INTRODUCTION

Gastrulation in sea urchin embryos has been reported to consist of two different phases (Dan and Okazaki, 1956; Gustafson and Kinnander, 1956): primary invagination (autonomous buckling of the vegetal plate) and secondary invagination. Secondary invagination can be further divided into two steps. In the first step, the archenteron elongates via a rearrangement of the cells in the rudimentary gut that was formed by the primary invagination (Ettensohn, 1985; Hardin and Cheng, 1986). In the second step, the secondary mesenchyme cells (SMCs) raise the gut by the filopodial constriction to complete gastrulation (Hardin, 1988). In this mode of gastrulation, no new cells are added to the arch-enteron after the primary invagination is completed (Kominami and Masui, 1996).

Although this gastrulation process has been observed in numerous echinoid species, a second type of gastrulation has been reported in the sand dollar, S. mirabilis. Gastrulation in this organism does not appear to have discernible phases (Kominami and Masui, 1996; Kominami and Takata, 2000). Convolution of the cells near the blastopore persists, even during the late gastrula stage.

At the 64-cell stage of the normal development of echinoid embryos, the rudimentary archenteron consists of descendants of part of the veg1 layer and all of the veg2 layer (Ruffins and Ettensohn, 1996; Logan and McClay, 1997). Kominami and Masui (1996) suggested that more cells derived from the veg1 layer contribute to the completion of the archenteron in S. mirabilis embryos than in H. pulcherrimus embryos, which exhibit the typical gastrulation process. Kominami and Masui (1996) also indicated that the number of cells in rudimentary gut was much higher in S. mirabilis embryos than in H. pulcherrimus embryos.

It has been hypothesized that a cascade of signals originating in the micromeres produced at the 16-cell stage flows from the vegetal to the animal direction for later cell specification (Davidson, 1989), and various experiments have supported this concept. Pluteus larvae develop from chimeric embryos consisting of micromeres and presumptive ectoderm (animal cap) (Horstadius, 1973; Amemiya, 1996; Minokawa et al., 1997; Minokawa and Amemiya, 1998). Both ectopic SMCs (McClay et al., 2000) and the secondary archenteron (Ransick and Davidson, 1993) have been found to be induced when exogenous micromeres are transplanted, indicating that the micromeres emit the inductive signal, although the timing of the micromere signal for arch-enteron induction might be much later than the 16-cell stage (Minokawa and Amemiya, 1999; Ishizuka et al., 2001). Micromere deletion at the 16-cell stage results in considerable reduction of the expression of the vegetal plate specific gene Endo16 (Ransick and Davidson, 1995) and in no specification of SMCs (Sweet et al., 1999; McClay et al., 2000).

At the 6th cleavage, macromere descendants divide horizontally to form veg1 and veg2 layers. The inductive effect of veg2 cells is necessary for part of the veg1 cells to differentiate into endoderm (Logan and McClay, 1999). In the present study, we investigated the inductive activity of the veg2 layers in H. pulcherrimus and S. mirabilis embryos as a means of determining whether the difference in the activity of the inductive effects of the veg2 layers is connected to the different modes of invagination. Our results showed that the ability of the veg2 layer to induce endoderm differentiation in the animal-cap cells of S. mirabilis embryos is greater than in H. pulcherrimus embryos.

MATERIALS AND METHODS

RESULTS AND DISCUSSION

Animal halves of sea urchin embryos isolated during the early cleavage stage differentiate exclusively into ectoderm to form a permanent blastula. However, the animal halves also have the potential to develop into pluteus-like larvae under the inductive influence of other blastomeres that have been recombined (Horstadius, 1973; Amemiya, 1996; Logan and McClay, 1999). We produced recombinant embryos [A+V2(8)] (Fig. 1) consisting of an animal half (A) and a veg2 layer (V2) that contained eight blastomeres (8) to determine the inductive potential of the veg2 layer. We judged the strength of the inductive effect of veg2 cells, by how many parts of the tripartite digestive tract originated from the animal-half cells. As the inductive effect increased in this system, cells derived from animal cap to have contributed to more parts of the tract could be assumed.

Fig. 1

Experimental procedure for producing recombinant embryos. The diagram shows the recombination of the animal cap with the veg2 layer (A). Bright field images (B, D) and fluorescent images (C, E) of the animal cap (arrowheads) and labeled veg2 layer (arrows) isolated from S. mirabilis embryos. Bar, 100 μm

All of the [A+V2(8)] embryos in H. pulcherrimus developed into pluteus larvae (25/25) with a differentiated digestive tract (Fig. 2A,B). The labeled cells showed that the foregut and midgut were composed of veg2 descendants, whereas the hindgut did not contain any labeled cells in nearly half of the recombinants (Table 1). In this half of the recombinants, the hindgut had differentiated from the animal-half cells under the inductive influence of the veg2 cells (Fig. 2A,B, double arrowheads). In the other recombinants, however, the labeled veg2 cells prevailed throughout the entire tripartite digestive tract, indicating that induction had not occurred in these cases (Table 1). In S. mirabilis, a number of [A+V2(8)] embryos also developed into pluteus larvae (22/29) with a differentiated digestive tract (Fig. 2C,D). In almost all of these recombinants, only the foregut (Fig. 2C,D, arrowheads) was composed of the veg2-descendant cells, with the midgut (Fig. 2C,D, arrows) and hindgut (Fig. 2C,D, out of focus) being derived from the animal-half cells (Table 1).

Fig. 2

Larvae with a digestive tract derived from recombinant embryos [A+V2(8)] in H. pulcherrimus (A, B) and S. mirabilis (C, D) two days after fertilization. A and C show bright field images, viewed from the left and dorsal sides, respectively. B and D show fluorescent images corresponding to A and B, respectively. Arrowheads, arrows and double arrowheads point to the foregut, midgut and hindgut, respectively. Bar, 100 μm

Table 1

The effect of veg2 layer on the induction of endoderm differentiation in animal-half cells. Each number is shown with percentage in parentheses

Alkaline phosphatase (APase) activity was examined to confirm whether the animal-half cells had functionally differentiated into a digestive tract. In both species, APase activity was expressed in the midgut and hindgut of the digestive tract derived from the animal-half cells in the recombinant larvae (Fig. 3), as in the normal larvae (Amemiya, 1996; Minokawa et al., 1997).

Fig. 3

Alkaline phosphatase (APase) activity in larvae derived from recombinant embryos [A+V2(8)], viewed from the left side. APase activity is detected in the induced hindgut of H. pulcherrimus (A) and the induced mid- and hindgut of S. mirabilis (B). Arrows and arrowheads point to the midgut and hindgut, respectively. Bar, 100 μm

The midgut and hindgut of S. mirabilis were entirely derived from the animal-half cells in all [A+V2(8)] chimeric embryos that developed the archenteron. In H. pulcherrimus, however, less than half of the chimeric embryos exhibited animal-half cells, and only in a limited region of the hindgut. These results indicate that the inductive potency of the veg2 layer in S. mirabilis is higher than in H. pulcherrimus. As described in the introduction, the mode of gastrulation in S. mirabilis embryos is different from that in H. pulcherrimus (Kominami and Masui, 1996). The number of cells in the archenteron in S. mirabilis embryos increased as gastrulation proceeded, whereas the numbers were almost constant in H. pulcherrimus. On the basis of the results obtained in the present study, the following appear to be possible explanations for the difference in modes of gastrulation in these two species. The inductive activity of the veg2 layer or the competence of the animal side cells in H. pulcherrimus embryos is weaker than in S. mirabilis, so that fewer cells in the animal side layer in H. pulcherrimus embryos are specified to endoderm than in S. mirabilis. It is also possible that the inductive signal of micromeres drives that of the veg2 layer, as reported by Davidson (1989) and Davidson et al. (1998). The difference in the intensity of the inductive effect of the veg2 layers may be associated with the difference in that of the micromeres between the two species. We are now producing the interspecies chimeric embryos derived from an animal cap isolated from one species recombined with a veg2 layer or a quartet of micromeres isolated from the other in order to investigate the above possibilities.

Next, we produced several types of chimeric embryos [A+V2(x)] consisting of an animal cap (A) recombined with four, two, and one blastomere(s) (number, x) of the veg2 layer (V2) in H. pulcherrimus. More anterior fractions of the digestive tract were induced to differentiate from the animal-half cells when fewer numbers of veg2 blastomeres were recombined (Fig. 4A), indicating that the inductive effect of the veg2 blastomeres rose as the number of veg2 blastomeres decreased. Chimeric S. mirabilis embryos [A+V2(1)] consisting of an animal cap (A) recombined with just one (1) of the veg2 (V2) blastomeres were produced, and the inductive potential of the veg2 blastomere was found to be higher than that of the veg2 layer (Fig. 4B). These results show that the inductive activity of veg2 blastomere(s) increased as the number of veg2 blastomeres decreased. The mechanism(s) responsible for the decreased number of veg2 blastomeres exerting stronger inductive activity than the increased number of blastomeres is still unknown. One possibility is that a lateral interaction between veg2 blastomeres inhibits the inductive activity of neighboring cells, as reported by Henry et al. (1989) in regard to the mesomeres.

Fig. 4

Difference in inductive effects related to variations in number of veg2 layer blastomeres of [A+V2(x)] embryos in H. pulcherrimus (A) and S. mirabilis (B). The induced region derived from animal-half cells within the digestive tract is shown diagrammatically under each group of columns. The parts of labeled veg2 cells are painted over. Arrows, arrowheads, and double arrowheads point to the hindgut, midgut, and foregut, respectively, of the induced animal-half cells.

* The labeled veg2 cells prevailed through all parts of the tripartite gut, so that no induction of veg2 layer blastomeres appeared to have occurred.

** Archenteron failed to segment into a tripartite tract, resulting in the formation of an incomplete gut.