Aminergic and Peptidergic Modulation of Motor Function at an Identified Neuromuscular Junction in Helisoma

Neurotransmitters, such as biogenic amines and neuroactive peptides, play pivotal roles in the control of animal behaviour. Not surprisingly, the sites and mechanisms of action of these neurotransmitters are diverse (Kaczmarek & Levitan, 1987). For example, serotonin can modulate ion channels, thus affecting membrane excitability, action potential size and duration, and the release of neurotransmitter. Such effects constitute the physiological basis of some simple forms of learning (Siegelbaum et al. 1982; Klein et al. 1982; Boyle et al. 1984). Serotonin also initiates and modifies neuronal discharges like those underlying cyclical feeding behaviour (Granzow & Kater, 1977; Weiss et al. 1978; Bulloch & Dorsett, 1979; Lent & Dickinson, 1984; Tuersley & McCrohan, 1988) and locomotion (Willard, 1981; Harris-Warrick & Cohen, 1985; Claassen & Kammer, 1986). In addition, serotonin can modulate the responses of postsynaptic targets such as the contractility of muscle cells associated with feeding arousal (Weiss et al. 1978, 1979) and postural behaviour (Kravitz et al. 1980).

Previous studies have described the modulatory effects of serotonin (5-HT), small cardioactive peptide B (SCP_B) and Phe-Met-Arg-Phe-NH₂ (FMRFamide) on the central nervous system of the gastropod mollusc Helisoma. 5-HT and SCP_B initiate and control patterned motor activity (PMA) associated with fictive feeding behaviour, whereas FMRFamide inhibits PMA (Granzow & Kater, 1977; Murphy et al. 1985b). In this study, we examine the peripheral (neuromuscular) modulatory actions of these neurotransmitters.

Neurone B19, located in the buccal ganglia of Helisoma, is a motoneurone known to innervate one of the radular muscles of the buccal mass (Kater, 1974). The objectives of this study were (1) to characterize through anatomical and electrophysiological techniques B19’s neuromuscular junction (NMJ), (2) to examine the modulatory actions of 5-HT, SCP_B and FMRFamide on an isolated neuromuscular preparation, (3) to dissociate target muscle cells into primary cell culture, and (4) to assay the direct effects of the same neuromodulators on single muscle fibres under more controlled environmental conditions. Some of these results have been published previously in preliminary form (Haydon et al. 1988).

Adult albino pond snails, Helisoma, were used in all experiments. Animals were laboratory reared in freshwater aquaria at ambient temperature (approximately 23 °C) and fed lettuce and trout chow. The animals were deshelled and pinned to a silicone rubber dish. A dorsal midline incision exposed the visceral organs, the head ganglia and the buccal mass. The salivary glands were removed and the oesophagus was separated from the buccal mass. The buccal ganglia were separated from the other head ganglia by severing the cerebrobuccal connective, and were removed from the body together with the buccal mass. The muscle mass was cut along the midline, through the anterior musculature, odontophore and radular sac, dividing the mass into two halves. Muscles surrounding the supralateral radular tensor (SLT) muscle (see Kater, 1974) were trimmed away leaving the intact buccal ganglia flanked on each side by an SLT muscle still innervated by its buccal nerves. Small pieces of odontophore and radular membrane were left attached to the ends of the SLT for pinning and transducer attachment. The buccal ganglia-SLT muscle preparation was then pinned to the silicone rubber pad of a recording chamber containing Helisoma saline (in mmoll ⁻¹: NaCl, 51·3; KCl,1·7; CaCl₂, 4·1; MgCl₂, 1·5; Hepes buffer, 10 at pH 7·3).

Morphology of the motor pathway was examined by ionophoretic injection of a 5% solution of the fluorescent dye Lucifer Yellow CH (Molecular Probes, Inc.) into the soma of neurone B19. After a 2-h period of dye migration, the preparations were fixed in 4% phosphate-buffered formaldehyde, dehydrated, and mounted on glass slides in methyl salicylate (Cohan et al. 1987). The cleared preparations were then viewed with a fluorescent microscope.

The reduced buccal ganglia–SLT muscle preparation described above was used for isolated studies of neuromuscular function in Helisoma. To measure muscle tension, the SLT muscle was anchored to the base of the recording chamber with pins placed through the odontophore, and a small hook, connected by a fine filament to the force transducer, was attached to the radula. The experimental arrangement for recording and stimulating neurone B19 and for simultaneously recording membrane potential and muscle tension of the SLT is illustrated in Fig. 1. Muscle tension was detected by a microdisplacement myograph transducer F-50 (Narco Bio-systems, Inc.) connected to the differential amplifier (Tektronix 5A22N) of an oscilloscope. Conventional intracellular recording and display methods were used to record from neurone B19 and SLT muscle fibres. The simultaneous records were stored on magnetic tape for later analysis. Borosilicate microelectrodes were filled with either 2 mol l⁻¹ potassium acetate or 1·5 mmol l⁻¹ potassium chloride and had d.c. resistances of 10–30MΩ. Irrespective of the Electrolyte used, the same data were obtained. All experiments were conducted at ambient temperature.

In all pharmacological experiments (unless otherwise indicated), 10 × calcium saline (containing 41mmoll⁻¹ CaCl₂) was used to minimize the unwanted excitation of other buccal neurones. In zero-calcium/high-magnesium experiments the concentrations of CaCl₂ and MgCl₂ were changed to 0 and 10 mmol l⁻¹, respectively. Physiological saline and experimental solutions were changed by perfusing 20 ml through the recording chamber (volume 2 ml). Unless otherwise specified, all solutions were made up in normal Helisoma saline. Acetylcholine chloride, d-tubocurarine chloride, hexaméthonium bromide and serotonin (5-HT) were obtained from Sigma Chemical Co. (St Louis, MO), Phe-Met-Arg-Phe-NH₂ (FMRFamide) was obtained from Bachem, Inc. (CA) and small cardioactive peptide (SCP_B) was purchased from Peninsula Laboratories, Inc. (CA).

Animals were dissected in physiological saline and the paired SLT muscle masses were removed and placed in 2 ml of defined medium [DM; 50% Leibowitz-15 (Gibco) with Helisoma salts; Wong et al. 1981] for 30min. The SLT muscles from approximately 10–15 animals were incubated for 2h in a 30 ml flask containing 800 µl of DM, 200 µ1 of sterile haemolymph (extracted from snails as described in Hadley & Kater, 1983) and 4 mg of collagenase (Worthington Biochemical Corp.; CLS IV) at 33·5°C in a shaking water bath (with gentle agitation). All glassware used in the dissociation was precoated with 0·5% bovine serum albumin (BSA) to prevent muscle fibres from adhering to the glass. Following incubation, the flask was shaken by hand to further disaggregate fibres, and the collagenase solution containing the dissociated fibres was removed with a pipette and transferred to a centrifuge tube. Fresh collagenase solution was then added to the flask containing the remaining undissociated muscle and the tissue was incubated for an additional 2h to promote further dissociation. The dissociated muscle fibres were collected by centrifugation in a clinical centrifuge set at a moderate speed (setting no. 4 for the International Equipment Co., IM174) for 3 min. The supernatant was discarded and the pellet was resuspended in antibiotic saline. This saline wash was repeated three times. Finally, the dissociated fibres were transferred to 2 ml of DM in BSA-coated Petri dishes (Falcon no. 1008) and stored at ambient temperature until used (up to 2 days).

Single, dissociated SLT muscle fibres were plated onto Formvar-coated, gold electron microscope grids that had been precoated with poly-L-lysine. Grids were used because muscle contractions on the Formvar film were not damaging to the muscle membrane as were contractions on more rigid substrata such as plastic culture dishes. Physiological Helisoma saline was continuously perfused over the preparation by gravity inflow. Saline volume (1 ml) in the recording chamber was maintained by continuous surface suction. The experimental solutions were changed by perfusing 20 ml through the chamber. All experiments used conventional intracellular recording techniques as described above. Puffs of acetylcholine (ACh) (200 µmoll⁻¹; 4–8 ms in duration) were delivered to muscle fibres by pressure ejection of ACh from a micropipette attached to a Picospritzer II (General Valve Corp.). The micropipette tip diameter was approximately 2 µm and the tip was positioned 10–30 µm from the fibre. The muscles were visualized under the phase-contrast optics of an inverted microscope (Nikon: Diaphot-TMD) to which a video microscopy camera (Hamamatsu C2400) was connected for video tape analysis of the SLT muscle fibre contractile responses. Hard copies of single video frames before and after contraction were made directly from the video tape by a video copy processor K70S (Mitsubishi). The single frames were then digitized and the muscle fibre lengths were calculated.

The effects of pharmacological agents on membrane potential and muscle contraction (or tension) in isolated preparations and cell culture experiments were compared to an initial (unity) control value in physiological saline. Data were quantified for three defined experimental intervals: immediately before and during application of the pharmacological solution and immediately after washout of the test solution with physiological saline. All data were normalized to the unity control and the normalized values for each interval in all experiments were then averaged. The reported N values for pharmacological experiments represent either the number of isolated preparations tested or, in cell culture experiments, the number of single muscle fibres tested. Significance of the pharmacological effects was tested using a two-tailed t-test.

Using intracellular recordings from motoneurones and extracellular recordings of muscle activity, Kater (1974) was able to conclude that neurone B19, located on the dorsal surface of the buccal ganglion, controlled rhythmic muscle activity in the supralateral radular tensor (SLT) muscle of the buccal mass. Simultaneous intracellular recordings, as well as tension recordings, from neurone B19 and the SLT muscle in isolated preparations confirmed that B19 is one of the motoneurones that innervate the SLT (Fig. 1). Stimulation of B19 evoked 1:1 excitatory junctional potentials (EJPs) of 4·l±l·8mV (mean±s.D.; A =19) amplitude at constant latency in the SLTfibres (Fig. 2A). The summation of EJPs in the SLT muscle was accompanied by overt muscle contractions. In addition, during patterned motor activity (PMA) associated with fictive feeding behaviour, action potentials in B19 evoked 1:1 EJPs in SLT muscle fibres (Fig. 2B). EJPs were frequently recorded during PMA, however, that did not correspond to activity in neurone B19, indicating that the SLTmuscle is innervated by more than one excitatory motoneurone (Fig. 2B). Throughout all experiments on isolated B19-SLT preparations, inhibitory (hyperpolarizing) potentials were never recorded from SLT muscle fibres.

To test the monosynaptic nature of the B19–SLT connection, the extracellular concentration of calcium was raised 10-fold to decrease neuronal excitability. As shown in Fig. 3A, action potentials in B19 evoked 1:1 EJPs in the SLT muscle which were at a constant latency. Taken together, these studies support the idea that B19 forms a monosynaptic, excitatory connection with fibres of the SLT muscle.

Neurone B19 possesses a stereotyped morphology consisting of bipolar axons projecting into the bilateral buccal musculature through heterobuccal and ventro-buccal nerves (Berdan et al. 1987). To examine further the relationship between neurone B19 and its target muscle, we conducted morphological studies of axonal projections using intracellular ionophoresis of the motoneurone with Lucifer Yellow. These axons enter several muscle groups of the buccal mass, most obviously the SLTs (data not shown). Owing to the endogenous fluorescence of the SLT muscle, the innervation pattern of B19’s peripheral arbor on the muscle was not reliably resolved. Nonetheless, the direct axonal projection of B19 to the SLT muscle supports the electrophysiological evidence that B19 innervates this musculature.

Experiments were performed to determine if the B19-SLT connection was chemical in nature. Current injection to hyperpolarize the SLT membrane increased EJP amplitude whereas depolarization of the postsynaptic cell decreased EJP amplitude. In addition, EJPs evoked by stimulation of neurone B19 were reversibly decreased in amplitude by replacing the physiological bathing saline with a calcium-deficient saline solution which blocks the Ca²⁺-dependent release of neurotransmitter (N = 4; Fig. 3B). Thus, the excitatory connection between B19 and the SLT muscle is chemical in nature.

To make a physiological identification of the neurotransmitter involved at the B19–SLT neuromuscular junction, a variety of neurotransmitters and their antagonists were applied to an isolated SLT muscle bundle, bathed in normal Helisoma saline and connected to a force transducer. Bath application of 10⁻⁶–10⁻³ mol l⁻¹5-HT, glutamate, FMRFamide and SCP_B had no direct effects on SLT muscle tension (N = 6). However, bath applications of 10⁻⁶–10⁻⁴moll⁻¹ ACh always caused changes in SLT tension (N =6; Fig. 4A). To test the cholinergic nature of the neuromuscular junction, we applied two cholinergic antagonists to the bathing solution. Both 10⁻⁶mol⁻¹ tubocurarine chloride (N = 5; Fig. 4B) and 10 ⁵ mol 1 ¹ hexamethonium bromide (N = 3) reversibly decreased the amplitude of the B19-evoked EJPs in the SLT muscle fibres. These experiments support the hypothesis that acetylcholine is the excitatory neurotransmitter at the B19-SLT neuromuscular junction.

Muscle action potentials were not required to evoke contractions in the SLT muscle fibres. Only on rare occasions were action potentials recorded from the SLTs in reduced NMJ preparations. Injection of d.c. depolarizing current through intracellular microelectrodes was ineffective at eliciting action potentials in the SLT fibres. This may have been a result of the strong electrical coupling that exists between the fibres of the SLT muscle (M. J. Zoran & P. G. Haydon, unpublished observation). B19-evoked SLTmuscle contractions were graded in nature; that is, the strength of the muscle tension produced by a burst of action potentials in B19 was dependent on the number of action potentials evoked in the motoneurone and consequently on the magnitude of the summated EJP.

To examine whether the B19-evoked SLT contractions could be modulated by substances in addition to cholinergic compounds, an investigation was made of the effects of several pharmacological agents known to have modulatory actions in molluscan systems. Bath applications of 10⁻⁶moll⁻¹ 5-HT and SCP_B both increased the magnitude of SLT muscle contractions produced by a constant number of action potentials in B19 (Fig. 5). Fig. 6 illustrates that 5-HT potentiated the B19-evoked muscle tension 6·3 times (N = 7; P < 0·05) and SCP_B increased the motoneurone-elicited tension 2-7 times (N = 6; P<0·05). In addition, the neuroactive peptide FMRFamide caused a small but significant (N = 10; P < 0·02) increase in B19-evoked muscle tension (Fig. 6). The magnitude of the FMRFamide effect, however, was much less than that of 5-HT and SCP_B. To test whether the effects of these modulators were mediated by changes in EJPs, the amplitudes of B19-evoked EJPs were analysed in the presence of these neuroactive agents. Fig. 7 shows that 5-HT, SCP_B and FMRFamide, at the same Concentration (10⁻⁶moll⁻¹) that enhanced B19-evoked contraction of the SLT, had no significant effect on the amplitude of B19-evoked EJPs.

Several problems might exist when attempting to draw conclusions about the effects of a neuromodulator on in vivo and reduced neuromuscular preparations owing to the complex and heterogeneous chemical environment at the neuromuscular junction. Therefore, an experimental system was designed to control more precisely the immediate environment of the STL muscle during ACL stimulation.

SLT muscle fibres were dissociated into culture using a 0·4% collagenase solution in a conditioned medium containing 20% haemolymph (see Materials and methods). Single SLT fibres maintained a resting potential of −64·5 ±19·1 mV (mean±s.D.; N=16) that was similar to resting potentials recorded in reduced preparations (−71·2 ± 11·7mV; N=19). The collagenase-dissociated SLT fibres responded to pressure-applied pulses of ACh with depolarizing potentials (Fig. 8) that were graded in amplitude depending on the duration of the pulse and the concentration of ACh in the pipette. Large ACh-evoked depolarizations were almost always accompanied by longitudinal shortening of the muscle fibres (Fig. 9). Experimental alteration of SLT muscle fibre membrane potential resulted in changes in the amplitude of ACh-evoked depolarizing potentials (Fig. 8A). Although it was difficult to change the SLTfibre’s membrane potential to very depolarized levels, in three preparations the ACh-evoked depolarizing potential could be reversed in sign by current injection. ACh-evoked membrane depolarizations were decreased in amplitude by 10⁻⁶mol l¹ tubocurarine (N = 6;Fig. 8B). Sequential assay of muscle responses to ACh applied at intervals along the length of the denervated fibre failed to reveal any evidence of hypersensitive spots for ACh-evoked membrane potential or contractile effects. Thus, in cell culture, single SLT muscle fibres maintain physiological properties seen in semi-intact preparations and represent an appropriate model to test the direct actions of neuromodulators on muscle membrane potential and contractility.

The specific modulatory effects of 5-HT, SCPB and FMRFamide on single SLT muscle fibre responses to controlled applications of ACh were examined. As shown in Fig. 10, 5-HT and SCP_B at 10⁻⁶moll⁻¹ increased ACh-evoked muscle fibre shortening by 3·9 (N = 7; P< 0·001) and 3·0 (N = 8; P< 0·001) times, respectively. FMRFamide, however, reduced ACh-elicited SLT muscle fibre shortening to 0-5 times that in normal saline (N =8; P< 0·005). Neither 5-HT (N= 14; Fig. 11) nor SCP_B (N= 11) had an effect on ACh-evoked depolarizing potentials in single SLT muscle fibres (Fig. 12). FMRFamide, however, caused small but significant (N = 16; P < 0·05) reduction in the amplitude of depolarizing membrane potentials in response to applications of ACh (Figs 11,12). The effects of 5-HT and SCPB on ACh-evoked muscle responses, in cell culture, closely reflect their effects on B19-evoked muscle responses. The data indicate that both 5-HT and SCP_B modulate the strength of SLT muscle fibre contractions at a site distal to the generation of EJPs. The actions of FMRFamide, however, appear to be paradoxical (see Discussion). FMRFamide slightly increased the magnitude of B19-evoked SLT muscle tension, yet in cell culture ACh-evoked contractions of single muscle fibres were reduced by this peptide.

In the present study, we have extended previous observations (Kater, 1974) and suggest that neurone B19 of Helisoma is a cholinergic motoneurone which monosynaptically elicits excitatory junctional potentials (EJPs) and contractions in the SLT muscle during patterned motor activity (PMA) associated with fictive feeding behaviour (Fig. 2B). Using this defined motor pathway we have examined the peripheral actions of neuromodulators on the neuromuscular junction.

Biogenic amines, such as serotonin, dopamine and octopamine, modulate the central and peripheral pathways underlying diverse types of animal behaviour (for examples see Evans & O’Shea, 1977; Weiss et al. 1979; Kravitz et al. 1980; Lent & Dickinson, 1984; Flamm & Harris-Warrick, 1986; Whim & Evans, 1988). Recent studies show that neuroactive peptides, such as small cardioactive peptide B (SCP_B; Morris et al. 1982) and Phe-Met-Arg-Phe-NH₂ (FMRFamide; Price & Greenberg, 1977), also have modulatory effects on invertebrate behavior (Cottrell et al. 1983; Murphy et al. 1985b; Lehman & Greenberg, 1987; Willows et al. 1988; Prior & Watson, 1988). In several systems, parallel aminergic and peptidergic pathways have been shown to mediate similar events underlying behavioural acts (Kravitz et al. 1980; O’Shea & Bishop, 1982; Greenberg et al. 1983; Lloyd et al. 1984, 1987; Cropper et al. 1987).

Previous studies on Helisoma have shown that both 5-HT and SCP_B can increase the patterned motor output of B19 during fictive feeding (Granzow & Kater, 1977; Murphy et al. 1985b). Peripherally, 5-HT and SCP_B had effects that appear to be consistent with their central actions. Both 5-HT and SCP_B increased the magnitude of B19-evoked contractions in SLT muscle (Figs 5, 6) without significant changes in EJP amplitude (Fig. 7). Thus, in addition to initiating and controlling the cycle rate of feeding PMA in Helisoma, 5-HT and SCP_B can also increase the strength of peripheral muscle contractions when feeding occurs.

An unambiguous identification of modulatory actions in vivo and in reduced preparations is difficult because some observed effects could be indirect products of neurosecretion from adjacent cells. Therefore, to identify the direct actions of a single neuromodulator, we reduced the complexity of the system to an isolated muscle fibre in the experimentally controlled environment of cell culture. Single dissociated SLTmuscle fibres, in culture, maintain normal physiological properties such as resting membrane potentials, depolarizing membrane potential responses to ACh, and contractile responses to ACh (Figs 8 and 9). Thus, cultured SLT muscle fibres represent an appropriate assay system for examining the direct modulatory effects of neurotransmitters on the muscle element of an identified neuromuscular junction. Both 5-HT and SCP_B had strikingly similar effects on ACh-evoked responses in single muscle fibres (Fig. 10) to those of B19-evoked muscle responses in isolated preparations (Fig. 6). Both modulators increased evoked muscle contractions without affecting depolarizing membrane potentials.

The site of modulatory action of 5-HT and SCP_B on muscle contraction in Helisoma is distal to the site of changes in membrane potential. There are several mechanisms which might bring about these modulatory effects; for example, the activation of a second messenger system. The potentiation of accessory radula closer (ARC) muscle contractions in Aplysia by 5-HT and SCP_B is mediated by increased cyclic AMP levels, perhaps via cyclic-AMP-dependent protein kinase phosphorylation of contractile proteins (Weiss et al. 1979; Lloyd et al. 1984).

The neuropeptide FMRFamide is contained within somata and processes of the Helisoma nervous system (Bulloch et al. 1988) and, unlike 5-HT and SCP_B, has been shown to reduce B19 discharge during PMA in this species (Murphy et al. 1985b). Further, FMRFamide had an opposite action to that of SCP_B on the modulation of the synaptic connection between neurone B4 and the salivary gland cells of Helisoma (Coates & Bulloch, 1985). Thus, one might predict that FMRFamide would have an antagonistic action to that of SCP_B on the B19-SLT Neuromuscular preparation. Nevertheless, in such preparations, FMRFamide slightly potentiated B19-evoked SLT muscle contractions (Fig. 6). In addition, previous molluscan studies have shown that FMRFamide can potentiate muscle contractions (Greenberg & Price, 1979; Richmond et al. 1986). However, when reduced to the level of a single muscle fibre in cell culture, FMRFamide reduced ACh-evoked contractions of SLT muscle cells (Fig. 10).

The differential effects exerted by FMRFamide on isolated preparations and cell culture may have arisen for several reasons. In our nerve-muscle experiments, it was necessary to bathe the preparation in 10 × Ca²⁺ saline to reduce overall neuronal excitability. This may have affected the action of FMRFamide on the SLT muscle. Alternatively, the effects of FMRFamide detected in such preparations may result from an interactive modulation with one or more endogenous neurotransmitters. Cell culture studies of neurone B19 have shown that 5-HT is capable of modulating the motility of growth cones and the extension of neurites (Haydon et al. 1984, 1987). However, when 5-HT is applied in the presence of ACh, the modulatory effects of 5-HT are blocked (McCobb et al. 1988). Thus, the action of a given neurotransmitter can be changed when it is presented to the target in combination with other neurotransmitters.

In this study, we have identified the action of specific neuromodulators on the B19-SLT neuromuscular system. The modulatory neurones which mediate these effects in vivo, however, are not known. Serotonergic metacerebral cells (MCCs) have been shown to modulate contractions of accessory radula closer (ARC) muscles in the buccal mass of Aplysia (Weiss et al. 1978, 1979). Excitation of the MCCs potentiates ARC muscle responses to action potentials in buccal moto-neurones B15 and B16. Serotonergic neurones (Cl) in Helisoma are homologous to the MCCs of Aplysia (Granzow & Rowell, 1981) and are known to modulate the motor output of B19 (Granzow & Kater, 1977) and have axonal projections in the buccal nerves that innervate the SLT muscles (Murphy et al. 1985a). Perhaps neurone Cl has parallel actions to those of MCC in Aplysia and mediates the serotonergic modulation of the SLT muscle. In addition, the cholinergic motoneurone B15 in Aplysia is known to contain SCP_B, which when applied exogenously potentiates B15-evoked contractions of the ARC muscle (Lloyd et al. 1984). Similarly, motoneurone B19 in Helisoma possesses SCP_B-like immunoreactivity (Lukowiak & Murphy, 1987), raising the possibility that this motoneurone is the source of this peptide. Perhaps, in vivo, direct interactions exist between endogenous SCP_B and ACh coreleased at B19-SLT neuromuscular junctions. The corelease of transmitters at neuromuscular junctions has been implicated in several other invertebrate systems (Adams & O’Shea, 1983; O’Shea & Schaffer, 1985; Cropper et al. 1987; Lloyd et al. 1988). At Aplysia buccal muscle NMJs, some neurotransmitters may have postsynaptic actions, and other coexisting modulators may have presynaptic actions (Cropper et al. 1988). This is particularly interesting since the type of transmitter released to regulate a postsynaptic event can be determined by the pattern of electrical activity in the presynaptic cell (Ip & Zigmond, 1984). Since we have recently been able to reconstruct the B19-SLT neuromuscular junction of Helisoma in culture (Zoran & Haydon, 1988), it may be possible to determine whether the corelease of neuroactive peptides can modulate the cholinergic effects of B19 on its target muscle cells.

We thank Drs S. Schacher and K. Lukowiak for valuable advice during the course of these experiments. We also thank Dr S. Shen and H. Man-Son-Hing for helpful comments on the manuscript and R. Doyle for technical assistance. This work was supported by NIH grant NS24233.