Differential Role of the α1C Subunit Tails in Regulation of the Cav1.2 Channel by Membrane Potential, β Subunits, and Ca2+ Ions*

Voltage-gated Cav1.2 channels are composed of the pore-forming α1C and auxiliary β and α2δ subunits. Voltage-dependent conformational rearrangements of the α1C subunit C-tail have been implicated in Ca2+ signal transduction. In contrast, the α1C N-tail demonstrates limited voltage-gated mobility. We have asked whether these properties are critical for the channel function. Here we report that transient anchoring of the α1C subunit C-tail in the plasma membrane inhibits Ca2+-dependent and slow voltage-dependent inactivation. Both α2δ and β subunits remain essential for the functional channel. In contrast, if α1C subunits with are expressed α2δ but in the absence of a β subunit, plasma membrane anchoring of the α1C N terminus or its deletion inhibit both voltage- and Ca2+-dependent inactivation of the current. The following findings all corroborate the importance of the α1C N-tail/β interaction: (i) co-expression of β restores inactivation properties, (ii) release of the α1C N terminus inhibits the β-deficient channel, and (iii) voltage-gated mobility of the α1C N-tail vis à vis the plasma membrane is increased in the β-deficient (silent) channel. Together, these data argue that both the α1C N- and C-tails have important but different roles in the voltage- and Ca2+-dependent inactivation, as well as β subunit modulation of the channel. The α1C N-tail may have a role in the channel trafficking and is a target of the β subunit modulation. The β subunit facilitates voltage gating by competing with the N-tail and constraining its voltage-dependent rearrangements. Thus, cross-talk between the α1C C and N termini, β subunit, and the cytoplasmic pore region confers the multifactorial regulation of Cav1.2 channels.

regulates Ca v 1.2 conductance by responding to Ca 2ϩ binding that shuttles it between two CaM-binding sites in the proximal half of the ␣ 1C subunit C-terminal tail (1)(2)(3)(4). CaM signals Ca 2ϩ for transcription activation (5) or Ca 2ϩ -induced intracellular Ca 2ϩ release (6) by the voltage-gated rearrangement of the ␣ 1C subunit C terminus, thus linking Ca 2ϩ -dependent inactivation (CDI) and Ca 2ϩ signal transduction (7). With these voltageand Ca 2ϩ -gated rearrangements, the role of the Ca v 1.2 cytoplasmic termini may be further defined when the ␣ 1C subunit tails are uncoupled from the channel regulation by transient immobilization in the plasma membrane.
The association of ␣ 1C with the auxiliary ␣ 2 ␦ and ␤ subunits is important for the functional expression of Ca v 1.2 channels. The cytoplasmic ␤ subunit binds to a conserved "␣-interaction domain" in the ␣ 1C subunit cytoplasmic linker between transmembrane repeats I and II (8,9). The extracellular ␣ 2 subunit is bound via an SS bridge to its post-translationally cleaved transmembrane ␦ peptide (10,11) that renders association with ␣ 1C . Both ␣ 2 ␦ (12)(13)(14) and ␤ subunits (15)(16)(17)(18)(19) modulate the channel. In particular, ␤ subunits affect the time course of the Ba 2ϩ current decay up to 3-fold depending on the type of the ␤ subunit.
To study conformational rearrangements in the channel in response to depolarization, measurements of differential changes in fluorescence resonance energy transfer (FRET) between the cyan (ECFP) and yellow (EYFP) fluorescent proteins fused to the ␣ 1C and ␤ subunit termini have been an effective approach. The current findings begin to specify the central features of conformational rearrangements associated with the transition of the channel from the resting (Ϫ90 mV) to the inactivated state of Ca v 1.2. With the (EYFP) N -␣ 1C,77 -(ECFP) C /␤ 1a /␣ 2 ␦ channel as a model, FRET microscopy showed reversible voltage-gated rearrangements between the ␣ 1C,77 tails and pointed to a role for the C-terminal mobile tail in intracellular Ca 2ϩ signal transduction (5). Another study (20) characterized the voltage-gated rearrangements between the N-terminal tails of the ␣ 1C and ␤ subunits associated with differential ␤ subunit modulation of inactivation and demonstrated limited rearrangements of both N-tails with regard to the plasma membrane.
To investigate further the role of voltage-gated mobility of the ␣ 1C,77 N-terminal tail for function of the Ca v 1.2 channel, here we have investigated the effects of N-terminal deletion or plasma membrane immobilization on Ca 2ϩ -and voltage-dependent inactivation, as well as ␤ subunit regulation of the channel. Most interestingly, if the ␣ 1C /␣ 2 ␦ channel is assembled without a ␤ subunit, either deletion or anchoring of the ␣ 1C,77 N-terminal tail results in high amplitude non-inactivating Ba 2ϩ or Ca 2ϩ currents. This provides a mechanism to explain how limited mobility of the ␣ 1C subunit N-terminal tail  Table 1.
integrates the ␤ subunit and the ␣ 1C subunit C terminus in inactivation of the channel.

MATERIALS AND METHODS
Molecular Biology-Reverse transcription-PCR cloning of human hippocampal ␣ 1C subunit (see Supplemental Material) showed substantial diversity of the transcripts because of alternative splicing. The exon-22 isoform (GenBank TM accession number Z34815) of the human hippocampus ␣ 1C subunit, known as ␣ 1C,77 (21), was selected for this study because it was identified in other human tissues and cells. (EY-FP) N -␣ 1C,77 , (PH-EYFP) N -␣ 1C,77 , ␣ 1C,77 -(PH-ECFP) C , and (ECFP) N -PH and (EYFP) N -PH (22) expression plasmids (N and C indicate the N and C terminus of a subunit, respectively) were prepared in the pcDNA3 vector for eukaryotic expression as described earlier (5) by using pEYFP and pECFP vectors (Clontech). To delete the N-terminal amino acids 2-120 of the ␣ 1C,77 subunit, the PCR product obtained by amplification of pHLCC77 (21) with the sense 5Ј-tggatccgccaccATGGTCGAATG-GAAACCATTTG-3Ј and antisense 5Ј-AGCCATGATCCCATCATACAT-CAC-3Ј primers (noncoding nucleotides are shown in lowercase letters) was digested by BamHI and SgrAI and ligated at the respective sites into 77-pcDNA3 (23). Nucleotide sequences of all PCR and ligation products were confirmed at the DNA sequencing facility of the University of Maryland. The ␤ 1a and ␣ 2 ␦ subunits in pcDNA3 vector were prepared as described by Soldatov et al. (24).
Transient Expression in COS1 Cells-COS1 cells were maintained at 37°C and 5% CO 2 in Dulbecco's modified Eagle's medium supplemented with 10% (v/v) fetal calf serum (Invitrogen). For transient Ca 2ϩ channel expression, cells were plated on poly-D-lysine-coated coverslips (MatTek) 18 h before transfection with cDNAs coding for the indicated ␣ 1C , ␤ 1a , and ␣ 2 ␦ subunits of the Ca 2ϩ channel (1:1:1, w/w) using an Effectene kit (Qiagen) according to the protocol described previously (20). Transfected cells were visualized by ECFP or EYFP tags genetically fused to the ␣ 1C subunits. In some experiments, the G␣ q Q209L mutant (The Guthrie Research Institute, Sayre, PA) was co-expressed with Ca 2ϩ channel subunits.
Electrophysiological Experiments-Whole-cell patch clamp recordings were performed at room temperature (20 -22°C) using the Axopatch 200B amplifier (Axon Instruments) 48 -72 h after transfection. The extracellular bath solution contained the following (in mM): 100 NaCl, 20 BaCl 2 , 1 MgCl 2 , 10 glucose, 10 HEPES, adjusted to pH 7.4, with NaOH. Borosilicate glass pipettes were fire-polished and showed a typical resistance of 3-6 megohms when filled with pipette solution containing the following (in mM): 110 CsCl, 5 MgATP, 10 1,2-bis(2aminophenoxy)ethane-N,N,NЈ,NЈ-tetraacetic acid, 20 tetraethylammonium, 0.2 cAMP, and 20 HEPES, adjusted to pH 7.4 with CsOH (24). Capacitance compensation and series resistance were set at 60%. Currents were sampled at 2.5-5 kHz and filtered at 1 kHz. Voltage protocols were generated, and data were digitized, recorded, and analyzed using pClamp 8.1 software (Axon Instruments). The holding potential (V h ) was Ϫ90 mV, and test pulses were applied at 30-s intervals. Current-voltage (I-V) curves were obtained by step depolarizations to test potentials in the range of Ϫ60 to ϩ50 mV (with 10-mV increments) applied from the holding potential. Statistical values are given as means Ϯ S.E.
Single channel recordings were carried out in cell-attached configurations at 24°C in the bath solution containing the following (in mM): 110 L-aspartic acid, 20 KCl, 2 MgCl 2 , 2 EGTA, and 20 HEPES, adjusted to pH 7.4 with KOH. The pipette solution contained the following (in mM): 110 BaCl 2 , 5 HEPES, adjusted to pH 7.4 with NaOH. The pipettes were heat-polished and showed resistances of 3-4.5 megohms.
Step depolarizations to the indicated potentials were applied from V h ϭ Ϫ90 mV for 1 s followed by inter-pulse intervals of 5 s. Currents were sampled at 40 kHz and low pass-filtered with a gauss filter at 500 Hz. Histograms with a bin size of 0.03 pA were fit with a sum of two gauss functions. Single channel data were analyzed using Clampfit 9.2 software (Axon Instruments) FRET measurements were combined with whole-cell patch clamp as described in detail by Kobrinsky et al. (5,20). Images were recorded with a Hamamatsu digital camera C4742-95 mounted on the Nikon epifluorescent microscope TE200 (60 ϫ 1.2 N.A. objective) equipped with an excitation 75-watt xenon lamp and multiple filter sets (Chroma Technology). Acquisition and analysis of FRET images were carried out with C-Imaging (Compix) and MetaMorph (Universal Imaging) software packages. Conditions for the measurements of corrected images of FRET between (ECFP) N -PH and (EYFP) N -␣ 1C,77 in the absence or on the presence of the ␤ 1a subunit closely matched those described recently (20).

RESULTS
Our goal in this study was to reveal the role of the ␣ 1C,77 subunit C-and N-terminal tails in inactivation of the Ca v 1.2 channel and its regulation by ␤ subunits. To simplify the interpretation of the results, in most of this series of experiments we used the protein kinase C-insensitive Ca v 1.2 assembled with the ␣ 1C,77 subunit identified in human hippocampus (see Supplemental Material). Our previous experiments (5,20,25) provided evidence that electrophysiological properties of the channel remained essentially unaltered by the N-and/or Cterminal fusion of the green fluorescent protein analogs to the ␣ 1C,77 subunit. Measurements of state-dependent FRET, as an indicator of the voltage-gated rearrangements of the ECFP/ EYFP-labeled channel, have demonstrated significant mobility of the ␣ 1C subunit C-terminal tail and its crucial role for Ca 2ϩ signal transduction, contrasting with a relatively weak ␤ subunit dependent mobility of the ␣ 1C N terminus.
Effect of Plasma Membrane Immobilization of the ␣ 1C C Terminus on CDI, Voltage-dependent Inactivation, and Ion Selectivity of the Channel-In the first set of experiments, the C terminus of the ␣ 1C,77 subunit was immobilized to the plasma membrane via the pleckstrin homology (PH) domain of phospholipase C␦1. To prepare the plasmid encoding ␣ 1C,77 -(PH-ECFP) C , the PH domain was genetically fused to the last Cterminal Leu-2138 residue of the ␣ 1C,77 subunit, and the reading frame was completed with the ECFP-coding sequence. The ␤ subunit-deficient ␣ 1C,77 -(ECFP) C /␣ 2 ␦ channel was predominantly localized in the cytoplasm and did not show substantial membrane targeting (for details, see Harry et al. (26)). The C-terminal fusion of the PH domain was sufficient to direct the plasma membrane insertion of the channel, as one can see from the distribution of the ECFP fluorescence across the cell (Fig. 1A, inset). When co-expressed in COS1 cells with the ␣ 2 ␦ subunit but in the absence of ␤ subunits (Fig. 1A), the ␣ 1C,77 -(PH-ECFP) C /␣ 2 ␦ channel generated only a minute Ba 2ϩ current in response to a ϩ10-mV depolarization applied from V h ϭ Ϫ90 mV. However, when ␣ 2 ␦ and ␣ 1C,77 -(PH-ECFP) C were co-expressed with the ␤ 1a subunit (Fig. 1B), the amplitude of the Ba 2ϩ current increased severalfold, and the current decay exhibited a distinctly prolonged plateau at approximately halfmaximum of the current. Fig. 1C shows a set of the representative traces of the Ba 2ϩ current evoked by 600-ms test pulses in the range of 0 to ϩ50 mV (10-mV increments) applied from V h ϭ Ϫ90 mV. The corresponding averaged I-V relation is presented in Fig. 1D. The most prominent feature of the Ba 2ϩ current through the channel with the plasma membrane-anchored ␣ 1C subunit C terminus was the sustained component of the current (Fig. 1D, open circles in the I-V curve) that, in a wide range of the current-evoking potentials, composed 35.5 Ϯ 3.5% (n ϭ 7) of the total Ba 2ϩ current (from 31.4% at Ϫ10 mV to 56.1% at ϩ50 mV). The sustained current was preceded by a rapidly inactivating component with the fast time constant f of 20.0 Ϯ 5.0 ms (at ϩ20 mV; n ϭ 9). Overall, the anchoring of the ␣ 1C subunit C terminus in the plasma membrane accelerated fast inactivation and inhibited slow inactivation of the Ba 2ϩ current.
Release of the ␣ 1C,77 subunit C-tail could be stimulated by the hydrolysis of the PH domain PIP 2 upon activation of phospholipase C. In previous studies, we demonstrated that activation of PIP 2 hydrolysis by epidermal growth factormediated stimulation of the co-expressed epidermal growth factor receptors helped to fully restore the CDI and voltagedependent slow inactivation of the channel (5). Because the epidermal growth factor-mediated recovery was transient and additionally complicated by inhibition of the channel activity (27), we analyzed the effects of the co-expression of the constitutively active mutant (Q209L) of the G␣ q protein that depletes the plasma membrane PIP 2 (e.g. see Howes et al. (28)). Fig. 1E (inset) shows the fluorescence image of a COS1 cell co-expressing the ␣ 1C,77 -(PH-ECFP) C , ␣ 2 ␦, and ␤ 1a subunits, the G␣ q Q209L mutant, and representative traces of the Ba 2ϩ current generated in response to step depolarizations in the range of ϩ10 to ϩ40 mV applied from V h ϭ Ϫ90 mV. Co-expression of the G␣ q Q209L mutant shifted inacti- Representative Ba 2ϩ currents were recorded in cells of approximately the same diameter in response to a ϩ10-mV test pulse applied for 600 ms from V h ϭ Ϫ90 mV. Dotted lines indicate zero current. C, typical traces of the Ba 2ϩ current through the ␣ 1C,77 -(PH-ECFP) C /␣ 2 ␦/␤ 1a channel evoked by indicated test pulses applied from V h ϭ Ϫ90 mV. D, I-V relations (n ϭ 13) measured for the peak Ba 2ϩ currents (closed circles) and sustained currents (mean values determined in the range of 400 -500 ms, open circles). The averaged curves were normalized to the maximum current. Control (filled squares) for the channel co-expressed with the G␣ q Q209L mutant was fitted (smooth line) by equation where G max is maximum conductance; E rev ϭ 68.9 Ϯ 4.6 mV is the reversal potential; V 0.5 ϭ Ϫ9.5 Ϯ 2.4 mV is voltage at 50% of I Ba activation; and k I-V ϭ Ϫ10.2 Ϯ 1.2 is slope factor. E, epifluorescent image of the expressing cell and representative traces of the Ba 2ϩ current evoked by indicated test pulses applied from V h ϭ Ϫ90 mV. F, steady-state inactivation curves for Ba 2ϩ currents through the ␣ 1C,77 -(PH-ECFP) C /␣ 2 ␦/␤ 1a channel assembled in the absence (filled circles; n ϭ 3) and presence (filled squares; n ϭ 5) of the constitutively active G␣ q Q209L mutant. A 1-s conditioning prepulse was applied from V h ϭ Ϫ90 mV (10-mV increments up to ϩ60 mV) followed by a 100-ms test pulse to ϩ10 mV. The intervals between each cycle were 15 s. The peak current amplitudes in each curve were normalized to the maximum value determined in the range of Ϫ60 to ϩ50 mV. The G␣ q Q209L mutant control (smooth line) was fitted by Boltzmann function: A and B are fractions of non-inactivating (15.1 Ϯ 2.4%) and inactivating channels, respectively; V is the conditioning prepulse voltage; V 0.5 ϭ Ϫ8.6 Ϯ 1.7 mV is the voltage at half-maximum of inactivation, and k ϭ 15.7 Ϯ 1.9 is a slope factor. G, typical traces of Ca 2ϩ currents through the ␣ 1C,77 -(PH-ECFP) C /␣ 2 ␦/␤ 1a channel activated by indicated test pulses applied from V h ϭ Ϫ90 mV. H, averaged I-V relationships (normalized to the maximum current) for the peak (filled circles) and sustained (determined as in D, open circles) Ca 2ϩ currents (n ϭ 10). Scaling bars, 3 m. vation of the channels to the normal phenotype with a prominent slow component. Single exponential fitting shows the time constant of the Ba 2ϩ current decay of 200 Ϯ 38 ms (at ϩ10 mV; n ϭ 9) and the lack of sustained component.
The inability of the channel with the immobilized C-terminal tail to complete inactivation was corroborated by the analysis of steady-state inactivation properties (Fig. 1F). Approximately 70% of the Ba 2ϩ current through the ␣ 1C,77 -(PH-ECFP) C /␣ 2 ␦/ ␤ 1a channel evoked by a ϩ10-mV depolarization remained noninactivated after a depolarizing prepulse in a range of Ϫ10 to ϩ50 mV was applied from V h ϭ Ϫ90 mV prior to the ϩ10-mV test pulse (Fig. 1F, closed circles). The release of the ␣ 1C,77 C-tail restored inactivation of the channel, as can be seen from a comparison of the Ba 2ϩ current decay (Fig. 1E) and steadystate inactivation curves recorded in the absence and in the presence of the constitutively active G␣ q Q209L mutant (Fig. 1F).
Ca v 1.2 channels classically inactivate by a combination of the voltage-and Ca 2ϩ -dependent mechanisms. One of the main consequences of the replacement of Ba 2ϩ for Ca 2ϩ as the charge carrier is an acceleration of the macroscopic current decay or CDI (29). However, this was not found to hold true for the C-terminal tail-anchored Ca 2ϩ channel. Indeed, the representative traces of the Ca 2ϩ current through the ␣ 1C,77 -(PH-ECFP) C /␣ 2 ␦/␤ 1a channel ( Fig. 1G), recorded from the same cell as Ba 2ϩ currents in Fig. 1C, show both inactivating and sustained components of the decay. Similar to the Ba 2ϩ current, the large sustained Ca 2ϩ current components lasted for the duration of depolarization at all indicated test pulses (ϩ20 to ϩ50 mV). Although the fast component of inactivation was not as prominent as in the case of the Ba 2ϩ current, the average fraction of the sustained currents (44.9 Ϯ 2.3%, from 33.3% at Ϫ10 mV to 48.6% at ϩ50 mV, Fig. 1H) was essentially the same as with Ba 2ϩ as the charge carrier. All these features of the C-terminal tail-anchored ␣ 1C,77 -(PH-ECFP) C channel closely matched the phenotype of the ␣ 1C,IS-IV channel, which has the slow inactivation mechanism inhibited by the specific mutation in the cytoplasmic pore region (30).
The I-V relations for the Ba 2ϩ -and Ca 2ϩ -conducting ␣ 1C,77 -(PH-ECFP) C channel show a number of common patterns ( Fig.  1, D and H). Immobilization of the ␣ 1C,77 subunit C-terminal tail caused a shift in activation of ion conductance by 10 -15 mV to more positive potentials. Currents reached the peak of the I-V relationship at ϩ20 (Ca 2ϩ ) or ϩ30 mV (Ba 2ϩ ), exhibiting a 10 -20-mV shift of the maximum toward more positive voltages as compared with the ␣ 1C,77 -(ECFP) C /␣ 2 ␦/␤ 1a channel. The apparent reversal potential was also notably changed. Both the Ba 2ϩ and Ca 2ϩ currents reversed direction at much higher voltages than in the wild-type channel. Although this effect was not investigated in detail, it may be due to altered ion selectivity and/or decreased permeability to Cs ϩ ions (introduced in electrodes) in the outward direction that contributed significantly to the apparent reversal potential of the wild-type channel (Ϸ ϩ65 mV) (31). Release of the ␣ 1C,77 -(PH-ECFP) C subunit C-terminal tail by co-expression of the G␣ q Q209L mutant reversed these changes in the I-V relationship to parameters that are more characteristic for the ␣ 1C,77 -(ECFP) C / ␣ 2 ␦/␤ 1a channel, including ion selectivity and the peak current voltage (Fig. 1D, filled squares). The permeability of the wild type ␣ 1C,77 channel to Ba 2ϩ is on average 2.8 times greater than that of the Ca 2ϩ ions (24,30). Immobilization of the ␣ 1C subunit C-tail reduced the difference between the maximum amplitudes of the Ba 2ϩ and Ca 2ϩ currents (compare traces in Fig. 1, C and G), again indicating that the anchoring of the ␣ 1C subunit C-terminal tail may affect the ion conductance of the channel. Taken together, these data support the model (7) that links CDI and slow inactivation of the Ca v 1.2 channel to specific folding of the ␣ 1C subunit C terminus vis à vis the cytoplasmic pore region.
␤ Subunit Facilitation of the Ca v 1.2 Channel Gating Is Revealed by Immobilization of the ␣ 1C Subunit N-terminal Tail-In the second set of experiments, the N-terminal tail of the (EYFP) N -␣ 1C,77 subunit was anchored to the plasma membrane via the PH domain. The (PH-EYFP) N -␣ 1C,77 channel was expressed in COS1 cells in different combinations with auxiliary ␤ 1a and ␣ 2 ␦ subunits. Fig. 2 shows a collection of representative traces of the Ba 2ϩ current elicited by depolarization to ϩ10 mV from V h ϭ Ϫ90 mV in a set of COS1 cells of approximately similar size. The insets in Fig. 2 are fluorescent images of the expressing cells showing subcellular localization of the EYFP-tagged ␣ 1C subunits. Confirming earlier data (32,33), COS1 cells did not show appreciable endogenous expression of Ca v 1.2 ( Fig. 2A). Transfection of COS1 cells with a mixture of cDNAs coding for the (EYFP) N -␣ 1C,77 , ␤ 1a and ␣ 2 ␦ subunits (Fig. 2B) renders the current with characteristics (Fig. 3, curves 3) closely resembling those of the wild-type channel (24). Anchoring of the ␣ 1C subunit N terminus by co-expression of (PH-EYFP) N -␣ 1C,77 with the ␤ 1a and ␣ 2 ␦ subunits ( Fig. 2C) accelerated the kinetics of inactivation of the Ba 2ϩ current by ϳ15% (5). An acceleration of inactivation was also seen when the PH domain was separately co-expressed with the (EYFP) N -␣ 1C,77 , ␤ 1a , and ␣ 2 ␦ subunits (Fig. 2D). Making no assumptions about the nature of these variations, we investigated the effect of the deletion of the ␤ 1a subunit from the expressed constituents of the channel. When (EYFP) N -␣ 1C,77 was expressed alone, the fluorescenttagged channel protein was diffusely distributed over the cytoplasm and did not generate measurable voltage-gated current (Fig. 2E). The N-terminal fusion of the PH domain caused robust surface membrane targeting by the labeled (PH-EY-FP) N -␣ 1C,77 protein, but the channel remained essentially silent (Fig. 2F). A similar result was obtained when ␣ 2 ␦ was co-expressed with the (EYFP) N -␣ 1C,77 subunit (Fig. 2G, note poor membrane targeting of the channel complex). However, the membrane anchoring of the ␣ 1C subunit N-terminal tail by co-expression of the (PH-EYFP) N -␣ 1C,77 and ␣ 2 ␦ subunits stimulated membrane targeting of the ␤-deficient channel that generated large slowly activating and non-inactivating Ba 2ϩ current in response to depolarization (Fig. 2H). This result is consistent with the idea that ␣ 1C and ␣ 2 ␦ are sufficient for the expression of a conducting L-type Ca 2ϩ channel if the N-terminal tail of the ␣ 1C subunit is immobilized by the plasma membrane anchoring. To find whether or not release of the PHtagged N-terminal tail would compromise the channel activity in the ␤ subunit-deficient channel, we investigated the effect of co-expression of the constitutively active Q209L mutant of G␣ q . A G␣ q -mediated depletion of PIP 2 in the plasma membrane did abolish the plasma membrane targeting by inhibiting the immobilization of the ␣ 1C,77 subunit N-terminal tail, and precluded expression of the conducting channel (Fig. 2I). Together, these data suggest that the ␣ 1C N-terminal tail may act as a silencer of the channel voltage-gated conductance and block it in the absence of the ␤ subunit. Results obtained with the ␣ 1C subunit having a genetically immobilized N-terminal tail further corroborate this suggestion by demonstrating robust current of the ␤ subunit-deficient Ca v 1.2 channel composed of the (PH-EYFP) N -␣ 1C,77 and ␣ 2 ␦ subunits.
To characterize further the inactivation characteristics of the Ba 2ϩ current through the (PH-EYFP) N -␣ 1C,77 /␣ 2 ␦Ϯ ␤ 1a channels, we measured steady-state inactivation curves by using a two-step voltage clamp protocol (Fig. 3C). Steady-state inactivation properties point to differences in voltage dependence of inactivation of the Ba 2ϩ current as compared with the control (EYFP) N -␣ 1C,77 /␤ 1a /␣ 2 ␦ channel (Fig. 3C, curve 3). Voltage at the half-maximum of inactivation was shifted from Ϫ7.9 Ϯ 0.7 mV (n ϭ 6) in control to Ϫ15.9 Ϯ 3.4 (n ϭ 3) in the (PH-EYFP) N -␣ 1C,77 /␤ 1a /␣ 2 ␦ channel (Fig. 3C, curve 1). Although the voltage dependence of availability (20.1 Ϯ 1.5%) was not significantly changed by the plasma membrane immobilization of the ␣ 1C,77 N-terminal tail (21.0 Ϯ 5.9%), the slope k of the steady-state inactivation curve (fitted by the Boltzmann equation) decreased from 6.3 Ϯ 0.6 to 13.5 Ϯ 3.8 (p Ͻ 0.05). Thus, cooperativity in voltage gating leading to inactivation is significantly affected by the plasma membrane immobilization of the ␣ 1C subunit N-terminal tail. Steady-state inactivation analysis also confirmed that the voltage dependence of inactivation of the ␤ subunit-deficient channel is inhibited in the range of conditioning pulses of up to ϩ60 mV (Fig. 3C, curve 2).
The single channel analysis corroborated the macroscopic data (Fig. 4, A and B). Recordings of the Ba 2ϩ current in the cell-attached configuration (110 mM Ba 2ϩ ) revealed rarely occurring activations of the (PH-EYFP) N -␣ 1C,77 /␣ 2 ␦ channel that were characterized by low probabilities of both opening and closing (Fig. 4B). Openings of the channel were preceded by a large number of blank traces. The loss of inactivation of this channel was evidenced by continuous bursting of a single Ca 2ϩ channel activity during long lasting depolarizations (mean burst length was 154 Ϯ 18 ms, n ϭ 39). In the presence of the ␤ subunit (Fig. 4A), the appearance of bursts was as low as 1 burst for 100 traces, whereas in the absence of the ␤ subunit, we observed bursts almost in every trace with the channel activity (Fig. 4B). The first latency was significantly prolonged ( ϭ 55.4 ms) compared with the control (EYFP) N -␣ 1C,77 /␤ 1a /␣ 2 ␦ channel ( ϭ 28.4 ms). This explains the delay in activation of the whole-cell current in the anchored (PH-EYFP) N -␣ 1C,77 /␣ 2 ␦ channel.
The characteristic, slowly activating and non-inactivating phenotype of the ␤ subunit-deficient (PH-EYFP) N -␣ 1C,77 /␣ 2 ␦ channel was also found to hold true with Ca 2ϩ as the charge carrier. Fig. 5A shows superimposed traces of the Ca 2ϩ current evoked by depolarization in the range of Ϫ20 to ϩ50 mV applied from V h ϭ Ϫ90 mV. A remarkable feature of the ␤-deficient channel is the striking similarity in the properties of the Ba 2ϩ and Ca 2ϩ currents. A single-exponential fit of the activation time course (ϩ20 mV) shows a marked increase (p Ͻ 0.05) in the time constant of activation of the Ca 2ϩ current through the ␤-deficient (PH-EYFP) N -␣ 1C,77 /␣ 2 ␦ channel ( a ϭ 69.0 Ϯ 12.5 ms, n ϭ 32) as compared with the (PH-EYFP) N -␣ 1C,77 /␤ 1a / ␣ 2 ␦ channel ( a ϭ 2.0 Ϯ 0.2 ms; n ϭ 5; Fig. 5C). These data are not significantly different (unpaired t test) from those obtained for the Ba 2ϩ current (see above). The ␤ subunit-deficient channel exhibits unusually high Ca 2ϩ current amplitude but did not reveal appreciable decay of the Ca 2ϩ current even when the test pulse duration was prolonged to 30 s (Fig. 5B). These data indicate that CDI was lost by the ␤ subunit-deficient Ca v 1.2 channel when its ␣ 1C subunit N-tail was immobilized in the plasma membrane. Release of the N terminus of the (PH-EYFP) N -␣ 1C,77 subunit by the co-expression of a constitutively active Q209L mutant of G␣ q to activate PIP 2 hydrolysis (Fig.  5D) sharply reduced the size of the Ca 2ϩ current through the ␤ subunit-deficient channel in response to a ϩ20-mV depolarization that corresponds to other controls discussed above (Fig. 2,  G and I). A small residual non-inactivating current with the average amplitude of 30 pA may be due to incomplete PIP 2 hydrolysis rendering plasma membrane immobilization of the ␣ 1C subunit N-tail in a small fraction of the channels (see also Fig. 2I). Taken together, these data suggest that inactivation by both the voltage-dependent and CDI mechanisms is missing in the ␤ subunit-deficient (PH-EYFP) N -␣ 1C,77 /␣ 2 ␦ channel.
Effect of the ␣ 1C Subunit N-terminal Deletion on the ␤ Subunit Modulation of the Ca 2ϩ Channel Gating-We have next explored another way of uncoupling the ␣ 1C subunit N-tail from its apparent silencing function by a genetically encoded deletion of 120 out of a total 124 N-terminal amino acids of ␣ 1C,77 (⌬ N -␣ 1C,77 ). Shistik et al. (34) made a similar attempt in Xenopus oocytes by expressing a rabbit cardiac ␣ 1C where all but 15 N-terminal tail amino acids were deleted. However, because Xenopus oocytes are known to express endogenous Ca 2ϩ channels (35)(36)(37)(38), there was ambiguity in their findings. Interpretation was also obscured by the continuous application of 1 M (Ϫ)-BayK8644 to artificially increase the amplitude of the single channel Ba 2ϩ current. More substantial deletion of the ␣ 1C subunit N-terminal tail as well as the use of COS1 cells, free of endogenous Ca 2ϩ channel subunits, helped us to eliminate these ambiguities. Fig. 6 shows Ca 2ϩ channel activity of the ⌬ N -␣ 1C,77 /␣ 2 ␦ channel assembled with (A) or without (B) the ␤ subunit. As was observed with the full size ␣ 1C,77 (for details see Refs. 20 and 26), the co-expression of ⌬ N -␣ 1C,77 and ␣ 2 ␦ increased the plasma membrane targeting by the (ECFP) N -␤ 1a subunit (Fig. 6A, panel a). The surface membrane targeting by the labeled ⌬ N -␣ 1C,77 -(ECFP) C subunit (and the (ECFP) N -␣ 2 ␦; Fig. 6B, panel a) was prominent independently on the presence of the ␤ subunit (Fig. 6, A and B, panel b). The presence of ␤ was not crucial for the functional expression of the ⌬ N -␣ 1C,77 / ␣ 2 ␦Ϯ ␤ 1a channels. In both channels, Ca 2ϩ currents were activated by stepwise depolarization from V h ϭ Ϫ90 mV (shown are traces recorded in response to ϩ20-mV test pulses). This result supports the data on the ␣ 1C,77 N-tail anchoring that a ␤ subunit is not essential for the expression of an active conducting channel if the N-terminal tail of the ␣ 1C subunit is uncou- FIG. 3. Effect of the ␤ subunit on inactivation properties of the Ca 2؉ channel with the N terminus of the ␣ 1C subunit immobilized in the plasma membrane. The (PH-EYFP) N -␣ 1C,77 and ␣ 2 ␦ subunits were co-expressed in COS1 cells with (curve 1) or without ␤ 1a (curve 2). The channel assembled of the (EYFP) N -␣ 1C,77 , ␣ 2 ␦, and ␤ 1a subunits was used as control (curve 3). A, superimposed curves of the Ba 2ϩ current evoked by stepwise depolarization to ϩ20 mV applied from V h ϭ Ϫ90 mV. Traces were normalized to the same current amplitude. B, the I-V relationships for the channels composed of (curve 1) (PH-EYFP) N -␣ 1C,77 /␣ 2 ␦/␤ 1a (n ϭ 6), (curve 2) (PH-EYFP) N -␣ 1C,77 /␣ 2 ␦ (n ϭ 13), and (curve 3) (EYFP) N -␣ 1C,77 /␣ 2 ␦/␤ 1a (n ϭ 6). Ba 2ϩ currents were measured with 15-s intervals between 600-ms test pulses in the range of Ϫ60 to ϩ50 mV applied from V h ϭ Ϫ90 mV. C, an averaged steady-state curves recorded with the channels composed of (curve 1) (PH-EYFP) N -␣ 1C,77 /␣ 2 ␦/␤ 1a (n ϭ 3), (curve 2) (PH-EYFP) N -␣ 1C,77 /␣ 2 ␦ (n ϭ 6), and (curve 3) (EYFP) N -␣ 1C,77 /␣ 2 ␦/␤ 1a (n ϭ 6). The indicated 1-s conditioning pulses were applied from V h ϭ Ϫ90 mV and followed by the 100-ms test pulse to ϩ10 mV. Interval between pulses was 15 s. Recorded peak currents were normalized to the maximum value determined in the range between Ϫ60 and ϩ40 mV. Curves 1 and 3 were fitted by Boltzmann function.
pled from the channel regulation. Although the ␤ subunit was crucial for inactivation of the (PH-EYFP) N -␣ 1C,77 /␣ 2 ␦/␤ 1a channel (Fig. 2C), this effect was essentially missing in the ⌬ N -␣ 1C,77 /␣ 2 ␦/␤ 1a channel (Fig. 6A). The Ca 2ϩ current evoked by the ϩ20-mV step depolarization had rather small amplitude and did not show accelerated decay that would be characteristic for CDI. This result indicates that the functional uncoupling of the ␣ 1C subunit N-terminal tail through the plasma membrane anchoring or deletion influences the interaction with the ␤ subunit in different ways.
Similar to the ␤-deficient (PH-EYFP) N -␣ 1C,77 /␣ 2 ␦ channel that shows non-inactivating Ca 2ϩ currents (Fig. 5B), inactivation of the ⌬ N -␣ 1C,77 /␣ 2 ␦Ϯ ␤ 1a channels was impaired (Fig. 6, A  and B). In addition, removal of the ␤ subunit from the channel FIG. 4. Effects of the ␣ 1C,77 subunit N terminus uncoupling by anchoring or deletion on single channel Ba 2؉ currents through the ␤ subunit-deficient Ca 2؉ channels expressed in COS1 cells. Ba 2ϩ currents were measured in 110 mM Ba 2ϩ and high K ϩ solution at a stimulation rate of 0.1 Hz. A and B, effect of anchoring of the ␣ 1C,77 subunit N terminus in the plasma membrane with (A) and without ␤ 1a (B) subunit expressed. (PH-EYFP) N -␣ 1C,77 /␣ 2 ␦ channel current was activated by a test pulse to 0 mV. Insets show detailed single channel activity. Single channel averages are depicted in gray under the original traces in A and B with exponential fit and mean level, respectively, (n ϭ 74 and n ϭ 34). Respective first latency histograms are also shown with their exponential fits. C and D, effect of the ␣ 1C,77 subunit N terminus deletion with (C) and without ␤ 1a (D) subunit expressed. The ⌬ N -␣ 1C,77 /␣ 2 ␦ channel current was activated by a test pulse to Ϫ10 mV. Length of bursts was 234.7 Ϯ 31.6 ms, (n ϭ 40) (C); and 361.9 Ϯ 70.3 ms, (n ϭ 23, p Ͻ 0.05) (D). complex significantly delayed activation of the Ca 2ϩ current (Fig. 6B). However, analysis of the respective I-V relations (Fig.  6C) showed that the voltage sensors responsible for activation of the ⌬ N -␣ 1C,77 /␣ 2 ␦Ϯ ␤ 1a channels were not strongly affected by the removal of the ␤ subunit from the oligomeric complex, despite other signature changes also seen with the N-tailanchored ␣ 1C,77 channel (Fig. 5A), including broadening of the I-V curves and a shift of the peak currents to more positive potentials (Fig. 6C). Single channel recordings (Fig. 4D) revealed a relatively large number of empty swipes corresponding to low probability of opening of the ␤ subunit-deficient ⌬ N -␣ 1C,77 /␣ 2 ␦ channel. First latencies were significantly shorter in the presence of the ␤ subunit, ϭ 48.5 ms (Fig. 4C), than without it, ϭ 139.4 ms (Fig. 4D). However, the important difference with the effect of the transient anchoring of the ␣ 1C subunit N-terminal tail (Fig. 4, A and B) is clearly visible; the N-tail uncoupling through a deletion essentially stabilized the open state of the ␤ subunit-deficient channel, which showed long opening of a single Ca 2ϩ channel activity during long lasting depolarizations (Fig. 4D). With either type of uncoupling of the ␣ 1C subunit N-terminal tail, a delay in the activation of the whole-cell current appears to be associated with prolongation of the first latency.
Effect of the ␤ Subunit on the Voltage-gated Mobility of the ␣ 1C Subunit N-terminal Tail-Finally, we assessed whether the ␤ subunit affects the voltage-gated mobility of the ␣ 1C subunit N-terminal tail. The (EYFP) N -␣ 1C,77 /␣ 2 ␦ channel was co-expressed with the (ECFP) N -PH in the absence (Fig. 7A) or in the presence (Fig. 7B) of the ␤ 1a subunit (for arrangement of fluorophores and subunits, see panels b). In this experiment, the EYFP-labeled N-terminal tail of the ␣ 1C,77 subunit reports its state-dependent position to the FRET partner, ECFP, fused to the plasma membrane-trapped PH domain. FRET was confined to the plasma membrane region (Fig. 7, yellow boxes in  panels a) and was set to be recorded within the time windows, marked by black or red bars under the actual Ba 2ϩ current traces on panels c, simultaneously with acquisition of the currents. An image recorded at the resting state (Ϫ90 mV, Fig. 7, black bars corresponding to panels d) was followed by an image recorded at the end of a ϩ20 mV depolarization test pulse (red bars corresponding to panels e). One can see that independent of the presence of the ␤ 1a subunit, a corrected FRET signal of different intensity was observed in Fig. 7, panels d and e. Ratios (Fig. 7, panels f) between the two consecutive images recorded at Ϫ90 and ϩ20 mV reflect voltage-dependent conformational rearrangements of the N-tail fluorophore fused to the ␣ 1C,77 subunit with regard to the relatively small membranetrapped PH domain probe (5). In the presence of the ␤ 1a subunit (Fig. 7B), the channel generated a Ba 2ϩ current in response to a ϩ20-mV depolarization (V h ϭ Ϫ90 mV) that showed complete inactivation within 500 ms (Fig. 7B, panel c). Thus, the ϩ20-mV corrected FRET image recorded at the end of the 600-ms depolarizing pulse (Fig. 7B, panel e) corresponds to a predominantly inactivated state of the channel. However, the ratio (Fig. 7B, panel f) shows very little if any change in corrected FRET between Ϫ90 and ϩ20 mV (0.90 Ϯ 0.06; n ϭ 6). A similar result was observed with another type of ␤ subunit, the rabbit ␤ 2 (20).
Deletion of the ␤ subunit (Fig. 7A) renders the channel to a voltage-insensitive silent state (see a zero current trace recorded in response to a ϩ20-mV test pulse in panel c). However, the ratio of corrected FRET images recorded at Ϫ90 and ϩ20 mV (1.70 Ϯ 0.10; n ϭ 5) at the time windows marked below the current trace by black and red bars, respectively, shows a substantial (p Ͻ 0.005) differential FRET signal generated in the plasma membrane by voltage-gated rearrangement of the ␣ 1C,77 subunit N-terminal tail. DISCUSSION Ca v 1.2 channels play an important role in initiation of Ca 2ϩ signaling in many cells, including neurons (39). Here we have investigated differential roles of the C-and N-terminal tails of one of the neuronal "short" (exon 1) ␣ 1C isoforms in voltage-, Ca 2ϩ -, and ␤ subunit-dependent regulation of the Ca v 1.2 channel.
Our study is an important step forward in understanding Ca v 1.2 channel regulation as a multifactorial process. Here for the first time we have analyzed differential regulation of the Ca 2ϩ channel by the C-and N-terminal tails of the ␣ 1C subunit with simultaneous assessment of the role of ␤ subunits. To deconvolute the contribution of these parts, transient anchoring to the plasma membrane via the PH domain fused to the ␣ 1C subunit tails has been used to uncouple the ␣ 1C tails from the channel regulation. To assess the potential interplay between the tails and a ␤ subunit, we compared properties of the channel expressed in the absence and in the presence of ␤. The accumulated body of data shows that inactivation of the Ca v 1.2 channel is mediated by a combinatorial input of these distinct parts, in which CDI is essentially additive to voltage-dependent inactivation.
Voltage-dependent slow inactivation and CDI were previously co-identified with the calmodulin-binding regions in the middle part of the ␣ 1C subunit C-tail (for review see Ref. 40). A recent picture (41) still includes an EF hand-like motif of the C-tail proximal region, although involvement of this motif in CDI has been contested by experimental evidence (29). Our findings show that CDI is not mediated solely by determinants of the ␣ 1C C-terminal tail. Investigation of other cytoplasmic constituents of the channel complex have been crucial in identifying the roles of the N-tail, the cytoplasmic pore region of ␣ 1C , and ␤ subunits for inactivation and in understanding how slow and fast voltage-dependent inactivation and CDI evolve from their interplay. A distinct advantage of COS1 cells lacking endogenous Ca 2ϩ channel subunits (32) over other expression systems, such as HEK293 cells or Xenopus oocytes, permits an unambiguous interpretation of our data on ␤ subunit modulation of the channel. Because no appreciable Ca 2ϩ channel activity was exhibited by the ␣ 1C,77 /␣ 2 ␦ channel (Fig. 2G), unless a ␤ subunit was co-expressed, we define the ␤ subunit modulation as a facilitation of voltage gating of the channel. Recognizing that the kinetics of inactivation of the Ba 2ϩ current through the "wild-type" ␣ 1C,77 /␤ 1a /␣ 2 ␦ channel is best fit by a sum of two (fast and slow) exponentials, the respective fractional components of the Ba 2ϩ current measured at the peak of I-V curves were analyzed to assess the voltage-dependent fast and slow inactivation mechanisms.
We have shown previously that four highly conserved amino acids of the transmembrane segment S6, constituting the cytoplasmic end of the pore, jointly form the ADSI. Their simultaneous mutation (S405I in IS6, A752T in IIS6, V1165T in IIIS6, and I1475T in IVS6) generates the ␣ 1C,IS-IV channel, which does not show the slow inactivation, CDI, or differential ␤ subunit modulation (30). Our findings have shown that the ␣ 1C subunit C-terminal tail is subject to voltage-gated conformational rearrangements that specifically deliver Ca 2ϩ signaling to downstream targets involved in cAMP-response elementbinding protein-dependent transcription activation (5). The plasma membrane anchoring of the C-tail interrupts Ca 2ϩ signal transduction despite robust Ca 2ϩ current through the channel. Thus, a specific position of the C-terminal tail vis à vis the polypeptide packing of the cytoplasmic channel constituents (including the ␣ 1C subunit tails and a ␤ subunit) is crucial for CDI and slow voltage-dependent inactivation.
We now find that plasma membrane immobilization of the ␣ 1C,77 C-terminal tail alters inactivation of the channel similar to the mutation of the ADSI (Fig. 1). An interesting distinction from the electrophysiological phenotype of the ␣ 1C,IS-IV channel, however, is a shift of the I-V relationship to more positive potentials. Because these dramatic changes do not involve structural alterations of the ADSI or the calmodulin-binding regions, and are completely reversible upon the release of the ␣ 1C,77 -(PH-ECFP) C subunit C-tail, stimulated by PIP 2 hydrol-ysis, it is reasonable to assume that the plasma membrane anchoring of the ␣ 1C C-terminal tail interferes with the functional folding of the cytoplasmic polypeptide bundle in the pore region, including ADSI. This may directly affect the pore-forming transmembrane segment IVS6, connected to the ␣ 1C C-tail, and, indirectly, the ensemble of the pore conformations defining the ion selectivity of the channel.
A very different role was found for the ␣ 1C subunit N-terminal tail. FRET microscopy combined with the patch clamp (20) revealed limited voltage-dependent mobility of the ␣ 1C N terminus in relationship with the plasma membrane (Fig. 7B). This may correspond to the restricted local dynamics of the adjacent S1 segment observed in other channels (42). Here we find that in the ␤ subunit-deficient (silent) channel, voltagegated conformational rearrangements of the ␣ 1C subunit N terminus vis à vis the plasma membrane are significantly increased (Fig. 7B). The functional importance of the conformational "rigidity" to voltage gating conferred by a ␤ subunit to the ␣ 1C subunit N terminus was confirmed here by anchoring of the N-tail to the plasma membrane, stabilized by the fused PH domain.
The ␤ subunit confers conformational rigidity to the ␣ 1C subunit N-tail in a manner that facilitates the channel response to voltage gating. We have found that the plasma membrane-anchoring of the ␣ 1C subunit N terminus, in the absence of ␤, completely inhibited inactivation of the Ba 2ϩ (or Ca 2ϩ ) current. At the same time, activation of the channel current in response to depolarization becomes slow independently of the charge carrier (Fig. 2H). The single channel study pointed to low probability of both opening and closing of the channel (Fig.  4B). Uncoupling of the N-terminal tail by its deletion from the ␣ 1C subunit also inhibited inactivation of the channel (Fig. 7). However, in this case retardation of the channel activation was enhanced to a greater extent, probably because of long lasting episodes of rare openings (Fig. 4D). Release of the plasma membrane-anchored ␣ 1C N terminus essentially blocked the channel current (Fig. 2I), leaving only some residual Ca 2ϩ channel activity that may be generated by the remaining small fraction of channels with immobilized ␣ 1C N-tails. Overall, these data correspond to the result obtained with the ␤-deficient Ca v 1.2 channel (Fig. 2G). Conversely, co-expression of ␤ with ␣ 2 ␦ subunits re-established both the voltage-and Ca 2ϩdependent inactivation of the channels despite uncoupling of the ␣ 1C subunit N-terminal tail from the regulation by its plasma membrane anchoring. These observations shed new light on the role of the ␤ subunit. It appears that there is a ␤ subunit-dependent desensitization of the ␣ 1C subunit N-tail to voltage, which is essential for the fast activation, CDI, and voltage-dependent inactivation of the channel.
Our data support the conclusion of Dascal and co-workers (34) that ␤ subunits prevent inhibition of the channel by the N-tail. The results of our work demonstrate that the N-terminal tail of the ␣ 1C subunit is a channel silencer, competing with the ␤ subunit facilitation of the channel gating. This may account for CDI and voltage-dependent inactivation and activation of the current.
Our findings add to the understanding of the ␤ subunit regulation of the Ca v 1.2 channel in a number of ways. The "chaperon" hypothesis implies that ␤ subunits ease functional expression of the Ca v 1.2 channel by binding to the ␣-interaction domain resulting in inhibition of an endoplasmic reticulum retention signal of the ␣ 1C subunit (43). The present study does not support this view. In COS1 cells, free of endogenous Ca 2ϩ channel subunits, an uncoupling of the ␣ 1C subunit N-terminal tail (by membrane anchoring or deletion) completely eliminates the requirement of ␤ subunits for the robust expression of functional Ca v 1.2 channels, but co-expression of ␣ 2 ␦ is critical.
Recent investigation of the human cardiac short ␤ 2f and ␤ 2g subunits (26) identified the C-terminal 153-amino acid sequence as essential for the functional modulation of the channel and interaction with the ␣ 1C subunit. This region includes Ser-574 that is subject to phosphorylation by phosphatidylinositol 3-kinase as necessary and sufficient to promote Ca 2ϩ channel trafficking to the plasma membrane (44). Although the mechanism of this effect is unknown, our data put the ␣ 1C subunit N-terminal tail in play as a possible determinant of the channel trafficking and a target of the ␤ subunit modulation.
Thus a new picture of the ␤ subunit modulation of the Ca v 1.2 channel has emerged (Fig. 8), showing how ␤ acts as a "molecular wedge" that prevents the N terminus of the ␣ 1C subunit from blocking the pore. Dependence of CDI on interaction between the ␣ 1C subunit N-tail and the ␤ subunit led us to a hypothesis that the ␤ subunit and the ␣ 1C C-tail may have an integrating role for the cytoplasmic polypeptide packing underlying the pore.
Ca v 1.2 channels couple membrane depolarization to distinct neuronal functions associated with regulation of exocytosis, gene expression, synaptic plasticity, cell survival, and other processes. Among the ensemble of mutually coordinated determinants of slow inactivation identified in our study, the ␣ 1C subunit Cterminal tail and ADSI are directly involved in this coupling, whereas the crucial correlates provided by ␣ 1C N terminus and ␤ subunits may add more specialization via genetic variation and alternative splicing in neuronal cells (see Supplemental Materials and Ref. 26). Whether neuronal specialization evolves from such variations is an interesting question to be addressed.