Modulation of Ca_V2.1 Channels by the Neuronal Calcium-Binding Protein Visinin-Like Protein-2

Abstract

Ca_V2.1 channels conduct P/Q-type Ca²⁺ currents that are modulated by calmodulin (CaM) and the structurally related Ca²⁺-binding protein 1 (CaBP1). Visinin-like protein-2 (VILIP-2) is a CaM-related Ca²⁺-binding protein expressed in the neocortex and hippocampus. Coexpression of Ca_V2.1 and VILIP-2 in tsA-201 cells resulted in Ca²⁺ channel modulation distinct from CaM and CaBP1. Ca_V2.1 channels with β_2a subunits undergo Ca²⁺-dependent facilitation and inactivation attributable to association of endogenous Ca²⁺/CaM. VILIP-2 coexpression does not alter facilitation measured in paired-pulse experiments but slows the rate of inactivation to that seen without Ca²⁺/CaM binding and reduces inactivation of Ca²⁺ currents during trains of repetitive depolarizations. Ca_V2.1 channels with β_1b subunits have rapid voltage-dependent inactivation, and VILIP-2 has no effect on the rate of inactivation or facilitation of the Ca²⁺ current. In contrast, when Ba²⁺ replaces Ca²⁺ as the charge carrier, VILIP-2 slows inactivation. The effects of VILIP-2 are prevented by deletion of the CaM-binding domain (CBD) in the C terminus of Ca_V2.1 channels. However, both the CBD and an upstream IQ-like domain must be deleted to prevent VILIP-2 binding. Our results indicate that VILIP-2 binds to the CBD and IQ-like domains of Ca_V2.1 channels like CaM but slows inactivation, which enhances facilitation of Ca_V2.1 channels during extended trains of stimuli. Comparison of VILIP-2 effects with those of CaBP1 indicates striking differences in modulation of both facilitation and inactivation. Differential regulation of Ca_V2.1 channels by CaM, VILIP-2, CaBP1, and other neurospecific Ca²⁺-binding proteins is a potentially important determinant of Ca²⁺ entry in neurotransmission.

Introduction

P/Q-type Ca²⁺ currents initiate exocytosis of neurotransmitters (Takahashi and Momiyama, 1993; Regehr and Mintz, 1994). Because the efficiency of synaptic transmission is proportional to the third power of the local Ca²⁺ concentration (Dodge and Rahamimoff, 1967; Mintz et al., 1995), small changes in Ca²⁺ influx and residual free Ca²⁺ cause multiple forms of short-term synaptic plasticity that have an important influence on synaptic function (Zucker and Regehr, 2002). Ca_V2.1 channels mediate P/Q-type currents (Llinás et al., 1989; Starr et al., 1991; Wheeler et al., 1994), and these channels are localized in high density in presynaptic active zones of central neurons (Westenbroek et al., 1995; Sakurai et al., 1996; Wu et al., 1999).

Ca²⁺ binding to calmodulin (CaM) causes facilitation and enhances inactivation of Ca_V2.1 channels through binding to a site in the C-terminal domain (Lee et al., 1999, 2000; Pate et al., 2000; DeMaria et al., 2001). Ca²⁺-binding protein 1 (CaBP1), a neurospecific CaM-like Ca²⁺ binding protein, accelerates inactivation, prevents facilitation, and binds at the same site as CaM in aCa²⁺-independent manner (Lee et al., 2002). These results raise the possibility that multiple CaM-related neuronal CaBPs (nCaBPs) may interact with the C-terminal regulatory site and differentially regulate Ca_V2.1 channels.

Visinin-like proteins (VILIPs) belong to the neuronal Ca²⁺ sensor superfamily (Haeseleer et al., 2000; Burgoyne and Weiss, 2001) and are expressed in retinal (Lenz et al., 1992) and brain neurons (Saitoh et al., 1995; Bernstein et al., 1999, 2003; Hamashima et al., 2001; Spilker et al., 2002). They sensitize G-protein signaling cascades to Ca²⁺ and enhance desensitization of G-protein-coupled receptors (De Castro et al., 1995; Ames et al., 1997; Sallese et al., 2000), which can regulate Ca_V2.1 channels through Gβγ subunits (Herlitze et al., 1996; Ikeda, 1996). VILIPs reversibly translocate to membranes in response to Ca²⁺ fluctuations in neurons (Spilker et al., 2002), suggesting that they may shuttle between the plasma membrane and intracellular compartments in response to Ca²⁺ channel activity.

The kinetics and voltage dependence of activation and inactivation of Ca²⁺ channels are strongly influenced by their Ca_Vβ subunits (Hofmann et al., 1999; Lee et al., 2000; Arikkath and Campbell, 2003). The β_1b subunit causes rapid voltage-dependent inactivation, similar to β₃ and β₄ (De Waard and Campbell, 1995), whereas the β_2a subunit slows channel inactivation (Olcese et al., 1994; Chien et al., 1996). The rapid inactivation of Ca_V2.1 channels with β_1b subunits substantially reduces Ca²⁺/CaM-dependent facilitation compared with β_2a subunits (Lee et al., 1999). Both β_1b and β_2a are coexpressed with Ca_V2.1 (Stea et al., 1994; Tanaka et al., 1995; Ludwig et al., 1997; Burgess et al., 1999) and are likely to form functional channels in vivo. In this report, we describe modulation of Ca_V2.1 channels having β_1b or β_2a subunits by VILIP-2 and compare these results with modulation by CaBP1. Our results show that VILIP-2 modulates Ca_V2.1 channels in a manner distinct from CaM and CaBP1, supporting the conclusion that Ca_V2.1 channel properties are fine-tuned by interaction of multiple neuronal Ca²⁺-binding proteins at a common C-terminal regulatory site.

Materials and Methods

Cloning of VILIP-2. A full-length cDNA clone of VILIP-2 was isolated from total rat brain mRNA. VILIP-2-specific cDNA was generated with reverse primer TCTAGACTACTTCTGCATG using Superscript First-Strand Synthesis system for reverse transcription-PCR (Invitrogen, San Diego, CA). Using the same reverse primer and the forward primer CTCGAGATGGGGAAGAACAATAGC, we amplified a single band that was cloned via the TOPO TA cloning kit (Invitrogen). This was sequence verified and subcloned into pcDNA 3.1+. Construction of a carboxyl myc-tagged form of VILIP-2 (VILIP-2 myc) was accomplished by PCR using the same forward primer and the reverse primer TCTAGACTACAAGTCCTCTTCAGAAATGAGCTTTTGCTCCTTCTGCATG, which was amplified and cloned via TOPO TA cloning kit and subcloned into pcDNA 3.1+. The myc tag was detected with an anti-myc antibody from Invitrogen.

cDNA expression constructs. Expression vectors for rat Ca_V2.1 channels were prepared by subcloning wild-type and mutant subunit cDNAs into the following plasmids: α₁2.1(rbA-II) in pcDNA 3.1+, α₁2.1ΔCBD (CaM-binding domain), α₁2.1 1965 stop, α₁2.1 IQ-AA, α₁2.1 IQ-AA/ΔCBD in pMT2, β_1b and β_2a in pMT2XS, and α₂δ in pZEM228 (Ellis et al., 1988; Starr et al., 1991; Perez-Reyes et al., 1992; Stea et al., 1994), as described previously (Lee et al., 1999, 2000, 2003). CaBP1 constructs were prepared as described previously (Lee et al., 2002).

Cell culture and transfection. tsA-201 cells were maintained in DMEM/F-12 supplemented with 10% fetal bovine serum (Invitrogen) at 37°C under 10% CO₂. Cells plated in 35 mm culture dishes for electrophysiological recording were grown to 70% confluence and transfected by the CaPO₄ method with a total of 5 μg of DNA, including a 1:1 molar ratio of Ca²⁺ channel subunits (α₁2.1, β_1b or β_2a, and α₂δ), 0.3 μg of CD8 expression plasmid for identification of transfected cells, and 1 μg of VILIP-2 cDNA. Biochemical studies used 150 mm dishes with a maximum of 70 μg of DNA using the same ratios as above, but without CD8.

Electrophysiology. tsA-201 cells were incubated with CD8 antibody-coated microspheres (Dynal Biotech, Oslo, Norway) to allow visual identification of transfected cells 48-72 h after transfection. Ca²⁺ currents were recorded in the whole-cell configuration of the patch-clamp technique using an Axopatch 200A patch-clamp amplifier (Molecular Devices, Union City, CA) and were filtered at 5 kHz. Voltage protocols were applied with Pulse software (HEKA Elektronik, Lambrecht/Pfalz, Germany), and data were analyzed using IGOR Pro 4.0 software (Wave-Metrics, Lake Oswego, OR). Leak and capacitive transients were subtracted using P/-4 protocol. The extracellular recording solution contained the following (in mm): 150 Tris, 1 MgCl₂, and 10 CaCl₂ or BaCl₂, adjusted to pH 7.3 with methanesulfonic acid, except where differences are noted in Figures 1, 2, 3, 4, 5, 6, 7, 8. Intracellular solutions contained the following (in mm): 120 N-methyl-d-glucamine, 60 HEPES, 1 MgCl₂, 2 Mg-ATP, and either 0.5 or 10 EGTA, adjusted to pH 7.3 with methanesulfonic acid. Because extracellular Ba²⁺ caused positive shifts in the voltage dependence of activation of 10 mV, voltage protocols were adjusted to compensate for this difference, as noted in Figures 1, 2, 3, 4, 5, 6, 7, 8. All averaged data represent the mean ± SEM. The current traces presented in Figures 1, 2, 3, 4, 5, 6, 7, 8 represent the means of normalized current traces from the indicated number of samples for each experimental condition.

Figure 1.

Effect of coexpression of VILIP-2 on Ca_V2.1 channels. A, Voltage dependence of activation. Ca_V2.1/β_2a channels was activated by 5 ms pulses to the indicated potentials from a holding potential of -80 mV, and Ca²⁺ tail currents were recorded after repolarization to -40 mV without (gray) or with (black) coexpression of VILIP-2 (±SEM; n ≥ 10). B, I_Ca was evoked by 1 s depolarizing test pulses to +10 mV from a holding potential of -80 mV in tsA-201 cells expressing Ca_V2.1/β_2a channels without (gray) or with (black) VILIP-2. Current records were normalized to the peak inward current and averaged (±SEM; n ≥ 10). C, I_Ba was evoked and analyzed as in B. D, I_Ca was evoked and analyzed as in B for Ca_V2.1ΔCBD/β_2a. E, I_Ca conducted by Ca_V2.1/β_1b channels was evoked and analyzed as in B. F, I_Ba conducted by Ca_V2.1/β_1b channels was evoked and analyzed as in B. G, I_Ba conducted by Ca_V2.1ΔCBD/β_1b channels was evoked and analyzed as in B. H, I, Residual current amplitude at the end of the 1 s pulse (I_Res) was divided by peak current from cells transfected with Ca_V2.1 alone (gray) or with VILIP-2 (black) for the indicated experimental conditions. WT, Wild type. Error bars represent SEM.

Figure 2.

Effect of coexpression of VILIP-2 on paired-pulse facilitation of Ca_V2.1/β_2a channels. Test pulses to potentials ranging from -40 to +80 mV were applied without a prepulse (open symbols) or with a 50 ms prepulse to +10 mV and an 8 ms period at -80 mV before the test pulse (filled symbols). A, Peak inward calcium currents for cells expressing Ca_V2.1/β_2a channels without (gray circles) or with (black squares) VILIP-2. B, Same protocol as that in A, with 10 mm barium as the charge carrier. Insets, Representative calcium or barium currents at 70 mV. Calibration: horizontal, 2 ms; vertical, A, B, left insets, 1 nA; A, B, right insets, 500 pA. Error bars represent SEM.

Figure 3.

Effect of VILIP-2 on facilitation and inactivation of Ca_V2.1 channels during trains of repetitive depolarizations. Test pulses to +10 mV (0 mV for I_Ba) for 5 ms at a frequency of 100 Hz were applied to transfected tsA-201 cells expressing Ca_V2.1 channels only (gray) or Ca_V2.1 channels plus VILIP-2 (black). Peak current amplitudes were normalized to the first pulse in the series and plotted against time (±SEM; n ≥ 10; every 10th SEM is plotted). A, Representative I_Ca measured for the first pulse, the pulse at 355 ms, and the last pulse in a train of stimuli for Ca_V2.1/β_2a alone. B, Representative I_Ca measured for the first pulse, the pulse at 355 ms, and the last pulse in a train of stimuli for Ca_V2.1/β_2a plus VILIP-2. C, I_Ca, Ca_V2.1/β_2a. D, I_Ca, Ca_V2.1/β_1b. E, I_Ba,Ca_V2.1/β_2a. F, I_Ba, Ca_V2.1/β_1b. Error bars represent SEM.

Figure 4.

Effect of VILIP-2 on Ca_V2.1/β_2a channels after trains of repetitive depolarizations. Ca_V2.1/β_2a channel currents were evoked by 1 s test pulses to +10 mV before (gray) or after (black) a 600 ms train of 5 ms depolarizations to +20 mV (+10 mV for I_Ba) at 100 Hz. A 2 min interval was maintained between applications of this protocol to allow the effects of the previous stimuli to return to baseline. Current amplitudes were normalized to the peak current of the first test pulse. Mean normalized currents during test pulse 1 (gray) and test pulse 2 (black) are overlaid. A, I_Ca, Ca_V2.1/β_2a alone. B, I_Ba, Ca_V2.1/β_2a alone. C, I_Ca, Ca_V2.1/β_2a plus VILIP-2. D, I_Ba, Ca_V2.1/β_2a plus VILIP-2. E, Bar graph presenting facilitation as the pulse ratio (P2/P1) at the time of the peak current in each panel (mean ± SEM). Error bars represent SEM.

Figure 5.

Effect of VILIP-2 on Ca_V2.1/β_1b channels after trains of repetitive depolarizations. Ca_V2.1/β_1b channel currents were evoked by 1 s test pulses to +10 mV before (gray) or after (black) a 600 ms train of repetitive depolarization to +20 mV (+10 mV for I_Ba) for 5 ms at 100 Hz. A 2 min interval was maintained between applications of this protocol to allow the effects of the previous stimuli to return to baseline. Current amplitudes were normalized to the peak current of the first test pulse. Mean normalized currents during test pulse 1 (gray) and test pulse 2 (black) are overlaid. A, I_Ca, Ca_V2.1/β_1b alone. B, I_Ba, Ca_V2.1/β_1b alone. C, I_Ca, Ca_V2.1/β_1b plus VILIP-2. D, I_Ba, Ca_V2.1/β_1b plus VILIP-2. E, Bar graph presenting facilitation as the pulse ratio (P2/P1) at the time of the peak current in each panel (mean ± SEM). Error bars represent SEM.

Figure 6.

Association of VILIP-2 with Ca_V2.1 channels via the IQ-like domain and CBD. Transfected cells were cross-linked in situ with DSP, as described in Materials and Methods. A, Lysates from cells transfected with Ca_V2.1 plus VILIP-2 myc or Ca_V2.1ΔCBD plus VILIP-2 myc were subjected to immunoprecipitation by anti-Ca_V2.1 antibodies or control IgG as indicated in the presence of 10 mm EGTA. B, Lysates from cells transfected with Ca_V2.1, Ca_V2.1/1965ST, Ca_V2.1/IM-AA, and Ca_V2.1ΔCBD/IM-AA plus VILIP-2 myc were subjected to immunoprecipitation by anti-Ca_V2.1 antibodies or control IgG in the presence of 10 mm EGTA. Blots were probed with either anti-Ca_V2.1 (top) or anti-myc (bottom). WT, Wild type.

Figure 7.

Effect of CaBP1 on Ca_V2.1/β_1b channels. A, Average, smoothed I_Ca conducted by Ca_V2.1 channels evoked by 1 s depolarizing pulses from a holding potential of -80 mV to +30 mV without (gray) or with (black) coexpression of CaBP1 (±SEM; n = 5-7). B, From a holding potential of -80 mV, test pulses to +10 mV for 5 ms at a frequency of 100 Hz were applied to transfected tsA-201 cells expressing Ca_V2.1 channels alone (gray) or Ca_V2.1 channels plus CaBP1 (black). Peak current amplitudes were normalized to the first pulse in the series and plotted against time of stimulation (±SEM; n = 8-10; every 10th SEM is plotted). C, Averaged, smoothed I_Ba was measured as in A with a test pulse to +20 mV for Ca_V2.1/β_1b channels without (gray) or with (black) coexpression of CaBP1 (±SEM; n = 5-8). D, From a holding potential of -80 mV, test pulses to 0 mV for 5 ms at a frequency of 100 Hz were applied to transfected tsA-201 cells expressing Ca_V2.1/β_1b channels without (gray) or with (black) coexpression of CaBP1. Peak current amplitudes were normalized to the first pulse in the series and plotted against time (±SEM; n = 6-8; every 10th SEM is plotted). Error bars represent SEM.

Figure 8.

Comparison of the structures and modulatory effects of CaM, VILIP-2, and CaBP1. A, Schematic illustration of the structures of CaM, CaBP1, and VILIP-2. EF-hands that are active in binding Ca²⁺ are indicated in gray, and inactive EF-hands are indicated in white. Circles denote the central α helix connecting EF-hands 2 and 3. Bent lines indicate N-terminal myristoylation. The lengths of the line segments approximately correspond to the length of the amino acid sequences. B, Ca²⁺-dependent inactivation of Ca_V2.1/β_2a channels with CaM, CaBP1, or VILIP-2. Overlapped Ca²⁺ and Ba²⁺ currents are plotted. The shaded area indicates the Ca²⁺-dependent increase in inactivation caused by CaM. C, Ca²⁺-dependent facilitation and inactivation of Ca_V2.1/β_2a channels with CaM, CaBP1, or VILIP-2. The CaM and CaBP1 data in B and C were modified from Lee et al. (2000) and (2002), respectively. Error bars represent SEM.

Biochemical studies. tsA-201 cells were cultured and transfected as described above and harvested 48 h after transfection. Dithiobis (succinimidyl propionate) (DSP; 2 mm; Pierce, Rockford, IL) was dissolved in DMSO and applied for 30 min in PBS, followed by quenching in Tris. Cells were harvested in lysis buffer containing 150 mm NaCl, 10 mm EGTA, 20 mm TrisHCl, and 1% NP-40. Lysates were passed through a syringe needle (3 ml of 25G 5/8) four times, followed by centrifugation at 14,000 rpm for 10 min to remove the nuclear fraction. Anti-CNA4 (10 μg) directed against α₁2.1 (Sakurai et al., 1996) or IgG control antibody was incubated for 1 h, followed by 15 min protein A-Sepharose bead precipitation. Beads were washed three times in lysis buffer with 0.1% NP-40, heated to 95°C for 10 min in loading buffer containing 0.1 m DTT, and separated by SDS-PAGE in a 4-20% acrylamide gel. Proteins were transferred onto nitrocellulose membranes, blocked in 3% nonfat milk and 0.5% SDS for 2 h, and probed with either anti-CNA4 (1:100) or anti-myc (1:5000), followed by secondary labeling with horseradish peroxidase-linked protein A (Amersham Biosciences, Piscatway, NJ) or horseradish peroxidase-linked sheep anti-mouse (Amersham Biosciences). Bound horseradish peroxidase was detected by ECL reaction (Amersham Biosciences) followed by exposure to film.

Results

Coexpression of VILIP-2 with Ca_V2.1/β_2achannels

We first examined the effect of coexpression of VILIP-2 on Ca²⁺ currents (I_Ca) in whole-cell voltage-clamp recordings of transfected tsA-201 cells expressing Ca_V2.1 channels with β_2a subunits. Ca²⁺ currents were observed in test pulses to potentials more positive than -10 mV, approached maximum at +40 mV, and inactivated slowly during long pulses (Fig. 1A,B, gray). Coexpression of VILIP-2 did not affect the voltage dependence of activation of Ca_V2.1 channels in response to depolarizing test pulses to different membrane potentials (Fig. 1A, black). I_Ca inactivates ∼53% during a 1 s test pulse (Fig. 1B, gray) compared with 31% for Ba²⁺ current (I_Ba) (Fig. 1C, gray). This Ca²⁺-dependent component of inactivation is caused by endogenous CaM, which binds entering Ca²⁺ and interacts with the CBD in the C terminus to inactivate Ca_V2.1 channels (Lee et al., 1999, 2003). When VILIP-2 is coexpressed with Ca_V2.1/β_2a channels, the rate of inactivation of I_Ca is slowed considerably (Fig. 1B, black). In contrast to I_Ca, VILIP-2 has no effect on the rate of inactivation of I_Ba conducted by Ca_V2.1/β_2a channels (Fig. 1C). The block of Ca²⁺-dependent inactivation by VILIP-2 suggests that it may prevent binding of endogenous Ca²⁺/CaM to the CBD of Ca_V2.1 channels.

A deletion mutant of Ca_V2.1 lacking the CBD (Ca_V2.1ΔCBD) conducts Ca²⁺ currents that do not undergo rapid CaM-dependent inactivation (Fig. 1D, gray), confirming that CaM association with the CBD is required for Ca²⁺-dependent inactivation. Expression of VILIP-2 with Ca_V2.1ΔCBD has no effect on its inactivation (Fig. 1D, black), supporting the conclusion that VILIP-2 acts primarily to negate the effect of endogenous Ca²⁺/CaM on Ca_V2.1/β_2a channels. Because VILIP-2 has no effect on inactivation of the Ca_V2.1ΔCBD channel, either VILIP-2 does not bind to Ca_V2.1ΔCBD or it binds but has no effect on inactivation.

Coexpression of VILIP-2 with Ca_V2.1/β_1b channels

When Ca_V2.1 is coexpressed with the β_1b subunit, voltage-dependent inactivation of I_Ca is faster than for Ca_V2.1/β_2a (Fig. 1, compare E, gray with B, gray), and it is unaffected by coexpression of VILIP-2 (Fig. 1E, black). Inactivation of I_Ba conducted by Ca_V2.1/β_1b is also more rapid than inactivation of I_Ba of Ca_V2.1/β_2a channels (Fig. 1, compare F, gray with C, gray), and it is substantially slowed by coexpression of VILIP-2 (Fig. 1F, black), without effect on the voltage dependence of activation (data not shown). In contrast, the Ca_V2.1ΔCBD/β_1b mutant channel is unaffected by VILIP-2 (Fig. 1G), indicating that VILIP-2 acts at the same site as CaM to control the rate of inactivation. Because VILIP-2 alters the rate of inactivation of I_Ba, it must associate with Ca_V2.1 channels at the resting level of intracellular Ca²⁺ and slow voltage-dependent inactivation. However, although VILIP-2 is bound to the Ca_V2.1/β_1b channel, it is unable to slow the inactivation of I_Ca appreciably, presumably because of the strong driving force for inactivation with both β_1b and Ca²⁺ bound to the channel complex. Overall, VILIP-2 appears to act by negating rapid Ca²⁺/CaM-dependent inactivation and slowing voltage-dependent inactivation in a β subunit-dependent manner.

Paired-pulse facilitation of Ca_V2.1/β_2a channels coexpressed with VILIP-2

I_Ca conducted by Ca_V2.1/β_2a channels is significantly facilitated after a previous depolarizing pulse, resulting from Ca²⁺/CaM interaction with the IQ-like domain and CBD (Lee et al., 2000, 2003; DeMaria et al., 2001). With 10 mm EGTA in the recording pipette, Ca²⁺ facilitation can be recorded in isolation without interference from Ca²⁺-dependent inactivation (Lee et al., 2000). Applying a similar protocol, we found that paired-pulse facilitation of I_Ca was unchanged by coexpression of VILIP-2 (Fig. 2A). As expected, no facilitation of I_Ba was observed without or with VILIP-2 (Fig. 2B). Evidently, although VILIP-2 modulates inactivation differently from CaM, it can mediate Ca²⁺-dependent paired-pulse facilitation similarly.

Ca²⁺-dependent facilitation and inactivation during trains of depolarizations

Previous work demonstrated that activity-dependent increases in Ca²⁺ entry cause facilitation attributable to a local increase in Ca²⁺, followed by inactivation attributable to a more global increase in Ca²⁺ during extended trains of depolarizations (Lee et al., 2000). These trains of brief depolarizations more closely resemble electrical activity in nerve terminals than single long depolarizing pulses. This dual regulation is caused by stepwise Ca²⁺ binding to C- and N-terminal lobes of CaM and sequential binding of Ca²⁺/CaM to the IQ-like domain and CBD of Ca_V2.1 channels (Lee et al., 1999, 2000, 2003; DeMaria et al., 2001). To investigate the effects of VILIP-2 coexpression on Ca_V2.1 channels during trains of brief stimuli, we depolarized tsA-201 cells expressing Ca_V2.1/β_2a channels repetitively to +10 mV for 5 ms at a frequency of 100 Hz and recorded I_Ca during each stimulation. Initially, we observed that VILIP-2 prevented both facilitation and subsequent inactivation (data not shown). However, on more careful analysis, we realized that the residual increase in Ca²⁺ concentration from previous depolarization protocols was occluding facilitation. By recording from cells that either had not experienced stimulated Ca²⁺ influx or were allowed to recover from previous depolarizations for at least 2 min, we were able to accurately record both facilitation and inactivation by CaM (Fig. 3A,C, gray) and facilitation by VILIP-2 (Fig. 3B,C, black). VILIP-2 coexpression increased the amount of facilitation and decreased the rate of pulse-wise inactivation of I_Ca compared with control cells not expressing VILIP-2 (Fig. 3C). As expected from the lack of effect of VILIP-2 on paired-pulse facilitation, it had insignificant effects during the first few pulses but reduced the pulse-wise inactivation of I_Ca thereafter. In contrast, when Ba²⁺ replaced Ca²⁺ as the charge carrier, insignificant facilitation was observed with endogenous CaM in control cells or after coexpression of VILIP-2 (Fig. 3E). These results are in agreement with our observation (Fig. 1) that VILIP-2 slows the rate of Ca²⁺-dependent inactivation of Ca_V2.1/β_2a channels during single long depolarizations and show further that VILIP-2 enhances facilitation during extended trains of stimuli in a Ca²⁺-dependent manner. Because VILIP-2 has no effect on facilitation by a single pulse, it is likely that the increase in facilitation during trains of stimuli reflects slowing of cumulative Ca²⁺-dependent inactivation.

When the β_1b subunit was coexpressed with Ca_V2.1, and Ca²⁺ was the charge carrier, trains of stimuli elicited brief facilitation followed by rapid, pulse-wise inactivation (Fig. 3D, gray). VILIP-2 had no effect on facilitation or inactivation of I_Ca during repetitive depolarizations (Fig. 3D, black). However, when Ba²⁺ replaced Ca²⁺ as the charge carrier, minimal facilitation of the current was observed, and inactivation was reduced. Coexpression of VILIP-2 reduced the rate of the remaining pulse-wise inactivation (Fig. 3F, black) compared with control cells (Fig. 3F, gray), similar to the reduced inactivation of I_Ba we observed during single long depolarizations (Fig. 1F). The effects of VILIP-2 on facilitation and inactivation of both Ca_V2.1/β_2a channels and Ca_V2.1/β_1b channels are consistent with a model in which VILIP-2 associates with the CBD at resting levels of Ca²⁺, blocks the effects of CaM, slows inactivation, and thereby enhances facilitation in a β subunit-dependent manner.

Modulation of Ca_V2.1 channels by VILIP-2 after trains of depolarizations

To further characterize the effects of trains of depolarizations on the modulation of Ca_V2.1 channels by VILIP-2, we used a double-pulse protocol in which a 1 s depolarizing test pulse was followed by a 600 ms train of depolarizations to +20 mV for 5 ms at 100 Hz and finally by a second identical test pulse. This protocol is expected to cause accumulation of Ca²⁺ in the vicinity of Ca²⁺ channels during the train of depolarizations, which will modulate the activation and inactivation of the Ca_V2.1 channels in the second test pulse compared with the first. For Ca_V2.1 channels containing the β_2a subunit, the rate of activation and the peak amplitude of I_Ca during 1 s test pulses to +10 mV are significantly increased after the pulse train (Fig. 4A, black, E) compared with before the train of stimuli (Fig. 4A, gray, E). These effects are not observed for I_Ba, indicating that they are caused by Ca²⁺ entering the cells during the train of stimuli (Fig. 4B,E). For cells coexpressing VILIP-2, the peak amplitude of I_Ca is further enhanced after the train of depolarizing stimuli (Fig. 4C), and this effect is also lost when Ba²⁺ is the current carrier (Fig. 4D,E). Thus, VILIP-2 increases the Ca²⁺-dependent facilitation of I_Ca conducted by Ca_V2.1/β_2a channels after a train of depolarizations, as observed in the experiments shown in Figure 2 for I_Ca recorded during the brief depolarizations within the train. As for depolarizations within the train, it is likely that the enhanced facilitation caused by VILIP-2 reflects slowing of cumulative inactivation of I_Ca during the pulse train.

We next examined the effect of a train of repetitive depolarizing stimuli on Ca_V2.1 channels containing the β_1b subunit (Fig. 5). In contrast to Ca_V2.1/β_2a channels, previous stimulation of Ca_V2.1/β_1b channels with a train of depolarizing pulses reduces peak I_Ca (Fig. 5A, black) compared with currents recorded during an identical test pulse immediately before the train (Fig. 5A, gray), yielding a pulse ratio [pulse 2 (P2)/P1] significantly <1.0 (Fig. 5E). Similar results are obtained for I_Ba (Fig. 5B), indicating that these effects of the train of prepulses are independent of Ca²⁺ and therefore reflect cumulative voltage-dependent inactivation. Coexpression of VILIP-2 has little effect on the response of I_Ba conducted by Ca_V2.1/β_1b channels to a train of depolarizations (Fig. 5C,D). Evidently, the effects of VILIP-2 on facilitation and inactivation after trains of prepulses are specific for Ca_V2.1 channels containing β_2a subunits, which have substantial Ca²⁺-dependent facilitation and inactivation.

Association of VILIP-2 with Ca_V2.1 channels

We measured the association of VILIP-2 and Ca_V2.1 channels in coimmunoprecipitation studies by coexpressing a myc-tagged form of VILIP-2 with Ca_V2.1 and the β_1b subunit. Our initial experiments failed to demonstrate interaction of VILIP-2 with Ca_V2.1 channels after detergent solubilization and immunoprecipitation, indicating that the complex formed is reversible and dissociates during isolation. In contrast to these results, in situ protein cross-linking with DSP applied to intact cells (see Materials and Methods), followed by detergent extraction and immunoprecipitation, resulted in detection of VILIP-2 binding to Ca_V2.1 channels at resting levels of Ca²⁺ (Fig. 6A, WT). Coimmunoprecipitation of VILIP-2 was specific, because VILIP-2 was not coimmunoprecipitated with control IgG in cells transfected with Ca_V2.1 and VILIP-2.

CaM binds to two closely spaced motifs in the C terminus of Ca_V2.1 channels, the CBD and the IQ-like domain (Lee et al., 1999, 2000, 2003; DeMaria et al., 2001; Erickson et al., 2003). Deletion of the CBD or deletion of the carboxyl-half of the C terminus containing the CBD at amino acid 1965 caused a variable decrease in VILIP-2 binding (Fig. 6A, ΔCBD, B, 1965ST), suggesting that VILIP-2 bound to a second site or subsite in addition to the CBD. Mutation of the first two amino acid residues of the IQ-like domain had little effect on binding of VILIP-2 (Fig. 6B, IMAA). In contrast, combined mutation of the CBD and the IQ-like domain consistently prevented VILIP-2 binding (Fig. 6B, ΔCBD-IMAA). These results indicate that VILIP-2, like CaM (Lee et al., 1999, 2000, 2003; DeMaria et al., 2001; Erickson et al., 2003), binds to both the IQ-like domain and the CBD of Ca_V2.1 channels.

Modulation of Ca_V2.1/β_1b channels by CaBP1

In previous studies, we examined the effects of CaBP1 on Ca_V2.1/β_2a channels and found that it enhanced voltage-dependent inactivation and prevented facilitation (Lee et al., 2002), in marked contrast to the effects of VILIP-2 reported here. To complete the comparison of modulation of Ca_V2.1 channels by CaBP1 and VILIP-2, we also studied the effects of CaBP1 on Ca_V2.1/β_1b channels. Under these experimental conditions, coexpression of CaBP1 had no effect on the rate of activation or inactivation of I_Ca conducted by Ca_V2.1/β_1b channels during 1 s depolarizations (Fig. 7A) and also had no effect on their voltage dependence of activation (data not shown). Although there was a trend toward increased decay of I_Ca during trains of repetitive pulses (Fig. 7B), this difference did not reach statistical significance. In contrast to these results for I_Ca, replacement of Ca²⁺ with Ba²⁺ eliminates the small effects of Ca²⁺/CaM-dependent inactivation of Ca_V2.1/β_1b channels and reveals Ca²⁺-independent effects of CaBP1. Coexpression of CaBP1 enhanced inactivation of I_Ba during single 1 s test pulses (Fig. 7C) and during extended trains of depolarizations (Fig. 7D). Thus, coexpression of CaBP1 has little effect on I_Ca conducted by Ca_V2.1/β_1b channels but accelerates inactivation of I_Ba, opposite to the effect of VILIP-2.

Discussion

Modulation of Ca_V2.1 channels by VILIP-2

Our results show that VILIP-2 modulates Ca_v2.1 channels in a biphasic manner that depends on the associated β subunit and on Ca²⁺ entry. Ca_V2.1 channels with associated β_2a subunits have slow voltage-dependent inactivation (Olcese et al., 1994; Chien et al., 1996). VILIP-2 decreases Ca²⁺/CaM-dependent inactivation of the Ca²⁺ current through Ca_V2.1/β_2a channels during 1 s depolarizations. Replacing Ca²⁺ with Ba²⁺ as the charge carrier eliminates Ca²⁺/CaM-dependent inactivation and eliminates the effect of VILIP-2. Similarly, Ca_V2.1ΔCBD/β_2a channels, which have impaired CaM binding, do not undergo rapid Ca²⁺/CaM-dependent inactivation, and inactivation of I_Ca is not further slowed by VILIP-2 coexpression. These results are consistent with the conclusion that VILIP-2 occupies the same binding site as endogenous CaM and thereby prevents rapid Ca²⁺/CaM-dependent inactivation. When bound in place of CaM, VILIP-2 reduces inactivation of Ca_V2.1/β_2A channels.

Coexpression of VILIP-2 does not alter Ca²⁺-dependent facilitation of Ca_V2.1/β_2a channels after a single pulse in paired-pulse facilitation experiments. Evidently, VILIP-2 can bind Ca²⁺ and induce facilitation like CaM. However, the effect of VILIP-2 to slow inactivation allows enhanced facilitation during and after extended trains of depolarizations by reducing cumulative inactivation during the train.

Ca_V2.1/β_1b channels exhibit rapid voltage-dependent inactivation (De Waard and Campbell, 1995; Hofmann et al., 1999; Lee et al., 2000; Arikkath and Campbell, 2003) and comparatively little acceleration of inactivation by Ca²⁺/CaM (Lee et al., 1999, 2000). In contrast to Ca_V2.1/β_2a channels, coexpression of VILIP-2 has no effect on the rapid inactivation of I_Ca but slows voltage-dependent inactivation of I_Ba to give a rate comparable with inactivation of I_Ba through Ca_V2.1/β_2a channels. Deletion of the CBD blocks the effect of VILIP-2 to slow the rate of inactivation of I_Ba, identifying this site as being necessary for VILIP-2 modulation of Ca_V2.1/β_1b. VILIP-2 coexpression also reduces inactivation of I_Ba during trains of depolarizations. These results are consistent with the conclusion that VILIP-2 occupies the same binding site as CaM. When bound in place of CaM, VILIP-2 slows inactivation of I_Ba but does not enhance facilitation of Ca_V2.1/β_1b channels during extended trains of depolarizations.

Interaction of VILIP-2 with Ca_V2.1 channels

CaM binds to at least two interacting subsites on the carboxyl tail of Ca_V2.1, the IQ-like domain and the CBD (Lee et al., 1999, 2003; DeMaria et al., 2001). Sequential Ca²⁺-dependent interactions with these subsites are required for Ca²⁺/CaM-dependent facilitation and inactivation (Lee et al., 2003). In contrast, we found that VILIP-2 binds to Ca_V2.1 channels at the resting level of Ca²⁺ by interaction with the IQ-like domain and the CBD. VILIP-2 binding was weaker than CaM and was not stable through solubilization and immunoprecipitation; however, covalent cross-linking in situ with DSP stabilized this in vivo interaction so that it could be measured by solubilization and coimmunoprecipitation. Because mutation of both the IQ-like motif and the CBD were required to prevent binding of VILIP-2, it likely displaces CaM interactions from both of those subsites on Ca_V2.1 channels. Evidently, binding of VILIP-2 at resting Ca²⁺ levels and displacement of CaM from the IQ-like domain and the CBD combine to produce the effects that we have recorded in our electrophysiological studies.

Colocalization of VILIP-2 and Ca_V2.1 channels in the brain

VILIP-2 is expressed in the neocortex, caudate-putamen, and hippocampus, including both pyramidal neurons and dentate granule cells, and in the Purkinje neurons of the cerebellum (Paterlini et al., 2000). Ca_V2.1, β_2a, and β_1b are also expressed in these regions (Stea et al., 1994; Tanaka et al., 1995; Westenbroek et al., 1995; Sakurai et al., 1996). The Ca_V2.1/β_2a currents reported here are similar to slowly inactivating P/Q-type Ca²⁺ currents recorded in hippocampal CA1 neurons (Hillyard et al., 1992; Mintz, 1994), cerebellar granule neurons (Randall and Tsien, 1995), and Purkinje neurons (Mintz et al., 1992a,b). Thus, VILIP-2 is present in neurons in which Ca_V2.1/β_2a is expressed and is therefore able to modulate these channels in vivo. Immunocytochemical studies of VILIP-2 localization show broad distribution in the cell bodies and dendrites of hippocampal pyramidal neurons, dentate granule neurons, cortical neurons, and cerebellar Purkinje neurons (Saitoh et al., 1994, 1995). In addition, punctate labeling in the neuropil of the hippocampus and cerebral cortex suggests localization of VILIP-2 in nerve terminals as well (Saitoh et al., 1994, 1995), together with Ca_V2.1 channels (Westenbroek et al., 1995). These results indicate that VILIP-2 is colocalized with Ca_V2.1 channels in neuronal cell bodies and dendrites and probably in nerve terminals. Our previous results show that CaBP1 and Ca_V2.1 channels are colocalized in cell bodies, dendrites, and nerve terminals in the hippocampus and cerebellum (Lee et al., 2002), thereby placing both nCaBPs and Ca_V2.1 channels in the same subcellular locations. Ca_V2.1 channels in these neuronal subcellular compartments would be modulated in a reciprocal calcium-dependent manner by CaM, CaBP1, and VILIP-2 acting at the same site.

Differential modulation of Ca_V2.1 channels by neuronal Ca²⁺-binding proteins

This study provides new evidence that different Ca²⁺-binding proteins fine-tune the functional properties of Ca_V2.1 channels. Figure 8 illustrates the structural features of CaM, CaBP1, and VILIP-2 and shows their effects on Ca_V2.1 channels during single depolarizations and repetitive stimulation as an example of their differential actions. CaM has four EF-hands, which are all functional in binding Ca²⁺.An α helix connects EF-hands 2 and 3 and divides the molecule into two halves, which function differentially in facilitation and inactivation of Ca_V2.1 channels (DeMaria et al., 2001; Lee et al., 2003). High-affinity binding of local Ca²⁺ to EF-hands 3 and 4 initiates facilitation by enhancing interaction with the IQ-like domain, whereas lower-affinity binding of globally increased Ca²⁺ to EF-hands 1 and 2 initiates Ca²⁺-dependent inactivation by interaction with the CBD (Lee et al., 2000, 2003; DeMaria et al., 2001).

CaBP1 and VILIP-2 have distinct molecular features and functional effects (Fig. 8). EF-hand 2 of CaBP1 has amino acid sequence changes that prevent high-affinity binding of Ca²⁺, the long N-terminal has a myristoyl lipid anchor, and the central α helix is lengthened by an additional turn compared with CaM (Fig. 8A). In contrast, EF-hand 1 of VILIP-2 does not bind Ca²⁺, the N-terminal is shorter but also has a myristoyl lipid anchor, and the central α helix is similar in length to CaM (Fig. 8A). These molecular differences cause different functional effects on Ca_V2.1 channels (Fig. 8B,C). CaM mediates Ca²⁺-dependent inactivation, CaBP1 causes rapid inactivation regardless of whether Ca²⁺ is the permeant ion, and VILIP-2 causes slow inactivation regardless of whether Ca²⁺ is the permeant ion (Fig. 8B). The differences in the effects of these neuronal Ca²⁺-binding proteins on inactivation during single long depolarizations (Fig. 8B) is mirrored in their effects on facilitation and inactivation during trains of stimuli (Fig. 8C). CaM causes facilitation followed by inactivation, VILIP-2 causes only enhanced facilitation, and CaBP1 causes only enhanced inactivation. These differential effects on presynaptic Ca²⁺ channels would alter the encoding properties of the presynaptic terminal by changing its response to repetitive stimuli from facilitation followed by inactivation to either facilitation only or inactivation only (Fig. 8). Thus, in nerve terminals where these Ca²⁺-binding proteins are colocalized with Ca_V2.1 channels, the input-output relationships of the corresponding synapses would be influenced in an important way by regulation of the presynaptic Ca²⁺ channels by differentially expressed neurospecific Ca²⁺-binding proteins.