Nitric oxide synthase in the thoracic ganglia of the locust: Distribution in the neuropiles and...

Nitric Oxide Synthase in the ThoracicGanglia of the Locust: Distribution in theNeuropiles and Morphology of Neurones

SWIDBERT R. OTT* AND MALCOLM BURROWS

Department of Zoology, University of Cambridge, Cambridge CB2 3EJ, England

ABSTRACTNitric oxide signaling is implicated in olfactory and visual pathways within the insect

brain. In contrast, little is known about the distribution and function of nitric oxide synthase(NOS) in the ventral nerve cord. This study uses NADPH diaphorase histochemistry todescribe the anatomy of NOS-containing neurones and the neuropilar distribution of NOS inthe thoracic nerve cord of the locust. It is shown for the first time that mechanosensoryneuropiles receive innervation from NOS-containing interneurones. Different cells innervateexteroceptive and proprioceptive projection neuropiles. In the projection neuropiles of tactileafferents, a dense meshwork of NOS-containing fibres is formed by collaterals of pairedintersegmental axons that run through the entire thoracic nerve cord, innervating exclusivelythese exteroceptive neuropiles. In neuropile areas where proprioceptive afferents terminate,stained fibres are comparatively sparse and originate from local interneurones. The protho-racic ganglion showed strongly stained dense fibres in the dorsal neuropile that were not seenin the other neuromeres. This differential NOS-expression can be related to the branchingpattern of a ventral group of neurones that was different in each neuromere. All thoracicneuromeres and the abdominal neuromeres A2 and A3 of the metathoraic ganglion containeda previously undescribed type of unpaired median neurone with bilaterally ascending anddescending intersegmental projections that stained strongly for NOS. The distribution of NOSfound in this study suggests a novel role for nitric oxide in an early stage of mechanosensoryinformation processing in all thoracic neuromeres and an additional role in the prothoracicganglion, which might be related to behavioural specializations of the forelegs. J. Comp.Neurol. 395:217–230, 1998. r 1998 Wiley-Liss, Inc.

Indexing terms: NOS; mechanoreception; NADPH diaphorase; histochemistry; insect nervous

system

The gas nitric oxide (NO) was first proposed as a neuralsignaling molecule by Garthwaite et al. (1988) and is nowwell established as a key messenger in both vertebrate andinsect nervous systems (for reviews, see Garthwaite andBoulton, 1995; Muller, 1997). NO is a short-lived moleculethat diffuses readily across cell membranes, a propertythat sets it apart from conventional transmitters. It doesnot require an elaborate synaptic machinery, and thepotential radius of action is limited only by physicaldiffusion and half-life. It can modulate synaptic transmis-sion, and there is evidence that this effect depends on theactivity state of the respective synapse itself. NO signalingcan, therefore, bypass classic cell-to-cell transmissionroutes and dynamically link sets of units in space and time(Edelman and Gally, 1992). In the nervous systems ofmammals (Bredt and Snyder, 1992) and insects (Elphick etal., 1993; Muller, 1994), NO is generated by a Ca21/calmodulin-stimulated nitric oxide synthase (NOS) in an

activation-dependent reaction. The principal action of NOin the nervous system appears to be the activation of asoluble guanylyl cyclase (sGC), resulting in increasedlevels of cyclic GMP (cGMP). The two key enzymes for anNO/cGMP signalling pathway, a Ca21/calmodulin-acti-vated NOS and an NO-activated sGC, are present in thelocust brain (Elphick et al., 1993).

Among the functions most frequently discussed for NOare roles in synaptic plasticity, learning, and olfactoryprocessing in mammals and insects (for reviews, see Breer

Grant sponsor: Balfour Funds; Grant sponsor: Furst Dietrichstein’scheStiftung; Grant sponsor: Josef Krainer Stiftung; Grant sponsor: NationalInstitutes of Health; Grant number: NS16058.

*Correspondence to: Swidert R. Ott, Department of Zoology, University ofCambridge, Downing Street, Cambridge CB2 3EJ, England.

Received 3 October 1997; Revised 22 January 1998; Accepted 26 January1998

THE JOURNAL OF COMPARATIVE NEUROLOGY 395:217–230 (1998)

r 1998 WILEY-LISS, INC.

and Shepherd, 1993; Schuman and Madison, 1994; Hol-scher, 1997; Muller, 1997). In the honeybee, NO is impli-cated both in plasticity at an early stage of olfactoryprocessing (Muller and Hildebrandt, 1995) and in memoryformation (Muller, 1996). In the locust brain, the NO/cGMP messenger system is present in the primary olfac-tory neuropile (Muller and Bicker, 1994; Bicker et al.,1996) and is also involved in visual processing in the opticlobe (Elphick et al., 1996; Bicker and Schmachtenberg,1997; Schmachtenberg and Bicker, 1997). All of this evi-dence suggests that in sensory systems of different modali-ties, NO plays a crucial role in early stages of processing.

Since NOS is present in the ventral nerve cord (Mullerand Bicker, 1994), the question arises as to whether NO isalso involved in sensory processing within this part of thecentral nervous system. Mechanosensory integration isparticularly well understood in locust thoracic ganglia(Burrows, 1996), which could thus provide a favourablesystem to study the actions of NO. What is missing,however, is information about the cellular distribution ofNOS in sufficient detail to guide physiological experi-ments. By using NADPH diaphorase histochemistry, thisstudy describes the anatomy of NOS-containing neuronesand their contribution to NOS-positive fibre meshworks inparticular thoracic neuropiles, thus providing a basis forfuture physiological studies. We show that NOS-positiveintersegmental neurones innervate exteroceptive neuro-piles, while local interneurones supply proprioceptive neu-ropiles suggesting a role for NO in mechanosensory path-ways. The NADPH diaphorase technique also allowed themorphological characterization of other previously unde-scribed neurones that are possible sources for NO in thethoracic nerve cord.

MATERIALS AND METHODS

NADPH diaphorase histochemistry

In vertebrates and insects, the neuronal Ca21/calmodu-lin-dependent NOS is responsible for aldehyde fixation-insensitive NADPH diaphorase (NADPHd) activity, so

that NOS-expressing neurones can be detected by usingNADPHd-histochemistry (Dawson et al., 1991; Hope et al.,1991; Muller, 1994; Muller and Bicker, 1994). The methodused here is a variation of the protocols given in Muller(1994) and Elphick et al. (1996). Unless stated otherwise,reagents were obtained from Sigma-Aldrich Co., Ltd.,Poole, UK. Adult locusts (Schistocerca gregaria Forskal) ofeither sex were obtained from our own crowded colony5–10 days after their final moult. Thoracic nerve cords orindividual thoracic ganglia were dissected out in locustsaline and fixed in 4% paraformaldehyde (PDH Chemicals,Ltd., Poole, UK) in 0.1 M phosphate-buffered saline (PBS,pH 7.2) for 2 hours at 4°C. Next, the tissue was infiltratedfor 16–20 hours with 20% sucrose/PBS at 4°C, embeddedin 20% gelatine, and frozen. Cryosections were cut at20–50 µm and collected on chrome alum-gelatine-coatedslides. To stain for NADPHd activity, sections were pre-soaked in 0.1 M PBS containing 0.2% Triton X-100 for 15minutes. They were then incubated with 0.25 mM or 0.1mM nitro blue tetrazolium (NBT) and 1.0 mM or 0.1 mMNADPH in 0.1 M Tris-HCl, pH 7.9, for 50–90 minutes. Incontrol experiments, NADPH was omitted from the stain-ing solution. After termination of the staining reaction indistilled water, sections were dehydrated in an ascendingalcohol series and mounted in DePeX (Serva, Heidelberg,Germany).

Neuroanatomy and image processing

The neuroanatomical terminology used here is that ofTyrer and Gregory (1982) for the fibre tracts and of Pflugeret al. (1988) for the sensory neuropiles.

For serial reconstructions, line drawings of individualsections were made by using a drawing tube attached to aZeiss Axiophot compound microscope. Drawings were digi-tized on an Epson GT-9000 flatbed scanner, and serialreconstructions were created by combining the single-section line drawings in Adobe Photoshop 3.0.4 or 4.0(Adobe Systems, Inc., Mountain View, CA) on a Macintosh8100/110 PowerPC (Apple Computers, Cupertino, CA). Forgreyscale images, conventional photomicrographs werescanned in on a Nikon LS-1000 film scanner (Nikon UKLtd., Kingston upon Thames, Surrey, UK), and histogramcorrections were performed in Photoshop when necessary.Alternatively, high-resolution video montages were ob-tained from serial sets of overlapping video images asdescribed in detail in Ott (1997). In short, a PanasonicWV-CL300 video camera (Panasonic UK Ltd., Bracknell,Berkshire, UK) was attached to the microscope and con-nected to the Power Macintosh. Video frames were cap-tured and automatically mosaicked in NIH Image 1.6.1running AutoMatch 3.0b. NIH Image was developed at theU.S. National Institutes of Health and is available via theInternet at ,ftp://zippy.nih.gov/pub/nih-image/.. Au-toMatch was developed by the author (Ott, 1997) and isavailable from ,ftp://zippy.nih.gov/pub/nih-image/contrib/..Both programs are in the public domain. For some figures,multiple focal planes or adjacent sections were superimposedin Photoshop.

RESULTS

General distribution of NADPHd activity

NADPHd-staining showed neural cell bodies, a largenumber of intersegmental fibres, and a complex pattern ofprocesses in the neuropile. Two general features of thepattern of NADPHd-positive fibres stand out: first, strong

Abbreviations

A2, . . . , A10 abdominal neuromere 1, 2, . . . , 10DCII, DCIII, DCIV Dorsal Commissure II, III and IVDCV-A1, DCVI-A1 Dorsal Commissure V and VI of A1DIT Dorsal Intermediate TractDMT Corsal Median TractiDUM intersegmental DUM neuroneLAC Lateral Association Centre

aLAC anterior LACpLAC posterior LAC

LDT Lateral Dorsal TractLVT Lateral Ventral TractMDT Median Dorsal TractMVT Median Ventral TractN1, N2, . . . , N6 paired segmental nerve 1, 2, . . . , 6R5 root of nerve 5SMC Supramedian CommissureT3 thoracic neuromere 3TT T-tractVAC Ventra Association Centre

aVAC anterior VAClVAC lateral VACmVAC medial VACvVAC ventral VAC

VC-A1 Ventral Commissure of A1VIT Ventral Intermediate TractVMT Ventral Median TractvVCLII Ventral part of Ventral Commissural Loop II

218 S.R. OTT AND M. BURROWS

staining in large parts of the Ventral Association Centres(VACs), which are prominent first-order mechanosensoryneuropiles. Stained fibres were also present in othermechanosensory neuropiles, including the Lateral Associa-tion Centres (LACs). Second, a differential distribution ofstained processes was present in the dorsal neuropiles ofthe three thoracic neuromeres. The focus of this paper is onthose neurones that contribute substantially to the neuro-pilar staining and which are therefore likely to release NOinto the neuropile. No stained fibres were seen in any ofthe paired segmental nerves or in the median nerves ofany of the three thoracic ganglia. All NADPHd-positivefibres must therefore be derived from interneurones. Stain-ing was also present in the cortices of the ganglia immedi-ately under the sheath and surrounding neuronal somata,presumably due to the presence of NOS in glial cells(Bicker and Hahnlein, 1995). The staining pattern wasunaffected by the reagent concentrations within the testedrange, although low reagent concentrations and longerincubation gave better definition of fine processes and lessbackground staining. No staining developed in the controlexperiments where NADPH was omitted from the stainingsolution.

NADPHd-positive cell bodies

The total number of NADPHd-positive cell bodies wasvariable from animal to animal and also from side to sidewithin a single ganglion. The latter observation indicatesthat at least some variability is not due to variations in theexperimental protocol. As with all histochemical methodsit is difficult to give a reasonable figure for the totalnumber of stained cells because there is no objectivethreshold criterion for positive staining. Typically, 70–100cell bodies were detected in each thoracic ganglion, ofwhich 40–60 showed intense NADPHd-reactivity. Figure 1shows cell bodies that were consistently stained andrecognized from animal to animal by their position and bythe pattern of their projections, together with the totaldistribution of other labeled cell bodies in an individualanimal. The numbering of the cell types is arbitrary andfollows the order in which the respective neurones aredescribed in the text.

Distribution of NADPHd-positive fibres inthe ventral neuropiles

The highest density of NADPHd-stained fibres occurs indistinct ventral neuropiles, the VACs (Fig. 2A–C). A densemeshwork of very fine fibres and heavily stained boutons ispresent in three morphological subdivisions of VAC, namely,the anterior VAC (aVAC), lateral VAC (lVAC), and ventralVAC (vVAC). The vast majority of these fibres is derivedfrom two pairs of prominent intersegmental axons (Fig. 2,large filled arrowheads; see also Fig. 6) that run throughthe entire thoracic nerve cord in the Ventral Median Tracts(VMTs). The location of their cell bodies is not known. Ineach ganglion, these axons give off collaterals (Fig. 2, filledarrows; see also Fig. 6) that penetrate into the VACs wherethey form profuse arborizations. Within VAC, the highestfibre densities tend to occur close to its ventral margin (seealso Fig. 2C).

Other stained axons in the VMT are smaller in diameterand have only minor branches in the neuropile surround-ing VMT; the most prominent among them (Fig. 2A, smallfilled arrowheads) is the ascending axon of one of theneurones in group ‘‘6A1’’ (see Fig. 1 and below). Fibres inthe Lateral Ventral Tract (LVT) have some very fine

collaterals near the ventral part of Ventral CommissuralLoop II (vVCLII) that run to the anterior ventral margin ofaVAC.

The medial VAC (mVAC; anterior ring tract neuropile ofTyrer and Gregory, 1982) receives no projections fromintersegmental axons in the VMTs, although an array ofcollaterals innervates a small median neuropile areaimmediately posterior to mVAC and dorsal to the VMTs(see Fig. 5C, star). In all thoracic neuromeres, however, ananterior lateral cluster of approximately 3–5 darkly stain-ing and 1–3 lightly staining cell bodies (12–20 µm indiameter; ‘‘1’’ in Figs. 1, 3A, 4) sends weakly stainedarborizations into mVAC (Fig. 5B, star). These neuroneshave additional branches in neuropile areas lateral andposterior to mVAC (Fig. 3A). In the metathoracic neuro-mere, mVAC is massively expanded as it receives theprojections of the auditory afferents. The projections of theNADPHd-positive cells are here restricted to the anterior-most region of mVAC. The main bulk of the metathoracicmVAC posterior to this region is totally void of stainedfibres.

Differential staining in the dorsal neuropilesof different thoracic ganglia

Dorsal neuropile regions also contain strongly stainedfibres, but in general they are much sparser than in theVACs. In part these dorsal processes belong to a segmen-tally repeated pair of local interneurones (‘‘8’’ in Fig. 1; seebelow). There is a striking difference, however, betweenthe prothoracic ganglion and the meso- and meta-thoracicganglia in the density of stained fibres in the dorsalneuropile. The prothoracic ganglion shows a dense aggre-gation of NADPHd-positive fibres in a distinct medialregion within the dorsal neuropile (Fig. 4A,C: star) that isnot seen in the other thoracic neuromeres (Fig. 4B,D). Themain source of this staining is 20–30 interneurones oneach side of the ganglion that have anterior ventral cellbodies (15–30 µm; ‘‘2’’ in Fig. 1A) and three distinct fields ofarborizations (Fig. 3B). The primary neurites gather toform a compact bundle of stout main processes (Fig. 5A,arrowhead) and give off a first, anterior array of branchesthat projects dorsomedially into the neuropile between thedorsal longitudinal tracts (A in Fig. 3B; Fig. 5A, filledarrows). The main processes then travel posteriorly in aconspicuous horizontal bundle (Fig. 5B, arrowhead). Ontheir way they give off a second, virtually two-dimensionalarray of projections spreading horizontally within thedorsalmost layer of VAC (V in Fig. 3B; Fig. 5A–C, openarrows). Then the main processes change direction acutely,turning dorsally into the I-tract (arrowhead in Figs. 3B,5C). They penetrate into the core of the dorsal neuropile toform a plethora of arborizations (D in Fig. 3B; Fig. 5C,filled arrows) that carry heavily stained boutons. In con-trast, the fibres that constitute the ventral field are smoothand less densely stained.

Clusters of strongly NADPHd-positive cell bodies arealso present in a similar position within the meso- andmeta-thoracic ganglia. They consist of fewer cells, how-ever, and their projections are different from those of theprothoracic cells. In the mesothoracic ganglion, there areapproximately 7–11 strongly and 2–4 weakly stainedsmall (12–17 µm) cells on each side, at the anteroposteriorlevel of vVCLII (‘‘2’’ in Fig. 1B). These are accompanied by1–2 larger, heavily stained cell bodies of elliptical shape(ca. 1533 µm), indicating the possible presence of differentcell types within the group. Many of the cells in the group

NOS IN LOCUST THORACIC GANGLIA 219

have very fine neurites that could not be traced. Thethicker neurites, including those of the large cell bodies,project into Dorsal Commissure III (DCIII; arrows in Fig.4D) where a bundle of darkly stained fibres crosses themidline (arrows in Fig. 4B). Faint fibres leave this commis-sure to join the Dorsal Intermediate Tract (DIT) andexit the ganglion via the anterior or posterior connectives.

It was not possible, however, to trace individual fibresfrom the cell bodies into the connectives, and it remainsuncertain whether these cells are local or interseg-mental neurones. Further faint arborizations emanatelaterally from DCIII to project into the surrounding neuro-pile. In the other thoracic ganglia, DCIII is void of stainedfibres.

The neurones that occupy corresponding positions in themetathoracic ganglion (‘‘2’’ in Fig. 1C and Fig. 6A,B) fallinto two classes. The primary neurites of 4–6 small (10–15µm) cells (‘‘2’’ in Fig. 6A) project dorsally into the neuropilebut could not be traced any further; DCIII, which in themesothoracic ganglion carries the neurites of at least someof the type ‘‘2’’ cells, is unstained. The primary neurites ofthe remaining 4–6 cells (‘‘2’’ in Fig. 6B) within the metatho-racic group project laterally to the anterior ventrolateralpart of the anterior Lateral Association Centre (aLAC;open arrows in Fig. 6A) where collaterals form a spatiallyrestricted array of fine arborizations (small open stars inFig. 6A). The main processes travel dorsally in a verticallateral fibre bundle (large open arrowheads in Fig. 6A;probably the Anterior Perpendicular Tract of Watson,1986) and then terminate in stout processes in the core ofthe dorsal neuropile (large open stars in Fig. 6A). Herethey branch primarily in the region lateral to DIT and VITincluding the dorsal aLAC, or run posteriorly within alongitudinal tract lateral to the Ventral Lateral Tract(small open arrowheads in Fig. 6B) to form additionalramifications in the posterior LAC (pLAC; not shown). Inall three thoracic neuromeres, DCII carries stained fibresbut their origin could not be resolved.

Intersegmental unpaired medianinterneurones

In addition to revealing the presence and origin ofNOS-containing fibre meshworks in the neuropiles, theNADPHd staining technique allowed the anatomical char-acterization of individual NOS-containing neurones. Alarge (40–50 µm) posterior unpaired median cell bodystains heavily for NADPHd in the pro- and mesothoracicganglia (‘‘3’’ in Fig. 1A,B; Fig. 7). The primary neurite joinsthose of the Dorsal Unpaired Median (DUM) neurones inthe DUM fibre bundle and bifurcates in or close to theSupramedian Commissure (SMC; small arrowheads inFig. 2C), giving rise to a pair of lateral secondary neuritesthat run to the ventral margin of the neuropile. Here theyenter the LVTs and bifurcate again into an ascending and adescending intersegmental axon. In both ganglia, this

Fig. 1. The distribution of NADPHd-positive cell bodies in thethoracic central nervous system (CNS) as reconstructed from serialsections. Dorsal cell bodies are shown as filled circles and ventral cellbodies as open circles. Black circles or circle outlines representstrongly stained cell bodies of neurones that were identified fromspecimen to specimen by their projections, size, and position. Greycircles or outlines represent other, weakly or inconsistently stainedcell bodies drawn from an individual specimen. Black and grey circleswithin a cluster indicate variability in number and staining intensity.The numbering scheme is according to the order in which the neuronesare discussed in the text. Identical numbers in the different thoracicneuromeres have been assigned only to ease the comparison ofneurones with similar cell body positions and do not imply serialhomology unless explicitly stated. (A) prothoracic, (B) mesothoracic,and (C) metathoracic ganglion. N1–N6, paired segmental nerves 1–6.Scale bar 5 200 µm.


Fig. 2. Dense NADPHd staining in the projection neuropiles of themechanosensory afferents originates from intersegmental fibres.A: Staining in the ventral region of the metathoracic ganglion, imagecombined from digital video mosaics of three adjacent horizontalsections; anterior is left. Two intersegmental fibres of extrathoracicorigin (large filled arrowheads) give off collaterals (filled arrows) intothe anterior and lateral Ventral Association Centres (aVAC and lVAC).Also shown are other types of intersegmentally projecting neurones.‘‘3A2’’ and ‘‘3A3’’ are the two unpaired median neurones in the secondand third abdominal neuromere, respectively, with bilaterally project-ing secondary neurites (large open arrowheads) that bifurcate intoascending and descending processes (small open arrowheads). ‘‘6A1’’ isone of the three lateral cell bodies in the first abdominal neuromere

with parallel ventromedial projections (stars) and intersegmentalaxons (small filled arrowhead). ‘‘7A3’’ is the soma of the ascendingneurone in the third abdominal neuromere. For cell bodies labelled ‘‘2’’and ‘‘8,’’ see Figure 1C and text. B,C: Transverse sections through amesothoracic ganglion, showing the profiles of prominent intersegmen-tal fibres (large arrowheads) and their collaterals (arrows) in aVAC (B)and lVAC (C). Dorsal is up; B is slightly anterior to Ventral Commis-sural Loop II; C is at the level of Dorsal Commissure V (DCIV) and theSupramedian Commissure (SMC). Small arrowheads in C: The twosecondary neurites of the large unpaired median neurone (‘‘3’’ in Fig.1B). Each of the two images is combined from two conventionalmicrophotographs taken at different planes of focus. Scale bars 5200 µm in A, 50 µm in B, and 100 µm in C.


large cell is accompanied by a small (20–25 µm) unpairedmedian neurone, which is often weakly stained. It was notpossible to trace the projections of this neurone beyond thebifurcation in SMC.

The mesothoracic ganglion contains an additional un-paired neurone with a smaller (20–30 µm) dorsal somaanterior to the cell bodies of the efferent DUM neurones(‘‘4’’ in Fig. 1B). The weakly stained projections (not shown)

are different from those of the unpaired median neuronesdescribed above. The primary neurite penetrates perpen-dicularly into the ganglion along the glial incision thatcarries the median tracheae. It terminates in a stoutX-shaped junction from which several processes emanateto travel dorsally and ventrally along the glial incision.

The metathoracic ganglion contains the metathoracicneuromere (T3) and abdominal neuromeres A1–A3. There

Fig. 3. A: Horizontal reconstruction of the projections of theanterior lateral cluster of NADPHd-positive cell bodies (‘‘1’’ in Fig. 1)into the medial Ventral Association Centre (mVAC) and neighbouringneuropile in the prothoracic ganglion. The prominent anterior maintracheae (Tr) are shown as a landmark. B: Horizontal reconstructionof the projections of the anterior ventral cluster of cell bodies in theprothoracic ganglion (‘‘2’’ in Fig. 1A). The neurones have three main

fields of arborizations: anterior (A), ventral (V), and dorsal (D). Forclarity, the dorsal field is omitted on the righthand side; the arrowheadindicates the profiles of the vertical processes in the I-Tract that linkthe dorsal and the ventral fields. Neurones with this branchingpattern are unique to the prothoracic ganglion. Horizontal lines referto the planes of section in Figure 5. Scale bar 5 100 µm.


are two equally large (45–55 µm) intersegmental unpairedmedian cells in T3 (‘‘3T3’’ in Fig. 1C). Their projectionsresemble those of the large unpaired median cell in thepro- and mesothoracic ganglion, but only one has itsintersegmental projections in the LVTs. The ascending anddescending intersegmental fibres of the other neurone runin the Median Ventral Tract (MVT). Abdominal neuro-meres A2 and A3 each contain two equally large (40–45µm) median cell bodies (‘‘3A2’’ and ‘‘3A3’’ in Figs. 1C, 2A)with bilaterally projecting secondary neurites (Fig. 2A:large open arrowheads) that bifurcate into ascending anddescending fibres (small open arrowheads). This patterncorresponds to that in the free abdominal neuromeresA4–A7 (R. Colbert and M. O’Shea, personal communica-tion). In A1, however, only one dorsal median cell body ofrelatively small size (20–30 µm; ‘‘4A1’’ in Fig. 1C) is present.The projection pattern is different from the unpairedneurones in A2 and A3 and resembles that of cell ‘‘4’’ in themesothoracic ganglion.

Paired intersegmental neurones

A conspicuous neurone with an anterior dorsolateral,elliptical cell body (ca. 25 3 45 µm; ‘‘5’’ in Figs. 1B, 4B) isidentifiable by its unique morphology in the mesothoracicganglion. The primary neurite bifurcates into two ascend-ing neurites, one in the ipsilateral and one in the contralat-eral connective; no local arborizations were seen. Thecontralaterally projecting neurite crosses the midline bypenetrating through aVAC. No comparable neurone ispresent in the prothoracic ganglion. The metathoracicneuromere contains a cell body in a similar position (‘‘5’’ in

Fig. 1C) that resembles the mesothoracic cell in having anascending projection in the contralateral connective andlacking local arborizations. No ipsilaterally ascendingcollateral was seen, however, and the contralateral projec-tion runs within vVCLII.

The metathoracic ganglion contains three heavily stainedlateral neurones at the anteroposterior level of the bound-ary between A1 and A2 (‘‘6A1’’ in Figs. 1C, 2A). Theprojections of their primary neurites indicate that the cellsbelong to A1. They enter the neuropile of A1 slightlyanterior to the boundary to A2 and form a tangle ofprocesses in the transverse plane (Fig. 8). One neurone (ca.30 3 60 µm; (a in Fig. 8) sends stout processes via variableroutes in between the longitudinal tracts to the dorsalmidline, where they terminate in local arborizations. Afurther prominent process runs to the contralateral side todescend intersegmentally in the Lateral Dorsal Tract(LDT). This neurone has also some local projections to theventral midline. The two other neurones (25–35 µm; b andg in Fig. 8) project exclusively to the ventral midline.Together with the ventral projections of the a-neuronethey form a parallel array of neurites (Fig. 2A, stars) thatarborizes around MVT/VMT at the anteroposterior level ofthe Ventral Commissure (VC-A1). Each of the two cells hasan intersegmental axon in the ipsilateral VMT; the axon ofone cell is ascending (Fig. 2A, small filled arrowheads),that of the other descending with an ascending collateralthat is restricted to the metathoracic ganglion.

A further strongly stained intersegmental neurone has alateral cell body (20–25 µm) and local arborizations in A3

Fig. 4. Differential distribution of NADPHd-activity in the dorsalneuropile of the pro- and mesothoracic ganglion in correspondinghorizontal (A,B) and transverse (C,D) sections. The prothoracicganglion (A,C) contains a well-defined region of dense staining in theneuropile (stars) that is not seen in the mesothoracic ganglion (B,D).

Stained cell bodies that also show up in the sections are numberedaccording to Fig. 1. Arrows 5 projections of the primary neurites ofmesothoracic type ‘‘2’’-neurones into Dorsal Commissure III (D), andstained fibres crossing the midline in this commissure (B). Scalebars 5 200 µm.


Figure 5


(‘‘7A3’’ in Figs. 1C, 2A). The ascending axon runs in theipsilateral Median Dorsal Tract (MDT) and can be tracedbeyond the prothoracic ganglion.

Other NADPHd-positive neurones

In all three thoracic neuromeres, a paired local interneu-rone with an anteromedial ventral, elliptical cell body(ca. 25 3 40 µm) stains strongly for NADPHd (‘‘8’’ in Fig. 1;Fig. 9). The fine primary neurite projects dorsolaterally toterminate in a prominent process close to the anteriormargin of the neuropile. Emanating from this process,major posteriorly directed neurites innervate three mainareas of neuropile: aLAC, pLAC, and dorsomedial neuro-pile. In the meso- and metathoracic ganglia, where theextensive projections of cluster ‘‘2’’ are absent, this neuroneis a main source for the few strongly stained processes inthe dorsal neuropile. In the prothoracic ganglion the dorsalneuropile receives NADPHd-positive innervation from bothcluster ‘‘2’’ and the type ‘‘8’’-neurone, and the projectionsoverlap partially.

The mesothoracic ganglion contains a pair of darklystained ventral midline cells (25 µm; ‘‘9’’ in Fig. 1B)

Fig. 5. The projections of the anterior ventral cluster of interneu-rones in the prothoracic ganglion (‘‘2’’ in Fig. 1A) as seen in transversesections at the levels indicated in Figure 3B. The main bundle ofneurites (arrowheads) runs posteriorly (A,B) before entering theI-Tract (C). An anterior array of collaterals (filled arrows in A) projectsdorsomedially between the longitudinal fibre tracts. Open arrows (inA–C) indicate the ventral field of arborizations close to the dorsal rimof the Ventral Association Centre (VAC). The main neurites arborizeextensively in the core of the dorsal neuropile (filled arrows in C). Alsovisible in the respective parts of the figure: A: Two cell bodies in theventral median group (‘‘9’’ in Fig. 1A); stained fibres in the anteriorLateral Association Centre (aLAC), predominantly branches of the type‘‘8’’ interneurone. B: Projections of the type ‘‘1’’ neurones that terminatein the medial Ventral Association Centre (mVAC; star). C: Collaterals ofintersegmental fibres in the VMTs innervate medial neuropil posteriorto mVAC (star). For abbreviations, see list. Scale bars 5 100 µm.

Fig. 6. The ventral cluster of interneurones in the metathoracicganglion (‘‘2’’ in this figure, in Fig. 1A, and in Fig. 2A) and theirprojections; digital video mosaics taken from transverse sections,dorsal is up. A is combined from two adjacent sections at the level ofDorsal Commissure IV (DCIV); B is from a single section slightlyposterior to A, at the level of Supramedian Commissure (SMC). Someof the neurones (‘‘2’’ in A) have faint neurites that project dorsally intothe neuropile. The remainder of the cell bodies (‘‘2’’ in B) have differentprojections: The primary neurites (open arrows in A) give off aspatially restricted ventral field (small open star in A) and form a

perpendicular bundle (large open arrowheads in A), terminating inarborizations in the dorsal neuropile (large open stars in A). From thelatter, posteriorly directed processes (small open arrowheads in B)form additional arborizations in the posterior Lateral AssociationCentre (not shown). Also seen in this figure are profiles of theintersegmental fibres (large filled arrowheads in A,B) and theircollaterals (filled arrows in B) which arborize in the Ventral Associa-tion Centres (VACs; cf. Fig. 2A). Small filled arrowheads show thesecondary neurites of the metathoracic unpaired median neurones(‘‘3T3’’in Fig. 1). For abbreviations, see list. Scale bars 5 100 µm.


immediately anterior to vVCLII. They have primary neu-rites in the ipsilateral T-tracts (TT) that terminate in thecore of the dorsal neuropile where they give rise to stoutprocesses. The faint staining of the latter prevented fur-ther tracing. Potential serial homologues in the metatho-racic ganglion (‘‘9’’ in Fig. 1C) were only weakly labelled.The prothoracic ganglion contains a variable number(6–10) of very small (10–15 µm) cell bodies in a correspond-ing position (‘‘9’’ in Figs. 1A, 5A), and some very faintprocesses are present in TT (‘‘TT’’ in Fig. 5B).

All thoracic neuromeres contain a heterogeneous assem-bly of stained cell bodies at their posterior margin (‘‘10’’ inFig. 1). In most preparations, one large (40–50 µm) pair ofcells is stained darkly in each neuromere (‘‘10’’ in Fig.4A,B), and up to ten smaller (15–30 µm) cell bodies occuron each side of the ganglion. The intensity of the labellingwas variable and sometimes further large somata werepresent. None of these cells could be traced and they aredifficult to establish as individuals on the sole basis ofNADPHd histochemistry. In the metathoracic neuromerethis posterior assemblage of stained somata is less obviousbecause the fusion between T3 and A1 increases thedistances between the cell bodies in the transverse plane.These metathoracic cells are anterior to and clearly differ-ent from the three ‘‘6A1’’ intersegmental neurones.

DISCUSSION

This study has shown that large parts of the mechanosen-sory Ventral Association Centres (VACs) receive massiveNADPHd-positive innervation from collaterals of interseg-mental axons. NADPHd-positive fibres in other mechano-sensory neuropiles originate from local interneurones.Striking differences were found in the distribution ofNADPHd staining between the dorsal neuropiles of thethree thoracic neuromeres. In previous studies NADPHd-histochemistry has been shown to detect NOS in themammalian and insect nervous system (Dawson et al.,1991; Hope et al., 1991; Muller, 1994; Muller and Bicker,1994) and it is thus very likely that the neurones described

here contain NOS and release nitric oxide. Nevertheless,some caution is necessary in the interpretation of thestaining pattern; the NADPHd method might not stain allNOS-containing neurones and/or might stain some neu-rones which do not contain NOS but some other enzymewith fixation-insensitive NADPHd activity.

Distribution of NOS in primary sensoryneuropiles

In locust thoracic ganglia, different classes of afferentsproject into different sensory neuropiles. Those from tac-tile hairs project exclusively into the Ventral AssociationCentres sensu stricto (aVAC, vVAC and lVAC) where theysynapse upon different classes of spiking interneurones(Siegler and Burrows, 1983, 1984; Nagayama and Bur-rows, 1989; Burrows and Newland, 1994). Proprioceptiveafferents terminate in the medial VAC (mVAC) and/or inthe Lateral Association Centres (LACs; Pfluger et al.,1988). The massive presence of NADPHd-positive fibreswithin projection neuropiles of exteroceptive afferentsstrongly suggests an important role for NO in an earlystage of tactile information processing. No branches ofmotor neurones or known nonspiking interneurones arepresent in the VACs; thus either the tactile afferents orfirst order spiking interneurones are likely targets for NOin these neuropiles. NO could also act within chemosen-sory pathways since the gustatory afferents from thebasiconic sensilla are thought to project into the dorsalpart of VAC (Burrows and Newland, 1994), but the strongstaining in the ventralmost parts of the VAC makes itunlikely that chemosensory pathways are a main targetfor NO released by the intersegmental neurones. Thepositions of their somata are not yet known. Possiblecandidates, however, are segmentally repeated pairs ofstrongly stained cell bodies with prominent ascendingaxons described by Muller and Bicker (1994) for A7.Similar pairs are present in A8 and A9 (and probably A10)of the terminal ganglion (R. Colbert and M. O’Shea,Personal Communication).

The observation that all thoracic VACs receive nitrergicinnervation from the same intersegmental neurones al-lows at least two hypotheses. First, the neurones might actas functional units in the sense that whenever they spikeNO is released in the VACs throughout the thoracic nervecord, modulating sensory processing in all thoracic seg-ments at the same time. Calcium necessary for the activa-tion of NOS could enter via voltage-dependent Ca21-channels. If the nitrergic neurones form output synapseswithin the VAC, such channels would need to be present inthe presynaptic terminals; alternatively, voltage-depen-dent Ca21-channels might be present in the axonal mem-brane, allowing Ca21- dependent spikes to effect the re-lease of NO.

Second, the branches of the intersegmental neuronescould receive synaptic inputs within the VACs whichtrigger a local release of NO in particular ganglia or inparticular regions of neuropile within a ganglion. In thisscheme, calcium could enter the neurones via channels inthe postsynaptic membrane or could be released fromintracellular stores following synaptic transmitter releaseand the activation of a second-messenger cascade. Thetactile afferents form a precise somatotopic map in theVACs (Newland, 1991); nitric oxide released locally withinthis map could thus selectively modulate sets of parallelsynaptic connections that convey information from the

Fig. 7. The large and small unpaired median cell body in theprothoracic ganglion that stain for NADPHd (‘‘3’’ in Fig. 1A). Horizon-tal section, anterior is top. Also visible are prominent intersegmentalaxons that enter the ganglion via the posterior connectives (arrows).Collaterals of these fibres are responsible for the dense staining in theVentral Association Centres (VACs; cf. Fig. 2). Image combined fromtwo conventional microphotographs taken at different planes of focus.Scale bar 5 100 µm.


same region of the body surface. As hypothesized forolfaction (Breer and Shepherd, 1993), NO might providecross-channel-signaling between parallel sensory chan-nels that results in plasticity at an early stage of process-ing. NO is known to be implicated in olfactory and visualprocessing in the insect brain (Bicker et al., 1996; Bickerand Schmachtenberg, 1997; Schmachtenberg and Bicker,

1997). The ability of the NOS/cGMP signaling pathway tointegrate activity over space and time might be a keyfactor for adaptive processing within topologically orderedarrays of parallel channels, such as are present in manysensory systems.

The NOS-containing intersegmental neurones do notinnervate the LACs or mVACs where proprioceptive affer-

Fig. 8. Transverse reconstruction of the projections of the threelateral NADPHd-positive cell bodies (‘‘6A1’’ in Fig. 1C) in the firstabdominal neuromere of the metathoracic ganglion. One neurone (a)has dorsal and ventral local arborizations and a descending axon inthe contralateral Lateral Dorsal Tract (LDT). The two other neurones

(b, g) have intersegmental axons in the ipsilateral Ventral MedianTract (VMT), and their local projections are restricted to the ventralmidline. MTr, glial incision carrying the median tracheae. For otherabbreviations, see list. Scale bar 5 50 µm.


ents terminate. The intersegmental innervation thus fol-lows the strict separation of exteroceptive and propriocep-tive afferent projections, underpinning the hypothesis of aspecific function of these interneurones in exteroceptiveinformation processing.

The comparatively weakly stained arborizations thatare detected in the mVACs stem from an anteriolateralgroup of local interneurones. The mVACs receive projec-tions from many of the proprioceptive chordotonal organs(see Pfluger et al., 1988). In the metathoracic ganglion, theauditory afferents of the tympanal organs also terminatein this neuropile (Rehbein, 1973; Romer and Marquart,1984). However, the latter project to the posterior part of

mVAC, whereas the proprioceptive terminals are re-stricted to the anterior part. The distribution of NOSseems to reflect this functional division, since within themetathoracic mVAC the staining is restricted to the anteri-ormost region. This differential distribution probably rulesout a role for NO in early auditory processing and pointstowards a specific function of these local neurones in theprocessing of signals from proprioceptive chordotonal or-gans. In all three thoracic neuromeres, the LACs receiveNOS-containing innervation from a segmentally repeatedpair of local interneurones (type ‘‘8’’). The presence ofstained fibres in the LACs and other dorsal neuropileregions is difficult to interpret functionally, however, since

Fig. 9. The projections of the anterior ventral interneurone (‘‘8’’ in Fig. 1B) in the mesothoracicganglion, reconstructed from horizontal serial sections. The neurone has extensive projections in thedorsal neuropile. Dashed line, midline of ganglion. Scale bar 5 100 µm.


they are ‘‘mixed’’ neuropiles that contain branches of manydifferent types of neurones.

Differential segmental expression of NOS

Afurther important result of this study is the demonstra-tion of considerable differences between the thoracic neuro-meres in the density and distribution of fibrous stainingwithin the dorsal neuropiles (Fig. 4). Much of this differen-tial staining can be related to the markedly differentprojections of interneurones with cell bodies in an anteriorventral cluster (‘‘2’’ in Fig. 1). Indeed the projections of theneurones are so different in the three neuromeres that it isquestionable whether they are segmental homologues. Inthe prothoracic ganglion they have varicose dorsal andsmooth ventral fields of arborizations, which are separatedby unbranched vertical neurites, similar to first-orderspiking local interneurones that process sensory inputsfrom the legs (Siegler and Burrows, 1984; Burrows andSiegler, 1984; Burrows and Watkins, 1986; Nagayama andBurrows, 1989). However, their main processes form aprominent horizontal bundle (Fig. 3B) and then enter theI-Tract. To our knowledge, such neurones have not yetbeen encountered in histochemical or intracellular stains.This new type of interneurone is not seen in the otherthoracic ganglia, either because they are unique to theprothoracic ganglion or because they do not express NOSin the other thoracic neuromeres. In either case, theidiosyncratic distribution of NOS in the prothoracic gan-glion calls for an explanation. In theory it should be relatedto some task that is predominantly or exclusively per-formed by the fore legs. When a locust walks, the protho-racic legs are first to encounter the terrain ahead, a factalso reflected in the density of chemoreceptors in the tarsi,which is almost doubled (Kendall, 1970). They are alsoused extensively in active sampling and in manipulatingfood to the mouthparts. It might be that the NOS-containing neurones in question are involved in thesebehavioural specializations of the fore legs.

NOS in individual intersegmental neurones

NADPHd-histochemistry revealed the presence of NOSin different types of unpaired median neurones within thethoracic CNS. The pro- and mesothoracic ganglia eachcontain a large and a small median cell, whereas two largemedian cells are present in the metathoracic neuromere(T3) and in A2 and A3. These neurones resemble interseg-mental Dorsal Unpaired Median neurones (iDUM neu-rones; Thompson and Siegler, 1991) but differ from iDUMneurones in two important points. First, their cell bodiesare more than twice as large as those of all previouslyreported iDUM neurones (10–15 µm), more closely resem-bling those of efferent DUM cells (40–70 µm). Second, theyhave their intersegmental projections in very ventral fibretracts, whereas previously describes iDUM neuronesproject via the Lateral Dorsal Tract. DUM neurones arethe progeny of a single identified neuroblast (Bate, 1976;Condron and Zinn, 1994); until the embryonic origin of theNOS-containing unpaired median cells is established, theycannot be counted among the DUM neurones. Neverthe-less it is possible that the progeny of the median neuro-blast includes a novel, anatomically distinct type of neu-rones that contains NOS. The additional neurones withunpaired median cell bodies that were found in the meso-thoracic ganglion and in A1 do not resemble any previouslydescribed type of neurone.

This study has shown that several different types of localand intersegmental NOS-containing interneurones arepresent in the thoracic CNS. In some cases, the anatomy ofthese neurones does not allow any functional conclusionsto be drawn. Other neurones, however, send NOS-containing fibres into well-defined sensory neuropiles,namely, into the projection neuropiles of exteroceptive andproprioceptive afferents. One principle in the organizationof the projections that emerges is that of exteroceptive andproprioceptive pathways receiving nitrergic innervationfrom separate sources. It is therefore likely that NO playsspecific roles in different sensory pathways within thethoracic CNS. The most interesting candidates for futurephysiological investigations are the intersegmental inter-neurones that provide the projection neuropiles of thetactile afferents with a dense meshwork of nitrergic fibres.Given the available knowledge about the components ofthe sensorimotor pathways in the locust, this model sys-tem should increase our understanding of the functions ofNO in early sensory processing.

ACKNOWLEDGMENTS

The authors thank Thomas Friedel, Michael Gebhardt,Tom Matheson, Phil Newland, Mark Wildman, and OliverMorris (all from the Department of Zoology, Cambridge,UK) for their helpful comments on earlier versions of themanuscript, and Richard Colbert and Prof. Michael O’Shea(University of Sussex) for constructive discussions.

LITERATURE CITED

Bate, C.M. (1976) Embryogenesis of an insect nervous system. I. A map ofthe thoracic and abdominal neuroblasts in Locusta migratoria. J.Embryol. Exp. Morphol. 35:107–123.

Bicker, G. and I. Hahnlein (1995) NADPH diaphorase expression inneurons and glial cells of the locust brain. Neuroreport 6:54–56.

Bicker, G. and O. Schmachtenberg (1997) Cytochemical evidence for nitricoxide/cyclic GMP signal transmission in the visual system of the locust.Eur. J. Neurosci. 9:189–193.

Bicker, G., O. Schmachtenberg, and J. DeVente (1996) The nitric oxide/cyclic GMP messenger system in olfactory pathways of the locust brain.Eur. J. Neurosci. 8:2635–2643.

Bredt, D.S. and S.H. Snyder (1992) Nitric oxide, a novel neuronal messen-ger. Neuron 8:3-11.

Breer, H. and G.M. Shepherd (1993) Implications of the NO/cGMP systemfor olfaction. Trends Neurosci. 16:5–9.

Burrows, M. (1996) The Neurobiology of an Insect Brain. New York: OxfordUniversity Press.

Burrows, M. and P.L. Newland (1994) Convergence of mechanosensoryafferents from different classes of exteroceptors onto spiking localinterneurons in the locust. J. Neurosci. 14:3341–3350.

Burrows, M. and M.V.S. Siegler (1984) The morphological diversity andreceptive fields of spiking local interneurons in the locust metathoracicganglion. J. Comp. Neurol. 224:483–508.

Burrows, M. and B.L. Watkins (1986) Spiking local interneurones in themesothoracic ganglion of the locust: Homologies with metathoracicinterneurones. J. Comp. Neurol. 245:29–40.

Condron, B.G. and K. Zinn (1994) The grasshopper median neuroblast is amultipotent progenitor cell that generates glia and neurons in distincttemporal phases. J. Neurosci. 14:5766–5777.

Dawson, T.M., D.S. Bredt, M. Fotuhi, P.M. Hwang, and S.H. Snyder (1991)Nitric oxide synthase and NADPH-diaphorase are identical in brainand peripheral tissues. Proc. Natl. Acad. Sci. U.S.A. 88:7797–7801.

Edelman, G.M. and J.A. Gally (1992) Nitric oxide—linking space and timein the brain. Proc. Natl. Acad. Sci. U.S.A. 89:11651–11652.

Elphick, M.R., I.C. Green, and M. O’Shea (1993) Nitric oxide synthesis andaction in an invertebrate brain. Brain Res. 619:344–346.


Elphick, M.R., L. Williams, and M. O’Shea (1996) New features of the locustoptic lobe: Evidence for a role for nitric oxide in insect vision. J. Exp.Biol. 199:2395–2407.

Garthwaite, J. and C.L. Boulton (1995) Nitric oxide signaling in the centralnervous system. Annu. Rev. Physiol. 57:683–706.

Garthwaite, J., S.L. Charles, and R. Chesswilliams (1988) Endothelium-derived relaxing factor release on activation of NMDA receptors sug-gests role as intercellular messenger in the brain. Nature 336:385–388.

Holscher, C. (1997) Nitric oxide, the enigmatic neuronal messenger: Its rolein synaptic plasticity. Trends Neurosci. 20:298–303.

Hope, B.T., G.J. Michael, K.M. Knigge, and S.R. Vincent (1991) NeuronalNADPH-diaphorase is a nitric oxide synthase. Proc. Natl. Acad. Sci.U.S.A. 88:2811–2814.

Kendall, M.D. (1970) The anatomy of the tarsi of Schistocerca gregariaForskal. Z. Zellforsch. 109:112–137.

Muller, U. (1997) The nitric oxide system in insects. Prog. Neurobiol.51:363-381.

Muller, U. (1996) Inhibition of nitric oxide synthase impairs a distinct formof long term memory in the honeybee, Apis mellifera. Neuron 16:541–549.

Muller, U. (1994) Ca21/calmodulin-dependent nitric oxide synthase in Apismellifera and Drosophila melanogaster. Eur. J. Neurosci. 6:1362–1370.

Muller, U. and G. Bicker (1994) Calcium-activated release of nitric oxideand cellular distribution of nitric oxide-synthesizing neurones in thenervous system of the locust. J. Neurosci. 14:7521–7528.

Muller, U. and H. Hildebrandt (1995) The nitric oxide/cGMP system in theantennal lobe of Apis mellifera is implicated in integrative processing ofchemosensory stimuli. Eur. J. Neurosci. 7:2240–2248.

Nagayama, T. and M. Burrows (1989) The morphology of a new populationof spiking local interneurones in the locust metathoracic ganglion. J.Comp. Neurol. 283:189–211.

Newland, P.L. (1991) Morphology and somatotopic organisation of thecentral projections of afferents from tactile hairs on the hind leg of thelocust. J. Comp. Neurol. 312:493–508.

Ott, S.R. (1997) Acquisition of high-resolution digital images in videomicroscopy: Automated image mosaicking on a desktop microcomputer.Microsc. Res. Tech. 38:335–339.

Pfluger, H.-J., P. Braunig, and R. Hustert (1988) The organization ofmechanosensory neuropiles in locust thoracic ganglia. Phil. Trans. R.Soc. Lond. B 321:1–26.

Rehbein, H.G. (1973) Experimentell-anatomische Untersuchungen uberden Verlauf der Tympanalfasern im Bauchmark von Feldheuschrecken,Laubheuschrecken und Grillen. Verh. dt. Zool. Ges. 66:184–189.

Romer, H. and V. Marquart (1984) Morphology and physiology of auditoryinterneurons in the metathoracic ganglion of the locust. J. Comp.Physiol. A 155:249–262.

Schmachtenberg, O. and G. Bicker (1997) Nitric oxide/cyclic GMP signaltransmission in the visual system of the locust. In: N. Elsner and H.Wassle (eds): From Membrane to Mind. Proceedings of the 25thGottingen Neurobiology Conference, Vol. 1. Stuttgart: Georg ThiemeVerlag, p. 171.

Schumann, E.M. and D.V. Madison (1994) Nitric oxide and synapticfunction. Annu. Rev. Neurosci. 17:153-183.

Siegler, M.V.S. and M. Burrows (1983) Spiking local interneurones asprimary integrators of mechanosensory information in the locust. J.Neurophysiol. 50:1281–1295.

Siegler, M.V.S. and M. Burrows (1984) The morphology of two groups ofspiking local interneurones in the metathoracic ganglion of the locust.J. Comp. Neurol. 224:463–482.

Thompson, K.J. and M.V.S. Siegler (1991) Anatomy and physiology ofspiking local and intersegmental interneurons in the median neuro-blast lineage of the grasshopper. J. Comp. Neurol. 305:659–675.

Tyrer, N.M. and G.E. Gregory (1982) A guide to the neuroanatomy of locustsuboesophageal and thoracic ganglia. Phil. Trans. R. Soc. Lond. B297:91–123.

Watson, A.D.H (1986) The distribution of GABA-like immunoreactivity inthe thoracic nervous system of the locust (Schistocerca gregaria). CellTissue Res. 246:331–343.


Nitric oxide synthase in the thoracic ganglia of the locust: Distribution in the neuropiles and...

Documents

Transcript of Nitric oxide synthase in the thoracic ganglia of the locust: Distribution in the neuropiles and...