A Ca-Induced Na-Current In Paramecium

The locomotion of a paramecium is regulated by its membrane potential. When the membrane is at rest, the cilia beat in their normal direction and propel the paramecium forward. Upon proper stimulation, a Ca-action potential can be triggered which increases the internal Ca²⁺ concentration which, in turn, causes the cilia to beat in the reverse direction and the cell to swim backward for a short distance (the ‘avoiding reaction’). Like the ‘tumbling’ in motile bacteria (Springer, Goy & Adler, 1979), the avoiding reaction is necessary for certain chemotaxes (Van Houten, 1978), thermotaxes (Hennessey & Nelson, 1979) and the response to mechanical stimulation at the anterior end (Naitoh & Eckert, 1969). Mutational blocks on the action potential also block the avoiding reaction and cause a failure in these taxes but not in normal ciliary motion (Nelson & Kung, 1978; Kung, 1979).

The electrophysiology of Paramecium has advanced greatly since the first electrode letration by Kamada in 1934 (see Naitoh & Eckert, 1974; Eckert & Brehm, 1979; Kung, 1979 for reviews). Much previous work concentrates on the Ca-action potential which involves a voltage-sensitive Ca-channel and a delayed rectifying K-channel However, several electrophysiological problems related to the behaviour of this organism have not been resolved. For example, the basis for the natural membrane discharges, which correspond to the ‘spontaneous’ avoiding reactions in the culture medium, is not fully understood. A paramecium can also spin backward rapidly and continuously for minutes upon certain stimuli. The electrophysiological basis of this continuous backing is also poorly understood.

Na⁺ is a natural cation in fresh water and in various culture media for Paramecium. Spontaneous avoiding reactions are easily observed in solutions containing Na⁺, such as the commonly used Cerophyl culture medium buffered with sodium phosphates (Sonneborn, 1970). The membrane discharges corresponding to such avoiding reactions have been recorded from paramecia bathed in Na⁺-containing solutions (Satow & Kung, 1974). Several mutants have been isolated which are best distinguished by their peculiar behaviour in these solutions (Kung, 1971a; Satow & Kung, 1974, 1976b; Satow, Hansma & Kung, 1976). ²²Na influx has been measured in Paramecium, particularly in the ‘paranoiac’ mutants which exhibit prolonged backward swimming and plateau depolarizations in solutions containing Na⁺ (Hansma & Kung, 1976; Satow et al. 1976; Hansma, 1979). Nevertheless, the questions of whether there is a Na-conductance in this membrane and what the functional role of Na⁺ is have not been settled electrophysiologically. We show in this paper that the Paramecium membrane has indeed a Na-conductance. The magnitude, kinetics and triggering mechanism of the Na-current under a voltage clamp are described and the functional significance and behavioural implications of this current are discussed.

P. tetraurelia, formerly P. aurelia species 4, was used. The stocks used included 51s (wild type), d4–9O (genotype PaA/PaA, a ‘paranoiac mutant’), d4–94 and d4–500 (pwA₉₄/pwA⁹⁴ and pwA⁵⁰⁰/pwA⁵⁰⁰, respectively, ‘pawn A mutants’), d4–95 and d4–501 (pwB⁹⁵/pwB⁹⁵and pwB⁵⁰¹/pwB⁵⁰¹, respectively, ‘pawn B mutants’) (see for the genetical descriptions: Kung, 1971b; Van Houten, Chang & Kung, 1977; Kung, 1979), and three constructed paranoiac-pawn double mutants (PaA/PaA pwA⁹⁴/ pwA⁹⁴, PaA/PaA pwB⁹⁵/pwB⁹⁵, PaA/PaA ptvB⁵⁰¹/pwB⁵⁰¹ (Van Houten et al. 1977; Chang, Thiede & Kung, unpublished)). d4–94 is a phenotypically slightly leaky, d4–500 a nearly non-leaky mutant of the PaA complementation group (Chang et ‘al. 1974; Satow & Kung, 1980a) and d4–95 and d4–501 are mutants of the pwB group with little leakiness (Chang et al. 1974; Schein, Bennett & Katz, 1976).

Paramecia were ordinarily cultured in Cerophyl medium buffered with Na-phosphates and bacterized with Enterobacter aerogenes (Sonneborn, 1970). To delete the cellular Na⁺, the enterobacters were first grown overnight in a defined glucóse-Tris medium without Na⁺ (Linn & Forte, personal communication). Paramecia were then transferred into a ten-fold dilution of this medium to a final concentration of 8–9 mMM-K⁺, 12–13 mm-Tris, 0·09 mm-Ca²⁺, 0·02–0·03 mm-Mg²⁺ and ca. 5 μg/ml stigmasterol. Only well-fed cells that had been grown in some medium for more than 20 hours were used for experimentation.

The methods of capturing, rinsing, immobilizing, penetrating and recording from paramecia were basically those described by Naitoh & Eckert (1972). Cells were withdrawn from their culture media, washed in the ‘K-solution’ (see below) and put into a drop hanging under a glass slip. The slip was then put on the experimental chamber on the stage of a compound microscope with electrodes poised for penetration. A syringe-controlled suction device was used to regulate the size of the hanging drop to immobilize the cell without undue desiccation. One electrode was then inserted by micromanipulation into the paramecium before the chamber was flooded with the K-solution to immerse the cell. A second electrode was then inserted. These electrodes were filled with 3 M-KCl with resistances around 50 MΩ.

The bath solution could be changed by a gravity-fed perfusion system. The first solution in our experiments was always the K-solution in which a fixed reference level was established for the entire experiment. After the penetration of the electrodes in the K-solution, the specimens gradually attained membrane potentials of about – 35 mV in 1–2 min. Experimentation began after this period of adjustment in this solution. The bath was changed several times as necessary.

The voltage clamp was that of Oertel, Schein & Kung (1977), with slight modifications. The holding potential was – 30 mV unless otherwise stated. The voltage stabilized within 1 msec after a command step change. However, a notch of less than 4 mV was usually observed when the transient Ca inward current was triggered by the depolatization. The open-loop gain was most often 100 × (80 × to 200 × ). The series resistance of the clamp circuit was less than 200 KΩ.

Unless noted otherwise, each set of experiments was repeated on at least three cells.

Ions were injected into the paramecium through a third electrode filled with the desired ion by a current of appropriate polarity. The iontophoresis was performed when the membrane was under the voltage clamp. Thus, current flowed largely between this third electrode and the current electrode of the clamp, both inside the cell. Since relatively little current actually passed through the membrane, we were able to inject large iontophoretic currents without damaging the specimens. The deflection of the membrane potential was as small as 6 mV (up to 20 mV in some cases) with a 10 nA injection by this method, even though the total membrane resistance of a paramecium was about 50 MΩ. A 10 nA, 1 min injection changes the internal concentration of the injected ion by 20 mm as caluclated by the method of previous workers (Satow & Kung, 1976 a).

Standard salt solutions used in this study were the ‘K-solution’ with 4 mm-K⁺, I mm-Ca²⁺, 1 mm-HEPES, 5–6 mm-Cl⁻, the ‘Na-solution’ with 8 mm-Na⁺ 1 mm-Ca²⁺, 1 mm-HEPES, 9–10 mm-Cl⁻ and the ‘Ca-solution’ with 1 mm-Ca²⁺, 1 mmHEPES, 2–4 mm-Cl^−. Solutions with different ionic compositions were also used as specified in the text. All solutions included 10μM-EDTA to chelate any possible toxic heavy-metal ions and were adjusted to pH 7·1–7·3.

Tris [Tris(hydroxymethyl)aminomethane], HEPES (N-2-hydroxyethylpiperazine-N’-2-ethanesulfonic acid), EDTA (ethylenediaminetetraacetic acid) were from Sigma. All chemicals were of reagent grade.

Experiments were performed at room temperatures (20–23 °C).

When paramecia are transferred to a solution of different ionic strength and species, they respond behaviourally. The nature, intensity and duration of the response depend on the ionic compositions of both of the solutions (Jennings, 1906; Naitoh, 1968). The paramecia that have been washed into and incubated in the ‘K-solution ‘(4 mm- K⁺, 1 mm-Ca²⁺; see Materials and Methods) swim forward in this solution, rarely showing an avoiding reaction. The mechanical disturbance of transferring them from one pool of K-solution to another also induces little change in this behaviour. Transferring these ‘adapted’ cells into solutions containing Na⁺, however, induces reversal of the ciliary beat and therefore backward swimming. When the wild-type paramecia in the K-solution are transferred into a solution of high Na⁺ or low Ca²⁺ concentrations, e.g. the ‘Na-solution’ (8 mm-Na⁺, 1 mm-Ca²⁺), the first few responses are variable but often consist of longer backward swimming for several seconds (up to 30 s in some cases). These prolonged backings are followed by jerks and do not recur later.

Sustained backward swimming is easily observed in the ‘paranoiac’ mutants of P. tetraurelia. Without any apparent trigger, these ‘paranoiac’ paramecia sporadically swim backward for up to 60 s or more in the culture medium (bacterized Cerophyl infusion buffered with Na phosphates). However, they show no such backward swimming in the K-solution. When transferred from the K-solution into various solutions containing Na⁺, they again swim backward up to 2 min or more. Such bouts of backward swimming are repeated as long as the cells are in the Na-containing solution. The following is a study of the electrophysiological correlates of the prolonged backward swimming in the Na-solution.

Corresponding to the simple forward swimming in the K-solution the membrane of the wild-type paramecium is found to be quiescent, resting at – 36·1±5·5 mV (mean ±S.D., n = 16) with no spontaneous discharge (Fig. 1A). Injection of an outward square pulse triggers an action potential followed by a steady-state depolarization (Fig. 1B), as has been previously reported (Naitoh & Eckert, 1968; Satow & Kung, 1976 a). There is an after-potential of about +5 mV (up to + 10 mV from the resting level) that follows the response to the large square pulses. The ionic nature of this small after-potential is not known.

In paramecia bathed in the K-solution, under voltage clamp, a step depolarization from the holding potential of – 30 mV to – 26 mV or higher, induces a transient inward current lasting about 5 ms (Oertel et al. 1977; Brehm & Eckert, 1978; Saton & Kung, 1979), and followed by an outward current sustained throughout the depolarization step (Fig. 2 A). The plot of current measured at 1 sec vs. the voltage step is nearly linear from –45 to around – 17 mV. The current is clearly rectified beyond this range (Fig. 2E). A step depolarization to above – 20 mV induces a slowly growing outward current after the inward transient (Fig. 2B). Correspondingly, a tail current appears after the step depolarization. This tail current is outward after a depolarization above – 20 mV and it decays nearly completely within 200 ms. This slowly growing outward current and the corresponding outward tail current are similar to those identified as the Ca-induced K-current and its tail by Satow & Kung (1980b). It has also been reported that there is a depolarization sensitive K-current with a fast activation kinetics (Oertel et al. 197; Satow & Kung, 1980a), though its voltage sensitivity is not studied here.

To investigate the effect of Na⁺ on the wild-type paramecium membrane, the bath of K-solution is replaced by solutions containing Na⁺. The unclamped membrane responds to this change of solution with a series of rapid discharges, i.e. it repeatedly depolarizes and repolarizes (Fig. 1C). These discharges correspond to the jerks be-haviourally observed (Satow & Kung, 1974). When 8 mm-Na⁺ (the Na-solution) is used to replace the K-solution, the membrane depolarizes to about – 10 mV (– 3 to – 23 mV) for tens of seconds. This first depolarization is often followed by a few isodes of shorter and smaller depolarizations from the new resting level ‘(– 26·0–3·6 mV, n = 14). These long depolarizations are highly variable in magnitude and iluration from cell to cell and are often separated by clusters of rapid discharges, each lasting less than 1 s (Satow & Kung, 1974). The long depolarization and the rapid discharges most likely correspond respectively to the backward swimming and the jerks when adapted paramecia are transferred to the Na-solution. Both the spontanous membrane activities and the behaviour upon the introduction of Na⁺ vary from specimen to specimen.

In the Na-solution, an outward current injected across the membrane induces an action potential and the steady-state depolarization (Fig. 1 C, D) as in the K-solution (Fig. 1 A, B). However, unlike those in the K-solution, the after-potentials in the Na-solution are always depolarizations that can be very long (up to 2 min) and very large (up to – 5 mV) (arrows, Fig. 1 C, D). The level and duration of such depolarizations vary greatly from cell to cell. The trigger needed to generate such depolarizations also varies. In some cases, plateau depolarizations spontaneously occur; in others, they cannot be triggered even with outward currents up to 1 nA for 1 s. It is, therefore, difficult to study the properties of the membrane during the prolonged depolarization induced by Na⁺, using the wild-type membrane.

The membrane currents under voltage clamp in the Na-solution are similar to those in the K-solution. Upon step depolarizations, the inward transient, the steady-state outward current, and the slowly developing outward currents are all observed in cells bathed in the Na-solution (Fig. 2C, D). The anomalous rectification is much stronger in the Na-solution, a phenomenon to be dealt with elsewhere.

Although there is no large difference in the depolarization half of the I-V curve (Fig. 2E) in the two solutions, the tail current after the step depolarization is more complex in the Na-solution. When the step is above – 17 mV, this tail has a prominent inward component (arrow, Fig. 2D). The inward current decays very slowly over hundreds of milliseconds. Before this inward tail current, an outward tail current which decays faster can be observed. The kinetics of this outward tail current are similar to those found in cells bathed in the K-solution. These complex tail currents are found in all wild-type paramecia bathed in the Na-solution and most likely represent the conductances responsible for the positive after-potentials which sometimes appear as plateau depolarizations of the unclamped wild-type membranes.

Although the membrane behaviour is better controlled under the voltage clamp, the variability and the small size of the inward tail current from wild-type specimens make the analyses more difficult. We therefore chose to study the ‘paranoiac’ mutant where prominent plateau depolarizations can always be triggered in the Na-solution and the inward tail currents are larger.

The electrical properties of the membrane of the paranoiac mutant bathed in the K-solution are similar to those of the wild type in the same solution. There is no spontaneous membrane activity and the membrane remains at – 39·3 ± 4·4 mV (n = 24). Proper current injections elicit action potentials, steady-state depolarizations and the small positive after-potentials in the mutant (Fig. 3, left) as in the wild type (Fig. 1, left). In the K-solution and under voltage clamp, the mutant membrane also behaves similar to that of the wild type. The inward transient, the steady-state outward current, the delayed rectification current and the tail current are seen in the mutant (Fig. 4A, B) as in the wild type (Fig. 2A, B). The I-V plot of the mutant at i s of the voltage step shows a rectification above – 17 mV as in the wild type. Correspondingly, slow tail outward currents are seen after such steps (Fig. 4B).

When the bath is changed from the K-solution to the Na-solution, the membrane of the paranoiac mutant first depolarizes to – 2·6 ± 4·9 mV (n = 24). Although in some cases, this first plateau depolarization can last for more than 5 min, typically the membrane repolarizes within 3 min to a lower ‘resting’ level (– 27·1 ±6·0 mV, n = 10). After the first episode, plateau depolarizations of up to several minutes take place spontaneously and can be triggered by injected current (Fig. 3, right). These repeated prolonged plateau depolarizations correspond to the bouts of backward swimming in the Na-solution, which constitute the diagnostic phenotype of paranoiac mutants (Kung, 1971 a, b).

Under voltage clamp, a step depolarization from the – 30 mV holding level induces the inward transient in the Na-solution. The peak inward transient and the voltage at which this peak occurs are similar to those in the K-solution. However, when the voltage step is less than – 14 mV, the late current in inward (Fig. 4C, arrow) instead of outward. This inward current develops very slowly during the depolarization (see below on kinetics). At the end of the voltage step, the sustained inward current is followed by a prominent inward tail current (Fig. 4C, double arrow), which decays slowly (see below and Fig. 7 for kinetics). A voltage step above –14 mV induces a large net outward current followed by a complex tail current. Nevertheless, a very slow inward component is clearly seen in the tail current after the step (arrow. Fig. 4D). This inward component is preceded by an outward component that decay faster (as estimated from the end of the voltage pulse to the peak of the tail current complex). The outward tail current seen in the Na-solution is similar in kinetics ta that in the K-solution.

A plot of the late current at 1 s vs. step voltage level shows a prominent region of negative resistance around –20 mV (Fig. 4E, arrow). This region of negative resistance is not seen in the K-solution or in the ‘Ca-solution’ devoid of both K⁺ and Na⁺. The N-shaped I-V plot suggests that depolarization of the unclamped membrane can become regenerative because of this slow inward current.

To study the kinetics of the slow inward current, we compare the membrane currents in solutions with and without Na⁺. In the K-solution and in the Na-free Ca-solution the I-V curve is nearly linear from –45 to about –17 mV (Fig. 4E), showing that the late outward current develops very little, while the slow inward current is induced (Fig. 4). Fig. 5 B is a typical result showing the magnitude of the inward current at different times during a 1 min voltage step from the holding level of – 30 mV to – 21 or – 17 mV. The inward current increases in the first 9 s (4–9 s in three cases) to about 1 nA (0·9–1·2 nA) and then slowly declines to less than 0·7 nA (0·3–0·7 nA) in 60 s during the clamp depolarization.

The development of the slow current was also studied by measuring the size of the tail inward current upon repolarization from the voltage step of various durations. Since the tail inward current decays exponentially after the return from a low voltage step (see below), we are presumably dealing with a single conductance in isolation (unlike the situation during the measurement of total current (Fig. 5 B) which may contain more than one component). The size of the tail current shows a rise and a fall (Fig. 5 A) when plotted against the duration of the voltage pulses which parallel those of the total slow inward current itself ; the maximal tail current is obtained after a voltage pulse of 4–9 s in duration (three cells). Therefore we conclude that the development of the slow inward current in the Na-solution is due to a change of one conductance which activates and inactivates very slowly.

The slow inward current begins its development within 10 ms of the onset of depolarization to a low (⩽– 17 mV) or a high (⩾ – 14 mV) potentials. On the other hand, the slow outward current, most probably the Ca-induced K-current (Satow & Kung, 19800), which is induced only at the high potential, appears > 200 ms after the onset of the voltage step (Fig. 6A-D). The outward component of the tail current following the pulse also occurs only when the step is long. When the step is < 200 ms (Fig. 6C), neither the outward current or its tail is observed. Thus, the slow outward current has a different threshold and different kinetics from those of the slow inward current.

The time course of the tail of the inward current during the step is also slow. The inward tail current (Fig. 4C) after a small depolarization (– 21 to–17 mV in different cases) decays almost exponentially, with a time constant of approximately 0·4 s (0·38 ± 0·15 s, n = 6) after the intial 20 ms during which an additional small inward component is occasionally seen. The slower inward component of the compound tail (Fig. 4D) seen after a larger depolarization (– 10 to –9 mV) also decays with similar kinetics (time constant 0·43 ±0·10 s, n = 6). An example of a semi-logarithmic plot of the tail currents is shown in Fig. 7. At higher voltage steps, the slow inward component of the tail is preceded by an outward component (Figs. 4D, 6 A and the upper curve of Fig. 7) which decays faster, with a time constant of less than 0·1 s. In contrast to these slow currents, the tail currents associated with the Ca-transient and the voltage-sensitive K-current decay so fast that they cannot be clearly discriminated from the capacitative surge.

It is clear, therefore, that during long depolarization a conductance is slowly activated for several seconds to pass the inward current, but only when the paramecium is bathed in a solution containing Na⁺. The slow current and its associated tail current, found in the K-solution and in the Ca-solution, are both outward and clearly differ from the inward currents in their kinetics and apparent voltage dependency. The results in this section further strengthen the notion that the slow inward current with a voltage range of negative resistance is due to the presence of external Na⁺.

Since the long plateau depolarization in the unclamped membrane and the slow inward current in depolarized membrane under the voltage clamp are all Na-dependent, we investigated the effects of external concentrations of Na⁺ on these parameters.

The level of the first plateau depolarization following Na⁺ perfusion, usually the longest and largest, is dependent on the external Na concentration. In 2, 4 and 8 mm - NaCl the depolarizations were – 9·1 ±6·1 (n = 18), –8·3 ±3·0 (n = 24) and –2·6 ± 4·9 mV (n = 24), respectively. (In 2 out of 20 cases, there was no typical plateau depolarization in the solution containing 2 mm-Na⁺. All cases gave plateau depolarizations with higher Na⁺ concentrations.) At 2 mm-Na⁺, there is little inward current and the region of negative resistance of this current is less prominent. This current and the region of negative resistance appear and become bigger in solutions of higher Na⁺ concentrations, although the variation is large. Changes in external Na⁺ concentration, however, have only minor effects on the fast inward transient (the Ca-current) and the region of negative resistance of this transient Ca-current (Table 1).

If the Na-dependent, slow, inward current is carried by Na⁺, one would expect internally added Na⁺ to inhibit this current. We, therefore, iontophoretically injected Na⁺ into paranoiac mutants bathed in Na-solution (8 mm-Na⁺) in which the slow inward current had been observed. A 10 nA injection of Na⁺ for 1 min suppresses the slow inward current during a small depolarization step, and the tail inward current after it. With a larger depolarization (⩾ –14 mV) Na⁺ injection increases the apparent outward current and decreases the tail inward current (Fig. 8 A, B), suggesting a suppression of the inward component in the complex. To isolate the suppressible inward component the membrane current after the Na-injection was subtracted from that before the injection. The isolated inward component, which corresponds to a lower depolarization step (–17 mV), is more than 0-2 nA and increases throughout the 1 s period. At a higher voltage (–14 mV) this current is larger ( > 0·6 nA) (Fig. 9 A). This isolated current is always associated with an inward tail current after the depolarization step. The magnitudes and the kinetics of the inward current and the tail current isolated by this method are similar to those induced by lower voltage steps (Fig. 4C, 5 B). In control experiments, iontophoresis of K⁺ Pith a current identical to the one used to inject Na⁺ (10 nA, 1 min) has little effect the slow membrane currents (Fig. 8C, D). Subtraction of the membrane current after the K⁺ injection from that before yields only a small change in currents (< 0·5 nA at – 14 mV) that changes little in time (Fig. 9B).

The effect of injected Na⁺ can also be demonstrated in cells bathed in the Na⁺-free ‘K-solution’. Iontophoresis of Na⁺for 1 min with 10 nA greatly increases the outward currents induced by voltage steps above –14 mV. Subtraction of currents reveals a slow outward current, which develops slowly during the 1 s step, and the corre-sponding slow outward tail current, which decays over several 100 ms (Fig. 9C). lontophoretic injection of K⁺ into cells bathed with the K-solution has virtually no effect (Fig. 9D).

Paramecia grown in the Na+-free medium can be expected to have very low internal Na⁺. The wild types grown in the Na⁺-free medium, show ‘paranoiac’ behavior (i.e. they swim backward repeatedly for tens to hundreds of seconds when exposed to external Na⁺). However, their normal behaviour is re-established within a few minutes, possibly due to the accumulation of internal Na⁺. This rapid readjustment and the complex nature of the tail current makes very difficult a quantitative analysis of the slow inward current and its tail current from ‘Na⁺-depleted cells’. Addition of 5 mm-Na⁺, in place of K⁺, in the ‘Na⁺-free medium’ (see Materials and Methods) suppressess part of the ‘paranoia’ induced by Na⁺ deficiency in the wild types. The size of the tail inward current of these cells is smaller than that of the Na⁺-depleted cells.

The results in this section are consistent with the view that prolonged depolarization induces a slowly developing current carried by Na⁺. This current is inward when the cell is bathed in the Na-solution but is outwardly directed in the Na-free K-solution. It inactivates slowly after the repolarization. Since depolarization first induces a fast Ca-response, which increases the internal concentration of Ca²⁺, it is possible that the Na-current is induced by internal Ca²⁺ and not by the depolarization directly. There-fore, we tested the effect of Ca²⁺ on this current.

To measure the external Ca²⁺ requirements for the inward Na⁺ current, we replaced most of the Ca²⁺ in the Na-solution with Mg²⁺. When this Mg-Na solution (8 mm-Na⁺, 0·99 mm-Mg²⁺, -0·01 mm-Ca²⁺) replaces the Na-solution (8 mm-Na⁺, 1 mm - Ca²⁺), the Ca-transient is reduced. However, the presence of an inward tail current after the step depolarizations indicates that a residual, though greatly reduced, slow inward current persists in cells bathed in the Mg-Na solution (data not shown). Correspondingly, most unclamped cells still show repeated spontaneous or triggered plateau depolarizations, albeit smaller and shorter.

For a more stringent test, paramceia were bathed in a 8 mm-Na⁺ solution with 01 mm-EGTA and no added Ca²⁺. The cells were thus in a critical condition. However, experiments could be performed, rapidly, before severe deterioration occurred. Step depolarizations with the voltage clamp, revealed little sign of the inward Catransient or of the slowly developing current, as judged by the lack of a tail current in this Na-EGTA solution (Fig. 10A). Upon return to the Ca-containing Na-solution, the Ca-transient and the slow currents are partially restored (Fig. 10 B). The re-appearance of the inward tail current indicates the slow inward current in the slow current complex.

The results above suggest that a small quantity of external Ca²⁺ is needed to initiate the Na-current and the plateau depolarizations. EGTA was, therefore, injected into the cells to discover whether internal Ca²⁺ induces the Na-current.

EGTA (injected with a current of 5 nA for 30 s) increases the inward Ca-transient slightly. In contrast, the slow Na inward current and its corresponding inward tail current are suppressed after EGTA injection (Fig. 11 A, B). The injected EGTA also suppresses the slow outward K-current, and its corresponding outward tail current, after the voltage step change (Fig. 11 C, D).

Larger EGTA injection (10 nA, 60 s) has an additional effect. It induces a sustaind inward current when the voltage step is at about –14 mV. Unlike the slow 1 inward current, that develops during seconds after the Ca inward transient, this sustained inward current appears immediately after the Ca-transient. This observation in P. tetraurelia is similar to that in P. caudatum by Eckert & Brehm (1979), who attribute this current to the residual Ca²⁺ flow through an incompletely inactivated Ca-conductance due to the low internal concentration of Ca²+. Unlike the Na-current this Ca-current is not followed by a slow tail after the step depolarization (Fig. 11 E, F).

The above experiments show that the slow Na-current requires a high internal Ca²⁺ concentration. It is, therefore, reasonable to suppose that this Na-current is not directly induced by depolarizations but by the high internal concentration of Ca²⁺ which may come from the Ca-transient triggered by the depolarization. To test the idea we examined mutants which do not have the Ca-transient.

‘Pawns’ are mutants that have little or no transient Ca²⁺ inward current (Kung & Eckert, 1972; Oertel et al., 1977 Satow & Kung, 1980a). Voltage-clamp experiments in four different pawn mutants show that, in all cases (⩾ 2 cells of each group), the mutations suppress not only the Ca inward transient, but also the slow Ca-induced Na inward current and its tail. The slow outward current (the Ca-induced K-current) is also suppressed.

As noted above, the inward Na-current is best observed in the paranoiac mutants. Therefore, we used pawn-paranoiac double mutants to test whether the Ca-transient is really a prerequisite for this Na-current in a genetic background more favourable for its appearance. As shown in Fig. 12, both the pawn A-paranoiac double mutant and the pawn B-paranoiac double mutants are also devoid of any late currents and their associated tail currents. In all cases of such double mutants, there is also no spontaneous or triggered prolonged depolarizations in the Na-solution characteristic of the paranoiac single mutant. These observations further support the view that the outward K-current and the inward Na-current that develop slowly upon depolarization in the wild type are induced by internal Ca²⁺ entered through the voltagesensitive Ca-channel.

We have discovered a slow inward current in Paramecium. It is most likely carried by Na⁺ because: (1) this current and the tail current associated with it are observed only when there is external Na⁺; (2) the current and its tail are larger when external Na⁺ is more concentrated; (3) they are smaller or absent when internal Na⁺ is increased by iontophoresis.

A large influx of Na⁺ associated with prolonged depolarizations has been demonstrated by ²²Na accumulation experiments in Paramecium (Hansma & Kung, 1976; Hansma, 1979). During step depolarizations of the paramecium membrane the Ca-current does not inactivate completely (Brehm & Eckert, 1978) and there is a Ca-induced K-current (Satow & Kung, 19806), although both currents should be small at voltage steps of less than –17 mV. This complexity, and the small size of the Na-current, make it very difficult to determine accurately the reversal potential of this Na-current. This reversal potential, though variable from cell to cell, is definitely above – 15 mV in the 8 mm-Na⁺-solution.,

The inward current has very slow kinetics. It develops, on depolarization for several seconds, to a peak and decays upon a repolarization with a time constant of hundreds of ms. The slow kinetics of this current exclude the possibility that the current flows through the voltage-sensitive Ca channel which activates and nearly completely inactivates in a few ms. Furthermore, EGTA-injection, which would be expected to enhance and prolong Ca-current or any current through the Ca-channel (Eckert & Brehm, 1979), in fact, reduces the slow inward current (Fig. 11). It is also unlikely that this inward current is a residual current after a hypothetical suppression of the 01 ward current by Na⁺. Such a hypothesis entails that Na⁺ suppresses a conductance in fact, during the Na-induced plateau depolarizations, the membrane is far more conductive (i.e. having very low resistance, Satow et al. 1976) and large fluxes of ions occur (Hansma & Kung, 1976; Hansma, 1979). Thus, the conductance for this slow inward current in Na-specific in physiological conditions. K⁺, Ca²⁺, Mg²⁺ and Ba²⁺ cannot substitute for Na⁺ to carry this current. Only Li⁺ can replace Na⁺ to trigger the characteristic behavioural responses and spontaneous or triggered membrane discharges of wild type and ‘paranoiac’ (unpublished data).

The Na-current in Paramecium is very different from the well-known Na-current in axons. Besides its slow kinetics, it is not affected by tetrodotoxin or saxitoxin (unpublished data). Furthermore, its triggering mechanism is very different from the conventional Na-current. It is not directly triggered by depolarization but by internal Ca²⁺, since depletion of external Ca²⁺ and injection of EGTA suppress or eliminate this Na-current even under depolarization. It appears to be indirectly depolarizationsensitive, because the increase in internal Ca²⁺ is normally induced by the activation of the voltage-sensitive Ca-conductance. This is demonstrated by the lack of this Na-current in the single or double mutants carrying a defect which makes the Ca-channel nonfunctional (Fig. 12).

The Ca-induced Na-current can have a region of negative resistance in the I-V plot (Fig. 4E). It is, therefore, potentially a mechanism for regenerative depolarization in the unclamped membrane. A form of slowly rising depolarization lasting tens to hundreds of seconds has been recorded in the paranoiac mutants (Fig. 3C; see also Satow et al. 1976). The slow rise and the long duration of such depolarizations of the unclamped mutant membrane parallel the slow activation and the persistence of the Na inward current under the voltage clamp (Fig. 5). Continuous ciliary reversal during these long depolarizations indicates a lasting increase in internal Ca²⁺ concentration which should be correlated with activations of the Ca-induced Na inward current as well as the Ca-induced K outward current. Such prolonged depolarizations have indeed been shown to be correlated with a large ²²Na influx and a large K⁺ loss (Hansma & Kung, 1976; Hansma, 1979). During the long depolarizations the membrane is electrically very conductive but gains resistance gradually over tens of seconds (Satow et al. 1976). This change parallels the gradual inactivation of the Na inward current, ∼ 10 s after the step depolarization under the voltage clamp (Fig. 6). A decrease in the Na-conductance could be part of the mechanism which terminates the long depolarization and, therefore, long backward swimming, probably augmented by the loss of the electromotive force for Na⁺ resulting from internal accumulation (Hansma, 1979).

The Ca-induced Na-current and the spontaneous long depolarizations are easily triggered in the paranoiac mutants. We have, therefore, exploited this opportunity to study these phenomena using these mutants. However, such phenomena also exist in the wild type. Although the observable slow inward current appears to be smaller and is often obscured by the outward current, the slow inward tail current after the repolarization is clearly seen in the wild type and such tails have similar decay kinetics as those of the paranoiac mutant (Fig. 2). Such tail inward currents become larger when the wild-type cells are cultured in a Na-free medium. Na-related long depolariRtion and prolonged backward swimming can also be observed in the wild type pecially when first exposed to a high concentration of external Na⁺ and/or after growing in Na⁺-free medium. Prolonged depolarizations in Na-solutions can be spontaneous or can be triggered by either a train of closely-spaced short pulses or prolonged current injection in the wild type. Such prolonged depolarizations are retained although the long backward swimming is seldom observed in free-swimming wild types, in near physiological conditions. This difference suggests that certain constraints or damage by the low-resistant electrodes favour the generation of the prolonged depolarizations. It is commonly observed that extreme ionic imbalances (e.g. relatively low external Ca²⁺ concentration), noxious chemicals (detergents, proteases, fixatives, etc.), excessive mechanical stress (microsurgery, microinjection, forcing through micropipette) or attacks of predators (suctorians, Didinium) all cause the wild-type paramecium to spin rapidly backward continuously for seconds to minutes in various Na⁺-containing culture media (Cerophyl, hay-infusions, fresh water) or wash solutions (Dryl’s, Pringsheim’s). A common trigger is likely to be increased internal Ca²⁺ which is prolonged by repeated Ca-action potentials or by membrane damage. The adaptive value of this rapid backward swimming could be in mediating direct escape from a source of threat. This behaviour is distinct from the transient avoiding reactions, which, like the tumblings in bacteria (Springer et al. 1979) accounts for various taxes through the biased random walk (Jennings, 1906; Dryl, 1973; Van Houten, 1978; Hennessey & Nelson, 1979). The continuous backward swimming and the transient avoiding reaction are mechanistically different. The former is based on the long depolarization in second or minute time scale, is due to Ca-induced conductances and involves external Na⁺, while the latter is in millisecond scale, due to voltage-sensitive channels and does not involve Na⁺.

This study takes advantage of ‘paranoiac’ mutant because it is able to generate prolonged depolarizations easily in the Na-solution and therefore allows us to investigate the electrophysiological parameters related to prolonged depolarizations. The findings apply to the wild type since it also exhibits such a current under certain circumstances. They do not, however, identify the specific defect in the paranoiac mutant since a variety of factors (from the retardation of the Ca-pump to alterations of the number or topography of the Na channel) can lead to ‘paranoia’ according to the above hypothesis of the generation of prolonged depolarization.

We thank Drs M. Forte, A. D. Murphy and Y. Satow for their comments and suggestions. This work was supported by NSF grants BNS 77-20440 and BNS 79-18554.