Coordinated Excitatory Effect of GABAergic Interneurons on Three Feeding Motor Programs in the Mollusc Clione limacina

248 words; Text: 27 pages; Figures: 10; Tables: 0 Contacts: Dr. Norekian Arizona State University, School of Life Sciences Tempe, AZ 85287-4501, USA Phone: (480)965-4027; FAX: (480)965-6899 E-mail: [email protected] Articles in PresS. J Neurophysiol (August 25, 2004). doi:10.1152/jn.00722.2004 Copyright © 2004 by the American Physiological Society. Coordination of Neural Networks 2 Abstract Coordination between different motor centers is essential for the orderly production of all complex behaviors. Understanding the mechanisms of such coordination during feeding behavior in the carnivorous mollusc Clione limacina is the main goal of the current study. A bilaterally symmetrical interneuron identified in the cerebral ganglia and designated Cr-BM neuron produced coordinated activation of neural networks controlling three main feeding structures: prey capture appendages called buccal cones, chitinous hooks used for prey extraction from the shell, and the toothed radula. The Cr-BM neuron produced strong excitatory inputs to motoneurons controlling buccal cone protraction. It also induced a prominent activation of the neural networks controlling radula and hook rhythmic movements. In addition to the overall activation, Cr-BM neuron synaptic inputs to individual motoneurons coordinated their activity in a phase-dependent manner. The Cr-BM neuron produced depolarizing inputs to the radula protractor and hook retractor motoneurons, which are active in one phase, and hyperpolarizing inputs to the radula retractor and hook protractor motoneurons, which are active in the opposite phase. The Cr-BM neuron used GABA as its neurotransmitter. It was found to be GABAimmunoreactive in the double-labeling experiments. Exogenous GABA mimicked the effects produced by Cr-BM neuron on the postsynaptic neurons. The GABA antagonists bicuculline and picrotoxin blocked Cr-BM neuron-induced PSPs. The prominent coordinating effect produced by the Cr-BM neuron on the neural networks controlling three major elements of the feeding behavior in Clione suggests that this interneuron is an important part of the higher-order system for the feeding behavior.Coordination between different motor centers is essential for the orderly production of all complex behaviors. Understanding the mechanisms of such coordination during feeding behavior in the carnivorous mollusc Clione limacina is the main goal of the current study. A bilaterally symmetrical interneuron identified in the cerebral ganglia and designated Cr-BM neuron produced coordinated activation of neural networks controlling three main feeding structures: prey capture appendages called buccal cones, chitinous hooks used for prey extraction from the shell, and the toothed radula. The Cr-BM neuron produced strong excitatory inputs to motoneurons controlling buccal cone protraction. It also induced a prominent activation of the neural networks controlling radula and hook rhythmic movements. In addition to the overall activation, Cr-BM neuron synaptic inputs to individual motoneurons coordinated their activity in a phase-dependent manner. The Cr-BM neuron produced depolarizing inputs to the radula protractor and hook retractor motoneurons, which are active in one phase, and hyperpolarizing inputs to the radula retractor and hook protractor motoneurons, which are active in the opposite phase. The Cr-BM neuron used GABA as its neurotransmitter. It was found to be GABAimmunoreactive in the double-labeling experiments. Exogenous GABA mimicked the effects produced by Cr-BM neuron on the postsynaptic neurons. The GABA antagonists bicuculline and picrotoxin blocked Cr-BM neuron-induced PSPs. The prominent coordinating effect produced by the Cr-BM neuron on the neural networks controlling three major elements of the feeding behavior in Clione suggests that this interneuron is an important part of the higher-order system for the feeding behavior. Coordination of Neural Networks 3 Introduction Any complex behavior requires a specific coordination between various neural networks controlling its different aspects in order to achieve a meaningful behavioral output. Such coordination ensures the orderly production of a complex behavior and represents a universal principle of CNS functioning in both vertebrate and invertebrate animals. The focus of the current study is the feeding behavior of the carnivorous pteropod mollusc Clione limacina and the neuronal mechanisms of coordination between the main elements of this complex behavior. Clione is a highly specialized carnivore, which feeds only on two species of shelled pteropod molluscs of the genus Limacina (Wagner 1885; Lalli 1970; Lalli and Gilmer 1989). To seize its prey, Clione rapidly protracts tentacle-like oral appendages, called buccal cones, which surround the Limacina shell and hold it during the subsequent phases of feeding. Once buccal cones have gripped the prey, they begin to manipulate it so that the shell aperture is pressed against the mouth of Clione. Then two other feeding structures, chitinous hooks and the radula, are used to extract the soft body of Limacina from its shell, to be swallowed whole. The radula is a specialized structure found in all gastropod molluscs. Its rhythmic protraction-retraction movement is designed to seize the food and bring it to the opening of the esophagus. Chitinous hooks, which normally are retracted inside two symmetrical muscular hook sacs, are unique to Clione and other molluscs from the order Gymnosomata and reflect a high food specialization (Lalli 1970; Lalli and Gilmer 1989). The functional role of the hooks is to grab the soft body of Limacina and pull it out of the shell into the buccal cavity. Hook activity is also rhythmic and consists of protraction and retraction phases. We have shown in the previous investigation that the rhythmic movements of radula and hooks are highly coordinated in a phase-dependent manner (Malyshev and Norekian 2002). This phasedependent coordination was observed on the behavioral level and was always recorded on the motoneuronal level during spontaneous and induced rhythmic activity. Hook protractor neurons Coordination of Neural Networks 4 were active in the same phase with radula retractor neurons, while hook retractor neurons were bursting in phase with radula protractor neurons (Malyshev and Norekian 2002). We describe in this study and discuss the role of a bilaterally symmetrical cerebral interneuron, designated Cr-BM cell, which produces a prominent excitatory influence on all neural networks that control three major feeding structures in Clione: buccal cones, chitinous hooks and the radula. A preliminary general description of some of these effects has been previously published (Norekian 1995). The overall strong excitatory influence produced by the Cr-BM neuron on all identified feeding neural networks is presumably important during the “prey extraction” period of the feeding behavior, following the prey capture. In addition to the overall activation, the Cr-BM neuron has a strong coordinating influence on the individual elements of each neural network, which contributes to the phase-dependent coordination of their rhythmic activity. The Cr-BM interneuron apparently uses GABA as its excitatory transmitter based on double-labeling experiments, mimicking effect of exogenous GABA and blocking effects of GABA antagonists. This represents another interesting example of the excitatory role of GABA in the feeding behavior of Clione (Norekian 1999; Arshavsky et al. 1993). Methods Adult specimens of Clione limacina were collected at Friday Harbor Laboratories, University of Washington (Friday Harbor, WA) in the Spring-Summer season and at the White Sea Marine Laboratory of the Zoological Institute (White Sea, Russia) in the Summer-Autumn season. The animals were held in 1-gallon jars in a refrigerator at 5-7°C. Prior to dissection, animals were anesthetized in a 1:1 mixture of seawater and isotonic MgCl2 and then tightly pinned to a Sylgard-coated Petri dish. Electrophysiological experiments were performed on reduced preparations consisted of the CNS, head, and wings. All central nerves innervating the Coordination of Neural Networks 5 head were intact. Prior to electrophysiological recording, the sheath of the central ganglia was softened by bathing the preparation in a 1 mg/ml solution of protease (Sigma, type XIV) for 5 minutes, followed by a 30-minute wash in filtered seawater. Intracellular recordings from individual neurons were made with glass microelectrodes (resistances 10-30 MΩ) filled with 2 M potassium acetate. Electrophysiological signals were amplified, displayed, and recorded using conventional electrophysiological techniques. Intracellular stimulation was achieved via an amplifier bridge circuit. To test for monosynaptic connections, a high divalent cation solution was used (in mM: 110 MgCl2, 25 CaCl2, 400 NaCl, 10 KCl, and 3 NaHCO3, pH 7.4). For morphological investigation of recorded neurons, a 5% solution of 5(6)-carboxyfluorescein (Sigma) prepared in 2 M potassium acetate was iontophoresed via the recording electrodes (resistances, 20-40 MΩ) with 1-10 nA negative current pulses for 5-30 minutes. Injected cells were observed live in the recording dish with a Nikon (Japan) epifluorescence microscope and with a BioRad (Hercules, CA) MRC 600 laser scanning confocal microscope. GABA was applied locally to the soma of identified neurons via pressure ejection or iontophoretic application. For pressure ejection, glass micropipettes were filled with 5 mM GABA solution in filtered seawater and connected to the air-pressure system (PV830 Pneumatic PicoPump, WPI, USA), which delivered pressure pulses of 40 to 60 p.s.i. and 200 ms duration. Fast Green dye (0.02 %) was added in the solution for monitoring the drug delivery. For iontophoretic application we used glass micropipettes with tip diameter 1-2 μm filled with 1 M GABA solution (pH 4). Iontophoretic currents had amplitude between 50 and 100 nA and were applied as brief pulses of 50-100 ms duration (Stimulator S4KR and Stimulus isolation unit SIU 4678, Grass instruments). GABA antagonists were applied with the use of a graduated 1-ml Coordination of Neural Networks 6 pipette. The final concentration was estimated from the known volume of injected solution and the known volume of saline in the recording dish. For the wholemount immunocytochemical procedure, the preparations were fixed for 3 hours in 4% paraformaldehyde and 0.1% glutaraldehyde in PBS (pH=6.5-7.0) at room temperature (15-20°C). To reduce high nonspecific fluorescence caused by glutaraldehyde fixation, the tissues were incubated overnight in 4% sodium borohydride in PBS (Kosaka et al. 1986). The preparations were washed for 12 hours in PBS and preincubated in PBS containing 0.1% triton X-100 to increase tissue permeability. The tissues were then exposed to 6% goat serum in PBS and 0.1% Triton X100 for 6 hours to reduce nonspecific staining, and incubated 36 hours at 5°C in GABA antibody (polyclonal GABA antisera raised in guinea pig; Eugene Tech International, Inc., Ridgefield Park NJ). The dilution of primary antibody was 1:500. After a 12-hour wash in PBS, the tissues were placed for 24 hours in fluorescein-labeled secondary antibody (working concentration 40 μg/ml; anti-guinea pig IgG produced in goat; Vector Labs, Burlingame CA). The secondary antibody was removed with several PBS exchanges, and preparations were washed overnight. The tissues were then cleared in xylene, mounted in DPX, and examined in wholemount with a Nikon fluoresence microscope and a Bio-Rad MRC 600 laser scanning confocal microscope. The first set of control experiments included preadsorption of the primary antibody with 50 μM GABA-BSA conjugate for 6 hours before processing the tissue. In the second set of controls, primary antibody was omitted from the procedure. No staining resulted in both sets of control experiments. For double-labeling experiments, interneurons were injected with neurobiotin (Vector Laboratories, Burlingame CA). The preparations were then fixed in 4% paraphormaldehyde and 0.1% glutaraldehyde in PBS, and incubated 12 hours in Texas Red-labeled avidin (Vector Laboratories, Burlingame CA) to visualize neurobiotin-filled interneurons. The preparations were Coordination of Neural Networks 7 then processed for immunocytochemical reaction described above. By switching filters in the fluorescence microscope or laser scanning confocal microscope for Texas Red and fluorescein, interneurons were identified as GABA-immunoreactive. Texas Red was not visible with the fluorescein filters, and fluorescein was not visible with the Texas Red filters, thus providing a clear comparison during filter switching. Results The Cr-BM interneuron morphology and overall effect on the neural networks The bilaterally symmetrical Cr-BM interneuron was located on the ventral surface of the cerebral ganglia in the anterior region between the head nerves N1 and N2 that innervate the head of the animal (Fig.1). The size of the cell body was 25-30 μm in diameter. A single large axon of the Cr-BM neuron exited the cerebral ganglia into the ipsilateral cerebro-buccal connective and innervated the neuropile of both buccal ganglia (Fig.1). A few small processes also branched in the neuropile of the ipsilateral cerebral ganglion. The morphological structure of 16 Cr-BM interneurons was studied in 12 preparations after intracellular injection with the fluorescent dye carboxyfluorescein. The intracelular activation of the otherwise silent Cr-BM interneuron produced a prominent effect in the semi-intact preparation opening of the skin folds that cover the mouth of Clione and rhythmic muscular contractions of the buccal mass, which included coordinated protraction-retraction movements of hooks and radula (N=24). Intracellular recordings from the motoneurons controlling radula and hook movements demonstrated the corresponding rhythmic activity (Fig.2). Induced by the Cr-BM neuron, rhythmic activity in the radula controlling network always started from the burst of spikes in the radula protractor Bc-RP motoneurons and Bc-PIN interneuron (N=20; Fig.2A). The Bc-PIN interneuron has been previously identified as a Coordination of Neural Networks 8 part of the radula rhythm generator, which is active in the radula protraction phase (Arshavsky et al., 1993). Radula retractor Bc-RR motoneurons were inhibited during this initial phase. In the hook controlling neural network, intracellular stimulation of the Cr-BM neuron produced initial burst of activity in the hook retractor Bc-HR motoneurons and inhibition of the hook protractor Bc-HP motoneurons, which were followed by a few cycles of rhythmic activity (N=18; Fig.2B). Stimulation of the Cr-BM interneurons also triggered a prominent spike activity in the Cr-A neurons, which control opening of the skin folds and buccal cone protraction (N=32; Fig.2C). Specific synaptic connections produced by the Cr-BM interneurons The Cr-BM interneuron induced rhythmic activity in the buccal neural network controlling radula movements. Radula protractor motoneurons received prominent excitatory PSPs from the Cr-BM neurons. Each induced Cr-BM neuron spike produced a single Bc-RP neuron EPSP, which persisted in the high divalent solution suggesting monosynaptic connection (N=14; Fig.3A). The Bc-PIN interneuron, which is active in the radula protraction phase, also received prominent excitatory inputs from the Cr-BM interneuron. Each Cr-BM spike generated a large-amplitude EPSP in the Bc-PIN cell, which persisted in high divalent solution (N=26; Fig.3A). Radula retractor Bc-RR motoneurons received inhibitory inputs from the Cr-BM interneurons. Each induced Cr-BM neuron spike triggered a single Bc-RR neuron IPSP, which persisted in high divalent solution suggesting monosynaptic connection (N=12; Fig.3B). Thus, having the overall excitatory effect on the radula controlling neural system, the Cr-BM neurons also produced coordinating inputs to the specific elements of this system ensuring that radula protractor and radula retractor neurons are active in the opposite phases of their rhythmic cycle. Stimulation of the Cr-BM interneuron induced rhythmic activity in the hook controlling neural system. The effect produced by the Cr-BM neuron on the buccal hook retractor Bc-HR Coordination of Neural Networks 9 neuron was excitatory (N=18; Fig.4A). However, there was always a short delay before the onset of the excitation. Close observation revealed that Cr-BM neuron activation always produced a combination of two responses in the Bc-HR neurons (N=15; Fig.4B). One response was a prominent slow depolarization, which was responsible for the overall activation of the Bc-HR neurons. The second response was represented by fast IPSPs, which appeared at a stable 1:1 spike to IPSP ratio and apparently added to the delay of the overall excitatory response in the BcHR neurons. Both types of responses persisted in high divalent solution suggesting monosynaptic connection and could be explained by the existence of two separate receptor types in the Bc-HR neurons. The hook protractor Bc-HP motoneurons received inhibitory inputs from the Cr-BM interneuron. Each induced Cr-BM neuron spike produced an individual IPSP, which persisted in high divalent solution (N=12; Fig.4C). Thus, the Cr-BM interneuron, in addition to the overall activation of the hook controlling neural system, provided fine coordination of its different elements ensuring their phase specific rhythmic activity. The Cr-A neurons that control opening of the skin folds and buccal cone protraction are located in the cerebral ganglia, with some of them in a very close proximity to the Cr-BM interneuron location (Norekian and Satterlie 1993a). The excitatory inputs from the Cr-BM neurons to the Cr-A neurons were highly efficient. However, they were apparently not monosynaptic; some of the induced Cr-BM neuron spikes failed in high divalent solution to produce EPSPs in these cells (N=17). In addition, in the reduced preparations that did not include buccal ganglia, the Cr-BM interneurons did not have any effect on the Cr-A neurons (N=6). These experiments suggested that a polysynaptic connection between the Cr-BM interneurons and the Cr-A neurons occur via the buccal ganglia. To confirm this suggestion the following experiments were performed. In the semi-intact preparations, several prominent excitatory inputs in the Cr-A neurons triggered by induced bursts of spikes in the Cr-BM interneurons were Coordination of Neural Networks 10 recorded. Then the ipsilateral cerebro-buccal connective was cut with fine scissors. After the cut, all Cr-BM neuron-induced responses in the Cr-A neurons completely disappeared (N=5; Fig.5A). One of the buccal interneurons, which could be involved in the transition of the excitatory inputs from the Cr-BM interneuron to the Cr-A neurons, was Bc-PIN interneuron. The Bc-PIN interneuron has been previously shown to send an individual axon to the cerebral ganglia and produce activation of the Cr-A neurons (“tentacle motoneurons” according to Arshavsky et al., 1993). We have found that these excitatory inputs from the Bc-PIN neuron to the Cr-A neurons were apparently monosynaptic. Each induced Bc-PIN neuron spike produced an individual EPSP in the Cr-A neurons, which persisted without failure in high divalent solution (N=11; Fig.5B). Simultaneous recordings from the Cr-BM neuron, Bc-PIN neuron and the Cr-A neurons revealed that the Bc-PIN neuron was always spiking when the Cr-BM interneuron activated Cr-A neurons (N=14). The Cr-BM neuron-induced EPSPs in the Cr-A neurons appeared only when corresponding spikes were observed in the Bc-PIN interneurons (N=8; Fig.5C). In addition, when the Bc-PIN interneurons were hyperpolarized and thus functionally removed from the network functioning, the Cr-A neuron response to the stimulation of the CrBM interneuron completely disappeared (N=7). This provided strong evidence that Bc-PIN interneuron is indeed the element of the neural network that relays the excitatory inputs from the Cr-BM interneuron to the Cr-A neurons controlling opening of the skin folds and buccal cone protraction. The noticeable characteristic of all Cr-BM neuron-induced PSPs in the buccal neurons was their prominent facilitation with repetitive Cr-BM neuron firing (Fig.6A,B). Facilitation was frequency-dependent with the highest values recorded during the Cr-BM neuron spike frequency of 20 Hz and almost disappearing at 1 Hz frequency (Fig.6A,B,C). Figure 6A provides the quantitative assessment of this facilitation by comparing the ratio between the amplitude of the Coordination of Neural Networks 11 second and first EPSPs in the Bc-PIN neuron at different firing frequencies of the Cr-BM neuron induced burst. Another example is a comparison between the first and the maximum PSP in the train. Such an increase in the EPSP amplitude calculated for the Bc-PIN neuron during the CrBM neuron-induced firing at 3-8 Hz frequency was 502% +/59% (N=20). The Cr-BM interneurons are GABAergic GABA-immunoreactivity in the Clione central ganglia has been studied before (Norekian, 1999; Arshavsky et al., 1993). Our detailed analysis of GABA-immunoreactivity in the cerebral ganglia indicated that a pair of brightly stained immunoreactive cell bodies had the same position and body size as the Cr-BM interneurons. Each bilaterally symmetrical neuron with strong GABA immunoreactivity had body size of 25-30 μm and was located on the ventral side of the cerebral ganglia in the anterior region between head nerves N1 and N2 (N=12; Fig.7A). The following double-labeling experiments undoubtedly demonstrated that these GABA-immunoreactive neurons are Cr-BM interneurons (N=6). One Cr-BM interneuron in each preparation, from the left or the right cerebral ganglion, was injected with neurobiotin and visualized with Texas Red (Fig.7B). The same preparation was then processed for GABA immunoreactivity with the fluorescein-labeled secondary antibody (Fig.7C). By switching filters in the same preparation, Cr-BM interneuron was identified as GABA-immunoreactive neuron. Texas Red and fluorescein had very distinct emission wavelengths and were visible only in their own set of filters, thus allowing complete separation of images and clear interpretation of data. Not only the cell body, but also the entire morphological structure of the neurons could be seen through both filters and served as additional matching characteristic. This included a large axon entering the cerebro-buccal connective, and a group of thin processes innervating the neuropile of the ipsilateral cerebral ganglia (Fig.7B,C). Coordination of Neural Networks 12 Exogenous GABA mimicked the effects produced by the Cr-BM interneuron on all postsynaptic buccal cells. The local application of GABA via pressure ejection or iontophoresis produced depolarization of the radula protractor Bc-RP neurons (N=4; Fig.8A) and hyperpolarization of the radula retractor Bc-RR neurons (N=6; Fig.8B). The Bc-PIN interneuron also responded with a prominent depolarization to GABA application (N=9; Fig.8C). Locally applied GABA also produced depolarization of the hook retractor Bc-HR neurons (N=5; Fig.8D) and hyperpolarization of the hook protractor Bc-HP neurons (N=10; Fig.8E). There was no indication of hyperpolarization in the Bc-HR neuron, even though Cr-BM neuron produced in this cell a fast IPSP/ slow EPSP combination, which can be explained by the masking effect of more powerful depolarization or the presence of only one receptor type on the surface of somata. The GABA antagonist, bicuculline, at 1 mM concentration produced a prominent and reversible blocking effect on the Cr-BM neuron-induced IPSPs in the Bc-HP and Bc-RR neurons (N=14; Fig.9A). Picrotoxin at 1 mM concentration also blocked the Cr-BM neuron-induced IPSPs in the Bc-HP and Bc-RR neurons, although its blocking effect was less effective than effect produced by bicuculline (N=10). Fast IPSPs in the Bc-HR neurons were also blocked by bicuculline and picrotoxin, with slow depolarization remaining intact as the main response to the Cr-BM neuron stimulation (N=4; Fig.9B). Both bicuculline and picrotoxin at concentrations 1 mM produced a prominent blocking effect on GABA-induced hyperpolarizations in the Bc-HP and Bc-RR neurons indicating that they function as GABA antagonists (N=7; Fig.8E). Bicuculline also reversibly blocked the Cr-BM neuron-induced fast EPSPs in the Bc-PIN neurons (N=4; Fig.9C). This effect was partial and less efficient than bicuculline blockade of IPSPs. Piperidine-4-sulfonic acid in concentration 1 mM did not block the Cr-BM neuron-induced EPSPs in the Bc-PIN and Bc-RP neurons, as well as GABA-induced depolarization in these cells, suggesting the different type of GABA receptors in the buccal neurons than in the Cr-A neurons (N=5; Norekian 1999). Incubation Coordination of Neural Networks 13 for 2-5 minutes in the 0.5 mM GABA solution blocked the Cr-BM neuron-induced EPSPs in the Bc-PIN and Bc-RP neurons presumably via occlusion of GABAergic receptors (N=4). These experiments strongly supported the idea that Cr-BM interneurons are GABAergic cells. Additional postsynaptic effects of the Cr-BM interneurons The Cr-BM interneuron produced a widespread effect in the buccal ganglia, which was mostly centered on the activity of the neural networks controlling radula and hook movements. However, the buccal ganglia include neurons with other functional roles. One cell, designated Bc-L neuron, has been identified during this investigation as a putative motoneuron that controls closing of the skin folds over the mouth of Clione. The cell body of the bilaterally symmetrical Bc-L neuron was 20-25 μm in diameter and located in the posterior-medial part of each buccal ganglion (Fig.10A). A single Bc-L neuron axon exited the ipsilateral buccal ganglion via the short hook nerve and crossed the entire muscular hook sac without any branching (N=4; Fig.10A). Then the axon crossed the connective tissue that linked the buccal mass with the head and entered the skin fold region. Stimulation of the Bc-L neuron always produced a prominent closing of the skin folds in the semi-intact preparation (N=9). Induced burst of spikes in the CrBM neuron triggered a prominent inhibition of the otherwise spontaneously active Bc-L neurons (N=7; Fig.10B). Each individual Cr-BM neuron spike produced a single IPSP in the Bc-L neuron that persisted in high divalent solution suggesting monosynaptic connection (N=5; Fig.10C). Arguably the best-known cell involved in the control of feeding behavior in the gastropod molluscs is the cerebral giant serotonergic MCC neuron. We have found that the Cr-BM interneuron produced excitatory inputs to the MCC neurons (N=8; Fig.10D). These inputs were relatively weak and polysynaptic in nature. Coordination of Neural Networks 14 Discussion Coordination between three major feeding structures and underlying neural networks In our previous study, we have described the phase-locked coordination between rhythmic activity of the hooks and radula, which both extract the prey from its shell and bring it to the gut (Malyshev and Norekian 2002). Clione does not bite small pieces from the prey during feeding, but pulls the entire body of the prey from its shell, which occurs in 20 to 40 minutes (Wagner 1885; Lalli 1970; Lalli and Gilmer 1989). The radula and hooks take turns in pulling the prey out of the shell moving in opposite phases of their protraction-retraction cycle and keeping a constant extracting pressure (Malyshev and Norekian 2002). On the neuronal level, Bc-HP and Bc-RR neurons were always firing in phase with each other (Malyshev and Norekian 2002). The Bc-HR and Bc-RP neurons were also firing in phase, though there was a slight phase-shift in the Bc-HR neuron bursting, which allowed a brief co-activation between Bc-HR and Bc-HP neurons. In that investigation, we identified some of the neuronal mechanisms, which were responsible for the phase-locked coordination between hook and radula controlling motoneurons (Malyshev and Norekian 2002). The Bc-HP and Bc-RR neurons, which fire in one phase, were electrically coupled to each other, which explained their synchronous activity. The Bc-HR and Bc-RP neurons were not electrically coupled, which allowed for the observed Bc-HR phase-shift and co-activation of the hook retractor and protractor neurons. The second mechanism of the phasedependent coordination between hook and radula controlling neurons was reciprocal inhibitory connections found between different motoneurons. The Bc-HP neurons produced inhibitory inputs to the Bc-RP neurons briefly terminating their activity, while action potentials in the BcRP neurons, in turn, produced inhibition of the Bc-HP neurons (Malyshev and Norekian 2002). In the current investigation, we significantly expand our understanding of the level of coordination between different feeding structures in Clione and the neuronal mechanisms of this Coordination of Neural Networks 15 coordination. The higher-order Cr-BM interneuron produced strong excitatory inputs to the neural networks that control hooks and radula rhythmic movements initiating their activity in quiescent preparations. In addition to this overall excitatory influence, Cr-BM interneuron produced fine-tuned, phase-specific coordination of individual radula and hook controlling motoneurons. The Cr-BM interneuron hyperpolarized both Bc-RR and Bc-HP motoneurons, which are active in one phase of the feeding rhythm. At the same time, Cr-BM interneuron depolarized Bc-RP motoneurons, Bc-PIN interneuron and Bc-HR motoneurons, which are all active in the other phase. Moreover, two types of the Cr-BM neuron synaptic inputs to the BcHR motoneurons were responsible for or at least contributed to the phase-shift in the Bc-HR neuron bursting, which allowed a brief co-activation between Bc-HR and Bc-HP neurons. The initial Cr-BM neuron-induced fast IPSPs provided a delay in the Bc-HR neuron burst onset, while slow depolarization eventually over-rode the IPSPs and triggered a burst of activity, which lasted longer than Cr-BM neuron induced burst of spikes. This phase-shift in the Bc-HR neuron bursting created a brief period of co-activation of the Bc-HR and Bc-HP neurons, which is presumably important for producing a fast and powerful protraction movement of the hooks (Malyshev and Norekian, 2002). Co-activation of the functionally opposite motoneurons followed by the inhibition of one type of cells is a well-known mechanism for a fast and powerful movement (Heitler and Burrows, 1977; Norekian and Satterlie, 1993b). The previously described mechanisms of coordination between hooks and radula rhythmic movements were internal to their underlying neural networks: electrical coupling and reciprocal inhibition between hooks and radula controlling motoneurons (Malyshev and Norekian, 2002). The Cr-BM interneuron influence is the external mechanism of this phase-locked coordination. In addition to initiating and controlling in the phase-dependent manner rhythmic activity of the radula and hook motoneurons, the Cr-BM interneuron also produced strong excitatory inputs to a large Coordination of Neural Networks 16 group of electrically coupled Cr-A neurons that control opening of the skin folds and hydraulic protraction of the buccal cones. Opening of the skin folds and buccal cone protraction are important components of the feeding behavior, which are present during the entire episode of feeding and responsible for capturing and holding the prey. In addition, if the skin folds are not open and buccal cones are not protracted, the hooks and radula can not physically reach the prey. Thus, Cr-BM interneuron apparently represents an important higher-order element of the feeding neural system with extensive coordinating influence on different participating neural networks. Excitatory role of GABA in feeding behavior One important characteristic of the Cr-BM interneurons, which was uncovered during the current investigation, was their GABAergic nature. It was demonstrated by the use of GABA antagonists, mimicking effect of exogenous GABA on postsynaptic neurons, and double-labeling experiments, which showed that Cr-BM neurons are GABA-immunoreactive. GABA was originally known as inhibitory neurotransmitter. However, GABA-induced depolarizing effects were found in several neural systems (Gallagher et al. 1978; Ogata 1987; Pfeiffer-Linn and Glantz 1989; ElBeheiry and Puil 1990; Michelson and Wong 1991; Mercuri et al. 1991; Goldmakher and Moss 2000; Swensen et al. 2000; Beg and Jorgensen 2003). In some gastropod molluscs, GABA was found to function as an excitatory transmitter producing activation of the buccal feeding motor programs (Bravarenko et al. 2001; Richmond et al. 1993). An especially strong case for the excitatory role of GABA in the feeding system was made in Clione. Exogenous GABA produced activation not only of the buccal feeding neural network, but also strong activation of the cerebral Cr-A neurons that control prey capture response of the buccal cones (Arshavsky et al. 1993; Norekian and Satterlie 1993c). The excitatory GABAergic Cr-Aint interneuron was later identified as a key element for the explosive and lasting discharge activity in the prey capture neural network Coordination of Neural Networks 17 (Norekian 1999). The excitatory responses in the Cr-A neurons induced by GABA had reverse potentials above zero and were sodium dependent (Norekian 1999). In this study, we have identified as GABAergic another important element of the feeding neural system with strong excitatory influence over several neural networks. Thus it appears that GABA acts as a potent excitatory transmitter in the feeding system of Clione. Cerebro-buccal interneurons in gastropod molluscs In most gastropod molluscs, the typical feeding response includes biting movements of jaws and rhythmic movements of the toothed radula, which are controlled by the neural networks located in the buccal ganglia. In the several species of molluscs, a group of higher-order interneurons has been identified in the cerebral ganglia that project their axons to the buccal ganglia and influence activity of the feeding neural network. These cerebro-buccal interneurons have been identified in the carnivorous marine mollusc Pleurobranchaea (Gillette et al. 1982), the pond snail Lymnaea (McCrohan and Kyriakides 1989) and the land slug Limax (Delaney and Gelperin 1990). But the most detailed investigation of their functional role has been done in Aplysia, in which more than a dozen of the cerebro-buccal interneurons have been identified and studied (Rosen et al. 1991; Xin et al. 1999; Hurwitz et al. 1999; Perrins and Weiss 1998; Jing and Weiss 2001). Two of these cerebrobuccal interneurons have been identified as GABAergic: CBI-3 and CBI-11 neurons (Diaz-Rios et al. 1999; Jing et al. 2003; Wu et al. 2003). The CBI-11 interneuron produced a prominent excitatory effect on the buccal motor program and was capable of initiating feeding rhythm, while CBI-3 interneuron had regulatory influence (Wu et al. 2003; Jing et al. 2003). Identified in the current study Cr-BM interneuron in Clione is also a GABAergic cerebro-buccal interneuron. We believe that Cr-BM interneuron is homologous to the Aplysia GABAergic cerebro-buccal neurons, presumably to the CBI-11 neuron based on its morphology and physiological effect. Coordination of Neural Networks 18 The transmitter nature, morphology and general excitatory role in feeding behavior appear to be similar for these two presumably homologous cerebro-buccal interneurons, Cr-BM interneuron in Clione and CBI-11 interneuron in Aplysia. However, their specific connections and physiological effects are quite different, and reflect the differences in the feeding habits of the animals. High food specialization in the carnivorous Clione is reflected in the appearance of two unique feeding structures (Lalli and Gilmer 1989). First, the buccal mass in Clione includes not only the radula and radula-controlling muscles, but also two muscular hook sacs with numerous chitinous hooks that are used to pull the prey from its shell. And second, Clione acquired additional feeding structure for prey capture, buccal cones, which represent the protrusions of the buccal skin folds. These two new feeding structures are controlled by specific neural networks: hook controlling network in the buccal ganglia and a large group of buccal cone controlling Cr-A neurons in the cerebral ganglia. These feeding structures and underlying neural networks are absent in Aplysia. And the cerebro-buccal GABAergic Cr-BM interneuron in Clione carries a new functional role of activating not only the neural network that controls rhythmic movements of radula, but also activation and phase-dependent coordination of the hook controlling neural network and neural network that controls the movements of the prey capture appendages, buccal cones. Coordination of Neural Networks 19 Acknowledgements This work was supported by the National Science Foundation grant IBN-0235107, NATO Collaborative Linkage grant 979205, and a grant from the Russian Foundation for Basic Research. Coordination of Neural Networks20 ReferencesArshavsky, Y.I., Deliagina, T.G., Gamkrelidze, G.N., Orlovsky, G.N., Panchin, Y.V., Popova, L.B.and Shupliakov, O.V. Pharmacologically induced elements of the hunting and feeding behavior inthe pteropod mollusc, Clione limacina. I. Effects of GABA. J. Neurophysiol. 69: 512-5521, 1993. Beg, A.A. and Jorgensen, E.M. EXP-1 is an excitatory GABA-gated cation channel. Nat. Neurosci.6(11): 1145-1152, 2003. Bravarenko, N.I., Ierusalimsky, V.N., Korshunova, T.A., Malyshev, A.Y., Zakharov, I.S. andBalaban, P.M. Participation of GABA in establishing behavioral hierarchies in the terrestrial snail.Exp. Brain Res. 141(3): 340-348, 2001. Delaney, K. and Gelperin, A. Cerebral interneurons controlling fictive feeding in Limax maximus. J.Comp. Physiol. A 16: 297-326, 1990. Diaz-Rios, M., Suess, E. and Miller, M.W. Localization of GABA-like imunoreactivity in thecentral nervous system of Aplysia californica. J. Comp. Neurol. 413: 255-270, 1999. El-Beheiry, H. and Puil, E. Unusual features of GABA responses in layers IV-V neurons ofneocortex. Neurosci Letters 119: 83-85, 1990. Gallagher, J.P., Higashi, H. and Nishi, S. Characterization and ionic basis of GABA-induceddepolarizations recorded in vitro from cat primary afferent neurons. J. Physiol. 275: 263-282,1978. Coordination of Neural Networks21 Gillette, R., Kovac, M.P. and Davis, W.J. Control of feeding motor output by paracerebral neuronsin brain of Pleurobranchaea californica. J. Neurophysiol. 47: 885-908, 1982. Goldmakher, G.V. and Moss, R.L. A subset of periglomerular neurons in the rat accessory olfactorybulb may be excited by GABA through a Na(+)-dependent mechanism. Brain Res. 871: 7-15, 2000. Heitler, W.J. and Burrows, M. The locust jump. I. The motor program. J. Exp. Biol. 66: 203-219,1977. Hurwitz, I., Perrins, R., Xin, Y., Weiss, K.R. and Kupfermann, I. C-PR neuron of Aplysia hasdifferential effects on “feeding” cerebral interneurons, including myomodulin-positive CBI-12. J.Neurophysiol. 81: 521-534, 1999. Jing, J. and Weiss, K.R. Neural mechanisms of motor program switching in Aplysia. J. Neurosci.21: 7349-7362, 2001. Jing, J., Vilim, F.S., Wu, J., Park, J. and Weiss, K.R. Concerted GABAergic actions of Aplysiafeeding interneurons in motor program specification. J. Neurosci. 23: 5283-5294, 2003. Kosaka, T., Nagatsu, I., Wu, J.Y. and Hama, K. Use of high concentrations of glutaraldehydefor immunocytochemistry of transmitter-synthesizing enzymes in the central nervous system.Neuroscience 18: 975-990, 1986. Coordination of Neural Networks22 Lalli, C.M. Structure and function of the buccal apparatus of Clione limacina (Phipps) with areview of feeding in Gymnosomatous pteropods. J. exp. Mar. Biol. Ecol., 4: 101-118, 1970. Lalli, C.M. and Gilmer, R.W. Pelagic Snails. The Biology of Holoplanktonic Gastropod Mollusks.Stanford, Stanford Univ.Press, p.167-213, 1989. Malyshev, A.Y. and Norekian, T.P. Phase-Locked coordination between two rhythmically activefeeding structures in the carnivorous mollusc Clione limacina. I. Motoneurons. J. Neurophysiol. 87:2996-3005, 2002. McCrohan, C.R. and Kyriakides, M.A. Cerebral interneurones controlling feeding motor output inthe snail Lymnaea stagnalis. J. Exp. Biol. 147: 361-374, 1989. Mercuri, N.B., Calabresi, P., Stefani, A., Stratta, F. and Bernardi, G. GABA depolarizes neurons inthe rat striatum: an in vivo study. Synapse 8: 38-40, 1991. Michelson, H.B. and Wong, R.K.S. Excitatory synaptic responses mediated by GABAA receptorsin the hippocampus. Science 253: 1420-1423, 1991. Norekian, T.P. Prey capture phase of feeding behavior in the pteropod mollusc, Clione limacina:neuronal mechanisms. J. Comp. Physiol. A 177: 41-53, 1995. Norekian, T.P. GABAergic excitatory synapses and electrical coupling sustain prolonged dischargesin a prey capture neural network of Clione limacina. J. Neurosci. 19(5): 1863-1875, 1999. Coordination of Neural Networks23 Norekian, T.P. and Satterlie, R.A. Cerebral neurons underlying prey capture movements in thepteropod mollusc, Clione limacina. I.Physiology, morphology. J. Comp. Physiol. A 172:153-169,1993a. Norekian, T.P. and Satterlie, R.A. Co-activation of antagonistic motoneurons as a mechanism ofhigh-speed hydraulic inflation of prey capture appendages in the pteropod mollusk, Clionelimacina. Biol. Bull., 185: 240-247, 1993b. Norekian, T.P. and Satterlie, R.A. FMRFamide and GABA produce functionally opposite effects onprey-capture reactions in the pteropod mollusk, Clione limacina. Biol. Bull. 185: 248-262, 1993c. Ogata, N. γ-Aminobutyric acid (GABA) causes consistent depolarization of neurons in the guinea pig supraoptic nucleus due to an absence of GABAB recognition sites. Brain Res. 403: 225-233,1987. Perrins, R. and Weiss, K.R. Compartmentalization of information processing in an Aplysia feedingcircuit interneuron through membrane properties and synaptic interactions. J. Neurosci. 18: 3977-3989, 1998. Pfeiffer-Linn, C. and Glantz, R.M. Acetylcholine and GABA mediate opposing actions on neuronalchloride channels in crayfish. Science 245: 1249-1251, 1989. Coordination of Neural Networks24 Richmond, J.E., Murphy, A.D., Lukowiak, K. and Bulloch, A.G.M. GABA regulates the buccalmotor output of Helisoma by two pharmacologically distinct actions. J. Comp. Physiol. A 174: 593-600, 1993. Rosen, S.C., Teyke, T., Miller, M.W., Weiss, K.R. and Kupfermann, I. Identification andcharacterization of cerebral-to-buccal interneurons implicated in the control of motor programsassociated with feeding in Aplysia. J. Neurosci. 11: 3630-3655, 1991. Swensen, A.M., Golowasch, J., Christie, A.E., Coleman, M.J., Nusbaum, M.P. and Marder, E.GABA and responses to GABA in the stomatogastric ganglion of the crab Cancer borealis. J. Exp.Biol. 14: 2075-2092, 2000. Wagner, N. (1885). Die Wirbellosen des Weissen Meeres: Zoologische Forschungen an der Kustedes Solowetzkischen Meerbusens in den Sommermonaten der Jahre. Leipzig, Verlag Von WilhelmEngelmann, pp.168. Wu, J., Jing, J., Diaz-Rios, M., Miller, M.W., Kupfermann, I. and Weiss, K.R. Identification of aGABA-containing cerebral-buccal interneuron-11 in Aplysia californica. Neurosci. Letters 341: 5-8,2003. Xin, Y., Hurwitz, I., Perrins, R., Evans, C.G., Alexeeva, V., Weiss, K.R. and Kupfermann, I.Actions of a pair of identified cerebral-buccal interneurons (CBI-8/9) in Aplysia that contain thepeptide myomodulin. J. Neurophysiol. 81: 507-520, 1999. Coordination of Neural Networks25