Possible Mediation of G-Proteins in Cold-Sensory Transduction in Paramecium Multimicronucleatum

The reception of thermosensory information is found in many living organisms (Hensel, 1974; Eckert et al. 1988), but the way in which intracellular transduction of such information occurs remains unknown. In the ciliated protozoan Paramecium, it has been shown that both warm and cold stimuli induce transient depolarization of the surface membrane, which triggers spike-like depolarizations and leads to directional changes in swimming (Hennessey et al. 1983; Nakaoka et al. 1987; Tominaga and Naitoh, 1992). In particular, cold-sensitive receptors are distributed on the anterior portion of the cell (Nakaoka et al. 1987), and their sensitivity is sufficiently high that they respond to an extremely slow rate of temperature decrease from the temperature at which the cell has been cultured (Nakaoka and Oosawa, 1977; Inoue and Nakaoka, 1990).

Heterotrimeric guanine-nucleotide-binding proteins (G-proteins) are involved in the pathways of a variety of sensory transduction processes in which the sensory response is amplified (Shepherd, 1991). In this study, we have examined whether G-proteins are involved in the cold-sensory response. Paramecium multimicronucleatum cells were voltage-clamped, and guanine nucleotide analogues were introduced into them. The results suggest the possibility that the cold-sensory transduction process in this organism involves some changes in G-protein activity.

Paramecium multimicronucleatum were cultured in a hay infusion inoculated with Klebsiella pneumoniae. The culture temperature was kept constant at 25 °C by incubation in a water bath. Cells at the stationary phase were collected by low-speed centrifugation and suspended in a solution containing 0.5 mmol l⁻¹ MgCl₂, 1 mmol l⁻¹ CaCl₂, 4 mmol l⁻¹ KCl and 2 mmol l⁻¹ Tris–HCl (pH 7.2). This solution was used as the experimental medium and, after suspension, cells were left for at least 1 h at 25 °C to enable them to adapt to these conditions. To carry out the experiments in the absence of action potentials, the cells were deciliated by incubation in experimental medium containing 5 % ethanol for 2–3 min before returning them to the experimental medium (Inoue and Nakaoka, 1990).

Methods of intracellular recording were similar to those described previously (Nakaoka and Iwatsuki, 1992). The cells were placed in a glass vessel mounted on an inverted microscope. Two electrodes were inserted from the upper side of the cell, one for injecting current and the second for recording the intracellular voltage. The electrodes usually contained 1 mol l⁻¹KCl and had a resistance of 20–50 MΩ. The temperature around the cell was controlled by changing the temperature of the water flowing beneath the vessel and was monitored with a thermistor probe placed close to the cell surface. The time constant of the thermistor was 0.4 s. The cells were initially voltage-clamped at 25 °C, then the temperature was lowered to 21 °C at an initial rate of 0.18 °C s⁻¹for 1 min, before returning the temperature to 25 °C at the same rate. The amplitudes of current responses were obtained from the maximal current changes that occurred during a 20 s period after the temperature drop. Membrane conductance was measured from the current evoked by application of a voltage step (−5 mV for 0.4 s) to the resting membrane. The resting potential was measured as the voltage difference between the internal and external electrodes after nulling the tip potentials in the bath.

Two methods were used for the intracellular application of reagents. In the first of these, chemicals were diffused through two microelectrodes for voltage-clamping. 10 mmol l⁻¹ guanosine-5′-O-3-thiotriphosphate (GTP-γS), 10 mmol l⁻¹ guanosine-5′-O-2-thiodiphosphate (GDP-βS), 0.1 mol l⁻¹ EGTA, 10 μmol l⁻¹ aluminium fluoride (AlF₄⁻; consisting of 10 μmol l⁻¹ AlCl₃ and 20 mmol l⁻¹ NaF), 10 mmol l⁻¹ adenosine-5′-triphosphate (ATP) or 10 mmol l⁻¹ adenosine-3′,5′-cyclic monophosphate (cyclic AMP) was added to the KCl electrodes and diffused into the cell. The second technique involved injection of chemicals through an additional glass microcapillary pipette using a pressure pulse of 30 Pa for 0.3 s (Picospritzer II, General Valve Co., Fairfield, NJ, USA). The capillary was filled with 5 mmol l⁻¹ Hepes (pH 7.0 adjusted with KOH) containing 5 mmol l⁻¹ GTP-γS, 5 mmol l⁻¹ GDP-βS, 25 mmol l⁻¹ EGTA or 30 mmol l⁻¹ LiCl. The volume injected was between 5 and 10 % of the cell volume, which was assumed to be 5×10⁻¹⁰ l. Both GTP-γS and GDP-βS (Boehringer-Mannheim) were in the form of lithium salts; ATP and cyclic AMP (Sigma) were sodium salts. EGTA (Dojindo, Japan) was neutralized with KOH.

When the membrane potential of cells at 25 °C was clamped at the resting level (approximately −21 mV) and the cells were subjected to a temperature drop for 1 min, an inward current was transiently induced at the time of maximum cooling (Fig. 1A). This observation was in agreement with previous results showing that the membrane transiently depolarized upon cooling (Nakaoka et al. 1987). Returning the temperature to 25 °C caused the inward current to disappear rapidly and sometimes induced a transient outward current (not illustrated).

When the holding potential was shifted to a more depolarized level than the resting potential, the amplitude of the inward current in response to cooling was increased. When the holding potential was shifted to a hyperpolarized level, the inward current decreased and disappeared at approximately −40 mV. Further hyperpolarization elicited only a slow outward current. The amplitude of the inward current was almost linearly proportional to the holding potential (Fig. 1B). The linear slope of the current–voltage plot was −9.41 pA mV⁻¹.

We examined the effects of chemicals diffusing into the cell from the electrodes on the cold-sensory response, soon after the voltage-clamp had been established for the cell.

GTP-γS (10 mmol l⁻¹), which may activate G-proteins, caused a large reduction in the inward current elicited in response to cooling (Fig. 2). This effect was particularly evident at the hyperpolarized holding potential below the resting potential (−24 mV). At depolarized potentials above the resting one, an inward current (whose amplitude was much smaller than that before the injection) was elicited. However, when the concentration of GTP-γS in the electrodes was decreased from 10 mmol l⁻¹ to 1 mmol l⁻¹, a reduction in inward current was not observed.

Application of AlF₄⁻, which may activate G-proteins, also reduced the current response (Fig. 3). This reduction was very similar to that caused by GTP-γS.

After the cell had been perfused with 10 mmol l⁻¹ GDP-βS, which is thought to inhibit G-protein activity, the inward current in response to cooling was reduced (Fig. 4). At a depolarized potential above the resting potential (−18 mV), this reduction was more pronounced than that resulting from application of either GTP-γS or AlF₄⁻. When the GDP-βS concentration was reduced from 10 mmol l⁻¹ to 1 mmol l⁻¹, the effects on the inward current virtually disappeared.

Intracellular application of EGTA, which should decrease the Ca²⁺ concentration within the cell, also caused a reduction of the inward current produced in response to cooling (Fig. 5). The results of this experiment also showed that the application of EGTA reduced the spontaneous fluctuations in current observed before the application of a cooling stimulus at various holding potentials (Fig. 5A).

Since the GTP-γS and GDP-βS used in the above experiments were lithium salts, the effect of intracellular application of LiCl was tested. In the presence of LiCl, no changes were detected in the inward current produced in response to cooling. Furthermore, neither ATP nor cyclic AMP had any effect on the current response (data not shown).

The chemicals that reduced the current response modulated both the resting potential and the membrane conductance (Table 1). Compared with the control condition, application of both GDP-βS and EGTA changed the resting potential to depolarized levels and decreased membrane conductance. In contrast, application of GTP-γS caused almost no changes in the resting potential and an increase in membrane conductance.

When the surface membrane of Paramecium multimicronucleatum was held under voltage-clamp conditions, the application of a cooling stimulus elicited an inward current. The amplitude of the inward current was linearly decreased by negative shifts of the holding potential and disappeared at approximately −40 mV, which is almost the equilibrium potential for K⁺ (Oka et al. 1986; Machemer, 1988). From the linear slope of the change in current amplitude with potential, the change in membrane conductance in response to a cooling rate of 0.18 °C s⁻¹ was estimated to be 9.4 nS, and this value was approximately half of the resting conductance at 25 °C (Table 1, control). These results are in agreement with the conclusion drawn from a previous current-clamp study that it is mainly the membrane conductance for K⁺ that changes in response to cooling (Nakaoka et al. 1987) and, since Paramecium multimicronucleatum membranes depolarize in response to cooling, K⁺ conductance must decrease under these conditions. This, in turn, suggests that K⁺ channels which respond to a cooling stimulus are open before the application of the stimulus.

Although the non-hydrolyzable guanosine analogues GTP-γS and GDP-βS affect G-protein activity in opposite directions (the former increases and the latter decreases G-protein activity), both reduced the current response upon cooling. Such effects indicate that cold-sensory transduction may include some transient changes in G-protein activity. These changes in G-protein activity might modulate effector systems, presumably resulting in the transient closing of K⁺ channels in response to cooling.

Two possible mechanisms can be suggested for the way in which G-protein regulation of the K⁺ channel activity occurs: first, a direct interaction with a G-protein subunit, which changes the K⁺ channel activity (Brown and Birnbaumer, 1990; Hille, 1992); second, changes in G-protein activity modulate the effector system that produces cyclic nucleotides or raises the intracellular Ca²⁺ concentration, and this in turn affects the channel activity (Simon et al. 1991; Kaupp and Koch, 1992; Berridge, 1993). Although the present study does not indicate which of these two mechanisms is involved in regulating K⁺ channels in Paramecium multimicronucleatum, the inhibitory effects of EGTA on the cooling-sensitive response suggest that Ca²⁺ levels contribute to the regulating effect. Intracellular application of EGTA, which reduces the intracellular Ca²⁺ concentration, decreased the membrane conductance and depolarized the membrane potential from the control value. Such changes in the membrane seem to be caused by a [Ca²⁺]-dependent decrease in K⁺ channel activity. Similar changes in the membrane conductance and the resting potential were induced by the application of GDP-βS. It follows therefore, that GDP-βS might decrease the activity of K⁺ channels through a decrease in the intracellular [Ca²⁺].

In support to this hypothesis, we have recently found that a mutant of Paramecium tetraurelia (cam¹²) with a substitution of an amino acid residue in the calmodulin molecule (Kink et al. 1990) shows a reduced sensitivity to the cooling stimulus and that this mutant has a decreased resting membrane conductance for K⁺(T. Kuriu, Y. Oosawa, Y. Watanabe and Y. Nakaoka, in preparation). These results support the view that K⁺channels whose activities are regulated by Ca²⁺ are involved in the response to the cooling stimulus.

Although in several Paramecium species, including P. multimicronucleatum, a β-subunit-like gene of the heterotrimeric G-protein has been found (Forney and Rodkey, 1992), a G-protein α-subunit has not yet been identified. Nevertheless, the cortical alveoli of Paramecium tetraurelia have been shown to act as a submembranous Ca²⁺ storage compartment and the alveolar membranes have Ca²⁺-pumping activity (Stelly et al. 1991, 1995; Lange et al. 1995). However, the regulatory mechanism controlling intracellular Ca²⁺ homeostasis has not been established.

The results of the present study indicate that G-proteins are involved in the transduction of cold-sensory information; however, it remains to be determined how they transmit the cooling stimulus and regulate the activity of ion channels.