Unpaired Median Neurones in a Lepidopteran Larva (Antheraea Pernyi) Ii

J. exp. Biol. 136, 333-350 (1988) 333 Printed in Great Britain © The Company of Biologists Limited 1988

UNPAIRED MEDIAN NEURONES IN A LEPIDOPTERAN LARVA (ANTHERAEA PERNYI) II. PERIPHERAL EFFECTS AND PHARMACOLOGY

BY S. J. H. BROOKES* Department of Zoology, University of Bristol, UK

Accepted 5 January 1988

Summary Two unpaired median cells (MCI and MC2) had a temporal pattern of firing that correlated with phasic muscular activity in preparations of larval Antheraea pernyi, and previous work has indicated that the axons of median cells are associated with nerve trunks innervating blocks of muscle. In spite of this, action potentials in median cells were not found to have any one-for-one effects on either the tension or the electrical activity of somatic muscle fibres. However, bursts of action potentials in MC2 were shown to modulate both tension production and electrophysiological activity of a number of motor units. These effects consisted of an increase in twitch tension, a relaxation of basal resting tension, an increase in relaxation rate following contractions, a hyperpolarization of some muscle fibres and an increase in amplitude of excitatory junction potentials. The relative potency of these different effects varied between fast and slow muscles. All of these effects were mimicked by the application of octopamine and synephrine, and in higher concentrations by a number of other biogenic amines and adrenergic agonists. The possibility that the effects of median cell activity were mediated by the release of endogenous octopamine was supported by the observation that phentolamine (10~5molP1) blocked the effects of both MC2 impulses and the application of exogenous octopamine, whereas propanolol affected neither set of responses. This observation also indicated a pharmacological similarity with a number of other octopamine-sensitive insect tissue preparations. MCI had similar effects to MC2 on the electrical activity of a number of muscles, suggesting that these two cells play a similar role. These observations provide strong evidence for the presence of an identifiable octopaminergic system of neurones, similar to the dorsal unpaired median (DUM) neurones that have been extensively studied in the locust.

Introduction The phenomenon of neuromodulatory control of muscular activity is now well- established in invertebrate preparations. Serotonergic neurones in Aplysia and in

•Present address: Department of Human Physiology, Flinders Medical Centre, Bedford Park, South Australia 5042.

Key words: unpaired median neurones, octopamine, modulation, Lepidoptera. 334 S. J. H. BROOKES leeches have been demonstrated to modify muscle contractions evoked by motoneurone activity (Weiss, Cohen & Kupfermann, 1975, 1978; Mason & Kristan, 1982). In insects, the DUM (dorsal unpaired median) cell system has been extensively studied in locusts and one neurone (DUMETi) has been shown to have neurosecretory-type nerve endings in a skeletal muscle (Hoyle, Dagan, Moberly & Colquhoun, 1974). DUMETi has been shown to modulate the myogenic activity of a specialized group of fibres in the extensor tibialis (Hoyle, 1975; Evans & O'Shea, 1978) and to modulate neuromuscular transmission and tension development (Evans & O'Shea, 1977; O'Shea & Evans, 1979). Suggestions that DUMETi acted through the release of octopamine have been confirmed (Morton & Evans, 1984), and several subtypes of octopamine receptors have been reported to mediate the effects on the muscle and nerve terminals of the extensor tibialis preparation (Evans, 1981). In the lepidopteran Manduca sexta, exogenous octopamine can also modulate neuromuscular transmission in the dorsal longitudinal flight muscles of immature and adult moths, in a manner similar to that reported in locusts (Klaassen & Kammer, 1985). Octopamine is present in the haemolymph of Manduca sexta in physiologically significant quantities (Klaassen & Kammer, 1985; Davenport & Wright, 1986). Octopamine can be synthesized in nervous tissue of Lepidoptera (Maxwell, Tait & Hildebrand, 1978) and can be detected in many parts of the central nervous system (Davenport & Wright, 1986). In spite of this circumstantial evidence for a physiological role for octopamine in Lepidoptera, the source of this biogenic amine in moths and caterpillars is a cause of speculation. The relative importance of circulating octopamine and the possibility of direct octopaminergic innervation of muscles remain in question. The aim of this study was to test the hypothesis that the unpaired median cells (MCI and MC2) were involved in the modulation of muscle activity in a manner similar to that of DUMETi in the locust extensor tibialis. Additionally, the possibility that the peripheral effects might be mediated through the release of octopamine was investigated.

Materials and methods Fifth-instar insect larvae were prepared for neurophysiological recording in the way described previously (Brookes & Weevers, 1988). The entire central nervous system (apart from the brain and the segmental ganglion being studied) was removed to minimize spontaneous muscular activity. The gut and silk glands were removed and the fat body and heart tissue were dissected away from the segment under study (usually segments 4 or 5). Care was taken to avoid disruption of the tracheal supply to the musculature, and a fine jet of air was blown at the spiracle. Standard saline (Weevers, 1966) was oxygenated and flowed at a rate of l-3mlmin~' through the body cavity of the caterpillar. The posterior end of the segment was pinned through the wall into a fixed Sylgard block and the anterior end of the muscle being investigated was attached by an array of fine hooks to a home-built transducer consisting of two Pixie Pharmacology of lepidopteran neurones 335 transducer elements (Endevco, Royston, Herts). Signals from this transducer were passed through a four-pole filter, differentiated by an operational amplifier differentiator (Buchan & Evans, 1980) and recorded on a Racal Store 4 DC tape recorder. The motoneurone innervating the muscle under study was stimulated by a fine pair of silver/silver chloride wires (insulated apart from their tips) that were laid against a single muscle fibre. Careful adjustment of electrode position, polarity and current allowed single motor units to be selectively activated. Muscle activity was monitored intracellularly by a low-resistance floating microelectrode (Woodbury & Brady, 1956). Motoneurone activity was also monitored by recording antidromic impulses in the root of nerve 1 with a silver hook electrode that could be drawn into a paraffin-filled glass sleeve. Median cell activity was controlled either by stimulating the neurone with intracellular depolarizing pulses or by extracellular stimulation of an axon sidebranch at a site remote from the muscle being studied. In the latter case, impulses of the median cell were recorded en passant between the site of stimulation and the muscle under study, using the paraffin hook electrode. This method could be used with confidence since the median cell extracellular impulse had a unique time course and its appearance always correlated with an antidromic impulse recorded when intracellular recording was also used. This preparation had the advantage that the segmental ganglion could be removed thus preventing all spontaneous muscular activity. Where drug applications to the muscle were intended, the entire preparation was tilted at an angle of 10° to the horizontal, to reduce the dead space of saline above the muscle. This meant that the volume of saline surrounding the muscle at any time could not be directly measured, but the changeover time (assessed visually with dye solutions) was of the order of a few seconds only. Solutions of drugs were made up on the day in normal saline. DL-octopamine-HCl, /S-phenylethylamine-HCl, DL-phenylethanolamine, dopamine-HCl, noradrenaline-HCl, adrenaline bitartrate, tyramine-HCl, 5-hydroxytryptamine creatinine sulphate, DL-synephrine-HCl, histamine dihydrochloride, naphazoline-HCl, tolazoline-HCl, clonidine-HCl and DL-propranolol hydrochloride were all obtained from Sigma Chemical Co. Phentolamine mesylate was obtained from Ciba and chlordimeform was a gift from ICI. Ninety-two preparations were used in the present study; in a few cases recordings were made from several segments in one animal.

Results Median cells normally showed a slow rate of spontaneous firing (from 005-0-5impulsess-1) during extended periods of recording from Antheraea pernyi larvae. Occasionally, however, the cell would fire bursts of impulses at much higher frequency, although rarely exceeding 2 spikes s~'. Observation of the animal at such times always revealed strong spontaneous muscular activity. When gross muscular activity was monitored by attaching a transducer to the tail, the 336 S. J. H. BROOKES

50 mV MCI -LU4

Tension -'

B ]50mV MCI-I— Tension

]50mV MC1- Tension - 10s

5mV MCI

Tension

Fig. 1. Recordings were made from MCI in ganglion 5 of a fifth-instar caterpillar. An uncalibrated transducer attached to the tail hook gave a reflection of the gross longitudinal muscle tone of the animal. It registered an upward deflection when the animal contracted. The nerve cord was intact in this preparation. (A) The rate of firing of the median cell increased with spontaneous jerks and during prolonged muscular activity of the animal. (B) A slow oscillation of tension recorded in an animal that did not appear to involve the phasic longitudinal muscles, was probably due to variations of firing rates of tonically active posture-maintaining muscles. Median cell firing did not correlate in any way with this slow oscillation. Note the slow basal rate of firing. (C) Touching the tail of the caterpillar (arrowhead) evoked a brief twitch during which the median cell fired several impulses at a relatively high frequency. (D) Touching the animal's head evoked two twitches both of which correlated with brief bursts of summating EPSPs. The median cell was hyperpolarized to prevent impulses from obscuring the synaptic activity. spiking rate of median cells was seen to increase during active movements of the animal, a single impulse coinciding with a single spontaneous jerk and a burst of impulses with a prolonged series of contractions (Fig. 1). During long periods of Pharmacology of lepidopteran neurones 337 apparent inactivity a regular oscillation of muscle tone was often seen. This did not correlate with visible contractions in the fast longitudinal muscles but probably reflected variations of tension in the many tonically active slow motor units. No correlation of median cell impulses was seen with these slow oscillations (Fig. IB). Muscular activity could be evoked in the preparation by a variety of sensory stimuli including blowing on, or touching, the head capsule, body wall, prolegs or tail. In each case the median cell initiated impulses at an increased frequency whether such evoked muscular activity was of short or long duration (Fig. 1). By hyperpolarizing the median cell it was possible to block impulse initiation and see the increase in frequency of excitatory synaptic potentials which was presumably responsible for the increased rate of firing during movement. To determine whether the increased firing rate was a cause or a result of muscular activity, relatively high rates of firing were induced by intracellular current pulses during periods of inactivity. On no occasion was stimulation of single median cells seen to evoke any contraction in the preparation. From these observations it was apparent that to understand the role of median cell impulses it would be necessary to record their effects on controlled contractions in single motor units.

The muscle A preparation One of the most consistent axon branches of the median cells was that of MC2 that entered nerve lj and innervated the dorsal longitudinal muscle A (see Brookes & Weevers, 1988). The axon could be reliably traced to enter and branch in this muscle block, but not to leave it. In spite of this, MC2 impulses had no one- for-one effects detectable on either the resting potential or the resting tension of the muscle. Single MC2 impulses also had no detectable effect on the size of twitches and EJPs evoked by motoneurone excitation whether the median cell impulse preceded or coincided with the twitch, and bursts of MC2 impulses had no immediate effects detectable on the muscle resting potential, suggesting that MC2 did not act as a classical inhibitory motoneurone (Usherwood & Grundfest, 1965). All effects of bursts of MC2 impulses took several seconds to appear. Bursts of impulses fired in the median cell by depolarizing current pulses (30 impulses at lHz) had long-term effects on the contractions evoked by motoneurone stimulation (Fig. 2). The resting tension of the muscle declined 15-60 s after the start of MC2 activity and gradually recovered over the next 3-8 min. A second effect was a variable increase in twitch tension of up to 15%, that also recovered over the next 10 min. Reducing the base tension of the muscle over a wide range (by moving the transducer) had little effect on twitch tension, indicating that the effects on base tension did not directly cause the increase in twitch size. In no cases were significant effects detected on the amplitude of EJPs or the rate of relaxation following twitch contractions in fast muscle A. It should be borne in mind that changes in the resting tension of other motor units that lay parallel to muscle A may have contributed to the apparent decline in tension 338 S. J. H. BROOKES A Tension

EJPsil ]50mV MC2- ]20mV 1 10 s

15 Hz 5 Hz

10 Hz 20 Hz

Fig. 2. (A) The effects of 30 MC2 impulses on twitches and EJPs of the fast muscle A. The motoneurone innervating this muscle was stimulated at approximately 1 Hz. About 30 s after MC2 activity (recorded here as axon spikes in the soma) the base tension declined and twitches increased in size. No effect was ever seen on either the resting potential or the amplitude of EJPs in muscle A fibres. (B) The effects of 30 MC2 impulses on tetanic contractions in muscle A. The motoneurone to the muscle was stimulated with gated pulse trains in a set series of frequencies; after 30 MC2 impulses contractions were enhanced in amplitude. In this particular case the effect was too small to be measured in twitches and at 5 Hz stimulation but was clear at 10,15 and 20Hz. In addition, at 20Hz stimulation, the MC2 impulses caused a small increase in the rate of relaxation (visible as the crossing-over of the 'before' and 'after' traces). This effect was not always seen even in high-frequency contractions of this muscle - in contrast with slow muscle QR, where a pronounced increase in relaxation rate was always caused by MC2 activity. Contractions after MC2 stimulation are marked with a filled circle.

following MC2 stimulation (especially in view of the widespread branching of MC2 in the dorsal musculature, Brookes & Weevers, 1988). When the motoneurone innervating muscle A was stimulated with gated pulse trains of different frequencies in a set sequence, contractions at all frequencies showed a small increase in peak amplitude following 30MC2 impulses (Fig. 2B, the increase is not clear on twitches or at 5Hz). A detectable effect on the Pharmacology of lepidopteran neurones 339 relaxation rate following the largest contractions was seen in a few preparations (shown by the crossing-over of the traces in Fig. 2B, 20Hz). This suggested that effects of MC2 activity on the relaxation rate of this muscle were rather limited in comparison with slower muscles (see Fig. 3C).

Muscle QR preparation Muscle A, the dorsal longitudinal motor unit, was one of the faster motor units in the caterpillar abdomen; it gave large, discrete twitches in response to single electrical stimuli. In contrast, muscle QR was a much slower muscle with very small-amplitude twitches that were slow to relax back to the resting level of tension; it too had a branch of the axon of MC2 consistently associated with it. As with the fast muscle A, no one-for-one short-latency effects of MC2 impulses could be detected on the resting potential, base tension, EJP size or twitch contractions of this motor unit, indicating once again that this neurone was not simply an inhibitory motoneurone. When the muscle was stimulated to contract repeatedly (varying frequencies in a set sequence of gated stimulus pulse trains every 15 s) median cell 2 activity (30 impulses evoked at 1 Hz) had five detectable effects (Fig. 3A): (1) a decrease in base tension similar to that seen in muscle A; (2) a small hyperpolarization of the muscle membrane (0-5 mV); (3) an increase in the size of EJPs; (4) an increase in peak tension of higher-frequency tetani and twitches; and (5) an increase in the rate of relaxation following contractions. The effects of MC2 activity on the peak tension of muscle QR depended on the frequency of stimulation of the motoneurone (Fig. 3C). With low-frequency tetani, peak tension was actually reduced after MC2 activity, whereas with higher- frequency tetani (2^10Hz) peak tension was consistently enhanced. This was presumably due to the interaction of the opposing effects of enhancement of both contraction size and relaxation rate. This is clearly seen in the comparison of twitches before and after MC2 activity (Fig. 3C). The relaxation rate following the cessation of stimulation was strongly enhanced at all frequencies, in contrast to the situation in muscle A (Fig. 3C). Excitatory junction potentials of muscle QR, which characteristically showed marked facilitation at the onset of repeated stimuli, were enhanced in amplitude at all levels of facilitation following MC2 impulses (Fig. 3B). In contrast, this effect on the amplitude of the electrical response to stimulation of the motoneurone to muscle A was never seen. It is possible that an effect was present in the fast muscle but was masked by its large active membrane response, a feature not shared by the slow motor unit QR. The increase in EJP amplitude in muscle QR was not dependent on the membrane hyperpolarization as it often occurred in fibres which were not hyperpolarized by MC2 activity, and the increase in EJP amplitude often exceeded the amplitude of the hyperpolarization. The increase in EJP amplitude had the fastest time course of all the MC2 effects, reaching its peak value within 60s of a burst of MC2 impulses. 340 S. J. H. BROOKES

The effects on base tension were rather variable in amplitude among different preparations and among repeated trials in the same preparation. Due to the widespread distribution of axon branches of MC2 in the dorsal musculature it is likely that many parallel motor units contributed to the overall decline in base tension as with fast muscle A. The small hyperpolarization was seen in about 50 % of preparations of muscle QR. In spite of the susceptibility of the floating microelectrode technique to movement artefact, this is unlikely to be the explanation for the effect, since a hyperpolarization was never seen in the fibres of muscle A despite the presence of a reduction in base tension. Also, the effect was always hyperpolarizing, whereas occasional depolarizing shifts would be expected if movement were responsible. It is possible that where no hyperpolarization was apparent in a muscle fibre following a burst of MC2 impulses, it may have been present in other (perhaps less damaged) fibres.

Tension • • till i i 111 11 \ 11. i I 111 11111 • 11111 *. 20mV EJPslilllHilllllilDllllllllllllllllllHll MC2- 60s

|20mV

EJPs J' ••'-'' i- ^ 10Qms

l\^ UOOmg Twitch v~*J J^VPC^ J

200 mg

25Hz' ' xv |500mg

5Hz—— ' lx |500mg Pharmacology of lepidopteran neurones 341

Median cell effects on other muscle groups Three other muscle groups in the dorsal musculature were studied for responses to MC2 activity. Two of these, groups G and I, were slow muscles similar to muscle QR; the other, muscle C, was a fast muscle similar to A. Both G and I responded similarly to QR, showing an increase in EJP amplitude, an increase in peak tension and a marked increase in relaxation rate following contractions. Muscle C showed a similar response to fast muscle A; twitches and tetani increased in amplitude without greatly increasing the relaxation rate. Interestingly, fast muscle C did show a significant increase in EJP size following MC2 activity, in contrast to the lack of this effect in muscle A. Effects similar to those of MC2 on the electrical activity of somatic muscles were also seen in muscles innervated by MCI in the ventral muscle block, suggesting that MCI has a comparable role to MC2. These effects were not studied systematically owing to technical difficulties in interrupting spontaneous motor input to the muscles without also interrupting the innervation by MCI. However, 30 MCI impulses were seen to cause both a hyperpolarization and an increase in the amplitude of spontaneous EJPs in slow muscle 8 that was directly comparable to that caused by MC2 in muscle QR.

Pharmacological mimicry of median cell effects Fourteen potential agonist substances were tested on the QR muscle preparation for their ability to mimic the effects of MC2 stimulation. Five of these, tyramine, phenylethylamine, noradrenaline, dopamine and histamine, were found to be ineffective when applied for 2min or longer at 10~6moll~1. Three amines, adrenaline, 5-hydroxytryptamine and phenylethanolamine, were weak agonists at 10~6mol 1~J and caused reversible changes in the electrical and contractile activity of muscle QR similar to those of MC2 stimulation. Two substances, octopamine Fig. 3. (A) The motoneurone innervating slow muscle QR was stimulated with gated pulse trains in a set sequence of 0-1 (twitch), 2-5, 5, 10, 15 and 20Hz. At the point marked by the black rectangle on the lowest trace, 30 MC2 impulses were fired at 1 Hz by intracellular depolarizing current pulses. The middle trace shows the EJPs recorded from a fibre of this motor unit with an intracellular microelectrode. After the MC2 activity there is a small relaxation of the muscle, some of the contractions become larger, there is a small hyperpolarization of the muscle resting potential and the EJPs increase in amplitude. (B) EJPs recorded from muscle QR were evoked by a gated stimulus pulse train. There is considerable facilitation of EJPs with repeated stimuli in this muscle. A second series of EJPs, recorded 1 min after the 30 MC2 impulses is superimposed (with a slight shift to the right to help visual comparison). At all levels of facilitation EJPs were enhanced in amplitude after MC2 impulses. (C) The effect of 30 MC2 impulses on contractions elicited by stimulation of the motoneurone of muscle QR at different frequencies. Contractions after MC2 activity are marked with a filled circle. A pronounced increase in the rate of relaxation after each contraction is seen. The enhancement of peak contraction amplitude is frequency-dependent - at 2-5 and 5 Hz the relaxation effect predominates and peak contraction amplitude is reduced. At 10, 15 and 20 Hz the effect on contraction amplitude predominates and a stronger contraction follows MC2 impulses. S. J. H. BROOKES

200 mg 200 mg

lgs" lgs"1

20 mV 20 mV

Fig. 4. The effects of 30 s pulses of octopamine (A) and synephrine (B), both at 10~8moir', and on the same preparation. Note the larger effect of synephrine. The figures were constructed by superimposing three records of contractions (elicited by stimulating at 20 Hz for 0-5 s) before and after drug application. The lowest set of traces in both cases shows EJPs evoked in the same muscle; traces after drug application are shifted slightly to the right to enable comparison of the effect of the agonist at different levels of facilitation. and synephrine, were found to be very potent agonists and had thresholds below 10~9moll~1. Synephrine was more potent than octopamine: a 30s pulse of synephrine had a larger peak effect than a similar dose of octopamine administered to the same preparation (Fig. 4). However, since synephrine is known to have a greater potency than octopamine in a number of octopamine-sensitive preparations, but has not been found in significant quantities in insect tissue (Orchard, 1982), attention was focused on the effects of octopamine. Three agonists of mammalian adrenergic receptors; naphazoline, tolazoline and clonidine have been shown to be effective agonists of octopamine receptors in the locust extensor tibialis preparation (Evans, 1981). They were all also effective at mimicking the effects of MC2 stimulation. The insecticide chlordimeform was also capable of mimicking the effects of both MC2 stimulation and small doses of octopamine when applied at lO^moll"1. This again reinforces the similarity of the receptors on caterpillar muscle to octopamine receptors in a number of insect preparations. Although dose-response curves for all the different agonists and effects were not constructed, it was apparent that at the concentrations tested either all or none of the effects of MC2 stimulation were mimicked.

Modulatory effects of octopamine on caterpillar muscle Very low doses of octopamine (10~8 mol I"1 for 30 s) caused repeatable effects of long duration on muscle QR. These effects were identical to those seen after a Pharmacology of lepidopteran neurones 343

50 mg Twitch

200 mg

10 Hz

Fig. 5. The effect of a 60s pulse of 10 8moll 1 octopamine on muscle QR. The motoneurone innervating the muscle was stimulated at different frequencies for 3 s in a set sequence. The tension traces before and after octopamine application were superimposed. Note that at low frequencies octopamine caused a reduction in peak tension (below 5 Hz). At higher frequencies peak tension was consistently enhanced (above 10Hz). As with the effects of MC2 stimulation, these differences are probably due to the interaction of opposing effects on peak tension and relaxation rate between twitches. burst of MC2 impulses and included a large consistent increase in the peak rate of relaxation, an increase of peak tetanic tension, an increase in EJP size, a small hyperpolarization and a small reversible decrease in base tension. All of these effects far outlasted the period of application; effects were detectable up to lOmin after such a dose. Similar to the effects of MC2 stimulation, the effects of small doses of octopamine on tension production in muscle QR were dependent on the frequency of motoneurone stimulation (Fig. 5). Again, this probably reflected the interaction of the opposing effects on peak tension and relaxation rate. In view of the different effects caused by MC2 stimulation on slow muscle QR and fast muscle A, it was of interest to see whether these differences were also mimicked by the application of octopamine. In muscle group A, the threshold for the appearance of effects of octopamine was below lO^moll"1, suggesting a slightly lower sensitivity than in muscle QR. Relatively low doses of octopamine (10~7moir' for 30s) mimicked the effects of MC2 stimulation, causing an increase in twitch tension and a decline in base tension with a smaller effect on relaxation rate and no detectable effects on membrane potential or EJP size (see 344 S. J. H. BROOKES

Fig. 6). However, at higher concentrations (e.g. 10 5moll l) octopamine clearly affected both the relaxation rate and the amplitude of the EJP. Thus it can be concluded that octopamine at low doses mimicked the effects of MC2 stimulation on both slow muscle QR and fast muscle A.

The effects of adrenergic antagonists Bursts of 10 MC2 impulses at regular intervals evoked consistent responses in muscle QR. When 10~5moll~1 phentolamine was superfused onto such a preparation, the effects of MC2 activity were reversibly blocked. Within lOmin of washing off, the effects of further bursts of MC2 impulses returned to their original level (see Fig. 7). Phentolamine is known to be an effective antagonist in a number of octopamine-sensitive preparations. However, it was necessary to show that this was also true for muscle QR. Sixty-second pulses of 10~8moir1 octopamine

Tension

50 mg 100 ms

50 mg Tension 60s 1 differential Jigs" Tension • 10 10"7moir' Control octopamine D 50 mg 20 mV

lgs- EJPs

20 mV

Fig. 6. The effect of octopamine on fast muscle A. In view of the different effects on MC2 activity on slow muscle QR and fast muscle A it was interesting that octopamine could mimic these effects on both muscles. (A) The muscle was made to twitch at approximately 1 Hz. A 30 s pulse of 1CT7 mol 1"' octopamine caused a reduction of base tension and an increase in twitch strength of muscle A with little effect on the relaxation rate. (B) Intracellularly recorded EJPs did not change in size with 10~7moir' octopamine. Three EJPs before and after octopamine are shown superimposed, the EJPs after octopamine were shifted slightly to the right to allow visual comparison of amplitude. These effects were similar to those seen when MC2 was stimulated. (C) Higher doses of octopamine had a variety of effects on fast muscle A. At 10~7moll~' relatively little effect on relaxation rate was seen, however at 10~5mol I"1 a clear increase in relaxation rate and EJP size was visible. (D) EJPs are shown before and after 10~5 mol I"1 octopamine application; those recorded after show a slight increase in peak amplitude. Pharmacology of lepidopteran neurones 345

50 mg

200 mgs"

60s

10 mV

EJPs 100 ms

50 mg

Tension differential 200 mgs"

10 mV

EJPs 100 ms

10 mol 1 ' phentolamine

Fig. 7. The effects of phentolamine on octopamine and MC2 modulation of muscle QR contractions. The motoneurone innervating the muscle was stimulated at 20 Hz for 0-5 s every 15 s. (A) Repeated bursts of 10 MC2 impulses given at 5 min intervals evoked consistent effects on the peak tension and relaxation rate of muscle QR. The lefthand series of traces shows this effect which was blocked (in the middle set of traces) after 4 min of supervision with lO^moll"1 phentolamine. The effects recovered after phentolamine had been washed out for 10 min (righthand traces). The EJP traces were taken from the same recordings - three traces were superimposed from before and after MC2 activity, the second set of traces being shifted slightly to the right to allow visual comparison. The point at which MC2 was fired is indicated by the arrowheads. (B) 60s pulses of 10~8molP' octopamine administered every 10min evoked consistent effects of the relaxation rate, peak tension and EJP size of muscle QR (lefthand traces). These effects were abolished after 6 min perfusion with 10~5moll~1 phentolamine (middle traces). The effects recovered (righthand traces) after 15 min of washing. The period of octopamine application is indicated by the solid bar. 346 S. J. H. BROOKES delivered at lOmin intervals evoked consistent MC2-like effects. Addition of 10~5moll"1 phentolamine completely blocked the effects of subsequent doses of octopamine. After 15min of washing with control saline the preparation fully recovered its sensitivity to octopamine (Fig. 7). Propranolol, a /3-adrenergic antagonist, was completely ineffective as an antagonist of the effects of both MC2 stimulation and octopamine application when administered in the same doses as phentolamine. Neither propranolol nor phentolamine had any direct effects on the electrophysiological or contractile activity of muscle QR.

Discussion MC2 has been demonstrated to have a number of modulatory effects on several skeletal muscles of the dorsal abdomen. From the effects of a burst of MCI impulses on spontaneous EJPs in two ventral slow muscles, it is highly likely that MCI plays a similar role to that of MC2. The effects were due to the direct action of the median cells since they persisted when the ganglion was removed and MC2 was stimulated extracellularly, thus removing the possibility of the recruitment of other neurones. Extreme variability in both the normal amplitude of contractions recorded in the muscles and the size of the change effected by MC2 stimulation or octopamine application made quantification of results difficult. The increase in amplitude of EJPs in slow muscles was a highly consistent effect of MC2 stimulation and was similar to the effects in the locust of octopamine (the transmitter of the modulatory neurone DUMETi) on the extensor tibialis muscle (O'Shea & Evans, 1979). In that preparation, the effect is probably presynaptic in action since the frequency, but not the size, of miniature endplate potentials is increased. In several other nerve-muscle preparations (lobster, Glusman & Kravitz, 1982; crab, Lingle, 1981; rat, Kuba, 1970) modulatory effects of amines (5-hydroxytryptamine, dopamine and adrenaline, respectively) have been shown to correlate with an increase of muscle membrane resistance. In Manduca sexta it has been shown that octopamine has both presynaptic effects on the frequency of miniature EJPs in the dorsal longitudinal flight muscle of the adult moth and a postsynaptic hyperpolarizing effect similar to that reported here (Klaassen, Kammer & Fitch, 1986). The lack of an effect on the EJPs in the fast muscle A is interesting since an enhancement of twitch amplitude was present. It is possible that an effect on the EJP was present but was masked by the large active membrane response seen in fibres of this muscle. However, since no broadening of the EJP was seen, it suggests that the enhancement of contractions in somatic muscles was not caused by presynaptic effects alone. The relative potency of the various effects of MC2 stimulation (and octopamine application) differed between fast muscles (A and C) and slow muscles QR, G and I. Thus large effects on both contraction amplitude and relaxation rate were seen in muscle QR whereas a relatively smaller effect was seen on the relaxation rate of the faster muscles. In the locust extensor tibialis (which receives both fast and slow innervation), octopamine (and DUMETi stimulation) increases both the peak Pharmacology of lepidopteran neurones 347 contraction and relaxation rate in slow twitches (similar to the effects of MC2 on muscle QR) but only affects the relaxation rate of fast twitches (O'Shea & Evans, 1979). Evans (1985) reported different sensitivities to octopamine of muscle fibres in various regions of this muscle that correlated to an extent with the innervation by SETi and FETi. The frequency-dependence of the effects of octopamine on muscle QR is directly comparable with that reported by Evans & Siegler (1982) in the locust for SETi-evoked contractions. In the present study such frequency- dependence was not seen in the fast muscle A where there was little effect on the relaxation rate; octopamine or MC2 stimulation enhanced the peak amplitude of all contractions irrespective of the frequency of stimulation of the motoneurone.

The receptors underlying median cell effects The results presented are consistent with the hypothesis that the effects of MC2 stimulation on the dorsal muscles are mediated by the release of an aminergic neuroeffector. In view of the similarity of the pharmacology of these effects to other preparations of octopamine receptors (see David & Coulon, 1985, for a review), the most likely candidate for the transmitter substance of the median cell system is octopamine. Confirmation of this hypothesis will require further studies. Synephrine was the only amine that was more potent than octopamine in stimulating MC2-like effects. However, this is unlikely to be the transmitter released by the median cells since it has not been found in significant quantities in insect tissue (Orchard, 1982) whereas octopamine is known to be synthesized and stored in various parts of the central nervous system of Lepidoptera (Maxwell et al. 1978; Davenport & Wright, 1986). Synephrine has been found to be a more potent agonist than octopamine in a number of other octopamine-sensitive preparations (Evans, 1981; Carlson, 1968; Batelle & Kravitz, 1978) although this is not the case in broken cell preparations (Evans, 1981). It is interesting that in the dorsal longitudinal flight muscle of adult Manduca sexta synephrine is less potent than octopamine in enhancing EJP amplitude and maximum following frequency (Klaassen & Kammer, 1985). In a preparation of the heart of larval Manduca sexta, octopamine and synephrine have been reported to be equipotent (Platt & Reynolds, 1986). These observations suggest developmental or interspecific differences, or that there are different subtypes of octopamine receptors in Lepidoptera, as has been demonstrated in the locust extensor tibialis preparation (Evans, 1981). The mimicry of octopamine effects by the insecticide chlordimeform (CDM) has also been reported in the locust preparation (Evans & Gee, 1980), although the putative breakdown product, demethylchlordimeform (DCDM), was an order of magnitude more potent in the extensor tibialis. In some other insect preparations CDM has been reported as being inactive whereas DCDM is highly potent (in the heart of Manduca sexta, Platt & Reynolds, 1986; in the firefly lantern, Nathanson & Hunnicutt, 1981). The effect of CDM is interesting in view of the susceptibility of some lepidopteran larvae to formamidine insecticides (Davenport & Wright, 1985). 348 . S. J. H. BROOKES

Evidence from several sources suggests that octopamine is an important neuroeffector substance in Lepidoptera, and the possibility that the median cells (which are found in at least five of the abdominal ganglia of A. pernyi) comprise an identifiable part of an octopaminergic system is fully compatible with this evidence. Circulating levels of octopamine (up to lO^moll"1, Davenport & Wright, 1986) in larval M. sexta are within the range shown here to be effective on both fast and slow muscles in A. pernyi. More direct effects, by local release close to the muscle (as in the MC2 stimulation experiments), may also be important. In Manduca sexta, neurosecretory-type axons similar to those of DUMETi in the locust extensor tibialis (Hoyle et al. 197r4) have been traced from the flight motor nerve to the dorsal longitudinal muscle to both peripheral cell bodies and to unpaired median neurones that are probably the thoracic homologues of MCI and MC2 (Wasserman, 1985).

The role of the median cells Strong muscular activity (when the median cells fired most impulses) correlated with the recruitment of phasically active fast motor units such as muscles A and C (see Brookes & Weevers, 1988). This is rather surprising at first sight, since the effects of median cell firing were more pronounced on contraction and relaxation in the tonically active slow muscles such as muscles QR, G and I. This apparent paradox would be explicable if the function of the median cells were to adapt all of the body wall musculature for the rapid and powerful movements required for walking or struggling. The effects of MC2 stimulation on reducing the base tension and increasing the rate of relaxation of slow muscles would reduce their resistance to the contraction of fast muscles in other parts of the body (assuming that the haemocoel and gut act together as a sort of hydrostatic skeleton). A similar role has been proposed for the action of DUMETi in the locust (Evans & Siegler, 1982). A temporary increase in speed of both contraction and relaxation of slow muscles for the duration of a bout of movement could increase the efficiency of muscular activity without the necessity for dual innervation of all of the muscles involved.

References BATELLE, B. A. & KRAVITZ, E. A. (1978). Targets of octopamine action in the lobster; cyclic nucleotide changes and physiological effects in hemolymph, heart and skeletal muscle. /. Pharmac. exp. Ther. 205, 438-448. BROOKES, S. J. H. & WEEVERS, R. DE G. (1988). Unpaired median neurones in a lepidopteran larva (Antheraea pernyi). I. Anatomy and physiology. /. exp. Biol. 136, 311-332. BUCHAN, P. B. & EVANS, P. D. (1980). Use of an operational amplifier signal differentiator reveals that octopamine increases the rate of development of neurally evoked tension in insect muscle. J. exp. Biol. 85, 349-352. CARLSON, A. D. (1968). Effects of drugs on luminescence of larval fireflies. /. exp. Biol. 49, 195-199. DAVENPORT, A. P. & WRIGHT, D. J. (1985). Toxicity of chlordimeform and amitraz to the Egyptian cotton leafworm, Spodoptera littoralis and the tobacco budworm Heliothis virescens. Pestic. Sci. 16, 81-87. Pharmacology of lepidopteran neurones 349

DAVENPORT, A. P. & WRIGHT, D. J. (1986). Octopamine distribution in the larvae and adults of two species of moth, Spodoptera littoralis and Manduca sexta. J. Insect Physiol. 32, 987-993. DAVID, J. C. & COULON, J. F. (1985). Octopamine in invertebrates and vertebrates. A review. Prog. Neurobiol. 24, 141-185. EVANS, P. D. (1981). Multiple receptor types for octopamine in the locust. /. Physiol., Lond. 318, 99-122. EVANS, P. D. (1985). Regional differences in responsiveness to octopamine within a locust skeletal muscle. J. Physiol., Lond. 366, 331-341. EVANS, P. D. & GEE, J. D. (1980). Action of formamidine pesticides on octopamine receptors. Nature, Lond. 287, 60-62. EVANS, P. D. & O'SHEA, M. (1977). The identification of an octopamine neurone which modulates neuromuscular transmission in the locust. Nature, Lond. 270, 257-259. EVANS, P. D. & O'SHEA, M. (1978). The identification of an octopaminergic neurone and the modulation of a myogenic rhythm in the locust. J. exp. Biol. 73, 235-260. EVANS, P. D. & SIEGLER, M. V. S. (1982). Octopamine mediated relaxation of maintained and catch tension in locust skeletal muscle. J. Physiol., Lond. 324, 93-112. GLUSMAN, S. & KRAVTTZ, E. A. (1982). The action of serotonin on excitatory nerve terminals in lobster nerve-muscle preparation. J. Physiol., Lond. 325, 223-241. HOYLE, G. (1975). Evidence that insect dorsal unpaired median (DUM) neurones are octopaminergic. /. exp. Zool. 193, 425-431. HOYLE, G., DAGAN, D., MOBERLEY, B. & COLQUHOUN, W. (1974). Dorsal unpaired median insect neurons make neurosecretory endings on skeletal muscle. J. exp. Zool. 187, 159-165. KLAASSEN, L. W. & KAMMER, A. E. (1985). Octopamine enhances neuromuscular transmission in developing adult moths, Manduca sexta. J. Neurobiol. 16, 227-243. KLAASSEN, L. W., KAMMER, A. E. & FITCH, G. K. (1986). Effects of octopamine on miniature excitatory junction potentials from developing and adult moth muscle. J. Neurobiol. 17, 291-302. KUBA, K. (1970). Effects of catecholamines on the neuromuscular junction in the rat diaphragm. /. Physiol., Lond. 211, 551-570. LINGLE, C. (1981). The modulatory action of dopamine on crustacean foregut neuromuscular preparations. /. exp. Biol. 94, 285-299. MASON, A. & KRISTAN, W. B. (1982). Neuronal excitation, inhibition and modulation of leech longitudinal muscle. /. comp. Physiol. 146, 527-536. MAXWELL, G. D., TAIT, J. F. & HILDEBRAND, J. G. (1978). Regional synthesis of neurotransmitter candidates in the CNS of the moth Manduca sexta. Comp. Biochem. Physiol. 61C, 109-119. MORTON, D. & EVANS, P. D. (1984). Octopaminergic release from an identified neurone in the locust. J. exp. Biol. 113, 269-287. NATHANSON, J. A. & HUNNICUTT, E. S. (1981). /V-Demethylchlordimeform. A potent partial agonist of octopamine sensitive adenylate cyclase. Molec. Pharmac. 20, 68-75. ORCHARD, I. (1982). Octopamine in insects; neurotransmitter, neurohormone and neuromodulator. Can. J. Zool. 60, 659-669. O'SHEA, M. & EVANS, P. D. (1979). Potentiation of neuromuscular transmission by an octopaminergic neurone in the locust. J. exp. Biol. 79, 169-190. PLATT, N. & REYNOLDS, S. E. (1986). The pharmacology of the heart of a caterpillar, the tobacco hornworm Manduca sexta. J. Insect. Physiol. 32, 221-230. USHERWOOD, P. N. R. & GRUNDFEST, H. (1965). Peripheral inhibition in skeletal muscle of insects. J. Neurophysiol. 28, 497-518. WASSERMAN, A. J. (1985). Central and peripheral neurosecTetory pathways to an insect flight motor nerve. J. Neurobiol. 16, 329-345. WEEVERS, R. DE G. (1966). A lepidopteran saline; the effects of inorganic cation concentrations on sensory reflex and motor responses in a herbivorous insect. J. exp. Biol. 44, 163-176. WEISS, K. R., COHEN, J. & KUPFERMANN, I. (1975). Potentiation of muscle contraction; a possible modulatory function of an identified serotinergic cell in Aplysia. Brain Res. 99, 381-386. 350 S. J. H. BROOKES

WEISS, K. R., COHEN, J. L. & KUPFERMANN, I. (1978). Modulatory control of buccal musculature by a serotinergic neuron (Metacerebral Cell) in Aplysia. J. Neurophysiol. 41, 181-203. WOODBURY, J. W. & BRADY, A. J. (1956). Intracellular recordings from moving tissues with a flexibly mounted microelectrode. Science 123, 100-101.