Microtubules and the Propagation of Bending Waves by the Archigregarine, Selenidium Fallax

Archigregarines of the genus Selenidium are found in the digestive tracts of many species of polychaete worms. The vegetative stage or trophozoite is itself vermiform and normally anchored by its anterior end to the intestinal epithelium of the host. The remainder of the trophozoite is thus free to move within the gut cavity.

When Giard (1884) named the archigregarine in Nerine cirratulus as Selenidium pendula he not only founded the genus but also set the trend for describing the movements characteristic of archigregarine trophozoites; ‘le nom rapelie les mouvements pendulaires caracteristiques de tout le group’. For nearly a century there has been little progress in furthering this essentially correct but oversimplified view of trophozoite motility, except perhaps in recognizing the organism’s ability to coil and uncoil like a watchspring (Ray, 1930; Vivier & Schrevel, 1964; Macgregor & Thomasson, 1965). Possibly Stebbings, Boe & Garlick (1974) came closest to recognizing the true potential of trophozoite movement by describing it as ‘whip-like’, and by noting the great degree of variation in amplitude of the beat form from one trophozoite to another.

Several authors have carried out ultrastructural studies describing a system of longitudinally aligned microtubules beneath the pellicle of the trophozoites (Vivier & Schrevel, 1964; Macgregor & Thomasson, 1965 ; Schrevel, 1971 ; Stebbings et al. 1974) and these constitute the only structures identified which could play a mechanistic role in the motility of the organism. So far, however, there has been no detailed analysis of the movement itself -a fundamental requirement if the function of the microtubules is to be accurately assessed - and this we present here.

It has become obvious from our cinematographic studies that in the trophozoites of the Selenidium sp. native to the intestine of Cirriformia tentaculata, the pendular movement is achieved by the anterior formation of bends and their propagation along the length of the trophozoite in a manner similar to that shown by beating cilia and sperm flagella. Waveforms of the latter have proved to be highly amenable to analysis by analogy with either travelling sine waves (Gray, 1955), travelling circular arcs and interconnecting straight lines with abrupt transitions between them (Brokaw & Wright, 1963; Brokaw, 1965; Goldstein, 1976) or meander-like waves (Rikmenspoel, 1971; Silvester & Holwill, 1972). More recently too, Rikmenspoel (1978) and Hira-moto & Baba (1978) have demonstrated that the changes in angular direction of beating sperm flagella can be accurately expressed as a sine function of time plus a constant, i.e. the flagella waveforms are ‘sine-generated’. This type of wave gives good approximations to both meander curves (Langbein & Leopold, 1966; Leopold & -Langbein, 1966) and arc-line curves (Sarashina, 1974). Although we have made use of the latter as an initial means of analysing the movement of trophozoites, we discuss the propriety of other wave analogues.

Cirriformia tentaculata, the polychaete host, was collected from the shore at Ladram Bay, Devon, cleaned, and maintained in aerated seawater at 10 °C. Providing injured worms were discarded, both the host and its archigregarine population could be kept for 10–14 days without any visible deterioration in their condition.

Small pieces of heavily infected intestine were excised by a longitudinal incision under seawater. Opening the gut often caused the release of unattached trophozoites and these were also collected, for observation, using Pasteur pipettes. Free trophozoites, or pieces of intestine with trophozoites attached, were mounted in seawater on a slide and examined using a Wild M20 microscope and bright field illumination. Cine film of six beating trophozoites was taken at 10 f.p.s. using a Vinten body and magazine mounted on the microscope. Ilford 16 mm Pan F was used and a Xenon arc lamp provided the illumination. Ultra-violet and infra-red filters were used in conjunction with the Xenon arc lamp at all times.

Other pieces of intestine were placed in small glass embryo cups with a plentiful covering of seawater and the trophozoites observed using a Vickers stereo binocular microscope. Approximately 100–300 trophozoites attached to a single piece of intestine could be examined by this method. The activity of the trophozoites was monitored at 30-min intervals by taking a sample of any six organisms and noting their beat frequency. The number of beats performed over a 30 s period was counted to the nearest quarter beat for each individual. All experiments were conducted at room temperature.

In order to analyse the film obtained, prints were made, to a final magnification of approximately 500 times, from individual frames selected at 0·5 or 1 s intervals. This was carried out over a 15–40 s period so that at least one full beat-cycle was obtained for analysis. A modification (Goldstein, 1976) of the arc-line method of analysis pioneered by Brokaw and co-workers (1963, 1965) was used to follow the development and propagation of bends (see Fig. 1).

Trophozoites attached to the gut and placed in embryo cups containing seawater could continue beating for up to 12 h, although 95% of any sample population had normally ceased to do so approximately 6 h after removal from the host. An initial beat frequency of 14–18 beats min⁻¹ (0·23–0·3 Hz) was occasionally recorded, but a rate half this was typically attained within the first hour. Thereafter, both the rate of movement and percentage motility declined. Many trophozoites beat in approximately one plane (Figs. 2, 3) but deviations from planarity can occur as a result of rotation and twisting at the anterior attachment end.

The apparent suitability of the arc-line model as a method of analysis for the planar waveforms is demonstrated by Fig. 4. Experimental values of 0·7–0·9 ( + 0·05) were commonly obtained for the ratio of the chord length at half amplitude to that at half wavelength for fully formed bends of different organisms. Since sinusoidal waves must, by definition, give a value of 0·667 for the same ratio (Brokaw & Wright, 1963), the waves of the trophozoites would not appear to be of this nature. Close inspection of the waveforms with the unaided eye, however, suggests a degree of continuity between the bent regions and, as will be discussed later, a meander or sine-generated wave may well be a more appropriate analogue than arc-line curves. In spite of this, developing bends could be located to within 5–10 μm of the anterior tip and followed along 7080% of the trophozoite, the remaining 20–30% being lost mainly at the distal extremity. If the total lengths of the analysed trophozoites were calculated from the model for comparison with their actual lengths, errors of no more than 6% arose.

Trophozoites are streamlined with two tapers, anterior and posterior, separated, except in very small trophozoites, by a mid-zone of constant diameter. They are up to 400 μm long, with a maximum diameter of about 20 μm. The proportion of the total length made up by the mid-zone (40–55 %) and its precise location along the length of the trophozoites varies between organisms (compare Figs. 2 b and 3 c). This midzone seems particularly important with regard to bend formation and propagation. Developing bends propagate slowly during formation and are two-thirds or fully formed on arrival at the beginning of the mid-zone, whereupon rapid propagation commences (Figs. 2 a and 5). Similar two-stage propagation rates have also been reported for sperm flagella of the sea-urchins Psammechinus miliaris (Gray, 1955), Lytechinus pictus (Brokaw, 1970) and Stronglyocentrotus purpuratus (Goldstein, 1977). In the trophozoites, the nucleus is normally positioned one-quarter to one-third of the way along the body from the anterior end and, as is illustrated by Figs. 2a, 36 and 5, does not constitute an obstacle to propagation, even though it occupies most of the diameter of the cell. When the posterior taper is reached, ‘unbending’ commences as shown by Figs. 2b and 6a. The propagation rate remains constant, as a rule, but the bends usually become indiscernible before they reach the posterior tip. Rapid propagation is 2·5–6·7 times faster than that during bend formation, and rates of 8·7–35 μms⁻¹ have been recorded for beat frequencies ranging from 2·4–10 beats min⁻¹ (0·04– 117 Hz). The correlation between the beginning and end of the mid-zone, and the onset of rapid propagation and unbending respectively could be pinpointed to within 10 μm.

In healthy organisms, consecutive bends of opposite sign are symmetrical, increasing to similar final angles at similar rates (Fig. 6a). As a result, a developing bend appears to be coordinated with the preceding bend and, later in the cycle, with the following bend. Instantaneous symmetry, however, may only be visible during parts of the cycle, an aspect which will ultimately depend upon the ratio of the mid-zone length to the wavelength measured along the trophozoite. Since the latter varies for different organisms between 160 and 300 μm (80–160 μm along the x-axis) the mid-zone : wavelength ratio is typically < 1, though exceptions do occur (Fig. 3 b). Depending upon the rate of bend formation and the final angle subtended by developing bends, the entire trophozoite oscillates with an angular velocity often approaching 0·25 rad s⁻¹.

The importance of the mid-zone to the constancy of the bend characteristics (Fig. 6) is also emphasized by the pattern of beating demonstrated by very small trophozoites, about 40 μm in length. These organisms are composed of an anterior and posterior taper only, their maximum diameter being at the junction of the two. Motility does, in this case, involve very simple pendular movements -bends of opposite sign are developed consecutively at the anterior end but propagation is very limited.

Coiling and uncoiling movements are sometimes exhibited by attached trophozoites but are more typical of free organisms (Fig. 3 a). The coiling motion is often accompanied by a helical twisting, larger trophozoites being capable of producing two to three full turns. Although coiling undoubtedly has a propagative basis it differs from the wave motion in that it is highly asymmetric, unless, as occasionally occurs, the entire protozoan rolls through 180° during uncoiling to give the appearance of symmetry. Careful measurement has shown that a length difference between the inner and outer curvature of over 100 μm may be generated during coiling. In trophozoites propagating waves, lesser differences in the region of 20 μm may occur, as in Fig. 2b where the organism maintained one bend only at times o, 12·0 and 24·0 s. Overall length changes of the trophozoites, however, were not observed. Stationary coils or bends may be held in position distally, while other bends continue to propagate normally from the anterior end. On approaching the non-propagating region, the travelling bends gradually increase in radius of curvature and decline in amplitude until they disappear completely.

The slow decline in beat frequency apparent under the experimental conditions used was accompanied by a gradual decrease in the extent to which bends were propagated along the trophozoites. This was as a result of fully formed bends no longer maintaining a constant amplitude along the length of the trophozoite from the onset of rapid propagation. The point at which the bends had completely decayed shifted slowly with time towards the attachment end, until the organisms were beating purely by the formation of non-propagating bends in the same fashion as very small trophozoites. Complete cessation of motility followed shortly after this stage. The shape of the protozoans also tended to change during this process, with the trophozoites becoming markedly shorter and broader as the decay continued.

The constancy of the bend amplitude along the mid-zone of the trophozoites suggests that they possess an energy source, and an active mechanism capable of utilizing it, along at least this entire region, since such waveforms cannot be duplicated by a system powered solely at one end (Gray, 1955; Machin, 1958). The fact that bends propagate through the nuclear region of the trophozoite without any loss of amplitude, angle, etc., is indicative of a peripheral location for the machinery of bend formation and propagation. Trophozoites contain up to 4000 peripheral longitudinal microtubules (Mellor & Stebbings, unpublished data) ordered into 2–3 rows parallel to the outer membrane complex of the cell, and the reversible depolymerization of these microtubules using 0–6 M urea has been demonstrated by Schrevel, Buissonnet & Metais (1974) with the trophozoites of S. hollandei. Cessation and recovery of motility was closely associated with the loss and repolymerization, respectively, of the microtubule complement, the trophozoites becoming linear in their absence.

The sliding microtubule mechanism by which the 9 + 2 axoneme produces ciliary motion (Satir, 1965, 1967, 1968) has also been adopted for flagella (Brokaw, 1971, 1972) and is now well supported both experimentally and theoretically (Summers & Gibbons, 1971, 1973; Gibbons & Gibbons, 1973, 1974; Warner & Satir, 1975; Sale & Satir, 1977; Hiramoto & Baba, 1978; Holwill, Cohen & Satir, 1979). That microtubule sliding to produce propagated bends is not a unique property of the 9 + 2 array has already been clarified by McIntosh, Ogata & Landis (1973) and McIntosh (1973) with the protozoan Saccinobaculus, an intestinal symbiont of the woodfeeding roach Cryptocercus punctulatus. In Saccinobaculus, the crystal-like lattice of several thousand closely packed microtubules forming the protozoan axostyle has been shown to propagate bends, at rates of up to 100 μm s⁻¹, with the aid of many of the same proteins (e.g. dynein, nexin) which have been found to be so important in the ciliary axoneme (Mooseker & Tilney, 1973; Bloodgood, 1975). Consequently, with the implication that microtubules are involved, it seems possible that a comparable sliding mechanism may be responsible for the trophozoite wave motion. In the trohozoites of S. fallax, the microtubules are separated by a centre to centre distance to approximately 40 nm, within and between the rows (Stebbings et al. 1974), a spacing which is not inconsistent with the proposal that microtubule interactions produce motility in this organism. However, the sites of mechanochemical transduction, the precise lengths of individual microtubules and the rearrangements they undergo with bending have yet to be thoroughly investigated.

The propagation of bends of one sign only (Figs. 3 c and 7) suggests the existence of two similar systems responsible for developing and transmitting bends of opposite sign. That consecutive bends in healthy organisms are symmetrical, developing and propagating at similar rates, on the other hand, implies that the two systems are normally highly coordinated. This is indeed reminiscent of the regulated activation of doublet sliding (e.g. Rikmenspoel, 1971) and coordinated production of shear resistance (Warner & Satir, 1974) occurring in the 9 + 2 axoneme, since, for instance, extracted reactivated sperm flagella of the sea-urchin L. pictus demonstrate a similar phenomenon (Goldstein, 1976).

In our analysis of the movement of trophozoites we have used the arc-line model because it provides a convenient means of demonstrating the propagation of bending waves in a system hitherto unstudied. Since the trophozoite comprises an entire cell and constituents, it seems likely that such a system would conform to the elastic rod envisaged by Silvester & Holwill (1972) in their discussion of the meander waveform. As Macgregor & Thomasson (1965) have indicated -‘Nothing in S. fallax…other than the folded multilayered pellicle is capable of acting as a stiff skeletal component, and the only structures to which we can attribute the property of “contractility” are the fibrils which lie beneath’. A completely objective analysis of the waveforms produced by the trophozoites is, however, beyond the realms of this article, and so sine-generated or indeed imperfect arc-line waves (Johnston, Silvester & Holwill, 1979) cannot yet be dismissed as suitable analogues.

As would be expected of an organism the size of Selenidium, a low value of the Reynolds number for the motion was obtained. Of the two external forces which could potentially oppose the beating trophozoites, the magnitude of the viscous resistance of the surrounding fluid is much greater than that of its inertial resistance, and so the latter may be neglected in comparison. Although further work is required in order to elucidate the relative importance of the external viscous resistance, the internal elastic and internal viscous resistance as the major forces opposing the formation and propagation of bends by archigregarines, the protozoan provides a novel opportunity to study the energetics of the movement of an entire cell by similar means to those already employed for cilia and flagella.

How do the movements performed by the trophozoites in seawater relate to those occurring in situ? When the host intestine was opened, it contained either a fairly solid matrix of sand, organic detritus and mucus, or a dark brown fluid which appeared to be of a higher viscosity than seawater. Clearly, seawater alone is quite a different environment to that which the trophozoites normally inhabit. The presence of a close packed matrix in the intestine during certain stages of the polychaete’s digestion probably does not allow the beating and coiling motions exhibited when trophozoites are placed in seawater. The initial instability of the beat frequency after removal from the intestine, the degeneration of the beat cycle, and the shortening the trophozoites all imply that seawater alone is not their optimum environment, a certain amount of difficulty was experienced in obtaining film of any one trophozoite demonstrating all the salient features of the beat cycle in perfect form, and the above factors seem largely responsible for this. The near century-long delay in the true characterization of trophozoite motility is probably attributable to these same factors.

The most obvious purpose of movement within the gut cavity must surely be that of maintaining a constant flow of nutrients over the trophozoites. The gut contents are sometimes closely packed and this suggests that the hypothetical nutrient flow is, at best, intermittent. However, the intestinal epithelium is equipped with cilia which are particularly active and well coordinated in the ventral groove of the intestine (Mellor & Hyams, 1978), though the mixing produced may not be sufficient to supply the nutritional needs of the trophozoites. It is conceivable then, that the beating and coiling motions may be induced by placing the trophozoites in seawater.

Nevertheless, the ability of the protozoan to propagate bending waves makes it a fascinating organism for further research on a number of counts. As well as providing a firm base from which to promote our knowledge of archigregarine motility, it provides an opportunity to improve the understanding of microtubule function per se, and, of equal importance, of microtubule interactions in systems where elastic forces external to the microtubules may be important. To quote Warner & Satir (1974) in referring to the importance of radial spokes in cilary axonemes: ‘Unfortunately the differences in bend form and propagation between other microtubule-based systems and true cilia and flagella, including 9 + 2 sperm tails, have not yet been analysed.’ Although a direct involvement of the microtubules in bend formation and propagation by the trophozoites still awaits adequate confirmation, they do indeed appear to possess a system which could be very usefully compared with the 9 + 2 axoneme.

Finally, it would seem probable that a single common mechanism producing both the coiling and wave motions is operative within the trophozoite. Any model which purports to explain this mechanism must also be able to account for the combinations of the two different movements which can occur and also for phenomena such as the coordination of developing bends. Therefore a high degree of versatility must be intrinsic to the model. Careful ultrastructural, manipulative, and waveform studies aimed at clarifying this mechanism are currently in progress.

We gratefully acknowledge Dr M. E. J. Holwill for interesting discussion of the biophysical aspects presented here and for his criticisms of an earlier version of this manuscript.

J. S. Mellor was supported by an S.R.C. postgraduate grant.