Scanning electron microscopy of gastrulation in a sea urchin (Anthocidaris crassispina)

The early morphogenetic processes involved in gastrulation in developing sea-urchin embryos have been investigated by means of light microscopy and transmission electron microscopy. Gastrulation was reported to consist of two phases (Gustafson & Wolpert, 1961); the first and slow phase is characterized by active pulsation and altered intercellular adhesiveness of cells in the vegetal plate following the emigration of primary mesenchyme cells (Gustafson & Wolpert, 1961), and the second and fast phase seems to be implemented by the action of pseudopodia of the secondary mesenchyme cells on the top of invaginating archenteron (Dan & Okazaki, 1956) or of tip cells of the archenteron itself (Gustafson & Kinnander, 1956; Kinnander & Gustafson, 1960). Based on their observations, Gustafson & Wolpert (1962) proposed a hypothetical model for gastrulation.

In the preceding paper (Akasaka, Amemiya & Terayama, 1980) we observed the inside of developing sea-urchin embryos (Pseudocentrotus depressus) with large blastocoel cavities by means of scanning electron microscopy, in the present study, we carried out observations of both the inside and the outside of developing Anthocidaris crassispina embryos with much smaller blastocoel cavities, finding that there may be some principal differences in the gastrulation processes between the two types of sea-urchin embryos.

Sea urchins, Anthocidaris crassispina, were caught near the Misaki Marine Biological Station. Eggs and sperm were collected by artificial spawning using a KC1 solution. Batches of eggs of which more than 95 % could be fertilized were used. Eggs were cultured in filtered sea water at 20 °C with gentle stirring.

Preparation of embryos for scanning electron microscopy was described in the earlier paper (Akasaka et al. 1980). A Hitachi HHS-2R scanning electron microscope was used for observing the surfaces of outside and inside of embryos.

Figure 1 shows the interior surface of an Anthocidaris embryo (blastula, at 10 ·5 h after insemination) dissected along the animal-vegetal axis. It should be noted that the length of columnar cells (or the depth of ectodermal wall) is rather long and the size of blastocoel cavity is extremely small. The intercellular contact in the vegetal plate appears to be loosened. A few cells with their heads somewhat protruded into the blastocoel, which may correspond to the basal lobe cells (Be) (Gibbins, Tilney & Porter, 1969), are seen in the periphery of the vegetal plate.

Moreover one can see some tadpole-like (Tc) or bottle-shaped cells in the vegetal plate in accordance with the observation by Katow & Solursh (1980). They appear to be presumptive primary mesenchyme cells emigrating from the vegetal plate into the blastocoel. Some gaps apparently remaining after the emigration of tadpole-like cells (primary mesenchyme cells) are also seen.

Probably reflecting the alteration of cell shape and intercellular adhesion in the vegetal plate and its vicinity, furrows (a big circular furrow surrounding the vegetal plate with long radial ones in the vicinity) were seen on the outer surface. The number of cells remaining in the vegetal plate after primary mesenchyme cell emigration appears to be about eight as judged from the small furrows apparently surrounding each cell in the vegetal plate (Fig. 2) in accord with the description of Endo (1966) and Katow & Solursh (1980).

At this stage, the cells surrounding the vegetal plate appear to expand their outer cell surfaces (on the hyaline layer side) towards the vegetal plate as judged from the apparently reduced density of microvilli in the area surrounding the vegetal plate as also shown in Fig. 2.

At 13 ·5 h after fertilization, the big circular furrow surrounding the vegetal plate as well as small ones surrounding each of the vegetal plate cells becomes less distinct as shown in Fig. 3 A. The dissected surface of an embryo (Fig. 3B) shows that the intercellular adhesion in the vegetal plate region is reestablished.

At 15 h after fertilization, small furrows surrounding each cell reappeared in the vegetal plate and its vicinity (Fig. 4X), suggesting that the intercellular contact was again loosened at this stage. At the same time, some cells in the vegetal pole region started to protrude their heads into the blastocoel (Fig. 4B). The phenomenon seems similar to the one observed at the time of primary mesenchyme cell formation, suggesting that active cellular movement is taking place.

The cells with surrounding furrows are not restricted to the vegetal plate and the number appears to increase continuously in the vegetal pole region (Fig. 5 A). Some of the cells near the centre of vegetal plate appear to show a characteristic deformation (Fig. 5B), suggesting that the cellular mass is transferring towards the blastocoel side. As a result the cell surface on the hyaline layer side became smaller and smaller while increasing on the blastocoel side.

At 18 h after fertilization, the outer surface area with furrows expanded further, and the furrows in the centre of the vegetal plate now became very small, leaving big furrows or gaps (Fig. 6 − 4). Accordingly the dissected surface of an embryo at this stage (Fig. 6B) clearly indicated that most of the cells in the vegetal plate were transformed into tadpole-like forms with their heads on the blastocoel side and their smaller tails on the hyaline layer side (or a little inside the vegetal plate wall) in contrast to the surrounding cells which had heads on the hyaline layer side. Some cells in the middle of vegetal plate appear to be translocated inwardly (the first sign of indentation).

At 19 ·5 h after fertilization, the indentation of the archenteron is clearly seen not only from the outside of an embryo (Fig. 7 A) but also from the inside (Fig. 7B). The cellular adhesion in the vegetal region appears to be reestablished (Fig. 7 B) and furrows are no longer seen on the outer surface of embryos (Fig. 7 – 4).

At 21 h after fertilization, the embryos seem to be in the mid-gastrular stage ; and some secondary mesenchyme-like cells are seen on the tip of invaginating archenteron. Some pseudopodia from these cells appear to extend to the animal pole, but others are linked to the primary mesenchyme cells (Fig. 8). These pseudopodia are much thinner and their number was smaller compared to those in the corresponding-stage embryos of Pseudocentrotus (Akasaka et al. 1980). The cells on the tip of the invaginating archenteron (Fig. 8) are smaller in size and irregular in shape as compared to those in the other parts of archenteron, apparently forming a bud-like structure.

At 22 ·5 h after fertilization, the tip of the invaginating archenteron nearly reached the animal pole (Fig. 9). The cells on the tip of archenteron now appear to be fully developed and the bud-like structure is no longer seen.

The columnar cells (or the depth of ectodermal wall), except those in the animal pole region, were found to contract up to the late gastrula stage. The tendency appeared to be more marked after the onset of gastrulation. The shortening of columnar cells may contribute to the enlargement of embryonic size or the blastocoel cavity. However, the number of columnar cells (except archenteron cells) on the dissected surfaces of embryos (Figs. 1, 8, 9) remained almost constant (about 50) in spite of the increase in embryonic size, the size of archenteron and the number of archenteron cells.

The invagination (or indentation) of the archenteron in sea-urchin embryos has been extensively investigated by Gustafson and his colleagues (Gustafson & Kinnander, 1956; Kinnander & Gustafson, 1960; Gustafson & Wolpert, 1961 ; Gustafson & Wolpert, 1962), and the two-phase mechanism was presented. The inward indentation of archenteron rudiment in the vegetal pole region (the first phase) is supposed to be due to the reduced intercellular adhesiveness without changing the adhesiveness to the hyaline layer, that may allow the pulsating activity of cells which leads to indentation. On the other hand, the elongation of the archenteron (the second phase) is supposed to be implemented by secondary mesenchyme cells (or archenteron tip cells) and their pseudopodia. The possible involvement of pseudopodia in the gastrulation processes has in general been accepted (Dan & Okazaki, 1956; Trinkaus, 1969; Akasaka et al. 1980), mainly based on the characteristic situation of secondary mesenchyme cells between the archenteron tip and the animal pole, their well-developed pseudopodia and the accelerated indentation rate after the development of pseudopodia. Alternative mechanisms have not been completely eliminated.

In the present study, using Anthocidaris embryos with small blastocoel cavities, we showed that furrows are observed twice on the outer surface (hyaline layer side) of embryos; the first occurrence is at the time of primary mesenchyme cell emigration and the second is at the onset of indentation of archenteron rudiment. The occurrence of furrows on the outer surface of the vegetal pole region seems to suggest that the reduction of intercellular adhesion extends even up to the hyaline layer in contrast to the earlier hypothesis (Gustafson & Wolpert, 1963).

The reduction of intercellular adhesion in the vegetal pole region may allow the deformation of cells (appearance of tadpole-shaped cells) and following cellular emigration (primary mesenchyme cell formation) or sliding (first sign of archenteron indentation, Fig. 6B). It should be noted that the furrow formation at the onset of gastrulation appears to be only temporary, and the intercellular adhesion appears to be reestablished already in the early gastrula (Fig. 1 A, B).

The observations on the Anthocidaris embryos question the positive role of pseudopodial contraction in the further elongation of archenteron. The number of pseudopodia was much smaller as compared to Pseudocentrotus (Akasaka et al. 1980). The pseudopodial linkage between the animal pole and the archenteron tip was observed only at stages later than the mid-gastrula, and the distance between the animal pole and the tip of archenteron did not change so much in gastrulation as in Pseudocentrotus. Moreover the rate of archenteron elongation appeared to be constant (about 4 ·6 μm/h) as roughly calculated from the figures at various stages (Figs. 6 B, 7B, 8, 9). Even though the pseudopodia may not contribute to the forces for archenteron elongation, the role of pseudopodia as a specific linker between the archenteron tip and the animal pole seems likely in Anthocidaris.

The shortening of columnar cells in the ectodermal wall in gastrulation was observed much more markedly in the Anthocidaris as compared to Pseudocentrotus embryos (Akasaka et al. 1980). The shortening of columnar cells may give rise to the enlargement of the outer surface of embryos, and henceforth the blastocoel cavity. In the case of Anthocidaris, the enlargement of embryonic size proceeds coherently with the elongation of the archenteron into the blastocoel cavity, so that the blastocoel always remains small. However, the number of ectodermal wall cells on the dissected surfaces of embryos at various stages (Figs. 1, 8, 9) remained almost unchanged, arguing against the possibility that a portion of columnar cells in the ectodermal wall near the vegetal pole may continuously be translocated into a developing archenteron.

The bud-like structure of the archenteron tip in the mid-gastrula (Fig. 8) as well as the increase in the number of archenteron cells (in contrast to the constant number of ectodermal wall cells) appears to suggest that the proliferation of tip cells may be involved in the gastrulation process. The elongation of the archenteron after the mid-gastrular stage appeared to be due to the development and rearrangement of archenteron cells not to increase in cell number.

Whether or not some of these findings in Anthocidaris (for instance, the occurrence of furrows or apparent constancy of columnar cells) may be true for other sea-urchin embryos with larger blastocoel cavities is under investigation.