ABSTRACT
The swimming behaviour of Paramecium was analyzed under natural and experimental hypergravity conditions. Paramecium that swam upwards (in the opposite direction to the gravitational force) along a straight path (straight swimmers) swam more slowly than those swimming downwards. This dependence of the swimming velocity on its direction relative to gravity can be partly interpreted as the consequence of sinking due to gravity if the propulsive force does not vary. The effect was different for Paramecium swimming along a circular path (curved swimmers). The difference in velocity between those swimming upwards and those swimming downwards was substantially smaller than would have been expected from sinking effects with invariant propulsion even after correcting for maximal hydrodynamic wall effects, indicating that Paramecium compensate for sinking caused by gravity by controlling their propulsion. The propulsive velocity evaluated by vector calculus increased both as Paramecium swam more sharply upwards and as the experimental gravitational force was increased. The dependence of propulsion on the swimming direction and on gravity was reduced in a high-density medium of nearly neutral buoyancy, suggesting that the site of gravireception is unlikely to be in the interior of the cell. The differences between straight and curved swimmers are discussed in terms of rapid adaptation of gravireceptors in the cell membrane, desensitization of mechanosensory channels and hyperactivation of ciliary activity in straight swimmers.
Introduction
It has been suggested that ciliates are capable of modulating their swimming behaviour by controlling ciliary activity in response to sensory input (Machemer, 1988; Naitoh, 1984; Poff et al. 1984). This sensory input may be chemical, tactile, thermal or optical. A possible role of sensory input from gravitational acceleration (gravireception) has also been postulated for loxodid Protozoa, which possesses vacuoles containing concretions of a barium compound (Finlay et al. 1983; Rieder et al. 1982) that bear a morphological resemblance to hydrozoan statoliths (Bedini et al. 1973; Horridge, 1969; Markl, 1974) and come into close contact with the vacuolar membrane when the cell is rotated in the gravitational field (Fenchel and Finlay, 1986). Capillary tube inversion experiments with a horizontal microscope have demonstated that individual cells of Loxodes striatus swimming upwards under anoxic conditions react to rapid upside-down inversion by turning them-selves back to their original direction of swimming; cells under aerobic conditions prefer downward swimming and show similar turning responses to changes in the direction of the cell body relative to gravity (Fenchel and Finlay, 1984).
Evidence for gravireception and candidates for gravireceptive organelles in non-loxodid Protozoa, e.g. Paramecium, have also been looked for (Bean, 1984; Machemer and de Peyer, 1977; Machemer et al. 1991; Markl, 1974; Mogami et al. 1988a). Our present analysis of the swimming velocity of Paramecium caudatum in experimental gravitational fields reveals gravity-induced changes in propulsion, and suggests that ciliary activity is modified by sensory input due to gravity. In view of the results from similar experiments but in a high-density medium, we propose a new model for gravireception in Paramecium: the anterior and posterior membranes react by depolarizing and hyperpolarizing, respectively, in response to stretching as a result of a local pressure difference caused by the density difference between the interior and the exterior of the cell when it is placed in a gravitational field. This shift in the membrane potential may modulate ciliary activity, as it does in voltage-clamped cells (Machemer, 1988). This causes an increase in speed, resulting from hyperpolarization, in cells swimming upwards and a decrease in speed, resulting from depolarization, in those swimming downwards in accordance with the present results and those of recent experiments with galvanotactically aligned Paramecium (Machemer et al. 1991). A possible role for gravity-induced modulation of the membrane potential and for its coupling to the ciliary beat polarity (Machemer, 1988) and to the pitch angle of helical trajectories (Machemer, 1989,1990; Machemer and Sugino, 1989; Crenshaw, 1990) in negative geotactic orientation is also discussed. Parts of this work have been presented elsewhere (Baba et al. 1989a,b;,Izumi-Kurotani et al. 1989; Ooya et al. 1988, 1989a, b, c).
Materials and methods
Culture and preparation
All experiments were carried out with Paramecium caudatum Ehrenberg (syngen 3, mating type V, a gift from Dr M. Takahashi of Tsukuba University), grown at 24°C in a hay infusion in Dryl’s medium (2 mmol l−1 sodium citrate, 1.2 mmol l−1 Na2HPO4,1.0 mmol l−1 NaH2PO4,1.5 mmol l−1 CaCl2, pH 7.1; Dryl, 1959). Cells grown to the stationary phase (10–14 days after incubation) were collected through a nylon mesh (pore size 42 μm), washed in fresh Dryl’s medium and adapted to room temperature (23–24.5 °C) before use (normally at least 1 h). In some experiments, cells were adapted in high-density Dryl’s medium containing Percoll prepared by diluting a 23 % (w/w) colloidal solution of Percoll (Sigma) to 25 % v/v (specific gravity 1.033, relative viscosity 1.12 at 24.5°C).
Paramecium in these media were transferred into a chamber made of a slide and/or cover slip and silicone rubber spacer (inner dimensions 13mm× 13 mm × 1mm) and kept free of air bubbles with no particular sealant. They remained healthy for several hours under these conditions.
Recording and analysis of the swimming paths
The swimming behaviour of Paramecium in the chambers was recorded either by multiple-exposure photography under dark-field stroboscopic illumination (Dagawa et al. 1986) or by video-microscopy with a charge-coupled device (CCD) camera (XC-77, Sony, Tokyo) and a video tape recorder (VTR, EDV-5000, Sony) under bright-field illumination with a light-emitting diode (MU02-2205, Stanley, Tokyo). The multiple-exposure photographs were analyzed as described pre-viously (Degawa et al. 1986). The video tape recordings were analyzed by means of digital image analysis as described in detail elsewhere (Baba et al. 1991). Successive frames from the VTR at constant time intervals (usually 0.6 s) were fed into an image processor (Nexus 6510, Nexus, Tokyo) to give a multiple-exposure photo-like image, which was analyzed with laboratory-made software (VIDEOB Ver. 1.13) on an IBM 4381-R24 computer.
Hypergravity experiments
Hypergravity was achieved by using a swing-basket-type centrifuge, which had two baskets with inner dimensions of 400 mm × 210 mm horizontal and 320 mm vertical at each edge of 2-m arms. One of the baskets carried Paramecium in the chamber described above and recording equipment, including the CCD camera with the VTR placed outside and connected through slip rings at the centre of the centrifuge. The wider wall of the chamber, to which the optical axis of the horizontal microscope was perpendicular, was vertical with respect to the floor of the basket and parallel to the arm. The resultant force of the earth’s gravitational force and the centrifugal force, which will be referred to as ‘experimental’ gravity or simply gravity, was designed to be perpendicular to the floor of the baskets and was monitored with a load cell carried in the second basket. It was found that the experimental gravity was steady at scheduled set-points in the range 2 – 5 g and was always directed perpendicular to the floor, and hence downwards in the picture taken through the horizontal microscope. Recording for 1g was performed inside the basket before and after operation of the centrifuge.
Sinking velocities under hypergravity conditions were also measured using nickel-immobilized cells. When the experimental gravity had reached a given value, the chamber was gently inverted, with the axis of rotation being kept on the optical axis of the horizontal microscope by means of a controlled d.c. motor. The sinking velocity was, however, routinely evaluated by extrapolation from the mean value measured under 1-g conditions, since that measured after the initial perturbation had settled was found to be exactly proportional to gravity at least up to 5g (not shown).
Results
Long-term adaptation
Paramecium swam at 1 – 2 mm s − 1 frequently (about 80 % of observations) along a straight line (short-diameter helical path) immediately after they were transferred into an experimental chamber. After adaptation (usually Ih), they swam along either a straight or a circular path (long-diameter helical path) with nearly equal frequency. The swimming velocity decreased to a steady level within this period of adaptation and remained steady, as reported by other authors (Machemer, 1989; Oka et al. 1986). It was this velocity that was measured.
Gravity was increased step by step with ample time for recording at each steady state and then decreased in a similar way. There was no systematic variation in the results between experiments in which gravitational force was increased and those in which it was decreased. Therefore, data from cells under experimental gravitational conditions were not thought to be affected by long-term hysteretic effects of gravistimulation (stimulation by gravity) and will be referred to only by the value of the gravitational force.
The effect of gravitational force on the swimming velocity of individual cells
Individual cells swimming along a circular path, usually viewed from the top or bottom of the helical axis, were found to maintain their average vertical position (Fig. 1A). Consistent downward drifts were not detected during the observation time. During the same period, immobilized cells fell a substantial distance (Fig. IB). These circular swimmers occasionally even drifted upwards (not shown).
These cells enabled us to analyze the swimming vector velocity for a full range of values of θ. As might be expected from the observations described above, ν remained nearly constant, far from the variation predicted by equation 1 (Fig. 2) This constancy in ν with respect to θ suggests that gravity-induced changes in propulsive force exist and can compensate for the effect of sinking due to gravity.
The effect of gravitational force on the swimming velocity of a population of Paramecium specified by path curvature
Preliminary analyses of ν vs θfor a given population of cells revealed that ν varies as an inverse sigmoid function of θ, as expected from equation 1. The discrepancy between this result and the finding described above for single cells led us to the idea that gravity-induced changes in propulsion may depend on path curvature. Under either natural or experimental gravitational conditions, a plot of ν vs θ for a population specified by a path curvature of less than 0.1 mm’1 (cells in this population will be referred to as straight swimmers) nearly matched that predicted by equation 1 (Fig. 3A), while v for cells with a curvature greater than 0.85 mm−1 (curved swimmers with about one-third of the curvature of the cell shown in Fig. 1A) remained nearly constant when plotted against θ (Fig. 3B). The critical curvatures were selected arbitrarily, but equal numbers of cells were included in each population. Sorting by curvature might select cells that tend to be close to the chamber wall and hence are more subject to a reduction in s, as a result of an increase in viscous resistance from hydrodynamic wall effects. However, a modified form of equation 1 with a 0.435-fold reduction in 5 did not match the data for curved swimmers. This reduction is the maximum possible assuming that all immobilized cells fell midway between opposite chamber walls and that all curved swimmers swam in contact with a wall (based on calculations using equations cited in Mogami et al. 1988b). In addition, preliminary analyses of the motion of cells far from chamber walls by means of three-dimensional video-microscopy demonstrated plots of ν vs θ similar to those in Fig. 3B (Izumi-Kurotani et al. 1989). It is therefore unlikely that the deviation from equation 1 of the measurements for curved swimmers is due to hydrodynamic wall effects.
The deviation of the data from equation 1 was measured by plotting (νdn–νup)/2 against s, with s being varied by changing the gravitational force (Fig. 4). If the propulsive force were invariant, irrespective of the swimming direction and gravity, (νdn– νup)/2 should be equal to s, as indicated by the dashed line in Fig. 4. The downward shift of the data points from the dashed line, Δ of Machemer et al. (1991), is assumed to be a measure of the gravity-induced active control of propulsion. It should be noted that this shift for curved swimmers under natural gravity conditions is about 750 μms−1 (from data shown in Fig. 4), ten times as large as that reported for cells whose orientation has been fixed galvanotactically (Machemer et al. 1991) (see Discussion).
In a high-density Percoll medium, which nearly neutralizes the buoyancy of cells, horizontal swimming velocities (θ =90±30°) of curved and straight swimmers under natural gravitational conditions were 0.461±0.081 mm s°1 and 0.701±0.135mms−1, respectively (mean±s.D., N=13 replicates of experiments similar to that shown in Fig. 3). These velocities were not significantly different from those measured in a normal-density medium, 0.506± 0.042mms−1 and 0.619±0.098mms−1 for curved and straight swimmers, respectively (N=18, P>0.05). Analyses similar to those shown in Fig. 3 were made with cells in the Percoll medium under natural and experimental gravitational conditions. However, we found that it was more informative to analyze propulsive velocities, as described below.
Gravity-induced changes in calculated propulsive velocity
The preceding analyses revealed that the propulsive velocity p tends to be either increased or decreased as cells swim upwards or downwards, respectively, probably by gravistimulation depending on the cell orientation. The magnitude of the propulsive velocity, p, increases with increasing values of ϕ (more-upward swimming) and decreases with decreasing values of ϕ (more-downward swimming) (Fig. 5). This tendency becomes more prominent as gravitational force increases and is always larger for curved swimmers (compare Fig. 5A and 5B), as might be expected. Regression lines for data from cells under natural and experimental gravitational conditions cross at a value of about 120°. This indicates that the degree of suppression of propulsion induced by gravistimulation of downward-orienting cells is larger than that of potentiation of propulsion of upward-orienting cells. It should be noted that this feature of the gravisensitivity of cells is consistent with the findings of Machemer et al. (1991), i.e. Δa>Δp using their terminology. The slopes of the regression lines for p vs ϕ, which is an overall measure of orientation-induced changes in propulsion and will be referred to as Δp/Δ ϕ, increased linearly with gravitational force (Fig. 6). This indicates that the orientation-induced changes in propulsive force are genuine responses to gravistimulation.
Fig. 6 also shows that Δp/Δ ϕ decreases significantly for curved swimmers when the cells are immersed in a high-density medium and that Δp/Δ ϕ for straight swimmers is relatively insensitive to the medium density. It should be noted that Δp/Δ ϕ for curved swimmers drops when buoyancy is neutralized to the level for straight swimmers.
Discussion
The present analysis of the swimming velocity of Paramecium under natural and experimental gravitational conditions revealed direction-dependent changes in propulsion that are responsible for the adjustment of the speeds of upward-and downward-swimming cells. One possible explanation is that the cells can sense their speed relative to the water from the changes in drag experienced by cilia (Mogami et al. 1988b). It has been demonstrated that a natural or experimental tether changes flow around swimming planktonic organisms (Emlet, 1990). It therefore seems plausible that changes in the forces on beating cilia could be the signal that the animal is sinking or rising. Using this signal, Paramecium could increase its propulsion when swimming upwards and decrease it when swimming downwards, thus keeping its speed constant. Regulation of this sort by negative feedback would be readily achieved when the loads are small. However, the compensation for sinking was more complete at 5 g than at 3–4g (see Fig. 4).
An alternative explanation is that the cells can sense their orientation relative to gravity by the changes in stress due to gravity experienced by the body. If the site of gravireception is in the interior of the cell, e.g. in Müller vesicles in Loxodes (Fenchel and Finlay, 1984, 1986), in the nucleo-nucleolar sensor (Moroz, 1984), in the concretion vacuole (Anderson and Dumont, 1966) or other statolith/statocyst-like structures, the magnitude of gravistimulation will be determined by the mass of the statolith-like structure, the density of the cellular fluid surrounding it and by gravity, and hence cannot be influenced directly by the external medium. If the site of gravireception is exposed to the exterior of the cell, the magnitude of stimulation will be determined by a balance between the internal and external fluid densities and gravity. When the external density approaches the internal density, the magnitude of stimulation, and hence the orientation-induced changes in propulsion, will decrease to zero. Therefore, the observation that Δp/Δ ϕ for curved swimmers decreased substantially when they were immersed in a nearly neutrally buoyant medium indicates that most of the sites of gravireception are exposed to the exterior of the cell. The plasma membrane itself is a primary candidate for this ‘exposed’ gravireceptor; its possible roles in gravireception have been discussed for cytoplasmic streaming in Nitellopsis obtusa (Wayne et al. 1990) and for geotactic reorientation in Paramecium (Machemer and de Peyer, 1977). We will discuss only this type of gravireception, although the residual Δp/Δ ϕ after neutralization of buoyancy for curved swimmers and the insensitivity of Δp/Δ ϕ to gravitational force in straight swimmers suggest that some internal gravireceptors may exist.
We propose a new model for the control of ciliary activity by the cell membrane through sensory input due to gravity, which will result in gravikinesis (regulation of propulsion by gravity) and gravitaxis (regulation of trajectory direction with respect to gravity).
Upward swimming
Up1. When the cell swims upwards or obliquely upwards, the posterior cell membrane is stimulated by gravity, because the membrane tension is increased as a result of the pressure difference between the exterior and the interior of the cell.
Up2. In response to this stimulation, mechanosensitive K+ channels, whose sensitivity is maximal at the posterior end of the cell (Ogura and Machemer, 1980), increase their open probability and, hence K+ efflux increases, resulting in hyperpolarization (Naitoh, 1984; Naitoh and Eckert, 1969).
Up3. Hyperpolarization induces changes in ciliary activity; an increase in beating frequency and a more posterior beating direction together increase propulsion (Machemer, 1976) (gravikinesis).
Up4. Hyperpolarization-induced changes in ciliary activity also cause changes in the rate of cell rotation (rolling, yawing and possibly pitching), resulting in a reduction of the pitch angle of the helical trajectory (Machemer, 1989, 1990; Machemer and Sugino, 1989), which will reorientate the axis of the path helix closer to the direction of the instantaneous swimming velocity vector (Crenshaw, 1989, 1990).
Downward swimming
Dnl. When the cell swims downwards or obliquely downwards, the anterior cell membrane is stimulated by gravity, as a result of increased membrane tension.
Dn2. In response to this stimulation, mechanosensitive Ca2+ channels, whose sensitivity is maximal at the anterior end of the cell (Ogura and Machemer, 1980), increase their open probability and, hence, Ca2+ influx increases, resulting in depolarization (Naitoh, 1984; Naitoh and Eckert, 1969).
Dn3. Depolarization induces changes in ciliary activity; a decrease in beating frequency and a more sideways beating direction together reduce propulsion (Machemer, 1976) (gravikinesis).
Dn4. Depolarization-induced changes in ciliary activity also cause changes in the rate of cell rotation, resulting in an increase in the pitch angle of the helical trajectory (Machemer, 1989, 1990; Machemer and Sugino, 1989), which will reorientate the axis of the path helix farther from the direction of the instan-taneous swimming velocity vector (Crenshaw, 1989, 1990).
Helical swimming
Hx1. Cells swimming along a helical path will experience continuous changes in the direction of the cell body relative to gravity. When the direction of the cell body is upwards, gravity-induced hyperpolarization will orientate the axis of the path helix closer to the instantaneous swimming direction, i.e. upwards (Up4), and when it is downwards, gravity-induced depolarization will orientate the axis of the path helix farther from the instantaneous swimming direction, i.e. downwards (Dn4).
Hx2. This will shift the helical axis of the cell trajectory farther upwards (gravitaxis).
Assumptions Up1 and Dn1 require extremely sensitive mechanoreceptors (stretch receptors?) in the cell membrane of Paramecium. The magnitude of stimulation by gravity can be estimated to be 0.08 Pa at 1 g, the pressure difference calculated from the average density of Paramecium measured with Percoll, 1.033gml−1 [a similar value of 1.04gml−1 has been reported by Taneda (1987) using deuterium oxide] and the cell length (230 μm). Sackin (1989) has demon-strated for Necturus proximal tubule that a negative pressure of 588.6 Pa (6cmH2O) is sufficient to increase the open probability of stretch-activated K+ channels by a factor of 4. Therefore, it is possible that Paramecium mechanorecep-tor channels have a sensitivity four orders of magnitude higher than that of Necturus proximal tubule. It should be noted, however, that evaluation of the proximal tubule’s sensitivity is based on measurements of an ionic current induced by tonic suction using the patch-clamp technique. It is intriguing in this context that the bullfrog sacculus is sensitive to vibratory accelerations four orders of magnitude smaller than the static acceleration due to gravity (Assad et al. 1989). This type of adaptation, if it exists, in gravireceptors may explain the much lower sensitivity to gravistimulation of Paramecium fixed in orientation either by galvanotaxis (Machemer et al. 1991) or swimming straight compared with those changing orientation by curved swimming. The trajectory and higher speed of straight swimmers may indicate a slightly more negative membrane potential compared with that of curved swimmers (Machemer, 1989, 1990; Machemer and Sugino, 1989). This shift in membrane potential might reduce gravisensitivity in straight swimmers by densensitizing either the anterior or the posterior mechano-receptors, since hyperpolarization is known to decrease the Ca2+-dependent depolarizing conductance (Machemer and Deitmer, 1985) and the K+ driving force (Machemer et al. 1991; Ogura and Machemer, 1980). The more hyperpolarized state in straight swimmers may also reduce gravisensitivity by shifting the working range to a saturated zone, where a given shift in the membrane potential induces a smaller change in ciliary activity (de Peyer and Machemer, 1983; Nakaoka and Machemer, 1990). Thus, adaptation, desensitization or saturation of gravireceptors in straight swimmers may underlie the different extent of gravity-induced control of propulsion found in the present study.
Gravity-induced changes in the membrane potential have recently been measured directly with a glass microelectrode inserted into elongating cortical cells of mung bean roots, although direct gravistimulation of these cells has not been verified (Ishikawa and Evans, 1990). Gravity-induced hyperpolarization and depolarization in Paramecium postulated in Up2 and Dn2 and by Machemer et al. (1991) can only be the result of direct gravistimulation of the cell but they remain to be measured directly.
Hyperpolarization- and depolarization-induced changes in ciliary activity have been demonstrated (Up3 and Dn3). However, the changes in the helical variables postulated in Up4 and Dn4 are based on observations of trajectories of cells immersed in hyperpolarizing or depolarizing media only under natural gravity conditions (Machemer, 1989, 1990; Machemer and Sugino, 1989). Measurements of these variables under microgravity (0g) and neutral-buoyancy conditions by means of three-dimensional video-microscopy (Baba et al. 1991) remain to be carried out.
In conclusion, the present study with Paramecium demonstrates a possible role of gravireception in inducing changes in ciliary activity that result in gravikinesis (Up3 and Dn3) and gravitaxis (Hx2). This fascinating adaptive behaviour was discussed more than a century ago (Verworn, 1889), but its mechanisms are still disputed (Fukui and Asai, 1985; Roberts, 1970; Taneda, 1987; Taneda et al. 1987; Winet and Jahn, 1974).