peripheral distribution of presynaptic sites of abdominal motor and modulatory neurons inmanduca...

Peripheral Distribution of PresynapticSites of Abdominal Motorand Modulatory Neuronsin Manduca sexta Larvae

CHRISTOS CONSOULAS,1* REBECCA M. JOHNSTON,2 HANS-JOACHIM PFLUGER,3

AND RICHARD B. LEVINE1

1Division of Neurobiology, University of Arizona, Tucson, Arizona 857212Department of Integrative Biology, University of California, Berkley, California 94720

3Freie Universitat Berlin, Institut fur Neurobiologie, D-14195 Berlin, Germany

ABSTRACTInsect muscle fibers are commonly innervated by multiple motor neurons and efferent

unpaired median (UM) neurons. The role of UM neurons in the modulation rather than rapidactivation of muscle contraction (Evans and O’Shea [1977] Nature 270:257–259) suggests thattheir terminal varicosities may differ structurally and functionally from the presynapticterminals of motor neurons. Furthermore, differences in the characteristics of UM neuron terminalvaricosities may be correlated with functional differences among their diverse target muscles.

Larval abdominal body wall muscles in the hawkmoth, Manduca sexta, consist of large,elongated fibers that are multiterminally innervated by one and occasionally two motorneurons (Levine and Truman [1985] J. Neurosci. 5:2424–2431). The fibers are also innervatedby one of two efferent UM neurons that bifurcate to innervate targets on both sides of theabdomen (Pfluger et al. [1993] J. Comp. Neurol. 335:508–522).

In this study, the intracellular tracer biocytin was used to identify the targets of the UMneurons and to distinguish their terminal axonal varicosities on the muscle fibers. Anantiserum to the synaptic vesicle protein, synaptotagmin, was used to label synaptic vesicles,and the styryl dye FM1–43 was used to demonstrate release and recycling. Most of theabdominal muscles in a given hemisegment were found to be supplied by one of the two UMneurons. Terminal varicosities of the excitatory motor neurons were large (3–7 µm) and werefound in rows of rosettes that extended to every aspect of the muscle fiber; these varicositieswere designated as type I terminals. The UM neuron terminal varicosities also occupied everyaspect of the fiber but were smaller (1–3 µm) and more separated from each other; these weredesignated as type II terminals. Both type I and type II terminals are synaptotagminimmunoreactive and, as shown by FM1–43 staining, are sites of synaptic vesicle recycling.The excitatory motor neuron terminals (type I) could easily be loaded and unloaded withFM1–43, which indicates their capacity for repeated vesicular exocytosis and recycling. Incontrast, the dye could not as readily be unloaded from UM neuron terminals (type II), whichmay indicate that they have a slower turnover of synaptic vesicles. J. Comp. Neurol 410:4–19,1999. r 1999 Wiley-Liss, Inc.

Indexing terms: insect; neuromuscular junction; motor terminal; octopamine; synaptotagmin;

synaptic; vesicle recycling

Insect skeletal muscle fibers often have polyneuronalinnervation. Furthermore, individual neurons make mul-tiple contacts with the muscle fibers through specializedperiodic axonal enlargements (varicosities), which are thelocations of neuromuscular synapses (Hoyle, 1983; Atwoodet al., 1993; Jia et al., 1993). The large number and widedistribution of varicosities along each muscle fiber oftenmake it difficult to determine which axon terminals belong

Grant sponsor: NIH; Grant number: NS 24822; Grant sponsor: NSF;Grant number: INT 9726330; Grant sponsor: Fogarty International Center;Grant number: TWO 4898.

*Correspondence to: Dr. Christos Consoulas, Division of Neurobiology,Room 611, Gould Simpson Building, University of Arizona, Tucson, AZ85721. E-mail: [email protected]

Received 1 December 1998; Revised 22 February 1999; Accepted 11March 1999

THE JOURNAL OF COMPARATIVE NEUROLOGY 410:4–19 (1999)

r 1999 WILEY-LISS, INC.

to a given neuron. This difficulty is compounded by thediversity of motor neuron and modulatory neuron typesthat project to insect muscle fibers. Thus, the terminalvaricosities of different neurons may be specialized forunique effects on muscle fibers. Furthermore, the morphol-ogy, spatial distribution, and functional characteristics ofterminal varicosities may differ in relation to the special-ized roles of different muscle groups within and amonganimals.

Axon terminal varicosities have been examined morpho-logically and physiologically in the abdominal muscles ofDrosophila melanogaster, where they have been catego-rized according to their size as type I or type II (Johansenet al., 1989; Budnik et al., 1990; Atwood et al., 1993; Jia etal., 1993). Type I varicosities, which belong to excitatorymotor neurons, are large, occupy a central area of themuscle fiber, and have ultrastructural and functionalcharacteristics (Kurdyak et al., 1994) that are similar tothose observed in phasic and tonic motor axons of crusta-ceans (Atwood and Wojtowicz, 1986). In most if not allcases, the excitatory neurotransmitter is glutamate (Janand Jan, 1976; Johansen et al., 1989; Broadie and Bate,1993) but may be colocalized with other putative neuro-transmitters or neuromodulators (Anderson et al., 1988;Cantera and Nassel, 1992; Gorczyca et al., 1993). Type IIvaricosities are smaller, span a much larger area along thelength of the muscle fibers, and are labeled by an antise-rum to the biogenic amine octopamine (Monastirioti et al.,1995).

Octopamine has been proposed to act as a neurotransmit-ter, a neuromodulator, and a neurohormone in inverte-brates (for reviews, see Orchard, 1982; Evans, 1985;Agricola et al., 1988). In several insect species, efferentoctopaminergic neurons have unpaired median (UM) so-mata and bilaterally symmetrical axons, which projectinto peripheral nerves to innervate skeletal and visceralmuscles, neurohemal organs, peripheral proprioceptors,and perhaps other targets (Hoyle et al., 1974; Evans andO’Shea, 1977; Hoyle, 1978; Orchard and Lange, 1985;Whim and Evans, 1988; Braunig et al., 1994; Stevensonand Sporhase-Eichmann, 1995; Braunig, 1997; Braunigand Eder, 1998). The peripheral branching pattern of thethree dorsal unpaired median (DUM) neurons that supplymuscles and peripheral nerves has been described in thelocust (Braunig et al., 1994; Braunig, 1997). These neuronsbranch extensively over several different muscles to formvaricose terminal processes. Whether these terminal vari-cosities represent sites of neurotransmitter or neuromodu-

lator release remains unknown. At the ultrastracturallevel, dendritic neuronal processes in the central nervoussystem and peripheral neuromuscular terminals of theUM neurons appear to contain large dense-core vesicles,together with less numerous, small clear vesicles. How-ever, dense-core vesicles, which are thought to containoctopamine, were never found in typical presynaptic spe-cializations, suggesting that octopamine may be releasednonsynaptically (Oertal et al., 1975; Hoyle et al., 1980;Kiss et al., 1984; Watson, 1984; Atwood et al., 1993;Rheuben, 1995; Pfluger and Watson, 1995).

Octopamine has been detected biochemically in lepidop-teran larval body wall muscles (Davenport and Wright,1986), and activation of a pair of efferent unpaired medianneurons in abdominal ganglia of Antherae pernyi larvaemodulates both muscle tension and the response to excita-tory motor neurons (Brookes, 1988; Brookes and Weevers,1988). Similarly, Pfluger et al. (1993) identified two effer-ent UM neurons in each abdominal ganglion of the hawk-moth Manduca sexta but did not identify their targets. Ourgoals in the present study were to identify the peripheraltargets of these two efferent UM neurons and to distin-guish between their terminals and excitatory motor termi-nals on muscle fibers that are innervated by both. Wedemonstrate that most of the muscles in the abdominalsegments of M. sexta larvae are supplied by one or theother of the two octopaminergic efferent UM neurons. Wealso show that the UM neuron terminals can be distin-guished from the terminal of the excitatory motor neuronson the basis of their size and distribution on the muscles.Finally, we demonstrate that both types of varicosities aresynaptotagmin immunopositive and are capable of synap-tic vesicle exocytosis and recycling.

MATERIALS AND METHODS

Third and fourth instar M. sexta (L.) larvae were ob-tained from a laboratory culture reared on an artificial diet(Bell and Joachim, 1976), under a long-day photoperiodregimen (17 hours light, 7 hours dark), and at 26°C andapproximately 60% relative humidity.

Biocytin staining

Biocytin (Sigma, St. Louis, MO) was used to demon-strate the peripheral branching of abdominal neurons(Horikawa and Armstrong, 1988; Consoulas et al., 1996).To fill the peripheral axons of the motor neurons, the

Abbreviations

A1, A2, A3 first, second, third abdominal segmentsAG1, AG2 first, second abdominal gangliaAM alary musclesATP anterior tergopleural muscleDEM dorsal external mediamDEO dorsal external oblique muscleDIL dorsal internal longitudinal muscleDIM dorsal internal median muscleDIO dorsal internal oblique muscleDN dorsal nerveDNa dorsal nerve, anterior branchDNl dorsal nerve, lateral branchDNp dorsal nerve, posterior branchEJP excitatory junction potentialLIO lateral internal oblique muscleLN link nerveML midline

MN median nerveP pleural musclePrtFlx pretarsal flexor musclePTP posterior tergopleural muscleSp spiracleSPM spiracular muscleTN transverse nerveTP tergopleural muscleUM unpaired mediam neuronUMa UM neuron with axon in DNaUMp/l UM neuron with axon in DNp and DNlVEM ventral external median muscleVN ventral nerveVEO ventral external oblique muscleVIL ventral internal logitudinal muscleVIM ventral internal median muscleVIO ventral internal oblique muscle

MOTOR AND MODULATORY PERIPHERAL RELEASE SITES 5

animals were first anesthetized by chilling on ice. Afterdecapitation, the animals were opened along the dorsalmidline and pinned down in saline (in mM): NaCl, 140;KCl, 5; CaCl2, 4; glucose, 28; HEPES, 5; final pH 7.4(Trimmer and Weeks, 1989) in a Sylgard-coated Petri dish.The cut ends of specific nerves or connectives (see Results)were isolated in a Vaseline pool to allow the infusion of abiocytin solution (3% w/v biocytin in distilled water). Thepreparations were stored at 7°C for a maximum of 2 days. Inother cases, the cell body of an abdominal UM neuron wasimpaled with a glass microelectrode (30–50-MV resistance)filled with 3% biocytin and 2 M potassium acetate. Depolariz-ing current pulses of 5–10-nA amplitude and 500-msec dura-tion were injected into the cell at a rate 1 Hz for 20 minutes to 1hour. To allow staining of the UM cell axons in the periphery,the preparations were left first under constant saline superfu-sion for 60 minutes and then kept at 7°C overnight.

After biocytin infusion, the preparations were dissectedand fixed overnight in a freshly prepared fixative solutionconsisting of 4% paraformaldehyde, 0.15% glutaralde-hyde, and 0.2% saturated picric acid in 0.1 M phosphatebuffer (pH 7.4) overnight (Sun et al., 1993). The tissueswere subsequently dehydrated in ethanol, permeabilizedin xylene or in propylene, rehydrated in ethanol, andwashed in 10 mM phosphate buffered saline (PBS; pH 7.4)for 3 3 15 minutes and in PBS containing 1% Triton X-100(PBSX) for 3 3 15 minutes. To reduce background staining,the preparations were incubated in 10% normal goatserum (NGS; Jackson ImmunoResearch Laboratories, WestGrove, PA) and 3% bovine serum albumin (BoehringerMannheim Biochemicals, Indianapolis, IN) in PBSX for 1hour. They were then incubated in 1:1,000 (w/v) Cy3-conjugated streptavidin (Jackson ImmunoResearch) inPBSX for 12 hours at 7°C. The preparations were thenwashed several times with PBS, dehydrated in ethanol,and cleared in methyl salicylate. To examine the morphol-ogy of the neuron terminals in relation to the target tissue,the neurons were filled with biocytin and processed asdescribed above; the labeling was demonstrated by incubat-ing the preparations in Cy3-conjugated streptavidin, andthe muscles were stained simultaneously by using 66 nMOregon-green Phalloidin (Molecular Probes, Eugene, OR)in PBSX to show filamentous actin.

The stained preparations were subsequently viewedwith the confocal microscope. By using the two shamblingchannels and dichromatic cubes (K1 and K2, Bio-Rad,Cambridge, MA; excitation wave lengths of 488 and 568nm), optical sections were simultaneously recorded throughthe depth of wholemount preparations for the two dyesused. The two digitized images then were merged by usingdifferent pseudocolors (red for Cy3-conjugated streptavi-din and green for Oregon-green Phalloidin). Images wereprepared by using Confocal Assistant (Bio-Rad) and Corel6 (Corel Corp., Ottawa, ON, Canada) and printed on aTectronix dye-sublimation printer.

Synaptotagmin immunostaining

To examine the type of presynaptic varicosities thatmotor and neurosecretory neurons make over the bodywall muscles, an antiserum raised against D. melanogas-ter synaptotagmin was used (DSYT2; Littleton et al., 1993;generously provided by J.T. Littleton and H.J. Bellen).Results were confirmed with an antiserum raised againstM. sexta synaptotagmin (generously provided by S.H.

Dubuque and L.P. Tolbert). The muscles were fixed in 4%paraformaldehyde for 1 hour at room temperature. Afterrinsing in PBSX for 2 hours, the preparations were blockedfor 1 hour in 10% NGS and incubated overnight in primaryantiserum (1:1,000) made up in PBSX (pH 7.4) with 1%NGS. After washing in PBSX and PBS for 2 hours, thetissues were incubated in Cy3-conjugated secondary anti-serum for 2–3 hours at room temperature. The prepara-tions were then rinsed in PBSX and PBS and cleared andstored in 80% glycerol.

FM1–43 staining

The fluorescent dye FM1–43 (Molecular Probes, Inc.)was used to monitor synaptic vesicle exocytosis and recy-cling (Betz and Bewick 1992; Betz et al., 1992; Ramaswamiet al., 1994). Abdominal or connective nerves were stimu-lated via a saline-filled suction electrode with a Grass S88stimulator (Grass Instruments, Quincy, MA) by using1–5-Hz pulses for 5 minutes in the presence of 4 µMFM1–43 in saline. To confirm that axons were stimulated,muscle contractions were observed or excitatory junctionalpotentials (EJPs) were recorded during the course ofstimulation with a microelectrode filled with 2 M potas-sium acetate. The preparation was then washed 2 3 30seconds and 4 3 5 minutes in Ca21-free saline (CaCl2 wasreplaced by 4 mM MgCl2 and 0.5 mM EGTA) to eliminatethe non–activity-dependent staining. Preparations wereviewed through a Zeiss 403 water-immersion objective ona Bio-Rad 600 Krypton/Argon confocal laser-scanning mi-croscope. A fluorescein excitation (488 nm) filter (BHS) wasused. No staining was observed when nerves were stimu-lated in the presence of FM1–43 in Ca21-free saline. As anadditional control for nonspecific staining, loaded termi-nals could be unloaded completely by restimulation innormal saline that was free of FM1–43.

RESULTS

Musculature and innervation of a typicalabdominal segment

Typical abdominal segments in M. sexta contain about50 pairs of muscles that are divided into internal andexternal groups (Fig. 1A–C). The present study concen-trated on the second abdominal segment (A2). Somegroups of muscles in A2 are innervated by motor neuronswith cell bodies in the segmental ganglion, whereas othersare innervated by motor neurons with axons that descendin the connectives from the next anterior ganglion (Fig.1E,F). The axons of all these motor neurons exit the segmentalganglion through the dorsal and ventral nerves (DN, VN; Fig.1D,E). A third pair of nerves, the transverse nerves, exit theganglia from a dorsomedial position, where the median nerveis attached (Fig. 1D,F). After exiting the ganglion, the DNdivides into three branches. The anterior branch (DNa) inner-vates lateral muscles, the posterior branch (DNp) suppliesventral intersegmental muscles, and the lateral branch (DNl)runs toward the dorsal region of the hemisegment, innervat-ing the internal, external, and alary muscles (Fig. 1D). Theventral nerve innervates ventral external and pleural muscles.Most of the motor neurons supplying the abdominal muscleshave been identified morphologically and electrophysiologi-cally in previous studies. They are all excitatory motor neu-rons; no inhibitory motor neurons have been identified (Taylor

6 C. CONSOULAS ET AL.

and Truman, 1974; Levine and Truman 1982, 1985; Weeksand Truman, 1984). In addition to the excitatory motor neu-rons, two UM neurons, which are reactive to an antiserumagainst the neuromodulator octopamine, have been identifiedin each abdominal ganglion (Pfluger et al., 1993). Theseneurons have large cell bodies located on the posterior ventralside of the ganglia and bilaterally symmetrical axons that exitthe ganglia through the DNs (Fig. 1F).

Types of nerve terminal varicosities onabdominal muscles

Anterograde biocytin filling of peripheral axons in themain nerves (DN, VN; see Fig. 1) demonstrated an exten-sive innervation supply to each abdominal body wallmuscle, with multiple terminals on each fiber (Fig. 2A).Immunostaining with an antiserum against the synapticvesicle membrane protein synaptotagmin also showed an

Fig. 1. Internal view of the neuromuscular system of the secondabdominal (A2) hemisegment of a third instar larva. A: Musculature ofthe second abdominal segment (A2) of a third instar larva. The f-actinof muscle fibers was labeled with Oregon-green Phalloidin. Theintersegmental (VIM, ventral internal medial; VIL, ventral internallateral; DIM, dorsal internal medial) and some of the external muscles(TP, tergopleural) are apparent. B,C: Schematic representation ofmost of the abdominal muscles of the second abdominal segment.C: The large dorsal internal median (DIM), dorsal internal longditudi-nal (DIL), VIM, and VIL muscles (shown in B) have been removed toshow the underlying external muscles (TP; P, pleural; VEO, ventralexternal oblique; LIO, lateral internal oblique; DEM, dorsal externalmedial). D: Main nerve pathways in A2. Three nerves (DN, dorsalnerve; VN, ventral nerve; TN, transverse nerve) exit the ganglion. TheDN gives off three branches (DNa, anterior; DNp, posterior; DNl,

lateral). The TN is connected to the DNl near the spiracle (Sp).E,F: Schematic representation of the methods used to show terminalsof neurons with axons within the major nerve branches. The nerves orinterganglionic connectives were filled with biocytin, as indicated bythe arrows. E: Some of the motor neurons with somata and dendritesin abdominal segment 1 (AG1) are shown. These motor neuronsinnervate muscles in the body wall of abdominal segment 2 (AG2). Thecontralateral cell body belongs to a neurosecretory neuron, with anaxon that descends the connective tissue, exits the ganglion in the DN,and then enters the TN via a link nerve (Taghert and Truman, 1982).F: Filling the DN on the right labels many neurons in AG2, but onlythe bilaterally symmetrical unpaired median (UM) neurons projectinto the contralateral nerve. MN, median nerve; ML, midline. Scalebar 5 1 mm in A.


Fig. 2. Confocal photomicrographs of abdominal muscles and theirinnervation. In both panels, the f-actin of muscle fibers was labeledwith Oregon-green Phalloidin (green). A: Biocytin-filled nerve axonsand terminals on the ventral internal lateral (VIL) muscle as shownwith Cy3-conjugated streptavidin (red). Some oblique fibers belong tothe ventral external oblique muscle, which is located underneath the

VIL. Nerve endings form either clusters of large terminal varicosities(arrowheads) or rows of smaller terminal varicosities (arrows).B: Anti-synaptotagmin staining (anti-DYST2) of VIL muscles asshown with a Cy3-conjugated secondary antibody. Terminal varicosi-ties (arrowheads and arrows as in A) but not axonal branches arelabeled. Scale bars 5 100 µm.

Figure 3

elaborate pattern of nerve terminals over the muscle fibers(Fig. 2B). Immunopositive terminals were oriented inelongated rows or rosettes along or perpendicular to thelongitudinal axis of the muscle fibers (Fig. 2A,B). In mostcases, large (Fig. 2A,B, arrowheads) and small terminalswere present in the same area of the muscle fiber, but rowsof small terminals were sometimes found alone in areas ofthe muscle fibers (Fig. 2A,B, arrows). As shown below, thelarge and small terminals belong to excitatory motorneuron and UM neuron axons, respectively.

Motor (type I) terminals

Excitatory motor terminals were stained with biocytinby filling the motor neuron axons running through the cutend of connectives between the first and second abdominalganglia (Fig. 1E, arrow). This approach made it possible tosimultaneously fill the cell bodies of the motor neurons inAG1 that innervate muscles in A2 (Fig. 3A) and theiraxons and terminal varicosities over the muscles (Fig. 3).Figure 3B,D shows two examples of the pattern of motorvaricosities over some of the ventral internal medial (VIM)and ventral internal lateral (VIL) muscle fibers, respec-tively. The VIM muscle is innervated by a single motorneuron, whereas VIL is innervated by two motor neurons(Levine and Truman, 1985). These three excitatory motorneurons have their cell bodies and dendritic processes inthe first abdominal ganglion (AG1), with their axonsprojecting through the AG1–AG2 connective tissue andexiting AG2 through the DN (Levine and Truman, 1985).Stimulation of the anterior ipsilateral connective withdifferent current strengths and simultaneous intracellularrecordings from different areas along the VIM musclefibers showed (not shown) a single EJP. Similar recordingsfrom VIL fibers showed that most of the fibers are singlyinnervated by only one of the two VIL motor neurons.However, in most lateral VIL fibers, variation of thestimulus intensity evoked two different sizes of EJPs (notshown), especially in the middle regions of these fibers.Collaterals from the VIM and VIL motor axons run in

different distal nerve branches before supplying the VIMand VIL muscle fibers extensively (Fig. 3B,D,G). For theVIM fibers and the singly innervated VIL fibers, thecollaterals terminated at nonoverlapping areas along thefibers. In the dually innervated, lateral VIL muscle fibers(Fig. 3D,G,H), collaterals from the two motor axons werelargely segregated along the longitudinal axis of the fibers(Fig. 3D, arrows). Even when they reached the middleregion of the fiber, the terminal varicosities of the twomotor axons remained spatially segregated on the upper orlower surface (Fig. 3G–J). Thus, in these large musclefibers, motor terminals were not restricted to the middleregion but extended to every part of the fiber, running inrows of rosettes along or perpendicular to the longitudinalaxis of the fibers (Fig. 3C,E,F). We have designated thesevaricosities as ‘‘type I’’ according to their size (2–7 µm;Johansen et al., 1989; Budnik and Gorczyca, 1992). Withinthis range, different sizes of terminals coexist along thesame nerve branch and even in the same row (Fig. 3C,E,F).

Peripheral targets and terminalsof the UM neurons

Two strategies were followed to show the innervationpattern and terminal varicosities of the UM neurons onthe abdominal muscles. The first approach was based onthe fact that the UM neurons are the only cells with axonsrunning through both right and left DNs of the abdominalganglia (Pfluger et al., 1993). Biocytin introduced into theleft DN of the second abdominal ganglion (AG2) filled theaxons, cell bodies, and central projections of the neurons inAG1, and the cell bodies and central projections of threemotor neurons and the two UM neurons in AG2 (Fig. 4A).With long diffusion times, biocytin traveled through thecontralateral axons of the UM neurons to the periphery,providing the only staining on muscles contralateral to thefilled DN. The second approach was to stain individual UMneurons intracellularly by penetrating their cell body withbiocytin-filled microelectrodes and allowing the dye todiffuse into the periphery (Fig. 4B). The axons of the twoUM neurons enter the DN and project into distinct pri-mary branches (Pfluger et al., 1993). A summary of theperipheral branching pattern of UM neurons based onexamination of 35 of the A2 hemisegments is shown inFigure 6.

The axon of one UM neuron bifurcates to send collater-als to the posterior (DNp) and lateral (DNl) DN branches(Pfluger et al., 1993). This neuron has been designated theUMp/l neuron (Figs. 4B, 5A, 6). The posterior collateral ofthe UMp/l axon projects into numerous side branches ofDNp, supplying all the ventral internal muscles (VIL, VIO,VIM), the ventral external oblique (VEO), and, in somecases, the ventral external medial (VEM) muscle (Figs.5A–D, 6). The other collateral of the UMp/l axon runs inDNl and, after reaching the dorsal region of the hemiseg-ment, projects into all of the subbranches of DNl toinnervate the internal and external dorsal muscles (DIL,DIM, DIO, DEO, DEM) and the lateral internal oblique(LIO) muscles (Figs. 5E, 6).

The second UM neuron runs in the anterior branch ofthe DN and has been designated the UMa neuron (Pflugeret al., 1993). The DNa nerve supplies muscles that arelocated in the lateral part of the hemisegments (Levineand Truman, 1985). The UMa axon divides four times justbefore reaching the spiracle. An anterior collateral inner-vates the tergopleural muscles (ATP, PTP; Figs. 4C,D, 5F)

Fig. 3. (overleaf) A: Cell bodies, central dendritic projections, andaxons of some of the motor neurons that supply muscles of the secondabdominal segment (A2) are demonstrated with biocytin filling of thedorsal nerve (DN) in A2 toward the central nervous system. The axonstravel through the interganglionic connective posterior to abdominalsegment (AG) 1 and exit the DN of AG2. B–F: Terminal varicosities ofmotor neurons that supply body wall muscles of the second abdominalsegment were labeled with biocytin after filling from the anteriorconnective of AG2, as shown in Figure 1E. Confocal photomicrographsat low magnification show the axons and terminal varicosities over theventral internal median muscle, which is innervated by one motorneuron (B), and the ventral internal lateral (VIL) muscle fibers, whichare supplied by two motor neurons (D, G). Motor axons branch severaltimes along the longitudinal axis of the muscles. D: Only terminalvaricosities that are on the upper surface of the fibers are shown.Varicosities that belong to different motor neurons occupy differentparts of the fiber (arrows and asterisks). C,E,F: Higher magnificationof motor terminals shown in B and D (see boxes). Motor neurons makedifferent sized (3–7 µm) varicosities over the same muscle fiber.G–J: Terminal varicosities of the VIL motor neurons on the VILmuscle fibers. The distal parts of the two VIL motor axons (indicatedby arrows and asterisks) follow different trajectories to approach thesame part of the fiber (H, see marked area in G), but each axonoccupies either the lower (I) or the upper (J) surface of the fiberwithout overlapping. I and J show different focal planes of theprojected image in H. ant, anterior. Scale bars 5 100 µm in A,B,D,G; 50µm in H–J; 25 µm in C,E,F.


and then extends toward the next anterior hemisegment.The second collateral supplies the spiracular muscles(SPM) and then projects towards the dorsoventral bound-ary to innervate the pleural muscle (P). The third collat-eral enters the DNl nerve and projects distally to inner-vate the alary muscles (AM). The fourth collateral projectsinto the transverse nerve (TN) and then enters the linknerve (LN) to the next anterior segment (Figs. 4C, 6).

A number of small external muscle fibers in the ventraland lateral regions of the hemisegment that are suppliedby VN and dorsal external muscles innervated by the DNare occasionally innervated by UMp/l. In these cases,however, the UM axons supply their targets either byreaching VN branches through an anastomosis with DNbranches or by projecting from internal to more superficialmuscles without following a nerve branch. The pattern ofUM neuron innervation over external muscles was notstereotypic. From the 35 hemisegments examined, 18showed no innervation for most of the external ventralmuscles and 14 showed no innervation for most of theexternal lateral and dorsal muscles.

UM neurons made characteristic rows of small terminalvaricosities (1–3 µm) over the surface of the muscle fibers(Fig. 5). UM terminal varicosities supplied every aspect ofthe fibers. They were designated as ‘‘type II’’ terminalsaccording to their size and distribution pattern (Johansenet al., 1989; Budnik and Gorczyca, 1992).

The most anterior nerve extending from the AG2 is thetransverse nerve (TN; Fig. 1D,E). With the exception of thetwo axons of the spiracular muscle motor neurons, theaxons running through the transverse nerve belong toneurosecretory neurons and do not supply the body wallmuscles (data not shown).

Synaptotagmin immunostaining labels type Iand II terminals

Anti-synaptotagmin immunostaining of body wallmuscles showed type I and type II terminals and theirdistribution over the muscle fibers. The spatial segrega-tion and the large size of excitatory motor terminals (typeI) made it possible to differentiate between these and the

Fig. 4. Confocal micrographs showing the branching pattern ofbiocytin-filled unpaired mediam (UM) neurons. A: Both UM neuronsand four motor neurons were filled through their axons running in thedorsal nerve (DN; Fig. 1C). Only the two UM neurons (arrow) haveaxons that also exit the contralateral DN. The filled DN is lying on topof the anterior connective at the top of this photomicrograph. B: SingleUM neurons were also filled intracellularly with biocytin, which wasthen allowed to diffuse through the axons into the periphery. C,D: Theaxon of the UM neuron with an axon in the DN (UMa) in the anteriorbranch of the DN (DNa) branches several times over the dorsal–lateral

region of the hemisegment supplying lateral tergopleural (TP) muscles(ATP, anterior; PTP, posterior) and the spiracular muscle (SpM). Italso supplies muscles in the next anterior segment by entering thetransverse nerve (TN), running toward the central nervous system,and then reaching the DN of the next anterior segment through a linknerve. Another collateral enters the lateral branch of the DN (DNl; seearrows in C) to supply the alary muscles. ant, anterior; dor, dorsal; A1,abdominal segment 1; UMp/l, UM neuron with axon in the posteriorand lateral branches of DN. Scale bars 5 50 µm in A; 100 µm in B–D.


type II terminals over singly or dually innervated musclefibers. Figure 7 shows several examples of muscles inner-vated by one or two motor neurons together with the UMneurons (Fig. 7A–D). Synaptotagmin immunostainingshowed a distribution of terminals similar to that shownby biocytin staining, but without the intervening axonalregions. The excitatory motor terminals (type I) were more

numerous in the middle region of the fibers (Fig. 7A). Theterminals seen along the axons of UM neurons (type II)were also apparent with the anti-synaptotagmin staining.Type II terminals were distributed both close to (Fig. 7B,C,see arrowheads) and apart from the type I motor terminals(Fig. 7B–D, arrows). No type II terminals were found onmany of the external muscles (Fig. 7E). Although not the

Fig. 5. A: Confocal photomicrographs showing the innervationpattern of the unpaired median (UM) neuron with an axon in theposterior and lateral branches of the dorsal nerve (UMp/l) over theventral internal lateral (VIL) muscle of the second abdominal hemiseg-ment. C,D: Higher magnification images show type II endings of theUM axons. The region in the rectangle in C is shown at higher

magnification in D. B,E,F: Confocal photomicrographs show thebranching pattern of UMp/l neuron over ventral external oblique(VEO), lateral internal oblique (LIO), dorsal internal median (DIM),and anterior tergopleural (ATP) muscles. Type II endings are appar-ent. Scale bars 5 100 µm in A,B,E,F; 50 µm in C; 25 µm in D.


primary focus of this study, UM neurons in thoracicganglia (Pfluger et al., 1993) projected to several muscles,including thoracic intersegmental and leg muscles (Fig.7F,G). Interestingly, synaptotagmin-positive UM neuronterminals were apparent along the peripheral nervesbefore they entered the leg muscles (Fig. 7G, see arrows).

Calcium and activity-dependent synapticvesicle exocytosis from motor and UM nerve

terminals

To determine whether the type I and type II terminals ofthe motor neurons and the UM neuron were functionalrelease sites, the styryl dye FM1–43 was used. This dyeallows the direct study of activity and calcium-dependentsynaptic vesicle exocytosis and recycling in nerve termi-nals (Betz and Bewick, 1992) and has been used success-fully in M. sexta (Consoulas and Levine, 1998). In the firstset of experiments, the interganglionic connectives project-ing anteriorly from AG2, which contain the axons ofexcitatory motor neurons supplying muscles in the A2segment but not the UM neurons, were electrically stimu-lated (5 Hz for 5 minutes) in the presence of 4 µM FM1–43in normal M. sexta saline containing 4 mM Ca21. The effectof stimulation was verified by either direct observation ofmuscle contractions or recording the muscle membraneresponse intracellularly (Fig. 8A, inset). In unwashedpreparations, FM1–43 stained many nonneuronal mem-branes in addition to the nerve terminals, includ-

ing muscle membranes, glial cells that enwrapped thenerve branches, trachea, and tracheoles. After washing thepreparations in Ca21-free saline for 30 minutes, labelingremained only in the terminals of the stimulated motorneu-rons (Fig. 8A). Staining of the nerve terminals remaineddespite washing for up to 1 hour. Nonstimulated terminalsover other abdominal muscles were devoid of staining afterwashing. No staining was observed in terminals that hadbeen stimulated in the presence of FM1–43 in Ca21-freesaline (not shown). Terminals could be loaded after theyhad been exposed to FM1–43 in high K1 saline (notshown), but the staining was generally weaker than thatin terminals of motor axons that had been electricallystimulated. To ensure that FM1–43 labeled the neuromus-cular junctions or presynaptic sites and not glial cells ortracheoles, the preparations were fixed and processed foranti-synaptotagmin immunostaining after making imagesof the loaded terminals (Fig. 8, compare A with B).Terminals loaded with the dye could be unloaded byrestimulating the same preparation in the absence ofFM1–43 in normal saline (not shown). These experimentsverified the usefulness of FM1–43 for studying vesicularexocytosis and recycling in the neuromuscular junctions ofbody wall muscles of M.sexta.

Vesicular exocytosis in type I and type II terminals wascompared in the proximal VIL muscle fibers, which areinnervated by a single motor neuron. The motor neuronaxons were stimulated (5 Hz for 5 minutes) in the presence

Fig. 6. Schematic overview of the peripheral branching pattern of unpaired median (UM) neurons ofthe second abdominal ganglion (AG2). UMa collaterals project to the next anterior segment, but theirtargets are unknown (question marks). For abbreviations, see list.


Fig. 7. Confocal photomicrographs of anti-synaptotagmin stainingof abdominal (A–E) and thoracic (F) intersegmental and leg (G)muscles. A: Synaptic terminals over the three more medial ventralinternal longitudinal muscle (VIL) fibers innervated by a single motorneuron. B–D: Two types of varicosities can be seen on abdominalmuscles that are innervated by unpaired median (UM) neurons(A: medial VIL fibers; B: VIL, see marked region in A; C: lateral VILfibers; D: anterior tergopleural muscle). Type II terminals (see arrowsin B–D) can be differentiated from the excitatory motor terminals

(type I) from their size and distribution pattern. E: Motor terminals(type I) but not UM terminals can be seen on ventral external medianmuscle fibers of this particular preparation. F: Synaptotagmin immu-nostaining of dorsal intersegmental muscles in the mesothoracicsegment. Arrows indicate the type II terminals. G: Synaptotagminstaining of thoracic nerves as they enter and branch over the pretarsalflexor muscle (PrtFlx) in the prothoracic legs. Synaptic terminals oftype II are apparent along the nerves (arrows) in areas where there isno muscle. Scale bars 5 100 µm in A,F,G; 50 µm in B,C; 10 µm in D,E.

of FM1–43, with a suction electrode placed on the anterioripsilateral connective. After washing with Ca21-free sa-line, active presynaptic sites corresponding to type Ivaricosities were demonstrated over the muscle fibers (Fig.8C,D). Motor and UM neuron axons to these fibers wererecruited simultaneously by a suction electrode placed onthe ipsilateral DN. Synaptic vesicle exocytosis was appar-ent in both type I and II terminals after stimulating theDN (1–5 Hz for 10 minutes) in the presence of FM1–43 innormal saline (Fig. 8E,F). However, in three of sevenpreparations, only the type I terminals were labeled. Thefailure in showing exocytosis of synaptic vesicles in type IIterminals in these preparations may represent a failure torecruit the UM neuron axon or a slower rate of vesiclerecycling. The UM neuron terminals on these fibers wererecruited alone by stimulation (1–5 Hz for 10 minutes) ofthe contralateral DN in the presence of FM1–43 in normalsaline. After washing of the preparations in Ca21-freesaline, only type II terminals were stained (Fig. 8G,H). Toensure that the UM neuron was properly stimulated inthese experiments, its action potential was recorded fromthe DNl branch (Fig. 8G, inset). In all of these prepara-tions (12 of 12), the UM neuron terminals (type II) werelabeled with FM1–43. Although type I terminals that hadbeen loaded with FM1–43 could be easily unloaded byrestimulating the same preparation for 5 minutes in theabsence of FM1–43, this was not the case with the type IIterminals. In only two of 12 preparations was the dyeclearly unloaded from the terminals after 10 minutes ofstimulation (see Discussion).

DISCUSSION

Branching pattern of motor neurons overabdominal muscles

The first goal of this study was to compare the character-istics of polyneuronal and multiterminal innervation amongmuscle groups of different functions in the abdomen ofManduca sexta larvae. Insect skeletal muscles can beinnervated by excitatory, inhibitory, and modulatory neu-rons. Excitatory motor neurons are categorized as fast orslow according to the size of EJPs and the speed of musclecontraction they produce. Inhibitory motor neurons pro-duce inhibitory junction potentials and muscle relaxation.Modulatory neurons affect muscle excitability and contrac-tility. Studies on several insect species have shown numer-ous examples of singly, dually, or polyneuronally inner-vated muscles by a combination of the three neuronaltypes mentioned above (for reviews, see Hoyle, 1974;Aidley, 1985). In polyneuronally innervated fibers, themotor terminals of different motor neurons are ordinarilyin close proximity to each other in the same region of thefiber (Hoyle, 1974; Aidley, 1985; Keshishian et al., 1993).Abdominal muscles in M. sexta larvae are singly or duallyinnervated by excitatory motor neurons (Taylor and Tru-man, 1974; Levine and Truman 1982, 1985; Weeks andTruman, 1984; Sandstrom and Weeks, 1996; present study).A novel feature of the few abdominal muscle fibers that aredually innervated is that the terminal varicosities ofdifferent motor neurons remain spatially segregated.

Similarly, D. melanogaster larval abdominal muscles aresingly or dually innervated only by excitatory motorneurons (Jan and Jan, 1976; Johansen et al., 1989). Thesize of the muscle fibers, the branching pattern of motoraxons, and the size and distribution of motor terminals,

however, are different from those in M. sexta. In D.melanogaster larvae, each muscle is a single isopotentialfiber, the motor axon(s) contact the muscle fiber throughseries of terminal varicosities that remain restricted to acentral area, and, in the case of dually innervated muscles,terminal varicosities of different motor neurons are colocal-ized. In contrast, most of the M. sexta body wall musclesconsist of several large, elongated muscles fibers thatfunction as a single unit, collateral branches of the samemotor neuron contact the muscle fibers at different pointsand make synaptic connections at areas that are separatedfrom each other, and a rather broad spectrum of differentsizes of varicosities are distributed in patches along thelength of individual fibers. The most common pattern ofinnervation for both singly or dually innervated muscles ofM. sexta is that of a single motor neuron supplying eachmuscle fiber. Even for the dually innervated VIL muscle,most of the fibers are separately innervated by one of thetwo VIL motor neurons. Interestingly, in the few duallyinnervated VIL muscle fibers, terminal varicosities thatbelonged to different motor neurons are spatially segre-gated along or perpendicular to the longitudinal axis of thefiber. Unlike D. melanogaster muscle fibers, the lengthconstant of the M. sexta fibers is short relative to the totallength. Thus, longitudinal separation of motor terminalsmay have functional consequences. Indeed, intracellularrecordings from the dually innervated fibers showed exci-tatory junctional potentials of two sizes at different posi-tions along the fiber. Differential activation of the motorneurons supplying dually innervated muscle fibers mayallow graded contractions during behavior. However, themotor terminals of different excitatory motor neurons aresimilarly segregated over dually innervated fibers of theM. sexta proleg muscles, but these motor neurons are notselectively activated during proleg retraction (Sandstromand Weeks, 1996).

Branching pattern and targets of UM neurons

A second goal of this study was to investigate theperipheral branching pattern and the targets of the twoUM neurons found in each abdominal ganglion of M. sextalarvae. Furthermore, it was of particular interest to com-pare the ramifications of UM neurons among abdominal,leg, and flight muscles because these functionally differentmuscle groups may have different neuromodulatory re-quirements.

UM neurons have somata located in the posterior part ofthe ganglion close to the midline, have symmetrical den-dritic arborazitions that occupy a large neuropilar area,have primary axons that bifurcate to exit the ganglionthrough the left and right dorsal nerves, and innervate thesame targets in both hemisegments. In addition, theneurons fire large, overshooting potentials, show promi-nent after hyperpolarization, and are immunopositive forthe biogenic amine octopamine (Pfluger et al., 1993).Therefore, they share common characteristics with abdomi-nal octopaminergic efferent UM neurons that have beenidentified in several other hemimetabolous insect species.The number of UM neurons per segment differs in differ-ent species; there are three in locust (Konings et al., 1988;Pfluger and Watson, 1988; Ferber and Pfluger 1990, 1992;Stevenson et al., 1992; 1994; Stevenson and Pfluger, 1994),two in crickets (Sporhase-Eichmann et al., 1992), bushcrick-ets (Consoulas and Theophilidis, 1992), and a giant silk-moth (Brookes and Weevers, 1988; Brookes 1988), and four


in cockroaches (Eckert et al., 1992). The axons of bothefferent UM neurons in M. sexta enter specific branches ofthe DN nerves to innervate most of the body wall musclesof each abdominal segment. The UMp/l neuron innervatesventral and dorsal muscles, and the UMa neuron inner-vates lateral muscles (Fig. 6). A similar branching patternhas been reported for DUM1a and DUM2 neurons innervat-ing several inspiratory and expiratory muscles in thelocust abdominal segments (Ferber and Pfluger 1990). Inthe D. melanogaster embryo and larva, three UM neuronshave been retrogradely labeled in each abdominal neuro-mere (Sink and Whitington, 1991; Landgraf et al., 1997),but only two axons are immunopositive for octopamine(Monastirioti et al., 1995). These axons run through bothsegmental and intersegmental nerves to supply most ofthe abdominal body wall muscles.

We have shown that the axons of the UM neurons of M.sexta span a large region, running anteriorly and posteri-orly along or perpendicular to the length of the muscles. Asimilar branching pattern on D. melanogaster larva bodywall muscle has been described (Monastirioti et al., 1995).For both species, however, abdominal muscles are poorlysupplied by UM neuron collaterals in comparison with theextensive pattern of ramifications seen over thoracic legand flight muscles in the locust (Brauning et al., 1994;Brauning, 1997). This difference in the branching of UMneurons may reflect different levels of muscle activity andrequirements for neuromodulatory control. We have yet toexamine the body wall muscles in the larval thorax or theadult flight muscles in M. sexta, although it is clear thatboth receive innervation from UM neurons (Duch, per-sonal communication).

Type I and type II nerve terminals

The third goal of this study was to compare the morphol-ogy of UM neuron and motor neuron terminal varicositiesand their ability to display vesicular exocytosis and recy-cling. Motor and UM neuron terminals were staineddifferentially by filling specific nerve pathways with biocy-tin, by using an antibody against synaptotagmin, and bylabeling sites of active vesicle recycling with FM1–43.Motor terminals are larger (3–7 µm) than UM terminals(1–3 µm) and, according to D. melanogaster nomenclature(Johansen et al., 1989; Budnik et al., 1990), can becategorized as type I and type II, respectively. For excita-tory motor terminals (type I), however, it was not possibleto differentiate between big (Ib) and small (Is) terminals,as has been done for D. melanogaster. Instead, a spectrumof sizes was observed, even on singly innervated fibers.

In M. sexta, the pattern and relative size of UM neuronterminals is reminiscent of type II octopamine-immunore-active terminals of D. melanogaster abdominal muscles(Johansen et al., 1989; Budnik et al., 1990; Monastirioti etal., 1995). In both cases, type II varicosities run betweenexcitatory motor varicosities but also extend beyond them.Efferent UM nerve terminals on thoracic and abdominalmuscles have been described in the locust (Braunig et al.,1994; Braunig, 1997), but their size and spatial relation-ship relative to the excitatory motor varicosities are notknown.

Synaptotagmin, which has been implicated in dockingand fusion of synaptic vesicles through a Ca21-dependentmechanism (Perin et al., 1990), is localized to presynapticvaricosities in vertebrate (Perin et al., 1990) and insect(Littleton et al., 1993) species. Furthermore, presynaptic

varicosities can be loaded with FM1–43 only in the pres-ence of Ca21 and neuronal activity (Betz and Bewick,1992), and FM1–43 has been used successfully to showactive motor varicosities in several vertebrate species(Betz et al,. 1992), D. melanogaster body wall (Ramaswamiet al., 1994), and M. sexta leg muscles (Consoulas andLevine, 1998). The present study has demonstrated thatboth motor and UM terminals are immunopositive forsynaptotagmin and are sites of Ca21-dependent vesicularrecycling. Similarly in D. melanogaster, both type I andtype II terminals of body wall muscles are synaptotagminimmunoreactive (Littleton et al., 1993), but FM1–43 hasnot been tested on type II terminals.

Our data support the hypothesis that UM nerve termi-nals are capable of synaptic vesicle recycling. It has beendemonstrated that locust DUM neurons are capable ofreleasing octopamine in a manner that is Ca21 sensitive, isdependent on the firing frequency of the DUM neuron, andis accelerated in the presence of high-potassium saline(Morton and Evans, 1984; Orchard and Lange, 1987).Electron microscopy of UM neuron varicosities on bodywall or visceral muscles (Hoyle et al., 1980; Kiss et al.,1984; Rheuben, 1995) and of type II terminal varicositieson D. melanogaster abdominal muscles (Atwood et al.,1993; Jia et all., 1993) have shown both dense-cored andless numerous, clear vesicles. In addition to octopamine,type II terminals in D. melanogaster contain glutamate(Johansen et al., 1989), but the postsynaptic specializa-tions surrounding these varicosities are not immunoposi-tive for either type of anti-glutamate receptor antiserum(DCluRIIA, DGluRIIB; Petersen et al., 1997). Althoughpoorly defined synaptic membrane specializations areapparent in the type II terminals, ‘‘omega-figures’’ sugges-tive of exocytosis of dense-core vesicles are occasionallypresent (Atwood et al., 1993; Jia et al., 1993). Alterna-tively, octopamine may be released from clear vesicles thatare apparent in type II terminals (Atwood et al., 1993; Jiaet al., 1993) or nonsynaptically (Schurmann et al., 1991).FM1–43 staining of type II terminals suggests that, in thepresence of neuronal activity and Ca21, the vesicularmembrane undergoes exocytosis and then is retrieved

Fig. 8. Confocal micrographs of fluorescent dye FM1–43 (A,C–H)and anti-synaptotagmin (B) staining of terminal processes over ven-tral internal longitudinal muscle (VIL) fibers. A,B: When the anteriorconnective of the second abdominal ganglion (AG2) was stimulated inthe presence of FM1–43, terminals were loaded with the dye (A). Theywere subsequently unloaded by stimulation in the absence of FM1–43(not shown). B: The same terminal varicosities were later shown inthis preparation with anti-synaptotagmin immunostaining. C,D: As inA, the anterior connective of AG2 was stimulated in the presence ofFM1–43. Presynaptic terminal varicosities of different sizes made bythe single motor neuron that supplies VIL fiber are apparent. Markedarea in C is shown at higher magnification in D. E,F: When the entiredorsal nerve (DN) was electrically stimulated in the presence ofFM1–43, both type I and type II terminals (arrows) were stained onthe ipsilateral side. Marked area in E is shown at higher magnificationin F. G,H: Type II terminals were stained exclusively when thecontralateral DN was stimulated in the presence of FM1–43. Markedarea in G is shown in H. Insets in A, C, and E show the evokedexcitatory junction potentials recorded from the VIL fibers by nervestimulation; in G, the extracellular action potential of the VUMneuron is shown (asterisk). The larger action potential belongs to amotor neuron with soma and dendrites in the AG2 that was spontane-ously active in the isolated ganglion. It did not project to this muscle.Scale bars 5 20 µm in A–C,E,G; 10 µm in D,F,H; 20 msec/20 mV ininsets in A,C,E; 10 msec in inset in G.


Figure 8

internally (recycled), but the identity of vesicles to whichthe recycled membrane belongs is unknown. Althoughexcitatory motor terminals (type I) can be easily loadedand unloaded with the dye, it was usually not possible tounload type II terminals. This difference may indicate thatthere is a slower turnover of synaptic vesicles in UMneuron terminals than in motor neuron terminals. Furtherknowledge of the expression and subcellular localization ofpresynaptic and postsynaptic proteins (Sudhof, 1995) andthe mechanism by which vesicle recycling is orchestratedfor UM nerve terminals will be necessary to address thisquestion. Nevertheless, our data suggest that exocytosisoccurs at terminals of UM neurons. A future challenge is todetermine whether the sites of vesicular recycling in UMterminals represent sites of octopamine release.

ACKNOWLEDGMENTS

We thank Drs. Norman Davis and Carsten Duch fortheir helpful comments on the manuscript. C.C. wassupported by Fogarty International Center FellowshipTWO 4898. H.J.P. and R.B.L. gratefully acknowledge aDAAD/NSF International Research Award.

LITERATURE CITED

Aidley DJ. 1985. Muscular contraction. In: Kerkut GA, Gilbert LI, editors.Comprehensive insect physiology, biochemistry and pharmacology.Volume 5. New York: Pergamon Press. p 407–437.

Anderson MS, Halpern ME, Keshishian H. 1988. Identification of theneuropeptide transmitter proctolin in Drosophila larvae: characteriza-tion of fiber-specific neuromuscular endings. J Neurosci 8:242–255.

Agricola H, Hertel W, Penzlin H. 1988. Octopamin—neurotransmitter,neuromodulator und neurohormon. Zool Jb Physiol 92:1–45.

Atwood HL, Wojtowicz JM. 1986. Short-term and long-term plasticity andphysiological differentiation of crustacean motor synapses. Int RevNeurobiol 28:275–362.

Atwood HL, Govind CK, Wu C-F. 1993. Differential ultrastructure ofsynaptic terminals on ventral longitudinal abdominal muscles inDrosophila larvae. J Neurobiol 24:1008–1024.

Bell RA, Joachim FA. 1976. Techniques for rearing laboratory colonies oftobacco hornworms and pink bollworms. Ann Entomol Soc Am 69:365–373.

Betz WJ, Bewick GS. 1992. Optical analysis of synaptic vesicle recycling atthe frog neuromuscular junction. Science 255:200–203.

Betz WJ, Mao F, Bewick GS. 1992. Activity-dependent fluorescent stainingand destaining of living vertebrate motor terminals. J Neurosci 12:363–375.

Braunig P. 1997. The peripheral branching pattern of identified dorsalunpaired median (DUM) neurones of the locust. Cell Tissue Res290:641–654.

Braunig P, Eder M. 1998. Locust dorsal unpaired median (DUM) neuronesdirectly innervate and modulate hindleg proprioceptors. J Exp Biol201:3333–3338.

Braunig P, Stevenson PA, Evans PD. 1994. A locust octopamine immunore-active dorsal unpaired median neurone forming terminal networks onsympathetic nerves. J Exp Biol 192:225–238.

Broadie KS, Bate M. 1993. Development of the embryonic neuromuscularsynapse of Drosophila melanogaster. J Neurosci 13:144–166.

Brookes SJH. 1988. Unpaired median neurones in a lepidopteran larva(Antheraea pernyi). II. Peripheral effects and pharmacology. J Exp Biol136:333–350.

Brookes SJH, Weevers RDEG. 1988. Unpaired median neurones in alepidopteran larva (Antheraea pernyi). I. Anatomy and physiology. JExp Biol 136:311–332.

Budnik V., Gotczyca M. 1992. SSB, an antigen that selectively labelsmorphologically distinct synaptic boutons at the D. melanogaster larvalneuromuscular junctions J Neurobiol 23:1054–1066.

Budnik V, Zhong Y, Wu C-F. 1990. Morphological plasticity of motor axonsin Drosophila mutants with altered excitability. J Neurosci 10:3754–3768.

Cantera R, Nassel DR. 1992. Segmental peptidergic innervation of abdomi-nal targets in larval and adult dipteran insects revealed with anantiserum against leucokinin I. Cell Tissue Res 269:459–471.

Consoulas C, Levine RB. 1998. Presynaptic function during muscle remod-eling in insect metamorphosis. J Neurosci 18:5817–5831.

Consoulas C, Theophilidis G. 1992. Anatomy, innervation and motor controlof the abdominal dorsal intersegmental muscles of Decticus albifrons(Orthoptera). J Insect Physiol 38:997–1010.

Consoulas C, Kent KS, Levine RB. 1996. Remodeling of the peripheralprocesses and presynaptic terminals of leg motoneurons during meta-morphosis of the hawkmoth, Manduca sexta. J Comp Neurol 372:415–434.

Davenport AP, Wright DJ. 1985. Toxicity of chlordimefon and amitraz to theEgyptian cotton leafworm, Spodoptera littoralis and the tobacco bud-worm Heliothis virescens. Pestic Sci 16:81–87.

Eckert M, Rapus J, Nurnberger A, Penzlin H. 1992. A new specific antibodyreveals octopamine-like immunoreactivity in cockroach ventral nervecord. J Comp Neurol 322:1–15.

Evans PD. 1985. Octopamine. In: Kerkut GA, Gilbert LI, editors. Compre-hensive insect physiology biochemistry and pharmacology. Oxford:Pergamon. p 499–530.

Evans PD, O’Shea M. 1977. An octopaminergic neurone modulates neuro-muscular transmission in the locust. Nature 270:257–259.

Ferber M, Pfluger H-J. 1990. Bilaterally projecting neurones in pregenitalabdominal ganglia of the locust: anatomy and peripheral targets. JComp Neurol 302:447–460.

Ferber M, Pfluger H-J. 1992. An identified dorsal unpaired median neuroneand bilaterally projecting neurones exhibiting bovine pancreatic poly-peptide/FMRFamide-like immunoreactivity in abdominal ganglia ofthe migratory locust. Cell Tissue Res 267:85–98.

Gorczyca M, Augart C, Budnik V. 1993. Insulin-like receptor and insulin-like peptide are localized at neuromuscular junctions in Drosophila. JNeurosci 13:3692–3704.

Horikawa K, Armstrong WF. 1988. A versatile means of intracellularlabeling: injection of biocytin and its detection with avidin conjugates. JNeurosci Methods 25:1–12.

Hoyle G. 1974. Neural control of skeletal muscle. In: Rockstein M, editor.The physiology of insecta. Volume 4. New York: Academic Press. p175–236.

Hoyle G. 1978. The dorsal, unpaired, median neurons of the locustmetathoracic ganglion. J Neurobiol 8:43–57.

Hoyle G. 1983. Muscles and their neural control. New York: John Wiley &Sons.

Hoyle G, Dagan D, Moberly B, Colguhoun W. 1974. Dorsal unpaired medianinsect neurons make neurosecretory endings on skeletal muscle. J ExpBiol 187:159–165.

Hoyle G, Colquhoun W, Williams M. 1980. Fine structure of an octopaminer-gic neuron and its terminals. J Neurobiol 11:103–126.

Jan YN, Jan LY. 1976. L-glutamate as an excitatory transmitter at theDrosophila neuromuscular junction. J Physiol 262:215–236.

Jia XX, Gorczyca M, Budnik V. 1993. Ultrastructure of neuromuscularjunctions in Drosophila: comparison of wild type and mutants withincreased excitability. J Neurobiol 24:1025–1044.

Johansen J, Halpern ME, Johansen KM, Keshishian H. 1989. Stereotypicmorphology of glutamergic synapses on identified muscle cells ofDroshophila larvae. J Neurosci 9:710–725.

Keshishian H, Chiba A, Chang T, Halfton M, Harkins E, Jarecki J, Wang L,Anderson M, Cash S, Halpern M, Johansen J. 1993. Cellular mecha-nisms governing synaptic development in Drosophila melanogaster. JNeurobiol 24:757–787.

Keshishian H, Broadie K, ChibaA, Bate M. 1996. The Drosophila neuromus-cular junction: a model for studying synaptic development and function.Annu Rev Neurosci 19:545–575.

Kiss T, Varanka I, Benedeczky I. 1984. Neuromuscular transmission in thevisceral muscle of locust oviduct. Neuroscience 12:309–322.

Konings PNM, Vullings HGB, Geffard M, Buijs RM, Diederen JHB, JansenWF. 1988. Immunocytochemical demonstration of octopamine-immuno-reactive cells in the nervous system of Locusta migratoria and Schisto-cerca gregaria. Cell Tissue Res 251:371–379.

Kurdyak HL, Atwood HL, Stewart BA, Wu C-F. 1994. Differential physiol-ogy and morphology of motor axons to ventral longitudinal muscles inlarval Drosophila. J Comp Neurol 350:463–472.

Landgraf M, Bossing T, Technau GM, Bate M. 1997. The origin, locationand projections of the embryonic abdominal motoneurons of Dro-sophila. J Neurosci 17:9642–9655.


Littleton JT, Bellen HJ, Perin MS. 1993. Expression of synaptotagmin inDrosophila reveals transport and localization of synaptic vesicles to thesynapse. Development 118:1077–1088.

Levine RB, Truman JW. 1982. Metamorphosis of the insect nervous system:changes in morphology and synaptic interactions of identified neurones.Nature 299:250–252.

Levine RB, Truman JW. 1985. Dendtritic reorganization of abdominalmotoneurons during metamorphosis of the moth, M. sexta sexta. JNeurosci 5:2424–2431.

Monastirioti M, Gorczyca M, Rapus J, Eckert M, White K, Budnik V. 1995.Octopamine immunoreactivity in the fruit fly Drosophila melanogaster.J Comp Neurol 356:275–287.

Morton DB, Evans PD. 1984. Octopamine release from an identifiedneurone in the locust. J Exp Biol 113:269–287.

Oertal D, Linberg KA, Case JF. 1975. Ultrastracture of the larval fireflylight organ as related to control of light emission. Cell Tissue Res164:27–44.

Orchard I. 1982. Octopamine in insects: neurotransmitter, neurohormone,and neuromodulator. Can J Zool 60:659–669.

Orchard I, Lange AB. 1985. Evidence for octopaminergic modulation of aninsect visceral muscle. J Neurobiol 16:171–181.

Orchard I, Lange AB. 1987. The release of octopamine and proctolin froman insect visceral muscle: effects of high-potassium saline and neuralstimulation. Brain Res 413:251–258.

Perin MS, Fried VA, Mignery GA, Sudhof TC, Jahn R. 1990. Phospholipidbinding by a synaptic vesicle protein homologous to the regulatorydomain of protein kinase C. Nature 345:260–263.

Petersen SA, Fetter RD, Noordermeer JN, Goodman CS, DiAntonio A.1997. Genetic analysis of glutamate receptors in Drosophila reveals aretrograde signal regulating presynaptic transmitter release. Neuron19:1237–1248.

Pfluger H-J, Watson AHD. 1988. Structure and distribution of dorsalunpaired median (DUM) neurones in the abdominal nerve cord of maleand female locusts. J Comp Neurol 268:329–345.

Pfluger H-J, Watson AHD. 1995. GABA and glutamate-like immunoreactiv-ity at synapses received by dorsal unpaired median neurones in theabdominal nerve cord of the locust. Cell Tissue Res 280:325–333.

Pfluger H-J, Witten JL, Levine RB. 1993. Fate of abdominal ventralunpaired median cells during metamorphosis of the hawkmoth, Man-duca sexta. J Comp Neurol 335:508–522.

Ramaswami M, Krishnan KS, Kelly RB. 1994. Intermediates in synapticvesicle cycling revealed by optical imaging of Drosophila neuromuscu-lar junctions. Neuron 13:363–375.

Rheuben MR. 1995. Specific associations of neurosecretory or neuromodula-tory axons with insect skeletal muscles. Am Zool 35:566–577.

Sandstrom DJ, Weeks JC. 1996. Novel dual innervation of a larval prolegmuscle by two similar motoneurons in the tobacco hornworm Manducasexta. J Exp Biol 199:775–791.

Schurmann F-W, Sanderman R, Sanderman D. 1991. Dence core vesiclesand non-synaptic exocytosis in the central body of the crayfish brain.Cell Tissue Res 265:493–501.

Sink H, Whitington PM. 1991. Location and connectivity of abdominalmotoneurons in the embryo and larva of Drosophila melanogaster. JNeurobiol 22:298–311.

Sporhase-Eichmann U, Vullings HGB, Buijs RM, Horner M. 1992. Octopa-mine-immunoreactive neurones in the central nervous system of thecricket Gryllus bimaculatus. Cell Tissue Res 268:287–304.

Stevenson PA, Pfluger H-J. 1994. Colocalization of octopamine and FMR-Famide related peptide in identified heart projecting (DUM) neuronesin the locust revealed by immunocytochemistry. Brain Res 638:117–125.

Stevenson PA, Sporhase-Eichmann U. 1995. Localization of octopaminergicneurones in insects. Comp Biochem Physiol 110A:203–215.

Stevenson PA, Pfluger H-J, Eckert M, Rapus J. 1992. Octopamine-likeimmunoreactive cell population in locust thoracic–abdominal nervoussystem. J Comp Neurol 315:382–397.

Stevenson PA, Pfluger H-J, Eckert M, Rapus J. 1994. Octopamine-likeimmunoreactive neurones in locust genital abdominal ganglia. CellTissue Res 275:299–308.

Sudhof TC. 1995. The synaptic vesicle cycle: a cascade of protein-proteininteractions. Nature 375:645–653.

Sun XJ, Tolbert LP, HildebrandJG. 1993. Ramification pattern and ultra-structural characteristics of the serotonin-immunoreactive neuron inthe antennal lobe of the moth Manduca sexta: a laser scanning confocaland electron microscopic study. J Comp Neurol 338:5–16.

Taghert PH, Truman JW. 1982. Identification of bursicon containingneurons in abdominal ganglia of the tobacco hornworm, Manduca sexta.J Exp Biol 98:385–401.

Taylor HM, Truman JW. 1974. Metamorphosis of the abdominal ganglia ofthe tobacco hornworm, Manduca sexta: changes in populations ofidentified motoneurons. J Comp Physiol 90:367–388.

Trimmer BA, Weeks JC. 1989. Effects of nicotinic and muscarinic agents onan identified motoneuron and its direct afferent input in larva Manducasexta during metamorphosis. J Exp Biol 144:303–337.

Watson AHD. 1984. The dorsal unpaired median neurons of the locustmetathoracic ganglion: neuronal structure and diversity, and synapsedistribution. J Neurocytol 13:303–327.

Weeks JC, Truman JW. 1984. Neural organization of peptide-activatedecdysis behaviors during the metamorphosis of Manduca sexta. II.Retention of the proleg motor pattern despite loss of the prolegs atpupation. J Comp Physiol A 155:423–433.

Whim MD, Evans PD. 1988. Octopaminergic modulation of flight muscle inthe locust. J Exp Biol 134:247–266.


peripheral distribution of presynaptic sites of abdominal motor and modulatory neurons inmanduca...

Documents