G protein βγ subunits mediate presynaptic inhibition of transmitter release from rat superior cervical ganglion neurones in culture

Abstract

The activation of presynaptic G protein-coupled receptors (GPCRs) is widely reported to inhibit transmitter release; however, the lack of accessibility of many presynaptic terminals has limited direct analysis of signalling mediators. We studied GPCR-mediated inhibition of fast cholinergic transmission between superior cervical ganglion neurones (SCGNs) in culture. The adrenoceptor agonist noradrenaline (NA) caused a dose-related reduction in evoked excitatory postsynaptic potentials (EPSPs). NA-induced EPSP decrease was accompanied by effects on the presynaptic action potential (AP), reducing AP duration and amplitude of the after-hyperpolarization (AHP), without affecting the pre- and postsynaptic membrane potential. All effects of NA were blocked by yohimbine and synaptic transmission was reduced by clonidine, consistent with an action at presynaptic α2-adrenoceptors. NA-induced inhibition of transmission was sensitive to pre-incubation of SCGNs with pertussis toxin (PTX), implicating the involvement of Gα_i/oβγ subunits. Expression of Gα transducin, an agent which sequesters G protein βγ (Gβγ) subunits, in the presynaptic neurone caused a time-dependent attenuation of NA-induced inhibition. Injection of purified Gβγ subunits into the presynaptic neurone inhibited transmission, and also reduced the AHP amplitude. Furthermore, NA-induced inhibition was occluded by pre-injection of Gβγ subunits. The Ca²⁺ channel blocker Cd²⁺ mimicked NA effects on transmitter release. Cd²⁺, NA and Gβγ subunits also inhibited somatic Ca²⁺ current. In contrast to effects on AP-evoked transmitter release, NA had no clear action on AP-independent EPSPs induced by hypertonic solutions. These results demonstrate that Gβγ subunits functionally mediate inhibition of transmitter release by α2-adrenoceptors and represent important regulators of synaptic transmission at mammalian presynaptic terminals.

Activation of presynaptic GPCRs represents an important regulatory mechanism for neurotransmitter release (Miller, 1998). Presynaptic inhibition by GPCRs has been proposed variously to involve an inhibition of Ca²⁺ channels, an activation of K⁺ channels, or a direct action on the exocytotic release machinery itself (Thompson et al. 1993; Wu & Saggau, 1997; Miller, 1998). Common to all of these mechanisms is the proposed mediation by Gβγ subunits, liberated from Gαβγ trimers by receptor occupation (Pierce et al. 2002). However, the functional contribution of Gβγ pathways to transmitter release from presynaptic terminals remains inferred largely from recordings of postsynaptic events. Due mainly to the lack of accessibility of most presynaptic terminals to exogenous peptides, few direct studies on Gβγ mechanism of action have been attempted. One exception amongst mammalian synapses is the calyx of Held preparation, in which unusually large, glutamatergic synapses occur in the medial nucleus of the trapezoid body (Forsythe, 1994). In these terminals, injection of Gβγ subunits was shown to inhibit Ca²⁺ currents and partially occlude GABA_B receptor agonist effects (Kajikawa et al. 2001). However, alternative mechanisms have been proposed. Purified Gβγ subunits have been introduced directly into lamprey reticulospinal–motoneurone synapses and shown to inhibit evoked transmitter release (Blackmer et al. 2001). Here, Gβγ subunits were proposed to interact directly with the exocytotic release machinery, independently of effects on Ca²⁺ channels (Blackmer et al. 2001; Takahashi et al. 2001). This mechanism was proposed to involve the binding of Gβγ subunits to syntaxin and/or SNAP-25 which may prevent the fusion of synaptic vesicles (see also Spafford & Zamponi, 2003). In addition, injection of GTPγS, a non-hydrolysable GTP analogue, into squid giant axon terminals was reported to inhibit synaptic transmission with few accompanying effects on intracellular Ca²⁺ levels or Ca²⁺ influx (Hess et al. 1993). Presynaptic inhibition by GPCRs independent of effects on Ca²⁺ influx has parallels at many central synapses, where inhibition of spontaneous, AP-independent transmitter release can be quantified as effects on ‘miniature’ postsynaptic events (Miller, 1998); such events are also mediated potentially by Gβγ subunits (Harvey & Stephens, 2004). Overall, these studies suggest potential differences in presynaptic Gβγ-mediated signalling mechanisms for the inhibition of transmitter release.

Here, we sought to extend studies on the functional consequences of direct, presynaptic injection of Gβγ subunits by examining effects on transmitter release in a mammalian model synapse for fast cholinergic transmission formed by SCGNs in culture. We took advantage of the unique, large presynaptic cells which are amenable to injection of exogenous peptides; these peptides can diffuse rapidly to nerve terminals contacting synaptically with adjacent neurones (Mochida et al. 1996). We show that α2-adrenoceptor inhibition of AP-evoked transmitter release is mediated by Gβγ subunits and also parallels the inhibition of somatic voltage-dependent Ca²⁺ channels (VDCCs); in contrast, α2-adrenoceptors had no clear effects on AP-independent transmitter release. These data demonstrate directly that Gβγ subunits act as important regulators of GPCR-mediated inhibition of synaptic transmission at mammalian presynaptic terminals.

Methods

Synaptic transmission between SCGNs

Wistar ST rats were decapitated on postnatal day 7 under diethylether anaesthesia according to the guidelines of the Physiological Society of Japan. SCGNs were isolated and maintained in culture for 5–6 weeks in a growth medium of 84% Eagle's minimal essential medium, 10% fetal calf serum, 50% horse serum, 1% penicillin–streptomycin (all from Gibco Industries Inc., Langley, OK, USA) and 25 ng ml⁻¹ nerve growth factor (2.5 S, grade II; Alomone, Jerusalem, Israel) as described (Mochida et al. 1995). Conventional intracellular recordings were made from two neighbouring neurones using microelectrodes filled with 1 m potassium acetate (KAc; 70–90 MΩ). Neurones were superfused with modified Krebs solution containing (mm): NaCl, 136; KCl, 5.9; CaCl₂, 5.1; MgCl₂, 1.2; glucose, 11; and Na-Hepes, 3; pH 7.4 with NaOH. EPSPs were recorded from one of the neurones, while APs were generated in the other neurone by passage of current (1–1.5 nA of 15 ms square pulse) through an intracellular recording electrode. The α2-adrenoceptor agents NA, clonidine or yohimbine (all from Sigma, St Louis, MO, USA), or CdCl₂ (Kanto Chemical Co. Inc., Tokyo, Japan) were dissolved in modified Krebs solution and superfused or drop-applied. In experiments to block the involvement of Gα_i/oβγ, SCGNs were incubated with 500 ng ml⁻¹ PTX (Sigma) in the growth medium for 12–36 h at 37°C, 95% O₂/ 5% CO₂. Gα transducin cDNA (0.25 mg ml⁻¹), a kind gift of Professor A. Dolphin (UCL, London, UK), was dissolved in 150 mm KAc, 5 mm Mg-ATP, 10 mm Hepes, pH 7.3 and injected into SCGN nuclei from a microglass pipette as described previously (Mochida et al. 2003). Dextran fluorescein (10 kDa, Molecular Probes, Eugene, OR, USA) was introduced via the pipette solution to monitor cDNA entry into nuclei. After incubation at 37°C, 95% O₂/ 5% CO₂ in the growth medium (see above) for 24–48 h, injected neurones were identified with an inverted microscope equipped with an epifluorescent unit. Gβγ subunits, a kind gift of Dr J. Pitcher (UCL), purified from bovine brain as described by Casey et al. (1989), or control buffer (20 mm Tris pH 8.0, 100 mm NaCl, 2 mm DTT, 0.1% Lubrol (Serva, Heidelberg, Germany)), were introduced into the presynaptic cell body by diffusion from a suction glass pipette for 2–3 min (GC150-F10, Harvard Apparatus, Kent, UK; 15–20 MΩ tip resistance). Fast Green FCF (5%, Sigma) was included in the injection solution to confirm entry into the presynaptic cell. AP-independent transmitter release was induced by puff application of 0.5 m sucrose onto synaptically coupled SCGNs (Mochida et al. 1998). Electrophysiological data were collected and analysed using software written by the Late Dr Ladislav Tauc (CNRS, France). Some recordings were monitored with a strip chart recorder (Nihon Kohden, Japan).

Patch clamp recordings in SCGNs

Whole-cell electrophysiological recordings were made from individual SCGNs, cultured for 3–4 weeks, using borosilicate glass electrodes (GC150-TF10, Harvard Apparatus) which, when filled with an intracellular solution containing (mm): caesium aspartate, 140; EGTA, 5; MgCl₂, 2; CaCl₂, 0.1; K₂ATP, 2; and Hepes, 10; pH 7.2 with CsOH, had tip resistances between 3 and 5 MΩ. Neurones were superfused with modified Krebs solution containing (mm): TEACl, 136; CsCl, 5.9; MgCl₂, 1.2; Na-Hepes, 3; and 3 μm tetrodotoxin (Wako Chemical Co., Osaka, Japan) (pH 7.4 with NaOH), together with 1 mm BaCl₂. Whole-cell currents were acquired using Clampex software via an Axopatch 200A amplifier (Axon Instruments, Union City, CA, USA). Data were filtered at 2–5 kHz, digitized at 5–20 kHz and analysed using pCLAMP and Origin (Microcal, Northampton, MA, USA) software. Series resistance was compensated between 50 and 70%. Current records are shown following leak and capacitance current subtraction (P/4 protocol) and correction for liquid junction potentials.

All data are given as mean values ± s.e.m. and statistical significance was determined using Student's paired or unpaired t tests as appropriate.

Results

α2-Adrenoceptor activation inhibits synaptic transmission in cultured SCGNs

To investigate the effects of activation of GPCRs on synaptic transmission in SCGNs cultured for 5–6 weeks, we examined the actions of NA, an agonist whose somatic actions have been characterized extensively in SCGNs in short-term culture (e.g. Schofield, 1990), and which is thought to modulate cholinergic transmission in the SCG (Libet, 1979). Transmission was examined by monitoring pre- and postsynaptic events in synaptically coupled SCGNs (Figs 1, 2, 3, 4). Bath application of NA caused a clear reduction in evoked EPSP amplitude (Fig. 1A and D (post)). NA-induced EPSP reduction was accompanied by presynaptic effects. As illustrated in Fig. 1A (pre), the amplitude of the AHP in the AP waveform recorded in presynaptic cells was reduced reversibly by 10 μm NA (from 15.6 ± 1.4 to 10.1 ± 1.0 mV, n = 3, P < 0.05; Fig. 1Ca). The time course of NA-induced reduction in the AHP amplitude is illustrated in a representative chart recorder trace in Fig. 1D (pre), which follows membrane potential and associated AHP in response to presynaptic APs. NA typically caused a rapid reduction in AHP amplitude, which recovered slowly. NA (10 μm) also reduced reversibly the AP duration at half of peak amplitude (from 2.47 ± 0.09 to 1.73 ± 0.15 ms, n = 3, P < 0.05; Fig. 1Cb), as illustrated in a typical expanded time course of the AP waveform in Fig. 1B.

Figure 1. Pre- and postsynaptic effects of noradrenaline (NA) and Cd²⁺ in SCGNs.

A, action potential (AP) waveform recorded in the presynaptic superior cervical ganglion neurone (SCGN; pre) and corresponding evoked excitatory postsynaptic potential EPSP (post) under different experimental conditions. NA and Cd²⁺ reduced EPSP and after-hyperpolarization potential (AHP) amplitude and also abolished a recurrent EPSP (*) in this example. B, expanded AP waveform corresponding to conditions in A. Arrows indicate AP duration. C, summary bar graphs illustrating mean (± s.e.m.) decrease in AHP amplitude (a) and AP duration by (b) 100 μm Cd²⁺ and 10 μm NA. Con, control. *P < 0.05. D, example chart recorder traces showing pre- and postsynaptic effects of Cd²⁺ (upper panel) and NA (lower panel). EPSPs were evoked at a frequency of 0.1 Hz.

Figure 2. NA effects are mediated by α2-adrenoceptors.

A, example chart recorder traces showing pre- and postsynaptic effects of NA (top panel), block of NA action by yohimbine (middle panel) and recovery of NA effects (bottom panel) recorded in the same cell. B, dose-related effects of NA on evoked EPSP amplitude; smoothed values of the normalized average EPSP with moving average algorithm plotted against time are shown. C, effects of clonidine on EPSP amplitude. Circles: averaged values of normalized EPSP. Continuous line: smoothed values of the normalized average EPSP (C and D). D, block of NA effects on EPSP amplitude by yohimbine (grouped data for yohimbine 1 μmversus 10 μm NA (3); 2 μmversus 5 μm NA (3); and 2 μmversus 10 μm NA (3). In A–D, α2-adrenoceptor agents were applied during indicated horizontal bars. Numbers in parentheses indicate number of replicants (n). EPSPs were evoked at a frequency of 0.1 Hz. E, summary bar graph illustrating mean (± s.e.m.) decrease in EPSP amplitude at 7 min after α2-adrenoceptor agent application. *P < 0.05 in comparison with effects of 10 μm NA in the same batch of cells (shown in D).

Figure 3. α2-Adrenoceptor effects are mediated by Gβγ subunits.

A, effects of 24–48 h pre-incubation with pertussis toxin (PTX) on NA-induced reduction of evoked EPSP amplitude. B, effects of 24 and 48 h expression of Gα transducin on NA-induced reduction of evoked EPSP amplitude. In A and B, NA was applied during indicated horizontal bars by superfusion (A) or drop-application (B). Numbers in parentheses indicate number of replicants (n). EPSPs were evoked at a frequency of 0.1 Hz. Circles: averaged values of normalized EPSP. Continuous lines: smoothed values of the normalized average EPSP. C, summary bar graph illustrating the effect of PTX and Gα transducin on mean (± s.e.m.) NA-induced decrease in EPSP amplitude. Effects of 10 μm NA measured at 7 min (PTX) and 25 min (Gα transducin). *P < 0.01 between each of the data sets indicated.

Figure 4. Pre- and postsynaptic effects of Gβγ subunits in SCGNs.

A, effects of presynaptic injection of Gβγ subunits (10 μm in the injection pipette) on evoked EPSP amplitude (n = 6). Inset: example EPSP recording traces before and after Gβγ subunits injection. EPSPs were evoked at a frequency of 0.1 Hz. Circles: averaged values of normalized EPSP. Continuous line: smoothed values of the normalized average EPSP (A and C). B, effects of Gβγ subunits on AP waveform recorded in the presynaptic cell. Example chart recorder traces showing effects of Gβγ on AHP before (a), during (b) and after (c) protein injection (top, middle panels), and AP duration (bottom panel). C, effects of NA (drop-applied during bar) on evoked EPSP amplitude in the presence (+Gβγ, 10 μm in the injection pipette, n = 6) or absence of Gβγ subunits (−Gβγ, n = 6). D, summary bar graph illustrating effects of Gβγ subunits, Gβγ in the presence of 1 mm Ba²⁺ added to the modified Krebs solution, and heat-denatured Gβγ controls, on mean evoked EPSP amplitude and effects of Gβγ on mean NA-induced decrease in EPSP amplitude (± s.e.m.). Effects of 10 μm NA measured at 25 min. *P < 0.001 between each of the data sets indicated.

The actions of NA on AP duration and AHP may be characteristic of agents that reduce Ca²⁺ entry into SCGNs in culture (Mochida & Kobayashi, 1986a, b). Therefore, we compared the effects of NA with 100 μm Cd²⁺, a general blocker of VDCCs, including Ca_V2.2 (N-type) channels which predominantly mediate cholinergic transmission in SCGNs (Mochida et al. 1995). Cd²⁺ (100 μm) greatly reduced EPSP amplitude (Fig. 1A and D (post)). Cd²⁺ strongly and rapidly reduced AHP amplitude from 14.7 ± 2.2 to 6.79 ± 1.4 mV (n = 5, P < 0.05) (Fig. 1A, Ca and D (pre)). Cd²⁺ also reduced AP duration at half of peak amplitude from 2.85 ± 0.22 to 1.93 ± 0.18 ms (n = 5, P < 0.05) (see Fig. 1B and Cb). All actions of Cd²⁺ were readily reversible. Together, the presynaptic actions seen were consistent with a reduction of Ca²⁺ influx into SCGNs. NA caused a reduction in EPSP amplitude over a range of concentrations: 200 nm (−13.8 ± 7.4%, n = 4), 1 μm (−25.1 ± 9.0%, n = 5) and 10 μm NA (−39.2 ± 6.1%, n = 10) gave a dose-related inhibition (Fig. 2A, B and E). Higher concentrations of NA (50–200 μm) caused no further inhibition (data not shown). The selective α2-adrenoceptor agonist, clonidine (10 μm), also produced a clear reduction in EPSP amplitude (−27.7 ± 6.8%, n = 6) (Fig. 2C and E). The NA-induced decrease in EPSPs was significantly reduced by the α2-adrenoceptor selective antagonist yohimbine (−12.9 ± 8.0%, P < 0.05) (Fig. 2A, D and E). Yohimbine also reduced the presynaptic effects of NA; for example, the NA-induced reduction in AHP amplitude was reversibly blocked by yohimbine (Fig. 2A (pre)). Together, these data were consistent with the activation of presynaptic α2-adrenoceptors causing an inhibition of cholinergic transmission in SCGNs.

Gβγ subunits mediate α2-adrenoceptor-induced inhibition of synaptic transmission

Agonist occupation of GPCRs releases GDP from a specific receptor-coupled Gα subunit which forms an activated Gα-GTP complex, resulting in release of free Gβγ dimers (Pierce et al. 2002). To examine the G protein contribution to NA inhibition, SCGNs were incubated with the G protein uncoupling agent PTX (500 ng ml⁻¹, 24–48 h) (Fig. 3A and C). In PTX-treated synapses, the NA-induced reduction of EPSP amplitude was significantly attenuated to −14.5 ± 3.5% (n = 6, P < 0.01), implicating the involvement of Gα_i/oβγ subunits. We next examined the effects of sequestering endogenous Gβγ subunits by expression of the Gβγ-binding protein Gα transducin, in individual presynaptic cells (Fig. 3B and C). In synapses examined 24 h after nuclear injection of Gα transducin cDNA, 10 μm NA still produced a clear reduction in EPSP amplitude (−35.0 ± 4.5%, n = 5), not significantly different to that seen in untreated cells. In contrast, in synapses examined at least 48 h after the cDNA injection, 10 μm NA-induced reduction in EPSP amplitude was significantly attenuated to −9.0 ± 3.8% (n = 6, P < 0.01) (Fig. 3B and C). This distinction may be due to the time taken for a sufficient concentration of Gα transducin to be translocated into the terminals and fully sequester Gβγ subunits. Overall, these data show that Gα transducin produced a time-dependent attenuation of NA action and implicate Gβγ subunits as a predominant mediator of NA action. We tested this hypothesis by injecting purified Gβγ subunits (10 μm in the injection pipette) directly into the presynaptic cell of synaptically coupled SCGNs. Gβγ, injected following a minimum of 20 min stable recording of evoked EPSPs, caused a clear reduction in EPSP amplitude (−29.0 ± 7.2% at 50 min after injection, n = 6) which took more than 30 min to peak and lasted for at least 2 h of recording (Fig. 4A and D). In contrast, injection of heat-denatured Gβγ subunits had only minor effects on EPSP amplitude, with a level of inhibition (−5.5 ± 2.4% at 50 min after injection, n = 5), significantly less than that seen with Gβγ (Fig. 4D). To investigate any possible contribution of presynaptic K⁺ channels to inhibition of transmitter release, Gβγ effects were examined in the continued presence of 1 mm Ba²⁺ (a general blocker of K⁺ channels, including G protein-coupled inward rectifier channels, which are potential targets of Gβγ-mediated activation, Clapham & Neer, 1997). In modified Krebs solution supplemented with Ba²⁺, injection of Gβγ subunits (10 μm in the injection pipette) caused a clear reduction in EPSP amplitude (−30.5 ± 9.3% at 50 min after injection, n = 5), which was no different to that seen for Gβγ in normal extracellular solution (P = 0.51) (Fig. 4D).

Gβγ also had effects on the AP waveform. As illustrated in Fig. 4B, Gβγ caused a reduction in AHP (b*) during and immediately following injection; AHP amplitude was reduced from 14.9 ± 1.6 to 11.5 ± 1.8 mV at 3 min after starting the injection (n = 7, P < 0.01). A similar reduction was not seen following injection of the carrier buffer (8.6 ± 0.2 to 8.4 ± 0.7 mV at 3 min after starting the injection, n = 3). In contrast to actions on AHP amplitude, Gβγ injection had no clear effects on AP duration at this time point (Fig. 4B, bottom traces) (1.6 ± 0.1 to 1.7 ± 0.2 ms, n = 7).

To confirm the contribution of Gβγ to NA action, we examined the effects of pre-injection of Gβγ subunits on NA-induced inhibition (Fig. 4C and D). Presynaptic cells were injected with Gβγ 3–5 min after starting EPSP recordings, and then, 50–60 min after injection, 10 μm NA was applied. In these synapses, the inhibitory effects of 10 μm NA on EPSP amplitude were now virtually abolished (−4.0 ± 3.8%, n = 6). This lack of effect was highly significant in comparison to time-matched controls in which 10 μm NA was added in the absence of Gβγ injection (−30.0 ± 6.1%, n = 6; P < 0.001). The occlusion of NA effects by Gβγ pre-injection is consistent with α2-adrenoceptor inhibition of synaptic transmission in SCGNs being mediated by Gβγ subunits.

Gβγ subunits and NA inhibit somatic Ca²⁺ currents

Synaptic transmission in SCGNs is mediated predominantly by presynaptic Ca_V2.2 (N-type) channels (Mochida et al. 1995), but direct measurement of Ca²⁺ currents in presynaptic terminals of SCGNs is technically difficult. Therefore, we examined the effects of NA and Gβγ subunits on somatic whole-cell Ca²⁺ currents in SCGNs in 3–4 weeks culture. A similar approach has been used to measure somatic Ca²⁺ currents in SCGNs in short-term culture (typically 1–4 days), leading to extensive characterization of an α2-adrenoceptor pathway (Horn & McAfee, 1980; Galvan & Adams, 1982; Schofield, 1990; Hille, 1994; Herlitze et al. 1996; Ikeda, 1996; Garcia et al. 1998; Delmas et al. 1998, 1999). However, it has been shown that the neurotransmitter released predominantly by SCGNs changes as time in culture increases. Thus, cells release NA initially (see Koh & Hille, 1997), and then, after 1–2 weeks, they switch to releasing acetylcholine predominantly (Mochida et al. 1994). Therefore, we investigated if a membrane-delimited pathway that mediates the modulation of Ca²⁺ currents was active at newly formed cholinergic presynaptic terminals in the long-term SCGN cultures used here.

The actions of the inorganic ion Cd²⁺, which caused a dramatic reduction in EPSP amplitude (Fig. 1A and D (post)), may suggest a role for presynaptic VDCCs in the inhibition of transmitter release. Cd²⁺ (100 μm) also abolished somatic Ca²⁺ current completely (Fig. 5A). Furthermore, 10 μm NA also caused a clear reduction in current amplitude of −32.6 ± 3.7% at +10 mV (n = 12) (see Fig. 5B). A major characteristic of agonist-induced G protein inhibition of VDCCs is a slowing in current activation kinetics, associated with Gβγ subunits binding to the channel to promote a ‘reluctant’ gating state (Bean, 1989; Patil et al. 1996). Therefore, current activation was quantified by a simple time-to-peak measurement: control values (12.4 ± 1.7 ms at +10 mV, n = 12) were significantly increased by 10 μm NA (22.4 ± 5.3 ms at +10 mV, n = 12; P < 0.05), as can be seen in Fig. 5B. Higher concentrations of NA promoted further inhibition, as also illustrated in Fig. 5B. A further characteristic of G protein inhibition of VDCCs is that it can be reversed partially by a large, depolarizing prepulse, described originally as a disruption of the voltage-dependent interaction between the G protein and the VDCC (Grassi & Lux, 1989; Bean, 1989). In cells subject to NA inhibition, application of a +120 mV prepulse caused an increase in whole-cell current amplitude and also an acceleration of activation kinetics (Fig. 5C and E). In the presence of NA, a clear increase in the ratio of peak current amplitude following the prepulse (+PP) to that before (−PP) was seen (+PP: −PP = 1.36 ± 0.08 at 0 mV, n = 5), representing a significant increase above the ratio in control cells (+PP: −PP = 1.01 ± 0.02 at 0 mV, n = 5, P < 0.05) (Fig. 5C, Db and E). The lack of facilitation seen in control cells suggests a lack of tonic modulation by endogenous Gβγ subunits. We also examined NA actions on somatic Ca²⁺ current in cells expressing Gα transducin. The NA-induced reduction of Ca²⁺ current amplitude was now attenuated significantly to −10.4 ± 3.2%, n = 5, P < 0.05. These data further implicate Gβγ subunits as mediators of NA effects and are in agreement with the reported attenuation of NA action on somatic Ca²⁺ currents by Gα transducin in SCGNs in short-term culture (Delmas et al. 1999). Overall, we saw no clear distinction between Gα transducin effects on somatic Ca²⁺ currents after 24 or 48 h, in comparison to actions on transmitter release. These findings suggest that extrapolation of events at the cell soma to those affecting presynaptic release may rely on a sufficient concentration of protein reaching synaptic terminals. Finally, introduction of Gβγ subunits (2–10 μm in the patch pipette) directly into SCGNs caused a clear slowing in the time course of activation of whole-cell current (Fig. 5Da). Time-to-peak was slowed dramatically to 168 ± 15 ms at +10 mV (n = 6, P < 0.001 compared to control currents). Application of a +120 mV prepulse revealed a large Gβγ-induced reduction in current amplitude and time-to-peak (Fig. 5Da and E), with a significant increase in peak current ratio to 1.47 ± 0.15 at 0 mV, n = 5 (P < 0.05 compared to control ratio) (Fig. 5E). Overall, these data indicate that Cd²⁺, NA and Gβγ caused clear reductions in somatic Ca²⁺ currents in SCGNs in long-term culture. These actions have close parallels with the effects of the same agents on presynaptic transmitter release described above.

Figure 5. Effects of Cd²⁺, NA and Gβγ subunits on somatic Ca²⁺ currents in SCGNs in 5–6 week culture.

Example effect of Cd²⁺ (A) and NA (B) on whole-cell Ca²⁺ currents. Currents were activated by a depolarizing step to +10 mV from a holding potential (V_H) of −70 mV. C, effect of a conditioning prepulse to +120 mV (dotted line, +PP) on NA-induced inhibition of whole-cell Ca²⁺ currents. D, example effect of a conditioning prepulse to +120 mV (+PP) on (a) Gβγ (6.7 μm in the patch pipette)-inhibited Ca²⁺ currents, and (b) control currents. Currents were activated by a depolarizing step to +10 mV; V_H=−70 mV. E, summary bar graph illustrating effect of NA (n = 5) and Gβγ subunits (n = 5) on control (n = 5) peak current amplitude ratio (+PP: −PP at 0 mV), mean (±s.e.m.),*P < 0.05.

NA has no effect on AP-independent EPSPs

A further hypothesis for the mechanism of Gβγ-mediated presynaptic inhibition is that action occurs independently of effects on Ca²⁺ influx through VDCCs, possibly via a direct interaction with the synaptic vesicle fusion machinery (Blackmer et al. 2001). In the SCGN model synapse, Ca²⁺ influx is triggered by generation of APs in the presynaptic SCGN; therefore, we investigated if activation of the identified α2-adrenoceptor pathway had any effect on transmitter release in the absence of APs, which may be consistent with alternative mechanisms of GPCR inhibition. AP-independent EPSPs were stimulated by focal application of hypertonic 0.5 m sucrose solution, puff-applied for 2 s onto SCGNs, in the absence of the passage of current in presynaptic cells (Mochida et al. 1998). Bath application of 10 μm NA, following a minimum of 8 min stable baseline in the sucrose response integral, had no clear effect on AP-independent EPSPs (n = 5 separate cells, Fig. 6A and B). These data suggest that, under these conditions, NA had no clear effect on acetylcholine release in SCGNs in the absence of APs. Overall, although we cannot rule out an action on the synaptic vesicle fusion machinery contributing to NA-induced inhibition of AP-evoked transmission, we find no confirmatory evidence from studies solely on transmitter release performed in the absence of APs.

Figure 6. Effects of NA on AP-independent transmitter release.

A, effects of NA on AP-independent EPSPs induced by 0.5 m sucrose (applied from a local puff pipette during period indicated by bar); raw traces before, during and after bath NA superfusion. B, normalized mean (± s.e.m.) sucrose response integral values illustrating lack of effect of NA (n = 5).

Discussion

Gβγ subunits mediate presynaptic inhibition

Although presynaptic inhibition of synaptic transmission by GPCRs is a well-known phenomenon, there is, as yet, a lack of direct demonstrations that Gβγ subunits functionally mediate inhibition and, also, if common mechanisms exist amongst central synapses. Here, we have taken advantage of the accessibility to exogenous peptides of the large, presynaptic SCGNs that form in culture, to show that modulation of fast cholinergic transmission by presynaptic α2-adrenoceptors is mediated by Gβγ subunits. Accordingly, direct injection of purified Gβγ subunits caused an inhibition of transmission between synaptically coupled SCGNs. Inhibition took more than 30 min to peak and lasted for at least 2 h of recording (Fig. 4A), consistent with Gβγ subunits diffusing from the soma to synaptic terminals to cause inhibition. Studies have concluded that transmitter secretion from SCGNs in culture occurs almost exclusively from varicosities and terminals, rather than the cell body itself (Wakshull et al. 1979; Przywara et al. 1993; Koh & Hille, 1997). Activation of the endogenous α2-adrenoceptor pathway confirmed the role of Gβγ. Accordingly, NA-induced inhibition of evoked transmitter release was (i) sensitive to PTX, (ii) reduced in a time-dependent manner by presynaptic over-expression of Gα transducin, and (iii) occluded by presynaptic injection of Gβγ subunits. At concentrations shown to inhibit evoked transmission, NA also inhibited somatic VDCCs, exhibiting characteristics of Gβγ-mediated modulation (Bean, 1989; Patil et al. 1996). Thus, application of NA (i) reduced whole-cell Ca²⁺ current amplitude, (ii) slowed current activation kinetics, and (iii) inhibition could be reversed partially by a large, depolarizing prepulse. Finally, introduction of Gβγ subunits into presynaptic SCGNs, at concentrations similar to those that inhibited synaptic transmission, also inhibited somatic VDCCs. In the SCG, NA may be released from dendrites and modulate cholinergic presynaptic neurones (Libet, 1979). The Gβγ-mediated presynaptic inhibition via α2-adrenoceptors demonstrated here in a cholinergic model synapse should therefore reflect the NA modulation of synaptic transmission in the SCG.

The presynaptic AP waveform

The shape of the presynaptic AP waveform is critical to the modulation of synaptic transmission, transmitter release being triggered by Ca²⁺ entry predominantly in the AP repolarization phase (Augustine, 1990; Wheeler et al. 1996; Sabatini & Regehr, 1997). SCGNs in culture display a mixed Na⁺–Ca²⁺ spike, followed by a long-lasting AHP (Mochida & Kobayashi, 1986a, b). Here, NA reduced both the duration and the AHP amplitude of the AP waveform. These effects are characteristic of agents that reduce Ca²⁺ entry into SCGNs in culture (Horn & McAfee, 1980; Mochida & Kobayashi, 1986a, b). In the present study, Cd²⁺ also reduced AP duration and AHP amplitude. The effects of NA and Cd²⁺ may therefore be attributed to inhibition of VDCCs, as also reported for somatic AP waveforms in other sensory neurones that express a mixed Na⁺–Ca²⁺ spike (Dunlap & Fischbach, 1978; Werz & MacDonald, 1983). In contrast, Cd²⁺ has been reported to block postsynaptic events with no concurrent effect on fast presynaptic AP waveforms (Borst et al. 1995; Sabatini & Regehr, 1997). Also, NA-mediated inhibition of glutamate release at calyx of Held synapses, which occurs via an effect on presynaptic VDCCs, proceeds without clear effects on the AP waveform (Leão & Von Gersdorff, 2002). AP duration in central presynaptic terminals is typically less than 1 ms and is not followed by a pronounced AHP (Augustine, 1990; Wheeler et al. 1996), whilst for SCGNs in culture, AP duration was 3–4 ms, followed by an AHP lasting > 150 ms under our recording conditions (Figs 1 and 4), such values being comparable for neurones in the SCG (Nishi et al. 1965). Therefore, the large Ca²⁺ component may contribute to differences in sensitivity of AP waveforms to VDCC blockers, and also to GPCR activation, between somatic SCGN recordings and those at central presynaptic terminals (Isaacson, 1998; Takahashi et al. 1998; Leão & Von Gersdorff, 2002). The present study suggests that synaptic transmission in sympathetic ganglia may be finely tuned by α2-adrenoceptors, since the duration of the AP modulates synaptic transmission; a broadening in AP leads to increased transmitter release and conversely a narrowing will decrease release and inhibit transmission (Wheeler et al. 1996; Geiger & Jonas, 2000).

Gβγ subunits also had effects on the somatic AP waveform in SCGNs in culture. Gβγ caused a decrease in AHP amplitude during and immediately following protein injection (see Fig. 4B), but as it did not persist during the Gβγ-mediated inhibition of synaptic transmission, this is likely to reflect a transient action of Gβγ at the soma. In contrast, we did not see any clear effects of Gβγ on AP duration. This may suggest that actions on AP duration and AHP amplitude represent distinct processes. However, as discussed above, our results suggest that agents that act by blocking Ca²⁺ channels would be expected to reduce both AP duration and AHP amplitude. One possible explanation is that Gβγ may concomitantly modulate additional conductances to further affect AP duration. For example, AP duration is also sensitive to modulation of K⁺ channels (Geiger & Jonas, 2000). Activation of K⁺ channels would shorten AP duration; conversely, inhibition or inactivation of K⁺ conductances would broaden the AP. Conceivably, the latter action may oppose the expected AP shortening due to the inhibition in Ca²⁺ influx by Gβγ. In this regard, signal transduction enzymes modulated by Gβγ include phospholipase A2, phospholipase C, mitogen-activated protein kinase, and several isoforms of adenylate cyclase (Clapham & Neer, 1997). Many of these Gβγ-activated enzymes have complex effects, including inhibition or inactivation of different types of K⁺ channels (e.g. Jonas & Kaczmarek, 1996), and thus have the potential to alter the shape of APs that invade nerve terminals.

Mechanism of Gβγ-mediated presynaptic inhibition

Having demonstrated that Gβγ subunits represent an important regulator of presynaptic inhibition by α2-adrenoceptors in SCGNs in culture, we examined potential mechanisms of action. Cd²⁺, a divalent cation that plugs the pore of VDCCs (Cloues et al. 2000), dramatically reduced transmitter release in SCGNs, and also effectively abolished somatic Ca²⁺ current. Similarly, NA reduced both transmitter release and somatic Ca²⁺ current. We examined the effects of NA in more detail and show a reduction in whole-cell Ca²⁺ current amplitude and a slowing of current activation, both effects being reversed by a large, depolarizing prepulse. These actions are characteristic of Gβγ-mediated modulation (Bean, 1989; Patil et al. 1996) and Gβγ subunits, introduced via the patch electrode, also caused a prominent inhibition of Ca²⁺ current. It is important to stress that somatic recordings may not necessarily correlate with presynaptic events (Geiger & Jonas, 2000), and that to implicate fully a Gβγ action on VDCCs at presynaptic terminals, direct recordings from these sites should be performed. However, VDCCs present at terminals contacting synaptically with adjacent neurones have been visualized recently. These VDCCs in distinct punctate clusters (Mochida et al. 2003), would also be expected to be subject to Gβγ-mediated inhibition, given that the injected Gβγ subunits reach these terminals, as the clear reduction in synaptic transmission seen here suggests. In support, studies in which NA released from short-term SCGN cultures was detected by amperometry have concluded that inhibition of transmitter release by NA showed pharmacological and kinetic similarities to the depression of somatic Ca²⁺ currents (Koh & Hille, 1997); it was proposed that a membrane-delimited action of Gβγ subunits may mediate this response. Thus, whilst presynaptic recordings are technically difficult due to the small size of the terminals, here we were able to use the SCGN system to introduce Gβγ subunits into presynaptic terminals via diffusion from somatic injection sites. It is clear from our findings that Gβγ subunits (in addition to NA and Cd²⁺) can inhibit somatic VDCCs and functionally inhibit transmitter release.

Ca_V2.2 (N-type) channels predominantly mediate fast transmission in SCGNs in long-term culture (Mochida et al. 1995). As discussed above, an α2-adrenoceptor pathway has been described extensively using somatic recordings from isolated SCGNs in short-term culture. For instance, N-type VDCCs are modulated by NA via multiple G protein subunits, using both voltage-dependent and -independent pathways in these cells (Hille, 1994). Furthermore, α2-adrenoceptors couple to Gα_o subunits to cause a voltage-dependent, PTX-sensitive inhibition of VDCCs, and to Gα_i subunits to cause a largely voltage-independent, PTX-resistant inhibition, mediated potentially by different Gβγ isoforms (Delmas et al. 1999). Similarly, release of NA from short-term SCGN cultures showed partial sensitivity to PTX (Koh & Hille, 1997). We demonstrate that the α2-adrenoceptor pathway is present in long-term-cultured cholinergic SCGNs and also is of functional significance in the inhibition of synaptic transmission. NA was able to cause a membrane-delimited, voltage-dependent inhibition of VDCCs and also to inhibit transmitter release by an action mediated by Gβγ subunits. Inhibition of transmitter release was reduced significantly by PTX and also in SCGNs expressing Gα transducin. Furthermore, NA-induced reduction of somatic Ca²⁺ current was also attenuated significantly by expression of Gα transducin, providing additional evidence for a link between actions on transmitter release and somatic Ca²⁺ channels.

These data confirm Gβγ subunits as an essential presynaptic inhibitory mediator for cholinergic synaptic transmission in sympathetic ganglia. As described above, the modulation of presynaptic transmitter release had close parallels with the modulation of somatic VDCCs; these pharmacological similarities represent correlative evidence that presynaptic inhibition may involve the same mechanism. This suggestion is supported by recordings at presynaptic terminals in the calyx of Held preparation, which demonstrated that injection of Gβγ subunits caused an inhibition of VDCCs (Kajikawa et al. 2001). However, alternative mechanisms have been proposed where Gβγ-mediated inhibition of evoked transmitter release occurs downstream of Ca²⁺ influx, possibly via a direct interaction with the synaptic release machinery (Blackmer et al. 2001). In relation to this general mechanism, numerous GPCRs, including α2-adrenoceptors (Starke, 2001), are reported to inhibit AP-independent ‘miniature’ postsynaptic events in mammalian synapses (Miller, 1998). Here, under conditions when the major drive for Ca²⁺ influx, i.e. the presynaptic AP, was removed, we saw no effect of NA on transmitter release stimulated by hypertonic sucrose application. Therefore, we have no evidence that the α2-adrenoceptor pathway identified in SCGNs had any effects on AP-independent transmitter release, although we cannot rule out fully a contribution to AP-evoked synaptic transmission.

Overall, we provide direct functional evidence that Gβγ subunits are important regulators of GPCR inhibition of transmitter release at mammalian presynaptic terminals. Results were consistent with Gβγ dimers representing important signalling molecules for presynaptic α2-adrenoceptor-mediated inhibition of transmitter release, supporting somatic studies in SCGNs. It will be of interest in the future to determine if other ion channel pathways modulated by multiple G protein subunits in SCGNs (Hille, 1994) can activate complementary pathways.