Localized Calcineurin Confers Ca²⁺-Dependent Inactivation on Neuronal L-Type Ca²⁺ Channels

Abstract

Excitation-driven entry of Ca²⁺ through L-type voltage-gated Ca²⁺ channels controls gene expression in neurons and a variety of fundamental activities in other kinds of excitable cells. The probability of opening of Ca_V1.2 L-type channels is subject to pronounced enhancement by cAMP-dependent protein kinase (PKA), which is scaffolded to Ca_V1.2 channels by A-kinase anchoring proteins (AKAPs). Ca_V1.2 channels also undergo negative autoregulation via Ca²⁺-dependent inactivation (CDI), which strongly limits Ca²⁺ entry. An abundance of evidence indicates that CDI relies upon binding of Ca²⁺/calmodulin (CaM) to an isoleucine–glutamine motif in the carboxy tail of Ca_V1.2 L-type channels, a molecular mechanism seemingly unrelated to phosphorylation-mediated channel enhancement. But our work reveals, in cultured hippocampal neurons and a heterologous expression system, that the Ca²⁺/CaM-activated phosphatase calcineurin (CaN) is scaffolded to Ca_V1.2 channels by the neuronal anchoring protein AKAP79/150, and that overexpression of an AKAP79/150 mutant incapable of binding CaN (ΔPIX; CaN-binding PXIXIT motif deleted) impedes CDI. Interventions that suppress CaN activity—mutation in its catalytic site, antagonism with cyclosporine A or FK506, or intracellular perfusion with a peptide mimicking the sequence of the phosphatase's autoinhibitory domain—interfere with normal CDI. In cultured hippocampal neurons from a ΔPIX knock-in mouse, CDI is absent. Results of experiments with the adenylyl cyclase stimulator forskolin and with the PKA inhibitor PKI suggest that Ca²⁺/CaM-activated CaN promotes CDI by reversing channel enhancement effectuated by kinases such as PKA. Hence, our investigation of AKAP79/150-anchored CaN reconciles the CaM-based model of CDI with an earlier, seemingly contradictory model based on dephosphorylation signaling.

Introduction

Voltage-gated Ca²⁺ channel activity is tuned to cellular need by a variety of processes, with Ca²⁺-dependent inactivation (CDI) providing one form of Ca²⁺-driven feedback control. Elimination of CDI from L-type Ca_V1.2 channels pathologically prolongs action potentials in cardiomyocytes (Alseikhan et al., 2002), illustrating the importance of feedback regulation of Ca²⁺ channels. CDI also reduces non-L-type Ca_V2.1 channel activity and thereby contributes to short-term synaptic depression during repetitive stimulation (Mochida et al., 2008).

The Ca²⁺ sensor for CDI is Ca²⁺/calmodulin (CaM) (Peterson et al., 1999; Zühlke et al., 1999), which at resting [Ca²⁺]_in is preassociated with the Ca_V1.2 C terminus (Pitt et al., 2001; Erickson et al., 2001, 2003). Ca²⁺ entering the cytoplasm through Ca_V1.2 binds to CaM, initiating Ca²⁺/CaM interaction with an isoleucine–glutamine (IQ) motif located adjacent to the CaM preassociation site (Zühlke et al., 1999, 2000; Erickson et al., 2003; Kim et al., 2004). Because disruption of the IQ motif impairs CDI, Ca²⁺/CaM interaction with the IQ motif has been proposed to induce directly a change in channel conformational that corresponds to CDI.

But two sets of observations suggest an alternative to this nonenzymatic model of CDI. First, Ca_V1.2-anchored Ca²⁺/CaM often acts via downstream enzymes. Examples include Ca²⁺-dependent facilitation (CDF) of Ca_V1.2 by CaM-dependent protein kinase II (CaMKII) (Noble and Shimoni, 1981; Zühlke et al., 2000; Hudmon et al., 2005; Erxleben et al., 2006; Grueter et al., 2006) and coupling of neuronal excitation to activation of transcription factors by kinases and phosphatases (Dolmetsch et al., 2001; Oliveria et al., 2007). Second, Chad and Eckert (1986) found for molluscan neurons that protein kinase A (PKA) activity slowed Ca_V channel inactivation, and that calcineurin (CaN) opposed this effect, suggesting an enzymatic mechanism for CDI (for review, see Armstrong, 1989). Modulation by PKA of Ca_V1.2 CDI has since been described in a variety of excitable cell types (Bean et al., 1984; Kalman et al., 1988; Hadley and Lederer, 1991; Meuth et al., 2002). Yet a role for CaN in CDI has remained controversial, in part owing to inconsistent responses to inhibition of CaN (Branchaw et al., 1997; Victor et al., 1997; Lukyanetz et al., 1998; Burley and Sihra, 2000; Zeilhofer et al., 2000; Meuth et al., 2002) and to a lack of appreciation of the role of scaffolding proteins in targeting CaN to Ca_V1.2.

In hippocampal neurons, Ca²⁺ and cAMP reciprocally regulate L channel activity by activation of CaN and PKA that are coanchored to Ca_V1.2 channels by the scaffolding protein AKAP79/150 (human/rodent) (Oliveria et al., 2007). AKAP79/150-anchored CaN suppresses enhancement of neuronal L-type Ca²⁺ channel activity by coanchored, activated PKA (Oliveria et al., 2007). CaN effectiveness can in part be attributed to its localization within the Ca_V1.2 macromolecular complex, where CaN is advantageously placed to interact with its activator, Ca²⁺/CaM. Here we describe experimental results indicating that CaM, the Ca²⁺ sensor for CDI, triggers CDI in hippocampal neurons by activating AKAP879/150-anchored CaN.

Materials and Methods

Cell culture and transfection of tsA201 cells.

tsA201 cells were transfected with equimolar ratios of cDNAs encoding Ca_V1.2 channel subunits (α₁1.2a, β_2b, α₂δ_1a) and other proteins, including AKAP79, using the Effectene transfection reagent (Qiagen). Cells were plated onto glass coverslips and studied 1–3 d after transfection.

Preparation and transfection of primary, short-term cultured hippocampal neurons.

Primary hippocampal neurons were prepared from neonatal Sprague Dawley rats [postnatal day 0 (P0)–P3] as described previously (Gomez et al., 2002; Smith et al., 2006). For Amaxa Nucleofector transfection, P0–P3 neurons were resuspended at 2 × 10⁶ to 3 × 10⁶ cells per transfection, electroporated with 6–8 μg of total DNA (including GFP or CaN_H151A-YFP cDNA and/or the 150RNAi construct), plated at 0.25 × 10⁶ to 0.5 × 10⁶ cells per 25 mm glass coverslip, and grown up to 6 d in vitro. Current waveform and tail current time course indicated that these short-term cultured neurons were electrotonically more compact than older neurons, permitting analysis of the inactivation kinetics of pharmacologically isolated L currents.

AKAP150ΔPIX knock-in mice.

AKAP150ΔPIX mice were developed in conjunction with the Rocky Mountain Neurological Disorders Transgenic and Gene Targeting Vectors Core and housed at the University of Colorado Anschutz Medical Campus Center for Comparative Medicine. Briefly, the AKAP150ΔPIX mutation, which removes the conserved PIAIIIT motif (amino acids 655–661), was introduced into the single AKAP150 coding exon in the AKAP5 gene by homologous recombination in C57BL/6J×129 F1 embryonic stem cells using a loxP-flanked neomycin cassette located downstream of the exon. Targeted ES clones were identified by G418 selection for the neomycin resistance gene and PCR screening of genomic DNA, expanded, and injected into blastocysts that were implanted into surrogate females. Heterozygous and homozygous mice carrying the mutation were derived from the resulting F0 chimeric founders by standard breeding with C57BL/6J. AKAP150ΔPIX mice housed in a home-cage environment had no obvious phenotypic alterations in body weight, life span, breeding, or behavior. A full description of the AKAP150ΔPIX knock-in targeting vector and additional details of generation and characterization of AKAP150ΔPIX mice was reported by Sanderson et al. (2012). Dissociated mouse neurons were cultured as for rat neurons.

Patch-clamp recording.

Borosilicate patch pipettes were heat-polished to a resistance, measured in the bath, of 3–6 MΩ. Voltage-clamped currents were measured with an Axopatch 1D amplifier (Molecular Devices), filtered at 2 kHz, and sampled at 10 kHz using Pulse software (HEKA) and an ITC-16 analog-to-digital and digital-to-analog interface (HEKA). Series resistance compensation, capacitance cancellation, and leak subtraction (−P/4 protocol) were used.

For tsA201 cells, the whole-cell pipette contained the following (in mm): 135 CsCl, 10 EGTA or 1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA), 10 HEPES, and 4 MgATP, pH adjusted to 7.5 with tetraethylammonium-hydroxide (TEA-OH). For recordings in Ca²⁺ or Ba²⁺, the bath solution contained the following (in mm): 125 NaCl, 15 CaCl₂ or 15 BaCl₂, 10 HEPES, and 20 sucrose, pH adjusted to 7.3 with TEA-OH. For recordings of Na⁺ currents carried by Ca²⁺ channels, the whole-cell pipette contained the solution described above, but supplemented with 10 mm BAPTA, and the bath solution contained the following (in mm): 140 NaCl, 5 CsCl, 5 EGTA, 10 HEPES, and 20 sucrose, pH adjusted to 7.3 with TEA-OH. Positive transfectants were identified by GFP, AKAP79-YFP, CaM-YFP, CaN-CFP, or CaN-YFP fluorescence. When unlabeled AKAP79 was transfected, only cells exhibiting membrane-localized CaN were selected for recording. In each tsA201 cell where inactivation rate was studied, the holding potential was −80 mV, and a sequence of 500-ms step depolarizations, from −60 to +70 mV in 10 mV increments, was applied for both Ca²⁺ and Ba²⁺ solutions. A fast perfusion system (Warner Instruments) and slower bath perfusion were used in combination to ensure rapid and complete solution exchange. Cells that exhibited an artifactual current–voltage relationship were excluded from further analysis. CdCl₂ (200 μm) was applied at the end of each experiment to isolate Ca²⁺ current and Ba²⁺ current from leak or other contamination. Na⁺ current recordings were performed similarly, but not in the same cells as the Ca²⁺ and Ba²⁺ current recordings. To provide a reliable basis for comparison of Ca²⁺ and Ba²⁺ inactivation rates, currents were generally allowed to stabilize for ∼5 min after establishment of whole-cell recording. In the experiments examining the time course of development for forskolin action inactivation (see Fig. 5), the stabilization period was necessarily omitted.

Figure 5.

Time course of forskolin action supports a role for AKAP79-anchored calcineurin in CDI of Ca_V1.2 channels expressed in tsA201 cells. A, Whole-cell recording with 5 μm forskolin in the pipette, with breakthrough into whole-cell mode at time 0 min. For cells transfected with 79ΔPIX, the rate of inactivation of Ca²⁺ current slowed as forskolin entered the cell. Cells transfected with wild-type AKAP79 (79) showed no slowing of CDI during intracellular perfusion with forskolin. ^#p < 0.05; ^##p < 0.01 vs 0.5 min (paired t test); *p < 0.05 vs AKAP79. Ca²⁺ current densities were as follows (in pA/pF): 79 + FSK, −29.9 ± 5.6; 79ΔPIX + FSK, −25.0 ± 5.0. B, Bath application of FK506 (10 μm) plus cyclosporin A (10 μm) significantly slowed inactivation of forskolin-enhanced Ca_V1.2 Ca²⁺currents in cells overexpressing AKAP79. ^#p < 0.05 (paired t test). Top right, Ca²⁺ current records before and during drug application are scaled to the same peak amplitude to facilitate comparison of inactivation rate.

For cultured hippocampal pyramidal neurons, the whole-cell pipette contained the following (in mm): 120 CsMeSO₄, 30 TEA-Cl, 10 EGTA, 5 MgCl₂, 5 Na₂ATP, and 10 HEPES, pH 7.2. The extracellular solution contained the following (in mm): 125 NaCl, 10 BaCl₂ or CaCl₂, 5.85 KCl, 22.5 TEA-Cl, 1.2 MgCl₂, 10 HEPES(Na), and 11 d-glucose, pH 7.4, as well as tetrodotoxin (1 μm). To isolate L-type currents, N- and P/Q-type Ca²⁺ channel currents were blocked by preincubating neurons in extracellular recording solution supplemented with ω-conotoxins (ω-CTx) GVIA (1 μm) and MVIIC (5 μm) for 30 min before recording (Tavalin et al., 2004; Oliveria et al., 2007). Neurons were used for <1 h after preincubation to minimize contamination from unblocked N- and P/Q-type currents [τ_off, ∼200 min for ω-CTx-MVIIC unbinding from P/Q-type channels (Sather et al., 1993); ω-CTx-GVIA binding to N-type channels was even more prolonged (McCleskey et al., 1987)]. A holding potential of −60 mV was chosen to inactivate R-type Ca²⁺ channel current (>80%) (Sochivko et al., 2003). Transfected neurons were identified by GFP or CaN_H151A-YFP fluorescence. Only neurons with an access resistance <10 MΩ were studied.

To analyze inactivation time course during 500 ms depolarizations to 0 mV, the decaying phase of each current record was best fit with a double-exponential function; in a very few cases, single exponential fits provided, by eye, equally good fits (PulseFit software; HEKA). For the double-exponential fits, the faster time constant (τ_fast) provided an index of Ca²⁺-dependent inactivation, and the slower component represented voltage-dependent inactivation (Adams and Tanabe, 1997). Currents were obtained in Ba²⁺ and Ca²⁺ in the same cell, with Ba²⁺ currents recorded before Ca²⁺ currents.

Where indicated, forskolin (5 μm), protein kinase inhibitor fragment (PKI; residues 6–22; 5 μm), cyclosporin A (CsA; 5 μm), the CaN autoinhibitory peptide (AIP; residues 457–482; 100 μm), or CaN anchoring antagonist peptide VIVIT (Oliveria et al., 2007; Li et al., 2012) (biotin-GPHPVIVITGPHEE, 10, 25 or 100 μm) was included in the pipette solution. To examine reversibility of CaN inhibition, cyclosporin and FK506 (10 μm each) were added to the bath in some experiments, as indicated in the text. Forskolin, PKI, cyclosporin A, and FK506 were obtained from Sigma-Aldrich; the CaN autoinhibitory peptide was obtained from Biomol; and the VIVIT peptide was obtained from Biomatik.

Ratiometric fluorometry.

Measurements of YFP⁵³⁵/CFP⁴⁷⁰ were made using a ratiometric fluorimeter (Solamere Technology Group) where the donor fluorophore, CFP, was excited (excitation bandpass, 436 ± 5 nm), and emissions from both CFP (CFP⁴⁷⁰) and YFP (YFP⁵³⁵) were collected using parallel photomultiplier tubes fronted by a dichroic mirror and bandpass filters (CFP bandpass, 470 ± 15 nm; YFP bandpass, 535 ± 15 nm). In all experiments, simultaneous emission ratio (YFP⁵³⁵/CFP⁴⁷⁰) and current (I_Ca) measurements were obtained. CFP was excited (shutter open) for 50 ms periods centered in time on the membrane depolarizations (20 ms steps; 0.067 or 1 Hz), and the emission ratio was measured as the mean YFP⁵³⁵/CFP⁴⁷⁰ value during each 50 ms depolarization. Concurrent I_Ca and YFP⁵³⁵/CFP⁴⁷⁰ signals were acquired and recorded using Pulse software (HEKA). The step depolarization frequency of 0.067 Hz was chosen to establish baseline emission ratio and current values because this was the highest frequency that was indistinguishable from the baseline determined for cells held at −80 mV for an extended period between depolarizing steps.

Fluorescence resonance energy transfer microscopy.

Fluorescence images were acquired 2–3 d after transfection from live tsA201 cells using a Nikon TE-300 inverted microscope equipped with a 175 W xenon illumination source, 100× oil-immersion objective lens, 16 MHz CCD camera (SensiCam QE; Cooke), and dual filter wheels (Sutter Instruments) controlled by SlideBook 4.0 software (Intelligent Imaging Innovation). For sensitized fluorescence resonance energy transfer (FRET) measurements [FRET^C; three filter (3F)], an 86002 (JP4) dichroic mirror (Chroma) and three different filter sets (donor/CFP, acceptor/YFP, and raw FRET) were used to capture serially a set of three images from a fixed image plane. The three images were captured using the same exposure time (100 ms in some experiments, 250 ms in others). The three filter sets used were donor [CFP; 436 center excitation wavelength and 10 nm bandwidth (436/10 nm); emission, 470/30 nm], acceptor (YFP; excitation, 500/20 nm; emission, 535/30 nm), and FRET (raw FRET; excitation, 436/10 nm; emission, 535/30 nm). Light that has passed through the FRET filter set is contaminated by donor bleed-through (average fraction, 0.50) and acceptor cross-excitation (average fraction, 0.02); fractional contamination by bleed-through and cross-excitation were determined in separate experiments using cells that expressed CFP- or YFP-tagged constructs alone. Corrected, sensitized FRET (FRET^C) images were obtained by subtracting the contamination components, pixel by pixel, from raw FRET images using the following equation (adapted from Gordon et al., 1998): FRET^C = rawFRET − (0.50 * CFP) − (0.02 * YFP).

To obtain estimates of effective FRET efficiency (E_eff) from images, Slidebook 4.0 was used to draw masks that isolated in-focus CaN-CFP fluorescence. From masked raw FRET, CFP, and YFP images, the FRET ratio (FR) was extracted as FR = [rawFRET − (0.5 * CFP)]/(0.02 * YFP).

E_eff was then calculated as E_{eff, 3F} = (FR − 1) * (ε_YFP440/ε_CFP440), where ε_YFP440 and ε_CFP440 represent the average molar extinction coefficients for YFP and CFP over the bandpass of the CFP excitation filter, respectively. The measurement E_eff takes into account cell-to-cell variation in expression of YFP and CFP, so that this FRET index is effectively independent of donor and acceptor concentration (Erickson et al., 2003; Oliveria et al., 2007).

FRET was additionally estimated using an acceptor photobleach (PB) method. PB measurements were obtained from the same cells used to make the 3F measurements. PB FRET was measured as the difference between CFP (donor) fluorescence intensity before and after photobleach of YFP (acceptor). Photobleach was achieved by several minutes of continuous illumination at 535 nm, and was virtually complete. E_eff was calculated from acceptor photobleach images as E_{eff, PB} = 1 − (F_DA/F_D), where F_DA and F_D indicate donor intensity before and after photobleaching, respectively [E_eff calculations are adapted from the study by Erickson et al. (2003)].

Statistical analysis.

Statistical analyses were performed using Student's t test, ANOVA, or Bonferroni's test. All error bars indicate SEM.

Results

AKAP79 and the Ca_V1.2 IQ motif sustain CaM–CaN signaling

We showed previously that Ca_V1.2 currents enhanced by the adenylyl cyclase agonist forskolin can be de-enhanced through CaN activation during episodes of increased stimulation frequency (Oliveria et al., 2007). To test whether CaM docked at the IQ motif is the activator of anchored CaN, we recorded forskolin-enhanced Ca_V1.2 currents from tsA201 cells while monitoring CaM–CaN interaction in response to an increase in step-depolarization frequency from 0.067 to 1 Hz. CaM stimulation of CaN was monitored by the CaM-YFP⁵³⁵/CaN-CFP⁴⁷⁰ fluorescence emission ratio in response to CFP⁴³⁶ excitation, which provides a FRET-based readout of changes in CaM–CaN interaction. Increasing depolarization frequency in voltage-clamped, transfected tsA201 cells caused a rapid increase in CaM-YFP⁵³⁵/CaN-CFP⁴⁷⁰, which appeared to proceed as fast as, or slightly faster than, de-enhancement of current (Fig. 1A). When intracellular EGTA was replaced with BAPTA, neither the change in YFP⁵³⁵/CFP⁴⁷⁰ ratio nor de-enhancement was observed upon increase in depolarization frequency (Fig. 1B). Differential sensitivity to BAPTA versus EGTA indicates that the change in YFP⁵³⁵/CFP⁴⁷⁰ occurs in response to increased [Ca²⁺] in the channel's nanoenvironment. These findings, along with previous ones showing that de-enhancement was fully prevented by intracellular dialysis with a peptide that mimics the CaN autoinhibitory domain and blocks CaN (Oliveria et al., 2007), together suggest that in our system Ca²⁺/CaM activates CaN, which in turn acts to de-enhance activity of Ca_V1.2 channels.

Figure 1.

In transfected tsA201 cells, AKAP79 and the IQ motif in Ca_V1.2 sustain CaM–CaN signaling during heightened channel activity. A–D, After establishing a steady baseline for both forskolin-enhanced Ca_V1.2 Ca²⁺ current amplitude (I_Ca) and YFP⁵³⁵/CFP⁴⁷⁰ emission ratio in response to step depolarization once every 15 s (0.067 Hz), depolarization frequency was increased to 1 Hz. Cells expressing channel subunits, AKAP79, CaN-CFP, and CaM-YFP (A), internally perfused with 10 mm intracellular BAPTA in place of 10 mm EGTA (B), expressing AKAP79PIX (79ΔPIX) in place of AKAP79 (C), and expressing Ca_V1.2_IQ/EQ in place of wild-type Ca_V1.2 (D) are shown. Left, Ca²⁺ current records evoked before (0 s) and after 600 s of 1 Hz depolarizations. Right, Time course of response to increased depolarization frequency for both Ca²⁺ current (black) and YFP⁵³⁵/CFP⁴⁷⁰ (gold). Ca²⁺ current was normalized to the value obtained at 0 s; when necessary, time courses were corrected for the rate of channel rundown at 0.067 Hz. ΔYFP⁵³⁵/CFP⁴⁷⁰ indicates the change in the emission ratio relative to the average value at 0.067 Hz. The white envelope against the shaded background represents SEM. E, Left, Micrographs showing subcellular localization of CaN-CFP, CaM-YFP, and corrected, sensitized FRET gated to donor intensity [FRET^C (CFP); pseudocolor]. Right, Pseudocolor images representing FRET donor intensity before (CFP_pre) and after (CFP_post) acceptor photobleaching. F, Effective FRET efficiency calculated from both 3F and also YFP-acceptor PB measurements in some conditions. Ca_V1.2, β_2b, α₂δ, and AKAP79 were coexpressed with fluorescent CaN-CFP and CaM-YFP, except when Ca_V1.2 was omitted (−Ca_V1.2). In the BAPTA-AM experiments, cells were bathed in 100 μm BAPTA-AM for 60 min. n > 10 for all conditions. *p < 0.001 (black) vs 3F control; *p < 0.001 (gray) vs PB control (by both ANOVA and Bonferroni's test). G, Left, Micrographs showing subcellular localization of CaN-CFP, CaM-YFP, and corrected, sensitized FRET gated to donor intensity [FRET^C (CFP); pseudocolor] in cells coexpressing unlabeled Ca_V1.2, β_2b, α₂δ, and AKAP79. Ca_V1.2_IQ/EQ and 79ΔPIX were expressed in place of wild-type Ca_V1.2 and AKAP79 as indicated. Right, Three-filter-set effective FRET efficiency. n > 10 for all conditions. **p < 0.001 vs control (by both ANOVA and Bonferroni's test).

Unlike the strong CaN activation signal we observed in cells expressing Ca_V1.2 channels with wild-type AKAP79, CaN activation was defective when the AKAP79–CaN interaction was disrupted via deletion of the CaN-binding AKAP79 PxIxIT-like motif (79ΔPIX). In cells expressing 79ΔPIX, the extent of apparent CaM–CaN interaction declined subsequent to the increase in depolarization frequency (Fig. 1C), revealing a detectable amount of CaM–CaN association even when the channel was only activated at low frequency. Currents recorded from channels lacking scaffolded CaN retained a residual degree of de-enhancement, as though they were weakly regulated by the phosphatase even in the absence of AKAP79 anchoring.

When the Ca_V1.2 IQ motif-CaM interaction was disrupted by the well-characterized IQ-to-EQ mutation (IQ/EQ) (Zühlke et al., 1999; Erickson et al., 2003; Kim et al., 2004), the strength of the CaM–CaN interaction signal also declined as depolarization frequency was increased (Fig. 1D). As for cells expressing 79ΔPIX, this decline reveals detectable CaM–CaN association near Ca_V1.2_IQ/EQ channels during low-frequency channel activation. Intriguingly, for channels containing the IQ/EQ mutation, increased stimulation frequency relieved channel inhibition that was present during the preceding period of low-frequency channel activity, as seen by a transient increase in current amplitude. The effects of the IQ/EQ and 79ΔPIX mutations indicate that IQ–CaM and AKAP79–CaN interactions are both necessary to sustain effective CaN-mediated opposition to PKA enhancement of channel activity.

CaM and CaN preassemble near Ca_V1.2 at resting Ca²⁺

Because CaN was able to partially suppress Ca_V1.2 activity during high-frequency depolarization in 79ΔPIX cells (Fig. 1C), we tested whether CaM and CaN are able to associate near the channel independent of AKAP79. To do this, we imaged FRET between CaM-YFP and CaN-CFP using two independent methods of measuring effective FRET efficiency: 3F and PB. When CaM-YFP and CaN-CFP were coexpressed with channel subunits and AKAP79, we observed FRET between CaM and CaN in live cells that was strongest near the plasma membrane (Fig. 1E,F). In contrast, FRET was absent when the pore-forming Ca_V1.2 channel subunit was not expressed or when cells were loaded with BAPTA-AM. Since the Ca_V1.2 C terminus contains both the IQ motif and a modified leucine zipper motif that helps support Ca_V1.2–AKAP79 interaction, one possible explanation for these observations is that the channel C terminus promotes CaM–CaN FRET between IQ-docked CaM and AKAP79-scaffolded CaN. However, CaN–CaM FRET persisted when FRET measurements were performed in cells expressing either Ca_V1.2_IQ/EQ or 79ΔPIX (Fig. 1G), indicating that in tsA201 cells, neither the CaN-binding site on AKAP79 nor the IQ motif is necessary for CaM–CaN preassemble within the Ca_V1.2 complex at resting [Ca²⁺]_in.

Calcineurin phosphatase activity triggers CDI of Ca_V1.2 channels

Given these results and the established role of CaM binding to the IQ motif in Ca_V1.2 CDI, we revisited the possibility that CaN might contribute to CDI. During 500-ms-step depolarization, AKAP79-associated Ca_V1.2 channels exhibited strong ion-dependent inactivation (measured as 1/τ_fast) with Ca²⁺ as the current carrier, but much less inactivation when current was carried by Ba²⁺ (Fig. 2A). CDI was most evident when Ca²⁺ influx was greatest (Brehm and Eckert, 1978; Lee et al., 1985): depolarization that evoked smaller Ca²⁺ current elicited little CDI, while depolarization that evoked maximal Ca²⁺ current (peak of the current–voltage relationship) elicited maximum CDI.

Figure 2.

Localized calcineurin phosphatase activity triggers CDI of Ca_V1.2 channels heterologously expressed in tsA201 cells. A, Top, Ca²⁺ (red) and Ba²⁺ (black) currents recorded from the same cell by a 500 ms depolarization from −80 to +10 mV. Ca²⁺ and Ba²⁺ currents are shown scaled to the same peak inward current amplitude to facilitate comparison of inactivation rate. In addition to the three channel subunits, AKAP79 was overexpressed either alone (79) or with CaN_H151A. Middle, Inactivation rates for Ca²⁺ current (red bars) and Ba²⁺ current (black bars). Number of individual cells recorded (n) is marked on the bars. Bottom, Ca²⁺ current inactivation rates over the range of test potentials that elicited inward current in cells overexpressing AKAP79 alone or AKAP79 + CaN_H151A. **p < 0.001 (Student's t test). B, As in A, except that endogenous AKAP79 was suppressed (79RNAi), either alone or in combination with CaN_H151A overexpression. *p < 0.05 (Student's t test). C, Ca_V1.2 Ca²⁺ (red) and Ba²⁺ (black) current records and inactivation rates with 10 mm intracellular BAPTA in place of 10 mm EGTA, or with the CaN AIP (100 μm) perfused into the cell via the patch pipette. Data for AKAP79-overexpressing cells, shown in gray, are reproduced from A for comparison. **p < 0.001 (ANOVA and Bonferroni's test). Rates of inactivation of Ca²⁺ current over the range of test potentials that elicited inward current in AKAP79-expressing cells that were exposed to the CaN AIP. Dotted red line reproduces from A the voltage dependence of Ca²⁺ current inactivation measured from control AKAP79-expressing cells without the CaN AIP. Ca²⁺ current densities were as follows (in pA/pF): 79, −37.1 ± 6.3; 79 + CaN_H151A, −33.8 ± 7.4; 79RNAi, −35.0 ± 13; 79RNAi + CaN_H151A, −27.4 ± 6.4; 79 + BAPTA, −31.1 ± 7.1; 79 + CaN AIP, −32.8 ± 8.6. No forskolin was used in the experiments.

CDI was significantly slowed when a catalytically inactive mutant of CaN (CaN_H151A) (Mondragon et al., 1997) was transfected along with AKAP79 and the Ca_V1.2 subunits. Suppression of AKAP79 expression by a well-characterized RNAi construct (Hoshi et al., 2005) also slowed CDI, but this was less effective than CaN_H151A (Fig. 2B). Combining CaN_H151A overexpression with RNAi suppression of AKAP79, thus disrupting both CaN activity and anchoring, prevented CDI (Fig. 2B). These results suggest that when it could be scaffolded to channels by AKAP79, endogenous CaN remained partially effective in regulating channel activity despite competing against overexpressed CaN_H151A. CDI was also reduced by an alternative means of inhibiting CaN: in cells transfected with AKAP79 and Ca_V1.2 subunits, intracellular perfusion with the CaN autoinhibitory peptide significantly depressed CDI (Fig. 2C). CDI was suppressed, albeit incompletely, by the Ca²⁺ chelator BAPTA (Fig. 2C), indicating that Ca²⁺ entry activates CDI and that the Ca²⁺ sensor that initiates CDI must be located close to the point of Ca²⁺ entry. Comparison of the full suppression by BAPTA of CaN-mediated channel de-enhancement during high-frequency stimulation (Fig. 1D) with the partial slowing of CDI during 500 ms depolarizations (Fig. 2C, middle) suggests that these two means of measuring CDI reveal different aspects of a single underlying mechanism. In either case, our results indicate that AKAP-anchored CaN, like IQ-bound Ca²⁺/CaM, signals within the channel nanodomain.

Ba²⁺ is a weak agonist for CDI

Inactivation of Ba²⁺ current carried by Ca_V1.2 is typically taken to represent voltage-dependent inactivation. But compared to inactivation of Ba²⁺ currents, Na⁺ currents through Ca_V1.2 channels exhibited much slower inactivation rates across the entire range of depolarizing pulses that evoked inward current (Fig. 3). Hence, it seems likely that Ba²⁺ acts as a weak agonist for CDI. This idea is supported by the observation that CaN_H151A slowed inactivation of Ba²⁺ current (Fig. 2A,B) but not of Na⁺ current (Fig. 3A).

Figure 3.

Ba²⁺ acts as a weak agonist for CDI of Ca_V1.2 channels in tsA201 cells. A, Top, Recordings of Na⁺ current (green) evoked by 500 ms depolarization from −80 to −10 mV in cells in which AKAP79 was overexpressed either alone (79) or with CaN_H151A. Control Ca_V1.2 Ca²⁺ (light gray) and Ba²⁺ (dark gray) current records reproduced from Figure 2A. Ca²⁺, Ba²⁺, and Na⁺ currents are shown scaled to the same peak inward current amplitude. Bottom, Inactivation rates measured for Ca²⁺ (light gray bar), Ba²⁺ (dark gray bar), and Na⁺ (green bars) currents carried by Ca_V1.2 channels in cells overexpressing AKAP79 (79) or AKAP79 and CaN_H151A. Ca²⁺ and Ba²⁺ data reproduced from Figure 2A. B, Families of Ca²⁺, Ba²⁺, and Na⁺ currents through Ca_V1.2 channels, recorded over a range of test potentials that elicited inward current in cells overexpressing AKAP79. C, For Ca²⁺, Ba²⁺, and Na⁺ currents carried by Ca_V1.2 channels, inactivation rates obtained over a range of test potentials applied to cells overexpressing AKAP79. No forskolin was used in the experiments.

AKAP79-targeted PKA modulates CDI

AKAP79/150-anchored PKA and CaN act as opponents in regulating L channel opening (Oliveria et al., 2007), predicting that active PKA might counter CaN-dependent CDI. Forskolin stimulation of PKA did not affect Ca_V1.2 inactivation (Fig. 4), perhaps because AKAP79 coanchoring of CaN allows the catalytically faster phosphatase to overcome PKA activity in the nanoenvironment of each channel (Oliveria et al., 2007). At a basal level of PKA activity, CDI was slightly slowed by expression of the CaN-anchoring-defective 79ΔPIX mutant, but stimulating PKA activity with forskolin in cells expressing 79ΔPIX slowed Ca²⁺ current inactivation nearly to the rate for Ba²⁺ (Fig. 4). Therefore it seems that in the presence of anchored and active PKA, CaN requires AKAP79/150 to overcome current enhancement by PKA and effectuate CDI. If CDI does indeed represent reversal of PKA phosphorylation of the channel by CaN, then CDI should also be sensitive to disruption of PKA enhancement. In accord with this idea, CDI was significantly slower when PKA activity was inhibited with PKI. However, inhibition of PKA did not eliminate CDI, as though PKA was not solely responsible for phosphorylation of the channel complex.

Figure 4.

AKAP79-anchored PKA opposes CDI of Ca_V1.2 channels in tsA201 cells. Top, Superimposed Ca²⁺ (red) and Ba²⁺ (black) currents carried through Ca_V1.2 channels. Bottom, Inactivation rates for Ca²⁺ and Ba²⁺ currents. Currents were recorded during forskolin (FSK; 5 μm) activation of PKA, with overexpressed mutant AKAP79 lacking the CaN binding site (79ΔPIX) with no forskolin present, with a combination of 79ΔPIX and forskolin, or during PKI (5 μm) inhibition of PKA activity (no forskolin). *p < 0.05 (Student's t test); **p < 0.001 vs AKAP79 + wild-type Ca_V1.2 (by both ANOVA and Bonferroni's test). Control data for AKAP79-overexpressing cells, shown in gray, are reproduced from Figure 2A for comparison. Ca²⁺ current densities were as follows (in pA/pF): FSK, −25.7 ± 7.2; 79ΔPIX, −28.5 ± 4.7; 79ΔPIX + FSK, −30.1 ± 3.9; PKI, −25.5 ± 6.5.

AKAP79-anchored calcineurin mediates CDI of Ca_V1.2 channels by opposing PKA

As an alternative means of studying PKA opposition to CaN, we studied the dynamics of development of this process in individual cells. Following patch rupture and initiation of whole-cell recording, internal perfusion with forskolin progressively slowed CDI in tsA201 cells transfected with 79ΔPIX, but not in cells transfected with wild-type AKAP79 (Fig. 5A). During intracellular perfusion of wild-type AKAP79 cells with forskolin, bath application of the CaN antagonists CsA and FK506, which inhibit CaN through distinct cyclophilin and FKBP pathways, was able to reduce CDI (Fig. 5B). The acute effect on CDI of this combination of CsA and FK506 was reversed upon washout. Together, the results support the idea that PKA, stimulated by elevation of cAMP with forskolin, can oppose CaN-mediated inactivation of Ca²⁺ current.

CaN contributes to CDI of L-type Ca²⁺ channels in cultured hippocampal neurons

As powerful as reconstitution in heterologous expression systems is for investigating mechanisms of channel regulation, models ultimately need to be tested in the biological system of interest, in this case, hippocampal neurons. We found previously that AKAP-anchored CaN powerfully regulates L-type Ca²⁺ currents in hippocampal neurons (Oliveria et al., 2007), so we next examined whether CaN might do so via CDI.

Pharmacologically isolated L-type Ca²⁺ current recorded from short-term cultured rat hippocampal neurons show robust CDI characterized by rapid inactivation in Ca²⁺ compared to Ba²⁺. Consistent with a role for AKAP79/150 in CDI, inactivation was much slower when expression of the rodent ortholog of human AKAP79, AKAP150, was suppressed with a previously characterized RNAi construct (Hoshi et al., 2005; Oliveria et al., 2007) (Fig. 6A). Suppression of CDI by 150RNAi was fully rescued by coexpression of AKAP79 in rat neurons, but no rescue of CDI was observed in neurons cotransfected with 79ΔPIX (Fig. 6B). In addition, the strong CDI observed in wild-type mouse neurons was completely absent in neurons cultured from 150ΔPIX knock-in mice (Sanderson et al., 2012) (Fig. 6C).

Figure 6.

AKAP79/150-anchored CaN is required for CDI of L-type Ca²⁺ channels in dissociated, short-term cultured hippocampal neurons. A, For rat neurons, pharmacologically isolated L channel Ca²⁺ (red) and Ba²⁺ (black) currents recorded from the same neuron, evoked by 500 ms depolarization from −60 to 0 mV. Peak Ca²⁺ current is normalized to peak Ba²⁺ current. Control neurons were transfected with GFP and the empty pSilencer vector. 150RNAi neurons were transfected with a short hairpin construct to suppress endogenous AKAP150 expression. Images of short-term cultured hippocampal neurons are a differential interference contrast (DIC) image of an untransfected hippocampal neuron (left) and a fluorescence image of hippocampal neuron expressing 150RNAi along with cytoplasmic GFP to mark transfected neurons (150RNAi; right). B, Cultured rat neurons were transfected with 150RNAi and either AKAP79 [79 wild-type (wt); rescue] or AKAP79ΔPIX (79ΔPIX; failed rescue). Images of cultured neurons are a fluorescence image of a hippocampal neuron expressing 150RNAi and AKAP79 (left) and a fluorescence image of a rat cultured hippocampal neuron expressing 150RNAi and either AKAP79ΔPIX or AKAP150ΔPIX. C, For mouse short-term cultured hippocampal neurons, pharmacologically isolated L channel Ca²⁺ (red) and Ba²⁺ (black) currents were recorded from the same neuron. DIC images of neurons from a wild-type mouse and a transgenic AKAP150ΔPIX mouse are shown. ***p < 0.0001 vs control (Student's t test). For inactivation rates, n values are indicated on the bars. Ca²⁺ current densities were as follows (in pA/pF): control, −32.2 ± 2.3; 150RNAi, −32.8 ± 2.0; 79 wt + 150RNAi, −29.2 ± 2.5; 79ΔPIX + 150RNAi, −28.7 ± 1.9; wt Mouse, −30.0 ± 2.1; 150ΔPIX Mouse, −31.9 ± 2.5.

To confirm a requirement for AKAP-anchored CaN in neuronal CDI, we used an acute means to antagonize CaN anchoring. Internal perfusion with a peptide that mimics the AKAP150 anchoring sequence for CaN, VIVIT (Oliveria et al., 2007; Li et al., 2012), competitively and dose-dependently suppressed CDI of Ca²⁺ current, leaving Ba²⁺ current unaltered (Fig. 7A). At the highest concentration tested (100 μm in the whole-cell pipette), VIVIT reduced the rate of inactivation in Ca²⁺ to that measured for Ba²⁺ current. To establish a requirement for CaN phosphatase activity in CDI of neuronal L channels, we used two different approaches: transfection of cultured neurons with CaN_H151A and internal perfusion with CsA. Both experimental manipulations strongly reduced the rate of inactivation of Ca²⁺ current carried by neuronal L channels (Fig. 7B).

Figure 7.

Acute inhibition of CaN anchoring and activity suppresses CDI in hippocampal neurons. A, VIVIT peptide, a competitive inhibitor of CaN anchoring on AKAP150, dose-dependently suppresses CDI in dissociated, short-term cultured hippocampal neurons. VIVIT was perfused into neurons via the whole-cell patch pipette. Control neurons were recorded with 0.1% DMSO in the patch pipette, which was the concentration of DMSO in the VIVIT experiments. B, Calcineurin phosphatase activity is required for Ca²⁺-dependent inactivation of L-type channels in dissociated, short-term cultured hippocampal neurons. Top, Comparison of scaled currents evoked by 500 ms depolarization in Ca²⁺ versus Ba²⁺ for control, neurons transfected with CaN_H151A, and neurons internally perfused with cyclosporin A (5 μm). Bottom, Corresponding inactivation rates (1/τ_fast). Control neurons were perfused with a whole-cell patch pipette solution containing 0.1% DMSO, which is the same concentration present in the cyclosporin A experiments. Results for CaN_H151A were compared to the control neurons in Figure 6 (no DMSO), while results with CsA were compared to the DMSO control. *p = 0.01; **p < 0.005; ***p < 0.0001 vs control neurons (Student's t test). Ca²⁺ current densities were as follows (in pA/pF): control, −31.3 ± 3.5; 10 μm VIVIT, −31.9 ± 5.0; 25 μm VIVIT, −31.5 ± 4.9; 100 μm VIVIT, −29.2 ± 4.0; CaN_H151A, −30.4 ± 2.2; and CsA, −32.6 ± 1.5.

Discussion

In neurons and a heterologous expression system that incorporates the important neuronal scaffolding molecule AKAP79/150, we found that Ca²⁺/CaM induces CDI of L channels through activation of CaN. A role for CaN in CDI is supported by our observations of significantly slowed inactivation by (1) overexpression of catalytically inactive CaN (CaN_H151A), (2) internal dialysis with the CaN autoinhibitory peptide, and (3) application of the CaN antagonist CsA. The fact that CaN_H151A had no effect on inactivation of Ca_V1.2 Na⁺ current indicates that this CaN mutant acts selectively to depress CDI. Expression of L channel CDI in neurons requires anchoring of CaN by AKAP79/150, whereas in the tsA201 expression system, CaN anchoring by AKAP79/150 is not absolutely required but is needed to support rapid, fully developed CDI. Evidence implicating CaN-anchoring by AKAP79/150 in CDI includes (1) elimination of CDI by 150RNAi knockdown and also by replacement with CaN-binding-defective AKAP79ΔPIX, (2) elimination of CDI in a 150ΔPIX knock-in mouse, and (3) competitive inhibition of CaN anchoring by VIVIT. Overall, interventions that impaired CaN action had more complete effects in neurons than in tsA201 cells, possibly reflecting more natural spatial organization and signaling stoichiometry, or yet-to-be identified regulatory factors.

Dephosphorylation signaling in CDI: permissive and direct models

Although Ca²⁺/CaM activation of CaN phosphatase activity appears to be a required step for CDI of L channels in hippocampal pyramidal neurons, our results do not exclude the possibility that conformational change of CaM docked to Ca_V1.2 can directly inactivate channel activity in some cellular contexts. An extensive body of work has highlighted the centrality of Ca²⁺/CaM in CDI (for review, see Calin-Jageman and Lee, 2008), but none of this work has taken into consideration the L channel scaffolding protein AKAP79/150, the absence of which minimizes CaN action on L channels. The logic underlying the two different experimental measures of CDI used here—cumulative inactivation during high-frequency depolarization reported by progressive reduction in Ca²⁺ current amplitude versus rate of inactivation during 500 ms depolarization—suggests a pair of alternative models for the action of dephosphorylation in CDI. In a permissive model, dephosphorylation by CaN must precede downstream steps that carry out CDI, such as conformational rearrangement of Ca²⁺/CaM on the C-terminal regulatory domains (PreIQ, IQ) of Ca_V1.2; dephosphorylation by CaN is a prerequisite for downstream steps to effectuate CDI. Alternatively, in a direct model of CaN′s role in CDI, dephosphorylation is envisioned as the terminal regulatory step in CDI, and Ca²⁺/CaM acts exclusively through CaN. In either case, the advantageous placement of CaN via anchoring on AKAP79/150 promotes CaN-dependent CDI. For highly abundant neuronal signaling molecules such as CaN (∼1% of protein in brain neurons), AKAP-based scaffolding may not only serve to localize enzymes to their targets, but via steric protection may also help shield substrates from extraneous phosphorylation signaling. It remains to be determined whether these models or related ones apply in excitable cells other than hippocampal neurons. For either model of CaN action, a dephosphorylation-based mechanism of CDI adds to the richness of signaling mechanisms that regulate L-type Ca²⁺ channels in neurons.

Ser/Thr kinases that oppose CaN-dependent CDI

In hippocampal pyramidal neurons, β₂ adrenergic receptors are known to regulate L-type Ca²⁺ current through PKA (Gray and Johnston, 1987; Davare et al., 2001; Hoogland and Saggau, 2004), and in thalamocortical relay neurons, β₂ adrenergic receptor activation slows CDI (Rankovic et al., 2011). But β₂ adrenergic receptor activity is not the only potential means of activating PKA: a major adenylyl cyclase isotype in hippocampal pyramidal neurons is the Ca²⁺-stimulable type 8, which is known to bind to AKAP79/150 (Willoughby et al., 2010). Thus, Ca²⁺ influx via L channels could locally stimulate adenylyl cyclase type 8 production of cAMP and hence PKA activity. Our results in cells expressing a CaN-anchoring-defective AKAPΔPIX construct, demonstrate that activation of PKA through forskolin stimulation of adenylyl cyclase reduced CDI, suggesting that phosphorylation by PKA opposes dephosphorylation by CaN. Inhibition of PKA by PKI also impacted CDI, but this treatment, too, slowed CDI. That both stimulation and inhibition of PKA slow CDI might be reconciled to some degree by the following: while PKA kinase activity reverses CaN phosphatase action and thus suppresses CDI, PKI inhibition of kinase activity may result in few channels remaining in a phosphorylated state and so few remain available to undergo CaN-dependent CDI. However, when compared to CaN inhibition, incomplete suppression of CDI by PKI suggests contributions from other signaling processes.

CDF of L channels is also a phosphosignaling-based process and one that overlaps CDI, as CDF can be detected as a reduction in CDI (Wu et al., 2001; Hudmon et al., 2005; Erxleben et al., 2006; Grueter et al., 2006). CDF is mediated by channel-associated Ca²⁺/CaMKII activity, which is not only congruent with a phosphatase-based mechanism for CDI, but also indicates that kinases other than PKA may be involved in CDI. Indeed, because our experiments showed that a high concentration of PKI was unable to prevent CDI entirely, as predicted if PKA were solely responsible for recovery from CaN-dependent inactivation, it seems likely that additional kinases may contribute to phosphorylation of the CaN-sensitive CDI substrate. CaMKII is one likely candidate, as it associates with Ca_V1.2 channels and enhances their activity. If CaMKII cooperates with PKA to oppose CaN action, it is intriguing to consider how Ca²⁺/CaM molecules residing in the channel's nanodomain are able to initiate signaling to CaMKII or CaN and thereby initiate either CaMKII-dependent facilitation or CaN-dependent inactivation (Saucerman and Bers, 2008).

Previous evidence for CaN-dependent CDI

Considering that an enzymatic mechanism for CDI of Ca_V channels was proposed over two decades ago (Chad and Eckert, 1986; for review, see Armstrong, 1989), why has a role for CaN in CDI fallen into disfavor? One likely reason is that channel-associated CaN activity is difficult to suppress. CsA and FK506 are the most commonly used CaN inhibitors, but CsA and FK506 were much less effective in preventing CDI when compared to the CaN AIP (Figs. 2C, 5B). This may be related to the mechanism of action of CsA and FK506, which must first bind to immunophilins and then inhibit CaN activity as drug-immunophilin complexes. Because such complexes are much larger (>18 kDa) than the CaN AIP (a 25-mer polypeptide of 2.9 kDa), it is possible that CaN may be protected from inhibition by drug-immunophilin complexes when the phosphatase is tightly associated with the channel. In some cases, the effectiveness of CaN-inhibiting drugs like CsA has also been found to be limited by immunophilin expression level (Kung and Halloran, 2000; Mitchell et al., 2002). In our experiments in tsA201 cells, CaN_H151A expression did not fully suppress CDI unless AKAP79/150 targeting of CaN was also disrupted, indicating that CaN-dependent channel inactivation is more resistant to perturbation when CaN is scaffolded to the channel. Although AKAP79/150 is expressed in every cell type where Ca_V1.2 is found and in most heterologous expression systems used to study Ca_V1.2 channels, including HEK293 and tsA201cells, the level of AKAP79/150 expression varies widely between cell types and might therefore explain partial or conflicting effects of CaN inhibitors on CDI.

Another confounding factor arises from the reliance upon Ba²⁺ to isolate CDI. But Ba²⁺ weakly supports CDI (Fig. 3), making it more challenging to work out the molecular mechanism of this process. Many divalent cations are able to activate CaM, their effectiveness as substitutes for Ca²⁺ largely depending on how closely they approximate Ca²⁺ in ionic radius (Chao et al., 1984). The difference in size between Ba²⁺ and Ca²⁺ generally makes Ba²⁺ a poor substitute for Ca²⁺, but several factors may facilitate Ba²⁺ binding to CaM within the channel complex: (1) [Ba²⁺], like [Ca²⁺], reaches high levels in the nanodomain near the channel mouth, (2) Ba²⁺ is poorly chelated by EGTA or BAPTA, and (3) Ba²⁺ is slowly sequestered by cellular transporters. Ba²⁺ has been found to activate other Ca²⁺-sensitive processes including neurotransmitter release (McMahon and Nicholls, 1993), Ca²⁺ release from internal stores (Satoh et al., 1987), contraction of skeletal (Stephenson and Thieleczek, 1986) and smooth muscle (Satoh et al., 1987), gene expression (Curran and Morgan, 1986), and cAMP production by Ca²⁺-sensitive adenylyl cyclases (Gu and Cooper, 2000).

Conclusions

In cultured hippocampal pyramidal neurons, CaN and PKA appear to work by opposing mechanisms of CDI/de-enhancement versus enhancement. Because multiple phosphorylation sites may be involved, and other kinases such as CaMKII or phosphatases such as PP2A (Xu et al., 2010), CaN versus PKA expressed as CDI versus enhancement may prove to be an oversimplification. The AKAP79/150 PxIxIT-like motif and Ca_V1.2 IQ motif play supporting roles in positioning CaN and CaM near the source of Ca²⁺, but preassembly of CaM with CaN may also help speed enzymatic CDI. The affinity of AKAP79/150 for CaN is relatively low (K_D ≈ 0.5 μm), perhaps anchoring CaN just tightly enough to target it in cells while also permitting sufficiently rapid release from the AKAP to promote regulation of nearby substrates (Li et al., 2012).