Nitric oxide (NO) and serotonin (5-HT) are two neurotransmitters with important roles in neuromodulation and synaptic plasticity. There is substantial evidence for a morphological and functional overlap between these two neurotransmitter systems, in particular the modulation of 5-HT function by NO. Here we demonstrate for the first time the modulation of an identified serotonergic synapse by NO using the synapse between the cerebral giant cell (CGC) and the B4 neuron within the feeding network of the pond snail Lymnaea stagnalis as a model system. Simultaneous electrophysiological recordings from the pre- and postsynaptic neurons show that blocking endogenous NO production in the intact nervous system significantly reduces the B4 response to CGC activity. The blocking effect is frequency dependent and is strongest at low CGC frequencies. Conversely, bath application of the NO donor DEA/NONOate significantly enhances the CGC-B4 synapse. The modulation of the CGC-B4 synapse is mediated by the soluble guanylate cyclase (sGC)/cGMP pathway as demonstrated by the effects of the sGC antagonist 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ). NO modulation of the CGC-B4 synapse can be mimicked in cell culture, where application of 5-HT puffs to isolated B4 neurons simulates synaptic 5-HT release. Bath application of diethylamine NONOate (DEA/NONOate) enhances the 5-HT induced response in the isolated B4 neuron. However, the cell culture experiment provided no evidence for endogenous NO production in either the CGC or B4 neuron suggesting that NO is produced by an alternative source. Thus we conclude that NO modulates the serotonergic CGC-B4 synapse by enhancing the postsynaptic 5-HT response.
Serotonin (5-HT) is an important modulatory neurotransmitter with significant effects on neuronal properties, synaptic function, network activity, and behavior (e.g., Bayliss et al. 1997; Birmingham and Tauck 2003; Castellucci and Schacher 1990; Dieudonne 2001; Katz 1998; Sillar et al. 2002; Weiger 1997). Recently, it has been recognized that there is significant morphological and functional overlap between 5-HT and nitric oxide (NO). Immunohistochemical studies, for example, have provided evidence for the co-localization of NO synthase (NOS) and 5-HT in dorsal raphe projections to the trigeminal somatosensory system in rats (Simpson et al. 2003). Co-localization of 5-HT and NOS has also been reported in other vertebrate species (e.g., Dun et al. 1994; Johnson and Ma 1993; Leger et al. 1998; Maqbool et al. 1995; Wang et al. 1995; Wotherspoon et al. 1994).
At the functional level, pharmacological studies have demonstrated that NO signaling can modulate 5-HT release from specific brain structures (Iuras et al. 2005; Prast and Philippu 2001; Segieth et al. 2001; Smith and Whitton 2000; Trabace and Kendrick 2000; Trabace et al. 2004). NO can also affect 5-HT re-uptake (Bryan-Lluka et al. 2004) and appears to interact with selective 5-HT re-uptake inhibitors used in the treatment of depression (Harkin et al. 2003, 2004; Inan et al. 2004). Furthermore, behavioral studies have suggested that NO modulates 5-HT1 receptor function (Chiavegatto and Nelson 2003; Pitsikas et al. 2005).
Despite this compelling evidence for interactions between 5-HT and NO, the direct effects of NO on an identified serotonergic synapse have not been studied. Here we report the results of an electrophysiological study of the effects of NO on the serotonergic synapse between the cerebral giant cells (CGCs) and B4 motoneurons in the feeding system of the pond snail Lymnaea stagnalis. The CGCs are an important pair of modulatory neurons with wide ranging effects on Lymnaea feeding behavior including arousal, “gating, ” frequency control, and long-term memory formation (Kemenes et al. 2006; Kyriakides and McCrohan 1989; Yeoman et al. 1994a,b). These effects are mediated via serotonergic synapses between the CGC and feeding interneurons and motoneurons including B4 (McCrohan and Benjamin 1980; Yeoman et al. 1996). The B4 motoneurons receive a monosynaptic excitatory input from the CGCs that increases their excitability and modulates a hyperpolarization-activated inward current, Ih, which enhances the B4 postinhibitory rebound property and activates an intrinsic bursting property (Straub and Benjamin 2001). These two effects are mimicked by application of 5-HT to B4 motoneurons (Straub and Benjamin 2001). NO has been reported to affect Lymnaea feeding both at the behavioral and circuit levels (Elphick et al. 1995; Kobayashi et al. 2000; Korneev et al. 2002; Moroz et al. 1993), but whether this is due to nitrergic co-transmission in the CGCs (Korneev et al. 1998) is unknown. Here, we demonstrate that NO enhances the postsynaptic effects of 5-HT released from the CGCs but this appears to be due to exogenous NO rather than NO originating from the CGCs.
All animals used in the experiments were raised in the Lymnaea breeding facilities at the University of Sussex and the University of Calgary. Snails were maintained at RT on a mixed diet of lettuce and dried fish food.
Intact nervous system preparation and intracellular recordings
All dissections were carried out in HEPES-buffered saline (normal saline, NS) containing (in mM) 50 NaCl, 1.6 KCl, 2 MgCl2, 3.5 CaCl2, and 10 HEPES, pH 7.9, in distilled water (Benjamin and Winlow 1981). The whole CNS consisting of the paired buccal ganglia and the circumesophageal ganglionic ring including the cerebral, pedal, pleural, parietal, and visceral ganglia was dissected from animals (shell length: 20–30 mm) and pinned out in a silicone elastomer (Sylgard)-lined recording chamber. The outer connective tissue was removed from the dorsal surface of the buccal ganglia and the ventral surface of the cerebral ganglia. Subsequently, the surfaces of the buccal and cerebral ganglia were treated with protease (Protease type XIV, Sigma, Poole, UK) for 60 s by placing small enzyme crystals on the connective tissue. The protease was washed off with an excess of NS.
Bridge and two-electrode voltage-clamp recording techniques were used to record simultaneously from pairs of CGC and B4 neurons. For this purpose, CGC neurons were routinely impaled with two electrodes, one for current injection and one for membrane potential recording. B4 neurons were impaled with single electrodes for bridge mode recordings and with two electrodes for voltage-clamp recordings of postsynaptic responses to CGC stimulation. During voltage-clamp experiments, the B4 membrane potential was clamped at the resting membrane potential. The recording electrodes were pulled from 2-mm glass capillaries with inner filament (GC200F-15; Harvard Apparatus, Edenbridge, UK) and filled with 3 M potassium acetate (electrode resistance, 15–35 MΩ). Signals were fed into intracellular recording amplifiers (AxoClamp-2B, Molecular Devices; NL102G, Digitimer, Welwyn Garden City, UK), digitized with a CED 1401plus interface (Cambridge Electronics Design, Cambridge, UK), and visualized and stored on a personal computer (PC) using Spike2 software (Cambridge Electronics Design).
During the recording, unless stated otherwise, whole nervous system preparations were superfused at a rate of ∼1.5 ml/min with a solution of hexamethonium chloride (1 mM) in a saline solution with an increased calcium and magnesium concentration (HiDi). HiDi saline contained (in mM) 35 NaCl, 2 KCl, 8 CaCl2, 14 MgCl2, and 10 HEPES, pH 7.9 in distilled water (Elliott and Benjamin 1989). This solution (HiDi/Hex) was used to isolate the monosynaptic excitatory CGC-B4 synapse from polysynaptic CGC-B4 interactions.
Other pharmacological agents that were added to the perfusion system were either dissolved directly in HiDi/Hex [methysergide (Tocris, Bristol, UK), DEA/NONOate (Axxora, Nottingham, UK), PTIO, l-NAME, d-NAME (all Sigma)] or first dissolved in DMSO [7-nitroindazole (7-NI; Sigma), ODQ (1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one) (Axxora)] before being diluted to the final concentration in HiDi/Hex. The maximum concentration of DMSO in the final solution was 0.05%. Control experiments with 0.05% DMSO in HiDi/Hex showed that this concentration of DMSO has no effect on the CGC-B4 interaction (data not shown).
Cell culture techniques and recording from isolated neurons
CGC and B4 neurons were isolated from the intact nervous system using procedures described previously (Koert et al. 2001; Straub and Benjamin 2001). B4 neurons that were isolated to test effects of NO on 5-HT induced currents in single cells were plated on 35 mm poly-l-lysine-coated BioCoat dishes (BD Biosciences, Bedford, MA) in defined medium (DM) containing 50% vol/vol Leibovitz L-15 (Invitrogen, Carlsbad, CA) and (in mM) 40 NaCl, 1.7 KCl, 4.1 CaCl2, 1.5 MgCl2, and 10 HEPES, pH 7.9 in distilled water. Single B4 neurons were recorded 24–48 h after isolation using two-electrode voltage-clamp techniques, whereas the culture dish was perfused with NS at a rate of ∼1 ml/min. Synaptic responses were mimicked by pressure application of 1-s 5-HT pulses (1 mM in NS) using a Picospritzer (General Valve, Fairfield, NJ).
For the study of effects of NO on re-constructed CGC-B4 synapses, pairs of CGC and B4 neurons were plated on poly-l-lysine-coated coverslips at a distance of ∼150–250 μm in brain-conditioned DM (Ridgway et al. 1991) so that they were not in direct contact. Under these conditions, neurons rapidly extended new processes and formed functional connections. Simultaneous recordings from pairs of neurons cultured for 24 h using intracellular electrodes showed that ∼70% of CGC-B4 pairs (n = 34) that had established contact had formed functional chemical excitatory synapses from the CGC to the B4 neuron. In addition, 50% of CGC-B4 pairs had formed electrical connections with an average coupling coefficient of 0.07 ± 0.01 (n = 17). There was no obvious correlation between the formation of chemical and electrical connections. During the recording culture dishes were perfused with NS at a rate of ∼1 ml/min. All pharmacological agents [DEA/NONOate (Axxora), PTIO (Axxora), l-arginine (Sigma)] were dissolved in NS and added to the perfusion system.
CGC activity has complex mono- and polysynaptic effects on B4 motoneurons
The CGC forms monosynaptic synaptic connections with a large number of buccal neurons involved in the control of feeding movements including the B4 motoneurons (Benjamin and Elliott 1989; McCrohan and Benjamin 1980; Yeoman et al. 1996). Consequently, CGC activity produces mono- and polysynaptic effects on the B4 motoneuron that result in a complex postsynaptic response when the CGC-B4 interaction is recorded in NS. The typical B4 response in NS consists of an initial depolarization due to the monosynaptic excitatory CGC-B4 synapse followed by variable hyperpolarizing responses due to polysynaptic inhibitory interactions (Fig. 1Ai). These compound synaptic effects complicate a quantitative analysis of the direct CGC-B4 connection. Incubating the preparation in a saline solution with a raised calcium and magnesium concentration (HiDi saline) that has previously been shown to suppress polysynaptic interactions (Elliott and Benjamin 1989) simplifies the situation but does not completely remove polysynaptic inhibitory synaptic interactions. The inhibitory synaptic inputs can be abolished by adding the cholinergic antagonist hexamethonium chloride (1 mM) to the HiDi saline (HiDi/Hex) leaving the direct monosynaptic excitatory connection between the CGC and B4 motoneuron intact (Fig. 1Aii). Under these conditions, the direct monosynaptic connection between the CGC and B4 motoneuron has two components, an initial transient depolarization that peaks ∼1.5 s after the start of CGC stimulation and a prolonged weak depolarization that persists for >10 s (Fig. 1Aii). If not stated otherwise, HiDi/Hex solution was used in all subsequent experiments that involved recording the CGC-B4 synapse in the intact nervous system.
5-HT is the primary transmitter at the CGC-B4 synapse
The CGCs have previously been shown to contain 5-HT (McCaman et al. 1984) and pharmacological studies of the CGC-B4 synapse have reported that the effects of CGC activity on B4 motoneurons can be blocked by serotonergic antagonists (Tuersley and McCrohan 1989). Here, we confirm this finding by using the nonspecific 5-HT1/2 receptor antagonist methysergide (50 μM). Bath application of methysergide reduced the average peak B4 depolarization induced by a burst of 10 CGC action potentials (APs) at 10 Hz from 7.6 ± 0.8 to 1.4 ± 0.2 mV (n = 7; Fig. 1Bi). This represents an 82% block of the initial response (Fig. 1C). Methysergide also reduces the late component of the B4 excitatory postsynaptic potential (EPSP) by 75% from 0.7 ± 0.2 to 0.2 ± 0.1 mV, which suggests that both the early and late component are predominantly due to the release of 5-HT. The remaining component is either mediated by a 5-HT receptor subpopulation that is insensitive to methysergide or by the co-release of neuropeptides (e.g., myomodulin, SPTR peptide, ERYM peptide) present in the CGCs (Koert et al. 2001; Santama et al. 1994). The methysergide induced block of the CGC-B4 synapse showed only some weak recovery and even after extended washout (>60 min) of the blocker the peak B4 EPSP amplitude increased only marginally to 1.6 ± 0.1 mV (n = 7). This is consistent with other reports that the methysergide block of serotonergic effects are irreversible (Doggrell 1987; Sakurai and Katz 2003).
Bath application of the NO synthase inhibitor 7-NI (0.25 mM) in combination with methysergide or after washout of methysergide caused a further small reduction in the peak B4 EPSP amplitude to an average of 1.0 ± 0.2 mV (n = 3; Fig. 1Bii). This effect was fully reversible.
The results of these experiments are summarized in Fig. 1C, which clearly illustrates the strong irreversible blocking effect of the serotonergic antagonist methysergide on the CGC-B4 synapse. It also shows that 7-NI causes a further small reversible reduction of the residual B4 EPSP from 18 ± 2 to 12 ± 2% of the control peak value. The peak B4 EPSP amplitudes during methysergide application, 7-NI application (after methysergide), and the wash period did not differ significantly from each other (ANOVA: F = 46.0, P ≤ 0.001; Tukey t-test: P > 0.05) but were significantly smaller than during the control period (Tukey t-test: P ≤ 0.001).
NO modulates serotonergic connection between CGC and B4 motoneurons
The previous results suggest a minor role of NO as a co-transmitter at the CGC-B4 synapse. However, when the effect of 7-NI on the CGC-B4 synapse was tested without prior block of the serotonergic component by methysergide, 7-NI (0.25 mM) had a far greater effect on the CGC-B4 interaction reducing the peak B4 EPSP amplitude in response to a burst of 10 CGC APs at 10 Hz from 8.5 to 4.7 mV in the example shown (Fig. 2A). This block was completely reversible.
The blocking effect of 7-NI increases strongly at concentrations between 0.05 and 0.25 mM from an average of 3 ± 3% to an average of 48 ± 7% (n = 3) reduction in the peak B4 EPSP amplitude (Fig. 2B). Raising the 7-NI concentration further did not significantly increase its total blocking effect. Even an increase in the 7-NI concentration to 1 mM only raised the blocking effect by 5% to a maximum value of 53 ± 4% (n = 3). The concentration dependence of the 7-NI blocking effect can be fitted with a modified Hill equation with an IC50 value of 137 μM and a Hill coefficient of 3.47.
These experiments were carried out in HiDi/Hex saline to suppress polysynaptic interactions, and to rule out the possibility that the increased concentration of divalent ions in HiDi saline alters the levels of endogenous NO production, we carried out additional control experiments in NS plus hexamethonium chloride (1 mM). Under these conditions, 7-NI also reduced the peak B4 EPSC recorded under voltage-clamp conditions in response to a CGC burst (10 Hz, 1 s) by 46 ± 4% (n = 4) from an average of −1.3 ± 0.1 to −0.7 ± 0.1 nA. This is not significantly different from the 7-NI (0.25 mM) reduction observed in HiDi/Hex saline (48 ± 7%; t-test: P > 0.05). Thus the use of HiDi saline does not significantly alter the relative contribution of NO modulation to the CGC-B4 synapse.
The observation that the NOS inhibitor 7-NI can block ∼50% of the CGC-B4 synapse is not compatible with the hypothesis that NO and 5-HT act as co-transmitters with simply additive effects at this synapse. If that was the case, blocking NO production should reduce the amplitude of the CGC-B4 synapse maximally by 20% as the serotonergic antagonist methysergide blocks 80% of this synapse. Thus these observations suggest that NO acts predominantly as a modulator of the serotonergic effects at the CGC-B4 synapse rather than a direct mediator of the synaptic interaction.
The observation that blocking the NO signaling pathway can modulate the CGC-B4 synapse was further confirmed using an alternative NOS inhibitor, l-NAME (1 mM), that also reduced the B4 EPSC amplitude in response to 10 Hz CGC stimulation by 40 ± 5% (paired t-test: P < 0.01, n = 6; Fig. 2C). In contrast, the inactive compound d-NAME (1 mM) did not significantly alter the B4 EPSC amplitude (paired t-test: P = 0.44, n = 6; Fig. 2C). Similarly, removal of extracellular NO by bath application of the NO scavenger PTIO also resulted in a reduction in the CGC-B4 interaction by 49 ± 15% (paired t-test: P < 0.05, n = 4; Fig. 2C).
NO modulation of CGC-B4 synapse is frequency-dependent
The modulatory effect of NO was further characterized by studying the effect of 7-NI on B4 EPSPs elicited by stimulating the CGC at frequencies between 1 and 10 Hz. In accordance with the previous results, 7-NI (0.25 mM) reduced the average peak B4 response triggered by a burst of 10 CGC APs at 10 Hz from 9.7 ± 0.7 to 5.3 ± 0.6 mV (n = 10; Fig. 3Ai, left). This effect was significant and fully reversible (ANOVA: F = 13.9, P ≤ 0.001; Tukey t-test: control vs. 7-NI: P ≤ 0.01, 7-NI vs. wash: P ≤ 0.001, control vs. wash: P = 0.31). Similarly, the peak B4 EPSP amplitude elicited by CGC stimulation at 1 Hz (4.5 ± 0.6 mV, n = 6) was significantly reduced by 7-NI application to 1.6 ± 0.2 mV (Fig. 3Ai, right; ANOVA: F = 7.1, P ≤ 0.01; Tukey t-test: control vs. 7-NI: P ≤ 0.05, 7-NI vs. wash: P ≤ 0.01, control vs. wash: P = 0.50).
A direct comparison of the blocking effects however revealed that 7-NI reduced the amplitude of the B4 EPSP by 64 ± 5% when the CGC was stimulated at 1 Hz, whereas it reduced the B4 EPSP amplitude by only 45 ± 4% when the CGC was stimulated at 10 Hz (Fig. 3Aii). Further analysis of 7-NI effects on the B4 EPSP amplitude at CGC stimulation frequencies of 5 (n = 6), 3.3 (n = 9), and 1.7 Hz (n = 3) revealed a significant correlation between the CGC stimulation frequency and the proportion of the CGC-B4 synapse that was blocked by 7-NI (Spearman's correlation coefficient: ρ = 0.43, P ≤ 0.01).
Similar results were obtained in independent experiments to test the effects of 7-NI on the CGC-B4 synapse when the membrane potential of the B4 neuron was voltage-clamped at the B4 resting membrane potential. Under these conditions, the average peak EPSC elicited by CGC stimulation at 1 and 10 Hz was −0.39 ± 0.09 nA (n = 5) and −1.05 ± 0.11 nA (n = 5), respectively (Fig. 3Bi). 7-NI application significantly reduced these currents to −0.14 ± 0.04 and −0.63 ± 0.06 nA, respectively (1-Hz CGC stimulation: ANOVA: F = 4.2, P ≤ 0.05; Tukey t-test: control vs. 7-NI: P ≤ 0.05, 7-NI vs. wash: P = 0.12, control vs. wash: P = 0.83; 10-Hz CGC stimulation: ANOVA: F = 7.1, P ≤ 0.01; Tukey t-test: control vs. 7-NI: P ≤ 0.01, 7-NI vs. wash: P ≤ 0.05, control vs. wash: P = 0.74). Thus 7-NI blocked 64 ± 6% of the B4 EPSC induced by CGC stimulation at 1 Hz, whereas the current induced by 10-Hz CGC stimulation was only reduced by 40 ± 3%. Additional tests of the effects of 7-NI on B4 EPSCs triggered by CGC stimulation at 3.3 (n = 5) and 1.7 Hz (n = 5) showed intermediate levels of blocking effects (Fig. 3Bii). Overall, a significant correlation existed between the CGC stimulation frequency and the proportion of the EPSC blocked by 7-NI (Spearman's correlation coefficient: ρ = 0.78, P ≤ 0.001). These results demonstrate that NO modulation of the serotonergic CGC-B4 interaction is more important at low than at high CGC firing frequencies (see also following text).
NO has differential effects on early and late component of CGC-B4 synapse
A closer inspection of the effects of 7-NI on the CGC-B4 synapse also revealed that although 7-NI partially blocks the peak EPSC amplitude, it does not significantly reduce the late B4 EPSC amplitude measured 8–10 s after the start of the CGC stimulation. The average amplitude of the EPSC at this point was −0.21 ± 0.07 nA (n = 5) prior to 7-NI application and −0.23 ± 0.02 nA during 7-NI application in response to a burst of 10 CGC APs at 10 Hz (Fig. 3Bi, left). The average amplitude of this late EPSC component in the presence of 7-NI also remained unchanged during the washout of 7-NI at −0.23 ± 0.05 nA (ANOVA: F = 0.1, P = 0.91). Similarly, comparisons of the average B4 EPSC amplitudes measured 8–10 s after the start of CGC stimulation at 3.3, 1.7, and 1 Hz prior, during, and after 7-NI application also revealed no statistically significant differences (ANOVAs: 3.3 Hz: F = 0.4, P = 0.67; 1.7 Hz: F = 0.6, P = 0.55; 1 Hz: F = 1.8, P = 0.20). This observation suggests that the interaction between the CGC and B4 motoneuron consists of multiple components that are differentially affected by NO.
Modulation of CGC-B4 synapse requires activation of soluble guanylate cyclase
Activation of soluble guanylate cyclase (sGC) resulting in the production of cGMP is the classical receptor mechanism for mediating the cellular effects of NO (Garthwaite and Boulton 1995). Here we tested whether suppressing the NO-induced production of cGMP using the sGC specific antagonist ODQ also affects the CGC-B4 synapse in Lymnaea. As in the previous experiments, bursts of 10 CGC APs were triggered at frequencies between 1 and 10 Hz. Application of ODQ reduced the average peak B4 EPSP amplitude triggered by 10-Hz CGC stimulation from 10.5 ± 0.5 to 5.0 ± 0.4 mV (n = 8; Fig. 4A, left; ANOVA: F = 34.6, P ≤ 0.001; Tukey t-test: control vs. ODQ: P ≤ 0.001, ODQ vs. wash: P ≤ 0.001, control vs. wash: P = 0.05). The average amplitude of B4 EPSPs in response to CGC stimulation at 1 Hz was reduced by ODQ from 3.6 ± 0.3 to 1.2 ± 0.2 mV (n = 7; Fig. 4A, right; ANOVA: F = 18.8, P ≤ 0.001; Tukey t-test: control vs. ODQ: P ≤ 0.001, ODQ vs. wash: P ≤ 0.001, control vs. wash: P = 1.00).
The proportion of the average peak B4 EPSP blocked by ODQ was again dependent on the CGC stimulation frequency and ranged from 53 ± 3% at 10 Hz to 67 ± 8% at 1 Hz (Fig. 4B). Further analysis showed a significant correlation between the CGC stimulation frequency and the ODQ blocking effect (Spearman's correlation coefficient: ρ = 0.44, P ≤ 0.05).
The blocking effect of ODQ, similar to the effects of 7-NI on the CGC-B4 synapse, was restricted to the early peak B4 response. The average EPSP amplitudes 8–10 s after the start of CGC stimulation were not significantly affected by ODQ application at the tested CGC stimulation frequencies (ANOVAs: 10 Hz: F = 1.2, P = 0.32; 3.3 Hz: F = 0.2, P = 0.81; 1.7 Hz: F = 1.0, P = 0.37; 1 Hz: F = 0.18, P = 0.84).
NO effects on CGC-B4 synapse are not due to modulation of the CGC AP shape
Previous studies have suggested that NO can modulate the synaptic release of 5-HT. This could be due to the modulation of the shape of presynaptic APs. Here we analyzed the effect of 7-NI on the shape of individual spontaneous CGC APs and the amplitude of unitary B4 EPSPs. Using spontaneous APs for this analysis has the advantage that the AP shape is not distorted by an underlying current pulse.
Single spontaneous CGC APs occurring at intervals >5 s elicited unitary EPSPs in B4 neurons with little temporal summation (Fig. 5A). Bath application of 7-NI (0.25 mM) had little effect on the spontaneous CGC firing rate (control: 0.17 ± 0.02 Hz; 7-NI: 0.16 ± 0.03 Hz; paired t-test: P = 0.70, n = 6). However, the average peak amplitude of CGC-induced unitary B4 EPSPs was reduced from 3.4 ± 0.5 to 0.6 ± 0.1 mV (n = 6; Fig. 5A) by the application of 7-NI. Again, this block is completely reversible (Fig. 5Bi; ANOVA: F = 13.1, P ≤ 0.001; Tukey t-test: control vs. 7-NI: P ≤ 0.01, 7-NI vs. wash: P ≤ 0.001, control vs. wash: P = 0.65). This represents a reduction in the EPSP amplitude of 82 ± 3%, which is even more pronounced than the effect of 7-NI on CGC stimulation at 1 Hz and further supports the earlier finding that NO modulation of the CGC-B4 synapse is particularly important at slow CGC firing rates.
Comparison of spike shapes prior to and during 7-NI application revealed no obvious differences in the depolarizing phase of the AP (Fig. 5A, inset). Application of 7-NI however appeared to cause a small reduction in the afterhyperpolarization (AHP) after the AP. A statistical comparison of key AP parameters confirmed these findings and showed no significant differences in CGC AP amplitude (Fig. 5Biii; ANOVA: F = 1.1, P = 0.36) and width (Fig. 5Biv; ANOVA: F = 1.4, P = 0.28). In contrast, the AHP reduction was statistically significant (Fig. 5Bii; ANOVA: F = 8.5, P ≤ 0.01; Tukey t-test: control vs. 7-NI: P ≤ 0.05, 7-NI vs. wash: P ≤ 0.01, control vs. wash: P = 0.65). These results strongly suggest that the modulation of the CGC-B4 synapse by NO is not due to alterations in the presynaptic AP shape. In fact, the reduction in the CGC AHP caused by 7-NI would be more consistent with an increase rather than a decrease in transmitter release.
CGC-B4 synapse is enhanced by NO donors
After the demonstration that interference with the NO signaling pathway partially blocks the CGC-B4 synapse, we also investigated whether exogenous NO application can boost the synaptic interaction between CGCs and B4 motoneurons. For this purpose, the NO donor DEA/NONOate (DEA) was added to the perfusion solution while recording the interaction between CGCs and B4 motoneurons. Preliminary results have shown that prolonged DEA application has a weak sustained depolarizing effect on the B4 membrane potential. Therefore we voltage-clamped the B4 soma to eliminate changes in the membrane potential during these experiments.
Under these conditions, 1 mM DEA application produced a small inward current with an amplitude of −0.12 ± 0.02 nA (n = 5). It also enhanced the B4 EPSC amplitude triggered by a burst of 10 CGC APs at 10 Hz from −0.88 to −1.39 nA (Fig. 6A). After washout of the DEA for 15 min, the amplitude of the EPSC returned to −1.02 nA. On average, 1 mM DEA caused a significant reversible increase in the B4 EPSC of 40 ± 9% (n = 5; ANOVA: P ≤ 0.01; Tukey t-test: control vs. DEA: P ≤ 0.01, DEA vs. wash: P ≤ 0.05, control vs. wash: P > 0.05; Fig. 6Bi).
Reducing the NO donor concentration to 0.1 mM has similar effects on the CGC-B4 synapse enhancing the B4 EPSC amplitude by 42 ± 13% (n = 3; ANOVA: P ≤ 0.05; Tukey t-test: control vs. DEA: P ≤ 0.05, DEA vs. wash: P ≤ 0.05, control vs. wash: P ≥ 0.05; Fig. 6Bii). In contrast, 1 and 10 μM DEA had no significant effects on the CGC-B4 synapse (ANOVAs: P ≥ 0.05, n = 3 for each).
NO modulates postsynaptic effects of 5-HT on isolated B4 motoneurons
The previous experiments demonstrate a clear modulatory effect of NO on the CGC-B4 synapse. Consequently, we attempted to localize the site of the modulatory action by using single isolated B4 motoneurons to test the hypothesis that NO modulates the CGC-B4 synapse by altering the postsynaptic 5-HT response. We have used isolated B4 motoneurons successfully in the past to characterize 5-HT effects on cellular properties (Straub and Benjamin 2001). The effects of CGC stimulation on a B4 motoneuron can be mimicked in cell culture by applying a puff of 5-HT from a pipette positioned close to the cell body of a single isolated B4 motoneuron. A 1-s puff of 5-HT (0.1 mM) triggered an inward current with an average amplitude of −0.15 ± 0.03 nA (n = 8; Fig. 6Ci). Similar to the effects in the intact nervous system, bath application of 1 mM DEA caused a small change in the holding current by −0.06 ± 0.02 nA (n = 8). It also enhanced the amplitude of the 5-HT-induced inward current to −0.24 ± 0.04 nA, which represents an increase of 61 ± 15% (Fig. 6Cii). The average amplitude of the 5-HT induced inward current remained enhanced by 45 ± 21% after washout of DEA for 10–15 min (ANOVA: F = 5.9, P ≤ 0.01; Tukey t-test: control vs. DEA: P ≤ 0.01, DEA vs. wash: P = 0.68, control vs. wash: P = 0.09). These results clearly demonstrate that NO can modulate the postsynaptic effect of 5-HT.
NO modulates CGC-B4 synapse in cell culture
After the demonstration that NO can modulate the response to 5-HT in isolated B4 neurons, we also studied whether NO has the same effect on reconstructed CGC-B4 synapses in cell culture. Recordings from pairs of CGC and B4 motoneurons that had established synaptic interactions in cell culture confirmed that NO can enhance chemical CGC-B4 interactions. In the example shown, the amplitude of the B4 EPSP measured just after the end of the CGC stimulation was 3.2 mV (Fig. 7Ai). Bath application of the NO donor DEA (0.1 mM) increased the postsynaptic effect of CGC activity so that CGC stimulation produced a more pronounced depolarization with an amplitude of 10.7 mV. The enhancement of the CGC-B4 synapse is reversed after washout of DEA. On average, bath application of DEA increased the CGC-B4 interaction in cell culture by 330 ± 67% (n = 7; ANOVA: F = 9.6, P ≤ 0.01; Tukey t-test: control vs. DEA: P ≤ 0.001, DEA vs. wash: P ≤ 0.05, control vs. wash: P = 0.30; Fig. 7Aii).
Subsequently, we studied whether endogenous NO production by the CGCs contributes to the modulation of CGC-B4 synapses in cell culture. For this purpose, the NO scavenger PTIO (0.25 mM) was added to the bath solution. We have previously shown that this treatment is very efficient at blocking nitrergic synaptic connections (Park et al. 1998). However, PTIO application had no significant effect on the amplitude of the chemical component of the CGC-B4 synapse in cell culture (Fig. 7Bi). The average B4 amplitude in the presence of PTIO remained at 96 ± 2% of the control value (paired t-test: P = 0.16, n = 5; Fig. 7Bii).
Furthermore, we also studied whether addition of l-arginine (1 mM) to the bath solution can boost CGC-B4 synapses in cell culture to exclude the possibility that NOS activity in cell culture is suppressed due to a lack of l-arginine, the substrate for NO production by NOS. However, l-arginine addition for ≤20 min did not enhance the CGC-B4 synapse (Fig. 7C; ANOVA: F = 1.7, P = 0.20). This result further supports the hypothesis that CGCs, at least in cell culture, do not express a sufficient amount of NOS to produce a high enough concentration of NO to modulate their serotonergic effects on B4 motoneurons.
The experiments presented here are the first description of the modulation of a serotonergic synapse by NO between two identified neurons, the CGC and B4 neurons, where both the pre- and postsynaptic neurons are accessible for electrophysiological recordings. This modulation is most significant at low CGC firing rates when blocking NO production reduces the amplitude of the CGC-B4 synapse by ≤80%. In contrast, at high CGC frequencies (10 Hz), modulation of the CGC-B4 synapse by NO only accounts for 40–50% of the EPSP amplitude. This frequency dependence of the NO modulation of the CGC-B4 synapse is probably due to saturation of the postsynaptic response by high concentrations of 5-HT that limits the B4 response amplitude. This observation is supported by the nonlinear relationship between the CGC firing rate and B4 response amplitude in the range from 1 to 10 Hz (Straub, unpublished observation). Thus NO has more potential to enhance the CGC-B4 synapse at low CGC frequencies/5-HT concentrations than at high CGC frequencies/5-HT concentrations.
Our results also indicate that the effects of NO are mediated via the sGC-cGMP pathway as application of the sGC antagonist ODQ mimics the effects of blocking NO production, consistent with the effects of NO on neuronal properties in other mollusks (Koh and Jacklet 2001; Schrofner et al. 2004; Van Wagenen and Rehder 2001). Furthermore, we present the first direct evidence that these effects are localized postsynaptically as NO donor application can enhance the response of isolated B4 motoneurons to 5-HT pulses. In contrast, there is no evidence for broadening of presynaptic action potentials, which could cause increased 5-HT release from the CGC. The overall effects of NO donors on 5-HT application to isolated B4 motoneurons and on CGC-B4 synapses in the intact nervous system are comparable, which suggests that the postsynaptic effects can fully account for the modulation of the CGC-B4 synapse by NO. Thus we conclude that NO alters the effects of activity-dependent 5-HT release from the CGCs by modulating predominantly postsynaptic 5-HT function.
This conclusion is consistent with behavioral pharmacological studies, which suggest that NO can alter behavior in mice and rats by modulating 5-HT1 receptor function (Chiavegatto and Nelson 2003; Pitsikas et al. 2005). Lymnaea also possesses a 5-HT1-like receptor (Sugamori et al. 1993) that is expressed in the buccal ganglia (Gerhardt 1996). Although it is not known whether this receptor is expressed in buccal B4 motoneurons, the block of the CGC-B4 synapse by the 5-HT1/2 antagonist methysergide is certainly consistent with a potential 5-HT1-like receptor on the B4 motoneuron. It is also noteworthy that the cloned Lymnaea 5-HT1-like receptor has two consensus sites for phosphorylation by cAMP/cGMP-dependent protein kinases (Sugamori et al. 1993). These are located within the predicted second and fourth cytoplasmic loops at amino acid residues 204–207 and 405–408, respectively. These two phosphorylation sites appear to be highly conserved among equivalent molluscan 5-HT receptors and are also found in 5-HT receptors cloned from another fresh water snail, Helisoma trivolis (Accession No. Q6TKQ2) and the marine mollusk Aplysia californica (Accession No. Q8MX83). A more general search of the Swiss-Prot database revealed consensus sites for phosphorylation by cAMP/cGMP-dependent protein kinases at positions equivalent to the phosphorylation site at position 204–207 in the molluscan 5-HT1-like receptors, among others, in mouse 5-HT1D/F receptors (Accession Nos. Q25414 and Q02284) and 5-HT1B/D receptors from the Japanese puffer fish (Accession Nos. O42384 and P79748). This suggests that phosphorylation of 5-HT1 receptors at this site by cAMP/cGMP-dependent kinases might be a widespread, evolutionary conserved mechanism for the regulation of 5-HT1 receptor function. In contrast, the second cAMP/cGMP-dependent protein kinase phosphorylation site appears to be unique to the molluscan 5-HT1-like receptors and is not found in vertebrate 5-HT receptors. Direct evidence for the modulation of postsynaptic receptor function by the NO-sGC-cGMP signaling pathway has also been reported for GABAA receptors (Cupello and Robello 2000; Fukami et al. 1998; Wexler et al. 1998) and NMDA receptors (Barnstable et al. 2004).
Alternatively, NO modulation of the CGC-B4 synapse could occur downstream of the postsynaptic 5-HT receptor. A potential target is the cAMP-gated sodium current (INa, cAMP) that has been found in various molluscan neurons including buccal feeding motoneurons in Lymnaea and Pleurobranchaea (Green and Gillette 1983; McCrohan and Gillette 1988). Similarly, a 5-HT activated, cAMP-mediated cation current has also been identified in Aplysia (Jacklet et al. 2006). In Pleurobranchaea, it has been shown that INa,cAMP is activated by 5-HT (Sudlow and Gillette 1995, 1997). Furthermore, recent work has shown that INa, cAMP can be enhanced by NO (Hatcher et al. 2006). However, the effect of NO on this current has been reported to be independent of the cGMP signaling pathway, whereas our results clearly demonstrate a role for the cGMP signaling pathway in the modulation of the CGC-B4 synapse by NO.
The conclusion that NO modulates postsynaptic 5-HT effects rather than presynaptic 5-HT release differs from a number of vertebrate studies that used microdialysis techniques to measure 5-HT concentrations in vivo (Iuras et al. 2005; Smith and Whitton 2000; Trabace and Kendrick 2000; Trabace et al. 2004; Wegener et al. 2000). These studies demonstrated changes in the concentration of 5-HT after NO application or NOS inhibition in various brain structures. This has prompted the suggestion that NO can modulate presynaptic 5-HT release. Similarly, NO has been shown to affect the release of various other neurotransmitters including acetylcholine, glutamate, catecholamines, and histamine (Prast and Philippu 2001). Although it is possible that some of these effects are indirect effects due to changes in the level of presynaptic activity, a number of studies have reported direct effects of NO on synaptic transmitter release that causes short- and long-term modulation of synaptic function (Arancio et al. 2001, 1995; Meffert et al. 1996; Wang et al. 2005). NO has also been shown to directly affect endocytosis and synaptic vesicle recycling (Micheva et al. 2003). The hypothesis that NO can directly modulate synaptic vesicle release is further supported by considerable evidence for the modulation of presynaptic transmitter release by cGMP either via the activation of cGMP-dependent kinase or of cyclic nucleotide-gated cation channels (Barnstable et al. 2004; Wang and Robinson 1997). Therefore we cannot completely rule out the possibility that modulation of presynaptic 5-HT release could also contribute to the effects of NO on the CGC-B4 synapse even though blocking NO production did not result in a change in spike width that could account for a reduction in 5-HT release. Interestingly, a cGMP-dependent enhancement of an action potential AHP, which is comparable to the effect of blocking NO production on the CGC spike AHP, has also been described for posterior pituitary terminals (Klyachko et al. 2001). This effect has been linked to a reduction in the number of action potential failures during trains of action potentials. However, no such effect was observed in our experiments as the stimulus parameters were set to values sufficient to trigger action potentials throughout the experiments without failures.
It has also been suggested that 5-HT and NO can directly interact with each other leading to the formation of 4-nitroso-serotonin and 4-nitro-serotonin that are unable to activate 5-HT receptors (Fossier et al. 1999). However, it is unlikely that this is contributing to the effects observed on the CGC-B4 synapse as blocking NO synthesis increases rather than decreases the effects of 5-HT on the B4 neuron.
Previous studies of NOS distribution in Lymnaea have reported that the CGCs contain NOS mRNA (Korneev et al. 1998). This suggested that NO could act as a co-transmitter at CGC synapses and could modulate directly the serotonergic output from the CGCs. However, our experiments in cell culture failed to reveal modulation of the CGC-B4 synapse by endogenous NO production as neither the NO scavenger PTIO nor addition of the NOS substrate l-arginine produced a significant effect. Nevertheless, CGC-B4 interactions could be enhanced significantly by the application of an exogenous NO donor, demonstrating the competence of reconstructed CGC-B4 synapses in cell culture for modulation by NO. Our results suggest that the level of functional NOS protein in the CGCs is too low for the production of NO at a level sufficient to modulate the CGC-B4 synapse. We cannot completely exclude the possibility that this is an artifact of the cell culture system. However, B2 neurons cultured under the same conditions maintain their ability to produce NO and form functional nitrergic synapses (Park et al. 1998). The apparent inability of CGCs to produce NO is consistent with previous reports that have shown that the majority of CGCs in the intact nervous system are NADPH diaphorese negative despite the presence of NOS mRNA (Korneev et al. 1999). The discrepancy between the presence of NOS mRNA and the apparent lack of functional NOS protein can be attributed to the co-expression of a NOS pseudogene in the CGCs, which contains a region of significant antisense homology to the NOS mRNA (Korneev et al. 1999). This antisense region enables the transcript of the pseudogene to form a heteroduplex with the NOS mRNA, which suppresses NOS mRNA translation. The suggestion that CGCs are unable to produce NO is also supported by a recent study, which showed a low concentration of NOS-related metabolites in the CGC (Moroz et al. 2005). However, it should be noted that the level of NOS and pseudo-NOS mRNA expression is dynamically regulated in the CGCs as reported in a recent publication showing that appetitive conditioning causes a transient upregulation in NOS expression and downregulation in pseudo-NOS expression in the CGCs (Korneev et al. 2005). Thus appetitive conditioning transiently shifts the balance between NOS and pseudo-NOS in favor of NOS expression. Furthermore, we cannot rule out the possibility that NOS expression is differently regulated in the soma and distal axonal processes as it has been shown that certain mRNAs are specifically targeted to axonal processes (e.g., Gioio et al. 2004; Mohr and Richter 2000; Smith et al. 2001; Sotelo-Silveira et al. 2006). Thus we cannot ignore the possibility that, at least in some cases, endogenous NO production in the CGC might contribute to the modulation of the CGC-B4 synapse.
The apparent lack of NO production in the CGCs raises the question about alternative NO sources that could modulate the CGC-B4 synapse in the intact nervous system. Studies of NOS distribution in the Lymnaea nervous system using NADPH diaphorese staining, anti-NOS antibodies, and in situ hybridization have shown that the buccal ganglia contain ∼100 NOS-expressing cells (Moroz et al. 1994; Park et al. 1998; Serfozo et al. 2002). Among those, the B2 neuron is the only identified neuron and the most prominent source of NO production in the buccal ganglia. B2 neurons have been shown to form nitrergic synapses with the B7nor neurons in the buccal ganglia (Park et al. 1998). They also exhibit a high level of spontaneous activity, which could be responsible for the production of a tonic background level of NO that underlies modulation of the CGC-B4 synapse. However, experimental manipulation of B2 activity has provided no evidence for an involvement of NO production by the B2 neuron in the modulation of the serotonergic CGC-B4 synapse (data not shown). Thus at present, the precise neuronal source of NO involved in the modulation of the CGC-B4 synapse is unknown.
However, NO has undoubtedly significant effects on molluscan feeding behavior, and NOS has been shown to be widely distributed throughout the neuronal networks that underlie the control and initiation of feeding behavior in mollusks (Moroz and Gillette 1995). The functional significance of NO in the molluscan feeding system has been demonstrated by pharmacological and behavioral studies that have revealed a role of NO in the activation of feeding and food-related learning paradigms (Elphick et al. 1995; Kemenes et al. 2002; Korneev et al. 2002; Moroz et al. 1993; Teyke 1996). NO and 5-HT have also been shown to complement each other in the regulation of procerebral lobe oscillations that are part of the olfactory processing network, which is involved in the activation of feeding behavior (Inoue et al. 2001).
At the cellular level, it has been demonstrated that NO can act as the sole anterograde messenger between two identified Lymnaea feeding motoneurons (Park et al. 1998) or as a co-transmitter at the synapse between the cerebral C2 neuron and the metacerebral giant cell (MGC) in Aplysia, which is homologous to the Lymnaea CGC (Jacklet 1995; Koh and Jacklet 1999). In Aplysia, NO enhances the excitability of the MGC, which potentially enhances the influence of the MGCs on feeding activity (Jacklet and Tieman 2004). Our results add to these findings and show that NO can also enhance the output from the CGCs providing an additional mechanism for the modulation of the effects of serotonergic cerebral giant cells on molluscan feeding behavior by NO.
This work was supported by grants from the Biology and Biotechnology Research Council (UK) to P. R. Benjamin and M. O'Shea.
We thank Prof. N. I. Syed, University of Calgary, for providing V. A. Straub with the opportunity to carry out the CGC-B4 co-culture experiments during a visit supported by a Royal Society (UK) travel grant. We also thank W. Zaidi for technical assistance with the cell culture procedure.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 by the American Physiological Society