Identification of candidate regions for a novel Usher syndrome type II locus

Family and clinical data

In this study, we investigated a Tunisian family with USH2. This family originates from centre of Tunisia. Two affected (1 male and 1 female aged 28 and 18 years, respectively) and six healthy family members (2 males and 4 females aged 21-61 years) attended our study. We also recruited 45 controls (22 males and 23 females aged 26-72 years) from different regions of Tunisia. Written informed consent was obtained from both parents, in accordance with the ethics committee of the University Hospital of Sfax. The pedigree was obtained upon interviews with parents (Figure 1). Clinical history and physical examinations of family members ruled out the implication of environmental factors in the etiology of HI and RP. Eight family members were subjected to audiologic examination, which consisted of otoscopic exploration and pure-tone audiometry. Testing of the vestibular system was performed by electron stagmography. Ocular examinations included fundus ophthalmoscopy, visual field examination, and Ganzfeld-electroretinogram (ERG). Blood samples were collected from eight family members. Genomic DNA was extracted from whole blood following a standard phenol-chloroform method.

Microsatellite genotyping and homozygosity mapping

For each gene and locus responsible for USH (Usher homepage) at least three microsatellite markers were selected on the basis of their map position (UCSC Genome Browser) and heterozygosity coefficient (HE; minimal HE of 0.7). Fluorescent dye-labeled microsatellite markers were genotyped for all the participating family members. Furthermore, a genome-wide scan was performed using 400 fluorescent dye-labeled microsatellite markers with an average spacing of approximately 10 cM (Prism Linkage Mapping Set, Applied Biosystems, Foster City, CA). We used the True Allele PCR Premix (Applied Biosystems) for PCR reactions according to the manufacturer’s instructions. Fluorescently labeled alleles were analyzed on an ABI Prism 3100-Avant automated DNA sequencer (Applied Biosystems).

We used homozygosity mapping to identify autozygous regions in the two affected children. Two-point and multipoint LOD scores were calculated with Genehunter software V2.1 version, using 100% penetrance, four alleles with equal allele frequencies, and no phenocopies.

Mutation analysis and single nucleotide polymorphism genotyping

Direct sequencing of candidate genes was performed using primers in the intronic regions (Table 1 and Table 2). Amplified products were directly sequenced using an ABI 3100-Avant automated DNA sequencer and Big Dye Terminator Sequencing V3.1 Kit (Applied Biosystems).

We also analyzed by direct sequencing three single nucleotide polymorphism (SNP) markers (rs155100, rs1157595, and rs1002207; dbSNP) spanning the ceramide kinase-like gene (CERKL) to check their cosegregation with the disease before proceeding to mutation screening. Primers and conditions were previously described [6].

Family and clinical data

The pedigree in Figure 1 displays a consanguineous Tunisian family segregating USH based on clinical history and audiometric and ophthalmologic tests. Audiometric test showed a moderate sloping bilateral sensorineural HI in USH patients (Figure 2). Severity of HI was similar in patients BT188 and BT189. Parents reported that HI was first noted in BT188 when the child was six years old, but observed it in BT189, when the older sibling was ten years old. For patient BT189 two audiograms were made at four-year intervals with no change in the profile (Figure 2). The father (60 years) had high frequency HI caused by bilateral presbycusis. No vestibular dysfunction was detected in both patients (BT188 and BT189) using the caloric test, nor was there any history of a delay in the age of walking. Patient BT189 reported having night blindness problem beginning at the age of 13 years. Fundus examination at in BT188 at age 14 and in BT189 at age 24 detected severe retinal degeneration. Visual fields (Goldmann targets III/4e) were significantly reduced to 5° concentric field and temporal island fields in BT189 for both eyes and 5° and 10° respectively in left and right eye in BT188. In BT189 and BT188, the nasal and temporal fields were not preserved, and only central field was maintained (Figure 3). The Ganzfeld-ERG recorded in BT189 showed an almost normal response flash visual-evoked potential in both eyes and a significant bilateral global retinal degeneration. Only cone flicker responses of less than 15% of the normal mean were recordable under photopic conditions while all other responses were below noise level (BT189), a typical finding for patients with RP (Figure 4). Nystagmus was noted in patient BT188 since her first examination at age 14 years. No other abnormalities were observed in these two patients. Taken together, the clinical signs observed in affected subjects indicate a form of USH2.

Genome-wide screening and homozygosity mapping

To localize the causative gene, we performed linkage analysis using polymorphic microsatellite markers bordering all described USH loci and genes. The USH phenotype segregating in the family was not found to be linked to the published USH loci. Table 3 shows statistical evidence for exclusion of USH2 genes. Therefore, a genome-wide screen using microsatellite markers was performed. Linkage was found with four markers D2S117 (2q32.3), D4S402 (4q26), D15S978 (15q21.1), and D15S117 (15q22.1). Additional markers were genotyped in all three regions to define the critical intervals. The homozygous region in chromosome 2 was delimited by two informative markers rs1002207 and D2S311; the chromosome 4 region was bordered by D4S1572 and D4S2938; and the chromosome 15 region stretched from D15S1039 to D15S120 (Figure 1). Analysis of these markers allowed us to refine the boundary of the critical linkage intervals respectively to 16 Mb, 25.5 Mb, and 55 Mb. In the CH15 region, six polymorphic microsatellite markers were found to be noninformative in the family (Figure 1). Investigation of the polymorphism of these repeats in the Tunisian population was performed in 40 unrelated individuals from different areas. Our results demonstrated that these microsatellites display a high degree of genetic polymorphism in the general Tunisian population. Microsatellite marker heterozygosity values were estimated using HET software version 1.8 and are as follows: 0.61 for D15S993, 0.89 for D15S153, 0.84 for D15S131, 0.86 for D15S205, 0.85 for D15S127 and 0.83 for D15S130. To rule out chromosome 15 aberrations, we performed G banded karyotype analysis on Phytohemagglutinin (PHA)-stimulated blood culture using standard procedures. Chromosome analysis of patient BT189 showed normal karyotype (data not shown).

We genotyped markers located in the three candidate regions in 40 healthy unrelated Tunisian individuals for more accurate estimation of allele frequencies and to determine the best candidate region. In the first region, we analyzed four markers (D2S148, D2S384, D2S364, and D2S117). In the first region, homozygous alleles were predominantly present in the population and the allele frequencies were 0.35 (D2S148), 0.125 (D2S384), 0.311 (D2S364), and 0.203 (D2S117). For the second region, three markers were analyzed and the frequencies of linked alleles were as follows: 0.025 for D4S2989, 0.122 for D4S402, and 0.125 for D4S2975. In contrast, the homozygous alleles of the chromosome 15 region were not frequent in controls. Allele frequencies of the polymorphic markers D15S992, D15S978, D15S117, and D15S1036 were assumed to be 0.05, 0.05, 0.1, and 0.075, respectively. These results suggest that the disease locus is most probably on chromosome 15. Multipoint LOD scores were calculated for the family data using Genehunter software. Maximum LOD scores (1.765 at θ=0) were identified for the candidate regions on chromosome 4 between D4S2989 and D4S2975, and chromosome 15 between D15S978 and D15S1036. On chromosome 2, a maximum LOD score of 1.51 was found for D2S117 microsatellite marker.

Candidate gene screening

The evaluation of the three homozygous regions revealed a large number of known and hypothetical genes (UCSC Genome Browser). More than 100 candidate genes in these three regions are expressed in the inner ear and in the retina. Although the region on chromosome 2 was not the best candidate locus (the lower LOD score and homozygous alleles of each linked marker are common in Tunisian population), we chose to investigate the CERKL gene, encoding a ceramide kinase, as candidate since it has been described to cause nonsyndromic autosomal recessive RP (RP26) [7]. The basis of this choice is that mutations in USH2A are responsible for USH2 as well as nonsyndromic recessive RP [5]. SNP (rs1157595 and rs155100) genotyping was compatible with linkage of the CERKL gene by cosegregation and homozygosity criteria (Figure 1). However, BT188 and BT189 were heterozygous for the rs1002207 (C/T), which was located at 0.8 Mb from CERKL gene. We screened this gene for mutations. Two affected children (BT188 and BT189) were compound heterozygous for two novel variants (Figure 1). The first change was a G>A (c.1073+34G>A) transition at position 34 from the donor splicing site of intron 8. The second was a c.242A>C transversion in exon 2, which leads to p.Asp81Ala substitution. Molecular modeling of the N-terminal region showed that the mutation p.Asp81Ala has no structural effect [8-13]. We detected this variant at heterozygous state in 2 out of 45 Tunisian controls. Taken together, these results exclude this variation to cause any functional defect on the encoded enzyme. Therefore, the locus on chromosome 2 was reduced to 14.7 Mb (Figure 1).

We also screened for mutations in another gene, MYO1E, encoding an unconventional myosin and representing a very good candidate on chromosome 15 locus. No nucleotide variant was detected in this gene.