Abstract
Glutamate release from rod photoreceptors is dependent on a sustained calcium influx through L-type calcium channels. Missense mutations in the CACNA1F gene in patients with incomplete X-linked congenital stationary night blindness implicate the Cav1.4 calcium channel subtype. Here, we describe the functional and pharmacological properties of transiently expressed human Cav1.4 calcium channels. Cav1.4 is shown to encode a dihydropyridine-sensitive calcium channel with unusually slow inactivation kinetics that are not affected by either calcium ions or by coexpression of ancillary calcium channel β subunits. Additionally, the channel supports a large window current and activates near -40 mV in 2 mM external calcium, making Cav1.4 ideally suited for tonic calcium influx at typical photoreceptor resting potentials. Introduction of base pair changes associated with four incomplete X-linked congenital night blindness mutations showed that only the G369D alteration affected channel activation properties. Immunohistochemical analyses show that, in contrast with previous reports, Cav1.4 is widely distributed outside the retina, including in the immune system, thus suggesting a broader role in human physiology.
Introduction
Neurotransmitter release is critically dependent on calcium entry through voltage-gated calcium channels. In most central and peripheral nerve synapses, the release process is triggered by the transient activation and opening of N-type and P/Q-type calcium channels in response to membrane depolarizations arising from action potentials arriving at the presynaptic nerve terminal (Wheeler et al., 1994). This contrasts with synaptic transmission in photoreceptor cells that operate via graded changes in membrane potential (Knapp and Dowling, 1987), where these cells are depolarized in the absence of light stimuli and tonically secrete neurotransmitter triggered by calcium entry predominantly through L-type calcium channels (Schneeweis and Schapf, 1995; Nachman-Clewner et al., 1999). Light stimuli hyperpolarize the photoreceptor cell and, thus, terminate calcium entry via L-type channels. Molecular cloning has led to the definitive identification of four genes encoding distinct L-type calcium channels. Cav1.1 (α1S) L-type channels are found exclusively in skeletal muscle and serve predominantly as a voltage sensor for excitation contraction coupling (Beam et al., 1992). In contrast, Cav1.2 (α1C) and Cav1.3 (α1D) L-type calcium channels are expressed in a variety of tissues, including most neurons, smooth and cardiac muscle, and endocrine cells (Seino et al., 1992; Williams et al., 1992a; Hell et al., 1993). The Cav1.2 and Cav1.3 channels can be distinguished based on their distinct pharmacological properties, with Cav1.3 channels exhibiting both a reduced sensitivity to dihydropyridine (DHP) antagonists (Koschak et al., 2001; Xu and Lipscombe, 2001) and a moderate sensitivity to the N-type calcium channel blocker ω-conotoxin GVIA (Williams et al., 1992b). More recently, the CACNA1F (Cav1.4) gene in humans has been identified as being associated with patients with incomplete X-linked congenital stationary night blindness (CSNB2) (Bech-Hansen et al.,1998a,b; Strom et al., 1998), although little is known about the detailed biophysical and pharmacological properties of Cav1.4 channels. Moreover, the consequence of missense mutations linked to CSNB2 on channel function remains to be described.
Salamander rod and cone photoreceptor L-type calcium channels (Corey et al., 1984; Wilkinson and Barnes 1996; Kourennyi and Barnes, 2000), as well as L-type calcium channels expressed in mammalian Muller and photoreceptor cells (Yagi and Macleish, 1994; Puro et al., 1996; de la Villa et al., 1998; Protti and Llano, 1998; Taylor and Morgans, 1998; Berntson et al., 2003), exhibit similar biophysical properties, including relatively slow kinetics, a low affinity for DHP antagonists, as well as a weak sensitivity to ω-conotoxin GVIA. Although these properties are primarily consistent with Cav1.3 channels comprising at least part of L-type currents in photoreceptors, the observation that Cav1.3 antibodies do not label rod photoreceptors suggests that they express a distinct L-type channel isoform (Taylor and Morgans, 1998; Morgans, 1999). Considering that impairment of vision in CSNB2 patients is linked to mutations in Cav1.4 channels, the evidence suggests that rods may rely predominantly on Cav1.4 for signal transduction.
We report the functional characterization of Cav1.4 calcium channels as well as the analysis of four mutations associated with CSNB2. Functional expression of the Cav1.4 calcium channel in human embryonic kidney (HEK) tsA-201 cells reveals that Cav1.4 encodes a DHP-sensitive calcium channel that preferentially conducts barium over calcium but, unlike other high-voltage-activated calcium channels (Liang et al., 2003), shows no discernable calcium-dependent inactivation. The Cav1.4 channels first activate at relatively negative potentials and, unlike other high-threshold calcium channels, display unusually slow voltage-dependent inactivation characteristics irrespective of the type of calcium channel β subunit coexpressed. This, together with the occurrence of a large window current, makes Cav1.4 channels ideally suited for maintaining tonic release at photoreceptor synapses. Immunohistochemical localization of Cav1.4 reveals that, in contrast with previous reports (Bech-Hansen et al. 1998a,b), Cav1.4 expression is not only confined to the retina but is also detected at significant levels in the adrenal gland, bone marrow, spinal cord, muscle, and spleen. Moreover, Cav1.4 appears to be present at high levels in plasma cells and mast cells, suggesting a potential role of Cav1.4 in mediating immune responses.
Materials and Methods
Cloning of the human Cav1.4 calcium channel. To isolate the full-length cDNA, we used the nucleotides from exons 47 and 48 of the human gene sequence to identify two expressed sequence tags (ESTs) (GenBank accession numbers AA317815 and AA019975) from the National Center for Biotechnology Information database. These ESTs were then used to screen a human retinal cDNA library (a generous gift from Dr. Jeremy Nathans, Johns Hopkins University, Baltimore, MD). Screening of 200,000 pfu from the library resulted in isolation of two clones that encoded for nucleotides 2864-5871 of the channel (nucleotides numbered as to GenBank accession number NM_005183). To isolate the remainder of the clones (clones 1 and 2), oligonucleotides were made to the deduced cDNA sequence, and PCR was used to amplify and clone the remainder of the Cav1.4 full-length cDNA. Oligonucleotides used to amplify the remaining regions were synthesized to regions corresponding to nucleotides 1-19, 1653-1672, 1541-1559, and 3221-3239. PCR reaction conditions consisted of a 50 μl reaction with 10× PCR buffer, 20 pmol of each oligonucleotide, 10 mm dNTP, 2.5 U of HotStart TaqDNA polymerase (Qiagen, Hilden, Germany), and 1 μl of retinal library heated for 94°C (15 min), followed by 35 cycles at 95°C (30 sec), 55°C (1 min), and 74°C (2 min). To confirm the cDNA sequences, all PCR products and isolated library clones were sequenced. Sequencing oligonucleotides corresponded to nucleotides 721-728, 1389-1406, 2152-2169, 3005-3021, 3695-3712, 4432-4450, and 5265-5282 (GenBank accession number NM_005183). The 5′ end PCR products were ligated together using the unique NarI and ClaI sites, and the full-length clone was then assembled in pBluescript KS using the isolated cDNA library cDNA clone and ClaI and XbaI. For expression studies, the Cav1.4 cDNA clone was transferred to pcDNA3-ZEO using SalI and XbaI, and the DNA sequence was again confirmed. Full-length cDNA clones that contained the reported CSNB2 mutations were synthesized using the Quick-Change Site-Directed Mutagenesis kit according to the manufacturer's instructions, and all mutated DNA sequences were confirmed via direct DNA sequencing.
Transient expression of the Cav1.4 cDNA clone. Cells used for transient transfection (tsA-201) were maintained in standard DMEM supplemented with 10% FBS and 50 U/ml penicillin-streptomycin to 80% confluence. Cells were maintained at 37°C in a humidified environment of 5% CO2 and every 2-3 d were dissociated enzymatically with trypsin-EDTA and plated on 15 mm coverslips in 110 mm dishes. A standard calcium phosphate procedure was used to transiently transfect the cells. The human Cav1.4 cDNA clone (6 μg) was cotransfected with one of the rat β subunits (6 μg) (i.e., β2a, β1b, β3, or β4) and the rat α2-δ1 (6 μg) subunit. In addition, we cotransfected pIRES-enhanced green fluorescent protein (EGFP) (1 μg; Clontech, Cambridge, UK) as a marker plasmid. Transfected cells were incubated in the media overnight, then the media was exchanged and the plates were incubated at 29°C for 3 d before evaluation by the whole-cell patch-clamp technique.
Electrophysiological recordings. Immediately before recordings, individual coverslips were transferred to a 3 cm culture dish filled with recording solution consisting of (in mm): 20 BaCl2, 65 CsCl, 40 TEA-Cl, 1 MgCl2, 10 glucose, and 10 HEPES, pH 7.2 (adjusted with TEA-OH). In some experiments, 20 or 2 mm CaCl2 replaced 20 mm BaCl2. Pipettes (3-4 MΩ; BF150-86-15 borosilicate glass; Sutter Instruments, Novato, CA) were pulled on a Sutter P-87 microelectrode puller, fire-polished with an MF-830 microforge (Narishige, Tokyo, Japan), and filled with filtered (0.22 μm) intracellular recording solution consisting of (in mm): 108 cesium methane sulfonate, 4 MgCl2, 9 EGTA, and 9 HEPES, pH 7.2 (adjusted with CsOH). For experiments examining calcium-dependent inactivation, and for recordings in low external calcium, “low EGTA” internal recording solution consisting of (in mm) 118.5 cesium methane sulfonate, 1 MgCl2, 0.5 EGTA, 9 HEPES, and 4 MgATP, pH 7.4 (adjusted with CsOH), was used. Recordings were performed using an Axopatch 200B amplifier (Axon Instruments, Union City, CA) linked to a personal computer with either a Digidata 1322A or a Digidata 1200 interface. pClamp version 9.1 was used for data acquisition. Currents were filtered at 1 kHz and digitized at a minimum sampling frequency of 2 kHz. Series resistance was compensated by 80%. The voltage dependence of inactivation was assessed by holding cells at various holding potentials for 10 sec before application of a test depolarization to +20 mV. Current-voltage relationships were obtained by holding cells at -100 mV before stepping to various test potentials or by ramping membrane voltage from -100 to +50 mV over 150 msec. ω-Conotoxin GVIA (Sigma, St. Louis, MO) was dissolved in water at a stock concentration of 1 mm. Nifedipine and BayK 8644 (Sigma) were dissolved in dimethylsulfoxide at a 10 mm stock concentration. Drugs were diluted from these stocks into the final recording solution immediately before recording and washed onto the cells using a gravity-driven microperfusion system.
Analysis of electrophysiological data was performed using Clampfit software and SigmaPlot 2000 (Jandel Scientific, Chicago, IL). Inactivation curves were fitted with modified Boltzman relationships I = x + (1 - x)/(1 + exp(-z(Vh - V)/25.6)), where I is the peak current amplitude, x is the noninactivating fraction of current, V is the test potential, Vh is the half-inactivation potential, and z is a slope factor that reflects the effective gating charge. Whole-cell I-V relationships were fitted with the equation I = G(V - Erev)/(1 + exp((Va - V)/S)), where G is the maximum slope conductance, I is peak current amplitude, V is the test potential, Va is the half-activation potential, and S is a slope factor. Error bars reflect SEs, and numbers in parentheses reflect the number of experiments. Statistical analysis was based on paired and unpaired Student's t tests at the 0.05 significance level.
Tissue expression reverse transcription-PCR reaction. To detect low levels of Cav1.4 mRNA from different tissues, we used a total human RNA panel that contained RNA from 24 different tissues (Clontech) and a reverse transcription-PCR (RT-PCR) protocol. The reverse transcriptase reaction consisted of 1 μg of total RNA, 1× reverse transcriptase buffer, 10 mm dNTPs, 5 U of SuperScript reverse transcriptase (Invitrogen, San Diego, CA), and 10 pmol of oligonucleotide (5′-ctgagatgcccaagggctgc) at 42°C for 90 min. From this reaction, 10 μl was removed and added to a 50 μl PCR. The PCR condition consisted of 10 μl of reverse transcriptase reaction, 10× PCR buffer, 20 pmol of each oligonucleotide (5′caatgcctgctactgggc and 5′ctgagatgcccaagggctgc), 10 mm dNTP, and 2.5 U of HotStar TaqDNA polymerase (Qiagen) heated for 95°C (15 min), followed by 35 cycles of 94°C (30 sec), 55°C (1 min), and 74°C (2 min). To confirm the PCR products were legitimate Cav1.4-amplified products, the PCR products were electrophoresed through a 1.8% agarose gel and then subjected to Southern blotting with a Cav1.4-specific 32P-labeled oligonucleotide (5′-attgtacggtctgggcc). In addition, to verify all amplified bands were correct, all products were cloned into pGemT-easy vector (Promega, Madison, WI) and sequenced.
Antibody production and immunohistochemical staining. To examine Cav1.4 distribution, a polyclonal antibody was produced using a fusion protein unique to the Cav 1.4 calcium channel as determined with a GenBank Basic Local Alignment Search Tool search (aa 1658-1723 of the human Cav1.4 calcium channel carboxyl tail; GenBank accession number NM_005183). The portion of the channel carboxyl tail containing the fusion protein was cloned into the pCX vector and expressed using the Caulobactor expression system (Invitrogen) tagged with the Caulobactor RSA protein. The fusion peptide was injected into rabbits, whole serum was collected, and Cav1.4/RSA antibodies were IgA purified. Specificity of the Cav1.4 antibody was determined by staining HEK cells transfected with each of the known neuronal calcium channel α1 subunits (Cav1.2, Cav1.3, Cav1.4, Cav2.1, Cav2.2, Cav2.3, Cav3.1, Cav3.2, and Cav3.3) plus β2a and α2-δ1 in the case of the high voltage-activated (HVA) channel complexes. Specific staining was seen only in cells transfected with the Cav1.4 channel complex, and staining was abolished in the Cav1.4 transfected cells when the antibody was preabsorbed with the blocking fusion protein (data not shown). For immunohistochemical staining, HEK tsA-201 cells were transfected with the Cav1.4, β2a, and α2-δ1 antisera and then stained with the Cav1.4 antisera. Briefly, transfected cells were grown for 3 d on poly-lysine-coated coverslips. The coverslips were washed two times with PBS, fixed in 3% paraformaldehyde (10 min), then followed by another two washes in PBS. To permeabilize the cells, the coverslips were transferred to 100% methanol for 2 min and then washed four times with fresh PBS. To hybridize the antibody to the fixed cells, 0.1 μg/ml of the Cav1.4 antibody was added to PBS supplemented with 5% powdered milk and incubated at room temperature for at least 1 hr. The coverslips had the primary antibody removed, followed by a wash (three times) for 5 min in PBS-1% Triton X-100. The FITC-tagged secondary goat-anti rabbit antibody (Amersham) was added and hybridized 30 min in PBS supplemented with 5% powdered milk. To remove any excess secondary antibody, the coverslips were washed three times in PBS-1% Triton X-100 solution, air dried, and mounted on slides, and the images were captured with a confocal microscope.
For Western blotting, adult female rats were euthanized and decapitated, and various organs were removed and either used immediately or frozen in liquid nitrogen for later use. Various organs were ground and homogenized in 1 ml of assay buffer (150 mm NaCl, 1%NP-40, 0.5% deoxycholate, 0.1% SDS, 50 mm Tris, ph 7.5, and a mixture of protease inhibitors). The solution was incubated on ice for 1 hr and centrifuged for 15 min (4°C) at 15,000 × g, and the supernatant was collected. Total protein concentration was determined with the modified Lowry assay, and 10 μg of protein was combined with 10 μl of 2× SDS loading buffer, boiled for 10 min, and loaded onto a 5% acrylamide SDS-PAGE minigel. Proteins were electroblotted onto Hybond ECL nitrocellulose membrane (Amersham), air dried, and prehybridized overnight in 5% powdered milk and 0.1% Tween 20 in PBS. Membranes were washed three times with PBS-0.1% Tween 20 (PBST), followed by the addition of anti-Cav1.4 antibody (1:10,000) in PBST with 5% powdered milk. Hybridization of the Cav1.4 antibody overnight at 4°C in PBST in 5% powdered milk was followed by three washes of PBST. A one hour incubation with the anti-rabbit antibody in PBST in 5% powdered milk was followed by three washes in PBST, and the antibody was detected according to the ECL+ kit protocol (Amersham). Western blots were repeated in triplicate, using 12 tissues each isolated from three separate rats.
Immunoperoxidase staining was performed on archival formalin-fixed, paraffin-embedded human tissue obtained from Vancouver Hospital and Health Sciences Centre. Tissues assessed for Cav1.4 calcium channel expression included retina, a retinoblastoma tumor, lymph node, thymus, spleen, and cervix. Eccentricity of the retina section could not be determined because the sectioned globe showed complete detachment of the retina by retinoblastoma tumor. Tissue sections were cut at 4 μm, placed on silanized glass slides, and baked at 60°C for 1 hr. They were deparaffinized in xylene and graded ethanol solutions, then rehydrated in double-distilled water. Antigen retrieval was performed by placing the slides in a Coplin jar filled with 10 mm citrate buffer, pH 6.0, and heating in a steamer at 60-90°C for 30-90 min. The slides were cooled to room temperature and rinsed in PBS, pH 7.2-7.4, then endogenous peroxidase activity was blocked by incubation in a 3% (v/v) aqueous hydrogen peroxide for 30 min. Before immunostaining, blocking was performed by incubating the sections with 3% BSA for 30 min. The primary antibody (rabbit polyclonal Cav1.4) was applied at a dilution of 1:10,000, and incubation was performed in a sealed immunochamber at 4°C overnight. The antigen was detected by incubation for 25 min with a universal biotinylated secondary antibody labeled streptavidin-biotin-HRP system; Dako, Carpinteria, CA), followed by a 25 min incubation with streptavidin-HRP (LSAB-HRP system; Dako). Nova Red (Vector Laboratories, Burlingame, CA) was used as the chromagen, and hematoxylin was used as the counterstain.
Results
The CACNA1F gene encodes a DHP-sensitive calcium channel with unique biophysical properties
Through a combination of library screening and RT-PCR, we isolated a cDNA clone from a human retinal cDNA library that encodes for a 1971 amino acid Cav1.4 calcium channel. Compared with the previously reported CACNA1F sequence (GenBank accession number AF15290), our Cav1.4 cDNA differed in several aspects: (1) there were four amino acid substitutions (K58R, E380G, R1282H, and I1306V); (2) there was an 11 amino acid insertion (Ser-Met-Ala-Glu-Glu-Gly-Arg-Ala-Gly-His-Arg) in the domain I-II linker region attributable to sliding of the intron/exon boundary by 33 nucleotides into the intron at the end of exon 9; and (3) because of exon skipping of exon 32, there was the deletion of seven residues (Asp-Gly-Gly-His-Leu-Gly-Glu) (Fig. 1).
To determine the functional characteristics of Cav1.4, the full-length cDNA was transiently expressed in HEK 293 tsA-201 cells. To confirm protein expression and localization, an antibody to the human Cav1.4 channel was designed against a unique stretch of residues (aa 1658-1723) contained within the C-terminal region. As shown in Figure 2, transfection of tsA-201 cells with Cav1.4 (plus ancillary β1b and α2-δ1 subunits) resulted in expression of the channel in the plasma membrane. In contrast, nontransfected cells did not exhibit any Cav1.4 immunoreactivity, indicating that tsA-201 cells do not endogenously express Cav1.4 protein.
To evaluate the functional properties of the human Cav1.4 channel, the Cav1.4 cDNA was transiently expressed in tsA-201 cells with β2a and α2-δ1 subunits plus an EGFP selection marker, and the characteristics of the channel were evaluated by whole-cell patch clamp. The β2a subunit was used initially because of reports that knock-out of this calcium channel subunit, but not that of other calcium channel β subunit isoforms, affects Cav1.4 channel distribution in the outer plexiform layer of the retina and results in abnormal b waves in ERGs (Ball et al., 2002). Unless otherwise noted, electrophysiology data were obtained using an intracellular recording solution with 9 mm EGTA. When using 20 mm barium as the charge carrier, expression of Cav1.4 resulted in robust, noninactivating inward currents that first activated near -30 mV and peaked at ∼+15 mV (Fig. 3A,B). The half-activation potential obtained from fits to individual whole-cell current-voltage relationships was -2.45 ± 0.54 mV (n = 75), comparable with that observed with Cav1.2 L-type calcium channels under similar recording conditions (Stotz et al., 2000). In contrast, the half-inactivation potential, Vh, obtained from fits to individual inactivation curves was -9.3 ± 2.8 mV (n = 13), which is ∼15 mV more positive than that observed with Cav1.2 channels (Olcese et al., 1994; Stotz et al., 2000). The significantly more depolarized half-inactivation potential results in an unusually large window current (Fig. 3C) that spans a voltage range of >40 mV, allowing the channel to be tonically active at these membrane potentials.
To test whether the Cav1.4 channel would likely to be functional at typical resting potentials of rod photoreceptors, 20 mm external barium recording solution was replaced by 2 mm calcium recording solution, and a low (0.5 mm) EGTA internal recording solution with ATP was used (see Materials and Methods). I-V relationships were obtained using a ramp protocol rather than a step protocol for convenience. As shown in Figure 3D, an ensemble of five ramp IV curves reveals that Cav1.4 channels carry a small inward current at membrane potentials as negative as -40 mV. The average half-activation potential obtained from the five cells was -9.5 ± 3.4 mV, and the slope factor, S, amounted to 8.42 ± 0.82 mV. The average current size for the five cells tested in 2 mm calcium was -80.7 ± 35.9 pA at peak. Taken together, these data suggest that transiently expressed Cav1.4 calcium channels are capable of mediating calcium influx at typical photoreceptor resting potentials.
Figure 3E examines the effects of coexpression of various types of calcium channel β subunits on the biophysical properties of Cav1.4 (bathed in 20 mm external barium). Interestingly, the calcium channel β subunit subtype mediated only relatively minor (∼5 mV) changes in half-inactivation and half-activation potentials. For β1b and β3, both Vh and Va shifted into the same direction, thus predicting a similar size of window current as that seen with β2a. For β4, there was a somewhat larger separation between Va and Vh, predicting less window current and, thus, perhaps slightly less tonic activation. Most surprisingly, however, the time constant for inactivation did not significantly depend on the nature of the β subunit coexpressed at any of the test potentials examined (ANOVA; 0.5 > p > 0.2). In each case, the time constant for inactivation was found to be in the range of several seconds [i.e., τinactivation at +10 mV β1b = 9.90 ± 0.73 sec (n = 5), β2a = 9.79 ± 1.48 sec (n = 11), β3 = 15.41 ± 3.81 sec (n = 5), β4 = 12.34 ± 1.68 sec (n = 9)]. This is, to our knowledge, the first incidence in which HVA calcium channel kinetics are not significantly regulated by the nature of calcium channel β subunits. In addition, these data imply that the slow inactivation properties illustrated in Figure 3A is not simply attributable to the presence of a β subunit that is known to generally antagonize the voltage-dependent inactivation of all other known HVA calcium channels (Olcese et al., 1994). Nonetheless, the observation that the time course of inactivation was slow irrespective of the type of β subunit coexpressed is consistent with the notion that in photoreceptors the calcium channels would need to be open for prolonged periods to allow for the tonic secretion of transmitter in the dark.
Figure 4A illustrates the effects of calcium ions on Cav1.4 channel activity. In the presence of low EGTA (0.5 mm) internal solution, switching from 20 mm Ba to 20 mm Ca resulted in a reduction in peak current amplitude. Strikingly, however, no calcium-dependent inactivation was apparent. The fraction of current remaining (R800) after an 800 msec depolarization to +20 mV did not differ significantly in barium and in calcium in seven cells examined (Fig. 4A, inset; paired t test). No significant decrease in R800 between barium and calcium was observed for any of the other test potentials examined (0, 10, or 30 mV; data not shown). This result contrasts with recent findings obtained with N-type and R-type calcium channels (Liang et al., 2003) and makes Cav1.4 the only HVA calcium channel that does not display calcium-dependent inactivation. Moreover, these data indicate that even in the presence of calcium ions, Cav1.4 channels are capable of sustained openings.
Figure 4B shows that the Cav1.4 channel was sensitive to low concentrations of the DHP antagonist nifedipine, such that at a holding potential of -100 mV, block occurred with an IC50 value of 944 ± 127 nm (n = 10). Holding the cells at a more depolarized membrane potential resulted in a small, but statistically significant, increase in the amount of block mediated by 300 nm nifedipine from 35.4 ± 5.5% (n = 11) to 43.4 ± 5.3% (n = 4, p = 0.01), consistent with the notion that DHP block of L-type channels is state dependent. Application of the DHP agonist BayK 8644 resulted in a pronounced upregulation of current activity (Fig. 4C), increasing the maximum slope conductance fourfold from 2.69 ± 0.37 to 9.48 ± 1.63 nS (n = 9) and shifting the half-activation potential from -1.28 ± 1.52 to -12.03 ± 1.87 mV (n = 9). Overall, the Cav1.4 channel indeed behaves pharmacologically similar to other native and cloned L-type calcium channels. Figure 4D summarizes other aspects of the pharmacological profile of Cav1.4. The channel was mildly inhibited by application of 3 μm ω-conotoxin GVIA (∼20% inhibition), consistent with findings in salamander rod photoreceptors (Wilkinson and Barnes, 1996). As expected for an HVA calcium channel, 10 μm cadmium completely inhibited current activity within 15 sec of application (Fig. 4D).
Effects of mutations associated with night blindness on Cav1.4 function
A number of mutations in the CACNA1F gene have been associated with CSNB2 (Boycott et al., 2000). Most of the mutations result in premature stop codons in the four major transmembrane domains or intracellular linker regions and would be predicted to result in the complete loss of Cav1.4 channel function in rod photoreceptors. Additionally, a number of patients also seem to carry missense mutations in the CACNA1F gene including substitution of glycine residues found in the IS6 and IIS5 segments and an alanine in the domain IIIS1-S2 linker region to aspartic acid (Boycott et al., 2000). To determine how the various mutations might contribute the clinical phenotype, we created three CSNB2 missense mutations in distinct regions in the Cav1.4 channel. In addition, we generated a CSNB2 mutation that results in the replacement of a tryptophan residue at the beginning of the carboxyl terminal tail region with a premature stop codon. Interestingly, two of the missense mutations (G674D and A928D) and the W1459 premature stop in the carboxyl tail appeared to exert no detectable changes in the activation, inactivation, or conductance properties of the Cav1.4 calcium channel (Table 1). In contrast, the G369D mutation resulted in statistically significant changes in the slope of activation (S) and in the difference in half-activation potential obtained in barium and calcium compared with the wild-type Cav1.4 channel. The notion that deletion of virtually the entire carboxyl terminus region had little effect on channel function is particularly surprising in view of previous work in other types of L-type calcium channels, showing that this region is an essential regulatory domain and required for proper function (Peterson et al., 1999; DeMaria et al., 2001). Taken together, these data suggest that the clinical phenotype associated with some alterations in the CACNA1F gene may not arise simply from altered biophysical properties of the channel.
Expression of Cav1.4 in human retina
To date, Cav1.4 protein distribution has only been examined in rodents and revealed Cav1.4 immunoreactivity in the outer plexiform layers and colocalization with protein bassoon (Morgans, 2001; Morgans et al., 2001; Berntson et al., 2003). There has been no description of Cav1.4 distribution in human retina, although, based on the notion that point mutations in the Cav1.4 calcium channel result in CSNB2, it might be expected that Cav1.4 channels comprise a major portion of the L-type calcium current in rod photoreceptors. Shown in Figure 5A is a tissue section of human retina stained with hematoxylin and eosin (H and E) demonstrating the constituents of the retina. Staining of human retina with the Cav1.4-specific antibody (Fig. 5B) revealed channel expression (brown precipitate) in the outer plexiform layer, consistent with localization of the channels in photoreceptors. Additional Cav1.4 staining was evident in the inner nuclear layer, inner plexiform layer, and nerve fiber layer. Cav1.4 staining is distinct and not as widespread from that reported for other HVA calcium channels with P/Q-, N-, and other L-type calcium channels present in the additional layers of the retina (Xu et al., 2002). Figure 5C is a negative control tissue section of retina and retinoblastoma tumor stained with preabsorbed Cav1.4 antibody, showing that the antibody reactivity has been abolished.
The retina sample depicted in Figure 5 was obtained from a retinoblastoma patient whose retina was displaced by the invasive tumor. Figure 6 illustrates Cav1.4 immunoreactivity in a retinal section from the same patient, but magnified to include the tumor tissue. Figure 6A is a tissue section of human retina stained with H and E demonstrating the constituents of the retinoblastoma tumor. As shown in Figure 6B, there was intense Cav1.4 staining throughout the retinoblastoma, except within the necrotic region and the blood vessel. As shown in Figure 6C, a negative control tissue section of retina and retinoblastoma tumor stained with preabsorbed Cav1.4 antibody and counterstained with H and E did not yield Cav1.4 immunoreactivity. This result suggests the possibility that Cav1.4 channel expression is either upregulated during tumor growth or, alternatively, that the channel might mediate a role in the development of retinoblastoma per se.
Tissue distribution of Cav1.4 mRNA and protein
To examine the distribution of Cav1.4 expression, we performed RT-PCR analysis of mRNA derived from 24 different human tissues. No Cav1.4 message was detected in RNA isolated from whole brain, heart, colon, fetal brain, fetal liver, kidney, lung, liver, mammary gland, pancreas, placenta, prostate, salivary gland, small intestine, stomach, testis, trachea, and the uterus. In contrast, Cav1.4 expression could be detected in RNA isolated from the spleen, thymus, adrenal gland, spinal cord, bone marrow, and skeletal muscle (Fig. 7A). These data indicate that mRNA encoding Cav1.4 is not exclusively transcribed in retinal tissue. To further examine these findings at the protein level, we performed Western blot analysis on protein homogenates isolated from 12 rat tissues, each from three separate rats. As shown in Figure 7B, Cav1.4 protein expression was detected by the antibody hybridizing to a single specific band (∼220 kDa) in whole-cell protein lysate isolated from rat retina as well as from the spleen, spinal cord, bone marrow, and thymus. No antibody hybridization was seen in the protein sample isolated from the heart, whole brain, liver, intestine, ovary, lung, and kidney.
The finding of Cav1.4 expression in both rat and human lymphoid tissues was of particular interest, and additional immunostaining was performed on tissue sections of human lymph node, thymus, and spleen. Figure 8A shows intense Cav1.4 positive staining (brown precipitate) of plasma cells within lymph node germinal centers. Other B lymphocytes (i.e., follicle center cells and mantle cells) and T lymphocytes did not interact with the Cav1.4 antibody. In addition, a subset of mast cells in the perinodal soft tissue stained strongly positive. Within thymic and splenic tissue, only rare cells stained positively, and definite morphological identification of these cells was not possible.
In surgical pathology, H and E staining is often used to assess tissue morphology and cytologic detail. Hematoxylin, with its net-positive charge, stains negatively charged material, including the genetic material in nuclei (blue/purple). Eosin, with its net-negative charge, stains positively charge intracellular material, such as cytoskeletal elements and cytoplasmic proteins (pink/red). With H and E staining, plasma cells appear as round to oval cells measuring ∼14-20 μm in size. The nucleus is small relative to the volume of cytoplasm, eccentrically placed, and characteristically has moderate amounts of condensed chromatin, giving an appearance reminiscent of a “clock face.” The cytoplasm of plasma cells is typically eosinophilic (pink) or amphophilic (purplish), and a pale paranuclear area corresponding to the Golgi apparatus is characteristic. Mast cells are similar in size or slightly larger than plasma cells and with H and E stain appear oval or polygonal in shape. The cytoplasm is characteristically granular and more eosinophilic (reddish) in comparison with plasma cells. The nucleus may be centrally or eccentrically placed, but the chromatin is typically more evenly distributed, and paranuclear clearing is not a feature in comparison with plasma cells. To confirm expression of Cav1.4 calcium channels in plasma and mast cells, a section of uterine cervix showing chronic cervicitis, with abundant plasma cell and mast cell infiltrates, was stained (Fig. 8B). Similar to that for lymph nodes, intense positive staining (brown precipitate) of plasma and mast cells was observed (Fig. 8B). As shown in Figure 8C, a control tissue section (i.e., uterine cervix) stained with preabsorbed Cav1.4 antibody and counterstained with H and E did not yield Cav1.4 immuoreactivity. Taken together, these data indicate that the Cav1.4 L-type calcium channel is more widely expressed than originally believed and suggest that it may play an as yet to be defined role in the immune system.
Discussion
Biophysical properties of the Cav1.4 calcium channel
Calcium currents in rod and cone inner segments have remained relatively poorly characterized because of the difficulty in dissociating and isolating photoreceptors with intact inner segments and because of their small size in most species used for physiological analysis. In particular, there is no information concerning the ionic currents in human photoreceptors. Perhaps the most complete description of ionic currents in photoreceptors comes from work in tiger salamander retina (Corey et al., 1984; Barnes and Hille, 1989; Barnes, 1994; Kurenny et al., 1994; Matthews, 1995; Wilkinson and Barnes, 1996; Akopian et al., 2000; Kourennyi and Barnes 2000; Thoreson and Stella, 2000). Corey et al. (1984) reported that whole-cell calcium currents in salamander rod photoreceptors first activated at ∼-40 mV and peaked at ∼0 mV. The current size increased when barium replaced calcium. In the presence of 10 mm internal EGTA, currents carried by both barium and calcium remained sustained over the course of several seconds, although a small degree of calcium-dependent inactivation became apparent on reducing internal EGTA to 0.1 mm. Kurenny et al. (1994) reported similar biophysical properties for rod L-type calcium currents and also showed a pronounced current enhancement in the presence of the nitric oxide donor S-nitrocysteine. Salamander rod photoreceptor currents are sensitive to DHP antagonists (Barnes, 1994; Thoreson et al., 1997) and show some sensitivity to the N-type calcium channel blocker ω-conotoxin GVIA (for review, see Barnes and Kelly, 2003). Although it has been suggested that L-type calcium currents in cone photoreceptors are carried primarily by Cav1.3 calcium channels (Taylor and Morgans, 1998), the notion that mutations in Cav1.4 calcium channels result in defects in rod vision (Bech-Hansen et al., 1998a,b; Strom et al., 1998) strongly suggests that the rod L-type currents are predominantly attributable to Cav1.4, an idea that is supported by our data showing Cav1.4 immunoreactivity in retina layers associated with human rod cells.
Our data in many ways parallel the results of Corey et al. (1984) and of findings in mammalian bipolar cells (de la Villa et al., 1998; Protti and Llano, 1998). Similar to salamander rod L-type currents, human Cav1.4 channels showed prolonged openings over several seconds, exhibited larger whole-cell currents in barium compared with calcium, and did not possess calcium-sensitive inactivation in the presence of low internal EGTA. The slight inhibition of Cav1.4 by ω-conotoxin GVIA and complete block by cadmium are also consistent with prior work in salamander rods (Barnes and Kelly, 2004). Calcium currents recorded by Protti and Llano (1998) from bipolar cells in rat retinal slices in 2 mm Ca2+ showed activation thresholds of -40 mV and negligible inactivation over the course of hundreds of milliseconds and were sensitive to DHPs, including nifedipine and BayK 8644. de la Villa et al. (1998) identified a transient current in mouse bipolar cells that activated at potentials near -40mV, showed no sign of time-dependent inactivation, was more permeable to barium ions than calcium, and showed high sensitivity to DHPs but not ω-conotoxin GVIA. Thus, human Cav1.4 currents show the same characteristics as native calcium currents recorded in mouse and rat bipolar cells, as well as salamander rods, and our staining of retinal sections show that Cav1.4 is localized in regions in human retina in which rod and bipolar cells are located. In support of the biophysical properties that we report here, Koschak et al. (2003) also found that Cav1.4 encodes for a novel type of DHP-sensitive calcium channel that lacks Ca-dependent inactivation; however, these authors did not use internal solutions with low calcium buffering capacity.
Photoreceptors tonically release glutamate in the dark because of sustained activity of L-type calcium channels (Taylor and Morgans, 1998). During a light stimulus, photoreceptors hyperpolarize, thus deactivating the L-type channels and terminating glutamate release. This necessitates that the L-type channels be tonically active at the typical photoreceptor resting potential of -40 mV (Schneeweis and Schapf, 1995). Like native rat and mouse bipolar cell L-type currents, transiently expressed Cav1.4 channels bathed in 2 mm external calcium first activated ∼-40 mV, demonstrating that this channel can indeed function in the range of typical photoreceptor resting membrane potentials. The large window current, together with the ultraslow inactivation displayed by Cav1.4 channels, make this channel ideally suited for maintaining tonic glutamate release in photoreceptors. The need for the capability of this channel to undergo prolonged openings may be underscored by the inability of calcium channel β subunits and calcium ions to accelerate inactivation kinetics (Figs. 3, 4). The molecular basis of this unique property remains to be explored. It is interesting to note that the Cav1.4 channel contains IQ and EF hand motifs in the C-terminal regions, which should, in principle, permit calcium-sensitive inactivation to occur (Peterson et al., 1999). Indeed, the original work of Corey et al. (1984) suggests that when the buffering capacity of the internal recording solution is reduced, some calcium-dependent inactivation may become apparent, similar to what has been described for other high-voltage-activated calcium channels (Lee et al., 1999; Zuhlke et al., 1999; Liang et al., 2003). However, we were unable to detect any calcium-dependent inactivation in the present study. Hence, it is yet to be determined as to whether calmodulin, the mediator of calcium-dependent inactivation in other HVA calcium channels, is capable of interacting with the Cav1.4 C terminus.
The calcium channel β subunit binding region in the domain I-II linker region of all high-threshold calcium channels is also conserved in Cav1.4. However, the I-II linker contains an additional 29 amino acids, and the remainder of the I-II linker sequence diverges considerably from that of other types of HVA calcium channels, both of which may affect β subunit interaction. In addition, it has been proposed that the domain I-II linker may serve as a physical inactivation gate of the channel and that β subunits regulate inactivation kinetics by regulating the mobility of this region (Stotz and Zamponi, 2001a,b). It is possible that the Cav1.4 I-II linker serves only poorly as an inactivation-gating particle to begin with, hence, rendering ancillary β subunits ineffective in regulating this process. We also cannot rule out the possibility that the ultraslow inactivation kinetics are not caused by a classical hinged-lid inactivation mechanism but instead reflect slow (C-type like) inactivation, or possibly a combination thereof.
Missense mutations implicated in night blindness
Examination of the functional consequences of naturally existing mutations in clinically affected individuals often represents an informative way to study both ion channel structure-function and to help understand the underlying clinical pathophysiology. In our analyses of three missense mutations and a carboxyl truncation associated with the CACNA1F gene and CSNB2, we found that only the G369D change in domain IS6 affected channel biophysical properties (Table 1). That only one missense mutation resulted in changes to biophysical properties and that the other two missense mutations and the premature stop did not seem to alter the biophysical properties of the channel is somewhat surprising, although not without precedence. For example, in some hypokalaemic periodic paralysis mutations found in skeletal muscle L-type calcium channels, the biophysical properties of the channels are only mildly altered, yet the clinical phenotype is quite severe (Lerche et al., 1996). Additionally, the effects of mutations in Cav2.1 P/Q-type channels implicated in familial hemiplegic migraine ranged from little, if any, change in channel properties to dramatic alterations of gating kinetics (Kraus et al., 1998, 2000; Wappl et al., 2002). In the case of the Cav1.4 channel, it is possible that channel activity in intact photoreceptors is regulated very tightly and that even subtle barely detectable changes in channel activity result in defective photoreceptor synaptic transmission. Alternatively, the effects of the mutations might not manifest themselves when the channel is removed from its native cellular environment. For example, it is conceivable that the mutations might interfere with the action of a regulatory protein that is present only in photoreceptors. Finally, in principle, it is also possible that some reported missense mutations might be polymorphisms found in the general population. Examination of additional mutations (Wutz et al., 2002) may reveal additional changes in Cav1.4 channel function.
Possible physiological function of Cav1.4 outside of the retina
Our data examining immunostaining in the human retina, together with the electrophysiological analysis, support a role of Cav1.4 in rod photoreceptor synaptic transmission. Cav1.4 was originally believed to be retinal specific (Bech-Hansen et al. 1998a,b; Strom et al., 1998); however, our data indicate that this is unlikely to be the case. Cav1.4 mRNA and protein were both detected in bone marrow, thymus, adrenal gland, and skeletal muscle. Interestingly, the slow inactivation kinetics of Cav1.4 would make this channel an ideal candidate for supporting hormone secretion, similar to what has been suggested for Cav1.3 channels in insulin-secreting cells, and the presence of Cav1.4 in the adrenal gland and thymus may be consistent with such a role. It is more difficult to ascribe a functional role of Cav1.4 in skeletal muscle and spinal cord, and future experiments examining more carefully the distribution of these channels in these tissues will be required.
Immunohistochemical staining of lymph tissue and cervical stroma displayed a strong cross-reactivity to the Cav1.4 antibody. Of particular note, the activation and release of both plasma cells and mast cell contents is regulated by the influx of calcium, although the mechanism by which this occurs is presently unknown (Rohlich et al., 1971; Sollner et al., 1993; Kim et al., 1997). Our data suggest the possibility that one means by which calcium enters plasma and mast cells is through the Cav1.4 L-type calcium channel. Previous reports have suggested that L-type calcium channels function on the surface of plasma and mast cells as treatment with the L-type calcium channel blocker nifedipine suppresses calcium transport into these cells (Briede et al., 1999). In addition, the presence of syntaxin peptides and synaptotagmin in mast cells suggest a support/scaffold role for L-type calcium channel-mediated exocytosis (Baram et al., 1998, 2001; Pombo et al., 2001). Interestingly, calcium-dependent exocytosis in mast cells is stimulated by the calcium sensor synaptotagmin I, whereas syntaptotagmin II negatively regulates MHC class II presentation by mast cells (Baram et al., 1999, 2002). In addition, in a rat basophilic leukemia cell (RBL-2H3), expression of the calcium sensor synaptotagmin has been shown to initiate the targeting of the secretory granules near the cell membrane. It would be interesting in future studies to examine whether synaptotagmin I and II regulate Cav1.4 function and whether the Cav1.4 channel colocalizes with either of the synaptotagmins within mast and plasma cells.
Footnotes
This work was supported by operating grants from the Canadian Institutes for Health Research (CIHR) to T.P.S. and G.W.Z. T.P.S. is a CIHR Senior Investigator. G.W.Z. is a CIHR Investigator, a Senior Scholar of the Alberta Heritage Foundation for Medical Research (AHFMR), and a National Alliance for Research on Schizophrenia and Depression Independent Investigator. J.E.M. was supported by a CIHR Postdoctoral Fellowship Award, M.H. by a Natural Sciences and Engineering Research Council (Canada) graduate studentship, and A.B. and C.J.D. hold AHFMR studentships.
Correspondence should be addressed to Dr. Terrance P. Snutch, Biotechnology Laboratory, University of British Columbia, 6174 University Boulevard, Vancouver, British Columbia, Canada V6T 1Z3. E-mail: Snutch{at}zoology.ubc.ca.
J. E. McRory's present address: Department of Physiology and Biophysics, Cellular and Molecular Neurobiology Research Group, University of Calgary, Calgary, Alberta, Canada T2N 4N1.
E. Garcia's present address: Centro de Investigaciones Biomedicas, Universidad de Colima, A.P. 199, Colima, 2800, Mexico.
Copyright © 2004 Society for Neuroscience 0270-6474/04/241707-12$15.00/0