous traces) in a P18 apical-coil (A) and a P22 basal-coil (D) IHC. Membrane currents were elicited in response to depolarizing voltage steps (10 mV increments) from −104 mV to the various test potentials shown by some of the traces, starting from a holding potential of −84 mV. Membrane potentials in italics refer to currents in the presence of Ca²⁺-free solution. B and E, steady-state I–V curves for the recordings shown in A and D, respectively, and the current sensitive to extracellular Ca²⁺ isolated by subtraction. C and F, enlarged representation of the small isolated Ca²⁺-sensitive K⁺ current. Panels A–F same cells as Fig. 3A–D. G, amplitude of the Ca²⁺-sensitive K⁺ current in apical-coil (P18-P22) and basal-coil (P22–P26) IHCs obtained by subtracting the current recorded close to −14 mV in the presence of a Ca²⁺-free solution or 30 μm nifedipine from the control current. H, examples of the Ca²⁺-sensitive current in apical and basal IHCs obtained by subtraction and fitted by a single exponential. Apical coil: same cell as A. Basal-coil IHC (P23): C_m 10.8 pF; R_s 1.8 MΩ; g_leak 2.0 nS; temperature 24°C. I, time constant of activation of the Ca²⁺-sensitive current isolated in apical and basal IHCs using either a Ca²⁺-free solution or 30 μm nifedipine.

Results

Kinetics and lack of dependence on extracellular Ca²⁺ of I_K,f

The highly specific BK channel blocker IbTx blocks the IHC current I_K,f at low nanomolar concentrations (Kros et al. 1998). We employed this specificity to isolate I_K,f from the total outward current of mature IHCs for the purpose of kinetic analysis. A saturating dose of IbTx (30 nm) was applied during repeated presentations of a series of depolarizing voltage steps, 50–170 ms in duration, from −104 mV in 10 mV increments from the holding potential of −84 mV. Figure 1A and B shows that extracellular application of 30 nm IbTx completely removed the fast I_K,f and left mainly the slowly activating delayed rectifier I_K,s (Kros & Crawford, 1990; Kros et al. 1998). The steady-state I–V relations for the control current and the current in the presence of IbTx are shown in Fig. 1C. To study the activation time course of the isolated I_K,f, the current remaining in the presence of IbTx (I_K,s) was subtracted from the control current (I_K,f+I_K,s, Fig. 2A). After subtraction, the IbTx-sensitive current (I_K,f) showed rapid activation followed by a decline (Fig. 2A, dotted trace) similar to that previously reported in guinea-pig IHCs when the same current was isolated using a similar procedure (Skinner et al. 2003; Kimitsuki et al. 2003). However, subtracting large currents with different activation kinetics tends to produce an artefactual representation of the subtracted current due to the voltage drop across the uncompensated R_s (see Methods). Therefore, traces recorded in control conditions and in the presence of IbTx were corrected point-by-point for the uncompensated R_s before subtraction. Figure 2B shows for the same cell the total current (I_K,f+I_K,s) and the current in the presence of IbTx (I_K,s) at −10 mV when the above procedure was applied. Under this condition, a true representation of the isolated I_K,f (dotted trace) could be obtained by subtracting I_K,s from the total current (Fig. 2B). The isolated I_K,f showed rapid activation and no signs of inactivation for voltage steps of up to 170 ms. The fitted single-exponential time constant for the activation of the isolated I_K,f was 0.41 ms.

An alternative way to isolate I_K,f would be to block all the other K⁺ currents expressed in mature IHCs. These consist of the delayed-rectifier K⁺ currents, I_K,s (80% of which is 4-AP sensitive) and the much smaller linopirdine-sensitive I_K,n (Marcotti et al. 2003a). We attempted to block these currents by using 15 mm 4-AP in the intracellular solution (Kros & Crawford, 1990) and superfusing IHCs with 100 μm linopirdine. Under these conditions the outward K⁺ current (Fig. 2C and E) resembled that obtained after the subtraction procedure (Fig. 2B, dotted trace). To determine whether 15 mm 4-AP and 100 μm linopirdine were sufficient to isolate I_K,f, we additionally superfused the same IHCs with a fully blocking concentration of IbTx (30 nm or 60 nm). In the presence of IbTx (Fig. 2D) a small residual IbTx-insensitive current was isolated which at 0 mV and at the steady state represented about 10.2 ± 2.1% (n = 6) of the current recorded in the presence of 4-AP and linopirdine alone. The composition of this residual current will be examined further on in the Results (Figs 7 and 8). This result suggests that the best method for isolating I_K,f was by subtracting the current in the presence of IbTx from the total outward K⁺ current (Fig. 2B) rather than by trying to block all the other K⁺ currents (Fig. 2C). However, since the kinetics of I_K,f are much faster than I_K,s, its time constant of activation, evaluated by using a single exponential fit, did not vary significantly between the methods of isolation used (for comparison see Fig. 2B and E). The time constant of I_K,f measured at room temperature and at different membrane potentials was found not to differ significantly between apical (P19–P22, n = 16) and basal (P22–P26, n = 11) IHCs (Fig. 2F). To compare the activation kinetics of I_K,f in mouse IHCs with those previously obtained in apical-coil guinea-pig IHCs (Kros & Crawford, 1990), we fitted the activation in six apical cells to the sigmoidal equation used by Kros & Crawford (1990):

where I(t) is the current amplitude at time t, I₀ is the current at t = 0, I_∞ is the steady-state current and τ₁ and τ₂ are the activation time constants. We found that in mice τ₁ (0.51 ± 0.04 ms, n = 6) and τ₂ (0.17 ± 0.02 ms) measured at a membrane potential around −37 mV were about two times faster than those recorded, also at room temperature, in the guinea-pig IHCs (Kros & Crawford, 1990). In these same mouse IHCs the time constant obtained by a single exponential fit (0.53 ± 0.05 around −37 mV, n = 6) was very similar to the principal time constant τ₁, for which it thus provided a good approximation. The kinetics of deactivation were studied for tail currents near −46 mV following 2 or 3 ms conditioning steps to more depolarized potentials and could be well fitted with a single exponential time constant of about 0.5 ms for apical-coil IHCs. The deactivation kinetics were about three times faster than in guinea-pig IHCs at room temperature (Kros & Crawford, 1990).

Although there is no doubt about the high IbTx sensitivity of I_K,f, one issue of controversy is its Ca²⁺ sensitivity. An early study reported that in guinea-pig IHCs I_K,f was not altered by the removal of extracellular Ca²⁺ (Kros & Crawford, 1990). However, a few years later it was shown, using the same preparation, that when L-type Ca²⁺ channels were selectively blocked by nifedipine a K⁺ current sensitive to Ca²⁺ could be isolated (Dulon et al. 1995). In order to determine whether I_K,f was directly affected by extracellular Ca²⁺ we measured its amplitude, before and during different Ca²⁺ manipulations, at a membrane potential of −25 mV and at 1.5 ms from the start of the voltage steps. Figure 3A and C shows examples of the total outward K⁺ current recorded from an apical- and basal-coil IHC, respectively, before and during the superfusion of a Ca²⁺-free solution. Current recordings were obtained by applying depolarizing voltage steps in 10 mV nominal increments from −104 mV starting from the holding potential of −84 mV. The I–V curves (measured at 1.5 ms) for the control current, the current in the presence of a Ca²⁺-free solution and the negligible isolated Ca²⁺-sensitive current, obtained by subtracting the current recorded during superfusion of a Ca²⁺-free solution from the control, are shown in Fig. 3B (apical coil) and D (basal coil). As shown in Fig. 3E the amplitude of I_K,f at −25 mV in both apical (P17–P22) and basal (P22–P24) IHCs was not significantly affected by the superfusion of a Ca²⁺-free solution suggesting that the Ca²⁺ current, known to be present in mature IHCs (see Fig. 9E in Marcotti et al. 2003b), is not directly involved in the activation of I_K,f. Although this result confirms earlier observations in isolated IHCs (Kros & Crawford, 1990), we considered whether the inability of the Ca²⁺-free solution to block I_K,f might be due to incomplete exchange of solution around the IHCs in this multicellular preparation. This was not the case since the Ca²⁺ current was rapidly abolished when cells were superfused with Ca²⁺-free solution (data not shown) and moreover I_K,f was completely blocked by nanomolar concentrations of IbTx (Fig. 1). The lack of a direct interaction between Ca²⁺ current and I_K,f was verified by superfusing nine IHCs with nifedipine and six IHCs with Bay K 8644, a known blocker and activator for the dihydropyridine (DHP)-sensitive L-type Ca²⁺ channel, respectively (Platzer et al. 2000). When 30 μm nifedipine or 5 μm Bay K 8644 was superfused onto IHCs (P17–P22 for nifedipine, Fig. 3F; P22–P26 for Bay K 8644, Fig. 3G) the size of I_K,f was not significantly affected in either apical or basal IHCs. Note that in basal-coil IHCs the current was slightly reduced (by 12%) by a Ca²⁺-free solution and increased (by 6%) by Bay K 8644, although the effects were not statistically significant. This is probably due to the presence of an additional Ca²⁺-sensitive K⁺ current that is different from I_K,f and is faster and larger in basally located cells than in apical cells (see Fig. 7).

Nature and modulation of the Ca²⁺-activated K⁺ current I_K,f

Since I_K,f was insensitive to extracellular Ca²⁺ a series of experiments was performed to evaluate its sensitivity to cytosolic Ca²⁺. We attempted to reduce any changes in free cytosolic Ca²⁺ concentration near the Ca²⁺-activated K⁺ channels either by using high concentrations of the fast Ca²⁺ buffer BAPTA (Neher, 1998) or by removing the contribution from the intracellular Ca²⁺ stores (Verkhratsky, 2002). First of all, BAPTA (0.1 mm, 1 mm and 30 mm) was used in the intracellular solution instead of EGTA (1 mm). Figure 4A and B shows basal hair cells that were recorded from using 0.1 mm or 30 mm BAPTA as the internal Ca²⁺ buffer. The membrane potential was stepped from −104 mV (0.1 mm BAPTA) or −108 mV (30 mm BAPTA) to different test potentials in 10 mV increments from the holding potential of −84 mV or −88 mV, respectively. Hair cells buffered with 30 mm BAPTA had smaller initial and steady-state currents than did those buffered with 0.1 mm BAPTA, suggesting that I_K,f has been substantially reduced. To test whether the delayed rectifier K⁺ current was also affected by BAPTA, we compared the size of the outward K⁺ current in P9 IHCs, a time at which I_K,f is not yet expressed (Kros et al. 1998; Marcotti et al. 2003a), using different intracellular BAPTA concentrations or EGTA. The amplitude of the delayed rectifier K⁺ current (measured 160 ms following a voltage step to −25 mV) was not significantly different between the above experimental conditions (0.1 mm BAPTA: 1.6 ± 0.2 nA, n = 2; 1 mm BAPTA: 1.3 ± 0.2 nA, n = 3; 30 mm BAPTA: 1.4 ± 0.3 nA, n = 4; 1 mm EGTA: 1.2 ± 0.1 nA, n = 6).

The I–V curves for the isolated I_K,f (measured 1.5 ms from the start of the voltage step) in Fig. 4A and B show that the high BAPTA concentration shifted the potential at which this current first became evident towards more depolarized potentials (Fig. 4C). The amplitude of I_K,f in the presence of different BAPTA concentrations was measured in both apical (P20–P22) and basal (P22–P25) IHCs (Fig. 4D). In both regions the size of I_K,f at −25 mV decreased significantly (P < 0.001) with increasing BAPTA concentration.

Steady-state activation of the isolated I_K,f in the presence of 0.1 mm, 1 mm and 30 mm of BAPTA was evaluated by analysing tail currents at a fixed membrane potential. From the holding potential of −84 mV or −88 mV (when 30 mm BAPTA was used), the membrane potential was stepped to a test potential near −47 mV after a series of short conditioning steps (2 ms or 3 ms in duration) in 10 mV increments from −104 mV or −108 mV (Fig. 5A and B). The activation curves in Fig. 5C were obtained by plotting the normalized instantaneous tail currents (measured at 0.1–0.2 ms from stepping to the test potential) against the different conditioning potentials. Data were fitted by a single first-order Boltzmann equation:

where I is the tail current amplitude, I_max is the maximal tail current, V_½ is the potential of half-maximal activation, V is the membrane potential of the conditioning voltage step and the slope factor S describes the voltage sensitivity of activation.

Increasing the intracellular BAPTA concentration from 0.1 mm to 30 mm shifted the activation curve by about 37 mV towards a more depolarized potential in apical IHCs (P < 0.0001). The slope factor and maximal tail current did not vary significantly. Figure 5D shows the mean values for V_½ (left panel) and S (right panel) when 0.1 mm, 1 mm and 30 mm BAPTA were used. The activation time constants also varied significantly between 0.1 mm, 1 mm and 30 mm BAPTA (P < 0.0001), with the slower time constants at a given potential in 30 mm BAPTA (Fig. 5E) corresponding to a depolarizing shift of about 20 mV. Between 0.1 mm and 1 mm the depolarizing shift was about 5 mV. Figure 5F shows the time constants of deactivating tail currents measured around −47 mV in the presence of the above BAPTA concentrations. The tail current kinetics were also significantly different between the three conditions (P < 0.0001) but becoming faster with increasing BAPTA concentration.

The above results show that although extracellular Ca²⁺ does not directly affect the activation of I_K,f, variations in cytosolic Ca²⁺ buffering shifted its steady-state activation, suggesting that I_K,f could potentially be modulated by Ca²⁺ released from intracellular stores. We investigated this hypothesis by adding either caffeine and thapsigargin or ryanodine to the intracellular solution. Caffeine is a xanthine that causes the release of Ca²⁺ from stores by affecting ryanodine receptors in such a way that their affinity for cytoplasmic Ca²⁺ is increased (Pozzan et al. 1994). On the other hand, thapsigargin is a known inhibitor of sarcoplasmic/endoplasmic reticulum Ca²⁺-ATPase (SERCA) (Thastrup, 1990), thus preventing the refilling of Ca²⁺ stores. Using both molecules together would initially empty the Ca²⁺ stores, due to the action of caffeine, which would not be able to refill, due to thapsigargin, and therefore remove any contribution that the stores have on cytosolic Ca²⁺ modulation. IHCs were dialysed for up to 14 min after reaching the whole cell configuration with a KCl or potassium gluconate based intracellular solution containing both caffeine (10 mm) and thapsigargin (1 μm). Typical examples of membrane currents recorded, using a potassium gluconate intracellular solution, from basal-coil IHCs in control conditions and when both caffeine and thapsigargin were used are shown in Fig. 6A and B, respectively. Ryanodine at high concentrations (greater than 10 μm) causes a block of the Ca²⁺ efflux from the stores (Meissner, 1986). Figure 6C shows membrane currents of a basal-coil IHC in the presence of 20 μm ryanodine in a KCl-based intracellular solution. Note that when caffeine and thapsigargin or ryanodine were present the fast initial current was reduced in size. The I–V curves (Fig. 6D), measured at 1.5 ms, indicate that in the presence of the drugs I_K,f started to activate at a more depolarized potential, similar to that found when 30 mm BAPTA was used (Fig. 4C). The amplitude of the isolated I_K,f in apical and basal IHCs (at 1.5 ms and at a membrane potential of −25 mV) in control conditions and in the presence of the drugs is shown in Fig. 6E. The amplitude of I_K,f was significantly reduced in the presence of caffeine and thapsigargin compared to that measured in control conditions (KCl based: Ac, P < 0.05; Bc, P < 0.01. Potassium gluconate based: Bc, P < 0.01). Ryanodine also reduced the size of I_K,f (KCl based: Ac, P < 0.01; Bc, P < 0.05). The amplitude of I_K,f at −25 mV recorded using 1 mm EGTA as the Ca²⁺ buffer under control conditions (Fig. 6E) was found not to differ significantly from that recorded with either 1 mm or 0.1 mm BAPTA (Fig. 4D). The modulation of the activation of I_K,f by the intracellular Ca²⁺ stores was further investigated in apical IHCs by comparing tail currents recorded in control conditions and when both caffeine and thapsigargin were added to the intracellular solution. Tail currents were recorded at a fixed membrane potential following short (2 ms or 3 ms in duration) depolarizing voltage steps that fully activated I_K,f (currents not shown), as in Fig. 5A and B. The activation curves in Fig. 6F were obtained by plotting the normalized instantaneous tail currents measured at around −44 mV against the different conditioning potentials. Data were fitted by a single first-order Boltzmann equation (eqn (2)). When both caffeine and thapsigargin were used the activation of I_K,f was shifted positively by about 22 mV with a V_½ (Fig. 6G, left panel) significantly more depolarized (P < 0.0001) than that in control conditions. Again the slope factor (Fig. 6G, right panel) and I_max did not vary significantly. Activation was slowed and deactivation speeded up by caffeine and thapsigargin, but to a lesser extent than in 30 mm BAPTA. Figure 6H shows V_½ for the current recorded using 1 mm EGTA as the intracellular Ca²⁺ buffer expressed as an equivalent BAPTA concentration. In control conditions V_½ corresponded to about 0.2 mm BAPTA. In the presence of caffeine and thapsigargin the V_½ was equivalent to about 3 mm BAPTA.

Do mature IHCs express another Ca²⁺-activated K⁺ current?

The experiments so far indicate that I_K,f is not directly activated by extracellular Ca²⁺. However, the total outward K⁺ current was not completely insensitive to extracellular Ca²⁺ since the steady-state current was reduced in amplitude in both apical and basal IHCs when a Ca²⁺-free solution was applied (Fig. 7A and D respectively, dotted traces). This became evident for voltage steps lasting longer than a few milliseconds (cf. Fig. 3A and C, which shows currents from the same apical and basal IHCs in response to shorter voltage steps). The steady-state I–V curves, measured at 160 ms, for the control, the current in the presence of a Ca²⁺-free solution and the isolated Ca²⁺-activated K⁺ current, obtained by subtracting the current in the absence of Ca²⁺ from the control current, are shown in Fig. 7B (apical) and E (basal). The Ca²⁺-activated currents are shown expanded in Fig. 7C and F, where the bell shape of the I–V curve can be appreciated. The maximal amplitude of the Ca²⁺-activated K⁺ current, measured around −14 mV, was obtained by subtraction when apical and basal IHCs were superfused with either a Ca²⁺-free solution or 30 μm nifedipine. Using either of these isolation procedures, the Ca²⁺-sensitive K⁺ current was significantly (P < 0.01) larger in basal IHCs compared to that isolated in apical cells (Fig. 7G). The Ca²⁺-sensitive current could be isolated by subtracting the current recorded in a Ca²⁺-free solution from the control current. Examples for an apical and a basal IHC are shown in Fig. 7H and their activation is fitted by a single exponential. The activation time constants of the isolated Ca²⁺-sensitive currents were significantly faster (P < 0.01) in basal than in basal IHCs (Fig. 7I). The range of time constants was quite wide: 2.2 ms to 11 ms in basal IHCs and 12 ms to 48 ms in apical IHCs. The larger size and faster activation of this current may account for the small reduction in the size of the outward current, measured at 1.5 ms, in basal IHCs in a Ca²⁺-free solution (Fig. 3E).

A reduction in the total outward K⁺ current in the presence of a Ca²⁺-free solution was also observed using very long lasting (4 s) voltage steps (Fig. 8A red traces). The ‘steady-state’I–V curves, measured at 4 s, for the control, the current in Ca²⁺-free solution and the isolated Ca²⁺-activated K⁺ current are shown in Fig. 8B. Note that the total outward K⁺ currents, also in the control condition, are considerably smaller than for shorter protocols of up to 170 ms duration (e.g. Fig. 7A). This may be due to a slow inactivation process on a time scale of seconds or to K⁺ accumulation and depletion effects, but was not investigated in detail for this study. Since previous results (Fig. 3) demonstrated that I_K,f is not significantly affected by extracellular Ca²⁺, we designed a series of experiments to find out whether indeed another type of Ca²⁺-activated K⁺ current was present or whether instead either one of the other K⁺ currents known to be expressed in mature IHCs (I_K,s and I_K,n: Marcotti et al. 2003a) was modulated by extracellular Ca²⁺. To differentiate between these possibilities we largely removed the delayed-rectifier current I_K,s, by including 15 mm 4-AP in the intracellular solution, and completely blocked I_K,n by superfusing IHCs with 100 μm linopirdine. The superfused solution also contained a fully blocking concentration of IbTx (30 nm or 60 nm) to remove I_K,f. Under these conditions a small outward current of between 300 pA and 800 pA remained (labelled ‘Control’ in Fig. 8C and D). Recordings were obtained by applying a voltage step to a nominal value of −14 mV from the holding potential of −84 mV. We stepped the voltage to −14 mV as it represents the potential at which the peak of the Ca²⁺-sensitive current occurs in 1.3 mm extracellular Ca²⁺ (Fig. 8B). When the current recorded in the presence of 30 μm or 50 μm nifedipine (Fig. 8C) or a Ca²⁺-free solution (data not shown) was subtracted from the control current, a Ca²⁺-activated current of a similar size to when all outward K⁺ currents were present could still be isolated. In contrast to the responses to 170 ms voltage steps (Fig. 7G), the size of the Ca²⁺-activated K⁺ current in response to 4 s voltage steps was not obviously different between apical and basal IHCs. The size of the Ca²⁺-sensitive current was directly dependent on the extracellular Ca²⁺ concentration since its amplitude increased when 5 mm Ca²⁺ was used instead of 1.3 mm Ca²⁺ (Fig. 8D and E). Any involvement of Ca²⁺ released from intracellular stores in the activation of this Ca²⁺-activated K⁺ current was excluded since similar results to those shown in Fig. 8C and D were obtained in five apical-coil IHCs when 10 mm caffeine and 1 μm thapsigargin were added to the intracellular solution (data not shown). These results suggest that a Ca²⁺-activated K⁺ current, sensitive to extracellular Ca²⁺ flowing through voltage gated Ca²⁺ channels, is expressed in mature IHCs in addition to I_K,f. Using the 4 s voltage step stimulus, the amplitude of the Ca²⁺- or nifedipine-sensitive current was similar, whether it was isolated from the total outward current (Fig. 8E, black column) or when using the different K⁺ channel blockers (Fig. 8E, hatched columns). The persistence of the current sensitive to extracellular Ca²⁺ in the presence of the blockers shows that it is not a component of I_K,s or I_K,n. From now on this Ca²⁺-activated current will be referred to as I_K(Ca). A careful examination of the time course of activation of I_K(Ca) (Fig. 8C and D) revealed that the slow activation was preceded by a much faster component. Figure 8F shows double exponential fits to the faster and the slower component for the isolated nifedipine-sensitive current (Fig. 8C). The two time constants of activation of I_K(Ca) are shown in Fig. 8G. The fast time constant τ₁ was not significantly different from that recorded in apical IHCs for the Ca²⁺- and nifedipine-sensitive current obtained using the 170 ms voltage steps (Fig. 7I). Because the fast and slow components shared the same pharmacological sensitivity we assume they represent a single population of channels. The most likely possibility is that the slow component reflects a slow increase in intracellular Ca²⁺ near the channels during long depolarizations.

The next step was to identify the type of channel responsible for I_K(Ca), in particular whether it is likely to be carried by BK or SK channels. In contrast to BK channels, SK channels are usually permeable to Cs⁺ (Park, 1994; Erostegui et al. 1994) and only blocked by high concentrations of TEA (> 20 mm, Sah, 1996). Under the same pharmacological and recording conditions described above (Fig. 8C and D control traces) we tested the sensitivity of this current in apical IHCs to extracellular TEA (5 mm) and to substitution of intracellular K⁺ by Cs⁺. Our findings show that I_K(Ca) is abolished by Cs⁺ and TEA. First of all, the control outward current was absent in Cs⁺ (Fig. 8H) or much reduced in TEA (Fig. 8I), compared to the control current in Fig. 8C and D. The average control current was −14 ± 20 pA (n = 4, P21) in Cs⁺ and +50 ± 8 pA (n = 4, P18) in TEA. Secondly, superfusion with a Ca²⁺-free solution now produced an increase (rather than a decrease as in Fig. 8C) in the outward K⁺ current, due to block of the inward Ca²⁺ current (I_Ca). I_Ca is evident after subtraction (Fig. 8H and I). The small outward current in the presence of a Ca²⁺-free solution (Fig. 8H and I, red traces) it is likely to be due to a residual I_K,s that was not fully blocked by 4-AP (see Fig. 2D). A similar increase rather than a decrease in outward current upon removal of extracellular Ca²⁺ was observed in four apical IHCs when 30 mm BAPTA replaced 1 mm EGTA as the intracellular Ca²⁺ buffer (data not shown), suggesting that strong buffering also abolished I_K(Ca). The block of I_K(Ca) by intracellular Cs⁺ and by low concentrations of TEA makes the presence of SK channels unlikely. The absence of SK channels has also been previously reported in mature IHCs of the guinea-pig cochlea (Kros & Crawford, 1990; Skinner et al. 2003). This was confirmed in a few experiments similar to that of Fig. 8C in which 300 nm apamin, a concentration that fully blocked the SK current in immature IHCs (authors' unpublished observation), did not reduce the control current during 4 s voltage steps (data not shown) but instead produced a small increase in the residual K⁺ current (of about 20%), possibly due to an interaction with the other K⁺ channel blockers present (4-AP, linopirdine and IbTx).

Effects of I_K,f and I_K(Ca) on voltage responses under current clamp

Since I_K,f is already partially activated at the resting membrane potential of mature IHCs (−73 ± 1 mV, n = 14, apical coil P22), we investigated the functional importance of the current's modulation by cytosolic Ca²⁺ by studying voltage responses to current clamp steps. This gives an approximate idea of the receptor potentials in response to steps of transducer current or to high-frequency auditory stimuli well above the corner frequency (f_3dB) of the low-pass RC filter due to the resistance and capacitance of the cell membrane (Kros, 1996). In control conditions (apical-coil IHCs, P22, 1 mm EGTA in the intracellular solution) hyperpolarizing and depolarizing current injections from the resting membrane potential, elicited fast, graded voltage responses (Fig. 9A) as previously observed in guinea-pig (Kros & Crawford, 1990) and mouse (Kros et al. 1998) IHCs. Current steps larger than about 1000 pA, but well inside the range of transducer currents expected in vivo (Kennedy et al. 2003), caused an initial fast voltage transient (Fig. 9A, 2000 pA) which declined to a steady level of around −45 mV. When age-matched IHCs were recorded in the presence of 30 mm BAPTA instead of 1 mm EGTA in the intracellular solution, both hyperpolarizing and depolarizing current steps caused larger but slower voltage responses (Fig. 9B). Although the reduced I_K,f affects the receptor potential, the resting membrane potential of IHCs was not significantly altered by changing the intracellular Ca²⁺ buffer (1 mm EGTA: −73 ± 1 mV, n = 14, P22; 30 mm BAPTA: −75 ± 1 mV, n = 14, P22). This is consistent with previous experiments in which the resting potential was also not significantly affected when I_K,f was fully blocked using IbTx or TEA (Marcotti et al. 2003a). Figure 9C shows currents under voltage clamp in the presence of 1 mm EGTA or 30 mm BAPTA for the same two IHCs shown in Fig. 9A and B. While I_K,f was large in the presence of 1 mm EGTA (Fig. 9C, top panel) it was almost completely abolished when 30 mm BAPTA was used as seen from the lack of the initial fast current (Fig. 9C, bottom panel). In this cell 30 mm BAPTA was particularly effective for an apical-coil IHC, but qualitatively similar results were obtained in all 14 cells tested under this buffering condition.

To quantify the effects of BAPTA on kinetics of the voltage responses we fitted single exponentials to their onsets, over a range of injected currents between −100 pA and +500 pA, when either 1 mm EGTA or 30 mm BAPTA were used. The fits shown in Fig. 9D are to the voltage responses elicited using a 100 pA depolarizing current injection for the same IHCs as in A and B. The average time constants of the voltage responses measured in apical-coil IHCs (Fig. 9E) in the presence of 1 mm EGTA (n = 11, P22) were significantly faster (P < 0.001) than those obtained when 30 mm BAPTA was used (n = 10, P22). With 30 μm ryanodine in the intracellular solution the time constants of apical-coil IHCs (n = 5, P22) were intermediate (data not shown) in that they were also slowed compared to 1 mm EGTA (P < 0.001), but were faster than in the presence of 30 mm BAPTA (P < 0.005). The relation between injected current and the resulting membrane potential, measured at the peak and at the steady state (40 ms) of the responses in the presence of the two Ca²⁺ buffers, is shown in Fig. 9F. Significantly larger voltage responses over the entire range of injected currents between −100 pA and +2000 pA were obtained when 30 mm BAPTA was used instead of 1 mm EGTA (P < 0.0001, both for steady-state and peak responses). When 30 μm ryanodine was used in the intracellular solution (apical-coil IHCs, n = 8, P22) the voltage responses were again intermediate relative to the other two conditions.

Because reducing the size of I_K,f by 30 mm BAPTA or blocking it altogether with TEA or IbTx had a negligible effect on the resting potential it might be assumed that its conductance at this potential is very small compared to that of, in particular, I_K,n, which is about 70% activated at the resting potential (Marcotti et al. 2003a; Oliver et al. 2003). This is not the case as can be appreciated from the effects of 30 mm BAPTA near the resting potential: it alters both the membrane time constant (Fig. 9E) and the slope of the voltage–current relationship (Fig. 9G). As the peak and steady-state slope conductances were near identical for small current injections, we only quote steady-state values: 9.2 ± 1.0 nS (n = 14) in 1 mm EGTA and 5.1 ± 0.2 nS (n = 14) in 30 mm BAPTA (P < 0.001), near the resting potential. Consistent with this difference in slope conductance, the average membrane time constants for small (±10 pA) current injections were 1.6 ms in EGTA and 3.1 ms in 30 mm BAPTA (from Fig. 9E). I_K,f thus contributes about 50% of the resting conductance of IHCs. The likely explanation for the lack of effect on the resting potential of blocking I_K,f is that its reversal potential (−71.0 ± 1.4 mV, n = 5, calculated from tail currents following a 2 ms step to around −20 mV) is closer to the resting potential than that for I_K,n (−78.8 ± 1.3 mV, n = 13, calculated from the current blocked by 100 μm linopirdine).

Finally we tested the contribution of I_K(Ca) under current clamp by comparing voltage responses in the same IHCs (n = 6, P21–P22, apical coil) before and during superfusion with a Ca²⁺-free extracellular solution. No significant differences were observed over the range of injected currents (40 ms steps) between −100 pA and +2000 pA, for either steady-state or peak responses, although there was a systematic 1 mV depolarizing shift in zero Ca²⁺ (data not shown). This negligible effect of I_K(Ca) is most likely due to its small size relative to the other K⁺ conductances.

Discussion

I_K,f is modulated by cytosolic but not by extracellular Ca²⁺

Around the onset of hearing, IHCs of the mouse cochlea start to express the rapidly activating outward K⁺ current I_K,f. This current appears to be of particular importance for mature mammalian cochlear IHCs since it is a major component of the outward K⁺ current of both guinea-pig and mouse IHCs (Kros & Crawford, 1990; Kros et al. 1998). Although I_K,f was found to be extremely sensitive to IbTx (Kros et al. 1998), therefore showing the typical pharmacology of Ca²⁺-activated K⁺ currents of the BK family (Kaczorowski et al. 1996), the apparent inability of extracellular Ca²⁺ to modulate its activation remained one of the unsolved questions in IHC physiology.

Membrane currents carried by BK channels have been recorded in hair cells from different species and are normally reduced in size when a Ca²⁺-free solution is superfused onto these cells (Hudspeth & Lewis, 1988; Fuchs & Evans, 1990; Fuchs & Sokolowski, 1990; Art et al. 1993; Armstrong & Roberts, 2001) or when a high concentration of intracellular BAPTA is used (Ricci et al. 2000). The fact that intracellular BAPTA affects the amplitude and kinetic properties of I_K,f indicates the direct dependence of K,f channels on cytosolic Ca²⁺. The apparent inability of extracellular Ca²⁺, flowing through voltage-gated Ca²⁺ channels, to affect I_K,f (Fig. 3 and Kros & Crawford, 1990) can be tentatively explained by a lack of colocalization between Ca²⁺ and K,f channels. Although it appears to be missing in mammalian IHCs, colocalization between Ca²⁺ and BK channels is an important requirement in lower vertebrate hair cells in order to set their intrinsic resonant frequency (Art & Fettiplace, 1987; Hudspeth & Lewis, 1988; Fuchs & Evans, 1990; Roberts et al. 1990). In mouse IHCs the absence of a direct interaction between Ca²⁺ and K,f channels is also suggested by the more hyperpolarized activation of I_K,f (starting at about −85 mV defined as 1% of g_max in the presence of EGTA as the intracellular Ca²⁺ buffer: Fig. 6F and Oliver et al. 2003) compared to that of the Ca²⁺ current (positive to −65 mV: Platzer et al. 2000; Marcotti et al. 2003b) recorded in the presence of 1.3 mm Ca²⁺.

The question arises whether the intracellular Ca²⁺ concentration around the K,f channels is constant or whether it varies as a function of membrane potential. A constant Ca²⁺ concentration is unlikely because the slope of the activation curves, at around 8–10 mV per e-fold change in potential (Figs 5D and 6G), is considerably steeper than that of BK channels studied in inside-out patches from IHCs at a constant Ca²⁺ concentration (18 mV: Oliver et al. 2003). The very hyperpolarized activation and the faster activation kinetics of I_K,f (Fig. 6F control and Fig. 2F, respectively) compared to those recorded for BK channels in lower-vertebrate hair cells (Art & Fettiplace, 1987; Hudspeth & Lewis, 1988; Fuchs & Evans, 1990) suggest that in mammalian IHCs I_K,f might be modulated in a voltage-dependent manner by Ca²⁺ coming from a source other than through L-type Ca²⁺ channels. Using ryanodine or a mixture of thapsigargin and caffeine in the intracellular solution we demonstrate a role for intracellular Ca²⁺ stores, known to be present in immature mouse cochlear IHCs (Kennedy & Meech, 2002), in regulating the Ca²⁺ concentration around the K,f channel and thus influencing the activation of I_K,f. Our findings demonstrate the continued presence of intracellular Ca²⁺ stores in mature IHCs. There may be some parallels with the coupling of BK channels and ryanodine receptors in smooth muscle (Pérez et al. 1999). Release of Ca²⁺ from intracellular stores can be induced by Ca²⁺ influx through the cell membrane, for example in smooth and cardiac muscle excitation–contraction coupling (Bolton et al. 1999; Lamb, 2000). As I_K,f is unaffected by the removal of extracellular Ca²⁺ another trigger for Ca²⁺ release from the stores has to be postulated. In skeletal muscle Ca²⁺ release from the stores does not depend on extracellular Ca²⁺ but is mediated by voltage sensors acting on RyR1 ryanodine receptors (Lamb, 2000). It is possible that a similar voltage-dependent mechanism operates in IHCs and it would be interesting to examine the subtypes of ryanodine receptors in these cells.

In the presence of 1 mm BAPTA I_K,f has a more depolarized activation than in 1 mm EGTA (Fig. 6H). Although BAPTA and EGTA have similar affinities for Ca²⁺, BAPTA has a much higher binding rate constant (Naraghi & Neher, 1997). The faster association kinetics enable BAPTA to buffer increases in intracellular Ca²⁺ closer to the source of Ca²⁺ (Neher, 1998). Therefore the different effects observed in the presence of the same concentration (1 mm) of either of the two buffers suggests that the increase in cytosolic Ca²⁺, required for the activation of I_K,f, originates in the close vicinity of the K,f channels. From Naraghi & Neher's (1997) estimate of the space constant for 2 mm BAPTA of 28 nm, the space constant for 1 mm BAPTA would be about greater (see eqn (2) in Neher, 1998). Considering then that 1 mm BAPTA significantly shifted the activation curve of I_K,f when compared to 0.1 mm BAPTA (Fig. 5) we estimate that the Ca²⁺ source is in the order of 40 nm away from the K,f channels. Such a distance is compatible with the very rapid submillisecond activation and deactivation kinetics of the K,f channels which may be in part Ca²⁺ dependent as they are slowed down and speeded up, respectively, by increasing the BAPTA concentration (Figs 2F and 5E and F). However, it is equally possible that the gating kinetics of I_K,f depend on a dynamic equilibrium for cytosolic Ca²⁺ in the vicinity of the channels. Store depletion would then cause a decrease in cytosolic Ca²⁺ at the K,f channels at steady state at any given potential, thus altering gating kinetics in the same fashion as increasing the BAPTA concentration. Although the results presented in this paper show that Ca²⁺ released from intracellular stores is involved in the activation of I_K,f at negative potentials, other mechanisms could also be involved. For example, protein kinase A (PKA) produced a hyperpolarizing shift in I_K,f activation (Jagger & Ashmore, 1999). PKA could potentially act on phosphorylation sites present on both α- and β-subunits and therefore increase the Ca²⁺ sensitivity of the BK channel directly (Toro et al. 1998) or indirectly via activating ryanodine receptors (Reiken et al. 2003).

It thus appears that IHCs are different from auditory hair cells in lower vertebrates in that the majority of BK channels are not coupled to the L-type Ca²⁺ channels that regulate neurotransmitter release (Moser & Beutner, 2000). This leads to a more hyperpolarized activation of I_K,f than that of I_Ca, by about 20 mV, and prevents electrical resonance, important for frequency tuning in auditory hair cells of various lower vertebrates, conceivably because electrical resonance would become progressively less efficient for frequencies above a few kilohertz (Wu et al. 1995). The contribution of I_K,f to the membrane's RC time constant at a given potential can be appreciated by comparing the steady-state slope conductances in 1 mm EGTA and 30 mm BAPTA (Fig. 9F and G). Due to the voltage-dependent activation of the membrane currents, the time constants become faster with depolarization of the membrane potential. Assuming a more depolarized resting potential in vivo (due to the resting transducer current) of say −55 mV, the membrane time constants were 0.30 ms in 1 mm EGTA and 0.82 ms in 30 mm BAPTA. This corresponds to corner frequencies of the low-pass membrane filter of about 530 Hz and 190 Hz, respectively, at this potential. Activation of I_K,f by intracellular Ca²⁺ is thus essential for high-frequency phase locking in the auditory nerve (Palmer & Russell, 1986). The corner frequency of 530 Hz is within the range of 480–940 Hz previously found for guinea-pig IHCs at this potential (Kros & Crawford, 1990).

Mature IHCs express an additional Ca²⁺-activated K⁺ current that is sensitive to extracellular Ca²⁺

An outward K⁺ current sensitive to Ca²⁺ influx through L-type Ca²⁺ channels was isolated even when all major K⁺ currents expressed in mature IHCs (I_K,s, I_K,n and I_K,f) were blocked (Fig. 8C and D). These results indicate that mature mouse IHCs express, in addition to I_K,f, a Ca²⁺-activated K⁺ current (I_K(Ca)) with different pharmacological and kinetic properties and sensitivity to extracellular Ca²⁺. A small current with activation kinetics slower than I_K,f has been identified in IHCs of the guinea-pig where it could be reduced in size when Ca²⁺ channels were blocked by nifedipine (Dulon et al. 1995). Similar to our findings, this small current also showed slower activation when compared with the IbTx-sensitive current (I_K,f) isolated in guinea-pig IHCs (Skinner et al. 2003). IHCs of the mouse and guinea-pig might thus express two distinct types of Ca²⁺-activated K⁺ current only one of which (I_K(Ca)) appears to be colocalized with voltage-gated Ca²⁺ channels (Fig. 10). The finding that I_K(Ca) was larger and activated more rapidly in basal than in apical IHCs in response to 170 ms voltage steps suggests there may be more L-type Ca²⁺ channels and a higher afferent innervation density in high-frequency basal IHCs, as found for turtle auditory hair cells (Ricci et al. 2000). A tonotopic gradient in afferent innervation has been found in the feline cochlea (Liberman et al. 1990), but it is not known whether this also applies to the mouse. The slow increase in I_K(Ca) on a time scale of seconds may be due to local buffer saturation and hence accumulation of Ca²⁺ near the I_K(Ca) channels (Ricci et al. 2000). Since I_K(Ca) could still be isolated when IbTx (up to 60 nm) and apamin (300 nm) were used, its exact nature remains unclear although BK channels with a very low sensitivity to IbTx have recently been described in hair cells of the bullfrog sacculus (Armstrong & Roberts, 2001). Additional evidence for the presence of BK rather than SK channels underlying I_K(Ca) comes from experiments in which extracellular TEA (5 mm) or a Cs⁺-based intracellular solution was used. TEA is a general blocker for a variety of K⁺ currents including the Ca²⁺-activated K⁺ current although its potency depends on the type of channel being investigated (Hille, 2001). While BK channels are usually blocked by submillimolar concentrations of TEA the block of the SK channel requires higher concentrations (> 20 mm: Sah, 1996). A similar variability in the blocking sensitivity for Ca²⁺-activated K⁺ channels also exists with respect to Cs⁺, with BK channels being blocked by the ion whereas a large Cs⁺ current can flow through SK channels of rat adrenal chromaffin cells (Park, 1994) and hair cells from the guinea-pig (Erostegui et al. 1994). In contrast to I_K,f this smaller and slower I_K(Ca) is influenced by Ca²⁺ entering the cell through the L-type Ca²⁺ channels and appears to be insensitive to Ca²⁺ released from intracellular stores. Its function in IHCs is at present not clear as it did not affect voltage responses to 40 ms current injections, but it may turn out to play some role on a longer time scale, for example in regulating neurotransmitter release.

Figure 10. Two mechanisms of activation of Ca²⁺-activated K⁺ channels in mature IHCs.

Schematic representation showing the presence in mature IHCs of two different Ca²⁺-activated K⁺ currents: I_K(Ca) and the larger and faster I_K,f. The experiments described in this paper are consistent with I_K,f being activated by cytoplasmic Ca²⁺ regulated by intracellular stores and I_K(Ca) being directly activated by Ca²⁺ flowing into the cell through voltage-gated Ca²⁺ channels.

Structure and modulation of BK channels

Large-conductance voltage and Ca²⁺-activated K⁺ channels can be expressed as either the integral membrane pore-forming α-subunit alone or as a complex of the α-subunit and the regulatory β-subunit (Vergara et al. 1998; Orio et al. 2002). When coexpressed with the α-subunit, the transmembrane β-subunit produces a hyperpolarizing shift in the voltage range of channel activation by increasing its apparent Ca²⁺ sensitivity although this is usually accompanied by a slowing of the activation and deactivation kinetics (Meera et al. 1996). Functional differences between BK channels (such as Ca²⁺ sensitivity and kinetics) can be explained either by the expression of splice variants of the same gene encoding the α-subunit (Saito et al. 1997) or by differential association with modulatory β-subunits (McManus et al. 1995; Meera et al. 1996). To date, four distinct auxiliary β-subunits for the BK channel have been cloned and are named β₁–β₄ (Orio et al. 2002). Of these subunits β₂ and β₃ confer fast inactivation with a time constant in the order of milliseconds (Xia et al. 2000; Armstrong & Roberts, 2001; Bentrop et al. 2001) upon the BK channels when coexpressed with the α-subunit, and are thus inconsistent with the non-inactivating (in response to voltage steps up to 170 ms) Ca²⁺-activated K⁺ currents in IHCs, I_K,f (Fig. 2B, C) and I_K(Ca) (Fig. 7H).

BK channels in hair cells

Recent findings in chick and turtle cochlear hair cells have shown the expression, along the tonotopic axis of the sensory epithelium, of various BK channel isoforms with different kinetic properties (Navaratnam et al. 1997; Rosenblatt et al. 1997; Jones et al. 1998; Ramanathan et al. 1999). Moreover, the coexpression of the different α-subunit splice variants with the modulatory β-subunit, predicted only to occur in the apical, low-frequency portion of the chick and turtle cochlea (Jones et al. 1999; Ramanathan et al. 1999, 2000), increased the Ca²⁺ sensitivity of the channel and slowed the current kinetics (Jones et al. 1999; Ramanathan et al. 2000).

Our electrophysiological findings are consistent with the expression of two different populations of Ca²⁺-activated K⁺ channels in IHCs that differ in their activation kinetics, sensitivity to IbTx and nature of Ca²⁺ dependence (Fig. 10). If I_K,f and I_K(Ca) both flow through BK channels the differences between the two channel populations might be related to splice variance of the α-subunit, the presence or absence of a β-subunit, or the association with different β-subunits. Mammalian IHCs show only a minimal internal splice variance of the α-subunit (Langer et al. 2003) but they express the modulatory β₁ and β₄ subunits and not β₂ and β₃, in agreement with our electrophysiological experiments indicating that neither I_K,f nor I_K(Ca) showed significant inactivation. While β₁ is expressed from around the onset of hearing, β₄ appears to be only present in immature mouse IHCs (Langer et al. 2003). Therefore, the most obvious hypothesis consistent with the literature is that the rapidly activating I_K,f is only composed of the α-subunit while the slower I_K(Ca) is carried by the coexpression of α and β₁. The β₁ subunit has been reported to confer a reduced sensitivity to IbTx (Dworetzky et al. 1996) when compared to the α-subunit alone.