Grxcr1 regulates hair bundle morphogenesis and is required for normal mechanoelectrical transduction in mouse cochlear hair cells

Ethical statement

Mouse studies were carried out in accordance with UK Home Office regulations and the UK Animals (Scientific Procedures) Act of 1986 (ASPA) under UK Home Office licences, and the study was approved by the Sanger Institute, King’s College London and University of Sheffield’s Ethical Review Committees. Mice were culled using cervical dislocation as approved under these licences to minimise any possibility of suffering.

Tasmanian devil mice

Tasmanian devil (Grxcr1^tde, referred to as tde) mice arose from microinjection of a human placental alkaline phosphatase genomic construct causing an insertional mutation [22]. The line was maintained on a mixed C57BL/6J; CBA/Ca genetic background. Both male and female mice were used as we observed no differences between them. The colony was generally maintained by homozygote by heterozygote matings, and experiments used homozygous mutants compared with heterozygous or wild type littermates as controls. Mice were group-housed in individually-ventilated cages with lights on a 12 hour on/12 hour off cycle, and experiments were conducted throughout the working day, from two hours to ten hours after lights were switched on. Mice were inspected for any signs of ill-health every day, as required by the UK Home Office, and were maintained in a Specific Pathogen Free facility. Genotypes were determined by abnormal behaviour indicative of a balance defect in mature mice or for mice younger than two weeks old by examination of the other ear by either scanning electron microscopy and detection of thin stereocilia in homozygotes by two independent viewers or by direct imaging in dissected organ of Corti, or by qRT-PCR of inner ear cDNA to detect reduced expression levels of Grxcr1 in the homozygotes. The Tasmanian devil mutants are available from the European Mouse Mutant Archive (EMMA).

Preparation of DNA samples and PCR analysis

Genomic DNA was extracted from ear biopsies or post-mortem tail skin using phenol/chloroform extraction with standard protocols. Complementary DNA (cDNA) was prepared after RNA extraction and reverse transcription. Whole inner ears from littermate heterozygotes and tde/tde mice and C57BL/6N wild-type mice were dissected and stored in RNA later stabilization reagent (QIAgen) at -20°C. Total RNA was extracted using QIAshredder columns and RNeasy mini kit (QIAgen) according to manufacturer’s instructions. DNAse I (Invitrogen) treatment was used to remove DNA. Reverse transcription reactions were performed using Superscript II First-strand synthesis system kit (Invitrogen).

Design of primers for mouse Grxcr1 gene

Primers for the genomic or the transcript sequence of the mouse Grxcr1 gene (ENSMUSG00000068082; Ensembl version 56) were designed using Primer3 software [23]. Transcript primers are located in exons 1 and 4, and are 134kb apart in the genome, ensuring that only cDNA could be amplified from them. The sequences of the primers are detailed in Table 1. PCR was performed using a touchdown protocol:

1. 94°C for 2 minutes

2. 94°C for 30 seconds

3. 64°C for 45 seconds (decreasing by 0.5C per cycle)

4. 72°C for 45 seconds

5. Repeat 2–4 15 times

6. 94°C for 30 seconds

7. 55°C for 45 seconds

8. 72°C for 45 seconds

9. Repeat 6–8 20 times)

10. 72°C for 7 minutes

Analysis of the sequences

PCR products were separated in agarose gels and purified (QIAgen) according to the manufacturer’s protocol. The purified samples were used for capillary sequencing and the results analysed using DNAStar Lasergene 8 software.

Scanning electron microscopy (SEM)

Mice at postnatal day 7 (P7; +/+ and +/tde n = 15; tde/tde n = 12 mice) or at P4 (+/+ and +/tde n = 9; tde/tde n = 3 mice) were selected for SEM. Inner ears were dissected and fixed in 2.5% glutaraldehyde in 0.1M phosphate buffer for 3 hours at room temperature, followed by washes in 0.1M phosphate buffer. Samples were then prepared using the osmium tetroxide-thiocarbohydrazide method (OTOTO) [24]. Processed cochleae were dehydrated followed by critical point drying using pressurized carbon dioxide. Samples were mounted and examined under a scanning electron microscope (Hitachi 4800 FE) at 5 kV. Images were processed using Adobe Photoshop, and all adjustments to settings were applied equally to control and mutant images.

Quantitative real-time PCR

Inner ears were dissected and RNA was extracted from whole inner ears at P7 (+/tde: n = 5; tde/tde: n = 11 mice) and at P12 (n = 2 +/tde; n = 4 tde/tde mice). Total RNA was isolated using RNeasy Mini Kit (Qiagen, Valencia, CA, USA) following the manufacturer’s instructions. RNA was first normalised, then treated with DNase I (Sigma, UK) to remove any remaining DNA, followed by reverse-transcription to cDNA using Superscript II-reverse transcriptase (Invitrogen, Carlsbad, CA, USA) and a poly(dT) primer (Invitrogen, cat no 18418012). Quantitative Real Time PCR (qRT-PCR) was carried out on an ABI Prism 7000 Sequence Detection System (Applied Biosystems) using TaqMan reagents and probes (Grxcr1 (Applied Biosystems: Mm00473947_m1), Hprt, (Mm01318747_g1), Kcnq4 (Mm01185500_m1) and Prestin (552740-C3)). Relative expression levels of each transcript were calculated using the 2^-ΔΔCt equation [25] and normalized to Hprt (a house-keeping gene). Each sample was run with 3 technical replicates and compared to littermates. The Wilcoxon rank sum test (Mann-Whitney U test) was chosen to determine significance, because it is a suitable test for small sample sizes and populations of unknown characteristics [26].

5’ RACE

5’ RACE PCR (5’ rapid amplification of 5’ cDNA ends) was performed to analyse the mRNA structure and expression of Grxcr1 by generating full-length cDNA from the mouse Grxcr1 gene. This technique involves three sequential enzymatic steps: reverse transcription from a primer designed to match the 3’ end of the transcript (3’UTR-exon 4, Table 1) followed by degradation of the mRNA template; addition of homopolymeric poly-A tail to the cDNA (5’ end of the transcript); and PCR using a poly(dT) primer to match the poly-A tail and a specific primer nested inside the first primer (Table 1). The resulting 5’ RACE products were cloned into an appropriate vector for sequencing and subsequent manipulation. We used a 5’/3’ RACE Kit, 2^nd generation (Roche) and specific primers (Table 1) including one located in the 3’UTR (exon 4) region of the mouse Grxcr1 transcript and the mRNA from wild-type C57BL/6N mice at P0 and P6 (n = 2 mice), and a heterozygous adult.

Immunofluorescence

Inner ears from P5-P12 mice were fixed in 4% paraformaldehyde for 2 hours at room temperature. After dissection in PBS, the organ of Corti was washed and permeabilized in 1% PBS/Triton-X-100 (PBT) before blocking with 10% sheep serum. The samples were incubated overnight at 4°C with the primary antibodies: EPS8 (Abcam, ab96144); MYO7A (Proteus Bioscience, 25–6790); ESPN (gift from Bechara Kachar and described in ref [27]); or PTPRQ (gift from Guy Richardson and described in ref [21]). Samples were then washed with PBT and incubated with anti-rabbit Alexa Fluor 488 secondary antibody (Invitrogen, anti-rabbit, diluted 1:300) and rhodamine/phalloidin (Invitrogen, diluted 1:200). Samples were mounted in Prolong Gold mounting medium (Molecular Probes). Images were acquired on either an LSM 510 Meta confocal microscope (Zeiss, Welwyn Garden City) or a Nikon A1R Confocal microscope, and post-acquisition image analyses were performed using Adobe Photoshop CS6.

Immunohistochemistry

Inner ears from P12 mice were fixed in 4% paraformaldehyde in PBS overnight at 4°C. Samples were embedded in paraffin wax, cut at 8 μm in a mid-modiolar plane and immunolabelled using the Ventana Discovery system (Ventana Medical Systems, Inc. Illkirch, France) according to manufacturer’s instructions. Antibodies to Kcnq4 (Santa Cruz, polyclonal goat, sc-9385) and prestin (Santa Cruz, polyclonal goat, sc-22694) were used as primary antibodies followed by anti-goat secondary antibody (Jackson ImmunoResearch, 705-065-147). For prestin, +/tde n = 2; tde/tde n = 5 mice. For Kcnq4, +/tde n = 2; tde/tde n = 2 mice.

Single-hair cell electrophysiology

Acutely dissected organs of Corti were used to study inner and outer hair cells (IHCs: n = 37; OHCs, n = 54) from Tasmanian devil mice and control littermates from postnatal day 7 (P7) to P13 for OHCs and P5 to P21 for IHCs, where the day of birth is P0. Animals were killed by cervical dislocation. Cochleae were dissected in normal extracellular solution (in mM): 135 NaCl, 5.8 KCl, 1.3 CaCl₂, 0.9 MgCl₂, 0.7 NaH₂PO₄, 5.6 D-glucose, 10 Hepes-NaOH. Sodium pyruvate (2 mM), MEM amino acids solution (50X, without L-Glutamine) and MEM vitamins solution (100X) were added from concentrates (Fisher Scientific, UK). The pH was adjusted to 7.5 (osmolality ~308 mmol kg^-1). Recordings were performed either at room temperature (22-24°C) or at body temperature (34-37°C).

Voltage and current clamp recordings were performed using the following intracellular solution in the patch pipette containing (in mM): 131 KCl, 3 MgCl₂, 1 EGTA-KOH, 5 Na₂ATP, 5 Hepes-KOH, 10 Na₂-phosphocreatine (pH 7.3; osmolality ~296 mmol kg^-1). Exocytosis and mechanoelectrical transduction recordings were measured using the following intracellular solution (in mM): 106 Cs-glutamate, 20 CsCl, 3 MgCl₂, 1 EGTA-CsOH, 5 Na₂ATP, 0.3 Na₂GTP, 5 Hepes-CsOH, 10 Na₂-phosphocreatine (pH 7.3). Patch pipettes were coated with surf wax (Mr Zoggs SexWax, USA) to minimise the fast patch pipette capacitance transient. An Optopatch (Cairn Research Ltd, UK) amplifier was used to perform electrophysiological recordings. Data acquisition was controlled by pClamp software using Digidata 1440A boards (Molecular Devices, USA).

The recorded basolateral membrane currents were low-pass filtered at 2.5 kHz (8-pole Bessel), sampled at 5 kHz and stored on computer for off-line analysis (Origin: OriginLab, USA). Membrane potentials in voltage clamp were corrected for the voltage drop across the uncompensated residual series resistance (R_s: 1.5 ± 0.2 MΩ, n = 63, after up to 80% compensation) and for a liquid junction potential (K⁺- and Cs⁺-based intracellular solution: –4 mV and –11 mV, respectively).

Real-time changes in membrane capacitance (ΔC_m) were measured using a hardware method (Optopatch). A 4 kHz sine wave of 13 mV RMS was applied to IHCs from −81 mV and was interrupted for the duration of the voltage step (100 ms and in 10 mV increments). The capacitance signal from the Optopatch was amplified (×50), filtered at 250 Hz, and sampled at 5 kHz. ΔC_m was measured by averaging the C_m trace over a 200 ms period following the voltage step. ΔC_m was recorded while applying 30 mM TEA and 15 mM 4AP (Fluka, UK) and additionally 60 μM linopirdine and 300 nM apamin (Tocris, UK) to reduce K⁺ currents. Synaptic transfer relations were fitted using a power function ΔC_m∝I_Ca^N (Eq 1), where N is the power. The average N values reported are from fits to all individual cells tested.

Mechanoelectrical transducer (MET) currents were elicited by stimulating the hair bundles of OHCs using a fluid jet from a pipette (tip diameter 8–10 μm) driven by a piezoelectric disc). The pipette tip of the fluid jet was positioned near to the bundles to elicit a maximal MET current. Mechanical stimuli were applied as force-steps or saturating 50 Hz sinusoids (filtered at 0.25 kHz, 8-pole Bessel) with driving voltages of ± 40 V. Membrane potentials were corrected for the liquid junction potential (Cs-based intracellular solution: –11 mV). Voltage clamp protocols are referred to a holding potential of –81 mV. The peak MET current-voltage curves were fitted according to a simple single-energy-barrier model [28]: I(V) = k [exp ((1 − γ)(V − V_r) /V_s) − exp (− γ(V − V_r) /V_s)] (eqn.2), where k is a proportionality constant, V_r is the reversal potential, V_s is a measure for the steepness of the rectification, and γ is the fractional distance within the membrane’s electrical field of an energy barrier, as measured from the outside.

Statistical comparisons of means were made by Student’s two-tailed t-test. Mean values are quoted ± s.e.m. where P < 0.05 indicates statistical significance.

Tasmanian devil mutants are deaf and have abnormally thin stereocilia

As described previously [10, 14, 15], Tasmanian devil (tde) homozygous mutants display hyperactivity, head-bobbing and circling, indicative of vestibular defects, and they lack a Preyer reflex (ear flick in response to noise) and cochlear responses to sound suggesting a severe hearing impairment. Scanning electron microscopy of the organ of Corti at P7 confirmed that the stereocilia of tde homozygous (tde/tde) mice were abnormally thin with a flaccid and disorganized appearance in both inner hair cells (IHC) and outer hair cells (OHC) compared to control heterozygous (+/tde) mice (Fig 1).

Reduced expression of Grxcr1 mRNA in the inner ear of Tasmanian devil mice

GRXCR1 has previously been shown to be specifically expressed in the cochlea and localized to the neuroepithelium containing hair cells and supporting cells of pre-hearing mice at P5 [10]. We used the gEAR (http://umgear.org) to look at Grxcr1 expression in single cell and bulk RNAseq datasets at P0 [29, 30] and P1 [31] and found it was strongly expressed in hair cells at these stages. We carried out a quantitative analysis of Grxcr1 mRNA levels in the inner ear of tde/tde and control littermates at P7. We found that the expression of Grxcr1 was almost completely abolished in tde/tde inner ears at P7 (Fig 2; P<0.001, Wilcoxon rank sum test, +/tde: n = 5; tde/tde: n = 11).

Molecular analysis of the Grxcr1 gene in Tasmanian devil mice

Tasmanian devil mice arose when a RSV-PALP transgene became inserted by accident into the first intron of Grxcr1 [10, 14, 22]. To analyse the effect of this transgene insertion on the genomic and transcript sequence of Grxcr1 in tde mutant mice, we performed PCR analysis and subsequent sequencing of the amplified bands (see Table 1 for the primers). One transcript of Grxcr1 is described in Ensembl: Grxcr1-201 (ENSMUST00000094715); 1,071 base pairs in 4 exons), which is translated into a protein of 296 amino acids (www.ensembl.org, accessed June 2020). We analysed the genomic sequence of all four coding exons of the Grxcr1 gene using primers specific for each exon (Table 1), covering the entire coding region and the splice sites, using genomic DNA from phenotyped adults (n = 1 +/+ (C57BL/6N); n = 12 +/tde; n = 16 tde/tde mice). Exons were sequenced by capillary sequencing and traces were analysed. No changes were found in the coding sequence of any exon of the Grxcr1 gene in the homozygous mutants compared to wild-type C57BL/6N and +/tde mice (Fig 3A). PCR amplification of total cDNA from adult inner ears and brains using primers located within Grxcr1 exons 1 and 4 showed that two bands of different size were amplified in both mutant and control samples (Fig 3B and 3C). To find out whether the two amplified bands corresponded to two different transcripts, we performed 5’ RACE-PCR (PCR for rapid amplification of cDNA ends) from inner ear samples of C57BL/6N wild-type mice at P0 and P6, using a primer in the 3’UTR for the initial reverse transcription. The sequencing of the two amplified bands revealed that the larger band contained all four exons whereas the smaller band skipped exon 2 and is therefore a novel transcript (Fig 3D and 3E; S1 File). This smaller, novel transcript, which we observed in cDNA made using a poly-d(T) primer as well as in the 5’RACE, had clearly undergone intron excision, so we conclude that it is likely to represent the mature mRNA sequence. Therefore, our results show the presence of two different transcripts of Grxcr1, which are identical in control and tde/tde inner ears. Exon 2 has 243 base pairs and starts and ends in phase 0, so its absence does not change the amino acid sequences of exons 1, 3 or 4, but the resulting protein would lack most of the glutaredoxin domain so likely has a different function to the full length GRXCR1 (Fig 3F; S1 File). In conclusion, while the transgene insertion causes a considerable reduction of Grxcr1 expression, it does not affect the transcript sequence.

Key hair bundle proteins are distributed normally in Tasmanian devil stereocilia

The main phenotype characterizing tde/tde is the presence of abnormally thin stereocilia with a flaccid appearance in cochlear and vestibular hair cells (Fig 1B, 1E and 1F). The stereocilia bundle shows progressive disorganisation from birth onwards, ultimately leading to extensive hair cell degeneration in adults [14]. Therefore, we analysed the expression patterns of several key proteins known to play a role in the organization of the stereociliary actin core in the development of the hair bundle: ESPN, MYO7A, EPS8 and PTPRQ. We used immunostaining and confocal microscopy on whole-mount preparations of the organ of Corti from P5-P12 mice to analyse the distribution of these proteins. The disorganisation of the stereocilia makes assessment of labelling challenging, so we have focussed on comparing the general features of labelling between mutants and controls.

ESPN is an actin bundling protein expressed in the core of stereocilia and plays an important role in hair bundle development [32, 33]. Our results indicate that ESPN is expressed in a punctate way along the entire length of the stereocilia (Fig 4A–4F) of tde/tde and control hair cells. The punctate labelling has recently been described elsewhere [34].

MYO7A is an unconventional myosin expressed in sensory hair cells and stereocilia, and known to be involved in developing the correct hair bundle architecture [19, 35]. Mice deficient in MYO7A (shaker1 mutants) show severe hair bundle disruption and abnormally elongated stereocilia [19, 36], which are also characteristics present in hair cells from tde/tde mice. The pattern of expression of MYO7A in the hair cells was broadly similar in tde/tde and control mice, with labelling evenly distributed within the top of the hair cell (Fig 4G–4L).

EPS8 is an evolutionarily conserved signal transducer known to have multiple functions in the control of actin dynamics. Eps8 mutant mice have shorter and more numerous stereocilia than those of control mice [37]. The localization of labelling was observed at the tips of stereocilia (Fig 4M–4R) in both tde/tde mutants and littermate controls.

PTPRQ is a protein generally concentrated at the base of stereocilia and is required for their normal structure and function [38, 39]. Mutations in Ptprq lead to different degrees of hair bundle disorganization and formation of giant stereocilia, and hearing loss [21, 39], as seen in tde/tde mice [14]. We found that the distribution of PTPRQ in the stereocilia of tde/tde mice was generally similar to that of littermate controls (Fig 4S–4X).

Mechanoelectrical transducer current is reduced in Tasmanian devil homozygotes

We then investigated whether the abnormal hair bundles of tde/tde mice had any physiological consequence by testing their ability to transduce mechanical stimuli. The mechanoelectrical transducer (MET) current was recorded from P7 apical-coil OHCs by displacing their hair bundles using a piezo-driven fluid-jet. Upon moving the bundles in the excitatory direction (i.e. towards the taller stereocilia) and at negative membrane potentials, a large inward MET current could be elicited in OHCs from both control (+/+) and mutant (tde/tde) mice (Fig 5A and 5B, respectively). However, the maximum MET current at −121 mV was significantly smaller in mutant (−837 ± 34 pA, n = 4) compared to control (−1463 ± 41 pA, n = 5, P<0.0001, t-test) mice. Any resting current flowing through open MET channels in the absence of mechanical stimulation was reduced when bundles were moved in the inhibitory direction (i.e. away from the taller stereocilia) in all control and mutant OHCs (Fig 5A and 5B, arrows). The fraction of the resting MET current was also significantly reduced in mutant (0.026 ± 0.002, n = 5, at −121 mV) compared to control (0.079 ± 0.008, n = 4, P<0.0005, t-test) mice. Because the MET current reverses near 0 mV, it became outward when excitatory bundle stimulation was applied during voltage steps positive to its reversal potential Fig 5A–5C). At positive potentials, the larger resting transducer current, which was more evident in control OHCs (e.g. +99 mV in Fig 5A and 5B: arrowheads), is due to an increased open probability of the transducer channel resulting from a reduced driving force for Ca²⁺ influx [40, 41].

We then investigated whether the adaptation properties of the MET current were affected in Tasmanian devil mice by stimulating the hair bundles of OHCs using mechanical step stimuli instead of sinusoids. Excitatory bundle movements with non-saturating stimuli elicited rapid inward currents at a holding potential of –81 mV that declined or adapted over time in both control and mutant P7 OHCs (Fig 5D, arrows). Inhibitory hair bundle stimulation shut off the small fraction of the current flowing at rest and at the offset of large inhibitory steps they caused a transient rebound (downward dip: Fig 5D, arrowheads).

We further tested the Ca²⁺ sensitivity of the transducer apparatus by deflecting the bundle in the presence of an endolymph-like solution containing 0.02–0.04 mM Ca²⁺ [42, 43]. Reducing Ca²⁺ influx into the transducer channel by either depolarizing hair cells to near the Ca²⁺ equilibrium potential (as shown in Fig 5A) or lowering the extracellular Ca²⁺ concentration from 1.3 mM to 0.04 mM increased both the resting and the maximal MET current in both control and mutant OHCs (Fig 5E), as previously described in non-mammalian vertebrates and mammals [41, 44, 45]. This is because Ca²⁺ is a permeant blocker of the MET channel and as such the increased MET current amplitude in low Ca²⁺ reflects the partial relief of this block. Moreover, extracellular Ca²⁺, by binding to the MET channel pore [41, 44, 45] or with the lipid environment [46] is able to directly regulate the open probability of the MET channel. Although the size of the MET current was significantly increased in low-Ca²⁺ (P<0.0001) in both control and mutant OHCs, the increment was much larger in the former (control: 2.5x; mutant: 1.5x, Fig 5F). The fraction of the resting open MET current was also significantly increased in control OHCs (P<0.001) but not in mutant cells (P>0.05, one-way ANOVA, Fig 5G). The above results indicate that the size and Ca²⁺ sensitivity of the MET current are affected in the OHCs from tde/tde mice.

IHCs from Tasmanian devil homozygous mice do not develop adult-type ion channels

We then investigated whether the biophysical properties of hair cells were affected in Tasmanian devil mice. The resting membrane potentials and size of K⁺ currents recorded from immature (<P12) control (+/+) IHCs were similar to those in tde/tde cells (Table 2). Moreover, all immature IHCs investigated were able to generate spontaneous repetitive Ca²⁺ action potentials (Fig 6A and 6B for control and tde/tde IHCs, respectively) as previously shown in wild-type cells [47]. These results show that the biophysical properties of pre-hearing IHCs (<P12 in most rodents) are not affected in early postnatal tde/tde mice. We then investigated whether IHCs were able to acquire functional maturity, which is associated with the down-regulation of immature ion channels and acquisition of adult-like electrical properties including a rapidly activating large conductance Ca²⁺-activated K⁺ current (I_K,f) and a current carried by KCNQ4 channels (I_K,n) with an unusually hyperpolarized activation range. Potassium currents in IHCs were elicited in response to depolarizing voltage steps in 10 mV increments from –154 mV (holding potential –64 mV). Both I_K,f and I_K,n were present in adult IHCs (P20-P21) from control (Fig 6C), but absent in cells from tde/tde mice (Fig 6D; see also Table 2), which instead retained an immature phenotype by expressing the inward rectifier K⁺ current I_K1 and the small conductance Ca²⁺ activated K⁺ current I_SK2. While control IHCs generate fast and graded voltage responses to current injection of up to 900 pA (Fig 6E), as previously described in adult IHCs of normal CD-1 mice [48], mutant cells retained the ability to generate an initial Ca²⁺ action potential (Fig 6F).

Another characteristic of IHC development is that they lose the ability to respond directly to the efferent neurotransmitter acetylcholine (ACh) by P14-P16 [49, 50]. The ACh-activated current in IHCs is mediated by Ca²⁺ entering hair cells through α9α10-nAChRs and activating SK2 channels [51, 52]. In agreement with previous findings, adult control IHCs did not respond to the extracellular application of 100 μM ACh (Fig 6G). In contrast, all mutant IHCs showed a large ACh-activated current at around the holding potential of −84 mV (Fig 6H, see also Table 2), which further supports the immature state of Tasmanian devil mutant IHCs.

We suggest that the immature state of IHCs in tde mutants is likely to be the consequence, not the cause, of reduced transduction current by analogy with other recent observations showing that defects in the MET current directly influence IHC development [53].

Exocytosis in adult IHCs from homozygous Tasmanian devil mice fails to mature

We investigated whether, in addition to the K⁺ currents, the development of I_Ca and the induced exocytosis fail to mature in tde/tde mice. Exocytosis was estimated by measuring increases in cell membrane capacitance (ΔC_m) following depolarizing voltage steps, which is generally interpreted as an indication of neurotransmitter release from presynaptic cells. The synaptic machinery of IHCs becomes more sensitive to Ca²⁺ as they mature, causing synaptic vesicles to be released linearly with increases in Ca²⁺ current. We found that in IHCs from tde/tde mice the developmental linearization of the exocytotic Ca²⁺ sensitivity did not occur. In adult tde/tde IHCs the maximal size of the Ca²⁺ current (I_Ca) was significantly larger (P<0.0005) than that of control cells (Fig 7A–7C), but similar to that normally seen in pre-hearing cells (e.g. [54]). However, the corresponding ΔC_m was similar between the two genotypes (Fig 7B and 7C, lower panel). As a consequence the exocytotic Ca²⁺ dependence, defined as the variation in ΔC_m as a function of I_Ca and displayed as a synaptic transfer function, was significantly less linear in IHCs from tde/tde mice (power of 2.02 ± 0.14, n = 4) than in control adult IHCs (power of 0.98 ± 0.07, n = 4: Fig 7D), and was more similar to that of immature cells [55, 56].

OHC basolateral membrane properties fail to mature completely in Tasmanian Devil mice

We finally investigated whether adult OHCs from Tasmanian Devil mutant mice were also unable to mature as in IHCs. Potassium currents were elicited by applying depolarizing voltage step of 10 mV increments from –124 mV (holding potential –84 mV). The resting membrane potential of mature OHCs was significantly more depolarized in mutant than in control cells (P<0.0005, see Table 2). I_K,n, the major current component in adult mouse OHCs, was largely reduced in mutant compared to control cells (Fig 8A and 8B; Table 2). We measured expression levels of two genes that are strongly expressed in outer hair cells, Slc26a5 (prestin) and Kcnq4, and both showed apparently reduced immunolabelling (although this is not a quantitative measure; Fig 8C–8F). qRT-PCR showed that Kcnq4 in the mutants was 63% of normal levels, although this was not a significant difference (n = 4 tde/tde; n = 2 +/tde, Wilcoxon rank sum test) (Fig 8G). Any reduced expression of Kcnq4 would be correlated with the reduced I_K,n current recorded from OHCs of tde/tde mice (Fig 8B).

Grxcr1 is down-regulated in Tasmanian devil mutant mice

In tde/tde mice, the chance insertion of the RSV-PALP transgene into intron 1 of the Grxcr1 gene disrupts its expression. In 5’ RACE PCR experiments, we showed the presence of two different Grxcr1 transcripts in the inner ear of wild-type mice, including a new transcript that skips exon 2. The sequences of these transcripts were not affected in tde mutants. However, quantitative analysis of Grxcr1 transcripts revealed a considerably reduced level in tde mutants. In pirouette (Grxcr1^pir) mice the same intron is affected [10], suggesting that this chromosomal region is critical for normal Grxcr1 expression. Low amounts of mRNA suggest residual GRXCR1 protein might be expressed in hair cells, but not enough to maintain normal function.

Indirect functional consequences associated with the down-regulation of Grxcr1

A functional MET current during pre-hearing stages of development, which contributes to setting the resting membrane potential of hair cells, is essential for establishing the biophysical and morphological properties of mature IHCs [53]. The development of OHCs also appear to be highly dependent on them having normal biophysical characteristics during immature stages [57, 58]. In Tasmanian devil mice, the stereociliary bundles of hair cells are abnormally thin from around birth and become progressively more disorganized with age (Fig 1; [14]). The morphological defects in the stereociliary bundle of tde/tde mice caused a largely reduced size of the MET current and of the resting open probability (Po) of the MET channels. While the smaller MET current could be caused by the loss of tip links between the disorganized stereocilia, the reduced resting Po and its Ca²⁺ sensitivity are most likely caused by a reduced tension in the remaining tip links. A reduced MET channel Po has also been reported in several mouse models affecting the bundle structure (e.g. Eps8 [37]; Mir96 [59]) or Ca²⁺ handling at the stereocilia (Atp2b2 [60]). The indirect functional consequence of the abnormal MET current apparatus was that Tasmanian devil mutant mice fail to acquire the adult-like basolateral membrane proteins, preventing them from becoming fully functional sensory receptors. This hypothesis is also supported by the observation of a similar failure in the acquisition of the mature IHC and OHC biophysical characteristics in mice lacking other stereociliary proteins associated with severe morphological hair bundle defects, such as TMC1 [61], EPS8 [37], USH1C and MYO7A [53].

Role of Grxcr1 in hair bundle growth

Gene expression and immunostaining assays have shown that GRXCR1 is present in the sensory hair cells of the inner ear from early stages of development onwards, and localizes in the stereociliary bundle [10]. Cochlear hair cells from mice homozygous for mutations in the Grxcr1 gene (pirouette and Tasmanian devil) have abnormally thin and long stereocilia, and their hair bundles become progressively more disorganized with age [10, 14]. Similar stereocilia defects were also seen in inner ear hair cells from Grxcr1 zebrafish mutants [16]. Despite these abnormalities, the hair bundles establish their usual orientation (Fig 1B) and a rudimental staircase structure [2, 14], indicating that GRXCR1 is unlikely to influence the initial bundle morphogenesis. Instead, GRXCR1 appears to be primarily involved in regulating the widening and the final elongation of stereocilia, which in mammals occur simultaneously during the final stages of hair bundle formation [62]. Changes in the stereocilia dimensions are thought to involve several actin-bundling and -capping proteins together with the unconventional myosin motors and their cargo [1, 63].

The widening of the stereocilia is determined by the addition of actin filaments to the parallel actin bundle [2, 64], a process that depends on the ability of the actin-bundling protein ESPN to bind to and cross-link the actin filaments [17, 65]. ESPN is also involved in the elongation of the stereocilia since mice deficient for the protein show both abnormally thin and short stereocilia [17, 32, 66]. ESPN seems to require MYO7A and MYO3B in order for it to be transported to the stereocilia tip so it can exert its role in stereocilia elongation [67, 68]. In addition to MYO3 and ESPN, several other protein complexes are needed to achieve the extremely precise regulation of the stereocilia actin-cytoskeleton required to establish the fine structure of the hair bundle, with stereocilia length being identical among rows, and between neighbouring hair cells, but changing along the cochlea and between rows within each bundle [62]. The deafness gene Eps8 (mice: [37]; human: [69]) encodes an actin-binding protein with multiple functions in the control of actin dynamics [70, 71]. In cochlear hair cells it is primarily localized at the tip of the longest row of stereocilia [18, 37, 72], and is key for their normal elongation by forming a complex with MYO15 and WHRN [18]. Another important protein complex is that formed by the unconventional MYO7A and its cargo TWF2, which is an actin capping protein able to restrict the elongation of the shorter rows of stereocilia [20, 73]. Stereocilia length can also be regulated by activity at the minus end of actin filaments at the base of the stereocilia [74]. PTPRQ is a receptor-like inositol lipid phosphatase required for normal stereocilia development [21], which requires the presence of MYO6 in order to localize to the stereocilia base [39]. Mutations in Ptprq lead to different degrees of hair bundle disorganization, including bundle shortening or giant stereocilia [21, 59].

In GRXCR1 deficient mice, we found ESPN, MYO7A, EPS8 and PTPRQ, which are involved in the four complexes listed above, surprisingly were distributed in a generally normal way within the mutant stereocilia despite the abnormalities visible in the hair bundles. GRXCR1 is an enzyme that has been shown to remove glutathione groups from target proteins, and it has been suggested to destabilize the interaction between the Usher Syndrome proteins USH1C and USH1G in zebrafish [16]. Although mice deficient for USH1C and USH1G do not exhibit thinner stereocilia, their hair bundles are profoundly disorganized [75]. Overall, our data indicate that GRXCR1 directly contributes to coordinating the final stages of hair bundle morphogenesis. This view is also supported by the observation that mice deficient for the paralogous vertebrate gene Grxcr2, which is also localized to the hair cell stereocilia, have profound hearing loss due to severe defects in the organization of their hair bundles [76].