A Homozygosity-Based Search for Mutations in Patients with Autosomal Recessive Retinitis Pigmentosa, Using Microsatellite Markers

Abstract

purpose. To identify possible mutations in known candidate genes in patients with autosomal recessive (ar) and simplex retinitis pigmentosa (RP), by using an established strategy of flexible, multiplexed, microsatellite-based homozygosity mapping.

methods. A total of 78 microsatellite markers corresponding to 16 genes known to be responsible for arRP were selected and used in 18 multiplex amplifications, followed by genotyping. Twelve consanguineous probands and 47 nonconsanguineous probands (59 patients with arRP or simplex RP) agreed to the screening.

results. Of the 59 probands examined, 24 had a mean of 1.4 genes showing homozygosity for all markers within the corresponding gene region. Subsequent direct sequencing revealed three homozygous mutations. Two of them were novel mutations in the genes TULP1 (c.1145T→C, F382S) and CNGB1 (c.3444+1G→A). The other was a mutation in RPE65 (c.1543C→T, R515W), which is known to cause Leber’s congenital amaurosis. The clinical features of each patient, together with the cosegregation analysis, strongly support the pathogenicity of these mutations.

conclusions. This systematic approach facilitated the identification of genes that cause arRP, and the results provide a widened spectrum of the mutation severity associated with a broader range of phenotypic manifestations of arRP.

Retinitis pigmentosa (RP) is a group of progressive hereditary disorders that cause retinal dystrophy which leads to blindness. Patients with RP show a wide variation in severity, mode of inheritance, age of onset, progression, and clinical manifestation. Approximately 1.5 million people are estimated to be affected by this disease worldwide. ¹ Patients experience night blindness, a gradual loss of peripheral visual field, and eventual loss of central vision, in most cases because of degeneration of the photoreceptor cells of the retina. It is a genetic disease with extremely high heterogeneity. ² There are autosomal dominant (adRP), autosomal recessive (arRP), X-linked (xlRP), and rare mitochondrial and digenic forms. Between them, adRP and arRP each account for approximately 20% of cases of RP, and xlRP accounts for 10%. The remaining 50% of forms of RP are described as simplex RP, although many of them may be unidentified types of arRP. ^{3 4 5} Although arRP is the most common form, its genetic nature is obscure, and a precise diagnosis is elusive because of numerous technical obstacles. 

Searching for a gene mutation known to cause RP is extremely time consuming and costly, and genetic testing as a clinical service is far from practical. ⁶ This is because 16 responsible genes have already been identified (according to RetNet: http://www.sph.uth.tmc.edu/RetNet/, 2003 version; provided in the public domain by the University of Texas Houston Health Science Center, Houston, TX), with >250 exons (>38,000 nucleotides of coding sequence), and scrutinizing all these targets is not easy. Furthermore, many genes remain to be identified, as the known loci do not account for many cases of RP. ⁷ Focusing on the target genes according to their clinical features is one possibility, but there is currently no database that provides a link between specific clinical phenotypes and mutations. No gene mutation is known to be the major cause of arRP, except in one population, ⁸ and family samples suitable for linkage analysis are largely unavailable. Thus, identifying the genes and their mutations that cause arRP remains methodologically challenging. 

Because of the genetic nature of recessive inheritance, the regions adjacent to the disease-causing locus are likely to be homozygous by descent in patients from consanguineous families. ^{9 10} Thus, homozygosity mapping is an efficient method for locating the responsible genes in such families. This approach is generally believed to be unsuitable when hunting for mutations responsible for a monogenic disease in nonconsanguineous families. However, homozygosity rather than compound heterozygosity, is frequently observed in the responsible gene in a patient with no apparent consanguinity in instances when the disease is extremely rare. ¹¹ This may be explained by the low mutation rate of the concerned gene and a high probability of homozygosity by descent of the gene in the patient, even though the common founder cannot be traced back in the available family history. As for arRP mutations in particular, responsible genes are likely to be very rare, although arRP as a whole is rather common because of its extensive heterogeneity. That at least some arRP mutations have been found in the homozygous state supports this idea. ^{12 13 14 15} Thus, homozygosity mapping may be a suitable method to focus on candidate genes for this disease, although some of the mutations may escape the search because of compound heterozygosity. 

We have previously reported a systematic method for the efficient diagnosis of RP with dominant inheritance (adRP) by simultaneously amplifying multiple microsatellite markers specific for the adRP gene, followed by linkage analysis. ¹⁶ In the current study, we applied the same technology to detect homozygosity of the known candidate genes for arRP for patients with or without consanguinity, and successfully identified three mutations. 

Methods

Designing Multiplex Gene-Specific Markers

All 16 arRP causative genes listed in RetNet (2003 version) were screened: photoreceptor cell-specific ATP-binding cassette transporter (ABCA4), ^{17 18} α-subunit of rod cGMP-gated channel (CNGA1), ¹⁹ β-subunit of rod cGMP-gated channel (CNGB1), ²⁰ Crumbs homologue 1 (CRB1), ²¹ lecithin retinol acyltransferase (LRAT), ²² c-mer proto-oncogene tyrosine kinase (MERTK), ¹⁵ photoreceptor cell-specific nuclear receptor (NR2E3), ²³ α-subunit of rod cGMP phosphodiesterase (PDE6A), ²⁴ β-subunit of rod cGMP phosphodiesterase (PDE6B), ²⁵ retinal G-protein-coupled receptor (RGR), ¹³ rhodopsin (RHO), ²⁶ cellular retinaldehyde-binding protein (RLBP1), ²⁷ RPE65 (RPE65), ²⁸ arrestin (SAG), ¹⁴ tubby-like protein 1 (TULP1), ^{12 29} and usherin (USH2A). ³⁰ The human genome draft sequences were retrieved from GenBank (http://www.ncbi.nlm.nih.gov/Genbank; provided in the public domain by the National Center for Biotechnology Information [NCBI], Bethesda, MD). The accession numbers of sequence contigs and transcripts of the respective genes are shown in Table 1 . Another gene, ceramide kinase-like protein (CERKL) was listed on RetNet (last updated January 27, 2004), but had not yet been cloned when this study began. The procedure for multiplex microsatellite genotyping was essentially the same as previously reported, ¹⁶ with minor modifications: di-, tri-, and tetranucleotide repeats were selected from a maximum 1000-kb segment that included the first nucleotide of the first exon of each gene at its center. Microsatellites of >10 repeats for dinucleotides or more than seven repeats for tri- and tetranucleotides were chosen. One of the 5′ ends of each primer pair was modified to contain either a GTT (for genes ABCA4, CNGA1, LRAT, MERTK, PDE6A, PDE6B, RGR and RLBP1) or ATT (for genes CNGB1, CRB1, NR2E3, RHO, RPE65, SAG, TULP1, and USH2A) sequence for two-color fluorescent labeling after polymerase chain reaction (PCR) amplification. ³¹ The other ends were modified to a TCC sequence, to protect them from being labeled. A total of 78 microsatellite markers were selected, and 18 multiplex PCRs were developed for all 16 genes. Detailed experimental information, including primer sequences, the expected length of the marker and the multiplex condition, and the number and frequency of the different alleles of each marker in the reference population, is listed in Supplementary Table S1. 

Homozygosity Mapping

Fifty-nine probands with RP (11 with arRP and 48 with simplex RP) were screened. Of these probands, 12 had a history of consanguinity (all patients were offspring of first-cousin marriages) and 47 were without such a history (Table 2) . The study adhered to the tenets of the Declaration of Helsinki. The research protocol was approved by the Ethic Review Board of Fukuoka University, and signed informed consent was obtained from each patient in Fukuoka University on the basis of the protocol. 

Genomic DNA was extracted from peripheral blood (QIAamp DNA Blood Midi Kit; Qiagen, Chatsworth, CA). Genotyping was performed as described. ¹⁶ If at least one marker was detected to be heterozygous within a given gene, the gene was regarded as heterozygous. For each patient, only genes homozygous for all markers were used for subsequent direct sequencing. 

Mutation Identification

Oligonucleotide primers were designed with Primer3 software (http://frodo.wi.mit.edu/primer3/primer3_code.html/; provided in the public domean by the Whitehead Institute, Massachusetts Institute of Technology, Cambridge, MA) ³² in such a way that each exon was bracketed by the appropriate intronic primers. The primer sequences are available on request. Sequencing reactions and then electrophoresis were performed according to standard protocols. ¹⁶ The sequence data were aligned with the respective reference sequences using phred/phrap/polyphred software. ³³ When nucleotide changes were detected, they were searched against the dbSNP database (http://www.ncbi.nlm.nih.gov/SNP; provided in the public domain by NCBI). 

Results

Multiplex Gene-Specific Markers

Eighteen multiplex PCRs were developed, which amplified 78 microsatellite markers for all 16 genes (an average of 4.9 markers per gene). The locations of markers in each gene are shown in Table 1 . The distance of each marker from RP gene regions ranged from 0 to 499 kb (average, 176 kb), where gene regions are defined to be from the first nucleotide of translation starting site to the last nucleotide of the stop codon. Detailed information on each marker and a summary for each gene are described in Table 1 and Supplementary Table S1. Homozygosity and the numbers of alleles of each marker were examined in eight normal subjects. As shown in Table 1 , homozygosity of the markers in each gene varied from 0% to 88%. Observed homozygosity was 0 for all genes, if all markers in each gene were combined. 

Homozygosity Mapping and Mutation Identification

Of the 59 patients tested, 35 were heterozygous for all the genes tested and were excluded from further screening. As expected, patients with consanguinity tended to have more homozygous regions than those without; homozygosity of the examined genes reached an average of 7% with consanguinity and 3% without (Table 3) . Samples from patients with homozygous genes were subjected to direct sequencing for the respective candidate genes. The number of patients searched for mutations was one for CNGA1 and RLBP1; two for CRB1, PDE6A, RHO, TULP1, and USH2A; three for ABCA4, NR2E3, PDE6B, and RPE65; four for CNGB1; and five for SAG. No patients were homozygous for the genes LRAT, MERTK, and RGR. Three genes revealed heterozygous single nucleotide polymorphisms (SNPs) in three patients: ABCA4 (A/G at c.635) in patient 00201, CNGB1 (A/G at c.299 and T/C at c.1783+23) in patient 02701, and USH2A (A/G at c.3157+34 and G/T at c.3812−8) in patient 02801. These genes were regarded as heterozygous and were excluded from further analysis. 

Subsequent sequencing revealed that 28 nucleotides in 10 genes from the patients with RP differed from the reference sequences. These were all homozygous substitutions, as summarized in Supplementary Table S2. Twenty-five were polymorphisms. There were 14 substitutions within introns at sites unlikely to be involved in splicing; 11 exonic substitutions (9 within coding regions and 2 within the untranslated region) were listed in dbSNP, identical with those of the respective NCBI reference sequence for mRNA or those found among normal subjects (our data). The remaining three sequence changes were considered to be mutations. These were c.1543C→T in RPE65 (patient 03901), c.1145T→C in TULP1 (patient 04601), and c.3444+1G→A in CNGB1 (patient 02201) (Fig. 1) . These sequence changes in the three genes were not present in the 159 control individuals (318 chromosomes). 

Family samples were available for two of the patients, and cosegregation of these mutations with the disease was searched for after sequencing. The changes in RPE65 and TULP1 cosegregated, respectively, with the disease among the families of patients 03901 and 04601 (Fig. 2) . A homozygous c.1145T→C mutation in TULP1 was also found in the affected brother of patient 04601. The other family members analyzed carried each mutant allele in a heterozygous manner. In patient 02201 with c.3444+1G→A in CNGB1, no family history of RP was known, and DNA for a family study was not available. 

An RPE65 Mutation Known to Cause Leber’s Congenital Amaurosis

Patient 03901 (c.1543C→T in RPE65), a 55-year-old Japanese woman, is the progeny of consanguineous parents (first cousins) who observed development of night blindness in her early childhood. She was free from visual disability until marriage at 24 years of age. Aged 40, she received a diagnosis of RP with loss of vision and constricted visual fields in both eyes. By the age of 54, she only had basic light–dark perception in both eyes. An examination of the fundus revealed pigmented lesions in the form of clumps or bony spicules involving the posterior retina and associated with a wide area of chorioretinal atrophy, which was prominent in the peripapillary area in both eyes (Fig. 3a) . These changes resembled those of some patients with RPE65 mutations who have been reported, ^{35 36} although the fundus appearance varies widely among patients with RPE65 mutations. An electroretinogram (ERG) showed no recordable rod or cone responses in both eyes. 

The RPE65 gene encodes for an abundant and evolutionally conserved 61-kDa protein in retinal pigment epithelium. This protein product plays a role in vitamin A metabolism in the retina, and the whole region of the protein is highly conserved in vertebrates. ^{37 38} The change c.1543C→T in RPE65 results in a missense mutation, R515W. Codon 515 is located in the conserved RPE65-specific region, and this amino acid is strictly conserved between species including human, dog, cow, mouse, rat, chicken, salamander, fog, and fruit fry (the respective GenBank or SwissProt accession numbers are NP_000320.1, Y16567, CAA46988.1, NP_084263.1, NP_446014.1, AB017594, AF047465, BC043751, NM_142050.2) (SwissProt is provided in the public domain by the Swiss Institute of Bioinformatics, Geneva, Switzerland, and is available at http://www.expasy.org). 

So far, more than 50 disease-causing mutations in the RPE65 gene have been reported, and these are known to cause diseases with a wide range of severity, from congenital blindness (Leber’s congenital amaurosis; LCA) to adult-onset RP. However, the most common phenotype is severe and early-onset retinal degeneration. ^{28 39 40 41 42} In most patients with RPE65 mutations, disease was diagnosed in infancy, with visual impairment frequently associated with nystagmus, night blindness, and a tendency to fixate on light. ⁴² In contrast, the visual performance of several patients in bright light was sufficient to permit attendance at regular school during the elementary years. At older ages, often during the secondary school years, visual acuity was greatly reduced. ⁴² Wada et al. (IOVS 2000;41:ARVO Abstract 617) reported two siblings with LCA who had compound heterozygous L450R and R515W mutations. Although details of clinical phenotypes were unavailable, our patient 03901, bearing a homozygous R515W mutation, comparatively exhibited a milder phenotype in childhood and a slower progression of the disease. Severe vision loss is believed to be the result of loss or considerable reduction of protein function due to mutations affecting both RPE65 alleles. Some homozygous missense mutations, such as M1T, A132T, R91W and P363T, are believed to result in loss or considerable reduction of protein function, causing severe and early-onset retinal degeneration. ^{28 42} The missense mutation found in this study affects highly conserved residues and may result in functionally impaired but milder reduction of protein function than other missense mutations, as mentioned earlier. 

A Novel TULP1 Mutation

Patient 04601 (c.1145T→C in TULP1) was a 20-year-old woman of Japanese origin. At 8 years of age, she began to exhibit myopic refractive errors and her corrected visual acuity was 0.3 in both eyes, without nystagmus. A diagnosis of RP was made because of poor dark adaptation and visual field defect. Funduscopy of both eyes showed attenuated retinal vessels and minimal pigmented retinopathy with temporal disc pallor and annuli of yellow deposits on the macula (Fig. 3b) . Her best corrected visual acuity was 0.3 with a refraction of −10 D, and visual fields were restricted to 20° in both eyes. The latest funduscopy showed a milder pigment change bilaterally. Dark-adapted flash ERGs showed nonrecordable patterns in both eyes. The affected brother also had a diagnosis of RP together with color vision alteration at 8 years of age. Fundoscopy of both eyes showed minimal pigmentary retinopathy with a bull’s-eye appearance of the macula. Dark-adapted flash ERGs showed nonrecordable patterns in both eyes. He had myopia with best corrected visual acuity of 0.5, and visual fields were restricted to 20°. Neither of the two patients was obese or showed other endocrine disorders or hearing impairment. 

TULP1 is a member of the tubby gene family, defined by its C-terminal half, which is highly conserved between the mouse Tub gene and the human orthologue TUB. The tubby-like genes TULP1 and TULP2 also carry this conserved region. ⁴³ The mutation c.1145T→C in TULP1 results in the missense change F382S, which is located in the tubby C-terminal domain, and the amino acid is highly conserved between human, mouse, rat, and chicken TULP1 as well as human TULP2 (the respective GenBank or SwissProt accession numbers are NP_003313.2, NP_067453.1, XM_228360.2, U92545.2, and NP_003314.1). 

The findings in this study are very similar to those in patients with TULP1 mutations (i.e., early-onset of symptom, myopia, central vision impairment, and milder pigmentary retinopathy together with signs of maculopathy). ^{44 45} So far, only nine TULP1 mutations (IVS2+1G→A, IVS4−2delAGA, c.937delC, IVS12+4A→G, R420P, I459L, K489R, F491L, and IVS14+1G→A) have been reported to cause arRP, with early-onset and severe forms. ^{12 29 45 46} Although the clinical features in the siblings with the disease match well with previously reported features, the patients lacked severe central visual dysfunction and nystagmus, which have been regarded as characteristic findings at all stages of the disease, ⁴⁴ suggesting that the mutation found in this study resulted in a milder reduction of protein function than the known mutations mentioned herein. 

A Novel CNGB1 Mutation

Patient 02201 (c.3444+1G→A in CNGB1), a 67-year-old Japanese man, noticed night blindness at school age and was diagnosed with RP at the age of 30 because of his visual field defects. The visual field defects slowly progressed bilaterally, whereas his visual acuity remained in the vicinity of 1.0. Twenty-five years later, after bilateral cataract surgery, dark-adapted flash ERGs were nonrecordable in both eyes. Presently, his corrected visual acuity is 0.07 in the right eye and 0.6 in the left. The visual field is restricted to 5° and 10° in the right and left eyes, respectively. A fundus examination revealed a typical bony spicule pigment lesion and nonremarkable macular changes in the left eye, whereas the degeneration involved the macula in the right eye (Fig. 3c) . 

The CNGB1 gene encodes the β-subunit of the rod photoreceptor cyclic nucleotide-gated (CNG) channel, which comprises a heterotetramer with the α-subunit of the channel encoded by CNGA1. ⁴⁷ The CNG channel conducts a cation current in response to changes in intracellular levels of cGMP and mediates the electrical response to light. ⁴⁸ Both genes share functional features, including the cGMP-binding domain to which CNG channels are activated by the direct binding of cGMP. We found a novel splice site mutation at a donor site of the second-to-last exon (exon 32 according to the exon numbering by Ardell et al. ³⁴ and Bareil et al. ²⁰ ). The mutation results in a frameshift, which leads to truncation of the protein with loss of the last 28 amino acids. 

Bareil et al. ²⁰ have reported one consanguineous family with a CNGB1 mutation, G993V, which affects the cyclic nucleotide-binding domain. The affected individuals presented typical and severe features of RP. The proband had had night blindness since early childhood, and visual impairment had progressed by age 30. The visual field defects were reduced to 10°, and the ERGs exhibited no rod response and severely reduced cone response. The fundus showed typical bone-spicule–shaped pigmentary deposits. Compared with these patients, patient 02201 showed a later onset. A possible explanation of this milder expression of RP is that the truncated C-terminal region is not directly involved in known functional domains. 

Discussion

We have identified three homozygous mutations in genes known to cause arRP. Two of them were novel mutations in very rare causes of arRP with a few cases documented worldwide so far. The other was a mutation known to cause LCA but first reported as a cause of arRP. Cosegregation of these mutations with the disease indicated that two of the mutations were responsible for the disease. The phenotypes associated with all mutations matched well the previous reports of the respective genes, although they were relatively milder. 

Screening only the proband by homozygosity mapping worked effectively in narrowing down the number of genes. Thus, the method has a remarkable advantage for the diagnosis of simplex cases of this disease. In this study, homozygosity was not detected for all genes in 60% of the probands and the candidates were narrowed to 1 to 5 (mean of 1.4) genes of 16 in total. Thus, the subjects to be analyzed were reduced to 3.5% of the population (1.4/16 × 40%) compared with the number required if all genes were to be searched at the time. 

Searching for homozygosity by descent has been applied to diagnose RP in consanguineous families. ^{20 23 49} As pointed out in the original paper of Lander and Botstein, ⁹ homozygosity mapping works best in genetically isolated populations and/or with disease of very low incidence and a small number of different alleles. The data obtained in this study robustly confirm this assumption. Two of the three genes in which mutations were identified among nonconsanguineous families represent very rare causes of arRP. Thus, homozygosity mapping using these gene-specific markers was effective in detecting mutations in rare causes of arRP genes, especially among nonconsanguineous families. 

Because this method was initially intended to diagnose consanguineous cases, the genomic interval of the markers was relatively large in some genes, such as CNGA1, CNGB1, LRAT, PDE6B, SAG, and USH2A, in which the farthest markers are >350 kb from coding sequence of the respective genes, as shown in Table 1 . This larger genomic interval may lead to erroneous exclusion of cases because of recombination in cases of nonconsanguinity. In this sense, the exclusion criteria used in this study (at least one marker had to be detected to be judged heterozygous) may have been too strict for the genes listed. Thus, using only the markers located in close vicinity or within the gene (e.g., including SNPs) may improve the detection of causative genes. In the two genes in which mutations were identified among nonconsanguineous patients, the farthest markers were more distantly located from the coding sequence of the genes (454 and 314 kb apart from the CNGB1 and TULP1 genes, respectively). Thus, exclusion criteria depend on how distantly common founders are expected on the basis of the genetic background. Among recent genetic founder populations (<200 generations) with low immigration and expansion rates, such as Finland, Iceland, Sardinia, and Japan, marker–marker linkage disequilibrium is regarded to be <1 cM. ⁵⁰ Furthermore, the chance that a given locus will be found homozygous by descent is less among nonconsanguineous cases than among consanguineous cases (the chances are 6.0% and 1.5%, for siblings of first- and second-cousin marriages, respectively ⁹ ). Searching among nonconsanguineous cases is productive, as there is a reduced likelihood of random homozygosity (false positives). In the present study, we found that 7% and 3% of the genes examined were apparently homozygous among the offspring of consanguineous and nonconsanguineous couples, respectively. The latter value is far above our expectation on the basis of the coefficient of inbreeding (F: random homozygosity) of this population. ⁵¹ Thus, many of the homozygous sites we detected may turn out to be heterozygous when further examined with additional highly informative markers. For instance, three subjects were subsequently revealed to carry heterozygous SNPs in ABCA4, CNGB1, and USH2A. Eight normal subjects were tested for the homozygosity of each marker as shown in Table 1 . Certainly, a more accurate representation of the number of different alleles in people and of the homozygosity fraction would yield a greater number of individuals. 

In contrast to the mutations in CNGB1 and TULP1, the situation is quite different for RPE65, in which mutations are more frequent causes of arRP and early-onset retinal degeneration. This gene shows extensive allelic heterogeneity or mutation hot spots with recurrent mutations. For such genes, the proportion of patients homozygous by descent should be low. For RPE65, the likelihood of homozygous mutations in a genetically open populations is 35% (7/20) according to the survey by Thompson et al., ⁴¹ who analyzed 453 unrelated patients with retinal dystrophy from the United States and Europe. In such genes, mutations in individuals homozygous by descent may be found mainly among consanguineous families, as shown in this study. 

In the instances of nonconsanguineous offspring, there is no clue to distinguish simplex patients with homozygous recessive mutations from those with other types of inheritance or nonmonogenic conditions. Forty-seven of the patients we examined here were from nonconsanguineous families, and only two mutations were detected. This figure is far below the level of practical application in the genetic diagnosis of patients with RP, and the focus may be primarily restricted to research applications. Only one mutation was detected among the 12 probands from consanguineous families. This was far below our expectation, because the responsible genes in these patients should be homozygous. We believe this is because of the presence of many unidentified responsible genes. During the preparation of this article, a new candidate gene, CERKL was identified. ⁵² This will be added to our subsequent testing. Some other genes responsible for related diseases, such as LCA, cone–rod dystrophy, and congenital stationary night blindness, or genes for RP with accompanying systemic abnormalities will also be included in future studies. It is also worthwhile to screen for genes exclusively expressed in the retina. 

We believe this approach provides a powerful tool for detecting candidate genes that are homozygous by descent and that excluding all such genes could be a great help in narrowing down novel genes responsible for arRP. This study thus contributes to the understanding of possible links between the variable phenotypes and their respective genes leading to arRP. 

 Supported by Grants-in-Aid 15591883 for Scientific Research, Japan; and the Japanese Retinitis Pigmentosa Society, 2002.

 Submitted for publication May 17, 2004; revised June 30, 2004; accepted July 1, 2004.

 Disclosure: H. Kondo, None; M. Qin, None; A. Mizota, None; M. Kondo, None; H. Hayashi, None; K. Hayashi, None; K. Oshima, None; T. Tahira, None; K. Hayashi, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Summary of Gene-Specific Microsatellite Markers

Gene	Contig ID*	Transcript ID*	Markers (n)	Marker Location (kb), †					Homozyosity (%), ‡
								Near	Far	Average	Range
ABCA4	NT_019284.5	NM_000350	4	0	256	120	13–63
CNGA1	NT_006238.10	NM_000087	4	0	395	187	13–38
CNGB1	NT_010406.7	NM_001297	6	6	454	186	13–63
CRB1	NT_004680.5	NM_120076	4	0	114	38	25–88
LRAT	NT_006258.11	NM_004744	7	20	481	246	38–75
MERTK	NT_022135.10	NM_006343	5	0	264	179	13–63
NR2E3	NT_010235.7	NM_016346	3	46	87	61	0–75
PDE6A	NT_027022.1	NM_000440	5	26	139	89	13–25
PDE6B	NT_006111.12	NM_000283	6	78	435	269	25–88
RGR	NT_024064.8	NM_002921	6	0	346	130	13–75
RHO	NT_006025.4	NM_000539	3	0	61	24	25–63
RLBP1	NT_010356.5	NM_000326	5	48	189	114	0–63
RPE65	NT_004464.7	NM_000329	5	2	276	87	25–63
SAG	NT_005414.5	NM_000541	5	91	499	261	13–63
TULP1	NT_007592.5	NM_003322	6	12	314	210	0–63
USH2A	NT_004612.12	NM_007123	4	134	388	228	0–25

*  Sequence data were retrieved from Genbank.

†  Distance from the marker to the closest end of the concerned gene, which is defined to be from the first nucleotide of the translation starting site to the last nucleotide of the stop codon. The distances for the markers within the gene (intron) are assumed to be 0.

‡  Data were from eight normal subjects (Japanese).

Table 2.

 

Summary of Probands: Mode of Inheritance and Consanguinity

Mode of Inheritance	Consanguinity (+)	Consanguinity (−)	Total
arRP	2	9	11
Simplex	10	38	48
Totals	12	47	59

 Data are number of patients.

Table 3.

 

Number of Probands Carrying a Complete Set of Homozygous Gene-Specific Markers

Gene	Consanguinity (+) (n = 12)			Consanguinity (−) (n = 47)
ABCA4	0		3*	(6%)
CNGA1	0		1	(2%)
CNGB1	2*	(17%)	2, †	(4%)
CRB1	0		2	(4%)
LRAT	0		0
MERTK	0		0
NR2E3	2	(17%)	1	(2%)
PDE6A	2	(17%)	0
PDE6B	2	(17%)	1	(2%)
RGR	0		0
RHO	1	(8%)	1	(2%)
RLBP1	0		1	(2%)
RPE65	1, ‡	(8%)	2	(4%)
SAG	4	(33%)	1	(2%)
TULP1	0		2, §	(4%)
USH2A	0		2*	(4%)
Average	0.9	(7%)	1.2	(3%)

*  Heterozygous single nucleotide polymorphisms were found: A/G at c.635 in ABCA4 (patient 00201); A/G at c.299 and T/C at c.1783+23 in CNGB1 (patient 02701), and A/G at c.3157+34 and G/T at c.3812−8 in USH2A (patient 02801). These were excluded from further analysis.

†  The mutation c.3444+1G→A was detected in patient 02201.

‡  The mutation c.1543C→T was detected in patient 03901.

§  The mutation c.1145T→C was detected in patient 04601.

Figure 1.

 

Sequencing analysis of PCR products from patients (m) and control subjects (wt), defining three novel homozygous mutations identified in individuals with RP. Exon numbering in CNGB1 was based on previous reports by Ardell et al. ³⁴ and Bareil et al., ²⁰ but not on the annotation of the respective genomic contig. Note that polymorphism (c.3444+7T→C) was also found (asterisk). Mutation in exon 14 of the RPE65 gene was analyzed using a 334-bp PCR fragment generated with the primers 5′-ATT CAT GCC AGG TGG TAC AAG A-3′ and 5′-GTT CCA AAA ACA TAT CTT GCT GGA GT-3′. Mutation in exon 11 of TULP1 gene was analyzed with a 365-bp PCR fragment generated using primers 5′-ATT GGG GAC ATG AAT TGC TCA GT-3′ and 5′-GTT CCT GGG AAG GAA ACA GAT G-3′. Mutation in exon 32 of the CNGB1 gene was analyzed using a 326-bp PCR fragment generated with primers 5′-ATT CCA GAG TGC GTC TGT GAT GT-3′ and 5′-GTT GGC CTT CTC CCT GAC ACA TA-3′. For these sequence pairs, the tagged sequences (5′-ATT and 5′-GTT) were added for postlabeling. ¹⁶

Figure 2.

 

Pedigree structure and segregation analysis of mutations in patients with RP. (a) Patient 03901, c.1543C→T in RPE65 and (b) patient 04601, c.1145T→C in TULP1. The genotypes of individuals with DNAs available are shown beneath their symbols: +, wild-type sequence; M, mutant sequences. Arrows: probands.

Figure 3.

 

Photographs illustrating different retinal appearances associated with mutations in three different genes. (a) Patient 03901 (c.1543C→T in RPE65) with pigment lesions in the form of clumps or bony spicules involving the posterior retina and associated with a wide area of chorioretinal atrophy, which was prominent in the peripapillary area. (b) Patient 04601 (c.1145T→C in TULP1) with minimal pigment retinopathy with temporal disc pallor. (c) Patient 02201 (c.3444+1G→A in CNGB1) with a typical bony spicule pigment lesion and nonremarkable macular changes in the left eye.

Supplementary Materials

Supplementary Table 1 - 335 KB 

Supplementary Table 2 - 93.2 KB 

The authors thank all the patients and families for their cooperation.