Compensatory Regulation of Ca_v2.1 Ca²⁺ Channels in Cerebellar Purkinje Neurons Lacking Parvalbumin and Calbindin D-28k

Abstract

Ca_v2.1 channels regulate Ca²⁺ signaling and excitability of cerebellar Purkinje neurons. These channels undergo a dual feedback regulation by incoming Ca²⁺ ions, Ca²⁺-dependent facilitation and inactivation. Endogenous Ca²⁺-buffering proteins, such as parvalbumin (PV) and calbindin D-28k (CB), are highly expressed in Purkinje neurons and therefore may influence Ca_v2.1 regulation by Ca²⁺. To test this, we compared Ca_v2.1 properties in dissociated Purkinje neurons from wild-type (WT) mice and those lacking both PV and CB (PV/CB^−/−). Unexpectedly, P-type currents in WT and PV/CB^−/− neurons differed in a way that was inconsistent with a role of PV and CB in acute modulation of Ca²⁺ feedback to Ca_v2.1. Ca_v2.1 currents in PV/CB^−/− neurons exhibited increased voltage-dependent inactivation, which could be traced to decreased expression of the auxiliary Ca_vβ_2a subunit compared with WT neurons. Although Ca_v2.1 channels are required for normal pacemaking of Purkinje neurons, spontaneous action potentials were not different in WT and PV/CB^−/− neurons. Increased inactivation due to molecular switching of Ca_v2.1 β-subunits may preserve normal activity-dependent Ca²⁺ signals in the absence of Ca²⁺-buffering proteins in PV/CB^−/− Purkinje neurons.

INTRODUCTION

In cerebellar Purkinje neurons, voltage-gated Ca_v2.1 channels conduct P-type Ca²⁺ currents that trigger dendritic Ca²⁺ spikes (Llinás and Sugimori 1980a,b; Tank et al. 1988) and regulate repetitive firing (Walter et al. 2006; Womack and Khodakhah 2004). Ca_v2.1 is concentrated in the soma and dendrites of Purkinje neurons (Westenbroek et al. 1995) and mediates >90% of the whole cell voltage-gated Ca²⁺ current (Jun et al. 1999; McDonough et al. 1997; Mintz et al. 1992). Mouse mutations that inhibit P-type current density inhibit the frequency and precision of spontaneous firing (Donato et al. 2006; Walter et al. 2006) and increase intrinsic excitability in Purkinje neurons (Ovsepian and Friel 2008). Given the importance of Purkinje neurons in the control of motor function (Ito 1984), these cellular defects invariably cause ataxia (Sidman 1965; Snell 1955).

Like other Ca_v channels, Ca_v2.1 is directly modulated by Ca²⁺ ions that permeate the channel. During repetitive depolarizations, Ca_v2.1 channels undergo Ca²⁺-dependent facilitation followed by inactivation (Chaudhuri et al. 2005; DeMaria et al. 2001; Lee et al. 2000). These effects rely on Ca²⁺ binding to calmodulin, which interacts directly with the pore-forming Ca_v2.1 subunit (α₁2.1), and are absent for Ca_v2.1 Ba²⁺ currents (DeMaria et al. 2001; Lee et al. 1999). Ca²⁺-dependent inactivation, but not facilitation, is blunted by intracellular dialysis with Ca²⁺ chelators such as ethylene glycol tetraacetic acid (EGTA) (Lee et al. 2000). These findings suggest that Ca²⁺-dependent facilitation depends on rapid, local Ca²⁺ signals through individual channels, whereas Ca²⁺-dependent inactivation relies on slower, global Ca²⁺ elevations supported by multiple neighboring channels (DeMaria et al. 2001; Liang et al. 2003; Soong et al. 2002).

The sensitivity of Ca²⁺-dependent inactivation to Ca²⁺ chelators has important implications for Ca_v2.1 channels in Purkinje neurons. These neurons have high endogenous Ca²⁺-buffering capacity due in part to the Ca²⁺-binding proteins parvalbumin (PV) and calbindin D-28k (CB) (Celio 1990; Fierro et al. 1998). By chelating free Ca²⁺, the two proteins and, more importantly, CB directly modulate the kinetics of synaptically evoked Ca²⁺ transients in Purkinje cell dendrites (Schmidt et al. 2003, 2007). PV and CB can also indirectly shape Ca²⁺ signals by controlling Ca²⁺ ions that are available for feedback regulation of Ca_v2.1. Like EGTA, coexpression of PV or CB with Ca_v2.1 in HEK293T cells can suppress Ca²⁺-dependent inactivation without affecting facilitation (Kreiner and Lee 2006). Therefore PV and CB may alter Ca²⁺ feedback regulation of Ca_v2.1 in Purkinje neurons.

To test this, we compared P-type currents in dissociated Purkinje neurons from wild-type (WT) mice and those lacking expression of PV and CB (PV/CB^−/−). Voltage-dependent inactivation but not Ca²⁺-dependent inactivation was greater in PV/CB^−/− than in WT neurons, which could be explained by down-regulation of the auxiliary Ca_vβ_2a subunit. Our findings suggest a new role for Ca²⁺-binding proteins in maintaining Ca_v2.1 function and also suggest a compensatory mechanism by which Ca²⁺ homeostasis may be achieved in the absence PV and CB.

METHODS

Purkinje cell dissociation

Animal procedures complied with National Institutes of Health guidelines and were conducted under a protocol approved by Emory Institutional Animal Care and Use Committee. PV and CB double-knockout mice (PV/CB^−/−) were characterized previously (Vecellio et al. 2000) and maintained on the Sv129×C57/BL6 strain, which served as the WT group. Postnatal day 14 (P14) to P21 mice were anesthetized with isoflurane and decapitated. Sagittal cerebellar slices (400 μm) were cut on a vibratome and held in Tyrode solution (in mM: 150 NaCl, 4 KCl, 2 CaCl₂, 2 MgCl₂, 10 HEPES, 10 glucose, adjusted to pH 7.4 with NaOH) at 34°C for 30 min before allowing them to cool to room temperature. Immediately prior to recording, slices were incubated for 10 min in papain (1 mg/ml; Worthington, Lakewood, NJ) dissolved in dissociation solution (in mM: 82 Na₂SO₄, 30 K₂SO₄, 5 MgCl₂, 10 HEPES, and 10 glucose, adjusted to pH 7.4 with NaOH). Slices were then washed in Tyrode solution, placed in a fresh tube containing 1 ml Tyrode solution, and dissociated by gentle trituration through a series of fire-polished pipettes. Supernatant containing dissociated cells was placed on poly-l-lysine–coated coverslips for electrophysiological recording. Dissociated Purkinje neurons were readily distinguished from surrounding cells by their large, pear-shaped soma and dendritic stump.

HEK293T cell culture and transfection

HEK293T cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum at 37°C under 5% CO₂. Cells plated in 35-mm tissue-culture dishes were grown to 65–80% confluency and transfected with GenePORTER transfection reagent (Gene Therapy Systems, San Diego, CA), according to the manufacturer's protocol, with a 1:1 molar ratio of cDNAs for Ca²⁺ channel subunits: rat α₁2.1 (rbA isoform; Stea et al. 1994), β_2a (Perez-Reyes et al. 1992) or β₄ (Castellano et al. 1993), α₂δ (Starr et al. 1991; total 5 μg), and 0.7 μg of a CD8 expression plasmid. At least 48 h after transfection, cells were incubated with CD8 antibody-coated microspheres (Dynal, Oslo, Norway) for identification of transfected cells for electrophysiological recording.

Electrophysiological recordings

Ca²⁺ or Ba²⁺ currents were recorded in transfected HEK293T cells and Purkinje neurons in whole cell patch-clamp configuration at room temperature. Extracellular recording solutions for HEK293T cells contained (in mM): 150 Tris, 1 MgCl₂, and 10 CaCl₂ or 10 BaCl₂. Intracellular recording solutions contained (in mM): 130 N-methyl-d-glucamine, 60 HEPES, 1 MgCl₂, 2 Mg-ATP, and 0.5 EGTA. The pH of extracellular and intracellular recording solutions was adjusted to 7.3 with methanesulfonic acid. For voltage-clamp recordings of Purkinje neurons, extracellular solutions were optimized for stable whole cell recordings of Ca_v2.1 currents and contained (in mM): 155 tetraethylammonium chloride (TEA), 10 HEPES, 1 EGTA, and 10 CaCl₂ or 2 or 10 BaCl₂, adjusted to pH 7.4 with TEA-OH. Prior to recording, tetrodotoxin (TTX, 1 μM) and nimodipine (5 μM) were included in the extracellular recording solution to block Na⁺ and L-type Ca²⁺ currents, respectively. Intracellular solutions contained (in mM): 140 Cs methanesulfonate, 4 MgCl₂, 0.5 EGTA, 9 HEPES, 14 creatine phosphate (Tris salt), 4 Mg adenosine 5′-triphosphate (ATP), and 0.3 Tris-GTP, adjusted to pH 7.4 with CsOH. For current-clamp recordings, Tyrode solution was used for the extracellular solution and the intracellular recording solution contained (in mM): 140 K-methyl sulfate, 10 KCl, 5 NaCl, 0.01 EGTA, 10 HEPES, and 2 Mg-ATP, adjusted to pH 7.4 with KOH. Electrode resistances in the recording solutions were typically 1–2 MΩ for Purkinje neurons and 2–4 MΩ for HEK293T cells. Reagents used for electrophysiological recordings were obtained from Sigma–Aldrich (St. Louis, MO).

Currents were recorded with an EPC-9 patch-clamp amplifier driven by PULSE software (HEKA Electronics, Lambrecht/Pfalz, Germany). Leak and capacitive transients were subtracted using a P4 protocol. Action potential (AP) waveforms were constructed by averaging traces corresponding to spontaneous APs recorded in current clamp from dissociated WT Purkinje neurons. The AP waveform was applied from a holding potential of −60 mV and had a half-width of 0.75 ms and peak amplitude of +25 mV for Purkinje neuron recordings. For transfected HEK293T cell recordings, the waveform was scaled to peak at +55 mV. Data were analyzed using routines written in IGOR Pro software (WaveMetrics, Lake Oswego, OR). Averaged data represent the means ± SE. Statistical differences between groups were determined by Student's t-test. Current–voltage (I–V) curves were fit with the function: g(V − E)/{1 + exp[(V − V_1/2)/k] + b}, where g is the maximum conductance, V is the test potential, E is the apparent reversal potential, V_1/2 is the potential of half-activation, k is the slope factor, and b is the baseline. Significant differences between I–V curve values were determined with two-way repeated-measures ANOVA.

PCR analysis

For endpoint polymerase chain reaction (PCR), total RNA was isolated from cerebellar tissue using Trizol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's protocol. Reverse transcription was performed on 2 μg RNA using random hexamers and Superscript II reverse transcriptase (Invitrogen) at 42°C for 50 min. For isolated Purkinje neurons, the entire neuron was harvested by suction into an electrode containing intracellular recording solution. The tip of the pipette was then pressurized and broken to expel the cell into a tube containing reverse transcription reagents. Following overnight incubation at 40°C, the samples were used immediately or stored at −80°C. PCR reactions were performed with GoTaq Green Master Mix (Promega, Madison, WI) or Taq DNA polymerase (New England BioLabs, Ipswich, MA) and a Mastercycler (Eppendorf, Westbury, NY). PCR products were visualized by electrophoresis on 1–2% agarose gels prestained with ethidium bromide.

For quantitative PCR, pools of 5–10 Purkinje neurons were isolated for reverse transcription as described earlier. Oligonucleotide primers were designed to amplify approximately 100 basepair mouse sequences for Ca_vβ_2a and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Amplification of a single band was confirmed initially by endpoint PCR and gel electrophoresis. Quantitative PCR was performed on reverse transcription (RT) reactions (2 μl) according to manufacturer's protocols with the DyNAmo HS SYBR Green qPCR kit (Finnzymes, Woburn, MA) and iCycler (Bio-Rad, Hercules, CA). Samples were run in duplicate for 40 cycles [95°C for 15 min (hot start), 94°C for 20 s, 60°C for 30 s, 72°C for 30 s]. Experiments were repeated at least three times using RNA collected from Purkinje neurons from different mice. Melt curve analysis revealed a single peak in SYBR green fluorescence, confirming amplification of a single product for each primer set. Efficiency of amplification was roughly 100% for each primer set and measured as [(10^−1/k − 1) × 100], where k is the slope of the relationship between cycle threshold [C(t)] and dilution of the RT reaction. Relative change in WT compared with PV/CB^−/− neurons was determined by the comparative C(t)[2^−ΔΔC(t)] method (Livak and Schmittgen 2001), in which GAPDH was used as the internal control gene. For WT and PV/CB^−/− neurons, ΔC(t) was determined as the difference in C(t) for the target (Ca_vβ_2a) and C(t) for GAPDH. Percentage WT expression was then determined as: {[ΔC(t) for PV/CB^−/−] − [ΔC(t) for WT]}× 100%. Primer sequences are listed in Supplemental Table S1.¹

RESULTS

Characterization of Ca²⁺-dependent inactivation and facilitation of P-type currents in Purkinje neurons

To allow for comparisons of neuronal P-type currents with Ca_v2.1 currents in transfected cells, recordings were made in whole cell patch configuration. Although this approach could cause dialysis of intracellular components, such as PV and CB, we did not observe significant differences in P-type current properties ≤20 min after membrane patch rupture. However, voltage protocols were initiated at the same time and run in the same order to minimize variability between cells. Although the absence of exogenous Ca²⁺ chelators in internal recording solutions might allow for maximal resolution of the effects of PV and CB, this approach did not allow stable whole cell recordings. However, the amount of EGTA in the intracellular recording solution was limited to 0.5 mM, which is permissive for Ca²⁺-dependent inactivation of Ca_v2.1 in transfected cells (DeMaria et al. 2001; Lee et al. 1999, 2000).

For our studies, we used Purkinje neurons from mice that were 14–21 days old. This age range was chosen as a compromise between the ability to obtain stable Ca²⁺ current recordings in young neurons and previous findings that mature expression levels of PV and CB are reached near P20 (Schwaller et al. 2002). Since Ca²⁺-dependent modulation of Ca_v2.1 has not been reported in mouse Purkinje neurons at this developmental stage, we first undertook a basic characterization of this process. To isolate P-type currents, whole cell patch-clamp recordings were conducted with extracellular Ca²⁺ (10 mM) and blockers of voltage-gated Na⁺ and K⁺ channels. Depolarizing steps were made from a holding voltage of −60 mV to minimize contribution of low-voltage–activated T-type currents in these cells (Raman and Bean 1999). Under these conditions, we observed large inward currents that did not inactivate significantly within the 50-ms test pulse. The currents were blocked about 80–90% on extracellular perfusion of the selective Ca_v2.1 blocker, ω-agatoxin type IVA (500 nM, Fig. 1 A). These properties were consistent with the P-type current in rat and mouse Purkinje neurons (Mintz et al. 1992; Raman and Bean 1999; Richards et al. 2007; Usowicz et al. 1992), which is mediated by Ca_v2.1 (Jun et al. 1999).

Fig. 1.

Ca²⁺-dependent inactivation of P-type currents in mouse cerebellar Purkinje neurons. A: current–voltage (I–V) relationships for a Ca²⁺ current (I_Ca) evoked by 50-ms steps from −60 mV to various voltages. Shown are representative current traces and voltage protocol (top) and I–V relationship (bottom) in cells before (control, filled circles) and after (+Aga-IVA, open circles) extracellular perfusion with ω-agatoxin-IVA (500 nM). Points represent means ± SE (n = 10). B: inactivation of P-type currents in Purkinje neurons and Ca_v2.1 (α₁2.1, β_2a, α₂δ) in transfected HEK293T cells. In Purkinje neurons, I_Ca and Ba²⁺ current (I_Ba) were evoked by 2-s pulses from −60 to −10 or −20 mV, respectively, and were evoked in transfected HEK293T cells by 2-s pulses from −80 to +10 or 0 mV, respectively. The −10-mV difference in test pulse used for I_Ba accounts for the shift in activation voltage when using Ba²⁺ as the charge carrier. Voltage protocol and current traces for I_Ca (black) and I_Ba (gray) are shown (top). I_res/I_pk (current amplitude at end of 2-s pulse normalized to peak current amplitude) is plotted below for Purkinje neurons (left) and transfected HEK293T cells (right). Parentheses indicate numbers of cells. *P < 0.01 by t-test. C: Ca²⁺-dependent facilitation during repetitive depolarizations in Ca_v2.1-transfected HEK293T cells. I_Ca and I_Ba were recorded during 5-ms steps from −80 to +10 mV (I_Ca) or 0 mV (I_Ba) delivered at 100 Hz. Fractional current represents test current amplitude normalized to the first in the train and is plotted against time. Every third data point is plotted. D: same as in C except that pulses were from −60 to 0 mV (I_Ca) or −10 mV (I_Ba) and applied to Purkinje neurons. In C and D, representative I_Ca and voltage protocols are shown above averaged data. Dashed line represents initial current amplitude.

We next compared inactivation of currents evoked by 2-s-step depolarizations using Ca²⁺ or Ba²⁺ as the permeant ion (Fig. 1B). Inactivation was measured as the current amplitude at the end of the test pulse normalized to the peak current amplitude (I_res/I_pk). With this protocol, Ca²⁺-dependent inactivation causes stronger inactivation (smaller I_res/I_pk) for Ca²⁺ currents (I_Ca) than for Ba²⁺ currents (I_Ba). By convention, inactivation of I_Ba is considered as proceeding by a purely voltage-dependent mechanism. As shown in Fig. 1B, P-type currents exhibited significant Ca²⁺-dependent inactivation similar to that of HEK293T cells transfected with cDNAs for Ca_v2.1 subunits (α₁2.1, β_2a, α₂δ).

In response to a train of 5-ms-step depolarizations, I_Ca in Ca_v2.1-transfected cells initially increases roughly 10–15% (Fig. 1C). This effect is not seen for I_Ba and represents Ca²⁺-dependent facilitation mediated by calmodulin (Lee et al. 2000). The same voltage protocol applied to mouse Purkinje neurons did not produce facilitation but only inactivation of I_Ca (Fig. 1D). Although the amplitude of facilitated I_Ca in Ca_v2.1-transfected cells declined to values approximating the initial (nonfacilitated) test current (Fig. 1C), I_Ca in Purkinje neurons inactivated roughly 60% by the end of the train (Fig. 1D). I_Ba also inactivated significantly more in Purkinje neurons than in Ca_v2.1-transfected cells during this protocol (Fig. 1, C and D), suggesting greater voltage-dependent inactivation in Purkinje neurons than that in transfected cells. Mechanistically, this could result from slower recovery from, or faster onset of, inactivation, which would increase the accumulation of inactivated channels during trains of depolarizations. Consistent with the latter possibility, inactivation of I_Ca in Purkinje neurons proceeded at a faster rate (τ = 0.77 ± 0.07, n = 24) than that in transfected cells (τ = 1.20 ± 0.06, n = 40; P < 0.001).

Voltage-dependent inactivation of the P-type current might obscure Ca²⁺-dependent facilitation during repetitive square-pulse stimuli since shorter depolarizations, such as with AP waveforms, have been shown to induce facilitation of P-type Ca²⁺ currents (Chaudhuri et al. 2005, 2007; Richards et al. 2007). Therefore we developed an AP waveform stimulus based on spontaneous APs measured first in current-clamp recordings (Fig. 2 A). At 200 Hz, AP waveforms consistently produced an initial facilitation of I_Ca (∼29 ± 4%) in Purkinje neurons, which decayed rapidly during the 1-s stimulus (Fig. 2, B and C). With the same protocol, I_Ba showed little facilitation (∼3 ± 2%), which may result from activity-dependent removal of Ca_v2.1 channels from G-protein inhibition (Brody et al. 1997; Park and Dunlap 1998). Facilitation of I_Ca was about tenfold greater than that for I_Ba, consistent with the magnitude of Ca²⁺-dependent facilitation in Ca_v2.1- transfected cells (Kreiner and Lee 2006; Lee et al. 2000). Facilitation of I_Ca was significantly greater with higher frequency stimulation (29 ± 4% for 200 Hz, 17 ± 2% for 100 Hz, and 11 ± 1% for 50 Hz; P < 0.001), which was not observed for I_Ba (3 ± 2% for 200 Hz, 2 ± 2% for 100 Hz, and 1 ± 1% for 50 Hz, P = 0.45; Fig. 2C). These results confirm that P-type currents undergo pronounced Ca²⁺-dependent facilitation in mouse Purkinje neurons, which would be greatest during periods of rapid firing.

Fig. 2.

Ca²⁺-dependent facilitation in Purkinje neurons during trains of action potential (AP) waveforms. A: spontaneous APs recorded in current-clamp (left) were used to construct an average AP waveform (right) with amplitude (+27 mV) and half-width (0.75 ms) indicated. B: representative I_Ca and I_Ba evoked by the AP waveform stimulus in A delivered at 200 Hz. Dashed line indicates current amplitude of the first pulse. C: increased Ca²⁺-dependent facilitation with higher frequency stimulation. Current responses to AP waveforms were measured at 50, 100, and 200 Hz. Fractional current represents test current amplitude normalized to the first in the train and plotted against time. For 100- and 200-Hz data, every third and fourth data point is plotted, respectively.

We next asked why voltage-dependent inactivation was stronger in Purkinje neurons than that in our Ca_v2.1-transfected cells, since this clearly influenced parameters for uncovering Ca²⁺-dependent facilitation (Fig. 1, C and D). We focused on the auxiliary Ca_vβ_2a subunit, used in the transfected cell experiments, which is unique among Ca_vβ subunits in conferring slow voltage-dependent inactivation to Ca_v2.1 (De Waard and Campbell 1995; Stea et al. 1994). Of the four Ca_vβ variants (β₁, β₂, β₃, β₄,) detected in rodent cerebellum, Ca_vβ_2a and Ca_vβ₄ are most highly expressed in Purkinje neurons (Ludwig et al. 1997; Richards et al. 2007). Relative to Ca_vβ_2a, Ca_vβ₄ causes Ca_v2.1 to undergo greater voltage-dependent inactivation (Bourinet et al. 1999; Sokolov et al. 2000; Stea et al. 1994). Therefore the inactivating profile of P-type currents in mouse Purkinje neurons may result from the combined contribution of Ca_vβ_2a and Ca_vβ₄. To test this, we compared I_Ca and I_Ba in Purkinje neurons and HEK293T cells transfected with Ca_v2.1 channels containing β_2a [Ca_v2.1(β_2a)] or β₄ [Ca_v2.1(β₄)] (Fig. 3, A–C). In contrast to square-pulse depolarizations, which caused facilitated I_Ca to decay within 0.6 s in cells transfected with Ca_v2.1(β_2a) (Fig. 1C), AP waveforms produced facilitation of I_Ca that remained maximal from about 0.2 to 0.6 s (Fig. 3B). Interestingly, maximal facilitation of I_Ca for Ca_v2.1(β₄) (28 ± 6%) was not different from that of Ca_v2.1(β_2a) (35 ± 5%, P = 0.42), but declined rapidly within 0.6 s (Fig. 3C). Thus the identity of the Ca_vβ isoform does not affect the maximal amount of Ca²⁺-dependent facilitation during AP waveform stimuli, but rather how long Ca²⁺-dependent facilitation lasts. The decline in the amplitude of the facilitated I_Ca in Ca_v2.1(β₄)-transfected cells was likely due to the onset of voltage-dependent inactivation, which was apparent in comparing the rapid decay of I_Ba for Ca_v2.1(β₄) with the maintained amplitude of I_Ba for Ca_v2.1(β_2a) (Fig. 3, B and C). With Ca²⁺ or Ba²⁺ as the charge carrier, the neuronal P-type currents evoked by AP waveforms inactivated to a level intermediate to that for Ca_v2.1(β_2a) and Ca_v2.1(β₄) (Fig. 3, D and E). Coexpression of β_2a and β₄ in HEK293T cells yielded Ca_v2.1 currents that inactivated to an extent similar to that of P-type currents in Purkinje neurons (Fig. 3, D and E). However, the initial rise time of the facilitated I_Ca was considerably faster in Purkinje neurons than in HEK293T cells transfected with any combination of β subunits, perhaps due to the expression of distinct α₁2.1 splice variants in Purkinje neurons (Chaudhuri et al. 2005; Richards et al. 2007). Nevertheless, these findings suggest that voltage-dependent inactivation and, subsequently, the maintenance of Ca²⁺-dependent facilitation depend on the ratio of Ca_vβ_2a and Ca_vβ₄ subunits in Purkinje neurons.

Fig. 3.

Ca_vβ subunits influence the profile of Ca_v2.1 currents during AP stimuli. I_Ca and I_Ba were evoked by AP waveforms at 200 Hz. Representative I_Ca (top) and fractional current measured as in Fig. 2C (bottom) in Purkinje neurons (A) and HEK293T cells transfected with α₁2.1, α₂δ, and β_2a (B) or β₄ (C). Parentheses indicate numbers of cells. Results in A–C and for cells cotransfected with β_2a and β₄ for I_Ca (D) and I_Ba (E) are superimposed on the same graph for comparison. For clarity, error bars were omitted in D and E and, for each graph, every third data point is plotted. For HEK293T cells cotransfected with both β_2a and β₄, results from exemplar cells are shown.

Comparison of P-type currents in WT and PV/CB^−/− neurons

Having established that P-type currents in mouse Purkinje neurons undergo Ca²⁺-dependent inactivation and facilitation, we next evaluated the role of endogenous Ca²⁺ buffers in Purkinje neurons. Due to their high expression levels in Purkinje neurons, the Ca²⁺-binding proteins, PV and CB, may inhibit Ca²⁺-dependent inactivation in a manner similar to that of EGTA and BAPTA [1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid] (Lee et al. 2000). Since Ca²⁺-dependent facilitation of Ca_v2.1 is mediated by local Ca²⁺ signals not affected by EGTA and BAPTA (DeMaria et al. 2001; Lee et al. 2000; Liang et al. 2003), PV and CB should not affect Ca²⁺-dependent facilitation. If so, then Ca²⁺-dependent inactivation but not facilitation should be greater in Purkinje neurons lacking PV and CB. To test this, we compared P-type currents in Purkinje neurons isolated from WT and PV/CB^−/− mice.

In general, P-type currents in WT and PV/CB^−/− neurons showed similar activation properties, although the amplitude of I_Ca was generally greater in PV/CB^−/− than that in WT neurons (Supplemental Table S2). There were no significant differences in cell capacitance (16.9 ± 0.6 pF for WT vs. 17.6 ± 0.7 pF for PV/CB^−/−, P = 0.45) or the extent of block by ω-agatoxin IVA (90.0 ± 1.6% for WT vs. 84.4 ± 8.3% for PV/CB^−/−; P = 0.20; Fig. 4). These results indicated that genetic deletion of PV and CB did not influence the contribution of Ca_v2.1 to the whole cell Ca²⁺ current in Purkinje neurons.

Fig. 4.

Characterization of P-type currents in PV/CB^−/− Purkinje neurons. A: representative current I_Ca (A) and I–V relationship (B) in cells before (filled circles) and after (open circles) exposure to ω-agatoxin IVA (500 nM). Points represent means ± SE (n = 12). CB, calbindin D-28k; PV, parvalbumin.

If PV and CB specifically attenuated Ca²⁺-dependent inactivation, we would expect an increase in inactivation of I_Ca in PV/CB^−/− compared with that in WT neurons; the decay of I_Ba due to voltage-dependent inactivation should not be affected. However, we found that both I_Ca and I_Ba underwent greater inactivation in PV/CB^−/− than that in WT neurons (Fig. 5, A and B). Since I_Ca inactivation contains both Ca²⁺-dependent and voltage-dependent components, we measured the difference between I_res/I_pk for I_Ca and I_Ba to isolate Ca²⁺-dependent inactivation and found that it was similar in WT and PV/CB^−/− neurons (Fig. 5A). Together, these results indicate that voltage-dependent inactivation is increased in PV/CB^−/− compared with that in WT neurons. Kinetic analyses showed significantly faster rates of inactivation for I_Ba in PV/CB^−/− (τ = 1.05 ± 0.06, n = 12) than those in WT (τ = 1.69 ± 0.23, n = 15; P = 0.02) neurons, although there was no difference in inactivation rate for I_Ca (τ = 0.77 ± 0.07, n = 24, for WT vs. τ = 0.72 ± 0.05, n = 23, for PV/CB^−/−; P = 0.5). Stronger voltage-dependent inactivation in PV/CB^−/− neurons was particularly apparent at high stimulation frequencies (Fig. 6, A–D): P-type currents during 200-Hz AP waveforms decayed to initial levels within 0.4 s in PV/CB^−/− neurons, whereas currents in WT neurons remained strongly facilitated (∼20%) at this time point (Fig. 6, C and D).

Fig. 5.

Increased voltage- but not Ca²⁺-dependent inactivation (CDI) of Ca_v2.1 currents in PV/CB^−/− Purkinje neurons. A: I_Ca and I_Ba were evoked by 2-s pulses from −60 to 0 mV (I_Ca) or −10 mV (I_Ba) in wild-type (WT) and PV/CB^−/− neurons. CDI represents the difference in I_res/I_pk for I_Ca and I_Ba in WT and PV/CB^−/− neurons. B: comparison of I_res/I_pk for I_Ca and I_Ba (measured with 2 mM extracellular Ba²⁺) in WT (black traces) and PV/CB^−/− neurons (gray traces). *P < 0.05 by t-test. Parentheses indicate numbers of cells.

Fig. 6.

Increased inactivation of P-type currents during high-frequency stimulation in PV/CB^−/− Purkinje neurons. I_Ca was evoked by AP waveform stimuli given at 50 (A), 100 (B), or 200 Hz (C) in WT (filled circles) and PV/CB^−/− (open circles) Purkinje neurons. Every other point is plotted for 200-Hz data. D: normalized I_Ca represents the grand average of current amplitude for the last 10 pulses (±SE) plotted against stimulation frequency for WT and PV/CB^−/− neurons. **P < 0.01; *P < 0.05 compared with WT by t-test.

Because the stronger voltage-dependent inactivation in PV/CB^−/− neurons could not be linked directly to the absence of Ca²⁺ buffers, we hypothesized that the differences in P-type currents in WT and PV/CB^−/− neurons could have arisen from a compensatory change in Ca_v2.1 subunit expression. As alluded to in Fig. 3, a decrease in the ratio of Ca_vβ_2a and Ca_vβ₄ could boost the number of Ca_v2.1 channels undergoing strong voltage-dependent inactivation. We tested this possibility by PCR of isolated Purkinje neurons with primers specific for the different Ca_vβ subunits (Fig. 7, A–D). Because dissociated preparations of Purkinje neurons also contain a vast number of other cell types, mainly granule cells, we performed control experiments to ensure that we were isolating nucleic acids mainly from Purkinje neurons. In these experiments, calbindin D-28k mRNA (gene symbol: Calb1) was amplified from cere bellum and Purkinje neurons from WT mice but, as expected, not from PV/CB^−/− neurons (Fig. 7A). In the cerebellum, calretinin (gene symbol: Calb2) is not expressed in Purkinje neurons but in unipolar brush cells, Lugaro cells, and predominantly in granule cells (Schwaller et al. 2002). As expected, calretinin mRNA was amplified only from total cerebellar RNA but not from RNA isolated from Purkinje neurons (Fig. 7A). These results validated the relative absence of contamination by granule cells in the samples of isolated Purkinje neurons.

Fig. 7.

Altered Ca_v2.1 subunit expression in PV/CB^−/− Purkinje neurons. A: control experiment in which cerebellar tissue or pools of 5–10 Purkinje neurons from WT or PV/CB^−/− mice were subjected to polymerase chain reaction (PCR) with primers specific for calbindin D-28k (Calb1, CB), calretinin (Calb2, CR), or GAPDH. B: representative experiment showing amplification of all 4 β subunits in WT Purkinje neurons (n = 5) and the absence of β₂ in PV/CB^−/− neurons (n = 5). C: PCR analysis of Ca_vβ subunits (β₁–β₄) in cerebellar tissue from WT and PV/CB^−/− mice. D: average results indicating percentage of individual neurons in WT and PV/CB^−/− mice in which β₂ was detected (left). Parentheses indicate numbers of cells. *P < 0.05 by t-test. Right: results from quantitative PCR from 5 to 10 Purkinje neurons using β_2a-specific primers. Relative levels of β_2a in WT and PV^−/− neurons were determined as described in methods. Shown are averaged results from 5 independent experiments. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

As shown previously (Richards et al. 2007), all four Ca_vβ variants were detected in whole cerebellum (Fig. 7C) and Purkinje neurons from WT mice (Fig. 7B). Similar results were obtained in extracts from total cerebellum of PV/CB^−/− mice (Fig. 7C), but not in the extracts from single isolated Purkinje neurons of null-mutant mice (Fig. 7B). Ca_vβ₂ was amplified from most (75%) WT Purkinje neurons and in a significantly smaller fraction (<50%, P < 0.05) of PV/CB^−/− Purkinje neurons (Fig. 7D). Of the Ca_vβ₂ splice variants expressed in mouse Purkinje neurons (β_2a–d), Ca_vβ_2a is the most prominent (Richards et al. 2007). Our results suggested that Ca_vβ_2a was specifically down-regulated in PV/CB^−/− neurons below the threshold for detection in a one-step PCR protocol. We explored this possibility by quantitative PCR with Ca_vβ_2a-specific primers (Supplemental Table S1). Because initial efforts to perform quantitative PCR with individual Purkinje neurons yielded inconsistent results, pools of 5–10 Purkinje neurons were analyzed with GAPDH as an internal reference. Consistent with the single-cell analyses, these experiments indicated levels of Ca_vβ_2a in PV/CB^−/− neurons that were only 53 ± 8% of that in WT Purkinje neurons (Fig. 7D). The nearly twofold decrease in Ca_vβ_2a in PV/CB^−/− neurons may therefore contribute to increased voltage-dependent inactivation of P-type currents in these neurons (Fig. 5B).

No differences in spontaneous firing properties in WT and PV/CB^−/− neurons

Similar to their behavior in vivo, acutely isolated Purkinje neurons fire spontaneous APs that are driven primarily by voltage-gated Na⁺ channels (Raman and Bean 1999). However, Ca_v2.1 channels can regulate this process through selective coupling to Ca²⁺-activated (K_Ca) K⁺ channels. By conducting Ca²⁺ signals that activate K_Ca channels, Ca_v2.1 channels control the afterhyperpolarization (AHP) and, subsequently, the spontaneous firing rate in Purkinje neurons (Edgerton and Reinhart 2003; Walter et al. 2006; Womack and Khodakhah 2002; Womack et al. 2004). To understand the physiological consequences of our findings, we asked whether increased voltage-dependent inactivation of P-type currents might affect spontaneous firing in PV/CB^−/− Purkinje neurons.

At first glance, limited Ca²⁺ influx due to faster inactivation of Ca_v2.1 channels might inhibit K_Ca contribution to the AHP such that the spontaneous firing rate would be faster in PV/CB^−/− neurons relative to that in WT. However, as endogenous Ca²⁺ buffers, PV and particularly the fast buffer CB could normally limit coupling between Ca_v2.1 Ca²⁺ signals and K_Ca channels. In this case, one would expect that in PV/CB^−/− neurons, spontaneous firing might be slower than that in WT due to stronger K_Ca-mediated AHPs. On the other hand, the molecular switch to Ca_v2.1 channels that undergo stronger voltage-dependent inactivation in PV/CB^−/− neurons might offset aberrant K_Ca activation, such that net excitability is unaltered. To test these possibilities, we compared spontaneous firing in isolated WT and PV/CB^−/− Purkinje neurons in current-clamp recordings with physiological solutions at room temperature. Under these conditions, nearly all Purkinje neurons from WT and PV/CB^−/− neurons fired spontaneously in the absence of injected current. There were no significant differences in AHP amplitude or other parameters of spontaneous APs in WT and PV/CB^−/− neurons (Fig. 8, A–F). The mean interspike interval for WT and PV/CB^−/− neurons was not significantly different (Fig. 8G), which indicated a similar firing rate in neurons from the two genotypes. We conclude that spontaneous firing of isolated Purkinje neurons is not perturbed in PV/CB^−/− neurons, which may be due in part to compensatory functional changes in Ca_v2.1.

Fig. 8.

No differences in spontaneous APs in WT and PV/CB^−/− Purkinje neurons. Spontaneous APs were recorded from (A) WT and (B) PV/CB^−/− neurons in current clamp without injected current. C: average AP waveform for WT (black trace, n = 11) and PV/CB^−/− (gray trace, n = 10) neurons. For each neuron, averages are from all APs recorded in 10 consecutive sweeps of 500-ms duration. D–G: parameters for spontaneous APs in WT (circles) and PV/CB^−/− (squares) neurons. Afterhyperpolarization (AHP) amplitude (D) was determined as the difference between the membrane potential measured 5 ms before the AP peak (double arrows in C) and the minimum membrane potential after the AP peak (single arrow in C). AP half-width (E) represents the duration of the AP at half-maximal amplitude from the threshold membrane potential. AP threshold (F) was measured at 5–15% of the tangential slope of the threshold depolarization. Mean interspike interval (ISI, G) is the average duration between peaks of the APs. Each symbol represents averaged data recorded as in A from one Purkinje neuron. Population averages are indicated with filled symbol (±SE).

DISCUSSION

Our results reveal multiple mechanisms governing the behavior of P-type currents in Purkinje neurons. First, P-type currents undergo Ca²⁺-dependent inactivation and facilitation, qualitatively similar to calmodulin-mediated regulation of recombinant Ca_v2.1 channels. Second, the inactivating profile of P-type currents during trains of AP stimuli may depend on the relative expression levels of different Ca_vβ subunits. Third, the Ca²⁺-binding proteins PV and CB are required for maintaining the normal properties of the P-type current. Multifaceted regulation of Ca_v2.1 in Purkinje neurons may ensure the temporal precision of Ca²⁺ signals required for the control of motor coordination.

Ca²⁺-dependent inactivation and facilitation of Ca_v2.1 in Purkinje neurons

Although the molecular details underlying Ca_v2.1 modulation by Ca²⁺/calmodulin have been elucidated at the atomic level (Kim et al. 2008; Mori et al. 2008), the prevalence of Ca²⁺ feedback regulation of Ca_v2.1 in neurons is less clear. Ca²⁺-dependent inactivation and facilitation were described for P-type currents in presynaptic terminals at the Calyx of Held (Cuttle et al. 1998; Forsythe et al. 1998) but were less robust for P/Q-type currents in chromaffin cells (Wykes et al. 2007). In rat Purkinje neurons, it was shown that P-type currents undergo Ca²⁺-dependent facilitation, whereas Ca²⁺-dependent inactivation was highly variable between cells (Chaudhuri et al. 2005). Ca²⁺-dependent inactivation was measurable in most of the mouse Purkinje neurons in our study (Figs. 1B and 5), which may be due to a species- or age-related difference in Ca_v2.1 properties since we used older mice (14 to 21 days old) rather than young (6 to 10 days old) rats as in the previous study. Considering the net hyperpolarizing function of P-type Ca²⁺ currents through coupling to K_Ca channels (Edgerton and Reinhart 2003; Llinás and Sugimori 1980b; Raman and Bean 1999; Womack et al. 2004), Ca²⁺-dependent inactivation of Ca_v2.1 may limit Ca²⁺ signals that would otherwise constrain the potential of Purkinje neurons to fire at high rates.

Our findings that Ca²⁺-dependent facilitation of the P-type current in Purkinje neurons increases with spike frequency (Fig. 2C) are consistent with previous work in rat and mouse (Chaudhuri et al. 2005; Richards et al. 2007). Mechanistically, Ca²⁺-dependent facilitation results from enhanced channel open probability due to Ca²⁺-dependent conformational changes in calmodulin associated with Ca_v2.1 (Chaudhuri et al. 2007). This process may be intrinsically favored by repetitive depolarizations. Alternatively, high-frequency AP trains may somehow derepress Ca²⁺-dependent facilitation of the P-type current. The presence of an endogenous inhibitor of Ca²⁺-dependent facilitation was suggested from findings that intracellular perfusion of Purkinje neurons with recombinant calmodulin significantly increased Ca²⁺-dependent facilitation of the P-type current (Chaudhuri et al. 2005). High-frequency stimulation may enhance the effectiveness of endogenous calmodulin in competing with this inhibitor, thus leading to greater Ca²⁺-dependent facilitation for currents evoked by 200- than that by 100- and 50-Hz AP stimuli (Fig. 2C). In Purkinje neurons, increased recruitment of Ca_v2.1 by high-frequency stimulation may underlie activity-dependent Ca²⁺ signals required for Ca_v2.1 coupling to activation of calmodulin-dependent protein kinase (Mizutani et al. 2008) and depolarization-activated Ca²⁺ signals leading to long-term depression (Jin et al. 2007).

Altered Ca_v2.1 properties and Ca²⁺ homeostasis in PV/CB^−/− Purkinje neurons

PV and CB act as mobile Ca²⁺ buffers, which can significantly alter the kinetics of Ca²⁺ signals in Purkinje neurons (Schmidt et al. 2003, 2007). We could not directly test whether PV and CB, like EGTA (Lee et al. 1999, 2000), might temper Ca²⁺-dependent inactivation of the P-type current since the intrinsic properties of the P-type current differed in WT and PV/CB^−/− Purkinje neurons (Figs. 5 and 6). Decreased contribution of Ca_vβ_2A subunits may permit greater voltage-dependent inactivation of P-type currents mediated by other Ca_vβ subunits in PV/CB^−/− neurons (Fig. 7, B and D). However, this interpretation does not rule out the involvement of other mechanisms. Multiple α₁2.1 splice variants have been detected in Purkinje neurons (Bourinet et al. 1999). Some have variable cytoplasmic C-terminal domains (Kanumilli et al. 2006; Richards et al. 2007; Tsunemi et al. 2002), which can increase voltage-dependent inactivation of Ca_v2.1 (Krovetz et al. 2000). Increased expression of these variants could also contribute to the more strongly inactivating P-type currents in PV/CB^−/− neurons.

Since PV and CB reach their mature expression levels in mice near P20 (Schwaller et al. 2002), it is tempting to speculate that increased Ca_v2.1 channel inactivation may be characteristic of the immature Purkinje neuron. Consistent with this possibility, in situ hybridization studies show that mRNA for Ca_vβ₂ subunits is weak in P1 rat Purkinje neurons, but increase from P7 to P14 (Tanaka et al. 1995). Interestingly, α₁2.1 splice variants that undergo greater Ca²⁺-dependent facilitation also increase in expression in rat Purkinje neurons during this age range (Chaudhuri et al. 2005). Developmental up-regulation of Ca_v2.1 subunits producing less inactivation and greater facilitation may increase the gain of Ca²⁺ signals because Ca²⁺ buffering strength increases due to the onset of PV and CB expression.

Although the spontaneous Purkinje cell firing properties were not different for WT and in PV/CB^−/− mice in vitro (Fig. 8), the firing rate of single spikes from Purkinje neurons in vivo are higher in PV/CB^−/− (∼70 Hz) than that in WT mice (∼45 Hz) (Servais et al. 2005). This difference may result from the influence of dendritic ion channels and synaptic inputs present in vivo that would not be retained in our in vitro recordings of dissociated Purkinje neurons. However, at the behavioral level, elimination of PV and CB results in only minor alterations in locomotor activity and coordination, not detectable in mice maintained under standard housing conditions (Bouilleret et al. 2000; Farre-Castany et al. 2007). In contrast to the severe ataxia exhibited by mice with loss-of-function mutations in the Ca_v2.1 α₁-subunit (Pietrobon 2005), the relatively modest motor phenotype in PV/CB^−/− mice may result from recruitment of multiple homeostatic mechanisms. The volume of mitochondria—organelles with a high Ca²⁺ uptake capacity, yet rather slow Ca²⁺ uptake kinetics—is increased selectively in a subplasmalemmal zone in Purkinje cells deficient for PV (Chen et al. 2006). In addition, the volume and length of dendritic spines are increased in Purkinje neurons from mice lacking CB (Vecellio et al. 2000). Together with Ca_v2.1 channels undergoing greater inactivation, these morphological changes may help Purkinje neurons cope with activity-dependent increases in Ca²⁺ that might otherwise degrade their information-encoding potential.

Our results are the first to demonstrate alterations in Ca_v2.1 channels as a consequence of loss of PV and CB. In addition to Purkinje neurons, PV and CB are highly expressed in distinct populations of interneurons (Celio 1990), where their expression is reduced following seizure activity (Kim et al. 2006; Magloczky et al. 1997). Modification of Ca_v2.1 properties through molecular switching of α₁2.1 splice variants and β subunits may represent a general mechanism to protect neurons against aberrant Ca²⁺ signaling and pathological Ca²⁺ overloads.

GRANTS

This work was supported by National Institutes of Health Grants R01-NS-044922 and R01-DC-009433 to A. Lee, F31-NS-04975 to L. Kreiner, and S11-NS-055883 to M. Benveniste; American Heart Association and Israel Binational Science Foundation grants to A. Lee; and Swiss National Science Foundation Grants 3100A0-100400/1 and 310000-113518/1 to B. Schwaller.