Abstract
Ca2+ signals associated with action potentials (APs) and metabotropic glutamate receptor (mGluR) activation exert distinct influences on neuronal activity and synaptic plasticity. However, it is not clear how these two types of Ca2+ signals are differentially regulated by neurotransmitter inputs in a single neuron. We investigated this issue in dopaminergic neurons of the ventral midbrain using brain slices. Intracellular Ca2+ was assessed by measuring Ca2+-sensitive K+ currents or imaging the fluorescence of Ca2+ indicator dyes. Tonic activation of metabotropic neurotransmitter receptors (mGluRs, α1 adrenergic receptors, and muscarinic acetylcholine receptors), attained by superfusion of agonists or weak, sustained (∼1 s) synaptic stimulation, augmented AP-induced Ca2+ transients. In contrast, Ca2+ signals elicited by strong, transient (50–200 ms) activation of mGluRs with aspartate iontophoresis were suppressed by superfusion of agonists. These opposing effects on Ca2+ signals were both mediated by an increase in intracellular inositol 1,4,5-trisphosphate (IP3) levels, because they were blocked by heparin, an IP3 receptor antagonist, and reproduced by photolytic application of IP3. Evoking APs repetitively at low frequency (2 Hz) caused inactivation of IP3 receptors and abolished IP3 facilitation of single AP-induced Ca2+ signals, whereas facilitation of Ca2+ signals triggered by bursts of APs (five at 20 Hz) was attenuated by less than half. We further obtained evidence suggesting that the psychostimulant amphetamine may augment burst-induced Ca2+ signals via both depression of basal firing and production of IP3. We propose that intracellular IP3 tone provides a mechanism to selectively amplify burst-induced Ca2+ signals in dopaminergic neurons.
- calcium (Ca)
- intracellular signaling
- IP3 receptor
- metabotropic glutamate receptor
- burst
- dopaminergic neuron
Introduction
Dopaminergic (DA) neurons in the ventral midbrain play a critical role in reward-based reinforcement learning, and maladaptive learning caused by excessive DA signals has been implicated in the development of drug addiction (Schultz, 1998; Redish, 2004). The firing of DA neurons in vivo switches between tonic, single-spike activity and phasic, glutamate-driven bursts (Hyland et al., 2002). DA neuron bursts giving rise to phasic DA signals in projection areas are thought to provide the “teaching signal” for reinforcement learning.
The firing pattern of neurons is determined by their intrinsic membrane properties and the synaptic inputs they receive. Intracellular Ca2+ ([Ca2+]i) is critically involved in both of these processes by regulating membrane conductances and synaptic plasticity. Ca2+ influx triggered by action potentials (APs) activates small-conductance Ca2+-sensitive K+ (SK) channels in DA neurons, generating large afterhyperpolarizations (AHPs) that control tonic firing frequency (Wolfart and Roeper, 2002). However, transient activation of metabotropic glutamate receptors (mGluRs) elicits release of Ca2+ from intracellular stores, producing a prolonged SK-mediated hyperpolarization. This hyperpolarization underlies a pause of activity that curtails phasic bursts driven by ionotropic glutamate receptors (iGluRs) (Fiorillo and Williams, 1998; Morikawa et al., 2003). Ca2+ signals associated with postsynaptic APs and mGluR activation also play important roles in the induction of plasticity at a variety of synapses in the CNS (Bortolotto et al., 1999; Linden, 1999; Nevian and Sakmann, 2006). Recent studies have shown that long-term potentiation (LTP) and long-term depression (LTD) of iGluR-mediated transmission can be induced in a manner dependent on postsynaptic bursts of APs and mGluR activation, respectively, in DA neurons (Bellone and Luscher, 2005; Liu et al., 2005), although other forms of plasticity have been reported using different induction protocols (Kauer, 2004; Jones and Bonci, 2005).
DA neurons receive multiple neurotransmitter inputs that activate metabotropic receptors coupled to phospholipase C (PLC)-mediated phosphoinositide (PI) hydrolysis, including mGluRs, α1 adrenergic receptors (α1ARs), and muscarinic acetylcholine receptors (mAChRs). Tonic activation of these receptors suppresses phasic mGluR-induced responses in DA neurons (Fiorillo and Williams, 1998, 2000; Paladini and Williams, 2004). Indeed, the psychostimulant amphetamine has been shown to inhibit mGluR-induced Ca2+ release via elevated extracellular DA tone activating α1ARs (Paladini et al., 2001). However, the exact intracellular mechanism mediating this inhibition remains unclear. Furthermore, it is not known how Ca2+ signals triggered by APs are affected by these PI-coupled neurotransmitter inputs in DA neurons.
In this study, we show that sustained activation of PI-coupled receptors augments AP-evoked Ca2+ transients while inhibiting phasic mGluR induced Ca2+ signals in DA neurons. Both of these effects are mediated by an increase in cytosolic inositol 1,4,5-trisphosphate (IP3) levels, which facilitates Ca2+-induced Ca2+ release (CICR) via IP3 receptors (IP3Rs) and, at the same time, reduces the size of Ca2+ stores. We further find that the facilitation of AP-induced Ca2+ signals is modulated by the pattern in which APs are generated, so that IP3 selectively amplifies Ca2+ signals triggered by bursts of APs, but not those induced by single APs evoked at low frequency.
Materials and Methods
Slices and solutions.
Horizontal slices (200–220 μm) of the ventral midbrain were prepared from adult Sprague Dawley rats (3–6 weeks old), as described previously (Cui et al., 2004). Slices were maintained at 35°C and perfused at a rate of 2–3 ml/min with physiological saline containing the following (in mm): 126 NaCl, 2.5 KCl, 1.2 NaH2PO4, 1.2 MgCl2, 2.4 CaCl2, 11 glucose, 21.4 NaHCO3, saturated with 95% O2 and 5% CO2, pH 7.4, 300 mOsm/kg. Unless noted otherwise, pipette solutions used for whole-cell and cell-attached recordings contained the following (in mm): 115 K-methylsulfate, 20 KCl, 1.5 MgCl2, 10 HEPES, 0.025 EGTA, 2 Mg-ATP, 0.2 Na2-GTP, and 10 Na2-phosphocreatine, pH 7.3, 280 mOsm/kg.
Electrophysiological recordings.
All recordings were performed in putative DA neurons, which were identified by their large cell bodies (>20 μm) visualized with infrared-differential interference contrast optics, spontaneous firing at 1–5 Hz, and the presence of large hyperpolarization-activated Ih currents. Most (∼90%) of the recordings were made in the substantia nigra pars compacta (SNc), whereas the remainder were in the ventral tegmental area (VTA). Whole-cell voltage-clamp recordings were made at a holding potential of −62 mV, corrected for a liquid junction potential of 7 mV. Series resistance (∼10–20 MΩ) was continuously monitored but left uncompensated. In whole-cell current clamp recordings, small hyperpolarizing currents were injected to maintain the membrane potential at ∼60 mV. MultiClamp 700A or 700B amplifiers (Molecular Devices, Union City, CA) were used to record the data, which were filtered at 1–2 kHz, digitized at 2–5 kHz, and collected using AxoGraph 4.9 (Molecular Devices) or AxoGraph X (AxoGraph Scientific, Sydney, Australia).
Iontophoretic pipettes (∼100 MΩ) containing l-aspartate (1 m; pH 7.4) were placed within 5 μm of the soma or proximal dendrites. Iontophoretic pulses (50–200 nA; 50–200 ms; 0–5 nA backing current) were applied once per minute. Synaptic responses were evoked with a bipolar tungsten electrode (tip separation 100 μm) placed at 50–100 μm rostral to the recorded cell. To isolate mGluR-mediated responses, these experiments were done in slices treated with 6,7-dinitroquinoxaline-2,3-dione (DNQX; 10 μm), dizocilpine maleate (MK-801; 50 μm), picrotoxin (100 μm), (2S)-3-{[(1S)-1-(3,4-dichlorophenyl)ethyl]amino-2-hydroxypropyl}(phenylmethyl)phosphinic acid (CGP55845; 10 μm), and eticlopride (100 nm) to block AMPA, NMDA, GABAA, GABAB, and DA D2 receptors, respectively.
Spontaneous AP firing was monitored using either perforated-patch or cell-attached recording configurations. Perforated-patch pipettes were filled with gramicidin (50–250 μg/ml) in a solution containing 135 mm KCl and 10 mm HEPES. The firing frequency within aspartate-induced bursts was obtained by calculating the average of the second and third interspike intervals after aspartate iontophoresis.
Ca2+ imaging.
Fluorescence imaging of [Ca2+]i was made using fluo-5F (Kd = 2.3 μm) or fluo-4FF (Kd = 9.7 μm) as Ca2+ indicators. These indicator dyes were loaded into the cell via the whole-cell pipette. Images were taken at 15–20 Hz using a Disk Spinning Unit confocal imaging system (Olympus, Melville, NY). Ca2+ signals from selected regions of interest (ROIs) were expressed as follows: %ΔF/F = 100 × (F − Fbaseline)/(Fbaseline − Fbackground).
Flash photolysis.
A 1 ms UV pulse was applied using a xenon arc lamp to elicit photolysis of caged IP3 or caged Ca2+ and the resulting SK-mediated outward current was measured. The concentration of compounds photolytically released is known to be proportional to the intensity of UV pulse, which is determined by the capacitance of the capacitor in the photolysis system (Cairn Research, Faversham, UK) supplying a current to the xenon arc lamp and the voltage to which the capacitor is charged. In this study, the voltage was constantly set at 300 V and the capacitance was varied (≤4000 μF) to adjust the UV pulse intensity. The peak of the IP3-evoked outward current (IIP3) trace was not rounded even with a supramaximal intensity of UV pulse, suggesting that SK channels were not saturated by Ca2+. Our previous study also demonstrated approximately linear relationship between the IIP3 amplitude and [Ca2+]i over a wide range in DA neurons (Morikawa et al., 2000). The amplitude of IIP3 elicited with a supramaximal UV intensity was 350–700 pA (515 ± 37 pA; n = 10). The threshold UV intensity was defined as the largest intensity that evoked IIP3 <20 pA, which was routinely achieved with the capacitance in the range of 50–150 μF. This frequently resulted in no detectable IIP3 because of the limited resolution in setting the capacitance of the photolysis system. Caged IP3 (100–200 μm) was loaded into the cell through the whole-cell pipette. For caged Ca2+, DM-nitrophen (1.5 mm) and CaCl2 (1.2 mm) were added to the whole-cell pipette.
Drugs.
DNQX, MK-801, (S)-3,5-dihydroxyphenylglycine (DHPG), 2-methyl-4-carboxyphenyglycine (LY367385), CGP55845, and cyclopiazonic acid (CPA) were obtained from Tocris Bioscience (Ellisville, MO). Heparin, ruthenium red, ryanodine, and DM-nitrophen were purchased from Calbiochem (La Jolla, CA). Tetrodotoxin (TTX) was obtained from Alomone Labs (Jerusalem, Israel). Fluo-5F, fluo-4FF, and caged IP3 were purchased from Invitrogen (San Diego, CA). All other chemicals were obtained from Sigma-RBI (St. Louis, MO).
Data analysis.
Data are expressed as means ± SEM. Statistical significance was determined by Student's t test or ANOVA followed by Bonferroni's post hoc test. The difference was considered significant at p < 0.05.
Results
APs trigger CICR in DA neurons
Whole-cell voltage clamp recordings were made from DA neurons in the VTA and SNc. To assess AP-induced Ca2+ signals, a 2 ms depolarizing pulse of 30–60 mV was applied from a holding potential of −62 mV to evoke an unclamped AP. This produced an outward tail current lasting 150–300 ms, which was inhibited by apamin (100 nm), a blocker of Ca2+-activated SK conductance (n = 33) (Fig. 1A). The apamin-sensitive component, termed IAHP, peaked at ∼20 ms after the test pulse, whereas the transient outward current insensitive to apamin mostly decayed within 20 ms. Therefore, we calculated the integral of the outward current from 20 to 300–600 ms after the test pulse to assess the charge transfer representing IAHP (called QAHP hereafter). IAHP was evoked in an all-or-none manner when the depolarizing pulse amplitude was varied up to 50–60 mV and was abolished by TTX (1 μm; n = 11) (supplemental Fig. S1, available at www.jneurosci.org as supplemental material). Further increasing the test pulse amplitude (n = 4) or prolonging the test pulse duration (n = 4) added a TTX-insensitive component to IAHP. Thus, we routinely used a 2 ms depolarizing pulse of 30–60 mV to evoke SK-dependent IAHP resulting from Ca2+ influx triggered by unclamped APs.
APs trigger Ca2+ release via both IP3Rs and RyRs. A, Representative traces of outward tail currents caused by an unclamped AP, evoked by a 2 ms depolarizing pulse in control and in apamin (100 nm). Apamin selectively eliminated the slow component of the tail current. The gray trace represents the IAHP obtained by current subtraction. The vertical dashed line is drawn at 20 ms after the test pulse. B, Representative traces illustrating the effect of CPA (20 μm) on IAHP recorded with a control (Con) internal solution, heparin (Hep; 1 mg/ml), or ruthenium red (RR; 200 μm). The gray trace in the left panel depicts the CPA-sensitive component of IAHP. C, Summary bar graph plotting QAHP in control and in CPA. Recordings were made with a control internal solution, heparin, ruthenium red, or both of these drugs together. *p < 0.05; **p < 0.01; ***p < 0.001; paired t test. D, Summary bar graph plotting the effect of CPA in cells recorded with a control internal solution, heparin, ruthenium red, or both of these drugs together. **p < 0.01; ***p < 0.001 versus control; ANOVA.
AP-induced Ca2+ transients can be amplified by CICR from intracellular stores via IP3Rs and/or ryanodine receptors (RyRs) (Berridge, 1998). Consistent with this, bath application of CPA (10–20 μm), which depletes endoplasmic reticulum Ca2+ stores (Smith et al., 1988), depressed QAHP by 44 ± 3% (n = 10) (Fig. 1B–D). The CPA-sensitive component of IAHP peaked at 50–100 ms after the test pulse. In spontaneously firing DA neurons recorded in current clamp, CPA also reduced the AHP amplitude (n = 3; perforated-patch recordings) and produced a small increase in the firing frequency (n = 9; cell-attached or perforated-patch recordings) (supplemental Fig. S2, available at www.jneurosci.org as supplemental material). The effect of CPA on QAHP was diminished by intracellular application of heparin (1 mg/ml; n = 5), an IP3R antagonist (Ghosh et al., 1988), or ruthenium red (200 μm; n = 8), a RyR antagonist (Zucchi and Ronca-Testoni, 1997), and was eliminated when these two drugs were applied together (n = 6) (Fig. 1B–D). In these experiments, inclusion of heparin and/or ruthenium red in the intracellular solution tended to reduce QAHP, although this reduction was not statistically significant because of the variability in QAHP among different cells. Ryanodine (20 μm), which locks RyR channels in a subconductance open state and thus depletes Ca2+ stores expressing RyRs (Zucchi and Ronca-Testoni, 1997), also produced 45 ± 9% inhibition of QAHP (n = 6), an effect that was blocked by ruthenium red (−1 ± 11% inhibition; n = 5; p < 0.01 vs control). These results demonstrate that both IP3Rs and RyRs are involved in CICR triggered by AP-induced Ca2+ influx.
Differential regulation of AP- and phasic mGluR-induced Ca2+ signals by tonic activation of PI-coupled receptors
Tonic activation of PI-coupled neurotransmitter receptors, such as mGluRs, α1ARs, and mAChRs, inhibits [Ca2+]i release produced by strong, phasic activation of these receptors in DA neurons (Fiorillo and Williams, 2000; Paladini et al., 2001; Morikawa et al., 2003). We thus tested whether sustained activation of these receptors could also affect the Ca2+ store-dependent component of IAHP. In agreement with the previous studies cited above, bath perfusion of an mGluR agonist DHPG (1 μm; n = 4), an α1AR agonist phenylephrine (10 μm; n = 7), and an mAChR agonist muscarine (1 μm; n = 7), all produced reversible inhibition of the mGluR-mediated outward current (ImGluR) evoked by aspartate iontophoresis in the presence of iGluR antagonists (Fig. 2A,B,E). In contrast, DHPG augmented IAHP by adding a slow component, which lasted up to 600 ms (Fig. 2A). This effect of DHPG was blocked by apamin (100 nm; n = 3) (Fig. 2C), consistent with the augmentation of AP-induced Ca2+ transients. On average, DHPG produced 108 ± 17% increase in QAHP (n = 21) (Fig. 2E). Furthermore, superfusion of phenylephrine and muscarine increased QAHP by 65 ± 10% (n = 11) and 116 ± 25% (n = 4), respectively (Fig. 2D,E). DHPG (n = 2) and phenylephrine (n = 2) also increased the AHP amplitude, measured in whole-cell current clamp, by ∼30% (Fig. 2D).
Sustained activation of PI-coupled receptors differentially regulates AP- and phasic mGluR-induced Ca2+ signals. A, Time course of the opposing effects of DHPG (1 μm) on QAHP and ImGluR recorded in the same cell. Representative traces of IAHP and ImGluR in control and in DHPG from the same experiment are shown on the right. ImGluR was evoked by iontophoretic application of aspartate (100 ms) at the time indicated by the arrow. B, Representative traces of ImGluR in control and in muscarine (1 μm). C, The effect of DHPG on IAHP was blocked by apamin (Apa; 100 nm). Con, Control. D, Left, Representative traces of IAHP in control and in phenylephrine (Phe; 10 μm). Right, Phenylephrine also augmented AHPs recorded in the whole-cell current clamp. APs were evoked by a 2 ms current injection. E, Summary bar graph showing that DHPG, phenylephrine, and muscarine all increased IAHP and inhibited ImGluR. F, DHPG enhanced the Ca2+ signal induced by a burst of APs while suppressing the phasic mGluR-mediated Ca2+ wave. Left, A confocal fluorescence image of a DA neuron loaded with fluo-5F (50 μm) is shown. Scale bar, 20 μm. Fluorescence changes were measured at the ROIs indicated by the boxes with different color codes (top traces), while currents were recorded in the same cell (bottom traces). Middle, A train of five test pulses at 20 Hz was used to evoke a burst of five unclamped APs. Right, Aspartate was iontophoresed at the soma, as indicated by the arrow in the image at left. G, Summary bar graph from five cells depicting the effects of DHPG on burst- and mGluR-induced Ca2+ signals measured in dendrites ∼20–50 μm away from the soma.
We next monitored [Ca2+]i by the fluorescence of fluo-5F (50 μm). In these experiments, we evoked a train of five unclamped APs at 20 Hz to mimic burst firing of DA neurons. The burst of APs produced a rise in [Ca2+]i and a summating outward current with a prolonged tail (Fig. 2F). The total charge transfer of this outward current, termed QAHP-burst, was calculated after removing a 20 ms window after each test pulse to isolate the apamin-sensitive component (Fig. 1A). The burst-induced Ca2+ signal appeared simultaneously throughout the cell but was invariably larger in proximal dendrites (∼20–50 μm from the soma) than in the soma. TTX (1 μm) abolished the burst-induced fluorescence change in dendrites and also reduced that in the soma by ∼80% (n = 7) (supplemental Fig. S3, available at www.jneurosci.org as supplemental material). Aspartate iontophoresis produced an mGluR-mediated wave of Ca2+ starting at the site of application, as reported previously (Morikawa et al., 2003). Bath application of DHPG (1 μm) increased the burst-induced fluorescence change by 75 ± 15% and decreased the mGluR-induced fluorescence change by 66 ± 12% when measured at proximal dendrites (n = 5) (Fig. 2G).
We further induced sustained activation of mGluRs by weak synaptic stimulation of glutamatergic fibers (60 stimuli at 50 Hz, starting 1 s before the onset of the burst), which evoked no detectable Ca2+ transient or outward current by itself (Fig. 3). Synaptic stimulation increased the burst-induced fluorescence change at the proximal dendrite close to the stimulating electrode by 59 ± 6% (n = 2) and QAHP-burst by 60 ± 11% (n = 3). The facilitatory effect of synaptic stimulation was blocked by bath application of LY367385 (50 μm), an mGluR1 antagonist (n = 3).
Synaptic activation of mGluRs facilitates burst-induced Ca2+ transients. A confocal fluorescence image of a DA neuron loaded with fluo-5F (50 μm) is shown on the right. Scale bar, 20 μm. A bipolar stimulating electrode was placed ∼50 μm rostral to the cell for synaptic stimulation. A train of five depolarizing pulses at 20 Hz, evoking a burst of APs, was applied with (w/; red) or without (w/o; black) synaptic stimulation (stim). The synaptic stimulation consisted of 60 stimuli at 50 Hz starting 1 s before the onset of the burst. The stimulation intensity was adjusted so that synaptic stimulation alone did not induce detectable fluorescence changes. Fluorescence changes were measured at the ROIs indicated by the numbered boxes. IAHP-burst is shown at the bottom. Synaptic stimulation enhanced IAHP-burst and the burst-induced Ca2+ signal at the rostral dendrite close to the stimulating electrode. This enhancement was blocked by the mGluR1 antagonist, LY367385 (LY; 50 μm).
Altogether, these results show that tonic activation of PI-coupled receptors augments AP-induced Ca2+ transients and suppresses phasic mGluR-mediated Ca2+ release.
IP3 mediates differential regulation of AP- and phasic mGluR-induced Ca2+ signals
To examine whether the facilitation of AP-induced Ca2+ signals is attained via CICR, we first used CPA to deplete Ca2+ stores. The effects of DHPG (1 μm; n = 10) and phenylephrine (10 μm; n = 4) on IAHP were abolished by CPA (10–20 μm) (Fig. 4A). Furthermore, the outward current produced by flash photolysis of caged Ca2+ (Icaged-Ca), which directly liberates Ca2+ inside of the cell, was also increased by DHPG and phenylephrine in a CPA-sensitive manner (Fig. 4B,C). Although CPA itself appeared to irreversibly suppress Icaged-Ca, we did not quantify this effect because Icaged-Ca showed rundown over time. Thus, tonic activation of PI-coupled receptors augments AP-induced Ca2+ signals by facilitating CICR.
Tonic activation of PI-coupled receptors augments CICR. A, Depleting [Ca2+]i stores with CPA blocked DHPG-induced facilitation of IAHP. DHPG (1 μm), CPA (10 μm), and apamin (100 nm) were perfused at the times indicated by horizontal bars. Traces of IAHP at the times indicated are shown on the right. B, Representative traces of Icaged-Ca depicting the facilitating effect of phenylephrine (Phe; 10 μm) and its CPA sensitivity. Photolytic release of Ca2+ was made at the time indicated by the arrow. Con, Control. C, Summary time graph showing that DHPG and phenylephrine augmented Icaged-Ca in a CPA-dependent manner. Icaged-Ca was evoked every 2 min. The amplitude of Icaged-Ca was normalized to the average of two values before drug application. DHPG (n = 4 in control; n = 2 in CPA) and phenylephrine (n = 3 in control; n = 3 in CPA) were perfused at the time indicated by the bar. DHPG and phenylephrine data were combined and averaged.
G-protein-coupled metabotropic receptors can mediate the production of two Ca2+-mobilizing messengers, IP3 and cyclic ADP-ribose (cADPR), which act on IP3Rs and RyRs, respectively (Cancela, 2001). To test the involvement of these messengers, we used heparin (1 mg/ml) to block the IP3–IP3R pathway and ruthenium red (200 μm) to block the cADPR–RyR pathway. DHPG-induced enhancement of IAHP and suppression of ImGluR were both suppressed by heparin but not by ruthenium red (Fig. 5A). Furthermore, the effects of phenylephrine (10 μm) and muscarine (1 μm) on IAHP and ImGluR were also blocked by heparin (n = 3 for each; data not shown). Superfusion of DHPG, phenylephrine, and muscarine all produced a small, sustained inward current (0–60 pA). This inward current was not affected by heparin or ruthenium red (data not shown), consistent with its independence from [Ca2+]i mobilization (Guatteo et al., 1999). These data suggest that an increase in IP3 tone mediates both facilitation of IAHP and suppression of ImGluR.
IP3 mediates differential regulation of AP- and phasic mGluR-induced Ca2+ signals. A, Summary bar graph showing the effects of DHPG (1 μm) on QAHP and ImGluR in cells recorded with a control internal solution, heparin (Hep; 1 mg/ml), or ruthenium red (RR; 200 μm). **p < 0.01 versus control; ANOVA. B, Left, Representative traces of IAHP with and without photolytic release of IP3. UV flash was applied 50 ms before the 2 ms depolarizing pulse (arrow). The current elicited by IP3 without the depolarizing pulse is also shown. The gray trace represents simple summation of control IAHP and the current evoked by IP3 alone. Right, Representative traces of ImGluR with and without repetitive photolytic release of IP3. Aspartate iontophoresis (200 ms) was made at the time indicated by the arrow. UV flash was applied at 5 Hz for 3 s, starting 1 s before aspartate iontophoresis. C, Representative traces of IAHP illustrating biphasic effects of DHPG at different concentrations. Con, Control.
To directly demonstrate the role of IP3, we next performed flash photolysis of caged IP3, which releases Ca2+ from intracellular stores and produces an SK-mediated outward current in DA neurons (Morikawa et al., 2000). IP3 released by a single UV pulse at threshold intensity, which barely produced an outward current by itself (<20 pA), facilitated IAHP in a manner similar to perfusion of receptor agonists (n = 5) (Fig. 5B). Furthermore, ImGluR was inhibited by repetitive photolytic applications of IP3 at threshold intensity for 3 s (n = 4) (Fig. 5B), which would gradually deplete [Ca2+]i stores by sustained Ca2+ release (Solovyova and Verkhratsky, 2003). Therefore, a small increase in cytosolic IP3 concentration reproduced the regulation of IAHP and ImGluR induced by sustained activation of PI-coupled receptors. A higher concentration of DHPG (30 μm) caused inhibition of IAHP instead of facilitation in five of eight cells tested (Fig. 5C), suggesting that higher IP3 levels produced by strong activation of mGluRs more fully deplete Ca2+ stores and thus can suppress CICR triggered by APs.
Activity-dependent regulation of AP-induced Ca2+ signals
DA neurons tonically fire APs at 1–5 Hz and also display phasic bursts comprising 2–10 spikes at 10–50 Hz in vivo (Hyland et al., 2002). To investigate the regulation of AP-induced Ca2+ signals under more physiological conditions, IAHP was evoked by a train of five test pulses at 2 Hz to mimic tonic firing. QAHP was rapidly reduced during the 2 Hz AP train (Fig. 6A1), from 17.3 ± 2.1 pC for the first IAHP (QAHP-1) to 11.2 ± 2.0 pC for the fifth one (QAHP-5) (n = 7; p < 0.01) (Fig. 6A2). The reduction in QAHP during the AP train was largely eliminated in the presence of CPA (9.3 ± 2.5 pC for QAHP-1 vs 8.2 ± 1.9 pC for QAHP-5; n = 5; p > 0.05). Although CPA produced an inhibition of QAHP-5, the magnitude of inhibition was significantly smaller compared with that of QAHP-1 (45 ± 4% inhibition for QAHP-1 vs 26 ± 5% inhibition for QAHP-5; n = 5; p < 0.05) (Fig. 6A3). These results demonstrate that evoking APs at 2 Hz causes a decrease in AP-triggered CICR.
Evoking APs at 2 Hz suppresses AP-induced CICR. A1, Representative traces of IAHP evoked by a train of five test pulses at 2 Hz in control (black), DHPG (1 μm; red), and CPA (10 μm; blue). A2–A4, QAHP, CPA-induced inhibition of QAHP, and DHPG-induced increase in QAHP for the first (AHP-1) and fifth (AHP-5) test pulses in the AP train are plotted for each cell. Note that DHPG had no effect on QAHP-5. B1, Representative traces of the outward current (IAHP-burst) evoked by a train of five test pulses at 20 Hz (burst) in control (black), DHPG (red), and CPA (blue). A burst was elicited alone (isolated burst; left) or 500 ms after a 2 Hz, five-pulse train (post-AP train burst; right). The first three IAHP in the AP train are not shown for clarity. Traces in A1 and B1 are from the same cell. B2–B4, QAHP-burst, CPA-induced inhibition of QAHP-burst, and DHPG-induced increase in QAHP-burst for the isolated burst and the post-AP train burst are plotted for each cell. Note that DHPG increased QAHP-burst even for the post-AP train burst. C, Evoking APs at 2 Hz suppressed the burst-induced Ca2+ transient. A burst was elicited alone or after a 2 Hz, five-pulse train in current clamp. APs were evoked by 2 ms current injections. A confocal fluorescence image of a DA neuron loaded with fluo-5F (50 μm) is shown on the left. Scale bar, 20 μm. Fluorescence changes were measured at the ROI indicated by the box, while the membrane potential was recorded in the same cell (bottom traces). D, Fluorescence changes induced by the isolated burst and the post-AP train burst are plotted for each cell. *p < 0.05; **p < 0.01; ***p < 0.001; paired t test.
Surprisingly, the facilitatory effect of DHPG (1 μm) on IAHP was abolished with the 2 Hz AP train (118 ± 23% increase for QAHP-1 vs −2 ± 4% increase for QAHP-5; n = 6; p < 0.001) (Fig. 6A1,A4). Consistent with this, the facilitatory effect of IP3, photolytically applied at threshold UV intensity, was also significantly depressed by the AP train (n = 6; p < 0.05) (Fig. 7A,B). The suppression of IP3 facilitation of CICR may be because of inactivation or reduced sensitivity of IP3Rs caused by [Ca2+]i elevations during the AP train (Taylor and Laude, 2002). Alternatively, AP-induced CICR may reduce the size of Ca2+ stores to prevent further CICR. Indeed, depolarization-induced Ca2+ influx can either increase or reduce the Ca2+ store size depending on the relative rates of Ca2+ uptake and release (Albrecht et al., 2001). To address these possibilities, we first asked whether the AP train alters the IP3 sensitivity of IP3Rs. Here, we performed flash photolysis of caged IP3 at 500 ms after the 2 Hz AP train using different intensities of UV pulse, thereby varying the concentrations of IP3 released. The AP train had no effect on the IP3-mediated outward current (IIP3) evoked with a supramaximal UV intensity. However, the same AP train caused significant depression of IIP3 elicited using lower UV intensities that produced ∼20% (EC20) or ∼50% (EC50) of the maximal current amplitude (Fig. 7C). The magnitude of AP train-induced reduction in IIP3 became smaller with an increase in the UV intensity [i.e., with an increase in the IP3 concentration (Fig. 7E)], indicating that high levels of IP3 can overcome the suppression caused by the AP train. We also found that a UV intensity that was subthreshold when applied 500 ms after the AP train was able to evoke a measurable outward current when applied without the AP train, indicating that the AP train elevated the threshold for evoking IIP3 (n = 3; data not shown). In contrast, bath application of DHPG (1 μm) produced comparable inhibition of IIP3 evoked by UV pulses at EC50 and supramaximal intensities (Fig. 7D,E). These data are consistent with the idea that (1) repetitive APs cause inactivation of IP3Rs by reducing their IP3 sensitivity and (2) DHPG-induced increase in IP3 tone inhibits IP3R-mediated Ca2+ release with no change in IP3 sensitivity, as would be expected if the driving force for Ca2+ release is diminished by a reduction in the store Ca2+ concentration. Furthermore, the reduction in IIP3 and QAHP caused by the 2 Hz AP train recovered over a period of several seconds (Fig. 7F,G), in good agreement with the recovery kinetics of Ca2+-dependent inactivation of IP3Rs (Parker and Ivorra, 1990; Finch et al., 1991), but much faster than the time course of Ca2+ store replenishment that takes place over minutes (Albrecht et al., 2001; Solovyova and Verkhratsky, 2003). Altogether, these results strongly suggest that repetitive AP-induced Ca2+ influx produces inactivation of IP3Rs during the AP train.
Repetitive APs at 2 Hz induce IP3R inactivation. A, Traces of IAHP evoked by a train of five test pulses at 2 Hz with (gray) and without (black) photolytic release of IP3. UV flash at threshold intensity was applied 50 ms before the first (left) or fifth (right) test pulse. B, The magnitude of increase in QAHP by photolytic release of IP3 at threshold intensity for the first (AHP-1) and fifth (AHP-5) test pulses in the AP train is plotted for each cell. *p < 0.05; paired t test. C, Representative traces of IIP3 with (gray) and without (black) a preceding 2 Hz AP train. A UV flash was applied 500 ms after the fifth pulse of the train. The intensity of the UV flash was varied to elicit maximal current amplitude (Max) and ∼20% (EC20) or ∼50% (EC50) of the maximal current. D, Representative traces of IIP3 in control (black) and in DHPG (1 μm; gray). E, Summary bar graph showing the magnitude of inhibition of IIP3, evoked with different UV intensities, by a 2 Hz AP train and by DHPG. Note that the AP train produced smaller inhibition as the UV intensity was increased. ***p < 0.001; repeated-measures ANOVA for post-AP train data and paired t test for DHPG data. F, G, The recovery time course of IIP3 (n = 3; F) and QAHP (n = 5; G) after the 2 Hz AP train. The intensity of UV flash was approximately EC50 for the experiments in F. The dotted lines represent single exponential fit to the data.
We further investigated the influence of the 2 Hz AP train on burst-induced Ca2+ signals. In these experiments, a burst was evoked by a train of five test pulses at 20 Hz, either in isolation or 500 ms after the 2 Hz train. As expected, QAHP-burst was significantly reduced when the burst was preceded by the AP train compared with the one elicited in isolation (64.3 ± 8.4 pC for the isolated burst vs 47.0 ± 5.3 pC for the post-AP train burst; n = 6; p < 0.05) (Fig. 6B1,B2). In line with this, the AP train caused a 33 ± 5% reduction in the burst-induced fluorescence change at proximal dendrites (∼20–50 μm from the soma), using fluo-5F (50 μm) or fluo-4FF (100 μm) as Ca2+ indicators (n = 6; four cells in voltage clamp and two cells in current clamp) (Fig. 6C,D). Furthermore, the effect of CPA on QAHP-burst was diminished by the AP train (38 ± 4% inhibition for the isolated burst vs 32 ± 4% inhibition for the post-AP train burst; n = 5; p < 0.05) (Fig. 6B3), consistent with a reduction in CICR. However, DHPG, which had no effect on QAHP-5 in the AP train (Fig. 6A1,A4), produced a significant increase in QAHP-burst after the AP train (30 ± 9% increase; n = 6) (Fig. 6B1,B4), although the magnitude of increase was smaller compared with that for the isolated burst (53 ± 12% increase; p < 0.01). These results suggest that, even if IP3Rs are inactivated by repetitive APs, a large Ca2+ influx associated with a burst of APs is able to trigger IP3R-dependent CICR when IP3 tone is elevated.
Potential dual mechanisms of amphetamine action on burst-induced Ca2+ signals
The findings above imply that suppression of basal AP firing can remove IP3R inactivation and increase the effect of IP3. The action of psychostimulants within the DA nuclei is thought to play an important role in the development of certain behavioral adaptations underlying addiction (Kauer, 2004; Jones and Bonci, 2005). It is well known that the psychostimulant amphetamine suppresses tonic, single-spike firing via activation of D2 autoreceptors (Mercuri et al., 1989), an effect that should increase burst-induced Ca2+ signals and their IP3-mediated facilitation by removing IP3R inactivation. However, it is not clear how amphetamine affects burst firing itself. To address this issue, we performed cell-attached recordings of DA neuron firing (Fig. 8A1). Iontophoretic application of aspartate elicited an iGluR-mediated burst followed by an mGluR-mediated pause, as reported previously (Morikawa et al., 2003). Bath application of amphetamine (10 μm) for ∼3 min decreased the basal firing frequency by 86 ± 3% (from 2.4 ± 0.2 to 0.3 ± 0.1 Hz; n = 5) (Fig. 8A1,A2). However, amphetamine produced only a small inhibition of the firing frequency within bursts (27 ± 4% inhibition, from 19.0 ± 3.1 to 13.8 ± 2.3 Hz) and the number of spikes per burst (26 ± 3% inhibition, from 7.4 ± 0.6 to 5.4 ± 0.3). These data demonstrate that amphetamine spares burst firing compared with its massive inhibition of tonic firing.
Amphetamine may augment burst-evoked Ca2+ signals via dual mechanisms. A1, Representative traces of DA neuron firing in control and in amphetamine (Amph; 10 μm) recorded with a cell-attached configuration. The burst was elicited by iontophoretic (ionto) application of aspartate (50 ms). A2, Summary bar graph illustrating that amphetamine suppressed basal firing frequency (FF) with relatively small effects on burst firing. The data are from five cells. B1, Representative traces of IAHP-burst generated by a control protocol simulating the firing pattern under control conditions (19 Hz; 7 AP burst preceded by 5 APs at 2.4 Hz; left) and by an amphetamine protocol mimicking the firing pattern in amphetamine (14 Hz, 5 AP burst preceded by 2 APs at 0.33 Hz; middle). A trace of IAHP-burst evoked by the amphetamine protocol in the presence of phenylephrine (10 μm) is also shown on the right. All three traces are from the same cell. B2, QAHP-burst of individual cells recorded under three different conditions as in B1. Switching from the control protocol to the amphetamine protocol induced a significant increase in QAHP-burst, which was further augmented by phenylephrine. ***p < 0.001; repeated-measures ANOVA.
Amphetamine-induced dopamine release can also activate α1ARs in DA neurons (Paladini et al., 2001; Cui et al., 2004). Therefore, it is possible that amphetamine augments burst-induced Ca2+ signals via both D2-mediated inhibition of basal firing and α1AR-mediated facilitation of CICR. To test this possibility, we first examined how amphetamine-induced change in the firing pattern affects burst-evoked Ca2+ signals. In these experiments, IAHP was elicited using two different protocols: (1) a control protocol simulating the firing pattern under control conditions (seven test pulses at 19 Hz for the burst, preceded by five test pulses at 2.4 Hz for the basal firing) and (2) an amphetamine protocol mimicking the firing pattern in amphetamine (five test pulses at 14 Hz for the burst, preceded by two test pulses at 0.3 Hz for the basal firing) (Fig. 8B1). QAHP-burst induced using the amphetamine protocol was significantly larger than that elicited with the control protocol (44 ± 11% increase; n = 11) (Fig. 8B2). Phenylephrine (10 μm) superfused, whereas evoking IAHP with the amphetamine protocol further increased QAHP-burst by 17 ± 5% (n = 5). These results suggest that amphetamine can augment burst-induced Ca2+ signals via both suppression of basal AP firing and IP3-mediated facilitation of CICR.
Discussion
By measuring the Ca2+-sensitive SK conductance and the fluorescence of Ca2+ indicator dyes, we have demonstrated that tonic activation of PI-coupled receptors enhances AP-evoked Ca2+ transients while inhibiting Ca2+ signals caused by phasic mGluR activation in DA neurons. A small rise in intracellular IP3 tone, causing (1) sensitization of IP3Rs to Ca2+-dependent activation and (2) partial depletion of [Ca2+]i stores, is responsible for these effects. This differential regulation is also tuned by the context in which APs are generated, in a manner that selectively amplifies burst-induced Ca2+ signals in tonically firing DA neurons. These findings provide important mechanistic insights into the regulation of Ca2+ signals by metabotropic neurotransmitter inputs and neuronal activity.
IP3 mediates differential regulation
Our results show that APs trigger CICR through both IP3Rs and RyRs under resting conditions. IP3Rs and RyRs are each coactivated by Ca2+ and another intracellular messenger: IP3 for IP3Rs and cADPR for RyRs (Berridge, 1998). Accordingly, a rise in IP3 or cADPR levels can augment AP-induced CICR via IP3Rs or RyRs (Hua et al., 1994; Nakamura et al., 2000). Previous studies have shown that both IP3 and cADPR contribute to Ca2+ release produced by strong, transient activation of mGluRs and α1ARs in DA neurons (Morikawa et al., 2003; Paladini and Williams, 2004). However, only the IP3 pathway was involved in the facilitation of CICR, as well as in the inhibition of transient mGluR-induced Ca2+ release, produced by relatively weak, sustained activation of these receptors. It is likely that the cADPR concentration did not reach levels necessary to coactivate RyRs with Ca2+ or to induce continuous Ca2+ leak via RyRs causing partial depletion of Ca2+ stores. It has been shown that pharmacological blockade of the PLC–IP3 pathway suppresses the enhancement of AP-induced Ca2+ responses by metabotropic receptors in other central neurons (Pan et al., 1994; Nakamura et al., 1999; Power and Sah, 2002). Our data showing that direct photolytic application of IP3 can reproduce the effect of receptor activation provide strong support for the role of IP3 in this process.
Strong, focal activation of mGluRs elicits a slowly propagating Ca2+ wave in DA neurons as well as in hippocampal and cortical pyramidal neurons (Nakamura et al., 1999; Larkum et al., 2003; Morikawa et al., 2003), likely reflecting the diffusion of IP3/cADPR or Ca2+. In contrast, Ca2+ transients evoked by a burst of APs occurred simultaneously at the soma and proximal dendrites, which were enhanced by DHPG in terms of both the amplitude and duration. This observation suggests that IP3 tone developed throughout the cell facilitated CICR triggered by rapidly propagating APs (Hausser et al., 1995). However, the amplitude of Ca2+ transients, as well as the magnitude of DHPG-induced enhancement, was larger in proximal dendrites than in the soma. Although the reason for this difference is uncertain, it is likely to result, at least partly, from the large volume of the soma diluting Ca2+ and/or IP3 (Watanabe et al., 2006). Consistent with this idea, weak, sustained synaptic stimulation of mGluRs facilitated burst-induced Ca2+ signals only in the proximal dendrite close to the stimulating electrode but not in the soma or the opposite dendrite.
Our result showing that DHPG inhibits IIP3 without changing the IP3 sensitivity suggests a reduction in the size of Ca2+ stores. It should also be noted that DHPG-induced inhibition of ImGluR slowly recovered over a period of ∼10 min after washout of DHPG, likely reflecting slow replenishment of Ca2+ stores (Solovyova and Verkhratsky, 2003). It has been shown that IP3Rs and RyRs are expressed on a common pool of Ca2+ stores in DA neurons (Morikawa et al., 2000). Recent evidence further suggests that the endoplasmic reticulum in DA neurons may actually be a single, interconnected lumen (Choi et al., 2006). Therefore, IP3 tone produced by sustained receptor activation likely affects the entire pool of Ca2+ stores. In line with this, our previous study has demonstrated that sustained activation of α1ARs suppresses RyR-mediated spontaneous Ca2+ responses observed in DA neurons of neonatal rats (Cui et al., 2004).
Repetitive APs inactivates IP3Rs
IP3Rs are biphasically regulated by cytosolic Ca2+ (Taylor and Laude, 2002). Compared with rapid Ca2+-dependent activation of IP3Rs, inactivation induced by Ca2+ has a slow rate of onset (hundreds of milliseconds) and lasts for seconds (Parker and Ivorra, 1990; Finch et al., 1991; Doi et al., 2005). A previous study has shown that repetitive firing of DA neurons produces elevations of [Ca2+]i up to ∼250 nm (Wilson and Callaway, 2000), which may well cause Ca2+-dependent inactivation of IP3Rs. Indeed, evoking APs at 2 Hz suppressed IP3R-mediated CICR in this study. It should be noted that during the 2 Hz AP train, AP-triggered CICR was fully suppressed when the second AP was evoked (i.e., 500 ms after the first AP) both in the control and in DHPG. Furthermore, this reduction fully recovered in 5–10 s. These observations are in good agreement with the kinetics of IP3R inactivation described above. We further found that IP3R inactivation caused by repetitive APs was associated with a decrease in IP3 sensitivity, consistent with previous studies demonstrating reduced IP3 binding affinity when IP3Rs are inactivated by Ca2+ (Joseph et al., 1989; Moraru et al., 1999).
It has been reported that IP3Rs inactivated by a low concentration of Ca2+ (250 nm) can still be activated by a higher concentration of Ca2+ (1 μm) in the presence of a constant level of IP3 (Finch et al., 1991). In line with this, a large Ca2+ influx produced by a burst of APs was able to trigger CICR through IP3Rs even when they were inactivated by repetitive APs. Importantly, this property enables tonic activation of PI-coupled receptors to selectively amplify burst-induced Ca2+ signals when DA neurons are constantly firing at low frequency.
In contrast to the suppression of AP-induced CICR and/or IP3R-mediated Ca2+ release caused by repetitive APs in this study, tonic firing of DA neurons for a relatively prolonged period (20–50 s) has been shown to augment mGluR-mediated Ca2+ release after the firing is stopped by a hyperpolarizing current injection (Fiorillo and Williams, 1998). This augmentation, which is most likely because of loading of Ca2+ stores by AP-induced Ca2+ influx (Stutzmann et al., 2003; Watanabe et al., 2006), is largest at ∼10 s after the injection of hyperpolarizing current and gradually declines over 1–2 min. The apparent discrepancy can be accounted for by the persistence of store loading, which can last for several minutes (Pozzo-Miller et al., 2000), much longer than the Ca2+-dependent IP3R inactivation that recovers in seconds. In our study, IIP3/IAHP reduced during the 2 Hz, five-AP train did not show much over-recovery above the control level after a 10 s interval, suggesting that the short (2 s) AP train produced little store loading.
In this study, amphetamine suppressed tonic firing with only a small inhibitory effect on iGluR-induced bursts, suggesting that the iGluR-mediated excitatory drive can largely overcome the D2-mediated inhibition. A recent report also showed that cocaine had similar differential effects on basal firing and evoked bursts in vivo (Almodovar-Fabregas et al., 2002). Amphetamine exerts two opposing actions on DA neuron activity: D2 autoreceptor-mediated inhibition and α1AR-mediated excitation (Mercuri et al., 1989; Shi et al., 2000; Paladini et al., 2001). Our study demonstrates that these two receptors can cooperate to selectively augment burst-induced Ca2+ signals. Here, D2-mediated suppression of basal firing removes Ca2+-dependent inactivation of IP3Rs, thereby boosting the facilitation of CICR via α1AR-mediated production of IP3 tone.
Functional significance: potential relevance to synaptic plasticity
It has been shown recently that LTP and LTD of iGluR-mediated transmission can be induced in a manner dependent on postsynaptic bursts of APs and mGluR activation, respectively, in DA neurons (Bellone and Luscher, 2005; Liu et al., 2005). A rise in [Ca2+]i is ubiquitously involved in the induction of LTP dependent on postsynaptic APs (Linden, 1999; Nevian and Sakmann, 2006). Furthermore, the mGluR-dependent LTD in DA neurons is blocked by the Ca2+ chelator BAPTA, suggesting the role of mGluR-mediated Ca2+ release (Bellone and Luscher, 2005). Thus, an increase in intracellular IP3 levels may shift the balance of DA neuron plasticity toward LTP by selectively amplifying burst-induced Ca2+ signals while suppressing mGluR-induced Ca2+ release.
Weak, sustained stimulation of glutamatergic fibers effectively facilitated burst-induced Ca2+ transients via mGluR activation in this study. DA neurons also receive other neurotransmitter inputs activating PI-coupled receptors. These include noradrenergic, cholinergic, and peptidergic inputs, which can all act through volume transmission (i.e., through a rise in extracellular tone) and are involved in behavioral arousals (Grace et al., 1998; Fiorillo and Williams, 2000; Ungless et al., 2003; Borgland et al., 2006). Recent studies further demonstrate the role of PI-coupled receptors on DA neurons in the behavioral and pharmacological responses to psychostimulants (Paladini et al., 2001; Ungless et al., 2003; Borgland et al., 2006). By promoting the potentiation of glutamatergic transmission onto DA neurons, the differential regulation of Ca2+ signals described in this study may contribute to enhanced reinforcement learning when animals are placed in behaviorally arousing environments or exposed to psychostimulants.
Footnotes
-
This work was supported by National Institutes of Health Grant DA015687. M.T.H. was supported by a National Science Foundation Graduate Research Fellowship. We thank Drs. Kamran Khodakhah, Christopher Fiorillo, and Nace Golding for comments on this manuscript and Dr. Tomokazu Doi for helpful discussions. We also thank Nicholas Gustafson for conducting some preliminary experiments.
- Correspondence should be addressed to Hitoshi Morikawa, Waggoner Center for Alcohol and Addiction Research, University of Texas, 2500 Speedway, Molecular Biology Building 1.150A, Austin, TX 78712. morikawa{at}mail.utexas.edu
