New insights from a high-resolution look at gastrulation in the sea urchin, Lytechinus variegatus

Abstract

Background

Gastrulation is a complex orchestration of movements by cells that are specified early in development. Until now, classical convergent extension was considered to be the main contributor to sea urchin archenteron extension, and the relative contributions of cell divisions were unknown. Active migration of cells along the axis of extension was also not considered as a major factor in invagination.

Results

Cell transplantations plus live imaging were used to examine endoderm cell morphogenesis during gastrulation at high-resolution in the optically clear sea urchin embryo. The invagination sequence was imaged throughout gastrulation. One of the eight macromeres was replaced by a fluorescently labeled macromere at the 32 cell stage. At gastrulation those patches of fluorescent endoderm cell progeny initially about 4 cells wide, released a column of cells about 2 cells wide early in gastrulation and then often this column narrowed to one cell wide by the end of archenteron lengthening. The primary movement of the column of cells was in the direction of elongation of the archenteron with the narrowing (convergence) occurring as one of the two cells moved ahead of its neighbor. As the column narrowed, the labeled endoderm cells generally remained as a contiguous population of cells, rarely separated by intrusion of a lateral unlabeled cell. This longitudinal cell migration mechanism was assessed quantitatively and accounted for almost 90% of the elongation process. Much of the extension was the contribution of Veg2 endoderm with a minor contribution late in gastrulation by Veg1 endoderm cells. We also analyzed the contribution of cell divisions to elongation. Endoderm cells in Lytechinus variagatus were determined to go through approximately one cell doubling during gastrulation. That doubling occurs without a net increase in cell mass, but the question remained as to whether oriented divisions might contribute to archenteron elongation. We learned that indeed there was a biased orientation of cell divisions along the plane of archenteron elongation, but when the impact of that bias was analyzed quantitatively, it contributed a maximum 15% to the total elongation of the gut.

Conclusions

The major driver of archenteron elongation in the sea urchin, Lytechinus variagatus, is directed movement of Veg2 endoderm cells as a narrowing column along the plane of elongation. The narrowing occurs as cells in the column converge as they migrate, so that the combination of migration and the angular convergence provide the major component of the lengthening. A minor contributor to elongation is oriented cell divisions that contribute to the lengthening but no more than about 15%.

1. Introduction

Gastrulation is a dynamic period in the development of an embryo and involves many different cell movements. Given the importance of early morphogenesis in establishing differences in animal body plans, it would be valuable to understand how those processes work. In the sea urchin, gastrulation is relatively simple, easy to observe, and thought to be the prototype of deuterostome gastrulation, all of which makes it a valuable model for investigating questions of morphogenesis, in particular, gastrulation.

At embryonic 4^th cleavage of the sea urchin, an unequal cell division gives rise to four macromeres in the vegetal hemisphere (Figure 1A). At 5^th cleavage, the tier of four macromeres divides meridianally to produce 8 cells. In the experiments below, one of these 8 macromeres was replaced with an identical but fluorescently labeled cell and the progeny imaged by timelapse later during gastrulation. At 6^th cleavage, the macromeres divide in an equatorial plane to become the lower Veg2 and upper Veg1 tiers of macromeres. Veg2 progeny give rise to endomesoderm, which later becomes the non-skeletogenic mesoderm plus endoderm of the foregut and midgut. The Veg1 tier gives rise to endoderm of the midgut and hindgut as well as contributing to posterior ectoderm (Logan & McClay, 1997). The gastrulation movements analyzed here will describe the movements primarily of the Veg2 endoderm progeny.

Figure 1.

Transplantation of cells in the gut lineage allows gastrulation to be captured at a higher resolution. (A) Progeny of the Veg tier of cells will comprise the entire gut. (B) Transplantation of a single membrane-GFP/histone2B-RFP labeled Veg macromere to an equivalent location in an unlabeled embryo of the same stage. (C) The transplanted cell produces 16–32 Veg progeny cells (depending on time of doubling) found in a patch at the beginning of gastrulation. Time-lapse visualization of the Veg lineage shows the cells slide by one another parallel to the axis of elongation—with intercalation occurring at an angle as one cell slips past another as both move along the axis of elongation.

At the beginning of gastrulation in Lytechinus variagatus, the Veg2 tier of cells is found at the vegetal plate, and the Veg1 tier is found adjacent to the vegetal plate (Logan & McClay, 1997). Gastrulation begins between ninth and tenth cleavage such that the Veg1 and Veg2 lineages constitute between 128 and 256 cells at the beginning of gastrulation. A small number of those cells soon go through an epithelial-mesenchymal transition to become pigment cells, while the remaining Veg1 and Veg2 cells then engage in archenteron invagination (Ettensohn, 1984; 1985; Hardin, 1988; 1989; 1987).

Primary invagination of the archenteron occurs by a series of cell shape changes to force the initial inbending of the flat vegetal plate. Many mechanical and cell biological properties have been observed and mechanistically proposed to account for the inbending of the cell sheet. Veg2 endomesoderm cells elongate along the apico-basal axis to form bottle cells, which results in a thickened vegetal plate (Amemiya, 1989). Other contributions in addition to bottle cells have been suggested by computational approaches using algorithms to predict the forces contributing to the inbending (L. A. Davidson, Koehl, Keller, & Oster, 1995; L. A. Davidson, Oster, Keller, & Koehl, 1999; Kimberly & Hardin, 1998; Lane, Koehl, Wilt, & Keller, 1993). The hyaline layer serves as a major mechanical component of the blastula wall and retains stiffness to promote the bending of the vegetal plate (L. A. Davidson et al., 1999). Osmotic pressure differences in the blastocoel, do not generate the major force of invagination (contrary to Rhumbler, 1902), but osmotic pressure could contribute to the process and ease the inward bending by decreasing before the onset of primary invagination (Takata & Kominami, 2001).

Secondary invagination is the main elongation step of the archenteron. It has long been thought that the main driving force of sea urchin gut elongation is by way of mediolateral convergent extension (Ettensohn, 1985; Hardin, 1989). Then, once the archenteron reaches about two thirds its final length a third process becomes a contributor to the extension. Filopodia at the tip of the archenteron extend and attach to the basement membrane lining the blastocoel roof and assist in pulling the archenteron to its final length near the animal pole (Hardin & McClay, 1990). Coincidentally, the Veg1 cells begin to make their way into the gut as a late population of cells contributing to the mass of the archenteron (Logan & McClay, 1997). Thus, much of the initial length of the gut is provided by the Veg2 endoderm morphogenesis with most of the Veg1 cells remaining near the vegetal plate until late in gastrulation. As will be described below, the cells of the Veg2 endoderm lineage contributes to the majority of extension by cell migration, cell division, and by convergence-extension. The goal of this analysis was to tease apart the relative contributions of each mechanism.

Convergent-extension is classically understood to be the polarized intercalation of cells that converge toward the dorsal midline (perpendicular to elongation) to enable a tissue to extend in the anterior-posterior direction. By observing such cell movements at a higher resolution, our aim was to gain more insights into how, or the extent to which, convergent-extension contributes to sea urchin archenteron elongation. By observing and tracking the movements of many individual cells at a higher resolution, we wanted to quantitatively analyze how and the extent to which, convergent-extension contributes to sea urchin archenteron elongation.

Another mechanism thought to contribute to directed morphogenesis in many organisms are regionalized cell divisions that bias the direction of animal tissue elongations. Divisions oriented preferentially to an elongating axis have the potential to facilitate polarized extension of that tissue. This mechanism has been reported for zebrafish gastrulation, for example (Gong, Mo, & Fraser, 2004; Quesada-Hernández et al., 2010). In chick primitive streak formation, cells divide at the midline to orient primitive streak development (Wei & Mikawa, 2000). Oriented cell divisions have also been well documented in Drosophila germ band extension (da Silva & Vincent, 2007).

Although studies have shown regionalized cell division in the sea urchin developing archenteron, the relative contribution of those divisions to the lengthening of the gut is not known (Nislow & Morrill, 1988). Previous studies in the species Lytechinus pictus report minimal divisions during gastrulation (a total 15% increase in cell number) (Hardin & Cheng, 1986). When DNA synthesis was inhibited using Aphidicolin, archenteron elongation was not significantly affected (Stephens, Hardin, Keller, & Wilt, 1986). In this same species, it was reported that cell rearrangement was a major contributor to archenteron elongation (Hardin, 1989; 1987; Hardin & Cheng, 1986). It was hypothesized by Nislow, et al (1988) that perhaps other species that have more cell divisions require fewer cell rearrangements, and those species that have fewer cell divisions have more cell rearrangements (Nislow & Morrill, 1988). This idea was based on measurements of cell division in gastrulation in which many more divisions were recorded in L. variegatus relative to the earlier measurements in L. pictus.

Here, we show in L. variegatus, that cells of the macromere lineage (fated to be the gut and non-skeletal mesoderm derivatives) contribute to invagination as a cluster that narrows along the anterior-posterior axis as elongation progresses. The cluster elongates with only modest changes in mediolateral position (convergence). Significant aspects of the gut lengthening are by cell migration and cell shape changes as cells migrate past one another to narrow the width of the cluster as they extend the length. Additionally, the cells of the archenteron increase by approximately one cell doubling that is biased along the plane of elongation. The consequence of that bias, however, contributes only modestly to archenteron elongation, while the main driving force appears to be the cell migrations.

Results

2.1 Cell migration is the main driving force of archenteron elongation

To examine the dynamics of gastrulation more closely, we turned to a cell recombination approach and produced mosaic embryos in which a single 32-cell stage macromere expressing membrane-GFP and histone2b-RFP replaced an equivalent macromere cell of an unlabeled host embryo at the 32-cell stage, or in some cases a single Veg2 cell replaced an unlabeled Veg2 cell at the 60-cell stage (Figure 1B). At the beginning of gastrulation, the macromere progeny from the 32-cell stage recombinants comprise 1/8 of the total cell mass of the future gut. The progeny of those transplants therefore began their morphogenesis as a cluster of approximately 16–20 cells (16 cells if the imaging began prior to the 10^th cleavage) at the vegetal plate before elongation (Figure 1C). Gastrulation was then followed by time lapse. To observe the behavior of the 20 cells shown in Figure 1C, we used a high-resolution DeltaVision deconvolution microscope for 9 hours (n=10 movies scored, Supplemental Figure 1). A series of images were taken every 3 minutes and compiled into time-lapse movies. By observing the cell movements over time, we saw that the endoderm progeny of the macromere, instead of moving perpendicular to the long axis of the archenteron, as in maximal convergence-extension, the cells actually migrated in a directed manner toward the anterior end of the embryo (Figure 2A). The initial mass of cells narrowed by one cell diameter as a column of cells, usually two cells wide, emerged from the fluorescent cell cluster. In numerical accounting, at 9^th cleavage there are 16 cells in the initial cluster, 8 Veg2 and 8 Veg1. Four of the Veg2 cells become non-skeletal mesoderm (NSM) and either transit gastrulation at the tip of the archenteron without converging or extending, or are released as pigment cells early in archenteron extension if the clone was on the dorsal side of the embryo (2–4 cells per cluster). The Veg2 cells divide at least once in gastrulation (see below) to provide roughly 8 Veg2 endoderm and 8 Veg2 mesoderm cells derived from each 32 cell-stage macromere. These cells move out of the initial mass as pairs of cells with the ~4–8 NSM cells leading. If cell divisions remain precisely equatorial alternating with meridian in direction (this ideal rarely occurs), Veg2 endoderm cells start approximately as a lateral line of 4 cells. There must be convergence of one cell diameter roughly at departure of invagination since the Veg2 mesoderm then endoderm cells tend to emerge from the cluster as pairs. The pairs of Veg2 endoderm cells then merge once more in transit, on average, to make a column of 1 × 8 Veg2 endoderm cells by the end of gastrulation (including the one cell division). From movies in which one cell was labeled at the 2-cell stage (data not shown), at the end of gastrulation there is a column of approximately 16–18 cells from the blastopore to the tip of the archenteron. Accounting for those 16 cells includes the 8 Veg2 endoderm cells, 4–6 NSM cells at the leading end (that do not appear to undergo significant convergence), and 2–4 Veg1 endoderm cells added late to the column in the archenteron extension (Suppl. Figure 1). Thus, in accounting for the number of diameters of convergent extension during gastrulation, in order to go from 4 to 1 cell width in the column, each Veg2 endoderm cell had to merge twice with its neighbor, once at departure near the blastopore and once in transit along the archenteron. The columns of labeled cells in most cases remained contiguous with their labeled neighbors throughout gastrulation, suggesting that there is very little additional convergence from outside the labeled patch.

Figure 2.

Elongation during gastrulation. (A) Displacement analysis using Imaris shows Veg2 endoderm-directed migration to be the main contributor to archenteron length. Veg2 cells were selectively tracked in either a transplanted Veg2 cell progeny (left) or a whole labeled gut (right) (B) Measurement of transplanted patches of cells at the beginning and end of gastrulation (from Suppl. Fig. 1) shows that each patch as a whole changes little in surface area, though there is a shift to an elongated patch of contiguous cells. (C) Ratios of convergence and extension confirm invagination to mostly take place parallel to the axis of elongation. Time-lapse movies were analyzed in Imaris software for the convergent and extension movements during gastrulation. A Measurements of the patches were made. For example, two parallel cells were labeled on the far ends of the Veg patch at the beginning of gastrulation and tracked for the duration of gastrulation (12–18hpf). Alternatively, two parallel cells that were adjacent to one another within the transplanted patch at the beginning were tracked throughout gastrulation (12–18hpf). It was found that with either way, the net convergence was no more than a cell’s width.

As the Veg2 endoderm tier elongated, cells within the patch migrated ahead of others, the cells changed shape from isotropic to anisotropic {Hardin, 1990 #1085} (Figure 1C, red brackets; Supplemental Movie 1), the cells changed shape from columnar to cuboidal allowing for increased surface area along the length of the archenteron, and some of the Veg1 cells contributed to the archenteron late in invagination adding to its completion (Figure 1C, yellow brackets). The remaining endoderm cells of the Veg1 tier moved into the archenteron after it had completed its full length {Logan, 1997 #1836}. Presumably, that final addition of Veg1 cells allowed the Veg2 endoderm cells to return to a compact, isotropic shape (Figure 1C, last panel). Strikingly, in each of the 10 time-lapse movies analyzed in detail, the Veg2 endoderm cells remained as neighbors with one another even as the column of cells narrowed to be 1–2 cells wide (Figure 2B). Thus, we conclude that two convergence events occur for each Veg2 endoderm cell, partially contributing to the lengthening of the archenteron. The movement of the cells in that convergence rather than being orthogonal, is angled toward the anterior direction as it occurs, and that directional migration appears to provide an important contribution to the lengthening process as cells slide past one another almost entirely in the direction of elongation and not in the mediolateral direction. That sliding narrows the column of cells from the initial cluster and extends the length of the archenteron. At the end of gastrulation, the archenteron is about 16–18 cells in length with an average of 8 nuclei around the perimeter of the gut, the fluorescent Veg2 endoderm cells contributing one of those nuclei. Convergence-extension in many vertebrates occurs with lateral movements of cells causing extension when the convergence forces cells to extend upon the intercalation. In contrast, in sea urchin gastrulation, while convergence occurs, it occurs as the cells are seen to move in the direction of elongation rather than perpendicular to it.

To reinforce this notion, we next wanted to score the relative positional changes within the Veg2 endoderm patch of cells during gastrulation. Since cells changed from columnar to cuboidal and since they also became anisotropic along the plane of elongation (Hardin, 1990), we wanted to determine just how much sliding of cells occurred to produce the file of cells observed at the end of elongation (Figure 2C). Accordingly, the position of two cells was followed over time and their relative displacements recorded. If we took 2 cells that were on opposing sides of the initial cluster of cells at the beginning of gastrulation, measured their change in position on the mediolateral axis (convergence) and their change in relative position on the anterior-posterior axis (extension), and calculated the ratio between the convergent distance and the extension distance, we found that overall convergence is no more than the width of a single nucleus (5μm), and extension also is modest (Figure 2C), confirming that in a Veg2 cluster there is initially only a convergence of about one cell diameter. The same can be said for two cells that were neighboring cells at the beginning of gastrulation. Their overall convergence was less than 1 cell’s width at 15 hr, and they moved apart by 18 hr, likely due to their being split by the mass of Veg2 NSM at the tip of the archenteron (Figure 2C). From following cells in a number of movies (Supplemental Figure 1), we conclude again that the cells within the contiguous patch of labeled Veg2 cells must migrate along the anterior-posterior direction while only slightly shifting position in the mediolateral direction, maximally a total of 2 cell diameters. This suggests again that an important contributing driving force for archenteron elongation is, in fact, anterior-posterior directed migratory displacement, and not extensive contributions from lateral intercalation as is the case in vertebrate convergent extension movements.

2.3 Cell divisions are prevalent throughout gastrulation

By labeling individual nuclei from our time-lapse movies using Imaris software, a transplanted patch of cells was tracked and the cell number determined over time. We observed that the number of cells increased linearly over time in both the Veg1 and Veg2 tiers so that at least one cell doubling occurred (Figure 3B). To see if cell divisions were present throughout the archenteron during gastrulation, we used a Click-iT EdU kit to label proliferating cells. Embryos were pulsed with EdU for one hour of gastrulation at every hour of archenteron elongation beginning at 11 hours and ending at 18 hours (Figure 3A). Cell proliferation occurred throughout gastrulation and was not restricted to a portion of the invagination time (Figure 3A). All cell lineages within the gut underwent cell divisions during the elongation. Cell counts revealed that during gastrulation the number of cells approximately doubled.

Figure 3.

Click-iT EdU assay confirms continuous proliferation throughout elongation and in all lineages of the gut. (A) Pulses of EdU were given every hour of archenteron elongation; embryos were then fixed, and stained with Hoechst to label all nuclei. Archenteron cells divided all through gastrulation without localized bias. (B) Imaris analysis shows cell division to be prevalent during gastrulation. Cells were labeled during a time-lapse movie in Imaris and proliferation was tracked throughout gastrulation (12–18hpf) within transplanted patches. Cell numbers throughout the time-lapse were graphed and a linear increase in the number of cells was observed. (C) Spindle orientation of mitotic gut cells contributes to archenteron elongation. Orientations of thirty mitotic spindles within the elongating archenteron were measured. We concluded from this that cell divisions are orientated (n=30, p<0.05); however, calculations of all cell divisions based on this bias indicate that the net contribution to elongation is about 15% of the total length (see text).

Earlier, based on inhibitor studies in L. pictus, it was concluded that cell division was not necessary for the archenteron to complete its invagination, although other investigations raised the issue that oriented cell divisions could influence invagination (Nislow & Morrill, 1988; Stephens et al., 1986). We repeated the inhibition using aphidocolin in L. variegatus at different concentrations and confirmed that at low concentrations there was a significant reduction in cell divisions yet the embryo completed archenteron invagination with little effect (Figure 4). Furthermore, we measured the surface area of the transplant pre- and post-gastrulation and determined there to be no significant growth in the size of the clone (Figure 2B). This was expected since at this time the embryo was not feeding, so it was unlikely that there would be a net increase in mass. We next measured directly whether cell divisions were oriented to facilitate elongation or if the convergence-extension plus cell migration were the sole contributors to gastrulation.

Figure 4.

Treatment with Aphidicolin resulted in an inhibition of cell proliferation. Embryos were treated with increasing doses of Aphidicolin and the numbers of cells within the archenteron were scored. The embryos were stained with an antibody to G-cadherin to label cell membranes and DAPI to label all nuclei. The embryos were still able to form a gut. Thus gastrulation reached completion despite the loss of the contribution of elongation-biased cell divisions.

2.4 Oriented cell divisions minimally contribute to archenteron extension

If the direction of cell divisions were oriented toward the direction of gastrulation, they could contribute to the total length of the archenteron or at least explain some of the elongation within a single Veg2 patch. That possibility prompted us to ask whether the direction of cell divisions during gut elongation were oriented relative to the axis of elongation. To quantify that possibility, we measured the angle of 30 divisions during anaphase at various times during multiple time-lapse movies of gastrulation to assess the orientation of the division. We found a significant bias towards anterior-posterior oriented divisions (n=30, p<0.05, Mann-Whitney U Test) (Figure 3C). 73% of the divisions occur at a 45° or greater angle perpendicular to the axis of elongation (Figure 3C). This suggests that the cell divisions bias their orientation along the anterior-posterior axis and contribute to elongation of the gut.

We next calculated the relative contribution of those oriented divisions to elongation of the gut. At the end of gastrulation, after further addition of Veg1 cells to the tube, the archenteron contains 18–20 cells along the A-P axis and 8–10 cells around a single perimeter depending on the location along the archenteron. By cell tracking in Imaris, we showed that there is approximately one cell doubling during gastrulation (Figure 3B). If the number of cells oriented perpendicular to the axis of elongation (therefore not contributing to augmented elongation) is subtracted from the number of cells oriented along the axis of elongation (above vs below 45°) there was a net of about 24 cells that oriented positively along the axis of orientation. Since each perimeter of the archenteron contained 8–10 cells, the 24 cells contributed positively to about 3 anterior-posterior oriented divisions per column of cells along the axis of elongation. The majority of those oriented divisions were between 75° and 90° relative to the plane perpendicular to the axis of elongation. Thus, maximally, the gain per cell division was 15μm per three cells per column (~5μm per cell given measured length vs height of sister cells). Since the archenteron length is about 100μm, oriented cell divisions therefore contribute no more than 15% of the total lengthening of the gut. This relatively small contribution is likely the reason that significant inhibitions of cell division did not prevent the archenteron from reaching its final length in the experiments (Figure 4) (Stephens et al., 1986). Apparently there is enough plasticity in the morphogenetic system to make up for the 15% loss due to absence of oriented cell divisions. After all, as invagination approaches completion, the mechanical pulling of the archenteron to the animal pole with filopodia contributes and is likely to help make up for the final lengthening in the absence of cell division (Hardin & McClay, 1990).

3. Discussion

Here, we report that invagination of the sea urchin archenteron involves sliding of endoderm cells relative to one another along the axis of elongation as the progeny of the Veg2 endoderm each converge a total of about 2 cell diameters. Those directed posterior-to-anterior movements plus the convergent-extension provide the major functional processes explaining archenteron elongation. These conclusions were the result of time-lapse analysis during gastrulation of the sea urchin, Lytechinus variegatus. Unlike gastrulation in amphibians, fish and chick where gastrulation includes a large number of perpendicular movements of cells relative to the axis of elongation, with intercalation that extends a column of cells, sea urchin cell Veg2 endoderm columns do not overtly move perpendicular to the axis of elongation. Instead, the cells of the invaginating archenteron move almost parallel to the axis of elongation such that the convergence results from cells of one column slipping past an adjacent column as both cells move in a posterior to anterior direction. That action leads to a narrowing of the perimeter and a lengthening of the archenteron as the columns of cells grow longer. As seen in the time lapse movies, the convergence extension also occurred with minimal to no mixing of labeled and unlabeled cells that were lateral to the labeled Veg2 patches. This meant the labeled cells tended to retain connections with their fluorescent neighbors throughout gastrulation even as some columns narrowed to one cell in width.

The cell tracking used in the present study also showed that during gastrulation there was roughly a doubling of the cells. Nevertheless, when the biased orientations of those divisions were considered, it was learned that they contribute to no more than about 15% of the total lengthening of the archenteron. Fifteen percent could be considered a substantial component of elongation, but because our study and earlier studies showed that if cell divisions were significantly inhibited, archenteron elongation could be completed anyway, we conclude that the system is plastic enough to complete gastrulation without that component.

3.2 Uncovering the transcriptional control of sea urchin gastrulation

While the movies show that the patches of endoderm cells extend in a directed manner along the plane of elongation during gastrulation and those movements of cells provide the means for invagination along with cell shape changes, extension of filopodia, and to a lesser extent cell divisions, the molecular mechanisms behind those morphogenetic events remain unresolved despite this and previous analyses of how gastrulation works (Ettensohn, 1985; Hardin & Cheng, 1986; Hardin & McClay, 1990). A number of research advances however have provided approaches that are now available to penetrate toward a far deeper mechanistic understanding of the gastrulation process, beginning with the control circuitry.

At the onset of gastrulation, an endomesoderm Gene Regulatory Network (GRN) model describes the distinct regulatory states of specification throughout the developing gut (Peter & Davidson, 2011). During cleavage, transcription factor sub-circuits emerge that drive specification, and as the beginning of gastrulation approaches transcription factors activated late in cleavage are likely important for activating a set of morpho-regulatory genes involved in the many cellular changes that are required for gut invagination. By perturbing those transcription factors proximal to the cell biological processes of gastrulation it will be possible to discover how individual components of gastrulation are controlled. This approach has already been used to identify circuits controlling the initial epithelial-mesenchymal transition of skeletogenic cells that occurs just prior to archenteron invagination (Saunders & McClay, 2014). Using assays that focus on specific processes of gastrulation movements, it will be possible to link distinct regulatory subcircuits to each of many distinct cell biological processes within gastrulation (i.e. invagination, cell migration, shape change, etc.).

Along with transcription factor involvement many signaling pathways have been identified as involved in aspects of gastrulation. Initially, non-canonical Wnt signaling, by means of Wnt signaling receptor, Frizzled5/8, is required for primary invagination (Croce, Duloquin, Lhomond, McClay, & Gache, 2006). Wnt/PCP small RhoGTPases, RhoA and Cdc42 are also seen to be cell biological players in the initial in-bending. RhoA overexpression induces precocious invagination and RhoA knockdowns are unable to invaginate upon blocking this small GTPase activity (Beane, Gross, & McClay, 2006). FGF signaling also has a role in timing of primary invagination: FGF translation-blocked morphants exhibit a severe delay in primary invagination (Röttinger et al., 2008). Wnt/β-catenin signaling initiates specification and later is necessary for the elongation of the primitive gut tube and, later, foregut and mid-gut composition and morphology (Croce & McClay, 2010; Croce et al., 2011; Peter & Davidson, 2011). At least 14 transcription factors are expressed to establish the Veg2 endoderm regulatory state, so at the time of gastrulation these transcription factors are primary candidates in learning how the upstream control of gut elongation works. FGF signaling, in addition to its timing role, apparently is necessary for invagination and compartmentalization of the archenteron (Röttinger et al., 2008). The parahox gene, lox, has been implicated in the regionalization of the gut tube into a functional, partitioned gut (Cole, Rizzo, Martinez, Fernandez-Serra, & Arnone, 2009). Lox is involved in hindgut specification and posterior sphincter formation, the boundary between the stomach and intestine. Lox morphants are seen to have a disorganized musculature in the foregut region and partial digestive function. Along with Parahox transcription factor, Cdx, circuitry of a GRN for gut regionalization is being uncovered (Annunziata & Arnone, 2014; Annunziata et al., 2014; Cole et al., 2009). What is necessary is to connect those known upstream regulators to the explicit cell behaviors they control.

The NSM, Veg2 endoderm and Veg1 endoderm specification states, at the beginning of gastrulation (12hpf), include approximately forty known transcription factors and six signaling pathways. The Veg2 NSM at the beginning of gastrulation is necessary for the initial in-bending of the gut tube at 12hpf. As described above, much of the gut tube elongation involves Veg2 endoderm, and Veg1 endoderm cells are added late in gastrulation. In each case advances in GRN assemblies will guide experiments to resolve how those GRN circuits contribute to control of gastrulation events.

As the circuits become better understood their value in learning about control of the cell biology of gastrulation will increase. Ultimately, a detailed understanding of the circuitry that controls gut morphogenesis will provide a powerful tool for dissection of the component processes of invagination.

The morphogenetic movements of gastrulation described in this paper can now be used to begin the perturbation analyses necessary to understand how gene regulatory network contributes to the proximal control of invagination. That analysis will enable a link between the cell biological activities of gastrulation and the transcriptional subcircuits responsible for controlling and coordinating distinct functions, including cytoskeletal dynamics, oriented cell divisions, and other cell biological events happening during archenteron elongation. Solutions to those problems will then allow connection of the GRN circuitry directly to the components of the cell biology that control morphogenesis. Those molecules that are transcriptionally controlled by the proximal GRN will then activate the remaining cell biological apparatus for that movement, assuming that most components of that apparatus are already constitutively present. In so doing, solutions of many of the remaining mechanistic questions will be accessed for the first time.

Methods

Adult animals and embryo culture

Adult Lytechinus variegatus were obtained from Reeftopia (Key West, FL, USA) or the Duke University Marine Lab (Beaufort, NC, USA). Gametes were obtained by injection of 0.5M KCl into the adult coelom. Embryos were cultured at 23°C in artificial seawater (ASW).

Live image acquisition

At 11hpf, embryos were de-ciliated in 2X hypertonic artificial seawater, mounted in 1X artificial seawater containing 10 μM p-methoxy-phenyl isoxazoline (Semenova et al., 2011) on a slide coated in 1% protamine sulfate and sealed with a combination of Vasoline, lanolin, and paraffin wax (also known as V.A.L.A.P.). Images were acquired using Coolsnap high-resolution CCD camera on the DeltaVision Elite with 40×/0.65–1.35 Oil UAPO40×0I3/340 DIC objective. Images were collected at 3-minute intervals beginning at 12hpf and ending at 18hpf or later and projected as movies using SoftWorx for DeltaVision. Movies were then analyzed using Imaris.

Analysis of oriented cell divisions

To measure cell divisions during timelapse movies, the angle of the division relative to the horizontal axis of the gut was measured in FIJI at anaphase. Statistical analyses were performed using GraphPad software. Statistical significance was measured using a Mann-Whitney U Test, and a p value < 0.05 was labeled as significant.

Drug treatments

Treatments with pharmacological drugs were carried out in 24-well plates at 23°C and drugs were diluted in artificial seawater. Aphidicolin was diluted to 0.2–10μg/mL and treated from 12–18hpf. Negative controls were carried out in the same manner as the experimental drug but using DMSO as the control treatment.

mRNA injections

mRNA for injection was transcribed in vivo using Ambion mMessage mMachine. Concentrations for mRNA injections: 300ng/μl Histone2B-GFP, 500 ng/μl Histone2B-RFP, 500 ng/μl membrane-RFP, and 300 ng/μl membrane-GFP. For injection, mRNAs were diluted in 20% glycerol in diethylpyrocarbonate-treated H2O.

Microsurgeries

Microsurgery was performed with fine glass needles and Narishige micromanipulators. Detailed methods of injections and transplants were followed as previously described (Logan, Miller, Ferkowicz, & McClay, 1999).

For Veg cell transplants: at 3 hpf, one Veg cell from an injected donor embryo was transplanted into an equivalent position in a 32-cell host embryo in the place of one discarded host Veg cell.

Click-iT EdU

Proliferating cells during gastrulation were assayed using the Click-iT EdU Imaging Kit (Invitrogen, C10085). The Click-iT EdU is a more sensitive and mild (to the sample) alternative to the classical BrdU assay for proliferating cells. EdU (5-ethynyl-2′-deoxyuridine) is a nucleoside analog of thymidine that is incorporated into DNA during synthesis. Detection of proliferation is achieved by a “click reaction”—a covalent reaction, catalyzed by Copper, between an azide in the Alex Fluor dye of the kit and an alkyne in the EdU.

A dilution of 20μM EdU solution was incubated at hour-long intervals during gastrulation (11–18hpf) before being washed out of the solution and fixed for detection. Fixed embryos were then incubated in a 3% BSA blocking step and a 0.5% Triton X-100 permeabilization step. EdU was detected by preparing a Click-iT reaction cocktail (as seen in manual) and incubating in the cocktail for 30 minutes protected from light. After incubation, embryos were washed with 1:2000 Hoechst to label all nuclei. EdU detection was visualized on a Zeiss 510 upright confocal microscope. Cell proliferation was counted in FIJI.

Supplementary Material

1. Supplemental Figure 1.

Ten movies were scored in detail for analysis in this manuscript. Still images are presented here at the beginning and end of each movie. Many movies scored were of clones that ended in the oral half of the archenteron and therefore contributed only to the archenteron. However, if the clone was found in the aboral half, many of the cells contribute to the pigment cells, which undergo an epithelial-mesenchymal transition early in invagination and rapidly move to the ectoderm through a mesenchymal-epithelial transition.

Highlights.

An analysis of sea urchin gastrulation shows that archenteron elongation occurs with minimal lateral convergence and a predominance of extension of clones of cells. Cell divisions during gastrulation approximately double the number of cells in the gut. That cell doubling contributes up to 15% of the total length of the gut due to divisions biased along the plane of elongation.