Blockade of the central generator of locomotor rhythm by noncompetitive NMDA receptor antagonists inDrosophila larvae

Blockade of the Central Generator of Locomotor Rhythm by Noncompetitive NMDA Receptor Antagonists in Drosophila Larvae Daniel Cattaert,1,2 Serge Birman3 1

Laboratoire Neurobiologie et Mouvements, CNRS, 31 Chemin Joseph Aiguier, 13402 Marseille Cedex 20, France 2

Laboratoire de Neurobiologie des Re´seaux, UMR 5816, CNRS - Universite´ de Bordeaux 1, Avenue des Faculte´s, 33401 Talence, France 3

Laboratoire de Neurobiologie Cellulaire et Fonctionnelle, CNRS, 31 Chemin Joseph Aiguier, 13402 Marseille Cedex 20, France

Received 27 April 2000; accepted 5 February 2001

ABSTRACT:

The noncompetitive antagonists of the vertebrate N-methyl-D-aspartate (NMDA) receptor dizocilpine (MK 801) and phencyclidine (PCP), delivered in food, were found to induce a marked and reversible inhibition of locomotor activity in Drosophila melanogaster larvae. To determine the site of action of these antagonists, we used an in vitro preparation of the Drosophila third-instar larva, preserving the central nervous system and segmental nerves with their connections to muscle fibers of the body wall. Intracellular recordings were made from ventral muscle fibers 6 and 7 in the abdominal segments. In most larvae, long-lasting (>1 h) spontaneous rhythmic motor activities were recorded in the absence of pharmacological activation. After sectioning of the connections between the brain and abdominal ganglia, the rhythm disappeared, but it could be partially restored by perfusing the muscarinic agonist oxotremorine, indicating that the activity was generated in

the ventral nerve cord. MK 801 and PCP rapidly and efficiently inhibited the locomotor rhythm in a dosedependent manner, the rhythm being totally blocked in 2 min with doses over 0.1 mg/mL. In contrast, more hydrophilic competitive NMDA antagonists had no effect on the motor rhythm in this preparation. MK 801 did not affect neuromuscular glutamatergic transmission at similar doses, as demonstrated by monitoring the responses elicited by electrical stimulation of the motor nerve or pressure applied glutamate. The presence of oxotremorine did not prevent the blocking effect of MK 801. These results show that MK 801 and PCP specifically inhibit centrally generated rhythmic activity in Drosophila, and suggest a possible role for NMDA-like receptors in locomotor rhythm control in the insect CNS. © 2001 John Wiley & Sons, Inc. J Neurobiol 48: 58 –73, 2001 Keywords: locomotor rhythm; glutamate; MK 801; phencyclidine; NMDA receptor; Drosophila melanogaster

INTRODUCTION

(CNS). Previous evidence showed that, in arthropods, rhythmic motor activities are controlled by local neural networks, the central pattern generators (CPGs), located in the ventral nerve cord ganglia. This was demonstrated in insects for adult walking (Ryckebusch and Laurent, 1994; Berkowitz and Laurent, 1996), larval feeding (Gorczyca et al., 1991), and crawling (Johnston and Levine, 1996), and in crustacea for swimmeret beating (Heitler and Pearson,

Little is known about the neuronal circuitry controlling locomotion in the insect central nervous system Correspondence to: D. Cattaert ([email protected]) or S. Birman ([email protected]). Contract grant sponsor: Centre National de la Recherche Scientifique. © 2001 John Wiley & Sons, Inc.

58

Inhibition of Locomotor Rhythms in Drosophila

1980), walking (Chrachri and Clarac, 1990), and respiration (Wilkens and DiCaprio, 1994; Dicaprio, 1997). However, the neuroactive transmitters or peptides directly involved in motor rhythm generation have not been identified yet. The availability of genetic, molecular, and biophysical techniques makes the fruit fly Drosophila an ideal system for the study of neuronal function, from ion channels (Jan and Jan, 1997; Engel and Wu, 1998) to networks (Helfrich, 1986; Wolf and Heisenberg, 1990; Gorczyca et al., 1991). Several locomotor activity mutants have been previously isolated in Drosophila (Sokolowski and Hansell, 1992; Hoshino et al., 1993; Strauss and Heisenberg, 1993; Varnam et al., 1996; Martin et al., 1998). Identification of the corresponding genes should help correlate mutant phenotypes with structural or functional alterations in the nervous system (Greenspan, 1997). L-Glutamate has long been known to be involved in excitatory neurotransmission at the neuromuscular junction (NMJ) in arthropods (Gerschenfeld, 1973) and Drosophila (Jan and Jan, 1976b; Johansen et al., 1989). The physiology and development of the glutamatergic NMJ in Drosophila late embryos and larvae have been extensively characterized (McLarnon and Quastel, 1988; Delgado et al., 1989; Broadie and Bate, 1993; Keshishian et al., 1993; Keshishian et al., 1994; Kidokoro and Nishikawa, 1994). In the ventral ganglia (segments A1–A7), about 34 motoneurons per hemisegment were identified using an antibody to glutamate and retrograde labeling methods (Sink and Whitington, 1991; Landgraf et al., 1997) that innervate in a stereotyped pattern 30 muscles per hemisegment (Keshishian et al., 1994; Bate et al., 1999). The body wall muscle cells in the Drosophila larva are easily identifiable and accessible for intracellular recording (Crossley, 1978). The glutamatergic motoneurons form large boutons (type I) structurally characterized by the subsynaptic reticulum, a differentiation of the postsynaptic membrane composed of several convoluted layers. In addition, glutamate is enriched in subsets of interneurons in the insect brain and ventral nerve cord (Na¨ssel, 1996), indicating that it is a transmitter also within the CNS. To further elucidate the physiological requirements for glutamate in locomotion control, we exposed Drosophila larvae to various antagonists of vertebrate glutamate receptors. We observed that two noncompetitive glutamate receptor antagonists, MK 801 and PCP, delivered in food, specifically induced a marked inhibition of locomotor activity in Drosophila larvae. These antagonists are potent and selective blockers of the activated NMDA receptor channel in vertebrates (Reynolds and Miller, 1990; Lipton, 1993). An in

59

vitro preparation of the larvae was used to analyze on which targets (CNS, nerve fiber, NMJ) these antagonists are active. We describe here the characteristic patterns of locomotor rhythm that were reproducibly observed in this preparation. Our results suggest that MK 801 and PCP inhibit a locomotor CPG located at a central site in the Drosophila ventral nerve cord.

METHODS Pharmacology The glutamatergic antagonists (⫹)-MK 801 maleate (MK 801, dizocilpine), CPP {3-[(RS)-2-carboxypiperazin-4-yl]propyl-1-phosphonic acid}, and AP-5 (DL-2-amino-5-phosphonopentanoic acid) were purchased from Tocris Cookson (Bristol, UK). Oxotremorine (oxo) and phencyclidine (PCP) were from Sigma (St. Louis, MO).

Drosophila Culture and Drug Feeding Wild-type Oregon R Drosophila melanogaster were raised on standard medium at 25°C on a 12-h light/dark cycle. The method used to feed antagonists to larvae was adapted from the procedure described by Neckameyer (1996). Flies were allowed to lay eggs for 4 h on grape juice-agar plates. Hatched first-instar larvae were collected 23 h later on yeast paste (0.5– 0.6 g baker’s yeast in 1 mL H2O) in the center of the plate. Larvae were maintained on yeast paste for an additional 24 h in a culture dish in a moist chamber. The second-instar larvae were collected, briefly washed in water, and weighed. Twenty-five to fifty mg of these larvae were placed in a paste made with 125 mg dry yeast and 200 –250 ␮L of a solution containing 0.1–10 mg/mL glutamatergic antagonist. The locomotion behavior of control and treated third-instar larvae was examined 24 h later.

In Vitro Preparation For electrophysiological experiments, we used wild-type third-instar larvae that were not previously exposed to glutamatergic antagonists. Larvae were dissected by making a longitudinal mid-dorsal incision and pinning the cuticle flat [Fig. 1(A)] on a Sylgard-lined Petri dish. The internal organs were carefully removed to expose the body wall muscles and the nervous system. The preparation was continuously superfused with an oxygenated haemolymph-like saline solution similar to HL3 of Stewart et al. (1994) (in mM): 70 NaCl, 5 KCl, 1.5 CaCl2, 20 MgCl2, 10 NaHCO3, 120 sucrose, and 5 HEPES at pH 7.6. Experiments were carried out at room temperature (20 –22°C).

Electrodes and Recordings Intracellular recordings from ventral longitudinal muscle fibers 6 and 7 (Crossley, 1978) (Fig. 1) in abdominal seg-

60

Cattaert and Birman

Figure 1 Experimental arrangement. (A) General schematic view of the dissected Drosophila third-instar larva showing the disposition of intracellular recording electrode from muscle fibers, suction electrode from the corresponding motor nerve, and glutamate pressure ejection pipette. Anterior is on top. (B) Detail of the body wall muscle innervation in an abdominal hemisegment (A2–A7) [redrawn from fig. 1B of Van Vactor et al. (1993)]. Muscle fibers are designated by numbers (Crossley, 1978). (C) Detail of a muscle fiber with experimental arrangement.

ments were made with thin-walled glass microelectrodes (25–30 M⍀ resistance) filled with 3 M KCl. Excitatory junctional potentials (EJPs) were evoked by stimulating the appropriate segmental nerve with a glass suction electrode [Fig. 1(A)] that had been head-polished to a 10-␮m inside diameter. In some experiments glutamate was applied by pressure ejection (50 ms, two bars) through a glass microelectrode (tip diameter 5–10 ␮m) using a Picospritzer II (General Valve Corporation, Fairfield, NJ). The microejections were delivered specifically very close to the recording intracellular electrode implanted in the muscle fiber. The signals were amplified by an Axoclamp 2B (Axon Instruments, Inc., Foster City, CA). Intracellular current pulses delivered through the recording microelectrode and nerve stimulation were controlled by an 8-channel digital stimulator (A.M.P.I, Jerusalem, Israel). All physiological recordings were monitored on a 4-channel digital oscilloscope (Yokogawa DL 1200, Tokyo, Japan) on a digital tape recorder (BioLogics DTR 1802, Claix, France), and digitized on a PC-based computer through an A/D interface (Cambridge Electronic Device, CED 1401Plus, Cambridge, UK). Intracellular and extracellular recordings were digitized at 5–10 kHz and written to disk. Signals were analyzed using the SPIKE2 CED software.

Statistical Analysis Statistical analyses were performed by the GraphPad PRISM program (GraphPad Software, Inc., San Diego, CA). All results are given as mean ⫾S.E.M. The significance of changes induced by drug application was measured by means of a one-way ANOVA followed by post hoc analysis by the Newman-Keuls multiple-comparison procedure.

RESULTS Effect of MK 801 and PCP on Living Drosophila Larvae Drosophila larvae were markedly and reproducibly paralyzed after exposure to various levels (0.5–20 mg/mL) of the glutamate receptor antagonists MK 801 maleate (six independent experiments) and PCP (four independent experiments) provided in the food. 0.5 mg/mL (approximately 1.5 mM) MK 801 was enough to induce akinesia, whereas lower doses

Inhibition of Locomotor Rhythms in Drosophila

(0.01– 0.05 mg/mL) had no effect. In addition, Drosophila larvae presented an aversion to PCP. After 24 h in the presence of the antagonists, most of the paralyzed larvae were still alive; they presented low amplitude tremor-like movements of the anterior part of the body and responded to external stimuli by very slow crawling. Treated larvae were able to recover and continue development up to the adult stage when transferred to fresh vials with no drugs present, showing that the effects of these NMDA receptor antagonists are reversible.

In Vitro Locomotor Activity Patterns of Drosophila Larvae We used an in vitro preparation of the larvae to investigate further the effects of MK 801 and PCP on locomotion. The experimental arrangement used to record neuromuscular and locomotor activities is presented in Figure 1 (see Methods). This preparation preserves the CNS and segmental nerves with their connections to muscle fibers of the body wall. Most of the dissected larvae (n ⫽ 44 out of 56, i.e., 78%) presented a spontaneous locomotor activity, characterized by muscle contraction waves propagated forward or backward. Intracellular recordings from ventral longitudinal muscle fibers 6 and 7 (Fig. 1) in abdominal segments revealed rhythmic bursts of EJPs correlated with the rhythmic contractions [Fig. 2(A)]. Care was taken to place the recording intracellular electrode in muscle fibers that did not move too much during contractions. Normally, the resting membrane potential stayed stable for 30 min to 3 h. Only stable resting potential recordings (⫺61.68 ⫾ 0.62 mV; n ⫽ 53) were retained for analysis. We observed two types of spontaneous rhythmic activities [Fig. 2(A)]. The first type consisted of long bursts (period 14.1 ⫾ 1.5 s; duration 12.4 ⫾ 1.4 s; n ⫽ 29 experiments) with a more or less continuous tonic inter-burst discharge [Fig. 2(A), pattern 1]. In the second type [Fig. 2(A), pattern 2], the period was generally shorter [6.15 ⫾ 0.21 s; Fig. 2(B)] and the burst duration was always markedly shorter [1.62 ⫾ 0.12 s; n ⫽ 34 experiments; Fig. 2(C)], with what was generally a silent inter-burst [Fig. 2(A), (2)]. The two rhythm patterns were frequently observed alternatively in the same muscle fiber during the experiment (Fig. 2). Each rhythmic activity generally lasted for up to several minutes. We observed that the first type of rhythmic activity corresponded to posteriorto-anterior waves of body wall muscle contraction, as occurs in normal anterograde crawling, and the second type to anterior-to-posterior waves of contraction, as in retrograde crawling. In four experiments, the

61

type 2 rhythm was uniquely present [Fig. 2(D)]. The rhythmic activity frequently lasted for more than 1 h. In the example presented in Figure 2(D), a continuous intracellular recording was made from a muscle fiber for 140 min; the five sequences presented show that the rhythm did not change markedly along time, except for a slight tendency to accelerate. During such rhythmic activities, the CNS was sectioned at the level of the brain hemispheres (n ⫽ 5), or immediately posterior to the subesophageal ganglia (n ⫽ 4) (Fig. 3). As a result, all rhythmic activity was immediately abolished (n ⫽ 9) and did not recover spontaneously even after more than 1 h (n ⫽ 3) [Fig. 3(A)]. However, if the muscarinic agonist oxo (10⫺5 M) was added to the bath 1 min after the section, a rhythmic activity was restored in four out of six experiments [Fig. 3(B,C)]. In the restored rhythm, type 1 sequences (period 14.51 ⫾ 0.77 s; n ⫽ 3 experiments) and type 2 sequences with shorter bursts (period 14.01 ⫾ 1.57 s; n ⫽ 4 experiments) generally alternated [three out of four experiments; Fig. 3(C)]. However, contrary to the initial motor rhythm, a tonic discharge was observed between bursts during the restored type 2 rhythm in two experiments [Fig. 3(B)]. This pharmacologically induced rhythm [Fig. 3(B,C)] was, however, more irregular than the control activity recorded before the section. The rhythm was significantly slower than its control value in two experiments for pattern 2 and in one experiment for pattern 1 (Table 1). The duration of burst during pattern 1 was not significantly different from its control value measured before the section in the three experiments where a comparison was possible (Table 1). By contrast, the pattern 2 induced by oxo in the isolated nerve cord displayed significantly shorter bursts in one experiment (Table 1). These results suggest that the motor rhythm generator is located in the ventral nerve cord, but requires an activation by (or an interaction with) the brain to elicit a locomotor activity. However, the fact that oxo elicited tonic activities in the isolated nerve cord in some experiments raises the possibility that oxo exerts a direct effect on the motoneurons themselves.

Blockade of the Locomotor Rhythm by MK 801 and PCP When the NMDA receptor antagonist MK 801 was added to the bath at a low concentration (0.01– 0.03 mg/mL) during a stable locomotor rhythm, the locomotor activity was slowed down and the burst duration was markedly reduced [Fig. 4(A,B)]. During wash, the rhythmic activity was progressively restored, but the reduction in burst duration generally

62

Cattaert and Birman

Figure 2 Spontaneous rhythmic activity recorded from a muscle fiber. (A) Intracellular recording from muscle fiber 6 shows alternation between two patterns of activity. The first pattern (1) is characterized by long duration bursts of EJPs in which the EJP frequency progressively increases (see detail on the right). The second pattern (2) is characterized by short EJP bursts occurring at a faster rhythm than pattern (1); the EJP frequency rapidly increases and decreases and successive bursts are separated by a silent period. Note that, although muscle fiber 6 is innervated by two different motor axons, the small EJP units were not clearly resolved in these recordings. (B) Evolution of the rhythm period with time over 20 min. (C) Evolution of burst duration with time. Alternation of the two activity patterns occurs randomly and is revealed by long duration and short duration (2) bursts. (D) Intracellular recording from muscle fiber 7 showing the persistence of rhythmic activity along time. Activity was recorded from the same muscle fiber over 2 h and is illustrated by five bursting sequences (40 s each) recorded every 30 min.

Inhibition of Locomotor Rhythms in Drosophila

63

Figure 3 Effect of sectioning the nerve cord on rhythmic activity. (A) Intramuscular recordings. After sectioning (line with arrows) between the ventral nerve cord (VC) and the cerebral ganglia (CG), the isolated nerve cord rapidly stops producing rhythmic bursts. (B,C) The perfusion of 10 ␮M oxotremorine (oxo), an agonist of muscarinic cholinergic receptors, can restore type 2 (B) or type 1 (C) rhythmic activities on an isolated ventral nerve cord.

persisted for a longer time [Fig. 4(A)]. In most of the intracellular recordings made from muscle fibers 6 and 7, we observed two distinct types of EJPs, large and small [Fig. 4(C)]. In some recordings, the amplitude of the small EJP units was very small compared to the large ones and was not clearly resolved [see Fig. 2(A)]. Figures 4(C) and 5(C) present examples in which both the large and small EJP units are clearly

visible. These two excitatory inputs are known to be generated respectively by the two major motor axons that innervate these muscles (Jan and Jan, 1976a; Kurdyak et al., 1994). The smaller EJPs appeared, in some cases, less phasic than the larger ones and often preceded the large EJPs and lasted longer in rhythmic bursts. However, MK 801 equally affected the rhythmic bursts of the two types of EJPs, because the time

64

Cattaert and Birman

Table 1

Parameters of the Locomotor Rhythm Induced by Oxotremorine in an Isolated Nerve Cord Control

Oxotremorine after Section

Type 1

Type 2

Type 1

Type 2

Period

Duration

Period

Duration

Period

Duration

Period

Duration

16.46 ⫾ 0.8

11.65 ⫾ 0.69

5.56 ⫾ 0.12

3.16 ⫾ 0.07

12.24 ⫾ 1.12

9.19 ⫾ 0.80

6.87 ⫾ 0.25

2.37 ⫾ 0.11

10.21 ⫾ 0.28

8.52 ⫾ 0.74

12.55 ⫾ 0.25

2.65 ⫾ 0.08

8.53 ⫾ 0.54

6.61 ⫾ 0.31

15.50 ⫾ 0.92 p ⫽ .44 15.04 ⫾ 1.05 p ⫽ .07 13.12 ⴞ 0.86 p ⫽ .009

11.10 ⫾ 0.12 p ⫽ .44 11.04 ⫾ 0.92 p ⫽ .14 10.53 ⫾ 0.85 p ⫽ .088

13.87 ⴞ 2.82 p ⫽.0075 11.72 ⴞ 1.80 p ⫽ .014 12.60 ⫾ 2.56 p ⫽ .984 18.45 ⫾ 1.42

3.05 ⫾ 0.60 p ⫽ .857 1.80 ⴞ 0.06 p ⫽ .0002 2.42 ⫾ 0.08 p ⫽ .054 3.57 ⫾ 0.21

Burst period and burst duration were measured during rhythm type 1 and type 2 before (control) and after section of the nerve cord and reactivation of the rhythm by oxotremorine. Each line summarizes the data from a different experiment. Results represent the mean ⫾S.E.M. of rhythm parameters determined on a train of 12 bursts. The statistical significance of the differences between the control values and the values determined in the presence of oxotremorine is indicated in italic (p value determined by the independent t test). Significantly different results are in bold.

course of instantaneous frequency for the two units was similarly reduced by the drug [Fig. 4(B)]. Comparable effects were observed with PCP, another noncompetitive NMDA receptor antagonist. 0.01– 0.03 mg/mL PCP slowed down the rhythm frequency and reduced burst duration [Fig. 5(A,B)]. Here again, the effect on rhythm frequency was reversible, but the reduction in burst duration persisted during wash [Fig. 5(A)]. As for MK 801, PCP similarly affected the two types of EJPs that can be recorded in these larval muscles [Fig. 5(B,C)]. Excitatory inputs to other muscles were also affected the same way and concomitantly by these two antagonists (not shown). When a higher concentration of MK 801 was used (0.1 mg/mL and above), an increase in burst frequency was consistently observed during the first 100 s of perfusion, followed by a complete blockade of the locomotor rhythm [Fig. 6(A,B)]. Sometimes a residual tonic activity persisted for up to 30 min and then stopped [Fig. 6(A)]. In other cases, no residual tonic activity was observed [Fig. 6(B)]. A very similar blocking effect was observed in the presence of 0.1 mg/mL PCP. With this antagonist, an increase in burst frequency was transiently observed before the complete blockade [Fig. 6(C)]. Therefore, in the statistical analysis, we have distinguished the immediate effect, within 2 min of exposure to the drug [Fig. 7(A)] from the stable longterm effect (between 2–7 min of exposure) [Fig. 7(B– D)]. Relative effects, expressed as percentage of control values, were cumulated to build dose response curves for burst frequency [Fig. 7(A)], burst duration, inter-burst duration, and the percentage of silent time (defined as total inter-burst durations divided by total time) [Fig. 7(B–D)]. The transitory increase in burst

frequency was significant only at 0.1 mg/mL for both antagonists [Fig. 7(A)]. In contrast, the reduction in burst duration was almost maximal at 0.01 mg/mL [Fig. 7(B)], whereas the inter-burst duration was further increased at 0.03 mg/mL. These effects were masked by the total blockade of the rhythmic activity at higher concentrations of the antagonists. The most consistent dose-response curves were obtained with the percentage of silent time [Fig. 7(D)], with a halfmaximal effect occurring at about 0.01 mg/mL with both drugs. In contrast to these results, perfusion of the preparation with two competitive antagonists of vertebrate NMDA receptors, CPP and AP-5, did not evoke any significant change in spontaneous locomotor rhythm recorded in larval muscles (0.1 mg/mL CPP, n ⫽ 4; 0.1 mg/mL AP-5, n ⫽ 3; data not shown). Accordingly, we observed that these antagonists had no effect on larval crawling when administered in food.

Lack of Effect of MK 801 on Neuromuscular Transmission We first hypothesized that MK 801 could act directly at the glutamatergic NMJ. In order to check whether or not muscular glutamate receptors are affected by MK 801, glutamate was directly applied by pressure (see Methods) in the vicinity of the recorded muscle fiber during a spontaneous locomotor rhythm (Fig. 8). Application of glutamate (10⫺2 M, 20 ms, one bar) evoked a 20 mV depolarizing response [Fig. 8(B)]. When MK 801 was bath-applied at a high level (1 mg/mL), the rhythmic EJPs disappeared after 10 s of exposure [Fig. 8(B,C)]. The comparison of a control burst [Fig. 8(C), left trace] with a burst occurring in the presence of MK 801 [Fig. 8(C), right trace] indi-

Inhibition of Locomotor Rhythms in Drosophila

65

Figure 4 Effect of MK 801 on rhythmic activity. (A) Perfusion of the noncompetitive antagonist of NMDA receptors, MK 801, at a concentration of 0.032 mg/mL (95 ␮M) slows down rhythmic activity recorded in muscle fiber 6 and reduces burst duration. The rhythm is progressively restored during wash but burst duration remains shorter. The traces show the frequency (Burst Freq.) and duration (Burst Dur.) of rhythmic activity. The two arrows indicate the time of occurrence of two bursts shown in (B); (1) control burst, (2) burst recorded in the presence of MK801. (B) Effect on the two types of excitatory inputs to muscle 6. A control burst (1) is presented on the left, and a burst recorded in the presence of MK801 (2) is shown on the right. The intracellular recordings (top traces) display two types of EJPs, a small one and a large one, due to innervation of the muscle fiber by two different motor axons. By using two levels for discriminating these events, it was possible to draw the time course of instantaneous frequency for all units undifferentiated (middle diagrams) and for the large units only (bottom diagrams). Both units were affected similarly by MK 801. However, the large unit displayed a more stable activity and was therefore used to define bursts (bottom traces). (C) Enlarged views of the start [see small horizontal segments under M. Fiber recordings in (B)] of a control burst (left) and of a burst in the presence of MK 801 (right), illustrating the reliability of EJP unit discrimination. Top traces: intracellular recordings; middle traces: time course of instantaneous firing frequency of the two units; bottom traces: time course of instantaneous firing frequency of the large unit.

cates that, at this concentration, MK 801 induced a marked decrease in EJP instantaneous frequency before blocking all EJPs. At the same time, the glutamate receptor of the muscle fiber was not affected, because no significant change in the response to pres-

sure-ejected glutamate was observed in the presence of MK 801 [Fig. 8(B,D)]. In this experiment, the blocking effect of MK 801 was not reversed during wash (not shown) as was the case in all the experiments with concentrations of MK 801 higher than 0.1 mg/mL.

66

Cattaert and Birman

Figure 5 Effect of PCP on rhythmic activity. Same disposition as in Figure 4. (A) Perfusion of 0.01 mg/mL (⬇35 ␮M) phencyclidine (PCP), another noncompetitive antagonist of NDMA receptors, has effects on rhythmic activity that are similar to MK 801. Activity recorded in muscle fiber 6 is rapidly slowed down and burst duration is reduced. The original rhythmic activity is restored after wash except for burst duration, which remains shortened. (B,C) PCP similarly affected the large and small EJP units. Left, control burst (1). Right, burst recorded in the presence of PCP (2). Same disposition as in Figure 4 (B,C).

In other experiments, an intracellular electrode was implanted into a muscle fiber and a suction electrode was placed on the corresponding motor nerve. Electrical stimulation (0.5 ms, 5 V recruiting the two motoneurons innervating muscle fibers 6 and 7) was used to evoke an EJP in the muscle fiber. In the presence of MK 801 (up to 5 mg/mL), no change in the evoked EJP was observed, even after long time exposure (⬎1 h) to the drug (not shown). However, during perfusion of high concentration of MK 801 (1–5 mg/ml), we observed some decrease in nerve fiber excitability, such that an EJP was not elicited by every stimulation. Overall, these results indicate that concentrations of MK 801 below 0.5 mg/mL have no major inhibitory effect on motor nerve conduction or neuromuscular transmission at

the Drosophila NMJ. Therefore, we concluded that MK 801 acts at a central site to block the locomotor rhythm generator.

Effect of MK 801 on the Rhythm Generated by Oxo Because the muscarinic agonist oxo is able to restore a locomotor rhythm in a sectioned ventral nerve cord preparation (Fig. 3), we hypothesized that muscarinic receptors could be involved in the activation of the locomotor generator in the ventral nerve cord. We tested to see if the noncompetitive NMDA antagonist MK 801 could still block this activity in the presence of oxo (10⫺5 M) in sectioned (data not shown) or in intact nerve cord [Fig. 9(A)]. Oxo was continuously

Inhibition of Locomotor Rhythms in Drosophila

Figure 6 Blockade of rhythmic activity by MK 801 and PCP. (A,B) Perfusion of MK 801 at a higher level (0.1 mg/mL) blocks rhythmic activity. The top traces represent the instantaneous frequency (Inst. Freq.) of EJPs intracellularly recorded from a muscle fiber. The traces below show the frequency of rhythmic activity (Burst Freq.). In the presence of MK 801, the rhythm is transiently accelerated and then stops. A residual tonic activity reappeared after a few seconds of drug application in some (A) but not all (B) preparations. (C) A very similar blockade is observed in the presence of 0.1 mg/mL PCP, also preceded by a transitory increase in rhythm frequency.

67

68

Cattaert and Birman

Figure 7 Statistical analysis of the effects of MK 801 and PCP on motor rhythm parameters. Results were accumulated from all experiments and represented as percentage of the corresponding control values. Each bar is the mean ⫾S.E.M. of five independent experiments. Statistical differences between the different drug concentrations were assessed using ANOVA followed by Newman-Keuls multiple comparison tests. Parameters that are significantly different from control conditions are indicated by an asterisk (*); ns ⫽ not significant. (A) Dose response curves of the transitory effect of MK 801 and PCP (measured during the first 2 min of drug application) on rhythm frequency. ANOVA results: p ⫽ .0032 for MK 801 and p ⫽ .0089 for PCP. (B,C,D) Dose response curves of the stabilized effects of the antagonists (measured 2 min after drug application and for 5 min) on burst duration (B), inter-burst duration (C), and the percentage of silent time (D). ANOVA results: (B) p ⫽ .001 (MK 801), p ⬍ .0001 (PCP); (C) p ⫽ .0049 (MK 801), p ⫽ .0007 (PCP); (D) p ⬍ .0001 (MK 801), p ⬍ .0001 (PCP).

perfused during the experiment. A comparison of the burst activity before and after exposition to 0.1 mg/mL MK 801 is presented in Figure 9(B). Three observations can be made: first, MK 801 is still able to dramatically and permanently affect the locomotor activity, because the 30 s period rhythm (1) disappears progressively in 3 min in the presence of MK 801; second, MK 801, however, does not totally block all the activity, and a fast rhythm is conserved (2); third, this fast rhythm is, however, different from the control and appears as a modulation of a continuous discharge, with no abrupt changes in the instantaneous firing frequency, which never exceeds 5 Hz (whereas it was ⬎10 Hz in control). Similar results were obtained when the nerve cord was sectioned (n ⫽ 3) or not (n ⫽ 2). These results show that MK 801 inhibits the locomotor rhythm generator located in the ventral nerve cord.

DISCUSSION Origin of the Locomotor Rhythm in Drosophila Larvae In the present work, we used an in vitro Drosophila larva preparation, which proved to be a very convenient system to study the neuronal circuitry involved in patterning locomotor rhythm. One of the first questions we asked is where are these specific neuronal centers located. In arthropods, motor activity is controlled by CPGs, which can be activated by cholinergic muscarinic agonists, as was demonstrated for CPGs controlling the pharyngeal muscles of Drosophila larvae (Gorczyca et al., 1991), locomotion in crayfish (Chrachri and Clarac, 1990), locust (Ryckebusch and Laurent, 1993, 1994; Baudoux et al., 1998), and stick insects (Bu¨schges et al., 1995), and fictive loco-

Inhibition of Locomotor Rhythms in Drosophila

69

Figure 8 Lack of effect of MK 801 on the postsynaptic glutamate receptor at the neuromuscular junction. (A) Intracellular recording from a muscle fiber in a rhythmic preparation. (B) During the experiment, the response of the postsynaptic glutamate receptor was tested by ejecting L-glutamate (10 mM) through a pressure ejection micropipette in the vicinity of the recorded muscle fiber. A high concentration of MK 801 (1 mg/mL) was perfused after the end of a glutamate-evoked response during a rhythmic burst. Although MK 801 suppresses all EJP activity, the muscle glutamate receptor does not seem to be affected, as indicated by the lack of change in the successive responses to pressure-ejected glutamate. (C) Enlargements of two details from the graph in (B), showing the rapid blockade of the burst in the presence of MK 801. (D) Comparison of the glutamate-evoked responses before and during application of 1 mg/mL MK 801. Each bar is the mean ⫾S.E.M. of five responses. No significant (ns) difference (unpaired Student’s t test; p ⫽ .195) exists between glutamate-evoked responses obtained in the two conditions.

motion in isolated larval nerve cords of the tobacco hawkmoth Manduca sexta (Johnston and Levine, 1996). In our experiments, when the ventral nerve cord is abruptly isolated from the brain hemispheres, the endogenous rhythm immediately stops, and never spontaneously recovers, indicating that neural networks located in these anterior regions of the nervous system are required for locomotor activity to be expressed. In agreement with this result, a reduced larval crawling behavior was observed in several Drosophila mutants characterized by morphological defects in the central complex, a prominent structure of the adult brain (Varnam et al., 1996). Likewise, in the ventral ganglia of Manduca larva, the rhythmic recruitment of the unpaired median neurons during fictive crawling depends, at least in part, on a source located

anteriorly in the subesophageal ganglion (Johnston et al., 1999). However, we also show that, on a silent ventral nerve cord completely isolated from the brain, a muscarinic agonist (oxo) is able to reactivate a clear, although more irregular, locomotor rhythm (Fig. 3). Similarly, Johnston et al. (1999) have shown that, after removal of the subesophageal ganglion from isolated Manduca nerve cord, the fictive locomotor rhythm generated by a muscarinic agonist persists in motor nerves, but presents a somewhat irregular and disrupted pattern. Thus, in Drosophila as in Manduca, the CPGs that control larval crawling can be directly stimulated by muscarinic agonists, are located in the abdominal and thoracic ganglia of the ventral nerve cord, and are

70

Cattaert and Birman

Figure 9 MK 801 blocks the rhythmic activity evoked by oxotremorine. (A) Evolution with time of the instantaneous frequency of EJPs in the presence of 10 ␮M oxotremorine in a nonsectioned nerve cord. In the presence of 0.1 mg/mL MK 801 in the perfusion medium, the rhythmic activity (period of 0.5 s) disappears, and is replaced by a continuously modulated tonic discharge. (B) Comparison of the EJP discharges before and after MK 801 application. The instantaneous frequency (top traces) reaches 10 Hz in a control burst (1), whereas in the presence of MK801 (2) the continuous EJP discharge oscillates rhythmically between 2 and 5 Hz. In each situation, the corresponding intracellular recording from the muscle fiber is presented in the bottom trace. The time of occurrence of the two enlarged views (1) and (2) are marked in (A).

controlled by descending regulatory pathways originating from more anterior centers.

Noncompetitive NMDA Antagonists Inhibit the Locomotor Rhythm We report here that the noncompetitive NMDA antagonists (MK 801 and PCP) decrease burst duration and ultimately block the locomotor rhythm in Drosophila larvae. All the excitatory inputs to larval body wall muscles were similarly and concomitantly affected by these antagonists, as shown for the two types of EJPs recorded in muscle 6 (Figs. 4, 5), while neuromuscular transmission was still active (Fig. 8). This suggests that MK 801 and PCP act at a central site generating rhythmic motor activity pattern. We

observed that the presence of oxo did not prevent the blocking effect of MK 801 on the rhythmic bursts, indicating that MK 801 inhibits CPGs located in the ventral cord. The fact that faster residual rhythm appears resistant to MK 801 (Fig. 9) could be due to a direct effect of oxo on the motoneurons. These results suggest the participation of NMDA or NMDA-like receptors in the control of locomotion in Drosophila. This hypothesis is, however, conditioned by the specificity of the noncompetitive blockers we used, keeping in mind that pharmacology is frequently different between vertebrates and invertebrates. MK 801 is a potent and selective noncompetitive NMDA receptor antagonist. PCP is a hallucinogenic drug in humans and is both a NMDA receptor antagonist and a ligand of the sigma receptors in

Inhibition of Locomotor Rhythms in Drosophila

vertebrates. The sigma receptors are orphan receptors, which appear to function as modulators of the activity of the NMDA receptors (Okuyama et al., 1996; Gronier and Debonnel, 1999). Therefore, the only common target of MK 801 and PCP known is the vertebrate NMDA receptor. Blockade or increase in burst frequency was generally resistant to wash at higher doses. This is consistent with previous observations that MK 801 and PCP are open-channel blockers that leave the channel very slowly (half-time equal to more than an hour) (Lipton, 1993). The lack of effect of the two NMDA competitive antagonists, CPP and AP-5, on motor rhythm could be due either to limited diffusion of these inhibitors in the nervous ganglia as compared to the much more hydrophobic noncompetitive antagonists, or to a different pharmacology of the Drosophila NMDA-like receptor.

Possible Role of an NMDA Receptor Homologue in Drosophila In vertebrates, NMDA receptors have been demonstrated to play a major role in the rhythmic activity in lamprey (Brodin and Grillner, 1986), Xenopus embryo (Dale and Roberts, 1984), chick (Barry and O’Donovan, 1987), and rat (Cazalets et al., 1992). Our results suggest that the akinesia displayed by noncompetitive NMDA antagonist-fed Drosophila larvae most likely results from the specific effect of these inhibitors on the CPGs. However, in the crustacean stomatogastric system, it has been demonstrated that rhythm production is based on potassium pacemaker potentials that progressively depolarize the neurons, plateau properties supported by noninactivating sodium channels, and calcium-dependent potassium channels that repolarize the membrane after each burst of activity (Buchholtz et al., 1992; Golowasch et al., 1992; Elson and Selverston, 1997). Therefore, our results are new and raise the possibility that the invertebrate counterpart of the NMDA receptors participates in the control of locomotion in Drosophila, as is the case in vertebrates. It is noted that MK 801 and PCP both cause tremor and ataxia at toxic doses in mammalians (Nakki et al., 1996), an effect strikingly similar to that displayed by living Drosophila larvae exposed to these antagonists. Several lines of evidence suggest that NMDA receptors are expressed in invertebrates. A functional NMDA receptor has been described in the crayfish optic lobe, which shares several features with those in vertebrates, including agonist specificity, inhibition by Mg2⫹, and glycine dependence (Pfeiffer-Linn and Glantz, 1991). Similarly, NMDA-like presynaptic au-

71

toreceptors have been recently identified at the crayfish NMJ (Feinstein et al., 1998). Homologues of the vertebrate NMDA-R1 (Ultsch et al., 1993) and NMDA-R2 subunits (Pellicena-Palle and Salz, 1995; Volkner et al., 2000) have been cloned in Drosophila, which are expressed in the CNS, but the function of these proteins in flies is still unknown. The membrane-spanning domains of the Drosophila and vertebrate NMDA-R1 subunits show a very high level of homology (Ultsch et al., 1993), suggesting that the channel-associated MK 801 and PCP binding sites have been conserved in the invertebrate receptor. Further work is needed to ascertain the physiological role of the Drosophila NMDA receptor homologue and its possible involvement in motor rhythm control. We thank Marie-The´re`se Besson and Be´atrice Reiniche for their help in the early stages of this work, and we are grateful to Franc¸ois Clarac and Pierre Meyrand for providing laboratory facilities.

REFERENCES Barry MJ, O’Donovan MJ. 1987. The effects of excitatory amino acids and their antagonists on the generation of motor activity in the isolated chick spinal cord. Brain Res 433:271–276. Bate M, Landgraf M, Ruiz Gomez Bate M. 1999. Development of larval body wall muscles. Int Rev Neurobiol 43:25– 44. Baudoux S, Duch C, Morris OT. 1998. Coupling of efferent neuromodulatory neurons to rhythmical leg motor activity in the locust. J Neurophysiol 79:361–370. Berkowitz A, Laurent G. 1996. Central generation of grooming motor patterns and interlimb coordination in locusts. J Neurosci 16:8079 – 8091. Broadie K, Bate M. 1993. Activity-dependent development of the neuromuscular synapse during Drosophila embryogenesis. Neuron 11:607– 619. Brodin L, Grillner S. 1986. Effects of magnesium on fictive locomotion induced by activation of N- methyl-D-aspartate (NMDA) receptors in the lamprey spinal cord in vitro. Brain Res 380:244 –252. Buchholtz F, Golowasch J, Epstein IR, Marder E. 1992. Mathematical model of an identified stomatogastric ganglion neuron. J Neurophysiol 67:332–340. Bu¨ schges A, Schmitz J, Ba¨ ssler U. 1995. Rhythmic patterns in the thoracic nerve cord of the stick insect induced by pilocarpine. J Exp Biol 198:435– 456. Cazalets JR, Sqalli-Houssaini Y, Clarac F. 1992. Activation of the central pattern generators for locomotion by serotonin and excitatory amino acids in neonatal rat. J Physiol (Lond) 455:187–204. Chrachri A, Clarac F. 1990. Fictive locomotion in the fourth

72

Cattaert and Birman

thoracic ganglion of the crayfish, Procambarus clarkii. J Neurosci 10:707–719. Crossley C. 1978. The morphology and development of the Drosophila muscular system. In: Ashburner M, Wright T, editors. The Genetics and Biology of Drosophila. New York: Academic Press, p 449 –559. Dale N, Roberts A. 1984. Excitatory amino acid receptors in Xenopus embryo spinal cord and their role in the activation of swimming. J Physiol (Lond) 348:527–543. Delgado R, Barla R, Latorre R, Labarca P. 1989. L-glutamate activates excitatory and inhibitory channels in Drosophila larval muscle. FEBS Lett 243:337–342. Dicaprio R. 1997. Plateau potentials in motor neurons in the ventilatory system of the crab. J Exp Biol 200:1725– 1736. Elson RC, Selverston AI. 1997. Evidence for a persistent Na⫹ conductance in neurons of the gastric mill rhythm generator of spiny lobsters. J Exp Biol 200:1795–1807. Engel JE, Wu CF. 1998. Genetic dissection of functional contributions of specific potassium channel subunits in habituation of an escape circuit in Drosophila. J Neurosci 18:2254 –2267. Feinstein N, Parnas D, Parnas H, Dudel J, Parnas I. 1998. Functional and immunocytochemical identification of glutamate autoreceptors of an NMDA type in crayfish neuromuscular junction. J Neurophysiol 80:2893–2899. Gerschenfeld HM. 1973. Chemical transmission in invertebrate central nervous systems and neuromuscular junctions. Physiol Rev 53:1–119. Golowasch J, Buchholtz F, Epstein IR, Marder E. 1992. Contribution of individual ionic currents to activity of a model stomatogastric ganglion neuron. J Neurophysiol 67:341–349. Gorczyca MG, Budnik V, White K, Wu CF. 1991. Dual muscarinic and nicotinic action on a motor program in Drosophila. J Neurobiol 22:391– 404. Greenspan RJ. 1997. A kinder, gentler genetic analysis of behavior: dissection gives way to modulation. Curr Opin Neurobiol 7:805– 811. Gronier B, Debonnel G. 1999. Involvement of sigma receptors in the modulation of the glutamatergic/NMDA neurotransmission in the dopaminergic systems. Eur J Pharmacol 368:183–196. Heitler WJ, Pearson KG. 1980. Non-spiking interactions and local interneurones in the central pattern generator of the crayfish swimmeret system. Brain Res 187:206 –211. Helfrich C. 1986. Role of the optic lobes in the regulation of the locomotor activity rhythm of Drosophila melanogaster: behavioral analysis of neural mutants. J Neurogenet 3:321–343. Hoshino M, Matsuzaki F, Nabeshima Y, Hama C. 1993. hikaru genki, a CNS-specific gene identified by abnormal locomotion in Drosophila, encodes a novel type of protein. Neuron 10:395– 407. Jan LY, Jan YN. 1976a. Properties of the larval neuromuscular junction in Drosophila melanogaster. J Physiol (Lond) 262:189 –214. Jan LY, Jan YN. 1976b. L-glutamate as an excitatory trans-

mitter at the Drosophila larval neuromuscular junction. J Physiol (Lond) 262:215–236. Jan LY, Jan YN. 1997. Voltage-gated and inwardly rectifying potassium channels. J Physiol (Lond) 505:267–282. Johansen J, Halpern ME, Johansen KM, Keshishian H. 1989. Stereotypic morphology of glutamatergic synapses on identified muscle cells of Drosophila larvae. J Neurosci 9:710 –725. Johnston RM, Consoulas C, Pfluger H, Levine RB. 1999. Patterned activation of unpaired median neurons during fictive crawling in Manduca sexta larvae. J Exp Biol 202:103–113. Johnston RM, Levine RB. 1996. Crawling motor patterns induced by pilocarpine in isolated larval nerve cords of Manduca sexta. J Neurophysiol 76:3178 –3195. Keshishian H, Chang TN, Jarecki J. 1994. Precision and plasticity during Drosophila neuromuscular development. FASEB J 8:731–737. Keshishian H, Chiba A, Chang TN, Halfon MS, Harkins EW, Jarecki J, Wang L, Anderson M, Cash S, Halpern ME, Johansen J. 1993. Cellular mechanisms governing synaptic development in Drosophila melanogaster. J Neurobiol 24:757–787. Kidokoro Y, Nishikawa K. 1994. Miniature endplate currents at the newly formed neuromuscular junction in Drosophila embryos and larvae. Neurosci Res 19:143– 154. Kurdyak P, Atwood HL, Stewart BA, Wu CF. 1994. Differential physiology and morphology of motor axons to ventral longitudinal muscles in larval Drosophila. J Comp Neurol 350:463– 472. Landgraf M, Bossing T, Technau GM, Bate M. 1997. The origin, location, and projections of the embryonic abdominal motorneurons of Drosophila. J Neurosci 17:9642– 9655. Lipton SA. 1993. Prospects for clinically tolerated NMDA antagonists: open-channel blockers and alternative redox states of nitric oxide. Trends Neurosci 16:527–532. Martin JR, Ernst R, Heisenberg M. 1998. Mushroom bodies suppress locomotor activity in Drosophila melanogaster. Learn Mem 5:179 –191. McLarnon JG, Quastel DM. 1988. A voltage clamp study of the glutamate responsive neuromuscular junction in Drosophila melanogaster. Can J Physiol Pharmacol 66:321– 327. Nakki R, Sharp FR, Sagar SM. 1996. FOS expression in the brainstem and cerebellum following phencyclidine and MK801. J Neurosci Res 43:203–212. Na¨ssel DR. 1996. Neuropeptides, amines and amino acids in an elementary insect ganglion: functional and chemical anatomy of the unfused abdominal ganglion. Prog Neurobiol 48:325– 420. Neckameyer WS. 1996. Multiple roles for dopamine in Drosophila development. Dev Biol 176:209 –219. Okuyama S, Imagawa Y, Tomisawa K. 1996. Behavioral evidence for modulation by sigma ligands of (⫹)MK801- induced hyperlocomotion in monoamine-depleted mice. Neuropharmacology 35:467– 474.

Inhibition of Locomotor Rhythms in Drosophila Pellicena-Palle A, Salz HK. 1995. The putative Drosophila NMDARA1 gene is located on the second chromosome and is ubiquitously expressed in embryogenesis. Biochim Biophys Acta 1261:301–303. Pfeiffer-Linn C, Glantz RM. 1991. An arthropod NMDA receptor. Synapse 9:35– 42. Reynolds IJ, Miller RJ. 1990. Allosteric modulation of N-methyl-D-aspartate receptors. Adv Pharmacol 21:101– 126. Ryckebusch S, Laurent G. 1993. Rhythmic patterns evoked in locust leg motor neurons by the muscarinic agonist pilocarpine. J Neurophysiol 69:1583–1595. Ryckebusch S, Laurent G. 1994. Interactions between segmental leg central pattern generators during fictive rhythms in the locust. J Neurophysiol 72:2771–2785. Sink H, Whitington PM. 1991. Location and connectivy of abdominal motoneurons in the embryo and larva of Drosophila melanogaster. J Neurosci 22:298 –311. Sokolowski MB, Hansell KP. 1992. The foraging locus: behavioral tests for normal muscle movement in rover and sitter Drosophila melanogaster larvae. Genetica 85: 205–209. Stewart BA, Atwood HL, Renger JJ, Wang J, Wu CF. 1994. Improved stability of Drosophila larval neuromuscular preparations in haemolymph-like physiological solutions. J Comp Physiol [A] 175:179 –191.

73

Strauss R, Heisenberg M. 1993. A higher control center of locomotor behavior in the Drosophila brain. J Neurosci 13:1852–1861. Ultsch A, Schuster CM, Laube B, Betz H, Schmitt B. 1993. Glutamate receptors of Drosophila melanogaster. Primary structure of a putative NMDA receptor protein expressed in the head of the adult fly. FEBS Lett 324: 171–177. Van Vactor D, Sink H, Fambrough D, Tsoo R, Goodman CS. 1993. Genes that control neuromuscular specificity in Drosophila. Cell 73:1137–1153. Varnam C, Strauss R, De Belle J, Sokolowski M. 1996. Larval behavior of Drosophila central complex mutants: interactions between no bridge, foraging and chaser. J Neurogenet 11:99 –115. Volkner M, Lenz-Bohme B, Betz H, Schmitt B. 2000. Novel CNS glutamate receptor subunit genes of Drosophila melanogaster. J Neurochem 75:1791–1799. Wilkens JL, DiCaprio RA. 1994. Effects of scaphognathite nerve stimulation on the acutely deafferented crab ventilatory central pattern generator. J Comp Physiol [A] 174:195–209. Wolf R, Heisenberg M. 1990. Visual control of straight flight in Drosophila melanogaster. J Comp Physiol [A] 167:269 –283.