|Home | About | Journals | Submit | Contact Us | Français|
Retinitis pigmentosa is a genetically heterogeneous form of retinal degeneration that affects ~1 in 3500 people worldwide. Recently we identified the gene responsible for the RP1 form of autosomal dominant retinitis pigmentosa (adRP) at 8q11–12 and found two different nonsense mutations in three families previously mapped to 8q. The RP1 gene is an unusually large protein, 2156 amino acids in length, but is comprised of four exons only. To determine the frequency and range of mutations in RP1 we screened probands from 56 large adRP families for mutations in the entire gene. After preliminary results indicated that mutations seem to cluster in a 442 nucleotide segment of exon 4, an additional 194 probands with adRP and 409 probands with other degenerative retinal diseases were tested for mutations in this region alone. We identified eight different disease-causing mutations in 17 of the 250 adRP probands tested. All of these mutations are either nonsense or frameshift mutations and lead to a severely truncated protein. Two of the eight different mutations, Arg677X and a 5 bp deletion of nucleotides 2280–2284, were reported previously, while the remaining six mutations are novel. We also identified two rare missense changes in two other families, one new polymorphic amino acid substitution, one silent substitution and a rare variant in the 5′-untranslated region that is not associated with disease. Based on this study, mutations in RP1 appear to cause at least 7% (17/250) of adRP. The 5 bp deletion of nucleotides 2280–2284 and the Arg677X nonsense mutation account for 59% (10/17) of these mutations. Further studies will determine whether missense changes in the RP1 gene are associated with disease, whether mutations in other regions of RP1 can cause forms of retinal disease other than adRP and whether the background variation in either the mutated or wild-type RP1 allele plays a role in the disease phenotype.
Retinitis pigmentosa (RP) is a heterogeneous group of inherited retinal diseases which affect more than one million people worldwide. Individuals affected with RP exhibit night blindness followed by progressive degeneration of the retina, often culminating in legal or complete blindness in the later decades of life. Bone spicule-like pigmentary deposits and an abnormal or absent electroretinogram (ERG) accompany photoreceptor degeneration (1,2). RP can be inherited in autosomal dominant, autosomal recessive, X-linked or digenic modes; to date, 11 autosomal dominant, 11 autosomal recessive, five X-linked and one digenic forms of retinitis pigmentosa have been identified through linkage studies or candidate gene screening. Several genes have been identified for each of these types, but the majority remain unknown (RetNet). Recently we identified the gene and mutations that are responsible for the RP1 form of autosomal dominant retinitis pigmentosa (adRP) located on chromosome 8q (3).
The RP1 locus was first identified using a large nine-generation family and was subsequently refined to a 4 cM interval between D8S601 and D8S285 using a second family (4–7). The RP1 gene has four exons with the initiation codon being present in exon 2. The RP1 mRNA is ~7000 bp long and is predicted to encode a protein 2156 amino acids in length with a calculated molecular weight of 240 kDa. The RP1 gene product shows partial sequence similarity to the human doublecortin gene (DCX) (8) and its expression is believed to be regulated by retinal oxygen levels (9); however, exact functional information is unknown at this time.
Three independent studies have reported a total of four different mutations in RP1 that cause adRP. One mutation, Arg677X, was found in all three studies, and appears to be quite common (3,9,10). An additional nonsense mutation, Gln679X (3), and two small deletions (Leu765 5 bp deletion; Asn763 4 bp deletion) (9) were also identified. All of these mutations result in a severely truncated protein approximately one-third the size of the wild-type (3,9,10).
The purpose of this study was to determine the types and frequency of mutations in RP1 that are responsible for autosomal dominant RP. Since previous studies have demonstrated that mutations within a single gene can cause different clinical phenotypes (11–16), we also wanted to establish whether RP1 mutations are the cause of other forms of retinal disease. We screened 659 unrelated probands with a wide range of clinical phenotypes for mutations. We identified eight different RP1 mutations in 17 of the 659 families tested. Six of the eight are novel mutations that have not been reported previously. All of these mutations lead to radically shortened proteins and were found in families with autosomal dominant RP.
In order to determine the types and frequencies of RP1 mutations responsible for retinal disease we tested a total of 659 affected probands (Table 1). The first set of 56 unrelated probands were screened for mutations in the entire RP1 gene. All 56 probands were members of American families with autosomal dominant RP who tested negative for mutations in rhodopsin, peripherin/RDS and CRX. This initial screen identified two different RP1 mutations in three families. One of these is the nonsense mutation in codon 677 (Arg677X) that was initially identified in the UCLA-RP01 and Australian families and was found in one of the 56 probands tested. The other disease-associated mutation, a 5 bp deletion of bases 2280–2284, was found in two families and results in 16 incorrect amino acids after codon 761 with a termination codon present at 777. This mutation, also referred to as Leu762 (5 bp deletion), has already been observed in one additional adRP family (9).
After preliminary results indicated that mutations seem to cluster in a small region of exon 4, we tested an additional 603 probands from American and European families for mutations in this small region. These probands have a range of retinal degenerations including autosomal dominant RP, autosomal recessive RP, isolated RP and autosomal dominant cone or cone–rod dystrophy. The segment tested spans 442 bp in exon 4 from nucleotide 1947 to 2388 (GenBank accession no. AF143222). Mutations in this small region of RP1 were found in 14 additional families with autosomal dominant RP.
We identified a total of eight different mutations in 17 of the 659 probands tested. Since all the mutations were found in probands with adRP, all frequency calculations are based on testing of the 250 probands with adRP. The remaining 409 probands are not included in these calculations since they have other forms of retinal degeneration. All disease-associated mutations and sequence variants identified in this study are listed in Table 2.
Mutation analysis revealed that five unrelated families had a nonsense mutation in codon 677 (Arg677X; 2029C→T) (Fig. 1A). This mutation is predicted to result in a severely truncated protein and is the same as the one originally identified in UCLA-RP01 and Australian family D. The Arg677X mutation was seen in both American and European samples and is responsible for ~2% (5/250) of adRP.
Families with the Arg677X mutation exhibit a range of disease phenotypes. The two original families, UCLA-RP01 and Australian family D, have mild forms of disease. In general, affected members of these families have late onset of night blindness and slow loss of visual acuity and visual fields. All of the families with the Arg677X mutation identified in this study have more severe forms of disease. These families have earlier onset of night blindness and more severe loss of visual acuity and fields. Despite a more severe disease phenotype, two families show variable penetrance and expressivity. One family, Moorfields1512, like UCLA-RP01 and Australian family D, has an instance of non-penetrance, while family UTAD103 has an elderly male with mild symptoms despite the severe disease in his daughter and grandchild.
Mutation testing also identified a 5 bp deletion of bases 2280–2284 (2280–2284del) in five unrelated families with adRP (Fig. 1B). This deletion results in 16 incorrect amino acids after codon 761 followed by a stop in codon 777. The 5 bp deletion was present in all affected family members that were tested and was not present in any of the unaffected members tested. This mutation is found in both American and European samples and appears to be responsible for ~2% (5/250) of adRP.
The majority of the families with the 2280–2284del mutation have mild disease with onset at 30–40 years of age, although a few family members have earlier onset and more severe disease. Of interest is the Moorfields1452 family. Even though the sister of the proband is affected also, this family was originally categorized as having an autosomal recessive pattern of inheritance since there was no other family history of disease. It is likely that one of the proband’s parents carries the 2280–2284del mutation, but due to non-penetrance or variable expressivity, has no discernable clinical symptoms.
Analysis of one affected member from two unrelated European families revealed a third type of RP1 mutation, a 14 bp deletion of bases 2168–2181 (2168–2181del) (Fig. 1C). This deletion results in 10 incorrect amino acids after codon 722 followed by a premature stop in codon 733. One family with this deletion, Moorfields1614, has another example of apparent non-penetrance.
The remaining RP1 mutations that were identified in this study were each found in only one proband. A 1 bp deletion of nucleotide 2303, was identified in one affected member of RFS103 (Fig. 1D). This deletion results in five incorrect amino acids after codon 768 with a premature termination codon at 774. This individual has mild disease with equal loss of rod and cone function. A second 1 bp deletion, 2029del, was found in seven affected members of Leeds356 (Fig. 1E). This deletion leads to four incorrect amino acids after codon 676 followed by a stop at codon 681. This deletion is of the same nucleotide that is mutated in the Arg677X mutation.
One insertion mutation in RP1 was identified in this study. One proband was tested and found to have a 1 bp insertion (G) between nucleotides 2169 and 2170 (Fig. 1F). This insertion leads to three incorrect amino acids after codon 724 followed by a premature stop codon. This individual has moderate disease with onset of visual field problems in childhood.
Two unique nonsense mutations were also identified. One affected member of Moorfields359 was tested and a nonsense mutation, Glu700X (2098G→T), was identified (Fig. 1G). This individual has mild retinal disease that was diagnosed at 41 years of age. The second nonsense mutation, Cys744X (2232T→A), was identified in one proband (Fig. 1H). This individual first noticed night blindness at 20 years of age.
Two missense variants were detected in affected probands. Leu1808Pro (5423T→C) was found in one affected member of RFS015 and a second variant, Lys663Asn (1989G→T), was detected in one affected member of Moorfields1474. In both instances no additional family members are currently available for testing. The Leu1808Pro was not seen in any of the other 56 probands tested through this region nor in a panel of 91 unaffected controls. The Lys663Asn variant was not seen in any of the other 658 samples tested.
Three previously unidentified variants were found in RP1. An Arg1595Gln (4784G→A) change in exon 4 was identified in one affected proband. Testing of unaffected controls revealed that this variant is present in ~1% of Caucasians, and therefore not a likely cause of disease. A silent substitution in exon 4, Ile2122Ile (6366T→C), was detected in one of the 56 probands tested through this region.
A third variant was identified in the 5′-untranslated region of exon 1. This variant, −315T→C was present in four affected probands, but not seen in any of the unaffected controls. Testing of additional family members in two of the four families revealed that the variant does not segregate with disease.
We designed PCR primers to flank four different microsatellite repeats that were located near the RP1 gene in BAC 18L28. These markers, D8S2607, D8S2608, D8S2609 and D8S2610, are located 3244, 25 314, 46 316 and 51 019 bp, respectively, from the 3′ end of the RP1 gene. Testing of these markers in CEPH parents showed that all four markers were polymorphic.
These four markers were analyzed in members of the families with the Arg677X, 2280–2284del and 2168–2181del RP1 mutations to determine whether families with the same mutation share a common ancestral haplotype or whether they represent independent mutation events. Due to the limited number of individuals from each family with the 2280–2284del and 2168–2181del RP1 mutations, the phase of marker alleles and the RP1 mutations could not be determined. The marker in phase with the Arg677X mutation could be determined in UCLA-RP01, Australian Family D, RFS137, UTAD103 and Moorfields1512. Three different D8S2607 and two different D8S2608 haplotypes were found among the five families tested (Table 3). These data suggest that the RP1 Arg677X mutations arose independently in some of the families or that this is an ancient mutation arising thousands of years ago. A mutation this old would allow for recombination between markers and the RP1 mutation, or for new marker alleles to arise, thus giving the appearance of a new mutation.
To determine the types and frequencies of disease-causing mutations in RP1, we tested 659 probands for mutations. We identified eight different disease-causing mutations in 17 of 250 apparently unrelated families with adRP. No mutations were identified in individuals with other forms of retinal degeneration. Each of these eight mutations results in a severely truncated RP1 protein about one-third the size of the wild-type protein.
We also identified two missense variants in one proband each. Because additional affected family members are not available at present, it is not possible to know whether these variants (Leu1808Pro and Lys663Asn) cosegregate with the disease phenotype. These variants may be associated with disease or they may just be benign variants. In this and previous studies we identified six polymorphic missense variants in RP1, so the RP1 protein may tolerate substantial variation at the amino acid level and only truncated RP1 proteins may cause disease. Residue 663 is conserved in the mouse, while residue 1808 is not conserved (GenBank accession no. AF155141; data not shown). These missense variants do not fall in the region of RP1 with sequence similarity to double-cortin and there is no functional information currently available on the remainder of the protein (3).
Although each of the eight different mutations identified in this study results in a similarly shortened RP1 protein, the phenotype of the affected individuals varies greatly. Some families have mild disease with late onset and slow progression, while others have more severe disease with onset in the teens followed by rapid vision loss. This variation is seen not only among families with different mutations, but also among families with the same mutation. Families with the Arg677X mutations have both mildly affected and severely affected individuals and there are several instances of non-penetrance. Furthermore, individuals in the same family can have a wide range of clinical symptoms and ages of onset. This is especially evident in the large UCLA-RP01 family (4,5). It is not known what modifying factors play a role in the severity of disease within and between families. It is possible that the haplotypes of the polymorphic amino acid variants, either in cis or in trans with the disease allele, play a role in the disease phenotype.
The clustering of mutations in such a small region of exon 4 is noteworthy. A larger sample size will be needed to determine whether disease-causing mutations are present in other regions of the RP1 gene. All 17 mutations found in this study were found in families with autosomal dominant RP. It is possible that mutations in other regions of the gene cause different types of retinal degeneration since we only examined a small region of the RP1 gene in diseases other than adRP. Despite these possibilities, this study does suggest that a large percentage of the RP1 mutations responsible for adRP lie within this small region.
RP1 mutations were found in 17 of 250 adRP families. However, of the 250 probands, only 56 were screened for mutations in the entire RP1 gene, so it is likely that more RP1 mutations are present in this population. These data suggest that RP1 mutations cause at least 7% of adRP cases. Furthermore, ~58% of these RP1 mutations are either the Arg677X or 2280–2284del mutations. Microsatellite markers tested in five families suggest that the Arg677X mutation has arisen independently several times or is an ancient mutation that arose so long ago that mutations have occurred in the microsatellite alleles or genetic recombination has occurred.
This study offers insight into the types and frequencies of RP1 mutations that cause inherited retinal diseases, in addition to providing screening methods for identifying new mutations. Future studies will determine whether missense mutations can cause retinitis pigmentosa, whether mutations in other regions of the RP1 gene can cause retinal disease, and whether the polymorphic amino acid variation seen on the ‘wild-type’ or mutated RP1 chromosome plays a role in disease expression.
Subjects tested in this study were diagnosed at one of the following sites: (i) the Anderson Vision Research Center, Retina Foundation of the Southwest, Dallas, TX; (ii) the Jules Stein Eye Institute, UCLA School of Medicine, Los Angeles, CA; (iii) Moorfields Eye Hospital, London, UK; and (iv) Eye Department, Leeds General Infirmary, Leeds, UK. Informed consent was obtained from all subjects tested. Families with RP1 mutations are shown in Figure 1. Individuals from whom DNA was available and tested are indicated in each pedigree.
Genomic DNA was extracted from peripheral blood using previously reported methods (6,17). The entire RP1 gene, except for amplimers 4F and 4P (which were sequenced), was screened in 56 individuals with adRP by using the primer pairs in Table 4. PCR was performed using Amplitaq Gold polymerase (Perkin Elmer, Branchburg, NJ) and standard cycling parameters. PCR products were radiolabeled by incorporating 1 μCi of [32P]dCTP. Most PCR products were digested with restriction enzymes (Stratagene, La Jolla, CA; New England Biolabs, Beverly, MA) as indicated in Table 1. PCR products were denatured and separated overnight on 0.6× MDE gels (FMC Bioproducts, Rockland, MD) at room temperature and 4°C. The gels were prepared in 0.6× TBE buffer and were dried and subjected to autoradiography after electrophoresis.
A targeted screen of Amplimer 4F+ was performed on 603 individuals using the forward primer from amplimer 4F and primer 4FR+ (5′-ACCGATTTCTCTTTTTTTAGGTG-3′). Genomic DNA was amplified and radiolabeled as described above. PCR products were either digested with HinfI (New England Biolabs) and run overnight on 0.6× MDE gels (FMC Bioproducts) as described above or were run without digestion on 1.0× SequaGel MD gels according to the manufacturer’s directions.
Two amplimers from exon 4 were screened for mutations by sequencing. Genomic DNA was amplified for amplimers 4F and 4P using standard cycling parameters and the primers described above. Any variants detected by SSCA or heteroduplex analysis were also amplified and sequenced. PCR products were sequenced using the original SSCP primers or the SSCP primers in conjuction with primers reported previously (GenBank accession no. AF143224–AF143226). Approximately 10–200 ng of PCR product was treated with shrimp alkaline phosphatase and Exonuclease I (United States Biochemical, Cleveland, OH), then sequenced according to the manufacturer’s protocols using the ABI BigDye cycle sequencing dye teminator kit (Applied Biosystems, Foster City, CA). Sequence reactions were purified using Sephadex G50 spin columns (Amika, Columbia, MD) and run on the ABI Prism 310 Genetic Analyzer (Perkin Elmer). Alternatively, PCR products were sequenced according to manufacturer’s protocols with the ThermoSequenase kit (Amersham Life Science, Piscataway, NJ).
All mutations and variants are designated based on the start of translation in the published cDNA sequence (GenBank accession no. AF143222) following nomenclature recommendations made by the Nomenclature Working Group (17).
In families where an RP1 mutation was found, additional family members were tested for mutations by sequencing or restriction digest. Amplimer 4F was amplified in family members with the Arg677X mutation and the PCR product was digested with TaqI. Fragments were separated and visualized on 2% agarose (Promega, Madison, WI) gels containing ethidium bromide.
The frequency of the Arg1595Gln variant in amplimer 4O was determined using a panel of 91 unrelated normal controls. Amplimer 4O was amplified and the PCR product was digested with HaeIII (New England Biolabs). Digested fragments were separated and visualized on a 2% agarose gel containing ethidium bromide.
The frequency of the −315T→C variant in Amplimer 1 was determined in a panel of 75 unrelated normal controls. Amplimer 1 was amplified, radiolabelled and digested as described above. Digested PCR product was separated overnight on 0.6× MDE (FMC Bioproducts) at 4°C. In order to determine whether the variant was cosegregating with disease, we took advantage of the fact that the −315T→C variant creates an EcoRII restriction site. Amplimer 1 was amplified from genomic DNA of family members from UTAD003 and UTAD076. PCR product was digested with EcoRII (Stratagene). Digested fragments were separated and visualized on 2% agarose gels containing ethidium bromide.
PCR primers were designed to flank repetitive regions contained in the same BAC as the RP1 gene (GenBank accession no. AF128525). DNA from 55 CEPH parents was typed to determine allele frequencies for each of the markers (GenBank accession nos 9865987, 9865975, 9865993 and 9865990). DNA from families with the Arg677X, 2280–2284del and 2168–2181del RP1 mutations were tested to determine which allele was associated with the disease. One PCR primer for each marker was labeled with [32P]ATP using polynucleotide kinase (Promega). Genomic DNA was amplified using standard cycling conditions and then separated on 6% denaturing acrylamide gels (Promega). Gels were dried and subjected to autoradiography after electrophoresis.
We are grateful to all of the family members for their continuing participation in this research study. Thanks go to Peter Lunt and Heather Skirton for access to RP patient DNAs and clinical information. We also thank Dr Richard Ruiz and Dr Richard Lewis for providing DNA samples and Ms Odessa June for technical assistance. This work was supported by grants from the Foundation Fighting Blindness and the George Gund Foundation, the William Stamps Farish Fund, the M.D. Anderson Foundation, the John S. Dunn Research Foundation, grants EY07142 and EY05235 from the National Eye Institute–National Institutes of Health and by a grant from the British Retinitis Pigmentosa Society and grants 035535/Z/96 and 054349/Z/98 from the Wellcome Trust.