Physical and Physiological Components Of The Graviresponses Of Wild-Type and Mutant Paramecium Tetraurelia

Two types of graviresponse of swimming protists have been experimentally verified, gravitaxis and gravikinesis. The more recent literature primarily reflects the search for physiological pathways of either gravitaxis or gravikinesis (Machemer and Bräucker, 1992; Häder and Hemmersbach, 1997; Machemer, 1998; Hemmersbach et al., 1999). This is due, in part, to the limited and still developing body of methods and tools for the investigation of graviresponses in cells, requiring that an experimenter concentrates on a few, well-established procedures. A further reason for the current emphasis on physiology is that gravitational signalling promises to be among the most sensitive mechanisms of mechanotransduction known among eukaryotes.

The existing separation of research into gravitaxis or gravikinesis, physiological or physical mechanisms, is artificial because protist cells are likely to use gravitaxis in conjunction with gravikinesis, and an ecologically useful graviresponse may be generated by mechanisms of physical torque or via sensory transduction. We have previously developed and refined methods documenting the orientation- and gravity-dependent speed of locomotion, the effects of sedimentation and gravikinesis in various ciliates, and most recently in a flagellate (Machemer-Röhnisch et al., 1999). Attempts were made to improve the assessment of cellular gravitaxis (Machemer and Bräucker, 1992), and we have proposed a method of screening gravitaxis for a possible physical mechanism (Machemer and Bräucker, 1996). Such screening suggested that the search for mechanisms of gravitaxis should not be limited to mutual exclusion of the physical or physiological signalling pathways. The responses of a single species and between species differed in directions taken, degree and time course. Observations of the precision of gravitaxis indicate that mechanisms exist that can antagonize a taxis so that at no time do all the cells swim precisely in one direction. It was suggested that brief kinks (‘reversals’) in swimming direction, which are ubiquitous in protists, randomize orientation to antagonize perfect gravitaxis.

Here, we present results of experiments aiming at the analysis of both physical and physiological mechanisms of graviresponses in Paramecium tetraurelia giving special attention to the role of swimming reversals in modulating gravitaxis. We use, apart from the wild type, a mutant known for local and global defects of cortex morphology and thought to have an abnormal distribution of mechanoreceptor channels along the antero-posterior cell axis (Jerka-Dziadosz et al., 1992). The data form part of a dissertation by U.N. at the Faculty of Biology, Ruhr-Universität Bochum, Germany (Nagel, 1998).

Paramecium tetraurelia wild type (Sonneborn; stock d4-2) and the mutant kin 241 (line 12A 2a) were kindly provided by Dr J. Beisson (Centre de Génétique Moléculaire, Gif-Sur-Yvette, France). The mutation is monogenic and recessive. The mutant kin 241 is homozygous and thermosensitive above 35 °C (Jerka-Dziadosz et al., 1992). Culture medium contained 0.8 % wheat straw in double-distilled water (w/v). It was autoclaved and buffered at 22 °C at pH 7.0 using Sörensen buffer (1.8 mmol l⁻¹ Na₂HPO₄ plus 0.2 mmol l⁻¹ NaH₂PO₄). The medium was bacterized with Enterobacter aerogenes 24 h before inoculation with Paramecium. Cells were cultured at 22 °C in a 14 h:10 h L:D photoperiod and harvested after 3 days in the early stationary phase. Cells were washed in the experimental solution (1 mmol l⁻¹ CaCl₂, 1 mmol l⁻¹ KCl/KOH, 1 mmol l⁻¹ Mops buffered at pH 7.0; 22 °C) and collected using gravitactic accumulation.

Cells were equilibrated for 4 h in an experimental chamber made from Plexiglas with a glass cover for viewing; the chamber had a swimming volume of 35 mm×35 mm×2 mm and a field of view of 13.5 mm×17.5 mm (Nagel et al., 1997). Chambers were slowly rotated, at 1.5 ° s⁻¹ (or at 3 ° s⁻¹ in the centrifuge), from the horizontal to the vertical position. After completion of turning and the passage of a predetermined time interval (60 s; or 0 s in the centrifuge), cells were monitored using a video camera under dark-field illumination (48 green light-emitting diodes, 565 nm, 700 lx corresponding to 4W m⁻²). Abrupt changes in swimming direction (‘reversals’ or ‘turns’) were documented if the swimming track preceding a reorientation was pursued for a sufficient time (1 s) to allow determination of orientation. Only those reorientations that were more than 30 ° off the axis of the previous path were recorded. With these restrictions, results were used in only 50–65 % of reorientation events.

We used a computer-controlled low-velocity centrifuge microscope (NIZEMI; Dornier, Germany; Jena Optronics, Germany) at accelerations between 1 and 7 g (for experimental arrangement, see Bräucker et al., 1994). The sequence of levels of acceleration was randomized. The centre of the experimental chamber was positioned in the radial plane at a distance of 20.1 cm from the centre of the horizontal turntable. The cells were monitored as described above. A motor-driven swing device mounted on a turntable turned the major plane of the fluid volume from a perpendicular to a parallel position with respect to the resulting g vector. While the acceleration of the centrifuge increased or decreased (at 0.05 g s⁻¹), the chamber was held in the perpendicular position to reduce an undesirable loss of cells from the viewing field. Prior to recording, the chamber was turned to the parallel position (at 3° s⁻¹; Nagel et al., 1997; Nagel, 1998). The experimental chamber swang out as a result of eccentric mass with rising acceleration. All experiments were carried out at room temperature (21–23 °C).

Approximations of the gravity-free propulsion rate of Paramecium tetraurelia were obtained during free-fall experiments in a drop tower (ZARM, Bremen, Germany). An experimental arrangement including devices for recording was maintained at normal atmosphere inside a drop capsule, which was released to fall for 4.6 s over a vertical distance of 110 m in an evacuated steel tube (10 Pa, effective gravity, 10⁻⁴g; Machemer et al., 1993a; Nagel, 1998).

Video images were digitized and superimposed to provide tracks from which swimming rates and track orientations were measured. Cell orientations were determined from track orientations by accounting for the sedimentation rate (Machemer et al., 1997). The numbers of reversals in upward and downward orientation were counted for all measured cells and were normalized to a time interval of 1 min.

The precision of orientation was determined using circular statistics. The coefficient of orientation (r_o; Machemer and Bräucker, 1992) integrates the coefficient of directivity (R value; Batschelet, 1981) and the error angle (the angle between the cell and stimulus vector); it corresponds to the mean cosine of all individual direction vectors of unity scalar value and varying error angle (Machemer and Bräucker, 1996). The orientation coefficient differentiates between the upward orientation (positive sign) and downward orientation (negative sign) of cells. We calculated orientation coefficients of a population (r_oC) after transforming observed individual track orientations to cell orientations by accounting for errors in reading orientations introduced by sedimentation (Machemer et al., 1997).

Possible physical mechanisms of gravitaxis in P. tetraurelia were screened by plotting orientation coefficients from experimental data as a function of applied acceleration and testing for fit with a torque-dependent sigmoidal relationship between r_oC and g (Machemer and Bräucker, 1996). Torque in a swimming cell may result from two physical conditions: eccentricity of cell density (static buoyancy mechanism) and/or asymmetry in shape (hydrodynamic mechanism). The distance between the geometric centre point of a cell (half-way between the anterior and posterior ends) and the centre of gravity (or centre of shape) and acceleration determine the resulting torque in the gravitational field. We determined the centre of shape from the intersection of a transverse plane separating equal anterior and posterior surface areas of the cell with its longitudinal axis. In P. tetraurelia, this plane separates a wider and shorter posterior area from a narrower and longer anterior area; the centre of shape is behind the geometric centre (Nagel, 1998). Because larger bodies of the same density settle faster than smaller ones (Roberts, 1970), the posterior part of P. tetraurelia swings below the anterior part.

We employ nonparametric statistics (medians with 95 % confidence intervals) because Gaussian distributions were not assured in all cases. In the case of normal distributions, the medians coincide with the means. Tests of statistical significance of speed differences are based on the Wilcoxon–Mann–Whitney U-test (level of significance 5 %). Orientational distributions were determined using the Rayleigh test and were statistically compared using the Kuiper test (Batschelet, 1981). For regression analysis, we applied the method of least squares. Significance of the linear relationship was tested (level of significance 5 %). The correlation coefficient (− 1::sR::s+1) is a relative measure of the strength and direction of a linear relationship. The square of the correlation coefficient, R₂, gives the coefficient of determination. The value of R₂ (for instance, 0.95) indicates how well a regression curve is in agreement with given data (for instance, at 95 %).

Wild-type and mutant cells tended to be oriented in the upward direction (negative gravitaxis). Negative gravitaxis was significantly more pronounced in the mutant (r_oC=0.24) than in the wild type (r_oC=0.20; Fig. 1), and this applied to accelerations of up to 7 g (wild-type r_oC=0.58; mutant r_oC=0.62).

The more pronounced gravitaxis of the mutant compared with the wild type might be due, in part, to differences in eccentricity of the centre of shape from the geometric centre of the cell (Table 1). Asymmetries in density distribution of the cell body were not tested by shape relationships, but are possible.

Cell orientation coefficients, plotted as functions of acceleration, are in satisfactory agreement with the r_oC–g transfer function predicted from a mechanism of orientation by mechanical torque (Machemer and Bräucker, 1996; Fig. 2A). At all accelerations tested, cellular gravitaxis was most pronounced immediately after completion of turning the chamber from the horizontal to the vertical position (Fig. 2B). Mutants were more oriented than wild-type cells (Fig. 2B). More than 50 s after completion of the turn to the vertical position, orientation coefficients were reduced at all g values (Fig. 2C). The observed time-dependent depression of orientation suggests that an unknown mechanism interfered with orientation to reduce gravitaxis. According to the coefficients of determination of the regression analysis (R₂; see Materials and methods), the early (Fig. 2B) and late (Fig. 2C) orientation data are in agreement (;:,94 %) with the hypothesis that gravitaxis is effected by a physical mechanism of mechanical torque.

Behavioural tests of Paramecium tetraurelia showed that the frequency of reversal responses varied. The frequency of reversals declined during the first 5 min following infusion of cells into the chamber (Table 2). During the first 3 min, reversals occurred significantly more often in downward swimmers than in upward swimmers (Table 2).

At 1 g and after 4 h of equilibration, the frequency of reversals was no longer orientation-dependent (Fig. 3). In hypergravity, however, downward swimmers were more likely to change swimming direction than upward swimmers (Fig. 3). The frequency of reversals of upward-swimming wild-type cells showed no significant relationship with increasing g, but the number of reversals rose significantly in downward swimmers (Fig. 3A). The same pattern was observed in the mutants. Increases in g caused a significant increase in reversal frequency in upward and, to a lesser degree, in downward swimmers (Fig. 3B).

The properties of the membrane in the wild-type and mutant cells are shown in Table 3. The resting membrane potentials of the wild-type and mutant cells were identical (−32 mV), whereas the input resistance in the mutant (29.7 MΩ) was significantly smaller than that of the wild type (46 MΩ). Part of this difference is due to the larger soma membrane surface area of the mutant (115 %) compared with the wild type (Table 1). The variable that most affects input resistance, however, is the specific resistance of the soma membrane, which in the mutant is reduced to 74 % compared with that of the wild type (Table 3). Input conductances of the wild type (21.7 nS) and the mutant (33.7 nS) were measured in horizontally oriented cells under 1 g conditions. These conductances, the membrane potential and the equilibrium potentials of Ca²⁺and K⁺ in the experimental solution (1 mmol l⁻¹ Ca²⁺, 1 mmol l⁻¹ K⁺) were used to calculate conductance ratios (g_Ca/g_K; Table 4).

Modulation of swimming speed in the wild type and mutant by gravity was investigated by determining the speed V and angle of inclination θ of swimming tracks, the median sedimentation rate (S) and the gravity-independent propulsion rate (P) (equation 2). In wild-type P. tetraurelia, the value of P was taken as the speed of horizontally swimming cells (V_H) at 1 g because the swimming speed during the weightless condition in drop-tower experiments did not differ significantly from V_H (Nagel, 1998). In the mutant, P was found to be 89 % of V_H. Moreover, V_H of the mutant was less (73 %) than V_H of the wild type (Nagel, 1998). Active propulsion is composed of P and the gravikinetic change (±Δ) from the level of P at the orientation (o) taken by the cell (P±Δ_o). The change in active propulsion rate was found to be a continuous function of cell orientation for both wild-type and mutant cells (Fig. 4). The sensitivity of the gravikinetic responses in the wild type and kin 241 was the same (0.64 μm s⁻¹ degree⁻¹). However, the gravikinetic response of the mutant was shifted in the positive direction by approximately 65 μm s⁻¹.

Despite identical sensitivity, only wild-type cells were able to generate a largely symmetrical gravikinetic response during downward and upward swimming (maxima near 50 μm s⁻¹, see Fig. 4). The mutant kin 241 showed no gravikinesis at 180 ° (oriented vertical downwards) and gave an extreme response at 0 ° (oriented vertically upwards; kinesis approximately 90 μm s⁻¹, see Fig. 4). The unilateral shift in gravikinesis of the mutant compared with the wild-type is equivalent to a hyperpolarization-dependent augmentation of the ciliary motor response.

Assessment of gravitational pressure

Approximating the shape of the cell by a rotational ellipsoid, the dimensions of Paramecium tetraurelia give a volume of 6.9×10⁻⁸ cm³ for the wild type and 8.6×10⁻⁸ cm³ for the mutant kin 241 (Table 1). With a cell density of 1.04 g cm⁻³ (Machemer and Bräucker, 1992), the apparent masses (cell mass minus mass of displaced body of water) are 2.8×10⁻⁹ g and 3.4×10⁻⁹ g, respectively. At normal acceleration (9.81 m s⁻²), the gross gravitational force can be calculated (wild type, 2.7×10⁻¹¹ N; mutant, 3.4×10⁻¹¹ N; Table 1). This force may be linked to gravisensitive receptors in the soma membrane as follows.

Two extreme assumptions regarding the physical properties of the cytoplasm approximate the local pressure: (i) if the body is an ideal fluid, the hydrostatic pressure on the surrounding membrane results from the height of the column of cytoplasm per unit area; (ii) if the cytoplasm including the cytoskeleton is a rigid body, static pressure on the lowest point of the membrane results from the total gravitational force acting on the horizontal cross section (shadow) of the cell (Fig. 5). The cytoplasm has viscoelastic properties (Todd, 1994; Ehrenstein and Iwasa, 1996). Assumptions i and ii above, therefore, mark the upper and lower extremes of a description of gravitational pressure at the membrane excluding additional cellular mechanisms of gravitational focusing. Fig. 5 illustrates, for three orientations, the lowest points on the membrane (arrow), the maximal pressure, as derived from the hydrostatic approach, and the minimal pressure, as derived from the static approach. The range of pressures applicable in Paramecium tetraurelia is therefore between 9 and 45 mPa (0.009–0.045 N m⁻²).

Inclusion of the values of maximal pressures for wild type and kin 241 in Fig. 4 shows that pressure alone does not explain the divergent motor responses of downward and upward swimmers, nor do differences in pressure between wild type and kin 241 explain the constant offset in the motor responses of these cells.

The speed of wild-type upward swimmers (V_U) decreased, and the speed of downward swimmers (V_D) increased with increasing gravitational acceleration (Fig. 6B). Linear regressions of these relationships intersected at 0 g at a common speed, 829 μm s⁻¹ (Fig. 6B). As for the swimming rate, the median sedimentation rates (S) of immobilized cells (wild-type and mutant) were a linear function of acceleration intersecting at the origin (Fig. 6A; wild type 67.4 μm s⁻¹; mutant 73.5 μm s⁻¹). To account for sedimentation during swimming in wild-type cells, the slope of S was added to the slope of V_U, and was subtracted from the slope of V_D (Fig. 6C). The resulting rates of upward propulsion increased and rates of downward propulsion decreased with increasing acceleration. The slopes of these functions give values of orientation-dependent gravikinesis, and the intersection of the slopes gives the rate of gravity-free propulsion (P). Referring gravikinesis to sedimentation as a vector (positive sign), negative gravikinesis of downward swimmers (−54 μm s⁻¹g⁻¹) significantly exceeded negative gravikinesis of upward swimmers (−25 μm s⁻¹g⁻¹; Fig. 6D). The viability of the mutant kin 241 suffered during the phasic run-up and run-down accelerations of the centrifuge so that, for these cells, data were obtained from sedimentation experiments only.

Our investigation of gravitaxis in Paramecium tetraurelia confirmed previous findings suggesting that hypergravity potentiates gravitaxis in ciliates (Bräucker et al., 1994) and Euglena gracilis (Häder et al., 1995) (Fig. 1). Moreover, gravitaxis to some degree relaxed over time (Fig. 2), a phenomenon commonly called adaptation. Adaptation of gravitaxis in protists is controversial in the literature. There are both reports of adaptation in P. caudatum and P. biaurelia (Bräucker et al., 1998; Hemmersbach-Krause and Häder, 1990; Machemer et al., 1993b) and claims that adaptation to gravity does not exist, at least under hypergravity (Häder et al., 1995; Hemmersbach et al., 1996; Machemer-Röhnisch et al., 1993). A stable acceleration/gravitaxis relationship (Figs 1, 2) implies a baseline reversal rate. In the absence of reversals, even a weak acceleration would perfectly orient all cells; this was not observed in the present study. The observed transition from an initially high degree of orientation (following completion of turning of the experimental chamber from horizontal to vertical at 1 g; Fig. 2B) to a reduced degree of orientation (Fig. 2C) confirms findings in P. caudatum (Bräucker et al., 1998). This phenomenon might be based on a time-dependent change in rates of reversals in Paramecium spp. as previously documented, for example, for adaptation following a K⁺-induced shift in the potential of the cell membrane (Machemer, 1989).

We believe that an interaction between gravitactic orientation and orientational randomization by reversals (Machemer and Bräucker, 1996) can explain the time-dependence of gravitaxis in the follwing ways.

(1) Reversal frequency depends on orientation

We have documented that reversals in downward swimmers occurred significantly more often than reversals in upward swimmers during the first few minutes after incubation (Table 2). This non-stochastic distribution of reversals will necessarily favour an improvement in upward orientation. The differences in the reversal frequency between upward and downward swimmers declined with time (Table 2). This decline logically explains the observed time-dependent relaxation of gravitaxis (the shift from initially large to lower k values; Fig. 2C). The observed orientational preference of reversals generates a bias favouring negative gravitaxis. Jennings (1906) was the first to note that reversals were rare in upward-swimming Paramecium spp. and more frequent during downward swimming.

(2) Reversal frequency links gravitaxis to membrane potential

The orientation and gravity-dependence of reversals (Fig. 3) suggests a coupling between cellular reversal rates and the membrane potential. A pronounced sensitivity of downward swimmers for reversals was postulated by Naitoh and Eckert (1974) because depolarizing mechanoreceptor channels accumulate in the anterior membrane of Paramecium caudatum (Naitoh and Eckert, 1969). In agreement with this, increasing acceleration augmented reversal rates in downward swimmers but not in upward swimmers (Fig. 3A). Moreover, the downward swimming speed was reduced compared with the gravity-free propulsion P (Fig. 4), in agreement with depolarization-dependent gravikinesis (Gebauer et al., 1999). The mutant kin 241 were more precisely oriented than the wild-type cells (Fig. 2), mutant reversals were more rare (74 % of the value for wild-type cells, data not shown), and mutant cells were more hyperpolarized under gravity (Table 4; see change in V_m from microgravity to 1 g). These observations agree with the view that the membrane potential can modulate gravitaxis by means of reversal frequency.

(3) Adapting reversal frequency modulates non-adapting gravitaxis

The established electrophysiological basis for reversal responses in Paramecium spp. necessarily links the observed time-dependence of gravitaxis (Fig. 2) with mechanisms of physiological adaptation in P. caudatum (Oka et al., 1986). The apparent contradiction between the documented adaptation of gravitaxis and the postulated non-adapting graviresponses is thereby resolved.

The observed rates of swimming and sedimentation were linearly related to the acceleration applied. These relationships allowed the gravikinetic components during upward and downward swimming to be determined using linear regressions (Fig. 6). The resulting values of gravikinesis (equations 3, 4; medians from ±45 ° sectors; Δ_U=−25 μm s⁻¹; Δ_D=−54 μm s⁻¹) largely agree with values obtained in noncentrifuged cells (equation 2; Fig. 4, medians from 10 ° sectors: Δ_U=−48 μm s⁻¹, Δ_D=−36 μm s⁻¹). Calculating the data of Fig. 4 for ±45 ° sectors, gives Δ_U=−34 μm s⁻¹, Δ_D=−44 μm s⁻¹. A 2 % error (i.e. of 9.5 μm s⁻¹) in assessment of the value of P under microgravity conditions would explain the remaining discrepancy between the hypergravity results (Fig. 6) and the 1 g results (Fig. 4). The results obtained using several levels of acceleration are likely to be superior to those derived from exclusively 1 g conditions because regression analysis reduces the effects of uncertainties regarding single data points.

The gravikinesis of wild-type P. tetraurelia neutralizes, in part, the effects of sedimentation by increasing the upward and reducing the downward speed of swimming (Fig. 4). This type of gravikinesis is common in ciliates and has been demonstrated previously in P. caudatum (Bräucker et al., 1998; Machemer-Röhnisch et al., 1996; Watzke et al., 1998), P. biaurelia (Hemmersbach-Krause et al., 1993), Tetrahymena pyriformis (Kowalewski et al., 1999), Loxodes striatus (Machemer-Röhnisch et al., 1993) and Bursaria truncatella (Krause, 1999). The mutant kin 241 of P. tetraurelia had the same orientational sensitivity as the wild-type. However, gravikinesis of upward swimmers was augmented, and gravikinesis of downward swimmers acted in the ‘wrong’ direction (Fig. 4). Thus, the well-balanced and quite efficient gravikinetic response of the wild type has apparently turned, in part, to a nonsense response in the mutant.

Among wild-type protists so far investigated, two species show an interesting correspondence with the monopolar gravikinetic response of the P. tetraurelia mutant. (i) Didinium nasutum can actively reduce forward swimming speed under gravitational stimulation (Bräucker et al., 1994) by using depolarizing mechanoreceptor channels in more anterior parts of the cell membrane. D. nasutum cannot augment ciliary propulsion (Pernberg and Machemer, 1995). Gravikinesis in D. nasutum thereby neutralizes sedimentation only during downward orientations, whereas P. tetraurelia kin 241 neutralizes sedimentation only during upward orientations. (ii) Euglena gracilis shows an apparently even more primitive gravikinetic response, augmenting swimming rates at all orientations in the gravity field, albeit more effectively while swimming upwards (Machemer-Röhnisch et al., 1999). A comparison of various types of gravikinesis suggests that Paramecium spp., by fine-tuning opposing gradients of mechanoreceptor channels, generates a useful kinetic response to the gravity vector. This balance appears to be lost in the mutant kin 241 of P. tetraurelia.

The linear gravikinetic curve of P. tetraurelia is at variance with the orientation-dependent gravikinetic curve of P. caudatum, which follows a sigmoidal course, where horizontally swimming cells occur in the centre of maximal orientational sensitivity (Gebauer et al., 1999). By comparison with P. caudatum, the curve for wild-type P. tetraurelia (Fig. 4) does not include orientations of pronounced sensitivity.

The assumption of a maximal hydrostatic pressure gradient of 13 mPa across the lower membrane of a horizontal cell, assuming ideal fluidity of the cell interior (Fig. 5), ignores the comparatively high viscosity of the cytoplasm and excludes a role of the cytoskeleton, via hydrodynamic coupling, as a potential agent of gravitational focusing. The minimal estimate of pressure, 9 mPa for a horizontal cell, is equally unrealistic because the cytoplasm is not a solid body, and gravitational focusing by such a body is also excluded. These hypothetical extreme values therefore simply delimit a range within which the value of gravitational pressure must lie.

We have included in Fig. 4 values of maximal pressure for the wild type and for the larger mutant (see Table 1) as functions of orientation. From Fig. 4, pressure is identical for cells at 0 ° and 180 ° orientation, and for cells at 45 ° and 135 ° orientation, etc. Gravikinetic responses, however, do not distribute symmetrically in this way but follow antero-posterior gradients. The comparison suggests that gravikinesis is determined by gravitational pressure in conjunction with cell orientation. A maximal transmembrane pressure difference of 13.3 mPa for a horizontal cell was sufficient to elicit a gravikinetic response (Fig. 4). Similar results have been obtained from smaller cells such as Tetrahymena pyriformis (9 mPa; Kowalewski et al., 1999) and Euglena gracilis (5 mPa; Machemer-Röhnisch et al., 1999). With channel gating distance in most sensitive mechanoreceptors presumably being restricted to no more than 3.5 nm (Howard et al., 1988), the primary energy derived from normal acceleration is very small; for a horizontal P. tetraurelia, it is 9×10⁻²⁰ N m (only a maximum of 45 times above the thermal noise level at room temperature).

Paramecium caudatum and Didinium nasutum transiently to maintain gravikinesis which, however, changed sign to act co-parallel with the gravity vector (Bräucker et al., 1998; Machemer-Röhnisch et al., 1998). This ‘gravikinetic paradox’ was interpreted as a viscoelastic transfer of gravity-dependent deformation of the lower membrane to deformation of the upper membrane, implicating a mechanically charged cytoskeleton. Such transfer of local deformation representing an early step in the relaxation from a gravity-induced deformation of the cell is, to our understanding, difficult to reconcile with the idea of the global transfer of cytoskeletal tension.

Inspection of basic electrophysiological data obtained under horizontal 1 g conditions (Table 3) revealed the same resting potential in the wild-type and mutant (−32 mV) cells, but a significantly lower specific resistance of the soma membrane in the mutant (4.55 kΩ cm²) than in the wild-type (6.15 kΩ cm²) cells. This change in resistance suggests a major increase in resting conductance, in particular of the K⁺ channels (Table 4). Fig. 7 illustrates how the resting and gravity-induced conductances of Ca²⁺ and K⁺ channels might be altered in the mutant compared with the wild-type cells. Using orthodox electrophysiological calculations, we suggest that, in the mutant, the mechanosensitive system of two

Deformation of the tension-dependent architecture of the cell by gravity has been suggested as a primary step in gravisensation (Ingber, 1999). According to this author, transmembrane mechanoreceptors could be activated by gravity via the cytoskeleton. In agreement with this general view, an observed potentiation of gravikinesis in Paramecium caudatum exposed to a weak direct current field was explained by postulating Ca²⁺-sensitive filaments linked to the gates of mechanoreceptor channels (Machemer-Röhnisch et al., 1996). Ingber’s (1999) suggested gravisensation relies on the presence of a three-dimensional link between the skeletal elements of the cytoplasm and the membrane, which can store mechanical tension and thereby spread a local cellular deformation (Wang et al., 1993). Direct application of this view of ‘tensegrity’ to gravitransduction in free-swimming protists is a problem because global deformations of the cell (elongations or rounding up) would necessarily cause events of global gravitransduction. Our data suggest, however, that gravity first activates channels of the lower membrane. These channels are part of antero-posterior gradients of mechanosensitivity that are an indispensable basis for active compensation of sedimentation of the cell irrespective of its orientation.

Experimental step transitions from normal acceleration to the weightless condition induced vertically oriented opposing and overlapping gradients of Ca²⁺ and K⁺ channels present in the wild type is shifted in favour of hyperpolarizing K⁺ channels (Table 4). An overall increase in the number of gravisensitive open K⁺ channels and/or reduction in the number of gravisensitive open Ca²⁺ channels could explain the shift in the gravikinetic curve (Fig. 4).

Gravisensory transduction in Paramecium tetraurelia profoundly interferes with the regulation of ciliary propulsion and with normal swimming behaviour; forward propulsion at any orientation of the cell is modulated in such a way as to counteract the passive settling effect of gravity. We show here for the first time that gravitransduction interferes indirectly with gravitaxis at 1 g and hypergravity in that reversals of the swimming direction occur more often in downward swimmers than in upward swimmers. Since reversals randomize orientation, upward orientation is favoured over downward orientation. Negative gravitaxis as a phenomenon of gradual upward orientation is shown to be compatible with static or hydrodynamic mechanisms of torque. The mutant kin 241 of Paramecium tetraurelia has abandoned, in part, a behaviour that neutralizes the settling effect of gravity. This is presumably due to disturbance of the fine-tuned topological balance of depolarizing and hyperpolarizing mechanoreceptors of the wild type, which is believed to be the basis of orientation-dependent up- and down-regulation of swimming speed in ciliates.

Financial support for this study was provided by the Deutsches Zentrum für Luft-und Raumfahrt (DLR), grant 50WB93193, and the Minister für Wissenschaft und Forschung of the state of Nordrhein-Westfalen, grant IV A1-21600588. We thank Professor H. J. Rath and his collaborators at ZARM, Bremen, for a most fruitful cooperation during the drop-tower experiments.