1 Institute for Zoology and Anthropology, University of Göttingen, ... 3 Department of Neurobiology, Harvard Medical School, 220 Longwood Ave., Boston, MA.
Synaptic connectivity of amine-containing neurosecretory cells of lobsters: Inputs to 5HT- and OCT- containing neurons Michael Hörner1, Ralf Heinrich1, Stuart I. Cromarty2 and Edward A. Kravitz3 1
Institute for Zoology and Anthropology, University of Göttingen, Berlinerstrasse. 28, D-37073 Göttingen, Germany 2 Division of Natural Sciences, Assumption College, Worcester, MA 01615, USA 3 Department of Neurobiology, Harvard Medical School, 220 Longwood Ave., Boston, MA 02115, USA Introduction Biogenic amines, such as serotonin (5HT) and octopamine (OCT), the invertebrate counterpart of norepinephrine in vertebrates, function as neuromodulators capable of influencing vital behaviors in a wide variety of animal species (Hen, 1992; Cases, 1995; Weiger, 1997; Edwards and Kravitz, 1997; Shiah and Yatham, 1998; Hörner, 1999a; Monastirioti, 1999; Kravitz, 2000). Amine actions have been examined at organizational levels ranging from behavioral, through systems, and ultimately to cellular and subcellular levels. Invertebrate preparations are used extensively for such studies because of their unique advantages: with these models the analyses can be brought to the level of identified neurons likely to be important in the behavior. In crustaceans, for example, individual identified 5HT and OCT neurons are readily accessible for electrophysiological, pharmacological and even molecular studies aimed at defining their function in complex behaviors like aggression. In lobsters the biogenic amines 5HT and OCT appear to be involved in the modulation of agonistic behavior (Livingstone et al., 1980; Kravitz, 1988; Huber et al., 1997). When injected into the hemolymph, 5HT or OCT provoke opposing flexed or extended postures resembling those seen in dominant and subordinate animals. These postural differences result from 5HT and OCT having opposite actions on central command systems concerned with postural regulation even though at the same time they have similar facilitating effects on peripheral neuromuscular targets (Harris-Warrick and Kravitz, 1984). 5HT injections also influence how long losing lobsters and crayfish are willing to fight (Huber et al., 1997). As described in the previous article (Heinrich et al., this volume), amine neuron systems have been completely mapped in the lobster nervous system (Beltz and Kravitz, 1983; Schneider et al., 1993; Cournil et al., 1994). Besides other amine-containing cells, the focus of these studies has been on the neurosecretory subsets of both the 5HT- and the OCT- containing neurons. With peripheral and central release sites, these cells appear capable of exerting widespread actions in lobsters, some of which are potentially related to their roles in complex behaviors like aggression. The intrinsic properties of these neurons were presented in the previous article. Here we review our studies on the search for presynaptic elements capable of influencing the activity of these neurosecretory neurons. In beginning this search, we relied on the findings of earlier immunocytochemical studies (Beltz and Kravitz, 1983; Beltz and Kravitz, 1987; Schneider et al., 1993) showing that amine-containing neurosecretory neurons have wide ramifications of processes and endings throughout the central and peripheral nervous systems of lobsters. These ramifications often overlapped with each other and with particular intersegmental fibers. In this article we present new data and also review our previous data on the synaptic connections to and between identified 5HT- and OCT- containing neurosecretory neurons in the thoracic and abdominal nerve cord of the lobster, Homarus americanus. Testing for synaptic inputs to neurosecretory amine-containing neurons -1-
The methods employed for electrophysiology have been described in detail elsewhere (Hörner et al., 1997; Fig.1). Briefly, in a typical experiment, the ganglia of thoracic segments 1-5 (T1-T5) and the complete abdominal ganglionic chain (A1-A6) of adult lobsters of both sexes were removed, pinned out in a sylgard coated recording chamber, and maintained under constant superfusion of oxygenated saline to which various drugs could be added. To expose the somata of the 5HT- and OCT- containing neurosecretory neurons, the outer sheath of the A1 and/or of various thoracic ganglia were removed by dissection with fine scissors. 5HTand OCT- containing neurosecretory neurons, motorneurons and giant interneurons were identified using criteria defined in earlier publications (Otsuka et al., 1967; Beltz and Kravitz, 1983; Ma et al., 1992; Schneider et al., 1993; Hörner et al., 1997). Standard electrophysiological techniques were used for extra- and intracellular recordings, stimulation of nerve roots, connectives and single axons, and for intracellular fluorescent dye injection (c.f., Hörner et al., 1997). Inputs to neurosecretory amine-containing neurons from intersegmental sources 5HT-neurosecretory neurons: As described in the previous paper, 2 pairs of 5HTneurosecretory cells, one in the 5th thoracic (T5-5HT neurons) and another in the A1 ganglion (A1-5HT cells) make up the serotonergic neurosecretory neuron system in the lobster ventral nerve cord (Beltz and Kravitz, 1983; Beltz and Kravitz, 1987). These interneurons are characterized by their wide ranging intersegmental projections ascending to the subesophageal ganglion (SEG) while giving off central branches and peripheral neurohemal endings on the surface of thoracic 2nd nerve roots in each anterior ganglion along the way (Beltz and Kravitz, 1983; Beltz, 1999). The A1- and T5-5HT neurons influence central circuitries concerned with posture (Livingstone et al., 1981; Harris-Warrick and Kravitz, 1984; Harris-Warrick, 1985; Ma et al., 1992) and peripheral targets likely related to postural control mechanisms (muscles: Kravitz, 1988; sensory neurons: Pasztor and Bush, 1987; heart: Batelle and Kravitz, 1978). The A1-5HT cells are reliably identified by their spontaneous spiking activity and the amplitude and shape of the action potential, which shows a pronounced hyperpolarizing afterpotential (Beltz and Kravitz, 1987; Ma et al., 1992; Hörner et al., 1997; Heinrich et al., 1999). Their endogeneous firing level is regulated by tonic inhibitory input, mainly coming from the A3 ganglion and to a lesser extent by unidentified excitatory synaptic inputs (Ma and Weiger, 1993; Weiger and Ma, 1993). In their role as “gain setters” modulating the output of postural command circuitries we observed that flexor command neurons excite the A1-5HT neurons whereas extensor commands inhibit their firing (Ma et al., 1992). The A1-5HT neurons also are excited by electrical stimulation of peripheral nerve roots that contain axons of primary sensory afferents. Such stimuli are likely to activate mechanosensory interneurons, which in turn will generate postural flexion reflexes and excite abdominal postural command systems (Kotak and Page 1987; Weiger and Kravitz 1994). OCT-neurosecretory neurons: From a total of 86 immunocytochemically identified OCT-containing neurons in the lobster ventral nerve cord, the neurosecretory subset comprises a group of 14 pairs of cells (Schneider et al., 1993). These mostly intersegmentally projecting OCT-containing interneurons are located in the anterior and posterior regions of all thoracic ganglia and in the posterior neuromeres of the SEG. As with the 5HT-neurosecretory neurons, the OCT-containing neurons also have sets of central and peripheral endings (Schneider et al., 1993) which can be further separated into two groups. One group of 3 pairs of OCT neurons (T3-anterior, T4-anterior, T4-posterior) have central endings and a set of peripheral endings that project selectively to nerves innervating the claws (claw octopamine cells: CLOCs; Heinrich et al., 2000). All the other pairs of OCT-neurosecretory neurons have central sets of endings and peripheral endings on 2nd thoracic roots (root octopamine cells: ROCs), comparable to those seen with the 5HT-neurosecretory neurons. The projections from -2-
the posterior pairs of ROCs in each ganglion are found within the 2nd roots of the same ganglion , while the anterior pairs of ROCs send ascending processes through the ipsilateral paramedial nerve running alongside the connectives between ganglia. These projections enter the 2nd nerve roots of the next anterior ganglia (Schneider et al., 1993). In contrast to the detailed knowledge available on 5HT neurons, less is known about the physiological properties of OCT-containing cells in the lobster central nervous system (CNS). The majority of the OCT-containing neurosecretory cells recorded from show no spontaneous spiking, but action potentials of large amplitude (40-55 mV) with hyperpolarizing afterpotentials typical of neurosecretory neurons usually can be elicited by intracellular current injection (Bräunig et al., 1994; Heinrich et al., 2000). In contrast to the constant barrage of synaptic input observed in A1-5HT neurons, few spontaneous synaptic events are observed in soma recordings from OCT cells (Bräunig et al., 1994; Heinrich et al., 2000). Despite their different branching patterns and their probably serving distinct functions, no striking differences are seen in the electrophysiological properties of ROCs and CLOCs (Heinrich et al., 2000). Sources of synaptic input to 5HT- and OCT- containing neurosecretory neurons: In our first attempts to identify possible sources of synaptic input to the A1-5HT and to the OCT cells, connectives and segmental nerve roots were stimulated. The largest and most reliable excitatory effect seen in A1-5HT cells was generated by stimulating the sensory nerve roots of the terminal (A6) abdominal ganglion. A single shock to the C nerve of the A6 ganglion evoked a compound EPSP that led to subsequent firing in A1-5HT cells (Hörner et al., 1997). In attempting to track down the interneurons mediating this excitation, we found that extracellular stimulation of large axons on the dorsal surface of connectives elicited shortlatency EPSPs in the 5HT cells, resembling those triggered by C root stimulation. Since the lateral (LG) and medial (MG) axons are amongst the largest of the axons within the ventral nerve cord, we asked whether LG or MG axons might provide synaptic input to A1-5HT cells. In the OCT neurons, similar stimulation of connectives and nerve roots evoked complex responses that could include EPSPs capable of triggering subsequent spiking. In these cells, however, no giant fiber related short-latency EPSPs were observed comparable to those seen in the A1-5HT cells (see below). Identified inputs to 5HT- but not OCT- containing neurons from the giant fiber system Serotonergic neurons: To further investigate whether A1-5HT cells received excitatory inputs from giant fiber pathways, we used double intracellular recording and stimulation techniques (Fig.1). It was necessary to go this route because extracellular stimulation of connectives did not selectively and reliably fire individual LG or MG axons. Intracellular stimulation of LG or MG axons triggered unitary EPSPs (Fig.2A, B) in ipsilateral, but not in contralateral A1-5HT cells. The LG spikes triggered by intracellular stimulation propagated along the ipsilateral chain of LG axons in either direction at 8.5-9ms-1. We observed similar EPSPs followed by action potentials in A1-5HT cells 6-7 ms after LG spikes were evoked either through intracellular stimulation at the level of the T5 ganglion or extracellular stimulation at the level of the A4 ganglion (see Fig.1). The EPSPs showed a steep rising phase and then a plateau lasting for at least 100 ms that in some cases did not decline to the baseline for at least another 100 ms (Fig.2A, B). EPSPs showed little variation in amplitude or slope during the rising phase, but during the plateau considerable variability was seen in EPSP amplitude and duration. Despite this variability the LG-evoked EPSPs appeared to be unitary events. Moreover, both the LG and MG spikes and the EPSPs evoked in the A1-5HT cells followed short trains of intracellular stimulation at frequencies of up to 100Hz without signs of fatigue or potentiation. Action potentials often arose from the EPSPs triggered by single LG spikes. -3-
To compare the variability of LG-evoked EPSPs in A1-5HT neurons, with that seen in the motorneuron targets of the LGs, evoked EPSPs were analyzed in fast flexor motorneurons and 5HT cells from the same ganglion (Fig.3A). The LG spikes generated EPSPs in motorneurons that showed steep rise times (half maximum at 3.5-4.5ms) and considerably larger amplitudes than those found in 5HT cells (range = 2-6mV). Most remarkably, the duration of LG and MG-evoked EPSPs seen in motorneurons (approximately 30 ms in duration) only were a fraction of the duration of the giant fiber-evoked EPSPs in A1-5HT cells. LG stimulation also caused spiking in some fast flexor motorneurons as indicated by the extracellularly recorded action potentials from the 3rd nerve roots (lowest trace Fig.3B). The strongest evidence that the synaptic inputs from the LG and MG to the A1-5HT neurons were excitatory came from 'reset' experiments carried out in spontaneously active A15HT cells. Single spikes were elicited in LG axons at random intervals following spontaneously occuring 5HT cell action potentials. The LG effects were quantified by comparing the interspike intervals between a spontaneous action potential and the LG-evoked potential (n) and between the LG-evoked potential and the next spontaneous event (n+1; Fig.4A). Phase response curves (Fig.4B) show the interspike intervals between the LGevoked spikes and the immediately preceding and following spontaneous action potentials. These interspike intervals then were compared to the mean spontaneous spike interval. The results show that single spikes triggered by LG activation can modify the spontaneous firing rhythm in A1-5HT cells. The greatest effects were observed when giant fiber-evoked EPSPs occurred within the first third of the spontaneous firing cycle (Fig.4C). Thus, the effects of driving LG were to: (i) shorten the interspike interval between a prior spontaneous action potential and the LG-generated spike (n); (ii) lengthen the interspike interval between the LGmediated spike and the next spontaneous action potential (n+1). Action potentials evoked by MG stimulation had similar effects on the firing rhythms of A1-5HT neurons (data not shown) and were additionally capable of resetting the spontaneous firing rhythm in A1-5HT cells (Hörner et al., 1997). Octopaminergic neurons: As mentioned above, EPSPs also could be evoked in thoracic OCT-containing neurons by stimulation of connectives (Fig.2C). For example, in the experiment illustrated in Fig.2C and D, electrical stimulation of the dorsal surface of the connectives between the A3 and A4 ganglion with a small tipped suction electrode, evoked a rapidly conducting large amplitude spike, which was recorded extracellularly at T3 (Fig.2C, D). The large extracellular spike, most likely originated from a giant fiber due to its fast conduction speed, and was correlated in Fig.2C with the appearance of a large EPSP in OCTcontaining neurons ipsilateral to the stimulated connective. However, after decreasing the stimulus intensity, the EPSP in the OCT cell disappeared while the extracellular spike persisted (Fig.2D). Thusfar, all EPSPs occurring in OCT cells after connective stimulation were of long latency and were not correlated with firing of either the MG or LG. This result was confirmed in spontaneously active OCT cells, where no giant fiber related effect on the firing rhythm could be observed, indicating the absence of synaptic connections between the giant fibers and the OCT-containing neurons. Thus, the two neurosecretory amine-containing neuron cell types seem to be differentially coupled to the giant fiber pathway: only the 5HT-containing neurosecretory neurons receive excitatory synaptic input from activation of the MG and LG axons, while the OCT-containing neurosecretory neurons appear to be coupled to non-giant pathways. Another difference between the 5HT- and OCT- containing neurosecretory neurons is the apparent absence of spontaneous EPSPs and IPSPs in the OCT cells, while both 5HT- and OCT cells receive synaptic input through connective stimulation. Synaptic coupling between amine-containing neurons
-4-
Serotonergic neurons: In experiments in which both A1-5HT neurons were penetrated with intracellular electrodes, the two cells always showed different spontaneous firing patterns with their action potentials out of phase. Even high frequency firing that sometimes occurred spontaneously, or that could be evoked through current injection into one of the cells, did not affect the spontaneous firing pattern of the other cell in the same ganglion (Fig.5A). Thus no evidence exists for direct chemical or electrical coupling between pairs of A1-5HT cells (Ma and Weiger, 1993; Weiger and Ma, 1993). The cells, however, do share some important inhibitory synaptic input from the 3rd abdominal ganglion (Weiger and Ma, 1993). Moreover, if picrotoxin (PTX) is used to block inhibitory GABAergic inputs, the firing frequencies gradually increase in both A1-5HT neurons, until both cells start bursting in a phase-coordinated fashion. These effects can be seen in pairs of A1-5HT cells recorded from preparations either with the entire nerve cord left intact (Fig.5B) or in isolated A1 ganglia (Fig.5D). The coupling of spiking activity occurring under PTX perfusion and also the Ca2+dependent hyperpolarizing afterpotential (Heinrich et al., 1999) typically seen in A1-5HT neurons is reversibly blocked by lowering the Ca2+ concentration in the saline superfusion solution (Fig.5C, D). These data suggest that a shared excitatory synaptic input exists within the A1 ganglion linked to the two A1-5HT cells and possibly helping to coordinate their firing patterns. Octopaminergic neurons: Comparable double-recording experiments could not easily be performed on OCT cells, because of the paucity of recordings in which neurons were spontaneously active. We could, however, trigger trains of action potentials in OCT neurons through stimulation of the paramedial nerves containing branches of these cells (Bräunig et al., 1994; Heinrich et al., 2000). Under these conditions, action potentials were evoked only in OCT cells ipsilateral to the stimulation site. In other experiments, intraganglionic pairs of OCT neurons, anterior and posterior OCT neurons and interganglionic pairs were tested using bidirectional current passing, and firing of single or trains of action potentials. Thusfar, we have found no suggestion that direct synaptic coupling exists among the OCT-containing neurosecretory cells. Possible interactions between OCT and 5HT neurons: In earlier studies we demonstrated that superfusion of OCT has an inhibitory effect on the spontaneous firing rhythm of A1-5HT neurons (Ma and Weiger, 1992). Therefore, we next searched for possible direct synaptic connections between 5HT- and OCT- containing neurons. In a series of experiments we recorded from OCT cells from the T3anterior, T4anterior, T4posterior, and T5anterior groups and simultaneously from T5- and/or A1-5HT cells. As above, neither bidirectional passing of current nor current-injection triggering trains of action potentials revealed any evidence of synaptic coupling between 5HT- and OCT- containing neurons (Heinrich et al., 2000). Differential synaptic activation of amine-containing neurons In addition to searching for evidence of direct synaptic coupling between OCT- and 5HT- containing neurons, we asked whether common inputs to these cells might originate from intersegmental sources. As already mentioned, stimulation of giant fibers elicited EPSPs in A1-5HT cells ipsilateral to the stimulated axons, while OCT cells received no excitatory or inhibitory synaptic input via these sources. Stimulation of non-giant interneurons, however, did trigger EPSPs in OCT cells. By contrast, in pairs of A1-5HT cells stimulation of non-giant interneurons regularly caused long-lasting inhibitory effects. This can be seen in Fig.6 in simultaneous recordings from both A1-5HT cells. These findings suggest that intersegmental non-giant pathways may act differentially on the two groups of amine neurons, causing inhibition of 5HT neurons and simultaneous excitation of OCT neurons. The experiment shown in Fig.6 demonstrates that in the same preparation, OCT- and 5HT- containing neurons receive synaptic input of opposite sign under identical stimulation -5-
conditions: both 5HT cells reproducibly showed inhibition, while an EPSP was observed in the single T4 CLOC cell we recorded from. The intersegmental sources mediating these effects have not yet been identified. The observations, however, are compatible with the assumption of differential control of the activity levels of 5HT- and OCT- containing neurosecretory neurons mediated via synaptic sources. That synaptic sources coordinate the firing of these cells is indicated by other preliminary experiments we have carried out. For example, electrical stimulation of one connective elicits action potentials in both OCT partner cells within a T4 ganglion (Fig.7A, B) and also in OCT cells of other thoracic ganglia (data not shown). Interestingly, an identical spiking pattern was observed no matter whether the stimulation site was anterior or posterior to the neurons we recorded from (Fig.7) suggesting that sets of OCT cells might be coactivated as a group by common pathways. Discussion and Conclusions General physiological features of neurosecretory amine neurons In this article we summarize the presently available data on the synaptic circuitry of the identified 5HT- and OCT- containing neurosecretory neurons in the lobster ventral nerve cord. The small number of amine-containing neurons, typically seen in other arthropods as well (Hörner, 1999a, b), has allowed for a fairly complete reconstruction of the central and peripheral projection fields of the two amines. Comparative studies demonstrate a considerable overlap of the central arborizations of 5HT- and OCT- containing neurons. Thus, the two amines could exert differential effects on underlying circuitries, which in turn could account for the opposing effects seen on behavior. Many electrophysiological, pharmacological and behavioral studies support the view that the neurosecretory amine neuron systems likely modulate the opposing body postures seen in dominant and subordinate lobsters (Fig.8). This was the starting point for our search for synaptic interactions between the aminergic neurosecretory neurons and the intersegmental pathways within the ventral nerve cord of the lobster they connect with. Our experiments provide evidence for the existence of intra- and intersegmental pathways capable of influencing amine-containing neurosecretory cells. Both 5HT and OCT cells receive input from intersegmental sources, as anticipated from the observed anatomical overlap between amine cell processes and the projections from intersegmental fibers within thoracic ganglia. Despite their overlapping central projections, however, no direct synaptic interactions were seen among pairs of 5HT- or OCT- neurosecretory neurons within a single ganglion or between 5HT and OCT cells within the same or adjacent ganglia. Instead, the results suggest that the activity of amine neurons may be orchestrated by intra- and intersegmental synaptic inputs that influence 5HT and OCT neurons in opposite directions, while driving groups of neurons of the same amine type in similar directions. Shared physiological properties of 5HT- and OCT- containing neurosecretory neurons Our immunocytochemical studies show that 5HT (Beltz and Kravitz, 1983; Beltz, 1999) and OCT (Schneider et al., 1993) neurons in the lobster CNS are morphologically distinct entities. A co-localization of 5HT and OCT within the same neuron has never been seen in the lobster CNS, and also -even though occasionally observed- seems to be a rare exception in other arthropods (Hörner, 1999a, b). At the electrophysiological level, however, a number of common features are seen in amine neurons. First, large overshooting action potentials with pronounced afterhyperpolarizations occur in both 5HT- and OCT- containing neurons. Spontaneous firing is seen more commonly in the 5HT neurons, but it is observed occasionally in OCT cells, suggesting that they retain the capacity for endogenous spike -6-
generation. Both types of neurons also show “autoinhibition” after periods of high frequency firing (Heinrich et al., 1999; Heinrich et al., 2000; Heinrich et al., this volume), which appears to be a general property of neurosecretory neurons in lobsters. Not only is the generation of action potentials suppressed during periods of autoinhibition, but slow excitatory and inhibitory synaptic inputs to amine neurons also are significantly reduced by high frequency firing of the cells. This was first shown for giant fiber-evoked EPSPs in A1-5HT cells (Hörner et al., 1997; Heinrich et al., 1999) but recently has been confirmed for OCT-containing interneurons as well (Heinrich et al., this volume). Finally, an inverse relationship exists between the initial rate of firing of both types of amine neurosecretory neurons and the duration of the period of autoinhibition following high frequency firing of the cells. The factors important in setting the firing rate of cells remain relatively unknown, but prior patterns of usage, like high frequency firing, might be important in this regard. For the serotonergic neurons, another possible way of influencing firing rates would be through a reduction in spontaneous inhibitory input from the A3 ganglion or from other sources. In insects, inhibitory mechanisms appear to play an important role in suppressing behavior (Huber, 1960; Kien, 1983; Sombati and Hoyle, 1984; Heinrich et al., 1998). In lobsters, blocking inhibition increase the rates of firing of A1-5HT neurons (Weiger and Ma, 1993) and the studies described above revealed the existence of a shared excitatory synaptic input that led to high frequency firing and bursting of A1-5HT cells when inhibition was blocked for a prolonged period of time. Such excitatory influences would become more important if reduced firing of inhibitory elements were a normal physiological mechanism involved in activating serotonergic neurons. Although we see no spontaneous inhibitory activity in OCT neurons, shared excitatory synaptic input could coordinate the firing of the entire group of OCT cells. Although not directly coupled to each other, the OCT-neurosecretory neurons were simultaneously activated above the spiking threshold in a phase-coupled manner by nongiant pathways of unknown origin (Fig.7). In insects the OCT-containing DUM cells have been shown to be synaptically coupled to common descending pathways leading to a concerted action of groups of OCT cells in specific behavioral contexts (Gras et al., 1990; Burrows and Pflüger, 1995; Baudoux et al., 1998; Hörner, 1999b). Differential activation of 5HT- and OCT- containing neurosecretory neurons We have sought, but not yet found opposing direct synaptic interactions between 5HT and OCT neurons (Ma and Weiger, 1993). This despite the fact that both 5HT and OCT inhibit the firing of A1-5HT cells. We have not yet tested the sensitivity of OCT cells to the two amines. Moreover, thusfar we have recorded only excitatory synaptic input in OCT cells, but it seems unlikely that OCT neurons are devoid of inhibitory influences. Weak inhibitory influences are difficult to detect in the non-spiking OCT neurons and IPSPs may not be observable within the somata of OCT neurons under our recording conditions. Thus it remains possible that firing of serotonergic neurons could exert weak inhibitory influences on OCT cells and we have not been able to measure this yet. Our evidence is stronger that there is opposition in the activation of 5HT and OCT neurons via synaptic sources. This comes from experiments in which central pathways were activated by electrical stimulation of connectives. Here we find that activation of intersegmental pathways can trigger synaptic inputs of opposite sign to 5HT- and OCTcontaining neurons (Fig.6). Moreover, excitatory input from the giant fiber pathways involved in triggering escape behavior, seem to be exclusively channeled to 5HT neurons. OCTcontaining cells receive no inputs from giant fibers and thus do not appear to be coupled to this important fast neuronal circuit. This further supports the view that sets of amine neurons are activated by different intersegmental pathways, with the proviso mentioned above, that inhibitory inputs to OCT neurons may not be detectable in our recordings.
-7-
Thusfar, the MG and LG axons are the only presynaptic elements to an amine neuron in the lobster CNS that have been positively identified (Fig.8). The rapid onset and long duration of the giant fiber-evoked EPSPs in A1-5HT cells sets the membrane potential of these cells at values close to the threshold for triggering spikes, holding it there for long periods of time. Even strong inhibition of A1-5HT neurons 'primed' in this way probably would not abolish giant fiber-evoked excitation but only modulate its duration. The input from the LG to A1-5HT cells has another interesting property. Yeh et al., (1996, 1997) have shown that in the last abdominal ganglion (A6) synaptic interactions between mechanosensitive afferents and the LG are modulated by 5HT in a manner determined by the social status of the animal. 5HT facilitates these synaptic contacts in dominant animals and suppresses them in subordinate animals (Yeh et al., 1996, 1997). Thus the interesting possibility is raised that that the giant fiber pathway functions more effectively in dominant than in subordinate animals. This could play some role in the behavioral differences seen in animals of opposite social status. In following up on these observations, it will be interesting to ask whether the synaptic coupling strength between giant axons and the A1-5HT or other amine cells changes along with changes in the behavioral status of animals. In summary our data suggest that the activity of amine-containing neurosecretory cells is controlled by multiple, complex synaptic mechanisms involving intrasegmental, intersegmental and sensory-motor pathways (Fig.8). Intersegmental pathways differentially influence the activity of 5HT and OCT neurons, while homologous sets of amine cells appear to function as units driven to similar levels of activity by both intra- and intersegmental pathways. In addition, the firing of amine cells appears to be under the control of sophisticated feedback mechanisms like autoinhibition and feed forward mechanisms like the “gain-setter” circuitry. All the experimental evidence we have gathered thusfar, including morphological, physiological and behavioral studies, supports the suggestion that aminergic neurons specifically modulate central and peripheral targets relevant to behavioral functions like aggression. They thus may serve as the interfaces between hormonal and neuronal functions that typically are closely linked in arthropods.
-8-
References Batelle, B.A. and Kravitz, E.A. (1978). Targets of octopamine action in the lobster: cyclic nucleotide changes and physiological effects in hemolymph, heart and exoskeletal muscle. J. Pharmacol. Exp. Ther. 205:38-481. Baudoux, S., Duch, C. and Morris, O.T. (1998). Coupling of efferent neuromodulatory neurons to rhythmical leg motor activity in the locust. J. Neurophysiol. 79:361-370. Beltz, B.S. (1999). Distribution and functional anatomy of amine-containing neurons in decapod crustaceans. in: The cellular distribution of histamine and other biogenic amines in the nervous system of arthropods. (eds J.E. Johnson, M. Hörner) Special issue of the journal Microscopy Research and Technique, Wiley and Sons, New York, Chichester, Vol 44 (2/3), p 105-120. Beltz, B.S. and Kravitz, E.A. (1983). Mapping of serotonin-like immunoreactivity in the lobster nervous system. J. Neurosci. 3:585-602. Beltz, B.S. and Kravitz, E.A. (1987). Physiological identification, morphological analysis, and development of identified serotonin-proctolin-containing neurons in the lobster nervous cord. J. Neurosci. 7:533546. Bräunig, P., Walter, H., Schneider, H. and Kravitz, E. A. (1994). A subdivision of octopamine neurosecretory cells in the lobster. Proc. Soc. Neurosci. 526.12, p1282. Burrows, M. and Pflüger, H.-J. (1995). Action of locust neuromodulatory neurons is coupled to specific motor patterns. J. Neurophysiol. 74:347-357. Cases, O., Seif, I., Grimsby, J., Gaspar, P. Chen, K., Pournin, S., Müller, U., Aguet, M. Babinet, C., Chen Shih, J. and De Maeyer, E. (1995). Aggressive behavior and altered amounts of brain serotonin and norepinephrine in mice lacking MAOA. Science 268:1763-1766. Cournil, I., Helluy, S.M., Beltz, B.S. (1994). Dopamine in the lobster Homarus gammarus. I. Comparative analysis of dopamine- and tyrosine hydroxylase immuno-reactivities in the nervous system of the juvenile. J. Comp. Neurol. 344:455-469. Edwards, D.H. and Kravitz, E.A. (1997). Serotonin, social status and aggression. Curr. Opin. Neurobiol. 7:812819. Gras, H., Hörner, M., Runge, L. and Schürmann, F.W.(1990). Prothoracic DUM-neurons of the cricket Gryllus bimaculatus - responses to natural stimuli and activity in walking behavior. J. Comp. Physiol. A 166:901914. Harris-Warrick, R.M. (1985). Amine modulation of extension command element-evoked motor activity in the lobster abdomen. J. Comp. Physiol. A 156:875-884. Harris-Warrick, R.M. and Kravitz, E.A. (1984). Cellular mechanisms for modulation of posture by octopamine and serotonin in the lobster. J. Neurosci. 4:1976-1993. Hen, R. (1992). Of mice and flies: commonalities among 5-HT receptors. Trends in Pharmacology 13:160165. Heinrich, R., Bräunig, P., Walter, I., Schneider, H. and Kravitz, E.A. (2000). Aminergic neuron systems of lobsters: morphology and physiology of octopamine-containing neurosecretory cells. J. Comp. Physiol. A in press. Heinrich, R., Cromarty, S. I., Hörner, M., Edwards, D.H. and Kravitz, E.A. (1999). Autoinhibition of serotonin cells: An intrinsic regulatory mechanism sensitive to the pattern of usage of the cells. Proc. Natl. Acad. Sci. USA 96:2473-2478. Heinrich, R., Hörner, M., Cromarty, S. I., and Kravitz, E.A. (this volume). Intrinsic properties of aminecontaining neurosecretory cells of lobsters: spontaneous activity and autoinhibition. Heinrich, R., Rozwod, K. and Elsner, N. (1998). Neuropharmacological evidence for inhibitory cephalic control mechanisms of stridulatory behavior in grasshoppers. J. Comp. Physiol. A 183:389-399. Hörner, M. (1999a). Introduction to the cellular distribution of histamine and other biogenic amines in the nervous system of arthropods. in: The cellular distribution of histamine and other biogenic amines in the nervous system of arthropods. (eds J.E. Johnson, M. Hörner) Special issue of the journal Microscopy Research and Technique, Wiley and Sons, New York, Chichester, Vol 44 (2/3), p 69. Hörner, M. (1999b). The cytoarchitecture of histamine-, dopamine- and serotonin-containing neurons in the cricket ventral nerve cord. in: The cellular distribution of histamine and other biogenic amines in the nervous system of arthropods. (eds J.E. Johnson, M. Hörner) Special issue of the journal Microscopy Research and Technique, Wiley and Sons, New York, Chichester, Vol 44 (2/3), pp 137-166. Hörner, M., Weiger, W.A., Edwards, H.D. and Kravitz, E.A. (1997). Excitation of identified serotonergic neurons by escape command neurons in lobsters. J. Exp. Biol. 200(14):2017-2033. Huber, F. (1960). Untersuchungen über die Funktion des Zentralnervensystems und insbesondere des Gehirnes bei der Fortbewegung und der Lauterzeugung der Grillen. Z. vergl. Physiologie 44:60-132.
-9-
Huber, R., Smith, K., Delago, A., Isakson, K. and Kravitz, E.A. (1997). Serotonin and aggressive motivation in crustaceans: altering the decision to retreat. Proc. Natl. Acad. Sci. USA 94:5939-5942. Kien, J. (1983). The initiation and maintenance of walking in the locust: an alternative to the command concept. Proc. R. Soc. Lond. B219:137-174. Kotak, V.C. and Page, C.H. (1987). Synaptic responses produced in lobster abdominal postural motor neurons by mechanical stimulation of the swimmeret. J. Comp. Physiol. 161:695-703. Kravitz, E.A. (1988). Hormonal control of behavior: amines and the biasing of behavioral output in lobsters. Science 241:1175-1781. Kravitz, E.A. (2000). Serotonin and aggression: insights gained from a lobster model system and speculations on the role of amine neurons in a complex behavior. J. Comp. Physiol. A 186:221-238. Livingstone, M.S., Harris-Warrick, R.M. and Kravitz, E.A. (1980). Serotonin and octopamine produce opposite postures in lobsters. Science 208:76-87. Livingstone, M.S., Schaeffer, S.F. and Kravitz, E.A. (1981). Biochemistry and ultrastructure of serotonergic nerve endings in the lobster: serotonin and octopamine are contained in different endings. J. Neurobiol. 12:27-54. Ma, P.M. and Weiger, W.A. (1993). Serotonin-containing neurons in lobsters: the actions of gammaaminobutyric acid, octopamine, serotonin and proctolin on neuronal activity. J Neurophysiol 69:20152029. Ma, P.M., Beltz, B.S. and Kravitz, E.A. (1992). Serotonin-containing neurons in lobsters: their role as “gain setters” in postural control mechanisms. J. Neurophysiol. 68:36-54. Monastirioti, M. (1999). Biogenic amine systems in the fruit fly Drosophila melanogaster. Microscopy Research and Technique 45(2):106-121. Otsuka, M., Kravitz, E.A. and Potter, D.D. (1967). Physiological and chemical architecture of a lobster ganglion with particular reference to gamma-aminobutyrate and glutamate J. Neurophysiol. 30:725752. Pasztor, V.M. and Bush, B.M.H. (1987). Peripheral modulation of mechanosensitivity in primary afferent neurons. Nature Lond. 326:793-795. Schneider, H., Trimmer, B.A., Rapus, J., Eckert, M., Valentine, D. and Kravitz, E.A. (1993). Mapping of octopamine-immunoreactive neurons in the central nervous system of the lobster. J. Comp. Neurol. 329:129-142. Shiah, I.S. and Yatham, L.N. (1998). GABA function in mood disorders: an update and critical review. Life Sci. 63(15):1289-1303. Sombati, S. and Hoyle, G. (1984). Generation of specific behaviors in a locust by local release into neuropil of the natural neuromodulator octopamine. J. Neurobiol. 15:481-506. Weiger, W.A. (1997). Serotonergic modulation of behaviour: a phylogenetic overview. Biol. Rev. 72:61-95. Weiger, W.A. and Kravitz, E.A. (1994). Serotonergic neurons in the lobster: excitatory inputs from the tail fan. Proc. Soc. Neurosci. 337.18, p816. Weiger, W.A. and Ma, P.M. (1993). Serotonin-containing neurons in lobsters: origins and characterization of inhibitory postsynaptic potentials. J. Neurophysiol. 69:2003-2014. Yeh, S.-R., Fricke, R.A. and Edwards, D.H. (1996). The effect of social experience on serotonergic modulation of the escape circuit of crayfish. Science 271:366-369. Yeh, S.-R., Musolf, B.E. and Edwards, D.H. (1997). Neuronal adaptations to changes in the social dominance status of the crayfish. J. Neurosci. 17:697-708.
- 10 -
Fig.1 Sample recordings from amine-containing neurons, connectives and single axons. Right side of Figure: A schematic representation of the preparation used for searching for evoked synaptic inputs. Various combinations of recording and stimulating electrodes were used in different experiments. A1-A6, ganglia of the first to the sixth abdominal segments; A1-5HT cells, serotonin neurons in the 1st abdominal ganglion; T4-CLOCs, claw octopamine neurons in the 4th thoracic ganglion; T3-T5, ganglia of the third to fifth thoracic segments; LG, lateral giant interneuron; MG, medial giant interneuron. Left side of Figure: Recordings showing an EPSP followed by an action potential in an A1-5HT neuron after extracellular stimulation of LG posterior to A4. Upper trace: intracellular recording from the soma of the A1-5HT (arrow indicates time of electrical stimulation of a hemi-connective posterior to A4); middle trace: intracellular recording from the LG axon in T5; lower trace: extracellular recording of the LG spike in T4 (modified after Hörner et al., 1997). Fig.2 Synaptic activity in 5HT- (A, B) and OCT- (C, D) containing neurosecretory neurons after firing of giant fibers. A. Intracellular stimulation of the LG axon at T5 elicits longlasting EPSPs in the ipsilateral A1-5HT cell. Upper trace: 10 superimposed LG-evoked EPSPs in A1-5HT neuron; lower trace: extracellular recording of the LG spike anterior to A4. B. Intracellular stimulation of the MG axon at T5. Upper trace: EPSPs in A1-5HT neuron (compare with A). Lower trace: extracellular recording of the MG spike at A4 (other details as in A). C. Intracellular recording from an OCT-containing neuron in T4. Stimulation at the subesophageal ganglion elicits an EPSP (upper trace) in anterior T4 CLOC accompanied by the activation of an action potential in MG recorded extracellularly in T3 (lower trace). D. Intracellular recording from the same OCT-containing neuron shown in C after a small reduction of the stimulus amplitude. The recording shows that the spike in MG and the evoked EPSP in the OCT neuron are not related. Upper trace: Intracellular recording from an OCT-containing neuron anterior in T4; lower trace: extracellular recording of the MG spike in T3. Fig.3 Effects of LG stimulation on motorneurons. A. Intracellular stimulation of LG elicits short-duration EPSPs in an M7 motorneuron in A1; upper trace: EPSPs in M7 (4 traces superimposed); middle trace: extracellular recording of the LG spike at A4; lower trace: extracellular recording of the LG spike at T4. B. Intracellular stimulation of LG at T5 elicits EPSPs in an A1-5HT neuron and spiking of a fast flexor motorneuron recorded in the 3rd nerve root between A3 and A4; upper trace: mean time course of EPSPs (8 traces averaged) in an A1-5HT neuron elicited by intracellular stimulation of LG; middle trace: extracellular recording of the LG spike at A4; lower traces: 8 superimposed extracellular recordings of fast flexor motorneuron activity in the 3rd nerve root between A3 and A4. Fig.4 Effects of EPSPs evoked by single LG spikes on the spontaneous firing activity in an A1-5HT neuron. A. Firing of LG reduces the time interval between a spontaneous and an evoked spike in an A1-5HT neuron and increases the interval between the evoked and the next spontaneous spike. Upper trace: spont. interval: time between two spontaneous spikes in an A1-5HT neuron; n: time between two spikes in A1-5HT during stimulation of LG; n+1: time between two spikes in A1-5HT in the immediately following period; lower trace: extracellular recording of LG spike at A4. B. Diagram showing the effects of the timing of the LG spike on the spontaneous activity pattern in A1-5HT. Solid line: mean time interval between spontaneously occurring spikes in A1-5HT (dotted lines: standard deviation); filled circles: time interval n, open squares: time interval n+1 as explained in A. C. Statistical evaluation of the effects of single LG spikes on the spike intervals in an A1-5HT neuron. Compared to spontaneous spikes (Spont.), single LG spikes significantly reduce (n; P
