• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of plosonePLoS OneView this ArticleSubmit to PLoSGet E-mail AlertsContact UsPublic Library of Science (PLoS)
PLoS ONE. 2009; 4(5): e5506.
Published online May 11, 2009. doi:  10.1371/journal.pone.0005506
PMCID: PMC2676505

Cell Lineage of the Ilyanassa Embryo: Evolutionary Acceleration of Regional Differentiation during Early Development

Patrick Callaerts, Editor


Cell lineage studies in mollusk embryos have documented numerous variations on the lophotrochozoan theme of spiral cleavage. In the experimentally tractable embryo of the mud snail Ilyanassa, cell lineage has previously been described only up to the 29-cell stage. Here I provide a chronology of cell divisions in Ilyanassa to the stage of 84 cells (about 16 hours after first cleavage at 23°C), and show spatial arrangements of identified nuclei at stages ranging from 27 to 84 cells. During this period the spiral cleavage pattern gives way to a bilaterally symmetric, dorsoventrally polarized pattern of mitotic timing and geometry. At the same time, the mesentoblast cell 4d rapidly proliferates to form twelve cells lying deep to the dorsal ectoderm. The onset of epiboly coincides with a period of mitotic quiescence throughout the ectoderm. As in other gastropod embryos, cell cycle lengths vary widely and predictably according to cell identity, and many of the longest cell cycles occur in small daughters of highly asymmetric divisions. While Ilyanassa shares many features of embryonic cell lineage with two other caenogastropod genera, Crepidula and Bithynia, it is distinguished by a general tendency toward earlier and more pronounced diversification of cell division pattern along axes of later differential growth.


Spiral cleavage is a form of early embryonic development that occurs widely within the lophotrochozoan superphylum [1]. Remarkably, ‘spiralian’ taxa have evolved disparate adult body plans, yet retain a close correspondence between early embryonic cell lineage (spatially defined with respect to the egg's primary animal-vegetal (AV) axis) and cell fate (i.e., clonal distribution among juvenile/larval body regions and tissues) [2][4]. Although spiral cleavage has been modified beyond recognition in some clades, most known mollusks, annelids, nemerteans, and polyclad flatworms show only subtle deviations from a generalized spiral cleavage pattern.

The term ‘spiral cleavage’ reflects a characteristic spatial pattern of cell division exhibited most conspicuously from the third through fifth cell cycles. Unlike other animal embryos, spiralian early cleavage planes are never perpendicular to the AV axis. Thus, while the third cleavage in a frog, sea urchin or jellyfish embryo separates four animal (‘northern’) cells from vegetal (‘southern’) sister cells, third cleavage in a spiralian embryo separates four ‘northwestern’ cells from ‘southeastern’ sisters (the reverse chirality is also observed in some taxa) (Figure 1A,B). The four (typically smaller) cells around the animal pole are called micromeres, and their vegetal sisters are called macromeres. The macromeres undergo two more rounds of concerted asymmetric division, budding two more quartets of micromeres toward the animal pole (Figure 1C–E). The chirality of the oblique macromere divisions alternates with each cell cycle: e.g., the first quartet is budded to the northwest, the second to the northeast, and the third to the northwest once more. The micromeres themselves continue to follow the rule of alternating division chirality for one or more cell cycles. Through synchronous reiteration of obliquely oriented divisions, four founder cells thus give rise to lineages (A, B, C, D) that represent ‘quadrants’ distributed in a rotationally symmetric pattern about the AV axis (Figure 1F). Owing to the practically identical cell division patterns in the four quadrant lineages, early cleavage generates multiple cell ‘tiers’, or sets of four synchronously formed and phenotypically similar cells that are radially disposed about the AV axis. Most spiralian embryos exhibit tier-specific patterns of cell division at early stages; the significance of this mitotic asynchrony is unknown, but presumably reflects differential inheritance of factors regulating cell division. Indeed, in 24-cell embryos of the marine snail Ilyanassa, each tier is distinguished by inheriting a unique set of cytoplasmic RNA species [5].

Figure 1
Spiral cleavage: elaborating fourfold rotational symmetry about the primary axis of polarity.

For the most part, each micromere clone in a spiralian embryo remains largely cohesive throughout embryogenesis, and adopts a highly predictable fate [2]. The entire ectoderm arises from the first, second and third micromere quartets (collectively, the ‘ectoblast’). Figure 2A shows a schematic summary of ectodermal morphogenesis, as inferred from fate mapping and other descriptive studies of embryogenesis in gastropod mollusks [6][11]; this scheme can be generalized in broad outline to other spiralians, though some details differ between higher-order taxa [2][4]. At the animal pole, the four-way junction of the first-quartet micromere clones persists to form the anterior apex of the juvenile animal. Oppositely, the vegetal extremities of the ectoblast (second and third quartet clones) converge at the vegetal pole to form the rudiment of the mouth; in many taxa (including Ilyanassa) this convergence occurs by epiboly, as illustrated in Figure 2A. During subsequent development, the AV axis bends extensively, displacing the mouth-forming vegetal cell group along one meridian toward the animal pole. The ectodermal territory centered on this meridian becomes shorter and wider as micromere clones spread bilaterally and/or become partially internalized; at the same time, the ectodermal region opposite to this territory elongates meridionally to form most of the trunk, including both the definitive dorsal and ventral ectoderm. In gastropods (as in most spiralians), the first-quartet micromere clones collectively form the animal's head; the boundary between the first-quartet and second-quartet clones forms the velum, a band of ciliated cells that functions in larval locomotion and feeding.

Figure 2
Morphogenesis and the secondary axis of the spiralian embryo.

The above described transformation of the spiral-cleaving embryo into a dorsoventrally polarized animal depends on an early intercellular signal that patterns the ectoblast along a secondary axis, orthogonal to the original (AV) axis of polarity. This secondary axis is commonly designated as ‘dorsoventral’ (a shorthand term whose inaccuracy is illuminated by Figure 2A). In gastropods, dorsoventral ectoblast patterning depends on one of the vegetal macromeres, 3D (and/or its daughter cell 4d) [12][17]. A substantial body of evidence suggests that the developmental program of each ectoblast micromere (or micromere daughter) is encoded combinatorially by its quartet/tier identity and its position relative to 3D/4d (reviewed in [16]). (Figure 2B). Consistent with this model, each micromere quartet includes at least one early-born bilateral pair of cells that form bilaterally paired regions of the larval ectoderm [6][10]. Ablation of any single micromere yields a defect corresponding largely to its fate, indicating early restriction of developmental pathways [18][20]. Single micromeres heterotopically transplanted prior to 3D formation develop according to tier identity and position [21]. Blocking the 3D/4d signal radializes morphogenesis, with each cell adopting a default ventral fate characteristic of its tier (Figure 2C); in this situation, the embryo's original four-fold rotational symmetry is maintained through gastrulation and larval organogenesis [12], [17], [22]. Differential specification of micromere lineages along the dorsoventral axis is correlated with early divergence of cell division rate and geometry [12]. This early mitotic differentiation is the first morphological sign of dorsoventral pattern in the ectoderm.

Our present understanding of spiralian cell fate specification owes much to studies of the mud snail Ilyanassa. In this species, the secondary axis is determined through an unusual mechanism: whereas in other gastropod taxa the 3D cell is specified inductively among initially equivalent macromeres, the 3D precursor in Ilyanassa is specified as the D cell by asymmetric segregation of vegetal cytoplasm during the first two cleavages [23] (Figure 3A). This segregation allows 1d, the first-quartet micromere of the D quadrant, to differ at birth from the three other first-quartet cells [12], [21]. In other respects, the ectoblast of Ilyanassa displays the same rotational symmetry seen in other gastropod embryos (Figure 3B), and a molecular marker of signaling from the D quadrant is not expressed until after the birth of the third quartet [16], [24]. As in other mollusks, the signaling 3D cell is mother to the mesentoblast 4d, which marks the definitive median plane at the 28-cell stage, and which gives rise to a bilateral pair of mesodermal and endodermal stem cell lineages [25]. Subsequent cell divisions among the ectodermal micromere lineages have not been systematically described.

Figure 3
Early cleavage in Ilyanassa obsoleta.

In this paper I follow the Ilyanassa embryo through its next fifty-six cell divisions, describing the spatiotemporal division pattern from first cleavage through the eighty-four-cell stage. This analysis reveals progressive differentiation of division patterns within cell tiers, especially along the secondary axis. Such within-tier differentiation is more widespread, and also tends to be more pronounced, in Ilyanassa compared to the related caenogastropod genera Crepidula [6], [11], [26] and Bithynia [27]. Although the D lineage is specified at later stages in these two genera, most of the accelerated regional differentiation in Ilyanassa occurs after 3D/4d signaling, and thus does not depend on early D specification as a logical precondition. Moreover, the global acceleration of differential growth in Ilyanassa is generated by a wide variety of changes in division geometry and timing among different micromere sublineages, implying the concerted evolution of diverse factors controlling cell division.

Materials and Methods

Animals, Microscopy, and Data Acquisition

Snails were obtained from the Marine Biological Laboratories at Woods Hole in the winter and early spring. Snail and embryo cultures were maintained as described by Collier [28]. Eight clutches of eggs were used for this study; each cell division was examined in 3–8 embryos, in most cases representing at least two different clutches. Developmental ages were recorded with reference to the onset of first, second, third or fifth cleavage, all of which occur with predictable relative timing. A small minority of embryos which initiated the reference cleavage outside of a ten-minute median time window were removed and cultured 5–6 days alongside controls; in every case, over 90% of both groups developed formed normal veliger larvae.

Descapsulated embryos were cultured at air-conditioned room temperature of 23±1°C in 108% Jamarin U artificial seawater (JSW; Jamarin Laboratories, Japan) that was 0.45 µm filtered, pasteurized, cooled and aerated by shaking. At 15- or 20-minute intervals during the desired time window, samples of 2–5 embryos were fixed for an hour at room temperature in 90% JSW containing 2.5% paraformaldehyde and 0.1% Tween-20. Fixed embryos were washed in water, incubated in methanol for 15–30 minutes, gradually rehydrated, and stained overnight at 4°C in 1 µg/ml Hoechst 33528 and 0.1% Tween-20. After another wash, they were cleared in glycerol and mounted under clay-feet-supported coverslips in 80% glycerol, 4% n-propyl gallate, 20 mM Tris pH 9. With this mounting method, embryos could be rolled around to permit observation from different sides. Embryos were observed with epifluorescence optics under UV illumination on a Leitz compound microscope with a 25× or 40× objective. Images were acquired by a ‘video camera lucida’ method in which fluorescence was imaged with a video camera, and the outlines of nuclei were traced onto transparent plastic sheets fixed to the video monitor. All cells were identified in each fixed embryo. The timing of each cell's division was estimated as roughly the mean age of embryos in which the cell was observed in metaphase, anaphase or telophase.


Most early cell divisions in Ilyanassa are oblique to the animal-vegetal axis. Such a division is described as dexiotropic if the clockwise daughter cell (in animal pole view) is the closer to the animal pole; in a laeotropic division the counterclockwise cell is closer to the animal pole. Some divisions are not oblique to the AV axis; these divisions are either transverse (spindle parallel to the embryo's equatorial plane) or longitudinal (spindle perpendicular to equator). Most cells are named following convention [6], [29], [30]. Micromere-derived cells are named according to their previous lineage history (encapsulated in the name of their immediate progenitor), and the relative positioning of sister cells along the egg axis. The upper sister (closer to the animal pole) is always given a terminal superscript 1, and the lower sister a terminal superscript 2. Where a transverse division places sister cells at the same level of the egg axis, they are usually numbered with reference to the chirality of the previous division: if the mother cell formed by a dexiotropic division, then the clockwise product of this cell's next division is given a superscript 2 and the counterclockwise product a superscript 1. Following Conklin [6], this nomenclature is replaced in the case of the third-quartet derivatives with terms relating cell identity to the secondary axis. In general, when referring collectively to cells formed by analogous divisions in two or three quadrants, I use terms such as 1abc1; this example signifies the three upper cells formed by the division of 1a, 1b, and 1c. Following Costello [31], analogous cells in all four quadrants are referred to collectively using “m” to denote micromeres and “M” to denote macromeres (1m1, for example, is used instead of 1abcd1). Finally, Conklin [6] named derivatives of the mesentoblast 4d on the basis of their supposed fates in Crepidula; these fates need to be confirmed using modern methods, and 4d derivatives are renamed here on the basis of lineal relationships and the positioning of sister cell nuclei [25]. As the 4d lineage seems to undergo identical division patterns in Crepidula and Ilyanassa, it is easy to refer back to the original cell names. Throughout this paper the secondary axis of the embryo will be called ‘dorsoventral’ according to a current, arbitrary convention; it should be noted that this axis does not correspond by cell lineage distribution to any definitive anatomical axis (Figure 2A). The ‘dorsal’ pole of the early embryo's secondary axis is marked by the center of the bilaterally symmetric 4d clone. Cells that lie on dorsal, ventral, left or right meridians are said to be positioned radially, while cells positioned forty-five degrees away from these meridians are positioned interradially.


First two cleavages and polar lobe segregation

Early cleavage in Ilyanassa has previously been described [12], [32], and is summarized in Figure 3A. Like the egg's two meiotic divisions, first cleavage is accompanied by polar lobe formation [33], [34]. The first cleavage furrow initiates at the animal pole and progresses vegetally in a roughly meridional plane. The polar lobe relaxes into the CD cell, which then forms another polar lobe at second cleavage, shunting it to its western daughter cell, D. Second cleavage is slightly asynchronous, CD entering mitosis before AB [32]. As in the first cleavage, AB and CD form meridional cleavage furrows that progress from animal to vegetal. The planes of second cleavage are almost orthogonal to the first cleavage plane, so the animal-pole surfaces of the four quadrant-founder cells (A, B, C, D) end up bordering each other in a regular fashion, close to the site of polar body formation. A slight laeotropic skew between the AB and CD cleavage planes place the A and C cells visibly higher on the AV axis compared to B and D; this geometry is common among gastropods and many other spiralians, and presages the overt dexiotropic chirality of the following cleavage. The interval between first and second cleavage lasts about 75–80 minutes. In subsequent generations, the same cell cycle length is observed in the quadrant founder cells (M) and in their macromere daughters (1M); all other cell cycles, with the exception of some cells in the 4d lineage, are longer. Temporal cell division patterns in the four micromere quartets are charted in Figures 4, 55, 66, 77.

Figure 4
Cell lineage table of the first micromere quartet.
Figure 5
Cell lineage table of the second micromere quartet.
Figure 6
Cell lineage table of the third micromere quartet.
Figure 7
Cell lineage table of the fourth micromere quartet.

First quartet (Figure 4)

At third cleavage, the quadrant founder cells divide asymmetrically to form the first quartet micromeres around the animal pole (Figure 3A). Mitotic spindles form synchronously in C and D, with A and B together lagging a few minutes behind. In spite of the apparent mitotic synchrony, the D cell consistently begins cytokinesis slightly before C. The dorsal micromere 1d is significantly smaller at birth than 1a, 1b and 1c [12], [35].

The first quartet cells are immediately distinguished from their macromere sisters by a retarded division schedule: the 1a, 1b, and 1c cells divide about 30 minutes after the macromeres (105 minutes after their formation), and 1d divides after an additional delay of 30–35 minutes. The division of each micromere is laeotropic and unequal, separating a large cell (1m1) at the embryo's apex from a much smaller ‘turret cell’ (1m2) distal to the animal pole. Following cytokinesis, the nuclei of the 1m1 cells grow as big as the nuclei in their mother cells, while the turret cell nuclei remain very small and condensed. The turret cells typically remain in interphase through the next eleven hours of development at least.

Compared to the first quartet micromeres, cell cycle lengths are more nearly uniform among their four apical daughters: the 1abc1 cycles last about 145 minutes (1a1 tending to divide first), and the 1d1 cycle about 155 minutes. Due to the delay in its formation, however, 1d1 divides about 45 minutes after its neighbors (450 minutes after first cleavage; Figures 8, 9).9). All four 1m1 cells divide dexiotropically. The 1abc1 cells each divide with a slight asymmetry, forming a smaller ‘apical cell’ (1abc11) and a larger ‘basal cell’ (1abc12). As previously noted [12], 1d1 divides with a more pronounced asymmetry which, compared to the other three divisions, is reversed in polarity: the apical daughter cell 1d11 is about the same size as the three other apical cells, while the basal cell 1d12 does not appear to be much bigger than the tiny turret cells. Within each quadrant, the dexiotropic division of 1m1 places apical and basal sister cells roughly forty-five degrees out of register with each other around the egg axis. With reference to the secondary axis (approximately defined at this stage by the ‘dorsal’ 3D macromere) the basal cells 1m12 are positioned on dorsal, ventral, and lateral radii of the head rudiment. Their apical sisters occupy interradial positions (1cd11 right- and left-dorsal, 1ab11 left- and right-ventral), in register with the turret cells (1m2).

Figure 8
27-cell embryo, 410 minutes.
Figure 9
35-cell embryo, 450 minutes.

The 1abc12 cells are the next first-quartet derivatives to divide, between 515 and 555 minutes after first cleavage (Figure 10). Invariably the 1a12 cell divides first. These divisions bring about the first sign of dorsoventral asymmetry among the 1abc lineages. According to the rule of alternating division chirality that predominates in early cleavage, all three cells should divide laeotropically. 1a12 tends to follow this rule, but the contralateral 1c12 cell tends to divide dexiotropically; the division of 1b12 is roughly longitudinal. Thus, 1a122 and 1c122 both tend to be shifted ventrally relative to their sisters.

Figure 10
41-cell embryo, 510 minutes.

About 30 minutes after the division of 1abc12, the apical cells 1abc11 begin to divide (Figure 11). The left- and right-ventral 1a11 and 1b11 cells divide more or less in synchrony, followed by 1c11 about 20 minutes later. The 1d11 cell divides after an additional delay of about 65 minutes. The cell cycles of 1ab11, 1c11, and 1d11 are respectively estimated as 160, 180, and 200 minutes long. All four cells divide laeotropically; as a result, the ‘apical rosette cells’ 1m111 end up in approximately radial positions around the animal pole, with the ‘peripheral rosette cells’ 1m112 approximately interradial like their mothers. This arrangement is not initially evident due to close spacing of the 1m11 daughter nuclei (Figures 11, 1212, 13),13), but is consistently seen at later stages (Figures 14, 1515, 1616, 1717, 18, 18, 2020, 21).21). None of these cells were observed to divide again (cell cycles>370 minutes).

Figure 11
45-cell embryo, 540 minutes.
Figure 12
54-cell embryo, 635 minutes.
Figure 13
59-cell embryo, 635 minutes.
Figure 14
67-cell embryo, 670 minutes.
Figure 15
67-cell embryo, 730 minutes.
Figure 16
67-cell embryo, 760 min.
Figure 17
71-cell embryo; 800 minutes. Dorsal view.
Figure 18
70-cell embryo; 811 minutes.
Figure 20
84-cell embryo, 955 minutes.
Figure 21
84-cell embryo, 955 minutes.

Prolonged cell cycles also occur in the daughters of 1abc12. In particular, the inner cells 1abc121 were never observed to divide (cell cycles>420 min). Their sisters 1abc122 divide 860–905 minutes after first cleavage (cell cycles≈340–360 min) (Figure 19). Unlike all previous divisions in the first quartet, these divisions appeared to be more or less perfectly transverse in all three cells. The 1b122 cell is consistently the first to divide, preceding 1ac122 by 10–20 minutes.

Figure 19
80-cell embryo; 890 minutes.

Second quartet (Figure 5)

The second quartet micromeres (2m) are formed by laeotropic division of the macromeres 1M. This round of divisions is slightly asynchronous, led by 1CD. The 2m cells are uniform in size, and perceptibly bigger than the 1abc cells. The second quartet cells divide at the same time as their macromere sisters, in the transition from the 16-cell to the 24-cell stage. Division of the 2m cells occurs with a slight asynchrony matching the timing of their birth; these divisions are all equal and dexiotropic, but nearly transverse [12] (Figure 8).

The second round of division in the second quartet takes place 445–495 minutes after first cleavage (Figure 9). The daughter cells of each second quartet micromere divide in approximate synchrony with each other. Between quadrants, however, a novel timing difference now appears. 2d1 and 2d2 always divide a few minutes ahead of the 2ac daughter cells, and 2b1 and 2b2 divide only after another 15–20 minutes. In the 2a, 2b, and 2c lineages, both sister cells divide dexiotropically and very unequally, but with opposite polarities: the upper 2abc1 cells bud off a small cell (2abc11) toward the animal pole, while the lower 2abc2 cells bud off a small cell (2abc22) toward the vegetal pole. In the D quadrant, the division geometry is different [12]. Both cells divide dexiotropically as in the other quadrants. The lower cell 2d2 divides very unequally like 2abc2, but the division of the upper cell produces daughters (2d11 and 2abc12) that are roughly equal in size (not shown, but note nuclei in Figures 10, 1111).

The small cells 2abc11 and 2m22 have small, condensed nuclei and display very long cell cycles (>475 min), similar to the turret cells and 1d12 in the first quartet. In contrast to the small 2abc11 cells, 2d11 divides approximately in synchrony with its equal-sized sister (cell cycles 165–175 minutes) (Figure 12). All the other large cells of the second quartet also divide around the same time. The wave of mitoses is consistently begun by 2c12. All of these divisions are equal and approximately longitudinal. As their precursors did in the last round of divisions, the 2b12 and 2b21 cells both lag behind their counterparts in other quadrants, this time by 15–40 minutes (Figure 13). After this series of divisions is complete the 2a, 2b, and 2c clones each number six cells, and 2d has formed seven. The arrangement of 2abc derivatives has at this stage become bilaterally symmetric. This is due to a second deviation of sister cell positioning from the rule of spiral cleavage: the upper division products in the A and C quadrants are all shifted ventrally with respect to their sisters, while in the B quadrant the upper cells are displaced to each side (Figures 12, 1313, 1515, 1616, 20).20). In the 2d clone, the six latest-born nuclei form a unique pattern consisting of three laeotropically skewed columns; as development proceeds, the lower-left nucleus in this array (2d122) tends to be displaced away from its sister toward the vegetal pole (Figures 14, 16162121).

Third quartet (Figure 6)

The third quartet micromeres are formed by the dexiotropic division of the 2 M macromeres. The 3d cell is always the first to form (315–330 minutes after first cleavage), and the others follow within five minutes. The third quartet micromeres are situated interradially, nestled between the 2m clones; thus, the quartet consists of two dorsolateral cells (3cd) and two ventrolateral cells (3ab) (Figures 8, 9).9). A distinct bilateral symmetry is immediately observed in the third quartet's collective division pattern [12], and no rotational symmetry about the egg axis is ever apparent. The first round of divisions is strikingly asynchronous, led by the simultaneous division of the dorsal cells 3c and 3d (480–500 minutes after first cleavage) (Figure 10). These divisions are longitudinal and very unequal, budding off small vegetal daughter cells (3cd2) with very small, condensed nuclei and long (>450 min) cell cycles. Approximately an hour after this division, the ventral cells 3a and 3b divide in near synchrony with each other (Figure 11). These divisions are also longitudinal, but equal (Figure 12).

The 3cd1 cells are next to divide, beginning with 3c1 (635–650 minutes after first cleavage) and followed 5–15 minutes later by 3d1 (Figures 12, 13).13). Each cell divides transversely and equally. This results in the formation of two bilateral pairs of cells: the more dorsal (medial) cells are called 3cd11 and their ventral (lateral) sisters are 3cd12 (Figures 14, 1515).

3cd11 and 3cd12 divide again about 3.5 hours later, before any other divisions in the third quartet have occurred. All four cells divide in the manner of 3cd, the lower daughter cells again being very small (Figures 1921).21). This round of divisions is led by 3c11 around 800–830 minutes after first cleavage (Figure 17); the subsequent division timing of 3d11 and the 3cd12 pair was not accurately determined, but all divisions were completed within the next 50 minutes.

The daughters of 3ab finally begin to divide around 890 minutes after first cleavage, nearly six hours after they were formed (Figure 19). In the few cases examined, the upper cells were seen to divide first. The division is transverse and equal in all four cells.

Fourth quartet/Mesentoblast 4d (Figure 7)

The 3D cell divides asymmetrically about 430 minutes after first cleavage, producing the large mesentoblast cell 4d and the very large vegetal cell 4D. The 4d cell is budded laeotropically toward the micromere cap, and occupies a radial position in line with the middle of the 2d clone (Figure 9). The 4abc cells are formed much later (635–675 minutes after first cleavage) in roughly transverse laeotropic divisions (Figures 15, 16).16). Neither the 4abc cells nor any of the remaining macromeres were ever seen to divide again (cell cycles>300 min). The 4d clone, by contrast, proliferates much faster than any other lineage during the next eight hours of development.

The division of 4d is equal and transverse, forming left and right daughter cells 4dL and 4dR (Figure 10) This division occurs around 490–510 minutes after first cleavage. Both left and right cells divide again in close synchrony 615–630 minutes after first cleavage (Figure 12). Both divisions are unequal [12], the lower cells 4dLR2 having much smaller nuclei than the upper cells 4dLR1 (Figure 13). Similar to the small, mitotically quiescent ectodermal cells with condensed nuclei, the lower cells were never seen to divide (cell cycle length>320 minutes). According to Conklin these small cells (termed E1 and E2) are the precursors of the hindgut in Crepidula [6].

The upper cells, representing the putative mesoblast (Me1 and Me2 in Crepidula) divide again after only about an hour (675–700 minutes after first cleavage). These divisions are synchronous and oriented at bilaterally symmetric angles to the egg axis. The more ‘ventral’ (or ‘anterior’) daughter cells 4dLR11 have somewhat smaller nuclei than their sisters 4dLR12 (Figure 15).

Both pairs of 4dLR1 daughter cells divide again very shortly (Figures 1719).19). In most cases the posterior pair (4dLR12) was seen to have completely divided before the onset of mitosis in the anterior pair. The reverse was seen in a few cases; in every case division synchrony was closer within than between bilateral cell pairs. The 4dLR12 pair divide along axes that are again roughly parallel to the median plane, though some variation is seen. A slight left–right asynchrony was consistently seen, as the right-sided 4dR12 generally had completed its division while its neighbor was in anaphase. Both 4dL12 and 4dR12 give rise to a more anterior cell with a larger nucleus (cell 4dLR121) and a more posterior cell with a much smaller nucleus (cell 4dLR122). The timing of these divisions ranged from 740 to 810 minutes after first cleavage. The anterior pair 4dLR11 always appeared to divide equally, along oblique axes that were roughly perpendicular to those of the previous division.

Finally, the large anterior nuclei (4dLR121) produced by the last anteroposterior division divide again within the next 1.5 hour (Figures 20, 21).21). Once again, the anterior daughter cells 4dLR1211 have much larger nuclei than their sisters. The 4dLR1211 cells have previously been reported to undergo two more stem-cell-like divisions over the next fourteen hours, budding two more pairs of small, slow-dividing cells toward the midline; in the meantime, the 4dLR112 lineages likewise give rise to three cells each [24].


I. Overview of developmental stages

The cell lineage tables presented here show the timing of all divisions from the four-cell to the 84-cell stage. Proliferation rates are heterogeneous among micromere lineages but tend to decrease over the first three or four cell cycles; thus, total cell number increases almost linearly, with irregular fits and starts (Figure 22). The 28-cell stage is marked by the birth of 4d [12], about seven hours after first cleavage. Over the next five hours, 37 more divisions occur in such close succession that few persistent stages can be defined in terms of total cell number. Divisions during this five-hour period include only the 1d1 cell and the 1m11 tier of the first quartet. In the second quartet, two rounds of division during this period bring the number of 2m derivatives to 25. The second round of 2m divisions (division of 2d11, 2m12, 2m21) is spread over an hour or more, an interval that also includes division of the bilateral 3cd1 and 4dLR pairs, as well as the 3ABC macromeres. Finally, division of the 4dLR1 pair brings the embryo to 65 cells. Two more rounds of division in the 4d lineage then mark well-defined 67-cell and 71-cell stages, collectively spanning a three-hour period during which the fifty-four ectoblast cells do not divide. The next cells to enter mitosis are the four 3cd1 derivatives and the two 4dLR121 cells. Shortly thereafter, division of 3ab1, 3ab2, and 1abc122 bring the total cell count to 84. The latest dividing cells among this group were consistently 3ab2. Less than an hour later, continuing proliferation in the 2d lineage obscured cell identities among 2d, 3cd and 4d derivatives.

Figure 22
Cell numbers charted over time.

During the last five hours of development followed here, the ectoblast begins to spread out over the macromeres. While this process of epiboly has still not been followed to completion, it seemed in this study to begin 10–11 hours after first cleavage (compare Figures 111313 with Figures 1421)21) After this stage, nuclei of the first quartet are widely spaced, and distributed in a more regular and stereotyped pattern than before. Another descriptive study reported a similarly abrupt flattening and spreading of the micromere cap at a stage (14 hours at 20°C) coinciding with division of the 3C cell [35]. The present study shows that this stage also corresponds closely with the divisions of 2d11, 2m12, 2m21, 3cd1, 3AB, and 4dLR1, and that the progress of epiboly over the next three or four hours coincides with mitotic quiescence of the ectoblast.

II. Reproducibility of cleavage pattern

Up to the 84-cell stage, spatial relationships of nuclei are reproducible enough that the trained eye may identify each cell unequivocally in almost any embryo. Recognition of cell patterns is greatly facilitated by a set of highly unequal divisions whose smaller daughter cells (1m2, 1d12, 2abc11, 2d22, and 3cd2) undergo prolonged cell cycle arrest and have small, condensed nuclei. Occasionally one of the 1m2 cells and/or 1d12 was found to have divided several hours earlier than usual (Figures 1416,16, ,18),18), but daughter cells of 2abc11, 2m22, or 3cd2 were never observed.

One might expect that due to metabolic noise, cell division timing should become progressively less predictable as development proceeds; such time-dependent decorrelation has indeed been noted in an early study of spiralian cell lineage [37]. Even at the latest stages examined in this study, however, the timing of at least some cell divisions among embryos within a brood was highly predictable. For example, between 855 and 905 minutes after first cleavage, only two embryos with an interphase 1abc122 nucleus occurred among nine samples (three embryos per sample) that also included the respective cell's postmitotic daughters. Such persistent correlation of division timing among distantly related cells might reflect intercellular coupling of cell cycle progression. Note that this analysis excluded individuals deviating from within-brood synchrony at the earliest stages (see Material and Methods).

It is noted that the overall rate of development observed in this study (which was consistent over a period of over three years) significantly exceeded that reported in a previous study of meiosis and early cleavage [32]. The cause of this discrepancy is unexplained, but might reflect differences among wild snail populations. In general, timing variation between broods from a given population is expected to be much greater than within a single brood, even at early stages. This is suggested by prior observations of between-brood variation in the timing of 3D signaling [16], [38].

III. Between-tier differentiation of cell cycle rates: preliminary comparisons within Caenogastropoda

To begin understanding the evolution of embryonic cell lineage in Ilyanassa, some elements of cell division pattern can be compared with related taxa. Ilyanassa belongs to the enormously successful superorder Caenogastropoda, whose monophyly is robustly supported [39]. Cleavage has been described to a comparable extent in two other caenogastropod genera, Bithynia [27], and Crepidula [6], [11], [26]. Although the relationship of these two genera and Ilyanassa is not yet well resolved by molecular phylogenetic methods [39], a comparison of cell lineage parameters is nonetheless illuminating.

One aspect of cell lineage that can be meaningfully compared between spiralian taxa is the relative timing of mitosis between cell tiers [40]. In gastropod species in which cleavages have been timed, all of the cells AB/CD, M, 1M exhibit a more or less uniform cell cycle length characteristic of the species. By normalizing to the common cell cycle length of these early macromeres, tier-specific cell cycles can be reasonably compared between taxa. Comparing the early cell cycles of Ilyanassa, Bithynia, and Crepidula shows that each genus has a unique pattern of tier-specific cell cycle lengths (Figure 23A). Ilyanassa exhibits the most complex pattern, with neither a single abrupt transition to longer cell cycles as in Bithynia, nor gradually increasing cell cycle lengths as in Crepidula. Extending the comparison of normalized cell cycle lengths between Ilyanassa and Bithynia suggests several generalities (Figure 23B). In each genus, some ectoblast tiers divide faster and others slower. Sister-cell pairs tend to exhibit concordantly shorter or longer cell cycles in a given genus (4/4 instances). In both genera, cell cycle length increases progressively in the second-quartet clones. Two first-quartet tiers (1m2 and 1m122) have extremely long cell cycles in Ilyanassa but not in Bithynia. Ignoring these two outliers, the mean normalized cell cycle length is very similar in the two genera (1.8 and 1.9, where the cell cycle length in AB/CD, M, 1M is defined as 1).

Figure 23
Between-tier cell cycle differentiation in Ilyanassa and two other caenogastropods.

Relative proliferation rates of the three micromere quartets can be further compared by using cell numbers in the whole embryo as a yardstick (Figure 23C). The only conspicuous difference among the three caenogastropod genera is a relatively high growth rate of the first quartet in Bithynia. This disparity, which largely reflects the unusually short cell cycles of 1m2 and 1m122 in Bithynia, becomes increasingly pronounced up to the latest stages examined; if it persists through later development, it might lead to heterochrony in organogenesis, akin to that recently described among basommatophoran pulmonate snails [41].

IV. Within-tier cell cycle differentiation: earlier departures from radial symmetry in Ilyanassa compared to other caenogastropods

Soon after signaling from 3D/4d, the form and timing of cell division begins to differentiate within most ectoblast tiers. Cells within a tier can be considered serially homologous, owing to their synchronous and topologically equivalent origins. The homology of cells within each tier is further supported by the tier-specific localization of numerous, randomly selected RNA species (by contrast, no quadrant-specific RNAs have yet been found) [5]. In the absence of D quadrant specification, the four cells of each tier divide with identical timing and geometry [12]. Within-tier differences in developmental potential and early mitotic behavior are evidently imposed on a homogeneous ground state by two sources. First, the polar lobe exclusively distinguishes the D lineage, and could in principle contribute unique information to the D lineage cell in each tier. Intrinsic control of D quadrant micromere development has been shown only for the 1d cell [12], [21] (see section VII). Second, dorsal and lateral cells (generally in the C and D quadrants, and in some tiers also the A quadrant) are influenced by signaling from 3D/4d [12], [21]. Thus, both lineage and position contribute to within-tier differentiation; both mechanisms are expected to generate differences correlated more or less exactly with cell position along the secondary axis. As described below, most within-tier differences in cell division pattern up to the 80+ cell stage are indeed correlated with dorsoventral cell position.

Cell positions within a tier are remarkably well defined in relation to the secondary axis. During the period when this axis is inductively patterned by the 3D macromere and/or its daughter 4d, the 1m1 and 2m cells occupy radial positions (dorsal, lateral, ventral) while the 1m2 and 3m cells occupy interradial positions (dorsolateral, ventrolateral) (Figure 24A; see also Figure 3B and Figure 8). These approximate geometric relationships between clones persist throughout the period of development studied here (with the secondary axis defined by the center of the 4d clone) (Figure 24B). Fate maps of Ilyanassa, Crepidula, and their distant relative Patella corroborate these assignments of relative axial position, at least within the bilaterally symmetric regions of the larval snail (head, velum, mouth, foot) [8][11]. Further divisions of the 1m1 cells generate both radially distributed tiers (1m111; 1m12) and the interradial 1m112 tier (Figure 24C–F). While the fates of these sublineages have not mapped by intracellular dye tracing, cell lineage analysis in Crepidula has shown them to make predictable contributions to larval and adult ectoderm [6], and their initial and ultimate axial dispositions are conserved outside of caenogastropods [7].

Figure 24
Disposition of micromere sublineages along the secondary axis.

Within-tier division asynchronies in Ilyanassa, Crepidula, and Bithynia are shown in Figure 25. Such division asynchrony is most prevalent in Ilyanassa, and least prevalent in Bithynia. After the birth of 3D, thirteen of fourteen examined cell tiers divide asynchronously in Ilyanassa. Only six of these tiers are reported to divide asynchronously in the same order in Crepidula. In each of these six cases, the magnitude of division asynchrony is greater in Ilyanassa, and no tier divides with a more pronounced asynchrony in Crepidula. In three cases (2m12, 2m21, 3m2) within-tier asynchronies occur with different order between Ilyanassa and Crepidula.

Figure 25
Within-tier cell cycle differentiation in three caenogastropods.

While most within-tier cell cycle diversification is correlated with cell position along the secondary axis, division asynchrony may also occur between left- and right-sided cells within a tier. Four subtle but stereotyped bilateral division asynchronies were found in Ilyanassa which have never been reported in other gastropod species. Curiously, the right-sided cell divides first in each of these cases. First, the right-dorsal 1c11 cell has a shorter cell cycle than its left-sided counterpart 1d11. Second, the right-sided 2c12 cell, though formed at about the same time as the 2a12 and 2a21 cells on the left, predictably leads the third round of second-quartet divisions. This asynchrony presages a quantitative difference in the morphogenetic contributions of 2c and 2a: the right-sided trunk ectoderm, formed by 2c, exhibits a higher growth rate during late gastrulation and organogenesis [42]; correspondingly, 2c forms a larger region of the mantle edge than 2a [10], and shell development can be severely disrupted by ablating either 2c or 2d, but not 2a or 2b [19]. Finally, two other right-hand cells in Ilyanassa (4dR12 and 3c11) were also found to divide precociously, suggesting a general right-sided growth advantage in multiple lineages. While none of these bilateral asynchronies have been reported in other gastropods, it is noted that the right-sided and dorsal 2cd22 cells in Bithynia divide before the left-sided and ventral 2ab22 (this division has not been described in Ilyanassa or Crepidula). Preferential right-sided growth and organogenesis is widespread and likely primitive among gastropods [43], suggesting that Ilyanassa has evolved an earlier onset of differential growth along the left-right axis as well as along the secondary axis.

V. Within-tier differentiation in cell division asymmetry: bigger differences in Ilyanassa

The geometry of cell division as well as its timing can vary predictably within cell tiers. Three tiers in Ilyanassa were found to exhibit quantitative within-tier differences in the geometric asymmetry of cell division; in each case the difference is precisely correlated with cell position along the secondary axis. Each of these differences is reduced or absent in Crepidula, Bithynia, or both. One case involves the dexiotropic division of the 1m1 cells. In Ilyanassa, the 1d1 cell divides with an asymmetry that is reversed with respect to the asymmetry of the 1abc1 divisions. In Crepidula, the 1abc1 cells divide as in Ilyanassa; however, the asymmetry of the 1d1 division is not reversed but is merely reduced compared to 1abc1 [26]. No difference in division geometry within the 1m1 tier was noted in Bithynia.

A second within-tier difference in division asymmetry occurs in Ilyanassa when 2d1 divides equally and the 2abc1 cells all divide very unequally. In Crepidula and Bithynia this differentiation is present in a reduced (or rudimentary) form: the 2abc1 cells divide as in Ilyanassa, while 2d1 divides with a slight asymmetry. In all three genera, cell cycle lengths of the 2m11 and 2m12 daughter cells are inversely correlated with inherited cell volume; thus, the 2d11 cell inherits a growth advantage which is greatest in Ilyanassa. Reduced asymmetry of the 2d1 division relative to 2abc1 may be an ancestral condition, as it is also reported in basal gastropods [30], [44] as well as in two other molluscan classes [45], [46]; however, all four 2m1 cells divide with the same extreme asymmetry in some other gastropod taxa [7], [47], [48], as well as in polyplacophorans [49], [50]. Ilyanassa is the only mollusk in which 2d1 is reported to divide equally.

Finally, in Ilyanassa the 3cd cells undergo highly asymmetric divisions while the 3ab cells divide equally; in Crepidula, by contrast, all four cells divide with the same slight asymmetry. Bithynia displays a third pattern: 3cd bud off small cells toward the vegetal pole, while 3ab bud small cells toward the animal pole. Both the Ilyanassa and Bithynia patterns have been reported in non-caenogastropod snails [7], [30], [44], [47], [48]; the Crepidula pattern is therefore most likely derived from one of the two others.

VI. Dorsoventral organization of cell division axes: breaking the golden rule of spiral cleavage

Another aspect of division geometry is the orientation of division axes relative to the embryo's spatial coordinates. At early stages, oblique division axes generate a chiral, rotationally symmetric pattern; later, division axes within many ectoblast tiers show a collective bilateral symmetry and dorsoventral asymmetry [6], [7]. In general, dorsoventral organization of cell division axes seems to be more prevalent in Ilyanassa than in Crepidula. For example, the collective division pattern of the 1abc12 group exhibits a dorsoventral asymmetry in Ilyanassa: although the spatial relationships of daughter nuclei vary considerably, both of the lower daughter nuclei (1ac122) are shifted ventrally in most cases (Figure 26). In Crepidula, the 1abc12 cells all divide dexiotropically (surprisingly, as their mothers do too) [6], [26]. In Bithynia the 1m12 cells are all reported to divide longitudinally.

Figure 26
Dorsoventrally polarized division axes in the 1abc12 cell group.

In several other instances, sister nuclei in related tiers seem to be distributed chirally in Crepidula but achirally in Ilyanassa (the 3ab daughters) or with a dorsoventral asymmetry in Ilyanassa but not in Crepidula (collectively, the 2ac12 and 2ac21 daughters). Comparison of these traits will require morphometric analysis beyond the scope of the present study.

VII. From fuzzy to rigorous during ontogeny and phylogeny: how early differentiation precedes, then preempts, cell fate decisions

The spatiotemporal geometry of cleavage in mollusks clearly has the potential to contain a great deal of information, but for the most part we do not know its developmental significance. In one instance, however, recent experimental studies in Ilyanassa and Crepidula have revealed an evolutionarily labile coupling between early differentiation and fate specification. Intriguingly, they suggest a stepwise evolutionary pathway in which accelerated phenotypic differentiation has functionally supplanted a primitive signal in regional specification.

In all examined gastropods, and in two other molluscan classes, part or all of the 1d12 clone enters a persistent cell cycle arrest while corresponding cells of the 1abc12 lineages proliferate to form definitive head structures [6], [7], [22], [27], [45], [47], [49], [53]. In both Ilyanassa and Crepidula, this arrest is anticipated by mitotic delays in the 1d, 1d1, and 1d12 cells [6], [12], [26] (Figure 25). In both genera, the 1d cell is born smaller than 1abc, indicating an intrinsically determined difference [12], [26]. Experiments on Ilyanassa have shown that the limited size and division rate of 1d depend on polar lobe segregation to the D cell [12]; the polar lobe also prevents 1d from forming an eye [21], evidently by limiting its inherited cytoplasmic volume [35].

Does early differentiation of 1d size and cell cycle length play any role in determining 1d fate in Crepidula? As in Ilyanassa, the mother cell of 1d inherits a polar lobe at second cleavage in Crepidula. By contrast to Ilyanassa, however, removal of the polar lobe in Crepidula does not appear to endow 1d with eye-forming potential, indicating that a polar lobe-independent mechanism must restrict 1d potential [15]. A cell-extrinsic mechanism of 1d fate restriction undoubtedly operates in more basal gastropods, where the D cell inherits neither a polar lobe nor any other determinant of developmental potential at second cleavage [52]. This ancestral mechanism of 1d fate restriction, which likely operates also in Crepidula, appears to have been completely replaced by an earlier, cell-intrinsic mechanism in Ilyanassa [21].

Specification of 1d is necessarily tied to that of the dorsal signaling cell 3D; the mechanism of this linkage, however, appears to have changed. In Ilyanassa, the early differentiation and fate restriction of 1d is preceded by determination of the D quadrant founder cell as precursor to 3D [12], [21], [51]. Although the polar lobe in Crepidula is likewise consistently inherited by the D cell, 3D identity is not specified until the 24-cell stage, as a result of inductive signaling from first-quartet micromeres to one of the 3M macromeres; this inductive mechanism, like the extrinsic specification of 1d, is primitive [50], [52]. The polar lobe in Crepidula evidently acts just to bias 3D-inductive signaling to one macromere [15]. The mechanism of this bias is unknown. The early differential behavior of the 1d lineage suggests that 1d itself could be involved in biasing 3D induction. Supporting this notion, selection of one macromere as 3D in the pulmonate Lymnaea and the basal gastropod Patella depends on the presence of a first-quartet micromere belonging to the same quadrant [55].

Our knowledge of the D lineage in Crepidula suggests a fuzzy-logical mode of lineage specification in which phenotypic divergence precedes a definitive cell fate decision. In theory, such an early differentiation event may itself act as a cue to bias cell fate; alternatively, it may depend incidentally on a preceding cue that acts in parallel to bias cell fate. In particular, further study is needed to find out whether an intrinsic quality of 1d biases 3D identity in Crepidula. Regardless of its primitive developmental role, such an early differentiation event provides an avenue for the future evolution of accelerated fate restriction. This is illustrated by the case of Ilyanassa, where D cell division asymmetry has been exapted to functionally replace a primitive head-patterning signal. Polar lobe segregation too can be regarded as an early cell differentiation event that has, in a gradual or punctuated manner, come to replace inductive specification of 3D during the evolution of Ilyanassa.

VIII. Overview and Prospects

Previous embryological studies of mollusks and other spiralians have shown wide variation in the complexity of early cleavage pattern. Some taxa, including representatives of two of the most basal gastropod clades, display almost no within-tier differentiation up to the latest stages examined (88+ cells)[29], [46], [47]. By contrast, a number of others (including scaphopod and some bivalve mollusks, as well as many annelids) deviate from rotational symmetry to a much greater degree than Ilyanassa [43], [44], [55], [56]; at the furthest extreme, the stereotyped cleavage patterns of cephalopods display no trace of spiral geometry [57]. In general, complex patterns of early lineage-specific differentiation seem to have evolved at the bases of several higher-level spiralian clades [50], and transitional states are largely unknown.

The variation documented here among caenogastropods suggests that it may be possible in some cases to reconstruct the evolution of complex early cell division patterns. By focusing on a quantitative aspect of differential behavior among serially homologous cells (i.e., cell cycle length within a tier), one can detect small evolutionary changes in the behavior of single cells; as shown in the case of Ilyanassa, many such small changes may summate along body axes to suggest a global evolutionary trend. Extending this analysis in the context of a robust, independent (i.e., molecular) phylogeny may shed light on the evolutionary dynamics of early developmental complexity.


Special thanks are due to Gary Freeman for support and guidance throughout the completion of a doctoral thesis and beyond. I thank B. Bowerman and C. Q. Doe for support during the completion of the manuscript. Additional thanks are owed to G. Freeman, S. Q. Schneider, J. D. Lambert, and two anonymous reviewers for critical reading of the manuscript.


Competing Interests: The author has declared that no competing interests exist.

Funding: This work was funded in part by the National Science Foundation (grant IBN-9982025 to G. Freeman), National Institutes of Health (grant R01GM049869 to B. Bowerman), and a Howard Hughes Medical Institute fellowship to Chris Q. Doe. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.


1. Dunn CW, Hejnol A, Matus DQ, Pang K, Browne WE, et al. Broad phylogenomic sampling improves resolution of the animal tree of life. Nature. 2008;452:745–9. [PubMed]
2. Henry JJ, Martindale MQ. Conservation and innovation in spiralian development. Hydrobiologia. 1999;402:255–265.
3. Maslakova S, Martindale MQ, Norenberg JL. Fundamental properties of spiralian cleavage program are displayed by the basal nemertean Carinoma tremaphoros (Palaeonemertea; Nemertea). Dev Biol. 2004;267:342–360. [PubMed]
4. Ackermann C, Dorresteijn A, Fischer A. Clonal domains in postlarval Platynereis dumerilii (Annelida: Polychaeta). J Morph. 2005;266:258–280. [PubMed]
5. Kingsley EP, Chan XY, Duan Y, Lambert JD. Widespread RNA segregation in a spiralian embryo. Evol Dev. 2007;9:527–539. [PubMed]
6. Conklin EG. The embryology of Crepidula. J Morph. 1897;13:1–226, plates I–IX.
7. Wierzejski A. Embryologie von Physa fontinalis L. Z Wiss Zool. 1905;83:502–706, plates XVIII–XXVII.
8. Render J. Fate maps of the first quartet micromeres in the gastropod Ilyanassa obsoleta. Development. 1991;113:495–501. [PubMed]
9. Dictus WJ, Damen P. Cell-lineage and clonal-contribution map of the trochophore larva of Patella vulgata (Mollusca). Mech Devel. 1997;62:213–226. [PubMed]
10. Render J. Cell fate maps in the Ilyanassa obsoleta embryo beyond the third division. Dev Biol. 1997;189:301–310. [PubMed]
11. Hejnol A, Martindale MQ, Henry JQ. High-resolution fate map of the snail Crepidula fornicata: the origins of ciliary bands, nervous system and muscular elements. Dev Biol. 2007;305:63–76. [PubMed]
12. Clement AC. Experimental studies on germinal localization in Ilyanassa. I. The role of the polar lobe in determination of the cleavage pattern and its influence in later development. J Exp Zool. 1952;121:593–626.
13. van den Biggelaar JAM. Dorsoventral polarity and mesentoblast determination as concomitant results of cellular interactions in the mollusk Patella vulgata. Dev Biol. 1979;68:462–471. [PubMed]
14. Martindale MQ. 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 Invertebr Reprod Dev. 1986;9:229–242.
15. Henry JQ, Perry KJ, Martindale MQ. Cell specification and the role of the polar lobe in the gastropod mollusc Crepidula fornicata. Dev Biol. 2006;297:295–307. [PubMed]
16. Lambert JD, Nagy LM. MAPK signaling by the D quadrant embryonic organizer of the mollusc Ilyanassa obsoleta. Development. 2001;128:45–56. [PubMed]
17. Gonzales EE, van der Zee M, Dictus WJ, van den Biggelaar J. Brefeldin A or monensin inhibits the 3D organizer in gastropod, polyplacophoran, and scaphopod molluscs. Dev Genes Evol. 2007;217:105–18. [PubMed]
18. Clement AC. The embryonic value of the micromeres in Ilyanassa obsoleta as determined by deletion experiments. I. The first quartet cells. J Exp Zool. 1967;166:77–88.
19. Clement AC. The embryonic value of the micromeres in Ilyanassa obsoleta as determined by deletion experiments. II. The second quartet cells. Int J Invert Reprod Dev. 1986a;9:139–153.
20. Clement AC. The embryonic value of the micromeres in Ilyanassa obsoleta as determined by deletion experiments. III. The third quartet cells and the mesentoblast cell, 4d. Int J Invert Reprod Dev. 1986b;9:155–168.
21. Sweet HC. Specification of first quartet micromeres in Ilyanassa involves inherited factors and position with respect to the inducing D macromere. Development. 1998;125:4033–4044. [PubMed]
22. Damen P, Dictus WJ. Organiser role of the stem cell of the mesoderm in prototroch patterning in Patella vulgata (Gastropoda, Mollusca). Mech Devel. 1996;56:41–60. [PubMed]
23. Crampton HE., Jr Experimental studies on gasteropod development. Arch Entw-mech. 1896;3:1–19, plates I–IV.
24. Clement AC. Development of Ilyanassa following removal of the D quadrant at successive cleavage stages. J Exp Zool. 1962;149:193–216.
25. Rabinowitz JS, Chan XY, Kingsley EP, Duan Y, Lambert JD. Nanos is required in somatic blast cell lineages in the posterior of a mollusk embryo. Curr Biol. 2008;18:331–336. [PubMed]
26. Conklin EG. Karyokinesis and cytokinesis in the maturation, fertilization and cleavage of Crepidula and other gasteropoda. J Acad Nat Sci Phil. 1902;12:1–122.
27. van Dam WI. Embryonic development of Bithynia tentaculata L. (Prosobranchia, Gastropoda). I. Cleavage. J Morph. 1986;188:289–302.
28. Collier J. Marine Invertebrates, Report on the Committee on Marine Invertebrates, Institute of Laboratory Animal Resources. Washington, DC: National Research Council, National Academy Press; 1981. Methods of obtaining and handling eggs and embryos of the marine mud snail Ilyanassa obsoleta. pp. 217–32.
29. Child CM. The early development of Arenicola and Sternaspis. Arch Entw-mech. 1900;9:587–723, plates XVIII–XXII.
30. Robert H. Embryologie des Troques. Arch de Zool Exper et Gen. 1902;3:18–538, plates XII–XLII.
31. Costello DP. Experimental studies of germinal localization in Nereis. I. The development of isolated blastomeres. J Exp Zool. 1945;100:19–46.
32. Cather JN. A time schedule of the meiotic and early mitotic stages of Ilyanassa. Caryologia. 1963;16:663–670.
33. Morgan TH. The formation of the antipolar lobe in Ilyanassa. J Exp Zool. 1933;64:433–467.
34. Conrad GW, Williams DC. Polar lobe formation and cytokinesis in fertilized eggs of Ilyanassa obsoleta. I. Ultrastructure and effects of cytochalasin B and colchicine. Dev Biol. 1974;36:363–78. [PubMed]
35. Goulding M. Cell contact-dependent positioning of the D cleavage plane restricts eye development in the Ilyanassa embryo. Development. 2003;130:1181–1191. [PubMed]
36. Craig MM, Morrill JB. Cellular arrangements and surface topography during early development in embryos of Ilyanassa obsoleta. Int J Invert Reprod Dev. 1986;9:209–228.
37. Wilson EB. The cell-lineage of Nereis. J Morph. 1892;6:361–480, plates XIII–XX.
38. Labordus V, van der Wal UP. The determination of the shell field cells during the first hour in the sixth cleavage cycle of eggs of Ilyanassa obsoleta. J Exp Zool. 1986;239:65–75.
39. Ponder WF, Colgan DJ, Healy JM, Nützel A, Simone LRL, et al. Caenogastropoda. In: Ponder WF, Lindberg DR, editors. Phylogeny and Evolution of the Mollusca. Berkeley and Los Angeles: University of California Press; 2008. pp. 331–384.
40. van den Biggelaar JAM, Haszprunar G. Cleavage patterns and mesentoblast formation in the gastropoda: an evolutionary perspective. Evolution. 1996;50:1520–1540.
41. Smirthwaite JJ, Rundle SD, Bininda-Emonds ORP, Spicer JI. An integrative approach identifies developmental sequence heterochronies in freshwater basommatophoran snails. Evol Dev. 2007;9:122–130. [PubMed]
42. Tomlinson SG. Intermediate stages in the embryonic development of the gastropod, Ilyanassa obsoleta: a scanning electron microscope study. Int J Invert Reprod Devel. 1987;12:253–280.
43. Page LR. Gastropod ontogenetic torsion: developmental remnants of an ancient evolutionary change in body plan. J Exp Zoolog B Mol Dev Evol. 2003;297:11–26. [PubMed]
44. van den Biggelaar JAM. Cleavage pattern in embryos of Haliotis tuberculata (Archaeogastropoda) and gastropod phylogeny. J Morphol. 1993;216:121–139.
45. van Dongen CAM, Geilenkirchen WLM. The development of Dentalium with special reference to the significance of the polar lobe. I, II, and III. Division chronology and development of the cell pattern in Dentalium dentale (Scaphopoda). Proc Kon Ned Akad Wet C. 1973;77:57–100.
46. Meisenheimer J. Entwicklungsgeschichte von Dreissensia polymorpha Pall. Z Wiss Zool Abt A. 1901;4:1–161, plates I–XIII.
47. Casteel DB. The cell-lineage and early larval development of Fiona marina, a nudibranch mollusk. Proc Acad Nat Sci Philad. 1904;56:325–401.
48. van den Biggelaar JAM. Development of dorso-ventral polarity and mesentoblast determination in Patella vulgata. J Morphol. 1977;154:157–186. [PubMed]
49. Heath H. The development of Ischnochiton. Zool Jahrb, Abt Anat. 1899;12:567–656.
50. van den Biggelaar JAM. Cleavage pattern and mesentoblast formation in Acanthochiton crinitus (Polyplacophora, Mollusca). Dev Biol. 1996;174:423–430. [PubMed]
51. Render J. Development of Ilyanassa obsoleta embryos after equal distribution of polar lobe material at first cleavage. Dev Biol. 1989;132:241–50. [PubMed]
52. Freeman G, Lundelius JW. Evolutionary implications of the mode of D quadrant specification in coelomates with spiral cleavage. J Evol Biol. 1992;5:205–247.
53. Holmes SJ. The early development of Planorbis. J Morph. 1900;16:369–450.
54. Conklin EG. The embryology of Fulgur: a study of the influence of yolk on development. Proc Acad Nat Sci Phila. 1907;59:320–359, plates XXIII–XXVIII.
55. Arnolds WJA, van den Biggelaar JAM, Verdonk NH. Spatial aspects of cell interactions involved in the determination of dorsoventral polarity in equal cleaving gastropods and regulative abilities of their embryos, as studied by micromere deletions in Lymnaea and Patella. Roux's Arch Dev Biol. 1983;192:75–85.
56. Dohle W. The ancestral cleavage pattern of the clitellates and its phylogenetic deviations. Hydrobiologia. 1999;402:267–283.
57. Watase S. Studies on cephalopods. I. Cleavage of the ovum. J Morph. 1891;4:247–302.

Articles from PLoS ONE are provided here courtesy of Public Library of Science


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...