J Comp PhysiolA (1992) 171 : 141-155

dourlml of lqaural, lind

9 Springer-Verlag1992

The morphology, innervation and neural control of the anterior arterial system of Aplysia californica M.E. Skelton 1 and J. Koester 1'2 Center for Neurobiology and Behavior1, Department of Psychiatry2, College of Physicians and Surgeons, Columbia University, 722 West 168 Street, New York, N.Y. 10032, USA Accepted April 3, 1992

Summary. The morphology, innervation, and neural control of the anterior arterial system of Aplysia californica were investigated. Immunocytochemical and histochemical techniques generated positive reactions in the anterior arterial system for several neuroactive substances, including SCPB, FMRFamide, R15cd peptide, dopamine and serotonin. Three neurons were found to innervate the rostral portions of the anterior arterial tree. One is the identified peptidergic neuron RI5 in the abdominal ganglion, and the other two are a pair of previously unidentified neurons, one in each pedal ganglion, named pedal arterial shorteners (PAs). The endogeneously bursting neuron RI5 was found to innervate the proximal anterior aorta. It also innervates a branch of the distal anterior aorta, the left pedal-parapodial artery. Activity in R15 causes constriction of the left pedal-parapodial artery. This effect is presumed to direct hemolymph towards the genital groove and penis on the right side in vivo. This vasoconstrictor action of R15 is mimicked by the Rl5el peptide. The PAS n e u r o n pair causes longitudinal contraction of the rostral anterior aorta and the pedal-parapodial arteries. In vivo, the pair is active during behaviors involving head withdrawal and turning. By adjusting the length of the arteries during postural changes, the PAS n e u r o n s may prevent disturbances in blood flow due to bending or kinking of the arterial walls. Key words: Aplysia californica - R15 - Egg-laying - Artery

Vasomotor

Introduction There is a considerable body of data on the neural control of various behaviors in Aplysia (e.g. Kandel 1979). This, coupled with the relatively simple circulatory and

Correspondence to : J. Koester

nervous systems of the genus, make Aplysia an excellent model in which to study the neural control of the cardiovascular system and its integration with behaviors such as feeding, reproduction and locomotion. In this paper we describe the morphology and innervation of the anterior arterial system and investigate the neural control of the anterior aorta, with emphasis on its previously unstudied distal trunk and side branches. The morphology of the circulatory systems of several Aplysia species have been described (Eales 1921 ; Winkler 1957; Wright 1960; Kandel 1979). Like all molluscs, apart from the evolutionarily advanced cephalophods, the circulatory system of Aplysia is open. The hemolymph is pumped to the tissues in closed arteries but does not stay within epithelia-lined capillaries as in higher animals. Instead, it empties into a system of sinuses and lacunae and thence into a large fluid-filled space, the hemocoel. The hemocoel contains many of the major organs including the gut. Hemolymph is collected from this cavity by sinuses and veins which lead via either the gill or kidney to the heart. The heart consists of a single auricle and ventricle. Hemolymph leaves the heart via 3 parallel pathways: the abdominal and gastroesophageal aortae, which supply the visceral tissues, and the anterior aorta, which supplies the somatic tissues. Aplysia relies on its circulatory system to transport oxygen, nutrients and waste products within the body. The circulation also has important hemodynamic functions associated with the maintenance and control of Aplysia's hydrostatic skeleton. Hemodynamic changes, probably resulting from the action of neurons on the cardiovascular system, may be instrumental in bringing about physiological changes that accompany behavior acts. During food-induced arousal there is significant swelling of the lips (Kuslansky et al. 1987), increase in heart rate (Dieringer et al. 1978), and increase in blood pressure (Koch and Koester 1982). During egg-laying in vivo the genital groove swells (Arch and Smock 1977); and following activation of the neuroendocrine bag cells in vitro, which triggers egg-laying behavior, there is an

142

M.E. Skelton and J. Koester: Morphology and neural control of Aplysia arterial system

increase in contractile activity of the anterior and gastroesophageal aortae (Ligman and Brownell 1985). Numerous substances present in the nervous system of Aplysia have been found to be active on the cardiovascular system, including small cardioactive peptides A and B (SCPA and SCPB; Lloyd et al. 1985), egg laying hormone (ELH; Ligman and Brownell 1985), serotonin (Hill 1964; Carpenter et al. 1971), glycine (Sawada et al. 1981 a), acetylcholine (Liebeswar et al. 1975), R1571 and R15~2 peptides (Alevizos et al. 1991 a), and dopamine (Hill 1964; Liebeswar et al. 1975). A number of neurons have been identified in the abdominal ganglion of Aplysia that innervate the cardiovascular system, some of which contain one or more identified cardioactive substances. Eight different types of motor or modulatory neurons that have cardiovascular effects have been reported (for review see Koester and Koch 1987). All of the cardiovascular neurons described thus far are located within the abdominal ganglion and appear to exert their effects on the heart and/or on the arteries close to the origin of the vessels from the ventricle. No data on the neural control of the distal anterior aorta or its branches have previously been reported. Hemodynamic changes that occur in the anterior portion of the body, remote from the influences of the cardioactive neurons so far described, may be under the control of neurons that either are located in the abdominal ganglion and project to the rostral portions of the body, or are located in the head ganglia. This paper describes data which indicate that several neuroactive substances are present in the anterior arterial system of Aplysia. It also characterizes the effects of 3 neurons on the arterial system supplying the rostral portions of the body. One of these cells is the previously identified peptidergic neuron R15 in the abdominal ganglion (Adams and Benson 1985). The other two, identified during this study, are a pair of neurons in the pedal ganglia, named pedal arterial shorteners (PAs)- The physiological role of the PAS neurons in the intact animal is examined. The morphology of the anterior arterial system is also described here. Some of these data have been previously reported in abstract form (Skelton et al. 1989). Methods Animals. Specimens of Aplysia californica weighing 5-700 g were supplied by Marinus, Long Beach, CA. Animals were kept in aquaria containing artificial seawater (ASW) (Tropic Marine) at 14-16 ~ until used.

Latex injections. To investigate the morphology of the anterior arterial system, colored latex medium (Carolina Biological Supply Company) was injected into the arterial vessels. Aplysia weighing 200400 g were anesthetized by injecting isotonic magnesium chloride (25% body weight) into the hemocoel. Anesthetized animals were cut open with a longitudinal incision along the mid-line of the foot. All the major branch points of the anterior aorta were tied off with the exception of one main pathway into which latex was injected. Other vessels were then untied and the process repeated until all of the main branches of the anterior aorta were filled with latex. The routes of the arterial branches were traced by further dissection and the morphology was drawn.

Glyoxylic acid-induced histofluorescence. A sucrose/potassium phosphate/glyoxylic acid (SPG) method, developed by De la Torre and Surgeon (1976) and modified by Goldstein and Schwartz (1989), Hawkins (1989) and Barber (1983), was used to detect monoaminergic nerve fibers within the anterior aorta. Portions of the anterior aorta were excised and treated according to the method of Barber (1983) and then photographed with a Leitz dialux microscope equipped with an incident illumination HBO 50 W super pressure mercury lamp, a BG12 excitation filter, and a K515 barrier filter to reveal induced fluorescence. This method allows the distinction between serotonin and dopamine: the former fluoresces yellow and the latter fluoresces blue/green. Immunocytochemistry. A modified version of the technique developed by Longley and Longley (1986) was used for wholemount immunocytochemistry. Portions of the anterior aorta, heart and ganglia were removed from Aplysia and pinned to the silastic-coated bottom of a small chamber and treated as described by Alevizos et al. (1989b). The primary antibodies used were anti-SCPB (Mahon et al. 1985) and anti-Lll peptide (Taussig et al. 1985) (obtained from R. Scheller); anti-FMRFamide (Peninsula); anti-peptide histidine isoleucine (PHI; Kuramoto et al. 1985) (Peninsula); antiserotonin (Incstar Corp.), anti-buccalin (Cropper et al. 1988) (obtained from K. Weiss); and antibodies raised to R15~2 peptide (Antibody I/II) and to the 16 amino acid sequence present in R15~1 peptide but not in R15~2 peptide (Antibody I) (Alevizos et al. 1991 a). As control against non-specific cross reaction of antibody and antigen, antisera that exhibit positive immunoreactivity (IR) in the tissues were pre-absorbed with the appropriate antigen. Control antisera were incubated for 12 h at 4 ~ with 10 -3 M antigen before they were applied to the tissues as described above. None of the preabsorbed antisera exhibited positive immunoreactivity.

Backfilling nerves. Nerves were backfilled with either CoG12 or NiC12 according to the method of Scott et al. (1991) and processed for wholemount visualization of backfilled neurons. Intra- and extracellular recording. Intracellular recordings from neurons and muscle fibers were carried out using standard electrophysiological recording techniques. Extracellular records were obtained using either cuff electrodes (Alevizos et al. 1991b) or focal electrodes, consisting of a glass macropipette. All in vitro experiments were carried out at 13-18 ~ To test for monosynapticity of connections, the threshold for action potential generation was raised by bathing preparations in high divalent cation solutions containing 165 m M Mg 2+ (3 • normal) and 30 m M Ca z+ (3 • normal). This solution was prepared by mixing isotonic CaCI2, isotonic MgClz and ASW. Identified neurons were labeled by iontophoretic injection of 4% Lucifer yellow (Sigma Chemical Co.) or 3% carboxyfluorescein (Kodak) with 1-5 nA hyperpolarizing current for 5-60 rain. In other cases, 2% neurobiotin (Vector) was iontophoretically injected using a 1-5 nA depolarizing current. Neurobiotin-labelled cells were visualized by incubation with either fluorescein conjugated to avidin or Vectorstain ABC kit reagents according to the methods of H. Kita supplied by Vector Laboratories, CA. Chronic in vivo recordings from nerves were made using the techniques described by Alevizos et al. (1991 b), with several modifications. In order to stabilize the recording electrode near to the appropriate nerve, the electrode was mounted on the outside of a short length (5 mm) of polyethylene tube. The tube was held in place by anchoring it to a portion of the arterial tree close to the nerve from which the recording was to be made. The appropriate artery was cut in half and the flanged tips of the tube were inserted into the cut ends of the vessel. The vessel was tied around the ends of the polyethylene tube with silk sutures (Ethicon 6-0), leaving the electrode exposed. The nerve was inserted into the electrode and insulated from the hemocoel by a latex membrane as described by Alevizos et al. (1991 b). To record from the right pedal arterial nerve (RPAn) the tube-mounted electrode was inserted into

M.E. Skelton and J. Koester: Morphology and neural control of Aplysia arterial system

143

To head Opaline gland a

r

t

e

r

y

~

Anterior aorta Abdominal ganglion

Opaline

Mantle artery Ganglionic artery - - Spermatheca - - S p e r m a t h e c a l artery - - C r i s t a e aorta

Genital artery White hemiduct Red hemiduct

""~-~

~.^,j

,

~

-

-

"'~ ...... /--+-7

Ventricle

Auricle

f Small hermaphroditic

.Mesentery Rostral

\ Posterior body wall Right 1 cm m

the adjacent portion of the rostral anterior aorta. To record from the left pedal arterial nerve (LPAn) the tube-mounted electrode was inserted into the proximal end of the left pedal-parapodial artery (Fig. 2).

Physiological recordings. The effects of PAS and R15 on the distal anterior aorta and its branches were measured by removing the appropriate portion of the aorta and ganglia with the innervation via the PAn intact. The preparation was placed in a chamber and the proximal end of the vessel was cannulated so it could be perfused with seawater from a reservoir, creating a perfusate pressure of 1-3 cm ASW. A T-shaped cannula was placed in line with the proximal aorta, between the inflow and the region innervated by the PAS cells. This cannula was connected to a pressure transducer which detected pressure increases brought about by constriction of the vessel. To detect changes in its length, the distal end of the aorta was attached to an isotonic movement transducer. The pedal and abdominal ganglia were pinned to the bottom of the chamber to obtain intracellular recordings from R15 and PAS. Pedal ganglia were desheathed in order to record from PAS, whereas impalement of R15 did not require desheathing. In the case of R15, hyperpolarizing current was injected for 1-2 h to prevent the cell from bursting. This silencing of R15 allowed the synaptic actions of the neuron, which are normally chronically desensitized, to be revealed when the cell was released from hyperpolarization and allowed to fire spontaneously (Alevizos et al. 1991 b~t). Potential agonists such as serotonin (Sigma), acetylcholine (Sigma) and R15~l peptide (HHMI Protein Core Facility, Columbia University) were injected as a bolus into the perfusion line, but the antagonist hexamethonium (Sigma) was added directly to the feed reservoir for a constant perfusion of the drug.

Results

Morphology of the anterior arterial system Our latex injections of the anterior aorta revealed its morphology in greater detail than had previously been

Left

Caudal

Fig. 1. The morphology of the proximal portion of the anterior aorta as revealed by latex injection, drawn from the ventral aspect. The gut has been removed and the position of the heart (dashed line) is shown underlying the pericardium

reported for Aplysia californica (Kandel 1979). Furthermore, we also observed some minor differences in morphology from the anterior arterial system of A. depilans and A. punctata as reported by Eales (1921). The anterior aorta of Aplysia californica is one of 3 main vessels that convey hemolymph from the ventricle. The other two, the gastroesophageal aorta and the abdominal aorta, supply the stomach and esophagus, and the digestive gland and ovotestis, respectively. These two aortae leave the ventricle directly, but between the anterior aorta proper and the ventricle lies a saccular structure of unknown function called the crista aortae. The genital artery branches off from the anterior aorta close to its origin from the crista aortae and extends right and caudally along the large hermaphroditic duct, to which it sends small branches (Fig. 1). The genital artery bifurcates just rostral to the accessory genital mass (involved in egg cordon formation). One branch travels ventral to this mass and onto the small hermaphroditic duct; the other branch extends dorsally to the body wall and the siphon. In several preparations, injection of the genital artery with latex filled vessels in the base of the gill pinnules, although these did not extend as far into the pinnules as the venous blood spaces. It was previously thought that the gill of Aplysia, like the kidney, possessed only a venous supply (Mayeri et al. 1974). A small vessel, termed the ganglionic artery, arises from the anterior aorta just rostral and to the left of the point where the genital artery originates from the anterior aorta. It supplies the abdominal ganglion and the spermatheca (a storage site for gametes). Also on the left, but more rostral, is the mantle artery, which leaves the anterior aorta and travels dorsally to the prox-

M.E. Skelton and J. Koester: Morphology and neural control of Aplysia arterial system

144

Tentacular

~ral artery

Buccal Penile art

~rior foot artery ad ganglia ~phalic artery

Cephalic arter

Left pedalparapodial artery

Right pedalparapodial artery-Anterior a o r t a - Opaline gland a r t e r y ~

-Mantle artery --Spermatheca -Abdominal ganglion Heart

Opaline g l a n d - - . ] Large hermaphroditic duct-Genital a r t e r y - -

- - Small hermaphroditic

duct

Rostral

Left Fig. 2. Morphology of the anterior arterial system of Aplysia californiea as revealed by latex injection. Drawn from ventral aspect after cutting longituCaudal dinally along the mid-line of the foot, spreading 2 cm out and pinning the cut edges, and removing the gut. Many of the distal ends of the vessels were embedded in the body wall and were revealed by further dissection of the tissues

Right +

imal end of the genital groove (which transports the egg cordon and sperm), the mantle shelf, and the dorsal body wall. To the right of the anterior aorta arises another small vessel which supplies the opaline gland. About mid-way between the abdominal ganglion and the head ganglia, the anterior aorta gives rise to the right pedal-parapodial artery (Fig. 2). This vessel supplies hemolymph to the right side of the body wall and the right parapodium via small ramifying branches which extend caudally all the way to the tail region. The left pedal-parapodial artery leaves the anterior aorta more rostrally, close to the head ganglia. Apart from this different point of origin and the presence of the left cephalic artery, which arises from the proximal portion of the left pedal-parapodial artery, the morphologies of both the left and right pedal-parapodial arteries are largely symmetrical. The right and left cephalic arteries give rise to small vessels that carry hemolymph to both the pedal and pleural ganglia. The hemolymph then enters the cerebral ganglion via small vessels contained within the sheath enclosing the cerebro-pedal and cerebro-pleural connectives. The cephalic arteries continue past the ganglia and supply the rostral body wall and the rhinophores. The right cephalic artery, which arises as a separate branch from the anterior aorta near but opposite the origin of the left pedal-parapodial artery, sends a branch caudally into the body wall and perfuses the rostral portion of

the genital groove. It also gives rise to the penile artery which supplies the penis and the penis retractor muscle (Fig. 2). At the point where the anterior aorta reaches the level of the buccal mass it gives rise to the anterior foot artery, which divides into two branches supplying the left and right sides of the anterior foot region. Just rostral to this artery, the anterior aorta supplies the buccal mass via a buccal mass artery. At its most rostral point the anterior aorta divides into two branches, which supply the tentacles. These tentacular arteries give rise to a pair of circumoral vessels, which perfuse the perioral zone (Fig. 2).

Neuroactive substances located within the anterior arterial system Glyoxylic acid histofluorescence. The glyoxylic acid histofluorescence technique of Barber (1983) allows the visualization of monoamines in wholemount preparations. It is possible to distinguish between serotonin and catecholamines (presumably dopamine in Aplysia; Carpenter et al. 1971) because under ultraviolet illumination the former fluoresces yellow and the latter blue/green. Goldstein (1984) has shown by microspectrofluorimetry that the emission spectra of tissues treated with glyoxylic acid agree well with serotonin and dopamine standards in Aplysia.

M.E. Skelton and J. Koester: Morphology and neural control of Aplysiaarterial system

145

Fig. 4A, B. The distal branches of the anterior aorta receive extensive peptidergic innervation. A FMRFamide-like IR and B SCPBlike IR in the tentacular arteries close to the origin of the circumoral arteries. Scale bar=A 75 I~m, B 150 ~m

Fig. 3A, B. The tentacular arteries are innervated by axons that stain for biogenic amines and serotonin-like material. A Glyoxilic acid-induced histofluorescence revealed densely distributed axons staining for biogenic amines. Most of the fibers fluoresced blue, indicating that they contained dopamine. Yellow fluorescing endings, containing serotonin, were present but much less prevalent. B Wholemount preparations stained with an antibody to serotonin revealed serotonin-like IR fibers in the tentacular arteries. The density of these IR fibers was similar to that of the yellow histofluorescent axons revealed by the glyoxylic acid technique in A. Scale bar = A 75 gm, B 25 gm

The glyoxylic acid histofluorescence technique revealed dopaminergic and serotonergic endings on the anterior aorta rostral to the origin of the anterior foot artery, on the tentacular arteries, and on the circum-oral arteries (Fig. 3 A). Fluorescent endings were particularly abundant in the tentacular arteries close to the origin of the circumoral vessels. In all areas exhibiting monoamine histofluorescence blue/green dopamine fluorescence always predominated over yellow serotonin fluorescence, which was relatively sparse.

Immunostaining. Several antibodies raised to neuroactive substances known to be present in Aplysia were applied to the anterior aorta and its branches. Serotonin-like, FMRFamide-like, and SCPR-like IR nerve terminals were seen in the tentacular arteries, close to the origin of the circumoral vessels (Figs. 3 B, 4A, B), and in the main aortal trunk between the origin of the tentacular arteries and the origin of the buccal mass artery. There was no IR to buccalin, L11 peptide, or peptide histidine

isoleucine (PHI) antibodies in the anterior arterial system. We used Antibody I and Antibody I/II (Alevizos et al. 1991a) to test for the presence of R15~l and R15a2 peptides (Buck et al. 1987). R15al peptide is translated from m R N A present in the neuron R15 o f Aplysia. The precursor of this m R N A is also present in several other neurons in the nervous system, but is alternatively spliced to encode a different peptide, R15a2 (Buck et al. 1987). Antibody I/II recognizes both R15al-like and R15a2-1ike peptides and stains several neurons in the CNS, including R15 (Alevizos et al. 1991a). Antibody I is raised to a 16 amino acid sequence which R15al possesses but R15a2 does not. Therefore, Antibody I recognizes only R15al-like material. The only cell body in the CNS that stains for antibody I is R15 (Alevizos et al. 1991 a). IR to both Antibody I and I/II was found in portions of the anterior aorta. The proximal part of the anterior aorta, from its origin at the crista aortae to the mantle artery, contained a dense network of Antibody I and I/II IR fibers (Figs. 5 and 6). Antibody I/II IR fibers extended onto the walls of the genital artery (including small branches supplying the large hermaphroditic duct), the opaline gland artery, the mantle artery and the ganglionic artery. Nerve fibers IR to Antibodies I and I/II extended rostrally along the anterior aorta to the origin of the mantle artery, but not beyond this point. The lack of staining immediately rostral to the mantle artery did not mean that all rostral parts of the anterior arterial system were devoid of Antibody I and I/II IR fibers. In all six of the arterial systems exposed to Antibody I, IR fibers were seen ramifying densely in the left pedal-parapodial artery (Fig. 7A). In two cases sparsely distributed IR fibers were seen in the anterior foot artery, and in one other case an axon IR to Antibody I entered the anterior aorta at the point of origin of the right pedal-parapodial artery and travelled ros-

146

M.E. Skelton and J. Koester: Morphology and neural control of Aplysia arterial system

Fig. 5. Axons IR to Antibody I ramify over the proximal end of the anterior aorta (AA) and its branches. Antibody I IR was visible in neuron R15, in the proximal anterior aorta (AA), in the mantle artery (MA), and in the ganglionic artery (GA) which perfuses the abdominal ganglion (AG). Viewed from the dorsal aspect. Scale bar = 300 gm trally up the anterior aorta and entered the left pedalparapodial artery. This axon did not enter the right pedal-parapodial artery. No other portions of the anterior arterial system rostral to the mantle artery were I R to Antibody I. The localization of Antibody I/II I R in the distal anterior arterial tree did not differ appreciably from that of Antibody I I R (Fig. 7 B). It seemed likely that the R15~l-like I R in the left pedal-parapodial artery is from an axon branch of R15, because Alevizos et al. (1991a) found that R15 is the only cell body in the CNS to stain with Antibody I. Observation of immunostained preparations, which included the head ganglia, indicated that Antibody I I R fibers reached the left pedal-parapodial artery via small pedal nerves, which we n a m e d the left and right pedal arterial nerves (LPAn and RPAn) (see Fig. 13). We therefore propose that the Antibody I I R axon travels from R15 in the abdominal ganglion to the head ganglia via the pleuroabdominal connectives and thence to the left pedal-parapodial artery via the PAn. This is consistent with the findings of Alevizos et al. (1991 a) who

Fig. 6 A-E. Distribution of antibody I/II IR demonstrates the presence of R15~l-like and/or R15ct2-1ike material in nerve terminals in the arterial system. A The distal portion of the genital artery. B A branch of the genital artery supplying the large hermaphroditic duct. C The opaline gland artery. D The mantle artery. E The ganglionic artery. Scale bar = 100 ~tm

reported that there is an Antibody I I R axon in each of the pleuro-abdominal connectives, and Rittenhouse and Price (1985), who found that R15 has an axon in the left pleuroabdominal connective. There were differences between the morphologies of the L P A n and the RPAn, and often considerable variation between individuals. Usually the left PAn originated from the parapedal commissure close to the pedal ganglion and travelled directly to the left pedal-parapodial

M.E. Skelton and J. Koester: Morphology and neural control of Aplysia arterial system

147

A

LPAn :

-

:

~ R15 soma~,_.~.,--"--"~

16 ~v - - - ] 2 0 mV 25 msec

B RPAn

Fig. 7A, B. Axons containing R15~-peptide-like material ramified over the left pedal-parapodial artery. The left pedal-parapodial artery contained terminals immunoreactive to (A) Antibody I and (B) Antibody I/II, both of which bind to the R15~l peptide synthesized by R15. Antibody I/II also binds to R15~2 peptide, so some of the terminals could come from other, R15~2 peptide-containing, neurons. Data are from two different wholemount preparations. Scale bar=(A) 100 ~tm, (B) 75 ~tm artery. However, in some preparations this nerve reached the artery by running along the left cephalic artery, making it difficult to identify. The right PAn usually divided soon after its origin from the parapedal commissure and sent one branch to the main aorta close to the origin of the right pedal-parapodial artery. The other branch travelled directly to the middle of the right pedal-parapodial artery (see Fig. 13).

Innervation of the anterior aorta by R15 The hypothesis that Antibody I I R present in the left pedal parapodial artery was due to innervation by R15 was tested physiologically. We recorded en passant from the left PAn and from portions of the anterior aorta with an extracellular electrode, while simultaneously monitoring R15 soma action potentials with an intracellular microelectrode. Extracellular potentials where recorded from the left P A n and from areas of the left pedal-parapodial artery that exhibited positive immunostaining. They followed R15 soma spikes 1:1 and with constant latency (Fig. 8 A). These recordings were made with the preparations bathed in high divalent cation solution to block polysynaptic transmission. Simultaneous recordings ( n = 4 ) from the L P A n and the R P A n revealed that an R15 axon was present in the L P A n but not in the R P A n (Fig. 8 B). Antibody I stained wholemount preparations suggested that R15 reached the left pedal-parapodial artery via the L P A n (but sometimes also via the RPAn). However, because o f the dense and widely distributed immunoreactive endings on the proximal portion of the anterior aorta it was not clear by which route R15 innervated the more caudal vessels. Rittenhouse and Price (1985)

R1 5

J~LLL~/

10 sec

12 pv 12 l i v 40 mV

Fig. 8A, B. RI 5 innervates the left pedal-parapodial artery via the left pedal arterial nerve (LPAn). Simultaneous recordings from the R15 soma and from the LPAn revealed electrical activity attributed to an axon branch of R15. A Seven superimposed soma spikes from R15 were followed 1:1 with a constant latency by extracellulady recorded spikes in the LPAn. The preparation was bathed in high divalent cation solution to block polysynaptic transmission. B The R15 spike was present only in the left PAn and not in the right PAn. The large units in the left and fight PAn were the PAS motoneurons, which caused the pedal-parapodial arteries and the anterior aorta to shorten (see below)

have shown that R15 axons extend out of the abdominal ganglion by way o f several peripheral nerves, including the pericardial nerve, which sends a small branch to the anterior aorta. Immunocytochemical evidence obtained here suggested R15 branching was more extensive than these authors report. By using extracellular suction electrodes to record R15 spikes in the proximal aorta during selective lesioning of possible R I 5 axon routes, it was established that the neuron innervated the aorta via at least two pathways. R15 spikes were recorded extracellularly from the main proximal anterior aorta, the mantle artery and the genital artery. These signals followed soma spikes 1:1 and with constant latency (Fig. 9A). However, latency of extracellular spikes in the mantle artery decreased when the suction electrode was m o v e d from the proximal end, close to the anterior aorta, to more distal portions o f the vessel (Fig. 9B). This suggested that R15 innervates the artery via a route other than from the anterior aorta. Lesioning of the pericardial nerve caused extracellularly recorded R15 spikes to disappear f r o m the anterior aorta and the genital artery, but they were still present in the mantle artery. The R15 spikes in this vessel were abolished by cutting a small branch of the vulvar nerve that travels in mesentery to the mid-point o f the mantle artery (Fig. 9 C). A special protocol was used to test whether R I 5 had m o t o r effects on the left pedal-parapodial artery. It had been reported earlier (Alevizos et al. 1991b, c, d) that

M.E. Skelton and J. Koester: Morphology and neural control of Aplysia arterial system

148

Anterior aorta 1.Gi,tV/ 20mVl..,_.

ID |

r-

E E 9

Genital artery ~.5~vt

"5

20mVl~ 5msec

|

t-

"5

Mantle Artery (proximal end) l~tvt

~nmV ~b--rmee

C

B 9

, "

~ ,1' ~

,

Anterior aorta

Opaline gland artery~ .St ~--.~,~.~ ~ V u l v a

Proximal

,~m~VL lOmsec

~, [Il ' ~ ,. ~ ~ M a n t l e "~ I~;~/f/~T~bBranchialnerve I

/0

nerve

artery

spermatneca

From heart Genital artery artery wascut, the RI 5 spikein this vesseldisappeared(right column). B Thelatencyof the extracellularlyrecordedspikes(10super-

Fig. 9A-C. R15 sends processes via two routes to the regions of the proximal anterior aorta and its branches that contain R15~l peptide-IR terminals. A Extracellular spikes (10 superimposed traces) recorded from 3 portions of the proximal anterior aorta followed R15 soma spikes 1:1 and with constant latency, suggesting that R15 processes terminated in the vessel (left column). However, R15 reached these parts of the aorta via two different routes. When the pericardial nerve was cut close to its origin from the abdominal ganglion, the extracellularly recorded R15 spike in the main trunk of the anterior aorta between the ganglionic artery and the mantle artery disappeared, as did the signal in the genital artery. Spikes in the mantle artery persisted (middle column). When a small branch of the vulvar nerve that innervated the mantle

some of the synaptic actions of R15 rapidly desensitize. In order to demonstrate them in a semi-intact preparation, R15 had to be prevented from spiking by injecting hyperpolarizing current for some time beforehand (approximately 1-2 h). Using this hyperpolarization protocol of Alevizos et el. (1991 b) it was found that spontaneous bursting by R15 following a 2 h period of enforced

imposed), recorded before the lesion experiments, increased when the recording electrode was moved towards the proximal end of the mantle artery. This result is consistent with the hypothesis that the axon of R15 reached the mantle artery from its distal end and travelled along the vessel towards its point of origin from the main aorta. C Schematic diagram of the abdominal ganglion and the proximal anterior aorta showing the recording sites in (A) and (B) (stars) and the axon routes by which R15 was found to innervate the anterior aorta. The points where the axon routes of R15 were lesioned are indicated by scissor symbols

silence resulted in constriction of two portions of the anterior arterial system that had been shown to contain antibody I and I/II immunoreactivity. The left pedalparapodial artery always responded to R15 activity (Fig. 10A) and the proximal portion of the anterior aorta responded in approximately one third of preparations (data not shown). This action was mimicked by perfu-

M.E: Skelton and J. Koester: Morphology and neural control of Aplysia arterial system

A

A Perfusate__~ pressure

149

Peffusate pressure

12.5 mm H20

12mmH20

Shortening~

111111111

/ 2 0 mV lOO sec LPAs

~~...J40

mV 5

sec

B p g e r f u s a t e ~ pressure A , C 0.1 ml 10-6 M R15o~1

-14 mm H20 40 sec

Fig. IOA, B. Activity in RI5 constricts the left pedal-parapodial artery and is mimicked by the perfusion of the artery with R15~1 peptide. A Bursting of R15 caused the artery to constrict, as measured by recording perfusate pressure upstream from the site of innervation. R15 was hyperpolarized for 2 h prior to the firing period to eliminate residual desensitization of the R15~1 receptors (Alevizos et al. 1991b, c, d). The action potentials recorded from R15 with an intracellular electrode were distorted because the amplifier was capacity-coupled. B The effect of R15 activity on the left pedal-parapodial artery was mimicked by the injection of a bolus ofR15~l peptide (bolus threshold 10 9 M) into the perfusion line connected to the artery. Brief pressure transients were caused by control injection (C) of a bolus of vehicle (ASW) or vehicle plus peptide

9 10-3 M Hex.

C Shortening~~~ 9 10-s M ACh

10

sec

D Aortal muscle ~ = m ~ e e ~ m m ~ fiber

L

LPAs

[1 mV

._JlO mv 20

Tlmissur8

~ 2 0 mV 5 sec

msec

Fig. 12A-D. When LPAs is fired at physiological rates, it generates brief cholinergic ejps in longitudinal muscle fibers that cause shortening of the anterior aorta. A Trains of action potentials in LPAs evoked longitudinal contraction of the anterior aorta between the origin of the right pedal-parapodial artery and the buccal mass artery (middle trace), but resulted in no significant constriction of the vessel as monitored by recording the perfusate pressure. B Contraction of the anterior aorta by LPAs activity was antagonized by perfusion of the artery with 1 0 - 3 M hexamethonium, a selective cholinergic blocker. C Injection of a bolus of acetylcholine into the artery (bolus threshold 10-7 M) caused shortening of the vessel, mimicking the effects of firing the neuron. D Intracellularly recorded ejps from single longitudinal muscle fibers of the anterior aorta followed LPAs soma spikes 1:1 and with constant latency in a preparation bathed in high divalent cation solution. Ejps evoked by PAS could be recorded only from regions of the arterial tree that contracted in response to activation of the cell

Innervation of the anterior aorta by arterial shortener neurons Fig. 11. The morphology of the left pedal arterial shortener (LPAs) neuron as revealed by carboxyfluorescein injection. The cell body is located in the pedal ganglion between the pedal commissure and the cerebro-pedal connective. Its axon leaves the pedal ganglion via the left pedal arterial nerve (LPAn). The RPAs neuron is bilaterally symmetrical to LPAs

sion of the arteries with R15~1 peptide with a threshold concentration of 10 9 M (Fig. 10B). The true threshold concentration is lower because the peptide was administered as an injected bolus and some dilution would have occurred. Both continuous R15 spiking and repeated peptide perfusion caused response decrement as reported for the other synapses made by R15 (Alevizos et al. 1991 b, c, d).

We used anatomical techniques to search for other neurons that might innervate rostral portions o f the anterior aorta. Cobalt backfills of the left and right PAn labelled a single pair of neurons situated in Cluster II of the pedal ganglion (Hening et al. 1979) between the pedal commissure and the cerebro-pedal connective (data not shown). Electrophysiological searches revealed a pair of neurons in Cluster II that, when stimulated do spike at physiological rates, caused selected portions of the anterior aorta to contract longitudinally. Dye filling of the neuron pair with Lucifer yellow or carboxyfluorescein revealed that each one send a single axon out the ipsilateral PAn (Fig. 11). Activity of each neuron caused longitudi-

150

M.E. Skeltonand J. Koester: Morphologyand neural control of Aplysia arterial system

Buccalmassartery -:>-"J /~ Anterior foot artery cC1 " ,, s S ~ ~ e b r a l ganglion P,.ur.,

~f'~,L\~L.~'~ --d]L~Pedal ganglion Plauroabdominalconnective~ - ~ 1 ~ /~/P9~ -P6 commissure LPA~L~ Parapedal ~ ~ 7 RPAn I ~ Rightpedal~,~ parapodial artery Leftpedalparapodialartery ~ ~_~ Anterior .

.

.

.

aorta~'v~"

Fig. 13. Diagram to show the regions of the arterial system that are innervatedby LPAsand RPAsneurons. When the neurons were fired by current injection they caused longitudinalcontraction of the ipsilateral pedal-parapodial artery and the anterior aorta from the right pedal-parapodial artery to the origin of the anterior foot artery [] RPAs, [] LPAs

nal contraction of the main aorta between the right pedal-parapodial artery and the buccal mass artery, and longitudinal contraction of the ipsilateral pedal-parapodial artery. The cells did not cause constriction of these vessels (Fig. 12A). These paired cells were named the left and right pedal arterial shorteners (LPAs and RPAs). The mean spike rate during evoked spike trains shown in Fig. 12A was 6.0 Hz. For comparison, examples of maximum spike rates in vivo are 8.2 Hz during head withdrawal, 9.0 Hz during escape stepping, 3.8 Hz during respiratory pumping episodes, 4.8 Hz during an ipsilateral head turn, and 1.2 Hz at rest (see below). Two types of evidence indicate that the PAS neurons are cholinergic. First, their motor effects were blocked by constant perfusion of the aorta with 10 -3 M hexamethonium (Fig. 12 B), a selective antagonist for excitatory cholinergic synapses in Aplysia (Ascher and Kehoe 1975). Second, the motor effects of RPAs and LPAs were mimicked by the perfusion of acetylcholine (Fig. 12C) (threshold measured by bolus injection was 10-7 M). The PAS neurons appeared to monosynaptically innervate the longitudinal muscle fibers of the anterior aorta and pedal-parapodial arteries. Intracellularly recorded excitatory junction potentials (ejps) in muscle fibers followed PAS soma spikes 1:1 and with constant latency, even in preparations bathed in high divalent cation solution (Fig. 12D). A distinct laterality of the motor effects of the left and right PAS neurons was revealed by observing the contractions each produced. Both caused the main anterior aorta to shorten. In addition, each PAS neuron caused the ipsilateral pedal-parapodial artery to shorten but had no effect on the contralateral vessel (Fig. 13). Extracellular recordings confirmed the dye filling re-

Fig. 14A, B. Spontaneous synaptic excitation of the PAS motoneurons coincides with activity in Int XIII, which strongly excites R15. A PAs was excited during intense Int XIII activity, as monitored by the giant Int XIII-generated epsp in R15. As the frequency of Int XIII epsps recorded from R15 increased, PASbecame increasingly depolarized. (epsp frequency was determined by playing the tape back on an expanded time scale and counting epsps by hand.) B In spontaneously active PAS neurons, strong Int XIII activity coincided with an increased spike frequency in PAS

sults: the PAS cells do not send axons out the contralateral PAn (see Figs. 11 and 15 D).

Electrophysiological character&tics of P As LPAs and RPAs did not make synaptic contact with each other (data not shown), although they received some common synaptic inputs. Both were excited during respiratory pumping, which is an episode of brief simultaneous contractions of the mantle organs accompanied by heart inhibition (Byrne and Koester 1978; Hening 1982) (See Fig. 15 C). They were also indirectly excited whenever Interneuron XIII fired spontaneously. Interneuron XIII is an unidentified, spontaneously active interneuron that produces epsps in R15, RBHE, L40 and the L9G neurons; ipsps in L10, L l l , and the LBvc neurons; and conjoint e-ipsps in L7 (Segal and Koester 1982; Koester 1991). Hyperpolarization of PAS by current injection during strong Interneuron XIII bursts revealed an increase in epsp input to PAS, which resulted in a depolarizing wave (Fig. 14A). If the Pns cell was not hyperpolarized, this wave of excitatory input evoked a significant increase in spike frequency (Fig. 14 B).

Activity of P As in vivo Chronic recording of extracellular signals from the PAns was used to determine the activity patterns of the PAS neurons in vivo during various behaviors. The interpretation of these records was based on the conclusion that the largest unit recorded from RPAn or LPAn was the ipsilateral PAS neuron. This conclusion was based upon

M.E. Skelton and J. Koester: Morphology and neural control of Aplysia arterial system

A LPAs

LPAn

.... ....... '~'~

!

10 mVI 6.Vl~ 2 msec

B LPAn --il 6,vL_ 2 msec

C LPAn 6 pV [ 2 msec Fig. 15A-C. The only large signal recorded extracellularly from the PAn is from the axon of the ipsilateral PAS neuron. A Firing an LPAs neuron, previously identified by its motor effect on the rostral anterior aorta, evoked a large extracellularly recorded unit in the LPAn (10 superimposed traces). The top trace was recorded intracellularly from the soma of LPAs, and the bottom trace from the axon in LPAn. B The proximal end of the LPAn was next cut and sucked into a stimulating electrode. When the stimulus voltage (0.1 ms, 1 Hz) was increased gradually, the axon with the lowest threshold was stimulated. This unit (a) was of similar size and shape as the identified PAS spike. Some small differences in the PAS spike and unit a were due to the superimposition of the stimulus artifact with the unit. The artifact decayed exponentially with a time constant of approximately 7 ms resulting in a lengthening of the apparent duration of unit a. Ten spikes are superimposed. C Further increases in the stimulus voltage caused only one other axon to fire. This unit (b) was much smaller than that of the lower threshold fiber and had a slower conduction velocity. It is presumed to be the axon of R15 (see Fig. 8B). Ten spikes are superimposed. All data recorded at 13 ~

two series of in vitro experiments. (1) Left or right PAS neurons were identified by their m o t o r effects on the anterior aorta and injected iontophoretically with either Lucifer yellow or neurobiotin. After allowing 4-12 h for diffusion of the dye into axonal processes the ipsilateral PAn was processed and sectioned. In these sections the axon with the largest cross-sectional area, which would produce the largest extracellularly recorded signal in the PAn, was the labelled PAS axon. In some cases the only axon that could be resolved at the light micro-

151

scope level was that of the labelled PAS neuron (data not shown). (2) In semi-intact preparations consisting of the head ganglia, the anterior aorta and the PAn, the soma of PAS was identified by impaling it and firing it, thereby generating its m o t o r effects on the vessel. The distal end of the PAn was then cut and attached to an extracellular electrode to record the PAS axon spike (Fig. 15A). Finally, the proximal end of the PAn was cut and connected to a stimulator. Axons in the nerve were stimulated to spike by gradually increasing the stimulus voltage (0.1 ms pulses, 1 Hz). The lowest threshold and largest amplitude extracellularly recorded spike in the PAn was found to be the same size and shape as the identified PAS unit (Fig. 15B). Further increases in the stimulus voltage caused only one other axon to fire (Fig. 15 C). This unit had a smaller amplitude and a slower conduction velocity than that generated by the PAS axon. For the L P A n it is likely that the smaller unit was the axon of R15, which electrophysiological and immunocytochemical evidence has suggested extends out of the L P A n (Figs. 7 and 8). In the R P A n there were also only two units - the larger generated by the RPAs axon and the smaller by an unidentified neuron (data not shown). In vivo, both PAS neurons generally showed low frequency activity, which changed with movements of the head and anterior body. A light tap to the animal's head evoked defensive withdrawal, which was accompanied by a large increase in the frequency of PAS spikes (Fig. 16A). More dramatic movements of the head, such as those accompanying escape locomotion, resulted in bursting of PAs. Bursts were in phase with body shortening when the posterior foot was detached from the substrate and pulled towards the anterior foot. Conversely, the interburst interval Of PAs was in phase with stretching out of the head region, with the anterior foot detached from the substrate (Fig. 16B). Spontaneous respiratory pumping, which is accompanied by a slight head withdrawal, also transiently increased PAS frequency (Fig. 16 C). Consistent with the findings of Alevizos et al. (1991 b), the lack of a smaller bursting unit in the recordings from L P A n suggest that R15 is normally silent in vivo. During head withdrawal both LPAs and RPAs were simultaneously active, but the activity patterns of LPAs and RPAs were different during asymmetric movements of the head. When the head turned to the left, LPAs spike frequency increased and RPAs frequency decreased, in some cases to zero during sharp head turns. Reciprocally, RPAs became m o r e active and LPAs became less active when the head turned to the right (Fig. 17 A, B). In general, the greater the degree of turning, the greater the ipsilateral excitation and contralateral inhibition of the PAS cells. In several preparations, an additional small unit was detected in the R P A n that was active during head movements. In contrast to RPAs, this small unit became active during a contralateral turn and was silent when the head was forward or during an ipsilateral turn to the right (Fig. 17B). No corresponding unit was found in LPAn.

M.E. Skelton and J. Koester: Morphology and neural control of Aplysia arterial system

152

A

Defensive wflhdrawal IIIIIJlltlllllll II llltlllllll

lll[llllllllllll II llllfll lll I A T~ch~

A T~h h ~

5 s~

B Escapestepl~ng UJPmm'u

llll

allB

i

....

JHl. . . . 16#V

Headb~aml

C

5sec

Respiratory pumping I

lip

5 sec

RP

6p.V

D I I r I i

i

i

iiiiiilliiiiiiiiiiiiiiiiil{~iillililiiiiiii

16p.v

12omV I:~kn

Ill

I

I

1qTq I

IdLIletll/lliltllilllltlhLletlhlr I l n HIIIll]HIHIItNNtlHIIII[IHIIIIItllliHliHIIIIIIIIIIIU]INIIHIII]II][] Ill ] I

RPAs

112 p-V

_.J 20 mV 2secs

Fig. 16A-D. Chronic, in vivo recordings from the LPAn reveal LPAs activity during various behaviors in the intact, unrestrained animal. A light touch to the head elicited defensive withdrawal and an increase in the frequency q LPAs spikes. B During escape stepping, evoke by placing salt crystals on the tail, the frequency of PAS spikes increased during shorter ing of the body associated with the movemen of the unattached tail region towards the sub strate-attached head. A decrease in spike frequency was associated with the movement of the head forward (bars). C During respiratou pumping episodes, which include weak, transient neck shortening, LPAs spike frequency ii creased. D In vitro experiments confirmed th~ both left and right PAS neurons evoked large extracellular signals from the ipsilateral pedal arterial nerve (PAn) but not from the contralateral nerve. In vitro data were recorded in two separate preparations

A tP~

10

B FPAn

Discussion

Morphology of the anterior aorta The m o r p h o l o g y o f the a n t e r i o r a o r t a o f Aplysia californica was f o u n d to be generally similar to that o f two

SeCS

Fig. 17A, B. The two PAS n e u r o n s exhibit reciprocal activity patterns during head turning in the intact, unrestrained animal. A During head turns to the left, LPAs became more active, but its spike frequency decreased during turns to the right. B In a recording from a di ferent animal, RPAs became more active duriJ head turns to the right, but its spike frequenc decreased when the head turned to the left. /~ smaller, barely detectable unidentified unit w~ also sometimes recorded in the right PAn. TI~ unit, which was active during contralateral turns but silent during ipsilateral turns, may 1 an unidentified inhibitory vasomotoneuron

other Aplysia species, A. depilans a n d A. punctata, studied by Eales (1921). The m a i n difference b e t w e e n the species was the p o i n t o f origin o f the genital artery, which in A. californica arose f r o m the a n t e r i o r a o r t a close to its origin f r o m the crista aortae. I n A. depilans the genital artery arose f r o m the a n t e r i o r a o r t a m o r e

M.E. Skelton and J. Koester: Morphology and neural control of Aplysia arterial system rostrally, close to, but independently from the origin of the opaline gland artery. In A. punctata the genital artery often gave rise to the opaline gland artery. Eales (1921) did not describe an aortal branch to the abdominal ganglion similar to that present in A. californica, although it seems unlikely that these closely related Aplysia species would differ in this respect. The rostral portions of the anterior aorta in A. punctata and A. depilans appeared to differ from A. californica only in the origin of the left cephalic artery. In the former two species the left cephalic arteries arose from the anterior aorta whereas in A. californica it usually arose from the base of the left pedal-parapodial artery.

Neuroactive substances within the anterior aorta Previous work and results presented here show that the cardiovascular system of Aplysia is richly and selectively innervated by several neuroactive substances, suggesting that the neural control of the vascular system is complex. Nerve terminals in the proximal aortal system have been found to contain acetylcholine (Liebeswar et al. 1975), dopamine histofluorescence (Koester and Koch 1987), serotonin-IR (Alevizos et al. 1989a), R15el peptide-IR (Alevizos et al. 1991 a), FMRFamide-IR (Alevizos et al. 1989 a), buccalin-IR (Miller et al., in press), and myomodulin-IR (Miller et al. 1991). In this study dopamine histofluorescence and serotonin-like, R15~l-like, SCPBlike and FMRFamide-like IR were also found in the rostral portions of the anterior aorta. The distribution of serotonergic innervation of the cardiovascular system is widespread. It was previously found to be present throughout the heart and in the abdominal and gastroesophageal arteries (Liebeswar et al. 1975; Alevizos et al. 1989a). This study demonstrates that it is also found in the rostral portions of the anterior aorta, the tentacular arteries, and the circumoral arteries. A source of serotonin in the heart is known to be the RBHE neuron in the abdominal ganglion, although recent evidence suggests not all the serotonin in the heart comes from this neuron (Skelton and Koester 1991). The source of serotonin in the rostral anterior aorta and its branches is unknown, but is unlikely to come from PAS or R15, because in studies using histochemical and immunological techniques for the visualization of serotonin-containing cells, no putative serotonergic cell bodies have been found in the region Of PAs neurons in the pedal ganglia or R15 in the abdominal ganglion (Hawkins 1989; Goldstein et al. 1984; Tritt et al. 1983). The distribution of dopamine histofluorescence is also extensive: it is present in the genital artery, the caudal portion of the anterior aorta (Koester and Koch 1987) and, from results presented here, in the rostral anterior aorta, the tentacular arteries and the circumoral arteries. The source of dopamine is unknown, although the positions of putative catecholamine-containing cells in the abdominal and head ganglia have been mapped (Goldstein and Schwartz 1989).

153

The significance of R15 innervation of the anterior aorta It has been hypothesized that R15 is involved in the regulation of egg-laying behavior. Many of the physiological actions of the neuron observed thus far can be related to changes which occur during egg-laying (Alevizos et al. 1991): (1) The neuroendocrine bag cells, which fire in a population burst to trigger egg-laying (Dudek et al. 1979), enhance the spontaneous bursting of R15 in vitro (Branton et al. 1978). (2) R15 increases respiratory pumping by increasing the activity in the R25/L25 network which initiates respiratory pumping episodes (Alevizos et al. 1991 b). (3) R15 contracts muscle fibers in the sheath of the pleuroabdominal connectives, by exciting neuron L7. This contraction of the pleuroabdominal connectives may modulate the flow of hemolymph past the neuroendocrine release sites of the bag cell terminals in the sheath (Alevizos et al. 1991c). (4) R15 increases the rate of anterograde peristalsis of the large hermaphroditic duct, presumably aiding egg cordon transportation from the accessory genital mass towards the genital groove (Alevizos et al. 1991 d). (5) Injection of R15~1 peptide into the hemocoel of Aplysia causes an increase in net water uptake (Weiss et al. 1989). The water content of gelatinous egg capsules is high, so R15 activity may serve to maintain the water content of Aplysia during egg-laying. The demonstration here that R15 activity also causes a constriction of the left pedal-parapodial artery is consistent with the hypothesis that R15 is involved in egglaying behavior. A constriction of the left pedal-parapodial artery in vivo will presumably shift hemolymph flow to the right side via the right pedal-parapodial and the right cephalic arteries. As reported here, portions of the right pedal-parapodial artery perfuse the genital groove. Arch and Smock (1977) reported that a consistent feature of egg-laying behavior is swelling of the genital groove, even before the eggs enter the genital groove. This swelling may be brought about by the redirection of hemolymph from the left side by the constrictive action of R15 on the left pedal-parapodial artery. As the penis and penis retractor muscles are also on the right side, it is possible that RI5 activity may also redirect hemolymph to the male genital organs during male copulation. Presently, it is unknown whether R15 is active during either copulation or egg laying. Alevizos et al. (1991 a, b) have found that R15 is not chronically active in vivo. They postulated that it is a conditional burster that becomes active only for relatively brief periods during certain behaviors. Ongoing experiments are directed at determining whether R15 bursts during egg-laying and/or copulation, when its synaptic actions may serve to facilitate these behaviors. The role of the dense R15~1 peptide-like IR material in the nerve terminals of the proximal anterior aorta and its branches is unclear. In this study and in previous studies (Alevizos, personal communication) it was found that R15 activity, following the hyperpolarization protocol, only occasionally caused a measurable change in proximal arterial tone. This may be because the neuron normally modulates the motor effects of another (un-

154

M.E. Skelton and J. Koester: Morphology and neural control of Aplysia arterial system

k n o w n ) neuron. Alternatively, R15 m a y use the proximal vessels as a n e u r o h e m a l release site in order to distribute peptide into the general circulation, a n d the occasional m o t o r effects observed m a y be caused by indirect stimulation o f a n o t h e r vasoactive neuron.

P As activity adapts the anterior arterial system to postural changes M o s t molluscs that rely on a hydrostatic skeleton, including Aplysia, u n d e r g o large changes in posture, presumably necessitating adjustment o f the length o f m a j o r vessels comprising the circulatory system. We f o u n d that the PAS neurons are reciprocally active during turns to the right and left and simultaneously active during behaviors involving head withdrawal, such as respiratory p u m p i n g , escape stepping and defensive withdrawal. The head and neck region o f Aplysia changes in length and f o r m dramatically when the animal makes postural changes such as those associated with l o c o m o t i o n (Hening et al. 1979), defensive withdrawal (Waiters and Erickson 1986), f o o d - i n d u c e d arousal (Teyke et al. 1990) or egg-laying behavior (Ferguson et al. 1989). We postulate that the function o f the PAS neurons is to maintain the distal p o r t i o n o f the anterior a o r t a and the pedalp a r a p o d i a l arteries at the a p p r o p r i a t e length during m o v e m e n t s o f the head and neck region. This would help to ensure that the flow o f h e m o l y m p h via these vessels is not disrupted due to arterial kinking. It m a y be significant that the innervation o f the longitudinal muscles occurred at regions where the anterior a o r t a gave rise to m a j o r branches (this data and S a w a d a et al. 1981 b). Therefore, prevention o f kinking o f u n b r a n c h e d portions o f the a o r t a m a y n o t be as i m p o r t a n t as near the b r a n c h e d portions, which m a y be m o r e susceptible to occlusion. The hypothesis that the LPAs neurons adjust the arterial system to changes in posture is supported by recent w o r k by Y. Xin and I. K u p f e r m a n n (personal c o m m u n i c a t i o n ) . They f o u n d that PAS neurons are indirectly inhibited in vitro by the cerebral-pedal regulator ( C P R ) neurons, which are c o m m a n d neurons that initiate head lifting (involving stretching o f the neck) during f o o d - i n d u c e d arousal (Teyke et al. 1990).

Acknowledgements. We thank Drs. A. Alevizos and D. Karagogeos for providing antibodies to R15c~-peptides, and Dr. I. Kupfermann and N. McKay for critically reviewing an earlier draft of this paper. This work was supported by Scope E of NIH grant. GM32099.

References Adams WB, Benson JA (1985) The generation and modulation of endogenous rhythmicity in the Aplysia bursting pacemaker neurone RI 5. Prog Biophys Mol Biol 46:1-49 Alevizos A, Bailey CH, Chen M, Koester J (1989a) Innervation of vascular and cardiac muscle of Aplysia by multimodal motoneuron L7. J Neurophysiol 61 : 1053-1063 Alevizos A, Weiss KR, Koester J (1989b) SCP-containing R20 neurons modulate respiratory pumping in Aplysia. J Neurosci 9: 3058-307I

Alevizos A, Karagogeos D, Weiss KR, Buck L, Koester J (1991 a) RI5~l and R15ct2 peptides from Aplysia: comparison of bioactivity, distribution and function of the two peptides generated by alternative splicing. J Neurobiol 22:405-417 Alevizos A, Skelton M, Karagogeos D, Weiss KR, Koester J (1991) Physiological role of bursting neuron R15 of Aplysia in the control of egg laying behavior. In: Kits KS, Boer HH, Joose J (eds), Molluscan neurobiology. North Holland Publishing Company, Amsterdam, pp 61-66 Alevizos A, Weiss KR, Koester J 0991 b) Synaptic actions of identified peptidergic neuron R15 in Aplysia. I. Activation of respiratory pumping. J Neurosci 11:1263-1274 Alevizos A, Weiss KR, Koester J (1991 c) Synaptic actions of identified peptidergic neuron R 15 in Aplysia. II. Contraction of pleuroabdominal connectives mediated by motoneuron L7. J Neurosci 11 : 1275-1281 Alevizos A, Weiss KR, Koester J (1991 d) Synaptic actions of identified peptidergic neuron R15 in Aplysia. III. Activation of the large hermaphroditic duct. J Neurosci 11:1282-1290 Arch S, Smock T (1977) Egg-laying behavior in Aplysia californica. Behav Biol 19:45 54 Ascher P, Kehoe JS (1975) Amine and amino acid receptors in gastropod neurons. In: Iverson LL, Iverson SD, Snyder SH (eds) Handbook of psychopharmacology, vol 4. Plenum, New York, pp 265-310 Barber A (1983) A monoamine-detecting histofluorescence technique for use on wholemounts of molluscan tissues. J Neurosci Meth 8 : 171-175 Branton WD, Arch S, Smock T, Mayeri E (1978) Evidence for mediation of a neuronal interaction by a behaviorally active peptide. Proc Natl Acad Sci USA 75:5732-5736 Buck LB, Bigelow JM, Axel R (1987) Alternative splicing in individual Aplysia neurons generates neuropeptide diversity. Cell 51:127-133 Byrne J, Koester J (1978) Respiratory pumping: neuronal control of a centrally commanded behavior in Aplysia. Brain 143 : 87105 Carpenter D, Breese G, Schanberg S, Kopin I (1971) Serotonin and dopamine: distribution and accumulation in Aplysia nervous and non-nervous tissues. Int J Neurosci 2 : 49-56 Cropper EC, Miller MW, Tenenbaum R, Kolks-Gawinowicz MA, Kupfermann I, Weiss KR (1988) Structure and action of buccalin: a modulatory neuropeptide localized to an identified small cardioactive peptide-containing cholinergic motoneuron of Aplysia californica. Proc Natl Acad Sci USA 85:6177-6181 De la Torre JC, Surgeon JW (1976) A methodological approach to rapid and sensitive monoamine histofluorescence using a modified glyoxylic acid technique: the SPG method. Histochemistry 49: 81-93 Dieringer N, Koester J, Weiss K (1978) Adaptive changes in heart rate of Aplysia californica. J Comp Physiol 123:11-21 Dudek FE, Cobbs JS, Pinsker HM (1979) Bag cell electrical activity underlying spontaneous egg laying in freely behaving Aplysia brasiliana. J Neurophysiol 42:804-817 Eales NB (1921) Aplysia. Liverpool Mar Biol Comm 1921, memoirs No. 24, pp 183-266 Ferguson GP, Ter Maat A, Parsons WD, Pinsker HM (1989) Egglaying in Aplysia. I. Behavioral patterns and muscle activity of freely behaving animals after selectively elicited bag cell discharges. J Comp Physiol A 164:835-847 Goldstein RS (1984) Immunocytochemical, histofluorescent, and ultrastructural studies ofmonoaminergic neurons and their processes in Aplysia. PhD Dissertation, Columbia University Goldstein RS, Schwartz JH (1989) Catecholamine neurons in Aplysia: Improved light-microscopic resolution and ultrastructural study using paraformaldehyde and glutaraldehyde (FAGLU) cytochemistry. J Neurobiol 20: 203-218 Goldstein SR, Kistler HB, Steinbusch HWM, Schwartz JH (1984) Distribution of serotonin-immunoreactivity in juvenile Aplysia. Neuroscience 11 : 535-547

M.E. Skelton and J. Koester: Morphology and neural control of Aplysia arterial system Hawkins RB (1989) Localization of potential serotonergic facilitator neurons in Aplysia by glyoxylic acid histofluorescence combined with retrograde fluorescent labeling. J Neurosci 9:42144226 Hening WA (1982) Central generation and coordination of a complex behavioral sequence in Aplysia californica. PhD Dissertation, New York University, New York Hening WA, Walters ET, Carew TJ, Kandel ER (1979) Motoneuronal control of locomotion in Aplysia. Brain Res 179:231-252 Hill RB (1964) Some effects of acetylcholine and of active amines on the isolated ventricles of Aplysia dactylomela and Aplysia fasciata. Pubbl Staz Zool Napoli 34:75-85 Kandel ER (1979) Behavioral biology of Aplysia. Freeman, San Francisco Koch UT, Koester J (1982) Time sharing of heart power: cardiovascular adaptations to food-arousal in Aplysia. J Comp Physiol 149 : 31-42 Koester J (1991) Identification of a network of neurons in the abdominal ganglion of Aplysia that are immunoreactive for the R15~ peptides. Soc Neurosci Abstr 17:1355 Koester J, Koch UT (1987) Neural control of the circulatory system of Aplysia. Experientia 43 : 972 80 Kuramoto H, Yui R, Iwanaga T, Fujita T, Yanaihara N (1985) PHI-like immunoreactivity in the nervous system of the cockroach (insect) and Aplysia (mollusc) with special reference to its relationship to VIP-like immunoreactivity. Arch Histolog Jpn 48:427-433 Kuslansky B, Weiss KR, Kupfermann I (1987) Mechanisms underlying satiation of feeding behavior of the mollusc Aplysia. Behav Neural Biol 48 : 278-303 Liebeswar G, Goldman JE, Koester J, Mayeri E (1975) Neural control of circulation in Aplysia III. Neurotransmitters. J Neurophysiol 38 : 767-779 Ligman SH, Brownell PH (1985) Differential hormonal action of the bag cell neurons on the arterial system of Aplysia. J Comp Physiol A 157:31 37 Lloyd PE, Kupfermann I, Weiss KR (1985) Two endogenous neuropeptides (SCPA and SCPB) produce a cAMP-mediated stimulation of cardiac activity in Aplysia. J Comp Physiol A 156:659667 Longley RD, Longley AJ (1986) Serotonin immunoreactivity of neurons in the gastropod Aplysia californica. J Neurobiol 17 : 339-358 Mahon AC, Lloyd PE, Weiss KR, Kupfermann I, Scheller RH (1985) The small cardioactive peptides A and B of Aplysia are derived from a common precursor molecule. Proc Natl Acad Sci USA 82 : 3925 3929 Mayeri E, Koester J, Kupfermann I, Liebeswar G, Kandel ER (1974) Neural control of circulation in Aplysia. I. Motoneurons. J Neurophysiol 37:458-475

155

Miller MW, Alevizos A, Cropper EC, Vilim F, Karagogeos D, Kupfermann I, Weiss KR (1991) Localization of myomodulinlike immunoreactivity in the central nervous system and peripheral tissues of Aplysia californica. J Comp Neurol 313:1-8 Miller MW, Alevizos A, Cropper EC, Kupfermann I, Weiss KR (1992) Localization of buccalin-like immunoreactivity in the central nervous system and peripheral tissues of Aplysia californica. J Comp Neurol (in press) Rittenhouse AR, Price CH (1985) Peripheral axons of the parabolic burster neuron R15. Brain Res 333 : 330-335 Sawada M, McAdoo D J, Blankenship JE, Price CH (1981 a) Modulation of arterial muscle contraction in Aplysia by glycine and neuron R14. Brain Res 207:486-490 Sawada M, Blankenship JE, MCAdoo DJ (1981 b) Neural control of a molluscan blood vessel, anterior aorta of Aplysia. J Neurophysiol 46: 967-986 Scott ML, Govind CK, Kirk MD (1991) Neuromuscular organization of the buccal system in Aplysia californica. J Comp Neurol 312: 207-222 Segal MM, Koester J (1982) Convergent cholinergic neurons produce similar postsynaptic actions in Aplysia: Implications for neural organization. J Neurophysiol 47 : 742-759 Skelton ME, Koester J (1991) Innervation of the heart of Aplysia by serotonin- and R15~l peptide-immunoreactive terminals, and by neuron L2. Soc Neurosci Abstr 17:1355 Skelton ME, Buck LB, Koester J (1989) Innervation of the anterior arterial system of Aplysia californica. Soc Neurosci Abstr 15:1298 Taussig R, Kaldany RR, Rothbard JB, Schoolnik G, Scheller R (1985) Expression of the Lll neuropeptide gene in the Aplysia central nervous system. J Comp Neurol 238 : 53-64 Teyke T, Weiss KR, Kupfermann I (1990) Appetitive feeding behavior of Aplysia: behavioral and neural analysis of directed head turning. J Neurosci 10:3922-3934 Tritt SH, Lowe IP, Byrne JH (1983) A modification of the glyoxilic acid induced histofluorescence technique for demonstration of catecholamines and serotonin in tissues of Aplysia californica. Brain Res 259:159-162 Waiters ET, Erickson MT (1986) Directional control and the functional organization of defense responses in Aplysia. J Comp Physiol A 159 : 339-351 Weiss KR, Bayley H, Lloyd PE, Tenenbaum R, Gawinowicz-Kolks MA, Buck L, Cropper E, Rosen SC, Kupfermann I (1989) Purification and sequencing of neuropeptides contained in neuron R15 of Aplysia californica. Proc Natl Acad Sci USA 86:29132917 Winkler LR (1957) The biology of California sea hares of the genus Aplysia. PhD Dissertation University of Southern California Wright DL (1960) Cardiac studies on Aplysia vaccaria. PhD Dissertation University of California, Los Angeles