SEARCH

SEARCH BY CITATION

Keywords:

  • spinal cleavage;
  • spiralian;
  • equal cleavage;
  • serpulid;
  • polychaete;
  • asymmetric cleavage;
  • sinistral cleavage;
  • laeotropic;
  • lophotrochozoan;
  • trochophore;
  • annelid

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES

Two major variants of the stereotypic spiral cleavage correlate with distinct developmental modes in polychaetes. Indirect development through a feeding trochophore larva correlates with development from four equal-sized blastomeres, whereas direct development correlates with unequal cleavage characterized by a large dorsal blastomere precursor maternally predetermined. The equal-size spiral cleavage of the indirectly developing serpulid Hydroides elegans has been reconstructed from serial sections of nuclei-stained embryos. The order of cell divisions has been determined from the 2-cell stage to the 80-cell stage, when gastrulation cell movements start to overlap with late spiral-cleavage divisions. In contrast to related species, the third cleavage in Hydroides elegans is invariably sinistral. The four quadrants remain indistinct until the 60-cell stage, when the small 2d22 and large 2d21 cells are generated. The developmental significance of the invariant spiral cleavage relates to the spatial distribution of gene functions that it partitions and their relation to blastomere fate commitments. The conservation and divergence of the cleavage pattern among spiralians is well suited to study the developmental control of the cell-cleavage machinery and its evolution. Developmental Dynamics 236:1611–1622, 2007. © 2007 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES

The relationship between spiral cleavage and the function and evolution of regulatory mechanisms that consolidate differential gene expression remains largely unknown. It is also uncertain the extent of maternal and/or zygotic developmental genetic control of the invariant cell divisions during cleavage. Spiral cleavage is relatively conserved among a diverse group of embryos (Wilson, 1898) and the deep understanding of its variants is relevant to trace homologies among animals as diverse as annelids, molluscs, sipunculids, echiurans, pogonophorans, nemerteans, gnatostomulids, turbelarian platyhelminths, and perhaps some arthropods (Wilson, 1898; Henry and Martindale, 1999). Spiral cleavage consists in alternating oblique cell divisions that are subsequent to the first two orthogonal divisions that generate blastomeres A, B, C, and D (Boyer and Henry, 1998). Two major spiral cleavage variants are associated with planktotrophic and lecithotrophic development in polychaetes (Anderson, 1966). The cleavage pattern of lecithotrophic polychaetes is associated with maternal specification of large blastomeres that immediately engage in terminal growth of the segmented adult worm, while planktotrophic cleavage has more blastomeres committed to the formation of the nonsegmented and microscopic feeding trochophore larva and terminal growth is relegated to the feeding dependent larval phase (Anderson, 1966). In both lecithotrophic and planktotrophic development, the segmented portion of the adult invariably derives from D quadrant descendants. In lecithotrophic development, D quadrant cytoplasmic endowment is enlarged to the expense of larval associated fates (Anderson, 1966). Thus, the cleavage associated with lecithotrophic development is unequal, that is, the D blastomere is larger than A, B, and C blastomeres, whereas planktotrophic associated cleavage is equal, at least in size. Nevertheless, equal-size spiral cleavage does not suggest equal-developmental potential, as demonstrated by the different fate of isolated blastomeres in the equal-cleaving nemertean Nemertopsis bivittata (Martindale and Henry, 1995). In equal-potential spiral cleavage, dorsoventral asymmetries are established by signaling events during cleavage; in contrast, quadrant differences in unequal spiral cleavage have maternal origin (Arnolds et al., 1983; Martindale et al., 1985; Henry, 2002; Lambert and Nagy, 2003). Equal-potential spiral cleavage has been proposed to be ancestral for spiralians, suggesting the convergent evolution of mechanisms maternally specifying dorsal fate (Freeman and Lundelius, 1992; Henry and Martindale, 1998; Henry, 2002).

Despite some detailed accounts on the spiral cleavage pattern, its relation to the mechanisms consolidating embryonic specifications remains limited by the scant studies on spiralian gene expression (Dorresteijn, 2005). It is therefore not currently possible to evaluate if the invariant spiral cleavage correlates with invariant gene expression patterns, or the extent to which differential gene expression during cleavage correlates with final fates. Blastomere dissociation experiments suggest that at least some individual blastomeres adopt their ultimate fate very early during cleavage and are capable of fulfilling their commitment in isolation, leading to the idea that spiral development corresponds to some sort of mosaic development (Wilson, 1904), characterized by fate determinants asymmetrically distributed that the invariant cleavage segregates. In the gastropod Illyanassa obsoleta, the mitotic machinery participates in the invariant early segregation of regulatory genes among blastomeres (Lambert and Nagy, 2002). If conserved, this segregation mechanism would explain, at least in part, the mosaic and invariant development observed in some spiralian embryos. Nevertheless, invariant development does not necessarily suggest cell autonomous specification mechanisms, because invariant blastomere interactions could also proceed between invariably positioned blastomeres. Therefore, a combination of autonomous and conditional specifications probably takes place in most spiralian embryos. Evidence for blastomere signaling events has been reported in association with the specification of dorsal fates in various spiralians (Arnolds et al., 1983; Martindale et al., 1985; Henry, 2002; Lambert and Nagy, 2003). Furthermore, the dynamic gene expression observed in various regulatory genes would be difficult to reconcile with purely mosaic mechanisms (Dorresteijn, 2005).

Despite the evolutionary-developmental relevance of equal spiral cleavage, the early embryogenesis of only one equal-size cleaving polychaete, Podarke obscura, has been described so far (Treadwell, 1901). Detailed anatomical studies in the species of Hydroides uncinatus have focused on trochophore larva formation rather than early cleavage (Hatschek, 1885; Shearer, 1911). The cleavage pattern of an unidentified species of the Hydroides genus was previously observed up to the 32-cell stage, and some general comments were made (Wilson, 1892), although no embryo illustrations were shown. Furthermore, not much is known about the gene regulatory machinery consolidating blastomere fates in spiralians and almost nothing in any lophotrochozoan that gastrulates by invagination. To better characterize the gene expression related to the early cleavage specifications, I report a detailed reconstruction of the early cleavage in Hydroides elegans with special attention to the order of intermediate cleavage rounds.

RESULTS AND DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES

Nomenclature follows the classic standard pattern for spiralian embryos (Wilson, 1892). Plain numerical notation instead of superscript or subscript following the quadrant letter is used thorough the text to enhance simplicity and readability. A diagram based on previous reports in other polychaetes and preliminary observations in Hydroides elegans may be useful to understand the spiral cleavage presented here (Arenas-Mena, 2006); however, the realistic account presented here supersedes this previous diagram, especially regarding individual blastomere positions past the 32-cell stage. No volumetric measurements have been performed, and only obvious blastomere size differences are mentioned in the text. Cell volume apparently correlates with nuclear volume when DNA is not highly condensed immediately before or after cell divisions. Nevertheless, the reader should be cautious about confusing small cell size with peripheral optical sections. No external differences could be appreciated among quadrants preceding the birth of 2d22 in 60-cell embryos. Therefore, the quadrant designations in all previous stages are arbitrarily assigned to indicate relative positions. A summary cleavage-sequence and cleavage diagram is included in Figures 10 and 11.

From First Cleavage to 16 Cells

The first two cleavages in Hydroides elegans are equal and almost perpendicular, as in other equal-cleaving embryos (Treadwell, 1901). This results in blastomeres A, B, C, and D of equal size. The 4′-6-diamidino-2-phenylindole (DAPI) nuclear staining reveals DNA condensation into compact heterochromatin during distinct phases of the cell cycle. For example, groups of sister chromatids are detected during their migration in anaphase cells (Fig. 1a); individual chromosome arms can be distinguished, although the haploid chromosome number could not be resolved by this method. The spiral-cleavage phase begins during the third cleavage (Fig. 1b–h). The metaphase chromosome plate of a four-cell embryo is tilted toward the right in side view (Fig. 1c); this is consistent with the subsequent laeotropic division, that is, seen from the animal pole, the animal-side blastomeres displace counterclockwise relative to the vegetal-side blastomeres (Fig. 1d–h). For example, in Figure 1e, 1d would be closer to the viewer than 1D. The reader should pay attention to whether the view is animal or vegetal, as indicated in the following figure captions, due to the possible confusion between the alternative mirror orientations.

thumbnail image

Figure 1. First cleavage to 16-cell stage. a: Animal pole view in an optical section that is perpendicular to the animal–vegetal axis of a dividing two-cell embryo. The 4′-6-diamidino-2-phenylindole (DAPI) -stained DNA is overlaid in red. Scale bar for this and subsequent panels. b: Animal–vegetal section intersecting opposing blastomeres of a four-cell embryo dividing into eight cells. c: Peripheral view of the embryo in b showing DNA overlay staining in red (false DAPI coloration). d: Peripheral view of an “almost” eight-cell embryo. e: Fluorescent DNA stain view of the embryo in d. The arrow indicates cleavage direction and points toward the viewer. f: 1Q (1A, 1B, 1C, and 1D) vegetal view. g: Differential interference contrast (DIC) image of the embryo in f. h: Optical section of the embryo in g at the level of 1q (1a, 1b, 1c, and 1d), polar cells out of focus and not shown. i: A 16-cell embryo at the 1q1 level seen from the animal pole, polar bodies out of focus, not shown. j: Embryo in i at the 1q2 level. k: Embryo in j at the 2q level. l: Embryo in k at the 2Q level. m–p: Brightfield images of the DAPI-stained embryos in i–l, respectively. q: Side view of a 16-cell embryo, animal to the top. r: Serial optical section of the embryo in q. s: Serial optical section of the embryo in r. One polar cell is seen to one side of 1d1. t: Animal view of a “late” 16-cell embryo at the 1q2 level. u–x: Brightfield images of the DAPI-stained embryos in q–s, respectively. y: Animal view of a “late” 16-cell embryo at the 1Q level. pc, polar bodies or polar cells.

Download figure to PowerPoint

The animal-side blastomeres, 1d, 1c, 1b, and 1a, of the eight-cell embryo rest on top of the boundaries between 1D-1C, 1C-1B, 1B-1A, and 1A-1D, respectively (Fig. 1g,h). Nevertheless, they lay closer to their corresponding vegetal-side sister cell (Fig. 1g,h); notice that the embryo in Figure 1f–h is seen from the vegetal side, and, therefore, 1d is seen to the left of 1D. The next division is dextral; all 1q and 1Q blastomeres divide clockwise and almost simultaneously to generate the 16-cell embryo (Fig. 1i–p). The individual cell contacts among blastomeres can be derived from the optical serial sections along the animal–vegetal axis (Fig. 1i–p) in combination with the side-view sections (Fig. 1q–s,u–x). In the late 16-cell embryos, the chromatin of 2Q and 1q2 blastomeres starts to condense (Fig. 1t,y) in preparation for the following round of divisions.

From 16 to 32 Cells

The first cells to divide in the 16-cell embryo are the 2Q blastomeres (Fig. 2a,b). The newborn 3Q macromeres and 3q quartet cells maintain highly compacted DNA just after their birth in the newly formed 20-cell embryo (Fig. 2a,b). The blastomeres of the 1q2 quartet also have highly compacted DNA (Fig. 2c,d) and are just about to divide into 1q21 and 1q22 cells, which brings the total number of blastomeres to 24 (1q21 and 1q22 derivatives can be seen in the 28-cell embryo of Fig. 2j,k). The division of 2Q just precedes or overlaps the division of 1q2. The next blastomere layer that will divide and generate the 28-cell embryo is 1q1. Just after the birth of 1q11 and 1q12 (Fig. 2i,j), the DNA of 3Q, 3q, 1q11, and 1q12 is decondensed, but the chromatin of 2q starts to compact in anticipation of the next quartet division (Fig. 2j–l,q,r). Around this stage, the polar cells start to migrate into the blastocoel and remain by the animal side in some embryos (Fig. 2i).

thumbnail image

Figure 2. At 16 to 28 cells. a–d: Serial sections of a 20-cell embryo seen from the vegetal side. e–h: Fluorescent nuclear staining of a–d, respectively. a: Optical section at the level of 3Q. DNA remains condensed just after the birth of the 3Q cells. The scale bar applies to this and subsequent panels. b: Optical section intersecting 3q and 2q. The DNA of the newly formed 3q cells remains condensed. c: Optical section at the level of 1q2 and 1q1. d: Optical section at the level of 1q1. Polar cells are seen in the center of the embryo by the base of the 1q1 cells. i–l,q,r: Serial sections of a 28-cell embryo seen from the animal pole. m–p, s and t: Respective brightfield images of i–l and q and r. i: Section at the level of 1q11. Three polar cells are seen in the center of the embryo. j: Optical section at the level of 1q12 and 1q21. k: Section at the level of 1q22. l: Section at the level of 2q; 3q cells are out of focus between 2q cells. q: Section at the level of 3q. r: Section at the 3Q level.

Download figure to PowerPoint

From 32 to 64 Cells

After the division of the four 2q blastomeres, the 32-cell embryo is formed, completing the round of divisions mediating the 16- to 32-cell transition. In Figure 3, a 32-cell embryo is shown, with the chromatin of all its nuclei relaxed. The relative position of each blastomere can be derived from the consecutive sections combined with the out of focus glow from blastomeres in other focal planes; not all the out of focus blastomeres are labeled for clarity (Fig. 3). The first cells that divide in the 32-cell embryo are the vegetalpole 3Q blastomeres into 4Q and 4q blastomeres (Fig. 4c,d); just after their birth, 4Q and 4q blastomeres still maintain their compacted DNA (Fig. 4c,d). Therefore, each cleavage round starts in the vegetal pole cells, i.e., the vegetal pole 2Q cells are the first blastomeres to divide in the 16-cell embryo (Fig. 2). The next cells that will immediately divide are 1q21 and 1q22; their DNA is already condensed just after the division of 3Q (Fig. 3a,b). Once the 1q21 and 1q22 pairs in each quadrant divide, the small 1q211 sits on top and between 1q212 and 1q221, and the also small 1q222 is located just underneath and between 1q212 and 1q221. Therefore, the complementary asymmetric divisions of 1q21 and 1q22 generate the relatively large 1q212 and 1q221 pairs in each quadrant, which are located just north of the embryo equator, that is, just above the boundary between the animal and vegetal halves (the boundary between blastomeres whose nomenclature starts with 1 and those whose nomenclature starts with 2 or higher; Arenas-Mena, 2006). The total of eight blastomeres 1q212 and 1q221 are the primary prototroch trochoblasts and will express several ciliary band markers soon after their generation (Arenas-Mena et al., 2007). The division of the equatorially located 1q21 and 1q22 after the division of 3Q in the 32-cell embryo (Fig. 4) parallels the division of the equatorial precursor 1q2 after the division of 2Q in the 16-cell embryo (Fig. 2). Thus, the transitions of 16- to 32-cell and 32- to 64-cell embryos begin with a vegetal pole quartet division followed by an equatorial quartet division.

thumbnail image

Figure 3. A 32-cell embryo. a–d,i and j: Serial sections of a 32-cell embryo seen from the vegetal point of view. e–h,k and l: Respective differential interference contrast optical sections of a–d, i and j. a: Section at the 3Q level. b: Section at the 3q-3Q level. c: Section at the 3q-2q2 level. d: Section at the 2q1-2q22 level. i: Section at the 1q21-1q12 level. Polar cells can be seen just underneath the 1q11 cells (out of focal plane) in the center of the blastocoel. j: Section at the 1q11-1q12 level.

Download figure to PowerPoint

thumbnail image

Figure 4. At 36 to 44 cells. a–d: Serial sections of a 36-cell embryo seen from the animal side. e–h: Brightfield images of the respective embryos above. a: Animal half showing 1q11, 1q12 and 1q21. DNA is more condensed in 1q21. The scale bar applies to this and subsequent panels. b: Optical section at the level of 1q1 and 1q22. DNA is more condensed in 1q22. c: Section at the level of 3q and 4q. The 4q DNA is more condensed. d: Section at the level of the newly formed 4Q blastomeres. i–l,p: Consecutive serial sections of a 44-cell embryo seen from the animal pole. i: Section at the level of 1q11, 1q12, and 1q211. j: Section at the level of 1q212 and 1q221 blastomere pairs. Polar cells can be seen just underneath the apical quartet. k: Section at the level of 1q222, 2q1, and 1q2. m: Brightfield view of i. n: Cleavage sequence diagram. Numbers in black circles indicate embryo cell number after the respective blastomere division. o: Brightfield view of p.

Download figure to PowerPoint

The quartet cleavage order in Hydroides elegans departs from the quartet cleavage order of Podarke obscura soon after the 16-cell stage. In Podarke obscura, the 2Q and 1q1 quartets divide almost simultaneously (Treadwell, 1901), whereas in Hydroides elegans 2Q and 1q2 divide almost simultaneously. The round of cell divisions that mediates the 32- to 64-cell transition start in the animal and vegetal poles in Podarke, that is, 1a11 and 3Q, respectively, whereas in Hydroides, the 32-cell embryo starts to divide its 3Q blastomeres, immediately followed by 1q21 and 1q22 blastomeres. In short, in Podarke obscura, almost simultaneous cell divisions start in both polar quartets in 16- and 32-cell embryos, but in Hydroides elegans, cell divisions start in the vegetal and equatorial quartets.

The next proliferating quartet is 3q, which divides in the eight 3q1 and 3q2 descendant cells; this brings the total embryo blastomere count to 48 (Fig. 5a,b). The newly formed 3q1 blastomeres lay almost just underneath the 1q222 blastomeres (Fig. 5c). In 48-cell embryos, it can be appreciated that the 2q2 blastomeres are larger than their animal side sisters 2q1. The 52-cell embryo is generated after division of 1q11 in the animal hemisphere (Fig. 5i-j). The apical-most 1q111 cells, that is, the rosette cells, sink toward the blastocoel in relation to their “vegetal” 1q112 sisters, that is, the cross cells. The transient sinking of the rosette cells is similar in the distantly related embryos of Podarke obscura Treadwell, 1901, which also exhibit equal-size spiral cleavage; however, equipotency among quadrants has not been experimentally demonstrated in Podarke obscura either. In 52-cell embryos, the DNA of blastomeres 1q12 and 2q2 is condensed (5,l)and 1q12 will divide subsequently to generate the 56-cell embryo (6i). Blastomeres 1q121 and 1q122 have similar sizes (Fig. 6a,b), but blastomeres 1q111 are much smaller than 1q112 blastomeres (Fig. 6a,e).

thumbnail image

Figure 5. At 48 to 52 cells. a–d: Serial sections of a 36-cell embryo seen from the vegetal side. e–h: Brightfield images of the respective embryos above. a: Optical section at the level of 3q2 and 4q blastomeres. The scale bar applies to this and subsequent panels. b: Optical section at the level of 2q2 and 3q1. c: Optical section at the level of 2q1, 3q1, and 1q22. d: Optical section at the level of 1q21, 1q12, and 1q11. i–l: Serial sections of a 52-cell embryo seen from the animal side. m–p: Brightfield images of the respective embryos above. i: Section at the level of 1q112 and 1q211; 1q111 cells are in the center and out of focus. j: Section at the level of 1q111 and 1q12. k: Section at the level of 1q1 and 1q222. l: Section that shows blastomeres from the 3q1 to the 4Q levels.

Download figure to PowerPoint

thumbnail image

Figure 6. At 56 to 60 cells. a–d,i–k: Serial sections of a 60-cell embryo seen from the animal side. e–g,m–o: Brightfield images of the respective embryos above. h: Cleavage sequence diagram. Numbers in black circles indicate embryo cell number after the respective blastomere division. a: Optical section at the level of 1q111, 1q112, and 1q211. The scale bar applies to this and subsequent panels. b: Optical section at the level of 1q211 and 1q122. c: Optical section at the level of the 1q221 and 1q212 pairs. The arrowhead marks the prospective dorsal midline. A couple of polar cells are seen in the blastocoel. d: Optical section at the level of 2q1 and 1q222. The DNA of 2q1 cells is partially condensed. i: Optical section at the level of 2q21. The blastomere 4d is the only 4q blastomere seen at this animal–vegetal level. j: Optical section at the level of 2q22, in close association with 4d. k: Optical section at the 4Q level. l: Embryo transitioning from 52- to 56-cell stage during the cleavage of 1q12 blastomeres. p: Serial section of the embryo in l at the 1q221 and 1q212 level.

Download figure to PowerPoint

The identity of the D-quadrant is revealed in 60-cell embryos after the division of the 2q2 blastomere quartet. The blastomere 2d21 is bigger than its counterparts in other quadrants, i.e., 2a21, 2b21, and 2c21 (Fig. 6d,i), and its sister 2d22 is smaller than its counterparts in other quadrants, i.e., 2a22, 2b22, and 2c22. In addition to the distinct cleavage pattern of 2d2, 4d occupies a clearly more animal position than 4a, 4b, and 4c (Fig. 6i–k); thus, it appears that 4d is intruding further into the blastocoel, perhaps a sign of an incipient invagination. This is seen by comparing the relative position of 4d in Figure 6j with that of 4a, 4b, and 4c in Figure 6k. The blastomere 4d will adopt mesodermal fate and contribute to great portion of the adult mesoderm, while its counterparts in other quadrants will assume endodermal fates. These D-quadrant–specific cleavage patterns and cell positions mark the future dorsal side, indicated by the arrowhead in Figure 6c.

After the division of the four 2q1 blastomeres, the 64-cell embryo is formed, completing the round of divisions mediating the 32- to 64-cell transition (Figs. 7, 8). The animal plate is relatively flat, and, therefore, almost all animal side blastomeres can be included in a single focal plane (Fig. 7a); the blastomere arrangements and individual cell contacts can be derived by combining the animal–vegetal serial sections (Fig. 7) with side view serial sections (Fig. 8).

thumbnail image

Figure 7. A 64-cell embryo. a–h: Serial sections of a 64-cell embryo seen from the animal side. a: View of the animal hemisphere. Note that 1q111, 1q112, 1q121, 1q122, and 1q211 are located in the same plane. The scale bar applies to this and subsequent panels. b: View at the level of 1q211 and 1q212 trochoblast precursor pairs and recently divided 1q11 blastomeres, which still have their DNA condensed. c: Section adjacent to b. d: Section at the level of 1q222. e: Section at the level of 1q21, 3q1 and 2q21. The blastomere 4d is seen “pushing” into the blastocoel. 2d21 is clearly larger than its counterparts in other quadrants. f: Section adjacent to e. Blastomere 2d22 is clearly smaller than its counterparts in other quadrants. g: Section at the level of 3q2. h: Section at the level of 4Q and 4q. i: Brightfield image of a. j: Brightfield image of b. k: Brightfield image of e. l: Brightfield image of h.

Download figure to PowerPoint

thumbnail image

Figure 8. A 64-cell embryo side view serial sections. a–f: Consecutive serial sections of a 64-cell embryo showing the blastomeres indicated. Animal to the top. The scale bar applies to all panels. g: Brightfield view of a. h: Brightfield view of d. i: Brightfield view of f.

Download figure to PowerPoint

From 64 to 77 Cells

The vegetal pole quartet also initiates the divisions necessary to generate the 68-cell embryo (Fig. 9c). The 72-cell embryo emerges after 1q112 divides into the relatively large 1q1121 and the inconspicuous 1q1122 (compare Fig. 9d with 9a), which apparently sinks between neighboring blastomeres (not shown). Similarly asymmetric divisions and cell behavior was reported in other 1q112 blastomere descendants of Podarke obscura (Treadwell, 1901). Nevertheless, in Podarke, the division of 1q112 blastomeres does not generate the inconspicuous cells and instead follows an asymmetric, bilateral pattern, which is not apparent in Hydroides elegans embryos. During later stages, the 1q122 blastomeres will divide too (not shown), and other cell divisions that are more difficult to relate to the spiral pattern will eventually ensue. In the D quadrant, the relatively large 2d21 is the only blastomere of its quartet that divides in 2d211 and 2d212 (Fig. 9i,j); this transiently brings the total number of cells to 73. The 2d21 division would represent a second, dextral division if considered within the spiral context, but it should be considered part of the process that secures a bilateral symmetry, i.e., the large 2d21 compensates with cellular division in the left side for the cellular void left by the midline positioning of 2d22. The next cells that divide are those of the 3q2 quartet (Fig. 9m–p), followed by 2a22, 2b22, and 2c22, whose DNA is seen condensed in Figure 9o, bringing the total number of cells to 80. Beyond this stage, gastrulation movements compound with late spiral cleavages, and cleavage transitions start to depart from the regular and symmetric spiral pattern (not shown). Therefore, lineage relations during cleavage–gastrulation stages become more and more difficult to reconstruct by the method used in this study, and alternative strategies will be required to link the invariant spiral cleavage to gastrulation and eventually to the consolidation of larval fates.

thumbnail image

Figure 9. A 68 to 77 cell embryo. a–c: Consecutive serial sections of a 68-cell embryo seen from the animal pole. a: Animal plate. The scale bar applies to this and subsequent panels. b: Section at the level of 4d. c: View of the vegetal pole. d,i–l: Serial sections of a 73-cell embryo. d: Animal plate of a 73-cell embryo. e: Brightfield view of c. f: Brightfield view of j. g: Brightfield view of k. h: Brightfield view of d. i: Section at the 2d11 level. j: Section at the level of 4q and 2d22. k,l: Sections at the level of the vegetal pole. m–p: Serial sections of a 77-cell embryo at the levels indicated.

Download figure to PowerPoint

Sinistral Cleavage

The third cleavage has been observed in several living embryos, and it has always been found to be laeotropic, that is, the animal blastomeres or the eight-cell embryo are displaced counterclockwise relative to the vegetal blastomeres when seen from the animal point of view (Figs. 10, 11a). Furthermore, no older embryos have been observed with blastomere orientations that would be consistent with a third dexiotropic cleavage. This is the first polychaete described that has a third sinistral cleavage. In Hydroides eupomatus (or Hydroides uncinatus), the third cleavage was reported dextral (Hatschek, 1885; Shearer, 1911), and yet another unidentified species of the genus Hydroides was also described as dextral (Wilson, 1892). Nevertheless, laeotropic spiral cleaving species have been described in mollusks, with some cases of alternating dextral and sinistral forms in the same species (Gerrier, 1970; Luetjens and Dorresteijn, 1995). The number of embryo anatomy reports remains small to properly evaluate the prevalence of dextral and sinistral modes among spiralians, although it seems that dextral species prevail. The existence of these alternating cleaving forms may be useful in the future to identify early cytological asymmetries that establish the direction of the third cleavage (Gerrier, 1970). The orientation of the third cleavage perhaps relates to left–right asymmetries in the larva such as the order of larval eye formation. In Xenopus, it seems that maternal organization of microfilaments predetermines first cleavage cortical rotations apparently upstream of left–right specifications (Danilchik et al., 2006). It is exciting to entertain the possibility of similar mechanisms for left–right axial specifications in spiralian and deuterostome eggs.

thumbnail image

Figure 10. Cleavage sequence diagram. The blastomere cleavage sequence from the 4-cell to the 80-cell stage. The total embryo cell number is indicated at the top; each number aligns with the corresponding newborn blastomeres, and horizontal separation approximates the timing between cell divisions. Time after fertilization for the stages above is indicated in hours at the bottom; this data should not be confused with an absolute time scale as explained in the text. Gray blocks underline 16- to 32-cell transition and 32- to 64-cell transitions. Numbers in black ovals at each division indicate the relative cleavage order within each transition. Diagrams of 12-, 32-, and 64-cell embryos at the bottom. Black circles indicate embryo cell number after the respective blastomere division.

Download figure to PowerPoint

thumbnail image

Figure 11. Summary diagram. Line drawings have been elaborated from optical sections to illustrate major features of Hydroides elegans cleavage. All images are or have been converted into animal point of view images. a: An eight-cell embryo drawn from optical sections in Figure 1g,h. Animal half in blue, and vegetal half in red discontinuous line. b: A 16-cell embryo drawn from optical sections in Figure 1m–p. Animal hemisphere to the left, and vegetal hemisphere to the right. Double black lines indicate sister blastomeres. c: A 28-cell embryo drawn from optical sections in Figure 2m–p,s,t. Left, animal hemisphere. Right, vegetal hemisphere. Color code matches the 16-cell precursors shown in b. The division of 2q blastomeres will generate the 32-cell embryo. d: Drawing of the animal hemisphere of the 64-cell embryo shown in Figure 7i. Sister blastomeres indicated with double lines along the cleavage plane. There is no obvious sign of bilateral symmetry or dorsal specific cleavages in the animal hemisphere. The quadrant designations are derived from vegetal side cleavage patterns in consecutive optical sections. Future ventral (V), dorsal (D), left (L), and right (R) sides are indicated. e: Composite drawing of vegetal optical sections of the 60-cell embryo shown in Figure 6m–o. The relatively smaller size of 2d22 reveals the future dorsal side of the embryo and marks bilateral symmetry. In the vegetal side, the dorsal side corresponds to the aboral side of the blastopore. Aboral–dorsal (A/D); oral–ventral (O/V). For diagrams illustrating aboral, dorsal, oral, ventral embryo–larva designations see previous report (Arenas-Mena et al., 2007). f: Side view drawing of the 64-cell embryo in Figure 1g of a previous report (Arenas-Mena et al., 2007). The large primary trochoblasts 1q221 and 1q212 (indicated in gray) have originated from complementary asymmetric divisions of 1q21 and 1q22. g: Side view drawing of the embryo in Figure 8g.

Download figure to PowerPoint

Equal-Size Spiral Cleavage and Bilateral Symmetry

Hydroides elegans displays equal-size cleavage, although equipotency among blastomeres has not been experimentally demonstrated. In equal-cleaving molluscs and annelids, and in contrast to unequal-cleaving species, dorsoventral asymmetry is not maternal but established by signaling events during later stages (Lambert and Nagy, 2003). Equal cleavage has been proposed to be ancestral for spiralians (Freeman and Lundelius, 1992; Henry and Martindale, 1998; Henry, 2002), and embryological studies reveal that animal micromere descendants provide an inductive signal to vegetal blastomeres in equal-cleaving gastropods (Arnolds et al., 1983; Martindale et al., 1985; Henry, 2002). It seems that signaling events activate the MAPK pathway in vegetal blastomeres of Hydroides and other spiralians (Lambert and Nagy, 2001, 2003). Those signaling events could precede the external manifestation of a D quadrant fate by the asymmetric division of 2d2, which defines bilateral symmetry (Fig. 11e). In parallel to the equal-quadrant size, the expression of the endoderm-mesoderm regulatory gene HeFoxA1 has equal intensity in all four 2q blastomeres of 16-cell embryos, and in parallel to the later cleavage asymmetries HeFoxA1 transcripts are depleted from prospective dorsal-side (D-quadrant) blastomeres (Arenas-Mena, 2006). Similarly, the earliest vegetal HeOtx expression is detected in 2a-c22 but not in 2d22 in 72-cell embryos (Arenas-Mena and Wong, 2007). Thus, although there are parallels between asymmetric gene expression and specific cleavage patterns by the dorsal side, the causal relation, if any, between cleavage and gene expression remains uncertain. No gene expression has been detected in the animal cap that would precede the putative signaling event in Hydroides elegans, and no clear external blastomere differences are appreciated in the animal hemispheres of 64-cell embryos, which are subsequent to the evident dorsal specific cleavages in the vegetal hemispheres of 60-cell embryos (Fig. 11d,e)

Genetic Control of the Invariant Spiral Cleavage

The invariant spiral cleavage of Hydroides elegans provides one more tractable system in which to study the developmental control of cleavage. In particular, Hydroides elegans is well suited to explore the genetic control of asymmetric cell division and spindle orientation. Blastomere size differences are obvious in the animal cap of a 64-cell embryo (Fig. 11d), where the animal-most 1q111 blastomeres are much smaller than their sister blastomeres 1q112. The earliest and more dramatic asymmetric cell divisions generate the eight primary trochoblasts, which are the first cells that stop dividing and differentiate in this polychaete (Arenas-Mena et al., 2007). Interestingly, complementary asymmetric divisions of 1q21 and 1q22 pairs in each quadrant locate the primary trochoblasts in the same animal–vegetal position (Figs. 3, 11f), in anticipation to the formation of an eight-cell ciliated prototroch arch that remains open by the dorsal side (Arenas-Mena et al., 2007).

The genes that control asymmetric cell division and spindle orientation seem to be conserved in Drosophila, Caenorhabditis elegans, and vertebrates (Doe and Bowerman, 2001; Gonczy, 2002; Betschinger and Knoblich, 2004), and one would expect fine control of such conserved cleaving machinery during the invariant spiral cleavage of Hydroides elegans. Spindle position is the major determinant of the relative size and positional relation of any two sister cells, and it is controlled by the growth dynamics of microtubules anchored to particular cortical domains in the dividing cells (Gonczy, 2002). The morphogenetic relevance of spindle orientation has been demonstrated during Drosophila wing development (Baena-Lopez et al., 2005). Perhaps not surprisingly, the authors identified upstream regulatory control of spindle orientation by some planar polarity genes (Baena-Lopez et al., 2005). The invariability of the spiral cleavage pattern within species does not suggest its independence from genetic control, and the cleavage variants among spiralians should certainly correspond to genetic modifications along evolutionary lines. Almost nothing is known about the genetic control of spindle position during stereotypic spiral cleavage. Zygotic regulatory inputs could control spindle position or maternal predetermination could outline the cleavage pattern in the egg. At least in the highly modified cleavage of leeches, it seems that zygotic gene expression is required to control spindle position in some blastomeres (Bissen and Smith, 1996).

The zygotic expression of HeFoxA1 in 2q blastomeres reveals zygotic differential gene expression as early as the 16-cell stage in Hydroides elegans (Arenas-Mena, 2006). Eventually, it will be interesting to analyze the effects of regulatory gene expression perturbations on the invariant cleavage pattern. Of more immediate relevance is the precise characterization of ongoing regulatory gene expression within the context of the invariant cleavage of Hydroides elegans. Eventually, the trochophore larval fates of the individual blastomeres that express distinct genes will be revealed by modern cell lineage methods. Given the temporal and spatial compact embryonic development of Hydroides elegans, which ends in the formation of a feeding trochophore in approximately 14 hr, it is expected that at least some the early-cleavage gene expression patterns will define the boundaries of specific larval fates.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES

Embryos were fertilized as previously described (Arenas-Mena, 2006), with the exception that half-hour sperm-egg incubations were used; therefore, stages are indicated by half-hour time ranges. The exact cleavage timing is uncertain beyond the first cleavage, which happens between 60 and 75 min after mixing the gametes. No in vivo temperature-controlled observations have been made for later cleavage events. Fixation of specimens was performed as previously described (Arenas-Mena, 2006), and DAPI staining was performed by a 5-min incubation in 10 μg/ml DAPI in MOPS buffer solution followed by 2 washes in MOPS buffer. Several hundred embryos from each half-hour time range were mounted and multiple embryos in the animal–vegetal orientation for each cleavage stage were observed. Full serial optical-section recordings were obtained for two to four embryos from each stage. The cleavage sequence has been further confirmed during parallel gene expression characterizations (Arenas-Mena and Wong, 2007; Arenas-Mena et al., 2007). Observation was performed with a 100× HCL-FLUOTAR oil immersion objective. Digital images were recorded with a Spot RT color camera. Image overlays of DAPI and brightfield images were done using Photoshop.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES

I thank Kimberly Suk-Ying Wong and Navid R. Arandi-Foroshani for providing and staining the embryos used in this study. Our laboratory is supported by CSUPERB funds.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES