INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Well-coordinated interactions between sensory and motor neuronal elements are essential for the orderly production of behavioral outputs in all animal species. The identification of both motor and sensory components and characterization of their synaptic and functional interactions is critical for understanding the cellular basis of animal behavior. Individual sensory neurons mediating various behaviors and their synaptic connections have been identified and described in a variety of invertebrate species, such as crustaceans (Calabrese 1976a,b; Fricke et al. 1982; Wiese 1976; Wine 1977), insects (Basarsky and French 1991; Callec et al. 1971; Davis and Murphey 1993; French et al. 1993; Wolf and Burrows 1995), annelids (Kristan 1982; Nicholls and Baylor 1986), and mollusks (Audesirk 1979; Byrne et al. 1974; Clatworthy et al. 1994; Inoue et al. 1996; Janse 1974; Mellon 1972; Olivo 1970; Rosen et al. 1982; Spray et al. 1980a,b; Walters et al. 1983). Different behavioral and neural aspects of the withdrawal reactions have been extensively studied in pulmonate snails (Arshavsky et al. 1994a; Sakharov and Rozsa 1989). Whole-body withdrawal motoneurons activated by multimodal sensory inputs have been identified in freshwater snail Lymnaea (Ferguson and Benjamin 1991a,b) and Planorbis (Arshavsky et al. 1994b,c).

Terrestrial snail Helix lucorum is a useful model for studying the cellular basis of different behaviors. Neural mechanisms of avoidance behavior had been intensively studied during last two decades, and the efferent part of the withdrawal neural circuit in Helix has been described in details (Balaban 1979, 1983; Balaban et al. 1987; Zakharov and Balaban 1987). However, almost nothing is known about the sensory neuronal elements recruited in avoidance reactions and sensory inputs to the withdrawal interneurons involved in triggering the withdrawal responses (Balaban 1979). Several putative mechanosensory cells presynaptic to withdrawal interneurons were described earlier; however, their receptive fields were not well defined and assigned to liver and other internal organs located under the shell (Arakelov et al. 1989). In this study, we have identified for the first time a group of mechanosensory neurons with receptive fields on the foot of the snail H. lucorum that monosynaptically activate interneurons for the withdrawal behavior. These neurons, termed pleural ventrolateral neurons (PlVL), presumably mediated part of the withdrawal response of the snail via activation of the withdrawal interneurons.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals

The experiments were performed on the mature specimens of H. lucorum L. weighting 30-35 g. Snails were originally collected in Crimea and maintained in an active state in the laboratory.

Behavioral experiments

Experiments on freely behaving snail were performed as described earlier (Balaban and Bravarenko 1993). Video records of experiments from a standard video tape recorder were digitized on the IBM-PC-compatible computer and then analyzed using video analysis software PhysVis (Kenyon College). The anterior tip of upper tentacles (or the anterior tip of the head, in cases when tentacles were completely retracted) was monitored, and its movements were plotted as a function of time.

Isolated CNS preparation

Animals were anesthetized by injection of isotonic MgCl₂ (~15% of the animal weight). The central ganglionic ring was removed from the animal and pinned to a silicone-elastomer (Sylgard)-coated dish. Connective tissue sheath was partially removed using fine forceps and scissors. To facilitate further desheathing, ganglia were treated with Protease (0.25 mg/ml; Type XIV, Sigma) for 30 min at room temperature and washed out, and then the fine sheath was completely removed. To expose ventrolateral surface of pleural ganglion, we usually transected the pedal commissure and rotated each pedal ganglion 180° outward. The CNS was bathed in a saline solution that contained (in mM) 100 NaCl, 4 KCl, 7 CaCl₂, 5 MgCl₂, and 10 Tris-HCl buffer (pH 7.8). Electrophysiological recordings were performed 60 min after the dissection.

To examine the monosynaptic connections, we used high-Mg²⁺/high-Ca²⁺ solution containing 28 mM Ca²⁺ and 20 mM Mg²⁺ to raise firing threshold and reduce the chances of activating interneurons (Berry and Pentreath 1976).

Reduced "CNS-foot" preparation

The reduced "CNS-foot" preparation consisted of the CNS connected with a foot dissected sagittaly in two halves positioned with the skin of the dorsolateral surface upwards. Half of the foot was placed in a separate compartment of the experimental chamber to prevent skin mucus to affect neuronal activity. All pedal cutaneous nerves remained intact. Ganglia were desheathed as it is described for an isolated CNS preparation. Tactile stimuli were delivered by means of electrical-mechanical tapper with fixed pressure intensity 1 g (tip diameter 0.05 mm). Central and peripheral synaptic connections were blocked by perfusing preparation with zero-Ca²⁺/high-Mg²⁺ saline (0 mM Ca²⁺/25 mM Mg²⁺).

Electrophysiology and dye injection

Intracellular recordings from either an isolated brain or reduced CNS-foot preparation were made using standard electrophysiological techniques. Identified withdrawal premotor interneurons of the parietal ganglia (Pa2 and Pa3) and pleural ganglia (Pl1) were penetrated with glass microelectrodes filled with 2 M potassium acetate (tip resistance, 10-15 M). For recording from PlVL neurons, electrodes with a fine tip (25-30 M) were used. Intracellular signals were recorded with preamplifiers (Neuroprobe 1600, A-M Systems), digitized, and stored on computer (Digidata 1200A A/D converter and Axoscope 8.0 software, both from Axon Instruments). All EPSP recordings were filtered with a 10-Hz low-pass filter built into the Axoscope program.

For morphological investigation of recorded neurons, a 5% neurobiotin solution in 2 M potassium acetate was iontophoretically injected via the recording electrodes with 1-nA positive current pulses for 30-50 min and after fixation of tissue was visualized with Texas-Red-labeled avidin (Vector Laboratories, Burlingame, CA). The preparations were then examined with an Axioscope fluorescence microscope and Bio-Rad MRC 600 laser scanning confocal microscope.

Double-labeling experiments

For double-labeling experiments, pleural ventrolateral neurons were injected with neurobiotin as described in the preceding text. Pa3 and Pl1 neurons in the same preparation were injected by using the air pressure system (PV830 Pneumatic PicoPump, WPI) with 3% Lucifer yellow in 0.1 M KCl in an approximate volume 10 pl. The preparations were then immunocytochemically processed with the rabbit anti-Lucifer yellow polyclonal antiserum (Chemicon International). Preparations were scanned with Bio-Rad MRC 600 laser confocal microscope with two different filters for Texas Red and FITC.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

During the pilot search of neurons monosynaptically connected with premotor withdrawal interneurons, several presynaptic cells were discovered in the pleural ganglia on the ventrolateral surface. In each experiment, after identifying the cell as presynaptic, we investigated its morphology, electrophysiological characteristics, and properties of the connections with several withdrawal interneurons.

Morphological characteristics of the pleural ventrolateral neurons

Between 15 and 30 PlVL monosynaptically connected to withdrawal interneurons were located on the ventrolateral surface of each pleural ganglion between pleuropedal and pleurocerebral connectives (Fig. 1). Most of PlVL cells had a soma diameter of ~20-30 µm. One cell from this group was bigger then others (soma diameter, ~40 µm) and could be visually identified; this cell has been designated as Pl4 neuron. Intracellular injection of neurobiotin revealed that all stained PlVL neurons (n = 12) sent their axons through the pleuropedal connective and pedal ganglion into the skin nerve (nervus cutaneus secundus; Fig. 2A). In addition, the Pl4 neuron had a second axon branch that projected into the cerebral ganglion through the pleurocerebral connective (n = 5, Fig. 2B). In two cases, axons from the pleural sensory neurons were traced to the skin.

View larger version (48K): [in this window] [in a new window]   Fig. 1. Schematic diagram of the ventral surface of the left pleural ganglion. Pleural ventrolateral (PlVL) cells () are situated on the ventrolateral surface of each pleural ganglion. Morphologically identifiable pleural withdrawal interneuron (Pl1) is shown.

View larger version (106K): [in this window] [in a new window]   Fig. 2. Morphology of PlVL neurons. A: all stained PlVL cells projected toward periphery via pleuropedal connective and then into the nervus cutaneus pedalis secundus. B: morphology of the Pl4 neuron. The Pl4 cell in addition had central projection into the pleurocerebral commissure.

Electrophysiological characteristics of PlVL neurons

The results described in the following text were based on examination of 263 PlVL cells in 52 preparations. The PlVL neurons were electrically silent, showing no spontaneous activity or fast synaptic potentials both in isolated brain and reduced CNS-foot preparations. The average membrane potential of the PlVL cells was 57.5 ± 3.1 mV. A gentle mechanical touch to the skin covering dorsal surface of the foot elicited spikes in the PlVL neurons without any noticeable prepotential activity (Fig. 3A). It is important to stress that action potentials appeared only at the beginning and the end of the mechanical stimulation. Mechanical stimulus was applied in wide range of duration (0.02-10 s). Such stimulation also produced a depolarizing response in ipsilateral pleural and all parietal interneurons for withdrawal (Pa2 and Pa3). Application of brief mechanical stimuli (20 ms) with high frequency (10 Hz) produced spike responses in the PlVL neurons to each stimuli showing no habituation or depression (Fig. 3B). Mechanical stimulus-induced action potentials in these cells persisted in high-Mg²⁺/zero-Ca²⁺ saline (Fig. 4). At the same time, depolarizing responses in the pleural withdrawal interneurons were completely abolished. These data suggest that PlVL neurons were primary mechanoreceptors. However, the possibility still exists that these cells receive excitatory inputs via electrical synapses or highly encapsulated chemical synapses in the periphery.

View larger version (24K): [in this window] [in a new window]   Fig. 3. Responses of PlVL cells to mechanical stimulation of the skin. A: mechanical stimulus applied to the skin covering lateral surface of snail body elicited in the Pl4 cell an action potential. Note that action potentials appeared only at the onset and the end of mechanical stimulus (ON and OFF responses). Increase of duration did not change the ON-OFF character of the PlVL neurons response. Mechanical stimulation also produced depolarizing synaptic responses in pleural withdrawal interneuron (Pl1) coincident in time with spikes in Pl4 cell. B: mechanical stimulation with brief (20 ms) mechanical stimuli () applied with frequency of 10 Hz produced spikes in Pl4 cell approximately with the stimulation rate. Pleural interneuron showed fast adapting response in such conditions.

View larger version (10K): [in this window] [in a new window]   Fig. 4. Effect of blockade of synaptic transmission in preparation. A: responses of PlVL and Pl1 neurons to the mechanical stimulus while preparation was bathed in normal Ringer solution. B: superfusion and maintenance of this preparation (60 min) in high-Mg2+/0-Ca2+ saline blocked response in withdrawal interneuron but failed to block spike response in PlVL neuron to the same mechanical stimulus.

To identify the individual receptive fields, PlVL neurons were impaled one after another in the same CNS-foot preparation, and receptive fields of each neuron were delineated by applying mechanical stimuli to sites over entire body surface. We found that each neuron had an oval-like shaped receptive field located on the body surface (Fig. 5). The receptive field of an identifiable Pl4 neuron was situated behind the rear tentacles and was ~20 mm in length, whereas receptive fields of other PlVL neurons were smaller (~5-12 mm). Partial overlapping of receptive fields of different neurons was often observed. We did not find any electrical or chemical interactions between different PlVL cell. Mechanical stimulation of the skin on the outer edge of a receptive field did not produce a hyperpolarization of neurons (data are not shown), as it was described in some other mollusks (Getting 1976; Mellon 1972; Spray et al. 1980; Walters et al. 1980).

View larger version (19K): [in this window] [in a new window]   Fig. 5. Receptive fields of PlVL neurons. Results are from a single animal. · · · , receptive field borders. Skin stripes pattern is shown for orientation.

Stimulation of the skin inside the receptive fields of PlVL neurons with various chemical agents, such as concentrated NaCl saline, 0.5 N solution of HCl and NaOH, failed to elicit spikes in these cells (n = 5 in 5 preparations) even though such a stimulation produced vigorous response in interneurons of avoidance behavior (Fig. 6A). This observation suggested that PlVL cells were specific mechanoafferent neurons, while other sensory neurons transferred inputs of chemical modality.

View larger version (22K): [in this window] [in a new window]   Fig. 6. Comparison of responses to mechanosensory and chemical stimulation of the skin. A: sensory selectivity of PlVL cells. Tactile stimulation of skin induced subthreshold depolarization in pleural withdrawal interneuron (Pl1) and spike response in PlVL cell in the CNS-foot preparation. Chemical stimulation of the same area did not produce action potentials in PlVL neuron but elicited a strong spike response in Pl1 cell. B: chemical stimulation produced stronger withdrawal response in freely behaving snails than mechanosensory stimuli. Mechanograms of withdrawal responces induced by mechanical stimulus (black arrows, dashed line) and by chemical stimulus (white arrow, solid line) are plotted together. Presented snail contour is illustrating the withdrawal response ratio.

To compare behavioral effects of mechanical and chemical stimulation, we performed a series of experiments in intact freely moving snails (n = 6). We found that short (300 ms) mechanical stimuli applied to the receptive field of Pl4 neuron induced transient, partial retraction of the upper tentacles (70% of their initial length), whereas chemical stimulation of the same region with a drop of 0.5 N NaOH induced complete withdrawal of tentacles immediately followed by retraction of body into the shell lasting tens of minutes (Fig. 6B). Thus chemical stimulation of skin produced much more stronger withdrawal response in the freely behaving snail than mechanosensory stimuli. This observation is in full accordance with electrophysiological data described in the preceding text.

PlVL neurons monosynaptically activate interneurons of avoidance behavior

To test the hypothesis that the PlVL neurons can activate the withdrawal interneurons, we performed a series of experiments with simultaneous intracellular recordings from PlVL and withdrawal interneurons. Each induced spike in any PlVL neuron was found to produce EPSPs in relation 1:1 in the ipsilateral Pl1, Pa2, and Pa3 and contralateral Pa2 and Pa3 neurons (Fig. 7, A and B). The average amplitude of EPSPs in pleural interneurons elicited by first spike in PlVL cells was 5.8 ± 2.1 mV (n = 12), whereas amplitudes of EPSPs in parietal interneurons were in general smaller (3.9 ± 1.7 mV, n = 10). A short burst of spikes induced in PlVL was found to elicit spike response in Pl1 neuron but only subthreshold EPSPs in all parietal withdrawal interneurons (Fig. 7C). PlVL neuron-induced EPSPs in the withdrawal interneurons persisted in high-Mg²⁺/high-Ca²⁺ saline (Fig. 7D). A high stability of spike-EPSP latency represented an additional confirmation of monosynaptic nature of the recorded connections. Figure 7E shows six superimposed traces of spikes in PlVL and responses in Pl1 neurons with constant latency 8 ms (Fig. 7E). These data suggest that the connections between PLVl cells and withdrawal interneurons appear to be monosynaptic.

View larger version (19K): [in this window] [in a new window]   Fig. 7. Synaptic connections between PlVL cells and parietal and pleural interneurons for withdrawal behavior. A: intracellulary induced spikes in LPlVL cell produced EPSPs in ipsilateral pleural and parietal withdrawal interneurons. Note that EPSP amplitude in Pl1 neuron is almost 2 times bigger than in Pa3 cell. B: Pl4 neuron excited both ipsi- and contrlateral Pa3 neurons. C: short burst of spikes induced in PlVL cell was able to produce spike response in pleural withdrawal interneuron but only subthreshold response in parietal interneuron. D: synaptic connection between PlVL cell and pleural interneuron persisted in high-Mg2+/high-Ca2+ solution. E: superposition of 6 traces of spikes in PlVL and responses in Pl1 neurons demonstrated high stability of latency of this synaptic connection.

Synaptic connections between the PlVL neurons and withdrawal interneurons showed a progressive depression with a repeated stimulation. Such depression has been observed reliably over a wide range of interstimulus intervals: from 1 to 300 s. Figure 8 summarizes the results of experiments in which individual PlVL neurons (7 cells in 7 preparations) were repeatedly activated with brief intracellular pulses at 60-s intervals. Averaged results of behavioral experiments in intact snails are also presented in Fig. 8 (n = 11). Tactile stimulus was delivered to the head region, 4-5 mm behind the ommatophores (approximately to the receptive field of Pl4 cell; Fig. 5). The initial increase in the behavioral response amplitude was apparently related to the sensitization process (Balaban 1983; Zakharov and Balaban 1987). The following dynamics of habituation of the behavioral responses and depression of EPSPs amplitude in the withdrawal interneurons were very similar.

View larger version (18K): [in this window] [in a new window]   Fig. 8. Dynamics of behavioral responses to tactile stimulation in intact snails (n = 11; ) and depression of monosynaptic connections between PlVL and withdrawal interneurons () with repeated 1/min stimulation. In behavioral experiments, mean amplitude of ommatophore withdrawal in percents of initial are shown. In neurophysiological experiments, mean responses ± SE to intracellular stimuli (each activating a single spike) applied at regular 60-s intervals are plotted (n = 7 cells in 7 preparations). EPSPs are normalized to the 1st EPSP elicited.

Location of synapses between mechanoafferent neurons and withdrawal interneurons

To determine the location of synapses between PlVL and withdrawal interneurons, we performed double-labeling experiments. PlVL and withdrawal interneurons were stained with different fluorescent dyes in the same preparation (see METHODS). We found that processes of PlVL cells and both Pa3 and Pl1 neurons were present in the same area only in the neuropile of the pleural ganglion (Fig. 9). Such synapse location can partially explain the difference in EPSP amplitude in parietal and pleural withdrawal interneurons: distance from possible synapse zone to soma is minimal for the Pl1 cell in which we recorded largest EPSPs.

View larger version (126K): [in this window] [in a new window]   Fig. 9. Possible site of a synaptic contact between PlVL and the withdrawal interneuron. The PlVL neuron was stained with neurobiotin and visualized with Texas Red (appears white). Ipsilateral withdrawal interneuron (LPa3) was injected with Lucifer yellow, then processed with anti-Lucifer antibodies, and then developed with FITC-conjugated secondary antibodies (appears black). Confocal microscopy study revealed that neuropile of the pleural ganglion is the only place in the CNS where processes of both cells are located together. Processes of the LPa3 neuron are marked by an arrow. Only few optical slices are presented (total thickness is ~40 µm).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Comparison of PlVL cells to other mechanoafferent neurons

In the present study, we identified mechanosensory neurons that produced monosynaptic excitatory inputs to the interneurons for the whole-body withdrawal behavior. Based on their partial morphological, electrophysiological, and functional similarities with other mechanosensory neurons described in various molluscan species, we believe that PlVL neurons are primary mechanosensory cells. In particular, PlVL neurons exhibited morphological and electrophysiological features that were somewhat similar to the mechanosensory neurons previously characterized in Aplysia (Byrne 1980a,b; Byrne et al. 1978; Walters et al. 1983), Tritonia (Getting 1976), and Lymnaea (Inoue et al. 1996). Like Aplysia VC mechanosensory neurons, Tritonia S cells, and Lymnaea RPD3, the PlVL neurons have its somata located in the central ring ganglia and have both peripheral and central (in case of Pl4 cell) projections. When mechanical stimulation was applied to the skin inside the receptive field of each PlVL neuron, neurons fired action potentials in the absence of observable prepotentials (Fig. 4). These responses persisted in the presence of various solutions known to block indirect polysynaptic transmission. Similar to other mechanoafferent and sensory interneurons in invertebrates (Byrne 1982; Byrne et al. 1974; Callec et al. 1971; Walters et al. 1983; Zucker 1972), synaptic connections between the PlVL neurons and withdrawal interneurons showed a progressive depression with repeated stimulation. The PlVL neurons and sensory neurons in other mollusks also have such common characteristics as the lack of impulse activity without mechanical stimulation and relatively low thresholds for their activation. However, in contrast to the mechanoafferent cells found in Aplysia (Walters et al. 1983) and Lymnaea (Inoue et al. 1996), which demonstrated slowly adapting responses to the maintained mechanical stimulation, the PlVL neurons showed single spike ON-OFF response in such conditions. It makes them more similar to the T cells of the leech (Nicholls and Baylor 1968), Tritonia S cells (Getting 1976), and mechanoafferent cells in razor clam (Olivo 1970). But in contrast to Tritonia S cells, the PlVL neurons were not activated by chemical stimuli. Unlike other mechanosensory cells (Getting 1976; Spray et al. 1980), the PlVL neurons lack a hyperpolarizing responses to stimuli applied outside of the excitatory receptive field. It is interesting to note that other putative sensory neurons (Pa7, 9, 11) described in Helix (Logunov and Balaban 1978) that innervate the liver and organs located inside the shell exhibited hyperpolarizing responses to mechanical stimulation of the skin and the mantle, whereas short mechanical stimulation of visceral organs elicited bursts of spikes lacking prepotentials in those cells (Arakelov et al. 1989). Unfortunately, authors did not apply maintained mechanical stimulation, so we cannot compare the adaptation response properties of PlVL and Pa7,9,11 neurons.

Functional role of PlVL neurons

Mechanosensory neurons in mollusks in some cases make direct excitatory chemical connections to the motor neurons, contributing to withdrawal (Walters et al. 1983), or may activate interneurons mediating whole-body withdrawal behavior (Arshavsky et al. 1994b; Inoue et al. 1996). Known structure of the network underlying withdrawal behavior in Helix includes nine giant premotor withdrawal interneurons receiving convergent multimodal sensory input and recruiting the appropriate motor neurons (Balaban 1983; Balaban and Zakharov 1992). It has been previously shown that intracellular stimulation of parietal interneurons (Pa3) elicited pneumostome closure as a component of withdrawal response, while intracellularly induced spike discharge of pleural interneurons (Pl1) induced muscular contraction of ipsilateral body wall (Balaban 1979) and withdrawal of ommatophores (I. S. Zakharov, personal communication). We demonstrated that even one PlVL neuron is sufficient to induce action potentials in the Pl1 neuron. Therefore it appears to be that PlVL neurons are capable of mediating at least part of the withdrawal behavior consisting of tentacle contraction. In addition, depolarization of parietal withdrawal interneurons elicited by action potentials in PlVL cells would decrease the threshold for withdrawal reactions and thus contribute to the behavioral output.