Mutations in CRB1 are a Relatively Common Cause of Autosomal Recessive Early-Onset Retinal Degeneration in the Israeli and Palestinian Populations

Abstract

Purpose.: We evaluated the role of Crumbs homolog 1 (CRB1) in autosomal recessive (AR) retinal diseases in the Israeli and Palestinian populations using homozygosity mapping.

Methods.: Clinical analysis included family history, ocular examination, full-field electroretinography (ERG), and funduscopy. Molecular analysis included homozygosity mapping using whole genome single nucleotide polymorphism (SNP) arrays and mutation analysis of CRB1.

Results.: We recruited over 400 families with AR nonsyndromic retinal degenerations, including retinitis pigmentosa (RP) and Leber congenital amaurosis (LCA). SNP array analysis was performed on 175 index cases, eight of whom carried a homozygous region on chromosome 1 harboring CRB1. A subsequent CRB1 mutation analysis of the eight families, followed by screening of candidate founder mutations in the whole cohort of patients, revealed a total of 13 mutations, six of which are novel, in 15 families. Nine mutations were family-specific, and four were founder mutations identified in patients of Arab-Muslim origin, and Jews originated from Iraq and Kurdistan. Interestingly, a null mutation on at least one of the two mutated CRB1 alleles results in the LCA diagnosis, whereas patients carrying missense mutations were diagnosed with either RP or LCA. The average age at which CRB1 patients were referred to ERG testing was young (11 years). Of the 30 identified CRB1 patients, five had Coats-like exudative vasculopathy.

Conclusions.: Our data show that CRB1 mutations are a relatively frequent cause of AR early-onset retinal degeneration in the Israeli and Palestinian populations (10% of LCA families), and causes severe retinal degeneration at an early age.

Introduction

Mutations in Crumbs homolog 1 (CRB1) are known to cause severe retinal dystrophies, ranging from Leber congenital amaurosis (LCA) to retinitis pigmentosa (RP). ^1–7 LCA is the most severe nonsyndromic retinal dystrophy, characterized by blindness or severe visual impairment from birth, nonrecordable electroretinogram (ERG), nystagmus, hypermetropia, sluggish or absent pupillary responses, and oculodigital reflexes. ^2,4–6,8 In contrast, RP is considered a milder and more heterogeneous disorder, with a later age of onset. It is characterized by night blindness followed by gradual loss of peripheral vision, progressive degeneration of photoreceptors, and eventually leads to visual impairment of variable severity that in rare cases can result in complete blindness. ^9–12 Patients with RP have impaired ERG responses with a rod > cone pattern of injury, and over time suffer characteristic funduscopic findings, including bone spicule–like pigmentary (BSP) changes, attenuation of retinal vessels, and waxy pallor of the optic discs. ^9–12  

Retinal dystrophies resulting from CRB1 mutations can be accompanied by additional specific features, including relative preservation of the para-arteriolar retinal pigment epithelium (PPRPE) and Coats-like vasculopathy. ^{1–5,8,13,14} RP with PPRPE is a form of RP characterized by preservation of the RPE that is adjacent to the retinal arterioles, while the rest of the RPE layer degenerates. Coats-like exudative vasculopathy is characterized by abnormal retinal vessels with increased permeability, leading to exudative retinal detachment that often is accompanied by massive subretinal lipid deposits. ^{3–5,7,13,15}  

CRB1 is a human homologue of the drosophila transmembrane crumbs protein, and is expressed in the brain and in the inner segments of mammalian photoreceptors. ^{2,3,7,16–19} The Crumbs protein is implicated in mechanisms that control cell-cell adhesion, intracellular communication and apicobasal cell polarity. For epithelial cells and photoreceptors, separation of their apical and basal compartments is critical for proper development and function of the cells and the tissue, including adhesion and signaling between and within cells. ^{2,7,13,16–19} Jacobson et al. suggested that CRB1 mutations underlie developmental defects in LCA, including thickening of the retina and lack of distinct layering in the fully developed adult retina. ¹³ Rashbass et al. postulated that CRB1 has a role in localizing phototransduction proteins to the apical membrane of the photoreceptors. ¹⁸ Thus, nonfunctional CRB1 may impede phototransduction, and lead to progressive dystrophy of the photoreceptors and the RPE, resulting in LCA or RP. 

CRB1 is composed of 12 exons, and may give rise to four known isoforms, the longest of which consists of 1406 amino acids and contains 19 epidermal growth factor (EGF)–like domains, three laminin A globular (AG)–like domains, and a signal peptide in the extracellular region, of which some residues are conserved throughout evolution. In addition, there is a C-terminal transmembrane domain. The intracellular region that has a highly conserved role in organizing a macromolecular protein scaffold contains a conserved FERM binding domain, which is involved in localizing proteins to the plasma membrane, and a PDZ binding motif (PBM), used in the process of anchoring to the cytoskeleton. ^{2,3,7,16,18,19} Although no binding partners for the extracellular region have been identified so far to our knowledge, almost all identified CRB1 mutations affect this region. ^1,3–5,8,14 Missense mutations affecting cysteine residues are predicted to affect the secondary structure of the protein due to disruption of disulfide bridges. ⁵ Laminin AG–like domains have been identified in different proteins; these domains have some residues that are highly conserved, and can serve as protein–protein interaction modules. Thus, missense mutations in those conserved residues might affect protein interaction. ²⁰ In addition, missense mutations in the laminin AG–like and EGF-like domains are predicted to affect protein–protein interactions, calcium binding, and protein folding ^2,16,19 and, thus, may affect proper retinal formation and/or function. 

In most cases, retinal dystrophies, such as RP and LCA, are inherited in an autosomal recessive (AR) manner. ^6,11,14 The Israeli and Palestinian populations are characterized by a relatively high rate of consanguineous marriages and marriages within specific ethnic groups, leading to a high proportion of AR diseases. ^21,22 In our study, we analyzed Israeli and Palestinian families with AR retinal degenerative diseases for CRB1 mutations, and show that homozygosity mapping is a powerful tool to identify causative genes and mutations in this population. 

Methods

Patient Recruitment and Clinical Evaluation

The tenets of the Declaration of Helsinki were followed and, before donation of a blood sample for DNA analysis, informed consent was obtained from patients and family members who participated in this study. Clinical evaluation included a detailed family history, full ophthalmologic exam, electrooculography (EOG), full-field electroretinography (FFERG), color vision testing using the Ishihara 38-plate test and Farnsworth D-15 panel, optical coherence tomography (OCT), color and infrared fundus photos, autofluorescence (AF) imaging, and fluorescein angiography (FA), performed as previously described. ²³ Briefly, FFERGs were recorded using monopolar corneal electrodes (Henkes-type; Medical Workshop B.V., Groningen, The Netherlands) and a computerized system (UTAS 3000; LKC, Gaithersburg, MD). Cone responses to 30 Hz flashes of white light were acquired under a background light of 21 cd/m². Scotopic responses, including a rod response to a dim blue flash and a mixed cone–rod response to an ISCEV standard white flash, were acquired following 30 to 45 minutes of dark adaptation. Between 2 and 4 sets of responses were recorded in each condition to verify repeatability. All ERG responses were filtered at 0.3 to 500 Hz, and signal averaging was used. 

Genetic Analysis

We enrolled 468 families with autosomal recessive LCA or RP in the study. Genomic DNA was extracted from peripheral blood of the participants using FlexiGene DNA kit (Qiagen, Valencia, CA). Whole genome single nucleotide polymorphism (SNP) analysis was performed using different Affymetrix platforms, including 10K, 250K, and Affymetrix 6.0 arrays. 

Primers were designed using the PRIMER3 program (available in the public domain at http://frodo.wi.mit.edu/) for mutation screening of the 12 CRB1 exons and exon-intron boundaries (NCBI Reference Sequence NM_201253.2) by PCR amplification (see Supplemental Material and Supplementary Table S1). PCR was performed in a 30 μL reaction with 35 cycles. Mutation analysis was performed by direct Sanger sequencing of purified PCR product. We also used the Asper Ophthalmic (available in the public domain at http://www.asperbio.com/asper-ophthalmics) LCA and RP APEX-based mutation detection arrays to screen for known mutations in some of the samples we analyzed. The possible pathogenicity of missense changes was evaluated using PolyPhen-2 (available in the public domain at http://genetics.bwh.harvard.edu/pph2/), MutationTaster (available in the public domain at http://www.mutationtaster.org/), and SIFT (available in the public domain at http://sift.jcvi.org/). 

Results

Homozygosity Mapping and Mutation Analysis

As part of our ongoing effort to characterize the set of genes that cause inherited retinal degenerations in the Israeli and Palestinian populations, we recruited 468 index cases with clear AR (328 families) or suspected AR (140 families) nonsyndromic early-onset retinal degenerations (EORD), including mainly RP, LCA, and cone–rod degeneration (CRD). We performed a preliminary mutation analysis using the Asper LCA and ARRP microarrays to screen 39 of the index patients, and identified a previously reported CRB1 mutation in only one case. Aiming to identify the causative genes in our cohort, we focused our analysis on 237 of the families in which the parents of the index case were consanguineous, as well as families in which the parents were unrelated, but shared the same ethnic origin. We performed homozygosity analysis using whole-genome SNP microarrays in 175 of the families (see Supplemental Material and Supplemental Table S2). A large homozygous region harboring the CRB1 gene on chromosome 1 was identified in 10 patients who belong to nine families (Table 1). Homozygosity mapping in each consanguineous patient revealed a relatively high number (8–48, an average of 17) of large homozygous regions (>10 Mb) per patient. Ranking of the genomic regions by size revealed that the region harboring the CRB1 gene ranged from first to 17th (an average ranking of 4.6). Sequencing analysis of the 12 CRB1 exons and exon-intron boundaries in the nine index cases revealed disease-causing mutations in six families, with an average CRB1 region size ranking of 3.7 (Table 1). A subsequent screen for the identified mutations in additional patients (based on the ethnic group in which each mutation was identified, Table 2) followed by CRB1 screening for a second mutation on the counter allele when needed, revealed a total of 13 mutations (Table 2), 6 of which were novel, in 15 families, 9 of which were consanguineous. The analysis scheme is presented as a flowchart in Supplementary Figure S1 (see Supplemental Material). 

Table 1. 

 

Homozygosity Mapping Data Using Whole Genome SNP Analysis

Patient #	Consanguinity*	N of Homozygous Regions†	Rank	Range, Mb	CRB1 Mutation
MOL0026-2	3:3	48	17	10.0–52.4	None
MOL0097-1	2:2	12	2	10.4–39.1	c.2236C > T
MOL0329-1	2:3	12	1	10.1–40.3	None
MOL0436-1	2:2	15	1	10.3–39.2	None
MOL0492-3	2:2	14	1	10.5–56.8	c.3307G > A
MOL0521-1	2:2	8	3	12.2–65.5	c.1842delT
MOL0650-1	2:2	20	1	10.8–40.1	c.2498G > A
MOL0695-1	2:2	14	5	10.8–46.8	c.3307G > A
MOL0757-1 and MOL0757-2	2:2	3	3	11.0–20.0	c.1148G > A

*  The number of generations that are needed to reach the common ancestor from each parent (for example, 2:2 denoted a first-cousin marriage).

†  A homozygous region is defined as a genomic region containing consecutive homozygous SNPs with a minimal physical distance of 10 Mb.

Table 2. 

 

Mutations Identified in CRB1 in Patients with Retinal Degenerations

Mutation	Clinical Diagnosis	Exon	Ethnic Group	N of Families	N of Screened Index Cases	Reference
c.424G > T, (p.Gly142*)	LCA	2	Jews-mixed origin	1	63	Novel
c.455G > A, (p.Cys152Tyr)	Early RP / LCA	2	Arab-Muslim	1	63	Novel
c.1148G > A, (p.Cys383Tyr)	Early RP + Coats	5	Jews-Kurdistan	3	60	Lotery et al.39
c.1576C > T, (p.Arg526*)	LCA	6	Jews-mixed origin	1	54	Seong et al.24
c.1735T > A, (p.Val578Glu)	LCA	6	Arab-Muslim	1	54	den Hollander et al.5
c.1842delT, (p.Gly614Glyfs*6)	LCA	6	Arab-Muslim	1	54	Novel
c.1846G > T, (p.Gly615Val)	RP / CRD	6	Arab-Muslim	2	54	Novel
c.2236C > T, (p.Thr745Met)	RP	7	Arab-Muslim	1	61	den Hollander et al.3
c.2498G > A, (p.Gly833Asp)	RP	7	Jews-Iraq	2	61	Novel
c.2555T > C , (p.Ile852Thr)	Early RP / LCA	7	Arab-Muslim	1	61	Hanein et al.34
c.2678-2682del5bpCCAAC, (p.Ser893Serfs*14)	LCA	8	Jews-Kurdistan	1	61	Novel
c.3307G > A, (p.Gly1103Arg)	Early RP	9	Arab-Muslim	2	32	Hanein et al.34
c.4121-4130del10, (p.Ala1374Glufs*20)	LCA	12	Arab-Muslim	1	36	Gerber et al.33

The six novel mutations included 3 missense c.455G > A, c.1846G > T, and c.2498G > A (p.Cys152Tyr, p.Gly615Val, and p.Gly833Asp, respectively), two frameshift c.1842delT and c.2678-2682del5bpCCAAC (p.Gly614Glyfs*6 and p.Ser893Serfs*14, respectively), and one nonsense c.424G > T (p.Gly142*). The three missense mutations were perfectly conserved among mammals (Fig. 1) and yielded a very high score for pathogenicity using online prediction programs (PolyPhen-2, MutationTaster, and SIFT). 

Figure 1

 

Evaluation of three novel CRB1 missense mutations. Evolutionary conservation of the altered amino acid of three novel missense mutations in CRB1 is presented at the top. The altered residues (marked in yellow) are fully conserved through all species. Accession numbers of protein sequences are as follows: homo sapiens (NP_957,705.1), chimpanzee (XP_525,009.2), dog (XP_547,379.3), mouse (NP_573,502.2), chicken (XP_003,641,718.1), and zebrafish (NP_001,038,408.1). The aa type is color-coded: small aa in red, acidic in blue, basic in magenta, and hydroxyl + amine + basic in green. The predicted effect of the mutation on the protein as estimated by online prediction programs, such as PolyPhen-2, MutationTaster, and SIFT, is depicted at the bottom. For each mutation, a very high score of pathogenicity for the altered amino acid was obtained.

Nine of the 13 mutations were family-specific, while four were shared among patients of the same ethnicity and, therefore, were likely to be founder mutations. Two of the shared mutations were found in Oriental Jews: c.1148G > A (p.Cys383Tyr) was identified in three families of Kurdistan Jewish origin and a novel missense mutation (c.2498G > A, p.Gly833Asp) was found in two Iraqi Jewish families. The remaining two shared mutations were found in Arab Muslims: A missense mutation (c.1846G > T, p.Gly615Val) was identified in a Palestinian family and in a family from the northern part of Israel (the Galilee). The second mutation (c.3307G > A, p.Gly1103Arg) was identified in two families from the vicinity of Jerusalem and haplotype analysis of the two index cases revealed an identical homozygous region of 21Mb harboring the CRB1 gene, indicating a true founder mutation in this population. Interestingly, an isolate patient (MOL0996-1) of mixed Jewish origin (his grandparents immigrated to Israel from Kurdistan, Greece, Poland, and Yemen) was a compound heterozygote for two nonsense mutations, one of which (p.Gly142*) was novel and the second (p.Arg526*) was reported in a Korean patient. ²⁴  

Clinical Analysis of Patients with CRB1 Mutations

We identified in our study 15 families, including 30 patients with CRB1 mutations. We examined clinically 20 of these patients and all of them were diagnosed with EORD (either LCA or RP). Visual acuity was decreased in all patients, ranging from 6/15 to light perception only, with no clear correlation with age, sex, or mutation type. Hyperopia was identified in the vast majority of cases (13 of 14 patients in whom refraction was available). Full field ERG amplitudes (including light-adapted cone flicker, dark-adapted rod, and mixed rod-cone responses) ranged from severely reduced to extinguished (see Supplemental Material and Supplemental Table S3), with a rod > cone pattern of involvement. The average age in which CRB1 patients were referred to ERG testing was relatively young (10.5 years, ranging from 6 months to 53 years), and the patients had an average 30Hz cone flicker amplitude of 9 μV (ranging from 0–23 μV, lower limit of normal being 60 μV). We compared the cone flicker 30 Hz ERG amplitudes of patients with CRB1 mutations to those who suffer from nonsyndromic RP or LCA due to mutations in other genes (Fig. 2). The group of CRB1 patients (average amplitude of 9μV) had on average a similar response to the group of patients with mutations in known RP genes in our population (including DHDDS, EYS, FAM161A, and MAK, for an average amplitude of 8.4 μV), but at a younger age (mean 10.5. vs. 31 years in the RP group). CRB1 patients had a higher response than the group of patients with mutations in known LCA genes (including AIPL1, GUCY2D, and RPE65, for an average amplitude of 2.5 μV). 

Figure 2. 

 

Cone flicker 30 Hz ERG amplitude versus age. Cone flicker 30 Hz ERG amplitude (log scale) versus the age at which ERG data were first obtained from 15 CRB1 patients (black hollow dots) versus 55 patients with mutations in other LCA- and RP-causing genes (AIPL1, DHDDS, EYS, FAM161A, GUCY2D, MAK, and RPE65). Each data point represents the average cone flicker amplitude of the two eyes of each patient.

Retinal imaging of CRB1 patients (Fig. 3) showed a wide range of phenotypes. Patient MOL0492-3 showed early stage of disease at the age of 7 years. Her fundus findings included pigmentary changes, macular atrophy (Figs. 3A, 3B), and tortuous veins (Figs. 3C–E). A more severe form of retinal degeneration is evident in two sisters of family MOL0757 (Figs. 3F–N). The index case (MOL0757-1) had EORD with Coats-like exudative vasculopathy, which also was evident in four additional patients in our study. The two sisters had atrophic changes and disorganization of the retina with pigmentary alterations (Figs. 3F–I), as well as tortuous and engorged retinal vessels in the periphery of each eye (Fig. 3F) accompanied by exudative retinal detachment (Fig. 3I). Fluorescein angiography (Fig. 3F) showed RPE atrophy as well as blocked fluorescence from pigmentary lesions in the posterior pole of each eye. OCT of the two sisters showed retinal edema with disorganized retinal layers and an epiretinal membrane in both eyes (Figs. 3J–N). An example of a more severe CRB1-associated retinal phenotype is demonstrated in patient MOL0082-1 (Fig. 3O). This patient had Coats-like disease in his left eye at the third decade of life, with massive subretinal lipid deposits and exudative retinal detachment. The patient was treated with laser photocoagulation, cryotherapy, and scleral buckling with drainage of subretinal fluid, eventually leading to flattening of the retina. 

Figure 3. 

 

Ocular phenotype of patients with CRB1 mutations. (A–E) Color fundus (A,B) and fluorescein angiography (C–E) of a 7-year-old patient (MOL0492-3) showing pigmentary changes, tortuous blood vessels, and macular atrophy. (F–K) Fluorescein angiogram (F), color fundus imaging (G–I), and OCT (J,K) of a 25-year-old patient (MOL0757-1) showing RP with Coats disease. Blocked fluorescence, pigment clumps, and an epiretinal membrane are apparent in the FA and color images, while the OCT shows an epiretinal membrane, cystoid macular edema, and disorganization of the retina with loss of the normal layering. (L–N) OCT of a 22-year-old patient (MOL0757-2) showing atrophic and disorganized retinal layers. (O) Fundus imaging of a 33-year-old patient showing massive subretinal lipid accumulation caused by a Coats-like disease component, along with pigmentary retinopathy.

Genotype–Phenotype Correlation in CRB1

Aiming to examine a possible genotype–phenotype correlation in CRB1, we collected information on 15 patients reported in our study, as well 173 previously reported patients (116 with LCA, 24 with other EORD, and 48 with RP, with or without PPRPE or Coats-like exudative vasculopathy). ^{1,3–5,8,13,14,25–47} We grouped the patients into the following subgroups (Fig. 4): Group 1 included 35 patients who were homozygous or compound heterozygous for expected null mutations (i.e., nonsense, frameshift, and splice-site mutations). Group 2 included 53 patients who were heterozygous for a missense mutation on one allele and an expected null mutation on the counter allele. Group 3 included 100 patients who were homozygous or compound heterozygous for a missense mutation. 

Figure 4

 

Genotype–phenotype correlation in patients with CRB1 mutations. (A) The percentage of patients with LCA, RP, or EORD who have null mutations on both alleles, missense mutations on both alleles, or one null and one missense mutation on each allele. The level of significance (P) is depicted by one asterisk (borderline significance) or two asterisks (significant). (B) Distribution of homozygous missense mutations along the CRB1 protein domains. The number of patients with LCA, RP, or EORD is represented by column heights.

Overall, most (62%, 116 of 188) CRB1 patients were diagnosed with LCA, 13% with EORD, and 26% with RP. Among group 1 patients, 30 were diagnosed with LCA, 3 with EORD, and 2 with RP (Fig. 4A). The two patients with RP were diagnosed at a relatively advanced age (of 52 ²⁷ and 38 ¹ years old) and, therefore, no accurate diagnosis could be made. Among group 2 patients, 33 were diagnosed with LCA, 7 with EORD, and 13 with RP (Fig. 4A). Among group 3 patients, 53 were diagnosed with LCA, 14 with EORD, and 33 with RP (Fig. 4A). A χ² analysis revealed a statistically significant difference (P = 0.0019) in the distribution of phenotypes between groups 1 and 3 (Fig. 4A). On the other hand, a borderline level of significance was obtained for the difference between groups 1 and 2, and no significant difference was found between groups 2 and 3 (Fig. 4A). These results indicated that patients with null mutations on both alleles are likely to suffer from a more severe form of retinal degeneration, while a missense mutation on at least one allele may confer a milder (e.g., RP) phenotype. 

The CRB1 protein contains a relatively large number of domains with known functional properties (e.g., laminin AG–like and EGF-like domains). Aiming to identify a possible correlation between disease severity and mutation location, we collected information on patients with missense mutations in CRB1. First, CRB1 missense mutations were distributed evenly along the protein with no preference for either laminin AG–like or EGF domains. Second, when we included only patients who are homozygous for missense mutations (Fig. 4B), a relatively high fraction of patients (54%–60%) with mutations in laminin AG–like domains 2 and 3 were diagnosed with RP compared to 12.5% in the remaining domains. 

Discussion

Mutations in CRB1 were reported previously as the cause of various forms of EORD in a large number of families of different ethnic origins. ^1,3–5,14 In our study we showed that mutations in CRB1 are a common cause of EORD among the Jewish and Arab-Muslim populations in Israel and the Palestinian Territories. Approximately 10% of the inherited LCA cases in our cohort were due to CRB1 mutations, a proportion that is in line with data obtained from other populations (ranging from 10%–15%). ^{4,5,8,27,34–36,39,40}  

The existence of different CRB1–associated retinal phenotypes and different types of mutations (missense, splice-site, and null) prompted other groups to search for genotype–phenotype correlations, but the results were nonconclusive. ^1,5 The lack of a clear correlation might be due to other modifying factors affecting the retinal phenotype in patients with CRB1 mutations, as suggested previously. ^1,5 Alternatively, the correlation does exist, but could not be detected in the studied cohort of patients. This could be explained, for example, by patients with compound heterozygous mutations, each having a different effect on the function of the CRB1 protein, thus masking any possible genotype–phenotype correlation. The populations we are studying have unique features, mainly a high level of consanguinity and a relatively large number of siblings per family. These allow us to identify a large proportion of patients who are homozygous for the disease-causing mutation. A clinical comparison between patients who are homozygous for different mutation types allows a more accurate genotype–phenotype analysis. In our study, 25 of 30 genotyped patients (83%) were homozygous for CRB1 mutations, compared to 66 of 171 (39%) in the cohort of patients originating mainly from Europe. ¹ As hypothesized previously by others, ^1,4,5,41 a more severe retinal phenotype (i.e., LCA) might be associated with null mutations (i.e., nonsense, frameshift, and canonical splice-site) on both gene copies. It not always is possible, however, to assess the effect of missense mutations on protein function, as some might lead to a nonfunctional protein. Taking this into account, we have revised the genotype–phenotype analysis performed by others, and showed a statistically significant difference between groups 1 and 3 (null mutations on both alleles versus missense mutations on both alleles), thus supporting the previously suggested hypothesis that the presence of two null mutations may confer a more severe retinal phenotype. 

The high number of functional domains in the CRB1 protein prompted others ^1,5,36,41 and us to inquire whether there is any correlation between disease severity and localization of missense mutations. Since some missense mutations were reported to show a more severe retinal phenotype compared to others, ^1,5,36,41 we decided to discard from the analysis compound heterozygous patients and include only patients who were homozygous for a missense mutation in CRB1. Our analysis indicated that patients with homozygous missense mutations within the Ca++ binding EGF–like domains are more likely to be diagnosed with a more severe phenotype (LCA or early RP), compared to patients with homozygous missense mutations in the laminin AG–like domains. The CRB1 protein contains 19 Ca²⁺ binding EGF–like domains and only 3 laminin AG–like domains. Despite the numerous copies of the Ca²⁺ binding EGF–like domain, each copy seems to be important for the function of the protein, since missense mutations in most of these domains are more likely to cause a severe phenotype. 

In accordance with data from other studies, we were not able to demonstrate any correlation between mutation type and associated clinical features, such as PPRPE and Coats disease. ^1,5 As others suggested previously, ^1,5,36,40,41 this can be due to modifying effects of other genetic or environmental factors. 

In summary, we reported the CRB1 mutation spectrum in the Israeli and Palestinian populations, and showed that mutations in CRB1 are a frequent cause of arLCA and EORD in these populations. 

Supplementary Materials

pdf S1 

Acknowledgments

The authors thank the patients and their families for their participation in the study, Israel Barzel, Kinga Bujakowska, and Michelle Grunin for excellent assistance. 

Footnotes

 Supported by the Foundation Fighting Blindness USA (BR-GE-0510-0490-HUJ to DS) and the Yedidut 1 Research Grant (EB).

Footnotes

³  These authors contributed equally to the work presented here and therefore should be regarded as equivalent authors.

Footnotes

 Disclosure: A. Beryozkin, None; L. Zelinger, None; D. Bandah-Rozenfeld, None; A. Harel, None; T.A. Strom, None; S. Merin, None; I. Chowers, None; E. Banin, None; D. Sharon, None