- Email: [email protected]
Conserved Mechanism of Dorsoventral Axis Determination in Equal-Cleaving Spiralians
Developmental Biology 248, 343–355 (2002) doi:10.1006/dbio.2002.0741
Conserved Mechanism of Dorsoventral Axis Determination in Equal-Cleaving Spiralians Jonathan J. Henry 1 Marine Biological Laboratory, Woods Hole, Massachusetts 02543
Many members of the spiralian phyla (i.e., annelids, echiurans, vestimentiferans, molluscs, sipunculids, nemerteans, polyclad turbellarians, gnathostomulids, mesozoans) exhibit early, equal cleavage divisions. In the case of the equal-cleaving molluscs, animal–vegetal inductive interactions between the derivatives of the first quartet micromeres and the vegetal macromeres specify which macromere becomes the 3D cell during the interval between fifth and sixth cleavage. The 3D macromere serves as a dorsal organizer and gives rise to the 4d mesentoblast. Even though it has been argued that this situation represents the ancestral condition among the Spiralia, these inductive events have only been documented in equal-cleaving molluscs. Embryos of the nemertean Cerebratulus lacteus also undergo equal, spiral cleavage, and the fate map of these embryos is similar to that of other spiralians. The role of animal first quartet micromeres in the establishment of the dorsal (D) cell quadrant was examined in C. lacteus by removing specific combinations of micromeres at the eight-cell stage. To follow the development of various cell quadrants, one quadrant was labeled with DiI at the four-cell stage, and specific first quartet micromeres were removed from discrete positions relative to the location of the labeled quadrant. The results indicate that the first quartet is required for normal development, as removal of all four micromeres prevented dorsoventral axis formation. In most cases, when either one or two adjacent first quartet micromeres were removed from one side of the embryo, the cell quadrant on the opposite side, with its macromere centered under the greatest number of the remaining animal micromeres, ultimately became the D quadrant. Twins containing duplicated dorsoventral axes were generated by removal of two opposing first quartet micromeres. Thus, any cell quadrant can become the D quadrant, and the dorsoventral axis is established after the eight-cell stage. While it is not yet clear exactly when key inductive interactions take place that establish the D quadrant in C. lacteus, contacts between the progeny of animal micromeres and vegetal macromeres are established during the interval between the fifth and sixth round of cleavage divisions (i.e., 32- to 64-cell stages). These findings argue that this mechanism of cell and axis determination has been conserved among equal-cleaving spiralians. © 2002 Elsevier Science (USA) Key Words: induction; cell interactions; unequal cleavage; Nemertea.
INTRODUCTION Spiralian embryos (including polychaete annelids, vestimentiferans, echiurids, molluscs, nemerteans, sipunculids, polyclad turbellarians, mesozoans, and gnathostomulids) exhibit a conserved pattern of quartet spiral cleavage in which the first two cell divisions lead to the formation of four basic cell quadrants, generally referred to as the A (left), B (ventral), C (right), and D (dorsal) quadrants (Figs. 1A and 1B). Subsequent divisions generate additional tiers of animal micromere quartets. Ultimately, specific cell fates and 1 Present address: University of Illinois, Department of Cell and Structural Biology, 601 S. Goodwin Avenue, Urbana, IL 61801. Fax: (217) 244-1648. E-mail: [email protected].
0012-1606/02 $35.00 © 2002 Elsevier Science (USA) All rights reserved.
the dorsoventral axis are determined via inductive interactions from the D quadrant, which serves as a dorsal organizer (Clement, 1962; Cather and Verdonk, 1979; Verdonk and Cather, 1983; van den Biggelaar and Guerrier, 1983; Verdonk and van den Biggelaar, 1983; Martindale, 1986). There are two principle mechanisms that operate to establish the D quadrant in spiralian embryos. In some species, the early cleavage divisions are unequal and this results in the differential partitioning of vegetally localized factors to the larger CD and D blastomeres during the first and second cleavage divisions, respectively. This can take place via shifting of the cleavage spindle, the formation of transient cytoplasmic protrusions called polar lobes, or a combination of these events (Henry, 1986; Boyer and Henry 1998; Henry and Martindale, 1999). Various experiments indicate
343
344
Jonathan J. Henry
FIG. 1. Diagram showing the patterns of early cleavage divisions and their relationships relative to the dorsoventral axis in spiralians. All views are from the animal pole. The first and second cleavage planes are shown, as labeled. (A) Typical spiralian pattern of equal cleavage leading to the eight-cell stage. Note the presence of the vegetal cross-furrow separating the future B and D blastomeres at the vegetal pole (dotted lines). Later during development, this packing arrangement predisposes these two cells toward the center of the embryo where they can make contacts with the progeny of the animal first quartet micromeres. The third order macromere derived from one of these two cells is ultimately induced to become the 3D macromere. (B) Oblique axial relationship between the early cleavage planes and the dorsoventral (D-V) axis in most spiralians (e.g., molluscs) in which the first quartet is formed in a dextral fashion as shown in (A). Note that in these cases the dorsoventral axis is always skewed 45 degrees counterclockwise relative to the first cleavage plane. (C) Condition seen in nemerteans, like C. lacteus and N. bivittata. Note that in these embryos there is no vegetal cross-furrow and hence each quadrant has an equal chance of becoming the D quadrant. Therefore, the dorsoventral axis can be oriented in either an oblique counterclockwise (50% of the cases) or clockwise (50% of the cases) direction relative to the first cleavage plane (Henry and Martindale, 1994, 1998). (D) Diagram depicting the anatomy of the helmet-shaped, nemertean, pilidium larva with its two ciliated lappets or flaps. Larval views and structures are as indicated. The dorsoventral (D-V), anteroposterior (A-P), and left–right (L-R) axes are also shown.
that these unequal divisions segregate key vegetal factors that impart the D quadrant cell with its unique role as an inductive organizer (Cather and Verdonk, 1979; Verdonk and Cather, 1983; van den Biggelaar and Guerrier, 1983; Render, 1983, 1989; Henry, 1986, 1989; Henry and Martindale, 1987; Dorresteijn et al., 1987). These events have been well documented in certain species of molluscs (e.g., Ily-
anassa obsoleta, Render, 1989; Dentalium vulgare, Guerrier et al., 1978; Pholas dactyluss and Spisula subtruncata, Guerrier, 1970b), as well as annelids (e.g., Sabellaria cementarium, Render, 1983; S. alveolata, Novikoff, 1940, Guerrier, 1970a; Platynereis dumerilii, Dorresteijn et al., 1987; Chaetopterus variopedatus, Tyler, 1930, Henry and Martindale, 1987).
© 2002 Elsevier Science (USA). All rights reserved.
Dorsoventral Axis Determination in Spiralians
345
FIG. 2. Various partial trees showing certain phylogenetic relationships amongst different spiralian phyla, classes, and orders, and the distribution of quartet spiral cleavage types, in which the first two divisions are known to be equal (E) vs unequal (U), based on morphological descriptions of cleavage patterns. Branches are labeled with various classes, orders, and suborders where specific data are available. Phyla are listed to the far right side of the figure. Astrices indicate that unequal cleavages involving the formation of polar lobes have been reported for some species. Diagram is expanded from that of Freeman and Lundelius (1992). Phylogenetic relationships are from Freeman and Lundelius (1992), Trumen and Clarke (1985), Fauvel et al. (1975), Fauchald (1975), and Salvini-Plawe´ n (1985). Dotted lines labeled 1 and 2 indicate possible alternative phylogenies within the Bivalvia (see Freeman and Lundelius, 1992), while 3 and 4 are suggested by the work of McHugh (1997).
© 2002 Elsevier Science (USA). All rights reserved.
346
Jonathan J. Henry
FIG. 3. Photomicrographs depicting typical development of vegetal half-embryos prepared by removal of the first quartet of micromeres at the eight-cell stage. This example is 72 h old. (A). The partial embryo appears to consist of an outer covering of ciliated ectoderm surrounding an inner mass of endodermal tissue. (B) Fluorescence micrograph showing presence of disorganized muscle cells stained with Bodipy–phallacidin. ec, ectoderm; en, endoderm; mf, muscle fibers. Scale bar, 20 m.
In other species, the early cleavage divisions are equal and all four blastomeres are the same size at the four-cell stage. In these forms, the D quadrant is actually specified later during development as a result of animal–vegetal inductive interactions (van den Biggelaar and Guerrier, 1979; Arnolds et al., 1983; Martindale et al., 1985; van den Biggelaar, 1996). These inductive interactions take place between the fifth and sixth cleavage divisions when the progeny of the animal first quartet micromeres contact one of the macromeres, which is induced to become the D quadrant macromere (3D). Normally, this is one of the two more centrally located vegetal cross-furrow macromeres; however, experiments indicate that any one of the four cell quadrants can be selected as the D quadrant if these interactions are shifted by ablating individual first quartet micromeres (van den Biggelaar and Guerrier, 1979; Arnolds et al., 1983). Furthermore, if these contacts are prevented by certain chemical treatments, the D quadrant is not induced, and the embryos develop in a radialized fashion (Martindale et al., 1985; Kuhtreiber et al., 1988). The existence of this second mechanism has only been demonstrated in equal-cleaving molluscs [e.g., Patella vulgata, Lymnaea stagnalis, L. palustris, van den Biggelaar and Guerrier, 1979; Arnolds et al., 1983; Martindale et al., 1985; Acanthochiton crinitus, van den Biggelaar, 1996; and the opistobranch Haminoea (callidegenita) vesicula, Boring, 1986]. A survey of cleavage types amongst the Spiralia indicates that equal cleavage is widely represented and generally present in the most basal groups (Fig. 2). On this basis, Freeman and Lundelius (1992; see also van den Biggelaar, 1996) argued that equal cleavage, and hence animal–vegetal inductive selection of the D quadrant, represents the ancestral condition among the spiralians. Verification of this assumption requires further phylogenetic analyses and experimental examinations to determine whether the same mechanism of cell fate and axis determination operates in other equal-cleaving spiralian phyla.
FIG. 4. Diagram showing the distributions of labeled cell quadrants following random injection of DiI into single blastomeres at the four-cell stage. DiI label is shown in red. Data in the upper right corner show distribution of cases in unoperated embryos, while those below show distribution of labeled quadrants following removal of one micromere at the eight-cell stage from specific positions relative to the single labeled quadrant (as shown). For each of the four different types of deletions, predicted results are shown to the left, as far as which cell quadrants are most likely to be induced to become the D quadrant. These predictions are based on the arrangement of cell contacts following removal of individual micromeres, which results in shifting the distribution of potential animal–vegetal cell contacts toward the two cell quadrants that lie opposite the site of ablation. Each of these two macromeres lies in contact with two of the remaining first quartet micromeres, while the other macromeres only maintain contacts with a single micromere. As there is no vegetal cross-furrow, one can predict that either of those two quadrants, with the maximal number of animal cell contacts, should be equally disposed to become the D quadrant.
© 2002 Elsevier Science (USA). All rights reserved.
Dorsoventral Axis Determination in Spiralians
347
FIG. 5. Paired DIC and fluorescence light micrographs showing the distribution of labeled progeny in embryos in which one of the first quartet micromeres had been removed at the eight-cell stage. These cells were removed from specific locations relative to the single labeled cell quadrant, as shown in Fig. 4. Larval anatomy is diagramed in Fig. 1D. (A–D) One first quartet micromere was removed from the labeled cell quadrant (deletion #1 in Fig. 4). Note the absence of label in the apical (animal) ectoderm and the apical organ. (A, B) Typical A quadrant labeling pattern (macromere 1A) with label in the gut, esophagus, left ciliary band, and ventral ectoderm associated with the left lappet. Label is also seen in some muscle fibers. (C, D) Typical B quadrant labeling pattern (macromere 1B) with label in the gut, esophagus, right ciliary band, and ventral ectoderm associated with the right lappet. Label is also seen in some muscle fibers. (E–H) One of the first quartet micromeres was removed from the quadrant located just counter-clockwise to the labeled quadrant (deletion #2 in Fig. 4). (E, F) Typical B quadrant labeling pattern with label in the apical organ, the gut, esophagus, right ciliary band, and ventral ectoderm, including part of the right lappet. Label is also seen in some muscle fibers. (G, H) Typical C quadrant labeling pattern with label in the apical organ, gut, esophagus, right ciliary band, and ventral ectoderm, including part of the right lappet. No label is seen in mesodermal cells. (I–L) One of the first quartet micromeres was removed from the quadrant located just clockwise to the labeled quadrant (deletion #3 in Fig. 4). (I, J) Typical A quadrant labeling pattern with label in the apical organ, gut, esophagus, left ciliary band, and ventral ectoderm, including part of the left lappet. Label is also seen in some muscle fibers. (K, L) Typical D quadrant labeling pattern with label in the apical organ, gut, left ciliary band, and dorsal ectoderm, including part of the left lappet. The mesentoblast bands are also labeled. (M–P) One of the first quartet micromeres was removed from the quadrant located opposite the labeled quadrant (deletion #4 in Fig. 4). (M, N) Typical C quadrant
© 2002 Elsevier Science (USA). All rights reserved.
348
Jonathan J. Henry
In the present study, experiments were undertaken to determine whether the first quartet and animal–vegetal inductive interactions play an important role in establishing the D quadrant in the equal-cleaving nemertean, Cerebratulus lacteus. The results indicate that the animal first quartet micromeres are essential for the establishment of the dorsoventral axis, and furthermore that these cells serve to specify the D quadrant. In addition, the central open cleavage cavity or blastocoel, which is present in early embryos, is obliterated during the interval between fifth and sixth cleavages (32- to 64-cell stages) when cell contacts are established between animal micromeres and vegetal macromeres. These findings are similar to those seen in equal-cleaving molluscs and suggest that similar mechanisms are employed in establishing the D quadrant in other equal-cleaving spiralians, such as the Nemertea.
MATERIALS AND METHODS Collection of Adults and Embryos Adult specimens of the nemertean worm C. lacteus were obtained by the Marine Resources Department of the Marine Biological Laboratory (Woods Hole, MA). Gametes were prepared as described by Martindale and Henry (1995).
Cell Lineage Analysis via Microinjected Fluorescent Tracer Single blastomeres were injected at the four-cell stage with DiI (Molecular Probes Inc., Eugene, OR) dissolved in vegetable oil (Terasaki and Jaffe, 1991) following the procedure of Henry and Martindale (1998). Embryos that survived microinjection continued to cleave and developed normally, at the same rate as uninjected controls. Following DiI labeling, specific first quartet micromeres were removed by using a fine glass needle at the eight-cell stage. The visible presence of the tiny intracellular DiI oil droplets and the use of a Leica epifluorescence dissecting microscope permitted the positive identification of the labeled cell quadrant, which served as a frame of reference for the ablation of specific cells. In addition, one can easily distinguish the animal first quartet micromeres in nemertean embryos since these cells are larger than the corresponding vegetal macromeres, and visible polar bodies remained attached at the animal pole in the denuded embryos at the eight-cell stage.
Examination of Internal Cell Contacts during Early Development To examine the formation of the cleavage cavity and subsequent cell contacts between animal micromeres and vegetal macromeres,
embryos were fixed at 30- to 42-min intervals between the 2-cell stage and the 256-cell stage in 4% formalin in sea water overnight at 4°C. These samples were washed in sea water and cleared in either 80% glycerin in sea water or a 1:2 mixture of benzyl alcohol and benzyl benzoate following a graded series of washes to 100% methanol (Klymkowsky and Hanken, 1991).
Culture and Preparation of Embryonic and Larval Specimens Embryos were raised at 18 –20°C in 0.22 m filtered sea water (FSW) for a period of 48 –72 h. Live pilidium larvae were examined and photographed following slight compression between Rain-Xcoated (Blue Coral-Slick 50, Ltd., Cleveland, OH) slides and coverslips supported by clay feet. Since the larvae are highly transparent, all the internal cell types could be clearly visualized in these whole mounts. Fluorescence staining of filamentous actin in muscle cells was accomplished in some fixed samples by using Bodipy– phallacidin (Molecular Probes) following the protocol of Martindale and Henry (1995). Images were captured by using a Nikon fluorescence light microscope and a Spot digital camera (Diagnostic Images, MI).
RESULTS Fate Maps of the Blastomeres within the Four-CellStage Embryo The embryonic fate map and larval anatomy of C. lacteus has already been published (Henry and Martindale, 1998; see also Fig. 1D). By labeling single blastomeres at the four-cell stage, one can utilize a very clear suite of cell fates to assign quadrant identities in addition to the characteristic dorsoventral and bilateral placements of their labeled ectodermal progeny. Only the A and B quadrants generate larval muscle cells (derived form the 3a and 3b micromeres, respectively), whereas the D quadrant forms the discrete pair of left and right mesodermal or mesentoblast bands (derived from the 4d micromere). Finally, the C quadrant does not generate any mesodermal cells. To determine a baseline for the frequency of the four cell quadrants in randomly labeled embryos in the present study, individual cells were injected with DiI at the four-cell stage. A total of 90 cases were examined. The relative frequencies of labeled quadrants in these injected control embryos was found to be: A, 30%; B, 27%; C, 18%; and D, 25% (also recorded in Figs. 4 and 6 for reference).
labeling pattern with label in the apical organ, gut, esophagus, right ciliary band, and ventral ectoderm, including part of the right lappet. No label is also seen in mesodermal cells. (O, P) Typical D quadrant labeling pattern with label in the apical organ, gut, left ciliary band, and dorsal ectoderm, including part of the left lappet. The mesentoblast bands are also labeled. (A, B) (I–L), and (O, P) are left-side views. (C–H) and (M, N) are right-side views. ao, apical organ; gt, stomach; ld, DiI lipid drop; llp, left lappet; mb, mesentoblast bands; mf, muscle fibers; rlp, right lappet. Scale bar, 50 m.
© 2002 Elsevier Science (USA). All rights reserved.
349
Dorsoventral Axis Determination in Spiralians
Removal of the Entire First Quartet at the EightCell Stage Previously, Freeman (1978) (see also Zeleny, 1904; Yatsu, 1910; Ho¨ rstadius 1937, 1971) separated animal and vegetal halves of C. lacteus embryos at the 8-cell stage to assess the developmental potential of these fragments. In his study, Freeman was concerned with the mechanisms that lead to the determination of cell fates along the animal–vegetal axis (e.g., the ectodermal apical organ and the endodermal stomach); hence, Freeman did not specifically examine the differentiation of dorsoventral polarity or cell fates. As discussed above, recent cell lineage analyses indicate that certain cell fates arise differentially along the dorsoventral axis, such as the larval muscles derived from the ventrally situated 3a and 3b micromeres, and the dorsal 4d-derived mesentoblast bands (Henry and Martindale, 1998). To investigate the potential role of the first quartet micromeres and animal–vegetal interactions in the establishment of dorsoventral polarity, a total of 21 vegetal halves were prepared at the 8-cell stage. In no case was any overt sign of dorsoventral polarity seen. The partial embryos developed in a radialized fashion, consisting of an outer ciliated ectodermal layer and a central endodermal mass, which in some cases protruded outside of the ectodermal covering (Fig. 3A). Although the outer ectoderm was ciliated, an apical tuft of elongated cilia was not formed, just as reported by Freeman (1978). It was not possible to visualize whether mesentoblast bands were formed in these cases; however, larval muscle cells could be readily identified. Twelve of the 21 cases were stained with Bodipy– phallacidin to examine larval muscle cell development. In 10 cases, abundant disorganized muscle fibers were formed (see Fig. 3B), which exhibited no signs of bilateral placement. The other two cases did not contain any labeled muscle fibers.
Ablation of Single Micromeres at the Eight-Cell Stage Individual micromeres were removed at the eight-cell stage in each of the four quadrants in specific positions relative to the single labeled quadrant (see Fig. 4). The contribution of the cells derived from the labeled quadrant was then ascertained in the pilidium larvae after 48 –72 h of development (Fig. 5). In general, these larvae appeared to be completely normal with no detectable defects. Normal symmetry properties and differentiated cell fates were evident in these larvae. The frequency distributions of the labeled quadrants are shown in Fig. 4 for each of the four types of deletions. In those cases where the labeled micromere was ablated (deletion #1; see also Figs. 5A–5D), there was no labeled ectoderm in the animal hemisphere or the apical tuft (normally all four of the first quartet micromeres generate the apical tuft; see Henry and Martindale, 1999); however, the remaining labeled vegetal progeny were sufficient to clearly ascertain the labeled quadrants’ identities. The results from these experiments (Fig. 4) indicate that
there is a strong bias in the distribution of labeled cell quadrants relative to the location of the ablated micromere.
Ablation of Two Adjacent Micromeres at the Eight-Cell Stage Further experiments were undertaken to shift inductive interactions mainly to a single cell quadrant by removing two adjacent micromeres at the eight-cell stage (see Figs. 6 and 7). Three of the four possible combinations of cell ablations were performed (deletions #5–7). A single case was also completed for the fourth type of deletion (deletion #8). In most cases, the macromere located under the two remaining micromeres became the D quadrant macromere (74 –95% of the cases). The distribution of these cases is shown in Fig. 6. In many cases, the clone of labeled animal ectodermal tissue extended over a larger surface area relative to that which would have been generated in intact embryos (Figs. 7B and 7D). Apparently these cells had compensated to generate the missing ectodermal tissues. The combination of the ectodermal labeling patterns and the differentiation of mesodermal cell fates, including larval muscle cells and the mesentoblast bands, were sufficient to ascribe the identity of the labeled cell quadrants.
Ablation of Two Opposing Micromeres at the Eight-Cell Stage The results reported above indicate that the animal micromeres play a role in the determination of the D quadrant. An attempt was made to generate twin embryos, containing duplicated dorsoventral axes, by separating inductive cell contacts toward opposite sides of the embryo. To test this possibility, two of the first quartet micromeres that lie opposite one another were removed at the eight-cell stage (Fig. 8). In most cases (82%, n ⫽ 45), the resulting pilidium larvae appeared to be fairly normal, displaying normal symmetry properties, including a single dorsoventral axis. In a few cases, however, twins were formed that contained two digestive tracts (9%; see Figs. 9A and 9B). Other cases exhibited a radialized phenotype with no clear dorsoventral asymmetry and no obvious mesentoblast bands (9%; see Fig. 9C).
Observations of Animal–Vegetal Cell Contacts during the Early Cleavage Stages Beginning at the 8-cell stage, a cleavage cavity can be visualized in the center of the embryo. This cavity or blastocoel persists through the 32-cell stage (Fig. 10A). After this stage, however, this cavity is obliterated as cells come in contact at the center of the embryo. In most cases, this cavity is no longer apparent at the 64-cell stage (Fig. 10B), when contacts between the animal micromeres and the vegetal macromeres could be observed.
© 2002 Elsevier Science (USA). All rights reserved.
350
Jonathan J. Henry
FIG. 6. Diagram showing the distributions of labeled cell quadrants following random injection of DiI into single blastomeres at the four-cell stage. DiI label is shown in red. Data in the upper right corner show distribution of cases in unoperated embryos, while those below show distribution of labeled quadrants following removal of two adjacent micromeres at the eight-cell stage from specific locations relative to the single labeled quadrant. For each of the different types of deletions, predicted results are shown to the left, as far as which cell quadrant is likely to be induced to become the D quadrant. These predictions are based on the arrangement of cell contacts following removal of two adjacent micromeres, which result in shifting the distribution of potential animal–vegetal cell contacts toward the cell quadrant that lies opposite the site of ablation. This is the only macromere that lies in contact with both of the remaining first quartet micromeres, and thus, should become the D quadrant in the majority of cases.
FIG. 7. Paired DIC and fluorescence light micrographs showing the distribution of labeled progeny in embryos in which pairs of adjacent first quartet micromeres had been removed at the eightcell stage. These cells were removed from specific locations relative to the single labeled cell quadrant, as shown in Fig. 6. (A, B) Left-side view of an example in which two first quartet micromeres were removed just clockwise to the labeled cell quadrant (deletion #5 in Fig. 6). Typical D quadrant labeling pattern with label in the gut, left ciliary band, and ventral ectoderm associated with the left lappet. Most of the left side of the larva is labeled. The mesentoblast bands are also labeled. (C, D) Case in which two first quartet micromeres were removed just counter-clockwise to the labeled cell quadrant (deletion #6 in Fig. 6). Typical C quadrant labeling pattern with label in the gut, esophagus, right ciliary band, and ventral ectoderm, including the right lappet. Most of the right side of the larva is labeled. No label is seen in the mesoderm. Larva is viewed from the right side and tilted somewhat to show the apical surface. (E, F) Right-side view of a case in which the labeled first quartet micromere was removed in addition to the one located counterclockwise to that cell (deletion #7 in Fig. 6). Typical B (1B) quadrant labeling pattern with label in the gut, esophagus, right ciliary band, and ventral ectoderm, including the right lappet. No label is present in the ectoderm (or the apical organ) except for a portion of the esophagus on the right side. DiI label is seen in numerous muscle cells. Labels are the same as those used in Fig. 5. Scale bar, 50 m.
© 2002 Elsevier Science (USA). All rights reserved.
351
Dorsoventral Axis Determination in Spiralians
FIG. 8. Diagram depicting deletions (“X”) of two opposing micromeres at the eight-cell stage in an attempt to segregate animal– vegetal inductive interactions that result in the induction of two D quadrants or “twins.” Predictions are made showing possible types of twins that might be achieved. “1D and 1d” signifies that these cells represent future D quadrant cells, while “1Q and 1q” signifies that these cells are fated to become other quadrant cells. Below is shown the actual distribution of cases observed in this experiment.
DISCUSSION Animal First Quartet Micromeres Play a Role in D Quadrant Specification The results indicate that the animal micromeres play an important role in the establishment of the dorsal
quadrant in C. lacteus. Removal of the first quartet prevents the formation of the dorsoventral axis. Likewise, one can consistently influence which cell quadrant becomes the D quadrant by altering the arrangement of the animal micromeres. This takes place in a highly predictable fashion such that those quadrants with the greatest number of overlying animal first quartet micromeres ultimately become the D quadrant. When one animal micromere is removed, the three remaining micromeres are situated such that two of the four vegetal macromeres lie in contact with two micromeres, while the other two macromeres each lie in contact with only a single micromere (Fig. 4). In the majority of cases (69 – 78%), one of the two macromeres with two overlying first quartet micromeres ultimately becomes the D quadrant. When two adjacent micromeres are removed, the two remaining micromeres overlie a single macromere, and this cell ultimately becomes the D quadrant in 74 –95% of the cases (Fig. 6). The results reported here cannot be accounted for by variations in the distribution of randomly labeled cell quadrants, which might be associated with either sample size or biases imposed by the intracellular labeling technique. In every experiment, there was a strong correlation between the cell quadrant that ultimately differentiated as the D quadrant and the side of the embryo that contained the greatest concentration of animal first quartet derivatives. These results are similar to those obtained in experiments performed with equal-cleaving molluscs (van den Biggelaar, and Guerrier 1979; Arnolds et al., 1983), and they indicate that inductive interactions from the progeny of the animal micromeres play a key role in establishing the D quadrant in equal-cleaving spiralians.
FIG. 9. Photomicrographs showing twin and radialized larvae resulting from the deletion of two opposing micromeres at the eight-cell stage. (A) Twin larva containing two separate digestive tracts, each with a separate mouth, esophagus, and stomach. Dotted lines show the axes of the two parallel digestive tracts. View is from the oral surface with the shared dorsal side to the right. (B) Twin larva with two separate digestive tracts, which share a common mouth. Twin ciliated, esophageal thick tissue plates (et) are seen leading up to the two different stomachs located on the dorsal sides of the larva. The axes of these two digestive systems are located nearly at right angles to one another (dotted lines). In this case, the dorsal surfaces would be located to the right and top of the figure, respectively. Like the example shown in (A), one of these digestive tracts is smaller in size. (C) Radialized larva. The single digestive tract extends straight up from the mouth to the apical organ and tuft. Also note there are no distinct lappets and no ciliated, thick esophageal plate of tissue is present within the esophagus. at, apical tuft; st, stomodeum. Other labels are the same as those used in Fig. 5. Scale bar, 50 m.
© 2002 Elsevier Science (USA). All rights reserved.
352
Jonathan J. Henry
FIG. 10. DIC photomicrographs showing fixed 32- and 64-cell-stage embryos. Note the presence of an open central cleavage cavity or blastocoel in the 32-cell stage embryo shown in (A). This cavity is no longer present in the 64-cell-stage embryo when cell contacts appear to be established between the micromere progeny and the vegetal macromeres (B). cc, cleavage cavity. Scale bar, 20 m.
Spatial Constraints in D Quadrant Induction In molluscs, such as Lymnaea and Patella, the oblique packing arrangement of the blastomeres normally predisposes either of the two vegetal “cross-furrow” macromeres to become the D quadrant cell by virtue of their more central position within the embryo (relative to the two peripheral noncross furrow macromeres; van den Biggelaar and Guerrier, 1979, 1983; Verdonk and Cather, 1983; Verdonk and van den Biggelaar, 1983; Arnolds et al., 1983; Martindale et al., 1985; see Figs. 1A and 1B). Hence, these cells are better situated to establish cell contacts with the progeny of the first quartet micromeres. In these species, the dorsoventral axis bears only one strict axial relationship relative to the first cleavage plane (Fig. 1B). In the case of some nemerteans, like C. lacteus and Nemertopsis bivittata, there are no distinct cross-furrows in the early embryo (Henry and Martindale, 1994a, 1998), and these vegetal blastomeres lie in essentially a single plane. Hence any one of the four vegetal macromeres has an equal chance of becoming the D quadrant cell. This is reflected in the fact that the first cleavage plane can have one of two different axial relationships relative to the future dorsoventral axis (see Fig. 1C; Henry and Martindale, 1994a, 1998; note however, that some other species of nemerteans apparently do have cross-furrows; S. Maslakova and M. Q. Martindale, personal communication).
Plasticity and Regulative Capacity of EqualCleaving Spiralian Embryos Although any one of the vegetal macromeres can become a D quadrant blastomere in equal-cleaving molluscs, and animal–vegetal and inductive interactions are involved in specifying cell fates in an epigenetic fashion, blastomeres isolated from two- and four-cell stage embryos seldom
display any signs of regulative development (e.g., van den Biggelaar et al., 1981; Verdonk and Cather, 1983). Only in extremely rare cases where the pattern of cleavage divisions reiterates that of the preceding stages does regulation take place (Morrill et al., 1973; Verdonk, 1979; Verdonk and Cather, 1983). These observations argue that the spatial arrangement of the blastomeres is very important for normal development in those embryos. On the other hand, the embryos of C. lacteus freely exhibit regulative development when blastomeres are isolated at the two- or four-cell stages (Ho¨ rstadius, 1937, 1971; Henry and Martindale, 1994b, 1996; Martindale and Henry, 1995; note however, this is not the case in the direct-developer, N. bivittata). Furthermore, the cleavage pattern in these partial embryos does not revert back from their normal course. These observations suggest that there are some interesting differences between different equal-cleaving embryos. In the case of unequal-cleaving spiralians, (e.g., species of annelids and molluscs), one can generate twin embryos. This is accomplished by experimental equalization of the first cleavage division, which subdivides vegetal factors required for determination of the D quadrant. When this event is followed by unequal divisions at second cleavage, two large D blastomeres are formed and these organize separate dorsoventral axes. In the case of equal-cleaving spiralians, no differential segregation of morphogenetic factors occurs as a result of the early cell divisions, and any one of the four blastomere quadrants is able to become the D quadrant. Here, an attempt was made to create twins in C. lacteus by biasing cell contacts toward opposing sides of the embryos. Twins did arise in a very small number of cases. Why they did not occur more frequently is unclear. Presumably, the progeny of the remaining two micromeres most often took up positions that focused cell contacts onto a single macromere. It is also possible that some form of
© 2002 Elsevier Science (USA). All rights reserved.
353
Dorsoventral Axis Determination in Spiralians
lateral inhibition prevents the occurrence of two adjacent D quadrants. However, the occurrence of some twins demonstrates further that any quadrant has the potential to become the D quadrant, and suggests that contact with a single micromere is sufficient to induce the formation of a D macromere. The occurrence of other cases of radialized larvae suggests that no dorsal cell quadrant had been induced in those cases. In fact, no recognizable mesentoblast bands were observed in those radialized embryos.
The Timing of D Quadrant and Dorsoventral Axis Specification in C. lacteus Consistent with this paper, an earlier study with C. lacteus concluded that the early cleavage divisions do not play a causal role in setting up dorsoventral axial properties (Henry and Martindale, 1996); however, evidence was also presented that the dorsoventral axis is established early, prior to the two-cell stage (Henry and Martindale, 1996). This interpretation came from experiments in which the first cleavage plane had been shifted, via compression, and its relationships to the dorsoventral axis was followed by tracing the progeny of blastomeres labeled in the resulting two-cell stage embryos. In that study, a significant number of cases exhibited normal development of pilidium larvae with atypical axial relationships. The present study clearly indicates that any one of the quadrants can become the D quadrant and their identities are normally established epigenetically within the “constrained” framework of the equal quartet spiral cleavage pattern. How is it possible that aberrant axial relationships, and hence, quadrant identities were generated in compressed embryos? One explanation might be that later cleavage divisions were significantly different in some compressed embryos such that the progeny of the four-cell quadrants ended up with altered spatial relationships, which generated altered embryonic ectodermal territories. An alternative explanation could be that there are factors that predispose the dorsoventral axis in early embryos, but this axis is not fully determined until later during development by virtue of micromere inductive interactions. Unlike the case in equal-cleaving molluscs, there is no prolonged resting stage prior to the formation of the mesentoblast (4d). For instance, the resting stage between the fifth and sixth divisions in the equal-cleaving mollusc L. palustris lasts nearly 3 h, which is two to three times longer than intervals between the preceding macromere divisions (Martindale et al., 1985). Experiments indicate that the key inductive interactions that specify the 3D macromere occur during this resting stage when contacts are established with the animal micromeres (van den Biggelaar and Guerrier, 1979; Arnolds et al., 1983; Martindale et al., 1985; Martindale, 1986). In C. lacteus, the time intervals leading to the formation of the first three micromere quartets is fairly uniform (40 – 45 min at 20°C), and the interval between the formation of the third and fourth quartets is only slightly longer (lasting only 1 h; Henry and Martindale, 1998). The
exact time interval during which the key animal–vegetal inductive interactions take place in C. lacteus is uncertain. Obviously, given the time of the ablations performed in the present study, this must occur sometime after formation of the eight-cell stage. It is even possible that these inductive interactions could occur at later stages of development following the sixth cleavage division. It seems likely, though, that these interactions take place between the fifth and sixth cleavage divisions when contacts are established between the animal micromeres and vegetal macromeres (Fig. 10). One could assess the exact timing of these events by preventing these contacts during specific periods of development. This could be accomplished using cytochalasin B (i.e., Martindale et al., 1985), or Monensin (i.e., Kuhtreiber et al., 1988) or perhaps by setting up a physical barrier (e.g., injecting an oil droplet into the cleavage cavity). Work is just beginning to elucidate the molecular mechanisms involved in D quadrant specification and dorsoventral axis determination in spiralians. Recently, Lambert and Nagy (2001) have shown that MAP kinase is activated in the 3D macromere, and the micromeres that respond to 3D inductive signals, in the unequal-cleaving gastropod mollusc Ilyanassa obsoleta. Furthermore, activated MAPK is required for normal development of the dorsoventral axis and proper cell fate specification within those cells. Obviously, further work needs to be done to fully understand these processes. In fact, some evidence indicates that interesting differences may be found when compared with the mechanisms involved in dorsoventral axis determination in other organisms. For instance, dorsal and snail homologs (Hro-d1, Hro-sna1, Hro-sna2) in the leech do not appear to be involved in dorsoventral axis or mesoderm formation, as is the case in Drosophila (Goldstein et al., 2001). Given the apparent differences in the timing and mechanisms by which the D quadrant is initially established, it will be important to determine how the molecular events compare between equal- and unequal-cleaving spiralians.
ACKNOWLEDGMENTS I thank the community of the Marine Biological Laboratory and especially, Scott Fraser, Marianne Bronner-Fraser, and Joel Rothman for their generous support. M.Q.M. and Brian Walter are also thanked for comments regarding this manuscript.
REFERENCES Arnolds, W. J. A., van den Biggelaar, J. A. M., and Verdonk, N. H. (1983). Spatial aspects of cell interactions involved in the determination of dorsoventral polarity in equally cleaving gastropods and regulative abilities of their embryos, as studied by micromere deletions in Lymnaea and Patella. Roux’s Arch. Dev. Biol. 192, 75– 85. Boring, L. (1986). Cell– cell interactions determine the dorsoventral axis in embryos of an equally cleaving opisthobranch mollusc. Dev. Biol. 136, 239 –253.
© 2002 Elsevier Science (USA). All rights reserved.
354
Jonathan J. Henry
Boyer, B. C., and Henry, J. Q. (1998). Evolutionary modifications of the spiralian developmental program. Am. Zool. 38, 621– 633. Cather, J. N., and Verdonk, N. H. (1979). Development of Dentalium following removal of the D quadrant at successive cleavage stages. Roux’s Arch. Dev. Biol. 187, 355–366. Clement, A. C. (1962). Development of Ilyanassa following removal of the D quadrant at successive cleavage stages. J. Exp. Zool. 149, 193–216. Dorresteijn, A. W. C., Borneswasser, H., and Fischer, A. (1987). A correlative study of experimentally changes first cleavage and Janus development in the trunk of Platynereis dumerilii (Annelida, Polychaeta). Roux’s Arch. Dev. Biol. 196, 51–58. Fauchald, K. (1975). Polychaete phylogeny: A problem in protostome evolution. Syst. Zool. 23, 493–506. Fauvel, P., Avel, M., Harant, H., Grasse´ , P., and Dawydoff, D. (1959). Embranchement des Anne´ lides. In “Traite de Zoologie” (P. Grasse´ , Ed.), Vol. 5, ptl. pp. 3– 686. Masson et Cie, Paris. Freeman, G. (1978). The role of asters in the localization of the factors that specify the apical tuft and the gut of the nemertine Cerebratulus lacteus. J. Exp. Zool. 206, 81–108. Freeman, G., and Lundelius, J. W. (1992). Evolutionary implications of the mode of D quadrant specification in coelomates with spiral cleavage. J. Evol. Biol. 5, 205–247. Goldstein, B., Leviten, M. W., and Weisblat, D. A. (2001). dorsal and snail homologs in leech development. Dev. Genes Evol. 211, 329 –337. Guerrier, P. (1970a). Les caracte`res de la segmentation et la de´ termination de la polarite´ dorsoventrale dans le de´ veloppement de quelques Spiralia. I. Les formes a` premier clivage e´ gal. J. Embryol. Exp. Morphol. 23, 611– 637. Guerrier, P. (1970b). Les caracte`res de la segmentation et la de´ termination de la polarite´ dorsoventrale dans le de´ veloppement de quelques Spiralia. III. Pholas dactylus et Spisula subtruncata (Mollusques Lamellibranches). J. Embryol. Exp. Morphol. 23, 667– 692. Guerrier, P., van den Biggelaar, J. A. M., van Dongen, C. A. M., and Verdonk, N. H. (1978). Significance of the polar lobe for the determination of dorsoventral polarity in Dentalium vulgare (da Costa). Dev. Biol. 63, 233–242. Henry, J. (1986). The role of unequal cleavage and the polar lobe in the segregation of developmental potential during the first cleavage in the embryo of Chaetopterus variopedatus. Roux’s Arch. Dev. Biol. 195, 103–116. Henry, J. (1989). Removal of the polar lobe leads to the formation of functionally deficient photocytes in the annelid Chætopterus variopedatus. Roux’s Arch. Dev. Biol. 198, 129 –136. Henry, J., and Martindale, M. Q. (1987). The organizing role of the D quadrant as revealed through the phenomenon of twinning in the polychaete Chaetopterus variopedatus. Roux’s Arch. Dev. Biol. 196, 499 –510. Henry, J., and Martindale, M. Q. (1994a). Establishment of the dorsoventral axis in nemerteans embryos: Evolutionary considerations of spiralian development. Dev. Genet. 15, 64 –78. Henry, J. Q., and Martindale, M. Q. (1994b). Inhibitory cell– cell interactions control development in Cerebratulus lacteus. Biol. Bull. 187, 238 –239. Henry, J., and Martindale, M. Q. (1996). The establishment of embryonic axial properties in the nemertean Cerebratulus lacteus. Dev. Biol. 180, 713–721. Henry, J., and Martindale, M. Q. (1998). Conservation of the spiralian developmental program: Cell lineage of the nemertean Cerebratulus lacteus. Dev. Biol. 201, 253–269.
Henry, J. J., and Martindale, M. Q. (1999). Conservation and innovation in the spiralian developmental program. Hydrobiologia 402, 255–265. Ho¨ rstadius, S. (1937). Experiments on determination in the early development of Cerebratulus lacteus. Biol. Bull. 73, 317–342. Ho¨ rstadius, S. (1971). Nemertinae. In “Experimental Embryology of Marine and Fresh-Water Invertebrates” (G. Reverberi, Ed.), pp. 164 –174. North Holland, Amsterdam. Klymkowsky, M. W., and Hanken, J. (1991). Whole-mount staining of Xenopus and other vertebrates. Methods Cell Biol. 36, 419 – 441. Kuhtreiber, W. M., van Til, E. H., and van Dongen, C. A. M. (1988). Monensin interferes with the determination of the mesodermal cell line in embryos of Patella vulgata. Dev. Biol. 132, 436 – 441. Lambert, J. D., and Nagy, L. M. (2001). MAPK signaling by the D quadrant embryonic organizer of the mollusc Ilyanassa obsoleta. Development 128, 45–56. Martindale, M. Q. (1986). The organizing role of the D quadrant in an equal-cleaving spiralian, Lymnaea stagnalis, as studied by UV laser deletion of macromeres at intervals between third and fourth quartet formation. Int. J. Invert. Reprod. Dev. 9, 229 –242. Martindale, M. Q., Doe, C. Q., and Morrill, J. B. (1985). The role of animal–vegetal interaction with respect to the determination of dorsoventral polarity in the equal-cleaving spiralian Lymnaea palustris. Roux’s Arch. Dev. Biol. 194, 281–295. Martindale, M. Q., and Henry, J. (1995). Novel patterns of spiralian development: Alternate modes of cell fate specification in two species of equal-cleaving nemertean worms. Development 121, 3175–3185. Martindale, M. Q., and Henry, J. (1999). Intracellular fate mapping in a basal metazoan, the ctenophore Mnemiopsis leidyi, reveals the origins of mesoderm and the existence of indeterminate cell lineages. Dev. Biol. 214, 243–257. McHugh, D. (1997). Molecular evidence that echiurans and pogonophorans are derived annelids. Proc. Natl. Acad. Sci. USA 94, 8006 – 8009. Morrill J. B., Blair, C. A., and Larsen, W. J. (1973). Regulative development in the pulmonate gastropod Lymnaea palustris as determined by blastomere deletion experiments. J. Exp. Zool. 183, 47–56. Novikoff, A. B. (1940). Morphogenetic substances or organizers in annelid development. J. Exp. Zool. 85, 127–154. Render, J. A. (1983). The second polar lobe of Sabellaria cementarium embryo plays an inhibitory role in apical tuft formation. Roux’s Arch. Dev. Biol. 192, 120 –129. Render, J. A. (1989). Development of Ilyanassa obsoleta embryos after equal distribution of polar lobe material at first cleavage. Dev. Biol. 132, 241–250. Salvini-Plawe´ n, L. (1985). Early evolution and the primitive groups. In “The Mollusca” (E. R. Trueman and M. R. Clarke, Eds.), Vol. 10, pp. 59 –150. Academic Press, Orlando. Terasaki, M., and Jaffe, L. (1991). Organization of the sea urchin egg endoplasmic reticulum and its reorganization at fertilization. J. Cell Biol. 114, 929 –940. Trueman, E. R., and Clarke, M. R., Eds. (1985). “The Mollusca,” Vol. 10, Evolution. Academic Press, Orlando. Tyler, A. (1930). Experimental production of double embryos in annelids and mollusks. J. Exp. Zool. 57, 347– 407. van den Biggelaar, J. A. M. (1996). Cleavage pattern and mesentoblast formation in Acanthochiton crinitus (Polyplacophoran, Mollusca). Dev. Biol. 174, 423– 430.
© 2002 Elsevier Science (USA). All rights reserved.
355
Dorsoventral Axis Determination in Spiralians
van den Biggelaar, J. A. M., Dorresteijn, A. W. C., de Laat, S. W., and Bluemink, J. G. (1981). The role of topographical factors in cell interaction and determination of cell lines in molluscan development. In “International Cell Biology 1980 –1981” (H. G. Schweiger, Ed.), pp. 526 –538. Springer, Berlin. van den Biggelaar, J. A. M., and Guerrier, P. (1979). Dorsoventral polarity and mesentoblast determination as concomitant results of cellular interactions in the mollusk Patella vulgata. Dev. Biol. 68, 462– 471. van den Biggelaar, J. A. M., and Guerrier, P. (1983). Origin of spatial information. In “The Mollusca” (N. H. Verdonk, J. A. M. van den Biggelaar, and A. S. Tompa, Eds.), pp. 179 –213. Academic Press, New York. Verdonk, N. H. (1979). Symmetry and asymmetry in the embryonic development of molluscs. In “Pathways in Malacology” (S. van der Spoel, A.C. vaan Bruggen, and J. Lever, Eds.), pp. 25– 45. Bohn, Scheltema & Holkema, Utrecht.
Verdonk, N. H., and Cather, J. N. (1983). Morphogenetic determination and differentiation. In “The Mollusca” (N. H. Verdonk, J. A. M. van den Biggelaar, and A. S. Tompa, Eds.), pp. 215–252. Academic Press, New York. Verdonk, N. H., and van den Biggelaar, J. A. M. (1983). Early development and the formation of the germ layers. In: “The Mollusca” (N. H. Verdonk, J. A. M. van den Biggelaar, and A. S. Tompa, Eds.), pp. 91–122. Academic Press, New York. Yatsu, N. (1910). Experiments on cleavage and germinal localization in the egg of Cerebratulus. J. Coll. Sci. Imp. Univ. Tokyo 27, 1–37. Zeleny, C. (1904). Experiments on the localization of developmental factors in the nemertine egg. J. Exp. Zool. 1, 293–329.
© 2002 Elsevier Science (USA). All rights reserved.
Received for publication April Revised May Accepted May Published online July
17, 24, 24, 22,
2002 2002 2002 2002
Conserved Mechanism of Dorsoventral Axis Determination in Equal-Cleaving Spiralians Jonathan J. Henry 1 Marine Biological Laboratory, Woods Hole, Massachusetts 02543
Many members of the spiralian phyla (i.e., annelids, echiurans, vestimentiferans, molluscs, sipunculids, nemerteans, polyclad turbellarians, gnathostomulids, mesozoans) exhibit early, equal cleavage divisions. In the case of the equal-cleaving molluscs, animal–vegetal inductive interactions between the derivatives of the first quartet micromeres and the vegetal macromeres specify which macromere becomes the 3D cell during the interval between fifth and sixth cleavage. The 3D macromere serves as a dorsal organizer and gives rise to the 4d mesentoblast. Even though it has been argued that this situation represents the ancestral condition among the Spiralia, these inductive events have only been documented in equal-cleaving molluscs. Embryos of the nemertean Cerebratulus lacteus also undergo equal, spiral cleavage, and the fate map of these embryos is similar to that of other spiralians. The role of animal first quartet micromeres in the establishment of the dorsal (D) cell quadrant was examined in C. lacteus by removing specific combinations of micromeres at the eight-cell stage. To follow the development of various cell quadrants, one quadrant was labeled with DiI at the four-cell stage, and specific first quartet micromeres were removed from discrete positions relative to the location of the labeled quadrant. The results indicate that the first quartet is required for normal development, as removal of all four micromeres prevented dorsoventral axis formation. In most cases, when either one or two adjacent first quartet micromeres were removed from one side of the embryo, the cell quadrant on the opposite side, with its macromere centered under the greatest number of the remaining animal micromeres, ultimately became the D quadrant. Twins containing duplicated dorsoventral axes were generated by removal of two opposing first quartet micromeres. Thus, any cell quadrant can become the D quadrant, and the dorsoventral axis is established after the eight-cell stage. While it is not yet clear exactly when key inductive interactions take place that establish the D quadrant in C. lacteus, contacts between the progeny of animal micromeres and vegetal macromeres are established during the interval between the fifth and sixth round of cleavage divisions (i.e., 32- to 64-cell stages). These findings argue that this mechanism of cell and axis determination has been conserved among equal-cleaving spiralians. © 2002 Elsevier Science (USA) Key Words: induction; cell interactions; unequal cleavage; Nemertea.
INTRODUCTION Spiralian embryos (including polychaete annelids, vestimentiferans, echiurids, molluscs, nemerteans, sipunculids, polyclad turbellarians, mesozoans, and gnathostomulids) exhibit a conserved pattern of quartet spiral cleavage in which the first two cell divisions lead to the formation of four basic cell quadrants, generally referred to as the A (left), B (ventral), C (right), and D (dorsal) quadrants (Figs. 1A and 1B). Subsequent divisions generate additional tiers of animal micromere quartets. Ultimately, specific cell fates and 1 Present address: University of Illinois, Department of Cell and Structural Biology, 601 S. Goodwin Avenue, Urbana, IL 61801. Fax: (217) 244-1648. E-mail: [email protected].
0012-1606/02 $35.00 © 2002 Elsevier Science (USA) All rights reserved.
the dorsoventral axis are determined via inductive interactions from the D quadrant, which serves as a dorsal organizer (Clement, 1962; Cather and Verdonk, 1979; Verdonk and Cather, 1983; van den Biggelaar and Guerrier, 1983; Verdonk and van den Biggelaar, 1983; Martindale, 1986). There are two principle mechanisms that operate to establish the D quadrant in spiralian embryos. In some species, the early cleavage divisions are unequal and this results in the differential partitioning of vegetally localized factors to the larger CD and D blastomeres during the first and second cleavage divisions, respectively. This can take place via shifting of the cleavage spindle, the formation of transient cytoplasmic protrusions called polar lobes, or a combination of these events (Henry, 1986; Boyer and Henry 1998; Henry and Martindale, 1999). Various experiments indicate
343
344
Jonathan J. Henry
FIG. 1. Diagram showing the patterns of early cleavage divisions and their relationships relative to the dorsoventral axis in spiralians. All views are from the animal pole. The first and second cleavage planes are shown, as labeled. (A) Typical spiralian pattern of equal cleavage leading to the eight-cell stage. Note the presence of the vegetal cross-furrow separating the future B and D blastomeres at the vegetal pole (dotted lines). Later during development, this packing arrangement predisposes these two cells toward the center of the embryo where they can make contacts with the progeny of the animal first quartet micromeres. The third order macromere derived from one of these two cells is ultimately induced to become the 3D macromere. (B) Oblique axial relationship between the early cleavage planes and the dorsoventral (D-V) axis in most spiralians (e.g., molluscs) in which the first quartet is formed in a dextral fashion as shown in (A). Note that in these cases the dorsoventral axis is always skewed 45 degrees counterclockwise relative to the first cleavage plane. (C) Condition seen in nemerteans, like C. lacteus and N. bivittata. Note that in these embryos there is no vegetal cross-furrow and hence each quadrant has an equal chance of becoming the D quadrant. Therefore, the dorsoventral axis can be oriented in either an oblique counterclockwise (50% of the cases) or clockwise (50% of the cases) direction relative to the first cleavage plane (Henry and Martindale, 1994, 1998). (D) Diagram depicting the anatomy of the helmet-shaped, nemertean, pilidium larva with its two ciliated lappets or flaps. Larval views and structures are as indicated. The dorsoventral (D-V), anteroposterior (A-P), and left–right (L-R) axes are also shown.
that these unequal divisions segregate key vegetal factors that impart the D quadrant cell with its unique role as an inductive organizer (Cather and Verdonk, 1979; Verdonk and Cather, 1983; van den Biggelaar and Guerrier, 1983; Render, 1983, 1989; Henry, 1986, 1989; Henry and Martindale, 1987; Dorresteijn et al., 1987). These events have been well documented in certain species of molluscs (e.g., Ily-
anassa obsoleta, Render, 1989; Dentalium vulgare, Guerrier et al., 1978; Pholas dactyluss and Spisula subtruncata, Guerrier, 1970b), as well as annelids (e.g., Sabellaria cementarium, Render, 1983; S. alveolata, Novikoff, 1940, Guerrier, 1970a; Platynereis dumerilii, Dorresteijn et al., 1987; Chaetopterus variopedatus, Tyler, 1930, Henry and Martindale, 1987).
© 2002 Elsevier Science (USA). All rights reserved.
Dorsoventral Axis Determination in Spiralians
345
FIG. 2. Various partial trees showing certain phylogenetic relationships amongst different spiralian phyla, classes, and orders, and the distribution of quartet spiral cleavage types, in which the first two divisions are known to be equal (E) vs unequal (U), based on morphological descriptions of cleavage patterns. Branches are labeled with various classes, orders, and suborders where specific data are available. Phyla are listed to the far right side of the figure. Astrices indicate that unequal cleavages involving the formation of polar lobes have been reported for some species. Diagram is expanded from that of Freeman and Lundelius (1992). Phylogenetic relationships are from Freeman and Lundelius (1992), Trumen and Clarke (1985), Fauvel et al. (1975), Fauchald (1975), and Salvini-Plawe´ n (1985). Dotted lines labeled 1 and 2 indicate possible alternative phylogenies within the Bivalvia (see Freeman and Lundelius, 1992), while 3 and 4 are suggested by the work of McHugh (1997).
© 2002 Elsevier Science (USA). All rights reserved.
346
Jonathan J. Henry
FIG. 3. Photomicrographs depicting typical development of vegetal half-embryos prepared by removal of the first quartet of micromeres at the eight-cell stage. This example is 72 h old. (A). The partial embryo appears to consist of an outer covering of ciliated ectoderm surrounding an inner mass of endodermal tissue. (B) Fluorescence micrograph showing presence of disorganized muscle cells stained with Bodipy–phallacidin. ec, ectoderm; en, endoderm; mf, muscle fibers. Scale bar, 20 m.
In other species, the early cleavage divisions are equal and all four blastomeres are the same size at the four-cell stage. In these forms, the D quadrant is actually specified later during development as a result of animal–vegetal inductive interactions (van den Biggelaar and Guerrier, 1979; Arnolds et al., 1983; Martindale et al., 1985; van den Biggelaar, 1996). These inductive interactions take place between the fifth and sixth cleavage divisions when the progeny of the animal first quartet micromeres contact one of the macromeres, which is induced to become the D quadrant macromere (3D). Normally, this is one of the two more centrally located vegetal cross-furrow macromeres; however, experiments indicate that any one of the four cell quadrants can be selected as the D quadrant if these interactions are shifted by ablating individual first quartet micromeres (van den Biggelaar and Guerrier, 1979; Arnolds et al., 1983). Furthermore, if these contacts are prevented by certain chemical treatments, the D quadrant is not induced, and the embryos develop in a radialized fashion (Martindale et al., 1985; Kuhtreiber et al., 1988). The existence of this second mechanism has only been demonstrated in equal-cleaving molluscs [e.g., Patella vulgata, Lymnaea stagnalis, L. palustris, van den Biggelaar and Guerrier, 1979; Arnolds et al., 1983; Martindale et al., 1985; Acanthochiton crinitus, van den Biggelaar, 1996; and the opistobranch Haminoea (callidegenita) vesicula, Boring, 1986]. A survey of cleavage types amongst the Spiralia indicates that equal cleavage is widely represented and generally present in the most basal groups (Fig. 2). On this basis, Freeman and Lundelius (1992; see also van den Biggelaar, 1996) argued that equal cleavage, and hence animal–vegetal inductive selection of the D quadrant, represents the ancestral condition among the spiralians. Verification of this assumption requires further phylogenetic analyses and experimental examinations to determine whether the same mechanism of cell fate and axis determination operates in other equal-cleaving spiralian phyla.
FIG. 4. Diagram showing the distributions of labeled cell quadrants following random injection of DiI into single blastomeres at the four-cell stage. DiI label is shown in red. Data in the upper right corner show distribution of cases in unoperated embryos, while those below show distribution of labeled quadrants following removal of one micromere at the eight-cell stage from specific positions relative to the single labeled quadrant (as shown). For each of the four different types of deletions, predicted results are shown to the left, as far as which cell quadrants are most likely to be induced to become the D quadrant. These predictions are based on the arrangement of cell contacts following removal of individual micromeres, which results in shifting the distribution of potential animal–vegetal cell contacts toward the two cell quadrants that lie opposite the site of ablation. Each of these two macromeres lies in contact with two of the remaining first quartet micromeres, while the other macromeres only maintain contacts with a single micromere. As there is no vegetal cross-furrow, one can predict that either of those two quadrants, with the maximal number of animal cell contacts, should be equally disposed to become the D quadrant.
© 2002 Elsevier Science (USA). All rights reserved.
Dorsoventral Axis Determination in Spiralians
347
FIG. 5. Paired DIC and fluorescence light micrographs showing the distribution of labeled progeny in embryos in which one of the first quartet micromeres had been removed at the eight-cell stage. These cells were removed from specific locations relative to the single labeled cell quadrant, as shown in Fig. 4. Larval anatomy is diagramed in Fig. 1D. (A–D) One first quartet micromere was removed from the labeled cell quadrant (deletion #1 in Fig. 4). Note the absence of label in the apical (animal) ectoderm and the apical organ. (A, B) Typical A quadrant labeling pattern (macromere 1A) with label in the gut, esophagus, left ciliary band, and ventral ectoderm associated with the left lappet. Label is also seen in some muscle fibers. (C, D) Typical B quadrant labeling pattern (macromere 1B) with label in the gut, esophagus, right ciliary band, and ventral ectoderm associated with the right lappet. Label is also seen in some muscle fibers. (E–H) One of the first quartet micromeres was removed from the quadrant located just counter-clockwise to the labeled quadrant (deletion #2 in Fig. 4). (E, F) Typical B quadrant labeling pattern with label in the apical organ, the gut, esophagus, right ciliary band, and ventral ectoderm, including part of the right lappet. Label is also seen in some muscle fibers. (G, H) Typical C quadrant labeling pattern with label in the apical organ, gut, esophagus, right ciliary band, and ventral ectoderm, including part of the right lappet. No label is seen in mesodermal cells. (I–L) One of the first quartet micromeres was removed from the quadrant located just clockwise to the labeled quadrant (deletion #3 in Fig. 4). (I, J) Typical A quadrant labeling pattern with label in the apical organ, gut, esophagus, left ciliary band, and ventral ectoderm, including part of the left lappet. Label is also seen in some muscle fibers. (K, L) Typical D quadrant labeling pattern with label in the apical organ, gut, left ciliary band, and dorsal ectoderm, including part of the left lappet. The mesentoblast bands are also labeled. (M–P) One of the first quartet micromeres was removed from the quadrant located opposite the labeled quadrant (deletion #4 in Fig. 4). (M, N) Typical C quadrant
© 2002 Elsevier Science (USA). All rights reserved.
348
Jonathan J. Henry
In the present study, experiments were undertaken to determine whether the first quartet and animal–vegetal inductive interactions play an important role in establishing the D quadrant in the equal-cleaving nemertean, Cerebratulus lacteus. The results indicate that the animal first quartet micromeres are essential for the establishment of the dorsoventral axis, and furthermore that these cells serve to specify the D quadrant. In addition, the central open cleavage cavity or blastocoel, which is present in early embryos, is obliterated during the interval between fifth and sixth cleavages (32- to 64-cell stages) when cell contacts are established between animal micromeres and vegetal macromeres. These findings are similar to those seen in equal-cleaving molluscs and suggest that similar mechanisms are employed in establishing the D quadrant in other equal-cleaving spiralians, such as the Nemertea.
MATERIALS AND METHODS Collection of Adults and Embryos Adult specimens of the nemertean worm C. lacteus were obtained by the Marine Resources Department of the Marine Biological Laboratory (Woods Hole, MA). Gametes were prepared as described by Martindale and Henry (1995).
Cell Lineage Analysis via Microinjected Fluorescent Tracer Single blastomeres were injected at the four-cell stage with DiI (Molecular Probes Inc., Eugene, OR) dissolved in vegetable oil (Terasaki and Jaffe, 1991) following the procedure of Henry and Martindale (1998). Embryos that survived microinjection continued to cleave and developed normally, at the same rate as uninjected controls. Following DiI labeling, specific first quartet micromeres were removed by using a fine glass needle at the eight-cell stage. The visible presence of the tiny intracellular DiI oil droplets and the use of a Leica epifluorescence dissecting microscope permitted the positive identification of the labeled cell quadrant, which served as a frame of reference for the ablation of specific cells. In addition, one can easily distinguish the animal first quartet micromeres in nemertean embryos since these cells are larger than the corresponding vegetal macromeres, and visible polar bodies remained attached at the animal pole in the denuded embryos at the eight-cell stage.
Examination of Internal Cell Contacts during Early Development To examine the formation of the cleavage cavity and subsequent cell contacts between animal micromeres and vegetal macromeres,
embryos were fixed at 30- to 42-min intervals between the 2-cell stage and the 256-cell stage in 4% formalin in sea water overnight at 4°C. These samples were washed in sea water and cleared in either 80% glycerin in sea water or a 1:2 mixture of benzyl alcohol and benzyl benzoate following a graded series of washes to 100% methanol (Klymkowsky and Hanken, 1991).
Culture and Preparation of Embryonic and Larval Specimens Embryos were raised at 18 –20°C in 0.22 m filtered sea water (FSW) for a period of 48 –72 h. Live pilidium larvae were examined and photographed following slight compression between Rain-Xcoated (Blue Coral-Slick 50, Ltd., Cleveland, OH) slides and coverslips supported by clay feet. Since the larvae are highly transparent, all the internal cell types could be clearly visualized in these whole mounts. Fluorescence staining of filamentous actin in muscle cells was accomplished in some fixed samples by using Bodipy– phallacidin (Molecular Probes) following the protocol of Martindale and Henry (1995). Images were captured by using a Nikon fluorescence light microscope and a Spot digital camera (Diagnostic Images, MI).
RESULTS Fate Maps of the Blastomeres within the Four-CellStage Embryo The embryonic fate map and larval anatomy of C. lacteus has already been published (Henry and Martindale, 1998; see also Fig. 1D). By labeling single blastomeres at the four-cell stage, one can utilize a very clear suite of cell fates to assign quadrant identities in addition to the characteristic dorsoventral and bilateral placements of their labeled ectodermal progeny. Only the A and B quadrants generate larval muscle cells (derived form the 3a and 3b micromeres, respectively), whereas the D quadrant forms the discrete pair of left and right mesodermal or mesentoblast bands (derived from the 4d micromere). Finally, the C quadrant does not generate any mesodermal cells. To determine a baseline for the frequency of the four cell quadrants in randomly labeled embryos in the present study, individual cells were injected with DiI at the four-cell stage. A total of 90 cases were examined. The relative frequencies of labeled quadrants in these injected control embryos was found to be: A, 30%; B, 27%; C, 18%; and D, 25% (also recorded in Figs. 4 and 6 for reference).
labeling pattern with label in the apical organ, gut, esophagus, right ciliary band, and ventral ectoderm, including part of the right lappet. No label is also seen in mesodermal cells. (O, P) Typical D quadrant labeling pattern with label in the apical organ, gut, left ciliary band, and dorsal ectoderm, including part of the left lappet. The mesentoblast bands are also labeled. (A, B) (I–L), and (O, P) are left-side views. (C–H) and (M, N) are right-side views. ao, apical organ; gt, stomach; ld, DiI lipid drop; llp, left lappet; mb, mesentoblast bands; mf, muscle fibers; rlp, right lappet. Scale bar, 50 m.
© 2002 Elsevier Science (USA). All rights reserved.
349
Dorsoventral Axis Determination in Spiralians
Removal of the Entire First Quartet at the EightCell Stage Previously, Freeman (1978) (see also Zeleny, 1904; Yatsu, 1910; Ho¨ rstadius 1937, 1971) separated animal and vegetal halves of C. lacteus embryos at the 8-cell stage to assess the developmental potential of these fragments. In his study, Freeman was concerned with the mechanisms that lead to the determination of cell fates along the animal–vegetal axis (e.g., the ectodermal apical organ and the endodermal stomach); hence, Freeman did not specifically examine the differentiation of dorsoventral polarity or cell fates. As discussed above, recent cell lineage analyses indicate that certain cell fates arise differentially along the dorsoventral axis, such as the larval muscles derived from the ventrally situated 3a and 3b micromeres, and the dorsal 4d-derived mesentoblast bands (Henry and Martindale, 1998). To investigate the potential role of the first quartet micromeres and animal–vegetal interactions in the establishment of dorsoventral polarity, a total of 21 vegetal halves were prepared at the 8-cell stage. In no case was any overt sign of dorsoventral polarity seen. The partial embryos developed in a radialized fashion, consisting of an outer ciliated ectodermal layer and a central endodermal mass, which in some cases protruded outside of the ectodermal covering (Fig. 3A). Although the outer ectoderm was ciliated, an apical tuft of elongated cilia was not formed, just as reported by Freeman (1978). It was not possible to visualize whether mesentoblast bands were formed in these cases; however, larval muscle cells could be readily identified. Twelve of the 21 cases were stained with Bodipy– phallacidin to examine larval muscle cell development. In 10 cases, abundant disorganized muscle fibers were formed (see Fig. 3B), which exhibited no signs of bilateral placement. The other two cases did not contain any labeled muscle fibers.
Ablation of Single Micromeres at the Eight-Cell Stage Individual micromeres were removed at the eight-cell stage in each of the four quadrants in specific positions relative to the single labeled quadrant (see Fig. 4). The contribution of the cells derived from the labeled quadrant was then ascertained in the pilidium larvae after 48 –72 h of development (Fig. 5). In general, these larvae appeared to be completely normal with no detectable defects. Normal symmetry properties and differentiated cell fates were evident in these larvae. The frequency distributions of the labeled quadrants are shown in Fig. 4 for each of the four types of deletions. In those cases where the labeled micromere was ablated (deletion #1; see also Figs. 5A–5D), there was no labeled ectoderm in the animal hemisphere or the apical tuft (normally all four of the first quartet micromeres generate the apical tuft; see Henry and Martindale, 1999); however, the remaining labeled vegetal progeny were sufficient to clearly ascertain the labeled quadrants’ identities. The results from these experiments (Fig. 4) indicate that
there is a strong bias in the distribution of labeled cell quadrants relative to the location of the ablated micromere.
Ablation of Two Adjacent Micromeres at the Eight-Cell Stage Further experiments were undertaken to shift inductive interactions mainly to a single cell quadrant by removing two adjacent micromeres at the eight-cell stage (see Figs. 6 and 7). Three of the four possible combinations of cell ablations were performed (deletions #5–7). A single case was also completed for the fourth type of deletion (deletion #8). In most cases, the macromere located under the two remaining micromeres became the D quadrant macromere (74 –95% of the cases). The distribution of these cases is shown in Fig. 6. In many cases, the clone of labeled animal ectodermal tissue extended over a larger surface area relative to that which would have been generated in intact embryos (Figs. 7B and 7D). Apparently these cells had compensated to generate the missing ectodermal tissues. The combination of the ectodermal labeling patterns and the differentiation of mesodermal cell fates, including larval muscle cells and the mesentoblast bands, were sufficient to ascribe the identity of the labeled cell quadrants.
Ablation of Two Opposing Micromeres at the Eight-Cell Stage The results reported above indicate that the animal micromeres play a role in the determination of the D quadrant. An attempt was made to generate twin embryos, containing duplicated dorsoventral axes, by separating inductive cell contacts toward opposite sides of the embryo. To test this possibility, two of the first quartet micromeres that lie opposite one another were removed at the eight-cell stage (Fig. 8). In most cases (82%, n ⫽ 45), the resulting pilidium larvae appeared to be fairly normal, displaying normal symmetry properties, including a single dorsoventral axis. In a few cases, however, twins were formed that contained two digestive tracts (9%; see Figs. 9A and 9B). Other cases exhibited a radialized phenotype with no clear dorsoventral asymmetry and no obvious mesentoblast bands (9%; see Fig. 9C).
Observations of Animal–Vegetal Cell Contacts during the Early Cleavage Stages Beginning at the 8-cell stage, a cleavage cavity can be visualized in the center of the embryo. This cavity or blastocoel persists through the 32-cell stage (Fig. 10A). After this stage, however, this cavity is obliterated as cells come in contact at the center of the embryo. In most cases, this cavity is no longer apparent at the 64-cell stage (Fig. 10B), when contacts between the animal micromeres and the vegetal macromeres could be observed.
© 2002 Elsevier Science (USA). All rights reserved.
350
Jonathan J. Henry
FIG. 6. Diagram showing the distributions of labeled cell quadrants following random injection of DiI into single blastomeres at the four-cell stage. DiI label is shown in red. Data in the upper right corner show distribution of cases in unoperated embryos, while those below show distribution of labeled quadrants following removal of two adjacent micromeres at the eight-cell stage from specific locations relative to the single labeled quadrant. For each of the different types of deletions, predicted results are shown to the left, as far as which cell quadrant is likely to be induced to become the D quadrant. These predictions are based on the arrangement of cell contacts following removal of two adjacent micromeres, which result in shifting the distribution of potential animal–vegetal cell contacts toward the cell quadrant that lies opposite the site of ablation. This is the only macromere that lies in contact with both of the remaining first quartet micromeres, and thus, should become the D quadrant in the majority of cases.
FIG. 7. Paired DIC and fluorescence light micrographs showing the distribution of labeled progeny in embryos in which pairs of adjacent first quartet micromeres had been removed at the eightcell stage. These cells were removed from specific locations relative to the single labeled cell quadrant, as shown in Fig. 6. (A, B) Left-side view of an example in which two first quartet micromeres were removed just clockwise to the labeled cell quadrant (deletion #5 in Fig. 6). Typical D quadrant labeling pattern with label in the gut, left ciliary band, and ventral ectoderm associated with the left lappet. Most of the left side of the larva is labeled. The mesentoblast bands are also labeled. (C, D) Case in which two first quartet micromeres were removed just counter-clockwise to the labeled cell quadrant (deletion #6 in Fig. 6). Typical C quadrant labeling pattern with label in the gut, esophagus, right ciliary band, and ventral ectoderm, including the right lappet. Most of the right side of the larva is labeled. No label is seen in the mesoderm. Larva is viewed from the right side and tilted somewhat to show the apical surface. (E, F) Right-side view of a case in which the labeled first quartet micromere was removed in addition to the one located counterclockwise to that cell (deletion #7 in Fig. 6). Typical B (1B) quadrant labeling pattern with label in the gut, esophagus, right ciliary band, and ventral ectoderm, including the right lappet. No label is present in the ectoderm (or the apical organ) except for a portion of the esophagus on the right side. DiI label is seen in numerous muscle cells. Labels are the same as those used in Fig. 5. Scale bar, 50 m.
© 2002 Elsevier Science (USA). All rights reserved.
351
Dorsoventral Axis Determination in Spiralians
FIG. 8. Diagram depicting deletions (“X”) of two opposing micromeres at the eight-cell stage in an attempt to segregate animal– vegetal inductive interactions that result in the induction of two D quadrants or “twins.” Predictions are made showing possible types of twins that might be achieved. “1D and 1d” signifies that these cells represent future D quadrant cells, while “1Q and 1q” signifies that these cells are fated to become other quadrant cells. Below is shown the actual distribution of cases observed in this experiment.
DISCUSSION Animal First Quartet Micromeres Play a Role in D Quadrant Specification The results indicate that the animal micromeres play an important role in the establishment of the dorsal
quadrant in C. lacteus. Removal of the first quartet prevents the formation of the dorsoventral axis. Likewise, one can consistently influence which cell quadrant becomes the D quadrant by altering the arrangement of the animal micromeres. This takes place in a highly predictable fashion such that those quadrants with the greatest number of overlying animal first quartet micromeres ultimately become the D quadrant. When one animal micromere is removed, the three remaining micromeres are situated such that two of the four vegetal macromeres lie in contact with two micromeres, while the other two macromeres each lie in contact with only a single micromere (Fig. 4). In the majority of cases (69 – 78%), one of the two macromeres with two overlying first quartet micromeres ultimately becomes the D quadrant. When two adjacent micromeres are removed, the two remaining micromeres overlie a single macromere, and this cell ultimately becomes the D quadrant in 74 –95% of the cases (Fig. 6). The results reported here cannot be accounted for by variations in the distribution of randomly labeled cell quadrants, which might be associated with either sample size or biases imposed by the intracellular labeling technique. In every experiment, there was a strong correlation between the cell quadrant that ultimately differentiated as the D quadrant and the side of the embryo that contained the greatest concentration of animal first quartet derivatives. These results are similar to those obtained in experiments performed with equal-cleaving molluscs (van den Biggelaar, and Guerrier 1979; Arnolds et al., 1983), and they indicate that inductive interactions from the progeny of the animal micromeres play a key role in establishing the D quadrant in equal-cleaving spiralians.
FIG. 9. Photomicrographs showing twin and radialized larvae resulting from the deletion of two opposing micromeres at the eight-cell stage. (A) Twin larva containing two separate digestive tracts, each with a separate mouth, esophagus, and stomach. Dotted lines show the axes of the two parallel digestive tracts. View is from the oral surface with the shared dorsal side to the right. (B) Twin larva with two separate digestive tracts, which share a common mouth. Twin ciliated, esophageal thick tissue plates (et) are seen leading up to the two different stomachs located on the dorsal sides of the larva. The axes of these two digestive systems are located nearly at right angles to one another (dotted lines). In this case, the dorsal surfaces would be located to the right and top of the figure, respectively. Like the example shown in (A), one of these digestive tracts is smaller in size. (C) Radialized larva. The single digestive tract extends straight up from the mouth to the apical organ and tuft. Also note there are no distinct lappets and no ciliated, thick esophageal plate of tissue is present within the esophagus. at, apical tuft; st, stomodeum. Other labels are the same as those used in Fig. 5. Scale bar, 50 m.
© 2002 Elsevier Science (USA). All rights reserved.
352
Jonathan J. Henry
FIG. 10. DIC photomicrographs showing fixed 32- and 64-cell-stage embryos. Note the presence of an open central cleavage cavity or blastocoel in the 32-cell stage embryo shown in (A). This cavity is no longer present in the 64-cell-stage embryo when cell contacts appear to be established between the micromere progeny and the vegetal macromeres (B). cc, cleavage cavity. Scale bar, 20 m.
Spatial Constraints in D Quadrant Induction In molluscs, such as Lymnaea and Patella, the oblique packing arrangement of the blastomeres normally predisposes either of the two vegetal “cross-furrow” macromeres to become the D quadrant cell by virtue of their more central position within the embryo (relative to the two peripheral noncross furrow macromeres; van den Biggelaar and Guerrier, 1979, 1983; Verdonk and Cather, 1983; Verdonk and van den Biggelaar, 1983; Arnolds et al., 1983; Martindale et al., 1985; see Figs. 1A and 1B). Hence, these cells are better situated to establish cell contacts with the progeny of the first quartet micromeres. In these species, the dorsoventral axis bears only one strict axial relationship relative to the first cleavage plane (Fig. 1B). In the case of some nemerteans, like C. lacteus and Nemertopsis bivittata, there are no distinct cross-furrows in the early embryo (Henry and Martindale, 1994a, 1998), and these vegetal blastomeres lie in essentially a single plane. Hence any one of the four vegetal macromeres has an equal chance of becoming the D quadrant cell. This is reflected in the fact that the first cleavage plane can have one of two different axial relationships relative to the future dorsoventral axis (see Fig. 1C; Henry and Martindale, 1994a, 1998; note however, that some other species of nemerteans apparently do have cross-furrows; S. Maslakova and M. Q. Martindale, personal communication).
Plasticity and Regulative Capacity of EqualCleaving Spiralian Embryos Although any one of the vegetal macromeres can become a D quadrant blastomere in equal-cleaving molluscs, and animal–vegetal and inductive interactions are involved in specifying cell fates in an epigenetic fashion, blastomeres isolated from two- and four-cell stage embryos seldom
display any signs of regulative development (e.g., van den Biggelaar et al., 1981; Verdonk and Cather, 1983). Only in extremely rare cases where the pattern of cleavage divisions reiterates that of the preceding stages does regulation take place (Morrill et al., 1973; Verdonk, 1979; Verdonk and Cather, 1983). These observations argue that the spatial arrangement of the blastomeres is very important for normal development in those embryos. On the other hand, the embryos of C. lacteus freely exhibit regulative development when blastomeres are isolated at the two- or four-cell stages (Ho¨ rstadius, 1937, 1971; Henry and Martindale, 1994b, 1996; Martindale and Henry, 1995; note however, this is not the case in the direct-developer, N. bivittata). Furthermore, the cleavage pattern in these partial embryos does not revert back from their normal course. These observations suggest that there are some interesting differences between different equal-cleaving embryos. In the case of unequal-cleaving spiralians, (e.g., species of annelids and molluscs), one can generate twin embryos. This is accomplished by experimental equalization of the first cleavage division, which subdivides vegetal factors required for determination of the D quadrant. When this event is followed by unequal divisions at second cleavage, two large D blastomeres are formed and these organize separate dorsoventral axes. In the case of equal-cleaving spiralians, no differential segregation of morphogenetic factors occurs as a result of the early cell divisions, and any one of the four blastomere quadrants is able to become the D quadrant. Here, an attempt was made to create twins in C. lacteus by biasing cell contacts toward opposing sides of the embryos. Twins did arise in a very small number of cases. Why they did not occur more frequently is unclear. Presumably, the progeny of the remaining two micromeres most often took up positions that focused cell contacts onto a single macromere. It is also possible that some form of
© 2002 Elsevier Science (USA). All rights reserved.
353
Dorsoventral Axis Determination in Spiralians
lateral inhibition prevents the occurrence of two adjacent D quadrants. However, the occurrence of some twins demonstrates further that any quadrant has the potential to become the D quadrant, and suggests that contact with a single micromere is sufficient to induce the formation of a D macromere. The occurrence of other cases of radialized larvae suggests that no dorsal cell quadrant had been induced in those cases. In fact, no recognizable mesentoblast bands were observed in those radialized embryos.
The Timing of D Quadrant and Dorsoventral Axis Specification in C. lacteus Consistent with this paper, an earlier study with C. lacteus concluded that the early cleavage divisions do not play a causal role in setting up dorsoventral axial properties (Henry and Martindale, 1996); however, evidence was also presented that the dorsoventral axis is established early, prior to the two-cell stage (Henry and Martindale, 1996). This interpretation came from experiments in which the first cleavage plane had been shifted, via compression, and its relationships to the dorsoventral axis was followed by tracing the progeny of blastomeres labeled in the resulting two-cell stage embryos. In that study, a significant number of cases exhibited normal development of pilidium larvae with atypical axial relationships. The present study clearly indicates that any one of the quadrants can become the D quadrant and their identities are normally established epigenetically within the “constrained” framework of the equal quartet spiral cleavage pattern. How is it possible that aberrant axial relationships, and hence, quadrant identities were generated in compressed embryos? One explanation might be that later cleavage divisions were significantly different in some compressed embryos such that the progeny of the four-cell quadrants ended up with altered spatial relationships, which generated altered embryonic ectodermal territories. An alternative explanation could be that there are factors that predispose the dorsoventral axis in early embryos, but this axis is not fully determined until later during development by virtue of micromere inductive interactions. Unlike the case in equal-cleaving molluscs, there is no prolonged resting stage prior to the formation of the mesentoblast (4d). For instance, the resting stage between the fifth and sixth divisions in the equal-cleaving mollusc L. palustris lasts nearly 3 h, which is two to three times longer than intervals between the preceding macromere divisions (Martindale et al., 1985). Experiments indicate that the key inductive interactions that specify the 3D macromere occur during this resting stage when contacts are established with the animal micromeres (van den Biggelaar and Guerrier, 1979; Arnolds et al., 1983; Martindale et al., 1985; Martindale, 1986). In C. lacteus, the time intervals leading to the formation of the first three micromere quartets is fairly uniform (40 – 45 min at 20°C), and the interval between the formation of the third and fourth quartets is only slightly longer (lasting only 1 h; Henry and Martindale, 1998). The
exact time interval during which the key animal–vegetal inductive interactions take place in C. lacteus is uncertain. Obviously, given the time of the ablations performed in the present study, this must occur sometime after formation of the eight-cell stage. It is even possible that these inductive interactions could occur at later stages of development following the sixth cleavage division. It seems likely, though, that these interactions take place between the fifth and sixth cleavage divisions when contacts are established between the animal micromeres and vegetal macromeres (Fig. 10). One could assess the exact timing of these events by preventing these contacts during specific periods of development. This could be accomplished using cytochalasin B (i.e., Martindale et al., 1985), or Monensin (i.e., Kuhtreiber et al., 1988) or perhaps by setting up a physical barrier (e.g., injecting an oil droplet into the cleavage cavity). Work is just beginning to elucidate the molecular mechanisms involved in D quadrant specification and dorsoventral axis determination in spiralians. Recently, Lambert and Nagy (2001) have shown that MAP kinase is activated in the 3D macromere, and the micromeres that respond to 3D inductive signals, in the unequal-cleaving gastropod mollusc Ilyanassa obsoleta. Furthermore, activated MAPK is required for normal development of the dorsoventral axis and proper cell fate specification within those cells. Obviously, further work needs to be done to fully understand these processes. In fact, some evidence indicates that interesting differences may be found when compared with the mechanisms involved in dorsoventral axis determination in other organisms. For instance, dorsal and snail homologs (Hro-d1, Hro-sna1, Hro-sna2) in the leech do not appear to be involved in dorsoventral axis or mesoderm formation, as is the case in Drosophila (Goldstein et al., 2001). Given the apparent differences in the timing and mechanisms by which the D quadrant is initially established, it will be important to determine how the molecular events compare between equal- and unequal-cleaving spiralians.
ACKNOWLEDGMENTS I thank the community of the Marine Biological Laboratory and especially, Scott Fraser, Marianne Bronner-Fraser, and Joel Rothman for their generous support. M.Q.M. and Brian Walter are also thanked for comments regarding this manuscript.
REFERENCES Arnolds, W. J. A., van den Biggelaar, J. A. M., and Verdonk, N. H. (1983). Spatial aspects of cell interactions involved in the determination of dorsoventral polarity in equally cleaving gastropods and regulative abilities of their embryos, as studied by micromere deletions in Lymnaea and Patella. Roux’s Arch. Dev. Biol. 192, 75– 85. Boring, L. (1986). Cell– cell interactions determine the dorsoventral axis in embryos of an equally cleaving opisthobranch mollusc. Dev. Biol. 136, 239 –253.
© 2002 Elsevier Science (USA). All rights reserved.
354
Jonathan J. Henry
Boyer, B. C., and Henry, J. Q. (1998). Evolutionary modifications of the spiralian developmental program. Am. Zool. 38, 621– 633. Cather, J. N., and Verdonk, N. H. (1979). Development of Dentalium following removal of the D quadrant at successive cleavage stages. Roux’s Arch. Dev. Biol. 187, 355–366. Clement, A. C. (1962). Development of Ilyanassa following removal of the D quadrant at successive cleavage stages. J. Exp. Zool. 149, 193–216. Dorresteijn, A. W. C., Borneswasser, H., and Fischer, A. (1987). A correlative study of experimentally changes first cleavage and Janus development in the trunk of Platynereis dumerilii (Annelida, Polychaeta). Roux’s Arch. Dev. Biol. 196, 51–58. Fauchald, K. (1975). Polychaete phylogeny: A problem in protostome evolution. Syst. Zool. 23, 493–506. Fauvel, P., Avel, M., Harant, H., Grasse´ , P., and Dawydoff, D. (1959). Embranchement des Anne´ lides. In “Traite de Zoologie” (P. Grasse´ , Ed.), Vol. 5, ptl. pp. 3– 686. Masson et Cie, Paris. Freeman, G. (1978). The role of asters in the localization of the factors that specify the apical tuft and the gut of the nemertine Cerebratulus lacteus. J. Exp. Zool. 206, 81–108. Freeman, G., and Lundelius, J. W. (1992). Evolutionary implications of the mode of D quadrant specification in coelomates with spiral cleavage. J. Evol. Biol. 5, 205–247. Goldstein, B., Leviten, M. W., and Weisblat, D. A. (2001). dorsal and snail homologs in leech development. Dev. Genes Evol. 211, 329 –337. Guerrier, P. (1970a). Les caracte`res de la segmentation et la de´ termination de la polarite´ dorsoventrale dans le de´ veloppement de quelques Spiralia. I. Les formes a` premier clivage e´ gal. J. Embryol. Exp. Morphol. 23, 611– 637. Guerrier, P. (1970b). Les caracte`res de la segmentation et la de´ termination de la polarite´ dorsoventrale dans le de´ veloppement de quelques Spiralia. III. Pholas dactylus et Spisula subtruncata (Mollusques Lamellibranches). J. Embryol. Exp. Morphol. 23, 667– 692. Guerrier, P., van den Biggelaar, J. A. M., van Dongen, C. A. M., and Verdonk, N. H. (1978). Significance of the polar lobe for the determination of dorsoventral polarity in Dentalium vulgare (da Costa). Dev. Biol. 63, 233–242. Henry, J. (1986). The role of unequal cleavage and the polar lobe in the segregation of developmental potential during the first cleavage in the embryo of Chaetopterus variopedatus. Roux’s Arch. Dev. Biol. 195, 103–116. Henry, J. (1989). Removal of the polar lobe leads to the formation of functionally deficient photocytes in the annelid Chætopterus variopedatus. Roux’s Arch. Dev. Biol. 198, 129 –136. Henry, J., and Martindale, M. Q. (1987). The organizing role of the D quadrant as revealed through the phenomenon of twinning in the polychaete Chaetopterus variopedatus. Roux’s Arch. Dev. Biol. 196, 499 –510. Henry, J., and Martindale, M. Q. (1994a). Establishment of the dorsoventral axis in nemerteans embryos: Evolutionary considerations of spiralian development. Dev. Genet. 15, 64 –78. Henry, J. Q., and Martindale, M. Q. (1994b). Inhibitory cell– cell interactions control development in Cerebratulus lacteus. Biol. Bull. 187, 238 –239. Henry, J., and Martindale, M. Q. (1996). The establishment of embryonic axial properties in the nemertean Cerebratulus lacteus. Dev. Biol. 180, 713–721. Henry, J., and Martindale, M. Q. (1998). Conservation of the spiralian developmental program: Cell lineage of the nemertean Cerebratulus lacteus. Dev. Biol. 201, 253–269.
Henry, J. J., and Martindale, M. Q. (1999). Conservation and innovation in the spiralian developmental program. Hydrobiologia 402, 255–265. Ho¨ rstadius, S. (1937). Experiments on determination in the early development of Cerebratulus lacteus. Biol. Bull. 73, 317–342. Ho¨ rstadius, S. (1971). Nemertinae. In “Experimental Embryology of Marine and Fresh-Water Invertebrates” (G. Reverberi, Ed.), pp. 164 –174. North Holland, Amsterdam. Klymkowsky, M. W., and Hanken, J. (1991). Whole-mount staining of Xenopus and other vertebrates. Methods Cell Biol. 36, 419 – 441. Kuhtreiber, W. M., van Til, E. H., and van Dongen, C. A. M. (1988). Monensin interferes with the determination of the mesodermal cell line in embryos of Patella vulgata. Dev. Biol. 132, 436 – 441. Lambert, J. D., and Nagy, L. M. (2001). MAPK signaling by the D quadrant embryonic organizer of the mollusc Ilyanassa obsoleta. Development 128, 45–56. Martindale, M. Q. (1986). The organizing role of the D quadrant in an equal-cleaving spiralian, Lymnaea stagnalis, as studied by UV laser deletion of macromeres at intervals between third and fourth quartet formation. Int. J. Invert. Reprod. Dev. 9, 229 –242. Martindale, M. Q., Doe, C. Q., and Morrill, J. B. (1985). The role of animal–vegetal interaction with respect to the determination of dorsoventral polarity in the equal-cleaving spiralian Lymnaea palustris. Roux’s Arch. Dev. Biol. 194, 281–295. Martindale, M. Q., and Henry, J. (1995). Novel patterns of spiralian development: Alternate modes of cell fate specification in two species of equal-cleaving nemertean worms. Development 121, 3175–3185. Martindale, M. Q., and Henry, J. (1999). Intracellular fate mapping in a basal metazoan, the ctenophore Mnemiopsis leidyi, reveals the origins of mesoderm and the existence of indeterminate cell lineages. Dev. Biol. 214, 243–257. McHugh, D. (1997). Molecular evidence that echiurans and pogonophorans are derived annelids. Proc. Natl. Acad. Sci. USA 94, 8006 – 8009. Morrill J. B., Blair, C. A., and Larsen, W. J. (1973). Regulative development in the pulmonate gastropod Lymnaea palustris as determined by blastomere deletion experiments. J. Exp. Zool. 183, 47–56. Novikoff, A. B. (1940). Morphogenetic substances or organizers in annelid development. J. Exp. Zool. 85, 127–154. Render, J. A. (1983). The second polar lobe of Sabellaria cementarium embryo plays an inhibitory role in apical tuft formation. Roux’s Arch. Dev. Biol. 192, 120 –129. Render, J. A. (1989). Development of Ilyanassa obsoleta embryos after equal distribution of polar lobe material at first cleavage. Dev. Biol. 132, 241–250. Salvini-Plawe´ n, L. (1985). Early evolution and the primitive groups. In “The Mollusca” (E. R. Trueman and M. R. Clarke, Eds.), Vol. 10, pp. 59 –150. Academic Press, Orlando. Terasaki, M., and Jaffe, L. (1991). Organization of the sea urchin egg endoplasmic reticulum and its reorganization at fertilization. J. Cell Biol. 114, 929 –940. Trueman, E. R., and Clarke, M. R., Eds. (1985). “The Mollusca,” Vol. 10, Evolution. Academic Press, Orlando. Tyler, A. (1930). Experimental production of double embryos in annelids and mollusks. J. Exp. Zool. 57, 347– 407. van den Biggelaar, J. A. M. (1996). Cleavage pattern and mesentoblast formation in Acanthochiton crinitus (Polyplacophoran, Mollusca). Dev. Biol. 174, 423– 430.
© 2002 Elsevier Science (USA). All rights reserved.
355
Dorsoventral Axis Determination in Spiralians
van den Biggelaar, J. A. M., Dorresteijn, A. W. C., de Laat, S. W., and Bluemink, J. G. (1981). The role of topographical factors in cell interaction and determination of cell lines in molluscan development. In “International Cell Biology 1980 –1981” (H. G. Schweiger, Ed.), pp. 526 –538. Springer, Berlin. van den Biggelaar, J. A. M., and Guerrier, P. (1979). Dorsoventral polarity and mesentoblast determination as concomitant results of cellular interactions in the mollusk Patella vulgata. Dev. Biol. 68, 462– 471. van den Biggelaar, J. A. M., and Guerrier, P. (1983). Origin of spatial information. In “The Mollusca” (N. H. Verdonk, J. A. M. van den Biggelaar, and A. S. Tompa, Eds.), pp. 179 –213. Academic Press, New York. Verdonk, N. H. (1979). Symmetry and asymmetry in the embryonic development of molluscs. In “Pathways in Malacology” (S. van der Spoel, A.C. vaan Bruggen, and J. Lever, Eds.), pp. 25– 45. Bohn, Scheltema & Holkema, Utrecht.
Verdonk, N. H., and Cather, J. N. (1983). Morphogenetic determination and differentiation. In “The Mollusca” (N. H. Verdonk, J. A. M. van den Biggelaar, and A. S. Tompa, Eds.), pp. 215–252. Academic Press, New York. Verdonk, N. H., and van den Biggelaar, J. A. M. (1983). Early development and the formation of the germ layers. In: “The Mollusca” (N. H. Verdonk, J. A. M. van den Biggelaar, and A. S. Tompa, Eds.), pp. 91–122. Academic Press, New York. Yatsu, N. (1910). Experiments on cleavage and germinal localization in the egg of Cerebratulus. J. Coll. Sci. Imp. Univ. Tokyo 27, 1–37. Zeleny, C. (1904). Experiments on the localization of developmental factors in the nemertine egg. J. Exp. Zool. 1, 293–329.
© 2002 Elsevier Science (USA). All rights reserved.
Received for publication April Revised May Accepted May Published online July
17, 24, 24, 22,
2002 2002 2002 2002