Developmental changes in the cochlear hair cell mechanotransducer channel and their regulation by transmembrane channel–like proteins

Hair cells, the sensory receptors of the cochlea, detect sounds through deflection of their stereociliary bundles, which activates mechanotransducer (MT) ion channels. MT channels, calcium-permeable cation channels (Corey and Hudspeth, 1979; Ohmori, 1985; Beurg et al., 2006) with large (>100-pS) unitary conductance (Crawford et al., 1991; Géléoc et al., 1997; Ricci et al., 2003; Beurg et al., 2006), are located at the tops of the stereocilia (Beurg et al., 2009). These channels are activated by tension in the tip links (Pickles et al., 1984; Furness and Hackney, 1985), fine extracellular strands that connect the top of each stereocilium with the side wall of its taller neighbor. Although analyses of mutations associated with deafness in humans and mice have enabled the elucidation of some of the molecular apparatus underlying transduction, including the composition and anchoring of the tip links (Richardson et al., 2011; Kazmierczak and Müller, 2012), the identity of the MT channel remains controversial. However, recent work has indicated that two isoforms of the transmembrane channel–like family, Tmc1 and Tmc2, play a central role in hair cell mechanotransduction (Kawashima et al., 2011). The “dn” and “Beethoven” mutations in the Tmc1 gene are associated with hearing loss (Steel and Bock, 1980; Kurima et al., 2002; Vreugde et al., 2002), but it has been unclear whether deafness stems from lack of MT channel function or is secondary to a developmental defect (Marcotti et al., 2006). Double knockouts of Tmc1 and Tmc2 abolish MT currents in both auditory and vestibular epithelia, a phenotype that can be rescued by transfection with either Tmc1 or Tmc2 mRNA (Kawashima et al., 2011). Although these results argue for a central role for Tmc1 and Tmc2 in hair cell mechanotransduction, it remains uncertain whether the MT current vanishes in Tmc1−/− (Marcotti et al., 2006; Kawashima et al., 2011).

The mammalian cochlea separates acoustic frequencies along its length so that auditory nerve fibers emanating from different regions are tuned to disparate frequencies, responding preferentially to high pitched tones at the base and low pitched ones at the apex (Fettiplace and Hackney, 2006). This systematic change in best frequency along the length of the cochlea, referred to as the tonotopic organization, stems from gradients in numerous properties including the dimensions of the cochlear partition and the ion channel complement of the outer hair cells (OHCs). Among these variables is the single-channel conductance of the MT channel, which, in mammals, increases from apex to base in OHCs but remains constant in inner hair cells (IHCs) (Beurg et al., 2006). This suggests that multiple isoforms of the MT channel protein may be differentially distributed along the mammalian cochlea. Here, we characterized MT channel properties in both wild-type mice and Tmc mutants and found that, in the first postnatal week, there is also a tonotopic gradient in the Ca²⁺ selectivity of the OHC MT channel that can be modified by mutations in Tmc1 or Tmc2. Our results suggest that the Tmc proteins specify the subunit composition of the MT channel.

MT currents were recorded in OHCs and IHCs in isolated organs of Corti of mice between 0 and 10 d postnatal (P0–P10, where P0 is the birth date) using methods described previously (Beurg et al., 2006). For initial characterizations in wild-type animals, both outbred CD-1 and inbred CBA/J mice were used, with the majority of measurements obtained on P4–P7 mice at the apex and P2–P5 mice at the base, ages at which the MT current was maximal at each position (Fig. 1, C and D). Mutation in Tmc1 was achieved with dn mice (CBA.Cg-Tmc1^dn/AjgJ; The Jackson Laboratory), which contain a deletion of exon 14 and are on a CBA/J background. The Tmc2 mutants (B6;129S5-Tmc2^tm1Lex/Mmucd), in which exon 1 is deleted, were obtained from the Mutant Mouse Regional Resource Center (University of California, Davis, Davis, CA). These were supplied as heterozygotes and bred with CBA/J mice through three generations to put them on the same background as the Tmc1 mutants, although the limited number of generations may be insufficient to ensure that both Tmc1 and Tmc2 had identical backgrounds. Both male and female mice of all strains were used. Mice were genotyped from tail clips using procedures and primer sets recommended by the suppliers.

Mice were killed by decapitation using methods approved by the Institutional Animal Care and Use Committee of the University of Wisconsin-Madison according to current National Institutes of Health guidelines. Excised apical, middle, or basal cochlear turns were fixed under ties of dental floss in a recording chamber mounted on a microscope (Axioskop FS; Carl Zeiss) and viewed through a 40× water-immersion objective (Carl Zeiss). The chamber was perfused with artificial perilymph of the following composition (mM): 150 NaCl, 6 KCl, 1.5 CaCl₂, 2 Na-pyruvate, 8 d-glucose, and 10 Na-HEPES, pH 7.4. The effect of endolymphatic Ca²⁺ was examined by changing the solution around the hair bundle using a nearby puffer pipette to one containing (mM): 150 KCl, 0.02 CaCl₂ (buffered with 4 HEDTA), 2 Na-pyruvate, 8 d-glucose, and 10 HEPES, pH 7.4. The puffer pipette was positioned ∼30 µm from the target and directed along the cochlear axis so that the flow did not directly stimulate the bundle.

Recordings were made from first or second row OHCs or from IHCs using borosilicate patch electrodes connected to an amplifier (Axopatch 200A; Axon Instruments), and currents were low-pass filtered at the amplifier output at 10 kHz. Patch electrodes were filled with a solution containing (mM): 135 CsCl, 3 MgATP, 10 Tris phosphocreatine, 1 EGTA, and 10 Cs-HEPES, pH 7.2; in some experiments, 3 mM BAPTA rather than EGTA was used as the intracellular Ca²⁺ buffer. Membrane potentials were corrected for a liquid junction potential and for the voltage drop across the uncompensated series resistance. Most voltage-clamp protocols are referred to a holding potential of −84 mV. Values are given as mean ± 1 SEM, and P < 0.05 indicates statistically significant difference on a two-tailed Student’s t test. All experiments were performed at room temperature of 21–25°C.

Reversal potentials (V_REV) were determined by using an extracellular solution containing (mM): 100 CaCl₂, 20 N-methylglucamine, 6 Tris, and 10 glucose, adjusted with HCl to pH 7.4, in both the fluid jet and the puffer pipette. Reversal potentials were corrected for a liquid junction potential of 9 mV. The relative Ca²⁺ permeability, P_Ca/P_Cs, was calculated from the Goldman–Hodgkin–Katz equation:

where RT/F has its usual meaning with a value at room temperature of 25.7 mV, [Cs⁺] and [Ca²⁺] are the concentrations of Cs⁺ intracellularly (140 mM) and Ca²⁺ extracellularly (100 mM), and a₁ and a₂ are the published activity coefficients for Cs⁺ (Partanen, 2010) and Ca²⁺ (Rard and Clegg, 1997), respectively. The analysis assumes that these are the only ions to which the MT channel is permeable and that the contributions of N-methylglucamine and Tris are negligible. Hair bundles were mechanically stimulated with a fluid jet from a pipette, with a tip diameter of 10–15 µm, driven by a 25-mm diameter piezoelectric disc (Johnson et al., 2011). The distance of the pipette tip from the bundle was adjusted to elicit a maximal MT current. When testing the effects of different Ca²⁺ concentrations, the fluid jet pipette was filled with a solution containing the relevant Ca²⁺, and the perfusate around the hair bundle was also exchanged for the same Ca²⁺ solution in the puffer. During fluid jet stimulation, bundle motion was monitored by projecting an image of the bundle onto a pair of photodiodes (LD 2–5; Centronics) at a total magnification of 340 (Johnson et al., 2011).

We initially characterized the properties of the wild-type MT channel, seeking features that might reveal tonotopic differences in the ion conduction pathway or “pore” of the channel. MT currents were recorded from hair cells in isolated turns of neonatal (P2–P10) mouse cochleas in response to deflections of their stereociliary bundle, implemented with a fluid jet (Kros et al., 1992; Johnson et al., 2011) to minimize damage to the short bundles (Fig. 1 A). Measurements were obtained from organ of Corti segments at fractional distances (the longitudinal distance from the apex divided by the total length of the cochlea) of 0.2, 0.6, and 0.9 (Fig. 1 B), which will be referred to as apex, middle, and base, respectively. MT currents in OHCs increased in peak amplitude from apex to base, whereas IHC MT currents varied little across the three locations (Fig. 1 B), similar to results found in other rodents (Beurg et al., 2006; Jia et al., 2007). For the IHC currents there was a small decrease from apex to base, which might be partly accounted for by greater metabolic vulnerability of the basal hair cells and damage to or block of the MT channels, for example by Ca²⁺ loading. If this also occurred in OHCs, then the apex to base gradient in their current will be even larger. The tonotopic variation in the MT current was conserved if the OHC hair bundles were perfused with endolymph (K⁺; 0.02 mM Ca²⁺) to which they are normally exposed in vivo rather than the perilymph (Na⁺; 1.5 mM Ca²⁺), but at all locations the current amplitude was larger (Fig. 1 B), with the increase averaged over the three locations being 1.66 ± 0.06 (mean ± SEM). Tonotopic variation in the size of the macroscopic MT current and conductance in OHCs may be attributable to a combined increase in the single-channel conductance and the number of stereocilia (and hence channels) per bundle from apex to base (Beurg et al., 2006).

Variation in the pore region of the MT channel may be manifested as a difference in its ionic selectivity, especially that for Ca²⁺, which is the most permeant ion. The Ca²⁺ permeability of the channel was determined from measurements of reversal potentials under conditions where Ca²⁺ was the sole permeant ion extracellularly and Cs⁺ was the intracellular cation. The experimental protocol constituted a saturating hair bundle stimulus superimposed on a voltage ramp (Fig. 2 A). Changes in the amplitude of the MT current in the voltage region around the reversal potential were approximately linear with membrane potential (Fig. 2 B), enabling accurate interpolation of the reversal potential. The Ca²⁺ permeability, P_Ca/P_Cs, was calculated from the reversal potential, V_REV, using the Goldman–Hodgkin–Katz equation (see Materials and methods). For the OHCs, the reversal potential and relative Ca²⁺ permeability showed small but significant decreases from apex to base, whereas the values were invariant with position for the IHCs (Fig. 2, C and D). Thus, for OHCs, the mean P_Ca/P_Cs (± SEM) = 6.1 ± 0.7 (n = 7) at the apex, 5.2 ± 0.1 (n = 4) at the middle, and 4.6 ± 0.5 (n = 12) at the base. The three means were significantly different from each other (t test; P < 0.02). A concern is that the smaller permeability at the base may be caused by Ca²⁺ loading, but the inferred permeability for basal OHCs was unaffected when 3 mM BAPTA was used instead of 1 mM EGTA as the Ca²⁺ buffer in the intracellular solution (P_Ca/P_Cs = 4.2 ± 0.6; n = 9). The results were also unaffected by reversing the polarity of the voltage ramp, starting at a depolarized potential where the initial Ca²⁺ flux would be outward. The results demonstrate a tonotopic gradient in OHC channel properties from apex to base, but IHCs lack such a gradient either in Ca²⁺ permeability or MT current size.

The measurements reported so far were made on animals in the first neonatal week. The MT current grows in amplitude over the first few neonatal days, with development at the apex lagging the base by about 2–3 d (Waguespack et al., 2007; Lelli et al., 2009). In our measurements on wild-type mice, the OHC current became maximal by P2 at the base and P4 at the apex (Fig. 1, C and D). MT currents at the two locations were recorded at a postnatal age when the current had become maximal, but at this stage, cochlear hair cells are still relatively immature. For example, OHCs lack the somatic motor prestin (Belyantseva et al., 2000; Abe et al., 2007) as well as the KCNQ4 voltage-dependent K⁺ conductance (Marcotti and Kros, 1999). These membrane proteins appear during the second postnatal week before the onset of hearing at about P12. During the second and third weeks, there are no further changes in the MT current amplitude (Fig. 1, C and D; Kennedy et al., 2003), but we did find changes in the OHC MT channel Ca²⁺ permeability: the apical permeability decreased with development from (mean ± SEM) 6.1 ± 0.7 (n = 7; P5–P6) to 4.6 ± 0.1 (n = 6; P8–P10), now making it similar to that at the base, which was 4.2 ± 0.2 (n = 4; P8). In the older animals, the Ca²⁺ permeabilities in apical and basal OHCs were no longer significantly different (t test; P = 0.38), implying that the gradient in OHC MT channel Ca²⁺ permeability is a developmental feature. In contrast to the OHCs, no such developmental change was seen in IHCs. The Ca²⁺ permeability in apical IHCs was 5.7 ± 0.4 (n = 13; P4–P7) and 5.8 ± 0.4 (n = 5; P8–P9). Thus, in the P8–P10 mice, after the developmental switch, the Ca²⁺ permeability of the MT channel in IHCs is larger than that in OHCs. This permeability difference between these two cell types agrees with previous results found in rat cochleas at a similar developmental age (Beurg et al., 2006).

We next investigated the effects on the tonotopic organization of Tmc1−/− and Tmc2−/− mice, the former (dn) being a recessive mutation that lacks exon 14 in the Tmc1 gene, and the latter being a recessive mutation with a deletion of exon 1 of the Tmc2 gene. In Tmc1−/−, there was little or no effect on the OHC MT current amplitude up until about P6 (Fig. 3 A), but beyond this time point, the current declined to zero by P10, but it was maintained in the heterozygotes (Fig. 3 B). For IHCs, significant though reduced MT currents could still be recorded in Tmc1−/− at P9: the mean current was 0.35 ± 0.02 nA (n = 3) in the knockout compared with 0.66 ± 0.07 nA (n = 3) in the heterozygotes. Although loss of transduction in OHCs may account for the deafness phenotype in dn mice, it disagrees with the observation of Marcotti et al. (2006), who were still able to record large OHC MT currents in dn mice at P8. The reason for this discrepancy is unclear, but it might result from differences in the genetic background: the mice used by Marcotti et al. (2006) were from the original dn stock and were maintained on an unknown background. However, in accord with our results, Kawashima et al. (2011) found an ∼40% reduction in the apical OHC current amplitude in P5–P7 Tmc1−/− mice.

The properties of the MT current before P6 in the Tmc1−/− mice were similar to those documented for the wild type or heterozygotes (Fig. 3, C and D). The current amplitudes at −84 mV were similar to wild type in perilymph (mean ± SEM = 0.86 ± 0.12 nA, n = 10, apex; 1.31 ± 0.09 nA, n = 5, base; Ca²⁺ = 1.5 mM), and also the current increased up to twofold upon exposing the bundle to endolymph (mean ± SEM = 1.68 ± 0.16 nA, n = 5, apex; 2.09 ± 0.4 nA, n = 4, base; Ca²⁺ = 0.02 mM). The tonotopic gradient in current amplitude was not significantly different between heterozygotes and Tmc1−/− mice over the time period before decline in the current in the knockout.

To explore further any differences in the Tmc1 knockouts, we measured the Ca²⁺ permeability of the MT channel and found that the gradient in OHC channel permeability was eliminated in the Tmc1−/− mice but not in the heterozygotes (Fig. 4, A–C). Abolition of the gradient occurred because the permeability at the base increased to become comparable to that at apex. Mean values (± SEM) for the Ca²⁺ permeability in the Tmc1−/− mice were 5.9 ± 0.2 (n = 3) at the apex and 6.3 ± 0.2 (n = 4) at the base. The Ca²⁺ permeability of the channel was also determined in Tmc2−/− mice. Robust MT currents could still be recorded in the neonatal Tmc2−/− mice. In perilymph at −84 mV, the apical OHC current was as follows: 0.35 ± 0.03 nA (n = 4) in Tmc2−/− compared with 1.02 ± 0.06 nA (n = 9) in Tmc2+/− at P4; 0.97 ± 0.05 nA (n = 4) in Tmc2−/− compared with 0.8 ± 0.07 nA (n = 3) in Tmc2+/− at P5; and 0.74 ± 0.10 nA (n = 3) in Tmc2−/− compared with 0.95 ± 0.05 nA (n = 2) in Tmc2+/− at P8. However, in Tmc2−/− mice, the MT channel Ca²⁺ permeability decreased at the apex and also to some extent at the base compared with the wild type or the heterozygotes (Fig. 4, D–F). Mean values (± SEM) for the Ca²⁺ permeability in the Tmc2−/− mice were 4.4 ± 0.1 (n = 5) at the apex and 3.7 ± 0.1 (n = 6) at the base. Thus, loss of Tmc2 reduces the Ca²⁺ permeability of the channel in both apical and basal OHCs. Knockout of Tmc2 had a similar effect on the apical IHCs both before (P4–P7) and after (P8–P9) the first postnatal week (Fig. 5). For example, the MT channel Ca²⁺ permeability in wild type was 5.7 ± 0.4 (n = 13) and in Tmc2−/− was 4.5 ± 0.2 (n = 6), which are significantly different (t test; P < 0.02) for P4–P7. A similar reduction in permeability with the knockout was seen in P8–P9 animals, indicating that Tmc2 is still present in the IHCs at that time. No such change was seen with the Tmc1−/− for which the IHCs retained their higher permeability (mean = 6.6 ± 0.6; n = 3).

We have characterized the currents flowing through the MT channels and their relative Ca²⁺ permeability in OHCs and IHCs along the mouse cochlea. Differences in Ca²⁺ permeability are likely to be the most robust indicator of a tonotopic variation in the ion conduction pathway or “pore” of the channel. Although small, the differences we observed under various conditions were reproducible and permit several conclusions. First, the Ca²⁺ permeability measurements show a gradation from apex to base in OHC MT channels, whereas this gradient is absent in IHCs, with similar permeabilities at the apex and base. Second, the gradient exists only during the first neonatal week, and thereafter (up to at least P10) the Ca²⁺ permeability decreases in apical OHCs to become the same as basal OHCs, whereas the IHCs are unchanged. Third, the Ca²⁺ permeability is susceptible to mutations in Tmc1 and Tmc2. Variations in the channel’s Ca²⁺ permeability were not accompanied by any striking changes in the MT current amplitude under any of the conditions studied. For example, once the MT currents became maximal at about P2 and P4 at the base and apex, respectively, they remained constant despite changes in Ca²⁺ permeability after the first neonatal week. The MT current amplitudes were similar to those reported in other rodents, with a tonotopic gradient for OHCs but no such gradient in IHCs (He et al., 2004; Beurg et al., 2006; Jia et al., 2007; Stauffer and Holt, 2007).

A clue to the variability arises from the results with the Tmc1 and Tmc2 mutants. Knockout of Tmc1 increased the Ca²⁺ permeability in basal OHCs, whereas knockout of Tmc2 reduced the permeability in apical (and to some extent also basal) OHCs and in IHCs. It is conceivable that the Tmc proteins are chaperones or trafficking proteins that specify channel composition. Indeed, Tmc6 and Tmc8 (also referred to as EVER1 and EVER2) are cytoplasmic proteins that interact with a zinc transporter to regulate Zn²⁺ levels in keratinocyte nucleoli (Lazarczyk et al., 2008). However, the results would also be consistent with the notion that Tmc1 and Tmc2 are in fact channel subunits able to contribute to the pore region, with Tmc1 being more important at the base and Tmc2 endowing the higher Ca²⁺ permeability at the apex and in IHCs. In line with this conclusion, the reduction in permeability of apical OHCs later than P6 may be accounted for by loss of Tmc2, which is down-regulated in the cochlea at this time (Kawashima et al., 2011). It is also consistent with the loss of the OHC MT current in Tmc1−/−, which occurs at about the same time. Indeed, the time course of disappearance of the current (Fig. 3 B) perhaps more precisely reflects loss of Tmc2 protein, which may be slightly delayed with respect to the decline in the message after P5 (see Fig. 1 B of Kawashima et al., 2011), with the delay reflecting the time course of MT channel turnover.

The results raise an important question concerning the functional significance of a change in MT channel composition and Ca²⁺ permeation during development. One possible explanation is that up until P6, cochlear hair cells are nonspecialized and the properties measured reflect those of the generic hair cell. In this context, it is pertinent that Tmc2 is retained in the vestibular end organs in the adult (Kawashima et al., 2011), and Tmc2 knockout, with or without Tmc1, substantially reduces MT currents in utricular type II hair cells, which might be regarded as less specialized than cochlear OHCs. The time point around P7 coincides with refining the OHC phenotype, including acquisition of prestin and the onset of electromotility (Belyantseva et al., 2000; Abe et al., 2007), the appearance of new voltage-dependent K⁺ currents (I_K,n; Marcotti and Kros, 1999), and changes in the pattern of innervation. Interestingly, failure to develop the adult hair cell voltage-dependent K⁺ channels was a conspicuous feature of the Tmc1 dn mutant (Marcotti et al., 2006). The causal relationship between a functional MT channel and subsequent maturation will be a worthwhile problem to address.

This work is supported by grant RO1 DC01362 from the National Institute on Deafness and other Communication Disorders to R. Fettiplace.

Edward N. Pugh Jr. served as editor.