Structural and Functional Abnormalities of Retinal Ribbon Synapses due to Cacna2d4 Mutation

Abstract

purpose. In a spontaneous mutant substrain of C57BL/10 mice, severely affected retinal ribbon-type synapses have been described. The retinopathy was accompanied by a substantial loss in the activities of the second-order neurons. Rod photoreceptor responses were maintained with reduced amplitude, whereas cone activities were absent. This study was conducted to identify the genetic defect underlying this hitherto unknown autosomal recessive cone–rod dysfunction.

methods. Genome-wide linkage analysis and screening of positional candidate genes were used to identify the causative mutation. Tissue-specific transcriptional activity of the defective gene was determined by Northern blot analysis and RT-PCR approaches. The number of cone photoreceptors was estimated by immunohistochemistry.

results. The mutation was localized to a 275-kb region of chromosome 6. Within this candidate interval, a homozygous frameshift mutation (c.2367insC) was identified in the Cacna2d4 gene of affected animals. This gene codes for an L-type calcium channel auxiliary subunit of the α2δ type. The mutation introduces a premature stop codon that truncates one third of the predicted Cacna2d4 protein. A severe reduction in Cacna2d4 transcript levels observed in mutant retinas probably results in the lack of Cacna2d4 protein. The mutation leads to significant loss of rods, whereas the number of cone cells remains unaffected until 6 weeks of age.

conclusions. The Cacna2d4 mutation underlies a novel channelopathy leading to cone–rod dysfunction in the visual system of mice and provides a new candidate gene for human retinal disorders including night blindness, retinitis pigmentosa, and cone–rod dystrophies.

Clinical symptoms of hereditary retinal diseases range from moderate night blindness to severe visual impairment and complete blindness in various forms of retinal degenerations. The elucidation of the underlying genetic defects helps to understand retinal function and pathogenic mechanisms. In a previous study, an autosomal recessive cone–rod dysfunction was discovered in a substrain of C57BL/10 mice. ¹ The phenotype is characterized by a disturbed signal transmission of photoreceptor cells to adjacent neurons. On dark- and light-adapted electroretinography, a profound loss of the b-wave, which requires synaptic transduction, has been detected in affected animals. The scotopic a-wave, which mainly reflects membrane potentials of rods, has a consistently reduced amplitude. Under photopic conditions however, cone-specific activity is absent. The substantial loss of postphotoreceptoral activities is associated with a severely compromised morphology of retinal ribbon-type synapses. The causative mutation of this disorder remains unknown. 

Ribbon-shaped synapses are known to involve the high-voltage-gated L-type calcium channels preferentially in their signaling pathways. ² These calcium channels are clustered at presynaptic membranes beneath ribbon organelles that guide synaptic vesicles to active zones for calcium-mediated fusion. ³ L-type calcium channels are heteromultimers composed of four independently encoded proteins, the pore-forming α1 subunit, which triggers calcium flow across the membrane, and the auxiliary subunits α2δ, β, and γ. ⁴ Mutations affecting the functionality in two of these calcium channel subunits lead to failure of retinal processing. Allelic variants of the retina-specific α1F subunit (CACNA1F) were found to be responsible for X-linked incomplete congenital stationary night blindness type 2 (CSNB2). ⁵ Mutations either result in complete loss of function or modify CACNA1F kinetics. ⁶ Certain variants of CACNA1F were associated with clinical symptoms of cone–rod dystrophies. ^{7 8} An abnormal morphology of the photoreceptor ribbon synapses and seriously diminished signal transmission from photoreceptor terminals to the second-order neurons were detected in mice lacking the β2 auxiliary subunit. ⁹ The recently reported mouse model for human CACNA1F-mediated CSNB2 revealed a cone–rod dysfunction due to a targeted deletion of the Cacna1f gene. ¹⁰ Of note, striking similarities were found between the phenotypic features of ribbon synapses in Cacna1f-mutant and affected C57BL/10 mice. 

We applied genetic mapping and screening of positional candidate genes to elucidate the causative mutation in affected C57BL/10 mice. By genome-wide linkage analysis, we localized the causal mutation on chromosome 6 and identified the defect in the gene encoding the fourth L-type calcium channel α2δ auxiliary subunit, Cacna2d4. This gene may be responsible for proper assembly and auxiliary modulation of biophysical properties of L-type calcium channels, similar to the previously described human CACNA2D4 orthologue. ¹¹ The mutation found in Cacna2d4 underlies a novel channelopathy leading to cone–rod disease in the visual system of mice. Our results provide insights into a new pathophysiology of retinal ribbon-type synapses and a new candidate gene for human retinal disorders including night blindness, retinitis pigmentosa, and cone–rod dystrophies. 

Materials and Methods

Animal Breeding

Mouse strains were obtained from regular providers of laboratory animals. The research was performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

To perform linkage, C57BL/10 mutant animals were crossed with animals of the AJ inbred mouse strain to generate heterozygosity in polymorphic markers. The heterozygous hybrids (F1) were subsequently mated for the second generation (F2). To refine linkage analysis, successive breeding strategy was repeated with new parental animals. Affected F2 mice were selected for mapping. 

Phenotypic Characterization

In an earlier study, extended electrophysiological examinations of the phenotype were performed in animals up to the age of 14 months. ¹ For screening purposes, the animals were characterized by scotopic ERG at the age of 3 to 6 weeks. Briefly, dark-adapted, anesthetized mice were placed in a Ganzfeld bowl exposing white-light flashes to the dilated eye of the mouse (flash energies ranging from 10⁻⁴ to 0.5 log cds/m²). A monopolar contact lens electrode served as the recording electrode. Light-induced retinal activities were visualized in ERG recordings. Representative negative scotopic ERG recordings of affected animals exhibited reduction of the a-wave amplitude and a distinct attenuation of the b-wave. Cone ERGs were nonrecordable in affected animals at any age. 

Retinal Morphology

To exclude modifying effects of the AJ genetic background, we looked for alterations in the retina of affected B10AF2 hybrids by light microscopy. Paraffin sections of 10 mutant and 10 wild-type eyes were stained with hematoxylin-eosin dye complex and evaluated with a microscopy station (Axioplan 2; Carl Zeiss Meditec, Inc., Feldbach, Switzerland). For this, eyes were enucleated from mice anesthetized and subsequently killed with CO₂ and fixed in Serra’s (60% ethanol, 30% formalin, 10% acetic acid) over night. After dehydration in an isopropanol series (70%, 80%, 90%, 100%, 100%), eyes were embedded in paraffin blocks and sectioned into 5-μm slides. 

Genotyping and Linkage Analysis

DNA was extracted from tail biopsy samples by standard methods. A whole-genome screen was performed with 75 informative microsatellite markers. Products of PCR assays with fluorescently labeled primers were analyzed by automated capillary genotyping (MegaBACE 1000, scored with genetic profiler analysis software; GE Healthcare). To identify linkage, homozygosity for C57BL/10 alleles was sought on chromosomes of 64 affected animals from the first B10A population. Significant linkage was confirmed by statistics computed by comparing the number of homozygous versus heterozygous mice and calculating χ² values. Genetic maps were constructed with the program Mapmaker. ¹² To narrow the linkage interval, we screened the critical region for new polymorphic markers by alignment of the locus spanning C57BL/6J genomic contigs NM_000264 and NM_000265 to corresponding AJ genomic sequences obtained from the Celera database (http://www. celera.com/ Celera Genomics, Rockville, MD). In addition, the Mouse Single Nucleotide Polymorphism (http://mousesnp.roche.com, Roche Molecular Biochemicals, Indianapolis, IN) and Entrez SNP Database (http://www.ncbi.nlm.nih.gov/ provided in the public domain National Center for Biotechnology Information, Bethesda, MD) were scanned for further markers. Identified polymorphisms (6 annotated microsatellite markers, 3 known and 22 new SNPs) were verified by allele-specific PCR, denaturing polyacrylamide gels, or direct sequencing and used to genotype affected F2 animals from both B10A populations (Supplementary Table S1). Primers were purchased from Microsynth (Balgach, Switzerland). (Genomic sequences including identified SNPs and primers are available on request.) Unaffected B10AF2 animals were genotyped for flanking markers of the final interval. 

Mutational Analysis

Candidate genes were selected on functional relevance and analyzed by direct sequencing of the open reading frame (ORF) or whole mRNA. Mainly, RT-PCR was directed from whole-eye cDNA of two mutants and two wild types to prescreen exon splicing and deletion or insertion defects, applying primers designed to amplify ORF/mRNA-spanning fragments of ∼550 bp. As no mutation was found in all investigated candidate genes, we proceeded with sequence analysis of the entire interval, defined by fine mapping. We performed primer design through the region to amplify overlapping fragments of 500 to 650 bp corresponding to the genomic reference sequence of C57BL/6J mice (http://genome.ucsc.edu/ Human Genome Browser, provided in the public domain by UCSC Genome Bioinformatics, University of California at Santa Cruz, Santa Cruz, CA). Purified fragments (ExoSap-IT Kit; USB, Cleveland, OH) were bidirectionally sequenced for two mutants and two wild types (Prism 3100 Genetic Analyzer in accordance with the Prism Big Dye Chemistry ver. 1.1 protocol; Applied Biosystems [ABI], Rotkreuz; Switzerland). Sequences were then aligned (PRISM SeqScape; ABI) and examined for nucleotide exchanges. Putative mutations were verified in all affected animals and wild types. 

RNA Isolation and cDNA Synthesis

After tissue homogenization with a glass pestle, total RNA extraction was performed (RNeasy Mini Kit; Qiagen, Basel, Switzerland). Through reverse transcription, 1.5 μg total RNA was transcribed into cDNA by random hexanucleotide priming (hexamer primers pd(N)₆, GE Healthcare, Otelfingen, Switzerland; and Superscript III; Invitrogen, Basel, Switzerland). cDNA synthesis efficiency and DNA contamination were examined by standardized Gapdh PCR including negative control experiments for each sample with no reverse transcriptase used during synthesis. Absent Gapdh product in the negative control experiments excluded DNA contamination. 

RT-PCR and Real-Time PCR

Gene transcription and the exonic composition of Cacna2d4 were verified by RT-PCR on total eye–retina RNA. For this, overlapping fragments spanning the entire mRNA were amplified in wild types and mutants and subsequently sequenced: (1) forward (for): 5′-ttgatcacctggggtggacc, reverse (rev): 5′-acaggatgattggcgtcttc; (2) for: 5′-ttgatcacctggggtggacc, rev: 5′-gtgttgtacagattcctgcc; (3) for: 5-ttgatcacctggggtggacc; rev: 5′-agcatggtctccatgtcctc; (4) for: 5′-ttgatcacctggggtggacc; rev: 5′-ccagccacggtttctgcagtc; (5) for: 5′-gcgtgtatatgtccgaagcc; rev: 5′-ccatgtaagcctctgtccag; (6) for: 5′-ggagtcatcgccttcgactgc; rev: 5′-agaaggcatagccatgcacc; (7) for: 5′-gtggccatgcctgtcttcag; rev: 5′-gtctctctgagatcaggacg; (8) for: 5′-cactgatctcgactgcttcg; rev: 5′-ttcttgtgcttgtgggagtg; and (9) for: 5′-tgttagagtggagtgcctgg; rev: 5′-gagacatcccacagcagtga. Cacna2d4 mRNA amounts in eyes were examined with real-time PCR using cDNAs from five wild-type, five heterozygous, and five mutant animals from the B10A strain at the age of 6 weeks. The experiment was repeated four times. For each animal, two independent cDNA syntheses were performed and pooled after quality control. Subsequently, two independent dilutions were prepared from each cDNA pool. Quantitative RT-PCR for retinal tissue was performed by pooling six retinas for RNA extraction from three homozygous wild-type, three heterozygous, and three mutant animals each. Three dilutions of two independent cDNA syntheses were examined for Cacna2d4 transcripts. The amplification (for: 5′-ctgccatgccttccatccag, rev: 5′-gagacatcccacagcagtga) was performed at 50°C for 2 minutes and 95°C for 10 minutes and then continued with 45 cycles of 15 seconds at 90°C and 1 minute at 60°C for detection of a 221-bp Cacna2d4-specific fragment (with SYBR Green nucleic acid stain; ABI) (exon 37–38) on a detection system (7900HT Prism; ABI). A melting curve was analyzed to verify single-product amplification. In addition, the specificity of the amplicons was confirmed by sequencing. Transcript levels were normalized to 18S rRNA (TaqMan Ribosomal RNA Control Reagent; ABI) and quantified to the mean level in homozygous wild types according to the ΔC_t-method. 

Northern Blot Analysis

For hybridization, a 550-bp cDNA fragment encoding the Cacna2d4-specific 3′end (exons 33–38; for: 5′-tgttagagtggagtgcctgg, rev: 5′-gagacatcccacagcagtga) was used as a probe. Labeling with [α-³²P]dCTP* (10 μCi/μL; Hartmann Analytic, Braunschweig, Germany) was performed according to the manufacturer’s instructions (Prime-It II Random Primer Labeling Kit; Stratagene, Amsterdam, The Netherlands). Denatured DNA probe was transferred to a hybridization tube (ULTRAhyb hybridization buffer; Ambion, Cambridgeshire, UK), Mouse Cot-DNA and a Northern blot preincubated for 50 minutes at 42°C. We hybridized a Poly A⁺ RNA commercial murine northern blot (Multi Choice; OriGene, Rockville, MD). After incubation for 15 hours at 42°C, the blot was washed twice with 2× SSC and 0.1% SDS for 5 minutes at 42°C and twice with 0.1× SSC and 0.1% SDS for 15 minutes at 65°C. Finally, the blot was exposed to film (BioMax; Eastman Kodak, Rochester, NY) at −80°C overnight. As the loading control, mouse β-actin probe (OriGene) hybridization was included. 

Cell Quantification in the Retina

Immunohistochemistry was preformed on 8-μm whole-eye cryosections of six affected and six wild-type littermates, as described elsewhere. ¹³ Gα cone transducin primary antibody (Santa Cruz Biotechnology, Santa Cruz, CA) was used to detect the cones. ¹⁴ As a secondary antibody, Cy3 labeled anti-rabbit IgG was applied (Dianova, Hamburg, Germany). Subsequently, tissue sections were embedded in mounting medium that contained DAPI (Vectashield; Vector Laboratories, Burlingame, CA), to visualize cell nuclei. Images were generated at 63× magnification by the fluorescence microscopy station (Axioplan2; Carl Zeiss Meditech) from three sections per animal, representing retinas including the optic nerve. Cone marker-positive cells were quantified in two 200 × 50-μm² subfields of the peripheral and two of the central outer nuclear layer per each image. In addition, the number of cells in the central outer and inner nuclear layers was counted in a 163 × 18-μm² subfield (n = 4) in DAPI-stained retinal cryosections for each animal. 

Statistics

Statistical relevance for ERG measurements, real-time PCR and cell number quantification was analyzed with the Mann-Whitney test (SPSS ver. 13 for Windows; SPSS Inc., Chicago, IL) and accepted if P < 0.05. Graphic data interpretations included the calculation of the SE visualized in confidence intervals. 

Results

Genome-Wide Linkage Analysis

To identify the mutation, we crossed the C57BL/10 mutant genome to the genetic background of the phylogenetically distant AJ mouse strain generating substantial heterozygosity in the alleles of genetic markers. Heterozygous (C57BL/10 x AJ)F1 mice revealed no phenotypic alterations. Subsequently, (C57BL/10 x AJ)F1 mice were intercrossed. In total, two independent C57BL/10 x AJ breedings were performed. 

In the B10AF2 generation, segregation into affected and unaffected phenotypes in a ratio of 25:75% was found in accordance with the originally suggested autosomal recessive transmission of the mutation. Among a total number of 580 F2 mice, 142 animals (24.48%) showed the characteristic negative ERG in scotopic conditions (Fig. 1) . 

Structural and functional alterations in affected B10AF2 animals were similar to the symptoms of the inbred line C57BL/10 (Fig. 2) . ¹ Affected B10AF2 offspring displayed significant reduction of the scotopic a-wave and profound loss of the b-wave. In affected B10AF2 mice, the a-wave amplitude was attenuated to 80.3% (Fig. 2A) . The severe reduction of the b-wave amplitude was reflected by the a/b-waves, calculated as the difference between the a- and b-wave maxima. The a/b-wave revealed a decrease to 25% in affected B10AF2 animals (Fig. 2A) . Corresponding to the defective signal transmission, the loss of the b-wave was associated with a significantly attenuated photoreceptor synaptic layer (outer plexiform layer, OPL; Fig. 2B ). Homozygous and heterozygous B10AF2 unaffected mice revealed no differences on scotopic electroretinography and morphology. Accordingly, the genetic background of the AJ strain did not apparently modify the retinal phenotype. 

For genetic localization of the mutation, 64 affected mice from the first B10AF2 breeding were included in a genome-wide linkage analysis with an initial marker-spacing of 20 cM. A homozygosity region was detected on chromosome 6 between the markers D6Mit389 and D6Mit218, comprising an interval of 15.6 Mb (Figs. 3A3B) . Linkage to chromosome 6 was evident through a distance of >30 cM, as revealed by allelic frequencies resulting in significant χ² values (Supplementary Table S2). We detected no further linkage elsewhere in the genome of the B10A hybrid strain (Supplementary Fig. S1). The linkage interval was refined by genotyping of additional SNPs and microsatellite markers to an interval of 7.83 Mb flanked by SNPs in Cacna1c and Akap3 genes. For further fine mapping, we proceeded by generating and genotyping a second B10AF2 hybrid population (total n = 319, affected n = 78) and reduced the interval to 275 kb (Fig. 3B) . 

Identification of Sequence Variants in Cacna2d4

In parallel experiments, we excluded sequence variations in 33 candidate genes from the initially identified 15.6-Mb region (Supplementary Table S3). In the final 275-kb interval, six potential candidate genes remained (Fig. 3B) . In their annotated mRNA sequences, no mutations were detected. Thus, considering the occurrence of a defect in an additional exonic or noncoding region, we sequenced the entire 275-kb locus. We identified two DNA sequence variations segregating with the phenotype. The first was found in a compound tetranucleotide repeat in intron 6 of Cacna2d4 and showed an insertion of an additional GAAA unit in a 12-unit GAAA replicon. However, RT-PCR examinations for a potential splicing defect revealed no exon skipping in affected mice (data not shown). In addition, sequencing of the orthologous repeat in the AJ mouse strain identified the variant of 14 GAAAs, strongly suggesting a polymorphic nature of the tetranucleotide repeat. 

The second sequence alteration was detected in a recently annotated 3953-bp long variant of Cacna2d4 mRNA (BK005394). This transcript consists of 38 exons and comprises 114.032 kb of genomic DNA. The 3350-bp ORF begins in exon 1 (ATG = 181 bp) and extends to exon 38 (TGA = 3531 bp), coding for a protein of 1116 amino acids. 

In a short C_{(n = 6)} monorepeat, affected animals displayed an insertion of a cytosine nucleotide (C_{(n = 7)}) in exon 25 of the Cacna2d4 gene (Fig. 4A) . The mutation was confirmed in all affected animals (n = 142, B10AF2 and n = 30, C57BL/10). In unaffected mice, exclusively homozygous (C_{(n = 6)}) or heterozygous (C_{(n = 6/7)}) allelic combinations were detected (n > 500). 

The identified insertion (c.2367insC) induces a frameshift that substitutes 14 amino acids and subsequently results in a premature stop in the corresponding protein sequence (Fig. 4B) . The mutation removes 327 amino acid residues from the carboxyl terminus of the Cacna2d4 subunit representing 29.3% of the protein. 

Qualitative and Quantitative Analysis of Cacna2d4 mRNA

We confirmed the predicted exonic structure and analyzed the expression of Cacna2d4 in wild-type and mutant retinas by RT-PCR (Fig. 5A) . RT-PCR of fragment 7 (exons 14–29) revealed a single product that included exon 25 (Fig. 5B) . The mutation was verified in mutant animals by direct sequencing of RT-PCR products. No additional fragments were detected in amplicon 7, which may correspond to alternatively spliced Cacna2d4 transcripts lacking exon 25 (Fig. 5B) . 

Real-time PCR experiments revealed a significant reduction of Cacna2d4 mRNA levels in eyes and retinas of affected and heterozygous animals (Figs. 6A 6B) . The relative amount of Cacna2d4 transcripts in eyes of homozygous mutant mice was diminished to 45% in comparison to wild-type animals. In the retina, a decline to 31% was observed. In the retina of heterozygous animals, a significant decrease to 73% was detected (Figs. 6A 6B) . In eyes of heterozygous mice, statistical analysis indicated significant reduction to 82% in comparison to wild-type control animals. 

We also examined Cacna2d4 expression in 13 additional mouse tissues. On a Poly A⁺ Northern blot including 12 tissues, we detected bands in the size range of 1.5 to 5 kb, most likely corresponding to alternative splice variants of Cacna2d4 as suggested from EST databases (GenBank accession no.: XM_132795 with 2.2 kb, AK030723 with 2.6 kb, AK044427 with 3.23 kb, XM_132794 with 4.8 kb, and AK137847 with 3.1 kb; Fig. 6C ). Strongest signals occurred at the size of ∼4.6 and ∼2.0 kb in most tissues. In brain, muscle and spleen, remarkably weaker signals were detected compared with other organs. Nevertheless, transcription of Cacna2d4 was evident in all tissues analyzed, indicating a ubiquitous expression pattern. Broad Cacna2d4 expression was confirmed by RT-PCR of fragment 9 (exons 33–38) in 12 tissues (Fig. 6D) . Additional bands detected in amplicon 9 occurred in the range of 300 to 400 bp in kidney, muscle, spleen, and stomach and may indicate alternative splicing at the 3′ end of Cacna2d4. 

Examination of Cone Numbers in Affected Mice

In previous studies, attenuated scotopic photoreceptor responses in mutants (scotopic a-wave) have been interpreted to reflect reduced rod photoreceptor number. ¹ The complete absence of cone-specific activities remained unexplained. To elucidate whether a preferential loss of cone photoreceptors occurs in mutant animals, immunohistochemical staining of cones was performed. No obvious alterations in the pattern of cones were detected. The cone density in the central and peripheral retina of 6-week-old mutant animals did not significantly differ from that of wild-type mice (Fig. 7A) . However, a ∼15% reduction in the number of cells in the outer nuclear layer was detected, indicating an early degeneration of rod photoreceptors (Fig. 7B) . 

Discussion

In the present study, we report genetic mapping of the respective trait and the identification of the disease-causing mutation in the voltage-gated L-type calcium channel auxiliary subunit Cacna2d4. This gene is ubiquitously expressed and shows significantly reduced transcript amounts in the retina of affected and heterozygous animals. The disease is caused by a homozygous single nucleotide insertion in exon 25 of Cacna2d4 which results in a frameshift that truncates one third of the predicted amino acid sequence. This defect represents a novel channelopathy that disturbs synaptic signal processing in the retina and leads to functional failure of cones and early onset degeneration of rods. 

The causative nature of the identified mutation is supported by several lines of evidence. Linkage analysis revealed a single locus limited to only 275 kb on chromosome 6. No further candidate regions were detected in the mouse genome on linkage analysis. The mutation perfectly segregates with the disease phenotype. Homozygosity for this mutation was detected only in affected animals (n = 172). In unaffected mice (n > 500), homozygosity for wild-type alleles or heterozygosity were observed. No sequence variations were found in any other gene from the linkage interval. Moreover, the expression of the respective Cacna2d4 mRNA was confirmed by RT-PCR analyses in the mouse retina and the mutation was verified in affected animals by direct sequencing of the relevant RT-PCR fragments. The premature termination signal may lead to selective mRNA degradation by nonsense mediated decay (NMD), to prevent translation of truncated proteins. ¹⁵ Indeed, we found significantly reduced mRNA levels in homozygous mutant mice by quantitative RT-PCR. The mutation introduces a premature stop codon 77 nucleotides upstream of the boundary of exons 26/27, a position that is efficiently recognized by the NMD machinery. It is noteworthy that we did not obtain evidence by RT-PCR for alternative splice variants of Cacna2d4 lacking exons 25 and 26. This observation suggests that these two exons are constitutively expressed in the retina. 

Nevertheless, residual mutant Cacna2d4 transcripts may undergo translation. However, this probably results in a nonfunctional product as it lacks almost one third of the protein at the C terminus. The deduced mouse sequence of Cacna2d4 shows a high degree of sequence identity and conserved domain structure in many species, as demonstrated by comparative protein alignments (part of data shown in Fig. 4C ). The human CACNA2D4 protein shares 79% identical amino acid residues with its mouse orthologue. The truncated part of Cacna2d4 displays 93% similarity in both species on the amino acid level and contains functionally important domains. Each α2δ subunit identified so far consists of a single-gene product cleaved posttranslationally into α2 and δ peptides linked by disulfide bridges. ^{4 16} The δ peptides possess a conserved single transmembrane segment at the carboxyl terminus for cell surface attachment. ¹⁷ In human CACNA2D4, a homologous highly hydrophobic region encoded by the last two exons (37 and 38) potentially acts as the transmembrane domain. ¹¹ This motif is removed in mutant mice by the premature stop codon (Fig. 5A) . Furthermore, the proteolytical cleavage site between the α2 and δ peptides is encoded by a conserved alanine residue at position 971 in exon 34. The premature translational stop occurs at amino acid position 802 (exon 26) and thus eliminates the entire δ subunit. Hence, membrane integration of Cacna2d4 is probably abolished as a major consequence. ¹⁸ An additional 170 amino acid residues (802-971) are truncated from the C terminus of the α2 peptide (exons 26–34). This deletion comprises one of the two Cache domains formed by amino acid residues 869-879. These domains, highly conserved in all α₂ subunits, are supposed to be involved in interaction with calcium channel complexes. ¹⁹ Finally, at least 11 cysteine residues, conserved in all murine and human α2δ subunits, are removed by the truncation. ²⁰  

Loss of Cacna2d4 function may have severe effects on the retinal physiology in C57BL/10 mutant animals. The α2δ subunits are determinants of auxiliary stimulation of calcium channel complexes. ^{17 21} Beyond this, they aid in augmentation of α1 subunits on the plasma membrane increasing the density of functional channels on the cell surface and thus accelerating calcium current amplitudes. ^{22 23} In cell lines, human CACNA2D4 enhances the α1C/β3-mediated calcium influx threefold. ¹¹ The precisely modulated calcium signaling and the overall integrity of α1 subunits into the presynaptic membranes may be severely affected in C57BL/10 mutant mice and therefore result in compromised neurotransmission of the retinal ribbon synapses. Loss of function of the retina-specific α1F subunit in the recently described Cacna1f mouse mutant revealed absence of synaptic signaling and was associated with substantial degeneration of photoreceptor ribbon terminals. ¹⁰ Transgenic mice deficient for the β2 auxiliary subunit demonstrate loss of synaptic processing due to lack of trafficking and proper assembly of α1F complexes in photoreceptor synaptic membranes. ⁹ Similar to our mutants, loss of ribbon-type synapses was detected. 

Two additional L-type α1 subunits, Cacna1c and Cacna1d, are involved in neurotransmission and synaptic plasticity of retinal ribbon synapses. ^{24 25 26 27 28} Strong expression of N- and P/Q-type calcium channels at plasma membranes of retinal neurons was also observed. ^{29 30} Because α2δ subunits have been identified as components of all voltage-gated calcium channel complexes, these channels may be subject to Cacna2d4 modulation. Likewise, other retinal channels may underlie Cacna2d4-mediated stimulation. ³¹ Preliminary results in immunofluorescence microscopy of retinal sections stained with antibodies detecting α1 subunits of voltage-gated calcium channels (anti-pan Cavα1 antibody (Alomine Laboratories, Munich, Germany), anti-α1C and anti-α1F (Santa Cruz Biotechnology, Santa Cruz, CA; data not shown) indicated reduced signals in the OPL of mutant animals. This may be a consequence of the reduced thickness of the OPL or may imply a decreased density of α1 subunits at the synaptic terminals. 

The phenotype in mutant mice is restricted to retina, as no further apparent abnormalities are detectable, although the expression is ubiquitous. Affected animals display normal behavior and body weight and regular breeding. In correlation to the cone–rod dysfunction of the Cacna1f-mutant, these mice are probably congenitally blind. ¹⁰ Yet, the elimination of ∼15% of photoreceptor cells in mutant animals may indicate additional degenerative processes. As minor alterations in calcium concentration modulate a diversity of molecular processes, the dysregulation of the calcium homeostasis may activate further pathogenic mechanisms responsible for the rod photoreceptor cell death. ³² The identification of the Cacna2d4 defect may have implications for corresponding retinal diseases in human patients. The deficiency of Cacna2d4 may be associated with rare human autosomal recessive incomplete CSNB. However, because a mutation in CACNA1F was also found in a family with a cone–rod dystrophy, ⁷ the CACNA2D4-mediated channelopathy may involve both night blindness and cone–rod dystrophy. Finally, the early degeneration of rod photoreceptors in mutant mice may indicate a causative role of mutations in this gene, also in patients with retinitis pigmentosa. 

³  Contributed equally to the work and therefore should be considered equivalent authors.

 Supported by the Foundation Fighting Blindness, Swiss National Science Foundation Grant 3100-067786 (WB), Grant DFG Ru 457/1-1 from Deutsche Forschungsgemeinschaft, and Grant 01GR0416 from the German Federal Ministry of Science and Education through the National Genome Research Network (PN).

 Submitted for publication March 14, 2006; revised April 19, 2006; accepted June 21, 2006.

 Disclosure: K.A. Wycisk, None; B. Budde, None; S. Feil, None; S. Skosyrski, None; F. Buzzi, None; J. Neidhardt, None; E. Glaus, None; P. Nürnberg, None; K. Ruether, None; W. Berger, None

 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.

 

Figure 1.

 

Phenotyping of B10A mice. Scotopic ERG response to a 0.5 log cd-s/m² white flash recorded from 6-week-old B10AF2 or C57BL/10 wild-type (wt), heterozygous (het) and affected mice (mut). In the mutant B10AF2 mouse, electronegative configuration of scotopic responses (absent b-wave) is observed, similar to that in affected animals of the original C57BL/10 strain.

Figure 2.

 

Phenotype verification in the B10A strain. (A) Significant reduction of the a- and a/b-wave amplitudes in 6-week-old B10AF2 mutants (mut) in comparison to wild-type (wt) mice (P < 0.05). (B) Confirmation of synaptic disorganization in retina of a B10AF2 mutant. Morphology was obtained by light microscopy of a retinal section stained with hematoxylin-eosin. B10AF2 mutants at 6 weeks of age display reduction of the photoreceptor synaptic layer (arrow). Outer segments (OS), inner segments (IS), outer nuclear layer (ONL), outer plexiform layer (OPL), inner nuclear layer (INL), inner plexiform layer (IPL), ganglion cell layer (GCL). Magnification ×63; scale bar, 20 μm.

Figure 3.

 

Linkage analysis. (A) Allele distribution on chromosome 6. Line aa: homozygous AJ; line bb: homozygous C57BL/10; line ab: heterozygous B10AT2 allele combinations. The number of mice homozygous for C57BL/10 alleles (bb) significantly increases between the markers D6Mit102 and D6Mit14. Shaded area: initial linkage interval of 15.6 Mb occurring between the cytogenetic bands F1–F3 on chromosome 6. (B) Physical map of the critical region on chromosome 6. The linkage interval is represented on top. The refined critical region is indicated in the bottom line. Selected DNA markers are included. The finally identified locus comprises 275 kb of genomic DNA and six genes.

Figure 4.

 

Mutation analysis. (A) Sequence examination of Cacna2d4 exon 25. The electropherograms indicate a cytosine monorepeat sequence of a wild-type (a), heterozygous (b), and mutant (c) animal. Genotypes are indicated by +/+ (wild-type), +/− (heterozygous), and −/− (mutant). (B) The mutant sequence (mut) shows an insertion of a single nucleotide (in bold) resulting in a frameshift and the substitution of 14 amino acids (indicated in bold, bottom panel). The 15th subsequent codon encodes a premature stop codon in mutant mice. (C) Partial sequence alignment of mouse Cacna2d4 and orthologous proteins from 12 species. Boxed area: The conserved region substituted in mutant animals. (GenBank accession no.: rat ENSRNOG00000008031; human ENSG00000151062; chimpanzee ENSPTRG00000004521, cow ENSBTAG00000021994; dog ENSCAFG00000015970; chicken ENSGALG00000013007; xenopus ENSXETG00000000900; zebrafish ENSDARG00000026855; fugu SINFRUG00000149170; spider GSTENG00014501001; and gnat ENSANGG00000018436; drosophila CG12295.

Figure 5.

 

Schematic drawing of Cacna2d4 transcript and RT-PCR analyses. (A) Exons are labeled by numbers (length of exons is not to scale). Overlapping RT-PCR fragments are represented by lines 1 to 9. These fragments were amplified and sequenced to confirm the exonic structure of the Cacna2d4 transcript. (★) Conserved cysteine residues in the truncated part of the protein. (B) RT-PCR of fragment 7 (exons 14–29) and 9 (exons 33–38) in retina of wild-type (wt), heterozygous (het), and mutant (mut) animals. L, the DNA ladder.

Figure 6.

 

Cacna2d4 transcript analysis. (A) Eye-specific expression of Cacna2d4 by real-time PCR. Confidence intervals (error bars) represent biological variability of five animals for each genotype. The Cacna2d4 transcript amount of each animal was normalized to the median value of the wild-type controls. The diagram displays a significant reduction of Cacna2d4 transcript abundance in eyes of affected animals compared with wild types (P < 0.05). In heterozygous animals, the confidence interval overlaps with wild type; however, statistical analyses indicate a significant tendency toward reduced transcript levels in those animals. (B) Retina-specific expression of Cacna2d4 by real-time PCR. Confidence intervals indicate experimental and biological variability of the quantitative RT-PCR performed with three cDNA dilutions from pooled retinas of three mice per genotype. Homozygous mutant and heterozygous animals show significantly reduced transcript levels in comparison to wild-type mice (P < 0.05) (C) Expression of Cacna2d4 in nonocular mouse tissues. Multitissue Poly A⁺ Northern blot analysis detects Cacna2d4 transcripts in 12 tissues: lane 1: brain; lane 2: heart; lane 3: kidney; lane 4: liver; lane 5: lung; lane 6: muscle; lane 7: skin; lane 8: small intestine; lane 9: spleen; lane 10: stomach; lane 11: testis; lane 12: thymus. Tissue-specific alternative splice variants were observed. (D) RT-PCR of fragment 9 comprising exon 33 to 38 in 12 mouse tissues: lane L: DNA ladder, lane 1: brain; lane 2: heart; lane 3: kidney; lane 4: liver; lane 5: lung; lane 6: muscle; lane 7: skin; lane 8: spleen; lane 9: stomach; lane 10: testis; lane 11: olfactory bulb; lane 12: eye. Bands occurring at the size of ∼100 to 200 bp are PCR primer dimers. Bands of 300 to 400 bp in kidney, muscle, spleen, and stomach may indicate alternatively spliced isoforms of Cacna2d4.

Figure 7.

 

Comparison of cell numbers in homozygous mutant and wild-type retinas. (A) The average density of cone Gα transducin-immunopositive photoreceptors examined at central (central) and peripheral (peri) retina in 163 × 18-μm² subfields of six wild-type (wt) and six mutant (mut) mice. (B) Number of cells in a 163 × 18-μm² subfield of the ONL (outer nuclear layer) and INL (inner nuclear layer) examined at central retina. In the ONL, significant reduction of photoreceptor cell number occurs (P < 0.05). Error bars, confidence intervals.

Supplementary Materials

Supplementary Table S1 - (PDF) Summary of relative allelic distribution for genetic markers. Locus determining markers are shaded in grey. Significance of linkage is visualized by Chi^2 values. 

Supplementary Table S2 - (PDF) List of analyzed candidate genes. 

Supplementary Table S3 - (PDF) Markers identified and applied in fine mapping. 

Supplementary Figure S1 - (PDF) Mouse genome scan. 

The authors thank Stefanie Derzsi, Ulrich Luhmann, Gabor Matyas, Gaby Tanner, Mariana Wittmer, and Christina Zeitz for discussions and experimental assistance.