Report

Mutation of a Gene Encoding a Protein with Extracellular Matrix Motifs in Usher Syndrome Type IIa

See allHide authors and affiliations

Science  12 Jun 1998:
Vol. 280, Issue 5370, pp. 1753-1757
DOI: 10.1126/science.280.5370.1753

Abstract

Usher syndrome type IIa (OMIM 276901), an autosomal recessive disorder characterized by moderate to severe sensorineural hearing loss and progressive retinitis pigmentosa, maps to the long arm of human chromosome 1q41 between markers AFM268ZD1 and AFM144XF2. Three biologically important mutations in Usher syndrome type IIa patients were identified in a gene (USH2A) isolated from this critical region. The USH2A gene encodes a protein with a predicted size of 171.5 kilodaltons that has laminin epidermal growth factor and fibronectin type III motifs; these motifs are most commonly observed in proteins comprising components of the basal lamina and extracellular matrixes and in cell adhesion molecules.

Usher syndrome, an autosomal recessive disorder, is the most frequent cause of combined deafness and blindness in adults and affects 3 to 6% of children born with hearing impairment. The frequency of the syndrome in the United States is ∼4.4/100,000 persons. Affected individuals have sensorineural hearing deficiencies at birth and later develop progressive retinitis pigmentosa (RP). Usher syndrome is both clinically and genetically heterogeneous (1). Three forms of Usher syndrome, types I, II, and III, can be distinguished from one another clinically. Type I differs from types II and III with respect to severity of hearing loss and vestibular involvement. Type I patients are profoundly deaf, whereas type II patients have mild hearing impairment. The hearing loss of type III patients is progressive. Type I patients have no vestibular responses, and type III individuals have variable vestibular dysfunction, whereas type II patients are normal in this regard. In all cases of Usher syndrome, deafness is associated with RP. The frequency of hearing impairment within the RP population is estimated to range between 8.0 and 33.3% (2), a substantial proportion of the hearing loss being due to the Usher syndromes. In fact, of the ∼16,000 deaf and blind people in the United States, more than half are believed to have Usher syndrome. Considering the tremendous burden imposed by the loss of both major senses and the fact that Usher syndrome is the major cause of deafness and blindness in technologically developed countries, progress in understanding the underlying pathological basis of this disorder will affect a large segment of the population.

Usher syndrome type II is the most common of the three Usher syndromes. Although Hallgren originally observed that Usher type II made up only ∼10% of all Usher cases that he encountered (3), more recent research shows that type II actually accounts for over half of all Usher cases (4). In our series of 560 families with Usher syndrome, 59% have Usher syndrome type II.

At least nine distinct genetic loci have been identified as being associated with the three clinical types of Usher syndrome (5); six loci correspond to the type I phenotype and two to the type II phenotype, and one locus has been mapped for Usher type III. At present, only the USH1B gene has been identified (6), a gene that codes for the unconventional myosin VIIa. Preliminary studies have suggested a role for myosin VIIa in trafficking of vesicles in photoreceptors and cochlear hair cells (7).

Recently, we localized the USH2A gene to a 1-mega–base pair interval between markers D1S474 and AFM144XF2 and accordingly generated a yeast artificial chromosome (YAC) map to encompass the region (8). Haplotype analysis to narrow the USH2A gene critical region placed the gene between markers AFM268ZD1 and AFM144XF2 (Fig. 1) (9). Before refinement of the USH2A region, we isolated two genes from within the D1S474 to AFM144XF2 interval, neither of which exhibited mutations in their coding regions. To identify a candidate gene for Usher syndrome type IIa, we constructed a bacterial artificial chromosome (BAC) contig of ∼300 kb between AFM268ZD1 and AFM144XF2 (10). During the process of generating sequence tagged sites (STSs) from the ends of the BAC clones, GenBank BLAST search analysis revealed that BAC 133c8 contained an open reading frame (ORF) on the centromeric end that showed substantial similarity to epidermal growth factor motifs present in the laminin family of proteins (11,12). To identify additional potential exons, the BAC 133c8 was digested with Hind IIl, the fragments were subcloned into pBluescript KS+ plasmid (Stratagene), and their sequences were determined. One of the subclones also contained an ORF with a sequence similar to that of laminin epidermal growth factor (LE) motifs, suggesting that the two ORFs originated from the same gene. Polymerase chain reaction (PCR) primer pairs generated from the DNA sequences of the ORFs were used to screen a lambda gt10 human retina cDNA library by a combination of PCR and hybridization techniques. Screening more than 5 × 106 plaques, we isolated three overlapping cDNA clones. In addition to cDNA library screening, 5′ rapid amplification of cDNA ends (5′ RACE) was performed on human retinal RACE-ready cDNA from Clontech. The library screening and 5′ RACE procedures allowed us to generate a contiguous cDNA sequence of 4782 base pairs (bp) (13). Subsequent BLAST search analysis identified a GenBank expressed sequence tag (EST) obtained from a retina cDNA library (GenBank accession number AF017021) that overlapped and extended our cDNA contig to 6330 bp. The consensus sequence of the cDNA contig has a 5′ untranslated region (UTR) sequence of 370 bp and a large ORF from nucleotides 371 to 5023, followed by a 3′ UTR of 1307 bp (GenBank accession number AF055580).

Figure 1

Refinement of the USH2Agene critical region. Nine polymorphic markers known to lie in the region were typed and examined in USH2A families. Homozygosity mapping in family 652 excluded the USH2A gene from the region centromeric to AFM268ZD1 in family 652, whereas in family 983 theUSH2A gene was excluded from the telomeric side of AFM144XF2.

To prove that the isolated gene is responsible for Usher syndrome type IIa, we designed PCR primers to screen DNA from 96 unrelated type II patients who showed linkage to 1q and from 96 control subjects without previous history of hearing or visual impairment. Heteroduplex analysis identified three different mutations—2314delG, 2913delG, and 4353-54delCT—which were found exclusively among the Usher type IIa patients (14) (Fig. 2A). All mutations were observed to segregate with the Usher II phenotype and resulted in frameshift mutations with premature terminations. 2314delG causes a frameshift at codon 772, after which the ORF continues for 20 codons and ends as TAG; 2913delG starts a frameshift at codon 972, continues for 43 codons, and ends as TGA; and 4353-54delCT causes a frameshift at codon 1452 and ends as TAA 28 codons downstream. The most frequent of the three mutations was the 2314delG (Fig. 2B). Of the 96 probands, 21 tested positive for the 2314delG mutation, 8 of which were homozygous and 13 heterozygous. All but two of these individuals had northern European ancestry (Swedish, Dutch, German, or English). The two non–northern European patients were both homozygous for 2314delG; one was from Spain and the other was an African American from Nebraska, USA. Examination of various haplotypes for markers AFM143XF10, AFM268ZD1, and AFM144XF2 failed to reveal any substantial disequilibrium with the 2314delG mutation. These data suggest that the 2314delG mutation did not arise in a common ancestor, although further studies of the origins of the 2314delG mutation are warranted. The 4353-54delCT was observed in a heterozygous Usher type IIa patient, one of the few found in the Louisiana Acadian population in our series. In total, the DNA tested for mutations accounted for 18% of the ORF, whereas the Usher type IIa mutations identified above accounted for 16% of the mutations expected to be found in the 96 Usher II cases.

Figure 2

Heteroduplex analysis and DNA sequencing ofUSH2A mutations. (A) Heteroduplex gels of three families segregating deletions of the USH2A gene. Family 791 segregates for the 2314delG mutation, 1074 segregates for the 2913delG mutation, and 536 segregates for the 4353-54delCT mutation. Gel pictures are labeled minus (−) and plus (+) to indicate inclusion of an amplification product from a control CRT-1 cDNA or 133c8 BAC templates for the USH2A gene, respectively. Inclusion of control PCR product allowed detection of mutations present in the homozygous state. Note the lack of heteroduplex formation in affected individuals II-1 and II-2 from family 791 and II-1 and II-4 from family 1074 when the PCR product from the cDNA or BAC template was not added before the heteroduplex formation. (B) DNA sequences surrounding the 2314delG mutation. The top panel shows the sequence of the PCR product from the cDNA clone, and the middle and bottom panels are sequences for PCR products from I-3 and II-1 of family 791. Note the deletion of guanine in the bottom panel. The 2314delG mutation was observed by both heteroduplex analysis and direct sequencing with both forward and reverse primers in 21 USH2A probands (8 homozygotes and 13 heterozygotes) and verified by heteroduplex analysis in all available family members (15).

To investigate the expression pattern of USH2A, we performed Northern blot analysis of RNA from various tissues, using the cDNA clone CRT-1 as a hybridization probe (15). Three transcripts of ∼6.5, 5.0, and 1.9 kb were detected in the human retina, but no discernable signal was found in other tissues (Fig.3A). We confirmed expression ofUSH2A in human fetal cochlea, eye, brain, and kidney and in adult retina and Y79 retinoblastoma cells, using reverse transcriptase–PCR (RT-PCR) with PCR primer pairs (Fig. 3, C and D). The PCR primer 1STS 1336 also amplified a product from monkey retina RNA (Fig. 3D). The sequence of the RT-PCR product from monkey RNA showed high sequence similarity with the corresponding human product, only two conservative changes being found within the 109 nucleotide sequence (16).

Figure 3

Expression pattern of USH2A in fetal and adult tissues. (A) Northern blot analysis, indicating the positions of 18S and 28S rRNA. (B) Hybridization of human β-actin probe. (C) Expression of USH2A in fetal human cochlea, eye, brain, and kidney. (D) Expression of USH2A in adult human retina, monkey retina, and Y79 human retinoblastoma cells (+, reactions with RT; −, reactions without RT). The PCR reaction products were run on 4% NuSieve 3:1 agarose gel (FMC) with 100-bp markers (M) from GIBCO-BRL.

The spectrum of mutations observed in the ORF and the expression of the gene in the target tissues establish this gene as the USH2Agene. Conceptual translation of the USH2A ORF results in a polypeptide of 1551 amino acid residues with a predicted size of 171.5 kD and an isoelectric point of 7.45 (Fig.4A). A BLAST search of GenBank with the deduced USH2A protein sequence revealed 32% identity and 47% similarity between amino acid residues 300 and 1050 to all laminin family members. The polypeptide chain contains 10 LE domains, each with ∼50 amino acid residues arranged in tandem (Fig. 4B). Laminins are one of the major components forming the extracellular matrix of basement membranes in all tissues, and the LE motif is present in other extracellular matrix proteins. Sequence similarity between the USH2A protein and the laminins ends at amino acid residue 1050, and analysis of the COOH-terminal region from residues 1050 to 1551 with the Paircoil program did not identify the characteristic coiled-coil domains present in all laminins identified thus far (17). From amino acid residues 1090 to 1500, however, the USH2A protein has four tandem repeats of ∼100 amino acids residues with a sequence similar to those of a variety of proteins containing fibronectin type III (F3) repeats. The greatest similarity is with several cell adhesion proteins and receptors, including neogenin and the leukocyte common antigen-related protein (18). Analysis of the F3 repeat region of the USH2A protein with the Transmembrane prediction program (TMpred) (17) identified two statistically significant potential transmembrane helices at amino acid residues 1366 to 1383 (outside to inside helix) and 1447 to 1465 (inside to outside helix). The first 20 amino acid residues of the USH2A protein are highly hydrophobic with characteristics of a signal peptide and may represent a signal for secretion. In addition, this protein contains 18 potential N-glycosylation sites.

Figure 4

Organization of the predicted USH2A protein. (A) Predicted amino acid sequence (25) and (B) schematic representation of the USH2A protein. The region of laminin homology and LE modules is shown in yellow. Blue denotes the region containing F3 modules with cell adhesion and receptor molecule sequence similarity. The solid black vertical box at the NH2-terminus identifies the potential signal peptide, whereas the thin black lines located on the top of the illustration denote potential N-glycosylation sites. 2314delG, 2913delG, and 4353-54delCT mutations identified in USH2A patients are shown as solid arrows, and the positions of the resulting stop codons are shown as dashed arrows; a.a., amino acid.

Collectively, sequence similarity of the USH2A protein with the sequences of laminins and cell adhesion molecules, the hydrophobic NH2-terminal 20 residues, and the potential N-glycosylation sites all suggest that the USH2A protein is either a novel tissue-specific extracellular matrix protein or a cell adhesion molecule. Of potential relevance, two regions of the human retina are particularly rich in extracellular matrix proteins (19); these areas include Bruchs membrane, which is the specialized basement membrane underlying the retinal pigment epithelium (RPE), and the interphotoreceptor cell matrix (IPM). The IPM is in direct contact with both photoreceptors and the RPE and provides structural and trophic support for development and maintenance of the neural retina. Given that other extracellular matrix proteins and cell adhesion molecules are known to influence neural-glial interactions and to be involved in synapse development and stabilization, the USH2A protein may be involved in similar processes (20). Extracellular matrix proteins also play a fundamental role in the cochlea; several collagen genes, including COL1A2,COL2A1, and COL3A1, are highly expressed in the membranous labyrinth of the cochlea, both in connective tissue elements and other cell types, including inner hair cells (21). In addition, both autosomal and X-linked forms of Alport syndrome are due to mutations in collagen type IV genes (22). Furthermore, patients with Kallmann syndrome, a disorder characterized by anosmia and hypogonadism, often develop sensorineural deafness and occasionally ocular abnormalities (23). The gene for the X-linked form of this syndrome, KAL1, is thought to encode an adhesion molecule that contains four F3 repeats and may direct neuronal migration during development (24).

The USH2A gene, which expresses a putative extracellular matrix protein, and the USH1B gene, responsible for an unconventional myosin, do not appear to have obvious functional correlations; however, little is known about the physiological roles of either protein, and a common molecular pathological process could conceivably be etiologic in both forms of the syndrome.

Identification of the USH2A gene will lead to the development of a differential diagnostic tool for patients with Usher syndrome. This is particularly important in that one mutation, G2314del, seems to be especially frequent. Studies are currently under way to determine the specific tissue distribution of the USH2A protein in the cochlea and retina. Defining the molecular events that result in deafness and blindness may provide further insight into processes involved in the development and maintenance of the retina and cochlea, as well as into the physiology of vision and hearing.

  • * These authors contributed equally to this work.

  • To whom correspondence should be addressed. E-mail: jsumegi{at}mail.unmc.edu

REFERENCES AND NOTES

View Abstract

Stay Connected to Science

Navigate This Article