Expression of the 5-HT1Apl(a) receptor in Aplysia pleural sensory neurons inhibited 5-HT-mediated translocation of the novel PKC Apl II in sensory neurons and prevented PKC-dependent synaptic facilitation at sensory to motoneuron synapses (Nagakura et al. 2010). We now demonstrate that the ability of inhibitory receptors to block PKC activation is a general feature of inhibitory receptors and is found after expression of the 5-HT1Apl(b) receptor and with activation of endogenous dopamine and FMRFamide receptors in sensory neurons. Pleural sensory neurons are heterogeneous for their inhibitory response to endogenous transmitters, with dopamine being the most prevalent, followed by FMRFamide, and only a small number of neurons with inhibitory responses to 5-HT. The inhibitory response is dominant, reduces membrane excitability and synaptic efficacy, and can reverse 5-HT facilitation at both naive and depressed synapses. Indeed, dopamine can reverse PKC translocation during the continued application of 5-HT. Reversal of translocation can also be seen after translocation mediated by an analog of diacylglycerol, suggesting inhibition is not through blockade of diacylglycerol production. The effects of inhibition on PKC translocation can be rescued by phosphatidic acid, consistent with the inhibitory response involving a reduction or block of production of this lipid. However, phosphatidic acid could not recover PKC-dependent synaptic facilitation due to an additional inhibitory effect on the non-L-type calcium flux linked to synaptic transmission. In summary, we find a novel mechanism downstream of inhibitory receptors linked to inhibition of PKC activation in Aplysia sensory neurons.
- membrane excitability
sensory neurons involved in the defensive withdrawal reflexes of the marine mollusc Aplysia californica have long been studied as modulatory targets for serotonin. Serotonin release onto sensory to motoneuron synapses increases membrane excitability through modulation of a variety of currents [see Barbas et al. (2003) for review] and produces both short-term and long-term facilitation of transmitter release at naive synapses through cAMP production followed by activation of PKA (Kandel 2001). Low-frequency stimulation of Aplysia sensory neurons results in a depression of synaptic transmission at sensory to motoneuron synapses (Castellucci et al. 1970). 5-HT-mediated facilitation following the low-frequency depression is a major mechanistic component of behavioral dishabituation of the defensive withdrawal reflexes (Antonov et al. 1999; 2010). Activation of the novel, calcium-independent PKC in Aplysia Apl II, not PKA, is required for this form of facilitation (Ghirardi et al. 1992; Manseau et al. 2001).
While much of the initial work on serotonin was done on the siphon sensory neurons in the abdominal ganglia, the ventrocaudal (VC) cluster of mechanoafferent somata of the pleural ganglia are most commonly used for culturing Aplysia mechanosensory neurons, as the cluster is quite distinct from the surrounding neurons, allowing easy access for isolation in culture. These sensory neurons innervate the foot, body, and parapodia of the animal and are involved in the head and tail-induced defensive withdrawal reflexes of the animal (Stopfer and Carew 1996; Walters et al. 1983a; b). 5-HT is reported to have very similar effects in both abdominal and pleural sensory neurons (Wright and Kirschman 1995).
Previously, our laboratory reported that overexpression of the 5-HT1Apl(a) receptor in pleural sensory neurons not only resulted in reversing the action of 5-HT on membrane excitability, but also had a profound and unexpected inhibitory effect on PKC activation with 5-HT, preventing the PKC-mediated reversal of synaptic depression (Nagakura et al. 2010). This receptor is known to be expressed only in a very small subset of pleural sensory neurons (Barbas et al. 2005); thus we wondered if any of the other known endogenous inhibitory responses on membrane excitability also had an inhibitory influence on PKC activation with 5-HT. At Aplysia sensory neurons, the inhibitory effects of both dopamine and FMRFamide on membrane excitability have been previously noted (Abrams et al. 1984; Barbas et al. 2006). Here we demonstrate that the inhibitory responses produced by both dopamine and FMRFamide, acting through endogenous receptors, are also capable of strongly inhibiting synaptic efficacy and the reversal of 5-HT-dependent facilitation at both naive and depressed synapses. Our results indicate that inhibition at depressed synapses occurs through at least two mechanisms: 1) a reduction in phosphatidic acid (PA), which is a requirement for the PKC Apl II activation at synapses (Farah et al. 2008), and 2) a reduction in the calcium flux that is responsible for triggering transmitter release.
MATERIALS AND METHODS
Animals and cell culture.
Mariculture-reared Aplysia californica were purchased from The National Resource for Aplysia at the University of Miami. Pleural sensory neurons were isolated from the VC cluster of the pleural ganglia following either a 2-h digestion at 37°C or an 18-h digestion at room temperature with a predetermined concentration of protease in L-15 media. Cells were cultured on poly-l-lysine-coated glass-bottomed culture dishes filled with a culture media of 50% Aplysia hemolymph and 50% L-15 media adjusted with salts to match Aplysia hemolymph. Cells were cultured for 72–96 h at room temperature in a high humidity chamber. Experiments with synapses required pairing isolated pleural sensory neurons with isolated LFS siphon motoneurons from the abdominal ganglion. The identification for isolation of LFS motoneurons from abdominal ganglia was based on location and morphology, and the identity of the LFS motoneurons was further confirmed electrophysiologically before assessing synaptic efficacy, through observing the occurrence of the “notch” potential following a hyperpolarizing pulse, as described in Chitwood et al. (2001). In experiments with dopamine, a 1 mM solution was prepared right before the start of the experiments and used for 1 h or until discoloration indicative of oxidation. FMRFamide (Phe-Met-Arg-Phe-NH2) was produced by Celtek Peptides (Celtek Bioscience, Nashville, TN); 1,2-dioctanoyl-sn-glycerol (DOG) and 1,2-dioctanoyl-sn-glycero-3-phosphate (DiC8-PA) were purchased from Avanti polar lipids (Alabaster, AL) and dissolved in DMSO. All other compounds used were purchased from Sigma-Aldrich. Enhanced green fluorescent protein (eGFP)-5-HT1Apl(a), eGFP-5-HT1Apl(b), and eGFP-PKC Apl II sequences in pNEX3 vectors were microinjected into sensory neuron nuclei 24 h prior to experimentation, as previously described in Farah and Sossin (2011).
Prior to electrophysiological recordings, the culture media was replaced with a Aplysia recording saline containing (in mM) NaCl 460, MgCl2 55, CaCl2 10, KCl 10, HEPES 10, pH 7.6. Presynaptic sensory neuron and postsynaptic motoneuron membrane potentials were recorded and manipulated with intracellular sharp electrodes backfilled with 2 M K-acetate, attached to an Axoclamp 900 (Molecular Devices, Sunnyvale, CA). Membrane potentials were held at −80 mV with current injection. Depolarizing current pulses of 50 ms were used to elicit action potentials in the sensory neuron, and the resultant postsynaptic potential (PSP) recorded in the motoneuron. Electrodes were periodically rebalanced, and input resistance measured with 500-ms, 0.5-nA hyperpolarizing pulses.
Fluo 4 imaging.
Sensory neurons were loaded with micropipettes backfilled with 1 μl of 6 mM membrane-impermeant fluo 4 using −1-nA current pulses. The fluo 4 load was monitored optically, and 15–30 min were allowed between loading and imaging/recording. Fluo 4 fluorescence was imaged with an ex. 470/40 nm-515/50 nm filter set through a 1.3 NA ×40 objective using a Photometrics (Tucson, AZ) QuantEM: 512SC EM CCD camera. Rapid image acquisition was achieved using the Fast Acquisition Solution from Zeiss (Carl Zeiss), involving hardware and software modifications that allow less time delay between subsequent frames. Exposure times were set to 20 ms, and regions of interest chosen to reduce frame size so that at least 20 frames/s were acquired. Fluorescence intensity was transformed into ΔF/F following background subtraction using the average intensity of the 10 frames preceding each action potential as the value for Fo. Peak fluorescence change values (occurring on the first or second frame following the action potential) for three successive action potential-induced fluorescence transients were averaged before and after addition of dopamine for comparison.
All PSP amplitudes are normalized as a percentage of the first PSP observed at the synapse (PSP#1). At synapses with PSPs large enough to activate voltage-dependent currents in the motoneuron, PSP rise rate was measured over 1 ms and normalized to PSP#1. eGFP-PKC Apl II cytosol to membrane translocation was measured and analyzed as described in Farah et al. (2008). For the dopamine and PA groups in Fig. 4, C and D, cells were treated first with DOG (10 μg/ml) for 5 min, OR with DOG (10 μg/ml) and PA (25 μg/ml) for 5 min. Since translocation in the presence of DOG and PA was not different from translocation in the presence of DOG alone, the two groups were pooled for data analysis. All statistical analysis was performed using t-tests unless stated otherwise, and all values are means ± standard error of the mean.
Modulation of pleural sensory neuron membrane excitability.
Previously, our laboratory reported that overexpression of the 5-HT1Apl(a) receptor in sensory neurons resulted in a reduction in excitability with 5-HT, requiring more current to reach action potential threshold and the inability, in many cases, to fire more than one action potential (Nagakura et al. 2010; Fig. 1, Aii, Bii, C, and D). This response with 5-HT is opposite to the primary endogenous response (Fig. 1, Ai, Bi, C, and D). Since the 5-HT1Apl(a) receptor had been previously characterized as negatively coupling to adenylate cyclase (AC) (Angers et al. 1998), we questioned whether the similar 5-HT1Apl(b) receptor, also observed to negatively couple to AC (Barbas et al. 2002) had similar effects, or whether this was a unique property of the 5-HT1Apl(a) receptor. The 5-HT1Apl(b) receptor is not known to be expressed in pleural sensory neurons (Barbas et al. 2005). Therefore, we expressed this receptor with an eGFP tag [eGFP-5-HT1Apl(b)], similar to the tag previously used with the 5-HT1Apl(a) receptor, for visualization of expression levels in pleural sensory neurons. Similar to what was observed with 5-HT1Apl(a) receptor expression, expression of eGFP-5-HT1Apl(b) greatly reduced membrane excitability with addition of 5-HT, increasing the amount of current required to reach threshold and reducing the number of action potentials per unit of depolarizing current (Fig. 1, Aiii, Biii, C, and D). Using a 500-ms hyperpolarizing pulse, membrane input resistance was also observed to decrease with both eGFP-5-HT1Apl(a) and eGFP-5-HT1Apl(b) receptor activation (Fig. 1E). As reported with eGFP-5-HT1Apl(a) expression in pleural sensory neurons paired with motoneurons in culture (Nagakura et al. 2010), eGFP-5-HT1Apl(b) expression in the sensory neuron blocked the facilitation of PSP amplitude with 5-HT following synaptic depression with low-frequency stimulation (Fig. 1F). Thus the inhibitory responses appear to be a general feature of these receptors, and we therefore searched for an endogenous transmitter that could also produce this response.
Inhibition of pleural sensory neuron excitability has been reported with the neuromodulatory peptide FMRFamide (Critz et al. 1991) and also with dopamine (Barbas et al. 2006). We found that both of these modulatory substances had a significant inhibitory effect on sensory neuron excitability: the inhibition with FMRFamide was generally observed to be less than that of dopamine (Fig. 1), which is also apparent when an individual sensory neuron responded to both FMRFamide and dopamine. A decrease in input resistance was also observed with dopamine and with FMRFamide, as with 5-HT when eGFP-5-HT1Apl(a) or eGFP-5-HT1Apl(b) was expressed (Fig. 1E). The increase in the S-K+ current in Aplysia sensory neurons with FMRFamide leads to hyperpolarization of the resting membrane potential (Belardetti et al. 1987). As the sensory neurons were held at −80 mV with negative-current injection, the effect of FMRFamide on the resting membrane potential manifest as a reduction in the amplitude of the holding current (reduction to 66 ± 6% of that before FMRFamide, n = 7). A similar reduction in the holding current was also observed with dopamine, consistent with the two substances acting on similar conductances (holding current reduced to 63 ± 6% of that before dopamine, n = 11).
Heterogeneity of pleural sensory neurons to dopamine and FMRFamide.
The reduction in pleural sensory neuron excitability with dopamine only occurs in some of the pleural sensory neurons (Fig. 2A). We found that the inhibitory response to dopamine was maximal at 500 nM, well below the reported activation of 5-HT1Apl(a) with dopamine, which has a minimal activation of the 5-HT1Apl(a) at concentrations above 10 μM (Angers et al. 1998). To ensure that the inhibitory response to dopamine is not a result of dopamine acting on endogenous 5-HT1Apl(a) receptors, dopamine up to 20 μM was applied to sensory neurons overexpressing eGFP-5-HT1Apl(a). While 5-HT led to extensive inhibition of excitability in all cells expressing eGFP-5-HT1Apl(a), dopamine only led to inhibition of excitation in some of the sensory neurons, indicating that 5-HT1Apl(a) receptor does not mediate the inhibitory dopamine response (Fig. 2B). Expression of the eGFP-5-HT1Apl(a) receptor in sensory neurons allowed us to confirm that this receptor is activated by nanomolar concentrations of the selective agonist 8-hydroxy-2(di-n-propylamino)tetralin (8-OH-DPAT), as previously reported (Angers et al. 1998). Activation of the eGFP-5-HT1Apl(a) receptor with 500 nM 8-OH-DPAT results in a large reduction in excitability in all neurons expressing the receptor (Fig. 2C), similar to what was observed with 5-HT on eGFP-5-HT1Apl(a)-expressing neurons (see Fig. 1). Conversely, 8-OH-DPAT on pleural sensory neurons not expressing the eGFP-5-HT1Apl(a) only rarely shows a decrease in excitability, which is likely the result of activation of an endogenous 5-HT1Apl(a) response (circled in Fig. 2C). The reduction in excitability with 8-OH-DPAT occurs in all cells expressing eGFP-5-HT1Apl(a), confirming the efficacy of the agonist and indicating that dopamine is not activating a potential endogenous 5-HT1Apl(a) response, but rather acting through another receptor.
Examination of many neurons isolated from the same pleural pedal ganglia revealed that the inhibitory dopamine response varied from 10 to 90% of the sensory neurons examined, with, on average of the 11 animals examined (99 sensory neurons in total), 58.2 ± 0.1% showing a reduction in excitation with dopamine (Fig. 2D). Next, greater numbers of neurons were examined with 500 nM 8-OH-DPAT to estimate the percentage of pleural sensory neurons with an endogenous 5-HT1Apl(a) response. In the animals used in this experiment, the percentage of pleural sensory neurons with the inhibitory dopamine response was similar to the previous estimates at 50%, while the percentage of neurons with an endogenous 5-HT1Apl(a) response (sensitive to 500 nM 8-OH-DPAT) was much lower at 5–7% (Fig. 2E). To estimate the percentage of pleural sensory neurons responsive to FMRFamide with a reduction in excitability, sensory neurons isolated from the same ganglion were examined first for FMRFamide sensitivity with 20 μM FMRFamide and then for dopamine sensitivity with 1 μM dopamine. Approximately one-half of the examined pleural sensory neurons were inhibited by FMRFamide (41.2 ± 5.9%), slightly less than the number of sensory neurons responsive to dopamine (59.4 ± 13.1% in this experiment). Many pleural sensory neurons respond to both FMRFamide and dopamine, although the extent of inhibition is generally weaker with FMRFamide (as in Fig. 1). These data indicate that pleural sensory neurons are heterogeneous with respect to their sensitivity to modulatory substances, with dopamine producing the most prominent inhibitory effect on membrane excitability.
Dopamine and FMRFamide regulate the recovery from synaptic depression.
Following the generation of 40 action potentials at low-frequency stimulation (0.05 Hz) of the sensory neuron, PSP amplitude decreases to ∼20% of the initial amplitude, and addition of 5-HT to the depressed synapse leads to facilitation. Both eGFP-5-HT1Apl(a) and eGFP-5-HT1Apl(b) expression in sensory neurons inhibited the PKC-mediated reversal of synaptic depression with 5-HT (Nagakura et al. 2010; Fig. 1). Furthermore, dopamine also blocked the reversal of depression with 5-HT (data not shown); thus we wondered if the 5-HT reversal of depression could be inhibited or reversed once already initiated. After only four 5-HT-facilitated PSPs, dopamine was added in combination with 5-HT. As observed in the above experiments, not all pleural sensory neurons (now paired with LFS motoneurons) responded to dopamine with a decrease in membrane excitability. Therefore, synaptic pairs could be grouped according to the change in sensory neuron excitability with dopamine (Fig. 3A). Addition of dopamine had no effect on 5-HT facilitation at depressed synapses if the sensory neuron did not also show a decrease in membrane excitability (Fig. 3, A–D). However, if membrane excitability of the sensory neuron was inhibited with 500 nM dopamine, then rapid synaptic depression also occurred (Fig. 3, A–D). Thus dopamine can reverse 5-HT-mediated facilitation at depressed synapses once already initiated (Fig. 3).
With the similar effects of 5-HT1Apl(a), 5-HT1Apl(b), and dopamine on membrane excitability and 5-HT facilitation at depressed synapses, we wondered whether FMRFamide, which also decreases membrane excitability (Fig. 1) and depresses synaptic transmission at naive synapses (Guan et al. 2003), could reverse 5-HT facilitation at depressed synapses. Again, since not all pleural sensory neurons respond to FMRFamide, individual synaptic pairs could be separated for analysis of the effect of FMRFamide on sensory neuron membrane excitability. Five of the sensory neurons showed no change in excitability with FMRFamide (95.1 ± 3.4% of current required before FMRFamide), while the other five synaptic pairs showed a large reduction in excitability requiring 230.2 ± 30.2% more current to reach action potential threshold after FMRFamide (represented in Fig. 3E, inset). And similar to what was observed with dopamine, 5-HT facilitation at depressed synapses was reversed at synaptic pairs with FMRFamide-sensitive sensory neurons only (Fig. 3E). Since the inhibitory responses produced by both dopamine and FMRFamide are variable between pleural sensory neurons, the correlation between the change in membrane excitability with the change in synaptic efficacy was examined (Fig. 3F). A significant correlation was found between the change in the amount of current required to reach action potential threshold (change in membrane excitability) and the reduction in PSP amplitude during 5-HT-mediated facilitation at depressed synapses with dopamine and FMRFamide (n = 17, P < 0.0001). Therefore, similar to activation of 5-HT1Apl(a), 5-HT1Apl(b), the endogenous inhibitory dopamine and FMRFamide responses are also capable of reversing the 5-HT facilitation at depressed synapses.
Dopamine reverses PKC Apl II translocation with 5-HT and DOG.
5-HT-mediated facilitation at naive synapses requires cAMP, whereas 5-HT facilitation at depressed synapses requires the calcium-independent PKC, Apl II (Ghirardi et al. 1992; Manseau et al. 2001). Our laboratory previously reported that eGFP-5-HT1Apl(a) overexpression inhibited 5-HT-dependent translocation of the novel PKC Apl II (Nagakura et al. 2010), suggesting the inhibitory effect of 5-HT1Apl(a) is upstream of PKC activation. Since dopamine could reverse the PKC-mediated reversal of depression, we wondered if dopamine could reverse the 5-HT-mediated eGFP-PKC Apl II translocation from the cytosol to the plasma membrane. Cultured pleural sensory neurons expressing eGFP-PKC Apl II show a redistribution of eGFP-PKC Apl II from the cytosol to the plasma membrane with application of 5-HT (the eGFP membrane-to-cytosol ratio after 5-HT in the control group was 1.87 ± 0.10 and 2.51 ± 0.18 in group yet to receive dopamine) (Fig. 4A). Subsequent application of dopamine in the continued presence of 5-HT resulted in a significant reduction in eGFP-PKC Apl II translocation compared with neurons not treated with dopamine (P < 0.001, Fig. 4). Translocation of PKC Apl II downstream of 5-HT requires production of both diacylglycerol (DAG) by phospholipase C (PLC) and PA by phospholipase D (PLD) (Farah et al. 2008). To test if dopamine was acting upstream of PLC, we examined whether dopamine could still reverse translocation when a cell-permeable, exogenous DAG analog, DOG, was used. Activation and translocation of eGFP-PKC Apl II with DOG could also be reversed with addition of dopamine, although this was slower than the dopamine reversal of 5-HT-dependent translocation (Fig. 4, C and D). These results suggest that dopamine was not acting on membrane DAG levels. To test if dopamine was acting on membrane PA levels, we examined whether dopamine could reverse translocation in the presence of a cell-permeable analog of PA, DiC8-PA. Unlike with DOG alone, the dopamine-induced reversal of eGFP-PKC Apl II translocation was inhibited by the combination of DOG and DiC8-PA (Fig. 4, C and D). This suggests that dopamine is inhibiting or inactivating PKC Apl II activation at least in part through a reduction in plasma membrane PA levels.
Since adding DiC8-PA rescued the inhibition of eGFP-PKC Apl II translocation, we attempted to prevent the dopamine-mediated inhibition of 5-HT facilitation at depressed synapses with preapplication of 25 μg/ml of DiC8-PA. Though this concentration was sufficient to prevent the inhibitory effect of dopamine on eGFP-PKC Apl II translocation with DOG, it failed to alter the inhibitory action of dopamine on the 5-HT-mediated reversal of depression (Fig. 4, E and F). This suggests that dopamine has additional actions on inhibiting transmitter release independently of PA production.
Facilitation and depression at naive synapses.
In the typical pleural sensory neuron, 5-HT not only increases excitability (as in Fig. 1i), and increases transmitter release at depressed synapses, but also increases transmitter release at naive synapses (no stimulation history since isolation and culture) through an independent mechanism (Ghirardi et al. 1992). In contrast, FMRFamide is reported to depress PSP amplitude at naive synapses (Edmonds et al. 1990; Guan et al. 2003; Montarolo et al. 1988). Whereas only a few published papers mention the short-term inhibitory effect of dopamine on synaptic efficacy (Abrams et al. 1984; Montarolo et al. 1988), none present any form of data as to the extent or mechanism of inhibition, and no reports involve the pleural sensory neurons. To examine the effect of dopamine on naive synapses, PSP amplitude was first measured with a single action potential, and then either 5-HT or dopamine was added to modulate PSP amplitude. Following 2 min, PSP amplitude was measured again with a second action potential in the sensory neuron and the amplitude of PSP#2 expressed as a percentage of PSP#1. While 5-HT produced facilitation, dopamine produced a dramatic depression in PSP amplitude. However, this occurred only if membrane excitability of the sensory neuron was also inhibited with dopamine (Inh DA Resp, had, on average, a 209 ± 38% change in the amount of current required to reach threshold, Fig. 5, A and B). In a second experiment, following measurement of PSP#1 and addition of 5-HT, facilitation was measured with PSP#2 after 5 min in 5-HT. Then dopamine was added, and a third PSP measured following an additional 2 min with now dopamine and 5-HT. When dopamine did not reduce excitability of the sensory neuron (No DA Resp, in this group the amount of current required to fire an action potential with a 50-ms pulse was 88 ± 7% of what was required before dopamine), PSP amplitude remained facilitated with dopamine for a third PSP (Fig. 5C). Conversely, if dopamine produced a reduction in membrane excitability (Inh DA Resp, in this group the amount of current required to fire an action potential in dopamine was, on average, 690 ± 216% of the amount required before dopamine), PSP amplitude was greatly depressed (Fig. 5C). Thus dopamine results in depression that can overcome or reverse 5-HT facilitation, much like that observed at depressed synapses. These data show that the pleural sensory neuron-specific inhibitory dopamine response includes a dramatic inhibitory effect on synaptic efficacy, independent of prior stimulation history. Similarly, eGFP-5-HT1Apl(a) expression alters the response of synapses to 5-HT from facilitation to depression [eGFP-5-HT1Apl(a) expression at two synaptic pairs resulted in a block of PSP amplitude with 5-HT, average initial PSP amplitude was 9.6 mV].
The inhibitory dopaminergic response includes a reduction in the non-L-type calcium flux.
The mechanism of FMRFamide mediated inhibition of synaptic transmission remains in contention. A previous paper (Edmonds et al. 1990) reported a reduction in the non-L-type calcium current with FMRFamide, and since this current is responsible for transmitter release, we wondered whether dopamine also affected this calcium flux. The L-type calcium current in Aplysia is not involved in transmitter release and can be selectively blocked with dihydropyridines (Braha et al. 1993). To selectively examine the relative calcium flux through the channels responsible for transmitter release, we iontophorectically preloaded the sensory neurons with membrane-impermeant fluo 4 and blocked the L-type flux with 5 μM nifedipine. With a sharp electrode in the sensory neuron to evoke single action potentials, fluo 4 imaging was conducted such that >20-Hz frame rates were achieved of regions of interest at axon ends or at the contacts between the sensory neuron axon/neurite and that of the LFS motoneuron (Fig. 6A). These subcellular regions were chosen as previous observations of calcium flux modulation with 5-HT observed the changes to be localized to these regions (Eliot et al. 1993; Leal and Klein 2009). A single action potential in the sensory neuron produced a fluorescence transient that was greater in some regions, and these were selected for measuring intensity change (Fig. 6B). Three successive action potentials, generated by brief 50-ms depolarizing pulses, produced individual fluo 4 fluorescent transients of similar amplitude (Fig. 6, C and D). As in previous experiments, sensory neurons were grouped according to the change in membrane excitability with dopamine for analysis. In sensory neurons where dopamine had no effect on membrane excitability, the fluo 4 fluorescence transient was unaffected when three more action potentials were generated following the addition of dopamine (Fig. 6C). However, when dopamine reduced excitability, the fluo 4 transients were also greatly reduced (Fig. 6D). Therefore, one of the actions of dopamine is to inhibit the calcium flux responsible for transmitter release.
The excitatory and facilitatory properties of 5-HT at Aplysia sensory to motoneuron synapses are usually the focus of examination, as these mechanisms underlie the sensitization of the behavioral defensive withdrawal reflexes with noxious stimuli, a very simple and highly accessible behavior for neurophysiological examination (Antonov et al. 1999). Although examined less frequently, the inhibitory responses are no less dramatic in affecting both membrane excitability and transmitter release.
Inhibitory (D2-like) dopaminergic reduction of transmitter releasing CaV2 calcium currents has been well documented, observed in a variety of preparations, indicating a conserved mechanism (Kline et al. 2009; Missale et al. 1998; Ramanathan et al. 2008; Salgado et al. 2005; Wikstrom et al. 1999; Zhang et al. 2004). In Aplysia, stimulation of a connective to the abdominal ganglia is known to reduce the calcium current in the abdominal neuron L10 and PSP amplitude in follower cells, suggesting that release of a substance reduces transmitter release by reducing calcium flux in this neuron (Shapiro et al. 1980). In abdominal LE sensory neurons, inhibition of excitability with both FMRFamide and dopamine was described by Abrams et al. (1984), and although they state that both transmitters also reduce synaptic transmission, no data are presented. Montarolo et al. (1988) also state that dopamine results in short-term depression of synaptic efficacy, but again present no data to describe the extent of the inhibition. Here we describe the extent of the inhibitory response and relative frequency of the different inhibitory responses produced by different neuromodulators in pleural sensory neurons.
We show the relative frequency of a variety of short-term inhibitory responses observed at Aplysia sensory neurons isolated from the VC cluster of the pleural ganglion. The strongest and most prominent inhibitory response observed was produced with dopamine in a majority of sensory neurons examined (Figs. 2 and 3). The inhibitory effect of dopamine on membrane excitability was very strong, increasing the amount of current required to reach threshold, while also reducing input resistance and the resting membrane potential. Thus, in the presence of dopamine, this subset of pleural sensory neurons is unlikely to generate more than one action potential per stimulus, regardless of the stimulus intensity. In addition, dopamine also had a profound inhibitory effect on synaptic efficacy (Figs. 3 and 5), further reducing the likelihood of a stimulus to the mechanosensory neuron exciting the motoneuron.
At both naive and depressed synapses, the inhibitory action of dopamine was dominant to excitatory effects of 5-HT, able to reverse the 5-HT-mediated facilitation and recovery from depression, with the former requiring PKA and the latter a form of synaptic plasticity requiring the activation of PKC (Ghirardi et al. 1992; Liu et al. 2004; Manseau et al. 2001). Therefore, as the typical response of the pleural sensory neuron to 5-HT is an increase in excitability and synaptic efficacy, dopamine produces the opposite response with a reduction in excitability and synaptic efficacy. While FMRFamide also produces a similar reduction in excitability and synaptic efficacy, the dopaminergic inhibitory response is stronger (Fig. 1) and more prevalent (Fig. 2) in pleural sensory neurons. The dominance of the inhibitory response of FMRFamide over 5-HT had been reported previously (Critz et al. 1991); here we extend this observation to depressed synapses and include the 5-HT1Apl(a) and 5-HT1Apl(b) responses and dopamine as producing a similar mechanism of inhibition.
Heterogeneity of the inhibitory responses in pleural sensory neurons.
The VC cluster pleural sensory neurons innervate the tail, foot, head, body, and the posterior parapodia of the animal, acting as primary mechanoreceptors (Walters et al. 2004; Walters et al. 1983a). Some of these sensory neurons innervate the tail, so that when the tail is stimulated, a defensive withdrawal of the tail occurs through direct excitation of the tail motoneurons in the pedal ganglion. The noxious stimulus to the tail results in serotonin release in local regions of a variety of ganglia, including the pleural ganglia (Marinesco and Carew 2002; Marinesco et al. 2004a; Marinesco et al. 2006; Marinesco et al. 2004b), resulting in increased excitability and facilitation of transmitter release from the tail sensory neurons (Ghirardi et al. 1992). Conversely, a noxious stimulus to the head (which leads to a defensive withdrawal of the head), inhibits the siphon withdrawal reflex through an inhibitory action of 5-HT at interneuron to motoneuron synapses (Marinesco et al. 2006). A small subpopulation of VC pleural neurons expresses the inhibitory 5-HT1Apl(a) receptor (Barbas et al. 2005). From the data presented in Fig. 2 and in Nagakura et al. (2010), these pleural sensory neurons would be inhibited with 5-HT, a response also reported in RF abdominal sensory neurons (Storozhuk and Castellucci 1999) and some cerebral sensory neurons (Rosen et al. 1989). In support of these observations, in situ hybridizations indicate 5-HT1Apl(a) or 5-HT1Apl(b) receptor expression in a variety of abdominal sensory neurons and ∼20% of the cerebral sensory neurons (Barbas et al. 2005). Thus, like the abdominal and cerebral ganglia, the pleural ganglia contain a small subset of mechanosensory neurons that are inhibited by 5-HT.
The heterogenous response of pleural sensory neurons to dopamine reported here was also observed between the different sensory neuron clusters of the abdominal ganglia, with the RF sensory neurons showing less of an inhibitory change in membrane excitability with dopamine than sensory neurons in the other clusters (i.e., LE, rLE, RE) (Dubuc and Castellucci 1991). Furthermore, our observation that about one-half of the pleural sensory neurons show no response to FMRFamide is supported by the observations of Buttner and Siegelbaum (2003); however, they suggest this may be due to a requirement for an initial excitation for inhibition to counter. Our data contradict this possibility as FMRFamide-insensitive sensory neurons were incapable of counteracting 5-HT-mediated facilitation following depression (Fig. 3E). We also observed that some pleural sensory neurons showed a response to dopamine, but not to FMRFamide, and vice versa, further indicating that the lack of response was more likely due to a lack of the specific receptor rather than the state of the neuron. The responses we observed to FMRFamide and dopamine were largely heterogeneous, indicating that, like the cerebral and abdominal sensory neurons, the pleural ganglion VC cluster of mechanosensory neurons is a heterogeneous population of neurons in their specific sensitivity to neuromodulatory substances. Thus the inhibitory changes to pleural sensory neuron excitability and synaptic efficacy will differ, depending on the specific neuron examined.
The strong inhibition of a majority of the pleural sensory neurons with dopamine predicts that dopamine release in the pleural ganglia would strongly inhibit body-induced defensive withdrawal reflexes. The appetitive behaviors of Aplysia are regulated by dopamine, such that dopamine in the hemocoel or from increased dopaminergic neuronal release induces feeding behaviors (Kabotyanski et al. 2000). Indeed, there is evidence that dopamine release mediates the reinforcement of both operant and associative conditioning to food in Aplysia (Baxter and Byrne 2006; Lorenzetti et al. 2008). Although habituation of the siphon and gill withdrawal reflexes are reported to be inhibited with perfusion of the gill with dopamine (Ruben and Lukowiak 1983), a variety of defensive responses of the animal that would involve pleural sensory neurons, including siphon withdrawal, locomotion after salt stimulation, and inking responses, were all observed to be highly attenuated 0.5 h after feeding (Advokat 1980). The cerebral-pedal regulator neurons are activated with food-induced arousal and act to reduce defensive reflexes, including suppression of the head withdrawal reflex through inhibitory action in the pleural/pedal ganglia (Teyke et al. 1990). Dopamine has been observed in the neuropile of pedal ganglia (McCaman et al. 1973), the location of the pleural sensory-to-tail motoneuron synapses (Wainwright et al. 2002); whether release is modulated by appetitive behavior has not been assessed. State-dependent changes in the defensive responses may represent the physiological relevance of the inhibitory responses observed in the pleural sensory neurons. Alternatively, the inhibitory responses may represent endogenous hypoalgesic mechanisms serving to prevent activation of withdrawal reflexes in favor of other defensive behaviors. A noxious stimulus to the tail of the animal alters subsequent defensive withdrawal of the siphon in favor of siphon flaring for directed inking toward the tail (Illich et al. 1994). Furthermore, extrinsic inhibition of pleural sensory neuron action potential discharge has been observed previously, indicating that, under some noxious stimulation patterns, inhibition of sensory neuron activity occurs (Clatworthy and Walters 1993).
The pleural sensory neuron inhibitory 5-HT response from endogenous 5-HT1Apl(a) only occurs in a very small subset of pleural sensory neurons (Fig. 2; Barbas et al. 2005). The majority of pleural sensory neurons respond in an excitatory manner with 5-HT, so the physiological function of the inhibitory response in this small subset of neurons is unclear, but may reflect a more complicated or alternate function, as suggested for the RF abdominal sensory neurons that are also inhibited by 5-HT (Storozhuk and Castellucci 1999). On the other hand, the inhibitory dopamine and FMRFamide responses likely function to modulate the body-induced defensive withdrawal reflexes in a state-dependent manner, as suggested above.
Reversing PKC activation and inhibition of the non-L-type calcium flux.
Facilitation with 5-HT following synaptic depression requires activation of calcium-independent PKC, Apl II (Manseau et al. 2001). A fluorescent protein tagged PKC Apl II, eGFP-PKC Apl II translocates from the cytosol to the plasma membrane when activated with 5-HT (Farah et al. 2008; Zhao et al. 2006). Recently, our laboratory reported that 5-HT-mediated translocation of eGFP-PKC Apl II is inhibited with overexpression of the 5-HT1Apl(a) receptor (Nagakura et al. 2010). Here we extend the list of receptors capable of inhibiting PKC activation to the very similar 5-HT1Apl(b) receptor, but also to the receptors producing the endogenous inhibitory responses to dopamine and FMRFamide. Application of dopamine following the reversal of depression with 5-HT demonstrates that the dopamine inhibitory response is able to reverse or overpower the 5-HT/PKC-mediated facilitation (Fig. 3). eGFP-PKC Apl II membrane translocation with 5-HT is reversed with dopamine, suggesting that the effect on facilitation following depression is a reversal of the facilitation through inactivation of the novel PKC Apl II (Fig. 4). Translocation of eGFP-PKC Apl II with 5-HT involves a synergistic interaction of the two membrane metabolites DAG and PA to localize the active kinase at the plasma membrane (Farah et al. 2008). PKC Apl II translocation with DOG is not inhibited with dopamine, if PA is included, indicating that the impact of dopamine on PKC Apl II translocation is likely through reducing plasma membrane PA levels. However, the same addition of PA was unable to prevent the inhibitory effect of dopamine on synaptic efficacy at both naive and depressed synapses. Similar to what has been previously observed with FMRFamide, dopamine strongly reduces the non-L-type calcium flux that is responsible for synaptic transmission. The mechanism of the calcium current inhibition was not examined here, but may be due to direct action on the channel or indirectly through a failure of the action potential to sufficiently depolarize the presynaptic terminals. While in many cases the inhibitory dopaminergic response is insufficient to fully block synaptic transmission, indicating that action potential propagation to at least some presynaptic terminals is maintained (as in the traces displayed in Fig. 3), the large reduction in synaptic efficacy with dopamine is likely a consequence of the large reduction in presynaptic calcium flux. We do not think the deficit of PKC Apl II translocation is due to the decrease in the calcium current, since PKC Apl II is not sensitive to calcium (Sossin et al. 1996) and since exogenous PA can rescue the translocation without rescuing transmitter release. Thus the inhibitory response is complicated, acting on membrane PA levels, either through PLD, or one of the lipid metabolic proteins, as well as acting on at least one more separate mechanism to reduce membrane excitability and calcium entry.
Dissociation of inhibitory responses from inhibition of AC.
Both the 5-HT1Apl(a) and 5-HT1Apl(b) receptors are known to negatively couple to AC when expressed in cell lines (Angers et al. 1998; Barbas et al. 2002), as expected for Gαi-coupled receptors (Taussig et al. 1993). This suggests that reducing cellular cAMP may result in the observed inhibitory response in these neurons. However, overexpression of an isoform of the cAMP phosphodiesterase apPDE4 that inhibits facilitation at naive synapses and long-term facilitation, both of which require cAMP increases, did not alter 5-HT facilitation at depressed synapses (Jang et al. 2011). Overexpression of the phosphodiesterase should reduce any basal cAMP levels and, therefore, reduce the modulatory range for any potential signaling through a reduction of cAMP levels. This is also consistent with the findings of Lee et al. (2009), where removal of the 5-HT receptor coupled to AC removed facilitation at naive synapses, but did not affect 5-HT facilitation at depressed synapses, and Chang et al. (2003), where expression of an exogenous AC-coupled octopamine receptor allowed for octopamine facilitation at naive, not depressed, synapses.
The work of Jang et al. (2011) also suggests that the inhibitory mechanism acting on synaptic strength and excitability in pleural sensory neurons is also unlikely to be signaled through reducing cAMP levels, as they report no decreases in initial synaptic strength or in membrane excitability with apPDE4 (only the 5-HT-mediated increases are impaired). Bernier et al. (1982) observed inhibitory changes in membrane excitability with both 5-HT and dopamine in abdominal neuron R15, concurrent with an increase in cAMP, also in support of an inhibitory mechanism that is not mediated through a downregulation in cAMP. Thus the evidence is that many of the actions noted here are independent of the inhibition of AC.
There is no direct evidence for Gαi inhibition of PKC activation, although inhibition of PLC-ϵ was noted with Gαi activation by muscarinic acetylcholine receptors expressed in human embryonic kidney-293 cells (vom Dorp et al. 2004). There is precedence for Gαi receptors coupling to Src → RAS pathway (Ma and Huang 2002) and to RAP → Erk pathway (Weissman et al. 2004) independent of cAMP modulation. Interestingly, neurite outgrowth in vertebrate neurons can also be regulated by 5-HT1 and inhibitory D2 dopamine receptors through a noncanonical Gαi/o coupling to RAP1GAP (reviewed in Ma'ayan et al. 2009). However, the relevance of these pathways for inhibiting excitability and blocking PKC Apl II translocation is not clear. In sensory neurons, the activated G proteins and subsequent second-messenger pathways that produce the common inhibitory responses described here remain unknown, and the mechanisms of these inhibitory responses appear to be complex, involving a variety of targets.
The molecular identity of the inhibitory dopamine receptor (or the inhibitory FMRFamide receptor) is not clear. A dopamine receptor positively coupled to AC has been cloned from Aplysia, but this receptor is not expressed in sensory neurons (Barbas et al. 2006). Bioinformatic screens of the Aplysia expressed sequence tags and genomic sequences resulted in identification of a second dopamine receptor orthologous to other invertebrate dopamine receptors (DinvApl) (Nagakura et al. 2010), but this receptor is also postulated to be positively coupled to AC (Han et al. 1996). No D2-like receptor was identified in the available Aplysia genomic sequences, but D2-like receptors are found in the mollusk Lottia (Nagakura et al. 2010), and D2-like responses are seen in the mollusk Lymnaea (Dobson et al. 2006; Werkman et al. 1987). Since the Aplysia genome is still incomplete, it is most likely that Aplysia also has an ortholog of a D2-like receptor, and this receptor most likely underlies the inhibitory responses observed.
This work was supported by Canadian Institutes of Health Research (CIHR) Grant MOP 12046 to W. S. Sossin, and a CIHR fellowship to T. W. Dunn. W. S. Sossin is a James McGill Scholar.
No conflicts of interest, financial or otherwise, are declared by the author(s).
Author contributions: T.W.D. and W.S.S. conception and design of research; T.W.D. and C.A.F. performed experiments; T.W.D. and C.A.F. analyzed data; T.W.D., C.A.F., and W.S.S. interpreted results of experiments; T.W.D. and C.A.F. prepared figures; T.W.D. drafted manuscript; T.W.D., C.A.F., and W.S.S. edited and revised manuscript; T.W.D., C.A.F., and W.S.S. approved final version of manuscript.
- Copyright © 2012 the American Physiological Society