Because Ca_V1.3 channels are alternatively spliced, we first sought to determine whether Ca_V1.3_S and Ca_V1.3_L channels are differentially regulated by [Ca²⁺]_i. Macroscopic currents were recorded from tsA-201 cells expressing either Ca_V1.3_S or Ca_V1.3_L channels in the presence of 2 mM Ba²⁺ or 2 mM Ca²⁺. Currents were activated by a depolarizing pulse (300 ms) from a holding potential of -80 mV to -10 mV. With Ba²⁺ in the external solution, membrane depolarization induced large Ca_V1.3_L currents that inactivated slowly (Figure 1A, center). Switching to a perfusion solution containing Ca²⁺ decreased the amplitude of Ca_V1.3_L currents by nearly 40% and increased the rate of inactivation. Like Ca_V1.3_L currents, Ca_V1.3_S currents inactivated faster when Ca²⁺ was used as a charge carrier however, in agreement with previous reports (Bock et al., 2011; Singh et al., 2008), we observed more pronounced CDI (defined as the difference between inactivation of I_Ba and I_Ca) in Ca_V1.3_S channels compared to the Ca_V1.3_L variant (Figure 1B, right versus Figure 1A, right). As discussed above, this difference in the magnitude of CDI has been attributed to the lack of the CTM domain in Ca_V1.3_S channels. Curiously, the amplitude of Ca_V1.3_S currents decreased to a lesser extent (only about 15% at -10 mV) upon changing the external solution from Ba²⁺ to Ca²⁺ (Figure 1B, center).

We investigated whether differences in the amplitude of elementary Ca_V1.3_L and Ca_V1.3_S channel currents could, at least in part, account for these disparities in macroscopic Ca²⁺ and Ba²⁺ currents. Single Ca_V1.3_L and Ca_V1.3_S channel currents were recorded from cell-attached patches with pipettes containing 20 mM Ca²⁺ or Ba²⁺. With Ca²⁺ as the charge carrier, the amplitudes of elementary Ca_V1.3_S and Ca_V1.3_L channel currents were similar. For example, at -30 mV they were -0.48 ± 0.07 and -0.49 ± 0.01 pA for Ca_V1.3_L and Ca_V1.3_S channels, respectively (Figures 1C–F). These values are in accordance with the unitary currents reported by Guia et al. for cardiac L-type channels using Ca²⁺ as charge carrier (Guia et al., 2001). Ensemble averages revealed currents that activated quickly and then inactivated likely due to Ca²⁺ and voltage-dependent mechanisms (Figure 1G). Furthermore, as is the case for other Ca_V1 channels, the single channel currents produced during the opening of Ca_V1.3_L and Ca_V1.3_S channels were both larger with Ba²⁺ as the charge carrier than with Ca²⁺ (-1.14 ± 0.02 pA and -1.10 ± 0.02 pA for Ca_V1.3_L and Ca_V1.3_S at -30 mV, respectively), but not significantly different from each other (Figure 1—figure supplement 1). The amplitude of the unitary currents with Ba²⁺ is also in accordance with previously reported values for these channels (Bock et al., 2011). These data suggest that the difference observed in the macroscopic currents between the Ca_V1.3_L and Ca_V1.3_S channels is not due to differences in unitary currents.

We then tested the hypothesis that Ca²⁺ enhances the activation of Ca_V1.3_S, but not Ca_V1.3_L channels by increasing the channel activity (NP_o). We performed a quantitative analysis of NP_o. The whole-cell current (I) is given by the equation I = i*N*P_o, where i is the amplitude of the elementary current, N is the number of functional channels, and P_o is the channel open probability. Since elementary Ca_V1.3_L and Ca_V1.3_S currents are larger when Ba²⁺ rather than Ca²⁺ is the charge carrier, for I_Ca to be similar to I_Ba, the NP_o of Ca_V1.3_S must be higher in the presence of Ca²⁺ than Ba²⁺. We estimated NP_o with Ca²⁺ and Ba²⁺ as charge carriers by dividing the amplitude of whole-cell Ca_V1.3_S and Ca_V1.3_L currents at -10 mV, by the values of unitary current. Our analysis relies on the assumption that the relatively larger peak of Ca_V1.3_S currents in the presence of Ca²⁺ was not due to faster activation kinetic of this channel than that of Ca_V1.3_L channels. This assumption is reasonable, as we found no significant difference in the activation time constants for Ca_V1.3_L and Ca_V1.3_S currents that were 1.60 ± 0.22 and 1.17 ± 0.051 ms (-10 mV, n = 6 for each channel, p = 0.112), respectively. We used elementary currents values recorded with 2 mM Ca²⁺ (-0.16 pA) and Ba²⁺ (-0.24 pA) at -10 mV (Guia et al., 2001). We found that Ca²⁺ ions entering the cell through the channels increased the NP_o nearly 1.5-fold for Ca_V1.3_S channels, but not at all for Ca_V1.3_L channels (Figures 1I and J). Thus, assuming that the number of functional Ca_V1.3 channels (N) in the membrane remained constant, a reasonable assumption given the short time lapse (~2 min) between recording I_Ba and I_Ca from the same cell, these data suggest that a Ca²⁺-dependent mechanism enhances inward Ca²⁺ currents by increasing the P_o of Ca_V1.3_S channels, but not that of Ca_V1.3_L channels.

One possible explanation for the Ca²⁺-influx–dependent increase in the P_o of Ca_V1.3_S channels is cooperative gating among channels in small clusters, as we have previously reported for Ca_V1.2 channels in cardiomyocytes and smooth muscle cells (Navedo et al., 2010). To test this possibility, we made optical recordings of individual Ca_V1.3_S-mediated Ca²⁺ influx events (called 'sparklets') in Ca_V1.3_L and Ca_V1.3_S-expressing tsA-201 cells loaded with 200 µM Rhod-2 using total internal reflection fluorescence (TIRF) microscopy. Ca_V1.3_S sparklets were recorded at a membrane potential of -80 mV in the presence of 20 mM external Ca²⁺; 10 mM EGTA was included in the patch pipette to confine the [Ca²⁺]_i signal to within ~1 µm of the point of Ca²⁺ entry (Zenisek et al., 2003). A quantal analysis of Ca_V1.3_S and Ca_V1.3_L sparklets revealed the presence of single-level (elementary) events with a mean amplitude of ~40 nM for both channels, in agreement with our previous study (Navedo et al., 2007). Interestingly, consistent with our single channel data, we found that multi-quantal sparklets, which presumably result from the simultaneous opening of several Ca_V1.3 channels, were more commonly observed in cells expressing Ca_V1.3_S than Ca_V1.3_L channels (Figure 2A). To determine whether channels in a cluster opened cooperatively or independently, we calculated the coupling coefficient (κ) among channels within a Ca_V1.3_S and Ca_V1.3_L sparklet site by applying a coupled Markov-chain model (Chung and Kennedy, 1996). Channels with κ > 0.1 were considered coupled (Navedo et al., 2010). Using this approach, we found that the average κ value for Ca_V1.3_S and Ca_V1.3_L channels was 0.21 ± 0.05 (n = 15) and 0.08 ± 0.04 (n = 12), respectively (Figure 2B). These results support the hypothesis that Ca_V1.3_S channels are more likely to undergo cooperative gating, generating persistent and greater Ca²⁺ influx than Ca_V1.3_L channels.

Figure 2

Download asset Open asset

If the signal for cooperative Ca_V1.3_S channel gating is a local increase in [Ca²⁺]_i, these channels must be in close proximity to one another. To test this hypothesis, we examined the spatial organization of endogenous Ca_V1.3 channels in hippocampal neurons using super-resolution localization microscopy (Figure 3A–D). Hippocampal neurons were immunostained against Ca_V1.3 using an antibody kindly provided by Dr. William Catterall and Dr. Ruth Westenbroek. This analysis showed that Ca_V1.3 channels form clusters occupying an average area of 3660 ± 80 nm² (n = 5). The antibody used in this study has been shown to not recognize the corresponding sequence of the closely related Ca_V1.2 α subunit in both, transfected cells and hippocampal tissue (Hell et al., 1993; 2013). However, we were not able to test the specificity of the antibody on Ca_V1.3-KO neurons and thus, pursued our analyses of channel clustering using a heterologous expression of Ca_V1.3 channels in tsA-201 cells.

Figure 3 with 2 supplements see all

Download asset Open asset

The specificity of the antibody in tsA-201 cells was tested by immunostaining untransfected and Ca_V1.3_S-transfected cells. No evidence of staining was observed in the untransfected cells (Figure 3—figure supplement 1A, n = 4). As our antibody cannot distinguish between Ca_V1.3 channels isoforms, we expressed Ca_V1.3_L or Ca_V1.3_S channels separately and found that both channel subtypes form clusters of similar size (Figure 3—figure supplement 1B–F). The mean areas of Ca_V1.3_L and Ca_V1.3_S channel clusters were 2543 ± 50 nm² and 2119 ± 73 nm², respectively (Figure 3—figure supplement 1G, n = 7). GSD images were acquired in the TIRF focal plane with a penetration depth of 130 nm.

To determine the number of channels within Ca_V1.3 channel clusters, we used step-photobleaching (Ulbrich and Isacoff, 2007), of expressed green fluorescent protein (GFP)-fused Ca_V1.3_S channels in hippocampal neurons (Figure 3E and F) and tsA-201 cells (Figure 3—figure supplement 1G and H). The rationale for using only Ca_V1.3_S-GFP channels in these studies is twofold. First, Ca_V1.3_S and not Ca_V1.3_L channels have the ability to undergo coupled activation. Second, the size of Ca_V1.3_L and Ca_V1.3_S puncta (at least in tsA-201 cells) is similar. Thus, Ca_V1.3_L and Ca_V1.3_S clusters are likely composed of the same number of channels. We identified and excited single Ca_V1.3_S channel clusters in hippocampal neurons and tsA-201 cells using TIRF microscopy with a penetration depth of 130 nm. Although some intracellular signal could be collected, the use of TIRF restricts the signal mainly to the plasmalemmal fraction of the channels. After continuous photobleaching we counted the number of step-wise decreases in fluorescence intensity of Ca_V1.3_S-GFP clusters. The majority (65%) of Ca_V1.3_S clusters underwent at least five stepwise decreases in fluorescence, with the remaining clusters showing fewer steps. A single photobleaching step was observed in only 4% of the Ca_V1.3_S clusters. The mean number of Ca_V1.3_S channels per cluster determined through sequential photobleaching was 8 ± 1 (n = 1105 clusters, 18 cells) in hippocampal neurons, this number is consistent with the average cluster area calculated from our super-resolution data and is similar to that reported in two different studies on the L-type channel distribution in hippocampal neurons using immunogold-labeling with electron microscopy and high-resolution immunofluorescence techniques (Leitch et al., 2009; Obermair et al., 2004). Step-photobleaching in tsA-201 cells revealed an average of 5 ± 1 Ca_V1.3_S channels per cluster (n = 585 clusters, 11 cells), consistent also with our super-resolution cluster area measurements (Figure 3—figure supplement 2).

Collectively, these results support our working hypothesis that multi-quantal events detected by imaging Ca²⁺ influx represent the simultaneous activity of multiple Ca_V1.3_S channels in a membrane microdomain. Furthermore, our data suggest that although channel clustering may be necessary for functional coupling of adjacent Ca_V1.3_S channels, physical clustering alone is not sufficient to induce functional coupling in Ca_V1.3_L channels.

To determine whether a physical interaction between Ca_V1.3_S channels induces I_Cafacilitation, we fused the C-terminus of Ca_V1.3_S and Ca_V1.3_L channels to an optogenetic light-induced dimerization system based on CIBN and CRY2 proteins (Kennedy et al., 2010). Blue-light illumination (488 nm), promotes CIBN and CRY2 fusion, which forces the C-termini of the attached channels to interact (Figure 4A).

Figure 4

Download asset Open asset

We transfected tsA-201 cells with Ca_V1.3_S-CIBN and Ca_V1.3_S-CRY2 or Ca_V1.3_L-CIBN and Ca_V1.3_L-CRY2 channels and measured I_Ca in response to a series of depolarizing voltage steps before and after a 30-s exposure to blue light (488 nm) (Figure 4B). As shown in Figure 4C, in cells expressing Ca_V1.3_S-CIBN and Ca_V1.3_S-CRY2 channels, I_Ca amplitude increased by 35% (n = 6, p<0.001) after illumination, whereas in cells expressing Ca_V1.3_L-CIBN and Ca_V1.3_L-CRY2 channels, there was no change in current amplitude (0.95 ± 0.03 n = 6). These results suggest that fusing adjacent channels at the tip of their C-tail increase the probability of functional coupling between Ca_V1.3_S adjoined channels but not between Ca_V1.3_L channels.

We used a second optogenetic approach that entailed fusing Ca_V1.3_S and Ca_V1.3_L channels with either the N- (VN) or C-terminus (VC) of the Venus fluorescent protein (Kodama and Hu, 2010; Shyu et al., 2006) (Figure 5A). Individual VN and VC fragments are non-fluorescent, but when they come into close proximity, they can reconstitute a full Venus protein, resulting in fluorescence. Venus reconstitution is irreversible and thus, the intensity of the fluorescence signal increases with time, proportionate to the number of Ca_V1.3_S/L-VN and Ca_V1.3_S/L-VC channels that physically associate.

Figure 5 with 2 supplements see all

Download asset Open asset

We simultaneously recorded I_Ca and obtained TIRF microscopic images from tsA-201 cells transfected with Ca_V1.3_S -VN and Ca_V1.3_S -VC. The first set of experiments was performed in the presence of 20 mM Ca²⁺ to mimic the experimental conditions used to record Ca²⁺ sparklets below. With 20 mM Ca²⁺ in the bathing solution, I_Ca and Venus fluorescence for Ca_V1.3_S channels increased in parallel in response to depolarizing pulses from a holding potential of -80 mV (Figure 5B). The normalized conductance and Venus fluorescence exhibited similar sigmoidal voltage dependencies (Figures 5C–D), which can be attributed to the irreversible nature of the Venus reconstitution, resulting in an increased number of fluorescent proteins with each successive depolarization as the open probability of Ca_V1.3_S channels increases.

Substituting 2 mM Ba²⁺ for 20 mM Ca²⁺ in the external solution led to a robust I_Ba during membrane depolarization, but caused no accompanying change in Venus fluorescence (Figures 5F–I). Importantly, Venus reconstitution was never observed in cells expressing Ca_V1.3_L-VN and Ca_V1.3_L-VC in (see Figure 8). These data indicate that Ca_V1.3_S channels, but not Ca_V1.3_L channels, can physically interact via their C-termini, that this association occurs in response to membrane depolarization, and that it is promoted by intracellular Ca²⁺.

Because Venus reconstitution is irreversible, once Ca_V1.3_S-VN and Ca_V1.3_S-VC channels fuse they must remain adjoined. Thus, we tested the hypothesis that Ca²⁺-induced fusion of Ca_V1.3_S-VN and Ca_V1.3_S-VC increases the activity of adjoined channels by recording Ca_V1.3_S sparklets at a holding potential of -80 mV, in the presence of Ba²⁺ or Ca²⁺, before and after applying the same depolarization protocol described above. Before depolarization, Ca_V1.3_S sparklet activity was very low (Figure 5J); after depolarization, Ca_V1.3_S sparklet activity (nP_s; Figure 5K) and sparklet density (Figure 5L) markedly increased in the presence of Ca²⁺, but not Ba²⁺, without a change in the amplitude of the elementary Ca²⁺ influx (Figure 5M).

To investigate the mechanism underlying the Ca²⁺-dependency of Ca_V1.3_S channel coupling further, we used the Venus system in conjunction with a mutagenesis approach, focusing on CaM, which is required for CDF and has been shown to bind Ca²⁺ and associate with the C-terminus of L-type Ca²⁺ channels (Zuhlke et al., 1999). tsA-201 cells were transfected with Ca_V1.3_S-VN/Ca_V1.3_S-VC and divided into three groups. Cells in the first group (controls) were dialyzed with a standard Cs⁺-based intracellular solution. Cells in the second group were dialyzed with a CaM-inhibitory peptide corresponding to a 15-aa fragment of the wild-type CaM-binding domain of myosin light chain kinase (MLCKp; 1 μM), which binds to CaM with high affinity (apparent dissociation constant, ~6 pM) in the presence of Ca²⁺ (Torok and Trentham, 1994) and has been used by others as a competitive inhibitor of CaM (Ciampa et al., 2011; Mercado et al., 2010; Piper and Large, 2004). The third group consisted of cells co-expressing a dominant-negative mutant form of CaM (CaM₁₂₃₄) that does not bind Ca²⁺ through its N- or C-terminal lobes. Dialysis of MLCKp or co-expression of CaM₁₂₃₄ prevented Ca_V1.3_S-VN and Ca_V1.3_S-VC fusion upon membrane depolarization (Figure 6A–D). Although MLCKp and CaM₁₂₃₄ were equally effective in preventing Venus reconstitution, they had differential effects on I_Ca inactivation (Figure 6E). Whereas the fraction of peak I_Ca remaining at 300 ms (r₃₀₀) in MLCKp-dialyzed cells (0.11 ± 0.03, n = 5) was similar to that of controls (0.14 ± 0.03, n = 5), the rate of inactivation of I_Ca was slower in cells expressing CaM₁₂₃₄, as reflected in a much higher r₃₀₀ value (0.75 ± 0.05, n = 5) (Figure 6F). Our interpretation of these findings is that the CaM molecules involved in CDI are distinct from those involved in functional coupling of Ca_V1.3_S channels. The results suggest that CaM molecules that mediate coupling could be both attached to the channels or recruited to the C-terminus during depolarization. This could explain why they are accessible to MLCKp blockade, unlike the CaM molecules that mediate CDI, which are tethered to the IQ domain of the channels (Pitt et al., 2001).

Figure 6

Download asset Open asset

To extend this analysis, we recorded Ca_V1.3_S sparklets in control and CaM₁₂₃₄ cells before and after depolarization to +20 mV (Figure 6G and H). Although multi-quantal Ca_V1.3_S sparklets were observed in control cells at rest (e.g., Figure 6G, trace 2 left), sparklets in CaM₁₂₃₄ cells prior to depolarization were long, single-quantal events (Figure 6H, traces 1 and 2, left). These long Ca_V1.3_S sparklets were likely due to decreased CDI of Ca_V1.3_S channels in cells expressing CaM₁₂₃₄ (Yang et al., 2014). Importantly, the overall Ca_V1.3_S sparklet density was lower in cells expressing CaM₁₂₃₄ than WT, suggesting that Apo-CaM does not increase Ca_V1.3_S channel activity.

The coupling coefficient for Ca_V1.3_S sparklet sites in WT cells was 0.07 ± 06, whereas that in CaM₁₂₃₄ cells was 0.02 ± 0.02 (n = 5). Membrane depolarization increased the coupling coefficient of Ca_V1.3_S channels within multi-quantal sparklet sites in control cells (0.18 ± 0.03, n = 5), but not in CaM₁₂₃₄ cells (0.08 ± 0.05, n = 5). The opposing effects of CaM₁₂₃₄ on Ca_V1.3_S sparklets — longer, but decoupled — resulted in Ca_V1.3_S sparklet activity before depolarization that was similar in control (nP_S = 0.07 ± 0.05; p>0.05) and CaM₁₂₃₄ (nP_S = 0.12 ± 0.09) cells (Figure 6I). However, depolarization increased Ca_V1.3 sparklet activity nearly 7-fold in control cells, but had no effect on sparklet activity in CaM₁₂₃₄ cells. Taken together with the I_Ca and Venus reconstitution results described above, these data strongly suggest that Ca²⁺ binding to CaM is required for physical and functional Ca_V1.3_S-to-Ca_V1.3_S channel coupling.

Finally, to establish the molecular mechanism by which CaM might mediate these effects, we mutated different CaM-binding domains of the Ca_V1.3_S. L-type channels have two binding sites for CaM in their C-terminus: the IQ domain (aa K1601-Q1621) and the pre-IQ domain (aa T1545-Q1587) (Fallon et al., 2009). These sites have different affinities for CaM; whereas CaM is “pre-associated” and binds tightly to the IQ domain, the association of CaM with the pre-IQ domain is seemingly weaker and likely transient. To determine which of these sites is necessary for CaM-mediated Ca_V1.3_S coupling, we generated two mutants of the Ca_V1.3_S-VN and Ca_V1.3_S-VC channels. The first contained a single point mutation I1608E (Ca_V1.3-I1608E) that disrupts CaM binding to the IQ domain (Zuhlke et al., 1999), and the second contained a triple mutation (L1569A, V1572A, and W1577E; Ca_V1.3_S-AAE) that prevents CaM binding to the pre-IQ domain and anti-parallel coiled-coil arrangement of the pre-IQ domains (Fallon et al., 2009) (Figure 7A). Ca_V1.3_S-I1608E channels showed a slower rate of inactivation than control and Ca_V1.3_S-AAE channels (Figure 7B), consistent with the lack of CDI. Interestingly, Ca_V1.3_S-I1608E-VN/VC, but not Ca_V1.3_S-AAE VN/VC, which were capable of reconstituting Venus during membrane depolarization (Figure 7C–G). These data suggest that binding of CaM to the pre-IQ domain is critically involved in Ca_V1.3_S channel coupling during membrane depolarization and further supports our previous assertion that the CaM pool involved in CDI is distinct from that involved in channel coupling.

Figure 7

Download asset Open asset

We investigated one of the possible molecular mechanisms that prevent functional coupling in Ca_V1.3_L channels. Singh et al (2008) showed that a Ca_V1.3_L mutant lacking the last 116 amino acids of the C-terminus (Ca_V 1.3_L∆116) had voltage-dependencies and kinetics similar to those of Ca_V1.3_S channels. The Ca_V1.3_L∆116 channels lack the distal regulatory domain (DCRD) that binds the proximal regulatory domain (PCRD) located downstream to the pre-IQ and IQ domains of the channel (Figure 8A). Interaction of these two regulatory domains interferes with CaM binding to the IQ domain and results in a reduction in CDI (Bock et al., 2011; Singh et al., 2008). If the DCRD interferes also with the binding of CaM to the pre-IQ domain, which we propose is important for channel-to-channel coupling, we would expect that removing the DCRD would allow coupling between Ca_V1.3_L channels. Thus, we investigated whether or not Ca_V1.3_L∆116 channels are capable of undergoing Ca²⁺-driven physical interactions. For these experiments, we created Ca_V1.3_L∆116 channels fused to the split Venus system. As expected, CDI of I_Ca was faster in cells expressing the Ca_V1.3_L∆116 channels compared to the full length Ca_V1.3_L channels (p<0.05; Figure 8B and C).

Figure 8 with 1 supplement see all

Download asset Open asset

We found that, with 2 mM Ca²⁺ in the bathing solution, cells expressing Ca_V1.3_L∆116 channels failed to reconstitute Venus fluorescence, similar to what we observed for the full length Ca_V1.3_L channel (Figure 8D–G). Even increasing the extracellular Ca²⁺ concentration to 20 mM was not enough to induce coupling in the Ca_V1.3_L∆116 channels (Figure 8—figure supplement 1, n = 5, p=0.245 -60 mV vs 20 mV). This result suggests that deletion of DCRD is not sufficient to allow coupling of Ca_V1.3_L channels. As Ca_V1.3_L∆116 channels still have a C-terminus (396 aa) that is considerably longer than that of Ca_V1.3_S channels, it is possible that another domain inside this region might be responsible for occluding the binding of CaM to the pre-IQ domain, and preventing Ca_V1.3_L∆116 channel coupling. Differential folding between the short and long C-terminal could be another explanation for the ability of Ca_V1.3_S channel-to-channel coupling during depolarization-induced Ca²⁺ entry.

We extended our investigation of the functional consequences of Ca_V1.3_S channel coupling to cultured rat hippocampal neurons. We began by recording spontaneous Ca_V1.3 sparklets from these cells. As in tsA-201 cells expressing Ca_V1.3_S channels, Ca²⁺ sparklets in neurons were restricted to specific sites and had multi-quantal amplitudes resulting from the simultaneous opening and closing of multiple channels. The quantal unit of Ca²⁺ influx was about 40 nM. The coupling coefficient (κ) of sparklet sites ranged from 0.14 to 0.33. The average κ value was 0.23 ± 0.04 (n = 6). Importantly, sparklet site activity was decreased by application of 300 nM of the dihydropyridine antagonist isradipine and completely eliminated when the concentration of the drug was increased to 10 µM (Figure 9A and B). This is consistent with the hypothesis that sparklets in hippocampal neurons were produced by L-type calcium channels. Both Ca_V1.2 and Ca_V1.3 channels are expressed in hippocampal neurons (Hell et al., 1993) and although it is impossible to distinguish between these two L-type Ca²⁺ channels using either electrophysiological or pharmacological methods, it has been shown that Ca_V1.3 channels have a reduced sensitivity to dihydropyridines compared to Ca_V1.2 channels (Lipscombe et al., 2004; Xu and Lipscombe, 2001). A previous study by Koschak et al found that 100% of Ca_V1.2 channels but only ~60% of Ca_V1.3 channels are inhibited by 300 nM isradipine (Koschak et al., 2001), given this, it is reasonable to assume that any sparklet remaining after application of 300 nM isradipine is more likely to be generated by Ca_V1.3channels than Ca_V1.2. In addition, near by 25% of the L-type current in hippocampal neurons is carried by Ca_V1.3 channels (Moosmang et al., 2005), this proportion is in agreement with the remaining sparklet site density we observed after the treatment with 300 nM isradipine (Figure 9C).

Figure 9

Download asset Open asset

As was the case for tsA-201 cells expressing Ca_V1.3_S channels, we found that conditioning membrane depolarization to 0 mV increased sparklet activity nearly 3-fold in the hippocampal neurons and induced persistent sparklet activity (Figure 9D and E). The coupling coefficient increased from 0.15 ± 0.04 to 0.43 ± 0.12 after membrane depolarization (n = 8). These results support the hypothesis that L-type channels undergo cooperative gating, generating persistent Ca²⁺ influx in hippocampal neurons. The persistent L-type channels sparklet activity evoked by membrane depolarization in the presence of Ca²⁺ bears a striking resemblance to the persistent cationic currents observed in several types of neurons (Fransen et al., 2006; Major and Tank, 2004; Moritz et al., 2007; Powers and Binder, 2001).

We found that the somatic and dendritic membranes of neurons expressing both Ca_V1.3_S-VN and Ca_V1.3_S-VC channels displayed prominent Venus fluorescence (Figure 10A), indicating fusion of Ca_V1.3_S-VN and Ca_V1.3_S-VC channel pairs and functional Venus reconstitution. Since Venus reconstitution is irreversible and neurons have spontaneous electrical activity, this observation suggests that the C-termini of Ca_V1.3_S channels make contact during normal neuronal firing. Although spontaneous self-assembly between Venus subunits might conceivably drive an interaction that would not otherwise occur, the improved Venus system used here has a mutation (I152L) that minimizes non-specific interactions (Kodama and Hu, 2010).

Figure 10 with 1 supplement see all

Download asset Open asset

We analyzed super resolution images of hippocampal neurons transfected with Ca_V1.3_S-VN and Ca_V1.3_S-VC and found that expression of these channels increased the number of Ca_V1.3 channel clusters in the cells, but not the size of the clusters (Figure 10—figure supplemental 1). These results indicate that the spontaneous fusion of Ca_V1.3_S channels was not a consequence of their being over-expressed in the plasma membranes of the hippocampal neurons.

A testable prediction of these observations is that Ca_V1.3_S channel fusion—forced coupling—should augment the inward I_Ca and consequently increase neural excitability and firing rate. Figure 10B shows the spontaneous action potentials recorded from a neuron transfected with Ca_V1.3_S-CIBN and Ca_V1.3_S-CRY2, before (right trace) and after (center trace) exposure to 488 nm light. Forced coupling of Ca_V1.3_S-CIBN and Ca_V1.3_S-CRY2 channel pairs increased the average firing rate by about 40% (p<0.01, Figure 10C). 488 nm illumination produced no effect on firing rate in control hippocampal neurons transfected only with Ca_V1.3_S-CIBN channels (0.90 ± 0.04, N = 4). Adding the L-type calcium channel blocker nifedipine 10 μM to the bathing solution abolished all AP activity after light-induced fusion of Ca_V1.3_S-CIBN and Ca_V1.3_S-CRY2 (Figure 10B, left trace and 10C). Together, these results highlight the crucial role that Ca_V1.3_S channels play in regulating the action potential firing in hippocampal neurons.

The pcDNA clones of the rat Ca_V1.3 isoforms were obtained from Addgene. We used two Addgene Ca_V1.3_S plasmids#26576 and #49333 (Xu and Lipscombe, 2001), the first one containing two single point substitutions, a glycine by a serine at position 244 and an alanine by a valine at position 1104. We found that they encoded channels with similar voltage-dependencies of activation and rate of activation (Figure 5—figure supplement 2A and B). The Ca_V1.3_S channels encoded by these plasmids were also capable of undergoing Ca²⁺-dependent dimerization and functional coupling (Figure 5—figure supplement 2C–F). On the basis of these data, we concluded that the G244S and A1104V substitutions in the plasmid #26576 are functionally silent with respect to voltage dependence, rate of activation and have no effect on the capacity of adjacent Ca_V1.3_S channels to undergo allosteric interactions. A prior study by Lieb et al. also reported no contribution of these mutations to the functional properties of Ca_V1.3_S channels (Lieb et al., 2012),. Both plasmids #26576 and #49333 were used to express and design the functional Ca_V1.3_S constructs used in this study. Ca_V1.3_L was also obtained from Addgene (plasmid #49332, (Xu and Lipscombe, 2001)). Auxiliary subunits Ca_Vβ₃ and Ca_Vα₂δ₁ were gift of Dr. Diane Lipscombe’s laboratory; Brown University, Providence, RI). The C-terminus of Ca_V1.3_S andCa_V1.3_L channels were fused to different proteins depending on the experimental approach. For bimolecular fluorescence complementation, they were fused to either VN or VC fragments of the Venus protein (Kodama and Hu, 2010) (Dr. Chang-Deng Hu, Addgene plasmids 27097, 22011); for photobleaching experiments, they were fused to monomeric GFP (mGFP_A206K), amplified from the pCGFP-EU vector (Kawate and Gouaux, 2006), kindly provided by Dr. Eric Goaux (Oregon Health and Science University, Portland, OR); and for the light induced cryptochrome system, they were fused with either CRY2 or CIBN (generous gifts from Dr. Pietro Di Camilli, Yale University, New Haven, CT). The CaM₁₂₃₄ plasmid was a generous gift from Dr. Johannes Hell (University of California, Davis, CA).

The tsA-201 cell line, used for heterologous expression of the constructs listed above, was maintained in Dulbecco’s modified Eagle medium supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin antibiotic solution. Cells were transiently transfected using jetPEI transfection reagent (Polyplus Transfection, New York, NY) and plated onto 25-mm coverslips (0.13–0.17-mm thick). Successfully transfected cells were identified on the basis of turbo red fluorescent protein (tRFP) fluorescence. Imaging and electrophysiology experiments were performed within 48 hr of transfection.

Hippocampal neurons were prepared from newborn (P1) Sprague-Dawley rats in accordance with University of Washington (UW) guidelines. Animals were decapitated and their tissue harvested according to a protocol approved by the UW Institutional Animal Care and Use Committee (IACUC). The hippocampi of six rat pups were dissected and cut into small pieces in cold dissection medium consisting of 12 mM MgSO₄ and 0.3% bovine serum albumen (BSA) in Hank’s balanced salt solution (HBSS). The pieces were incubated for 30 min at 37ºC in dissection medium containing 25 U/ml papain. The digested tissue was washed with warm Neuronal medium consisting of Minimal Essential Medium (MEM) supplemented with 10% horse serum, 2% B27, 25 mM HEPES, 20 mM glucose, 2 mM GlutaMAX, 1 mM sodium pyruvate, and 1% penicillin/streptomycin antibiotic solution. The tissue was then gently homogenized in fresh Neuronal medium using a long Pasteur pipette. Neurons were plated on poly-D-lysine-coated coverslips (0.2 mg/ml for 2 hr) at a density of 2 x 10⁵ cells/coverslip. After incubating neurons at 35ºC for 24 hr, unattached cells were removed by replacing the medium with fresh Neuronal medium. Every 5 day, one-third of the medium was replaced with fresh Neuronal medium supplemented with the anti-mitotic agents fluorodeoxyuridine (20 μM) and uridine (50 μM).

After 14 d in culture, rat hippocampal neurons were transfected using 2.4 μg DNA, 4.8 μl of Lipofectamine LTX and 4.8 μl PLUS reagent (Life Technologies, Grand Island, NY) in a final volume of 1 ml of a 1:1 mixture of Neuronal medium and Opti-MEM. After 1 hr of incubation, the medium was replaced with fresh Neuronal medium. Experiments were performed within 48 hr of transfection. Successful transfection was corroborated by detection of tRFP.

Ca²⁺ currents were recorded using the whole-cell configuration of the patch-clamp technique in voltage-clamp mode. Currents were sampled at a frequency of 10 kHz and low-pass filtered at 2 kHz using an Axopatch 200B amplifier. During the experiments, tsA-201 cells were superfused with a solution containing 5 mM CsCl, 10 mM HEPES, 10 mM glucose, 140 mM N-methyl-D-glucamine, 1 mM MgCl₂ and 2 mM CaCl₂ or 2 mM BaCl₂, depending on the experiment. pH was adjusted to 7.4 with HCl. For the experiments using 20 mM CaCl₂, the osmolarity was adjusted by decreasing the concentration of NMDG to 113 mM. Borosilicate patch pipettes with resistances of 3–6 MΩ were filled with an internal solution containing 87 mM cesium aspartate (CsAsp), 20 mM CsCl, 1 mM MgCl₂, 10 mM HEPES, 10 mM EGTA and 5 mM MgATP, adjusted to pH 7.2 with CsOH. A voltage offset of 10 mV, attributable to the liquid junction potential of these solutions, was corrected offline. Current–voltage relationships were obtained by subjecting cells to a series of 300-ms depolarizing pulses from a holding potential of -80 mV to test potentials ranging from -70 to +50 mV. The voltage dependence of channel activation (G/G_max) was obtained from the resultant currents by converting them to conductances using the equation, G = I_Ca/(test pulse potential – reversal potential of I_Ca); normalized G/G_max was plotted as a function of test potential. We found that the reversal potential for Ca_V1.3_L (+51 ± 4 mV) and Ca_V1.3_S (+54 ± 3 mV) currents in the presence of 2 mM Ca²⁺ were not significantly different (p=0.49). The same was true with 2 mM Ba²⁺ in the bath (Ca_V1.3_L = +60 ± 5 mV; Ca_V1.3_S = +65 ± 2 mV; p=0.31). Thus, while E_rev is similar in Ca_V1.3_L and Ca_V1.3_S channels with the same permeating ion, the reversal potential of Ca²⁺ and Ba²⁺ currents through these channels differ. In our patch clamp experiments in which the external solution was switched from Ba²⁺ to Ca²⁺, 2-min intervals were inserted between the onset of the whole cell configuration to the first pulse and again after switching the external solution, to rule out any effect of the run-up of the I_Ca (Tiaho et al., 1993).

Single-channel currents (i_Ca) were recorded from tsA-201 using the cell-attached configuration. Cells were superfused with a high K⁺ solution to fix the membrane potential at ∼0 mV. The bathing solution had the following composition: 145 mM KCl, 2 mM MgCl₂, 0.1 mM CaCl₂, 10 mM HEPES and 10 mM glucose; pH was adjusted to 7.3 with KOH. Pipettes were filled with a solution containing 10 mM HEPES and either 20 mM CaCl₂ or 20 mM BaCl₂; pH was adjusted to 7.2 with CsOH. The dihydropyridine agonist BayK-8644 (500 nM) was included in the pipette solution to promote longer channel open times. A voltage-step protocol from a holding potential of −80 mV to a depolarized potential of −30 mV was used to elicit currents. The single-channel event-detection algorithm of pClamp 10.2 was used to measure single-channel opening amplitudes. We generated all-points histograms from our cell-attached patch-clamp recordings. These histograms were fit using Prism 5.0a software (GraphPad software Inc. La Jolla, CA) with a multi-Gaussian function that included a quantal (q; i.e., elementary current) parameter using the following equation:

where N is the number of events, a and b are constants, i_Ca is the amplitude of the current measured and q is the quantal elementary current of the channel.

Spontaneous discharge of cultured hippocampal neurons was recorded in current-clamp mode using the perforated-patch configuration. Neurons were superfused with a solution containing 140 mM NaCl, 5 mM KCl, 10 mM HEPES, 10 mM glucose, 1 mM MgCl₂, 2 mM CaCl₂ and 1 mM Na-pyruvate, adjusted to pH 7.4 with NaOH. Borosilicate patch pipettes with resistances of 3–6 MΩ were filled with an internal solution containing 5 mM NaCl, 140 mM KCl, 15 mM HEPES and 7 mM MgATP, adjusted to pH 7.2 with KOH; 60 μM amphotericin B was added to the solution before starting the recording. Series resistances lower than 30 MΩ were obtained within 5 min of seal formation. The sampling frequency was 10 kHz filtered at 2 kHz.

For immunostaining Ca_V1.3 in tsA-201 cells or hippocampal neurons, cells were fixed by incubating in phosphate-buffered saline (PBS) containing 3% paraformaldehyde and 0.1% glutaraldehyde for 15 min. After washing with PBS, cells were incubated with 50 mM glycine at 4ºC for 10 min (aldehyde reduction), washed again with PBS, and blocked by incubating with 20% SEA BLOCK (Thermo Scientific) and 0.25% v/v Triton X-100 in PBS (blocking buffer) for 1 hr. The cells were incubated overnight at 4ºC with primary antibodies recognizing the residues 809 to 825 located at the intracellular II-III loop of the Ca_V1.3 channel (DNKVTIDDYQEEAEDKD, rabbit; provided by Drs. William Catterall and Ruth Westenbroek) and the neuronal marker MAP2 (mouse; Abcam), diluted in blocking buffer to a concentration of 10 μg/ml. Cells were then washed with PBS, incubated for 1 hr with Alexa Fluor 647-conjugated donkey anti-rabbit (2 µg/ml; Molecular Probes) and Alexa Fluor 488-conjugated chicken anti-mouse (2 µg/ml; Molecular Probes) secondary antibodies, and washed again with PBS. Our antibody was designed to bind to the intracellular loop linking the 2^nd and 3^rd membrane domains of Ca_V1.3. It cannot distinguish between Ca_V1.3_S and Ca_V1.3_L.

For super-resolution microscopy, coverslips were mounted on microscope slides with a round cavity using MEA-GLOX imaging buffer (NeoLab Migge Laborbedarf-Vertriebs GmbH, Germany) and sealed with Twinsil (Picodent, Germany). The imaging buffer contained 10 mM MEA, 0.56 mg/ml glucose oxidase, 34 μg/ml catalase, and 10% w/v glucose in TN buffer (50 mM Tris-HCl pH 8, 10 mM NaCl).

A super resolution ground-state depletion system (SR-GSD, Leica) based on stochastic single-molecule localization was used to generate super-resolution images of Ca_V1.3 in hippocampal neurons and tsA-201 cells. The Leica SR-GSD system was equipped with high-power lasers (488 nm, 1.4 kW/cm²; 532 nm, 2.1 kW/cm²; 642 nm, 2.1 kW/cm²) and an additional 30 mW, 405 nm laser. Images were obtained using a 160× HCX Plan-Apochromat (NA 1.43) oil-immersion lens and an EMCCD camera (iXon3 897; Andor Technology). For all experiments, the camera was running in frame-transfer mode at a frame rate of 100 Hz (10 ms exposure time). Fluorescence was detected through Leica high-power TIRF filter cubes (488 HP-T, 532 HP-T, 642 HP-T) with emission band-pass filters of 505–605 nm, 550–650 nm, and 660–760 nm.

Super-resolution localization images of Ca_V1.3 channel distribution were reconstructed using the coordinates of centroids obtained by fitting single-molecule fluorescence signals with a 2D Gaussian function using LASAF software (Leica). A total of 50,000–100,000 images were used to construct the images. The localization accuracy of the system is limited by the statistical noise of photon counting. Thus, assuming the point-spread functions are Gaussian, the precision of localization is proportional to DLR/√N, where DLR is the diffraction-limited resolution of a fluorophore and N is the average number of detected photons per switching event (Dempsey et al., 2011; Folling et al., 2008). Accordingly, we estimated a lateral localization accuracy of 16 nm for Alexa 647 (~1900 detected photons per switching cycle). Ca_V1.3 cluster size was determined using binary masks of the images in ImageJ software (NIH).

The number of Ca_V1.3_S channels in clusters along the surface membrane was estimated using a single-molecule bleaching approach similar to that described by Ulbrich and Isacoff (2007). Briefly, TIRF images of tsA-201 or hippocampal neurons expressing Ca_V1.3_S channels fused to the monomeric GFP were acquired using our Leica GSD microscope in TIRF mode. Cells were fixed with 4% paraformaldehyde for 10 min prior to the acquisition. Images were acquired using an oil immersion 160x objective (NA 1.43) and an Andor iXON EMCCD camera. GFP was excited with a 488 nm laser and image stacks of 2000 frames were acquired at 30 Hz. During analysis, the first 5 images of the stack were averaged and a rolling-ball background subtraction was applied using ImageJ (NIH). This image was then low-pass filtered with a 2 pixel cut-off and high-pass filtered with a 5 pixel cut-off. Thresholding was then applied to identify connected regions of pixels that were above threshold. The ImageJ plugin ‘Time Series Analyzer v2.0’ was used to select 4x4 pixel ROIs, centered on the peak pixel in each spot. These ROIs were used to plot Z-axis intensity profiles (where z is time) of the entire image stack to manually detect the bleaching steps.

The spontaneous interaction between the C-terminus of Ca_V1.3_S channels was assessed using the split Venus system (Shyu et al., 2006). In this approach, Ca_V1.3_S channels were fused to either the VN fragments (N_1–154) or the VC fragment (C_155–238) of the Venus fluorescent protein. The Venus protein emits fluorescence only when the two fragments are close enough to interact and reconstitute the whole protein, providing a measure of the proximity between the C-terminus of the Ca_V1.3_S channels. Ca_V1.3_S -VN and/or Ca_V1.3_S -VC constructs were expressed at a 1:1 ratio in hippocampal neurons; tRFP fluorescence was used as an indicator of successful transfection. Confocal images were acquired with a Fluoview FV1000 microscope (Olympus, Center Valley, PA) equipped with a UPlanS-Apochromat 60× (NA 1.2) water-immersion objective. The Venus protein was excited using a 488 nm laser line. The calcium dependence of Ca_V1.3 spontaneous interactions was studied in tsA-201 cells expressing Ca_V1.3_S -VC alone or Ca_V1.3_S -VN and Ca_V1.3_S -VC in a 1:1 ratio. Using the whole-cell configuration of the patch-clamp technique, cells were depolarized from a holding voltage of -80 mV to test potentials ranging from -60 to +60 mV, administered as 9-s pulses. Maturation of newly reconstituted Venus protein takes some time, hence the long depolarizing pulse (Nagai et al., 2002). Images were acquired at a frequency of 100 Hz during each depolarizing pulse using a through-the-lens TIRF microscope built around an inverted microscope (IX-70; Olympus) equipped with a Plan-Apochromat (60×; NA 1.49) oil-immersion lens (Olympus) and an electron-multiplying charge-coupled device (EMCCD) camera (iXON; Andor Technology, UK). The last 10 images of each stack were averaged, and total fluorescence was quantified using ImageJ software (NIH). The images were pseudo-colored using the ‘red hot’ lookup table in ImageJ. F₀ was calculated by dividing the total fluorescence for each voltage by the initial fluorescence at -80 mV. The change in F/F₀ was plotted against voltage and compared to G/G_max curves constructed as described in the electrophysiology section. The Ca²⁺ dependence of Venus reconstitution was tested in cells bathed in an external solution containing 20 mM Ca²⁺ or 2 mM Ba²⁺.

Ca_V1.3_S sparklets were recorded using the TIRF microscopy system described above. [Ca²⁺]_i was monitored by adding the Ca²⁺ indicator Rhod-2 (200 µM) to the pipette solution and exciting with a 568 nm laser. The much slower Ca²⁺ buffer EGTA (10 mM) was included with the relatively fast Ca²⁺ indicator, Rhod-2, to restrict Ca²⁺ signals to the vicinity of the Ca²⁺ entry source. Sparklets were detected in tsA-201 cells expressing Ca_V1.3_S channels. The driving force for Ca²⁺ entry was increased by holding the membrane potential at -80 mV using the whole-cell configuration of the patch-clamp technique.

TIRF images were acquired at a frequency of 100 Hz using TILL Image software. Sparklets were detected and analyzed using custom software written in MATLAB (Source code 1). Fluorescence intensity values were converted to nanomolar units as described previously (Navedo et al., 2007). Event amplitude histograms were generated from [Ca²⁺]_i records and fitted with a multicomponent Gaussian function. We determined the activity of sparklets by calculating the nP_s of each sparklet site, where n is the number of quantal levels and P_s is the probability that a quantal sparklet event is active. A detailed description of this analysis can be found in Navedo et al. (Navedo et al., 2006).

In split Venus experiments, sparklets were acquired at -80 mV before and after a depolarizing protocol from -60 to +60 mV with 9-s pulses. Sparklet images were always acquired in a solution containing 20 mM Ca²⁺, whereas depolarization protocols were run in 20 mM Ca²⁺ or 2 mM Ba²⁺ to compare the effect of Ca²⁺-dependent Ca_V1.3_S dimerization on sparklet activity.

The degree of coupling between Ca_V1.3 Ca²⁺ sparklet sites was assessed by further analyzing sparklet recordings using a binary coupled Markov chain model, as first described by Chung and Kennedy (1996) and previously employed by our group (Navedo et al., 2010; Cheng et al., 2011; Dixon et al., 2012). The custom program (Source code 2), written in the MATLAB language, assigns a coupling-coefficient (κ) to each record, where κ can range from 0 (purely independently gating channels) to 1 (fully coupled channels). Elementary event amplitudes were set at 38 nM.

Light-induced dimerization of Ca_V1.3 channels was accomplished by fusing Ca_V1.3 channels with the optogenetic light-induced dimerization system based on CRY2 and CIB1 proteins of Arabidopsis thaliana (Kennedy et al., 2010). Upon blue-light illumination (488 nm), CRY2 absorbs a photon, causing a conformational change in one of its domains that promotes binding to the N-terminal region of CIB1 (CIBN). Ca_V1.3-CIBN and/or Ca_V1.3-CRY2 constructs were expressed in a 1:1 ratio in hippocampal neurons. Forty-eight hours after transfection, spontaneous action potential firing was recorded in current-clamp mode. One minute after initiating recordings, neurons were exposed to a 30-s blue light pulse to induce dimerization, and the changes in the firing pattern were measured. In tsA-201 cells, Ca²⁺ currents were recorded before and after light illumination in response to a 20-ms depolarizing pulse at +10 mV from a holding potential of -80 mV. Experiments were performed on a Nikon (Eclipse TE2000-S) Swept Field confocal system controlled with Elements software and equipped with a 488 nm laser line and a Plan Apo 60× 1.45 N.A. oil-immersion objective.

Data were collected from at least five independent experiments in each series. The data included in this paper were normally distributed. Accordingly, parametric statistics were performed and mean ± SEM are used to provide a description of the data set. Student’s t-test was used to test for statistical significance using Prism 5.0 a software (GraphPad software Inc. La Jolla, CA). We decided, a priori, that p values <0.05 were indicative a statistical significance difference between or among data groups. The number of cells used for each experiment and p values are detailed in each figure legend. Paired t- tests were used to test for statistical significance of paired observations. Comparisons between three or more conditions were made by one-way ANOVA test using Prism software.

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

We thank Dr. Bertil Hille and Dr. Manuel F Navedo for comments and Dr. Joshua Vaughan for help with super-resolution analysis. Dr. William Catterall and Dr. Ruth E Westenbroek provided antibodies against Ca_V1.3.

Animal experimentation: This study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All of the animals were handled according to approved institutional animal care and use committee (IACUC) protocols (#3374-01) of the University of Washington and (#18896) of the University of California, Davis.

1. Received: March 3, 2016
2. Accepted: April 9, 2016
3. Version of Record published: May 17, 2016 (version 1)

© 2016, Moreno et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.