Neuronal and neurohormonal control of the heart in the stomatopod crustacean, Squilla oratoria

Alexandrowicz (1934) first showed that three pairs of central cardioregulatory nerves (α, βand γ) control the heart of the stomatopod Squilla mantis. Of these, β and γ are most probably nerves originating from the 3rd and 4th segments of `the large thoracic ganglionic mass', while α is a nerve originating from a more anterior segment of the ganglionic mass. He proposed that the α nerve was a cardio-inhibitor while the β andγ nerves were cardio-accelerators. The functions of α, β andγ nerves in control of the heart were verified by means of intracellular recording from neuronal cell bodies in the cardiac ganglion of Squilla oratoria (Watanabe et al., 1968, 1969). It is known that the heartbeat of stomatopods is triggered by the cardiac ganglion, located on the outer surface of the heart (reviewed by Maynard, 1960). Neither the anatomical studies of Alexandrowicz(1934) nor the functional studies of Watanabe et al.(1968, 1969), have been followed by more detailed studies of stomatopod neuroanatomy, and until the present study,a detailed description of the anatomy of the central ganglia and of the ganglionic nerve roots from which the cardioregulatory axons emerge has not been published, nor have their cell bodies been identified in the central ganglia.

The neurotransmitter candidates for the intrinsic and extrinsic heart neurons have been identified in a few species within the sub-class Malacostraca from the results of pharmacological and immunocytochemical studies. Dealing with the decapods, Yazawa and Kuwasawa(1994) proposed thatγ-amino-butyric acid (GABA) and dopamine (DA) are the extrinsic neurotransmitters of the cardio-inhibitory and cardio-acceleratory nerves,respectively, in the hermit crab Aniculus aniculus. They also proposed that acetylcholine (ACh) and DA are the intrinsic neurotransmitters of the small and large neurons, respectively, in the cardiac ganglion of that species (Yazawa and Kuwasawa,1992). On the other hand, glutamate (Glu) has been proposed for the neurotransmitter of both of the small and large neurons in the cardiac ganglion of the lobster Panulirus argus(Delgado et al., 2000). In isopods, it has been proposed that Glu is the intrinsic excitatory neurotransmitter released by the motoneurons of the cardiac ganglion of Bathynomus doederleini (Yazawa et al., 1998) and Ligia exotica(Sakurai et al., 1998). Furthermore, GABA has been proposed as the extrinsic cardio-inhibitory neurotransmitter, while ACh has been proposed as the extrinsic cardio-acceleratory neurotransmitter in B. doederleini(Tanaka et al., 1992).

In stomatopods, epinephrine (E), norepinephrine (NE), and ACh have cardioexcitatory effects on the heart of S. mantis(Alexandrowicz and Carlisle,1953; Florey and Rathmayer,1990). Additionally, pharmacological experiments indicate that GABA may be a neurotransmitter of the extrinsic cardio-inhibitor nerves of S. oratoria (Watanabe et al.,1968). However, candidate neurotransmitters of the cardiac ganglion neurons and cardio-accelerator nerves have not yet been proposed for the stomatopod (recently reviewed by Cooke, 2002).

In this study, we describe the central ganglia and the nerve roots from which the cardioregulatory axons emerge, and identify their cell bodies in the ganglia. Furthermore, we identify neurotransmitter candidates for all the cardioregulatory neurons and neurons of the cardiac ganglion, from the results of electrophysiological, pharmacological and immunocytochemical studies. Preliminary reports have appeared elsewhere in abstract form(Ando and Kuwasawa, 1993; Ando and Kuwasawa, 1994; Ando et al., 1995).

The stomatopods, Squilla oratoria De Haan used in this study were about 10–15 cm in body length and about 20 g in mass. Specimens were collected in fishing trawl nets in Tokyo Bay and kept for a few days to several months in a laboratory aquarium maintained at 13–15°C.

Before dissection, the animals were anesthetized by injection of an isotonic MgCl₂ (0.36 mol l^–1) solution, through a syringe inserted into the thorax. After anesthetization, they were pinned to the Silpot (Dow Corning, Kanagawa, Japan)-lined bottom of a chamber filled with cold seawater, kept in a refrigerator. The carapace in the cephalo-thorax was carefully peeled off to expose the terminal region of the cardioregulatory nerves on the heart inside the pericardium. The peripheral cardioregulatory nerves were stained by means of a vital staining technique, in Methylene Blue filtered seawater (SW). The stained preparations were fixed in a 4% ammonium molybdate–SW solution overnight and washed in running tapwater for several hours (Alexandrowicz,1932). Then, the fixed preparations were stained in a Methylene Blue–tapwater solution (Kihara and Kuwasawa, 1984), and dissection was continued to trace the cardioregulatory nerves up to the central ganglia.

According to their functions, as elucidated by physiological experiments(see Electrophysiology section below), we refer to the nerves named α,β and γ by Alexandrowicz(1934) as, respectively, the cardio-inhibitor (CI) nerve, and the 1st and 2nd cardio-accelerator (CA1 and CA2) nerves. Those are terms that have been commonly used in a variety of crustacean species: in the decapods(Maynard, 1953; Yazawa and Kuwasawa, 1984) and in the isopods (Kihara and Kuwasawa,1984; Sakurai and Yamagishi,1998).

After the heart and cardioregulatory nerves were exposed, the nerves were cut in the pericardial cavity. The isolated heart was pinned dorsal side up to a Silpot-lined experimental bath (3.5 ml). The distal cut-stumps of the nerves, extending to the heart, were introduced one by one into a glass capillary suction electrode containing a Ag–AgCl wire, used to apply stimulus pulses to the nerve. Heartbeat of the whole heart was monitored using either a mechanogram, with a strain gauge transducer (TB-611T Nihon Kohden,Tokyo, Japan) or an electrocardiogram (ECG), which was recorded from the outer surface of the heart with a glass capillary suction electrode.

Myocardial intracellular potentials were recorded from the isolated heart,which was opened by longitudinal incision of the ventral wall of the heart. After incision, the heart was spread out and then pinned dorsal side up to a Silpot-lined experimental bath (1.5 ml). Intracellular potentials were recorded with a conventional glass capillary microelectrode filled with 3 mol l^–1 KCl (tip resistance, 20–35 MΩ). Excitatory junctional potentials (EJPs) in myocardial cells were evoked by electrical stimuli applied to the cardiac ganglion at the cut stump of the trunk of the ganglion. Preparations were routinely perfused with Squilla saline(in mmol l^–1: NaCl, 450; KCl, 15; CaCl₂, 10;MgCl₂, 20; Hepes, 5; pH 7.8) by gravity-feeding through a cannula at the rate of 3–4 ml min^–1. Saline was suctioned for removal. The saline was modified from the saline of Watanabe et al.(1967). The perfusate was maintained at 17–20°C by a cooling device (EC-201, Scinics, Tokyo,Japan).

The following agents were used: acetylcholine chloride (ACh), atropine sulfate monohydrate, dopamine hydrochloride (DA),γ-amino-N-butyric acid (GABA), sodium l-glutamate monohydrate, histamine dihydrochloride, joro spider toxin 3 (JSTX), l-noradrenaline bitartrate, picrotoxin (PTX),serotonin–creatinine sulfate, serotonin (5-HT), d-tubocurarine hydrochloride (Wako Pure Chemical Industries, Osaka,Japan), adrenaline bitartrate, chlorpromazine hydrochloride, octopamine hydrochloride (Sigma Chemicals, St Louis, MO, USA), phentolamine mesylate,Regitin (Ciba-Geigy, Basle, Switzerland).

The salines containing the agents were applied to preparations when the cock of a three-way valve was turned in the perfusion line, except for JSTX,when saline containing the agent was added to the experimental bath by a pipette.

Pieces of nerve, isolated from the CI, CA1 and CA2, were pre-fixed in a 3%(w/v) glutaraldehyde solution and post-fixed in a 0.1% (w/v) osmium solution. After washing in the buffer solution and dehydrating in a graded ethanol series, they were embedded in Quetol-812 (Nissin EM, Tokyo, Japan). Sections were cut at 1 μm and stained with the mixed 1% (w/v) Methylene Blue and 0.1% (w/v) Azur II solutions. They were observed under a microscope and photographed (BX-51, Olympus, Tokyo, Japan).

In order to locate the cell bodies of the CI, CA1 and CA2 neurons in the central nervous system (CNS), a proximal cut-stump of the cardioregulatory nerve was introduced into a glass capillary filled with a mixed solution of 1 mol l^–1 CoCl₂ and 1 mol l^–1NiCl₂. The preparation was incubated for 4–8 days at 4°C,and then rinsed with SW. Some drops of a saturated Rubeanic Acid–ethanol solution were added to precipitate a sulfate compound(Quicke and Brace, 1979). The stained preparation was fixed with 4% (w/v) formaldehyde, dehydrated with a graded ethanol series, and cleared with methyl salicylate. The preparation was observed under a microscope and photographed (BX-51, Olympus).

We examined immunoreactivity of neural processes in the heart and extrinsic nerves against four kinds of antibodies: anti-GABA, anti-glutamate,anti-serotonin and anti-histamine. In immunocytochemistry using anti-GABA and anti-glutamate, preparations were processed as described for the isopod by F.-Tsukamoto and Kuwasawa(2003). For controls,preparations were processed in the same manner but without treatment with the primary antibodies. For each antibody, no immunoreactivity was detected in the control preparations. The preparations were observed under a microscope and photographed (BX-51, Olympus).

To locate GABA-like immunoreactive neural processes, specimens were fixed with a solution of 4% (w/v) paraformaldehyde–0.1% glutaraldehyde in 0.1 mol l^–1 phosphate buffer (pH 7.2) containing 15% (w/v)sucrose for 2–3 h at 4°C. After fixation they were rinsed in a 0.1 mol l^–1 phosphate buffer solution containing 15% (w/v)sucrose over 1 day at 4°C. The specimens were embedded in paraffin and sectioned serially at 10 μm. The sections were dried on the slides,deparaffinized, rehydrated and immersed in distilled water (DW). In order to eliminate intrinsic peroxidase activity in the sectioned tissues, the slides were treated with 0.3% (w/v) H₂O₂ in DW for 30 min at room temperature. They were incubated with a 0.1 mol l^–1phosphate buffer containing 0.1% (w/v) Triton X-100 (pH 7.2) (0.1% PBT) for 15 min, then with the primary antibody (rabbit anti-GABA; cat. no. 20094,Incstar, Stillwater, MN, USA), diluted 1:2000 in a 0.1% PBT for 24 h and rinsed in a 0.1% PBT for 1 h. The specimens were treated with the secondary antibody (goat anti-rabbit IgG; Sigma), diluted 1:200 in a 0.1% PBT for 1.5–2 h at room temperature and rinsed in a 0.1% PBT for 1 h. They were treated with the third antibody [rabbit peroxidase–antiperoxidase (PAP)complex; Sigma], diluted 1: 200 in a 0.1% PBT for 2 h and rinsed in a 0.1% PBT for 1 h. 0.03% (w/v) 3,3′-diaminobenzidine-tetrahydrochloride (DAB) in 0.05 mol l^–1 Tris buffer (pH 7.6) containing 0.006% (w/v)H₂O₂ was applied for about 10 min at room temperature in dark. The peroxidase reaction was stopped by transferring the slides to DW. The sections were counterstained with Methyl Green, dehydrated in a graded ethanol series, cleared in xylene and mounted in Bioleit (Oken Shoji, Tokyo,Japan).

For whole-mount preparations, fixed specimens were immersed in 0.2% (w/v)collagenase in SW, rinsed in 0.3% PBT, and immersed in 1% PBT overnight. They were incubated with the primary antibody diluted 1:1000 in a 0.3% PBT for 48 h, with the secondary antibody for 2–4 h andfurther with PAP complex for 2–4 h. After each incubation they were rinsed in 0.3% PBT for 1 h. The specimens were stained with DAB solutions, dehydrated with a graded ethanol series, cleared in methyl salicylate and then in xylene and mounted on glass slides in Bioleit.

Specimens were treated as described for GABA immunocytochemistry before treatment with the antibodies. Specimens were treated with the primary antibody (mouse anti-glutamate; cat. no. 22523, Incstar or cat. no. G9282,Sigma) diluted 1:1000 (Incstar) or 1:20000 (Sigma) in a 0.1% PBT. The specimens were treated with the secondary antibody (goat anti-mouse IgG;Sigma), diluted 1:200 in a 0.1% PBT and with the third antibody (mouse PAP complex; Sigma), diluted 1:200 in a 0.1% PBT. Then the specimens were processed and examined as described above.

To locate serotonin-like immunoreactive neural cells and processes,specimens were fixed with a solution of 4% (w/v) paraformaldehyde in 0.1 mol l^–1 phosphate buffer containing 15% (w/v) sucrose (pH 6.5)for 30 min at room temperature, followed by a 4% (w/v) paraformaldehyde in 0.1 mol l^–1 sodium tetraborate buffer containing 15% (w/v)sucrose (pH 9.0) for 2.5–3 h at 4°C. The specimens were treated with the primary antibody (rabbit anti-serotonin; cat. no. 2008, Incstar), diluted 1:2000 in a 0.1% PBT for 24 h and rinsed in a 0.1% PBT. Then, the specimens were processed and examined as described above.

To locate histamine-like immunoreactive neural cells and processes,specimens were fixed with a solution of 4% (w/v) ethyl-dimethylcarbodiimide in 0.1 mol l^–1 phosphate buffer (pH 7.3) for 60 min at 4°C,followed by a 4% (w/v) paraformaldehyde in 0.1 mol l^–1phosphate buffer containing 15% sucrose (pH 7.2) overnight at 4°C. The specimens were treated with the primary antibody (rabbit anti-histamine; cat. no. 22939, Incstar), diluted 1:1000 in a 0.1% PBT for 24 h and rinsed in a 0.1% PBT. Then, the specimens were processed and examined as described above.

In stomatopods, the cephalon and the 1st to 4th thoracic segments are fused to form the cephalo-thorax, which is followed by the 5th to 8th thoracic segments, the 1st to 6th abdominal segments and the telson(McLaughlin, 1980). The topographical relationship between the exoskeleton, the heart and CNS in S. oratoria is shown in Fig. 1. The heart lies over a range from the 2nd maxilla in the cephalo-thorax to the 5th abdominal segment. The anterior artery (AA), 15 pairs of the lateral arteries (LA1-15) and the posterior artery (PA) arise from the heart. The subesophageal ganglion (SEG) is connected with the cerebral ganglion by the circumesophageal nerve ring and is followed by the thoracic ganglia and abdominal ganglia.

The cardiac ganglion (CG), which lies on the outer surface of the heart,consists of 13–16 ganglion cells in stomatopods(Alexandrowicz, 1934; Irisawa and Irisawa, 1957; Brown, 1964a; Watanabe et al., 1967; Florey and Rathmayer, 1990). We used Methylene Blue staining techniques to determine the number of neurons in the CG of the present material. We counted 15 cell bodies in the area of the CG between the 2nd and 14th lateral arteries for more than 50 preparations, which were well stained with the dye. Three of the 15 cell bodies formed a cluster at the base of the 2nd arteries, and the other 12 cell bodies were individually located at each base of the lateral arteries. The number of cell bodies coincides with that of pairs of the lateral arteries. Fig. 2 shows the cardiac ganglion cells in the CG in the anterior part of the heart. Three anterior cell bodies of the ganglion lie forming a cluster at the 2nd lateral artery.

Three pairs of cardioregulatory nerves join the CG at the level of the origins of the 2nd, 3rd and 4th lateral arteries in the cephalo-thorax, for CI, CA1 and CA2 nerves, respectively (Fig. 2A–C).

Anatomically, we traced the nerves all the way to the central nervous system from which they arose, and located their origins in the SEG. We decided that the CI, CA1 and CA2 nerves emerged from, respectively, the 10th the 16th and 19th nerve roots (10th NR, 16th NR and 19th NR) of the SEG(Fig. 3A). In order to verify these findings, we observed the effects of stimulation of their roots on heart beat.

When the 10th NR was electrically stimulated at a distal cut-stump of the nerve, heart rate decreased (Fig. 3B). The effect was intensified with increased stimulus frequency(Fig. 3C). By contrast, when either the 16th or 19th NR was electrically stimulated at a distal cut-stump of each nerve, heart rate and amplitude of action potentials increased(Fig. 3B). The acceleratory effect was intensified with increased stimulus frequency(Fig. 3C). Fig. 3C shows the effects of stimulation of the nerve roots on heart rate. Heart rate in controls was 33.2±8.9 for the 10th NR, 30.0±7.7 for the 16th NR and 26.6±6.1 for the 19th NR (mean ± s.d.; N=7 for the 10th NR, N=9 for the 16th NR, N=10 for the 19th NR). The increased amplitude of action potential, during the stimulations, may indicate an increase of contraction force of the heart. Thus, the anatomical results described above were confirmed by these results.

In order to examine the axonal composition of the cardioregulatory nerves,extracellular impulses were recorded from the nerves. Only one kind of orthodromic (Fig. 4A orth) or antidromic unit impulse (Fig. 4A ant) was recorded for CI, CA1 and CA2. Additional impulse units never appeared, even when stimulated at higher intensity. These results were confirmed in seven preparations. Resin cross sections obtained from the CI,CA1 and CA2 nerves were observed under a microscope(Fig. 4B), and showed that each of the CI, CA1 and CA2 nerves contains one axon (6–8 μm in diameter)wrapped in the heavily stained perineurium (about 1 μm in thickness), which is surrounded by epineurium (10–20 μm in thickness), thus forming the nerves.

The axons were back-filled with Ni²⁺ and Co²⁺ ions at their proximal cut-stumps. The cell body (about 30 μm in diameter) of the CI axon was found stained at a site near the midline on the side contralateral to the nerve root through which the CI axon emerged in the 1st SEG segment(Fig. 5). The cell body (about 20 μm in diameter) of the CA1 axon was found stained at a site near the midline on the side ipsilateral to the nerve root through which the CA1 axon emerged in the 3rd SEG segment. The cell body (20 μm in diameter) of the CA2 axon was found stained at a site near the midline on the side ipsilateral to the nerve root through which the CA2 axon emerged in the 4th SEG segment.

In order to examine the effects on the heart of putative neurotransmitter substances known in crustaceans (Cooke and Sullivan, 1982), mechanograms of heartbeat were recorded from isolated hearts treated with the substances.

Effects of GABA and histamine on the heart are shown in Fig. 6. GABA(>10^–6 mol l^–1) decreased heart rate and contraction force depending on the dose(Fig. 6A). Similar results were reported by Watanabe et al.(1968) for GABA application to the heart of S. oratoria. Histamine (HA) (>10^–6mol l^–1) also exerted inhibitory effects on the heart rate and contraction force (Fig. 6B). Picrotoxin, a GABAergic antagonist, at 10^–4mol l^–1, completely blocked the cardiac inhibition induced by stimulation of the CI (Fig. 6C).

5-HT, DA, OA, ACh and Glu were applied to the isolated heart, and dose–response curves are shown in Fig. 7. 5-HT initially transiently decreased both heart rate and contraction force, and then increased both heart rate and contraction force(Fig. 7A). The effects of 5-HT on heart rate and contraction force increased in a dose-dependent manner. Threshold concentrations for 5-HT in both the first and second phases were between 10^–9 and 10^–8 mol l^–1. DA (Fig. 7B), OA (Fig. 7C),ACh (Fig. 7D) and Glu(Fig. 7E) increased heart rate and force. Threshold concentrations for both DA and OA were approximately 10^–8 mol l^–1, and those for both ACh and Glu around 10^–6 mol l^–1. Epinephrine (E) and norepinephrine (NE) showed weak excitatory effects compared to the effects induced by stimuli of 5-HT, DA and OA (data not shown).

Antagonists for various putative neurotransmitter substances were tested in order to predict the neurotransmitters of the extrinsic CA axons and cardiac ganglion neurons. The catecholaminergic blockers, chlorpromazine 10^–5 mol l^–1 and phentolamine 10^–5 mol l^–1, did not antagonize the cardio-acceleratory effects of CA nerves and force of the myocardium beats induced by the cardiac ganglion. The cholinergic blockers, atropine 10^–4 mol l^–1 and d-tubocurarine 10^–4 mol l^–1, were also ineffective on the acceleratory effects produced by stimulation of the CA nerves (data not shown). Three preparations were used for each of the agents.

JSTX (10^–5 mol l^–1), known as a glutamate antagonist (Kawai et al.,1982; Chiba and Tazaki,1992; Sakurai et al.,1998; F.-Tsukamoto and Kuwasawa, 2003), blocked cardio-acceleratory effects of stimulation of the CA1 and CA2 axon (Fig. 8). Acceleratory effects of stimulation of CA1 or CA2 decreased to 38.5±12.7% of control value for CA1 and 24.5±17.5% for CA2(means ± s.d., N=3).

5×10^–6 mol l^–1 JSTX blocked EJPs in cardiac muscle, evoked by stimulation of the cardiac ganglion trunk(Fig. 9). The amplitude of EJPs decreased to 29.1±14.6% of control values (mean ± s.d., N=9). EJPs partially recovered at 3 h after washout of JSTX.

No immunoreactivity was detected in a negative control preparation of wholemount preparations or paraffin sections incubated without the primary anti-GABA and anti-glutamate antibodies.

The CI axon showed GABA-like immunoreactivity to anti-GABA antibodies(Fig. 10A). An axon showing GABA-like immunoreactivity runs in the CG(Fig. 10B). Although many histamine-like immunoreactive neuronal processes, as well as GABA-like immunoreactive ones, were observed in the SEG, the CI nerves showed no histamine-like immunoreactivity to anti-histamine antibodies (data not shown).

Fig. 11 shows glutamate-like immunoreactive neuronal processes, evoked in response to anti-glutamate antibodies. Both the CA1 and CA2 axons and the CG showed glutamate-like immunoreactivity (Fig. 11A). A glutamate-like immunoreactive cell body was observed in Fig. 10B. Glutamate-like immunoreactivity was shown in a major motor branch extending from the CG(Fig. 11C). Glutamate-like imunoreactivity was observed in all cells in the CG (data not shown). The immunoreactivity was not observed in the cardioregulatory nerves and in the heart (data not shown), but it was observed that serotonin-like immunoreactive neuronal processes ran out of the SEG and extended on the ventral skeletal muscles.

The cell bodies of cardioregulatory neurons have not yet been identified in the CNS of any decapod species, although it has been shown that a pair of the CI nerves and two pairs of the CA nerves emerge from the thoracic ganglion in Panulirus argus (Maynard,1953), Sesarma dehaani(Ikeya et al., 1984) and Aniculus aniculus (Miyazaki et al., 1985). In isopods, the neuronal cell bodies of a pair of CI axons, and of two pairs of CA (CA1 and CA2) axons, have been shown to be located in the 1st, 2nd and 3rd thoracic ganglia of Bathynomus doederleini (Tanaka and Kuwasawa, 1991a,b),and Ligia exotica (Sakurai and Yamagishi, 1998; Sakurai et al., 1999). In S. oratoria, we refer to the 1st ganglionic mass of the ventral nerve cord as the SEG, which consists of six segments. The cell bodies of the CI neurons are located in the 1st SEG segment. They extend their axons out of the SEG through the 10th nerve root contralateral to the cell bodies (Fig. 5), so that the axons cross over the midline. The cell bodies of both the CA1 and CA2 neurons are located in the 3rd and 4th SEG segments,respectively. The neurons extend their axons out of the SEG through the 16th and 19th nerve roots, respectively, ipsilateral to their cell bodies. These anatomical arrangements for cardioregulatory neurons in S. oratoriaare similar to those suggested previously in the isopods mentioned above. In both the isopods and stomatopods, the cardioregulatory cell bodies of the inhibitory, 1st and 2nd acceleratory neurons are located in successive separate central ganglia. The cell bodies of the CI neuron are located contralaterally to their axons, and the cell bodies of the CA1 and CA2 neurons are located ipsilaterally to their axons. If the posterior part of the 1st thoracic ganglion were to be regarded as the 2nd thoracic ganglion in the isopod, B. doederleini, then the ganglia in which the cell bodies of the CI, CA1 and CA2 neurons are located could be expressed as the 1st, 3rd and 4th thoracic ganglia, respectively (cf. Tanaka and Kuwasawa, 1991a,b),as they are in the present animal. On the other hand, the subesophageal ganglion of stomatopods consists of the ganglia that extend nerves not only to the mouth-parts, but also to the first five pairs of thoracopods(McLaughlin, 1983). Therefore,the portion anterior to the 1st thoracic ganglion in the present material might be regarded as the subesophageal ganglion, and the 1st, 3rd and 4th segments of subesophageal ganglion in S. oratoria as the thoracic ganglia. It is likely that the central neuronal arrangements for cardioregulatory neurons are conserved between the taxa.

In the present material, GABA and histamine produced inhibitory effects on the heartbeat (Fig. 6). Picrotoxin blocked neurally induced cardiac inhibition induced by impulses of the CI nerve (Fig. 6C). Though picrotoxin is not a GABA-specific antagonist but a chloride channel blocker,the CI nerves showed immunoreactivity to anti-GABA antibodies(Fig. 10) not to anti-histamine antibodies (data not shown). These results show that GABA may be a neurotransmitter for the CI neurons in S. oratoria. Neurotransmitters for cardioregulatory neurons and cardiac ganglion neurons have been proposed for some crustaceans. It has been proposed that GABA may be a neurotransmitter of the CI neuron in the isopod, B. doederleini(Tanaka et al., 1992; F.-Tsukamoto and Kuwasawa,2003), and in the decapods, A. aniculus(Yazawa and Kuwasawa, 1994)and Panulirus argus (Delgado et al., 2000). It is concluded that GABA may be commonly a cardio-inhibitory neurotransmitter in all those crustacean taxa.

The heart of S. mantis is activated by application of E, NE(Alexandrowicz and Carlisle,1953) and ACh (Florey and Rathmayer, 1990). While NA and DA are reported to produce cardio inhibitory activity on the heart of an isopod, Bathynomus doederleini, 5-TH, OA and Glu are reported to produce cardio-acceleratory activity (Tanaka et al., 1992; Yazawa et al., 1998; F.-Tsukamoto and Kuwasawa,2003). ACh has been proposed to be a neurotransmitter for cardio-accelerator nerves in B. doederleini(Tanaka et al., 1992), and DA to be a neurotransmitter for cardio-accelerator nerves in the decapod species, A. aniculus (Yazawa and Kuwasawa,1994) and the crab (Fort and Miller, 2001; reviewed by Cooke, 2002). In the present study the heart was activated by 5-HT, DA, OA and Glu, as well as E, NE and ACh (Fig. 7). No serotonin-like immunoreactivity was observed in CA axons and cardiac ganglion neurons (data not shown), although 5-HT activated heartbeat(Fig. 7A). Cardiac acceleration induced by stimulation of the CA nerves was not antagonized by cholinergic blockers, atropine and d-tubocurarine, or by catecholaminergic blockers,chlorpromazine and phentolamine (data not shown). Thus, ACh, 5-HT and catecholamines may be excluded from the list of transmitter candidates for the CA neurons.

Glu increased not only contraction force but also heart rate(Fig. 7E). JSTX blocked cardiac acceleration induced by the CA1 and CA2 nerves(Fig. 8). Glutamate-like immunoreactivity was observed in both the CA1 and CA2 axons(Fig. 11). JSTX is known to antagonize glutamatergic actions on the skeletal muscle of the lobster Palinurus japonicus (Kawai et al., 1982), on the stomach muscle of S. oratoria(Chiba and Tazaki, 1992), the heart of L. exotica (Sakurai et al., 1998) and on the arterial valve in B. doederleini(F.-Tsukamoto and Kuwasawa,2003). These results may indicate that, in S. oratoria,both the CA1 and CA2 neurons are glutamatergic. It has been suggested that DA may be a neurotransmitter of the CA nerves in the hermit crab, A. aniculus (Yazawa and Kuwasawa,1994) and the crab (Fort and Miller, 2001; reviewed by Cooke, 2002), while Glu had little effect on the cardiac ganglion (Yazawa and Kuwasawa, 1992, 1994). In the isopod, B doederleini, ACh may be a neurotransmitter of the CA neurons(Tanaka et al., 1992).

Thus, it seems that the transmitters of cardio-accelerator nerves may be varied among the crustacean taxa.

EJPs in the myocardium induced by motor impulses of the cardiac ganglion were blocked by JSTX (Fig. 9). Glutamate-like immunoreactivity was observed in all of the 15 cardiac ganglion neurons (Fig. 11). On the other hand, heartbeat triggered by the cardiac ganglion neurons was not antagonized by the cholinergic blockers and catecholaminergic blockers (data not shown). These results may indicate that, in S. oratoria, the cardiac ganglion neurons are glutamatergic. Pharmacological experiments suggested that ACh and DA may be neurotransmitters of small (pacemaker) and large (motor) neurons of the cardiac ganglion, respectively, although Glu exerted little effect on the cardiac ganglion and cardiac muscle in the hermit crab A. aniculus (Yazawa and Kuwasawa, 1992, 1994). On the other hand,glutamate-like immunoreactivity has been observed in the small and large neurons of the cardiac ganglion in the lobster P. argus(Delgado et al., 2000). A neurotransmitter of the cardiac ganglion neurons appears to differ in the anomuran (A. aniculus) and palinuran (P. argus) decapod species examined. Since Glu increased contraction force but not heart rate in the isopod, B. doederleini, Glu was proposed to be a neurotransmitter of the cardiac ganglion neurons (Yazawa et al., 1998). Glutamate-like immunoreactivity has been observed in the motor neurons of the cardiac ganglion in the isopod, L. exotica(Sakurai et al., 1998). Glutamate may be a neurotransmitter of the cardiac ganglion neurons of S. oratoria as of the case in a decapod, P. argus, and in isopods, B. doederleini and L. exotica.

Alexandrowicz (1952)observed that, in S. mantis, unpaired nerves ran in the connectives between the subesophageal, thoracic and abdominal ganglia, and extended toward the ventral body muscles. He called this nervous system `the system of the median connectives' and suggested that it had a neurosecretory function. He(Alexandrowicz, 1953) also observed neuropile-like networks on the inner surface of the pericardium in S. mantis and referred to it as the pericardial organ. The pericardial organ contained neuronal cell bodies and endings of nerves extending from the CNS. Pericardial organ extracts increased heart rate and contraction force in S. mantis(Alexandrowicz and Carlisle,1953; Brown,1964b). Substances contained in the pericardial organ and in the system of the median connectives have not yet been examined in any stomatopod species. The decapod pericardial organ contains 5-HT, DA, OA and several neuropeptides (Cooke and Sullivan,1982; Cooke, 1988; Christie et al., 1995;reviewed by Cooke, 2002). In the present species, 5-HT, DA and OA increased heart rate and contraction force (Fig. 7A–C). Since we observed serotonin-like immunoreactivity in the system of the median connectives as well as the CNS (data not shown), 5-HT may be one of the cardioexcitatory neurohormones. ACh may be a neurotransmitter for small neurons of the pacemaker in the cardiac ganglion of the hermit crabs(Yazawa and Kuwasawa, 1992)and for the CA axons of the isopod (Tanaka et al., 1992). It is suggested that in the stomatopod a transmitter of CA neurons is Glu but not ACh, although ACh increased both heart rate and contraction force (Fig. 7D). It is of interest for further study to determine whether ACh is contained as a neurohormonal substance in the pericardial organ of S. oratoria.

Cardio-inhibitory substances other than GABA have not previously been reported in any stomatopod species. This study showed that HA induced inhibitory effects on the heart (Fig. 6B). However, no histamine-like immunoreactivity was observed in the CI neurons. Since many histamine-like immunoreactive neurons were located throughout the CNS in the present material (data not shown), it cannot be excluded that HA is possibly a cardio-inhibitory neurohormone, as proposed for Homarus americanus(Hashemzadeh-Gargari and Freschi,1992).

List of abbreviations

 
• AA

  anterior artery

 
• ACh

  acetylcholine

 
• AG1-6

  1st to 6th abdominal ganglia

 
• CA1 and CA2

  1st and 2nd cardio-accelerator

 
• Cer. G

  cerebral ganglion

 
• Ceph.-Thorax

  cephalo-thorax

 
• CG

  cardiac ganglion

 
• CGC

  cardiac ganglion cell

 
• CGT

  cardiac ganglion trunk

 
• CI

  cardio-inhibitor

 
• CNS

  central nervous system

 
• DA

  dopamine

 
• DAB

  3,3′-diaminobenzidine-tetrahydrochloride

 
• DW

  distilled water

 
• E

  epinephrine

 
• ECG

  electrocardiogram

 
• EJP

  excitatory junctional potential

 
• GABA

  γ-amino-butyric acid

 
• Glu

  glutamate

 
• HA

  histamine

 
• 5-HT

  serotonin

 
• JSTX

  joro-spider toxin

 
• LA1-15

  1st to 15th lateral arteries

 
• MB

  major branch

 
• NB

  neural branch

 
• NE

  norepinephrine

 
• NR

  nerve root

 
• PA

  posterior artery

 
• PAP

  peroxidase–antiperoxidase

 
• PBT

  phosphate buffer containing Triton X-100

 
• PTX

  picrotoxin

 
• SEG

  subesophageal ganglion

 
• SW

  seawater

 
• TG7-9

  7th to 9th thoracic ganglia

This study was partly supported by the grant from JSPS No. 12304050 (K.K.). We thank Mr K. Shishikura for generously collecting animals. We would like to thank Prof. R. B. Hill, University of Rhode Island, for revision of this manuscript. Contribution from the Shimoda Marine Research Center, No. 707.