ABSTRACT
The vertebrate GABAB receptor (GABABR) agonists 3-aminopropylphosphonous acid (3-APPA), SK&F97541 and 3-aminopropylphosphonic acid (3-APA) are able to induce a hyperpolarization on the cell body of motor neurone Df in the metathoracic ganglion of the cockroach (Periplaneta americana), although the classic vertebrate GABABR agonist L-baclofen fails to induce any responses on the same neurone. Consistent with the findings on vertebrate GABABRs, the 3-APPA-induced responses on Df were also insensitive to the GABAAR antagonists bicuculline and picrotoxin. However, the GABABR antagonists saclofen and CGP35348 also failed to block this GABAB-like response. These results indicate a novel pharmacology for the GABAB-like receptors on the Df neurone. The reversal potential indicates that these GABAB-like receptors may be coupled to K+ channels.
INTRODUCTION
Vertebrate GABAB receptors (GABABRs) were discovered by Bowery and colleagues in the early 1980s (Bowery et al. 1980; Hill and Bowery, 1981). Initially identified at peripheral nerve terminals of the rat atrium, GABABRs were found to mediate presynaptic inhibition (Bowery et al. 1980). Later, GABABRs were found in many regions of brain, ganglia and smooth muscles. The presynaptic inhibition of transmitter release and postsynaptic inhibitory actions mediated by GABABRs appear to involve a decrease in Ca2+ conductance (Dunlap et al. 1987) and an increase in K+ conductance (Gähwiler and Brown, 1985; Dutar and Nicoll, 1988; Gage, 1992). There is evidence that the link between GABABRs and effector channels is via G-proteins (Andrade et al. 1986; Thalmann, 1988) coupled to second messenger systems, notably adenylyl cyclase (Wojcik and Neff, 1984; Karbon et al. 1984; Hill, 1985). Vertebrate GABABRs are activated specifically by L-baclofen and 3-aminopropylphosphonous acid (3-APPA) and are blocked by phaclofen, saclofen and CGP35348 (Bowery, 1993).
Multiple subtypes of GABABR may exist in the vertebrate central nervous system (CNS). Electrophysiological studies on hippocampal neurones have indicated that presynaptic GABAB responses are insensitive to phaclofen and pertussis toxin (Colmers and Williams, 1988), whereas postsynaptic responses are blocked by application of either of these agents (Dutar and Nicoll, 1988). In rat neocortical slices, the hyperpolarizing actions of baclofen are not mimicked by 3-APPA (Ong et al. 1990). In contrast, 3-APPA is a potent inhibitor of GABA autoreceptors in spinal cord, whereas L-baclofen is ineffective (Bonanno and Raiteri, 1993), indicating the presence of pharmacologically distinct subtypes of GABABRs in vertebrate nervous systems.
Little is known about invertebrate GABABRs. This may be due in part to the finding by several laboratories that L-baclofen is ineffective in many invertebrate preparations (Sattelle et al. 1988; Benson, 1989). However, receptors of this general type have been observed at lobster neuromuscular synapses of both inhibitory and excitatory axon terminals where they appear to be coupled to K+ channels via a G-protein (Miwa et al. 1990). Using the potent agonist 3-APPA (Dingwall et al. 1987), GABAB-like responses have been observed in the neuropile of the cockroach terminal abdominal (A6) ganglion (Hue, 1991). 3-APPA has also been reported to inhibit the heart rate of Limulus polyphemus (Benson, 1989).
The present study, using an identified cockroach neurone, the fast coxal depressor motor neurone (Df), describes aspects of the pharmacology and the ionic basis of insect GABABRs. This neurone is readily identified and is amenable to electrophysiological studies, including conventional two-electrode voltage-clamp. It is a convenient preparation with which to characterize putative GABAB-like receptors. GABA-gated chloride channels have been characterized in detail on the same neurone (Sattelle et al. 1988; Pinnock et al. 1988).
MATERIALS AND METHODS
Adult male cockroaches (Periplaneta americana L.), reared at 27°C with freely available food and water, were used throughout the investigation.
3-Aminopropylphosphonous acid (3-APPA; alternatively termed 3-aminopropylphosphinic acid), 3-aminopropylphosphonic acid (3-APA) and saclofen were purchased from Tocris-Neuramin (Bristol, UK). CGP-35348 was generously donated by Ciba-Geigy (Basel, Switzerland). SK&F97541 was a gift from SmithKline Beecham Pharmaceuticals (Harlow, England). Picrotoxin, bicuculline methiodide and other compounds are products of Sigma Chemical Co. (St Louis, MO, USA).
GABA, 3-APPA, 3-APA and SK&F97541 were applied via a syringe pump (Bai et al. 1991) unless otherwise stated. The outlet of the syringe pump was inserted into the perfusion line via a Y-shaped tube. The final drug concentration to which the preparation was exposed was calculated from the flow rate of the perfusion pump and the flow rate of the syringe pump. This method offered a more rapid application and wash-out of the agonists than that provided by conventional bath application. Antagonists were delivered to the preparation by bath application. Micropipette pressure application (Pneumatic Pressure System, PPS-2) was used for the direct local application of 3-APPA to the cell body of Df. In these studies, the tip of the micropipette was positioned about half a cell diameter above the motor neurone cell body. Brief pressure pulses (140 kPa, 1 s) were employed to apply 3-APPA (pipette concentration 10−2 mol l−1 in saline).
A portion of the nerve cord containing the meso-and metathoracic ganglia and the first two abdominal ganglia was removed and the ventral surface of the metathoracic ganglion was desheathed under a dissecting microscope using sharpened forceps. The preparation was then mounted under saline in a Perspex experimental chamber (volume 0.1 ml). Isolation of the soma of motor neurone Df was achieved by cutting around and under the cell with a tungsten hook electrode (Hancox and Pitman, 1991). Detailed membrane potential recording methods have been reported elsewhere (Bai et al. 1991). The cell body of the fast coxal depressor motor neurone (Df) was located visually and impaled with a microelectrode filled with 2.0 mol l−1 filtered KCl (resistance 10–15 MD). The membrane potential was amplified by a d.c. amplifier (Axo-clamp 2A) and recorded with a chart recorder (Gould BS-272). The preparation was perfused at a constant rate (2.7 ml min−1) with either normal saline or saline containing a test drug. The perfusion saline used throughout these experiments had the following composition (in mmol l−1): NaCl, 214; CaCl2, 9.0; KCl, 3.1; Tes, 10 (pH 7.2, adjusted with 2.0 mol l−1 NaOH) and sucrose, 50.
A two-electrode voltage-clamp (Axo-clamp 2A) was used to estimate the reversal potential for the GABAB-like responses. The current electrode (resistance 8–15 MD) and the voltage electrode (resistance 10–30 MD) were filled with 2.0 mol l−1 KCl. The holding potential was varied over the range -70 mV to -120 mV. 3-APPA was introduced via a syringe pump and the current was recorded on a chart recorder (Gould BS-272).
RESULTS
A picrotoxin-insensitive component of the response to GABA recorded from cockroach motor neurone Df
The use of low-resistance, KCl-filled electrodes will normally load the cell with chloride ions. Typically, the increase in intracellular chloride concentration changes the equilibrium potential for chloride from -80 mV to about -50 to -60 mV. Under these conditions, the activation of GABA-gated chloride channels induces large-amplitude, highly reproducible depolarizing responses better suited for analysis of pharmacological data than the smaller hyperpolarizing responses to GABA recorded from cells not loaded with chloride ions (see Pinnock et al. 1988). The responses to GABA usually became stable about 30 min after the initial impalement. GABA was applied to the preparation via a syringe pump. In many cases, a small-amplitude, slower, GABA-induced, picrotoxin-insensitive hyperpolarization was observed following blockade of the chloride-mediated response by a high dose of picrotoxin (10−5 mol l−1, Fig. 1A; N=3). The potent GABABR agonist 3-APPA (10−3 mol l−1) induced a similar hyperpolarizing response, which was also insensitive to picrotoxin (10−5 mol l−1, N=3) as shown in Fig. 1B.
Responses to 3-APPA were also recorded during pressure application of the agonist via a micropipette (pipette concentration 10−2 mol l−1) locally to the motor neurone Df (Fig. 2A; N=3). In addition, the isolated motor neurone Df (see Hancox and Pitman, 1991) responded to syringe-pump-applied 3-APPA (10−3 mol l−1, N=4), as shown in Fig. 2B, indicating that receptors mediating these GABAB-like responses are on the cell body membrane of motor neurone Df. The differences in onset and recovery phase of 3-APPA-induced responses detected by these two methods may reflect the differences in drug concentration changes at the cell surface.
Actions of GABABR agonists
Syringe-pump application of the newly developed, potent GABABR agonist SK&F97541 resulted in a hyperpolarizing response of very slow onset (time-to-peak of the rising phase 2.7±0.2 min; N=6) and of long duration (Fig. 3A). Dose–response curves for SK&F97541, 3-APPA and 3-APA were constructed (Fig. 3B). Of the agonists tested, SK&F97541 was the most potent with an estimated EC50 of 1.4X10−6 mol l−1. 3-APPA induced a similar response with a more rapid onset (time-to-peak of the rising phase 1.7±0.1 min; N=9) and of shorter duration compared with the SK&F97541-induced responses (Fig. 3A); both responses saturated at about the same level. 3-APPA-induced responses were less potent (estimated EC50 2.2X10−5 mol l−1), and 3-APA, an analogue of 3-APPA, was ineffective at concentrations below 10−4 mol l−1. Only at very high doses of 3-APA were GABAB-like hyperpolarizations observed (Fig. 3B).
Suppression by GABABR agonists of 3-APPA-induced responses
Responses to 3-APPA were reversibly reduced by bath-application of SK&F97541 (10−5 mol l−1) (Fig. 4). High doses of 3-APA (10−3 mol l−1) also partially supressed the 3-APPA-induced responses (Fig. 4B). This reduction was not entirely due to the hyperpolarizing actions of these drugs since, even when the membrane potential was depolarized to resting level by injecting constant current (as in the example shown in Fig. 4A), the reduction was still observed. The inhibitory actions of SK&F97541 and 3-APA on 3-APPA-induced responses provide evidence that these GABABR agonists either act on the same receptor or share common interacting sites.
Actions of vertebrate GABAAR and GABABR antagonists
3-APPA was selected as the agonist for further studies in view of the relatively rapid onset and recovery of its effects. The vertebrate GABABR antagonists saclofen (10−4 mol l−1) and CGP35348 (10−4 mol l−1) failed to block 3-APPA-induced responses (Fig. 4B) recorded from motor neurone Df. The vertebrate GABAAR blockers bicuculline (10−4 mol l−1; N=3) and picrotoxin (10−5 mol l−1; N=3) also failed to block 3-APPA-induced hyperpolarizations (Fig. 4B).
Ionic basis of 3-APPA-induced currents
The membrane potential of the cell body of motor neurone Df was voltage-clamped at various membrane potentials between -70 mV and -120 mV, and current traces were recorded in response to syringe-pump-applied 3-APPA (Fig. 5A,B; N=4). When membrane potentials were in the range -70 mV to -90 mV, 3-APPA induced an outward current, whereas between -100 mV and -120 mV, inward currents were recorded. A reversal potential of -95 mV was estimated, a value close to the value for the potassium equilibrium potential (EK) of -97 mV previously determined for this neurone (David and Sattelle, 1990).
DISCUSSION
Although, under physiological conditions, both GABA-gated chloride channels (Sattelle et al. 1988) and 3-APPA-induced responses (present study) of motor neurone Df are in the same hyperpolarizing direction, there are clear differences in their pharmacology and signal transduction mechanisms. First, responses through GABA-gated chloride channels have a rapid onset and a short duration (Sattelle et al. 1988), whereas 3-APPA-induced hyperpolarizations are of slower onset and have a longer duration. Second, picrotoxin, an effective blocker of GABA-gated chloride channels of vertebrates (Simmonds, 1983) and insects (Sattelle, 1990), fails to block 3-APPA-induced responses. Third, GABA-gated chloride channels reverse at the chloride equilibrium potential, approximately -50 mV when KCl electrodes are used (Sattelle et al. 1988), whereas 3-APPA-induced responses reverse at the potassium equilibrium potential. These results are consistent with previous findings on an identified cockroach interneurone (Hue, 1991), in which hyperpolarization responses are observed when applying GABA (in the presence of picrotoxin) and 3-APPA. Under the actions of picrotoxin the GABA-induced hyperpolarizations on the interneurone have a reversal potential close to the potassium equilibrium potential (Hue, 1991).
The dose–response curves for GABABR agonists presented here indicate that their order of potency on this cockroach motor neurone is as follows: SK&F97541>3-APPA. 3-APA and baclofen (Sattelle et al. 1988) are less active on this neurone. This agonist profile is different from that obtained from functional studies on substantia nigra neurones in rat brain, where the potency order of GABABR agonists is: SK&F97541>baclofen=3-APPA (Seabrook et al. 1990) (Table 1) and from binding studies on rat brain membranes, where SK&F97541>3-APPA>baclofen (Bowery, 1993). The differences in agonist profile may indicate the existence of different subtypes of GABABR.
In the present study, SK&F97541-and 3-APPA-induced responses appear to be saturable at similar levels. The blocking actions of 3-APA and SK&F97541 on 3-APPA-induced responses recorded from the cell body of motor neurone Df are possibly the result of a combination of these GABABR agonists (a) competing for a single site and (b) acting on different sites that either interact or share common intracellular pathways.
Advantages of pipette-pressure application of drugs include local delivery and rapid application. However, a limitation is that the concentration of the test molecule at the cell surface is unknown. In the present study, the pulse of agonist is delivered into a chamber filled with saline that is constantly being perfused, and the resulting concentration at the cell surface will be much lower than the concentration in the pipette.
Pharmacological differences have been detected between the GABAB-like receptors on the cockroach motor neurone Df and all vertebrate GABABRs reported to date. The widely used vertebrate GABABR agonist L-baclofen is inactive on this insect GABABR (Sattelle et al. 1988) and on dissociated neurones from a locust, Locusta migratoria (Benson, 1989), as well as in situ giant interneurones of the cockroach Periplaneta americana (Hue, 1991). The pharmacology of GABAB-like receptors on motor neurone Df also showed novel properties, for example, insensitivity to the vertebrate GABABR antagonist CGP35348 (10−4 moll−1), a potent blocker of GABABRs in rat brain neurones at this same concentration (Seabrook et al. 1990; Olpe et al. 1990). This GABABR antagonist has not previously been used to study insect GABAB-like responses, and the results presented here point to the existence of a novel type of GABABR on cockroach motor neurone Df. When the molecular cloning of GABABRs has been achieved, it may be possible to account in molecular terms for differences observed in receptor pharmacology.
This identifiable neurone offers a convenient preparation for further studies of receptor signal transduction, pharmacology and function. It may also offer a single cell PCR (polymerase chain reaction) approach to the molecular cloning of a GABABR, which has not been achieved to date.