The ribbon synapses of auditory inner hair cells (IHCs) undergo morphological and electrophysiological transitions during cochlear development. Here we report that myosin VI (Myo6), an actin-based motor protein involved in genetic forms of deafness, is necessary for some of these changes to occur. By using post-embedding immunogold electron microscopy, we showed that Myo6 is present at the IHC synaptic active zone. In Snell's waltzer mutant mice, which lack Myo6, IHC ionic currents and ribbon synapse maturation proceeded normally until at least post-natal day 6. In adult mutant mice, however, the IHCs displayed immature potassium currents and still fired action potentials, as normally only observed in immature IHCs. In addition, the number of ribbons per IHC was reduced by 30%, and 30% of the remaining ribbons were morphologically immature. Ca2+-dependent exocytosis probed by capacitance measurement was markedly reduced despite normal Ca2+ currents and the large proportion of morphologically mature synapses, which suggests additional defects, such as loose Ca2+-exocytosis coupling or inefficient vesicular supply. Finally, we provide evidence that Myo6 and otoferlin, a putative Ca2+ sensor of synaptic exocytosis also involved in a genetic form of deafness, interact at the IHC ribbon synapse, and we suggest that this interaction is involved in the recycling of synaptic vesicles. Our findings thus uncover essential roles for Myo6 at the IHC ribbon synapse, in addition to that proposed in membrane turnover and anchoring at the apical surface of the hair cells.
Hearing relies primarily on the ability of cochlear inner hair cells (IHCs) to encode sound information with an extreme sensitivity and a high temporal precision (1). Sound-evoked mechanical stimuli are transduced by the mechanoelectrical transduction machinery of the IHC hair bundles into graded membrane potential variations that drive neurotransmitter release. This release can be maintained at high rate in response to sustained stimulation (2). Each mature IHC has 10 to 20 synaptic active zones that are characterized by the presence of an electron-dense structure of submicron diameter, the synaptic ribbon, to which synaptic vesicles are tethered, and which faces the dendrite of a bipolar auditory neuron (3,4).
During the first 2 post-natal weeks in the mouse, the IHCs undergo morphological and electrophysiological changes. The magnitudes of both IHC Ca2+ currents and Ca2+-dependent exocytosis increase until post-natal day 6 (P6) (5,6). The Ca2+ currents then decline progressively during IHC maturation, reaching their mature steady-state level by P12, but leaving the size of Ca2+-induced exocytosis almost unaffected. In parallel, the number of ribbons per IHC decreases, and the ribbon shape evolves from spherical to plate-like (3,7). Concomitant with the onset of hearing (P12), the IHCs also acquire fast membrane electrical properties, enabling phase locking of the auditory nerve firing pattern to the periodicity of the auditory stimulus up to several kilohertz (kHz), mainly through the expression of the large and fast Ca2+-activated K+ current, Ik,f (6,8). To date, little is known about the molecular players that are involved in these transitions.
Myosin VI (Myo6) defects underlie the dominant and recessive forms of human hearing impairment, DFNA22 and DFNB37 (9–12), and the recessive deafness of Snell's waltzer mice (Myo6sv/sv) (13,14). Myo6 can function either as an actin-based transporter or as an anchoring protein (15). This myosin is unique in that it moves toward the minus end of actin filaments (16), which tend to be oriented toward the cell center. Because Myo6 is abundant in the cuticular plate and the pericuticular necklace of auditory hair cells (17), and mutant mice (14) and zebrafish (18) lacking Myo6 display disorganized hair bundles, most of the structural and functional analyses carried out so far have focused on the role of this myosin in the apical part of the hair cells (19–22). Myo6 is indeed essential for the structural integrity of the hair bundle (19,20). Here, we report that Myo6 is involved in the morphological and functional maturation of IHC ribbon synapses and is required for efficient Ca2+-dependent exocytosis in mature IHC ribbon synapses. We also show that Myo6 is necessary for the maturation of the electrophysiological membrane properties of the IHCs to occur.
Myo6 is present at the IHC ribbon synapse
Two rabbit immune sera were raised against the His-tagged C-terminal part of murine Myo6 (amino acids 1047–1253), and the specific antibodies were affinity-purified against Myo6 globular tail produced as a glutathione S-transferase (GST) fusion protein (GST-Myo6-gt, amino acids 981–1253). The specificity of these antibodies was established by the immunoreactivity of the inner ear sensory epithelia observed in the wild-type mice, but absent in Myo6sv/sv mice. The labeling was also absent when the primary antibody was omitted or pre-adsorbed on the antigen (Supplementary Material, Fig. S1c, f, g and h). In the inner ear of wild-type mice, Myo6 immunoreactivity was restricted to the sensory cells of the cochlea and vestibule, in agreement with previous reports (14,17) (Fig. 1A, and Supplementary Material, Fig. S1c and f). Immunolabeling experiments on whole-mount preparations of the organ of Corti analyzed by confocal microscopy at P6, P15 and P21, that is before and after the onset of hearing, showed that Myo6 is present throughout the hair cell soma (Fig. 1B). A strong labeling was observed in the two IHC regions known for their intense vesicular trafficking, namely the pericuticular necklace (23) and the basolateral region, where the afferent boutons of the primary auditory neurons are located (Fig. 1B) (24). We then carried out post-embedding immunogold electron microscopy experiments on sections of organs of Corti from P10 and P15 mice. In the IHCs, the gold particles were not only associated with the cuticular plate and the peri-cuticular necklace as reported (17,20), but were also seen in the synaptic region of the cells (Fig. 1C–F). In all the ribbon-containing sections examined (10 sections from 3 mice), most of the gold particles (n = 73) were located at the edge of the active zone (48% of the gold particles), whereas 16% of the particles were nearby or associated with the ribbon, and 36% were seen within the cytoplasm. Notably, many of the gold particles seen at the active zone (31%) were associated with tubular features that have been considered to be endocytotic structures (Fig. 1F) (25), thus suggesting that Myo6 is involved in synaptic vesicle recycling. The presence of the protein throughout the IHC soma, just like other key synaptic proteins including SNAP25, syntaxin1, otoferlin and VGLUT3 (26–28), may reflect the intense synaptic vesicular trafficking of the IHCs.
Calcium-dependent exocytosis is impaired in mature IHCs lacking Myo6
To investigate a possible role of Myo6 in the synaptic exocytotic process, we studied IHC Ca2+-dependent exocytosis in Myo6sv/sv mutant mice that lack Myo6 expression. We monitored voltage-gated Ca2+ currents and membrane capacitance changes (ΔCm) in P6-P7 (immature) and P26–P31 (mature) cochlear IHCs (Fig. 2A–C). At both stages, the Ca2+ currents (Fig. 2A and B, middle panels) were not significantly different between Myo6sv/sv mice and their heterozygous littermates. Current–voltage curve of peak ICa were fitted with the equation:2A, middle panel). Fits for P26–P31 Myo6+/sv IHCs yielded gmax = 7.8 nS, Vrev = 18 mV, V1/2 = −33 mV, S = 9.9 mV, and for P26–P31 Myo6sv/sv IHCs gmax = 6.2 nS, Vrev = 15 mV, V1/2 = −36 mV, S = 8.9 mV (Fig. 2B, middle panel). Reversal potentials in the mature IHCs were about 20 mV less depolarized than in immature IHCs, most likely due to incomplete block of residual outward currents through the K+ channels (6). In P6–P7 Myo6sv/sv IHCs, peak maximum inward ICa was −389 ± 30 pA, n = 4 (versus Myo6+/sv −396 ± 25 pA, n = 5), whereas in P26−P31 Myo6sv/sv IHCs maximum peak ICa was −184 ± 15 pA, n = 4 (versus Myo6+/sv −232 ± 15 pA, n = 5; P > 0.05 at both stages). The Ca2+ currents of Myo6+/sv and Myo6sv/sv IHCs declined significantly with age (P < 0.001 for both genotypes), indicating that the normal developmental downregulation of Ca2+ influx did occur in the IHCs from adult mutant mice (5,6). The Ca2+-induced exocytosis of IHCs from mature homozygous mutant mice, however, was markedly reduced compared with the heterozygotes (Fig. 2B, right panel): the maximum ΔCm was 11.3 ± 4.1 fF, n = 4, for Myo6sv/sv, versus 27.0 ± 2.8 fF, n = 5, for Myo6+/sv (P < 0.02). Such a difference was not found in P6–P7 immature homozygous mutants (maximum ΔCm 51.8 ± 8.0 fF, n = 4, for Myo6sv/sv, versus 59.3 ± 13.1 fF, n = 5, for Myo6+/sv; P > 0.05) (Fig. 2A, right panel). Next, we compared synaptic transfer functions (6) that were obtained by plotting ΔCm against peak ICa for 100 ms voltage steps over a range of membrane potentials from −81 mV up to −11 mV (Fig. 2C). Fitted lines are according to the power function: 2C). Ca2+-efficiency at the maximum inward ICa (6), which occurred near a membrane potential of −21 mV, in mature Myo6sv/sv IHCs (0.066 fF/pA) was about half that of IHCs of their heterozygous littermates (0.119 fF/pA). For smaller depolarizations, the effect became progressively less pronounced due to the reduced Ca2+-dependence, but it was maintained for Ca2+ currents >20 pA, so it would affect neurotransmitter release for physiologically relevant receptor potentials, spanning a membrane potential range of about −60 to −20 mV (Fig. 2B and C) (29). Taken together, these findings suggest that Myo6 is essential for efficient synaptic vesicle exocytosis in IHCs of the mature cochlea.
IHC ribbon synapse maturation is delayed in Myo6sv/sv mice
To investigate whether an abnormal structural development could contribute to the observed exocytotic defect, we first carried out quantitative analyses of the IHC ribbons in mutant and wild-type mice at P6, P15 and P21. In the apical part of the cochlea where exocytosis was studied, the ribbons were visualized by using anti-CtBP2/ribeye antibodies that label the ribbons and the cell nucleus (26). The ribbons were quantified after 3D reconstruction of the IHC immunolabeling analyzed by confocal microscopy (Fig. 3A). At P6, the number of ribbons harbored by the IHCs of Myo6+/+ and Myo6sv/sv mice was similar (27.3 ± 0.8, n = 31, versus 24.7 ± 0.7, n = 29, respectively; P > 0.05). Likewise, we did not observe a significant difference in the number of ribbons per IHC between the two genotypes at P15 (10.9 ± 0.5, n = 71, versus 12.3 ± 0.3, n = 65, respectively; P > 0.05). This result shows that the normal developmental downregulation of the number of ribbons (3) occurred also in the mutant mice. At P21, however, a moderate but statistically significant reduction in the number of ribbons was found in the IHCs of Myo6sv/sv mice (13.4 ± 0.3 ribbons per IHC, n = 70 from 3 Myo6+/+ cochleas, versus 8.9 ± 0.3 ribbons per IHC, n = 91 from 3 Myo6sv/sv cochleas; P < 0.05) (Fig. 3A). Ultrastructural analysis showed that at P6, the IHC synapses of Myo6+/+ and Myo6sv/sv mice were indistinguishable (Fig. 3B and C). The same analysis carried out at P21 confirmed our immunofluorescence results, showing that the number of ribbon synapses in Myo6sv/sv IHCs was lower than that in the wild-type mice (8 ± 0.7 per IHC of Myo6+/+, n = 10; versus 4.5 ± 0.4 per IHC of Myo6sv/sv, n = 21; P < 0.05). Notably, 70% of the ribbons encountered in P21 Myo6sv/sv IHCs (n = 57) exhibited a mature shape and were decorated with synaptic vesicles (Fig. 3H). However, 30% of the ribbons still had an immature shape, indicating a delay in the ribbon maturation process. The immature phenotype of the ribbons in Myo6sv/sv IHCs was even more pronounced at P15, at which stage most of the synapses had round-shaped ribbons anchored to the active zone through two tubular rodlets (Fig. 3E). In P15 wild-type mice, 46 of 51 ribbons examined (91%) had a plate-like shape characteristic of mature synapses (Fig. 3D), whereas in the mutant mice, such mature ribbons were rarely encountered (11 out 83 ribbons examined in 20 IHCs from 4 cochleas, i.e. 13%). Nevertheless, the immature ribbons (87%) had matured beyond the P6–P8 stage, as their shape was comparable with that of wild-type P10–P12 ribbons (3). Notably, many P15 Myo6sv/sv IHCs were still contacted by efferent nerve fibers, an immature feature that was not observed in the wild-type IHCs of the same stage (Fig. 3F and G). These results show that Myo6 is not necessary for the formation of the IHC ribbon synapse, but is required for its proper maturation and maintenance. This is consistent with what has been reported in neuronal synapses. In cultured hippocampal neurons derived from Myo6sv/sv mice, synapses seemed to develop normally, but were also 21% fewer than in cultures from wild-type mice (30,31). Notably, the loss is even greater when Myo6 was acutely disrupted (30,31).
IHCs of adult Myo6sv/sv mice exhibit immature electrophysiological membrane properties
At the onset of hearing (P12), the mature IHCs display the IK,f current, also called BK current, which enables the cells to switch from spiking pacemakers to high-frequency signal transducers (8,32,33).
The presence of evoked and spontaneous spiking in the absence of Myo6 was studied by recording voltage responses to current injection in three IHCs from three P5–P6 Myo6sv/sv mice. Spontaneous and evoked spiking activity was recorded in all three cells (Supplementary Material, Fig. S2), indicating that at least up to P6, the IHC basolateral currents develop normally in Myo6sv/sv mice, which is consistent with our morphological observations. Surprisingly, we found that IHCs from adult Myo6sv/sv mice also fired action potentials (Fig. 4A right), just like immature IHCs (8). This contrasts with the normal graded voltage responses recorded in the mature Myo6+/sv IHCs (Fig. 4A left). The resting potentials of mature Myo6sv/sv IHCs were hyperpolarized compared with those of their heterozygous littermates (−75.2 ± 2.2 mV, n = 6, versus −62.7 ± 2.2 mV, n = 5; P < 0.01). This suggested that mutant and heterozygote IHCs have differences in membrane currents, which we studied under voltage clamp. Myo6sv/sv IHCs lacked IK,f recorded in IHCs of mature heterozygous littermates (IK,f measured 3 ms following a voltage step to −25 mV: Myo6+/sv 2.27 ± 0.37 nA, n = 6, versus Myo6sv/sv –0.10 ± 0.02 nA, n = 6; P < 0.0001) (Fig. 4B). In addition, the delayed rectifier potassium current IK,s (8,33) recorded in P21–P31 Myo6sv/sv IHCs was smaller than that recorded in IHCs from heterozygotes (IK,s, measured by subtracting IK,f from the total current at the end of a voltage step to −25 mV: Myo6sv/sv 1.74 ± 0.14 nA, n = 5, versus Myo6+/sv 2.82 ± 0.31 nA, n = 5, respectively; P < 0.01) (Fig. 4C). Notably, Myo6sv/sv IHCs also lacked the deactivating inward K+ current IK,n, characterized by a very negative activation range, which appears in mature wild-type IHCs (34,35) and was found in Myo6+/sv mice. The amplitude of the IK,n current recorded was measured as the difference between the peak inward current and the steady-state current in response to a voltage step to −124 mV from a holding potential of −84 mV (36), and was −94 ± 24 pA (n = 3). The inward current in the Myo6sv/sv IHCs was distinctly different from that of the Myo6+/sv mice. Currents increased rather than deactivated during hyperpolarizing voltage steps in the Myo6sv/sv IHCs (Fig. 4D), thus resembling those of the inward rectifier IK1 found in immature IHCs. IHCs normally increase in size during development, which is reflected as an increase in resting membrane capacitance (Cm) (34). This increase failed to occur in Myo6sv/sv IHCs. Cm was 8.0 ± 0.1 pF in immature Myo6+/sv IHCs (P4–P6, n = 17), increasing to 9.9 ± 0.4 pF upon maturation (P25–P31, n = 17). In Myo6sv/sv IHCs, Cm was 7.5 ± 0.2 pF (P4–P7, n = 17) early in development, and 7.4 ± 0.3 pF later on (P21–P31, n = 15). Mature Myo6+/sv IHCs had a significantly larger Cm than the other groups (all P < 0.001). In conclusion, although the electrophysiological development of Myo6+/sv IHCs appeared entirely normal, as investigated in normal CD-1 mice (32,34), the Myo6sv/sv IHCs retained immature electrophysiological properties.
Myo6 and otoferlin interact in vivo
By using a yeast two-hybrid assay and molecular and biochemical analyses, we have identified Myo6 as a potential otoferlin-interacting protein. Otoferlin, a six C2-domain transmembrane protein of synaptic vesicles (Fig. 5A), is responsible, when mutated, for a recessive form of deafness (DFNB9) and has been proposed to be a major Ca2+ sensor at IHC ribbon synapses (26). The otoferlin N-terminal fragment, containing the first three C2 domains (amino acids 1–761), was used as a bait to screen a library constructed from P3–P5 mouse inner-ear sensory epithelia (Fig. 5A) (37). The Myo6 globular tail (Fig. 5A), a domain regarded as a cargo-binding region (22), was among the preys identified. The direct interaction between otoferlin and Myo6 was confirmed by in vitro binding assays. In addition to Myo6 full length (Myo6-fl), two different fragments of Myo6, its globular tail (Myo6-gt amino acids 981–1253) and the protein without its globular tail (Myo6-Δgt amino acids 1–981) (Fig. 5A), produced in vitro and labeled with 35S-methionine, were incubated with the N-terminal fragment of otoferlin produced as a GST fusion protein. Myo6 full-length and its globular tail domain did bind to the otoferlin fragment, whereas Myo6 deleted from its globular tail domain did not (Fig. 5B). We then assessed the co-localization of otoferlin and Myo6 in IHCs. Whole-mount preparations of the organ of Corti from P21 mice labeled for otoferlin and Myo6 were analyzed by confocal microscopy. Myo6 and otoferlin labelings largely overlapped in the IHCs: in particular, many co-labeled varicosities were visible in the basolateral region of these cells (Fig. 5C). We next carried out double-immunogold electron microscopy on sections of the organ of Corti, which showed the presence of the two proteins at the IHC active zones, but with slightly different distributions (Fig. 5D). Indeed, Myo6 was preferentially detected at the edge of the active zone (Fig. 1C–F), whereas otoferlin was mainly localized around the ribbon, as described previously (Fig. 5D) (26). In some locations, the two types of gold particles were found just a few nanometers apart, including at the edge of the active zone (Fig. 5D), suggesting that the two proteins can physically interact in vivo. This was confirmed by co-immunoprecipitation experiments carried out on protein extracts from transfected human embryonic kidney (HEK)-293 cells producing otoferlin and Myo6 (Fig. 5E), and on protein extracts from inner ears of P15 mice (Fig. 5F). The anti-otoferlin antibody immunoprecipitated not only Myo6, but also syntaxin1 and SNAP25, two members of the SNARE complex (26). The anti-Myo6 antibody, however, immunoprecipitated only otoferlin. This suggests that the SNARE proteins do not physically interact with Myo6, and that otoferlin interaction with Myo6 is independent of the interaction between otoferlin and the SNARE proteins. To determine which otoferlin domain(s) is involved in the binding to Myo6, the precise boundaries of each of the C2 domains present in the otoferlin N-terminal fragment, namely, C2A-C, were determined. This was achieved by combining programs of secondary elements search (HCA and PSI) and multiple sequence alignment (Dialign and Clustal W), followed by in vitro expression. The C2A, C2B and C2C domains were produced as GST-fusion proteins and purified by affinity chromatography. These proteins were then incubated with Myo6-ct (amino acids 835–1253) or Myo6-gt. Only C2B domain was able to specifically bind to Myo6 carboxy terminal tail (Fig. 5G and data not shown). To determine whether this interaction could be modulated by Ca2+ ions, we repeated this test in the presence of 2 mm EDTA or 1 mm free Ca2+. Modifying the free Ca2+ concentration did not affect the binding of C2B to Myo6 carboxy terminal tail (Fig. 5G). Indeed, C2B sequence lacks most of the residues known to be involved in Ca2+-binding in other C2 domains (38,39). It is noteworthy that a different otoferlin–Myo6 interaction has recently been reported, which involves the otoferlin C2D domain (40) that can bind to Ca2+ ions (26). The Ca2+-dependence of this interaction, however, remains to be determined.
We found that Myo6 is present at the IHC synaptic active zone and that its absence prevents the IHC ribbon synapse maturation from proceeding normally. We also found that Myo6 defect leads to a markedly reduced synaptic exocytosis and to the lack of several K+ currents in adult IHCs. Finally, we provide evidence that Myo6 and otoferlin, a putative Ca2+ sensor of synaptic exocytosis also involved in a genetic form of deafness, DFNB9, interact at the IHC ribbon synapse.
The implication of Myo6 in IHC exocytosis is consistent with several studies showing that this myosin is involved both in endocytotic and exocytotic membrane-trafficking pathways (41). This was shown not only in non-neuronal cells, in which Myo6 was found to be required for efficient secretion and for the maintenance of the structure of the Golgi apparatus (42), but also in synaptic vesicle exocytosis at hippocampal synapses (30,31). In hippocampal neurons, both spontaneous and evoked synaptic vesicle exocytosis were reduced by 35% in Myo6sv/sv mice (31,41). The more dramatic reduction of synaptic exocytosis reported here (∼60%) may reflect a specific requirement of the IHC ribbon synapse for a high rate of exocytosis and recycling, and/or the inability of mutant IHCs to develop compensatory mechanisms as efficient as hippocampal neurons do in the absence of Myo6. The marked decrease of IHC synaptic exocytosis observed in mature Myo6sv/sv mice could, in principle, result from a failure of IHC synaptic development, a vesicular trafficking defect, an inefficient Ca2+-exocytosis coupling, or a combination of these defects.
The finding that the IHCs of adult Myo6sv/sv mice harbor morphologically mature ribbon synapses excludes a major ultrastructural abnormality as the sole cause of the exocytosis defect. Likewise, the moderate reduction of the number of ribbons (30%) is unlikely to fully account for the dramatic decrease of IHC exocytosis, since in the bassoon mutant mice that have over 90% of ribbon loss, IHC exocytosis is largely unaffected (4).
Part of the IHC exocytosis defect of Myo6sv/sv mice could be the outcome of the delayed maturation of the synapse we observed, which ultimately leads to IHCs in which mature and immature ribbon synapses coexist. Such a configuration may hinder IHC transmitter release, as the two types of synapses have different dependences on local Ca2+ concentration. It is generally believed that during development, IHC exocytosis switches from a mode of Ca2+-dependence in which each release event requires the cooperative action of overlapping Ca2+ microdomains, to a nanodomain configuration in which very few channels are required for the release of a single vesicle (6,43,44). It has been suggested that the more efficient coupling between Ca2+ entry and vesicle exocytosis in the mature IHCs arises from a developmental tightening between synaptic vesicles and the Ca2+ channels, which would compensate the lower Ca2+ influx in mature IHCs (44–46). Therefore, the immature ribbon synapses may not function properly in IHCs that display smaller adult-type Ca2+ currents suitable for a release machinery operating in a nanodomain configuration.
Insufficient supply of synaptic vesicles due to an endocytosis defect could also contribute to the observed reduction of exocytosis. Indeed, Myo6 has been found to be associated with clathrin-coated pits/vesicles and to be able to modulate endocytosis in the kidney (42), a mechanism that is likely to also take place at glutamatergic synapses during synaptic exocytosis and the recycling of AMPA receptors (30,31). Myo6 is also required for the efficient transportation of nascent endocytotic vesicles from the actin-rich periphery of the retinal pigment epithelium cell to supply the early endosome compartment with endocytotic vesicles (47). In the IHC, endocytosis at the synaptic active zone and at the peri-cuticular necklace has been proposed to be the main source for IHC synaptic vesicle pool replenishment (24,25). The abundance of Myo6 in these two regions, its association with tubular structures, likely endocytotic structures, at the edge of the active zone, and the increasing evidence that the protein is involved in endocytotic membrane-trafficking together suggest that Myo6 is implicated in the transport of vesicles on actin filaments from the IHC apical region to the ribbon synapse and/or in the retrieval of IHC synaptic vesicles after exocytosis (41,48,49). Our results suggest that at the IHC ribbon synapse, Myo6 could achieve such a task through its binding to the vesicular membrane protein otoferlin. Two findings support this proposal. First, Myo6 and otoferlin are co-localized at the edge of the synaptic active zone, where IHCs endocytosis is suggested to take place (24,25). Second, the amount of otoferlin is apparently reduced in the Myo6sv/sv IHCs (40).
In addition to its role at the synapse, we report that Myo6 is necessary for the maturation of the electrophysiological membrane properties of the IHCs. Mature Myo6sv/sv mice lack IK,f and IK,n, whereas they exhibit a reduced amplitude of the delayed rectifier potassium current IK,s. IK,f current not only prevents the mature IHCs from firing action potentials, but also lowers their membrane electrical time constant to well under 1 ms (8), enabling phase-locking of transmitter release for sound frequencies up to a few kHz. In the absence of this current in Myo6sv/sv IHCs, the membrane time constant is expected to be abnormally large, which will prevent phase-locking of action potentials in the auditory nerve for sound frequencies above a few tens of Hz. Mature Myo6sv/sv mice are therefore expected to have defects in the precision of spike timing in the auditory nerve, as reported for BK knockout mice which lack IK,f (50). Somewhat unexpectedly, the immature electrophysiological membrane properties of the IHCs recorded in adult mutant mice did not include the synaptic Ca2+ currents, suggesting that maturation of K+ and Ca2+ currents in IHCs proceed independently. The absence of adult-type K+ currents in the Myo6sv/sv IHCs could in principle result from a gene expression defect or from a mistargeting of the corresponding channels. The latter possibility is the most likely, since a missorting of the BK channel α-subunit has recently been reported in the Myo6sv/sv IHCs (40).
Finally, it is worthy of note that although otoferlin and Myo6 directly interact, otoferlin and Myo6 mutants display major differences in their phenotypes. The lack of otoferlin does not interfere with IHC ribbon or K+ currents maturation, whereas it leads to an almost completely abolished Ca2+-induced exocytosis, despite normal Ca2+ currents (26). Moreover, this failure of Ca2+-induced exocytosis is already present at P6 (26), whereas in the absence of Myo6, the defect of Ca2+-induced exocytosis is partial and only appears in mature IHCs. This suggests that the retrieval of synaptic vesicles, where otoferlin may act as a cargo adapter, is just part of the variety of cellular functions wherein Myo6 is involved (51).
Together, our results broaden the roles of Myo6 in hair cells and uncover an unexpected and essential role of this protein in the maturation of the IHC electrophysiological membrane properties and in the morphological and functional maturation of the ribbon synapse.
MATERIALS AND METHODS
Snell's waltzer mice (Myo6sv/sv), which have an intragenic deletion of Myo6 leading to a null allele (13,14), were bred with C57BL/6J mice. Genotyping at the Myo6 locus was conducted by PCR as described in what follows. The day of birth was denoted P0. All experiments reported were carried out according to INSERM, Institut Pasteur and UK Home Office welfare guidelines.
Genomic DNAs from Myo6+/sv inter-cross litters were amplified using primers (5′-TGGTGAAAAGAGTCAACCTGTG-3′ and 5′-GCTTCAGCTCGATATTTTATT-3′), flanking the deletion in the Myo6sv allele (14). The reaction mixture contained 200 ng of genomic DNA, LA PCR Buffer II with 2.5 mm MgCl2, 0.4 mm of each dNTP, 500 nm of each primer and 0.05 unit/ml of reaction mixture of TaKaRa LA Taq polymerase (TaKaRa Bio Inc). PCR was performed as follows: 5 min denaturation at 95°C, followed by 30 cycles of 1 min denaturation at 94°C, 1 min annealing at 60°C, 2.5 min extension at 72°C and finally an additional 10 min at 72°C. Wild-type mice displayed a PCR product of 2255 bp, whereas Myo6sv/sv mice displayed a 1245 bp fragment, and DNA from Myo6+/sv mice generated both fragments (Supplementary Material, Fig. S2).
Yeast two-hybrid screening
We used the yeast two-hybrid cDNA library constructed from a P3–P5 inner-ear sensory epithelia previously reported (37). Yeast two-hybrid screenings were carried out as reported (52,53), using otoferlin amino terminal region (Otof-Nt, amino acids 1–761) as bait (Fig. 5A). The interacting ‘prey’ fragments of the positive clones were PCR-amplified and sequenced. The resulting sequences were used to identify the corresponding gene in the GenBank database (NCBI) using a fully automated procedure.
The Myo6 full-length cDNA (NM_001039546) was cloned by three successive rounds of RACE-PCR/subcloning on a BALB/c mouse cochlear cDNA library. The predicted protein sequence was 1253 amino acid long and lacks both inserts described in the Myo6 C-terminal domain (Fig. 5A). The full-length Myo6 cDNA was then cloned into the plasmid vectors pCDNA3.1-V5 His (Invitrogen). The cDNAs encoding different Myo6 fragments (Myo6-ct: amino acid 835–1253, Myo6-gt: 981–1253 and Myo6-Δgt: 1–981) were then PCR-amplified and subcloned into pGST parallel 1 [modified from GEX-4T1, glutathione-S-transferase (GST) tag, Amersham Biosciences] and pCDNA3.1-V5 His.
The mouse cDNA encoding otoferlin amplified from the same cDNA library (26) was used as a template to subclone cDNAs encoding otoferlin full-length or truncated forms into pcDNA (No tag, Invitrogen) and pGST parallel 1. All constructs were verified by DNA sequencing.
Protein expression in Escherichia coli
GST-tagged proteins were expressed in Escherichia coli BL21-CodonPlus (DE3)-RIPL (Stratagene). Production was made in 2YT medium (17 g of bacto-tryptone, 10 g bacto-yeast extract and 5 g NaCl per liter) overnight at 14°C, following a 0.1 mm isopropyl-1-thio-beta-d-galactopyranoside induction at optical density OD600nm = 0.6–0.7.
In vitro binding assays
35S-methionine-labeled Myo6 protein fragments were produced by in vitro transcription/translation using pCDNA3.1-V5 His constructs and the TNT-coupled reticulocyte lysate system (Promega). The 35S-labeled products were equally divided between the samples and incubated with GST alone or GST-Otof-Nt immobilized on Glutathione Sepharose 4B beads (Pharmacia) at 4°C on a rotating wheel for 4 h in binding buffer [20 mm Tris–HCl, pH 7.4, 150 mm NaCl, 0.5% Triton X-100, 5 mm β-mecaptoethanol, EDTA-free protease inhibitor cocktail (Roche)], supplemented with 1 mg/ml bovine serum albumin (BSA). The beads were then washed once with this buffer and three times in the binding buffer without BSA. The bound proteins were separated by SDS–PAGE. Levels of GST fusion proteins were assessed by Coomassie blue labeling (Pharmacia Biotech). The gel was then incubated in signal enhancement buffer (Amplify, Amersham Biosciences), dried and processed for autoradiography.
Otoferlin C2A, C2B and C2C domains (amino acid 1–127, 262–408, 423–593, respectively) were purified as GST fusion proteins, treated with Benzonase (Merck) and washed with high-salt buffers as described (54) to remove the bacterial contaminants that stick to the C2 domains and alter their properties (54), before being eluted from the resin using 20 mm glutathione. The purified proteins were dialyzed overnight at 4°C against buffer D (25 mm Tris–HCl, pH 7.4, 300 mm NaCl, 0.1% Triton X-100, 5 mm β-mecaptoethanol) and then ultracentrifugated for 30 min at 100 000g at 4°C. Myo6-ct and Myo6-gt were also purified as GST fusion proteins, then cleaved from their GST tag with rTEV, ultracentrifugated for 30 min at 100 000g at 4°C, before being subjected to gel filtration using a Superdex 200 in 25 mm Tris–HCl, pH 7.4, 300 mm NaCl, 1 mm 1,4-dithio-dl-threitol (DTT). The monodispersity of Myo6 recombinant proteins was attested by dynamic light scattering. About 8 µg of GST and GST-C2 domains were immobilized on Glutathione Sepharose 4B beads equilibrated with buffer D supplemented with EDTA-free protease inhibitor cocktail, 1 mg/ml BSA and either 2 mm EDTA to chelate free Ca2+ ions in solution or 2 mm EDTA plus 3 mm CaCl2. About 8 µg of Myo6-ct or Myo6-gt was then added to each sample and incubated for 5 h. Sepharose beads were then washed three times in the corresponding buffer without BSA before being processed for SDS–PAGE. The proteins electrotransferred to nitrocellulose sheets were stained with 0.2% Ponceau-S in 10% acetic acid before being processed for immunodetection using antibody directed against Myo6.
Cell culture, immunoprecipitation and immunoblotting
HEK-293 cells were cultured in high-glucose Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum, 2 mm of Glutamax-1 (Invitrogen), 50 units/ml penicillin, 50 µg/ml streptomycin. The cells were transfected with otoferlin full length cDNA alone or in combination with Myo6 full length (10 µg total DNA/10 cm dish), using Lipofectamin with Plus reagents (Invitrogen) according to the manufacturer's instruction. After 36 h, cells were rinsed in cold PBS, harvested and pelleted by centrifugation. Cell extracts were prepared using 1% Triton X-100, 0.5% deoxycholate and 0.2% SDS, 5 mm DTT supplemented with 3 mm ATP/MgCl2 and EDTA-free cocktail of protease inhibitors (Roche). Immunoprecipitation was carried out using the polyclonal antibody directed against Myo6 or the anti-otoferlin monoclonal antibody, pre-incubated with protein G-agarose (Pharmacia) for 1.5 h at room temperature. The resin was pelleted by centrifugation, washed twice with 50 mm Tris–HCl, pH 7.4, and 0.1% Triton X-100, and three times in the same buffer complemented with 150 mm NaCl. Proteins were transferred electrophoretically to nitrocellulose sheets. Blots were blocked with 5% non-fat dry milk in Tris-buffered saline containing 0.05% Tween-20 and probed with the following antibodies: anti-Myo6 1:500, anti-otoferlin 1:500, anti-SNAP25 1:2500, anti-syntaxin1 1:2500. Horseradish peroxidase-conjugated goat anti-rabbit or anti-mouse antibodies (BioRad) and the ECL chemiluminescence system (Amersham) were used for detection.
Cochlear tissue preparation and immunohistochemistry for light microscopy were carried out as reported previously (26). Cochlear whole-mount preparations were fixed with 4% paraformaldehyde (PFA) in PBS, permeabilized with 0.3% Triton X-100 in PBS containing 20% normal goat serum for 1 h at room temperature and incubated with the polyclonal antibody directed against Myo6 (1:400), otoferlin (1:500) or CtBP2 (Transduction Laboratories, France) (1:100), overnight at 4°C. After three washes in PBS, the cochleas were incubated for 1 h with F(ab)′2 fragment of goat anti-rabbit IgG antibody conjugated with Alexa488 fluorescein (Interchim, France) alone or with goat anti-mouse IgG antibody conjugated with Cy3 fluorophore (Jackson ImmunoResearch Laboratories) diluted at 1:500 in PBS. Rhodamine phalloidin (1:2000, Invitrogen) was used to label F-actin. The preparations were then washed three times in PBS and finally mounted in one drop of Fluorsave medium (Biochem Laboratories, France).
The samples were analyzed using a confocal laser scanning microscope, LSM510 Meta (Zeiss, Pasteur Institute, Imageople). The images taken with a step size of 0.1 µm were used to generate 3D reconstructions of the IHCs using Osirix software (Antoine Rosset, Department of Radiology, Geneva University Hospital, Switzerland). The total number of ribbons counted was divided by the number of IHCs analyzed. All comparative images between wild-type and mutant mice were done in the same conditions of preparation, acquisition and analysis.
Mouse cochleas were fixed as described (26). The organs of Corti were microdissected and processed by the progressive temperature-lowering technique as reported (26). Ultrathin sections (70 nm) were cut with a Leica Ultracut S microtome and transferred to formvar-coated single-slot grids. Immunogold-labeling was carried out as described (26). The sections were incubated overnight with the anti-Myo6 polyclonal antibodies alone or in combination with the anti-otoferlin monoclonal antibody, diluted at 1:200 and 1:100, respectively. The sections were washed and then incubated for 2 h with 10 nm gold-conjugated goat anti-mouse and 5 nm gold-conjugated goat anti-rabbit antibodies (1:50, Tebu, France). The sections were then stained with uranyl acetate and lead citrate and examined under a Jeol1200EX electron microscope. Gold particle distribution in the synaptic region was quantified compared with the ribbon considered as the center of the active zone. A gold particle was counted as being at the edge when it was associated with the presynaptic density over 100 nm away from the ribbon. A gold particle was counted as being nearby the ribbon when it was not associated with the membrane but located within 50 nm from it. For morphological analyses, cochleas were perfused with 4% PFA and 2% glutaraldehyde in PBS at pH 7.4 and immersed in the fixative solution for 2 h. They were then post-fixed by overnight incubation in 1% osmium tetraoxide at 4°C, dehydrated in graded acetone concentrations and embedded in Spurr's low-viscosity epoxy resin hardened at 70°C. Ultrathin sections were transferred to formvar-coated single-slot grids, stained with uranyl acetate and lead citrate and examined under a Jeol1200EX electron microscope (Pasteur, Imageopole).
Electrophysiological recordings and capacitance tracking
Voltage- and current-clamp recordings were obtained from apical-coil IHCs, using an Optopatch amplifier (Cairn Research). Voltage-clamp experiments were done at room temperature (20–24°C), current-clamp and capacitance tracking close to body temperature (32–35°C). The extracellular solution contained (in mm): 135 NaCl, 5.8 KCl, 1.3 CaCl2, 0.9 MgCl2, 0.7 NaH2PO4, 2 Na pyruvate, 5.6 d-glucose, 10 HEPES–NaOH, amino acids and vitamins for Eagle's MEM. Osmolality was around 308 mOsm/kg and pH 7.5. The pipette filling solution for voltage- and current-clamp experiments was (mm) 131 KCl, 3 MgCl2, 5 Na2ATP, 1 EGTA–KOH, 5 HEPES–KOH (osmolality 282 mOsm/kg, pH 7.3). Currents under voltage clamp were corrected offline for linear leak conductance calculated from 10 mV steps from the holding potential, usually −84 mV, including a −4 mV liquid junction potential correction. Membrane potentials were corrected for voltage drop across the residual series resistance (Rs) of 1.7 ± 0.3 MΩ (n = 12) after compensation of around 80%. Sizes of the K+ currents IK,f and IK,s were determined as reported before (8).
Real-time changes in membrane capacitance (ΔCm) were studied using the Optopatch (6), applying a 2.5 kHz sine wave of 13 or 18.5 mV amplitude around a holding potential of −81 or −91 mV (including −11 mV correction for the liquid junction potential), to avoid activating significant voltage-dependent membrane currents, as accurate capacitance tracking requires a membrane resistance (Rm) that is relatively constant and high compared with Rs (6). For the capacitance-tracking experiments, no Rs compensation could be applied. Nevertheless, no offline correction of membrane potentials for Rs (5.0 ± 0.1 MΩ, n = 18) was necessary, as maximum voltage errors were in the order of 1 to 2 mV. Rm did not differ significantly between the four groups of IHCs (mature versus immature, heterozygote versus mutant) and averaged 803 ± 108 MΩ (n = 18). The extracellular superfusion solution for capacitance tracking and the associated Ca-current recordings contained 30 mm TEA–Cl (Fluka) to help block the large currents through the K+ channels. Some of the apparent inactivation of the inward Ca2+ currents (Fig. 2A and B, left panels) is likely to be due to residual unblocked currents. We therefore measured Ca2+ current size as that of the initial peak current. In the superfusion solution, amino acids and vitamins were omitted and NaCl was reduced to 115 mm to keep osmolality near 308 mOsm/kg. The intracellular solution was (mm) 140 Cs-glutamate, 3 MgCl2, 5 Na2ATP, 0.3 Na2GTP, 1 EGTA–CsOH, 5 Hepes–CsOH (osmolality around 282 mOsm/kg and pH 7.3).
Electrophysiological recordings used for the data analysis were mostly in response to single presentations of the stimulus protocol, but occasionally averaged from two successive presentations. All quantitative data are presented as mean ± SEM. Student's t-test or one-way ANOVA followed by the Tukey post-test was used to analyze statistical significance (criterion P < 0.05).
This work was supported by grants from EMBO ALTF 952–2006 to I.R., the French Ministry of Research and the European Commission FP6 Integrated Project EuroHear LSHG-CT-2004–512063 to C.P., Human Frontiers Science Program and the Agence Nationale de la Recherche (ANR-07-Neuro-036–01) to S.S. and a BBSRC Case studentship sponsored by Pfizer to S.H. and C.J.K.
We thank Dr Jean-Pierre Hardelin for critical reading of the manuscript. We are grateful to Drs Elisabeth Verpy and Michel Leibovici for the cochlear library, and to all staff members of the intramural platforms for excellent technical assistance.
Conflict of Interest statement. None declared.