|Molecular Vision 2006;
Received 13 August 2006 | Accepted 15 November 2006 | Published 6 December 2006
Novel triple missense mutations of GUCY2D gene in Japanese family with cone-rod dystrophy: Possible use of genotyping microarray
Yoko Yamaji,1 Ayako Yoshida,1 Rumi Kuwahara,1, Ken
Yamamoto,2 Toshiaki Kubata,3 Tatsuro Ishibashi1
1Department of Ophthalmology, Graduate School of Medical Sciences, 2Department of Molecular Genetics, Medical Institute of Bioregulation, Kyushu University, 3Department of Ophthalmology, School of Medicine, University of Occupational and Environmental Health, Fukuoka, Japan
Correspondence to: Shigeo Yoshida, MD, PhD, Department of Ophthalmology, Kyushu University Graduate School of Medical Sciences, Fukuoka, 812-8582, Japan; Phone: 81-92-642-5648; FAX: 81-92-642-5663; email: firstname.lastname@example.org
Purpose: To report a novel mutation in the GUCY2D gene in a Japanese family with autosomal dominant cone-rod dystrophy (adCORD), and to examine the possible use of arrayed primer extension (APEX)-based genotyping chip in detecting mutations.
Methods: Genomic DNA was extracted from the peripheral blood of family members with adCORD. It was PCR-amplified, fragmented, and hybridized to APEX-based genotyping microarrays on which known disease-associated sequence variations were arrayed for patients with early-onset retinal dystrophy. All coding exons of the GUCY2D gene were directly sequenced. The PCR amplicon carrying a novel mutation was subcloned, and each clone was sequenced.
Results: Five single nucleotide polymorphisms in AIPL1, RPGRIP1, and GUCY2D were detected in the proband by microarray screening, and all were validated by direct sequencing. A novel heterozygous triple missense mutation of c.2540_2542delinsTCC (p.Gln847_Lys848delinsLeuGln amino acid substitutions) was found in both the proband and his father, and the three nucleotide changes were located on the same chromosome. Electroretinography (ERGs) demonstrated a significant reduction in rod function and a complete absence of cone function in both affected individuals.
Conclusions: A novel heterozygous triple consecutive missense mutation in the GUCY2D gene has been linked to adCORD. Our study demonstrates that the APEX-based gene screening can be used to identify simultaneously disease-modifying sequence changes as well as disease-causing mutations, once proper and comprehensive sites of sequence variations of the disease are arrayed.
Cone dystrophy (COD) and cone-rod dystrophy (CORD) belong to a subgroup of inherited chorioretinal dystrophies that is characterized by an initial degeneration of cone photoreceptors, causing an early decrease of visual acuity and color vision. The cone degeneration is followed by the degeneration of rod photoreceptors leading to progressive night blindness and peripheral visual field loss [1,2].
At present, COD and CORD are known to be genetically heterogeneous, and dominant, recessive, and X-linked inheritance patterns have been reported. The disease displays phenotypic and genotypic heterogeneity, and recent genetic studies have implicated a number of causative genes for CORD and COD, e.g., the CRX, GUCY2D, AIPL1, GUCA1A, RIMS1, and UNC119 genes for autosomal dominant (ad) CORD; the ABCA4 and RDH5 genes for autosomal recessive CORD; and the RPGR gene for X-linked recessive CORD .
Screening for mutations responsible for CORD has yet to become a routine procedure in clinical practice. This is mainly due to the large genetic heterogeneity. Because the current detection technologies are labor-intensive involving a screening of at least 100 amplicons (exons) that encompass the entire open reading frames of several disease-causative genes. Therefore, it would be more convenient if a rapid and efficient method is developed to identify disease-causing genes responsible for CORD.
The capabilities of genotyping microarrays have greatly improved during the past decade . These microarrays display hundreds of specific oligonucleotide probes that are precisely located on a small-formatted solid support. The array-based technologies have both research and potential clinical applications due to their ability to examine multiple genotypes from an individual simultaneously.
Among a number of microarray genotyping devices, the arrayed primer extension (APEX) is a method based on an array of oligonucleotides, immobilized at the 5' end on a glass surface . A patient's DNA is amplified by PCR, digested enzymatically, and annealed to the immobilized primers. This promotes sites for template-dependent DNA polymerase extension reactions using four unique fluorescently-labeled dideoxy nucleotides. This technique for genotyping microarrays has been used to detect different genotypes and mutations, including those for retinal diseases such as Stargardt disease  and Leber's congenital amaurosis (LCA) .
These microarrays are commercially available, and an LCA chip containing 307 sequence variants previously identified in eight genes associated with LCA as well as early-onset retinal degeneration: AIPL1, CRB1, CRX, GUCY2D, RPE65, RPGRIP1, MERTK, and LRAT is available . Because LCA is considered to be a kind of congenital stationary "cone-rod dystrophy" with high hypermetropia, panretinal degeneration, and greatly reduced visual acuity , and because AIPL1, CRX, and GUCY2D are also causative genes for adCORD, we hypothesized that the LCA microarray could be used as an initial screening tool for patients with CORD to identify the disease-causing mutation(s). During our examinations with the LCA chip, we coincidentally identified a novel heterogeneous triple mutation in the GUCY2D gene. The possible use of genotyping microarray is discussed.
We examined two affected and two nonaffected members of a family with adCORD (Figure 1). Full medical histories were taken from each individual, and ophthalmologic examinations, including best-corrected visual acuity, slit-lamp biomicroscopy, kinetic visual field examination, fundus examination, fluorescein angiography, and electroretinography (ERG) were performed on each subject. Color vision testing was performed with the panel D-15 and the Ishihara pseudoisochromatic plates.
Standardized full-field, photopic, flicker, and scotopic ERGs were recorded as recommended by the Standardization Committee of the International Society for Clinical Electrophysiology of Vision (ISCEV). After pupil dilatation with 0.5% tropicamide and 0.5% phenylephrine hydrochloride and 30 min of dark-adaptation, the scotopic ERGs were recorded with a white stimulus at an intensity of 0.12 cd.s/m2. Rod-cone mixed single-flash ERGs were elicited by a white stimulus at an intensity of 20 cd.s/m2. The photopic single-flash ERGs and the 30-Hz flicker ERGs were elicited with a white stimulus at an intensity of 1.2 and 0.6 cd.s/m2, respectively, on a white background of 25 cd/m2. Ten responses were averaged for the scotopic, the rod-cone mixed, and photopic ERGs, and 20 responses for the 30-Hz flicker ERGs.
Genomic DNA samples
This study was conducted in compliance with tenets of the Declaration of Helsinki, and approved by the Ethics Committee of the Kyushu University Hospital. All patients gave their informed consent prior to their inclusion in the study.
Genomic DNA was extracted from the blood of the patients and from non-affected individuals in the family using standard protocols to screen for genetic mutations [10,11].
Arrayed primer extension-based analysis
A detailed description of the APEX-based analysis is available at www.asperbio.com; Asper Biotech, Ltd. Briefly, selected exons of the AIPL1, CRB1, CRX, GUCY2D, RPE65, RPGRIP1, MERTK, and LRAT genes were amplified by polymerase chain reaction (PCR) following steps previously described . After the amplified products were concentrated and purified they were fragmented by adding thermolabile uracil N-glycosylase (Epicentre Technologies, Madison, WI) with heat of 95 °C . One-sixth of each amplified product was used in the primer extension reaction on the LCA genotyping microarray. The APEX mixture consisted of 10 μl of fragmented products, 4 units of Thermo Sequenase DNA polymerase (Amersham Pharmacia, Pittsburgh, PA), 2 μl of Thermo Sequenase reaction buffer (260 mM Tris HCl, pH 9.5/65 mM MgCl2; Amersham Pharmacia), and 2 μM final concentration of each fluorescein-labeled ddNTP: Texas Red-ddATP, Cy3-ddCTP, fluorescein-ddGTP, Cy5-ddUTP (Amersham Pharmacia) and NEN Life Science Products (Boston, MA). The entire mixture was applied to slides warmed to 58 °C. The reactions were allowed to proceed for 20 min under parafilm and stopped by washing at 95 °C 2 times at 90 s each in MilliQ water. A droplet of SlowFade Light Antifade Reagent (Molecular Probes, Carlsbad, CA) was applied to the microarrays to limit bleaching of the fluorescein. The slides were imaged with the Genorama imaging system (Asper Biotech, Tartu, Estonia) at 20 μm resolution. Gene sequence and mutations were identified by GENORAMA 3.0 genotyping software by using clustered signal patterns from a sequenced control DNA as the statistical reference (Figure 2). The PCR amplification of the samples, hybridization, and image analysis were performed by Asper Biotech (Tartu, Estonia).
The extracted array-identified variants were confirmed by direct sequencing with the Taq Dyedeoxy Terminator Cycle Sequencing Kit (Applied Biosystems, Foster City, CA), according to the manufacturer's instructions. Sequencing reactions were resolved on an ABI 3100 automated sequencer.
Molecular analysis of GUCY2D gene
All 18 coding exons of the GUCY2D gene, exon 2 of AIPL1, and exons 4 and 7 of RPGRIP1 genes were amplified by PCR and directly sequenced using the Taq Dyedeoxy Terminator Cycle Sequencing Kit as described in the literature [11,12]. Primer sequences used for amplification of GUCY2D were obtained from published sequences . To detect single nucleotide polymorphisms, we directly sequenced exons 10 and 13 of the GUCY2D gene from 136 chromosomes (35 men and 33 women) of unrelated normal Japanese individuals.
Subcloning of GUCY2D alleles
The PCR fragments of exon 13 of the GUCY2D gene in the patient were subcloned into TOPO-2 (Invitrogen, San Diego, CA) according to the manufacturer's protocol [14-16]. The cloned inserts were sequenced using T7 primer.
The proband (III-1) was a 50-year-old man who was referred to our hospital because of reduced vision. His best-corrected visual acuity was 0.4 in each eye with mild myopia. He stated that he had noted a decrease in his vision approximately 10 years earlier. Family history revealed that his father, paternal uncle, and grandmother also had depressed central vision (Figure 1). Slit-lamp examination showed that, except for mild cataracts in both eyes, the anterior segments of his eyes were normal. However, fundus examination showed bull's eye maculopathy similar to macular degeneration in both eyes (Figure 3A). There were no apparent abnormalities of the peripheral retina. The proband (and his father) did not have pendular nystagmus.
Fluorescein angiography revealed a granular hyperfluorescence corresponding to the macular degeneration (Figure 3B). He failed all of the Ishihara color plates and also failed the panel D-15 test with a tritan axis. Goldmann kinetic perimetry showed a central scotoma of about 10-20 ° (Figure 3C). The full-field photopic ERGs and the 30 Hz flicker ERGs were almost unrecordable in this proband. The amplitude of the scotopic b-wave was significantly reduced, and the a- and b-waves of the bright-flash mixed rod-cone ERGs, as well as the oscillatory potentials, were reduced (Figure 4).
The proband's 76-year-old father (II-1) showed a more severe clinical phenotype. His best-corrected visual acuity was 0.01 in the right eye and 0.02 in the left. He had chorioretinal atrophy in the macular region with a central scotoma in both eyes (Figure 3D,F). Fluorescein angiography showed a hypofluorescence around the macula, reflecting the retinochoroidal atrophy (Figure 3E). Because of poor visual acuity, the father could not take the Ishihara color vision and panel D-15 tests. His ERGs were similar to those of the proband, although they were more severely reduced (Figure 4).
Both unaffected family members had good best-corrected visual acuities of >1.0 in both eyes without any ocular abnormalities. All affected and unaffected family members had myopia of approximately -3.0 diopters.
Microarray and genetic analyses
A typical autosomal dominant hereditary pattern through three generations in this family prompted us to screen genes responsible for adCORD. We decided to use LCA genotyping microarray screening first because LCA is considered to be a type of congenital stationary "cone-rod dystrophy" with high hypermetropia, panretinal degeneration, and greatly depressed visual acuity . In addition, the sequences for the AIPL1, CRX, and GUCY2D genes were also represented on the microarray and have been reported to be causative genes for adCORD.
The screening by genotyping the proband revealed five sequence alterations, one of which was c.2101C>T (p.Pro701Ser) in the GUCY2D gene (Table 1). This mutation has been called pathogenic by the GENORAMA software (Figure 1B) because it has been reported as a disease-causative gene when the mutation is homozygous in patients with LCA . The other sequence alterations were called polymorphisms by the software. All 5 microarray-identified variants were then confirmed by direct sequencing.
Because a sequence alteration in the GUCY2D gene was called pathogenic, we directly sequenced all coding region of the GUCY2D gene. Direct sequencing revealed three consecutive novel heterozygous missense mutations of c.2540_2542delinsTCC that would predict p.Gln847_Lys848delinsLeuGln amino acid substitutions. These changes were found in the proband and his affected father (Figure 5A). None of the mutations was detected in the proband's unaffected mother and sister, or in 136 control individuals (data not shown).
To determine whether the three sequence alterations are located on the same chromosome or on two chromosomes, we subcloned the PCR product of exon 13 of the gene from the proband into a TOPO 2 vector and each clone was sequenced. Five of eight clones had the triple sequence changes (Figure 5C), whereas the remaining three showed the wild-type sequence (Figure 5B). Thus, the three consecutive nucleotide sequence alterations were concluded to be present in one chromosome.
We have identified complex novel missense mutations in the GUCY2D gene in two members of a Japanese family with adCORD. The GUCY2D gene encodes retinal guanylate cyclase, RetGC-1, which is a photoreceptor-specific enzyme that is involved in recovery during the phototransduction cascade. Its function is to synthesize cyclic guanosine monophosphate (cGMP) from 5'-GMP in the retina and is responsible for increasing the proportion of open cGMP-gated channels in the dark-adapted state.
Our study confirmed the idea that a heterozygous mutation of GUCY2D not involving codon 838 can also be linked to CORD , although it has been reported that codon 838 appeared to be particularly prone to mutational changes in the heterozygous state.
RetGC-1 is believed to exist in a dimeric state , and mutations at sites 847-848, as observed in our case, are located within the putative dimerization domain of the RetGC-1 protein . Moreover, alignment of part of this domain of human RetGC-1 and other members of the subgroup (human RetGC-1 , RetGC-2 , rat GC-E , GC-F , mouse GC-E , and bovine ROS-GC ) showed that 848 and 847 are highly conserved among the sensory cyclase family members (Figure 5D), suggesting that both mutations are critical. In support of this, In Silico analysis using SIFT and PolyPhen, which are Web-based applications that use phylogenetic and structural information from homologous proteins, showed that both sequence alterations might be pathogenic (data not shown).
It appears that the GUCY2D gene is prone to complex missense mutations. Thus far, three research groups have reported heterozygous complex missense mutations in the GUCY2D gene: a triple mutation of p.Glu837_Arg838_Thr839delinsAspCysMet , a complex mutation of p.Glu837_Arg838delinsAspSer [24,25], and a complex missense mutation of p.Ile915Thr and p.Gly917Arg in a Japanese family . It was suggested that the high mutability of the CpG sequences in the GUCY2D gene may account for the multiple mutations . However, this is not always the case because the multiple mutations in our case were not located in the CpG sequence. Thus, the mechanism responsible for the generation of multiple mutations in the GUCY2D gene remains unclear.
Electrophysiological examination showed that both the proband and his father had significant loss of the scotopic system in addition to the absence of cone responses with the reduction greater in the father. This is consistent with an earlier idea that a moderate to severe loss of rod function was present in families with multiple mutations compared with families with a single mutation showing a marked loss of cone function with only minimal rod involvement [19,24,25,27]. Therefore, different mutations in this dimerization domain of the GUCY2D gene can result in differing severities of CORD, especially that of the scotopic system.
Our study demonstrated the potential use of genotyping microarrays for the simultaneous detection of not only the causative but also the modifying sequence alterations in one test. Although the heterozygous c. 2101C>T (p.Pro701Ser) variant was considered to be not causative of the disease because it was detected in normal control individuals heterozygously (data not shown), we assume that the heterozygous c. 2101C>T might modify the disease phenotype as an additive effect, with the c.2540_2542delinsTCC change being the major disease-causing change in our family. This is because homozygous c. 2101C>T can be a causative alteration in LCA. Similarly, it is also possible that the other four silent mutations in the retinal disease causing genes, AIPL, GUCY2D, and RPGRIP1, detected as polymorphisms, could serve as disease modifier genes with another major gene defect occurring simultaneously, even if those changes do not significantly affect healthy individuals. Moreover, the technology that allows the examination of multiple genes simultaneously might also reveal digenic or multigenic inheritance mechanism of the CORD .
Because the mutation spectrum of GUCY2D in LCA patients is significantly different from that in CORD patients, and because the mutation variations in Japanese patients with CORD possibly may be significantly different from that in Caucasian patients, we were not able to directly detect the disease-causing mutations; however, because all of sequence alterations detected by the microarray were confirmed by independent direct sequencing, it is likely that the APEX-based gene test platform itself provided an accurate and efficient means for detecting genotypes in each individual .
Generating a custom-made APEX-based genotyping microarray on which almost all of the CORD disease-causative genes are arrayed by collecting data from the ever growing Web-based mutation database of GUCY2D, as well as information on mutations obtained in individual laboratories in ethnically diverse populations, may eventually offer a unique and reliable diagnostic tool. This should then enhance the detection rate of not only disease-causative but also of modifying-sequence changes and may accelerate our understanding of the basic mechanisms underlying CORD and its phenotypic variability which facilitate prospective diagnosis.
This work was supported in part by grants from the Ministry of Education, Science, Sports and Culture, Japan (T.I. and S.Y.), Japanese Retinitis Pigmentosa Society (S.Y.), Clinical Research Foundation (S.Y.), The Japan Health Foundation (S.Y.), and Japan National Society for the Prevention of Blindness (S.Y.).
1. Moore AT. Cone and cone-rod dystrophies. J Med Genet 1992; 29:289-90.
2. Rabb MF, Tso MO, Fishman GA. Cone-rod dystrophy. A clinical and histopathologic report. Ophthalmology 1986; 93:1443-51.
3. Michaelides M, Hunt DM, Moore AT. The genetics of inherited macular dystrophies. J Med Genet 2003; 40:641-50.
4. Mockler TC, Chan S, Sundaresan A, Chen H, Jacobsen SE, Ecker JR. Applications of DNA tiling arrays for whole-genome analysis. Genomics 2005; 85:1-15. Erratum in: Genomics 2005; 85:655.
5. Kurg A, Tonisson N, Georgiou I, Shumaker J, Tollett J, Metspalu A. Arrayed primer extension: solid-phase four-color DNA resequencing and mutation detection technology. Genet Test 2000; 4:1-7.
6. Jaakson K, Zernant J, Kulm M, Hutchinson A, Tonisson N, Glavac D, Ravnik-Glavac M, Hawlina M, Meltzer MR, Caruso RC, Testa F, Maugeri A, Hoyng CB, Gouras P, Simonelli F, Lewis RA, Lupski JR, Cremers FP, Allikmets R. Genotyping microarray (gene chip) for the ABCR (ABCA4) gene. Hum Mutat 2003; 22:395-403.
7. Zernant J, Kulm M, Dharmaraj S, den Hollander AI, Perrault I, Preising MN, Lorenz B, Kaplan J, Cremers FP, Maumenee I, Koenekoop RK, Allikmets R. Genotyping microarray (disease chip) for Leber congenital amaurosis: detection of modifier alleles. Invest Ophthalmol Vis Sci 2005; 46:3052-9.
8. Allikmets R. Leber congenital amaurosis: a genetic paradigm. Ophthalmic Genet 2004; 25:67-79.
9. Perrault I, Hanein S, Gerber S, Lebail B, Vlajnik P, Barbet F, Ducroq D, Dufier JL, Munnich A, Kaplan J, Rozet JM. A novel mutation in the GUCY2D gene responsible for an early onset severe RP different from the usual GUCY2D-LCA phenotype. Hum Mutat 2005; 25:222.
10. Yoshida S, Yamaji Y, Yoshida A, Noda Y, Kumano Y, Ishibashi T. Rapid genotyping for most common TGFBI mutations with real-time polymerase chain reaction. Hum Genet 2005; 116:518-24.
11. Yoshida S, Kumano Y, Yoshida A, Hisatomi T, Matsui H, Nishida T, Ishibashi T, Matsui T. An analysis of BIGH3 mutations in patients with corneal dystrophies in the Kyushu district of Japan. Jpn J Ophthalmol 2002; 46:469-71.
12. Yoshida S, Yamaji Y, Kuwahara R, Yoshida A, Hisatomi T, Ueno A, Ishibashi T. Novel mutation in exon 2 of COL2A1 gene in Japanese family with Stickler Syndrome type I. Eye 2006; 20:743-5.
13. Perrault I, Rozet JM, Calvas P, Gerber S, Camuzat A, Dollfus H, Chatelin S, Souied E, Ghazi I, Leowski C, Bonnemaison M, Le Paslier D, Frezal J, Dufier JL, Pittler S, Munnich A, Kaplan J. Retinal-specific guanylate cyclase gene mutations in Leber's congenital amaurosis. Nat Genet 1996; 14:461-4.
14. Yoshida S, Yoshida A, Ishibashi T, Elner SG, Elner VM. Role of MCP-1 and MIP-1alpha in retinal neovascularization during postischemic inflammation in a mouse model of retinal neovascularization. J Leukoc Biol 2003; 73:137-44.
15. Yoshida S, Yoshida A, Matsui H, Takada Y, Ishibashi T. Involvement of macrophage chemotactic protein-1 and interleukin-1beta during inflammatory but not basic fibroblast growth factor-dependent neovascularization in the mouse cornea. Lab Invest 2003; 83:927-38.
16. Yoshida A, Yoshida S, Khalil AK, Ishibashi T, Inomata H. Role of NF-kappaB-mediated interleukin-8 expression in intraocular neovascularization. Invest Ophthalmol Vis Sci 1998; 39:1097-106.
17. Ito S, Nakamura M, Nuno Y, Ohnishi Y, Nishida T, Miyake Y. Novel complex GUCY2D mutation in Japanese family with cone-rod dystrophy. Invest Ophthalmol Vis Sci 2004; 45:1480-5.
18. Yang RB, Garbers DL. Two eye guanylyl cyclases are expressed in the same photoreceptor cells and form homomers in preference to heteromers. J Biol Chem 1997; 272:13738-42.
19. Kelsell RE, Gregory-Evans K, Payne AM, Perrault I, Kaplan J, Yang RB, Garbers DL, Bird AC, Moore AT, Hunt DM. Mutations in the retinal guanylate cyclase (RETGC-1) gene in dominant cone-rod dystrophy. Hum Mol Genet 1998; 7:1179-84.
20. Yang RB, Foster DC, Garbers DL, Fulle HJ. Two membrane forms of guanylyl cyclase found in the eye. Proc Natl Acad Sci U S A 1995; 92:602-6.
21. Lowe DG, Dizhoor AM, Liu K, Gu Q, Spencer M, Laura R, Lu L, Hurley JB. Cloning and expression of a second photoreceptor-specific membrane retina guanylyl cyclase (RetGC), RetGC-2. Proc Natl Acad Sci U S A 1995; 92:5535-9.
22. Goraczniak RM, Duda T, Sitaramayya A, Sharma RK. Structural and functional characterization of the rod outer segment membrane guanylate cyclase. Biochem J 1994; 302:455-61.
23. Perrault I, Rozet JM, Gerber S, Kelsell RE, Souied E, Cabot A, Hunt DM, Munnich A, Kaplan J. A retGC-1 mutation in autosomal dominant cone-rod dystrophy. Am J Hum Genet 1998; 63:651-4.
24. Gregory-Evans K, Kelsell RE, Gregory-Evans CY, Downes SM, Fitzke FW, Holder GE, Simunovic M, Mollon JD, Taylor R, Hunt DM, Bird AC, Moore AT. Autosomal dominant cone-rod retinal dystrophy (CORD6) from heterozygous mutation of GUCY2D, which encodes retinal guanylate cyclase. Ophthalmology 2000; 107:55-61.
25. Payne AM, Morris AG, Downes SM, Johnson S, Bird AC, Moore AT, Bhattacharya SS, Hunt DM. Clustering and frequency of mutations in the retinal guanylate cyclase (GUCY2D) gene in patients with dominant cone-rod dystrophies. J Med Genet 2001; 38:611-4.
26. Ito S, Nakamura M, Ohnishi Y, Miyake Y. Autosomal dominant cone-rod dystrophy with R838H and R838C mutations in the GUCY2D gene in Japanese patients. Jpn J Ophthalmol 2004; 48:228-35.
27. Downes SM, Payne AM, Kelsell RE, Fitzke FW, Holder GE, Hunt DM, Moore AT, Bird AC. Autosomal dominant cone-rod dystrophy with mutations in the guanylate cyclase 2D gene encoding retinal guanylate cyclase-1. Arch Ophthalmol 2001; 119:1667-73.
28. Kajiwara K, Berson EL, Dryja TP. Digenic retinitis pigmentosa due to mutations at the unlinked peripherin/RDS and ROM1 loci. Science 1994; 264:1604-8.