Structure and development of the egg of the glossiphoniid leech Theromyzon rude: characterization of developmental stages and structure of the early uncleaved egg

Much of the early and late literature on glossiphoniid leeches indicates that the embryos may be advantageously used for the study of developmental processes (Whitman, 1878; Schleip, 1936; Dawydoff, 1959; Fernández & Olea, 1982). Recent studies in developmental neurobiology have successfully utilized the leech embryo in the exploration of the role of cell lineage and cell-cell interactions in the development of the nerve cord (Weisblat, 1981; Stent & Weisblat, 1982, 1985; Weisblat & Stent, 1982; Weisblat, Kim & Stent, 1984).

Another interesting feature of the leech embryo is related to the possibility of tracing the cytoplasm of many of its cells to specific regions of the uncleaved egg. Thus, the poles of the egg include organelle-rich domains of ooplasm (called teloplasms) destined to be utilized in the formation of ectodermal and mesodermal cells. Hence, the animal and vegetal teloplasms may be repositories of certain morphogenetic determinants. Completion of meiosis and formation of teloplasm are accompanied by prominent deformation movements of the uncleaved egg. Thus, the glossiphoniid leech egg represents a potentially valuable material for the study of the mechanism underlying both mosaic development and motility in nonmuscle cells. The manner of formation of ooplasmic domains in the egg of certain glossiphoniids, such as those of the genera Glossiphonia and Theromyzon, are only comparable to those presented by the egg of the oligochaete Tubifex (see Penners, 1922; Shimizu, 1982a, 1986). It is not surprising then that the study of their eggs attracted the attention of many early investigators of metazoan development such as Whitman (1878). His work deserves special mention because he presented a surprisingly detailed and accurate description of the structure and development of the uncleaved egg of some Glossiphonia. Some preliminary information on the structure and development of the uncleaved egg of the glossiphoniid Theromyzon rude is found in Fernández (1980) and in Fernández & Olea (1982).

This paper is the first of a series intended to analyse the structure and development of the uncleaved egg of T. rude. In this presentation we examine some aspects of the reproductive biology of the leech under laboratory conditions and also provide a description of the structure and of the stages of development of the uncleaved egg.

Adult specimens of the leech Theromyzon rude were collected in ponds of the Golden Gate Park (San Francisco, California) and around Calgary (Canada), during June and July. Animals were found under rocks or hard objects, either alone or in groups. They were placed in large plastic jars containing artificial spring water (see Fernández & Olea, 1982). Breeding was performed in a thermoregulated chamber at 14°C in complete darkness. If breeding needs to be accelerated, a group of animals may be removed and placed at higher temperature. In this manner, the availability of mature eggs was controlled. The characteristic social behaviour of leeches during the breeding season is also seen under laboratory conditions. Thus, solitary and grouped animals are found in the breeding jars. The former are unfed, have recently fed or have already laid eggs. Grouped leeches constitute breeding communities of variable size. Animals enter the community as males, with guts full of blood, and leave as females, with empty guts and highly developed ovisacs. At 14°C, gravid leeches usually leave the breeding community 1–2 days before oviposition.

Eggs are considered ready to be laid when their colour turns brown and when they become clustered within the ovisacs. Each cluster presumably comprises the set of oocytes destined to be housed within a cocoon. Animals lay four to six cocoons per clutch, with each cocoon including 15–30 eggs. Accordingly, one clutch includes 60–180 eggs.

In order to postpone egg laying, gravid females may be continuously agitated in a shaker or their suckers can be rendered useless by piercing them with minute insect pins. The latter procedure was used when mechanical agitation failed to block egg laying. Uncleaved eggs may be obtained from laid cocoons or directly from the ovisacs. Since development of the egg starts as soon as it leaves the ovisacs and takes a couple of hours to complete oviposition, eggs within a single cocoon develop asynchronously when laid naturally (for more details see Fernández, 1980; Fernández & Olea, 1982). Removal of ripe eggs from the ovisacs has the advantage that all of them will develop synchronously. For this purpose, unanaesthetized leeches were cut and pinned open in a Sylgard-filled Petri dish. The paired ovisacs were removed and opened in artificial spring water, which is a convenient culture medium for uncleaved eggs.

Development of the uncleaved egg may be studied at temperatures ranging from about 8 to 25 °C. Development seems to be reversibly blocked at approximately 3°C. Thus, eggs may be stored at that temperature for a couple of hours if needed. A total of approximately 2000 eggs, coming from 50 pregnant leeches, was used in this study.

Development of the uncleaved egg may be conveniently studied by combining the observation of live with fixed/cleared eggs. Eggs were fixed in ALFAC (95 % ethyl alcohol, 85ml; 40% formaldehyde, 10ml; glacial acetic acid, 5 ml) for 12–24 h at room temperature. After dehydration in absolute ethanol (two changes of 1 h each), eggs were cleared and stored in methyl benzoate. Live and cleared unmounted whole eggs were examined against a dark background using reflected cold light.

Eggs were fixed 1·5–2 h in 50% Kamovsky solution (see Fernández & Stent, 1980) or in 3 % glutaraldehyde in 0·1 M-phosphate buffer pH7·4, containing 0·15 % of tannic acid. The latter fixation, performed at room temperature, gave reasonably good preservation of microtubules and micro-filaments. After rinsing in buffer at room temperature, eggs were postfixed in 1 % OsO₄ for 1 h at room temperature in the dark, dehydrated in graded ethanol and slowly embedded under vacuum in Epon 812. For light microscopy, 1 μm sections of the embedded egg were stained with 1 % toluidine blue in 1 % sodium borate. For electron microscopy, thin sections of the embedded egg were placed on collodion-coated copper grids and double stained with alcoholic uranyl acetate and lead citrate. Sections were viewed in a Philips EM 300 electron microscope.

Eggs were fixed in 2·5 % glutaraldehyde in 0·1 M-cacodylate buffer pH 7·4 for 2h at 4°C. After rinsing in the same buffer, eggs were dechorionated (removal of the perivitelline envelope) with fine tweezers and postfixed in 1 % OsO₄ at room temperature and darkness for 2–3 h. Tissues dehydrated in graded acetone were critical-point dried from CO₂, mounted and then coated with a layer of gold approximately 30nm thick. For this purpose, a Polaron E5000 sputter apparatus was used. Samples were examined in a Philips EM300 electron microscope, equipped with a scanning device.

To explore the organization of the ectoplasmic cytoskeleton, eggs were treated with extraction buffer containing the nonionic detergent Triton X-100. Successful preparations of whole-mounted cytoskeletons of early uncleaved eggs were obtained with the extraction buffer designed by Schliwa (1980, see also Schliwa & van Blerkom, 1981). This buffer includes Mg²⁺, Pipes, Hepes, EGTA and 0·15% Triton X-100 at pH 7. Dechorionated eggs were briefly washed in extraction buffer without detergent and then treated for 5–10 min in the same buffer with detergent at room temperature. After a brief rinse in buffer without detergent, eggs were fixed in 3 % glutaraldehyde in 0·1 M-cacodylate buffer for Ih at room temperature. After rinsing in cacodylate buffer for Ih, eggs were postfixed in 1 % OsO₄ for another hour at room temperature. Tissues were then dehydrated in graded acetone, critical-point dried from CO₂ and covered with gold.

Eggs treated with the Schliwa extraction buffer were briefly rinsed and then fixed in 4 % paraformaldehyde in 0·1 M-phosphate-buffered saline (PBS) pH 7·4, for 15 min at room temperature. After rinsing in PBS another 15 min, permeabilized eggs were incubated in a solution of approximately 1 μg ml⁻¹ of RLP (kindly provided by Dr Schliwa) in PBS for 20 min at room temperature. Eggs were again rinsed in PBS for about 30 min and finally mounted under coverslip in glycerol-PBS (9:1) containing 0· 1 M-n-propyl gallate. To avoid compression of eggs, the comers of the coverslip were provided with small plasticine stoppers. The samples were examined in a Zeiss microscope equipped with epifluorescence illumination and utilizing the rhodamine excitation filter.

To determine the distribution of actin filaments across the endoplasm and their connections with the ectoplasmic actin network, stained eggs were either fractured or slightly compressed under the coverslip.

To prepare isolated egg cortices, dechorionated eggs were placed on coverslips coated with 0· 1% L-polysine (Sigma). After breaking the egg with tweezers, jets of extraction buffer with detergent were directed against the preparation. The extraction continued for 5– 10 min. After buffer rinsing, fixation and staining with RLP, egg peels were similarly mounted under coverslips without stoppers and viewed under epifluorescence optics.

Observations on the behaviour of T. rude breeding in the laboratory, and also in the field, indicate that copulation probably takes place during the period of several weeks in which animals remain together in breeding comtnunities. Copulation in T. rude is known to last for several days (Wilkialis & Davies,1980). Since at 14°C animals leave the community a few days before egg laying, the final steps of egg maturation and perhaps fertilization take place in solitary animals. Data presented under Materials and methods strongly suggest that fertilization in T. rude occurs in the ovisacs. This conclusion can also be extended to other glossiphoniids because fertilized eggs may be obtained directly from the ovisacs of several different species used in our laboratory (unpublished observations). Sometimes eggs removed from the ovisacs appear to lack a fertilization membrane, the perivitelline chamber is absent and development fails to start. These features probably indicate that the eggs were not fertilized and there-fore the cortical reaction has not yet taken place. This result strongly suggests that fertilization may occur a few hours before the initiation of egg laying.

When pregnant leeches are subjected to continuous agitation in a shaker or their suckers are immobilized, egg laying may be postponed for hours or days. Egg laying may also be interrupted and resumed later if during delivery the mother is disturbed. Examination of freshly laid eggs and of eggs that have remained in the ovisacs for varying lengths of time reveals that development of their meiotic nuclei has not progressed beyond metaphase I. Thus, it is clear that intraovarian development of the fertilized egg is subjected to meiotic arrest.

Stage 1 of development has been subdivided into six substages designated la-lf. Each stage is defined as that moment in which the process diagnostic for that stage has been completed. A diagrammatic representation of the development of the uncleaved egg is shown in Fig. 1. The following characterization of the developmental stages of the uncleaved egg is based on observations of both live (Fig. 2) and fixed/ cleared (Fig. 3) eggs developing at 20°C.

Fertilized eggs removed from the ovisacs or spontaneously laid by the mother have a perivitelline space filled with fluid. This fluid includes tiny particles and becomes opalescent when the eggs are properly illuminated. Fresh eggs are spherical and yellow and measure 650– 700μ mi in diameter (Figs 2A, 3A). The first sign that the meiotic arrest has been released and that development is under way is the appearance of a shallow circular depression, 25– 50μ m in diameter, at the egg surface. This region, called the grey spot, corresponds to the outer pole of the meiotic spindle. It marks the site of release of the pole cells and accordingly the animal pole of the egg. Between 1.15 and 1.45 h, the egg develops two lobes as an equatorial or slightly supraequatorial ring of contraction appears (Fig. 2B). As this ring of contraction moves towards the animal pole, the first pole cell emerges at the centre of the grey spot (Fig. 2C). This is a transparent cell, 25– 50 μ m in diameter, that remains connected to the surface of the grey spot until discharge of the second pole cell.

Deformation of the animal hemisphere of the egg often ceases shortly after emergence of the first pole cell. Although the egg resumes its spherical shape (Fig. 2C), it soon enters into a new period of deformation. Flattening of both hemispheres, along the animal-vegetal axis, results in the egg resembling a biconvex lens. After about 30min, the egg again recovers its spherical shape. At 2.30–2.45 h the animal hemisphere of the egg is deformed into a coneshaped structure which persists for about 15 min (Fig. 2D). During this time the second pole cell emerges from the animal pole immediately below the first pole cell. These two cells are similar in size and appearance and lie one on the top of the other. As release of the second pole cell is completed, the vegetal hemisphere deforms into a cone-shaped structure similar to that produced shortly before in the animal hemisphere. This deformation lasts for approximately 15 min and is followed by the protrusion of the vegetal pole. This constitutes the vegetal protuberance (Fig. 2D).

Inspection of cleared eggs at late stage l b shows that a mass of opaque ooplasm appears at the centre of the egg. Moreover, after singamy is completed this central mass of ooplasm is destined to house the cleavage nucleus. For that reason, the central domain of ooplasm is designated as perinuclear plasm (Fig. 3B).

After termination of the animal and vegetal hemisphere deformation, many eggs suffer a brief dorsoventral flattening of their animal hemisphere. Meanwhile, the first pole cell comes down to the egg surface. Division of the first pole cell does not seem to occur.

Another episode of deformation movements takes place at both hemispheres of the egg 30–45 min later. First, a ring of contraction appears at the upper third of the animal hemisphere which bulges out in the form of a dome-shaped structure about 100 μm high (Fig. 2E). By this time, the animal pole is marked by a shallow depression that may include the two de-generating pole cells. The ring of contraction of both live and fixed/cleared eggs becomes increasingly conspicuous as whitish or opaque ooplasm gradually accumulates at its walls. The animal polar ring, thus formed, is about 50μm thick and 250–500 μm in diameter (Fig. 3C). The vegetal polar ring forms similarly, but later, at the other hemisphere of the egg. The vegetal polar ring is often smaller (150–400 μm in diameter) and less marked on the egg surface (Fig. 2E).

Meridional bands of contraction are initiated at the equator of the egg in the form of short and shallow grooves first seen by midstage 1c. Grooves become prominent as they grow in length and depth and whitish ooplasm accumulates at their walls (Fig. 3C). Since the majority of the grooves reach both polar rings, a system of meridians forms. By the end of the stage Id, ten to fifteen such meridians (20–60μm thick) have formed and the egg now has a pumpkin-like appearance. Meanwhile, fixed/cleared eggs show that the perinuclear plasm has grown into a disk-shaped structure which reaches 150–300 μm in diameter (Fig. 3C).

The grooves diminish first at the equator and then progressively toward the poles of the egg. This causes a shortening of the meridians towards the poles and concentration of ooplasm in the vicinity of the polar rings. Concomitantly, increasing constriction of the polar rings leads to gradual accumulation of their ooplasm at both poles of the egg (Fig. 3D). Formation of the teloplasm terminates when the system of meridians disappears from the egg surface. Since this event first happens in the vegetal hemisphere of the egg, the vegetal teloplasm forms first and the animal teloplasm forms last (Fig. 3E).

Another change in the overall shape of the egg, from spherical to biconcave, is initiated during stage le and is completed during stage If (Fig. 2F). This conformational change is brought about by dorsoventral flattening of the egg and deepening of its poles. Fixed/cleared eggs show that during this process the teloplasm undergoes two important topographical changes. First, it is displaced at that side of the egg from which the blastomere CD will originate. Second, the teloplasm begins to be internalized and thus less and less of it remains adjacent to the surface. Simultaneously, the elongated dumb-bell-shaped perinuclear plasm moves toward the region of the egg destined to form the AB blastomere. The first cleav-age furrow appears at the dorsal surface of the egg, just beside the animal teloplasm. In this manner, the deepening cleavage furrow subdivides the perinuclear plasm into two symmetrical halves (Fig. 3F). More-over, the CD blastomere encloses most of the egg teloplasm and thus is larger than the AB blastomere.

The preceding description of developmental stages of T. rude uncleaved eggs reveals that this cell is engaged in a stereotyped sequence of peculiar deformation movements leading to completion of meiosis and ooplasmic segregation. As a first step towards the understanding of these processes, our attention was directed to the study of the structure of the early stage la egg.

Examination of whole-mounted fixed/cleared eggs with the dissecting microscope revealed that most of the egg cytoplasm corresponds to yolk. The figure-eight-shaped opaque body, visualized at one pole of the egg, corresponds to the meiotic female nucleus (Fig. 3A). Further details on the structure of the egg were obtained by examination of sectioned material under the light and electron microscopes. These studies show that the early uncleaved egg consists of endoplasm, a layer of ectoplasm and the meiotic female nucleus (Fig. 4). Since examined eggs were all fertilized, they must also include the sperm nucleus. Unfortunately it has not yet been possible to detect it in early uncleaved eggs.

The endoplasm consists of numerous darkly stained yolk platelets scattered across a lightly stained matrix (Fig. 4). Yolk platelets are about 0·1–15 μm in diameter, lack a limiting membrane and contain a regulararray of highly osmiophilic granules. Some platelets have outer and inner regions that stain differently. The outer region stains darker because its granules are more closely packed (Fig. 8). However, very small patches of less-densely packed granules may be found across the outer region (Fig. 7). Yolk platelets are not segregated by size and their distribution is generally random. Alignment of yolk platelets occurs around the poles of the meiotic spindle (Fig. 16). The vitelloplasmic matrix includes moderate numbers of small mitochondria, an assortment of dense bodies and numerous granules about 25 nm in diameter (Figs 8–10). Some dense bodies have limiting membranes and may include vesicles and clumps of electron-dense material. These bodies are 1–2μm in diameter and may correspond to lysosomes (Fig. 10). Other dense bodies exhibit a spongy structure due to the presence of numerous vesicular profiles. The nature of these spongy bodies is unknown, but they may correspond to transformed yolk platelets (Fig. 10). Finally, some dense bodies have very irregular outlines and resemble lipid droplets. Granules of 25 nm are also abundant in other regions of the egg and may correspond to glycogen particles (our unpublished observations). Many bundles of thin filaments, 7-8 nm in diameter, have been found in lacunae-type expansions of the matrix containing accumulations of dense bodies, small yolk platelets and mitochondria (Figs 9, 10). Interestingly these filament bundles are often seen making close contacts with mitochondria and dense bodies (Fig. 9).

An irregular layer of ooplasm (2–10μm thick) that has few small yolk platelets is seen around the egg surface. Although this layer is continuous with the endoplasmic matrix, it differs markedly from it. The main difference is concerned with the accumulation of vesicular and fibrillar material in the ectoplasm (Figs 5–7). The layer of ectoplasm may be subdivided into two sectors. The thin outer sector, which corresponds to the egg cortex, includes few granules and numerous 7–8 nm thick filaments. Some of these filaments form a subplasmalemmal fine network (Figs 6, 7) and the rest is arranged in bundles running more or less parallel to the surface. The thick inner sector, which corresponds to the subcortical layer, is more granular than fibrillar and includes numerous vesicles and various organelles. A fine network of filaments is much less evident, but bundles of filaments are seen travelling in various planes. Obliquely oriented bundles are sometimes seen passing into the outer layer, whereas perpendicularly oriented bundles may be seen entering microvilli. Therefore, bundles of filaments seem to form a coarse network of filaments which extends throughout the entire ectoplasm. Rounded or oval vesicles appear scattered across the inner sector of the ectoplasm. These vesicles sometimes enclose granular or membranous material. The plasmalemma of the early uncleaved egg forms numerous long (1–2 μm) and some short (approximately 0·5μm) microvilli. These two types of microvilli intermingle and their distribution is largely isomorphic. Microvilli are found on the egg surface singly or in groups. In the latter case, two or more microvilli are seen to originate from a short and broad common stalk (Figs 5, 11). A bundle of thin filaments extends from the tip of the microvillus to the inner or outer sector of the layer of ectoplasm. A fine network of filaments may be found in some microvilli, but granules are rarely seen. The plasmalemma between microvilli generally displays irregular outlines and may form coated or uncoated pits (Fig. 6). In some cases vesicles are seen clustered at the neighbourhood of such pits.

Further details on the organization of filaments in the early uncleaved egg have been obtained by examination, under the scanning electron microscope, of whole-mounted permeabilized eggs. Low-magnification electron micrographs (Fig. 12) show filaments arranged into strands or bundles that constitute a three-dimensional network throughout the ectoplasm. The strands differ in length and thickness, and the spaces between them are thus of different sizes and shapes. The spaces of the network probably contained small yolk platelets, dense bodies, vacuoles or mitochondria removed during the extraction procedure. High-magnification electron micrographs (Fig. 13) show that filaments frequently change direction as they pass from one strand into another. In this manner, a much finer network with tiny spaces is formed. These spaces probably contained smaller structures such as vesicles or granules, extracted from the ectoplasm during permeabilization.

Staining of permeabilized eggs with the actin probe, rhodamine-labelled phalloidin, demonstrates that the early uncleaved egg contains a large amount of F-actin. As shown in Figs 14 and 15, the ectoplasm includes an actin network that probably corresponds to the network of filaments seen under the transmission and scanning electron microscopes. As expected, changes in the geometry of the network generate variations in its staining with the actin probe (see Fig. 15). Another interesting feature of the ectoplasmic F-actin network is that it extends into the endoplasm, where it forms a tenuous network around the yolk platelets.

The main axis of the meiotic nucleus of the early uncleaved egg lies parallel or slightly oblique to the egg surface. It consists of two voluminous and symmetrical poles linked to one another by a strand of ooplasm (about 60pm long), containing the spindle and the chromosomes (Figs 3A, 4). The latter are very small and display a metaphase arrangement. Such arrangement of the chromosomes has also been detected in ripe eggs forced to remain in the ovisacs for varying lengths of time. The anaphase movement of the chromosomes, however, is initiated soon after the eggs abandon the ovisacs. Therefore, it is quite clear that eggs are subjected to meiosis I metaphase-arrest. Changes in the orientation of the meiotic spindle take place during anaphase I (our unpublished observations), presumably as a result of changes in the organization of the spindle poles. Hence, it was of interest to determine the structure and relationships of the poles of the early meiotic spindle. Toluidine-blue-stained sections show that the poles of the meiotic spindle consist of centrosome, centrosphere and astrosphere (Fig. 16). The centrosome appears as a rounded unstained central region of the pole enclosing a pair of darkly stained granules, about 1 μm thick, interpreted to be the centrioles. The centrosphere may be subdivided into inner and outer sectors that stain differently. The inner sector is 10–20 μm in diameter and includes particles that stain with toluidine blue. The outer sector is 40–50 μm in diameter and stains much lighter. The astrosphere may reach 250 gm in diameter and is made of numerous straight or curved fibres that may be more than 100gm long. The astral fibres of the two spindle poles extend both into the endo-plasm and ectoplasm. The electron microscope shows that the inner centrosphere includes granules and numerous microtubules, which are seen to travel away from the centrosome periphery in all directions (Fig. 17). In the outer centrosphere, microtubules gather in bundles that constitute radial fibres (Fig. 18). The cytoplasm between these fibres contains fewer granules, mitochondria and small yolk platelets. As radial fibres leave the centrosphere, they are funnelled between the yolk platelets to constitute the astral fibres (Fig. 19). The centrosome has not yet been examined under the electron microscope.

Examination of numerous thin sections reveals that the majority, if not all, of the microtubules of the early uncleaved egg are part of the female meiotic nucleus; microtubules are absent across most of the ectoplasm and endoplasm.

Results presented in this paper indicate that the uncleaved egg of T. rude is very convenient material for the study of cytoplasmic movements associated with the expulsion of the pole cells and with ooplasmic segregation. Deformation movements are prominent and stereotyped. Since the ooplasm has a distinct whitish appearance and its redistribution mostly takes place across the egg surface, ooplasmic segregation may be studied in live eggs. Furthermore, very simple clearing techniques allow detailed observations of both deep and superficial translocation of ooplasm during teloplasm and perinuclear plasm formation.

Development of the uncleaved egg of T. rude may be subdivided into two periods. The early period (stages la and 1b) deals with completion of meiosis and initiation of the perinuclear plasm formation. The late period (stages 1c to 1f) deals with formation of the teloplasm and completion of the first cleavage division. A similar developmental sequence is seen in Tubifex eggs, which also form prominent accumulations of ooplasm (pole plasms) at both egg poles (Shimizu, 1982a). Moreover, the egg of this oligochaete includes a prominent actin lattice which seems to generate the motive force for the ooplasmic segregation (Shimizu, 1982b, 1984). Recent observations by the same investigator (Shimizu, 1986) indicate that accumulation and displacement of superficial ooplasm in the form of rings and meridians also seem to occur in Tubifex eggs. However, the eggs of Tubifex and T. rude differ in several respects including the types of deformation movements accompanying release of the pole cells. Thus, Tubifex eggs form meridional bands of contraction during release of the two pole cells. T. rude eggs, on the other hand, form rings of contraction during the same process. Finally, Tubifex eggs differ from T. rude eggs in that the former do not form a vegetal protuberance.

Meiotic arrest may be produced by the presence in the oocyte cytoplasm of a substance that blocks meiosis beyond metaphase I. Such a substance may be a cytostatic factor, perhaps similar to that partially purified from unfertilized Xenopus eggs by Meyerhof & Masui (1979a,b) and Masui, Meyerhof & Miller (1980). The factor might be quickly inactivated when T. rude eggs are exposed to the cocoon fluid or culture medium.

There are two reasons why it was of interest to study the structure of the ectoplasm. First, because the force-generating elements triggering the deformation movements of the egg probably reside at the cell surface. Second, because it is known that after fertilization the egg surface is subjected to profound remodelling. An extensive framework of actin filaments seems to be present throughout the ectoplasm of the early uncleaved egg. Since microtubules are not found across most of the ectoplasm and endoplasm, the cytoskeleton of the early uncleaved egg may largely consist of actin filaments. Of course the presence of intermediate filaments remains to be investigated. As it occurs in the egg of sea urchins (Burgess & Schroeder, 1977; Begg & Rebhun, 1979; Mabuchi, Hosoya & Sakai, 1980; Carron & Longo, 1982) and amphibia (Charbonneau & Picheral, 1983), assembly of cortical actin filaments and elongation of microvilli in T. rude eggs may also be induced at the time of fertilization. The fine network component of the ectoplasmic cytoskeleton may be considered to represent a sort of three-dimensional actin gel, that perhaps contributes to the mechanical properties of the egg surface (Kane, 1983). The coarse network component of the ectoplasmic cytoskeleton, on the other hand, would not only provide rigidity to the egg surface but also probably represents the potential contractile machinery responsible for the deformation movements of the egg. The bundle of filaments present in the microvilli is considered to be part of the coarse actin network and is probably engaged in shortening of microvilli taking place during stage la (our unpublished observations). Arrangement of filaments in fine or coarse networks may depend on the type of protein or factor associated with F-actin (see Kane, 1982, 1983; Schliwa, 1981). Moreover, it is quite possible that the two types of networks are interconvertible.

The available evidence indicates that the endoplasmic actin cytoskeleton seems to be composed of fewer elements and this situation might be related to the greater fluidity of the vitelloplasm in these eggs. However, there are regions of the endoplasm where large bundles of filaments appear to build up a complex cytoskeleton closely related to the surface of mitochondria as well as other structures. This observation suggests that an actin-based system for the translocation of organelles may be present across the endoplasm. Such systems may be responsible, for example, for the concentration of ooplasm at the egg centre during perinuclear plasm formation.

Presence of pits along the plasmalemma and of many vesicles across the cortical ooplasm of leech early uncleaved eggs suggests that membrane may be flowing into and out of the egg surface. This flow may be related to membrane retrieval provoked by microvilli resorption. Presence of some short microvilli across the surface of Oh eggs indicates that such processes may have already started in the intra-ovarian oocyte. Membrane retrieval is known to take place in other invertebrate eggs after cortical granule exocytosis which has expanded the area of the plasmalemma (Fisher & Rebhun, 1983). Thus, vesicles scattered throughout the ectoplasm would be formed by endocytosed plasmalemma.

We thank Víctor Guzmán, Víctor Monasterio, Ana Valdés and Lilio Yañez for technical assistance, Cecilia Fernández for preparing the illustrations, Gunther Stent and Manfred Schliwa for the supply of various reagents and John Gerhart for critical reading of the manuscript. Margery Hoogs, Duncan Stuart and Ronald Davies kindly provided many of the animals used in this study. This investigation was supported by grants B 1987-8415, 8525 and 8636, from Departamento de Investigación y Bibliotecas, Universidad de Chile and grant 1218, from Fondo Nacional de Investigación Científica y Tecnológica.