Mutated Alleles of the Rod and Cone Na-Ca+K-Exchanger Genes in Patients with Retinal Diseases

Abstract

purpose. To study the possible involvement of the rod (SLC24A1) and cone (SLC24A2) Na-Ca+K exchanger (NCKX) genes in retinal diseases.

methods. DNA was collected from unrelated patients with retinal disease, mainly from North America. A human genomic library was screened with the cone NCKX cDNA, and hybridizing clones were sequenced to determine the genomic organization of the SLC24A2 gene. The single-strand conformation polymorphism (SSCP) technique and direct sequencing were used to screen the patients’ DNA for mutations in SLC24A1 and SLC24A2. The effect of selected missense changes on protein function was tested by measuring potassium-dependent Na-Ca exchange of the mutant proteins expressed in insect cells.

results. Twenty-seven novel sequence changes were found in the rod NCKX gene, 21 of which are unlikely to be pathogenic, because they did not cosegregate with the disease or did not affect conserved regions of the protein. Of the remaining six, two were frameshift mutations found in one patient each. If translated, these alleles would encode nonfunctional proteins. Three of the six possibly pathogenic mutations were missense changes located in conserved regions, and their protein functions were assayed. Only one (Ile992Thr) had a significantly low level of exchanger function, but it was found in two unrelated patients who were heterozygotes with different retinal diseases, and this mutation could not be unequivocally associated with either disease. The last of the six changes is likely to create a new splice acceptor site. The genomic organization of the cone NCKX gene was determined, and it contained 11 exons with a few splice variants. Fifteen novel sequence changes were identified in the cone exchanger gene in patients with a cone dysfunction or degeneration. Only three of these sequence changes, all missense changes found in heterozygous patients, were considered possibly pathogenic. Functional analysis showed only a slight reduction in the activity of the corresponding mutant proteins.

conclusions. Although variant alleles of the rod and cone NCKX genes were found, none could be definitively associated with a specific retinal disease. The human phenotype associated with mutant exchanger alleles remains unknown.

Changes in calcium and sodium concentrations within the photoreceptor outer segments are crucial for phototransduction. In the dark, calcium and sodium ions enter the rod and cone outer segments through open cyclic guanosine monophosphate (cGMP)–gated cation channels and are extruded by the rod and cone Na-Ca+K exchangers (NCKXs). The extrusion is coupled with an inward sodium gradient and an outward potassium gradient. In response to light, cGMP-gated cation channels close, resulting in a reduction in the calcium concentration within the photoreceptor cell because of continued function of the exchanger. This leads to hyperpolarization of the plasma membrane and modulation of neurotransmitter release. 

Rods and cones express similar exchanger proteins encoded by distinct genes: SLC24A1 and SLC24A2, respectively. mRNA transcripts of SLC24A1 have been found to date only in retinal rod photoreceptors and in a megakaryotic cell line. ¹ The gene is within chromosome 15q22. ² The cone exchanger gene SLC24A2 has been found to be abundantly expressed in cone photoreceptors and retinal ganglion cells and is expressed at a lower level in the brain. ³ SLC24A2 has been assigned to 9p22. There is no reported evidence of unidentified retinal disease genes in the chromosomal regions containing SLC24A1 or SLC24A2. 

Although the overall length of the rod and cone NCKX proteins is quite different, hydropathy analysis and sequence alignment indicate two sets of putative transmembrane-spanning segments with a high degree of conservation. The difference in length is accounted for by the difference of the two large hydrophilic loops, ⁴ one located extracellularly at the N terminus and one located in the cytosol between the two sets of putative transmembrane-spanning segments. The transmembrane domains are highly conserved between human and bovine rod exchangers (>94% amino acid identity), whereas the two large loops are poorly conserved in these two species. ⁴ An interaction between the rod exchanger and the rod α subunit of the cGMP–gated channel has been reported. ⁵ A cation transport assay has been developed recently by using a dolphin rod or a bovine rod exchanger with shortened extracellular and cytosolic loops. ⁶  

Mutations in many other genes encoding proteins that take part in phototransduction are known to be the cause of stationary or progressive retinal diseases. Examples are the genes encoding rhodopsin, the α subunit of rod transducin, the α and β subunits of cGMP phosphodiesterase, the rod α-cGMP–gated channel, guanylyl cyclase, rhodopsin kinase, arrestin, and the α and β subunits of the cone cGMP-gated channel. Mutations in these genes can cause forms of retinal degeneration, such as retinitis pigmentosa (RP), or forms of retinal malfunction, such as stationary night blindness. It is reasonable to propose that mutations in the rod and cone exchanger genes might also be the cause of some hereditary retinal diseases. To evaluate this possibility, we conducted a broad screen of patients with various forms of RP or some other retinal disease for mutations in these genes, and we studied the function of selected mutant exchanger proteins. 

Methods

Ascertainment of Patients

This study, which involved human patients, conformed to the tenets of the Declaration of Helsinki. All index cases had retinal disease diagnosed through ophthalmologic examination, including electroretinography (ERG). Most patients resided in the United States or Canada. Ninety-three unrelated individuals without symptoms or a family history of retinal disease were used as normal control subjects. Informed consent was obtained from all participants before they donated 10 to 50 mL of venous blood for this research. Leukocyte nuclei were prepared from the blood samples and stored at −70°C before DNA was purified from them. 

Genomic Library Screen for Human SLC24A2

Approximately 6.6 × 10⁵ clones from a human placental genomic library in phage vector lambda FIX II (9–22 kb insert length; Stratagene, La Jolla, CA) were screened separately with the following two probes: (1) a partial cDNA sequence 1383 nucleotides (nt) long and corresponding to codons 141 to 601 (accession number NM_020344), which was obtained by RT-PCR amplification of retinal mRNA with the primers GAATGATTTACATGTTCATAGCCTTAGC and AGTGGCTGTCAGCAGCAATGGC; (2) a cDNA sequence, including 15 bp at the 3′ end of exon 4, exon 5, and 15 bp at the 5′ end of exon 6, which was obtained by RT-PCR amplification of brain mRNA with the primers TATTGGTACCTCTGAACCTCCCTTCTTCAGTCAT and CCAGGATCCACTCGCCGAAGAACT (accession number NT_008335). The hybridization was performed under medium-stringency conditions (100 mM sodium phosphate [pH 7.2], 0.1% SDS, 65°C). Eleven phage colonies that hybridized with probe 1 and one that hybridized with probe 2 were chosen, purified, and amplified. The phage DNA was isolated by the plate lysate method and subsequent polyethylene glycol-NaCl precipitation, phenol-chloroform extraction, and ethanol precipitation. All oligonucleotides were synthesized on a commercially available DNA synthesizer (Oligo 1000; Beckman, Carlsbad, CA). To establish the 5′ and 3′ untranslated regions (UTRs), we used 5′ and 3′ rapid amplification of cDNA ends (RACE) and PCR techniques. First-strand cDNA was made from total RNA from the two tissues that express this gene: retina and brain. RACE products were directly sequenced or sequenced after cloning into plasmids. Cone NCKX was also amplified from adaptor-ligated full-length cDNA of human whole brain (Marathon-Ready cDNA; Clontech Laboratories; Palo Alto, CA). 

Screening for Mutations

The single-strand conformation polymorphism (SSCP) technique was used to screen the open-reading frames of the rod and cone exchangers, as well as the immediately flanking intron sequences, for point mutations and other small-scale sequence changes. Each exon was individually amplified from leukocyte DNA samples by PCR with primer pairs (primer sequences available at http://eyegene.meei.harvard.edu/ provided in the public domain by the Ocular Molecular Genetics Institute, Massachusetts Eye and Ear Infirmary). PCRs were performed in the wells of 96-well microtiter plates. In each well was 20 ng leukocyte DNA in 20 μL of a buffer containing 20 mM tris-HCl (pH 8.4 or 8.6), 0.25 to 1.5 mM MgCl₂, 50 mM KCl, 0.02 mM dATP, 0.02 mM dTTP, 0.02 mM dGTP, 0.002 mM dCTP, 0.6 μCi [α-³²P or α-³³P]dCTP (3000 Ci/mmol), 0.1 mg/mL bovine serum albumin (BSA), 0% or 10% dimethyl sulfoxide, and 0.25 U Taq polymerase. The pH, Mg⁺⁺ concentration, annealing temperature, and presence or absence of 10% dimethyl sulfoxide were tailored to each primer pair to yield optimal amplification. After an initial denaturation (94°C for 5 minutes), 25 cycles of PCR amplification were performed. Each cycle consisted of denaturation (94°C for 30 seconds), primer annealing (50–58°C for 30 seconds), and extension (71°C for 30 seconds). The final extension was at 71°C for 5 minutes. Some SLC24A1 amplicons were digested with a restriction enzyme (exon 2A, SmaI; exons 2B and 2C, AluI; exons 2D and 2E, DpnII; exon 4, AvaI; exon 7A, BsmAI; and exon 10, EarI) by adding 5 U of enzyme and incubating for 3 hours at the recommended temperature. The amplified DNA fragments were heat denatured, and aliquots of the single-stranded fragments were separated through polyacrylamide gels. At least two different polyacrylamide gels (of the three described herein) were used for each exon: 6% polyacrylamide in Tris-borate-EDTA (TBE) buffer, 6% polyacrylamide with 10% glycerol in TBE buffer, and 6% polyacrylamide in Tris-2-(N-morpholino)ethane sulfonic acid (MES)-EDTA (TME) buffer (30 mM Tris, 35 mM MES, 1 mM Na₂EDTA; pH 6.8). Gels were run at 6 to 18 W for 5 to 18 hours at room temperature before drying and autoradiography. Variant bands were analyzed by sequencing the corresponding PCR-amplified DNA segments using dye terminators (Dye terminator cycle sequencing kit; PE Biosystems, Foster City, CA) on an automated sequencer (ABI 377; PE Biosystems) or with a radiolabeled cycle sequencing kit (Thermosequenase; Amersham, Arlington Heights, IL). Participating relatives of index patients with selected sequence anomalies were evaluated for the same sequence changes by SSCP or sequence analysis of DNA. Methods for PCR amplification and sequencing of the rod α-cGMP–gated channel gene have been described previously. ⁷  

Construction of Mutant Cone NCKX cDNAs

The short splice variant ³ of the human retinal cone NCKX cDNA (AF177987) was used in all our mutant constructs. The human c-myc tag (EQKLISEEDL) was inserted at the BstEII site between bases 241 and 242 (at amino acid 81) of the human cone NCKX sequence. Mutations in the myc-tagged human cone NCXK cDNA were prepared by generating mutated cassettes and reintroducing the cassette into the wild-type sequence. Mutations were prepared by synthesizing oligonucleotide primers containing the desired mutation and generating two overlapping fragments (upstream and downstream) by PCR. The two fragments were combined, amplified by PCR again, and digested with appropriate restriction endonucleases before cloning the cassette back into the wild-type sequence. The mutants I917V, I992T, and A994T were cloned as BsrGI fragments, F282C and V284I as XbaI-NdeI fragments, and N601S as a BsrGI-BamHI fragment. All PCR reactions were performed using commercial DNA polymerase (Platinum Pfx; Life Technologies, Burlington, Ontario, Canada) according to the manufacturer’s directions, and all fragments generated by PCR were thoroughly sequenced after cloning to ensure that no unwanted mutations were generated by PCR errors. Plasmid DNA to be transfected into insect cells was prepared using a kit (EndoFree Maxi Prep; Qiagen, Mississauga, Ontario, Canada). 

Transient Expression of Human Mutant and myc-Tagged Cone NCKX cDNAs in Insect Cells

A novel lepidopteran insect cell expression system was used for transient transfection of BTI-TN-5B1-4 insect cells (High Five; Invitrogen, San Diego, CA) with human cone NCKX cDNA. ⁸ High Five cells were subcultured at 28°C in IPL-41 insect medium (Life Technologies) supplemented with 0.2 g/L l-glutamine, 1.0 g/L d-glucose, 10% heat-inactivated fetal bovine serum (GibcoBRL, Grand Island, NY) and penicillin-streptomycin-amphotericin B (GibcoBRL). Transient transfection of cells was performed with lipofectin, as described elsewhere. ⁸ Epitope tags used were human c-myc (Roche Molecular Diagnostics, Laval, Quebec, Canada) or FLAG M2 (Eastman Kodak, New Haven, CT). 

Measurement of ⁴⁵Ca²⁺ Uptake

Our functional assay for NCKX activity took advantage of the fact that NCKX proteins are bidirectional and mediate both calcium efflux (forward exchange) and calcium influx (reverse exchange). Reverse Na-Ca+K exchange was measured as potassium-dependent ⁴⁵Ca²⁺ uptake in sodium-loaded High Five cells, as described in more detail elsewhere. ⁶ Control cells were either untransfected High Five cells or High Five cells transfected with empty vector, both of which resulted in very similar background ⁴⁵Ca²⁺ uptake. ³ Potassium-dependent ⁴⁵Ca²⁺ uptake in sodium-loaded cells represented reverse Na-Ca+K exchange. Sodium loading using the ionophore monensin and subsequent removal of monensin by washing cells with monensin-free solutions containing BSA were performed as described before. ⁶ The final cell pellet was resuspended in 150 mM choline chloride, 80 mM sucrose, 20 mM HEPES (pH 7.4), and 0.05 mM EDTA and stored at 25°C. ⁴⁵Ca²⁺ (0.5–1.0 μCi per experiment; Amersham Pharmacia Biotech, Piscataway, NJ) uptake experiments were performed in a medium containing 140 mM KCl, 80 mM sucrose, 20 mM HEPES (pH 7.4), and 0.05 mM CaCl₂ in addition to ⁴⁵Ca²⁺. This medium optimized ⁴⁵Ca²⁺ uptake through reverse Na-Ca+K exchange. Background ⁴⁵Ca²⁺ uptake was measured when KCl in the uptake medium was replaced with NaCl, under which condition ⁴⁵Ca²⁺ uptake through reverse Na-Ca+K exchange is completely inhibited. ⁶ External ⁴⁵Ca²⁺ was removed from ⁴⁵Ca²⁺ taken up by cells by a rapid filtration and washing procedure with the use of borosilicate glass fiber filters as described previously. The washing medium contained 140 mM KCl, 80 mM sucrose, 20 mM HEPES (pH 7.4), 5 mM MgCl₂, and 1 mM EGTA. NaCl, KCl, LiCl, and choline chloride were all SigmaUltra grade (Sigma-Aldrich, Oakville, Ontario, Canada). Protein content of cell samples was determined with a protein assay (Bio-Rad Laboratories, Mississauga, Ontario, Canada). Sample preparation, SDS gel electrophoresis, and Western blot analyses with the human c-myc antibody were performed as described. ⁶  

Results

Sequence Changes in the Rod Exchanger

The genomic organization of the rod NCKX gene has been published. ² The gene contains 10 coding exons. The coding exons and the intron–exon boundaries were PCR-amplified and screened for mutations in DNA samples from 815 patients from separate families with a retinal disease (Table 1) . Twenty-seven novel sequence changes were identified (Table 2) , and 20 of these would presumably affect the rod exchanger amino acid sequence (Fig. 1) . Fifteen of the 20 changes were missense, two were small frameshift deletions, one was an inframe deletion, one an inframe insertion, and one an isocoding change that possibly created a novel splice site. We obtained evidence that five of the 20 sequence changes were unlikely to be pathogenic: Two (Thr37Ser and Leu313Val) were polymorphisms (>1% minor allele frequency), and three (2626-2652del, Glu776Gln and Phe1054Val) did not cosegregate with the disease in the respective patients’ families (Table 2 , Fig. 2 ). Eight of the remaining 15 changes (Arg239Thr, Val311Leu, Ser331Cys, Met389Lys, Val459Ile, Val608Ile, 2658-2659insGAG, and Val962Ala) affected nonconserved parts of the protein or were conservative amino acid changes (Fig. 3) and thus were not likely to be pathogenic. One missense change, Trp922Arg, was heterozygous in three patients with dominant RP: three with recessive RP and one with recessive congenital retinal blindness (Leber congenital amaurosis). We considered it unlikely that this change causes dominant RP, because it was also found in patients with recessive RP. We also considered it unlikely to cause recessive RP or Leber congenital amaurosis, because in none of these cases was a second hit found. Six sequence changes found in seven patients remained as possibly pathogenic mutations: Two were frameshift mutations (754-755delAT and 2186delC), three were missense changes (Ile917Val, Ile992Thr, and Ala994Thr), and one was an isocoding change (c87G→A in codon 29) that was likely to affect intron splicing (description follows). 

The two frameshift mutations were found in one patient each, both of which were heterozygotes. Patient 243-002 carried the mutation 754-755delAT (Met252delAT) and had age-related macular degeneration diagnosed at age 71. Patient 088-005 carried the 2186delC (Pro729delC) mutation (Fig. 4) and had optic atrophy but normal photoreceptor function, as determined by ERG at age 20. We could not find a mutation on the second allele in both patients. If expressed, these mutant alleles would encode truncated, nonfunctional proteins in which at least half of the transmembrane domains would be absent. Because these frameshift mutations are probably null, both patients would be expected to have the same, or a similar, retinal disease. However, the patients have two different diseases, neither of which are directly attributable to rod photoreceptor dysfunction. Thus, these two frameshift mutations are unlikely to be the cause of disease in these two patients. 

Three missense changes (Ile917Val, Ile992Thr, and Ala994Thr) were selected for additional evaluation (i.e., analysis of the corresponding mutant proteins’ function; described later) because they affected residues that were highly conserved in evolution, were too infrequent to be polymorphisms, and cosegregated with disease in the available relatives of the index patients who carried them. All three mutations affected residues in transmembrane domains (H6, H8, and H8; Fig. 3 ). Patient 063-008 with congenital stationary night blindness (CSNB) was heterozygous for the Ile917Val missense change. The affected patient and his affected brother both had the Ile917Val allele derived from their unaffected father. In addition, both had inherited the same allele, determined by the polymorphism Thr37Ser, from their unaffected mother (Fig. 2) . The apparent recessive pattern of inheritance suggested that mutations of both alleles should be present, but no mutation of the maternally derived allele was found. Two patients were heterozygous for Ile992Thr: 274-011 with CSNB (Fig. 4) and 121-067, with isolate RP. Three family members of 274-011 were available for segregation analysis, which revealed that Ile992Thr was not a dominant mutation causing night blindness (Fig. 2) . No family members of patient 121-067 were available for segregation analysis. Patient 041-001 with unilateral RP was heterozygous for the Ala994Thr change, but no family members were available for segregation analysis. 

Patient 038-051 was heterozygous for the isocoding change c87G→A in codon 29. The patient had pericentral RP, ^{9 10} with reduced rod-plus-cone full-field ERG amplitudes in response to 0.5-Hz flashes of light (average between the two eyes, 185 μV; normal, ≥350 μV) and reduced 30-Hz ERG amplitudes (29 μV; normal, ≥50 μV) at age 51. A computer analysis (http://www.fruitfly.org/seq_tools/splice.html) of the sequence revealed a potential splice site created by this change (wild-type sequence had a less than 0.1 probability of being a splice acceptor site, whereas the mutant sequence had a 0.95 probability). If the novel acceptor splice site is used in vivo, a protein without the first 29 amino acids may be produced. The N terminus of the rod exchanger might act as a noncleaved signal peptide, ¹¹ and thus the absence of these 29 residues may produce a nonfunctional protein. We did not find any other likely pathogenic mutations in 20 other unrelated patients with pericentral RP, and the patient had no relatives available for segregation analysis. 

All seven patients with the frameshift mutations (754-755delAT and 2186delC), the missense mutations (Ile917Val, Ile992Thr, and Ala994Thr), and the likely splice-site change (c87G→A) were heterozygotes. We performed an intensive screen for mutations of the homologous allele in each of these seven patients. This included sequencing all SLC24A1 exons (including 5′ and 3′ UTRs), and the 5′ end of the gene, including the presumed promoter region. No pathogenic mutations were identified. The α subunit of the rod cGMP-gated channel was shown by others to interact with the rod exchanger. ⁵ Thus, we hypothesized that a possible digenic form of inheritance with mutations in both the rod exchanger gene and the α-cGMP–gated channel gene may be the genetic mechanism causing disease in some of the seven families. We sequenced the coding exons of the rod α-cGMP–gated channel in DNA from these seven index patients but could not identify any mutations that were likely to be pathogenic. 

Genomic Organization of SLC24A2

The genomic organization of the cone exchanger (SLC24A2) gene was determined by sequence analysis of the genomic fragments derived from phages that hybridized with a human cone NCKX cDNA probe. The gene consists of 11 exons (Table 3) . The genomic organization of SLC24A2 resembles that of the human rod exchanger. ² However, SLC24A2 has a splice-variant that contains one additional exon (exon 5; 51 nt) which would encode residues in the third cytoplasmic loop. The predominant isoform in the retina does not contain exon 5. The third cytoplasmic loop is not well conserved when compared with mammalian rod and cone exchangers. Because SLC24A2 product is also expressed in brain, ^{3 12} we tested hippocampal total RNA for hybridization with the human cone NCKX cDNA. A Northern blot analysis revealed the expected 10.5-kb band with a weaker hybridization to hippocampal RNA than retinal RNA. A screen of a human genomic library with an exon 5 sequence as a probe resulted in the isolation of one hybridizing phage. This exon is flanked by introns that are more than 8 kb in size each. 

Using RACE and PCR we found three different versions of exon 1 in the cone NCKX mRNA transcripts that were isolated from human retina. The predominant transcript found in the retina contained an exon 1 of 154 nt (exon 1A: 7 of 10 cDNA clones). The other two types of 5′ UTR were exon 1B of 284 nt length (1 of 10 clones) and exon 1C of 158 nt length (2 of 10 clones). Exon 1C is located directly upstream of exon 2 and thus is part of it. The results of the 5′ RACE of brain mRNA showed that exons 1A and 1B were present, although at a frequency different from that found in the retina. There are several consensus binding sites for transcription factors in the regions upstream of the three different versions of exon 1 (a CCAAT box upstream of exon 1A, an SP1 binding site upstream of exon 1B, and CCAAT and TATA boxes upstream of exon 1C). The upstream regions of exons 1A and 1B have a high GC content. 

To study the 3′ UTR of human cone NCKX mRNA, we used Northern blot analysis, which revealed an mRNA (10.5 kb) in the retina and the brain much longer than the open reading frame size (2 kb). Thus, we amplified the 3′ UTR of cone NCKX from brain RNA and compared it with the genomic sequence. The comparison revealed a 3′ UTR in exon 11 of approximately 8.7 kb. Only the first 250 nt of exon 11 are coding sequence. Polyadenylation started 15 nt downstream of an AAUAAA polyadenylation signal sequence. 

Sequence Changes in the Cone Exchanger

One hundred sixty-six patients with retinal diseases affecting cone photoreceptors were screened for mutations in the SLC24A2 gene encoding the cone Na-Ca+K exchanger (Table 1) . Fourteen sequence changes were encountered (Table 4) . Only three of these were considered to be possibly pathogenic changes because they were rare and they affected conserved residues in the protein sequence. All three changes were missense and a different patient was heterozygous for each change: Patient 162-006 with macular malfunction carried Phe282Cys, patient 065-010 with incomplete achromatopsia carried Val284Ile, and patient 286-002 with complete achromatopsia carried Asn601Ser. The residues affected by the Phe282Cys, Val284Ile, and Asn601Ser mutations are in transmembrane domains H5, H5, and H10, respectively (Fig. 1) . We sequenced the coding exons of the cone exchanger gene to screen more intensively for a second sequence change in these three patients, but no likely pathogenic mutations were identified. None of these three missense changes were found in 93 normal control subjects. 

Protein Function of Missense Changes

The function of three rod exchanger mutants (Ile917Val, Ile992Thr, and Ala994Thr) and three cone exchanger mutants (Phe282Cys, Val284Ile, and Asn601Ser) was studied in a heterologous cell system. Each of the six mutations affected residues in transmembrane domains of the encoded proteins (Fig. 1) . 

A previous attempt to study the function of a full-length human rod NCKX cDNA ⁴ did not produce measurable NCKX function, and thus we were not able to examine the functional consequences of missense mutations in human rod NCKX directly. In contrast, human cone NCKX cDNA produced strong potassium-dependent Na-Ca exchange when expressed in High Five cells. ³ Therefore, we used the human cone NCKX construct to analyze the consequences of the missense changes in both the rod and the cone NCKX mutants. All these mutations affected residues in regions conserved between human rod and cone NCKX. For each of the missense mutations examined, protein expression levels were indistinguishable from that observed with the myc-tagged wild-type cone NCKX (Fig. 5A) . The functional consequences of the different NCKX missense mutations on reverse Na-Ca+K exchange are summarized in Figure 5B . Of the mutations examined, only Ile992Thr (a rod exchanger mutant) led to a substantial reduction of reverse Na-Ca+K exchange activity (<20% of wild type). The other five mutations produced proteins with moderately less activity or roughly equal activity compared with wild type. 

Discussion

We found four mutations that are likely to affect the function of the encoded rod exchanger protein. These were found in five patients, all of whom were heterozygotes. Two mutations were frameshifts and were found in patients with age-related macular degeneration (754-755delAT) and optic atrophy (2186delC). One was a missense change (Ile992Thr) that produced a poorly functional protein and was found in two patients: one with CSNB and the other with isolate RP. Finally, one change was likely to create a new splice acceptor site and was found in a patient with pericentral RP. However, in none of the index patients could we find a mutation in the other allele. Thus, our data are insufficient to conclude that these mutations caused the diseases of the index patients. 

The patient with CSNB and the Ile992Thr mutation deserves further mention. This patient’s ERG showed abnormalities of both the rod a-wave (probably reflecting a problem in photoreceptors) and the rod b-wave. ¹³ The expressed mutant protein is likely to have a very low level of function, which may contribute to the abnormal a-wave in this patient. The results of segregation analysis make it unlikely that Ile922Thr is a dominant allele, but the possibility remains that it is a recessive pathogenic allele. The index patient’s mother has RP and also carries the missense mutation, but whether the rod exchanger mutation contributes to her disease remains unclear. The same mutation was heterozygous in a patient with isolate RP (patient 121-067), and it is conceivable that Ile992Thr contributes to the phenotype in both of these patients. In neither of these cases could we find a mutation affecting the homologous allele. 

The genomic organization of the cone NCKX gene resembles that of the rod NCKX. The main difference is the presence of a splice variant that does not contain exon 5 and that is more abundant in the retina than in the brain. The full-length variant is the predominant isoform in the brain. Exon 5 codes for 17 amino acid residues that would be located in the cytoplasmic loop of the protein between H5 and H6. It is tempting to speculate that the absence of these residues in cone NCKX may have an effect on the protein’s function similar to that reported for the calcium channel α1A and α1B subunits, which also undergo alternative splicing to generate functionally distinct channels that may be essential for fine tuning the release of neurotransmitters at central and peripheral synapses. ^{14 15} We screened the cone NCKX gene for mutations and identified three missense changes. All showed little reduction in function and thus are unlikely to be pathogenic. 

The rod and cone exchanger proteins are well characterized and known to be necessary for phototransduction. The mutant alleles we found in the rod NCKX gene may represent rare variants in the population that are fortuitously present in our patients and thus do not cause the retinal diseases that affect these patients. Our failure to find patients whose disease can be completely explained by rod or cone exchanger mutations may be due to one of the following reasons. First, patients with retinal disease caused by exchanger mutations may exist but at a very low frequency, and we had no examples in our set of patients. Second, these exchangers are expressed in other tissues, and they may have important roles there. Thus, exchanger mutations may be embryonic lethals or cause syndromes that result in death at a young age. Third, other ion exchangers may function in the photoreceptors and compensate for the malfunction of the rod or cone NCKX proteins. However, no such proteins are yet known in photoreceptors. Fourth, some forms of retinal degeneration are the result of the coexistence of mutations in two genes. Thus, it is possible that some of the mutations identified here are part of a biallelic ¹⁶ or triallelic ¹⁷ digenic mode of inheritance. It will be interesting to see whether retinal disease develops in a knockout mouse with a disrupted rod or cone exchanger gene. 

²  Department of Ophthalmology, Kobe University, School of Medicine, Kobe, Japan.

 Supported by National Eye Institute Grants EY08683 and EY00169, the Foundation Fighting Blindness, Canadian Institutes for Health Research, the Alberta Heritage Foundation for Medical Research, and the Canadian Foundation Fighting Blindness.

 Submitted for publication October 24, 2001; revised January 8, 2002; accepted January 25, 2002.

 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.

 

List of Diagnostic Categories Screened for Mutations in the Rod and Cone NCKX Genes

Diagnosis	Patients (n)
Rod NCKX gene
Dominant RP	191
Recessive RP	230
Leber congenital amaurosis	48
Sector RP	40
Bardet-Biedl syndrome	35
Isolate RP	26
RP, genetic type unknown	25
Clumped pigmentary RP	23
Generalized choroidal sclerosis	22
Pericentral RP	21
Paravenous RP	20
Congenital stationary night blindness	12
Other categories*	122
Total	815
Cone NCKX gene
Cone–rod degeneration	87
Cone dysfunction syndromes, †	79
Total	166

*  Unilateral RP, cone-rod degeneration, age-related macular degeneration, fundus albipunctatus, optic atrophy, for example.

†  Cone degeneration or achromatopsia, for example.

Table 2.

 

Sequence Anomalies Found in the Rod NCKX Gene in Patients with Retinal Disease

DNA Change	Protein Change*	Exon	Allele Count, †	Patient	Coseg/Poly/Prot/Cons, ‡
Frameshift
754-5delAT	Met252; Ter@253	2	1	243-002	ND/−/null/+
2186delC	Pro729; Ter@821	6	1	088-005	ND/−/null/+
Inframe change
2626-52del27bp	876-884del	7	1	003-185	−/−/ND/−
2658-9ins3bp	Glu887insGlu	7	1	233-010	ND/−/ND/−
Putative splice site
G87A	Leu29Leu	2	1	038-051	ND/−/ND/ND
Missense
A109T	Thr37Ser	2	263		ND/+/ND/−
G676C	Arg239Thr	2	1	187-001	ND/−/ND/−
G931C	Val311Leu	2	3	001-346, 088-004, 099-026	−/−/ND/−
T937G	Leu313Val	2	92		−/+/ND/−
A991T	Ser331Cys	2	1	048-036	ND/−/ND/−
T1166A	Met389Lys	2	2	003-112, 003-114	ND/−/ND/−
G1375A	Val459Ile	2	1	001-217	+/−/ND/−
G1822A	Val608Ile	2	7		ND/−/ND/−
G2326C	Glu776Gln	7	1	048-069	−/−/ND/−
A2749G	Ile917Val	7	1	063-008	+/−/+/+
T2764C	Trp922Arg	7	7		ND/−/ND/+
T2885C	Val962Ala	9	1	121-245	ND/−/ND/+
T2975C	Ile992Thr	9	2	274-011, 121-067	+/−/−/+
G2980A	Ala994Thr	9	1	041-001	ND/−/+/+
T3160G	Phe1054Val	10	1	003-128	−/−/ND/−
Isocoding
C1653T	Leu551	2	2	001-412, 001-459	ND/−/NA/NA
C2778T	Pro926	7	10		ND/+/NA/NA
G3180A	Ala1062	10	1	003-062	ND/−/NA/NA
Intronic
IVS2-6T→C	tgtctgc	int2	1	003-156	ND/−/NA/NA
IVS3-22ins7bp	aggcctg	int3	424		ND/+/NA/NA
IVS3-58G→A	ctcatgt	int3	2	003-147, 274-008	ND/−/NA/NA
IVS4-6T→C	catctgc	int4	1	009-012	ND/−/NA/NA

*  Missense changes that were selected for protein function analysis are in bold.

†  Number out of a total of 1630 alleles surveyed.

‡  Coseg, the change cosegregated with disease (+), did not cosegregate (−), or was not tested (ND); poly, the minor allele had an allele frequency of more (+) or less (−) than 1%; prot, the protein function is probably null (null), was studied and does not to differ from wild-type (+), was studied and has substantially reduced function (−), or was not studied in vivo (ND); cons the change affects a conserved (+) or a nonconserved (−) region of the protein (NA, not applicable).

Figure 1.

 

Location of the sequence changes that are predicted to affect the primary structure of the rod and cone exchanger proteins. The transmembrane domains are numbered 0 to 11. The sizes of the cytoplasmic (top) and extracellular (bottom) loops are drawn approximately to scale for the rod exchanger. The cone exchanger protein differs mainly in the sizes of the first cytoplasmic loop and the third extracellular loop. Underscored mutations are those found in the cone exchanger gene; the remainder were found in the rod exchanger gene. The difference in the sizes of the rod and cone proteins explains why some mutations are closer to each other than would be expected based on codon number (e.g., cone Val284Ile is near rod Val608Ile). In bold-faced type are the sequence changes that are considered likely to produce a nonfunctional or poorly functional protein. All of these were found in the rod exchanger gene.

Figure 2.

 

Segregation analysis of sequence changes in the rod Na-Ca+K exchanger and retinal phenotypes. Filled symbols: affected individuals; open symbols: unaffected individuals; arrows: index patients. Underneath the symbol of each family member is the SLC24A1 genotype with letters indicating the alleles designated at the bottom of each pedigree and + indicating the wild-type sequence We could not determine whether the mutations Val311Leu and Leu313Val were in cis or trans in family E558. The phenotypes of the families are as follows: 5382 and B024, autosomal recessive RP; 9425, CSNB; E558 and 6721, autosomal dominant RP; OMGL-0165, Leber congenital amaurosis. In family OMGL-0066 the index patient had CSNB and his mother RP.

Figure 3.

 

An amino acid sequence alignment of human, rat, bovine, and dolphin rod exchanger, and the human cone exchanger (when applicable) homologues in the vicinity of missense mutations that remained as possible pathogenic mutations after segregation analysis. Twenty amino acids surrounding each mutation are presented. The affected amino acid is in bold.

Figure 4.

 

Genomic sequences of two SLC24A1 mutations. The mutation 2186delC is shown in the sense direction, and Ile992Thr is in the antisense direction.

Table 3.

 

Genomic Organization of the Human Cone NCKX gene ( SLC24A2 )

Exon	Exon Size (bp)	Exon (3′ end)	Intron (5′ end)	Intron Size (kb)	Intron (3′ end)	Exon (5′ end)	No. of the Subsequent Exon
1	154*	… TCTTATCCTG	gtaagtgcgg …	1.7	… tttatttaag	CATAAATGAT …	2
2	1083	… CCAAGCAAAG	gttggtggtg …	n.d.	… tcttatgcag	CCATCTGCAG …	3
3	39	… [A/G]ACTCTACCG	gtaggaagca …	2.6	… ctctgctcag	GCTAAGCCGC …	4
4	109	… CTCGCCGAAG	gtaaacatca …	>8.0	… gatgttgtag	AACTTGGATC …	5
5	51	… ACTGAAGAAG	gtaagaatga …	>8.0	… tttgcttcag	GGAGGTTCAG …	6
6	99	… AACCACGTGG	gtgagtgcag …	3.2	… tatatcacag	AAAAAATTGA …	7
7	119	… AGAAGCCCAG	gtatggcttt …	n.d.	… tctaccacag	ACCGCTGATG …	8
8	132	… TCGCAAACCT	gtaagtaaag …	n.d.	… tgtctttcag	TCATCGAGGA …	9
9	90	… GGCGCACCAG	gtaagatttt …	6.0	… tgacatttag	GTTGGAGAGA …	10
10	167	… TCACTGTAGG	gtgagtagag …	4.0	… acttctgcag	GCTCCCACTG …	11
11	>7000

 The exons and introns are shown in uppercase and lowercase letters, respectively. Sizes of exons and most of the introns are listed.

*  We identified alternative use of three different first exons of 154, 284, and 158 bp in size.

Table 4.

 

Sequence Anomalies Found in the Cone NCKX Gene in Patients with Retinal Disease

DNA Change	Protein Change*	Exon	Allele Frequencies, †			Patient
							Patients	Normal Subjects
Missense
845T→G	Phe282Cys	2	1	0	162-006
850G→A	Val284Ile	2	1	0	065-010
1802A→G	Asn601Ser	11	1	0	286-002
Silent
648C→T	Ile216	2	7	4
960G→A	Pro320	3	83	50
1201C→A	Arg401	6	65	38
1221C→T	Asn407	6	4	0
1404C→A	Arg468	8	3	0
1758G→C	Leu586	11	1	0
1857C→G	Leu619	11	2	1
1875G→A	Lys625	11	1	1
Intronic
IVS2+21A>G		2	1	0
IVS6-35(TG)n, ‡		7
IVS9-4T→C		10	3	6

*  Missense mutations that were selected for protein-function analysis are in bold.

†  Number in a total of 332 patients and 186 normal subjects.

‡  A dinucleotide simple sequence repeat with 10-25(TG) repeats; variant alleles have been observed among normal control subjects and patients.

Figure 5.

 

Function of selected mutant alleles. (A) Western blot of proteins expressed by High Five cells transfected with human cone NCKX mutants. Membrane proteins (10 μg) extracted from cells expressing the indicated human cone NCKX missense mutants were run in each lane. The exchanger protein was revealed with a human anti-c-myc antibody. (B) ⁴⁵Ca²⁺ uptake through reverse Na-Ca+K exchange was measured in cells expressing the indicated NCKX mutants and compared with that observed in cells expressing wild-type human cone NCKX. Data are expressed as mean uptake compared with wild-type uptake; error bars: SE from 11 (Ile992Thr), 9 (Ile917Val), 8 (Val284Ile), 5 (Asn601Ser), or 7 (Phe282Cys) independent determinations obtained from separate transfection experiments.

The authors thank Peggy Rodriguez and Carol Weigel-DeFranco for technical assistance and Michael A. Sandberg, PhD, for helpful discussions.