Control of Ciliary Activities by Adenosinetriphosphate and Divalent Cations in Triton-Extracted Models of Paramecium Caudatum

The locomotor behaviour of Paramecium and many other protozoans depends on the movements performed by cilia distributed over the cell surface (Jennings, 1906). Recent evidence indicates that the movements of the cilia are regulated by electric events in the surface membrane. A depolarization of the membrane, which occurs in response to an injection of outward current (Naitoh, 1958; Naitoh & Eckert, 1968; Eckert & Naitoh, 1970; Machemer & Eckert, 1973), or to mechanical stimulation of the front end of the specimen (Naitoh & Eckert, 1969a, b) or spontaneously (Kinosita, 1954; Naitoh, 1966), is always followed by a transient reversal in the direction of the effective power stroke of cilia (i.e. ciliary reversal). Reversal of the ciliary beat causes the animal to swim backward. On the other hand, a hyperpolarization induced by an injection of inward current (Naitoh, 1958; Naitoh & Eckert, 1968) or by mechanical stimulation of the rear end of the animal (Naitoh & Eckert, 1969a, b; 1973) is followed by an inhibition of ciliary reversal and an increase in the beat frequency of the cilia in the normal direction. This causes the animal to swim forward with increased velocity.

Depolarization of the surface membrane by injected current or mechanical stimulation of the cell anterior induces a voltage-dependent increase in calcium conductance, which permits Ca²⁺ to flow inward down its electrochemical gradient. This results in a regenerative depolarizing action potential (Naitoh, Eckert & Friedman, 1972; Eckert, Naitoh & Friedman, 1972).

Ciliary reversal also occurs when the external cationic conditions are changed so as to liberate bound calcium from the membrane (Jahn, 1962; Naitoh, 1968). The liberated calcium is believed to activate a reversal mechanism in the cilia (Naitoh, 1969).

In order to understand how modifications of ciliary movement are coupled to electric and/or ion exchange events of the surface membrane, it is important to know the effects on the ciliary apparatus of the ions involved in the membrane events. We therefore examined the effects of various cations on the ATP-reactivated cilia of detergent-extracted models of Paramecium. In such models the ciliary membrane is functionally disrupted so that the externally applied cations have direct access to the ciliary apparatus.

Our preliminary report (Naitoh & Kaneko, 1972) showed that external application of calcium ions led to a reversal of the orientation of effective power stroke in ATP-Mg²⁺-reactivated cilia. Thus, calcium ions cause the models to swim backward as do live specimens upon receiving a depolarizing stimulus.

Specimens of P. caudatum (mating type I of syngen 1) reared in a hay infusion were washed thoroughly with 2 mm-CaCl₂ solution buffered to pH 7·2 by 1 HIM Tris-HCl buffer. A concentrated suspension of the specimens was cooled in an ice bath and centrifuged gently to make a loose pellet. The pellet was then resuspended in a cold (0–1 °C) extraction medium which consisted of 0·01 % (by volume) octylphenoxy polyethoxyethanol (Triton X-100 obtained from Wako Pure Chemicals Co., Tokyo), 20 mm-KCl, 10 HIM ethylen-diaminetetraacetic acid (EDTA : neutralized by KOH) and 10 DIM Tris (hydroxymethyl) aminomethane (Tris)-maleate buffer (pH was adjusted.to 7·0 with NaOH). The suspension was then stored in an ice bath (0–1 °C) for 35–45 min. The extracted specimens thus obtained were washed gently three times with a cold (0–1 °C) washing medium which consisted of 50 mm-KCl, 2 mm EDTA, and 10 mm tris-maleate buffer (pH 7·0), and kept in this medium for 15 min to remove the Triton. They were then washed again with a cold (0-1 °C), buffered (pH 7·0 by 10 mm Tris-maleate) 50 mm-KCl solution to remove the EDTA. The washed models were equilibrated in the KC1 solution for at least 15 min at 0–1 °C before experimentation.

About 500 cell models with a minute amount of KC1 solution (about 10⁻⁴ ml) were pipetted into a large amount of a reactivation medium (about 1 ml) in a depression slide and stirred gently by a blunt glass needle at a room temperature of 19–21 °C. The reactivation media consisted of 50 mm-KCl containing test substances (pH was adjusted to 7·0 by 10 mm Tris-maleate buffer). After an appropriate exposure in the reactivation solution the models were observed and photographed.

About 200 reactivated models were introduced into a thin (70 μm) space between a glass slide and a coverslip. This spacing was sufficient for the models to swim freely in parallel with the glass plates but too shallow for locomotion in the vertical direction (median size of the models was 250 μm in length and 50 μm in width). The models were then photographed in this space with five consecutive xenon flashes (1 flash/sec) with dark-field illumination. The first image of the sequence was brighter than the others because the first flash was more intense, and so the direction of swimming is readily evident (Text-fig. 1). Swimming velocity was calculated from the distance between each photographic image of a swimming model. Slowly swimming models produced an overlapping bead-like image. A bright single image indicates that the model did not swim at all.

The frequency of metachronal waves passing a point on the model was monitored photometrically. Several models in a reactivation medium were pipetted on to a glass slide and compressed gently by a coverslip to prevent free swimming. The image of metachronal waves (×200), was focused on a screen through a phase-contrast objective (Nikon bright contrast ; × 40), and a small part of the image went on to a photomultiplier (Toshiba MS-9S, anode voltage, 1000 V) through a small hole (0·5 mm in diameter) in the screen. Cyclic changes in the light intensity due to passage of metachronal waves were displayed by this means on a cathode-ray tube and photographed (Text-fig. 2).

Long (250 msec) electric current pulses in both inward and outward directions were introduced through a glass microcapillary electrode (less than 1 μm in tip diameter; filled with 0·1 M-KCI; 100–200 MΩ; see Naitoh & Eckert, 1972) into a live nonextracted specimen in a dilute saline solution. Resulting intracellular potential changes were monitored through a recording electrode of the same type and displayed on a cathode-ray tube. The external solution was then replaced with cold (1–2 °C) extraction medium. Electrical characteristics, such as resting membrane potential and responses to the current pulses, were monitored during the extraction.

Changes in the amounts of adenosinetriphosphate (ATP), adenosinediphosphate (ADP) and adenosinemonophosphate (AMP) in the specimens were enzymically determined during the course of extraction. About 0 ·2 ml of loosely packed models were homogenized and extracted in 4% perchloric acid for 30 min and then centrifuged in the cold. The neutralized supernatant was subjected to the enzymic determinations of adenine nucleotides by the method described by Bergmeyer (1963).

When specimens of Paramecium were immersed into the extraction medium they swam forward without showing any avoiding reactions, either spontaneously or upon collision with an obstacle, which are characteristic of non-extracted live specimens in their culture medium or in dilute saline solutions. The swimming velocity gradually decreased, and the specimens stopped swimming after 20 min due to a slowing of ciliary activity. Complete cessation of ciliary beating occurred after 30–40 min.

Non-extracted live specimens in a saline solution (4 mm-KCl +1 mm-CaCl₂, buffered to 7·2 by 1 HIM Tris-HCl) showed graded regenerative calcium spikes (Naitoh et al. 1972) followed by damped potential oscillations in response to long (250 msec) depolarizing current pulses (10⁻⁹ A) (Text-fig. 3 A). The surface membrane behaved as an ohmic resistance (about 1·5 ×10⁷Ω) in parallel with a capacitance (about 0·001 μF) to a hyperpolarizing pulse (Text-fig. 3B). A delayed anomalous rectification was observed in response to stronger hyperpolarization.

Upon replacement of the saline with the extraction medium a sudden depolarization and a transient (several seconds) sign reversal of the membrane potential (about + 20 mV) occurred, then the potential gradually approached the reference (zero) level. The membrane thereafter never showed any kind of electrical activity.

An injection of a square pulse did not induce a measurable potential shift across the membrane (Text-fig. 3 C, D), indicating a great increase in the membrane conductance.

The concentration of ATP was 0·87 mm in non-extracted live specimens (a mean of three measurements on different groups of the specimens ; approximate volume of specimens was 5·6 ×10⁻⁷ ml; see Fortner, 1925). As shown in Text-fig. 4, the ATP concentration decreased gradually with extraction time and became less than one-fifth the original value in 30 min. It continued to decrease during washing and storage of the models in the KC1 solution.

The amounts of ADP and AMP were very small (less than 0·05 mm) and became undetectable after 20 min of extraction.

A few seconds after the models were transferred into a reactivation medium, which contained ATP, Mg²⁺ and EGTA (see later section), the cilia began to beat, first slowly and soon quickly, exhibiting antiplectoidal (Machemer, 1972) metachronal waves (Pl. 1 B). Metachronal waves could be seen even in cilia which beat very slowly (less than 1 beat/sec) at the beginning of their reactivation. Whereas some models, especially those extracted for long periods or stored for a long time before reactivation, showed lashing-type beatings with the frequency higher than that in the good re-activated models but without metachronal co-ordination. The effective stroke was directed toward the rear and somewhat to the right, causing the models to swim forward as the live specimens do under non-stimulated conditions (Text-fig. 1A, B).

The beating frequency of the cilia reached its maximum about 2 min after the transfer, then decreased slowly. The cilia finally stopped beating in 40–60 min (Text-fig. 5). Frequent renewals of the reactivation medium did not prolong the time of the ciliary reactivation. All the measurements of beating frequency and swimming velocity, therefore, were made two minutes after the transfer.

In an optimum reactivation medium (4 mm ATP, 4 mm-MgCl₂, 3 mm EGTA) more than 95% of the models were reactivated and swam immediately after 15 min storage in the cold KC1 solution. However, only 50 % of the models were reactivated after 3 h of storage in the KC1 solution.

The beating frequency, as well as the swimming velocity, in a reactivation medium differed among specimens obtained from different cultures. Therefore, a series of experiments was carried out on fresh models (within 1 h after the extraction) made at the same time from one culture.

Sometimes, especially in a medium with low ATP and/or Mg²⁺ concentrations, only cilia on a limited area of the model were reactivated. This caused a curved or circular swimming of the model (Text-fig. 1 A). However, there was no tendency for cilia over a definite region to exhibit greater sensitivity to ATP and Mg²⁺ than the rest. The local reactivation occurred indiscriminately on any portion of the model in a diluted ATP-Mg²⁺ medium.

The beating frequency was found to be almost identical in all the cilia over the surface of one model when they were well reactivated in an optimum medium (Table 1), although local differences in beating frequency are normally observed in live specimens.

The beating frequency and swimming velocity of the models were determined in nine ATP-Mg²⁺ media (4 mm ATP, 4 mm-MgCl₂, 3 mm EGTA) with the pH controlled by 10 mm Tris-maleate-NaOH buffer between pH 5 and 9.

As shown in Text-fig. 6 (beating frequency) and Text-fig. 7 (swimming velocity), maximum values were obtained in a pH range from 6·5 to 7·0. For that reason all the following experiments were carried out at pH 7·0.

Beating frequency and swimming velocity were determined in a series of reactivation media with different ATP concentrations (0·25–8·0 HIM), while other ionic compositions were kept constant (4 mm-MgCl₂, 3 mm EGTA). Increase of frequency was sigmoidal when plotted against a logarithmic scale of ATP concentration between 0·25 and 8·0 mm (Text-fig. 8). The response was most sensitive in the concentration range of 1–4 mm. With further increase in ATP concentration the frequency tended toward a plateau value. Swimming velocity showed the same pattern of change in response to increased ATP concentration (Text-fig. 9).

Guanosine triphosphate (GTP), uridine triphosphate (UTP), cytocine triphosphate (CTP), ADP and AMP were tested to determine whether these can reactivate the models in the presence of Mg²⁺.

Only ADP was found to induce ciliary beating. The time course of the reactivation by 8 mm ADP was similar to that by 4 mm ATP as shown in Text-fig. 10, but the maximum swimming velocity in the ADP solution was significantly smaller than that in an ATP solution of the same concentration (8 mm).

Beating frequency and swimming velocity were determined in a series of media with Mg²⁺ concentrations ranging from 0·25 to 32 mm at a constant ATP concentration (4 mm).

Both beating frequency (Text-fig. 11) and swimming velocity (Text-fig. 12) increased with increasing Mg²⁺ concentration, reached a maximum at 4–8 mm, then decreased with further increase in the concentration.

Addition of Ca²⁺ (more than 10⁻⁵ M; without either Mg²⁺ or ATP) always induced longitudinal shortening of the model (i.e. contraction; see Weis-Fogh & Amos, 1972) (Pl. 1 C) and discharge of trichocysts; then the cilia came off the body and finally the models disintegrated.

When ATP was applied in association with Ca²⁺, non-beating cilia, which pointed toward the rear in the KC1 solution (Pl. 1 A), swung forward so as to point toward front (Pl. 1 D). Ciliary beating usually was not observed in the Mg²⁺-free ATP-Ca²⁺ medium. However, cilia sometimes beat for a short period (10–20 sec) after they became detached from the body of the model.

The most striking effect of Ca²⁺ on the ciliary orientation occurred when Ca²⁺ was applied to the reactivated, forward-swimming models in an ATP-Mg²⁺ medium. The orientation of the effective power stroke was reversed when the calcium concentration exceeded 10⁻⁶ M, causing the models to swim backward like backwardswimming live specimens in response to an appropriate stimulus (Text-figs. iC, D; Pl. IE; see also fig. 1 in Naitoh & Kaneko, 1972).

An increase in Ca²⁺ concentration up to 5×10⁻⁴ M did not affect the beating frequency of ATP-Mg²⁺-reactivated cilia. Ciliary beating, however, slowed with further increase in Ca²⁺ concentration and stopped at 10⁻³M (see fig. 2 in Naitoh & Kaneko, 1972).

Symplectoidal metachronal waves were observed in the Ca²⁺-induced backwardswimming models (Pl. 1E).

Effects of Ba²⁺ and Sr²⁺ on the models were essentially the same with those of Ca²⁺. They induced backward swimming in the ATP-Mg²⁺-reactivated models. Velocity of the backward swimming was somewhat slower than that produced by Ca²⁺ (66·8 ± 6·8% with Sr²⁺, mean of 222 models; 89-216-5% with Ba²⁺, mean of 325 models; 100 ± 5·7% with Ca²⁺ and 35·2 ± 2·0μm/sec, mean of 421 models).

The models were reactivated in a medium with ATP (4 mm) and Mn²⁺ (4 mm) with the swimming velocity of less than 45 % of that in ATP-Mg²⁺ of the same concentration.

Co²⁺ (4 mm) was also effective in reactivating cilia of the models in the presence of ATP (4 mm), but it was less effective than Mn²⁺, so that most of the reactivated models did not swim freely.

In a medium with ATP (4 mm) and Fe²⁺ (4 mm) only delayed (1-2 min after the treatment) weak reactivation of ciliary beating was observed. Beating frequency was less than 2 beats/sec.

When a small amount of Ni²⁺ ions (0–5 mm) was applied to the ATP-Mg²⁺ reactivated models, cilia ceased to beat within 1 min. The pointing direction of the stopped cilia was toward the rear. Addition of Ca²⁺ (more than 10⁻⁶ M) in the Ni²⁺ containing ATP-Mg²⁺ medium caused the stopped cilia to point toward the front. Ni²⁺, therefore, inhibits only Mg²⁺-dependent beating but not Ca²⁺-dependent change in the orientation of cilia.

Effects of several metabolic inhibitors, such as NaF (5 mm), ouabain (1 HIM), salyrgan (0·2 HIM), and NaCN (5 mm), on the ATP-Mg²⁺-reactivated ciliary beating were examined.

As shown in Text-fig. 13, salyrgan, which is a potent SH-blocking reagent, strongly depressed ciliary activity. Washing the salyrgan-inhibited models with 5 mm cystein caused a small recovery of the ciliary activity. Fluoride and ouabain reduced ciliary activity a little, but the models continued to swim for as long a time as the models in normal reactivation medium. Cyanide showed no effect.

Electron microscopy showed that the membrane remained at least partially visible around each cilium and covering the cell surface of the present models (Y. Naitoh, 1972 unpublished). However, loss of the electrical potential across the membrane and large increase in the electrical conductivity (Text-fig. 3) indicate that the Tritonextraction rendered the membrane very permeable to ions. The large decrease in cellular adenine nucleotides during the extraction (Text-fig. 4) indicates that the extracted membrane is also permeable to these substances. Therefore, externally applied inorganic ions and ATP (or other nucleotides phosphates) have direct access to the cell interior and can influence the function of the ciliary apparatus without much restriction by the surface membrane.

In the present models the reactivation of ciliary beating by ATP required Mg²⁺ or Mn²⁺ ions. The other divalent cations tested (Ca²+, Co²⁺, Fe²⁺ and Ni²⁺) were far less effective than Mg²⁺ and Mn²⁺. The swimming velocity in Mg²⁺ was double that in Mn²⁺. A high degree of specificity of Mg²⁺ for ciliary reactivation has been reported in the extracted models of other ciliary systems (spermatozoa: Bishop, 1962; Gibbons & Gibbons, 1972. Protozoan flagella: Brokaw, 1961, 1963. Protozoan cilia: Seravin, 1961 ; Gibbons, 1965. Ciliated epithelium: Child & Tamm, 1963 ; Satir & Child, 1963 ; Eckert & Murakami, 1972. For detailed references, see Arronet, 1971).

Recently Summers & Gibbons (1971) demonstrated that Mn²⁺ is as effective as Mg²⁺ in producing disintegration of isolated trypsin-digested axonemes from seaurchin spermatozoa in the presence of ATP, caused by active sliding between the peripheral tubules of the axonemes. The sliding is thought to be the underlying mechanism of ciliary bending in intact cilia. Reactivation of ciliary beating by ATP-Mg²⁺ in the presence of EGTA indicates that Ca²⁺ ions are not primarily necessary for the cyclic bending mechanism in the cilia (Gibbons, 1965).

High ATP-specificity for the ciliary reactivation found in the present models has also been reported in other ciliary models (Brokaw, 1961 ; Gibbons, 1965; Gibbons & Gibbons, 1972). The fact that ADP is effective for ciliary reactivation suggests that an adenylate kinase system is present in the models (Brokaw, 1961 ; Gibbons, 1965; Winicur, 1967; Naitoh, 1969). Some investigators (Gibbons, 1965; Winicur, 1967) reported a delay in reactivation of ciliary beating by ADP. However, in the present models the time course of reactivation by ADP (8 mm) was essentially similar to that by ATP (4 mm) (Text-fig. 10).

Optimum pH for the ATP-Mg²⁺ reactivation was 6·5–7·0 (Text-figs. 6, 7); the value is similar to that in isolated glycerinated cilia of Tetrahymena (Gibbons, 1965), and somewhat lower than that in Triton-extracted sperm models (Gibbons & Gibbons, 1972).

Similarity between beating frequency and swimming velocity curves as a function of concentration of H⁺, ATP or Mg²⁺ (cf. Text-figs. 6 and 7, 8 and 9, 11 and 12) indicates that the swimming velocity of the reactivated models is primarily dependent on the beating frequency of their cilia.

The beating frequency in an optimum reactivation medium (4 mm ATP, 4 mm-MgCl₂ and 3 mm EGTA) was 12·3 ± 0·6 beats/sec (mean and standard errors for 50 models), approaching the value for the unextracted live specimen (Sleigh, 1962 ; Kinosita, Dryl & Naitoh, 1964b). The mean swimming velocity (240 models) was 162 ± 12 μm/sec; this was less than half that of the normal live specimen (Kinosita et al. 1964b). The lower swimming velocity might be due to somewhat diminished co-ordination in the reactivated cilia.

The reactivated cilia usually showed metachronal co-ordination. This finding indicates that metachronal co-ordination is basically independent of the membrane function. It depends most probably on mechanical interactions between the cilia (Sleigh, 1962; Kinosita & Murakami, 1967). The idea that metachronal waves are propagated by impulses passing along the cell membrane (Grebecki, 1965) is no longer credible (see Eckert & Naitoh, 1970). But this should not be confused with the fact that bioelectric events in the membrane regulate the frequency and the orientation of ciliary beating (Kinosita, 1954; Kinosita et al. 1964a, b, 1965), thereby secondarily affecting the metachronal wave pattern (see the next section).

In Ca²⁺ concentrations below 10⁻⁷ M the direction of effective power stroke was toward the rear of the model, and the swimming motion was forward. The power stroke gradually shifted toward the front as the Ca²⁺ concentration was increased. This caused decreased forward velocity of swimming at Ca²⁺ concentrations of 10⁻⁷ to 10⁻⁸ M, and reversal of the swimming direction occurred at concentrations above 10⁻⁶ M. This corresponds to the backward swimming of non-extracted live animals in response to an appropriate stimulus (Pl. 1E; see also fig. 1 in Naitoh & Kaneko, 1972).

Treatment of the models with a mixture of Ca²⁺ and ATP (without Mg²⁺) induced no ciliary beating, but a reorientation of the non-beating cilia toward the front. This position of the immobile cilia corresponds to backward swimming with motile cilia (Pl. 1D).

Naitoh (1969) found that cilia of Paramecium lost their ability to beat in response to ATP and Mg²⁺ during extraction by glycerol (10–15 days at –15 °C); however, they still reversed their pointing direction in response to ATP and Ca²⁺ similarly to the present models. Sr²⁺ and Ba²⁺ also induced reversed orientation of nonbeating cilia.

These findings strongly support the proposal (Naitoh, 1966) that the ciliary system of Paramecium has at least two kinds of motile component; one component concerned with the cyclic bending of the cilium and the other regulating the orientation of the effective power stroke in beating cilia, or the pointing direction in non-beating cilia. Both use ATP as the energy source. The former requires Mg²⁺ and the latter Ca²⁺ as co-factors for the ATP-energized reactivation.

Since the end of the last century it has been known that ciliary reversal of Paramecium and other ciliated protozoans occurs on the cell surface facing the cathode when the specimen is placed in an electric field (Verworn, 1889; Jennings, 1906). More recently ciliary reversal was found to occur in close association with a membrane depolarization (Kinosita, 1954; Kinosita et al. 1964a, 1965; Naitoh, 1966; Eckert & Naitoh, 1970). It was also noted that external Ca²⁺ is indispensable for ciliary reversal (Bancroft, 1906; Kamada, 1940; Kinosita, 1954; Okajima, 1954). Recent electrophysiological examination of the Paramecium membrane showed that a regenerative depolarization is mediated by an influx of Ca²⁺ in response to an increase in the calcium conductance of the membrane (Naitoh et al. 1972). Thus, the idea was put forward that an increase in cytoplasmic calcium concentration within the cilium due to this calcium influx is responsible for the manifestation of ciliary reversal (Eckert, 1972; Eckert & Naitoh, 1972). The present finding that the extracted models of Paramecium show ciliary reversal in the presence of Ca²⁺ (above 10⁻⁶ M) strongly supports the proposal (see also ‘calcium hypothesis’ in Naitoh, 1968,1969). Eckert (1972) calculated the approximate increment in free Ca²⁺ concentration within a cilium corresponding to a i mV depolarization, which approximates to a threshold membrane depolarization for the least ciliary reversal in live specimens (Machemer & Eckert, 1972). The value was about 10⁻⁶ M, which is comparable to an effective concentration for the induction of ciliary reversal in the present models. Furthermore, a behavioural mutant of Paramecium aurelia, which never shows ciliary reversal in response to stimuli, fails to exhibit the increased Ca²⁺ conductance in response to a stimulus (Kung & Eckert, 1972). Triton-extracted models of the same mutant do, however, exhibit normal ciliary reversal in response to applied calcium (Kung & Naitoh, 1973). These lines of evidence strongly indicate Ca²⁺ –mediated membrane regulation of ciliary orientation in Paramecium.

Hyperpolarization of the membrane inhibits ciliary reversal (i.e. it normalizes the direction of the power stroke) and increases the beating frequency (Naitoh, 1958; Kinosita et al. 19646; Eckert & Naitoh, 1970). Present results clearly indicate that the beating frequency is a direct function of both Mg²⁺ and ATP concentrations (Textfigs. 8, 11). It might therefore be conjectured that changes in the Mg²⁺ concentration associated with bioelectric events in the membrane are responsible for the mechanism underlying the control of the beating frequency.

An increase in the beating frequency is also observed in association with ciliary reversal by membrane depolarization (Kinosita et al. 1965 ; Machemer & Eckert, 1973). Eckert & Murakami (1972) found an increase in beating frequency by ionophoretic application of Ca²⁺ to the ciliated cell of amphibian oviduct epithelium. They suggested an activation of ATP-yielding enzyme systems by Ca²+ as a cause of the frequency increase. However, this is not likely in the case of Paramecium because an increase in the frequency following a depolarization (and also following a hyperpolarization) occurs with very short latency (5–10 msec). Perhaps an antagonism between Mg²⁺ and Ca²⁺ in the ciliary motile system and/or calcium (or magnesium) sequestering sites (possibly in the membrane; see Naitoh, 1968) might be involved in the mechanism.

Cilia on the anterior half of a live specimen respond to membrane depolarization with shorter latency and stronger shifts to the anterior than do cilia on the posterior half (Okajima, 1953; Eckert & Naitoh, 1970). The similar localized differences of the sensitivity and the degree in changing the ciliary orientation in response to applied Ca²⁺ are seen in the models (Naitoh, 1969). This suggests that the differences are due to factors intrinsic in the ciliary apparatus. On the other hand, the present finding that reactivity of the extracted cilia to ATP and/or Mg²⁺ is the same almost all over the model (Table 1) supports the possibility that the localized difference in beating frequency found in a live animal results from localized differences in the regulatory function of the surface membrane. Some examples of localized differentiation in membrane function are well demonstrated in Opalina and Paramecium, namely, specialized mechanoreceptive (Naitoh & Eckert, 1969 a) and chemosensitive (Naitoh, 1961) areas.

Most of these experiments were carried out at the Zoological Institute, Faculty of Science, University of Tokyo, with a partial support by a grant for Fundamental Scientific Research from the Ministry of Education, Japan, to one of the authors (Y. N.). We would like to thank Drs T. Nakazawa and T. Yamada of National Institute of Radiological Sciences, Chiba, Japan for their kind technical help and for giving us reagents for the enzymic determination of ATP and ADP. We also thank Dr R. Eckert of U.C.L.A. for his critical reading of the manuscript which was prepared in his laboratory with the support of N.I.H. grant NS 8364.

Photographs of the models in various solutions taken by single xenon flashes through a bright-contrast phase objective (Nikon ×20) (Magnification about x 230). (A) 50 mm-KCl solution. Cilia did not beat. Pointing direction of the cilia was toward posterior or perpendicular to the surface. (B) A forwardswimming model in the KC1 solution with 4 mm ATP, 4 mm-MgCl₂ and 3 mm EGTA. Black arrow indicates the swimming direction. Small white arrows indicate metachronal waves, which are propagated in the direction of large white arrow (antiplectoidal). (C) 50 mm-KCl solution with 0·05 mm-CaCl₂. Cilia do not beat and point perpendicular to the surface. (D) 50 mm-KCl solution with 0·05 mm-CaCl₂ and 4 mm ATP. Cilia do not beat, but had reoriented once to point toward the front. Small white arrows show approximate pointing direction of cilia. (E) A backward-swimming model in the KC1 solution with 4 mm ATP, 4 mm-MgCl₂ and 0·05 mm-CaCl₂. Black arrow shows the swimming direction. Small white arrows indicate the metachronal waves which were conducted in the direction of the large white arrow (symplectoidal). (a) Anterior end of the models.