The fine structure of Comanthus japonica (Echinodermata: Crinoidea) from zygote through early gastrula

The fine structure of the early embryo of Comanthus has been described by scanning and transmission electron microscopy at approximately 20-min intervals from zygote (20 min) through early gastrula (260 min). In normally developing (and presumably monospermic) embryos, some non-fertilizing sperm were invariably trapped in the perivitelline space; this suggests that there is an effective block to polyspermy at the level of the plasma membrane. No trace of a hyaline layer is encountered in the pervitelline space. At first cleavage, which begins unilaterally at the animal pole, the contractile ring filaments are rather thick (50–150 Å) in comparison to those known for other marine invertebrates. From first cleavage through early gastrula, the lateral surfaces of the blastomeres are broadly adherent, and there is an intercellular material, presumably an adhesive, in the intercellular space. The blastocoel first appears during the four-cell stage. From the eight-cell stage through the start of gastrulation, only one opening, the vegetal pore, connects the blastocoel with the perivitelline space. Gastrulation begins at the 50–100-cell stage, while the vegetal pore is still open, and a clearly defined blastula stage is bypassed. Gastrulation is by a novel process, which I have called holoblastic involution. At gastrulation the eight most vegetal blastomeres, which encircle the vegetal pore, shoot out erect, unbranched filopodia for many microns through the blastocoel. The filopodia adhere to the blastocoelic surfaces of the animal blastomeres and contract, pulling the vegetal blastomeres into the blastocoel. The migrated vegetal blastomeres adhere to one another, forming the entoderm in the vegetal region of the embryo; the remaining blastomeres become the ectoderm. Soon after the completion of cell migration, the entodermal blastomeres appear to cast off their contractile microappendages and adhesive membranes into the blastocoel.     

Distribution and redistribution of pigment granules in the development of sea urchin embryos

Summary Pigment granules (PGs) are embeded in the cortex of embryos of the Japanese sea urchins,Hemicentrotus pulcherrimus and Anthocidaris crassispina. PGs in the cortex actively retreated from the vegetal pole area at the 4-cell stage and then a notable PG-distribution gradient formed along the egg axis (the polar redistribution of PGs). The polar redistribution of PGs in the cortex occurred at the same time after fertilization even in solutions of microtubule disrupting reagents such as Colcemid, vinblastine sulfate or griseofulvin. Consequently, the polar redistribution of PGs was not associated with the microtubules. However, the polar redistribution of PGs was interrupted in seawater containing cytochalasin B (CB), dithiothreitol (DTT) or tetracaine, and the distribution pattern of PGs in the cortex was definitely disturbed. Moreover, CB, DTT and tetracaine altered the division pattern of vegetal blastomeres at the 4th cleavage which is normally unequal so that all the blastomeres divided equally. Microtubule disrupting reagents did not have such an effect on the cleavage pattern. Thus the cortical movement along the egg axis reflected by the polar redistribution of PGs seems to correlate with the micromere formation.     

Removal of vegetal yolk causes dorsal deficencies and impairs dorsal-inducing ability of the yolk cell in zebrafish.

To examine the nature of cytoplasm determinants for dorsal specification in zebrafish, we have developed a method in which we remove the vegetal yolk hemisphere of early fertilized eggs (vegetal removed embryos). When the vegetal yolk mass was removed at the 1-cell stage, the embryos frequently exhibited typical ventralized phenotypes: no axial structures developed. The frequency of dorsal defects decreased when the operation was performed at later stages. Furthermore, the yolk cell obtained from the vegetal-removed embryos lost the ability to induce goosecoid in normal blastomeres while the normal yolk cell frequently did so in normal and vegetal-removed embryos. These results suggested that the vegetal yolk cell mass contains the dorsal determinants, and that the dorsal-inducing ability of the yolk cell is dependent on the determinants.     

Animal and vegetal poles of the mouse egg predict the polarity of the embryonic axis, yet are nonessential for development

Recent studies suggest early (preimplantation) events might be important in the development of polarity in mammalian embryos. We report here lineage tracing experiments with green fluorescent protein showing that cells located either near to or opposite the polar body at the 8-cell stage of the mouse embryo retain their same relative positions in the blastocyst. Thus they come to lie on either end of an axis of symmetry of the blastocyst that has recently been shown to correlate with the anterior-posterior axis of the postimplantation embryo (see R. J. Weber, R. A. Pedersen, F. Wianny, M. J. Evans and M. Zernicka-Goetz (1999). Development 126, 5591-5598). The embryonic axes of the mouse can therefore be related to the position of the polar body at the 8-cell stage, and by implication, to the animal-vegetal axis of the zygote. However, we also show that chimeric embryos constructed from 2-cell stage blastomeres from which the animal or the vegetal poles have been removed can develop into normal blastocysts and become fertile adult mice. This is also true of chimeras composed of animal or vegetal pole cells derived through normal cleavage to the 8-cell stage. We discuss that although polarity of the postimplantation embryo can be traced back to the 8-cell stage and in turn to the organisation of the egg, it is not absolutely fixed by this time.     

beta-Catenin in early development of the lancelet embryo indicates specific determination of embryonic polarity

The lancelet (amphioxus) embryo develops from a miolecithal egg and starts gastrulation when it is approximately 400 cells in size, in a fashion similar to that of some non-chordate deuterostomes. Throughout this type of gastrulation, the embryo develops characteristics such as the notochord and hollow nerve cord that commonly appear in chordates. beta-Catenin is an important factor in initiating body patterning. The behavior and developmental pattern of this protein in early lancelet development was examined in this study. Cytoplasmic beta-catenin was localized to the animal pole after fertilization and then was incorporated asymmetrically into the blastomeres during the first cleavage. Asymmetric distribution was observed at least until the 32-cell stage. The first nuclear localization was at the 64-cell stage, and involved all of the cells. At the initial gastrula stage, however, concentrated beta-catenin was found on the dorsal side. LiCl treatment affected the asymmetric pattern of beta-catenin during the first cleavage. LiCl also changed distribution of nuclear beta-catenin at the initial gastrula stage: distribution extended to cells on the animal side. Apparently associated with this change, expression domains of goosecoid, lhx3 and otx also changed to a radially symmetric pattern centered at the animal pole. However, LiCl-treated embryos were able to establish embryonic polarity. The present study suggests that in the lancelet embryo, polarity determination is independent of dorsal morphogenesis.     

Electron microscopic study of the development of amphioxus,Branchiostoma belcheri tsingtauense: The gastrula

The gastrulae of amphioxus were investigated by means of scanning and transmission electron microscopy (SEM and TEM) during 7 arbitrary stages that were seen about 4 to 10 hr after fertilization. Throughout gastrulation, SEM revealed subtle differences in cells of the blastoporal lip. In fractured specimens at early and middle stages, two opposing zones different in shape, size, and connection of the component cells were found: one which consists of columnar smaller cells in close contact in animal region and the other which is composed of round or polygonal larger cells in looser association in vegetal region. The polar body was found unexpectedly on the concave vegetal surface of the early gastrula in about 25% of cases. This might be the result of migration of the polar body. A short cilium that later elongated was recognized on each cell at mid-gastrula stage. The cilia on the dorsal surface (the neural ectoderm) of the final-stage gastrula became shorter than those on the epidermal ectoderm. TEM of thin sections demonstrated that the cytoplasmic components of gastrula cells are essentially the same as those of cleavage cells. But, the homogeneous nucleus seen during cleavage changed into a heterogeneous structure in which a nucleolus and dense particles were seen. Until the late stage, regional characteristics of the gastrulae indicating definitively the anterior-posterior and dorso-ventral polarity were not detected in the present SEM and TEM study.     

Mesoderm formation by isolated and cultivated 8-cell stage blastomeres of the teleost, Leucopsarion ptersii (shiro-uo).

Isolation of cleavage-stage blastomeres and the study of their developmental potential has been used extensively for analyzing the mechanisms of embryogenesis in vertebrates, including amphibians and echinoderms. We devised a method to isolate 8-cell stage blastomeres in the teleost, shiro-uo, by utilizing its unique cleavage pattern of the horizontal 3rd cleavage plane. Removal of all the upper blastomeres at the 8-cell stage allowed almost normal embryogenesis from the remaining lower blastomeres and yolk cell mass. Isolated upper or lower blastomeres formed vesicles and spherical bodies, which later showed morphological changes during cultivation. Mesoderm formation was detected not only in the cultivated lower blastomeres or whole blastomeres but also in the upper blastomeres isolated from the yolk cell mass at the 8-cell stage, although at a lower frequency than the lower blastomeres. These results indicated the presence of very early signaling for mesoderm induction, which is independent from the currently postulated signals from the yolk syncytial layer at later stages. This also indicated non-equivalence or differentiation of the blastomeres from the very early cleavage stage in teleost embryos.     

'METACHRONOUS' CLEAVAGE AND INITIATION OF GASTRULATION IN AMPHIBIAN EMBRYOS

The cleavage pattern in the egg of Xenopus laevis has been investigated with the aid of time-lapse cinematography. From the 5th cleavage onward, divisions of the surface blastomeres are not synchronous but metachronous. A few blastomeres in a very restricted region which is situated in most cases in the dorsal side of the animal hemisphere, slightly distant from the median line and near the equatorial junction of the animal and vegetal hemispheres, divide before the other blastomeres, and a wave-like propagation of the divisions travels along the surface from that region toward the animal and vegetal poles. The wave-like propagation ends in the vegetal pole region. In the animal hemisphere, this pattern of cleavage is continued until the 13th cleavage and thereafter the divisions of surface blastomeres become asynchronous. In the vegetal pole region, however, the 14th metachronous division of blastomeres is clearly observed in the film. Gastrulation begins after 14 cleavages.     

A Stereometric Analysis of Karyokinesis, Cytokinesis and Cell Arrangements during and following Fourth Cleavage Period in the Sea Urchin, Lytechinus variegatus

Fourth cleavage of the sea urchin embryo produces 16 blastomeres that are the starting point for analyses of cell lineages and bilateral symmetry. We used optical sectioning, scanning electron microscopy and analytical 3-D reconstructions to obtain stereo images of patterns of karyokinesis and cell arrangements between 4th and 6th cleavage. At 4th cleavage, 8 mesomeres result from a variant, oblique cleavage of the animal quartet with the mesomeres arranged in a staggered, offset pattern and not a planar ring. This oblique, non-radial cleavage pattern and polygonal packing of cells persists in the animal hemisphere throughout the cleavage period. Contrarily, at 4th cleavage, the 4 vegetal quartet nuclei migrate toward the vegetal pole during interphase; mitosis and cytokinesis are latitudinal and subequatorial. The 4 macromeres and 4 micromeres form before the animal quartet divides to produce a 12-cell stage. Subsequently, macromeres and their derivatives divide synchronously and radially through 8th cleavage according to the Sachs-Hertwig rule. At 5th cleavage, mesomeres and macromeres divide first; then the micromeres divide latitudinally and unequally to form the small and large micromeres. This temporal sequence produces 28-and 32-cell stages. At 6th cleavage, macromere and mesomere descendants divide synchronously before the 4 large micromeres divide parasynchronously to produce 56- and 60-cell stages.     

Overexpression of S-adenosylmethionine decarboxylase (SAMDC) in Xeno-pus embryos activates maternal program of apoptosis as a "fail-safe" mechanism of early embryogenesis

In Xenopus, injection of S-adenosylmethionine decarboxylase (SAMDC) mRNA into fertilized eggs or2-cell stage embryos induces massive cell dissociation and embryo-lysis at the early gastrula stage due toactivation of the maternal program of apoptosis. We injected SAMDC mRNA into only one of the animal side blastomeres of embryos at different stages of cleavage, and examined the timing of the onset of theapoptotic reaction. In the injection at 4- and 8-cell stages, a considerable number of embryos developed intotadpoles and in the injection at 16- and 32-cell stages, all the embryos became tadpoles, although tadpolesobtained were sometimes abnormal. However, using GFP as a lineage tracer, we found that descendant cellsof the blastomere injected with SAMDC mRNA at 8- to 32-cell stages are confined within the blastocoel atthe early gastrula stage and undergo apoptotic cell death within the blastocoel, in spite of the continueddevelopment of the injected embryos. These results indicate that cells overexpress     

Overexpression of S-adenosylmethionine decarboxylase.（SAMDC） in Xeno-pus embryos activates maternal program of apoptosm as a “fail-safe”mechanism of early embryogenesis

In Xenopus, injection of S-adenosylmethionine decarboxylase (SAMDC) mRNA into fertilized eggs or 2-cell stage embryos induces massive cell dissociation and embryo-lysis at the early gastrula stage due toactivation of the maternal program of apoptosis. We injected SAMDC mRNA into only one of the animalside blastomeres of embryos at different stages of cleavage, and examined the timing of the onset of theapoptotic reaction. In the injection at 4-and 8-cell stages, a considerable number of embryos developed intotadpoles and in the injection at 16-and 32-cell stages, all the embryos became tadpoles, although tadpolesobtained were sometimes abnormal. However, using GFP as a lineage tracer, we found that descendant cellsof the blastomere injected with SAMDC mRNA at 8-to 32-cell stages are confined within the blastocoel atthe early gastrula stage and undergo apoptotic cell death within the blastocoel, in spite of the continued development of the injected embryos. These results indicate that cells overexpressed with SAMDC undergo apoptotic cell death consistently at the early gastrula stage, irrespective of the timing of the mRNA injection.We assume that apoptosis is executed in Xenopus early gastrulae as a “fall-safe“ mechanism to eliminate physiologically-severely damaged cells to save the rest of the embryo.     

Epiboly connected with cleavage in morula and early blastula stages of Xenopus laevis,a study using time-lapse photography

Summary Measurements were made of the external surface areas occupied by animal and vegetal blastomeres and their daughter cells at successive cleavage cycles in 15 embyros of Xenopus laevis. On the animal side, after each cleavage a general area increase (epiboly) occurs from cycle 4 (16-cell stage, stage 5) to cycle 10 (stage 8 1/2), while on the vegetal side there is a slight general area decrease after each cleavage from cycle 6 to cycle 10. The comparison between the external surface areas of individual animal blastomeres and those of their daughter cells, visible at the next cycle, shows a significantly larger increase after radial than after tangential cleavage, a difference that may be connected with the insertion of new membrane into the external surface at radial cleavage.     

The effect of local cortical microfilament disorganization on ooplasmic segregation in the loach (Misgurnus fossilis) egg

Injections of cytochalasin D (CD) or DNase I under the surface of fertilized loach egg result in local disorganization of microfilamentous cortex (MC) as revealed by transmission electron microscopy. This effect correlates with the loss of the cortex ability to contract in vitro. The disorganization of MC in the vegetal hemisphere of the egg does not affect the ooplasm segregation or blastodisk cleavage. Injection under the animal pole suppresses blastodisk formation and results in the autonomous separation of ooplasm in the central part of the egg. The experiments suggest that (1) autonomous separation of ooplasm from the yolk granules can proceed in the central part of the egg without the participation of MC; (2) normal segregation of ooplasm at the animal pole requires that the structures of microfilaments in the animal hemisphere (but not in the vegetal one) be preserved.     

Cell differentiation during early development of Nassarius reticulatus L. (Gastropoda Prosobranchia)

Summary The ultrastructure of the zygote and of early segmentation stages, up to the 16-cell stage, was investigated in normal Nassarius reticulatus from Roscoff (France). This study deals predominantly with structures that remain morphologically unchanged throughout this period, such as yolk granules, lipid droplets, and multimembranous vesicles. These organelles do not change in position, fine structure, or quantity from the egg to the 16-cell stage. The cortex and the vitelline layer also remain almost unchanged until the 16-cell stage. Cortical granules could not be observed and thus the thin vitelline layer is not transformed to a thick fertilization membrane. This phenomenon seems to be related to the facts that the eggs are protected by a tough capsule and that the cells of the embryo have to incorporate extraembryonic nutrient substances. From the 2-cell stage until the 16-cell stage micropinocytotic vesicles 500–2000 Å in diameter are associated with the plasmalemma. Perhaps they help to incorporate the extraembryonic nutrient reserves. Desmosomes first appear at the 4-cell stage and are common at the 8-cell stage, when a small transitory blastocoel appears. Normally the cell borders are in close proximity during interphase. Syncytial connections between the blastomeres were not observed. Beginning at the trefoil stage the mitochondria increase in number and many apparent division stages are observed. The increase in mitochondria occurs in the perinuclear region throughout the embryo and does not result in a polar lobe especially rich or poor in mitochondria. From the 4-cell stage onwards, the number of ribosomes increases differentially in different blastomeres.Dedicated to Prof. B. Rensch on his 75th birthday. We wish to thank Mrs. C. Mehlis for her valuable technical assistance and Professor J. Bergérard for the excellent working conditions at the Station Biologique at Roscoff (France).     

Ultraviolet irradiation of eggs and blastomere isolation experiments suggest that gastrulation in the direct developing ascidian, Molgula pacifica, requires localized cytoplasmic determinants in the egg and cell signaling beginning at the two-cell stage

Gastrulation in the maximum direct developing ascidian Molgula pacifica is highly modified compared with commonly studied "model" ascidians in that endoderm cells situated in the vegetal pole region do not undergo typical invagination and due to the absence of a typical blastopore the involution of mesoderm cells is highly modified. At the gastrula stage, embryos are comprised of a central cluster of large yolky cells that are surrounded by a single layer of ectoderm cells in which there is only a slight indication of an inward movement of cells at the vegetal pole. As a consequence, these embryos do not form an archenteron. In the present study, ultraviolet (UV) irradiation of fertilized eggs tested the possibility that cortical cytoplasmic factors are required for gastrulation, and blastomere isolation experiments tested the possibility that cell signaling beginning at the two-cell stage may be required for the development of the gastrula. Irradiation of unoriented fertilized eggs with UV light resulted in late cleavage stage embryos that failed to undergo gastrulation. When blastomeres were isolated from two-cell embryos, they developed into late cleavage stage embryos; however, they did not undergo gastrulation and subsequently develop into juveniles. These results suggest that cytoplasmic factors required for gastrulation are localized in the egg cortex, but in contrast to previously studied indirect developers, these factors are not exclusively localized in the vegetal pole region at the first stage of ooplasmic segregation. Furthermore, the inability of embryos derived from blastomeres isolated at the two-cell stage to undergo gastrulation and develop into juveniles suggests that important cell signaling begins as early as the two-cell stage in M. pacifica. These results are discussed in terms of the evolution of maximum direct development in ascidians.     

Maternal information and localized maternal mRNAs in eggs and early embryos of the ascidian Halocynthia roretzi

Summary

The mosaic behavior of blastomeres isolated from ascidian embryos has been taken as evidence that localized ooplasmic factors (cytoplasmic determinants) specify tissue precursor cells during embryogenesis. Experiments involving the transfer of egg cytoplasm have revealed the presence and localization of various kinds of cytoplasmic determinants in eggs of Halocynthia roretzi. Three cell fates, epidermis, muscle and endoderm, are fixed by cytoplasmic determinants. The three kinds of tissue determinants move in different directions during ooplasmic segregation. Prior to the onset of the first cleavage the three kinds of determinants reside in egg regions that correspond to the future fate map of the embryo and then they are differentially partitioned into specific blastomeres. In addition to tissue-specific determinants, there is evidence suggesting that ascidian eggs contain localized cytoplasmic factors that are responsible for controlling the cleavage pattern and morphogenetic movements. Transplantation of posterior-vegetal egg cytoplasm to an anterior-vegetal position causes a reversal of the anterior-posterior polarity of the cleavage pattern. Localized cytoplasmic factors in the posterior-vegetal region are involved in the generation of a unique cleavage pattern. When vegetal pole cytoplasm is transplanted to the animal pole or equatorial position of the egg, ectopic gastrulation occurs at the site of transplantation. This finding supports the idea that vegetal pole cytoplasm specifies the site of gastrulation. Recently, we started a cDNA project to analyze maternal mRNAs. An arrayed cDNA library of fertilized eggs of H. roretzi was constructed, and more than 2000 clones have been partially sequenced so far. To estimate the proportion of the maternal mRNAs that are localized in the egg and embryo, 150 randomly selected clones were examined by in situ hybridization. We found eight mRNAs that are localized in the eight-cell embryo, of which three were localized to the myoplasm (a specific region of the egg cytoplasm that is partitioned into muscle-lineage blastomeres) of the egg, and then to the postplasm of cleavage-stage embryos. These results indicate that the proportion of localized messages is much higher than we expected. These localized maternal messages may be involved in the regulation of various developmental processes.     

A reexamination of cleavage patterns in eutherian mammalian eggs: rotation of blastomere pairs during second cleavage in the rabbit.

The pattern of cleavage was examined during second and third furrowing of the rabbit egg. Two-cell eggs, collected just prior to onset of second cleavage, were continuously observed in a culture chamber, which was kept at 37 degrees C. Semi-cinematographic techniques were used to photograph progressive stages of cleavage. It was demonstrated that the pattern of cleavage in the rabbit differs from that in the sea urchin, because the blastomeres at the 4-cell stage are arranged crosswise in the former, while they are situated next to each other in the latter. The crosswise arrangement of the blastomeres in the rabbit at the 4-cell stage is a consequence of a 90 degree rotation of the polar axis in one hemisphere of the egg. Subsequently, due to the rotation of the original polar axis in one hemisphere, the third cleavage plane through one half of the egg is transverse to the third cleavage plane through the other half. Evidence is provided to show that the cross wise configuration of blastomeres at the 4-cell stage occurs in other eutherian mammals. It is proposed that this rotational cleavage pattern be recognized as distinct from those of radial, spiral and bilateral.     

Dorsal and ventral cells of cleavage-stage Xenopus embryos show the same ability to induce notochord and somite formation

Cells in the dorsal marginal zone of the amphibian embryo acquire the potential for mesoderm formation during the first few hours following fertilization. An examination of those early cell interactions may therefore provide insight on the mechanisms important for organization of axial structures. The formation of mesoderm (notochord, somites, and pronephros) was studied by combining blastomeres from the animal pole region of Xenopus embryos (32- to 512-cell stages) with blastomeres from different regions of the vegetal hemisphere. The frequency of notochord and somite development was similar in combinations made with dorsal or ventral blastomeres, or with both. Our results show that during early cleavage stages the ventral half of the vegetal hemisphere has the potential to organize axial structures, a property previously believed to be limited to the dorsal region.     

SPICULE FORMATION IN VITRO BY THE DESCENDANTS OF PRECOCIOUS MICROMERE FORMED AT THE 8-CELL STAGE OF SEA URCHIN EMBRYO*

The micromeres at the 16-cell stage of sea urchin embryo have already been endowed with a faculty to self-differentiate into spicule-forming cells (11). The present experiment was designed to test whether the factor(s) necessary for such self-differentiation had already been localized at the 8-cell stage in an area corresponding to the presumptive micromere region in Hemicentrotus pulcherrimus. Since the blastomeres at the 8-cell stage are all equal in size in normal embryo, unequal 3rd cleavage, by which small blastomeres are pinched off toward the vegetal pole (precocious micromeres), was experimentally induced either by treatment with 4NQO (4-nitroquinoline-1-oxide) at the 2-cell stage or by continuous culture in Ca-free sea water. The precocious micromeres were cultured in vitro in natural sea water containing horse serum. Descendants of the precocious micromeres formed spicules. In comparison their spicule formation with that by the descendants of the micromere of normal embryo, no differences were found regarding 1) time of initiation of spicule formation, 2) rate of growth of spicule, 3) size and shape of resultant spicule and 4) percentage of clones which formed spicule. The fact indicates that factor(s) indispensable for self-differentiation into spicule-forming cells have already been localized near the vegetal pole as early as the 8-cell stage.     

Unequal cleavage in the early Tubifex embryo

Unequal cleavage that produces two blastomeres of different size is a cleavage pattern that many animals in a variety of phyla, particularly in Spiralia, adopt during early development. This cleavage pattern is apparently instrumental for asymmetric segregation of developmental potential, but it is also indispensable for normal embryogenesis in many animals. Mechanically, unequal cleavage is achieved by either simple unequal cytokinesis or by forming a polar lobe at the egg's vegetal pole. In the present paper, the mechanisms for unequal cytokinesis involved in the first three cleavages in the oligochaete annelid Tubifex are reviewed. The three unequal cleavages are all brought about by an asymmetrically organized mitotic apparatus (MA). The MA of the first cleavage is monastral in that an aster is present at one pole of a bipolar spindle but not at the other. This monastral form, which arises as a result of the involvement of a single centrosome in the MA assembly, is both necessary and sufficient for unequal first cleavage. The egg cortex during the first mitosis is devoid of the ability to remodel spindle poles. In contrast to the non-cortical mechanisms for the first cleavage, asymmetry in the MA organization at the second and third cleavages depends solely on specialized properties of the cell cortex, to which one spindle pole is physically connected. A cortical attachment site for the second cleavage spindle is generated de novo at the cleavage membrane resulting from the first cleavage; it is an actin-based, cell contact-dependent structure. The cortical microtubule attachment site for the third cleavage, which functions independently of contact with other cells, is not generated at the cleavage membrane resulting from the second cleavage, but is located at the animal pole; it may originate from the second polar body formation and become functional at the 4-cell stage.