Giant Fibre Activation of Direct Flight Muscles in Drosophila

A number of behavioural events act to initiate escape in many flies including Drosophila (Nachtigall, 1966; Nachtigall & Wilson, 1967; Kaplan & Trout, 1974). The most prominent events involve the mesothoracic legs and the wings. The mesothoracic legs drive a powerful escape jump, lifting the animal into the air. At the same time the wings are quickly elevated (Kaplan & Trout, 1974). The wings also move horizontally forward from the closed rest position into the opened flight position. Full wing opening is achieved over the course of the first six wingbeats (Nachtigall, 1966; Nachtigall & Wilson, 1967).

Escape may apparently be initiated by any of several neural pathways (Thomas, 1981). These pathways are, for the most part uncharacterized. However, recent studies have shown that the cervical giant fibres of Drosophila mediate a visually induced escape response by driving potentials in several escape muscles (Tanouye & Wyman, 1980; Thomas, 1980, 1981). The cervical giant fibres are a bilaterally symmetrical pair of large axons which run from the brain to the thoracic ganglion where their motor outputs occur (Power, 1948; Coggshall, Boschek & Buchner, 1973,King & Wyman, 1980; Koto et al. 1981). Giant fibre activation of the tergotrochanter muscle (TTM) appears to account for several of the behaviour patterns observed in escape initiation (Tanouye & Wyman, 1980). The TTM is a large tubular muscle which functions as the main extensor of the middle leg and is responsible for the escape jump (Williams & Williams, 1943; Nachtigall & Wilson, 1967; Mulloney, 1969). Anatomically, the TTM could function as an indirect wing elevator (Williams & Williams, 1943; Nachtigall & Wilson, 1967). Thus, its activation might be sufficient to explain the first wing elevation seen in escape initiation. The giant fibre also activates the dorsal longitudinal muscle (DLM) and dorsoventral muscles (DVMs) which respectively provide the main power for wing depression and wing elevation during flight (Tanouye & Wyman, 1980). Since these muscles are fibrillar muscles, providing powerful contractions only when stretched (Pringle, 1949; Machin & Pringle, 1959), they do not make substantial contributions to wing position during escape initiation (Nachtigall & Wilson, 1967).

Previous studies on the giant fibres of Drosophila have focused on the TTM, DLM and DVMs, muscles which control wing position indirectly through distortions of the thoracic walls (Tanouye & Wyman, 1980). The present study extends the earlier analyses by describing giant fibre activation of direct flight muscles, tubular muscles which attach directly to the wing base. In particular, all direct wing elevators and wing openers were examined for giant fibre activation. The relationship between these giant fibre-driven wing movements and movements seen during escape initiation is discussed.

Wild type Drosophila melanogaster of the Canton-Special strain were obtained from the collection of Dr S. Benzer, Division of Biology, California Institute of Technology, Pasadena, CA, U.S.A. Adult females (aged 4–6 days posteclosion) were used.

For examining wing and leg movements, the tip of an insect pin was glued to the dorsal surface of the head and thorax with Eastman 910 adhesive. The pin was held with the fly suspended in the flight position. For experiments requiring dissection, flies were mounted so that the left side, facing up, was exposed to Ringer’s solution while the right side, including all right side spiracles, was in contact with air circulating beneath the fly. Softseal tackiwax (Cenco Scientific) was used to keep the animal stationary and maintain the air/water interface. To facilitate placement of stimulating electrodes, the head was rotated 90 ° about the cervix so that the ventral surface of the head was facing up. The position of the head was stabilized with tackiwax.

The direct flight muscles were exposed by dissecting free and removing the mesopleurum (episternum) and part of the pteropleurite (Ferris, 1950). Particular care was taken not to damage air sacs and tracheoles. The Drosophila Ringer’s solution (Jan & Jan, 1976) contained 128 mm-NaCl, 2mm-KCl, 4mm-MgCl2, 35·5 mm-sucrose and l·8 mm-CaCl₂. Solutions were buffered to pH 7·1 with 5mm-Hepes (Sigma).

Electrical stimuli were delivered to the brain using insulated tungsten electrodes. Intracellular recordings of direct flight muscle potentials were made using glass micropipettes filled with a 4% aqueous solution of the dye Pontamine sky blue 6BX (G.T. Gurr Ltd, London). Pipettes were pulled on a Brown and Flaming puller (Sutter Instruments) and had resistances of 80–120 MΩ. Similar electrodes had resistances of 40–60 MΩ when filled with 3M-KCL Dorsal longitudinal muscle (DLM) potentials were recorded with insulated tungsten electrodes. Potentials were recorded through M4-A preamplifiers (WP Instruments) and displayed on a storage oscilloscope. Indifferent electrodes were Ag-AgCl in dissected preparations. In undissected preparations, indifferent electrodes were uninsulated tungsten electrodes located in the abdomen.

Response latencies were measured according to Tanouye & Wyman (1980). Measurements were made from the stimulus artifact (0·01 ms duration) to the first detectable voltage deflection of the evoked response. To facilitate determination of the latter, measurements were made using threshold stimuli and ineffective stimulus traces were used as baselines. Measurements were made directly from the storage oscilloscope screen or from a Polaroid photograph of the screen.

Giant fibre activation was identified physiologically as described previously (Tanouye & Wyman, 1980; Koto et al. 1981; Tanouye, Ferrus & Fujita, 1981). Identification relies on the fact that the giant fibre drives the DLM in a one-to-one manner. The muscle apparently is not driven one-to-one by any other axon descending from the brain (Tanouye & Wyman, 1980; Koto et al. 1981; Tanouye et al. 1981). At the start of each experiment, an electrical stimulus delivered to the brain was used to drive the DLM. Threshold voltage to this stimulus was determined. A similar threshold voltage for each penetrated direct flight muscle fibre was determined and compared with the DLM threshold. Results from muscle fibres which could not be driven by brain stimulation were discarded. Ten to twenty stimuli were usually used to determine threshold for each direct flight muscle fibre examined. An additional 10·20 stimuli were used to measure latency relationships between the evoked direct flight muscle potential and the DLM potential. After threshold voltage and latency measurements had been made, stimulus polarity was reversed and the values were remeasured. The change in polarity controlled, to a certain extent, for differences in stimulating electrode position.

Evoked direct flight muscle potentials with the same threshold as the DLM were considered to be driven by the giant fibre, and hence part of the giant fibre system. Previous studies on the giant fibre with dye-filled intracellular microelectrodes have confirmed that this physiological identification is sufficient for reliable identification of giant fibre activation (Tanouye & Wyman, 1980; Koto et al. 1981).

Fig. 1 shows a schematic representation of the muscles referred to in this report.

The dorsal longitudinal muscle (DLM) and the dorsoventral muscles (DVMs) are the indirect fibrillar muscles which respectively depress and elevate the wing during flight (Harcombe & Wyman, 1977, 1978; Tanouye & Wyman, 1981). The tergotrochanter muscle (TTM) is a large tubular muscle used to extend the mesothoracic leg, providing the main power for jumping (Williams & Williams, 1943; Nachtigall & Wilson, 1967; Mulloney, 1969). The TTM may also act as an indirect wing elevator (Williams & Williams, 1943; Nachtigall & Wilson, 1967). The nomenclature of Zalokar (1947) is used for the direct flight muscles. Equivalent names for anterior pleural muscles pal, pa2,pa3 and pa4 may be found in King & Tanouye (1983).

Electrode placement in the DLM was guided by its insertion site on the dorsal surface of the thorax (Tanouye & Wyman, 1980, 1981). Recording sites in direct flight muscle fibres were marked by a small amount of dye iontophoresed from the recording electrode. If more than one recording site was injected in a single specimen, the relative positions of these sites were mapped. At the end of each experiment, the thorax was fixed in Kahle’s fixative, dehydrated in an ethanol series and embedded in Epon via propylene oxide. Each thorax was serially sectioned at 5 μm. Recording sites were identified by the iontophoresed dye, which appeared on unstained sections as faint blue spots confined within single muscle fibres (Fig. 2).

The giant fibre was stimulated electrically on the brain, and wing movements were observed through the dissecting microscope. Prior to stimulation, the wings were folded over the abdomen in the normal resting position. Following a single stimulus, the wings quickly moved laterally from the closed position to a partially opened position, making an angle of about 45 ° with respect to the long axis of the fly. The wings did not remain open, but closed back to the resting position. The closing occurred in two phases. A rapid initial closing brought the wings to an intermediate position about halfway back to resting. A final movement back to the resting position was slower and in some cases took several seconds. In response to high frequency (10 Hz) stimulation of the giant fibre, the wings remained tonically open (45 ° relative to the body axis) for several seconds. In spite of continued stimulation, however, the wings did not remain open but closed to an intermediate position similar to that described above. This intermediate position was maintained throughout the duration of the high frequency stimulation. After cessation of the stimulation, the wings moved slowly back to the resting position.

In addition to wing opening, giant fibre activation caused a very slight wing elevation. This movement was not as prominent as wing opening and was best observed during high frequency (10 Hz) stimulation of the giant fibre. Unlike wing opening which occurred tonically during high frequency stimulation, wing elevations occurred as small discrete movements correlated with the delivery of each stimulus. Interestingly, mesothoracic leg extensions were similarly discrete movements during high frequency stimulation (see below). Wing elevations were only observed while the wing was opened about 45° (above). They were not observed when the wing had closed to the intermediate position.

The remainder of this paper deals with the problem of how the giant fibre system generates wing elevation and opening movements. In particular, intracellular microelectrode recordings of direct wing elevator muscle fibres and direct wing opener muscle fibres were made during giant fibre activation. These recordings defined which of the direct flight muscles were driven by the giant fibre system. An earlier study described giant fibre activation of indirect wing elevators (Tanouye & Wyman, 1980). Thus, taken together, all major muscles responsible for wing elevation and opening have been tested and a complete description of giant fibre system contributions of these movements is presented here.

Intracellular microelectrode recordings from direct flight muscle fibres gave resting potential values of —80 to —90 mV and evoked potentials of 70–90 mV amplitude. Spontaneous activity was rarely observed. Action potential and resting potential amplitudes in direct flight muscle fibres declined rapidly by 20–40 mV following the delivery of three-to-five effective stimuli. This was probably due to damage at the microelectrode recording site caused by contraction of the muscle fibre. The changes in amplitude had no effect on the voltage thresholds required to evoke muscle potentials and the latencies of the potentials. Thus, the amplitude changes do not alter the main conclusions presented in this report and no attempt was made to minimize muscle fibre contractions.

The anterior pleural muscles pal and pa2 (Fig. 1) are the direct flight muscles responsible for wing elevation (Williams & Williams, 1943; Nachtigall & Wilson, 1967). Fig. 3 shows evoked potentials recorded in the DLM and pal. The DLM and pal have different thresholds to brain stimulation suggesting that pal is not driven by the giant fibre. In Fig. 3A stimulus voltage was at threshold for the DLM response (9 V). Two stimuli were delivered. One stimulus was effective and evoked a potential in the DLM. A second stimulus of the same intensity was ineffective and failed to evoke the DLM response. The ineffective trace was superimposed on the effective trace and serves as a baseline. Note that neither stimulus evoked a response in the pal fibre.

Fig. 3B shows responses for the same two fibres where stimulus voltage was at threshold for the pal response (17 V). One stimulus was effective in evoking a pal potential. A second stimulus was ineffective. Since the stimuli were above threshold for the DLM response, both stimuli evoked DLM potentials. These results indicate that the pal has a different threshold to brain stimulation than the DLM. The suggestion is that pal is not driven by the giant fibre. Similar results were obtained for seven pal fibres in five animals. The pal is, however, activated one-to-one by another, higher threshold, descending pathway which can be activated by brain stimulation. This descending pathway was not identified in the present analysis.

Similar results were obtained for the other direct wing elevator muscle, pa2. Like pal,pa2 was activated by brain stimulation but at a threshold of stimulation different from DLM activation. Thus, pa2 is likewise not driven by the giant fibre but rather is activated one-to-one by another, non-giant fibre, descending pathway, pal andpa2 are the only direct wing muscles thought to subserve wing elevation (Nachtigall & Wilson, 1967). Thus, the results presented here indicate that giant fibre-driven wing elevation movements are not mediated by direct wing muscles.

The anterior pleural muscles pa3 and pa4 (Fig. 1) are the direct wing muscles responsible for wing opening (Williams & Williams, 1943; Nachtigall & Wilson, 1967; Heide, 1971). Results presented in this section show that pa3 is activated by the giant fibre; pa4 is not. Thus, giant fibre-driven wing opening movements appear to be mediated solely via activation of pa3.

Fig. 4 shows evoked potentials recorded in the DLM and a direct flight muscle fibre of pa3. The evoked muscle potentials had the same threshold and were never evoked independently. In Fig. 4, two threshold stimuli were delivered to the brain. One stimulus was effective and evoked potentials in both the DLM and thepa3 fibre. One stimulus of the same intensity was ineffective and failed to evoke potentials in either fibre. The ineffective trace was superimposed on the effective trace and appears as a baseline. Similar results were obtained for fivepaJ fibres in four animals. Thus, in response to electrical stimulation of the brain, pa3 and the DLM share the same threshold. Since the DLM is activated via stimulation of the giant fibre (Tanouye & Wyman, 1980; Koto et al. 1981), the suggestion is that pa3 muscle potentials are also activated by the giant fibre andpaJ is part of the giant fibre system of escape initiation.

As well as having identical thresholds, evoked muscle potentials in the DLM and pa3 showed characteristic latency relationships. In experiments directly comparing responses in the DLM and pa3, the average latency of the evoked DLM response 1·13 ± 0·05ms(s.D.) (five recordings). This value was similar to those reported by Tanouye & Wyman (1980). The average latency of the evoked pa3 response was 1·28 ± 0·10 ms (S.D.). The average latency difference recorded between evoked DLM andpa3 responses was 0·15 ± 0·10 ms (S.D.). Based on anatomical (King & Wyman, 1980; Koto et al. 1981) and physiological (Tanouye & Wyman, 1980) arguments, only one chemical synapse appears to be interposed between the giant fibre and the DIM motoneurone. The similarity in latencies between the pa3 and DLM responses suggests that only one, or possibly two chemical synapses, might be interposed between the giant fibre and the pa3 motoneurone.

The results for the direct wing opener pa4 were similar to those reported earlier for the direct wing elevators. That is, pa4 was activated by brain stimulation but at a threshold of stimulation different from DLM activation. Thus, pa4 is not driven by the giant fibre but is activated one-to-one by another, non-giant fibre, descending pathway. pa3 and pa4 are the only muscles thought to subserve wing opening (Nachtigall & Wilson, 1967). Taken together, the results presented here suggest that giant fibre-driven wing opening movements during escape initiation are mediated solely through activation of pa3.

The giant fibre system has proved to be an ideal subject for physiological and genetic study of motor pathways in Drosophila. Several studies have shown that the giant fibre system is responsible for rapid activation of the fibrillar muscles and the TTM. However, these muscles alone are not capable of producing the wing opening movements that accompany giant fibre stimulation. It has also not been clear whether the wing elevation that occurs upon giant fibre stimulation is produced by the TTM (an indirect wing elevator) or by direct flight muscles. Our observations were designed to give a more complete understanding of the role of the giant fibre system in escape behaviour. We have reached the following conclusions: (1) wing opening during giant fibre stimulation is due to activation of muscle pa3 alone. No other wing opener is closely linked to the giant fibre pathway. (2) Since no direct wing elevators are activated by giant fibre stimulation, the slight wing elevation which occurs must be powered by an indirect muscle. The TTM is the only candidate with appropriate stimulus-response properties. (3) Other muscles served by other descending pathways may participate in escape-flight initiation under natural conditions.

Electrical stimuli delivered to the brain evoke potentials in the DLM at a sharply defined threshold of stimulation. The present experiments examined several phenomena which had the same threshold. Specifically, the electrical stimuli drove wing opening movements, wing elevation movements, and potentials in the pa3 wing opener muscle. Since the DLM is activated via the giant fibre (Tanouye & Wyman, 1980; Koto et al. 1981), the suggestion is that the wing movements and pa3 muscle potentials are also activated by the giant fibre. In contrast, the muscles pal, pa2 and pa4 were activated at different thresholds of stimulation from the DLM. Thus activation of these muscles was via non-giant fibre descending pathways. All thoracic muscles previously examined (i.e. TTM, DLM, DVMs) have been found to be drive By the giant fibre. The muscles pal, pa2 and pa4 are the first thoracic muscles analysed which are not activated by giant fibre stimulation.

The above conclusions are dependent on the reliability of DLM activation threshold to brain stimulation as an accurate predictor of giant fibre activation. As argued by Tanouye & Wyman (1980), an electrical stimulus grossly applied to the brain might not be able to discriminate between pathways with slightly different thresholds. Such a problem would seriously alter the conclusions presented in this report. Arguing against this possibility is the finding that different thresholds were always observed for pal, pa2 and pa4 activation, but were never observed for pa3 and wing movement activation. Thus, electrical stimulation of the brain can, apparently, distinguish between different neural pathways descending to the thorax. Also, a number of recent investigations examining the giant fibre with intracellular microelectrodes (Tanouye & Wyman, 1980; Koto et al. 1981; Tanouye et al. 1981) have shown that DLM activation is a reliable indication of giant fibre activation. Thus, it is suggested here that pa3 and the wing movements are driven by the giant fibre and thus, part of the giant fibre system of escape initiation. The muscles pal, pa2 and pa4 are not part of the giant fibre system and any role they play in escape initiation must be via other, non-giant fibre neural pathways.

The most prominent wing movement driven by the giant fibre was an opening of the wings. Only two muscles in the fly, pa3 and pa4, subserve wing opening (Williams & Williams, 1943; Nachtigall & Wilson, 1967; Heide, 1971). Giant fibre stimulation was found to drive pa3 but not pa4. Thus, giant fibre-driven wing opening movements appear to be generated solely by pa3.

pa3 activity in flight has been best described for larger flies (Nachtigall, 1966; Nachtigall & Wilson, 1967; Heide, 1971). pa3 is most prominently active during extreme turning movements (Nachtigall & Wilson, 1967; Heide, 1971).pa3 activity and wing opening movements are also observed at the start of flight (Nachtigall, 1966). However, considerable variability was observed, and pa3 was not active in all starts. Since giant fibre activity was not monitored in these studies, it is not known ilpaJ is not driven by giant fibres in larger Diptera, or if only a small fraction of starts examined were giant fibre-driven. The present study shows that for Drosophila, the giant fibre always drives pa3, and thus must participate in all starts initiated by giant fibre activity.

The failure of the giant fibre to drive pa4 is somewhat surprising since anatomical investigation (King & Tanouye, 1983) has shown that the motor axon innervating pa4 is one of the largest in the Drosophila nervous system. Thepa4 motor axon is comparable in size to the giant fibre and TTM motor axon, both integral parts of the giant fibre system (King & Wyman, 1980; Tanouye & Wyman, 1980). Thepa4 motor axon is the only large calibre fibre thus far examined which is not part of the giant fibre system. In contrast, the motor axon innervating pa3, which is driven by the giant fibre, is of relatively modest diameter (King & Tanouye, 1983).

In addition to wing opening movements, the giant fibre also drove wing elevation. Giant fibre-driven wing elevation was not as prominent as wing opening and was only reliably observed during high frequency giant fibre stimulation experiments. Only two direct wing muscles, pal and pa2, are known to be wing elevators (Nachtigall & wilson, 1967). Neither was found in the present study to be driven by giant fibre stimulation. The only elevators known to be driven by the giant fibre are indirect wing elevators, the TTM and the three DVM muscles (Tanouye & Wyman, 1980). The DVMs are fibrillar muscles and contract strongly only when stretched (Pringle, 1965). Thus, the likely candidate to explain wing elevation movements due to giant fibre stimulation is the TTM.

The results of Kaplan & Trout (1974) showed a very prominent wing elevation at the start of escape initiation in Drosophila. In contrast, only slight wing elevation movements were found to be driven by the giant fibre in the present study. Direct comparisons between the present results and those of Kaplan & Trout (1974) are difficult, however, since giant fibre activation was not monitored in the earlier study. Also, it is possible that the giant fibre drives a larger wing elevation under more natural conditions than in the present report. In particular, observations in larger flies (Nachtigall, 1966; Nachtigall & Wilson, 1967) have shown that 15–30ms prior to escape initiation, the pleurosternal muscle is often activated tonically. The pleuro-sternal muscle improves the effectiveness of indirect flight muscle contractions by controlling the lateral stiffness of the thorax. Since pleurosternal muscle activity was not monitored in the present experiments, the influence of its activation on these observations is not known.

The present paper provides an initial physiological examination of direct flight muscles in Drosophila. To our knowledge, such recordings have not been presented previously. Electrical activity in each of the muscles is easily monitored since they are all activated by inputs descending from the brain. In the case oipa3, activation is via the giant fibre, thus extending our understanding of the giant fibre system of escape initiation. The short and characteristic latency of the response suggests a relatively simple pathway between the giant fibre and the pa3 motoneurone. Subsequent analysis will provide a description of the pathway similar to that provided for the giant fibre-DLM and giant fibre-TTM pathways (King & Wyman, 1980; Tanouye & Wyman, 1980). pal, pa2 and pa4 are activated via unknown pathways descending from the brain. This activation, however, should facilitate identification and characterization of these pathways and is an important next step toward the long range goal of providing a complete understanding of escape in Drosophila.

We thank Dr Seymour Benzer for the use of his facilities and his support during these experiments. We thank our colleagues Drs L. Kauvar, S. Shotwell, B. Tempel, T. Venkatesh, C.-F. Wu and L. Zipursky for discussion. This research was supported by National Science Foundation Grant PCM 7911771 to S. Benzer and the Pew Memorial Trust. M. A. T. was supported by National Institutes of Health Training Grant GM 7401-02.