Allatostatin-like-immunoreactive neurons of the tobacco hornworm,Manduca sexta, and isolation and identification of a new neuropeptide related to cockroach allatostatins

Allatostatin-Like-ImmunoreactiveNeurons of the Tobacco Hornworm,Manduca sexta, and Isolation and

Identification of a New NeuropeptideRelated to Cockroach Allatostatins

N.T. DAVIS,1* J.A. VEENSTRA,2 R. FEYEREISEN,2 AND J.G. HILDEBRAND1

1Arizona Research Laboratories, Division of Neurobiology, University of Arizona, Tucson,Arizona 85721

2Department of Entomology, University of Arizona, Tucson, Arizona 85721

ABSTRACTThe YXFGLamide C-terminus serves to define most members of a family of structurally

related neuropeptides, the YXFGLamides. These peptides have been identified from thenervous system of various insects and include the allatostatins of cockroaches and crickets,the schistostatins of locusts, and the callatostatins of blowflies. The YXFGLamides have beenshown to have various functions, including inhibition of juvenile hormone biosynthesis incockroaches and crickets and inhibition of contraction of certain insect visceral muscles. Wewanted to know if these peptides occur inManduca sexta and what functions they might have.A new peptide, AKSYNFGLamide, was isolated and identified from M. sexta and has beennamed ‘‘lepidostatin-1’’; this is the first YXFGLamide to be found in a lepidopteran, and thereare indications that additional YXFGLamides occur in M. sexta. An antiserum to cockroachallatostatins (YXFGLamides) was shown to recognize lepidostatin-1 ofM. sexta and was usedto map YXFGLamide-immunoreactive neurons in larvae. Because immunoreactive interneu-ronswere found to form an extensive neuropil,YXFGLamides probably function as neuromodu-lators inM. sexta. Neuroendocrine cells in the brain, abdominal ganglia, and their respectiveneurohemal organs were YXFGLamide immunoreactive and appear to release YXFGLamidesas neurohormones. Immunoreactivity to YXFGLamides and M. sexta diuretic hormone werefound to be colocalized and appear to be coreleased in these neuroendocrine cells, indicatingthat YXFGLamides may be involved in regulation of fluid transport. Innervation of thecorpora allata by YXFGLamide-immunoreactive processes was very sparse, suggesting thatthis innervation does not play an important role in allatostasis. Many thoracic motor neuronswere YXFGLamide immunoreactive, suggesting that YXFGLamides may have a myomodula-tory or myotrophic function in larvae. However, this immunoreactivity disappeared early inmetamorphosis and did not reappear in the adult. The YXFGLamide-immunoreactiveneurons in the terminal abdominal ganglion were found to innervate the hindgut, indicatingthat YXFGLamides may be involved in the control of the rate of myogenic contractions of thelarval hindgut. J. Comp. Neurol. 385:265–284, 1997. r 1997 Wiley-Liss, Inc.

Indexing terms: insects; immunocytochemistry; SCPB; FMRFamide; diuretic hormone

The YXFGLamide peptide family is usually character-ized by a Tyr-Xaa-Phe-Gly-Leu-amide C-terminus, inwhichXaa may be Gly, Ala, Ser, or Asn. The first YXFGLamideswere identified from brain extracts of the cockroach,Diploptera punctata, and because these peptides wereshown to inhibit biosynthesis of juvenile hormone (JH) bycorpora allata (CA) in vitro, they were named allatostatins(Pratt et al., 1989, 1991; Woodhead et al., 1989, 1994).Immunocytochemical studies of D. punctata showed thatallatostatin-immunoreactive neuroendocrine cells in the

brain innervate the corpora allata, and these cells appar-ently are responsible for inhibition of JH biosynthesis(Stay et al., 1992).

Grant sponsor: NIH; Grant numbers: AI-23253, AI-33429, DK-34549.*Correspondence to: Dr. Norman T. Davis, ARL Division of Neurobiology,

University of Arizona, PO BOX 21007, TucsonAZ 85721-0077.E-mail: [email protected] October 1996; Revised 14 March 1997; Accepted 1April 1997

THE JOURNAL OF COMPARATIVE NEUROLOGY 385:265–284 (1997)

r 1997 WILEY-LISS, INC.

Allatostatins similar to those of D. punctata occur in,and inhibit JH biosynthesis by the American cockroach,Periplaneta americana, the German cockroach, Blattellagermanica, and the cricket, G. bimaculatus (Belles et al.,1994; Weaver et al., 1994; Lorenz et al., 1995). AlthoughYXFGLamides, known as schistostatins, have been identi-fied from the locust Schistocerca gregaria, their inhibitionof JH synthesis has not yet been shown in this species(Veelaert et al., 1995, 1996, 1997).YXFGLamides, known as callatostatins, have been iden-

tified in the blowfly, Calliphora vomitoria, but do notinhibit JH biosynthesis in this insect (Duve et al., 1993,1994, 1995a). A YXFGLamide has been identified in thehoneybee, Apis mellifera, and, as in the blowfly, thispeptide has not been found to be allatostatic (H. Kaatz,unpublished observation). Similarly, in the present study,we have identified a YXFGLamide fromM. sexta, but testsindicate that this peptide is not allatostatic in M. sexta(P.E.A. Teal, unpublished data). The evidence, therefore,suggests that YXFGLamides do not have an allatostaticfunction in holometabolous insects. Moreover, in M. sextathis function has been shown to be under the control of anallatostatic peptide unrelated to peptides of the YXFGLam-ide family (Kramer et al., 1991).In addition to the role of YXFGLamides in the inhibition

of JH biosynthesis in some insects, these peptides mayhave several other functions. In the German cockroach, anallatostatin has been shown to inhibit vitellogenesis (Mar-tin et al., 1996). Studies show that in cockroaches, crickets,locusts, blowflies, and fruitflies many interneurons areYXFGLamide immunoreactive (Agricola et al. 1992; Stayet al., 1992; Duve et al., 1993, 1994; Duve and Thorpe,1994; Neuhauser et al., 1994; East et al., 1995; Veelaert etal., 1995; Yoon and Stay, 1995; Vitzthum et al., 1996), andthis immunoreactivity of interneurons is an indicationthat YXFGLamides act as neuromodulators. Also, theinnervation of the hindgut of cockroaches, locusts, andblowflies isYXFGLamide immunoreactive, andYXFGLam-ides modulate the rate of contraction of the hindgut andantennal heart muscles (Hertel and Penzlin, 1992; Langeet al., 1993, 1995; Duve and Thorpe, 1994; Duve et al.,1994, 1995b; East et al., 1995; Veelaert et al., 1995, 1996).Endocrine cells that are YXFGLamide immunoreactivehave been found in the posterior midgut of cockroaches,

blowflies, and fruitflies (Duve et al., 1994; Duve andThorpe, 1994; Reichwald et al., 1994; Yu et al., 1995; Yoonand Stay, 1995), indicating that YXFGLamides of thesecells have an endocrine or paracrine function.It is clear that the YXFGLamides are multifunctional,

but much is to be learned of these functions. Our study wasundertaken becauseM. sexta appeared to be ideally suitedfor investigating the YXFGLamides. Because the neuro-anatomy of M. sexta has been studied in depth, themapping of neurons with allatostatin-like (ASL) immuno-reactivity in this species offered considerable promise inshowing this immunoreactivity in previously identifiedneurons. Moreover, it seemed likely that identification ofASL-immunoreactive (ASL-ir) neurons that project to pe-ripheral targets or neurohemal organs would be useful inindicating possible functions of YXFGLamides. To helpvalidate these immunocytochemical studies, we also under-took the isolation and identification ofYXFGLamides inM.sexta. For practical reasons, our initial study has beenlimited to the identification and synthesis of one YXFGL-amide-type peptide from M. sexta and to the mapping ofASL-ir neurons in larvae. In the future we hope to identifyadditional YXFGLamides, and we have undertaken stud-ies of ASL-ir cells in the adult central and stomatogastricnervous systems and in the enteric endocrine system.Because a synthetic YXFGLamide of M. sexta is nowavailable, we will perform bioassays on the functions ofthis peptide.

MATERIALS AND METHODS

Animals

Manduca sexta (Lepidoptera: Sphingidae) were rearedat 25°C and 50–60% relative humidity under a long-dayphotoperiod regimen (17 hours light-7 hours dark). Larvaewere fed on an artificial diet adapted from that of Bell andJoachim (1976).

Immunocytochemistry

Primary antisera. Allatostatin-like (ASL) immunore-activity was shown by using an antiserum raised to theoctadecapeptideallatostatin,ASB2, (AYSYVSEYKALPVYN-FGL-NH2) of D. punctata (Pratt et al., 1991). Productionand testing of this antiserum (No. 368T) was reported byReichwald et al. (1994). Because the antiserum immuno-stains the same cells that are recognized by several otherantisera raised to allatostatins (N.T. Davis, unpublisheddata), it probably recognizes the common C-terminus(YXFGLamide) of the YXFGLamides. The capacity of thisantiserum to recognizeAKSYNFGLamide, the YXFGLam-ide identified from M. sexta, was tested by liquid-phasepreadsorption (24 hours at 4°C) of the working dilution(1:2,000) of the antiserum with 100 nM synthetic AKSYN-FGLamide.All specific immunostaining was eliminated bythis test, indicating that the antiserum recognizes AKSYN-FGLamide ofM. sexta. In addition, the antiserum probablyrecognizes other allatostatin-like peptides in this insect,and so, neurons immunoreactive to the anti-allatostatinserumwill be referred to asASL-immunoreactive (ASL-ir).Antisera to leucokinin-IV and to M. sexta diuretic hor-

mone were used to show colocalization of their immunore-activities with that of ASL immunoreactivity. Productionand testing of the DH and LK antisera are described byChen et al. (1994a) and Veenstra and Hagedorn (1991),

Abbreviations

IIa4, IIa5 types of median neuroendocrine cells of the brainAG1-6 abdominal ganglia of segments one through sixASL-ir allatostatin-like-immunoreactiveCA corpus allatumCAP2b cardioacceleratory peptide 2bDH diuretic hormone ofM. sextaFMRFamide molluscan cardioactive tetrapeptideJH juvenile hormoneL1-4 lateral neurosecretory cells of abdominal ganglia one

through sixLK leucokininM5 median neurosecretory cell five of abdominal gangliaMA9 median neurosecretory cell of the ninth neuromere of the

terminal abdominal ganglionNCC112 fused nerves one and two of the corpus cardiacumPM1-3 proctodeal median neurons one through threePvO perivisceral organSCP molluscan small cardioactive peptideSeG subesophageal ganglionTG1-3 thoracic ganglia one through threeTAG terminal abdominal ganglion

266 N.T. DAVIS ET AL.

respectively, and cells immunoreactive to these antiseraare referred to here as DH-ir and LK-ir.Cross-reactivity between the anti-allatostatin and anti-

diuretic hormone sera was tested by preadsorption of eachantiserum (working dilution) with 100 nM of the heterolo-gous antigen. These tests showed that the antisera are notcross-reactive at the dilutions used in this study.An antiserum (No. 232) to the molluscan cardioactive

peptide, FMRFamide (a gift fromW.H. Watson III, Univer-sity of New Hampshire, Durham, NH) and a monoclonalantibody to SCPB (Monoclonal Laboratory, University ofWashington, Seattle, WA) were used to help establish theidentity of certain ASL-ir neurons. The anti-FMRFamideserum has been characterized by O’Donahue et al. (1984).The SCPB antibody was characterized byMasinovsky et al.(1988) and Arbiser and Beltz (1991). Cells immunostainedby the anti-FMRFamide serum will be referred to asFMRFamide-ir, and those immunostained by the SCPBantibody are referred to as SCP-ir.Cardioacceleratory peptide 2b (CAP2b), which was iden-

tified from M. sexta by Huesmann et al. (1995), has aC-terminus (YLAFPRMamide) similar to that of SCPB(YLAFPRVamide), and, therefore, this peptide was used ina preadsorption control of the SCPB antibody. Preadsorp-tion of a 1:1,000 dilution of the antibody supernatant with100 nM synthetic CAP2b eliminated all specific immuno-staining by the antibody, indicating that this monoclonalantibody can recognize CAP2B.Tissue preparation. Insects were anesthetized by

chilling on ice, and dissections were performed in a physi-ological saline solution formulated for M. sexta (Chris-tensen and Hildebrand, 1987). Brains and ventral nervecords of larvae of various stages were removed and fixedfor approximately 18 hours at 4°C in 4% (w/v) paraformal-dehyde in sodium phosphate buffer (0.1 M, pH 7.4).Immunostaining protocols. Whole-mounted prepara-

tions were studied by an immunofluorescence methodadapted from that of Davis (1987). Briefly, phosphate-buffered (0.01 M, pH 7.4) normal saline containing 0.5%(v/v) Triton X-100 was used for washing tissues, and thissolution, containing 10% (v/v) normal goat serum, wasused as a blocking solution and as a diluent for the primaryand secondary antisera. At room temperature and withgentle agitation, the tissues were treated for approxi-mately 18 hours with the primary antiserum diluted1:2,000 (anti-allatostatin, anti-leucokinin, anti-diuretic hor-mone, and anti-FMRFamide) or 1:1,000 (anti-SCPB). Forlabeling by the secondary antiserum, the tissues, afterthorough washing, were treated with a 1:200 dilution ofgoat anti-rabbit IgG conjugated to one of the followingfluorophores: rhodamine; fluorescein; indocarbocyanine(Cy3); or indodicarbocyanine (Cy5; Jackson ImmunoRe-search Laboratories, West Grove, PA). Immunostainedtissues were mounted on slides in 80% (v/v) glycerin in0.05 M NaHCO3 buffer (pH 9.5).The brains of late-stage larvae were too thick to be used

effectively in whole mounts. Therefore, this material wasimmunostained as above and then embedded in 5% (w/v)agarose (Sigma type I-A) and sectioned at 50 µM on aVibratome. These serial sections were then mounted onslides as above.Demonstrations of colocalized immunoreactivities were

performed by double immunostaining. Colocalization ofSCP and FMRFamide immunostaining was shown byusing fluorescein-conjugated goat anti-mouse to label the

SCP-ir cells and Cy5-conjugated goat anti-rabbit IgG tolabel FMRFamide-ir cells. Colocalization of diuretic hor-mone and ASL immunoreactivities was shown by labelingthe DH-ir cells with Cy5-conjugated goat anti-rabbit IgG,followed by immunostaining with rhodamine-conjugatedrabbit IgG anti-allatostatin. The anti-allatostatin wasconjugated to rhodamine, as described by Veenstra et al.(1995).Triple immunostaining was achieved by treating the

above diuretic hormone-ASL double-stained tissues withanti-SCP labeled by fluorescein-conjugated goat anti-mouse IgG.

Backfills

To provide positive identification of certain efferentASL-ir neurons, these cells first were stained by backfill-ing through the stump of a nerve branch containing theiraxons. Dissected ganglia were bathed in a Weevers-typephysiological saline solution (Davis et al., 1989), and theisolated nerve stump was bathed in a 2% (w/v) solution ofNeurobiotin (Vector Laboratories, Burlingame, CA) indistilled water. After backfilling for 2 or 3 days at 4°C, thetissue was fixed, washed (as for immunocytochemistry),and treated with a 1:1,000 dilution of streptavidin-rhodamine (Jackson ImmunoResearch Laboratories) forapproximately 18 hours at 4°C. In addition, 2% (w/v)rhodamine-dextran (MW 3,000; Molecular Probes, Eu-gene, OR) was used for some backfills, and backfilling andfixation were performed overnight as described above. Byeither method, successfully stained backfills were thendouble-stained forASL immunoreactivity, by using fluores-cein- or Cy5-conjugated goat anti-IgG for immunostaining.

Confocal microscopy

Images of the immunostained and backfilled prepara-tions were obtained with a Bio-Rad MRC 600 laser-scanning confocal microscope with a BHS, YHS, or RHSfilter block with the argon-krypton laser or a GHS filterblock with the argon laser. Images were processed onAdobe Photoshop and PowerPoint programs, and thepictures of the digitized images were made with a Laser-graphics slide maker. The color prints were made fromEktachrome Lumiere 100 transparencies. Where needed,the digitized images were modified only to enhance con-trast, to merge images of double-stained tissue, and toprovide pseudocolor.

Peptide isolation, identification,and synthesis

Isolation of M. sexta YXFGLamides was performedaccording to the same general purification schedule success-fully applied for the isolation of several mosquito peptides(Veenstra, 1994; Veenstra et al., 1997; Veenstra, unpub-lished data) and a cockroach midgut peptide (Veenstra andLambrou, 1995). Five hundred ventral nerve cords of latepharate-adultM. sextawere dissected and stored frozen at280°C. Then they were extracted in Bennett’s mixture (1%NaCl, 5% formic acid, 1% trifluoroacetic acid [TFA], and 1M HCl in water; Bennett et al., 1981) and prepurified on aC18 reversed-phase cartridge (Waters Corporation, Mil-ford, MA). The HPLC separations were performed on anEconosil C-18 column (10 3 250 mm, 10 µm) from AlltechAssociates, Inc. (Deerfield, IL) and a Microsorb phenylcolumn (4.6 3 250 mm, 5 µm) from Rainin InstrumentCompany, Inc. (Woburn, MA). Aliquots of fractions were

ASL-ir NEURONS IN M. SEXTA 267

analyzed by using a competitive ELISA for allatostatinASB2 of Diploptera punctata (AYSYVSEYKRLPVYNFGL-amide), as described by Reichwald et al. (1994). Sequenceanalysis was performed by the Division of Biotechnology ofthe University of Arizona, by using an Applied Biosystemsmodel 477A pulsed-liquid nitrogen protein sequencer withan online Applied Biosystems model 120A phenylthiohy-dantoin amino acid analyzer.AKSYNFGLamide, the YXFGLamide-type peptide iden-

tified from M. sexta and CAP2B (pELYAFPRVamide) weresynthesized by the Division of Biotechnology of the Univer-sity of Arizona, by using a Gilson model 422 multiplepeptide synthesizer and 9-fluorenylmethoxy-carbonylchemistry. The crude peptides were purified by HPLC andtheir sequences were confirmed by tandem mass spectralanalysis on a Finnigan TSQ7000 (Palo Alto, CA) instru-ment at the Southwest Environmental Health SciencesCenter of the University of Arizona.

RESULTS

Background: Anatomy of the larval CNS

A brief summary of the anatomy of the larval CNS ofM.sexta is necessary to establish the locations and projectionsof various ASL-ir neurons. The CNS consists of a dorsalbrain (Br) and a ventral nerve cord formed by a chain ofeleven segmental ganglia, namely, the subesophageal gan-glion (SeG), three thoracic ganglia (TG1-3), six unfusedabdominal ganglia (AG1-6), and the terminal abdominalganglion (TAG; Fig. 1A–C). The SeG is a composite gan-glion consisting of the neuromeres of the labial, maxillary,and mandibular segments. The TAG is formed by thefusion of the neuromeres of the seventh through ninthabdominal segments, plus almost indistinguishable rem-nants of the neuromere of the greatly reduced tenthsegment.Most efferents and afferents of each unfused segmental

ganglion travel via two pairs of lateral nerve trunks, thedorsal and ventral nerves (Figs. 1A,B, DN, VN). In addi-tion, a third pair, the intersegmental nerves, arises fromthe TG1-2 and TG2-3 connectives (Fig. 1A, IsN). Thesenerves are so named because they contain the motor axonsthat innervate most of the longitudinal and oblique trunkmuscles of the next posterior segment. Thus, most motorneurons of trunk muscles are located in the ganglion of thesegment anterior to that of theirmuscles. The intersegmen-tal nerve of most segments is incorporated completely intothe adjacent connective and is confluent with the dorsalnerve of the next posterior segment (Fig. 1A, DN of TG1;Fig. 1B, DN ofAGs); only in themesothoracic andmetatho-racic segments can portions of intersegmental nerves bedistinguished. However, studies of motor neurons of thetrunk muscles have shown that the axons of these neuronsextend posteriorly through their ipsilateral connective andinto the dorsal nerve of the next segment (Taylor andTruman, 1974; Casaday and Camhi, 1976). This pathwayis the remnant of the intersegmental nerve and is desig-nated here as the intersegmental motor tract.In addition to the pairs of lateral nerve trunks, a median

nerve extends posteriorly from each unfused ganglion andthen divides to form the transverse nerves (Fig. 1A,B,MRN, TN). The median and transverse nerves containmotor axons of the spiracular muscles of the next posteriorsegment (Taghert and Truman, 1982). Each transversenerve also contains the terminal processes of severalneuroendocrine cells, and much of the surface of this nerve

appears to be neurohemal. In the abdomen, the proximalportion of each transverse nerve is a slightly thickenedstructure called the perivisceral organ (PvO; Truman,1973; Fig. 1B,C). Neuroendocrine cells project to the PvOsvia three pathways (a,b,c in Fig. 2), and their somata aremedian or lateral in the ganglia (Taghert and Truman,1982; L1-4, M4-5 in Fig. 2).

Immunocytochemistry

A very large number of ASL-ir neurons and processeswere immunostained throughout the larval CNS. Most ofthese cells appeared to be interneurons, and becausenumerous processes in the connectives were stained (seeFig. 5A, arrow), many of these ASL-ir cells must beascending or descending interneurons.An extensiveASL-irneuropil was stained in the brain and ganglia, and theprocesses were mostly very fine and uniform, rather thanthick and varicose. Relatively few ASL-ir processes werestained in peripheral nerves, and these processes projectto neurohemal organs, to the hindgut, and to thoracicmuscles. The abundance of ASL immunostaining madeidentification of individual neurons and their processesvery difficult, and therefore, the scope of this study had tobe limited to ASL-ir neuroendocrine cells and certaindistinctive interneurons and motor neurons.ASL-ir interneurons of the larval brain. Staining for

ASL immunoreactivity resulted in labeling of extensiveareas of neuropil and numerous somata in the brain (Fig.3A,B). Several ASL-ir interneurons are conspicuous oridentifiable by their respective positions. The first is agroup of approximately seven somata (Fig. 3A,E, straightarrow), of which one is especially prominent (Fig. 3C), inthe anterior cortex of each lobe of the tritocerebrum; thesecells appear to contribute ASL-ir fibers to the tritocerebralneuropil, and some of the cells project through the tritocere-bral commissure to arborize in the contralateral tritocere-bral neuropil. Nearby, there are rather diffuse ASL-irprocesses in the neuropil of the larval antennal center anda few, weakly staining ASL-ir cells that may be the originof these processes (Fig. 3D).Densely arrayed ASL-ir processes are found throughout

most of the protocerebrum (Fig. 3F,G,H), but, with theexception of the larval optic center (Fig. 3H, open arrow),specific areas of organized neuropil are difficult to distin-guish. Several intensely stainedASL-ir interneurons appar-ently contribute to much of the ASL-ir neuropil of theprotocerebrum. One pair of these neurons is located in theposterior lateral cortex of each protocerebral hemisphere(Fig. 3B,H, arrowhead), and as shown in Figure 3B, theinitial projections of these interneurons can be distin-guished from the rest of the neuropil. A strongly stained,ASL-ir interneuron that contributes many processes to theanterior and dorsal regions of the protocerebral neuropil isfound in the dorsal cortex of each protocerebral hemi-sphere (Fig. 3A,F, arrowhead). Often this cell is accompa-nied by one or two weakly stained somata. Finally, a pair ofASL-ir interneurons lies in a paramedian position in theanterior ventral cortex, and these cells send processes tothe ventral median region of the neuropil (Fig. 3A,E,curved arrow).ASL-ir neuroendocrine cells in the brain and colocal-

ization of diuretic hormone immunoreactivity in these

cells. There are two pairs of ASL-ir cells in the parsintercerebralis (Fig. 3A, open arrow), each of which hasprojections that can be followed ventrally through a me-


Fig. 1. Depiction of the central nervous system and nerve trunks ofanM. sexta larva (dorsal view; anterior is up).A:Brain, subesophagealganglion, and thoracic ganglia (arrow, motor nerve of muscle VO1; Br,brain; CC, corpus cardiacum; CeC, circumesophageal connective; DN,dorsal nerve; FG, frontal ganglion; IsN, intersegmental nerve; MRN,median recurrent nerve; NCC3, third nerve of the CC; TcC, tritocere-bral commissure; TG1-3, thoracic ganglia; TN, transverse nerve; VN,ventral nerve). B: Unfused abdominal ganglia representing ganglia of

the first through sixth segments (AG, abdominal ganglion; DN, dorsalnerve; MRN, median recurrent nerve; TN, transverse nerve; VN,ventral nerve.C: Terminal abdominal ganglion (DN7&8, dorsal nervesof the seventh and eighth abdominal segments; NM7, neuromere ofthe seventh abdominal segment; PN, proctodeal nerve; TmN, terminalnerve; VN7&8, ventral nerves of seventh and eighth abdominalsegments). Also seeAbbreviations list.


dian chiasma and then contralaterally into nerve NCC112, which extends to the corpus cardiacum (Fig. 3B,arrow; Fig. 1A, NCC 112, CC). From this nerve theASL-irprocesses extend into an extensive, varicose neurohemalmeshwork in the corpus cardiacum (Fig. 4C, CC). Inaddition, a few fine ASL-ir processes are found in thecorpus allatum (Fig. 4C, CA), but, as will be shown below,these processes probably come from neuroendocrine cellsin the pars lateralis, rather than in the pars intermedialis.The location and projections of the ASL-ir cells in the

pars intermedialis indicate that they are neuroendocrinecells. These cells occur in a cluster known as the IIa group(terminology of Buys and Gibbs, 1981, as elaborated byCopenhaver and Truman, 1986; Homberg et al., 1991;Zitnan et al., 1995). On the basis of their numbers, relativesizes, and immunoreactivities, five different subtypes ofneuroendocrine cells are recognized in group IIa (Zitnan etal., 1995). Two of these five subtypes, IIa4 and IIa5, can bedistinguished by the fact that each is represented by onlytwo pairs of neuroendocrine cells, and so the two pairs ofASL-ir neuroendocrine cells must be one of these twosubtypes. The IIa4 cells are labeled by the anti-diuretichormone serum (Veenstra and Hagedorn, 1991; Chen etal., 1994b; Zitnan et al., 1995), and the IIa5 cells arelabeled by the SCPB antibody (Homberg et al., 1991).Therefore, to identify the ASL-ir cells, the two subtypes ofneuroendocrine cells were investigated by triple stainingfor diuretic hormone, ASL, and SCP immunoreactivities.Because ASL immunoreactivity was colocalized with thatof diuretic hormone in the IIa4 cells (Fig. 4A, arrows) but

not with that of SCP in the IIa5 cells (Fig. 4A, arrowheads),the results indicated that the ASL-ir cells are the same asthe IIa4 cells.Double immunostaining showed colocalization of di-

uretic hormone and ASL immunoreactivities in processesof the IIa4 cells only in the corpora cardiaca allata (Fig. 4C,CC), and the aforementionedASL-ir processes that extendinto the corpora allata did not show DH-ir (Fig. 4C, CA).Therefore, these ASL-ir processes of the CA probably donot originate from the IIa4 neuroendocrine cells. Althoughwe were unable to identify origin of these processes, itseems likely that they are from small neuroendocrine cellsin the pars lateralis.ASL-ir interneurons of AG1 through AG6. In the

unfused abdominal ganglia (AG1-6), there are four distincttypes of ASL-ir interneurons, identifiable as such becausethey do not have immunoreactive processes projecting intoperipheral nerves. Each type is recognized on the basis ofits respective location, projection, and immunostainingcharacteristic. These four types serve as a basis for recog-nizing comparableASL-ir interneurons of other ganglia. Inaddition to these interneurons, a few other interneuronswere immunostained in AG1-6, but their staining wasweak or inconsistent.The type-1 ASL-ir interneurons are a pair of relatively

large, weakly stained cells in the anterior ventral cortex;their projections are difficult to distinguish (Fig. 5A, T1).Representing the second type of ASL-ir interneurons is agroup of three somata in each anterior, lateral cortex (Fig.5A, T2). One of these somata is more dorsal and alwaysstains more intensely than the others. The type-3 ASL-irinterneurons are a triplet of cells found in the posteriordorsolateral cortex, one of which is always less intenselyimmunostained than the others (Fig. 5A, T3). The neuritesof these interneurons extend in a transverse tract leadingto the contralateral neuropil, where they arborize. Thefourth type of ASL-ir interneuron has a median, unpairedsoma located in the posterior ventral cortex (Fig. 5A, T4).Abifurcated neurite extends anteriorly from this cell, but itsfiner arborizations could not be distinguished (Fig. 5B).ASL-ir neuroendocrine cells in AG2-6. In AG2-6 (but

not inAG1), ASL-ir staining revealed a lateral neuron thatprojects ipsilaterally into each ventral nerve and fromthere into the PvO of the next posterior segment (Fig. 5A,L3, curved arrow). The terminal processes of these ASL-ircells have many prominent varicosities located near thesurface of the PvO (Fig. 5A, PvO), a characteristic ofneurohemal release sites and an indication that theseASL-ir cells have a neuroendocrine function.The pattern of projection of the ASL-ir neuroendocrine

cell is comparable with that of the L2-4 neuroendocrinecells of abdominal ganglia and previously described byTaghert and Truman (1982), Davis et al. (1993), and Chenet al. (1994b). Unlike neuroendocrine cell L1, cells L2-4(and theASL-ir neuroendocrine cell) project to the PvO viathe ipsilateral ventral segmental nerve (Fig. 2, L2-4, a). Toconfirm that the ASL-ir neuroendocrine cell is one the ofthe L2-4 cells, these cells were backfilled with Neurobiotinand then stained with streptavidin-fluorescein. Next, theganglionwas stained forASL immunoreactivity and rhoda-mine-conjugated goat anti-rabbit IgG used for visualiza-tion. One of the L2-4 cells was doubly stained by thisprocedure, indicating that the ASL-ir neuroendocrine cellbelongs to the L2-4 group (Fig. 4B, arrow). In addition, onepair of median neuroendocrine cells, the M5 cells (Fig. 4B,

Fig. 2. Depiction of the neuroendocrine cells of a larval abdominalganglion and their three projection patterns to the perivisceral organ(dorsal view; anterior is up; a, projection of lateral neuroendocrinecells to the PvO via the ventral nerve; b, projection of the M4 cellsanterior to the PvO; c, projection of L1 neuroendocrine cell posteriorlyto reach the PvO via the dorsal nerve of the next ganglion; DN, dorsalnerve; M4&5, median neuroendocrine cells four and five; MRN,median recurrent nerve; TN, transverse nerve; VN, ventral nerve).Also seeAbbreviations list.


Fig. 3. ASL immunostaining of cells and neuropil of the larvalbrain (frontal views; dorsal is up). Photographs in this and otherfigures are digitized, confocal microscopic images of merged opticalsections.A:Anterior half of the third instar brain, showingASL-ir cellsand neuropil of the tritocerebrum (straight arrow), protocerebralinterneurons characterized by strong immunostaining (curved arrowand arrowhead), and type IIa neurosecretory cells of the pars interme-dialis (open arrow).B: Posterior half of the third instar brain, showingASL-ir cells and neuropil (arrowhead), one of several protocerebral

interneurons characterized by strong immunostaining; arrow, ASL-irprocesses in nerve NCC 112). C: ASL-ir cells and neuropil of thetritocerebrum. D: ASL-ir cells and neuropil of the olfactory center.E–H:ASL-ir cells and neuropil in serial Vibratome sections (50 µm) ofbrain of fifth instar larva (straight arrow, ASL-ir cells and neuropil ofthe tritocerebrum; curved arrow and arrowheads, interneurons of theprotocerebrum characterized by strong immunostaining; open arrow,ASL-ir immunostaining of neuropil of the optic center). Scale bars 5100 µm forA, B, 50 µm for C, D, 200 µm for E–H.

Figure 4


M5), also was stained by this backfilling procedure but didnot exhibit ASL immunoreactivity.Colocalization of ASL, leucokinin, and diuretic hor-

mone immunoreactivities in the L3 neuroendocrine

cells. Next, we were interested in determining which ofthe three lateral neuroendocrine cells is ASL-ir. A previousstudy had shown that L3 and L4, but not L2, can beimmunostained by an anti-leucokinin serum (Chen et al.,1994b). Therefore, abdominal ganglia were doubly stainedfor leucokinin and ASL immunoreactivity. Leucokinin andASL immunoreactivities were colocalized in one of the twoLK-ir neuroendocrine cells (Figs. 5C,D, arrow), indicatingthat the ASL-ir neuroendocrine cell is not L2 but could beL3 or L4.Because the study by Chen et al. (1994b) also had shown

that the L3 cells can be immunostained by an antiserumagainst M. sexta diuretic hormone, abdominal gangliawere doubly stained for diuretic hormone and ASL immu-noreactivity. The results showed that diuretic hormoneandASL immunoreactivities are colocalized in the L3 cellsand in the varicosities in the PvO (Figs. 5E,F, arrow, PvO).Therefore, the ASL-ir neuroendocrine cells of the abdomi-nal ganglia are the L3 cells.ASL-ir interneurons in the terminal abdominal gan-

glion. A constriction in the terminal abdominal ganglion(TAG) serves to distinguish the seventh neuromere fromthe remainder of the ganglion (Fig. 1C, Nm7), and ASL-irinterneurons found in the seventh neuromere are mostlycomparable with those of AG1-6 (Fig. 6A, T3; Fig. 6B, T2,T4).The terminal region of the TAG contains a number of

ASL-ir interneurons, but it was not possible to relate theseneurons to the types of ASL-ir interneurons found inAG1-6 (Figs. 6A,B; note unlabeled ASL-ir somata). Doubleimmunostaining, initially done in a study of ASL-ir neuro-endocrine cells of the TAG (see below), revealed colocaliza-tion of diuretic hormone and ASL immunoreactivities in aprominent pair of paramedian interneurons at the poste-rior end of the TAG (Figs. 4D,E; 6A, IN). These are localinterneurons, and they have extensive arborizations in theneuropil of the ninth neuromere (see Fig. 8A). The diuretic-

hormone immunoreactivity of these interneurons wasdescribed by Chen et al. (1994b).ASL-ir neuroendocrine cells of the TAG. The L3

neuroendocrine cells of neuromere seven of the TAG areASL-ir (Fig. 6A, L3), and, as in the unfused abdominalganglia, double staining showed that ASL, diuretic hor-mone, and leucokinin immunoreactivities are colocalizedin these cells (Fig. 4E, L3A7).The presence of ASL-ir processes in the terminal nerves

and in the ventral segmental nerves of the eighth segment(Fig. 6A, arrows) indicated that there are efferent ASL-irneurons in the TAG or, possibly, in gangliamore anterior toit.AsingleASL-ir process in the ventral nerve of the eighthsegment was traced proximally to an intensely immuno-stained, ipsilateral cell in the eighth neuromere of the TAG(Fig. 6A, curved arrow, L3A8). Because this neuron appearsto be a segmental homolog of the L3 neuroendocrine cellsof other abdominal ganglia, it is designated L3A8. TheASL-ir process of this L3 cell was traced distally throughthe eighth ventral nerve and into a nerve formed by theconfluence of a branch from the eighth ventral nerve andthe terminal nerve (Fig. 7A, VN8, TmN, arrow). By way ofthis nerve, the ASL-ir process enters the cryptonephridialchamber of the rectum and extends varicose branchesthat, along with other ASL-ir processes (see below), form adiffuse neurohemal-like system in the chamber.The location and projection pattern of the L3 neurons of

the eighth neuromere are very similar to those of diuretichormone-immunoreactive neuroendocrine cells previouslydescribed by Chen et al. (1994b). The results of double-immunostaining experiments indicated that diuretic hor-mone andASL immunoreactivities are, indeed, colocalizedin these cells (Figs. 4D,E, L3A8).Staining forASL immunoreactivity also showed a pair of

cells in the posterior, midventral cortex of the TAG, justanterior to the DH/ASL-ir interneurons noted above (Figs.4E, 6A, MA9). These cells, which are median and projectbilaterally, appear to be comparable with the medianneuroendocrine cells of the unfused abdominal ganglia,but because they are not clearly related to any one type ofthe median abdominal neuroendocrine cells, they are

Fig. 4. Double and triple staining of the brain (frontal view) andganglia (dorsal view, anterior is up) of third instar larvae. Note thatred, green, and blue are used as the primary pseudocolors and thatyellow results from the overlap of red and green of the merged images.Yellow coloration of the same cell or process indicates colocalization ofstaining, but some yellow also may be the result of the partial overlapof red and green of structures that are not the same (see overlappingcells on the left side of Fig. 4G). A: Triple immunostaining of the brainfor YXFGLamides (red), diuretic hormone (green), and SCP (blue).Yellow coloration indicates that ASL immunoreactivity is colocalizedwith that of DH in the IIa4 neuroendocrine cells (arrows) rather thanthe SCP-ir IIa5 cells (arrowheads). B: Neurobiotin backfill (green) oflateral neuroendocrine cells (L2-4) and of a pair of median neuroendo-crine cells (M5) in the fourth abdominal ganglion. Immunostaining forYXFGLamides (red) shows that one of the lateral cells (yellow) isASL-ir (arrow) and that the M5 cells are not ASL-ir (the yellow areasover the M5 cells is due to the overlap of YXFGLanide-ir neuropil). C:Double immunostaining of the corpus cardiacum (CC) and corpusallatum (CA) shows that diuretic hormone (red) and ASL (green)immunoreactivities are colocalized in processes in the CC (yellow) butnot in the CA (green). D: Triple immunostaining of the caudal portionof the terminal abdominal ganglion shows the position of SCP-ir (blue)lateral proctodeal neurons (PL1&2) and shows that DH (green) andASL (red) immunoreactivities are colocalized in the lateral neuroendo-crine cells (L3A8) and in a pair of interneurons (IN). E: Double

immunostaining of the caudal portion of the TAG shows colocalizationof DH (green) andASL (red) immunoreactivities in lateral neurosecre-tory cells of the seventh (L3A7) and eighth (L3A8) neuromeres and inthe neuropil and a pair of local interneurons (IN) of the ninthneuromere. The median neuroendocrine cells of the ninth neuromere(MA9) show only ASL immunoreactivity. F: Dextran-rhodamine back-filling (green) of the left proctodeal nerve followed by SCP immunostain-ing (red) of the TAG shows that double staining (yellow) occurs in thethree contralateral proctodeal neurons (PL1-3) but not in the SCP-irinterneuron (IN). On the side ipsilateral to the backfill, the four SCP-irneurons (red) correspond to the three proctodeal lateral neurons andthe interneuron.G:Dextran-rhodamine (red) backfill of the proctodealnerve followed by ASL immunostaining (green) shows that doublestaining (yellow) occurs in the proctodeal median cells of the seventh(PM2A7) and eighth (PM2A8, PM3) neuromeres of the TAG.H: Dextran(red) backfilling of the motor neuron of mesothoracic muscle VO1,followed by ASL immunostaining (green). Double staining (yellow)indicates that the motor neuron identified by backfilling (arrow) isASL-ir. I:Double immunostaining of the TAG forYXFGLamide (green)and leucokinin (red) shows that ASL and LK immunoreactivities arecolocalized (yellow) in the lateral neuroendocrine cells of the seventhneuromere (L3A7) and in the median neurosecretory cells of the ninthneuromere (MA9). Lateral neuroendocrine cells of the eighth neuro-mere (L3A8) showASL but not LK immunoreactivity. Scale bars 5 100µm inA, 50 µm in B–I.


designated simply as median neuroendocrine cells of theninth abdominal neuromere, or MA9. The processes of theMA9 cells project bilaterally into each terminal nerve (Fig.8B) and then into a branch that, as noted above, enters thecryptonephridial chamber (Fig. 7A, arrow). There the

ASL-ir processes of the MA9 cells extend with ASL-irbranches from the L3A8 neuroendocrine cell to form theneurohemal-like structure noted above.The earlier demonstration (Chen et al., 1994b) of a

similar pair of leucokinin-ir neuroendocrine cells in the

Fig. 5. ASL immunostaining of neurons and processes in thefourth abdominal ganglion of a third instar larva. In this and thefigures that follow, the ganglia are viewed dorsally and anterior is up.A: ASL immunoreactivity is seen in many ascending and descendingprocesses in the connective (arrow), in the perivisceral organ (PvO), infour types of ASL-ir interneurons (T1-4), in a lateral neuroendocrinecell (L3), and in the projection of that cell into the ventral nerve(curved arrow). B: A type-4 interneuron showing the characteristic

bifurcation of its neurite. C,D: Lateral neuroendocrine cells immuno-stained for leucokinin (C) and the same tissue stained for ASL-ir (D);note that colocalization of LK and ASL immunoreactivities occur inone soma (curved arrow).E,F:Lateral neuroendocrine cell L3 immuno-stained for diuretic hormone (E), and the same tissue stained forASL-ir (F). The results show that colocalization of DH and ASLimmunoreactivities occur in the L3 cell (curved arrow) and in varicosi-ties in the PvO. Scale bars 5 50 µm.


TAG prompted our use of double staining for ASL andleucokinin immunoreactivities. These experiments showedthat ASL and leucokinin immunoreactivities are colocal-ized in the MA9 cells (Fig. 4I, MA9) and in the L3 cells of theseventh neuromere, but not in the L3 cells of the eighthneuromere (Fig. 4I, L3A7, L3A8).ASL-ir visceromotor neurons of the TAG. As de-

picted in Figure 7A, the proctodeal nerve innervates theanterior rectum, ileum, ileal bulb, pylorus, and ampullaeof the Malpighian tubules (terminology of Reinecke et al.,1973; Fig. 7A, PN, Im, Py, MA).We observed immunostain-ing of two or three ASL-ir processes that extend into theproctodeal nerve via the terminal nerve (Fig. 6E) andbranching processes on the visceral muscles of the pyloricand ileal regions of the hindgut (Fig. 6F). We were unableto follow these ASL-ir processes back to their somata.Therefore, as a first step in identifying the ASL-ir neuronsthat innervate the hindgut, the proctodeal nerve wasbackfilled with dextran-rhodamine or dextran-fluorescein.Among the cells stained by bilateral backfilling were twomedian pairs of somata in the dorsal cortex of AG5 andAG6 and in neuromere 7 of the TAG (Figs. 6C, 7B,C, PM1,PM2A7; Fig. 7B, PM1, PM2; Fig. 7C, PM1A7, PM2A7). Intheir respective ganglia, the smaller of these neurons aredesignated the PM1 (proctodeal median-1) neurons, andthe larger pair are the PM2 neurons (Fig. 7B, PM1, PM2).(Where needed for clarity, subscripts are used to designatethe respective abdominal ganglion or neuromere of theseneurons, as for example in Figure 6C, PM1A7 and PM1A8).The PM1 and PM2 neurons are also present in neuro-mere-8 of the TAG, and, additionally, a pair of largemedian somata, the PM3 neurons are located in thisneuromere (Figs. 6B, 7C, PM3). The PM3 somata areusually somewhat smaller than those of the PM2 neuronsand are in the ventral rather than dorsal cortex.In neuromere-9 of the TAG, there are two more median

clusters of neurons that innervate the proctodeum (Fig.7C, arrows), but, because these neurons were not immuno-reactive to our antisera, they were not studied further.Backfilling also resulted in the demonstration of theterminal ramifications of an afferent projection, presum-ably from a sensory receptor on the hindgut (Fig. 6C,D,curved arrow). Unilateral backfilling of the proctodealnerve resulted in the staining of only the ipsilateralmember of each of the respective pairs of the PM neurons,showing that the efferent projections of these cells are notbilateral (Fig. 6D, PM1, PM2A7, PM2A8, PM3).When the backfilled proctodeal neurons were stained for

ASL immunoreactivity, only PM2A7, PM2A8, and PM3 weredouble-stained (Fig. 4G, PM2A7, PM2A8, PM3), indicatingthat these are the neurons that provide theASL-ir innerva-tion of the anterior hindgut. However, in larvae later thanthe third instar, the PM2A7 neurons failed to show ASLimmunoreactivity, and ASL immunoreactivity of PM3 andof PM2A8 also was lost by the wandering stage of fifthinstar larvae.SCP/FMRFamide-ir visceromotor neurons of the TAG.

In addition to the proctodeal median neurons, three pairsof lateral somata were stained by backfilling through theproctodeal nerve; these cells are designated the proctodeallateral neurons (PL1-3). Neurons PL1 and PL2 appear tobe located in neuromere-8, and PL3 in neuromere-9 (Figs.6C,D, 7C, PL1-3). Unilateral backfilling of the proctodealnerve showed that these cells project contralaterally intothe proctodeal nerves (Fig. 6D, PL1-3)

SCP immunostaining of the TAG backfilled through theproctodeal nerve resulted in the double staining of thePL1-2 neurons and sometimes PL3 as well (Fig. 4F,PL1-3). The PL1 soma is located just anterior, and PL2 justposterior, to L3A8 (Fig. 4D, PL1, PL2, L3A8). A weaklystained SCP-ir cell, which appears to be an interneuron, isalso located between PL1 and PL2 (Fig. 4F, IN). Two orthree SCP-ir processes were observed in the proctodealnerve, and scattered SCP-ir endings were observed on thepyloric region of the proctodeum. Double immunostainingshowed that SCP and FMRFamide immunoreactivities arecolocalized in neurons PL1-2 and that PL3 is FMR-Famide-ir but usually not SCP-ir (Fig. 6G, PL1-2; Fig. 6H,PL1-3). In addition, neuroendocrine cell L3A8 and otherneurons of the TAG are FMRFamide-ir but not SCP-ir(compare Fig. 6G with 6H). Unlike the immunoreactivityof theASL-ir median cells innervating the hindgut (PM2A7,PM2A8, PM3), the PL1-3 cells did not lose their SCP/FMRFamide immunoreactivity at the onset ofmetamorpho-sis.ASL-ir interneurons of the thoracic ganglia. ASL

immunostaining of the thoracic ganglia (TGs) resulted inthe labeling of many more somata than were found in theunfused abdominal ganglia (compare Fig. 5A with Figs.8C,D), and some of these additional somata appeared to bemotor neurons (see below). Somata comparable with thetype-1 abdominal neurons (paramedian in the anterior,ventral cortex) are present, and occasionally a cell compa-rable with the type-4 neuron (unpaired, median in theposterior ventral cortex) can be distinguished (Fig. 8D, T1,T4). Somata comparable with the type-2 neurons (in theanterior dorsolateral cortex) are present, as are the so-mata of the type-3 neurons (triplet in the posterior dorso-lateral cortex; Fig. 8C, T2, T3).Absence of ASL-ir neuroendocrine cells in the TGs.

Lateral neuroendocrine cells of each of the thoracic gangliaproject via the median recurrent nerve to neurohemalrelease surfaces on the next posterior transverse nerve(Wall and Taghert, 1991). These lateral neuroendocrinecells are labeled by the SCP antibody (Mesce et al., 1993),but double immunostaining for SCP andASL immunoreac-tivity showed that these cells are not ASL-ir. Moreover,there are no ASL-ir processes in the median and trans-verse nerves of the thoracic and first abdominal segment(Fig. 8F, TN). Thus, we conclude that theASL-ir cells of thethoracic ganglia do not have a neuroendocrine function.ASL-ir motor neurons of the TGs. Several ASL-ir

processes extend into the ventral nerve trunks of TG1, 2,and 3 and enter nerve branches innervating the legmuscles (Fig. 8E, arrows). The ASL-ir processes alsoextend from TG1 and 2 into the intersegmental nerve (Fig.8F, IsN, double arrow) and could be traced into nervebranches innervating various ventral longitudinal andoblique muscles (Fig. 8H). In TG2 and 3, ASL-ir processesin the intersegmental nerve extend into the dorsal nervevia the connection between the intersegmental and dorsalnerve (Fig. 8F, DN, arrowhead), but no ASL-ir processesenter the dorsal nerve via the trunk of this nerve (Fig. 8F,DN).Most of the motor neurons in the intersegmental nerve

of TG2 have not been identified, and that identificationwas beyond the scope of this study. Nevertheless, wewished to identify at least oneASL-ir motor neuron for thisstudy. Fortunately, we were able to locate a single ASL-irprocess that extends from the metathoracic intersegmen-


Figure 6

tal nerve into a nerve branch that innervates the metatho-racic ventral oblique muscle 1 (VO1; terminology of Eaton,1982; Figs. 1A, 8F, single arrow). Moreover, ASL-ir pro-cesses are found on this muscle (Fig. 8H). The motorneuron that innervates muscle VO1was shown by backfill-ing its nerve with dextran-rhodamine (Fig. 8G); doublestaining indicated that the soma of this motor neuron isASL-ir (Fig. 4H, arrow).As indicated in the Background: Anatomy of the Larval

CNS section (above), the intersegmental nerve of TG3 hasbecome incorporated into the connective to, and the dorsalnerve of AG1, forming the intersegmental motor tract. Afew ASL-ir processes from TG3 could be distinguished inthis intersegmental motor tract; one of these processesprojects to the metathoracic VO1 muscle and others, todorsal muscles of the first abdominal segment.SomeASL immunoreactivity of processes in the thoracic

nerves could be distinguished in all larval stages butappeared to become more pronounced late in the fifthinstar. This immunoreactivity was lost very early in pupaldevelopment and did not reappear in the thoracic nerves ofadults.ASL-ir neurons of the subesophageal ganglion. As

shown in Figures 8I and J, immunostaining resulted in thelabeling of many ASL-ir somata and processes in thesubesophageal ganglion (SeG). Because this is a highlymodified, composite ganglion, the identities of most of theASL-ir neurons of three neuromeres of the SeG were notapparent. None of these neurons projects to the corporacardiaca or to other neurohemal organs, and, therefore,none is a neuroendocrine cell. The maxillary nerves con-tain ASL-ir processes (Fig. 8J, arrow), and this immuno-staining suggests that some of the ASL-ir neurons of theSeG aremotor neurons of maxillarymuscles. In contrast tothe maxillary nerve,ASL-ir processes were not observed inthe labial and mandibular nerves.

The dorsal nerve of TG1 containsASL-ir processes of theintersegmental motor tract, which descends from thelabial neuromere of the SeG, and so some of the ASL-irneurons of this neuromere are probably motor neurons ofmuscles of the prothoracic segment.

Peptide isolation and identification

Five hundred abdominal ventral nerve cords of latepharate adults were homogenized in Bennett’s mixture,prepurified on a C-18 Seppak, and the peptide fractioninjected onto a C-18 column. Analysis of allatostatinimmunoreactivity by ELISA indicated two major immuno-reactive peaks and three or four minor ones (Fig. 9). Thetwo major immunoreactive peaks were injected on thesame C1-column used previously to purify various insectneuropeptides (Veenstra, 1994; Veenstra and Lambrou,1995; Veenstra et al., 1997; Veenstra, unpublished data),but no immunoreactivity was recovered. The major differ-ence between this and previous purifications was thetissue used for isolation. In previous efforts rather largeamounts of complex material were used for purification,whereas here we used cleanly dissected nerve cords. Ourresults suggested that recovery of peptides from thiscolumn is severely compromised if the injected material isa limited quantity of an already relatively pure peptide.Therefore, to purify the immunoreactive peak elutingbetween 46 and 48 minutes on the first HPLC column, weused aMicrosorb column and 0.1% heptafluorobutyric acidas a pairing ion. A single immunoreactive peak wasrecovered and submitted to a final purification on the samecolumn, by using 0.1% TFA as the pairing ion. The finalchromatogram showed a single UV-absorbing peak thatwas immunoreactive (Fig. 10). We estimate that less than100 pmol of the peptide were recovered, and virtually all ofthis material had to be used for sequence analysis. There-fore, an insufficient amount of the peptide was availablefor later comparison with the synthetic peptide.Analysis yielded the following unambiguous sequence:

Ala-Lys-Ser-Tyr-Asn-Phe-Gly-Leu. The largest amount ofany other amino acid in the first cycle was 3.6 pmol of Gly;carryover was similarly very limited, the largest amountbeing 4.9 pmol of Asn in the sixth cycle (Table 1).All insect peptides related to this sequence of amino

acids have an amidated C-terminus, and the anti-allato-statin serum used in the ELISA requires an amidatedC-terminus for recognition. The peptide, therefore, wassynthesized with a C-terminal amide. The synthetic pep-tide had the same retention time on reverse-phase HPLCas the natural peptide, whereas the retention times ofC-terminally amidated oligopeptides and their analogswith a C-terminal acid differ profoundly under theseconditions (Veenstra, 1994). It was, therefore, concludedthat the structure of the peptide isAKSYNFGLamide. TheYNFGLamide C-terminus of this peptide indicates that itbelongs to the YXFGLamide family of peptides. Thispeptide is the first one in the YXFGLamide family to beidentified from the Lepidoptera.

DISCUSSION

ASL-ir interneurons

We have shown that there are many ASL-ir interneu-rons in the brain and ventral ganglia ofM. sexta larvae andthat some of these cells are ascending or descendinginterneurons. Because synaptic neurotransmitters are

Fig. 6. Immunostaining and backfills of cells and projections of theterminal abdominal ganglion of third instar larvae.A,B:ASL immuno-reactivity in merged sections of the dorsal two-thirds (A) and ventralone-third of the TAG. In the seventh neuromere (anterior half) there isimmunostaining of many of the same ASL-ir cells identified in otherabdominal ganglia (T2-4, L3) and of large median somata, whichproject to the proctodeum (PM2). ASL-ir cells identified in the poste-rior half (eighth and ninth neuromeres) of the TAG include the largeproctodeal median cells (PM2-3), the L3 (L3A8) and MA9 neuroendo-crine cells, and a pair of DH/ASL-ir interneurons (IN). ASL-ir pro-cesses are seen in ventral nerve 8 (curved arrow) and in the terminalnerve (arrow). C: Bilateral backfill of the proctodeal nerves withdextran-rhodamine (right nerve) and dextran-fluorescein (left nerve),showing the array of cells in the TAG that innervate the proctodeum.These cells include medial cells in the seventh and eighth neuromeres(PM1-3) and lateral cells in the eighth and ninth neuromeres (PL1-3).In addition, an afferent projection has been filled (curved arrow). D:The same ganglion as in (C), showing only the dextran-rhodaminebackfill; the staining shows that the projections of the PL1-3 cells arecontralateral and that the projections of the PM1-3 cells are unilat-eral. The afferent projection from the proctodeal nerve can also be seenin this figure (curved arrow). E: ASL-ir of processes in the proctodealnerve in the pyloric region of the hindgut.F:ASL-ir endings associatedwith visceral muscles of pyloric region of the hindgut. G,H: Doubleimmunostaining of the posterior part of the TAG for SCP (G) andFMRFamide (H) immunoreactivities. Note that immunoreactivities toSCP and FMRFamide are colocalized in the PL1 and PL2 cells but notin the PL3 cells. In addition, the L3A8 neuroendocrine cells and othersare FMRFamide-ir but not SCP-ir. Scale bars 5 50 µm in A–E, G, H,100 µm in F.


small, fast-acting molecules, it is likely that YXFGLam-ides of the ASL-ir interneurons function as neuromodula-tors rather than as synaptic neurotransmitters. In theory,such functions could be manifested through a generalparacrine release, i.e., via nonsynaptic exocytosis and

broad diffusion to relatively distant receptors in the neuraltissue. Alternatively, the release could be parasynaptic,i.e., involving focal diffusion over relatively short distancesto pre- or postsynaptic receptors. These two types ofrelease probably represent the extremes of a continuum,

Fig. 7. Depiction of the larval hindgut and its innervation. A:Innervation of the hindgut by the proctodeal and terminal nerves(arrow shows the location of a nerve that enters the cryptonephridialchamber; Im, ileum; ImB, ileal bulb; MG, midgut; MA, Malpighianampulla; MT, Malpighian tubule; PN, proctodeal nerve; Py, pylorus;RC, rectal cryptonephridium; TmN, terminal nerve; Tr, trachea; VN8,ventral nerve of the eighth abdominal segment). B: Representation ofabdominal ganglia 5 or 6 and cells (PM1, PM2) that descend toinnervate the hindgut via the proctodeal nerve. C: Terminal abdomi-

nal ganglion showing the median (PM1-3 [subscripts denote abdomi-nal neuromeres]) and lateral (PL1-3) cells that innervate the hindgutvia the proctodeal nerve. Stippled cells show the relationship of thelateral neurosecretory cells (L3, L3-4) to the proctodeal lateral neu-rons. Cells that project through the right proctodeal nerve are solid,and those that project through the left nerve are open. Arrows point tounidentified proctodeal median cells of the ninth neuromere (PN,proctodeal nerve; TmN, terminal nerve). Also seeAbbreviations list.


Fig. 8. A: Diuretic hormone immunoreactivity of the soma andneuropil of a local interneuron of neuromere-9 of the TAG. These cellsare also ASL-ir. B: Leucokinin immunoreactivity of paired medianneurosecretory cells of neuromere-9 and their bilateral projection intothe terminal nerve. These cells are also ASL-ir. C,D: Frontal sectionsof the upper third (C) and lower two-thirds of the metathoracicganglion of a third instar larva, showing four types ofASL-ir interneu-rons (T1-4). Some of the unlabeledASL-ir somata in (D) are believed tobe motor neurons. E: ASL-ir processes (arrows) in branches of theventral nerve of the mesothoracic ganglion of a fourth instar larva. F:ASL-ir processes (double arrows) in the intersegmental nerve (IsN) ofthe mesothoracic ganglion of a fourth instar larva. ASL-ir processes

also extend from this nerve into the dorsal nerve (arrowhead) and intoa branch (arrow) that innervates ventral oblique muscle 1. No ASL-irprocesses are found in the neurohemal transverse nerve (TN) and inthe base of the dorsal nerve (DN). G: Motor neuron of ventral obliquemuscle 1, shown by backfilling the motor nerve with dextran-rhodamine of a fourth instar larva. H: ASL-ir processes on ventraloblique muscle 1 of a fourth instar larva. I,J: ASL-ir processes andsomata in frontal sections of the dorsal two-thirds (I) and ventralone-third of the subesophageal ganglion of a third instar larva. Someof the ASL-ir processes extend into the labial nerve (arrow). Scalebars 5 50 µm inA–D, G, I, J, 200 µm in E, F, H.

and there may be instances in which parasynaptic releasecan affect synaptic domains of various sizes. Examples ofboth nonsynaptic and parasynaptic release have beenfound in nervous systems (see reviews by Hokfelt, 1991;Golding, 1994), but these phenomena have received littleattention in insects.Some peptidergic interneurons of insects are character-

ized by arborizations that are thicker, more irregular, andmore varicose than most processes in the neuropil. Theappearance of these branches is very similar to that of theneuroendocrine processes found in neurohemal organs,suggesting that peptidergic interneurons that have neuro-secretory-like processes provide a general paracrine (non-synaptic) release within the nervous system. Examples ofthis type of peptidergic interneuron include in the Arg-vasopressin-immunoreactive, descending interneurons ofthe SeG of cockroaches and grasshoppers (Davis andHildebrand, 1992; Tyrer et al., 1993) and the PBAN-irdescending interneurons and CCAP-ir interneuron 704 ofM. sexta (Davis et al., 1993, 1996). Moreover, nonsynapticexocytosis of dense-core vesicles has been shown in theneuropil of cockroaches (Buma and Roubos, 1986).Our study has shown that the ASL-ir neuropil of the

brain and ganglia is very extensive, regular, and usuallyfine textured. Thus, unlike the examples cited above, mostASL-ir interneurons appear not to be general nonsynapticparacrine cells but probably provide parasynaptic releaseof their neuropeptide. The coexistence of small-molecule,fast-acting synaptic transmitters with one or more pep-tides has been reported in many neurons and may becharacteristic of most neurons; in addition, there is evi-dence that parasynaptic release of neuropeptides is activ-ity dependent (see reviews by Hokfelt, 1991; Golding,1994). The demonstration of colocalization of a classic

synaptic transmitter and of a neuropeptide, therefore, maybe an identifying feature of interneurons capable of bothsynaptic modulation and synaptic transmission. It may bethat there is a class of peptidergic paracrine neurons thatlack classic synaptic transmitters, but we suspect thatmost, if not all, of the ASL-ir interneurons also contain aclassical synaptic transmitter.It may be argued, moreover, that the extent of modula-

tory function of an interneuron is reflected in the level ofproduction of its neuropeptide(s) and that interneuronsthat have a high peptide content may be expected to havemajor modulatory functions. Accordingly, many of theASL-ir interneurons appear to serve major modulatoryfunctions.

ASL-ir neuroendocrine cells

We have shown allatostatin-like immunoreactivity inthe IIa4 neuroendocrine cells of the brain and the L3neuroendocrine cells of the abdominal ganglia. All of thesecells also have colocalized diuretic hormone and leucokininimmunoreactivities (Chen et al., 1994b), and, therefore,

Fig. 9. Separation of 500 ventral nerve cords of M. sexta on anEconosil C-18 HPLC column (10 3 250mm, 10 5m;AlltechAssociates,Deerfield, IL). The column was equilibrated in 6.5% acetonitrile and0.1% HFBA in water. After injection of the prepurified sample, thecolumn was eluted for 10 minutes isocratically and, subsequently, by alinear gradient over 60 minutes to 39% acetonitrile and 0.1% in waterat a flow rate of 4 ml per minute. One-minute fractions were collectedand analyzed by competitive ELISA forASL immunoreactivity. Fig. 10. Final purification ofAKSYNFGLamide on a C-18 reversed-

phase column. A single UV-absorbing peak (arrow) contained all ASLimmunoreactivity. AU214: Absorbance units at 214 nm.

TABLE 1. Sequence Analysis of theM. sexta YXFGLamide

Cycle Amino acid Amount found (pmol)

1 Ala 852 Lys 963 Ser 484 Tyr 915 Asn 706 Phe 667 Gly 298 Leu 119 None


the IIa4 and L3 neuroendocrine cells appear to function inthe regulation of fluid transport. Because allatostatin-likeimmunoreactivity is found in these neuroendocrine cellsand in neuroendocrine cells projecting into the cryptone-phridial chamber (L3A8, MA9), YXFGLamides also mayhave functions directly or indirectly related to the regula-tion of fluid transport. Bioassays now are needed todetermine if these peptides do, in fact, have such func-tions.A peptide unrelated to peptides of the YXFGLamide

family has been shown to have an allatostatic function inM. sexta (Kramer et al., 1991). Zitnan et al. (1995) haveshown that several neuroendocrine cells of the pars latera-lis of the protocerebrum are immunoreactive to this allato-static peptide and that these cells terminate in extensivearborizations in the corpora allata. In contrast we havefound that there are very few ASL-ir processes in thecorpora allata; the paucity of ASL-ir innervation of the CAappears to indicate that this innervation has no major rolein the regulation of juvenile hormone biosynthesis in M.sexta larvae. There is a possibility that inhibition of JHbiosynthesis in the larva could be humoral rather thanneural, but it has been shown that the YXFGLamideidentified in this study (AKSYNFGLamide) does not in-hibit JH biosynthesis by the CA ofM. sexta in vitro (P.E.A.Teal, unpublished data).

ASL-ir skeletomotor neurons

Many of the motor neurons innervating the thoracic andmaxillary muscles are ASL-ir, and, except for a few motorneurons innervating the first abdominal segment, none ofthe motor neurons of the abdominal muscles are ASL-ir.This ASL-ir innervation can be seen in all larval stages,and therefore, it may be that YXFGLamides modulate thecontraction of larval thoracic muscles. In cockroaches andlocusts, such a function has been shown for the peptide,proctolin (Adams and O’Shea, 1983; Baines et al., 1990).However, most thoracic and abdominal trunk muscles ofM. sexta larvae have essentially the same locomotor func-tions, and, if YXFGLamides do serve as myomodulators, itseems remarkable that only the thoracic muscles haveASL-ir innervation. It remains to be determined if thisdifference is related to the fact that there is putativerelease of YXFGLamide-like neurohormones in the abdo-men but not in the thorax, and that YXFGLamides,therefore, might not be a suitable agents for modulation ofabdominal muscles.The ASL-ir of the thoracic motor neurons disappears

early in pupal development and does not reappear in themotor neurons of adults. If YXFGLamides serve as myo-modulators, this change could reflect a difference betweenlarval and adult control of the thoracic muscles. It isknown that neurons, which provide slow-type motor inner-vation of larval muscles, change during metamorphosis toprovide fast innervation of the flight muscles of adults(Rheuben and Kammer, 1980), and the loss of YXFGLam-ide immunoreactivity in thoracic motor neurons at meta-morphosis might possibly be related to such a change infunction.Another possibility is that a YXFGLamide serves as a

trophic factor for the thoracic muscles. Myotrophic func-tions have been shown for a number of neuropeptides invertebrates (Hokfelt, 1991), and trophic functions havebeen proposed for the neurosecretory axons associatedwith certain insect muscles (reviewed by Rheuben, 1995).

At the start of metamorphosis the thoracic muscles ofLepidoptera regress and are later replaced by a set of adultmuscles (reviewed by Nuesch, 1985), but many or all of themotor neurons of the adult thoracic muscles persist andare remodeled from those of the larval stage (reviewed byKent et al., 1995). If the thoracic muscles are denervatedjust before metamorphosis, the adult thoracic muscles failto develop normally, and this effect of denervation is takento indicate that the normal development of the adultthoracic muscles depends on a trophic effect provided bytheir motor neurons (Nuesch, 1985; Kent et al., 1995).Therefore, YXFGLamides possibly have a trophic influ-ence on the development of the adult thoracic muscles.Immunostaining indicates that a YXFGLamide is presentin thoracic motor neurons late into the prepupal stage, butthat this peptide disappears from the motor neurons by aday after pupal ecdysis and does not return in the adult.Therefore, if YXFGLamides do exert a trophic effect on thedevelopment of the adult muscles, that effect would haveto be at a very early stage (prepupal) in the development ofthe adult muscles.It is noteworthy that the larval motor neurons of M.

sexta are FMRFamide-ir, that the number of these immu-noreactive motor neurons increases with larval develop-ment, that this immunoreactivity declines duringmetamor-phosis, and that the motor neurons of adults are notFMRFamide-ir (Witten and Truman, 1996). Thus, thepattern of developmental change of inASL-ir immunoreac-tivity of motor neurons is very similar to that of FMR-Famide-ir motor neurons, but, as we have shown, ASL-irmotor neurons are limited to the thoracic ganglia, whereasFMRFamide-ir motor neurons are found in the thoracicand abdominal ganglia (Witten and Truman, 1996).

ASL-ir visceromotor neurons of the hindgut

As in other insects, the rhythmic contractions of thehindgut ofM. sexta are myogenic (Tublitz et al., 1992). Thedemonstration that the PM2A7, PM2A8, and PM3 neuronsareASL-ir and that they terminate on the visceral musclesof the hindgut of M. sexta suggests that YXFGLamidesmight modulate contractions of the hindgut of this species.Inhibition of the rate of hindgut contractions by YXFGLam-ides has been shown in cockroaches and blowflies (Langeet al., 1993, 1995; Duve et al., 1994, 1995a). In a continuingpart of our study of M. sexta, we also have found thatapplication of a physiological concentration (10-7 M) oflepidostatin-1 resulted in inhibition of the rate and ampli-tude of contractions of larval hindgut in vitro (N.T. Davis,unpublished data).Soon after the commencement of the active movements

of gut emptying in the last instar larva, staining of theASL-ir hindgut motor neurons can no longer be detected.Thus, this loss of ASL-ir comes at a time when there maybe no need for inhibition of hindgut movements.

SCP/FMRFamide-ir visceromotorneurons of the hindgut

The array of neurons innervating the proctodeum thatwe have shown by backfilling through the proctodeal nerveis comparable with, but somewhat more extensive than,that described previously by Thorn and Truman (1989)and Tublitz et al. (1992).We have shown that the hindgut is innervated by the

PL1 and PL2 neurons of the TAG and that these cells areSCP- and FMRFamide-ir. The immunoreactivity of these


cells suggests that myogenic hindgut contractions alsomay be modulated by SCP- and/or FMRFamide-like pep-tides. Because SCP and FMRFamide immunoreactivitiesare colocalized in the PL1 and PL2 cells, the questionarises as to whether this immunoreactivity is due to thecolocalization of distinct peptides or the colabeling of thesame peptide(s). Colocalization of SCP and FMRFamideimmunoreactivities often have been observed in arthro-pods, but Arbiser and Beltz (1991) have shown that, inlobsters, immunostaining by the SCP antibody can beabolished by preadsorption with a heptameric FLR-Famide. Therefore, the colocalized SCP and FMRFamideimmunostaining can be due to colabeling of the sameFLRFamide. Similar heptameric FLRFamides have beenidentified from ventral nerve cords of larval M. sexta(Kingan et al. 1996), and so these peptides could beresponsible for the colocalized SCPandFMRFamide immu-nostaining observed in the PL1 and PL2 cells. Colabelingof FLRFamides in these cells, however, does not eliminatethe possibility that the SCP immunostaining could be duein part to an SCP-like peptide. This possibility is morethan theoretical because an SCP-like peptide, CAP2b, hasbeen identified in M. sexta (Huesmann et al., 1995). Thesequence of the last five amino acids of the C-terminus(AFPRVamide) of CAP2b is identical to that of SCPB, exceptthat the terminal amino acid of SCPB is methionine ratherthan valine. Masinovsky et al. (1988) concluded that theepitope of the SCP antibody includes all or part of a sixamino acid sequence located at the C-terminus of SCPs.We have shown that the SCP antibody can be preadsorbedwith synthetic CAP2b; these results probably are due to thesimilarity of the C-terminus of CAP2b to that of the SCPB.Consequently, the SCP immunostaining of the PL1 andPL2 cells could be due to CAP2b.Indirect evidence suggests that the function of the PL1

and PL2 cells is to accelerate the rate of hindgut contrac-tions and that this effect may be mediated by CAP2b and/orFLRFamides. The supporting evidence with regard toCAP2b is as follows. Tublitz and Truman (1985) isolated apeptide fraction known as CAP2, and Tublitz et al. (1992)found that hindgut contractions of larval M. sexta areaccelerated by this CAP2 fraction. The peptide, CAP2b, wasidentified from the CAP2 fraction and has been shown to becardioacceleratory (Huesmann et al., 1995). Therefore, theCAP2b in the CAP2 fraction could contribute to the accelera-tory effect that this fraction has on hindgut contractions.Because the SCP immunoreactivity of the PL1 and PL2cells could be due to CAP2b, these neurons might functionin accelerating the rate of hindgut contractions.The supporting evidence regarding FLRFamides is as

follows. Kingan et al. (1990, 1996) have identified threeextended FLRFamides (Mas-FLRFamide I, II, and III)fromM. sexta, and Mas-FLRFamide II and III were foundto be the most abundant FLRFamides of the ventral nervecord of larvae. In M. sexta hindgut bioassays, these twopeptideswere very effective in accelerating hindgut contrac-tions (Kingan et al., 1996). Because the FMRFamide-ir ofthe PL1 and PL2 cells could be due to Mas-FLRFamide IIand/or III, acceleration of the rate of hindgut contractionsmight be mediated by the release of these FLRFamidesfrom the PL1 and PL2 cells.In any case, the ASL-ir (PM2 and PM3) and SCP/

FMRFamide-ir (PL2 and PL3) visceromotor neurons of thehindgut appear to be antagonistic systems regulatinghindgut contractions. Some of the somata shown by back-

filling the proctodeal nerve were neither ASL-ir or SCP/FLRFamide-ir, and therefore, the control of hindgut move-ments may also involve release of neuroactive substancesother than those that are SCP/FMRFamide- andASL-ir.

Identification of AKSYNFGLamide(lepidostatin-1)

The peptide that we have identified fromM. sexta is yetanother novel member of the YXFGLamide family ofpeptides. It differs from all other YXFGLamides that havebeen isolated from cockroaches, crickets, locusts, blowflies,and mosquitoes (Pratt et al., 1989, 1991; Woodhead et al.,1989, 1994; Duve et al., 1993, 1994, 1995a; Belles et al.,1994; Neuhauser et al., 1994; Weaver et al., 1994; Veelaertet al., 1996; J.A. Veenstra et al., 1997), and it differs fromYXFGLamides predicted from the sequences of theircDNA’s (Donly et al., 1993; Ding et al., 1995; J.A Veenstraet al., 1997).Although only one peptide was identified, there are at

least two, and probably four, more YXFGLamides presentin the ventral nerve cord of M. sexta. It seems very likelythat the same or very similar peptides will be found inother lepidopterans. Therefore, ‘‘lepidostatin-1’’ is pro-posed as the trivial name of this peptide, and it is hopedthat as other lepidopteran YXFGLamides are identified,they can be so named in numerical sequence.The identification of lepidostatin-1 will be helpful in the

isolation of the cDNA encoding it and all other YXFGLam-ides of M. sexta. In addition, the availability of thesynthetic peptide will enable the use of various bioassaysin the study of its functions.The structure of this peptide is very similar to that of

cockroach allatostatins, but designation of it as an allato-statin could cause confusion because an allatostatic pep-tide unrelated to the YXFGLamides has already beenidentified fromM. sexta and has been named ‘‘allatostatin’’(Kramer et al., 1991). Moreover, the YXFGLamide of M.sexta is not be allatostatic inM. sexta (P.E.A. Teal, unpub-lished data). In response to a similar indication thatYXFGLamides of the blowfly Calliphora vomitoria are notallatostatic, these peptides have been named ‘‘callatostat-ins’’ (Duve et al., 1993), and this precedence has beenfollowed by (Veelaert et al., 1997a, 1997b) in the use of‘‘schistostatin’’ for naming the YXFGLamides of Schisto-cerca gregaria. This generic-based method of namingYXFGLamides avoids the implication that these peptidesare allatostatic, but given the very large number of insectgenera and the probability that identical YXFGLamideswill be found even in unrelated genera, the list of peptidesnamed by this method may eventually prove to be veryextensive and full of synonyms. This probability is illus-trated by what is known of the peptides of adipokinetichormone family; so far, the primary structures of approxi-mately 31 different peptides have been reported, theseoccur in one or more of approximately 60 genera (Gade,1996), and the numbers are still growing.Perhaps a suitable alternative to the genera-based

naming of the YXFGLamides would be to use a prefixreferring to the insect order in which the respective‘‘statin’’ (YXFGLamide) has been reported and to followthe ordinal base name with the numerical order of theidentification of the peptides in each order. Such a systemwould limit the list of names applied to the YXFGLamidesto a comprehensible size and minimize the synonymy.


ACKNOWLEDGMENTS

We thank Wallace Clark for the sequence analysis andThomas McClure for the mass spectrometry.

LITERATURE CITED

Adams,M.E., andM. O’Shea (1983) Peptide cotransmitter at a neuromuscu-lar junction. Science 221:286–289.

Agricola, H., K. Schildberger, A. Schmidt, W. Naumann, S. Reissmann, F.Huber, and H. Penzlin (1992) Immunocytochemical distribution ofallatostatin in the nervous system of the cockroach Periplaneta ameri-cana. In N. Elsner and D.W. Richter (eds): Rythmogenesis in Neuronsand Networks. Stuttgart: Thieme, p. 494.

Arbiser, Z.K., andB.S. Beltz (1991) SCPB- and FMRFamide-immunoreactivi-ties in lobster motor neurons: Colocalization of distinct peptides orcolabeling of the same peptide(s)? J. Comp. Neurol. 306:417–424.

Baines, R.A., A.B. Lange, and R.G.H. Downer (1990) Proctolin in theinnervation of locust mandibular closer muscle modulates contractionsthrough the elevation of inositol triphosphate. J. Comp.Neurol. 297:479–486.

Bell, R.A., and F.A. Joachim (1976) Techniques for rearing laboratorycolonies of tobacco hornworms and pink bollworms. Ann. Entomol. Soc.Am. 69:365–373.

Belles, X., J.L. Maestro, M.D. Piulachs, A.H. Johnson, H. Duve, and A.Thorpe (1994) Allatostatic neuropeptides from the cockroach Blattellagermanica (L.) (Dictyoptera, Blattellidae). Identification, immunolocal-ization, and activity. Regul. Pept. 53:237–247.

Bennett, H.P.J., C. A. Browne, and S. Solomon (1981) Purification of twomajor forms of rat pituitary corticotropin using only reversed-phaseliquid chromatography. Biochemistry 20:4530–4538.

Buma, P., and E.W. Roubos (1986) Ultrastructural demonstration ofnon-synaptic release sites in the central nervous system of the snailLymnaea stagnalis, the insect Periplaneta americana and the rat.Neuroscience 17:867–878.

Buys, C.M., and D. Gibbs (1981) The anatomy of neurons projecting to thecorpus cardiacum from the larval brain of the tobacco hornworm,Manduca sexta (L.). Cell Tissue Res. 215:505–513.

Casaday, G.B., and J.M. Camhi (1976) Metamorphosis of flight motorneurons in the mothManduca sexta. J. Comp. Physiol. 112:143–158.

Chen, Y., J.A. Veenstra, N.T. Davis, and H.H. Hagedorn (1994a)A compara-tive study of leucokinin-immunoreactive neurons in insects. Cell TissueRes. 276:69–83.

Chen, Y., J.A. Veenstra, H. Hagedorn, and N.T. Davis (1994b) Leucokininand diuretic hormone immunoreactivity of neurons in the tobaccohornworm,Manduca sexta, and colocalization of this immunoreactivityin lateral neurosecretory cells of abdominal ganglia. Cell Tissue Res.278:493–507.

Christensen, T.A., and J.G. Hildebrand (1987)Male-specific sex pheromone-selective projection neurons in the antennal lobes of the mothManducasexta. J. Comp. Physiol. 160A:553–569.

Copenhaver P., and J.W. Truman (1986) Metamorphosis of the cerebralneuroendocrine system of the moth Manduca sexta. J. Comp. Neurol.249:186–204.

Davis, N.T. (1987) Neurosecretory neurons and their projection to theserotonin neurohemal system of the cockroach Periplaneta americana(L.), and identification of mandibular and maxillary motor neuronsassociated with this system. J. Comp. Neurol. 259:604–621.

Davis, N.T., and J.G. Hildebrand (1992) Vasopressin-immunoreactiveneurons and neurohemal systems in cockroaches andmantids. J. Comp.Neurol. 230:381–394.

Davis, N.T., S.G. Velleman, T.G. Kingan, and H. Keshishian (1989) Identifi-cation and distribution of a proctolin-like neuropeptide in the nervoussystem of the gypsy moth, Lymantria dispar, and in other Lepidoptera.J. Comp. Neurol. 283:71–85.

Davis, N.T., U. Homberg, H. Dircksen, R.B. Levine, and J.G. Hildebrand(1993) Crustacean cardioactive peptide-immunoreactive neurons in thehawkmoth Manduca sexta and changes in their immunoreactivityduring postembryonic development. J. Comp. Neurol. 338:612–627.

Davis, N.T., U. Homberg, P.E.A. Teal, M. Altstein, and J.G. Hildebrand(1996) Neuroanatomy and immunocytochemistry of the neurosecretorysystem of the subesophageal ganglion of the tobacco hawkmoth, Man-duca sexta: Immunoreactivity to PBAN and other neuropeptides.Microsc. Res. Tech. 35:201–229.

Ding, Q., B.C. Donly, S.S. Tobe, andW.G. Bendena (1995) Comparison of theallatostatin neuropeptide precursors in the distantly related cock-

roaches Periplaneta americana and Diploptera punctata. Eur. J. Bio-chem. 234:737–746.

Donly, B.C., Q. Ding, S.S. Tobe, andW.G. Bendena (1993) Molecular cloningof the gene for the allatostatin family of neuropeptides from thecockroach Diploptera punctata. Proc. Natl. Acad. Sci. USA 90:8807–8811.

Duve, H., and A. Thorpe (1994) Distribution and functional significance ofLeu-callatostatins in the blowfly Calliphora vomitoria. Cell Tissue Res.276:367–379.

Duve, H., A.H. Johnsen, A.G. Scott, C.G. Yu, K.J. Yagi, S.S. Tobe, A. Thorpe(1993) Callatostatins: Neuropeptides from the blowfly Calliphora vomi-toria with sequence homology to cockroach allatostatins. Proc. Natl.Acad. Sci. USA 90:2456–2460.

Duve, H., A.H. Johnsen, A.G. Scott, P. East, and A. Thorpe (1994)[Hyp3]Met-callatostatin: Identification and biological properties of anovel neuropeptide from the blowfly Calliphora vomitoria. J. Biol.Chem. 269:21059–21066.

Duve, H., A.H. Johnsen, A.G. Scott, and A. Thorpe (1995a) Isolation,identification and functional significance of [Hyp(2)Met-callatostatinand des Gly-Pro-Met-callatostatin, two further postranslational modifi-cations of the blowfly neuropeptide Met-callatostatin. Regul. Pept.57:237–245.

Duve, H., P. Wren, and A. Thorpe (1995b) Innervation of the foregut of thecockroach Leucophaea maderae and inhibition of spontaneous contrac-tile activity by callatostatin neuropeptides. Physiol. Entomol. 20:33–44.

East, P.D.,A. Thorpe, and H. Duve (1995) Leu-callatostatin gene expressionin the blowflies Calliphora vomitoria and Lucilia cuprina studied by insitu hybridization: Comparison with Leu-callatostatin confocal laserscanning immunocytochemistry. Cell Tissue Res. 280:355–264.

Eaton, J.L. (1982) Exoskeleton and musculature morphology of the larvalthorax of Manduca sexta L. (Lepidoptera: Sphingidae). Ann. Entomol.Soc. Am. 75:313–322.

Gade, G. (1996) The revolution in insect neuropeptides illustrated by theadipokinetic/red pigment-concentrating hormone family of peptides. Z.Naturforsch. 51C:607–617.

Golding, D.W. (1994)Apattern confirmed and refined—Synaptic, nonsynap-tic, and parasynaptic exocytosis. Bioessays 16:503–508.

Hertel, W., and H. Penzlin (1992) Function and modulation of the antennalheart of Periplaneta americana (L.). Acta Biol. Hung. 43:113–125.

Huesmann, G.R., C.C. Cheung, P.K. Loi, T.D. Lee, K.M. Swiderek, and N.J.Tublitz (1995) Amino acid sequence of CAP2b, an insect cardioaccelera-tory peptide from the tobacco hawkmoth Manduca sexta. FEBS Lett.371:311–314.

Hokfelt, T. (1991) Neuropeptides in perspective: The last ten years. Neuron7:867–879.

Homberg, H., N.T. Davis, and J.G. Hildebrand (1991) Peptide-immunohis-tochemistry of neurosecretory cells in the brain and retrocerebralcomplex of the sphinx moth, Manduca sexta. J. Comp. Neurol. 303:35–52.

Kent, K.S., C. Consoulas, K. Duncan, R.M. Johnston, R. Leudeman, andR.B. Levine (1995) Remodelling of neuromuscular systems duringmetamorphosis. Am. Zool. 35:578–584.

Kingan, T.G., D.B. Teplow, J.M. Phillips, J.M. Riehm, K.R. Rao, J.G.Hildebrand, U. Homberg, A.E. Kammer, I. Jardine, P.R. Griffin, andD.H. Hunt (1990) A new peptide in the FMRFamide family isolatedfrom the CNS of the hawkmoth,Manduca sexta. Peptides 11:849–856.

Kingan, T.G., J. Shabanowitz, D.F. Hunt, and J.L. Witten (1996) Character-ization of two myotropic neuropeptides in the FMRFamide family fromsegmental ganglia of the moth Manduca sexta: Candidate neurohor-mones and neuromodulators. J. Exp. Biol. 199:1095–1104.

Kramer, S.J., A. Toschi, C.A. Miller, H. Kataoka, G.B. Quistad, J.P. Li, R.L.Carney, and D.A. Schooley (1991) Identification of an allatostatin fromthe tobacco hornworm Manduca sexta. Proc. Natl. Acad. Sci. USA88:9458–9462.

Lange, A.B., K.K. Chan, and B. Stay (1993) Effect of allatostatin andproctolin on antennal pulsatile organ and hindgut muscle in thecockroach, Diploptera punctata. Arch. Insect Biochem. Physiol. 24:79–92.

Lange, A.B., W.G. Bendena, and S.S. Tobe (1995) The effect of the thirteenDip-allatostatins on myogenic and induced contractions of cockroach(Diploptera punctata) hindgut. J. Insect Physiol. 41:581–588.

Lorenz, M.W., R. Kellner, and K.H. Hoffmann (1995) Identification of twoallatostatins from the cricket, Gryllus bimaculatus de Geer (Ensifera,Gryllidae): Additional members of a family of neuropeptides inhibitingjuvenile hormone biosynthesis. Regul. Pept. 57:227–236.


Martin, D., M.D. Piulachs, and X. Belles (1996) Inhibition of vitellogeninproduction by allatostatin in the German cockroach. Mol. Cell. Endocri-nol. 121:191–196.

Masinovsky, B., S.C. Kempf, J.C. Callaway, and A.O.D. Willows (1988)Monoclonal antibodies to the molluscan small cardioactive peptideSCPB: Immunolabeling of neurons in diverse invertebrates. J. Comp.Neurol. 273:500–512.

Mesce, K.A., K.A. Klukas, and T.C. Brelje (1993) Improvements for theanatomical characterization of insect neurons in whole mount: The useof cyanine-derived fluorophores and laser scanning confocal microscopy.Cell Tissue Res. 271:381–397.

Neuhauser, T., D. Sorge, B. Stay, and K.H. Hoffmann (1994) Responsive-ness of the adult cricket (Gryllus bimaculatus and Acheta domesticus)retrocerebral complex to allatostatin-1 from a cockroach, Diplopterapunctata. J. Comp. Physiol. (B) 164:23–31.

Nuesch, H. (1985) Control of muscle development. In G.A. Kerkut and L.I.Gilbert (eds): Comparative Insect Physiology and Biochemistry. Oxford:Pergamon, pp. 425–452.

O’Donahue, T., J.F. Bishop, B.M. Chronwall, J. Groome, and W.H. WatsonIII (1984) Characterization and distribution of FMRFamide immunore-activity in the rat central nervous system. Peptides 5:563–568.

Pratt, G.E., D.E. Farnsworth, N.R. Seigel, K.F. Fok, and R. Feyereisen(1989) Identification of an allatostatin from adult Diploptera punctata.Biochem. Biophys. Res. Commun. 163:1243–1247.

Pratt, G.E., D.E. Farnsworth, N.R. Seigel, A.L. McCormack, J. Shabanow-itz, D.F. Hunt, and R. Feyereisen (1991) Identity of a second type ofallatostatin from cockroach brains: An octapeptide amide with atyrosine-rich address sequence. Proc. Natl. Acad. Sci. USA 88:2412–2416.

Reichwald, K., G.C. Unnithan, N.T. Davis, H. Agricola, and R. Feyereisen(1994) Expression of the allatostatin gene in endocrine cells of thecockroach midgut. Proc. Natl. Acad. Sci. USA 91:11894–11898.

Reinecke, J.P., B.J. Cook, and T.S. Adams (1973) Larval hindgut ofManduca sexta (L.) (Lepidoptera: Sphingidae). Int. J. Insect Morphol.Embryol. 2:277–290.

Rheuben,M.B. (1995) Specific associations of neurosecretory or neuromodu-latory axons with insect skeletal muscles. Am. Zool. 35:566–577.

Rheuben, M.B., and A.E. Kammer (1980) Comparison of slow larval andfast adult muscle innervated by the same motor neuron. J. Exp. Biol.84:103–118.

Stay, B., K.K. Chan, and A.P. Woodhead (1992) Allatostatin-immunoreac-tive neurons projecting to the corpora allata of adult Diplopterapunctata. Cell Tissue Res. 270:15–23.

Taghert, P.H., and J.W. Truman (1982) Identification of bursicon containingneurons in abdominal ganglia of the tobacco hornworm,Manduca sexta.J. Exp. Biol. 98:385–401.

Taylor, H.M., and J.W. Truman (1974) Metamorphosis of the abdominalganglia of the tobacco hornworm, Manduca sexta: Changes in popula-tions of identified motor neurons. J. Comp. Physiol. 90:367–383.

Thorn, R.S., and J.W. Truman (1989) Sex-specific neuronal respecificationduring metamorphosis of the genital segments of the tobacco hornwormmothManduca sexta. J. Comp. Neurol. 284:489–503.

Truman, J.W. (1973) Physiology of insect ecdysis. III. Relationship betweenthe hormonal control of eclosion and tanning in the tobacco hornworm,Manduca sexta. J. Exp. Biol. 57:805–816.

Tublitz, N.J., and J.W. Truman (1985) Identification of neurons containingcardioactive peptides (CAPs) in the ventral nerve cord of the tobaccohawkmoth,Manduca sexta. J. Exp. Biol. 116:395–410.

Tublitz, N.J., A.T. Allen, C.C. Cheung, K.K. Edwards, D.P. Kimble, P.K. Loi.and A.W. Sylwester (1992) Insect cardioactive peptides: Regulation ofhindgut activity by cardioacceleratory peptide 2 (CAP2) during wander-ing behavior inManduca sexta larvae. J. Exp. Biol. 165:241–264.

Tyrer, N.M., N.T. Davis, E.A. Arbas, K.J.S. Thompson, and J.P. Bacon(1993) Morphology of vasopressin-like immunoreactive (VPLI) neuronsin many species of grasshopper. J. Comp. Neurol. 329:385–401.

Veelaert, D., L. Schoofs, S.S. Tobe, C.G. Yu, H.G.B. Vullings, F. Couillaud,and A. De Loof (1995) Immunological evidence for an allatostatin-likeneuropeptide in the central nervous system of Schistocerca gregaria,Locusta migratoria, andNeobellieria bullata. Cell Tissue Res. 279:601–611.

Veelaert, D., B. Devreese, L. Schoofs, J. Van Beeumen, J. Vanden Broeck,S.S. Tobe, and A. De Loof (1996) Isolation and characterization of eightmyoinhibiting peptides from the desert locust, Schistocerca gregaria:New members of the cockroach allatostatin family. Mol. Cell. Endocri-nol 122:183–190.

Veelaert, D., B. Devreese, J. Vanden Broek, C.G. Yu, L. Schoofs, J. VanBeeumen, S.S. Tobe, and A. De Loof (1997) Isolation and characteriza-tion of schistostatin-211–18 from the desert locust, Schistocerca gregaria:A truncated analog of schistostatin-2. Regul. Pept. 67:195–199.

Veenstra, J.A. (1994) Isolation and identification of three leucokinins fromthe mosquito Aedes aegypti. Biochem. Biophys. Res. Commun. 202:715–719.

Veenstra, J.A., and H.H. Hagedorn (1991) Identification of neuroendocrinecells producing a diuretic hormone in the tobacco hornworm moth,Manduca sexta. Cell Tissue Res. 266:359–364.

Veenstra, J.A., and G. Lambrou (1995) Isolation of a novel RFamide fromthe midgut of the American cockroach, Periplaneta americana. Bio-chem. Biophys. Res. Commun. 213:519–524.

Veenstra, J.A., G.W. Lau, H.J.Agricola, and D.H. Petzel (1995) Immunohis-tochemical localization of regulatory peptides in the midgut of thefemale mosquito Aedes aegypti. Histochem. Cell Biol. 104:337–347.

Vitzthum, H., U. Homberg, and H. Agricola (1996) Distribution of Dip-allatostatin I-like immunoreactivity in the brain of the locust Schisto-cerca gregaria with detailed analysis of immunostaining in the centralcomplex. J. Comp. Neurol. 369:419–437.

Wall, J.B., and P.H. Taghert (1991) The timing of initial neuropeptideexpression by an identified insect neuron does not depend on interac-tions with its normal peripheral target. J. Neurobiol. 22:935–956.

Weaver, R.J., Z.A. Freeman, M.G. Pickering, and J.P. Edwards (1994)Identification of two allatostatins from the CNS of the cockroachPeriplaneta americana: Novel members of the neuropeptide inhibitorsof insect juvenile hormone biosynthesis. Comp. Biochem. Physiol. (C)107:119–127.

Witten, J.L., and J.W. Truman (1996) Developmental plasticity of neuropep-tide expression in motoneurons of the moth, Manduca sexta: Steroidhormone regulation. J. Neurobiol. 29:99–114.

Woodhead, A.P., B. Stay, S.L. Seidel, M.A. Kahn, and S.S. Tobe (1989)Primary structure of four allatostatins: Neuropeptide inhibitors ofjuvenile hormone synthesis. Proc. Natl. Acad. Sci. USA 86:5997–6001.

Woodhead, A.P., M.A. Khan, B. Stay, and S.S. Tobe (1994) Two newallatostatins from the brains of Diploptera punctata. Insect Biochem.Mol. Biol. 24:257–263.

Yoon, J.G., and B. Stay (1995) Immunocytochemical localization of Diplop-tera punctata allatostatin-like peptide in Drosophila melanogaster. J.Comp. Neurol. 363:475–488.

Yu, C.G., B. Stay, Q. Ding, W.G. Bendena, and S.S. Tobe (1995) Immuno-chemical identification and expression of allatostatins in the gut ofDiploptera punctata. J. Insect Physiol. 41:1035–1043.

Zitnan, D., T.G. Kingan, S.J. Kramer, and N.E. Beckage (1995) Accumula-tion of neuropeptides in the cerebral neurosecretory system ofManducasexta larvae parasitized by the braconid wasp Cotesia congregata. J.Comp. Neurol. 356:83–100.


Documents

Allatostatin-like-immunoreactive neurons of the tobacco hornworm,Manduca sexta, and isolation and identification of a new neuropeptide related to cockroach allatostatins