Mechanical Stimulation of Starfish Sperm Flagella

Although in recent years considerable progress has been made in studies of flagellar movement, from both structural and functional points of view (cf. Sleigh, 1974), details of mechanisms of the initiation of bending and of the conduction of the bending wave are still unknown. In sea-urchin spermatozoa, the bending wave is initiated at the basal region of the flagellum and propagates along the flagellum without decrement (Gray, 1955). However, if an obstacle in the path of the wave impedes the movement of the flagellum, the bending wave cannot propagate past the obstacle. Machin (1958) tried to explain the observed waveform of the flagellum by assuming that active contractile elements are distributed along the entire length of the flagellum and that they can be activated by local passive bending. This ‘mechanical propagation hypothesis’ has been extended by Machin (1963) and Brokaw (1966).

Recently, active bending force has been presumed to be generated by active sliding among the outer doublet microtubules (Satir, 1968; Summers & Gibbons, 1971; Warner & Satir, 1974; Lindemann & Gibbons, 1975). Brokaw (1971, 1972) and Lubliner & Blum (1972) attempted to explain the flagellar movement by a ‘sliding microtubule hypothesis’ based on mechanical responsiveness of the motile elements distributed along the flagellum. The legitimacy of this assumption as the basis of the above theoretical works can be proved by an experiment showing that an active mechanical response is initiated in the motile elements of the flagellum by a mechanical strain, and that a strain sufficiently large for the initiation of an active response is generated by mechanical activity of adjacent elements of the flagellum.

In the present paper, responses of starfish sperm flagella to various mechanical stimuli were investigated in order to examine whether or not the initiation and propagation of the flagellar bending wave can be explained by a mechanism based on the responsiveness of the flagellum to mechanical stimulation.

Spermatozoa of the starfish Asterina pectinifera were suspended in artificial sea water containing 1 mm histidine, which activated sperm motility (Fujii et al. 1955). The sperm suspension was poured into a trough (Fig. 1). A coverslip, coated with a thin layer of agar gel on its under surface, was put on the trough. (The gel consisted of 0·5% agar in artificial sea water containing 10 mm histidine). Before placing on the trough, part of the gel was removed by cutting with a knife at right angles to the coverslip surface. Some of the spermatozoa adhered with their heads to the cut surface of agar, beating their flagella. Only the spermatozoa whose flagella were beating in the plane of the microscopic field were used in experiments.

A microneedle fixed to a piezo-electric bender element (B in Fig. 1) was inserted with a micromanipulator into the trough through one of the side openings of the trough. The tip of the microneedle could be moved in the direction of its axis by supplying pulses to the bender element from an electronic pulse generator (Nihon-koden MSE-3). Rectangular pulses (0·3–0·4 s duration) were used in Expts 1, 2 and 3 and saw-toothed pulses (15–50 Hz) in Expt 4. For Expts 1, 2 and 3, a micro-needle with a rectangular cross-section (2 μm × 20 μm at the tip) was used. For Expt 4, a microneedle with a square cross-section (5–20 μm thick at the tip) was used. In order to amputate a motile flagellum from a spermatozoon in Expt 3, the flagellum was cut by pressing against the coverslip with an ordinary microneedle of 1 μm tip diameter. The amputation was made in a solution containing 0·4 M sucrose, 0·1 M-KCI, 4 mm-MgSO₄, 10 mm-Tris-HCl buffer (pH 8-2), 1 mm ethylene diamine tetra acetate (EDTA), 5 mm dithiothreitol (DTT) and 1 mm-ATP.

A Nikon microscope with phase contrast optics (BM 40 × objective and HKW 10x ocular) was used for observation and recording. Spermatozoa were illuminated by light from a high-pressure mercury vapour lamp (Ushio, USH 100D-3) passing through a thick layer of Mohr’s salt solution. Movement of flagella was recorded with a 16 mm high-speed cine-camera (HYCAM 20 S1) on Kodak plus × or 4X films at rates of 200 to 2300 frames/sec. The frequency of the flagellar beating was determined either from the cinematographic record or by means of stroboscopic observation, in which case the illumination was interrupted by vibrating a mirror in the optical path. Vibration was by means of an electromagnet operated by another electronic pulse generator (Nihon-koden MSE-3).

Curvature of the flagellum, and the velocity and bend angle of the bending wave, were determined from the cinematographic record. Curvature was defined as the reciprocal of the radius of the arc which described that part of the flagellum which was maximally bent. The amplitude of the bending wave was determined as a half of the breadth of the envelope of beat of the flagellum. Bend angle was defined as the angle between the tangent at two adjacent points of inflexion. The velocity of the wave was determined as the velocity of the point of maximum curvature along the flagellum.

As reported by Brokaw (1965) in spermatozoa of other species, the waveform of the flagellum of a freely swimming starfish spermatozoon (Fig. 2a) was considerably different from the waveform of one that was attached to the agar gel, or coverslip, at its head (Fig. 2 b). The waves of the attached spermatozoon were compressed in the direction of the wave propagation. Both the curvature and bend angle in the attached spermatozoon were larger than those in a freely swimming one. The curvature was large at the proximal region and decreased as the wave propagated to the distal portion. The wave length increased as the wave propagated, too. The amplitude of a wave attained a maximum in the middle region, as shown by the example in Fig. 2(b), in which amplitude attained a maximum of 5·6 μm at a point 15 μm from the head.

When the amplitude of a beating flagellum was restricted by a microneedle slowly approached to within a few μm from the wave axis of the flagellum, both the amplitude and the bend angle decreased over the entire length of the flagellum, as shown in Fig. 2. In the case of Fig. 2(c), the amplitude was reduced from 4·5 to 2·5 μm at a point 10 μm from the head. The amplitude attained a maximum value of 3·9 μm at a point 18 μm from the head, smaller than the maximum amplitude (5·6μm) of the unrestricted wave. The curvature and the bend angle of the freely beating flagellum (Fig. 2,b) and those of the restricted flagellum (Fig. 2,c) are shown in Fig. 3. The curvature of the unrestricted flagellum was about 3·5 × 10³ cm⁻¹ at 10 μm from the head, and it decreased to 2·2 × 10³ cm⁻¹ beyond 30 μm from the head. The curvature of the restricted flagellum was almost the same as that of the unrestricted flagellum, being 3·1 × 10³ cm⁻¹ at 10μm from the head where the microneedle restricted the amplitude, and 2·2 × 10³ cm⁻¹ at 30 μm from the head. The bend angle of the unrestricted flagellum first increased as the distance from the head increased, attaining a value of 2·8 radian at 10 μm from the head as a maximum value, and then it gradually decreased to 2·4 radian at a point 30 μm from the head. In the restricted flagellum, the bend angle was smaller, keeping a constant value of 1·9 radian in the region beyond 10 μm from the head. Both the number of waves contained in the flagellum, and the wave length measured along the flagellum, were slightly affected by restriction of the amplitude: the number of waves contained in the flagellum was 1·7 in the case of Fig. 2(b) and 1·9 in the case of Fig. 2(c); the wave length in the middle region of the flagellum was 27 μm in the former and 24 μm in the latter.

When the amplitude was further restricted as shown in Fig. 2 (d), the decrease in amplitude over the entire length of the flagellum was more pronounced. In many cases, the bending waves became irregular and asymmetrical. The number of waves contained in the flagellum scarcely changed and the bend angle was almost constant in the portion distal to the arrested part.

In the case shown in Fig. 2, the beating frequency was 33 Hz before restriction. As the microneedle approached the flagellum and began to make contact with it, the frequency increased up to 44 Hz (Fig. 2,c) within three beats. Further approach resulted in a slight frequency decrease to 39 Hz (Fig. 2 d). The frequency recovered its initial value within three beats after the removal of the microneedle. An increase in frequency was not always observed. In some cases, the frequency was decreased by restriction. The beating frequency of a restricted flagellum did not remain constant.

When a part of a beating flagellum was rapidly hit by a microneedle, the movement of the part in contact with the microneedle was arrested as shown in Fig. 4. The bending wave which had passed the arrested part before the microneedle touched propagated in the distal part with a speed similar to that of the wave in the flagellum without arrest (1·3 × 10³μm/s for Fig. 4,a and 1·0× 10³μm/s for Fig. 4,b), while the wave which was still in the part between the base and the arrested part was extinguished at the arrested part (cf. Figs. 4 a and 5). The curvature and the bend angle of the wave which had passed the arrested part were propagated to the end of the flagellum without appreciable decrement. After the waves which had existed at the time of arrest had passed over the entire length, the distal part became straight and immobile. The direction of the axis of the flagellum after it became immobile often coincided with the direction of the mean axis of beating before the arrest, although this was not always the case.

The wave which existed in the region between the arrested part and the base of the flagellum continued to propagate within this region with decreased speed and frequency of vibration. In some cases, the flagellum ceased to vibrate, having a wavy form. In the former case, new waves were formed at the base and propagated towards the arrested part, where they were extinguished. An example is shown in Fig. 5, in which the wave length was smaller and the maximum bend of the flagellum was larger than in a normally beating one. Propagation of waves was never observed when the distance between the base and the arrested part was smaller than 10 μm. When the distance was larger than 10 μm, propagation of waves was sometimes observed in the region proximal to the arrested point. It seems likely that whether the wave can propagate or not depends not only on the length of the flagellum between the base and the arrested point but also on the degree of constraint of the flagellum in that region.

When the microneedle was removed from the flagellum after an arrest of short duration (0·3 – 0·4 s), the bending waves which existed between the base and the arrested part propagated beyond that part to the distal end, as did also those initiated at the base, with a speed slightly larger than that in normally beating flagella (cf. about 2·0 × 10³μm/s for Fig. 6,a). The velocities of the second and subsequent waves recovered to the value in normally beating flagella. A bending wave could also be initiated from the fixed bend in the part between the base and the arrested part mentioned above, and from the bend passively formed by the removal of the microneedle, as well as from the base, as shown in Fig. 6(b). The bending waves always propagated from the proximal end to the distal end. New bending waves were never generated de novo in the distal part beyond the arrested point when the microneedle was removed.

The head of a spermatozoon attached to the coverslip was cut off the flagellum by pressing a microneedle against the coverslip at a point close to the head. When the operation was made in sea water, the flagellum stopped beating after making a few irregular vibrations. Application of an external force made a sharp bend in the flagellum at the point of application, other parts remaining straight. It did not recover its original shape after the removal of the external force. It was as if the flagellum had become brittle, losing its flexibility and elasticity.

When the head was cut off in the medium containing 1 mm-ATP (see Materials and Methods), beating was observed in the amputated flagellum for one hour or more after the amputation, although the beating was irregular, the amplitude was small (2 – 3 μm or less) and the beating frequency was low (about 10 Hz). A typical example is shown in Fig. 7. The direction of bend propagation was from the proximal end, where the amputated flagellum was attached to the coverslip, toward the distal end. Propagation in the reverse direction was never observed. The beating form was asymmetrical and both the curvature and the bend angle were lower on one side of the axis than on the other. DTT and EDTA in the medium seemed to be favourable for retaining motility for a long period but over a period of about 5 min they were not necessary.

If the operation was made in a medium containing ADP (1 mm) instead of ATP, the amputated flagellum vibrated several times just after amputation and soon became immobile, but it was not so stiff as the flagellum amputated in sea water. In a high concentration of ADP (above 5 mm) the amputated flagellum continued to beat in just the same manner as that in the medium containing ATP.

If the proximal region of an amputated flagellum affixed to the coverslip at its cut end was bent with a microneedle, this passive bend propagated along the entire length of the flagellum as shown in Fig. 8. The propagation velocity (about 100 μm/s) was generally smaller than that in normal beating. Although the curvature of the bend often increased in the middle region of the flagellum in the course of propagation, the bend angle was always constant from the bend initiation to the extinction at the distal end of the amputated flagellum. As shown in Fig. 8, a bend of opposite sign to the first (double arrows in the figure) was often initiated in the middle region and propagated along the flagellum, although both the curvature and the bend angle were smaller than those of the first wave. The bending wave could also be initiated at the middle region of the amputated flagellum. In this case, the direction of the bend propagation was always towards the distal end. The propagation velocity, the curvature and the bend angle of a bending wave initiated in the middle region of the flagellum were almost the same as those of a bending wave initiated in the proximal region.

In the 5 mm-ADP medium, the amputated flagellum showed the same responses to passive bending as those in the 1 mm-ATP medium.

An attempt was made to change the frequency of flagellar beating by applying rhythmic forces with a frequency different from the intrinsic frequency of the flagellum, employing the microneedle with a square cross-section (see Materials and Methods). This was done in three different ways: (a) with the tip of a microneedle vibrating with the same frequency as the beating flagellum, the needle was advanced to the flagellum and then the frequency of vibration was gradually increased or decreased; (b) after the needle was advanced as above, the frequency of vibration was abruptly changed ; (c) with the microneedle vibrating with a frequency different from the beating frequency of the flagellum, it was gradually advanced to the flagellum. In all three cases, when the tip of the microneedle was approached within 5 – 10 μm from the flagellum (not touching it), the frequency of the flagellar beating could be modulated, within a limited range, to a frequency the same as that of the vibration of the microneedle. Some representative results are shown in Fig. 9.

The modulation was most effectively made by applying rhythmic forces to the region within 10 μm from the base of the flagellum; however, the effective region depended on the shape and thickness of the microneedle and the distance between the microneedle and the flagellum. Modulation was most easily achieved when the first vibration of the microneedle occurred in phase with the flagellar beating and in a region close to the flagellum. When the vibrations of a microneedle situated near the distal portion of the flagellum were not in phase with the flagellar beat, the beat of the flagellum often became irregular as a result of its touching the microneedle. Synchronization of the flagellar beating with the vibration of the microneedle was observed to be independent of the angle between the direction of vibration of the microneedle and the mean axis of the flagellar beat, within a range of ± 0·7 radian from parallel. The maximum range of modulation of the frequency was ± 8 Hz. With the above method, it was not possible to adjust the frequency of flagellar beating to a value higher than 43 Hz nor lower than 23 Hz. Outside this range, beating was put into disorder by touching of the flagellum with the microneedle. This frequency range almost coincided with the range of beating frequency observed in normally beating flagella (20 – 45 Hz). Stroboscopic observation showed that the frequency of normal flagellar beating in each spermatozoon was practically unchanged during an experiment (within ±1 %).

There are three stages in the rhythmic beating activity of flagella: initiation of bending ; the bending itself ; and the propagation of bending waves along the flagellar shaft. It has been shown in the present study that the bending waves which propagate along the flagellum are spontaneously initiated at the proximal region of the amputated flagellum in a medium containing Mg-ATP (Expt 3). The bending waves of the amputated sperm flagella are characterized by small amplitude, small curvature, small bend angle and irregular beating. Bending waves with large amplitude, large bend angle and large curvature can only be initiated by applying mechanical strain to a part of the flagellum by external force. This result may indicate that the centriole is not indispensable for initiation and propagation of bending waves in flagella, though a possibility still remains that the presence of a centriole or other structures at the basal region is indispensable to generate normal regular bending waves. Lindemann & Rikmenspoel (1972 a, b) reported that coordinated waves could be initiated by Mg-ATP, under certain conditions, in bull and Drosophila spermatozoa in the absence of a centriole, although the flagella were generally soon deactivated by impalement or dissection with a microneedle in an ATP-free medium. Brokaw & Gibbons (1973) showed that bending can be initiated in any region of a sperm flagellum by applying ATP to partially demembranated sperm flagella. In sea-urchin and starfish spermatozoa the flagellum amputated by laser microbeam irradiation does not continue beating (Goldstein, 1969), whereas in Crithidia oncopelti the amputated flagellum can propagate bending waves in either direction, both from proximal to distal ends, and from distal to proximal ends (Goldstein, Holwill & Silvester, 1970). The result with C. oncopelti flagella may indicate that the flagellar shaft possesses the ability to initiate bending and to propagate bending waves in the absence of the centriole. The difference between echinoderm spermatozoa and C. oncopeltimay mainly be due to the existence of a wave initiating centre at the distal end of the flagellum in the latter material. The effects of mechanical constraint of part of a flagellum in the present study are similar to those found in C. oncopelti (Holwill & MacGregor, 1974).

It has been shown in the present study (Expt 2) that the propagation of the bending wave is blocked at the part of the flagellum which is arrested with a microneedle. It is unlikely that the structures responsible for initiation and propagation of the bending wave are broken by arresting, because the propagation of bending waves beyond the arrested part recovers if the microneedle is removed. It seems likely that blockage of propagation of the bending wave is due to the suppression of that deformation of the axoneme (including sliding and/or bending of microtubules) which is necessary to initiate active bending in the distal part by active sliding and/or contraction of microtubules. A new bending wave can start either from the wave which has arrived at the arrested part, or, in the case of intact flagella (Expt 2), from the passive bend formed at the time of the removal of the microneedle. The above observations, and the observation that it is necessary for an amputated flagellum to be bent with a microneedle to initiate large bending waves, strongly suggest that the propagation of bending waves depends on the responsiveness of elementary sections of the axoneme to passive mechanical strain which is normally generated by active mechanical activity of an adjacent section of the axoneme, as assumed by previous investigators (Machin, 1958, 1963; Brokaw, 1966, 1971, 1972; Lubliner & Blum, 1972). Decrease in amplitude of beating over the entire length of the flagellum by restriction of the amplitude at a distal point (Expt 1) may be due to restriction of the amplitude bringing about a decrease in mechanical strain in the flagellum, and thus affecting mechanical activity.

In normally beating flagella, both curvature and bend angle of the bending wave change in the proximal part (within 10 μm from the head). In the distal part beyond 10 μm from the head, the curvature gradually decreases as the distance from the head increases, but the bend angle is almost constant. In all the experiments of the present study, the bend angle tended to be constant in the course of the wave propagation, namely in the wave passing beyond the restricted part (Expt 1 ), in the first wave on recovery (Expt 2), and in the passively initiated bending wave in the amputated flagellum (Expt 3), whereas the curvature in these cases varied. These results may suggest that an upper limit exists to the amount of sliding among microtubules, because the bend angle is considered to be proportional to the amount of sliding.

It has been shown by Expt 3 that the flagellum becomes brittle, losing flexibility, if it is severed from the head by amputating with a microneedle in a medium lacking ATP, whereas the flagellum keeps flexibility if the operation is made in a medium containing ATP. It seems likely that the state of rigour was due to a lack of ATP. Decrease in flexural rigidity of the flagellum by addition of ATP to the medium was shown by Lindemann, Rudd & Rikmenspoel (1973) in bull spermatozoa impaled with a microneedle, and by Okuno (1974) in demembranated flagellar models of sea-urchin spermatozoa. It is considered that rigid bindings are formed between adjacent doublet microtubules by dynein arms in the absence of ATP (cf. Gibbons & Gibbons, 1974).

Spontaneous beating activity was observed in flagella amputated in the ATP medium (Expt 3). Similar observation was made by Lindemann & Rikmenspoel (1972a) in bull sperm impaled with a microneedle. It is supposed that exogenous ATP is utilized in flagellar movement as in the case of sperm models. The observation that ADP can induce motility in amputated flagella has also been made with starfish (Expt 3) and mammalian spermatozoa (Lindemann & Rikmenspoel, 1972b).

It is known that the beating of flagella close to each other tends to synchronize sperm or spirochaetes (Gray, 1928). Mechanical interference among the cilia on ciliated epithelium is considered to be one of the most important factors in ciliary metachronism (Sleigh, 1974). Tsuchiya (1969) succeeded in modulating the beating frequency of the large abfrontal cilium of Mytilus by rhythmically hitting the basal part of the cilium with a microplate. In the present study (Expt 4), modulation of beating frequency was achieved in sperm flagella by applying rhythmic forces to flagella by moving the medium with a vibrating microneedle close to, but not touching, the flagella. The synchrony of the beating of flagella close to each other, mentioned above, may be explained by the modulation of frequency by rhythmic external forces. However, it is still unknown whether modulation of the frequency is due to mechanical stimulation of the pacemaker which exists at the basal region of the flagellum or if it is due to a direct mechanical effect on the flagellar shaft.

The authors wish to thank Professor H. Sugi, Teikyo University for putting at the authors’ disposal his high-speed cine-camera for the present experiments. Thanks are also due to the Director and the Staff of Misaki Marine Biological Station for supplying the materials. This work was supported in part by Research Expenditure of the Ministry of Education, Funds from Cooperative Program (Nos. 74108, 75116) provided by the Ocean Research Institute, University of Tokyo and a Mainichi Science Encouragement Award, given to one of the authors (Y.H.).