Functional impact of a congenital stationary night blindness type 2 mutation depends on subunit composition of Ca_v1.4 Ca²⁺ channels

Abstract

Voltage-gated Ca_v1 and Ca_v2 Ca²⁺ channels are comprised of a pore-forming α₁ subunit (Ca_v1.1-1.4, Ca_v2.1-2.3) and auxiliary β (β_1-4) and α₂δ (α₂δ−1−4) subunits. The properties of these channels vary with distinct combinations of Ca_v subunits and alternative splicing of the encoding transcripts. Therefore, the impact of disease-causing mutations affecting these channels may depend on the identities of Ca_v subunits and splice variants. Here, we analyzed the effects of a congenital stationary night blindness type 2 (CSNB2)-causing mutation, I745T (IT), in Ca_v1.4 channels typical of those in human retina: Ca_v1.4 splice variants with or without exon 47 (Ca_v1.4+ex47 and Ca_v1.4Δex47, respectively), and the auxiliary subunits, β_2X13 and α₂δ-4. We find that IT caused both Ca_v1.4 splice variants to activate at significantly more negative voltages and with slower deactivation kinetics than the corresponding WT channels. These effects of the IT mutation, along with unexpected alterations in ion selectivity, were generally larger in channels lacking exon 47. The weaker ion selectivity caused by IT led to hyperpolarizing shifts in the reversal potential and large outward currents that were evident in channels containing the auxiliary subunits β_2X13 and α₂δ-4 but not in those with β_2A and α₂δ-1. We conclude that the IT mutation stabilizes channel opening and alters ion selectivity of Ca_v1.4 in a manner that is strengthened by exclusion of exon 47 and inclusion of β_2X13 and α₂δ-4. Our results reveal complex actions of IT in modifying the properties of Ca_v1.4 channels, which may influence the pathological consequences of this mutation in retinal photoreceptors.

Voltage-gated Ca²⁺ (Ca_v) channels are comprised of a pore-forming α₁ subunit and two auxiliary subunits, β and α₂δ (reviewed in Ref. 1). The α₁ subunit is comprised of 4 homologous domains (I-IV), each containing 6 alpha-helical transmembrane-spanning segments (S1–S6); the S1–S4 segments form a voltage-sensing domain, and S5–S6 forms the pore (2). In contrast to the diverse complement of Ca_v channels expressed in many neurons, the pore-forming α_1F subunit (referred to as Ca_v1.4 from here on), encoded by the Cacna1f gene, appears to be the major Ca_v subtype localized in the synaptic terminals of photoreceptors in the retina (3, 4) where it co-assembles with β₂ and α₂δ-4 subunits (5). Within photoreceptor synaptic terminals, Ca_v1.4 channels are activated at the relatively depolarized voltage of these cells in darkness, causing the tonic release of glutamate. At the sign-inverting synapse formed between photoreceptors and depolarizing bipolar cells, the termination of Ca_v1.4-dependent glutamate release by light stimuli enables disinhibition of a nonselective cation channel, initiating excitation of the ON pathway in the retina (6, 7). Thus, the voltage-dependent properties of Ca_v1.4 are critical parameters for controlling the dynamic range of visual signaling.

More than 140 mutations in Cacna1f have been identified and are linked to vision disorders including congenital stationary night blindness type 2 (CSNB2) (reviewed in Ref. 8). The sequelae of these mutations are not entirely clear because, when analyzed in heterologous expression systems, they can weaken, enhance, or have no impact on the function of Ca_v1.4 (9–13). Understanding the pathological consequences of CSNB2 mutations is complicated by the functional diversity of retinal Ca_v1.4 conferred in part by alternative splicing of the pre-mRNAs corresponding to each subunit (5, 14–16). The β₂ variant that is most highly expressed in human retina contains an alternatively spliced exon 7B (β_2X13) and causes stronger voltage-dependent inactivation of Ca_v1.4 than β₂ variants with exon 7A (β_2A) (5). The Ca_v1.4 α₁ pre-mRNA also undergoes alternative splicing, particularly in the sequence encoding the large cytoplasmic C-terminal domain (15, 16). We previously characterized a Ca_v1.4 variant lacking exon 47 (Ca_v1.4Δex47) that is highly expressed in human retina (16). Exon 47 encodes a portion of a C-terminal modulatory domain (CTM) in Ca_v1.4 that suppresses Ca²⁺-dependent inactivation (CDI) and causes depolarizing shifts in the voltage-dependence of activation (10, 17). When expressed in a human embryonic kidney cell line (HEK293T), Ca_v1.4Δex47 exhibits more negative activation thresholds and stronger CDI than Ca_v1.4 variants containing exon 47 (Ca_v1.4+ex47) (16, 18).

Studies investigating the electrophysiological consequences of Cacna1f mutations have focused on the Ca_v1.4+ex47 variant coexpressed with auxiliary subunits other than β_2X13 and α₂δ-4 (9–13). Alternative splicing can affect the severity of disease-causing mutations in Ca_v channel genes (19). Thus, analysis of Cacna1f mutations in the context of Ca_v1.4 variants expressed in photoreceptors in human retina is necessary for understanding the visual phenotypes associated with such mutations.

Here, we investigated the effects of a CSNB2-causing mutation on the properties of Ca_v1.4+ex47 and Ca_v1.4Δex47 channels containing β_2X13 and α₂δ-4. The mutation results in the replacement of isoleucine 745 with a threonine (IT) in the S6 helix of domain 2 (IIS6, Fig. 1A). In Ca_v1.4+ex47 coexpressed with α₂δ-1 and β₃ or β₂, the IT mutation causes a large hyperpolarizing shift (>30 mV) in the voltage-dependence of activation (12). Our results indicate that, when coexpressed with β_2X13 and α₂δ-4, Ca_v1.4+ex47 channels bearing the IT mutation (Ca_v1.4+ex47^IT) show hyperpolarized activation voltages compared with wild-type (WT) channels. The gain-of function effect is more severe for Ca_v1.4Δex47 channels with the IT mutation (Ca_v1.4Δex47^IT), which showed more negative activation thresholds and slower deactivation kinetics than Ca_v1.4+ex47^IT. An unexpected finding is that IT alters the ion selectivity of both Ca_v1.4 splice variants in a manner that varies with the identity of the α₂δ subunit. Our findings highlight the importance of splice variation and auxiliary subunit composition as potential modifiers of disease-causing mutations affecting Ca_v channels.

Figure 1.

IT enhances voltage-dependence of activation of Ca_v1.4+ex47. A, schematic of Ca_v1.4 pore-forming α₁ subunit with 4 transmembrane spanning domains (I-IV; blue), and β_2X13 (tan) and α₂δ-4 (green) subunits. The CTM of Ca_v1.4 (purple) contains exon 47 (orange). The red star illustrates the location of IT in domain II. B, representative I_Ba family of traces for Ca_v1.4+ex47 or Ca_v1.4+ex47^IT. C, I-V plots for I_Ba current density (pA/pF) in cells transfected with Ca_v1.4+ex47 (black) or Ca_v1.4+ex47^IT (red). I_Ba was evoked by 50-ms pulses from −100 mV to various voltages. Here and in all graphs of electrophysiological data, parentheses indicate number of cells, and symbols and error bars represent mean ± S.E., respectively.

Results

IT mutation enhances activation and slows deactivation of Ca_v1.4+ex47

Exon 47 resides in the CTM of Ca_v1.4 (Fig. 1A); deletion of this exon, like the IT mutation, causes a large negative shift in the voltage dependence of channel activation (16, 18). Thus, the effect of IT on Ca_v1.4 activation could be additive, or alternatively, could be occluded by exon 47 deletion. To distinguish between these possibilities, we compared the activation properties of Ba²⁺ currents (I_Ba) mediated by Ca_v1.4+ex47 and Ca_v1.4Δex47, and the corresponding IT mutant channels, in transfected HEK293T cells. Ba²⁺ rather than Ca²⁺ was used as the charge carrier to minimize the complicating effects of CDI which, whereas negligible in Ca_v1.4+ex47, is prominent in Ca_v1.4Δex47 (18). Because previous analyses of Ca_v1.4+ex47^IT were performed primarily with β₃ or β_2a, and α₂δ-1 (12), we first characterized the effect of IT on Ca_v1.4+ex47 coexpressed with auxiliary subunits representative of Ca_v1.4 complexes in the retina (i.e. β_2X13 and α₂δ-4 (5)). Although there was no effect of IT on the slope factor (k), Boltzmann fits of current-voltage (I-V) plots showed that the half-maximal voltage of activation (V_h) of Ca_v1.4+ex47^IT was significantly more negative than that of Ca_v1.4+ex47 (Fig. 1, B and C, Table 1).

Table 1.

Parameters from I–V relationships of Ca_v1.4+ex47 and Ca_v1.4Δex47 with or without IT mutation

V_h, k, E_rev, and peak I_Ba values (mean ± S.E.) were determined from Boltzmann fits of the I-V data in Figs. 1, 3, and 9 and as described under “Experimental procedures.”

	Peak IBa (pA/pF)	p Value	Vh (mV)	p Value	k	p Value	Erev (mV)	p Value
Cav1.4+ex47 + β2x13 + α2δ-4	−5.6 ± 1.4	–	−8.6 ± 0.4	–	−6.2 ± 1.6	–	58.3 ± 3.7	–
Cav1.4+ex47IT + β2x13 + α2δ-4	−5.3 ± 1.2	0.904a	−28.5 ± 2.0	<0.0001a	−8.5 ± 1.2	0.253b	37.1 ± 3.7	<0.0001a
Cav1.4Δex47 + β2x13 + α2δ-4	−10.5 ± 1.9	–	−24.8 ± 1.3	–	−6.6 ± 0.4	–	54.3 ± 1.5	v
Cav1.4Δex47IT + β2x13 + α2δ-4	−2.6 ± 0.4	0.002b	−43.7 ± 2.1	<0.0001b	−11.2 ± 1.0	<0.001b	16.7 ± 5.0	<0.0001a
Cav1.4+ex47 + β2A + α2δ-1	−11.2 ± 2.0	–	−2.7 ± 1.2	–	−7.1 ± 0.2	–	52.7 ± 0.8	–
Cav1.4+ex47IT + β2A + α2δ-1	−9.2 ± 3.0	0.581b	−28.3 ± 2.0	<0.0001b	−7.4 ± 1.1	0.527a	50.1 ± 2.4	0.237b
Cav1.4Δex47 + β2A + α2δ-1	−17.8 ± 0.9	–	−19.2 ± 0.5	–	−4.6 ± 0.8	–	50.4 ± 3.2	–
Cav1.4Δex47IT + β2A + α2δ-1	−0.6 ± 0.1	0.082b	−36.1 ± 2.5	0.003b	−7.7 ± 0.5	0.009b	41.0 ± 3.2	0.121b

^a Mann-Whitney test.

^b Student's t test.

Exponential fits of the rising phase of the peak currents yielded time constants for activation (τ_act) that were significantly longer (Table 2) and with weaker voltage dependence for Ca_v1.4+ex47^IT (v = −50.3 mV) than for Ca_v1.4+ex47 (v = −26.9 mV; F_2,7 = 16.4, p = 0.002; Fig. 2, A and B). To analyze rates of channel closure, the time constant for deactivation (τ_deact) was obtained from exponential fits of the decay phase of the tail current evoked upon repolarization of the membrane voltage. τ_deact was significantly greater at the most positive repolarization voltage tested (-60 mV, Table 2) and the voltage-dependence of τ_deact was significantly steeper for Ca_v1.4+ex47^IT (v = 43.1 mV) than for Ca_v1.4+ex47 (v = 169.9 mV; F_2,22 = 59.2, p < 0.0001; Fig. 2, C and D). Thus, as has been shown for Ca_v1.2 channels bearing the analogous IT mutation (20), IT slows the activation and deactivation of Ca_v1.4 containing exon 47 in a highly voltage-dependent manner.

Table 2.

Time constants for activation and deactivation of Ca_v1.4+ex47 and Ca_v1.4Δex47 with or without IT mutation

τ_act was obtained from single exponential fits of the rising phase of currents evoked by the voltages indicated in parentheses. τ_deact was obtained from single exponential fits of the rising phase of currents evoked upon depolarization from 10-ms steps to voltages evoking peak inward I_Ba indicated in parentheses and repolarization to −60 mV.

Channel	τact (ms)	p Value	τdeact (ms)	p-Value
Cav1.4+ex47 + β2x13 + α2δ-4	1.8 ± 0.2 (0 mV)	–	4.4 ± 0.8 (0 mV)	–
Cav1.4+ex47IT + β2x13 + α2δ-4	2.8 ± 0.2 (−20 mV)	0.019a	14.2 ± 1.8 (−20 mV)	<0.001a
Cav1.4Δex47 + β2x13 + α2δ-4	2.0 ± 0.2 (−10 mV)	–	3.1 ± 0.3 (−10 mV)	–
Cav1.4Δex47IT + β2x13 + α2δ-4	2.5 ± 0.4 (−30 mV)	0.343a	37 ± 9.7 (−30 mV)	0.004b

^a Student's t test.

^b Mann-Whitney test.

Figure 2.

IT alters kinetics of activation and deactivation for Ca_v1.4+ex47. A, voltage protocol for measuring activation kinetics (left) and representative I_Ba family of traces (right). I_Ba was evoked by 50-ms pulses from −100 mV to various test voltages. Current traces are color-coded according to the depolarizations used to evoke them in voltage protocol. Exponential fits (black lines) are overlaid on corresponding current traces. B, activation time constants (τ_act) were obtained from exponential fits of I_Ba and plotted against test voltage in cells transfected with Ca_v1.4+ex47 (black symbols) or Ca_v1.4+ex47^IT (red symbols). C, voltage protocol for measuring deactivation kinetics (left) and representative family of I_Ba traces (right). Tail I_Ba was evoked by 10-ms pulses to voltages evoking peak inward I_Ba (see Table 2) followed by repolarizations to various voltages. Exponential fits of tail I_Ba are color-coded according to the repolarization voltage used to evoke them in voltage protocol. D, deactivation time constants (τ_deact) were obtained from exponential fits of tail I_Ba and plotted against repolarization voltage in cells transfected with Ca_v1.4+ex47 (black symbols) or Ca_v1.4+ex47^IT (red symbols). In B and D, solid lines represent exponential fits of the averaged data. Ca_v1.4 variants were co-expressed with β_2X13 and α₂δ-4.

Deletion of exon 47 augments effects of the IT mutation on voltage-dependent gating of Ca_v1.4

We next investigated how deletion of exon 47 affects the impact of the IT mutation (Fig. 3A). As for Ca_v1.4+ex47 (Fig. 1C), IT caused a negative shift in V_h for Ca_v1.4Δex47 (Fig. 3, B and C, Table 1). The net hyperpolarizing effect of IT (ΔV_h) was not significantly different between Ca_v1.4+ex47 (median ΔV_h= 19.9 mV, n = 11) and Ca_v1.4Δex47 (median ΔV_h= 18.9 mV, n = 8; Mann-Whitney U = 44, p > 0.999). However, the additive effects of the IT mutation and deletion of exon 47 resulted in an extremely negative activation threshold of Ca_v1.4Δex47^IT (∼ −70 mV, Fig. 3C). Moreover, IT enhanced rather than weakened the voltage-dependence of τ_act in the absence of exon 47 (v = −21.4 mV for Ca_v1.4Δex47^IT versus v = −33.4 mV for Ca_v1.4Δex47; F_2,8 = 5.3, p = 0.03; Fig. 4, A and B).

Figure 3.

IT enhances voltage-dependence of activation of Ca_v1.4Δex47 and decreases current density. A–C, same as described in the legend to Fig. 1, except for cells transfected with Ca_v1.4Δex47 or Ca_v1.4Δex47^IT (black and red symbols, respectively in C).

Figure 4.

IT significantly alters the activation and deactivation kinetics of Ca_v1.4Δex47. A–D, same as described in the legend to Fig. 2 except for cells transfected Ca_v1.4Δex47 or Ca_v1.4Δex47^IT (black and red symbols, respectively in B and D).

Similar to its effects in the presence of exon 47 (Fig. 2, C and D), IT strengthened the voltage-dependence of τ_deact (v = 104.1 mV for Ca_v1.4Δex47 versus v = 27.6 mV for Ca_v1.4Δex47^IT; F_2,22 = 151.6, p < 0.0001; Fig. 4, C and D). However, IT increased τ_deact more than 10-fold for Ca_v1.4Δex47 versus ∼4-fold for Ca_v1.4+ex47 upon repolarization to −60 mV (Table 2). These results indicate that deletion of exon 47 augments the gain-of-function effects of IT by modifying the kinetics and voltage-dependence of channel activation and deactivation.

Unique effects of IT on Ca_v1.4Δex47

An effect of IT that was not reported previously was a reduction in current density, which was only seen in the absence of exon 47 (Figs. 1C and 3C, Table 1). We first tested the possibility that IT impaired the stability of the channel in ways that diminished overall levels of the Ca_v1.4Δex47 protein. However, Western blots indicated similar levels of total channel protein in cells transfected with either Ca_v1.4Δex47 or Ca_v1.4Δex47^IT (Fig. 5A). Moreover, biotinylation and streptavidin pulldown of cell-surface proteins revealed no significant difference in the levels of Ca_v1.4Δex47 or Ca_v1.4Δex47^IT in the plasma membrane (Fig. 5B). Thus, impaired trafficking of the mutant channels to the cell surface was unlikely to be the major cause of the decrease in current density. A second unexpected effect of IT was an apparent decrease in ion selectivity based on the development of large outward currents at positive voltages and hyperpolarizing shift in the reversal potential (E_rev) (Figs. 1C and 3C, Table 1). The outward currents and median change in E_rev (ΔE_rev) were significantly larger for Ca_v1.4Δex47^IT (−37.2 mV, n = 8) than for Ca_v1.4+ex47^IT (−16.6 mV, n = 11; Mann-Whitney U = 14, p = 0.01) relative to the corresponding WT channels. Therefore, we probed the underlying mechanism with an emphasis on Ca_v1.4Δex47.

Figure 5.

IT does not alter the expression levels or cell-surface density of Ca_v1.4Δex47. A, representative Western blotting images of lysates from HEK293T cells that were untransfected (Control) or transfected with either Ca_v1.4Δex47 (Δex47) or Ca_v1.4Δex47^IT (Δex47^IT) as well as β_2X13 and α₂δ−4. Blots were probed with antibodies against Ca_v1.4, Na/K-ATPase, or GAPDH. The percentage of lysates used for total protein (left 3 lanes) and biotinylated cell-surface proteins (right 3 lanes) were 10 and 90%, respectively. B, densitometric analysis of total and cell-surface Ca_v1.4 protein normalized to those for GAPDH and Na/K-ATPase, respectively. The use of these proteins as normalization controls was justified because there was no effect of transfection on their levels (p = 0.84 for GAPDH and p = 0.99 for Na/K-ATPase, both by analysis of variance). Each point represents result from an independent experiment. Bars represent mean; p values were determined by t test.

The nature of the outward currents was mysterious considering that the major intracellular cation in our recording solutions was NMDG⁺ (N-methyl-d-glucamine), a large organic cation that does not permeate most voltage-gated ion channels. However, Ca_v1.2 and Ca_v1.3 are permeable to NMDG⁺ under some conditions (21, 22). If IT enabled NMDG⁺ efflux through Ca_v1.4 channels, then the outward currents in cells transfected with Ca_v1.4Δex47^IT should be reduced by known blockers of Ca_v channels such as Cd²⁺ (23) and by decreasing the chemical gradient of NMDG⁺ across the membrane. Consistent with these predictions, Cd²⁺ abolished outward currents in cells transfected with Ca_v1.4Δex47^IT as well as the inward currents in cells expressing either WT or mutant channels (Fig. 6). To test the effects of altering the NMDG⁺ concentration, we compared E_rev using extracellular solutions containing 5 or 130 mm NMDG⁺ ([NMDG⁺]₅ and [NMDG⁺]₁₃₀, respectively, Fig. 7A). Although having no effect on E_rev of Ca_v1.4Δex47 (66.5 ± 2.3 mV with [NMDG⁺]₅, n = 4 versus 68.3 ± 1.4 mV with [NMDG⁺]₁₃₀, n = 4, p = 0.532 by t test; Fig. 7, B and C), increasing extracellular NMDG⁺ caused a positive shift in E_rev (31.2 ± 2.7 mV with [NMDG⁺]₅, n = 4 versus 53.1 ± 3.6 mV with [NMDG⁺]₁₃₀, n = 3, p = 0.004 by t test) and diminished outward currents in cells expressing Ca_v1.4Δex47^IT (Fig. 7, B and C). In addition, increasing the extracellular [NMDG⁺] had no effect on the permeability of Ba²⁺ versus NMDG⁺ (P_Ba/P_NMDG) for Ca_v1.4Δex47 (351.4 ± 53.4, n = 4, for [NMDG⁺]₅ versus 391.7 ± 36.9, n = 4, for [NMDG⁺]₁₃₀, p = 0.558 by t test) but significantly increased that for Ca_v1.4Δex47^IT (27.4 ± 5.3, n = 4, for [NMDG⁺]₅ versus 133.4 ± 37.0, n = 3, for [NMDG⁺]₁₃₀, p = 0.020 by t test). We further assessed the effect of IT on selectivity of Ca_v1.4Δex47 by measuring E_rev and P_Ba/P_x under other bi-ionic conditions. With intracellular solutions containing Na⁺ or K⁺, IT caused a negative shift in E_rev and lowered P_Ba/P_x (Fig. 8, Table 3). Taken together, these results signified a reduction in the ionic selectivity of Ca_v1.4Δex47^IT compared with WT channels.

Figure 6.

Cd²⁺ suppresses inward and outward currents in cells transfected with Ca_v1.4Δex47^IT. A and B, representative traces (left) for I_Ba evoked by 50-ms pulses to the indicated voltages before (black) and after perfusion of extracellular solution containing Cd²⁺ (red; 100 μm). I-V plots (right) obtained before (Ba²⁺, black symbols) and after (Cd²⁺, red symbols) perfusion of Cd²⁺. Ca_v1.4 variants were co-expressed with β_2X13 and α₂δ-4.

Figure 7.

Increasing NMDG⁺ in the extracellular recording solution minimizes alterations in E_rev and outward currents caused by IT mutation. A, schematic showing composition of [NMDG]₅ and [NMDG]₁₃₀ recording solutions. B and C, representative current traces evoked by 50-ms pulses from −100 mV to the indicated voltages (left) and I-V plots (right) in cells transfected with Ca_v1.4Δex47 (top) or Ca_v1.4Δex47^IT (bottom). I/I_Max represents I_Ba normalized to peak inward current amplitude. Ca_v1.4 variants were co-expressed with β_2X13 and α₂δ-4.

Figure 8.

IT impairs selectivity of Ca_v1.4Δex47. A, representative current traces evoked by 50-ms pulses from −100 mV to the indicated voltages. B, I-V plots in cells transfected with Ca_v1.4Δex47 or Ca_v1.4Δex47^IT. I/I_max represents I_Ba normalized to peak inward current amplitude. C, expanded view of I-V plots in B with E_rev indicated by arrows for Ca_v1.4Δex47 (black) or Ca_v1.4Δex47^IT (red). Data were fit by linear regression. Intracellular solution contained 140 mm Na⁺ (top panels) or K⁺ (bottom panels). Ca_v1.4 variants were co-expressed with β_2X13 and α₂δ-4.

Table 3.

Parameters for monovalent and Ba²⁺ permeability for Ca_v1.4Δex47 and Ca_v1.4Δex47^IT

E_rev and P_Ba/P_x were determined from data shown in Figs. 7 and 8 and from equations described under “Experimental procedures,” p values were determined by unpaired t tests. Intracellular solutions contained 140 mm NMDG⁺, Na⁺, or K⁺. Significance was determined by Student's t test.

Erev (mV)	Cav1.4Δex47	Cav1.4Δex47IT	p Value
NMDG+	66.5 ± 2.3	31.2 ± 2.7	<0.001
Na+	65.5 ± 1.6	51.0 ± 2.7	0.004
K+	60.9 ± 3.4	50.8 ± 3.1	0.064
PBa/Px
NMDG+	351.4 ± 53.4	31.2 ± 2.7	0.0009
Na+	316.6 ± 34.6	113.8 ± 19.1	0.0009
K+	243.7 ± 58.8	116.2 ± 28.4	0.074

Although smaller for Ca_v1.4+ex47 than for Ca_v1.4Δex47 (Fig. 1C, Table 1) the effects of IT on E_rev were, nevertheless, not reported for Ca_v1.4+ex47 in a previous study (12). A key difference was in the choice of auxiliary subunits (β_2X13 and α₂δ-4, this study) versus β₃ or β_2A and α₂δ-1 (12)). Therefore, we tested the impact of IT on the Ca_v1.4 variants containing β_2A and α₂δ-1. Consistent with the previous study, IT caused a large negative shift in V_h in these experiments. Although the mutation strongly reduced current densities of Ca_v1.4Δex47 + β_2A + α₂δ-1, IT did not affect E_rev (Fig. 9, A–C, Table 1). Thus, the identity of the auxiliary β and α₂δ subunits critically determines the effects of IT on selectivity of Ca_v1.4.

Figure 9.

IT does not alter selectivity in Ca_v1.4 channels containing β_2Aδ and α₂δ-1. A, schematics of Ca_v1.4 (left). Representative current traces evoked by 50-ms pulses from −100 mV to the indicated voltages (right). B, I-V plots in cells transfected with Ca_v1.4+ex47, Ca_v1.4+ex47^IT, Ca_v1.4Δex47, or Ca_v1.4Δex47^IT. I_Ba (pA/pF) represents I_Ba normalized to peak inward current amplitude. C, expanded view of I-V plot in B. Data were fit by linear regression. Extracellular and intracellular solutions were the same as those used in Figs. 1–4.

Discussion

Our study provides new insights about how IT affects the biophysical properties of Ca_v1.4. First, we show that IT produces a large negative shift in voltage-dependent activation of Ca_v1.4 channels containing the major auxiliary Ca_v subunits in the retina, β_2X13 and α₂δ-4 (Figs. 1 and 3, Table 1), as well as Ca_v1.4 channels comprised of other auxiliary subunits (Fig. 9, Table 1, and see Ref. 12). Second, deletion of exon 47 exacerbates the gain of function effects of IT: Ca_v1.4Δex47^IT activates at more negative voltages and exhibits stronger voltage-dependent alterations in the kinetics of activation and deactivation than Ca_v1.4+ex47^IT (Figs. 1–4, Tables 1 and 2). Third, IT weakens the selectivity of Ca_v1.4 for Ba²⁺ in a manner that varies with the identity of the auxiliary β and α₂δ subunits (Figs. 1, 3, and 9, Tables 1 and 3). Our findings highlight the importance of splice variation and auxiliary subunit composition as potential modifiers of disease-causing mutations affecting Ca_v channels.

Conserved role of Ile-745 in activation gating

The S5 and S6 pore-lining helices give rise to the selectivity filter (2), with the four S6 helices (IS6–IVS6) converging at the intracellular side of the membrane in the closed state of the channel (2). Ile-745 of Ca_v1.4 corresponds to Ile-781 in IIS6 of Ca_v1.2, which lies in a cluster of hydrophobic residues (Leu-779–Ala-782, LAIA) in the S6 bundle-crossing region that are conserved among Ca_v1 and Ca_v2 channels (24). Disruptive mutations of these residues also cause hyperpolarizing shifts in activation and slowing of deactivation of Ca_v1.2 and Ca_v2.3 (20, 25). Our study is the first to show that the IT mutation causes similar effects on Ca_v1.4. Based on the correlation of their hydrophobicity and the negative shift in V_h (20, 25), the distal S6 residues are likely buried within a hydrophobic environment in the closed channel and become exposed to the aqueous milieu upon pore opening. By analogy to the model of Ca_v1.2 (26), contacts between Ile-745 with a corresponding hydrophobic residue in IIIS6 may stabilize helix-helix interactions, which support the closed conformation in Ca_v1.4, and are disrupted by the IT mutation.

Functions of exon 47 in regulating the impact of IT on Ca_v1.4 activation

Exon 47 encodes the initial 47 amino acids of the CTM, a modular domain present in both Ca_v1.4 and Ca_v1.3 that interacts with a region in the proximal C-terminal domain (10, 17, 27). The CTM nearly abolishes CDI of Ca_v1.4 by competing with calmodulin (CaM) for binding to the channel (10, 17). Deletion of the CTM enables CDI by allowing CaM binding to the channel, but also causes a negative shift in V_h (10). In Ca_v1.4, exon 47 is critical for the modulatory function of the CTM in that Ca_v1.4Δex47 exhibits similar alterations in V_h and CDI as those caused by deletion of the entire CTM (16, 18). Our findings that IT and deletion of exon 47 are additive with respect to hyperpolarizing V_h (Table 1) suggest distinct mechanisms by which Ile-745 and the CTM facilitate activation. In Ca_v1.3, deletion of the CTM leads to stronger pairing of voltage sensor charge movement and channel opening (28). In Ca_v2.3, the IIS4-S5 loop and the cytoplasmic end of IIS6 are thought to functionally interact in the activation pathway (see Ref. 29). In Ca_v1.4, partial deletion of exon 47 might disinhibit such intramolecular interactions, allowing IT to more freely destabilize closed channels and promote channel opening at more negative voltages than in channels with a complete CTM. Interactions of S4–S5 with S6 have been studied by homology modeling and molecular dynamics simulations of K_v channels (30). Similar approaches would be useful in dissecting the relationships of the corresponding regions, and of the CTM, with respect to activation gating of Ca_v1.4.

The effect of IT on hyperpolarizing V_h, whereas decreasing the peak current density of Ca_v1.4Δex47 (Table 1), parallels the effect of the S218L migraine-causing mutation in Ca_v2.1 expressed in HEK293 cells. In the latter case, the reduction in current density was determined to be an artifact of overexpression and related to a reduction in the number of functional channels in the membrane rather than changes in unitary current amplitudes (31). Because IT did not affect the total or cell-surface levels of Ca_v1.4Δex47 protein (Fig. 5), the reduced current density of Ca_v1.4Δex47^IT could result from a decrease in single channel conductance, and/or the functionality of the mutant channels within the membrane. Alternatively, the extremely negative activation properties of Ca_v1.4Δex47^IT could have compromised cell health such that outward leak currents compromised I_Ba amplitudes and caused the negative shift in E_rev. This scenario seems unlikely given that IT reduced current density but did not produce outward currents or alterations in E_rev in Ca_v1.4Δex47 channels containing β_2A and α₂δ-1 (Fig. 9, Table 1). Single channel recordings will be necessary to fully uncover the impact of IT on the elementary properties of Ca_v1.4Δex47.

Effects of IT on the ion selectivity of Ca_v1.4Δex47

The exquisite selectivity of Ca_v channels is largely determined by Ca²⁺ binding with high affinity to the selectivity filter (32, 33). Thus, the increased permeability of Na⁺, K⁺, and particularly NMDG⁺ caused by a mutation outside of the selectivity filter was unexpected. However, in the absence of Ca²⁺, Na⁺ and large organic cations such as tetramethylammonium are capable of permeating Ca_v1 channels (34). These results suggest that the pore of Ca_v channels is at least 6 Å in diameter, an interpretation that has been verified in structural analyses (2, 35). Indeed, despite being a relatively large cation (∼6.4 Å wide × 12 Å long; ∼7.3 Å mean diameter (36)), NMDG⁺ can permeate Ca_v1.2 channels containing pore mutations (37) and Ca_v1.3 channels exposed to the dihydropyridine agonist FPL 64176 (FPL) (21). Functional interactions between the selectivity filter and the inner S6 helix bundle are involved in K_v channel gating transitions (38) and may be conserved among Ca_v channels. For example, CaM binding to the cytoplasmic domain promotes conformational changes in the selectivity filter of Ca_v1 channels that lead to CDI (39). Thus, IT could alter positioning of IIS6 and its contributions to the Ca²⁺ (or Ba²⁺) binding affinity within the selectivity filter, allowing monovalent ions including NMDG⁺ and Na⁺ to permeate even in the presence of significant extracellular concentrations of Ba²⁺.

Our findings that impaired selectivity was specific to IT mutant channels containing β_2X13 and α₂δ-4 explain why previous analyses did not uncover any alteration in selectivity in these channels containing β_2A and α₂δ-1 (12). Unlike β_2A, β_2X13 lacks exon 7B, which causes increased voltage-dependent inactivation of Ca_v1.4 (5). Although it is unclear how this difference could affect ion selectivity of the IT mutant channels, there is evidence that structural alterations in α₂δ could affect the permeation properties of Ca_v channels. For example, CACHD1 is an α₂δ-like protein that has a disrupted metal-ion adhesion site that is critical for structural and functional interactions of α₂δ with the channel (2, 40, 41). When co-expressed with Ca_v2.2, CACHD1 impairs the ion selectivity of Ca_v2.2 (42). In the cryo-EM structure, α₂δ-1 forms multiple extracellular contacts with Ca_v1.1 including the extended loops between S5 and P1 helices in domains II and III (2). The L5 loops of each of the 4 domains form a domed window above the selectivity filter that direct Ca²⁺ ions into the pore (2). Differences in how α₂δ variants may interact with these extracellular sites, in concert with those produced by β subunits at intracellular sites, could determine the impact of IT on selectivity in the context of Ca_v1.4.

Significance for visual phenotypes of Ca_v1.4 channelopathies

CSNB2 is a nonprogressive retinal disorder with variable clinical features including reduced visual acuity, myopia, and nystagmus (43). A hallmark feature of this disorder is a reduced b-wave in electroretinograms, which is consistent with a defect in transmission from photoreceptors to second-order bipolar neurons (43, 44). Of the numerous CSNB2 mutations affecting Cacna1f, the IT mutation causes the most severe form of visual impairment (45). Despite the reduced current density of Ca_v1.4Δex47^IT in our experiments, the mutation enabled significant inward I_Ba at voltages negative to the activation thresholds of WT channels (Fig. 3C). Due to charge screening effects (46), our use of 20 mm Ba²⁺ in the external recording solutions would cause activation voltages ∼20 mV more positive than those expected in the retina; however, the relative differences in the voltage-dependent properties of the WT and IT mutant channels should be preserved under our recording conditions. Even in the presence of reduced current density, the negative shift in V_h and slow deactivation of Ca_v1.4Δex47^IT would lead to aberrant Ca²⁺ influx during light-dependent hyperpolarization of photoreceptors, thus degrading the fidelity of visual transmission to second-order neurons. However, our study also raises the possibility that the aberrant conductance of monovalent cations by Ca_v1.4Δex47^IT could lead to alterations in the excitability of photoreceptors that could lead to degenerative changes. Photoreceptor degeneration, as well as altered retinal ganglion cell activity and morphological and functional defects in photoreceptor synapses, are characteristic of an IT knock-in mouse line (47–50). However, Ca_v1.4 splice variants lacking exon 47, although abundant in human and monkey retina, are conspicuously absent from mouse retina (16). An understanding of the pathological consequences of Ca_v1.4Δex47^IT could therefore benefit from analyses of the mutant channels in human stem-cell derived photoreceptors in the context of retinal organoids (51).

Experimental procedures

cDNAs and molecular biology

The following cDNAs were used: Ca_v1.4 (GenBank AF201304), β_2A (GenBank AF465485), β_2X13 (GenBank NM_053851), α₂δ-1 (GenBank M86621), and α₂δ-4 (GenBank NM_172364) in pcDNA3.1. The construct encoding Ca_v1.4Δex47 was described previously (16). To incorporate the IT mutation into Ca_v1.4 (Ca_v1.4+ex47^IT and Ca_v1.4Δex47^IT), the upstream and downstream cDNA regions flanking the codon corresponding to I756 were amplified with Q5 High-Fidelity DNA polymerase (New England Biolabs) using Ca_v1.4+ex47 as the template and primers incorporating the mutation. PCR products were digested with DpnI, column purified, and cloned into Ca_v1.4+ex47 and Ca_v1.4Δex47 between AgeI and ClaI with the NEBuilder HiFi DNA Assembly kit (New England Biolabs) following the manufacturer's protocol. All constructs were verified by DNA sequencing before use.

Cell culture and transfection

Human embryonic kidney (HEK) 293 cells transformed with SV40 T antigen (HEK293T, CRL-3216, RRID:CVCL_0063; ATCC) were cultured in Dulbecco's modified Eagle's medium (Life Technologies, Grand Island, NY) with 10% fetal bovine serum (Atlantic Biologicals) at 37 °C in 5% CO₂. Cells were not used after they were passaged 15 times. At 70–80% confluence, the cells were co-transfected with cDNAs encoding human Ca_v1.4 α₁ (1.8 μg; Ca_v1.4+ex47, Ca_v1.4+ex47^IT, Ca_v1.4Δex47, or Ca_v1.4Δex47^IT), β_2A or β_2X13 (0.6 μg), α₂δ-4 or α₂δ-1 (0.6 μg), and enhanced GFP in pEGFP-C1 (0.1 μg) using FuGENE 6 transfection reagent (Promega) according to the manufacturer's protocol. In some experiments, cells were co-transfected with a plasmid encoding SK-1 Ca²⁺-activated K⁺ channel (0.1 μg) in an effort to reduce toxicity (there were no differences in results obtained in cells transfected with or without SK-1 and thus data were combined). Cells treated with the transfection mixture were incubated at 37 °C for 24 h. After 24 h, cells were incubated at 30 °C for at least 24 h prior to whole-cell patch clamp recordings.

For Western blotting and cell-surface biotinylation assays, HEK293T cells were transfected using Lipofectamine 3000 reagent (Life Technologies). Plasmid DNA (Ca_v1.4Δex47 or Ca_v1.4Δex47^IT (1.8 μg), β_2X13, and α₂δ-4 (0.6 μg each)) was diluted in Opti-MEM (50 μl, Life Technologies) and 4 μl of P3000 reagent. This was added to a mixture of Opti-MEM (50 μl) and Lipofectamine 3000 reagent (3 μl) and incubated for 10 min at room temperature. The DNA mixture was added incubated with the cells for 24 h at 37 °C in 5% CO₂ after which the cell culture medium was replaced with fresh medium.

Electrophysiology

Whole-cell patch clamp recordings were performed at room temperature between 48 and 72 h after transfection. Data were obtained under voltage-clamp with an EPC-9 patch clamp amplifier operated by Patchmaster software (HEKA Elektronik). The composition of recording solutions contained as follows (in mm): for Figs. 1–4 and 9, external solution contained NMDG (140), BaCl₂ (20), and MgCl₂ (1); internal solution contained NMDG (140), HEPES (10), MgCl₂ (2), Mg-ATP (2), and EGTA (5). For Fig. 6, Cd²⁺ (100 μm) was added to the external solution; pH was adjusted to 7.3 with methanesulfonic acid. For Fig. 7, the external solution contained Tris (130), NMDG (5 or 130), and BaCl₂ (20); internal solution contained NMDG (140), EGTA (10), HEPES (5), Tris (5); pH was adjusted to 7.3 with methanesulfonic acid. For Fig. 8, the external solution contained TEA-Cl (130), BaCl₂ (20), HEPES (5), pH 7.3, with TEA-OH); internal solution contained KCl or NaCl (140 mm), EGTA (5), HEPES (5), Tris (5), pH 7.3, with KOH or NaOH. Pipette resistances were typically 2-6 megaohms in the bath solution, and series resistance compensated up to 70%. Leak subtraction was conducted using a P/-4 protocol.

To measure current density, I_Ba was evoked by 50-ms pulses from a holding voltage of −100 mV to various voltages and normalized to the cell capacitance. I-V data were fitted with the Boltzmann equation: I = G_max × (V_m – E_rev)/(1 + exp(V_h – V_m)/k, where I is the measured current at each test voltage (V_m), V_h is the voltage of half-maximal activation, k is the slope factor, and G_max is the maximal conductance. Peak current density was determined by dividing the maximal I_Ba by the cell capacitance. Kinetic parameters for I_Ba activation (τ_act) and deactivation (τ_deact) were obtained by fitting the test current and tail current, respectively, with a single exponential function (y₀ + A (exp(−t/τ)), where y₀ is the offset (asymptote), t is time, τ is the time constant, and A is the amplitude. The voltage-dependence of τ_act and τ_deact was described by: y₀ + A (exp(−v/v)), where y₀ is the asymptote, v is voltage, v is the voltage constant, and A is the amplitude. Relative permeability of Ba²⁺ versus different monovalent cations (x) was calculated as: P_Ba/P_x = [x]_i/4[Ba²⁺]_o × exp (E_revF/RT){1 + exp (E_revF/RT)}. Data were analyzed offline with Igor Pro (Wavemetrics) or Origin Pro (OriginLab Corporation) software. Statistical analysis and preparation of graphs were performed using GraphPad Prism software. The data were initially analyzed for normality using the Shapiro–Wilk or D'Agostino-Pearson omnibus test. For parametric data, significant differences were determined by Student's t test. For nonparametric data, Mann-Whitney test was used. Significant differences in the curve fits of τ_act and τ_deact versus voltage relationships were determined by F tests. Data were incorporated into figures using GraphPad and Adobe Illustrator software. Unless otherwise indicated, averaged data represent mean ± S.E. from at least 3 independent transfections.

Biochemical analysis of cell-surface Ca_V1.4 protein

Transfected HEK293T cells were subject to cell-surface biotinylation and Western blotting as described previously (52). Cell-surface proteins were biotinylated according to the manufacturer's protocol. Briefly, cells were washed with ice-cold PBS (PBS, in mm: 2.5 KCl, 136 NaCl, 1.5 KH₂PO₄-Na₂HPO₄ 6.5, pH 7.4), prior to incubation with sulfo-NHS-SS-biotin (Thermo Scientific) for 30 min at 4 °C. The cells were then incubated with biotin quenching solution (in mm: 50 glycine, 2.5 CaCl₂, 1 MgCl₂, pH 7.4), scraped off the plate in PBS, pelleted by centrifugation, and resuspended in lysis buffer containing in mm: 150 NaCl, 25 Tris–HCl, pH 7.6, with 1% Nonidet P-40, 1% sodium deoxycholate, 0.1% SDS, and 0.5 phenylmethylsulfonyl fluoride and other protease inhibitors. After 10 min on ice, cell lysates were subject to centrifugation (16,000 × g for 10 min at 4 °C) and biotinylated proteins recovered with NeutrAvidin gel. The bound proteins were eluted in SDS-PAGE sample buffer (in mm: 58 Tris-Cl, 50 DTT, with 1.7% SDS, 5% glycerol, 0.002% bromphenol blue, pH 6.8) and subject to electrophoresis using Novex^TM WedgeWell^TM 4–20% Tris glycine gel (Invitrogen) and transfer to nitrocellulose blotting membranes.

For Western blotting, the membranes were incubated in blocking buffer containing milk (5%) in TBS-T (100 mm Tris–HCl, 0.15 m NaCl, 0.05% Tween 20) followed by incubation with the following antibodies diluted in blocking buffer: Ca_V1.4 (1:4,000 (48)); Na⁺/K⁺ ATPase (1:700, Developmental Studies Hybridoma Bank, University of Iowa, RRID:AB_2314847), GAPDH (1:10,000; Cell Signaling catalog number 14C10). Horseradish peroxidase (HRP)-conjugated secondary antibodies used were anti-rabbit HRP (1:3000; GE Healthcare catalog number NA934-1ML) and anti-mouse HRP (1:3000; GE Healthcare catalog number NA931V) followed by chemiluminescent detection (Thermo Scientific; SuperSignal West Pico catalog number 34080). The Western blotting signals were visualized with the Odyssey Fc Imaging System (LI-COR). The results shown were obtained from at least 3 independent experiments. Densitometric analysis was performed with Image Studio Lite software (LI-COR).

Data availability

All data relevant to this work are contained within this manuscript or available upon request.