Evidence for multiple, developmentally regulated isoforms of Ptprq on hair cells of the inner ear

Abstract

Ptprq is a receptor-like inositol lipid phosphatase associated with the shaft connectors of hair bundles. Three lines of evidence suggest Ptprq is a chondroitin sulfate proteoglycan: (1) chondroitinase ABC treatment causes a loss of the ruthenium-red reactive, electron-dense particles associated with shaft connectors, (2) chondroitinase ABC causes an increase in the electrophoretic mobility of Ptprq, and (3) hair bundles in the developing inner ear of wild-type mice, but not those of Ptprq^−/− mice, react with monoclonal antibody (mAb) 473-HD, an IgM that recognizes the dermatan-sulfate-dependent epitope DSD1. Two lines of evidence indicate that there may be multiple isoforms of Ptprq expressed in hair bundles. First, although Ptprq is expressed throughout the lifetime of most hair cells, hair bundles in the mouse and chick inner ear only express the DSD1 epitope transiently during development. Second, mAb H10, a novel mAb that recognizes an epitope common to several avian inner-ear proteins including Ptprq, only stains mature hair bundles in the extrastriolar regions of the vestibular maculae. MAb H10 does not stain mature hair bundles in the striolar regions of the maculae or in the basilar papilla, nor does it stain immature hair bundles in any organ. Three distinct, developmentally regulated isoforms of Ptprq may therefore be expressed on hair bundles of the chick inner ear. Hair bundles in the mature chick ear that do not express the H10 epitope have longer shaft connectors than those that do, indicating the presence or absence of the H10 epitope on Ptprq may modulate the spacing of stereocilia. © 2010 Wiley Periodicals, Inc. Develop Neurobiol 71: 129-141, 2011

INTRODUCTION

Our senses of hearing and balance depend on hair cells' mechanosensory receptors located in the auditory and vestibular end organs of the inner ear. Each hair cell has a hair bundle, an ensemble of height-ranked stereocilia and usually a single kinocilium, located at its apical pole. Deflections of the hair bundle caused by sounds or head movements modulate the gating of transducer channels located at the tips of the stereocilia, thereby generating a receptor potential in the hair cell. These receptor potentials then drive the release of transmitter from ribbon synapses situated around the basolateral pole of the hair cell. This activates the afferent nerve fibers innervating the hair cell, and information about the nature of the stimulus is thereby relayed to the central nervous system.

A number of distinct cell-surface specializations known as links or connectors are associated with the elements comprising the stereociliary bundles of inner-ear hair cells (Hillman, 1969; Bagger-Sjoback and Wersall, 1973; Hirokawa and Tilney, 1982; Neugebauer and Thurm, 1984; Pickles et al., 1984; Furness and Hackney, 1985; Ernstson and Smith, 1986; Csukas et al., 1987; Jacobs and Hudspeth, 1990; Tsuprun and Santi, 1998). In the avian inner ear, up to five types of link can be distinguished on hair bundles: kinocilial links, oblique tip links, horizontal top connectors, shaft connectors, and ankle links (Goodyear and Richardson, 1992; Goodyear and Richardson, 2003). The oblique tip links are considered to gate the hair cell's elusive mechanotransducer channel (Pickles et al., 1984), whereas the other links are generally thought to maintain the structural integrity of the hair bundle. Many of the molecules that are associated with or form these different connectors have been identified over the last 10 years. Protocadherin 15 (originally identified as the tip-link antigen, TLA, in the bird ear) and cadherin 23 interact to form the tip and kinocilial links (Goodyear and Richardson, 2003; Siemens et al., 2004; Sollner et al., 2004; Ahmed et al., 2006; Kazmierczak et al., 2007), the receptor-like inositol lipid phosphatase Ptprq (previously the hair-cell antigen, HCA) is a component of the shaft connectors (Richardson et al., 1990; Goodyear et al., 2003), and the very large G-protein coupled receptor Vlgr1 (originally identified as the ankle link antigen, ALA, in the chick ear) and usherin are associated with the ankle links (Goodyear and Richardson, 1999; Adato et al., 2005; McGee et al., 2006; Michalski et al., 2007). The proteins that form the horizontal top connectors observed in the hair cells of birds, frogs, and fishes have yet to be identified, but a protein known as stereocilin, a product of the DFNB16 deafness locus, may be a component of these structures in the outer hair cells of the mouse cochlea (Verpy et al., 2008).

The HCA was originally identified with a monoclonal antibody (mAb) as a large, nonionic detergent-soluble protein that was associated with the shaft connectors of sensory hair bundles in the avian inner ear (Richardson et al., 1990; Goodyear and Richardson, 1992). Subsequent studies revealed the HCA was Ptprq (Goodyear et al., 2003), a type III receptor-like phosphatase cloned from a transcript that was up-regulated in kidney mesangial cells in response to experimentally induced glomerular nephritis (Wright et al., 1998). Although originally predicted to be a receptor-like protein tyrosine phosphatase (RPTP), in vitro studies have shown the intracellular, catalytic domain of Ptprq has low activity against phosphorylated tyrosine residues but can dephosphorylate a variety of inositol phospholipids, catalyzing the removal of phosphate groups from the 3′ and 5′ positions in the inositol ring (Oganesian et al., 2003). The over-expression of the intracellular domain of Ptprq in cultured cells inhibits proliferation and stimulates apoptosis (Oganesian et al., 2003). In mice that are effective null mutants for Ptprq, hair bundles in the inner ear fail to mature properly and the hair cells in the basal, high-frequency end of the cochlea eventually die (Goodyear et al., 2003). The loss of Ptprq typically leads to the loss and fusion of stereocilia in cochlear inner hair cells, and it has been suggested that shaft connectors may function as spacers, elements that prevent the close apposition and fusion of stereocilia, rather than acting as cohesive structures (Goodyear et al., 2003). Recent studies have revealed that mutations in PTPRQ underlie recessive, autosomal, nonsyndromic hearing loss at the DFNB84 locus, and that such mutations are, in some individuals, associated with vestibular dysfunction (Schraders et al., 2010; Shahin et al., 2010).

The distribution of shaft connectors varies according to hair-cell type in the avian ear (Goodyear and Richardson, 1992). Hair bundles in the extrastriolar regions of the maculae and the peripheral regions of the cristae have shaft connectors and Ptprq distributed fairly uniformly over their surface, whereas those in the striolar regions of the maculae, central regions of the cristae and the auditory papilla have shaft connectors and Ptprq concentrated in the lower, basal region of the bundle. In the mouse inner ear, Ptprq is only expressed transiently on the hair bundles of basal-coil outer hair cells, during a period from E17.5 through to ∼P15 (Goodyear et al., 2003). Inner hair cells, apical-coil outer hair cells, and vestibular hair cells in the mouse inner ear continue to express Ptprq throughout life, as do all hair cells in the bird ear. On inner hair cells of the mature cochlea, Ptprq expression is restricted to the basal region of the hair bundle. This distribution has been suggested to be dependent on the activity of a minus end-directed actin-based motor, myosin VI (Sakaguchi et al., 2008).

In cells that have been fixed in the presence of tannic acid, shaft connectors form a dense meshwork of fine filaments interspersed between the stereocilia. In preparations that have been fixed in the presence of ruthenium red, shaft connectors appear as electron dense particles suspended between the membranes of adjacent stereocilia by a number of fine strands (Goodyear and Richardson, 1992). Reactivity with ruthenium red suggests that proteoglycan (PG) is associated with these connectors, and the electron dense particles that are observed may result from the cation-induced collapse of their negatively charged glycosaminoglycan chains. In this study, we present evidence indicating that Ptprq is a chondroitin sulfate (CS) PG. Furthermore, we show that there are several glycosylation variants of Ptprq expressed in sensory hair cells, one of which is expressed transiently during the development of both mouse and chick hair cells and carries the dermatan sulfate-dependent epitope DSD1, an epitope common to the protein-tyrosine phosphatase receptor type Z (Ptprz, also known as PTPζ or RPTPβ). Ptprz is preferentially expressed in the brain as a major CS PG in which three splice variants encoding two receptor isoforms and one secretory isoform are known (Maeda et al., 1994; Chow et al., 2008a, b). The dermatan sulfate-dependent epitope DSD1 was originally identified on DSD-1-PG, a CSPG that is involved in axon outgrowth in the developing brain (Faissner et al., 1994) and was later revealed to be the secretory isoform of Ptprz (Garwood et al., 1999).

MATERIALS AND METHODS

Animals

One-day old chicks were obtained from Joice and Hill Poultry (Peterborough, UK) and housed in accordance with U.K. Home Office regulations. Fertile chicken eggs were obtained form Henry Stewart, Lincolnshire. Mice were obtained from in-house breeding colonies. All animal procedures were performed in accordance with U.K. Home Office guidelines and with the approval of the local ethical committee.

Antibodies

Rat mAb 473-HD was obtained from Santa Cruz. Rabbit antiserum to Ptprz (anti-Ptprz-S) and mouse mAb D10 recognizing chick Ptprq were prepared as described previously (Richardson et al., 1990; Goodyear et al., 2003; Chow et al., 2008a, b). Mouse mAb H10 was obtained from a mouse that was immunized with a mixture of formaldehyde-fixed shark inner ear tissue and unfixed membranes from chick sensory organs. The shark inner ear tissues were collected with the help and advice of Dr. Jeffrey Corwin from the discarded heads of specimens caught during a fishing competition run out of Martha's Vineyard, MA, USA in the summer of 2006. Hybridomas were prepared as described previously (Richardson et al., 1990; Goodyear and Richardson, 1999) and screened on cryosections of the chick inner ear, selecting for clones secreting antibodies that stained hair bundles. MAb H10 was cloned three times by limiting dilution and isotyped as an IgM class antibody using isotype specific anti-antibodies. Rabbit antibodies to Ptprq were raised to a fragment of the recombinant, intracellular domain of the protein. In brief, an 883-base pair DNA fragment was amplified from a chick Ptprq clone using primers 8F1 (5′-ATGCTCGAGGATTTTGAAGACCTTGCT-3′) and 8R1 (5′-AGCAGCCGGATCCCATGGTAGTTTCTTCCCA-3′) and subcloned into the pET15b expression vector. His-tagged protein was expressed in E. coli BL21 codon plus cells, purified on a Ni+ column, dialyzed into water, and lyophilized. Serum from a rabbit (R27) that was immunized with this protein was affinity purified on a column to which the recombinant protein was covalently crosslinked.

Chondroitinase ABC Treatment of Inner Ear Tissues

Utricular maculae were dissected from 2-day posthatch chicks in Hepes-buffered (10 mM, pH 7.2) Hanks' balanced salt solution (HBHBSS), and the otoconial membranes were gently removed with fine forceps. Mouse cochlear cultures were prepared from the inner ears of 1–2-day postnatal CD1 mice on collagen-coated, round glass coverslips as described previously (Russell and Richardson, 1987). Chick maculae and mouse cochlear cultures were transferred to HBHBSS (as a control) or HBHBSS containing chondroitinase ABC at 1.0 mg/mL or 0.1 mg/mL (for chick maculae and mouse cochlear cultures, respectively) and incubated at 37°C for 1 h in a humid environment. Samples were then washed twice in HBHBSS and fixed for either immunofluorescence or electron microscopy. For immunofluorescence microscopy, tissues were fixed in 3.7% formaldehyde in 0.1M sodium phosphate buffer pH 7.4 for 1 h. For electron microscopy, samples were fixed in 2.5% glutaraldehyde in 0.1M sodium cacodylate buffer containing 0.5% ruthenium red for 2 h, washed three times in buffer and postfixed for 1 h in 1% osmium tetroxide. After a brief wash in buffer, samples were dehydrated through a series of ascending concentrations of ethanol, equilibrated with propylene oxide, infiltrated with and imbedded in Epoxy resin. Resin was polymerized at 60°C for 2 days. Thin sections were cut from the polymerized blocks with glass and diamond knives, mounted on copper mesh grids, double stained with uranyl acetate and lead citrate, and viewed in a Hitachi 7100 microscope operating at 100 kV. Images were captured with a Gatan digital camera.

Immunofluorescence Microscopy

Inner ear tissues were obtained from wild-type mice (CD1 and C57Bl/6J), Ptprq CAT-KO mice, and Ptprz null mutant mice. Ptprq-CAT-KO mice were the progeny of matings set up between heterozygous and homozygous Ptprq-CAT-KO mice on a mixed (50:50) C57BL/6J:129/Sv background. The Ptprq-CAT-KO mice have a deletion of two exons encoding a region of the cytoplasmic domain that includes the catalytic site and are functional nulls (Goodyear et al., 2003). They were originally produced by, and obtained from, Drs. Dan Bowen-Pope and Ron Seifert (University of Seattle, Washington). Inner ears from wild-type CD1 mice and Ptprq-CAT-KO mice were fixed by immersion in 3.7% formaldehyde in 0.1M sodium phosphate buffer pH 7.4 for 1 h at room temperature and then washed three times in PBS.

Wild-type and Ptprz^−/− mouse pups (Shintani et al., 1998) on a C57BL/6 background were fixed by perfusion with 4% paraformaldehyde in PBS pH 7.4, and the heads were immersion fixed for a further 24 h at 4°C. The heads were washed three times in Tris-buffered saline (TBS, 10 mM Tris-HCl pH 7.2, 150 mM NaCl) and the inner ear tissues were subsequently removed by dissection in PBS.

For cryosectioning, tissues were decalcified in 0.5M EDTA, equilibrated with 30% sucrose in PBS and imbedded in 1% low gelling temperature agarose in 18% sucrose. Sections of 10 μm thickness were cut at −30°C in a cryostat and mounted on gelatin-coated slides. Slides were preblocked in TBS containing 10% heat inactivated horse serum (TBS/HS). Wholemounts were preblocked in TBS/HS containing 0.1% Triton X-100 (TX-100). Wholemount preparations and cryosections were stained overnight with primary antibodies diluted in preblock, washed with TBS, and labeled with the appropriate secondary antibodies (Alexa 555 donkey anti-rabbit Ig, Alexa 555 goat anti-mouse IgG₁, and FITC goat anti-rat IgM) containing 1:300 Alexa 647 phalloidin for 2 h. Following washing in TBS, samples were mounted in Vectashield and imaged with a Zeiss LSM 510 Meta confocal microscope.

Immunoprecipitations

Sensory organs were dissected from the inner ears of early posthatch chicks (1–4 days after hatching) in cold PBS containing a cocktail of protease inhibitors (1 mM PMSF, 2 mM benzamidine, 1 μg/mL leupeptin, and 1 μg/mL pepstatin) and frozen to −80°C. Frozen tissues were thawed in ∼1.0 mL cold TBS containing 1% TX-100 and the same cocktail of inhibitors, were homogenized, and were centrifuged at 14,000 rpm for 10 min in a cold microcentrifuge. The supernatant was collected and 10 μg of affinity purified antibody directed against the intracellular domain of chick Ptprq was added, followed by 20–50 μL of a 1:1 slurry of protein G-Sepharose that had been preblocked in TBS containing 25 mg/mL BSA. After overnight mixing at 4°C, the protein-G Sepharose beads were washed three times with TBS containing 0.1% TX-100, divided into aliquots, washed 1x with the appropriate digestion buffer, and incubated in either buffer alone or buffer containing chondroitinase ABC (1 mg/mL), endo-β-galactosidase (5 U/mL) or heparinase I (1 mg/mL) for 1 h at 37°C. Beads were washed 1x with TBS/0.1% TX-100 and were eluted by heating to 100°C for 4 min in 10 μL of 2x concentrated SDS-PAGE sample buffer. Eluted samples were separated on 5% polyacrylamide gels and transferred to PVDF using semidry blotting. PVDF membranes were preblocked in 3% low-fat dried milk powder in TBS/0.05% Tween-20 and incubated with affinity purified rabbit anti-Ptprq or mAb H10 supernatant overnight. Bound antibodies were detected with alkaline phosphatase-conjugated goat anti-rabbit Ig (Dako) or alkaline phosphatase goat anti-mouse Ig (Dako).

Immunogold Labeling

Utricular maculae from 2 day-old posthatch chicks were collected and fixed in phosphate buffered 3.7% formaldehyde as described above, washed in TBS, preblocked in 10% horse serum for 1 h, and incubated overnight in mAb H10 supernatant diluted 1:5 in TBS/10% horse serum. Following washing in TBS, maculae were incubated with rotation for 2 days at 4°C in a 1:10 dilution of 10 nm colloidal gold-conjugated goat anti-mouse IgM in TBS/0.05% Tween with 1 mM EDTA, washed extensively in PBS, refixed in 2.5% glutaraldehyde in 0.1M sodium cacodylate pH 7.2 containing 0.5% ruthenium red for 2 h and further processed for electron microscopy as described above.

RESULTS

Shaft Connectors and Ptprq Are Chondroitinase Sensitive

The structure of shaft connectors in hair bundles of the early posthatch chick utricle that were incubated in saline for 1 h at 37°C before fixation for electron microscopy is shown in Figure 1(a). As described previously for samples fixed in the presence of ruthenium red (Goodyear and Richardson, 1992), shaft connectors are comprised of dense particles that are located between the membranes of the stereocilia and connected to these membranes by a number of fine strands. In grazing, en-face profiles of the stereocilia, these particles [arrowheads in Fig. 1(a)] appear much darker. In samples that were incubated in saline containing 1 mg/mL chondroitinase ABC for 1 h at 37°C before fixation in the presence of ruthenium red, the dense particles are no longer visible, but the fine strands interlinking the stereocilia are still present [Fig. 1(b)]. Immunoblots reveal Ptprq undergoes a small, but distinct, shift in electrophoretic mobility following treatment with chondroitinase ABC [Fig. 1(c)]. In some experiments, a concomitant increase in the densitometric intensity of the immunoreactive band was also noted but this increase, when observed, was small (<50% of control). A shift in mobility of Ptprq is not observed following treatment with either endo-β-galactosidase (keratanase) or heparinase I [Fig. 1(c)]. Chondroitinase ABC treatment does not reduce or abolish the immunostaining observed with mAb D10, a mAb that recognizes an epitope located in the subtilisin-sensitive ectodomain of Ptprq [Fig. 1(d,e)]. The loss of the dense particles observed by electron microscopy is therefore unlikely to be due to contaminating proteolytic activity. Together these results indicate that Ptprq is a CS PG, although the shift in mobility observed following chondroitinase treatment is not large and the number and/or length of the associated CS GAG chains are likely to be small.

Figure 1. Chondroitinase ABC removes shaft-connector associated dense particles from chick hair bundles and causes a shift in the electrophoretic mobility of Ptprq. (a, b) Electron micrographs from control (a) and chondroitinase ABC treated (b) extrastriolar hair bundles of the early posthatch chick utricle. Arrowheads indicate the electron dense particles associated with the shaft connectors, and arrows indicate residual shaft connector filaments observed following chondroitinase ABC treatment. Filaments (AL) in the tapered region shown in panel b are ankle links. (c) Western blots of Ptprq immunoprecipitates treated with buffer alone (Control), or buffer containing either chondroitinase ABC, heparinase or keratanase. Samples were immunoprecipitated from Triton X-100 soluble fractions of the early posthatch chick utricle with mAb D10, digested, and immunoblotted with a rabbit antibody raised to the recombinant intracellular domain of Ptprq. (d, e) Confocal images of control (d) and chondroitinase ABC treated (e) hair bundles from the early posthatch chick utricle immunolabeled with mAb D10. Bars = 200 nm (a,b), 5 μm (d, e).

A dense cell coat is a transient feature of developing mouse cochlear outer hair cells and its loss in basal-coil outer hair cells correlates with the loss of Ptprq immunoreactivity from the hair bundle (Goodyear et al., 2003, 2005). Particles are also seen to be associated with the surface of early postnatal, mouse cochlear hair bundles that have been fixed in the presence of ruthenium red [Fig. 2(a)]. These particles are no longer observed following treatment with 0.1 mg/mL chondroitinase for 1 h at 37°C [Fig. 2(b)]. mAb 473-HD, a rat IgM that recognizes the dermatan sulfate-dependent epitope DSD1 of the RPTP Ptprz (Garwood et al., 1999), reacts intensely with hair bundles in the early postnatal cochlea [Fig. 2(c)]. This immunoreactivity with mAb 473-HD is abolished by treatment with 0.1 mg/mL chondroitinase ABC for 1 h at 37°C [Fig. 2(d)]. A similar loss of mAb 473-HD immunoreactivity was also seen in cochlear cultures that had been treated with chondroitinase ABC at lower concentrations (10 and 1.0 μg/mL, data not shown).

Figure 2. Crondroitinase ABC causes a loss of shaft-connector densities and the DSD1 epitope from early potsnatal mouse cochlear hair bundles. (a, b) Transmission electron micrographs of hair bundles from apical-coil mouse cochlear cultures that were incubated in buffer (a) or buffer containing chondroitinase ABC (b) before fixation in the presence of ruthenium red. Arrows indicate particles associated with the shaft connectors. (c, d) Confocal images from the middle of apical-coil mouse cochlear cultures that were incubated in buffer (c) or buffer containing chondroitinase ABC (d) before fixation and double labeling with mAb 473-HD (c, d) and Texas Red conjugated phalloidin (c′, d′). I = inner hair cells, O1, O2, and O3 = outer hair cells in rows 1, 2, and 3, respectively. Bars = 200 nm (a, b), 10 μm (c, d).

The DSD1 Epitope Is Absent from Hair Bundles of Ptprq^−/− Mice

Hair-bundle staining with mAb 473-HD is still observed in the cochleae of Ptprz null mutant mice [Fig. 3(a,b)] and therefore cannot be due to the presence of this CS-containing RPTP. No obvious hair-bundle defects were observed in the cochleae of the Ptprz null mutant mice [Fig. 3(a,b)], and the hair bundles in wild-type mice were not stained with the anti-Ptprz-S antibody that is specific to the extracellular region of Ptprz (not shown). Double immunofluorescence labeling reveals the distribution of the DSD-1 epitope in the early postnatal mouse cochlea is remarkably similar to that of Ptprq, as judged by an antibody (R27) raised to the intracellular domain of this phosphatase [Fig. 3(a′,b′)]. Neither mAb 473-HD nor antibody R27 stains the hair bundles of mice that are homozygous for the deletion of the Ptprq catalytic domain (Ptprq^−/− mice) [Fig. 3(c,c′)]. These observations therefore suggest that the DSD1 epitope recognized by mAb 473-HD is associated with Ptprq.

Figure 3. The DSD1 epitope is present in the hair bundles of the Ptprz^−/− mouse but absent from those of the Ptprq^−/− mouse. Confocal images from the basal-coil mouse cochlear wholemounts from (a) a Ptprz knockout mouse, (b) a wild-type mouse, and (c) a Ptprq^−/− mouse double labeled with mAb 473-HD (a–c) and with a rabbit antibody to the intracellular domain of Ptprq (a′–c′). I = inner hair cells, O1, O2, and O3 = outer hair cells in rows 1, 2, and 3, respectively. Bar = 10 μm.

Developmental Expression Patterns of Ptprq and DSD1 Are Distinct

The temporal expression patterns of DSD1 and Ptprq in the mouse and chick inner are, however, distinct. In the basal-coil of the developing mouse cochlea, hair-bundle staining with mAb 473-HD can be first detected in inner hair cells at E16.5 [Fig. 4(a)] and with R27 (anti-Ptprq) at E18.5 [Fig. 4(b′)]. By P3, staining with both antibodies is observed in inner and outer hair cells throughout most of the length of the cochlea [Fig. 4(c,c′)]. Between P3 and P9, mAb 473-HD immunoreactivity in basal-coil hair bundles diminishes to undetectable levels, but R27 staining remains roughly constant [Fig. 4(d,d′)]. Hair bundles of apical-coil outer hair cells and inner hair cells present throughout the length of the cochlea lose immunoreactivity with mAb 473-HD during this period of development, but Ptprq continues to be expressed in the hair bundles of both of these types of hair cell into maturity. In the vestibular system, staining with mAb 473-HD and R27 is first observed at E17. Staining with mAb 473-HD declines in intensity between P9 and P16 and is no longer detectable by P100 (not shown).

Figure 4. Temporal expression patterns of the DSD1 epitope and Ptprq in mouse cochlear hair bundles are distinct but overlap. (a–d) Confocal images from basal-coil mouse cochlear wholemounts at E16.5 (a, a′), E18.5 (b, b′), P3 (c, c′) and P9 (d, d′) double labeled with mAb 473-HD (a–d) and rabbit antibody to the recombinant intracellular domain of Ptprq (a′–d′). Hair-bundle labeling with mAb 473-HD is detected at E16.5 (a), becomes intense by P3 (c), and diminishes to nearly undetectable levels by P9 (d). Hair-bundle labeling with rabbit anti-Ptprq is detected at E18.5 (b′) and is still observed in both inner and outer hair cells at P9 (d′). The ratios of 473-HD reactive hair bundles and R27 reactive hair bundles (473-HD:R27) present in the panels shown for each stage are as follows: E16.5, 8:0; E18.5 6:6; P3 7:7; P9 0:7 (Inner hair cells); E16.5 0:0; E18.5 16:7; P3 24:24; P9 0:22 (Outer hair cells). I = inner hair cell, IP = inner pillar cell, OP = outer pillar cell, DC = Dieters' cell. Scale bar = 10 μm.

Immunoreactivity to mAb 473-HD is also only seen transiently during the development of the basilar papilla, the auditory end organ of the bird inner ear. mAb 473-HD staining is first detected in the distal end of the papilla at E7 (not shown). At E8 many hair bundles are reactive with mAb 473-HD, but only a proportion of these are labeled by R27 [Fig. 5(a)]. By E10 most hair bundles react with both mAb 473-HD and R27 [Fig. 5(b)], but reactivity to mAb 473-HD subsequently declines [Fig. 5(c)]. By E14, all hair bundles present are labeled by R27, but only a proportion is also labeled by mAb 473-HD [Fig. 5(c)]. Expression of DSD-1 diminishes to nearly undetectable levels in all auditory hair bundles by the early posthatch stage of development [Fig. 5(d)].

Figure 5. Temporal expression patterns of the DSD1 epitope and Ptprq in auditory and vestibular hair bundles of the chick inner ear are distinct but overlap. Confocal images from wholemount preparations of the basilar papilla (a–d) and utricle (e–h) of the chick inner ear at E8 (a, e), E10 (b, f), E14 (c, g), and early after hatching (d, h) double labeled with mAb 473-HD (green, a–h) and rabbit antibody (R27) to the recombinant intracellular domain of Ptprq (red, a′–c′, e′–h′). At E8 (a, e), more hair bundles are labeled by mAb 473-HD than by anti-Ptprq. At E14 (c, g), more hair bundles are labeled by anti-Ptprq than by mAb 473-HD. None of the hair bundles in the basilar papilla are labeled with mAb 473-HD by the early posthatch stage of development (d). In the posthatch utricle, the small hair bundles of immature hair cells are labeled (h, arrows). In panels a and e, arrowheads indicate examples of hair bundles double labeled by both antibodies, arrows indicate examples of those that are only labeled by mAb 473-HD. In panels c and g, arrowheads also indicate examples of hair bundles double labeled by both antibodies, whereas arrows indicate examples of hair bundles labeled by anti-Ptprq but not by mAb 473-HD. The ratios of 473-HD reactive hair bundles and R27 reactive hair bundles (473-HD:R27) present in the panels shown for each stage are as follows: (Basilar papilla) E8, 106:27; E10, 56:54; E14, 30:38; PH 0:34; (Utricle) E8, 36:21; E10, 53:51; E14, 22:39; PH 15:46. Bar = 10 μm.

A similar developmental expression profile is observed in the developing utricle of the chick inner ear [Fig. 5(e–h)]. At E8, mAb 473-HD recognizes more hair bundles than R27 [Fig. 5(e)]. At E10, approximately similar numbers of hair bundles are detected by both antibodies [Fig. 5(f)], but by E14 more are labeled by R27 [Fig. 5(g)]. In the early posthatch chick utricle, mAb 473-HD immunoreactivity is only observed in small immature hair bundles and in a small subset of mature hair bundles where the distribution of DSD1 within the hair bundle is more restricted than that of Ptprq [Fig. 5(h)]. Assuming that mAb 473-HD does indeed recognize Ptprq, these observations suggest that there is a developmentally regulated glycosylation variant of Ptprq that carries the DSD1 epitope and is expressed at high levels in immature hair bundles.

mAb H10 Recognizes an Extrastriolar Variant of Ptprq

mAb H10 provides evidence that there may be an additional variant of Ptprq expressed by mature avian hair cells. mAb H10 is a mouse IgM that stains avian, but not murine, hair bundles. mAb H10 is not specific for the hair bundle, stains other structures in the ear including the tectorial membrane, and it reacts with multiple protein bands on western blots of inner ear lysates (not shown) suggesting it recognizes a shared and widely distributed epitope. Immunoprecipitation experiments reveal mAb H10 recognizes Ptprq, and that this immunoreactivity is retained following chondroitinase ABC treatment (see Fig. 6). Although Ptprq is expressed by all hair bundles in the chick inner ear, and although mAb H10 recognizes Ptprq, mAb H10 only reacts with mature hair bundles in the extrastriolar regions of the utricle [Fig. 7(a)]. It does not react with hair bundles in the striolar region of the utricle or with those in the basilar papilla [Fig. 7(b,c)]. Although staining is seen with mAb H10 at the tips of some basilar papilla hair bundles, this is most likely due to residual-attached tectorial membrane. In the striolar region of the utricle and the basilar papilla, but not in the extrastriolar region of the utricle, mAb H10 stains the apical surface of supporting cells [Fig. 7(a–c)]. mAb H10 does not stain the immature hair bundles that are scattered throughout the utricle of the posthatch bird [Fig. 7(a)], nor does it stain hair bundles in the developing basilar papilla (not shown). Immunogold labeling reveals that mAb H10 does not stain the shaft connectors of striolar hair bundles, but reacts intensely with those of extrastriolar hair bundles [Fig. 8(a–d)]. Together these observations suggest mAb H10 may recognize a variant of Ptprq this is restricted to relatively mature extrastriolar hair bundles.

Figure 6. mAb H10 recognizes a chondroitinase ABC insensitive epitope associated with Ptprq. A western blot of immunoprecipitates that were treated with buffer alone (−ABCase) or buffer containing chondroitinase ABCase (+ABCase) before electrophoresis and staining with mAb H10. Samples were immunoprecipitated from a Triton X-100 soluble fraction from the early posthatch chick utricle with rabbit anti-Ptprq (R27) or nonimmune rabbit IgG (control IgG).

Figure 7. mAb H10 only recognizes the hair bundles of mature extrastriolar hair cells. Confocal images from the extrastriolar region of the utricle (a), the striolar region of the utricle (b), and the basilar papilla (c) of the early posthatch inner ear double labeled with mAb H10 (a-c) and either mAb D10 (a′, b′) or Texas Red conjugated phalloidin (c′). Corresponding merges are shown in panels a′′-c′′. Small arrows (a, a′, a′′) indicate examples of immature hair bundles that are not labeled by mAb H10, but are labeled by mAb D10. Arrowheads in c indicate staining at tectorial membrane attachment sites on the tallest row of stereocilia. Bar = μm.

Shaft Connectors Are Shorter in Extrastriolar Hair Bundles

Electron microscopy of hair bundles from the extrastriolar and striolar regions of the vestibular maculae reveals that there are differences in the structure of shaft connectors that correlate with the presence and absence of the Ptprq isoforms that react with mAb H10. Hair bundles in the striolar regions usually have stereocilia that are thicker and more widely spaced than those in the extrastriolar regions and have shaft connectors that are concentrated in the more basal regions of the stereocilia. The connectors in the basal regions of these striolar hair bundles are typically long (up to 160 nm) and have a distinct, centrally located density [Figs. 8(a,b) and 9(a–d)]. In the extrastriolar regions, the stereocilia are narrower and more tightly packed and the shaft connectors are distributed over most of the hair-bundle surface. In these cells, the shaft connector density is also centrally located, but the length of the connectors (up to 80 nm maximum) is distinctly shorter [Fig. 9(e–h)] than that observed in the striolar region.

Figure 8. mAb H10 labels the shaft connectors in the hair bundles of extrastriolar hair cells. Immunogold labeling of striolar (a, b) and extrastriolar (c, d) chick utricular hair bundles with mAb H10 (a, c) and irrelevant IgM mAbs A46 (b) and F32 (d). Bar = 200 nm.

Figure 9. Shaft connectors of striolar hair bundles are longer than those of extrastriolar hair bundles. Transmission electron micrographs of shaft connectors from the striolar (a–d) and extrastriolar (e–h) regions of the early posthatch chick utricle. Samples were fixed in the presence of ruthenium red. Scale bar = 200 nm.

DISCUSSION

The results of this study provide evidence that the shaft connector protein, Ptprq, is a CS PG, and that there is likely to be a form of Ptprq that is recognized by mAb 473-HD and is only expressed transiently during the early stages of hair-bundle development in the mouse cochlea and chicken inner ear. Furthermore, the data presented suggest that there may be two variants of Ptprq expressed in the mature avian inner ear, one of which is recognized by mAb H10 and the other of which is not, and that the distribution of these variants correlates with the spacing observed between the stereocilia.

Ptprq immunoprecipitated from the early posthatch chicken utricle undergoes a discrete shift in electrophoretic mobility following digestion with chondroitinase ABC. This observation provides the best direct evidence that Ptprq is a CS PG. Most of the Ptprq that is immunoprecipitated from the posthatch utricle is likely to be derived from the extrastriolar hair bundles, as these are present in far greater number than striolar hair bundles in this organ. As only a few extrastriolar hair bundles in the posthatch utricle, most of which are probably immature, stain with mAb 473-HD [see Fig. 5(h)], and as Ptprq derived from this tissue does not react with mAb 473-HD on immunoblots (data not shown), the band shift observed on chondroitinase ABC treatment is likely to be due to the loss of CS glycosaminoglycan chains that are distinct from those recognized by mAb 473-HD.

Ptprq is likely to be a major component of the shaft connectors, as these structures are completely absent in Ptprq^−/− mice (Goodyear et al., 2003). The loss of the ruthenium red-stained dense particles from the shaft connectors that occurs on chondroitinase ABC treatment provides further evidence that Ptprq is a CS PG, as does the absence of mAb 473-HD immunoreactivity in Ptprq^−/− mice. Although the presence of an additional, unidentified CS PG that interacts with Ptprq cannot be formally excluded, a parsimonious explanation for the combined observations would be that Ptprq is itself a CS PG, and that there is a variant that carries the DSD1 epitope and is only expressed at high levels during a brief window of developmental time in the mouse cochlea and the chick ear.

In this respect, it is interesting to note that the DSD1 epitope is also associated with another RPTP, Ptprz, a RPTP that plays an important role in brain development (Hayashi et al., 2005; Faissner et al., 2006). Several isoforms of Ptprz are expressed, including phosphacan (Ptprz-S), a secreted ectodomain variant that has mAb 473-HD reactive CS GAG chains. The presence of the mAb 473-HD epitope is thought to modulate the neurite outgrowth promoting properties of Ptprz (Faissner et al., 1994; Clement et al., 1998), and it is tempting to speculate that the presence of the DSD1 epitope on Ptprq somehow modulates its function during hair-bundle development. Although Ptprq is not required for the very early, embryonic stages of hair-bundle development in the mouse cochlea, it is required for the postnatal maturation of these hair bundles, with defects becoming apparent within a day after birth in Ptprq^−/− mice, at a stage when there is considerable expression of the DSD1 epitope in the hair bundle. The presence of the DSD1 epitope may influence ligand binding and ectodomain interactions.

mAb H10 provides evidence that there may, additionally, be at least two distinct Ptprq variants in the mature avian inner ear. As yet the H10 epitope remains to be identified. It is, however, associated with many proteins and could be a common modification generated by glycosylation. Previous studies have shown that hair bundles in the bird ear can be distinguished on the basis of their relative reactivity to the lectin peanut agglutinin (PNA), which has a specificity for the Gal1-β(1-3)-GalNAc disaccharide. For example, PNA stains extrastriolar hair bundles intensely, striolar hair bundles very weakly and does not react at all with those in the basilar papilla (Goodyear and Richardson, 1994), much in the same way that mAb H10 behaves. Furthermore, Griffonia simplificolia lectin isoform IB(4) with a specificity for terminal α-D-galactosyl residues, only stains immature hair bundles in the avian inner ear (Warchol, 2001), a situation that is similar to that observed with mAb 473-HD. Whatever the H10 epitope is, it is clearly only expressed on the mature hair bundles of extrastriolar hair cells, and immunoblotting shows that it is associated with Ptprq.

The presence and absence of H10 immunoreactivity on hair bundles correlates with the presence and absence of short and long shaft connectors, respectively. How could a putative glyco-epitope modulate the length of the shaft connectors? Also, what might be the function of the CS GAG that is associated with the Ptprq ectodomain? The GAG chains of the Ptprq ectodomain could be located at any point along the length of the Ptprq ectodomain and may provide a negatively charged glycocalyx that extends up to ∼70 nm (18 × 4 nm long FN3 repeats) from the cell surface. Theoretical studies (Dolgobrodov et al., 2000) have indicated electrostatic interactions between the glycocalyces of opposing stereocilia could provide either adhesive or repellent forces depending on the relative disposition of the charged groups (see Fig. 10). The presence of GAG (or other charged structural side chains) located centrally as opposed to distally within the Ptprq ectodomain (i.e., closer to the plasma membrane) may allow for a closer spacing between stereocilia, as is seen in the extrastriolar region where mAb H10 immunoreactivity is found. Absence of such centrally located side chains may allow for greater spacing, as is seen in the striolar regions of the maculae or the basilar papilla. Such a scenario would require that hair cells differentially modify the type of Ptprq that they express. One possibility would be that exon splicing removes the sites at which carbohydrates are added to the ectodomain. However, it is known that modification of a protein with GAG chains can be differentially regulated in both a stage- and region-specific manner (e.g., Hamanaka et al., 1997). Therefore, the possibility that the same core protein is differently modified during development cannot be excluded.

Figure 10. Models for adhesive and repellent shaft connector interactions. Diagram illustrating how differences in the location of GAG chains along the ectodomain of Ptprq could influence the spacing of stereocilia when mediating putative repellent (top) or adhesive (bottom) interactions between stereocilia.

Overall the results imply there may be at least three variants of Ptprq in the chick, an immature form substituted with the DSD1 epitope and two mature forms, one with and one without the mAb H10 epitope. The functional consequences of these differences remain to be explored. Although it has been argued previously that shaft connectors may act as spacers as opposed to adhesive links (Goodyear et al., 2003), a role for Ptprq in maintaining hair-bundle cohesion cannot be ruled out. The coherence of hair bundles during large deflections has not been tested in Ptprq^−/− mutants at stages before hair-bundle fusion occurs, and the functional consequences of chondroitinase ABC treatment remain to be examined. In the short term (within 1 h), the loss of CS GAG seems to have little effect, as no obvious differences in stereocilia spacing have been seen between control and chondroitinase treated hair bundles. The effects of imposing large deflections on these hair bundles or of growing them for longer time periods now need to be explored.

Acknowledgements

Gowri Nayak was supported by a PhD studentship from the Royal National Institute for Deaf People.