Mutations in ABCR (ABCA4) in Patients with Stargardt Macular Degeneration or Cone-Rod Degeneration

Abstract

purpose. To determine the spectrum of ABCR mutations associated with Stargardt macular degeneration and cone–rod degeneration (CRD).

methods. One hundred eighteen unrelated patients with recessive Stargardt macular degeneration and eight with recessive CRD were screened for mutations in ABCR (ABCA4) by single-strand conformation polymorphism analysis. Variants were characterized by direct genomic sequencing. Segregation analysis was performed on the families of 20 patients in whom at least two or more likely pathogenic sequence changes were identified.

results. The authors found 77 sequence changes likely to be pathogenic: 21 null mutations (15 novel), 55 missense changes (26 novel), and one deletion of a consensus glycosylation site (also novel). Fifty-two patients with Stargardt macular degeneration (44% of those screened) and five with CRD each had two of these sequence changes or were homozygous for one of them. Segregation analyses in the families of 19 of these patients were informative and revealed that the index cases and all available affected siblings were compound heterozygotes or homozygotes. The authors found one instance of an apparently de novo mutation, Ile824Thr, in a patient. Thirty-seven (31%) of the 118 patients with Stargardt disease and one with CRD had only one likely pathogenic sequence change. Twenty-nine patients with Stargardt disease (25%) and two with CRD had no identified sequence changes.

conclusions. This report of 42 novel mutations brings the growing number of identified likely pathogenic sequence changes in ABCR to approximately 250.

Stargardt macular degeneration is characterized by the onset of central vision loss, usually by 20 years of age, progressive bilateral atrophy of the retinal pigment epithelium in the macula, accumulation of a lipofuscin-like substance in the retinal pigment epithelium, and a reduced foveal cone ERG. ^{1 2 3 4} It is recessively inherited and has an incidence of approximately 1 in 10,000. Mutations in the ABCR gene, located within chromosome 1p13, have been identified as a cause of recessive Stargardt macular degeneration. ^{5 6 7 8} Fundus flavimaculatus is allelic and is regarded by some as the same disorder with a later onset of symptoms and slower progression. ^{3 6 9 10 11} Unidentified loci linked to dominant forms of juvenile macular degeneration have been mapped to chromosomes, 6q11-14 (STGD3) and 4p (STGD4). ^{12 13 14}  

Recessive mutations in ABCR have been found in patients with Stargardt macular degeneration, fundus flavimaculatus, cone–rod degeneration (CRD; a panretinal photoreceptor degeneration with predominant loss of cone function that affects the macula early in its course) and retinitis pigmentosa (a panretinal photoreceptor degeneration usually associated with intraretinal pigmentary deposits). ^{5 15 16 17 18 19 20 21 22 23} To date, approximately 200 mutations in ABCR have been found in patients with these diseases. 

We report the results from an analysis of ABCR in 118 patients with juvenile macular degeneration and 8 with CRD. 

Materials and Methods

Ascertainment of Patients

The methods used in this study conformed to the tenets of the Declaration of Helsinki and received approval from the Institutional Review Boards at the Massachusetts Eye and Ear Infirmary and Harvard Medical School. Informed consent was obtained from all patients and family members who participated in the study. Patients with Stargardt macular degeneration were recruited from the Berman-Gund Laboratory (83 patients) and the Bascom Palmer Eye Institute (35 patients). All patients with Stargardt macular degeneration had unaffected parents by history or clinical evaluation. All had reduced central visual acuity before age 30 and characteristic fundus changes of Stargardt macular degeneration. In addition, all patients had diagnostic fluorescein angiograms and/or ERGs consistent with the diagnosis of Stargardt macular degeneration. ERGs showed full-field rod responses that were normal and full-field cone responses that were either normal or slightly reduced and delayed. Foveal ERGs were abnormal in every patient in whom they were measured. Fluorescein angiography revealed a dark choroid in those patients so evaluated. 

We also included eight patients with CRD in this mutation screen (five from the Berman-Gund Laboratory, two from the Bascom Palmer Eye Institute, and one from Schepens Retina Associates). Patients with CRD had unaffected parents. Ophthalmoscopy revealed panretinal degeneration affecting the macula more severely. Patients had severely reduced full-field cone ERG amplitudes (reduced 90% or more), moderately reduced cone–rod ERG amplitudes (reduced approximately 50% or more), and markedly delayed cone implicit times (≥40 msec; normal, ≤32 msec). Mixed cone–rod responses were typically 30 times larger than cone-isolated responses in young patients with this condition. 

Venous blood samples were obtained from participating patients and some of their relatives after informed consent was received. Leukocyte DNA was purified according to standard methods. 

Detection of Sequence Changes

DNA samples were screened with the single-strand conformation polymorphism (SSCP) technique for sequence changes in the 50 exons of ABCR (primer sequences are available at a Web site provided by the Massachusetts Eye and Ear Infirmary http://eyegene.meei.harvard.edu/OMGI/ABCR/primers.html). Each exon and 6 to 100 bp of flanking intron sequence were amplified from 20 ng of leukocyte DNA in 20 μl of a solution containing 20 μM dATP, dTTP, and dGTP; 2 μM dCTP, including 0.6 μCi[ -³²P]-dCTP (3000 Ci/mmol); 20 mM tris hydrochloride (pH 8.4 or 8.6); 0.25 to 5.0 mM MgCl₂; 50 mM KCl; 0.1 mg/ml bovine serum albumin; 20 pmol of each primer; 0.25 U Taq polymerase; and 0% or 10% dimethyl sulfoxide (DMSO). Conditions for the amplification of each exon were optimized for [MgCl₂], 0% or 10% DMSO, annealing temperature, and pH. In some cases, primer sets for two amplicons with a difference in size of at least 50 bp were combined in the same amplification reaction mixture. Samples were heated to 95°C for 4.5 minutes and incubated for 22 to 27 cycles of the following temperature sequence: 30 seconds at 94°C, 30 seconds at 52°C to 60°C, and 40 seconds at 72°C. Samples underwent a final incubation at 72°C for 5 minutes. After amplification, samples were diluted 1:1 with a solution of 40% formamide, 5 mM EDTA, 0.05% SDS, 0.25% bromphenol blue, 0.25% xylene cyanol, and 0.5× TBE (45 mM Tris-base, 45 mM boric acid, 1 mM disodium EDTA, pH 8.3). The amplified DNA was heat denatured (95°C for 3 minutes), and the resultant single-stranded fragments were separated by gel electrophoresis through 6% polyacrylamide TBE gels, with or without 10% glycerol or through a 5% polyacrylamide gel with TME (30 mM tris, 35 mM 2-[N-morpholino]ethanesulfonic acid, and 1 mM EDTA) ²⁴ at 8 to 16 W for 8 to 16 hours. Gels were transferred to Whatman paper (Whatman, Inc., Clifton, NJ), dried, and analyzed by autoradiography. DNA fragments that migrated at rates different from wild-type fragments were evaluated by direct genomic sequencing, according to standard methods. 

Individuals without a history of retinal degeneration or blood relatives without retinal degeneration were used as control subjects. Normal control subjects were screened for every likely pathogenic sequence change found in at least 6 of the 126 patients screened. 

Neutral and intron sequence changes not affecting the canonical splice-acceptor or splice-donor sites were analyzed for their likelihood of creating or destroying splice sites, by using the neural network software that is available at a Web site provided by the Berkeley Drosophila Genome Project, University of California, Berkeley (www.fruitfly.org/seq_tools/splice.html). ²⁵  

Results

Sequence Changes Identified

On analyzing 118 patients with Stargardt macular degeneration and 8 with CRD, we identified a total of 118 sequence changes (Table 1) , 77 of which were likely to be pathogenic (Fig. 1) . Forty-two of these 77 likely pathogenic sequence changes have not been previously reported. Twenty-one of the changes (including 15 novel changes) were interpreted as obviously null mutations: One was a missense change affecting the initiation codon (Met1Val), four were nonsense mutations, four changed canonical splice-site donor or acceptor sites, and 12 were frameshifts caused by the insertion or deletion of one or more base pairs (Table 1) . We have categorized the splice-site mutation IVS13+2T→C as a likely null mutation, although it may only reduce splice donor function because the dinucleotide GC is occasionally found as a splice donor in some genes. 

There were a total of 61 missense changes (26 novel) and a novel in-frame deletion of a consensus glycosylation site (deletion of Lys13-Trp15). Ten missense changes were found in 6 or more patients each. From 95 to 190 normal control individuals were evaluated for the presence of these 10 missense changes. Five of these 10 changes were statistically significantly less frequent in control subjects and were thus considered likely to be pathogenic. Specifically, Leu541Pro, Pro1380Leu, Gly1961Glu, and Leu2027Phe were not identified in any of the control individuals (P < 0.04). Gly863Ala was detected in 9 of 252 patient alleles and 2 of 380 normal control alleles tested (P = 0.009), and it was also considered pathogenic. Four missense changes, Arg212His, His423Arg, Arg943Gln, and Pro1948Leu, were found at approximately equal frequency among patients and normal control subjects (P > 0.05 by Fisher’s two-tailed analysis) and were thus categorized as nonpathogenic polymorphisms. Asn1868Ile was less frequent in patients with Stargardt disease than in control subjects (42/252 patient alleles and 50/170 control alleles; P = 0.002 by Fisher’s two-tailed analysis) and was therefore considered nonpathogenic. The abundance of this allele in control subjects may reflect an unappreciated difference in the ethnic ancestry of the patients versus the control subjects in this study. 

A rarely encountered missense change, Ala1637Thr, was interpreted as nonpathogenic because it was found in a patient (032-066) who also had two obviously null mutations determined to be allelic by segregation analysis, Lys356Ter and Gln1513(insC). Another change, Ala1038Val, was found in two patients (032-023 and 034-035), and in both cases segregation analysis showed that it was in cis with Leu541Pro, a combination also reported by Rivera et al. ²³ We interpreted Leu541Pro as pathogenic because it was found without the Ala1038Val change more frequently in patients than in control subjects. However, Sun et al. ²⁶ have shown abnormal ABCR function associated with either Ala1038Val or Leu541Pro, and it is therefore possible that both these changes are pathogenic in isolation. The remaining 49 missense changes were each found in five or fewer patients with Stargardt-CRD and were categorized as likely to be pathogenic. In all, 26 of the 55 likely pathogenic missense changes were novel. 

We detected 35 isocoding substitutions and intron changes. These were all interpreted as nonpathogenic. Only one of these, Val2244Val, was predicted to possibly change a splice site. This sequence change affects the first codon of exon 49. Splice-site prediction software identifies the expected splice-acceptor site at the end of intron 48 in the wild-type sequence (probability score 0.98). This changes to a predicted splice-donor site in the mutant sequence (probability score 0.70). Despite this computer-based prediction of an effect on intron splicing, the Val2244Val change was interpreted as nonpathogenic because one of the two patients who were heterozygous for this change (032-066) also had two obvious null mutations determined to be allelic by segregation analysis. 

Likely Pathogenic Sequence Changes Found in 95 of 126 Patients

Forty-nine patients with Stargardt and three with CRD had at least two likely pathogenic sequence changes. An additional three with Stargardt and two with CRD were homozygotes for a likely pathogenic sequence change (Leu244Pro, Pro1380Leu, Arg1640Gln, Cys2150Tyr, or Val1973[delG]). We conducted segregation analyses in 20 of these 57 patients’ families, and the results showed that the identified sequence changes segregated as expected for pathogenic alleles (Fig. 2) . 

One family had an index member (034-045) who was heterozygous for the missense changes Ile824Thr and Gly1961Glu. The Gly1961Glu allele ²⁷ was found to have been inherited from the patient’s mother, but the Ile824Thr allele was not detected in either parent or in any sibling. The results of genotyping analysis of six microsatellite markers in the patient and her parents were consistent with the designated paternity and we interpreted the Ile824Thr allele as a de novo mutation in this patient. The markers were D13S1316 (The Genome Database [GDB] accession number 614907), D13S1325 (GDB 615130), D13S1236 (GDB 601655), D13S1275 (GDB 604479), and D19S254 (GDB 189257; RB1.20) ²⁸ (results not shown). We did not conduct further studies to determine whether the Ile824Thr and Gly1961Glu changes were allelic in this index patient. 

Thirty-seven patients with Stargardt disease and one with CRD each had only one detectable sequence change that we categorized as likely to be pathogenic. There were a total of 24 unique sequence changes among these 38 patients. Five were null mutations, one was an in-frame deletion, and 18 were missense changes. Eight of the missense changes were found in at least one other index patient who was a compound heterozygote with another likely pathogenic sequence change. Of the remaining 10 missense changes, 8 affect amino acid residues that are identical in the mouse abc1 ²⁹ and human ABCR proteins (Table 2) . Four of these missense changes also occur in consensus sequences for functional motifs (Table 2) . 

Twenty-nine patients with Stargardt disease and two with CRD had no detectable sequence changes that were considered likely to be pathogenic. 

Nucleotide 2588G →C (Gly863Ala) and Nucleotide 2828G→A (Arg943Gln)

One likely pathogenic missense change, Gly863Ala, was frequently associated with a presumed nonpathogenic missense change, Arg943Gln. In fact, all 9 patients who were heterozygous carriers of Gly863Ala also carried Arg943Gln. Maugeri et al. ²¹ also found an association between the Gly863Ala and Arg943Gln changes, but they were unable to determine whether Gly863Ala by itself was pathogenic. In our study, the Arg943Gln change was present in five other patients without Gly863Ala, and it was present without Gly863Ala in 9 of 190 control alleles. In addition, a recently reported evaluation of the Gly863Ala mutant protein has shown that it has abnormal function in vitro. ²⁶ Taken together, these results indicate that Gly863Ala by itself is likely to be pathogenic and Arg943Gln by itself is not. 

Seven of the nine patients with Stargardt disease who were carrying Gly863Ala heterozygously also carried another missense change. Segregation analysis was conducted in the families of four of these 7 patients and the results in all four indicated that the two changes were allelic. Two patients who were heterozygous for the Gly863Ala allele had no other detectable changes likely to be pathogenic. 

Nullizygosity Associated with Panretinal Degeneration

We found putative null mutations (e.g., frameshifts, nonsense mutations, or intron splice-site alterations) in 26 (10%) of the 252 alleles screened in this study. Eleven of these patients were compound heterozygotes with one null mutation and a second missense mutation. All had Stargardt macular degeneration. Only four had two allelic null mutations and all four of these patients had the diagnosis of CRD. The ERGs recorded from three of these patients (007-014, 035-002, and 032-066 at ages 24, 21, and 19 years, respectively) showed severely reduced cone amplitudes in response to 30-Hz flickering light (0.2, 0.4, and 6.8 μV, respectively; normal, ≥50 μV) and showed moderately reduced rod-dominated amplitudes in response to single flashes of light (129, 170, and 180 μV, respectively; normal, ≥350μ V). The fourth patient (032-081) declined evaluation with an ERG. This is in contrast to patients with Stargardt disease who typically have full-field rod and cone ERG amplitudes in the normal or near-normal range at comparable ages. 

Discussion

We report 118 ABCR sequence changes, including 77 categorized as pathogenic. The 77 likely pathogenic changes include 21 null mutations (15 novel), 55 missense mutations (26 novel), and one novel deletion of a consensus glycosylation site. Of the 236 mutant alleles in the 118 patients with Stargardt macular degeneration, we found mutations in 141 (assuming that all patients with two likely pathogenic mutations were compound heterozygotes). The percentage of Stargardt alleles with an identified ABCR mutation (60%) is comparable to that found in another survey of a large group of patients: Lewis et al. ²⁰ found mutations in 57% of the 300 alleles in a set of 150 unrelated subjects. 

The absence of detected mutations in 29 patients with Stargardt in our survey is not strong evidence for a second recessive Stargardt disease macular degeneration gene besides ABCR. It is more likely that we missed mutations that lie outside the regions of the gene that were screened (e.g., intron sequence far from flanking exons, the promoter region and the 5′ and 3′ untranslated regions) or that the SSCP mutation screening technique failed to detect some mutations, especially large deletions or insertions that encompass one or both primer sites used for a PCR amplification. Furthermore, we conducted a search for examples of Stargardt macular degeneration not linked to the ABCR locus. Of the 29 index patients with no detected ABCR mutations, only one (032-070) had both an affected sibling who was willing to participate in our research and an informative polymorphism in the ABCR gene. The index patient and the affected sibling had identical ABCR alleles (data not shown). Thus, even in this multiplex family with Stargardt macular degeneration and no identified ABCR mutations, segregation analysis was consistent with ABCR being the disease locus. 

The criteria we and others used to classify sequence variants as“ likely” to be pathogenic are not perfect. It is possible that some of the likely pathogenic mutations, especially the missense changes, are nonpathogenic. In addition, because segregation analysis was not always possible or was not always informative, some of the patients with two likely pathogenic changes may not be compound heterozygotes but rather may have a complex allele with both changes in cis. Some of the missense mutants previously associated with Stargardt macular degeneration are reported to have abnormal adenosine triphosphatase (ATPase) activity stimulated by all-trans retinal in vitro. ²⁶ It is not yet clear whether this assay reliably distinguishes pathogenic missense variants from those that are nonpathogenic. 

Of the eight patients carrying Gly863Ala reported by Maugeri et al., ²¹ five were compound heterozygotes with a null mutation affecting the other allele. This led to speculation that the Gly863Ala sequence change is only pathogenic when present in compound heterozygotes who also carried a null allele. However, none of our patients with this change had a detected null allele. Rather, seven had another missense change and two had no other detectable changes likely to be pathogenic. 

ABCR mutations have been reported to cause a spectrum of vision disorders including Stargardt macular degeneration, CRD, and atypical retinitis pigmentosa. ^{15 16 17 18 19 20 21 27} Maugeri et al. ²¹ and others have proposed a model in which two ABCR alleles with severe (null) mutations result in a visual disorder with features that are more severe than typical Stargardt macular degeneration and that they have called atypical retinitis pigmentosa. According to this model, compound heterozygosity for a severe (null) and a moderately severe mutation causes CRD, whereas two moderately severe mutations or a mild and a severe allele together cause Stargardt macular degeneration. Our findings support this model because all four patients with allelic null mutations (035-002, 032-066, 032-081, and 007-014) whom we encountered had CRD, a panretinal degeneration much more severe than typical Stargardt macular degeneration. A fifth patient with CRD (032-030) was homozygous for the missense change Arg1640Gln and a sixth patient (007-009) had the missense change Gly2146Asp and no other sequence changes that we were able to detect. (The two remaining patients with CRD had no detected ABCR mutations.) Although we have determined that all six of these patients have CRD, the late stages of CRD can have fundus features and extinguished ERGs that are indistinguishable from the late stages of retinitis pigmentosa. Therefore, it may be clearer to consider together in one category both CRD and atypical retinitis pigmentosa caused by severe or null ABCR mutations. 

 Supported by Grants EY08683 and EY00169 from the National Eye Institute, Bethesda, Maryland; the Foundation Fighting Blindness, Owings Mills, Maryland; The Ruth and Milton Steinbach Fund, New York, New York; the V. Kann Rasmussen Foundation, Boston, Massachusetts; and Research to Prevent Blindness, New York, New York.

 Submitted for publication January 3, 2001; revised April 25, 2001; accepted May 25, 2001.

 Commercial relationships policy: N.

 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.

 

ABCR Sequence Changes Found in 118 Patients with Stargardt and 8 with CRD

Patient ID	Mutations (Amino Acid Based)	Sequence Change (Nucleotide Based)	Het/Hom	Other Sequence Changes
21 Null Mutations
071-004	Met1Val	ATG → GTC	Het	None
035-002*	Ser84(insCAAA) 30	251ins4	Het	IVS36 + 1G → A
034-039	Ser84(insCAAA) 30	251ins4	Het	Gly1961Glu
032-018	Arg152Ter 23	CGA → TGA	Het	Arg2107Cys
032-005	Ala222(del13bp)	666del13 [AAAGACGGTGCGC]	Het	None
032-039	Ala222(del13bp)	666del13 [AAAGACGGTGCGC]	Het	None
032-060	[Ser278(delT); Arg1300Gln]	[832delT; CGA → CAA]	Het	Pro1486Leu
032-066*	Lys356Ter	AAG → TAG	Het	Gln1513(insC)
032-072	—	IVS13+ 2T → C	Het	Val77Glu
032-073	Arg681Ter 21	CGA → TGA	Het	Leu1388Pro
034-016	Ser1071(insGT) 31	3212insGT	Het	None
032-065	Ser1071(insGT) 31	3212insGT	Het	None
035-003	Ile1114(delC) 5	3340delC	Het	Pro1380Leu
007-014*	—	IVS26+ 1G → A	Het	Asn1345(insCA)
007-014*	Asn1345(insCA)	4034insCA	Het	IVS26+ 1G → A
032-066*	Gln1513(insC)	4538insC	Het	Lys356Ter
032-010	Gln1513(insC)	4538insC	Het	None
032-024	Pro1570(delC) 16	4710delC	Het	Gly1961Glu
032-016	Thr1721 (delAC)	delete AC @ nt 5161	Het	Thr1525Met
035-002*	—	IVS36+ 1G → A 23	Het	Ser84(insCAAA)
034-031	Leu1741(del11)	5194del11 [GTGGTGGGCAT]	Het	Gly1961Glu
032-051	Trp1772Ter	TGG → TGA	Het	None
032-022	—	IVS41-2delA	Het	Gly1961Glu
032-081*	Val1973(delG)	5917delG	Hom	None
034-017	Gly2100(delG)	6300delG	Het	Gly1961Glu
55 Missense and One In-Frame Deletion
032-020	Cys54Tyr 15	TGC → TAC	Het	Gly863Ala
035-012	Cys54Tyr 15	TGC → TAC	Het	Arg1108Cys
071-007	Cys54Tyr 15	TGC → TAC	Het	Val935Ala
071-003	Asn58Lys	AAC → AAG	Het	Leu1201Arg
032-069	Ala60Val 15	GCG → GTG	Het	None
032-028	Gly65Glu 16	GGA → GAA	Het	None
032-072	Val77Glu	GTG → CAG	Het	IVS13+ 2T → C
034-013	Gln190His	CAG → CAC	Het	Gly1961Glu
032-076	Leu244Pro	CTG → CCG	Hom	None
032-012	Pro309Arg	CCA → CGA	Het	Arg1300Gln
032-054	Phe525Cys	TTT → TGT	Het	Ile1846Thr
032-046	Arg537Cys	CGT → TGT	Het	Val989Ala
034-038	Arg537Cys	CGT → TGT	Het	Gly863Ala
032-095	Leu541Pro 18	CTA → CCA	Het	None
034-022	Leu541Pro 18	CTA → CCA	Het	Leu2027Phe
035-001	Leu541Pro 18	CTA → CCA	Het	None
032-009	Leu541Pro 18	CTA → CCA	Het	None
032-023	[Leu541Pro 18 ; Ala1038Val 27 ]	[CTA → CCA; GCC → GTC]	Het	Gly863Ala
034-035	[Leu541Pro 18 ; Ala1038Val 27 ]	[CTA → CCA; GCC → GTC]	Het	Gly863Ala
032-011	Ala549Pro	GCC → CCC	Het	Gly1961Glu
032-044	Gly550Arg	GGA → AGA	Het	None
032-085	Arg602Gln	CGG → CAG	Het	Val643Met
032-090	Gly607Arg	GGG → AGG	Het	Leu2027Phe
032-085	Val643Met	GTG → ATG	Het	Arg602Gln
032-042	Val767Asp 30	GTC → GAG	Het	Pro1486Leu
071-006	Val767Asp 30	GTC → GAG	Het	Ile1562Thr
032-014	Leu797Pro	CTG → CCG	Het	Pro1486Leu
032-038	Trp821Arg 18	TGG → AGG	Het	None
034-045	Ile824Thr	ATC → ACC	Het	Gly1961Glu
032-056	Gly863Ala 5	GGA → GCA	Het	None
032-091	Gly863Ala 5	GGA → GCA	Het	None
032-020	Gly863Ala 5	GGA → GCA	Het	Cys54Tyr
032-023	Gly863Ala 5	GGA → GCA	Het	[Leu541Pro; Ala1038Val]
034-011	Gly863Ala 5	GGA → GCA	Het	Cys1488Arg
034-015	Gly863Ala 5	GGA → GCA	Het	Thr1525Met
034-035	Gly863Ala 5	GGA → GCA	Het	[Leu541Pro; Ala1038Val]
034-036	Gly863Ala 5	GGA → GCA	Het	Cys2150Arg
034-038	Gly863Ala 5	GGA → GCA	Het	Arg537Cys
071-007	Val935Ala	GTA → GCA	Het	Cys54Tyr
032-043	Arg943Trp	CGG → TGG	Het	Arg1108Leu
032-046	Val989Ala	GTT → GCT	Het	Arg537Cys
071-005	Arg1108Cys 18	CGC → TGC	Het	None
035-012	Arg1108Cys 18	CGC → TGC	Het	Cys54Tyr
032-043	Arg1108Leu 5	CGC → CTC	Het	Arg943Trp
032-097	Glu1122Lys 18	GAG → AAG	Het	None
035-019	Glu1122Lys 18	GAG → AAG	Het	None
071-003	Leu1201Arg 15	CTG → CGG	Het	Asn58Lys
032-012	Arg1300Gln	CGA → CAA	Het	Pro309Arg
032-068	Arg1300Gln	CGA → CAA	Het	None
032-013	Pro1380Leu 15	CCG → CTG	Het	Gly1961Glu
032-015	Pro1380Leu 15	CCG → CTG	Het	Gly1961Glu
032-027	Pro1380Leu 15	CCG → CTG	Het	Gly1961Glu
071-001	Pro1380Leu 15	CCG → CTG	Hom	None
034-020	Pro1380Leu 15	CCG → CTG	Het	Leu2027Phe
034-028	Pro1380Leu 15	CCG → CTG	Het	Gly1961Glu
034-044	Pro1380Leu 15	CCG → CTG	Het	Leu2027Phe
034-048	Pro1380Leu 15	CCG → CTG	Het	Gly1961Glu
035-003	Pro1380Leu 15	CCG → CTG	Het	Ile1114(delC)
032-073	Leu1388Pro	CTG → CCG	Het	Arg681Ter
034-040	Trp1408Arg 15	TGG → CGG	Het	Arg1640Trp
035-013	Trp1408Arg 15	TGG → CGG	Het	Arg1640Trp
032-060	Pro1486Leu 20	CCA → CTA	Het	[Ser278(delT); Arg1300Gln]
032-014	Pro1486Leu 20	CCA → CTA	Het	Leu797Pro
032-025	Pro1486Leu 20	CCA → CTA	Het	Asp1531Asn
032-042	Pro1486Leu 20	CCA → CTA	Het	Val767Asp
034-011	Cys1488Arg 15	TGC → CGC	Het	Gly863Ala
032-034	Cys1490Tyr 15	TGC → TAC	Het	Ile1846Thr
032-084	Thr1525Met 15	ACG → ATG	Het	Arg2139Trp
032-016	Thr1525Met 15	ACG → ATG	Het	Thr1721(delAC)
032-021	Thr1525Met 15	ACG → ATG	Het	None
032-041	Thr1525Met 15	ACG → ATG	Het	None
034-015	Thr1525Met 15	ACG → ATG	Het	Gly863Ala
032-049	Asp1531Asn 15	GAC → AAC	Het	Gly1961Glu
034-019	Asp1531Asn 15	GAC → AAC	Het	None
032-025	Asp1531Asn 15	GAC → AAC	Het	Pro1846Leu
071-006	Ile1562Thr 27	ATT → ACT	Het	Val767Asp
034-040	Arg1640Trp 18	CGG → TGG	Het	Trp1408Arg
035-013	Arg1640Trp 18	CGG → TGG	Het	Trp1408Arg
032-030*	Arg1640Gln	CGG → CAG	Hom	None
032-019	Pro1776Leu	CCC → CTC	Het	Gly1961Glu
032-034	Ile1846Thr 21	ATT → ACT	Het	Cys1490Tyr
032-054	Ile1846Thr 21	ATT → ACT	Het	Phe525Cys
032-011	Gly1961Glu 27	GGA → GAA	Het	Ala549Pro
032-013	Gly1961Glu 27	GGA → GAA	Het	Pro1380Leu
032-015	Gly1961Glu 27	GGA → GAA	Het	Pro1380Leu
032-019	Gly1961Glu 27	GGA → GAA	Het	Pro1776Leu
032-022	Gly1961Glu 27	GGA → GAA	Het	IVS41-2delA
032-024	Gly1961Glu 27	GGA → GAA	Het	Pro1570(delC)
032-027	Gly1961Glu 27	GGA → GAA	Het	Pro1380Leu
032-040	Gly1961Glu 27	GGA → GAA	Het	None
032-049	Gly1961Glu 27	GGA → GAA	Het	Asp1531Asn
034-013	Gly1961Glu 27	GGA → GAA	Het	Gln190His
034-017	Gly1961Glu 27	GGA → GAA	Het	Gly2100(delG)
034-021	Gly1961Glu 27	GGA → GAA	Het	None
034-025	Gly1961Glu 27	GGA → GAA	Het	None
034-028	Gly1961Glu 27	GGA → GAA	Het	Pro1380Leu
034-031	Gly1961Glu 27	GGA → GAA	Het	Leu1741(del11)
034-033	Gly1961Glu 27	GGA → GAA	Het	None
034-039	Gly1961Glu 27	GGA → GAA	Het	Ser84(insCAAA)
032-050	Gly1961Glu 27	GGA → GAA	Het	None
034-045	Gly1961Glu 27	GGA → GAA	Het	Ile824Thr
034-048	Gly1961Glu 27	GGA → GAA	Het	Pro1380Leu
032-003	Gly1977Ser 15	GGC → AGC	Het	Leu2027Phe
032-003	Leu2027Phe 5	CTC → TTC	Het	Gly1977Ser
032-090	Leu2027Phe 5	CTC → TTC	Het	Gly607Arg
034-006	Leu2027Phe 5	CTC → TTC	Het	None
034-020	Leu2027Phe 5	CTC → TTC	Het	Pro1380Leu
034-022	Leu2027Phe 5	CTC → TTC	Het	Leu541Pro
034-044	Leu2027Phe 5	CTC → TTC	Het	Pro1380Leu
035-011	Leu2027Phe 5	CTC → TTC	Het	None
032-063	Arg2030Gln 15	CGA → CAA	Het	None
032-093	Arg2030Gln 15	CGA → CAA	Het	None
071-002	Leu2035Pro	CTT → CCT	Het	None
032-064	Val2050Leu 5	GTT → CTT	Het	None
032-061	Arg2107His 18	CGC → CAC	Het	None
032-018	Arg2107Cys	CGC → TGC	Het	Arg152Ter
032-084	Arg2139Trp	CGG → TGG	Het	Thr1525Met
007-009*	Gly2146Asp	GGC → GAC	Het	None
032-045	Cys2150Tyr 16	TGT → TGC	Hom	None
034-036	Cys2150Arg	TGT → CGT	Het	Gly863Ala
034-026	Deletion Lys13-Trp15	38del9 [AGAACTGGA]	Het	None
034-037	Deletion Lys13-Trp15	38del9 [AGAACTGGA]	Het	None

Polymorphisms and Rare Variants	Sequence Change	Alleles Among 126 Patients (n)	Alleles Among Controls (n)	P , †
6 Nonpathogenic Missense Changes
Arg212His 23	CGC → CAC	8	10/(188)	0.3
His423Arg 23	CAC → CGC	34	14/(178)	0.09
Arg943Gln 27	CGG → CAG	15	9/(190)	0.7
Ala1637Thr	GCC → ACC	1	—	Not determined
Asn1868Ile 21	AAT → ATT	41	50/(170)	0.002
Pro1948Leu 21	CCA → CTA	10	4/(190)	0.4
35 Intron and Isocoding Changes
Pro47Pro	CCG → CCA
—	IVS3− 71delA
—	IVS3+ 20C → T
—	IVS3+ 26A → G
—	IVS3+ 92A → G
—	IVS6− 32T → C 23
Thr311Thr	ACC → ACT
—	IVS9− 14C → T
—	IVS10+ 6insC
—	IVS10+ 11delG
Ala626Ala	GCG → GCA
Leu988Leu	CTC → CTT
—	IVS22− 34A → G
Pro1401Pro 21	CCC → CCA
—	IVS32− 38C → T 23
—	IVS32− 15C → T
—	IVS33− 16delGT 23
—	IVS35− 32G → A
—	IVS38− 50delA
—	IVS38− 10T → C
—	IVS39+ 6del12
	[TGGTAGCCGAGG]: ins11
	[CGGTCGAGGGC]
—	IVS40 − 25A → C
Leu1894Leu 21	CTG → CTC
—	IVS41− 10A → G
Leu1938Leu 23	TTA → TTG
Pro1948Pro 21	CCA → CCG
Ile2023Ile	ATC → ATT
Ile2083Ile 27	ATC → ATT
—	IVS45+ 7G → A 32
Asp2095Asp 21	GAT → GAC
—	IVS48+ 21 C → T
—	IVS48− 3 T → C
—	IVS49− 85 C → T
—	IVS49+ 28 G → C
Val2244Val	GTG → GTA

 References for previously reported changes are indicated in superscript in the first column (for exon sequence changes) or the second column (for intron sequence changes). Those without references are novel. Patients who are compound heterozygotes are listed twice, once per mutation. Het, heterozygote; Hom, homozygote.

*  Patients with CRD. All others had Stargardt macular degeneration.

†  By the two-tailed Fisher’s exact test.

Figure 1.

 

Distribution of 77 ABCR mutations found in our patients. Numbers immediately above the bar indicate each exon number. Numbers within the boxes of the bar indicate the number of codons per exon. Numbers below the bar indicate the codon number at exon boundaries.

Figure 2.

 

Pedigrees of 20 patients in whom segregation analysis was conducted. Arrows: index cases. The ABCR alleles are indicated beneath or to the side of the symbol for each family member whose DNA was evaluated. The clinical findings of the three affected siblings in the family of the proband 032-030 with the R1640Q mutation have been previously reported. ³³

Table 2.

 

Missense Changes Found in Patients with No Other Detected ABCR Changes

Patient ID	Missense Change	Mouse abc1 34	Mouse abc2 34	Human ABCC 35
032-069	Ala60Val	Ala	N/A	Glu
032-028	Gly65Glu	Gly	N/A	Leu
032-044	Gly550Arg*	Gly	N/A	N/A
032-038	Trp821Arg, ‡	Trp	N/A	Trp
035-019, 032-097	Glu1122Lys	Glu	Glu	Glu
032-063, 032-093	Arg2030Gln, †	Arg	Arg	Arg
071-002	Leu2035Pro	Phe	Leu	Met
032-064	Val2050Leu	Phe	Val	Cys
032-061	Arg2107His	Arg	Arg	Arg
007-009	Gly2146Asp, ‡	Gly	Gly	Gly

 Residues at homologous locations in other ABCR proteins. N/A denotes that there is no homologous residue in this protein.

*  Predicted to reside in a transmembrane-spanning region.

†  Predicted to reside within a consensus protein kinase C phosphorylation site.

‡  Predicted to reside within a consensus N-myristoylation site.

The authors thank Ileana Cantillo, Terri McGee, and Jennifer McEvoy for technical assistance.