INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

It has been well established that the neural circuitry in the animal brain is flexible and will be reconfigured according to various kinds of behaviors. One of the important general questions in neurobiology is how a specific sensory modality can alter behaviors by reconfiguring a well-defined neural network. The cellular mechanism for the modulation of neural circuits by mechanoafferent sensory inputs has been well studied in both vertebrate and invertebrate central pattern generator (CPG) networks (Grillner et al. 1995; Katz and Frost 1996; Marder and Calabrese 1996; Pearson 1993, 2000; Rosen et al. 2000a,b). However, the way in which specific types of sensory inputs, such as odors or tastes, can modulate the neural network has not yet been elucidated because of the complex processing mechanism.

In food preference behaviors, animals recognize an odor or a taste, decide between liking and aversion often with help of their memory, and sometimes change their preference by learning the taste. After this processing, the animals show multiple behaviors such as ingestion or rejection. Therefore the study of the neural mechanism for food preference can potentially contribute to our understanding of the reconfiguration of neural circuits according to a specific sensory modality. The gastropod mollusks are very useful models for the study of the neuronal basis of such complex behaviors (Kandel 1979). The behaviors have been well studied for odor preference (Audesirk 1975; Croll and Chase 1977, 1980; Peschel et al. 1996; Teyke 1995; Teyke et al. 1992; Willows 1978; Yamada et al. 1992) and taste preference (Chang and Gelperin 1980; Culligan and Gelperin 1983; Kemenes et al. 1986; Kupfermann and Carew 1974). For taste preference, the learning behaviors that associate specific tastes and textures with inedibility have been well studied (Botzer et al. 1998; Schwarz and Susswein 1986; Schwarz et al. 1988; Susswein et al. 1986). The processing neural network has also been well studied for odor preference (Gelperin and Tank 1990; Gelperin et al. 1993; Schütt and Basar 1994). In contrast, the neuronal basis of taste preference has been little studied except for the neural pathway for the chemosensory inputs of the primary food (Bicker et al. 1982; Delaney and Gelperin 1990; Kemenes et al. 1986, 2001; Rosen et al. 1991).

The marine gastropod Aplysia has been reported to show clear food preferences (Audesirk 1975; Carefoot 1967, 1970; Kupfermann and Carew 1974, Nagahama and Shin 1998). In Aplysia californica, the responses of the cerebral neurons to chemical taste stimulation have also been explored (Chiel et al. 1986; Fredman and Jahan-Parwar 1980; Jahan-Parwar 1972; Rosen et al. 1991; Teyke et al. 1990; Weiss et al. 1978, 1986a). Moreover, the neural mechanisms generating ingestion and rejection have been well studied. In this case, however, rejection is usually induced by tactile stimulation with objects such as polyethylene tubing (Chiel et al. 1986; Church and Lloyd 1994; Hurwitz et al. 1996; Kabotyanski et al. 1998; Kupfermann 1974; Morton and Chiel 1993a,b; Rosen et al. 1991, 2000b; Weiss et al. 1986a). In a Japanese species, A. kurodai, we have recently found that the animals feed well on Ulva but reject Gelidium or Pachydictyon with distinctive rhythmic patterned movements of the jaws and radula (Nagahama and Shin 1998). These responses can be induced not only by the natural seaweeds but also by extract solutions, suggesting that ingestion or rejection could be induced by taste alone. Therefore our system may be appropriate for the study of the neural mechanism generating taste preference behaviors.

We have previously demonstrated that the difference in patterned jaw movements during ingestion and rejection might be caused by a change in the onset time of the firing of the jaw-closing (JC) motor neurons at each depolarizing phase of the rhythmic responses (Nagahama and Shin 1998). During ingestion, delay of the onset time of the JC firing may be produced by the firing of multiaction MA neurons which monosynaptically inhibit the JC motor neurons (Nagahama and Takata 1988-1990). The MA neurons are probably equivalent to the B4/5 in A. californica (Gardner 1971). In an earlier report (Nagahama et al. 1999), we also demonstrated that advance of the onset time of the JC firing during rejection of Gelidium might be caused by transient suppression of the synaptic transmission from the MA neurons to the JC motor neurons. Our results also suggested that some cerebral neurons might directly modulate this synaptic transmission. Therefore in the current experiments we tried to find such cerebral neurons contributing to the reconfiguration of the basic feeding circuits for rejection of Gelidium. The purpose of the present study is to understand how the CPG network generating the firing patterns in the JC motor neurons is modulated by particular neurons activated by the taste of the seaweed.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals and seaweed

A. kurodai, weighing 30-400 g, were collected from the coast of Awaji Island or Fukui City. Animals were maintained at 14-16°C in aquaria filled with aerated and filtered artificial seawater (ASW). Two species of seaweed, Ulva pertusa (Ulva) and Gelidium amansii (Gelidium) were also collected at a location with many animals. These were maintained at room temperature in aquaria filled with ASW or stored below 20°C in a freezer, respectively. Ulva was used for animal food and was supplied once every day.

Preparations

In the present experiments, the reduced preparation was used as previously described (Nagahama et al. 1999). The animals were anesthetized by injection of isotonic MgCl₂ solution (25% of body weight) into the body cavity and dissected at room temperature. The preparation consisted of the head structures, cerebral ganglia, buccal ganglia, and buccal mass. The head structures, innervated by the upper labial (ULAB), anterior tentacular (AT), and lower labial (LLAB) nerves arising from the cerebral ganglia, included the anterior tentacles, lips, and perioral zone. The cerebral nerves, ULAB, AT, LLAB, and cerebral-buccal connectives (CBC), were left intact and other nerves were severed. The buccal mass was removed from the head structures and then cut into halves along the midline to separate the paired symmetrical buccal musculature innervated by buccal nerves 2 and 3 (n2 and n3). The peripheral buccal nerves, except n2 and n3, were severed. The nomenclature of the buccal nerves followed Gardner (1971). The head structures, cerebral ganglia, buccal ganglia, and paired buccal mass were separately pinned to the silicone elastomer (Sylgard, Dow) surface of each compartment of a Lucite recording chamber, and petroleum jelly (Vaseline) was placed on the partitions between the compartments as previously reported (Nagahama et al. 1999). The head structures were pinned with the ventral surface upward in a way that revealed the inner portion of the lips and the perioral zone. The cerebral ganglia and the buccal ganglia were pinned with the ventral surface upward or caudal surface upward, respectively. The sheath overlaying the cerebral and buccal ganglia was surgically removed. The anterior aorta supplying the tentacle-lip region was cannulated with polyethylene tubing (tip diameter, 50-100 µm) and perfused with fresh ASW (0.5-0.7 ml/min).

Taste solutions

For taste stimulation of the lips, solutions of freshly prepared seaweed extract mixed with ASW were used as previously described (Nagahama and Shin 1998; Nagahama et al. 1999). A 20 g sample of fresh seaweed was ground in a mortar and soaked for 30 min in 30 ml (Ulva) or 40 ml (Gelidium) of ASW at room temperature. The mixture was centrifuged (12,000 g, 20 min), and the supernatant was used for the taste stimulation. In the text, they are described as the original extract solutions (solutions with the original concentration). The Ulva extract was used for favorite taste that induced the ingestion-like response, whereas the Gelidium extract was used for aversive taste that induced the rejection-like response. These original extract solutions were sufficient to induce the stable rhythmic responses, whereas solutions diluted with ASW often failed to induce the rhythmic responses (Nagahama and Shin 1998). In the present experiments, we used two sizes of chambers depending on the size of the preparation, and the free bath sizes for taste stimulation around the head structure were 2-3 and 3-6 ml, respectively. A constant volume (1-2 ml) of seaweed extract solutions for each preparation was gently applied into the bath solution near the lips with a Pasteur pipette to avoid tactile stimulation. In this procedure, the stimulus solutions were gradually diluted by diffusion, and the resultant responses also tended to decrease with time. Therefore we usually compared the initial responses for 5 s just after onset of stimulation. After the end of the recording, the lip region was completely washed with a sufficient amount of fresh ASW by aspirating the overflow and the next chemical stimulation was given 25-30 min later. The two kinds of extract solutions were applied in a random order and more than two trials each were usually performed. In these experiments, we could usually repeat chemical stimulation four to five times in the same preparation.

Electrophysiology

Recordings from individual neurons were performed using conventional electrophysiological techniques. For intracellular recording and stimulation, neurons were impaled with glass microelectrodes filled with 2 M potassium acetate. These electrodes were beveled to 5-7 M. For intracellular recording and dye injection for the calcium imaging, glass microelectrodes filled with 4% Calcium Green-1 hexapotassium salt (Molecular Probes) in distilled water (10-20 M) were used. Extracellular recordings of the nerve activity were made with flexible polyethylene suction electrodes (tip diameter, 100-150 µm) placed on the CBC. Similar suction electrodes were used for electrical stimulation of nerve trunks. The JC motor neuron was also voltage-clamped by a two-microelectrode method. Two electrodes were filled with 2 M potassium acetate (voltage recording electrode, 5-7 M; current-passing electrode, 2-4 M). Membrane currents were read as the voltage drop across a resister (1 M) interposed in the feedback loop (chassis ground mode). In these experiments, the reversal potential of the inhibitory postsynaptic currents (IPSCs) produced in the JC motor neurons maintained a stable level throughout the experiment.

The composition of ASW used in the experiments was as follows (in mM): 470 NaCl, 11 KCl, 11 CaCl₂, 25 MgCl₂, 25 MgSO₄, and 10 Tris-HCl (pH 7.8-7.9). In some experiments, polysynaptic pathways were suppressed using a 5 × Ca²⁺, 2 × Mg²⁺ solution or a 3 × Ca²⁺, 3 × Mg²⁺ solution, in which the Ca²⁺ and Mg²⁺ concentrations were raised five- and twofold, respectively, or threefold in both by replacement of Na⁺ to maintain the osmotic balance. In the 5 × Ca²⁺, 2 × Mg²⁺ solution, we obtained stable firing of the MA neuron with constant depolarizing current and large IPSCs in the JC motor neurons, but the firing of the CBM1 frequently induced the MA firing resulting from the large EPSPs (see RESULTS). Therefore this solution was incomplete for demonstrating direct effects of the CBM1 on the synaptic modulation. On the other hand, in the 3 × Ca²⁺, 3 × Mg²⁺ solution, the stable firing of the neuron with constant depolarizing current was hard to obtain, but the firing of the CBM1 could not induce the MA firing, suggesting that this solution could completely suppress the polysynaptic pathways. Therefore in the present experiments, we mainly used a 3 × Ca²⁺, 3 × Mg²⁺ solution to demonstrate the direct effect, and subsequently showed that a similar effect of the CBM1 could be obtained even in a 5 × Ca²⁺, 2 × Mg²⁺ solution.

Biocytin backfilling and paraformaldehyde-glutaraldehyde histofluorescence

The locations of the cell bodies of the neurons in the cerebral ganglia that send axons to the buccal ganglia were determined by backfilling the CBC with biocytin following the method of Xin et al. (1999). The cerebral ganglia were dissected from the animal and pinned to the Sylgard bottom of a chamber filled with ASW. The cut end of the CBC attached to the cerebral ganglia was sucked into the polyethylene tubing and the ASW in the tubing was replaced with 3% biocytin (Molecular Probes) in 50 mM NaHCO₃ (pH 8.0). The ganglia were maintained in ASW containing 1 mM probenecid at 4°C for 2 days. After a brief wash in ASW, the ganglia were fixed in 4% paraformaldehyde and 30% sucrose in 0.1 M phosphate-buffered saline (PBS, pH 7.6) at room temperature for 2 h. The ganglia were washed several times in 0.1 M PBS, each time for >30 min at 4°C. The preparation was permeabilized by overnight incubation in 0.1 M PBS containing 2% Triton X-100 at 4°C and then incubated in 50 µg/ml streptavidin rhodamine conjugate (Molecular Probes) in PBS Triton at 4°C for 12-24 h. After washing the ganglia in 0.1 M PBS for 2-3 days, the preparation was cleared in PBS/glycerol (1:6). The ganglia were viewed and photographed with an Olympus fluorescence microscope fitted with a rhodamine filter set (excitation filter BP520-550, dichroic mirror DM565, emission filter BA580IF). To observe catecholamine-containing neurons, the cerebral ganglia were fixed in 4% paraformaldehyde, 0.5% glutaraldehyde, and 30% sucrose in 0.1 M PBS [paraformaldehyde-glutaraldehyde (FaGlu) method] (Furness et al. 1977; Goldstein and Schwartz 1989) for 2-4 h at room temperature. The preparations were then viewed with the appropriate filter set (excitation filter BP400-440, dichroic mirror DM455, emission filter BA475) under which catecholamine fluorescence appears blue-green. In some preparations, the biocytin backfilling method and the FaGlu method were performed simultaneously. The FaGlu fixative solution was then used instead of the normal one, and the preparations were viewed with different filter sets for rhodamine and catecholamine.

Calcium imaging

When we explored the spike response of the newly identified CBM1 to taste stimulation of the lips with seaweed extract in a single preparation, the application was repeated four to five times after alternate washing with ASW for >2 h. The electrophysiological procedure is very difficult to use for this because a microelectrode cannot be kept in the cell body for a long time. However, we recently demonstrated that the calcium imaging method was very useful for detecting the spike activity of Aplysia neurons (Yoshida et al. 2001). Once a calcium-sensitive dye had been iontophoretically introduced into a cell body of the neuron with a microelectrode containing the dye, we could detect the spike activity of the neuron for a long time. The spike activity simply induced a fluorescence increase in the cell body, in which the slope of the fluorescence increase was almost relative to the spike frequency and the period of the increase almost corresponded to the firing period. We also demonstrated that the fluorescence intensity and the maximum change in the intensity after the spike activity tended to depend on the intracellular dye concentration. However, these differences did not affect the quantitative detection of the spike activity in each cell. Therefore in the present experiments, this method was used for the detection of the CBM1 spike activity to taste stimulation, and after the experiments were complete, the slope and the period of the fluorescence increase were analyzed in detail.

The calcium imaging was performed as described in our earlier report (Yoshida et al. 2001). A glass microelectrode filled with 4% Calcium Green-1 hexapotassium salt solution was impaled into a cell body of the neuron. After identification of the neuron with the electrophysiological procedure, a hyperpolarizing current (5-10 nA) was passed through the membrane for 2-5 min to introduce the dye into the cell body. The recording chamber was mounted on the stage of a fluorescence microscope (Olympus, BX50WI). A Spectra MASTER (Olympus) was used as the light source at 488 nm. The emission fluorescence passed through a dichroic reflector (DM510) and a BA515 barrier filter was detected. The fluorescence image was visualized with a ×10 objective (0.3 NA) and acquired by a cooled CCD camera (Olympix FE250). A sequence of images was stored on the MERLIN imaging system (Olympus), in which each image was usually binned in 3 × 3 blocks to enhance sensitivity, and images of 128 × 128 pixels were acquired at 20 frames/s for 120-150 s. The brightness and contrast levels were adjusted to achieve minimal background fluorescence and sufficient dynamic range for increases in fluorescence. In the data analyses, circular region of interests (ROIs) were defined at the region of a cell body of the CB neuron. A change in fluorescence was expressed as F/F (as a percentage), where F is the average fluorescence intensity at the resting state before stimulation and F is the time-dependent change in fluorescence. In plotting the change in fluorescence against the time course, the representative F/F every 0.5 s was obtained by averaging the F/F over 11 images, including five images before and after the 0.5 s image, to reduce data dispersion owing to apparatus noise. When the initial slope for 5 s was calculated, the representative F/F every 0.25 s was obtained by averaging the F/F over five images before and after the 0.25 s image, and each slope obtained from the increment of F/F for the neighboring representative images was averaged over 5 s. After the calcium imaging experiments, the ganglia were fixed with a FaGlu solution to confirm whether the measured neurons contained catecholamine.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Candidates for cerebral modulatory neurons

Our previous results (Nagahama et al. 1999) suggested that the synaptic transmission from the MA neurons to the JC motor neurons may be selectively suppressed during rejection of Gelidium and that the modulatory action of some cerebral neurons projecting axons to the buccal ganglia may directly contribute to the decrement of the synaptic efficacy. In the present experiments, we tried to identify such cerebral modulatory neurons. The locations of the cell bodies of the candidates were determined by backfilling the CBC with biocytin. The cell bodies of Aplysia cerebral neurons have been reported to be concentrated into several clusters (Jahan-Parwar and Fredman 1976; Ono and McCaman 1980; Rosen et al. 1979). In the present experiments, the cell bodies in the G, M, and E clusters were mostly stained (n = 25). We also obtained mirror images for the results of backfilling either the left or right CBC. A typical result is shown in Fig. 1A. In the ipsilateral G cluster located in the anterior portion of the ganglia, one large cell body (diameter >200 µm) immunoreactive to serotonin (unpublished data) was stained. This neuron may be equivalent to the metacerebral cells (MCCs) in A. californica (Weiss and Kupfermann 1976; Weiss et al. 1978) on the basis of the location of the cell body and the axonal projection to the CBC and the LLAB nerves. A small cell body of the neuron was also stained on the dorsal side within the same cluster. In the ipsilateral M cluster located at the separation between the ULAB and the AT nerves, four cell bodies were stained on the ventral side and one cell body on the dorsal side (Fig. 1D). In addition, one cell body was also stained on the ventral side in the contralateral M cluster (Fig. 1C). In the ipsilateral E cluster, several cell bodies were stained on both the ventral and dorsal sides. No cell bodies could be stained in the J and K clusters (n = 25) where the cell bodies of some mechanosensory neurons (ICBMs) sending axons to the CBCs are located in A. californica (Rosen et al. 1982). In the present experiments, we termed the backfilled cerebral neurons in the G, M, and E clusters CB neurons (cerebral-buccal neurons) except for the large serotonergic neuron. These neurons may be equivalent to the cerebral-to-buccal interneurons (CBIs) in A. californica (Hurwitz et al. 1999; Perrins and Weiss 1998; Rosen et al. 1991; Xin et al. 1999). Moreover, the CBIs in the M cluster have been reported to have command-like or modulatory functions (Hurwitz et al. 1999; Rosen et al. 1991, 2000a). Therefore in the present experiments, we addressed the CB neurons in the M cluster (CBM) as the candidates and explored whether either of these CBM neurons has modulatory functions for suppressing the synaptic transmission at the MA-JC connections.

View larger version (49K): [in this window] [in a new window]   Fig. 1. Views of the cerebral ganglia stained with backfilling of the cerebral-buccal connective with biocytin. A: a ventral view of the cerebral ganglia at low magnification where the neurons projecting axons to the buccal ganglia are stained. The left cerebral-buccal connective was backfilled with biocytin. B: a schematic drawing of the stained cell clusters () when the left cerebral-buccal connective () was backfilled (ventral view). C: a ventral view of the contralateral M cluster at high magnification. A single cell body of the neuron that sent axons to the bilateral cerebral-buccal connectives is stained. D: a ventral view of the ipsilateral M cluster where 4 cell bodies of the neurons are stained. ULAB n., upper labial nerve; AT n., anterior tentacular nerve; LLAB n., lower labial nerve; C-B conn., cerebral-buccal connective; C-P conn., cerebral-pedal connective; C-Pl conn., cerebral-pleural connective.

Catecholamine-containing neurons

A pair of catecholaminergic neurons (CBI-1) has been reported in the M cluster of A. californica (Rosen et al. 1991). In A. kurodai, we explored the location of the cell bodies of the catecholamine-containing CBM neurons by the use of the FaGlu method in combination with the biocytin backfilling (n = 24). Among four cell bodies of the ipsilateral CBM neurons on the ventral side (Fig. 2B), only one cell body in the most anterior portion exhibited catecholamine-like histofluorescence (blue-green fluorescence, Fig. 2D), and one cell body of the contralateral CBM also exhibited similar fluorescence (Fig. 2, A and C). Considering the mirror images for backfilling of the left or right CBC, these ipsilateral and contralateral neurons may belong to the same type of neurons, and they may send their axons to the buccal ganglia via the bilateral CBCs. These neurons may be equivalent to the CBI-1 in A. californica, and we termed these neurons CBM1s to avoid confusion between the animal species (Fig. 2, E and F). We termed the other CBM neurons CBM2a, CBM2b, and CBM3 (Fig. 2F), probably equivalent to CBI-2, CBI-12, and CBI-3 in A. californica, respectively. These CBM and the CB neurons in the other clusters did not exhibit catecholamine-like histofluorescence.

View larger version (59K): [in this window] [in a new window]   Fig. 2. Views of the cerebral ganglia to which the paraformaldehyde-glutaraldehyde (FaGlu) method was applied in combination with the biocytin backfilling of the cerebral-buccal connective. A and C: a cell body of the biocytin backfilled neuron (A) simultaneously exhibiting the FaGlu histofluorescence (C) in the contralateral M cluster. B and D: 4 cell bodies of the neurons stained by backfilling (B) and a cell body exhibiting FaGlu histofluorescence among them (D) in the ipsilateral M cluster. *, the same cells in each pair of upper and lower images. E and F: a schematic drawing of the cell bodies of the neurons stained in the contralateral (E) and ipsilateral (F) M clusters by backfilling the cerebral-buccal connective (). , the positions of the cell bodies of the catecholamine-containing neurons.

Identification of CBM1

Figure 3A shows examples of fluorescence images of the CBM1 filled with 5(6)-carboxyfluorescein which was injected iontophoretically into the cell body from a micropipette. The CBM1 sent bifurcated axons to the ipsi- and contralateral CBCs and had many dendritic trees extending toward the central region of the hemiganglion. The structure of the CBM1 was useful for us to identify by the electrophysiological procedure. Intracellular stimulation of the CBM1 by passing a depolarizing current through the membrane simultaneously produced extracellular nerve activity in the bilateral CBCs following the evoked neuronal spikes one for one (Fig. 4A). In addition, we found that firing of the CBM1 produced excitatory postsynaptic potentials (EPSPs) in the MA and JC neurons in the ipsilateral buccal ganglion. Even when the cerebral and buccal ganglia were bathed in a 3 × Ca²⁺, 3 × Mg²⁺ solution, these EPSPs followed the neuronal spikes one for one with constant latency [25.0 ± 0.2 (SE) ms, n = 8 in Fig. 4B], suggesting a monosynaptic connection with the ipsilateral MA and JC neurons. It was noted that the size of the EPSPs induced in the MA neurons was always larger (~3 times) than that in the JC neurons. Additional certification was often performed after the double-labeling procedure when only electrophysiological experiments were performed. In this procedure, we impaled the cell body of the CBM1 with a microelectrode filled with a rhodamine-lissamine dye. After the electrophysiological experiments, the dye was iontophoretically injected into the cell body and the tissue was subsequently fixed with a FaGlu mixture. As shown in Fig. 3, the positively identified CBM1 showed both rhodamine (red; B) and catecholamine-like (blue-green; C) fluorescence.

View larger version (85K): [in this window] [in a new window]   Fig. 3. Histological features of the catecholaminergic neuron in the bilateral cerebral M clusters (CBM1 neuron). A: a fluorescence image of the CBM1 neuron filled with 5(6)-carboxyfluorescein. B and C: a cell body of the CBM1 showing rhodamine fluorescence (B) and catecholamine-like fluorescence (C) with the appropriate filter sets. The sample was fixed with a FaGlu mixture after a rhodamine-lissamine dye was iontophoretically introduced into the cell body of the CBM1. *, the same cell body of the CBM1.

View larger version (18K): [in this window] [in a new window]   Fig. 4. Postsynaptic effects of the CBM1 neuron on the identified buccal neurons. A: excitatory postsynaptic potentials (EPSPs) produced in an ipsilateral multiaction neuron (MA1) by the spike activity of a CBM1. Firing of the CBM1 also elicited the extracellularly recorded nerve activity following the spikes one for one in the bilateral cerebral-buccal connectives (C-B conn.). B: EPSPs produced in an ipsilateral MA1 and an ipsilateral JC2 by current-induced spikes in a CBM1 when the buccal ganglia were bathed in a 3 × Ca2+, 3 × Mg2+ solution. - - -, correspondence between the CBM1 spikes and the EPSPs produced in the MA1 with constant latency.

Modulatory effects of CBM1 on MA-induced IPSCs

We first explored whether the CBM1 neuron had modulatory effects on the MA-induced IPSPs in the JC motor neurons. In these experiments, the cerebral and buccal ganglia were bathed in a 3 × Ca²⁺, 3 × Mg²⁺ solution to suppress polysynaptic activity. The JC motor neurons were voltage-clamped to remove the effects of membrane potential changes induced by CBM1 firing on the IPSP sizes. The holding potential was 40 mV because the inhibitory effects of the MA neurons on the JC motor neurons during the rhythmic feeding response usually appeared near the firing threshold of the JC motor neurons. A single spike was repetitively evoked in the MA neuron with a constant depolarizing current every 10 s, and the size of the inhibitory postsynaptic currents (IPSCs) produced in the JC motor neuron was compared before and after firing of the CBM1 neuron for 30 s. A typical result (n = 19) is shown in Fig. 5A. When the CBM1 neuron was fired, the MA neuron largely depolarized, and inward currents were also induced in the JC motor neurons, corresponding to the monosynaptic excitation of these neurons by the CBM1. Some examples of the faster recordings for each individual IPSC in the same experiment are shown in Fig. 6B, together with those for the control experiment in the absence of the CBM1 firing in the same preparation (Fig. 6A). The test immediately prior to the onset of the CBM1 firing was assigned as time 0 and the size of the sampled IPSC relative to that at time 0 was obtained every 10 s. During depolarization induced by the CBM1 firing, MA neuron often evoked more than two spikes (Fig. 5B) and in those cases the first IPSC was sampled. Figure 7 shows time courses of the relative IPSC sizes for two trials of the CBM1 firing (, 41.1 spikes/s; , 42.3 spikes/s on average) together with the control data in the same preparation. As previously reported (Nagahama et al. 1999), even in the control the relative IPSC size gradually decreased with time (). In contrast, after the CBM1 firing, the relative IPSC size rapidly decreased and attained the minimum level (<70%) within 40 s. In comparison with the control data, it was found that this suppressive effect was often prolonged to >100 s. Similar experiments were repeated in nine preparations when the CBM1 was fired at ~40 spikes/s for 30 s, and the minimum size of the IPSCs relative to the control value at that time and the minimum-attaining time after onset of the CBM1 firing were averaged over all preparations. In the CBM1 firing at 40.1 ± 1.1 (SE) spikes/s, the average minimum size of the IPSCs was 67.9 ± 2.6% and the average minimum-attaining time was 40.8 ± 3.7 s.

View larger version (14K): [in this window] [in a new window]   Fig. 5. Effects of the CBM1 firing on inhibitory postsynaptic currents (IPSCs) produced in the JC motor neuron by firing of the MA1. A JC2 was voltage-clamped at 40 mV, and a single IPSC was repetitively evoked by a current-induced spike of a MA1 every 10 s (A). The firing of a CBM1 induced depolarization of the MA1 and inward currents in the JC2, corresponding to the monosynaptic excitatory inputs from the CBM1. B, 1 and 2: the faster recordings of the IPSCs numbered 1 and 2 in A, respectively. During the depolarization of the MA1, the constant depolarizing current often evoked >2 spikes in the MA1 (B2). The preparation was bathed in a 3 × Ca2+, 3 × Mg2+ solution to suppress a polysynaptic pathway.

View larger version (15K): [in this window] [in a new window]   Fig. 6. Modulation of the MA-induced IPSC in the JC motor neuron by spike activity of the CBM1. A time series of the IPSCs in the absence (control, A) or presence (B) of the CBM1 firing in the same preparation. These data were obtained from the same preparation as shown in Fig. 5.

View larger version (27K): [in this window] [in a new window]   Fig. 7. Time courses of the relative sizes of the MA-induced IPSCs in the JC motor neuron before and after firing of the CBM1 (bar). The peak size of the IPSCs was measured every 10 s for 2 trials in the preparation as shown in Fig. 5.  and , the results obtained from the 1st and 2nd trials, respectively. The 2nd trial was performed 10 min after the 1st trial.

These results suggest that the CBM1 neuron may have a direct modulatory effect on the synaptic transmission at the MA-JC connections and may largely suppress the inhibitory effects of the MA neurons on the JC motor neurons.

Effects of the firing frequency of the CBM1 on the synaptic modulation

We further explored the effects of the firing frequency of the CBM1 neuron on the change in IPSC size by bathing the preparation in a 3 × Ca²⁺, 3 × Mg²⁺ solution. In seven preparations, the effects of the CBM1 firing at low frequency and high frequency for 30 s were compared. Then the strength of the depolarizing current was adjusted for 30 s to evoke the CBM1 firing at 15-25 and 35-45 spikes/s on average in each preparation. A typical result for the two firing frequencies is shown in Fig. 8A. The CBM1 firing at high frequency (44.3 spikes/s on average) induced a large decrease in the IPSC size, whereas the firing at low frequency (19.0 spikes/s) induced a slight decrease. Figure 8B shows the minimum IPSC size averaged over all preparations for the two types of frequencies. The average values were 94.0 ± 2.1% at 18.9 ± 1.3 spikes/s and 68.0 ± 2.6% at 40.1 ± 1.1 spikes/s, and the difference was significant (P < 0.001, paired t-test, n = 7). In contrast, the difference between the control and the minimum IPSC value for the low frequency was less significant (P < 0.05, paired t-test). The average value of the minimum-attaining time was 41.3 ± 2.3 s for the high frequency and 40.0 ± 3.6 s for the low frequency with no significant difference (P > 0.1, paired t-test). These results suggest that the synaptic modulation largely depends on the firing frequency of the CBM1 neuron with the constant minimum-attaining time.

View larger version (18K): [in this window] [in a new window]   Fig. 8. Dependence of synaptic modulation on the firing frequency of the CBM1. A: time courses of the relative sizes of the MA-induced IPSCs in the JC motor neuron before and after the CBM1 firing (bar) at low and high frequencies (19.0 spikes/s, ; 44.3 spikes/s, ) in the same preparation. , the control data in the absence of the CBM1 firing. B: comparison of the modulatory effects of the CBM1 averaged over all preparations (n = 7) for low (18.9 ± 1.3 spikes/s) and high (40.1 ± 1.1 spikes/s) firing frequencies. Bars show the SE. C: relationship between the average frequency of the CBM1 firing and the minimum IPSC size. These data were obtained from preparations bathed in a 3 × Ca2+, 3 × Mg2+ solution () and a 5 × Ca2+, 2 × Mg2+ solution ().

To explore the dependence of the synaptic modulation on the firing frequency of the CBM1 in more detail, the minimum IPSC sizes obtained from all preparations bathed in a 3 × Ca²⁺, 3 × Mg²⁺ solution were plotted against the raw data of the firing frequency in Fig. 8C (). It appears that the modulatory effect was very small at firing frequencies <25 spikes/s and largely increased above this frequency. This figure also shows data obtained from the preparations bathed in a 5 × Ca²⁺, 2 × Mg²⁺ solution (). As described in the METHODS, in this solution the CBM1 firing at high frequency frequently induced spike activity of the MA neuron, indicating the incomplete effect of this solution in blocking polysynaptic pathways. Figure 8C shows little difference between the data for the two kinds of solution, suggesting that direct modulatory effects were mainly obtained even in a 5 × Ca²⁺, 2 × Mg²⁺ solution.

Modulatory chemicals released from the CBM1 neuron

In the next step, the chemicals contributing to the synaptic modulation were explored. The results obtained from the FaGlu experiments suggested that the CBM1 contained catecholamines and may release these for the synaptic modulation. The most likely candidate is dopamine, by analogy with the CBI-1 in A. californica (Rosen et al. 1991), and we therefore explored the effects of this chemical on the IPSC size. Bath application of 5 × 10⁵ M dopamine largely reduced the size of the IPSC, mimicking the CBM1 effects, and washing out with the control solution almost recovered the size (Fig. 9A). The average values of the relative minimum size of IPSCs reduced by dopamine was 73.5 ± 1.5% (n = 10). It was found that the modulatory effects of dopamine were obviously induced at concentrations of >10⁵ M. 

View larger version (13K): [in this window] [in a new window]   Fig. 9. Effects of D1 antagonist SCH23390 on the dopamine-induced synaptic modulation. A: effect of dopamine application on the size of the MA-induced IPSCs in the JC2. In these experiments, the preparations were bathed in a 3 × Ca2+, 3 × Mg2+ solution to suppress a polysynaptic pathway. The external solution of the buccal ganglia was replaced with a solution containing 5 × 105 M dopamine and the effect was explored 10 min after application. B: blockage of the dopamine modulation by a D1 antagonist, SCH23390. The preparation was initially bathed in a solution containing 105 M SCH23390 for 20 min. The external solution was then replaced with a solution containing 105 M SCH23390 and 5 × 105 M dopamine and the dopamine effect was explored 10 min after application. C: comparison of the relative minimum sizes of the IPSCs induced by dopamine in the absence (DA, n = 10) or the presence (SCH23390 + DA, n = 7) of SCH23390 averaged over all preparations. Bars represent ±SE.

Several dopamine antagonists were also used to verify the contribution of dopamine. In A. californica, methylergonovine has been reported to affect dorpaminergic transmission (Ascher 1972; Nargeot et al. 1999; Teyke et al. 1993). However, we found that it did not have any effect on the dopamine-induced suppression of the IPSC even after methylergonovine in sufficient concentration (5 × 10⁵ M) was applied for 30 min (n = 4). Therefore we further explored specific antagonists for the D1 and D2 dopamine receptors, SCH23390 and sulpiride, previously used in other gastropods (Audesirk 1989; Magoski et al. 1995; Stoof et al. 1985). Pretreatment with a sufficient concentration of sulpiride (5 × 10⁵ M) for 30 min had no effect on the dopamine-induced suppression of the IPSC size (n = 4). In contrast, the dopamine-induced modulation was obviously reduced when the preparation was pretreated with the D1 antagonist SCH23390 (10⁵ M) for 20 min prior to dopamine application (Fig. 9B). The blockage was obviously induced at a concentration of 3 × 10⁶ M, but a solution at a higher concentration was used to induce sufficient effects in these experiments. Figure 9C shows a comparison of the average relative minimum sizes of the IPSCs reduced by dopamine application in the absence or the presence of SCH23390. The average value was 95.3 ± 1.7% for dopamine application with pretreatment of SCH23390 (n = 7), and there was a significant difference from the average value for the application of dopamine only described in the preceding text (P < 0.001, 2-sample t-test). In addition, the difference from the control was low (P < 0.05, 2-sample t-test). These results suggest that the D1 receptors may contribute to the suppression of the IPSC by dopamine.

Moreover, to investigate whether a similar mechanism participates in the synaptic modulation by the CBM1, we explored the effect of SCH23390 on the suppression of the IPSC by the CBM1. After the application of 10⁵ M SCH23390 for 20 min, the CBM1 effects were obviously, but not completely, blocked in the same preparation and a typical result is shown in Fig. 10, A-C (n = 4). The average values of the relative minimum IPSC size reduced by the CBM1 firing in the absence or the presence of SCH23390 were 68.1 ± 1.6 and 90.8 ± 1.0%, respectively (Fig. 10D), and the difference was significant (P < 0.005, paired t-test). In this case, there was also a significant difference between the sizes in the control and the pretreatment with SCH23390 (P < 0.005, paired t-test). These results suggest that the CBM1 may partly release dopamine that acts directly on the D1 receptors in the MA-JC synaptic sites and may suppress the transmission. On the other hand, similar application of 10⁵ M SCH23390 did not affect the sizes of the EPSPs monosynaptically produced in the MA neuron by the CBM1 firing (Fig. 11), suggesting that the CBM1 may use some chemical other than dopamine as a transmitter producing these EPSPs.

View larger version (28K): [in this window] [in a new window]   Fig. 10. Effects of D1 antagonist SCH23390 on the modulatory effects of the CBM1. The suppression of the synaptic transmission between the MA and JC neurons induced by the CBM1 firing for 30 s (A) is obviously reduced in the presence of the SCH23390 (B). C: time courses of the relative sizes of the IPSCs before and after firing of the CBM1 (bar) in the absence () and the presence () with SCH23390 in the same preparations as shown in A and B. The modulatory effect induced by the CBM1 firing for 30 s (39.7 spikes/s on average) was initially explored by bathing the preparation in a 3 × Ca2+, 3 × Mg2+ solution. The CBM1 was fired for 30 s (41.0 spikes/s on average) again after the preparation was bathed in a solution containing 105 M SCH23390 for 20 min. D: comparison of the average relative minimum sizes of the IPSCs induced by the CBM1 firing in the absence (CBM1) or the presence (SCH23390 + CBM1) of pretreatment with SCH23390 (n = 4). Bars represent ±SE.

View larger version (12K): [in this window] [in a new window]   Fig. 11. Effects of SCH23390 on the EPSPs monosynaptically produced in the MA1 by firing of the CBM1. These EPSPs are not affected by pretreatment with 105 M SCH23390 for 20 min. - - -, correspondence between the CBM1 spikes and the EPSPs produced in the MA1.

Responses of the CBM1 to taste stimulation of the lips with seaweed extracts

In the next step, we explored the response of the CBM1 neuron to taste stimulation of the lips with the seaweed extracts in the reduced preparations. In the present experiments, we found that it was very difficult to detect spike activity of the CBM1 for a long time (>2 h) during repetitive application (4-5 times) of the extract using the electrophysiological method, and we therefore used the calcium imaging method for this purpose (see METHODS).

After identification of the CBM1 neuron, a calcium-sensitive dye, Calcium Green-1, was iontophoretically introduced into the cell body of the CBM1, and the change in fluorescence in the cell body region was explored when the seaweed extract was applied to the lips. In the present experiments, we used the Ulva and Gelidium extracts (original extract solutions, see METHODS), which can induce the patterned jaw movements for ingestion or rejection when these solutions are applied to the lips as previously shown (Nagahama and Shin 1998; Nagahama et al. 1999). Figure 12 shows typical results for changes in the pseudocolor images of the CBM1 cell body after application of the Ulva (A) or Gelidium (B) extract after a control recording for 10 s in the same preparations (n = 13). The time courses of the changes in fluorescence (F/F, see METHODS) averaged over the whole region of the cell body are shown in Fig. 12C. Application of ASW scarcely induced a change in fluorescence. In contrast, just after stimulation with either of the two kinds of seaweed extract, the F/F began to increase and attained the peak values at ~10 s. The peak value for the Gelidium extract was larger than that for the Ulva extract. After recording the fluorescence images for 120-150 s, the lips were completely washed with a sufficient amount of fresh ASW, and the fluorescence intensity usually decreased to the control resting level within 5 min. In the present experiments, the two kinds of taste stimulation were presented in a random order at 25-30 min intervals and more than two trials each were usually performed. Figure 13 shows one example of the time courses of the F/F after alternate application of the Ulva (A) and Gelidium (B) extracts for two trials each. In all preparations, the fluorescence responses to taste stimulation with the same species of seaweed were very similar, while the responses for Gelidium were always larger in comparison with those for Ulva. Moreover the results indicated that the slope of the F/F increase was always larger in taste stimulation with the Gelidium extract compared with the Ulva extract, but the maximums of the F/F were attained at almost the same time. Considering the relationships between the spike activity and the fluorescence change (see METHODS), application of either kind of extract may induce spike activity of the CBM1 neuron with a similar firing duration, but the neuron may fire at a higher frequency after application of the Gelidium extract compared with the Ulva extract.

View larger version (44K): [in this window] [in a new window]   Fig. 12. Calcium imaging for the detection of the spike response of the CBM1 after application of seaweed extract to the lips. A and B: a series of pseudocolor images of the CBM1 cell body following application of the Ulva (A) or Gelidium (Geli.) (B) extract after 10 s control recordings in the same preparation. The number in each image shows the time before and after the application of the extracts and the color bar shows the fluorescence intensity in pseudocolor. C: time courses of the change in fluorescence of the CBM1 cell body after application of ASW and extracts of the 2 species of seaweeds. Seaweed extracts were applied to the animal lips at an arrowhead.

View larger version (23K): [in this window] [in a new window]   Fig. 13. Time courses of the changes in fluorescence of the CBM1 cell body after the application of the Ulva (A) or Gelidium (Geli.) (B) extract for 2 trials each in the same preparation. In this case, the Ulva and Gelidium extracts were alternately applied to the lips at arrowheads after washing with ASW, at intervals of 25-30 min.

Dependence of the fluorescence response on the concentration of the extract solution

In the present study, we initially expected that the cerebral modulatory neurons contributing to rejection would show spike activity only following application of the Gelidium extract. However, the results obtained indicated that the application of the Ulva extract also induced the spike activity of the CBM1 neuron. It was therefore necessary to consider the effects of the concentration of the extract solutions on the CBM1 responses. In our previous study, the concentration of the original extract solutions was determined so as to induce stable rhythmic responses for ingestion or rejection in semi-intact preparations, and dilution of the original extract solutions often failed to induce the rhythmic responses (Nagahama and Shin 1998). However, we have not explored the concentration dependence of the response in detail. Therefore we explored how the fluorescence response depended on the concentration of the seaweed extract solutions in the same preparation (n = 7 for Ulva, n = 5 for Gelidium). Examples of the responses to taste stimulation with the Ulva and Gelidium extract solutions at various concentrations are shown in Fig. 14, A and B. Application of the extract solutions diluted with ASW induced transient fluorescence responses depending on the concentration. We compared the initial slopes for 5 s just after the onset of stimulation, relating to the firing frequency (see METHODS). Figure 14, C and D, shows the average initial slopes for all preparations plotted against the relative concentrations to those of the original Ulva and Gelidium extract solutions. The initial slopes were normalized against the slope for the original extract solution in each preparation. In both kinds of taste stimulation, the relative initial slopes increased at more than 0.02 times the original concentrations and attained the maximum at half the original concentrations. In four of seven preparations, the fluorescence responses following the application of the Ulva extract solutions at twofold the original concentration were also explored. For the Ulva extract solutions at half and twofold the original concentration, the average values of the relative initial slope were 0.97 ± 0.03 and 1.05 ± 0.04, respectively, and there were no significant differences between the initial slopes of the original solution and the solutions with either of the other concentrations (P > 0.1 for half concentration, n = 7; P > 0.1 for twofold concentration, n = 4, paired t-test). In addition, for the Gelidium extract solution at half the original concentration the average value of the relative initial slope was 0.97 ± 0.03, and there was no significant difference from that for the original concentration (P > 0.1, n = 5). These results suggest that the original extract solutions of Ulva and Gelidium may be sufficient to induce the maximum spike response of the CBM1 neuron and the use of more concentrated extract solutions will not increase the responses.

View larger version (39K): [in this window] [in a new window]   Fig. 14. Dependence of the fluorescence response in the CBM1 cell body to taste stimulation on the concentration of the seaweed extracts. A: responses in the CBM1 cell body after application of the Ulva extract at different concentrations (0.01, 0.1, 0.2, 1 times original extract solution) and ASW. B: responses in the CBM1 cell body after application of the Gelidium (Geli.) extract at different concentrations (0.05, 0.2, 0.5, 1 times original extract solution) and ASW. The extracts were applied to the lips at an arrowhead. C and D: dose-response curves for the average relative initial slopes of the fluorescence changes against the relative concentration of the Ulva (C, n = 7 for 0.01-1 times, n = 4 for 2 times) or Gelidium (D, n = 5) extracts. Bars represent ±SE.

In the other cerebral neurons we further compared the fluorescence responses to taste stimulation with the original extract solutions of Ulva and Gelidium. Several neurons in the M, G, and C clusters responded to the taste stimulation (the response of the G and C clusters was also preliminarily reported in Yoshida and Nagahama 2001). The responses were almost the same for the extract solutions of the two species of seaweed, although a very few neurons showed larger responses following the application of the Ulva extract compared with the Gelidium extract. However, of all the cerebral neurons studied in the present experiments, only the CBM1 neuron obviously showed larger responses to taste stimulation with the Gelidium extract compared with the Ulva extract. Figure 15 shows an example of the fluorescence responses of a CBM2a/b (either of CBM2a or -2b) neuron after the application of the Ulva and Gelidium extracts to the lips (n = 9). Taste stimulation with the two kinds of extracts induced similar long-lasting rhythmic changes in fluorescence in the same neuron. The average initial slopes for 5 s were 0.70 ± 0.13%/s for Ulva and 0.69 ± 0.12%/s for Gelidium, and there was no statistical difference (P > 0.1, paired t-test). There was also no significant difference in the successive rhythmic responses for the two kinds of extracts in all preparations. This neuron may be equivalent to the feeding command-like neurons, CBI-2/12, reported in A. californica on the basis of the location of the cell body and the functional role that the firing of the neuron induced the rhythmic response in many buccal neurons (Rosen et al. 1991; unpublished data). These results suggest that there is little difference of the spike responses to taste stimulation with the two kinds of extracts in many cerebral neurons including the feeding command-like neurons. Therefore the application of the two kinds of original extract solutions may produce similar basic rhythmic responses of the buccal neurons, probably caused by the similar response of the command-like neurons, and the application of the diluted extracted solution may induce a weaker response. This may explain our finding that the application of the diluted solution of the original extract solutions often fails to induce the rhythmic responses.

View larger version (23K): [in this window] [in a new window]   Fig. 15. Fluorescence responses in the CBM2a/b cell body after application of the Ulva (A) or Gelidium (Geli.) (B) extract for 2 trials each in the same preparation. The extracts were applied to the lips at an arrowhead.

These results support the validity of the application of the original extract solutions of Ulva and Gelidium for comparison of the CBM1 responses, and we used these solutions in the following experiments.

Averaged fluorescence responses of CBM1 to taste stimulation

The fluorescence responses of the CBM1 to taste stimulation with the original extract solutions of Ulva and Gelidium were explored in 13 preparations. In the present experiments, application of the two kinds of extract solutions produced a similar time course of the change in fluorescence, and we compared only the initial slopes. Figure 16 shows the comparison of the average initial slopes of the responses to taste stimulation with the Ulva and the Gelidium extracts for all preparations. The values were 0.78 ± 0.13%/s for the Ulva extract and 1.42 ± 0.21%/s for the Gelidium extract, and the difference between them was significant (P < 0.001, paired t-test). The average initial slope for ASW application (0.10 ± 0.02%/s, n = 12) is also shown in Fig. 16, and the differences from that for the Ulva extract and the Gelidium extract were significant (P < 0.001 for Ulva, P < 0.001 for Gelidium, 2 sample t-test). These results suggest that the difference between the firing frequencies of the CBM1 spike responses to taste stimulation with the Ulva and Gelidium extracts is clear.

View larger version (17K): [in this window] [in a new window]   Fig. 16. Average initial slopes of the change in fluorescence induced by application of ASW, the Ulva extract or the Gelidium (Geli.) extract. The fluorescence responses to taste stimulation of the 2 species of seaweed were explored in the same preparations and averaged over all preparations (n = 13). In 12 preparations, the responses to ASW application were also averaged. Bars represent ±SE.

Firing frequency of CBM1 response induced by taste stimulation with seaweed extracts

From the calcium imaging alone, we could not show the real difference in the firing frequencies of the CBM1 spike activity induced by taste stimulation with the two species of seaweed. To estimate the real firing frequency of the CBM1 spike response to taste stimulation, the CBM1 was fired by repetitive electrical stimulation of the CBC, and the resultant changes in fluorescence were compared with the taste-induced responses in the same neuron. In these experiments, the strength of each short electrical pulse was adjusted to evoke a single antidromic spike in the CBM1 neuron. Figure 17 shows an example of the change in fluorescence after application of the Ulva and Gelidium extract solutions (A) and electrical stimulation of the CBC (B) in the same CBM1 neuron (n = 8). Electrical stimulation of the ipsilateral CBC for 5 s at various frequencies induced increases in fluorescence depending on the stimulus frequency. The relationship between the stimulus frequency and the initial slope of the change in fluorescence for 5 s is shown in Fig. 17C. The initial slopes were almost proportional to the stimulus frequency, i.e., the firing frequencies of the CBM1. Therefore we could estimate the firing frequency of the CBM1 spikes after taste stimulation by the use of this relationship for the initial slope of the response to taste stimulation in the same preparation, and the obtained frequencies were averaged over all explored preparations. Figure 17D shows a comparison of the average spike frequencies for the two kinds of taste stimulation. The values were 19.7 ± 2.4 spikes/s for the taste of Ulva and 31.8 ± 3.4 spikes/s for the taste of Gelidium and the difference was significant (P < 0.001, n = 8, paired t-test).

View larger version (29K): [in this window] [in a new window]   Fig. 17. Estimation of the firing frequency of the CBM1 spike activity after taste stimulation. A: fluorescence responses in a CBM1 cell body following the application of the Ulva (Ulva) or the Gelidium (Geli.) extracts to the lips. B: fluorescence responses in the same cell after repetitive electrical stimulation of the cerebral-buccal connective for 5 s with short square pulses (5 ms, 0.8-1.0 V) of different frequencies (10, 20, and 30 Hz). C: initial slopes of the fluorescence changes plotted against the stimulus frequency or the firing frequency. Arrows show the initial slopes obtained from the taste responses shown in A (within the broken rectangle) and the firing frequencies were obtained from this relationship (11.8 spikes/s for Ulva and 28.2 spikes/s for Gelidium). D: comparison of the average spike frequencies in the CBM1 firing induced by the application of the Ulva extract or the Gelidium extract (n = 8). Bars represent ±SE.

Moreover, we explored the CBM1 spike activity induced by the application of the seaweed extracts electrophysiologically, although a pair of responses for the two kinds of extracts could not be compared in the same preparation. Examples of the spike responses to taste stimulation with the two kinds of extract solutions in separate preparations are shown in Fig. 18, A and B. In these experiments, there was an obvious difference between the firing frequencies of the spike activity just after extract application. The spike frequency averaged for 5 s just after the onset of stimulation was calculated and the values for all preparations were averaged. Figure 18C shows a comparison of the average spike frequencies of the CBM1 responses induced by taste stimulation with Ulva and Gelidium. The values were 20.2 ± 1.5 spikes/s for the taste of Ulva (n = 6) and 30.4 ± 1.1 spikes/s for the taste of Gelidium (n = 6) and the difference was significant (P < 0.001, 2-sample t-test). The average spike frequency for ASW application (3.9 ± 0.5 spikes/s, n = 6) is also shown in Fig. 18C, although the value was positive because the CBM1 tended to fire spontaneously even before stimulation. The differences of the spike frequencies between ASW and the Ulva extract application and between ASW and the Gelidium extract application were significant (P < 0.001 for Ulva, P < 0.001 for Gelidium, 2-sample t-test).

View larger version (22K): [in this window] [in a new window]   Fig. 18.