A Mutation in CCDC50, a Gene Encoding an Effector of Epidermal Growth Factor–Mediated Cell Signaling, Causes Progressive Hearing Loss

Mutation Screening of the Human CCDC50 Gene

Primers to PCR amplify all the exons and intron-exon boundaries of the human CCDC50 gene were designed using Oligo 4.0 (Molecular Biology Insights). PCRs were performed in standard conditions, as described elsewhere.12 PCR products were purified with the QIAquick PCR purification kit (Qiagen) and were directly sequenced using the BigDye Terminator v3.1 Cycle Sequencing Ready Reaction Kit in an ABI PRISM 3100 genetic analyzer (Applied Biosystems). To isolate the c.1394_1401dupCACGGCAT mutation, the exon 11 amplimer obtained from patient III:1 (fig. 1) was cloned into pUC19 (New England Biolabs), was transformed into XL1-Blue competent cells, and was grown on Luria Bertani plates with isopropyl-β-D-thiogalactopyranoside and X-gal under ampicillin selection. Plasmid DNA obtained from several white colonies was sequenced using pUC (-47) and pUC (-48) universal primers. Mutant sequences contained a CACGGCAT 8-bp tandem duplication. For the detection of the mutation in the human population, we developed a size-based screening test in which exon 11 is amplified with a pair of primers, one of which is fluorescently labeled. Allele size was obtained after capillary electrophoresis in an ABI PRISM 3100 genetic analyzer (Applied Biosystems), by analysis with the GeneScan program (Applied Biosystems). The wild-type allele is 8 bp smaller than that of the mutant.

RT-PCR Analysis for Transcript Variants of the Murine Ccdc50 Gene

Total RNA was prepared from the mouse cochleae at embryonic day (E) 18.5, postnatal day (P) 0, and P14 and from mouse lung, brain, and liver at P58, and the contaminating genomic DNA was eliminated using RNase-free DNase treatment. Approximately 1 μg of total RNA was reverse transcribed using SuperScriptase II (Invitrogen) and oligo-dT primer. The sequences of primers used to RT-PCR amplify the different transcripts of the mouse Ccdc50 gene are as follows: F1, 5′-CCATGGCCGACGTGAGTG-3′; F5, 5′-TCTTTGTACCCATGTCATGAAGAAAA-3′; F6, 5′-ATTGCAAGAAGAAGAACTTTTGGAA-3′; F7, 5′-CGTGAGTGTAGATCAGTCCAAGTTG-3′; F8, 5′-CCATGTCATGAAGAAAATAATCTCTTG-3′; F10, 5′-CGCGGGGAGGTATTTCTGAG-3′; R1, 5′-CCAGAATGGCACACAGAAGGAC-3′; R2, 5′-GCCGATCCCTTTTTGTTTGTATT-3′; R4, 5′-AGGCTGTTTGTTCTTTAATTTTGAG-3′; R8, 5′-CAAGAGATTATTTTCTTCATGACATGG-3′; R9, 5′- CATTAGAAGTCGAGCAATTTCCAA-3′; and R14, 5′-CCATGGTGATGAGACTGCTTTTC-3′. Table 1 summarizes the primer pairs used in the amplification of the different Ccdc50 transcript variants. Primary RT-PCRs were performed in a thermocycler (Perkin-Elmer 9700) with the use of 0.75 U of FastStart Taq polymerase (Roche Applied Science) and 1:50 of the cDNA preparation in a final volume of 15 μl; 40 cycles (at 96°C for 30 s, 58°C for 30 s, and 72°C for 1 min/kb) were used. Secondary amplifications were done using nested primers (see table 1) and 1.5 μl of a 1:100 dilution of the previous PCR (F1-R4); 30 cycles (at 96°C for 30 s, 63°C for 15 s, and 72°C for 30 s) were used. Direct sequencing of PCR products was performed following the standard procedure described above. RACE (rapid amplification of cDNA ends) analyses were performed on total RNA from E17.5 whole-embryo mouse, with use of the Marathon cDNA Amplification kit (BD Biosciences-Clontech).

Northern Blot Analysis

Northern analysis was performed on a commercial blot (FirstChoice Mouse Blot I [Ambion]) containing mRNA from 10 different mouse tissues (heart, brain, liver, spleen, kidney, embryo, lung, thymus, testes, and ovary), in accordance with the manufacturer’s instructions. Mouse Ccdc50 ORF cDNA (transcript 1) was amplified with the F1-R1 primer pair, was cloned into pGEM-5Zf(+) (Promega), and was linearized. Antisense transcripts were generated using T7 polymerase (Strip-EZ RNA probe synthesis [Ambion]) and were nonisotopically labeled (BrightStar Psoralen-Biotin [Ambion]). The northern blot was hybridized overnight at 60°C with the full-length antisense mouse Ccdc50 probe and was washed at 58°C three times at low stringency and two times at high stringency. Nonisotopic detection was performed according to the manufacturer’s directions (BrightStar Biodetect [Ambion]). The blot was stripped and reprobed with antisense probe generated from pTRI-β-actin mouse control template (Ambion).

Antibodies and Immunoblot Analysis

Immune serum against two nonoverlapping synthetic peptides based on the predicted sequence of the mouse Ymer protein located in the N-terminal (peptide I, RIQEKKDEDIARLL) and C-terminal (peptide II, NQHSTTWHLPKSES) regions of the protein (see fig. 2) were raised in rabbit (CovalAb). After three boosts of immunization, the antiserum was affinity purified. The specificity of the polyclonal antibody was tested by immunoblotting of protein extracts prepared from the inner ear. In brief, cochleae from P2 mice were homogenized and sequentially extracted with Tris-buffered saline (TBS) (150 mM NaCl and 10 mM Tris-HCl [pH 7.4]), high-salt solution (1.0 M NaCl and 10 mM Tris-HCl [pH 7.4]), low-salt solution (10 mM Tris-HCl [pH 7.4]), and 1% nonionic detergent (1% Triton X-100 and 10 mM Tris-HCl [pH 7.4]). High-speed centrifugation after each extraction was used to produce soluble supernatants from which the protein was precipitated with trichloroacetic acid or acetone before gel electrophoresis. The insoluble fraction remaining after centrifugation in TX-100 solution was solubilized in SDS-PAGE sample buffer. Protein samples were run under reducing conditions on 10% polyacrylamide gels and were transferred to a polyvinylidene fluoride Hybond-P membrane (Amersham Biosciences). Membranes were blocked for 1 h with 3% semiskimmed milk in TBS Tween (TBST) (150 mM NaCl, 10 mM Tris-HCl [pH 7.2], and 0.05% Tween 20) and were incubated overnight with the primary antibody at 1:400 dilution. After three washes with TBST, the membrane was incubated with alkaline phosphatase–conjugated goat anti-rabbit secondary antibody (DAKO) at 1:1,000 dilution, and the bound antibodies were visualized with nitro-blue tetrazolium chloride/5-bromo-4-chloro-3′ indolyl phosphate p-toluidine solution.

Transfections

NIH-3T3 and Hela cells were grown in Dulbecco's modified Eagle medium (Sigma) containing 10% fetal bovine serum, 1% glutamine, and 1% penicillin/streptomycin under standard conditions, were transiently transfected with JetPEI (Qbiogene), and were processed for immunofluorescence as reported elsewhere.13 For transfection purposes, mouse Ccdc50 transcripts 1 and 2 were PCR amplified (forward primer, 5′-CCATGGCCGACGTGAGTG-3′; reverse primer, 5′-CCAGAATGGCACACAGAAGGAC-3′) with the use of mouse brain cDNA as a template and were cloned into the dicistronic expression vector pIRES-hrGFP-1a (Stratagene), which is driven by the cytomegalovirus promoter. Similarly, human CCDC50 transcripts 1 and 2 were amplified with the forward 5′-CGGGCTCCGGATATTTGGTA-3′ and reverse 5′-TGAGTCCATTTTCAAGGCAGATTC-3′ primers with the use of cDNA from Epstein-Barr virus–transformed human B lymphocytes and were cloned into the expression vector pIRES-hrGFP-1a (Stratagene), and the 8-bp duplication was introduced at exon 11 by mutagenesis (Quick Change Site-Directed Mutagenesis Kit [Stratagene]).

Immunohistochemistry

Mouse inner ears were dissected and fixed in 3.7% formaldehyde in 0.1 M sodium phosphate buffer (pH 7.4). Tissues obtained from P5 mice onwards were decalcified with 0.5 M EDTA (pH 8.0) for 5 d before sectioning. Inner ears were cryoprotected in 30% sucrose in PBS, were embedded in 1% low-gelling agarose in 18% sucrose in PBS, and the agar blocks were mounted for cryosectioning onto microtome chucks with use of optimal cutting temperature compound (Tissue Tek [Miles Scientific]). Sections were cut in a cryotome at −30°C.

For immunostaining, cryosections were blocked with 10% horse serum in TBS containing 1 mM sodium azide for 1 h and were incubated overnight with 1:100 dilution of the primary antibody. After three TBS washes, the slides were incubated with a 1:100 dilution of the fluorescein isothiocyanate (FITC)–conjugated swine anti-rabbit secondary antibody (DAKO) for 1 h. Sections were mounted in Vectashield H-1000 (Vector Laboratories). Staining specificity was confirmed by substituting the primary antibody with nonimmune rabbit immunoglobulin G (IgG) at the same concentration or by preabsorbing antibodies with the two peptide antigens. Preimmune sera from the rabbit used to generate the polyclonal anti-Ymer antibody was also used to check specificity. Tissue-section slides were visualized in an Axioplan 2 fluorescence microscope (Carl Zeiss), and the images were captured with a Spot-Junior Camera (Diagnostic Instruments).

For colocalization experiments, the following primary antibodies were used: mouse monoclonal antibodies to α-tubulin (A-11126 [Molecular Probes] and TU-16 [Abcam]), acetylated α-tubulin (clone 6–11B-1 [Sigma]), polyglutamylated tubulin (T9822 [Sigma]), and β-tubulin I and II (clone JDR.3B8 [Sigma]). Primary antibodies were detected with Alexa Fluor 594-conjugated goat anti-rabbit IgG (H+L) (Molecular Probes) and FITC-conjugated goat anti-mouse F(ab′)₂ IgG fragment (Sigma). Actin was detected with rhodamine or Alexa-Fluor 350–conjugated phalloidin. For colocalization studies, whole-mount and single-cell preparations were analyzed on a Zeiss LSM510 laser scanning confocal microscope.

Identification of the Mutation in CCDC50 Causing DFNA44 Hearing Loss

Sequence analysis of all exons and flanking intronic sequences of CCDC50 in the affected subject III:1 (fig. 1A) revealed eight sequence variations (table 2). Seven of these changes, located in exons 3, 6, 7, and 10, included four silent changes and three nonsynonymous amino acid substitutions that, with the exception of the synonymous change at exon 3, corresponded to previously reported nonpathogenic SNPs. The remaining change corresponded to a heterozygous mutation in exon 11 (c.1394_1401dupCACGGCAT, with respect to the longer transcript [GenBank accession number NM_178335]). This 8-bp tandem duplication leads to a frameshift that replaces the last 15 aa of the protein by a novel 36-aa sequence (p.Phe468HisfsX37) (fig. 1B–1D). We took advantage of the size difference between the wild-type and mutant alleles to design a screening test to detect this mutation (see fig. 1E and the “Material and Methods” section). This test showed that the mutation segregated with the hearing impairment in the family. In addition, it was not present in 100 unrelated Spanish individuals with normal hearing. These findings indicate that c.1394_1401dupCACGGCAT is the pathogenic mutation responsible for DFNA44 hearing loss in the family. We screened for this mutation in probands from 280 unrelated Spanish families displaying nonsyndromic sensorineural hearing loss. Of these patients, 135 belonged to families segregating autosomal dominant hearing impairment, and the remainder belonged to families in which the pattern of inheritance could not be defined unambiguously. The mutation was not detected in any of these probands.

Identification of Ccdc50 Transcripts in the Mouse Inner Ear

The orthologous mouse Ccdc50 gene (UniGene Mm.258985) is also predicted to encode two isoforms. Transcript 1 (GenBank accession number NM_026202.2)5 lacks exon 6 and corresponds to the shorter human variant. It is expressed in several tissues including the organ of Corti (GenBank accession number BQ567072.1) and is predicted to encode a protein of 305 aa. Transcript 2 (GenBank accession number NM_001025615.1), generated by exon 8 skipping, is expressed in the testis of the adult mouse (GenBank accession number AK016827.1) and is predicted to encode a protein of 290 aa (fig. 2A). Ccdc50 is widely expressed in most tissues tested, but there is little information available concerning the different transcripts that are present in these tissues and their expression at the different stages of mouse inner ear development. We therefore performed RT-PCR analysis on mRNA from the mouse cochlea at E18.5, P0, and P14. This analysis detected the presence of transcripts 1 and 2 at all three developmental stages (fig. 3A and 3B). In addition, we used 5′-RACE-PCR to investigate the presence of putative novel exons upstream of the described Ccdc50 exon 1 (data not shown). Although no other exons were found toward the 5′ region of the gene, sequence analysis of the 5′-RACE-PCR products revealed the existence of a novel exon between exons 3 and 4. This novel coding sequence, exon 3b, comprises a 46-bp sequence located 243 bp downstream of exon 3 in the genomic sequence (fig. 2B). A set of specific primers was therefore designed to amplify overlapping fragments of the full-length Ccdc50 coding sequence and to investigate the presence of putative novel transcripts expressed in the inner ear containing this novel exon (see the “Material and Methods” section, figs. 2A and 3A, and table 1). This analysis led us to characterize two additional transcripts (transcripts 3 and 4) that are derived, respectively, from transcripts 1 and 2 by the inclusion of exon 3b. The presence of this exon in both transcripts leads to a frameshift that generates a premature stop codon at the beginning of exon 4. Transcripts including exon 3b are predicted to generate a shorter protein isoform of 95 aa. The last 15 aa of this isoform would correspond to the predicted protein fragment encoded by exon 3b. Similar to transcripts 1 and 2, the variants 3 and 4 were identified in the inner ear at all three stages tested (fig. 3A).

Although a BLAST analysis (BLASTN, TBLASTN, BLASTX, or cross-species megaBLAST performed against the Human Genome Database) did not reveal the existence of this exon in humans, a ClustalW alignment performed between the human CCDC50 intron 3 and the mouse Ccdc50 exon 3b sequences detected a stretch of human sequence with ∼71% similarity to mouse exon 3b (fig. 2D). Conserved donor and acceptor splice sites, however, were not present, suggesting that it is unlikely that there are human splice variants containing the orthologous exon 3b. Further studies at the RNA level are necessary for a more definitive conclusion.

Although the mouse gene was not reported to contain an exon homologous to human exon 6, cross-species megaBLAST analysis, with the use of human exon 6 as a query, revealed a putative mouse exon 6, and the conceptual translation was ∼60% identical to the human protein (hmm192144). We therefore designed mouse exon 6–specific primers (F10 and R14), and different RT-PCRs were performed. This strategy allowed us to identify up to four novel transcripts that include this exon (figs. 2A and 3A and table 1). The presence of exon 6 in these transcripts was confirmed by direct sequencing of different overlapping fragments (fig. 2C). RT-PCR analysis of lung, brain, and liver mRNA from a P58 mouse indicates that the expression of none of these transcripts, variants 1–8, is restricted to the mouse inner ear.

Analysis of a northern blot with mRNA from 10 different tissues (see the “Material and Methods” section) detected two bands of ∼4.6 and 7.1 kb (fig. 3C). These findings indicated that the gene is ubiquitously expressed. The band sizes do not correspond with the expected length for mouse transcripts 1 and 2, reported to be ∼3.5 kb. They are much longer and may correspond to transcripts with alternate 3′ ends. Indeed, the 7.1-kb band is similar to the predicted sizes of hmm192700 and hmm192422 (models for transcripts 1 and 2, respectively), which contain a 3′ UTR of up to 6 kb.

A processed Ccdc50 pseudogene showing a 93% similarity with mouse transcript 1 is present on mouse chromosome 4 (GenBank accession number NG_005190.2). The expression of the pseudogene has not been reported so far, and it has a premature stop codon at exon 5. The specific amplification of Ccdc50 gene transcripts was ensured by designing RT-PCR primer pairs that take into account the sequence variations that exist with respect to the pseudogene, and the PCR products were also verified by direct sequencing.

Ymer Is a Soluble Cytoplasmic Protein

A polyclonal antibody to Ymer was raised by injecting a rabbit simultaneously with two peptides based on the predicted sequence of the mouse protein, one located in the N-terminal half of the protein (residues 103–116) and the other in the C-terminal region (residues 283–296) (see the “Materials and Methods” section and fig. 2A). The antibody was affinity purified on a column to which both peptides had been conjugated, and its specificity was assessed by western blot analysis of protein extracts prepared from three different tissues of the P2 mouse inner ear: the greater epithelial ridge (GER), the stria vascularis, and the modiolus. This analysis revealed a single prominent band of ∼38 kDa in all three tissues. This molecular mass is similar to that predicted from murine Ymer isoforms 1 and 2 (fig. 4A). A minor band of approximately double the mass was also detected and may correspond to a stable dimeric form of the protein. These two bands were not observed when blots were probed with nonimmune rabbit IgGs at the same concentration (data not shown), indicating the specificity of the purified antibody. On the basis of the location of the two peptides used to generate the polyclonal antibody, it should be possible to detect the predicted isoforms 5 and 6 (fig. 2A). They have not been detected, however, which suggests they are expressed at levels that escape the sensitivity of this technique.

The solubility of the Ymer protein was estimated by immunoblotting proteins prepared by sequentially extracting cochlear tissues with normal saline, buffers of high and low ionic strength, and a buffer containing nonionic detergent (see the “Material and Methods” section). The 38-kDa band was readily extracted in TBS (fig. 4B). The subcellular localization of Ymer was investigated in NIH-3T3 and Hela cells transiently transfected with the murine wild-type transcripts 1 and 2. Transfected cells showed that Ymer is globally distributed in the cytoplasm and is not present in the nucleus (fig. 4C). Staining was not observed in nondividing cells transfected with the empty (mock) vector. Collectively, these results indicate that Ymer is likely to be a soluble cytoplasmic protein.

To evaluate whether the identified mutation alters the subcellular distribution of the protein, NIH-3T3 and Hela cells were transiently transfected with mutated human transcripts 1 and 2. Mutated protein is expressed but, unlike the wild-type protein, is restricted to the perinuclear area (fig. 4D).

Expression Pattern of the Ymer Protein in the Mouse Inner Ear

The spatiotemporal expression pattern of the Ymer protein was investigated in the developing and adult mouse inner ear by immunohistochemistry with the affinity-purified antibody. We compared Ymer expression in the mouse inner ear at embryonic stages E14.5 and E17.5 and at several postnatal stages from P2 to P69.

At E14.5, strong punctate immunostaining was detected in the otic mesenchyme. Moreover, expression was detected in the cells lining the lumen of the primitive cochlear duct and in the nerve fibers of the spiral ganglion (fig. 5A). Immunolabeling of nerve fibers invading the cochlear epithelium was also observed at this stage. At E17.5, a strong signal was detected only in the mesenchyme around the cochlear duct (fig. 5B).

By P2, the mesenchymal staining became weaker, except in the region of the spiral limbus, and Ymer expression was observed for the first time in the apical region of the developing pillar cells (PCs). Staining was also observed in the cytoplasm of the outer hair cells (OHCs), in their innervating nerve fibers, and in the marginal cells of the stria vascularis at this stage (fig. 5C). At P9 and P12, Ymer was detected strongly through the entire length of the inner and outer PCs, and the staining in tissues of mesenchymal origin was restricted to two major structures, the spiral limbus and the spiral ligament. Also at P9 and P12, labeling became more evident in the stria vascularis (fig. 5D and 5E). By P14 and P16, strong staining was still evident in the PCs and in the marginal cells of the stria vascularis. Staining was weak in the spiral limbus and spiral ligament at this stage and had begun to appear in Deiter’s cells (fig. 5F and 5G). At P19 and P22, staining was still intense in the PCs and stria vascularis and was strong in the cell bodies and the processes of Deiter’s cells. Weaker staining was observed in the spiral limbus and spiral ligament at these stages (fig. 5H and 5I). At P31, when the inner ear is functionally mature, Ymer was expressed only in the stria vascularis and PCs (fig. 5J). At P69, the staining was still strong in PCs and less intense in stria vascularis (fig. 5K).

At all stages, for negative controls, we substituted the primary antibody with nonimmune rabbit IgGs at the same concentration or used preimmune sera from the rabbit used to generate the polyclonal antibody. Similar results were obtained in both controls, with nonspecific staining only in one region of the cochlea corresponding to the cells of the spiral ligament in the later stages of development (fig. 5L). These results indicate that the cochlear localization and expression levels of Ymer vary dynamically with time from embryonic stages onward (table 3).

The expression of Ymer in the vestibular maculae and the cristae ampullaris was similar to that observed in the cochlea (fig. 6)—that is, with a strong signal observed in the mesenchyme at the initial embryonic stages that progressively became weaker and was less prominent in the adult (P33). Staining of the nerve fibers innervating the sensory epithelia was seen at E17.5 and P2 (fig. 6A–6C). Staining was initially detected in some cells of the vestibular epithelium at E17.5 and became more evident at P2 (fig. 6B–6D). Later, the staining increased and persisted to adult stages in which it was mainly restricted to the apical cytoplasm of the epithelial cells (fig. 6E and 6F).

Ymer Colocalizes with Microtubules of the Cytoskeleton and Mitotic Apparatus

The staining pattern of Ymer through the entire length of PCs and Deiter’s cells suggested this protein could be associated with their prominent microtubule-based cytoskeleton. This possibility was investigated by double immunolabeling for Ymer and different types of tubulins in postnatal mouse inner ear sections. This analysis showed that Ymer exhibits overlap with the microtubule bundles of PCs and Deiter’s cells and with the complex microtubule structure in the interdigitating processes of the marginal and intermediate cells in the stria vascularis (fig. 7A and 7C). Moreover, a clear colocalization was observed in the apical cytoplasm of vestibular epithelial cells (fig. 7B). The subcellular localization of Ymer was similarly investigated in the mouse NIH-3T3 cell line. Double labeling of nontransfected cells revealed that Ymer colocalized with microtubules of the mitotic apparatus of cells undergoing division. During the nuclear division, or karyokinesis, Ymer colocalizes with microtubules of the mitotic spindle. In the final steps of cell division, during cytokinesis, Ymer overlaps with the midzone, a region of bundled microtubules between the migrated chromosomes, and then with the microtubule midbody contained in the intercellular bridge that is created as the cleavage furrow ingresses and compresses the midzone (fig. 7D). Taken together, these results suggest that Ymer is associated with microtubule-based structures.