The Identification of An Octopaminergic Neurone and the Modulation of A Myogenic Rhythm in the Locust

Octopamine is a biogenic amine which occurs naturally in nervous tissue of vertebrates and invertebrates (Axelrod & Saavedra, 1977; Robertson & Juorio, 1976). In vertebrates octopaminergic neurones have not yet been identified, but in the central nervous system of the marine mollusc Aplysia (Saavedra et al. 1974) and the peripheral nervous system of the lobster (Wallace et al. 1974; Evans et al. 19766) octopamine is contained in specific neurones. In the lobster a calcium dependent release of octopamine into the haemolymph occurs from the peripherally located cells (Evans, Talamo & Kravitz, 1975; Evans, Kravitz & Talamo, 1976a), suggesting a neurosecretory role, though the precise function of octopamine remains unknown.

Octopamine is widely distributed within the nervous systems of insects (Robertson & Steele, 1973; Robertson, 1976; Evans, 1978 a) and specific octopamine activated adenylate-cyclases occur (Nathanson & Greengard, 1973; Harmer & Horn, 1977). But octopamine has not been demonstrated in identified neurones and is not established as a transmitter substance in insects. It has been suggested, however (Hoyle & Barker, 1975; Hoyle, 1975) that octopamine is the natural transmitter substance of a group of prominent neurones whose cell bodies lie dorsally and near the midline of central ganglia in the locust. These dorsal neurones were first described in the locust by Plotnikova (1969). Neurones of apparently similar morphology occur in the cockroach (Crossman et al. 1971). They are unpaired in the locust ganglia; that is, they have bifurcating axons which project symmetrically into left and right peripheral nerve roots of the ganglion. In this respect they differ from all known locust moto-neurones that are bilaterally paired and have axons leaving the central ganglia in the roots of one side only. For these reasons the dorsal cells are referred to as Dorsal Unpaired Median (or DUM) neurones (Hoyle et al. 1974). One of the cells of this group projects to the main jumping muscles (the extensor tibiae or ETi muscles) of both locust metathoracic or hind legs and has been called DUMETi (Hoyle et al.1974).

The notion that DUM neurones, including DUMETi, are octopaminergic is based on two pieces of evidence. First, tissue removed from the dorsal region of the locust metathoracic ganglion, which included somata of some DUM cells, synthesized octopamine when incubated with radioactive tyrosine (Hoyle & Barker, 1975). This suggested to Hoyle & Barker (1975) that some DUM neurones are octopaminergic. Second, the only known physiological effect of stimulating DUM cells (the slowing of an intrinsic rhythm of contraction and relaxation found in the extensor tibiae muscle of the locust metathoracic leg), can be mimicked by applying octopamine to the muscle (Hoyle, 1974, 1975). This was taken by Hoyle (1974, 1975) as evidence that the DUM neurone innervating this muscle (DUMETi) is octopaminergic. The above evidence, however, is not sufficient to establish the presence of octopaminergic neurones. In the biochemical studies (Hoyle & Barker, 1975) the specific ability of DUM cells to synthesize octopamine was inferred from analysis of heterogenous tissue containing unidentified cells. Furthermore, the only products of the incubations of whole ganglia were tyramine and octopamine; dopamine and noradrenaline are present in locust nervous tissue (Robertson, 1976) and should have been synthesized. Their absence from the labelled products (Hoyle & Barker, 1975) is therefore puzzling and suggests that the incubation conditions did not permit normal metabolism. The reported correspondence between the physiological effects of octopamine application and DUM cell stimulation on the intrinsic rhythm is also difficult to interpret. Many of the cells in the DUM cell group will cause a slowing of the rhythm (Hoyle, 1975,1975). In these experiments the DUMETi neurone was not uniquely identified and was therefore not shown to affect the rhythm. In addition, the effects of octopamine on the rhythm were not shown to be specific; indeed in the method employed, physiological saline was an equally good inhibitor of the rhythm (Hoyle, 1975).

The first aim of this paper is to demonstrate the presence of an octopaminergic neurone in insects. This is acheived by performing biochemical and physiological experiments on an individual DUM neurone, DUMETi, from the locust metathoracic ganglion. DUMETi neurones are identified by physiological criteria and are shown to contain octopamine by a radio-enzymatic assay. We show that octopamine is the most likely transmitter of DUMETi by comparing the effect on the rhythm of stimulating DUMETi with that of applying octopamine and a wide range of other monoamines. We eliminate the possibility that the effects of DUMETi and octopamine on the rhythm are mediated indirectly, via peripheral effects on the moto-neurones to the extensor muscle, and conclude that they act directly on the muscle fibres.

The second aim is to show that the effects of octopamine are mediated by at least two aminergic receptors. One has a high-affinity for octopamine and mediates a slowing of the intrinsic rhythm, and the other has a low-affinity for octopamine and accelerates the rhythm. A blood-borne factor, possibly 5-hydroxytryptamine, is implicated as the natural source of stimulation for the accelerating receptor. The possibility of blood-borne modulation is discussed in relation to the possible function of the rhythm.

The preparation we describe here includes three identified motoneurones (Pearson & Bergman, 1969; Cochrane, Elder & Usherwood, 1972; Hoyle & Burrows, 1973) and a single DUM neurone, DUMETi (Hoyle et al. 1974); it therefore provides a simple system on which to investigate the action and function of an identified aminergic neurone. In a subsequent paper (O’Shea & Evans, in preparation) we will show that DUMETi not only affects the rhythm in the extensor tibiae muscle but also functions as a modulator of neuromuscular transmission. A preliminary account of some of this work has already been published (Evans & O’Shea, 1977).

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

Animals were restrained on a dish filled with periphery wax (Surgident; Lactona Corp.) such that their dorsal surfaces were uppermost and their metathoracic tibiae could move freely. The ventral nerve cord was exposed by dissection from the dorsal surface. The physiological recording arrangements are shown in Fig. 1. Intracellular recordings were made from DUM neurone somata of the metathoracic ganglion using glass microelectrodes filled with 2 M-potassium acetate and having resistances of 20−50 MΩ. Current could be injected through the electrode via a bridge circuit. Paired silver hook electrodes were used to record differentially from nerves and to stimulate them. Paired electrodes were attached to the right, and the left extensor tibiae nerve (nerve 5b₁ ; nomenclature of Pringle, 1939), which lie on the surface of the exposed metathoracic extensor-tibiae muscles. A third pair of electrodes attached to nerve 3 inside the body cavity was used to stimulate the slow extensor motoneurone (SETi) to the extensor-tibiae muscle. The intrinsic rhythm of contraction and relaxation in the extensor tibiae muscle of the locust metathoracic leg was monitored by recording muscle tension almost isometrically with a force transducer attached to the distal end of the muscle tendon (apodeme). The effect of stimulating an identified DUMETi neurone (see Results for identification of DUMETi) on the frequency and amplitude of the intrinsic rhythm was investigated. The effect of stimulating SETi on the rhythm was also examined. Results were recorded on magnetic tape and filmed after the experiment.

DUMETi somata were identified physiologically as described in the Results section. In experiments where the DUMETi soma was to be isolated for octopamine analysis, the microelectrodes used in the identification were coated with Indian ink. A black spot was left on the dorsal surface of the metathoracic ganglion at each recording site, when the electrode was withdrawn. The DUMETi recording site was drawn in relation to the landmarks provided by the tracheal supply to the ganglion. When the isolated ganglion was subsequently viewed under darkfield illumination the DUM neurone somata were clearly visible and the soma of DUMETi was located by reference to the recording site. The ganglion was desheathed without disturbing the DUM neurones and the DUMETi soma teased away from neighbouring neurons using hand-held tungsten needles which had been electrolytically sharpened. The isolated soma was cleaned of adhering cells, pigment granules and connective tissue. An isolated DUMETi soma is shown in Fig. 2. This soma is pear shaped with a maximum diameter of 65 μm. Alongside it is placed, for comparison, the soma of a physiologically identified motoneurone, from the same ganglion, the fast extensor of the tibia (FETi). The FETi soma has a diameter of 100 μm measured along its longest diameter. This is the largest motoneurone in the metathoracic ganglion and its soma has a markedly flattened shape (Burrows & Hoyle, 1973). The volumes of the isolated somata were calculated from measurements made on the identified cells immediately after isolation.

Isolated identified somata were transferred by use of a fine glass capillary tube (Otsuka, Kravitz & Potter, 1967) to an extraction tube containing 10 μl of 0·01 N-formic acid. This procedure was performed under a dissecting microscope so that the soma was visible at all times during the transfer. The extraction tube was freeze-thawed several times before being stored at – 20°C prior to analysis. Octopamine assays were performed using a modification (Evans, 1978 a) of the radio-enzymatic assay method described by Molinoff, Landsberg & Axelrod (1969). This gave the assay a twice background sensitivity of 10−15 Pg (∽0·1 pmol) based on a standard curve devised for authentic DL-octopamine. The identity of the assay products was examined for each tissue, by chromatographic separation in several different solvent systems (see Evans, 1978 a for details of solvent systems used) and by comparison with authentic standards.

Neutral red-stained preparations were made by incubating desheathed locust metathoracic ganglia in isotonic locust saline containing 0·01 mg/ml neutral red (Sigma) (Stuart, Hudspeth & Hall, 1974) for 3 h at room temperature or overnight at 4 °C. Neutral red solutions were filtered immediately before use.

The response of the intrinsic rhythm to octopamine and other related amines was investigated in an isolated leg where solutions of known concentration could be superfused directly onto the myogenic portion of the extensor muscle. The intrinsic rhythm persists when the leg is detached from the animal, showing a comparatively constant frequency ( ∽ 0·2 Hz). This contrasts with the rhythm observed when the leg is attached to the animal, for under these circumstances it accelerates, slows or even stops periodically. Although the rhythm continues in the absence of connexion to the central nervous system, and is not dependent on the muscle’s innervation, it is nevertheless capable of being modulated (Hoyle & O’Shea, 1974). The intrinsic rhythm is confined to a bundle of tonic muscle fibres at the proximal end of the extensor muscle, where the tendon attaches to the cuticle of the femur (Burns & Usherwood, pers. comm.). The experimental arrangement is shown in Fig. 3. The isolated locust metathoracic leg was positioned with its ventral surface uppermost. A small flap of cuticle was then cut out over the myogenic muscle bundle and the flap raised without damaging the underlying muscle fibres. This exposed the surface of the myogenic bundle which was then superfused with saline at a constant rate of 1 ml/min by use of a peristaltic pump. Known concentrations of octopamine, other amines, adrenergic blocking agents and putative motoneurone transmitters were then introduced into the superfusate for varying lengths of time and their effects on the intrinsic rhythm monitored. This has the advantages over the method employed by Hoyle (1975), applying drugs in a small drop of saline through a hole in the leg, that known concentrations of drugs can be applied directly to the myogenic bundle for known lengths of time.

Isotonic physiological saline contained: 140 mm-NaCl; 10 mm-KCl; 4 mm-CaCl₂; 4mm-NaHCO₃; 6 mm-NaH₂PO₄ (Usherwood & Grundfest, 1965) plus 90 mm-sucrose (pH = 6·8).

We would like to acknowledge the following gifts: phenoxybenzamine HC1 (Smith, Kline & French); N,N-dimethyloctopamine (Sterling-Winthrop Res. Inst.); and gramine (Dr M. J. Berridge). All other drugs were obtained from Sigma Chemical Co. except for phentolamine mesylate (Ciba) and phenylethanolamine HC1 and 2-phenyl-ethylamine HC1 (Regis Chemical Co.).

The extensor tibiae (ETi) muscle of the hindleg of the locust is innervated by the branched extensor nerve (nerve 5b1) (see Fig. 4). Proximally this nerve contains the axons of three motoneurones in addition to that of the DUMETi neurone (Hoyle et al. 1974). These are the fast and slow extensor motoneurones (FETi and SETi) and an inhibitory motoneurone which innervates many muscles and is therefore called the common inhibitory (CI) motoneurone (Usherwood & Grundfest, 1965; Pearson & Bergman, 1969; Cochrane et al. 1972; Hoyle & Burrows, 1973). The motoneurone cell bodies lie in the metathoracic ganglion and are paired; one cell from each pair innervates the left and the other the right ETi muscle. In contrast, DUMETi is an unpaired cell and its axon bifurcates to send branches to the ETi muscle in both the right and left legs (Hoyle et al. 1974). Extracellular hook electrodes placed on the proximal regions of one extensor nerve therefore record activity from four axons (i.e. those of the three motoneurones and one branch of DUMETi). The bilateral distribution of DUMETi axons can be used as a criterion for the physiological identification of its soma. To locate DUMETi, DUM somata were penetrated randomly with an intracellular glass electrode until one was found where spikes initiated in its soma, by depolarizing current passed through the microelectrode, were transmitted 1:1 to both left and right extensor nerves (Fig. 5 A). This distinguishes the DUMETi neurone soma from that of other DUM cells since the activity in the identified cell can account for all the spikes found simultaneously in the left and right extensor nerves. If the oscilloscope is triggered by spikes initiated and recorded in the soma of DUMETi, then the spikes in the left and right axons superimpose (Fig. 5 B). The difference in latency between spikes in the left and right nerves is caused by differences in the distance from the soma of the recording site. DUMETi soma spikes can be initiated antidromically by stimulating either the left or right ETi nerves (Fig. 5 C) indicating a direct connexion between the soma and axons rather than a synaptic connexion. If the soma is hyperpolarized the active soma spike (Kerkut, Pitman & Walker, 1969) fails to initiate in response to antidromic stimulation (R1 lower trace, Fig. 5C), and the smaller positive potential seen in the soma is a recording of the axon spike conducted passively from its site of initiation. In repeated experiments in which many DUM cells were penetrated in single metathoracic ganglia, only one was found to meet the above criteria for the identification of DUMETi. This, in addition to the fact that antidromic stimulation of one extensor tibiae nerve initiates a single spike in the other extensor nerve, suggests that in each animal there is one DUM neurone that projects to the metathoracic extensor-tibiae muscles.

Two small parts of the ETi muscle (commonly known as the ‘accessory extensor muscle ‘) arise in the distal part of the femur from the dorsal wall and are inserted onto the base of the main extensor tendon (Snodgrass, 1935) (see Fig. 4). The accessory extensor muscle is separated from the main part of the ETi muscle by about 3 mm. We have only been able to record activity from the axons of SETi and CI in the accessory branch of the extensor nerve. It is not possible to excite the soma of DUMETi by antidromic stimulation or to record its axon spike from this accessory branch. FETi activity is absent from the accessory branch. Thus the axons of the FETi and DUMETi neurones are confined to the main proximal part of the extensor nerve and the accessory branch contains only the axons of SETi and CI. This fortunate anatomical arrangement allowed us to demonstrate the exclusive association between octopamine and the DUMETi axon (see Results, section 3).

The sites of penetration of DUMETi somata in 50 different preparations is shown in Fig. 6. The variation of position shown is far greater than that found for the cell bodies of identified motoneurones in the same ganglion (Burrows & Hoyle, 1973). Physiological recording sites alone, however, do not distinguish between individual variation in the position of the DUM cell group, within which the cells retain a fixed relation, and individual variation of the relative positions of the somata within the group.

DUM cells in the locust stain with neutral red (Evans & O’Shea, 1977) a dye which selectively stains monoamine containing neurones in the leech (Stuart et al. 1974) and the lobster (Wallace et al. 1974). Neutral red-stained metathoracic ganglia (Fig. 7A-E) show that variation exists in both the position and organization of the DUM cell group. Two size classes of DUM cell bodies are revealed by this staining procedure (Fig. 7). The large somata are 40−80 μm in diameter and the small somata are about 30 μm in diameter. Typically 8 to 11 large and 14 to 21 small cells are stained in each preparation. The variability in cell number might reflect a true variation in the total cell numbers of different ganglia or result from a variability in staining properties. Since only rarely did we fail to locate the DUMETi soma in the DUM cell group we favour the latter explanation. There is, however, a precedent for variations in the numbers of octopamine cells. In the lobster peripheral octopamine cells are known to be variable in this respect (Wallace et al. 1974; Evans et al. 1976a). Thus physiological identification of the DUMETi soma was required in each preparation because soma size, position in the ganglion and soma position in relation to the other DUM somata are all variable and cannot be used as criteria for identification.

A single DUMETi soma, when assayed for octopamine, gave a reading which was just twice the background of the assay. In contrast, a pool of 47 FETi somata gave no reading above background. The mean amount of octopamine per DUMETi soma ± S.E. was determined by assaying pools of DUMETi somata (see Table 1) and is 0·099 ± 0·02 pmol. As a control, 2 μl of the saline used to isolate the DUMETi somata was assayed, and no octopamine was detected. The authenticity of the assay products was confirmed by chromatography in several solvent systems, the results from one of which are shown in Fig. 8. Most of a 1·0 ng standard of authentic DL-octopamine, when assayed, was N-methylated to synephrine and a small amount was dimethylated to N,N-dimethyloctopamine (Fig. 8B). The bulk of the radioactivity of the DUMETi pool assay product (Fig. 8 A) was found to co-chromatograph with authentic synephrine. Less than 12 % of the recoverable radioactivity co-chromatographed with dimethyloctopamine. This suggests that the assay is measuring true endogenous octopamine in the DUMETi pool and that the pool contains little or no endogenous synephrine which would have shown up as N,N-dimethyloctopamine in the assay product.

Octopamine assays were also performed on pieces of the extensor nerve from both the proximal and distal regions of the extensor muscle (see Fig. 4). Chromatography of the assay products (Fig. 8C) confirms the presence of octopamine in the proximal region of the extensor nerve branch. We have shown that the latter branch contains axons of SETi and CI (see Results, section 1). Thus the octopamine detected in the proximal extensor nerve must be associated with the axon of DUMETi since the only remaining possibility, that it is associated with the axon of FETi, has been eliminated on the grounds that octopamine is undetectable in that cell’s soma.

If we assume an average diameter of 65 μm for a single DUMETi soma we can calculate that the concentration of octopamine in a DUMETi soma will be 8·9×10⁻⁴M. Also assuming a 1μm average diameter for the peripheral axon of DUMETi (unpublished observations) in the region of the extensor nerve analysed, we can calculate the octopamine concentration in the DUMETi axon to be 3·3×10⁻³ M. Thus octopamine appears to be about four times more concentrated in the axon of DUMETi than in its soma.

The case for octopamine being the natural transmitter of the DUMETi neurone would be strengthened if octopamine could be shown to mimic the physiological effects of stimulating DUMETi. The only known physiological effect of stimulating DUM cells is the slowing of an intrinsic rhythm of contraction and relaxation found in the extensor-tibiae muscle of the locust metathoracic leg (Hoyle, 1974). The rhythm is brought about by a small discrete bundle of muscle fibres attached to the proximal end of the extensor apodeme. This bundle corresponds to the bundle of tonic fibres described by Cochrane et al. (1972) which receive motor innervation from only the slow excitor and the common inhibitor. It has been reported previously that the rhythm is abolished by saline (Hoyle & O’Shea, 1974; Hoyle, 1975). We have now overcome this problem by the use of isotonic saline and have therefore been able to use the rhythm to assay the effect of various drugs. In fact it is even possible to remove the bundle of contractile fibres from the rest of the muscle without affecting the contractile rhythm (Burns & Usherwood, pers. comm; Berridge, pers. comm. ; and our unpublished observations.). The intrinsic rhythm has been proposed to be myogenic (Hoyle & O’Shea, 1974; Burns & Usherwood, pers. comm.).

When the DUMETi neurone is stimulated at 1 Hz both the frequency and the amplitude of the myogenic contractures are reduced (Fig. 9) and there is a slight reduction in the basal tension which is normally present in the extensor tibiae muscle. Stimulation of unidentified DUM cells (Hoyle, 1974, 1975) can have similar but weaker effects with more prolonged and higher frequencies of stimulation. A single spike in DUMETi causes a significant lengthening of the period of the rhythm. The interval between contractures during which a DUMETi spike was given (arrowed Fig. 9) was 21·8 s compared to a mean interval + s.E. for the five preceding contractions of 20·8 ± 0·3 s.

To test the hypothesis that octopamine released from the endings of DUMETi is responsible for the above effects, we superfused octopamine directly on to the myogenic bundle of fibres. The effect of a 2 min pulse of 10⁻⁶ M DL-octopamine superfused on to the extensor muscle is shown in Fig. 10A. The rhythm is inhibited and returns 56 s after the muscle is washed with saline. In zero calcium saline the rhythm is speeded up and a pulse of 10⁻⁶ M DL-octopamine now fails to abolish the rhythm completely (Fig. 11). Instead the frequency and amplitude become highly irregular suggesting that the normal response to octopamine is dependent on the presence of calcium (Fig. 11). In normal saline, low concentrations of octopamine (10⁻⁹ − 10⁻⁷ M) reduce the frequency of the rhythm, while higher concentrations (10⁻⁷−8 × 10⁻⁷ M) reduce both the frequency and the amplitude of the rhythmic contractures and produce a small decline in basal tonus. All effects of DUMETi stimulation upon the rhythm can therefore be mimicked by the application of octopamine. The doseresponse curve for the effect of superfused octopamine on the frequency of the rhythm is shown in Fig. 12. The apparent threshold for octopamine action on the rhythm lies between 10⁻⁹ and 10⁻¹⁰ M and the myogenic contractions are completely abolished by 8 × 10⁻⁷ M. The discontinuity seen in the dose-response curve at about 10⁻⁷ M shows that at higher concentrations octopamine is relatively less effective in slowing the rhythm. One possible explanation of this effect is that a second receptor which accelerates the rhythm is stimulated by octopamine at high concentrations thereby competing with the slowing effect. Indeed in experiments in which the normal slowing response to octopamine is blocked (see below) the presence of an accelerating receptor is revealed.

The nature of the octopamine receptor mediating the slowing effect on the myogenic rhythm was investigated by examining the ability of various drugs to block the action of superfused octopamine. The drugs used were the α-adrenergic blocking agents phentolamine and phenoxybenzamine and the β-adrenergic blocking agents propranolol and dichloroisoproterenol.

The effect of octopamine on the myogenic rhythm was not blocked when octopamine (10⁻⁶ M) was applied to the muscle in the presence of 10⁻⁶ M DL-propranolol (Fig. 10; trace B). When the same concentration of octopamine was subsequently applied in the presence of 10⁻⁸ M-phentolamine the slowing action of octopamine was blocked and a slight initial accelerating effect was observed (Fig. 10; trace C). The magnitude of the accelerating effect seen under these conditions is variable. A more pronounced example is shown in Fig. 15. In other experiments phentolamine at concentrations as low as 10⁻¹⁰ M gave some protection to the rhythm against 10⁻⁶ M-octopamine. Phentolamine once applied to the preparation becomes tightly bound to the receptor and to wash it off requires long periods of superfusion (see Fig. 15). Phenoxybenzamine (10⁻⁸ M) was found also to block the action of 10⁻⁶ M octopamine whereas 10⁻⁶ M-dichloroisoproterenol was not an effective blocking agent.

The above experiments indicate that α-adrenergic blocking agents are capable of blocking the slowing response of the myogenic rhythm to octopamine and show that the receptor mediating this must have some structural similarities to the vertebrate α-adrenergic receptor. The experiments also revealed the presence of a receptor not blocked by the α- and β-blocking agents; in response to high concentrations of octopamine this receptor mediates an acceleration of the myogenic rhythm (see Results, section 7).

Octopamine can mimic the effects of DUMETi stimulation which suggests that it is the natural neurotransmitter of DUMETi. Further support for this idea would be provided if the response were shown to be specifically sensitive to octopamine. To investigate this, we superfused the muscle with a series of amines related to octopamine and tested their ability to mimic the effect of DUMETi stimulation. The results are expressed as the percentage reduction in rhythm frequency produced by the application of each amine at a concentration of 10⁻⁶ M (Table 2). Only DL-syn-ephrine and phenylethanolamine, besides DL-octopamine, completely abolished the rhythm at a concentration of 10⁻⁶ M. The relative potencies of these three amines were determined from dose-response curves. The most potent was DL-synephrine (threshold between 10⁻¹⁰ and 10⁻¹¹ M), followed by DL-octopamine, then phenylethanolamine (thresholds between 10⁻⁹ and 10⁻¹⁰ M, and between 10⁻⁶ and 10⁻⁷ M respectively). The naturally occurring D(– ) isomer of octopamine is, however, over 200 times as potent as the L( +) isomer in activating an adenylate cyclase in cockroach brain (Harmer & Horn, 1977). Thus the thresholds of the response of the myogenic rhythm to D(–) synephrine and to D(– ) octopamine will be much lower than those observed with the DL-compounds used in the present study. Other amines were much less potent or as in the case of 5-hydroxytryptamine (5HT), an amine present in large quantities in insect nervous tissue (Hiiipi & Rózsa, 1973; Kusch, 1975), had the opposite effect and produced an acceleration of the rhythm (Fig. 13).

At a concentration of 10⁻⁶ M, 5HT dramatically increases the frequency of the rhythm (an 8-fold stimulation) and slightly increases the amplitude of contractures (Fig. 13 A). Its ability to accelerate the rhythm at lower concentrations is shown in a dose-response curve (Fig. 14). At higher concentrations 5HT markedly increases the amplitude as well as the frequency of the contractions. At 2 × 10⁻⁶ M (Fig. 13 B) and 5×10⁻⁶ M (Fig. 13C) ‘giant’ contractions are initiated which in duration appear to be multiples of the normal contractions. When they occur in long trains they therefore appear to divide the frequency of the rhythm (Fig. 13 C). Maintained contractures of the muscle were induced by io^-5 M-5HT, an effect which was reversed when the muscle was washed in saline.

These results show, first, that the receptor mediating the slowing response of the rhythm is maximally sensitive to monophenolic amines and, with the exception of synephrine which was undetectable in DUMETi (see Results, section 3), that octopamine was the most potent amine tested. Secondly, that acceleration of the rhythm can be caused by both octopamine and 5HT. In the next section both accelerating effects are shown to be mediated by the same receptor.

In experiments in which the muscle was first exposed to the a-adrenergic blocking agent phentolamine the subsequent application of octopamine produced an increase in frequency of the myogenic rhythm (see Figs. 10C and 15). This suggests that when the receptor mediating the slowing of the rhythm is blocked, octopamine activates another receptor which accelerates the rhythm. Either a second octopamine receptor is unmasked or octopamine under these circumstances is activating the same receptor as 5HT. The hypothesis that the accelerating response to octopamine and 5HT is mediated by the same receptor was tested by investigating the accelerating effect of octopamine in the presence of gramine, a known 5HT blocking agent (Berridge,1972). When the octopamine receptor mediating slowing is blocked by phentolamine and the 5HT receptor is then blocked with gramine, octopamine does not produce an acceleration and when gramine is washed from the muscle the accelerating effect returns (Fig. 15). It appears therefore that when the high-affinity binding site for octopamine is blocked by phentolamine, an effect of octopamine on the 5HT receptor is revealed. The affinity of octopamine for the 5 HT receptor is much lower than for its own receptor, so normally its action on the 5HT receptor is masked. This effect may explain the ‘shoulder’ on the octopamine dose-response curve (Fig. 12), i.e. at higher concentrations some effect of octopamine on the 5HT receptor is being expressed. Indeed when the effect of octopamine was tested in the presence of 10⁻⁶ M-gramine the shoulder on the dose-response curve was reduced (Fig. 12).

The action of octopamine on the myogenic rhythm could be directly on the muscle fibres or, alternatively, mediated presynaptically by affecting endings of the slow extensor (SETi) or common inhibitory (CI) motoneurones, which also innervate the fibres of the myogenic bundle (Cochrane et al. 1972; Burns & Usherwood, pers. comm.). To distinguish between these possibilities the effects on the myogenic rhythm of the putative excitatory neuromuscular transmitter of insects, glutamate, and the inhibitory neuromuscular transmitter, gamma-aminobutyric acid (GABA) were examined (Fig. 16). At high concentrations of glutamate (10⁻⁴ M) the myogenic rhythm was completely abolished, and a series of contractures elicited. At lower concentrations (5×10⁻⁵ M) the amplitude of myogenic contractions was reduced but the frequency of the rhythm increased (Fig. 16, trace A). This is similar to the effect of firing SETi (Fig. 17) and strengthens the evidence that glutamate is the neurotransmitter of this motoneurone. Lower concentrations of glutamate (10⁻⁶ M) slightly decreased both the frequency and amplitude of the rhythm.

At 5×10⁻⁵ M, GABA also reduced the amplitude of the contractures but slightly decreased the frequency of the rhythm (Fig. 16, trace C). The action of GABA could be blocked by picrotoxin (5×10⁻⁴ M) (Fig. 16, trace B) without affecting the ability of octopamine (10⁻⁶ M) to slow the rhythm. The time course for the actions of both GABA and glutamate is quite different to that of octopamine. All produce an immediate effect on the rhythm which for GABA and glutamate is terminated rapidly when the muscle is washed with saline. In contrast, octopamine has an effect which long outlasts its presence in the superfusate. The most likely site for the action of octopamine is, therefore, directly on the muscle fibres of the myogenic bundle, because the action of octopamine cannot be mimicked by glutamate or GABA.

We have shown the presence of an octopaminergic neurone in an insect: the dorsal unpaired median neurone (DUMETi) which projects to the extensor tibiae muscles of the locust metathoracic legs (Hoyle et al. 1974). Its soma was physiologically identified, isolated and shown to contain about 0·1 pmol of endogenous octopamine. Octopamine is about four times more concentrated in the axon than in the soma. This finding agrees well with the distribution of amines between axon and soma in other adrenergic neurones, such as the distribution of noradrenaline in the cell bodies and terminal regions of vertebrate sympathetic neurones (Norberg & Hamberger, 1964; Dahlstrom & Haggendal, 1966). In comparison with the soma of an identified motoneurone (FETi: Hoyle & Burrows, 1973), octopamine is at least 800 times more concentrated in the soma of DUMETi.

We have demonstrated that the firing of DUMETi at low frequencies slows a myogenic rhythm in the extensor tibiae muscle of the locust metathoracic leg and than exogenously applied octopamine mimics this effect. This together with the finding of endogenous octopamine in DUMETi argues strongly for the positive identification of DUMETi as an octopaminergic neurone.

The myogenic bundle which produces the rhythm of contraction and relaxation is shown to possess at least two types of aminergic receptors, one which slows and one which accelerates the rhythm, in addition to receptors for GABA and glutamate. The aminergic receptor which slows the rhythm is maximally sensitive to monophenolic amines and exhibits some of the characteristics of the vertebrate α-adrenergic receptor. This finding contrasts with that of Hoyle (1975) who found that α-adrenergic blocking agents were more effective than α-adrenergic blocking agents in inhibiting the effects of amines on the rhythm. We cannot explain this discrepancy because Hoyle (1975) did not specify the blocking agents used or their concentrations. Our findings are, however, in agreement with two studies on specific octopamine-activated adenylate cyclases, one in lobster blood cells (Battelle & Kravitz, in preparation) and one in cockroach brain (Harmer & Horn, 1977), both of which showed that α-adrenergic blocking agents were more effective than β-adrenergic blocking agents in blocking octopamine receptors.

The structure specificity for the receptor mediating the slowing of the rhythm is similar to that found for the receptor mediating the activation of the octopamine-sensitive adenylate cyclases mentioned above (Battelle & Kravitz, in preparation; Harmer & Horn, 1977). This suggests the possibility that the effects on the myogenic rhythm are also mediated via the activation of a specific octopamine-sensitive adenylate cyclase. In our system, however, DL-synephrine is a more potent agonist than DL-octopamine, whereas in those of Battelle & Kravitz (in preparation) and Harmer & Horn (1977), synephrine is not as potent as octopamine. One possible explanation for this discrepancy is the presence of a more efficient inactivation mechanism for octo - pamine in our relatively intact system. It is known that insect nerve cord possesses a high-affinity uptake mechanism for octopamine and that the receptor mediating the uptake is decreasingly sensitive to the monophenolic amines in the order tyramine, octopamine and synephrine (Evans, 1978b). Thus it is possible that DL-synephrine is a more potent agonist of the octopamine receptor than octopamine, because octopamine is more rapidly inactivated than synephrine. The discrepancy therefore does not argue against the presence of an octopamine sensitive adenylate cyclase in our preparation. Indeed the activation of an adenylate cyclase by an a-adrenergic receptor has been suggested in the fire-fly light organ (Oertel & Case, 1976). This system is similar to ours in that synephrine is a potent agonist (Carlson, 1968) but is not an endogenous constituent of the light organ containing segments as is octopamine (Robertson & Carlson, 1976).

The aminergic receptor which accelerates the rhythm has a low affinity for octopamine and a high affinity for the indolalkylamine, 5-hydroxytryptamine. The function of the accelerating receptor is not clear. None of the neurones innervating the myogenic bundle appear significantly to activate it and it is therefore likely to be the target of a blood-borne factor. This factor may be an amine like 5HT, which at low concentrations is known to accelerate the beating of the insect heart (Miller, 1975) and is thought to be a neurohormone in insects (Berridge & Prince, 1972). This evidence alone, however, does not mean that 5HT is the natural activator of the myogenic rhythm. The presence of naturally occurring modulation of the rhythm is suggested by the observation that in the intact animal, in the absence of activity in the extensor nerve, the frequency of the rhythm is highly variable (Hoyle & O’Shea, 1974; our unpublished observations). In addition, marked acceleration of the rhythm occurs when locust blood is applied to the myogenic bundle in saline superfused preparations (unpublished observation). The possible role of modulation can only be understood in relation to the function of the rhythm.

Two types of function have been suggested for the rhythm. Hoyle (1974) suggested it might be an intrinsic exercise rhythm. Usherwood’s suggestion that it aids blood-flow along the metathoracic leg (Usherwood, 1974), however, seems more likely to us. We observe that the myogenic rhythm causes a cycle of inflation and deflation of a proximal air sac associated with the myogenic bundle (see Fig. 3) and of distal tracheae (unpublished observation) suggesting that the rhythm, in addition, assists in the ventilation of the leg. The most likely function for the myogenic rhythm is probably related to the specialized form of the locust hind leg-rhythms are not found in the equivalent muscles of the other legs. The slender, elongated hind leg might require more blood- and air-flow than can be provided by the heart and abdominal pumping. An accessory pumping structure in the leg would fulfil this requirement. Indeed it is to be noted that myogenic pumping structures are found at the bases of many insect appendages including the antennae and wings (see Wigglesworth, 1965 for references). Pulsatile myogenic organs have also been described in the legs of many other insects, including the tibiae of aquatic Hemiptera (Behn, 1835) and the tarsus of Tettigonia (Verlooren, 1847). The function of pulsatile myogenic organs in the legs of insects is possibly best summarized by a quote from Mitchell (1859) referring to an insect of the family Nepadae: ‘but if we sought for a reason why this insect is furnished with such an unusual organ, I think it might be found in the slow pulsation of the dorsal vessel and the languid circulation in the body, rendering some additional force necessary to impel the blood to the extremity of its long and slender limbs’. So the function of frequency modulating receptors on the myogenic bundle may be to enable the rhythm to respond to the respiratory and/or metabolic needs of the leg.

Our evidence suggests that the effects of DUMETi and octopamine on the rhythm occur by direct activation of aminergic receptors on the surface of the muscle fibres of the myogenic bundle. We were able to reject as a possible mechanism a transmitter release from motoneurone terminals. The time course of the rhythm’s response to GABA and glutamate is different from that to DUMETi or octopamine. Furthermore, picrotoxin at concentrations which block the action of GABA does not affect the response of the rhythm to octopamine. The effects of SETi stimulation and glutamate application were similar (see Fig. 18). In contrast Usherwood (1974) reports that SETi stimulation results in an increase in the amplitude of the rhythm. Stimulation of the CI results in a decrease in the amplitude and an acceleration of the rhythm (Hoyle, 1974; Usherwood, 1974). In our experiments, the putative transmitter of CI (GABA) produces a decrease in amplitude but at 5×10⁻⁵M does not increase the rhythm frequency. This is not a serious discrepancy because both increases and decreases in frequency have been observed in response to different concentrations of glutamate (see Results). The effects of various transmitters and neurohormones together with the sources of the active agents and blocking drugs for their responses are summarized in Fig. 18.

It should be emphasized here that the function of DUMETi and other DUM cells is unlikely to be concerned solely with the modulation of the myogenic rhythm. First, on physiological and biochemical evidence we have shown that the axon of DUMETi projects to parts of the metathoracic extensor tibiae muscle that do not exhibit the myogenic rhythm. Second, although DUM cells exist in other thoracic ganglia of the locust, the rhythm in the extensor tibiae muscle is confined to the metathoracic segment (Hoyle & O’Shea, 1974). Third, DUM cells in the cockroach must have a different function because myogenic rhythms are apparently absent from the leg muscles of this animal (Hoyle & O’Shea, 1974). Thus it seems likely that DUMETi has other roles in different parts of the extensor muscle ; roles which might be shared by other DUM cells. A neuromodulatory role for octopamine released from DUMETi at the locust neuromuscular junction has already been briefly described (Evans & O’Shea, 1977) and will be reported in more detail in a subsequent publication (O’Shea & Evans, in preparation).

We thank Drs M. Burrows and M. Siegler for their critical reading of the manuscript and Mr J. W. Rodford for drawing Figs. 3 and 4. Much of the equipment used in this study was supplied from a Nuffield Foundation Grant to Dr M. Burrows. Other equipment was loaned by the ARC Unit of Invertebrate Chemistry and Physiology for which we are grateful to the Director, Dr J. E. Treherne. We would like to acknowledge financial support from the ARC (PDE) and SRC (MO).