Novel Null Mutations in the EYS Gene Are a Frequent Cause of Autosomal Recessive Retinitis Pigmentosa in the Israeli Population

Abstract

Purpose.: To characterize the role of EYS, a recently identified retinal disease gene, in families with inherited retinal degenerations in the Israeli and Palestinian populations.

Methods.: Clinical and molecular analyses included family history, ocular examination, full-field electroretinography (ERG), perimetry, autozygosity mapping, mutation detection, and estimation of mutation age.

Results.: Autozygosity mapping was performed in 171 consanguineous Israeli and Palestinian families with inherited retinal degenerations. Large homozygous regions, harboring the EYS gene, were identified in 15 of the families. EYS mutation analysis in the 15 index cases, followed by genotyping of specific mutations in an additional 121 cases of inherited retinal degenerations, revealed five novel null mutations, two of which are founder mutations, in 10 Israeli and Palestinian families with autosomal recessive retinitis pigmentosa (arRP). The most common mutation identified was a founder mutation in the Moroccan Jewish subpopulation. The ESTIAGE program produced an estimate that the age of the most recent common ancestor was 26 generations. The retinal phenotype in most patients was typical yet relatively severe RP, with an early age of onset and nonrecordable ERGs on presentation.

Conclusions.: The results demonstrate that EYS is currently the most commonly mutated arRP gene in the Israeli population, mainly due to founder mutations. EYS mutations were associated with an RP phenotype in all patients. The authors concluded that the gene plays only a minor role in causing other retinal phenotypes.

Retinitis pigmentosa (RP), with a worldwide prevalence of approximately 1:3500, ^1–3 is a group of hereditary degenerative diseases of the retina and is considered one of the most heterogeneous genetic diseases in humans. The disease appears with different modes of inheritance including autosomal recessive (50%–60%), autosomal dominant (30%–40%), and X-linked (5%–15%). ⁴ A total of 25 genes have so far been described to cause nonsyndromic autosomal recessive RP (arRP), and four additional loci have been identified by linkage studies (see the RetNet database at http://www.sph.uth.tmc.edu/RetNet/home.htm/ provided in the public domain by the University of Texas Houston Health Science Center, Houston, TX). In 1998, homozygosity mapping led to the identification of the RP25 locus in Spanish families. ⁵ Subsequent linkage analyses showed that additional arRP families of various origins, including Pakistani ⁶ and Chinese, ⁷ were linked to the RP25 locus. Extensive fine mapping and sequencing efforts in the RP25 region led to a refinement of the linked region ⁸ and eventually to the identification of the causative gene simultaneously by two groups that used different gene-hunting strategies. ^9,10 Abd El-Aziz et al. ¹⁰ used a multistep analysis in which they excluded 60 of the 110 genes in the RP25 region, refining the locus to a 2.67-cM region, and identified a large genomic deletion in one of the linked families. Collin et al., ⁹ on the other hand, applied homozygosity mapping in nonconsanguineous families, using genome-wide single-nucleotide polymorphism (SNP) genotyping. The RP25 gene identified by both studies, termed EYS (eyes shut homolog), spans almost 2 Mb of genomic sequence and includes 44 exons coding for a 10,475-nucleotide transcript. The gene is abundantly expressed in the human retina, ^9,10 and the protein is localized to the outer segments of the photoreceptor cells. ¹⁰ The retinal function of the EYS protein is still unknown. It is a large protein, composed of 3165 amino acids, containing a signal peptide and 28 EGF-like and 5 laminin A G-like domains. The human EYS protein is a homolog of the Drosophila eyes shut (spacemaker) protein, which is an extracellular matrix protein essential for photoreceptor development and morphology of the insect eye. ^11,12  

Eight EYS mutations, all of which are null, have been reported to date in eight arRP families. ^9,10 Most patients with EYS mutations had the clinical diagnosis of RP with an autosomal recessive inheritance pattern, ^9,10 but it is notable that one patient had a cone–rod pattern of photoreceptor loss. ⁹  

In a study conducted to estimate the prevalence of RP in the Israeli Jewish population, 341,175 individuals aged 17 to 20 years were screened for visual acuity problems, and those who had acuity less than 20/25 were further examined by an ophthalmologist. ¹³ The prevalence in this study was estimated as 1:4610. ¹³ Since many RP patients in this age group have relatively preserved visual acuities, this estimate is probably low. The Israeli and Palestinian populations serve as important genetic sources for identifying autosomal recessive disease genes since the Muslim population has been enriched with consanguineous marriages and the Jewish population has divided into isolated ethnic groups during history, leading to relatively high inbreeding levels. A study conducted in schools for the blind in the West Bank and Gaza in 1992 concluded that 44% to 85% of the affected children (depending on location) were the offspring of consanguineous marriages. ¹⁴ Only limited information is currently available on the genetic causes of nonsyndromic RP in the Israeli and Palestinian populations, with only a few mutations reported thus far in the CERKL, CRB1, NR2E3, RDH12, TULP1, and USH2A genes. ^15–19  

As part of our genetic analysis of Israeli and Palestinian families with autosomal recessive retinal degenerations, we used autozygosity mapping as the major mapping tool and identified families with arRP linked to the EYS region. Sequence analysis revealed five novel null mutations, two of which were founder mutations, in 10 families with arRP. The phenotype of most affected individuals was typical for arRP, although some clinical variation was evident. 

Methods

Patient Recruitment

The protocol adhered to the tenets of the Declaration of Helsinki, and informed consent was obtained from all patients who participated in the study, before donation of a blood sample. Blood samples for DNA analysis were obtained from the index patient as well as other affected and unaffected family members. 

Genetic Analysis

Genomic DNA was extracted from peripheral blood samples of index cases and their family members (FlexiGene DNA kit; Qiagen, Valencia, CA). Whole-genome SNP microarray analysis was performed (GeneChip Human Mapping 10K Xba 142 2.0; Affymetrix, Santa Clara, CA). Primers for the screening of EYS by PCR amplification have been reported previously. ⁹ PCR with 35 cycles was performed in a volume of 25-μL reaction. Mutation analysis was performed by direct sequencing of purified PCR products. Genotyping of specific mutations was performed as follows: Two mutations, c.403delA,406G>T,410_424del15 (exon 4) and c.4361_4362CC>AG (exon 26), were genotyped by using the restriction enzymes DdeI and MnlI, respectively. For the genotyping assay of the c.3715G>T (exon 25) mutation, we designed allele-specific primers to distinguish between the mutant and normal alleles (primer sequences available on request) and used the NlaIII enzyme for restriction analysis. The c.1211_1212insA (exon 8) and c.8218_8219delCA (exon 43) mutations were screened by sequence analysis. The possible effect of sequence changes on the splicing of the corresponding exon was estimated by using Splice Site Prediction by Neural Network at http://www.fruitfly.org/seq_tools/splice.html. ²⁰  

Estimation of Mutation Age

We used the ESTIAGE program ²¹ to calculate the age (in generations) of the most recent common ancestor of patients carrying the c.403delA,406G>T,410_424del15 mutation. ESTIAGE is a likelihood-based method that uses multilocus marker data from patients carrying the same mutation, assuming that they descended from a common ancestor who introduced the mutation. An estimate of the number of generations since the most recent common ancestor is obtained from the size of the haplotype shared by the individuals on each side of the disease locus. The method uses the haplotype information in patients carrying the mutation and in control subjects to identify the most likely positions of recombination events on the ancestral haplotype. This method has been reported to be effective in identifying a very small number of affected individuals in rare diseases. We used SNP genotyping data from 36 markers on eight chromosomes carrying the mutation and 46 population-matched control chromosomes. Allele frequencies were calculated based on the control population. 

Clinical Evaluation

Complete ophthalmic examinations were performed, including refraction, visual acuity, biomicroscopic slit lamp examination, and funduscopy. Retinal imaging included color fundus photographs, optical coherence tomography (OCT), and autofluorescence. Best corrected visual acuity was measured in each eye using an ETDRS chart and presented as a decimal ratio. In patients with visual acuity lower than 0.05, the distance at which fingers could be reliably counted was recorded. Retinal function was evaluated according to the level of patient cooperation by full-field electroretinography (ERG) or Goldmann kinetic and/or static perimetry (Humphrey; Carl Zeiss Meditec, Inc., Dublin, CA) as previously described. ²² Briefly, full-field ERGs were recorded with corneal electrodes and a computer system (UTAS 3000; LKC, Gaithersburg, MD). In the dark-adapted state, rod responses to a dim blue flash (Wratten 47b; Eastman Kodak, Rochester, NY) and mixed cone–rod responses to a white flash (2.35 cd · s/m²) were acquired. Cone responses to 30-Hz flashes of white light (9.4 cd · s/m²) were acquired under a background light of 21 cd/m². All ERG responses were filtered at 0.3 to 500 Hz, and signal averaging was used. 

Results

Autozygosity Mapping and EYS Mutation Analysis

We recruited 712 Israeli and Palestinian families with hereditary nonsyndromic retinal disease, 307 (including 121 with RP and 43 with Leber congenital amaurosis [LCA]) of which were consanguineous. In seeking to identify the causative gene in the consanguineous families, we used a 10K SNP microarray system (Affymetrix, Santa Clara, CA) for autozygosity mapping in patients from 171 of the families (82 with RP, 28 with LCA, and 61 with cone-dominated diseases). Autozygosity analysis revealed 18 patients from 15 different families (Table 1) who had a large homozygous region harboring the recently identified EYS gene and fulfilled two criteria: the genetic size of the homozygous region was at least 10 cM, and it was among the 10 largest homozygous regions in the affected individual (Table 1). Patients from three of the families (MOL0318, MOL0501, MOL0662; Fig. 1) were of Moroccan Jewish ancestry and shared a large region covered by 44 consecutive homozygous SNP markers. Sequence analysis of the 44 EYS exons and exon–intron boundaries in the 15 index cases revealed two novel homozygous null mutations in four families (Table 1, Fig. 2). Patients from the three Moroccan Jewish families were homozygous for a novel complex frameshift mutation in exon 4, c.403delA,406G>T,410_424del15 (hereafter referred to as p.Thr135LeufsX25), which causes a premature stop codon at position 160 due to a single base-pair deletion followed by a 15-bp deletion (Fig. 2A). In addition, an isolated case from a consanguineous Palestinian Muslim family (MOL0410; Fig. 1, Table 1) was homozygous for a novel nonsense mutation, c.4361_4362CC>AG (p.Ser1454X), in exon 26 of the EYS gene (Fig. 2D). Seeking to assess the frequency of these mutations among patients from the corresponding subpopulations, we screened a set of 56 Jewish families of North-African ancestry (Morocco, Libya, Tunisia, and Algeria) with RP for the p.Thr135LeufsX25 mutation and 47 Muslim families with RP for the p.Ser1454X mutation. The analysis revealed four additional index patients who were either heterozygous (TB59 III:1 and MOL0626 III:1; Fig. 1, Table 1) or homozygous (MOL0792 II:2 and MOL0806 III:7) for the p.Thr135LeufsX25 mutation. 

Table 1.

 

Data Regarding Patients Screened for Mutations in the EYS Gene

Patient Number*	Inheritance Pattern and Clinical Diagnosis	Level of Consanguinity†	Origin	Number of Consecutive Homozygous SNPs	Size of Homozygous Region (in Mb and cM)‡	Rank among Homozygous Regions (Based on Genetic Size)	Mutation 1 Mutation 2§
MOL0139–1	arRP	2:1	Moroccan Jew	169	50 Mb; 39 cM	Second	None
MOL0139–2		2:1		332	90 Mb; 70 cM	First
MOL0144–1	arRP	3:2	Oriental Jew	113	33 Mb; 27 cM	Second	None
MOL0161–1	sSTGD	2:2	Israeli Muslim	236	60 Mb; 43 cM	First	None
MOL0260–1	sRP	2:2	Israeli Muslim	277	72 Mb; 53 cM	First	None
MOL0264–1	sRP	2:2	Israeli Muslim	118	34 Mb; 16 cM	Fourth	None
MOL0318 II:13	arRP	2:2	Moroccan Jew	243	65 Mb; 72 cM	First	p.Thr135LeufsX25
MOL0318 II:9		2:2		200	55 Mb; 49 cM	First	p.Thr135LeufsX25
MOL0318 III:5		None		9	3.2 Mb; 1.5 cM	> Thirtieth
MOL0318 III:2		None		ND	ND
MOL0399–1	sLCA	2:2	Oriental Jew	396	103 Mb; 80 cM	First	None
MOL0400–1	sRP	2:2	Ashkenazi Jew	107	30 Mb; 13 cM	Third	None
MOL0410 II:1	sRP	2:2	Palestinian Muslim	122	37 Mb; 20 cM	Fifth	p.Ser1454X
							p.Ser1454X
MOL0501 II:1	arRP	2:2	Moroccan Jew	175	44 Mb; 24 cM	First	p.Thr135LeufsX25
							p.Thr135LeufsX25
MOL0615–1	sLCA	2:2	Israeli Muslim	157	42 Mb; 20 cM	First	None
MOL0626 III:1	arRP	None	Moroccan Jew	ND	ND	ND	p.Thr135LeufsX25
							p.Asn404LysfsX2
MOL0631–1	arRP	2:2	Moroccan Jew	254	70 Mb; 71 cM	First	None
MOL0640 II:1	sRP-sector	None	Iraqi Jew	ND	ND	ND	p.His2740TyrfsX27
							p.His2740TyrfsX27
MOL0652–1	arRP	2:2	Israeli Muslim	102	28 Mb; 18 cM	Fourth	None
MOL0662 IV:1	sRP	2:2	Moroccan Jew	60	20 Mb; 10 cM	Seventh	p.Thr135LeufsX25
							p.Thr135LeufsX25
MOL0788–1	sRP	2:2	Ashkenazi Jew	69	24 Mb; 12 cM	Ninth	None
MOL0792 II:2	sRP	2:2	Moroccan Jew	ND	ND	ND	p.Thr135LeufsX25
							p.Thr135LeufsX25
MOL0806 III:7	arRP	2:1	Moroccan Jew	ND	ND	ND	p.Thr135LeufsX25
							p.Thr135LeufsX25
TB24 III:2	sRP	None	Moroccan/Iraqi Jew	ND	ND	ND	p.Glu1239X
							p.His2740TyrfsX27
TB59 III:1	sRP	None	Iraqi Jew	ND	ND	ND	p.Thr135LeufsX25
							p.His2740TyrfsX27

 arRP, autosomal recessive retinitis pigmentosa; sRP, simplex retinitis pigmentosa; sSTGD, simplex Stargardt disease; sLCA, simplex Leber congenital amaurosis; ND, not done.

*  Bold type indicates those patients in whom EYS mutations were identified. The corresponding family trees are shown in Fig. 1.

†  Level of consanguinity is measured by the number of generations separating the spouse from the common ancestor (e.g., 2:2 designates first cousins; 2:1 designates marriage between an uncle and his niece).

‡  Data are based on 10K SNP analysis (see Materials and Methods).

§  The predicted effect of the mutation on the protein sequence is presented. Please see Supplementary Table S1 for additional information.

Figure 1.

 

Israeli and Palestinian families with EYS mutations. The numbers above the family trees indicate the family serial number, numbers within symbols represent the number of siblings, and numbers below symbols indicate the generation and individual numbers. Arrows indicate index cases; double horizontal line: consanguinity with an indication of the consanguinity level (e.g., 2:2 designates first-cousin marriages). For each recruited individual, the EYS genotype is depicted below the individual symbol. M1, p.Thr135LeufsX25; M2, p.Ser1454X; M3, p.Asn404LysfsX2; M4, p.His2740TyrfsX27; M5, p.Glu1239X.

Figure 2.

 

The chromatograms of the five novel EYS mutations. For each mutation, the wild-type (wt) sequence is depicted with the homozygous or heterozygous sequences when available. The mutation location is indicated by either an arrow (for a nucleotide change) or a horizontal line (for a sequence change affecting more than a single nucleotide). (A) c.403delA, 406G>T,410_424del15;p.Thr135 LeufsX25 (exon 4), (B) c.1211_1212insA; p.Asn404LysfsX2 (exon 8), (C) c.3715G>T; p.Glu1239X (exon 25), (D) c.4361_4362CC>AG; p. Ser1454X (exon 26), and (E) c.8218_8219delCA; p.His2740TyrfsX27 (exon 43).

The two heterozygous patients were from nonconsanguineous families, and we therefore assumed that the patients are compound heterozygous. Screening of the EYS gene indeed revealed novel heterozygous null mutations in these patients. Patient MOL0626 III:1 (from Moroccan Jewish ancestry) was heterozygous for a frameshift mutation, c.1211_1212insA (p.Asn404LysfsX2; Table 1, Fig. 2B), in exon 8. No additional family members were available to confirm the compound heterozygous state. To evaluate the frequency of the c.1211_1212insA mutation in North African Jewish RP patients, we sequenced exon 8 in 47 additional index RP patients, but none of them carried the c.1211_1212insA mutation or any other mutation in this exon. Patient TB59 III:1 was heterozygous for a novel frameshift mutation, c.8218_8219delCA (p.His2740TyrfsX27; Fig. 2E), in exon 43. Cosegregation analysis in family TB59 (Fig. 1) revealed that the p.Thr135LeufsX25 mutation was on the maternal allele (of Moroccan Jewish ancestry) and the c.8218_8219delCA mutation was on the paternal allele (of Iraqi Jewish ancestry). We therefore evaluated the frequency of the c.8218_8219delCA mutation in 31 Oriental Jewish RP patients (mainly from Iran, Iraq, and Afghanistan) and identified two nonconsanguineous patients of Iraqi Jewish origin who were either homozygous (MOL0640 II:1) or heterozygous (TB24 III:2) for the mutation. All the alleles carrying the c.8218_8219delCA mutation were of Iraqi Jewish origin and shared an identical haplotype within EYS, indicating a founder mutation. Sequence analysis of the remaining exons in TB24 III:2 revealed yet another novel nonsense mutation, c.3715G>T (p.Glu1239X), in exon 25 (Table 1, Fig. 2C). The mutation could not be identified in any of 31 additional RP patients from Oriental Jewish origin. 

In addition, we identified 29 sequence changes in our genetic screen, of which 14 were nonsynonymous changes which are known polymorphisms in the SNP database (Supplementary Table S1). None of these sequence changes is likely to be pathogenic. 

Haplotype Analysis and Estimation of Mutation Age

The most common EYS mutation identified was the p.Thr135LeufsX25 mutation in seven RP families, all of Moroccan Jewish ancestry, indicating a founder effect. Screening of normal individuals of Moroccan origin revealed that 1 of 94 control individuals was heterozygous. Haplotype analysis of four patients who were homozygous for this mutation revealed a shared homozygous region composed of nine SNP markers covering 3.2 Mb or 2.1 cM (Supplementary Fig. S1). The shared homozygous region in three of these patients from consanguineous families was much larger, harboring 44 markers along a region of 17 Mb (6.3 cM). Two pieces of evidence suggest that the mutation is relatively young: the relatively large size of the shared homozygous region and the fact that the mutation was found only in Moroccan Jews and not in Jews originating from neighboring countries (Algeria, Libya, and Tunisia). To estimate the mutation age, we used genotyping data from 36 SNP markers (8 within the EYS gene and 28 flanking it) on 8 chromosomes carrying the mutation (MOL0318 II:13, MOL0318 III:2, MOL0501 II:1, and MOL0662 IV:1) and 46 population-matched control chromosomes. The data set was analyzed with the ESTIAGE software (see the Methods section) and the number of generations since the most recent common ancestor was estimated as 26 (95% CI of 12 to 56 generations), spanning 650 years (based on a mean generation time of 25 years). 

Ocular Phenotype Associated with EYS Mutations

We clinically examined 15 patients with EYS mutations and performed ERG testing in 11 (Table 2). All patients with EYS mutations showed characteristic manifestations of RP with a relatively severe course of disease. Night vision, visual fields, and ERG responses were markedly impaired already at early ages. Both myopic and hyperopic refractive errors were present. Fundus appearance was typical for RP, including bone-spicule–like pigmentation, pallor of the optic discs, and narrowing of blood vessels (representative fundus mosaics of the left eyes of patient TB24 III:2 at age 63 and patient MOL0501 II:3 at age 52 are shown in Figs. 3A, 3E, respectively). Macular involvement is evident also on autofluorescence imaging (Figs. 3B, 3F) and OCT, with retinal thinning and atrophy (Figs. 3D, 3H). Visual fields in these two patients are markedly constricted (Figs. 3C, 3G). In almost all patients for whom ERG data were available (ages 19–51 years), both scotopic and photopic full-field ERG responses were extinguished, indicating severe retinal involvement (Table 2). Only in one patient (MOL0640 III:1) scotopic as well as photopic ERG responses were measurable, although also severely reduced. This patient initially presented with sector RP at the age of 25 (based on funduscopic and visual field findings) which later progressed to widespread, generalized retinal involvement. The mean 30-Hz cone flicker ERG amplitude measured at our center on the first ERG test performed in each patient was significantly lower in the EYS group than were the pooled, first test ERG amplitudes in RP patients in our population (Fig. 4). The EYS group included 11 patients with a mean age of 34 years and mean ± SD amplitude of 0.65 ± 2.17 μV. The control RP group included 240 patients with a mean age of 29.8 years and mean ± SD amplitude of 17.7 ± 22.2 μV (P < 0.01 for the cone flicker ERG amplitude difference). 

Table 2.

 

Clinical Data of Patients Carrying Mutations

Patient Number	Sex (Age, y)	Age at Diagnosis (y)	Refraction*	Visual Acuity*	ffERG* †
MOL0318 III:1	M (19)	19	NA	NA	Extinguished
					Extinguished
MOL0318 II:9	F (51)	36	−4.25/−1.00 × 35°	CF 2m	Extinguished
			−6.00/−0.75 × 175°	CF 2m	Extinguished
MOL0318 II:11	M (32)	32	−4.50/−1.00 × 35°	0.15	NP
			−6.00/−0.50 × 170°	0.15
MOL0318 II:13	F (36)	36	NA	NA	Extinguished
					Extinguished
MOL0318 III:2	M (21)	19	NA	NA	Extinguished
					NP
MOL0318 III:5	F (19)	18	NA	0.7	NP
				0.7	Extinguished
MOL0410 II:1	M (21)	21	−2.0/−0.25 × 35°	0.5	Extinguished
			−2.0/−1.00 × 140°	0.8	Extinguished
MOL0501 II:1	F (33)	33	NA	0.1	Extinguished
				0.25	Extinguished
MOL0501 II:3	F (52)	37	+1.50/−1.00 × 16°	0.06	NP
			+1.50/−0.75 × 160°	0.05
MOL0626 III:1	F (50)	50	NA	0.4	Extinguished
				0.4	Extinguished
MOL0640 II:1	M (47)	20	NA	0.5	27; 53/97; 8, 36
				0.6	29; 31/81; 7, 36
MOL0662 IV:1	M (41)	41	NA	0.15	NP
				0.15
MOL0792 II:2	M (57)	57	−0.75/−3.00 × 40°	CF 2.5m	NP
			−2.50/−1.50 × 50°	CF 1m
TB24 III:2	F (55)	40	−0.75/−0.25 × 18°	0.5	Extinguished
			+0.25/−1.50 × 132°	0.4	Extinguished
TB59 III:1	F (19)	19	+3.50/−0.75 × 160°	0.8	Extinguished
			+4.25/−0.50 × 44°	0.8	Extinguished

 ffERG, full field electroretinogram; CF, counting fingers; F, female; M, male; m, meter; NA, not available; NP, not performed.

*  For each entry, data on first line represent right eye; data on second line represent left eye.

†  ffERG including the following: rod response b-wave amplitude (normal, >200 μV); mixed cone–rod a-/b-wave (normal a-wave, >90 μV; normal b-wave, >400 μV); cone response (normal, >60 μV), implicit time (normal, ≤33 ms).

Figure 3.

 

Fundus imaging and visual fields. Representative fundus images and Goldmann visual fields of the left eyes of patient TB24 III:2 (A–D) at age 63 and patient MOL0501 II:3 (E–H) at age 52 are shown. The typical funduscopic findings of RP are present, including bone-spicule–like pigmentation, pallor of the optic discs, and narrowing of blood vessels (A, E). Macular involvement is evident also on autofluorescence imaging (B, F) and OCT, with retinal thinning and atrophy (D, H). Note the more severe foveal atrophy in patient MOL0501 II:3, which correlates with the lower visual acuity of this patient (Table 2). Visual fields were markedly constricted in both patients (C, G).

Figure 4.

 

A graph representing cone 30-Hz ERG amplitude versus the age at which ERG data were first obtained from 11 patients with EYS mutations (♦) versus 240 patients with diagnosed RP (□). Each data point represents the average cone flicker amplitude of the two eyes of each patient.

Discussion

EYS represents a major arRP gene, as can be appreciated from the widespread ethnic origins of families with arRP, due to EYS mutations thus far reported. ^9,10 In the present study, we show that mutations in EYS are currently the most frequent cause of nonsyndromic arRP in the Israeli population, mainly due to founder mutations in two ethnic groups: Moroccan and Oriental Jews. The five novel EYS mutations reported herein account for at least 7% of arRP families (10/141 families) we have thus far recruited from the Israeli and Palestinian populations. The most common mutation we identified in this study (p.Thr135LeufsX25) is a founder mutation in the Moroccan Jewish population, accounting for approximately 19% (12/64 chromosomes) of nonsyndromic arRP alleles in this population and is therefore currently the most common cause of RP in Jews of Moroccan ancestry. 

Similar to previous studies, most of our patients with EYS mutations manifested typical and rather severe arRP. One exception is an isolated case in which the patient initially presented with a milder phenotype, sector RP, and was homozygous for a null mutation in exon 43. Although one might expect that the nonsense-mediated mRNA decay (NMD) mechanism would prevent protein production of this mutant allele, it is tempting to speculate that some transcripts escape degradation, as reported in many other cases (see Holbrook et al. ²³ for review), manifesting a milder, more localized phenotype. Mutations that are likely to result in a similar effect, however, were reported to cause the typical widespread and severe RP phenotype. ⁹ Only one patient to date was reported to manifest a different retinal phenotype, cone–rod dystrophy (CRD), due to a null mutation located at the very end of the carboxyl terminus of EYS, thereby resulting in the absence of only the last 10 amino acids. His older sib, though, displayed RP, suggesting the possibility of variable intrafamilial expression of certain mutations, ⁹ probably due to other factors influencing clinical expression. We included 61 families with AR cone-dominated diseases (mainly CRD) in our autozygosity analysis, but only one (in which no EYS mutations were identified) was homozygous at the EYS locus, indicating that mutations in EYS are a rare cause of CRD in our population. The clinical variability in patients with null mutations in EYS suggests that other factors contribute to the severity of disease in these cases, as previously suggested. ⁹  

It is interesting to note that all EYS mutations thus far reported, ^9,10 as well as the five novel mutations reported here, are null mutations (either frameshift or nonsense). A large number of EYS missense changes are known (some of which appear as entries at dbSNP), but none of which is currently considered to be a cause of disease. Such missense changes, in combination with a null mutation on the counter allele, may result in a milder retinal phenotype. Another possible explanation for the lack of pathogenic missense mutations in EYS is the ability of the protein to tolerate single amino-acid alterations, affecting only a specific protein domain, due to the multiple functional domains (e.g., EGF-like and laminin A G-like domains) in the EYS protein. This phenomenon resembles the mutation spectrum of well-studied retinal genes, including CHM ²⁴ and the RPGR-ORF15 terminal exon, ^25–27 in which only null mutations have been reported so far. Such genes therefore cause a retinal phenotype only if no protein is generated. One can also assume that some of the polymorphic missense changes may result in an expressed protein with reduced activity. This makes EYS an attractive candidate for modulating retinal disease severity in cases caused by mutations in other retinal genes, mainly those encoding proteins that interact with EYS, such as PROM1. ⁹ Since EYS was only recently identified, further EYS mutation analyses are likely to result in a more accurate and comprehensive EYS mutation spectrum. 

The most common mutation that we identified (p. Thr135LeufsX25) is a founder mutation in the Moroccan Jewish subpopulation. The history of the Jewish population in North Africa (including Morocco, Algeria, Libya, and Tunisia) is ancient and complicated. The population was founded approximately 2000 to 2600 years ago and subsequently underwent several historic events that had a dramatic effect on the population size, including a few waves of immigration (during the first temple period ∼600 years BCE and in 1492 due to the expulsion from Spain), the conversion of Berber tribes to Judaism (during the 6th and 7th centuries), and the persecution of a major part of the community in 1033 and 1232. By 1948, the Jewish Moroccan population (the major North African Jewish population) was estimated to contain 270,000 individuals, most of whom immigrated to Israel once the state was established in 1948. Using the ESTIAGE program, which was designed for rare diseases, we estimated the age of the most recent common ancestor to be relatively young, 26 generations (or 650 years) ago. This young age is supported by the fact that the mutation was found in Jews originating from Morocco (and not other North African countries), and by the relatively large size of the shared homozygous region in most of the studied patients. Up to date, the estimated ages of only two founder mutations in the Moroccan Jewish population have been reported and found to be relatively ancient (2600–2700 years ago). ^28,29  

In summary, in this study we report the mutation spectrum of EYS in a cohort of Israeli and Palestinian RP patients that includes five novel mutations and shows that EYS is the most frequently mutated arRP gene currently known in the Israeli population. 

Supplementary Materials

Supplementary Table S1 - (PDF) 

Supplementary Figure S1 - (PDF) 

Footnotes

 Supported by Foundation Fighting Blindness Grant BR-GE-0607-0395-HUJ, the Stichting Wetenschappelijk Onderzoek Oogziekenhuis Prof. Dr. H. J. Flieringa, Rotterdam, and the Yedidut 1 Research Fund.

Footnotes

 Disclosure: D. Bandah-Rozenfeld, None; K.W. Littink, None; T. Ben-Yosef, None; T.M. Strom, None; I. Chowers, None; R.W.J. Collin, None; A.I. den Hollander, None; L.I. van den Born, None; M.N. Zonneveld, None; S. Merin, None; E. Banin, None; F.P.M. Cremers, None; D. Sharon, None

The authors thank the patients and their families for participating in the study, and Lina Bida, Ruhama Neis, Inbar Erdinest, and Alex Obolensky for excellent technical assistance.