The effects of Tmc1 Beethoven mutation on mechanotransducer channel function in cochlear hair cells

Our sense of hearing depends on sound stimuli evoking submicrometer displacements of bundles of stereocilia or hairs that project from the upper surface of the eponymous hair cells (Schwander et al., 2010; Fettiplace and Kim, 2014). Hair bundle motion is then transformed into an electrical signal by activation of mechanically gated ion channels located at the tips of the shorter stereocilia (Beurg et al., 2009). During bundle displacement, force is applied to the mechanotransduction apparatus by tension in tip links (Pickles et al., 1984; Hackney and Furness, 2013) extending from the top of each stereocilium to the side wall of its taller neighbor, which modulates the open probability of the mechano-electrical transducer (MT) channels. Because the channels detect at most a few hundred nanometers of motion at the tops of the stereocilia, transduction is endowed with multiple phases of adaptation to ensure that the MT channels are poised at their maximum sensitivity (Eatock, 2000; Fettiplace and Kim, 2014). The prevailing view has been that all phases of adaptation are regulated by a change in intracellular Ca²⁺ after influx of the ion via the Ca²⁺-selective MT channels. However, there has been recent dispute about this mechanism in mammalian hair cells and whether, as in nonmammals, Ca²⁺ entry is required (Peng et al., 2013; Corns et al., 2014). The two groups reached different conclusions about the effects of altering Ca²⁺ influx or changing the concentration or buffering of intracellular Ca²⁺. In one case (Peng et al., 2013), these manipulations had little effect on adaptation, whereas in the other (Corns et al., 2014), all modulated adaptation.

Although the molecular identity of the MT channel is still uncertain, recent studies have suggested that transmembrane channel–like protein isoform 1 (TMC1) and isoform 2 (TMC2) are possible pore-forming subunits of the channel (Kawashima et al., 2011; Kim and Fettiplace, 2013; Pan et al., 2013). The Beethoven (Tmc1^Bth) mutant is a mouse model for progressive hearing loss (DFNA36) caused by a point mutation producing a methionine to lysine substitution at position 412 of TMC1 (Vreugde et al., 2002), which may be located in the ion conduction pathway (Pan et al., 2013), and might therefore influence channel properties. We have characterized the MT channel in the Beethoven mutant, confirming that, as with the Tmc1 mutant, it has an altered selectivity for Ca²⁺ (Kim and Fettiplace, 2013; Pan et al., 2013), and we then exploited this feature to investigate the role of Ca²⁺ in MT channel adaptation. These results, taken together with those derived from other manipulations of Ca²⁺ influx and its cytoplasmic concentration, are consistent with a major component of adaptation being mediated by a change in intracellular Ca²⁺.

Beethoven (Tmc1^Bth/Bth) mice were obtained from K. Steel (King’s College, London, England, UK) and W. Marcotti (Sheffield University, Sheffield, England, UK). Tmc1^Bth/Bth encodes a point mutation at residue 412, predicted to cause a methionine to lysine substitution (Vreugde et al., 2002). Tmc2 mutant mice (B6.129S5-Tmc2^tm1Lex/Mmucd) were obtained from the Mutant Mouse Regional Resource Center (University of California, Davis, Davis, CA) and were used to generate double mutants Tmc1^Bth/Bth:Tmc2^−/−. Mice were genotyped from tail clips taken after dissection for the electrophysiology recordings or immunolabeling (Marcotti et al., 2006; Kim et al., 2013). In both cases, the experimenter was blind to the genotype. Mice, postnatal days (P) 3–9 were killed by decapitation, as specified by the Institutional Animal Care and Use Committees of the University of Wisconsin-Madison according to current National Institutes of Health guidelines. The cochlea was removed, and the organ of Corti was dissected in a solution composed of (mM): 150 NaCl, 6 KCl, 1.5 CaCl₂, 2 Na-pyruvate, 8 d-glucose, and 10 Na-HEPES, pH 7.4 (osmolarity of 315 mOsm). Either apical or basal coils of the excised organ of Corti were fixed in a microscope chamber under strands of dental floss and viewed through a long working-distance 63× water-immersion objective on a microscope (DMLFS; Leica). The effect of endolymph-like Ca²⁺ (0.04 mM) (Bosher and Warren, 1978; Ikeda et al., 1987) on transduction was studied by local application from a nearby puffer pipette containing (mM): 150 NaCl, 0.04 CaCl₂ (buffered with 4 HEDTA), 2 Na-pyruvate, 8 d-glucose, and 10 Na-HEPES, pH 7.4. For each extracellular solution, the fluid jet for bundle stimulation was filled with the same solution.

MT currents were recorded from outer hair cells (OHCs) and inner hair cells (IHCs) at room temperature (21–23°C) as described previously (Beurg et al., 2006, 2014). Patch electrodes were filled with an intracellular solution of composition (mM): 130 CsCl, 3 MgATP, 10 Tris phosphocreatine, 1 EGTA (1 BAPTA for the two-pulse adaptation experiment), 0.5 GTP, 0.5 cyclic AMP, and 10 Cs-HEPES, pH 7.2 (∼295 mOsm/l). They were connected to an amplifier (Axopatch 200B; Molecular Devices) with a 5-kHz output filter. Membrane potentials were corrected for a liquid junction potential of −4 mV. Single MT channel currents were isolated by brief treatment with extracellular saline containing 5 mM BAPTA to sever the majority of the tip links (Beurg et al., 2006; Kim et al., 2013).

MT currents were elicited as described previously (Kim and Fettiplace, 2013; Corns et al., 2014) using a fluid jet in which saline is ejected or withdrawn through the 10-µm tip of a glass pipette, with the flow generated by flexure in a piezoelectric disk. The flow, and hence the force applied to the hair bundle, is proportional to the driving voltage across the piezoelectric disk. The amplitude and time course of hair bundle motion during fluid jet stimulation were quantified by projecting an image of the OHC bundle onto a pair of photodiodes (LD 2–5; Centronics) at 240× (Crawford and Fettiplace, 1985; Kim and Fettiplace, 2013; Fig. S1). Because of the difficulty of the procedure, which requires optimal orientation of the preparation to produce a bright bundle, the bundle displacement was not calibrated in every experiment. The fluid jet was preferred to a rigid probe (Beurg et al., 2006; Peng et al., 2013) because it does not require any prior contact with the hair bundle, which could affect its resting position, and is also likely to evoke a more uniform stereociliary displacement (Corns et al., 2014); however, it suffers from the drawback that stimulus onsets are filtered, and therefore slower, and so adaptation kinetics could not be determined accurately. The relationship between MT current, I, and bundle displacement, ΔX, was fitted with a single Boltzmann equation: I/I_MAX = 1/(1 + exp(−(ΔX − X_0.5)/X_S)), where I_MAX = maximum current, X_0.5 = bundle displacement for half-maximum current, and X_S = slope factor expressing width of relation.

The calcium selectivity of the MT channel was determined as described previously (Beurg et al., 2006; Kim and Fettiplace, 2013) by measuring the Ca²⁺ reversal potential (V_rev) using a CsCl-based intracellular solution containing (mM) 135 CsCl, 3 MgATP, 10 Tris phosphocreatine, 1 EGTA-CsOH, and 10 HEPES, pH 7.2 (osmolarity of 293 mOsm/l), and a high Ca²⁺ extracellular solution containing (mM) 100 CaCl₂, 20 N-methylglucamine, 6 Tris, and 10 d-glucose, pH 7.4, which was applied as described above and also included in the fluid jet. Reversal potentials were corrected for the liquid junction potential of −9 mV. The relative permeability, P_Ca/P_Cs, was calculated from the Goldman–Hodgkin–Katz equation:

where θ = (V_revF/RT); RT/F has its usual meaning with a value at room temperature of 25.6 mV; [Cs⁺] is the intracellular Cs⁺ concentration (140 mM); [Ca²⁺]_o is the extracellular Ca²⁺ concentration (100 mM); [Ca²⁺]_i is the intracellular Ca²⁺ concentration (0, 1, or 2.5 mM); and a1, a2, and a3 are the activity coefficients for Cs⁺ (Partanen, 2010) and Ca²⁺ (Rard and Clegg, 1997) with values of a1 = 0.73 for intracellular Cs⁺, a2 = 0.52 for 100 mM extracellular Ca²⁺, and a3 = 0.89 or 0.84 for 1 or 2.5 mM of intracellular Ca²⁺, respectively. For those internal solutions containing elevated Ca²⁺, MgATP was reduced to 0.1 mM to minimize Ca²⁺ buffering by ATP.

Immunofluorescence labeling for PMCA2, the plasma membrane CaATPase pump, was performed as described previously (Chen et al., 2012), using an affinity-purified rabbit polyclonal antibody (NR2; Thermo Fisher Scientific). P5–P10 mice were decapitated, and cochleas were fixed in 4% paraformaldehyde in phosphate buffer for 40 min at room temperature. Isolated cochlear coils were treated with 0.5% Triton X-100, immersed for 1 h in 10% goat serum (Invitrogen) to block nonspecific labeling, and incubated overnight at 4°C with primary NR2 antibody. Specimens were labeled with Alexa Fluor 488 goat anti–rabbit IgG secondary antibody followed by Alexa Fluor 568 phalloidin (Invitrogen), which defines the hair bundle by its actin labeling. Mounted tissue was viewed under a 60× (NA = 1.4; ELWD Planfluor; Nikon) oil-immersion objective in a laser scanning confocal microscope (A1; Nikon). Four cochleas were characterized from Tmc1^+/+ and four from Tmc1^Bth/Bth mice. In each, the fluorescence intensities were quantified by averaging over areas of interest drawn around the hair bundle (Chen et al., 2012) in confocal images taken under identical conditions and photomultiplier settings using ImageJ (Fiji) software (National Institutes of Health). Fluorescence intensities were measured for 50 OHC bundles chosen at random in each preparation.

Unless otherwise stated, all values are quoted as means ± SD. Statistical comparisons of means were made by Student’s two-tailed t test.

Fig. S1 shows calibration of the fluid jet–evoked hair bundle motion during a two-pulse adaptation experiment by imaging the bundle on a dual photodiode.

MT currents from OHCs of wild-type (Tmc1^+/+) and littermate mutant (Tmc1^Bth/Bth) Beethoven mice were recorded during hair bundle displacements with a piezoelectric fluid jet (Kros et al., 1992; Kim and Fettiplace, 2013). The maximum amplitude of the MT current was similar at least up to postnatal day (P) 9 in all genotypes tested, indicating that, unlike dn mutation in Tmc1 (Kim and Fettiplace, 2013), the Beethoven mutation does not impair mechanotransduction in apical OHCs. Because early on the bundles have a normal shape, a corollary is that the single–MT channel conductance was unaffected by the Beethoven mutation (Fig. 1, A–E). Although, as observed previously, there was an increase in OHC MT channel conductance from apex to base (Beurg et al., 2014, 2015), the conductance at neither location was reduced relative to wild type in Tmc1^Bth/Bth. The basal values are most diagnostic, because they show a marked twofold reduction in the Tmc1^−/− (Beurg et al., 2014), but here there was no significant difference between the wild-type and Beethoven mutant. There was also no effect of the Beethoven mutation on the MT channel conductance in IHCs (Fig. 1 E).

Another manifestation of the properties of the MT channel pore is the ionic permeability, in particular its selectivity for Ca²⁺ over other cations. The Ca²⁺ selectivity of the MT channel in Beethoven was assayed by measuring the reversal potential of the transducer current under ionic conditions in which Ca²⁺ and Cs⁺ were the only permeant cations in the extracellular and intracellular solutions, respectively (Beurg et al., 2006; Kim and Fettiplace, 2013). This enabled the permeability ratio, P_Ca/P_Cs, to be calculated (see Materials and methods). The reversal potential and relative Ca²⁺ permeability in apical OHCs were both significantly reduced (P < 0.001) in Tmc1^Bth/Bth mice compared with Tmc1^+/+ (Fig. 1, F–H). In the first postnatal week, there are concerns about the presence of TMC2, which is down-regulated at about P6 (Kawashima et al., 2011), and the knockout of which reduces the channel Ca²⁺ selectivity (Kim and Fettiplace, 2013). Although the Ca²⁺ selectivity in wild-type OHC MT channels decreases slightly with age up to P10 (Kim and Fettiplace, 2013), to exclude any contribution from TMC2 in the present experiments, double mutant mice (Tmc1^Bth/Bth:Tmc2^−/−), in which TMC2 was also not present, were also examined and a similar reduction in Ca²⁺ selectivity was observed (Fig. 1 H). In both the presence and absence of TMC2, there was a half to two thirds reduction in the Ca²⁺ selectivity of the OHC MT channel in Tmc1^Bth/Bth relative to wild type; a comparable reduction in Ca²⁺ selectivity was reported for IHCs of Tmc1^Bth/− relative to wild type (Pan et al., 2013).

Through its action on transducer channel adaptation, Ca²⁺ modulates the fraction of the transducer current activated at the bundle’s resting position (i.e., the position in the absence of stimulation). Increasing Ca²⁺ influx through the MT channels decreases channel open probability and thus promotes adaptation, whereas reducing Ca²⁺ influx, by lowering its extracellular concentration or depolarizing the cell to near the Ca²⁺ equilibrium potential, increases the channel open probability (Corey and Hudspeth, 1983; Assad et al., 1989; Crawford et al., 1991; Ricci et al., 1998). The effect of Ca²⁺ on adaptation of the MT current in Beethoven mutants was tested by comparing experiments in which bundles were bathed in saline containing 1.5 mM Ca²⁺ as in perilymph, and low extracellular Ca²⁺, 0.04 mM, similar to endolymph to which they are exposed in vivo (Bosher and Warren, 1978; Ikeda et al., 1987). In OHCs from Tmc1^+/+ mice, reducing the extracellular Ca²⁺ concentration from 1.5 to 0.04 mM at −84 mV increased MT channel resting P_open (Fig. 2, A and B) from (mean ± SD) 0.02 ± 0.01 (n = 8) to 0.43 ± 0.02 (n = 5), as reported previously (Johnson et al., 2011). However, in OHCs of Tmc1^Bth/Bth mice, there was a smaller increase in the MT channel resting P_open (Fig. 2, C and D), from 0.03 ± 0.01 (n = 6; 1.5 mM Ca²⁺) to 0.19 ± 0.07 (n = 5; 0.04 mM Ca²⁺). The values for P_open in the low Ca²⁺ were significantly different between genotypes (P < 0.001).

Apart from regulating adaptation, Ca²⁺ is also known to block inward current through the MT channels from the external surface, with a half-blocking concentration of ∼1 mM (Ricci and Fettiplace, 1998; Kim et al., 2013); therefore, lowering extracellular Ca²⁺ to its concentration in endolymph increases the current by relieving Ca²⁺ block of the channel (Corey and Hudspeth, 1983; Ricci and Fettiplace, 1998; Fettiplace and Kim, 2014). In the present experiments, reducing extracellular Ca²⁺ concentration from 1.5 to 0.04 mM increased I_max, the peak MT current amplitude: I_max (0.04/1.5) = 1.44 ± 0.17 (n = 5) in Tmc1^+/+, and I_max (0.04/1.5) = 1.45 ± 0.05 (n = 5) in Tmc1^Bth/Bth mice, with no significant difference in I_max (0.04/1.5) values between genotypes. This indicates that the site of Ca²⁺ blockade within the channel is unaffected by the point mutation in Tmc1^Bth/Bth.

To confirm whether OHCs from Tmc1^Bth/Bth mice possess reduced adaptation, a paired-pulse protocol was used in which two series of brief mechanical test steps (4 ms) were delivered to the hair bundle, one before and the other during a nonsaturating adapting step (10 ms). This paired-pulse protocol was applied to OHCs held at −84 (Fig. 3 A) and +96 mV (Fig. 3 B). The displacement of the hair bundle evoked by a given fluid jet stimulus was in some experiments calibrated by imaging the hair bundle on a photodiode pair (Crawford and Fettiplace, 1985; Fig. S1). Because the procedure requires optimal orientation of the preparation to produce a bright bundle, the bundle displacement was not calibrated in every experiment. In those cells where it was, the fluid jet calibration, expressed as nanometers of OHC bundle displacement per volt of piezoelectric driving voltage, was 3.6 ± 0.4 nm/V (n = 16) in Tmc1^+/+ and 3.2 ± 0.5 nm/V (n = 10) in Tmc^Bth/Bth. The two means are not significantly different, implying that there is no major change in OHC bundle stiffness in the Beethoven mutant. The calibrations were used to derive the relationship between MT current and displacement. After the adapting step at a −84-mV holding potential, the relationship was shifted to the right along the displacement axis compared with that preceding the adapting step (Fig. 3 C), but at +96 mV, the adaptive shift was absent (Fig. 3 D); retesting after the holding potential was returned to −84 mV showed that the adaptive shift returned (Fig. 3 E). The existence of an adaptive shift, present at −84 mV and absent at +96 mV, was seen in all cells studied whether the bundle displacement was measured directly in the experiment (Fig. S1) or was inferred from piezoelectric driving voltage, V_piezo, using calibration values given above (Fig. 3, C–E). Furthermore, the same voltage-dependent behavior of adaptation was observed in the OHCs of Tmc1^Bth/Bth as well as of Tmc1^+/+ mice. The adaptive shift in the current–displacement relationship is usually thought to result from a change in intracellular Ca²⁺ after influx of the ion through the MT channels; such influx will be small or absent on depolarizing to large positive potentials (here, +96 mV) that approach the Ca²⁺ equilibrium potential (Assad et al., 1989; Crawford et al., 1991).

To examine the efficacy of adaptation, the same two-pulse protocol was used, but the size of the adapting step was varied (Fig. 3, F–H). For each size of adapting step (A), the shift in the current–displacement relationship (ΔX) was determined and plotted against A (Fig. 3 H). The linear fit to these plots in five OHCs gave a slope that was significantly (P < 0.02) smaller in Tmc1^Bth/Bth (0.33 ± 0.01; n = 5) than in Tmc1^+/+ mice (0.52 ± 0.02, n = 5; mean ± SD). It might be argued that such a difference reflects different hair bundle mechanics between wild type and mutant, but this is unlikely because the slope (ΔX/A) is dimensionless (ΔX and A are measured in the same units), and furthermore, there was no difference in fluid jet sensitivity between OHC hair bundles of Tmc1^+/+ and Tmc1^Bth/Bth (see above). The results demonstrate that OHCs in the Beethoven mutant still retain adaptation but at a reduced level.

The results so far are most easily explained by postulating that impairment of MT adaptation in Beethoven is attributable to a smaller Ca²⁺ influx through MT channels with reduced Ca²⁺ permeability. Doubt has previously been cast on the role of cytoplasmic Ca²⁺ in regulating adaptation because, among other observations, adaptation was unaffected by raising intracellular Ca²⁺ to a level that might be expected to exceed that normally present during adaptation (Peng et al., 2013). We confirmed this result, obtaining a slope factor (ΔX/A) of 0.52 ± 0.03 (n = 11), but then found that the adaptation observed with a patch-pipette solution containing 2.5 mM Ca²⁺ was in all OHCs significantly reduced or abolished (P < 0.001) on depolarizing to +96 mV (Fig. 4): at +96 mV, the slope factor (ΔX/A) was 0.17 ± 0.10 (n = 11; range of 0.002 – 0.310). This result suggests that a major component of transducer adaptation is mediated by Ca²⁺ entry, but it does not rule out that an additional process independent of Ca²⁺ influx can make some contribution (Peng et al., 2013).

A possible explanation for the lack of effect of high intracellular Ca²⁺ introduced via the patch pipette solution is that there is significant accumulation of Ca²⁺ at the inner face of the channel when it is open, but Ca²⁺ influx is still much reduced on depolarizing toward the Ca²⁺ equilibrium potential. Such a hypothesis is consistent with two other observations in which the two-pulse adaptive shifts in the current displacement relation were documented under different combinations of intracellular and extracellular Ca²⁺ (Fig. 5). First, if the hair bundle was bathed in saline containing an endolymph-like, 0.04-mM Ca²⁺, an adaptive shift in the current–displacement relation still occurred when recording with an intracellular solution buffered with 1 mM EGTA (Fig. 5, C and D). In low extracellular Ca²⁺, the efficacy of adaptation, expressed as ΔX/A, was 0.49 ± 0.11 (mean ± SD; n = 5). However, the shift was much reduced on recording with 2.5 mM of intracellular Ca²⁺ (Fig. 5, E and F), with a ΔX/A value of 0.11 ± 0.14 (n = 7), not significantly different from 0 (P = 0.08). Evidence that the high, 2.5 mM of intracellular Ca²⁺ reached the channel was that it produced partial block of the MT current: the mean amplitude of the current was 1.44 ± 0.15 nA when recording with submicromolar intracellular Ca²⁺ buffered with 1 mM EGTA, and was reduced to 0.61 ± 0.21 nA with 2.5 mM of intracellular Ca²⁺.

If, during recording with 2.5 mM of intracellular Ca²⁺, a comparably high concentration is achieved at the internal face of the MT channel, it should theoretically reduce the reversal potential of the MT current determined as described above (see Materials and methods). The reversal potential was measured while bathing the hair bundle in 100 mM of extracellular Ca²⁺, and was not significantly altered when either 1 or 2.5 mM Ca²⁺ was in the recording pipette solution (Fig. 5 H). The expected reduction in reversal potential, V_rev, was estimated from the constant field equation, Eq. 1 (see Materials and methods), assuming a constant permeability ratio P_Ca/P_Cs; for 2.5 mM of intracellular Ca²⁺, the predicted decrease in reversal potential compared with zero intracellular Ca²⁺ was roughly 10 mV.

An explanation for this discrepancy is that when the preparation is immersed in 1.5 mM Ca²⁺-perilymph, there is stereociliary accumulation of the divalent cation before and during performing the measurement with 100 mM. To address this possibility, before determining the reversal potential, the hair bundle was bathed in 0.04 mM Ca²⁺ for 5 or more minutes to lower bundle Ca²⁺. With this added manipulation, a significant change in reversal potential was seen experimentally, similar to that predicted by Eq. 1 (Fig. 5, G and H). An important conclusion is that the stereociliary Ca²⁺ is substantially elevated when the bundles are bathed in perilymph. A corollary is that the reversal potential, and hence the permeability ratio P_Ca/P_Cs, is underestimated in such conditions. Even when the internal solution was buffered with 1 mM EGTA or BAPTA, the reversal potential was increased to 30.0 ± 1.5 mV (n = 5) after exposure to 0.04 mM Ca²⁺ saline (Fig. 5 H). Using this value for reversal potential and assuming zero intracellular Ca²⁺, a permeability ratio P_Ca/P_Cs of 6.7 ± 0.2 was calculated. With this permeability ratio, and a reversal potential of 24.6 mV obtained with exposure to 1.5 mM Ca²⁺ saline (Fig. 5 H), the intracellular Ca²⁺ is estimated to be in at least 3 mM. The equivalent measurement performed on the Beethoven mutant after prior exposure to 0.04 mM Ca²⁺ saline gave a reversal potential ∼2 mV more positive than that obtained under conditions where the preparation was first exposed to normal 1.5 mM Ca²⁺ saline. The reversal potential after low Ca²⁺ exposure was 11.2 ± 1.9 mV (n = 7), from which a P_Ca/P_Cs of 1.9 ± 0.2 was calculated, compared with the P_Ca/P_Cs value of 1.5 ± 0.2 in 1.5 mM Ca²⁺ saline, with both measurements being made on a Tmc2^−/− background.

Ca²⁺ ions entering OHC stereocilia via the MT channels bind to calcium-buffering proteins, such as oncomodulin (Hackney et al., 2005), and are extruded via the PMCA2a isoform of the plasma membrane CaATPase pump (Dumont et al., 2001; Beurg et al., 2010), which is present at high density on the OHC stereociliary membrane. As reported previously (Chen et al., 2012), PMCA2 labeling is greater in bundles of OHCs than in IHCs; however, there are additional pumps in the IHC lateral membrane, probably conferred by the PMCA1 isoform, to cope with Ca²⁺ influx via the voltage-dependent Ca²⁺ channels in IHCs (Dumont et al., 2001; Chen et al., 2012). Changes in OHC expression of the PMCA2 with developmental age and cochlear location closely follow the functional appearance of the MT current (Chen et al., 2012). Because the Beethoven mutation causes a reduction in the Ca²⁺ selectivity of the MT channel, we hypothesized that the pump density would be affected in the mutant. Immunofluorescent labeling for PMCA2 confirmed it was strongly expressed in hair bundles of OHCs (Fig. 6, A and B). In Tmc1^Bth/Bth, PMCA2 expression was reduced in both apical and basal OHCs (Fig. 6 C), with the mean reduction being 40% at the apex and 58% at the base, and the differences with respect to the control both being significant (P < 0.001).

The molecular identity of the MT channel is still unclear, but recent work has identified two molecules (LHFPL5 and TMIE) implicated in mechanotransduction (Xiong et al., 2012; Zhao et al., 2014) that interact with protocadherin-15 at the lower end of the tip link (Kazmierczak et al., 2007). However, each of these deafness-associated membrane proteins is small and unlikely to be a pore-forming subunit of the channel (Xiong et al., 2012; Zhao et al., 2014; Beurg et al., 2015). In contrast, TMC1 and TMC2 are larger and could be pore-forming subunits of the MT channel (Kawashima et al., 2011; Pan et al., 2013). Here, we have examined the effects of a point mutation in TMC1, Beethoven, which is potentially in the ion conduction pathway (Pan et al., 2013) and so might influence MT channel properties. The primary outcome of the Beethoven mutation on MT channel function was an approximate halving in P_Ca/P_Cs, the relative Ca²⁺ selectivity of the channel (Fig. 1 H). However, we found no effect on the single MT channel conductance in either apical or basal OHCs or in IHCs, or on the block of the channel by extracellular Ca²⁺. Both findings disagree with an earlier report focused on IHCs (Pan et al., 2013), which claimed a change in the single-channel conductance in a Beethoven mutant. The lack of effect of the Beethoven point mutation on either MT channel conductance or on blockade by external Ca²⁺ weakens the conclusion that the M412K point mutation resides in the pore region of the MT channel, a line of argument used to maintain that TMC1 is a component of the pore (Pan et al., 2013). This type of behavior contrasts with localization of the pore in the mechano-sensitive channel protein Piezo by a single glutamate substitution (E2133A in piezo1 and E2416 in piezo2) to halve channel conductance (Coste et al., 2015). The Tmc1^Bth/Bth mutation resembles Tmc2^−/− in causing a change in the Ca²⁺ selectivity with no alteration in unitary conductance (Kim and Fettiplace, 2013; Kim et al., 2013). Our results do not exclude TMC1 as a pore component but merely suggest that the Beethoven mutation is unlikely to reside in the pore region.

A secondary consequence of the Beethoven mutation was an ∼40% reduction in the sensitivity of OHC adaptation (ΔX/A; Fig. 3 H). This consequence would be consistent with a halving in the Ca²⁺ influx through the channel, provided the Ca²⁺-binding site for regulating adaptation was on the cytoplasmic side of the channel. In addition, there were smaller shifts in the current–displacement relationship for reduced extracellular Ca²⁺ (Fig. 2) in the mutant. In those experiments, the resting position of the current–displacement relationship, reflected in the control resting P_open of the channel, was normal, but either manipulation drew a reduced response. This implies that, besides the component of adaptation dependent on Ca²⁺ influx, some other adaptive mechanism must exist (Eatock, 2000; Peng et al., 2013) and be resistant to the Beethoven mutation or to elevating cytoplasmic Ca²⁺. Such a mechanism might be mediated by tension in the tip link, adjusting its upper attachment point or the force it delivers to the MT channel. Nevertheless, because the adaptive shift revealed in paired-pulse experiments was largely eliminated on depolarization to near the Ca²⁺ equilibrium potential (Fig. 3), one component of MT channel adaptation is likely to be regulated by a change in intracellular Ca²⁺.

Three lines of evidence attest to the role of intracellular Ca²⁺ in regulating adaptation. The suppressive effects of depolarization and increased Ca²⁺ buffering have been contested (Peng et al., 2013; Corns et al., 2014), but both are indirect. The finding that adaptation is preserved when recording with a pipette solution containing 1.5 mM Ca²⁺ (Peng et al., 2013) seems definitive in arguing against a role for intracellular Ca²⁺. We reproduced this observation with 2.5 mM of internal Ca²⁺ but then discovered that the adaptation was still suppressed by depolarization, and furthermore, if measurements were made when bathing the hair bundle in saline containing endolymph-like Ca²⁺ (40 µM), then adaptation was now abolished with the same 2.5 mM of internal Ca²⁺ (Fig. 5). The simplest explanation for this result is that there is accumulation of Ca²⁺ at the cytoplasmic face of the MT channel, reaching a concentration of several millimolar in external saline (1.5 mM Ca²⁺). It should be noted that the internal concentration of Ca²⁺ can be higher than the external because of the additional effect on the driving force of a −80-mV holding potential that can theoretically produce a 500-fold concentration of the divalent ion.

The origin of the progressive hearing loss phenotype in the Beethoven mutant mice (Vreugde et al., 2002; Marcotti et al., 2006) is unknown. One possibility from the results here is that the decreased Ca²⁺ permeability of the OHC MT channel is the main causative factor. Related to the reduced Ca²⁺ influx through the MT channels in Beethoven mutants, there was a smaller shift in the MT current–displacement relation in endolymphatic Ca²⁺ concentration, causing a smaller fraction of the current to be activated at the resting position of bundles bathed in low Ca²⁺ endolymph. The large resting open probability generates an inward current that normally depolarizes OHCs (Johnson et al., 2011), and as a consequence, its reduction in Beethoven mutants will hyperpolarize these cells. This hyperpolarization will shift the membrane potential negative to that optimal for prestin activation (about −50 mV), thereby decreasing cochlear sensitivity attributable to OHC amplification (Ashmore, 2008; Johnson et al., 2011). This does not account for the abnormality and depletion of IHCs, which precedes that of OHCs (Vreugde et al., 2002). However, the overall consequence is likely to be an increase in the compound action potential threshold despite the presence of functional MT channels.

The Beethoven mutation also results in a reduction in the stereociliary density of the PMCA2 pump responsible for extruding Ca²⁺ ions that have entered via the MT channels. The reduced expression of the pump may reflect homeostatic regulation in the face of a decreased Ca²⁺ influx through MT channels with half their normal permeability to that ion. Because several types of deafness have been attributed to defects in PMCA2 expression or pumping capacity in OHCs (Street et al., 1998; Bortolozzi et al., 2010), it is conceivable that this is an additional factor in the progressive deafness in Beethoven. However, if the decreased expression of PMCA2 results from a smaller influx and cytoplasmic concentration of Ca²⁺, then it is possible that other genes are down-regulated in Beethoven, and these contribute to the deafness phenotype.

We would like to thank Laura Corns and Walter Marcotti of Sheffield University, UK, for helpful discussions and for sharing unpublished results.

This work was supported by a grant from the National Institutes of Deafness and other Communication Disorders (RO1 DC01362 to R. Fettiplace).

The authors declare no competing financial interests.

Richard W. Aldrich served as editor.