Cellular Distributions and Functions of Histamine,Octopamine, and Serotonin in the Peripheral Visual System,Brain, and Circumesophageal Ring of the Horseshoe CrabLimulus polyphemusB.-A. BATTELLE,* B.G. CALMAN, AND M.K. HARTWhitney Laboratory and Department of Neuroscience, University of Florida, St. Augustine, Florida 32086

KEY WORDS biogenic amines; histamine; serotonin; octopamine; retina; photoreceptors; circa-dian rhythms; neuromodulation

ABSTRACT The data reviewed here show that histamine, octopamine, and serotonin areabundant in the visual system of the horseshoe crab Limulus polyphemus. Anatomical andbiochemical evidence, including new biochemical data presented here, indicates that histamine is aneurotransmitter in primary retinal afferents, and that it may be involved in visual informationprocessing within the lateral eye. The presence of histamine in neurons of the central nervoussystem outside of the visual centers suggests that this amine also has functions unrelated to vision.However, the physiological actions of histamine in the Limulus nervous system are not yet known.Octopamine is present in and released from the axons of neurons that transmit circadianinformation from the brain to the eyes, and octopamine mimics the actions of circadian input onmany retinal functions. In addition, octopamine probably has major functions in other parts of thenervous system as octopamine immunoreactive processes are widely distributed in the centralnervous system and in peripheral motor nerves. Indeed, octopamine modulates functions of theheart and exoskeletal muscles as well as the eyes. A surprising finding is that although octopamineis a circulating neurohormone in Limulus, there is no structural evidence for its release into thehemolymph from central sites. The distribution of serotonin in Limulus brain suggests this aminemodulates the central processing of visual information. Serotonin modulates cholinergic synapses inthe central nervous system, but nothing further is known about its physiological actions. Microsc.Res. Tech. 44:70–80, 1999. r 1999 Wiley-Liss, Inc.

INTRODUCTIONThe horseshoe crab Limulus polyphemus is used

extensively by neurobiologists to study basic mecha-nisms of vision and the modulation of visual functionsby a circadian clock. Other studies have focused on theamine modulation of neuronal functions in severalpreparations including the central nervous system(CNS), neuromuscular junctions of the walking legs,and the neurogenic heart. Therefore, most studies ofthe distribution of the biogenic amines histamine (HA),octopamine (OA), and serotonin (5HT) in Limulus focuson the peripheral visual system, the protocerebrum(brain), which contains the major visual centers, andthe circumesophageal ring (CER), which is the source ofprocesses that innervate the heart and the walkinglegs.

We review here current knowledge of the cellulardistributions of HA, OA, and 5HT in the peripheralvisual system, brain, and CER of Limulus, and presentnew data concerning HA uptake into and release fromcentral visual tissues of Limulus. The known andproposed functions of these amines are discussed, and,where possible, results from Limulus are comparedwith those from other chelicerate arthropods.

MATERIALS AND METHODSAdult intermolt horseshoe crabs, collected in the

Indian River near Oak Hill, Florida, were maintainedin running, natural seawater (15–18°C) on a 12 hourlight:12 hour dark cycle. Animals were fed once a weekand adapted to these conditions for at least 2 weeksbefore they were used. Unless otherwise specified,reagents were purchased from Sigma Chemical Co. (St.Louis, MO) or Fisher Scientific (Pittsburgh, PA). Immu-nocytochemistry and Neurobiotin tract tracing proto-cols have been described in detail elsewhere (Battelle etal., 1991; Calman and Battelle, 1991; Chamberlain etal., 1986; Lee and Wyse, 1991).

For autoradiographic analyses of HA uptake andbiochemical analyses of HA release, tissues were dis-sected from the animal and desheathed during themid-afternoon in normal Limulus saline (Chamberlainet al., 1986) and then incubated overnight at 12°C in

*Correspondence to: B.-A. Battelle, Whitney Laboratory, 9505 Ocean ShoreBlvd., St. Augustine, FL 32086. E-mail: battelle@whitney.ufl.edu

Received 14 November 1997; accepted in revised form 6 March 1998Contract grant sponsor: Neuroscience Programs and Research Experience for

Undergraduates Program of the National Science Foundation; Contract grantsponsor: National Institutes of Health.

MICROSCOPY RESEARCH AND TECHNIQUE 44:70–80 (1999)

r 1999 WILEY-LISS, INC.

Limulus saline containing 20 µCi/ml 3H-HA (NEN-Dupont, Boston MA, 5.8 Ci/mmol). The final concentra-tion of HA in the saline was 3.4 µM. The next morning thetissues were rinsed for 1 hour with frequent changes ofsaline (Battelle, 1980) alone or saline containing 2 x 10-4 Mpyrilamine, a drug that blocks HA uptake into terminalsof barnacle photoreceptors (Stuart and Mekeel, 1990).

To determine whether pyrilamine blocks HA uptakein Limulus, two laminae were preincubated for 30minutes in saline containing 2 x 10-4 M pyrilamine, andthen 3H-HA (20 µCi/ml) was added and the incubationcontinued for 4 hours. At the end of the incubation, thetissues were rinsed briefly, extracted in acid (Battelle,1980) and an aliquot of the acid extract was analyzed byliquid scintillation spectroscopy (LKB/Wallac 1214RACKBETA Counter, Gaithersburgh, MD). The acidinsoluble pellet was solubilized in 1M NaOH and analiquot of this was used to determine total tissueprotein (Lowry et al., 1951).

Sites of 3H-HA uptake and storage were identified inbrains that had been incubated with saline containing3H-HA and rinsed in saline plus pyrilamine. Brainswere fixed overnight at 4°C in 0.1 M phosphate buffer,pH 7.2 containing 4.5% sucrose, 3% NaCl, 5% parafor-maldehyde, and 1% glutaraldehyde (Polysciences Inc., War-rington, PA), rinsed in 0.1M phosphate buffer, pH 7.2,containing 8% sucrose, then incubated for 1 hour in therinse buffer containing 1% OsO4. The tissue was rinsedfurther in cold distilled water, dehydrated and embed-ded in Epon. Sections (2.5 µm thick) were dipped inNTB2 photographic emulsion (Kodak, Rochester, NY), ex-posed for 2 weeks at 4°C, then developed in D-19 (Kodak).

3H-HA release was analyzed from pairs of laminaeincubated together with 3H-HA. After the tissues wererinsed for 1 hour as described above, the volume of thesaline rinse was reduced to 200 µl and collected andreplaced at 5-minute intervals. The amount of radioac-tivity released into the saline during each 5-minuteinterval was determined by analyzing an aliquot byliquid scintillation spectroscopy. To depolarize the prepa-ration, the concentration of KCl in the saline wasincreased to 100 or 200 mM as indicated in Figures 1through 7, and the amount of NaCl was reducedcomparably to maintain osmotic strength. To blockCa11 influx, the Ca11 concentration in the saline wasreduced from 26 to 1 mM and 40 mM CoCl2 was added.Experiments were done in the presence or absence ofpyrilamine as indicated in Figures 1–7.

Radioactive substances stored in tissues or releasedinto the saline were separated by either high voltagepaper electrophoresis (Battelle et al., 1991) or thinlayer chromatography (TLC) using silica-gel G250 mi-cron glass backed plates (Universal Scientific Inc. At-lanta, GA) and chloroform:methanol:ammonia (2:2:1)as the solvent system. Histamine and its potentialmetabolites (imidazol acetic acid, N-acetyl histamine,and g-glutamyl histamine) were visualized with sulfa-nilic acid spray. The distribution of radioactivity alongthe electropherogram was determined as describedpreviously (Battelle, 1980). To recover the radioactivityfrom the TLC plates, Strip Mix (Alltech Associates, Inc.,Deerfield, IL) was spread thinly over the plates andpermitted to dry so that the silica could be peeled from

the plates. The sample lanes, 3 cm wide, were cut into0.5-cm sections from the origin to the solvent front, andeach section was placed into a scintillation vial andincubated for at least 1 hour in 1 M HCl before liquidscintillation fluid (ScintiVerse BD) was added and thesamples were analyzed. Recovery of radioactivity fromthe plates was about 30%, and it varied within 5% fromlane to lane on the same plate.

RESULTS AND DISCUSSIONLimulus brain structures are described using the

nomenclature of Chamberlain and Wyse (1986). Theterms of Patten (1912) and Bullock and Horridge (1965)are used to describe regions of the CER.

HistamineDistribution in the Peripheral Visual System.

HA-immunoreactivity (IR) is abundant in structuresof the peripheral visual system. In the lateral eye,intense HA-IR is seen in photoreceptor cell bodies andin eccentric cell bodies, dendrites, and processes (Fig.1A and E). Eccentric cells are electrically coupled tophotoreceptors, depolarize and generate action poten-tials in response to light, and project through the opticnerve to the brain (Hartline et al., 1952). The presenceof HA-IR in the eccentric cell axon collaterals that formthe lateral plexus of the lateral eye suggests that HAhas a role in lateral inhibition, a fundamental processin vision that is mediated by eccentric cell processes inthe lateral plexus and that enhances contrast at edges(Hartline et al., 1952; Fahrenbach, 1985). All cells inthe photoreceptor cell layer of the median eye showHA-IR. This layer contains cell bodies of UV light- andvisible light-sensitive photoreceptors and the arhabdo-meric cells. Arhabdomeric cells of the median eye, likethe eccentric cells of the lateral eye, are electricallycoupled to photoreceptors, depolarize and generateaction potentials in response to light, and project to thebrain (Nolte and Brown, 1969, 1972).

Intense HA-IR is also detected over photoreceptorcell bodies and axons of the ventral eye (not shown;Battelle et al., 1991), which consists of large, bilobedphotoreceptors scattered along optic nerves that extendanteriorly from the ventral side of the brain and end ata wart-like structure on the ventral cuticle just in frontof the mouth. The ventral eye is one of three types ofrudimentary eyes with structurally similar photorecep-tors (Fahrenbach, 1970; Calman and Chamberlain,1982). The rudimentary eyes appear early in Limulusdevelopment, before the development of the more com-plex lateral compound eyes and median ocelli (Watase,1889). Lateral rudimentary eyes are located at theposterior edge of each lateral compound eye, and themedian rudimentary eye is located under the carapacebetween the median ocelli. The projections of the rudi-mentary photoreceptors will be important to considerwhen analyzing the distribution of HA-IR and sites ofHA uptake in the brain.

Distribution in Central Visual Structures. Muchof the HA-IR in the brain (Figs. 2A and 3) correlateswith known projections of photoreceptors and second-ary visual neurons (Fig. 4). HA-IR is intense in thelamina, which contains the terminals of photoreceptorsfrom the lateral eye, and in the ocellar ganglion, which-

71BIOGENIC AMINES IN LIMULUS

Fig. 1.

contains the terminals of photoreceptors from the me-dian eyes. The lamina and the ocellar ganglion alsocontain terminals of axon collaterals of the HA-immunoreactive eccentric cells from the lateral andarhabdomeric cells from the median eyes, respectively.Projections of these secondary visual cells can alsoaccount for the HA-immunoreactive processes observedin the medulla and the optic tract. Regions of intenseHA-IR seen in sections that include the ventral medulla(not shown) contain terminals of ventral and lateralrudimentary photoreceptors. The region of the centralbody where the median rudimentary photoreceptorsterminate has not been examined for HA-IR.

Distribution Outside of the Visual System. Onlythree groups of HA-immunoreactive cell bodies areobserved outside of the visual centers of the brain (Fig.2A): three pair of large cells (65 µm diameter) withinthe curve of the central body among cells of the dorsalmedian group, small (9 µm diameter) cells scatteredwithin the anterior portion of the ganglion cell layer ofthe central body, and a cluster of 15–20 cells (20 µmdiameter) in the cheliceral ganglion. Initial portions ofthe axons of HA-immunoreactive cells in the dorsalmedian group project in an posterior-lateral direction,and cells in the cheliceral ganglion project toward theprotocerebrum. None of these processes has been tracedto its termination. Other labeled structures include finebeaded processes (0.5 µm diameter) in the neuropil ofthe central body and find beaded processes (0.5 µmdiameter) and processes with more uniform diameters,ranging from 1–2 µm, in the neuropil posterior to thecentral body and extending into the circumesophagealconnectives. No HA-immunoreactive processes are de-tected in the neuropil of the corpora pedunculata.

HA Uptake Into and Release From Visual Cen-ters of the Brain. The distribution of silver grainsshowing sites of 3 H-HA uptake and storage is similar tothe distribution of HA-IR (compare Figs. 5 and 3A). Thedensity of silver grains is highest over the lateral opticnerve and lamina. The medulla, although less intenselylabeled overall, contains heavily labeled fiber tracts.There is also an intense accumulation of silver grainson the ventral side of the medulla where lateral rudi-mentary photoreceptors terminate (Calman et al., 1991).Background labeling is seen over the corpora peduncu-lata. More than 90% of the radioactivity that accumu-lates in laminae incubated and rinsed exactly like thetissue analyzed for autoradiography is identified bio-

Fig. 1. HA-immunoreactive structures in lateral and median eyes.A: Lateral eye ommatidia. Frozen section of two lateral eye ommatidiacut parallel to the cuticle near the base of the rhabdom of the retinularcell (compare with B). Retinular cell labeling is observed at theperiphery of the ommatidia, but obscured by pigment closer to thecenter. The eccentric cell dendrite and cell body are labeled.B: Schematic of the structure of an ommatidium of the lateral eye cutparallel to the cuticle. The eccentric cell body is not included in thisdrawing. C: Longitudinal section through the median ocellus (comparewith D). HA-IR is seen in cells of the photoreceptor layer and inphotoreceptor axons. D: Schematic longitudinal section through amedian ocellus. E: Transverse section at the level of the lateral plexusbelow the ommatidia (compare with F). Labeled fiber bundles run intoand between glomeruli, and the central portion of each glomerulus isstrongly labeled. This section was taken from tissue that was incu-bated overnight in organ culture medium before fixation. A similar,although less intense, pattern of immunostaining was seen in tissuesfixed immediately after dissection. F: Schematic of a transversesection of the lateral eye showing the lateral plexus below theommatidia. AX, axons; ECB, eccentric cell body; ECD, eccentric celldendrite; G, glomerulus; OL, position of the ocellar lens; P, photorecep-tor cells; PF, plexus fibers; R, retinular cell. Scale bar 5 50 µm. A, C,and E are reproduced from Battelle B-A, Calman BG, Rews AW GriecoFD Mleziva MB Callaway JC, Stuart AE. 1991. Histamine: A putativeafferent neurotransmitter in Limulus eyes. J Comp Neurol 305:527–542. with permission of the publisher; F is modified from Bullock andHorridge (1965).

Fig. 2. Schematic of major (A) HA-, (B)OA-, and (C) 5HT-immunoreactive structuresin brain and CEG. See text for details. CB,central body; ChelG, cheliceral ganglion;ChilG, chilarial ganglion; CP, corpora pedun-culata; DHN, dorsal hemal nerve; DMG, dor-sal median group; L, lamina; LON, lateraloptic nerve; M, medulla; MON, median opticnerve; OG, ocellar ganglion; OperG, opercu-lar ganglion; OT optic tract; PCB, protocere-bral bridge; PN, pedal nerve; POC, prosomal-opisthosomal connective; VON, ventral opticnerve. A is drawn from the results of Battelleet al. (1991). B is modified from Lee and Wyse(1991). C is modified from Chamberlain et al.(1996).

73BIOGENIC AMINES IN LIMULUS

chemically as 3 H-HA; the remaining radioactivity isassociated with the HA metabolites g-glutamyl HA andimidazol acetic acid (Hart and Battelle, 1991).

Depolarizing the lamina with a high concentration ofextracellular K1 in the presence of pyrilamine producesan enhanced efflux of radioactivity (Fig. 6A). LittleK1-stimulated increase in the efflux of radioactivity is

observed in the absence of pyrilamine (Fig. 6B). Prelimi-nary studies showed that pyrilamine blocks 80% of theuptake of 3 H-HA into the lamina. Ca11-dependence ofthe K 1 -stimulated HA release is shown in Figure 6C.Chromatographic analysis (Fig. 6D) demonstrates thatK1 stimulated the efflux of 3H-HA and not its metabo-lites.

Fig. 3. HA-IR in optic ganglia. A: Terminals of the lateral opticnerve. HA-immunoreactive processes in the lateral optic nerve projectto the lamina, medulla, and the optic tract. All staining distal to thelamina can be accounted for by axons in the lateral optic nerve.B: Projections of the median optic nerve. The nerve and the ocellarganglion are brightly labeled. C, chiasma; L, lamina; LON, lateral

optic nerve; M, medulla; MON, median optic nerve; OG, optic gan-glion; OT, optic tract. Scale bar 5 50 µm. Reproduced from BattelleB-A, Calman BG, Rews AW Grieco FD Mleziva MB Callaway JC,Stuart AE. 1991. Histamine: A putative afferent neurotransmitter inLimulus eyes. J Comp Neurol 305:527–542, with permission of thepublisher.

74 B.-A. BATTELLE ET AL.

Functions. To establish that a molecule is a neuro-transmitter at a particular synapse synthesis, storage,release, removal, and physiological actions of the mol-ecule must be demonstrated. The immunocytochemicaland biochemical studies reviewed here show that HA issynthesized and stored in Limulus retinal afferents. Asimilar association of HA with photoreceptor cell bodiesand axons is observed in the spider Cupiennius salei(see Schmid and Becherer, this issue), another chelicer-ate arthropod, and in insects and crustaceans (Nassel,1988; Callaway and Stuart, this issue). Thus, thepresence of HA in photoreceptors is a general feature ofarthropods.

New data presented here show that depolarizationenhances the Ca11-dependent release of HA from thelamina in Limulus, and that the lamina and medullatake up HA with apparent high affinity. Since silvergrains are concentrated over the visual centers ofbrains incubated with a low concentration of 3H-HA,and the distribution of silver grains in the visualcenters is similar to the distribution of HA-IR, wepropose that HA is taken up into photoreceptor andeccentric cell processes and terminals. High affinityuptake may provide a mechanism for removing HAfrom the synapse. In this regard, Limulus appears moresimilar to the barnacle, where there is unambiguousevidence for HAuptake into the terminals of photorecep-tors (Callaway and Stuart, this issue), than to insects,where HA uptake is into glia (Elias and Evans, 1984;Sarthy, 1991). Presynaptic uptake of HA is also lackingin mammals (Schwartz et al., 1991). In Limulus, it isparticularly interesting that depolarization-induced re-lease of HA is detected only in the continuous presence

of an HA uptake blocker. One interpretation of thisfinding is that in the absence of the blocker, avidreuptake of HA into presynaptic sites prevents theaccumulation of HA in the saline bath. Convincingphysiological evidence that HA is a photoreceptor neu-rotransmitter comes from studies done in barnaclesand flies (Callaway and Stuart, this issue). No compa-rable evidence exists for Limulus since, in this animal,the direct central effects of activating visual afferentpathways remain unclear (Snodderly, 1971). A moreapproachable and equally interesting question in Limu-lus, that has not yet been addressed in a systematicway, is whether HA mediates lateral inhibition in thelateral eye.

The functions of the three groups of HA-immunoreac-tive cell bodies observed outside of the visual centers ofthe Limulus brain are not known. The three pairs of cellbodies within the curve of the central body in Limulusmay be structurally homologous to three pairs of HA-immunoreactive cells found in a comparable location inthe spider (Schmid and Duncker, 1993; Schmid andBecherer, this issue). In the spider, these cells arethought to modulate central sensory and motor path-ways. The two other groups of HA cells in Limulus,those in the cheliceral ganglia and those within theganglion cell layer of the central body appear not tohave counterparts in most spiders. However, in the birdspider Psalmopoeus cambridgei, HA-immunoreactivecells have been observed in the cheliceral ganglia, andthese may be analogous to the ones found in a similarlocation in Limulus (Schmid and Becherer, this issue).The functions of these cells in P. cambridgei are not yetknown.

Fig. 4. Schematics of the projections ofphotoreceptors and secondary visual cells asderived from results of studies of the distribu-tion of a photoreceptor specific antigen(Calman et al., 1991) and from cobalt backfillsof optic nerves (Chamberlain and Barlow,1980). A: Photoreceptors of the lateral rudi-mentary eye project to the ganglion cell layeron the ventral side of the medulla. The projec-tions of several cells are shown schematicallyto indicate the area innervated. The actualnumber of cells is much higher. B: Retinularcells terminate in the lamina. C: Eccentriccells project to the lamina, medulla, and optictract. D: Photoreceptors of the median ocelliterminate in the ocellar ganglion. E: Photore-ceptors of the median rudimentary eye inner-vate the ocellar ganglion then continue intothe optic tract to terminate near the centralbody. F: Arhabdomeric cells of the median eyeproject to the ocellar ganglion then throughthe optic track to terminate in the medulla.Reproduced from Calman BG, Lauerman MA,Rews AW, Schmidt M, Battelle B-A. 1991.Central projections of Limulus photoreceptorcells revealed by a photoreceptor-specificmonoclonal antibody. J Comp Neurol 313:553–562, with permission of the publisher. Forabbreviations, see Figure 2.

75BIOGENIC AMINES IN LIMULUS

OctopamineDistribution. The distribution of OA-IR in Limulus

brain and CER at the level of the light and the electronmicroscope was described by Lee and Wyse (1991).Results of their light microscopic studies are summa-rized in Figure 2B. They identified eight bilateral pairsof OA-immunoreactive cell clusters, each containingone to two dozen neurons ranging in size from 40–100µm. The first, most anterior cluster pair is locatedlaterally within the cheliceral ganglia; the second mostanterior cluster pair lies within the first of five pedalganglia that innervate the walking legs. Cluster pairsnos. 3–5 border the esophageal opening and cluster

pairs nos. 6–8 lie in the chilarial/opercular region of theCER near the midline. All of the OA-immunoreactivecells are below the dorsal surface of the CER. Clustersnos. 3–8 are relatively compact and centered aboutmidway through the ring. The cells in the cheliceral andpedal ganglia have a wider dorso-ventral distribution.In CER preparations with a total thickness of 2–2.5mm, the cells were found from about 1.0 to 1.6 mmbelow the dorsal surface.

OA-immunoreactive processes within the pedal gan-glia run in an anterior-posterior direction throughoutthe CER neuropil. Coarse immunoreactive axons crossthe midline through the post-esophageal commissures,

Fig. 5. LM autoradiographic analysis ofsites of 3H-HA uptake and storage in thelamina and medulla. Brains were incubated,rinsed, fixed, and processed for LM autoradi-ography as described in Materials and Meth-ods. The most heavily labeled regions are theincoming lateral optic nerve (LON), the lamina(L), and the region of the medulla (M) wherethe lateral rudimentary photoreceptors termi-nate (LRP). Most of the labeling over themedulla appears concentrated over fibertracts, which is similar to the appearance ofthe HA-IR in this structure.

76 B.-A. BATTELLE ET AL.

and numerous processes project out of the CER into theventral nerve cord and abdominal ganglia. Large fibertracts also enter the brain posteriorly through thecircumesophageal connective and some cross to thecontralateral side. Some of these processes probablyarise from the cell clusters in the cheliceral ganglia (seebelow). OA-immunoreactive processes are widely dis-tributed within the brain itself with the heaviest label-ing found in the optic ganglia (lamina and medulla) andthe central body. The neuropil of the corpora peduncu-lata is moderately to sparsely innervated, and a loosebundle of OA-immunoreactive fibers is detected in eachlateral optic nerve.

Examination of OA-immunoreactive processes inLimulus brain at the level of the electron microscopeled to the interesting finding that they can be identifiedunequivocally by the presence of large, unique, crystal-

line, cylindrical granules (100–150 by 150–400 nm)with a prominent indentation at one end. These gran-ules accumulate at synapses together with clear vesicles,but their function is not known. Because of theseunusual granules, OA-containing axons and terminalscan be distinguished from all other axons and terminalseven in the absence of antibody labeling.

Functions as a Modulator of Vision. The large,unique granules characteristic of OA-immunoreactiveaxons and processes in Limulus CNS are also character-istic of the axons and terminals of neurons that projectfrom the CNS to all of the eyes of Limulus. Theseretinal efferent neurons innervate all cell types in thelateral compound eye (Fahrenbach 1975, 1981), themedian eye (Fahrenbach, 1975), lateral rudimentaryphotoreceptors (Fahrenbach, 1970), and the rhabdom ofventral photoreceptors (Clark et al., 1965; Battelle et

Fig. 6. Potassium stimulated release of radioactivity from laminaeincubated with 3H-HA. Isolated laminae were incubated together with3H-HA as described in Materials and Methods and rinsed withfrequent changes of saline until the efflux of radioactivity stabilized.The radioactivity plotted is the total radioactivity released into 200 µlof saline during a 5-minute interval. A: Release measured in thecontinued presence of 2 x 10-4M pyrilamine. Net efflux of radioactivitywas increased when laminae were challenged with an elevatedconcentration of extracellular K1 . B: Release measured in the absenceof pyrilamine. Net efflux of radioactivity did not reproducibly increase

when laminae were challenged with elevated K1. C: The K1-stimulated increase in the net efflux of radioactivity from laminae wasinhibited in the presence of Co11 and a reduced concentration of Ca11.D: TLC analysis of the radioactive molecules released into fractionnos. 9 and 10 in the experiment shown in A (*). The distribution ofradioactivity in the chromatograph was determined as detailed inMaterials and Methods. In both fractions only histamine (HA), and nothistamine metabolites (M), were detected. Thus, the enhanced netefflux of radioactivity induced by elevated K1 is due to the efflux of3H-HA.

77BIOGENIC AMINES IN LIMULUS

al., 1982; Evans et al., 1983). Retinal efferents are ofparticular interest because they are activated by anendogenous circadian clock and drive dramatic changesin the structure and function of the eyes that result inan increase in retinal sensitivity and responsiveness tolight during the animal’s subjective night (Barlow andChamberlain, 1980). Biochemical and autoradiographicstudies have demonstrated unequivocally that retinalefferent neurons synthesize, store, and release OA, andphysiological studies demonstrate that OAmimics manyof the effects of the endogenous efferent neurotransmit-ter (Battelle, 1991). Thus, there is convincing evidencethat OA is a neurotransmitter in these retinal effer-rents.

The neuronal cell bodies that give rise to retinalefferents were located by backfilling the optic nerveswith Neurobiotin (Calman and Battelle, 1991). Back-fills of each type of optic nerve, lateral, median, andventral, consistently stain cells, between 40 and 80 µmin diameter, located in the cheliceral ganglion ipsilat-eral and contralateral to the side of the fill, aboutmidway between the dorsal and ventral surfaces (Fig.7A). These cells match the OA-immunoreactive cellsdescribed by Lee and Wyse (1991) in number, size andlocation. Thus the most complete backfills, like the onereconstructed in Figure 7A, provide detailed informa-tion about the projections of one subset of OA-immunoreactive cells. A number of inferences can bemade from these results about the source of OA-immunoreactive processes observed in different parts ofthe brain and CER.

Many of the OA-immunoreactive processes in Limu-lus brain are probably projections of the retinal efferentneurons. These include the OA-immunoreactive pro-cesses that enter the brain posteriorly through thecircumesophageal connective, cross the midline belowthe central body, and are present in the lamina andmedulla, and within the lateral optic nerves. Neurobio-tin backfills show that efferent axons filled from anyone optic nerve have branches that project to the othertypes of optic nerves. This suggests that retinal efferentneurons branch widely within the brain and that asingle neuron innervates more than one eye. Further-more, filling any one of the three types of optic nerveson one side of the animal consistently labels axons andcells on both sides of the CER, showing that a singleeye, of any type, is innervated by cells located in boththe ipsilateral and contralateral cheliceral ganglion.The proposed major projections of a single OA-immuno-reactive retinal efferent neuron are diagrammed inFigure 7B. It is not clear from the Neurobiotin fillswhether retinal efferent neurons give rise to the OA-immunoreactive processes within the neuropil of thecorpora pedunculata, but the OA-immunoreactive tractswithin the neuropil of the CER and those projecting outthe pedal or dorsal hemal nerves or into the ventralnerve cord probably do not arise from these cells.

Other Functions. The wide distribution of OA-immunoreactive processes in the brain and CER sug-gests that this amine has major functions throughoutthe CNS where it may modulate, among other things,the central motor pattern for feeding (Lee and Wyse,1986). The many projections of OA-immunoreactiveprocesses toward the periphery, through the pedalnerves, which innervate the muscles of the walking

legs, and the dorsal hemal nerves, which innervate theheart, guts, glands, and sensory hairs, suggest that OAhas a wide variety of effects in the periphery as well.Indeed, effects of OA on the heart and exoskeletalmuscle have been demonstrated (Augustine and Fet-terer, 1985; Augustine et al., 1982; Rane et al., 1984),and OA has a role in osmoregulation (Edwards andPierce, 1986). OA is also a circulating hormone inLimulus (Edwards and Pierce, 1986). Therefore, it issurprising that none of the OA-containing terminalsidentified in the CNS are located near blood spaces (Leeand Wyse, 1991). These findings suggest that in Limu-lus, circulating OA enters the blood from peripheralsites.

Comparison of the Distribution of OA in Limu-lus With That in Scorpions and Spiders. There aresimilarities, but there are also major differences. Reti-nal efferent neurons that probably contain OA arefound in the scorpion Androctonus australis (Fleissner,1983; Fleissner and Heinrichs, 1982; Heinrichs andFleissner, 1987) but not in the spider C. salei (Seyfarthet al., 1993). A circadian neural input drives structuralchanges in the eyes of the scorpion that are similar insome ways to those that occur in Limulus eyes inresponse to circadian clock input. Just as in Limulus,the cell bodies that give rise to the retinal efferents inthe scorpion are located bilaterally in the cheliceralganglion. However, in the scorpion, different subsets ofthese cells project to the median and lateral eyes.Injections of OA into the scorpion mimics the effects ofthe circadian neural input (Fleissner and Fleissner,1988), but the identification of OA as a neurotransmit-ter in the retinal efferents of the scorpion must beverified by further biochemical, immunocytochemical,and functional studies.

Two distinct clusters of OA-immunoreactive cells,4–5 large (25 µm) cells and a greater number of smaller(10 µm) cells, are present in the cheliceral ganglion of C.salei (Seyfarth et al., 1993). Some neurites originatingfrom these cells project into the brain, and, as inLimulus, OA-immunoreactive processes are found inmost brain neuropils. However, the optic lobes and theoptic nerves of C. salei lack OA-immunoreactive pro-cesses, suggesting that the eyes of this spider are notinnervated by OA-containing efferent neurons. Anothermajor difference between the distribution of OA inLimulus and C. salei is that in this spider no OA-immunoreactive processes project out the pedal orhemal nerves and many terminals in the CNS arelocated near blood spaces. This observation suggeststhat in spiders circulating OA is probably released fromcentral sites, and that peripheral actions of OA areprobably neurohormonal.

SerotoninDistribution. Figure 2C summarizes the distribu-

tion of the major 5HT-immunoreactive cells and pro-cesses in Limulus brain (Chamberlain et al., 1986). No5HT is found in the eyes, optic nerves, or the corporapedunculata. However, an extensive network ofbranched 5HT-immunoreactive processes is evidentwithin the optic ganglia, numerous processes are withinthe central neuropil of the brain, and a sparse networkinvades the proximal portions of the stalks of thecorpora pedunculata, the optic track, and the circum-

78 B.-A. BATTELLE ET AL.

esophageal connective. Two separate zones of 5HT-immunoreactive processes are detected in the centralbody; an intermediate zone contains a meshwork ofsmall diameter, branching, beaded processes and aventral fiber zone contains large diameter fibers.

The brain contains four bilateral clusters of 5HT-immunoreactive cell bodies. Most (25–50 on each side)lie within the dorsal median group and project to thecentral body neuropil and to the central neuropil of thebrain (Fig. 2C). Not included in Figure 2C are 3–4 cellsfound at the ventral pole of each medulla that project tothe proximal stalks of the corpora pedunculata, and5–10 cells in each ventral lateral posterior group no. 2and 10–15 cells in each ventral medial group thatproject to the central neuropil of the brain. It is notclear which cell groups innervate the optic ganglia.Observations of the distribution of 5-HT-IR in Limulushave not included the cheliceral ganglia, so it is notknown whether cells in these clusters contribute 5HT-immunoreactive processes to the circumesophageal con-

nective. 5HT-immunoreactive cells are found in thecheliceral ganglia of the spider (Seyfarth et al., 1990).

Each Limulus abdominal ganglion also contains twobilateral clusters of 5HT-immunoreactive neurons, and5HT-immunoreactive processes are present within eachganglion, in the connectives between ganglia, andprojecting out the neural and hemal nerves (Washing-ton et al., 1994).

Functions. The physiological effects of 5HT in Limu-lus have not been studied in detail. However, thisamine may modulate neurons in the CNS (Walker andRoberts, 1982) with cholinergic synapses a major target(Ford et al., 1995). The dense serotonergic innervationof the optic neuropils indicates 5HT probably modu-lates the central processing of visual information andthe presence of 5HT-immunoreactive processes withinthe neural and hemal nerves projecting from the ab-dominal ganglia suggests it modulates peripheral neu-ral functions. The absence of 5HT from the peripheral

Fig. 7. A: Composite camera lucida draw-ing of the efferent neurons and axons thatbecame labeled in the brain when one lateraloptic nerve was filled with Neurobiotin. Thenerve was exposed to Neurobiotin for 7 days.The drawing is compiled from 19 serial100-µm horizontal sections comprising mostof the dorsal half of the brain, and includingall of the labeled projections. The left lateraloptic nerve was filled, and the areas of thebrain filled with afferent terminals are indi-cated by the stippled filling. The outlines ofthe major structures are taken from a horizon-tal level approximately halfway through thereconstructed area. This figure includes all ofthe major branches of the efferents, but doesnot include small collateral fibers. The recon-struction shows that after one lateral opticnerve is filled, efferent cells on both sides ofthe CER are labeled, and labeled efferentfibers project into the contralateral lateraloptic nerve, the ipsilateral and contralateralventral optic nerves and the ipsilateral andcontralateral median optic nerves. B: Sche-matic of proposed branch points of an ideal-ized efferent neuron and its projections outthe optic nerves. The number of branches inthe median and ventral optic nerves from oneefferent neuron is undetermined, and onebranch at each point is shown for diagram-matic purposes. Small collaterals projectinginto the neuropil are not shown. A: Mainefferent axon divides to project bilaterally. B:Branch point of efferent fibers projecting tothe median optic nerves. C: Branch point ofefferent fibers projecting to ventral opticnerves. Reproduced from Calman BG, Bat-telle B-A. 1991. Central origin of the efferentneurons projecting to the eyes of Limuluspolyphemus. Vis Neurosci 6:481–495, withpermission of the publisher.

79BIOGENIC AMINES IN LIMULUS

visual system indicates this amine is not a neurotrans-mitter in retinal afferents or efferents.

A preliminary description of 5HT-IR in the spider C.salei shows that, as in Limulus, the central brainneuropils and the optic ganglia are innervated by5HT-immunoreactive processes (Seyfarth et al., 1990).Some of the 4 bilateral clusters of 5HT-immunoreactivecells observed in the brain of this spider also may bestructural homologues of those in Limulus.

Do Amines Co-Localize in Limulus? Both 5HT-and HA-containing neurons are found among the cellsof the dorsal median group, and both HA- and OA-containing neurons are located within the cheliceralganglion. Co-localization of 5HT and HA among cells ofthe dorsal median group cannot be ruled out withoutdirect double label comparisons. However, comparisonsof the distributions of HA-immunoreactive processeswith the projections of the OA-containing retinal effer-ent neurons indicate that these are distinct populations.

ACKNOWLEDGMENTSWe thank Lynn Milstead for preparing the figures

and Jim Netherton for preparing the photographs.

REFERENCESAugustine GJ, Fetterer RH. 1985. Neurohormonal modulation of the

Limulus heart: Amine actions on the cardiac ganglion neurons. JExp Biol 118:53–69.

Augustine GJ, Fetterer RH, Watson WH III. 1982. Amine modulationof the neurogenic Limulus heart. J Neurobiol 13:61–74.

Barlow RB Jr, Chamberlain SC. 1980. Light and a circadian clockmodulate structure and function in Limulus photoreceptors. In:Williams TP, Baker BN, editors. Effects of constant light on visualprocesses. New York: Plenum Publishing, p 247–269.

Battelle B-A. 1980. Neurotransmitter candidates in the visual systemof Limulus polyphemus: Synthesis and distribution of octopamine.Vis Res 20:911–922.

Battelle B-A. 1991. Regulation of retinal functions by octopaminergicefferent neurons in Limulus. In: Osborne N, Chader J, editors. Prog.in retinal res, 10. Oxford: Permagon Press, p 333–355.

Battelle B-A, Evans JA, and Chamberlain SC. 1982. Efferent fibers toLimulus eyes synthesize and release octopamine. Science 216:1250–1252.

Battelle B-A, Calman BG, Andrews AW Grieco FD Mleziva MBCallaway JC, Stuart AE. 1991. Histamine: A putative afferentneurotransmitter in Limulus eyes. J Comp Neurol 305:527–542.

Bullock TH, Horridge GA. 1965. Structure and function in the nervoussystems of invertebrates II. San Francisco: W.H. Freeman and Co.

Calman BG, Battelle B-A. 1991. Central origin of the efferent neuronsprojecting to the eyes of Limulus polyphemus. Vis Neurosci 6:481–495.

Calman BG, Chamberlain SC. 1982. Distinct lobes of Limulus ventralphotoreceptors. II. Structure and ultrastructure. J Gen Physiol80:839–862.

Calman BG, Lauerman MA, Andrews AW, Schmidt M, Battelle B-A.1991. Central projections of Limulus photoreceptor cells revealed bya photoreceptor-specific monoclonal antibody. J Comp Neurol 313:553–562.

Chamberlain SC, Barlow RB Jr. 1980. Neuroanatomy of the visualafferents in the horseshoe crab (Limulus polyphemus). J CompNeurol 192:387–400.

Chamberlain SC, Wyse GA. 1986. An atlas of the brain of thehorseshoe crab Limulus polyphemus. J Morphol 187:363–386.

Chamberlain SC, Pepper J, Battelle B-A, Wyse GA, Lewandowski TJ.1986. Immunoreactivity in Limulus. II. Studies of serotonin-likeimmunoreactivity, endogenous serotonin and serotonin synthesis inthe brain and lateral eye. J Comp Neurol 251:363–375.

Clark AW, Millechia R, Mauro A. 1969. The ventral photoreceptor cellsof Limulus I: The microanatomy. J Gen Physiol 54:289–309.

Edwards SC, Pierce SK. 1986. Octopamine potentiates intracellularNa1 and Cl- reductions during cell volume regulation in Limulusexposed to hypoosmotic stress. J Comp Physiol B 156:481–489.

Elias MS, Evans PD. 1984. Autoradiographic localization of 3H-histamine accumulation by the visual system of the locust. CellTissue Res 238:105–112.

Evans JA, Chamberlain SC, Battelle B-A. 1983. Autoradiographiclocalization of newly synthesized octopamine in retinal efferents inthe Limulus visual system. J Comp Neurol 219:369–383.

Fahrenbach WH. 1970. The morphology of the Limulus visual systemIII: The lateral rudimentary eye. Z Zellforsch 105:303–316.

Fahrenbach WH. 1975. The visual system of the horseshoe crabLimulus polyphemus. Int Rev Cytol 41:285–349.

Fahrenbach WH. 1981. The morphology of the horseshoe crab (Limu-lus polyphemus) visual system VII: Innervation of the photoreceptorneurons by neurosecretory efferents. Cell Tissue Res 216:655–659.

Fahrenbach WH. 1985. Anatomical circuitry of lateral inhibition in theeye of the horseshoe crab limulus polyphemus. Proc R Soc Lond B225:219–249.

Fleissner G. 1983. Efferent neurosecretory fibers as pathways forcircadian clock signals in the scorpion. Naturwissenschaften 70:366.

Fleissner G, Fleissner G. 1988. Efferent control of visual sensitivity inarthropod eyes: With emphasis on circadian rhythms. In: LindauerM, editor. Information processing in animals 5. New York: FisherVerlag, p 1–67.

Fleissner G, Heinrichs S. 1982. Neurosecretory cells in the circadianclock system of the scorpion Androctonus australis. Cell Tissue Res224:233–238.

Ford BD, Dorsey WC, Townsel JG. 1995. Neurotransmitter andneuropeptide modulation of high affinity choline uptake in Limulusbrain. Comp Biochem Physiol A 111:147–153.

Hart MK, Battelle B-A. 1991. Histamine: Metabolism and release inthe Limulus visual system (Abstr.) Invest Ophthalmol Vis Sci32:1151.

Hartline HK, Wagner HG, MacNichol Jr EF. 1952. The peripheralorigin of nervous activity in the visual system. Cold Spring HarborSymp Quant Biol 17:124–141.

Heinrichs S, Fleissner G. 1987. Neuronal components of the circadianclock in the scorpion Androctonus australis: Central origin of theefferent neurosecretory elements projecting to the median eyes. CellTissue Res 250:277–285.

Lee H, Wyse GA. 1986. Octopamine stimulates a central patterngenerator for feeding in Limulus. Soc Neurosci Abstr 12:791.

Lee H, Wyse GA. 1991. Immunocytochemical localization of octopa-mine in the central nervous system of Limulus polyphemus: A lightand electron microscopic study. J Comp Neurol 307:683–694.

Lowry OH, Rosenbrough HJ, Farr AL, Randall RJ. 1951. Proteinmeasurements with the Folin phenol reagent. J Biol Chem 193:265–275.

Nassel DR, Holmquist HH, Hardie RC, Hakanson R, Sundler F (1988)Histamine-like immunoreactivity in photoreceptors of the com-pound eyes and ocelli of the flies Calliphora erythrocephala andMusca domestica. Cell Tissue Res., 253:639–646.

Nolte J, Brown JE. 1969. The spectral sensitivities of single cells in themedian ocellus of Limulus. J Gen Physiol 54:636–649.

Nolte J, Brown JE. 1972. Electrophysiological properties of cells in themedian ocellus of Limulus. J Gen Physiol 59:167–185.

Patten W.1912. The evolution of the vertebrates and their kin.Philadelphia: Blakistons.

Rane SG, Gerlach PH, Wyse GH. 1984. Neuromuscular modulation inLimulus by both octopamine and proctolin. J Neurobiol 15:207–220.

Sarthy PV. 1991. Histamine: a neurotransmitter candidate for Dro-sophila photoreceptors. J Neurochem 57:1757–1768.

Schmid A, Duncker M. 1993. Histamine immunoreactivity in thecentral nervous system of the spider Cupiennius salei. Cell TissueRes 273:533–545.

Schwartz J-C, Arrang J-M, Garbarg M, Pollard H, Ruat M. 1991.Histaminergic transmission in mammalian brain. Physiol Rev 71:1–51.

Seyfarth E-A, Hammer K, Grunert U. 1990. Serotonin-like immunore-activity in the CNS of spiders. Verh Dtsch Zool Ges 83:640.

Seyfarth E-A, Hammer K, Sporhase-Eichmann U, Horner M, VullingsHGB. 1993. Octopamine immunoreactive neurons in the fusedcentral nervous system of spiders. Brain Res 61:197–206.

Snodderly DM Jr. 1971. Processing visual inputs by the brain ofLimulus. J Neurophysiol 34:588–611.

Stuart AE, Mekeel HE. 1990. Uptake of histamine into the presynapticterminals of barnacle photoreceptors (Abstr.) Invest Ophthalmol VisSci 31:335.

Walker RJ, Roberts CJ. 1982. The pharmacology of Limulus centralneurons. Comp Biochem Physiol C 72:391–401.

Washington B, Higgins DE, McAdory B, Newkirk RF. 1994. Serotonin-immunoreactive neurons and endogenous serotonin in the opistho-somal ventral nerve cord of the horseshoe crab Limulus polyphemus.J Comp Neurol 347:312––324.

Watsae S. 1889. On the structure and development of the eyes of theLimulus. Johns Hopkins University Circulars 8, 70:34–37.

80 B.-A. BATTELLE ET AL.