Association of Unconventional Myosin MYO15 Mutations with Human Nonsyndromic Deafness DFNB3

See allHide authors and affiliations

Science  29 May 1998:
Vol. 280, Issue 5368, pp. 1447-1451
DOI: 10.1126/science.280.5368.1447


DFNB3, a locus for nonsyndromic sensorineural recessive deafness, maps to a 3-centimorgan interval on human chromosome 17p11.2, a region that shows conserved synteny with mouse shaker-2. A human unconventional myosin gene, MYO15, was identified by combining functional and positional cloning approaches in searching forshaker-2 and DFNB3. MYO15 has at least 50 exons spanning 36 kilobases. Sequence analyses of these exons in affected individuals from three unrelated DFNB3 families revealed two missense mutations and one nonsense mutation that cosegregated with congenital recessive deafness.

Nonsyndromic recessive deafness accounts for about 80% of hereditary hearing loss (1). To date, 20 loci responsible for this form of deafness have been mapped and three have been identified (2-4). DFNB3, first identified in families from Bengkala, Bali, initially was mapped to a 12-centimorgan (cM) region near the centromere of chromosome 17 (5) and subsequently was refined to a 3-cM region of 17p11.2 (6). Congenital hereditary deafness in two unrelated consanguineous families from India is also linked to DFNB3(6), indicating that the contribution of DFNB3alleles to hereditary deafness is likely to be geographically widespread.

On the basis of conserved synteny and similar phenotypes, we proposed that the autosomal recessive mouse mutation shaker-2 was the homolog of DFNB3 (5, 6). In the accompanying paper, we describe the bacterial artificial chromosome (BAC)–mediated transgene correction of the deafness and circling phenotype of homozygous shaker-2 mice (7). DNA sequence analyses of this BAC revealed an unconventional myosin gene,Myo15. Myosins are a family of actin-based molecular motors that use energy from hydrolysis of adenosine triphosphate (ATP) to generate mechanical force. The classic, two-headed filament-forming myosins that provide the basis for muscle contraction are referred to as conventional myosins. Other members of the myosin superfamily, the unconventional myosins, have functions that are less well understood but in some cases are thought to mediate intracellular trafficking events (8). All myosins share a common structural organization consisting of a conserved NH2-terminal motor domain followed by a variable number of light-chain binding (IQ) motifs and a highly divergent tail. In theshaker-2 mouse, an amino acid substitution was found in a conserved residue in the motor domain of Myo15 (7). Here we report the identification of human MYO15 (9) and describe three mutations of this gene that cause hereditary deafness in three DFNB3 families (10).

To isolate MYO15, we used primers to predict exons of the mouse homolog to amplify human genomic DNA. Sequence analyses of four polymerase chain reaction (PCR) products showed 99% identity toMyo15 at the amino acid level (11), indicating that the isolated PCR products were derived from MYO15. When these human sequences were used as starting points, a partialMYO15 cDNA sequence of ∼2.3 kilobases (kb) was identified by RACE (rapid amplification of cDNA ends) and reverse transcription–PCR (RT-PCR) (12). To obtain additional MYO15 sequences, we isolated genomic clones from a human chromosome 17–specific cosmid library, and the 35.9-kb insert from one clone was completely sequenced (13) (GenBank accession number AF051976). Coding regions were identified by means of gene structure prediction programs (14), homology search (BLASTX) (15), and Pustell DNA matrix analysis (MacVector 6.0) (16), which together predicted the presence of 49 exons in this cosmid. Of these, 45 MYO15 exons have thus far been identified in cDNA clones (12, 17). A 6-base pair (bp) exon 6 was not predicted but is present in a MYO15cDNA clone (18). The longest MYO15 open reading frame deduced from the overlapping cDNAs is 4757 bp, which comprises 45 exons (Fig. 1A) (19).

Figure 1

(A) Predicted amino acid sequence of MYO15. A partial MYO15 cDNA (4757 bp) is predicted to encode 1585 amino acids with a motor domain (blue), two IQ motifs (green), and a tail region containing a MyTH4 and a talin-like domain (red). A consensus ATP binding site and two putative actin binding sites within the motor domain are indicated (red and pink, respectively). Mini-exon 6 in the motor domain is also shown (red). The corresponding mouse Myo15 sequence is shown below the human sequence. Amino acid identities are indicated by dots. The threeDFNB3 mutations (N890Y, I892F, and K1300X) are highlighted in yellow above the sequence, and the shaker-2mutation (C610Y) is highlighted in yellow below the sequence. (B) Genomic structure of MYO15. Relative positions of 50 MYO15 exons in 35.9 kb of genomic DNA sequence are indicated. Intronic sequences are drawn to scale but exons are not. Solid black or red vertical lines represent exons that have been identified in cDNA clones. Dashed vertical lines represent exons that are predicted based on GENSCAN, GRAIL, and conservation between human and mouse sequences. Alternative exon 24 (red vertical line) was found in a brain cDNA clone. Exon 24 has stop codons in all three reading frames and, if included in a transcript, would result in a MYO15 isoform with a shorter tail. The three DFNB3mutations and the domain organization of the encoded MYO15 protein are also shown.

To determine whether MYO15 maps to the DFNB3critical region, we used a primer pair derived from MYO15intron 15 to amplify DNA from somatic cell hybrid lines containing various deletions of chromosomal region 17p11.2 (6). The results demonstrated that MYO15 maps to the 3-cMDFNB3 critical region (20).

Searches of public nucleotide and protein databases with theMYO15 cDNA sequence (4757 bp) revealed no exact matches, and the highest significant matches were to actual or predicted unconventional myosins (21). In MYO15, a motor domain from codon 21 to 696 was identified by alignment against chicken skeletal muscle myosin II (GgFSK) (Fig. 1, A and B) (22). Alignments with other myosins reveal a consensus ATP binding site (GESGSGKT) (exon 5) (23) and two putative actin binding sites (exon 15 and 19) (24). Two IQ motifs adjacent to the motor domain are encoded in exon 22 (25). The tail region of MYO15 contains a myosin tail homology 4 (MyTH4) domain (26), encoded in exons 27 and 28, similar to those present in unconventional myosins ofAcanthamoeba (Myo4), Caenorhabditis elegans(Myo12/HUM-4 and HUM-6), Bos taurus (Myo10), and human (MYO7A) (27). A talin-like sequence was also found in the MYO15 tail region spanning exons 42 to 47 (26).

Myosins are classified on the basis of sequence divergence of their motor domains. To date, 14 classes have been defined (28). In a ClustalW alignment with the motor domains of MYO15 and other myosins, the highest amino acid identity was 42% with C. elegans HUM-6 (GenBank accession number U80848) and 41% with MYO7A (GenBank accession number U39226) (29). The extent of sequence divergence of MYO15 motor domain from other reported myosins qualifies MYO15 as a new branch of the myosin superfamily (myosin-XV) (7).

To search for MYO15 mutations in the Bengkala kindred and the two unrelated consanguineous Indian families (M21 and I-1924), we amplified and sequenced the 50 identified MYO15 exons and flanking intronic sequences from DNA of affected individuals (30). In each of the families, a single mutation was identified that cosegregates with deafness (Fig.2, A and B) (31). These mutations were not found in 390 chromosomes from 95 unrelated Indians and 100 unrelated Caucasians (31). In the Bengkala kindred, an A-to-T transversion (2674 A → T) in codon 892 (exon 28) is predicted to result in an Ile-to-Phe (I892F) substitution at a conserved position within the MyTH4 domain. The mutation identified in Indian family M21 (2668 A → T, exon 28) also is predicted to result in a substitution within the MyTH4 domain (Asp-to-Tyr; N890Y) and is found just two codons upstream of the mutation in the Bengkala families. Although a function for a MyTH4 domain is not known, the presence of these mutations suggests a critical role of this region of MYO15 in sensory transduction within the human inner ear. In contrast to these two missense mutations, the mutation identified in the other Indian family (I-1924) is a nonsense mutation in exon 39 (3898 A → T; K1300X) and is predicted to result in either a truncated protein or no protein at all (32).

Figure 2

(A) MYO15 mutations in the Bengkala kindred and two unrelated Indian families, M21 and I-1924. Portions of MYO15 DNA sequences are shown for an individual with normal hearing (control) and for an affected individual from each of the three DFNB3 families. The position of each mutation and the corresponding normal allele is indicated by an arrow and an asterisk. (B) Cosegregation of I892F, N890Y, and K1300X point mutations ofMYO15 with deafness in three Bengkala nuclear families and two Indian families (M21 and I-1924, respectively). Genomic DNA from individuals in these families was PCR amplified with primer pairs specific for normal (N) or mutant (M) alleles (31). Amplification products in each lane correspond to numbered individuals directly above in the pedigrees. In each family, all deaf individuals are homozygous for the mutantMYO15 allele. Their hearing parents are heterozygotes. Cosegregation of the mutant allele with deafness in the second Indian family I-1924 was demonstrated by RFLP analysis. Genomic NA was PCR amplified and digested with Xmn I. PCR products obtained from the normal alleles are digested ith Xmn I to yield fragments of 143 and 136 bp, whereas PCR products from the mutant alleles are not digested, yielding a 279-bp fragment only. All deaf individuals in family I-1924 are homozygous for the mutant allele and show the 279-bp fragment only. Obligate carriers have Xmn I-digested and undigested fragments. Each member of the threeDFNB3 families gave consent to publish unaltered family relationships, which are excerpted from the complete pedigrees published elsewhere (6).

MYO15 is expressed in human fetal and adult brain as evaluated by Northern blot analysis (Fig.3B). As shown by dot-blot analysis of poly (A)+ RNA from a variety of tissues, MYO15is expressed in ovary, testis, kidney, and pituitary gland (33,34). We also observed expression of MYO15 by RT-PCR of poly (A)+ RNA from cochlea of 18- to 22-week fetuses (Fig.3A). Sequencing of the RT-PCR product confirmed that it corresponds toMYO15.

Figure 3

Expression ofMYO15. (A) Expression of MYO15 in human fetal cochlea by RT-PCR analysis. RNA from human fetal cochlea and human placenta was reverse transcribed from an oligo(dT) primer and a portion of the first-strand cDNA was PCR amplified with primers derived from MYO15 exons 21 and 27 (37). Lane 1, RT-PCR product from 10 ng of fetal cochlea poly (A)+ RNA; lane 2, RT-PCR product from 20 ng of fetal cochlea poly (A)+ RNA; lane 3, RT-PCR product from 1 μg of total placenta RNA; lane 4, PCR amplification of a mock reverse transcription reaction (no RNA). The primers will amplify a 688-bp product from cDNA and a 2903-bp product from genomic DNA. The identity of the 688-bp RT-PCR products from fetal cochlea and placenta was confirmed by sequence analysis. The PCR was run in a 1% agarose gel with 100-bp markers (lane M) (Gibco-BRL). (B) Northern blot analysis using a MYO15 RT-PCR product from exons 29 to 47 as a probe. Each lane contains approximately 2 μg of poly (A)+ RNA from a human adult (left, lanes 1 to 8) or from fetal tissue (right, lanes 9 to 12) (MTN blots 7760-1 and 7756-1; Clontech Laboratories). Lane 1, heart; lane 2, brain; lane 3, placenta; lane 4, lung; lane 5, liver; lane 6, muscle; lane 7, kidney; lane 8, pancreas; lane 9, brain; lane 10, lung; lane 11, liver; lane 12, kidney. The most intense hybridization signals (4.2 to 5.5 kb) were observed in poly (A)+ RNA from adult and fetal brains (lanes 2 and 9). The same filters were rehybridized to a β-actin control probe for assessment of equal poly (A)+ RNA loading and transfer efficiency and are shown in the lower panel. Hybridization conditions are described in (13).

Our data show that MYO15 may be expressed in a number of tissues in addition to the inner ear. However, there is no obvious consistent clinical abnormality other than profound deafness in affected individuals from these three DFNB3 families. A possible explanation for the absence of pleiotropy of theseMYO15 mutations is the presence of functional redundancy provided by unconventional myosins expressed in other tissues. Alternatively, the region where the three DFNB3 mutations occurred may be functionally significant only in the auditory system. The isoform of MYO15 in the inner ear identified by RT-PCR does not contain exon 24, although this exon has been observed in cDNA clones derived from other tissues (35). Exon 24 contains translation stop codons in all three reading frames. Therefore,DFNB3 missense and nonsense mutations in exons 28 and 39 are not likely to have a functional consequence for MYO15isoforms that include exon 24.

We have identified three mutations of MYO15 in two geographically and ethnically diverse populations and are now in a position to evaluate the contribution of DFNB3 to hereditary deafness worldwide. Additionally, our findings demonstrate thatMYO15 encodes an essential mechanoenzyme of the auditory system. Mutations in two other unconventional myosins, Myo6 and MYO7A, also cause hereditary deafness (4, 36). This implies that unconventional myosins play crucial and nonredundant roles in auditory hair cell function. The discovery of MYO15 provides another entry point toward an integrated understanding of auditory signaling pathways.

  • * To whom correspondence should be addressed. E-mail: tfriedman{at}


View Abstract

Stay Connected to Science

Editor's Blog

Navigate This Article