Modulation of G_q-Protein-Coupled Inositol Trisphosphate and Ca²⁺ Signaling by the Membrane Potential

Abstract

G_q-protein-coupled receptors (G_qPCRs) are widely distributed in the CNS and play fundamental roles in a variety of neuronal processes. Their activation results in phosphatidylinositol 4,5-bisphosphate (PIP₂) hydrolysis and Ca²⁺ release from intracellular stores via the phospholipase C (PLC)–inositol 1,4,5-trisphosphate (IP₃) signaling pathway. Because early G_qPCR signaling events occur at the plasma membrane of neurons, they might be influenced by changes in membrane potential. In this study, we use combined patch-clamp and imaging methods to investigate whether membrane potential changes can modulate G_qPCR signaling in neurons. Our results demonstrate that G_qPCR signaling in the human neuronal cell line SH-SY5Y and in rat cerebellar granule neurons is directly sensitive to changes in membrane potential, even in the absence of extracellular Ca²⁺. Depolarization has a bidirectional effect on G_qPCR signaling, potentiating thapsigargin-sensitive Ca²⁺ responses to muscarinic receptor activation but attenuating those mediated by bradykinin receptors. The depolarization-evoked potentiation of the muscarinic signaling is graded, bipolar, non-inactivating, and with no apparent upper limit, ruling out traditional voltage-gated ion channels as the primary voltage sensors. Flash photolysis of caged IP₃/GPIP₂ (glycerophosphoryl-myo-inositol 4,5-bisphosphate) places the voltage sensor before the level of the Ca²⁺ store, and measurements using the fluorescent bioprobe eGFP–PH_PLCδ (enhanced green fluorescent protein–pleckstrin homology domain–PLCδ) directly demonstrate that voltage affects muscarinic signaling at the level of the IP₃ production pathway. The sensitivity of G_qPCR IP₃ signaling in neurons to voltage itself may represent a fundamental mechanism by which ionotropic signals can shape metabotropic receptor activity in neurons and influence processes such as synaptic plasticity in which the detection of coincident signals is crucial.

Introduction

G-protein-coupled receptors (GPCRs) convert a wide variety of extracellular signals into intracellular messages by binding to and activating G_q/11-, G_s-, G_i/o-, or G_12/13-proteins (Neves et al., 2002). Stimulation of receptors coupled to G_q/11-proteins results in activation of phospholipase C (PLC). This leads to hydrolysis of phosphatidylinositol 4,5-bisphosphate (PIP₂) and generation of diacylglycerol and inositol 1,4,5-trisphosphate (IP₃), resulting in activation of protein kinase C (PKC) and Ca²⁺ release from intracellular stores, respectively (Berridge, 1993). G_qPCRs are widely expressed in the brain and implicated in a variety of neuronal processes, such as synaptic plasticity, regulation of gene expression, and neuronal excitability (Berridge, 1998; Augustine et al., 2003).

Although a considerable amount is known about the effect GPCR signaling can have on membrane excitability, the converse, how the membrane potential affects GPCR signaling, has hardly been examined. Interestingly, a synergy between neuronal electrical activity and G_qPCR-evoked IP₃ and Ca²⁺ signaling has been demonstrated in a number of systems, including hippocampal neurons in situ (Nakamura et al., 1999, 2000, 2002) and in vitro (Nash et al., 2004; Young et al., 2005), neocortical neurons in situ (Larkum et al., 2003), and cerebellar Purkinje cells in situ (Okubo et al., 2001, 2004). Depolarization was suggested to affect G_qPCR signaling indirectly, by evoking Ca²⁺ influx across the plasma membrane through ion channels. This can (1) promote store filling and thus influence the magnitude of the subsequent response to G_qPCR activation (Irving and Collingridge, 1998; Rae et al., 2000), (2) stimulate Ca²⁺ release from intracellular stores by shifting the sensitivity of the IP₃ receptor to IP₃ (Berridge, 1998; Nakamura et al., 1999), and (3) stimulate the IP₃ signaling cascade by positive modulation of receptor-stimulated PLCβ (Eberhard and Holz, 1988; Hashimotodani et al., 2005). However, because G_qPCRs and some of their effectors are located in the plasma membrane of cells with dynamic membrane potentials (e.g., neurons), changes in membrane potential may also influence signaling directly. Indeed, in some non-neuronal cells, G_qPCR signaling appears to be sensitive to voltage independently of Ca²⁺ influx. For example, voltage exerts a graded, bipolar effect on the muscarinic receptor-evoked Ca²⁺ release in guinea pig coronary artery smooth muscle cells (Ganitkevich and Isenberg, 1993) and on the purinergic P2Y receptor-evoked Ca²⁺ release in rat megakaryocytes (Mahaut-Smith et al., 1999; Mason et al., 2000; Mason and Mahaut-Smith, 2001; Martinez-Pinna et al., 2004, 2005).

The aim of this study was to investigate whether G_qPCR-evoked IP₃ and Ca²⁺ signaling in neurons can be directly modulated by changes in membrane potential. Using combined patch-clamp and imaging methods, our data demonstrate that depolarization per se potentiates muscarinic signaling at the level of the IP₃ production pathway. The sensitivity of G_qPCR signaling to voltage itself may represent a fundamental mechanism by which ionotropic signals can shape metabotropic receptor activity in neurons and influence processes such as synaptic plasticity in which coincidence detection of signals is crucial.

Materials and Methods

Cell culture.

To prepare primary cerebellar granule cell cultures, Lister Hooded rats (postnatal days 5–7) were decapitated, the cerebellar hemispheres were discarded, and the cerebellar vermis was diced and incubated at 37°C for 15 min in trypsin solution [PBS containing 0.25 mg/ml trypsin, 1.5 mg/ml bovine serum albumin (BSA) fraction V, 7 mm glucose, and 0.75 mm MgSO₄]. The trypsin digestion was stopped by addition of trypsin inhibitor solution (PBS containing 0.4 mg/ml trypsin inhibitor from soybean, 3 mg/ml BSA fraction V, 14 mm glucose, 7.5 mm MgSO₄, and 40 U of DNase 1). The digested tissue was centrifuged at 250 × g for 3 min, and the pellet was resuspended in disaggregation buffer (containing PBS, 0.4 mg/ml trypsin inhibitor from soybean, 3 mg/ml BSA fraction V, 14 mm glucose, 7.5 mm MgSO₄, and 120 U of DNase 1) and triturated 10–12 times using a fire-polished Pasteur pipette. Larger tissue clumps were allowed to settle for 1–2 min, and the remaining cells in suspension were transferred to a sterile tube and centrifuged at 250 × g for 3 min. The cell pellet was then resuspended in serum-free minimal essential medium (MEM), and the cells were counted and centrifuged at 250 × g for 2 min before DNA plasmid transfection (see below). The cells were then plated on poly-d-lysine-coated coverslips in MEM supplemented with 10% fetal calf serum, 2 mm l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, 25 mm KCl, and 33 mm glucose. After 24 h, three-quarters of the media was exchanged with media containing 10 μm cytosine arabinoside, and, 2 d later, half of the media was exchanged again. The cultured cerebellar granule cells were used for experiments at 4–7 d in vitro. SH-SY5Y human neuroblastoma cells were grown in MEM supplemented with 10% fetal calf serum, 2 mm l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. Human embryonic kidney 293 cells stably expressing the cloned human muscarinic m₃ receptor (Tovey and Willars, 2004) (HEK–m₃) were grown in MEM α medium supplemented with 10% fetal calf serum, 100 U/ml penicillin, 100 μg/ml streptomycin, and fungizone. Chinese hamster ovary cells stably expressing the cloned human m₃ receptor (CHO–m₃) were grown in MEM α medium supplemented with 10% newborn calf serum, 100 U/ml penicillin, 100 μg/ml streptomycin, and fungizone. For experiments, these cells were plated onto 13 mm glass coverslips and used within 1–3 d. All cells were kept at 37°C in humidified air containing 5% CO₂.

Solutions.

Before recordings, the cells were usually perfused with a solution containing the following (in mm): 134 NaCl, 4.7 KCl, 1.2 MgCl₂, 1.3 CaCl₂, 4.2 NaHCO₃, 1.2 KH₂PO₄, 10 HEPES, and 11.7 glucose. The pH was adjusted to 7.4 with NaOH and the osmolarity to ∼315 mOsm with sucrose. All recordings were performed in solution lacking Ca²⁺, and the cells were perfused for at least 8 min with Ca²⁺-free solution before recording started. For experiments using sodium-free extracellular solution, NaHCO₃ was replaced by an equimolar concentration of KHCO₃, NaCl was replaced by N-methyl-d-glutamine (NMDG) chloride, and the pH was adjusted to 7.4 with HCl. The standard intracellular solution contained the following (in mm): 140 KCl, 2 MgCl₂ 10 HEPES, 0.1 EGTA, 2 Mg-ATP, and 0.05 Na₂-GTP. The osmolarity was ∼300 mOsm, and the pH was adjusted to 7.3 with KOH. For experiments using potassium-free intracellular solution, KCl was replaced with NMDG-Cl and the pH was adjusted to 7.3 with HCl. All experiments were performed at physiological temperature (35–37°C).

Chemicals.

Fura-2 and fluo-4 (pentapotassium salts) were obtained from Invitrogen (Paisley, UK); 1-(2-nitro-phenyl)ethyl caged-IP₃ (NPE–IP₃), NPE caged-glycerophosphoryl-myo-inositol 4,5-bisphosphate (NPE–GPIP₂), forskolin, 3-isobutyl-1-methylxanthine (IBMX), and thapsigargin were from Calbiochem (Nottingham, UK); PBS and MEM were from Invitrogen; and bradykinin, oxotremorine-M, ryanodine, caffeine, pertussis toxin (PTX), nifedipine, and all other chemicals were from Sigma (Poole, UK).

Electrophysiology.

Whole-cell or amphotericin-B perforated voltage-clamp recordings were made from cells using thick-walled glass pipettes (GC150F-7.5; Clark Electromedical, Reading, UK) with either an Axopatch 200B amplifier (Molecular Devices, Palo City, CA), low-pass filtered at 5 kHz (eight-pole Bessel filter), and sampled at 20 kHz using a 1320 Digidata and pClamp 8.2 software (Molecular Devices) or an EPC-10 amplifier using PatchMaster software (HEKA Elektronik, Lambrecht/Pfalz, Germany). The open tip resistance of the patch pipettes used in whole-cell and perforated-cell experiments was 4–6 and 8–12 MΩ, respectively, and the whole-cell and perforated-cell access resistance was usually 10–15 and 30–40 MΩ, respectively, and compensated for by >70%. Unless indicated otherwise, the membrane potential was clamped at −70 mV.

Ca²⁺ imaging.

Cells were visualized with infrared differential interference contrast (DIC) optics on an upright Nikon (Tokyo, Japan) E600FN microscope with a 60×, 1.0 numerical aperture water immersion fluor objective. To image intracellular Ca²⁺, 50 μm fura-2 (pentapotassium salt) was included in the patch pipette. Fura-2 was excited at 350 and 380 nm (100 ms light exposure per wavelength per frame) with a Polychrome II monochromator (xenon lamp based; T.I.L.L. Photonics, Martinsried, Germany). Emitted light was separated by a 400 nm dichroic mirror and filtered with a 420 nm long-pass emission filter. The fluorescent signals were acquired every 500 ms with a PentaMAX cooled charge-coupled device camera via a Gen IV image intensifier (Princeton Instruments, Trenton, NJ) and analyzed with MetaFluor software (Universal Imaging, West Chester, PA). All data were background subtracted using the 350 and 380 nm fluorescence values of a neighboring, unpatched cell as background. The fura-2 fluorescence ratio R = F_350nm/F_380nm (where F_350nm and F_380nm are the measured fluorescence intensities when fura-2 is excited at 350 and 380 nm, respectively) was converted into [Ca²⁺] levels using the following equation (Grynkiewicz et al., 1985):

The K_D of fura-2 in intracellular patch solution was determined using a calibration kit from Invitrogen. R_min and R_max are the ratios, and F^380nm_min and F^380nm_max are the background-subtracted fluorescence intensities at 0 (minimum) and saturating (maximum; 39 μm) free Ca²⁺, respectively, as measured in intracellular patch solution using the Invitrogen calibration kit.

Flash photolysis.

In some experiments, 100 μm NPE–IP₃ or 200 μm NPE–GPIP₂ [the slowly metabolized, less potent analog of IP₃ (Bird et al., 1992)] was included in the whole-cell patch pipette. To photolyze the caged compounds, the cell was exposed to 360 nm light for 100 ms. In these experiments, fura-2 could not be used as Ca²⁺ indicator, because the fura-2 excitation light (350/380 nm) photolyzes the caged compounds. Instead, fluo-4 (50 μm) was used as Ca²⁺ indicator during these experiments. Fluo-4 was excited at 475 nm (10 ms light exposure per frame), a wavelength that does not interfere with the NPE cage, and emitted light was separated by a 505 nm dichroic mirror, filtered with a 520 nm long-pass emission filter, and sampled at 500 ms intervals. Control experiments omitting NPE–IP₃/GPIP₂ from the pipette showed that the 360 nm light flash did not interfere with the fluo-4 fluorescence signal. The fluo-4 fluorescence signal was background subtracted, and graphs were presented as F/F_o ratios, where F is the background-subtracted fluorescence intensity, and F_o is the background-subtracted fluorescence intensity at the start of the recording.

Plasmid transfection and enhanced green fluorescent protein–pleckstrin homology domain–PLCδ1 imaging.

Cerebellar granule neurons and SH-SY5Y cells were transfected with enhanced green fluorescent protein (eGFP)–pleckstrin homology domain–PLCδ1 (PH_PLCδ1) plasmid DNA (Stauffer et al., 1998) by electroporation (Amaxa, Cologne, Germany), using the Rat Neuron Nucleofector kit and program G-13, according to the instructions of the manufacturer, and the cells were used for experiments 2–7 d after the electroporation. Alternatively, SH-SY5Y cells were transfected with 0.5 μg/ml eGFP–PH_PLCδ1 plasmid DNA using 1.5 μl/ml Lipofectamine2000 (Invitrogen), according to the instructions of the manufacturer, the transfection medium was replaced after 4 h, and cultures were used for experiments 1 and 2 d later. eGFP was excited at 488 nm (100 ms light exposure per frame) with a Cairn Optoscan monochromator (xenon lamp based; Cairn Research, Faversham, UK), emitted light was separated by a 505 nm dichroic mirror and filtered with a 520 nm long-pass emission filter, and the fluorescent signals were acquired every 500 ms with a Cascade 512B cooled CCD camera (Photometrics, Tucson, AZ). The eGFP fluorescence signal was background subtracted, and graphs were presented as F/F_o ratios. For illustration purposes only, the images in Figures 9Aa–Ad and 10, Ab and Ac, were deconvolved using a no-neighbor algorithm implemented in MetaMorph (Universal Imaging).

Figure 9.

Muscarinic IP₃ production in SH-SY5Y cells is potentiated by depolarization. A, eGFP–PH_PLCδ fluorescence of a whole-cell patch-clamped SH-SY5Y cell. Application of 10 μm oxotremorine-M leads to the production of IP₃ and the translocation of the GFP-tagged biosensor into the cytoplasm, which can be seen as an increase in cytoplasmic fluorescence. a–d, The specimen images show the biosensor fluorescence before (a), at the peak (b), and after muscarinic receptor stimulation (d) at −70 mV and during the oxotremorine-M response at +50 mV (c). The image in e is a difference image created by subtracting image a from image b and highlights the increase of cytoplasmic biosensor fluorescence during the peak of the oxotremorine-M response. The dotted line shows the position of the cell. B, The cytoplasmic fluorescence signal (F/F_o) was measured in the whole-cell patch-clamped SH-SY5Y cell shown in A for the duration of the experiment. The arrows indicate the times at which the images displayed in A were chosen. Oxotremorine-M at 10 μm (black bar) evoked an increase in cytoplasmic eGFP–PH_PLCδ fluorescence that was potentiated by depolarization. In contrast, depolarization before the muscarinic response did not affect the fluorescence signal. C, Mean ± SEM of the eGFP–PH_PLCδ fluorescence response (F/F_o) to 10 μm oxotremorine-M at the time of depolarization, normalized to the fluorescence response without depolarization. Depolarization (10 s from −70 to +50 mV) significantly potentiated the fluorescence response in all cells tested (n = 9; p < 0.01).

Figure 10.

Depolarization potentiates the muscarinic IP₃ production in cerebellar granule neurons. A, DIC image of a perforated patch-clamped cerebellar granule neuron and its eGFP–PH_PLCδ fluorescence before (a) and at the peak (b) of the oxotremorine-M-evoked muscarinic response at −70 mV. Note that no fluorescence increases were observed in the large portion of the cytoplasm that is occupied by the nucleus. The image in c is a difference image created by subtracting image a from image b and highlights the increase of cytoplasmic biosensor fluorescence during the peak of the oxotremorine-M response. The dotted lines show the position of the cells. B, The cytoplasmic fluorescence signal (F/F_o) was measured in the perforated patch-clamped cerebellar granule neuron shown in A for the duration of the experiment. The arrows indicate the times at which the images displayed in A were chosen. Oxotremorine-M at 10 μm (black bar) evoked an increase in cytoplasmic eGFP–PH_PLCδ fluorescence that was potentiated by depolarization, demonstrating that muscarinic IP₃ production in these neurons is voltage sensitive. In contrast, depolarization before the muscarinic response did not affect the fluorescence signal. C, Mean ± SEM of the eGFP–PH_PLCδ fluorescence response (F/F_o) to 10 μm oxotremorine-M at the time of depolarization, normalized to the fluorescence response without depolarization. Depolarization (10 s from −70 to +50 mV) significantly potentiated the fluorescence response in all cells tested (n = 5; p < 0.01).

Statistics.

All data are given as mean ± SEM, and statistical analysis was performed using paired or unpaired two-tailed t tests as appropriate and regarded as significant if p < 0.05.

Results

The M₃ receptor Ca²⁺ signal is potentiated by depolarization in SH-SY5Y cells

Cultured SH-SY5Y cells (Fig. 1A), a human neuronal cell line closely resembling sympathetic ganglion cells, endogenously express G_q-coupled muscarinic acetylcholine receptors, mainly of the M₃ subtype (Lambert et al., 1989). Activation of these receptors with a muscarinic agonist, such as oxotremorine-M, leads, via the activation of PLC and the generation of IP₃, to Ca²⁺ release from intracellular stores. In Ca²⁺-free solution, to prevent Ca²⁺ influx and record purely store Ca²⁺ release, this was measured in whole-cell patch-clamped SH-SY5Y cells as a transient Ca²⁺ increase from a basal Ca²⁺ concentration of 73 ± 6 to 313 ± 26 nm at −70 mV (Fig. 1B) (n = 39; 10 μm oxotremorine-M) and 76 ± 12 to 232 ± 13 nm at −70 mV (n = 5; 1 μm oxotremorine-M). Because the experiments were performed in the absence of extracellular Ca²⁺, intracellular Ca²⁺ stores will gradually run down without the possibility of refilling. In accordance, the magnitude of the Ca²⁺ transient varied between cells, depending greatly on the amount of releasable Ca²⁺ left in the stores. Depolarization during the oxotremorine-M-evoked Ca²⁺ transient potentiated the Ca²⁺ signal for the duration of the depolarization, an effect that could be repeated several times during the muscarinic response (Fig. 1C–F) and that we term “depolarization-evoked potentiation” (DEP). Depolarization of whole-cell patch-clamped SH-SY5Y cells in Ca²⁺-free solution in the absence of muscarinic agonist (control) did not affect the intracellular Ca²⁺ concentration (n = 130), confirming that there is indeed no voltage-dependent Ca²⁺ influx under these conditions.

Figure 1.

Muscarinic Ca²⁺ signaling in the human neuronal cell line SH-SY5Y is potentiated by depolarization. A, DIC image of cultured SH-SY5Y cells. These cells were whole-cell patch clamped and intracellular Ca²⁺ imaged using the Ca²⁺ indicator fura-2 applied to the cytoplasm via the patch pipette. B, Mean ± SEM of the basal intracellular Ca²⁺ levels and the muscarinic receptor-mediated Ca²⁺ transient evoked by 10 μm oxotremorine-M in whole-cell patch-clamped SH-SY5Y cells (n = 39). C, D, Ca²⁺ responses of whole-cell patch-clamped SH-SY5Y cells to a maximal (10 μm; C) and submaximal (1 μm; D) concentration of oxotremorine-M (black bar). Depolarization (10 s from −70 to +50 mV) in the presence of oxotremorine-M potentiated the muscarinic Ca²⁺ response. E, F, Mean ± SEM of the intracellular Ca²⁺ concentration rise normalized in each cell to the peak response to 10 μm (E) and 1 μm (F) oxotremorine-M at −70 mV (black bars). Control, Normalized Ca²⁺ rise during the depolarization before oxotremorine-M application; Peak, normalized maximum Ca²⁺ increase evoked by muscarinic receptor stimulation at −70 mV; first to sixth, normalized Ca²⁺ rises at the time of the first to the sixth depolarization in the presence of oxotremorine-M; wash, normalized Ca²⁺ rises after oxotremorine-M application. Depolarization (10 s to +50 mV; white bars) significantly potentiated the Ca²⁺ responses to 10 μm (E; n = 38; p < 0.01) and 1 μm (F; n = 5; p < 0.01) oxotremorine-M in all cells tested.

Bidirectional effect of depolarization on G_q-coupled receptor signaling

In addition to G_q-coupled muscarinic receptors, SH-SY5Y cells also endogenously express G_q-coupled bradykinin (B₂) receptors (Willars and Nahorski, 1995). Activation of these receptors also leads to Ca²⁺ release from intracellular stores via the PLC–IP₃ pathway [1 μm bradykinin, n = 6 (Fig. 2A); 0.1 μm bradykinin, n = 10 (Fig. 2B)]. In contrast to the muscarinic Ca²⁺ response, depolarization did not potentiate the bradykinin-evoked Ca²⁺ transient. In 6 of 16 cells (3 of 6 and 3 of 10 for 1 and 0.1 μm bradykinin, respectively), depolarization did not appear to influence the bradykinin-evoked Ca²⁺ transient at all, whereas in the remaining 10 of 16 cells (3 of 6 and 7 of 10 for 1 and 0.1 μm bradykinin, respectively), the bradykinin-evoked Ca²⁺ signal decreased for the duration of the depolarization. Interestingly, depolarization still potentiated the Ca²⁺ signal evoked by activation of muscarinic receptors in the same cell (Fig. 2C). These data show that depolarization can bidirectionally affect G_q-coupled receptor signaling, enhancing the muscarinic receptor-evoked Ca²⁺ signal but attenuating the bradykinin-evoked Ca²⁺ signal, even in the same cell.

Figure 2.

Bidirectional effect of depolarization on muscarinic and bradykinin receptor-mediated Ca²⁺ signaling in SH-SY5Y. A, B, Ca²⁺ responses of whole-cell patch-clamped SH-SY5Y cells to a maximal (1 μm; A) and submaximal (0.1 μm; B) concentration of bradykinin (black bar). Depolarization (10 s to +50 mV) attenuated the bradykinin-evoked Ca²⁺ response. C, Specimen trace of an SH-SY5Y cell in which the muscarinic Ca²⁺ signal (10 μm oxotremorine-M; first black bar) was potentiated (black arrows) by depolarization, whereas the bradykinin receptor-mediated Ca²⁺ signal was attenuated (white arrows) by depolarization (0.1 μm bradykinin; second black bar). The black dotted lines show the predicted Ca²⁺ signal if the cell was not depolarized.

The muscarinic Ca²⁺ response to activation of recombinant M₃ receptors in CHO and HEK cells is not affected by voltage

Being a neuronal cell line, SH-SY5Y cells are relatively complex, expressing a variety of typically neuronal proteins (see Discussion). To investigate whether voltage affects the M₃ receptor-evoked muscarinic Ca²⁺ signal even in the absence of these neuronal components, we examined the effect of voltage on the muscarinic Ca²⁺ signal in two non-neuronal cell lines, stably expressing recombinant M₃ receptors: CHO–m₃ and HEK–m₃ cells. A CHO–m₃ clone was chosen (Fig. 3A) that expresses the M₃ receptor at a comparable level with that present in SH-SY5Y cells (200–250 fmol/mg). Activation of muscarinic M₃ receptors in whole-cell patch-clamped CHO–m₃ cells in the absence of extracellular Ca²⁺ evoked transient Ca²⁺ increases that were not potentiated by depolarization (Fig. 3C) (10 μm oxotremorine-M, n = 5; 0.6 μm oxotremorine-M, n = 6). To check that this was not simply a characteristic of CHO–m₃ cell Ca²⁺ signaling, we repeated the experiments in another non-neuronal cell background. Again, the muscarinic Ca²⁺ signal in response to 0.6 μm oxotremorine-M in whole-cell patch-clamped HEK–m₃ cells (Fig. 3B) in the absence of extracellular Ca²⁺ was not affected by depolarization (Fig. 3D) (n = 6). These results show that the muscarinic Ca²⁺ signaling pathway is not necessarily affected by depolarization and suggest that the M₃ receptor itself is not sensitive to voltage. We thus investigated the DEP of the muscarinic Ca²⁺ signaling pathway in the neuronal SH-SY5Y cell line further to narrow down the site of voltage sensitivity.

Figure 3.

The muscarinic Ca²⁺ signal in CHO and HEK cells expressing recombinant M₃ receptors is not affected by voltage. DIC images of cultured CHO–m₃ (A) and HEK–m₃ (B) cells, two cell lines stably expressing recombinant M₃ receptors. C, D, Ca²⁺ responses of whole-cell patch-clamped CHO–m₃ (C) and HEK–m₃ (D) cells to 10 and 0.6 μm oxotremorine-M (black bars), respectively. Depolarization (10 s to +50 mV) did not affect the Ca²⁺ signal in the two expression systems.

The effect of voltage does not involve the G_i/o or G_s pathway

In addition to G_q-protein-coupled receptors, SH-SY5Y cells also express G_s and G_i/o-protein-coupled receptors (Klinz et al., 1987), which are positively and negatively coupled to adenylyl cyclase (AC), respectively. The different G-protein signaling pathways can interact, and cross talk between G_i/o- and G_q- as well as G_s- and G_q-coupled receptor signaling has been shown in a variety of systems (Werry et al., 2003). In SH-SY5Y cells, activation of some pertussis toxin-sensitive G_i/o-coupled receptors enhances the Ca²⁺ signal produced by activation of G_q-coupled M₃ muscarinic receptors (Connor and Henderson, 1996). Given that G_i/o-protein-coupled receptor signaling can be sensitive to voltage (Ben-Chaim et al., 2003), we investigated whether the DEP of the muscarinic Ca²⁺ signal was via an effect of voltage on G_i/o-protein signaling, which could crosstalk to enhance the muscarinic Ca²⁺ signal. However, even when G_i/o-protein signaling was blocked by incubation with pertussis toxin (20 h, 100 ng/ml), the muscarinic Ca²⁺ signal evoked by oxotremorine-M was still potentiated by depolarization (Fig. 4A,B) (1 μm, n = 5; 10 μm, n = 5), indicating that the DEP of the muscarinic Ca²⁺ signal does not involve the G_i/o pathway.

Figure 4.

The DEP of the muscarinic Ca²⁺ signal is not mediated by G_i/o-proteins or the G_s–AC pathway and does not involve L-type Ca²⁺ channels or ryanodine receptors. Ca²⁺ responses of whole-cell patch-clamped SH-SY5Y cells to 1 μm (A) or 10 μm (C) oxotremorine-M (black bars). The cells were incubated for 20 h in 100 ng/ml PTX to inhibit G_i/o-protein signaling (A) or exposed to 100 μm ryanodine for >6 min and during the course of the experiment to block ryanodine receptors (C). Even in these conditions, depolarization potentiated the muscarinic Ca²⁺ signal. B, Mean ± SEM of the intracellular Ca²⁺ rise evoked by the first depolarization in 1 μm (1 oxo-M) and 10 μm (10 oxo-M) oxotremorine-M in untreated (ctrl.) cells and in cells treated with PTX or forskolin/IBMX (fors.). In all conditions, depolarization potentiated the oxotremorine-M-evoked Ca²⁺ signal significantly (1 μm oxotremorine-M: control, n = 5, p < 0.01; PTX, n = 5, p < 0.01; 10 μm oxotremorine-M: control, n = 38, p < 0.01; PTX, n = 5, p = 0.03; forskolin/IBMX, n = 3, p = 0.03). The depolarization-evoked Ca²⁺ increase is not significantly different in control, PTX-, and forskolin/IBMX-treated cells (p > 0.05). D, Mean ± SEM of the intracellular Ca²⁺ rise evoked by the first depolarization in 10 μm oxotremorine-M in untreated cells (control), in cells treated with 10 μm nifedipine (nifedip.), and in cells exposed to 100 μm ryanodine (ryanod.). In all conditions, depolarization potentiated the muscarinic Ca²⁺ signal significantly (control, n = 38, p < 0.01; nifedipine, n = 6, p < 0.01; ryanodine, n = 7, p < 0.01). The depolarization-evoked muscarinic Ca²⁺ increase is not significantly different in control and in the presence of nifedipine or ryanodine (p > 0.05).

In some systems, the G_s–AC–cAMP pathway interacts with the G_q–PLC–IP₃ pathway to enhance Ca²⁺ signaling (Werry et al., 2003). One possible mechanism may be the phosphorylation and sensitization of IP₃ receptors to IP₃ by cAMP-dependent protein kinase (Wojcikiewicz and Luo, 1998). Because the activity of AC may be influenced, directly or indirectly, by the membrane potential (Reddy et al., 1995; Cooper et al., 1998), we investigated whether the DEP of the muscarinic Ca²⁺ signal was via an effect of voltage on the G_s–AC–cAMP signaling pathway, which could crosstalk to enhance the muscarinic Ca²⁺ signal. We exposed the cells for >10 min before and throughout the recording to 50 μm forskolin, to maximally activate AC and thus saturate cAMP production, and 100 μm IBMX, to inhibit phosphodiesterase and thus limit cAMP breakdown. Under these conditions, the G_s–AC–cAMP signaling pathway should be maximally activated and insensitive to any additional stimulation by, for example, voltage. However, even in these conditions, depolarization still potentiated the oxotremorine-M-evoked muscarinic Ca²⁺ signal (Fig. 4B) (10 μm oxotremorine-M, n = 3), indicating that the DEP of the muscarinic Ca²⁺ signal is not via an effect of voltage on the G_s–AC–cAMP pathway.

The DEP of the muscarinic Ca²⁺ signal does not involve L-type Ca²⁺ channels or ryanodine receptors

Depolarization-evoked Ca²⁺ release from intracellular stores has been demonstrated in a variety of systems. In skeletal muscle, depolarization of the sarcolemma is sensed by L-type Ca²⁺ channels, and the resulting conformational change is mechanically transferred by physical coupling to ryanodine receptors, leading to Ca²⁺ release from intracellular stores. A similar mechanism has also been proposed for terminals of hypothalamic neurons, in which plasmalemmal depolarization increases the open probability of intracellular ryanodine receptors, causing increased Ca²⁺ release from intracellular stores (De Crescenzo et al., 2004). In addition, depolarization sensed by L-type Ca²⁺ channels has also been shown to increase IP₃ mass in skeletal muscle (Araya et al., 2003) and activate G-proteins and the PLC–IP₃–Ca²⁺ pathway to induce Ca²⁺-induced Ca²⁺ release via ryanodine receptors in vascular myocytes (del Valle-Rodriguez et al., 2003). These effects of depolarization could be blocked by L-type Ca²⁺ channel inhibitors.

Because SH-SY5Y cells express functional voltage-gated L-type Ca²⁺ channels (Forsythe et al., 1992) as well as ryanodine receptors (Mackrill et al., 1997), we investigated whether the DEP of the muscarinic Ca²⁺ signal could be produced by a mechanism similar to the one in skeletal muscle, hypothalamic nerve terminals, or vascular myocytes. Although none of these were shown to require GPCR activation, constitutive activity may still have taken place. However, the Ca²⁺ response to 10 μm oxotremorine-M in the presence of the L-type Ca²⁺ channel blocker nifedipine (>6 min preincubation in 10 μm nifedipine; n = 6) or in the presence of 100 μm ryanodine (>6 min preincubation with ryanodine; n = 7) (Fig. 4C), a concentration that blocks ryanodine receptors (Verkhratsky, 2005), was still potentiated by depolarization (Fig. 4D). This indicates that neither functional L-type Ca²⁺ channels nor ryanodine receptors are required for the DEP of the muscarinic Ca²⁺ signal in SH-SY5Y cells and rules out a conformational coupling model, such as occurs in skeletal muscle cells.

The voltage dependence of the modulation of the muscarinic Ca²⁺ signal

Apart from L-type Ca²⁺ channels, SH-SY5Y cells possess a variety of other voltage-gated ion channels. These include voltage-gated sodium and potassium channels (Forsythe et al., 1992), currents through which can be recorded in whole-cell patch-clamped SH-SY5Y cells (see Fig. 6A–C). Theoretically, either of these channels could serve as voltage sensor and transfer the information of plasma membrane voltage changes to the muscarinic Ca²⁺ signaling pathway. To get an indication of the type of voltage sensor involved, we characterized the voltage dependence of the muscarinic Ca²⁺ signal.

Figure 6.

The DEP of the muscarinic Ca²⁺ signal is not via an effect of voltage on K⁺ efflux or NCX. A–C, Current responses of whole-cell patch-clamped SH-SY5Y cells to voltage steps from a holding potential of −70 mV to maximally +50 mV in 10 mV increments. In control solution (A), the cells exhibited transient voltage-gated inward and slowly inactivating voltage-gated outward currents. The outward currents vanished when intracellular potassium was replaced by NMDG (B), indicating that they were mediated by voltage-gated potassium channels, isolating the inward (sodium) current, whereas the inward current vanished when extracellular sodium was replaced by NMDG (C), indicating that they were mediated by voltage-gated sodium channels, isolating the outward (potassium) current. D–F, Ca²⁺ responses of whole-cell patch-clamped SH-SY5Y cells to 10 μm oxotremorine-M (black bars) in the absence of intracellular potassium (D) or extracellular sodium (E). Depolarization potentiated the muscarinic Ca²⁺ signal in both conditions. F, Mean ± SEM of the intracellular Ca²⁺ rise evoked by the first depolarization in 10 μm oxotremorine-M in untreated cells (control), in the absence of extracellular sodium (0 [Na⁺]_o), and in the absence of intracellular potassium (0 [K⁺]_i). In all conditions, depolarization potentiated the muscarinic Ca²⁺ signal significantly (control, n = 38, p < 0.01; 0 [Na⁺]_o, n = 6, p < 0.01; 0 [K⁺]_I, n = 5, p < 0.01). The depolarization-evoked muscarinic Ca²⁺ increase was not significantly different in control and in the absence of extracellular Na⁺ or intracellular K⁺.

The effect of different voltage step amplitudes (activation)

In whole-cell patch-clamped SH-SY5Y cells, 10 μm oxotremorine-M evoked Ca²⁺ transients at −70 mV in Ca²⁺-free solution, which were potentiated by depolarization to +50 mV. This equates to a voltage step (ΔV_step) of 120 mV amplitude. Subsequent voltage steps to +50 mV, given ∼1 min apart during the oxotremorine-M application, evoked Ca²⁺ increases with reduced amplitude (Fig. 1C), which may be attributable to gradual emptying of the stores. The amount of potentiation varied between cells, making it difficult to assess Ca²⁺ increases in response to different size voltage steps. To overcome variability between cells, we depolarized each cell from −70 to +50 mV (ΔV_step of 120 mV) before and after depolarizing it from −70 to x mV (ΔV_step of y mV). We then normalized the Ca²⁺ increase evoked by depolarization to x mV (i.e., the second depolarization) to the average of the two Ca²⁺ increases evoked by depolarization to +50 mV (i.e., the first and third depolarizations). Depolarizations of different amplitude caused different amounts of potentiation, with larger voltage steps evoking greater potentiations of the Ca²⁺ signal (Fig. 5A–C). No clear activation threshold could be observed, and depolarization to +80 mV (Fig. 5A,C) evoked even greater potentiation than depolarization to +50 mV (Fig. 1C, 5C). This suggests that the voltage sensor is not simply a voltage-gated ion channel, which should show a clear activation threshold and would be expected to be fully activated at +50 mV. Instead, the voltage dependence of the muscarinic Ca²⁺ signal was gradual, with no apparent upper limit.

Figure 5.

The DEP of the muscarinic Ca²⁺ signal is graded with no apparent upper limit, bipolar and non-inactivating. Ca²⁺ responses of whole-cell patch-clamped SH-SY5Y cells to 10 μm oxotremorine-M (black bars). A–C, The cells were depolarized for 10 s from −70 to +50 mV (ΔV_step of 120 mV) before and after depolarizing them from −70 mV to a certain step potential, e.g., from −70 to +80 mV (A; ΔV_step of 150 mV) or from −70 to 0 mV (B; ΔV_step of 70 mV). C, Mean ± SEM of the Ca²⁺ increases evoked by depolarization to a certain step potential (i.e., the second depolarization) normalized to the average of the two Ca²⁺ increases evoked by depolarization to +50 mV (i.e., the first and third depolarizations) in each cell. Larger voltage steps evoked greater potentiations of the Ca²⁺ signal, with no apparent upper limit (−100 mV, n = 7; −40 mV, n = 6; 0 mV, n = 9; +50 mV, n = 21; +80 mV, n = 6). D–F, The cells were depolarized for 10 s from −70 to +20 mV (ΔV_step of 90 mV) before and after depolarizing them slowly over ∼15 s to 0 mV and then for 10 s from 0 to +90 mV (D; ΔV_step of 90 mV) or from 0 to −90 mV (E; ΔV_step of −90 mV). F, Mean ± SEM of the Ca²⁺ increases evoked by depolarization or hyperpolarization from a holding potential of 0 to +90 mV or −90 mV (i.e., the second depolarization) normalized to the average of the two Ca²⁺ increases evoked by depolarization from −70 to +20 mV (i.e., the first and third depolarizations) in each cell.

The effect of depolarization from different holding potentials (inactivation) and the effect of hyperpolarization

To investigate whether the voltage sensor of the muscarinic Ca²⁺ signal could be inactivated by depolarization, we performed similar experiments as described above. However, after the initial first depolarization from −70 mV by a certain ΔV_step, we now depolarized the cells slowly over ∼15 s to a new holding potential before the second depolarization by the same ΔV_step, this time from the new holding potential. Interestingly, the slow depolarization to the new holding potential of −25 mV (n = 5) or 0 mV (n = 17) gradually increased the muscarinic Ca²⁺ signal (Fig. 5D,E). This confirms, as already suggested above, that the voltage sensor does not have a clear activation threshold. It also shows that the voltage sensor is not inactivated by slow depolarization, unlike most voltage-gated ion channels. Moreover, depolarization from the new holding potential of −25 mV (to +95 mV, i.e., ΔV_step of 120 mV; n = 5) or 0 mV (to +120 mV, i.e., ΔV_step of 120 mV, n = 5; or to +90 mV, i.e., ΔV_step of 90 mV, n = 7) still potentiated the muscarinic Ca²⁺ signal (Fig. 5D,F), indicating that the voltage sensor of the muscarinic Ca²⁺ signaling pathway is not inactivated at these potentials. Furthermore, hyperpolarization from a holding potential of 0 to −90 mV (ΔV_step of −90 mV) decreased the muscarinic Ca²⁺ signal (Fig. 5E,F) (n = 5), revealing a bipolar effect of voltage on the muscarinic signaling pathway.

These data unveil a gradual (with no apparent upper limit), non-inactivating, bipolar voltage dependence of the muscarinic Ca²⁺ signal in SH-SY5Y cells, excluding traditional voltage-gated ion channels as primary voltage sensors. This profile of voltage dependency is similar to the one observed on P2Y receptor signaling in megakaryocytes (Martinez-Pinna et al., 2004).

The DEP of the muscarinic Ca²⁺ signal is not via efflux and extracellular accumulation of K⁺

It has been suggested recently that the activity of G_q-protein-coupled P2Y receptors can be directly modulated by small changes of the extracellular K⁺ concentration, independent of Ca²⁺ influx or changes in membrane potential (Pitt et al., 2005). The modulation by extracellular K⁺ might not be restricted to P2Y receptors but may also occur in other G_q-protein-coupled receptors.

Under normal physiological conditions, depolarization increases whereas hyperpolarization decreases the driving force for K⁺ leaving the cell. Depolarization can thus shift K⁺ out of the cell and lead to accumulation of extracellular K⁺, which could modulate the activity of G_q-protein-coupled receptors. The gradual (with no apparent upper limit), non-inactivating, bipolar voltage dependence of the muscarinic Ca²⁺ signal in SH-SY5Y cells excludes traditional voltage-gated ion channels as primary voltage sensors but fits in with the passive movement of certain ions across the plasma membrane along their electrochemical gradients. We therefore investigated whether the DEP of the muscarinic Ca²⁺ signal in SH-SY5Y cells was attributable to a shift of K⁺ out of the cell, leading to extracellular K⁺ accumulation and potentiation of the M₃ receptor activity. SH-SY5Y cells were whole-cell patch-clamped and dialyzed with intracellular solution lacking K⁺ (replaced by NMDG). The dialysis of K⁺ from the cytoplasm was confirmed by the disappearance of voltage-gated K⁺ outward currents several minutes after establishing the whole-cell configuration (Fig. 6B). However, even in the absence of intracellular K⁺, 10 μm oxotremorine-M evoked a Ca²⁺ signal that was enhanced by depolarization (n = 6) (Fig. 6D,F), indicating that the DEP of the muscarinic Ca²⁺ signal is not via extracellular K⁺ accumulation and modulation of M₃ receptors by K⁺.

The modulation of the muscarinic Ca²⁺ signal by voltage is not via an effect of voltage on the Na⁺–Ca²⁺ exchanger

The cytoplasmic Ca²⁺ signal is a balance between Ca²⁺ increase (i.e., release from intracellular stores and influx from the extracellular space) and Ca²⁺ removal. In our experimental protocol, the Ca²⁺ increase evoked by G_q-coupled receptor activation and depolarization is via Ca²⁺ release from intracellular stores, because the experiments were performed in the absence of extracellular Ca²⁺, excluding the possibility of Ca²⁺ influx across the plasma membrane via voltage-, store-, or receptor-operated Ca²⁺ channels. Cells have a variety of mechanisms to remove Ca²⁺ ions from the cytoplasm, and effects of voltage on the activity of any of these processes will alter the shape of the Ca²⁺ transient. We thus set out to investigate whether the DEP of the muscarinic Ca²⁺ signal is via an effect of voltage on the Ca²⁺-clearing mechanisms, such as mediated by the Na⁺–Ca²⁺ exchanger NCX (Mason et al., 2000). This transporter requires Na⁺ and Ca²⁺ ions as substrates and is electrogenic, with depolarization inhibiting the removal of Ca²⁺ from the cell. To test whether the DEP of the muscarinic receptor-evoked Ca²⁺ signal is via an effect of voltage on NCX, we repeated the experiments in the absence of extracellular Na⁺ (confirmed by the disappearance of the fast inward current) (Fig. 6C) to block the transporter, of which there are two isoforms (NCX-1 and NCX-3) in SH-SY5Y cells (Magi et al., 2005). However, depolarization enhanced the Ca²⁺ signal evoked by 10 μm oxotremorine-M even in the absence of extracellular Na⁺ and Ca²⁺, when NCX is blocked (Fig. 6E,F) (n = 7), indicating that the DEP of the muscarinic Ca²⁺ signal does not require extracellular Na⁺ and is not via an effect of voltage on NCX. A potential effect of voltage on other Ca²⁺-clearing mechanisms was excluded using flash photolysis approaches and is discussed below.

The role of Ca²⁺ stores in the DEP of the muscarinic Ca²⁺ signal

Many intracellular organelles can serve as Ca²⁺ stores, such as the sarcoplasmic (SR) and endoplasmic reticulum (ER), the nuclear envelope (which is a physical continuation of the ER), the Golgi body, the mitochondria, an NAADP-sensitive store (proposed to be the lysosome or lysosome-related organelle), and secretory vesicles. Of these organelles, only the SR/ER, the nuclear envelope, and the cis-Golgi contain IP₃ receptors and accumulate Ca²⁺ from the cytoplasm via the thapsigargin-sensitive SERCA [sarco(endo)plasmic reticulum Ca²⁺ ATPase] pump (Michelangeli et al., 2005). It further appears that the SR/ER, although forming one single continuous tubular network, can be made up of two spatially and functionally distinct Ca²⁺ stores, one of which is IP₃ sensitive, whereas the other one is sensitive to caffeine and ryanodine (Golovina and Blaustein, 1997; Verkhratsky, 2005). In addition, these two stores may also differ in their sensitivity to thapsigargin, with the IP₃-sensitive store showing a higher sensitivity to the SERCA inhibitor than the caffeine-sensitive store (Garavito-Aguilar et al., 2004). Whereas virtually all SH-SY5Y cells contain an IP₃-sensitive Ca²⁺ store, which is accessed by muscarinic stimulation, only a very small percentage of SH-SY5Y cells appear to contain a functional caffeine-sensitive Ca²⁺ store (Riddoch et al., 2005), consistent with the low level of ryanodine receptor expression in these cells (Mackrill et al., 1997). However, the caffeine-sensitive store may act as Ca²⁺ source after muscarinic stimulation (Riddoch et al., 2005). We therefore investigated the role of the different Ca²⁺ stores in the DEP of the muscarinic Ca²⁺ signal and examined whether depolarization recruits Ca²⁺ from another store (e.g., a caffeine-sensitive store) during the muscarinic-evoked Ca²⁺ release, thus potentiating the muscarinic Ca²⁺ signal.

A caffeine-sensitive store is not evident in SH-SY5Y cells and is not required for the DEP of the muscarinic Ca²⁺ signal

Application of 10 mm (n = 10) or 30 mm (n = 4) caffeine, a ryanodine receptor agonist, to whole-cell patch-clamped SH-SY5Y cells in Ca²⁺-free solution did not evoke a Ca²⁺ response (Fig. 7A), indicating that these cells do not possess a functional caffeine-sensitive Ca²⁺ store. Depolarization from −70 to +50 mV before, during, or after the caffeine application did not affect the intracellular Ca²⁺ concentration (n = 14).

Figure 7.

The DEP of the muscarinic Ca²⁺ signal does require a functional thapsigargin-sensitive Ca²⁺ store but not a caffeine-sensitive Ca²⁺ store. Ca²⁺ responses of whole-cell patch-clamped SH-SY5Y cells to 10 μm oxotremorine-M (black bars) are shown. A, Caffeine at 10 mm (gray bar) failed to evoke a Ca²⁺ signal, showing that these cells do not contain a functional caffeine-sensitive store. After depletion of any putative caffeine-sensitive store, oxotremorine-M (black bar) evoked a Ca²⁺ signal that was potentiated by depolarization, indicating that voltage does not recruit Ca²⁺ from a caffeine-sensitive store. B, Oxotremorine-M failed to evoke Ca²⁺ responses in cells that were incubated for 20–30 min in 1 μm thapsigargin, to deplete thapsigargin-sensitive Ca²⁺ stores. Depolarization did not affect the Ca²⁺ signals in these cells, indicating that a functional thapsigargin-sensitive store is necessary for the effect of voltage. C, Mean ± SEM of the intracellular Ca²⁺ rise evoked by the first depolarization in 10 μm oxotremorine-M in untreated cells (control), in cells treated with 1 μm thapsigargin (thapsi.), and in cells pretreated with 10 mm caffeine. Depolarization potentiated the muscarinic Ca²⁺ signal significantly in control (n = 38; p < 0.01) and after caffeine treatment (n = 7; p < 0.01) but had no effect on the Ca²⁺ signal when thapsigargin-sensitive stores were depleted.

In addition to activating ryanodine receptors, millimolar concentrations of caffeine also block IP₃ receptors (Missiaen et al., 1994). We thus removed caffeine from the solutions before eliciting muscarinic Ca²⁺ responses. Although we could not detect a Ca²⁺ response to caffeine (see above), all Ca²⁺ from any putative caffeine-sensitive store should have been released during the 3 min exposure to 10 or 30 mm caffeine. Moreover, the experiments were performed in Ca²⁺-free solution, ensuring that, once emptied, the Ca²⁺ stores cannot refill and remain empty for the remainder of the experiment. Even in these conditions, several minutes after caffeine washout, 10 μm oxotremorine-M evoked a Ca²⁺ response in whole-cell patch-clamped SH-SY5Y cells at −70 mV in Ca²⁺-free solution, which was potentiated by depolarization (Fig. 7A,C) (n = 7). This shows that (1) the Ca²⁺ signal evoked by oxotremorine-M is not dependent on a caffeine-sensitive store, and (2) voltage does not recruit Ca²⁺ from a caffeine-sensitive store during the oxotremorine-M response. The DEP of the muscarinic Ca²⁺ signal is therefore not via an effect of voltage on Ca²⁺ release form a caffeine-sensitive store.

A functional thapsigargin-sensitive store is necessary for the DEP of the muscarinic Ca²⁺ signal

The thapsigargin-sensitive SERCA pump is located on the SR/ER, the nuclear envelope, and the cis-Golgi but not on mitochondria, lysosomes, or secretory vesicles (Michelangeli et al., 2005). The Ca²⁺ stores in these organelles can thus be classified as “thapsigargin-sensitive” and “thapsigargin-insensitive” stores. When the SERCA pump is inhibited by thapsigargin, the thapsigargin-sensitive Ca²⁺ stores rapidly deplete by leakage of Ca²⁺ out of these stores (Verkhratsky, 2005). To investigate the role of the thapsigargin-sensitive Ca²⁺ stores in the DEP of the muscarinic Ca²⁺ signal, we depleted these stores using thapsigargin to see whether depolarization evokes release of Ca²⁺ from a thapsigargin-insensitive store. After SH-SY5Y cells were incubated for 20–30 min in 1 μm thapsigargin, oxotremorine-M (10 μm; n = 4) failed to evoke a Ca²⁺ response in whole-cell patch-clamped SH-SY5Y cells at −70 mV in the absence of extracellular Ca²⁺ (Fig. 7B). This confirms that the muscarinic receptor-evoked Ca²⁺ response in SH-SY5Y cells is via Ca²⁺ release from a thapsigargin-sensitive store. More importantly, depolarization before and during the oxotremorine-M application did not affect the intracellular Ca²⁺ levels (n = 4) (Fig. 7B,C), indicating that depolarization does not evoke release of Ca²⁺ from a thapsigargin-insensitive store (i.e., from mitochondria, lysosomes, or secretory vesicles) but that a functional thapsigargin-sensitive store is necessary for the DEP of the muscarinic Ca²⁺ signal.

A functional IP₃-sensitive store is not sufficient for the DEP of the Ca²⁺ signal

The results above show that the DEP of the muscarinic Ca²⁺ signal (1) requires a functional thapsigargin-sensitive store and (2) does not involve a caffeine-sensitive store or ryanodine receptors, suggesting that a functional IP₃-sensitive store is necessary for the effect of voltage. To investigate whether Ca²⁺ release from an IP₃-sensitive store was sufficient for the DEP of the Ca²⁺ signal, we activated the IP₃-sensitive store directly, without the need for muscarinic stimulation, by photolytic release of caged IP₃ or caged GPIP₂, the less potent, slowly hydrolysable analog of IP₃ (Bird et al., 1992).

Uncaging IP₃ by a 100 ms UV flash caused a fast rising Ca²⁺ transient in whole-cell patch-clamped SH-SY5Y cells in Ca²⁺-free solution at −70 mV (Fig. 8A), which returned to baseline with a half-decay time of 8.1 ± 0.8 s (n = 7). Longer-lasting UV flashes release more caged IP₃ and evoked longer-lasting Ca²⁺ transients (Fig. 8E, white bars). Importantly, none of these Ca²⁺ responses were affected by depolarization. This is not attributable to the inability to release more Ca²⁺ at the time of depolarization, because a second UV flash given at that time was always able to evoke additional Ca²⁺ rise (Fig. 8B) (n = 3, 4, and 4 for two UV flashes of 100 ms, 500 ms, and 1 s duration, respectively).

Figure 8.

IP₃ receptor activation is not sufficient for the effect of voltage on the Ca²⁺ signal. A–D, Ca²⁺ responses (F/F_o) of whole-cell patch-clamped SH-SY5Y cells to one (A, C) or two (B, D) 100 ms flashes (arrows) to photolyze caged IP₃ (A, B) or GPIP₂ (C, D). Depolarization did not affect the IP₃ or the GPIP₂-evoked Ca²⁺ signals, showing that IP₃ receptor activation is not sufficient for the effect of voltage. E, Mean ± SEM of the half-decay of the Ca²⁺ response to flash release of caged IP₃ (white bars) or GPIP₂ (black bars) in relation to the flash duration. Longer flashes generate more IP₃/GPIP₂, resulting in prolonged IP₃ receptor activation and longer-lasting Ca²⁺ signals (IP₃, 100, 500, and 1000 ms, n = 7, 5, and 5, respectively; GPIP₂, 100, 500, and 1000 ms, n = 7, 5, and 6, respectively). The GPIP₂-evoked Ca²⁺ responses were significantly longer lasting than the IP₃-evoked Ca²⁺ responses, at all flash lengths (p < 0.01).

Because IP₃ is rapidly metabolized, the decay of the IP₃-evoked Ca²⁺ transient, during which the effect of voltage was examined, is likely to reflect mainly removal of Ca²⁺ from the cytoplasm. Thus, if voltage affects IP₃ receptor activation, depolarization may fail to alter the decay of the IP₃-evoked Ca²⁺ transient. To investigate the effect of voltage on the Ca²⁺ transient during IP₃ receptor activation, we used caged GPIP₂, a slowly metabolized analog of IP₃, to activate IP₃ receptors continuously. Ca²⁺ transients evoked by flash photolysis of caged GPIP₂ decayed much more slowly than the ones evoked by IP₃ (Fig. 8C,E), consistent with sustained activation of IP₃ receptors by GPIP₂. In addition, longer-lasting flashes produce more GPIP₂, leading to continued IP₃ receptor activation and longer-lasting Ca²⁺ responses (Fig. 8E, black bars). Like for IP₃, none of the GPIP₂-evoked Ca²⁺ responses were affected by depolarization. Again, this was not attributable to the inability to release more Ca²⁺ at the time of depolarization, because a second UV flash given at that time was able to evoke an additional Ca²⁺ rise (Fig. 8D) (n = 5). Whereas Ca²⁺ removal seems to be a major factor in shaping the IP₃-evoked Ca²⁺ transient, sustained IP₃ receptor activation and associated Ca²⁺ release appear to shape the Ca²⁺ transient evoked by GPIP₂. The observation that none of the IP₃- or GPIP₂-evoked Ca²⁺ transients were affected by depolarization suggests that neither Ca²⁺ removal from the cytoplasm nor Ca²⁺ release from the IP₃-sensitive stores are, under these conditions, sensitive to voltage, in accordance with voltage effects on IP₃ receptor signaling in megakaryocytes (Martinez-Pinna et al., 2005). Our data therefore show that (1) Ca²⁺-clearing mechanisms in SH-SY5Y cells are not affected by depolarization and (2) thapsigargin/IP₃-sensitive stores are necessary but not sufficient for the DEP of the muscarinic Ca²⁺ signal. The voltage sensor of the muscarinic Ca²⁺ signaling pathway must thus be before the level of the Ca²⁺ store and resides most likely within the IP₃ production pathway itself.

The muscarinic IP₃ production pathway is sensitive to voltage per se

To investigate whether the IP₃ production pathway itself is indeed sensitive to voltage, we used eGFP–PH_PLCδ (Stauffer et al., 1998) to visualize muscarinic receptor-evoked IP₃ production. Whereas under basal conditions this biosensor binds to PIP₂ and is thus located at the plasma membrane, it translocates to the cytoplasm during production of IP₃ and/or depletion of PIP₂ (Nahorski et al., 2003), which can be recorded as a cytoplasmic fluorescence increase. In agreement with previous studies in SH-SY5Y cells (Nash et al., 2001), stimulation of muscarinic receptors led to translocation of the biosensor into the cytoplasm (Fig. 9A,B). Depolarization before muscarinic receptor stimulation (control) did not enhance the cytoplasmic fluorescence signal in SH-SY5Y cells transfected with the biosensor (Fig. 9B) (n = 9), indicating that depolarization is not sufficient to trigger a measurable production of IP₃/depletion of PIP₂. However, the cytoplasmic fluorescent signal evoked by application of 10 μm oxotremorine-M was increased by depolarization to 170 ± 14% (Fig. 9C) (n = 9; p < 0.01), demonstrating that the muscarinic receptor-evoked IP₃ production pathway itself is indeed sensitive to voltage per se.

The muscarinic IP₃ production pathway in cerebellar granule neurons is voltage sensitive

We also investigated the voltage sensitivity of the IP₃ production pathway in cerebellar granule neurons, which endogenously express functional G_q-coupled muscarinic M₃ receptors (Whitham et al., 1991). Cultured granule cells (Fig. 10A) were transfected with eGFP–PH_PLCδ and perforated patch clamped to control the membrane voltage. Application of oxotremorine-M caused a translocation of the biosensor from the membrane to the cytoplasm (Fig. 10A,B), indicating increased IP₃ production/PIP₂ depletion. Depolarization before muscarinic receptor stimulation did not affect the cytoplasmic fluorescence signal in granule neurons (Fig. 10B) (n = 5), indicating that depolarization is not sufficient to trigger a measurable production of IP₃. However, depolarization during the muscarinic receptor response increased the cytoplasmic fluorescence to 163 ± 11% (n = 5; p < 0.01) (Fig. 10C), indicating substantial voltage sensitivity of the IP₃ production pathway in these primary neurons.

Discussion

Our results demonstrate that G_q-protein-coupled receptor signaling in neurons is directly sensitive to changes in membrane potential at the level of the IP₃ production pathway. Depolarization has a bidirectional effect on G_qPCR signaling, potentiating Ca²⁺ responses to muscarinic receptor activation but attenuating those mediated by bradykinin receptors. These effects represent a mechanism by which the electrical activity of the cell can directly shape G_qPCR-mediated IP₃ and Ca²⁺ signaling and thus provide a means for coincidence detection of ionotropic and metabotropic signals.

Mechanism of the depolarization-evoked potentiation

The voltage effects on G_qPCR signaling were observed in rat cerebellar granule neurons and the human neuronal cell line SH-SY5Y. Cerebellar granule neurons are the most numerous neurons in the brain and are thought to play an important role in motor learning (Marr, 1969). The cerebellar cortex is innervated by cholinergic fibers, which project mainly to the granule cell layer of the vestibulo-cerebellum (Ojima et al., 1989; Jaarsma et al., 1997), but the functional significance of this is not well understood. Cerebellar granule cells in situ and in vitro endogenously express functional G_q-coupled muscarinic receptors, almost exclusively of the M₃ subtype (Whitham et al., 1991; Tice et al., 1996; Takayasu et al., 2003). Alternatively, data were also obtained from SH-SY5Y cells. These cells closely resemble human sympathetic ganglion cells (Ross et al., 1983) and endogenously express a variety of ion channels (Fig. 6) (Forsythe et al., 1992), transporters (Magi et al., 2005), and certain GPCRs, including muscarinic M₃ receptors (Lambert et al., 1989; Willars and Nahorski, 1995). Like other neurons, they express the same complement of IP₃ receptor (type I), ryanodine receptor (type 2), and SERCA pump (SERCA-2B) (Mackrill et al., 1997; Bollimuntha et al., 2005) isoforms that prevail in the brain (Verkhratsky, 2005) and are thus a good model system to investigate the modulation of neuronal IP₃ and Ca²⁺ signaling. In contrast to these neuronal cells, agonist stimulation of HEK and CHO cell lines stably expressing recombinant muscarinic M₃ receptors produced intracellular Ca²⁺ signals that were not potentiated by voltage, indicating that muscarinic Ca²⁺ signaling is not necessarily voltage sensitive but requires a neuronal cell background or endogenous expression of other components of the signaling pathway.

The precise mechanism of the DEP of the muscarinic signaling in neuronal cells has yet to be determined, but it can be narrowed down to an effect of voltage on the IP₃ production pathway itself. In support of this, caged release of IP₃/GPIP₂ produces Ca²⁺ responses that are insensitive to depolarization, placing the voltage sensor before the level of the Ca²⁺ store. Furthermore, experiments using the IP₃/PIP₂ biosensor eGFP–PH_PLCδ directly demonstrate that the IP₃ production pathway itself is sensitive to voltage. Other potential mechanisms have also been ruled out, including (1) an effect of voltage on Ca²⁺ clearing mechanisms and (2) a conformation-coupling model involving L-type voltage-gated Ca²⁺ channels and ryanodine receptors. Moreover, the DEP has no apparent upper limit, is graded, bipolar, and non-inactivating, excluding traditional voltage-gated ion channels as the primary voltage sensors. PTX, forskolin, IBMX, and K⁺ substitution experiments have shown that the DEP is not via an effect of voltage on the G_i/o or G_s pathway and subsequent cross talk of these with the G_q pathway (Werry et al., 2003) or via an effect of voltage on K⁺ efflux, which could locally accumulate and enhance G_qPCR signaling (Pitt et al., 2005).

The components of the basic G_qPCR IP₃ production pathway are located in or associated with the plasma membrane, making them potentially susceptible to changes in membrane potential: (1) the G_qPCR itself is a heptahelical transmembrane protein (Pierce et al., 2002), (2) the α-subunit of the heterotrimeric G_q-protein is linked to the plasma membrane via palmitoylation (Wedegaertner et al., 1995), (3) PLCβ is loosely attached to the plasma membrane via its PH domain and C-terminal basic amino acids (Rhee, 2001), and (4) PIP₂, the substrate of PLCβ, is a highly negatively charged membrane lipid in the inner leaflet of the plasma membrane (McLaughlin et al., 2002). Although our data show that the IP₃ production pathway is modulated by voltage, they suggest that this effect is downstream of the G_qPCR, for the following reasons. Both maximal and submaximal concentrations of agonist produced Ca²⁺ responses in SH-SY5Y cells that were potentiated by depolarization, demonstrating that the voltage effects are not attributable to changes in agonist binding as has been suggested previously for the muscarinic response in acinar cells isolated from rat lacrimal glands (Marty and Tan, 1989). This, together with our findings that muscarinic Ca²⁺ signals in CHO–m₃ and HEK–m₃ cell lines were not potentiated by voltage, indicates that the action of voltage on the signaling cascade is downstream of the muscarinic G_qPCR and does not reside within the receptor. This is in agreement with the voltage effect on muscarinic Ca²⁺ signaling in guinea pig coronary smooth muscle (Ganitkevich and Isenberg, 1993) but in contrast to the effects on Ca²⁺ signaling via the P2Y receptor in rat megakaryocytes (Martinez-Pinna et al., 2005) and the signaling mediated by muscarinic M₁/M₂ and metabotropic glutamate 1/3 receptors recombinantly expressed in Xenopus oocytes (Ben-Chaim et al., 2003; Ohana et al., 2006). Our data thus complement these studies and demonstrate an alternative mechanism by which voltage is able to influence G_qPCR signaling.

It is unclear at present which component of the IP₃ production pathway downstream of the G_qPCR is voltage sensitive. One possibility is that the activity of PLCβ can be affected by voltage, such that the rate of PIP₂ hydrolysis, and hence IP₃ production, increases when the cell is depolarized. Although depolarization has been shown to indirectly affect the activity of PLCβ, by means of Ca²⁺ influx through voltage-gated ion channels and positive modulation of PLCβ (Hashimotodani et al., 2005), this cannot account for our results, because (1) all of our experiments were performed in the absence of extracellular Ca²⁺, (2) no increase in cytoplasmic Ca²⁺ was observed when the cells were depolarized in the absence of muscarinic agonist, and (3) the activation/inactivation profile of the DEP did not agree with a traditional voltage-gated ion channel as the primary voltage sensor. We also found that depolarization has a bidirectional effect, even in the same cell, potentiating the Ca²⁺ signal produced by muscarinic receptor activation but attenuating that produced by bradykinin receptor activation, although both receptors couple to the G_q–PLCβ–IP₃ pathway. The differential effects of voltage on the bradykinin and the muscarinic Ca²⁺ signals may be attributable to their ability to activate different PLCβ isoforms (Haley et al., 2000), some of which may be sensitive to voltage whereas others may not. Another possibility is that the availability of the PLCβ substrate PIP₂ is affected by voltage, and, at least in rat superior cervical ganglion neurons, muscarinic and bradykinin receptors appear to have different signaling microdomains (Delmas and Brown, 2002), possibly containing separate PLC isoforms and pools of PIP₂. Interestingly, the presence of a voltage-sensing domain in a PIP₂-producing enzyme Ci-VSP has been found recently in ascidian Ciona intestinalis sperm tails (Murata et al., 2005), demonstrating a possible link between membrane voltage and PIP₂ availability.

Physiological context

The observation that muscarinic IP₃ and Ca²⁺ signaling in neurons is directly sensitive to voltage could have considerable implications for neuronal function. It directly links the electrical activity of neurons to muscarinic signaling and may provide a fundamental mechanism by which ionotropic signals can shape metabotropic activity, thus opening a gateway to coincidence detection of ionotropic and metabotropic signaling. Interestingly, ionotropic and metabotropic signaling has been shown to synergize and increase IP₃ production and Ca²⁺ store release in a variety of neuronal preparations (Nakamura et al., 1999, 2000, 2002; Okubo et al., 2001, 2004; Larkum et al., 2003; Nash et al., 2004; Young et al., 2004, 2005). However, this synergism is thought to be attributable to Ca²⁺ entry via ionotropic and voltage-gated ion channels, leading to (1) increased store filling (Irving and Collingridge, 1998; Rae et al., 2000), (2) sensitization of the IP₃ receptor (Berridge, 1998; Nakamura et al., 1999), and (3) positive feedback onto PLCβ (Eberhard and Holz, 1988; Nash et al., 2004; Hashimotodani et al., 2005). Our results suggest an additional mechanism of synergism in which the ionotropic receptor-evoked depolarization itself, without the need of Ca²⁺ influx, may be able to increase IP₃ production and thus Ca²⁺ release from intracellular stores in neurons.

The DEP of muscarinic signaling could be important in a variety of physiological and pathological conditions. For example, muscarinic signaling in neurons can influence a variety of ion channels, thus modulating neuronal excitability (Brown et al., 1997; Suh and Hille, 2005). In the hippocampus, this results in cholinergic activation generally depolarizing neurons and increasing action potential discharge (Cobb and Davies, 2005; Young et al., 2005), and, in the cerebellum, the activation of muscarinic M₃ receptors leads to the depolarization of granule neurons and increases the frequency of spontaneous synaptic events onto Purkinje cells (Takayasu et al., 2003). Our data imply that the muscarinic-evoked depolarization could form a positive feedback onto the muscarinic signal, enhancing its actions further. Muscarinic signaling has also been shown to play a pivotal role in the enhancement of endocannabinoid release and thus depolarization-induced suppression of inhibition (DSI), via postsynaptic muscarinic receptors (Ohno-Shosaku et al., 2003). Interestingly, the enhancement of DSI is facilitated when muscarinic activation is combined with depolarization, an effect attributed to Ca²⁺ influx through voltage-gated Ca²⁺ channels and positive modulation of PLCβ by Ca²⁺ (Hashimotodani et al., 2005). Our results provide an additional mechanism in which depolarization per se enhances the muscarinic signaling pathway, thus contributing to the coincidence detection of muscarinic receptor activation and depolarization. The effect of voltage on G_qPCR signaling thus adds yet another layer of complexity to the neuronal signaling mechanisms mediated by these receptors.