Myosin XVa localizes to the tips of inner ear sensory cell stereocilia and is essential for staircase formation of the hair bundle

Methods

Construction of the [-N]Myo15a-GFP Expression Vector. A 6.9-kb cDNA encoding mouse myosin XVa (Myo15a) and containing the entire ORF of the short isoform (2,306-aa residues; National Center for Biotechnology Information accession no. AY331133; see Fig. 6, which is published as supporting information on the PNAS web site) was synthesized from 200 ng of mouse inner ear polyA⁺ RNA by using SuperScript II reverse transcriptase (Invitrogen) and specific primers in exon 1 and 66 (primer sequences are provided in Table 1, which is published as supporting information on the PNAS web site). Three overlapping cDNA fragments were synthesized by using Pfu Turbo (Stratagene) and sequenced (see primer sequences in Table 2, which is published as supporting information on the PNAS web site). These cDNA fragments were cloned individually into pGEM-3Zf(–) vector (Promega) and then transformed into XL10-Gold Ultracompetent Cells (Stratagene), excised with appropriate restriction endonucleases (see Supporting Methods, which is published as supporting information on the PNAS web site), gel purified, sequentially ligated into the EcoRI and SalI sites of the pEGFP-C2 vector (Clontech), and sequenced. This cDNA insert corresponds to an endogenously expressed isoform of Myo15a that lacks exon 2 encoding 1,187 amino acids preceding the motor domain (N-terminal extension) and exon 26, which encodes 18 amino acids between the light chain-binding motifs (IQs) and the first MyTh4 domain (Fig. 6). The construct was designated as [-N]Myo15a-GFP. Because mouse inner ear sensory epithelia explants transfected with the isoform containing the N-terminal extension exhibited a low efficiency of transfection, we focused this study on [-N]Myo15a-GFP. The pEGFP-C2 without an insert was used as a control.

Evaluation of the Specificity of Antimyosin XVa Antibodies. To confirm the specificity of the previously described TF1 and PB48 antibodies (Fig. 1A and ref. 28), a colocalization assay was performed by using [-N]Myo15a-GFP transfected COS7 cells, which lack endogenous expression of myosin XVa. The cells were plated on glass coverslips in DMEM media with 10% FBS (Invitrogen) at a density of 4 × 10⁶ cells per cm², cultured until 80% confluency, and transfected by using Lipofectamine 2000 (Invitrogen). After 36 h of incubation at 37°C and 5% CO₂, transfected cells were fixed with 4% paraformaldehyde in PBS for 30 min, permeabilized for 15 min in PBS with 0.5% Triton X-100, blocked with 2% BSA and 5% goat serum in PBS for 30 min, and incubated with TF1 or PB48 antibodies (≈5 μg/ml) for 2 h followed by incubation with Alexa 594 anti-rabbit secondary antibody (Molecular Probes) for 30 min. After several washes in PBS, samples were mounted by using a ProLong Antifade Kit (Molecular Probes).

Immunofluorescence Study of Inner Ear Sensory Epithelia. The cochleae of adult C57BL/6 mice (Charles River Breeding Laboratories), adult (120–150 g) Sprague–Dawley rats (Taconic Farms), adult pigmented guinea pigs (Elm Hill Laboratories, Chelmsford, MA), postnatal day 21 (P21) shaker 2 mice, and C57BL/6 mice ranging from embryonic day 14.5 (E14.5) embryos to P21 pups were fixed with 4% paraformaldehyde in PBS for 2 h. Dissected vestibular and organ of Corti (OC) sensory epithelia were permeabilized in 0.5% Triton X-100 in PBS for 30 min followed by 3 × 10-min washes in PBS. Nonspecific binding sites were blocked by 5% goat serum and 2% BSA in PBS for 1 h at room temperature. Samples were incubated for 2 h with TF1 or PB48 antibodies (≈5 μg/ml) followed by several rinses in PBS and incubation with the anti-rabbit FITC-conjugated secondary antibody (Amersham Pharmacia Biosciences) for 30 min. F-actin was visualized by rhodamine–phalloidin staining (26).

Culture and Transfection of Inner Ear Sensory Epithelia. Inner ear sensory epithelium cultures were prepared from OC, saccule, utricle, and ampule of P3–P5 C57BL/6 mice and from OC of P3 rats. The OC spiral was dissected away from the modiolus in Leibowitz cell culture medium (Invitrogen). The apical and middle turns of the OC were separated from the basal turn and cut into four pieces. The sensory maculae of the utricle, saccule, and ampule were dissociated from the nonsensory tissue, and the otoconia layers were removed. Each piece of sensory epithelium was attached to a glass-bottom Petri dish (MatTek, Ashland, MA) coated with CellTak (Collaborative Biomedical Products, Bedford, MA) and maintained at 37°C and 5% CO₂ in DMEM supplemented with 7% FBS for 1–3 days. Cultures were then transfected by using a Helios gene gun (Bio-Rad). Gold particles of 1.0-μm diameter (Bio-Rad) were coated with plasmid DNA at a ratio of 2 μg of plasmid DNA to 1 mg of gold particles and precipitated onto the inner wall of Tefzel tubing, which was cut into individual cartridges containing ≈1 μg of plasmid DNA. Cartridges containing the pEGFP-C2 vector with [-N]Myo15a insert and control cartridges containing the pEGFP-C2 vector alone were prepared separately. Samples were bombarded with the gold particles from one cartridge per culture by using 120 psi of helium. After an additional 8 h to 4 days in culture, samples were fixed in 4% paraformaldehyde, stained with rhodamine–phalloidin, and observed by using a Zeiss LSM510 confocal microscope equipped with a ×100, N.A. = 1.45 objective.

Results

Antimyosin XVa Antibody Staining Is Colocalized with [-N]Myo15a-GFP Fluorescence at the Tips of Filopodia in COS7 Cells. At 36 h after transfection with the [-N]Myo15a-GFP construct (Fig. 1 A), myosin XVa-GFP was detected in the cell bodies and at the tips of filopodia of COS7 cells (Fig. 1 B–D), which do not express a detectable level of endogenous myosin XVa (Fig. 7E, which is published as supporting information on the PNAS web site). Immunolabeling of transfected COS7 cells revealed that the TF1 antibody staining pattern overlapped with the green fluorescent signal of myosin XVa-GFP (Fig. 1 B–D). PB48 antibody staining produced similar results (Fig. 7). No fluorescent signal was detected at the tips of filopodia in nontransfected or COS7 cells transfected with empty pEGFP-C2 vector with either antibody (Fig. 7).

Myosin XVa Is Localized at the Tips of Stereocilia of Cochlear and Vestibular Hair Cells. In confocal fluorescence images of whole-mount preparations of adult mouse, rat, and guinea pig OCs, we observed myosin XVa immunoreactivity at the top of the stereocilia bundles (Figs. 2 B and C and 4D). In the mouse, OC hair cell stereocilia bundles acquire their mature staircase shape during late embryonic and early postnatal development. By using TF1 antibody, myosin XVa in the developing mouse OC hair cell stereocilia was first detected at E18.5 in the basal turn of the cochlea (Fig. 2G), where both inner hair cell (IHC) and outer hair cell (OHC) hair bundles begin to form a staircase-like pattern. In the middle turn, only the more mature IHC stereocilia were positive for myosin XVa (Fig. 2H). Within the apical turn, no developing hair cell stereocilia exhibited staining (Fig. 2I). Myosin XVa immunoreactivity was present at the tips of stereocilia in the cochlear hair cells during mouse postnatal development (Fig. 2 A) and in the adult (Fig. 4D). The brightest staining was consistently observed at the tips of the tallest stereocilia of both IHCs and OHCs. Myosin XVa-specific staining at the tips of short stereocilia was barely detectable by confocal fluorescence microscopy (Fig. 2 A). However, when the hair cell stereocilia were spread apart, it was possible to see that myosin XVa was indeed localized at the tip of each individual stereocilium (Fig. 2C). Myosin XVa localization extended ≈0.5 μm from the tip downward. Myosin XVa immunofluorescence did not completely overlap with F-actin, as revealed by rhodamine–phalloidin staining (Fig. 2 D–F), although we cannot rule out the possibility that there is a low level of F-actin at the top of the stereocilia core.

A similar distribution of myosin XVa staining was observed at the tips of all stereocilia of the vestibular hair cells of the three ampule (Fig. 3B), the utricle (Fig. 4C), and the saccule (Fig. 3C). The staining of hair cell stereocilia of ampule was detectable as early as E14.5 (Fig. 3A), corresponding to the time in development when stereocilia bundles of vestibular hair cells begin to exhibit a staircase-like arrangement (29). Myosin XVa also continued to be present at the tips of hair cell stereocilia of utricle and saccule from E14.5 onward. Independent of stereocilia length, the area of myosin XVa localization was ≈0.5 μm in length from the top of all stereocilia (Figs. 2 and 3). Both TF1 and PB48 antisera produced a similar pattern of staining in cochlear (Figs. 2 and 4D) and vestibular (Fig. 3 B and C) hair cell stereocilia.

Shortened Hair Cell Stereocilia of Shaker 2 Mice Do Not Possess Myosin XVa at the Tips and Do Not Develop the Staircase Shape of a Mature Hair Bundle. No myosin XVa immunoreactivity was detected at the stereocilia tips of Myo15^sh2/sh2 mice (Fig. 4 A and B). By counterstaining inner ear sensory epithelia from adult Myo15^sh2/sh2 mice with rhodamine–phalloidin, we observed that their abnormally short hair cell stereocilia do not exhibit the characteristic staircase organization of hair bundles, observed in the wild type. This was particularly evident in the vestibular end organs. In the utricle of P21 Myo15^sh2/sh2 mice, all hair cells had stereocilia of nearly equal length within and between the bundles (Fig. 4A).

Targeting of Endogenous Myosin XVa and Myosin XVa-GFP to the Tips of Stereocilia. Hair cell bundles in the OC of P0 wild-type mice already have a well developed staircase pattern (29) and generate a mechanotransduction current of an adult-like amplitude (30). To assess the ability of these hair cells to target myosin XVa to the tips of stereocilia, we transfected hair cells of P4 and P6 mouse inner ear sensory epithelia cultures with [-N]Myo15a-GFP. At 8–96 h after transfection, myosin XVa-GFP protein was concentrated at the tip of every stereocilium of a transfected cell (Fig. 5). From 40 inner ear sensory epithelia explants (four to eight explants per mouse), we observed 106 transfected hair cells. Some vestibular sensory epithelium cultures had up to seven transfected hair cells with slightly different levels of myosin XVa-GFP expression; possibly due to unequal binding of [-N]Myo15a-GFP to the gold bullets (Fig. 5C). Nontransfected cells were used as indicators of the extent to which preservation of the structural integrity of the hair cell bundles was maintained. Among the 40 explants, we observed no systematic differences in bundle morphology between transfected and adjacent non-transfected cells. Hair cells transfected with an empty pEGFP-C2 vector show uniformly distributed green fluorescence along the stereocilia length and in the cell body (data not shown).

A close-up view of the tip of a stereocilium revealed a reproducible difference between the immunostaining of the native myosin XVa and the localization of myosin XVa-GFP, overexpressed in hair cells. Native myosin XVa was localized precisely in the capping region of the stereocilium, between the upper end of the actin core and the apical plasma membrane, slightly overlapping the actin core (Fig. 2 D–F). In contrast, the myosin XVa-GFP appeared to surround the upper portion of the actin core (Fig. 5D). The images collected 24–96 h after transfection revealed that the tip of every stereocilium of transfected hair cells acquired a bulbous shape (Fig. 5 B–E; Fig. 8, which is published as supporting information on the PNAS web site). No patches of myosin XVa were observed along the length of stereocilia (Fig. 5).

Discussion

Our data show that myosin XVa is localized precisely to the tips of mammalian hair cell stereocilia, under the apical plasma membrane and overlapping with the barbed ends of the actin filaments of the stereocilia core (Figs. 2 and 3). This is the location of growth and remodeling of the actin core (6). We first detect myosin XVa at the tips of mouse E14.5 vestibular and E18.5 cochlear hair cell stereocilia, coinciding with the onset of the formation of a characteristic staircase arrangement of the stereocilia in the hair bundle. In homozygous shaker 2 mice (Myo15^sh2/sh2), no myosin XVa was detected at the tips of the inner ear hair cell stereocilia, and bundle elongation seems to have been arrested at the stage when stereocilia are of approximately equal height. This finding is consistent with the timing of the appearance of myosin XVa within the tips of the stereocilia in the wild-type mouse (Fig. 3A). Taken together, these data suggest that myosin XVa is necessary for the differential growth of the stereocilia to form their staircase relationship.

Targeting and Tethering of Myosin XVa to the Tips of Stereocilia. Colocalization of myosin XVa-GFP with antimyosin XVa antibody staining in transfected COS7 cells not only confirmed the specificity of the two antibodies (TF1 and PB48) but also revealed targeting of myosin XVa to the tips of filopodia (Fig. 1 B–D). This is similar to the localization of unconventional myosin X in transfected HeLa cells (31, 32). In mammalian inner ear hair cells, myosin XVa seems to be the only molecular motor thus far shown to be concentrated exclusively at the tip of a stereocilium, which is one of the proposed locations of the hair cell transduction machinery.

The observation that myosin XVa only partially colocalizes with the most distal detectable rhodamine–phalloidin staining of filamentous actin at the tips of the stereocilia (Fig. 2D) suggests that domains of myosin XVa, other than the actin-binding site of the motor domain, may be important for its localization at the apex of stereocilia. This possibility is also supported by the observation of an apparently identical phenotype of Myo15^sh2/sh2 and Myo15^sh2j/sh2j mice (27), which results from a missense mutation in the motor domain and a deletion of the last six exons encoding the C-terminal 268 amino acids of the myosin XVa tail, respectively (27). In mice homozygous for either of the mutant shaker 2 alleles, myosin XVa is not present in the stereocilia but is localized ectopically under the cuticular plate in the hair cell body, suggesting that an intact motor and tail are both required for correct stereocilia tip localization of myosin XVa.

Previously elucidated structures and protein partners of other myosins (33–35) suggest the possibility that some myosin XVa tail domains could be important for targeting and tethering at the tips of stereocilia. The tail of myosin XVa contains an SH3 domain, which in myosin I was implicated in directed polymerization of actin (33, 36) and two pairs of tandemly arranged MyTh4 and FERM domains (Fig. 1 A). MyTh4 and FERM domains are also similarly grouped in the tail of myosins VIIA and X (31, 33), which were also found in actin-rich protrusions such as microvilli, filopodia, and stereocilia (21, 31). Moreover, both the long and the short isoforms of myosin XVa (Fig. 6) have predicted class 1 consensus PDZ-binding ligands (X-S/T-X-V/L) that are formed by the last four amino acids at the C terminus (I-T-L-L), suggesting the possible interaction of myosin XVa with a PDZ-containing protein such as whirlin (37) and in a manner similar to the interaction of myosin VIIA with harmonin in stereocilia (23).

Mutations of the motor and of various regions of the myosin XVa tail, including the MyTh4 and FERM domains, cause deafness in humans and mice (24, 25, 38). To date, no mutations that cause hearing loss have been identified in the alternatively spliced exon 2, which encodes ≈1,200-aa residues of an N-terminal extension preceding the motor domain. The exact function of the long proline/tyrosine rich N-terminal extension is currently unknown. Our transfection experiments using [-N]Myo15a-GFP show that the N-terminal extension of myosin XVa is not essential for this molecular motor to be specifically targeted and tethered to the tips of stereocilia.

Accumulation of Myosin XVa-GFP at the Tips of Stereocilia. When GFP-tagged β-actin monomers are added to the barbed apical ends of actin filaments, the existing actin core shifts toward the base of stereocilia. After 48 h, β-actin of the stereocilia core seems to be replaced by β-actin-GFP, and stereocilia become fluorescent along their length (6). In contrast, myosin XVa-GFP appears only at the tips of the stereocilia and remains concentrated there for at least 96 h after transfection (Fig. 5). Excessive accumulation of newly synthesized myosin XVa-GFP at the upper end of a stereocilium causes distension of the tip (Figs. 5 B–E and 8). We presume that in transiently transfected hair cells, the cytomegalovirus promoter driving [-N]Myo15a-GFP expression results in an abnormally high concentration of myosin XVa-GFP at the tips of stereocilia. In addition or alternatively, the GFP epitope tag may interfere with a steady-state recycling process that prevents excessive accumulation of endogenous myosin XVa at the tip of each stereocilium. Whatever the reason, this excessive accumulation of myosin XVa does not confound the demonstration by our data that native and GFP-tagged myosin XVa are both targeted and confined to the tip of each stereocilium. There appears to be only a slight overlap of endogenous myosin XVa with the barbed ends of actin filaments (Fig. 2 D–F), which may mean that a significant fraction of myosin XVa at the tip of a stereocilium is detached from the actin core and may not be entrained by the β-actin treadmill (32).

Myosin XVa and Mechanoelectrical Transduction. In our experiments, myosin XVa was not detected by confocal immunofluorescence microscopy near the insertion sites of the tip-link on the side of the stereocilium (Fig. 2 D–F), which is a proposed location of the adaptation motor (39). However, myosin XVa at the tip of the shorter stereocilium may influence the tension applied to the tip-link and, therefore, participate in the process of adaptation. Also, tip-links appear to be absent from stereocilia of Myo15^sh2/sh2 mice (26), suggesting a possible requirement of myosin XVa for tip-link assembly and/or maintenance. The molecular identities of the tip-link, the mammalian hair cell mechanotransduction channel, and the additional elastic elements involved in the gating of this channel are still unknown. Some of these molecules may specifically bind to myosin XVa and may be transported by this motor to the tips of stereocilia, as was suggested for myosin molecular motors (11, 15).

Myosin XVa and the Formation of Hair Bundle Staircase Shape. The graded height of stereocilia in a bundle is a characteristic feature of auditory hair cells of the modern amniotes including reptiles, birds, and mammals (40). The structural integrity of this arrangement appears to be essential for auditory transduction, and the directional sensitivity of mechanotransduction has been shown to be associated with the gradation in height of the stereocilia (41). When the formation of a characteristic staircase pattern of the bundle is first evident, we are able to detect myosin XVa at the tips of stereocilia. In mouse, this occurs as early as E18.5 in cochlear hair cells and at E14.5 in the vestibular end organs (Figs. 2 G–I and 3A). The absence of functional myosin XVa at the tips of stereocilia in Myo15^sh2/sh2 mice results in a failure of stereocilia bundles to form the expected staircase arrangement (Fig. 4 A and B), although there is no noticeable influence on the previous stages of bundle development. The stereocilia of Myo15^sh2/sh2 mice are spatially well organized but short and of approximately equal height within the bundle. In summary, myosin XVa is localized precisely to the tips of inner ear hair cell stereocilia and is critical for their graded elongation during formation of the staircase arrangement of mammalian hair bundles.

Supplementary Material