INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The feeding behavior of Aplysia consists of a sequence of appetitive and consummatory behaviors (Kupfermann 1974a,b; Kupfermann et al. 1991). First, food (seaweed) in the immediate environment of the animal activates appetitive behaviors, such as locomotion and head waving, which bring the animal in contact with the food (Teyke et al. 1990, 1992). Second, contact with the food activates consummatory behaviors, such as ingestion (biting and swallowing) or rejection (Kupfermann 1974a; Susswein et al. 1976). Consummatory behaviors involve rhythmic movements of feeding organs, including the lips, jaws, buccal mass, odontophore, and esophagus (Drushel et al. 1997; Kupfermann 1974a; Morton and Chiel 1993a; Rosen et al. 1997; Weiss et al. 1986). During a bite, the odontophore is rotated forward (i.e., protraction) as the jaws open. Initially, the two halves of the radula (toothed grasping surfaces of the odontophore) are separated during protraction. Before the peak of protraction, however, the radula begins to close and grasp the food. The radula remains closed as the odontophore retracts (backward rotation), which brings the food into the buccal cavity, and the jaws close. Swallowing consists of rhythmic movements of the odontophore and radula, which are similar to those during biting, and is associated with peristaltic contractions of the esophagus. Unlike biting, swallowing is not associated with opening of the jaws. In addition to biting and swallowing, the feeding apparatus can produce rejection movements in response to inedible material in the buccal cavity. During rejection, the radula is closed as the odontophore protracts and open as it retracts, causing the unwanted material to be ejected from the buccal cavity. Thus the motor programs that mediate consummatory feeding behaviors can be described, in general, as having two phases: a protraction phase followed by a retraction phase. During ingestion, the radula is open during the initial phase of protraction and closed during retraction, whereas during rejection, the radula is closed during protraction and open during retraction. Thus the two-phase feeding motor program of Aplysia, which is a browser, differs from other herbivorous gastropods (e.g., Helisoma, Helix, Limax, Lymnaea, Planorbarius, etc.), which are raspers and have three phases of feeding cycle (for review, see Kupfermann 1974a,b; Willows 1985).

Using a variety of intact, semi-intact, and reduced preparations, recent studies have begun to relate specific patterns of neural activity recorded in vivo and in situ to aspects of consummatory feeding behaviors (Chiel et al. 1986; Church and Lloyd 1994; Cropper et al. 1990; Evans and Cropper 1997; Evans et al. 1996; Fiore et al. 1992; Hurwitz et al. 1996; Jahan-Parwar and Fredman 1983; Kabotyanski et al. 1997, 1998a; Kupfermann and Weiss 1982; Morton and Chiel 1993a,b; Nagahama and Takata 1987, 1988, 1990; Nargeot et al. 1997, 1999a-c; Perrins and Weiss 1996, 1998; Rosen et al. 1991, 1997; Scott et al. 1995; Susswein et al. 1996; Weiss et al. 1978, 1986). Two types of buccal motor programs (BMPs) have been characterized during recordings in vivo. One type of BMP was associated primarily with ingestion, whereas the other was associated primarily with rejection (Morton and Chiel 1993a). The two BMPs were distinguished by the timing of large-unit activity in the radula nerve (R n.) relative to the onset of large-unit activity in buccal nerve 2 (n.2). During ingestion, large-unit activity in the two nerves overlapped to a large degree. In contrast, during rejection, large-unit closer activity in R n. preceded large-unit activity in n.2. Moreover, some motor neurons that contribute to the extracellular recordings have been identified (e.g., Hurwitz et al. 1996; Morton and Chiel 1993b; Nargeot et al. 1997). Some of the large units that were recorded in the R n. corresponded to activity in radula-closer motor neuron B8, whereas some of the large units recorded in the n.2 corresponded to activity in radula-retractor motor neurons B10 and B9. Thus the timing of activity in radula-closure motor neurons relative to activity in radula-retractor motor neurons was predictive of the type of consummatory behavior (i.e., ingestion or rejection) in the intact animal.

The in vivo studies have provided a framework, or set of criteria, that can be used to evaluate the potential behavioral relevance of BMPs observed in more reduced preparations and/or in response to different stimuli or experimental conditions. Using these criteria, several studies have found that the isolated buccal ganglia retain the circuitry necessary to produce at least two behaviorally relevant patterns (i.e., rejection-like and ingestion-like BMPs) and that the central pattern generator (CPG) in the buccal ganglia can switch between different functional configurations (e.g., Hurwitz and Susswein 1996; Hurwitz et al. 1997; Kabotyanski et al. 1997, 1998a; Morton et al. 1991; Nargeot et al. 1997, 1999a,b; Plummer and Kirk 1990; Rose 1972; Sossin et al. 1987; Susswein and Byrne 1988). The mechanisms underlying the generation of the different BMPs and the switching between different BMPs is not well understood, however.

Several lines of evidence suggest that the generation and switching among different BMPs may be mediated, in part, by the actions of catecholamines (e.g., dopamine, DA) and/or the indolamine serotonin (5-HT). First, the CPG of the buccal ganglia contains an intrinsic system of putative DA-containing cells (e.g., cells B20 and B65) (Kabotyanski et al. 1994, 1998a; Teyke et al. 1993) and receives extrinsic catecholaminergic inputs (e.g., cell CBI-1) (Rosen et al. 1991). Second, direct depolarization of these cells elicits rhythmic neural activity in the CPG and, in some instances, may selectively elicit ingestion-like BMPs. Third, the CPG of the buccal ganglia receives extrinsic serotonergic inputs (e.g., the serotonergic metacerebral cell, MCC) (Weiss et al. 1978, 1981; see also Goldstein and Schwartz 1989; Rathouz and Kirk 1988; Salimova et al. 1987; Soinila and Mpitsos 1991; Susswein et al. 1993). Fourth, firing MCC increases the rate of BMPs under some experimental conditions (Kupfermann et al. 1979; Weiss et al. 1978). Finally, in vivo recordings from freely behaving animals indicate that MCC is active during feeding (Fiore et al. 1992; Kupfermann and Weiss 1982; Weiss et al. 1978) and that lesions of the MCC induced deficits in biting (Rosen et al. 1983, 1989). These results suggest that DA and 5-HT may play roles in initiating rhythmic activity in the CPG and, more specifically, in organizing those BMPs that underlie aspects of ingestion.

To further characterize the roles of DA and 5-HT in the feeding system of Aplysia, the present study examined the effects of DA and 5-HT in progressively more reduced preparations. Perfusing a semi-intact head preparation with DA (50 µM) induced feeding-like movements that were similar to bites in the intact animal. Similarly, perfusing isolated ganglia preparations with DA (50 µM) elicited ingestion-like BMPs. Serotonin (5 µM) also induced feeding-like movements when perfused into the semi-intact preparation. The 5-HT-induced movements were similar to swallows, however. Perfusing isolated buccal ganglia with 5-HT (5 µM) did not significantly change the number of BMPs nor did it significantly bias the output of the CPG toward ingestion-like BMPs. Finally, the effects of DA and 5-HT on several cells (e.g., B4/5, B8, B31/32, B34, B35, B51, B63, and B64) and synaptic connections within the CPG were examined. The results of the present study suggest that DA and 5-HT regulate the functional configuration of the CPG and thereby play distinctive roles in organizing different aspects of feeding. Preliminary reports of some of these results have appeared in abstract form (Baxter and Byrne 1993; Baxter et al. 1995; Cushman et al. 1995; Kabotyanski et al. 1993, 1995, 1998b).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Aplysia californica (150-300 g) were obtained from Alacrity Marine Biological Services (Redondo Beach, CA), Marine Specimens Unlimited (Pacific Palisades, CA), Marinus (Long Beach, CA), and Pacific Biomarine (Venice, CA). Animals were kept in aerated aquaria containing artificial seawater (ASW; Instant Ocean, Aquarium Systems, Mentor, OH), which was maintained at 15°C. Animals were fed dried seaweed (Hang Loong Marine Products, Japan) three times a week. Before dissections, animals were anesthetized by an injection of isotonic MgCl₂.

Semi-intact head preparation

The semi-intact head preparation (Fig. 1) consisted of the anterior portion of the animal (i.e., an isolated head) that retained, in situ, the external structures of the head (e.g., lips, tentacles, and rhinophores), the feeding organs (e.g., jaws, buccal mass, and esophagus), and the buccal, cerebral, pleural, and pedal ganglia (see also Chiel et al. 1986; Drushel et al. 1997; Nagahama and Takata 1987, 1988; Rosen et al. 1991; Weiss et al. 1986). Before each experiment, animals were food-deprived for 2 days. The head was cut from the rest of the animal slightly rostral to the parapodia (dorsally) and slightly caudal to the beginning of the foot (ventrally). The esophagus was severed rostral to the crop and was canulated with a polyvinyl tube, 80 mm long and 3 mm in diameter. The anterior aorta was severed slightly caudal to the pedal artery. Another polyvinyl tube, 80 mm long and 2 mm in diameter, was inserted in the aorta and protruded rostrally into the buccal artery. One ligature secured the buccal artery to the tube, and another constricted the anterior aorta so that the perfusion of the buccal artery bypassed the pedal and cephalic arteries. The peripheral nerves from the pleural-pedal ganglia that projected caudally (e.g., PL2, pleuroabdominal connectives, P7, P8, P9, and P10) were severed. Once the gross dissection was completed, the head was transferred to an experimental perplex chamber, which contained ASW. Within the chamber, the preparation was mounted on a polyurethane tube, which had a diameter similar to that of the "neck" of the isolated head. The head was mounted on the tube by pinning the cut edges of the neck to the tube as well as by tying a ligature around the overlapping portions of the neck and tube. The polyurethane tube was attached on a perplex tube that was attached to one wall of the experimental chamber. This wall had sliding gate/door (not shown) through which the two canula were drawn outside the chamber. The gate then was closed and sealed off with petroleum jelly (Vaseline). Thus the preparation had two isolated compartments: one outside the head and another inside (Fig. 1). The volume of the internal compartment (30-40 ml) depended on the size of the head.

View larger version (26K): [in this window] [in a new window]   Fig. 1. Schematic drawing of the semi-intact preparation. Preparation consisted of the anterior portion of the animal (i.e., an isolated head) that retained, in situ, the external structures of the head, the feeding organs, and the buccal, cerebral, pleural, and pedal ganglia. Esophagus was canulated with a polyvinyl tube. Another polyvinyl tube was inserted in the aorta, and the buccal artery was canulated bypassing the pedal and cephalic arteries. Peripheral nerves from the pleural-pedal ganglia that projected caudally were severed. Head was mounted on a polyurethane tube by pinning the cut edges of the neck to the tube as well as by tying a ligature around the overlapping portions of the neck and tube. Polyurethane tube was attached to a perplex tube that was attached hermetically to 1 wall of an experimental perplex chamber containing artificial seawater (ASW). This wall had sliding gate (not shown), through which the 2 canuli were drawn outside the chamber. Gate then was closed and sealed off with Vaseline. Thus the preparation had 2 isolated compartments: 1 outside the head and another inside. ASW was pumped into the arterial system via inflow tube connected to the canulated buccal artery. To maintain hydroskeletal inflation, positive hydrostatic pressure was provided inside the head by elevating the outflow tube 30-40 mm above the surface of ASW in the chamber. A thread was used to measure feeding movements. It was attached to a lever of isotonic transducer and was drawn into the mouth, between the 2 halves of the radula and through the esophagus and its canulaoutside the chamber. A counterbalance weight was attached to the opposite arm of the transducer's lever so that a small constant outward force of 0.2 g was applied to the thread, ensuring that the thread always was stretched and transduced movements in both directions. At the beginning of each experiment, the transducer's lever was positioned so that it was possible to measure both inward and outward movements.

The canulated buccal artery was used to pump ASW into the arterial system and thereby perfuse the tissue and simulate the normal hydroskeletal inflation of the animal. The rate of perfusion was 1-2 ml/min. To maintain hydroskeletal inflation, positive hydrostatic pressure was provided inside the head by elevating the open end of the outflow tube 30-40 mm above the surface of ASW in the chamber. In addition, the canulated aorta was used to perfuse drugs (e.g., DA or 5-HT) into the preparation. Note that the drugs in chosen concentrations (see following text) were applied directly to the buccal ganglia, buccal mass, and the body wall. They reached the cerebral, pedal, and pleural ganglia much diluted and delayed. (If the volume of internal head compartment was 30 ml, perfusing it with a drug at 1 ml/min rate would reach half the final concentration in ~30 min)

The esophagus was canulated for two reasons: first, to prevent exchange of fluids between internal and external compartments of the head preparation through the buccal orifice and, second, to control a thread that was fed through the radula. The thread was used to measure feeding movements. It was attached to a lever of Isotonic Transducer (Model 60-3000, Harvard Apparatus, Natick, MA) and was drawn into the mouth, between the two halves of the radula and through the esophagus (Fig. 1). It was adjusted before each experiment to set the initial point of recording. A counterbalance weight was attached to the opposite arm of the transducer's lever so that a small constant outward force of 0.2 g was applied to the thread, ensuring that the thread always was stretched and transduced movements in both directions. The signal from the force transducer was amplified further with an oscilloscope, displayed on a chart recorder, and stored on magnetic tape. In addition, two video cameras and a video mixer were used to make a video tape record of the semi-intact preparation. One camera was focused on the mouth of the isolated head preparation, and the other was focused on the signal from the force transducer, which was displayed on the chart recorder. The two simultaneous video images were combined by the video mixer into one image, which was displayed and stored on video tape for subsequent analysis. Experiments were performed at 15°C.

In vitro preparation

An isolated buccal ganglia preparation was used to make extracellular recordings of BMPs and intracellular recordings from identified cells. Typically the peripheral nerves of the buccal ganglia were severed at the point where they entered the buccal mass, esophagus, and salivary glands. The cerebral-buccal connectives (C-B conn.) were severed close to the cerebral ganglion. The buccal ganglia were removed from the animal and pinned to the floor of a recording chamber, which was coated with a silicone elastomer (Sylgard 184, Dow Corning, Midland, MI). In those preparations where only extracellular recordings were made, the connective tissue sheath that covers the buccal ganglia was left intact and the cut ends of the R n. and buccal nerves 1, 2, and 3 (n.1, n.2, and n.3) were drawn into extracellular electrodes. (The nerve designations are from Gardner 1971; see also Church and Lloyd 1994; Morton and Chiel 1993b; Scott et al. 1991.) In those preparations in which both intracellular and extracellular recordings were made, the buccal ganglia were pinned to the floor of the recording chamber so that the rostral surface of the right buccal ganglion and caudal surface of the left buccal ganglion were accessible. The ganglia were desheathed, and only the R n. was drawn into an extracellular electrode. For intracellular recordings, neurons were identified by their position, size, projection of axons in peripheral nerves, nature of synaptic connectivity, characteristic patterns of firing during BMPs, and intrinsic biophysical properties (see RESULTS for details).

Conventional techniques were used for extracellular and intracellular recordings of neural activity. Briefly, extracellular recordings of neural activity were made with polyethylene suction electrodes and AC coupled amplifiers (Model 1700, AM Systems, Everett, WA). The suction electrodes also were used to stimulate buccal nerves. Intracellular electrodes were filled with a solution of 3 M potassium acetate and 100 mM potassium chloride and had an impedance of 6-15 M. The signals from the intracellular electrodes were amplified using DC-coupled amplifiers (models Axoclamp 2A and Axoprobe 1A, Axon Instruments, Burlingame, CA). The extracellular and intracellular signals were amplified further and displayed on a chart recorder, and an oscilloscope and were recorded on magnetic tape.

Solutions

Generally, control saline consisted of ASW, to which 10 mM of either TRIZMA or HEPES buffer (Sigma Chemical, St. Louis, MO) was added and the pH was adjusted to 7.4. Solutions containing L-3,4-dihydroxyphenylalanine, a metabolic precursor of catecholamines (DOPA; Calbiochem, La Jolla, CA), 3-hydroxytyramine HCl (dopamine, DA; Calbiochem), and 5-hydroxytryptamine creatinine sulfate complex (serotonin, 5-HT; Sigma Chemical) were made immediately before use. Solutions of DA and DOPA also contained an equimolar concentration of an antioxidant (ascorbic acid; Sigma Chemical). In those experiments in which either DA or DOPA was used, the control saline also contained ascorbic acid.

Unless otherwise noted, 50 µM DA, 250 µM DOPA, or 5 µM 5-HT was used. These concentrations were selected on the basis of preliminary studies that examined the effects of 1-500 µM DA, 0.5-50 µM 5-HT, and 100-1,000 µM of DOPA. Concentrations were selected that elicited reliable, stable, but not maximal responses. Moreover these concentrations were similar to concentrations used in previous studies (e.g., Ascher 1972; Gospe and Wilson 1980; Kabotyanski et al. 1994; Kramer and Levitan 1990; Ocorr and Byrne 1985; Shozushima 1984; Sossin et al. 1987; Teyke et al. 1993; Weiss and Drummond 1981; see also Gospe 1983; Trimble and Barker 1984; Yeoman et al. 1994). In previous studies, levels of DA and 5-HT were measured in the CNS, specific ganglia and their neuropils, monoamine-containing neurons, and even vesicles (e.g., Chien et al. 1990, 1995; McCaman et al. 1973, 1979). These measurements, however, do not provide information about effective concentrations acting at synaptic contacts. It is possible to estimate such concentrations by comparing cellular responses to exogenous transmitters with responses elicited by stimulating DA- or 5-HT-containing neurons or by studying binding properties of DA and 5-HT receptors. These concentrations are on the scale of 10⁶ M for 5-HT, and 10⁵ M for DA (e.g., Ascher 1972; Fox and Lloyd 1998; Gospe 1983; Magoski et al. 1995; Shozushima 1984). Thus we believe that the concentrations used in the present study are in the physiological range.

Summary data were represented as means ± SE or as medians with interquartile range (1st quartile to 3rd quartile). The statistical methods of analyses were indicated in the RESULTS and P < 0.05 was taken as indicating significant differences.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

DA and 5-HT coordinate different feeding-like movements a semi-intact head preparation

As a first step toward investigating the role of biogenic amines in feeding, it was important to determine the behavioral significance of the DA- and 5-HT-induced activity. To address this issue, we developed semi-intact head preparations in which either DA or 5-HT could be applied and behavior could be measured in quantitative and controllable manner.

Results of pilot experiments

The final design of the semi-intact preparation was an outcome of extensive preliminary study. First, we assessed the behavioral effects of DA and 5HT by injecting them into the intact animals (n = 5). These results were unclear. For example, DA and DOPA induced changes in posture, local skin contractions, and foot disattachment as well as brief movements of the buccal mass. Because the main focus of this study was the modulation of consummatory feeding, we then used a more reduced preparation of head with only buccal ganglia attached to segregate consummatory feeding component from the rest of behavioral effects of the transmitters. This approach was based previous findings suggesting the buccal ganglia contain the circuit sufficient to generate consummatory feeding (e.g., Kirk 1989; Kupfermann 1974b; Morton et al. 1991). In this preparation (n = 4), DA elicited periodic movements of the buccal mass, but they were weak, did not appear as bites, or swallows or rejections, and produced only weak inward displacements of thread placed on radula. Successful feeding involves coordination of many muscle groups besides the buccal mass (foot, body wall, and extrinsic buccal muscles), redistribution of hydroskeleton, etc. In the next set of preparations (n = 4), we retained the cerebral, pedal, and pleural ganglia. We also canulated the anterior aorta and perfused it under small hydrostatic pressure to inflate tissues and provide some hydroskeletal support. When DA was perfused first through the cephalic and pedal arteries, however, we observed a mixture of effects: local contractions and withdrawals, penis protrusion as well as abortive feeding movements. We then adopted the preparation described in METHODS. In this preparation, the pedal, pleural, and cerebral ganglia were preserved, but bypassed during perfusion of ASW or monoamines. The advantage of this preparation was that the isolated head was able to assume a feeding-like "posture" and exhibit recognizable and strong feeding movements. Bypassing the cephalic and pedal arteries and perfusing DA or 5-HT through the buccal artery allowed the drugs to have their predominate effects on the buccal ganglia. After this preparation was tested (n = 9), quantitative experiments were performed in the final setup (Fig. 1).

The role of periphery also was addressed in the pilot experiments (n = 3) in which we removed all the CNS and perfused the head and the buccal mass with DA. No feeding movements were observed; this indicated that the DA-induced activity in the head preparation was not of peripheral origin.

DA induced bite-like movements

Figure 2 compares a bite that was elicited by presenting food (seaweed) to a freely behaving animal (Fig. 2A) to feeding-like movements that were elicited by perfusing a semi-intact preparation with DA (Fig. 2B). In the freely behaving animal, the bite began with the jaws opening to accommodate the protraction of the odontophore (Fig. 2A, 1 and 2). During this initial phase of protraction, the two halves of the radula were open. At the peak of the protraction phase (Fig. 2A3), the two halves of the radula were closed, and they remain closed as the odontophore retracted and the jaws closed (Fig. 2A, 4 and 5).

View larger version (42K): [in this window] [in a new window]   Fig. 2. Feeding-like movements in freely behaving animals and semi-intact preparations. A: line drawings depicting a sequence of movements of the jaws and radula/odontophore during a bite in an intact, freely behaving animal. Drawings were generated by tracing individual frames of a video record of a single bite, which was elicited by presenting the animal with a small strip of seaweed. (Note, the animal was not allowed to grasp the seaweed, thus it does not appear in the drawings.) A1: initially, the jaws were closed. A2: the jaws opened as the radula/odontophore protracted and the 2 halves of the radula were open. A3: approximate peak of protraction and the 2 halves of the radula have closed. A4: jaws began to close as the radula/odontophore were retracted. A5: finally, the jaws closed. B: line drawings of a dopamine (DA)-induced feeding-like movement in a semi-intact preparation. Insets: output of the isotonic force transducer, which was displayed simultaneously on a chart recorder and combined with the video record of the behavior (see METHODS). B1: initially, the jaws were closed. B2: jaws opened as the radula/odontophore protracted and the 2 halves of the radula were open slightly. B3: approximate peak of protraction and the 2 halves of the radula began to close. B4: jaws began to close as the closed radula/odontophore were retracted. B5: finally, the jaws were closed. Dashed line in the insert indicates the thread displacement recorded during this sequence of movements. Calibration: 10 s, 5 mm. B, botom: complete isotonic force transducer recording that monitored the displacement of the thread in the semi-intact preparation illustrated in B, 1-5. Upward deflections represent inward movements (i.e., ingestion), whereas downward deflections represent outward movements (i.e., rejection). Dashed line indicates position of the thread at the onset of drug application (at time = 0). The solid line indicates when 50 µM DA was added to the saline that perfused the semi-intact preparation. Before application of DA, several relatively slow displacements were recorded. Video record indicated that these slow displacements represented mainly movements of the entire preparation (i.e., head waving) rather than consummatory feeding-like movements. In the presence of DA, the activity of the semi-intact preparation increased and the video record indicated that many of the rapid movements were feeding-like. For example, the activity outlined by the shaded box was a series of bite-like movements, 1 of which was illustrated in B, 1-5. C, 1-4: line drawings from a video recording of a sequence of serotonin (5-HT)-induced feeding-like movements. Output of the isotonic force transducer (insets) indicated a large inward displacement of the thread (dashed line and calibration in C4 insetas in B5 inset). Throughout this sequence, however, the jaws of the semi-intact preparation remained closed. C, bottom: complete record of the isotonic force transducer recording that monitored the displacements of the thread in the semi-intact preparation illustrated in C, 1-4. During perfusion with control saline, little or no displacement of the thread occurred. Soon after perfusing with 5-HT (as indicated by the solid line), large inward displacements of the thread were recorded. Examination of the accompanying video record indicated that the jaws remained closed during these inward movements. Shaded box indicates the feeding-like movement depicted in C, 1-4.

DA induced a similar sequence of movements in the semi-intact preparation (Fig. 2B). These movements began with the jaws opening as the odontophore protracted (Fig. 2B, 1 and 2). During this initial phase of protraction, the two halves of the radula were open. At the peak of the protraction phase (Fig. 2B3), the two halves of the radula began to close, and they remained closed as the odontophore retracted and the jaws closed (Fig. 2B, 4 and 5). The recording from the isotonic force transducer for the sequence of DA-induced movements (Fig. 2, B, 1-5, insets) indicated that the thread was drawn into the mouth ~4 mm. Figure 2B, bottom, illustrates the complete record of that experiment from the isotonic force transducer. Before perfusion with DA, feeding-like movements occurred infrequently and there was little net displacement of the thread. Soon after the preparation was perfused with DA, the frequency of feed-like movements increased, and ~13 min into the perfusion, these movements began to produce a net inward displacement of the thread. By the end of the experiment, ~100 mm of thread had been drawn into the foregut and examination of the video record indicated that the jaws opened and closed rhythmically during these feeding-like movements. These results suggest the DA induced biting-like movements.

It should be noted that the effects of perfusing preparations with a transmitter, such as DA, may not reflect the normal role of the transmitter. For example, perfusion will activate all receptors for the transmitter that under normal conditions might not be active together. To partially address this issue, semi-intact head preparations were perfused with the metabolic precursor of catecholamines, DOPA. Precursors, such as DOPA, are presumed to be accumulated selectively and metabolized by neurons that use the transmitter and to act via increased release of the transmitter from synaptic terminals of those cells (e.g., Kabotyanski and Sakharov 1988, 1991; Kabotyanski et al. 1994; McCaman et al. 1984; Mitchell et al. 1992). As a result, precursors affect neural circuitry via endogenous mechanisms, reach their targets in physiologically relevant order and timing, and produce gradual and long-lasting effects.

Perfusing with DOPA increased the frequency of feeding-like movements, produced a net inward displacement (i.e., ingestion) of the thread (Fig. 3B), and the DOPA-induced feeding-like movements were similar to those induced by DA (not shown). Both the DA- and DOPA-induced inward displacements of the thread were accompanied by coordinated opening and closing of the jaws as the odontophore rhythmically protracted and retracted (i.e., movements similar to biting in freely behaving animals). The only notable difference between perfusing with DA and DOPA was the greater time required for the feeding-like movements to develop in presence of DOPA. The onsets of sustained feeding rhythm (Fig. 3C1) and net inward displacement of thread (i.e., ingestion; Fig. 3C2) were delayed for ~15 min longer in DOPA than in DA (Fig. 3C). This presumably reflected the time required for dopaminergic cells to accumulate DOPA and metabolize it into DA. Nevertheless a weak agonist also can have delayed effects. If this was the case, the effects would be weaker, however. Yet the effects of DOPA, although delayed, were notably stronger than effects of DA (Fig. 3, A and B). Thus delayed but stronger effects support the hypothesis that DOPA acts as the metabolic precursor of DA and not as a weak agonist.

View larger version (33K): [in this window] [in a new window]   Fig. 3. Ingestive feeding-like movements were induced by perfusing semi-intact preparations with either DA, L-3-4-dihydroxyphenylalanine (DOPA), or 5-HT. For each semi-intact preparation, the frequency (Hz) of feeding-like movements as well as the direction and amount of displacement of the thread (mm/h) was measured. Positive values for the displacement of the thread represented inward movement (i.e., ingestion), whereas negative values represented outward movement (i.e., rejection). Measurements were monitored while the preparation was perfused with control saline and after the addition of 50 µM DA, 250 µM DOPA, or 5 µM 5-HT to the saline. A: in 3 semi-intact preparations, perfusing with DA significantly increased the frequency of feeding-like movements (A1) and induced a significant inward displacement of the thread (A2). Average observation periods were 50 ± 10 min in control saline and 55 ± 13 min in the presence of DA. B: in 3 preparations, perfusing with DOPA significantly increased the frequency of feeding-like movements (B1) and induced a significant inward displacement of the thread (B2). Preparations were observed for an average of 70 ± 5 min in control saline and 80 ± 10 min in the presence of DOPA. C: onsets of sustained feeding-like movements (C1) and ingestion (i.e., net inward displacement of thread; C2) were delayed both in DA and DOPA. Effects of DOPA were delayed for ~15 min longer than those of DA, which presumably reflected the time required to metabolize DOPA into DA. D: in 4 preparations, perfusing with 5-HT did not increase the frequency of feeding-like movements (D1) but did induce a significant inward displacement of the thread (D2). Average observation periods were 56 ± 4 min in control saline and 60 ± 0 min in the presence of 5-HT.

5-HT induced swallow-like movements

Perfusing semi-intact preparations with 5-HT also induced a net inward displacement of the thread (i.e., ingestion; Fig. 2C). However, 5-HT-induced ingestion was different from DA- or DOPA-induced movements. As illustrated by the video record in Fig. 2C, 1-4, the jaws remained closed while the thread was drawn into the mouth (net inward displacement for this sequence was ~10 mm). Figure 2C, bottom, illustrates the complete record from the isotonic force transducer. Before perfusing with 5-HT, there was no net change in the position of the thread. Within ~15 min of perfusing with 5-HT, large inward displacements of the thread were recorded (i.e., ingestion-like movements). By the end of the experiment, ~80 mm of thread had been drawn into the foregut. On average, one 5-HT-induced movement displaced more thread than one DA-induced bite-like movement (2.730 ± 0.546 vs. 0.614 ± 0.076 mm; 2-sample t-test, t₅ = -3.26, 2-tailed P < 0.025). There were fewer movements in 5-HT than in DA (Fig. 3), however, so the resulting average total net displacement was smaller in serotonin (Fig. 3). Examination of the video record indicated that the jaws remained closed during these feeding-like movements. In addition, we could clearly see through transparent experimental chamber that each inward displacement of the thread was associated with strong movement of the buccal mass inside the "head." On the basis of definitions of swallowing during behavioral observations (Kupfermann 1974b), these results suggest that 5-HT induced swallow-like movements.

Figure 3 summarizes data from semi-intact preparations that were perfused with DA, DOPA (Fig. 3, A-C) or 5-HT (Fig. 3D). DA and DOPA had similar effects. First, DA and DOPA increased the frequency of feeding-like movements from 0.0056 ± 0.0029 to 0.0343 ± 0.0071 Hz and from 0.0086 ± 0.0021 to 0.0452 ± 0.0010 Hz, respectively. Paired t-tests indicated that these increases were significant (DA: t₂ = 4.54, 1-tailed P < 0.025; DOPA: t₂ = 15.32, 1-tailed P < 0.0025). [As mentioned in the preceding text, before these experiments, we conducted a series (n = 17) of pilot experiments using a simplified design. Thus we had specific hypotheses for the direction of the effects, which warranted 1-tailed tests used in this section]. Second, the feeding-like movements that were induced by DA or DOPA produced a net inward displacement of the thread. Before perfusion with DA, the average displacement of the thread was 0.3 ± 2.0 mm/h. During perfusion with DA, the average displacement was 78.8 ± 25.7 mm/h. Similarly, before perfusion with DOPA, the average displacement of the thread was 3.2 ± 2.2 mm/h, whereas during perfusion with DOPA, the average displacement was 117.5 ± 31.6 mm/h. Paired t-tests indicated that both of these changes were significant (DA: t₂ = 3.11, one-tailed P < 0.05; DOPA: t₂ = 3.59, 1-tailed P < 0.04). Although effects of DA and DOPA were similar in amplitude, the time courses of their development were different. DA elicited sustained feeding-like movements with a latency of 59 ± 33 s after the start of perfusion, whereas the effects of DOPA had a latency of 890 ± 18 s (Fig. 3C1). The difference is statistically significant (2-sample t-test, t₄ = 22.09, 1-tailed P < 0.00002). Moreover the number of feeding-like movements elicited during first 10 min after drug application was 25 ± 2.6 for DA and 6.3 ± 1.3 for DOPA; this was significantly different (t₄ = 6.30, 1-tailed P < 0.002). Net inward thread displacement started after 532 ± 130 s of DA perfusion versus 1,410 ± 206 s after DOPA perfusion (Fig. 3C2; t₄ = 3.60, 1-tailed P < 0.02). Because the latter effect of DOPA began ~20 min after start of perfusion, the observation times were ~1 h for DA and 1.33 h for DOPA. At 1 h, though, the average total net displacements were similar for the both compounds (90 ± 6.3 mm for DOPA and. 86.5 ± 43.3 mm for DA). This results suggests that DOPA-induced movements, although delayed, appeared to be more vigorous than DA-induced movements.

Perfusing semi-intact preparations with 5-HT did not significantly increase the frequency of feeding-like movements (Fig. 3D1). In control saline, the average frequency of feeding-like movements was 0.00375 ± 0.00083 Hz, and in the presence of 5-HT, the average frequency was 0.00667 ± 0.00170 Hz. A paired t-test indicated that this difference was not significant (t₃ = 1.21). In contrast, perfusing semi-intact preparations with 5-HT significantly increased the inward displacement of the thread (Fig. 3D2). In control saline, the average displacement of the thread was 3.7 ± 1.5 mm/h, and in the presence of 5-HT, the average displacement was 60.1 ± 11.2 mm/h. A paired t-test indicated that this difference was significant (t₃ = 5.58, 1-tailed P < 0.006).

The results from the semi-intact preparations indicated that both DA and 5-HT induced and organized ingestion-like movements. However, the results indicated that the two transmitters organized different aspects of ingestion. Bites appeared to be organized by DA, whereas 5-HT appeared to organize swallows. The actions of these two transmitters can result from modulation of either elements in the CNS (e.g., cellular and/or synaptic properties within the CPG) and/or elements in the periphery (e.g., the properties of muscle fibers, release at neuromuscular junctions, and/or sensory feedback). To investigate the neural mechanisms that underlie the actions of these two transmitters, we began by characterizing the motor programs that were expressed in preparations of isolated buccal ganglia.

Characterization of spontaneously occurring BMPs in vitro

Besides the buccal ganglia, other central ganglia play a role in integrating a correct feeding response in the head preparation, and their removal may change the parameters of feeding. However, previous studies showed that rhythmic movements of the odontophore and radula are mainly supported by the buccal ganglia (e.g., Kupfermann 1974b), and the monoamines in our experiments with the semi-intact preparation were applied primarily to the buccal ganglia via the buccal artery. Hence the analysis that follows focuses on the effects of DA and 5-HT on the CPG for rhythmic movements of the odontophore and radula in the isolated buccal ganglia.

During feeding, the rhythmic movements of the radula and odontophore are produced by coordinated contractions of muscles in the buccal mass, which in turn, are innervated by motor neurons located in the buccal ganglia. These motor neurons project to the buccal mass via four peripheral nerves; the R n. and n.1, n.2, and n.3, respectively (e.g., Nargeot et al. 1997; Scott et al. 1991). Thus extracellular recordings from these four nerves provide a comprehensive monitor of centrally generated buccal motor output. Before investigating the effects of DA and 5-HT on BMPs, we examined the patterns of neural activity that were generated in preparations of isolated buccal ganglia, which were perfused with control saline.

Figure 4 illustrates an extracellular recording of spontaneous rhythmic neural activity from a preparation of isolated buccal ganglia. A BMP was defined as a sequence of bursts of large-unit activity in all four buccal nerves. Spontaneous BMPs occurred most often as discrete, individual patterns (e.g., the neural activity indicated by the box in Fig. 4A), but occasionally BMPs occurred as a "chain" of multiple cycles of bursting activity (e.g., the neural activity indicated by the bar in Fig. 4A). Although all BMPs had some features in common (e.g., each began and ended with a burst of large-unit activity in n.1), several types of BMPs could be distinguished, in part, by the phase relationship between activity in R n. and n.2. For example, in Fig. 4B, the bursts of large-unit activity in R n. terminated before the burst of large-unit activity in n.2 (see also Nargeot et al. 1997). This temporal relationship was similar to the neural correlate of rejection described by Morton and Chiel (1993a,b).

View larger version (42K): [in this window] [in a new window]   Fig. 4. Characterization of buccal motor programs (BMPs) in preparations of isolated ganglia. A: buccal ganglia were perfused with control saline and efferent activity was recorded extracellularly from the radula nerve (R n.) and buccal nerves 1, 2, and 3 (n.1, n.2 and n.3, respectively).BMPs were defined as sequential bursts of large-unit activity in all 4 buccal nerves. (Occasionally, bursts of activity were observed in individual nerves or in subsets of nerves. These "incomplete" patterns were not included in the present study.) BMPs occurred most often as individual patterns (e.g., activity outlined by the dashed box in A), but occasionally a pattern occurred as a "chain" of multiple bursts (e.g., the activity indicated by the bar in A). Using the criteria developed in previous studies (Kabotyanski et al. 1998a; Morton and Chiel 1993a; Nargeot et al. 1997), individual BMPs were categorized into 3 groups. B1: rejection-like BMPs were characterized by large-unit activity in R n. ending before the onset of large-unit activity in n.2. C1: ingestion-like BMPs were characterized by a substantial overlap of the large-unit activity in R n. and n.2. D1: transitional-like BMPs were characterized by 2 bursts of large-unit activity in R n., 1 before the onset of large-unit activity in n.2 and a 2nd burst that overlapped with large-unit activity in n.2. B2, C2, and D2: schematic diagram summarizing the sequences of large-unit activity in the 4 buccal nerves during spontaneously occurring BMPs. Boxes represent the averaged (± SE) start and stop times of large-unit activity that was measured in a pseudorandom sample of 100 individual BMPs in 20 preparations. (Selection process was not truly random because chain BMPs were excluded and only 5 BMPs were selected from each preparation.) Filled boxes correspond to large-unit activity in n.1, n.2, and n.3, and the open boxes correspond to large-unit activity in R n. All times were relative to the start of large-unit activity in n.1. B2: averaged data from 47 rejection-like BMPs. C2: averaged data from 20 ingestion-like BMPs. D2: averaged data from 33 transitional-like BMPs. Note the phase shifts of the large-unit activity in R n. (open boxes) that distinguished the different types of BMPs. E: a chain pattern, which was indicated by the horizontal bar in A, plotted with an expanded time scale. Asterisk indicate the medium-sized units that were recorded in n.1 and that appeared to be unique to the chain BMP.

In addition to the rejection-like BMP (Fig. 4B), the isolated ganglia spontaneously expressed several other types of BMPs. A second type of BMP was characterized by a substantial overlap of large-unit activities in R n. and n.2 (Fig. 4C) (see also Nargeot et al. 1997). The temporal relationship was similar to the neural correlate of ingestion described by Morton and Chiel (1993a,b). A third type of BMP was characterized by two bursts of large-unit activity in R n. (Fig. 4D). The first burst in R n. occurred before large-unit activity in n.2 and the second burst overlapped with large-unit activity in n.2. This temporal relationship was similar to a pattern that was observed in vivo mainly during transitions to or from other types of responses (Morton and Chiel 1993a,b). A fourth type of BMP was the chain pattern (Fig. 4E). The chain BMP had some features similar to those of the individual BMPs. For example, the chain BMP began and ended with bursts of large-unit activity in n0.1, and the large-unit activity in R n. preceded that in n.2. Some differences were observed constantly, however. First, the chain BMP contained a burst of medium-unit activity in n.1 that overlapped with the burst of large-unit activity in n.2. Whenever this medium-unit activity was present, the rhythmic activity continued, and whenever this medium-unit activity was absent, the chain pattern terminated. Second, the bursts of large-unit activity in n.2 and n.3 did not overlap substantially. At present, the behavioral relevance of the chain BMP is unknown.

Extracellular recordings indicated that individual preparations could spontaneously express several different types of BMPs. For example, the rejection- and transitional-like BMPs illustrated in Fig. 4, B1 and D1, were recorded from the same preparation. Preparations did not express the different types of BMPs in equal numbers, however. Figure 5 illustrates the distribution of spontaneously expressed BMPs that was observed in 55 preparations. A median of 43% [interquartile range (IR): 9.5-69.5%] of the BMPs were rejection-like, 20% (IR: 0-48.5%) were transitional-like, 0% (IR: 0-21%) were ingestion-like, and 0% (IR: 0-17.5%) were categorized as other (e.g., chain patterns). For analyses of the distribution of different types of BMPs, we used Friedman's testa nonparametric test analogous to the two-factor ANOVA, repeated-measures design (Zar 1996). The types of BMPs were the fixed-level factor, and subjects were the random-effect factor with within subjects' repeated measures. The test does not depend on normal distribution assumptions and on scale of measurements. Medians were used to characterize the populations. The data are ranked and then ² test statistic is calculated. Friedman's test indicated a statistically significant difference in the median values among the four categories (₃² = 29.896, P < 0.001). Post hoc Dunnett's all pairwise multiple comparison test indicated that preparations of isolated buccal ganglia spontaneously expressed significantly more rejection-like BMPs than any other type (2-tailed P < 0.05). Thus although the CPG in the isolated buccal ganglia preparation could generate several different BMPs, its spontaneous activity appeared to be biased toward generating the rejection-like BMP.

View larger version (24K): [in this window] [in a new window]   Fig. 5. Distribution of types of spontaneously occurring BMPs in isolated buccal ganglia that were perfused with control saline. BMPs were categorized as being either rejection-like (e.g., Fig. 4B), transitional-like (e.g., Fig. 4D), ingestion-like (e.g., Fig. 4C), or as failing to correspond to these previously characterized motor programs (i.e., other; e.g., Fig. 4E). For each preparation, the 4 types of BMPs were expressed as a percent of total number of BMPs that were observed and the median of these values were illustrated in this bar graph. Preparations were observed for an average of 57 ± 3 min in control saline and a total of 605 BMPs were recorded in 55 preparations.

Effects of DA and 5-HT on spontaneous BMPs in vitro

MODULATING THE FREQUENCY OF SPONTANEOUSLY OCCURRING BMPS. To determine whether the frequency of spontaneously occurring BMPs changed over time in vitro, 16 preparations of isolated buccal ganglia were perfused with control saline for 2 h. The average frequency of BMPs was 0.0019 ± 0.0003 Hz during the first hour and 0.0021 ± 0.0003 Hz during the second hour (Fig. 6A). A paired t-test indicated that this change was not significant (t₁₅ = 0.84). Thus the frequency of spontaneously occurring BMPs was relatively constant for several hours in vitro. The effects of DA (50 µM) on the frequency of spontaneously occurring BMPs were examined in 19 preparations (Fig. 6B). Before application of DA, the average frequency of spontaneous BMPs was 0.00276 ± 0.00049 Hz. During perfusion with DA, the average frequency increased to 0.00701 ± 0.00109 Hz. A paired t-test indicated that this increase was significant (t₁₈ = 4.09, 2-tailed P < 0.001).

View larger version (42K): [in this window] [in a new window]   Fig. 6. DA but not 5-HT increased the frequency of spontaneously occurring BMPs in isolated buccal ganglia. BMPs were monitored via extracellular recordings from the R. n and n.1, n.2, and n.3. In each preparation, the frequency (Hz) of BMPs was calculated by dividing the total number of BMPs that were observed in a given saline (e.g., control saline, saline containing DA or saline containing 5-HT) by the duration, in seconds, of exposure to that particular saline. Chain patterns (e.g., Fig. 4E) were scored as a single pattern. A1: in control saline, BMPs occurred at a fairly consistent frequency. (Sample trace was taken from the 2nd 1-h period.) A2: in 16 control preparations, no significant change in the frequency of spontaneous BMPs was observed when ganglia were perfused with control saline for 2 h. B1: in the presence of DA (50 µM), the frequency of spontaneously occurring BMPs was increased. B2: in 19 preparations, perfusing with DA (50 µM) significantly increased the frequency of spontaneously occurring BMPs. Average durations for the observation periods were 51 ± 4 min in control saline and 39 ± 4 min in the presence of DA. C1: in the presence of 5-HT (5 µM), the frequency of spontaneously occurring BMPs was not changed. There was an increase in the general level of nonpatterned activity recorded from the nerves, however. C2: in 12 preparations, perfusing with 5-HT (5 µM) did not significantly change the average frequency of spontaneously occurring BMPs. Average durations for the observation periods were 53 ± 4 min in control saline and 43 ± 3 min in the presence of 5-HT.

View larger version (35K): [in this window] [in a new window]   Fig. 7. DA also increased the frequency of spontaneous BMPs during simultaneous intracellular and extracellular recordings in desheathed isolated buccal ganglia. A1: in control saline, spontaneous BMPs occurred at a low frequency. A2: in the same preparation, bath-applied DA (50 µM) elicited sustained fictive feeding of a higher frequency. Recording was made 30 min after the start of application of DA. Note that spikes gradually disappear in B34, the neuron associated with the generation of rejection-like BMPs. B: summary data from 9 preparations. DA (50 µM) significantly increased the average frequency of spontaneous BMPs by about eightfold during 1 h of application.

DA BUT NOT 5-HT BIASED THE SPONTANEOUS OUTPUT OF THE CPG TOWARD INGESTION-LIKE BMPS. In addition to increasing the frequency of spontaneously occurring BMPs, DA modulated the type of BMP that was expressed. Two examples of this second action of DA are illustrated in Fig. 8. During the 30 min that the preparation illustrated in Fig. 8A was perfused with control saline, five spontaneous BMPs were recorded and all five of these were rejection-like BMPs (e.g., Fig. 8A1). During the 30 min that this preparation was perfused with DA (50 µM), 11 BMPs were observed and all of these BMPs were ingestion-like (e.g., Fig. 8A2). Similar results were obtained in preparations in which simultaneous extracellular and intracellular recordings were used to monitor BMPs (Fig. 8B). Figure 8B1 illustrates a rejection-like BMP that was recorded in control saline, and Fig. 8B2 illustrates an ingestion-like BMP that was recorded from the same preparation after application of DA (50 µM). Recent studies have indicated that high levels of spiking activity in cells B4/5 and B34 were correlated with rejection-like BMPs (Hurwitz et al. 1997; Kabotyanski et al. 1997, 1998a). The level of spiking activity of B4/5 and B34 appeared to be reduced during the DA-induced ingestion-like BMP (Fig. 8B2), which suggested that DA may bias the functional configuration of the CPG toward generating ingestion-like BMPs by decreasing the excitability of these cells (see following text).

View larger version (45K): [in this window] [in a new window]   Fig. 8. DA biased the output of the central pattern generator (CPG) toward ingestion-like BMPs. - - -, presumed transition from the protraction phase of the BMP (to the left) to the retraction phase (to the right). , radula closure activity (i.e., the burst(s) of large-unit activity in the R n. and firing of radula-closer motor neuron B8). A: extracellular recordings were used to monitor BMPs in control saline (A1) and after bath application of 50 µM DA (A2). A1: rejection-like BMP (i.e., large-unit activity in R n. preceded large-unit activity in n.2) recorded in control saline. A2: ingestion-like BMP (i.e., large-unit activity in the R n. coincided with large-unit activity in n.2) recorded in the presence of DA. B: simultaneous intracellular and extracellular recordings during spontaneous BMPs in control saline (B1) and in the presence of 50 µM DA (B2). B1: rejection-like BMP (from Fig. 7A1). As previously reported, B8 has an axon in the R n. and contributed to the large-unit activity recorded extracellularly from R n. Closure activity occurred primarily during the protraction phase (i.e., coinciding with activity in cell B34). B2: ingestion-like BMP (the 3rd cycle in Fig. 7A2). In the presence of DA, the majority of closure activity was shifted to the retraction phase (i.e., coinciding with activity in cell B64).

View larger version (16K): [in this window] [in a new window]   Fig. 9. Modulation of the types of spontaneously occurring BMPs in isolated buccal ganglia. A: in 16 control preparations, no significant changes in the distribution of BMPs were observed between the 1st and 2nd hours of recordings. B: in 19 preparations, bath application of DA (50 µM) biased the output of the CPG away from rejection-like BMPs and toward ingestion-like BMPs. C: in 12 preparations, perfusing isolated buccal ganglia with 5-HT (5 µM) did not significantly alter the distribution of the types of BMPs.

Modulation of cellular and synaptic properties by 5-HT and DA

To understand the mechanisms and the role of modulation of the CPG for feeding by DA or 5-HT, we have begun to examine sites in the CPG at which the transmitters might exert their actions. As a first step, it was necessary to characterize the cellular events associated with the action of modulators that occur in the CPG and output neurons under the same pharmacological conditions that led to fictive feeding (i.e., while existing synaptic connections were maintained). Figure 10 illustrates some of the loci within the feeding neural circuitry that were modulated by application of 5-HT. At least three biophysical properties of B31/32 were consistently modulated in the presence of 5-HT (n = 8 preparations). First, the excitability of B31/32 was reduced (Fig. 10A). Second, an oscillation in the resting membrane potential of B31/32 was induced (Fig. 10B). Third, the plateau-like potential in B31/32 appeared to be reduced (Fig. 10C). Similar 5-HT-induced changes were observed in B33 (not shown). In the presence of 5-HT, the membrane potential of B33 oscillated and its excitability was decreased (n = 3 preparations). In contrast to its actions on B31/32 and B33, 5-HT did not induce oscillations or any other observable change in the membrane potential of cell B35. It did, however, appear to increase the excitability of B35 (Fig. 10D) and increased the strength of the B35 synaptic connection to B4/5 (n = 4 preparations). Finally, although the presence of 5-HT did not induce a change in the resting membrane potential of cells B4/5, it did decrease the excitability of these cells and decreased the strength of the B4/5 synaptic connection to B31/32 (Fig. 10E; n = 7 preparations).

View larger version (28K): [in this window] [in a new window]   Fig. 10. Effects of bath application of 5-HT on elements of the CPG. A: excitability of B31/32 was decreased in the presence of 5-HT. Two electrodes were inserted into B31/32: 1 in the soma for injecting depolarizing current pulses () and another in the proximal segment of the axon for recording spiking activity. (Soma of B31/32 do not support action potential.) (see Hurwitz et al. 1994; Susswein and Byrne 1988.) In control saline, a 1-s, 5-nA current pulse elicited 23 spikes (A1), whereas in the presence of 5-HT, an identical stimulus elicited only 6 spikes (A2). B: subthreshold oscillations in the resting potential of B31/32 were observed consistently in the presence of 5-HT. C: plateau-like potential in B31/32 was reduced in the presence of 5-HT. Using 2 electrodes, a series of 2-s depolarizing current pulses () of 5, 10, and 15 nA were injected into B31/32, and the approximate threshold of the plateau-like potential could be monitored as a slowly developing and decaying depolarization (C1). (Duration and amplitude of the stimuli were adjusted to avoid eliciting a BMP. Smaller of the spike-like potentials are believed to represent spikes in the electrically coupled cells such as B63, whereas the larger spike-like potentials are believed to represent axonal spikes in B31/32) (see Hurwitz et al. 1997.) In the presence of 5-HT, identical stimuli elicited far less of the plateau-like potential (C2). D: excitability of B35 and the strength of its excitatory synaptic connection to B4/5 were enhanced in the presence of 5-HT. Two electrodes, 1 for passing current and another for monitoring membrane potential, were placed in both the presynaptic and postsynaptic cells (B35 and B4/5, respectively). In control saline (D1), a 2-nA, 1-s depolarizing current pulse () elicited 5 spikes in B35 and a train of excitatory postsynaptic potentials (EPSPs) in B4/5, the first of which was ~2 mV in amplitude. In the presence of 5-HT (D2), an identical stimulus elicited 11 spikes in B35 and the amplitude of the first EPSP in B4/5 was increased to ~3 mV. Membrane potentials of both cells were adjusted to control levels. E: bath application of 5-HT led to a decrease in the excitability of B4/5 and a decrease in the strength of synaptic connections from B4/5 to B31/32. In control saline, a 2.5-nA, 3-s current pulse () elicited a train of 31 spikes in B4/5 and PSPs in B31/32. In the presence of 5-HT, an identical stimulus elicited only 6 spikes in B4/5 and no PSPs were observed in B31/32. In both cells, only a single electrode was used to monitor the membrane potential and/or inject current pulses.

Some of the same sites that were modulated by bath application of 5-HT also were modulated by presence of DA, albeit in different ways. For example, bath application of DA depolarized the membrane potential of cells B31/32 (n = 7 preparations). This depolarization immediately preceded and appeared to induce the sustained rhythmic activity in the CPG. The depolarizations of membrane potentials in the presence of DA also were observed in cells B8 (n = 8 preparations), B34 (n = 5 preparations), and B65 (n = 2 preparations). In contrast, in the presence of DA, the membrane potentials were hyperpolarized in cells B4/5 (n = 14 preparations), B51 (n = 8 preparations), B63 (n = 2 preparations), and B64 (n = 5 preparations).

Figure 11 illustrates some additional effects of DA. In the presence of DA, the excitability of cell B34 appear to be reduced (Fig. 11A; n = 6 preparations) and the strength of the excitatory synaptic connection from B34 to B31/32 was decreased. Similarly the excitability of cells B4/5 (Fig. 11B; n = 14 preparations) (see also Kabotyanski et al. 1994, 1997, 1998a) was reduced. At the same time, B4/5 exhibited a lower rate of activity in presence of DA, and in ~20% of preparations, it produced only one to two spikes per cycle. In addition, bath application of DA decreased