D2-like dopamine autoreceptors regulate dopamine release and are implicated in important actions of antipsychotic drugs and rewarding behaviors. To directly observe the effects of D2 autoreceptors on exocytic neurotransmitter release, we measured quantal release of dopamine from pheochromocytoma PC12 cells that express D2 and D4 autoreceptors. High potassium-evoked secretion in PC12 cells produced a unimodal population of quantal sizes. We found that exposures to the D2-like agonist quinpirole that inhibited tyrosine hydroxylase activity by ∼50% also reduced quantal size by ∼50%. The reduced quantal size was blocked by the D2 antagonist sulpiride and reversed byl-DOPA. Quinpirole also decreased the frequency of stimulation-evoked quantal release. Together, these findings indicate effects on quantal neurotransmission by D2-like dopamine autoreceptors previously distinguished as synthesis-modulating autoreceptors that regulate tyrosine hydroxylase activity versus impulse-regulating autoreceptors that modulate membrane potential. The results also provide an initial demonstration of a receptor-mediated mechanism that alters quantal size.
D2-like dopamine (DA) autoreceptors that inhibit stimulation-dependent DA release are present in mesencephalic DA dendrites (Saller and Salama, 1984; Nissbrandt et al., 1985; el Mestikawy and Hamon, 1986; Strait and Kuczenski, 1986) and presynaptic sites of mesolimbic terminals (Goldstein et al., 1990;Tepper and Groves, 1990; Westerink et al., 1990; Chen et al., 1991;Mercuri et al., 1992; Sesack et al., 1994). Because D2-like autoreceptors are thought to play important roles in the action of antipsychotics (Hetey et al., 1991), psychostimulant drugs (Brodie and Dunwiddie, 1990; Bernardini et al., 1991; Seutin et al., 1991; Yamada et al., 1991; Giambalvo, 1992; Zhang et al., 1992; Pitts et al., 1993;Silvia et al., 1994; Xue et al., 1994), and sexual behavior (Hull et al., 1990), the mechanisms by which autoreceptors mediate synaptic DA release have received much attention. However, brain D2-like receptors are located both presynaptically on DA afferents and postsynaptically on cortical and striatal neurons, making it difficult to isolate effects of autoreceptor activation on neurotransmitter release (Palij et al., 1990; Timmerman et al., 1990;Santiago and Westerink, 1991). Some reports suggest that D2autoreceptors on mesocortical DA afferents may primarily hyperpolarize membrane potential (impulse-regulating autoreceptors) (Chiodo et al., 1984; Bean and Roth, 1991; Fedele et al., 1993), whereas mesostriatal afferents and cell body autoreceptors may modulate both membrane potential (Bernardini et al., 1991; Cass and Zahniser, 1991; Kennedy et al., 1992; Tanaka et al., 1992; Cardozo, 1993) and the affinity of tyrosine hydroxylase (TH) for tyrosine (synthesis-modulating autoreceptors) (Bohmaker et al., 1989; Onali and Olianas, 1989;Salah et al., 1989; Johnson et al., 1992; Tissari and Lillgèls, 1993).
Recently, advances in electrochemical techniques have enabled the detection of vesicular release of monoamine transmitters in real time from chromaffin cells (Wightman et al., 1991; Chow et al., 1992), mast cells (Alvarez de Toledo et al., 1993), PC12 pheochromocytoma cells (Chen et al., 1994), invertebrate neurons (Bruns and Jahn, 1995; Chen et al., 1995), pancreatic cells (Zhou and Misler, 1996), mammalian sympathetic neurons (Zhou and Misler, 1995), and mammalian midbrain neurons (Pothos and Sulzer, 1998; Pothos et al., 1998). Zhou et al. (1994) have used these techniques to demonstrate a direct relationship between norepinephrine release and adrenergic autoreceptor activation in isolated chromaffin cells. Because PC12 cells express D2-like receptors that can inhibit DA release after exposure to the agonist quinpirole (Courtney et al., 1991), we adapted this technique to examine the effect of D2 autoreceptor activation on quantal DA release. We report that receptor activation reduces the amount of transmitter released per quantum, a previously unexplored potential mechanism of synaptic plasticity.
MATERIALS AND METHODS
PC12 culture. PC12 cells were obtained from Dr. Lloyd Greene (Columbia University) and maintained as described (Greene et al., 1991). For electrochemical experiments, 40,000 cells were plated per culture on glass coverslips (Carolina Biological Supply, Burlington, NC), coated with 40 μg/ml poly-d-lysine (Sigma, St. Louis, MO), and recoated with 10 μg/ml laminin (Collaborative Research, Bedford, MA). Cultures were maintained in media containing phenol red-free RPMI 1640 media supplemented with 10% heat-inactivated horse serum, 5% fetal bovine serum (JRH Biologicals, Lenexa, KS), and 50 U/ml penicillin and streptomycin (Life Technologies, Grand Island, NY) in 7.5% CO2 atmosphere at 37°C. Experiments were performed on undifferentiated cells at least 5 d after plating, and compared cultures were derived from the same source culture. HPLC measurements of the cultures, which are grown in the absence of the DA β-hydroxylase cofactor ascorbic acid, indicate that <5% of catecholamines are norepinephrine and that the remainder is DA (Pothos et al., 1996).
RT-PCR methods. For total RNA extraction from PC12 cultures, a 60-mm-diameter tissue culture dish (Falcon) coated with 10 μg of Vitrogen-100 collagen/cm2 (Collagen Biomedical, Palo Alto, CA) was plated with 2 × 106 PC12 cells. The cells were cultured for 1 week. The monolayer was rinsed with PBS, and 5 × 105 cells were harvested by trituration in 1 ml of the saline. One milliliter of Trizol reagent (Life Technologies) was added to the cell suspension. The tissue was homogenized by trituration through a 28.5 gauge needle, 200 μl of chloroform was added, and the RNA was precipitated from the aqueous phase after centrifugation (12,000 × g for 15 min) with isopropanol.
For total RNA extraction from brain tissue, 230 mg of tissue from rat ventral striatum was dissected from postnatal day 1 pups and washed in PBS. RNA extraction with Trizol reagent was performed as above, using 7 ml of Trizol and 1.4 ml of chloroform, and the tissue was homogenized by trituration with an 18 gauge needle.
mRNA from the total RNA sample was isolated using the Poly ATtract mRNA isolation system (Promega, Madison, WI) following the manufacturer’s instructions. Reverse transcriptase and reagents for cDNA preparation from mRNA were obtained from the Superscript preamplification system kit (Life Technologies). Twenty microliters of the reaction mix were incubated at 42°C for 50 min, and the reaction was terminated by heating to 70°C for 15 min. The sample was then chilled on ice for 2 min, and the remaining RNA was degraded by incubation at 37°C for 20 min with RNase H.
For PCR amplification, we used custom primers (Life Technologies) with published sequences (Surmeier et al., 1996). For thermal cycling we used a Perkin-Elmer (Emeryville, CA) thermal cycler. Incubation times and temperatures were as follows: (1) 94°C for 4 min; (2) 45 cycles of 94°C (1 min), 56°C (1 min), and 72°C (1.5 min); and (3) 72°C for 10 min. Samples were then chilled at 4°C. Reagents were from Promega.
Tyrosine hydroxylase assay. To estimate the effects of drug exposure on TH activation in cultured cells, we adapted methods to measure DOPA formation (Hayashi et al., 1990). In brief, sister cultures were rinsed twice with 1 ml of the physiological medium used for the DA uptake assay (see below) and then incubated with DA agonists or antagonists for 40 min at 37°C in 100 μm l-tyrosine and 500 μm brocresine (Lederle Laboratories, Pearl River, NY), a selective DOPA decarboxylase inhibitor. After the incubation, medium was discarded, the cells were covered with 1 ml of 0.1N PCA, collected, and sonicated. An aliquot (50 μl) was used to determine protein concentration (Lowry assay), and the remainder was centrifuged at 15,000 × g at 4°C for 10 min. A 20 μl aliquot of the supernatant was injected onto a BAS Biophase ODS column (250 × 4.6 mm; 5 μm). Mobile phase consisted of (in mm): 50 potassium phosphate, 0.1 EDTA, 0.2 sodium octylsulfonate, and 10% methanol, pH 2.6. l-DOPA peaks were detected with an ESA Coulochem 5100A with a 5011 analytical cell.
DA uptake assay. We measured uptake using3H-labeled DA (Koide et al., 1986) in a physiological medium (in mm: 125 NaCl or choline chloride, 4.8 KCl, 1.2 potassium phosphate, 1.3 CaCl2, 1.2 MgSO4, 1 sodium ascorbate, 5.6 glucose, and 25 HEPES, pH 7.3). Sister cultures of PC12 cells were incubated for 40 min at 37°C in the appropriate physiological medium with or without receptor agonists and antagonists. The cultures were rinsed twice with physiological medium containing 125 mm NaCl for total uptake and 125 mm choline chloride for nonspecific uptake and then incubated in the same medium for 10 min at 37°C. This was replaced with the same medium containing a mixture of3H-labeled and unlabeled DA (0.3 nmol and 0.7 μCi/ml solution) for 5 min at 37°C. The reaction was stopped by placing the cultures on ice, followed by two 1 min rinses with ice-cold medium. Then, 2 ml of 1N NaOH were added to each dish, and the dishes were shaken overnight. An aliquot of the resulting solubilized mixture was mixed with 5 ml of scintillation cocktail, and radioactivity was counted using liquid scintillation. Specific uptake was determined by subtracting the nonspecific count and measuring total protein by the Lowry assay. DA uptake is expressed as picomoles per minute per milligram of protein.
Amperometric electrodes and apparatus. Carbon fiber electrodes were constructed by aspirating 5 μm carbon fibers (Amoco, Greenville, SC) into 1.2 × 0.68 mm glass capillary tubes (A-M Systems, Everett, WA) that were then pulled with a Flaming–Brown micropipette puller (Sutter Instruments, Novato, CA). The electrode tip was dipped into epoxy (Epo-Tek 301; Epoxy Technology, Billerica, MA) and cured at 100°C for 15 hr. The electrodes were back-filled with 3m KCl. Electrode tips were polished at 40° on a beveler (World Precision Instruments, New Haven, CT). Electrode response was tested by cyclic voltammetry in a freshly prepared, nitrogen-bubbled 10 μm DA solution, and those with unstableI–V curves were rejected. The electrode was gently pushed against the cell body with a Huxley-style micropositioner (Newport Instruments, Irvine, CA).
A +700 mV voltage versus Ag–AgCl ground was applied to the carbon fiber electrode using an Axon 200A amplifier (Axon Instruments, Foster City, CA). The output was digitized at 50 kHz and low-pass-filtered at 10 kHz using an internal four-pole Bessel filter. The traces were digitally filtered at 2.5 kHz and analyzed using a locally written Superscope II program (GWI Instruments, Medford, MA). The average background current in the vicinity of the spikes was subtracted from the signal. Spikes were identified if their amplitude was 4.5 times greater than the rms background current, typically 0.3 pA. The number of molecules oxidized at the electrode face was determined by the relation N = Q/nF, whereQ is the charge of the spike, n is the number of electrons transferred (shown to be two for catecholamines when used in a similar experimental configuration; Ciolkowski et al., 1994),N is the number of moles, and F is Faraday’s constant (96,485 coulombs/equivalent).
For electrochemical recordings, physiological incubation medium contained (in mm):150 NaCl, 2 KCl, 1.2 CaCl2, 1 MgCl2, 1 NaH2PO4, 25 glucose, and 10 HEPES, pH 7.3. Drug exposures were for the times indicated at 37°C. The drug-containing medium or control medium was then removed and replaced with physiological incubation medium, and cells were recorded at room temperature (∼25°C). The stimulation medium was composed of (in mm): 72 NaCl, 80 KCl, 6 CaCl2, 21 glucose, and 10 HEPES, pH 7.3. The cells were stimulated by 6 sec superfusion of ∼20 nl of stimulation medium at a distance of 18 μm from the recorded cell (Picospritzer; General Valve, Fairfield, NJ). Low pressure (<7–8 psi) was applied to avoid mechanical stimulation. Five second baseline durations were recorded before stimulation to identify spontaneously active cells. Two stimulations, spaced 20 sec apart, were administered. Interspike intervals and the n of events are only reported for the initial 20 sec period.
Statistics. Values are expressed as mean ± SEM. Kolmogorov–Smirnov statistics, exponential fits, regression analysis, and ANOVA were performed using GB-Stat (Dynamic Microsystems, Silver Springs, MD). Box diagrams were used to display the range of all mean values for the cell populations; median values were indicated as a point within the box, 25 and 75% values of the distribution were indicated by the box edges, and 10 and 90% values were indicated by the ends of the upper and lower lines. Cells that released less than five events within the 20 sec after stimulation were excluded from analysis. χ2 values were determined by the formula [(observed − expected)2/expected] with means as the expected values.
Identification of DA autoreceptor mRNA
There are three known D2-like DA receptors, classified as D2, D3, and D4receptors (O’Dowd et al., 1994). To determine the expression of D2 DA autoreceptors expressed in PC12 cells we used RT-PCR to identify cDNA transcripts, adapting methods used in the striatum (Surmeier et al., 1996). PC12 cells expressed message for D2 and D4 receptors at the predicted molecular weights (D2 long and short forms, 404 and 317 bp; D4, 223 bp) with no detectable D3message (Fig. 1). As a control, we also examined D2-like mRNA in neonatal rat ventral striatum using the same primers. The striatum displayed D3 message (461 bp) in addition to D2 and D4 message.
Effect of autoreceptor activation on TH activity
We recently found that the TH product l-DOPA increases quantal size in PC12 cells by increasing cytosolic DA synthesis (Pothos et al., 1996). Interestingly, TH activity can be modulated by a D2 receptor-mediated pathway (Chiodo et al., 1984;Goldstein et al., 1990; Booth et al., 1994). To directly observe modulation of TH activity by D2 activation, the TH productl-DOPA was measured in the presence of an aromatic acid decarboxylase inhibitor that blocked l-DOPA conversion to DA (Arbogast et al., 1993). The D2 agonist quinpirole (0.5 μm for 40 min) decreased TH activity to 59% of control levels (Fig. 2;F (3,8) = 35.5; p < 0.0001 by one-way ANOVA). The effect of quinpirole was completely blocked by coincubation with the D2 antagonist sulpiride. We chose this level of TH inhibition as appropriate for further study.
Effect of autoreceptor activation on DA uptake
DA uptake in synaptosomes (Krueger, 1990) and Xenopusoocytes transfected with DA transporter (Zhu et al., 1995; Sonders et al., 1997) is regulated by transmembrane potential. Therefore, another possible result of D2 activation could be to modulate uptake by the plasma membrane catecholamine uptake transporter by altering ionic currents. We examined potential D2modulation of DA uptake by measuring uptake of low levels (300 fmol) of tritiated DA in the cultures; this low exposure would not saturate the transporter or significantly affect tonic extracellular DA in the cultures, which is ∼10 nm (Pothos et al., 1996). We found that neither quinpirole nor sulpiride significantly altered accumulation of labeled DA (Fig. 3).
Quantal release from PC12 cells
To directly observe quantal release from undifferentiated PC12 cells, we used amperometric recording after low-pressure superfusion of high-K+ medium. This resulted in current spikes (Fig. 4 A) that were clearly resolved as individual events at an expanded time scale (Fig.4 B). Spikes were characterized by the number of molecules represented (molecules × 103), amplitude (I max), width at half-height (t ½), and the interval between successive events (intspike) for those spikes within 20 sec after stimulation (Table 1). A low-amplitude “foot” was occasionally observed to precede the full current spike, particularly for higher-amplitude events (Fig.4 C), presumably because the events were not disguised in baseline noise. However, in all cases, the detectable feet contributed to <1% of release (Table 1). Such events have been observed in recordings of chromaffin, mast cell, and pancreatic β-cell exocytosis (Chow et al., 1992; Alvarez de Toledo et al., 1993; Zhou and Misler, 1996) and are suggested to represent transmitter release from an intermediate stage of fusion pore formation.
As expected, if release events were independent (Alvarez de Toledo and Fernandez, 1990), there was a good fit of intspike values during the first 1500 msec after stimulation to an exponential decay [Fig. 4 D; y = (0.256)e(−0.0027/t);r 2 = 0.96]. Given that the average spike width at half-height (t ½) in controls is 6.7 msec, the probability that two or more independent events would occur within this period can be estimated as: where t = 6.7 msec; k = rate constant = 0.0027; and τ = 1/k = 370 msec. Therefore, randomly overlapping spikes are rare.
Effect of D2 autoreceptor activation on quantal release
In preliminary liquid chromatography experiments that confirmed the previous findings of Courtney et al. (1991), we found that incubation of PC12 cultures with 0.5 μm quinpirole for 25 min reduced DA release in high-potassium saline to 54 ± 13% of control levels, and that this effect was completely inhibited by 40 μm sulpiride (data not shown).
As an initial approach, we analyzed quantal release by examining the response of individual cells to successive stimulations (80 mm K+, 6 sec) at 0, 5, 15, 25, and 35 min in the presence of normal recording medium or 0.5 μmquinpirole. Although >90% of PC12 cells displayed quantal release after the first stimulation, only 19% (4 of 21 cells recorded) of the control cells that responded to the first stimulation responded to a second stimulation 5 min later. This was further reduced at three subsequent 10 min intervals to 14% (3 of 21) of control levels at 15 min after the first release, to 10% (2 of 21) at 25 min, and to 5% (1 of 21) at 35 min; the decrease was significant at p < 0.0001; χ2 = 44.97. Moreover, the number of events per stimulation decreased in successive stimulations (mean, 40 events per cell at 5 min, 7 events at 15 min, 4 events at 25 min, and 3 events at 35 min; the decrease was significant by ANOVA with repeated measures (F (4,54) = 4.597; p < 0.01). Similar results were found for quinpirole-treated cells (data not shown). We conclude that although the recording configuration is stable, refill of the releasable vesicle pool in PC12 cells is very slow. This contrasts with bovine adrenal chromaffin cells, which, as we and others have observed, continue to exhibit stable release after numerous stimulations (Finnegan et al., 1996). This appears to be because bovine chromaffin cells are densely filled with granules that provide rapid replacement of docked vesicles (Steyer et al., 1997), whereas PC12 cells have a smaller presence of mobile granules, requiring increased time for replenishment of the releasable pool (Burke et al., 1997).
An alternate successful approach was to compare bins of large groups of quanta from multiple cells after single applications of secretagogue, particularly because the quantal size distribution was found to fit a lognormal distribution. To determine the population distribution of quantal parameters in PC12 cells, we examined 2316 quantal events elicited from 48 control cells (Table 1). As reported previously, untransformed PC12 quanta and other amperometrically detected quanta do not show a normal size distribution but are skewed to the right (Fig.5 H, left). Earlier studies have shown that either a cubed root or log transformation of the quantal amplitudes of postsynaptically recorded quanta (Van der Kloot, 1991) or amperometrically recorded quantal sizes (Finnegan et al., 1996) results in a normal distribution. The suggested rationales are that (1) vesicle diameters are distributed normally; therefore, if transmitter filling is proportional to volume, the population distribution should produce a cubed root normal distribution (Bekkers et al., 1990); and (2) if multiplicative deviations from the mean occur, this would produce a lognormal distribution (Van der Kloot, 1991); an example would be if integral differences in the number of uptake transporters per vesicle resulted in multiplicative transmitter accumulation rates. In the present case, we found that the lognormal transformation resulted in a closer fit of the individual values to a normal probability function than the cubed root transformation (Table2). Therefore, the quantal size populations are lognormal and unimodal. Although a small number of values at the extreme edges are not well described by this distribution, they are typically <1% of the events. In the case of larger events that deviate from the lognormal distribution, this may be attributable to overlapping multiple events. In a second statistical test of the normal distribution of quantal sizes, the Kolmogorov–Smirnov (K–S) test of normality showed a close fit of the lognormal distribution to a Gaussian distribution (K–SZ score = 2.0216; p < 0.05).
By comparing binned quantal sizes from multiple cells, we found that quinpirole reduced the average quantal size to 51% of control values (K–S Z = 2.886; p < 0.0001; Fig.5 B, Table 1), close to the inhibition of TH reported above. The distribution of the mean quantal sizes from each cell measured is indicated (Fig. 5 C). Quinpirole was also observed to decrease the maximum amplitude, width at half-height, and number of molecules represented by the foot when detected (Table 1). We reproduced this experiment four additional times and observed quinpirole-induced reduction of quantal size to 49, 54, 62, and 74% of control levels. In an experiment in which cells were exposed to 0.5 μm quinpirole for 150 min rather than 40 min, quantal size was reduced to 31% of control levels (K–S Z = 4.499; p < 0.0001; n = 144 events from 13 control cells; n = 169 events from 25 quinpirole-treated cells).
To examine whether the effect of D2-like activation on quantal size could be blocked, we used the D2-like antagonist sulpiride. Sulpiride (40 μm) alone had no effect on quantal size but blocked the effects of quinpirole (K–SZ = 3.619; p < 0.0001); the combination of quinpirole and sulpiride was not different from controls (Table 1). In preliminary experiments, we found that lower concentrations of sulpiride (2 and 10 μm) produced only partial blockade (data not shown) (the affinity of quinpirole is 3.9 nm for D2 and ∼12 nm for D4; the affinity for sulpiride is 14.7 nm for D2 and 52 nm for D4; O’Dowd et al., 1994).
Quinpirole also had a significant (K–S Z = 2.258;p < 0.0001) effect on the frequency of release events as measured by the interspike intervals (Fig. 5 D; Table 1). The reduction in frequency was also blocked by sulpiride (K–SZ = 2.756; p < 0.0001; Fig.5 E,G). Sulpiride and sulpiride–quinpirole treatments were not significantly different from controls, although sulpiride exposure induced spontaneous release in a fraction of PC12 cells, nearly the only instances of this observed. Of the cells recorded after sulpiride exposure, 23% (9 of 39) showed more than one release event in a 5 sec period before application of the secretagogue; these were excluded from the analysis of interspike intervals (Fig. 5, Table 1). The effect of sulpiride on release suggests that low tonic levels of DA in these cultures may serve to block spontaneous release by D2-like receptor activation.
Effects of l-DOPA on quinpirole action
If quinpirole reduces quantal size by inhibiting TH, the effect should be reversed by the TH product l-DOPA. We exposed PC12 cultures to 50 μm l-DOPA for 30 min, which elevated the average quantal size by 250% (Fig.6 G–J; the total population was increased, K–S Z = 14.628;p < 0.0001) as observed in a previous study (Pothos et al., 1996). The l-DOPA and l-DOPA–quinpirole treatments did not produce different effects. The combination ofl-DOPA and quinpirole increased the average quantal size over controls by 213% (K–S Z = 12.631;p < 0.0001) and the average quantal size over quinpirole-only treatments by 416% (K–S Z = 5.508,p < 0.0001). This effect is consistent with inhibition of l-DOPA synthesis by D2 activation and also demonstrates that quinpirole does not reduce quantal size by blocking vesicular uptake of cytosolic transmitter. Surprisingly,l-DOPA blocked the decreased frequency of release elicited by quinpirole (p < 0.0001; K–SZ = 3.3015; l-DOPA–quinpirole was not different from controls; Fig.6 D–F).
Although it might be assumed that DA itself would have quinpirole-like effects, we found that 50 μm DA actually elevated quantal release (425,400 ± 18,300 molecules; n = 1691 events recorded from 49 cells), to levels similar to 50 μm l-DOPA. We presume that whereas DA itself is an effective agonist at D2 receptors, it has multiple actions including activation of D1-like receptors, which are also present in PC12 cells (Inoue et al., 1992). In addition, it is a substrate for the plasma membrane catecholamine uptake transporter of these cells (Ramachandran et al., 1993) and the vesicular monoamine transporter VMAT1 (Liu et al., 1994) and thus would be loaded into the vesicles. Therefore, DA activation of D2-like receptors does not provide a straightforward approach to analyzing D2-mediated alteration of quantal size, mostly because of the elevation of cytosolic stores of DA. We also found that including sulpiride (40 μm) with DA had no effect on the DA response (460,100 ± 20,400 molecules; n = 1365 events from 30 cells) and did not unmask D2-like mediated effects on quantal size.
Modulation of synaptic neurotransmitter release can be attributable either to effects on the number of quanta released or the quantal size. The activation of a variety of neurotransmitter receptors is known to alter the frequency of quantal release events (Thompson et al., 1993). Although autoreceptors inhibit stimulation-dependent neurotransmitter release from a range of secretory cells, it is unclear whether such effects are attributable solely to a reduction in the number of quanta released, or if reduced quantal size might also play a role. Here we report clear evidence indicating that autoreceptor activation can reduce quantal size.
We are aware of four mechanisms by which D2-like autoreceptors might mediate apparent quantal size: (1) D2activation could decrease the occurrence of simultaneous release events that are misidentified as single quanta; (2) D2 activation could induce morphological changes resulting in an increased distance of release from the electrode, causing the spikes to appear smaller because of diffusion; (3) D2 activation could promote release of a subpopulation of vesicles with lower levels of transmitter; and (4) by inhibiting DA synthesis or vesicular accumulation, autoreceptors could reduce the total amount of neurotransmitter sequestered within vesicles and hence the number of molecules released during the expression of the fusion pore.
Our results suggest that overlapping events do not explain altered quantal size, because the exponential decay of the distribution of the interspike intervals suggests that <2% of spikes in the control preparations would overlap. The results are also inconsistent with release from more distal sites. In the protocol used in this study, the electrode was placed directly on the cell surface. If amperometric current spikes were not directly under the electrode, they would have shorter amplitudes and increased durations because of DA diffusion (Schroeder et al., 1992); however, quantal release from quinpirole-treated cells exhibited shorter amplitudes and decreased duration. Release of a smaller subpopulation of vesicles is inconsistent with the unimodal lognormal distribution of the populations. Although PC12 cells have both small (∼50 nm diameter) and large (∼150 nm diameter) vesicles, the small vesicles typically contain acetylcholine rather than DA (Schubert et al., 1980; Bauerfeind et al., 1993), which would not be detected. Our data support the hypothesis that quinpirole decreases quantal size by reducing the amount of DA packaged and released per vesicle. A difference in transmitter packaging is suggested by the leftward shift of the cubed roots of quantal sizes, which is consistent with a reduction of the amount of DA per vesicle volume. Moreover, this finding is congruent with decreased release from an entire culture, as previously reported (Courtney et al., 1991), confirming a genuine reduction of DA release.
Mechanism of altered quantal size
A potential mechanism by which D2-like receptor activation might alter the amount of DA packaged per vesicle is by modulation of the activity of VMAT1. VMAT1 has been reported to be inhibited by cAMP-dependent phosphorylation (Nakanishi et al., 1995); however, the effect we observed is in the opposite direction of that which would be predicted after D2 activation, which decreases intracellular cAMP (Goldstein et al., 1990).
An alternate possibility stems from the idea that, because vesicular transmitter accumulation is dependent on the transvesicular electrochemical gradient and substrate concentration gradient, D2 receptor activation might reduce the available pool of cytosolic DA. We examined two mechanisms that could underlie altered cytosolic pools, decreased reuptake and decreased synthesis. We found that although D2 activation has been found to modulate DA reuptake (McElvain and Schenk, 1992; Meiergerd et al., 1993; Cass and Gerhardt, 1994; Wieczorek and Kruk, 1994), under the conditions used, quinpirole did not inhibit the uptake of labeled extracellular DA by the PC12 catecholamine transporter.
The second mechanism addressed was suggested by a large body of evidence suggesting that D2 receptor activation inhibits DA synthesis by blocking cAMP-dependent PKA-mediated phosphorylation of TH. PKA-mediated phosphorylation of the serine 40 residue activates TH, whereas reduced PKA activity after D2 autoreceptor activation causes a reduction in ser40 phosphorylation (Harada et al., 1996). The reduced phosphorylation of ser40 in turn increases binding affinity of catecholamines to the iron moiety of TH resulting in a feedback inhibition of enzymatic activity (Griffith and Schulman, 1988;Haycock and Haycock, 1991; Daubner et al., 1992; Kumer and Vrana, 1996). Consistently, D2 autoreceptor effects have long been known to inhibit DA synthesis (Chiodo et al., 1984; Goldstein et al., 1990; Booth et al., 1994). The quinpirole exposures we used inhibited TH activity to nearly the same level as the reduction in quantal size. Therefore, we suggest that D2 activation decreases quantal size by inhibiting TH activity and reducing cytosolic DA available for vesicle accumulation.
We also noted a decreased quantal release frequency after D2 activation that was blocked by sulpiride. Surprisingly,l-DOPA also blocked the quinpirole-mediated decrease in frequency. The mechanism by which l-DOPA produces this response is not clear. It is possible that, rather than a reduction in the rate of exocytosis, vesicle fusion after quinpirole proceeds at the normal rate but includes quanta with small amounts of transmitter that are below detection. These empty vesicles might then be filled if elevated cytosolic DA were available after l-DOPA. However, the ability of sulpiride to promote spontaneous release without a change in quantal size is consistent with a D2 influence on vesicle exocytosis. Additional potential explanations for thel-DOPA reversal of the effects of quinpirole on frequency are that very high levels of l-DOPA act as an antagonist at D2-like receptors or that upregulated DA afterl-DOPA provides additional effects. An elegant approach to this in the future may be provided by simultaneous patch capacitance and amperometric recording of monoquantal release (Albillos et al., 1997).
Cellular and behavioral effects
Recent work indicates that the quantal size of catecholamine release can be regulated by various pharmacological interventions. (1) Transmitter accumulation is dependent on the magnitude of the vesicular electrochemical gradient. Amphetamine reduces the quantal size of DA release by collapsing the vesicular proton gradient (Sulzer et al., 1995). (2) Quantal size can be increased by elevating cytosolic neurotransmitter levels, and l-DOPA increases the quantal size of DA release (Pothos et al., 1996, 1998). (3) We now provide evidence that autoreceptor-mediated inhibition of neurotransmitter synthesis decreases the quantal size of DA release. More generally, although quantal size is often modeled as invariant (cf. Van der Kloot, 1991; Frerking et al., 1995) and is rarely invoked as a mechanism of synaptic plasticity, we and others have found that regulation of vesicular transporters (Song et al., 1997) and exposure to neurotrophic factors (Wang et al., 1995; Pothos et al., 1998) alters quantal size.
These findings may provide insights into the synaptic changes induced by psychostimulant and antipsychotic drugs. The psychostimulants cocaine and amphetamine elevate extracellular DA levels and activate mesolimbic and mesocortical D2 autoreceptors. Therefore, psychostimulants may reduce appropriate stimulation-dependent perisynaptic DA input to postsynaptic targets by decreasing quantal size (Pothos and Sulzer, 1998) while increasing stimulation-independent overflow of extrasynaptic DA. This has important implications for the signal-to-background ratio of DA input; indeed, a precise level of DA input is required for maintenance of working memory at mesocortical connections, and it is striking that a feature of amphetamine psychosis is an impairment in working memory (Williams and Goldman-Rakic, 1995).
Classically, D2-like autoreceptors have been classified as displaying impulse-regulating or synthesis-modulating responses. The data in this study are consistent with the possibility that both impulse-regulating and synthesis-modulating responses occur in the same cell population. D2 autoreceptors of mesolimbic DA neurons, perhaps including D3 autoreceptors (Meller et al., 1993;Rivet et al., 1994), inhibit firing by hyperpolarization and increasing input resistance (Grace and Bunney, 1984a) in addition to inhibiting DA synthesis. In the high-affinity state, D2 receptors have a binding constant of ∼7.5 nm (O’Dowd et al., 1994). Because background extracellular striatal DA levels are ∼4 nm (Parsons and Justice, 1992), there may be significant inhibition of DA release at tonic levels, and indeed the D2antagonist sulpiride alone elevates DA release in awake, behaving animals (Tanaka et al., 1992). Antipsychotic drugs that act as D2-like antagonists elevate extracellular striatal DA levels and increase DA cell firing rates after acute administration but reduce extracellular DA and firing rate after longer term exposure. A decreased quantal size would be expected to reduce stimulation-dependent diffusional overflow of DA from the release site. Because monoamine synapses provide extrasynaptic transmitter overflow (Garris et al., 1994), it is likely that alteration of quantal size plays a more important role for monoamine release than fast-acting transmitters.
In summary, these findings directly demonstrate that activation of neurotransmitter receptors alters quantal size; this apparently occurs by inhibition of transmitter synthesis. If a similar mechanism exists in neurons, alteration of quantal size by autoreceptor activation would appear to play a role in modulating synaptic plasticity.
This work was funded by National Institute on Drug Abuse (NIDA) Grants 07418 and 10154 (D.S.), an NIDA Shannon Award (D.S.), the Aaron Diamond Foundation (E.N.P., D.S.), National Alliance for Research on Schizophrenia and Depression (NARSAD) (E.N.P., D.S.), and the Parkinson’s Disease Foundation. E.N.P. is an Aaron Diamond Foundation Fellow and recipient of a 1995 NARSAD Young Investigator Award. We are grateful to Dr. Lloyd Greene for providing parent PC12 cultures, Dr. Tjoeben Nygaard for use of his thermal cycler, Drs. James Surmeier and Zhen Yan for advice on RT-PCR techniques, Dr. Michael Neystat for advice on molecular techniques, and Dr. Steven Siegelbaum for discussion of this manuscript.
Correspondence should be addressed to David Sulzer, Black Building 305, 650 West 168th Street, Columbia University, New York, NY 10032.