WO2001070972A2 - Multiple human and mouse otoferlin isoforms - Google Patents

Abstract

Using a candidate gene approach, we identified a novel human gene, Otoferlin, as underlying an autosomal recessive nonsyndromic prelingual deafness, DFNB9. Otoferlin is the second member of a mammalian gene family related to Caenorhabditis elegans fer-1. Further, we describe the search for the cDNA encoded by the mouse orthologous gene, Otof. A comprehensive analyses of the human and mouse otoferlin gene organization is also reported as well the characterization of several alternatively spliced forms.

Description

MULTIPLE HUMAN AND MOUSE OTOFERLIN ISOFORMS

The present invention relates to nucleic acids and protein implicated in deafness.

Approximately 1 in 1000 children is affected by deafness at birth or before 2 years of age, i.e. in the prelingual period. Although a variety of syndromes including hearing loss have been described, the hearing loss is the sole symptom in the majority of cases; these forms are referred to as isolated or nonsyndromic deafness. In developed countries, approximately two thirds of the prelingual nonsyndromic deafness cases are of a genetic origin, 85% of which are inherited in an autosomal recessive mode (DFNB forms). These forms most frequently cause severe or profound hearing loss, which impedes speech acquisition (ref. 1, for a review). The DFNB forms have been predicted to be monogenic diseases which are genetically highly heterogenous; 20 different DFNB loci have been reported to date. Five of the corresponding genes have been identified, namely GJB2 (or Cx26) encoding the gap junction protein connexin26 (DFNBl)², MY07A andMY015 encoding two unconventional myosins, myosin NHA and myosXN (DFΝB2 and DFNB3, respectively)^3"5, PDS encoding pendrin, a putative sulfate⁶ or iodide⁷ transporter (DFNB4), and TECTA encoding a component of the rectorial membrane, a-tectorin (DFNB21)⁸. With the exception of DFNBl which accounts for approximately half of all non-syndromic prelingual inherited deafness cases in the populations studied (France, Italy, Spain, United Kingdom and New Zealand)^9"1 ', only one or a few affected families have been reported for each deafness locus. As a result, in most instances, the chromosomal candidate gene interval which can be defined is too large to undertake the search for the corresponding gene exclusively by a positional cloning strategy. In order to circumvent this difficulty, a candidate gene approach based on the isolation of genes specifically or preferentially expressed in the inner ear has been developed; the proteins encoded by these genes are indeed likely to play a crucial role in auditory function¹.

It is an object of the present invention to provide nucleic acid sequences which are implicated in deafness. Here, using a combination of the candidate gene strategy that we previously described¹²'¹³ and a positional cloning strategy, we describe the gene underlying DFNB9 (OMM 601071).

Using a candidate gene approach, we identified a novel human gene, Otoferlin, as underlying an autosomal recessive nonsyndromic prelingual deafness, DFNB9. The same nonsense mutation was detected in four unrelated Lebanese affected families. Otoferlin is the second member of a mammalian gene family related to Caenorhabditis elegans fer-1. It encodes a predicted 1230 aminoacid cytosolic protein, with three C2 domains and a single C-terminal transmembrane domain. The sequence homologies and predicted structure of otoferlin suggest its involvement in vesicle membrane fusion. In the inner ear, the expression of the murine orthologous gene mainly in the sensory hair cells indicates that such a role could apply to synaptic vesicles.

Further, we describe the search for the cDNA encoded by the mouse orthologous gene, Otof. It led us to detect a cDNA form predicted to encode a protein isoform longer than the one we initially reported. A comprehensive analyses of the human and mouse otoferlin gene organization is also reported as well the characterization of several alternatively spliced forms.

The study described above demonstrates a human gene, OTOF, as underlying an autosomal recessive form of sensorineural prelingual non-syndromic deafness, DFNB9. The 5-kb (kilo base pairs) cDNA we isolated, was predicted to encode a 1230 aa C-terminal membrane anchored cytosolic protein with three C2 domains, which was named otoferlin in reference to its homology with Caenorhabditis elegans FER-1.

Another human FER-1 like protein, dysferlin, is responsible for two types of muscular dystrophies. We now cloned a cDNA derived from the mouse orthologous gene, Otof. Interestingly, in cochlea (the auditory sensory organ) and brain, a 7-kb cDNA was detected which was predicted to encode a 1992 aa and a 1977 aa protein, respectively.

We seeked for the human long otoferlin cDNA form in a variety of tissues. Only in brain, a cDNA expected to encode a 1977 aa protein could be isolated. The mouse/human long otoferlin isoform is predicted to contain five C2 domains as dysferlin. The complete gene structures of OTOF and Otof wer determined; they extend on approximately 90 kb and 80 kb, respectively, and are composed of 48 coding exons. Primers to amplify each of the 19 additional 5' OTOF exons were selected for further analyses of DFNB9 patients. The variety of alternatively spliced forms detected predict several otoferlin isoforms which differ by N-terminal region, first and third inter-C2 domain as well as very C- terminal end. Future works will aim to understand the role of these various tissue specific otoferlin isoforms. A more complete appreciation of the invention and many of the attendant advantages thereof will be readily obtained as the same becomes better understood by reference to the following detailed description when considered in connection with the accompanying drawings, wherein:

Figure 1. Genetic linkage analysis of the DFNB 9 affected family AB. Individuals with prelingual deafness are indicated by filled symbols and unaffected individuals by open symbols. Segregation analysis with the microsatellite polymorphic markers located on chromosome 2p23.1 limited the DFNB9 candidate interval to between D2S158 and D2S165. The haplotype associated with the mutated DFNB9 allele in this family (AB) is indicated by a vertical bar. Combining these data with the previous mapping data obtained in family F (ref. 14) permitted the narrowing of the candidate gene interval to between markers D2S158 andD2S174, Le. a distance of approximately I cM (see Fig. 2).

Figure 2. Physical map of the DFNB9 region.

The candidate region is limited by loci D2S158 and D2S174. Poly- morphic markers corresponding to D2S2223 andD2S2350 axe homozygous in all deaf individuals from the DFNB9 affected families, AB, F, Kl and K2. YACs, BACs and PACs are represented by hatched, grey plain and dotted lines, respectively. PCR amplification using primers derived from the 3' ends of cDNAs permitted the mapping of 4 genes, namely PPIL1 (peptidylprolyl isomerase (cyclophilin)-like 1), HADHA and HADHB (hydroxyacyl-Coenzyme A dehydrogenase/3-ketoacyl-Coenzyme A thiolase/enoyl-Coenzyme A hydratase (trifunctional protein) and β subunits), and CENPA (centromere protein A) to this contig. Six ESTS, ie. RH57018, RH20012, RH1296, RH25229, RHI2053 and RH26192 could also be assigned to this contig. The ten corresponding genes could be ordered, with the exception of HADHA/RH20012, HADHB/RH1296 and CENPA/RH26192. The position of HADHA and of the gene encoding RH20012 with regards to D2S158 could not be determined. RH12053 corresponds to the 3' end of the Otoferlin cDNA. PCR amplification performed on the BACs of the contig orientated the 5' end of the gene centromeric to its 3' end. The deduced aa sequence corresponding to RH57018 presented similarity with those of ras- related GTP binding proteins.

Figure 3 a. Deduced amino acid sequence of human otoferlin. The predicted C-terminal transmembrane domain is indicated in bold letters. The three predicted C2 domains are underlined. The five aspartyl residues which presumably bind Ca²⁺ in the three C2 domains²⁰ are indicated in bold, two of them are located in the loop between the predicted β strands 2 and 3 and the three others in the loop between the predicted β strands 6 and 7. The C2A, C2B and C2C domains share only 20% aa identity and 50% aa similarity on average. For the C2A (aa position 196 to 329) and C2B (aa position 709 to 833) domains, the similarity with SYT-1 C2A does not extend to the βl and β8 strands; however, βl and β8 strands are predicted by the secondary structure analysis softwares (see Methods). The C2C domain (aa position 949 to 1104) shows similarity with SYT-1 C2A all along the eight β strands, except for the loop between β5 and β6 which is longer (37 aa instead of 14 aa). The position of the premature stop codon in the DFNB9 patients is indicated by an arrow head (Y730).

Figure 3b. Schematic representation of the otoferlin, dysferlin and fer- 1 structures.

The transmembrane domain is indicated by a vertical bar. Dashed boxes indicate the sequences related to C2 domains, partial or full. The C2 domains with putative Ca²⁺ binding motifs are indicated by *.

The three proteins are mainly similar in their C-terminal regions, where otoferlin (aa position 196 to C-terminus) showed a 39.3% and 30. 1 % identity and a 65.6% and 57.5% similarity with dysferlin (aa position 1135 to C-terminus) and fer-1 (aa position 954 to Cterminus), respectively. The N-terminal 195 aa fragment of otoferlin is weakly similar to dysferlin (aa position 697 to 889) and to fer-1 in two different regions (aa positions 13 to 217 and 550 to 745). In dysferlin, the two N-terminal sequences related to C2 domains are located at aa positions 1-108 and 363-504, respectively. In fer- 1, the three C-terminal sequences related to C2 domains show only 16.4% to 21.9% identity and 48.4% to 49.2% similarity with the SYT-1 C2A.

Figure 4. Sequence analysis of the mutation present in Otoferlin exon 18 in family F. Genomic sequence of a control individual (a), a heterozygous parent (b) and an affected individual (c). The position of the T to A transversion, which results in a premature stop codon at aa position 730, is indicated by an arrowhead.

Figure 5. RT-PCR analysis of otoferlin expression in murine tissues. (a) Otoferlin. Primers were 75A2MH-P5 (3'-LJTR exon) for reverse transcription, and 75A2MH-P1 (exon 27) and 75A2NM-P3 (exon 25) for PCR. (b) Gapdh, used as a positive control. +/- indicates presence or absence of reverse transcriptase in the

RT reaction.

Figure 6. In situ hybridization analysis of otoferlin expression in the mouse inner ear.

(a, b) Cochlea, a: At P2, otoferlin is strongly expressed in the inner hair cells (ihc). A weak signal is seen in the outer hair cells (ohc) and spiral ganglion (sg) neurons, b: At PI 2, within the organ of Corti, otoferlin-is detected in the inner hair cell. No signal is observed in the outer hair cells and supporting cells. (c-f) Vestibule. At P2, otoferlin is highly expressed in the neuro- epithelia of the utricle (c) and lateral and superior crista arnpullaris (d). c, d: The orientation of the sections is indicated in (d): (A)= anterior, (P)= posterior, (D)= dorsal, (V)= ventral, e, f: The detail of the utricular neuroepithehum boxed in (c) shows otoferlin labelling in some hair cells (e), and myosinNIIA staining in hair cells (t). Otoferlin is highly expressed in type I hair cells (arrowheads), whereas only weak labelling is observed in type II hair cells (arrows). The quenching of the fluorescence signal within type I cells is due to the co-localization of the otoferlin and myo7a labellings. rm: Reissner's membrane, sv: stria vascularis. Bar = 50 μm (a), 20 μm (b), 10 μm (c, d), 35 μm (e,f). Figure 7. Schematic representation OTOF gene structures and cDNA.

(Bottom) The originally described (heart and fetus) cDNA form, the brain short and long forms of OTOF. The predicted coding regions are indicated by filled boxes; the predicted untranslated regions by open boxes.

Figure 8. Deduced amino acid sequence of the brain long human otoferlin isoform. The predicted C-terminal transmembrane domain is indicated in bold letters. The five C2 domains are underlined, in which the five aspartyl residues (D) that presumably bind Ca²⁺ (13, 20) are indicated in bold. The C2 domains are designated as C2A (aa 1-105), C2B (aa 401-537), C2C (aa 943-1076), C2D (aa 1476-1600) and C2E (aa 1716-1870). Previously, C2C, C2D and C2E were designated as C2A, C2B and C2C, respectively (2).

Figure 9. Schematic representation of the short and the long forms of otoferlin, human dysferlin (11, 12) and nematode FER-1 (10). The C-terminal transmembrane domain is indicated by a vertical bar. Dashed boxes indicate the predicted C2 domains. The C2 domains with a putative Ca²⁺ binding motif (13, 20) are indicated by an asterisk (*).

Figure 10. RT-PCR analysis of alternative splicings in mouse cochlea and brain. Lane 1, 3, 4 for cochlea, and lane 2, 4 and 6 for brain. Lane 7 is loaded by a size marker, fX174 Hae III digest. (Lane 1 and 2) Amplifications were performed using primers from exon 4 and exon 10 to detect the skipping of exon 6. A 561-bp band and a

516-bp band were observed in cochlea and only the shorter band was observed in brain.

(Lane 3 and 4) For identification of two forms of exon 31, primers were selected from exon 29-30 and exon 32-33. Two bands of 282 bp and 342 bp exist in cochlea and brain, however in cochlea the longer band is only a faint one. (Lane 5 and 6) Amplifications were performed using primers from exon 44 and 3'-UTR in exon 48 to detect the skipping of exon 47. Single 688-bp band was observed in cochlea, and single 891-bp band in brain. Figure 11. Alternative splicings of Otof. The coding regions are indicated by filled boxes, whereas the untranslated regions are indicated by open boxes.

(a) Exon 6 is not present in brain transcript, and both forms are observed in cochlea, (b)

Most of the cochlea transcript only contain the 3 '-part of exon 31 (previously described as exon 12), whereas brain contain both of complete and partial exon 31. (c) The exon 47 (previously described as exon 28) is skipped in cochlea and the 5' part of exon 48

(previously described as 3'-UTR exon) encodes the C-terminal 60 aa (including a transmembrane domain) of this protein. On the other hand, brain transcript possesses the sequence derived from exon 47.

Figure 12. Comparison of the deduced amino acid sequences encoded by exon 47 and exon 48. (a) human otoferlin; (b) mouse otoferlin. The predicted transmembrane domains are underlined. Identical residues are indicated by a vertical bar and residues with similarity are indicated by a colon. Figure 13: (a) murine otoferlin cDNA (double stranded) from cochlea cells; (b) murine otoferlin cDNA (double stranded) from brain cells. Figure 14: (a)-(h) sequences of deposited materials. The present invention provides nucleic acids and polypeptides related to deafness. The term «nucleic acid» includes RNA and DNA and may be single- or double-stranded. Based on this amino acid sequence of a protein described herein and the degeneracy of the genetic code one of ordinary skill in the art can immediately envision all of the nucleic acids which would encode the amino acid sequence, and, therefore, these other sequences are also within the scope of the invention. Also included within the scope of the invention are nucleic acids which encode a polypeptide which have the same activity as the polypeptides described herein. Preferably, these nucleic acid sequences hybridize to the nucleic sequence described herein under stringent conditions. Stringent conditions are, for example 0.2X SSC, 0.1% SDS, as defined in Current Protocols in Molecular Biology, Volumes 1-4, Eds. Ausubel et al, lohn Wiley and Sons, Copyright 1994-1998.

The nucleic acid may be incoφorated into a vector using well- established techniques known to those skilled in the art. The vector may be a plasmid. The plasmid preferably contains the necessary control and regulatory sequences.

The vector may be used to transform a host cell. The nature of the host cell is not particularly limited and may be any of the host cells routinely used in the field of recombinant DNA technology. The host cell may be a bacterial cell. One particularly preferred host cell is E. coli. Transformation of host cells is described in detail in Current Protocols in Molecular Biology, Volumes 1-4, Eds. Ausubel et al, John Wiley and Sons, Copyright 1994-1998. The polypeptide may be produced by culturing the transformed host cell in a liquid culture medium. The produced polypeptide may then be isolated from the medium. Such culturing procedures are also described in Current Protocols in Molecular Biology, Volumes 1-4, Eds. Ausubel et al, John Wiley and Sons, Copyright 1994-1998. EXAMPLES PARTI

The chromosomal interval defined for DFNB9

We previously reported the segregation of the DFNB9 locus in a consanguineous family (family F) living in Northern Lebanon and affected with a profound sensorineural prelingual hearing loss ¹⁴. The causative gene had been assigned to a 2 cM interval, limited by D2S2303 and D2S174, on chromosome 2 at p23. 1. A recent genetic linkage study on 30 Lebanese families presenting with deafness and large enough to provide significant lodscores, led us to identify three additional unrelated consanguineous DFNB9 affected families (AB, Kl and K2). Families F, AB and Kl were living in geographically distinct regions in Northern Lebanon, and family K2 in Southern Lebanon. The ten affected members of family F, as well as the eleven of the three other families, all suffered from a prelingual severe to profound form of sensorineural deafness. The results of the segregation analysis performed in family AB (Fig. 1), combined with the previous mapping data obtained in family F, permitted the refinement of the size of the DFNB9 interval to between markers D2S158 and D2S174, ie. approximately 1 cM (Figs. 1 and 2).

Identification of a candidate gene for DFNB9 A contig of yeast artificial chromosomes (YACs), bacterial artificial chromosomes (BACS) and PI phage artificial chromosomes (PACs) covering the defined chromosomal interval was constructed. From the size of the various inserts, and particularly that of YAC 876bl2 (720 kb) which covers the entire interval, the candidate interval was estimated to extend less than 700 kb (Fig. 2). At least two genes, HADHB (trifunctional protein P subunit) and CENPA (centromere protein A) and four expressed sequence tags (ESTs), RH1296, RH25229, RH12053 and RH26192, belonging to different transcription units, could be assigned to this interval. The two genes were not considered as candidate genes for deafness due to the putative function of their encoded proteins. The ESTs were submitted to round(s) of extension by 5'-RACE (Rapid Amplification of CDNA Ends)-PCR on total fetus MPNA. The predicted aminoacid (aa) sequences they encode were compared with those derived from the clones previously isolated from two subtracted murine cochlear cDNA libraries¹²'¹³. The deduced aa sequence of one of these extended clones, RHI2053 (isolated from infant brain), showed 89.7%) identity and 97.1% similarity with the predicted 205 aa sequence encoded by one of the murine cDNA clones, 75A2NM. These two homologous sequences detected a single 9.5 kb and 9.0 kb band on EcoRl digested human and mouse DNA, respectively. This indicates that RH12053 and 75A2NM are derived from orthologous genes. We thus considered the corresponding human gene as a promising candidate gene for DFNB9.

Northern blot analysis of human polyA⁺ RNA probed with RH12053, failed to detect a transcript in the various adult tissues tested (heart, brain, placenta, lung, liver, skeletal muscle, kidney and pancreas). Northern blot analysis of adult mouse polyA⁺ RNA probed with 75A2NM, also failed to detect a band in heart, brain, spleen, lung, liver, skeletal muscle, kidney and testis. The human cDNA was reconstructed by 5'RACE-PCR performed on total fetus mRNA (see Methods). Successive rounds of extension resulted in a 4954 bp polyA⁺ cDNA sequence. The reconstituted cDNA sequence was verified by reverse transcription (RT)-PCR and sequencing of the full- length coding region. The translation initiation site was identified within a weak Kozak consensus site¹⁵ (CAGGAGatgA) at position 227, preceded by an in- frame stop codon located 51 bp upstream. The initiation codon is followed by a 3690 bp open reading frame (ORF) and a 1038 bp 3' untranslated region (UTR) with a polyadenylation signal (AATAAA) at position 4934. Sequence analysis predicts a 1230 aa protein with a molecular mass of 140.5 kD (Fig. 3a). Because of the sequence homology detected between this protein and the spennatogenesis factor, fer-1, described in Caenorhabditis elegans ¹⁶ (and see below), the encoded protein is hereafter referred to as otoferlin (gene symbol OTOF).

A stop mutation in the DFN-B9 affected patients To search for mutations in the Otoferlin gene, its complete structure was determined. The gene was found to extend over 21 kb, and to contain 28 coding exons, one 5' LTTR exon and one 1018 hp 3'LTR exon. Primers flanking each of the coding exons and adjacent splicing sites were selected (see Table 1). All Otoferlin coding exons were amplified and sequenced in family F. A transversion, in exon 18 at position 2416, from T to A which substitutes a stop codon for a tyrosine codon (Y730X), was detected (Fig. 4). This nonsense mutation is expected to lead to a truncated 729 aa protein. The same mutation was detected in families AB, Kl and K2. This mutation was homozygous in all affected individuals (21 individuals) and heterozygous in their parents (11 individuals). It was not detected in 106 unaffected individuals living in Lebanon who were unrelated to these families. These results identify Otoferlin as the causative gene for DFNB9. The existence of the same mutation in the four unrelated DFNB9 families with different geographical origins, argues in favour of a single mutation which has spread in the Middle East. Hence, the only other DFNB 9 family reported so far, who originates from Eastern Turkey¹⁷, might carry the same mutation.

Analysis of predicted otoferlin aminoacid sequence Analysis of the deduced aa sequence of otoferlin reveals a highly hydrophobic 33 aa C-terminus, including a stretch of leucine residues which is predicted to form a transmembrane domain (aa position 1198 to 1214) (Fig. 3a). No leader peptide nor any other protein targeting signal could be detected. The rest of the protein (aa position 1 to 1197) was predicted to have a cytoplasmic location. Four putative N- glycosylation sites and 13 potential protein kinase C phosphorylation sites were found. Sequence comparisons showed that otoferlin is homologous not only to the nematode fer- 1 protein¹⁶, but also to a newly identified human protein, dysferlin. Dysferlin has recently been reported to underly Miyoshi myopathy (MM) and limb girdle muscular dystrophy type 2B (LGMD2B)¹⁸'¹⁹. Otoferlin is 38.1% and 28.0% aa identical and 64.0% and 52.9% aa similar to dysferlin and fer-1, respectively. However, dysferlin (2080 aa) and fer-I (2034 aa) are longer than human otoferlin. The three proteins are mainly similar in their C-terminal regions, ie. from aa position 196 to the C-terminus of otoferlin, whereas the N-terminal 195 aa sequence of otoferlin presents only weak similarly with dysferlin and fer- 1 (see Fig. 3b). Interestingly, three sequences with homology to C2 domains were recognized in otoferfm, namely C2A (aa position 196 to 329), C2B (aa position 709 to 833) and C2C (aa position 949 to 1104) (Fig. 3). About a hundred C2 domains have now been described. According to the three-dimensional structures established for four of them, C2 domains are composed of two four-stranded β-sheets with high structural homology (although two distinct topologies have been reported)²⁰'²¹. The C2A domain of rat synaptotagmin-1 (SYT-1 C2A) is presently the most extensively characterised. The three otoferlin C2 domains showed a 25.0% to 29.7% identity and a 55.4% to 58.6% 22 similarity with SYT-1 C2A²². Each of them contains five aspartyl residues located at positions similar to those which bind Ca²⁺ in SYT-1 C2A ²⁰'²¹ (Fig. 3).

Analysis of the inner ear expression pattern of otoferlin

We investigated the expression of the otoferlin gene in several murine tissues, by RT-PCR using primer 75A2NM-P5 (3'-UTR exon) for reverse transcription and primers 75A2MH-P3/P1 (spanning exons 25 to 27) for PCR (see Methods). An RT- PCR band was observed in all the tissues tested, i.e. the eye, cochlea, vestibule, brain, heart, skeletal muscle, liver, kidney, lung and testis. However, a strong band was only obtained in the cochlea, vestibule and brain (Fig. 5). The expression pattern of Otoferlin in the inner ear was analysed by in situ hybridization. The sections were also labelled with an antimyosinNHa antibody, which exclusively stains the sensory cells i.e., in the cochlea, the inner hair cells (IHC) and the outer hair cells (OHC) and, in the vestibular apparatus, the type I and type II hair cells. At embryonic day 19.5 (E19.5), birth (PO) and postnatal day 2 (P2), in the cochlea, an intense otoferlin labelling was observed in the IHC. The OHC and the spiral ganglion cells were faintly labelled (Fig. 6a). At P12 and P20, a strong otoferlin signal persisted in the IHCs (Fig. 6b). At El 9.5, PO and P2, in the vestibular apparatus, the neuroepithelia of the utricle (Fig. 6c), the saccule (not shown) and the semicircular canals (Fig. 6d) strongly expressed otoferlin. The sensory cells identified as type I cells, according to their morphology and their position in the neuroepithelium, intensively expressed otoferlin, whereas most of the type II sensory cells did not appear, or were faintly positive (Figs. 6e, f); no signal was detected in the supporting cells (Fig. 6). At P20, the same expression pattern was still observed in the vestibule.

Discussion

Otoferlin is the second human protein described as related to the C elegans fer-1 protein. The first one, dysferlin, was also identified on the basis of its implication in human diseases, ie. two muscular dystrophies^18,19. According to sequence analysis, these two human proteins are predicted to be C-terminus membrane-anchored cytosolic proteins and to contain C2 domains. We could identify three full C2 domains in otoferlin. The reexamination of the dysferlin sequence, in light of the deduced characteristics of otoferlin, also revealed the presence of three full C2 domains at corresponding positions (Fig. 3b). Some C2 domains bind Ca²⁺, and this binding has been shown to primarily involve aspartyl side chains which act as bidenrate, ligands for these ions. In otoferlin and dysferlin, the three common C2 domains possess these five aspartyl residues suggesting that their interactions with other molecules are Ca2+-dependent²⁰'²¹. Dysferlin has two additional sequences related to C2 domains in its Ν-terminal end (Fig. 3b); the most Ν-terminal one is only partial (lack of the predicted βl strand) and in both, some aspartyl residues are missing. The fer-1 protein presents substantial differences with otoferlin and dysferlin. The three C2 related sequences, detected at the same positions as C2A, C2B and C2C in otoferlin and dysferlin, show only weak similarity with C2 domains, and lack most of the aforementioned Ca²⁺ binding aspartyl residues (Fig. 3b); moreover the cysteine rich region located near the C-terminus of fer-1 is absent from the human proteins. According to the existence of at least two additional human ESTs with fer-1 similarity (Hs8076 from Unigene and H71264 from GenBank), the mammalian fer-1-like family should comprise several other members¹⁶; their study should help clarify the relationship between the mammalian protein family and fer-1.

The C2-domain proteins are known to interact with phospholipids and proteins²¹. They fall into two functional categories: they are implicated in either (i) the generation of the lipid second messengers involved in transduction pathways or (ii) membrane trafficking. To the first category belong the cytoplasmic phospholipases A2, to the second, several proteins such as the synaptotagmins, rabphilin 3 A, munc 13, DOC2 proteins and RIM, involved in the docking of the synaptic vesicles to the plasma membrane and/or their fusion^23"29. In this fusion process, the C2 domains can interact with the negatively charged phospholipids and proteins, as a result of the modification of their surface electrostatic potential due to Ca²⁺ binding²¹'³⁰. Along the same line, although the function ofC. elegans fer-1 is not yet entirely elucidated, this protein also seems to be required in vesicle membrane fusion. Several different fer-1 mutants have been described, and in all of them, the fusion between large vesicles (termed membranous organelles) and the plasma membrane in spermatids was defective¹⁶. Based on the well established interactions of C2 domains with phospholipids, and the impaired cellular process in C elegans fer-1 mutants, we hypothesize that otoferlin is involved in Ca^2+" triggered vesicle membrane fusions. In the inner ear, otoferlin is mainly expressed in the cochlear IHC and vestibular type I sensory hair cells. The synapses of these cells, termed ribbon synapses, have not only particular structural features (an electron-dense matrix surrounded by neurotransmitter vesicles), but also specific biochemical³¹ and functional³² characteristics. Therefore, it is tempting to speculate that otoferlin acts in synaptic vesicular trafficking. Consistently, the mouse brain at PO-P2 is one of the sites of stronger otoferlin expression. The absence of neurological symptoms in DFNB9 affected patients, however suggests that otoferlin may be dispensable in neurons. We also observed a weak expression of otoferlin in a variety of non neuronal murine tissues, suggesting that the encoded protein may in addition be implicated in a relatively ubiquitous vesicle membrane trafficking pathway. The here-postulated dual function for otoferlin is reminiscent of that described for some synaptotagmins. These proteins are known to be involved in synaptic vesicle exocytosis. However some isoforms, with an ubiquitous expression, are thought to play also a part in endocytosis³³'³⁴. The genes responsible for hearing impairement identified to date (refs.

35,36 and Website http://hgins.uia.ac.be/dnalab/hhh) have been implicated in different processes, including the maturation of the sensory hair cells for POU4F3 (refs. 37-39), the K⁺ inner ear homeostasis⁴⁰ for IsK/KCNQl (forming together a K⁺ channel in the stria vascularis)^41"43, KCNQ4 (proposed to contribute to the basolateral K⁺ conductance in outer hair cells)⁴⁴ and Cx26 (presumably involved in K⁺ recycling)^2,11, as well as the formation of tectorial membrane for TECΣ4⁸'⁴⁵'⁴⁶. So far, no deafness gene has been implicated in vesicle membrane fusion processes. However, myosin NIIA has been suggested to be involved in the trafficking of synaptic vesicles⁴⁷. Moreover, hair cells from homozygous mutant Myo7a^δj mice have been demonstrated to be resistant to aminoglycoside ototoxicity due to a defect in the intracellular accumulation of these drugs; this result indicates a role of this motor protein in the regulation of apical membrane recycling⁴⁸.

Further studies including the definition of the subcellular localisation of otoferlin, identification of its ligands and gene inactivation, should help clarify the functions of this protein and test its potential involvement in some of the same cellular trafficking processes as myosin NIIA.

Methods

Patients. Thirty unrelated large consanguineous Lebanese families affected with an isolated form of sensorineural deafness were analysed. All individuals underwent a clinical investigation including a general examination and audiometric tests. Three families were found to fulfill the criteria corresponding to the DFΝB9 form, ie. severe to profound prelingual hearing loss and segregation to the previously defined locus on chromosome 2p23.1 (ref. 14). Informed consent was given by all the subjects, adults and parents of under-aged children. The study was in accordance with the rules edicted by the French Consultative Committee for People Protection in Biomedical Research (CCPPRB) (agreement 95-08-02).

Linkage analysis. Genotyping was carried out with the polymorphic microsatellite markers from the Genethon collection. Southern and northern blot analysis. Southern blots containing EcoRl-digested mouse and human genomic DNA were hybridised (in Church buffer at 65°C) with a probe which extends from bp 3290 to 3879 on the reconstituted human Otoferlin cDNA and with a probe which covers the homologous region in the murine cDNA clone 75A2MH. Adult human and mouse multiple tissue northern blots (Clontech) were hybridised with the human and mouse cDNA probes respectively, the same as those used for the Southern blots, in Express Hyb solution (Clontech) according to the manufacturer's instructions. The filter was then exposed to Kodak BioMax X-ray film for 7 days at -80°C. Identification of YACS, BACs and PACs. YAC clones containing the polymorphic markers D2S158 and/or D2S174 were selected from the WC2.3 contig (Whitehead Institute for Biomedical Research: http://www-genome.wi.mit.edu). BAC clones containing the polymorphic markers D2S158, D2S2223, D2S2350, D2S174, and the EST RH26192 were selected from the "down to the well" human BAC pools (Genome Systems). To bridge the gaps of the BAC contig, a high density filter from the chromosome 2 PAC library LL02NP04 obtained from UK HGMP Resource Center (http://www.hgmp.mrc.ac.uk/) was hybridised with three probes: RH26192-PI derived from RH26192 cDNA clone, and the end sequences of BAC clone 88f 14 (88f 14-T7 and 88f 14-SP6) obtained by direct sequencing. Identification of Otoferlin.. Human Otoferlin cDNA was obtained by

5' RACE-PCR on total fetus polyA⁺ RNA, according to the supplier's instructions (Marathon cDNA amplification kit, Clontech); random hexamers or cDNA specific primers were used in the reverse-transcription. The reconstruction of the complete ORF of Otoferlin was checked by RT-PCR performed on total human fetus mRNA using the primers RH 12053-NGSP-5' (5'- TCCTCATGATGACCGATACTCAGG-3') and RH12053-NGSP-3' (5'-AGGGTTGAGGAACCAGACGAAGG-3'), located in the 5' UTR and 3' UTR, respectively. The amplification product was cloned into pGEM-T vector (Promega) and sequenced using the Thermo Sequenase dye terminator cycle sequencing pre-mix kit version 2.0 (Amersharn Life Science), on an ABI 377 DNA sequencer. YAC clone 876bl2 was subcloned into a λgtl 1 vector and direct sequencing of the exons and flanking introns was performed using primers derived from the Otoferlin cDNA, thus allowing the determination of the exon-intron structure of the gene. Mutation detection. Each Otoferlin coding exon was PCR-amplified on 50 ng of genomic DNA extracted from blood samples of DFNB9 affected family members, using the primers listed in Table 1. Exonuclease I and shrimp alkaline phosphatase-treated PCR products (500 ng) were sequenced using the same primers as those for PCR.

Protein sequence analysis. Sequence comparison were carried out using the BLAST (http://www.ncbi.nhn.nih.gov/BLAST/) and FASTA (http:/twww.ebi.ac.uk/searches/fasta.hml) programs. The hydrophilicity plot was performed using the Kyte-Doolittle method (http://bioinformatics.weizmann.ac.il/ hydroph/). The search for protein motifs was carried out by MOTIF (http://www.motif.genome.adjp). The PSORT II progam was used to predict the subcellular localization of the protein (http://psort.nibb. acjp:8800/). The secondary structure of the protein was predicted by the Pole Bio-informatique Lyonnais server (http://pbil.ibcp.fr/NPSA/npsa_server.html). RT-PCR. Total RNAs from two day-old mouse eye, cochlea, vestibule, brain, heart, liver and kidney, and adult mouse skeletal muscle and testis were prepared by the guanidium isothiocyanate procedure. RT-PCR was performed with 500 ng total RNA from each tissue, according to the GeneAmp RNA PCR kit protocol (Perkin Elmer Cetus). Otoferlin primer, 75A2MH-P5 (3'-UTR exon) (5'- GGGAGGCTGTAAAGGAAGA-3') was used for reverse transcription, and primers 75A2MH-P1 (5'-GCCAGGCCCACAGGGTTCTTCTC-3') (exon 27) and 75A2MH-P3 (5'-GGAGTCTATGTTCTCCTGGGATGAGAC-3') (exon 25) for PCR. It resulted in a 348 bp product. RT-PCR analysis of Gapdh (glyceraldehyde-3-phosphate dehydrogenase) was used as a positive control; oligo-dT primer was used for reverse transcription, and primers 5'-AACGGGAAGCCCATCACC-3' and 5'- CAGCCTTGGCAGCACCAG-3' for PCR amplification: it resulted in a 442 bp product. PCR reactions were run during 35 cycles.

In situ hybridization. In situ hybridization was performed using digoxigenin-11-UTP labelled RNA probes, as in ref. 49. The mouse otoferlin cDNA fragment from 75A2MH, was cloned into pGEM-T vector (Promega). Sense and antisense probes were transcribed using SP6 and T7 RNA polymerases after appropriate linearisation. After DNase I digestion, the probes were ethanol precipitated twice with 0.4 M LiCl. Mouse inner ears were fixed for 1 h at 4°C in 4% paraformaldehyde-PBS.

After three washing steps in PBS, they were immersed in 20% sucrose overnight at 4°C.

Cryostat sections (10-14 μm) were postfixed and rinsed in PBS. Following prehybridization at room temperature for at least 3 h, they were hybridised overnight at 56°C in a humid chamber. The sections were then washed and incubated with sheep anti- digoxigenin antibody coupled to alkaline phosphatase. Staining by NBT/BCIP

(Boehringer Mannheim) was performed for 2 h at 37°C and overnight at room temperature. Some sections were directly mounted in Aquatex (Merck) and observed using an optical microscope (Leica). Other sections were, in addition, stained by immunofluorescence with an antibody to myosin VIIA, as previously described⁴⁷. The care of experimental animals was in accordance with the institutional European guidelines.

GenBank accession number. Otoferlin CDNA and deduced protein sequences, AF 107403.

Table 1 Primers for PCR amplification of human otoferlin exons

Exon Forward primer (5¹ - 3") Reverse primer (5' - 3^*) Product size (bp) '-UTR GTGACGTCAGGATCTTGGC TTCAGGCCTTCTTCCTGTGA 378

1 CCTAGCGAGAGCTCCCAG TTGTCAGCGAGGGAGAGGA 389

2 TGCCAGGGCTGGGCAGAT GACAGCΓCGGGCCATGAC 261

3 GCCTGGTTGTGAGAAGGTG GGGTCTAGCCTCCTGATTG 264

4 CCTCAGGATCAGGGGGCT TGTCACTCAGGCTTCCAGC 316

5 TGGTGACCCCATGCCCAC GGCCTGGTACATGTGCGC 344

6 CTCACTTGAAGCC CACCT TCCAGTCCCCACAGGCTC 272

7 GTTCCGTGGAAGCTGAGTG TCAGCGCAGGTGGAGTGC 307

8 ATΘGCGCTGTGGTAAG6AC GCCCCAGCCTTCAGGGT 335

9 GATCATGGCAAACAACTCATG GGGGATGACAAGCCACTTC 278

10 TGGTCCAGCACCCCCATG AAGAAGGGGCAGAGGAAGC 317

11 GCCAGTGTGGCΓCCACTC CTTGGACTGGGCGGAGAC 327

12 TCTGTCCATCTATCCATCTGT GCAGATAGTCTGGTTCACAG 236

13 CCACACCCCTGTGCTAGG TCTGT ICTGATGCTGGAC 202

14 GCTGACA6ATGGCGGAATG CGTGGGAAAGAAGCTGGAC 291

15 TAAGACTGGTCTGGGCCCA AAAAGAGAAGCAGGTGATGAG 243

16 GCGACAGAGCTGGGTCTC CCCCACCTGCAGGATCAC 291

17 GGCGCAGTGATCCTGCAG GCAGTGGTGGGAGGTGAG 299

18 CTCCTGGTGCTGTTAGCTAT GATGAGGAGACTTGCAAGGAG 258

19 CCCTCAGGCAGCCTGGC AGGGCTTAAGGATTGGCTAG 293

20 ACTTCTCXΛACCCAGAACAC . AATCACXΛGGATCTGMTCTC 300

21 CCACGACCAGCTGTCATC AGGTTCCCCAGGGAAGTGT 318

22 TGTGCTCCTAAGTCCCTAGT TTCΓACXTΓTTACTGACTCAGG 304

23 TTAGGAGGGAGAGGAGAGC AGGGATGCCAACTGGCCAA 286

24 ACCCCAAACXΛCTCCTCCA CCTCCCATGCAGGGACTG 278 2S GAGGAGGCAGAGGGAAGG GATGGACTGGAAGCAATGAC 444

26 GAGCTGACACTGAGGTTGC AGAGGACAGACAGGTCCCA 303

27 AGTAGGCAGGGCTGGGAC GΠΓCΓGGGGATCGTCTCCTT 259

28 ACACAGCCCAGGACAGCTT TTCTGAGAGTCTATTCTGTCAT 551 3'-UTR TTTGGGGGTCCAGAAGGAC AACAGCCAACAGAAACCCAG 1052 References from Part I

1. Petit, C. Genes responsible for human hereditary deafness: symphony of a thousand. Nature Genet. 14,385-391 (1996).

2. Kelsell, D.P. et al. Connexin 26 mutations in hereditary non-syndromic sensorineural deafness. Nature 387, 80-83 (1997).

3. Liu, X.-Z. et al. Mutations in the myosin VIIA gene cause non-syndromic recessive deafness. Nature Genet. 16, 188-190 (1997).

4. Weil, D. et al. The autosomal recessive isolated deafness, DFNB2, and the Usher IB syndrome are allelic defects of the myosin- VIIA gene. Nature Genet. 16, 191- 193 (1997).

5. Wang, A. et al. Association of unconventional myosin MY015 mutations with human nonsyndromic deafness DFNB 3. Science 280, 1447-1451 (1998).

6. Li, X.C. et al. A mutation in PDS causes non-syndromic recessive deafness. Nature Genet. 18, 215-217 (1998). 7. Scott, D.A., Wang, R, Kerman, T.M., Sheffield, V.C. & Karniski, L.P. The Pendred syndrome gene product functions as an iodine, chloride transporter, in Molecular Biology of Hearing and Deafness (Bethesda, USA) 63 (1998).

8. Mustapha, M. et al. An α-tectorin gene defect causes a newly identified autosomal recessive form of sensorineural pre-lingual non-syndromic deafness, DFN-B21. Hum. Mol. Genet. 8,409-412 (1999).

9. Zelante, L. et al. Connexin26 mutations associated with the most common form of nonsyndromic neurosensory autosomal recessive deafness (DFNB 1) in Mediterraneans. Hum. Mol. Genet. 6, 1605-1609 (1997).

10. Denoyelle, F. et al. Prelingual deafness: high prevalence of a 3 OdelG mutation in the connexin 26 gene. Hum. Mol. Genet. 6, 2173-2177 (1997).

11. Denoyelle, F. et al. Connexin26 gene linked to a dominant deafness. Nature 393, 319320 (1098).

12. Cohen-Salmon, M., El-Amraoui, A., Leibovici, M. & Petit. C. Otogelin: a glycoprotein specific to the acellular membranes of the inner ear. Proc. Natl Acad. Sci. USA 94, 1445014455 (1997).

13. Verpy, E., Leibovici, M. & Petit, C. Characterization of otoconin-95, the major protein of murine otoconia, provides new insights into the formation of these inner ear biominerals. Proc. Natl Acad Sci. USA 96, 529-534 (1999).

14. Chaib, H. et al. A gene responsible for a sensorineural nonsyndromic recessive deafness maps to chromosome 2p22-23. Hum. Mol. Genet. 5, 155-158 (1996).

15. Kozak, M. Interpreting cDNA sequences: some insights from studies on translation. Mamm. Genome 7, 563-574 (1996).

16. Achanzar, W.E. & Ward, S. A nematode gene required for sperm vesicle fusion. J. Cell Sci. 110, 1073-1081 (1997).

17. Leal, S.M. et al. A second Middle Eastern kindred with autosomal recessive nonsyndromic hearing loss segregates DFNB9. Eur. J. Hum. Genet. 6, 341-344 (1998).

18. Liu, J. et al. Dysferlin, a novel skeletal muscle gene, is mutated in Miyoshi myopathy and limb-girdle muscular dystrophy. Nature Genet. 20, 31-36 (1998).

19. Bashir, R. et al. A gene related to Caenorhabditis elegans spertnatogenesis factor fer-1 is mutated in limb-girdle muscular dystrophy type 2B. Nature Genet. 20, 37-42 (1998).

20. Surton, R.B., Davletov, B.A., Berghuis, A.M., Sϋdhof, T.C. & Sprang, S.R. Structure of the first C2 domain of synaptotagrnin I: a novel Ca²⁺/phosphohpid- binding fold. Cell 80, 929-938 (1995).

21. Rizo, J. & Sϋdhof, T.C. C2-domain s, structure and function of a universal Ca²⁺- binding domain. J. Biol. Chem. 273,15879-15882 (1998).

22. Matthew, W.D., Tsavaler, L. & Reichardt, L.F. Identification of a synaptic vesicle-specific membrane protein with a wide distribution in neuronal and neurosecretory tissue. J Cell Biol. 91, 257-269 (198 1).

23. Perin, M.S., Fried, V.A., Mignery, G.A., Jahn, R. & Sϋdhof, T.C. Phospholipid binding by a synaptic vesicle protein homologous to the regulatory region of protein kinase C. Nature 345, 260-263 (1990).

24. Clark, J.D. et al. A novel arachidonic acid-selective cytosolic PLA2 contains a Ca(²⁺)- dependent translocation domain with homology to PKC and GAP. Cell 65, 1043-1051 (1991). 25. Shirataki, H. et al. Rabphilin-3A, a putative target protein for smg p25A/rab3A p25 small GTP-binding protein related to synaptotagrnin. Mol. Cell. Biol 13, 2061-2068 (1993). 26. Li, C. et al Synaptic targeting of rabphilin-3A, a synaptic vesicle Ca²⁺/phospholipidbinding protein, depends on rab3A/3C. Neuron 13, 885-898 (1994).

27. Brose, N., Hofmann, K., Hata, Y. & Sϋdhof, T.C. Mammalian homologues of Caenorhabditis elegans unc-13 gene define novel family of C2-domain proteins.

J BioL Chem. 270, 25273-25280 (1995).

28. Nerhage, M. et al. DOC2 proteins in rat brain: complementary distribution and proposed function as vesicular adapter proteins in early stages of secretion. Neuron 18, 453-461 (1997). 29. Wang, Y., Okamoto, M., Schmitz, F., Hofmann, K. & Sϋdhof, T.C. Rirn is a putative Rab3 effector in regulating synaptic- vesicle fusion. Nature 388, 593-598 (1997). 30. Shao, X. et al Synaptotagmin-syntaxin interaction: the C2 domain as a Ca²⁺- dependent electrostatic switch. Neuron 18, 133-142 (1997). 31. Safieddine, S. & Wenthold, R.J. Snare complex at the ribbon synapses of cochlear hair cells: analysis of synaptic vesicle- and synaptic membrane-associated proteins. Eur. J. Neurosci. 11, (in press) (1999).

32. Parsons, T.D., Lenzi, D., Aimers, W. & Roberts, W.M. Calcium-triggered exocytosis and endocytosis in an isolated presynaptic cell: capacitance measurements in saccular hair cells. Neuron 13, 875-883 (1994).

33. Li, C. et al. Ca(²⁺)-dependent and -independent activities of neural and non-neural synaptotagmins. Nature 375, 594-599 (1995).

34. Sϋdhof, T.C. & Rizo, J. Synaptotagmins: C2-domain proteins that regulate membrane traffic. Neuron 17, 379-388 (1996). 35. Avrahain, K.B. Hear come more genes! Nature Med. 4, 1238-1239 (1998).

36. Kalatzis, N. & Petit, C. The fundamental and medical impacts of recent progress in research on hereditary hearing loss. Hum. Mol. Genet. 7, 1589-1597 (1998).

37. Erkman, L. et al. Role of transcription factors Bm-3.1 and Bm-3.2 in auditory and visual system development. Nature 381, 603-606 (1996). 38. Xiang, M., Gao, W.-Q., Hasson, T. & Shin, JJ. Requirement for Brn-3c in maturation and survival, but not in fate determination of inner hair cells. Development 125, 3935-3946 (1998). 39. Vahava, 0. et al. Mutation in transcription factor POU4F3 associated with inherited progressive hearing loss in humans. Science 279, 1950-1954 (1998).

40. Spicer, S.S. & Schulte, B.A. Evidence for a medial K+ recycling pathway from the inner hair cells. Hearing Res. 118, 1-12 (1998). 41. Vetter, D.E. et al. Inner ear defects induced by null mutation of the isk gene. Neuron 17, 1251-1264 (1996). 42. Neyroud, N. et al. A novel mutation in the potassium channel gene KVLQTI causes the Jervell and Lange-Νielsen cardioauditory syndrome. Nature Genet. 15, 186-189 (1997). 43. Schulze-Bahr, E. et al. KCΝ-EL mutations cause Jervell and Lange-Νielsen syndrome. Nature Genet. 17, 267-268 (1997).

44. Kubisch, C. et al. KCΝQ4, a novel potassium channel expressed in sensory outer hair cells, is mutated in dominant deafness. Cell 96 437-446 (1999).

45. Legan, P.K., Rau, A., Keen, J.N. & Richardson, G.P. The mouse tectorins. Modular matrix proteins of the inner ear homologous to components of the spertn- egg adhesion system. J. Biol. Chem. 272, 8791-8801 (1997).

46. Verhoeven, K. et al. Mutations in the human a-tectoiin gene cause autosomal dominant non-syndromic hearing impairment. Nature Genet. 19, 60-62 (1998).

47. El-A raoui, A. et al. Human Usher IB/mouse shaker-I; the retinal phenotype discrepancy explained by the presence/absence of myosin NIIA in the photo- receptor cells. Hum. Mol. Genet. 5,1171-1178 (1996).

48. Richardson, G.P. et al. Myosin NIIIA is required for aminoglycoside accumulation in cochlear hair cells. J. Neurosci. 17, 9506-9519 (1997).

49. Schaeren-Wiemers, Ν. & Gerfm-Moser, A. A single protocol to detect transcripts of various types and expression levels in neural tissue and cultured cells: In situ hybridization using digoxigenin-labelled cRΝA probes. Histochemistry 100, 431- 440 (1993). PARTII

In our initial study of otoferlin mRΝA, we failed to detect a signal on Νorthern-blots of poly(A)+ RΝA from various adult human and mouse tissues (9). Here, using a probe derived from 3'-UTR of the mouse and human otoferlin cDΝA and upon a long exposure of the Northern blot (see Materials and Methods), a weak 5-kb band was detected in various adult human tissues (heart, placenta, liver, skeletal muscle, kidney and pancreas), and a weak 7-kb transcript in adult mouse brain (data not shown).

Characterization of cDNA encoded by mouse otoferlin gene (Otof) To explore the basis of the transcript size discrepancy between mouse and human, the cDNA clone "75A2MH" (encoding C-terminal 205 aa of mouse otoferlin) that had been obtained from a subtractive cochlear cDNA library (14, 15), was submitted to rounds of extension by 5'- and 3'-RACE (rapid amplification of cDNA ends)-PCR on mouse fetal head mRNA (see Materials and Methods below). Reconstituted cDNA sequences were verified by RT-PCR on cochlea and brain total RNA. It resulted in 6872- bp (in cochlea) and 7090-bp (in brain) poly(A)+ cDNA sequences (16, 17). The translation initiation site was identified in an appropriate Kozak consensus site (ACCAGCatgGCC) (18) at position 105, preceded by in frame stop codon located 66 bp upstream. The initiation codon is followed by a 5976 (5-991)-bp ORF and a 792 (995)-bp 3'-UTR with a polyadenylation signal (AATAAA) at position 6856 (7074) (in parenthesis for the brain form). The sequence analysis predicted a 1992 aa cochlear and a 1997 aa brain protein with molecular mass of 226.1 kD and 227.0 kD, respectively, thus 761 aa and 746 aa longer than the human otoferlin form for the cDNA previously isolated from human total fetus RNA (9); their sizes are similar to those of FER-1 (2034 aa) and dysferlin (2080 aa). Two murine cDNA clones were deposited with the C.N.C.M.

(Collection Nationale De Cultures De Microorganisms, Institut Pasteur, 28, rue du Dr Roux, 75724 Paris Cedex 15, France) on March 17, 2000 under the accession number I- 2399 and 1-2400.

The murine clones were constructed as follows: The fist clone named MOTOF 1 (1-2399) is derived from the cochlear coding sequence (GenBank AF-183183), is a fragment comprising between 65 bp and 2584 bp. The fragment is cloned in the strain XLl-Blue which contains the vector pGEM-TA and an ampicilline resistant marker.

The second clone named MOTOF 2 (1-2400) is derived from the brain coding sequence (GenBank AF-183184), is a fragment comprising between 2343 bp and 6096 bp. The fragment is cloned in the strain XLl-Blue which contains the vector pEGFP-N3 and a kana ycine resistant marker. Characterization of brain cDNA forms derived from the human otoferlin gene (OTOF)

In order to search for a long human otoferlin cDNA form, we performed 5'-RACE-PCR, with primers derived from sequence encoding aa 63 to 70 (see Materials and Methods below) in the initially reported human otoferlin, using various human mRNA from total fetus, adult heart, kidney and brain. No RACE-PCR product were obtained in all tissues tested but brain. Successive rounds of extension permitted the reconstitution of a 7156-bp poly(A)+ cDNA sequence (19). The translation initiation site we identified at position 128 (ACCAGCatgGCC) belongs to a Kozak consensus sequence (18), and is preceded by in frame stop codon located 51 bp upstream. The initiation codon is followed by a 5991-bp ORF (1997 aa) and a 1038-bp 3'-UTR with a polyadenylation signal (AATAAA) at position 7146. The sequence analysis predicted a protein with molecular mass of 226.8 kD (Fig. 8).

Analysis of the predicted human brain long otoferlin isoform The mouse and human predicted long otoferlin isoforms (both of 1997 aa) showed 94.9 % identity and 98.4 % similarity. Five sequences with homology to C2 domains were recognized in this isoform. We thus renamed the C2 domains, C2A (aa 1- 105), C2B (aa 401-537), C2C (aa 943-1076), C2D (aa 1476-1600) and C2E (aa 1716- 1870) (Figs. 8 and 9). The additional N-terminal 747 aa sequence of human otoferlin brain long isoform showed 30.7 % identity and 58.2 % similarity with the N-terminal part of dysferlin (aa 1-694); the total 1997 aa sequence presented 31.4 % identity and 54.5 % similarity with the total aa sequence of dysferlin. The sequence from aa 238-747 of the human brain long otoferlin isoform showed 18.6 % identity and 52.7 % similarity with nematode FER-1 (aa 84-546) (Fig. 9). As whole, this otoferlin isoform presented 23.2 % identity and 48.8 % similarity with FER-1.

The C2 domains are composed of two four-stranded b sheets; these domains share a high structural homology (13, 20). The extensive analysis of the C2A domain of rat synaptotagrnin- 1 (Syt-1 C2A) has permitted the recognition of five aspartyl residues which bind Ca^24"; two are located in the loop between β strand 2 and 3, and three others in the loop between β strands 6 and 7. Among the additional C2 domains detected in the predicted long otoferlin isoform, the C2A domain is only a partial one (the βl strand is missing as in dysferlin) and in addition it lacks four of the five aforementioned aspartyl residues. Amino acid sequence of the C2A domain showed 20.6 % identity and 58.9 % similarity with that of Syt-1 C2A. The C2B domain is predicted to be a full- structure C2 domain and showed 25.0 % identity and 50.0 % similarity with Syt-1 C2A. It contains all the five aspartyl residues located at the same emplacement as those that bind Ca²⁺ in Syt-1 C2A. The subsequent three C2 domains are also complete domains predicted to bind Ca²⁺ (9); they are followed by a transmembrane domain (see Fig. 9).

Eight human brain cDNA.clones were deposited with the C.N.C.M. on March 17, 2000 under the accession number 1-2401 ,1-2402, 1-2403, 1-2404, 1-2405, 1- 2406, 1-2407 and 1-2408. The human brain cDNA clones were constructed as follows: The clone named HOTOF 3 (1-2401) is derived from the human brain coding sequence (GenBank AF-183185), is a fragment comprising between 62 bp and 1299 bp. The fragment is cloned in the strain XLl-Blue which contains the vector pMOS-blue.

The clone named HOTOF 4 (1-2402) is derived from the human brain coding sequence (GenBank AF-183185), is a fragment comprising between 918 bp and 2002 bp. The fragment is cloned in the strain XLl-Blue which contains the vector pMOS-blue.

The clone named HOTOF 5 (1-2403) is derived from the human brain coding sequence (GenBank AF-183185), is a fragment comprising between 1708 bp and 2436 bp. The fragment is cloned in the strain XLl-Blue which contains the vector pGEM-TA.

The clone named HOTOF 6 (1-2404) is derived from the human brain coding sequence (GenBank AF-183185), is a fragment comprising between 2414 bp and 3282 bp. The fragment is cloned in the strain XLl-Blue which contains the vector pMOS-blue.

The clone named HOTOF 7 (1-2405) is derived from the human brain coding sequence (GenBank AF-183185), is a fragment comprising between 3124 bp and 3985 bp. The fragment is cloned in the strain XLl-Blue which contains the vector pMOS-blue. The clone named HOTOF 8 (1-2406) is derived from the human brain coding sequence (GenBank AF-183185), is a fragment comprising between 3682 bp and 4903 bp. The fragment is cloned in the strain XLl-Blue which contains the vector pGEM-TA.

The clone named HOTOF 9 (1-2407) is derived from the human brain coding sequence (GenBank AF-183185), is a fragment comprising between 4549 bp and 5656 bp. The fragment is cloned in the strain XLl-Blue which contains the vector pGEM-TA.

The clone named HOTOF 10 (1-2408) is derived from the human brain coding sequence (GenBank AF-183185), is a fragment comprising between 5358 bp and 6425 bp. The fragment is cloned in the strain XLl-Blue which contains the vector pGEM-TA. These E. coli recombinant strains were selected for using an ampicilin resistant marker.

Gene structure of OTOF and Otof

To determine the exon/intron boundaries and to estimate the size of introns, long-range PCR amplifications were performed on genomic DNA with primers selected within the cDNA sequence, and the PCR products were directly sequenced. Direct sequencing of BAC (bacterial artificial chromosome) DNA was also performed (with human BAC clone 93m21 and mouse BAC clone 233j 15). From these experiments, we could conclude that OTOF and Otof are composed of 48 coding exons (the last exon also encodes 3'-UTR and the first one also encodes 5'-UTR), extending over 90 kb and 80 kb, respectively (Fig. 7). We renamed the exons of otoferlin as shown in Table 2 (i.e. exon 20 was previously designated as exon 1). The acceptor and donor splicing sites of each exon of OTOF are listed in Table 2; all the exon/intron boundaries follow the GT/AG rule (21). Primers flanking each of the additional 19 OTOF exons and adjacent splicing sites were designed (Table 3) for the further mutation research of DFNB9 patients.

Multiple protein isoforms predicted from alternatively spliced transcripts

In all mouse tissues tested (cochlea, vestibule, brain and testis), only long cDNA forms were detected. Conversely as mentioned above, in human tissues (total fetus, adult brain, heart and kidney) only short cDNA forms were detected except in brain where both short and long cDNA forms were found. In human brain, we also isolated a short cDNA form (22), which differ from the initially described one at the level of the 5'- region; the initiation codon site was identified, with a weak Kozak consensus (TTCCTCatgATG) and is preceded by in frame stop codon 57 bp upstream. (Fig. 7). The corresponding otoferllin isoform is expected to have additional 57 N-terminal aa, which do not present any similarity to known proteins. This predicts an encoded protein of 1309 aa (148.9 kD). Based on genomic sequence analysis, we could conclude that this sequence and following 25 aa are derived from exon 20 and that previously described short cDNA form uses internal 5'-donor and 3'-acceptor splice sites as alternative splice sites (see Fig. 7 and Table 2). The 3'-acceptor splice site is common to the 3'-acceptor sites of exon 20 in long otoferlin cDNA form. In the course of Cto/cDNA extension, we also found cDNA products with skipping of exon 6. RT-PCR performed on mouse cochlea and brain total RNA using primers derived from exon 4 and exon 10, revealed two bands of 534 bp and 489 bp in cochlea, and only the shorter one in brain (Figs. 10 and 11a). Exon 6 sequence was absent from the short PCR product. RT-PCR performed on human adult brain RNA using primers derived from exon 5 and exon 8 revealed only a product with skipping of exon 6, although exon 6 could be detected on human genomic DNA sequence. This exon encodes the 5' part of the first inter C2 domain (aa position 169-184 in mouse cochlea form, AF183183).

We could also identify brain Otof cDNA products including an additional 60-bp sequence (20 aa) which is followed by the sequence encoded by exon 31 (initially named exon 12). We could ascribe this 60 bp to exon 31 and correlates the other forms with the existence of a consensus 3 '-acceptor splice site (Table 2), which can act as a cryptic splice site (23). RT-PCR performed with primers derived from Otof sequences encoded by exons 29-30 and exons 32-33 revealed 2 bands (of 282 bp and 342 bp) in cochlea and brain related to presence of either the entire or only the 3'-part of exon 31, however the longer band in both organs was weaker especially in cochlea (Figs. 10 and 1 lb). Similarly, RT-PCR using the same primer pair on human RNAs (because these sequences are almost identical between human and mouse) revealed that brain RNA express the two forms of exon 31, and the other RNAs tested (total fetus, heart and kidney) only contain the 3'-part of exon 31 (data not shown). 5' part of exon 31 submitted to alternative aplicing encodes the third inter C2 domain (aa position 1244-1264 in mouse brain form, AF 183184) Moreover, the skipping of exon 47 (previously described as exon 28) was also observed in mouse transcripts. Both of exon 47 and 5'-part of exon 48 encodes a 60 aa peptide showing simlarity and both containing a predicted transmembrane domain located at the very C-terminal end (Fig. 12). RT-PCR amplification on mouse RNA using primers derived from exon 44 and 3'-UTR (in exon 48) gave rise to a 688-bp band in cochlea and a 891-bp band in brain showing that the C-terminal end of cochlea transcript is encoded by exon 48 and that of brain by exon 47 (Figs. 10 and l ie). Similarly, RT-PCR using a primer pair derived from human exon 45 and 3'-UTR (in exon 48) showed a 749-bp band in brain and a 551-bp band in total fetus, adult heart and kidney (data not shown).

DISCUSSION

So far the expression of exon 6 has only been detected in cochlea. We could not identify any short forms in mouse mRNA, and Otof intron 19 (434 bp) did not present a similarity to OTOF intron 19 (1512 bp including an ALU sequence) (Fig. 7). Considering also the results of Northern-blot analysis, the long form is a dominant form in mouse, and the short form in human. The long form in human is considerably expressed only in brain. We do not know any other examples similar to this situation. In the sense of molecular evolution, a change of protein size within mammalians is quite an interesting matter. A switching of the C-terminal 60 aa, including a transmembrane domain, between cochlea and brain is also an interesting finding. Most of the changes in amino acid residues are conserved between human and mouse. This will surely influence the subcellar localization and the ligand selection of otoferlin.

Based on the finding of otoferlin gene structure and organization described in this paper, we are in progress on the further studies (i.e. the definition of the subcellar localization of otoferlin, identification of its ligands and gene inactivation in animal models). This attempt will help clarify the functions of this protein. MATERIALS AND METHODS Northern-blot analysis Adult human and mouse multiple-tissue northern blots (Clontech) were incubated with human and mouse cDNA probes derived from 3'-UTR sequence (at position 6722-7103 in AF183185 for human and position 6615-7005 in AF183184 for mouse), respectively, in Express Hyb solution (Clontech) according to the manufacturer's instruction. The filter was then exposed to Kodak Bio Max X-ray film for 14 days at - 80°C. The tissues tested were human heart, brain, placenta, lung, liver, skeletal muscle, kidney and pancreas, and mouse heart, brain, spleen, lung, liver, skeletal muscle, kidney and testis.

RACE-PCR, RT-PCR and long-range PCR Oligo-dT and random primed cDNA libraries were constructed using a Marathon cDNA amplification kit (Clontech) from poly(A)+ mRNA from human total fetus, adult brain, heart, kidney, and mouse fetal head. RACE-PCR performed on these libraries used linker primers and a series of primers selected from the cDNA. The amplificated products were directly cloned into pGEM-T Easy vector (Promega) and sequenced. RACE-PCR primer to detect a human long cDNA form was selected from sequence encoding aa 63-70 in the initially reported human otoferlin (5'- TTCACCTGGGCCCGCAGCATCCT-3') . Total RNA from two-day-old mouse cochlea and brain were extracted by the guanidium isothiocyanate procedure. RT-PCR were performed in various mouse and human RNA source, according to the GeneAmp RNA PCR kit protocol (Perkin Elmer Cetus). For reconstirution of mouse cDNA derived from cochlea and brain, we used two primer pairs; one is from 5'-UTR in exon 1 (5'- AGGCGTGTGAGCCACACTCCACCA-3') and exon 22 (5'-

CATAACCTCAGCTTGTCCCGAACA-3'), and the other from exon 18-19 (5'- GGCCCCAGATCACGGACAGGAAC-3') and 3'-UTR in exon 48 (5'- GGCCAGTACACCTGATTCACACT-3'). To reconstitute the additional part in human brain long cDNA form, primers derived from 5'-UTR in exon 1 (5'- GGAGGAGGCAGCGGCAGAGAAGA-3') and exon 22 (5'-TTCACCTGGGCCC GCAGCATCCT-3') were used. Alternative splicings were also detected by RT-PCR (Fig. 10). To show the skipping of exon 6, the primers from exon 4 (5'-AATCGGG TAGAGGTGACCGACAC-3') and exon 10 (5'-CCGAGCCTCAATCACTGTGATGC- 3') were selected in mouse and from exon 5 (5'-GTGGAGGTCCGGTATCAGGCCAC- 3*) and exon 8 (5'-ACACCGAGTCGGGATCCAGTCCA-3') in human. A primer pair from mouse exon 29-30 (5^,-CAAGTGGTTTGAAGTGGACCTCCC-3^,) and exon 32-33 (5'-GCCACATCCACCTTGACCACAGC-3') was used to reveal the different forms of exon 31; this pair was also used for amplification on human RNA as the human and mouse sequences were almost identical. To show the skipping of exon 47, the primers from exon 44 (5'-GGAGTCTATGTTCTCCTGGGATGAGAC-3') and 3'-UTR in exon 48 (5'-GTCTT GCTCAAGGCTGGCAGGCG-3') were selected for mouse mRNA and from exon 45 (5'-GACAGCCAAGCAGTGCACCATGG-3') and 3'UTR in exon 48 (5'- AGGCAGGCTC GGCCCAAGGCATG-3') for human mRNA.

Long-range PCR was performed using Expand Long Template PCR Systems. Selecting the PCR primers with melting temperature (Tm) more than 70°C, two-step PCR (92-94°C; 15 sec and 68°C; 4-30 min, 35 cycles) was adopted. DNA Sequencing

PCR products (500 ng) were treated with exonuclease I and shrimp alkaline phosphatase (Amersham Life Science), and subsequently sequenced using a

DYEnamic ET terminator cycle sequencing premix kit-(Amersham Pharmacia Biotech) on an ABI 377 DNA sequencer. For the direct sequencing of BAC DNA, we used a BigDye terminator RR mix (PE Applied Biosystems) .

DNA and protein sequence-analysis tools

Sequence comparison analysis was carried out using BLAST (website; http://www.ncbi.nlm.nih.gov/BLAST/) and FASTA (website; http://www.ebi.ac.uk/ searches/fasta.html). The search for protein motifs was carried out by SMART (http://smart.embl-heidelberg.de/) and Pfam (http://www.sanger.ac.uk/Software/Pfarn/).

PSORT II was used to predict the subcellular localization of the proteins

(http://psort.nibb. ac.jp:8800/). The secondary structure of the protein was predicted by the Pole Bio-informatique Lyonnais server (http://pbil.ibcp.fr/NPSA/npsa_server.html).

Putative enzyme post-translational modification sites were predicted using Prosearch2.0 program (http:^ioweb.pasteur.fr/seqanal/interfaces/prosearch-simpIe.html).

ABBREVIATIONS aa, amino acids; BAC, bacterial artificial chromosome; IHC, inner sensory hair cells; LGMD2B, limb-girdle muscular dystrophy type 2B; MM; Miyoshi myopathy; OHC, outer hair cell; ORF; open reading frame; PCR, polymerase chain reaction; RACE, rapid amplification of cDNA ends; RT, reverse transcription; Syt-1, synaptotagrnin- 1; UTR, untranslated region. The exon/intron boundaries of OTOF

Table 2

exon number ze , ,. ., ,., ,,. intron size

(previous si acceptor splice site (5 '-3') doner splice site (5 -3 ) ,.,-, designation) (bp)

1 206 λCTTTCCC3VG/gu-ggg»agccgctgxctca 11

2 59 tattCccrcctttccccacag/GGCλλTCCTT L'l XUλTGAG/g-cagg-Lgaccca-.ga.ccecs 8J

3 89 aartg ccct tcCCCCcag/ACATTTCGGT TCΛGCλACAAygτι»gcgtgggacgggggag 7

4 100 g cccctctoccatcccceg/GC CXTCGGG VT ATCλΛG/gtgggtatgccccaLCgyagt 2.5

5 182 ca rtcgcccccctcc tecag/ ACCλGCCTGT U l'iX.'UjCaG/gtaacagucauuacccctctt 6

6 45 catttettecceetactcag/CλλΛCGCλGλ ACt-ACCA A-WgCaggtigccccaggcgaggg 9

7 74 CCtjjq LtfU-H-XCCtlca rng/AGCCGGGAGG CAAAGACCλfi/gtaggaeaggtgccgcgtga 1J2

8 127 cagtccttcatc tccccag/X'H-AACDSGC CCΛACXλGCG/gτgagτggacggg*gcectg 0.492

9 55 tBβ^*c«ttc cCt:cctc«g/ATCTλλCCCA (-GATTACCAG/gtaggtgggcacctggagge 7

10 132 cceecttctcc gccgcag/CTCλGCλTCA TXACJ CGAG/gtcagtggccctgtgggga* 6

11 - 63 ctggΛCcct cccctccj-g/TA TTαrrC CλλGΛTTTCG/gcgagtggggagcagcccex 0.383

12 82 gctgtggttctgcctg-LB-g/G GAI CACT CCXGCCλG/gcgagΛCC' g g acgg c 5

13 160 cccmocccccaitcattgcig/λGCACCλTrT ACATTi-AGGa/gtgaggcccagctaccceag 0.8

14 187 ctgget.ge--tgc,ge oge«g/GAACl'l\J ^,m TCGCCλGλλG/gt»ccggggt-ιtg*ggtJ-ca 0.9

15 187 tgtgctcccccttccaceβg/GGCλλGΛCTT OOλαλ λλλQ/gCsag agcggyagac cgg 1.397

16 224 Ctt gttcccactgccacag/GCTTCCTCCC CλTCTCCC-AG/gtgagaccccaggcaegεg 0.474

17 109 tt. g* tg ctxcccacag/λGCπrrSCλQ CTCΛCCΛIλα/gttfag gctgggcccacggg 0 48

18 181 ctgggccc ttctgctccag/GCAACTATGG TCλCCGACλG/gtαggcccagceteccatcc 0.081

19 121 cu gcetctgcce-teeee»g/caAJTΛCTrC C-aCAAGCTO/gccaguuucaggccaggggc 1.512

20int (5--UTR) 199 AGΛ Cλ GΛG/gtcccccactctcccgatca

20 (1) 101 cccacactcctgaεt cag/GλλGλAGSSCC (j-^aUL^^-CTC/gτgagaagugaguugijggca 0.142

21 (2) 91 ccggctc^ecta∞cccag/CCCXTTCCTC GWSOOAGCTC/gtgagg-icgcaactggacgg 0.127

22 O) 117 cgm »cc cctcc .t gcag/αλλλλ λ. GG GG GGACGAC/gCgcggccc aggggtcggg 0.128

23 (4) 153 ggccaccceccaJicccccag/C C λGCX λ l'l'L . iaΛG grtgcLgua auo sw 034

24 ^"(5) 190 gaccccatgcccacececag/CTGCCXGGGA σiCTΛCAC λ/gtgagtgaggacccctcact 0.089

25 (6) 125 caggctgcccttccccacag/AGAAGCXGGC σrσcACACJ»G/gtgagggeetuuuauu«gUU 0.42

26 O) 135 ctecccatcccatgc gcag/CT3CrGAA_rG _GOλϊ _CCM_Cσ/gtΛtgggtggg tctggtgc 0.684

27 (8) 162 tgggttgcctgttgegtcag/GGCλAλGCTG G TCCTGCAG/gtgggactgcagggaagagg 0.402

28 (9) 120 cCgcgacccccjitirccc ag/λTTCGλC λα CCGAGTGGΛG/gtgcgBgggctc gtgtgg ' 0.426 30-1

Table 2 (followed)

exon number

(previous size acceptor splice site (5 '-3') doner splice site (5 '-3') intron size designation) (bp) (kb)

29 (10) 162 caccctccaacctctcccag/GTGCTCTTCT QTTTC-Aλ-STC/gtgagtgcaggcccfcggcgg 0.111

30 an 163 tcctgtcgccaacaccccag/GλCCl'Cα.ΛG AλCλ CλCCS/gcatggccacacccaccccg 0.482

31 131 ccatic aCceac gcccag/TCλuu TlC QCTGGACC G/gCaagvcguu guua6c*gc 0.542

3 lint (12) 71 cceccαec εcactceacag/GGCλCiTTTCT αoeu above 1.369

32 (13) 30 cctecttggttcctctgcag/ACTTCTCAAG GGTGaA,TGTG / ff gagtg -gggggccatgca ■ 0.399

33 (14) 129 tggaectggcca»cctgcag/ GCTGΛGGλGG CATGXACGAG / gcgagaccfccggg t-igggtg 2.119

34 05) 67 ttcatctcaccttttggcag/CAA T CCAC Aλ λCCCλGG/gtgagccctggaacctggaa 0.905

35 (16) 137 cctftttetgcttetggccag/GCCIGMGGG TGλ-JCTTλλG / grtgagagcctaggagcagac 0.131

36 (17) 135 ajL agcctgg-t tgrctcag/GTλlλCCCCA ACGCTTCλAG/g cagflccagg-igcacgggc 0.247

37 (18) 138 j ccc»c*cttςrtcτ.tgccc«g/GG CCC CT TCTOGTOCGG/ gtgagacteccgtcgccttc 0.637

38 (19) 13. agccctecxc ccattccag/GCCλCGαλCC TC'l'..'GU ΛA/g taggceecSCggtgcecc 0.106

39 (20) 171 tccC-UtcCCCaccccaccag/Grt TI αλC CCTλ T CAC/gtatgtggggcatggggcag 0.642

40 (21) 161 ctg gtigtecccteeca ag/ACλTOGCIAC SλCGλCλλCG/ gCaatggggcae ggatgca 0.766

41 (22) 143 tgagccgccggcaccracag/GTCΛGAGGAA CXTCGAGCAG/ g gaeacetgca.tggccaag .0.392

42 (23) 89 gccacaaceclictteetLcag/gaXGCCITJS λCCCCλλGXΛ/gCgagcgg c ggggcccag 1.287

43 (24) 99 ccttctctccctgygcccag/G ACGΛGC G TCGTGAGGGQ ,' gtggytgagcagtccc tgca 0.145

44 C5) 242, ggagacg ggcgeccacag/π'U- I'ljΛAC GλCTTCCTGG / g- gcagβgcagggcagggat 0.664

45 (26) 179 Cgac ccgag ctgc a ag/CCGCCλ αλ τσλG ICACG/gtgcgcaccccttcctgctc 0.104

46 (27) 101 gacctgtctgfcectctgcag/GGCAAGGTGG ΛCλλλ C λλ/ gtgo cgccctgccggccca 0.441

47 (28) 198 ctccaccettcacactccag/CCGGCCCOAC CGCCTCCIGα / gtctgaC CCtcttacctct 1.695

46 (.T-UTR) 1018 tccccecεtecacϊtcccag/CCGGCCCCλC

31 Table 3. Primers for PCR amplification of OTOF additional exons

exon forward primer (5' - 3') reverse primer (5' - 3') product size (bp)

1 AGGCAGCGGCAGAGAAGA CTCCCAGCCCTGTCCTAT 270

2 CACGAGGTCCCATGTTGCA CGGCCAGTGCCTGGGATT 295

3 TCCCTGGGGAGCACTGG AGGTTGGGAG GTAGGTCC 334

4 CCAAGCAGTCACAGCCC ACCTCGCCATGCATGAGAG 302

5 TCCAGTGAGGCAAGGGTGT C TGGATGTCTC CCAGAAG 434

6 TC GCAGACCTAGGCTΓGC CGACAGCCCACTCC GAG 334

7 TCTCTCA G TTGGCTCTTC GAGGGCCACGCATCACTG 304

8 TAACTCTCAGC TC GGA G TACCCAAATTCCAATCATGGC 303

9 G GC TGAG GTTTAAAGACC AGTATAGTGGATAA GCACATC 265

10 CAGATGAGAGGCAGTGGTC GCTCCΓ GACTTCCAGCCA 385

11 AGGGAGGGGCCCAACACT CTCTTTA CA GGGTCTAGAC 239

12 TCCCAC TCACCACAAAGCT CAGTCGCCAGAC G GGT 301

13 GACAGCATTTGGTTCTGCCA G 3GCAGGTGCTCTCAGC 351

14 GTGCCAGGACCCAGGAGT GA CCAGCC GTCTAC 387

15 CCCACGCCCTCACC GT TGAAGAGAGGGCATCTCACA 302

16 TCAGCACCCAGGAGCTGG CC GGGACCCAGGGACT 408

17 CGGCCTGTCTG GAGACG GAGCCTCACACTTACCACC 290

18 AAGGCAGCACCGGAT GGA CGCTCCAGGTAGGGCAG 371

19 TCC CCAC CCACCAATGC CCTCTGACAGCGCCGTCT 300

32 References from Part II

1. Denoyelle, F. , Marlin, S., Weil, D., Moatti, L., Chauvin, P., Garabedian, .-N. and Petit, C. (1999) Clinical features of the prevalent form of childhood deafness, DFNBl, due to a connexin-26 gene defect: implications for genetic counselling. Lancet, 353, 1298-1303.

2. Petit, C. (1996) Genes responsible for human hereditary deafness: symphony of a thousand. Nature Genet, 14, 385-391.

3. Kelsell, D.P., Dunlop, J., Stevens, H.P., Lench, N.J., Liang, J.N., Parry, G., Mueller, R.F. and Leigh, I.M. (1997) Connexin 26 mutations in hereditary non- syndromic sensorineural deafness. Nature, 387, 80-83.

4. Liu, X.-Z., Walsh, J., Mburu, P., Kendrick- Jones, J., Cope, M.J.T.V., Steel, K.P. and Brown, S.D.M. (1997) Mutations in the myosin VIIA gene cause nonsyndromic recessive deafness. Nature Genet., 1-6, 188-190.

5. Weil, D., Kϋssel, P., Blanchard, S., Levy, G., Levi-Acobas, F., Drira, M., Ayadi, H. and Petit, C. (1997) The autosomal recessive isolated deafness, DFNB2, and the Usher IB syndrome are allelic defects of myosin-VJiA gene. Nature Genet., 16, 191-193.

6. Wang, A., Liang, Y., Fridell, R.A., Probst, F.J., Wilcox, E.R., Touchman, J.W., Morton, C.C., Morell, R.J., Noben-Trauth, K., Camper, S.A. and Freidman, T.B. (1998) Association of unconventional myosin MY015 mutations with human nonsyndromic deafness DFNB3. Science, 280, 1447-1451.

7. Li, X.C., Everett, L.A., Lalwani, A.K., Desmukh, D., Friedman, T.B., Green, E.D. and Wilcox, E.R. (1998) A mutation in PDS causes non-syndromic recessive deafness. Nature Genet., 18, 215-217. 8. Mustapha, M., Weil, D., Chardenoux, S., Elias, S.₅ El-Zir, E., Beckmann, J.S.,

Loiselet, J. and Petit, C. (1999) An a-tectorin gene defect causes a newly identified autosomal recessive form of sensorineural pre-lingual non-syndromic deafness, DFNB21. Hum. Mol. Genet., 8, 409-412.

9. Yasunaga, S., Grati, M., Cohen-Salmon, M., El-Amraoui, A., Mustapha, M., Salem, N., El-Zir, E., Loiselet, J. and Petit, C. (1999) A mutation in OTOF, encoding otoferlin, a FER-1 -like protein, causes DFNB9, a nonsyndromic form of deafness. Nature Genet., 21, 363-369. 33

10. Achanzar, W.E. and Ward, S. (1997) A nematode gene required for sperm vesicle fusion. J. Cell Sci., 110, 1073-1081.

11. Liu, L, Aoki, M., Ilia, L, Wu, C, Ferdie M., Angelini, C, Serrano, C, Urtizberea, J.A., Hentati, F., Hamida, M.B., Bohlega, S., Culper, E.J., Amato, A.A., Bossie, K., Oeltjen, J., Bejaoui, K., McKenna-Yasek, D., Hosier, B.A., Schurr, E.,

Arahata, K., de Jong, P.J. and Brown, R.H. Jr (1998) Dysferlin, a novel skeletal muscle gene, is mutated in Miyoshi myopathy and limb girdle muscular dystrophy. Nature Genet., 20, 31-36.

12. Bashir, R., Britton, S., Strachan, T., Keers, S., Nafiadaki, E., Lako, M., Richard, I., Marchand, S., Bourg, Ν., Argov, Z., Sadeh, M., Mahjneh, I., Marconi, G.,

Passos-Bueno, M.R., Moreira, S. de E., Zatz, M., Beckmann, J.S. and Bushby, K. (1998) A gene related to Caenorhabditis elegans spermato genesis factor fer-1 is mutated in limb-girdle muscular dystrophy type 2B. Nature Genet., 20, 37-42Λ

13. Rizo, J. and Sudhof, T.C. (1998) C2 domains, structure and function of a universal Ca²⁺-binding domain. J. Biol. Chem., 273, 15879-15882.

14. Cohen-Salmon, M., El-Amraoui, A., Leibovici, M. and Petit, C. (1997) Otogelin: A glycoprotein specific to the acellular membranes of the inner ear. Proc. Natl. Acad. Sci. USA, 94, 14450-14455.

15. Nerpy, E., Leibovici, M. and Petit, C. (1997) Characterization of otoconin-95, the major protein of murine otoconia, provides insights into the formation of these inner ear biominerals. Proc. Natl. Acad. Sci. USA, 96, 529-534.

16. Yasunaga, S., Grati, M. and Petit, C. (GenBank AF183183).

17. Yasunaga, S., Grati, M. and Petit, C. (GenBank AF183184).

18. Kozak, M. (1996) Interpreting cDΝA sequences: some insights from studies on translation. Mamm.Genome, 7, 563-574.

19. Yasunaga, S., Grati, M. and Petit, C. (GenBank AF 183185).

20. Sutton, R.B., Davletov, B.A., Berghuis, A.M. and Sϋdhof, T.C. (1995) Structure of the first C2 domain of synaptotagrnin I: a novel Ca²⁺/phospholipid-binding fold. Cell, 80, 929-938. 21. Shapiro, M.B. and Senapathy, P. (1987) RΝA splice junctions of different classes of eukaryotes: Sequence statistics and functional implications in gene expression. Nucleic Acids Res., 18, 2887-2903. 34

22. Yasunaga, S., Grati, M. and Petit, C. (GenBank AF183186).

23. Green, M.R. (1986) Pre-mRNA splicing. Annu. Rev. Genet, 20, 671-708. Obviously, numerous modifications and variations of the present invention are possible in light of the above teachings. It is therefore to be understood that within the scope of the appended claims, the invention may be practiced otherwise than as specifically described herein.

The publications cited above are incoφorated herein by reference in their entirety, unless otherwise noted.

Claims

35 CLAIMS

1. A purified polynucleotide encoding the protein sequence shown in Figure 3 a.

2. A vector comprising the nucleic acid of Claim 1.

3. A host cell transformed with the vector of Claim 3.

4. A method of producing the protein sequence shown in Figure 3a, comprising culturing the transformed host cell of Claim 3 in a culture medium.

5. An isolated nucleic acid encoding the protein sequence shown in Figure 8.

6. A vector comprising the nucleic acid of Claim 5.

7. A host cell transformed with the vector of Claim 6.

8. A method of producing the protein sequence shown in Figure 8, comprising culturing the transformed host cell of Claim 7 in a culture medium.

9. An isolated protein selected from the group consisting of (A) the short human otoferlin isoform,

(B) the long human otoferlin isoform,

(C) the short murine otoferlin isoform, and

(D) the long murine otoferlin isoform.

10. A purified polynucleotide encoding the protein of Claim 9.

11. A vector comprising the nucleic acid of Claim 10.

12. A host cell transformed with the vector of Claim 11.

13. A method of producing a protein, comprising culturing the transformed host cell of Claim 7 in a culture medium.

14. A purified polynucleotide according to Claim 1, containing at position 2416 a stop codon instead of a Tyrosine residue in the non-mutated form.

15. A purified polynucleotide containing at least 12 nucleotides between the nucleotides no. 2400 and no. 2500 of the sequence shown in Figure 3a.

16. A purified polynucleotide according to Claim 15, consisting of the nucleotides between no. 2410 and 2420 of the sequence as Figure 3a.

17. A purified polynucleotide according to Claim 16, wherein position

2416, corresponding in the wild-type sequence to a codon for tyrosine, is replaced by a sequence for a stop codon at position 730 of the Figure 4. 36

18. A purified primer consisting of nucleotide 2410 to nucleotide 2420 as shown in Figure 13.

19. A process for detection of a deafness disease corresponding to a mutation of at least one nucleotide in the nucleotide sequence shown on Figure 13 or to a deletion of at least one nucleotide of the said sequence, said processing comprising: a) treating a sample from a patient containing nucleic acid to make the nucleic acid accessible to a reagent for testing the mutation deletions suspected in the sequence of the wild-type amino acid of otoferlin protein; b) contacting the nucleic acid to be tested with a probe on a primer having a mutation or deletion or with the wild-type sequence; and c) revealing the hybridization reaction.

20. A process according to Claim 19, wherein a) is followed by sequencing of the nucleic acid corresponding to the nucleotide no. 2410 to 2420.

21. A kit for detecting a mutation or deletion in the nucleotide sequence encoding the otoferlin protein containing the reagents necessary for the visualization of the mutation or deletion in the sample.

22. A kit for detecting a truncated protein containing a fragment of the sequence shown in Figure 8.

23. A kit for detecting a deafness disease corresponding to a mutation or deletion in the polynucleotide encoding the otoferlin protein.

24. A plasmid containing an insert having a part of the polynucleotide sequence encoding the otoferlin gene and deposited in CNCM on March 17, 2000 under the accession numbers 1-2399, 1-2400, 1-2401, 1-2402, 1-2403, 1-2404, 1-2405, 1-2406 or 1-2407.

25. The primers having the sequences recited in Table 1 and Table 3.