Association of Whirlin with Ca_v1.3 (α_1D) Channels in Photoreceptors, Defining a Novel Member of the Usher Protein Network

Abstract

Purpose.: Usher syndrome is the most common form of hereditary deaf-blindness. It is both clinically and genetically heterogeneous. The USH2D protein whirlin interacts via its PDZ domains with other Usher-associated proteins containing a C-terminal type I PDZ-binding motif. These proteins co-localize with whirlin at the region of the connecting cilium and at the synapse of photoreceptor cells. This study was undertaken to identify novel, Usher syndrome-associated, interacting partners of whirlin and thereby obtain more insights into the function of whirlin.

Methods.: The database of ciliary proteins was searched for proteins that are present in both the retina and inner ear and contain a PDZ-binding motif. Interactions with whirlin were evaluated by yeast two-hybrid analyses and validated by glutathione S-transferase pull-down assays, co-immunoprecipitation, and co-localization in the retina with immunofluorescence and immunoelectron microscopy.

Results.: The L-type calcium channel subunit Ca_v1.3 (α_1D) specifically interacts with whirlin. In adult photoreceptors, Ca_v1.3 (α_1D) and whirlin co-localize in the region of the connecting cilium and at the synapse. During murine embryonic development, the expression patterns of the Whrn and Cacna1d genes show significant overlap and include expression in the eye, the inner ear, and the central nervous system.

Conclusions.: The findings indicate that Ca_v1.3 (α_1D) is connected to the Usher protein network. This conclusion leads to the hypothesis that, in the retina, whirlin scaffolds Ca_v1.3 (α_1D) and therefore contributes to the organization of calcium channels in the photoreceptor cells, where both proteins may be involved in membrane fusions.

Usher syndrome is the most common form of human hereditary deaf-blindness and has an autosomal recessive pattern of inheritance. Three clinical types, USH1, -2, and -3, are distinguished based on the progression and severity of the hearing loss and the presence or absence of vestibular dysfunction, with visual loss due to retinitis pigmentosa (RP) in all three types. ¹ To date, the affected genes have been identified for 9 of the 12 described Usher loci. ^2–10 Besides being causative of Usher syndrome, mutations in the Usher genes MYO7A, USH1C, CDH23, PCDH15, and DFNB31 are also associated with nonsyndromic hearing loss. ^2,6–10 In addition, mutations in the USH2A gene can lead to nonsyndromic autosomal recessive RP. ^2,8,11,12 All USH1 and -2 proteins are integrated in an Usher protein network in both the inner ear and the retina, with a central scaffolding role for the proteins whirlin and harmonin. ^6,13–15 These proteins contain multiple PDZ domains, homologous to domains identified in postsynaptic density protein 95 (PSD-95), disc large (Dlg), and zonula occludens-1 (ZO-1), which link them to the intracellular moieties of transmembrane Usher proteins with a C-terminal PDZ-binding motif (PBM). The USH1 and -2 proteins are prominently located in the ciliary transition zone (connecting cilium) and periciliary region of the photoreceptor cells, rendering Usher syndrome a ciliopathy. In addition, most of the proteins localize to the synapse of photoreceptor cells. ^{6–8,14–16} At these specific sites, the Usher proteins are suggested to function in protein transport, structural support, and organization of ion channels. ^6,8,14,17,18  

To identify novel members of the Usher protein complex and thereby obtain more clues on the function of whirlin and harmonin, we used a bioinformatic approach. This approach revealed the L-type calcium channel pore–forming subunit Ca_v1.3 (α_1D) as a potential new member of the Usher protein complex. It is a transmembrane protein that is predicted to contain four transmembrane and four ion transport domains. ¹⁹ Ca_v1.3 (α_1D)-deficient mice are deaf due to the complete absence of L-type currents in cochlear inner hair cells and degeneration of both the inner and outer hair cells. The mice do not show any signs of vestibular dysfunction or retinal degeneration. ²⁰ The zebrafish Ca_v1.3a (α_1D) mutant does exhibit, besides hearing loss, a vestibular (circling) phenotype, but no retinal degeneration has been described, although the latter may be explained by the presence of a second copy of the gene which is differentially expressed in the retina. ²¹  

In this study, we sought to identify a specific PDZ–PBM-based interaction between whirlin and Ca_v1.3 (α_1D) and their co-localization in different layers of the retina, including the outer limiting membrane (OLM), the outer plexiform layer (OPL), and the region of the connecting cilium. We suggest that in the retina, the interaction of whirlin with Ca_v1.3 (α_1D) contributes to the organization of voltage-gated L-type calcium channels. 

Materials and Methods

Bioinformatics

Swiss-Prot protein sequences were downloaded from UniProt (www.uniprot.org/ ²² ) in FASTA format, and proteins matching the C-terminal class I PDZ-binding motif (PBM; [ST].[VIL]$) (http://elm.eu.org/elmPages/LIG_PDZ_1.html/ ELM [Eukaroytic Linear Motif] Functional Sites in Proteins, provided in the public domain by a consortium funded by the European Union) where extracted. The resultant dataset was filtered for proteins associated with ciliary function, by using the Ciliary Proteome database ver. 3.0 (www.ciliaproteome.org/ ²³ ), study selection: all except for Liu et al. 2007, database type: all; cutoff E-value: 30. Subsequently, a filter for proteins that are encoded by genes expressed in the eye or ear (UniGene; http://www.ncbi.nlm.nih.gov/UniGene; provided in the public domain by the National Center for Biotechnology Information, Bethesda, MD) was applied. A final selection of candidate Usher proteins was made by selecting for homologues of mouse mutant proteins that are associated with the vision/eye (MP:0005391) or the hearing/vestibular/ear (MP:0005377) phenotype, as contained in the Mouse Genome Informatics database (www.informatics.jax.org/ provided in the public domain by The Jackson Laboratory, Bar Harbor, ME). 

Animals

The Wistar rats and C57BL6 JOlaHsD mice (Harlan, Horst, The Netherlands) used in this study were housed in standard cages and received water and food ad libitum. All experiments were conducted according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and international and institutional guidelines. 

Cloning

Human brain cDNA (Marathon; Clontech, Palo Alto, CA) was used as a template to amplify the cDNA encoding part of the C-terminal intracellular domain of CACNA1D (amino acids [aa] 1938-2181). The aa numbers are according to the following GenBank entries: Ca_v1.3 (α_1D) NP_000711, whirlin NP_056219, harmonin NP_710142, and Plk1 NP_005021 (http://www.ncbi.nlm.nih.gov/Genbank; provided in the public domain by the National Center for Biotechnology Information, Bethesda, MD). All constructs were generated with commercial cloning technology (Gateway; Invitrogen, Carlsbad, CA), according to the manufacturer's instructions. 

Yeast Two-Hybrid Analysis

To test whether there is an interaction between Ca_v1.3 (α_1D) and whirlin, a Gal4-based yeast two-hybrid system (HybriZAP; Stratagene, La Jolla, CA) was used according to methods previously described. ^15,24  

Antibodies

The antibodies against whirlin and centrins have been described. ^14,15,25 Anti-HA and anti-Ca_v1.3 were derived from Sigma-Aldrich (Munich, Germany). Anti-GFP (Roche, Mannheim, Germany) was used to detect eCFP-tagged proteins. As secondary antibodies, goat-anti-guinea pig Alexa Fluor 488, goat-anti-rat Alexa Fluor 568, IRDye800 goat-anti-guinea pig IgG, and IRDye800 goat-anti-mouse were used (all from Molecular Probes-Invitrogen, Carlsbad, CA). 

GST Pull-Down

The GST-fusion proteins were produced by transforming Escherichia coli BL21-DE3 with plasmids pDEST15-whirlin PDZ1 (aa 138-233), PDZ2 (aa 279-360), PDZ1+2 (aa 138-360), PDZ3 (aa 819-907), pDEST15-Ca_v1.3 (aa 1938-2181), or pDEST15-Ca_v1.3ΔPBM (aa 1938-2175), according to methods previously described. ¹⁵ HA-tagged whirlin full-length (aa 1-907), eCFP-Ca_v1.3 (aa 1938-2181), and eCFP-Ca_v1.3ΔPBM (aa 1938-2175) were produced by transfecting COS-1 cells with pcDNA3-HA-Whirlin fl, pDEST501-Ca_v1.3, and pDEST501-Ca_v1.3ΔPBM, respectively, using a transfection reagent (Effectene; Qiagen, Hildene, Germany), according to the manufacturer's instruction. The follow-up of the experiment is described elsewhere. ¹⁵  

Co-immunoprecipitation in COS-1 Cells

Full-length eCFP-tagged whirlin (aa 1-907) was expressed by using the expression vector pDest501. HA-tagged Ca_v1.3 (aa 1938-2181) and Plk1 (aa 1-603) were expressed by using the pcDNA3-HA/Dest expression vector. COS-1 cells were transfected (Effectene; Qiagen) according to the manufacturer's instructions. Twenty-four hours after transfection, the cells were washed with phosphate-buffered saline (PBS) and subsequently lysed on ice in IP lysis buffer. ¹⁵ eCFP-tagged whirlin was immunoprecipitated from cleared lysates overnight at 4°C by using the anti-whirlin antibody. ¹⁵ HA-tagged Ca_v1.3 and Plk1 were immunoprecipitated by using anti-HA monoclonal antibody (Sigma-Aldrich) and protein A/G PLUS-Sepharose beads (Santa Cruz Biotechnology, Santa Cruz, CA). The protein complexes were washed four times with IP lysis buffer and subsequently analyzed by Western blot analysis (Odyssey Infrared Imaging Systemp; LI-COR, Lincoln, NE). 

Immunohistochemistry in Rat Retina

Unfixed eyes of 20-day-old (P20) Wistar rats were isolated and frozen in melting isopentane. Seven-micrometer-thick cryosections were treated and immunolabeled as described previously. ¹⁵  

Digoxigenin Labeling of cRNA In Situ Hybridization Probes

A probe corresponding to nucleotides 556-1673 (GenBank NM_001083616), which recognizes mouse Cacna1d transcripts was generated from mouse retina cDNA (Marathon; Clontech), with 5′-gcgaacgaggcaaactatg-3′ used as a forward primer and 5′-ccaaggatgatcagactaac-3′ used as a reversed primer. The labeling was performed as previously described. ¹⁵  

RNA In Situ Hybridization

C57Bl6 Jlco mouse embryos were collected at various embryonic stages (E12.5-E18.5), as well as postnatal day (p)7 (P7) and P90 mouse eyes. RNA in situ hybridization was performed as previously described. ¹⁵ To increase structural detail, several slides were incubated in nuclear Fast red (Sigma-Aldrich) for 10 seconds before embedding in rapid mounting medium (Entellan; ProSciTech, Kirwan, QLD, Australia). Images were recorded with a microscope (Axioskop2; Carl Zeiss Meditec, Inc., Oberkochen, Germany) equipped with a 3CCD color video camera (HAD DXC-950P; Sony, Tokyo, Japan). 

Pre-embedding Immunoelectron Microscopy

Vibratome sections of mouse retina were stained by a 100× diluted antibody against Ca_v1.3 (α_1D) and visualized by a secondary antibody (Vectastain ABC-Kit; Vector Laboratories, Peterborough, UK). After fixation with 0.5% OsO₄, the specimens were embedded in araldite, and ultrathin sections were analyzed with a transmission electron microscope (Tecnai 12; FEI, Hillsboro, OR). The procedure is described in detail elsewhere. ¹⁷  

Results

Identification of Candidate Members of the Usher Proteome Complex

To obtain more clues to the function of whirlin and harmonin, we used a bioinformatic approach to identify novel members of the Usher protein complex. We developed a script to select for human proteins from the Swiss-Prot database (www.uniprot.org) that match the class I PBM, as many of the interactions in the Usher protein network are PDZ–PBM based. ^6,15 Of the resultant 955 proteins, 147 are present in the ciliary proteome database. ²⁶ As our candidate Usher genes should be expressed in the eye or inner ear, a filter for expression in eye or ear (www.ncbi.nlm.nih.gov/UniGene/) yielded 117 genes. We then used the Mouse Genome Informatics (MGI) database (www.informatics.jax.org) to restrict our selection of candidate Usher genes to genes that are associated with the vision/eye (MP:0005391) or the hearing/vestibular/ear phenotypes (MP:0005377) in mice. This query yielded 18 genes, including the three of the known Usher genes CDH23, PCDH15, and USH2A, which validates our approach. Of these 18 genes, we selected the gene encoding the L-type calcium channel pore–forming subunit Ca_v1.3 (α_1D) as a very promising candidate Usher gene for further analysis, since the phenotype of early hair cell degeneration in the cognate mouse mutant ²⁰ is most similar to the phenotypes of the known Usher mouse mutants (reviewed in Ref. 27). 

Interaction of Ca_v1.3 (α_1D) and Whirlin

A yeast two-hybrid assay was used to assess whether the PDZ domains of whirlin and harmonin associate with part (aa 1938-2181) of the intracellular C-terminal tail of Ca_v1.3 (α_1D), which contains a class I PBM (aa 2178-2181). We identified an interaction between the cytoplasmic region of Ca_v1.3 (α_1D) and the N-terminal two PDZ domains of whirlin (Fig. 1A). Weak associations of Ca_v1.3 (α_1D) with PDZ1 and -3 of harmonin were suggested by limited growth of the yeast colonies. No interaction between the Ca_v1.3 (α_1D) C terminus and whirlin PDZ3 or harmonin PDZ2 was detected (Fig. 1A). 

Figure 1.

 

Validation of Ca_v1.3 (α_1D)–whirlin interaction. (A) Yeast two-hybrid assays were performed with a part of the Ca_v1.3 (α_1D) C terminus fused to the activation domain (AD) and whirlin PDZ1, PDZ2, PDZ1+2, and PDZ3 or harmonin PDZ1, PDZ2, and PDZ3 fused to the DNA-binding domain (BD) of the GAL4 reporter gene, respectively. Ca_v1.3 (α_1D) interacts with the PDZ1 and/or PDZ2 domains, but not with the PDZ3 domain of whirlin. Ca_v1.3 (α_1D) interacts with the PDZ1 and PDZ3, but not with the PDZ2 domain of harmonin, but this interaction was weak (+) compared with the Ca_v1.3 (α_1D)–whirlin (++) interaction. (B) GST pull-down assays showing that HA-tagged whirlin was efficiently pulled down by GST- Ca_v1.3 (α_1D) C terminus, but not by Ca_v1.3 (α_1D)ΔPBM or GST alone, as detected by an anti-HA antibody. The first lane shows 2% of the input of COS-1 cell lysate. (C) GST pull-down assays, showing that eCFP-tagged Ca_v1.3 (α_1D), detected by an anti-GFP antibody, was efficiently pulled down by GST-whirlin PDZ1, PDZ2, and PDZ1+2, but not by GST-whirlin PDZ3 or GST alone. The first lane shows 1.5% of the input eCFP- Ca_v1.3 (α_1D) protein lysate. (D) Co-immunoprecipitation assay from COS-1 cell lysates, showing that eCFP-whirlin co-immunoprecipitated with the HA-tagged C-terminal region of Ca_v1.3 (α_1D), but not with the HA-tagged protein Plk1. (E) GST pull-down assay, showing that HA-tagged harmonin not was pulled down by GST- Ca_v1.3 (α_1D) C terminus with the PBM or GST alone, as detected by an anti-HA antibody. The first lane shows 2% of the input of COS-1 cell lysate.

The association of whirlin and Ca_v1.3 (α_1D) was confirmed by using a glutathione S-transferase (GST) pull-down assay. Full-length HA-tagged whirlin was efficiently pulled down from COS-1 cell lysates by the GST-fused C-terminal part of Ca_v1.3 (α_1D), but not by GST alone (Fig. 1B). Deletion of the predicted C-terminal class I PBM in Ca_v1.3 (α_1D) disrupts the binding, indicating that this interaction is indeed based on a PDZ–PBM association (Fig. 1B). 

To confirm that specifically the PDZ domains 1 and 2 of whirlin are involved in this interaction, we expressed the domains separately as GST fusion proteins to pull down the enhanced cyan fluorescent protein (eCFP)–tagged cytoplasmic tail of Ca_v1.3 (α_1D). As shown in Figure 1C, whirlin PDZ1, PDZ2, and a peptide containing both PDZ domains were able to bind to the C-terminal part of Ca_v1.3 (α_1D), but not PDZ3 or GST alone. 

As a further validation of the interaction in vivo, a co-immunoprecipitation assay from COS-1 cells was performed, which showed that full-length eCFP-tagged whirlin specifically co-immunoprecipitates with the HA-tagged C terminus of Ca_v1.3 (α_1D), but not with the unrelated HA-tagged protein Plk1 (Fig. 1D). In contrast, HA-tagged harmonin was not pulled down from COS-1 cell lysates with the GST-fused C-terminal part of Ca_v1.3 (α_1D) (Fig. 1E). Therefore, the interaction between harmonin and Ca_v1.3 (α_1D) was not further investigated. 

Ca_v1.3 (α_1D) and Whirlin Co-localize in the Retina

To investigate whether Ca_v1.3 (α_1D) and whirlin co-localize in the retina, we co-immunostained retinal cryosections with antibodies against whirlin (green; Fig. 2A) and Ca_v1.3 (α_1D) (red; Fig. 2B), revealing expression and co-localization of these proteins (Fig. 2C) at the region of the connecting cilium, the OLM, and the OPL. Immunostaining of Ca_v1.3 (α_1D) was also observed at the inner plexiform layer (IPL) (Figs. 2B, 2C). Double immunostaining of Ca_v1.3 (α_1D) and centrins, markers for the accessory centriole, basal body, and connecting cilium, ²⁵ further confirmed the localization of Ca_v1.3 (α_1D) in this region, since a (partial) co-localization was observed (Figs. 2D–F). 

Figure 2.

 

Whirlin and Ca_v1.3 (α_1D) co-localize in rat photoreceptor cells. Subcellular localization of whirlin and Ca_v1.3 (α_1D) in retina cryosections of adult (P20) rat. Immunostaining with (A) anti-whirlin (green) and (B) anti-Ca_v1.3 (α_1D) (red) and (C) an overlay (yellow) indicate co-localization at the region of the connecting cilium (CC; open arrowheads), the OLM (filled arrowheads), and the OPL. (D–F) Immunofluorescence with anti-Ca_v1.3 (α_1D) (green) and anti-pan centrin (red) as a marker for the connecting cilium, centriole, and basal body confirmed the localization in this region, since partial co-localization (yellow) was observed.

Recently, the subcellular localization of whirlin in the mouse photoreceptor was determined by immunoelectron microscopy. ^6,17 Here, we present the results for Ca_v1.3 (α_1D). Pre-embedding labeling with an antibody directed against Ca_v1.3 (α_1D) revealed its localization in the collarlike extension of the apical inner segment (Figs. 3A, 3B), at the basal body complex (Figs. 3B, 3C), in the connecting cilium (Figs. 3B–D), in the basal region of the outer segments (Figs. 3A–C), and at the synapse (Figs. 3E, 3F) of the mouse photoreceptor cells. As whirlin was also detected at these exact subcellular sites, with the exception of the outer segments and the synapse, ^6,17 these results confirm the co-localization of Ca_v1.3 (α_1D) and whirlin. 

Figure 3.

 

Localization of Ca_v1.3 (α_1D) by immunoelectron microscopy. Micrographs of anti-Ca_v1.3 (α_1D) labeling in longitudinal (A–C, E–F) and cross (D) sections of rod mouse photoreceptor cells. Ca_v1.3 (α_1D) was detected in the collarlike extension of the apical inner segment (CE; A, B), at the basal body complex (BB; B, C), and in the connecting cilium (CC; B–D). Ca_v1.3 (α_1D) was also detected in the base of the outer segment (OS; A–C) and at the synapses (S) of mouse photoreceptor (E, F). IS, inner segment. Scale bars: (A–C) 0.25 μm; (D, F) 0.1 μm; (E) 0.5 μm.

Cacna1d Expression during Murine Development

The expression of Cacna1d in development was assessed by RNA in situ hybridization using mouse embryos of gestational days E12.5 to E18.5 and eyes of mice at P7 and P90. Cacna1d was widely expressed during murine embryonic development. From E12.5 to E16.5, expression was observed in the eye, inner ear, thalamus, neopallial cortex, midbrain, choroid plexus of the fourth ventricle, spinal cord, jaw, olfactory bulb, olfactory epithelium, lung, tongue, trigeminal (V) ganglion, duodenum, umbilical cord, venous heart region, kidney, adrenal gland, and stomach wall (Fig. 4A, and data not shown). At E18.5, the expression became more restricted and was detected in the lung, kidney, spinal cord, and olfactory epithelium, and intense staining was observed in the brain, inner ear (data not shown), and eye (Fig. 4E). In the developing retina, Cacna1d was transcribed from E12.5 onward (Figs. 4B–H). At E12.5, there was a strong signal in the complete neuroblastic layer (Fig. 4B). The signal at E14.5 was comparable with that at E12.5, but no staining was detected at the outermost cells. At E16.5, expression was detected at the inner neuroblastic layer of the retina, with the highest signal intensity in the area around the developing lens (Fig. 4D). At E18.5, the Cacna1d expression could be clearly distinguished in a subset of the ganglion cells and the inner nuclear layer (INL), and low-intensity staining was present in the neuroepithelium surrounding these layers (Fig. 4E). In juvenile P7 and P90 eyes, Cacna1d transcripts were detected in the INL, the outer nuclear layer (ONL), and the inner segments. A high-intensity signal, suggestive of a high level of Cacna1d expression, was seen in a subset of the ganglion cells and in the INL (Figs. 4F–H). In the developing inner ear, distinct expression was observed from E14.5 onward, which became more pronounced at E16.5 (Fig. 4I). At higher magnification, we show the expression of Cacna1d in the vestibular system and cochlea of the E16.5 inner ear (Fig. 4J). To obtain more structural detail for the inner ear, we counterstained several sections of an E16.5 embryo with nuclear Fast red, revealing that Cacna1d expression was situated in the inner hair cells (IHCs) of the cochlea and in the spiral ganglion cells (Figs. 4L, 4M). Also, expression in the developing sensory cells of the macula of the utricle and the cristae ampullaris of the semicircular canals was detected (Figs. 4N, 4O). Hybridizations with a sense cRNA probe revealed no signals, indicating the specificity of the antisense cRNA probe used in these experiments (Fig. 4K). 

Figure 4.

 

RNA in situ hybridization of Cacna1d mRNA in embryonic and adult mouse. Cacna1d was widely expressed during development (E12.5-E16.5), with most intense signals in the following structures (indicated by numbers and arrowheads). (A) Neopallial cortex (1), midbrain (2), lung (3), adrenal gland (4), spinal cord (5), stomach wall (6), tongue (7), and olfactory epithelium (8). Expression was also observed in the kidney, choroid plexus of the fourth ventricle, the lower and upper jaws, olfactory bulb, trigeminal (V) ganglion, duodenum, thalamus, umbilical cord, and venous heart region (data not shown). (B–H) A strong signal for Cacna1d was observed in the eye. Embryonic development of the eye at E12.5 (B) and E14.5 (C), in which expression was observed in the whole neuroblastic layer of the retina. At E16.5 (D), Cacna1d was expressed in the inner neuroblastic layer of the retina, and at E18.5 (E) expression was observed in a subset of the cells and the inner nuclear layer (INL). Expression of Cacna1d was maintained at postnatal days 7 (F) and 90 (G, H). A strong signal, indicative of a high level of expression, was observed in the INL, and a subset of the GCL. Furthermore, expression was seen in the ONL and inner segments (IS; H). (I, J, L–O) From E14.5 on, Cacna1d expression in the developing inner ear was observed and became more pronounced at E16.5 (I, J, L–O). Sections of the developing inner ear at E16.5 with Cacna1d expression in the inner hair cell (L, 1; M, 1), the spiral ganglion cells (M, 2), developing sensory cells of the macula of the utricle (N, 3), and in the crista ampullaris of the semicircular canals (O, 4). To increase structural detail, several sections were counterstained with nuclear Fast red (L–O). (K) Sections hybridized with the sense Cacna1d cRNA probe revealed no staining, indicating the specificity of antisense cRNA probe used in these experiments.

Discussion

In this study, we demonstrated that the C terminus of the calcium channel subunit Ca_v1.3 (α_1D) interacts with PDZ1 and -2 of whirlin and that these proteins co-localize at distinct sites in photoreceptor cells. A functional calcium channel is a heteromultimeric protein complex, composed of a pore-forming α₁ subunit, such as Ca_v1.3 (α_1D), and the auxiliary subunits β, γ, and α₂δ. Ten α₁-subunit pore–forming isoforms (α_1a-α_1I, and α_1S) of voltage-activated calcium channels are described. ²⁸ The α₁ subunit imparts most of the conductive properties of the channel, whereas the accessory subunits modulate calcium currents and channel activation/inactivation kinetics. ^29–31  

Ca_v1.3 (α_1D), encoded by the CACNA1D gene, is the most abundant isoform in hair cells and fulfills distinct physiological roles in the inner ear. ^20,32 In the retina of mouse, rat, and zebrafish, Ca_v1.3 (α_1D) mRNA has been observed in the ONL, the INL, and the ganglion cell layer (GCL). ^21,33–35 At the protein level, Ca_v1.3 (α_1D) has been detected in the Müller cells, the OPL, and photoreceptor cell bodies of the rat retina. ³⁵ In the salamander retina, immunoreactivity is observed in the GCL, Müller cells, the IPL, the OPL, and the photoreceptor inner segment. ^36,37  

During development, the murine Cacna1d gene is widely expressed (e.g., in the eye, the inner ear, and the CNS). Of note, Whrn expression has also been detected in these tissues. ¹⁵ We and others have shown that whirlin is connected to the dynamic Usher protein interactome, and we suggested that whirlin mediates multiple biological processes in the inner ear and the retina. ^6,15 Our findings indicate that through whirlin, Ca_v1.3 (α_1D), a voltage-gated calcium channel subunit, is connected to the Usher protein network, underlining the molecular diversity of this interactome. 

Cacna1d Expression during Murine Development

At early embryonic stages (E12.5-E14.5) Cacna1d expression was detected in the eye, the inner ear, the midbrain, spinal cord, tongue, lung, choroid plexus of the fourth ventricle, and kidney, in which Whrn RNA was also shown to be present. ¹⁵ This overlap in expression may indicate a function of both proteins in the same complex, even at these early embryonic stages in several tissues. In the developing inner ear Cacna1d expression was observed in the cochlear IHCs, the spiral ganglion cells, and the sensory cells of the vestibular part. Also, Whrn is expressed in these cells, although in the sensory epithelium of the cochlea it is only detectable from E18.5 onward. ¹⁵ In a recent study, RNA in situ hybridization was performed on wholemount preparations of mature mouse organs of Corti, and Cacna1d mRNA was detected in both IHCs and outer hair cells (OHCs), but mainly in the OHCs. ³⁸ We were able to show Cacna1d expression only in the IHCs. This discrepancy may be explained by the difference in age of the animals—embryonic stages in our study versus P19 in the study of Knirsch et al. ³⁸  

At early developmental stages of the retina, both Cacna1d (present data) and Whrn (described by van Wijk et al. ¹⁵ ) are expressed, although Whrn transcripts are mainly present in the innermost layers, whereas Cacna1d transcripts were detected in the entire developing retina. From E16.5 on, the retinal expression patterns of Cacna1d and Whrn have a higher similarity. Cacna1d expression in the adult mouse retina was observed in all three nuclear layers (GCL, INL, ONL), which is consistent with the findings in a previous study. ³⁴ The difference in relative signal strength between the study of Xiao et al. ³⁴ and the data we present may be due to a difference in antibody concentration and colorimetric reaction time. In postnatal stages of retinal development and in adult retina, Cacna1d expression was most prominent in the INL, whereas Whrn expression was highest in the ONL. The overlap in Cacna1d and Whrn expression at several stages of mouse development and in several organs, although at an apparently different quantitative distribution, together with the interaction at the protein level, indicates that Ca_v1.3 (α_1D) and whirlin serve in the same processes of neuronal differentiation. 

Ca_v1.3 (α_1D) Localization and Function in the Photoreceptor

In mature photoreceptor cells, Ca_v1.3 (α_1D) and whirlin co-localize at regions where also Usher syndrome-associated proteins such as USH2A, SANS, and GPR98 (VLRG1) co-localize with whirlin. ^14,15,17 We previously suggested that whirlin may contribute to the organization of ion channels in the pre- and/or postsynaptic membranes of hair cells and photoreceptor cells. ^6,14,15 Our present data indicate that Ca_v1.3 (α_1D) is one of these ion channels at the OPL of the photoreceptor. In contrast to Ca_v1.3 (α_1D), whirlin was not observed in the IPL by immunohistochemistry, which indicates that whirlin does not play a role in Ca_v1.3 (α_1D) organization in the synapses of this layer. Expression of whirlin, synonymously named CIP98, was detected, not only in the synaptic region of the photoreceptor cells, but also in the synaptic regions of the cochlear sensory cells ¹⁵ and possibly in neurons in the brain. Whirlin interacts with calmodulin-dependent serine kinase (CASK), and both proteins are co-immunoprecipitated from brain extracts. The whirlin-CASK complex has been suggested to play a role in trafficking of synaptic vesicles for transmission. ³⁹ Of note, CASK is also detected at the OPL of the retina, ⁴⁰ and therefore whirlin may serve as an adaptor protein linking CASK and Ca_v1.3 (α_1D) in the synaptic regions of the retinal sensory cells. The complex may participate in the regulation of neurotransmission via the organization of Ca_v1.3 (α_1D) channels in the photoreceptor cell synapses. Also, in the inner ear, Ca_v1.3 (α_1D) channels are involved in the regulation of exocytosis of synaptic vesicles in IHCs and probably also in OHCs. ^38,41–43 Although whirlin was not detected in the synaptic region of IHCs, it has been found in the synaptic region of OHCs, ¹⁵ and therefore the whirlin–Ca_v1.3 (α_1D) interaction may play a role in synaptic transmission in these cells as well. However, to confirm this hypothesis, the localization of whirlin and its co-localization with Ca_v1.3 (α_1D) in the synaptic region of OHCs should be studied in more detail. 

The role of Ca_v1.3 (α_1D) and its possible interaction with whirlin at the basal body and in the connecting cilium remains elusive. It is thought that the basal body functions in the organization of transport of cytoplasmic and transmembrane proteins into and through the connecting cilium. Therefore, the presence of whirlin and Ca_v1.3 (α_1D) in the basal body may be explained by the fact that they are transported from there into the connecting cilium, where both proteins have been detected. In the connecting cilium, the whirlin-Ca_v1.3 (α_1D) interaction may function in the organization of the channels and thereby in the Ca²⁺ homeostasis in this region and the regulation of the Ca²⁺-dependent interaction of centrins and the visual G-protein transducin. ²⁵ The centrin-transducin interaction is thought to regulate the light-dependent translocation of transducin through the connecting cilium. ²⁵  

Recently, our group and others described the molecular and ultrastructural homology between the periciliary structures in mammalian and amphibian photoreceptor cells. In mammals, the membrane domain of the apical inner segment collar corresponds to the periciliary ridge complex (PRC) of the amphibian photoreceptor. ^17,44 Besides providing structural support of the connecting cilium, this region is thought to function in docking trans-Golgi–derived cargo vesicles containing proteins that are essential for outer segment formation and renewal and phototransduction. ^6,17,45,46 We previously identified whirlin as one of the major molecules of the PRC ¹⁷ and subsequently Mazelova et al. ⁴⁷ used an antibody against whirlin as a marker for the PRC, demonstrating the association of the SNARE proteins syntaxin 3 and SNAP25 with this region. ⁴⁷ Also, the interaction of whirlin and Ca_v1.3 (α_1D) may have a role in the transport to or organization of the calcium channel at the plasma membrane. The role of the Ca_v1.3 (α_1D) channel at these subcellular sites may be the regulation of docking and fusion of cargo vesicles, involving SNARE proteins, through mediation of the Ca²⁺ concentration, and Ca_v1.3 (α_1D) was shown to be in a complex with SNARE proteins. ^48,49 As is shown for synaptic vesicle fusion and exocytosis, the increase in Ca²⁺ concentration via Ca²⁺ influx through voltage-gated calcium channels initiates these processes. ^30,50 Of note, Ca_v1.3 (α_1D) was also detected in the pericuticular region of hair cells in the inner ear, ⁴¹ and in this region SNARE proteins were also detected. ⁵¹ This region is thought to be the site of vesicular trafficking, endocytosis, exocytosis, and membrane recycling for stereocilia repair ⁵² and therefore may have functional similarities to the periciliary region of photoreceptor cells. 

We detected Ca_v1.3 (α_1D) also at the base of the outer segments. At this site, the process of outer segment disc morphogenesis occurs. Rab and SNARE proteins, present in this region are suggested to contribute to this process by mediating membrane fusion of vesicles containing outer segment proteins. ^53,54 For vesicles loaded with rhodopsin, the targeting and fusion was indicated to be regulated by the protein SARA through its direct interaction with rhodopsin, syntaxin 3, and PI3P. ⁵³ The Ca²⁺-dependence of the process of membrane fusion, ^55–57 as discussed in the previous paragraph, may explain the presence of Ca_v1.3 (α_1D) in this region. 

CACNA1D in Disease

Voltage-gated L-type calcium channels contribute to retinal signal transmission. ⁵⁸ Recently, it has been shown that a mutation in the CACNA2D4 gene, which encodes one of the auxiliary subunits (α₂δ) of the L-type calcium channels, causes autosomal recessive progressive cone dystrophy. ⁵⁹ The gene encoding the L-type calcium channel α₁-subunit, CACNA1F, is associated with other types of retinal dysfunction—namely, incomplete X-linked congenital stationary night blindness (CSNB2) ^60,61 and X-linked cone-rod dystrophy (CORDX3). ⁶² The main cellular function of Ca_v1.4 (α_1F) is thought to be the mediation of neurotransmitter release from photoreceptor and bipolar cells, ⁶³ a function that could be assumed for Ca_v1.3 (α_1D) as well. Therefore, and because of the Ca_v1.3 (α_1D) expression in the developing retina, CACNA1D is a candidate gene for retinal dysfunction in humans. However, no mutations have been described in patients with retinal disorders, and animal models with defects in Cacna1d do not exhibit a retinal phenotype. Despite the absence of an obvious retinal phenotype in a Ca_v1.3^−/− mouse model, there is a reduced electroretinogram (ERG) light peak (LP) amplitude. ⁶⁴ The LP of the ERG reflects the depolarization of the basolateral plasma membrane of the retinal pigment epithelium (RPE), resulting from changes in the activity of one or more calcium-sensitive chloride channels. ⁶⁵ However, we did not observe immunostaining of Ca_v1.3 (α_1D) in the RPE (Fig. 2B). The lack of retinal phenotypes in the animal models may be due to compensation by other α-subunits. In zebrafish, Ca_v1.3b is one of the obvious candidates ²¹ and, in the mouse, it could be Ca_v1.4 (α_1F). ^60–62 The absence of a retinal dysfunction of the Ca_v1.3^−/−-defective mice does not rule out a retinal dysfunction in humans with defects in the orthologous gene, since several mouse mutants with defects in genes involved in Usher syndrome do not exhibit retinal degeneration, but only mild ERG abnormalities in some of them. ^66–69 For the whirler mouse with defects in the Whrn gene, no retinal phenotype has been reported. However, progressive retinal degeneration has been described in a Whrn–knockout mouse (Yang J, et al. IOVS 2008;49:ARVO E-Abstract 4405). Whether the Ca_v1.3 (α_1D) localization or function is disturbed in this mouse mutant remains to be determined. To the authors' knowledge, no locus for syndromic or nonsyndromic retinal dysfunction or deafness has been reported for the CACNA1D locus. The locus for Usher syndrome type 2b, which harbored the CACNA1D gene, has recently been withdrawn. ^70,71  

Footnotes

 Supported by the Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands; The British Retinitis Pigmentosa Society Grant GR552; and the Heinsius Houbolt Foundation.

Footnotes

 Disclosure: F.F.J. Kersten, None; E. van Wijk, None; J. van Reeuwijk, None; B. van der Zwaag, None; T. Märker, None; T.A. Peters, None; N. Katsanis, None; U. Wolfrum, None; J.E.E. Keunen, None; R. Roepman, None; H. Kremer, None