Cleavage pattern and fate map of the mesentoblast, 4d, in the gastropod Crepidula: a hallmark of spiralian development
© Lyons et al.; licensee BioMed Central Ltd. 2012
Received: 22 May 2012
Accepted: 24 July 2012
Published: 19 September 2012
Animals with a spiral cleavage program, such as mollusks and annelids, make up the majority of the superphylum Lophotrochozoa. The great diversity of larval and adult body plans in this group emerges from this highly conserved developmental program. The 4d micromere is one of the most conserved aspects of spiralian development. Unlike the preceding pattern of spiral divisions, cleavages within the 4d teloblastic sublineages are bilateral, representing a critical transition towards constructing the bilaterian body plan. These cells give rise to the visceral mesoderm in virtually all spiralians examined and in many species they also contribute to the endodermal intestine. Hence, the 4d lineage is an ideal one for studying the evolution and diversification of the bipotential endomesodermal germ layer in protostomes at the level of individual cells. Little is known of how division patterns are controlled or how mesodermal and endodermal sublineages diverge in spiralians. Detailed modern fate maps for 4d exist in only a few species of clitellate annelids, specifically in glossiphoniid leeches and the sludge worm Tubifex. We investigated the 4d lineage in the gastropod Crepidula fornicata, an established model system for spiralian biology, and in a closely related direct-developing species, C. convexa.
High-resolution cell lineage tracing techniques were used to study the 4d lineage of C. fornicata and C. convexa. We present a new nomenclature to name the progeny of 4d, and report the fate map for the sublineages up through the birth of the first five pairs of teloblast daughter cells (when 28 cells are present in the 4d sublineage), and describe each clone’s behavior during gastrulation and later stages as these undergo differentiation. We identify the precise origin of the intestine, two cells of the larval kidney complex, the larval retractor muscles and the presumptive germ cells, among others. Other tissues that arise later in the 4d lineage include the adult heart, internal foot tissues, and additional muscle and mesenchymal cells derived from later-born progeny of the left and right teloblasts. To test whether other cells can compensate for the loss of these tissues (that is, undergo regulation), specific cells were ablated in C. fornicata.
Our results present the first fate map of the 4d micromere sublineages in a mollusk. The fate map reveals that endodermal and mesodermal fates segregate much later than previously thought. We observed little evidence of regulation between sublineages, consistent with a lineage-driven cell specification process. Our results provide a framework for comparisons with other spiralians and lay the groundwork for investigation of the molecular mechanisms of endomesoderm formation, germ line segregation and bilateral differentiation in Crepidula.
Endomesoderm is an ancient cell type that gives rise to a wide array of tissues in bilaterians and its origin and diversification are the subjects of great interest and debate [1–3]. One of the largest assemblages of bilaterians is the Spiralia, which includes mollusks, annelids, sipunculans, echiurans, myzostomids, nemerteans, platyhelminths, entoprocts and gnathostomulids . Despite the profound disparity in their adult body plans, most spiralians share a conserved early cleavage program that allows one to draw direct homology at a single cell level between taxa that have been diverging for hundreds of millions of years. Yet within this conserved framework, variation exists between species, which is documented in the literature going back to the late 19th century [5–10]. This provides unparalleled opportunities for asking how cells, tissue layers and organ systems have evolved within a stereotyped cleavage program . These attributes make spiralians a powerful system for studying the evolution of different tissues, including endomesoderm.
Here we report a comprehensive fate map of 4d in the slipper snail Crepidula fornicata and a closely related species, C. convexa. Crepidula is a model system for studying the spiralian cleavage program and subsequent development [22, 23]. It was in Crepidula, in fact, that the term mesentoblast was first applied to the 4d micromere [5, 23]. A detailed fate map has already been described for the larval contributions of the first through fourth quartet micromeres and the four fourth quartet macromeres . In that study, 4d was found to give rise to many cell types including the bilateral mesodermal bands, the intestinal hindgut, the left and right velar retractor muscles, and the adult kidney and larval kidney complexes . Henry et al.  presented preliminary descriptions related to the cleavage pattern of the 4d cell in C. fornicata up through the formation of the first four pairs of teloblast daughters (when 12 cells are present in the 4d sublineage). Here we extend and refine that analysis up through the formation of the 6mL/R cells and describe the specific fates of 1mL/R through 5mL/R. We followed the behavior of these 4d sub-clones through gastrulation and early organogenesis stages, and report the ultimate contributions of specific sublineages to preveliger and veliger tissues. We have found that the hindgut comes from the 1m, 3m, 4m and 5m cells, more than were previously thought [23, 24]. The 2m cells give rise to many different cell types including cells located beneath the apical organ, muscles surrounding the esophagus, retractor muscles, the absorptive and crystal cells of the larval kidney complex and the presumptive germ cells. Additional tissues, including heart and muscle cells, come from the 6M teloblasts (that is, their later-born daughter cells). We have found that the cleavage pattern and contributions of 4d through the birth of the 4m/4M cells are identical in the direct-developing congeneric species C. convexa. We carried out ablations of specific 4d daughter cells in C. fornicata and found little evidence of regulation for missing sublineages. These data provide the framework for understanding the mechanistic underpinnings of endomesoderm differentiation, and provide a point of comparison for similar studies in other spiralians.
Collection of adults and embryonic material
Adult C. fornicata were collected by the Marine Resources Department of the Marine Biological Laboratory (Woods Hole, MA, USA). Adult C. convexa were obtained at low tide from Waquoit Bay near the Waquoit Bay National Estuarine Research Reserve Headquarters (Falmouth, MA, USA). Adults were maintained in sea tables under running natural seawater. Embryos were obtained as previously described [15, 22, 25, 26]. Briefly, embryos were removed from their egg sacs and reared in 0.2-μm-filtered seawater (FSW) at room temperature (approximately 20°C) in gelatin-coated dishes. Penicillin (100 U/mL) and streptomycin sulfate (200 μg/mL) were added to prevent bacterial growth.
Fixation and staining
Some embryos and larvae were fixed in 3.7% formaldehyde in FSW for 30 minutes to one hour at room temperature and rinsed in 1X PBS (1.86 mM NaH2PO4, 8.41 mM Na2HPO4, 175 mN NaCl, pH 7.4) twice before storage at 4°C. To prevent bacterial growth, 0.01% NaN3 was added. Cases were stained with BODIPY FL phallacidin (Molecular Probes, Eugene, OR, USA), which selectively binds the filamentous actin present in differentiated muscle cells. Care was taken to not expose embryos to methanol or ethanol, which is not conducive to phallacidin staining. Specimens were washed three times in 1X PBS and 0.2% Triton X-100 in glass depression dishes. Phallacidin was diluted 1:500 in 1X PBS and 0.2% Triton X-100 by placing the required volume of phallacidin in the empty well of a depression dish, and allowing the methanol diluent to evaporate. The phallacidin was re-eluted in PBS and 0.2% Triton X-100 and specimens were added to this mixture to incubate in the dark at 4°C for 72 to 96 hours. The longer incubation time ensures complete staining of muscles, including those attached deep within the shell of older larvae. Specimens were rinsed three times in 1X PBS and 0.2% Triton X-100 and mounted in 80% glycerin diluted in 1X PBS. Hoechst (1:10,000 dilution) was added to the 80% glycerol in 1X PBS to visualize nuclei.
Live DiI- and dextran-labeled embryos were mounted in filtered seawater between Rain-X-coated (ITW Global Brands, Houston, TX, USA) glass slides and coverslips supported by clay feet. Embryos were imaged with an inverted LSM 700 confocal microscope (Carl Zeiss Inc., Munich, Germany). Confocal Z-stacks were processed to make maximum projections with ImageJ software (National Institutes of Health, Bethesda, MD, USA).
For example, with a distance of 250 μm, IR laser power is reduced to 49.7%, meaning that roughly half of the energy is lost when the light travels through this much water. Therefore, to obtain effective and consistent results, the height of the coverslip must be uniform and the depth of the media kept as shallow as possible. To achieve these conditions and to prevent the embryos from being crushed, the glass coverslips were supported by glass feet of uniform thickness (made from slivers of coverslips of the required thickness, see Figure 4A). These coverslips may be stacked in any configuration to obtain the desired height. The coverslip feet were cut using a diamond scribe and fixed to opposite edges of an intact coverslip using paraffin wax melted on a hot plate. Care must be taken to not use too much wax, which can increase the height of the feet. Any excess wax was removed using a razor blade and the coverslip was then polished clean using tissue paper. In this study we used number 1.5 glass coverslips for both the supporting feet and the top coverslip, which ensured a minimal water path for the laser light to reach the target.
As the XYClone objectives were originally intended for use on inverted microscopes, they are optically corrected for the thickness of plastic culture dishes (approximately 1 mm). To obtain an optimal high-resolution image using an upright microscope, one needs to add a secondary coverslip made from a glass slide (approximately 1m thick, which can also be cut using a diamond scribe). This coverslip can be placed on top of the coverslips described above or used in lieu of a thinner coverslip when affixing the glass feet (see above). The requirement for this secondary coverslip is much less critical when using the 20× XYClone objective, but is essential if one instead uses the 40× XYClone objective, which has a higher numerical aperture.
To visualize the progeny of 4d for laser ablation, the mesentoblast was initially injected with either DiI or rhodamine dextran, which is inherited by all of its progeny. Embryos were viewed via the fluorescence microscope equipped with the laser. Laser target acquisition was done using the computer’s camera display and XYClone software, or the laser system’s virtual ‘heads-up’ ‘RED-I’ spot visible directly through the microscope eyepieces. Live image acquisition for documentation was done using the Luminera camera controlled by the XYClone software (Hamilton-Thorne, Beverly, MA). Other images were captured using a Zeiss Axiocam (Carl Zeiss Inc., Munich, Germany).
Overview of C. fornicata development
Early development of C. fornicata has been described in detail [22–24]. The first two cleavages are essentially equal and parallel to the animal-vegetal axis, establishing four embryonic quadrants, the macromeres A, B, C and D. A small transient vegetal cytoplasmic protuberance called a polar lobe is formed during each of these cleavages and, by the four-cell stage, its contents and plasma membrane are shunted into one macromere, which will become the D or dorsal quadrant [15, 23]. Although inheritance of the polar lobe is predictive of which blastomere becomes the D quadrant, its influence on dorsal-ventral axis formation is not completely understood in Crepidula. In some species, including C. fornicata, polar lobe removal generally has little effect on development or second (dorsoventral) axis formation [15, 32]. By contrast, for many other species, the embryo becomes radialized when the polar lobe is removed [33–50]. At the third cleavage each of the four macromeres divides asymmetrically in a dexiotropic direction, giving rise to a quartet of smaller cells at the animal pole called micromeres 1a, 1b, 1c and 1d. The larger, vegetal daughter macromeres are renamed 1A, 1B, 1C and 1D. The macromeres undergo two additional rounds of asymmetric division producing the second and third quartet of micromeres, 2a to 2d and 3a to 3d, in a leotropic and dexiotropic direction, respectively. During this period, the first and second quartet micromeres themselves divide again, so by 18 h past fertilization (hpf) a 24-cell stage is reached consisting of 20 smaller, yolk-deficient micromeres forming an animal cap on top of the four large and yolk-rich 3A-3D macromeres.
Gastrulation proceeds over the course of approximately 3 days and during this process the progeny of 4d become fully internalized . Gastrulation begins with the epiboly of the animal cap micromeres to cover the macromeres. By 72 hpf, the embryo also begins to flatten in the animal-vegetal axis so that the equator is the largest circumference of the embryo. Gastrulation is complete by approximately 3.5 to 4 days post fertilization with the remnant of the blastopore prominent on the vegetal side. Organogenesis continues over several days. By approximately 4.5 to 5 days post fertilization, the embryo reaches the preveliger stage when many organ rudiments are distinguishable, including those of the intestine, larval kidney complex, foot, shell gland, velar lobes and stomodeum (Figure 5A-E). The alimentary canal consists of a ventral stomodeum, esophagus, stomach with associated paired midgut glands, style sac, intestine or hindgut, and terminal anus. The external absorptive cells of the larval kidney complex are transient paired lateral structures that protrude laterally from behind the velum and are generally lost by hatching. Each complex contains an external absorptive cell that extends from the surface of the larva, and internal crystal, pore and other supporting cells enclosing the flagellum of the deeper protonephridium [54, 55]. The crystal cells harbor green autofluorescent material (Figure 5A-B,G,I,K). The distinctive young veliger stage is reached by 7 to 8 days post fertilization with a prominent shell, large velar lobes, right and left ocelli, foot and operculum (Figure 5F-L). Veligers hatch from the egg case at around 4 weeks and the juvenile stage follows metamorphosis at approximately 6 to 8 weeks [22, 23, 56, 57].
4d lineage nomenclature
We employ a new nomenclature for naming 4d progeny in C. fornicata (Figure 1), originally introduced by Henry et al. , as a modification of Conklin’s system . The 4d micromere divides bilaterally to give rise to the paired teloblasts ML and MR located to either side of the midline, where L and R refer to left and right, respectively (Figure 1C). With each successive division, these teloblasts generate subsequent daughter cells, signified by the lower case letter m (for example, 1mL and 1mR; 2mL and 2mR). As each daughter cell is generated, the teloblast is correspondingly numbered in sequence, (for example, 1ML and 1MR; 2ML and 2MR) so that the birth order, age and relationship of these cells can be followed. Superscript numbers (for example, 2mL1, 2mL2, 2mL1.1, 2mL1.2) are used in accordance with the standard spiralian nomenclature system  for the sequential progeny of the teloblast daughters, as these begin to divide in turn. Daughter cells born closer towards the animal pole or towards the midline are given the superscript 1 and those born closer towards the vegetal pole or more laterally are given the superscript 2.
Early cleavage pattern of the 4d lineage and behavior of the clone during gastrulation
To follow division patterns in detail, fluorescent lineage tracer was pressure-injected into 4d soon after its birth and embryos were raised to various stages of development and observed live with confocal microscopy (Figure 2). The lineage tracer tends to accumulate in nuclei making it easier to distinguish individual cells, even when they are tightly packed. The slow cell cycles of the teloblast and its daughter cells (Figure 6) made it convenient to study their behavior and provided ample time to inject individual daughter cells within each clone with a second lineage tracer to uniquely label particular sublineages. This allowed one to decipher the order and orientation of the subsequent divisions (examples shown in Figure 3).
In total, 4d gives rise to a tremendous array of differentiated cell fates, which are illustrated in Figures 5 and 6. These include the hindgut (intestine); the left and right larval retractor muscles; the external absorptive cells and their underlying autofluorescent crystal cells (components of the larval kidney complex); a pair of cells that underlie the apical organ (of unknown function, labeled ‘sac’ in Figures 5 and 7G,I); muscles of the heart, esophagus and foot; widely distributed mesenchyme; and a pair of putative germ cells.
Contributions of specific 4d progeny and ablation results in C. fornicata
To understand the specific contribution of different 4d daughter cells to larval tissues, 4d was injected with rhodamine green dextran and allowed to develop to various stages when one daughter cell was then injected again with DiI (Figure 3). These tandem injected embryos were raised to preveliger and veliger stages to determine the contributions of each cell (Figure 7). For each 4d daughter cell injection, at least 10 cases were examined and similar results were seen for each cell type. In this study, we uniquely labeled the first five teloblast progeny (1mL/R through 5mL/R), and in some instances, their daughter cells (Figure 6). The teloblasts continue to form many additional progeny (6mL/R, 7mL/R, and so on), but the exact number and fates of these cells were not followed in this study. Though generally formed in a synchronous fashion, there may be some slight asynchrony in terms of the relative birth order of the correspondingly numbered left and right daughter cells.
Fates of 1mL/R
The first pair of daughter cells formed by the paired ML and MR teloblasts (Figures 1C and 2C) are the 1mL and 1mR cells, respectively (Figures 1D, 2D and 6). The cells are rather large, approximately 50 μm in diameter, and are given off towards the vegetal pole. In general, these cells are formed synchronously, but in some cases the 1mL cell appears slightly larger than that of the 1mR cell. This may be due to the relative packing arrangement and location of these cells. Their sister cells are the teloblasts, and like the initial teloblasts (ML and MR), 1ML and 1MR contact one another along the dorsal midline (Figures 1D and 2D). Injections of fluorescent lineage tracers reveal that both 1mL and 1mR ultimately give rise to proximal and deeper regions of the hindgut endoderm (Figures 2E,F and 7B,C,E,F,K) as has been described previously for Crepidula and other mollusks . At the time of their formation, the 1mL and 1mR cells are characterized by the presence of a highly granular, yolky cytoplasm with a vacuolated appearance, previously described by Conklin . This is because there is an animal-vegetal (A-V) gradient of yolk granules within the cytoplasm of 4d and subsequently the two ML and MR teloblasts, with a higher concentration of yolk granules located towards the vegetal pole. At later stages, 1mL/R and their subsequent daughters become elongated along the ventral midline. As the intestinal rudiment begins to form, the proximal region extends to the right side (Figure 7E,F). The clone of cells derived from 1mL lies to the left of and slightly deeper to that derived from 1mR and does not extend all the way to the tip of the extending proximal intestinal rudiment (Figure 7E,F). By contrast, the progeny of the 1mR cell form an elongated spike that extends towards the right side of the embryo during the fifth day of development (see Figure 7E). These cells, along with the progeny of 1mL and additional teloblast daughter cells (see below), are later drawn up the right side of the embryo where they dive deeper to join the visceral endoderm derived from the macromeres , as rotation of the visceral mass take place. Throughout this process, the endodermal progeny of 1mL and 1mR remain as a fairly coherent mass of cells.
Ablation of 1mL and/or 1mR
Removal of either 1mL or 1mR alone does not have a tremendous impact on the appearance of the hindgut, however removal of both cells (Figure 4C-F) leads to certain defects (Figure 8A-C). If the 1mL and 1mR cells are removed, the remaining 3mL and 3mR cells take a position closer to the vegetal blastopore, occupying the former location of the 1mL and 1mR cells during gastrulation. This leaves an apparent void where the 3mL and 3mR cells would normally be situated. The rightward pointed proximal end of the hindgut does not form, and the deeper fluorescing contributions of 1mL and 1mR to the visceral mass appear to be reduced or absent, but the progeny of the remaining 3mL and 3mR cells (along with the 4mL/R and 5mL/R cells) do organize an elongated hindgut tube with an external opening at the anus (Figure 8A-C).
Fates of 2mL/R
The second pair of daughter cells formed by the paired 1ML and 1MR teloblasts are the 2mL and 2mR cells, respectively (Figures 1E and 2E). Typically these two cells form synchronously though one may form before the other. As noted by Conklin , they do not appear as vacuolated as do the 1mL and 1mR cells, described above (Figure 2E). The 2mL and 2mR cells are smaller than the 1mL and 1mR cells, approximately 30 μm in diameter and are formed bilaterally, towards the animal pole (Figures 1E, 2E and 3B,C). As these cells are formed they lie deep beneath the advancing cap of ectoderm derived from the animal micromeres. To inject the 2mL and 2mR cells one must pierce through this overlying ectodermal sheet. These two cells are already present by the time the other fourth quartet micromeres (4a to 4c) are born. At this stage the 1mL and 1mR cells and the teloblasts (1ML and 1MR) still lie in contact with one another along the dorsal midline. The 2mL/R cells and their subsequent progeny make up the tips of the elongating mesodermal bands (Figure 3G,H,J,K,Q,T).
The 2mL/R cells give rise to a large number of progeny (Figures 7G-J and 6). These include progenitors of the left and right bands of the main larval retractor (Figure 7H,J) and specific components of the larval kidney complex that include the external absorptive cells (Figure 7G,I) and the autofluorescent crystal cells (that eventually lie just deep to the absorptive cells; Figure 7G,I). In addition, these give rise to a pair of cells (of unknown function) that become situated beneath the apical organ (labeled ‘sac’ in Figure 7G,I), cells adjacent to the esophagus, and a large quantity of internal mesenchyme. With rotation of the visceral mass the external absorptive cell and underlying crystal cell on the right side become elongated and located at a higher (more anterior) position compared with those on the left (compare Figures 5I and 7H,S,DD).
Daughters of 2mL/R (2mL1/2mR1)
As the 2mL/R cells divide, they give rise to 2mL1/L2 and 2mR1/R2, respectively (Figures 1G, 2G, 3G,H,J-L,Q,Tand 6). The 2mL1/R1 daughters are somewhat larger (approximately 25 μm in diameter) and lie more animal and medial relative to their corresponding 2mL2/R2 sister cells (which are approximately 20 μm in diameter, Figures 1G, 2G, 3G,H,K,L). Subsequently, these cells undergo a series of divisions. The 2mL2/R2 cells divide first, and the 2mL1/R1 cells divide several hours later, after the formation of 5mL/R (Figures 1I-J, 2I-L and 6). The more animal 2mL1.1/R1.1 cells are generally larger than the vegetal 2mL1.2/R1.2 cells (Figures 1I, 2L,M and 3T). The differential contributions of these latter four cells have not been distinguished. After the birth of 4mL/R the progeny of 2mL/R (2mL1, 2mL2, and 2mR1, 2mR2) begin to separate from the other progeny of the teloblasts (Figures 2L-R and 3S,T). It is not clear if this separation is to due active migration or the result of being passively moved by surrounding tissues. Later, as the clones of cells derived from the 2mL/R cells begin to separate away from the other progeny of 4d (by the time that 5mL/R and/or 6mL/R are formed, see below), they are brought to the lateral sides of the embryo and move towards the future anterior pole (Figures 2M and 3Q,T). Initially, these spatially isolated clones contain three and then four cells (2mL1.1/R1.1, 2mL1.2/R1.2, 2mL2.1/R2.1 and 2mL2.2/R2.2; see Figures 1I-J, 2I and 3T).
The 2mL1/2mR1 cells give rise to numerous progeny (Figure 7L-P). These cells contribute to the formation of the retractor muscles. The retractor muscle precursor cells move dorsally and join other retractor muscle precursors derived from the 5ML/R teloblasts to give rise to the left and right bands of the retractor muscles. In the process, some of these cells may cross the midline to lie on opposite sides. The 2mL1/2mR1 cells also give rise to the autofluorescent crystal cells of the larval kidney complex that are lost at the time of hatching. In addition, the 2mL1/2mR1 cells also give rise to a pair of cells that become situated beneath the apical organ. In 2mL1/2mR1-injected embryos, cells (including muscles) that surround the esophagus and a large quantity of internal mesenchyme are also labeled.
Daughters of 2mL/R (2mL2/2mR2)
In contrast to the 2mL1/R1 cells, the 2mL2/R2 cells only undergo a single division, giving rise to 2mL2.1/R2.1 and 2mL2.2/R2.2 (Figures 1I and 21). After their birth, these two daughters were never observed to divide again (Figure 6). The 2mL2.1/R2.1 cells are relatively large cells (Figures 2I and 3T) that give rise to the external absorptive cell of the larval kidney complex (Figures 6 and 7Q-U). These protruding, rounded cells are directly exposed to the capsular environment and are not covered by ectodermal tissue. Formerly referred to as the external larval kidneys, the exact function of these cells is unclear [54, 55]. Like the deeper crystal cells derived from 2mL1/R1, the absorptive cells are lost at the time of hatching.
The more vegetal, 2mL2.2/R2.2 cells are small (Figures 2I and 3T) and, as mentioned, were never observed to divide again during the larval stage. This is supported by the fact that the fluorescence in these cells remains very bright, presumably from the lack of dilution through subsequent cell division. The behavior and location of 2mL2.2/R2.2 cells suggests that they represent the primordial germ cells, which we refer to here as presumptive germ cells (Figure 6, see Discussion). During gastrulation, the 2mL2.2/R2.2 cells become displaced towards the cells of the intestinal rudiment along the ventral midline below the rudiment of the developing foot (Figures 2M,O-R, 5C-G and 7Q-V). It is unclear if they undergo directed active migrations to reach this location. As they migrate they form numerous long filopodia (inset in Figure 7R) and appear to move in an amoeboid fashion towards the cells of the intestinal rudiment (Figure 2O-R). As differential growth and rotation of the visceral mass occurs (separate from the later process of torsion), 2mL2.2/R2.2 are brought up with the intestinal primordium along the right side of the larva (Figure 7S-U). Eventually these two cells come to lie in contact with one another, next to the intestine and deep near the heart on the right side, adjacent to one of the gut glands (Figures 5F,G and 7V).
Ablation of 2mL, 2mR and their daughters 2mL1/R1 and 2mL2/R2
Additional file 1: Minus 2mL and 2mR Heart. This movie depicts a typical veliger larva that forms after removal of both the 2mL and 2mR derivatives of 4d. The larva is depicted as a right lateral view with the shell that surrounds the opaque visceral mass located to the left and the more transparent head and velum to the right. Note the presence of a beating heart towards the top of the larva, which is located between the visceral mass and the head. (MOV 1 MB)
Fates of 3mL/R
The third pair of daughter cells formed by the paired 2ML and 2MR teloblasts are the 3mL and 3mR cells, respectively (Figures 1F and 2F). The cells are rather large, about 40 μm in diameter, but not quite as large as the 1mL and 1mR daughters. The appearance of 3mL and 3mR is similar to that of 1mL and 1mR, characterized by the presence of a granular, yolky cytoplasm, but not quite as yolky or vacuolated in appearance. 3mL and 3mR are formed from the vegetal and medial sides of the 1ML/R teloblasts and lie in a more animal and superficial location relative to the 1mL/R cells, and in intimate contact with them (Figure 2F). In some cases these cells may lie somewhat more lateral to the 1mL/R cells, and separated from one another (Figure 2G,I). At this stage the teloblasts are no longer in contact with one another as they are separated by the 3mL/R and 1mL/R endodermal precursors. Typically the 3mL/R cells form synchronously, though one can observe one forming before the other. As these cells are formed they lie deep to the advancing cap of ectoderm derived from the animal micromeres that extends via epiboly. Thus, to inject these cells one needs to pierce through the overlying ectodermal sheet. As epiboly takes place the 3mL and 3mR cells may ride up to a slightly more superficial location over the 1mL and 1mR cells.
Like the 1mL/R cells, these cells also do not divide for a few days, around the same time that the 1mL and 1mR cells begin to divide (Figure 6). Injections of fluorescent lineage tracers reveal that both the 3mL and 3mR cells give rise to distal regions of the hindgut endoderm (intestine), as has been described previously [22, 23]. Their progeny subsequently generate elongated columns of cells that make up the tube of the hindgut (Figure 7D,W-Y). These progeny extend along most of the length of the intestine. This differs from the behavior of the 1mL/R progeny, which contribute mainly to deeper cells of the digestive tract at the more proximal end of the hindgut tube (Figure 7E,F) and remain as a tighter mass of cells towards the tip of the intestinal rudiment until later in development. The 3mL and 3mR clones appear to undergo the greatest elongation early during gastrulation. These cells extend the furthest to the animal dorsal (posterior) side of the hindgut intestinal rudiment. As the intestine elongates, these progeny become stretched out as discontinuous columns as other progeny (derived from subsequent progeny of the teloblasts, see below) presumably intercalate between these cells. Elongation may be driven by convergent extension and/or directed proliferation. These cells contribute much of the distal length of the hindgut tube that leads all the way to the opening of the anus. The anus begins its formation in a posterior location along the ventral midline (data not shown). As the intestinal tube elongates, the terminal opening (anus) is located in a posterior position below the level of the absorptive cell and to the side of the foot on the right side of the early veliger larva. Eventually the anus will be brought up to a more anterior location above the head when further rotation of the visceral mass and torsion takes place.
Ablation of 3mL and/or 3mR
Removal of either cell alone does not have a tremendous influence on the appearance of the hindgut, however removal of both cells leads to the absence of an elongated tube and disrupts formation of the anus (Figure 8D-F). The clones of cells derived from the 1mL and 1mR (and other endodermal progenitors, see below) spread out somewhat and do not organize into a coherent mass. Scattered labeled cells can be seen in the embryo. The overall morphology of the larvae appears somewhat abnormal and development is stunted.
Fates of 4mL/R
The fourth pair of daughters derived from the 3ML and 3MR teloblasts is the 4mL and 4mR cells, respectively (Figures 1H, 2H-M,P, 3M-P,R, and 6). These are very small cells, 10 μm in diameter, formed adjacent to the progeny of the 2mL and 2mR cells (Figures 1H and 2H). They are much smaller than the 2mL and 2mR cells, being approximately the same size or smaller than 2mL2 and 2mR2. 4mL/R make contributions to the hindgut intestinal tube (Figure 7Z-DD). These cells are typically formed laterally and towards the animal pole off of the 3M teloblasts (Figures 2H,I and 3M-O) and come to lie just vegetal to the progeny of the 2mL/R cells. They may be somewhat shifted laterally or medially relative to the progeny of 2mL/R. Typically these cells form synchronously, though one can observe one forming before the other. The 4mL/R cells typically are born just before the 2mL2/R2 cells undergo their division to form 2mL2.1/R2.1 and 2mL2.2/R2.2 (Figures 1H,I, 2H,I and 6). As these cells are formed they lie deep to the advancing cap of ectoderm derived from the animal micromeres that extends via epiboly. As development progresses, these cells move vegetally (ventrally) relative to the teloblasts; as they join the mass of hindgut endoderm they subsequently pass towards the vegetal tip of the intestinal rudiment derived from the 3mL/R and 1mL/R cells (Figures 2I-M,P and 7Z-CC). After the 4mL/R cells come to lie to the sides of the intestinal mass and have migrated towards the vegetal tip of the hindgut rudiment (adjacent to the blastopore), they begin to divide. By the time the embryo has reached the preveliger stage, they have each formed two progeny (4mL1/R1 and 4mL2/R2, Figure 6). Eventually, these cells form an elongated chain of cells along most of the length of the hindgut intestinal tube (Figure 7Z,AA). Progeny of these cells extend all the way to the anus. This behavior is somewhat similar to that seen for the progeny of 3mL/R, as described above. Note that earlier, Henry et al.  reported incorrectly that 4mL/R contributed to muscles.
Ablation of 4mL and 4mR
Removal of both 4mL and 4mR leads to abnormal development of the intestine, and the larvae may look somewhat stunted (slowed in development, data not shown). The mass of intestinal cells appears to be smaller in these larvae and the hindgut does not develop normally with the intestinal tube failing to elongate.
Fates of 5mL/R
The fifth pair of daughters derived from the 4ML and 4MR teloblasts is the 5mL and 5mR cells (Figure 1I). Like the 4mL and 4mR cells, these are small cells, approximately 10 μm in diameter, formed adjacent to the 4mL and 4mR cells (Figures 2J and 3S). In some cases 5mL/R may appear larger than 4mL/R. As these cells form they may lie proximal to the 4mL/R cells or may lie medial to these cells, situated somewhat closer to the vegetal pole (Figure 2J,K). They may be seen to form as part of a coherent band of cells extending from the teloblasts, but their formation may be somewhat variable in location. 5mL/R also contributes to the hindgut intestine (Figure 7EE-HH). These cells end up lying at the posterior base of the intestinal rudiment, to the left and right sides respectively (opposite of the vegetal tip; Figure 7EE,GG). Eventually these cells begin to divide and form a small chain of cells on opposite sides of the terminal hindgut tube (Figure 7FF,HH). Their behavior is somewhat similar to that seen in the progeny of 3mL/R and 4mL/R, as described above (compare with Figure 7W-AA). These cells extend to the opening of the anus as a discrete chain, but do not extend as far as those derived from 3mL/R and 4mL/R. The 5mL and 5mR cells were not removed in this study.
Later contributions of the teloblasts, 5ML/R
The bilateral teloblasts (for example, 5ML and 5MR) continue to divide and produce a large number of additional progeny (Figures 1I,J, 2J-R, 6 and 7II-NN). The teloblasts continue to generate numerous small daughter cells such as 6mL/R, which form loosely organized chains or clusters (Figures 1J and 2L,M). The 6mL and 6mR cells form next to the 5mL/R and 4mL/R cells. We have not traced the fates of these cells individually in this study. Fates of the later-born teloblast progeny include additional retractor muscles that join those from 2mL1/R1 (Figure 7II-NN). These fates must also include additional mesenchyme, the heart, the adult kidney and muscles of the foot, as well as other cells that may contribute to the larval kidney complex (for example, protonephridia). Removal of the teloblasts (for example, 4ML/R and 5ML/R) leads to a set of defects including abnormal formation of the retractor muscles, which are not exclusively derived from the 2mL1 and 2mR1 cells, and loss of a beating heart (data not shown).
Cleavage and fates of 4d in C. convexa
The 4d lineage in Crepidula
Examples of varied germ layer contributions from the 4d lineage
Primordial germ cells
A new nomenclature for 4d
To keep better track of the progeny of 4d, we developed a new nomenclature for the 4d lineage in Crepidula, modified from Conklin’s original system (Figures 1 and 6; [22, 23]). Conklin referred to the 4d cell as ‘ME’ for ‘MesEntoblast’, and distinguished the left and right progeny with superscripts, ME1 and ME2, respectively. For subsequent divisions, his original nomenclature referred to mesodermal (M/m) and endodermal (E/e) progenitors with those combinations of letters, respectively. Taking the left, ME1 lineage (now ML in our nomenclature, Figure 1) as an example, in Conklin’s system its daughter cells were E1 (now 1mL), the primary enteroblast, and Me1 (now 1ML), the teloblastic progenitor of mesoderm and endoderm. When Me1 divides, it gives rise to m1 (now 2mL), the primary mesoblast, and M1e1 (now 2ML), the mother cell of the mesodermal teloblast and additional endoderm. With the division of M1e1, Conklin assumed endoderm and mesoderm were fully segregated, and thus the daughter cells were named e1 (now 3mL) for the secondary enteroblast and M1 (now 3ML) for the mesoblastic teloblast. Hence, Conklin’s nomenclature emphasized the endodermal contributions of 4d, which had been underappreciated or completely overlooked by previous studies, and the emergence of the mesodermal teloblasts (, see also ).
However, for modern lineage studies, Conklin’s system is cumbersome and impractical for extending the lineage. For example, when m1 divides, Conklin names the animal daughter m1.1 and the vegetal daughter m1.2, in line with the standard nomenclature to distinguish animal versus vegetal daughters in the early embryo . His use of numbers to denote left versus right and animal versus vegetal daughter cells is confusing. Instead, we use the postfix L or R to distinguish left and right (for example, ML/MR), and save superscripts for distinguishing the animal and vegetal daughter cells of divisions in non-teloblast lineages (for example, 1mL1, 1mL1.1; Figure 1).
Another weakness of Conklin’s system is that there is no useful convention for following sequential progeny of the teloblasts or for distinguishing the teloblasts from their daughter cells. Similar weaknesses are present in the system used by Woods  where, in fact, two entirely different progeny may carry the same designations. Goulding [20, 80] used a modified naming system based on Conklin’s nomenclature that applied the prefix 4d and superscripts for left/right 4dR/4dL (equal to ML and MR), which were then followed by numbers to denote the further divisions, with 1 being applied to the more animal daughter and 2 applied to the more vegetal daughter. A specific nomenclature exists for 4d sublineages in clitellate annelids, which develop via a modified spiral cleavage program that gives rise to exceptionally large ectodermal and mesodermal stem cells called teloblasts [9, 17, 62, 81, 82]. For example, in the leech Helobdella the homolog of 4d, DM" divides bilaterally giving rise to two mesodermal stem cells, the ML and MR teloblasts. Each M teloblast undergoes iterated and highly asymmetric divisions to give rise to a coherent column (bandlet) of primary ‘m’ blast cell cells that each serve as a founder for repeating elements of the segmental mesoderm. The leech has a fixed number of 32 segments, but the M teloblasts produce additional blast cells that also contribute to anterior (early mesoderm or em cells) and posterior (supernumerary) nonsegmental tissue. For example, the first six cells born from the M teloblasts give rise exclusively to nonsegmental (em1 to em4), or a mix of nonsegmental and segmental (em5 and em6), tissue. The seventh cell born from the M teloblast (equivalent to 7mL/R in our nomenclature) is the first bona fide segmental founder cell and is called sm1, and subsequent founder cells are numbered in succession . In our system, the teloblast is labeled with a capital letter, and the non-teloblast daughter cell with a lower case letter (for example, 1ML versus 1mL). Additionally, a prefix number is used to keep track of birth order and, with each division, the teloblast is renumbered correspondingly (Figure 1).
Lastly, Conklin’s use of E/e and M/m for endodermal and mesoderm made assumptions about the ultimate fate of these blastomeres, and is limiting. The data presented here clearly show that the separation between endoderm and mesoderm occurs much later in Crepidula than Conklin had inferred (Figure 6). In fact, in some species, 4d may not contribute to endoderm formation, such as in Capitella (see Table 1). Thus, our naming system builds in no assumptions about ultimate cell fate, and should allow for easier comparisons between different species.
Contributions to the intestine
In Crepidula, the hindgut or intestine is a tube of cells of endodermal origin that contacts the style sac at its proximal (anterior) end and the anus at its distal (posterior) end. Multiple cells born from the teloblasts contribute to the hindgut. The 1mL/R cells give rise to deeper cells of the digestive tract at the proximal end of the hindgut. 3mL/R give rise to the hindgut endoderm, and contribute to most of the distal end of the hindgut tube that leads to the opening of the anus. 4mL/R and 5mL/R also make contributions to the hindgut intestine. There are clearly much more extensive contributions to the formation of the hindgut than previously recognized by Conklin , and the separation of endoderm and mesoderm occurs at a later stage of development.
The hindgut is a relatively shorter structure in annelids than in molluscs . In annelids, the hindgut is conventionally considered a proctodeal inturning of ectodermal origin situated posterior to the endodermally-derived midgut, and includes the rectum and anus [84, 85]. However, in the polychaete Capitella, Meyer et al.  argue that 4d gives rise to a few cells on the surface of the presumptive anus, which they interpret to be ectodermal (Table 1). In this species, 4d does not give rise to endoderm at all. Conklin reported that cells of the anus or proctodeum are derived from the second and third quartet micromeres in Crepidula. He describes that these large cells remain on the surface of the posterior end of the embryo at least through the preveliger stage, that the anus forms much later in development and that the proctodeum is very short. We observed that the 3mL/R through 5mL/R lineages contribute to the most distal end of the intestine leading up to the anus, and these cells are certainly endodermal in origin and clearly not located on the surface of the embryo.
Contributions to the larval kidney complex
4d contributes to cells of the larval kidney complex, including the superficial external absorptive cell and the underlying crystal cell. The larval kidney complex contains additional deeper cells that include the protonephridia, the pore cell and various support cells (this complex has been well described by  and ). Notably, the absorptive cells are only present in caenogastropods and are lost at hatching. We have not yet detected whether 4d contributes to the formation of the protonephridia or the adult kidney. One may question whether the absorptive cells, which are exposed on the surface of the larva, represent mesodermal versus ectodermal cells (Table 1). If it is the latter, the 4d cell should be considered an ‘ectomesentoblast’ cell in Crepidula.
Contributions to the heart
The origin of heart tissue in mollusks is controversial, and has been ascribed to the 2c and/or 4d micromeres in different studies [21, 23, 24, 59, 77, 86–88]. Conklin reported that the second and third quartet micromeres contribute to the larval heart in Crepidula. Hejnol et al.  found that 2c gave rise to both the ectodermal covering and the muscles of the larval heart. Those analyses did not report any contribution of 4d to the heart. In this study, we observe fluorescence in the heart following injection of 4d, which is particularly noticeable in the nuclei of the elongated muscle cells. As the heart beats, it protrudes from the larva between the head and the dorsal edge of the mantle and is covered by a thin layer of ectoderm that appears to be derived from 2c. The heart muscles are very thin elongated cells, and it is difficult to detect label in these cells, which is one reason they might have been missed previously. Furthermore, much of the analysis of Hejnol et al.  was carried out on fixed specimens and some fluorescence may have been lost.
In Ilyanassa, Render  reported that the heart was labeled in 2c micromere injections, while removal of the 2c micromere leads to the absence of a beating heart . Yet it has also been reported that removal of the 4d micromere leads to the absence of a beating heart in Ilyanassa[59, 76, 90]. More recently, Rabinowitz et al.  also observed that the heart was absent when 4d was ablated in Ilyanassa. 4d ablations in Ilyanassa are particularly informative because, mostly, only 4d-derived tissue is affected, as most (but not all) of the organizing activity of the D quadrant resides in the 3d macromere [21, 91]. By contrast, the 4d micromere of Crepidula serves as the principle embryonic organizer and when it is ablated the embryo becomes radialized and does not form a heart [15, 92]. Consistent with the sublineage fates reported here, we found that sequential removal of paired 1mL/R through 4mL/R daughters does not affect the development of a beating heart, but removal of 4ML/R and 5mL/R do. Additionally, injections of 5mL only label the intestine; therefore, later-born progeny of 4d (derived from 5ML/R) must contribute to this structure.
Contributions to the germline
It is generally accepted that the germline arises from the 4d lineage in spiralians [13, 21, 51–53, 65, 66, 72, 73, 93–96], but exactly when it arises and from what sublineage are not well understood for most species. We argue that the 2mL2.2/R2.2 cells are the progenitors of the germline in Crepidula. This claim is based on several lines of evidence, including comparisons with other model systems, their behavior during gastrulation and their position in the larva. There are claims that the germ cells arise from the pericardial mesoderm such as in the gastropod Viviparus. Their close proximity might suggest such an association. However, in Crepidula we now know that the heart tissue must arise from a different set of precursor cells than those that give rise to the germline (see above).
Like many germ cell precursors, the 2mL2.2/R2.2 cells are mitotically quiescent; we never observed either of these two cells dividing up through the veliger stage. The 2mL2.2/R2.2 cells break away from the mesodermal bands during gastrulation and appear to make amoeboid-like movements while forming dynamic filopodia ( Figures 2Q,R, 7R and 9V). Their behavior is remarkably similar to the situation encountered for the germ cells of other organisms, including mammals and Drosophila. Furthermore, the 2mL2.2/R2.2 cells take a position in close association, but not necessarily in direct contact with the intestinal rudiment and end up close to the heart. In some other systems, the germ cells migrate along endodermal tissues to reach the gonads . It is possible that the 2mL2.2/R2.2 cells are following unidentified cues produced by the endodermal cells of the intestine in Crepidula.
Another common characteristic of primordial germ cells is a specialized cytoplasmic germ plasm or nuage . In Crepidula, there is a distinctive cytoplasm that is associated with the polar lobe that some authors describe as being similar to germ plasm [100–103]. While the polar lobe is known to fuse with the D macromere, it is not known whether the 4d cell inherits the aforementioned specialized cytoplasm. At the eight-cell and later stages the remnants of the polar lobe contents can be detected as a portion of the vegetal cytoplasm - situated adjacent to the vegetal cross-furrow in the D quadrant macromere - that has a different refractive property from the rest of the cell . This cytological feature is not observable directly in the 4d cell. In the course of our present studies we have not observed any unique properties of the cytoplasm of the 2mL2.2/R2.2 cells that would suggest that they inherit a specialized cytoplasm. Weak and transient vasa mRNA staining is seen in 4d in Crepidula, but it does not persist into later stages . The localization of Pl10 mRNA has also been studied and no enrichment in the D quadrant or 4d was seen . Other genes in the germline gene regulatory network such as nanos and piwi have not been examined.
Interestingly, the tips of the bands, including the precursors of the 2mL2.2/R2.2 cells, preserve beta-catenin protein expression for a long time, suggesting the first evidence of differential molecular signatures among the 4d lineage in Crepidula. In the sea urchin embryo, beta-catenin expression persists in the small micromeres as it clears from the rest of the vegetal plate between hatching and archenteron formation . Since the sea urchin small micromeres are known to contribute to the germline , this similarity might suggest a conserved role for beta-catenin signaling in germline cell fate.
The origin of the germline within the 4d clone has been examined in a few other spiralians. In Ilyanassa, Swartz et al.  reported that IoVasa mRNA is initially ubiquitous and then becomes restricted to the 4d micromeres and ultimately restricted to the 2mL/R progeny and the 3ML/R teloblasts (Figure 6; cells 4dL121 and 4dR121 according to their nomenclature), after which staining disappears. Rabinowitz et al.  also examined IoNanos in situ as well as with antibody staining and found that both the mRNA and the protein were initially expressed ubiquitously, then became restricted to the 4d cell. IoNanos protein is mostly associated with the teloblasts, but interestingly it persists longest in the 4dL112 and 4dR112 cells, equivalent to 2mL2/R2 cells in our nomenclature (Figure 6). This is suggestive evidence that these cells may represent the germ cell precursors in caenogastropod snails. However, Swartz et al.  suggested that it is the teloblasts that may represent the progenitors of the germ cells. This assumption was based in part on the earlier studies of Woods [52, 53]. In the freshwater bivalve clam Sphaerium, Woods [52, 53] did a careful study of the emergence of the germline from 4d (which he referred to as d.4 or M). Sphaerium possesses a mitochondria-rich and histologically distinct cytoplasm that is present in the egg that segregates to the 4d cell and its two bilaterally symmetrical daughters M and M.1 (equal to ML and MR, respectively). After these teloblasts each form three rounds of mesodermal progenitors, the two remaining teloblasts (called G and G.1, equal to 3ML and 3MR, respectively) each divide equally giving rise to two cells, and these in turn enter a long period of quiescence. Woods [52, 53] indicates that these latter four cells are the primordial germ cells, which were followed through late stage development (by virtue of their inheritance of the distinctive population of mitochondria) to eventually appear to join the gonad . Thus, in Sphaerium, the germline comes from the equivalent of the 3 M teloblasts.
In Helobdella and Tubifex, a combination of lineage tracing, ablation and in situs for nanos and vasa have shown that the germline arises from M teloblast-derived segmental mesoderm (sm) blast cell clones contributing to midbody segments [93, 95]. Thus, these germ cells are born much later in the 4d lineage relative to those in Crepidula, Ilyanassa or Sphaerium. The germline is known to arise from the 4d lineage in the polychaetes Capitella[13, 65, 66] and Platynereis, but for the former it is not known when the germline precursors are specifically born. Recently, Rebscher et al. showed that, in Platynereis, the germline comes from the first two daughters born from the teloblasts (equivalent to 1mL/R and 2mL/R in Crepidula by birth rank), which come to reside in the posterior growth zone of the larva.
By all accounts, it appears that the germline arises from 4d in spiralians (Table 1), but the exact sublineage contributions appear to differ between species. In each of these cases, the germline appears to be set aside relatively early during development, and one could argue that it is specified autonomously. Except for Crepidula, these embryos exhibit early unequal cleavage divisions and the D quadrant is specified at the four-cell stage. In Crepidula, however, the early cleavage divisions appear equal and the D quadrant is specified later in development by virtue of inductive interactions from the animal micromeres. There is, however, a bias in terms of which quadrant normally becomes the D quadrant, as only one blastomere (the presumptive D cell) inherits the small polar lobe during the second cleavage division. Whether or not these small polar lobes play a role in germline specification remains to be determined.
Render  reported some asymmetries for the 4d lineage in Ilyanassa in terms of the progeny of the ML (which she referred to as ME1) and MR (which she referred to as ME2) cells. ME1 (ML) contributed to the formation of the larval retractor muscle and the intestine. ME2 (MR) contributed to the formation of the retractor muscle, intestine, the kidney and the heart. Here we have not been able to detect any obvious differences in the fates between the two teloblasts. In fact, the earlier lineages exhibit a great deal of symmetry. It is unclear, however, if there are any later asymmetries that may arise from progeny of the 5ML and 5MR teloblasts after the birth of 5mL/R.
Induction and regulation within the 4d lineage
We have observed that, in general, when 4d daughter cells are removed, the structures derived from them are missing and little effect is seen in other lineages (Figure 8). However, removal of 3mL/R and/or 4mL/R leads to loss of the elongated hindgut tube, which is derived from 1m, 3m, 4m and 5m cells (Figure 8D-F; note that we did not manage to remove 5mL/R). The retractor muscles are formed mainly by 2m cells and some later-born progeny of the 5M teloblasts. Removal of 2mL and 2mR cells leads to disorganized larval retractor muscles, but they are still visible (Figure 8M). Thus, there could be some inductive interactions controlling the development of these structures as well, but little evidence of any regulation was found to take place. We have not followed 2mL/R2.2-ablated embryos to a late enough stage to assess whether definitive germ cells are missing or if other tissues can compensate. The data show that there is very little regulation for missing tissues when specific 4d sublineages are ablated, suggesting that each lineage may be specified autonomously. Elegant cell ablation studies in several clitellate species demonstrated that segmental identity of primary m blast cells (as well as primary blast cells from ectodermal lineages) is specified autonomously at or very near their birth [75, 107, 108].
In Ilyanassa, the 4d cell gives rise to intestine, heart and various other structures . When 4d is ablated these structures are missing, but there are ectodermal defects as well . The shells were smaller and the foot mass was abnormal. Furthermore, Cather  suggests that endoderm/archenteron is necessary for the shell gland to develop in Ilyanassa. In this species, the 3d macromere carries out a majority of the organizing activity, but the 4d cell has some organizing activity as well, which might explain some of the ectodermal defects in 4d ablations . Since the 4d mesentoblast carries out all organizing activity in Crepidula, one might expect to see analogous ectodermal defects when it is ablated. On the contrary, we have not observed dramatic ectodermal defects in our ablations of 4d sublineages, including no obvious reduction of the shell.
Comparisons of 4d lineage in C. fornicata and C. convexa
C. convexa is a direct-developing species [56, 60, 61]. The embryos do not form well-developed velar lobes, which are instead vestigial. Unlike in C. fornicata, there is accelerated development of the head, foot and the adult shell. Furthermore, we and others have observed that the foot does not form an operculum . This is seen in embryos of large yolky-egged species with direct development. By contrast, species with larval development, such as C. fornicata, do form an operculum, as do direct-developers produced from small eggs that increase in size by consumption of nurse eggs (Lesoway, personal observation), but these are subsequently lost at metamorphosis. No adult calyptraeids have an operculum . C. convexa does form some transient larval structures, such as components of the larval kidney complex, that include the external absorptive cells and the autofluorescent crystal cells, which are eventually lost at or shortly after hatching. The cleavage pattern of the teloblasts is identical to that in C. fornicata, as has been observed through the formation of the 4mL/R and 4ML/R teloblasts (compare Figures 1, 2 and 9).
Evolution of 4d cleavage patterns in spiralians
One of the enduring paradoxes of studying spiralian development is how such a conserved early cleavage program generates radically different adult body patterns, which may or may not include a hardened exoskeleton (for example, gastropod mollusks versus nemerteans), segmentation (for example, annelids versus nemerteans) and the formation of a coelomic cavity (for example, annelids versus polyclad turbellarians). A crucial intermediate step between the spiral cleavage program and the clade-specific generation of the bilaterally symmetric adult (or larval) body plan is the bilateral division of the 4d cell. Although formation and bilateral division of the 4d cell is often considered one of the most conserved features of spiralian development ([4, 10]; see exceptions in [13, 79, 110, 111]), there appears to be a great amount of variation in the cleavage patterns of the 4d daughter cells, the mesodermal teloblasts ([21–23, 52, 58] and this work). Studying this variation should provide keys to understanding the morphological diversity in spiralians.
In clitellate annelids, DM" (4d) divides making exceptionally large M teloblasts that undergo highly asymmetric divisions and the daughter cells are always born at the same spot off of the teloblast, generating a coherent chain of cells [17, 62, 81, 82]. The 4d cell division pattern in Crepidula is very different from this situation. The early cleavages of another caeneogastropod, Ilyanassa, are similar overall to Crepidula, but exhibit interesting signs of divergence in the arrangement of cells in the clone [20, 21]. For example, in both species the first three divisions of the teloblast are reminiscent of spiral cleavage, with the 1m, 2m and 3m cells being born off of alternating sides. Then at its fourth division the teloblast starts to produce a more or less coherent chain of cells. In Crepidula this chain extends away from the teloblasts, while in Ilyanassa it wraps around the teloblast . The variation in 4d cleavage patterns in mollusks makes an interesting test case for future studies of the evolution of cell division and cell fate within the endomesoderm. For example, in other systems, such as vertebrates and sea urchins, endomesoderm specification happens by inductive signals in a field of cells [105, 112], while in spiralians specification is lineage-driven and might rely more heavily on cell autonomous mechanisms such as differential inheritance of cytoplasmic determinants [104, 113, 114].
Evolution of novel cell types
Studying spiralian development in a phylogenetic framework has the potential to reveal how novel characters arise within a stereotyped cleavage program. The 4d lineage gives rise to numerous tissues, some of which are almost certainly homologous across spiralian phyla such as the heart, germline and intestine. Other 4d-derived tissues are specific to particular groups such as the retractor muscles in mollusks, the larval kidney complexes in caenogastropods, or the proboscis of glossiphoniid leeches. Clearly, these evolutionarily novel tissues can be readily incorporated into the early division of the 4d lineage, since the founder cells of each arise within the first few divisions of the teloblasts (this work and ). Having detailed fate maps for the clonal origin of these tissues is a crucial step to understanding the molecular basis of their evolution.
Evolution of 4d germ layer contributions
Our data show that, in Crepidula, multiple teloblast sublineages give rise to endoderm, while other sublineages give rise to mesodermal fates. Likewise, in Ilyanassa, 4d gives rise to both mesoderm and endoderm [21, 59]. Although 4d is often referred to as a mesentoblast in the spiralian literature, it does not give rise to endoderm in all species [10, 13, 18, 24]. For example, in the bivalve Sphaerium, separation of endoderm and mesoderm happens at the birth of M (4d) and D (4D); 4d gives rise exclusively to mesoderm (and germline) and 4D gives rise exclusively to endoderm (Table 1; [52, 53, 78]).
Among annelids, 4d’s contribution to the endoderm is more controversial. Several older studies reported no contribution to endoderm (for example, in Bdellodrilus and Nereis, see Table 1). More recently in Platynereis, Ackerman et al. observed 4d-derived cells as a tight ball at the end of the hindgut, but could not confirm whether these were endodermal cells. These cells might, in fact, be germline precursors [72, 73]. The 4d lineage in Capitella is particularly unusual because, while it does give rise to the germline, it does not generate discrete mesodermal bands or endoderm . In Helobdella, it is traditionally reported that 4d does not contribute to endoderm directly. Rather it does so indirectly when spent M teloblasts and supernumerary m blast cells fuse with the endodermal syncytial yolk cell, which subsequently cellularizes to form the gut epithelium [17, 67–69]. However, a recent study found that em1 and em2 (equivalent to 1m and 2m by birth rank) contribute to epithelium that lines the digestive tract directly, and therefore may be considered endodermal cells .
The insights mentioned above emphasize the importance of undertaking careful lineage analyses using long-lived lineage tracers and methods like live imaging using confocal microscopy [13, 18, 63, 115]. A further demonstration is the fact that recent high-resolution lineage studies have revealed that 4d might also produce ectodermal cells. Meyer et al.  argue that Capitella’s 4d cell represents a ‘mesectoblast’ because it gives rise to ectodermal cells that make up the anus (Table 1). In Helobdella, 4d appears to contribute to the formation of all three germ layers, because it gives rise to cells in the provisional integument and elements of the nervous system [16, 18, 71]. Our analysis of Crepidula revealed that 4d forms specialized cells exposed on the surface of the larva (the external absorptive cells derived from 2mL2.1/R2.1[54, 55]). If these cells are ectodermal, then the 4d blastomere should be considered an ‘ectomesentoblast’ in Crepidula as well (Table 1).
Together, these studies uncover significant diversity in terms of the contributions of 4d in these different spiralian embryos. In reviewing the information summarized in Table 1, which includes a more basal representatives (a polyclad turbellarian in the Platyhelminthes), one might argue that the ancestral condition is one in which 4d gave rise to both endoderm and mesoderm, in addition to the primordial germ cells. Therefore, contributions to ectodermal fates may represent traits more recently derived within certain organisms.
The authors thank the community of the Marine Biological Laboratory and especially Drs. Nipam Patel and Lee Niswander for their support. The authors also acknowledge generous assistance and advice from Hamilton-Thorne. DCL thanks DR McClay for his support. Professional Development for MPL was partially supported by EDENRCN, NSF Grant #105-0955517. This material is based upon work supported by the National Science Foundation under Grant No. 1121268 to JQH (JJH). Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.
- Hinman VF, Yankura KA, McCauley BS: Evolution of gene regulatory network architectures: examples of subcircuit conservation and plasticity between classes of echinoderms. Biochim Biophys Acta. 2009, 1789: 326-332. 10.1016/j.bbagrm.2009.01.004.PubMedGoogle Scholar
- Maduro MF: Structure and evolution of the C. elegans embryonic endomesoderm network. Biochim Biophys Acta. 2009, 1789: 250-260. 10.1016/j.bbagrm.2008.07.013.PubMed CentralPubMedGoogle Scholar
- Martindale MQ, Hejnol A: A developmental perspective: changes in the position of the blastopore during bilaterian evolution. Dev Cell. 2009, 17: 162-174. 10.1016/j.devcel.2009.07.024.PubMedGoogle Scholar
- Hejnol A: A twist in time - the evolution of spiral cleavage in the light of animal phylogeny. Integr Comp Biol. 2010, 50: 695-706. 10.1093/icb/icq103.PubMedGoogle Scholar
- Wilson EB: Considerations on cell-lineage and ancestral reminiscence. Ann N Y Acad Sci. 1898, 11: 1-27. 10.1111/j.1749-6632.1898.tb54960.x.Google Scholar
- Verdonk NH, van den Biggelaar JAM: Early development and the formation of the germ layers. The Mollusca. 1983, San Diego, CA: Academic, 91-122. 3Google Scholar
- 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. 10.1046/j.1420-9101.1992.5020205.x.Google Scholar
- Henry J, Martindale M: Conservation and innovation in spiralian development. Hydrobiologia. 1999, 402: 255-265.Google Scholar
- Dohle W: The ancestral cleavage pattern of the clitellates and its phylogenetic deviations. Hydrobiologia. 1999, 402: 267-283.Google Scholar
- Lambert JD: Mesoderm in spiralians: the organizer and the 4d cell. J Exp Zool (Mol Dev Evol). 2008, 308B: 15-23.Google Scholar
- Lambert JD: Developmental patterns in spiralian embryos. Curr Biol. 2010, 20: R72-R77. 10.1016/j.cub.2009.11.041.PubMedGoogle Scholar
- Boyer BC, Henry JJ, Martindale MQ: The cell lineage of a polyclad turbellarian embryo reveals close similarity to coelomate spiralians. Dev Biol. 1998, 204: 111-123. 10.1006/dbio.1998.9084.PubMedGoogle Scholar
- Meyer NP, Boyle MJ, Martindale MQ, Seaver EC: A comprehensive fate map by intracellular injection of identified blastomeres in the marine polychaete Capitella teleta. EvoDevo. 2010, 1: 8-10.1186/2041-9139-1-8.PubMed CentralPubMedGoogle Scholar
- Lambert JD, Nagy LM: The MAPK cascade in equally cleaving spiralian embryos. Dev Biol. 2003, 263: 231-241. 10.1016/j.ydbio.2003.07.006.PubMedGoogle Scholar
- 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. 10.1016/j.ydbio.2006.04.441.PubMedGoogle Scholar
- Weisblat DA, Kim SY, Stent GS: Embryonic origins of cells in the leech Helobdella triserialis. Dev Biol. 1984, 104: 65-85. 10.1016/0012-1606(84)90037-X.PubMedGoogle Scholar
- Weisblat DA, Huang FZ: An overview of glossiphoniid leech development. Can J Zool. 2001, 79: 218-232. 10.1139/z00-199.Google Scholar
- Gline SE, Nakamoto A, Cho SJ, Chi C, Weisblat DA: Lineage analysis of micromere 4d, a super-phylotypic cell for Lophotrochozoa, in the leech Helobdella and the sludgeworm Tubifex. Dev Biol. 2011, 353: 120-133. 10.1016/j.ydbio.2011.01.031.PubMed CentralPubMedGoogle Scholar
- Goto A, Kitamura K, Shimizu T: Cell lineage analysis of pattern formation in theTubifexembryo I. Segmentation in the mesoderm. Int J Dev Biol. 1999, 43: 317-327.PubMedGoogle Scholar
- Goulding MB: Comparative and experimental analysis of precocious cell-lineage diversification in the embryonic dorsoventral axis of the gastropod Ilyanassa. 2001, Austin: University of Texas, PhD thesisGoogle Scholar
- Rabinowitz JS, Chan XY, Kingsley EP, Duan Y, Lambert JD: Nanos is required in somatic blast cell lineages in the posterior of a mollusc embryo. Curr Biol. 2008, 18: 331-336. 10.1016/j.cub.2008.01.055.PubMedGoogle Scholar
- Henry JJ, Collin R, Perry KJ: The slipper snail, Crepidula: an emerging lophotrochozoan model system. Biol Bull. 2010, 218: 211-229.PubMedGoogle Scholar
- Conklin EG: The embryology of Crepidula, a contribution to the cell lineage and early development of some marine gastropods. J Morphol. 1897, 13: 1-226. 10.1002/jmor.1050130102.Google Scholar
- 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. 10.1016/j.ydbio.2007.01.044.PubMedGoogle Scholar
- Henry JQ, Perry KJ, Martindale MQ: beta-catenin and early development in the gastropod, Crepidula fornicata. Integr Comp Biol. 2010, 50: 707-719. 10.1093/icb/icq076.PubMedGoogle Scholar
- Henry Lab Home Page: Crepidula Methods.http://www.life.illinois.edu/henry/crepidula_methods.html,
- Henry JJ, Martindale MQ: Conservation of the spiralian developmental program: cell lineage of the nemertean, Cerebratulus lacteus. Dev Biol. 1998, 201: 253-269. 10.1006/dbio.1998.8966.PubMedGoogle Scholar
- Henry JQ, Martindale MQ, Boyer BC: Axial specification in a basal member of the spiralian clade: lineage relationships of the first four cells to the larval body plan in the polyclad turbellarian Hoploplana inquilina. Biol Bull. 1995, 189: 194-195.Google Scholar
- Martindale MQ, Henry JQ: Diagonal development: establishment of the anal axis in the ctenophore Mnemiopsis leidyi. Biol Bull. 1995, 189: 190-192.Google Scholar
- Tadir Y, Douglas-Hamilton DH: Laser effects in the manipulation of human eggs and embryos for in vitro fertilization. Methods Cell Biol. 2007, 82: 409-431.PubMedGoogle Scholar
- Douglas-Hamilton DH: Laser Heating of Embryos and Cells in the ZILOS/ZYclone system. 2008, Bererly, MA: Hamilton Thorne BiosciencesGoogle Scholar
- Henry JJ: The role of unequal cleavage and the polar lobe in the segregation of developmental potential during first cleavage in the embryo of Chætopterus variopedatus. Roux's Arch Dev Biol. 1986, 195: 103-116. 10.1007/BF00456106.Google Scholar
- Crampton HE: Experimental studies on gastropod development. Wilhelm Roux Arch Entwickl Mech Org. 1896, 3: 1-19.Google Scholar
- Novikoff AB: Embryonic determination in the annelid, Sabellaria vulgaris. II. Transplantation of polar lobes and blastomeres as a test of their inducing capacities. Biol Bull. 1938, 74: 211-234. 10.2307/1537755.Google Scholar
- 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. 10.1002/jez.1401210310.Google Scholar
- Rattenbury JC, Berg WE: Embryonic segregation during early development of Mytilus edulis. J Morphol. 1954, 95: 393-414. 10.1002/jmor.1050950302.Google Scholar
- Atkinson JW: Organogenesis in normal and lobeless embryos of the marine prosobranch gastropod Ilyanassa obsoleta. J Morphol. 1971, 133: 339-352. 10.1002/jmor.1051330307.Google Scholar
- Cather JN, Verdonk NH: The development of Bithynia tentaculata (Prosobranchia, Gastropoda) after removal of the polar lobe. J Embryol Exp Morphol. 1974, 31: 415-422.PubMedGoogle Scholar
- van Dongen CAM, Geilenkirchen WLM: The development of Dentalium with special reference to the significance to the significance of the polar lobe. 3. Division chronology and development of cell pattern in Dentalium dentale (Scaphopoda). Proc Koninklijke Nederlandse Akademie van Wetenschappen Series C Biol Med Sci. 1974, 77: 85-100.Google Scholar
- van Dongen CAM, Geilenkirchen WLM: The development of Dentalium with special reference to the significance to the significance of the polar lobe. 1. Division chronology and development of cell pattern in Dentalium dentale (Scaphopoda). Proc Koninklijke Nederlandse Akademie van Wetenschappen Series C Biol Med Sci. 1974, 77: 57-70.Google Scholar
- van Dongen CAM, Geilenkirchen WLM: The development of Dentalium with special reference to the significance to the significance of the polar lobe. 2. Division chronology and development of cell pattern in Dentalium dentale (Scaphopoda). Proc Koninklijke Nederlandse Akademie van Wetenschappen Series C Biol Med Sci. 1974, 77: 71-84.Google Scholar
- van Dongen CAM, Geilenkirchen WLM: The development of Dentalium with special reference to the significance to the significance of the polar lobe. 4. Division chronology and development of cell pattern in Dentalium dentale after removal of polar lobe at 1st cleavage. Proc Koninklijke Nederlandse Akademie van Wetenschappen Series C Biol Med Sci. 1975, 78: 358-375.Google Scholar
- van Dongen CAM: The development of Dentalium with special reference to the significance to the significance of the polar lobe. 7. Organogenesis and hostogenesis in lobeless embryos of Dentalium vulgare (Da Costa) as compared to normal development. Proc Koninklijke Nederlandse Akademie van Wetenschappen Series C Biol Med Sci. 1976, 79: 454-465.Google Scholar
- van Dongen CAM: The development of Dentalium with special reference to the significance to the significance of the polar lobe. 6. Differentiation of cell pattern in lobeless embryos of Dentalium vulgare (Da Costa) during late larval development. Proc Koninklijke Nederlandse Akademie van Wetenschappen Series C Biol Med Sci. 1976, 79: 256-266.Google Scholar
- van Dongen CAM: The development of Dentalium with special reference to the significance to the significance of the polar lobe. 5. Differentiation of cell pattern in lobeless embryos of Dentalium vulgare (Da Costa) during late larval development. Proc Koninklijke Nederlandse Akademie van Wetenschappen Series C Biol Med Sci. 1976, 79: 245-255.Google Scholar
- Cather JN, Verdonk NH, Dohmen MR: Role of the vegetal body in the regulation of development in Bithynia tentaculata (Prosobranchia, Gastropoda). Am Zool. 1976, 16: 455-468.Google Scholar
- Render JA: The second polar lobe of Sabellaria cementarium plays an inhibitory role in apical tuft formation. Roux's Arch Dev Biol. 1983, 192: 120-129.Google Scholar
- Wilson EB: Experimental studies on germinal localization. I. The germ-regions in the egg of Dentalium. J Exp Zool. 1904, 1: 1-72. 10.1002/jez.1400010102.Google Scholar
- Wilson EB: Experimental studies in germinal localization. II. Experiments on the cleavage-mosaic in Patella and Dentalium. J Exp Zool. 1904, 1 (Wilson EB): 197-268.Google Scholar
- Hatt P: Essais expérimentaux sur les localizations germinales dans l'oeuf d'un annelide (Sabellaria alveolata L.). Arch D'anat Micros. 1932, 28: 81-98.Google Scholar
- Swartz SZ, Chan XY, Lambert JD: Localization of Vasa mRNA during early cleavage of the snail Ilyanassa. Dev Genes Evol. 2008, 218: 107-113. 10.1007/s00427-008-0203-6.PubMedGoogle Scholar
- Woods FH: History of the germ cells in Sphaerium striatinum (Lam.). J Morphol. 1931, 51: 545-10.1002/jmor.1050510208.Google Scholar
- Woods FH: Keimbahn determinants and continuity of the germ cells in Sphaerium striatinum (Lam.). J Morphol. 1932, 53: 345-365. 10.1002/jmor.1050530207.Google Scholar
- Ruthensteiner B, Schaefer K: On the protonephridia and 'larval kidneys' of Nassarius (Hinia) reticulatus (Linnaeus) (Caenogastropoda). J Molluscan Stud. 1991, 57: 323-329. 10.1093/mollus/57.3.323.Google Scholar
- Rivest BR: Studies on the structure and function of the larval kidney complex of prosobranch gastropods. Biol Bull. 1992, 182: 305-323. 10.2307/1542251.Google Scholar
- Collin R: Worldwide patterns in mode of development in calyptraeid gastropods. Mar Ecol Prog Ser. 2003, 247: 103-122.Google Scholar
- Pechenik J: The relationship between temperature, growth rate, and duration of planktonic life for larvae of the gastropod Crepidula fornicata (L.). J ExpMar Biol Ecol. 1984, 74: 241-257. 10.1016/0022-0981(84)90128-X.Google Scholar
- Wilson EB: A cell-lineage of Nereis. A contribution to the cytogeny of the annelid body. J Morphol. 1892, 6: 361-481. 10.1002/jmor.1050060301.Google Scholar
- Render JA: Cell fate maps in the Ilyanassa obsoleta embryo beyond the third division. Dev Biol. 1997, 189: 301-310. 10.1006/dbio.1997.8654.PubMedGoogle Scholar
- Hendler G, Franz DR: Population dynamics and life history of Crepidula convexa Say (Gastropoda: Prosobranchia) in Delaware Bay. Biol Bull. 1971, 141: 514-526. 10.2307/1540265.Google Scholar
- Hoagland KE: Patterns of encapsulation and brooding in the calyptraeidae (Prosobranchia: Mesogastropoda). Am Malacol Bull. 1986, 4: 173-183.Google Scholar
- Whitman CO: The embryology of Clepsine. Q J Microsc Sci. 1878, 18: 213-315.Google Scholar
- Wilson EB: Cell-lineage and ancestral reminiscence. Biological Lectures from the Marine Biological Laboratory, Woods Hole, Massachusetts. 1898, Boston: Ginn & Co, 21-42.Google Scholar
- Tannreuther GW: The embryology of Bdellodrilus philadelphicus. J Morphol. 1915, 26: 143-216. 10.1002/jmor.1050260202.Google Scholar
- Dill KK, Seaver EC: Vasa and nanos are coexpressed in somatic and germ line tissue from early embryonic cleavage stages through adulthood in the polychaete Capitella sp. I. Dev Genes Evol. 2008, 218: 453-463. 10.1007/s00427-008-0236-x.PubMedGoogle Scholar
- Giani VC, Yamaguchi E, Boyle MJ, Seaver EC: Somatic and germline expression of piwi during development and regeneration in the marine polychaete annelid Capitella teleta. EvoDevo. 2011, 2: 10-10.1186/2041-9139-2-10.PubMed CentralPubMedGoogle Scholar
- Desjeux I, Price DJ: The production and elimination of supernumerary blast cells in the leech embryo. Dev Genes Evol. 1999, 209: 284-293. 10.1007/s004270050254.PubMedGoogle Scholar
- Liu N-JL, Isaksen DE, Smith CM, Weisblat DA: Movements and stepwise fusion of endodermal precursor cells in leech. Dev Genes Evol. 1998, 208: 117-127. 10.1007/s004270050162.PubMedGoogle Scholar
- Nardelli-Haefliger D, Shankland M: Lox10, a member of the NK-2 homeobox gene class, is expressed in a segmental pattern in the endoderm and in the cephalic nervous system of the leech Helobdella. Development. 1993, 118: 877-892.PubMedGoogle Scholar
- Gline SE, Kuo DH, Stolfi A, Weisblat DA: High resolution cell lineage tracing reveals developmental variability in leech. Dev Dyn. 2009, 238: 3139-3151. 10.1002/dvdy.22158.PubMed CentralPubMedGoogle Scholar
- Weisblat DA, Shankland M: Cell lineage and segmentation in the leech. Philos Trans R Soc Lond B Biol Sci. 1985, 312: 39-56. 10.1098/rstb.1985.0176.PubMedGoogle Scholar
- Rebscher N, Zelada-González F, Banisch TU, Raible F, Arendt D: Vasa unveils a common origin of germ cells and of somatic stem cells from the posterior growth zone in the polychaete Platynereis dumerilii. Dev Biol. 2007, 306: 599-611. 10.1016/j.ydbio.2007.03.521.PubMedGoogle Scholar
- Rebscher N, Lidke AK, Ackermann CF: Hidden in the crowd: primordial germ cells and somatic stem cells in the mesodermal posterior growth zone of the polychaete Platynereis dumerillii are two distinct cell populations. EvoDevo. 2012, 3: 9-10.1186/2041-9139-3-9.PubMed CentralPubMedGoogle Scholar
- Ackermann C, Dorresteijn A, Fischer A: Clonal domains in postlarval Platynereis dumerilii (Annelida: Polychaeta). J Morphol. 2005, 266: 258-280. 10.1002/jmor.10375.PubMedGoogle Scholar
- Kitamura K, Shimizu T: Analyses of segment-specific expression of alkaline phosphatase activity in the mesoderm of the oligochaete annelid Tubifex: implications for specification of segmental identity. Dev Biol. 2000, 219: 214-223. 10.1006/dbio.1999.9602.PubMedGoogle Scholar
- 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 Invertebr Reprod Dev. 1986, 9: 155-168. 10.1080/01688170.1986.10510193.Google Scholar
- Okada K, Sci Rep Tohoku Imp Univ Ser: The development of the primary mesoderm in Sphaerium japonicum biwaense Mori. Sci Rep Tohoku Imp Univ Ser 4. 1939, 14: 25-47.Google Scholar
- Zeigler HE: Die Entwicklung von Cyclas cornea Lam. 1885, Bd: Zeit f wiss Zool, 41-Google Scholar
- Boyer BC, Henry JQ, Martindale MQ: Dual origins of mesoderm in a basal spiralian: cell lineage analyses in the polyclad turbellarian Hoploplana inquilina. Dev Biol. 1996, 179: 329-338. 10.1006/dbio.1996.0264.PubMedGoogle Scholar
- Goulding MQ: Cell lineage of the Ilyanassa embryo: evolutionary acceleration of regional differentiation during early development. PLoS One. 2009, 4: e5506-10.1371/journal.pone.0005506.PubMed CentralPubMedGoogle Scholar
- Shimizu T, Nakamoto A: Segmentation in annelids: cellular and molecular basis for metameric body plan. Zoolog Sci. 2001, 18: 285-298. 10.2108/zsj.18.285.Google Scholar
- Weisblat DA: Leeches. Curr Biol. 2003, 13: R752-10.1016/j.cub.2003.09.011.PubMedGoogle Scholar
- Brusca RC, Brusca GJ: Invertebrates. 1990, Sunderland, MA: SinauerGoogle Scholar
- Tzetlin A: G P: Pharynx and intestine. Hydrobioilogia. 2005, 535/536: 199-225. 10.1007/s10750-004-1431-z.Google Scholar
- Michel C: The ultrastructure of Polychaeta. Microfauna Marina. Edited by: Westheide W, Hermans CO. 1988, Stuttgart: Gustav Fischer Verlag, 5-175. 4Google Scholar
- Verdonk NH, Cather JN: Morphogenetic determination and differentiation. The Mollusca. 1983, San Diego, CA: Academic, 215-252. 3Google Scholar
- Demian ES, Yousif F: Embryonic development and organogenesis in the snail Marisa cornuarietis. III. Development of the circulatory and renal systems. Malacologia. 1973, 12: 175-194.PubMedGoogle Scholar
- Okada K: Some notes on Sphaerium japonicum biwaense Mori, a freshwater bivalve. IV. Gastrula and fetal larva. Sci Rep Tohoku Imp Univ Ser 4. 1936, 11: 49-68.Google Scholar
- Clement AC: The embryonic value of the micromeres in Ilyanassa obsoleta, as determined by deletion experiments. II. The second quartet cells. Int J Invertebr Reprod Dev. 1986, 9: 139-153. 10.1080/01688170.1986.10510192.Google Scholar
- Clement AC: Development of the Ilyanassa embryo after the removal of the mesentoblast cell. Biol Bull. 1960, 119: 310-Google Scholar
- Lambert JD, Nagy LM: MAPK signaling by the D quadrant embryonic organizer of the mollusc Ilyanassa obsoleta. Development. 2001, 128: 45-56.PubMedGoogle Scholar
- Henry JJ, Perry KJ: MAPK activation and the specification of the D quadrant in the gastropod mollusc, Crepidula fornicata. Dev Biol. 2008, 313: 181-195. 10.1016/j.ydbio.2007.10.019.PubMedGoogle Scholar
- Kang D, Pilon M, Weisblat DA: Maternal and zygotic expression of a nanos-class gene in the leech Helobdella robusta: primordial germ cells arise from segmental mesoderm. Dev Biol. 2002, 245: 28-41. 10.1006/dbio.2002.0615.PubMedGoogle Scholar
- Fabioux C, Huvet A, Lelong C, Robert R, Pouvreau S, Daniel JY, Minguant C, Le Pennec M: Oyster vasa-like gene as a marker of the germline cell development in Crassostrea gigas. Biochem Biophys Res Commun. 2004, 320: 592-598. 10.1016/j.bbrc.2004.06.009.PubMedGoogle Scholar
- Oyama A, Shimizu T: Transient occurrence of vasa-expressing cells in nongenital segments during embryonic development in the oligochaete annelid Tubifex tubifex. Dev Genes Evol. 2007, 217: 675-690. 10.1007/s00427-007-0180-1.PubMedGoogle Scholar
- Kranz AM, Tollenaere A, Norris BJ, Degnan BM, Degnan SM: Identifying the germline in an equally cleaving mollusc: Vasa and Nanos expression during embryonic and larval development of the vetigastropod Haliotis asinina. J Exp Zool B Mol Dev Evol. 2010, 314: 267-279.PubMedGoogle Scholar
- Griffond B: Individualization and organogenesis of embryonic gonad in gonochoric prosobranch mollusc Viviparus viviparus (L.). Roux Arch Dev Biol. 1977, 183: 131-147. 10.1007/BF00848783.Google Scholar
- Richardson BE, Lehmann R: Mechanisms guiding primordial germ cell migration: strategies from different organisms. Nat Rev Mol Cell Biol. 2010, 11: 37-49. 10.1038/nrm2815.PubMed CentralPubMedGoogle Scholar
- Ikenishi K: Germ plasm in Caenorhabditis elegans, Drosophila and Xenopus. Dev Growth Differ. 1998, 40: 1-10. 10.1046/j.1440-169X.1998.t01-4-00001.x.PubMedGoogle Scholar
- Dohmen MR, Lok D: The ultrastructure of the polar lobe of Crepidula fornicata (Gastropoda, Prosobranchia). J Embryol Exp Morphol. 1975, 34: 419-428.PubMedGoogle Scholar
- Dohmen MR, Verdonk NH: The ultrastructure and role of the polar lobe in development of molluscs. Determinants of Spatial Organization. Edited by: Subtelny S, Konigsberg IR. 1979, New York: Academic, 3-Google Scholar
- Dohmen MR, Verdonk NH: Cytoplasmic localizations in mosaic eggs. Maternal Effects of Development. Edited by: Newth DR, Balls M. 1979, Cambridge: Cambridge University Press, 127-145.Google Scholar
- Raven CP: Morphogenesis: the Analysis of Molluscan Development. 1966, Oxford: Pergamon, 2Google Scholar
- Henry JJ, Perry KJ, Fukui L, Alvi N: Differential localization of mRNAs during early development in the mollusc, Crepidula fornicata. Integr Comp Biol. 2010, 50: 720-733. 10.1093/icb/icq088.PubMedGoogle Scholar
- Logan CY, Miller JR, Ferkowicz MJ, McClay DR: Nuclear β-catenin is required to specify vegetal cell fates in the sea urchin embryo. Development. 1999, 126: 345-357.PubMedGoogle Scholar
- Yajima M, Wessel GM: Small micromeres contribute to the germline in the sea urchin. Development. 2011, 138: 237-243. 10.1242/dev.054940.PubMed CentralPubMedGoogle Scholar
- Martindale MQ, Shankland M: Intrinsic segmental identity of segmental founder cells of the leech embryo. Nature. 1990, 347: 672-674. 10.1038/347672a0.PubMedGoogle Scholar
- Gleizer L, Stent GS: Developmental origin of segmental identity in the leech mesoderm. Development. 1993, 117: 177-189.PubMedGoogle Scholar
- Cather JN: Cellular interactions in the development of the shell gland of the gastropod, Ilyanassa. J Exp Zool. 1967, 166: 205-223. 10.1002/jez.1401660204.PubMedGoogle Scholar
- Surface FM: The early development of a polyclad, Planocera inquilina Wh. Proc Acad Nat Sci Phila. 1907, 59: 514-559.Google Scholar
- van den Biggelaar JAM: Cleavage pattern and mesentoblast formation in Acanthochiton crinitus (Polyplacophora, Mollusca). Dev Biol. 1996, 174: 423-430. 10.1006/dbio.1996.0085.PubMedGoogle Scholar
- Fukuda K, Kikuchi Y: Endoderm development in vertebrates: fate mapping, induction and regional specification. Dev Growth Differ. 2005, 47: 343-355. 10.1111/j.1440-169X.2005.00815.x.PubMedGoogle Scholar
- Kingsley EP, Chan XY, Duan Y, Lambert JD: Widespread RNA segregation in a spiralian embryo. Evol Dev. 2007, 9: 527-539. 10.1111/j.1525-142X.2007.00194.x.PubMedGoogle Scholar
- Lambert JD, Nagy LM: Asymmetric inheritance of centrosomally localized mRNAs during embryonic cleavages. Nature. 2002, 420: 682-686. 10.1038/nature01241.PubMedGoogle Scholar
- Meyer NP, Seaver EC: Cell lineage and fate map of the primary somatoblast of the polychaete annelid Capitella teleta. Integr Comp Biol. 2010, 50: 756-767. 10.1093/icb/icq120.PubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.