Cell Motility and the Cytoskeleton 44:85–95 (1999) Movement of Turritella Spermatozoa: Direction of Propagation and Chirality of Flagellar Bends Sumio Ishijima,1* Sanae A. Ishijima,2 and Björn A. Afzelius3 1Biological Laboratory, Faculty of Bioscience and Biotechnology, Tokyo Institute of Technology, O-okayama, Meguro-ku, Tokyo, Japan 2Department of Biology, Japan Women’s University, Mejirodai, Bunkyo-ku, Tokyo, Japan 3Department of Ultrastructure Research, Wenner-Gren Institute, University of Stockholm, Stockholm, Sweden The marine snail, Turritella communis, produces two types of spermatozoa, named apyrene and eupyrene. Eupyrene spermatozoa are usually paired, but unpaired ones are involved in fertilization. Movements of these spermatozoa were analyzed using a video camera with a high-speed shutter. The eupyrene spermatozoa usually swim with the head foremost but are able to swim flagellum foremost. A reversal of the direction of their swimming was found to be the result of a change in the direction of flagellar bend propagation, which changed with calcium concentration. Reversal of the direction of bend propagation was accompanied by a reversal of direction of the rotational movement of the spermatozoa around their long axis, suggesting that the bending waves keep the sense of their three-dimensional form. The swimming speed of apyrene spermatozoa in natural seawater was about one-eighth of that of the eupyrene ones and remained almost constant in highly viscous medium.The swimming speed of conjugated eupyrene spermatozoa was the same as that of unpaired spermatozoa over a wide viscosity range (⬍3,000 cP). No advantage of swimming by two spermatozoa could be detected in Turritella spermatozoa. Cell Motil. Cytoskeleton 44:85–95, 1999. r 1999 Wiley-Liss, Inc. Key words: bidirectional swimming; conjugated spermatozoa; marine snail; 9 ⴙ 2 axoneme; rotational movement; sperm dimorphism INTRODUCTION Turritella spermatozoa are unusual in that two types exist, apyrene and eupyrene spermatozoa, and that the eupyrene ones, which are usually paired, can swim backward as well as forward [Afzelius and Dallai, 1983]. These peculiarities are useful for an understanding of the mechanism generating flagellar bends and of the meaning of sperm dimorphism and sperm conjugation. A change in the swimming direction has been reported in several protozoa and spermatozoa [Holwill and McGregor, 1976; Afzelius, 1983; Baccetti et al., 1989; Curtis and Benner, 1991; Ishida et al., 1991; Ishijima et al., 1994]. A reversal of the swimming direction of Crithidia cells and spermatozoa of tephritid flies and of Myzostomum has been shown to be due to a change in the direction of bend propagation and to be r 1999 Wiley-Liss, Inc. regulated by intracellular Ca2⫹ concentrations [Holwill and McGregor, 1976; Baccetti et al., 1989; Ishijima et al., 1994]. Furthermore, a reversal of the swimming direction of the tephritid spermatozoa was accompanied by a reversal of rotation of the spermatozoa around the long axis [Baccetti et al., 1989]; the significance of these phenomena, however, remained to be studied. It has long been known that two types of spermatozoa are produced in several phyla (sperm dimorphism), especially Lepidoptera and Gastropoda [Wilson, 1925]. *Correspondence to: Dr. Sumio Ishijima, Biological Laboratory, Faculty of Bioscience and Biotechnology, Tokyo Institute of Technology, 12-1, O-okayama 2-chome, Meguro-ku, Tokyo 152-8551, Japan. E-mail: sishijim@bio.titech.ac.jp Received 30 November 1998; accepted 14 July 1999 86 Ishijima et al. Fig. 1. Phase-contrast video micrographs of unpaired eupyrene Turritella spermatozoa propagating bending waves toward the base of the flagellum (A) and toward the tip (B). The time interval between successive images is 1/30 s (A) and 1/60 s (B). Arrowheads indicate the corresponding bends. Bar ⫽ 30 µm. TABLE I. Movement Characteristics of Eupyrene Turritella Spermatozoa Free-swimming spermb Sperm attached to coverslipa Direction of bend propagation Base to tip Unpaired spermatozoa Paired spermatozoa Tip to base Unpaired spermatozoa Amplitude (µm) Wavelength (µm) Beat frequency (Hz) Swimming speed (µm/s) 7.2 ⫾ 1.7 7.3 ⫾ 2.2 52.9 ⫾ 4.2 51.2 ⫾ 2.3 30.0 ⫾ 10.0 30.0 ⫾ 14.1 144.0 ⫾ 27.2 169.0 ⫾ 33.4 3.6 ⫾ 1.5 25.6 ⫾ 2.0 8.1 ⫾ 1.7 11.8 ⫾ 5.1 ⫾ S. D. for samples of 23 spermatozoa propagating bends from base to tip from four experiments and of 20 spermatozoa propagating bends from tip to base from five experiments. bMean ⫾ S. D. for samples of 15 spermatozoa propagating bends from base to tip from four experiments and of 18 spermatozoa propagating bends from tip to base from five experiments. aMean This phenomenon has attracted many investigators, and light and electron microscopic studies have been carried out on the morphology of the apyrene and the eupyrene spermatozoa and their spermatogenesis [Koike and Nishiwaki, 1980; Braidotti and Ferraguti, 1982]. The eupyrene spermatozoa are known to be essential for fertilization [Kennedy and Keegan, 1992], but the function of the apyrene spermatozoa remains obscure. Conjugated spermatozoa have been described from four or five animal groups; American marsupials [Biggers and Creed, 1962; Biggers and DeLamater, 1965; Phillips, 1972; Temple-Smith and Bedford, 1980; Taggart et al., 1993], the snail family Turritella [Afzelius and Dallai, 1983], the primitive apterygote insect group Thysanura [Dallai and Afzelius, 1984], dytiscid water beetles [Dallai and Afzelius, 1985, 1987], and the diplopod myriapods [Reger and Fitzgerald, 1979]. These studies suggest the pairing is an adaptation for a more effective swimming capacity, although many other suggestions as to the function of the conjugated spermatozoa have been offered [Afzelius and Dallai, 1987]. Movement of Turritella Spermatozoa 87 Fig. 3. A eupyrene Turritella spermatozoon swimming in a relatively deep suspension. Micrographs of a spermatozoon swimming in a circular path close to a coverslip when an observer views the cell from above. Because the beating plane is parallel to the optical axis, segments of the sperm flagellum are in focus. The time interval between successive images is 1/30 s. Arrowheads indicate the corresponding bends. Bar ⫽ 30 µm. TABLE II. Percentage of Spermatozoa With Flagellar Waves Propagating Proximally* Medium Fig. 2. The time-course of flagellar movement of an unpaired eupyrene Turritella spermatozoon. Bar ⫽ 30 µm. See text for details. The purpose of this study was to investigate in detail the observation by one of us (B.A.A.). The movements of Turritella spermatozoa, not only the eupyrene spermatozoa but also the apyrene spermatozoa, were recorded by means of a charged-coupled device (CCD) video camera with a highspeed shutter. Movement characteristics of conjugated eupyrene spermatozoa were compared with those of unpaired spermatozoa. Effects of free calcium ions and of viscosity variations on these movements were also examined. The results obtained suggest that changes in the calcium concentration affect the direction of bend propagation, which in turn determines the swimming direction of the spermatozoa. From Standard solution ⫹ 0.3 mM EGTA Standard solution ⫹ 0.1 mM EGTA Standard solution ⫹ 1 mM CaCl2 Standard solution ⫹ 3 mM CaCl2 % n 0 16 88 100a 30 37 32 29 *Standard solution consists of Ca-free seawater (470 mM NaCl, 10 mM KCl, 54 mM MgSO4 , and 10 mM Tri-HCl, pH 8.2) and 20 µM calcium ionophore A23187. aSpermatozoa ceased their movement within one minute. these results, it is argued that the rotation of spermatozoa around their long axis is related to the chirality of the bending waves. MATERIALS AND METHODS Turritella communis Lamarck snails were collected near the Kristineberg Marine Biological Station, Fiske- 88 Ishijima et al. Fig. 4. Phase-contrast video micrographs of the rotational movement of Turritella spermatozoa observed along the sperm axis. Spermatozoa propagating bending waves distally roll counterclockwise when viewed from the head (A), whereas spermatozoa propagating bending waves proximally roll clockwise (B). The time interval between successive images is 1/60 s (A) and 1/2 s (B). Arrows indicate the direction of the beating planes determined from the parts of the flagellum in focus. Bar ⫽ 10 µm. bäckskil, Sweden, in the summers of 1990 and 1996. Snails having a total length of 2.3–3.0 cm were used. Spermatozoa were obtained by micropipettes from the gonad after removing the shell. To prepare the unpaired eupyrene spermatozoa, the conjugated spermatozoa were separated by pipetting several times. The movements of the sperm and their flagella were observed in natural seawater, artificial seawater consisting of 470 mM NaCl, 10 mM KCl, 10 mM CaCl2, 54 mM MgSO4, and 10 mM Tris-HCl, pH 8.2, and artificial seawater with the viscosities of 110, 230, 780, and 3,000 cP, which were increased with methylcellulose (viscosity of 2% aqueous solution: 4,000 cP, Tokyo Kasei Kogyo Co., Ltd., Tokyo). The viscosity of artificial seawater with methylcellulose was measured by a falling ball viscometer [Ishijima et al., 1986] at 18°C. In experiments examining the effect of Ca2⫹ on the sperm movement, CaCl2 or [ethylenebis(oxyethylenenitrilo)]tetraacetic acid [EGTA] was added to a standard solution, which consists of Ca-free artificial seawater and 20 µM calcium ionophore A23187 (ICN Biomedicals, Inc., Aurora, OH). A stock solution of 10 mM calcium ionophore A23187 was prepared in dimethyl sulfoxide (E. Merck, Darmstadt, Germany) and stored at 4°C. The presence of dimethyl sulfoxide did not change the movement characteristics of the spermatozoa. For examination and recording of sperm movement, samples were transferred onto a glass slide beneath a raised coverslip (0.18 mm depth). In the experiments examining the rotational movement of the spermatozoa around their long axis, samples were placed into a 0.30-mm-deep observation chamber and covered with a coverslip precoated with 0.5% agar and dried [Ishijima et Movement of Turritella Spermatozoa al., 1992]. This treatment with agar increased the number of spermatozoa attached vertically to the coverslip and did not change the rolling direction and the rotation frequency of the spermatozoa (data not shown). The movements of Turritella sperm and their flagella were recorded using a Leitz Dialux 22 microscope equipped with a phase-contrast condenser, objectives (Neofluor 40/0.75), and 12.5⫻ eyepieces, or a Leitz Laborlux S microscope with a phase-contrast condenser, objectives (EF 40/0.65 and 25/0.50), and 10⫻ eyepieces. Images were recorded on VHS 1⁄2-inch cassette videotape with a Panasonic CCD video camera with a shutter of 1/500 s (WV-BD 400, Matsushita Communication Industrial Co., Ltd., Yokohama). For preparation of the figures, images on the video monitor were photographed with a frozen field using Fuji Neopan F 35-mm film. Observations and recording were made at 17–20°C. The movements of the sperm and their flagella were analyzed by methods previously described [Ishijima et al., 1994]. The direction of sperm rotation around its longitudinal axis was determined from the direction of rotation of flagellar images [Ishijima et al., 1992]. RESULTS Bidirectional Propagation of Flagellar Waves of Eupyrene Spermatozoa in Natural Seawater In natural seawater, both forward and backward swimming spermatozoa were seen in a microscopic field. All spermatozoa (more than 100 spermatozoa) that propagated the flagellar waves from the base of the flagellum to the tip swam with the head foremost, whereas all spermatozoa (more than 100 spermatozoa) that propagated the waves from the tip to the base swam with the flagellum foremost. These spermatozoa exhibited one characteristic feature: the distal end of the flagella hardly exhibited transverse movement while the sperm flagella beat (Fig. 1). Movement characteristics of these two types were, however, quite different; spermatozoa swimming with the head foremost moved fast and had flagellar waves of large amplitude and high beat frequency, compared with spermatozoa swimming with the flagellum foremost (Table I and Fig. 1). Turritella spermatozoa changed their swimming direction by a spontaneous reversal of bend propagation. Generally, the spermatozoa kept propagating distally the flagellar waves for approximately 10 min (Fig. 2A). The proximally propagating flagellar waves appeared first in the tip region of the flagella, then gradually extended toward the base of the flagellum (Fig. 2B and C), and finally covered the whole flagellum (Fig. 2D) by approximately 5 min. Usually the proximally propagating waves kept beating for more than 10 min. The sequence of these changes in flagellar waves was not so exact; the spermato- 89 Fig. 5. Relationship between the swimming speed and the viscosity of the medium. Each point represents the average of 11 to 27 spermatozoa. The vertical bars are standard deviations. The lines are weighted least square regression lines with a slope of ⫺0.35 for paired (䊉) and ⫺0.33 for unpaired (䊊) spermatozoa. zoa that had flagellar waves propagating proximally often recovered the flagellar waves of large amplitude and of high beat frequency and changed the direction of wave propagation. Addition of 10 mM caffeine or 8-bromoadenosine 3’:5’-cyclic monophosphate (Sigma Chemical Co., St. Louis, MO) to the sperm suspensions prolonged the movement of the distally propagating waves and restored the active movement of sperm flagella that had propagated proximally. Therefore, three types of sperm flagella could be seen in the field of viewing: (1) flagella propagating large bending waves distally; (2) flagella propagating short bending waves proximally; and (3) flagella propagating waves simultaneously in two directions with the distally propagating waves in the proximal region of a flagellum and the proximally propagating ones in the distal region. The propagation velocity of the bending waves was not always the same in a flagellum; the propagation velocity of distally propagating waves in the distal region was faster than that in the proximal region. When the eupyrene spermatozoa were swimming in a deep observation chamber, they moved in spiral paths. Spermatozoa that came in contact with the coverslip or the slide glass moved almost in circles. Their beating plane was curved and perpendicular to the glass surface (Fig. 3). A curved beating plane was also noticed in spermatozoa sticking to the glass surface by the head; the entire length of a sperm flagellum was not completely in focus (Fig. 1). The flagellar waves were fairly symmetrical with a constant amplitude over the entire flagellum when viewed from the direction perpendicular to the beating plane. No significant difference in the movement characteristics (amplitude, wavelength, and beat frequency of 90 Ishijima et al. Fig. 6. Typical bending waves of unpaired eupyrene Turritella spermatozoa swimming in media of increased viscosity. A: 110 cP. B: 230 cP. C: 3,000 cP. Bar ⫽ 30 µm. flagellar waves and swimming speed of spermatozoa) could be detected between paired and unpaired spermatozoa (Table I). Effects of Calcium Ion on the Propagating Direction of Flagellar Waves To elucidate the factor influencing the direction of flagellar wave propagation, the effects of calcium ion on the propagation direction of flagellar waves were examined by altering the concentration of CaCl2 or EGTA in the presence of 20 µM calcium ionophore A23187. A decrease in EGTA and an increase in CaCl2 in the standard solution increased the percentage of the spermatozoa propagating flagellar waves proximally (Table II). The parameters of flagellar waves propagating proximally in the standard solution did not differ essentially from that in natural seawater (data not shown) except for the duration of motility, which rapidly decreased with high calcium concentrations. Rotational Movement of an Unpaired Eupyrene Spermatozoon Around Its Axis The rotational movement of spermatozoa around their long axis is due to the three-dimensional geometry of their flagellar waves [Taylor, 1952; Gray, 1962; Ishijima et al., 1992]. Therefore, valuable information on the three-dimensional geometry of flagellar waves will be obtained from the parameters of the rotational movement of spermatozoa; for example, knowledge of the direction of sperm rotation gives us the sense of the threedimensional geometry in flagellar waves. To examine whether or not the conformational change of flagellar waves of spermatozoa is caused by changing the propagation direction of bending waves, the rotational movement of spermatozoa around their axis was examined using unpaired eupyrene spermatozoa in standard solution containing either 0.3 mM EGTA or 1 mM CaCl2. In the former solution, all spermatozoa had distally propagating flagellar waves, whereas in the latter solution almost all spermatozoa had proximally propagating flagellar waves (Table II). All spermatozoa (n ⫽ 58) propagating bending waves from the base of the flagellum to the tip rotated counterclockwise, as viewed from the head to the tail (Fig. 4A). The average rotation frequency was 10.1 ⫾ 3.1 Hz (n ⫽ 15). On the other hand, all spermatozoa (n ⫽ 22) propagating bending waves from the tip to the base rotated clockwise with the rotation frequency of 1.2 ⫾ Movement of Turritella Spermatozoa 91 Fig. 7. Phase-contrast video micrographs of paired eupyrene spermatozoa beating in natural seawater (A) and in viscous seawater of 110 cP (B). The time interval between successive images is 1/60 s. Arrowheads indicate the corresponding bends. Bar ⫽ 30 µm. 0.49 Hz (n ⫽ 12) (Fig. 4B). These results show that the chirality of flagellar waves does not change when the direction of wave propagation along the sperm flagella changes. Effects of Viscosity on Eupyrene Sperm and Their Flagellar Movements Turritella communis has internal fertilization; therefore, their spermatozoa will have to swim in fairly viscous medium in situ. To investigate the meaning of pairing of spermatozoa and the change in the movements of spermatozoa and their flagella in viscous medium, the sperm and their flagellar movements were examined in viscous seawaters of 110, 230, 500, and 3,000 cP. The swimming speed was found to be inversely proportional to the one-third power of the viscosity (Fig. 5). No significant difference in the swimming speed could be detected between paired and unpaired spermatozoa in seawater or in the four viscous seawaters. Typical flagellar waves of unpaired spermatozoa in viscous seawaters are shown in Figure 6. It is well known that two individual flagella beat synchronously with one another even though they do not have physical contact with one another, as spirochetes also do [Gray, 1928; Taylor, 1952]. To test the synchronization in the paired eupyrene Turritella spermatozoa, the flagellar movements were examined in seawater and the four viscous seawaters. In natural seawater, the two spermatozoa in a pair beat with different phases (Fig. 7A), whereas in the seawater with viscosities over 110 cP, 92 Ishijima et al. swimming in natural seawater. Some fragments moved proximally (Fig. 9A), but most moved distally (Fig. 9B). The movements of flagellar fragments were less active than that of whole flagella. DISCUSSION Bidirectional Swimming of Eupyrene Spermatozoa Fig. 8. Phase-contrast video micrographs of the flagellar movement of apyrene spermatozoa. A spermatozoon swims in natural seawater (A) and in viscous seawater of 3,000 cP (B). Bar ⫽ 30 µm. paired spermatozoa were aggregated and beat in union (Fig. 7B). Movements of Apyrene Spermatozoa It has been suggested that the apyrene spermatozoa may aid in the transport of the eupyrene spermatozoa due to their greater motility [Friedländer and Gitay, 1972]. Apyrene spermatozoa from Turritella were hence recorded with respect to their movement in natural seawater or the seawaters with different viscosities. An apyrene Turritella spermatozoon consists of a head (without a nucleus) and eight or more flagella (8.6 ⫾ 1.7, mean ⫾ S.D., n ⫽ 12). One of these flagella is longer (0.23 ⫾ 0.11 mm, n ⫽ 11) than the others (0.04 ⫾ 0.02 mm, n ⫽ 11). The average swimming speed of the apyrene spermatozoa was 19.3 µm/s (S.D. ⫽ 12.0, n ⫽ 18) in natural seawater and 11.2 µm/s (S.D. ⫽ 4.1, n ⫽ 20) in the 3,000 cP seawater. The swimming speed of the apyrene spermatozoa in natural seawater was rather slower than that of the eupyrene spermatozoa but was almost the same as that in the 3,000 cP seawater. Many flagella beat independently in natural seawater (Fig. 8A), whereas they beat in union in the 3,000-cP seawater (Fig. 8B). Movements of Distal Tail Fragments of Eupyrene Spermatozoa An unexpected finding in this study was that of the distal tail fragments of the eupyrene sperm flagella Turritella spermatozoa can move forward and backward. The bidirectional swimming seems to be a general characteristic of the snail spermatozoa because the eupyrene spermatozoa of the mud creeper (Terebralia palustris) and the black snail (Semisulcospira bensoni) also showed bidirectional swimming (Ishijima and Furukawa, unpublished observations). There are basic differences in the movement parameters between waves propagating toward the tip of the flagella and those moving toward the base, as shown in Table I and Figure 1. The distally propagating waves were more effective than the proximally propagating ones: the swimming speed of spermatozoa with distally propagating waves was more than 12 times that of spermatozoa with proximally propagating waves, and the distally propagating waves generally preceded the proximally propagating ones in the time course of the flagellar movement of eupyrene spermatozoa, even though the reverse transition occasionally happened. Thus, it is unlikely that backward swimming in Turritella spermatozoa plays an important role. The presence of calcium ion increased the percentage of spermatozoa with the proximally propagating waves and seemed to inhibit the distally propagating waves; thereby, the proximally propagating waves were able to appear on the sperm flagella. This interpretation seems to be supported by several findings. Caffeine or 8-bromoadenosine 3’:5’-cyclic monophosphate, both of which activate the sperm motility [Mann and Lutwak-Mann, 1981], prolonged the duration of flagellar movement of distally propagating waves and restored the active movement of sperm flagella that had proximally propagated bending waves. Only at unusually high concentrations of calcium did all the spermatozoa exhibit the proximally propagating waves. Turritella spermatozoa can move backward as well as forward like some other spermatozoa and many protozoa [Holwill and McGregor, 1976; Baccetti et al., 1989; Ishijima et al., 1994]. Crithidia cells and tephritid and Myzostomum spermatozoa, in which the backward swimming is usual, abruptly change their swimming direction, but such a rapid change in swimming direction could not be detected in Turritella spermatozoa, although spermatozoa with proximally propagating waves restored the flagellar waves propagating distally within a short time. Thus, the role of a change in the swimming direction of Turritella spermatozoa seems to differ from that of these cells. Movement of Turritella Spermatozoa 93 Fig. 9. Phase-contrast video micrographs of distal fragments of unpaired eupyrene spermatozoa. A fragment that moves toward the base of the flagellum (A) and one toward the tip (B). The swimming speed of the flagellar fragment (with a total length of 51.4 µm) was 6.6 µm/s in A and that of the flagellar fragment (the total length of 67.5 µm) was 3.9 µm/s in B. Amplitude of the flagellar fragment in B was 2.5 µm; beat frequency, 4.6 Hz; and wavelength, 25.2 µm. The time interval between successive images is 2 s. Bar ⫽ 30 µm. Eupyrene Turritella spermatozoa swam almost in circles when moving freely near a coverslip, as sea urchin spermatozoa also do [Ishijima and Hamaguchi, 1992]. However, the circular rotation of Turritella spermatozoa is due to the bending of the beating plane of flagellar waves (Fig. 3), rather than to an asymmetry of the flagellar waves (a difference in curvature between the principal and reverse bends) as seen in other spermatozoa, including those of sea urchins and mammals. Because the beating plane of Turritella spermatozoa is in the direction perpendicular to the line connecting the central microtubules [Afzelius and Dallai, 1983] as the others do, the bending of the beating plane in the Turritella spermatozoa is perhaps caused by a mechanism that is different from that of other spermatozoa. The bending of the beating plane has also been described for the spermatozoa of the humpbacked fly [Curtis and Benner, 1991]. Direction of Wave Propagation and Chirality of Flagellar Waves The eupyrene Turritella spermatozoa reversed the direction of their swimming as a result of a change in the direction of bend propagation, this direction apparently being regulated by calcium ions. When the spermatozoa were viewed from head to tail, a reversal of direction of the bend propagation accompanied the reversal of the sperm rotating around their long axis. Baccetti et al. [1989] reported the same findings in spermatozoa of tephritid flies; thus their flagellar waves, which propagate from the tip of the flagellum toward the head, induce the 94 Ishijima et al. sperm to rotate clockwise with respect to the advancement direction, whereas flagellar waves that propagate from the head toward the tail tip induce the sperm to rotate clockwise with respect to the advancement direction as well. These findings in the spermatozoa of both tephritid flies and Turritella snails suggest that the sense of the three-dimensional flagellar waves of these spermatozoa does not change when the direction of bend propagation changes (Fig. 10). This response to calcium ions differs from that of sea urchin spermatozoa; these do not change the propagation direction along the flagella by altering the calcium ion concentration, but they change the sense of flagellar waves [Ishijima and Hamaguchi, 1993]. These differences in the response to calcium ions between sea urchin and Turritella spermatozoa can be explained by a different behavior of active sliding between the doublet microtubules. The rotational movement of spermatozoa around their long axis is due to the three-dimensional movement of their flagella, and the direction of sperm rotation is determined by the sense of their three-dimensional geometry [Gray, 1962; Ishijima and Hamaguchi, 1993]. The three-dimensional flagellar waves are generated by localized active sliding between doublet microtubules transmitting around the flagellar circumference as well as propagating along the flagellum [Ishijima and Hamaguchi, 1993]. The two propagation directions, around the flagellum and along it, in response to Ca2⫹ differ between sea urchin and Turritella spermatozoa. In sea urchin spermatozoa, transmitting the localized active sliding between doublet microtubules around the flagellum in an alternative direction changes the chirality of the three-dimensional bending waves. On the other hand, in Turritella spermatozoa, the propagation of a localized sliding along the flagella will change the direction of propagation of the bending waves. In sea urchin spermatozoa, changing the propagation direction around the flagella seems to be easier than that along the flagella, whereas in Turritella spermatozoa, the reverse is found, possibly because of mechanical conditions at the distal region of the sperm flagellum. The doublet microtubules may, for instance, attach to the plasma membrane [Afzelius and Dallai, 1983], maybe as an adaptation to enable the spermatozoa to change the direction of bend propagation along rather than around the flagellum [Machin, 1958; Kaneda, 1965]. The finding that distal fragments of the eupyrene sperm flagella could swim both forward and backward may support this idea because such fragments usually cannot move in other spermatozoa [Okuno and Hiramoto, 1976; Woolley and Bozkurt, 1995]. The important feature in the flagellar movement of Turritella spermatozoa in which the distal end of the flagella hardly showed transverse movement also implies that active sliding between the doublet microtubules is inhibited at the distal end of the Turritella sperm flagella. Fig. 10. Hypothetical diagram explaining the rotational direction of the spermatozoa and the propagation direction of their bending waves. Two flagella (open and dotted) with a phase difference of about a quarter cycle are shown. Distally propagated bending waves of a left-handed helix rotate a spermatozoon counterclockwise (curved arrow) when viewed from the head because the direction of rotation of the spermatozoon is opposite to that (arrow) of the angular movement of segments of the flagellum (A), whereas proximally propagated bending waves of the same sense rotate a spermatozoon clockwise (B). The flagellar waves are tentatively illustrated as flattened helices because there is no detailed information on it. Conjugated Spermatozoa There was little difference in the movement characteristics between the paired and the unpaired eupyrene Turritella spermatozoa in either normal or viscous seawater (Table I and Fig. 5). This finding differs from the previous reports on cooperative swimming in the American opossum in which two spermatozoa are more forceful in their swimming than are single spermatozoa [Biggers and Creed, 1962; Moore and Taggart, 1995]. This may be due to a difference in the orientation of the two conjugated spermatozoa in Turritella and in the opossum. In the Turritella spermatozoa, the two spermatozoa in a pair face the same direction [Afzelius and Dallai, 1983], whereas in the American opossum the two spermatozoa face each other [Temple-Smith and Bedford, 1980; Taggart, et al., 1993]. A difference in the sperm alignment is also observed in the beating pattern of paired spermatozoa; the paired Turritella spermatozoa beat with a phase difference that is less than a quarter cycle, whereas the paired opossum spermatozoa beat with a phase difference of approximately half a cycle [compare Fig. 7 in the present paper with fig. 12 in Phillips, 1972, or fig. 3 in Movement of Turritella Spermatozoa Moore and Taggart, 1995]. The difference in orientation of the two spermatozoa may induce different responses to the viscosity; the paired Turritella spermatozoa beat in union in highly viscous seawaters, whereas the paired opossum spermatozoa beat independently in a medium of approximately the same viscosity [Moore and Taggart, 1995]. The two flagella of paired Turritella spermatozoa did not beat in phase in natural seawater, in contrast to the situation of closely adjacent sea urchin spermatozoa. This is probably due to the accessory fibers of Turritella spermatozoa. In fact, spermatozoa of the American opossum also did not beat in phase [Phillips, 1972; Moore and Taggart, 1995] and neither do adjacent mammalian spermatozoa [Ishijima, 1992], although large-scale wavy motions can be seen in a dense sperm suspension of bull or ram spermatozoa [Rothschild, 1949; Mann and LutwakMann, 1981; Ishijima, 1992]. ACKNOWLEDGMENTS We thank the director, Prof. J.-O. Strömberg, and the staff of the Kristineberg Marine Biological Station, Fiskebäckskil, Sweden, where this work was carried out, for their generous hospitality. We also thank Dr. Y. Hamaguchi for his active interest and advice. REFERENCES Afzelius BA. 1983. The spermatozoon of Myzostomum cirriferum (Annelida, Myzostomida). J Ultrastruct Res 83:58–68. Afzelius BA, Dallai R. 1983. The paired spermatozoa of the marine snail, Turritella communis Lamarck (Mollusca, Mesogastropoda). J Ultrastruct Res 85:311–319. Afzelius BA, Dallai R. 1987. Conjugated spermatozoa. In: Mohri H, editor. New horizons in sperm cell research. Tokyo: Japan Scientific Societies Press/New York: Gordon and Breach Science Publishers, p 349–355. Baccetti B, Gibbons BH, Gibbons IR. 1989. Bidirectional swimming in spermatozoa of Tephritid flies. J Submicrosc Cytol Pathol 21:619–625. Biggers JD, Creed RFS. 1962. Conjugate spermatozoa of the North American opossum. Nature 196:1112–1113. Biggers JD, DeLamater ED. 1965. Marsupial spermatozoa pairing in the epididymis of American forms. Nature 208:402–404. Braidotti P, Ferraguti M. 1982. Two sperm types in the spermatozeugmata of Tubifex tubifex (Annelida, Oligochaeta). J Morphol 171:123–136. Curtis SK, Benner DB. 1991. Movement of spermatozoa of Megaselia scalaris (Diptera:Brachycera:Cyclorrhapha:Phoridae) in artificial and natural fluids. J Morphol 210:85–99. Dallai R, Afzelius BA. 1984. Paired spermatozoa in Thermobia (Insecta, Thysanura). J Ultrastruct Res 86:67–74. Dallai R, Afzelius BA. 1985. Membrane specializations in the paired spermatozoa of dytiscid water beetles. Tissue Cell 17:561–572. Dallai R, Afzelius BA. 1987. Sperm ultrastructure in the water beetles (Insecta, Coleoptera). Boll Zool 4:301–306. Friedländer M, Gitay H. 1972. The fate of the normal-anucleated spermatozoa in inseminated females of the silk worm Bombyx mori. J Morphol 138:121–130. Gray J. 1928. Ciliary movement. London: Cambridge University Press. 95 Gray J. 1962. Flagellar propulsion. In Bishop DW, editor. Spermatozoan motility. Washington, DC: American Association for the Advancement of Science. p 1–12. Holwill MEJ, McGregor JL. 1976. Effects of calcium on flagellar movement in the trypanosome Crithidia oncopelti. J Exp Biol 65:229–242. Ishida S, Yamashita Y, Teshirogi W. 1991. Analytical studies of the ultrastructure and movement of the spermatozoa of freshwater triclads. Hydrobiologia 227:95–104. Ishijima S. 1992. Methods for analyzing sperm motility. In Mohri H, Morisawa M, Hoshi M, editors. Spermatology. Tokyo: University of Tokyo Press. p 212–224 (in Japanese). Ishijima S, Hamaguchi Y. 1992. Relationship between direction of rolling and yawing of golden hamster and sea urchin spermatozoa. Cell Struct Funct 17:319–323. Ishijima S, Hamaguchi Y. 1993. Calcium ion regulation of chirality of beating flagellum of reactivated sea urchin spermatozoa. Biophys J 65:1445–1448. Ishijima S, Oshio S, Mohri H. 1986. Flagellar movement of human spermatozoa. Gamete Res 13:185–197. Ishijima S, Hamaguchi MS, Naruse M, Ishijima SA, Hamaguchi Y. 1992. Rotational movement of a spermatozoon around its long axis. J Exp Biol 163:15–31. Ishijima S, Ishijima SA, Afzelius B. 1994. Movement of Myzostomum spermatozoa: calcium ion regulation of swimming direction. Cell Motil. Cytoskeleton 28:135–142. Kaneda Y. 1965. Movement of sperm tail of frog. J Fac Sci Univ Tokyo, Sec.IV, 10:427–440. Kennedy JJ, Keegan BF. 1992. The encapsular developmental sequence of the mesogastropod Turritella communis (Gastropoda: Triitellidae). J Mar Biol Ass UK 72:783–805. Koike K, Nishiwaki S. 1980. The ultrastructure of dimorphic spermatozoa in two species of the Strombidae (Gastropoda: Prosobrachia). Venus 38:259–275. Machin KE. 1958. Wave propagation along flagella. J Exp Biol 35:796–806. Mann T, Lutwak-Mann C. 1981. Male reproductive function and semen. Berlin: Springer-Verlag. Moore HDM, Taggart DA. 1995. Sperm pairing in the opossum increases the efficiency of sperm movement in a viscous environment. Biol Reprod 52:947–953. Okuno M, Hiramoto Y. 1976. Mechanical stimulation of starfish sperm flagella. J Exp Biol 65:401–413. Phillips DM. 1972. Comparative analysis of mammalian sperm motility. J Cell Biol 53:561–573. Reger JF, Fitzgerald ME. 1979. The fine structure of membrane complexes in spermatozoa of the millipede, Spirobolus sp., as seen by thin-section and freeze-fracture techniques. J Ultrastruct Res 67:95–108. Rothschild Lord 1949. Measurement of sperm activity before artificial insemination. Nature 163:358. Taggart DA, Johnson JL, O’Brien HP, Moore HDM. 1993. Why do spermatozoa of American marsupials form pairs? A clue from the analysis of sperm-pairing in the epididymis of the grey short-tailed opossum, Monodelphis domestica. Anat Rec 236:465–478. Taylor G. 1952. The action of waving cylindrical tails in propelling microscopic organisms. Proc R Soc London Ser A 211:225–239. Temple-Smith PD, Bedford JM. 1980. Sperm maturation and the formation of sperm pairs in the epididymis of the opossum, Didelphis virginiana. J Exptl Zool 214:161–171. Wilson EB. 1925. The cell in development and heredity. New York: MacMillan Inc. Woolley DM, Bozkurt HH. 1995. The distal sperm flagellum: its potential for motility after separation from the basal structures. J Exp Biol 198:1469–1481.