A Novel Gene for Autosomal Dominant Stargardt-Like Macular Dystrophy with Homology to the SUR4 Protein Family

A Novel Gene for Autosomal Dominant Stargardt-like Macular Dystrophy with Homology to the SUR4 Protein Family

Albert O. Edwards,1 Larry A. Donoso,2 and Robert Ritter, III1

PURPOSE. To describe a novel gene causing a Stargardt-like fatty acid biosynthesis in the pathogenesis of macular degen- phenotype in a family with dominant macular dystrophy and eration. The PCR-based assay for the 5-bp deletion will facilitate the exclusion of all known genes within the disease locus. more accurate genetic counseling and identification of other METHODS. Meiotic breakpoint mapping in a family of 2314 branches of the family. (Invest Ophthalmol Vis Sci. 2001;42: individuals enabled refinement of the location of the disease 2652–2663) gene. The genomic organization and expression profile of known and putative genes within the critical region were determined using bioinformatics, cDNA cloning, and RT-PCR. acular dystrophies with subretinal flecks may arise from The coding sequence of genes expressed within the retina was Mmutations at multiple disease loci,1,2 the most common scanned for mutations, by using DNA sequencing. of which is the ABCA4 gene on chromosome 1 that gives rise 3,4 RESULTS. The disease-causing gene (STGD3) was further local- to Stargardt disease (STGD1) or fundus flavimaculatus. We 2,5 ized to 562 kb on chromosome 6 between D6S460 and a new previously reported a founder effect for a dominant condi- polymorphic marker centromeric to D6S1707. Of the four tion (STGD3) phenotypically similar to Stargardt disease local- genes identified within this region, all were expressed in the ized to chromosome 6q.2,6 A genealogical and molecular inves- retina or retinal pigment epithelium. The only coding DNA tigation of several families with autosomal dominant Stargardt- sequence variant identified in these four genes was a 5-bp like macular dystrophy led to the recognition that all studied deletion in exon 6 of ELOVL4. The deletion is predicted to lead families were related through a common founder that immi- to a truncated protein with a net loss of 44 amino acids, grated to North America in 1730.2,5–8 We also showed that the including a dilysine endoplasmic reticulum retention motif. Stargardt-like disease (STGD2), previously reported to be The ELOVL4 gene is the fourth known example of a predicted linked to chromosome 13,9 was actually a branch of the family human protein with homology to mammalian and yeast en- we had been studying on chromosome 6.5 zymes involved in the membrane-bound fatty acid chain elon- Disease-causing mutations segregating within single families gation system. The genomic organization of ELOVL4 and can be difficult to identify with certainty, because all affected primer sets for exon amplification are presented. patients share the identical genomic DNA segment on which CONCLUSIONS. ELOVL4 causes macular dystrophy in this large the founding mutation arose. Thus, all DNA sequence variation family distributed throughout North America and implicates within the disease locus or critical region is shared by all affected patients and segregates with the trait. In addition to the customary criteria for distinguishing between polymor- From the 1Department of Ophthalmology, University of Texas phisms and mutations, exclusion of other potential disease- Southwestern Medical Center, Dallas; and the 2Henry and Corinne causing sequence variation within the disease locus can con- 10 Bower Laboratory, Wills Eye Hospital, Philadelphia, Pennsylvania. firm the mutation. Therefore, it is critical to exclude Supported in part by Grant EY12699 from the National Institutes sequence variation in all genes within the disease locus. of Health (AOE, LAD), Career Development Awards from Research to The disease locus on chromosome 6q14 for this autosomal Prevent Blindness and the Foundation Fighting Blindness (AOE), and dominant Stargardt-like macular dystrophy family has been unrestricted departmental funds from Research to Prevent Blindness (University of Texas Southwestern and Wills Eye Hospital). The Scholl- progressively refined to between novel markers within the maier Foundation, Fort Worth, Texas; the Anne Marie and Thomas B. region defined by D6S1625 and D6S1707 by us and between 11,12 Walker, Jr. Fund for Age-Related Macular Degeneration, Dallas, Texas; D6S460 and D6S391 by other groups. While this manu- and the Walter Center for Macular Degeneration, Dallas, Texas pro- script was in preparation, a 5-bp deletion in the ELOVL4 gene vided additional support at University of Texas Southwestern. The in three branches of this family was reported by Zhang et al.11 Henry and Corinne Bower Laboratory for Macular Degeneration; the In addition to ELOVL4, they reported exclusion of two other Elizabeth C. King Trust; the estates of Margaret Mercer, Harry B. genes within their 3100-kb critical region between D6S460 and Wright, and Martha W. S. Rogers; the Association for Macular Diseases; 11 and Macular Degeneration International, all of Philadelphia, Pennsyl- D6S391. We are aware of eight additional genes within their vania, provided additional support at Wills Eye Hospital. LAD is the reported critical region that were not excluded and thus could Thomas D. Duane Professor of Ophthalmology, Wills Eye Hospital and harbor a disease-causing sequence variation. Herein, we report Jefferson Medical College, Thomas Jefferson University, Philadelphia, our independent identification and characterization of the Pennsylvania. gene for autosomal dominant Stargardt-like dominant macular Submitted for publication March 26, 2001; revised May 25, 2001; accepted June 5, 2001. dystrophy and the exclusion by DNA sequencing of the com- Commercial relationships policy: N. plete coding region of all candidate genes within our 562-kb The publication costs of this article were defrayed in part by page critical region, defined by D6S460 and a novel marker charge payment. This article must therefore be marked “advertise- (260P22.A) centromeric to D6S1707. The expression profile in ment” in accordance with 18 U.S.C. §1734 solely to indicate this fact. human retina, genomic organization, and known functional Corresponding author: Albert O. Edwards, Department of Oph- information are presented for 15 genes in or near the critical thalmology, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, TX 75390-9057. region for STGD3 that are candidates for other chromosome 6q [email protected] retinopathies.

Downloaded from iovs.arvojournals.org on 09/24/2021 IOVS, October 2001, Vol. 42, No. 11 Gene for Dominant Stargardt-like Dystrophy 2653

FIGURE 1. Color photograph of the retina of a 52-year-old family member with autosomal dominant Stargardt-like macular dystrophy, showing RPE atrophy, subretinal ﬂecks, and temporal pallor of the optic nerve.

MATERIALS AND METHODS 74°C for 5 minutes. Amplified products were denatured for 4 minutes at 94°C in the presence of formamide stop dye (6.5 ␮l), snap cooled on DNA Extraction ice, resolved at room temperature on 8% denaturing polyacrylamide White blood cells were isolated by centrifugation of whole blood at gels at 80 W, and transferred to nylon membranes (Roche Molecular 1000g and DNA purified with a kit (Masterpure Genomic DNA; Epi- Biochemicals, Indianapolis, IN) overnight. DNA was cross-linked to the centre Technologies, Madison, WI), according to the manufacturer’s membranes for 5 minutes (Photoprep I; Photodyne, Newberlin, WI). instructions. This research was approved by the institutional review The immobilized products were denatured for 10 minutes in 0.4 M boards and followed the tenets of the Declaration of Helsinki. NaOH and neutralized in 2ϫ SSC. An oligonucleotide containing an AC15 repetitive sequence was Polymorphic Marker Identification, 3Ј-tailed with digoxigenin (DIG)-11-dUTP and terminal transferase ac- Amplification, and Analysis cording to the manufacturer’s instructions (Roche Molecular Biochemi- cals) and hybridized to the resolved PCR products at 42°C in 5ϫ SSC, Previously reported primers for short tandem repeat (STR) amplifica- 1% blocking reagent, 0.1% N-laurosarkosine, 0.02% SDS. Hybridized tion were purchased from Research Genetics (Huntsville, AL). Novel membranes were washed in 6ϫ SSC and 0.1% SDS three times for 5 dinucleotide (AC)n repeat motifs were identified using a BLAST 2 Sequences search (available publicly from the National Center for minutes, followed by a single wash in 0.1 M maleic acid, 0.15 NaCl, and Biotechnology Information [NCBI] at http://www.ncbi.nlm.nih.gov/ 0.3% Tween-20 at pH 7.5 for 1 minute. The products were visualized BLAST/) of finished and unfinished genomic sequence within the by incubating membranes with alkaline phosphatase-conjugated anti- STDG3 critical region produced by the Human Chromosome 6 Se- DIG antibody for 60 minutes, washing twice in 0.1 M maleic acid and quencing Group at the Sanger Centre (Cambridgeshire, UK) with an 0.15 M NaCl at pH 7.5 for 15 minutes, followed by a single equilibra- tion in 100 mM Tris-HCl, 100 mM NaCl, and 50 mM MgCL at pH 9.5 (AC)60 sequence motif. Primers were designed in flanking sequence for 2 amplification. Polymerase chain reactions (PCRs) were performed on for 2 minutes. The membranes were incubated with the chromogenic genomic DNA (12.5 ng) in the presence of 1.5 mM MgCl, 1.25 nM each substrate (nitroblue tetrazolium chloride-5-bromo-4-chloro-3-indoyl dNTP, 2.5 pM each primer, and 0.11 U Taq polymerase in a final phosphate 4-toludine sale [NBT/BCIP]) until the desired band intensity volume of 6.5 ␮l under the following conditions: initial denaturation at was achieved. All results were analyzed by at least two independent 95°C for 4 minutes, followed by 30 cycles at 94°C for 10 seconds, 55°C investigators and compared with a control sample (CEPH 1331-01) for for 10 seconds, and 74°C for 10 seconds, and then a final extension at registering genotypes between assays.

Downloaded from iovs.arvojournals.org on 09/24/2021 2654 Edwards et al. IOVS, October 2001, Vol. 42, No. 11

TABLE 1. Meiotic Breakpoint Mapping of Autosomal Dominant Stargardt-like Macular Dystrophy

Patients*

B1001 B1025 B2003 S93 S110 B3015 B2078 B2017 B2022 S111 Marker (A) (A) (A) (A) (A) (A) (A) (A) (N/17) (N/29)

D6S430 X X X X X D D D6S313 U U X X X D U D D6S1681 X X X X X D D D D6S280 U U D X X D U D D6S1596 D U D X X D D D D6S1659 D D D U X D U D D6S406 D D D X X D D D D6S1622 U U D D X D D D D6S456 D D D D U D D D U U 234P15.C D U D D 234P15.A U U U U 234P15.B U U U D6S1589 D D D D D D D U U D 474L11.A D D D D 474L11.B U D U U 351K21.A D D D D 134M13.A U U U 134M13.B U U U 472A9.A D D D D D6S1625 D D D D D U D U U D 501M23.A U U U U 501M23.B U U U U 501M23.C U U U U 551A13.C U D U D 551A13.A D D X D 1007B16.A U U U U 1007B16.B U U U U 411F9.A D D X X 411F9.B D D X X 424E5.A D U X U 136A11.A U D X U D6S284 U U D U D U U U U U D6S286 U U D D U D U U X U D6S460 D U D D D D D X X 75K24.A U D U U 75K24.B U U X U 159G19.A D D U X 357D13.A U U U U 357D13.B D D X X 130P1.A D U X X 260P22.A D X U X D6S1707 U U U D D X D U U X D6S463 U U U U U X U U U D6S1646 U U D D D U U U X D6S445 D D U D D U D D X U D6S1652 D D D D D U U U U X D6S1634 D U D D D U X D U X D6S1609 D D D D D X U D X U D6S1627 D D D D D D X X D6S1601 U U D D U X X D6S1644 U U U D D X U U D6S1613 D D U D D X X X

Genotypes were not determined if table cells are empty. Shaded region is excluded based on informative recombinant markers. Marker order is centromeric to telomeric and based primarily on genomic sequence and mapping data from the Sanger Centre. D, informative alleles from the disease haplotype; U, uninformative alleles; X, informative alleles from haplotypes not associated with disease. * Subjects are identiﬁed as affected (A) or normal (N)/age at examination.

Preparation of Human Retina and Retinal isolated with minor modiﬁcations of a previously described method.13 Pigment Epithelium Brieﬂy, the RPE surface was treated with 0.25% trypsin and 0.5% EDTA in Hanks’ balanced salt solution at 37°C for 20 minutes or until the RPE Adult human eyes were purchased from the Lions Eye Bank (Omaha, NE). The anterior segment was removed by circumferential incision cell layer appeared marbled. The trypsin solution was removed, and posterior to the ora serrata. The retina was removed and snap frozen in the RPE cells were dislodged and suspended with a stream of Ham’s liquid nitrogen. The retinal pigment epithelium (RPE)–choroid com- F-12 medium without serum from a pipet. The cells were collected in plex was rinsed with sterile Hanks’ balanced salt solution. The RPE was sterile tubes and centrifuged at 50g for 3 minutes and then washed

Downloaded from iovs.arvojournals.org on 09/24/2021 IOVS, October 2001, Vol. 42, No. 11 Gene for Dominant Stargardt-like Dystrophy 2655

TABLE 2. New Polymorphic Dinucleotide (AC)n Short Tandem Repeat Markers

Product Observed Name* Forward Primer Reverse Primer Size (bp) Heterozygosity

234P15.C CAAGCTGGCTGGTGTTTGTA TCCTCAGTCAACAACTCCGC 162 0.90 234P15.A CTGGAACCCCATCCCTATTT CTATTTGAAAGGGCAGGAACT 169 0.63 234P15.B CAGCCAAATTTTCTGGTCTGA ATGCAAAAAGACGCAGAGAT 170 0.78 474L11.A GGCAAGGAGGAGAATTGCTT TTCGATATTTGCCCTGTTGA 156 0.78 474L11.B CCACAGCTGGCTCTTACACA TTTTATCTCCCACTGCACCC 208 0.87 351K21.A AAGGATAAGAAAAATGAGCATTGA ACTTCCCAATAAGCCAAGGG 168 0.76 134M13.A GCCTGGTGGCTGCATATACT ATGACATTCTGGGCCAGTTC 184 0.46 134M13.B CACCTGGTTGCTCTTCCCTA CTTGAGGCCAGGAATTTGAG 203 0.40 472A9.A CAGGAAAATTTCACACTTTTCACA TTTGGGTCTCTTTCTCTCCG 187 0.75 501M23.A GCTGCCAAAAACCCTAATGA TTCTTGCAGAATTGGCCTCT 194 0.90 501M23.B AGGCAGAAAATTTCACTAACAGA GCTCTTGGATTTTCCATCTCC 177 0.57 501M23.C TGGCTAGACATGTGCATGAAA TGCATTCCAGCCTGAGAG 172 0.78 551A13.C TGAATGGAGCAATAAGTGCAA TCCACAGTACCATCCTGGGT 182 0.71 551A13.A CTCTCCCCACAAAACCCTTC ACCCTTCCTCGGAAAAGAAA 141 0.76 1007B16.A CCCCAGAGAAACAGCCATTA TTTGGCATCTGAAGGAGGTC 161 0.50 1007B16.B TGCATGAGCATGAACCAAGT CCAGAAAGGAATTTGCATTG 203 0.31 411F9.A TCTCTTTTCTCCATATTCTTGCTAT TCAAGTGTAGACATACAAGTTTTCAA 197 0.65 411F9.B ATGAGAACAGGGCTTGAGGA ACATGTGTGTTTCTGGGTGC 206 0.82 424E5.A CCATTCACATGTTATTTGCTGAAA CACACCAGTGAACAACTAAGTAAACA 195 0.61 136A11.A TTCCCCTCAGATTTCCATTG ACTCCTGCCACCTGTCAGTC 204 0.73 75K24.A AACGTCTGAGTGAACCTCTTGA TCCATTCCATCTTGACTCCA 182 0.87 75K24.B TGTGGTTTAGGCCTTTAGATTGA CAGCACTGAACACTTCTGCTTT 195 0.83 159G19.A GTGGAAGTGCCTGTTTCCTC TAGGCTAAATGTCCGGTGCT 178 0.88 357D13.A TTGGGAGTCTCAGGTGGAAG GAAAGACATTTAAATTCCATCCAAA 196 0.66 357D13.B TGGTTTCATAGCACCAAAATG TGAAGAACCAAGTGAGCAAGAA 184 0.85 130P1.A TGAAGAACCAAGTGAGCAAGAA TGGTTTCATAGCACCAAAATG 194 0.75 260P22.A ACAGAGACCTACCCTGTGCC CACACCTGGTTGAGACCAAA 195 0.76

* The primers are named after the P1 artiﬁcial chromosome (PAC) within which they are located (e.g., 234P15) in the order they were studied (e.g., A, B, C). Searching the PAC sequence for the oligonucleotides used for ampliﬁcation can identify the location of the markers.

once in Hanks’ balanced salt solution and snap frozen in liquid nitro- Mutation Scanning gen. These RPE preparations may contain photoreceptor-derived mRNA. Exons were amplified with primers designed in flanking intronic sequences. Direct DNA sequencing of the amplified products was performed in one affected and one unaffected patient to scan for DNA Total RNA Isolation sequence variants. Total RNA was prepared from retina and RPE by mincing with mortar and pestle in liquid nitrogen. The minced powder was collected into Genomic and Bioinformatic Resources an RNA extraction reagent (RNA Stat-60; Tel-Test, Inc., Friendswood, TX). Total RNA was extracted as described by the manufacturer. Genomic DNA sequence data referred to in this article were produced by the Chromosome 6 Sequencing Group at the Sanger Centre and can be obtained at http://www.sanger.ac.uk. A combination of the RT-PCR and 5؅-RACE BLASTn, BLASTp, BLASTx, and BLAST 2 Sequences programs were Total RNA (1 ␮g) was used to make cDNA with a first-strand synthesis used to identify sequences of candidate genes within the critical system for RT-PCR (Superscript; Gibco-BRL, Rockville, MD), as de- region. Putative CpG islands (regions of DNA at 200 bp in length with scribed by the manufacturer. PCR was performed from cDNA using GϩC content and a ratio of CpGs of 0.6 or above) within genomic gene-specific primers as follows: 94°C for 5 minutes; 35 cycles of 94°C sequences were identified using CpG Island software (available in the for 30 seconds, 55°C for 30 seconds, 72°C for 30 seconds; and a final public domain from European Bioinformatics Institute [EBI], Cam- extension for 5 minutes at 72°C. Amplification products were analyzed bridgeshire, UK, at http://www.ebi.ac.uk/cpg/). Transmembrane helix by ethidium bromide-agarose gel electrophoresis. 5Ј-Rapid amplifica- location and topology predictions were made by computer (Predict- tion of cDNA ends (5Ј-RACE) was performed with total RNA (10 ␮g) Protein software; provided in the public domain by European Molec- from human retina using the a kit (FirstChoice RLM-RACE; Ambion, ular Biology Laboratory, Heidelberg, Germany, and available at http:// Austin, TX), as described by the manufacturer. www.embl-heidelberg.de/predictprotein.html) based on a previously described algorithm.14 Potential phosphorylation sites were identified 15 Retina cDNA Library based on a primary- and secondary-structure–based algorithm. A cDNA library from two pooled human adult retinas was constructed using a predigested vector kit (ZAP Express; Stratagene, La Jolla, CA), RESULTS as described by the manufacturer. Clones derived from the amplified library were identified using probes labeled with an enhanced chemi- Meiotic Breakpoint Mapping luminescence system (ECL Direct Nucleic Acid Labeling and Detection System; Amersham Pharmacia Biotech, Piscataway, NJ). Isolated clones We previously reported genealogical and molecular genetic were sequenced, and exon sequences were identified by a BLAST 2 studies demonstrating a common founder for all studied North Sequences search of the clone sequence against finished and unfin- American families with the autosomal dominant Stargardt-like ished genomic sequences. phenotype.2,5 Out of this family of 2314 individuals, 171 were

Downloaded from iovs.arvojournals.org on 09/24/2021 2656 Edwards et al. IOVS, October 2001, Vol. 42, No. 11

FIGURE 2. Genomic structure of critical region containing the autosomal dominant Stargardt-like macular dystrophy gene (ELOVL4). The larger critical region based on affected patients extends from D6S1622 to 260P22.A (Table 1). Using unaffected patients, a smaller critical region was deﬁned between 551A13.A and 260P22.A (Table 1). A previously reported unaffected patient enabled further reﬁnement to between D6S460 and 260P22.A.11,12 (f), Predicted CpG islands; (E), pseudogenes.

clinically examined and 145 samples genotyped. Figure 1 recombinant markers localizing the disease gene are presented shows the typical phenotype in a 52-year-old patient. in Table 1. The 1.8-cM larger critical region defined by affected We used both affected and unaffected individuals to refine patients was between centromeric marker D6S1622 at 88.63 the critical region for the STGD3 gene. Because the penetrance cM (patient S110) and telomeric marker D6S1707 at 90.43 cM of autosomal dominant Stargardt-like macular dystrophy is (patient B3015). Using unaffected patients, the centromeric nearly complete by the teenage years, observed in ophthalmo- border was further refined to D6S1625 at 89.23 cM.5 The scopic examination,2,6 unaffected patients with recombinant approximate locations of these boundaries were confirmed by markers were used in the refinement.2,5 Because of the possi- more than one patient with recombinant markers (Table 1). bility of a rare nonpenetrant individual, we used two critical Additional polymorphic STR markers were identified to regions: a larger critical region based only on affected patients further refine the critical region. Dimeric tandem repeats in the and a smaller critical region based on affected and unaffected Sanger Centre chromosome 6 sequence were located and patients. Our positional candidate gene strategy was to screen flanking oligonucleotides synthesized for amplification of these first within the smaller region. sites from 10 control DNA samples. The polymorphic subset The initial 20-centimorgan (cM) critical region6 was pro- of these dinucleotide markers is presented in Table 2. The gressively refined during the course of the study by using critical region was further refined to between 551A13.A and existing STR polymorphic markers.16,17 The individuals with 260P22.A, based on all available patients with recombinant

Downloaded from iovs.arvojournals.org on 09/24/2021 IOVS, October 2001, Vol. 42, No. 11 Gene for Dominant Stargardt-like Dystrophy 2657

TABLE 3. Genes Within the Critical Region Based on Affected with Recombinant Marker Patients

Identifying Information Gene (Accession Number of Expression and Mutation Sequence Variants Between Subjects and Name Numbers) Exons Scanning* Published Sequence in GenBank

COL12A1 NM_004370 65 Not expressed in retina; Not screened expressed in RPE; excluded by unaffected meiotic recombinant COX7A NM_001864 4 Not expressed in retina or Not screened RPE; excluded by unaffected meiotic recombinant dJ234P15.3 Unnamed gene within 7 Expressed in retina and RPE; Intronic SNP: C/T at positon 77601 in AL080250 PAC HSJ234P15 all exons sequenced; (AL080250) excluded by unaffected meiotic recombinant SSP1 AF196304 24 Expressed in retina and RPE; Not screened excluded by unaffected meiotic recombinant MYOVI U90236 34 Expressed in retina and RPE; Intronic SNPs: NM_004999—Splice 32 exons 1–7, 9, 11, 13–34 A/T at position 110771 of AL109897 variant sequenced; excluded by A/T at position 116611 of AL109897 unaffected meiotic C/T at position 11303 of AL136093 recombinant G/A at position 28771 of AL136093 C/T at position 61507 in AL136093 IMPG1 AF047492 17 Expressed in retina; all exons None observed sequenced; excluded by unaffected recombinant stSG30792 EST No information Expressed in retina; excluded Not screened by unaffected recombinant HTR1B NM_000712 1 Expressed in retina and RPE; 5Ј UTR SNP: G/T at position 3420 of AL049595; all exons sequenced; 3Ј UTR SNP: A/G at position 4861 of AL049595; excluded by unaffected 3Ј UTR SNP: C/T at position 5120 of AL049595 recombinant SG67324 EST No information Expressed in retina and RPE; Not screened excluded by unaffected recombinant11 Unnamed AK000712 13 Expressed in retina and RPE; Intronic SNP: C/T at position 30991 of AL356776; protein all exons sequenced; CDS SNP: T/C at position 1011 in AK000712 excluded by unaffected (exon 9; no change in predicted amino acid at recombinant11 codon 271) TRIP7 L40357 Ͼ6 Expressed in retina and RPE; None observed six known 3Ј exons screened; excluded by unaffected recombinant11 75K24.1 Unnamed gene within Ն3 Expressed in retina and RPE; None observed PAC HS75K24 all known exons sequenced; (AL035700) potential unknown centromeric exons; excluded by unaffected recombinant11 ELOVL4 AF277094 6 Expressed in retina and RPE; CDS 5-bp deletion in exon 6 all known exons sequenced; cDNA clone sequenced TTK M86699 22 Expressed in retina and RPE; Intronic SNPs: all exons sequenced C/T at position 79371 of AL133475 A/T at position 79380 of AL133475 G/A at position 79683 of AL133475 A/G at position 82950 of AL133475 G/A at position 108928 of AL133475 BCKDHE1 NM_000056 11 Expressed in retina and RPE; Intronic SNPs: all exons sequenced G/A at position 25398 of AL138732 G 3 A at position 3005 of AL049696 A 3 T at position 1424 of AL049696

SNP, Single nucleotide polymorphism; PAC, P1 artiﬁcial chromosome; CDS, coding sequence; UTR, untranslated region. * The human RPE preparations may contain small amounts of photoreceptor-derived or illegitimate transcripts that can be easily detected by RT-PCR.

markers (Table 1). Meiotic breakpoint data from unaffected allowed the exclusion of several additional genes, leaving patients previously described11,12 enabled the critical region to four genes within the critical region of 562 kb, as will be be further reﬁned to between D6S460 and 260P22.A. This described.

Downloaded from iovs.arvojournals.org on 09/24/2021 2658 Edwards et al. IOVS, October 2001, Vol. 42, No. 11

Evidence for a Disease-Causing Mutation in the ELOVL4 Gene The segregation of the 5-bp deletion within the extended family was examined by amplification of exon 6 to confirm that the sequence variant was located on the disease chromosome (data not shown). Four additional patient samples were sequenced to verify the presence of the 5-bp deletion in other affected individuals (data not shown). The 5-bp deletion would be predicted to cause a frameshift with a truncated protein. The truncation would delete 44 amino acids including the dilysine endoplasmic reticulum retention motif. This would be predicted to have a major effect on the function of the protein (Figs. 3, 4). To exclude this sequence variant as a polymorphism, exon FIGURE 3. DNA sequence from exon 6 of ELOVL4 in unaffected (left) and affected (right) family members showing a heterozygous affected 6 from 144 ethnically matched control individuals was ampli- patient. Registration of the two DNA sequences in the affected patient fied. All individuals exhibited the 334-bp normal band with no demonstrates deletion of the 5-bp segment AACTT. other alleles present, demonstrating the absence of this sequence variant on 288 control chromosomes. Given the expression of ELOVL4 in the retina, the pre- Genomic Structure of the Critical Region dicted severity of the 5-bp deletion on protein function, and The larger critical region based on affected patients spans the absence of this sequence variant on 288 control chro- approximately 6000 kb on the Sanger Centre chromosome 6 mosomes, this sequence variant appears to be the cause of database with four gaps, whereas the smaller critical region autosomal dominant Stargardt-like macular dystrophy. Be- based on unaffected patients with recombinant markers spans cause any sequence variant within the critical region would 562 kb, as noted. Within these regions, we sought to identify segregate with disease in the family, and no unrelated fam- genes based on the presence of homology to expressed se- ilies are known to exist, all other genes within the critical quence tags (ESTs), CpG islands, and a bioinformatic search for region were screened, and no coding sequence variations genes in genomic DNA sequence. Within the larger region we were identified (Table 3). identified evidence for 15 potential and/or known genes based on multiple exons and EST clusters. Additionally, nine pseudogenes based on the absence of exons, presence of a poly(A) Expression of ELOVL4 tail, and homology to a gene on another chromosome were Predicted and known genes within the disease locus for STGD3 identified (Fig. 2). Twelve CpG islands were identified, of were assayed for expression in RPE and neural retina isolated which two were not associated with a predicted or known directly from human eyes using RT-PCR. Two of the expressed gene (Fig. 2). sequences were grouped together in a cluster that was ulti- Screening of Candidate Genes mately identified as the STGD3 gene and both were expressed in both the neural retina and RPE preparations by RT-PCR (data Expression of the candidate genes in the retina and RPE was not shown). No expression was identified in the kidney, dem- determined by RT-PCR. Of the original 15 genes in the critical onstrating tissue-restricted expression of ELOVL4 according to region based on affected individuals, 14 were expressed in one RT-PCR (data not shown). No expressed pseudogenes have or both of these tissues (Table 3). A combination of bioinfor- been identified for ELOVL4; thus, these expression results matic gene analysis, cDNA cloning, and 5Ј-RACE was used to based on RT-PCR are likely to indicate the presence or absence determine the genomic structure of these genes. The exons of gene transcription. Because primary human RPE prepara- and their flanking sequence within each of these genes were tions may be contaminated with the overlying photoreceptors, scanned for sequence variation by DNA sequencing in one Northern blot analysis for mRNA expression was performed. affected and one unaffected patient (Table 3). All the exons within all identifiable genes in the smaller Northern blot analysis of human retina, RPE, and kidney re- critical region based on unaffected and affected patients vealed two transcripts in human retina only, demonstrating (D6S460 to 260P22.A) were sequenced. The only coding-se- that ELOVL4 is expressed only in the retina at detectable levels quence variant identified in these four genes was a 5-bp dele- (Fig. 5). A search for alternative splicing using overlapping tion (Fig. 3) within exon 6 of a cDNA cloned using ESTs primer sets along the message and RT-PCR, failed to identify SG24780 and SG53542 as hybridization probes. The DNA se- any evidence for size variation within the amplified products quence for the clone and predicted protein sequence for the (data not shown). Based on these data, ELOVL4 transcripts wild type and mutant are given in Figure 4. Primers for the appear to share identical coding sequences in the human amplification of the exons and their intronic borders for the retina. Thus, the size variation observed on the Northern blot ELOVL4 gene are given in Table 4. analysis most likely results from alternative polyadenylation.

‹

FIGURE 4. Sequence of a cDNA clone for ELOVL4 isolated from a retinal cDNA library using ESTs as probes (GenBank accession AY037298). The predicted protein sequence is illustrated along with the exon boundaries (arrows). The predicted effect of the 5-bp deletion (bold and underlined sequence) on the primary structure of the ELOVL4 protein is shown under the wild-type sequence. Four putative sp1-binding domains are underlined in the promoter region obtained from the Sanger Centre (the ﬁrst six rows of sequence). The translation start (ATG) and stop (TAA) codons are bold and underlined, as are putative polyadenylation consensus sequences. The differences in the start of transcription and three DNA sequence changes in the 3Ј untranslated region between this clone and previously described sequence data from RT-PCR products11 are shown in bold.

Downloaded from iovs.arvojournals.org on 09/24/2021 IOVS, October 2001, Vol. 42, No. 11 Gene for Dominant Stargardt-like Dystrophy 2659

Downloaded from iovs.arvojournals.org on 09/24/2021 2660 Edwards et al. IOVS, October 2001, Vol. 42, No. 11

TABLE 4. Primer Sets for Mutation Scanning of ELOVL4

Exon Forward Primer Reverse Primer Size

1a ctctgggtctccgctttct gaagtcagcggctttacctg (70)40(17) 1b gagccgggtagtgtcctaaa gagcgaggatggggaagt (0)82(30) 2a ttaggaaattaagttgaaacatcttg aggttctcggtccttcatcc (97)116(0) 2b tgtgtggctgggtccaaa tgatggttttacacattctcattt (0)112(75) 3 ttgtgtttttaatgctgtgtttacta ggggcctataaaaatcaaacc (61)81(57) 4a tttggtgtatataacacgctttcc ccaccacaaggtaaacatcg (34)141(0) 4b ttgagtatttggacacagtgttttt tgcaattaaacgcaagcagt (0)132(40) 5 catctcagtggcttactgcct ccagaagaaactttggaagca (66)128(31) 6 tgtcaacaacagttaaggccc tgttttgtgctgtcatttctgtt (27)278(29)

Exons 1, 2, and 4 are divided into two segments. The size column provides the base pairs of 5Ј flanking intronic sequence, coding sequence, and 3Ј flanking intronic sequence inclusive of the primers in the following format: (5Ј flanking)coding(3Ј flanking). There are 99 bp of overlap for exon 1, 40 bp for exon 2, and 101 bp for exon 4 between the two primer sets.

macular dystrophy in a large family distributed throughout North America. This novel gene is called elongation of very- long-chain fatty acids–like 4 (ELOVL4) because of its homology to the SUR4 family of yeast and mammalian proteins involved in the synthesis of long chain fatty acids. Without independent families to confirm that mutations within ELOVL4 cause disease, we screened the coding sequence and intron–exon borders of all known and identifiable genes within the critical region. The disease gene was ulti- mately localized within 562 kb of completely sequenced genomic DNA on chromosome 6q14. Within this region we were able to identify four genes, all of which were expressed FIGURE 5. Northern blot analysis of human retina, RPE, and kidney, in the retina or RPE. No DNA sequence variants were identified demonstrating tissue-restricted expression of ELOVL4. Total RNA (10 within the exons or their flanking sequences in our affected ␮g) was loaded onto each lane, and the cDNA clone described in patients, except for the 5-bp deletion in ELOVL4. Although Figure 4 was 32P labeled and used as a probe. bioinformatics and homology to expressed sequences led to the identification of several novel genes within this region ELOVL4 Protein and Elongation of the Very-Long- (Table 3), we cannot exclude the possibility that one or more Chain Fatty Acids–Like Gene Family genes may have gone undetected. Nonetheless, the extensive genomic and proteomic resources available for eucaryotic ge- The STGD3 disease gene is the fourth predicted human protein nomes and available bioinformatic software make it unlikely to be identified with homology to a yeast family of proteins that unidentified genes are located within this 562-kb region. (FEN1/Elo2 and SUR4/Elo3) involved in the elongation of The 5-bp deletion within exon 6 of ELOVL4 is predicted to very-long-chain fatty acids based on a bioinformatic search of cause a frameshift starting at amino acid 264 and truncated public genomic and proteomic databases. The human genes protein product at amino acid 271. The truncation would lead identified in a search of the genomic databases include to the loss of 44 amino acids. The C-terminal region of the ELOVL1, ELOVL2, ELOVL4, and HELO1. Mouse Elovl3 exists, normal 314-amino-acid protein is partially conserved within and presumably a human homologue will be identified. The members of the ELOVL protein family and contains a dilysine homology of these predicted proteins is striking with conser- motif required for retention in the endoplasmic reticulum (Fig. vation of the dideoxy binding site, the dilysine endoplasmic 6). Thus, the mutation is predicted to disrupt function of the reticulum retention motif, and potential tyrosine phosphoryla- protein. That the 5-bp deletion is not a rare polymorphism was tion sites (Fig. 6). Each of these proteins has been predicted to 18 demonstrated by its absence within 288 race–matched control have five transmembrane domains (Fig. 6). chromosomes. Genomic Organization of ELOVL4 The genomic localization of the STGD3 gene has been progressively refined by several investigators.2,5–7,9,11,12,20,21 Comparison of the sequence of our cDNA clone to the The reported localizations have all been in agreement, with the genomic sequence of chromosome 6 from the Sanger Centre, exception of unaffected individual III-4 who was reported to enabled the unambiguous localization of the intron–exon bor- exclude D6S284 and telomeric markers by Kniazeva et al.20 ders of ELOVL4 (data not shown). Each of the intron–exon The critical region reported herein is based on multiple af- borders conformed to the GT/AG consensus for spice donor– fected and unaffected patients, and the data on this subject acceptor sequences.19 reported by Kniazeva et al.20 most likely represent a genotyp- ing error, diagnostic misclassification, or a nonpenetrant indi- DISCUSSION vidual. This patient illustrates the difficulties in using unaffected subjects for disease localization. We describe a novel human gene with an ancestral 5-bp dele- No experimental data are available on the function of the tion in exon 6 giving rise to autosomal dominant Stargardt-like human ELOVL1, ELOVL2,orELOVL4 genes. The human ho-

‹

FIGURE 6. Multiple protein sequence alignment of the human and mouse predicted protein products of the elongation of the very-long-chain fatty acid–like gene family. Each of these gene products has homology to the SUR4 gene family involved in elongation of fatty acid biosynthesis in yeast. The amino acid sequence resulting from the 5-bp mutation is shown italicized.

Downloaded from iovs.arvojournals.org on 09/24/2021 Downloaded from iovs.arvojournals.org on 09/24/2021 2662 Edwards et al. IOVS, October 2001, Vol. 42, No. 11

mologue of mouse Elovl3 has not been identified to date. 3. Allikmets R, Singh N, Sun H. et al. A photoreceptor cell-specific Experimental data are available for Elovl3, which is also called ATP-binding transporter gene (ABCR) is mutated in recessive Star- Cig30, and a human gene called HELO1, with homology to gardt macular dystrophy (see comments). Nat Genet. 1997;15: ELOVL4 (Fig. 6). Cig30 was reported to complement the phe- 236–246. notype of the fen1/elo2 mutant that has reduced levels of fatty 4. Azarian SM, Travis GH. The photoreceptor rim protein is an ABC acids in the C to C range, demonstrating the involvement of transporter encoded by the gene for recessive Stargardt’s disease 20 24 (ABCR). FEBS Lett. 1997;409:247–252. the Elovl3 protein in the elongation of very-long-chain fatty 18 22 5. Donoso LA, Frost AT, Stone EM. et al. Autosomal dominant Star- acids. HELO1 was identified by Leonard et al., based on gardt-like macular dystrophy: founder effect and reassessment of homology to yeast fen1/elo2. The HELO1 protein was ex- genetic heterogeneity. Arch Ophthalmol. 2001;1:19:564–570. pressed in Saccharomyces cerevisiae 334 and was found to 6. Stone EM, Nichols BE, Kimura AE. et al. Clinical features of a result in increased elongation of monounsaturated and polyun- Stargardt-like dominant progressive macular dystrophy with ge- saturated fatty acids compared with controls.22 Based on the netic linkage to chromosome 6q (see comments). Arch Ophthal- homology to this family of proteins, we hypothesize that mol. 1994;112:765–772. ELOVL4 is part of the membrane-bound enzymatic complex 7. Lagali PS, MacDonald IM, Griesinger IB. et al. Autosomal dominant that executes the four enzymatic steps in the elongation of Stargardt-like macular dystrophy segregating in a large Canadian very-long-chain fatty acids in the neural retina.23,24 The mRNAs family. Can J Ophthalmol. 2000;35:315–324. for both of these proteins were expressed in a tissue-restricted 8. Cibis GW, Morey M, Harris DJ. Dominantly inherited macular pattern, as is ELOVL4 (Fig. 5). dystrophy with flecks (Stargardt). Arch Ophthalmol. 1980;98: Seven hereditary retinal diseases have been localized to 1785–1789. chromosome 6q14.25–31 Each of these diseases is within a 9. Zhang K, Bither PP, Park R. et al. A dominant Stargardt’s macular dystrophy locus maps to chromosome 13q34 (see comments). 30-cM region on chromosome 6q14 between coordinates Arch Ophthalmol. 1994;112:759–764. 80.34 and 111.17 cM. Based on our expression data (Table 3 10. Stone EM, Lotery AJ, Munier FL. et al. A single EFEMP1 mutation and unpublished) a large number of genes expressed in the associated with both Malattia Leventinese and Doyne honeycomb retina and/or RPE are potential candidate genes for these and retinal dystrophy. Nat Genet. 1999;22:199–202. other hereditary ocular disorders. The evolutionary bases for 11. Zhang K, Kniazeva M, Han M. et al. A 5-bp deletion in ELOVL4 is this chromosomal region with a high density of retinal disease associated with two related forms of autosomal dominant macular genes are unknown. dystrophy. Nat Genet. 2001;27:89–93. Identification of a mutation in ELOVL4 as the cause of 12. Griesinger IB, Sieving PA, Ayyagari R. Autosomal dominant macu- autosomal dominant Stargardt-like macular dystrophy impli- lar atrophy at 6q14 excludes CORD7 and MCDR1/PBCRA loci. cates fatty acid biosynthesis in the pathogenesis of macular Invest Ophthalmol Vis Sci. 2000;41:248–255. dystrophies and degenerations for the first time. The essential 13. Feldman EL, Del Monte MA, Stevens MJ, Greene DA. Establishment dietary ␣-linolenic fatty acid is converted into arachidonic acid and maintenance of in vitro cultures of human retinal pigment (20:4n-6) and docosahexaenoic acid (22:6n-3) by the fatty acid epithelium. In: Jones GE, ed. Methods in Molecular Medicine: chain elongation system.32 Docosahexaenoic acid accounts for Human Cell Culture Protocols. Totowa, NJ: Humana Press; 1996: 517–524. 33% of the fatty acid content of human photoreceptor outer 14. Rost B, Fariselli P, Casadio R. Topology prediction for helical segments and the lipid environment of the outer segment 33–38 transmembrane proteins at 86% accuracy. Protein Sci. 1996;5: influences photoreceptor function. These data are consis- 1704–1718. tent with an important role for the fatty acid chain elongation 15. Blom N, Gammeltoft S, Brunak S. Sequence and structure-based system in retinal function and disease. prediction of eukaryotic protein phosphorylation sites. J Mol Biol. Recently, the results of a prospective study examining the 1999;294:1351–1362. hypothesis that ARMD is influenced by dietary lipids showed a 16. Edwards A, Civitello A, Hammond HA, Caskey CT. DNA typing and positive association between total fat consumption and devel- genetic mapping with trimeric and tetrameric tandem repeats. opment of ARMD with visual loss.39 This study and another Am J Hum Genet. 1991;49:746–756. have reported a negative association between consumption of 17. Broman KW, Murray JC, Sheffield VC, White RL, Weber JL. Com- fish and ARMD.39,40 Although these results lead to the hope prehensive human genetic maps: individual and sex-specific vari- that the incidence or severity of macular degeneration may be ation in recombination. Am J Hum Genet. 1998;63:861–869. decreased by dietary modulation, the observation that the 18. Tvrdik P, Westerberg R, Silve S. et al. Role of a new mammalian dietary precursor to docosahexaenoic acid (linolenic acid) ap- gene family in the biosynthesis of very long chain fatty acids and sphingolipids. J Cell Biol. 2000;149:707–718. pears to explain the majority of the risk for macular degener- 39 19. Mount SM. A catalogue of splice junction sequences. Nucleic Acids ation demonstrates the need for a better understanding of Res. 1982;10:459–472. retinal fatty acid metabolism before making clinical recommen- 20. Kniazeva MF, Chiang MF, Cutting GR. et al. Clinical and genetic dations to patients. studies of an autosomal dominant cone-rod dystrophy with features of Stargardt disease (see comments). Ophthalmic Genet. 1999;20:71–81. Acknowledgments 21. Zhang K, Kniazeva M, Hutchinson A. et al. The ABCR gene in recessive and dominant Stargardt diseases: a genetic pathway in The authors thank the patients and members of their extended families macular degeneration. Genomics. 1999;60:234–237. for continued support and participation in this research. 22. Leonard AE, Bobik EG, Dorado J, et al. Cloning of a human cDNA encoding a novel enzyme involved in the elongation of long-chain References polyunsaturated fatty acids. Part 3. Biochem J. 2000;350:765–770. 23. Cinti DL, Cook L, Nagi MN, Suneja SK. The fatty acid chain 1. Krill AE. Flecked retina diseases. In: Krill AE, ed. Krill’s Hereditary elongation system of mammalian endoplasmic reticulum. Prog Retinal and Choroidal Diseases. Hagerstown, MD: Harper Row, Lipid Res. 1992;31:1–51. 1977:750–762. 24. Kohlwein SD, Eder S, Oh CS. et al. Tsc13p is required for fatty acid 2. Edwards AO, Miedziak A, Vrabec T. et al. Autosomal dominant elongation and localizes to a novel structure at the nuclear-vacuo- Stargardt-like macular dystrophy. I: clinical characterization, longi- lar interface in Saccharomyces cerevisiae. Mol Cell Biol. 2001;21: tudinal follow-up, and evidence for a common ancestry in families 109–125. linked to chromosome 6q14. Am J Ophthalmol. 1999;127:426– 25. Ruiz A, Borrego S, Marcos I, Antinolo G. A major locus for auto- 435. somal recessive retinitis pigmentosa on 6q, determined by ho-

Downloaded from iovs.arvojournals.org on 09/24/2021 IOVS, October 2001, Vol. 42, No. 11 Gene for Dominant Stargardt-like Dystrophy 2663

mozygosity mapping of chromosomal regions that contain gamma- 36. Mizota A, Sato E, Taniai M, Adachi-Usami E, Nishikawa M. Protec- aminobutyric acid-receptor clusters. Am J Hum Genet. 1998;62: tive effects of dietary docosahexaenoic acid against kainate-in- 1452–1459. duced retinal degeneration in rats. Invest Ophthalmol Vis Sci. 26. Khaliq S, Hameed A, Ismail M. et al. Refinement of the locus for 2001;42:216–221. autosomal recessive retinitis pigmentosa (RP25) linked to chromo- 37. de Turco EB, Jackson FR, Parkins N, Gordon WC. Strong associa- some 6q in a family of Pakistani origin (letter). Am J Hum Genet. tion of unesterified [3H]docosahexaenoic acid and [3H: 1999;65:571–574. docosahexaenoyl]phosphatidate to rhodopsin during in vivo label- 27. Dharmaraj S, Li Y, Robitaille JM. et al. A novel locus for Leber ing of frog retinal rod outer segments. Neurochem Res. 2000;25: congenital amaurosis maps to chromosome 6q. Am J Hum Genet. 695–703. 2000;66:19–26. 38. Weisinger HS, Vingrys AJ, Bui BV, Sinclair AJ. Effects of dietary n-3 fatty acid deficiency and repletion in the guinea pig retina. Invest 28. Kelsell RE, Gregory-Evans K, Gregory-Evans CY, et al. Localization Ophthalmol Vis Sci. 1999;40:327–338. of a gene (CORD7) for a dominant cone-rod dystrophy to chromo- 39. Cho E, Hung S, Willett WC, et al. Prospective study of dietary fat some 6q (letter). Am J Hum Genet. 1998;63:274–279. and the risk of age-related macular degeneration. Am J Clin Nutr. 29. Kniazeva M, Traboulsi EI, Yu Z, et al. A new locus for dominant 2001;73:209–218. drusen and macular degeneration maps to chromosome 6q14. 40. Smith W, Mitchell P, Leeder SR. Dietary fat and fish intake and Am J Ophthalmol. 2000;130:197–202. age-related maculopathy. Arch Ophthalmol. 2000;118:401–404. 30. Small KW, Udar N, Yelchits S, et al. North Carolina macular dystrophy (MCDR1) locus: a fine resolution genetic map and PPENDIX haplotype analysis. Mol Vis. 1999;5:38. A A 31. Kelsell RE, Godley BF, Evans K, et al. Localization of the gene for Database Accession Numbers progressive bifocal chorioretinal atrophy (PBCRA) to chromosome 6q. Hum Mol Genet. 1995;4:1653–1656. ELOVL4 (AY037298), 234P15.C (AF391300), 234P15.A (AF391303), 234P15.B (AF391299), 474L11.A (AF391304), 32. Fliesler SJ, Anderson RE. Chemistry and metabolism of lipids in the vertebrate retina. Prog Lipid Res. 1983;22:79–131. 474L11.B (AF391310), 351K21.A (AF391311), 134M13.A 33. Rapp LM, Maple SS, Choi JH. Lutein and zeaxanthin concentrations (AF391312), 134M13.B (AF391313), 472A9.A (AF391314), in rod outer segment membranes from perifoveal and peripheral 501M23.A (AF391301), 501M23.B (AF391302), 501M23.C human retina. Invest Ophthalmol Vis Sci. 2000;41:1200–1209. (AF391298), 551A13.C (AF391315), 551A13.A (AF391317), 34. Suh M, Wierzbicki AA, Lien EL, Clandinin MT. Dietary 20:4n-6 and 1007B16.A (AF391316), 1007B16.B (AF391318), 411F9.A 22:6n-3 modulates the profile of long- and very-long-chain fatty (AF391319), 411F9.B (AF391320), 424E5.A (AF391321), acids, rhodopsin content, and kinetics in developing photorecep- 136A11.A (AF391322), 75K24.A (AF391309), 75K24.B tor cells. Pediatr Res. 2000;48:524–530. (AF391305), 159G19.A (AF391306), 357D13.A (AF391307), 35. Organisciak DT, Darrow RM, Jiang YL, Blanks JC. Retinal light 357D13.B (AF391308), 130P1.A (AF391323), 260P22.A damage in rats with altered levels of rod outer segment docosa- (AF391324). GenBank is provided by NCBI, Bethesda, MD, and hexaenoate. Invest Ophthalmol Vis Sci. 1996;37:2243–2257. is available at http://www.ncbi.nlm.nih.gov/

Downloaded from iovs.arvojournals.org on 09/24/2021