Potentiation of Neuromuscular Transmission by An Octopaminergic Neurone in the Locust

Biogenic amines are capable of modulating transmission at vertebrate neuromuscular junctions (Orbeli, 1923; Bowman & Zamis, 1958; Kuba, 1970). However, the natural release sites of the modulating amines and the physiological significance of the experimental effects have yet to be established (Bowman & Nott, 1969). In contrast, in invertebrates skeletal muscle has recently been shown to receive a direct aminergic input in addition to its motor innervation, for instance in the marine mollusc, Aplysia (Weiss, Cohen & Kupfermann, 1975, 1978) and in locusts (Evans & O’Shea, 1977, 1978; O’Shea & Evans, 1977).

In the locust the aminergic input to the extensor muscle of the tibia of the hindleg is provided by one of the neurones of the dorsal unpaired median (DUM) group.^* The particular neurone that innervates this muscle is unpaired in the metathoracic ganglion and has been called DUMETi (Dorsal Unpaired Median cell to Extensor Tibiae muscle) (Hoyle et al. 1974). Both left and right extensor tibiae muscles receive an axon from the single unpaired DUMETi neurone. Each extensor muscle is also innervated by three paired motoneurones (Pearson & Bergman, 1969; Cochrane, Elder & Usherwood, 1972; Hoyle & Burrows, 1973), whose cell bodies lie in the meta-thoracic ganglion.

In previous publications (Evans & O’Shea, 1977, 1978) we have shown that DUMETi is an octopaminergic neurone. One physiological effect of stimulating DUMETi is to lower the frequency of a myogenic rhythm of contraction found in some of the fibres of the extensor muscle. This effect can be mimicked by superfusing the muscle with low concentrations of octopamine (Evans & O’Shea, 1978). This slowing is mediated by receptors on the myogenic muscle fibres with a high affinity for octopamine. These fibres comprise only a very small fraction of the total, and the above finding is not therefore evidence for the widespread presence of octopamine receptors on the extensor muscle.

The aim of the present study is to determine the action of the octopamine released from the terminals of DUMETi in the non-myogenic parts of the extensor muscle. Here we show the absence of a dramatic direct effect of DUMETi stimulation on the muscle. This finding led us to suspect that the function of DUMETi might only be revealed by its interactions with the effects produced by the motoneurones. Here we show that DUMETi is able to potentiate the strength of contractions produced by the slow excitatory motoneurone (SETi) and that this effect can be mimicked by the application of low concentrations of octopamine. The presynaptic terminals of the SETi motoneurone possess octopamine receptors capable of mediating at least a part of the observed potentiation of neuromuscular transmission. Other octopamine receptors, present on the muscle fibres themselves, increase the rate at which the muscle relaxes after being excited by either the slow or the fast excitatory motoneurone. Octopamine can also reduce the effectiveness of the common inhibitory motoneurone to the extensor muscle (Evans & O’Shea, 1977). Here we show that this effect is not responsible for the potentiating effect of DUMETi and octopamine on SETi neuromuscular transmission.

The results suggest that the DUMETi neurone functions as a modulator of the effectiveness of a synaptic junction, between a slow motoneurone and a muscle. This conclusion, and the implications of the involvement of a biogenic amine in neural modulation are discussed. Brief accounts of some of this work have already been published (Evans & O’Shea, 1977; O’Shea & Evans, 1977).

Experiments were performed at room temperature (21°C) on adult Schistocerca americana gregaria (Dirsch, 1974) (formerly S. gregaria) of either sex. The animals were obtained from our own crowded laboratory cultures fed on wheat seedlings.

Dorsal dissections to expose the ventral nerve cord were performed on animals which were restrained on a dish. Full details of the physiological techniques used are given in Evans & O’Shea (1978). Intracellular recordings were made from the somata of DUM neurones in the metathoracic ganglion using glass microelectrodes filled with 2 M-potassium acetate and having resistances of 20–50 MΩ. DUMETi was identified by showing that its soma spikes always corresponded with its axon spikes recorded extracellularly in both nerves innervating the left and right metathoracic extensor tibiae muscles (nerve 5b₁; nomenclature of Pringle, 1939). Spikes in DUMETi were elicited either by passing depolarizing current into the soma through the recording microelectrode via a bridge circuit, or by antidromic stimulation from one pair of silver hook electrodes placed on the extensor tibiae nerve (5b₁) of one side. The latter technique allowed us to elicit a higher frequency of spikes in the DUMETi axons.

Motoneurones of the extensor tibiae muscle were excited by stimulating the peripheral nerve roots which contain their axons. The motor axon of SETi was stimulated by a pair of silver hook electrodes placed around nerve 3b. This nerve also contains a branch of the CI motoneurone but this was not excited at threshold for extracellular stimulation of SETi. This was confirmed by intracellular recording from muscle fibres innervated by both motoneurones. The axon of the fast extensor motoneurone of the tibiae (FETi) was stimulated by paired hook electrodes placed around nerve 5. This nerve contains the axons of many neurones, including that of DUMETi, but that of FETi is the largest in diameter and has the lowest threshold to extracellular stimulation.

The effect of stimulating DUMETi on the muscle tension developed by the moto-neurones was investigated. Known concentrations of drugs were superfused directly on to the extensor tibiae muscle. The muscle was exposed by opening the metathoracic femur along the dorsal ridge. Solutions were superfused at a flow rate of 1 ml/min using a peristaltic pump. Tension was measured almost isometrically with a force transducer attached to the distal tendon of the muscle. Intracellular recordings were made from the muscle fibres innervated by the SETi motoneurone. All results were recorded on magnetic tape and filmed after the experiments.

The amines and blocking agents were dissolved in physiological isotonic saline (pH 6·8) containing 140 mm-NaCl, 10 mm-KCl, 4 mm-CaCl₂, 4 mm-NaHCO₃, 6 mm-NaH₂PO₄ (Usherwood & Grundfest, 1965) plus 90 mm-sucrose.

We would like to acknowledge the gift of N,N, dimethyl-octopamine (Sterling Winthrop Res. Inst.). Other drugs were obtained from the Sigma Chemical Co. except for phentolamine mesylate (Ciba) and phenylethanolamine HC1 (Regis Chemical Co.).

Stimulation of DUMETi produces a small hyperpolarization ( ∼ 3–5 mV) in the resting membrane potential of certain muscle fibres of the extensor tibiae muscle. This effect is quite variable and is often not present. While monitoring the direct effects of DUMETi stimulation (and also of direct octopamine application) on the muscle, we have often attempted to measure associated changes in conductance of intracellularly recorded muscle fibres. We have consistently been unable to measure such changes and assume therefore that the muscle’s receptors for octopamine do not mediate large conductance changes.

DUMETi stimulation produces a reduction in the overall muscle tonus. This relaxation can be marked although it is again highly variable. It can be mimicked by superfusion of octopamine on to the muscle. Low concentrations of DL-octopamine (10⁻⁸M) produce small and variable effects whereas high concentrations (10⁻⁶M) produce consistently large effects. These actions of octopamine on the extensor muscle persist in the presence of picrotoxin at a concentration (5×10⁻⁴ M) which blocks the effect of the inhibitory motoneurone on the muscle (Usherwood & Grundfest, 1965). The direct superfusion of GABA (5×10⁻⁵ M) on to the extensor muscle produces no reduction in basal tonus. The above findings suggest that the relaxation in basal tonus is not caused by a release of GABA from the terminals of the common inhibitory (CI) motoneurone. The relaxing effect on muscle tonus is therefore a direct post-synaptic effect of octopamine on the extensor muscle.

When the axon of SETi is stimulated at 6 Hz, each spike evokes a small twitch in the muscle fibres it innervates. At 6 Hz these twitches are discrete, constant in amplitude and do not summate. Stimulating DUMETi at 12 Hz while stimulating SETi at 6 Hz produces a fall in basal tension and potentiates the amount of tension produced by the slow motoneurone. In Fig. 1A the basal tension of the muscle is reduced during the 12 s burst of DUMETi spikes and returns to its original value after about 30 s. SETi induced twitches are increased in amplitude by about 30% at maximum and potentiation is detectable for at least 2 min. This suggests that the spikes in DUMETi induced a long-term change somewhere in the process from the mobilization of the motoneurone transmitter substance to the activation of the contractile mechanism.

The excitatory junctional potentials (EJPs) elicited in the muscle fibres by the SETi motoneurone are also potentiated by the stimulation of the DUMETi neurone at 20 Hz (Fig. 1 B). The increase outlasts the stimulation of DUMETi by about 2 min and has a time-course similar to the potentiating effect on the amplitude of the tension. The effect on EJP amplitude is generally small and variable, whereas the potentiating effect on tension is larger and highly reproducible.

The effects of stimulating DUMETi can be mimicked by the application of low concentrations of octopamine to the extensor muscle (Fig. 2). A 30 s pulse of 10⁻⁶ M-DL-octopamine potentiates the amplitude of SETi-induced EJPs, increases the amplitude of the individual twitches by approximately 30% and causes a small reduction in basal tonus (Fig. 2 A). The potentiating effects on contractions and EJPs are prolonged, and outlast the octopamine pulse by several minutes. The effect on EJP size of any muscle fibre was much more variable and generally smaller than the effect of octopamine on SETi-induced tension recorded from the whole muscle. Fig. 2B shows another example of the potentiating effect of octopamine on EJP size, in which a 30 s pulse of DL-octopamine (10⁻⁶M) produces about a 10% potentiation. This concentration of octopamine typically produces a 30% increase in the amplitude of the contraction induced by SETi.

One possible explanation for the potentiating effect of octopamine is that the effect is secondary to a suppression of spontaneous release of transmitter (GABA) from the inhibitory motoneurone. This explanation seemed plausible since 10⁻⁶ M-DL-octopamine can reduce the amplitude of IJPs induced by the common inhibitory motoneurone (Evans & O’Shea, 1977). The potentiating effect of 10⁻⁶ M-DL-octopamine, however, is not blocked in the presence of the GABA blocking agent, picrotoxin, at a concentration (5 ×10⁻⁴ M) which is known to block the effect of exogenously applied GABA (10⁻⁸ M). The potentiating effect of octopamine on SETi-induced contractions appears, therefore, to be a direct effect on SETi neuromuscular transmission and not a secondary consequence of reducing tonic, spontaneous release of GABA.

To examine the characteristics of the potentiating effect of octopamine on the tension produced by the extensor muscle we varied both the length of the octopamine pulse and its concentration. A 30 s pulse induces a larger potentiation than a 10 s pulse (cf. Fig. 3 A, B). The half-time of the maximal response to continuous application of octopamine was 1·2 min (Fig. 3 C) where a 6 25 min pulse of 10⁻⁶ M-DL-octopamine was applied to the muscle. The potentiating effect of superfused octopamine on the extensor muscle is thus proportional to the. length of the octopamine pulse. Increasing the concentration of the octopamine in the superfusate from 10⁻⁶ M to 10⁻⁵ M increased magnitude of the potentiating response when the length of the pulse is kept constant at 30 s (cf. Fig. 4A with Fig. 3B). At 10⁻⁵ M-octopamine, prolonging the exposure time to 220 s produced only a slight increase in the extent of the potentiation (Fig. 4B). A comparison of the effect of short pulses of octopamine (Figs. 3 A, 3B, 4 A) with those of longer pulses (Figs. 3 C, 4B) indicates that the maximal tension response takes longer to develop in the extended exposures suggesting that under these com. dirions a desensitization of the octopamine receptors may be occurring.

The threshold for the potentiating effect of octopamine on SETi-induced twitch tension occurs between 10⁻⁷ and 10⁻⁸ M-DL-octopamine (Fig. 5A-C). The exact threshold concentration varies with the individual preparation. The effect of octopamine on basal tension is also concentration dependent. It is particularly marked at 10⁻⁵ M-DL-octopamine (Fig. 5C) but can still be seen in some preparations at concentrations down to 10⁻⁸ M (Fig. 5 A).

In an attempt to determine the structural specificity of the potentiation of SETi-induced tension, amines related to octopamine were superfused over the extensor muscle at a concentration of 1 μM. Fig. 6 shows the results of an experiment in which the effects of DL-octopamine, dopamine and DL-synephrine were compared. In the same preparation 1 μM DL-synephrine (Fig. 6C) produced a more marked potentiation of SETi-induced tension than that induced by an equivalent 30 s pulse of DL-octopamine (Fig. 6A). Dopamine at this concentration had no effect (Fig. 6B). In a similar series of experiments the following amines were also shown to be without effects on SETi-induced twitch tension when applied for 30 s at 10⁻⁶ M; tyramine, N,N,dimethyloctopamine, L-noradrenaline, L-adrenaline, phenylethanolamine, and 5-hydroxytryptamine (5-HT). Thus the receptor mediating the potentiating effect on SETi-induced twitch tension is very specific for the β-hydroxylated monophenolic amines at concentrations below 10⁻⁸ M.

Octopamine increases the rate at which the muscle relaxes following a twitch contraction induced by both SETi and FETi thereby decreasing the duration of individual contractions (Fig. 7 A, C). The structural specificity of the relaxation response appears to be similar to that of the potentiating response on SETi-induced contractions. DL-Synephrine at 10⁻⁶M (Fig. 7B) was again a potent agonist whilst the following amines were ineffective at this concentration: dopamine, L-noradrenaline, L-adrenaline, phenylethanolamine, N,N,dimethyloctopamine, tyramine and 5-HT.

The effect on the rate of relaxation differs in three respects from the potentiating effect of octopamine on the amplitude of the twitch. First, it has a much lower threshold being around 10⁻⁸ M for DL-octopamine. The same is true for DL-synephrine, the most potent agonist for this response (Fig. 8 A-C). The threshold for the potentiating effect of DL-synephrine on twitch amplitude is between 10⁻⁸ M and 10⁻⁷ M (Fig. 8B) whereas the threshold for the effect on relaxation is between 10⁻¹⁰ M and 10⁻⁹ M (Fig. 8 A). Second, the effects differ in relative potency. In the normalized data shown in Fig. 9A, B, 10⁻⁶ M-octopamine increases the rate of relaxation by 1·5 × and the twitch tension by 1·25 ×. Third, the time-course for the development of the two responses differs. The relaxation effect is initiated more rapidly (Fig. 9B), being detectable about 15 s after the application of 10⁻⁶ M-DL-octopamine. The effect on twitch tension is detected later, typically after about 25 s. The effect on relaxation develops more rapidly and reaches its peak sooner than the effect on twitch tension (Fig. 9B).

These differences between the threshold, potency and time course of the effect of octopamine on rate of relaxation and amplitude of twitch suggest different underlying mechanisms or different sites of action of octopamine.

We investigated the characteristics of the receptors mediating the effects of octopamine on tension induced by stimulating SETi. Pulses of octopamine were superfused over the muscle in the presence of α- and β-adrenergic blocking agents (Fig. 10). The potentiating effect of 10⁻⁰ M-DL-octopamine (Fig. 10A, D) was completely blocked in the presence of 10⁻⁸M phentolamine (a-adrenergic blocking agent) (Fig. 10C) but was not affected in the presence of DL-propranolol (β-adrenergic blocking agent) (Fig. 10B). Similarly the effect of octopamine and synephrine on the rate of relaxation was blocked by phentolamine, but was unaffected by propranolol. In this case, however, 10⁻⁵ M-phentolamine produced an incomplete block (Fig. 11). The inability of 10⁻⁶ M-phentolamine completely to block the effect on relaxation, and its ability to eliminate the effect on twitch tension is probably related to the greatly differing thresholds for the two effects; the rate of relaxation is an order of magnitude more sensitive to octopamine than is the amplitude of the SETi-induced twitches.

Recordings from resting, unstimulated muscle fibres which receive innervation from the slow extensor and CI motoneurones revealed a low frequency of small, transient, depolarizing potentials (Fig. 12, upper). We assume these to be quantal events representing spontaneous release of transmitter from the terminals of either the SETi or CI motoneurone. Their frequency is not affected by application of picrotoxin (5×10⁻⁴ M) and they therefore represent mepps released from SETi. We found no spontaneous hyperpolarizing, potentials in muscle fibres known to be innervated by the CI motoneurone.

In the presence of 10⁻⁵ M-DL-octopamine, the frequency of mepps increases (Fig. 12 lower), but the amplitude distribution and duration of the individual events is unchanged. This effect of octopamine is blocked by the α-adrenergic blocking agent phentolamine (10⁻⁸M), but not by the β-adrenergic blocking agent DL-propranolol (10⁻⁸ M) (Fig. 13). The effect on frequency of 10⁻⁸ M-DL-octopamine is also unaffected by the presence of picrotoxin at a concentration of 5×10⁻⁴ M. Thus octopamine appears to be acting to facilitate the amount of transmitter released from the pre-synaptic terminals of the SETi motoneurone by an interaction with a presynaptic receptor which has some of the characteristics of the vertebrate a-adrenergic receptor.

We have shown that an octopaminergic neurone which innervates a locust skeletal muscle modulates the force of contraction generated by an excitatory motoneurone innervating the same muscle. The finding of a direct aminergic modulatory input to a muscle has a precedent in the marine gastropod Aplysia in which the serotonergic metacerebral cells potentiate the force of contraction caused by motoneurones which innervate the muscles of the buccal mass (Weiss, et al. 1975, 1978).

The pharmacological properties, of octopamine receptors on myogenic fibres of the extensor tibiae muscle of the locust hindleg have been described (Evans & O’Shea, 1978). The myogenic fibres, however, are few in number and confined to a discrete bundle at the proximal end of the muscle. Here we have shown the presence of octopamine receptors on the remaining non-myogenic parts of the extensor muscle. This finding together with the presence of the DUMETi axon in the non-myogenic parts of the muscle (Evans & O’Shea, 1978) suggests that octopamine released from the DUMETi terminals has a more widespread functional importance in the extensor muscle than inhibition of the myogenic rhythm.

Stimulating DUMETi lead to a marked, long-lasting, potentiation of the twitch tension produced by the slow motoneurone (SETi). It also produced a slight decline in the basal tonus of the muscle. Intracellular recordings from individual muscle fibres innervated by SETi revealed that the EJPs were potentiated by DUMETi stimulation. Both potentiating effects far outlasted the period of DUMETi stimulation. The effect on EJP amplitude however was smaller than that on twitch tension and much more variable. This might suggest that the increased EJP amplitude is not responsible for the potentiated twitch tension. The effects on EJP size, however, were recorded from single muscle fibres and the effect on twitch tension was recorded from the whole muscle. Variation in the amount of EJP potentiation among single muscle fibres may be caused by variation in recording sites and proximity to neuromuscular junctions. The evidence of variation alone therefore cannot eliminate the possibility that potentiation of twitch tension is mediated by the EJP potentiation. There are post-synaptic effects of octopamine (see below) and we cannot yet say whether twitch potentiation is at least in part caused by direct postsynaptic action of octopamine on the muscle fibres.

Direct application of low concentrations of octopamine mimics the effects of DUMETi on SETi neuromuscular transmission. Concentrations of DL-octopamine as low as 10⁻⁷ M produced a significant potentiation of SETi-induced twitch tension. The octopamine induced potentiation of both SETi twitch tension and amplitude of SETi EJPs were both of long duration, far outlasting the stimulating pulse. We take the similarities between these two effects to indicate that the potentiation induced by stimulating DUMETi is mediated by the release of octopamine from its terminals on the muscle.

The mechanism of the potentiation in the system we describe is as yet unknown. Although we have shown the presence of both pre- and post-synaptic receptors for octopamine, we do not know which causes the enhancement of contraction. Octopamine causes an increase in the size of the EJP and an enhanced contraction but we have not shown that the enhanced contraction is caused by the increased EJP amplitude. Indeed, in Aplysia, Weiss et al. (1978) have shown that enhanced contraction response to 5-HT can occur independently of the parallel increase in EJP size.

Our only evidence on this issue is indirect. The effects which seem clearly to be mediated by post-synaptic receptors (slowing of the myogenic rhythm, increase in the rate of relaxation) have a lower threshold than the presynaptic effects of octopamine (increase in mepp frequency). The thresholds of the former are lower than the threshold for the potentiating effect on twitch tension which is close to that for the effect on mepp frequency. This evidence argues for a presynaptic mechanism for at least a part of the potentiation of twitch tension.

The octopamine receptors mediating the potentiating effects on SETi twitch tension and rate of relaxation have many similarities to those mediating the effects on the myogenic rhythm (Evans & O’Shea, 1978). The most potent agonist in all cases is synephrine, the N-methylated analogue of octopamine. The threshold effects for potentiation of SETi twitch tension and relaxation were respectively between 10⁻⁸ M and 10⁻⁷ M and between 10⁻¹⁰ M and 10⁻⁹ M for DL-synephrine. Synephrine is also a potent agonist at several other octopamine receptors such as those mediating the activation of adenylate cyclase in insect brain (Harmer & Horn, 1977) and in lobster blood cells (Battelle & Kravitz, 1978). Synephrine, however, could not be detected in significant amounts in assays of pooled DUMETi somata or in pooled lengths of extensor tibiae nerves (n 5) (Evans & O’Shea, 1978). Other biogenic amines related to octopamine were without effect on SETi twitch tension and rate of relaxation when administered at 10⁻⁶ M for 30 s. Thus of the naturally occurring biogenic amines in the insect nervous system, therefore octopamine is the most potent agonist of the receptors mediating the potentiating effects on SETi neuromuscular transmission.

Octopamine has been shown in the lobster to increase the tension generated in skeletal muscle by an excitatory motoneurone (Evans, Talamo & Kravitz, 1975; Kravitz et al. 1976). It is released into the blood from a group of peripheral octopaminergic neurones (Wallace et al. 1974; Evans et al. 1975, 1976b; Evans, Kravitz & Talamo, 1976a). Lobster skeletal muscles, however, unlike those of the locust (Evans & O’Shea, 1977, 1978) and of Aplysia (Weiss et al. 1975, 1978) are not known to receive any direct aminergic innervation but they are affected by a variety of amines which must presumably reach their targets through the blood (Dudel, 1965; Kravitz et al. 1976). Some of the active amines in the lobster such as 5-HT are better at potentiating the tension produced by the stimulation of the excitatory motoneurone than octopamine (Kravitz et al. 1976). This non-specificity of the potentiation effect is in contrast to the situation we describe here in the locust.

We examined the effects of 5-HT on neuromuscular transmission in the locust because of its known facilitatory effect on several other preparations (e.g. in lobster, Dudel, 1965; Kravitz et al. 1976; in Aplysia, Weiss et al. 1975, 1978) and because the muscle fibres of the myogenic component of the locust extensor tibiae muscle possess receptors with a high affinity for 5-HT which accelerate the frequency and increase the amplitude of the myogenic contractions (Evans & O’Shea, 1978). Serotonin however, had no measurable effect on the twitch tension or the EJP generated by the SETi motoneurone. Therefore, the receptors which mediate stimulation of the myogenic rhythm do not appear also to affect neurally evoked contractions. There is no evidence therefore that these receptors, in contrast to the octopamine receptors, extend beyond the myogenic muscle fibres.

The octopamine effects described here were all blocked by the a-adrenergic blocking agent phentolamine and not by the β -blocking agent DL-propranolol. The receptors thus possess some of the characteristics of classical vertebrate a-adrenergic receptors. In this respect they are similar to octopamine receptors on the myogenic fibres of the extensor muscle (Evans & O’Shea, 1978), those that mediate the activation of adenylate cylase (Harmer & Horn, 1977; Battelle & Kravitz, 1978) and to those suggested to be involved in the autoregulation of the peripheral octopamine cells in lobsters (Konishi & Kravitz, 1978).

Pre-synaptic a-adrenergic receptors mediate the potentiating effects of noradrenaline at vertebrate skeletal neuromuscular junctions (Bowman & Nott, 1969; Kuba, 1970) and these effects have many parallels with the action of octopamine at the SETi-neuromuscular junction of the locust. Both increase twitch tension, increase EJP size and affect the frequency but not the amplitude of spontaneous mepps. Adrenaline increases the maximum twitch tension and the time to peak tension in vertebrate fast twitch muscle (Lewis & Webb, 1976) but the rate of development of tension is not significantly affected (Bowman & Nott, 1969). This again is similar to the effect of octopamine on twitch tension at the SETi neuromuscular junction in the locust.

The effects of stimulating DUMETi, and of superfusing octopamine, are long lasting and far outlast the period of stimulation. This feature was also found in the potentiating effects of the metacerebral cells in Aplysia (Weiss et al. 1975; 1978) and for the modulatory effects of octopamine and other amines at the lobster neuromuscular junction (Kravitz et al. 1976). The modulating receptors stimulated by amines in these examples seem, therefore, to mediate processes very different from those which are stimulated by conventional neurotransmitters. The receptors stimulated by the octopamine released from DUMETi terminals may for example have access to metabolic functions, possibly by being linked to adenylate cyclase. This possibility is suggested by the similarities between octopamine receptors on locust muscle and those known to be linked to adenylate cyclase in cockroach (Harmer & Hom, 1977) and in lobster (Battelle & Kravitz, 1978). Furthermore, in both vertebrate smooth muscle (Gomperts, 1976) and cardiac muscle (Tsien, 1977) the degree of relaxation is directly proportional to the accumulation of cAMP mediated by catecholamines. It is thus tempting to speculate that both the potentiation of twitch tension and the increased rate of relaxation caused by octopamine in locust muscle are mediated via the stimulation of adenylate cyclase.

We think that the regulation of motoneurone effects is an important function of DUMETi, which may be shared by other DUM neurones in other muscles. It seems unlikely to us that the effect of DUM neurones on the myogenic rhythm found in the ( extensor muscle (Hoyle, 1974; Hoyle & O’Shea, 1974) is an important role for these cells. The octopamine receptors extend beyond the small number of myogenic fibres and the effect of octopamine on the myogenic rhythm (a slowing) is less dramatic than the accelerating effect of 5-HT (Evans & O’Shea, 1978). This fact, together with the evidence that the 5-HT receptors do not affect neurally evoked muscle activity suggesting they are confined to the myogenic fibres, is added evidence that the primary role of octopamine receptors is not the control of the myogenic rhythm; this may be the primary function of the 5-HT receptors.

The modulatory function is an important and general function which has been confirmed for biogenic amines in both vertebrates and invertebrates, in both the central and peripheral nervous systems. The various long-lasting modulatory effects of the biogenic amines have led to the idea that they may function in the long-term regulation of behavioural responsiveness or arousal (Weiss et al. 1975, 1978; Susswein, Kupfermann & Weiss, 1976). This idea is supported by recent evidence showing that amines can function in the vertebrate CNS to bias or modulate the effects of output produced, by conventional neurotransmitters (Freedman et al. 1977; Hokfelt & Fuxe, 1969; Hoffer et al. 1973). The example we describe is probably the simplest way in which the intensity of behaviour can be increased; it is done without altering the motor output but by a bias placed directly at the neuromuscular junction. We know nothing about the possible functions of DUMETi or other DUM neurones in the central nervous system, but we speculate that the neuromuscular function provides a simple model for central effects.

We thank Drs M. Burrows, M. Siegler and P. Simmons for their critical reading of the manuscript. Some of the equipment used in this study was supplied from a Nuffield Foundation Grant to Dr M. Burrows. We would like to acknowledge financial support from the SRC (MO) and ARC (PDE).