Augmentation Of Bursting Pacemaker Activity By Serotonin In An Identified Achatina Fulica Neurone: An Increase In Sodium- And Calcium-Activated Negative Slope Resistance Via Cyclic-Amp-Dependent Protein Phosphorylation

In molluscan neurones, such as those of Aplysia californica (Barker and Smith, 1976, 1978; Smith et al. 1975; Wilson and Wachtel, 1978), and Euhadra peliomphala neurones (Onozuka et al. 1986, 1988a, 1991a,b), it has been demonstrated that bursting activity is accompanied by a negative slope resistance (NSR) region in the steady-state current–voltage (I–V) curve of these neurones. It has been proposed that this NSR region plays an essential role in the maintenance of slow membrane oscillations during bursting activity (Wilson and Wachtel, 1978), although the reason for the involvement of an NSR in bursting activity is not understood (Epstein and Marder, 1990). Recently, we identified in the suboesophageal ganglion of Achatina fulica a spontaneously bursting neurone (PON) possessing a similar I–V relationship (Funase, 1990a). In this neurone, serotonin (5-HT) has been found to bring about an increase in the magnitude of this NSR and the induction of another NSR at more positive voltages on the I–V curve, as well as an enhancement of bursting pacemaker activity (for a brief report, see Funase, 1990b). However, the underlying intracellular mechanism(s) responsible for 5-HT augmentation of bursting activity with NSRs has yet to be elucidated.

It has been shown that 5-HT elevates the level of adenosine 3’,5’-cyclic monophosphate (cyclic AMP) in some molluscan neurones (Greenberg et al. 1987; Sweatt et al. 1989). An increase in the level of intracellular cyclic AMP ([cyclic AMP]_i), in turn, causes an activation of a protein kinase that has multiple effects on the membrane properties of these neurones (Baxter and Byrne, 1989, 1990; Levitan et al. 1987). Since one of the major effects of cyclic AMP is mediated by protein phosphorylation (for reviews, see Browning et al. 1985; Greengard, 1976; Kostyuk, 1986; Krebs and Beavo, 1979), it is conceivable that 5-HT may be involved in the augmentation of the bursting pacemaker activity through cyclic-AMP-dependent protein phosphorylation. In the present study, we tested this possibility by examining the effects of cyclic-AMP-dependent protein kinase inhibitors on 5-HT action on the membrane properties of the PON from Achatina fulica, using several pharmacological agents and the voltage-clamp technique in combination with pressure injection.

The suboesophageal ganglion of Achatina fulica was isolated and pinned to the Sylgard base of a 0.1ml perfusion chamber. After a 15-min incubation with 0.7% trypsin (type III, Sigma Chemical Co.) at room temperature (22±1°C), the connective tissue sheath overlying the dorsal surface of the ganglion was dissected with microforceps. All experiments were performed on the PON neurone (Funase, 1990a,b). In some experiments the neurone was axotomized by cutting the septum between the visceral and right parietal ganglion to exclude synaptic inputs. This axotomy was performed at the time the preparation was set up for recording.

The neurone was superfused at a flow rate of 3mlmin⁻¹ with normal saline containing (in mmoll⁻¹): 80 NaCl, 4 KCl, 10 CaCl₂, 5 MgCl₂, 10 Tris–HCl (pH7.4). Na⁺-free saline was made by substituting Tris for Na⁺ and Ca ²⁺-free saline by replacing Ca²⁺ with an equimolar amount of Co²⁺.

The arrangements for recording, stimulating and voltage-clamp circuitry were conventional. A voltage detection electrode and current injection electrode using a Nihon-Kohden voltage-clamp amplifer (CEZ-1100) were employed as described elsewhere (Onozuka et al. 1986, 1988a). A neurone was kept at a holding voltage (V_h) of –50mV and stepped to various membrane potentials (5s in duration). The time between successive voltage steps was at least 30s to minimize possible interference from the preceding change of voltage. Steady-state current was measured at the end of the 5-s step, and leakage current was subtracted, assuming that the leak was linear over the voltage steps used.

The third microelectrode was used for pressure injection of test materials into the neurone according to our previous method (Funase, 1990a,b; Onozuka et al. 1988b, 1991b). The electrode was filled with 0.1mol l⁻¹ KCl (pH7.2) containing cyclic-AMP-dependent protein kinase inhibitor isolated from bovine muscle (PKI; type V) or cyclic AMP and 0.1% (w/v) Fast Green dye. A pressure of 2kgcm⁻² of 1s duration was used, except when otherwise specified. To test the effect of pressure per se, a solution without test material was injected into the neurone. No significant change was noted in the I–V curve or the electrical activity.

Isobutylmethylxanthine (IBMX), cyclic AMP, tetraethylammonium chloride (TEA⁺) and PKI were obtained from Sigma Chemical Co. and isoquinolinesulphonamide (H-8), a membrane-permeable inhibitor of cyclic-AMP-dependent protein kinase, from Seikagaku Kogyo Co. IBMX and TEA⁺ were dissolved in normal saline.

In agreement with a previous finding (Funase, 1990a), a PON neurone in normal saline usually exhibited bursting pacemaker activity characteristic of a slow potential oscillation (Fig. 1A). Under voltage-clamp, this neurone (N=15) showed an inward current over a wide range of membrane potentials, giving a dip in the I–V curve at membrane potentials between –40 and –20mV, characteristic of an NSR (Fig. 1B). This NSR will hereafter be referred to as the first NSR.

When the PON neurone (N=10) was superfused with 5-HT, the amplitude of both the depolarizing and the post-burst-hyperpolarizing phases of the bursting cycle was increased, although the duration of both phases was irregular (Fig. 2A). The I–V relationship shifted downwards with increasing concentrations of 5-HT, and the peak inward current, which gives the first NSR, became more pronounced (Fig. 2B). In addition, 5-HT produced another downward dip in the I–V relationship (hereafter referred to as the second NSR) at membrane voltages between –20 and 0mV. The magnitude of the second NSR was also dose-dependent. These effects of 5-HT were virtually reversible upon rinsing with 5-HT-free normal saline (d in Fig. 2A,B). Essentially similar results were obtained with axotomized cells (N=5) (data not shown), indicating that the action of 5-HT upon the neurone is direct and not mediated synaptically. Levels of 5-HT below 0.5 μmol l⁻¹ did not affect either the membrane potential or the I–V curve. In subsequent experiments, 10 μmol l⁻¹ 5-HT was used, unless otherwise specified, since this concentration always brought about both first and second NSRs.

The ionic mechanism of the action of 5-HT on bursting activity was examined by removing extracellular Na⁺ or Ca²⁺, since both Na⁺ and Ca²⁺ influx contribute to the rising phase of the action potential in molluscan neurones (Adams et al. 1980a,b). In addition, the I–V relationship of the net 5-HT current (I_5-HT) was obtained from the difference between the currents in the presence and absence of this bioamine in the perfusate. Filled circles in Fig. 3 show the I–V relationships in which the second NSR was larger than the first NSR. When the neurone (N=7) was superfused with Na⁺-free saline containing 5-HT, the first NSR completely disappeared, whereas the second NSR was reduced by about 50% (filled triangles in Fig. 3). In contrast, the second NSR disappeared in response to perfusion with Co²⁺-substituted Ca²⁺-free saline, while the first NSR was reduced (filled squares in Fig. 3).

Since it has been suggested that augmentation of the post-burst hyperpolarization by 5-HT in the bursting pacemaker neurone of Aplysia californica might be due to an increase in the outward K⁺ conductance (DePeyer et al. 1982), we also examined the effect of TEA⁺ on the 5-HT-induced I–V relationship. As shown in Fig. 4, TEA⁺ (N=5) slightly increased the magnitude of the first NSR and enhanced the size of the second NSR in a dose-dependent manner. In addition, the membrane potential at the peak of the second NSR was progressively shifted with increasing concentrations of TEA⁺ towards the voltage at which the largest Ca²⁺ current has been shown to be generated in snail neurones (Onozuka et al. 1990).

Because several actions of 5-HT in molluscan neurones are mediated through cyclic AMP which, in turn, activates a protein kinase (Deterre et al. 1981; Drummond et al. 1980; Levitan et al. 1987; Levitan and Levitan, 1988; Swandulla and Lux, 1984; Siegelbaum et al. 1982; Taussig et al. 1989), we hypothesized that cyclic AMP may mimic the action of 5-HT on the I–V curve. To test this hypothesis, we measured the currents when [cyclic AMP]_i was elevated by either extracellular application of 100 μmol l⁻¹ IBMX (N=10) or pressure injection of 1mmol l⁻¹ cyclic AMP into the neurone (N=10). Fig. 5A shows the I–V curve of the net IBMX current ( I_IBMX) obtained by substracting the currents measured before from those measured 20min after IBMX exposure. The pattern of this I–V curve was comparable to that of I_5-HT (see filled circles in Fig. 3). A similar result was obtained in the neurones when cyclic AMP was applied through the intracellular microelectrode (Fig. 5B). However, when neurones (N=4) had been superfused with Na⁺-and Ca²⁺-free saline, 5-HT and cyclic AMP had no effect (data not shown).

We have previously found that the cyclic-AMP-dependent protein kinase inhibitors H-8 and PKI both depress cyclic-AMP-dependent protein phosphorylation in extracts from snail ganglia (Onozuka et al. 1988a; Watanabe and Funase, 1991). Accordingly, we examined the involvement of the phosphorylation in the effect of 5-HT on the I–V relationship of the neurone using these two inhibitors. As seen in Fig. 6A, I_5-HT was inhibited in a time-dependent manner by injecting PKI into the neurone (N=10). The I–V relationships were also obtained in another experiment (N=4), first in normal saline after injection with PKI, then during perfusion with saline containing 5-HT, and finally during recovery in normal saline. PKI did not cause any shift of the resting I–V relationship; however, I_5-HT was completely abolished by PKI (data not shown). Similarly, another inhibitor, H-8, also suppressed I_5-HT in a dose-dependent manner, when added to the perfusate (Fig. 6B).

In molluscan neurones, the NSR has been shown to be evoked after application of a convulsant drug, pentylenetetrazole (Onozuka et al. 1991a) and several peptides or neurotransmitters, such as oxytocin (Funase, 1990a), vasopressin (Onozuka et al. 1986), FMRFamide (Ichinose and McAdo, 1988, 1989), egg-laying hormone (Kirk and Scheller, 1986; Kirk et al. 1988; Levitan et al. 1987) and dopamine (Matsumoto et al. 1988). An essentially similar response is seen in the Aplysia californica R15 neurone, which has a spontaneous bursting pacemaker potential (Wilson and Wachtel, 1978). It is widely accepted that the NSR is caused by an increase in Na⁺ conductance (Barker and Smith, 1978; Carnivale and Wachtel, 1980; Smith et al. 1975; Wilson and Wachtel, 1978), Ca²⁺ conductance (Eckert and Lux, 1975; Gorman et al. 1982) or both (Johnston, 1976).

In the present experiments, the PON neurone always exhibited spontaneous bursts of firing. Moreover, the steady-state I–V relationship of this neurone showed an NSR with a pattern similar to that of the Aplysia californica R15 neurone (Levitan et al. 1987; Levitan and Levitan, 1988). It was also similar to the NSR present during bursting activity induced by pentylenetetrazole (Onozuka et al. 1991a) and by vasopressin (Onozuka et al. 1986) in Euhadra peliomphala neurones. The PON neurone also showed 5-HT augmentation of the amplitude of both the depolarizing and hyperpolarizing phases of the burst cycle. Furthermore, the first NSR disappeared on removal of Na⁺ from the saline, while the second NSR disappeared on removal of Ca²⁺, suggesting that 5-HT augmentation of the depolarizing phase may be due to its stimulatory effect on the opening of both Na⁺ and Ca²⁺ channels associated with the NSR. In contrast, an augmentation of the post-burst hyperpolarization during exposure to 5-HT is attributed to an increase in the Ca²⁺-dependent K⁺ conductance coupled to the elevation of Ca²⁺ levels during the membrane depolarization.

Our data also show that both extracellular application of IBMX and intracellular injection of cyclic AMP simulate the 5-HT-induced effect on the I–V relationship in normal saline but that the effects are not detectable in Na⁺-and Ca ²⁺-free saline. These results imply that 5-HT may augment the bursting activity by elevating [cyclic AMP]_i, which, in turn, affects Na⁺-and Ca²⁺-dependent NSRs. In Euhadra peliomphala neurones, Onozuka et al. (1986, 1988a) found that the NSR at membrane potentials between -40 and -20mV, which is completely dependent on extracellular Na⁺, could be elicited by cyclic AMP. In addition, in the present experiments, the effect of 5-HT on the I–V relationship was abolished after intracellular injection of PKI or extracellular application of H-8, suggesting that both 5-HT augmentation of the Na⁺-activated NSR and 5-HT development of the Ca²⁺-activated NSR may be mediated by cyclic-AMP-dependent protein phosphorylation.

The mechanisms of the 5-HT response have been extensively studied in different Aplysia californica neurones. In the Aplysia californica sensory neurones, 5-HT produces a slow excitatory postsynaptic potential mediated by cyclic AMP, which contributes to presynaptic facilitation and behavioural sensitization of the animal’s gill withdrawal reflex (Kandel and Schwartz, 1982). Work by Kandel and his co-workers (Klein et al. 1982) led to the identification of a K⁺ current whose magnitude is reduced by 5-HT. This was named the S current. They also found that, in the cell-free membrane patch, this channel is modulated by cyclic-AMP-dependent protein phosphorylation (Shuster et al. 1986). In contrast, recent experiments carried out using the Aplysia californica R15 neurone, which is an endogenous burster, have shown that 5-HT augments the depolarizing and post-hyperpolarizing phases of the bursting cycle in this neurone and they suggest that this augmentation is due to an increase in the Ca²⁺ and K⁺ currents via cyclic-AMP-dependent protein phosphorylation (Lemons et al. 1986; Levitan and Levitan, 1988). The results from the present study, therefore, may demonstrate a mechanism of 5-HT action that has not yet been reported.

One question arises: how does 5-HT develop the Ca²⁺-dependent second NSR? Based both on the results obtained by application of IBMX or intracellular injection of cyclic AMP and on the observation that 5-HT increases [cyclic AMP] in Aplysia californica neurones (Greenberg et al. 1987; Sweatt et al. 1989), two explanations are possible. One is that 5-HT may activate a cyclic-AMP-dependent protein kinase that, in turn, is linked to the opening of Ca²⁺ channels. The other explanation is that 5-HT-activated protein kinase may be involved in the development of a Na⁺-dependent NSR, followed by voltage-dependent Ca²⁺ channel activation. The experiment carried out in Ca²⁺-free saline resulted in the disappearance of the second NSR, which was reduced by removal of Na⁺. Taken together, the results suggest that 5-HT causes the second NSR by phosphorylating either a Ca²⁺ channel directly or a channel-associated protein. This assumption may be indirectly supported by the observation that, when the outward K⁺ current was suppressed by TEA⁺, the voltage at the peak of the second NSR was shifted closer to the membrane potential at which the largest Ca²⁺ current is recorded in Aplysia californica (Adams et al. 1980a,b) and Euhadra peliomphala neurones (Onozuka et al. 1990), and that this phenomenon did not occur in the presence of a general protein kinase inhibitor 1-(5-isoquinolinylsulphonyl)-2-methylpiperazine (H-7) in the bathing medium (K. Funase, K. Watanabe and M. Onozuka, unpublished observation).

From the present results, we conclude that 5-HT augments the depolarizing phase of the burst cycle by its stimulatory effect on both the slow Na⁺ channels and the Ca²⁺ channels through cyclic-AMP-dependent protein phosphorylation. In contrast, the post-burst hyperpolarization may be augmented by an increase in the Ca²⁺-activated K⁺ conductance. Our recent evidence for the activation of K⁺ channels during bursting activity in snail neurones shows that intracellular injection of a Ca²⁺/calmodulin-dependent protein kinase II causes both a prolonged hyperpolarization of the membrane and an enhancement of the outward K⁺ current during pentylenetetrazole-induced bursting activity (Onozuka and Tsujitani, 1991).

This work was supported by grants from the Ichiro Kanehara Foundation, the Japan Brain Foundation, the Nukada Foundation, the Tokai Foundation of Encouragement of Academy, the Naito Foundation, the Narishige Neuroscience Trust and the Mochida Memorial Foundation for Medical and Pharmacological Research.