Bistable Membrane Potential of the Ciliate Coleps Hirtus

Within the eight different classes of ciliates, approximately 8000 species have so far been described (Corliss, 1994). However, electrophysiological properties have only been studied in a surprisingly small number of species, mainly within the genera Paramecium (Eckert and Naitoh, 1972), Stylonychia (Machemer and Deitmer, 1987), Didinium (Pape and Machemer, 1986; Pernberg and Machemer, 1989) and Euplotes (Lueken et al., 1996). The presence of voltage-sensitive Ca²⁺ channels is a general feature in these ciliates, and the direction of the ciliary beat is controlled by the cytosolic Ca²⁺ concentration (Machemer and Sugino, 1989; Pernberg and Machemer, 1995a,b). The detailed electrical membrane properties, however, show great variation among different ciliate classes. In most cases, depolarising current injections evoke regenerative Ca²⁺ spikes lacking all-or-nothing characteristics in normal solution, for instance in Paramecium (see Machemer, 1988), belonging to the class Oligohymenophorea. A striking exception to this pattern is found in Bursaridium, from the class Colpodea, which generates spontaneous all-or-nothing action potentials with a distinct threshold in normal solution (Berg and Sand, 1994). Spontaneous action potentials with a plateau phase lasting up to 3–4 s have also been reported in Stylonychia in solutions containing artificially high concentrations of Ca²⁺ (Machemer, 1970).

The ciliate genus Coleps belongs to the class Prostomatea, and no species within this class has previously been studied using electrophysiological methods. The swimming pattern of C. hirtus in normal solution consists of forward swimming covering relatively large distances interrupted by periods of circular swimming within a small area. We have made current-clamp recordings from C. hirtus using standard microelectrodes, and report exceptional electrophysiological properties of this ciliate. The membrane potential showed spontaneous shifts between two semistable levels. The more negative level corresponds to the traditional resting potential, whereas the less negative level is a Ca²⁺-dependent plateau potential. The transition between these membrane potential levels was correlated with the reversal of the direction of the ciliary beat and may explain the characteristic swimming behaviour of this ciliate.

Coleps hirtus Nitzsch is a barrel-shaped freshwater ciliate feeding on dead or living algae, flagellates, rotifers or other protozoa. The genus is characterised by an elaborate endoskeleton consisting of superficial rows of calcium carbonate plates (Fig. 1). The animals were obtained from a lake near Oslo and kept in culture. They were fed on the planktonic alga Rhodomonas minuta, as described by Klaveness (1984). The algae were grown in standard growth medium (Guillard and Lorenzen, 1972) at 17 °C with a 14 h:10 h light:dark photoperiod. The mean length of the animals in our cultures was 42 μm and the mean width was 29 μm (N=60). The cultures of Coleps hirtus and Rhodomonas minuta were generously provided by Dr Dag Klaveness.

Prior to the recordings, the animals were transferred from the growth medium to a recording solution containing (in mmol l⁻¹): CaCl₂, 1; NaCl, 1; KCl, 1. The solution was adjusted to pH 7.2 with 1 mmol l⁻¹ Tris/HCl. In some experiments, the CaCl₂ in this normal recording solution was replaced with BaCl₂, CoCl₂ or MgCl₂. In one series of experiments, 1 mmol l⁻¹ CsCl was added to the normal solution. During the experiments, the animals were viewed through an inverted microscope.

The electrical properties of the surface membrane were studied using conventional microelectrodes and standard recording equipment. The electrodes were filled with 4 mol l⁻¹ potassium acetate adjusted to pH 7.2 with acetic acid, and the electrode resistance was 40–80 MΩ. During the recordings, the animals were kept in position using a microsuction pipette with a tip diameter of 10–15 μm. The suction pipette was connected to ambient, subatmospheric or superatmospheric pressure via a solenoid valve (Jonsson and Sand, 1987). A selected animal was sucked onto the tip of the pipette by activating the solenoid valve, using a manual trigger. The subatmospheric catching pressure was between −1.0 and −1.5 kPa, whereas the holding pressure during the recordings was reduced to −0.5 kPa. The animals are coated with a gel-like material, which tended to clog the pipette. Between each catch, the pipette was therefore cleaned using a short flush of medium through the tip. This was achieved by connecting the pipette to a pressure of approximately 5 kPa.

The movements of both freely swimming animals and ciliates penetrated by microelectrodes were studied using a video camera recording at 25 frames s⁻¹. The general direction of the ciliary beat of penetrated cells was determined in a suspension of carmine particles. A rough estimate of the beat direction was obtained from the movements of the insoluble dye grains (diameter 0.5–1 μm) surrounding the cell. Simultaneous recordings of the membrane potential and animal and particle movements were stored on separate channels of the same recorder.

Results are reported as means ± S.D.

In normal recording solution, the swimming behaviour of C. hirtus consisted largely of two alternating main patterns: forward swimming and circular movements. The forward swimming usually displayed a component of lateral oscillations of small amplitude. Swimming speed was 0.4–1.0 mm s⁻¹. At irregular intervals, the periods of forward swimming were interrupted by circular movements within a restricted area (Fig. 2). The duration of these two swimming patterns varied from a few seconds to more than a minute.

The main purpose of the present investigation was to elucidate the cellular mechanisms behind this characteristic swimming pattern. In other ciliate species, the direction of the ciliary beat is controlled by the cytosolic Ca²⁺ concentration and, hence, the membrane potential. The natural approach was therefore to study the electrophysiological properties of C. hirtus.

The initial recordings revealed a bimodal distribution of the membrane potential in normal solution. In approximately 70 % of the cells, the membrane potential was between −50 and −60 mV (deep membrane potential), whereas the remaining cells displayed membrane potentials between −20 mV and −30 mV (shallow membrane potential). To study the resistance, capacitance and rectifying properties of the membrane, hyper- and depolarising current pulses of varying magnitude and 250–500 ms duration were injected, and the data were plotted as current/voltage (I/V) diagrams. Fig. 3 presents the I/V relationship for a cell displaying a resting potential of −48 mV during the recording period. The potentials recorded 200 ms after the start of current injections are plotted in the graph, and the inset shows a sample of superimposed original recordings. The slope of the I/V relationship was nearly linear between −60 and −30 mV, corresponding to a constant membrane resistance. The mean input resistance of the cells in this potential range was 386±141 MΩ (mean ± S.D., N=61). The mean value of the membrane time constant measured from potential deflections within this range was 48±19 ms (N=42), giving a mean membrane capacitance of 114±35 pF (N=42).

The I/V relationship shows marked outward rectification for potentials above −25 mV and pronounced inward rectification for potentials below −60 mV. The potential deflections with large hyperpolarising current injections reached a peak value before settling at a stable level towards the end of the pulse (inset) because of the time-dependence of the activation of the ion channels responsible for the inward rectification. These peak values are also included in the graph. The inward rectification was only moderately depressed by 1 mmol l⁻¹ Cs⁺ (data not shown), but was completely blocked by 1 mmol l⁻¹ Ba²⁺ (see Fig. 9).

On the basis of previous studies of the electrophysiological properties of ciliates, current-induced, graded Ca²⁺ spikes or all-or-nothing Ca²⁺-dependent action potentials might have been expected in response to depolarising current injections. However, in C. hirtus, depolarising current injections evoked membrane responses more like passive potential deflections than active, regenerative responses. Fig. 4 presents recordings from two different cells, with stable membrane potentials of −50 mV (Fig. 4A) and −20 mV (Fig. 4B). It is evident from Fig. 4A that the response to depolarising current injections cannot be completely explained by the passive electrical properties of the membrane. The cell displayed an after-depolarisation outlasting the current injection by several hundred milliseconds; a similar active component of the membrane response was evident in approximately 90 % of the cells with a deep membrane potential. After-depolarisations were never observed in cells with a shallow membrane potential. Depolarising current injection into these cells were instead usually followed by a moderate after-hyperpolarisation (Fig. 4B).

The shape and duration of the after-depolarisation varied greatly among cells. Fig. 5 presents a recording from a cell with a stable membrane potential of approximately −60 mV. A small, depolarising current injection caused a passive membrane response, whereas larger current injections evoked after-depolarisations of increasing duration. At a level of current injection exceeding 70 pA, the post-stimulus membrane potential stayed at a stable, depolarised level for approximately 19 s. Depolarising current injection may therefore induce a prolonged, depolarised plateau phase of the membrane potential. It is therefore reasonable to explain the bimodal distribution of the membrane potential in C. hirtus in the following manner. The proper resting potential is relatively deep and is within the range of resting potentials observed in other ciliates. Under certain conditions, C. hirtus may generate active membrane depolarisations of extremely long duration. These plateau potentials are dependent on voltage-sensitive ion channels and correspond to the action potentials observed in some species of ciliate. However, in C. hirtus, the plateau phase of the action potential may be prolonged to such anextent that the membrane potential of the cell is best described as bistable or biphasic.

During long-lasting recordings of several minutes, spontaneous shifts between shallow and deep membrane potential levels were frequently observed, supporting the theory outlined above. An example of a shift from the shallow to the deep level is shown in Fig. 6, which also presents potential responses to depolarising current injections at both the shallow and the deep membrane potentials. Initially, the membrane potential was approximately −30 mV, and depolarising current injection evoked a nearly passive response lacking an after-depolarisation. Shortly after the injection, the membrane potential spontaneously hyperpolarised and stabilised at a level of approximately −50 mV. From this level, depolarising current injection evoked a potential response with the characteristic after-depolarisation. The time course of the spontaneous shift from the shallow to the deep membrane potential was similar to the shape of the after-depolarisation, supporting the idea that the shallow membrane potential level and the after-depolarisation are dependent on a similar, active component. Depolarisations from the shallow level will then lack an after-depolarisation, since the active component is already activated.

In the cell depicted in Fig. 7, hyperpolarising current injections from a shallow membrane potential induced a shift of the membrane potential to a semi-stable deep level. In normal medium, recordings were obtained from 94 cells, and shifts between the shallow and deep membrane potential levels could be induced by depolarising or hyperpolarising current injections in approximately 50 % of the cells. The mean value of the deep potential level, corresponding to the resting membrane potential, was −52±6 mV (N=69). The mean value of the shallow membrane potential level, or the plateau potential, was −29±6 mV (N=65). The mean value of the observed shifts between these two semi-stable levels of the membrane potential was 23±6 mV (N=47). In recordings in which current injections were not performed, the membrane potential was at the deep level for 78 % of the total recording time.

Fig. 8 shows potential responses to hyperpolarising current injections. The first injection was performed at the deep membrane potential level, and the potential response was passive. However, after termination of the hyperpolarising response, the membrane potential spontaneously shifted to a stable, shallow level. The hyperpolarisation evoked by current injection at this plateau potential clearly deviated from a passive response. Closure of voltage-sensitive ion channels during the hyperpolarizing phase reduced the membrane conductance and accelerated the hyperpolarization (Fig. 8, arrow a). Opening of voltage-sensitive ion channels during the repolarising phase (Fig. 8, arrow b) induced the regenerative response responsible for the plateau potential.

On the basis of the known electrophysiological properties of other ciliates, it is reasonable to suggest that the depolarisation from the deep membrane potential to the plateau potential is due to Ca²⁺ influx through voltage-sensitive Ca²⁺ channels. To test this hypothesis, the Ca²⁺ in the recording solution was replaced with either Co²⁺, which blocks Ca²⁺ channels, or Mg²⁺, which cannot permeate the channels. In solutions containing Co²⁺, the resting membrane potential was −57±5 mV (N=4), and in solutions containing Mg²⁺, the resting membrane potential was −50±4 mV (N=7). Potential responses to depolarising current injections were passive, with no sign of after-depolarisations. Neither electrically induced plateau potentials nor spontaneous shifts to a shallow membrane potential were observed. The complete elimination of plateau potentials and membrane potential shifts in these solutions supports the hypothesis that voltage-sensitive Ca²⁺ channels are responsible for the biphasic membrane properties of C. hirtus.

The voltage-sensitive Ca²⁺ channels are permeable to Ba²⁺, which is also a general inhibitor of K⁺ channels. Furthermore, while Ca²⁺-dependent inactivation of Ca²⁺ channels is an important element in the termination of Ca²⁺-dependent depolarisations, Ba²⁺ does not inactivate Ca²⁺ channels. If Ca²⁺ channels were involved in the activation of the plateau potential, this shallow membrane potential would be expected to dominate in a recording solution in which Ca²⁺ is replaced with Ba²⁺. In such a solution, the membrane potential was −14±9 mV (N=27) and spontaneous shifts to a deeper potential level never occurred. Fig. 9 presents the I/V relationship for a cell in Ba²⁺-containing solution. Compared with the I/V relationship in Fig. 3, the general blocking of K⁺ channels is evident from the elimination of both inward and outward rectification.

The hypothesis described above predicts that the shallow membrane potential is due to a continuous inward Ba²⁺ current through the Ca²⁺ channels and it should, therefore, be possible to terminate the plateau potential with a brief injection of hyperpolarising current. Fig. 10 shows an example in which such a treatment shifted the membrane potential to a deep level for a few seconds, after which the cell spontaneously depolarised to the stable plateau potential. Similar experiments were performed on 14 additional cells, and the mean value of the short-lasting, deep membrane potential following the hyperpolarising current injection was −65±14 mV (N=15). The mean value of the shifts between this deep level and the plateau potential was 52±12 mV. In other ciliates that exhibit graded Ca²⁺ spikes, such regenerative responses may be transformed to all-or-nothing action potentials in Ba²⁺-containing solution. We therefore investigated whether Ba²⁺-dependent action potentials can also be evoked in C. hirtus. Since the membrane potential in Ba²⁺-containing solution was stable at the shallow level, at which the Ca²⁺ channels are activated, these experiments were performed during continuous injection of hyperpolarising current. This treatment forced the membrane potential to stay at a hyperpolarized level, at which the Ca²⁺ channels are closed. Brief depolarising current pulses superimposed on the direct holding current then evoked all-or-nothing action potentials with a prolonged plateau coinciding with the shallow potential (Fig. 11).

To determine the direction of the ciliary beat, intracellular recordings were performed after the addition of the hydrophobic dye Carmine Red to the solution. In the initial experiments, the membrane potential was monitored without injecting current, and movements of the cilia were correlated with potential level. At the deep membrane potential, the ciliary power stroke had a backward direction, corresponding to forward swimming of unrestricted cells. When the membrane potential spontaneously shifted to the shallow level, the power stroke reversed to a forward direction. Similar changes in the ciliary beat were observed when the membrane potential was shifted between the two semi-stable levels by current injections. It is reasonable to assume that the periods of forward swimming, during which the animals may cover fairly large distances, are correlated with the deep membrane potential level, whereas the circular movements are associated with the plateau potential. To clarify this question, swimming behaviour was studied in a solution in which Ca²⁺ was replaced with Mg²⁺. The plateau potentials are absent in this solution, and the membrane potential is stable at the deep level.

In this solution, the animals only displayed forward swimming, without the characteristic, brief periods of circular swimming observed in normal medium.

The mean resting membrane potential (deep potential level) of C. hirtus in normal recording solution is −52 mV. Although the reported resting membrane potentials in most ciliates are between −30 mV and −40 mV, the value in C. hirtus is not exceptional. Other ciliates with rather negative resting membrane potentials are Bursaridium difficile (−45 mV; Berg and Sand, 1994), Stentor coeruleus (−49 mV; Wood, 1992), Stylonychia mytilus (−51 mV; Deitmer, 1981) and Didinium nasutum (−53 mV; Pape and Machemer, 1986).

The mean membrane resistance of C. hirtus is 386 MΩ. At the resting potential, the conductance of the ciliary membrane is negligible compared with that of the soma membrane (Pape and Machemer, 1986). The surface area of a barrel-shaped body with the dimensions of C. hirtus, 42 μm×29 μm, is approximately 4.7×10⁻⁵ cm². The specific resistance of the soma membrane is therefore estimated to approximately 1.8×10⁴ Ω cm². This value is comparable with the membrane resistance of Paramecium caudatum (1.9×10⁴ Ω cm²; Machemer and Ogura, 1979), Euplotes vannus (2.3×10⁴ Ω cm²; Krüppel and Lueken, 1988) and Didinium nasutum (2.8×10⁴ Ω cm²; Pape and Machemer, 1986), but considerably smaller than the specific membrane resistance of Stentor coeruleus (5.3×10⁴ Ω cm²; Wood, 1982).

The mean membrane time constant of C. hirtus is 48 ms, corresponding to a membrane capacitance of 114 pF. The membrane time constant shows little variation among different ciliate species, and previously reported values are in the range 40–60 ms (Naitoh et al., 1972; Pape and Machemer, 1986; Machemer and Deitmer, 1987; Berg and Sand, 1994).

C. hirtus showed both outward and inward membrane rectification, in common with other ciliates. In Paramecium, the outward rectification is mainly due to Ca²⁺-activated conductances and is therefore secondary to Ca²⁺ influx through voltage-gated channels. The inward rectification in Paramecium is due to at least five distinct ion conductances (Oertel et al., 1978; Richard et al., 1986; Saimi, 1986; Hennessey, 1987; Preston et al., 1990, 1992), some of which are Ca²⁺-dependent and secondary to Ca²⁺ influx through channels activated by hyperpolarisation. The ion conductances responsible for the rectification in C. hirtus were not investigated further in the present study. However, it is interesting to note that Cs⁺, which inhibits most types of inward rectification in metazoans (Castle et al., 1989), did not block the inward rectification in C. hirtus.

We have shown that the membrane potential in C. hirtus has a bimodal distribution, with a deep level of approximately −50 mV and a shallow level of approximately −30 mV. In normal recording solution, spontaneous shifts between these levels occur regularly, with the animals spending approximately 80 % of the time at the deep level. Similar biphasic membrane properties in ciliates have not been reported previously. The direction of the ciliary beat is backwards at the deep potential, corresponding to forward swimming, and the direction is reversed at the shallow potential. These findings may explain the swimming behaviour of C. hirtus, which displays irregular periods of linear forward swimming interrupted by periods of circular movements within a small area. It is reasonable to suggest that favourable environmental conditions, such as an abundance of food, may promote the latter swimming pattern, but this hypothesis has not been tested in behavioural experiments.

The shallow membrane potential level was absent in Ca²⁺-free solution containing Mg²⁺ or Co²⁺, which do not permeate the Ca²⁺ channels, and the animals spent 100 % of the time at the shallow potential level in a solution in which Ca²⁺ had been replaced with Ba²⁺. This ion may act as a charge carrier through Ca²⁺ channels and is also a general inhibitor of K⁺ channels. The graded, regenerative Ca²⁺-dependent depolarisations reported in other ciliates may thus be transformed to all-or-nothing action potentials in the presence of Ba²⁺ (Paramecium caudatum, Naitoh and Eckert, 1968; Stylonychia mytilus, de Peyer and Deitmer, 1980; Didinium nasutum, Pape and Machemer, 1986; Euplotes vannus, Krüppel and Leuken, 1988), and a similar phenomenon was observed in C. hirtus. If the membrane potential was kept at a hyperpolarised level in Ba²⁺-containing solution by direct holding current, depolarising current injections evoked prolonged action potentials with distinct thresholds.

On the basis of the electrophysiological data, we conclude that the deep membrane potential in C. hirtus corresponds to the traditional resting potential, while the shallow potential level is a plateau potential maintained by non-inactivating inward Ca²⁺ current through voltage-sensitive Ca²⁺ channels. It is interesting to note that, although similar electrophysiological properties have not been observed in other ciliates, bistable membrane potentials have previously been reported in vertebrate motoneurones (Hounsgaard and Kiehn, 1989). In these neurones, serotonin induces a sustained, Ca²⁺-dependent plateau potential associated with a state of enhanced excitability. The plateau potential is initiated by a combination of reduced K⁺ permeability and a non-inactivating Ca²⁺ current. The primary effect of serotonin is probably inhibition of Ca²⁺-dependent K⁺ channels.

We thank Dr Dag Klaveness for providing the cultures of Coleps hirtus and for valuable discussions.