Intrinsic Rhythm and Basic Tonus in Insect Skeletal Muscle

The large, anterior, dorsal, unpaired median (DUM) neurones of orthopterous insects (Plotnikova, 1969) were first studied physiologically by Kerkut, Pitman & Walker (1969), who found their cell bodies to have overshooting action potentials. Crossman et al. (1971) filled some of them with dye in the cockroach and like Plotnikova (1969), found tentative evidence for their sending an axon into a leg nerve. This possibility was fully substantiated by Hoyle et al. (1974) for the locust Schistocerca gregaria DUM neurone, DUMETi, whose axons were followed into its final terminals and found to make neurosecretory-type endings in the jumping muscle.

In testing the possible action of DUMETi on muscle properties and neuromuscular transmission I was unable to find any effect of the axon on transmission, but the extensor tibiae would sometimes very slowly extend spontaneously in a rhythmical manner. Such movements had been noticed in Dr P. N. R. Usherwood’s laboratory, who had demonstrated them to the Scottish Electrophysiological Society. In a note to the Physiological Society, Anderson et al. (1970) state: ‘tonic fibres… undergo spontaneous rhythmic contractions in vivo and in vitro. These contractions are possibly myogenic and could conceivably be synchronized by electrical coupling between the fibres.’

The movement is of variable amplitude and time-course, both in different preparations or different legs of the same animal and for the same leg at different times, although it can be constant over several hours. When at its most pronounced, the tibia extends through an arc of 30°. How was it possible to have missed if for so long? The answer is that it is actively suppressed when the animal is disturbed and handled, and the suppression is usually long-lasting. Also, it is seldom evident in Locusta migratoria preparations, on which the early neuromuscular research was carried out, unless the leg nerve has been cut for some hours. In all insects that show it, a few days of denervation, followed by excision of the leg, are sufficient to reveal that it is an intrinsic feature, at least of the extensor tibiae muscle. It will be referred to hereafter as the intrinsic rhythm (IR). An example, recorded as isometric force by a transducer attached to the extensor tibiae of an isolated Schistocerca gregaria femur, is shown in Fig. 2. Another reason for the late discovery of IR is that bathing the muscle in insect Ringer’s solutions tends to suppress it. It has now been found that a hidden, or Ringer-suppressed IR can be recovered by adding acetylcholine (10⁻⁸ M) and eserine to the saline (G. Hoyle & E. Florey, in preparation).

In the intact animal the IR movements are sufficiently strong to initiate resistance reflexes (Hoyle & O’Shea, 1974). During the extension phase the antagonistic flexor tibiae (FITi) muscle is excited and the slow extensor (SETi) is inhibited. During relaxation the flexion movement reflexly causes SETi to be excited. The net result is that there is very little actual movement in the intact animal, which is another reason why IR failed to be noticed. The animal is, in effect, continually performing a set of isometric exercises.

The locust IR has acquired a special interest since it was found that it is suppressed by DUMETi (Hoyle, 1974). The suppression is mimicked by some biogenic amines, the most effective of which is octopamine, which works at a concentration as low as 2 × 10⁻¹⁰ M (Hoyle, 1975).

These observations also bring into focus the neglected question of the generation of tonus and its control by central nervous systems. Before about 1930 there had been many theories about the basis of tonus in skeletal muscle. A favourite hypothesis had been that of Langelaan (1915, 1922), known as the dual theory of tonus, which claimed that there are two separate aspects to tonus: a basic, or ‘plastic’ element that is modulated by the sympathetic nervous system, and a ‘contractile’ element dependent upon ordinary motor innervation. The ideas were reviewed by Stanley Cobb in 1925 (Cobb, 1925), who was impressed only by the elegant, eventually to become classical, work of Liddell & Sherrington (1924, 1925) on the myotatic reflex. He concluded ‘that tonus is a beautifully graded series of proprioceptive reflexes.’ And that was that; even invertebrate myo-neural physiologists have never chosen to challenge this conclusion except in the case of the ‘catch’ mechanism of some molluscan obliquely striated muscles.

In arthropods, studies of the membrane potential-tension relationships of individual muscle fibres revealed some potential anomalies. Atwood, Hoyle & Smyth (1965) found some fibres in the crab claw closer muscle whose resting membrane potential was essentially at the excitation-contraction coupling threshold (Ec). In the levator of the eyestalk of the crab Podophthalmus, Hoyle (1968) found several muscle fibres with resting potentials below Ec. They are in a constant state of partial contraction that is their normal state, not a pathological one, which serves the useful function of keeping the eyestalks raised. This is so in the absence of excitatory nerve input. When only the inhibitory axon supplying the muscle was stimulated, the muscle relaxed. The slow excitatory axon acts reflexly to supplement this intrinsic tonus and move the relevant eyestalk, in conjunction with the inhibitor. This system is clearly a dual tonus control mechanism.

The present work, though principally a study of IR, supports the prematurely discarded notion of dual mechanisms of tonus in striated muscles. As in the crustaceans, not all muscle fibres of a muscle contribute: only specialized slow fibres are involved.

The species studied were the locusts Schistocerca gregaria Forskål = S. americana (Dirsch, 1974), Locusta migratoria and Schistocerca vaga =S. nitens nitens (Dirsch, 1974), and the grasshoppers Romalea microptera and Brachystola magna. The principal preparation used (Fig. 1) consisted of a dorsal dissection in which the wings were removed and the thorax was bisected longitudinally followed by removal of the heart and gut. The wing bases were clamped firmly and the body and legs, but not the abdomen, firmly bedded in soft dental wax. A hard-wax-coated platform was micro-manipulated under the metathoracic ganglion. The tibiae were cut off just below the knee joint, and the latter opened up. The extensor tibiae apodemes were cut close to their distal insertions and seized by fine forceps tips attached directly to micromanipulated RCA 5734 mechano-electronic transducer tubes, providing isometric force measurement. The apodeme position corresponded to a femero-tibial angle of approximately 40°. Some recordings were also obtained of tension, or movement of the tibia, from whole animals and preparations in which the whole leg was left intact. In these, force was registered by tying the tibia to the force transducer; movement was registered by a photocell, with a light flag attached to the tibia to interrupt a light beam.

The intrinsic rhythm is sometimes altered, or even abolished, by perfusing the leg with saline, so the present experiments were all carried out with the muscle bathed in its own haemolymph. A petroleum jelly cup was fashioned around the metathoracic ganglion so that it could be bathed independently in locust saline without the latter affecting the leg muscle.

Ganglion cells were excited either directly, by an intracellular electrode, or indirectly via small insulated wires or a suction electrode applied to the ganglion surface. The slow extensor tibiae axon (SETi) was excited via small hook electrodes placed under nerve N3b, the common inhibitor (CI) by similar electrodes placed under N3C, and the fast extensor (FETi) by electrodes touching N5.

IR was found in most preparations whilst the muscles were still bathed in haemolymph, and the patterns observed were substantially similar in all genera and species.

It is not possible to say, even from a long-term recording, which species a given record came from. Some preparations did not show IR within the first 2 h of setting up, but then it began to appear. In these, the muscle was in a state of strong tonic contraction after being set up, and the indication that a rhythm would appear was an abrupt relaxation having a time-course similar to that of an IR relaxation. It might then be minutes before a contraction followed, and likewise before the next relaxation but gradually the rhythm speeded up to a typical one. IR persisted, essentially unchanged, when recorded from an isolated leg or femur (Fig. 2).

Hoyle & Florey (in preparation) studying similar Locusta preparations found that the majority did not show either a strong tonus or IR. However, they found that IR could quickly be induced by adding acetylcholine (10⁻⁸g/ml threshold) and eserine to ordinary locust saline and perfusing the leg with it. Other tricks that revealed IR were isolation from other locusts for 2 weeks, and denervation.

Each IR contraction consists of three components: (1) a slow rising phase, when tension is increasing ; (2) a peak, which is sometimes sharp but usually extended into a plateau; and (3) relaxation, which is relatively fast compared with the rise time. A pause, or interval, follows each contraction. The three principal variables encountered were: the amplitude of the excursion, the total duration of the contraction and the interval between successive contractions.

The basic IR for a given muscle can be determined only in the isolated leg. Not only is IR affected by direct neural output from the ganglion, it is also affected by neurohumoral material released into the blood. Ideally, the isolated limb should be perfused by a simple standard saline solution. Unfortunately, all salines so far tested have reduced or stopped the rhythm completely in some specimens. An improvement in IR stability in saline occurred in Locusta, which is particularly sensitive to saline, by using lowered K+ (2 mm/1, Hoyle & Florey, in preparation), but no improvement was noticed in either Schistocerca or Romalea. In the isolated femur, the extensor tibiae showed a strong IR that persisted with little change in amplitude, duration and frequency for several hours. The tension developed was maximally about 0·2 g. For the example shown in Fig. 3, the mean duration of an IR contraction was 4·8 s, the mean frequency 5·6 per min. During a period of 3 h the same preparation showed a 4 s minimum and a 7 s maximum duration, and 4 per min minimum and 6 per min maximum frequencies. During the whole second hour, however, the variance in these parameters was below 15%.

In the intact animal, the amplitude of extensor tibiae IR contractions fluctuated, in all species, in an unpredictable manner. The longest single IR lasted for almost 4 h, and the shortest one, 1 s. The frequency varied from 1 per 4 h to 8 per min. The time to peak contraction varied from 0·8 to 40 s for the same muscle at different times. The corresponding times from peak contraction to 70% relaxation were 0·8−1·4s. Examples from different species are illustrated in Fig. 3, taken, in each case, about 1 h after setting up the preparation. However, the detailed patterns for a particular species are meaningless in view of the extensive variations at different times, as the records shown in Fig. 4 clearly demonstrate. These records were samples of a continuous record from one S. gregaria at various times during a 24 h period. During this time there were many changes in amplitude, frequency and duration of IR contractions, and also in the presence or absence and size when present, of relaxation undershoot.

It will be shown that frequency, amplitude and undershoot are under central nervous control, probably by neurohumoral activity. It is not known what factors determine the durations of IR contractions, which change slowly, in a manner which suggests neurohumoral action, over a remarkably wide range.

In addition to slow, smooth transitions, many occurred abruptly (Fig. 5). In light of experience, to be described below, of the effects of ordinary motor nerve activity on IR, it seems that abrupt changes in IR may be caused by effects of either SETi, CI or both axons.

It was anticipated that the source of IR would be an oscillatory depolarization, but since during earlier studies on locust muscles only stable membrane potentials were recorded in widespread sampling, it was likely that only a relatively small number of fibres are involved. There could have been multiple sites, or there might be a single region involved. Very long, flexible glass capillary microelectrodes filled with 2 M-K-acetate, first tested for their ability to bend without giving artifacts, were used. A few muscle fibres showing depolarization waves at the same time as IR (Fig. 6) were located in a large bundle of fibres located on the outside of the apodeme between 6 and 3 mm distant from the proximal border of the femur. Fibres examined elsewhere in the muscle showed no depolarization waves. The active fibres were in the same region found by Hoyle & Florey (in preparation) to be the one where drugs applied in small drops most readily affected IR. There is an obvious shortening of fibres in this region during IR in the isolated or denervated leg. This bundle was isolated from the whole of the distal part of the muscle, and also the more proximal fan-shaped region, and continued to develop IR tension equal to at least 80% of that developed by the whole leg. Hence, fibres in this region are largely, or entirely, responsible for IR. The depolarizations shown by these fibres generally matched rather closely the shape of IR contractions. They ranged from the barely detectable, to ones having 24 mV peak amplitude. It is presumed that the larger waves were recorded from pacemaker fibres and the smaller from one weakly electrically coupled to them. As IR amplitude and duration fluctuated naturally, so did those of the depolarization wave. That the wave was not an artifact was demonstrated when strong SETi contractions and FETi twitches were interposed in the records.

DUMETi and other DUM neurones were stimulated whilst recording intracellularly from an IR ‘pacemaker’ muscle fibre. The amplitude of the IR contractions was reduced following the neural stimulation and it was found that the depolarizing waves were correspondingly reduced (Fig. 6).

The onset of depolarization in presumed pacemaker fibres preceded the onset of tension development in the whole bundle by a few milliseconds, but occurred in some fibres showing a small wave after the onset of tension. The onset of relaxation was not well correlated with tension decline, and in some fibres the onset of relaxation occurred before repolarization. But this anomaly could be explained on the basis of differences between contraction times of individual fibres. It will be shown below that the rhythmicity can be markedly different in different regions of the same muscle.

When the slow axon discharged, as in Fig. 6A, following the second ganglion stimulation, and also Fig. 6C, the SETi contractions were accompanied only by a very small depolarization. FETi fired during the experiment shown in Fig. 6E, as shown by the twitch at the arrow, and there was no e.j.p. Neighbouring fibres generally had relatively large, facilitating SETi e.j.p.s and also large CI i.j.p.s, but some also responded to FETi with large e.j.p.s and overshooting spikes. There is no branch of DUMETi to the region.

The following muscles were examined for signs of IR: flexor tibiae, tarsal levator, tarsal depressor, retractor unguis, anterior coxal rotator, coxal adductors, coxal levators, trochanteral levators and dorsal longitudinal flight. Only in the flexors was evidence obtained for the existence of an IR. During isometric recordings made by attaching the force transducer to the tip of the tibia, two tension rhythms, with different amplitudes and durations occurred. One of them was an enhanced tension, attributable to the extensor. The other was of tension reductions and could only have been produced by the flexor (Fig. 7 A). On another occasion (Fig. 7B) there was strong interaction from reflexes, but the dominant rhythm was a flexion. However, rhythmic contractions were not obtained from any preparation during direct isometric recording from the flexors, so intrinsic rhythmicity is either rarely present in flexors or is easily suppressed there.

In a few preparations complex polyrhythms occurred. This condition was seen to arise out of a regular rhythm in three preparations, and in two others it gradually changed into a smooth IR. Intracellular recordings made from pacemaker muscle fibres at these times showed normal-looking IR depolarizations. It seems probable, then, that a highly irregular baseline occurs when IRs of individual fibres of fibre clusters cease to be synchronized, and vice versa.

Examples of polyrhythm, in which two or more units repeating at different frequencies could be recognized, are shown in Fig. 8. Figure 8 A shows one in which a large IR and a smaller one in the same muscle were very similar in shape and frequency. They summated together in simple algebraic fashion, showing clearly that they were generated by two independent pacemakers affecting two different sets of muscle fibres.

Another example, shown in Fig. 8B, is more complex: there were at least four components. One occurred at the same frequency as the stronger unit and was synchronized with it (1), leading to a step in the rising phase (arrows). Another (2) was small, and of short duration whilst a third (3) was small but of long duration. The latter had a ripple on its plateau phase. The muscle bundle had a strong basic tonus, and following the unit 1 IR contraction there was a large relaxation undershoot (R). But when the onset of unit 3 overlapped the unit 1 relaxation the undershoot was missing. The sub-rhythms commonly summed simply: more complex interaction sometimes occurred, and all sometimes merged into a single normal IR. Thus there are a few IR pacemakers in the jumping muscle or perhaps every muscle fibre involved has its own intrinsic rhythmicity, but somehow the fibres are normally synchronized. The question of how synchrony is brought about is of considerable interest in itself, and will be the object of future studies.

A further type of aberration could not so readily be explained. It consisted of a periodic fluctuation in amplitude, that was at times highly regular (Fig. 8C,D). It was as if there was a basic period causing equal contractions in about half the fibres, with progressive recruitment of other fibres on the second, third and fourth repetitions only.

When IR contractions had a long plateau phase, this tended to be interrupted by shallow quick relaxations, suggesting that some contributing units had a higher frequency than the dominant one.

It is to be expected that intrinsic rhythms of the two sides be completely independent, but in the intact or nearly-intact insect the possibility exists of neural coordination. There were indeed times when the left and right legs had perfectly synchronized IRs, of similar amplitudes and widths, but synchrony did not last for long, and it was concluded that in all cases it resulted from co-incidence. At times there was a marked disparity on the two sides, in regard to all parameters. There was, however, a strong tendency for the same types of overall fluctuation, frequency or amplitude shift, to occur in IRs of the two sides at about the same times. This was presumably because many central nervous influences affect both sides at the same time, as has to be the case for actions caused by unpaired neurones. Nevertheless, the magnitudes of the effects were seldom exactly matched on the two sides. On no occasion was an IR observed to be synchronized to respiratory movements.

In some preparations there was an undershoot of the tension baseline during the relaxation phase of each IR contraction, with slow return to the starting tension. In the intact animal, over a long period, the undershoot would come and go (see Fig. 4). The undershoot could occur only because there is a basic weak background tonus in the muscle fibres in which IR occurs. The undershoot was greatest when the background tonus was moderately large. When this relaxed, the tension undershoot disappeared. The background tonus appeared to be developed independently of IR, but when it was very large, as was occasionally the case, IR was reduced until, in some cases, it vanished altogether, to re-appear after the tonus had subsided somewhat. It became obvious during experiments on intact animals that basic tonus is under some kind of central nervous control. Octopamine, or stimulation of DUM neurones, not only inhibits IR but also affects the basic tonus, the common reaction being relaxation (see Fig. 1 in Hoyle, 1975), but sometimes contraction occurs (Fig. 9).

Whilst exploring the effects of a small electrode applied to the dorsal surface of the ganglion, a site was consistently found that had a dramatic relaxing action on basic tonus (BT) of the extensor tibiae (Fig. 10). This effect, unfortunately, has not been seen whilst stimulating any individual neurone intracellularly. The effective site was in region C4 (Burrows & Hoyle, 1973) and it affected BT on the stimulated side only. At the same time the amplitude of IR was reduced, but only if the frequency and total number of stimuli were above critical levels. There was a delayed, rebound increased in amplitude and frequency of IR if a sufficiently high frequency was used (Fig. 10E).

A pair of shocks at a separation of 4 s was sufficient to produce a detectable fall in BT, that lasted for about 70 s. Five shocks at the same frequency produced a larger decrease in BT, of 70 mg, that lasted for 180 s. Ten shocks at the same frequency relaxed BT by 120 mg for 210 s. At this frequency the amplitude of IR contractions was enhanced (as they started from a lower BT level) and there was only a very small slowing of IR frequency. At a shorter interval (2 s) a single IR was attenuated, and at a yet shorter interval (0·5 s) IR was completely suppressed for three cycles, attenuated for a further three, but followed by a period of enhancement (Fig. 10E). A 10 s stimulation at 2/s led to an abrupt relaxation of BT that lasted 396 s.

The rate of relaxation of BT, as well as its maximum extent, was also a function of the frequency of the stimulation. The duration of the period of relaxation of BT was a function of both the frequency and the number of shocks (Figs. 10, 11). As is also true for neurones enhancing ER, it is especially desirable to locate the specific neurones responsible for the extraordinary graded relaxation of BT.

BT is clearly present in the muscle fibres responsible for IR. Is it only present in these fibres? At the present time this question cannot be answered and further research will be needed to resolve it, but it seems unlikely, in principle. DUM neurone activity generally relaxes BT, but there have been exceptions. The same was true in the experiments in which octopamine caused relaxation of BT in one leg but enhancement in the other. Since the undershoot was not changed, the increase in BT was probably in muscle fibres other than those responsible for IR.

Nevertheless, as was mentioned earlier, IR is small when BT is high and may not be visible at all until BT relaxes. Occasionally, individual IR contractions were seen to lengthen, erratically, until an individual IR failed to relax (Fig. 12). This state was indistinguishable from one in which there was relatively strong BT.

No individual neurones were identified that caused IR amplitude to increase. But during localized stimulation of the dorsal surface of the ganglion, a site was found in quadrant B 5 of the map in Burrows & Hoyle (1973) that produced a marked, long-lasting enhancement of the amplitude of IR contractions (Fig. 12). The amplitude increase was not accompanied by a change in IR frequency, for low frequencies of stimulation, but at frequencies above about 10 Hz the frequency was increased at the same time as the amplitude was enhanced.

The amplitudes of both sides of the animal were affected at the same time, within i s of the onset of stimulation when this was commenced at about the time an IR contraction was anticipated. A frequency of about 10 Hz for 2−3 s was required to produce the effect. The results and stimulation requirements are compatible with their being associated with a DUM neurone. However, all DUM neurones that have been penetrated and directly stimulated with a microelectrode have produced inhibition, not enhancement, of IR, probably via octopamine release. It is possible that one of the smaller DUM neurones was responsible. However, only one of them travels in the extensor nerve, namely DUMETi, which is a strong inhibitor.

The enhancement occurred at exactly the same time on both sides provided the effective dorsal site was stimulated at a time when IRs of the two sides were expected to occur together (Fig. 13B). The delay in action cannot have been more than about 0·5 s. This suggests that the sites of release of the neurohumoral agent responsible are close to the extensor muscles.

The decay of the enhancement was a function of the duration of stimulation. Maximum enhancement following a 20 s stimulation at 10 Hz was 8·25 ×, and this decayed to one-half (4·1 ×) 35 s after cessation of stimulation. Following a 5 s stimulation at 10 Hz, the maximum amplitude enhancement was 6·4 ×, and this decayed to one-half in 15 s, i.e. by the next contraction. Return to normal took 4 min in the former experiment and 2 min in the latter. Thus the amount of enhancement, and its decay rate, are a function of the period of stimulation.

Whenever the whole animal was being studied, there was the possibility of spontaneous and reflex neural outputs occurring in SETi or CI that might interact with, and affect IR. Experiments were undertaken to investigate the effects of various frequencies, durations and timing of CI and SETi on IR.

CI does not share innervation of the muscle fibres on which DUMETi endings occur (Hoyle, 1978). If the latter act only on the fibres they innervate, a close direct interaction between the two neurones would not be expected. However, DUMETi makes neurosecretory-type terminals (Hoyle et al. 1974), so its action may be on more distant targets. No influence of DUMETi on CI action was observed, but CI was found to have strong effects on IR. When CI was stimulated at the same time as an individual IR contraction, it reduced its amplitude and duration. But unlike a DUM neurone, its action consisted only of that on the immediate contraction, there was no persistent effect reducing later contractions. When stimulated between bursts, at exactly the same repetition rate as the IR frequency, there was no long-term action either. Such bursts sometimes caused a small relaxation, sometimes a small contraction, or were without direct mechanical effect. Stimulation of CI for periods longer than a single IR cycle consistently altered IR, even at frequencies below 10 Hz, by reducing both the amplitude and duration of individual IR contractions (Fig. 14A). The reductions were slight and they declined with time under continued CI stimulation. This was true for CI frequencies below about 25 Hz, above which IR suppression was total.

Interposition of a strong CI burst during the course of an IR contraction quickly and fully suppressed it. However, if a second similar or even longer CI burst was applied during subsequent IR contractions, it was generally much less effective than the first (Fig. 14B(i)), although preparations varied considerably in this respect and a few showed the opposite effect (Fig. 14B(v)), first series). Several examples are shown in Fig. 14 in order to illustrate the extent to which inhibitability of IR by CI varied from moment to moment (Fig. 14B(i), second and third series). When an IR contraction was not completely inhibited a series of small contractions occurred at irregular intervals (Fig. 14B (ii), second series).

At frequencies below 15 Hz, amplitude and also duration, were at first reduced, but the attenuation waned rapidly. At higher frequencies all the effects were further enhanced until, at 25−30 Hz complete suppression occurred. Following inhibition of IR by CI, the first contraction is larger than normal and longer in duration. Its onset immediately follows cessation of CI stimulation. There was never a pause, such as would occur sometimes following only partial suppression of IR.

Single impulses in the slow extensor axon (SETi) cause a minute twitch in most locust/grasshopper preparations. This is generally quite small, even compared with an IR contraction, and is often not visible on IR records, but when SETi was excited at a low rate it nevertheless greatly influenced IR. The first SETi impulse to arrive tended to initiate a contraction (Fig. 15), even if the next IR contraction to be expected in the normal sequence was not due to start for several seconds. This triggering action will be described in detail later. It results in synchronizing of IR with SETi impulses at rates of 1 per 10 s to 1 per 5 s. At somewhat higher, but still low, frequencies a scaling process occurs, such as one IR contraction for every second, third or fourth SETi impulse, depending on the SETi stimulation rate (Fig. 15 A). Also, there is a tendency for SETi impulses to become progressively less effective with time, in triggering IR.

The amplitude of individual IR contractions was decreased slightly at the same time as IR frequency was increased by SETi action. Also the duration, as well as the amplitude, of an individual IR contraction was reduced. The extent of this reduction was increased the greater the SETi frequency, but some preparations were affected to a much greater extent than others. In general, if the IR frequency was low, and the duration of an individual IR contraction was long, the latter was but little changed by SETi. But if IR frequency was higher and duration of individual contraction short the latter was greatly shortened. Examples of these different effects are shown in Fig. 16 in which D(i) should be compared with Fig. 15E. It was in many ways surprising to find that SETi, which itself causes depolarizing junctional potentials and tension, reduces the amplitude of IR.

A few preparations did not show IR, as noted also by Floyle & Florey (in preparation) who found that addition of eserine and 10⁻⁶ACh to saline bathing inactive Locusta preparations initiated IR. In the present experiments, it was found that in such preparations IR could be initiated by stimulating SETi at a moderate frequency (Fig. 16B). IR thus evoked continued, however, only as long as SETi was excited and it was quelled by combining CI with SETi action (Fig. 15 B(ii)). The effects of stimulation of SETi at frequencies of 5−15 Hz were different in different preparations and generally very interesting. In some preparations stimulation of SETi at to Hz led only to a slight acceleration of IR (Fig. 15D(i)). In others, it produced an erratic contraction (Fig. 15 D (ii)) reminiscent of those sometimes initiated by causing repeated suppression of IR by CI bursts (Fig. 14B(iv, v)). In others there was a strong suppression of IR (Fig. 15 F). The explanation of these various effects seems to be that they are due to a combination of two actions. One is that a SETi impulse tends to initiate an IR prematurely during an inter-IR contraction period. Secondly, when an IR contraction is actually under way, a SETi impulse tends to cause it to terminate prematurely. Both effects will be described separately in a later section. Variations in the relative strengths of these two independent actions, between preparations, can account for the differences seen. The balance between the tendency of a SETi impulse to initiate an IR contraction prematurely and the tendency to cause its early termination once under way, also explains the effects of increasing frequency of stimulation of SETi on IR. As SETi frequency is increased, the frequency of IR rises but the duration of individual IR contractions falls (Fig. 15 C).

In some respects, the actions of SETi and CI on IR are similar: both lead to attenuation and acceleration. But during long-continued low-frequency CI stimulation the suppressed IR breaks through and eventually resumes nearly normal amplitude, duration and frequency. By contrast, the reduction in amplitude, shortening of duration and increase in frequency of IR that occur during SETi stimulation continue indefinitely.

The major difference is that following a period of CI inhibition of IR there is an immediate rebound; an IR contraction starts as soon as CI stimulation ceases. It will be recalled that both the amplitude and the duration of this rebound IR contraction are greater than normal (Fig. 16B). By contrast, following cessation of SETi action, there is no immediate rebound IR. Furthermore, the first IR contraction following SETi action tends to be of less than average amplitude (Fig. 16A, dep.).

Since the principal actions of CI and SETi, stimulated at low frequencies, on IR are similar, namely increased IR frequency but decreased IR duration, it might be expected that they would reinforce each other. On the other hand, the membrane potential changes that each effects are basically opposed to each other. The principal difference in their actions is that CI produces more powerful attenuation than does SETi.

When CI and SETi were stimulated together at low frequencies the dominant action was slow contraction, with suppression of IR. It required a CI : SETi ratio of 3 : 1 or greater for suppression of the SETi contraction (Fig. 15 C(iii)). At this ratio IR was still apparent, whereas with CI alone at this frequency it was completely suppressed. Thus SETi slightly antagonizes the inhibitory effect of CI on IR.

Little interaction was observed when DUM neurones were stimulated at the same time as SETi and CI. Stimulation of SETi after first causing a strong burst in a DUM neurone that inhibited the extensor tibiae IR led to contractions that resembled IR rather than normal slow ones. SETi counteracts the combined suppressing actions of both CI and DUMETi and restores IR if the frequency is not too high. In the experiment shown in Fig. 17A, a DUM neurone burst was elicited, that completely suppressed IR. During the inhibition, a brief SETi train at 6 Hz was applied. This immediately initiated IR, at a faster than normal rate. Thus SETi is easily able to overcome DUM neurone inhibition and restore IR.

In another experiment SETi was stimulated at 10 Hz, initiating a contraction that was then inhibited by stimulation of CI at the same time, at 20 Hz (Fig. 17B). A weak slow contraction slowly built up, with superimposed, very weak, IR contractions at increased frequency. During this period, a burst was elicited from a DUM neurone that inhibited IR. No additional attenuation of the abortive IR was caused. Normal IR was restored earlier than it would normally have been following such a burst. The longer period of SETi stimulation opposed the long-term action of DUMETi.

Whenever the whole animal is being studied, or in preparations in which the nerves have not been severed, there is a possibility of SETi, CI or any of the DUM neurones affecting IR. Examples of such action were shown in Fig. 5 and can now be interpreted in light of the effects obtained on IR whilst stimulating these axons.

In Fig. 5 A, the interpretation of the abrupt changes in IR are that excitation of the slow extensor tibiae (SETi) to the right leg only occurred, followed soon after by a prolonged SETi burst to the left leg. At the same time as the right SETi was excited the dorsal unpaired median neurone that innervates both extensor tibiae, DUMETi, fired. The DUMETi action was responsible for the abrupt attenuation of the left IR contraction and for the long interval before the next, small right IR contraction following the right SETi burst. How, though, can the acceleration in frequency of the left IR be explained, when we know DUMETi had released material that decreases IR frequency. The answer lies in the difference between the two SETi discharges, combined with the fact that SETi also interacts with IR. The left SETi discharge continued at a low frequency after the initial burst, as evidenced by the continued tension plateau. When a SETi neurone fires, the frequency of IR in that leg is increased and the amplitude is diminished (Hoyle, 1978): these actions evidently override the frequency-diminishing action of the DUMETi secretion.

In the middle set of IRs shown in Fig. 5 B, a right SETi burst was matched by a simultaneous burst in the common inhibitor. This reduced the left IR, but continuing, low-frequency right SETi firing after the burst caused acceleration of the right IR. Finally, in Fig. 5C is shown common events: variable attennuation of IR amplitude, with slight increase in frequency. These are both attributable to the presence of an accelerating prolonged discharge in the common inhibitor (CI). The erratic nature of the changes in IR reflects a fluctuating frequency in CI.

Following a FETi twitch the muscle does not relax fully, but instead goes into a premature IR contraction. Provided this was not caused within the first 2 s after termination of a preceding IR contraction it was similar in size and duration to the preceding one; having only been initiated early. The FETi impulse would trigger an IR contraction of normal amplitude and duration as early as 1·8 s following the termination of a preceding one, whatever the natural IR frequency. IR could routinely be driven synchronously at frequencies up to 5 × normal by repeated FETi triggering (Fig. 19) and sometimes even higher, though with a slight reduction in amplitude and duration.

If a FETi impulse occurred during the middle of an IR contraction, relaxation to the baseline occurs, i.e. the IR contraction is aborted. A critical time must follow its onset, before which it is not so terminated. This time was greater the lower the frequency of IR, or the longer the duration of individual IR contractions, which tend to have a long duration when their frequency is low. A linear relationship was found between the IR rest interval and the critical time for early termination by a FETi impulse. After the frequency of IR has been increased by FETi triggering for several cycles, upon cessation the normal interval is resumed right away.

Single SETi impulses also caused triggering and early termination in some preparations (see Figs. 16, 18). Others did not, but in these a SETi burst was effective if it was sufficiently long and of sufficiently high frequency (Fig. 19 B). The ones that did respond to a single SETi impulse gave unusually large SETi e.j.p.s and twitches. Single CI impulses never initiated early IR contractions, but it will be recalled that repeated CI impulses increase IR frequency, which could mean that there is at least a slight tendency for CI junctional potentials to trigger them. Nevertheless, single CI impulses were effective in initiating early termination (Fig. 20 A). The critical times for initiating early termination were slightly longer for SETi and CI than for FETi in the same preparations.

The quantitative relationship between the mean periods of IR contractions and duration of an IR contraction for naturally different durations are shown in Fig. 21B. The relationship between the duration of IR contraction (in one preparation) and the time before the next IR contraction started is shown in Fig. 21C. The duration was cut short by exciting FETi at various periods after the onset of a natural (i.e. not triggered) IR contraction. Both of these graphs are linear, with similar slopes, and both hint at an underlying relaxation oscillator-type of mechanism that is controlled by slow chemical accumulation. If an IR contraction is cut short the next one is bound to occur earlier, by an amount that is strictly linearly related to the abbreviated duration. The early onsets are not quite sufficient to keep the interval between termination constant and normal: there is slight overcompensation, so that the interval becomes less than normal following curtailment. In the normal unstimulated preparation, the duration of individual IR contractions slowly alters. There is a small decrease in frequency as the duration of contraction increases (Fig. 21B, lower line).

The present results show that in some insect skeletal muscles there are three independent mechanisms determining tonus. Each of the three types of tonus mechanism will be given a descriptive title. Only one category of tonus mechanism, from among several once thought to exist, survived ‘critical’ appraisals in the mid-1920s. The favoured one is low-frequency tetanus in one or a few motor units innervated by ordinary excitatory motor neurones. This will be termed tetanus tonus and it is one of the three forms present in locust and grasshopper jumping muscles. Arthropods in general have specialized slow motor neurones, often innervating specialized slow muscle fibres, that serve a similar function, possibly more efficiently, by way of graded, frequency-dependent synaptic depolarizations at distributed, multiple terminals on single muscle fibres. The other two tonus mechanisms are intrinsic to muscle fibres: basic tonus (BT) (Langelaan, 1922) and the intrinsic rhythm (IR). The former is the steady tension that is developed by some muscle fibres in the absence of motor nerve synaptic excitation. The latter is the slow myogenic rhythm described in this paper. There is some interaction between BT and IR, and also between IR and tetanus tonus. This is to be expected since all three act via the same muscle fibres and all act by way of membrane potential shifts.

During studies on the action of octopamine, which is considered to be released by DUM neurones (Hoyle, 1975), it was found that BT is relaxed in many preparations, but enhanced in a few and unaffected in others. Neither octopamine action nor DUM neurone action exactly mimic the actions of the specific BT relaxing neural elements encountered (see p. 185).

The first example of BT to be discovered in an arthropod muscle was the pupal moth spiracular muscle (Beckel & Schneiderman, 1957). Here it is sufficiently great that the spiracular valves are closed in the absence of neural input ; CO₂ is the natural relaxing agent. Similar phenomena also exist in adult locust spiracular muscles (Hoyle, 1961). BT was later found in a crustacean eye-stalk muscle (Hoyle, 1968) where it can be relaxed, very rapidly, by the peripheral inhibitory axon. There was no evidence, from the present work, that the locust peripheral inhibitory axon can relax BT. In fact, stimulation of CI led to the development of tension in most preparations. It might have been concluded that CI does not innervate muscle fibres responsible for BT, but since CI often gives depolarizing i.p.s.p.s this could not be ascertained without further tests.

The frequency of stimulation of the key neurone required to achieve a marked reduction in BT was only 1 per 2 s and even a pair of stimuli produced a detectable effect. The low frequency and small number of stimuli that were effective in relaxing BT together with a long duration of the action, strongly suggest that secretion of a neurohumoral type of material is responsible for the relaxation.

IR is reduced when the BT level increases and is lost completely when it is high. Conversely, as BT falls IR amplitude is enhanced. It must be presumed that there are functions for IR and BT in the jumping muscle. Although the strength of these is small, not more than 0·3 g, it should be borne in mind that a whole mature locust weighs barely 2·0 g. BT in the muscle is probably seldom sufficiently large to put the locust into a raised posture, but it does prevent complete slackness when SETi is silent which is often the case when the locust is at rest.

IR is masked in the intact animal by the resistance reflexes it evokes, though movements caused by it are still recordable from the whole animal (Hoyle & O’Shea, 1974). The peak tension developed in a large IR contraction is about equalled by a steady discharge in SETi at 10 Hz.

The existence of BT, and perhaps also IR, provides a previously unsuspected possible reason for the existence of common peripheral inhibitory neurones. Because both can be expected to interfere with the generation of behaviour, they need to be suppressed at times. However, CI, although it does reduce BT, does not suppress it completely, perhaps because fast muscle fibres are not innervated by it. The as yet unidentified dorsal neurones that suppress BT are very much more effective.

The results described have shown that there is a subtle interaction between the mechanism generating IR and effects associated with each of the ordinary motor axons. There are three ways in which the interactions might occur. They could be caused indirectly by the independent mechanical actions of the axons. This seems unlikely, because SETi and CI exert miniscule effects on tension individually, or at the low frequencies that affect IR. A second way in which the mediation might occur is by a direct action of the SETi and CI transmitter substances on the muscle fibres active in IR. This could be mediated either by conventional close synaptic action or after diffusion through haemolymph. If the muscle fibres in which IR originates are innervated by CI and SETi then DUMETi must exert its action on the pacemaker after diffusing, or being carried, in the haemolymph, since DUMETi terminals apparently only accompany FETi (Hoyle, 1978).

There is no obvious function for IR. Usherwood (personal communication) has suggested that it may aid the flow of haemolymph. Since it must interfere with normal reflex and centrally programmed behaviour, it is desirable to be able to turn it off when normal locomotor activity is called for, and that is something DUM neurones do.

The minimum latency for attenuation by DUMETi is about 100 ms (Hoyle, 1974), so the IR initiation site must be fairly close to the terminals. However, the muscle fibres on which these terminals occur are of fast type and they are not innervated by either the common inhibitor or the slow extensor. Since the latter affect IR with even shorter latencies, of about 50 ms, either IR is initiated in slow muscle fibres receiving CI and SETi innervation, or else all the relevant transmitter substances affect IR after diffusing a short distance to the generator sites.

The three types of terminals release different substances. CI is associated with a probable release of GABA and causes increased chloride conductance, and either hyperpolarizing, or weakly depolarizing i.j.p.s, whilst SETi causes only depolarizing e.j.p.s, possibly by release of L-glutamate (Usherwood & Cull-Candy, 1975). It is particularly surprising that SETi impulses attenuate, rather than simply sum with, IR. So the interaction between the motorneurone effects and IR are unlikely to be simple indirect consequences of membrane potential shifts but more probably represent actions on the ion conductance changes of the pacemaker.

The effects of two of the three axons in initiating an anticipated IR contraction prematurely, or of all three in terminating it early, are subtle. FETi is the more powerful in both respects: a single FETi impulse will trigger an IR contraction as little as 2 s after termination of a preceding one when the anticipated natural interval is from 7 s to more than 30 s. An IR having a natural frequency of i/min will follow for 1 h at a rate as high as 6/min when triggered by FETi impulses although the amplitude will diminish slightly. Single SETi impulses would trigger an early IR contraction in preparations with large SETi e.j.p.s, whereas a burst was required in those with small e.j.p.s. Neither a single, nor a burst of CI impulses triggers an IR contraction but during a low-frequency train of CIs the IR frequency always increased slightly, which may represent a similar, though weak, effect.

The critical time for termination of an IR contraction is directly proportional to its frequency. IR contractions are themselves briefer at higher frequencies though there is no simple relationship between IR frequency and duration. The critical time for SETi was slightly longer than for FETi, and that for CI about the same as that for SETi. Since a single CI impulse can terminate an IR contraction early, the action cannot be mechanical. The only aspect the three impulses always have in common is that they initiate some outward current flow. That may be sufficient to start repolarization of the pacemaker muscle fibres after the critical period. FETi and SETi first initiate inward current flow, and that may be what initiates the early IR contractions. The capability to generate an IR depolarization wave clearly resets faster than, and is independent of, the pacemaker to initiate it. Duration of IR contraction, frequency, critical time for attenuation and amplitude are all interrelated, although the first three can vary independently of the others to a considerable extent. The problems associated with these phenomena are familiar ones to students of cardiac pacemaking. This new preparation for studying such matters may therefore have a wider potential interest than its relevance to insect muscle physiology and behaviour.

This research was supported by National Science Foundation Researach Grant No. BNS 75-00463.

Whilst this paper was in press T. Piek & P. Mantel (1977) published an article on the extensor tibiae of Locusta migratoria in which they noted the normal absence of IR in locust saline. However, they found that addition of 10⁻⁹ mole per litre proctolin to the saline evoked IR. They also noted (their Fig. 3) two types of relaxation of tonus. One (ist asterisk) was equivalent to termination of a very long-lasting IR (see Fig. 12 of the present paper); after this relaxation IR was uncovered. The other (2nd asterisk) was clearly a relaxation of BT. Both were induced by proctolin.