A single CaVcan reconstitute both trafficking and macroscopic conductance of voltage-dependent calcium channels
http://www.100md.com
《生理学报》
1 Calcium Signals Laboratory, Department of Biomedical Engineering, Johns Hopkins University School of Medicine, Baltimore, MD 21205, USA
Abstract
Voltage-dependent calcium-channel subunits (CaV) strongly modulate pore-forming 1 subunits by trafficking channel complexes to the plasma membrane and enhancing channel open probability (Po). Despite their central role, it is unclear whether binding of a single CaV, or multiple CaVs, to an 1 subunit governs the two distinct functions. Conventional experiments utilizing coexpression of 1 and CaVsubunits have been unable to resolve the ambiguity due to difficulties in establishing their stoichiometry in functional channels. Here, we unambiguously establish a 1: 1 stoichiometry by covalently linking CaV2b to the carboxyl terminus of 1C (CaV1.2), creating 1C·2b. Recombinant L-type channels reconstituted in HEK 293 cells with 1C·2b supported whole-cell currents to the same extent as channels reconstituted via coexpression of the individual subunits. Analysis of gating charge showed 1C·2b fully restored channel trafficking to the plasma membrane. Co-transfecting CaV2a with 1C·2b had little further impact on function. To rule out the possibility that fused CaV2b was interacting in trans with neighbouring 1 molecules, 1C·2b was cotransfected with 1B (CaV2.2), and pharmacological block with nimodipine showed an absence of 1B trafficking. These results establish that association of a single CaVwith a pore-forming 1 subunit captures the functional essence of HVA calcium channels, and introduce 1–CaVfusion proteins as a powerful new tool to probe structure–function mechanisms.
Introduction
High-voltage-activated (HVA) calcium channels are hetero-multimeric protein complexes comprised minimally of a pore-forming 1-subunit (CaV1.1–1.4; CaV2.1–2.3), and auxiliary (CaV1–CaV4), 2 (2-1–2-3), and sometimes subunits (Catterall, 2000). These proteins control Ca2+ fluxes that underlie critical biological functions ranging from muscle contraction to hormone and neurotransmitter release. Auxiliary CaVsubunits play a crucial role in the functional maturation of HVA calcium channels: each of the four CaVs can interact promiscuously with any of the seven known HVA calcium-channel 1 subunits to promote membrane trafficking of the associated 1 (Chien et al. 1995; Brice et al. 1997; Yamaguchi et al. 2000; Colecraft et al. 2002; Takahashi et al. 2004), to increase channel Po (Singer et al. 1991; Costantin et al. 1998; Hullin et al. 2003; Takahashi et al. 2004), and to produce hyperpolarizing shifts in the voltage-dependence of channel activation (Singer et al. 1991; Perez-Reyes et al. 1992; De Waard et al. 1994; Jones et al. 1998).
The powerful impact of CaVs on HVA calcium-channel operation logically focuses attention on the interaction between 1 and CaVsubunits as a promising locus to modify calcium-channel behaviour. Indeed, this interaction is the target for physiological down-regulation of calcium currents by the Rem, Rad, Kir/Gem (RGK) family of Ras-like GTPases (Beguin et al. 2001; Finlin et al. 2003), and serves as the basis for high-throughput identification of novel pharmacological inhibitors of HVA calcium-channel activity (Young et al. 1998). A deeper mechanistic understanding of the HVA 1–CaVinteraction is important for both appreciating its physiological ramifications, as well as potentially hastening the discovery of new calcium-channel therapeutics.
A fundamental ambiguity of HVA calcium-channel structure relates to the functional stoichiometry of 1 to CaV(Birnbaumer et al. 1998; Canti et al. 2001; Jones, 2002; Dolphin, 2003). Specifically, it is unclear whether the fully functional channel is comprised of an 1 subunit associated with single or multiple CaVsubunits. Biochemical studies initially established that 1 and CaVsubunits copurify in an equimolar ratio (Witcher et al. 1993); the uncertainty arises from studies which indicate that trafficking and gating-modulation are separable functions dependent on distinct CaVconcentrations. In Xenopus oocytes, expression of 1 without exogenous CaVleads to recovery of low-amplitude currents that activate at relatively depolarized potentials, compared to channels obtained by coexpressing 1 and CaV(Singer et al. 1991; Perez-Reyes et al. 1992; Birnbaumer et al. 1998). However, the presumed 1-alone currents are abolished by antisense oligonucleotides against a subsequently identified endogenous Xenopus CaV(CaVXO) (Tareilus et al. 1997). These results suggested that there is sufficient endogenous CaVXO in the Xenopus expression system to fully traffick 1 subunits to the plasma membrane, but not to modulate gating (Tareilus et al. 1997; Birnbaumer et al. 1998). In agreement, careful titration of the amount of CaV3 coexpressed with CaV2.2 in Xenopus oocytes indicated that the concentration of CaV3 required to traffick channels to the plasma membrane is nearly an order of magnitude less than that required to modulate channel gating (Canti et al. 2001). Furthermore, acute application of purified CaVprotein was found to modulate the gating of pre-existing calcium channels in the plasma membrane (Yamaguchi et al. 1998; Garcia et al. 2002).
These data are consistent with two mutually exclusive interpretations of the functional stoichiometry of the 1–CaVinteraction (Birnbaumer et al. 1998; Jones, 2002; Dolphin, 2003). In the ‘single-CaV-binding’ model, CaVbinds to 1 at the endoplasmic reticulum (ER) to promote translocation to the plasma membrane. At low CaVconcentrations, reversible unbinding of CaVfrom 1 favours a significant steady-state fraction of low-activity CaV-less channels in the plasma membrane. At high CaVconcentrations, the equilibrium is shifted towards high-activity (CaV-bound) channels (Fig. 1A). Thus, a single CaVis responsible for both trafficking and enhancing the channel Po. Alternatively, in the ‘multiple-CaV-binding’ model, trafficking and gating-modulation are mediated by distinct CaVsubunits (Fig. 1B). As before, a CaVbinds to an 1 to promote channel trafficking. After insertion into the plasma membrane, the high-affinity interaction between 1 and CaVpersists. By contrast to the single-CaVmodel, these 1 + CaVchannels exhibit low gating activity, and binding of a second CaVto a distinct lower affinity site is required for high gating activity. Distinguishing between these two models is fundamental to mechanistic understanding of CaVfunction, and has important implications for rational manipulation of the 1–CaVinteraction to regulate calcium-channel activity.
A, the single-CaV-binding model. A single CaVsubunit binds to the 1 subunit at the endoplasmic reticulum (ER) membrane resulting in trafficking of the 1–complex to the plasma membrane. At the plasma membrane, a reversible 1: 1 interaction between 1 and CaVsubunits switches the channel between low- and high-Po modes. B, the multiple-CaV-binding model. A CaVbinding to an 1 subunit at the ER membrane is also responsible for trafficking the 1-channel complex to the plasma membrane. However, this CaVdoes not dissociate from the channel complex, which gates in a low-Po mode. Binding of an additional CaVsubunit (or subunits) switches the channel complex to a high-activity gating mode.
A key obstacle to discriminating between the two models is that in traditional coexpression studies it is practically impossible to distinguish the precise stoichiometry of 1 and CaVsubunits present in functional channels. Here, we take the novel approach of covalently linking an 1C subunit (CaV1.2) to a CaV2b subunit, thus ensuring a 1: 1 stoichiometry in transfected HEK 293 cells. We find that a single CaVis necessary and sufficient to completely reconstitute both channel trafficking and macroscopic conductance. We conclude that the single-CaVmodel captures the functional essence of HVA calcium channels.
Methods
Molecular biology
A truncation mutant of 1C (1C[1905]), and 1C–CaVfusion proteins (1C[1905]·2b, 1C·2b, 1C·2b[P235R]) were constructed using rabbit 1C/pGW1 (Wei et al. 1991) and human CaV2b/pAd CIG (Takahashi et al. 2003) expression vectors as backbones. To generate 1C[1905]/pGW1, we inserted a stop codon after residue E1905 of 1C by polymerase chain reaction (PCR) amplifying 1C/pGW1 using the following upstream and downstream primers, respectively: (primer 1) CACCATCTGGATGAATTTAAAAGAATC and (primer 2) TCTAGACTACTCCTCCGGGGGCGCCAGTTG. The resulting PCR fragment was ligated into 1C/pGW1 using BstEII/XbaI sites, generating 1C[1905]/pGW1.
The 1C[1905]·2b construct was generated by overlap extension PCR as follows. 1C[1905]/pGW1 was used as a template in a PCR reaction with primer 1 and the following downstream primer: (primer 3) CTCCTCCGGGGGCGCCAGTTGTCG. A second PCR using human CaV2b/pAd CIG as template was carried out using the following upstream and downstream primers, respectively: (primer 4) CTGGCGCGCCCGGAGGAGGGTGGTGGTGGTATGCTTGACAGACGCCTT and (primer 5) GCGTACAAGGGTACCGCAATACCG. The two PCR fragments were used as templates in an overlap-extension PCR reaction using primers 1 and 5. A BstEII/Xba1 cassette from the overlap extension PCR product was transferred to 1C[1905]/pGW1, generating 1C[1905]·2b/pGW1 in which CaV2b is attached to 1C[1905] via a 4-glycine linker.
The 1C·2b construct was generated using a strategy similar to that described above for 1C[1905]·2b. The 1C/pGW1 construct was used as template in a PCR reaction with the following upstream and downstream primers, respectively: (primer 6) GCAGTGGCGGGCCTGAGTCCCCTC, and (primer 7) CAGGCTGCTGACGCCGGCCCTGCG. Human CaV2b/pAd CIG was used as template in a second PCR using the following upstream primer: (primer 8) GCCGGCGTCAGCAGCCTGGGTGGTGGTGGTATGCTTGACAGACGCCTTATAG in conjunction with primer 5. The two PCR products were used in an overlap extension reaction with primers 6 and 5. An RsrII/XbaI cassette from the overlap extension PCR product was transferred into 1C/pGW1, generating 1C·2b/pGW1, in which CaV2b is attached to the carboxyl terminus of 1C via a 4-glycine linker.
The point mutant fusion protein, 1C·2b[P235R], was generated by overlap extension PCR using 1C·2b as template. The first PCR reaction was performed using primer 8 and the following downstream primer: (primer 9) CTTCAGAGAACGGCCCACTAGGAC (mutated base underlined). The second PCR reaction was performed using the following upstream and downstream primers, respectively: (primer 10) GTCCTAGTGGGCCGTTCTCTGAAG and (primer 11) AGGTCGACTCTAGAGCGGCCGCCA. The two PCR products were used in an overlap extension reaction, and a SfuI/XbaI cassette from the resulting product transferred into 1C·2b/pGW1 to generate 1C·2b[P235R]/pGW1. Pfu polymerase (Strategene, La Jolla, CA, USA) was used to increase fidelity in all PCR reactions. All PCR products were verified by sequencing.
Cell culture and transient channel expression
Low passage number human embryonic kidney (HEK 293) cell cultures (< 20 passages) were maintained as previously described (Brody et al. 1997). Cells were transiently transfected 24 h after passage using the calcium-phosphate precipitation method with 8 μg each of CMV expression plasmids encoding the indicated calcium-channel subunits [rabbit 1C (Wei et al. 1991); 1C[1905]; human CaV2b (Takahashi et al. 2003); 1C·2b; and rat 2 (Tomlinson et al. 1993)] and 3 μg of T antigen.
Electrophysiology
Electrophysiological experiments were performed as previously described (Takahashi et al. 2004). Whole-cell recordings were performed at room temperature, 2–3 days after transfection, using an EPC8 patch-clamp amplifier (HEKA Electronics, Lambrecht/Pfalz, Germany) controlled by PULSE software. Patch pipettes were fashioned from 1.5-mm thin-wall glass with filament (WPI, Waltham, MA). Patch pipettes typically had a series resistance of 2–4 M, compensated 50–70%, when filled with an intracellular solution containing (mM): 135 caesium methanesulphonate (Cs-MeSO3), 5 CsCl, 5 EGTA, 1 MgCl2, 4 MgATP (freshly added on the day of patch-clamp experiments), and 10 Hepes (pH 7.3, adjusted with CsOH). Cells were continuously perfused with an external solution containing (mM): 140 tetraethylammonium (TEA) MeSO3, 10 Hepes, 5 BaCl2 (pH 7.4, adjusted with TEA-OH). The standard electrophysiological protocol consisted of a family of 20-ms test pulse depolarizations (from –40 to +120 mV) used to evoke currents from a –90-mV holding potential, followed by a repolarization to –50 mV to measure tail currents. Current signals were sampled at 25 kHz and filtered at 10 kHz, with leak and capacitive transients subtracted using a P/8 protocol. Traces were acquired at a repetition interval of 15 s.
Data and statistical analysis
Data were analysed off-line using PULSEFIT (HEKA Electronics). Peak current amplitudes during the test pulse were normalized by cell capacitance to provide the current density (Jpeak). Integration of ON gating currents recorded at the reversal potential to provide gating charge (QON) was performed in Origin (OriginLab Corp., Northampton, MA, USA). J–V relationships were fitted to the following equation:
where J is the whole-cell current density (pA pF–1), G is the specific conductance (pA pF–1 mV–1), Vrev is the reversal potential (mV), V1/2 is the voltage of half-maximal activation (mV), and k is a slope factor (mV). Statistical analyses were performed in Microsoft Excel using built-in functions. Pooled data are represented as means ± S.E.M.; P-values were calculated using Student's two-tailed t test, with P < 0.05 considered statistically significant.
Results
Robust modulation of truncated CaV1.2 (1C[1905]) by coexpressed CaV2b in HEK 293 cells
The traditional experimental approach towards investigating the functional effects of CaVon HVA calcium channels has been to compare the properties of recombinant channels reconstituted by expressing HVA calcium-channel 1 subunits, either with or without CaVin heterologous expression systems. Figure 2 shows the characteristic powerful impact of CaVon the functional expression of recombinant calcium channels, as reported by the conventional coexpression approach. Transfection of HEK 293 cells with cDNA for a truncated CaV1.2 (1C[1905]) subunit and 2 resulted in the recovery of small-amplitude recombinant L-type calcium-channel currents in 70% of selected cells (Fig. 2A and B). Co-expression of CaV2b with 1C[1905]/2 dramatically increased the amplitude of recorded calcium-channel currents across the entire range of membrane voltages as indicated by exemplar traces (Fig. 2C and D), and population peak current density vs. voltage (Jpeak–V; Fig. 2E) and tail-current vs. voltage (Itail–V; Fig. 2F) curves. Overall, peak current density obtained with 1C[1905] + CaV2b channels were over 10-fold larger than those obtained with CaV-less 1C[1905] (current density at 0 mV, Jpeak,0mV = 4.05 ± 0.39 pA pF–1, n = 7, for 1C[1905]/2; Jpeak,0mV = 47.9 ± 11.8 pA pF–1, n = 6, for 1C[1905]/CaV2b/2; P < 0.05).
A, schematic diagram of the 1C[1905] subunit. B, exemplar whole-cell currents obtained from a HEK 293 cell transfected with 1C[1905] + 2. C, schematic diagram of 1C[1905] + CaV2b. D, exemplar whole-cell currents obtained from a cell cotransfected with 1C[1905] + CaV2b + 2. Co-expression with CaVresults in a marked increase in whole-cell current amplitude compared to 1C[1905] + 2 channels. E, plots of peak current density vs. test pulse voltage (Jpeak–V) for cells transfected with 1C[1905] + 2 () or 1C[1905] + CaV2b + 2 (). F, plots of tail current amplitude vs. voltage (Itail–V) from channels expressed without () or with () cotransfected CaV2b. G, top, exemplar gating currents obtained from recombinant 1C[1905] channels expressed without (left) or with (right) cotransfected CaV2b. G, bottom, time integral of the gating current obtained at the reversal potential (QON) obtained for channels reconstituted without or with CaV2b.
It has been previously established that CaVs elevate whole-cell currents by increasing both the number of channels expressed in the plasma membrane (N) (Chien et al. 1995; Brice et al. 1997; Gao et al. 1999), and the channel Po (Costantin et al. 1998; Hullin et al. 2003). In our electrophysiological protocol, N can be estimated by calculating the integral of the ON gating current (Qmax = N x q, where q is single-channel gating charge) isolated at the reversal potential (usually +50 mV) (Jones et al. 1998; Colecraft et al. 2002; Takahashi et al. 2004). For cells transfected with 1C[1905]/2, gating currents were essentially unresolved (Fig. 2A and G; QON = 0.05 ± 0.04 fC pF–1, n = 7), clearly indicating a dearth of channels with movable voltage sensors at the plasma membrane. By contrast, coexpressing 1C[1905] with CaV2b resulted in relatively large and clearly resolved gating currents (Fig. 2D and G; QON = 5.7 ± 2.3 fC pF–1, n = 7), thus demonstrating that CaV2b promotes targeting of 1C[1905] to the plasma membrane. Overall, these results establish the baseline functional properties of recombinant L-type channels obtained via cotransfecting HEK 293 cells with 1C[1905] and CaV2b.
Covalently linking 1C[1905] and CaV2b recapitulates robust functional expression of recombinant L-type channels
Unfortunately, traditional coexpression experiments as described in Fig. 2 cannot be used to determine the functional stoichiometry of 1 and CaVspecies, since their relative expression levels cannot be tightly controlled. To control the stoichiometry between the subunits, coding sequence for CaV2b was fused to the carboxyl terminus of 1C[1905] by subcloning. The resulting plasmid construct, when transfected and expressed in HEK 293 cells would result in a fusion protein (1C[1905]·2b) in which a 1: 1 ratio of 1C[1905] to CaVis assured (Fig. 3A). Whole-cell records from these fusion experiments showed that 1C[1905]·2b clearly supports robust currents, as demonstrated by exemplar traces (Fig. 3B) and population Jpeak–V (Fig. 3C) and Itail–V (Fig. 3D) curves. Importantly, these experiments prove that CaV2b was still functional in the fused-channel configuration.
A, schematic diagram of the 1C[1905]·2b construct. B, exemplar whole-cell currents obtained at different test potentials for a cell transfected with 1C[1905]·2b + 2. C, population Jpeak–V plot obtained for 1C[1905]·2b + 2 channels (). Data for 1C[1905] + 2 channels have been reproduced (grey trace) to facilitate visual comparison of differences in peak current density amplitude. D, population Itail–V plot for 1C[1905]·2b + 2 channels (). The corresponding data for 1C[1905] + 2 channels are reproduced (grey trace). E, QON values.
How did currents obtained with 1C[1905]·2b channels compare with those obtained by the customary approach of coexpressing 1C[1905] and CaV2b as independent subunitsIn fact, whole-cell current amplitudes were essentially identical to those obtained with 1C[1905]/CaV2b/2 (Jpeak,0mV = 33.9 ± 4.8 pA pF–1, n = 13, for 1C[1905]·2b/2 channels; P = 0.27, compared to 1C[1905]/CaV2b/2 channels). Moreover, the fused CaV2b promoted channel trafficking to the same extent as free CaV2b, as determined by gating current analyses (Fig. 3E; QON = 4.9 ± 0.7 fC pF–1, n = 13, for 1C[1905]·2b/2; P = 0.75, compared to 1C[1905]/CaV2b/2 channels). Overall, the similarity in whole-cell current amplitude and QON suggests that CaV2b has virtually identical effects on microscopic channel properties (N and Po) irrespective of whether it is directly fused to 1C[1905], or operating independently.
Modulation of full-length CaV1.2 by free and covalently linked CaV2b
We conducted the initial series of experiments with the truncated CaV1.2 subunit, 1C[1905], to maximize the chances of obtaining a positive result. We theorized that the shortened carboxyl terminus of 1C[1905] would facilitate the interaction of the fused CaV2b with the primary CaVinteraction site in the cytoplasmic linker between domains I and II of the 1C subunit (I–II linker) (Pragnell et al. 1994). Having established the efficacy of the fused CaV2b in the truncated channel configuration (Fig. 3), we sought to determine whether the experimental paradigm yielded equally robust results in the context of full-length 1C.
Similar to observations with the truncated 1C[1905] subunit, coexpression of full-length 1C and CaV2b/2 (Fig. 4D–F) resulted in significantly larger whole-cell currents than obtained with 1C/2 (Fig. 4A–C). Fits of the J–V relations to a modified Boltzmann relation revealed that CaV2b resulted in a four-fold increase in the macroscopic channel conductance (Table 1). Moreover, channels reconstituted with free CaV2b displayed another classic hallmark of CaVmodulation compared to 1C/2: namely a hyperpolarizing shift in the voltage-dependence of channel activation (as gauged by V1/2-values; Table 1).
A, schematic diagram of full-length 1C. B, exemplar currents obtained from cell transfected with 1C + 2. C, population Jpeak–V plot for 1C + 2 channels. D–F, data for 1C + CaV2b + 2 channels. Same format as for A–C. Data for 1C + 2 channels in C are reproduced in F, I and L (grey trace) to facilitate direct visual comparisons. G–I, data for 1C·2b + 2; format same as above. J–L, data for 1C·2b[P235R] + 2; same format as above for A–C.
Reassuringly, whole-cell currents produced by 1C·2b/2-reconstituted channels (Fig. 4G–I) were also four-fold larger than obtained with CaV-less 1C/2 channels, and essentially identical to 1C/CaV2b/2 currents (Table 1). Moreover, channel trafficking was promoted to the same extent by CaV2b irrespective of whether it was free or fused, as indicated by similar QON values that were significantly greater than observed with 1C/2-reconstituted channels (Table 1). The quantitative J–V analyses did reveal an important difference between channels reconstituted with free vs. fused CaV2b. The 1C·2b/2 channels did not recapitulate the hyperpolarizing shift in the voltage-dependence of channel activation (Table 1, P < 0.01 when comparing V1/2-values for 1C + CaV2b + 2, 1C + 2, and 1C·2b + 2 channels by one-way ANOVA). This distinction could reflect an intrinsic property of the fused 1C·2b protein itself, or could signal the requirement of a second CaVto reconstitute the shift-in-activation gating function. Results provided in the next section appear more consistent with the former interpretation. Overall, the finding that robust currents could be successfully reconstituted with both 1C[1905]·2b and 1C·2b reveals that the design criteria for creating functional HVA calcium channels by fusing CaVs to the carboxyl terminus of 1 subunits are quite relaxed; the fused CaVappears equally effective when placed at different positions along the carboxyl terminus. By contrast, fusing CaV2b directly to the amino-terminus of 1C did not yield robust currents (not shown), suggesting that geometric constraints may preclude the formation of functional channels under this condition. Alternatively, amino-terminus-fused CaVconstructs may require more extensive optimization to produce functional channels (e.g. varying linker lengths) than appears necessary with attachment to the carboxyl terminus.
To discount the possibility that observed channel modulation might be due to nonspecific effects of attaching CaV2b to the 1C carboxyl terminus, we tested the effect of a proline to arginine (P235R) mutation in the fused CaV2b subunit, 1C·2b[P235R] (Fig. 4J–L). The analogous mutation in other CaVsubunits has been shown to disrupt binding to the 1 interaction domain (AID), and ablates the capacity of CaVto modulate 1 subunits in cotransfection experiments (De Waard et al. 1994; Takahashi et al. 2004), possibly due to derangement of the three-dimensional structure of the CaV1-binding pocket (ABP) (Chen et al. 2004; Opatowsky et al. 2004; Van Petegem et al. 2004). Reassuringly, 1C·2b[P235R] yielded whole-cell currents that were indistinguishable from those obtained with CaV-less 1C/2 (Fig. 4J–L and Table 1), clearly ruling out any unanticipated confounding effects generated by the fused CaV2b subunit.
Additional CaVs do not appreciably enhance 1C·2b channel currents
The results thus far indicated that a single CaV2b fused to 1C was necessary and sufficient to fully recapitulate CaVeffects on channel trafficking and macroscopic conductance (Figs 3 and 4). Could the presence of additional CaVsubunits further enhance the channel modulationWe directly examined this question by coexpressing 1C·2b/2 with CaV2a and recording whole-cell currents (Fig. 5A–C). Previous studies have established that CaV2a targets autonomously to the plasma membrane and robustly modulates recombinant L-type Ca2+ channel currents. Therefore, CaV2a was used in these coexpression experiments because we reasoned that its high effective local concentration at the membrane would provide the most favourable circumstance to observe a functional effect due to a second CaVsubunit. Exemplar whole-cell current traces for 1C·2b/CaV2a/2 were similar to those obtained for 1C·2b/2. Population data confirmed that there was no significant difference in macroscopic conductance between 1C·2b currents in the presence or absence of CaV2a (Fig. 5C and Table 1). By contrast, coexpressing 1C·2b[P235R] with CaV2a (Fig. 5D–F) resulted in significantly enhanced whole-cell current amplitudes compared to 1C·2b[P235R] (Fig. 5F and Table 1). Importantly, the presence of free CaV2a did not result in a significant change in the voltage-dependence of activation in 1C·2b/2 channels (Table 1; V1/2 = –6.1 ± 2.1 for 1C/2; V1/2 = –6.4 ± 1.5 mV for 1C·2b/2; V1/2 = –9.9 ± 1.8 mV for 1C·2b/CaV2a/2 channels, P = 0.28 by one-way ANOVA). This suggests that the relatively right-shifted voltage-dependence of activation of 1C·2b/2 channels reflects an intrinsic property of the fusion protein, and does not signal the requirement for a second CaVto reconstitute this functional property. Together, these results raised confidence that the fused wild-type CaV2b subunits were sufficient to fully modulate channel properties, and did not require the collaboration of additional CaVsubunits.
A, schematic diagram of 1C·2b + CaV2a. B, exemplar whole-cell currents from a cell transfected with 1C·2b + CaV2a + 2. C, population Jpeak–V relationship for 1C·2b + CaV2a + 2 channels (). Data for 1C·2b + 2 are reproduced to permit direct visual comparison (grey trace). D, schematic diagram of 1C·2b [P235R] + CaV2a. E, exemplar whole-cell currents from 1C·2b[P235R] + CaV2a + 2 channels. F, Jpeak–V relationship for 1C.2b + CaV2a + 2 channels (). Data for 1C.2b[P235R] + 2 are reproduced (grey trace).
Ruling out the presence of confounding trans interactions between distinct 1C·2b fusion-protein molecules
Our interpretation of the data obtained with 1C·2b as indicating that a single CaVis necessary and sufficient to fully modulate channel trafficking and macroscopic conductance hinges critically on the assumption that the fused CaVs act solely on the 1C subunits to which they are attached. This assumption might not be true if the fused CaVcould act in trans on a neighbouring 1C molecule. To evaluate the extent to which such intermolecular interactions took place, we coexpressed 1C·2b with 1B (Fig. 6). Our strategy was based on the idea that in the absence of free CaVthe only avenue for appreciable 1B transport to the membrane would be via an intermolecular interaction with the fused CaV2b subunit. Pharmacological block with nimodipine could then be used to distinguish relative amounts of 1C (L-type) and 1B (N-type) channels at the plasma membrane.
A, exemplar 1C.2b + 2 currents in the absence (black trace) and presence (grey trace) of 5 μM nimodipine. B, time course of nimodipine effect on exemplar current. Arrows denote time of nimodipine addition. C, bar graph showing degree of nimodipine inhibition in population of 1C.2b + 2 currents. D–F, data for 1B + CaV2b + 2 channels; same format as panels A–C. G–I, data for 1B + 1C.2b + 2 channels; same format as panels A–C. J–L, data for 1B + CaV2b + 1C.2b + 2 channels; same format as panels A–C.
As expected for recombinant L-type currents, channels reconstituted with 1C·2b/2 were strongly inhibited by the dihydropyridine antagonist, nimodipine (Fig. 6A–C). For these channels, 5 μM nimodipine inhibited 90% of the whole-cell current shortly after exposure (Fig. 6A–C). By contrast, 5 μM nimodipine had minimal effect on channels recorded from cells transfected with 1B + CaV2b (Fig. 6D–F). In cells transfected with 1C·2b/1B/2, application of 5 μM nimodipine inhibited whole-cell currents by 90% (Fig. 6G–I), identical to what was observed with 1C·2b. As a control to ensure that nimodipine-insensitive currents can be observed when a mixed population of 1B and 1C·2b channels are present at the membrane, we transfected cells with 1B/CaV2b and 1C·2b/2. In this regime, 5 μM nimodipine had an intermediate effect, inhibiting whole-cell currents by 50% on average (Fig. 6J–L). Together, these result provided compelling evidence against any significant contribution of intermolecular interactions involving fused CaV2b subunits.
Discussion
We investigated the functional stoichiometry of HVA calcium-channel 1 and subunits to address a long-standing debate over whether a fully functional HVA calcium channel is comprised of an 1 subunit associated with a single, or multiple CaVsubunits. Our results indicate that a 1: 1 ratio of 1 and CaVcaptures the functional essence of mature HVA channels. We discuss the significance of our findings in the context of previous results relevant to the issue of the stoichiometry of the calcium-channel 1–subunit interaction.
Previous studies on the functional stoichiometry of calcium-channel 1 and subunits
Biochemical studies of HVA calcium channels initially established that 1 and subunits copurified in an equimolar ratio, suggesting a 1: 1 stoichiometry of these proteins in native calcium-channel complexes (Witcher et al. 1993). Structure–function studies and recent high-resolution crystal structures show that CaVs bind to a conserved stretch of amino acids, termed 1 interaction domain (AID) located in the cytoplasmic region linking the first and second transmembrane domains of HVA 1 subunits (De Waard et al. 1994; Pragnell et al. 1994; Chen et al. 2004; Opatowsky et al. 2004; Van Petegem et al. 2004). For their part, the core region of CaVs contains two protein interaction motifs—a src homology 3 (SH3) domain, and a guanylate kinase-like (GK) domain (Chen et al. 2004; Opatowsky et al. 2004; Van Petegem et al. 2004), both of which are necessary for CaVfunction (Opatowsky et al. 2003; McGee et al. 2004; Takahashi et al. 2004). The GK domain exhibits a high-affinity interaction with the AID peptide.
However, the assumption of a 1: 1 1 to CaVstoichiometry was challenged by electrophysiological studies of recombinant channels expressed in Xenopus oocytes. Calcium-channel 1-subunit RNA injected alone into Xenopus oocytes results in low-amplitude whole-cell currents that activate at relatively depolarized potentials compared to those coexpressed with CaV(Singer et al. 1991; Perez-Reyes et al. 1992; Birnbaumer et al. 1998). Such low-level currents were attributed to CaV-less ‘1-alone’ channels. However, Xenopus oocytes were subsequently found to possess an endogenous subunit (XO). Antisense inhibition of this protein ablated so-called 1-alone currents (Tareilus et al. 1997), indicating that the CaVXO participated in the trafficking of the non-native 1 subunits. The question, then, was: why did channels obtained by injecting Xenopus oocytes with 1 alone exhibit clear-cut deficiencies in their gating properties despite the presence of CaVXOTwo mutually exclusive scenarios were proposed that could account for these data, as presented in Fig. 1 (Tareilus et al. 1997; Birnbaumer et al. 1998; Jones, 2002).
Several studies have affirmed the independence of the CaVtrafficking and gating roles without resolving the fundamental issue of whether the two functions are mediated by a single, or multiple CaVsubunits. Dolphin and colleagues (Canti et al. 2001) explicitly demonstrated concentration-dependent effects for CaVon recombinant N-type channels by carefully titrating the amount of CaV3 expression in Xenopus oocytes (Canti et al. 2001). Relatively low CaV3 levels (Kd = 17 nM) were sufficient to increase whole-cell conductance, whereas substantially higher concentrations (Kd = 120 nM) were required to elicit gating effects. Gerster et al. used a mutation of a conserved tyrosine in the AID (Y467S in 1C), previously identified to disrupt the high-affinity 1–CaVinteraction (Pragnell et al. 1994), to examine the separate trafficking and gating effects. They found that CaVs no longer trafficked 1C(Y467S) to the membrane upon cotransfection in tsA201 cells; but single-channel experiments indicated that CaVs still increased Po in 1C(Y467S) channels (Gerster et al. 1999). More recently, truncated splice variants of CaVs that lack a GK domain have been found to increase single-channel Po, while being unable to reconstitute the CaVtrafficking function (Hullin et al. 2003; Cohen et al. 2004). These data suggest the existence of secondary 1–CaVinteraction sites beyond the high-affinity interaction between the -binding pocket (ABP) of the GK domain and the AID. Indeed, lower affinity interactions between some CaVsubunits and the carboxyl termini of 1A and 1E have been identified in biochemical experiments (Qin et al. 1997; Birnbaumer et al. 1998; Walker et al. 1998, 1999; Walker & De Waard, 1998). However, none of these experiments could discriminate whether CaVs modulate channel trafficking and gating according to a single- or multiple-CaV-binding model (Jones, 2002).
Acutely applied CaVsubunits have also been shown to affect the gating activity of channels pre-existing in the plasma membrane. In one study, purified CaV3 protein was introduced into Xenopus oocytes previously injected with 1C cRNA. It was found that CaV3 produced effects on the gating of channels pre-existing at the membrane (leftward shift in the voltage-dependence of channel activation, increased current amplitude presumably due to elevated single-channel Po) long before increasing trafficking of 1C subunits from intracellular sites to the membrane (Yamaguchi et al. 1998). Similarly, acute introduction (via patch pipette dialysis) of purified CaV1a into spherical vesicles derived from skeletal muscle plasma membrane resulted in up-regulation of whole-cell current amplitude without any changes in gating current size (Garcia et al. 2002). These results indicate that under these experimental conditions, acute application of CaVs can modify the gating properties of channels pre-existing at the membrane. Unfortunately, these results could also be well explained by either the single- or multiple-CaV-binding models (Jones, 2002).
In light of these previous results, our finding that fusing a 1 subunit with a single CaVcaptures the functional essence of channels permits an unambiguous interpretation of the structure–function relationship of the calcium-channel 1–CaVinteraction. Although we do not directly measure channel Po in this study, we infer from the similarities in channel trafficking (as gauged by QON) and macroscopic conductance that channels reconstituted with 1C/CaV2b/2 and 1C·2b have similar Po-values. Future single-channel studies will provide more direct insights into microscopic gating properties of 1C·2b vs. 1C/CaV2b-reconstituted L-type calcium channels. From the current data, we cannot unambiguously rule out the notion that a second CaVcould possibly modulate some gating properties of the channel. Overall, the refined interpretations afforded by our results suggest new physiological dimensions to the 1–CaVinteraction that we discuss next.
Reversible unbinding of calcium-channel 1–CaVsubunits
Our results are consistent with a single-CaV-binding model, and in conjunction with previously published studies, an important implication of this model is that at low CaVconcentrations reversible unbinding of CaVensures an appreciable fraction of CaV-less 1 channels in the plasma membrane. The notion that CaVs can reversibly unbind from 1 subunits suggests that CaVmodulation of channel gating could be a more dynamic process than previously appreciated. Because CaV-less channels have a lower Po than CaV-bound channels, reversible unbinding could serve to physiologically switch calcium channels between low- and high-activity modes of signalling. Recently, it was reported that in invertebrate Lymnaea stagnalis neurones, CaVs (LCaV) are associated with LCaV2 channels in mature neurones (Spafford et al. 2004). However, LCaVwas physically uncoupled from LCaV2 channels located in the leading edge of neuronal growth cones (Spafford et al. 2004). Moreover, such uncoupled LCaV2 channels were shown to be important for growth cone projections. Reversible unbinding of CaVfrom 1 subunits could feature prominently in the appearance of such CaV-less channels at the membrane, and could thus serve to significantly enrich the physiologic dimensions of calcium-channel signalling in cells.
Utility of 1–CaVfusion-proteins as tools to probe calcium-channel structure–function relationships
Beyond the use of 1C·2b to resolve the functional stoichiometry of HVA calcium channels as demonstrated here, 1–fusion proteins have tremendous potential to elucidate other currently challenging questions relating to the structure–function of HVA calcium channels. A key example relates to the fact that in many excitable cells multiple HVA calcium-channel 1 and CaVsubunits coexist, and appear to interact in a rather promiscuous manner (Witcher et al. 1995; Scott et al. 1996; Ludwig et al. 1997). What is the physiological significance of the molecularly diverse combinations of calcium-channel 1–subunits in cellsDo distinct 1–combinations transduce unique biological responsesSuch questions are difficult to address given the multiplicity of calcium-channel 1 and CaVsubtypes in cells, and the promiscuity of their interactions. The 1–fusion-protein approach could provide a novel avenue to investigate the physiological significance of unambiguously identified 1–combinations.
References
Beguin P, Nagashima K, Gonoi T, Shibasaki T, Takahashi K, Kashima Y, Ozaki N, Geering K, Iwanaga T & Seino S (2001). Regulation of Ca2+ channel expression at the cell surface by the small G-protein kir/Gem. Nature 411, 701–706.
Birnbaumer L, Qin N, Olcese R, Tareilus E, Platano D, Costantin J & Stefani E (1998). Structures and functions of calcium channel subunits. J Bioenerg Biomembr 30, 357–375.
Brice NL, Berrow NS, Campbell V, Page KM, Brickley K, Tedder I & Dolphin AC (1997). Importance of the different subunits in the membrane expression of the 1A and 2 calcium channel subunits: studies using a depolarization-sensitive 1A antibody. Eur J Neurosci 9, 749–759.
Brody DL, Patil PG, Mulle JG, Snutch TP & Yue DT (1997). Bursts of action potential waveforms relieve G-protein inhibition of recombinant P/Q-type Ca2+ channels in HEK 293 cells. J Physiol 499, 637–644.
Canti C, Davies A, Berrow NS, Butcher AJ, Page KM & Dolphin AC (2001). Evidence for two concentration-dependent processes for -subunit effects on 1B calcium channels. Biophys J 81, 1439–1451.
Catterall WA (2000). Structure and regulation of voltage-gated Ca2+ channels. Annu Rev Cell Dev Biol 16, 521–555.
Chen Y, Li M, Zhang Y, He L, Yamada Y, Fitzmaurice A, Shen Y, Zhang H, Tong L & Yang J (2004). Structural basis of the 1–subunit interaction of voltage-gated Ca2+ channels. Nature 429, 675–680.
Chien AJ, Zhao X, Shirokov RE, Puri TS, Chang CF, Sun D, Rios E & Hosey MM (1995). Roles of a membrane-localized subunit in the formation and targeting of functional L-type Ca2+ channels. J Biol Chem 270, 30036–30044.
Cohen RM, Foell JD, Balijepalli RC, Shah V, Hell JW & Kamp TJ (2004). Unique modulation of L-type Ca2+ channels by short auxiliary 1d subunit present in cardiac muscle. Am J Physiol Heart Circ Physiol 288, H2363–H2374.
Colecraft HM, Alseikhan B, Takahashi SX, Chaudhuri D, Mittman S, Yegnasubramanian V, Alvania RS, Johns DC, Marban E & Yue DT (2002). Novel functional properties of Ca2+ channel subunits revealed by their expression in adult rat heart cells. J Physiol 541, 435–452.
Costantin J, Noceti F, Qin N, Wei X, Birnbaumer L & Stefani E (1998). Facilitation by the 2a subunit of pore openings in cardiac Ca2+ channels. J Physiol 507, 93–103.
De Waard M, Pragnell M & Campbell KP (1994). Ca2+ channel regulation by a conserved subunit domain. Neuron 13, 495–503.
Dolphin AC (2003). subunits of voltage-gated calcium channels. J Bioenerg Biomemb 35, 620.
Finlin BS, Crump SM, Satin J & Andres DA (2003). Regulation of voltage-gated calcium channel activity by the Rem and Rad GTPases. Proc Natl Acad Sci U S A 100, 14469–14474.
Gao T, Chien AJ & Hosey MM (1999). Complexes of the 1C and subunits generate the necessary signal for membrane targeting of class C L-type calcium channels. J Biol Chem 274, 2137–2144.
Garcia R, Carrillo E, Rebolledo S, Garcia MC & Sanchez JA (2002). The 1a subunit regulates the functional properties of frog and mouse L-type Ca2+ channels of skeletal muscle. J Physiol 545, 407–419.
Gerster U, Neuhuber B, Groschner K, Striessnig J & Flucher BE (1999). Current modulation and membrane targeting of the calcium channel 1C subunit are independent functions of the subunit. J Physiol 517, 353–368.
Hullin R, Khan IFY, Wirtz S, Mohacsi P, Varadi G, Schwartz A & Herzig S (2003). Cardiac L-type calcium channel subunits expressed in human heart have differential effects on single channel characteristics. J Biol Chem 278, 21623–21630.
Jones SW (2002). Calcium channels: when is a subunit not a subunitJ Physiol 545, 334.
Jones LP, Wei SK & Yue DT (1998). Mechanism of auxiliary subunit modulation of neuronal 1E calcium channels. J General Physiol 112, 125–143.
Ludwig A, Flockerzi V & Hofmann F (1997). Regional expression and cellular localization of the 1 and subunit of high voltage-activated calcium channels in rat brain. J Neurosci 17, 1339–1349.
McGee AW, Nunziato DA, Maltez JM, Prehoda KE, Pitt GS & Bredt DS (2004). Calcium channel function regulated by the SH3-GK module in subunits. Neuron 42, 89–99.
Opatowsky Y, Chen C-C, Campbell KP & Hirsch JA (2004). Structural analysis of the voltage-dependent calcium channel subunit functional core and its complex with the 1 interaction domain. Neuron 42, 387–399.
Opatowsky Y, Chomsky-Hecht O, Kang MG, Campbell KP & Hirsch JA (2003). The voltage-dependent calcium channel subunit contains two stable interacting domains. J Biol Chem 278, 52323–52332.
Perez-Reyes E, Castellano A, Kim HS, Bertrand P, Baggstrom E, Lacerda AE, Wei XY & Birnbaumer L (1992). Cloning and expression of a cardiac/brain subunit of the L-type calcium channel. J Biol Chem 267, 1792–1797.
Pragnell M, De Waard M, Mori Y, Tanabe T, Snutch TP & Campbell KP (1994). Calcium channel -subunit binds to a conserved motif in the I–II cytoplasmic linker of the 1-subunit. Nature 368, 67–70.
Qin N, Platano D, Olcese R, Stefani E & Birnbaumer L (1997). Direct interaction of Gwith a C-terminal G-binding domain of the Ca2+ channel 1 subunit is responsible for channel inhibition by G protein-coupled receptors. Proc Natl Acad Sci U S A 94, 8866–8871.
Scott VE, De Waard M, Liu H, Gurnett CA, Venzke DP, Lennon VA & Campbell KP (1996). subunit heterogeneity in N-type Ca2+ channels. J Biol Chem 271, 3207–3212.
Singer D, Biel M, Lotan I, Flockerzi V, Hofmann F & Dascal N (1991). The roles of the subunits in the function of the calcium channel. Science 253, 1553–1557.
Spafford JD, Van Minnen J, Larsen P, Smit AB, Syed NI & Zamponi GW (2004). Uncoupling of calcium channel 1 and subunits in developing neurons. J Biol Chem 279, 41157–41167.
Takahashi SX, Miriyala J & Colecraft HM (2004). Membrane-associated guanylate kinase-like properties of subunits required for modulation of voltage-dependent Ca2+ channels. Proc Natl Acad Sci U S A 101, 7193–7198.
Takahashi SX, Mittman S & Colecraft HM (2003). Distinctive modulatory effects of five human auxiliary 2 subunit splice variants on L-type calcium channel gating. Biophys J 84, 3007–3021.
Tareilus E, Roux M, Qin N, Olcese R, Zhou J, Stefani E & Birnbaumer L (1997). A Xenopus oocyte subunit: evidence for a role in the assembly/expression of voltage-gated calcium channels that is separate from its role as a regulatory subunit. Proc Natl Acad Sci U S A 94, 1703–1708.
Tomlinson WJ, Stea A, Bourinet E, Charnet P, Nargeot J & Snutch TP (1993). Functional properties of a neuronal class C L-type calcium channel. Neuropharmacology 32, 1117–1126.
Van Petegem F, Clark KA, Chatelain FC & Minor DL (2004). Structure of a complex between a voltage-gated calcium channel -subunit and an -subunit domain. Nature 429, 671–675.
Walker D, Bichet D, Campbell KP & De Waard M (1998). A beta 4 isoform–specific interaction site in the carboxyl-terminal region of the voltage-dependent Ca2+ channel 1A subunit. J Biol Chem 273, 2361–2367.
Walker D, Bichet D, Geib S, Mori E, Cornet V, Snutch TP, Mori Y & De Waard M (1999). A new subtype–specific interaction in 1A subunit controls P/Q-type Ca2+ channel activation. J Biol Chem 274, 12383–12390.
Walker D & De Waard M (1998). Subunit interaction sites in voltage-dependent Ca2+ channels: role in channel function. Trends Neurosci 21, 148–154.
Wei XY, Perez-Reyes E, Lacerda AE, Schuster G, Brown AM & Birnbaumer L (1991). Heterologous regulation of the cardiac Ca2+ channel 1 subunit by skeletal muscle and subunits. Implications for the structure of cardiac L-type Ca2+ channels. J Biol Chem 266, 21943–21947.
Witcher DR, De Waard M, Liu H, Pragnell M & Campbell KP (1995). Association of native Ca2+ channel subunits with the 1 subunit interaction domain. J Biol Chem 270, 18088–18093.
Witcher DR, De Waard M, Sakamoto J, Franzini-Armstrong C, Pragnell M, Kahl SD & Campbell KP (1993). Subunit identification and reconstitution of the N-type Ca2+ channel complex purified from brain. Science 261, 486–489.
Yamaguchi H, Hara M, Strobeck M, Fukasawa K, Schwartz A & Varadi G (1998). Multiple modulation pathways of calcium channel activity by a subunit. Direct evidence of subunit participation in membrane trafficking of the 1C subunit. J Biol Chem 273, 19348–19356.
Yamaguchi H, Okuda M, Mikala G, Fukasawa K & Varadi G (2000). Cloning of the 2a subunit of the voltage-dependent calcium channel from human heart: cooperative effect of 2 and 2a on the membrane expression of the 1C subunit. Biochem Biophys Res Commun 267, 156–163.
Young K, Lin S, Sun L, Lee E, Modi M, Hellings S, Husbands M, Ozenberger B & Franco R (1998). Identification of a calcium channel modulator using a high throughput yeast two-hybrid screen. Nat Biotechnol 16, 946–950.(Stanislava Dalton, Shoji )
Abstract
Voltage-dependent calcium-channel subunits (CaV) strongly modulate pore-forming 1 subunits by trafficking channel complexes to the plasma membrane and enhancing channel open probability (Po). Despite their central role, it is unclear whether binding of a single CaV, or multiple CaVs, to an 1 subunit governs the two distinct functions. Conventional experiments utilizing coexpression of 1 and CaVsubunits have been unable to resolve the ambiguity due to difficulties in establishing their stoichiometry in functional channels. Here, we unambiguously establish a 1: 1 stoichiometry by covalently linking CaV2b to the carboxyl terminus of 1C (CaV1.2), creating 1C·2b. Recombinant L-type channels reconstituted in HEK 293 cells with 1C·2b supported whole-cell currents to the same extent as channels reconstituted via coexpression of the individual subunits. Analysis of gating charge showed 1C·2b fully restored channel trafficking to the plasma membrane. Co-transfecting CaV2a with 1C·2b had little further impact on function. To rule out the possibility that fused CaV2b was interacting in trans with neighbouring 1 molecules, 1C·2b was cotransfected with 1B (CaV2.2), and pharmacological block with nimodipine showed an absence of 1B trafficking. These results establish that association of a single CaVwith a pore-forming 1 subunit captures the functional essence of HVA calcium channels, and introduce 1–CaVfusion proteins as a powerful new tool to probe structure–function mechanisms.
Introduction
High-voltage-activated (HVA) calcium channels are hetero-multimeric protein complexes comprised minimally of a pore-forming 1-subunit (CaV1.1–1.4; CaV2.1–2.3), and auxiliary (CaV1–CaV4), 2 (2-1–2-3), and sometimes subunits (Catterall, 2000). These proteins control Ca2+ fluxes that underlie critical biological functions ranging from muscle contraction to hormone and neurotransmitter release. Auxiliary CaVsubunits play a crucial role in the functional maturation of HVA calcium channels: each of the four CaVs can interact promiscuously with any of the seven known HVA calcium-channel 1 subunits to promote membrane trafficking of the associated 1 (Chien et al. 1995; Brice et al. 1997; Yamaguchi et al. 2000; Colecraft et al. 2002; Takahashi et al. 2004), to increase channel Po (Singer et al. 1991; Costantin et al. 1998; Hullin et al. 2003; Takahashi et al. 2004), and to produce hyperpolarizing shifts in the voltage-dependence of channel activation (Singer et al. 1991; Perez-Reyes et al. 1992; De Waard et al. 1994; Jones et al. 1998).
The powerful impact of CaVs on HVA calcium-channel operation logically focuses attention on the interaction between 1 and CaVsubunits as a promising locus to modify calcium-channel behaviour. Indeed, this interaction is the target for physiological down-regulation of calcium currents by the Rem, Rad, Kir/Gem (RGK) family of Ras-like GTPases (Beguin et al. 2001; Finlin et al. 2003), and serves as the basis for high-throughput identification of novel pharmacological inhibitors of HVA calcium-channel activity (Young et al. 1998). A deeper mechanistic understanding of the HVA 1–CaVinteraction is important for both appreciating its physiological ramifications, as well as potentially hastening the discovery of new calcium-channel therapeutics.
A fundamental ambiguity of HVA calcium-channel structure relates to the functional stoichiometry of 1 to CaV(Birnbaumer et al. 1998; Canti et al. 2001; Jones, 2002; Dolphin, 2003). Specifically, it is unclear whether the fully functional channel is comprised of an 1 subunit associated with single or multiple CaVsubunits. Biochemical studies initially established that 1 and CaVsubunits copurify in an equimolar ratio (Witcher et al. 1993); the uncertainty arises from studies which indicate that trafficking and gating-modulation are separable functions dependent on distinct CaVconcentrations. In Xenopus oocytes, expression of 1 without exogenous CaVleads to recovery of low-amplitude currents that activate at relatively depolarized potentials, compared to channels obtained by coexpressing 1 and CaV(Singer et al. 1991; Perez-Reyes et al. 1992; Birnbaumer et al. 1998). However, the presumed 1-alone currents are abolished by antisense oligonucleotides against a subsequently identified endogenous Xenopus CaV(CaVXO) (Tareilus et al. 1997). These results suggested that there is sufficient endogenous CaVXO in the Xenopus expression system to fully traffick 1 subunits to the plasma membrane, but not to modulate gating (Tareilus et al. 1997; Birnbaumer et al. 1998). In agreement, careful titration of the amount of CaV3 coexpressed with CaV2.2 in Xenopus oocytes indicated that the concentration of CaV3 required to traffick channels to the plasma membrane is nearly an order of magnitude less than that required to modulate channel gating (Canti et al. 2001). Furthermore, acute application of purified CaVprotein was found to modulate the gating of pre-existing calcium channels in the plasma membrane (Yamaguchi et al. 1998; Garcia et al. 2002).
These data are consistent with two mutually exclusive interpretations of the functional stoichiometry of the 1–CaVinteraction (Birnbaumer et al. 1998; Jones, 2002; Dolphin, 2003). In the ‘single-CaV-binding’ model, CaVbinds to 1 at the endoplasmic reticulum (ER) to promote translocation to the plasma membrane. At low CaVconcentrations, reversible unbinding of CaVfrom 1 favours a significant steady-state fraction of low-activity CaV-less channels in the plasma membrane. At high CaVconcentrations, the equilibrium is shifted towards high-activity (CaV-bound) channels (Fig. 1A). Thus, a single CaVis responsible for both trafficking and enhancing the channel Po. Alternatively, in the ‘multiple-CaV-binding’ model, trafficking and gating-modulation are mediated by distinct CaVsubunits (Fig. 1B). As before, a CaVbinds to an 1 to promote channel trafficking. After insertion into the plasma membrane, the high-affinity interaction between 1 and CaVpersists. By contrast to the single-CaVmodel, these 1 + CaVchannels exhibit low gating activity, and binding of a second CaVto a distinct lower affinity site is required for high gating activity. Distinguishing between these two models is fundamental to mechanistic understanding of CaVfunction, and has important implications for rational manipulation of the 1–CaVinteraction to regulate calcium-channel activity.
A, the single-CaV-binding model. A single CaVsubunit binds to the 1 subunit at the endoplasmic reticulum (ER) membrane resulting in trafficking of the 1–complex to the plasma membrane. At the plasma membrane, a reversible 1: 1 interaction between 1 and CaVsubunits switches the channel between low- and high-Po modes. B, the multiple-CaV-binding model. A CaVbinding to an 1 subunit at the ER membrane is also responsible for trafficking the 1-channel complex to the plasma membrane. However, this CaVdoes not dissociate from the channel complex, which gates in a low-Po mode. Binding of an additional CaVsubunit (or subunits) switches the channel complex to a high-activity gating mode.
A key obstacle to discriminating between the two models is that in traditional coexpression studies it is practically impossible to distinguish the precise stoichiometry of 1 and CaVsubunits present in functional channels. Here, we take the novel approach of covalently linking an 1C subunit (CaV1.2) to a CaV2b subunit, thus ensuring a 1: 1 stoichiometry in transfected HEK 293 cells. We find that a single CaVis necessary and sufficient to completely reconstitute both channel trafficking and macroscopic conductance. We conclude that the single-CaVmodel captures the functional essence of HVA calcium channels.
Methods
Molecular biology
A truncation mutant of 1C (1C[1905]), and 1C–CaVfusion proteins (1C[1905]·2b, 1C·2b, 1C·2b[P235R]) were constructed using rabbit 1C/pGW1 (Wei et al. 1991) and human CaV2b/pAd CIG (Takahashi et al. 2003) expression vectors as backbones. To generate 1C[1905]/pGW1, we inserted a stop codon after residue E1905 of 1C by polymerase chain reaction (PCR) amplifying 1C/pGW1 using the following upstream and downstream primers, respectively: (primer 1) CACCATCTGGATGAATTTAAAAGAATC and (primer 2) TCTAGACTACTCCTCCGGGGGCGCCAGTTG. The resulting PCR fragment was ligated into 1C/pGW1 using BstEII/XbaI sites, generating 1C[1905]/pGW1.
The 1C[1905]·2b construct was generated by overlap extension PCR as follows. 1C[1905]/pGW1 was used as a template in a PCR reaction with primer 1 and the following downstream primer: (primer 3) CTCCTCCGGGGGCGCCAGTTGTCG. A second PCR using human CaV2b/pAd CIG as template was carried out using the following upstream and downstream primers, respectively: (primer 4) CTGGCGCGCCCGGAGGAGGGTGGTGGTGGTATGCTTGACAGACGCCTT and (primer 5) GCGTACAAGGGTACCGCAATACCG. The two PCR fragments were used as templates in an overlap-extension PCR reaction using primers 1 and 5. A BstEII/Xba1 cassette from the overlap extension PCR product was transferred to 1C[1905]/pGW1, generating 1C[1905]·2b/pGW1 in which CaV2b is attached to 1C[1905] via a 4-glycine linker.
The 1C·2b construct was generated using a strategy similar to that described above for 1C[1905]·2b. The 1C/pGW1 construct was used as template in a PCR reaction with the following upstream and downstream primers, respectively: (primer 6) GCAGTGGCGGGCCTGAGTCCCCTC, and (primer 7) CAGGCTGCTGACGCCGGCCCTGCG. Human CaV2b/pAd CIG was used as template in a second PCR using the following upstream primer: (primer 8) GCCGGCGTCAGCAGCCTGGGTGGTGGTGGTATGCTTGACAGACGCCTTATAG in conjunction with primer 5. The two PCR products were used in an overlap extension reaction with primers 6 and 5. An RsrII/XbaI cassette from the overlap extension PCR product was transferred into 1C/pGW1, generating 1C·2b/pGW1, in which CaV2b is attached to the carboxyl terminus of 1C via a 4-glycine linker.
The point mutant fusion protein, 1C·2b[P235R], was generated by overlap extension PCR using 1C·2b as template. The first PCR reaction was performed using primer 8 and the following downstream primer: (primer 9) CTTCAGAGAACGGCCCACTAGGAC (mutated base underlined). The second PCR reaction was performed using the following upstream and downstream primers, respectively: (primer 10) GTCCTAGTGGGCCGTTCTCTGAAG and (primer 11) AGGTCGACTCTAGAGCGGCCGCCA. The two PCR products were used in an overlap extension reaction, and a SfuI/XbaI cassette from the resulting product transferred into 1C·2b/pGW1 to generate 1C·2b[P235R]/pGW1. Pfu polymerase (Strategene, La Jolla, CA, USA) was used to increase fidelity in all PCR reactions. All PCR products were verified by sequencing.
Cell culture and transient channel expression
Low passage number human embryonic kidney (HEK 293) cell cultures (< 20 passages) were maintained as previously described (Brody et al. 1997). Cells were transiently transfected 24 h after passage using the calcium-phosphate precipitation method with 8 μg each of CMV expression plasmids encoding the indicated calcium-channel subunits [rabbit 1C (Wei et al. 1991); 1C[1905]; human CaV2b (Takahashi et al. 2003); 1C·2b; and rat 2 (Tomlinson et al. 1993)] and 3 μg of T antigen.
Electrophysiology
Electrophysiological experiments were performed as previously described (Takahashi et al. 2004). Whole-cell recordings were performed at room temperature, 2–3 days after transfection, using an EPC8 patch-clamp amplifier (HEKA Electronics, Lambrecht/Pfalz, Germany) controlled by PULSE software. Patch pipettes were fashioned from 1.5-mm thin-wall glass with filament (WPI, Waltham, MA). Patch pipettes typically had a series resistance of 2–4 M, compensated 50–70%, when filled with an intracellular solution containing (mM): 135 caesium methanesulphonate (Cs-MeSO3), 5 CsCl, 5 EGTA, 1 MgCl2, 4 MgATP (freshly added on the day of patch-clamp experiments), and 10 Hepes (pH 7.3, adjusted with CsOH). Cells were continuously perfused with an external solution containing (mM): 140 tetraethylammonium (TEA) MeSO3, 10 Hepes, 5 BaCl2 (pH 7.4, adjusted with TEA-OH). The standard electrophysiological protocol consisted of a family of 20-ms test pulse depolarizations (from –40 to +120 mV) used to evoke currents from a –90-mV holding potential, followed by a repolarization to –50 mV to measure tail currents. Current signals were sampled at 25 kHz and filtered at 10 kHz, with leak and capacitive transients subtracted using a P/8 protocol. Traces were acquired at a repetition interval of 15 s.
Data and statistical analysis
Data were analysed off-line using PULSEFIT (HEKA Electronics). Peak current amplitudes during the test pulse were normalized by cell capacitance to provide the current density (Jpeak). Integration of ON gating currents recorded at the reversal potential to provide gating charge (QON) was performed in Origin (OriginLab Corp., Northampton, MA, USA). J–V relationships were fitted to the following equation:
where J is the whole-cell current density (pA pF–1), G is the specific conductance (pA pF–1 mV–1), Vrev is the reversal potential (mV), V1/2 is the voltage of half-maximal activation (mV), and k is a slope factor (mV). Statistical analyses were performed in Microsoft Excel using built-in functions. Pooled data are represented as means ± S.E.M.; P-values were calculated using Student's two-tailed t test, with P < 0.05 considered statistically significant.
Results
Robust modulation of truncated CaV1.2 (1C[1905]) by coexpressed CaV2b in HEK 293 cells
The traditional experimental approach towards investigating the functional effects of CaVon HVA calcium channels has been to compare the properties of recombinant channels reconstituted by expressing HVA calcium-channel 1 subunits, either with or without CaVin heterologous expression systems. Figure 2 shows the characteristic powerful impact of CaVon the functional expression of recombinant calcium channels, as reported by the conventional coexpression approach. Transfection of HEK 293 cells with cDNA for a truncated CaV1.2 (1C[1905]) subunit and 2 resulted in the recovery of small-amplitude recombinant L-type calcium-channel currents in 70% of selected cells (Fig. 2A and B). Co-expression of CaV2b with 1C[1905]/2 dramatically increased the amplitude of recorded calcium-channel currents across the entire range of membrane voltages as indicated by exemplar traces (Fig. 2C and D), and population peak current density vs. voltage (Jpeak–V; Fig. 2E) and tail-current vs. voltage (Itail–V; Fig. 2F) curves. Overall, peak current density obtained with 1C[1905] + CaV2b channels were over 10-fold larger than those obtained with CaV-less 1C[1905] (current density at 0 mV, Jpeak,0mV = 4.05 ± 0.39 pA pF–1, n = 7, for 1C[1905]/2; Jpeak,0mV = 47.9 ± 11.8 pA pF–1, n = 6, for 1C[1905]/CaV2b/2; P < 0.05).
A, schematic diagram of the 1C[1905] subunit. B, exemplar whole-cell currents obtained from a HEK 293 cell transfected with 1C[1905] + 2. C, schematic diagram of 1C[1905] + CaV2b. D, exemplar whole-cell currents obtained from a cell cotransfected with 1C[1905] + CaV2b + 2. Co-expression with CaVresults in a marked increase in whole-cell current amplitude compared to 1C[1905] + 2 channels. E, plots of peak current density vs. test pulse voltage (Jpeak–V) for cells transfected with 1C[1905] + 2 () or 1C[1905] + CaV2b + 2 (). F, plots of tail current amplitude vs. voltage (Itail–V) from channels expressed without () or with () cotransfected CaV2b. G, top, exemplar gating currents obtained from recombinant 1C[1905] channels expressed without (left) or with (right) cotransfected CaV2b. G, bottom, time integral of the gating current obtained at the reversal potential (QON) obtained for channels reconstituted without or with CaV2b.
It has been previously established that CaVs elevate whole-cell currents by increasing both the number of channels expressed in the plasma membrane (N) (Chien et al. 1995; Brice et al. 1997; Gao et al. 1999), and the channel Po (Costantin et al. 1998; Hullin et al. 2003). In our electrophysiological protocol, N can be estimated by calculating the integral of the ON gating current (Qmax = N x q, where q is single-channel gating charge) isolated at the reversal potential (usually +50 mV) (Jones et al. 1998; Colecraft et al. 2002; Takahashi et al. 2004). For cells transfected with 1C[1905]/2, gating currents were essentially unresolved (Fig. 2A and G; QON = 0.05 ± 0.04 fC pF–1, n = 7), clearly indicating a dearth of channels with movable voltage sensors at the plasma membrane. By contrast, coexpressing 1C[1905] with CaV2b resulted in relatively large and clearly resolved gating currents (Fig. 2D and G; QON = 5.7 ± 2.3 fC pF–1, n = 7), thus demonstrating that CaV2b promotes targeting of 1C[1905] to the plasma membrane. Overall, these results establish the baseline functional properties of recombinant L-type channels obtained via cotransfecting HEK 293 cells with 1C[1905] and CaV2b.
Covalently linking 1C[1905] and CaV2b recapitulates robust functional expression of recombinant L-type channels
Unfortunately, traditional coexpression experiments as described in Fig. 2 cannot be used to determine the functional stoichiometry of 1 and CaVspecies, since their relative expression levels cannot be tightly controlled. To control the stoichiometry between the subunits, coding sequence for CaV2b was fused to the carboxyl terminus of 1C[1905] by subcloning. The resulting plasmid construct, when transfected and expressed in HEK 293 cells would result in a fusion protein (1C[1905]·2b) in which a 1: 1 ratio of 1C[1905] to CaVis assured (Fig. 3A). Whole-cell records from these fusion experiments showed that 1C[1905]·2b clearly supports robust currents, as demonstrated by exemplar traces (Fig. 3B) and population Jpeak–V (Fig. 3C) and Itail–V (Fig. 3D) curves. Importantly, these experiments prove that CaV2b was still functional in the fused-channel configuration.
A, schematic diagram of the 1C[1905]·2b construct. B, exemplar whole-cell currents obtained at different test potentials for a cell transfected with 1C[1905]·2b + 2. C, population Jpeak–V plot obtained for 1C[1905]·2b + 2 channels (). Data for 1C[1905] + 2 channels have been reproduced (grey trace) to facilitate visual comparison of differences in peak current density amplitude. D, population Itail–V plot for 1C[1905]·2b + 2 channels (). The corresponding data for 1C[1905] + 2 channels are reproduced (grey trace). E, QON values.
How did currents obtained with 1C[1905]·2b channels compare with those obtained by the customary approach of coexpressing 1C[1905] and CaV2b as independent subunitsIn fact, whole-cell current amplitudes were essentially identical to those obtained with 1C[1905]/CaV2b/2 (Jpeak,0mV = 33.9 ± 4.8 pA pF–1, n = 13, for 1C[1905]·2b/2 channels; P = 0.27, compared to 1C[1905]/CaV2b/2 channels). Moreover, the fused CaV2b promoted channel trafficking to the same extent as free CaV2b, as determined by gating current analyses (Fig. 3E; QON = 4.9 ± 0.7 fC pF–1, n = 13, for 1C[1905]·2b/2; P = 0.75, compared to 1C[1905]/CaV2b/2 channels). Overall, the similarity in whole-cell current amplitude and QON suggests that CaV2b has virtually identical effects on microscopic channel properties (N and Po) irrespective of whether it is directly fused to 1C[1905], or operating independently.
Modulation of full-length CaV1.2 by free and covalently linked CaV2b
We conducted the initial series of experiments with the truncated CaV1.2 subunit, 1C[1905], to maximize the chances of obtaining a positive result. We theorized that the shortened carboxyl terminus of 1C[1905] would facilitate the interaction of the fused CaV2b with the primary CaVinteraction site in the cytoplasmic linker between domains I and II of the 1C subunit (I–II linker) (Pragnell et al. 1994). Having established the efficacy of the fused CaV2b in the truncated channel configuration (Fig. 3), we sought to determine whether the experimental paradigm yielded equally robust results in the context of full-length 1C.
Similar to observations with the truncated 1C[1905] subunit, coexpression of full-length 1C and CaV2b/2 (Fig. 4D–F) resulted in significantly larger whole-cell currents than obtained with 1C/2 (Fig. 4A–C). Fits of the J–V relations to a modified Boltzmann relation revealed that CaV2b resulted in a four-fold increase in the macroscopic channel conductance (Table 1). Moreover, channels reconstituted with free CaV2b displayed another classic hallmark of CaVmodulation compared to 1C/2: namely a hyperpolarizing shift in the voltage-dependence of channel activation (as gauged by V1/2-values; Table 1).
A, schematic diagram of full-length 1C. B, exemplar currents obtained from cell transfected with 1C + 2. C, population Jpeak–V plot for 1C + 2 channels. D–F, data for 1C + CaV2b + 2 channels. Same format as for A–C. Data for 1C + 2 channels in C are reproduced in F, I and L (grey trace) to facilitate direct visual comparisons. G–I, data for 1C·2b + 2; format same as above. J–L, data for 1C·2b[P235R] + 2; same format as above for A–C.
Reassuringly, whole-cell currents produced by 1C·2b/2-reconstituted channels (Fig. 4G–I) were also four-fold larger than obtained with CaV-less 1C/2 channels, and essentially identical to 1C/CaV2b/2 currents (Table 1). Moreover, channel trafficking was promoted to the same extent by CaV2b irrespective of whether it was free or fused, as indicated by similar QON values that were significantly greater than observed with 1C/2-reconstituted channels (Table 1). The quantitative J–V analyses did reveal an important difference between channels reconstituted with free vs. fused CaV2b. The 1C·2b/2 channels did not recapitulate the hyperpolarizing shift in the voltage-dependence of channel activation (Table 1, P < 0.01 when comparing V1/2-values for 1C + CaV2b + 2, 1C + 2, and 1C·2b + 2 channels by one-way ANOVA). This distinction could reflect an intrinsic property of the fused 1C·2b protein itself, or could signal the requirement of a second CaVto reconstitute the shift-in-activation gating function. Results provided in the next section appear more consistent with the former interpretation. Overall, the finding that robust currents could be successfully reconstituted with both 1C[1905]·2b and 1C·2b reveals that the design criteria for creating functional HVA calcium channels by fusing CaVs to the carboxyl terminus of 1 subunits are quite relaxed; the fused CaVappears equally effective when placed at different positions along the carboxyl terminus. By contrast, fusing CaV2b directly to the amino-terminus of 1C did not yield robust currents (not shown), suggesting that geometric constraints may preclude the formation of functional channels under this condition. Alternatively, amino-terminus-fused CaVconstructs may require more extensive optimization to produce functional channels (e.g. varying linker lengths) than appears necessary with attachment to the carboxyl terminus.
To discount the possibility that observed channel modulation might be due to nonspecific effects of attaching CaV2b to the 1C carboxyl terminus, we tested the effect of a proline to arginine (P235R) mutation in the fused CaV2b subunit, 1C·2b[P235R] (Fig. 4J–L). The analogous mutation in other CaVsubunits has been shown to disrupt binding to the 1 interaction domain (AID), and ablates the capacity of CaVto modulate 1 subunits in cotransfection experiments (De Waard et al. 1994; Takahashi et al. 2004), possibly due to derangement of the three-dimensional structure of the CaV1-binding pocket (ABP) (Chen et al. 2004; Opatowsky et al. 2004; Van Petegem et al. 2004). Reassuringly, 1C·2b[P235R] yielded whole-cell currents that were indistinguishable from those obtained with CaV-less 1C/2 (Fig. 4J–L and Table 1), clearly ruling out any unanticipated confounding effects generated by the fused CaV2b subunit.
Additional CaVs do not appreciably enhance 1C·2b channel currents
The results thus far indicated that a single CaV2b fused to 1C was necessary and sufficient to fully recapitulate CaVeffects on channel trafficking and macroscopic conductance (Figs 3 and 4). Could the presence of additional CaVsubunits further enhance the channel modulationWe directly examined this question by coexpressing 1C·2b/2 with CaV2a and recording whole-cell currents (Fig. 5A–C). Previous studies have established that CaV2a targets autonomously to the plasma membrane and robustly modulates recombinant L-type Ca2+ channel currents. Therefore, CaV2a was used in these coexpression experiments because we reasoned that its high effective local concentration at the membrane would provide the most favourable circumstance to observe a functional effect due to a second CaVsubunit. Exemplar whole-cell current traces for 1C·2b/CaV2a/2 were similar to those obtained for 1C·2b/2. Population data confirmed that there was no significant difference in macroscopic conductance between 1C·2b currents in the presence or absence of CaV2a (Fig. 5C and Table 1). By contrast, coexpressing 1C·2b[P235R] with CaV2a (Fig. 5D–F) resulted in significantly enhanced whole-cell current amplitudes compared to 1C·2b[P235R] (Fig. 5F and Table 1). Importantly, the presence of free CaV2a did not result in a significant change in the voltage-dependence of activation in 1C·2b/2 channels (Table 1; V1/2 = –6.1 ± 2.1 for 1C/2; V1/2 = –6.4 ± 1.5 mV for 1C·2b/2; V1/2 = –9.9 ± 1.8 mV for 1C·2b/CaV2a/2 channels, P = 0.28 by one-way ANOVA). This suggests that the relatively right-shifted voltage-dependence of activation of 1C·2b/2 channels reflects an intrinsic property of the fusion protein, and does not signal the requirement for a second CaVto reconstitute this functional property. Together, these results raised confidence that the fused wild-type CaV2b subunits were sufficient to fully modulate channel properties, and did not require the collaboration of additional CaVsubunits.
A, schematic diagram of 1C·2b + CaV2a. B, exemplar whole-cell currents from a cell transfected with 1C·2b + CaV2a + 2. C, population Jpeak–V relationship for 1C·2b + CaV2a + 2 channels (). Data for 1C·2b + 2 are reproduced to permit direct visual comparison (grey trace). D, schematic diagram of 1C·2b [P235R] + CaV2a. E, exemplar whole-cell currents from 1C·2b[P235R] + CaV2a + 2 channels. F, Jpeak–V relationship for 1C.2b + CaV2a + 2 channels (). Data for 1C.2b[P235R] + 2 are reproduced (grey trace).
Ruling out the presence of confounding trans interactions between distinct 1C·2b fusion-protein molecules
Our interpretation of the data obtained with 1C·2b as indicating that a single CaVis necessary and sufficient to fully modulate channel trafficking and macroscopic conductance hinges critically on the assumption that the fused CaVs act solely on the 1C subunits to which they are attached. This assumption might not be true if the fused CaVcould act in trans on a neighbouring 1C molecule. To evaluate the extent to which such intermolecular interactions took place, we coexpressed 1C·2b with 1B (Fig. 6). Our strategy was based on the idea that in the absence of free CaVthe only avenue for appreciable 1B transport to the membrane would be via an intermolecular interaction with the fused CaV2b subunit. Pharmacological block with nimodipine could then be used to distinguish relative amounts of 1C (L-type) and 1B (N-type) channels at the plasma membrane.
A, exemplar 1C.2b + 2 currents in the absence (black trace) and presence (grey trace) of 5 μM nimodipine. B, time course of nimodipine effect on exemplar current. Arrows denote time of nimodipine addition. C, bar graph showing degree of nimodipine inhibition in population of 1C.2b + 2 currents. D–F, data for 1B + CaV2b + 2 channels; same format as panels A–C. G–I, data for 1B + 1C.2b + 2 channels; same format as panels A–C. J–L, data for 1B + CaV2b + 1C.2b + 2 channels; same format as panels A–C.
As expected for recombinant L-type currents, channels reconstituted with 1C·2b/2 were strongly inhibited by the dihydropyridine antagonist, nimodipine (Fig. 6A–C). For these channels, 5 μM nimodipine inhibited 90% of the whole-cell current shortly after exposure (Fig. 6A–C). By contrast, 5 μM nimodipine had minimal effect on channels recorded from cells transfected with 1B + CaV2b (Fig. 6D–F). In cells transfected with 1C·2b/1B/2, application of 5 μM nimodipine inhibited whole-cell currents by 90% (Fig. 6G–I), identical to what was observed with 1C·2b. As a control to ensure that nimodipine-insensitive currents can be observed when a mixed population of 1B and 1C·2b channels are present at the membrane, we transfected cells with 1B/CaV2b and 1C·2b/2. In this regime, 5 μM nimodipine had an intermediate effect, inhibiting whole-cell currents by 50% on average (Fig. 6J–L). Together, these result provided compelling evidence against any significant contribution of intermolecular interactions involving fused CaV2b subunits.
Discussion
We investigated the functional stoichiometry of HVA calcium-channel 1 and subunits to address a long-standing debate over whether a fully functional HVA calcium channel is comprised of an 1 subunit associated with a single, or multiple CaVsubunits. Our results indicate that a 1: 1 ratio of 1 and CaVcaptures the functional essence of mature HVA channels. We discuss the significance of our findings in the context of previous results relevant to the issue of the stoichiometry of the calcium-channel 1–subunit interaction.
Previous studies on the functional stoichiometry of calcium-channel 1 and subunits
Biochemical studies of HVA calcium channels initially established that 1 and subunits copurified in an equimolar ratio, suggesting a 1: 1 stoichiometry of these proteins in native calcium-channel complexes (Witcher et al. 1993). Structure–function studies and recent high-resolution crystal structures show that CaVs bind to a conserved stretch of amino acids, termed 1 interaction domain (AID) located in the cytoplasmic region linking the first and second transmembrane domains of HVA 1 subunits (De Waard et al. 1994; Pragnell et al. 1994; Chen et al. 2004; Opatowsky et al. 2004; Van Petegem et al. 2004). For their part, the core region of CaVs contains two protein interaction motifs—a src homology 3 (SH3) domain, and a guanylate kinase-like (GK) domain (Chen et al. 2004; Opatowsky et al. 2004; Van Petegem et al. 2004), both of which are necessary for CaVfunction (Opatowsky et al. 2003; McGee et al. 2004; Takahashi et al. 2004). The GK domain exhibits a high-affinity interaction with the AID peptide.
However, the assumption of a 1: 1 1 to CaVstoichiometry was challenged by electrophysiological studies of recombinant channels expressed in Xenopus oocytes. Calcium-channel 1-subunit RNA injected alone into Xenopus oocytes results in low-amplitude whole-cell currents that activate at relatively depolarized potentials compared to those coexpressed with CaV(Singer et al. 1991; Perez-Reyes et al. 1992; Birnbaumer et al. 1998). Such low-level currents were attributed to CaV-less ‘1-alone’ channels. However, Xenopus oocytes were subsequently found to possess an endogenous subunit (XO). Antisense inhibition of this protein ablated so-called 1-alone currents (Tareilus et al. 1997), indicating that the CaVXO participated in the trafficking of the non-native 1 subunits. The question, then, was: why did channels obtained by injecting Xenopus oocytes with 1 alone exhibit clear-cut deficiencies in their gating properties despite the presence of CaVXOTwo mutually exclusive scenarios were proposed that could account for these data, as presented in Fig. 1 (Tareilus et al. 1997; Birnbaumer et al. 1998; Jones, 2002).
Several studies have affirmed the independence of the CaVtrafficking and gating roles without resolving the fundamental issue of whether the two functions are mediated by a single, or multiple CaVsubunits. Dolphin and colleagues (Canti et al. 2001) explicitly demonstrated concentration-dependent effects for CaVon recombinant N-type channels by carefully titrating the amount of CaV3 expression in Xenopus oocytes (Canti et al. 2001). Relatively low CaV3 levels (Kd = 17 nM) were sufficient to increase whole-cell conductance, whereas substantially higher concentrations (Kd = 120 nM) were required to elicit gating effects. Gerster et al. used a mutation of a conserved tyrosine in the AID (Y467S in 1C), previously identified to disrupt the high-affinity 1–CaVinteraction (Pragnell et al. 1994), to examine the separate trafficking and gating effects. They found that CaVs no longer trafficked 1C(Y467S) to the membrane upon cotransfection in tsA201 cells; but single-channel experiments indicated that CaVs still increased Po in 1C(Y467S) channels (Gerster et al. 1999). More recently, truncated splice variants of CaVs that lack a GK domain have been found to increase single-channel Po, while being unable to reconstitute the CaVtrafficking function (Hullin et al. 2003; Cohen et al. 2004). These data suggest the existence of secondary 1–CaVinteraction sites beyond the high-affinity interaction between the -binding pocket (ABP) of the GK domain and the AID. Indeed, lower affinity interactions between some CaVsubunits and the carboxyl termini of 1A and 1E have been identified in biochemical experiments (Qin et al. 1997; Birnbaumer et al. 1998; Walker et al. 1998, 1999; Walker & De Waard, 1998). However, none of these experiments could discriminate whether CaVs modulate channel trafficking and gating according to a single- or multiple-CaV-binding model (Jones, 2002).
Acutely applied CaVsubunits have also been shown to affect the gating activity of channels pre-existing in the plasma membrane. In one study, purified CaV3 protein was introduced into Xenopus oocytes previously injected with 1C cRNA. It was found that CaV3 produced effects on the gating of channels pre-existing at the membrane (leftward shift in the voltage-dependence of channel activation, increased current amplitude presumably due to elevated single-channel Po) long before increasing trafficking of 1C subunits from intracellular sites to the membrane (Yamaguchi et al. 1998). Similarly, acute introduction (via patch pipette dialysis) of purified CaV1a into spherical vesicles derived from skeletal muscle plasma membrane resulted in up-regulation of whole-cell current amplitude without any changes in gating current size (Garcia et al. 2002). These results indicate that under these experimental conditions, acute application of CaVs can modify the gating properties of channels pre-existing at the membrane. Unfortunately, these results could also be well explained by either the single- or multiple-CaV-binding models (Jones, 2002).
In light of these previous results, our finding that fusing a 1 subunit with a single CaVcaptures the functional essence of channels permits an unambiguous interpretation of the structure–function relationship of the calcium-channel 1–CaVinteraction. Although we do not directly measure channel Po in this study, we infer from the similarities in channel trafficking (as gauged by QON) and macroscopic conductance that channels reconstituted with 1C/CaV2b/2 and 1C·2b have similar Po-values. Future single-channel studies will provide more direct insights into microscopic gating properties of 1C·2b vs. 1C/CaV2b-reconstituted L-type calcium channels. From the current data, we cannot unambiguously rule out the notion that a second CaVcould possibly modulate some gating properties of the channel. Overall, the refined interpretations afforded by our results suggest new physiological dimensions to the 1–CaVinteraction that we discuss next.
Reversible unbinding of calcium-channel 1–CaVsubunits
Our results are consistent with a single-CaV-binding model, and in conjunction with previously published studies, an important implication of this model is that at low CaVconcentrations reversible unbinding of CaVensures an appreciable fraction of CaV-less 1 channels in the plasma membrane. The notion that CaVs can reversibly unbind from 1 subunits suggests that CaVmodulation of channel gating could be a more dynamic process than previously appreciated. Because CaV-less channels have a lower Po than CaV-bound channels, reversible unbinding could serve to physiologically switch calcium channels between low- and high-activity modes of signalling. Recently, it was reported that in invertebrate Lymnaea stagnalis neurones, CaVs (LCaV) are associated with LCaV2 channels in mature neurones (Spafford et al. 2004). However, LCaVwas physically uncoupled from LCaV2 channels located in the leading edge of neuronal growth cones (Spafford et al. 2004). Moreover, such uncoupled LCaV2 channels were shown to be important for growth cone projections. Reversible unbinding of CaVfrom 1 subunits could feature prominently in the appearance of such CaV-less channels at the membrane, and could thus serve to significantly enrich the physiologic dimensions of calcium-channel signalling in cells.
Utility of 1–CaVfusion-proteins as tools to probe calcium-channel structure–function relationships
Beyond the use of 1C·2b to resolve the functional stoichiometry of HVA calcium channels as demonstrated here, 1–fusion proteins have tremendous potential to elucidate other currently challenging questions relating to the structure–function of HVA calcium channels. A key example relates to the fact that in many excitable cells multiple HVA calcium-channel 1 and CaVsubunits coexist, and appear to interact in a rather promiscuous manner (Witcher et al. 1995; Scott et al. 1996; Ludwig et al. 1997). What is the physiological significance of the molecularly diverse combinations of calcium-channel 1–subunits in cellsDo distinct 1–combinations transduce unique biological responsesSuch questions are difficult to address given the multiplicity of calcium-channel 1 and CaVsubtypes in cells, and the promiscuity of their interactions. The 1–fusion-protein approach could provide a novel avenue to investigate the physiological significance of unambiguously identified 1–combinations.
References
Beguin P, Nagashima K, Gonoi T, Shibasaki T, Takahashi K, Kashima Y, Ozaki N, Geering K, Iwanaga T & Seino S (2001). Regulation of Ca2+ channel expression at the cell surface by the small G-protein kir/Gem. Nature 411, 701–706.
Birnbaumer L, Qin N, Olcese R, Tareilus E, Platano D, Costantin J & Stefani E (1998). Structures and functions of calcium channel subunits. J Bioenerg Biomembr 30, 357–375.
Brice NL, Berrow NS, Campbell V, Page KM, Brickley K, Tedder I & Dolphin AC (1997). Importance of the different subunits in the membrane expression of the 1A and 2 calcium channel subunits: studies using a depolarization-sensitive 1A antibody. Eur J Neurosci 9, 749–759.
Brody DL, Patil PG, Mulle JG, Snutch TP & Yue DT (1997). Bursts of action potential waveforms relieve G-protein inhibition of recombinant P/Q-type Ca2+ channels in HEK 293 cells. J Physiol 499, 637–644.
Canti C, Davies A, Berrow NS, Butcher AJ, Page KM & Dolphin AC (2001). Evidence for two concentration-dependent processes for -subunit effects on 1B calcium channels. Biophys J 81, 1439–1451.
Catterall WA (2000). Structure and regulation of voltage-gated Ca2+ channels. Annu Rev Cell Dev Biol 16, 521–555.
Chen Y, Li M, Zhang Y, He L, Yamada Y, Fitzmaurice A, Shen Y, Zhang H, Tong L & Yang J (2004). Structural basis of the 1–subunit interaction of voltage-gated Ca2+ channels. Nature 429, 675–680.
Chien AJ, Zhao X, Shirokov RE, Puri TS, Chang CF, Sun D, Rios E & Hosey MM (1995). Roles of a membrane-localized subunit in the formation and targeting of functional L-type Ca2+ channels. J Biol Chem 270, 30036–30044.
Cohen RM, Foell JD, Balijepalli RC, Shah V, Hell JW & Kamp TJ (2004). Unique modulation of L-type Ca2+ channels by short auxiliary 1d subunit present in cardiac muscle. Am J Physiol Heart Circ Physiol 288, H2363–H2374.
Colecraft HM, Alseikhan B, Takahashi SX, Chaudhuri D, Mittman S, Yegnasubramanian V, Alvania RS, Johns DC, Marban E & Yue DT (2002). Novel functional properties of Ca2+ channel subunits revealed by their expression in adult rat heart cells. J Physiol 541, 435–452.
Costantin J, Noceti F, Qin N, Wei X, Birnbaumer L & Stefani E (1998). Facilitation by the 2a subunit of pore openings in cardiac Ca2+ channels. J Physiol 507, 93–103.
De Waard M, Pragnell M & Campbell KP (1994). Ca2+ channel regulation by a conserved subunit domain. Neuron 13, 495–503.
Dolphin AC (2003). subunits of voltage-gated calcium channels. J Bioenerg Biomemb 35, 620.
Finlin BS, Crump SM, Satin J & Andres DA (2003). Regulation of voltage-gated calcium channel activity by the Rem and Rad GTPases. Proc Natl Acad Sci U S A 100, 14469–14474.
Gao T, Chien AJ & Hosey MM (1999). Complexes of the 1C and subunits generate the necessary signal for membrane targeting of class C L-type calcium channels. J Biol Chem 274, 2137–2144.
Garcia R, Carrillo E, Rebolledo S, Garcia MC & Sanchez JA (2002). The 1a subunit regulates the functional properties of frog and mouse L-type Ca2+ channels of skeletal muscle. J Physiol 545, 407–419.
Gerster U, Neuhuber B, Groschner K, Striessnig J & Flucher BE (1999). Current modulation and membrane targeting of the calcium channel 1C subunit are independent functions of the subunit. J Physiol 517, 353–368.
Hullin R, Khan IFY, Wirtz S, Mohacsi P, Varadi G, Schwartz A & Herzig S (2003). Cardiac L-type calcium channel subunits expressed in human heart have differential effects on single channel characteristics. J Biol Chem 278, 21623–21630.
Jones SW (2002). Calcium channels: when is a subunit not a subunitJ Physiol 545, 334.
Jones LP, Wei SK & Yue DT (1998). Mechanism of auxiliary subunit modulation of neuronal 1E calcium channels. J General Physiol 112, 125–143.
Ludwig A, Flockerzi V & Hofmann F (1997). Regional expression and cellular localization of the 1 and subunit of high voltage-activated calcium channels in rat brain. J Neurosci 17, 1339–1349.
McGee AW, Nunziato DA, Maltez JM, Prehoda KE, Pitt GS & Bredt DS (2004). Calcium channel function regulated by the SH3-GK module in subunits. Neuron 42, 89–99.
Opatowsky Y, Chen C-C, Campbell KP & Hirsch JA (2004). Structural analysis of the voltage-dependent calcium channel subunit functional core and its complex with the 1 interaction domain. Neuron 42, 387–399.
Opatowsky Y, Chomsky-Hecht O, Kang MG, Campbell KP & Hirsch JA (2003). The voltage-dependent calcium channel subunit contains two stable interacting domains. J Biol Chem 278, 52323–52332.
Perez-Reyes E, Castellano A, Kim HS, Bertrand P, Baggstrom E, Lacerda AE, Wei XY & Birnbaumer L (1992). Cloning and expression of a cardiac/brain subunit of the L-type calcium channel. J Biol Chem 267, 1792–1797.
Pragnell M, De Waard M, Mori Y, Tanabe T, Snutch TP & Campbell KP (1994). Calcium channel -subunit binds to a conserved motif in the I–II cytoplasmic linker of the 1-subunit. Nature 368, 67–70.
Qin N, Platano D, Olcese R, Stefani E & Birnbaumer L (1997). Direct interaction of Gwith a C-terminal G-binding domain of the Ca2+ channel 1 subunit is responsible for channel inhibition by G protein-coupled receptors. Proc Natl Acad Sci U S A 94, 8866–8871.
Scott VE, De Waard M, Liu H, Gurnett CA, Venzke DP, Lennon VA & Campbell KP (1996). subunit heterogeneity in N-type Ca2+ channels. J Biol Chem 271, 3207–3212.
Singer D, Biel M, Lotan I, Flockerzi V, Hofmann F & Dascal N (1991). The roles of the subunits in the function of the calcium channel. Science 253, 1553–1557.
Spafford JD, Van Minnen J, Larsen P, Smit AB, Syed NI & Zamponi GW (2004). Uncoupling of calcium channel 1 and subunits in developing neurons. J Biol Chem 279, 41157–41167.
Takahashi SX, Miriyala J & Colecraft HM (2004). Membrane-associated guanylate kinase-like properties of subunits required for modulation of voltage-dependent Ca2+ channels. Proc Natl Acad Sci U S A 101, 7193–7198.
Takahashi SX, Mittman S & Colecraft HM (2003). Distinctive modulatory effects of five human auxiliary 2 subunit splice variants on L-type calcium channel gating. Biophys J 84, 3007–3021.
Tareilus E, Roux M, Qin N, Olcese R, Zhou J, Stefani E & Birnbaumer L (1997). A Xenopus oocyte subunit: evidence for a role in the assembly/expression of voltage-gated calcium channels that is separate from its role as a regulatory subunit. Proc Natl Acad Sci U S A 94, 1703–1708.
Tomlinson WJ, Stea A, Bourinet E, Charnet P, Nargeot J & Snutch TP (1993). Functional properties of a neuronal class C L-type calcium channel. Neuropharmacology 32, 1117–1126.
Van Petegem F, Clark KA, Chatelain FC & Minor DL (2004). Structure of a complex between a voltage-gated calcium channel -subunit and an -subunit domain. Nature 429, 671–675.
Walker D, Bichet D, Campbell KP & De Waard M (1998). A beta 4 isoform–specific interaction site in the carboxyl-terminal region of the voltage-dependent Ca2+ channel 1A subunit. J Biol Chem 273, 2361–2367.
Walker D, Bichet D, Geib S, Mori E, Cornet V, Snutch TP, Mori Y & De Waard M (1999). A new subtype–specific interaction in 1A subunit controls P/Q-type Ca2+ channel activation. J Biol Chem 274, 12383–12390.
Walker D & De Waard M (1998). Subunit interaction sites in voltage-dependent Ca2+ channels: role in channel function. Trends Neurosci 21, 148–154.
Wei XY, Perez-Reyes E, Lacerda AE, Schuster G, Brown AM & Birnbaumer L (1991). Heterologous regulation of the cardiac Ca2+ channel 1 subunit by skeletal muscle and subunits. Implications for the structure of cardiac L-type Ca2+ channels. J Biol Chem 266, 21943–21947.
Witcher DR, De Waard M, Liu H, Pragnell M & Campbell KP (1995). Association of native Ca2+ channel subunits with the 1 subunit interaction domain. J Biol Chem 270, 18088–18093.
Witcher DR, De Waard M, Sakamoto J, Franzini-Armstrong C, Pragnell M, Kahl SD & Campbell KP (1993). Subunit identification and reconstitution of the N-type Ca2+ channel complex purified from brain. Science 261, 486–489.
Yamaguchi H, Hara M, Strobeck M, Fukasawa K, Schwartz A & Varadi G (1998). Multiple modulation pathways of calcium channel activity by a subunit. Direct evidence of subunit participation in membrane trafficking of the 1C subunit. J Biol Chem 273, 19348–19356.
Yamaguchi H, Okuda M, Mikala G, Fukasawa K & Varadi G (2000). Cloning of the 2a subunit of the voltage-dependent calcium channel from human heart: cooperative effect of 2 and 2a on the membrane expression of the 1C subunit. Biochem Biophys Res Commun 267, 156–163.
Young K, Lin S, Sun L, Lee E, Modi M, Hellings S, Husbands M, Ozenberger B & Franco R (1998). Identification of a calcium channel modulator using a high throughput yeast two-hybrid screen. Nat Biotechnol 16, 946–950.(Stanislava Dalton, Shoji )